No title

Annual Review of Immunology Volume 7 1989 How One Thing has Led to Another George Klein and Eva Klein.Vol. 7: 1–34 Dec...

Author: The Annual Reviews

7 downloads 846 Views 39MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Annual Review of Immunology Volume 7 1989

How One Thing has Led to Another George Klein and Eva Klein.Vol. 7: 1–34

Decay-Accelerating Factor: Biochemistry, Molecular Biology, and Function D M Lublin, and J P Atkinson. Vol. 7: 35–58

Heterogeneity of Mast Cells and Phenotypic Change Between Subpopulations Y Kitamura. Vol. 7: 59–76

The Cellular Basis of T-Cell Memory J C Cerottini, and H R MacDonald. Vol. 7: 77–89

Microanatomy of Lymphoid Tissue During Humoral Immune Responses: Structure Function Relationships A K Szakal, M H Kosco, and J G Tew. Vol. 7: 91–109

Cells and Molecules that Regulate B Lymphopoiesis in Bone Marrow P W Kincade, G Lee, C E Pietrangeli, S I Hayashi, and J M Gimble. Vol. 7: 111–143

TH1 and TH2 Cells: Different Patterns of Lymphokine Secretion Lead to Different Functional Properties T R Mosmann, and R L Coffman. Vol. 7: 145–173

The Structure, Function, and Molecular Genetics of the gamma/delta T Cell Receptor D H Raulet. Vol. 7: 175–207

V-Region Connectivity in T Cell Repertoires P Pereira, A Bandeira, A Coutinho, M A Marcos, M Toribio, and C Martinez-A. Vol. 7: 209–249

The Immune System of Xenopus L D Pasquier, J Schwager, and M F Flajnik. Vol. 7: 251–275

Molecular Genetics of Chronic Granulomatous Disease S H Orkin. Vol. 7: 277–307

Cell Biology of Cytotoxic and Helper T Cell Functions: Immunofluorescence Microscopic Studies of Single Cells and Cell Couples A Kupfer, and S J Singer. Vol. 7: 309–337

The Leukocyte Common Antigen Family M L Thomas. Vol. 7: 339–369

T Cell Receptors in Murine Autoimmune Diseases H Acha-Orbea, L Steinman, and H O McDevitt. Vol. 7: 371–405

Manipulation of T-Cell Responses with Monoclonal Antibodies H Waldmann. Vol. 7: 407–444

Clonal Expansion Versus Functional Clonal Inactivation: A Costimulatory Signalling Pathway Determines the Outcome of T Cell Antigen Receptor Occupancy D L Mueller, M K Jenkins, and R H Schwartz. Vol. 7: 445–480

Immunogenetics of Human Cell Surface Differentiation W J Rettig, and L J Old. Vol. 7: 481–511

Probing the Human B-Cell Repertoire with EBV: Polyreactive Antibodies and CD5+ B Lymphocytes P Casali, and A L Notkins. Vol. 7: 513–535

Stable Expression and Somatic Hypermutation of Antibody V Regions in B-Cell Developmental Pathways C Kocks, and K Rajewsky. Vol. 7: 537–559

T-Cell Responses and Immunity to Experimental Infection with Leishmania Major I Muller, T Pedrazzini, J P Farrell, and J Louis. Vol. 7: 561–578

The Biologic Roles of CD2, CD4, and CD8 in T-Cell Activation B E Bierer, B P Sleckman, S E Ratnofsky, and S J Burakoff. Vol. 7: 579–599

Antigen Recognition by Class I-Restricted T Lymphocytes A Townsend, and H Bodmer. Vol. 7: 601–624

The Biology of Cachectin/TNF -- A Primary Mediator of the Host Response B Beutler, and A Cerami. Vol. 7: 625–655

The T-Cell Receptor Repertoire and Autoimmune Diseases V Kumar, D H Kono, J L Urban, and L Hood. Vol. 7: 657–682

T-Cell Recognition of Minor Lymphocyte Stimulating (MLS) Gene Products R Abe, and R J Hodes. Vol. 7: 683–708

Annu. Rev. Immunol. 1989.7:1-34. Downloaded from arjournals.annualreviews.org by HINARI on 08/29/07. For personal use only.

Annual Reviews

Annual Reviews Ann. Re~. lmmunol. 1989. 7: I 33 Copyright © 1989 by Annual Rev&wsInc. All rights reserved


HOW ONE THING HAS LED TO ANOTHER George Klein

and Eva Klein

Department of TumorBiology, Karolinska Institutet, S-104 01 Stockholm, Sweden, and Lautenberg Center for General and Tumor Immunology, Hadassah Medical School, Jerusalem, Israel GEORGE

KLEIN

WRITES:

Dawn This story starts on the 10th of January, 1945, when I emerged from a cellar on the outskirts of Budapest where I had been hiding, with false papers, during the last weeks of the Germanoccupation. With a totally newfeeling about the sunshine that was floating over the snow, the ruined houses, the dead and frozen soldiers, civilians, and horses, I suddenly realized, with a mixture of surprise, guilt, and delight, that I had survived in spite of an 80%chance that I wouldend my19 years in the gas chambers or in a military slave labor camp. After a few quick walks in the newly liberated area of the still besieged capital, I decided that it was time to start mymedical studies, already delayed by almost two years. During the first year after mygraduation from middle school, it was impossible for a Jewish boy to enter medical school. After the Germanoccupation nothing mattered except survival. Wewere free at last, but it was a complicated freedom. After a few more days, the Eastern side of the city, Pest, was all in Russian hands. I moved around relatively freely but I was caught twice, like other young menwho were automatically regarded as disguised soldiers. In comparison with my earlier escape from a Nazi labor camp, it was an easy matter to run away from the improvised, loosely organized Russian patrols. It was a wise move. Several friends of mine whowent out to get a loaf of bread returned years later from Russia. As soon as the streets were open, I walked to the University to see 1 0732--0582/89/04104)001 $02.00

Annual Reviews


2

KLEIN & KLEIN

whether it would open its doors for me now. I found deserted buildings, broken windows, and dead soldiers. Together with a friend we therefore decided that we should try to reach Szeged. The journey of less than 300 kmtook more than five days. Wewalked long stretches, hitched on horsedrawn carriages and every other vehicle that we could get on, including a Russian military truck. Wearrived in Szeged on February 4. It was a cold and beautiful morning. The city was intact, and we were admitted to the University on the same day. It was a strange place. All the professors had fled to the West. An assistant professor of forensic medicine with a Christlike head and very sad eyes was teaching anatomy, pathology, and forensic medicine all by himself. Students kept arriving from all former theaters of war, labor camps, and illegal hiding. Cadavers were abundant. The large dissection hall of the AnatomyDepartment was crowded. The smell of formalin, the half dissected or fully prepared body parts, and even the continually tipsy attendant appeared to meas parts of a magic, enchanting landscape, a previously forbidden paradise that was nowall mine. Twoyears passed as a single wave of febrile activity. I finished three terms during three months in Szeged and returned to Budapest when the university reopened there. I wanted to start research work, but the departments were still paralyzed. They had no resources and the routine workconsumed the energyof all staff. Still, I got a first decisive inspiration from the professor of histology, Tivadar Huzella, one of the few internationally knownscientists in Hungaryand als O one of the few true liberals amongthe medical professors of his generation. In spite of his consistent anti-Fascist stance, and his strong opposition to any form of discrimination during the war, he became a suspected person in the eyes of the new rulers. His uncompromisingindividualism and his democratic value system invited the enmity of the political opportunists whowanted to see a more compromisingperson in his position. His arch-enemy, the professor of anatomy, a political opportunist and a scientific nonentity who had resented Huzella’s international fame for manyyears, delivered a list of accusations against him to the "people’s court." The sympathies of all the students were on Huzella’s side. The crucial trial, where all the absurd accusations--exemplified by the charge that Huzella ate eggs ordered for tissue culture--were readily dismissed, ended in tragedy when the presiding lay judge asked whether Huzella still believed a sentence he wrote during the war. Huzella had stated (an act of great courage at the time) that Hitler, Stalin, and Salazar were equally abominable dictators. If he would have been willing to exempt Stalin and admit his "mistake," he would have been cleared. But he stuck to his words and was summarily dismissed. He died a few years later. Today he has been "rehabilitated."

Annual Reviews


HOWONE THINGHAS LED TO ANOTHER 3 His homeand laboratory are kept as a public memorial. They also house the leading immunological laboratory of Hungary. Huzella had an exceptional ability to convey his own deep interest in biology to his students. He was convinced that the time had come when biology could be converted from "metaphysical speculation" into a natural science with precision and dignity similar to those of chemistry and physics. He believed that the biology of the interstitial space would turn into detailed biochemistry in a few decades but that the cell interior would remain a black box during the rest of the century. Before blaming him for a lack of foresight, we must realize that most biologists of the time were unwilling to accept his "optimistic" view even about the connective tissue. I learned some tissue culture, but my practical experience remained rudimentary, and I compensatedonly slightly by avid reading in the still quite deficient library. After Huzella’s removal, I realized that I could not learn more in the nowlargely nonfunctional department, and so I movedto Pathology. After a few weeksI found myself totally immersedin autopsies. There was a great abundance of cadavers here and very few pathologists. The large postwar classes of medical students had to be taught quickly. I greatly enjoyed the double task of teaching the little I knewand trying to explain to the rushed and often very nervous clinicians what their patients had died of. In the early spring of 1947, one of "my"students approached me after an autopsy. He said something appreciative about my demonstration and asked whether I would be interested to visit Swedenwith a student group. I was amused by his naivet6. Whowould not like to visit Sweden? But were we not all aware of the fact that foreign travel was the exclusive privilege of important functionaries and people with much moneyand many good connections? He replied that he was currently organizing a trip for students and that he would include me. Hungarystill had an elected coalition government at this time. It was possible to get a passport, but this was not sufficient to leave the country. A special exit permit had to be issued by the "Allied" forces, i.e. the Soviet Army.It was very difficult to get this permit, and it was nearly impossible to obtain foreign currency. I mailed mypapers to my student who was interested in Sweden and totally forgot about our conversation. Decisive

Summer

In June 1947 myboss, Professor Balr, told me that I would be responsible for the autopsies during the comingmonth, virtually alone. I was happy, proud and frightened. I was not yet 22, far from being an MD,but the night’s sleep of a professor in surgery could depend on what I was going

Annual Reviews


4

KLEIN’ & KLEIN

to find. The combined feeling of responsibility and awe turned every autopsy into an exciting detective story. During myminimal "spare hours" I also started myfirst attempts to do someexperiments. I was sitting in a corner of the laboratory with a small water bath and a stalagrnometer, trying to follow a lead that had been opened up by mychief. The most important messengers of my future destiny appeared in the shape of two house painters in the middle of July. They had been ordered to repaint the laboratories. I was chased from room to room with my water bath, but I refused to give up. Finally, I was squeezed into a small corner in a tiny windowless alcove that I refused to leave. The painters complained to Professor Balr. With an irritated "you can take two weeks vacation for once" he ordered meto leave myparadise. A senior colleague was to take care of the autopsies. I was angry and disappointed. Whatwas I to do during two whole weeks? By coincidence I learned that some fellow students, two couples from the Pharmacology Department, were planning to spend the forthcoming weekat the Lake Balaton. I was also told that they had invited someother friends and that I was welcometo join them. Wewere allowed to use the terrace of a bombedsummerhouse and were going to sleep on mattresses, spread out on the terrace. It was quite warmduring the first week in August, and we would have a roof over our head. After considerable hesitation, I decided to join them, but I felt ambivalent and uninterested. The place was unexpectedly pleasant and myfellow students were much nicer in private life than at the University. On the second day, the two other boys went down to the train to meet another student from the Pharmacology Department, who was to join us. I did not know who it was, and since the Hungarian language does not dist.inguish between he and she, I did not even knownwhether we were expecting a boy or a girl. After a while I saw them walking up the hill with the new guest: a dark girl with a strange, breathtaking beauty. I perceived a most unusual combination of hilarity and sorrow, seriousness and play in her eyes. It was Eva, myfuture wife and colleague until this day. I had seen her before at the university, but myobsessive preoccupation with work prevented me from giving her or any other girl muchattention. Still, I could remembervery well how I met her the first time. On the second day of mymedical studies in Szeged, I was standing in the Dean’s office, to get mypapers. She entered, dressed in a skiing outfit, having arrived in the city after a long and adventurous trip from Budapest, like myown. She asked me howto get papers. I saw that she was very beautiful. Her direct way of talking to a strange boy--very unusual for a girl in Hungary at the time--struck me as original and sympathetic. During the forthcoming weeks I saw her at some lectures, but then she disappeared.

Annual Reviews


HOWONE THING HAS LED TO ANOTHER 5 Later I saw her name on the posters of the city theater. She was playing small roles in Pirandello and Moli&eplays. Halfa year later I saw her again in Budapest. She had returned to medical studies and came sometimes to my autopsy demonstrations. I knew that she belonged to the same group of students in the Pharmacology Department as mymarried friends and temporary hosts. Their "gang" treated me with friendly tolerance, and even with a trace of respect for my"knowledge"--in spite of their "objections" to the "dead morphology"that pathology represented in their eyes. I respected their intelligence and their dynamicexperimentation and could therefore forgive their blatant ignorance of pathology and clinical medicine. But this time everything was different. There was one table but only three intact chairs in the ruined villa, and we were six. Wehad to place a board on each chair to hold two. Eva and I were placed on the same board and had to coordinate our movementsto prevent each other from falling down. This trivial problem initiated a contact that metamorphosedafter only a few hours into a passion that conquered myentire consciousness with the force of an elementary power. All other interests and problems vanished as if they had never existed. I spent eight days at the lake, intoxicated, overwhelmed,cut-off from all earlier reality. An unexpected telegram arrived on the seventh day. Everything was settled for the trip to Sweden!Myformer pathology student or, as we were soon to call him, Our Leader, had succeeded against all odds. He had pursued his plan with obstinate ingenuity and obtained all the exit permits for a group of seventeen students selected by himself with the arbitrariness of a sovereign. Wecamefrom different faculties and were to visit Stockholm and Gothenburgas the guests of the Jewish Student Club there, in order to see a country that was saved from the war. NowI did not have the slightest wish to go. I felt very bitter about having to leave the person who had becomemore important than anything else in mylife so far. The week at the Balaton appeared as an eternity; everything before was unreal. But vague feelings 6f responsibility and premonition commandedme to go. I left at dawn on a Sunday morning. Eva told me later that she heard the train whistle while half asleep and thought that a beautiful summerepisode was nowover. She did not believe that I would ever come back from Swedenor that she would see me again. Cell Biology

1947

The first International Congress of Cell Biology had just terminated when I arrived in Stockholm. I was told that Torbjrrn Caspersson was one of the most important figures at the Congress. His recent development of ultraviolet microspectrophotometryon fixed cells created muchattention.

Annual Reviews


6

KLEIN & KLEIN

The method was based on his doctoral thesis, written in 1936 in German and largely unavailable to English speaking readers during the war years. It was the first major attempt to combine morphologyand cytochemistry. Cells were photographed in monochromatic UVlight under standardized conditions. A semiquantitative method was developed to map the localization of nucleic acids and proteins in different cell types. Jack Schultz, one ofJ. H. Morgan’slast disciples, was the first Americangeneticist who saw the potentialities of the new approach. He traveled to Stockholm to work with Caspersson shortly before the outbreak of the war. He brought genetic thinking to the biophysically oriented group. His studies with Caspersson on the banding patterns of polytenic insect salivary gland chromosomesgave the first information about the distribution of nucleic acids and chromosomal proteins and set the conceptual basis for the development of the chromosome banding technique by Caspersson and Zechthree decades later. The chemistry of the genetic material was still unknownat the time of the Cell Biology Congress in Stockholm. Most biologists believed that only proteins could provide the necessary diversity. Nucleic acids were considered as repetitive, boring molecules. Levene and Bass pronounced the death sentence on the coding capacity of the nucleic acids already in the 1930s. The mistaken analogy between the "4-letter alphabet" of the nucleic acids and the phonetic alphabet served as a roadblock: howcould one build a language from four letters? Caspersson’s semiquantitative measurementsof nucleic acids and proteins in different cell organelles led him to conclude that there was a definite relationship betweennucleic acid and protein synthesis and that the former might actually govern the latter. This visionary insight was widely disbelieved, however. The idea that nucleic acids might carry genetic information that could be translated into proteins was totally foreign, even to Caspersson. The fundamental discovery of Avery, McLeod, and McCarthy on DNA-mediated transformation in Pneumococcus, published in 1944, was widely ignored or discarded as an artefact. The Cell Research Department of Karolinska Institute had just moved to the newly built campuson the northern edge of the city; there I was to spendall myscientific years, up to the present day. I visited it first in the middle of August, 1947, the peak of the vacation season and soon after the Congress participants had left town. Membersof the Department who happened to be in town were frantically trying to get settled in the new building. As I mademyentry, tall, blond, 37-year-old Torbjorn Caspersson was lying under a large instrument in a blue overall, trying to fix the wires. I thought that he was an electrician or a technical assistant. His identity was not revealed to me and I was not introduced to him. After I had

Annual Reviews


HOWONE THING HAS LED TO ANOTHER 7 learned the difficult art of protecting him from uninvited visitors a few years later, I could understand the reasons. In 1947, I was desolate when I had learned the next day that he had left fo~ the USA.Only after a long series of complications did I get in touch with him, several weeks later. But my first conversation with him was decisive. Thanks to the rudimentary and largely theoretical knowledgeof tissue culture, acquired in the Huzella laboratory two years earlier, I got the best-paid job of mylife (if the importanceof the salary is considered). I was employedas a junior research assistant, on 500 SwCrs (about US$100) per month. I still rememberthe mixture of ecstatic happiness and enormousanxiety. Mysituation appeared totally hopeless. I knew virtually nothing. I was halfway through mymedical studies, still far removedfrom an MD.I was desperately in love with a girl whomI had only knownduring a summer vacation of eight days and who was on the other side of an increasingly forbidding political barrier. I did not knowa word of Swedish. Still, I was firmly decided to resist the more comfortable possibility of continuing my studies in Hungary. Mymotivation was reinforced by a series of articles that kept appearing in the major Swedish daily, Da#ens Nyheter, translated for me by my temporary host. The Prime Minister of Hungary, Ferenc Nagy (not to be confused with Imre Nagy) of the Smallholder’s Party has just fled to the West, and he gave a series of interviews to the Swedishpaper. In contrast to the rosy optimism that prevailed amongmy friends in Budapest who hoped that Hungary would become a democratic country, Nagy’s statements had an ominous ring. He said that the influence of the Communist Party was increasing continuously behind the scenes. The Stalinist party leader, Rfikosi, was acting under the protection of the Russian forces. The politicians of the other parties were frightened. Several of their leading representatives were arrested on false charges and deported to unknown destinations. Those whoremained were increasingly inclined to give in. The police were infiltrated by party members. Nagy did not have the slightest doubt that a Communisttakeover was imminent. Similar signals reached me indirectly from one of myteenage idols, Nobel Prize winning biochemist Albert Szent-Gy6rgyi. He was still holding manyhigh posts in Hungary at the time, but he had told his nephew, who was a friend of mine, that the days of freedom were numbered. If you were young and wantedto have a future in science, you should get your degree as soon as possible and leave the country. Farewell,

My Native

Land

In mid-September,I decided to go back to Budapest and try to get out for good. Mymost important acquisition was safely tucked awayin mybreast

Annual Reviews


8

KLEIN & KLEIN

pocket: a re-entry visa to Sweden and a labor permit for continued work in Caspersson’s department. Mypassport was still valid for a few months. The reunion with Eva confirmed what we both knew already: we wanted to live and work together. The day after myarrival, someof our friends gathered at myhome to hear the latest news from the "great world." I told them about Nagy’s report and the iron curtain that was about to descend over Hungary. The reaction was mixed. Those who were already preparing to leave believed me. Others wanted to stay and hoped that my report was exaggerated. One of them--still a good friend today--declared that I was probably right, and for that reason, he was going to break all further contact with me. This was his country, Hungarianwas his language, his historical roots were here. I should leave, if I felt so inclined, but he had to stay and do the best he could. Today he is the foremost medical historian of Hungary. I had none of his historical perspectives. I had only one goal, to get married and leave the country. But howto get married? It had to be in secret, because nobody would understand why two 22-year-old students who had known each other for only a short time and had no income would want to get married. And how could myfuture wife join me? She had no passport and the difficulties in getting one were now increasing day by day. Weagreed that I would go back to Stockholm before my own passport expired and try to obtain letters of invitation for Eva that could help her to get a passport. The last weekday before my trip was a Friday. Eva and I met outside the pharmacological institute to go to the day’s lecture. I suggested that we should go to the prefecture instead and ask howone gets married. We got a list of the manydocumentsyou needed. It looked hopeless. It would take months to get them. I suggested that we ask for the first document, a certificate to showthat we had no police records. Wewent to the police station. "It takes at least three weeks." SuddenlyI acted on impulse. I had always heard others tell of such things but I myself had neither seen nor done it. I pulled a fairly modest bill out of mypocket and put it in the policeman’s hand. "Pardon me, how muchtime was it, you said? .... I’ll go and get it at once," he answered. It was now11 Ara. Wecontinued from office to office. Everywherethe same answer: one week, four weeks, six weeks. A little bill in the hand--the certificate was completed within a few minutes. I was amazed to find that the shyness I usually exhibited before persons of authority vanished completely. I learned a lesson about the importance of motivation and the unsuspected possibilities it mayopen to surpass one’s limitations.

Annual Reviews


HOWONE THING HAS LED TO ANOTHER

9

By 3 PMonly one document was missing: a medical certificate that neither of us had venereal disease. The tests would take several weeks. What to do now? Wewent to a slightly older colleague who had recently finished his medical studies. He had just started his first assignment in the Children’s Hospital. Wetold him, in the strictest confidence, about our situation. He had a good laugh and wrote the certificate on the hospital stationary. By 4 PMwe were at the prefecture again. Wehad all the papers and wanted to get married that second. Twoother friends, swornto the highest secrecy, camealong as witnesses to the wedding.The official had just finished the day’s work and had taken off the broad Hungarian tricolor from his corpulent chest when we rushed in. Weheard him telling his wife on the phone that he was on his way homefor dinner. Marry us at this time of day? Not a chance! Comeback on Monday! I started to appeal to his humanfeelings. I had to leave the country on Sunday. Howcould 1 leave myyoung bride alone if we didn’t get married? He was noticeably irritated and doubted that we had all the papers. While leafing through the documents,he caught sight of the doctor’s certificate that had been drawnup at the Children’s Hospital. He laughed until tears ran downhis cheeks. This was the funniest thing he had seen during his whole time in service. Nowhe was in splendid spirits. The flag resumedits place on the large body. Wepromised to love one another til death us did part. Afterwards we ate our wedding dinner on the hall bench together with our witnesses. There was only one dish: mymother’s carefully packed goose liver sandwiches. In the evening we went back to our parents’ homes where no one suspected anything. That SundayI returned alone to Stockholm. Eva joined me, after many complications, in March 1948, after the Iron Curtain had already descended over the country. GEORGE

AND EVA

The Genetics

WRITE:

Congress

In August 1948, several months after we were happily settled in our rented room and Eva had also started to work in Caspersson’s department, the International Congress of Genetics took place in Stockholm. The presidential address of J. H. Muller was a scathing denunciation of the abuse of genetics in the Soviet Union. The scientific world was still largely unawareof the fact that the "theories" of a charlatan, Lysenko, had been declared "official" by the Central Committee of the CommunistParty, meaning that it became essentially illegal to do any scientific work in

Annual Reviews


10

KLEIN & KLEIN

genetics. Mullerhimselfhadbeenthe first to introduceDrosophilagenetics into Russia, and he wasstill a member of the Soviet Academy of Sciences at this time. He called Lysenko"a paranoic and half educated young demagoguewhohad done some work in raising plants but whowas in fact ignorant of scientific principles andincapableof understandingthem." He added that many of the outstanding Russian geneticists had disappeared,and somehad lost their lives in unexplainedways.His speech ended with his resignation from the Soviet Academy.The reaction from Moscow camethe day after. Theyrefused to accept his resignation and expelled him. Onthe last day of the congress, the Bulgarian delegate asked to make a statement at the concludingplenary session. Speakingin the nameof the delegates from Bulgaria, Roumania,Poland, and Czechoslovakiahe delivered a strong protest against Muller’s introductory speechthat was "ill-suited to favor international understanding."His protest wastaken to the protocol. After the session wasclosed, the representative of Hungarycameto us. Hedid not understand English well, and wehad previously helped him. Hewantedto knowwhat the Bulgarian delegate said. Whenhe heard our interpretation he becameextremelyupset. It wastypical for the Slavic delegates to leave out the Hungarians!Hehad to join the protest, he had to think of his family! Howcould he return without havingsigned it! But he wasout of luck, the congresswasover, nothing could be donefor him. His panic showedus howthe fear imposedby Stalinism had descendedon the country wehad left only a few monthsearlier. It wasalso a reminder of the eternal strife amongthe nations that have risen from the ruins of the Hapsburgmonarchy. Duringthe Congresswelearned about the startling progress in microbial genetics. Bacteriologyhad been the last citadel of Lamarckism.At this time, whenproteins wereregardedas the vehicles of genetic information, whennotions about a bacterial nucleuswereregardedwith great suspicion, inducedenzymeadaptations and drug resistance werewidelyattributed to the inheritance of acquired characteristics. But the rapidly growingevidenceof clonal variation and Darwinianselection wasnowdefinitely gaining ground,integrating microbiologywith the rest of biology. Ourknowledgeabout cancer was rudimentary,but westarted nevertheless wondering whether the population dynamicsof microorganismsand the phenomenon Of cancer might share some commondenominator(s). Cancer cells are resistant against growth control of the organism. Canthey be compared to drug resistant microorganisms? Couldcancers also arise by a series of mutations?Theseseeminglypuerile notions wereto play an importantrole for our worklater on.

Annual Reviews HOW ONE THING HAS LED TO ANOTHER


The Cell

Research

11

Department

Back at the laboratory, we found ourselves in an exciting environment but facing another impossible situation. Wewere still medical students in midcourse. Wewere struggling hard to get into a Swedish medical school and finish our studies. At first, this looked impossible but eventually we succeeded, one by one, taking turns between the school and the lab work. Worsethan that, the project given us turned out to be quite unmanageable. Caspersson’s methodologywas based on the absorption of monochromatic ultraviolet light in fixed cells. Shortly before our arrival, it was heavily criticized by Barry Commonerand other biophysicists. They suggested that the loss of UVlight registered by Caspersson’s optical system was not due to absorption but to light scattering from the denatured proteins. Due to this artefact, part of the nonabsorbed light would never reach the objective, leading to false conclusions about the localization of nucleic acids and proteins. Their distribution in living cells could be totally different from the pattern suggested by Caspersson’s measurements. Our task was to measure light absorption in living cells. But this was more easily said than done. Tissue culturing of the times followed the dogmaslaid downby Alexis Carrell. The plasma clot and the embryonic extract were regarded as essential substrates. Nobodyin his fight mind wouldhave thought of culturing cells directly on glass, even less on quartz slides. The plasma clot was not transparent to UV. It turned out that mysudden and unexpected employmentafter my first conversation with Caspersson was due to the fact that I had someexperience of growing cells on collagen, Huzella’s favorite method. Collagen is poor in aromatic amino acids, and it was therefore expected to provide less of a problem for UVmicroscopy. Westruggled frantically to obtain someresults. Wehad no experience, no assistance, and virtually no apparatus. The large UV-equipmentwas not suitable. The cells were killed by UVlong before we could take a picture. Washingand sterilization of the glassware, preparation of the embryonicextract, and most difficult of all, collecting plasma from the carotid of our single rooster with a primitive paraffin oil canule were a neverending struggle. At last, we managedto take a few pictures before the cells died, but we were still far from the number of monochromatic exposures needed for a spectrum. Our future looked dim. Salvation came in the form of two unexpectedevents. The first was a lecture given by Hans Lettr6 of Heidelberg on the Ehrlich ascites tumor, which he used for biochemical studies. Eva immediately pointed out that we might obtain homogenouspopulations of living cells from the peritoneal cavity of the mouse, without having to do any tissue culture at all! Our first attempt to

Annual Reviews


12

KLEIN

& KLEIN

propagate the tumor, kindly sent by Lettr6 in the form of a single mouse, ended in total failure, however.The first of our inoculated mice developed a nice round belly that turned out to carry a lovely litter of eight, instead of the expected tumor cells. But the second mousedeveloped a tumor, and we were in business. But as we were getting ready for the UVpictures, a paper was published by Brumberg & Larionov in the USSR.They used a new, reflecting optical system that avoided the killing of the cells during UV-exposure. They had done all the experiments that we had planned and showed that Caspersson was right and the critiques were wrong. UVmicroscopy did measure nucleic acids, and they were localized exactly in the organelles where Caspersson had found them in fixed cells. Our project had becomeobsolete overnight. What were we going to do? Weexpected the worst, but Caspersson suggested that we continue to work with ascites tumors and try to formulate our ownproject. The early experience of the Genetics Congress came to our rescue. Whywas the Ehrlich ascites carcinoma unique? Whycould other tumors not be propagated in this freely dissociated "fluid" form? Did most tumor cells require a solid substrate and/or the microenvironmentof a solid tissue? Wehad some ideas about how to start looking at this, but our mouse and tumor facilities were very limited. Inbred mice were totally unknownin Sweden at this point. Salvation came again unexpectedly. In the summer of 1950 we participated in the International Cancer Congress in Paris. The week was occupied by frantic and hopeless efforts to get acquainted with the entire cancer field, interspersed with meetings with old friends who had left Hungaryafter us. At the end of the week, we felt definitely reassured about two earlier, disparate but equally important, conclusions: (i) It was very fortunate that we had left Hungaryin time, for the Stalinistic system nowhad a firm grip, and (ii) tumor cells could be definitely regarded genetically heterogeneous populations with extensive subclonal variation. Wealso felt proud as contributors to the Congress. George lectured in broken and very slow English in the 32nd parallel section, the late afternoon of the last day, with six persons in the audience (1). It turned out later, however, that this was an extremely important event. One of Otto Warburg’sassistants had been in the audience and brought homethe great news that the mouseascites tumor cell is an equally good tool for largescale experimentation on such relatively homogeneouscell populations as the famous Chlorella algae of the Great Master. Warburg immediately requested the cells and was later very helpful in supporting us. In a letter written in 1956, he stated that we had madea very important contribution, because we had sent him the cells that made it possible for him to solve the cancer problem. All bureaucrats were deeply impressed!

Annual Reviews


HOWONE THING HAS LED TO ANOTHER 13 During the Congress week, we also enjoyed frantic, colorful, decadent, exhilarating, and slightly putrescent Paris. Sitting at a cafe on Boulevard St. Michel the evening of Bastille Day we exclaimed: "Howwonderful! howcrazy! how can one possibly live in a sterile country like Sweden?" Rattling home a week later in a third-class wagon across strike-tom Belgium we said: "Howmarvelous that we can return to quiet, boring, aseptic, polite Swedenwith its thousands of lakes, endless forests, and luminescent nights!" Wehad hardly opened the door to our rented room in Stockholm when I saw the new miracle: an express letter in Caspersson’s ownhar~dwriting. "Get in touch with me immediately on arrival." One of the main private research foundations in Sweden, established in the memory of Knut and Alice Wallenberg, had asked Caspersson to choose two young men for an urgent mission. They were to go to the United States for several months and report about recent advances in cancer research. Caspersson chose one of myolder colleagues and myself. But Eva had to stay home--there was not enough money. Mycolleague was to travel around from center to center. Mytask was to work with Jack Schultz at the Institute for Cancer Research in Fox Chase, Philadelphia, on myownproject, and to make short visits to some of the major centers in the neighborhood. Wereceived the newswith a mixture of joy and sorrow. It was a fantastic opportunity. But the sorrow and anxiety of being separated again from myyoung wife, and for quite some time, were further aggravated by the sudden outbreak of the Korean war with its forebodings of a possible world war. Weshared our vision of an approaching Apocalypse with most other survivors of the Second WorldWarand the Holocaust. Our officially stateless status addedfuel to the nightmares. Still, I knewI had to go. The Statue

of Liberty

The Institute for Cancer Research has developed from a small private research group at LankenauHospital, due to the great foresight of Stanley Reimann.WhenI got there, they had just finished a major expansion from a small group of scientists to a large research center in a magnificent new building. Several prominent biologists had joined the laboratory. The leitmotiv was to look at the cancer problem from the biological point of view. Myownboss was Jack Schultz, a lively little manin his sixties. Jack exuded boundless curiosity, joy of life, and great humanwarmth. He received me as if I were his long lost, finally recovered son. During my stay he often gave me a lift from myrented room to the laboratory. Most of what I knowabout genetics can be traced to those car tides. But the trip was not over when we arrived. Jack’s office was at the far end of a

Annual Reviews


14

KLEIN & KLEIN

long corridor. Walking down the hallway he would stick his head into every lab and stop and talk with people on the way. He asked them about everything, the health of their kids, mother’s broken leg, the weekend excursion, but first and foremost about the latest experiment. The people brightened visibly when they saw him and were always ready to stop for a chat or to ask him to come in and look into the microscope, at a bacterial plate or at a Drosophilaprogeny. Jack looked, listened, discussed, interpreted, proposed new experiments. Under his arm he carried his briefcase with all the papers he planned to finish during the day. Sometimes half a day passed before we arrived at his office where his secretary waited in despair! I visited Jack’s office 25 years later, long after his death. It has been refurnished as a conference room. It bears Jack’s name. A silver plate on the wall reminds us of the unselfish inspiration he provided to everybody in his environment. Jack succeeded in communicating the notion that biology is the most exciting science. He told me about whole worlds I had never heard about. Barbara McClintock’s discovery of transposons in maize was one of them. Jack was one amongthe dozen or even fewer geneticists who understood what McClintock was talking about. He already knew, 10-15 years before most others, that her findings were going to revolutionize biology. Jack’s corridor was a wonderland for me at the age of 25. Briggs and King experimented with nuclear transplantation to enucleated frog’s eggs. The question was whether the nuclei remained totipotent during differentiation. This was also relevant for cancer research. Could cancer cells contain a totipotent nucleus? The question was answered several decades later, by Beatrice Mintz at the sameinstitute--at least answeredso far as diploid teratoma cells were concerned. A cartoon appeared on the wall of the same corridor where Jack and I so often walked in the morning. Two mice were talking to each other. One of them said: "Myfather was a cancer, what does your father do for a living?" The Mintz experiment is still unique in showing that at least some cancers can develop by epigenetic changes. The majority are no doubt due to changes at the DNAlevel, however. Myother important master at the ICR in Philadelphia was the mouse geneticist Theodore Hauschka. Through him I became acquainted with the inbred mouse. He had also taken a direct interest in myexperiments. He gave me myown room in the perfectly organized mouse colony where I was in full operation for days on end. I comparedthe ability of different solid tumors to grow in the fluid form in the abdominal cavity. Whenthey were reluctant to behave according to mywishes, I tried to select variants that would. At the same time I began to wonder whether the "histo-

Annual Reviews


HOW ONE THING HAS LED TO ANOTHER

15

compatibility genes" that were shown to govern the transplantability of tissues might provide me with the right system to substantiate our speculations on variation and selection within populations of tumor cells? Despite myloneliness and separation from Eva, alleviated only somewhat by the letters I mailed her daily, I enjoyed being in America. In addition to the positive attitude of Schultz and Hauschka,the environment of the whole laboratory was highly supportive for a young man. There was a wholesome difference compared to European laboratories, particuarly with regard to teacher-student relationships. It can best be summarized by a statement of the Danish biochemist, Lindestrrm-Lang: "The greatest accomplishmentof the Americanrevolution was to establish the right of youngstudents to ask foolish questions." During mystay in the United States I lost part of myemigrant complex. Hungarian emigrants in a comparable situation commentedlater that my initial shyness has turned into its opposite in the Americansetting. It may have seemedso. I was no longer afraid to ask questions, to inject myself into the conversation of learned professors, to speculate, and to risk makinga fool of myself. I began to feel that this was not only mynatural right but myresponsibility.

TumorProgression by Variation and Selection The work on the conversion of solid into ascites tumors turned out to be quite interesting (2). Lymphomas and leukemias often converted immediately, while carcinomas and sarcomas refused to grow in the ascites form at first. Someof them could be converted gradually, however, by passaging the few desquamated,freely floating tumor cells in the peritoneal fluid. Using a simplified form of the Luria-Delbrfick fluctuation analysis, I could show that this conversion was due to the selective enrichment of a small number of spontaneously occurring variants. After four months in the United States, I returned to Swedenwith 200 mice, anxiously guarded in myNewYork hotel room overnight and during the plane trip of more than 24 hours, to the great displeasure of myfellow passengers. Back in Stockholm, the ascites tumor variants turned out to be stable. Theyretained the ability to growin the peritoneal fluid immediatelyafter inoculation, even after reconversion to the solid form and subcutaneous propagation over extended periods of time. The ascites adapted tumors were also more metastatic, less adhesive, and had a higher surface charge than their original nonadapted counterparts (3, 4). A comparison of our findings with Leslie Foulds’ (5) work on tumor progression and Jacob Furth’s studies (6) on the change of hormone dependent to autonomous tumors convinced us that we had hit an unusually well-defined case of

Annual Reviews 16

KLEIN & KLEIN

progression (7). It appearedto have a certain clinical relevance, at least the conceptual level, because it showedt.hat tumor cell populations were heterogeneous, and subpopulations could differ in their metastatic properties. But where did we go from here?


Tumor

Immunolofy

To study variation and selection in tumorcell populations, it was obviously necessary to study variation first. Wewere looking for cellular markers, determined by knowngenes that could be detected at the cellular level. Wefound them in the recently discovered H-2 antigens of the mouse. George Snell has just started to distribute his first H-2 congenic mouse strains. Wehad induced tumors in H-2 heterozygous but otherwise congenic F1 hybrids and isolated haplotype-loss variants by transplantation to the parental strains (8). Single haplotype-loss variants could be readily obtained, but in frequencies that varied widely between different tumors, even if they had been induced by the same agent and in the same host genotype. This biological variability was no longer a surprise to us, after the variations in ascites convertibility that we had encountered previously. Double H-2 haplotype losses were extremely rare. Around this time, in the mid-1950s, a former colleague from medical school started to makeextravagant claims concerning the prospects for the immunological prevention and cure of humancancer. He had immunized a horse with pooled tumor tissue and was firmly convinced that his serum reacted with a universal tumor antigen. He advocated immediate vaccination against cancer. The newspapers made a big splash. He was supported by some of the most powerful professors of microbiology and virology whohad no experience in cancer. He injected himself with a HeLa cell derived "cancer vaccine" on TV. The public regarded him as a hero, particularly since the newspapers started to accuse the "cancer establishment" of lacking any concern for preventing cancer, due to vested interests. The few of us who actually worked with cancer cells were profoundly sceptical. This could just not be true. But what was the real situation? Wouldtumors elicit immunityin their own inbred strain of origin? Myprevious work with Hauschka left me imbued with a healthy scepticism against most earlier research in tumor immunology.The field was dominated by misinterpreted artefacts of experimentation with noninbred mice. The confusion between transplantation immunology and tumor immunologyprevailed during the entire first part of the century. Only several decades after the developmentof the inbred mice by Little, Strong, Tyzzer, McDowelland others, and after the formulation of the "transplantation laws" by George Snell, was it gradually realized that the so-

Annual Reviews



17

called "transplantable tumors" violated histocompatibility barriers because serial homografting had selected them to outpace the rejection response. If the balance was tilted in favor of the host, e.g. by preimmunizationwith attenuated tumor cells, it could reject the tumor. This easily won immunitycould not be reproduced with tumors that had arisen in homozygousmice and were tested within their own strain. But what would happen at a more modest level of ambition? Could immunityprotect the syngeneic host against near-threshold numbersof tumor cells? Clinicians and pathologists have always maintained that only a small proportion of disseminated tumor cells could grow into metastases in the humanpatient. Could an immuneresponse that fell short of protecting the host against an established tumor still reject disseminated cells, in analogy with concomitant immunityin antiparasite responses? Just as we started to think about these matters, Foley (9) and Prehn Main (10) suggested that chemically induced mouse sarcomas, but not spontaneous mammarycarcinomas, could elicit a state of immunity in syngeneic mice. The data were persuasive but still not fully convincing. Did chemically induced tumorcells really possess a distinct antigenicity of their own, or did these experimentsmerely reflect a residual heterozygosis in the inbred strains? It was obvious that the question could be decisively settled if it could be shown that the primary host could be immunized against its owntumor. Using a combined scheme of tumor induction, operative removal, immunization with irradiated autologous tumor cells, and challenge with graded numbers of viable cells, we could show that methylcholanthreneinduced sarcoma cells were indeed capable of inducing true rejection reactions in the original host (11). Different tumors varied in their immunogenicity over a 5 log range of cell doses, required to break the state of immunity. Another and even more striking manifestation of biological individuality concernedthe individual distinctness of the tumor antigens, also noted by Prehn, Baldwin, and Old (12-14). Each tumor could only immunizeagainst itself. Cross-reactions were rare and irregular. The total number of possible specificities is still not known.Wefound no crossreactions among more than 20 tumors. Hellstrrm could not immunize mice against MC-carcinogenesis by using pools of a dozen tumors for immunization, while Old reported a certain preventive effect after the use of nonspecific immunomodulatorsthat acted presumably by boosting the host’s own responsiveness. The nature of the carcinogen was not immaterial in determining the immunogenicity of the chemically induced tumors. Amongthe aromatic hydrocarbons, MC, BP, and DMBAinduced sarcomas with decreasing immunogenicity, in that order. Sarcomas induced by the implantation of

Annual Reviews 18

KLEIN & KLEIN

cellophane film were hardly immunogenicat all (15). In the rat, Baldwin found that most azo dye-induced tumors were highly immunogenic, whereas acetylaminofluorene-induced tumors and spontaneous fibrosarcomas were not immunogenicat all (16). Several decades have passed since these findings, but the nature of the TSTA(tumor specific transplantation antigen) of the chemically induced tumorsis still a mystery.


Antigenicity of Virus-Induced Tumors In 1958 I went to the Canadian Cancer Conference, in Honey Harbor, Ontario. Stewart and Eddy’s pioneering work on the polyoma virus was still very new. Most participants were flabbergasted by the number and variety of the tumors that arose after the inoculation of the virus into newborn mice. Burnet was one of them. "Sir Mac" had recently shifted from virology to immunologyand had developed a very negative view of the role of viruses in cancer in the course of this transition; he considered all virus-induced tumors as laboratory artefacts. Viruses were essentially cytopathic, and he saw no place for any true tumor inducing effect. Confronted with the polyoma story, he formulated immediately a new hypothesis. It was based on the only observation of Stewart and Eddy that turned out to be incorrect. They claimed that polyoma tumors were not transplantable. This was due to the accidental use of heterozygous mice, however. Burnet suggested that polyoma virus may destroy some unknown, systematic "growth-controlling center," a possible "hypothalamus-like" homeostatic regulator of cell renewal in manydifferent tissues. This would explain the ability of the virus to cause tumors in manydifferent tissues. These tumors would not be transplantable to mice that have not been similarly conditioned by polyomavirus. Hans-OlofSjrgren had just started to work with us at this time. Stimulated by Burnet’s idea, I asked him to test the transplantability of polyoma tumors in unmanipulated and polyoma-infected syngeneic mice. The result was the exact opposite of what was predicted by Burnet’s hypothesis: the tumors were readily transplantable to untreated mice, while small, graded numbers of cells were rejected by virus-inoculated syngeneic mice (17). The resistance of the virus-infected mice could be transferred adoptively with lymphocytes but not with serum. Both Karl Habel and our group later showedthat antiviral immunitywas neither necessary nor sufficient to induce rejection. Polyoma-inducedtumors or transformed cells induced rejection, whether they released virus or not. All polyoma-inducedtumors were rejected by the immunizedmice, irrespective of tissue origin, but they did not reject tumors induced by other viruses or by chemical agents. We

Annual Reviews


HOWONE THING HAS LED TO ANOTHER 19 have therefore developed the concept ofa polyoma-specific transplantation antigen (TSTA)that was present in all tumors induced by. polyoma, but not in tumors induced by other agents. Weand others later found that similar group-specific rejection-inducing antigens were present.on other virus-induced tumors (18). The retrovirus induced leukemias were particulady useful for the study of both humoraland cell-mediated reactions, as was shownby Old et al (19), and by our group. Moloneyvirus-induced lymphomaswere particularly useful for these studies, since they gave a brilliant membranefluorescence reaction with the sera of preimmunized, syngeneic animals. Nevertheless, it was not possible to distinguish the rejection inducing antigen from the viral glycoprotein that accumulated on the surface. Different Moloneylymphomas induced in the sameinbred strain differed in their rejection inducing potential. This system has permitted a distinction between immunogenicity and immunosensitlvity, and we could show that they were independent variables. The former correlated with virus release, while the latter did not. The Department

of Tumor Biology

During the years of our transition from H-2 antigens to tumor immunology, our department developed rapidly. It was formally established in 1957, against all odds. Previously, both George and Eva had becomeassistant professors in Caspersson’s Department of Cell Research (in 1951 and 1955, respectively), but our appointments were limited to maximum of 6 years. Unless one acquired a tenured position, one was out of the research system. But no tenured positions were available in our field, which had not been previously represented at the Swedishuniversities. To circumvent the inflexibility of the university system, a numberof"personal professorships" had been established for individual scientists, but some years before this time, the governmentdecided to stop creating new positions of this type. Science was too expensive already for a country of eight million, they said; and it was also undesirable to continue the traditional recruitment of medical students into research. There was a shortage of doctors that madethe authorities very sensitive about this, particularly since all higher education was financed by the taxpayers, and there was heavy competition for admission to the medical schools. I had a good offer from the ICR in Philadelphia, and we seriously considered movingto the United States. Meanwhile,the Karolinska Insti~ tute, the Medical Research Council and the SwedishCancer Society joined forces to initiate a parliamentry move,requesting the establishment of a Department of TumorBiology with George Klein as its first head. This move was supported by representatives from the four major political

Annual Reviews


20

KLEIN & KLEIN

parties, but it failed to convince the Government.Decisions about budgetary matters rest with the Parliament, however. A parliamentary committee dealt with the matter on April 30, 1957. The odds were against us. The committee had 13 members of the ruling Labour Party and 12 members. from the three major opposition parties. It was expected that the movewouldfail with a majority of at least one vote. In fact, the opposite happened. One Labour Party member(unknown to us) decided to vote with the opposition parties and the Departmentwas established, as of July 1, 1957. Numerousmedical and PhDstudents interested in research joined our group. With the support of the National Institutes of Health of the United States and the Swedish Cancer Society, the department expanded rapidly. The accumulation of married couples who pursued research together was a peculiar feature of the lab that has remained with us ever since. In the early years, the Hellstrrms, the Mfllers, the Sjfgrens, the Nordenskjflds, the Nadkarnis, and the Ozers were some of the examples. At one point we had seven married couples working at the lab at the same time, surely a world record. The problem of space became overwhelming in the late 1950s. Again, there was no provision or precedent for the type of support that was needed. Wewere facing the possibility of having to return the first major NIHgrant we had received under the Virus Cancer Program. I turned to the SwedishCancer Society, although with little hope since the statutes of this essentially private organization explicitly discouraged the support of building facilities. But the Chairman,Professor Hilding Bergstrand, drove through a positive resolution against all odds. A new laboratory building was constructed in 1961. It houses the Department even today. Burkitt’s

Lymphoma

Sometimein the mid-1960s, Eva suggested that we should use our experience on virus-induced murine lymphomas to examine a human lymphoma with a presumptive viral etiology. Could we detect group specific antibody responses that might be helpful in tracing a virus? Burkitt’s lymphoma (BL) was the obvious choice. The recent description of the highly endemic occurrence of the African form which is climate-dependent strongly supported the idea of a possible viral etiology. I wrote letters to numerous hospitals in Africa and to international organizations, explaining our project and asking for tumor, blood, and serum. I received somepolite letters in reply, promises of material, and lovely stamps which made myson happy. But the material was not forthcomingat all, apart from an occasional shipment that arrived broken or infected. Then somebody--I have forgotten who--advised me to write to Peter Clifford, ENTsurgeon at the Kenyatta National Hospital in Nairobi.

Annual Reviews



21

I got no letter and no stamps in reply, but the material started comingin a continuous flow. It arrived with chronometric precision on the single direct flight from Nairobi, late Tuesday afternoon. Large dry ice boxes carried hundreds of sera, and a special wet ice package contained fresh biopsy material. There was always a long list in Clifford’s ownhandwriting with all the essential details and a brief "good luck" message. Weworked together with Peter over a period of more than 10 years. Wehave published 45 joint papers, the first in 1966(20), the last in 1974 (21). Weworked and published together for several years before we had a chance to meet in person. This taught us a new lesson. For collaborative studies, we tried to find a colleague who was motivated to study the problem and to collaborate with us, no matter where he or she resided. But we hasten to add that we have never encountered another clinical collaborator like Peter Clifford. He had a profound interest in BL, ever since he introduced chemotherapy in the treatment of the disease and became fascinated by the remarkably good regression in most of the patients. Their long-term survival eventually turned out to be complete cure in 15-20%of patients, including those whohad only received incomplete chemotherapy. This was quite different from the effect of chemotherapy on other types of B-cell lymphomas.Clifford was convinced that the immunologicalresponse of the patient was decisive. If it was effective, even incomplete chemotherapy could induce total and long-lasting remission. If it was not, even more effective forms of chemotherapywere ultimately unsuccessful. Peter hoped that we would find evidence for an antitumor response in his patients. We changed our working habits. Every Tuesday night was "Burkitt night." Wemade living cell suspensions from the fresh tumors, reacted them with the patient’s own serum and other sera, and tried to read the tests immediately to obtain clues for the continued work. It was not difficult to motivate our personnel to work through the night every Tuesday. Eventually, numerous other laboratories requested material, in the United States, England, and Japan, and some of them became engaged in collaborative projects. Wecould identify a membraneantigen (MA)that was expressed in some Burkitt lymphoma~terived cultures, but not in others (20). WhenI presented these data at an ACSConference in Rye, NewYork, in 1967 (22), WernerHenle gave a talk in the same session. reported his results, obtained with an immunofluorescencetest on fixed BL cells that he and Gertrud Henle had recently developed, later known as the VCA(viral capsid antigen) test (23). They already knew that reaction was due to structural antigens of a newly discovered herpes virus, first seen by Epstein, Barr, and Achongin the electron microscope. Henle

Annual Reviews


22

KLEIN & KLEIN

showedthat it was antigenieally distinct from previously knownherpes viruses (24). Wedecidedto call it EBV. The Henles’ VCAtest and our MAtest showeda certain concordance. Thesamelines appearedto react or failed to react in both tests. At the Ryemeetingweagreed to collaborate. This initiated a highly productive association that has lasted for 20 years, terminatedonly by WernerHenle’s death in 1987. Alreadyin the beginningof this workweobtaineddefinite evidencethat MAwasencodedby EBV(25). It is nowknownas one of the viral envelope glycoproteins. It assembleswithin the membrane of the virus-producing ceils, andafter virus release, it can also attach to other cells in the same culture if they carry EBV-reeeptors.WithJondal, Yefenof,and Oldstone, welater identified the B-cell specific C3d(CR2)receptor as the attachment site of the viral glyeoprotein(26, 27). By1970, it was clear that Epstein, the Henles, and ourselves had only seen the top of the iceberg whenwelookedat viral particles, VCA or MA. Theyonly appear in virus-producingcell lines, and only in someof the cells. With Harald zur Hausen, we found in 1970, however, that more than 90%of the African BLsand all low differentiated or anaplastie nasopharyngeal carcinomas (NPC)contained multiple EBV-genomes r cell, no matter whetherthey producedvirus or not (28). In 1973I [GK] have found with Beverly Reedmanthat 100%of the cells in EBV-DNA positive BLbiopsies and cell lines contained an EBV-encoded nuclear antigen, which wedecided to call EBNA (29). Todaywe knowthat EBNA consistsof a familyof at least six different proteins(30). Several important discoveries have been madeby others in the meanwhile. The Henles, Pope et al, and Nilsson et al found that EBVcould readily immortalize normal B cells in vitro (31-33). Departing from serendipitous observationon a laboratory assistant, the Henlesdiscovered (75) that EBVis the causative agentof infectious mononucleosis (IM). Svedmyr,wecould readily detect EBNA-positive cells in the peripheral blood of mononucleosispatients (34), and the Henles and GeorgeMiller foundthat the saliva of these patients containedtransformingvirus. Transformation was thus a natural property of the virus, not a laboratory artefact due to the accidental isolation of a defective strain, as our colleaguesin the lytic herpesvirus fields initially surmised.Miller &Epstein have also shownthat EBVcan cause lethal lymphoproliferativedisease in immunologicallynaive marmosetand owl monkeys(35, 36). Mononucleosis appearedas an acute rejection reaction of the "immunologically prepared" humanhost, selectively conditionedby a nearly symbiotic relationship with EBVover millions of years, against the virally transformedB cells. Wefound that the peripheral blood of the acute IM

Annual Reviews



23

patient contains activated killer cells that can lyse EBV-carrying and other target cells (37, 38). Moreover,autologous mixedlymphocytecultures between EBV-transformedB-cell lines and T cells of the same normal donorgenerated a proliferative and cytotoxic response equally as strong as that of MHC-incompatible allogeneic MLC(39). Later, Rickinson, Mossand Pope showedthat the autologous mixed cultures generated specific MHC class 1-restricted CTLsby repeated stimulation (40). Eva’s group, Sigurbj6rg Torsteinsdottir, and MariaGrazia Masucciin particular, showedthat theCTLresponse was heterogenous, directed against different target epitopes (41). Thenature and specificity of the relevant targets have not beenclearly defined yet in terms of the knownvirally encodedproteins, althoughcurrent evidenceby Mosset al and by ThorleyLawson,respectively, indicates that both EBNA-2 and LMP epitopes may servein this capacity(42, 43). Since the work of Townsendet al (44, 45) has shownthat MHC class I-associated peptides of processedendogenous or viral proteins can serve as immunogens and CTLtargets, it wouldnot be surprising if even more amongthe seven knowngrowth transformation-associated EBVproteins could serve as CTLtargets. A similar reasoning can be applied to the polyomavirus-induced TSTA,discussed above. The recent work of Dalianis et al in our laboratory suggests that all three polyomaencoded T-antigenscan elicit rejection responsesof the TSTA-type. Thehypothesisthat T cell-mediatedresponsesinhibit the proliferation of EBV-carryingB cells in healthy seropositives and in IMpatients was reaffirmed whenwefound with DavidPurtilo (46) that most and perhaps all lymphoproliferativediseases that appear in congenitally or iatrogenically immunodefective patients, like children with the X-linkedlymphoproliferative syndromeor organ transplant recipients, carry EBV-genomes. Hanto, Ho, and others have later shownthat these initially polyclonal immunoblastie proliferations mayprogress to monoelonal lymphoma (47, 48). Whilethe tumorigenicpotential of EBVwasclearly established by these and related findings, its lifelong, innocuouslatent presencein morethan 80%of all humanpopulations has also suggested that disease occurs only as an accident. Even mononucleosisappears as an "accident" of civilization. Modernhygienicconditions haveapparently interfered with the normal,disease-free ecologyof the virus-host relationship, with its predominantsymptom-freeearly childhood infection. The "accident" of the EBV-associatedtumors has nowbeen largely clarified for Burkitt’s lymphoma,as described below, while the pathogenesis of nasopharyngealcarcinoma, the most regularly EBV-carrying humantumor, is still not understood.

Annual Reviews 24

KLEIN& KLEIN


Oncogene Activation

by Chromosomal Translocation

By 1970, it was clear that some important element was missing from the BLscenario. EBVhas clearly contributed to the genesis of the high endemic form of the disease, since 97%of the African BLs carried the viral genome, whereas non-BLlymphomasdid not (49). Moreover, the prospective study of Geser and de The (reviewed in 50) showed that children with a high EBV-loadare at a greater risk to develop BLthan are their brothers and sisters with a low EBV-load,as indicated by the antibody titers. Since the numberof EBV-infected B cells represents only a minor fraction of the total B-cell population even in persons with a high EBV-load, the presence of the virus in the majority of the African BLs can only be interpreted to mean that an EBV-carrying B cell runs a greater risk of turning into a BL cell under the conditions prevailing in the "high BL belt" of Africa than does its EBV-negativecounterpart. This is to say that EBVcontributes to the etiology of the tumor. But this is still not a satisfactory explanation; someessential element is obviously missing. BLs differ from the true EBV-inducedlymphoproliferative diseases like fatal mononucleosisor the immunoblastic iymphoproliferative diseases in organ transplant recipients, with regard to their cellular phenotype (51). The latter resemble the EBV-transformedB-cell lines of nonneoplastic origin (LCLs). LCLsare permanently growing immunoblasts that express a set of activation markers but not CALLAor BLA. BL cells, on the other hand, carry surface antigen and glycoprotein markers that resemble resting B-cells, rather than immunoblasts (52, 53). They express CALLAand BLAbut no activation markers (unless they drift to a more LCL-like phenotype during prolonged cultivation). Recently, Gregory et al found normal B cells with a corresponding phenotype in tonsil germinal centers (54). For the understanding of BL pathogenesis, it is also important to remember that approximately 3% of the African BLs, and 80% of the sporadic BLs that oc~:ur all over the world are EBV-negative. Amongthe recent, AIDS-associated BLS, the incidence of the EBV-carrying form is currently estimated as 40-50%. The discovery of the "missing factor" in the "Burkitt equation" started when Manolovand Manolova reported in 1972 (55) that a 14q + chromosomal marker was present in about 80% of the tumors. The Manolovs came from Sofia, Bulgaria, to work with us in 1970, at the time whenthe chromosomebanding technique was discovered by Caspersson and Zech. I suggested that they apply the banding technique to the cytogenetically unexplored BL that kept coming in from Clifford every Tuesday in excellent condition. Theyagreed rather reluctantly since they had hoped to learn

Annual Reviews


HOWONE THINGHASLEDTO ANOTI-iER

25

some immunology. But their cytogenetic work soon picked up momentum, particularly after Albert Levan agreed to consult and guide them. When George Manolovshowed me the extra band that he found attached to the distal part of the long arm of one chromosome14 in a BLbiopsy, I first suspected sometrivial reason, perhaps a constitutional variation (isochromosome),and suggested that the Manolovsshould take a look at the fibroblasts of the patient. So they did, but they found that the anomaly was totally restricted to the clonal tumor. After the Manolovsreturned to Bulgaria, we continued the work with Lore Zech. She soon showed that the "extra piece" was derived from chromosome8; the 14q + marker was thus a product of a reciprocal 8; 14 translocation (56). Several groups found subsequently that approximately 20% of the BLs that had no 14q+ marker carried one of two variant translocations instead (for review, see 57). Chromosome 8 broke at the same site (8q24) and entered into a reciprocal translocation either with the short arm of chromosome2 or with chromosome22. All BLs were found to carry one of the three translocations, no matter whether they were high endemic or sporadic, EBV-positive or negative. The same translocations were only exceptionally found in non-BL-lymphomas, although 14q+ markers are quite common;they usually arise by reciprocal translocations between chromosome 14 and some other chromosome, with 11 and 18 as the most frequent participants. But BL-typetranslocations were also found in the form of B cell-derived ALLthat resembles Burkitt lymphomacells phenotypically and is often called Burkitt leukemia. Meanwhile,another, quite independent cytogenetic study, entirely confined to mousetumor cells, was progressing in our laboratory. It started when the Hungarian-Rumanian pathologist, Francis Wiener joined our group in 1970. He is still one of our closest coworkers. Wiener became interested in the role of chromosome15 trisomy in mouseT-cell leukemia, and he was also the main cytogeneticist in the somatic hybrid studies, together with Henry Harris, mentioned below. In the late 1970s Wiener examined a series of pristane oil-induced mouse plasmacytomas (MPCs); he was working together with a Japanese guest worker, Shinsuke Ohno, and in collaboration with Michael Potter’s group at the NIH. Our 1979 Cellpaper described the MPC-associatedtypical (12; 15) and variant (6; translocations (58). Mouse plasmacytomas are very different from Burkitt lymphomas. The only commondenominator is that both originate from cells of the Blymphocyte series. Wenever expected to find anything in commonbetween the two. Therefore, the fact that two apparently unrelated research projects, carried out by different cytogeneticists, led to the discovery of a commonpathogenetic mechanism, based on almost exactly homologous

Annual Reviews


26

KLEIN& KLEIN

chromosomaltranslocations, was one of the greatest and most pleasant surprises of myentire scientific career. It was even more surprising that the highly speculative workinghypothesis, formulated to explain the mechanism wherebythe translocations contribute to the tumorigenic process in such a decisive fashion, turned out to be essentially correct. The hypothesis was built on the fact that the recipient murine chromosomes of the dislocated fragment from chromosome 15 were known to carry the IgH (chromosome 12) and the kappa (chromosome 6) gene, respectively. Likewise, human chromosome 14 was known to carry the IgH cluster. Wehave therefore speculated that a proto-oncogene and probably the same proto-oncogene could be localized at the breakpoint of the murine chr 15 and the humanchr 8. Accidental translocation of the putative gene to one of the immunoglobulin loci might have led to the constitutional activation of the gene, in analogy with the retroviral activation of the c-myc gene by the insertion of an ALV-derivedLTRin the chicken bursal lymphoma,as described by Haywardet al. I started to expose the hypothesis to the test of peer criticism in 1979. An outstanding molecular biologist, a good friend of mine, called it the "most hair-raising extrapolation from the centimorgans to the kilobases." It was. Still, the hypothesis was published in Nature in 1981 (59), but was not fully convinced of it myself, until the critical momentduring the summerof 1981, when I was waiting for a plane at Washingtonairport to take me to Tokyo. The waiting hall was full of people, mostly Japanese. There were only two telephones on the other side of the security check. They were busy most of the time. The plane was called up. Finally, one of the telephones was free. I tried to get hold of Philip Leder at the NIH. I wanted to hear whether he knew anything about the chromosomallocation of the immunoglobulin light chain genes in humans. Leder came to the telephone. No, he hadn’t heard anything; it was still unknown.But one of his colleagues had just come back from the recently held HumanChromosome Mappingmeeting in Oslo. If I waited, he would try to ask if the colleague had heard anything. "Final call." The last Japanese walked aboard, and I had to leave. At the momentwhen I was about to hang~up the phone, Leder’s voice came back: Yes, there were two small reports in Oslo. An English group had found that kappa is on chromosome 2. An American group had proved that lambda is on chromosome22. I ran on board. It was an intoxicating feeling! I knew for certain that the hypothesis was correct. The molecular confirmation and clarification camein a virtual avalanche during 1982. Taking off from quite different points, Jerry Adamswith Susan Cory in Australia and Kenneth Marcu in New York showed for

Annual Reviews


HOW ONE

THING

HAS

LED

TO ANOTHER

27

MPC,and Carlo Croce and Phil Leder for BL, that the translocations resulted in the juxtaposition of donor chromosomederived sequences and immunoglobulin gene sequences. Michael Cole’s group has identified the transposed gene as c-myc (for review, see 60). The subsequent development has led to many new insights, but it has also created some puzzles and paradoxes with regard to myc-regulation, constitutive activation, and certain details of the timing and regulation of Ig-gene rearrangement (for review see 65). With Francis Wiener and Janos Siimegi, we have also found a third Ig/myc translocation system (61-63), the spontaneous immunocytoma of the Louvain rat (RIC), developed by Herv~ Bazin. A comparison of the translocations in MPC, RIC, and BLat the molecular level reveals more similarities than differences. In fact, it wouldbe hard to find a comparablesituation in cancer biology where three pathogenetically different tumors that arise from the same cell lineage in three different species show a similarly close pathogenetic mechanismat the molecular level. The causal, i.e. rate limiting, involvementof constitutive rnyc activation in the genesis of the three tumors was deduced from the regularity of the Ig/myc juxtaposition that extended to cryptic translocations and complex rearrangements, where two or three successive genetic events had occurred (61, 64). Further confirmation came from recent facsimile experiments. Michael Potter and Francis Wiener showed (66) that introduction of activated rnyc gene within a retroviral (J3) construct into pristane oiltreated Balb/c mice induced plasmacytomasthat did not carry any translocations, provided they expressed the inserted (v-myc) gene. Meanwhile, Adams& Cory’s group generated transgenic mice that carried the mycgene coupled to the IgH enhancer (67). The mice developed more than 90%pre-B- or B-cell-derived lymphomas.Using the Australian transgenic mice, Francis Wiener recently found that Abelson virus infection, already knownto increase the incidence and shorten the latency period of pristane oil-induced mouse plasmacytoma, has led to the appearance of plasmacytomas in the Emu-myctransgenic mice. The virus has obviated the pristane requirement and lifted the genetic restrictions to MPCsusceptibility. These plasmacytomaswere also translocation free. That introduction of an activated myc construct was tumorigenic for B cells and obviated the need for the translocations could be only interpreted to mean that the naturally occurring constitutive activation of myc by the Ig-translocations provided an essential, rate-limiting step within the carcinogenic process. But it is not the only step. All tumors were monoclonal, even in the transgenic mice where myc was activated in all B and pre-B cells. Sequential activation of several oncogenesor, alternatively, loss of suppressor genes mayprovide additional steps. Feedbackinhibition

Annual Reviews


28

KLEIN & KLEIN

by the clone that happensto get the upper hand first wouldbe another alternative. The Burkitt lymphoma story has also developedfurther in the meanwhile and has posed somenewfascinating questions. Wehave suggested, for both conceptualand factual reasons, that the BLprogenitoris a longlived B-memory cell. In this scenario, antigenicallystimulatedB-cell clones that have previously expandedas immunoblasts,were in the process of switching their phenotype to CALLAand BLA-positive, activation markernegative memory cells when,uponthe waningof the antigenic stimulus, the translocation accident occurred. Dueto the linking of rnyc to a constitutively active Ig-locus, the cells were unableto leave the cycling compartment,however.It could be shownthat the translocation carrying "suspendedresting cell" had several additional phenotypicproperties that could facilitate its evasion from immunologicalcontrol. Certain MHC elass I polymorphic specificities weredown-regulated in the BLcells, compared to EBV-transformed B-cell lines of normal origin. The BLcells also failed to express certain adhesionmoleculespresent on the LCLsor expressed them at a low level. Even the EBV-encoded,growth-transformation associated nuclear and membraneantigens were down-regulated in the BLcells, with the exceptionof EBNA-1. This wasparalleled by a relative resistance of the BLcell to CTL-mediated lysis (56). It thus appears that the myc/Ig translocation promotesthe malignant growthof the BLcell by several mechanisms.This mayexplain the extraordinaryregularity of its presencein all typical BLsso far studied. Tumor Suppressor

Genes

I haverecently reviewedthis field in somedetail (68) and concludedthat weare probably approaching an era whenthe study of genes that can antagonizetumorigenicbehaviorwill be equally as, if not morerewarding than, the study of the oncogenes.Ourowncommitment to this field started with a decadeof another long distance collaboration, initiated by Henry Harris in 1969(69). Wehave inoculated a large numberof somatic cell hybrids, derived from the fusion of high malignantwith normalor with low malignantcells, into genetically compatibleand/or immunosuppressed mice. The hybrids were generated by Harris, and Wienerexaminedtheir chromosomes.These studies have firmly established the notion that tumorigenicityis suppressedby fusion with normalcells. It reappearsafter somecritically important chromosomes, contributed by the normalcell, have been lost. Others have extended this work to human/human hybrids morerecently and obtained similar results. Chromosomes that carry tumor suppressor genes have been identified by Stanbridge, Klinger, Sager and their associates (70-72). Thefield is nowmovingtowardsa morereduc-

Annual Reviews HOWONE THINGHAS LED TO ANOTHER

29

tionistic analysis where microcell hybrids are taking the place of whole cell hybridization and c-DNAtransfections are initiated to identify the suppressor genes and their products. Meanwhile,evidence for tumor antagonizing genes has also emerged from the study of revertants and particularly from the rapidly movingfield of "recessive cancer genes" that contribute to tumorigenesis by their loss (73, 74). It is not clear if or what extent there is a relationship betweenthe genes identified by these three approaches. Annu. Rev. Immunol. 1989.7:1-34. Downloaded from arjournals.annualreviews.org by HINARI on 08/29/07. For personal use only.

Whither Tumor Immunology? It is often asked if or to what extent the spectacular developmentof the oncogene field during the last decade may provide some new handles for targeting the antitumor response. The answer may differ in relation to oncogenes activated by regulatory or by structural changes, respectively. Up-regulation of a structurally normaloncoprotein is less likely to provide a rejection target than oncoproteins activated by structural changes, e.g. the products of the ras-mutations or the truncated growth factor receptors, exemplified by the tumorigenic variants of erbB or fins. Following Townsend’s discovery that intracellular, endogenousproteins can be processed to peptides that combinewith class I or class II molecules and can then serve as immunogensand/or as CTLtargets, the structurally changed oncoproteins deserve serious consideration. Progress will depend on the expression of mutation-activated (compared to normal) oncogenes in nonimmunogenic tumor cells--of which there are many--followed by the assessment of their immunogenicityand rejectability in syngeneic hosts.

Epilo#ue As each of us is moving towards the approaching darkness, the sun is never setting over the vast oceans of science. It has been a rare privilege to live and work through the times whenthe genetic material turned from protein to DNA,when adaptive changes in cell populations--including antibody production--were unmaskedas Darwinian variations and selection, when GODbecame the rearrangement of immunoglobulin genes, violating the dogmathat all somatic cells have the same DNA.Another central dogma was abolished when the RNAtumor viruses became DNA proviruses. Followingclosely in the wakeof this discovery, the enthusiastic retrovirologists, searching for the universal cause of cancer, permitted the great cuckoo egg, the oncogenes, to hatch--almost imperceptibly at first, but with a rapidly increasing crescendo, towards the triumphant emphasis on the regulatory genes of the cells and their dysfunction as the key factor in the oncogenic process. Departing from even greater obscurity, the MHC system, once the esoteric pet of a few mousegeneticists, nowoccupies a

Annual Reviews 30

KLEIN & KLEIN

central place in virtually every area of immunology.It was a great time, and it still is, but it is only the stumbling, stuttering, premature foreshadowingof what lies ahead. Wehave barely scratched the surface.


Literature Cited 1. Klein, G., Klein, E. 1951. The nucleic acid content and the growth rate of mouse ascites tumor cells. Acta Union Int. Contra Cancrum7:376-85 2. Klein, G. 1951. Comparative studies of mouse tumors with respect to their capacity for growth as "ascites tumors" and their average nucleic acid content per cell. Exp. Cell. Res. 11:518-73 3. Purdom, L,, Ambrose, E. J., Klein, G. 1958. A correlation between electrical surface charge and some biological characteristics during the stepwise progression of a mouse sarcoma. Nature 181:1586-87 4. Ringertz, N., Klein, E., Klein, G. 1957. Histopathological studies of peritoneal implantation and lung metastasis at different stages of the gradual transformation of the MCIMmouse sarcoma into ascites form. J. Natl. CancerInst. 18:173-99 5. Foulds, L. 1954. The experimental study of tumor progression: A review. Cancer Res. 14:32%39 6. Furth, J. 1953. Conditioned and autonomous neoplasms: A review Cancer Res. 13:477-92 7. Klein, G., Klein, E. 1957. Evolution of independence from growth stimulation and inhibition... Symp. Soc. Exp. Biol. I1:305 8. Klein, G., Klein, E. 1958. Histocompatibility changes in tumors. J. Cell Comp.Physiol. Suppl. 1, 52:125-68 9. Foley, E. J. 1953. Antigenic properties of methylcholanthrene-induced tumors in mice of the strain of origin. Cancer Res. 13:835-37 10. Prehn, R. T., Main, J. M. 1957. Immunity to methylcholanthrene-induced sarcomas. J. Natl. Cancer Inst. 18:769-78 11. Klein, G., Sj6gren, H. O., Klein, E., Hellstrrm, K. E. 1960. Demonstration of resistance against methylcholanthrene-induced sarcomas in the primary autochthonous host. Cancer Res. 20: 1561-72 12. Baldwin, R. W. 1955. Immunity to methylcholanthrene-induced tumors in inbred rats following atrophy and regression of implanted tumors. Br. J. Cancer 9:652-57 13. Old, L. J., Boyse, E. A., Clarke, D. A.,

Carswell, E. A. 1. 1962. Antigenic properties of chemically induced tumors. Ann. NY Acad. Sci. 10l: 80-106 14. Prehn, R. T. 1962. Specific isoantigenicities among chemically induced tumors. Ann. NY Acad. Sci. 101: 10713 15. Klein, G., Sj6gren, H. O., Klein, E. 1963. Demonstration of host resistance against sarcomas induced by implantation of cellophane films in isologous (syngenic) recipients. Cancer Res. 23: 84~92 16. Baldwin, R. W. 1973. Immunological aspects of chemical carcinogenesis. Adv. Cancer Res. 18:1-75 17. Sjrgren, H. D., Hellstrrm, I., Klein, G. 1961. Transplantation ofpolyoma virusinduced tumors in mice. Cancer Res. 21: 329-37 18. Klein, G. 1966. Tumor antigens. Ann. Rev. Microbiol. 20:223-52 19. Old, L. J., Boyse, E. A., Stockert, E. J. 1963. J. Natl. Cancer Inst. 31:97786 20. Klein, G., Clifford, P., Klein, E., Stjernsw/ird, J. 1966. Search for tumor specific immunereactions in Burkitt lympfioma patients by the membrane immunofluorescence reaction. Proc. Natl. Acad. Sci. USA 55:1628-35 21. Gunven, P., Klein, G., Clifford, P., Singh, S. 1974. Epstein-Barr virus-associated membrane-reactive antibodies during long term survival after Burkitt’s lymphoma.Proc. NatL Acad. Sci. USA 71:1422-26 22. Klein, G., Klein, E., Clifford, P. 1967. Search for host defenses in Burkitt lymphoma: Membrane immunofluorescence tests on biopsies and tissue culture lines. Cancer Res. 27:2510-20 23. Henle, G., Henle, W. 1966. Immunofluorescence in cells derived from Burkitt’s lymphoma.J. BacterioL 91: 124856 24. Henle, G., Henle, W. 1967. Irnmunofluorescence, interference and complement fixation technics in the detection of herpes-type virus in Burkitt tumor cell lines. Cancer Res. 27:2442-46 25. Klein, G., Pearson, G., Nadkarni, J. S., Nadkarni, J. J., Klein, E., Henle, G., Henle, W., Clifford, P. 1968. Relation between Epstein-Barr viral and cell

Annual Reviews



31

membraneimmunofluorescenceof Bur- 35. Epstein, M. A., Hunt, R., Rabin, H. kitt tumorcells. I. Dependence of cell 1973.Pilot experimentswith EBvirus in membraneimmunofluorescenceon presowl monkeys(aortus-trivigatus). Reticulo proliferative disease in an ence of EB virus. J. Exp. Med. 128: 1011-20 inoculated animal. Int. J. Cancer12: 26. Jondal, M., Klein, G., Oldstone, M.B. 309-18 A., Bokish,V., Yefcnof,E. 1976.Surface 36. Shope, T., De Chiaro, D., Miller, G. markers on human B and T lympho1973. Malignant lymphoma in cottoncytes. VIII. Association betweencomtop marmosetsafter inoculation with Epstein-Barr virus. Proc. Natl. Acad. plementand Epstein-Barr virus recepSci. USA70:2487-91 tors on human lymphoidcells. Scand. J. Immunol. 5:401-10 37. Adams,J. M., Cory, S. 1985. Myconco27. Yefenof,E., Klein, G., Jondal, M., Oldgene activation in B and T lymstone, M.B. A. 1976.Surfacemarkerson phoid tumours. Proc. R. Soc. London humanB and T-lymphocytes. IX. TwoSer. B 226:59-72 color immunofluorescence studies on the 38. Svedmyr,E.,Jondal, M.1975. Cytotoxic association betweenEBVreceptors and effectorcells specificfor Bcell linestranscomplementreceptors on the surface of formedby Epstein-Barrvirus are present lymphoidcell lines. Int. J. Cancer17: in patients with infectious mono693-700 nucleosis. Proc. Natl. Acad. Sci. USA 28. zur Hausen, H., Schulte-Holthausen, 72:1622-26 H., Klein, G., Henle, W., Henle, G., 39. Svedmyr,E. A., Deinhardt, F., Klein, G. 1974.Sensitivity of different target Clifford, P., Santesson, L. 1970. EBVDNA in biopsies of Burkitt tumorsand cells to the killing action of peripheral anaplastic carcinomas of the nasolymphocytesstimulated by autologous pharynx. Nature 228:1056-58 lymphoblastoid cell lines. Int. J. Cancer 29. Reedman,B. M., Klein, G. 1973. Cellu13:891-903 lar localization of an Epstein-Barrvirus 40. Rickinson, A. B. 1966. Cellular (EBV)-associatedcomplement-fixing animmunological responses.In The Epsteintigen in producer and non-producer BarrVirus: RecentAdvances,pp. 77-125. lymphoblastoid cell lines. Int. J. Cancer London: HeinemannMedical 11:499-520 41. Torsteinsdottir, S., Masucci, M. G., 30. Ricksten,A., Kallin, B., Alexander,H., Ehlin-Henriksson, B., Brautbar, C., Dillner, J., F~hraeus,R., Klein,G., LerBen-Bassat, H., Klein, G., Klein, E. 1986. Differentiation-dependent senner, R., Rymo,L. 1988. The BAMHIE region of the Epstein-Barrvirus gesitivity of human B-cell derivedlines to nomeencodes three transformationmajor histocompatibility complexassociatednuclear proteins. Proc.Natl. restricted T-cellcytotoxicity.Proc.Natl. Acad. Sci. USA85:995-99 Acad. Sci. USA83:5620-24 31. Henle, W., Diehl, V., Kohn, G., zur 42. Moss,D. J., Misko,I. S., Burrows,S. R., Burman,K., McCarthy,R., Sculley, Hausen, H., Henle, G. 1967. Herpestype virus and chromosomemarker in T. B. 1988. CytotoxicT-cell clones disnormal leukocytes after growth with criminatebetweenA- andB-typeEpsteinirradiated Burkitt cells. Science 157: Barr virus transformants. Nature331: 1064-65 719-21 32. Nilsson, K., Klein, G., Henle, W., 43. Thorley-Lawson, D. A., Israelsson, E. S. Henle, G. 1971. The establishment of 1987.Generationof specific cytotoxic T lymphoblastoid lines from adult and cells with a fragmentof the Epstein-Barr fetal humanlymphoidtissue and its virus-encoded p63/latent membrane dependenceon EBV.Int. J. Cancer8: protein. Proc. Natl. Acad.Sci. USA84: 443-50 5384-88 33. Pope, J. I-I., Hornc,M. K., Scott, W. 44. Townsend,A. R. M., McMichael,A. J., 1968. Transformationof foetal human Carter, N. P., Huddeston, J. A., leukocytes in vitro by filtrates of a Brownlee,G. G. 1984. CytotoxicT-cell humanleukaemia cell line containing recognition of the influenza nucleoherpes-likevirus. Int. J. Cancer3: 857protein and hemagglutininexpressedin 66 transfected mouseL-cells. Cell 39: 1334. Klein, G., Svedmyr,E., Jondal, M., 25 Persson, P. O. 1976. EBVdetermined 45. Townsend,A. R. M., Rothbard, J. M., nuclear antigen (EBNA)-positive cells Frances, M., Gotch, G., Bahadur,J., the peripheral bloodof infectious monoWrath,D., McMichael,A. J. 1986. The nucleosispatients. Int. J. Cancer17: 21epitopes of influenza nucleoproteins 26 recognized by cytotoxic lymphocytes

Annual Reviews


32

KLEIN & KLEIN

can be definedwith short synthetic peptides. Cell 44:959~58 46. Purtilo, D. T., Klein, G. 1981. Introduction to Epstein-Barr virus and lymphoproliferative diseases in immunodeficient individuals. Cancer Res. 41:4209 47. Hanto,D. W.,Frizzera, G., Gajl-Peczalska, K.J., Sakamoto, K., Purtilo, D. T., Balfour, H. H., Simmons, R. L., Najarian,J. S. 1982.Epstein-Barrvirusinduced B-cell lymphoma after renal transplantation. N. Engl. J. Med.306: 913 18 48. Ho,M., Jaffe, R., Miller, G., Breining, M. K., Dummer,J. S., Makowka.,L., Atchison, R. W., Karrer, F., Nalesnik, A., Starzl, T. E. 1988.Thefrequencyof Epstein-Barrvirus infection andassociated lymphoproliferativesyndromeafter transplantation andits manifestationin children. Transplantation45:719-27 49. Klein, G. 1975. Studies on the EpsteinBarr virus genomeand the EBV-determinednuclear antigen in humanmalignant disease. Cold Spring HarborSymp. Quant.Biol. 39:783-90 50. De The, G. 1980. Role of Epstein-Barr virus in humandiseases: Infectiousmononucleosis, Burkitt’s lymphoma andnasopharyngeal carcinoma.In Viral Oncoloyy, ed. G. Klein, pp. 769-97.NewYork: Raven 51. Nilsson, K., Klein, G. 1982. Phenotypic andcytogeneticcharacteristics of human B-lymphoid cell lines andtheir relevance for the etiology of Burkitt’s lymphoma. Adv. CancerRes. 37:319-80 52. Ehlin-Henriksson,B., Klein, G. 1984. Distinction betweenBurkitt lymphoma subgroups by monoclonalantibodies: relationships between antigen expression and type of chromosomal translocation. Int. J. Cancer33:459-63 53. Rowe, M., Rooney, C. M., Edwards, C. F., Lenoir, G. M., Rickinson,A. B. 1966.Epstein-Barrvirus status and turnour cell phenotypein sporadic Burkitt’s lymphoma. Int. J. Cancer37:367-73 54. Gregory, C. D., Tursz, T., Edwards, C. F., Tetaud, C., Talbot, M., Caillou, B., Rickinson,A. B., Lipinski, M.1987. Identification of a subset of normalB cells with a Burkitt’s lymphoma (BL)like phenotype.J. Immunol.139:313-18 55. Manolov,G., Manolova,Y. 1972. Marker band in one chromosome14 from Burkitt lymphomas.Nature 237:33-34 56. Zech, H., Haglund, U., Nilsson, K., Klein, G. 1976. Characteristic chromosomal abnormalities in biopsies and non-Burkitt lymphomas,lnt. J. Cancer 17:47-56

57. Bernheim,A., Berger, R., Lenoir, G. 1981. Cytogenetic studies on African Burkitt’s lymphoma cell lines: t(8; 14), t(2; 8) andt(8; 22) translocations. Cancer Genet.Cytogenet.3:307-15 58. Ohno, S., Babonits, M., Wiener, F., Spira, J., Klein, G., Potter, M. 1979. Nonrandom chromosome changes, involving the Ig gene-carryingchromosomes 12 and 6 in pristane-induced mouseplasmacytomas.Cell 18:1001-7 59. Klein, G. 1981. Therole of genedosage and g.enetic transpositions in carcinogenesis. Nature294:313-18 60. Klein, G. 1983. Specific chromosomal translocationsandthe genesis of B-cellderived tumorsin mice and men. Minireviews. Cell 32:311-15 61. Pear, W. S., Wahlstrrm, G., Nelson, S. F., Axelson,H., Szeles, A., Wiener, F., Bazin,H., Klein,G., Sumegi,J. 1988. 6-7 chromosomal translocation in spontaneously arising rat immunocytomas: evidencefor c-mycbreakpointclustering and correlation betweenisotypie expression andthe c-myctarget. MoLCell. Biol. 8:441-51 62. Sumegi,J., Spira, J., Bazin,H., Szpirer, J., Levan,G., Klein, G. 1983.Rat c-myc oncogene is located on chromosome-7 and rearranges in immunocytomas with t(6; 7) chromosomal translocation. Nature306: 497- 98 63. Wiener, F., Babonits, M., Spira, J., Klein, G., Bazin, H. 1982. Nonrandom chromosomal changes involving chromosomes-6 and 7 in spontaneous rat immunocytomas_._~/_nt~_J. Cancer29:43137 64. Fahrlander,P. D., Sumegi,J., Yang,J. Q., Wiener,F., Marcu,K. B., Klein, G. 1985. Activation of the c-myc oncogene by the immunoglobulin heavy-chain gene enhancer after multiple switch region-mediated chromosome rearrangements in a routine plasmacytoma. Proc. NatLAcad. Sci. USA82: 3746-50 65. Klein, G., Klein, E. 1985. Myeflgjuxtaposition by chromosomal translocations. ImmunoLToday 6:208-15 66. Potter, M., Mushinski,J. F., Mushinski, E. B., Brust, S., Wax,J. S., Wiener,F., Babonits, M., Rapp, U. R., Morse, H. C. III. 1987. Avian v-myc replaces chromosomaltranslocation in murine plasmacytoma-genesis. Science 235: 787-89 67. Adams,J. M., Hard.s, A. W., Pinkert, C. A., Corcoran, L. M., Alexander, W.S., Cory,S., Palmiter,R. D., Brinster, R. L. 1985. Thec-myc oneogenedriven by immunoglobulinenhancers induces

Annual Reviews HOW ONE THING HAS LED TO ANOTHER

68. 69.

70.


71.

lymphoid malignancy in transgenic nfice. Nature 318:53938 Klein, G. 1987. The approaching era of the tumor suppressor genes. Science 238: 1539-45 Harris, H., Miller, O. J., Klein, G., Worst, P., Tachibana, T. 1969. Suppression of malignancy by cell fusion. Nature 223:363~8 Klinger, H. P. 1982. Suppression of tumorigenicity. Sixth International workshop on human gene mapping. Cytogenet. Cell Genet. 32:68-84 Sager, R. 1985. Genetic suppression of tumor formation. Adv. Cancer Res. 44:

33

43~8 72. Stanbridge, E. J. 1987. Genetic regulation of tumorigenic expression in somatic cell hybrids. Adv. Viral Oncol. 6:83-87 73. Benedict, W. F. 1987. Recessive human cancer susceptibility genes (retinoblastoma and Wilms’loci). Adv. Viral Oncol. 7:19-34 74. Knudson, A. G. 1987. A two-mutation model human cancer. Adv. Viral Oncol. for 7:1-17 75. Henle, W., Henle, G. 1973. Epstein-Barr virus and infectious mononucleosis. New En,ql. J. Med. 288:263-64



Annual Reviews www.annualreviews.org/aronline


Ann. Rev. Immunol. 1989. 7:35-58 Copyright © 1989 by Annual Reviews Inc. All rights reserved

DECAY-ACCELERATING FACTOR: BIOCHEMISTRY, MOLECULAR BIOLOGY, AND FUNCTION Douglas M. Lublin* and John P. Atkinson~f Departmentsof Pathology* and Medicine*’~and the HowardHughes MedicalInstitute Laboratoriest, WashingtonUniversity School of Medicine,St. Louis, Missouri 63110 INTRODUCTION Complement componentsregularly becomeanchoredto host cells as well as to microbes.Thesecomponents mustbe allowedto promotethe reaction cascadeon microbesbut mustbe inhibited on self-tissue. Thus,it is critical for the host cell to downregulate the complementpathwayon its own membrane.To accomplishthis, cells possess several membraneproteins that can modulate complementcomponentsdeposited on their surface (reviewedin 1). Thestep at whichmuchof this control is directed is the formationof the enzymesthat cleave C3(so-called C3convertases). These bimolecular enzymecomplexesconsist of a cell-bound component to which a serine protease is noncovalentlyattached. Oncelarge amountsof C3 fragmentsare deposited, then the cell or microbemaybe ingested through interaction with phagocyticcells bearing C3receptors or lysed through engagement of the terminal (membraneattack complex) components. Consequently,muchof the regulatory activity is directed at modulating the C3-cleavingenzymes. Decay-acceleratingfactor (DAF)is one of these regulatory membrane proteins of the complement system. As its namestates, it can dissociate (decay-accelerate) both the classical and alternative pathwayC3convertases, as well as serve to preventtheir assembly.It is a widelydistributed membrane glycoproteinthat wasfirst describedas a functional entity over 35 0732-0582/89/0410-0035502.00

Annual Reviews


36

LUBLIN & ATKINSON

20 years ago(2, 3), althoughit wasnot isolated andpurified until 1981(4, 5). DAF is of special clinical interest becauseit is deficient in paroxysmal nocturnal hemoglobinuria (PNH)(6, 7), an acquired donal hemolytic disorder of man.Theincreased sensitivity to complement-mediated lysis of erythrocytes of PNHpatients is causally related to DAFdeficiency and is partially corrected by supplying the cells with DAF(8). Further, DAF has recently been shownto be a glycophospholipid-anchoredmembrane protein (9, 10), and this interesting structural feature mayfacilitate movement in the membraneand thereby permit DAFto downregulate complementactivation moreeffectively. In this chapter wereviewthe structure of DAFat the protein and DNA levels, and highlight its role in protecting cells fromdamageby.autologous complement.

COMPLEMENT SYSTEM Complement is a major effector system of the humoralimmuneresponse. This group of 20 plasma proteins and over a half-dozen cell membrane receptor and regulatory proteins serves for clearance or lysis of foreign particles or cells (reviewedin 11, 12). . Complement can be activated by either of two pathways.The classical pathwayis triggered by antigen-antibodycomplexes,specifically IgMand IgG. Thealternative pathwayis activated by foreign surfaces including bacteria, fungi, viruses, and tumorcells, as well as by immune complexes containing IgG, IgA, and IgE. Activation by either pathwayleads to production of a bimolecularcomplexdesignated C3convertase, C4b2afor the classical pathwayand C3bBbfor the alternative pathway,whichhas the ability to cleave C3 to C3b and C3a. Hence, the two pathways of activation convergeat the C3step. Thefact that C3is itself a component of C3convertase in the alternative pathwayresults in an amplification loop. At this C3step of the complement cascade, manyof the effector functions of the systemare brought into play. The released C3a, as well as C4aand C5a, are anaphylatoxins that serve as important mediators of inflammation.C3bis covalently boundto the target membrane or immune complex,promotingits clearance by phagocyticcells bearing C3breceptors (CR1). Finally, the terminal membrane attack complexcan be assembled on the target membrane,with the formation of transmembranechannels that can lead to cell death. This destructive potential of the complement systemrequires tight control so that host tissues are not damaged. In particular, there is a constant, low-level activation or tick-over of the alternative pathway,leading to

Annual Reviews


DECAY-ACCELERATING FACTOR

37

deposition of C3b on all surfaces. Activators such as bacteria cannot control the formation of C3 convertase, whereas host cells and other nonactivators rapidly inactivate the C3 convertase. The plasma proteins factor I (I), factor H (H), and C4-binding protein (C4bp) function to end. Additionally, cells possess a numberof membraneproteins to regulate complementthat is deposited on their surfaces; the largest group, focused on C3 convertases, includes CR1 (13), DAF, and membrane cofactor protein (MCP)(14). These plasma and membraneproteins inactivate convertase by dissociating its two components(decay-accelerating activity) or by serving as a eofactor for the proteolytic inactivation of C3b or C4b by plasma factor I (cofactor activity). Thus, the complementsystem remains focused on its proper (foreign) target.

DECAY-ACCELERATING FACTOR (DAF) Identification and Purification In 1969 Hoffmannreported that a substance remaining in the aqueous phase of an extract of humanerythrocyte stroma with n-butanol inhibited the complement-mediated hemolysis of antibody-coated sheep erythrocytes (2). He further showedthat this inhibition involved an acceleration in the decay of EAC14b2ato EAC14b(3). Over a decade later NicholsonWeller and colleagues purified an intrinsic membraneglycoprotein from guinea pig and human erythrocyte (E) stroma by butanol extraction, followed by sequential chromatography on DEAE-Sephacel, hydroxylapatite, phenyl-Sepharose, and trypan blue-Sepharose (4, 5). The protein was purified during this schemeby monitoring its ability to accelerate the decay of the classical pathway C3 convertase; hence it was named decayaccelerating factor. The purified componentwas a single chain membrane protein, DAF, with a Mr of 60,000 (guinea pig) or 70,000 (human) SDS-PAGEunder reducing conditions. Staining of human DAFwith periodic acid Schiff reagent demonstratedthat it is a glycoprotein. Biochemical

Activities

and Physiological

Roles

Several lines of evidence indicate that DAFprotects cells from damageby autologous complementproteins deposited on their surfaces. Specifically, DAFprevents the assembly of the C3 and C5 convertases of the classical or alternative pathways, which act as amplification steps in the complement cascade. DAF(all references are to humanDAFunless otherwise noted) was initially purified based on its ability to accelerate the spontaneous decay of the preformed classical C3 convertase, C4b2a(5). Blocking with antibodies to DAFshowed that it was responsible for this function on

Annual Reviews


38

LUBLIN & ATKINSON

intact E, for both the classical andalternative pathways,but that it lacked any cofactor activity for I-mediatedcleavageof C4bor C3b(7). Pivotal insights into the role of DAFcame from studies using DAF reincorporated into sheep E (15). These studies demonstratedthat DAF inhibits the formationof the C3andC5convertases;this effect wasreversible, as DAFdid not affect the structure of C4bor C3b.In addition, DAF only exerted this effect intrinsically, i.e., on C3convertasesassembledon the samecell as the DAF.Thefunctional activity of DAFis schematically shownin Figure 1. Anothergroup of investigators narrowedthe site of action further by showingthat DAFdoes not prevent the initial binding of C2or B to the cell (containing C4bor C3b,respectively), but that rapidly dissociates C2aor Bbfromtheir bindingsites, thus preventingthe assemblyof the C3convertase (16). The precise mechanismunderlying this interference with the C3convertase,and the specific bindingsites for DAFon the C3 convertase, are still unclear. One group used a homobifunctional cross-linking reagent to showan endogenousassociation of DAFwith C4b and C3bon the cell surface (17). Anotherinvestigation

DAF

Bb

Bb

DAF

Figure 1 Functional decay-accelerating activity as demonstrated by DAF.C3b is shown covalently attached to a cell surface through an ester or amide bond to a glycoprotein or glycolipid. DAFis present in the membrane,anchored by a glycophospholipid structure;, the four short consensus repeat domainsof DAFare shownby striped circles. (See later sections for details of these structures.) (Top) DAFbinds to a C3band prevents formation of the C3 convertase, C3bBb. This might be the most important role of DAFon a cell. (Bottom) DAFdissociates a preformed C3 convertase. It is not knownwhich short consensus repeat domain(s) binds to C3b. DAFshows the same activities with Cgb and the classical pathway C3 convertase, C4b2a.

Annual Reviews DECAY-ACCELERATING FACTOR

39

using fluid-phase competitive inhibition suggested that the primary interaction of DAFwith C3 convertases is with the C2a or Bb components (18). The experimental systems of these two groups are quite different, and the discrepancy in their results has not been resolved.


Biosynthesis

and Glycosylation

of DAF

DAFundergoes several posttranslational modifications to attain its final overall structure in the cell membrane.These have been elucidated for DAFby studying the biosynthesis of DAFin tissue culture and by chemical and enzymatic analysis of purified DAF. Weanalyzed the oligosaccharide structure of DAFby endo- and exoglycosidase digestions. The 74,000 Mr DAFfrom E was lowered ,-~ 3000 by endoglycosidase F, whereas endoglycosidase H had no effect, indicating one N-linked complex-type oligosaccharide (Figure 2) (19). Treatment with endo-~-N-acetylgalactosaminidase to remove O-linked oligosaccharides dropped the apparent Mr on SDS-PAGE by ,-~26,000, with two thirds of the shift due to sialic acid (Figure 2/. Thus, E DAFpossesses multiple, highly sialylated O-linked oligosaccharides (19). Similar results were obtained for DAFfrom peripheral blood granulocytes and cell lines such as HL-60, except for partial resistance to enzymatic removal of Olinked oligosaccharides. This suggests that the higher Mrof DAFon white blood cells versus E (see below) might arise from differences in O-linked glycosylation, but this point requires further investigation. A study of biosynthesis in the HL-60 cell line demonstrated two DAFintracellular species of 43,000 and 46,000 Mr(Figure 3) 09). Both species possess N-linked high-mannose unit, added cotranslationally, but no O-linked oligosaccharides. The lower Mrform is only seen in brief pulse labelings of 5-10 min; longer biosynthetic labelings only reveal the 46,000 Mrspecies and the mature form of DAF. Pulse-chase experiments indicate that the 43,000 Mr species is the earliest biosynthetic form of DAF,and that it undergoes an early posttranslational modification to a 46,000 Mrspecies (19). This change occurs before DAFenters the central region of the Golgi and does not appear to involve N- or O-linked glycosylation. The nature of this modificationis still unknown,as is its possible relation to the other major known DAFposttranslation modification, addition of a glycophospholipid anchor (see below). The 46,000 Mr species of DAFproceeds through the Golgi, where the one N-linked oligosaccharide is modified to a complex type, and the multiple O-linked oligosaccharides are added to produce the mature form of the protein seen on the cell surface. All forms of DAFhave a slower migration on SDS-PAGE under reducing, compared to nonreducing, conditions; this indicates the presence of intrachain disulfide bonds.


Annual Reviews

40 LUBLIN & ATKINSON


41

Similar results were found in a study of DAFbiosynthesis in the HeLa epithelial cell line (10). A single DAFprecursor of 48,000 Mr (equivalent to the 46,000 Mr DAFspecies discussed above) was identified using biosynthetic labelings of 30 min or greater. This DAFprecursor incorporated ethanolamine, a componentof the glycophospholipid anchor (10). This discussed later in more detail.


Glycophospholipid

Anchor

of Membrane DAF

Whenpurified DAFfrom the E membraneis added back to a cell suspension, it reincorporates in the membrane,apparently as an integral membrane protein, and displays functional activity (15). This property prompted an examination of the membrane-anchoring domain of DAF by two groups. DAFwas found to belong to a recently described class of membraneproteins (reviewed in 20, 21) whose carboxy terminus covalently attached to a glycophospholipid containing phosphatidylinositol (PI) inserted in the outer leaflet of the lipid bilayer. This anchoring was first shown by Davitz and colleagues, who demonstrated the release of DAFfrom peripheral blood cells following treatment with phosphatidylinositol-specific phospholipase C (PI-PLC) (9). Specifically, 60-80% of membrane DAFwas released from leukocytes by PI-PLC, although only 10% of E membrane DAF was removed. This partial resistance to PI-PLC has been found in other glycophospholipid-anchored proteins, and it mayrepresent structural modifications in the PI (22). These investigators and others found that DAFreleased by PI-PLChad lost its hydrophobiccharacter and its ability to reincorporate into cell membranes, and thus it could not intrinsically inhibit assembly of the C3 convertase on the cell surface (9, 10). However,this hydrophilic form of DAFcould still accelerate the decay of preformedC4b2aon a surface, albeit at a much reduced efficiency (10). A different hydrophilic fragment of DAF,also lacking the glycophospholipid anchor, demonstrated equal efficiency to purified membraneDAFin dissociating the fluid-phase C3 convertases

Figure 2 N- and O-linked oligosaceharide structure of erythrocyte membrane DAF. Erythrocytes were prepared from peripheral blood of a healthy humandonor and then were surface-labeled with ~zsI. DAFwas immunoprecipitated from detergent lysates and then was divided into equal aliquots for treatment with enzymes. The samples were analyzed by SDSPAGE(on a 9% gel run under reducing conditions) and autoradiography. Enzymetreatments are neuraminidase to remove sialic acid (lane 2), endo-ct-N-acetylgalactosaminidase remove O-linked oligosaccharides (lane 3), endoglycosidase H to remove high-mannose linked oligosaccharides (lane 4), endoglycosidase F to remove high-mannose and complex N-linked oligosaccharides (lane 5), or buffer alone (lanes 1 and 6). Reprinted from (19), permission.


Annual Reviews 42 LUBLIN & ATKINSON

Annual Reviews DECAY-ACCELERATING FACTOR O Prot ein~)-~-NH-CH~,~CH~ 0 "O-,P=O 0

~

an~l-2Man-l,

43

I-IO~-x OH "1£01"~

~Mandl-4GIcNHa~l-O~-~ OH


9 R~ R~ Figure 4 Common backbone structure of glycophospholipid anchor from trypanosomc variant surface glycoprotein (VSG)andrat brain Thy-1 glycoprotein. Thecompletestructure of the anchors from VSG(24) ~ad Thy-I (25) has been dcte~ined, and they contain common structure shownhere. The~-carboxy group of the carboxy-te~inal aminoacid of the proteia is linked,to aa ethanolaminephosphate.This is then attach~ to a glycan group containing three mannose(Man) residues and a glucosamine (GlcNH~) and ends in inositol phospholipiS.R~and R2arc fatty acids that form a di(acyl/alkyl)glyccrol moiety whichis inserted in the lipid bilaycr. Theglycangroup hasvarying additional substituents in the VSGand Thy-1 anchors. TheDAFglycophospholipid anchor has not beencompletely dete~ined,but the data (10, 26) are consistent with this structure. Figure is basedon data

in(24,25). (23). Thusthe functional site on DAFis separate from the glycophospholipidanchor. Thecompletestructureof the glycophospholipid anchorhas beendeterminedfor the trypanosomevariant surface glycoprotein (24) andfor Thy-1antigenfromrat brain (25), andthese anchorsshowan identical backbonewith variation in the side chain groups(Figure 4). Chemical analysis of the anchorfromE DAF,thoughless detailed, is consistent with these structures. Thesestudies by Medofandcoworkers(10, 26) demonstrated the presenceof ethanolamine andglucosamine(1.8 and0.8 Figure 3 Biosynthetic labeling of DAFin HL-60cells. HL-60cells (differentiated for 48 hr with vitamin D to increase DAFexpression) were biosynthetically labeled with [35S] methionine during a 10-rain pulse (P) followed by a 60 min chase (C) with unlabeled methionine. The detergent lysate from each condition was divided in half and was immunoprecipitated with either anti-DAF antibody (~DAF)or nonspecific (NS) control nonimmune rabbit Ig, and then was analyzed by SDS-PAGE (under reducing conditions) and fluorography. Another aliquot of HL-60cells was ~25I surface labeled, was immunoprecipitated with anti-DAF antibody and was analyzed by SDS-PAGEand autoradiography (lane 5). An arrow marks the position of mature DAF(80,000 Mr), and open and solid arrowheads mark the positions of the DAFspecies of 43,000 and 46,000 Mr, respectively. Reprinted from (19), with permission.

Annual Reviews


44

LUBLIN& ATKINSON

moles per mole of DAFprotein, respectively) in the carboxy terminus of the protein, as well as inositol (0.7 mole) and a mixture of saturated and unsaturated fatty acids (0.7 and 1.2 moles, respectively). In addition, analysis by thin layer chromatography of labeled anchor fragments released by nitrous acid deamination (which cleaves at the nonacetylated glucosamine)revealed the presence ofinositol phospholipids other than PI. This could explain the partial resistance to PI-PLCdescribed previously. The presence of the glycophospholipid anchor in DAFmight be expected to bestow some advantage in its functional role. Measurement of the lateral mobility of DAFin HeLacells using the fluorescence photobleaching recovery technique gave a mean diffusion coefficient of 1.61 +0.17 × 10-9 emE/s (27). This mobility is close to that exhibited membranelipids and an order of magnitude higher than most cell surface proteins. Physiologically, this increased mobility could enhancethe ability of a limited number of DAFmolecules to contact a large number of C3b and C4b fragments on the cell surface. Other possible roles for glycophospholipid anchors include serving as a means to release the protein from the cell membraneand transducing an intracellular signal (reviewed in 20, 21). AlthoughDAFis found in plasma and other body fluids, it is not knownwhether it arises from the membrane form via endogenousphospholipases. It has been found that antibodies to DAFinduce activation of human T cells, and removal of DAFby PIPLCabrogated the response (28). This has also been noted with two glycophospholipid-anchored murine proteins, Thy-1 (29) and T-cell-activating protein (30), but the role of the anchor itself in these processes not known. Sites

of Expression

and Alternate

Forms of DAF

DAFis present on virtually all peripheral blood cells: E, granulocytes, T and B lymphocytes, monocytes, and platelets (31, 32). The DAFmolecule from leukocytes has a 3000-9000 higher Mr than E DAF. Quantitation by iinmunoradiometric assay demonstrated 3000 DAFmolecules per E, 85,000 per neutrophil, 68,000 per monocyte, 33,000 per lymphocyte (B cell > T cell), and 2000 per platelet (31). Surface expression of DAF neutrophils can be doubled within minutes of exposure to activators such as N-formyl-methionyl-leucyl-phenylalanine; this occurs by translocation of an intracellular pool to the surface (33). Interestingly, DAFis absent on natural killer cells (34). DAFhas also been found on bone marrow mononuclear cells and erythroid progenitors (35). It is present on the epithelial surface of cornea, conjunctiva, oral and gastrointestinal mucosa, exocrine glands, renal tubules, ureter and bladder, cervical and uterine mucosa, and pleural, pericardial, and synovial serosa (36), as well as

Annual Reviews


DECAY-ACCELERATING

FACTOR

45

cultured umbilical vein endothelial cells (37). It is clear that DAFis widely distributed, and this supports its important role in controlling the complement system. Soluble forms of DAFhave been found in extracellular fluids and tissue culture supernatants. With the use of a two-site immunoradiometricassay (36), DAFantigen was detected in plasma, urine, tears, saliva, synovial fluid, and cerebrospinal fluid, with levels ranging from 40-400 ng/ml. Analysis by immunoprecipitation and Western blotting showed that the DAFfrom plasma, tears, and saliva had an apparent Mr of ~ 100,000, whereas that in urine had a Mr of 67,000, slightly lower than E membrane DAF. Urinary DAF was less hydrophobic than membrane DAF and did not inhibit the intrinsic assemblyof C3 convertases on the cell surface, but it could accelerate the decay of preformed C4b2awith an efficiency comparable to C4bp. Urinary DAFis thus similar to DAFreleased from membranesby PI-PLC (10). A species similar in size to urinary DAFwas also detected in the culture supernatants of the HeLaepithelial cell line (36), prompting the suggestion that urinary DAFis synthesized by the adjacent urethelium. Alternate forms of the membrane DAFmolecule have also been described. A larger variant, designated DAF-2, was detected on E membranes by Western blotting (38). DAF-2 possesses a 140,000 Mr and represents less than 10%of membraneDAF.This variant accelerates the decay of C3 convertase and shares with DAFthe ability to reincorporate into E membranes, suggesting the presence of the glycophospholipid anchor. The apparent Mr of DAF-2raises the possibility that it is a dimer of DAF, although neither reduction with 2-mereaptoethanol nor denaturation in SDS could separate DAF-2 into two components. The structure of DAF-2thus remains unexplained. Degradation fragments of membraneDAFhave been produced in vitro by treatment of DAFwith PI-PLC(9, 10, 23), a PI-specific phospholipase D from serum(39), or papain (10, 23). Interestingly, incubation of surfacelabeled E with leukocytes led to release of a fragment of equal size to a papain-derived fragment (23). It is unknownwhether any of these degradative processes are relevant in vivo. Blood

Group Antiyens

on DAF

It was recently demonstrated that the Cromer-related humanblood group antigens Cr" and Tc" reside on the DAFmolecule (40). Antibodies to ~ and Tc" recognized purified DAFon Western blots, and these antisera had reduced or absent reactivity with PNHE that lack DAF(see below). Moreover, cells of the rare Cromer-related null phenotype Inab did not react with antiserum to DAFby direct binding or Western blotting. The

Annual Reviews

46

LUBLIN& ATKINSON


reason underlying lack of expression of DAFin this null phenotypeis unknown. DAF In Other Species DAF’srole in protecting host tissues leads to the expectationthat a DAF or DAF-like molecule would exist in any species with a complement system. To date, only the DAFfromguinea pig (4) and rabbit (41, 42) have been isolated. The guinea pig DAFwas actually purified before humanDAF(4); the same investigators then used this schemeto isolate humanDAF(5). The guinea pig DAFhas a Mr of 60,000 on SDS-PAGE under reducing conditions. Decay-acceleratingactivity wasalso found on rabbit erythrocytes (41, 42). The rabbit DAFwas purified, yielding protein with a Mr of 66,000 on SDS-PAGE under nonreducingconditions (42). Rabbit DAFhas an amino acid composition resembling human DAF,and it can spontaneouslyand selectively reincorporate into sheep E, whichsuggests that it possessesa glycophospholipidanchorsimilar to human DAF. Cloning

of DAF eDNA

Twogroups have independently cloned DAFcDNAs(43, 44). Both used oligonucleotide probes based on the amino-terminalsequenceof immunoaffinity-purified E DAF.The clones were derived from libraries constructed with mRNA from either the HeLaepithelial cell line (43, 44) or the HL-60promyelocyticleukemiacell line (43). Thenucleotide and derived aminoacid sequencesfor DAFare shownin Figure 5. There is a single long open reading frame beginning with an initiation methionine codonand extending 1143bp. This is surroundedby 5’- and 3’-untranslated regions, the latter endingin a poly(A)track. Thededucedaminoacid sequencepredicts a protein of 381 aminoacids including a 34 aminoacid signal peptide. Starting at the aminoterminusof the matureprotein, there are four contiguous short consensus repeat (SCR)units of ~ 60 amino acids (Figure 6). EachSCRcontains four cysteines, .as well as conserved residues of proline, tryptophan, glycine, and several other aminoacids, and are homologousto domains found in other complementregulatory proteins, including CR1,CR2,C4bp, and H, as well as in several noncomplementproteins (reviewed in 45). The SCRsare followed by a 70aminoacid region that is rich in serine and threonine residues (45%). similar serine- and threonine-richregion, located just extracellular to the plasmamembrane,is the site of clustered O-glycosylation in the lowdensity lipoprotein receptor (46), with single O-linkedunits located other regions of the protein. This is consistent with the large amountof O-linkedoligosaccharide previously identified in DAF(19); the deduced


47

G~TGCGACTCGGCGGAGTCCCGGCGGCGCGTCCTTGTTCTAACCCGGCGCGCCATGACCGTCGCGCGGCCGAGCGTGCCCGCGGCGCTG 89 MTVARPSVPAAL-23 °34 ~~cCT~cTcGGGGAG~TGC~CCGGCTGCTG~TGCTGGTGCTGTTGTG~CTG~CGGCCGTGTGGGGTGA~TGTGGC~TT~CCCCAGATGTA 179 PLLG~LPRLLLLVLLCLpAVWGOCGLPPDV

Cc TGGCGAGAAGGACTCAGTGATCTGCC TTAAGGGCAGTC AATGGTCAGATATTGAAGAGTTC TGGAATCGTAGCTGCGAGGTGCCAACA P G E K D S V I C L K G S Q W $ D I E E F C N R S C E V P T AGGCT A AATTC TGCATCCCTCAAACAGCCT TATA TCACTCAGAATTATTTTCCAGTCGGTACTGTTGTGGAATATGAGTGCCGTCC^GGT 44~ R L N $ A S L K O P Y I T O N Y F P V G T V V E Y E C R P G


~"AC AGA AGAGAACC T TCTCTATCACCAAAACTAACT TGGCTTCAGAATTTAAAATGGTCCACAGCAGTCGAATTT TGTAAAAAGAAATCA $39 ¥ R R E P $ k $ P K k T C k O N L K W 8 T A V E F G K K K S 128 829 1~ GGGTA~AAATTATTTGG~T~~A~TTCTAGTTTTTGT~TTATTT~AGG~AG~T~TGT~~AGTGGAGTGA~~CGTTGCcAGAGTGcAGAGAA 71 g G Y K L F G $ T $ $ F C L I S G 8 8 V ~ W ~ D P L P E G R E 133 ATT T ATTGT~CAGCA~~A~CA~AAATTGACAATGGAATAATT~AAGGGGAACGTGA~~ATTATGGATATAGACAGTCTGTAACGTATGCA 809 ~ Y C P A P P O I D N G I I 0 G E R D H Y G ¥ FI O $ V ’r Y A 213 T GT AA T AAAGGA T TCACCA TGATTGGAGAGCACTCTA TTT ATTGTACTGTGAATAATGATGAAGGAGAGTGGAGTGGCCCACCACCTGAA 899 248 C N K O F T M I G E H $ I Y G T V N N D E G E W 8 Q P P P E TGCAGAGGAAAATCT~TAA~TT~~AAGGTC~~A~CAA~AGTT~AGAAA~~TAC~A~AGTAAATGTT~CAA~TACAGAAGT~T~A~cAACT 989 C R G K S L T $ K V P P T ¥ Q K P T T V N ¥ P T T E V S P T 278 T C T C AG A A AACCACCAGAA AA ACCACCACACCAAATGCTCAAGCAACACGGAGTACACCTGT TTCCAGGACAAGCAAGGATTTTCATGAA ’t079 $ Q K T T T K T T T P N A O A T R $ T P V 8 R T T K H F H E 30E ACA ACCCCAAA TAAAGGAAGTGGAACCAC TTCAGGTACTACCCGTGTTCTATCTGGGCAC AGGTGTTTCACGTTGAC AGGTTTGCTTGGG1 139 338 T T P N K G S G T T 8 G T T R |L L S (3 H T C F T L T G L L A CGCTAG T A ACCATGGGC T TGCTGACTTAGCCAAAGAAGAGTTAAGAAGAAAA TACACACAAGTATACAGACTGTTCCTAGTTTC TTAGA 125~ T L V T M (3 k L T I ° 347 1~49 C T T A T C T GCAT A T TGGATA A.A.A.T.A.A..ATGC^^T TGTGCTC T TCAT T TAGGATGC TTTCATTGTCTTTA^GATGTGTTAGGAATGTCAAC^G AGCAAGGAG^^^^^^GGGA~T~~T~GA^T~^~^TTGTT^G¢^~^~~TA~A~~T~TTGAAÂTAGÂGA^GTTG~^GÂTTG^G^~TGATT 1438 CC T T TCCT A A AAGTGTAAGAAAGC AT AGAGAT TTGTTCGT ATT TAGAATGGGATCACGAGGAAAAGAGAAGGAAAGTGAT TT TTTTCGAC 152~ A AGATC T GTAATGTTATTTCGAC TTATAAAGGA.A.A.T.A.A.A.A AATGAAAAAGAT TATTTGGATATCAAAAGGA.A.A..T.A.A.AAAGGGAATTGAGT 1§19 C TG T TC T AAGCAAAAT TGCTAAAGAGAGATGAACGACATTATAAAGTAATCTTTGGCTGTAAGGCA TTTTCATCTTTCCTTGGGGTTGGG 170~ A AA A T AT T TT AAAGGTAAA ACATGGTGGTGAACCAGGGGTGT TGATGGTGATA ^GGGAGGAATATAGAATGAAAGAGTGAATGTTCGTTT 1799 1089 GT TGCAC A_A.A.T.A~G_AGT TTGGAAAAAGGC TO TQAAAGGTGTCTTC TT TGACTTAATQTCTTTAAAAGTATCCAGAGATACTACAATATTAA CATAAGAAAAG^TTATATATTAT TTCTGÂTCGAG^TGTCC^TAGTCAAATTTGTAAATCTTATTCTTTTGTAATATTTATTTATATTT^ 1979 T T T ATGACAGTGAACATTCTGATTTTACATGT~AAAcAAGAAAAGTTGAAGAAGATATGTGAAGAAAAATGTATTTTTCCTAAATAGAA-A20~ " A_A_A_TGATCCCATTTTTTGGTAAAAAAAAAAA

2101

Figure 5 Nucleotide and derived amino acid sequences of DAFeDNA. The nucleotide sequence is numbered from the most 5’ nucleotide, and the derived amino acid sequence, numberedfrom the first aminoacid of the mature protein, is shownbelow, using single-letter codes with an asterisk denoting the stop codon. The single N-glycosylation site is marked with an arrow, an S/T-rich region (probable site of O-linked glycosylation) is markedwith an underline, a carboxy-terminal hydrophobic region (replaced posttranslationally with glycophospholipid anchor) is boxed, and potential polyadehylation signals are markedwith dashed lines. Data from (43; D. M. Lublin, unpublished).


Annual Reviews

48 LUBLIN & ATKINSON

Annual Reviews


DECAY-ACCELERATING FACTOR

49

protein sequence also showsone site for N-linked glycosylation, again as expected (19). The deduced protein structure ends in a hydrophobic 24-amino acid segment. The series of basic residues (that act as a stop anchor sequence) and the cytoplasmic tail that are present in polypeptid¢-anchored membrane proteins are not seen in DAF. However, this carboxy-terminal hydrophobic peptide is similar to extension peptides encoded by the cDNAsfor other glycophospholipid-anchored membraneproteins such as the trypanosome variant surface glycoprotdns (47) and Thy-1 (48). These extension peptides of 17-31 amino acids are removedposttranslationaily and the carboxy-terminal glycophospholipid anchor is attached (reviewed in 20, 21). A similar processing presumably leads to the attachment of the DAFglycophospholipid anchor, perhaps with the hydrophobic extension peptide acting as a transient membraneanchor in the endoplasmic reticulum. Studies in HeLacells demonstrated that the major intracellular DAFprecursor of 48,000 Mr incorporated ethanolamine, a component of the glycophospholipid anchor (10). In addition, partitioning of proteins into the detergent phase of Triton X-114 (49) was used to show that treatment of this DAFprecursor species with PI-PLC removedits hydrophobic domain (D. M. Lublin, unpublished). These results demonstrate that the glycophospholipid anchor is already attached to this DAFprecursor. Each of the SCRdomains in f12 glycoprotein I (50) and several that have been studied in C4bp(51) have the four cysteine residues disulfide bonded as cys 1-cys 3 and cys 2-cys 4. This pattern most likely holds for the SCRsin DAF. Secondary structure predictions for DAFbased on the methods of Chou-Fasman(52) or Robson (53) indicate predominantly structure. Similar predictions for many of the SCRsalong with spectroscopic data on H, suggest that the SCRforms a compact domain with antiparallel fl-sheets (54). At this point the structure of DAFat the cDNAand protein level can be summarized. The cDNAsequence shown in Figure 5 is organized into structural regions in Figure 7. Translation of this sequence into protein, coupled with the co- and posttranslational modifications discussed previously, results in a structural model of the membraneDAFglycoprotein (Figure 7). The signal for attachment of the glycophospholipid anchor to a given protein is still unclear. However, two groups have made and expressed mutant cDNAscontaining the carboxy-terminal segment of DAFattached to the amino-terminal segment of another protein (55, 56). These mutant cDNAs,whenexpressed in transfected cell lines, led to the production of a membrane protein that was anchored by a glycophospholipid, thus

Annual Reviews 50

LUBLIN

& ATKINSON

Consensus Repeats

s’--I

I 1 I 2 I 3 I

! Signal Peptide

Ser- and ThrRich Region

4 I

II

l HydrophobicRegion

A n

NH 2


4 Repeat Domains

O-Linked Carbohydrate Domain

CHO-O~ CHO-O.O-CHO CHO-O,~O-CHO , -O-CHO

Glycophospholipid Anchor

~

Membrane

Figure 7 Structure of DAF cDNA and DAF membrane glycoprotein. (Top) This DAF cDNA structure corresponds to the sequence in Figure 5. Coding regions are shown by boxes, and 5’- and 3’-untranslated regions by lines. The ser-and thr-rich region is the probable site of most of the extensive O-linked glycosylation of DAF. The carboxy-terminal hydrophobic region of DAFis replaced posttranslationally with a glycophospholipid anchor. (Bottom) This model of the membrane DAF glycoprotein is based on the above cDNA and the biochemical in the text.

studies

of DAF glycosylation

and the glycophospholipid

anchor discussed

establishing that the carboxy-terminal aminoacids (either 37 [55] or 91 [56] aminoacids) containthe signal for attachment of a glycophospholipid anchor.Thenatureof that signal is still unknown. TheeDNA for DAFdetects several bandson Northernblot analysis of mRNA fromvarious cell lines. The majorspecies are reported as 2.0 and2.7 kb (43) or 1.5 and2.2 kb (44), apparentlysimplyreflecting standardizationdifferences. Thesetwo species of mRNA are productsof alternative polyadenylation(44). Relative levels of DAFmRNA in the HeLa,HL-60,and HSB-2cell lines correlated with the levels of DAF proteindetectedby immunoradiometric assay (43), suggestingtissue-specific transcriptionalcontrol of DAF expression. One group has found a second class of DAFcDNAclones which containeda 118 bp insertion near the endof the codingregion(44). The resulting frameshiftpredicts a longer encodedprotein (440 aminoacids

Annual Reviews


DECAY-ACCELERATING

FACTOR

51

including signal peptide) that wouldhave a hydrophilic carboxy terminus. The authors speculate that the 118 bp represents an unspliced intron. A probe based on this sequence detected a minor species of DAFon Northern analysis of HeLacell RNA.Transfection of these two types of cDNAinto Chinese hamster ovary (CHO)cells resulted in the production of DAF immunoreactive material, but only the cDNAending in the hydrophobic extension peptide produced surface DAF.It was suggested that the spliced and unspliced species of cDNAencode membrane and secreted DAF species, respectively. However, subsequent work has cast doubt on this hypothesis (D. M. Lublin, unpublished; V. Nussenzweig, personal communication). CHOcells transfected with the spliced (regular or hydrophobic) DAFcDNAproduce both DAFattached to the membrane by glycophospholipid anchor and a secreted form of DAFapproximately 5000 lower in Mr. Antibodies raised against the carboxy-terminal hydrophilic peptide (encoded only by the alternate, unspliced cDNA)did not recognize the soluble DAFspecies in HeLacell culture supernatants. The physiological relevance of this alternate DAFcDNAspecies, along with the origin of the secreted form of DAF,remains unclear. DAF Gene Southern blot analysis of human DNAshows that the DAFgene is~ approximately 35 kb in length (57-59). The relatively simple pattern generated from restriction digests suggests that DAFis a single copy gene, and this was supported by hybridizations with DAF-specific oligonucleotide probes (59). The structure and organization of the DAFgene is not yet known. Three RFLPshave been identified in the DAFgene: two for the enzyme Hind III and one for BamHI (58, 59). All are located in the noncoding region of the gene. The chromosomallocation of the DAFgene is on the long arm of humanchromosome1, band q3.2. This was derived from analysis of a panel of hamster x humansomatic cell hybrids and by in situ hybridization of the DAFcDNAto human metaphase cells (57). The same result was obtained by segregation analysis of the DAFRFLPs in families that are informative for segregation of alleles at the CR1, C4bp, and H loci (58). The latter three complementproteins were already known to be located at thg regulator of complementactivation (RCA)gene cluster at lq3.2, so the DAFgene is added to this group. Furthermore, recombinations within the RCAlocus demonstrated that DAFmaps closer to the CR1/C4bploci than to the H locus. Subsequent detailed mapping of the RCAgene cluster by pulsed-field gel electrophoresis has shown the order of the genes to be MCP-CR1-CR2-DAF-C4bp, located within an 800-kb segment of DNAon the long arm of chromosome1 (60, 61; N. S. Bora, J. P. Atkinson, submitted).

Annual Reviews 52

LUBLIN & ATKINSON


PAROXYSMAL NOCTURNAL HEMOGLOBINURIA (I’NJ) The physiological role of DAFhas been elucidated by studies of the disease paroxysmal nocturnal hemoglobinuria (PNH), an acquired clonal disorder of hematopoietic stem cells (reviewed in 62). Circulating blood cells that arise from the affected pluripotent stem cell show increased complement sensitivity, leading to episodes of hemolysis that are a hallmark of PNH. By use of a quantitative complementlysis sensitivity assay (63), three populations of blood cells are detected in the circulation of PNHpatients: PNHI cells with normal sensitivity, PNHII cells with 3-5-fold increased sensitivity, and PNHIII cells with 15-25-fold increased sensitivity to lysis by complement. PNHII and PNHIII cells show increased C3b uptake (64); in addition, PNHIII cells showincreased susceptibility to bystander or reactive lysis by the terminal complement pathway components C5b-9 (65). Twogroups of investigators found that the affected E of PNHpatients lack DAF(6, 7). This was also found for PNHleukocytes and platelets (31, 66). Furthermore,this defect was causally related to increased sensitivity to complement, since reincorporation of purified DAFinto these E normalized their C3b uptake and partially corrected the sensitivity (8). Specifically, incorporation of DAFinto PNHII cells completely corrected their complement sensitivity, whereas PNHIII cells with DAFincorporated, although taking up C3b normally, still had markedly increased susceptibility to reactive lysis (67). NormalE possess a membraneprotein, homologousrestriction factor (HRF) (68, 69), which can inhibit transmembrane channel formation by the terminal complement components. PNHE have been shown to lack HRF, and thus to have increased susceptibility to reactive lysis, which can be corrected by reincorporation of purified HRFinto the cell membranes (70). Thus, both DAFand HRF appear to be critical in vivo for protection of host cells from damageby autologous complement. Investigations of DAFhave also shed light on the underlying lesion in PNH. Southern and Northern analysis utilizing leukocytes from PNH patients revealed a normal DAFgene and mRNA transcripts (59). Indeed, affected cells of PNHpatients lack not only DAF, but also acetylcholinesterase (71), alkaline phosphatase (72), lymphocyte functionassociated antigen 3 (73), Fc receptor type III (74), and HRF(70). these proteins except HRFhave been directly shown to be anchored to the cell membraneby a glycophospholipid anchor [Table 1 in reference (21); for FcRIII see 74-76]. Purified HRFcan spontaneously reincorporate into cell membranes(70), suggesting that it also possesses this form

Annual Reviews DECAY-ACCELERATING

FACTOR

53

membraneanchor. The fact that these otherwise unrelated proteins are all absent in a clonal disorder strongly suggests that the lesion.in PNHmust involve their only commonelement, the glycophospholipid anchor (9). The nature of this defect in the pathwayfor biosynthesis or attachment of the anchor structure is unknown,as is the reason for the existence of more than one type of affected cell, PNHII versus PNHIll, in this clonal disorder.


CONCLUSIONS The work reviewed here has provided a picture of DAFwith its polypeptide backbone, extensive glycosylation, and glycophospholipid anchor (Figure 7). The bulk of the extracellular part of the protein is organized as four contiguous short consensus repeat domains, thus putting DAFin a family of complement (and noncomplement) proteins that share this 60 amino acid structural unit. Furthermore, DAFis genetically linked to a subgroup of these (CR1, CR2, H, C4bp, and MCP)that are all C3 regulatory receptor proteins located on the long arm of humanchromosome1, band q3.2, at the regulator of complementactivation locus. This group probably arose by a process of gene duplication from an ancestral C3 binding protein. The study of the organization and expression of the DAFgene will help in understanding the evolution of this gene family and the role that DAFplays in inflammation. The glycophospholipid anchor, which is unique to DAFamong this group, might help DAFserve its function by increasing its mobility in the cell membrane.Future workproviding further characterization of the fine structure of the glycophospholipid anchor in DAFand assessing its contribution to the function of DAFwill be important not only for an understanding of the role of DAFin the complement system, but also for elucidating the disease PNHin which there is a defect in the synthesis or attachment of this anchor. DAFdown regulates complement activation by preventing C3 convertase formation on cell surfaces as well as by dissociating preformed convertases. The physiological importance of DAFis suggested by its wide tissue distribution and is further highlighted by the increased complement sensitivity and lysis of cells from PNHpatients, which lack DAF.DAFis present on the membranesof invading inflammatory cells and on the tissue at sites of inflammation, and it protects both from bystander lysis due to complementactivation. More generally, there is a constant low-level activation of the alternative pathway of complement; this nonspecific initiation is targeted to and amplified on foreign surfaces and avoids host tissues because the latter can control complementon their surfaces. Although the plasma proteins H and C4bp can control the C3 convertase,

Annual Reviews


54

LUBLIN & ATKINSON

this appearsto be importantmainlyfor the fluid phase, whereason cell surfaces the membrane proteins are critical. DAFdoes not permanently modifythe C3bor C4battachedto the cell surface, so they can reforman active C3convertase.Thus, membrane proteins that can act as cofactors for the permanent inactivationof the C3convertaseby the serine protease factorI also play a role in protectinghost cells. Theimportant proteinin this regardis probablymembrane cofactor protein (MCP) (14, 77), which sharesthe widetissue distribution of DAF,whereasthe principalrole of CR1,whichhas both decay-acceleratingactivity andcofactor activity, probablyinvolves processing immune complexes.In this way, DAFand other complement regulatoryproteins not only protect host tissues from inflammation initiated by antibody,but also serve to target the constant low-level activation of the alternative pathwayof complement awayfrom self-tissues andtowardnonself(78). ACKNOWLEDGMENT The authors

thank Avtar Khalsa for excellent

secretarial

assistance.

Literature Cited 1. Holers, V. M., Cole, J. L., Lublin, D. M., Seya, T., Atkinson, J. P. 1985. Human C3b- and C4b-regulatory proteins: A new multi-gene family. Immunol. Today 6:188-92 2. Hoffmann, E. M. 1969. Inhibition of complement by a substance isolated from humanerythrocytes. I. Extraction from human erythrocyte stromata. Immunochemistry 6:391M03 3. Hoffmann, E. M. 1969. Inhibition of complement by a substance isolated from humanerythrocytes. II. Studies on the site and mechanism of action. Immunochemistry 6:405-19 4. Nicholson-Weller, A., Burge, J., Austen, K. F. 1981. Purification from guinea pig erythrocyte stroma of a decay-accelerating factor for the classical C3 convertase, C4b,2a. J. Immunol. 127: 203539 5. Nicholson-Weller, A., Burge, J., Fearon, D. T., Weller, P. F., Austen, K. F. 1982. Isolation of a humanerythrocyte membrane glycoprotein with decay-accelerating activity for C3convertases of the complement system. J. Immunol. 129: 184-89 6. Nieholson-Weller, A., March, J. P., Rosenfeld, S. I., Austen, K. F. 1983. Affected erythrocytes of patients with paroxysmal nocturnal hemoglobinuria

are deficient in the complement regulatory protein decay-accelerating factor Proc. Natl. Acad. Sci. USA 80:5066-70 7. Pangburn, M. K., Schreiber, R. D., Miiller-Eberhard, H. J. 1983. Deficiency of an erythrocyte membrane protein with complement regulatory activity in paroxysmal nocturnal hemoglobinuria. Proc. Natl. Acad. Sci. USA 80:5430-34 8. Medof,M. E., Kinoshita, T., Silber, R., Nussenzweig, V. 1985. Amelioration of lytic abnormalities of paroxysmal nocturnal hemoglobinuria with decay-accelerating factor. Proc. Natl. Acad. Sci. USA 82:2980-84 9. Davitz, M. A., Low, M. G., Nussenzweig, V. 1986. Release of decay-accelerating factor (DAF)from the cell membrane by phosphatidylinositol-specific phospholipase C (PIPLC). Selective modification of a complement regulatory protein. J. Exp. Med. 163: 115061 10. Medof, M. E., Walter, E. I., Roberts, W. L., Haas, R., Rosenberry, T. L. 1986. Decay accelerating factor of complementis anchored to cells by a C-terminal glycolipid. Biochemistry 25:6740-47 11. Mfiller-Eberhard, H. J., Miescher, P. A., eds. 1985. Complement.Berlin: SpringerVerlag 12. Ross, G. D., ed. 1986. lmmunobiologyof

Annual Reviews


DECAY-ACCELERATING FACTOR the ComplementSystem. An Introduction for Research and Clinical Medicine. London: Academic Press 13. Fearon, D. T. 1980. Identification of the membraneglycoprotein that is the C3b receptor of the human erythrocyte, polymorphonuclear leukocyte, B lymphocyte and monocyte. J. Exp. Med. 152:20-30 14. Seya, T., Turner, J., Atkinson,J. P. 1986. Purification and characterization of a membraneprotein (gp45-70) which is cofactor for cleavage of C3b and C4b. J. Exp. Med. 163:837-55 15. Medof, M. E., Kinoshita, T., Nussenzweig, V. 1984. Inhibition of complement activation on the surface of cells after incorporation of decay-accelerating factor (DAF) into their membranes. J. Exp. Med. 160:1558-78 16. Fujita, T., Inoue, T., Ogawa, K., Iida, K., Tamura, N. 1987. The mechanismof action of decay-accelerating factor (DAF). DAFinhibits the assembly C3 convertases by dissociating C2a and Bb. J. Exp. Med. 167:1221-28 17. Kinoshita, T., Medof, M. E., Nussenzweig, V. 1986. Endogenousassociation of decay-accelerating factor (DAF)with C4b and C3b on cell membranes. J. Immunol. 136:3390-95 18. Pangburn, M. K. 1986. Differences between the binding sites of the complement regulatory proteins DAF, CR1, and factor H on C3 convertases. J. Immunol. 136:2216-21 19. Lublin, D. M., Krsek-Staples, J., Pangburn, M. K., Atkinson, J. P. 1986. Biosynthesis and glycosylation of the human complement regulatory protein decay-accelerating factor. J. Immunol. 137:1629-35 20. Low, M. G. 1987. Biochemistry of the glycosyl-phosphatidylinositol membrane protein anchors. Biochem. J. 244: 1 13 21. Low, M. G., Saltiel, A. R. 1988. Structural and functional roles of glycosylphosphatidylinositol in membranes. Science 239:268 75 22. Roberts, W. L., Kim, B. H., Rosenberry, T. L. 1987. Differences in the glycolipid membrane anchors of bovine and human erythrocyte acetylcholinesterases. Proc. Natl. Acad. Sci. USA84: 7817-21 23. Seya, T., Farries, T., Nickells, M., Atkinson, J. P. 1987. Additional forms of human decay-accelerating factor (DAF). J. Immunol. 139:1260~57 24. Ferguson, M. A. J., Homans, S. W., Dwek, R. A., Rademacher, T. W. 1988. Glycosyl-phosphatidylinositol moiety

55

that anchors trypanosoma brucei variant surface glycoprotein to the membrane. Science 239:753-59 25. Homans, S. W., Ferguson, M. A. J., Dwek, R. A., Rademacher, T. W., Anand, R., Williams, A. F. 1988. Complete structure of the glycosyl phosphatidylinositol membrane anchor of rat brain Thy-1 glycoprotein. Nature 333:269-72 26. Walter, E. I., Roberts, W. F., Rosenberry, T. L., Medof, M. E. 1987. Analysis of fatty acids and inositol in the membrane anchor of human erythrocyte decay accelerating factor (DAF). Fed. Proc. 46:772 (Abstr.) 27. Thomas, J., Webb, W., Davitz, M. A., Nussenzweig, V. 1987. Decay accelerating factor diffuses rapidly on HeLaAEcell surfaces. Biophys. J. 51: 522a (Abstr.) 28. Ritter, A. R., Davis, L. S., Patel, S. S.~ Atkinson, J. P., Lipsky, P. E. 1988. An antiserum to decay-accelerating factor (DAF) activates human T cells. Fed. Proc. 47:A871 (Abstr.) 29. Gunter, K. C., Malek, T. R., Shevach, E. M. 1984. T cell activating properties of an anti-Thy-I monoclonal antibody. Possible analogy to OKT3/Leu-4. J. Exp. Med. 159:716-30 30. Rock, K. L., Yeh, E. T. H., Gramm,C. F., Haber, S. I., Reiser, H., Benacerraf, B. 1986. TAP,a novel T-cell-activating protein involved in the stimulation of MHC-restricted T lymphocytes. J. Exp. Med. 163:315-33 31. Kinoshita, T., Medof, M. E., Silber, R., Nussenzweig, V. 1985. Distribution of decay-accelerating factor in the peripheral blood of normal individuals and patients with paroxysmal nocturnal hemoglobinuria. J. Exp. Med. 162: 7592 32. Nicholson-Weller, A., March, J. P., Rosen, C. E., Spicer, D. B., Austen, K. F. 1985. Surface membraneexpression by humanblood leukocytes and platelets of decay-accelerating factor, a regulatory protein of the complement system. Blood 65:1237~,4 33. Berger, M., Medof, M. E. 1987. Increased expression of complement decay-accelerating factor during activation of humanneutrophils. J. Clin. Invest. 79:214-20 34. Nicholson-Weller, A., Russian, D. A., Austen, K. F. 1986. Natural killer cells are deficient in the surface expression of the complement regulatory protein, decay-accelerating factor (DAF). Immunol. 137:1275-79 35. Moore, J. G., Frank, M. M., Miiller-

Annual Reviews


56

LUBLIN & ATKINSON

Eberhard, H. J., Young,N. S. 1985. organization, and regulation of the Decay-acceleratingfactor is present on complementgenes. Ann. Rev. Immunol. paroxysmal nocturnal hemoglobinuria 6:161-95 erythroid progenitors and lost during 46, Davis, C. G., Elhammer,A., Russell, erythropoiesisin vitro. J. Exp.Med.162: D. W., Schneider, W.J., Kornfeld, S., 1182-92 Brown,M. S., Goldstein, J. L. 1986. 36. Medof,M. E., Walter, E. I., Rutgers, Deletion of clustered-O-linked carboJ. L., Knowles, D. M., Nussenzweig, hydratesdoes not impairfunction of low V. 1987. Identification of the compledensity lipoprotein receptor in transment decay-accelerating factor (DAF) fected fibroblasts. J. Biol. Chem.262: on epithelium and glandular cells and 2828-38 in bodyfluids. J. Exp. Med.165: 848- 47. Boothroyd,J. C., Paynter,C. A., Cross, G. A. M., Bernards,A., Borst, P. 1981. 37. Asch,A. S., Kinoshita,T., Jaffe, E. A., Variant surface glycoproteins of TryNussenzweig, V. 1986. Decay-accelpanosomabrucei are synthesized with erating factor is present on cultured cleavable hydrophobicsequencesat the humanumbilical vein endothelial cells. carboxy and amino termini. Nucleic J. Exp. Meal. 163:221-26 Acids Res. 9:4735-43 38. Kinoshita,T., Rosenfeld,S. I., Nussen- 48. Tse, A. G. D., Barclay,A. N., Watts,A., zweig, V. 1987. ,6 high m.w. form of Williams, A. F. 1985. A glycophosphodecay-acceleratingfactor (DAF-2)exhilipid tail at the carboxylterminusof the bits size abnormalities in paroxysmal Thy-1 glycoprotein of neurons and nocturnal hemoglobinuriaerythrocytes. thymocytes.Science 230:1003-8 J. Immunol.138:2994-98 49. Bordier, C. 1981. Phase separation of 39. Davitz, M. A., Hereld, D., Shak, S., integral membrane proteins in Triton XKrakow,J., Englund, P. T., Nussen114 solution. J. Biol. Chem.256: 563zweig, V. 1987. A glycan-phospha67 tidylinositol-specific phospholipaseDin 50. Lozier, J., Takahashi, N., Putnam,F. humanserum. Science 238:81-84 W.1984. Completeaminoacid sequence 40. Telen, M.J., Hall, S. E., Green,A. M., of humanplasma fl2-glycoprotein I. Moulds,J. J., Rosse,W.F. 1988.IdentiProc.Natl. Acad.Sci. USA8 I: 3640-44 fication of humanerythrocyte blood 51. Janatova,J., Reid, K. B. M., Willis, A. group antigens on decay-accelerating C. 1988. Involvementof the disulfide factor (DAF)and an erythrocyte phenobonds in the structure of complement type negative for DAF.J. Exp. Med. regulatory protein C4bp.Fed. Proc. 47: 167:1993-98 A1832(Abstr.) 41. Horstmann,R. D., M/Jller-Eberhard,H. 52. Chou,P. Y., Fasman,G. D. 1978. PreJ. 1986. Demonstration of C3breceptordiction of the secondary structure of like activity and of decay-accelerating proteins fromtheir aminoacid sequence. factor-like activity on rabbit erythroAdv. Enzymol. 47:45-148 cytes. Eur. J. ImmunoL 16:1069-73 53. Gamier,J., Osguthorpe,D. J., Robson, 42. Sugita, Y., Uzawa, M., Tomita, M. B. 1978. Analysis of the accuracy and 1987. Isolation of decay-accelerating implications of simple methodsfor prefactor (DAF)from rabbit erythrocyte dicting the secondarystructure of globumembranes.J. Immunol. Methods104: lar proteins. J. Mol.Biol. 120:97-120 123G0 54. Perkins,S. J., Haris, P. I., Sim,R. B., Chapman, D. 1988.A study of the struc43. Medof,M.E., Lublin, D. M., Holers, V. M., Ayers,D. J., Getty, R. R., Leykam, ture of humancomplementcomponent J. F., Atkinson,J. P., Tykocinski,M.L. factor H by fourier transform infrared 1987. Cloning and characterization of spectroscopy and secondary structure cDNAsencoding the complete sequence averaging methods.Biochem.27: 4004of decay-accelerating factor of human 12 complement.Proc. Natl. Acad.Sci. USA 55. Caras, I. W.,Weddell,G. N., Davitz, M. 84:2007-11 A., Nussenzweig,V., Martin, D. W.Jr. 44. Caras, I. W., Davitz, M. A., Rhee,L., 1987. Signal for attachment of a Weddell,G., Martin,D. W.Jr., Nussenphospholipid membrane anchor in zweig,V. 1987.Cloningof decay-acceldecay-acceleratingfactor. Science238: eratingfactor suggestsnoveluse of splic1280-83 ing to generatetwo proteins. Nature325: 56. Tykocinski, M. L., Shu, H. K., Ayers, 54%49 D.J., Walter,E. I., Getty,R. R., Groger, 45. Campbell,R. D., Law,S. K. A., Reid, R. K., Hauer, C. A., Medof,M. E. 1988. K. B. M., Sim, R. B. 1988. Structure, Glycolipid reanchoringof T-lymphocyte

Annual Reviews


DECAY-ACCELERATING FACTOR surface antigen CD8using the 3’ end sequence of decay-accelerating factor’s mRNA.Proc. Natl. Acad. Sci. USA 85: 3555 59 57. Lublin, D. M., Lemons, R. S., LeBeau, M. M., Holers, V. M., Tykocinski, M. L., Mcdof, M. E., Atkinson, J. P. 1987. The gene encoding decay-accelerating factor (DAF)is located in the complement-regulatory locus on the long arm of chromosome 1. J. Exp. Med. 165: 1731-36 58. Rey-Campos,J., Rubinstein, P., Rodriguez de Cordoba, S. 1987. Decay-accelerating factor. Genetic polymorphism and linkage to the RCA(regulator of complementactivation) gene cluster in humans. J. Exp. Med. 166:246-52 59. Stafford, H. A., Tykocinski, M. L., Lublin, D. M., Holers, V. M., Rosse, W. F., Atkinson, J. P., Medof, M. E. 1988. Normal polymorphic variations and transcription of the decay-accelerating factor gene in paroxysmal nocturnal hemoglobinuria cells. Proc. Natl. Acad. Sci. USA 85:880-84 60. Rey-Campos,J., Rubinstein, P., Rodriguez de Cordoba, S. 1988. A physical mapof the humanregulator of complement activation gene cluster linking the complement genes CR1, CR2, DAF, and C4BP. J. Exp. Med. 167:664-69 61. Carroll, M. C., Alicot, E. M., Katzman, P. J., Klickstein, L. B., Smith, J. A., Fearon, D. T. 1988. Organization of the genes encoding complement receptors Type 1 and 2, decay accelerating factor, and C4-binding protein in the RCA locus on humanchromosome1. J. Exp. Med. 167:1271-80 62. Rosse, W. F., Parker, C. J. 1985. Paroxysmal nocturnal haemoglobinuria. Clin. Haematol. 14:105~5 63. Rosse, W. F., Dacie, J. V. 1966. Immune lysis of normal human and paroxysmal nocturnal hemoglobinuria (PNH) red blood cells. I. Thesensitivity of PNH red cells to lysis by complementand specific antibody. J. Clin. Invest. 45:736-48 64. Logue, G. L., Rosse, W. F., Adams, G. P. 1973. Mechanismof immunelysis of red cells in vitro. Paroxysmalnocturnal hemoglobinuriacells. J. Clin. Invest. 52: 1129-37 65. Packman, C. H., Rosenfeld, S. I., Jenkins, D. E. Jr., Thiem, P. A., Leddy, J. P. 1979. Complementlysis of human erythrocytes: Differing susceptibility of two types of paroxysmal nocturnal hemoglobinuria cells to C5b-9. J. Clin. Invest. 64:428-33 66. Nicholson-Weller, A., Spicer, D. B.,

57

Austen, K. F. 1985. Deficiency of the complementregulatory protein, "decayaccelerating factor," on membranesof granulocytes, monocytes, and platelets in paroxysmal nocturnal hemoglobinuria. N. Engl. J. Med. 312:1091-97 67. Medof, M. E., Gottlieb, A., Kinoshita, T., Hall, S., Silber, R., Nussenzweig,V., Rosse, W. F. 1987. Relationship between decay accelerating factor deficiency, diminished acetylcholinesterase activity, and defective terminal complementpathway restriction in paroxysmal nocturnal hemoglobinuria erythrocytes. J. Clin. Invest. 80:165-74 68. Zalman, L. S., Wood, L. M., MfillerEberhard, H. J. 1986. Isolation of a human erythrocyte membrane protein capable of inhibiting expression of homologous complement transmembrane channels. Proc. Natl. Acad. Sci. USA 83:6975-79 69. Sch6nermark, S., Rauterberg, E. W., Shin, M. L., L6ke, S., Roelcke, D., H~inch, G. M. 1986. Homologousspecies restriction in lysis of humanerythrocytes: a membrane-derivedprotein with C8-binding capacity functions as an inhibitor. J. Immunol. 136:1772 76 70. Zalman, L. S., Wood,L. M., Frank, M. M., Miiller-Eberhard, H. J. 1987. Deficiency of the homologousrestriction factor in paroxysmal nocturnal hemoglobinuria. J. Exp. Med. 165:572 77 71. Auditore, J. V., Hartmann, R. C. 1959. Paroxysmal nocturnal hemoglobinuria II. Erythrocyte acetylcholinesterase defect. Am. J. Med. 27:401-10 72. Lewis, S. M., Dacie, J. V. 1965. Neutrophil (leucocyte) alkaline phosphatase in paroxysmal nocturnal haemoglobin~ uria. Br. J. Haematol. 11:549-56 73. Selvaraj, P., Dustin, M. L., Silber, R., Low, M. G., Springer, T. A. 1987. Deficiency of lymphocyte functionassociated antigen 3 (LFA-3) in paroxysmal nocturnal hemoglobinuria. J. Exp. Med. 166:1011-25 74. Selvaraj, P., Rosse, W. F., Silber, R., Springer, T. A. 1988. The major Fc receptor in blood has a phosphatidylinositol anchor and is deficient in paroxysmal nocturnal haemoglobinuria. Nature 333:565~i7 75. Simmons, D., Seed, B. 1988. The Fcv receptor of natural killer cells is a phospholipid-linked membraneprotein. Nature 333:568~0 76. Huizinga, T. W. J., van der Schoot, C. E., Jost, C., Klaassen, R., Kleijer, M., von demBorne, A. E. G. Kr., Roos, D., Tetteroo, P. A. T. 1988. The PI-linked

Annual Reviews 58

LUBLIN & ATKINSON


receptor FcRIII is released on stimu(MCP):Evidence for inclusion in the lation of neutrophils. Nature333: 667multi-gent family of complement-regu69 latory proteins. J. Exp. Med.168: 18177. Lublin, D. M., Liszewski,M. K., Post, 94 T. W., Arce, M. A., Le Beau, M.M., 78. Atkinson, J. P., Farries, T. 1987. Rebentisch,M.B., Lemons,R. S., Seya, Separationof self from non-self in the complementsystem. Immunol.Today 8: T., Atkinson,J. P. 1988.Molecularcloning and chromosomallocalization of 212-15 human membrane cofactor protein





Ann. Rev. Immunol. 1989. 7: 59-76 Copyright © 1989 by Annual Reviews Inc. All rights reserved

HETEROGENEITY OF MAST CELLS AND PHENOTYPIC CHANGE BETWEEN SUBPOPULATIONS Yukihiko Kitamura Division of Cancer Pathology, Biomedical Research Center, Osaka University Medical School, Nakanoshima4-3-57, Kita-ku, Osaka, 530 Japan

INTRODUCTION Mast cells are not seen in routine histological sections stained with hematoxilin and eosin, but a considerable numberof mast cells are found in various tissues that are properly fixed and stained with dyes such as toluidine blue and Alcian blue. The substances in granules that stain specifically with these dyes are proteoglycans, whichare negatively charged and thought to form complexes with positively charged proteases and histamine. Mast cells have high affinity IgE receptors on their surface, and the immunologicalactivity of mast cells is mediated through these IgE receptors (1). Binding of antigens to IgE molecules results in the formation of linkages between IgE receptors, and then the release of the granules themselves or chemical mediators in the granules (2). This process constitutes an important step in the immediatehypersensitivity reaction that occurs in allergic diseases such as urticaria, bronchial asthma, and allergic rhinitis. In addition to havinga role in allergic diseases, mastcells have a physiological role as an effector of host defense mechanismsin intestinal helminth infection (3-5) and dermaltick infestation (6, Although some of cytochemical and functional characteristics of mast cells are commonalso to basophils, these two types of cells can be distinguished with an electron microscope(8). Moreover,their differentiation 59 07324)582/89/04104)059502.00

Annual Reviews


60

KITAMURA

processesare different, in spite of the fact that both mast cells and basophils originate from multipotential hematopoietic stem cells. Like neutrophils and eosinophils, basophils complete their differentiation within the bone marrow,then circulate in the blood, and finally function and die in tissues (8-10). In contrast, undifferentiated precursors of mast cells leave the bone marrow, migrate in the blood, invade tissues, and then proliferate and differentiate into mast cells (9, 10). The life span of at least somemast ceils appears to be muchlonger than that of basophils. In contrast with basophils which lose proliferative potential after differentiation, some morphologically differentiated mast cells can proliferate extensively (9, 10). Mastcells in various tissues differ in their phenotype. The heterogeneity has either biological or clinical significance, and it has been investigated from morphological, biochemical, immunological, and functional points of view (reviewed in 11-14). In the present review, I describe the heterogeneity and the differentiation processes of mast cells and attempt to interpret the heterogeneity from the perspective of their unique differentiation process. HETEROGENEITY Morphology Maximow (15) was probably the first to recognize that certain mast cells in the rat intestinal mucosawere atypical in their staining characteristics and differed from those of the mast cells observed in other anatomical sites. Enerbfick (16) greatly extended these observations and defined conditions of fixation and histochemical staining that discriminated between such atypical or mucosal mast cells (MMC) and the connective tissue-type mast cells (CTMC)of the skin, peritoneal cavity, and muscularis propria of the digestive canal, amongother sites. In the mid-1960swhen Enerb~ck started his investigations, the concentration of mast cells in the intestinal mucosa was considered to be very low (16), since granules of MMC are not fixed with the most commonlyused fixative, 10%formalin. However, after adequate fixation and staining, Enerb~ick found that in rats the gastrointestinal mucosais one of the tissues richest in mast cells. Carnoy’s solution (containing methanol, chloroform, and acetic acid) or a combination of 0.6%formaldehyde and 0.5% acetic acid is necessary for the fixation of MMC (16). Whenappropriately fixed sections are stained, rat MMC stain blue with Alcian blue, while the granules of rat CTMC stain red with safranin. MMC are smaller than CTMC,are more variable in shape, and as a rule contain fewer granules of more variable size and shape. Although MMC are never found in the epithelium in normal rats,

Annual Reviews


MAST CELL DIFFERENTIATION

61

numerous MMC are observed between the individual epithelium cells in infections of intestinal parasites (16). Berberine sulfate is a fluorescent dye suitable for identification of CTMC in rats and mice. In these animals this dye forms a strong fluorescent complexwith heparin in CTMC granules (17, 18). The prior digestion with heparinase abolished the fluorescence, but digestion by chondroitinase ABCdid not (19). Since the proportion of CTMC of which granules stain red with safranin is lower in mice than in rats, I prefer berberine sulfate to safranin for staining mouseCTMC.In spite of the presence of heparin, granules of humanmast cells do not stain with berberine sulfate. In addition to tissues from which mast cells are harvested, the species and age of animals also influence the phenotypesof mast cells. Even within the peritoneal cavity of rats, the morphologyof mast cells changes with age (20). The nomenclature of MMC and CTMC is based on observations of rat tissues, but the samecriteria are applicable to mast cell populations of mice. Althoughthe heterogeneity of mast cells is detectable in humans, the distinction between MMC and CTMC is not so clear as that observed in rats and mice. Therefore, I will hereafter confine the terms, MMC and CTMC,to mast cell populations of’rats and mice. Cultured

Mast Cells

The characteristics of individual mast cells can be investigated with morphological techniques. However,for biochemical and functional investigations, it is necessary to obtain pure suspensionsof mast cells belonging to each subpopulation. Both CTMCand MMC can be purified in rats, but purification of MMC from mouse intestinal mucosa has not been accomplished. For this reason mice have not been used for the study of mast cell heterogeneity until recently. Developmentof a simple and easy technique to culture mast cells from mouse hematopoietic cells changed the situation. Lzrge numbers (> 108) of bone marrow~lerived cultured mast cells (BMCMC) can be generated as virtually homogeneous populations or as clones. Attempts to culture mast cells have been carried out at least since 1963. Ginsburg & Sachs (21) cultured thymus cells on a feeder layer composed of mouseskin fibroblasts and observed the development of mast cells. Ishizaka et al obtained similar results by using rat thymus, and they demonstrated the presence of histamine and IgE receptors in mast cells developing in such a system (22, 23). About20 years after the first paper of Ginsburg&Sachs (21), six groups of investigators independently reported that mast cells developed when hematopoietic cells of mice were cultured in suspension with growth factors (24-29). Five of these six groups obtained the growth factor from stimu-

Annual Reviews


62

KITAMURA

lated T cells, whereas Nagaoet al (26) used the culture mediumof a mouse myelomonocytic leukemia cell line (WEHI-3) as the source of growth factor. Ihle et al (30) purified a mast cell growth factor from the culture medium of WEHI-3cells and designated it interleukin 3 (IL-3). Fung et al (31) isolated cDNAencoding IL-3 by using WEHI-3cells, whereas Yokota et al (32) isolated cDNAof a mast cell growth factor from a mouse T cell line stimulated by concanavalin A and demonstrated that this mast cell growth factor was identical to IL-3. Smith &Rennick (33) identified another mast cell growth factor in the culture mediumof the same T-cell line that was used for the isolation of cDNAencoding IL-3; Lee et al (34) isolated and characterized a cDNA clone that encodedthis mast cell growth factor. Unexpectedly, the nucleotide sequence of the cDNAwas identical with that of cDNAencoding the IgG-l-inducing factor isolated by Nomaet al (35). Moreover, the terminal amino acid sequence of the peptide inferred from this cDNAwas in agreement with the amino acid sequences of B-cell stimulating factor-1 (BSF-1), which were partially determined by Grabstein et al (36) and Paul, Ohara and coworkers (37). Nowthis peptide is designated as BSF-1/IL-4 (hereafter IL-4). Pure mast cell suspensions can be obtained from rat hematopoietic cells by a similar method (38). However, when humanhematopoietic cells are cultured in the presence of humanIL-3, basophils but not mast cells develop (39). Mediators PROTEOGLYCANS Differences

in histochemical reactions of mast cell granules are attributed to differences of their proteoglycans. Proteoglycans consist of a protein core and attached side chains of glycosaminoglycans. Although CTMC in the peritoneal cavity of rodents (rats and mice) and mast cells in the humanlung contain heparin proteoglycans, the relative molecular mass (Mr) of heparin proteoglycans is considerably different (600,000 to 1,000,000 in rodents and 60,000 in humans)(11-14, 40). Somesubpopulations of mast cells don’t synthesize heparin proteoglycan but do synthesize chondroitin sulfate proteoglycans. Rat MMC purified from the small intestine infected with a helminth, Nippostrongylus brasiliensis (41, 42), and rat BMCMC (43) incorporate 35S into a highly sulfated chondroitin sulfate. The type of proteoglycan contained in or synthesized by mouseMMC has not been determined, since it is difficult to collect enough of these cells for chemical analysis. However, mouse BMCMC have been shown to incorporate 3~S into an oversulfated chondroitin sulfate (44).

Annual Reviews MASTCELLDIFFERENTIATION 63


Synthesizing a type of proteoglycan does not necessarily meanthat the cell contains only this type of proteoglycan. Although rat peritoneal CTMC incorporate 3~S exclusively into heparin proteoglycan, staining by a monoclonal antibody showed that someof the cells contained chondroitin sulfate as well (45). PROTEASES Apparent heterogeneity of proteases has been observed in rat mast cell populations (ll-14). Rat CTMC contain a chymotrypsin-like neutral protease (rat mast cell protease I, RMCP-I)and carboxypeptidase A, an enzyme that cleaves C-terminal aromatic amino acids, whereas MMCand BMCMC of rats contain another chymotrypsin-like neutral protease (rat mast cell protease II, RMCP-II)(43, 46, 47). Although RMCP-!and RMCP-II have substantial homology in their amino acid sequences, antibodies to each of them do not cross-react (48). During infection by intestinal parasites, the serumlevel of RMCP-IIrises remarkably (49). Carboxypeptidase A is detectable in both CTMC and BMCMC of mice, but the concentration is more than 100 times higher in CTMC than in the BMCMC(50). AMINES Histamine is contained in mast cells of all the mammalianspecies so far examined. However, the concentration of histamine differs among mast cell subpopulations (11-14). Rat peritoneal CTMC contain 15 pg histamine per cell whereas rat MMC contain only 1-2 pg per cell. Mouse peritoneal CTMC also contain l0 pg of histamine per cell, and mouse BMCMC 0.1 pg per cell (19). Moreover, a significant difference in histamine concentration is detectable among peritoneal CTMCharvested from different inbred strains of mice (T. Nakano, U. Waki, J. Fujita, A. Yamatodani, H. Asai, Y. Kitamura, unpublished data). Unlike histamine, serotonin showsa remarkable species difference (1114). Both CTMCand MMCof rats contain serotonin, and CTMCand BMCMC of mice contain serotonin as well. In contrast, serotonin is not present in mast cells isolated from humanlung. ARACHIDONIC ACIDMETABOLITES Stimulated mast cells produce pharmacologically active metabolites from arachidonic acid. Since arachidonic acid metabolites are produced by various types of cells other than mast cells, purification of mast cells is a prerequisite for the chemical analysis. Purified rat peritoneal CTMC almost exclusively metabolize arachidonic acid through the cyclo-oxygenase pathway; the chief product is prostaglandin D2 (13, 51). On the other hand, rat MMC produce leukotriene C4, leukotriene B4, and prostaglandin D2 (51). Arachidonic acid metabolism in mouse BMCMC is similar to that in rat MMC,that is, they

Annual Reviews 64

KITAMURA

generate and release leukotriene C4 along with some leukotriene B4 and small amounts of prostaglandin D2 (52).


Surface

Antigen

Katz et al developed rat monoclonal antibodies that are useful for distinguishing mast cell subpopulations of mice (53). One antibody, designated B1. l, recognizes a neutral glycosphingolipid, globopentaosylceramide (Forssman glycolipid), on the surface of mouse peritoneal CTMC, but fails to detect this glycosphigolipid in mouse BMCMC. Since mouse BMCMC express the direct precursor of globopentaosylceramide, namely, globotetraosylceramide (globoside) (54), Katz et al proposed mouse BMCMC differed from mouse CTMCin either lacking or having an inactive form of the glycosyltransferase nccdcd to synthesize the more complex globopentaosylceramide 02). Functional

Heterogeneity

Antigens and anti-IgE antibodies that aggregate between IgE molecules induce histamine release from both CTMCand MMC of rats. Whereas the secretagogue compound48/80 does induce histamine release from rat CTMC,it fails to induce release from rat MMC.Similarly, compound 48/80 activates mouse CTMCbut does not activate mouse BMCMC (13, 14). Inhibitors of release of mast cell mediators have practical significance as antiallergic drugs. Theophylline and disodium cromoglycate prevent antigen-stimulated histamine release by rat peritoneal CTMC but not by rat MMC (13, 14). Heterogeneity

in Human Mast Cells

Staining with dyes is not so useful for discriminating between human mast cell subpopulations as it is in the cases of rats and mice, but the immunohistochemicalstaining of tryptase (trypsine-like neutral protease) and chymase(chymotrypsin-like neutral peptidase) distinguishes two mast cell populations in humantissues (55). Mastcells in the skin contain both tryptase and chymase(TC-positive), whereas mast cells in the lung and the intestinal mucosacontain only tryptase (T-positive) (56, 57). Functional heterogeneity is also found between TC-positive and T-positive mast cells. TC-positive mast cells purified from the skin release histamine in response to compound48/80, morphine, and substance P (58, 59), but T-positive mast cells from the lung and intestinal mucosado not (60). Extensive characterization of humanmast cells purified from various tissues has been performed. Characteristics first reported in rodent mast cells have been found in human mast cells as well. For example, the

Annual Reviews MAST CELL

DIFFERENTIATION

65

presence of chondroitin sulfate proteoglycan in purified humanlung mast cells was recently demonstrated (61, 62). Although the terms CTMC and MMC are not suitable for humanmast cell populations, heterogeneity does exist amongthem.


ORIGIN

AND DIFFERENTIATION

The generation of mast cells from hematopoietic cells was first shownby using giant granules of beige (C57BL/6-bg-’/bg~, Chediak-Higashi syndrome) mice as a marker. Transplantation of bone marrow cells from C57BL/6-bg~/bg~ to irradiated normal (C57BL/6-+/+) mice resulted development of beige-type mast cells with giant granules (63). Twoother mutant mice are also useful for investigations of mast cell differentiation. v (64) or WCB6Ft-SI/SId (65) are deficient Mice of either WBB6F~-W/W in mast cells. In the first description, only CTMC were counted, but later the depletion of MMC was also reported (66). The absence of mast cells ~ mice and to is attributed to a defect of precursor cells in WBB6F1-W/W a defect of the tissue environment necessary for differentiation of mast d mice (64, 65). cells in WCB6F~-SI/SI There are various types of cells in the bone marrow. In mylaboratory, we demonstrated that mast cells are the progeny of multipotential hemato~ mice as recipients of the cells poietic stem cells by, using WBB6F~-W/W to be tested. In the first experiment, we enucleated spleen colonies which were produced by individual stem cells in irradiated mice. The intravenous transfer of cell suspensions from a single spleen colony resulted in the ~ mice (67). development of mast cells in tissues of WBB6F1-W/W the second experiment, various in vitro hematopoietic cell colonies were produced from the bone marrow of normal WBB6F~-+/+mice, and then cell suspensions from individual colonies were directly injected into the ~ mice (68). The appearance of a mast cell colony skin of WBB6Fl-W/W at the injection site was evidence for the presence of mast cell precursors in the injected cell suspension. More than 40%of the mixed colonies tested, which contained erythroblasts, megakaryocytes, neutrophils, and macrophages,contained mast cell precursors as well (68). Wedid not find morphologically identifiable mast cells in the mixed colonies. However, Nakahata & Ogawa(69) reported mixed colonies that did contain mast cells. This observation also indicates that mast cells can originate from the multipotential hematopoietic stem cells. Most of the progeny of multipotential stem cells such as erythrocytes, platelets, neutrophils, eosinophils, and basophils leave the hematopoietic tissue after they differentiate. However,mast cells do not complete their differentiation in hematopoietic tissue. Nomast cells are detectable in the

Annual Reviews


66

KITAMURA

blood, but when blood mononuclearcells are plated in methylcellulose in the presence of IL-3, large mast cell colonies containing more than 500 mast cells appear (70, 71). The mast cell precursors that produce such large colony (CFU-Mast) resemble lymphoid cells, by light microscopy (71). Although the density of mouse peritoneal CTMC is significantly greater than that of small lymphocytes, the density of CFU-Mastis comparable to that of small lymphocytes(71). CFU-Mastinvade the connective or mucosal tissues, proliferate, and differentiate into morphologically identifiable mastcells (72, 73). Since the increase and the subsequent decrease of MMC at the site of helminth infection occur within three weeks(16, 49), the life span of most MMC appears to be limited to 1 or 2 weeks. In contrast, CTMC in the skin (74) and peritoneal cavity (71) of mice have a long life (probably than one year), and some humanlung mast cells produce granules again after degranulation (75). Although most progenies of multipotential hematopoietic stem cells lose the proliferative potential whenthey differentiate fully, some morphologically identifiable CTMC have an appreciable proliferative potential. Sonodaet al (76) identified murine peritoneal CTMCunder the phase contrast microscope, picked up a single CTMC vW/W with the micromanipulator, and injected it into the skin ofWBB6F1mice. Mast cell colonies containing about 2000 mast cells developed in 10 of 168injection sites (76). Proliferation of peritoneal CTMC was also confirmed by using an in vitro culture technique. Nakahata et al (77) and Hamaguchiet al (78) plated purified peritoneal CTMC of mice in methylcellulose; about 20% to 50%CTMC showed clonal growth. In this condition, colonies produced by differentiated CTMC are smaller than colonies produced by CFU-Mast (71, 77). Although CFU-Mastrequired only IL-3 for colony formation (77), both IL-3 and IL-4 are necessary for development of colonies from morphologically identifiable CTMC (78). Colony-forming activity is not limited to peritoneal CTMC; the recent result of Kanakuraet al suggests that CTMC in the skin of mice can also produce mast cell colonies (79). REGULATION OF DIFFERENTIATION

PROLIFERATION/

IL-3 elaborated by helper/inducer T cells appears to induce the mast cells in the intestinal mucosaof mice and rats. Such a mast cell accumulation does not occur in T cell depleted mice (80) and rats (81). On the other hand, a significant increase of MMC (82) and subsequent repulsion of helminth, Strongyloides ratti (83), occurred whenT cell~tepleted nude mice were injected with IL-3. Since IL-4 stimulates the proliferation of both

Annual Reviews


67

mast cells and B cells and the synthesis of IgE (37), it is possible that newly formed mast cells may be armed by lgE antibodies synthesized in their vicinity.


Fibroblast

Dependent Proliferation

Although the proliferation and differentiation mediated by IL-3 and IL-4 is the only regulatory mechanismunderstood at the molecular level, other regulatory mechanismsless clearly defined may also influence the proliferation and differentiation of mast cells. For example, the concentration of CTMC in the skin of genetically T cell-depleted nude mice is comparable to that of normal congenic mice, implicating the presence of a regulatory mechanismwithout T-cell involvement (84, 85). Fujita et al (86) recently investigated the role offibroblasts on the proliferation of mast cells. Without the addition of IL-3 and IL-4, mouse BMCMC may continue to proliferate on a monolayerof the NIH-3T3fibroblast cell line. Since the fibroblasts synthesize neither IL-3 nor IL-4, and since direct contact of mast cells with fibroblasts is necessary for the proliferation, the supportive effect of fibroblasts appears to be mediated neither by knowngrowth factors (such as IL-3, IL-4) nor by unknowndiffusible substances (86). Fujita et al analyzed the mechanismof mast cell deficiency in WBB6F1W/V~ (86, 87) and WCB6FI-SI/SU (J. Fujita, H. Onoue, Y. Ebi, H. Nakayama, Y. Kanakura, Y. Kitamura, unpublished data) mice by cocul~ turing of BMCMC and fibroblast cell lines. T cells of both WBB6F~W/W and WCB6F~-SI/SUmice may produce IL-3 and IL-4 (J. Fujita, H. Onoue, Y. Ebi, Y. Kitamura, unpublished data). BMCMC develop when bone ~ or marrow cells or blood mononuclear cells of either WBB6F~-W/W WCB6F~-SI/SU mice are cultured in the mediumcontaining IL-3. Therefore, IL-3 and IL-4 do not appear to be involved in the actions of the W and S1 mutantgenes. In addition, several fibroblast cell lines were screened; all six fibroblast cell lines derived from mouseembryos (including the NIH-3T3 cell line) supported the growth of BMCMC derived from WBB6Ft-+/÷ mice. In contrast, none of these mouse embryo-derived v mice (87). fibroblast cell lines supported BMCMC of WBB6F~-W/W synchronizing BMCMC at the G~ phase of the cell cycle, the defect of ~ BMCMC W/W was further characterized as a failure to transit G~ and enter the S phase upon contact with fibroblasts. This suggests that W gene product expressed on the surface of BMCMC is mandatory for the fibroblast-dependent proliferation (87) (Figure BMCMC derived from WCB6F~-SI/SIa mice were maintained as well by the NIH-3T3 cell line as were BMCMC of WCB6F~-+/+mice, indicating the normal function of SI/SU BMCMC. Then, 3T3 fibroblast cell ~ and the control WCB6F lines were established from WCB6F1-SI/SI ~-+/+

Annual Reviews 68

KITAMURA

(+)

Proliferation (-)

(-)

= W Gene Product


- SI Gene Product Figure i Scheme showing possible expression of W gene product on the surface of BMCMC and SI gene product on the surface of fibroblasts. These molecules are indispensable for flbroblast-dependent proliferation of BMCMC.

embryos. The 3T3 cell lines derived from +/+ embryos induced the G~ to S transition in synchronized ÷/+ BMCMC upon contact, but the 3T3 cell lines derived from Sl/Sl a embryosdid not. This suggests that Sl gene product expressed on the surface of 3T3 fibroblasts is indispensable for the fibroblast-dependent proliferation of BMCMC (J. Fujita, H. Onoue, Y. Ebi, H. Nakayama, Y. Kanakura, and Y. Kitamura, unpublished data) (Figure 1). Suppression

of Differentiation

Unpleasant symptoms accompany overproduction of mast cells such as are observed in patients with urticaria pigmentosa, a benign skin tumor of mastcells. This raises the possibility that suppressionas well as induction of mast cell differentiation may be an important normal regulatory mechanism. Recently Kanakura et al (71) investigated a mechanism for the suppression of mast cell differentiation. Peritoneal CTMC of mice were eradicated by intraperitoneal injection of distilled water, and the regeneration process was analyzed by estimating the changes in numbers of CFU-Mastand morphologically identifiable mast cells. CFU-Mastincreased after the injection of distilled water. Whenpurified peritoneal CTMC were injected two days after the water injection, the increase of CFU-Mastdid not occur (71). In the peritoneal cavity of WBB6F~+/ mice that had been lethally irradiated and rescued by bone marrowinjection of C57BL/6-bg~/b9~ mice, CFU-Mastwere of the bg~/b9J type, but the morphologicallyidentifiable mast cells were of + / + type. Injection of distilled water into the radiation chimeras resulted in development of bg~/bgS-type CTMC with giant granules. The presence of morphologically identifiable CTMC appears to suppress the invasion of CFU-Mastfrom the blood and to inhibit the differentiation of CFU-Mastinto morpho-

Annual Reviews MAST CELL DIFFERENTIATION

69

logically identifiable mast cells (71). The concentration of CTMC in tissues is too low for the CTMC to be in direct contact with each other. Therefore, the inhibitory effect of differentiated CTMC on invasion and differentiation of CFU-Mastis considered to be mediated by diffusible substance(s) rather than the direct contact. PHENOTYPIC


Experiments

CHANGES Using

WBB6F~- W/Wv Mice

As already mentioned, mast cell populations are heterogeneous. However, the interrelation between subpopulations had not been systematically investigated until Nakanoet al (19) demonstrated the phenotypic change from BMCMC to CTMC.BMCMC were used since they can be obtained as a homogeneouspopulation and since some of the characteristics of BMCMCare shared by MMC. Nakano et al (19, 88) cultured BMCMC from the bone marrow WBB6F~-+/+mice and transferred them into the peritoneal cavity of v mice. At various times after genetically mast cell~deficient WBB6FIW/W the intraperitoneal transfer, mast cells were recovered from the peritoneal cavity. The density of the original BMCMC is significantly less than that of CTMCharvested from the peritoneal cavity of WBB6F~-+/+mice, but the density increased and becamecomparable to that of the peritoneal CTMC 10-30 weeks after the transfer (88). The recovered mast cells acquired the electron microscopic features of CTMC (19). Furthermore, the histamine content increased more than 20-fold after the transfer. Although the starting BMCMC did not stain with berberine sulfate, the recovered mast cells stained with this fluorescent dye. This suggests that BMCMC acquired the ability to synthesize and store heparin proteoglycan after the intraperitoneal transfer (19, 88). Recently, the change from BMCMC to CTMCwas confirmed by biochemical and immunochemical criteria (40). BMCMC derived from WBB6F~-+/+mice synthesized 350,000-Mr proteoglycan which contained 55,000-Mr chondroitin sulfate glycosaminoglycans. Fifteen weeks after the intraperitoneal transfer, transferred mast cells were recovered ~ mice; these from the peritoneal cavity of the recipient WBB6F~-W/W mast cells synthesized 650,000-Mr proteoglycan containing 105,000-Mr heparin glycosaminoglycans. Although globopentaosylceramide ~Forssman glycolipid) recognized by the B I.I rat monoclonal antibody was weakly expressed on the surface of BMCMC, it was strongly evident 15 ~ mice weeks after the transfer into the peritoneal cavity of WBB6FI-W/W (40). The phenotypic change occurs in the opposite direction as well (89).

Annual Reviews


70 KITAMURA Whenpurified peritoneal CTMC of WBB6F~+ / + mice were cultured in methylcellulose with IL-3 and IL-4, about 25%of the CTMC formed colonies, all of whichcontainedboth berberine sulfate-positive and berberine sulfate-negative mast cells. Whenthese mast cells weregrownin suspensionculture, they generated populations that were 100%berberine sulfate-negative, and that synthesized predominantlychondroitinsulfate proteoglycans (89). Whenthese MMC-like cultured mast cells derived from WBB6F~-÷/+ peritoneal CTMC were injected into the peritoneal v mice, the adoptively transferred mast cell popucavity of WBB6F~W/W lation became100%berberine sulfate-positive. In methylcelluloseculture, these "second generation peritoneal CTMC" formedclonal colonies containing both berberine sulfate-positive and -negative mast cells. Thus, clonal mastcell populations, initially derived from a single peritoneal CTMC, exhibited multiple and bidirectional alterations between CTMClike and MMC-like phenotypes(89). The fate of CTMC derived from WBB6F~-+/+ mice was investigated v mice (90). After the injection of in the stomachwall of WBB6F~-W/W single CTMC, mast cells mayappear both in the mucosaand the muscularis propria. Mastcells that appearedin the mucosashowedthe histochemical and electron microscopic features of MMC, whereas the cells that appeared in the museularis propria showedthe features of CTMC (90). In Vitro Experiments Galli et al (91) attempted to induce a phenotypic change in murine BMCMC by adding sodium butyrate to the culture medium.The electron density of cytoplasmicgranules greatly increased, and the histaminecontent increased by up to 50-fold. However,the BMCMC treated with sodiumbutyrate incorporated 35S into chondroitin sulfate proteoglycan but not into heparin proteoglycan (91). So far no one has succeeded switchingthe type ofproteoglycansynthesis by simpleaddition of a defined substance to the medium of a suspensionculture. Stevens and coworkers cultured mouse BMCMC with the Swiss albino mouse-skin~lerivedfibroblast cell line and demonstratedthat the BMCMC acquired CTMC-likephenotype (92). The cells becamesafraninpositive, showedincreased amountsof histamine and carboxypeptidaseA, and synthesized considerably moreheparin proteoglycan (50, 92). These investigators also comparedthe arachidonic acid metabolites produced by BMCMC before and after the coculture with the fibroblasts. When stimulated with antigen, BMCMC cocultured with fibroblasts produce more leukotriene B4 and prostaglandin D2 than did the starting BMCMC (52). Theseresults cannot be explained by the selective expansionof

Annual Reviews



71

particular subpopulation on the fibroblasts, since comparableresults have been obtained by using cloned BMCMC as well (93). Stevens and coworkers added IL-3 to their coculture system (92, 93). Mouse embryo-derived cell lines support the proliferation of BMCMC without IL-3, whereas somefibroblast cell lines derived from the tissues of adult mice do not (87). It is possible that the fibroblast cell line used Stevens and coworkers (which was derived from the skin of Swiss albino mice) does not support BMCMC by itself. However, the presence of IL-3 is not mandatoryfor the phenotypic change, since Fujita et al reported a similar phenotypic change when mouse BMCMC were cocultured with the NIH-3T3 fibroblast cell line in the absence of IL-3 (86). CONCLUSION Heterogeneity of mast cells is found in rats, mice, and humans. The difference between CTMC and MMC observed in rats and mice is a good exampleof such heterogeneity, but likely additional differences are present amongmast cell populations. For example, electron microscopic differences are detectable amongthe CTMC population in the skin of rats (94) and mice (95). The presence of heterogeneity in mast cell populations maybe explained by their unique differentiation process. Mast cell precursors (CFU-Mast) migrating in the blood do not appear to be committed to any subpopulations. Since CFU-Mast differentiate after invading particular tissue, the phenotype of mast cells is influenced by the tissue environment in which differentiation occurs (Figure 2). The mechanismby which the phenotype is determinedby tissue factor(s) remains to be clarified. The phenotypic change between subpopulations is possibly due to the other uniquecharacteristics of mastcells, that is, the extensiveproliferative potential of differentiated mast cells. Somedifferentiated mast cells can BMCMC CFUMast / (~)

~

MMC (~

#~7~ (a)

Environment Suspension Culture Mucosa

(b)

k~_~ CTMC

Muscularis

propria

Peritoneal

Cavity

Skin

Figure 2 Phenotypic changes between subpopulations. Proliferation does not appear necessary for the phenotypic change in direction (a) but it appears necessary for the phenotypic change in direction (b).

Annual Reviews 72

KITAMURA

proliferate in an environmentthat is different from the original environment. The phenotype of the resulting progeny is determined by the tissue environment in which the second differentiation occurs. In other words, some mast cells can experience cycles of proliferation and phenotypic changedue to their extensive proliferative potential (Figure 2).


ACKNOWLEDGMENTS

The studies described here were supported by grants from the Ministry of Education, Science, and Culture, the Ministry of Health and Welfare, the Mitsubishi Fou!adation, the Asahi Fund for Science and Culture, and the HoanshaFoundation, and the Cell Science Foundation. The author thanks his scientific colleagues whoaided in the preparation of this review by providing preprints and allowing him to quote from their unpublished works, and Drs. Stephen J. Galli and Jun Fujita for reviewing the manuscript.

Literature Cited ¯ 1. Metzger, H., Alcaraz, G., Hohman,R., Kinet, J. P., Pribluda, V., Quarto, R. 1986. The receptor with high affinity for immunoglobulin E. Ann. Rev. Immunol. 4:419~0 2. Ishizaka, T., Ishizaka, K. 1984. Activation of mast cells for mediator release through IgE receptors. Prog. Allergy 34: 188-235 3. Ha, T. Y., Reed, N. D., Crowle, P. K. 1983. Delayed expulsion of adult Trich&ella spiral& by mast cell-deficient ~ mice. Infect. Immun. 41:445-47 W/W 4. Oku, Y., Itayama, H., Kamiya. M. 1984. Expulsions of Trichinella spiralis from ~ mice reconstituted the intestine of W/W with haematopoietic and lymphopoietic cells and origin of mucosal mast cells. Immunology 53:337-44 5. Nawa, Y., Kiyota, M., Korenaga, M., Kotani, M. 1985. Defective protective ~ mice against Stroncapacity of W/W gyloides ratti infection and its reeonstitution with bone marrowcells. Parasite Immunol. 7:429-38 6. Matsuda, H., Fukui, K., Kiso, Y., Kitamura, Y. 1985. Inability of genetically v mice to acquire mast cell-deficient W/W resistance against larval Haemaphysalis longicornis ticks. J. Parasitol. 71:443 48 7. Matsuda, H., Nakano, T., Kiso, Y., Kitamura, Y. 1987. Normalization of anti-tick of mast cell deficient ° miceresponse W/W by intracutaneous injection

of cultured mast cells. J. Parasitol. 73: 155-60 8. Galli, S. J., Dvorak, A. M., Dvorak, H. F. 1984. Basophils and mast cells: morphologicinsights into their biology, secretory patterns, and function. Pro#. Allergy 34:1-141 9. Kitamura, Y., Sonoda, T. 1985. Differentiation of mast cells and basophils. In Hematopoietic Stem Cells, ed. D. W. Golde, F. Takaku, pp. 65-80. New York: Marcel Dekker. 379 pp. 10. Kitamura, ¥., Kanakura, Y., Fujita, J., Nakano, T. 1987. Differentiation and transdifferentiation of mast ceils; a unique memberof the hematopoietic cell family, lnt. J. Cell Clon. 5:108 21 11. Barrett, K. E., Metcalfe, D. D. 1984. Mast cell heterogeneity: evidence and implications. J. Clin. Immunol. 4: 25361 12. Katz, H. R., Stevens, R. L., Austen, K. F. 1985. Heterogeneity of mammalian mast cells differentiated in vivo and in vitro. J. Allergy Clin. ImmunoL76: 25059 13. Lee, T. D. G., Swieter, M., Bienenstock, J., Befus, A_ D. 1985. Heterogeneity in mast cell populations. Clin. Immunol. Rev. 4:14349 14. Pearce, F. L. 1986. Onthe heterogeneity of mast cells. Pharmacology.32:61-71 15. Maximow,A. 1906. Ober die Zellformen des lockeren Bindesgewebes. Arch. Mikrosk. Anat. 67:68(~757

Annual Reviews


MAST CELL DIFFERENTIATION 16. Enerb~ick, L. 1986. Mast cell heterogeneity: the evolution of the concept of a specific mucosalmast cell. In MastCell Differentiation and Heterogeneity, ed. A. D. Befus, J. Bienenstock, J. A. Denburg, pp. 1~6. NewYork: Raven. 426 pp. 17. Enerb/ick, L. 1974. Berberine sulfate binding to mast cell polyanions: a cytofluorometric method for the quantitation of heparin. Histochemistry 42: 30113 18. Dimlich, R. V. W., Meineke, H. A., Reilly, F. D., McCuskey,R. S. 1980. The fluorescent staining of heparin in mast cells using berberine sulfate: compatibility with papaformaldehyde or ophthalaldehyde induced-fluorescence and metachromasia. Stain Technol. 55:212 23 19. Nakano, T., Sonoda, T., Hayashi, C., Yamatodani, A., Kanayama, Y., Yamamura T., Asai, H., Yonezawa,T., Kitamura, Y., Galli, S. J. 1985. Fate of bone marrow-derivedcultured mast cells after intracutaneous, intraperitoneal, and intravenous transfer into genetically v mast cell-deficient W/W mice: evidence that cultured mast cells can give rise to both "’connective tissue type" and "mucosal" mast cells. J. Exp. Med. 162: 1025-43 20. Combs,J. W., Lagunoff, D., Benditt, E. P. 1965. Differentiation and proliferation of Biol. embryonic mast cells of the rat. J. Cell 25:57792 21. Ginsburg, H., Sachs, L. 1963. Formation of pure suspension of mast cells in tissue culture by differentiation of lymphoid cells Inst. from 31 the: 1-39 mouse thymus. J. Natl. Cancer 22. Ishizaka, T., Okudaira, H., Mauser, L. E., Ishizaka, K. 1976. Development of rat mast cells in vitro. I. Differentiation of mast cells from thymus cells. J. Immunol. 116:747-54 23. Ishizaka, T., Adachi, T., Chang, T. H., Ishizaka, K. 1977. Developmentof mast cells in vitro. II. Biologic function of cultured mast cells. J. Immunol. 118: 211-17 24. Hasthorpe, S. 1980. A hemopoietic cell line dependent upon a factor in pokeweedmitogen-stimulated spleen cell conditioning medium.J. Cell. Physiol. 105: 379-84 25. Schrader, J. W., Lewis, S. J., ClarkLewis, I., Culvenor, J. G. 1981. The persisting (P) cell: histamine content, regulation by a T cell-derived factor, origin from a bone marrow precursor, and relationship to mast cells. Proc. NatL Acad. Sci. USA 78:323~7 26. Nagao, K., Yokoro, K., Aaronson, S. A.

73

1981. Continuous lines of basophil/mast cells derived from normal mouse bone marrow. Science 212:333-35 27. Nabel, G., Galli, S. J., Dvorak, A. M., Dvorak, H. F., Cantor, H. 1981. Inducer T lymphocytes synthesize a factor that stimulates proliferation of cloned mast cells. Nature 291:332-34 28. Tertian, G., Yung, Y. P., Guy-Grand, D., Moore, M. A. S. 1981. Long-term in vitro culture of murinemast cells. I. Description of a growth-factor dependent culture technique. J. Immunol.127: 78894 29. Razin, E., Cordon-Cardo, C., Good, R. A. 1981. Growth of a pure population of mousemast cells in vitro with conditioned medium derived from concanavalin A-stimulated splenocytes. Proc. Natl. Acad. Sci. USA 78:2559~61 30. Ihle, J. N., Keller, J., Oroszlan, S., Henderson, L. E., Copeland, T. D., Fitch, F., Prystowsky, M. B., (3oldwasser, E., Schrader, J. W., Paraszynski, E., Dy, M., Lebel, B. 1983. Biologic properties of homogeneousintcrleukin 3. I. Demonstration of WEHI-3growthfactor activity, mast cell growth-factor activity, P cell-stimulating factor activity, colony-stimulating factor activity and histamine-producing cell-stimulating factor activity. J. lmmunol. 131: 28~87 31. Fung, M. C., Hapel, A. J., Ymer, S., Cohen, D. R., Johnson, R. M., Campbell, H. D., Young, I. G. 1984. Molecular cloning of cDNAfor murine interleukin-3. Nature 307:233 37 32. Yokota, T., Lee, F., Rennick, D., Hall, C., Arai, N., Mosmann,T., Nabel, G., Cantor, H., Arai, K. 1984. Isolation and characterization of a mousecDNAclone tttat expresses mast-cell growth-factor activity in monkeycells. Proc. Natl. Acad. Sei. USA 81:1070-74 33. Smith, C. A., Rennick, D. M. 1986. Characterization of a murine lymphokine distinct from interleukin 2 and interleukin 3 (IL-3) possessing a T-cell growth factor activity and a mast-cell growth factor activity that synergizes with IL-3. Proc. Natl. Aead. Sci. USA 83:1857-61 34. Lee, F., Yokota, T., Otsuka, T., Meyerson, P., Villaret, D., Coffman, R., Mosmann, T., Rennick, D., Roehm, N., Smith, C., Zlotnik, A., Arai, K. 1986. Isolation and characterization of a mouse interleukin cDNAclone that expresses B-cell stimulatory factor 1 activities and T-cell- and mast-cellstimulating activities. Proc. Natl. Acad. Sei. USA 83:2061-65

Annual Reviews


74

KITAMURA

35. Noma,Y, Sideras, P., Naito, T., Bergstedt-Lindquist, S., Azuma,C., Severinson, E., Tanabe,T., Kinashi, T., Matsuda, F., Yaoita, Y., Honjo,T. 1986. Cloning of eDNAencoding the murine IgG1induction factor by a novel strategy using SP6 promoter. Nature 319: 640-46 36. Grabstein, K., Eiseman,J., Moehizuki, D., Shanebeek,K., Conlon, P., Hopp, T., March,C., Gillis, S. 1986. Purification to homogeneity of B cell stimulating factor. Amoleculethat stimulates proliferation of multiple lymphokinedependentcell lines. J. Exp. Med.163: 1405-14 37. Paul, W.E., Ohara, J. 1987. B-cell stimulatoryfactor-l/interleukin 4. Ann. Rev. Immunol.5:429-59 38. Haig, D. M., MeKee,T. A., Jarrett, E. E. E., Woodbury,R., Miller, H. R. P. 1982. Generationof mucosalmast cells is stimulatedin vitro by factors derived from T cells of helminth-infectedrats. Nature 300:188-90 39. Saito, H., Hatake, K., Dvorak,A. M., Leiferman, K. M., Donnenberg,A. D., Arai, N., lshizaka,K., lshizaka,T. 1988. Selective differentiation and proliferation of hematopoietiecells inducedby recombinanthumaninterleukins. Proc. Natl. Acad. Sei. USA85:2288-92 40. Otsu, K., Nakano, T., Kanakura, Y., Asai, H., Katz, H. R., Austen, K. F., Stevens,R. L., Galli, S. J., Kitamura,Y. 1987. Phenotypicchangesof bone marrow-derived mast cells after intraperitoneal transfer into W~Wv micethat are geneticallydeficientin mastcells. J. Exp. Med. 165:61:%27 41. Enerbfick,L., Kolset, S. O., Kusche,M., ~jerpe, A., Lindahl, U. 1985. Glycosaminoglycans in rat mucosalmast cells. Biochem.J. 227:66168 42. Stevens,R. L., Lee, T. D.G., Seldin, D. C., Austen,K. F., Befus,A. D., Bienenstock, J. 1986. Intestinal mucosalmast cells from rats infected with Nippostrongylusbrasiliensis containproteaseresistant chondroitin sulfate di-B proteoglycans. J. Immunol.137:291~5 43. Jarrett, E. E. E., Haig, D. M. 1984. Mucosalmastcells in vivo andin vitro. Immunol. Today 5:115 19 44. Razin, E., Stevens, R. L., Akiyama,F., Schmid,K., Austen,K. F. 1982. Culture from mousebone marrowof a subclass of mastcells possessinga distinct chondroitin sulfate proteoglycanwith glycosaminoglycansrich in N-acetylgalactosamine-4,6-disulfate.J. Biol. Chem. 257:7229-36 45. Katz, H. R., Austen, A. F., Caterson,

B., Stevens,R. L. 1986. Secretorygranules of heparin-containingrat serosal mastcells also possesshighly sulfated chondroitin sulfate proteoglycans. J. Biol. Chem.261:13393-96 46. Lagunoff,D., Pritzl, P. 1976. Characterization of rat mast cell granule proteins. Arch. Biochem.Biophys. 173: 554-63 47. Woodbury,R. G., Gruzenski, G. M., Lagunoff, D. 1978. Immunofluorescent localization of a serine proteasein rat small intestine. Proc. Natl. Acad.Sci. USA 75:2785-89 48. Woodbury,R. G., Everitt, M. T., Neurath, H. 1981. Mast cell proteases. MethodsEnzymol. 80:588-609 49. Woodbury,R. G., Miller, H. R. P., Huntley, J. F., Newlands, G. F. J., Palliser, A. C., Wakelin,D. 1984. Mucosal mastcells are functionallyactive during spontaneousexpulsionof intestinal nematodeinfections in rat. Nature312: 450-52 50. Serafin, W.E., Dayton,E. T., Gravallese, P. M., Austen,K. F., Stevens, R. L. 1987. CarboxypeptidaseA in mouse mastcells: identification, characterization, anduse as a differentiation marker. J. lmmunol.139:3771-76 51. Heavey,D. J., Ernst, P. B., Stevens, R. L., Befus, A. D., Bienenstock, J., Austen, K. F. 1988. Generationof leukotriene C4, leukotriene B4, and prostaglandin D2by immunologically activated rat intestinal mucosa mastcells. J. lmmunol. 140:1953-57 52. Levi-Schaffer, F., Dayton, E. T., Austen,K. F., Hein,A., Caulfield,J. P., Gravallese, P. M., Liu, F. T., Stevens, R. L. 1987. Mousebone marrow-derived mast cells cocultured with fibroblasts: morphologyand stimulation-induced release of histamine, leukotriene B4, leukotriene C4,and prostaglandinD2. J. Immunol. 139:3431-41 53. Katz, H. R., LeBlanc,P. A., Russell, S. W.1983.Twoclasses of mousemastcells delineated by monoelonalantibodies. Proe. Natl. Acad. Sci. USA80:5916-18 54. Katz, H. R., Austen,K. F. 1986. Plasma membrane and intracellular expression of globotetraosylceramide (globoside) mousebone marrow-derivedmast cells. J. Immunol.136:3819-24 55. Irani, A. A., Sehechter,N. M., Craig, S. S., DeBlois,G., Schwartz,L. B. 1986. Twotypes of humanmast cells that have subsets with distinct neutral protease compositions. Proc. Natl. Acad. Sci. USA 83:4464-68 56. Schwartz,L. B., Irani, A.A., Roller, K., Castells, M.C., Scheehter,N. M.1987.

Annual Reviews


MAST CELL DIFFERENTIATION Quantitation of histamine, tryptase, and cttymase in dispersed truman T and TC mast cells. J. Immunol. 138: 261115 57. Irani, A. A., Craig, S. S., DeBlois, G., Elson, C. O., Schechter, N. M., Schwartz, L. B. 1987. Deficiency of the tryptase-positive, chymase-negativemast cell type in gastrointestinal mucosaof patients with defective T lymphocyte function. J. Immunol. 138:4381 86 58. Benyon, R. C., Lowman, M. A., Church, M. K. 1987. Humanskin mast cells: their dispersion, purification, and secretory characterization. J. Immunol. 138:861~7 59. Lawrence, I. D., Warner, J. A., Cohan, V. L., Hubbard, W. C., Kagey-Sobotka, A., Lichtenstein, L. M. 1987. Purification and characterization of human skin mast cells: evidence for humanmast cell heterogeneity. J. Imrnunol. 139: 3062-69 60. Befus, A. D., Dyck, N., Goodacre, R., Bienenstock, J. 1987. Mastcells from the humanintestinal lamina propria: isolation, histochemical subtypes, and functional characterization. J. Immunol.138: 2604-10 61. Thompson,H. L., Schulman, E. S., Metcalfe, D. D. 1988. Identification ofchondroitin sulfate E in humanlung mast cells. J. Immunol. 140:2708-13 62. Stevens, R. L., Fox, C. C., Lichtenstein, L. M., Austen, K. F. 1988. Identification of chondroitin sulfate E proteoglycans and heparin proteoglycans in the secretory granules of humanlung mast cells. Proc. Natl. Acad. Sci. USA85: 2284-87 63. Kitamura, Y., Shimada, M., Hatanaka, K., Miyano, Y. 1977. Development of mast cells from grafted bone marrow cells in irradiated mice. Nature 268: 44243 64. Kitamura, Y., Go, S., t-latanaka, K. 1978. Decrease of mast cells in °W/W mice and their increase by bone marrow transplantation. Blood 52:447-52 65. Kitamura, Y., Go, S. 1979. Decreased production of mast cells in Sl/Sl d anemic mice. Blood 53:492-97 66. Crowle, R. K., Reed, N. D. 1984. Bone marroworigin ofmucosal mast cells. Int. Archs. Allergy Appl. lmmun. 73:242-47 67. Kitamura, Y., Yokoyama, M., Matsuda, H., Ohno, T., Mori, K. J. 1981. Spleen colony-forming cell as common precursor for tissue mast cells and granulocytes. Nature 291 : 159~50 68. Sonoda, T., Kitamura, Y., Haku, Y., Hara, H., Mori, K. J. 1983. Mast cell precursors in various haematopoietic

75

colonies of mice producedin vivo and in vitro. Br. J. Haematol. 53:6ll-20 69. Nakahata, T., Ogawa, M. 1982. Identification in culture of a class of hemopoietic colony-formingunits with extensive capability to self-renew and generate multipotential hemopoietic colonies. Proc. Natl. Acad. Sci. USA 79:3843-47 70. Kobayashi, T., Nakano, T., Nakahata, T., Asai, H., Yagi, Y., Tsuji, K., Komiyama, A., Akabane, T., Kojima, S., Kitamura, Y. 1986. Formation of mast cell colonies in methylcellulose by mouse peritoneal cells and differentiation of these cloned cells in both the skin and the gastric mucosa of W~Wv mice: evidence that a commonprecursor can give rise to both "connective tissue-type" and "mucosal" mast cells. J. Immunol. 136: 1378-84 71. Kanakura, Y., Kuriu, A., Waki, N., Nakano, T., Asai, H., Yonezawa, T., Kitamura, Y. 1988. Changes in numbers and types of mast cell colony-forming cells in the peritoneal cavity of miceafter injection of distilled water: evidencethat mast cells suppress differentiation of bone marrow-derived precursors. Blood 71:573-80 72. Kitamura, Y., Matsuda, H., Hatanaka, K. 1979. Clonal nature of mast cell clusv mice after bone ters formed in W/W marrow transplantation. Nature 281: 154-55 73. Hatanaka, K., Kitamura, Y., Nishimune, Y. 1979. Local development of mast cells from bone marrow-derived precursors in the skin of mice. Blood53: 142-47 74. Sonoda, T., Tsuyama, K., Kitamura, Y., Tanooka, H. 1982. Different effects of dimethylbenz(~)anthracene and tetradeeanoylphorbol acetate on differentiation of mast cells in the skin of mice. Am. J. Pathol. 106:312-17 75. Dvorak, A. M., Schleimer, R. P., Lichtenstein, L. M. 1988. Humanmast cells synthesize new granules during recovery from degranulation. In vitro studies with mast cells purified from human lungs. Blood 71:76-85 76. Sonoda, T., Kanayama, Y., Hara, H., Hayashi, C., Tadokoro, M, Yonezawa, T., Kitamura, Y. 1984. Proliferation of peritoneal mast cells in the skin of ~ W]W mice that genetically lack mast cells. J. Exp. Med. 160:138-51 77. Nakahata, T., Kobayashi, T., Ishiguro, A., Tsuji, K., Naganuma,K., Ando, O., Yagi, Y., Tadokoro, K., Akabene, T. 1986. Extensive proliferation of mature connective-tissue type mast cells in vitro. Nature 324:65q57

Annual Reviews


76

KITAMURA

78. Hamaguchi, Y., Kanakura, ¥., Fujita, J., Takeda, S., Nakano, T., Tarui, S., Honjo, T., Kitamura, Y. 1987. Interleukin 4 as an essential factor for in vitro clonal growth of murine connective tissue type mast cells. J. Exp. Med. 165: 268-73 79. Kanakura, Y., Sonoda, S., Nakano, T., Fujita, J., Kuriu, A., Asai, H., Kitamura, Y. 1987. Formation of mast cell colonies in methylcellulose by mouse skin cells and development of mucosallike mast cells from the cloned cells in the gastric mucosa of W/W"mice. Am. J. Pathol. 129:168-76 80. Ruitenberg, E. J., Elgersma, A. 1976. Absenceof intestinal mast cell response in congenitally athymic mice during Trichinella spiralis infection. Nature 264: 258~50 81. Mayrhofer, G., Fisher, R. 1979. Mast cells in severely T-cell depleted rats and the response to infestation with Nippostrongylus brasiliensis. Immunology37: 145-55 82. Abe, T., Ochiai, H., Minamishima, Y., Nawa, Y. 1988. Induction of intestinal mastocytosis in nude mice by repeated injection of interleukin-3. Int. Archs. Allergy Appl. lmmun. 86:35~58 83. Abe, T., Nawa, Y. 1988. Wormexpulsion and mucosal mast cell response induced by repetitive IL-3 administration in Strongyloides ratti-infected nude mice. lmmunolo.qy 63:181 85 84. Viklick~, V., ~ima, P., Prichard, H. 1973. Onthe origin of mast cells in adult life. Folia Biol. 19:247-51 85. Keller, R., Hess, M. W., Riley, J. F. 1976. Mast cells in the skin of normal, hairless and athymic mice. Experientia 32:171 72 86. Fujita, J., Nakayama, H., Onoue, H., Kanakura, Y., Nakano, T., Asai, H., Takeda, S., Honjo, T., Kitamura, Y. 1988. Fibroblast-dependent growth of mousemast cells in vitro: duplication of mast cell depletion in mutant mice of v genotype. J. Cell. Physiol. 134: W/W 78-84 87. Fujita, J., Nakayama, H., Onoue, H., Ebi, Y., Kanakura, Y., Kuriu, A., Kita° mousemura, Y. 1988. Failure of W/W derived cultured mast cells to enter S phase upon contact with NIH/3T3 fibroblasts. Blood 72:463 68 88. Nakano, T., Kanakura, Y., Asai, H.,

Kitamura, Y. 1987. Changing processes from bone marrow-derived cultured mast cells to connective tissue-type mast cells in the peritoneal cavity of mastcell° mice: association of prodeficient W/W liferation arrest and differentiation. J. Immunol. 138:544~9 89. Kanakura, Y., Thompson, H. L., Nakano, T., Yamamura, T., Asai, H., Kitamura, Y., Metcalfe, D. D., Galli, S. J. 1988. Multiple bidirectional alterations of phenotype and changes in pro" liferative potential during the in vitro and in vivo passage of clonal mast cell populations derived from mouse peritoneal mast cells. Blood 72:877 85 90. Sonoda, S., Sonoda, T., Nakano, T., Kanayama,Y., Kanakura, Y., Asai, H., Yonezawa, T., Kitamura, Y. 1986. Developmentofmt/cosal mast cells after injection of a single connective tissuetype mast cell in the stomach mucosa of genetically mast cell-deficient °W~W mice. J. Immunol. 137:131%22 91. Galli, S. J., Dvorak, A. M., Marcum,J. A., Ishizaka, T., Nabel, G., Der Simonian, H., Pyne, K., Goldin, J. M., Rosenberg, R. D., Cantor, H., Dvorak, H. F. 1982. Mast cell clones: a modelfor the analysis of cellular maturation. J. Cell Biol. 95:435-44 92. Levi-Schaffer, F., Austen, K. F., Gravallese, P. M., Stevens, R. L. 1986. Coculture of interleukin 3-dependent mousemast cells with fibroblasts results in a phenotypic change of the mast cells. Proe. Natl. Aead. Sei. USA83:6485-88 93. Dayton, E. T., Pharr, P., Ogawa, M., Serafin, W. E., Austen, K. F., LeviSchaffer, F., Stevens, R. L. 1988. 3T3 fibroblast induce cloned interleukin 3dependent mousemast cells to resemble connective tissue mast cells in granular consistency. Proc. Natl. Acad. Sci. USA 85:569 72 94. Aldenborg,F., Enerb~ick, L. 1988. Histochemical heterogeneity of dermal mast cells in athymic and normal rats. Histochem. J. 20:19-28 95. Yamamura, T., Nakano, T., Fukuzumi, T., Waki, N., Asai, H., Yoshikawa, K., Kitamura, Y. 1988. Electron microscopic changes of bone marrow-derived cultured mast cells after injection into the skin of genetically mast cell-deficient W/W~ mice. J. Invest. Dermatol.91 : 26973



Annual Reviews www.annualreviews.org/aronline Ann. Rev. lmmunol. 1989. 7:7749 Copyright © 1989 by Annual Reviews Inc. All rights reserved


THE CELLULAR BASIS OF T-CELL MEMORY Jean-Charles Cerottini and H. Robson MacDonald Ludwig Institute for Cancer Research, Lausanne Branch, 1066 Epalinges, Switzerland

INTRODUCTION One of the major characteristics of the immunesystemis its ability to react more quickly and more intensely on the second exposure to an antigen. This characteristic, usually referred to as immunologicmemory,was first observed in studies of antibody production in vivo (reviewed in 1). The secondary, or anamnestic, response to an antigen not only results in the production of increased levels of antibodies, but these antibodies also generally have a higher affinity for antigen comparedto those antibodies produced during the primary response (2). Memoryat the B-cell level generally acceptedas the result of selection and specific expansionof clones secreting high affinity antibody molecules (3, 4). Molecularstudies suggest that somatic hypermutation of antibody variable-region genes during the primary response maybe responsible for the occurrence of B cells expressing high affinity receptors (reviewed in 5). However,the differentiation pathway involved in the generation of memoryB cells is still poorly understood. In particular, it is not clear how, in the primary response, someof the B cells differentiate into terminal plasmacells and others into memorycells. While there is ample evidence that immunologicmemoryalso involves T cells, our understanding of the molecular and cellular basis of T-cell secondary responses is quite limited. Following early studies demonstrating that cytolytic T lymphocytes(CTL)are not necessarily terminal cells but maydifferentiate, at least in vitro, into cells exhibiting the expected properties of memorycells (6), attempts have been made to define quantitative and qualitative differences betweenmemory T cells and T cells that have not yet been stimulated by antigen (which are often designated as 77 0732~)582/89/0410~077502.00

Annual Reviews 78

CEROTTINI

& MACDONALD

naive or virgin T cells). Until recently, detailed analysis of these differences has been hamperedby the lack of suitable surface markers allowing identification of memoryT cells. As discussed in this chapter, recent studies indicate that naive and memory T cells can now be distinguished phenotypically by surface markers.


CHARACTERISTICS

OF

T-CELL

MEMORY

Muchof the information concerning the differentiation and ultimate fate of T cells in primary or secondary responses has been obtained in Studies + T cells appear to exhibit cytolytic potenconcerning CTL. As most CD8 tial, it is likely that the results of these studies are applicableto the majority of T cells expressing this surface phenotype. Muchless is knownon the properties of CD4+ T cells involved in memory. That some form of memoryexisted in CTLresponses was first indicated in studies concerning the formation of CTL in mice immunized with allogeneic tumor cells (reviewed in 6). The level of CTLmemoryresulting from primary immunization was further analyzed by measuring the development of CTLactivity in cultures of lymphoid cells from immunized mice restimulated with the appropriate antigens (7). Althoughaccelerated kinetics and higher peak levels of CTLactivity could be readily demonstrated in these secondary responses in vitro, these studies provided little insight into the cellular basis of T-cell memory.Withthe development of appropriate limiting dilution microculture systems (8), however, became possible to determine the frequencies of antigen-specific CTL precursors (CTL-P). Using this approach, direct evidence has been obtained that the frequency of CTL-Pagainst a given antigen is increased after primary immunization (9). In someinstances, this increase is quite substantial. For example, primary immunization of female mice with male cells bearing the H-Yminor transplantation antigen has been found to result in a > 20-fold increase in the frequency of H-Yspecific CTL-P.In most viral systems studied, primary immunization leads to a > 10-fold increase in CTL-Pfrequencies. In contrast, primary immunization against major histocompatibility complex (MHC)antigens generally results in modest( < 3-fold) increase in CTL-Pfrequencies. Whileit is clear that T-cell memorymayinvolve increased frequencies of antigen-specificT cells, there is also evidence,albeit indirect, for qualitative differences between memoryand naive T cells. For example, it has been shownthat the antigenic requirements for the generation of alloreactive CTLin vitro are quite different between normal and alloimmunelymphoid cell populations (6). Althoughthe molecular basis for these differences has yet to be defined, it is conceivablethat they are related, at least in part, to

Annual Reviews


T-CELL

MEMORY 79

the affinities of the antigen receptors expressed by naive and memory T cells. Presently, there is no direct methodto measurethe affinities of Tcell receptors (TCR).However, relative affinities of TCRcan be indirectly comparedby varying the concentration of antigen/MHCcomplexon the stimulator cell (10, 11) or by neutralizing the avidity-enhancingfunction of accessorymoleculessuch as CD4or CD8(12, 13). Usingthis approach, it has beenfoundthat the proportionof alloreactiveCTL-P bearingrelatively high affinity TCRis increasedafter appropriateimmunization in vivo (12). As discussed below, there is evidence that memoryT cells express increased levels of the surface moleculesthoughtto be involvedin T-cell adhesion or activation (such as CD2,LFA-1,and LFA-3).The enhanced responsivenessof memory T cells to antigenic stimulation maythus involve several other properties, in additionto expressionof high affinity TCR.In this context, it is noteworthythat alloreactive or virus-specific memory CTL-P have beenreported to respondto activation by nonspeeific stimuli such as phorbol ester and calcium ionophore in conditions under which alloreactive CTL-Pfrom normal mice are nonresponsive(14-16). Moreover, humanmemory T cells appear to respond better than naive T cells to activation by anti-CD3antibodies, although both cell types express comparable amountsof CD3on the cell surface (17). As these latter results havenot beenconfirmedin similar studies performedin the mouse,further workis neededto ascertain whetherdifferentiation into memory cells is accompanied by qualitative or quantitative changesin the T-cell requirementfor activation. In addition to increased proliferation, secondaryT-cell responses to antigen maybe characterized by production of larger amountsof lymphokines. In most instances, it has beendifficult to distinguish whether this increased production can be accountedfor solely by the increased frequencyof antigen-specific T cells in immunizedlymphoidcell populations. As discussed below, there is nowevidence that memory T cells produce, on a per cell basis, moreIFN-yor IL-4 than do naive T cells uponstimulationwith specific or nonspecificstimuli. As sucha difference does not applyto the productionof IL-2, these results raise the possibility that the expression of the IFN-7and IL-4 genes can be modulateddifferently in naive and memory T cells and independentlyof the expression of the IL-2 gene. PGP-1 (LY24) AS MARKER OF MURINE MEMORY T CELLS Pgp-1wasoriginally described as a major integral membrane glycoprotein on murinefibroblasts andperitoneal phagocyticcells (henceits designation

Annual Reviews


80

CEROTTINI

& MACDONALD

phagocyte glycoprotein 1) (18, 19). Also designated Ly24 (20), Pgpa mol wt of 80,000-95,000 and is encoded by a locus on chromosome2 (21). Twoallelic forms have been defined using mousealloantibodies (22). The glycoprotein has been identified in a wide variety of tissues including brain, liver, kidney, and lung (23). In hematopoietictissues, Pgp-1is found in highest amounts in the bone marrow. Whenbone marrow cells are stained with antibody to Pgp-1, mature myeloid cells stain very brightly, committed progenitors brightly, and pluripotent stem cells and lymphoid cells relatively weakly(23, 24). Whileit is expressed in most prothymocytes (25), only about 5%of adult thymocytes are Pgp-1+, at least in certain mousestrains (23). Early reports on Pgp-1 expression amongthymocytes suggested that the antigen was restricted to the minor (CD4- CD8-) subset knownto contain immature thymocytes (23, 26, 27). Unlike adult thymocytes, the great majority of fetal thymocytes in AKRmice are Pgp1+ + at day 13-14 of gestation (27). Thereafter, the proportion of Pgp-1 thymocytes declines, reaching adult levels by day 19 of gestation. On the basis of these findings, it has been proposed that at least someof the PgpI + cells within the thymus are progenitors of mature thymocytes (27). However, direct evidence for such a developmental relationship has not been obtained yet (28). It is nowclear that Pgp-1 + expression is not confined to subpopulations of immature thymocytes but can be found on a small percentage of mature (CD4+ CD8+ and CD4- CD8+) cells in the thymus (28-31). Moreover, peripheral T cells contain a significant proportion of Pgp-1 + ceils (29, 30, 32). It should be emphasizedthat the difference in Pgp-1 expression among peripheral T cells is quantitative rather than qualitative. Two-colorflow microfluorometry analysis of peripheral T cells reveals two well-defined subpopulations among CD4+ and CD8+ cells, i.e. a major subpopulation that stains weakly (designated Pgp-1 ) and a minor (Pgp-1 +) subpopulation that stains brightly. Pgp-1+ T cells have been identified in blood, spleen, and lymphnode (29). While there is no significant difference in the percentages of positive cells within these lymphoidcell compartments, it has been shownthat the relative numberof Pgp-1+ cells increases + and CD8 + subsets. This progressively as a function of age in both CD4 increase can be accelerated by surgical thymectomyof adult mice (29). A possible relationship between Pgp-1- and Pgp-1 + peripheral T cells was first suggested by the demonstration that Pgp-1- CD8+ lymph node cells became rapidly Pgp-1 + after mitogenic or antigenic stimulation in vitro (29). Moreover,analysis of Pgp-1expression by these stimulated cells revealed that this phenotype was stable for as long as the cultures could be viably maintained without restimulation (3 weeks). As it is well established that CD8+ CTLgenerated in allogeneic MLCcultures revert to

Annual Reviews


T-CELL MEMORY

81

small-sized quiescent lymphocytes under similar culture conditions (6), these findings suggested that Pgp-1 is a marker not only of recently activated T cells but also of the progeny of these activated cells, including memorycells. Direct evidence that Pgp-1 identifies recently activated T cells in vivo was obtained by examining the surface phenotype of the CTLgenerated in mice undergoingrejection of an intraperitoneal tumorallograft. In this + Pgp-1 + (29). It system, virtually all of the CTLwere found to be CD8 noteworthy that the CTLgenerated under these conditions represent a relatively homogeneous noncycling population of small-sized lymphocytes. These data thus suggest that Pgp-1 identifies recently activated T cells independent of the usual criteria of blast transformation and DNA synthesis. As discussed previously, immunization may result in increased frequencies of antigen-specific CTL-P.Several studies indicate that these cells are mainly found in the Pgp-1 + subpopulation (29, 32, 33). For example, the frequency of H-Yspecific CTL-Pin CD8+ spleen cells from female mice immunizedwith male cells has been found to be enriched 30-fold in + antigen-specific the Pgp-1 + subpopulation (29). Data pertaining to CD4 T cells are more limited at present; however, there is already evidence in several antigenic systems that the Pgp-1+ CD4+ subpopulation from immunizedmice is enriched in precursor cells directed against the immunizing antigen (33). Further evidence in favor of the expression of Pgp-1 by memoryT cells has been provided by studies of the surface phenotype of alloimmune CTL-Pbearing high affinity TCR.As mentioned previously, measurement of the resistance of CD8+ CTLclones to inhibition of cytolytic activity by anti-CD8 antibodies provides a way to indirectly assess the frequency of CTLwith high affinity TCR.Using this approach, it has been found that the CD8+ Pgp-I ÷ subpopulation from alloimmune spleen cells was highly enriched for such cells compared to the CD8+ Pgp-1- subpopulation (32). In contrast, no difference in susceptibility to inhibition by anti-CD8 antibodies was observed when Pgp-l + and Pgp-1- subpopulations from normal mice were analyzed by the same method. Taken together with earlier studies (12), these data strongly suggest that the Pgp-I + subset in immunizedmice is enriched for antigen-specific T cells bearing high affinity TCR. Evidence that affinity (rather than density) of TCRaccounts for this difference comes from the observation that CD3staining (and hence presumablyTCRdensity) is not significantly different in the Pgp-1 + versus + or CD8 + T cells (33). Pgp-l- subpopulations of either CD4 Additional functional differences between Pgp-1 + and Pgp-1- T cells have been revealed by studies of lymphokine production by these cells

Annual Reviews


82

CEROTTIN|

& MACDONALD

upon stimulation with nonspecific stimuli (33, 34). While IL-2 production was similar in the two subpopulations, Pgp- 1 ÷ cells (irrespective of their CD4or CD8phenotype) produced greater amounts of IFN-~, IL-4, and (to a lesser extent) IL-3 comparedto Pgp-1- cells. Since the difference lymphokine production was observed after stimulation with nonspecific stimuli, it is likely that these data reflect enhancedsecretion of IFN-7and IL-4 by individual Pgp-1 ÷ T cells. Moreover,these data indicate that there is no apparent relationship between the subpopulations of CD4÷ T cells ÷ defined by Pgp-1 and the recently described Thl and ThEsubsets of CD4 T cell clones (35) since IFN-~and IL-4, which are both producedin greater + Pgp-l + cells, are expressed by mutually exclusive subamounts by CD4 sets of T-cell clones. Taken together, the results discussed in this section indicate that + and CD8+ expression of Pgp-1 identifies discrete subpopulations of CD4 T cells that have both the quantitative and qualitative properties of memory T cells. Accordingly, a simplified modelfor the cellular differentiation pathway leading to the formation of memoryT cells can be proposed. Mature T ceils produced in the thymus are Pgp-l-. Whenthese (naive) cells seed the peripheral lymphoid tissues, they remain Pgp-1- until they are triggered by antigen. Those with sufficiently high affinity TCRto be triggered acquire Pgp-1expression as they differentiate into effector cells and (ultimately) memoryT cells. Once acquired, Pgp-1 expression would be stable for the lifetime of the cell. A prediction of this modelis that the Pgp-1+ subpopulation of peripheral T cells should include recently activated lymphocytes as well as memory + Pgp-1+ T cells cells. In this context, it is noteworthythat 10-20%of CD8 are medium- to large-sized lymphocytes, whereas the vast majority of + Pgp-1- T cells are small (29). However, most of the cells in both CD8 groups are not actively cycling. Whetherthe large Pgp-1+ T cells represent recently activated cells has yet to be determined. Finally, despite current lack of knowledgeof the function of the Pgp-1 glycoprotein, it is clear that Pgp-1is a useful markerfor the identification and separation of T lymphocytes in the mouse. It should be noted, however, that Pgp-1 expression by mature T cells exhibits considerable heterogeneity amongmousestrains. While the conclusions presented here regarding the usefulness of this marker apply to C57BL/6mice, it appears that, in certain other strains, expression of Pgp-I does not resolve T cells into two distinct subsets (28, 31). In these strains, further study is needed to identify other surface markers of memoryT cells. As discussed in the next section, a variety of such markers have been identified in humans. It is thus likely that, in the near future, the phenotypic identification of memoryT cells will be feasible in all mousestrains.

Annual Reviews T-CELL


PHENOTYPIC IDENTIFICATION T CELLS IN OTHER SPECIES

MEMORY

83

OF MEMORY

Human studies Human T lymphocytesalso express the Pgp-1 glycoprotein (36). Although + and there is quantitative heterogeneity in Pgp-1 expression amongCD4 + T cells, the differencesare too small to allow definition of distinct CD8 subpopulations. However,two-color flow microfluorometryanalysis using antibodies against a variety of surface moleculesexpressedby T cells has revealed coordinate enhancedexpression of Pgp-1and five other markers (LFA-I, LFA-3, CD2, CDw29and UCHL1)on a subpopulation of cells (37). Amongthem, LFA-3and UCHL1 appear to discriminate two subpopulations better than the other markers because their level of expressionon adult T cells exhibits a clear bimodaldistribution. It should be pointed out that the subpopulationsdefined by these coor+ T cells, to dinately expressed moleculescorrespond, at least in CD4 the "helper-inducer" and "suppressor-inducer"T-cell subsets previously identified by the monoclonalantibodies 4B4(directed against CDw29) and 2H4(directed against CD45R),respectively (38, 39). Whileit has generallyassumed that these subsetsrepresent different lineagesof T cells, recent data strongly suggestthat they correspondto different maturational stages (17). Accordingto this newinterpretation, naive T cells are found + subset, whereasactivated and memory in the CD45R T cells are confined + to the CDw29subset. A major finding in favor of this contention is the demonstrationthat + subset convert to the phenotype peripheral blood T cells of the CD45R + of the CDw29subset upon stimulation in vitro, whereasconversion in the opposite direction is not observed.Thesechangesaffect in particular the expression of someof the CD45polypeptides. The CD45complex, previously knownas humanleukocyte common antigen or T200, consists of four molecularspecies of 220, 205, 190, and 180 kd (reviewedin 40). Geneticstudies indicate that a single geneproducesthe different formsof CD45by alternate splicing of RNA(41) and by posttranslational processing (42). The 205/220-kdmolecular species have been designated CD45R becauseof their restricted distribution on B cells, NKcells, and subsets of T cells. Whilesomemonoclonalantibodies recognizeepitopes on all forms of CD45,others react only with CD45R (such as 2H4)(38) or the 180-kdmolecular species (UCHL1) (43), respectively. Expression of CD45R and UCHL1 on T lymphocytesis mutually exclusive: while 40+ or UCHL1 50%of peripheral blood T cells are CD45R +, < 1%of these (small-sized)cells are positive for both markers.Moreover,clonal analysis + CD45R + T cells, after mitogenic stimuhas shownthat individual CD4

Annual Reviews


84

CEROTTINI& MACDONALD

lation, lose cell surface expression of CD45Rand acquire UCHL1 (and CDw29) with time in culture (44, 45). As is the case for Pgp-1 in the mouse, expression of UCHL1,once acquired, appears to be stable. The switch in + to UCHL1+ is prothe phenotype of stimulated T cells from CD45R gressive and unidirectional. For example, about 40%of the cells obtained + T cells with PHAwere found to be 3 days after stimulation of CD45R + +, CD45R UCHL1 compared to < 1% and < 10% at day 0 and day 7, respectively (45). In contrast, no double staining was observed in PHA+ cells which remained essentially "= after stimulated UCHLI CD45R extended culture in vitro. In line with these data is the observation that cord blood T cells consist + cells (which convert to the CDw29phenoalmost exclusively of CD45R type after in vitro stimulation with PHA)(37). Also, a significant fraction + cells (44). Interestingly, of mature thymocytes is represented by CD45R + + most, if not all, immature thymocytes are CD45R- UCHL1 CDw29 (44, 45), suggesting that several changes in the expression of these markers + CDw29 + thyoccur during T-cell development. Immature UCHL1 mocytes may lose these markers and gain CD45Ron maturation in the thymus. These mature nai’ve T cells, when they seed peripheral lymphoid + UCHL1-CDw29-until they are triggered by tissues, remain CD45R antigen. Uponactivation, these cells lose CD45Rand reexpress UCHL1 and CDw29,this time in a stable manner. + CDw29 + subset of peripheral blood Functional analysis of the UCHL1 T cells supports the notion that it contains memory(previously activated) cells (reviewed in 17). In agreement with the data obtained in the mouse, the frequencyof antigen-specific cells is enhancedin this subset. Similarly, enhanced IFN-y production has been observed after PHAstimulation, whereas IL-2 production appears not to be significantly different between + and CD45R + T cell subsets (37). Consistent with this latter UCHL1 finding is the observation that cord blood T cells (which are virtually + cells) produce significant amounts of IL-2, but little devoid of UCHL1 IFN-~, in response to PHAstimulation (46). Since differentiation from naive to memoryT cells appears to be accompanicdby increased expression of several surface molecules which are (presumably) involved in T-cell function (such as LFA-1, LFA-3 CD2), it has been proposed that such an increase mayrender memorycells moresusceptible to activation signals (37). As virtually all peripheral T cells are clonogenic under optimal stimulation conditions (47), this hypothesis maybe tested experimentally by determining the frequencies of T cells in the various subsets that proliferate in response to a limiting stimulus. It should be noted that the data presented in this section only concern T cells expressing tlae ~//3 TCRheterodimer. Little is knownabout the

Annual Reviews


T-CELL MEMORY

85

differentiation and fate of y/6 TCR-expressingT cells. Since they represent only a minor (1-5%) population of peripheral T cells, detailed analysis their surface phenotype is difficult. Moreover, the ability of y/6 TCRexpressing T cells to undergo clonal growth in vitro appears more limited than that of ~///TCR-expressing T cells (48). There is evidence in the mousethat ~/6 TCR-expressing T cells in lymph node are Pgp-1 ÷ (H. R. MacDonald,unpublished results). As reported recently, the majority of mouse T cells residing in intestinal epithelium express y/6 TCR(49). Although their Pgp-lphenotype is not known, it is noteworthy that most of the human T cells in normal gut mucosa have been reported to be ÷ (50). If these cells turn out to express y/~i TCR,it would CDw29 interesting to determine whether they are memorycells.

Rat Studies Rat CD45also exists in different molecular weight species (51, 52). monoclonal antibody, OX-22, has been shown to react with the higher molecular weight forms of rat CD45but not with the 180 kd-form (53). ÷ cells This antibody stains approximately two-thirds of peripheral CD4 + (51). Functional data support the view that OX-22-CD4 T cells in the ÷ CD4÷ T cells in humans. In rat are homologous to UCHL1÷ CDw29 particular, it has been shownthat, in immunizedrats, this subset was much ÷ CD4 ÷ subset in providing help for more potent than the reciprocal OX-22 B cells (54). Whileinitial studies suggested that alloreactive proliferating precursor cells were confined to the OX-22÷ CD4÷ subset, more recent work using different culture conditions shows no difference between the two subsets in this respect (D. W. Mason, personal communication). Moreover, a direct lineage relationship between the two CD4÷ T cell subsets has been demonstrated using in vivo transfer experiments. It is ÷ cells become OX-22- upon differentiation, now clear that OX-22+ CD4 ÷ whereas OX-22- CD4 cells remain OX-22-. While comparable studies concerning CD8÷ ceils have yet to be done, it is likely that a similar phenotypic change occurs amongthese cells, since rat CTLappear to be OX-22- (D. W. Mason, personal communication). Although many points remain to be established, the analogy betweenthe data obtained in the rat, ÷ T cells are naive the mouse, and the humanstrongly suggest that OX-22 ceils, whereas OX-22-T cells are memory(previously activated) cells.

CONCLUDING

REMARKS

Both B and T cells participate in the establishment of immunologic memory.Little is knownabout the differentiation pathwaythat results in the generation of memoryB cells, in addition to terminally differentiated

Annual Reviews


86

CEROTTINI

& MACDONALD

plasma cells. In contrast, there is evidence, at least in vitro, that memory T cells are the progeny of antigen-induced effector cells. MemoryT cells in mice, rats, and humanscan be phenotypically distinguished from naive T cells. Expression of most of the surface markers that fulfill this purpose is low in naive cells but is rapidly enhancedupontriggering by antigen. In contrast to other surface activation antigens, which are only transiently expressed, the enhanced expression of these markers is stable. A particularly useful antigenic system in this respect is the leukocyte common antigen encoded by the CD45gene. Differentiation of naive T cells into activated (and memory)cells is accompanied by the progressive loss expression of higher molecular weight forms and gain of a lower molecular weight form. Antibodies against these markers define two subsets in per+ and CD8 + T cells. In contrast to a previous interpretation, ipheral CD4 these two subsets appear to correspond to different maturational stages rather than two different functional lineages. ~ Phenotypic identification of memoryT cells has allowed a more definitive analysis of their functional properties. It nowseemsclear that memory cells produce greater amountsof lymphokines (particularly IFN-7 and IL4) than their naive counterparts. This property of memorycells probably accounts (at least in part) for their enhancedability to help B cells and provoke inflammation in vivo. Furthermore, indirect evidence suggests that antigen-specific memorycells express higher affinity TCRthan most naive T cells. This finding (if confirmedat the molecularlevel) is a direct verification of the clonal selection theory. Further insights into the nature of T cell selection by antigen should be forthcoming with the availability of specific markers for activated and memoryT cells. ACKNOWLEDGMENT Wewish to thank Josiane Ducfor her excellent secretarial assistance. Literature Cited 1. Celada, F. 1971. The cellular basis of the immunologicmemory.Pro~7. Aller~Ty 15: 22347 2. Eisen, H. N., Siskind, G. W. 1964. Variations in affinities of antibodies during the immuneresponse. Biochemistry 3: 9961008 3. Andersson, B. 1970. Studies of the regulation of avidity at the level of the single antibody-formingcell. The effect of antigen dose and time after immunization. J. Exp. Med. 132:77-88 4. Davie, J. M., Paul, W. E. 1972. Receptors on immunocompetent cells. V.

Cellular correlations of the "maturation" of the immuneresponse. J. Exp. Med. 135:660-74 5. Mtller, G., ed. 1987. Role of somatic mutation in the generation of lymphocyte diversity. Immunol. Rev. 96:5-162 6. Engers, H. D., MacDonald,H. R. 1976. Generation of cytolytic T lymphocytes in vitro. In ContemporaryTopics in Immunobioloyy, ed. W. Weigle, pp. 145-90. New York: Plenum 7. Cerottini, J.-C., Engers, H. D., MacDonald, H. R., Brunner, K. T. 1974. Generation of cytotoxic T lymphocytes

Annual Reviews


T-CELL in vitro. I. Response of normal and immunemousespleen cells in mixedlymphocyte cultures. J. Exp. Med. 140: 70317 8. Ryser, J.-E., MacDonald, H. R. 1979. Limiting dilution analysis of alloantigen-reactive T lymphocytes.III. Effect of priming on precursor frequencies. J. lmmunol. 134:128-32 9. MacDonald, H. R., Cerottini, J.-C., Ryser, J.-E., Maryanski, J. L., Taswell, C., Widmer,M. B., Brunner, K. T. 1980. Quantitation and cloning of cytolytic T lymphocytes and their precursors. Immunol. Rev. 51:93-123 10. Marrack, P., Endres, R., Shimonkevitz, R., Zlotnik, A., Dialynas, D., Fitch, F. W., Kappler, J. 1983. The major histocompatibility complex-restricted antigen receptor on T cells. II. Role of the L3T4product, J. Exp. Med. 158: 107791 11. Shimonkevitz, R., Luescher, B., Cerottini, J.-C., MacDonald, H. R. 1985. Clonal analysis of cytolytic T lymphocyte-mediated lysis of target cells with inducible antigen expression: correlation between antigen density and requirement for Lyt-2/3 function. J. Immunol. 135. 892-99 12. MacDonald, H. R., Glasebrook, A. L., Bron, C., Kelso, A., Cerottini, J.-C. 1982. Clonal heterogeneity in the functional requirement for Lyt-2/3 molecules on cytolytic T lymphocytes (CTL): possible implications for the affinity of CTLantigen receptors. Immunol. Rev. 68:89-115 13. Swain, S. L. 1983. T cell subsets and the recognition of MHC class. Immunol. Rev. 74:129-42 14. Truney, A., Albert, F., Golstein, P., Schmitt-Verhulst, A.-M. 1985. Calcium ionophore plus phorbol ester can substitute for antigen in the induction of cytolytic T lymphocytes from specifically primed precursors. J. Immunol.135: 2262q57 15. Isakov, N., Altman, A. 1985. Tumor promoters in conjunction with calcium ionophores mimic antigenic stimulation by reactivation of alloantigen-primed murine T lymphocytes. J. Immunol. 135: 3674-80 16. Tabi, Z., Lynch, F., Ceredig, R., Allan, J. E., Doherty, P. C. 1988. Virus-specific memoryT cells are Pgp-1÷ and can be selectively activated with phorbol ester and calcium ionophore. Cell. Immunol. 113:268 77 17. Sanders, M. E., Makgoba,M. W., Shaw, S. 1988. Human naive and memory T cells: reinterpretation of helper-inducer

MEMORY 87

and suppressor-inducer subsets. ImmunoL Today 9:195-98 18. Hughes, E. N., Mengod, G., August, J. T. 1981. Murine cell surface glycoproteins. Characterization of a major componentof g0,000 daltons as a polymorphic differentiation antigen of mesenchymalcells. J. Biol. Chem.256: 702327 19. Hughes, E. N., Colombatti, A., August, J. T. 1983. Murine cell surface glycoproteins. Purification of the polymorphic Pgp-I antigen and analysis of its expression on macrophagesand other myeloid cells. J. Biol. Chem.258: 101421 20. Morse, H. C. III, Shen, F.-W., H/immerling, U. 1987. Genetic nomenclature for loci controlling mouselymphocytes antigens. Immunogenetics 25:71-78 21. Colombatti, A., Hughes, E. N., Taylor, B. A., August, J. T. 1982. Gene for a major cell surface glycoprotein of mouse macrophages and other phagocytic cells is on chromosome2. Proc. Natl. Acad. Sci. USA 79:1926-29 22. Lesley, J., Trowbridge, I. 1982. Genetic characterization of a polymorphic murine cell-surface glycoprotein, lmmunogenetics 15:313-20 23. Trowbridge,I. S., Lesley, J., Schulte, R., Hyman, R., Trotter, J. 1982. Biochemical characterization and cellular distribution of a polymorphic, murine cell-surface glycoprotein expressed on lymphoid cells. Immunogenetics15: 299312 24. Bauman, J. G. J., Wagemaker, G., Visser, J. W. M. 1986. A fractionation procedure of mouse bone marrow cells yielding exclusively pluripotent stem cells and committedprogenitors. J. Cell. Physiol. 128:133-42 25. Lesley, J., Hyman,R., Schulte, R. 1985. Evidence that the Pgp-1 glycoprotein is expressed on thymus-homing progenitor cells of the thymus. Cell Immunol. 91: 397-403 26. Trowbridge,I. S., Lesley, J., Trotter, J., Hyman, R. 1985. Thymocyte subpopulation enriched for progenitors with an unrearranged T-cell receptor/~chain gene. Nature 315:666-69 27. Lesley, J., Trotter, J., Hyman,R. 1985. The Pgp-1 antigen is expressed on early fetal thymocytes. Immunoyenetics 22: 149-57 28. Lynch, F., Ceredig, R. 1988. Ly-24 (Pgp1) expression by thymocytes and peripheral T cells. Immunol. Today 9:7-10 29. Budd, R. C., Cerottini, J.-C., Horvath, C., Bron, C., Pedrazzini, T., Howe, R. C., MacDonald,H. R. 1987. Distinc-

Annual Reviews


88

CEROTTINI

& MACDONALD

tion of virgin and memory T lymphocytes. Stable acquisition of the Pgp-1 glycoprotein concomitant with antigenic stimulation. J. Immunol.138:312029 30. Lynch, F., Chaudhri, G., Allan, J. E., Doherty, P. C., Ceredig, R. 1987. Expression of Pgp-I (or Ly24) by subpopulations of mouse thymocytes and activated peripheral T lymphocytes. Eur. J. Immunol. 17:137~,0 31. Lesley, J., Schulte, R., Trotter, J., Hyman, R. 1988. Qualitative and quantitative heterogeneity in Pgp-1 expression among murine thymocytes. Cell. Immunol. 112:40-54 32. Budd, R. C., Cerottini, J.-C., MacDonald, H. R. 1987. Phenotypic identification of memory cytolytic ÷ T lymphocytes in a subset of Lyt-2 cells. J. Immunol. 138:1009-13 33. MacDonald, H. R., Budd, R. C., Cerottini, J.-C. 1989. Pgp-1 (Ly24) as marker of murine memory T lymphocytes. In Curr. Top. Microbiol. Immunol. In press 34. Budd, R. C., Cerottini, J.-C., MacDonald, H. R. 1987. Selectively increased production of interferon-~, by subsets of Lyt-2 + and L3T4+ T cells identified by expression of Pgp-1. J. Immunol. 138:3583 86 35. Mosmann,T. R., Coffman, R. L. 1987. Twotypes of mousehelper T-cell clone. lmmunol. Today 8:223-27 36. Isacke, C. M., Sauvage, C. A., Hyman, R., Lesley, J., Schulte, R., Trowbridge, I. S. 1986. Identification and characterization of the human Pgp-I glycoprotein. Immunogenetics 23:326-32 37. Sanders, M. E., Makgoba, M. W., Sharrow, S. O., Stephany, D., Springer, T. A., Young, H. A., Shaw, S. 1988. Human memory T lymphocytes express increased levels of three cell adhesion molecules (LFA-3, CD2, and LFA-I) and three other molecules (UCHLI, CDw29,and Pgp-1) and have enhanced IFN-y production. J. Immunol. 140: 1401-7 38. Morimoto, C., Letvin, N. L., Distaso, J. A., Aldrich, W. R., Schlossman, S. F. 1985. The isolation and characterization of the humansuppressor inducer T cell subset. J. Immunol. 134:1508-15 39. Morimoto, C., Letvin, N. L., Boyd, A. W., Hagan, M., Brown, H. M., Kornacki, M. M., Schlossman, S. F. 1985. The isolation and characterization of tl~e human helper inducer T cell subset. J. lmmunol. 134. 3762~59 40. Beverley, P. C. L., Merkenschlager, M., Terry, L. 1988. Phenotypic diversity

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

of the CD45antigen and its relationship to function. Immunology(Suppl.) 3-5 Ralph, S. J. M., Thomas, M. L., Morton, C. C., Trowbridge, I. S. 1987. Structural variants of humanT200 glycoprotein (leukocyte-common antigen). EMBOJ. 6:1251-57 Lefranqois, L., Puddington, L., Machamer, C. E., Bevan, M. J. 1985. Acquisition of eytotoxic T lymphocyte-specific carbohydrate differentiation antigen. J. Exp. Med. 162:1275-93 Terry, L. A., Brown, M. H., Beverley, P. C. L. 1988. The monoclonal antibody, UCHL1, recognizes a 180,000 MWcomponent of the human leucocyte-common antigen, CD45. Immunology 64:331-36 Serra, H. M., Krowka,J. F., Ledbetter, J. A., Pilarski, L. M. 1988. Loss of CD45R (Lp220) represents a postthymic T cell differentiation event. J. Immunol. 140:143541 Akbar, A. N., Terry, L., Timms, A., Beverley, P. C. L., Janossy, G. 1988. Loss of CD45R and gain of UCHLI reactivity is a feature of primed T cells. J. Immunol. 140:2171-78 Lewis, D. B., Larsen, A., Wilson, C. B. 1986. Reduced interferon-gamma mRNAlevels in human neonates. Evidence for an intrinsic T cell deficiency independent of other genes involved in T cell activation. J. E.~cp. Med. 163: 1018-23 Moretta, A., Pantaleo, G., Moretta, L., Cerottini, J.-C. and Mingari, M. C. 1983. Direct demonstration of the clonogenic potential of every human peripheral blood T cell. Clonal analysis of HLA-DR expression and cytolytic activity. J. Exp. Med. 157:743-54 Moretta, L., Pende, D., Bottino, C., Migone, N., Ciccone, E., Ferrini, S., Mingri, M. C. and Moretta, A. 1987. Human CD3÷4-8-WT31 - T lymphocyte populations expressing the putative T cell receptor y-gene product. A limiting dilution and clonal analysis. Eur. J. Immunol. 17:1229-34 Goodman, T., Lefran~;ois L. 1988. Expression of the y-6 T-cell receptor on intestinal CD8+ intraepithelial lymphocytes. Nature 333:855-58 James, S. P., Fiocchi, C., Graeff, A. S., Strober, W. 1986. Phenotypic analysis of lamina propria lymphocytes. Gastroenterology 91:1483-89 Wooltett, G. R., Barclay, A. N., Puklavec, M., Williams, A. F. 1985. Molecular and antigenic heterogeneity of the rat leukocyte-common antigen from thy-

Annual Reviews

T-CELL


mocytes and T and B lymphocytes. Eur. J. Immunol. 15. 168-73 52. Barclay, A. N., Jackson, D. I., Willis, A. C., Williams, A. F. 1987. Lymphocyte specific heterogeneity in the rat leukocyte commonantigen (T200) is due to differences in polypeptide sequences near the NH2-terminus. EMBOJ. 6: 1259-64 53. Spickett, G. P., Brandon, M. R., Mason, D. W., Williams, A. F., Wollett, G. R. 1983. MRCOX-22, a monoclonal anti-

MEMORY

89

body that labels a new subset of T lymphocytes and reacts with the high molecular weight form of the leukocytecommonantigen. J. Exp. Med. 158: 795810 54. Arthur, R. P., Mason, D. 1986. T cells that help B cell responses to soluble antigen are distinguishable from those producing interleukin 2 on mitogenic or allogeneic stimulation. J. Exp. Med. 163. 774-86





Ann. Rev. lmmunol. 1989. 7.’91-109 Copyright © 1989 by Annual Reviews Inc. All rights reserved

MICROANATOMY OF LYMPHOID TISSUE DURING HUMORAL IMMUNE RESPONSES: Structure Function Relationships A. K. Szakal Department of Anatomyand Division of Immunobiology, MedicalCollege of Virginia, Richmond,Virginia 23298 M. H. Kosco and J. G. Tew Departmentof Microbiology and Immunology,Medical College of Virginia, Richmond,Virginia 23298 INTRODUCTION The complexcellular interactions involved in the regulation of immune responseshavegenerally beenstudied in vitro. Animplicit assumptionis that cells performthe sameroles in culture as they do in vivo. However, it shouldbe appreciatedthat secondarylymphoidtissues are highly organized and that this functional architecture is destroyedin the process of suspendingcells. The aim of this review is to relate novel structures observedin the lymphoidnoduleor follicle with inductionof the secondary antibody(Ab)response.Special attention is focusedon antigen (Ag)recognition, Agprocessing, and Agpresentation steps that appear to occur in the lymphoidnodule. In contrast with the primaryresponse, Abis present whenthe booster immunizationis given. This Ab binds the immunogen, and immunecomplexes are rapidly formed. These immunecomplexes localize in lymphoidnodulesin minutes, germinalcenters developrapidly, and Ab production and memoryB-cell production are accelerated. These 91 0732-0582/89/0410-0091 $02.00

Annual Reviews

92

SZAKAL,

KOSCO

& TEW

events are documented in draining.lymph nodeswith reference to other lymphoid tissue for comparison.


KINETICS OF THE GERMINALCENTER REACTION Basedon histochemical observations of Ag localization and germinal center (GC)development in mouselymphnodes, the kinetics illustrated in Figure 1 emerge. Horseradishperoxidase (HRP)immunemice were boostedin the hind feet with HRP.Notethat the popliteal lymphnode morethandoubledin size in the first three days.Thisincreaseis largely due to development of the Ag-retainingreticulum(ARR)andedemawhich resolvesby day5. Byday10 the nodeis againenlarged,but this is dueto increasedcellularity attributablemostlyto a proliferativeexpansion of B cells in germinalcenters. TheAg-retaining reticulumrepresentssites of Ag localization on follicular dendritic cells (FDCs)in the light zone germinalcenters (Figure2a). Notethat Ag-retainingreticulumreachtheir maximum number(Figure 1) and size between days 1 and 3 after injection. Byday5 these parameters decreaseto a plateauof aboutfive ARR per node. Theincrease in Ag-retainingreticulumcoincides with the initial peakin lymphnodesize, andit is followedby a size reduction 1.60

10

T

111 Z

1.28-

-8

Z

"2

:~

0.96-

0.~4-

0.32"

0.00

÷

0 Ag 15’ D:I

.3

5

TIME AFTER ANTI(~EN INJECTION (rain.

1’0 & days)

Figure ] This chart illustrates size changesin pop]Jtea] lymphnodesas related to the kinetics of antigen retaining reticulum (ARR) and de novo germinal center (GC) development after antigen (Ag) injection in the hind feet of immune C57BL/6 mice. Note: lymph node volume reflects changes in ARRand PNA + GC numbers and that ARRdevelopment (Ag trapping) precedes B-cell proliferation (development of PNA + GC).

Annual Reviews


LYMPHOID NODULE STRUCTURE FUNCTIONS

93

Figure 2 Light micrographs of (a) an antigen retaining reticulum (ARR)and (b) a GCin the popliteal lymph node 5 days after Ag injection in immunemice. (a) Note the location of the ARRin the GCindicated by arrow heads. Faintly stained TBMare also visible below the ARR;(b) Observe the highly PNA÷cells in the GC. Note that the upper half stains darker than the lower half which also contain some PNA+ TBM(arrows). This staining is the reverse of the convention on light and dark zones of GCs. The ARRis located in the upper darker staining half. ( × 70.)

paralleling the drop in the number of ARR.Using HRP-labeled peanut agglutinin (PNA)to bind activated germinal center lymphocytes (1), novo germinal center development can be monitored (Figure 2b). Germinal center development, dependent on Ag trapping in the ARR(2) begins between the third and fifth day after challenge, but new germinal centers develop only after preexisting germinal centers dissociate (3). Note (Figure 1) that the maximumnumber of germinal centers corresponds with the maximumnumber of ARR.The number of germinal center B cells peaks about day 10 and then begins to decline (1).

KEY EVENTS IN THE SECONDARY RESPONSE In primed animals the immunogen encounters Ab and is rapidly complexed. These complexes are quickly trapped by macrophages and Ag transporting cells in the lymphatic vessels and lymphnode (4). Figure 3 a model relating events in the lymphoid nodule to the induction and maintenance of the secondary Ab response. These events are described in subsequent sections.

ANTIGEN TRAPPING

AND ANTIGEN TRANSPORT

In the secondary response, the majority of complexesare phagocytized by sinusoidal macrophages, degraded and eliminated from the lymph node by 24-48 hr. Someimmunecomplexesare transported via an "alternative" pathway to lymphoid nodules. Early studies noted an apparently "purposeful movement"of Ag-bearing cells from-the splenic red pulp to the

Annual Reviews 94

SZAKAL,

KOSCO

& TEW

ANTIGEN TRAPPING, TRANSPORT, PROCESSING,

IMMUNOLOGICAL EVENTS

PRESENTATION ANDFEEDBACK

-- ANTIGEN TRAPPING BY ATCAND~ ANTIGEN TRANSPORT ATCs ~ EDCs


ANTIGEN PRESENTATION TO B-CELLS BY FDCs ICCOSOME DISPERSION __ ANDUPTAKE BY B-CELLS ANDTBM ANTIGEN PROCESSING BY -- GERMINAL CENTER B-CELLS, TBM.ANDPRESENTATION TO T-CELLS

]

’~ANTIBODY FEEDBACK REGULATION

Figure 3 Immunological events in the "alternative pathway" of the Ag-stimulated draining lymph nodes of immune mice. ATC, antigen transport cells; SS, subcapsular sinus; M, macrophage; FDC, follicular dendritic cell; ICS, immune complex coated body or "iccosome"; Bm, B memory cell; TBM, tingible body macrophage; Ab, antibody; pAg, processed antigen; Time designations: 1 r~irq-5 HR; DI, day 1; DI-3, days 1 to 3; D5, day 5; D14, day 14; --indicate events prominent at these times.

peripheral aspects of germinal centers (5). In 1983, based on light and studies, we described the transport of horseradish peroxidase in lymph nodes (4). The Ag was transported on the surface of cells of varying dendritic morphology, from the subcapsular sinus (SS) to lymphoid nodules. The Agin transit was recognizable as early as 1 min after injection (Figure 4a). Monocyte-like cells with immunecomplex coated veil-like processes and cells referred to as penetrating "frilly" or "veiled" cells (6) (Figure 4b) appeared to transport the immunecomplexes through pores in the floor of the subcapsular sinus. Werefer to these nonphagocyticcells collectively as Ag transport cells (ATCs). One min after HRPinjection, Ag-transport cells are found near one another below the subcapsular sinus floor with their processes extending through pores into the subcapsular sinus, forming a "mitten-like" dendritic configuration. These short processes, coated with immunecomplexes, appear to be in different stages of withdrawal from the sinus (Figure 4c). SomeAg-transport cells located below the subcapsular sinus had dendritic processes in contact with ATCs in transit through the floor of the subcapsular sinus and with processes of less differentiated FDCsdeeper at the periphery of lymphoid nodules (Figure 4d). Thus, a progression of the immunecomplexes, detectable

Annual Reviews


LYMPHOII) NODULE STRUCTURE FUNCTIONS

95

Figure 4 Light (a) and electron microscopic (b-d)histochcmical illustration of the PO+ inverted "funnel-like" Ag transport site (a),extending from the SS toward the future site of

ARR in the lymphoid nodule. (b) and (c), ATC (arrow heads) in transit through a pore (arrows) in the floor of the SS. Note the immune complex coated surface (black, PO+) of convoluted dendritic cell processes (P) extending into the SS lumen. (d) A chain (arrow heads) of ATC (pre-FDC?) in the Ag transport site with PO+ immune complex coated dendritic processes. Note, only the surface of dendrites but not the surface of lymphocytes are P O + . ATC,antigen transport cell; CA,capsule; FL, SS floor; L, lymphocyte; P, P1, P2, ATC cell processes; G,Golgi; N, nucleus. Magnification: (a) x 225; (b) x 1500; (c) x 1990; (d) x 2250 (Szakal et al, in Ref. 4).

on the surface of a chain of ATCs and FDCs, forms the profile, light microscopically visible, of an inverted funnel between the subcapsular sinus and lymphoid nodule (Figure 4a). As Ag-retaining reticulum develop, Ag transport cells diminish and Ag transport paths disappear. By day 1 most of the transported Ag is retained on the dendrites of FDCs in the fully developed ARR. However, not all ARR develop simultaneously, and Ag transport may continue into the first day at some sites (7).

Annual Reviews 96

SZAKAL~ KOSCO & TEW


Somestudies implicate lymphocytes in Ag transport (8-10), and immune complexes reportedly maydrain passively into lymphoid nodules (11). general, preformed immunecomplexes were used, and this may relate to differences in observations. Weinject Ag for booster immunization, and we routinely detect transport by Ag transport cells. The single exception is in old mice where the ATC-mediatedtransport is defective. In aged animals Ag persists on cells near the subcapsular sinus and never reaches the lymphoid nodule (7). ANTIGEN-RETAINING

RETICULUM

Onceit was accepted that Agis retained extracellularly on dendritic processes of nonphagocytic cells, "antigen retaining reticulum" (ARR)was proposed (12) to replace the term "phagocytic reticulum." The ARR refers to the three-dimensional network formedby the interdigitating, Agretaining dendritic processes of neighboring follicular dendritic cells as visualized by autoradiographic or histochemical detection of Ag (Figure 5a) (2, 7, 13). The de novo formation of Ag-retaining reticulum is clearly observable in lymph nodes 0f mice (Figure 1). The dendritic processes trapping Ag have been shownto belong to FDCs(14-16), and it is accepted that Ag is retained in the form of Ag-Abcomplexes(17-19). Descriptions of lymph node and splenic FDCsat the EMlevel agree about the cytological features (15, 20-23). Typically, follicular dendritic cells are described as having a cell body with an irregular, sometimes bilobed, euchromatic nucleus containing a distinct nucleolus. A few multinucleate cell bodies also exist (23). The scanty cytoplasm contains few profiles of mitochondria, RER,Golgi, and vesicles. Emanating from the cell body are the dendritic processes, some of which appear attenuated, with folds and intermittent thickenings forming pleiomorphic cytoplasmic extensions. Other dendrites form more uniform, highly convoluted, labyrinthine configurations. These dendritic processes interdigitate with those of neighboring FDCs and are the primary structures involved in the retention of Ag. Recent scanning EMon follicular dendritic cells identified two major FDCtypes (24, 25): (a) FDCswith filiform dendrites (Figure 5b), and FDCswith "beaded" dendrites (Figure 5d). In addition, intermediate forms (Figure 5c) between filiform and beaded type FDCsand FDCswith pleiomorphic thickening of dendrites and veil-like processes were seen. Culture studies indicated that the filiform dendrites mature into "beaded" dendrites. High resolution EMstudies revealed a spiraling periodicity (440490 ~) of immunecomplexes on FDCprocesses (25). Lymphocytes were seen

Annual Reviews



97

Figure 5 The antigen retaining reticulum (ARR). and its component FDC types. (a) Light microscopic view of the Ag (HRP) positive ARR at day 1 after Ag challenge. Scanning electron microscopy shows two types of FDC (arrows): (b) FDC with filiform dendrites; (d) FDC with beaded dendrites, and (c) intermediate form FDC. Magnification: (u) x 140; (b) x 2000; (c) x 2000; (d) x 2800 (Szakal et al, in Ref. 25).

bind to immune complexes on FDC dendrites in vivo, but they did not sequester free complexes on their surface. The periodicity of immune complex binding may facilitate the interaction between B cells and immune complexes. This interpretation is consistent with the report that in order to bind and directly stimulate B cells, an immunogen (e.g. DNP-polymer) must have a certain number of epitopes spaced 120-670 8, apart (26). Phenotypic studies indicate that follicular dendritic cells are a distinct cell type. Monoclonal Ab that selectively bind FDCs include MRC Ox 2 (27, 28), and Ki-M4R in the rat (29) and R4/23 or DRC-I (30), and KiM4 (3 1) in the human. The follicular dendritic cells express class I and I1 MHC Ags as well as common leukocyte Ag and receptors for C3b, Ig,

Annual Reviews

98

SZAKAL, KOSCO & TEW

and Fc (32). However, they lack macrophageor T-cell Ags such as Mac1, -2, -3; F4/80; Thy-1, CD5,or CD8(28, 32).


ICCOSOME FORMATION, DISPERSION, ATTACHMENT, AND ENDOCYTOSIS This year we reported the discovery of a 0.3~0.4/~m diameter spherical particle, or "immunecomplex coated body," which we termed iccosome (13). Iccosomes are produced by follicular dendritic cells between day and day 3 after secondary challenge. Their discovery by EMwas facilitated by the use of the Ag horseradish peroxidase which after histochemical developmentis seen to outline these particles (Figure 6a). The existence of iccosomes in developing germinal centers appears to be limited to a period of hours at an early phase of germinal center development. Since the development of one germinal center may lag behind another by 24 hr, their visualization by EMis somewhatfortuitous. In addition, Ag dose and the level of circulating Abs appear to influence iccosome production. The EMsuggests (13) that iccosomes are formed through interaction two FDCtypes. The first population with pleiomorphic dendrites binds large accumulations of immunecomplexes. "Beaded" dendrites of the second FDCtype bind to areas of sequestered complexes. This binding proceeds in a way similar to a receptor-mediated endocytosis, resulting in the formation of an immunecomplex layer between the membraneof the pleiomorphic dendrite surrounding the "bead" and the limiting membrane of the "bead." Thus, the "bead" becomescoated with a uniform thickness of immunecomplexes and also with a layer of the pleiomorphic dendrite (cytoplasmic layer) (Figure 6b). This cytoplasmic layer remains incomplete; therefore the process does not result in true endocytosis. Next the "bead" is apparently penetrated by small globules of immunecomplexes, seen as PO+ material in the "bead" (Figure 6c). Subsequently, the cyto-

Figure 6 Electron micrographs of iccosome formation and dispersion in developing GC. (a) Shows iccosomes (PO +, small immunecomplex coated bodies) dispersed amongthe lymphocytes; note edema; (b) an FDCbead embeddedin the pleiomorphic dendrite is in the early stage of the formation of the immunecomplex layer (arrow); (c) an iccosome in intermediate stage of formation. Note membraneslimiting the (black) immunecomplex layer (arrow heads), the globular immunecomplexes in the center and the cytoplasmic layer (arrow); (d) free iccosomeattaching to the surface of a lymphocyte;(e) a recently endocytosed iccosome in a GClymphocyte. Magnification: (a) ×4380; (b) × 108,100; (c) × 101,500; (d) × 103,500; (e) × 12,150.


Annual Reviews

Annual Reviews

100

SZAKAL, KOSCO & TEW

plasmic layer disintegrates, releasing the iccosome which is coated by and contains varying amounts of immune complexes (Figure 6 4 . Around day 3, iccosomes are dispersed in the developing germinal center, an event associated with edema (Figure 6a) that appears to facilitate iccosome dispersion. The free iccosomes become attached to germinal center B cells via their immune complex layer and are endocytosed (Figure 6e). By day 5, iccosomes are rarely visible in germinal centers, but the majority of the lymphocytes contain PO endocytosed HRP complexes (Figure 7a). These events culminate in the delivery of the Ag to germinal center B cells, for endocytosis, degradation (1 3), and, potentially, Ag presentation (35).


+

ANTIGEN PRESENTATION BY GERMINAL CENTER B CELLS It is now well known that B cells can present Ag to T cells and that their specific surface Ig receptors make them remarkably efficient. The iccosome-derived Ag appears to be processed by germinal center B cells as indicated by the decreasing intensity of the HRP reaction product at the Golgi apparatus (Figure 7a). The Golgi is known to be associated with

Figure 7 Endocytosis and presentation of in vivo obtained Ag by GC B lymphocytes: (a) an electron micrograph showing the Ag, HRP in endocytic vesicles 5 days after Ag injection in immune mice. Note the reduced PO activily of Ag associated with the Golgi (arrow); (0) bar graph illustrating the results of the presentation of Ag (OVA) to 3DO-54.8 T cells in vitro. Note Ag presentation is maximal at day 5 when the majority of GC B cells endocytose and process Ag. Magnification: (a) x 20,600 (Szakal et al, in Ref. 13); (b) Kosco et al, in Ref. 35.

Annual Reviews



101

production of plasma membrane and could serve as a mechanism for delivering the Ag-Ia complexto the cell surface (13). Wesought to determine whether the germinal center B cells were able to present the Ag they internalized in vivo to T cells that are knownto be in germinal centers (33, 34). Germinalcenter B cells were isolated using peanut agglutinin panning (35). The germinal center B cells were obtained at various times after Ag challenge and assessed for their ability to stimulate T cells to produce IL-2. The results of the Ag presentation experiments (Figure 7b) correlate with the kinetics of Aguptake in vivo (13). Shortly after challenge, whenfew iccosomes are available to germinal center cells (i.e. Day0, Day1), very little IL-2 production is elicited. At day 5, when the maximumamount of Ag appears to be processed by germinal center B cell in vivo, maximum IL-2 production is stimulated. Finally, as the response is winding downin vivo, the ability to stimulate T cells also declines. It appears that germinal center B cells can efficiently process and present Ag they obtained from FDCs. ROLE

OF

TINGIBLE

BODY MACROPHAGES

(TBM)

Tingible bodymacrophages(36) are large cells containing the characteristic "tingible bodies" of Flemming (37). Tingible body macrophages are thought to derive from macrophages entering the lymph node via the afferent lymphatics (38). Kinetic analysis at the light and EMlevels during de novo formation of germinal centers revealed that: (a) by day 1, close association with the antigen retaining reticulum a few tingible body macrophages were present with low numbers of endocytosed cells; (b) day 3, TBMwere seen endocytosing iccosomes and immune complexes from FDCdendrites (13); (c) by day 5, Ag+TBMcontaining increased numbers of endocytosed cells were present in large numbers in the dark zone of germinal centers (Figure 8a). The observed uptake of iccosomes suggested that tingible body macrophages maypresent Ag to germinal center T cells. This prompted tingible body macrophage isolation and phenotyping studies, and TBMwere found to be unique. In addition to macrophagemarkers (Ia, F4/80, Mac-l, Mac2) they were Tby-I positive (39). The significance of Thy-I (Figure 8c) unknown,although it may act as an adhesion molecule (39) to help bind cells for endocytosis. Tingible body macrophagecould endocytose T cells and thus dampen Ag presentation. Alternatively, TBMexpress Ia and could present Ag. Wehope to test these possibilities. Proliferation and cell death are features of germinal centers (40). Tingible body macrophagesappear to play a major role in disposing of dead cells or cells destined to die. Originally, Cottier proposed(41) that "tingible

Annual Reviews


102

SZAKAL, KOSCO & TEW

Figure 8 Illustrates the morphology of Ag (HRP), Thyl, and Ia positive tingible body macrophages (TBM). (a) TBM in a GC day 5 after Ag injection of immune mice. Note

the Ag+ TBM (black, PO+) associated with the PO+ ARR retaining HRP-anti-HRP complexes. TBM below the AKR are less positive (arrows). (b) An isolated TBM stained with methyl green to show nuclei of phagocytized GC lymphocytes; N, TBM nucleus; (c) electron (EM) and light (inset) micrographs of TBM showing membrane and cytoplasmic labeling for Thy-I with anti-Thy-1 and PO conjugate (EM) and glucose oxidase conjugate (inset); (d) Ia+ TBM labeled with anti-la and PO conjugate (EM) and anti-Ia and glucose oxidase conjugate (inset). Magnification: (a) x 95; (b) x 730; (c) x 3150; inset, x 570; ( d ) x 3038; inset, x 540 (Smith et al, in Ref. 39).

bodies” may originate from the breakdown of germinal center cells. DNA labeling studies (40, 41) indicated that “tingible bodies” derive from germinal center cells dying before prophase when intact “tingible bodies” show a 4n DNA content. Concurrent EM studies revealed that “tingible

Annual Reviews 103



bodies" represent lymphocytesand differentiating plasma cells in various stages of degradation (36). The significance of tingible body macrophagesmayrelate to recent data on affinity maturation of B cells. In strain A mice, B cells use a single heavy-chain V-gene segment to make most p-azophenyl-arsonate (Ars)specific Ab (42). In the secondary response, B cells are present that have mutatedderivatives of this V region. Mostof these B cells have even higher affinity for Ars than do the original B cells. This phenomenonhas been explained by an affinity-based selection mechanism(42). Antigen directs both the expression of the immunerepertoire and the amplification of V region diversity by selecting mutated cells with the highest affinity. However,manymutations would result in surface Ig on B cells with lower affinity for the antigen. Since these cells wouldnot bind Ageffectively, a mechanismfor eliminating these cells wouldbe desirable. Germinalcenters represent an in vivo structure with features matching those needed for affinity maturation (43). These features include: (a) Availability of Ag follicular dendritic cells in the light zone for selection of appropriate B cells; (b) rapid proliferation of B cells in the dark zone allowing for expression of mutations; and (c) destruction of B cells in large numbers by tingible body macrophages. Phagocytosis by TBMwould provide a mechanismfor eliminating mutated cells with low affinity receptors that cannot successfully compete for Ag at low concentration. These B cells may be programmedto die in the absence of stimulation by antigen, and this would be consistent with the observation that TBMare phagocytosing dead or dying cells (39). Thus, affinity maturation could be explained terms of the germinal center microenvironment. GERMINAL ANTIBODY

CENTERS FORMING

AND THE CELLS

INDUCTION

OF

The lymphoid nodule is recognized as a center for production of memory B cells (1). However, several studies, including one of our own, have reported that cells of the plasmacyticseries are also within germinalcenters as is indicated in Figure 9 (21, 44~6). The number of antibody forming cells (AFC) peaks about 3 to 4 days after Ag challenge. Note that the amount of anti-HRP antibodies released into the endoplasmic cisternae increases with antibody forming cell maturity. The observation of antibody forming cells within the germinal center led us to investigate whether these cells were capable of secreting Ab. At various times after challenge, germinal center B cells were isolated and cultured. The germinal center B cells obtained 3 or 4 days after challenge spontaneously produced (no Ag added) high amounts of HRP-specific

Annual Reviews


104

SZAKAL, KOSCO & TEW

Flqure 9 Immunoperoxidase localization of anti-HRP antibodies (black) in cells of the GC plasmocytic series (see lower left corner and insets) and on the convolutions of FDC dendrites retaining HRP-anti-HRP complexes (see upper right: arrows). Note (lower left) arrows heads indicate the beginning differentiation of a plasmocytic cell with anti-HRP antibodies in its nuclcar cnvelopc (scc also insct a). Insets b and c show a proplasmocyte and a plasmocyte (G, Golgi) respectively. Magnification: x 3200; insets: x 4900 (Szakal et al, in Ref. 44).

IgG which could not be augmented by T-cell factors. In contrast, germinal center B cells obtained 7-12 days after challenge produced small amounts of anti-HRP. However, production of anti-HRP by the 7-12 day population was dramatically enhanced when Con A supernatant was added (44; M. H. Kosco, unpublished findings). Two distinct phases of the germinal center reaction exist. During the first phase, germinal center B cells receive signals needed to differentiate into antibody forming cells. In the second phase, which peaks about day 10, cell proliferation results in the restoration and expansion of the memory B-cell pool. We observe many PNA positive cells leaving germinal centers

Annual Reviews LYMPHOID NODULE STRUCTURE FUNCTIONS

105

during the first phase. Recent data suggest that these germinal center cells home to the medullary cords and to bone marrow where they become mature plasma cells which produce much of the Ab associated with the anamnestic response (47).


FEEDBACK REGULATIONIN MAINTENANCEOF THE ANTIBODY RESPONSE The role of follicular dendritic cells, persisting Ag, and specific Abin the maintenance and regulation of humoral Ab production was reviewed in 1980 (48) and is not covered in detail here. The concept is that in the maintenance phase of immuneresponses, multiple dynamicequilibria exist between persisting Ag on FDCs, free specific Ab, and Ag-Abcomplexes of various ratios. The immunogenicityof the complexesis directly related to the Ag-Abratio, and alterations in serum Ablevels result in formation or dissociation of these complexes. WhenAb levels in the circulation decline, Agis exposed, memoryB cells are stimulated, and a new cycle of Ab synthesis is induced. This newly produced Ab feeds back and terminates the immunogenicstimulus. Repetition of this cycle serves to maintain serum Ab levels within narrow limits. Evidence supporting the role of FDCsand associated Ag in the maintenance mechanismis summarized in Table 1.

Table 1 Evidence supporting a role for FDCsand retained antigen in the maintenance of serum antibody 1. Specific Abtiters "spontaneously" rebound to the previous levels after severe bleeding (49). 2. Specific antigen persists intact on FDCsfor years and retains its immunologicalspecificity

(50). 3. Persisting Agis restricted to draining lymphoidorgans. The further the organ is from the injection site the less Agit retains (50). Correspondingly,specific AFCare localized the draining organs. AFCare most numerous in organs which contain the most FDC associated Ag(51). 4. As Agis depleted, the AFCbecomemore restricted in distribution. One year after immunizationin the hind foot, the AFCare almost exclusively localized to the popliteal lymph nodes (51). 5. Removingpopliteal lymphnodes with their high levels of persisting Agfrom mice injected in the hind feet results in a markeddecrease in serumAb(48). 6. Removinglymph nodes from the Abrich environment in vivo and culturing results in a "spontaneous Ab response". This "spontaneous response" is subject to Abfeedback inhibition (48). The "spontaneous Abresponse" only occurs in nodes with retained

(48).

Annual Reviews 106

SZAKAL,


SUMMARY

KOSCO

AND

& TEW

SIGNIFICANCE

Secondary responses require no more immunogen than do primary responses, and lower doses are often adequate. In view of the feedback inhibitory effects of specific Ab, one might predict that secondaryresponses would require more Ag. However, a special pathway is operative in immune animals for handling immunogen. This Ag pathway, which efficiently utilizes the immunogen,is described in this review. A brief summaryof key events in this Ag pathway is described below and illustrated in Figure 3. In immuneanimals, immunogenencounters specific Ab and is rapidly converted into immunecomplexes. Most of these complexes are quickly trapped, endocytosed, and catabolized by macrophagcs. This occurs quickly, and by 48 hr after immunizationvery little Agpersists in or on these macrophages. However, an alternative Ag pathway exists which tends to preserve the immunogen. Immunecomplexes in the alternative pathwayare trapped on the surface of nonphagocytic cells which transport the immunogento the follicular dendritic cells in the lymph node outer cortex. During the next 3 days FDCdendrites bead and form immune complex-coated bodies or "iccosomes." The iccosomes are dispersed in the developing germinal centers where they are trapped and endocytosed by germinal center B cells. The germinal center B cells process the iccosome derived Ag, and this Ag appears to be presented to T cells in the area. During this early period of iccosomerelease, endocytosis, and Agprocessing, the lymph node is swollen and edematous. The edema appears to facilitate movementof antibody forming cells and B cells which have developed and are leaving the germinal center microenvironment. These departing AFCand stimulated germinal center B cells appear to be going to the medullary cords, and some may migrate to the bone marrowwhere large amounts of specific Ab are produced. The edemain the lymph node resolves somewhatby day 5, but the germinal center reaction continues with a second lymph node enlargement apparent by day 10. This enlargement of the lymph node is associated with a proliferative expansion of germinal center B cells. These germinal center B cells are memorycells, and an Ag-basedselection of high affinity B cells is thought to be taking place. Cycles of Ab production regulated by an antibody feedback mechanism maintain circulating levels of specific Abfor monthsor years. Cyclic release oficcosomesmaybe responsible for the induction of specific Ab. Iccosomes have not yet been identified in the maintenanceof Ab production; however, the concept that a mechanisminvolving iccosomes is responsible for both the induction and maintenance of the secondary response is appealing.

Annual Reviews LYMPHOID NODULE STRUCTURE FUNCTIONS

107


Literature Cited 1. Coico, R. F., Bhogal, B. S., Thorbecke, G.J. 1983. Relationship of germinal centers in lymphoid tissue to immunologic memory. VI. Transfer of B cell memory with lymph node cells fractionated according to their receptors for peanut agglutinin. J. Immunol. 131:2254-57 2. Klaus, G. G., Humphrey,J. H., Kunkl, A., Dongworth, D. W. 1980. The follicular dendritic cell: its role in antigen presentation in the generation of immunological memory. Immunol. Rev. 53:3-28 3. Hanna, M. G. Jr., Congdon, C. C., Wust, C. J. 1966. Effect of antigen dose on lymphatic tissue germinal centers. Proc. Soc. Exp. Biol. Med. 121:286-91 4. Szakal, A. K., Holmes,K. L., Tew, J. G. 1983. Transport of immune complexes from the subcapsular sinus to lymph node follicles on the surface of nonphagocytic cells, including cells with dendritic morphology. J. Immunol. 131: 1714-27 5. Nossal, G. J., Austin, C. M., Pye, J., Mitchell, J. 1966. Antigens in immunity. XII. Antigen trapping in the spleen. Int. Arch. Alleryy 29:368-83 6. Fossum,S. 1980. The architecture of rat lymph nodes. 1I. Lymphnode compartments. Scand. J. Irnmunol. 12:411-20 7. Szakal, A. K., Taylor, J. K., Smith, J. P., Kosco, M. H., Burton, G. F., Tew, J. G. 1988. Morphometry and kinetics of antigen transport and developing antigen retaining reticulum of follicular dendritic cells in lymphnodes of aging immunemice. Aging: Immunol. and Inf. Dis. 1: 7-22 8. Brown, J. C., Harris, G., Papamichail, M., Slijvic, V. S., Holborow,E. J. 1973. The localization of aggregated human gammaglobulin in the spleens of normal mice. Immunolo#y 24:955-1003 9. Gray, D., McConnell, I., Kumararatne, D. S., MacLennan, I. C., Humphrey, J. H., Bazin, H. 1984. Marginal zone B cells express CR1 and CR2 receptors. Eur. J. lmmunol. 14:47-52 10. Heinen, E., Braun, M., Coulie, P. G., Van Snick, J., Moeremans, M., Cormann, N., Kinet Denoel, C., et al. 1986. Transfer of immune complexes from lymphocytesto follicular dendritic cells. Eur. J. Immunol. 16:167-72 11. Kamperdijk, E. W. A., Dijkstra, C. D., Dopp, E. A. 1987. Transport of immune complexes from the subcapsular sinus into the lymphnode follicles of the rat. ImmunobioL 174:395~405 12. Mitchell, J., Abbot, A. 1965. Ultra-

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

structure of the antigen-retaining reticulum of lymph node follicles as shown by high resolution autoradiography. Nature 208:500-2 Szakal, A. K., Kosco, M. H., Tew, J. G. 1988. A novel in vivo follicular dendritic cell-dependent iccosome-mediated mechanism for delivery of antigen to antigen-processing cells. J. Immunol. 140:341-53 Nossal, G. J., Abbot, A., Mitchell, J. 1968. Antigens in immunity. XIV. Electron microscopic radioautographic studies of antigen capture in the lymph node medulla. J. Exp. Med. 127:263-76 Szakal, A. K., Hanna, M. G. Jr. 1968. The ultrastructure of antigen localization and viruslike particles in mouse spleen germinal centers. Exp. Mol. Pathol. 8:75-89 Tew, J. G., Thorbecke, G. J., Steinman, R. M. 1982. Dendritic cells in the immuneresponse: characteristics and recommendednomenclature. J. ReticuloendotheL Soc. 31:371-80 Nossal, G. J., Ada, G. L., Austin, C. M., Pye, J. 1965. Antigens in immunity.VIII. Localization of 125-I-labelled antigens in the secondary response. Immunolo#y 9:349-57 Humphrey, J. H., Frank, M. M. 1967. The localization of non-microbial antigens in the draining lymphnodes of tolerant, normal and primed rabbits. Immunoloyy 13:87-100 Chen, L. L., Frank, A. M., Adams, J. C., Steinman, R. M. 1978. Distribution of horseradish peroxidase (HRP)anti-HRP immune complexes in mouse spleen with special reference to follicular dendritic cells. J. Cell Biol. 79:184-99 Nossal, G. J., Abbot, A., Mitchell, J., Lummus,Z. 1968. Antigens in immunity. XV. Ultrastructural features of antigen capture in primary and secondary lymphoid follicles. J. Exp. Med. 127: 277-90 Hanna, M. G. Jr., Szakal, A. K. 1968. Localization of ~25I-labeled antigen in germinal centers of mousespleen: histologic and ultrastructural autoradiographic studies of the secondary immune reaction. J. Immunol. 101:949-62 Chen, L. L., Adams, J. C., Steinman, R. M. 1978. Anatomyof germinal centers in mousespleen, with special reference to "follicular dendritic cells". J. Cell Biol. 77:148~4 Fossum, S., Vaaland, J. L. 1983. The architecture of rat lymphnodes. I. Combined light and electron microscopy of

Annual Reviews


108

SZAKAL, KOSCO & TEW

lymph node cell types. Anat. Embryol. 167:229.46 24. Schnizlein, C. T., Kosco, M. H., Szakal, A. K., Tew, J. G. 1985. Follicular dendritic cells in suspension: identification, enrichment, and initial characterization indicating immune complex trapping and lack of adherence and phagocytic activity. J. Immunol. 134:1360 68 25. Szakal, A. K., Gieringer, R. L., Kosco, M. H., Tew, J. G. 1985. Isolated follicular dendritic cells: cytochemicalantigen localization, Nomarski, SEM, and TEMmorphology. J. Immunol. 134: 1349-59 26. Dintzis, R. Z., Middleton, M. H., Dintzis, H. M. 1983. Studies on the immunogenicityand tolerogenicity of Tindependent antigens. J. Immunol. 131: 2196-2203 27. Barclay, A. N. 1981. The localization of populations of lymphocytes defined by monoclonal antibodies in rat lymphoid tissues. Immunology 42:593 600 28. Humphrey, J. H., Grennan, D. 1982. Isolation and properties of spleen follicular dendritic cells. Adv. Exp. Med. BioL 149:823~7 29. Wacker, H. H., Radzun, H. J., Mielke, V., Parwaresch, M. R. 1987. Selective recognition of rat follicular dendritic cells (dendritic reticulum cells) by a new monoclonal antibody Ki-M4Rin vitro and in vivo. J. Leukocyte. Biol. 41: 7077 30. Naiem, M., Gerdes, J., Abdulaziz, Z., Stein, H., Mason, D. Y. 1983. Production of a monoclonal antibody reactive with humandendritic reticulum cells and its use in the immunohistological analysis of lymphoid tissue. J. Clin. Pathol. 36:167-75 31. Parwaresch, M. R., Radzun, H. J., Feller, A. C., Peters, K. P., Hansmann, M. L. 1983. Peroxidase-positive mononuclear leukocytes as possible precursors of humandendritic reticulum cells. J. Immunol. 131:2719-25 32. Kosco, M. H., Tew, J. G., Szakal, A. K. 1986. Antigenic phenotyping of isolated and in situ rodent follicular dendritic cells (FDC)with emphasis on the ultrastructural demonstration of Ia antigens. Anat. Rec. 215: 201-13, 219-25 33. Nieuwenhuis, P., Opstelten, D. 1984. Functional anatomyof germinal centers. Am. J. Anat. 170:421 35 34. Rouse, R. ¥., Ledbetter, J. A., Weissman, I. L. 1982. Mouselymph node germinal centers contain a selected subset of T cells--the helper phenotype. J. Immunol. 128:2243M6 35. Kosco, M. H., Szakal, A. K., Tew, J. G.

1988. In vivo obtained antigen presented by germinal center B cells to T cells in vitro. J. Immunol. 140:354-60 D. C., Congdon, 36. Schwartzendruber, C. C. 1963. Electron microscope observations on tingible body macrophagesin mousespleen. J. Cell Biol. 19:641-46 37. Flemming,W. 1885. Studien uber regeneration der Gewebe. Arch. Mikrosk. Anat. 24:50 38. Hoefsmit, E. C. M., Kamperdijk, E. W. A., Balfour, B. M. 1980. Reticulum cells and macrophages in the immune response. In Mononuclearphagocytes: Functional aspects, Part II, ed. R. van Furth, pp. 1809-36. Boston: Martinus Nijhoff 39. Smith, J. P., Kosco, M. H., Tew, J. G., Szakal, A. K. 1988. Thy-1 positive tingible body macrophages (TBM)in mouse lymph nodes. Anat. Rec. 222: In press 40. Odartchenko, N., Lewrenze, M., Sordat, B., Roos, B., Cottier, H. 1966. Kinetics of cellular death in germinal centers of mouse spleen. In Germinal centers in immune responses, ed. H. Cottier, D. Odartchenko, R. Schindler, C. C. Congdon, pp. 212-17. New York: Springer-Verlag 41. Fliedner, T. M. 1966. On the origin of tingible bodies in germinal centers. See Ref. 40, pp. 218-21 42. Manser, T., Gefter, M. L. 1986. The molecular evolution of the immune response: idiotope-specific suppression indicates that B cells express germ-line encoded V genes prior to antigenic stimulation. Eur. J. Immunol. 16: 143944 43. MacLennan,I. C., Gray, D. 1986. Antigen driven selection of virgin and memory B cells. Immunol. Rev. 91:61-85 44. Szakal, A. K., Kosco, M. H., Burton, G. F., Tew, J. G. 1988. Germinal center antigen presenting B cells in the induction and maintenance of the secondary antibody response. Progr. Leuk. Biol. 7: 281-90 45. Sordat, B., Sordat, M., Hess, M. W., Stoner, R. D., Cottier, H. 1970.~ Specific antibody within lymphoid germinal center cells of mice after primary immunization with horseradish-peroxidase: a light and electron microscopic study. J. Exp. Med. 131:7741 46. Terashima, K., Imai, Y., Kasajima, T., Tsunoda, R., Takahashi, K., Kojima, M. 1977. An ultrastructure study on antibody production of the lymph nodes of rats with special reference to the role of germinal centers. Acta Pathol. Jpn. 27:1~4 47. Benner, R., Hijmans, W., Haaijman, J. J. 1981. The bone marrow: the major

Annual Reviews


LYMPHOIDNODULESTRUCTUREFUNCTIONS 109 source of serum immunoglobulins,but Progress in Immunology,p. 628. New still a neglectedsite of antibodyformaYork: AcademicPress tion.. Clin. E,,cp. Immunol. 46:1-8 50. Mandel,T. E., Phipps,R. P., Abbot,A., 48. Tew,J. G., Phipps,R. P., Mandel,T. E. Tew,J. G. 1980.Thefollicular dendritic 1980. The maintenanceand regulation cell: long term antigen retention during of the humoralimmuneresponse: perimmunity. Immunol.Rev. 53:29-59 sisting antigenandthe role of follicular 51. Donaldson,S. L., Kosco,M. H., Szakal, antigen-bindingdendritic cells as accesA. K., Tew,J. G. 1986. Localizationof sory cells. Immunol.Rev. 53:175-201 antibody-formingcells in draining lym49. Bystryn,J. C., Schenkein,I., Uhr,J. W. phoid organs during long-term mainten1971.Amodelfor the regulationof antiance of the antibodyresponse.J. Leukobody synthesis by serumantibody. In cyte. Biol. 40:147-57





Ann. Rev. Immunol. 1989. 7.’111~43 Copyright © 1989 by Annual Reviews Inc. All rights reserved

CELLS AND MOLECULES THAT REGULATE B LYMPHOPOIESIS IN BONE MARROW Paul W. Kincade, Grace Lee, Carolynn E. Pietrangeli, Shin-Ichi Hayashi and Jeffrey M. Gimble Oklahoma Medical Research Foundation, 825 Northeast Thirteenth Street, Oklahoma City, Oklahoma 73104

Introduction Mature and immature forms of eight different types of blood cells are tightly packed within the spaces of bone marrow, complicating investigation of howit is efficiently regulated. Hundredsof billions of cells of the various lineages are produceddaily in this vital organ and exported to other tissues via processes which are responsive to inflammation and other systemic events. Because of advances in cell separation and culture techniques, rapid progress is nowbeing made in resolving steps in each differentiation pathway. It is even moreinteresting that regulatory interactions betweencells are being appreciated and defined in molecular terms. This review focuses on cells of the humoral immunesystem and those steps involved in their formation that can be observed and manipulated in culture. Purified recombinantmolecules can nowbe used to elicit particular responses in cultured lymphocyte precursors, and the probable source of such regulatory substances is becoming clearer. Moreover, cells that comprise the inductive microenvironment of bone marroware themselves subject to exquisite regulation. In somecases, this occurs via factors they can themselves make, i.e. autocrine regulation. Information of this kind is relevant to the treatment of a numberof diseases. However, while cell culture modelshave madeit possible to identify manymolecules that affect proliferation and differentiation of B-cell precursors, they do not provide a completely accurate representation of intact bone marrow. For example, there are indications that Compensatorycircuits and "quality control" 111 0732-0582/89/0410-0111502.00

Annual Reviews 1-12

KINCADE ET AL


mechanismsthat operate in vivo are not present in established tissue cultures. Thus,recognition of the deficiencies in experimentalapproaches is informative with respect to the complexityof bonemarrowand should highlightareas for future investigation. Organization Within Bone Marrow Transmissionand scanning electron microscopicstudies have provided a general picture of bonemarrowarchitecture and areas of hemopoiesis(1 5). Someof the space within bonesis occupiedby fat (yellow marrow) is not engagedin blood cell production. Active red marrowcan expand into such areas in circumstances of unusual demand, and a molecular explanation of this response maybe near (see below). Stem cells and hemopoieticprogenitor cells are in extravascular areas and tend to be concentratedin the subendostealarea, i.e. near the bone cortex (6, 7). Newlyformedblood cells leave the marrowafter traversing the endothelium of venoussinuses, whichare morecentrally located within bones

(8). Platelets are madeand released from megakaryocytesand are conspicuousbecauseof their large size in marrowsections, while erythrocytes are madein close association with a centralized macrophage-likecell in what is knownas an "erythroblastic island" (9). Thus, these two blood cell types are madein highly specialized locations. Theorigin of other bloodcells is moredifficult to discern andrequires extrapolationfromin vitro studies. Maturinggranulocyteshavebeenidentified in close association with large "adventicular reticular cells," whichradiate awayfrom venoussinuses (10). Early B-lineage lymphocytesare moreabundant the subendostealarea of marrow;however,areas of focal proliferation comparable to the tbllicles in the avian bursa of Fabricius are not readily apparent(l 1, 12). B-lineageand myeloidcells growin close association with large "stromar’cells in culture (see below),and it seemslikely that they wouldbe near the extendedmembranes of similar cells in situ. Theincidence of multipotential stem cells amongtheir differentiated progenyis extremelylow in bonemarrow(less than 0.1%), whereasnearly a quarter of the cells are B-lineagelymphocytes.Therefore,a considerable amountof division mustaccompany the transition of one cell type to the other. Indeed,the kinetics andextent of proliferation in B-cell precursors is well documented (13). Since large clusters of lymphocytesare not conspicuous, it seemslikely that differentiating cells must be continually -.moving,perhaps in a "conveyorbelt" relationship along the convoluted ¯ membranes of stromal cells. Directed migration of maturingB-lineage cells could be passive, e.g. with movement resulting frompulse pressure and modulationof adhesionmolecules,or alternatively, chemotacticprocesses

Annual Reviews B-CELL

FORMATION

1 13

might be involved. Further details of the three-dimensional structure of bone marrowshould be revealed when distinctive markers are found for critical components of the lymphopoietic microenvironment. Such tools are likely to be developed by exploitation of long-term cell culture techniques.


Long-Term Bone Marrow Cultures Short-term culture systems madeit possible to conclude that close cellular interactions favor B lymphocyte formation, and a number of potential regulators of lymphopoiesishave been identified with this relatively simple approach (14, 15). However, more impressive advances in understanding bone marrow function followed the development of long-term bone marrow culture techniques (16). Dexter and colleagues first found conditions suitable for maintenance of multipotential stem cells and sustained granulopoiesis (17). Whitlock &Witte (18;-19) successfully adapted these methods for selective growth of B-lineage lymphocytes, and other innovations facilitated analysis of the cells involved (20, 21). Theseimportant technical advances have been reviewed elsewhere (15, 22). Only a brief account is given here, based primarily on our personal experiences. Hemopoietic cells from bone marroware maintained in long-term cultures by close physical association with a complexlayer of adherent cells. Usually a majority of the adherent cells are macrophages,whichare i~eadily distinguished on the basis of marker expression, phagocytosis, and uptake of acetylated low density lipoprotein (LDL)(23, 24). A number of stances that influence lymphopoiesis are elaborated by macrophages, and several potential regulatory mechanisms can be proposed that involve them. However, macrophage-depleted, long-term cultures sustain the growth of lymphocytes that have already beenadapted to growth, and this proliferation is supported by one or more types of "stromal cells" (see below). In some respects, lymphocytes propagated in culture resemble their normal counterparts in bone marrow. Despite a relatively high mitotic index, they are notably small to mediumin size and die within three days when removedfrom the adherent layer (18, 22, 23). Also, at least some cells taken from pooled cultures are capable of normal differentiation following transplantation (25-28). However,we have not consistently been able to elicit responses in long-term cultured lymphocyteswith a variety of stimuli knownto be effective with freshly isolated lymphocytes from bone marrow, or pre-B tumor cell lines. Pre-B cells maintained for extended periods in vitro often have unusual phenotypes. For example, the BP-1 antigen can be absent or expressed in abnormally high density (29; G. Lee et al, submitted) and,~unlike normal bone marrowpre-B cells,

Annual Reviews


114

KINCADE ET AL

mayfail to stain with 14.8 antibody (23). It is possible that a majority the lymphocytesmaintainedfor an extended period in vitro are intrinsically defective, and studies with bone marrowfrom immunodeficient SCIDmice are consistent with this view (29, 30). That is, lymphocyteswith aberrantly rearranged immunoglobulingenes expand in culture, but not in vivo. Ceils growing in individual culture dishes seem to arise from a small numberof surviving lymphocytes, and rare abnormal cells might have a growth advantage (23, 31). Therefore, populations of cultured lymphocytes could become more abnormal with time. Whensteroids and/or horse serum are used in the culture mediumas originally described by Dexter et al (17), multipotential stem cells are maintained and granulocytes predominate. Reduction of incubator temperature to 33°C is also an advantage. Use of relatively low concentrations of selected fetal calf serumand addition of 2-mercaptoethanol, as described by Whitlock & Witte, favor lymphocyte growth (18, 19). Shifting established Dexter-type culture to Whitlock-Witte conditions results in a switch from myeloid to lymphoid cell production (20, 32). Additional procedures add flexibility to the use of long-term cultures. For example, hemopoietic cells can be selectively eliminated by treatment with mycophenolic acid or 5-fluorouracil, leaving a functional adherent layer (21, 33). Fresh adherent layers can also be established with an initial low numberof bone marrowcells (19). Cells that can readily form an adherent layer in culture can be dramatically reduced by passage of suspensions through G-10 Sephadex or nylon wool columns (21). However, it. should be cautioned that precursors of macrophagesand stromal cells are probably not removedby this treatment (see below). While there are limitations to the existing long-term culture methodology, these approaches have madeit possible to dissect an extraordinarily complex tissue and identify at least some of the cell types critical to lymphocyte formation. Moreover, lines established from these cultures provide a means to clone genes for unique regulatory molecules. Stromal

Cell

Clones

Adherent layers of primary long-term bone marrow cultures are complex and include macrophages,endothelial cells, fat cells, and fibroblasts, in addition to the lympho-hemopoieticcells that are in close physical association with them (15, 24, 34, 35). Manyinvestigators have sought to simplify this by establishing cloned cell lines that can be subcultured manytimes. These are then used to investigate which cells are responsible for specific microenvironmental functions. Salient characteristics of some clones known to influence B lymphopoiesis are summarized in Table 1. Considerable phenotypic and functional diversity has been found in adherent

Annual Reviews B-CELL FORMATION

115

Table 1 Someroutine stromal cell clones that influence B lymphopoiesis


MBA-2.4 Endothelial-like line; makesM-CSF; membrane associated factors stimulate a transformed pre-B line and inhibit plasmacytoma growth(36) B.AD Fibroblastoid,phagocytic,pre-adipocyteline; makesCSFbut not IL-3; requires close contact with target cells to supportcompactmultipotentcolonies, includinglymphocyte precursors (38-40) TC-1 Fibroblast/endothelialline; makesM-CSF anda high MW synergistic activity for stemcells, anda factor whichstimulates pre-B formation,but not IL-3 (41, 42) S 17 Clonedslromal line; allows nylon woolpassagedbone marrowto makemyeloidcells and support B lineage precursors, whichexpandunder Whitlock-Witteculture conditions; Cells are la-, Thy-1-, Mac-I- and BP-1-; producesa factor whichstimulates pre-B cell formation and IL-4 (43-45) AC 4 Endothelial-like, pre-adipocyteclone; maintainsgrowthof a clonedpre-Bline and sorted bonemarrow cells withoutterminal differentiation; these andrelated clonesexpressthe 6C3antigen, but not Thy-1and makeM-CSF,but not IL-3 (37) ALC Fibroblastoid,pre-adipocyteclone; supportspre-Bclones as well as cultured myeloidcells; makesa 30-40Kpre-B growth factor, M-CSF and G-CSF,but not GM-CSF or IL-3 (46) 30R Stromalcell clone; supports growthof a clonedpre-Blymphocyteline; has IL-4 receptors and 1-2%are 6C3+ (47, 48) IxN/A6 SV40transformedstromalcell clone; used to purify andclone IL-7 (49, 50) BMS-2 Pre-adipocytestromalcell line; supportsgrowthof five clonedlymphocyte lines; part of a series whosephenotypeswerecompared(Table 2) and studied with respect to factor responsiveness;cells are Thy-1÷ and makeM-CSF (33) Stromalcell clone; supports growthof Blineage progenitorsfromvery early (9.5 day) embryos(51)

cell clones, and this was particularly notable when large numbers were isolated and studied in a single laboratory (33, 36, 37; D. Rennick, personal communication). Although many morphologic designations have been used, the term "stromal cell" is convenient because it does not imply a known lineage derivation. Our cloned stromal cells, like others reported in the literature, all have a large oval nucleus, with one or more nucleoli.

Annual Reviews


116

KINCADEET AL

The cytoplasmcan be highly spread, giving a total cell diameter of at least 150 microns. Hemopoieticcells placed in low numberson a stromal cell clone survive and grow in a typical functional assessment. However,even when highly purified bone marrowstem cells are added, it should be rememberedthat a simple precursor-adherent cell interaction maynot be taking place. For example, myeloid progenitors rapidly give rise to macrophages in this situation, and the adherent layer thus includes them in addition to the cloned line. Manylaboratories have also isolated cloned lymphocytelines from longterm cultures and used them to define stromal cell functions (33, 37, 46, 47, 49, 52). In our experience, the characteristics of these clones are as unusual as lymphocytes in the long-term cultures from which they were derived. They display unusual phenotypes and have been unresponsive to inductive stimuli (G. Lee, unpublished observations). Nonetheless, they provide a more homogeneous tool for the assessment of cytokine production and other stromal cell activities. Our panel of stromal cell clones was namedon the basis of the tissue of origin and interactions with cloned lymphocytes (33). If the lymphocytes quickly died when placed on the adherent layer, we designated them "nonsupport," as opposed to "support" stroma. For example, BMNS1 was isolated from bone marrow and did not support the growth of a particular lymphocyte clone. Impressive proliferation of lymphocytes in liquid or methyl cellulose cultures occurred when "support" clones were used, and this was usually improved by prior irradiation of the adherent layer. Indeed, division of the stromal cell clones could be influenced by mesenchymalgrowth factors, and this tended to relate inversely to good lymphocyte growth support (33). Functional heterogeneity of our stromal cell clones becameapparent by use of multiple clones of lymphocytes as indicators (C. E. Pietrangeli, manuscript in preparation). For example, one stromal cell clone (BMS2) supported all of five lymphocyte clones and another (SNS1) supported none of them. However,other stromal cells efficiently supported the proliferation of only certain of the lymphocyteclones. This important observation indicates that multiple lymphocytegrowth stimuli potentially can be provided by stromal cells. Somelymphocyte clones need more than one of these, whereas others could have simpler requirements. An alternative explanation might be that particular clones make inhibitors that affect only certain of the lymphocyte clones. As previously noted by Whitlock (22), lymphocyte clones occasionally becomestromal cell independent and would presumably then be tumorigenic. This has also happened occasionally in our lab, particularly with clones derived from BALB/cmice.

Annual Reviews 117


B-CELL FORMATION

Thestability of stromalcell clonesis of practicalandtheoreticalinterest. Wehave preparednonsupportsubclones of BMS2 andexpect these variants will be veryvaluablefor associatinggeneexpressionwithfunctional capability. Certaincharacteristics of the stromalcells seemto change duringcloning (Table 2). For example,the Thy-1antigenwasnot demonstrable in primaryculturesbut is expressedby mostof our stromalcell clones. This mayindicate an intrinsic phenotypicinstability of stromal cells, whichis made obviousby cloning.Also, veryrarecells in the primary cultures mayhavebeenexpandedby cloning. However, a moreinteresting possibility is that their geneexpressionis regulatedby interactionwith macrophages andother cells in primarycultures. As is discussedin more detail below,stromalcell clonesare responsiveto a number of regulatory stimuli. Marrowsuspensions that have been passed through G-10 Sephadex columnsandplatedat low densities simplydie withoutestablishment of an adherentlayer. When placedon a supportstromalcell clone, hemopoiesis is initiated. Macrophages are alwaysproduced,andin manycases, verylarge colonies of themare noticeable. Granulocytopoiesis can also occur, even whenWhitlock-Witte conditions are used (C. E. Pietrangeli, manuscript in preparation).Publishedaccountsof other stromalcell clones suggest Table 2 Comparisonof clones to stromal cells in primarycultures

Markers

Primary culture

SSI

E Cadherin Lgp 100 Class II LFA1 Macl ...... 14.8 (CD45R) Ly5/T200 -dH-2D + _ BP 1 .... Actin + + N-CAM + + + M 1/69 + M 1/87 + -Mac 2 + Mac3 ++ Collagen (IV) + + + + Thy 1 + ++ Qa2

SNS1

BMS1

BMS2

BMNS1

...... ...... ...... ......

+ --

...... + _

+ + + + + +++ _+ ++ +

+ + + + + + + ++ + + + ++ +

+ _+ + + + + +++ + + + ++ +

-+ _ + + + +++ -_+ ++ +

Annual Reviews 118

KINCADEET AL


this is more noticeable whenDexter-type culture conditions are employed (37, 43, 46). With certain of our clones, under Whitlock-Witteconditions, B-lineage lymphocytesare produced, and there is only a minimal tendency for these cells to give rise to surface-Ig-bearing B cells. This is in accord with reports from other laboratories (46, 53) in that the support of terminal maturation is less markedthan stimulation of growth. While there is good reason to think that a single type of stromal cell can sustain growth of lymphoidand myeloid cells (37, 43, 46), we do not yet knowif this occurs simultaneously in vivo (see below).

Cytokines Madeby Stromal Cells At least eight regulatory macromolecules can be elaborated by cloned stromal cells under some circumstances. Macrophagecolony stimulating factor (M-CSF;CSF-1) is almost always produced. This has been detected by direct coculture of fresh bone marrowwith stromal cells, by assay of conditioned medium,by inhibition with specific neutralizing antibodies, and by analysis of messenger RNAs. Likewise production of G/M-CSF and/or G-CSFhas been demonstrated with bioassays and RNAanalysis (Table 1, and J. M. Gimble, manuscript in preparation). Wehave found that with somestromal cell clones, CSFcan be readily detected in culture supernates, whereas with others, close contact with the adherent cells is necessary (C. E. Pietrangeli, manuscript in preparation). This could be due to differential production of extracellular matrix componentsthat absorb such factors (54, 55, and see below). Interleukin 3 is a T cell-derived multipoietin which, together with IL-6, can stimulate multipotential stem cells and more mature progenitors of several lineages (56, 57). Therefore, it might be expected that stromal cells or some other component of bone marrowwould elaborate this cytokine. Thus far, no one has found evidence of IL-3 production in cloned stromal cells ~r implicated its function in typical long-term bone marrowcultures (Table 1; and 58). Our experience has been that this factor alone does not sustain proliferation of cells that are recognizablypart of this lineage (59). In addition, there is a marked genetic polymorphism of IL-3 responsiveness, and virtual nonresponder animals make adequate numbers of lymphocytes (see below). Therefore, this mediator mayonly affect hemopoiesis during unusual circumstances. A novel 25-kd factor, termed interleukin 7 (IL-7), has been purified and obtained in recombinant form from a transformed bone marrow stromal cell line (49, 50). Asis detailed below,IL-7 is a potent proliferative stimulus for large pre-B cells and maysignificantly contribute to lymphocytegrowth in long-term cultures. Indeed, IL-7 mRNA is detectable in our best support stromal cell clones, as are other growth factors. Althoughlong-term cul-

Annual Reviews


B-CELL FORMATION

119

tured lymphocytesproliferate in response to IL-7, it is rare for them to grow indefinitely (G. Lee, submitted, and A. Namen, personal communication). Together with other findings that indicate a complexpattern of lymphocyte-stromal cell interactions, this suggests that more than one replieative stimulus is usually required for lymphopoiesisin culture. Wehave detected mRNA for IL-6, TGF-/~, and neuroleukin in all of our stromal cell clones (J. M. Gimble, manuscript in preparation). IL-6 of interest for a numberof reasons. Amongeffects on manydifferent cell types, IL-6 prepares multipotential hemopoietic stem cells for responses to other stimuli (57, 60). As is detailed below, IL-6 appears to autoregulate its ownsynthesis in stromal cells. TGF-fl is madeby manycell types in an inactive precursor form (61, 62). Whenactivated by acidification, it is potent inhibitor of lymphopoiesis and hemopoiesis (see below). Neuroleukin, a B lymphocyte activator, has recently been shown to be highly homologousto glucose-6-phosphate isomerase (63, 64). Stromal Cell Participation Responses

in Autocrine

and Paracrine

Hemopoiesis is dramatically affected by inflammation and exogenous stimulation (65), and it is nowclear that several componentsof the bone marrow are potentially involved in such responses. Macrophages can respond to bacterial endotoxin by release of IL-1 and TNF, which can in turn stimulate endothelial cells to produce IL-1 (66-68). Moreover, these two cell types, when induced, make M-CSF, G-CSF, and G/M-CSF(66, 67, 69, 70). Similarly, stromal cells are capable of constitutive and induced CSFproduction (71, 72). Constitutive production of M-CSFseems to a commonfeature of stromal cell clones, whereas monocyteor endothelial ceil-derived IL-1 causes them to increase expression of G/M-CSFand GCSFgenes (71-73 and Table 1). Wehave monitored the relative abundance of various mRNAs in a bone marrowderived stromal cell clone exposed to a numberof stimuli (J. M. Gimble, manuscript in preparation). For example, while the cells contain low levels of IL-6 message, this is increased following exposure to LPS, IL-1, 1L-7, or tumor necrosis factor (TNF). This type of data, using multiple cDNAprobes, indicates that functional receptors for IL-6, TGFfl, IL-4, and IL-7 are expressed on stromal ceils. Experiments involving cyclohexamide treatment suggest that some of these mRNA transcripts maybe produced under the influence of labile repressors or subject to a rapid degradation pathway(66, 74, 75). Growthof stromal cells in culture can be enhanced by addition of epidermal cell growth factor (EGF) and inhibited by LPSor interferon y (33). The function and ability of stromal cells to accumulate fat is modulated by TGF-fl’s, cytokines also knownto

Annual Reviews


120

KINCADE ET AL

affect synthesis of extracellular matrix constituents (76, 77; S.-I. Hayashi, manuscript in preparation). Other studies indicate that platelet-derived growth factor (PDGF) can also influence bone marrow microenvironmental cells (78). These observations suggest that paracrine regulatory networks could potentially regulate growth, differentiation, and function of bone marrow stromal cells (Figure 1). Activated T lymphocytes can make IL-6 and interferon ~, in addition to myeloidcolony stimulating factors (79). Macrophages are a potential source ofTNF,IL-1, and TGF-fl (66). T cell-derived IL-4 can directly influence hemopoietic progenitors (47, 80). Stromal cell

Fi#ure 1 Stromal cells play a pivotal role in regulating progression of stem cells down lymphoid and myeloid lineages. They are also involved in paracrine and autocrine interactions with T cells, macrophages,other stromal cells, and endothelial cells (not shown).

Annual Reviews


B-CELL FORMATION

121

clones display receptors for IL-4 and may also be influenced by this lymphokine (48, and K. Landreth, personal communication). Therapy with recombinant factors may soon become routine, and it is important to understand direct effects on stromal cells, as well as indirect responses achieved via a cytokine cascade (81). Information about the effects of hydrocortisone and other drugs on these cells is also being obtained with long-term culture approaches (82-85). Indirect evidence might be used to argue that stromal cells are regulated by neighboring cells in vivo. Stromal cell clones that support pre-B cells can be isolated from cultures of adult spleen (33). However,the spleen not an active site for B lymphopoiesis in normal adult mice. Stromal cell function could be markedly influenced by cytokines and close cellular interactions which together constitute "he, mopoietic inductive microenvironments"in discrete sites (86). Coordination of responses by cells in the bone marrowmicroenvironment may be achieved via an autocrine mechanism (J. M. Gimble, manuscript in preparation). This follows from findings that our stromal cell clone responds to factors that it can produce, including IL-6, TGF-fl, and IL-7. Precedence exists for autocrine regulation in studies of T cells, macrophages,endothelial cells, fibroblasts, and B cells (79, 87-91). Cells each of these lineages mayrespond to factors which they, or their cohorts, make. Such networks can recruit cells of the same type to participate in particular responses. Further study might also reveal sometype of feedback inhibition through which individual stromal cells are sensitive to the concentration of their soluble products. All stromal cell clones that support lymphopoiesis can, under some circumstances of culture, also stimulate myelopoiesis, and it is interesting to consider that these functions might be linked in vivo. B lymphocyte precursor production can be markedly dysregulated in cyclic neutropenia (92). Pre-B cell numbersincreased 50-fold while myeloid progenitors were low, consistent with a competition because these two lineages for some commonfeature. B lymphocyteproduction is also compromised with a granulocytosis-inducing tumor (93). Stromal cells could be a pivotal site for regulation in bone marrow, ensuring that appropriate numbersof each blood cell type are produced. For example, stromal cells actively engaged in lymphopoiesis might make little CSF, relative to IL-7, and we should soon knowbowexogenousstimuli such as IL-1 affect this balance (71, 94). This information could have implications for managementof a number of bone marrowdisorders. While it is not yet clear which elements of the bone marrow microenvironment are routinely established in transplant recipients, the feasibility of stromal cell engraftment has been experb ment~il!y, demonstrated(95, 96).

Annual Reviews 122

KINCADE

ET AL


Oriyin, Derivation, Stromal Cells

and Differentiation

Potential

of

The bone marrowstroma may be mesenchymalin origin, based on embryologic and other studies. During mammalian development, mesenchyme derived cells invade the marrowcavity (97). Withtime, these cells differentiate into osteoblasts that form a bone matrix. This step is then followed by the formation of the bone marrowstroma or adventitia and the appearance of lymphohematologicprecursor cells (97). In the adult, the bone marrowstroma consists of at least three morphologically distinct phenotypes: fibroblasts or endothelial-like cells, macrophages, and adipocytes. Transplantation, and immunochemical and mRNA studies performed on mixed and cloned stromal cell cultures, indicate that fibroblast/endothelial and adipocyte stromal cells are not derived from hematologic precursors (24, 33, 98, and J. M. Gimble, manuscript in preparation). Unlike macrophages, they do not express a number of commonleukocyte genes (24, 33). In our view, immunohistochemical analysis alone does not provide a sharp distinction between fibroblasts and endothelial cells. However,our clones lack scavenger receptors for acetylated low density lipoprotein and are similar to pericytes in expression of a smooth muscle isoform of actin (33, 99, 100, and unpublished observations with P. D’Amore).Thy-1 antigen expression is typical of our clones but was not found in primary cultures and is not on all stromal cell clones isolated by other investigators (Tables 1 and 2). It is interesting that a cloned epithelial component of the human thymus microenvironment expresses this marker (101). Additional evidence for the mesenchymalderivation of stromal cells comes from LTBMCstudies. Under Dexter conditions, two non-bone marrowderived cell lines can substitute as stroma to support myelopoiesis, the fibroblasts Swiss 3T3 and CH3/10T1/2(102, and J. M. Gimble, unpublished observations). Eachof these lines also has characteristics of multipotential mesodermalprecursors. Whentreated with azacytidine, either 3T3 or CH3/10T1/2can give rise to myocytes, adipocytes, and chondrocytes. Mechanisms regulating the differentiation of these fibroblasts during myogenesishave been attributed to the expression of a single gene (103105). That these cells can act as typical myeloid stroma in coculture experiments is consistent with a mesodermalorigin for the stromal cells. Manycloned bone marrow stromal cells are pre-adipocytes (106, 107, and Table 1). After a period of several weeksin culture, they can accumulate lipids and take on a fat cell morphology.The adipocyte differentiation process is markedly influenced in vitro by exogenousgrowth factors such as TGF-fl, an inhibitor of adipogenesis, and hydrocortisone, an enhancer

Annual Reviews


B-CELLFORMATION 123 of adipogenesis (10g, 109, and S.-I. Hayashi, manuscript in preparation). However, unlike extramedullary adipocytes, stromal cells do not respond to insulin (108, 110). In vivo, the natural history of the bone marrow the distal extremities is a progression from "red" hematopoietic marrow to "yellow" adipocyte laden, nonproductive marrow. With conditions of anemia, this process can be altered, with the reversion of "yellow" to "red" marrow (l 1 l). Bone marrow stromal cells may thus represent a type of multipotent mesenchymalcells, capable of further differentiation into adipocytes and possibly osteoblastic cells. Moreover,this differentiation process, while influenced by the age of the animal, can be reversed under certain conditions of hematologic stress. Developmentof gene probes and antibodies that detect unique products of these cells should be informative with respect to their lineage derivation and mayprovide insight into their normal lifespan and functional capabilities. Cell

Adhesion

Molecules

Somelymphocytes crawl beneath the stromal cell layer, in a curious phenomenon known as "pseudo-emperipolesis," and are flattened into foci which have a cobblestone appearance (24, 112). Other lymphocytes adhere to the surface of stromal cells, with sufficient strength to withstand the pull of gravity in inverted cultures (P. W. Kincade, unpublishedobservations). There is somespecificity to this attraction, because lymphocytes never bind to macrophages, which comprise a majority of the adherent cells (24). Specific recognition and adhesion betweenparticular cells is essential to normal bone marrow function, and molecules that could be involved in such functions are now being defined. In one well-studied example, a numberof different N-CAM (neural cell adhesion molecules) glycoproteins are madefrom a single complexgene (113, 114). Selective use of exons the N-CAM gene leads to the production of at least seven different mRNA species, and someof the resulting proteins are tissue specific (115, 116). For example, embryonicor regenerating muscle cells express a unique form of N-CAM,which has a domain not found in other tissues, and neural cells express a type of N-CAM not found in muscle cells. N-CAM was the first known adhesion molecule to be demonstrated on stromal cells derived from long-term cultures (117). The extracellular portions of N-CAM molecules can be extensively modified by carbohydrate addition, altering the overall charge of the molecule and its potential interactions with adhesion molecules on other ceils (118). However, anchorage of the carboxy terminal portions of the molecules is of particular interest. Most cell surface glycoproteins have one or more stretches of hydrophobic amino acids that float in the phos-

Annual Reviews


124

KINCADE ET AL

pholipids of the cell membrane.However,a subset uses a unique glycosylphosphatidylinositol (G-PI) linkage between the protein and cell (119). There is a great deal of flexibility in howcells use the G-PI anchor. In the case of N-CAM, only molecules derived from a particular exon become G-PI linked 020, 121). Others either have a short, or very long, C terminal intracellular domain. The latter might interact with cytoskeletal elements in cells, whereas G-PI linked N-CAM is subject to release via action of a specific enzyme. Phosphatidylinositol specific phospholipases (PI-PLC) have been isolated from bacterial, parasitic, and mammalian sources (122, 123). Experimentally, PI-PLCs can be used to remove G-PI anchored proteins selectively from cell surfaces 024). Such molecules might also be liberated from cells by endogenous phospholipases generated as a consequenceof particular stimuli (125). It seems possible that attachment, and subsequent release of adhering cells is controlled in part by this mechanism, and it is interesting that PI-PLC treatment causes many lymphocytesto detach from stromal cells in long-term cultures (24). N-CAM molecules are structurally related to other members of the "immunoglobulingene superfamily," which includes T cell receptors for antigen and manyother molecules knownto be involved in specific recognition functions (126, 127). It is thought that these all derived during evolution from a single primordial gene. Another gene family includes adhesion molecules that function in a quite different way 028-130). They are called cadherins because an extracellular domainrequires a divalent cation such as Ca++ or cell-cell recognition, and at least one memberof this family is expressed on B lineage lymphocytes (P. W. Kincade, unpublished observations). Still other receptors on pre-B cells mayrecognize fragments prepared from fibronectin 031). It is important to learn howall of these molecules participate in the orientation and movementof lymphocytes within the microenvironment. The association of granulocytes with stromal cells maybe mediated by a recently described adhesion molecule termed "haemonectin" (132). Immobilized

Cytokines

and the Extracellular

Matrix

Complexproteoglycans encompassall of the cells in bone marrowand, in addition to structural functions, these substances may play an important regulatory role, Heparan sulfate is knownto bind and stabilize a number ofmesenchymalgrowth factors, and this has recently been found to be true for GM-CSF (55, 133). Therefore, some of the most important regulatory molecules maynot alwaysact as diffusible soluble substances. Extracellular matrix components made by stromal cells or other specialized components of the marrowmay bind factors to achieve very high local concentrations in discrete microenvironments (54). There can also be transmembrane

Annual Reviews B-CELL FORMATION

125

forms of some mediators (134-137), and this provides a mechanismwhereby cells that make a substance can directly present it to adjacent cells. Experimentally it is mucheasier to detect mediators that are soluble than those that are cell bound. However, studies employing permeable membraneseparation of cells, conditioned medium, or matrix extracts demonstrate the importance of positive and negative regulators of hemopoiesis that are closely associated with adherent cells (54, 138, 139).


B-Lineage

Precursors

in the Mouse

The development ofmonoclonal antibodies and cell separation techniques considerably advanced our understanding of B-lymphocyte lineage progenitors, particularly the relatively abundant cells near the end of this differentiation pathway. Thoroughphenotyping has been done with large numbers of patient samples as well as established tumor lines, and the results are usually depicted by placing them into a single lineage (15, 140). However,it is still difficult to determine the degree of overlap and incidences of very small subsets of lymphoidprecursors. Webriefly summarize our experience with cells taken from murine bone marrow and advise somecaution in strict application of lineage diagrams that we and others have published. Anindirect subtraction technique was used to infer the composition of marrow from normal, young adult (B6xDBA2)F1mice as summarized Figure 2 (59). Samples were first depleted of mature B cells, and then negatively selected with one of several antibodies. The results of flow cytometric and microscopic analyses were then used along with absolute cell recoveries to calculate the incidences of cells bearing various markers. Essentially all precursors express Ly5-200 (equivalent to CD45Ron humans), Lym19, and J11D. Substantial display of class II antigen (Ia), M1/75heat stable, Mel 14, and Thy 1 antigens as well as receptors for transferrin and Fc~2b/7l receptor (2.4G2) was also documented. A very small subpopulation also appeared to bear IL-2 receptors. The BP-1 antigen is uniquely associated with late B-lineage precursors (141). The sizes oflymphocytes expressing various markers were also evaluated in terms of low angle light scatter by flow cytometry (59). Whilea majority of the Lyb-2÷ cells in marrowwere small, surface Ig populations bearing other markers generally included greater numbersof large cells (Figure 3). In another study, done with a different strain of mice, it appearedthat ThB was preferentially expressed on small lymphocytes (142). However, nearly half of the slg ,Th-B+ cells in our analyses have been large. Detailed kinetic studies have been done of large and small pre-B cells, cells containing terminal deoxynucleotidyl transferase (TdT), and cells bearing lineage antigens in rat and mouse bone marrow (143-148). Our flow

Annual Reviews


126

KINCADE ET AL

OtherMariners: Qa2 (15%) tL2R(3%l LessIhan1%: Lyt2 L3T4(GK1,5) M1/87 F4/80 Mac1 Mac2 Mac3 Grnt.2

Figure 2 Numerousantigen and receptors characterize the final three stages of B cell formation in murine bone marrow. Percentages of cells (range) expressing certain markers are given to indicate a certain degree of population heterogeneity.

SIZE PROFILESOF s lg7 Ly 5 (220)+ADULTBM CELLS small cells

= large cells

small cells

lar0e cells

I

~T~

y 5 (220) total

ThB+ + .,.~= - ,\

+

BP-1 +

Forward Angle Light Scatter (FALS) Figure 3 Size heterogeneity of B lymphocyte lineage precursors expressing particular markers is reflected in forward angle light scatter. Cells lacking surface immunoglobulin were purified from bone marrow with monoclonal 14.8 antibody and evaluated with other antibodies as indicated.

Annual Reviews


B-CELL FORMATION

127

cytometric analyses of cellular DNAcontent have been consistent with those findings, i.e. small lymphocytes are noncycling, and most of the mitotic activity is associated with large B-lineagecells. Manymonoclonal antibodies have been prepared to epitopes on the Ly5 (CD45) commonleukocyte antigen family (15). The large literature involving these reagents provides an example of complex gene regulation in lympho-hemopoieticcells. Multiple mRNA species are transcribed from a single large gene, and expression is controlled in a tissue-specific manner (149 151). The function of the resulting transmembraneglycoproteins not known, but their importance is suggested by substantial conservation during evolution. As has been described for CD45Rin the human(152), our monoclonal 14.8 antibody detects virtually all B-lineage lymphocytes and a subset of peripheral T cells in mature mice 045, 153, 154). This includes all B-cell precursors that can quickly maturein culture, and most that can do so after transfer to immunodeficient recipients (153, 155). However,pre-B cells in early embryos, sometransformed B-cell lines, and some of the B lineage lymphocytes that grow for prolonged periods in culture lack the epitope detected by this antibody (23, 153, 156, 157). This illustrates the importance of using more than one marker and establishing a "composite" phenotype before positioning any given normal or transformedcell type within the lineage. It also demonstrates that a strictly ordered series of differentiation steps need not always be followed as the progeny of stem cells progress downthis pathway. Physical differences in bone marrowcells, such as buoyant density and size, have been exploited along with lectin receptors, drug sensitivity, and marker expression in attempts to obtain enriched preparations of stem cells (53, 158-161). With multiparametercell sorting and sequential separation with various monoclonal antibodies, it has now becomepossible to obtain multipotential stem cells in virtually pure form (162). This remarkable achievement should soon lead to definitive diagrams of early differentiation steps as the cells are placed into all of the available assays with recombinantfactors and cloned stromal cell lines. The possibility has often been raised of branchedand/or parallel lineages of B-cell differentiation that result in functionally restricted cells. Most recently this has followed intriguing findings with Ly-I(CD-5) bearing cells (163, 164). As far as we are aware, none of the available data lymphocyte subpopulations preclude that possibility in bone marrow, or any other tissue. Factors That Stimulate B Lineage Precursors in Culture Extraordinary progress is being madein defining eytokines that are potentially important for regulating B lymphopoiesis. Responses detected in

Annual Reviews 128

KINCADE ET AL


culture studies suggest mechanisms through whichfactors might directly interact with specific receptors on B lineageprogenitors,or indirectly via effects on macrophages, endothelial cells, and stromal cells (Figure 1). Somecould be involved in normalsteady state B-cell production, whereas others maybe important only in situations such as obtain in trauma, inflammation, leukemia, or autoimmunity. Antagonists maybe no less importantthan agonists, and a numberof these have already been identified. It wouldbe surprising if these findings do not quicklylead to therapeutic applications, as is already occurring with a numberof cytokines

(81). Theactive substancesin "tumornecrosis serum," whichwas first found to influence B-cell formationin culture, are likely to be tumornecrosis factor (TNF)and interleukin 1 (IL-I) (15, 165). IL-1 induces kappa chain expression on a pre-B cell tumor line and maturation of normal precursors in cultures of B cell depleted bone marrow(166, 167). Recent findingssuggestthat the latter responsecan proceedat least partly via an indirect mechanism (45). IL-1 stimulated a cloned stromal cell line release IL-4, whichin turn inducedmaturationof small pre-Bcells. A low pHsensitive cytokine in T cell-conditioned mediumwhich had similar effects was subsequently demonstratedto be immuneinterferon (IFN-~) (166, 168-170).Selectivestimulation of cells near the end of the B lineage was also demonstratedwith factors present in the serumof youngNZB strain mice (171). Whiletwo of these substances have been purified homogeneity, their cellular origin is not known(172). Relatively early B lineage precursors in humanand murinebone marrow respondto substancesexcreted duringdiscrete intervals by cyclic neutropenic patients (173). A standardized assay for this type of activity was constructed with routine bone marrowdepleted of all mature B cells and most other lymphocyteswith monoclonal14.8 antibodies. The subsequent appearanceof large 14.8÷ cells and pre-B cells was then monitoredin short-term cultures of such cell suspensions.At least one substancewith similar activity wasdetectable in the supernatantsof a murinestromalcell clone and then obtained in recombinantform (44, and K. S. Landreth, personal communication).The factor has no homologywith previously publishedmediatorsand is designatedIL-? in Figure 1. A transformed stromal cell clone was used to isolate another unique factor, IL-7, whichdramaticallyeffected replication ofmurinecells in longterm bone marrowcultures (49, 50). Since such lymphocytes may abnormal in manyrespects, the effects of recombinant IL-7 were thoroughlystudied using normalbone marrowcells (G. Lee, A. G. Namen, S. Gillis, P. W.Kincade, submitted). A promptproliferation followed addition of IL-7 to wholebone marrowcultures, and a series of cell

Annual Reviews


B-CELL FORMATION

129

separation experimentsshowedthat the immediatelyrespondingcells were large B-cell progenitors. This property wasexploitedto construct a selective cloningprocedurefor pre-Bcells in agar. Alinear doseresponsecurve for colonies resulting fromtitrated numbers of sorted marrowcells suggests that IL-7 directly stimulateslarger pre-Bcells to produceclonesof as many as 2000cells withinfive days. In contrast, highly enrichedpreparationsof smallpre-Bcells died in IL7, and 14.8- bone marrowsuspensions did not respond until after a considerable delay. It is possible that early progenitors spontaneously acquire IL-7 receptors, or that small amountsof IL-7 eventually stimulate upregulation of receptors to a functional level. Experimentswere then done to monitor population changes in bone marrow cultures maintainedin IL-7. BP-1antigen expressionwasvery high on these cells, as is the case for most transformedpre-B lines (141). Significant numbers of cells were also found that expressedclass II and/or Thy-1antigens (G. Lee, A. E. Namen,S. Gillis, P. W.Kincade,submitted). IL-7 primarily a replication factor, and little or no progression towardsurface Ig expressionoccurredduring5 days of semisolidor several weeksof liquid culture. B-cell precursors normallyundergoonly a finite numberof divisions under the influence ofIL-7(G. Lee, A. E. Namen, S. Gillis, P. W.Kincade, submitted). AlthoughIL-7 dependentlymphocyteclones have been isolated from long-termbonemarrowcultures, they must represent extremely rare variants (49, and A. Namen,personal communication).The phenotypes of such clones are unusual; they are not responsive to normal inductive stimuli, and they can spontaneously becomefactor independent (G. Lee, K. Medina, unpublished observations). These might be pre-malignant cells that have lost some normal regulatory responses, and it will be extremelyinteresting to learn howhumancells grownin IL-7 relate to those involved in leukemias (140). Problems with production, or receptor recognition of IL-7, mightbe involved in suchdiseases. Potential Antagonists of B Lymphopoiesis Normalmouseserumblocks most responses of an inducible pre-B tumor cell line in culture (G. Lee, unpublishedobservations).However,TGF-fl’s providedthe first clear exampleof potential antagonists for B lymphopoiesis (174). Thesefactors selectively blockedkappaexpressionon maturing normalpre-Bcells, or certain inducedresponsesin a pre-Bcell line. ClassII expressionis upregulatedon pre-Bcells by IL-4, and this response is blockedby prostaglandinE, but not by TGF-fl’s(174-177). Surface expressioncan be inducedon pre-Bcells by several factors, but only some

Annual Reviews


130

KINCADE

ET AL

of these are completely blocked by TGF-~’s 074). For example, LPS induction, but not IFN-~ induction, was effectively inhibited at the ~c mRNA level. This correlates with findings that variant clones of a pre-B cell line are differentially responsive to these agents (168, 178). A kappa enhancer binding protein, NFg-b, is induced by LPS (179), but not IFN~ although transcription of g mRNA is caused by both types of stimuli (M. Briskin, R. Wah, submitted). This suggests that multiple transmembrane signaling pathways and second messengers can lead to expression of this critical gene and that each type of response can be independently controlled by particular antagonists. TGF-fl’s are knownto be made and released in an inactive precursor form (61, 62). However, it was surprising to find that immunohistochemically detectable quantities of TGF-//werepresent on the surface of manynormal pre-B cells (174, and G. Lee, unpublished observations). little as 10-~° Mconcentrations of the active factor should block maturation of these cells. The mechanismwherebylatent TGF-fl’s are activated in vivo is not known. Production of myeloid or lymphoid cells in long-term cultures was completely blocked by adding TGF-/~’s to the medium(S.-I. Hayashi, manuscript in preparation). While these factors can directly interact with hemopoieticprogenitors (180), the effect was mediated in part by effects the microenvironment(S.-I. Hayashi, manuscript in preparation). TFG-/~’s block adipogenesis in long-term cultures and are knownto regulate synthesis of collagen and other componentsof the extracellular matrix (76, 77, 109, 181). Determining thc site of action of a cytokine can thus be difficult whenit is addedto a complexculture system. Clonal analysis of bone marrowcells with IL-7 plus recombinantfactors revealed other potentially important antagonists (G. Lee, A. E. Namen, S. Gillis, P. W. Kincade, submitted). A subpopulation of IL-7 responsive cells were sensitive to TGF-~inhibition, but complete proliferation arrest was achieved with IL-1 or high concentrations of IL-2. Recombinant IL2 is already being employedin clinical trials with certain malignancies, and it is intriguing to speculate that such cytokines might have efficacy for treatment of B lineage leukemias. IL-7 induced proliferation was unaffected by interleukins, 3, 4, 5, and 6 as well as IFN-y, G-CSF,G/M-CSF,TNF, or NZBserum derived factors (G. Lee, A. E. Namen, S. Gillis, P. W. Kincade, submitted). Thus, replication of B lineage precursors is under discrete control of agonists and antagonists. Other molecules have been defined that may influence B lymphocyte formation. For example, there is evidence that IL-3 directly (182, 183) indirectly (184) influences replication of someearly precursors. However, this thoroughly studied multipoietin has not been detectable in long-term

Annual Reviews


B-CELLFORMATION 131 culture systems, and we have other reasons for thinking that it is not essential for normal lymphopoiesis (see below). The bone marrowis a site not only for primary blood cell formation, but also for secondary immune responses(185). In this context, it is interesting that a critical growthfactor for antibody forming cells, IL-6, is made. by stromal cells (72, 186, 187, and J. M. Gimble, manuscript in preparation). IL-6 and other stimuli increased IL-6 mRNA levels in stromal cells and fibroblasts (72, 89, and unpublished observations). Thus, stromal cells might potentially contribute to "peripheral" as well as "central" lymphoid tissue microenvironments. Genetically Determined Abnormalities and Polymorphisms Reveal Additional Complexity in Bone Marrow Long-termculture models and their utility have been emphasized in this review. If these approaches provided a perfect representation of normal bone marrow, one would always expect to see a direct correlation between in vivo and in vitro studies. For example, if a genetic abnormality prevents lymphocyte formation in culture, it should have the same influence on B-cell numbers in the affected mice. Reciprocally, bone marrow from immunodeficient mice which have few lymphocytes should not be able to form them in culture. Wehave found exceptions to both of these types of predictions. Also, genetic polymorphismsthat have little or no effect on blood cell formation in intact animals can profoundly influence hemopoiesis in vitro. These findings suggest that intact bone marrowmayhave "fail safe" regulatory mechanismsnot always duplicated in vitro. However, the relative simplicity of our experimental approaches makes it possible to appreciate the delicate balance and redundancypresent in this organ. Mice with severe combined immunodeficiency disease (SCID) have normal numbersof stem cells, myeloid cells, and NKcells, although B and T lymphocytes are virtually absent (188-191). The molecular basis of this mutation is thought to involve improper recognition of appropriate sequences during rearrangement of Ig and T cell receptor genes (192). Althoughthe rearrangements occur, this only rarely results in functional lymphocytes. However, long-term lymphocyte cultures can be readily established with bone marrow from SCID mice and subsequently maintained for at least eight months (29, 30). Therefore, some component intact marrowmay be able to sense and eliminate intrinsically defective lymphocytes. Our long-term cultures must lack positive and negative selection mechanismswhich operate to achieve this type of "quality control." Lymphoid and myeloid cells are made in severely immunodeficient/autoimmune mice with motheaten mutation (193). However, addition of bone marrowcells from these animals to long-term cultures

Annual Reviews


132

KINCADE ET AL

established with normal littermate cells prevented production of nonadherent cells under three different types of culture conditions (139, 194, 195). The mechanismof this suppression is not clear, but it seems to involve abnormalexpansion of macrophage-like cells in primary cultures. The imbalance was not reflected in the release of soluble inhibitory substances; motheaten cells had to be closely associated with normal cells to prevent lympho-hemopoiesis. Thus, an antagonist that is cell- or matrixbound could be abnormally active in long-term cultures of motheaten marrow. An understanding of this phenomenoncould be instructive with respect to a regulatory mechanism which is normally coordinated and potentially important in vivo. The X chromosome-linked immunodeficiency of CBA/Nmice does not influence the rate of pre-B cell formation in bone marrow (195, 196). However, this process is unusually dependent on T lymphocytes (197199). Athymic nude mice congenic for the Xid mutation have few lymphocytes of any kind, and the same is true of thymectomized recipients of (Xid) bone marrow. Cells with obvious T lineage markers are not demonstrable in our long-term cultures (23). Thus, one might predict that cultures established from Xid marrow would be poor lymphocyte producers. However, the mutation actually caused accelerated lymphopoiesis, and myelopoiesis in vitro (S.-I. Hayashi, P. L. Witte, P. W.Kincade, submitted). Xid positively affected an early step in culture formation, presumablyinvolving the establishment of an adherent layer. IL-3 has been shownto influence cells of manylineages (56), including early B lymphocyte precursors (182, 183, 200, 201). However, we found enormousvariation in responsiveness of bone marrowcells from d’,~,erent inbred strains to this lymphokine (59). Cells from mice that taave obvious difficulty with blood cell formation were virtually unresponsiveto IL-3 in several culture assays. Thus, this multipoietin is probably not a necessary part of the regulatory network. Otherwise, polymorphisms affecting IL-3 responses would lead to markeddifferences in blood cell numbers. One nonlymphoid cell line (WEHI-3) makes large amounts IL-3. However,this resulted from insertion of a transforming retrovirus into the IL-3 gene, and T helper cells are probably the normal source for this mediator (202). Sensitive bioassays and mRNA analyses have not revealed IL-3 production by stromal cells, or any other component of long-term cultures (58). However, strom~ cells can support the growth and differentiation of stem cells that have been maintained for a time in IL-3 containing medium(102, 203,204). All of these findings indicate that one or moreother factors duplicate the functions of IL-3 in vivo. They also suggest that responses to treatment with certain recombinant cytokines can vary substantially amongindividuals.

Annual Reviews


B-CELLFORMATION 133 Other types of polymorphismsmarkedly influence the behavior of cells in long-term bone marrow cultures. For example, BALB/cbone marrow cells are superior to CBA/H derived cells in terms of the ease of culture establishment and continued lymphocyte production (S.-I. Hayashi, P. L. Witte, P. W.Kincade, submitted). Virtually all of the nonadherentcells in cultures initiated with equal part mixtures of the two marrowtypes were BALB/c,suggesting that their hemopoietic cells have an intrinsic proliferative advantage. Wedo not knowthe molecular basis for these differences, and similar influences on lymphopoiesis are not obvious in vivo. Again, this suggests that long-term culture modelslack the full complexity of intact bone marrow. However, there are circumstances when results of short- and longterm cultures correlate with those predicted from in vivo behavior. As mentionedabove, factors active on early B lineage precursors appeared in the urine of cyclic neutropenia patients just whennumbersof bone marrow pre-B cells were markedly elevated (92, 173). This made it practical induce formation of humanpre-B cells in culture, and the experimental approach led to the molecular cloning of a unique cytokine (44, and K. S. Landreth, personal communication). Likewise, lymphopoiesis is premature and exaggerated in embryonic and young NZBstrain mice. Cytokines that maycontribute to this dysregulated situation were purified from the serum of these animals, and evidence obtained that older NZBmice make autoantibodies to them (171, 172). As the animals mature, pre-B cells virtually disappear from the bone marrowin what could be a related phenomenon(205). Abnormalhemopoiesis of this strain of mice was also reflected in long-term bone marrowcultures (206). Concludin9 Remarks Our understanding of B lymphocyte progenitors should increase with concerted application of improvedcell sorting protocols, monoclonalantibodies, purified cytokines, and gene probes. The latter are becoming increasingly available through subtractive hybridization and screening techniques (207) and will considerably facilitate our understanding differentiation events. Similar experimental approaches should be informative about stromal cells and other microenvironmental elements, about which several questions seem crucial. Weneed to know how many cytokines they are capable of making, what physical form they have in vivo, and which ones are relevant to the B lymphocyte lineage. Positive and negative regulatory interactions betweencells in bone marrowcultures are amenable to study, and this information should be extended by injection of recombinant materials in vivo. Further technical advances might lead to better duplication of the bone marrowin culture, and this would have

Annual Reviews

134

KINCADE ET AL

major implications for diagnosis as well as basic research. Bonemarrow transplantation has becomea therapeutic option for manydisorders. However, the molecular and cellular basis for diseases such as myelofibrosis, anemias, autoimmunediseases, and malignancies differ. It may someday be possible to engraft only the relevant components(96) or treat with the appropriate cytokines.


ACKNOWLEDGMENTS

Our research is supported by grants AI 20069 and AI 19884 from the National Institutes of Health and a Special Fellowship (to Carolyn G. Pietrangeli) from the LeukemiaSociety of America. Weare grateful for expert technical assistance provided by Ms. Annette Dorheim, Anna Henley, Kay Medina, and Margaret Robinson.

Literature Cited I. Weiss, L., Chert, L. T. 1975. The organization ofhematopoietic cords and vascular sinuses in bone marrow. Blood Cells 1:617-38 2. Weiss, L., Sakai, H. 1984. The hematopoietic stroma. Am. J. Anat. 170: 44763 3. Lichtman, M. A. 1981. The ultrastructure of the hemopoietic environment of the marrow: A review. Exp. Hematol. 9:391~410 4. Deldar, A., Lewis, H., Weiss, L. 1985. Bone lining cells and hematopoiesis: An electron microscopic study of canine bone marrow. Anat. Rec. 213: 187 20l 5. Weiss, L. 1976. Hematopoietic microenvironment of the bone marrow: An ultrastructural study of the stroma in rats. Anat. Ree. 186:161-84 6. Shackney,S. E., Ford, S. S., Wittig, A. B. 1975. Kinetic-microarchitectural correlations in the bone marrowof the mouse. Cell Tissue Kinet. 8:505-16 7. Lord, B. I., Testa, N. G., Hendry,J. H. 1975. The relative spatial distributions of CFU-s and CFU-c in the normal mouse femur. Blood 46:65 8. Tavassoli, M., Shaklai, M. 1979. Absence of tight junctions in endothelium of marrow sinuses: Possible significance for marrowcell egress. Br. J. Haematol. 41:303-7 9. Bessis, M. C., Breton-Gorius, J. 1962. Iron metabolism in the bone marrowas seen by electron microscopy: a critical review. Blood 19:635 63 10. Westen, H., Bainton, D. F. 1979.

Association of alkaline-phosphatasepositive reticulum cells in bone marrow with granulocytic precursors. J. Exp. Med. 150:919-37 11. Batten, S. J., Osmond, D. G. 1984. The localization of B lymphoeytes in mouse bone marrow: Radioautographic studies after in vivo perfusion of radiolabelled anti-IgM antibody. J. Immunol. Methods 72:381-99 12. Hermans, M. J. A., Hartsuiker, H., Opstelten, D. 1988. Anin situ study of B-lymphocytopoiesis in rat bone marrow: topographical arrangement of terminal deoxynueleotidyl transferase positive cells and pre-B cells. J. lmmunol. In press 13. Osmond, D. G. 1986. Population dynamics of bone marrow B lymphocytes. Immunol. Rev. 93:10354 14. Kincade, P. W., Lee, G., Paige, C. J., Scheid, M. P. 1981. Cellular interactions affecting the maturation of murine B lymphocyte precursors in vitro. J. Immunol. 127:255-60 15. Kincade, P. W. 1987. Experimental models for understanding B lymphocyte formation. Adv. Immunol. 41: 181267 16. Dexter, T. M., Lajtha, L. G. 1974. Proliferation of hematopoietic stem cells in vitro. Br. J. Haematol.28: 5253O 17. Dexter, T. M., Allen, T. D., Lajtha, L. G. 1977. Conditions controlling the proliferation of haemopoictic stem cells in vitro. J. Cell Physiol. 91: 335-

Annual Reviews


B-CELL FORMATION 18. Whitlock, C. A., Witte, O. N. 1982. Long term culture of B lymphocytes and their precursors from murine bone marrow. Proc. Natl. Acad. Sci. USA 79:3608-12 19. Whitlock, C. A., Robertson, D., Witte, O. N. 1984. Murine B cell lymphopoiesis in long-term culture. J. lmmunol. Methods 67:353-69 20. Dorshkind, K. 1986. In vitro differentiation of B lymphocytesfrom primitive hemopoietic precursors present in long-term bone marrow cultures. J. lmmunol. 136:422-29 21. Dorshkind, K., Johnson, A., Collins, L., Keller, G. M., Phillips, R. A. 1986. Generation of purified stromal cell cultures that support lymphoid and myeloid precursors. J. Immunol. Method~ 89:37-47 22. Whitlock, C., Denis, K., Robertson, D., Witte, O. 1985. In vitro analysis of murine B-cell development. Ann. Rev. lmmunol. 3:213-35 23. Witte, P. L., Kincade, P. W., Vetvicka, V. 1986. Interculture variation and evolution of B lineage lymphocytes in long-term marrow culture. Eur. J. lmmunol. 16:779-88 24. Witte, P. L., Robinson, M., Henley, A., Low, M. G., Stiers, D. L., Perkins, S., Fleischman, R. A., Kincade, P. W. 1987. Relationships between B-lineage lymphocytes and stromal cells in long term bone marrow cultures. Eur. J. Immunol. 17:1473-84 25. Kurland, J. I., Ziegler, S. F., Witte, O. N. 1984. Long-term cultured B lymphocytes and their precursors reconstitute the B-lymphocytelineage in vivo. Proc. Natl. Acad. Sci. USA81: 755458 26. Nagasawa, R., Kanagawa, O., Tittle, T. V., Chiller, J. M. 1985.In vivo maturation of pre-B cells derived from longterm cultured bone marrow. J. lmmunol. 135:965-72 27. Dorshkind, K., Denis, K. A., Witte, O. N. 1986. Lymphoid bone marrow cultures can reconstitute heterogeneous B and T cell-dependent responses in severe combined immunodeficient mice. J. Immunol. 137:3457-63 28. Denis, K. A., Dorshkind, K., Witte, O. N. 1987. Regulated progression of B lymphocyte differentiation from cultured fetal liver. J. Exp. Med. 166: 391-403 29. Witte, P. L., Burrows, P. D., Kincade, P. W., Cooper, M. D. 1987. Characterization of B lymphocyte lineage progenitor cells from mice with severe combined immune deficiency disease

135

(SCID) madepossible by long term culture. J. Immunol. 138:2698-2705 30. Hirayoshi, K., Nishikawa, S., Kina, T., Hatanaka, M., Habu, S., Nomura, T., Katsura, Y. 1987. lmmunoglobulin heavy chain gene diversification in the long-term bone marrowculture of normal mice and mice with severe combined immunodeficiency. Eur. J. Immunol. 17:1051-57 31. Yoshida, N., Radbruch, A., Rajewsky, K. 1988. Ig gene rearrangement and expression in the progenyof B-cell progenitors in the course of clonal expansion in bone marrow cultures. EMBO J. 6:2735-41 32. Denis, K. A., Witte, O. N. 1986. In vitro development of B lymphocytes from long-term cultured precursor cells. Proc. Natl. Acad. Sci. USA 83: 44145 33. Pietrangeli, C. E., Hayashi, S.-I., Kincade, P. W. 1988. Stromal cell lines which support lymphocyte growth: Characterization, sensitivity to radiation and responsiveness to growth factors. Eur. J. Immunol. 18:863~2 34. Allen, T. D., Dexter, T. M. 1984. The essential cells of the hemopoieticmicroenvironment. Exp. Hematol. 12: 51721 35. Dorshkind, K., Schouest, L., Fletcher, W. H. 1985. Morphologic analysis of long-term bone marrow cultures that support B-lymphopoiesis or myelopoiesis. Cell Tissue Res. 239:375-82 36. Zipori, D., Duksin, D., Tamir, M., Argaman, A., Toledo, J., Malik, Z. 1985. Cultured mouse marrow stromal cell lines. II. Distinct subtypesdiffering in morphology, collagen types, myelopoietic factors, and leukemic cell growth modulating activities. J. Cell. Physiol. 122:81-90 37. Whitlock, C. A., Tidmarsh, G. F., Muller-Sieburg, C., Weissman, I. L. 1987. Bone marrowstromal cell lines with lymphopoietic activity express high levels of a pre-B neoplasia-associated molecule. Cell48:1009-21 38. Li, C. L., Johnson, G. R. 1985. Stimulation of multipotcntial, erythroid and other murine haematopoietic progenitor cells by adherentcell lines in the absence of detectable multi-CSF (IL3). Nature 316:633-36 39. Li, C. L., Culter, R. L., Johnson, G. R. 1987. Characterization of hemopoietic activities in media conditioned by a murine marrow-derived adherent cell line, B.Ad. Exp. Hematol. 15:373-81 40. Leung, L. C., Johnson, G. R. 1987. In vitro maintenance of hemopoietic stem

Annual Reviews

136

41.


42.

43.

44.

45.

46.

47.

48.

49.

50.

KINCADE ET AL

cells with lymphoidand myeloid rep. opulating ability by a cloned murme adherent bone marrowcell line. Exp. Hemtaol. 15:989-94 Song, Z. X., Shadduck, R. K., hines, D. J., Waheed,A., Quesenberry, P. J. 1985. Hematopoietic factor production by a cell line (TC-1) derived l¥om adherent murine marrow cells. Blood 66:273-81 Quesenberry, P., Song, Z. X., McGrath, E., McNiece, I., Shadduck, R., Waheed, A., Baber, G., Kleeman, E., Kaiser, D. 1987. Multilineage synergistic activity produced by a murine adherent marrow cell line. Blood 69: 827-35 Collins, L. S., Dorshkind, K. 1987. A stromal cell line from myeloid longterm bone marrowcultures can support myelopoiesis and B lymphopoiesis. J. Immunol. 138:1082-87 Landreth, K. S., Dorshkind, K. 1988. Pre-B cell generation potentiated by soluble factors from a bone marrowstroreal cell line. J. Immunol.140:845-52 King, A. G., Landreth, K. S., Wierda, D. 1988. Bone marrow stromal cell regulation of B-lymphopoiesis: l. The role of macrophages,interleukin-1, and interleukin-4 in pre-B cell maturation. J. ImmunoLIn press Hunt, P., Robertson, D., Weiss, D., Rennick, D., Lee, F., Witte, O. N. 1987. A single bone marrow-derived stromal cell type supports the in ~itro growth of early lymphoid and myeloid cells. Cell 48:997-1007 Rennick, D., Yang, G., Muller-Sieburg, C., Smith, C., Arai, N., Takabe, Y., Gemmell, L. 1987. Interleukin 4 (B-cell stimulatory factor 1) can enhance or antagonize the factor-dependent growth of hemopoietic progenitor cells. Proc. Nail Acad. Sci. USA 84:6889-93 Lowenthal, J. W,, Castle, B. E., Christiansen, J., Schreurs, J., Rennick, D., Arai, N., Hoy, P., Takebe, Y., Howard, M. 1988. Expression of high affinity receptors for murineinterleukin 4 (BSF-I) on hemopoietic and nonhemopoietic cells. J. Immunol. 140: 456-64 Namen,A. E., Schmierer, A. E., March, C. J., Overell, R. W., Park, L. S., Urdal, D. L., Mochizuki, D. Y. 1988. B cell precursor growth-promoting activity. Purification and characterization of a growth factor active on lymphocyte precursors. J. Exp. Med. 167:988-1002 Namen, A. E., Lupton, S., Hjerrild, K., Wagnall, J., Mochizuki, D. Y.,

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

Schmierer, A., Mosley, B., March, C. J,, Urdal, D., Gillis, S., Cosman,D., Goodwin, R. G. 1988. Stimulation of B-cell progenitors by cloned murine interleukin-7. Nature 333:571-73 Ogawa, M., Nishikawa, S., Ikuta, K., Yamamura, F., Naito, M., Takahashi, K., Nishikawa, S.-I. 1988. B cell ontogeny in murine embryo studied by a culture system with the monolayer of a stromal cell clone, ST2: B cell progenitor develops first in embryonal body rather than in yolk sac. EMBO J. 7:1337~13 Whitlock, C. A., Ziegler, S. F., Treiman, L. J., Stafford, J. I., Witte, O. N. 1983. Differentiation of cloned populations of immature B ceils after transformation with Abelson murine leukemia virus. Cell 32:903-11 Muller-Sieburg, C. E, Whitlock, C. A., Weissman,I. L. 1986. Isolation of two early B lymphocyte progenitors from mouse marrow: A committed pre-preB cell ~nd a clonogenic Thy-1~° hematopoietic stem cell. Cell 44:653-62 Gordon, M. Y,, Riley, G. P., Watt, S. M., Greaves, M. F. 1987. Compartmentalization of a haematopoietic growth factor (GM-CSF) by glycoaminoglycans in the bone marrow microenvironment. Nature 326:403-5 Roberts, R., Gallagher, J., Spooncer, E., Allen, T. D., Bloomfield, F., Dexter, T. M. 1988. Heparan sulphate bound growth factors: A mechanism for stromal cell mediated haemopoiesis. Nature 332:376-78 Schrader, J. W. 1986. The panspecific hemopoietin of activated T lymphocytes (Interleukin-3). Ann. Rev. Immunol. 4:205-30 Ikebuchi, K., Wong, G. G., Clark, S. C., Ihle, J. N., Hirai, Y., Ogawa, M. 1987. Interleukin 6 enhancement of interleukin 3-dependent proliferation of multipotential hemopoietic progenitors. Proc. Natl. Acad. Sci. USA 84:9035-39 Zipori, D., Lee, F. 1988. Introduction of interleukin-3 gene into stromal cells from the bone marrow alters hemopoietic differentiation but does not modify stem cell renewal. Blood 71: 586-96 Lee, G. 1988. Characterization of routine B lymphocyte precursors and factors which regulate them. PhDthesis, Univ. Okla., OklahomaCity, Okla. Van Snick, J., Cayphas, S., Szikora, J.-P., Renauld, J.-C., Van Roost, E., Boon, T., Simpson, R. J. 1988. cDNA cloning of murine interleukin-HPl:

Annual Reviews


B-CELL FORMATION Homology with human interleukin 6. Eur. J. Immunol. 18:193-97 61. Sporn, M. B., Roberts, A. B., Wakefield, L. M., de Crombrugghe,B. 1987. Somerecent advances in the chemistry and biology of transforming growth factor-beta. J. Cell Biol. 105:1039-45 62. Massague,.J. 1987. The TGF-ft family of growth and differentiation factors. Cell 49:437-38 63. Gurney, M. E., Apatoff, B. R., Spear, G. T., Baumel, M. J., Antel, J. P., Bania, M. B., Reder, A. T. 1986. Neuroleukin: A lymphokine product of lectin-stimulated T cells. Science 234: 574-81 64. Chaput, M., Claes, V., Portetelle, D., Cludts, I., Cravador, A., Burny, A., Gras, H., Tartar, A. 1988. The neurotrophic factor neuroleukin is 90% homologous with phosphohexose isomerase. Nature 332:454-55 65. Metcalf, D. 1971. Acute antigeninduced elevation of serum colony stimulating factor (CSF) levels. Immunology 21: 427-36 66. Thorens, B., Mermod,J.-J., Vassalli, P. 1987. Phagocytosis and inflammatory stimuli induce GM-CSF mRNAin macrophages through posttranscriptional regulation. Cell 48:671-79 67. Seelentag, W. K., Mermod,J.-J., Montesano, R., Vassalli, P. 1987. Additive effects of interleukin 1 and tumour necrosis factor-or on the accumulation of the three granulocyte and macrophage colony-stimulating factor mRNAsin human endothelial cells. EMBOJ. 6:2261-65 68. Nawroth, P. P., Bank, I., Handley, D., Cassimeris, J., Chess, L., Stern, D. 1986. Tumornecrosis factor/cachectin interacts with endothelial cell receptors to induce release of interleukin 1. J. Exp. Med. 163:1363-75 69. Broudy, V. C., Kaushansky, K., Segal, G. M., Harlan, J. M., Adamson,J. W. 1986. Tumor necrosis factor type ~ stimulates humanendothelial cells to produce granulocyte/macrophage colony-stimulating factor. Proc. Natl. Acad. Sci. USA 83:7467-71 70. Segal, G. M., McCall, E., Stueve, T., Bagby, G. C. Jr. 1987. Interleukin 1 stimulates endothelial cells to release multilineage humancolony-stimulating activity. J. ImmunoL138:1772-78 71. Rennick, D., Yang, G., Gemmell, L., Lee, F. 1987. Control of hemopoiesis by a bone marrowstromal cell clone: Lipopolysaccharide- and interleukin- 1inducible production of colonystimulating factors. Blood 69:682-91

137

72. Yang, Y.-C., Tsai, S., Wong, G. G., Clark, S. C. 1988. Interleukin-1 regulation of hematopoietic growth factor production by human stromal fibroblasts. J. Cell. Physiol. 134:292-96 73. Broudy, V. C., Zuckerman, K. S., Jetmalani, S., Fitchen, J. H., Bagby,G. C. Jr. 1986. Monocytes stimulate fibroblastoid bone marrowstromal cells to produce multilineage hematopoietic growth factors. Blood 68:530-34 74. Wall, R., Briskin, M., Carter, C., Govan, H., Taylor, A., Kincade, P. W. 1986. Labile inhibitor blocks gene transcription in a pre-B leukemia cell line. Proc. Natl. Acad. Sci. USA 83: 29598 75. Shaw, G., Kamen, R. 1986. A conserved AU sequence from the 3’ untranslated region of GM-CSF mRNAmediates selective mRNAdegradation. Cell 46:659-67 76. Seyedin, S. M., Segarini, P. R., Rosen, D. M., Thompson, A. Y., Bentz, H., Grayzar, J. 1987. Cartilage-inducing factor-fl is a uniqueprotein structurally and functionally related to transforminggrowth factor-ft. J. Biol. Chem. 262:1946-49 77. Bassols, A., Massague, J. 1988. Transforming growth factor ft regulates the expression and structure of extracellular matrix chondroitin/dermatan sulfate proteoglycans. J. Biol. Chem. 263:3039-45 78. Rosenfeld, M., Keating, A., BowenPope, D. F., Singer, J. W., Ross, R. 1985. Responsiveness of the in vitro hematopoietic microenvironment to platelet-derived growth factor. Leuk. Res. 9:427-34 79. Miyajima, A., Miyatake, S., Schreurs, J., De Vries, J., Arai, N., Yokota, T., Arai, K.-I. 1988. Coordinate regulation of immuneand inflammatory responses by T cell-derived lymphokines. FASEB 2:2462-73 80. Peschel, C., Paul, W. E., Ohara, J., Green, I. 1987. Effects of B cell stimulatory factor- 1/interleukin 4 on hematopoietic progenitor cells. Blood 70: 254-63 81. Clark, S. C., Kamen, R. 1987. The humanhematopoietic colony-stimulating factors. Science 236:1229-37 82. Zalman, F., Maloney, M. A., Patt, H. M. 1979. Differential response of early erythropoietic and granulopoietic progenitors to dexamethasone and cortisone. J. Exp. Med. 149:67-72 83. Hirata, J., Kaneko, S., Nishimura, J., Motomura,S., Ibayashi, H. 1988. Hydrocortisone modulates colony-stim-

Annual Reviews


138

KINCADE ET AL

ulating activity produced by human bone marrow-derived adherent cells. Eur. J. Haematol. 40:215-21 84. Ku, G., Witte, O. N. 1986. Corticosteroid-resistant bone marrowderived B lymphocyte progenitor for long term in vitro cultures. J. lmmunol. 137:2802-7 85. Quesenberry, P. J., Coppola, M. A., Gualtieri, R. J., Wade,P. M., Song, Z., Doukas, M. A., Shideler, C. E., Baker, D. G., McGrath, E. H. 1984. Lithium stimulation of murine hematopoiesis in liquid culture: Aneffect mediated by marrowstromal cells. Blood 63: 12127 86. Trentin, J. J. 1971. Determination of bone marrowstem cell differentiation by stromal hemopoietic inductive microenvironments (HIM). Am. J. Pathol. 65:621-28 87. Stanley, E. R., Guilbert, L. J., Tushinski, R. J., Bartelmez, S. H. 1983. CSF1A mononuclear phagocyte lineagespecific hemopoietic growth factor. J. Cell. Biochem. 21:151-59 88. Van Obberghen-Schilling, E., Roche, N. S., Flanders, K. C., Sporn, M. B., Roberts, A. B. 1988. Transforming growthfactor fl I positively regulates its own expression in normal and transformed cells. J. BioL Chem. 263: 774146 89. Kohase, M., May, L. T., Tamm, I., Vilcek, J., Sehgal, P. B. 1987. A cytokine network in human diploid fibroblasts: Interactions of /~-interferons, tumor necrosis factor, platelet-derived growth factor, and interleukin-l. Mol. Cell. Biol. 7: 27380 90. Brooks, K. H., Uhr, J. W., Vitetta, E. S. 1984. A B cell growth factorlike activity is secreted by cloned, neoplastic B cells. J. lmmunol. 133: 313337 91. Ambrus,J. L. Jr., Fauci, A. S. 1985. Human B lymphoma cell line producing B cell growth factor. J. Clin. Invest. 75:732-39 D., Landreth, K. S., 92. Engelhard, Kapoor, N., Kincade, P. W., DeBault, L. E., Theodore, A., Good, R. A. 1983. Cycling of peripheral blood and marrow lymphocytesin cyclic neutropenia. Proc. Natl. Acad. Sci. USA 80: 573438 93. Fulop, G., Lee, M. Y., Rosse, C. 1985. A granulocytosis-inducing tumor inhibits the production of B lymphocytes in murine bone marrow. J. lmmunoL 135:4266-72 94. Dorshkind, K. 1988. IL-1 inhibits B

cell differentiation in long term bone marrowcultures. J. Immunol. 141: 53138 95. Dexter, T. M. 1982. Is the marrow stroma transplantable? Nature 298: 222-23 96. Anklesaria, P., Kase, K., Glowacki,J., Holland, C. A., Sakakeeny, M. A., Wright, J, A., FitzGerald, T. J., Lee, C.-Y., Greenberger,J. S. 1987. Engraftment of a clonal bone marrowstromal cell line in vivo stimulates hematopoietic recovery from total body irradiation. Proc. Natl. Acad. Sci. USA 84:7681-85 97. Tavassoli, M., Crosby, W. H. 1970. Bone marrowhistogenesis: A comparison of fatty and red marrow. Science 169:291-93 98. Perkins, S., Fleischman, R. A. 1988. Hematopoietic microenvironment. Origin, lineage, and transplantability of the stromal cells in long-term bone marrowcultures from chimeric mice. J. Clin. lnvest. 81:1072-80 99. Charbord, P., Gown, A. M., Keating, A., Singer, J. W. 1985. CGA-7 and HHF,two monoclonal antibodies that recognize muscle actin and react with adherent cells in humanlong-term bone marrowcultures. Blood 66:1138-42 100. Herman, I. M., D’Amore, P. A. 1985. Microvascular pericytes contain muscle and nonmuscleactins. J. Cell Biol. 101: 43 52 101. Mizutani, S., Watt, S. M., Robertson, D., Hussein, S., Healy, L. E., Furley, A. J., Greaves, M. F. 1987. Cloning of human thymic subcapsular cortex epithelial cells with T-lymphocytebinding sites and hemopoietic growth factor activity. Proc. Natl. Acad. Sci. USA84: 4999-5003 102. Roberts, R. A., Spooncer, E., Parkinson, E. K., Lord, B. 1., Allen, T. D., Dexter, T. M. 1987. Metabolically inactive 3T3cells can substitute for marrow stromal cells to promote the proliferation and development of multipotent haemopoietic stem cells. J. Cell. Physiol. 132:203-14 103. Davis, R. L., Weintraub, H., Lassar, A. B. 1987. Expressionof a single transfected cDNAconverts fibroblasts to myoblasts. Cell 51:987-1000 104. Lassar, A. B., Paterson, B. M., Weintraub, H. 1986. Transfection ofa DNA locus that mediates the conversion of 10T1/2 fibroblasts to myoblasts. Cell 47:649-56 105. Konieczny, S. F., Emerson, C. P. Jr. 1984. 5-azacytidine induction of stable mesodermal stem cell lineages from

Annual Reviews


B-CELL FORMATION

139

tural and functional determinants. II. 10T1/2 cells: Evidence for regulatory Placement of monoclonal antibody genes controlling determination. Cell epitopes. J. Cell Biol. 103:1729-37 38:791-800 106. Lanotte, M., Metcalf, D., Dexter, T. M. 117. Thomas,P. S., Pietrangeli, C. E., Hay1982. Production of monocyte/macashi, S.-I., Schachner, M., Goridis, C., Low, M. G., Kincade, P. W. 1988. rophage colony-stimulating factor Demonstration of neural cell adhesion by preadipocyte cell lines derived from murine marrow stroma. J. Cell. Phymolecules on stromal cells which support lymphopoiesis. Leukemia 2: 171siol. 112:123-27 107. Kodama, H., Sudo, H., Koyama, 75 118. Edelman, G. M. 1986. Cell adhesion H., Kasai, S., Yamamoto,S. 1984. In vitro hemopoiesis within a micromolecules in neural histogenesis. Ann. environment created by MC3T3-G2/ Rev. Physiol. 48:417-30 119. Low, M. G. 1987. Biochemistry of the PA6preadipocytes. J. Cell. Physiol. glycosyl-phosphatidylinositol mem118:233-40 108. Greenberger, J. S. 1979. Corticosbrane protein anchors. Biochem.J. 244: teroid-dependent differentiation of 1 13 humanmarrowpreadipocytes in vitro. 120. He, H.-T., Barbet, J., Chaix, J.-C., Goridis, C. 1986. Phosphatidylinositol In Vitro 15:823-28 109. Ignotz, R. A., Massague,J. 1985. Type is involved in the membraneattach/~ transforming growth factor controls ment of NCAM120, the smallest comthe adipogenic differentiation of 3T3 ponent of the neural cell adhesion fibroblasts. Proc. Natl. Acad. Sci. USA molecule. EMBOJ. 5:2489-94 82:8530-34 121. Hemperly, J. J., Edelman, G. M., Cunningham, B. A. 1986. cDNAclones of 110. Greenberger, J. S. 1978. Sensitivity of cor ticosteroid-dependent insulin-resisthe neural cell adhesion molecule (Ntant lipogenesis in marrow preadipCAM)lacking a membrane-spanning ocytes of obese-diabetic (db/db) mice. region consistent with evidence for membraneattachment via a phosphatiNature 275:752-54 dylinositol intermediate. Proc. Natl. 111. Tavassoli, M. 1984. Marrowadipose Acad. Sci. USA 83:9822-26 cells and hemopoiesis: Aninterpretative review. Exp. Hematol. 12:139-46 122. Low, M. G., Prasad, A. R. S. 1988. A 112. Nishi, Y., Yoshikawa, K., Hiai, H., phospholipase D specific for the phosphatidylinositol anchor of cell surface Notake, K., Shisa, H., Yasuaki, N. 1982. Formation of symbiotic comproteins is abundant in plasma. Proc. Natl. Acad. Sci. USA 85:980-84 plex by microenvironment-dependent mouseleukemias and thymic epithelial 123. Low, M. G., Saltiel, A. R. 1988. Strucreticular cells. JNCI69:627-33 tural and functional roles of glycosylphosphatidylinositol in membranes. 113. Owens, G. C., Edelman, G. M., Cunningham, B. A. 1987. Organization of Science 239:268-75 the neural cell adhesion molecule (N124. Low, M. G., Kincade, P. W. 1985. CAM)gene: Alternative exon usage as Phosphatidylinositol is the membranethe basis for different membraneanchoring domain of the Thy-1 glycoassociated donmins. Proc. Natl. Acad. protein. Nature 318:62q54 Sci. USA 84:294-98 125. Ishihara, M., Fedarko, N. S., Conrad, 114. Cunningham, B. A., Hemperly, J. J., H. E. 1987. Involvement of phosphatiMurray, B. A., Prediger, E. A., Brackdylinositol and insulin in the coenbury, R., Edelman, G. M. 1987. ordinate regulation of proteoheparan Neural cell adhesion molecule: strucsulfate metabolism and hepatocyte ture, immunoglobulin-like domains, growth. J. Biol. Chem. 262:4708-16 cell surface modulation, and alter126. Hunkapiller, T., Hood, L. 1986. The native RNAsplicing. Science 236: 799growing immunoglobulin gene super806 family. Nature 323:1%16 115. Dickson, G., Gower, H. J., Barton, 127. Williams, A. F. 1987. A year in the life of the immunoglobulin superfamily. C. H., Prentice, H. M., Elsom, V. L., Moore, S. E., Cox, R. D., Quinn, C., Immunol. Today 8:298-303 Putt, W., Walsh, F. S. 1987. Human 128. Takeichi, M., Hatta, K., Nagafuchi, A. muscle neural cell adhesion molecule 1985. Selective cell adhesion mechanisms: Role of the calcium-dependent (N-CAM):Identification of a musclespecific sequence in the extracellular cell adhesion system. In Molecular domain. Cell 50:1119-30 Determinants of Animal Form, ed. 116. Frelinger, A. L. III, Rutishauser, U. G. M. Edelman, pp. 223-33. New 1986. Topography of N-CAMstrucYork: Alan R. Liss, Inc.

Annual Reviews


140

KINCADE ET AL

129. Hatta, K., Takagi, S., Fujisawa, H., Takeichi, M. 1987. Spatial and temporal expression pattern of N-cadherin cell adhesion molecules correlated with morphogenetic processes of ~hieken embryos. Dev. Biol. 120:215-27 130. Nagafuchi, A., Shirayoshi, Y,, Okazaki, K., Yasuda, K., Takeidtii, M. 1987. Transformation of cell adhesion properties by exogenously introduced E~cadhcrin cDNA.Nature 329:341-43 131. Bernardi, P., Patel, V. P., Lodish, H. F. 1987. Lymphoidprecursor cells adhere to two different sites on fibronectin. J. Cell Biol. 105:489-98 132. Campbell, A. D., Long, M. W., Wicha, M. S. 1987. Haemonectin, a bone marrow adhesion protein specific for cells of granulocyte lineage. Nature 329: 74446 133. Lobb, R. R., Harper, J. W., Fett, J. W. 1986. Purification of heparin-binding growth factors. Anal. Bioehem. 154: 114 134. Kurt-Jones, E. A., Beller, D. I., Mizel, S. B., Unanue, E. R. 1985. Identification of a membrane-associated interleukin 1 in macrophages. Proe. Natl. Acad. Sci. USA 82:1204-8 135. March, C. J., Mosley, B., Larsen, A., Cerretti, D. P., Braedt, G., Price, V., Gillis, S., Henney, C. S., Kronheim, S. R., Grabstein, K., Conlon, P. J., Hopp, T. P., Cosman, D. 1985. Cloning, sequence and expression of two distinct humaninterleukin-1 complementary DNAs. Nature 315:641 ~47 136. Gough, N. M., Metcalf, D., Gough, J., Grail, D., Dunn,A. R. 1985. Structure and expression of the mRNAfor murine granulocyte-macrophage colony stimulating factor. EMBO J. 4:645-52 137. Rettenmier, C. W., Roussel, M. F., Ashmun,R. A., Ralph, P., Price, K., Sherr, C. J. 1987. Synthesis of membrane-bound colony-stimulating factor 1 (CSF-1) and downmodulation CSF-1 receptors in NIH3T3cells transformed by cotransfection of the human CSF-1 and c-fms (CSF-I receptor) genes. Mol. Cell. Biol. 7:2378-87 138. Kierney, P. C., Dorshkind, K. 1987. B lymphocyte precursors and myeloid progenitors survive in diffusion chamber cultures but B cell differentiation requires close association with stromal cells. Blood 70:1418-24 139. Hayashi, S.-I., Witte, P. L., Shultz, L. D., Kincade, P. W. 1988. Lymphohemopoiesis in culture is prevented by interaction with adherent bone marrow cells from mutant viable motheaten mice. J. Immunol. 140:2139-47

140. Greaves, M. F. 1986. Differentiationlinked leukemogenesis in lymphocytes. Science 234(4777): 69%704 141. Cooper, M. D., Mulvaney, D., Coutinho, A., Cazenave, P.-A. 1986. A novel cell surface molecule on early Blineage cells. Nature 321:616-18 142. Coffman, R. L. 1982. Surface antigen expression and immunoglobulin gene rearrangement during mousepre-B cell development. Immunol. Rev. 69: 523 143. Osmond,D. G., Nossal, G. J. V. 1974. Differentiation of lymphocytes in mouse bone marrow. II. Kinetics of maturation and renewal of antiglobulin-binding cell studies by double labeling. Cell Immunol. 13:132M5 144. Landreth, K. S., Rosse, C., Clagett, J. 1981. Myelogenous production and maturation of B lymphocytes in the mouse. J. Immunol. 127:2027-34 145. Park, Y. H., Osmond, D. G. 1987. Phenotype and proliferation of early B lymphocyte precursor cells in mouse bone marrow. J. Exp. Med. 165: 44458 146. Opstelten, D., Osmond,D. G. 1983. PreB cells in the bone marrow: lmmunofluorescence stathmokinetic studies of the proliferation of cytoplasmic #-chain bearing cells in normal mice. J. Immunol. 131:2635M0 147. Opstelten, D., Deenen, G. J., Rozing, J., Hunt, S. V. 1986. B lymphocyteassociated antigens on terminal deoxynucleotidyl transferase-positive cells and pre-B cells in bone marrowof the rat. J. Immunol. 137:76-84 148. Deenen, G. J., Hunt, S. V., Opstelten, D. 1987. A stathmokinetic study of B lymphocytopoiesis in rat bone marrow: Proliferation of cells containing cytoplasmic #-chains, terminal deoxynucleotidyl transferase and carrying HIS24 antigen. J. Immunol. 139: 70~ 10 149. Thomas, M. L., Barclay, A. N., Gagnon,J., Williams, A. F. 1985. Evidence from cDNAclones that the rat leukocyte-common antigen (T200) spans the lipid bilayer and contains a cytoplasmic domain of 80,000 Mr. Cell 41:8343 150. Saga, Y., Tung, J. S., Shen, F. W., Boyse, E. A. 1986. Sequences of Ly-5 cDNA:Isoform-related diversity of Ly5 mRNA.Proc. Natl. Acad. Sci. USA 18:6940-44 15 I. Streuli, M., Hall, L. R., Saga, Y., Schlossman, S. F., Saito, H. 1987. Differential usage of three exons generates at least five differenti mRNAs encoding

Annual Reviews


B-CELL FORMATION humanleukocyte common antigens. J. Exp. Med. 166:1548-66 152. Dalchau,R., Fabre,J. W.1981. Identification with a monoclonalantibodyof a predominatelyB lymphocyte-specific determinant of the humanleukocyte common antigen. Evidence for structural andpossible functional diversity of the humanleukocyte common molecule. J. Exp. Med.153:753-65 153. Kincade,P. W., Lee, G., Watanabe,T., Sun, L., Scheid, M.P. 1981. Antigens displayed on murineB lymphocyteprecursors. J. Immunol.127:2262~8 154. Scheid, M. P., Landreth,K. S., Tung, J. S., Kincade,P. W.1982.Preferential but nonexclusiveexpression of macromolecularantigens on B-lineagecells. Immunol.Rev. 69:141-59 155. Landreth,K. S., Kincade,P. W., Lee, G., Medlock,E. S. 1983. Phenotypic and functional characterization of murine B lymphocyteprecursors isolated fromfetal and adult tissues. J. Immunol. 131:572-80 156. Velardi, A., Cooper,M. D. 1984. An immunofluorescenceanalysis of the ontogenyof myeloid, T and B lineage ceils in mousehemopoietictissues. J. Immunol. 133:672-77 157. Medlock,E. S., Landreth,K. S., Kincade, P. W.1984. Putative B lymphocyte precursorcells in early murineembryos. J. Dev.Comp.Immunol.8: 88794 158. Visser, J. W. M., Bauman,J. G. J., Mulder,A. H., Eliason, J. F,, Deleeuw, A. M.1984. Isolation of murinepluripotent hemopoieticstemcells. J~ Exp. Med. 159:1576-90 159. Bertoncello, I., Bartelmez, S. H., Bradley,T. R., Stanley, E. R., Harris, R. A., Sandrin,M.S., Kriegler, A. B., McNiece,I. K., Hunter, S. D., Hodgson, G. S. 1986. Isolation andanalysis of primitive hemopoietic progenitor cells on the basis of differential expression of Qa-m7antigen. J. Immunol. 136:3219-24 160. Goldschneider, I., Metcalf, D., Battye, F., Mandel,T. 1980. Analysis of rat henaopoieticcells on the fluorescenceactivated cell sorter. I. Isolation of pluripotent hemopoieticstem cells and granulocyte-macrophage progenitor cells. J. Exp. Med.152:419-37 161. Muller-Sieburg,C. E., Townsend,K., Weissman,I. L., Rennick, D. 1988. Proliferation and differentiation of highly enriched mousehematopoietic stem cells and progenitor cells in responseto defined growthfactors. J. Exp. Med. 167:1825-40

141

162. Spangrude,G. J., Heimfeld,S., Weissman, I. L. 1988. Purification and characterization of mouse hematopoietic stem cells. Science 241: 5862 163. Hardy, R. R., Hayakawa,K. 1986. Developmentand physiology of Ly-1 B and its humanhomolog, Leu-1 B. Immunol.Rev. 93:53-79 164. Herzenberg,L. A., Stall, A. M., Lalor, P. A., Sidman, C., Moore, W. A., Parks,D. R. 1986.TheLy-1B cell lineage. Immunol.Rev. 93:81-102 165. Hoffman,M. K., Oettgen, H. F., Old, L. J., Chin, A. F., Hammerling,U. 1977. Endotoxin-inducedserumfactor controlling differentiation of bonemarrow-derived lymphocytes. Proc. Natl. Acad. Sci. USA74:1200-3 166. Girl, J. G., Kincade,P. W.,Mizel,S. B. 1984.Interleukin 1-mediatedinduction of x-light chain synthesis and surface immunoglobulinexpression on pre-B cells. J. Immunol.132:223-28 167. Stanton, T. H., Maynard,M., Bomsztyk, K. 1986.Effect ofinterleukin-I on intracellular concentrationof sodium, calcium,and potassiumin 7OZ/3cells. J. Biol. Chem.261:5699-5701 168. Weeks, R. S., Sibley, C. H. 1987. Molecularanalysis of immu~loglobulin expressionin variants of murineB lymphoma, 70Z/3. Somatic. Cell Mol. Genet. 13:205-19 169. Paige, C. J., Schreier, M., Sidman, C. L. 1982. Mediators from cloned T helper cell lines affect immunoglobulin expressionby B cells. Proc.Natl. Acad. Sci. USA79:4756-60 170. Sidman,C. L., Paige, C. J., Schreier, M. H. 1984. B cell maturation factor (BMF): A lymphokine or family lymphokines promoting the maturation of B lymphocytes.J. Immunol.132: 209 22 171. Jyonouchi, H., Kimmel,M. D., Lee, G., Kincade,P. W., Good,R. A. 1985. Humoralfactors in very young NZB mice that enhance the maturation of normalB cell precursors:Partial purification and characterization. J. Immunol. 135:1891-99 172. Jyonouchi, H., Voss, R. M., Good, R. A. 1988. The presence of autoantibodies specific for NZBserumfactors in adult NZBmice and the establishment of monoclonal autoantibodies against these humoral factors. Cell Immunol.113:158-74 173. Landreth,K. S., Engelhard,D., Beare, M. H., Kincade, P. W., Kapoor, N., Good, R. A. 1985. Regulation of humanB lymphopoiesis: Effect of a uri-

Annual Reviews


142

KINCADE ET AL

nary activity associated with cyclic neutropenia. J. Itnmunol. 134:2305-9 174. Lee, G., Ellingsworth, L. R., Gillis, S., Wall, R., Kincade, P. W. 1987. fl transforming growth factors are potential regulators of B lymphopoiesis. J. Exp. Med. 166:1290-99 175. Polla, B. S., Poljak, A., Ohara, J., Paul, W. E., Glimcher, L. H. 1986. Regulation of class II gene expression: Analysis in B cell stimulatory factor 1-inducible murine pre-B cell lines. J. Immunol. 137:3332-37 176. Polla, B. S., Poljak, A., Geier, S. G., Nathenson, S. G., Ohara, J., Paul, W. E., Glimcher, L. A. 1986. Three distinct signals can induce class II gene expression in a murine pre-B cell line. Proc. Natl. Acad. Sci. USA83: 487882 177. Polla, B. S., Ohara, J., Paul, W. E., Nabavi, N., Myer, A., Liou, H.-C., Shen, F.-W., Gillis, S, Bonventre, J. V., Glimcher, L. H. 1988. Differential induction of class II gene expression in murine pre-B-cell lines by B-cell stimulatory factor-1 and by antibodies to Bcell surface antigens. J. Mol. Cell. lmmunol. 3:363-73 178. Mains, P. E., Sibley, C. H. 1983. LPSnonresponsive variants of mouseB cell lymphoma,70Z/3: Isolation and characterization. Somatic Cell Genetics 9: 699-720 179. Sen, R., Baltimore, D. 1986. Inducibility of x immunoglobulin enhancer binding protein NF-xbby a post-transcriptional mechanism. Cell 47: 92128 180. Ohta, M., Greenberger, J. S., Anklesaria, P., Bassols, A., Massague, J. 1987. Two forms of transforming growth factor-/~ distinguished by multipotential haematopoietic progenitor cells. Nature 329:539-41 181. McKearn, J. P., McCubrey, J., Fagg, B. 1985. Enrichment of hematopoietic precursor cells and cloning of multipotential B-lymphocyte precursors. Proc. Natl. Acad. Sci. USA82: 741418 182. Kinashi, T., Inaba, K., Tsubata, T., Tashiro, K., Palacios, R., Honjo, T. 1988. Differentiation of an interleukin 3-dependent precursor B-cell clone into immunoglobulin-producing ceils in vitro. Proc. Natl. Acad. Sci. USA85: 4473-77 183. Paige, C. J. 1985. Analysis of the requirements for murine B cell differentiation. Lymphokines 10:143-63 184. Koch, G., Benner, R. 1982. Differential requirement for B-memory and T-

memorycells in adoptive antibody formation in mouse bone marrow. Immunoloyy 45:697-704 185. Muraguchi, A., Hirano, T., Tang, B., Matsuda, T., Horii, Y., Nakajima, K., Kishimoto, T. 1988. The essential role of B cell stimulatory factor 2 (BSF2/IL-6) for the terminal differentiation of B cells. J. Exp. Med. 167:332-44 186. Kishimoto, T., Hirano, T. 1988. Molecular regulation of B lymphocyte response. Ann. Rev. Immunol. 6:485 512 187. Bosma, G. C., Custer, R. P., Bosma, M. J. 1983. A severe combined immunodeficiency mutation in the mouse. Nature 301:527-30 188. Dorshkind, K., Pollock, S. B., Bosma, M. J., Phillips, R. A. 1985. Natural killer (NK) cells are present in mice with severe combined immunodeficiency (SCID). J. Immunol. 134:3789 189. Dorshkind, K., Keller, G. M., Phillips, R. A., Miller, R. G., Bosma, G. C., O’Toole, M., Bosma,M. J. 1984. Functional status of cells from lymphoidand myeloid tissues in mice with severe combined immunodeficiencydisease, J. Immunol. 132:1804-8 190. Bosma,G. C., Fried, M., Custer, R. P., Carroll, A., Gibson, D. M., Bosma, M. J. 1988. Evidence of functional lymphocytes in some(Leaky) scid mice. J. Exp. Med. 167:1016-33 191. Schuler, W., Weiler, I. J., Schuler, A., Phillips, R. A., Rosenberg, N., Mak, T., Kearney, J. F., Perry, R. P., Bosma, M. J. 1986. Rearrangement of antigen receptor genes is defective in mice with severe combined immune deficiency. Cell 46:963-72 192. Schultz, L. D., Sidman, C. L. 1987. Genetically determined murine models of immunodeficiency. Ann. Rev. Immunol. 5:367-403 193. Greiner, D. L., Goldschneider, I., Komschlies,K. L., Medlock, E. S., Bollure, F. J., Schultz, L. 1986. Defective lymphophoiesis in the bone marrowof motheaten (me/me) and viable mothV) mutant mice. 1. Analyeaten (meV/me sis of the development of prothymocytes, early B lineage cells and terminal deoxynucleotidyl transferase-positive cells. J. Exp. Med. 164:1129-44 194. Medlock, E. S., Goldschneider, I., Greiner, D. L., Shultz, L. 1987. Defective lymphopoiesis in the bone marrow of motheaten (me/me) and viable v) mutant mice. II. motheaten (meV/me Description of a microenvironmental defect for the generation of terminal deoxynucleotidyltransferase - positive

Annual Reviews


B-CELL FORMATION bonemarrow cells in vitro. J. lmtnunol. 138:3590 195. Kincade, P. W., Jyonouchi, H., Landreth, K. S., Lee, G. 1982. B-lymphocyte precursors in immunodeficient autoimmune and anemic mice. Immunol. Rev. 64:81-98 196. Reid, G. K., Osmond,D. G. 1985. B lymphocyte production in the bone marrowof mice with X-linked immunodeficiency (xid). J. Immunol. 135:2299-2302 197. Mond,J. J., Scher, I., Cossman,J., Kessler, S., Mongini,P. K. A., Hansen, C., Findelman,F. D., Paul, W.E. 1982. Role of the thymusin directing the developmentof a subset of B lymphocytes. J. Exp. Med.155:924-36 198. Sprent, J., Bruce,J. 1984.Physiology of B ceils in mice with x-linked immunodeficiency(xid). II1. Disappearance of xid B cells in double bone marrowchimeras. J. Exp. Med. 160:711-23 199, Wortis, H. H., Burkly,L., Hughes,D., Roschelle, S., Waneck,G. 1982. Lack of matureBcells in nude micewith Xlinked immune deficiency. J. Exp. Med. 155:903-13 200. Palacios, R., Henson,G., Steinmetz, M., McKearn,J. P. 1984. Interleukin3 supports growthof mousepre-B cell clones in vitro. Nature309:126 201. Spalding, D. M., Griffin, J. A. 1986. Differentpathwaysof differentiation of pre-Bcell lines are inducedby dendritic cells and T cells from different lymphoidtissues. Cell 44:507-15 202. Ymer,S., Tucker,W.Q. J., Sanderson,

143

C. J., Hapel, J., Campbell, H. D., Young,I. G. 1985. Constitutive synthesis of interleukin-3byleukemiacell line WEHI-3B is due to retroviral insertion near the gene. Nature317:255-58 203. Spooncer,E., Heyworth,C. M., Dunn, A., Dexter, T. M. 1986. Self-renewal and differentiation of interleukin-3dependentmultipotent stem cells are modulatedby stromal cells and serum factors. Differentiation31:111-18 204. Anklesaria,P., Klassen,V., Sakakeeny, M.A., FitzGerald,T. J., Harrison,D., Rybak,M.E., Greenberger,J. S. 1987. Biological characterization of cloned permanent stromal cell lines from anemic S1/Sld mice and +/+ littermates. Exp. Hematol.15:636-44 205. Jyonouchi, H., Kincade, P. W., Landreth, K. S., Lee, G., Good,R. A., Gershwin,M. E. 1982. Age-dependent deficiencyof Blymphocyte lineage precursors in NZBmice. J. Exp. Med. 155:1665-78 206. Yoshida,S., Dorshkind,K., Bearer,E., Castles, J. J., Ahmed, A., Gershwin,M. E. 1987. Abnormalities of B lineage cells are demonstrable in long term lymphoid bone marrow cultures of NewZealand Black mice. J. Immunol. 139:1454-58 G. G., Eisenberg, D., Kin207. Hermanson, cade, P. W., Wall,R. 1988.A newmember of the immunoglobulin superfamily exclusivelyexpressedon Blineagecells. Proc. Natl. Acad. Sci. USA85: 689094





Ann. Rev. Immunol. 1989. 7:145 73 Copyright © 1989 by Annual Reviews Inc. All rights reserved

TH1 AND TH2 CELLS: Different Patterns of Lymphokine Secretion Leadto Different Functional Properties T. R. Mosrnann and R. L. Coffman DNAX Research Institute, California 94304

901 California Avenue, Palo Alto,

Introduction Effector functions in the immunesystem are carried out by a variety of cell types, and as our understanding of the complexity of the system expands, the numberof recognized subdivisions of cell types also continues to increase. B lymphocytes, producing antibody, were initially distinguished from T lymphocytes, which provide help for B cells (1, 2). The T-cell population was further divided when surface markers allowed separation of helper cells from cytotoxic cells (3). Althoughthere were persistent reports of heterogeneity in the helper T-cell compartment (reviewed below), only relatively recently were distinct types of helper cells resolved. In this review we describe the differences betweentwo types of cloned helper T cells, defined primarily by differences in the pattern of lymphokinessynthesized, and we also discuss the different functions of the two types of cells and their lymphokines. Patterns of lymphokinesynthesis are convenient and explicit markers to describe T-cell subclass differences, and evidence increases that manyof the functions of helper T cells are predicted by the functions of the lymphokines that they synthesize after activation by antigen and presenting cells. The separation of manymouse helper T-cell clones into these two distinct types is nowwell established, but their origin in normal T-cell populations is still not clear. Further divisions of helper T cells may have to be recognized before a complete picture of helper T-cell function can be obtained. 145 0732-0582/89/0410-0 145502.00

Annual Reviews


146

MOSMANN & COFFMAN

Lymphokine Activities--The Need for Monospecific Assays Duringthe last several years, our understandingof lymphokinestructure and function has progressedenormously.Dueto the availability of purified proteins, reco~nbinant cDNA clones and monoclonalantibodies, manyof the knownlymphokineactivities can nowbe unambiguously attributed to well-characterized proteins. All knownlymphokinesaffect morethan one cell type and often havediverse effects evenon cells of the samelineage (reviewedin 4, 5, 6). Further complexityis addedby the fact that each type of cell in the immunesystemresponds to morethan one lymphokine. Becauseof this multiplicity of lymphokineaction, monospecificbioassayshavebeendifficult to establish. Monoclonal antibodies(7, 8, 9, 10, 11) havebeenused to improvethe specificity of bioassays, and to measure lymphokines directly by ELISAassays (Table 1). For example,in the T-cell growthassay, both Interleukin 2 (IL-2) and IL-4 can cause proliferation mostT-cell lines, althoughthe dose-response relationships are different (6, 12, 13, 14, 15, 16). Usingmonoclonalantibodies that neutralize the biological activities of IL-2 (6) and IL-4 (9), these bioassays can be monospecificfor either lymphokine(11). TH1 and TH2 Lymphokine Secretion

Patterns

Whenstringent, apparently monospecificassays are used for evaluating lymphokine synthesis, mousehelper T-cell clones fall into twomaingroups. Earlyresults showed that in a panelof clones, eachclonesynthesizedeither IL-2 andInterferon 7 (IFNT),or IL-4 (6). Usinga further set of bioassays and particuarly by evaluating mRNA synthesis by hybridization, the differences in lymphokine synthesis were extendedto a numberof lympho-

Table 1 "Monospecific" assays for lymphokines aBioassays IL-2 IL-4 IL-5 IFN~/ IL-3 GM-CSF

HT2+anti-IL-4 (9) HT2+ anti-IL-2 (7)

MC/9 +anti-IL-4

bELISAs $4B6 (7)+rabbit alL-2 -TRFK2+TRFK5 (10) XMGI.2(1 l)+rabbit alFN3, 8FS+43D11 (8) c31 G6+ 22E9

"Proliferation assaysusingthe indicatedtarget cell line, andblockingmonoclonal antibodiesas indicated. bTwo-sitesandwichassays in whichthe first antibodyis boundto the plate, and the second usedin solution. cj. Abrams,personal communication.

Annual Reviews


THI AND TH2 HELPER CELLS

147

kines and other secreted proteins (11). Table 2 lists our current knowledge of the lymphokinepatterns of the two types of clone, TH1 and TH2. TH1 clones synthesize IL-2, IFNy, and lymphotoxin (LT), whereas these lymphokinesare not detectably expressed in TH2clones. Conversely, only TH2clones synthesize detectable amountsof IL-4, IL-5 (11), and probably IL-6 (F. Lee, T. Mosmann,unpublished). An additional marker for TH2 clones was obtained with the discovery of the induction-specific cDNA clone P600 in cDNAlibraries from an induced TH2done (K. D. Brown, S. M. Zurawski, T. R. Mosmannand G. Zurawski, submitted). The synthesis of IL-2, IFNy, IL-4 and IL-5 is tightly controlled, because induced supernatants of the appropriate cell type contain at least 2,00010,000-fold more of the lymphokine than do induced supernatants of the

Table 2 Properties of mouseT cell clones TH 1

CTL

TH2

Surface markers: LY1 L3T4 LYT2 aLymphokines: Interferon ), Interleukin 2 Lymphotoxin GM-CSF Tumornecrosis factor TY5 P500 H400 Interleukin 3 Met-enkephalin Interleukin 4 Interleukin 5 lnterleukin 6 P600 B cell help: IgM, IgGl, IgA IgG2a IgE Delayed type hypersensitivity: Macrophageactivation: a Lymphokineexpression was evaluated by bioassays, ELISAsand RNA hybridization. b Somebut not all CTLclones produceIL-2.

Annual Reviews


148

MOSMANN & COFFMAN

other cell type. Several other genes were expressed by all clones tested. These included three lymphokines, granulocyte macrophage-colony stimulating factor (GM-CSF), tumor necrosis factor (TNF) and IL-3, neuropeptide, preproenkephalin (ppENK), and three other inductionspecific genes of unknownfunction, TYS, PS00, and H400. Careful analysis of mRNA levels revealed that THl cells expressed relatively more GMCSF, TNF, TY5 (11), H400, and PS00 (K. D. Brown, S. M. Zurawski, T. R. Mosmannand G. Zurawski, J. Immunol. In press), whereas TH2 clones expressed relatively more ppENK(11). These moderate differences (approximatelyfive-fold in most cases) appear to be statistically significant, but we do not yet understand their biological importance. These patterns are characteristic of the majority of long-term T-cell ¯ Clones tested. Someexceptions have also been seen, e.g. expression of IFN7 by a clone that was originally a good example of a TH2. It is not yet knownif these exceptions could be due to aberrations arising in tissue culture. IL-2 synthesis maybe regulated independently of the synthesis of other lymphokines, since THI clones can often lose the ability to produce IL-2 after lo.n_g periods in culture, while the synthesis of other lymphokines appears to be more stable (11). This mayalso account for reports of number of T-cell clones that secrete IFN7 and LT, but not IL-2 (17). Recent evidence from clones that have been grown in culture for short periods (e.g. 4-8 weeks) suggests that there mayalso be other states of differentiation preceding the TH1and TH2phenotypes (T. R. Mosmann, N. Street, H. Bass~and.J. Schumacher,unpublished). As discussed below, the in vivo representation of THI, TH2, and possible precursor phenotypes remains to be established. Lymphokine

Synthesis

Patterns of Other T-Cell Types MouseT-cell clones of the Lyt2+, cytotoxic phenotypes also express a pattern of lymphokines that corresponds closely to the TH1pattern (18, 19; T. A. T. Fong, T. R. Mosmann,unpublished). The CTLclones that we have tested all showed good levels of~ IFN~, and TY5expression, and they produced moderate amounts ofGM-CSF, IL-3, LT and TNFmRNA. No IL-4 or IL-5 synthesis could be detected, and ppENKexpression was detectable at low levels in some clones. These mRNA results have been confirmed by lymphokine assays for IL-3, IL-4, IL-5, GM-CSF,and IFN~. Someclones synthesized moderate amounts of IL-2, whereas IL-2 protein or mRNA was undetectable in other clones. The variability in IL-2 synthesis has been.reported previously (17) and may be physiologically relevant, or it may"be an extreme manifestation of the in vitrolinstability of IL~’2~synthegisin THl.clones. A fourth type of T c~ll, the ~TCI~I+ cell, expresses the 76 form of the T-cell antigen/MHCreceptor and has the

Annual Reviews


TH1 AND TH2 HELPER CELLS

149

surface antigen phenotypeThyl+, L3T4-,Lyt2 . A subset of these T cells constitutes the dendritic T cells found in the epidermis (20), and one exampleof a dendritic TCRl÷ T-cell clone expressedthe THIpattern of lymphokines(T. R. Mosmann, R. E. Tigelaar, unpublished). The expression of the TH1pattern of lymphokinesynthesis in three different T-cell types and the stability of the TH1 andTH2phenotypesin culture suggest that these two patterns of lymphokinegene expression constitute two tightly controlled "cassettes" of regulation. However, there are examplesof T-cell clones, especiallyat early times after establishment in culture, that produce a mixture of TH1and TH2patterns (T. R. Mosmann, N. Street, H. Bass and J. Sehumacher,unpublished; A. Glasebrook, personal communication).These clones and manyhumanclones (21) suggestthat further patterns of lymphokine expressionexist. Hybrids between a TH2clone and a CTL(expressing IFNy)were able to produceboth IL-4 and IFNyin response to stimulation with the antigen recognizedby either parent (22). This suggests that the lack of synthesis of IL-4 in a TH1,or IFNyin a TH2clone, is not due to a suppressive mechanism, but rather is dueto the lack of positive induction. In addition, the data indicate that the phenotypefor secretion of lymphokines is not linked to a particular specificity of the antigenreceptor but is probablya functionof the cell’s state of differentiation. Functions of TH1 and TH2 Cells and Their Lymphokines a-CELLHEL~The ability of THclones to function as helpers for Bcell responseshas beenstudied in vitro using two different experimental strategies. Thefirst is the traditional hapten-carriersystemused for many years to characterize normal THpopulations, In this system, haptenspecific responsesare measuredin cultures containing primedor unprimed B cells, a carrier-specific THcloneandhaptenatedcarrier protein (or cells) as the antigen. In such cultures, the frequencyof B cells that respond specifically is low, typically 10-3 to 10-5 , althoughthe frequencycan be greatly enhancedusing hapten-enrichedB cells (23). Thesecondstrategy is to use THclones specific for antigens such as Ia, Mls, or H-Y(malespecific antigen) whichare expressedon mostor all B cells. In this way, mostB cells are, in essence, antigen-specific,antigen-presentingcells and appear to respond to cell-mediated and lymphokine-mediated signals in the samewayas do antigen-specific B cells in the hapten-carrier systems. Asignificant refinementof this strategy is to use THclones specific for rabbit IgGand rabbit anti-mouseIgMor IgD antibodies as the antigens (24, 25). This system dependsupon a specific interaction betweenthe antigen andIg on the surfaceof the B cells, andit requires processingand Ia-restricted presentationof the rabbit IgG. Amajority of splenic B cells

Annual Reviews 150

MOSMANN & COFFMAN


can be activated and induced to proliferate and differentiate in this way (26). In this discussion, we refer to this secondstrategy as "polyclonal stimulation" and distinguish it from another form of polyclonal response, which we term "bystander stimulation" and which does not require a THB interaction (see below). The recent purification and gene cloning many lymphokines and the development of neutralizing antilymphokine antibodies have madeit possible at last to define whichT-cell products are important in regulating B-cell growth and differentiation and to study the specific functions of each of these products. TH2help for B cells It is generally agreed that murine TH2clones can be excellent helpers, both in antigen-specific (27, 28, 29; DeKruyff, submitted) and polyclonal (25, 30, 31, 32) in vitro cultures. TH2clones induce growth and Ig secretion by 50-80%of B cells in limiting dilution cultures (26) and can efficiently induce responses in populations of small, resting B cells (27, 30, 32). The activation of resting B cells by TH2clones appears to require at least three "signals." The first of these is provided by direct contact with the activated helper cell. A/though a few mouse TH2clones may secrete a soluble factor that can induce growth and differentiation of resting B cells (33, 34), in mostcases, this processrequires direct contact with the TH cell and cannot be achieved with TH2supernatants (35, 36, 37). Optimumproliferation and differentiation require both IL-4 and IL-5, in addition to this cell-mediated activation. This requirement has been defined in two types of experiments. In the first, the addition of neutralizing anti-IL-5 antibodies to cultures of B cells and TH2 clones causes a substantial inhibition of Ig production (10, 30, 37). The addition of anti-IL-4 antibodies also inhibits Ig production but to a lesser and more variable extent (28, 30, 31), whereas the combination of both antibodies inhibits Ig production almost totally. Similar conclusions have been reached in experiments in which TH2 products are used to induce the differentiation of B cells polyclonally activated (but not induced to differentiate) by direct interaction with a TH1 clone. In these experiments, both IL-4 and IL-5 are required for optimal proliferation and Ig production, and no other TH2product was active in this system (31; R. L. Coffman, J. Christiansen, B. Seymour, D. Hiraki, H. Cherwinski, R. Schreiber, M. Bond and T. Mosmann,in preparation). Thus, IL-4 and IL5 are the major "helper factors" produced by TH2cells, and both act to enhance the growth and differentiation of activated B cells. The important exceptionto this is the IgE response, for whichIL-4 but not IL-5 is essential (see below). Large B cells, unlike resting B cells, do not require TH-Bcontact and can proliferate and differentiate in response to TH2supernatants. The

Annual Reviews


TH1 AND TH2 HELPER CELLS

151

active componentin these supernatants has been shownto be IL-5 (38). This response can account for muchof the unlinked "bystander" response observed in TH2-stimulated cultures at high antigen concentrations (37). The response to IL-5, however, is much smaller than the response of the same large B-cell population to direct interaction with a polyclonally stimulating TH2clone; this suggests either that IL-5 alone is a weak stimulus or that it stimulates only a subpopulation of large B cells (R. L. Coffman, unpublished). Nevertheless, this bystander response may be quite significant in some situations. For example, very young NZB/NZW F 1 mice, whichlater in life develop a severe lupus-like autoimmune disease, appear to have a much higher p~oportion of IL-5-responsive B cells. Culture of these cells with IL-5 leads to substantially higher production of autoantibodies than culture of cells from nonautoimmunemice (38). TH1help for B cells The helper function of TH1cells is more uncertain since they have been shownto provide antigen-specific help in some, but not all, in vitro systems. Several groups have reported TH1clones that can help antigen-specific secondary responses in primed B-cell populations (39; DeKruyff, submitted), and the ability of one TH1clone to stimulate primary antihapten antibody responses in unprimed, hapten-purified Bcell populations has been characterized (29). However,Bottomly, Janeway and their colleagues have described manyTH1 clones that cannot provide help for a specific primary response to the phosphorylcholine hapten, although many of these clones can induce polyclonal proliferation and differentiation of B cells at high antigen concentrations (27, 28). However, these authors have presented evidence that some of these clones produce no detectable IL-2 (17, 28), so their results maynot reflect the activity IL-2-producing TH1 clones. Similarly, Abbas and colleagues report that TH2,but not TH1,clones specific for rabbit ~-globulin can induce polyclonal proliferation and Ig production from dense, resting B cells in the presence of rabbit anti-mouse Ig antibodies (30). In our hands, most TH1clones reactive with self- or allo-Ia or with Mls antigens, efficiently stimulate proliferation, but not Ig production, by B cells bearing the appropriate surface molecule (31). The defect in differentiation to Ig production, however, is not caused by an inherent inability of TH1products to induce differentiation, but by insufficient production of IL-2 in vitro. Thus, addition of exogenousIL-2 to such cultures enhances Ig production. This demonstrates that helper function is possible with products of only TH1clones and suggests that IL-2 is the most important helper factor made by THI cells (31). Further enhancement can often be achieved blocking part of the IFN-~activity with anti-IFN-~ antibodies. In other words, manyTHI clones can stimulate activation and proliferation of B

Annual Reviews 152

MOSMANN

& COFFMAN


cells in vitro and can also stimulate Ig secretion independently of the T-B contact-mediated activation signal if the levels of IL-2 and IFN-7 are optimized. It should be noted that the differentiation of THl-activated B cells can also be induced by TH2products, IL-4 and IL-5 (31). TH1cells can also suppress B-cell responses The assessment of the role of TH1cells in B-cell responses is complicated by two things: (a) the fact that high concentrations of IFN-7 can be quite generally immunosuppressive (40, 41, 42), whereas low concentrations can enhance certain types of responses (42), and (b) by the observations that most TH1clones are directly cytotoxic for activated B cells (43, 44). Not surprisingly, activated TH1clones have been shownto be directly suppressive in cultures optimally stimulated by TH2cells (45, 46, 47). The helper activity TH1clones in vitro appears to dominate at low antigen concentrations (suboptimal TH activation) or at low T:B ratios (

INHIBITION

Figure 3 Inter-regulation of THI and TH2. Heavy arrows show the predominant effects expected during the early and late phases of responses dominated by either TH 1 or TH2 responses. The lymphokinedesignated by "?" is possibly the TH2-derivedinhibitor described by Horowitzet al (92).

sites and tumor cells (99, 100), and increased expression of Fc receptors for IgG2aantibodies (101). These receptors could then bind the increased IgG2a levels produced in response to IFN~ (42), leading to increased antibody-dependent macrophage cytotoxicity. Lymphotoxin and IFNy synergize in the killing of target cells (102), and IgG2acan kill target cells by complementlysis. TH1clones also cause effective DTHreactions. All of these effector mechanismsare appropriate for dealing with intracellular (viral and parasite) infections. In general, a strong TH1response in the absence of any TH2response might be expected to result in DTHbut little or no antibody. THETH2RESPONSE Preferential activation of TH2cells should lead to high general antibody levels. IL-4 should cause increased IgE production, as well as increased levels of IgE Fc~ receptor on B cells (103) and Ia antigens on macrophages (104). IL-3 and IL-4 would be expected to result mucosalmast cell proliferation (6), and IL-5 wouldcause proliferation eosinophils (105). Thus, several features of an allergic response are increased by TH2activation (Figure 6). In contrast, TH1clones inhibit the pathway by decreasing TH2growth and inhibiting IgE production by B cells.

Annual Reviews THl AND TH2 HELPER CELLS


GM, LT,~ IL2, IFN’¢ ~

163

ENVIRONMENT FAVORABLE FOR TH1 RESPONSES

ELL II C FOR MHC

IgG2a Figure 4 Immuneresponses against viral antigens. Someof the ex aected features of the immuneresponse during a viral infection are shown. Compared to a response against a protein antigen, the anti-viral response should produce more IFN,~ and LT because of the activation of CTLs. This should then lead to an enhancement of the THl-like side of the response, resulting in a bias towards IgG2aproduction.

MIXEDTHI ANDTH2 RESPONSES Although strongly biased TH1 or TH2 responses would be expected to result in clearly distinguishable immune responses, as described above, many normal responses may involve a mixture of the two types of cell, especially in cases where the response is neither strong nor prolonged. Under these conditions, we might expect that IgE would not be produced due to the dominant suppression by IFN~,, and a DTHresponse might also not occur because of possible inhibition of DTHby a TH2 response (94; T. A. T. Fong, T. R. Mosmann,unpublished). Since both THtypes can activate B cells, which are then responsive to lymphokines produced by either THtype, antibody responses would be strongly supported in a mixed TH1and TH2response. The isotype patterns may depend on the ratio of TH1and TH2 activation, with a THI bias giving preferentially IgG2a, and a TH2preference resulting in more IgG1. These expected properties of a mixed THl and TH2response are clearly compatible with the majority of immuneresponses, i.e. variable IgG isotype responses, without significant IgE or DTHreactions. The

Annual Reviews 164

MOSMANN

& COFFMAN

~

IgM, IgG1, etc.

ANTIGEN+ MNC


DTH

LYMPHOKINE" MEDIATED CYTOTOXICITY

~

~

)

A~TIBODY-DEPENDENT CEI~L-MEDIATED CYTOTOXICITy, PHAGOCYTOSIS, INTRACELLULAR KILLING

Fi#ure 5 TH]-mcdiatcd cffcctor

COMPLEMENTMEDIATED CYTOTOXICITY

functions.

patterns of normal immuneresponses could be even more complexwhen it~ is realized that TH1-biasedor TH2-biased responsescouldtheoretically occur simultaneouslyin different anatomicallocations, providedthat the responseswere not strong enoughto provok~systemic regulatory effects. Immune Responses in Which TH1 and TH2 Ratios May Be Important Several antigens, particularly with certain adjuvants, characteristically induce particular classes of immuneresponse, e.g. alum adjuvant, especially with Bordetellapertussis, provokesgoodIgE responses, whereas completeFreund’sadjuvantgives high antibodylevels but not IgE. Several other biased immune responsessuggest selective activation of the T~I~or TH2pathways(reviewedin.reference 44). A notableexampleis’th~ i.mrnune.,: response to collagen type IV, whichproduces an apparent T~f response in H2~ mice and a TH2response in other mice (106). Several infectious agents mayalso induce biased responses, such as the antiviral response discussed above, and ~t:numberof protozoanand metazoanparasites. NIPPOSTRONGYLUSBRAS1LIKNSIS Mosthelminth parasite infestations induce significant IgE responses, often accompanied by substantial productionof

Annual Reviews THI AND TH2 HELPER CELLS

165

ANTIGEN + MHC


~,~

Eosinophils~ANTIGEN

MEDIATOR RELEASE

IFNY~

~-~ Nast cells

MEDIATOR RELEASE

Figure 6 TH1 and TH2regulation of IgE. Stippled arrows indicate inhibitory effects, and solid arrows showstimulatory effects.

polyclonal IgE. In addition to the high IgE levels, these infestations are often associated with eosinophilia and intestinal mast cell hyperplasia. A well-studied exampleis Nippostrongylus brasiliensis (Nb, 107, 108, 109), in w.hiqh, these effects can be exp!a~_nedby the preferential activation of TH2cells (Table 3). The high IgE levels can be inhibited by in vivo administration of anti-IL-4 antibody (110), and the eosinophilia can inhibited by anti-IL-5 antibody (R. L. Coffman, D. M. Rennick, unpublished). Since t.lae~ T-cell response(possibly irtyplving mast cells) has been implicated in expulsion of wormsfrom the..gut (111), the TH2response can be considered appropriate for this parasite. Wehave recently studied the lymphokinepatterns synthesized by spleen and lymph node cells from Nb-infected mice, and we find that IL-2 and IFN7~Ie)~els are suppressed ,below normal levels, a~d..t~, at IL-4 and IL-5 levels are greatly elevated. , This is probably due both~to selective,amplification of TH2. cells, and regulation of activation of THI and TH2’(N..Stre,et, T.~.R. Mosmann, unpublished). LEISHMANIA ~ Inf_egtion of.~i~ce by Leishmania major ~(Lm)results~in one

Annual Reviews 166

MOSMANN & COFFMAN


Table 3 Immuneresponses against parasites Nippostrongylus brasiliensis High IgE levels (polyclonal) Eosinophilia Mast cell hyperplasia High IL-4 levels High IL-5 levels LowIL-2 levels Low IFN7 levels Leishmania major in mice High IL-4 Low IFN? Balb/c High IgE levels ~High antibody No DTH Low IL-4 High IFN2 C57B1/6 Low IgE Low antibody Strong DTH

Progressive, fatal disease

Limited disease, cure

Nippostrongylus eliminationis associatedwith a TH2-1ike pattern. Leishmaniaeliminationis associatedwith a TH1response.

of two responses (Table 3). In susceptible strains, such as Balb/c, the response has the features expected of TH2activation, such as high antibody levels (including IgE; M. Sadick, R. L. Coffman, unpublished), high IL-4 and low IFN~, expression, and no DTH.The Balb/c mice develop a severe and progressive disease and die (112). In contrast, resistant strains such as C57B1/6, develop strong DTH,low antibody levels with no elevation of serum IgE, high IFN~,, and low IL-4 expression, and the infection is local and ultimately is cured (112). Theseresults and others suggest that a TH1response is effective in eliminating this parasite (an intracellular parasite in macrophages), possibly because of the ability of IFN~to activate macrophages. The most direct evidence for the TH1requirement has recently been obtained by Scott et al (113), who have prepared TH1and TH2cell lines and clones specific for Lmantigens. Whenthese cells are injected back into Lm-infected mice, the TH1line completely cures the infection, whereas the TH2line actually exacerbates the course of the disease. Thus for Lminfection, the THI response is the appropriate response that leads to elimination of the parasite. The Lmsystem in mice is particularly interesting since Leishmania donovani infection in humans also produces two alternative forms of the immuneresponse, either DTH leading to local containment of the infection and elimination of the

Annual Reviews THI AND TH2 HELPERCELLS

167


parasite, or the Kala-Azarresponse in which high antibody levels are associated with low DTHreactions and a severe, disseminated disease (114). It remainsto be seen whetherthese responses are linked to TH1and TH2responses in humans. Concludin 9 Remarks Althoughwe still lack conclusive proof that TH1and TH2cells exist in vivo in the mouse,and especially in humans,the weight of evidence now suggests that, at least in the mouse,these two types of helper cell exist, and because of their profoundlydifferent functions, they are important regulators of the class of immuneresponse. Several major immune responses, especially against parasites, showa remarkably goodfit with the features expected of either TH1 or TH2responses. Becausethe appropriate response(i.e. the responsethat eliminates the infection) can be either TH1 or TH2,dependingon the infectious agent, it is obviously importantto consider the interregulationof these types of cell wheninducingtherapeutic immuneresponses. In closing, someof the outstandingquestions in this area are these: Are TH1and TH2cells found in vivo? Whatare the precursors of TH1and TH2cells? Whatare the lymphokinesecretion patterns of the precursors? Is the commitmentto one of the two lymphokinesecretion patterns made + cells have a similar before or after exposure to antigen? Do humanCD4 dichotomyof lymphokinesynthesis and function? Whichcell regulates the preferential activation of TH1 or TH2by certain antigens, andhowis this achieved? Howis the regulation of TH1 and TH2activation connected to the ongoing pattern of response? Since excellent tools are nowavailable to explore these possibilities further, we look forwardto answersfor many of these questionsin the next few years.

Literature Cited 1. Claman, H. N., Chaperon, E. A., Triplett, R. F. 1966. Thymus-marrow cell combinations. Synergism in antibody production. Proc. Soc. Exp. Biol. Med. 122:1167-71 2. Mitchell, G. F., Miller, J. F. A. P. 1968. Cell to cell interaction in the immune response. II. Thesource of hemolysinforming cells in irradiated mice given bone marrowand thymus or thoracic duct lymphocytes. J. Exp. Med. 128: 821-37 3. Cantor, H., Boyse, E. A. 1975. Functional subclasses of T-lymphocytes bearing different Ly antigens. I. The generation of functionally distinct T-

4.

5.

6.

7.

cell subclasses is a differentiative process independent of antigen. J. Exp. Med. 141:1376-89 Mosmann,T. R., Coffman, R. L. 1987. Twotypes of mousehelper T cell clone: Implications for immuneregulation. lmmunol. Today 8:223-27 O’Garra, A., Umland, S., De 1;rance, T., Christiansen, J. 1988. B cell factors are pleiotropic. Immunol.Today 9: 4554 Mosmann,T. R. 1988. Helper T cells and their lymphokines.In T Cells, ed. M. Feldmann, J. Lamb, M. Owen. New York: Wiley. In press Mosmann, T. R., Cherwinski, H.,

Annual Reviews


168

MOSMANN & COFFMAN

Bond, M. W., Giedlin, M. A., Coffman, R. L. 1986. Twotypes ofmurine h.e!per T cell clone: I. Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol. 136: 2348-57 8. Abrams, J. S., Pearce,. M. K. 1988. Development of rat anti-mouse interleukin 3 monoclonal antibodies which neutralize bioactivity in vitro. J. Immunol. 140:131-37 9. Ohara, J., Paul, W. E. 1985. Production of a monoclonal antibody to and molecular characterization of B-cell stimulatory factor-1. Nature 315:333 36 10. Schumacher, J. H., O’Garra, A., Shrader, B., van Kimmenade, A., Bond, M. W., Mosmann,T. R., Coil’man, R. L. 1988. The characterization of four monoclonalantibodies specific for mouse IL-5 and development of mouse and human IL-5 enzyme-linked immunosorbent assays. J. Immunol. 141:1576-81 11. Cherwinski, H. C., Schumacher, J. H., Brown, K. D., Mosmann,T. R. 1987. Twotypes of mousehelper T cell clone: 3. Further differences in lymphokine synthesis between TH1 and TH2 clones revealed by RNAhybridization, functionally monospecific bioassays and monoclonal antibodies. J. Exp. Med. 166:122944 12. Lichtman, A.H., Kurt-Jones, E. A., Abbas, A. K. 1987. B-cell stimulatory factor 1 and not interleukin 2 is the autocrine growth factor for some helper T lymphocytes. Proc. Natl. Acad. Sci. USA 84:824-27 13. Fernandez-Botran, R., Sanders, V.M., Oliver, K. G., Chert, Y. W., Krammer, P. H., Uhr, J. W., Vitetta, E. S. 1986. Interleukin 4 mediates autocrine growth of helper T cells after antigenic stimulation. Proc. Natl. Acad. Sci. USA 83:9689-93 14. Grabstein, K., Eisenman, J., Mochizuki, D., Shanebeck, K., Conlon, P., Hopp, T., March, C., Gillis, S. 1986. Purification to homogeneity of B cell stimulating factor. A molecule that stimulates proliferation of multiple lymphokine-dependent cell lines. J. Exp. Med. 163:1405-14 15. Smith, C. A., Rennick, D. M. 1986. Characterization of a murine lymphokine distinct from IL-2 and IL-3 possessing a TCGFactivity and an MCGF activity that synergizes with IL-3. Proc. Natl. Acad. Sci. USA 83:1857-61 16. Mosmann,T. R., Bond, M. W., Coffman,R. L., Ohara, J., Paul, W. E. 1986.

17.

18.

19.

20.

21.

22.

23.

24.

25.

.26.

T cell and mast cell lines respond to B cell stimulatory factor-1. Proc. Natl. Acad. Sci. USA 83:5654-58 Woods, A., West, J., Rasmussen, R., Bottomly, K. 1987. Granulocytemacrophage colony stimulating factor produced by cloned L3T4+, class IIrestricted T cells induces HT-2cells to proliferate. J. lmmunol. 138:4293-97 Prystowsky, M. B., Ely, J. M., Belier, L. E., Goldman, J., Goldman, M., Goldwasser,E., Ihle, J., Quintans, J., Remold,H., Vogel, S. N., Fitch, F. W. 1982. Alloreactive cloned T cell lines. Multiple lymphokineactivities secreted by helper and cytolytic lymphocytes. J. Immunol. 129:2337-44 Kelso, A., Glasebrook, A. L. 1984. Secretion ofinterleukin 2, macrophageactivating factor, interferon and colony-stimulating factor by alloreactive T lymphocyteclones. J. Irnmunol. 132: 2924-31 Bonyhadi, M., Weiss, A., Tucker, P. W., Tigelaar, R. E., Allison, J. P. 1987. The y/6 antigen receptor of Thy- 1 + dendritic epidermalcells: Identification of ~ as the Cx gene product. Nature 330:574-76 Paliard, X., de WaalMalefijt, R., Yssel, H., Blanchard, D., Chretien, I., Abrams,J., de Vries, J., Spits, H. 1988. Simultaneous production of IL2, IL4, ÷ and IFN~, by activated human CD4 and CD8÷ T cell clones. J. Immunol. 141:84%55 Havran, W., Fitchg F. W. 1987. Characterization of murine cytolytichelper hybrid T cell clones. Nature 325: 65 67 Vitetta, E. S., Bossie, A., FernandezBotran, R., Myers, C. D., Oliver, K. G., Sanders, V. M., Stevens, T. M: 1987. Interaction and activation of antigenspecific T and B cells. Immunol. Rev. 99:193-239 Tony, H.-P., Parker, D. C. 1985. Major histocompatibility complex-restricted B cell responses resulting from helper T cell recognition of anti-immunoglobulin presented by small B lymphocytes. J. E:~-p. Med. 161:223-41 Tony, H.-P., Phillips, N. E., Parker, D. C. 1985. Role of membrane immunoglobulin (Ig). cross-linking membrane Ig-mediated major histocompatibility-restricted T cell-B cell cooperation. J. Exp. Med. 162: 16951708 Lebman, D. A., Coffman, R. L. 1988. Interleukin 4 causes isotype switching to IgE in T cell-stimulated clonal.B cell cultures. J. Exp. Med. .168: 853-62

Annual Reviews


TH1 AND TH2 HELPER CELLS 27. Kim, J., Woods, A., Becker-Dunn, E., Bottomly, K. 1985. Distinct functional phenotypes of cloned Ia-restricted helper T cells. J. Exp. Med. 162:188-201 28. Killar, L., MacDonald,G., West, J., Woods, A., Bottomly, K. 1987. Cloned, Ia-restricted T cells that do not produce interleukin 4 (IL-4)/B cell stimulatory factor 1 (BSF-1) fail to help antigenspecific B cells. J. Immunol.138: 167479 29. Stevens, T. L., Bossie, A., Sanders, V. M., Fernandez-Botran, R., Coffman, R. L., Mosmann,T. R., Vitetta, E. S. 1988. Subsets of antigen-specific helper T cells regulate isotype secretion by antigen-specific B cells. Nature. In press 30. Boom,W. H., Liano, D., Abbas, A. K. 1988. Heterogeneity of helper/inducer T lymphocytes. II. Effects of interleukin 4- and interleukin 2-producing T cell clones on resting B lymphocytes. J. Exp. Med. 167:1352~53 31. Coffman, R. L., Seymour, B. W., Lebman, D. A., Hiraki, D. D., Christiansen, J. A., Shrader, B., Cherwinski, H. M., Savelkoul, H. F. J., Finkelman, F. D., Bond, M. W., Mosmann, T. R. 1988. The role of helper T cell products in mouseB cell differentiation and isotype regulation. lmmunol. Rev. 102:5-28 32. Tite, J. P., Kaye, J., Jones, B. 1984. The role of B cell surface Ia antigen recognition by T cells in B cell triggering: Analysis of the interaction of cloned helper T cells with normal B cells in differing states of activation and with B cells expressing the xid defect. Eur. J. Immunol. 14:553 61 33. Leclercq, L., Bismuth, G., Theze, J. 1984. Antigen-specific helper T cell clone supernatantis sufificient to induce both polyclonal proliferation and differentiation of small resting B lymphocytes. Proc. Natl. Acad. Sci. USA 81:6491-95 34. Roth, C., Moreau, J.-L., Korner, M., Jankovic, D., Theze, J. t988. Biochemical characterization and biological effects of partially purified B cell-activating factor (BCAF).Eur. J. lmmunol. 18:577-84 35. Snow, E. C., Noelle, R. J. 1987. Thymus-dependent antigenic stimulation of hapten-specific B lymphocytes. Immunol. Rev. 99:173-92 36. Julius, M. H. 1987. Reciprocity in lymphocyte interactions. Immunol. Rev. 95:177-94 37. Rasmussen, R., Takatsu, K., Harada, N., Takahashi, T., Bottomly, K. 1988.

169

T cell-dependent hapten-specific and polyclonal B cell responses require release of interleukin 5. J. lmmunol. 140:705-12 38. Herron, L. R., Coffman, R. L., Bond, M. W., Kotzin, B. L. 1988. Increase autoantibody production by NZB/ NZW B cells in response to interleukin 5. J. ImmunoL 141:84248 39. Giedlin, M. A., Longenecker, B. M., Mosmann, T. R. 1986. Murine T cell clones specific for chicken erythrocyte alloantigens. Cell. Immunol. 97:357-70 40. Coffman,R. L., Carty, J. 1986. AT cell activity that enhances polyclonal IgE production and its inhibition by interferon-~. J. Immunol. 136:949-54 41. Reynolds, D. S., Boom, W. H., Abbas, A. K. 1987. Inhibition of B lymphocyte activation of interferon-y. J. lmmunol. 139:76%73 42. Snapper, C. M., Paul, W. E. 1987. Interferon-3, and B cell stimulatory factor-1 reciprocally regulate Ig isotype production. Science 236:944-47 43. Tite, J. P., Janeway, C. A. Jr. 1984. Cloned helper T cells can kill B lymphomacells in the presence of specific antigen: la restriction and cognate vs. noncognate interactions in cytolysis. Eur. J. Immunol. 14:878-86 44. Janeway, C. A. Jr., Carding, S., Jones, B., Murray, J., Portoles, P., Rasmussen, R., Rojo, J., Kaizawa, K., ÷ T West, J., Bottomly, K. 1988. CD4 cells: Specificity and Function. lmmunol. Rev. 101:39-80 45. Bottomly, K., Kaye, J., Jones, B., Jones, F. III, Janeway, C. A. Jr. 1983. Acloned, antigen-specific, la-restricted Lyt-l÷,2 - T cell with suppressive activity. J. Mol. Cell lmmunol.1: 4249 46. Friedman, S., Sillcocks, D., Rao, A., Faas, S., Cantor, H. 1985. A subset of Ly-I inducer T cell clones activates B cell proliferation but directly inhibits IgG secretion. J. Exp. Med. 161:785 804 47. Asano, Y., Hodes, R. J. 1983. T cell regulation of B cell activation. J. Exp. Med. 158:1178-90 48. Coffman, R. L., Ohara, J., Bond, M. W., Carty, J., Zlotnik, A., Paul, W. E. 1986. B cell stimulatory factor-1 enhances the lgE response of lipopolysaccharide-activated B cells. J. Immunol. 136:453841 49. Finkelman, F. D., Katona, I. M., Mosmann, T. R., Coffman, R. L. 1988. Interferon-7 regulates the isotypes of immunoglobulin secreted during in vivo humoral immune responses. J. Imrnunol. 140:1022-27

Annual Reviews


170

MOSMANN & COFFMAN

50. Finkelman, F. D., Katona, 1. M., Urban, J. F., Snapper, C. M., Ohara, J., Paul, W. E. 1986. Suppression of in vivo polyclonal IgE responses by monoclonal antibody to the lymphokine B cell stimulatory factor-1. Proc. Natl. Acad. Sci. USA83: 967578 51. Savelkoul, H. F. J., Lebman, D., Benner, R., Coffman, R. L. 1988. Increase of precursor frequency and clonal burst size of murine immunoglobulin E-secreting cells by interleukin-4. J. Immunol. 141:749-55 52. Itaya, T., Ovary, Z. 1979. Suppression of lgE antibody production in SJL mice. IV. Interaction of primed and unprimed T cells. J. Exp. Med. 150: 507-16 53. Vitetta, E. S., Ohara, J., Myers, C., Layton, J., Krammer, P. H., Paul, W. E. 1985. Serological, biochemical and functional identity of B cell-stimulatory factor-I and B cell differentiation factor for IgG1. J. Exp. Med. 162:1726-30 54. Erard, F., Corthesy, P., Nabholtz, M., Lowenthal, J. W., Zaech, P., Plaetinck, G., MacDonald, H. R. 1985. Interleukin 2 is both necessary and sufficient for the growth and differentiation of lectin-stimulated cytolytic T lymphocyte precursors. J. Immunol. 134: 1644-52 55. Simon, M. M., Landolfo, S., Diamantstcin, T., Hochgeschwender, U. 1986. Antigen- and lectin-sensitized murine cytolytic T lymphocyte-precursors require both interleukin 2 and endogenously produced immune (gamma) interferon for their growth and differentiation into effector cells. Curr. Top. Microbiol. Immunol. 126: 173-85 56. Pfeifer, J. D., McKenzie,D. T., Swain, S. L., Dutton, R. W. 1987. B cell stimulatory factor 1 (interleukin 4) sufficient for the proliferation and differentiation of lectin-stimulated cytolytic T lymphocyteprecursors. J. Exp. Med. 166:1464-70 57. Takatsu, K., Kikuchi, Y. H., Takahashi, T., Honjo, T., Matsumoto, M., Harada, N., Yamaguchi, N, Tominaga, A. 1987. Interleukin 5, a T cell derived B cell differentiation factor also induces cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. USA 84:4234-38 58. Vadas, M. A., Miller, J. F., McKenzie, I. F., Chism,S. E., Shen, F. W., Boyse, E. A., Gamble, J. R., Whitelaw, A. M. 1976. Ly and Ia antigen phenotypes of T cells involved in delayed-type hyper-

sensitivity and in suppression. J. Exp. Med. 144:10-19 59. Cher, D. J., Mosmann,T. R. 1987. Two types of murine helper T cell clone: 2. Delayed-Type Hypersensitivity is mediated by TH1 clones. J. Immunol. 138:3688-94 60. Zinkernagel, R. 1976. H-2 restriction of virus-specific T-cell-mediated effector functions in vivo. II. Adoptive transfer of delayed-type hypersensitivity to murine lymphocytic choriomeningitis virus is restricted by the K and D regions of H-2. J. Exp. Med. 144:776-87 61. Issekutz, T. B., Stoltz, J. M., van der Meide, P. 1988. Lymphocyte recruitment in delayed-type hypersensitivity. The role of IFN),. J. Immunol. 140: 2989-93 62. Van Loveren, H., Kato, K., Meade, R., Green, D. R., Horowitz, M., Ptak, W., Askenase, P. W. 1984. Characterization of two different Lyt-I ÷ T cell populations that mediate delayed-type hypersensitivity. J. lmmunol. 133: 2402-11 63. Maggi, E., Del Prete, G., Macchia, D., Parronchi, P., Tiri, A., Chretien, I., Ricci, M., Romagnani,S. 1988. Profiles of lymphokine activities and helper function for IgE in humanT cell clones. Eur. J. lmmunol. 18:1045-50 64. Budd, R. C., Cerottini, J.-C., MacDonald, H. R. 1987. Selectively increased production of interferon 7 by _ subsets of Lyt2 + and L3T4+ T cells identified by expression of Pgp-1. J. lmmunol. 138:3583-86 65. Raft, M. C., Cantor, H. 1971. Subpopulations of thymus cells and thymus-derived lymphocytes. In Proc. First Int. Congr. lmmunol. Washington, DC. New York: Academic 66. Kappler, J. W., Hunter, P. C., Jaeobs, D., Lord, E. 1974. Functional heterogeneity among the T-derived lymphocytes of the mouse.I. Analysis by adult thymectomy. J. Immunol. 113:27-38 67. Araneo, B. A., Marrack (Hunter), P. C., Kappler, J. W. 1975. Functional heterogeneity among the T-derived lymphocytes of the mouse. II. Sensitivity of subpopulations to antithymocyte serum. J. Immunol. 114: 747-51 68. Araneo, B. A., Marrack, P., Kappler, J. W. 1977. Functional heterogeneity among the T-derived lymphocytes of the mouse. VII. Conversion ofT1 cells to T2 cells by antigen. J. Immunol.119: 765 71 69. Simpson, E., Cantor, H. 1975. Regu-

Annual Reviews


THl AND TH2 HELPER CELLS

171

lation of the immuneresponse by subBoyse, E. 1987. Alternative use of 5’ classes of T lymphocytes. The effect of exons in the specification of Ly-5 isoadult thymectomy upon humoral and forms distinguishing hematopoietic cell cellular responses in mice. Eur. J. lineages. Proc. Natl. Acad. Sci. USA Immunol. 5:337~13 84:5364-68 70. Janeway, C. A. Jr. 1975. Cellular co81. Arthur, R. P., Mason, D. 1986. T cells operation during in vivo antihapten that help B cell responses to soluble antibody responses. The effect of cell antigen are distinguishable from those number on the response. J. Immunol. producing interleukin 2 on mitogenic 114:1394-1401 or allogeneic stimulation. J. Exp. Med. 71. Swain, S. L., McKenzie,D. T., Dutton, 163:774-86 R. W., Tonkonogy,S. L., English, M. 82. Morimoto, C., Letvin, N. L., Boyd, 1988. The role of IL4 and IL5: A. W., Hagan, M., Brown, H. M., Characterization of a distinct helper T Kornacki, M. M., Schlossman, S. F. cell subset that makes IL4 and IL5 1985. The isolation and characteriza(TH2) and requires priming before tion of the humanhelper inducer T cell induction of lymphokine secretion. subset. J. Immunol. 134:3762-69 Immunol. Rev. 102:77-105 83. Morimoto, C., Letvin, N. L., Distaso, 72. Powers, G. D., Abbas, A. K., Miller, J. A., Alrich, W. A., Schlossman,S. F. R. A. 1988. Frequencies of IL2 and 1985. The isolation and characterIL4-secreting T cells in naive and ization of the human suppressor inantigen-stimulated lymphocyte popuducer T cell subset. J. lmmunol. 134: lations. J. Immunol. 140:3352 57 1508-15 73. Marrack, P. C., Kappler, J. W. 1975. 84. Bottomly, K. 1988. A functional diAntigen-specific and nonspecific medichotomy in CD4 T lymphocytes, lmators of T cell/B cell cooperation. 1. munol. Today 9:268-74 Evidence for their production by dif85. Powrie, F., Mason, D. 1988. Phenoferent T ceils. J. Imrnunol.114:1116-25 typic and functional heterogeneity of ÷ T cells, lmmunol. Today 9: 27474. Swierkosz, J. E., Marrack, P. J., CD4 Kappler, J. W. 1979. Functional analy77 sis of T cells expressing Ia antigens. 86. Kurt-Jones, E. A., Liano, D., Hayglass, Demonstration of helper T-cell heteroK. A., Benacerraf, B., Sy, M.-S., geneity. J. Exp. Med. 150:1293-1309 Abbas, A. K. 1988. The role of antigen75. Waldman,H., Lefkovits, I., Feinstein, presenting B cells in T cell priming in A. 1976. Restrictions in the functions vivo. Studies of B cell-deficient mice. J. of single helper T cells. Immunology 31 : Immunol. 140:3773-78 353-62 87. Rock, K. L., Haber, S. I., Liano, D., 76. Tada, T., Takemori, T., Okumura,K., Benacerraf, B., Abbas, A. K. 1986. Nonaka, M., Tokuhisa, T. 1978. Two Antigen presentation by hapten-specidistinct types of helper T cells involved fic B lymphocytes. III. Analysis of the in the secondary antibody response: immunoglobulin-dependent pathway independent and synergistic effects of of antigen presentation to Interleukin ÷ Ia and Ia helper T cells. J. Exp. Med. I-dependent T lymphocytes. Eur. J. 147:446-58 Immunol. 16:1407-12 77. Janeway, C. A. Jr., Murgita, R. A. 88. Greenbaum, L. A., Horowitz, J. B., Weinbaum,F. I., Asofsky, R., Wigzell, Woods, A., Pasqualini, T., Reich, H. 1977. Evidence for an immunoE.-P., Bottomly, K. 1988. Autocrine ÷ T cells. Differential globulin-dependentantigen-specific helgrowth of CD4 per T cell. Proc. Natl. Acad. Sci. effects of IL-1 on helper and inflamUSA 74:4582-86 matory T cells. J. Immunol. 140: 155578. Bottomly, K., Mosier, D. E. 1979. Mice 60 whose B cells cannot produce the T15 89. Fernandez-Botran, R., Sanders, V. M., idiotype also lack an antigen-specific Mosmann,T. R., Uhr, J. W., Vitetta, helper T cell required for TI5 exE. S. 1988. Lymphokine-mediated pression. J. Exp. Med. 150: 1399regulation of the proliferative response 1409 of clones of TH1and TH2cells. J. Exp. 79. Thomas,M. L., Reynolds, P. J., Chain, Med. 168:543-58 A., Ben-Neriah, Y., Trowbridge, I. S. 90. Muhlradt, P. F., Opitz, H. G. 1982. 1987. B cell variant of mouseT200(LyClearance of interleukin 2 from the 5): Evidence for alternative mRNA blood of normal and T cell-depleted splicing. Proc. Natl. Acad. Sci. USA mice. Eur. J. Immunol. 12:983-85 84:5360-65 91. Kurt-Jones, E. A., Hamberg, S., 80. Saga, Y., Tung, J.-S., Shen, F.-W., Ohara, J., Paul, W. E., Abbas, A. K.

Annual Reviews


172

MOSMANN& COFFMAN

1987. Heterogeneityof helper/inducer T lymphocytes. I. Lymphokineproduction and lymphokine responsiveness. J. Exp. Med.166:1774-87 92. Horowitz, J. B., Kaye, J., Conrad, P. J., Katz, M. E. 1986. Autocrine growthinhibition of a cloned line of helper T cells. Proc. Natl. Acad.Sci. USA 83:1886-90 93. Gajewski, T. F., Fitch, F. W. 1988. Anti-proliferative effect of IFN~, in immune regulation. I. IFN~,inhibits the proliferation of TH2but not TH1 murine HTLclones. J. Immunol.140: 4245-52 94. Parish, C. R. 1972. The relationship between humoral and cell-mediated immunity.Transplant Rev. 13:35-66 95. Katsura, Y. 1977. Cell-mediated and humoral immuneresponses in mice. III. Dynamicbalance betweendelayed type hypersensitivity and antibody response. Immunology32:227-35 96. Coulie, P. G., VanSnick, J. 1985. Enhancementof IgG anti-carrier responses by IgG2 anti-hapten antibodies in mice. J. Exp. Med.15: 79398 96a. Ptak, W., Janeway,C. A. Jr., Flood, P. M. 1988. Immunoregulatory role of Ig isotypes.-II. Activationof cells that blockinduction of contact sensitivity responses by antibodies of IgG2aand IgG2bisotypes. J. Immunol.141: 76~ 73 97. Murgita,R. A., Vas,S. I. 1972.Specific antibody-mediatedeffect on the immuneresponse. Suppression and augmentation of the primary immune responsein miceby different classes of antibodies. Immunology22:31%31 97a. Ptak, W., Flood, P. M., Janeway, C. A. Jr., Marcinkiewicz,J., Green, D. R. 1988. Immunoregulatory role of Ig isotypes. I. Inductionof contrasuppressor T cells for contact sensitivity responses by antibodies of the IgM, IgGl and IgG3isotypes. J. Immunol. 141:756-64 98. Coutclier, J. P., van der Logt, J. T. M., Heessen,F. A., Warnier, G., Van Snick, J. 1987. IgG2arestriction of murine antibodies elicited by viral infections. J. Exp. Med.165:64~69 99. Murray, H. W., Spitalny, G. L., Nathan, C. F. 1985. Activation of mouseperitoneal macrophagesin vitro andin vivoby interferon-~. J. Immunol. 134:1619 22 100. Pace, J. L., Russell, S~ W., Torres, B. A., Johnson, H. M., Gray, P. W. 1983. Recombinant mousey interferon induces the primingstep in macrophage

activation for tumorcell killing. J. Immunol. 130:201113 101. Warren, M. K., Vogel, S. N. 1985. Opposingeffects of glucocorticoids on interferon-y-induced murinemacrophage Fc receptor and Ia antigen expression. J. Immunol.134:2462-69 102. Lee, S. H., Aggarwal,B. B., Rinderknecht, E., Assisi, F., Chiu,H. 1984. Thesynergisticanti-proliferativeeffect of gamma-interferonand humanlymphotoxin. J. Immunol.133:1083-86 103. Hudak,S. A., Gollnick,S. O., Conrad, D. H., Kehry, M. R,~ t987. MurineB cell stimulatoryfactor 1 (interleukin4) increases expressionof the Fc receptor for IgE on mouseB cells. Proc. NatL Acad. Sci. USA84:4606-10 104. Zlotnik, A., Fischer, M., Roehm,N., Zipori, D. 1987.Evidencefor effects of interleukin 4 (B cell stimulatoryfactor 1) on macrophages: Enhancementof antigen presentingability of bonemarrow-derivedmacrophages.J. Immunol. 138:4275-~79 105. Sanderson,C. J., O’Garra,A., Warren, D. J., Klaus, G. G. 1986. Eosinophil differentiation factor also has B-cell growthfactor activity: proposedname interleukin 4. Proc. Natl. Acad.Sci. USA 83:437-40 106. Tite, J. P., Foellmer, H. G., Madri, J. A., Janeway,C. A. Jr. 1987.Inverse Ir gene control of the antibody and T cell proliferative responsesto human basement membranecollagen. J. Immunol. 139:2892-98 107. Ogilvie, B. M., Jones, V. E. 1969.Reaginic antibodies and helminthinfections. In Cellular and HumoralMechanismsin Anaphylaxisand Allergy, ed. H. Z. Movat. Basel/New York: R. Karger 108. Kelly, J. D., Ogilvie,B. M.1972.Intestinal mast cell and eosinophil numbers during wormexpulsion in nulliparous and lactating rats infected with Nippostrongylusbrasiliensis. Int. Arch. Allergy 43:497-509 109. Mayrhofer,G., Fisher, R. 1979. Mast cells in severelyT cell depletedrats and the responseto infestation with Nippostrongylusbrasiliensis. Immunology 37: 145 52 110. Finkelman, F. D., Katona, I. M., Urban,J. F. Jr., Holmes,J., Ohara,J., Tung, A. S., Sample, J. vG., Paul, W.E. 1988. Interleukin 4 is required to generate and sustain in vivo IgE responses.In press 111. Askenase, P. W. 1980. Immunopathologyof parasitic diseases: Involvement of basophils and mast

Annual Reviews THI AND TH2 HELPER CELLS


cells. Springer Semin. lmmunopathol.2: 417-42 112. Locksley, R. M., Heinzel, F. P., Sadick, M. D., Holaday, B. J., Gardner, K. D. Jr. 1987. Murine cutaneous Leishmaniasis: Susceptibility correlates with differential expansion of helper T-cell subsets. Ann. Inst. Past./Immunol. 138: 744-49 113. Scott, P., Natovitz, P., Coffman,R. L., Pearce, E., Sher, A. 1988. Immuno-

173

regulation of cutaneous leishmaniasis: T cell lines which transfer protective immunity or exacerbation belong to different THsubsets and respond to distinct parasite antigens. J. Immunol. Submitted 114. Sacks, D. L., Lal, S. L., Shrivastava, S. N., Blackwell, J., Neva, F. A. 1987. Ananalysis of T cell responsiveness in Indian Kalaazar. J. Immunol. 138:908-13





Ann. Rev. lmmunol. 1989. 7:175-207 Copyright © 1989 by Annual Reviews Inc. All rights reserved

THE STRUCTURE, FUNCTION, AND MOLECULAR GENETICS OF THE 7/3 T CELL RECEPTOR David H. Raulet Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139

INTRODUCTION Distinct subclasses of functional T lymphocytesare involved in induction of antibody production by B cells (helper T lymphocytes) and cytolysis pathogen-infected host cells (cytolytic T lymphocytes, CTL).These T-cell subclasses share several properties: They are largely thymus-derived, bear the Thy-1 antigen (in the mouse), and express a clonally variable, cellsurface antigen-receptor (1-4). The receptor is noncovalently associated with a complexof several invariant polypeptides collectively called CD3 proteins. The two subclasses can be distinguished by expression of two - cells, and most other surface proteins: most helper T cells are CD4+CD8 + CTLare CD4-CD8 cells (5). Unlike B cells, T cells generally recognize fragments of protein antigens only when they are bound to major histocompatibility complex (MHC)encoded glycoproteins on the surface of host cells. Helper T cells and CTLgenerally recognize antigens boundto structurally different classes of MHC proteins, called class 2 and class 1, respectively. Recent studies have demonstratedthat recognition by most T cells of both the foreign antigenic peptide and the MHCprotein is accomplished by a single heterodimeric antigen-receptor structure, composedof disulfide-linked c~ and fl chains (6, 7). Both chains are clonally variable and are encodedby families of variable (V), diversity (D; only for the/~ chain) and joining (J) gene segments assemble by gene rearrangements during T-cell differentiation (3, 4, 8). Although helper T cells and CTLrecognize antigens associated with strut175 0732~0582/89/0410-0175502.00

Annual Reviews


1.76

RAULET

rurally distinct MHC proteins, they utilize the same pool of ~ and/3 gone segmentsto assemble their receptors (8, 9). In t.he face of the considerable progress in. understandingthe recognition systems of functionally defined T lymphocytes, immunologists have recently been confronted with a novel T-cell subclass. Cells of this subclass differ in several respects from "conventional" T cells, but their most striking characteristic is the use of a distinct antigen-receptor-likestructure, composedof clonally variable, CD3-associated 7 and 6 chains. In common with T cells .that express ~/fl receptors (hereafter called ~/fl cells), 7/3 cells express several markers characteristic of T cells and exhibit T-cell functional activities such as cytolysis and lymphokinerelease in .vitro. Progress in defining ~/j~ T cells proceeded in a "conventional" manner after initial detection of T-cell functional activities. The properties of the cells, the structure of their antigen-receptors, and the characteristics of the corresponding gone families followed sequentially. In contrast, progress in understanding 7/6 cells is proceeding in the reverse order: the first discovery was of a gone encoding an unknownantigen-receptor-like chain (10), followed by discovery of a novel T-cell subclass bearing the chain (11). The characteristics of these cells, and the specificity of their receptors is currently under intense scrutiny. As of this writing, the biological function and specificity of 7/3 cells is unknown.7/6 cells are found in seyeral if not all vertebrate species; t.hey express clonally variable receptor mo._lecules encodedby complexgone families and exhibit various functional activiti.es similar to those of e//3 T cells--all points to be discussedlater. Theresee.ros little doubt that these cells function in vivo, perhaps to mediate heretofore unrecognized immunereactions. The purpose of this review is to assemble current knowledgeof the structure of 7/6 receptors and the corresponding genes, the phenotypic and functional properties of 7/6 cells, and the localization and ontogeny of this unique lymphocyte subclass. Murine, human, and avian systems are discussed later,, although a bias to the murinesystem is evident. This information represents most of the available pieces of the 7/6 puz.zle~ whichthe reader is invited to help solve.

HISTORICAL

OVERVIEW

In early 1984 cDNA clones encoding a T-cell receptor (TCR)subunit (the /3 chain) were firstisolated by subtractive hybridization (l 2) and differential screening (13) methods.:Thes~.meth0ds are, based onthe assumption that TCRmRNAs are prese.nt inT-cells but,not in-B cells and that the genes encoding TCRsundergo T cell-specific, clonally variable gene rearrangements. Subsequently, a candidate ~ chain cDNAclone was reported in

Annual Reviews


GAMMA AND DELTA

TCR

177

mid-1984, again by subtractive hybridization cloning (10). The deduced amino acid sequence of the latter cDNAhad clear similarities to immunoglobulin and TCR-fl chains in structure and sequence; the corresponding gene was rearranged in T-cell clones; and corresponding mRNAwas detectable in cloned CTLlines. But it soon becameclear that this cDNA could not encode the u-chain, because a-chains are glycosylated on asparagine residues (1), while th~ protein sequence deduced from the cDNAcontained no consensus sites for asparagine-linked glycosylation (N-x-S/T). Bona fide a-chain cDNAs were isolated shortly thereafter (14, 15). The original candidate u chain cDNAwas renamed "~," and efforts to determine the function and structure of the ~ chain were underway. Over the next two years, the diversity, ontogeny of expression, and genomicstructure of ~ genes were investigated in considerable detail. The discovery that ~ mRNA levels were highest in the thymocyte subset that includes the least mature thymocytes (CD4-CDS-or "double-negative" thymocytes), and very low in most resting T lymphocytes, provokedspeculation that the ~ polypeptideis expressedin the progenitors of u/fl T cells and plays a role in their ontogeny (16-18). Subsequent reports demonstrating a preponderance of nonproductive ~-rearrangements in cloned ~/fl T cells raised doubts about these models and suggested that ~ genes do not function in u/fl T cells (19-25). During this period, human7 genes were first characterized (26-28), and the diversity of V~ genes in both murine and humansystems was found to be more extensive than previously thought. Altogether seven murine (2932) and seven or eight functional humanV~ genes (24, 25, 34-36) have been identified; this suggests a relatively limited repertoire of V~genes in both species. In the summerof 1986, evidence was first provided for a T-cell subset that expressed a second heterodimeric T cell receptor (11, 37). The use monoclonal antibodies reactive with all u/fl receptors on humanT cells + cells that allowed the identification of a small subset of peripheral CD3 do not express u/fl receptors. These cells, as well as certain cell lines and lymphomas, were shown to express a CD3-associated heterodimer of a ~ chain with a second unidentified chain (11, 3741). Subsequently, these findings were extended by the detection of 7-containing TCRson a subset of murine adult double-negative thymocytes (42). Expression of ?-containing heterodimeric receptors was also detected on a subset of 5-10% of murine fetal thymocytes, beginning around day 14 of ontogeny, long before u/fl receptor expression can be detected (ca. day 17, 18) (43, 44). In both humanand murine systems, evidence provided that the second chain of the new T-cell receptor was not u or/3,

Annual Reviews


178

RAULET

thus arguing against modelsin which ~ pairs with either the ~ or/~ chains and fueling the expectation that a fourth chain, called 6, remained to be cloned (11, 42). The derivation by several groups of ?/6 T-cell lines and hybridomas that carry out typical T-cell effector functions (cytolysis of target cells, lymphokinerelease) strengthened the notion that ~/6 cells maybe immune effector cells as opposed to ontogenic intermediates (37, 39, 40, 45-49). Other models, in which ?/6 T cells are a separate lineage that regulate the developmentof ~/fl T cells, were also proposed (42). In the summerof 1987, a surprising turn of events yielded the 6 chain gene. A study of unusual rearrangements occurring just upstream of the J~ gene segments revealed the presence of novel TCR-like C, J, and D gene segments that rearrange in double-negative thymocytes (50, 51). Subsequently, protein sequencing data and serological evidence confirmed the assignment of this novel gene as the 6 gene (52-55). In addition highly related human 6 cDNAhomologue was isolated independently by subtractive hybridization procedures (56). Around the same time, it was determined that a previously detected population of Thy-1 ÷ marrow-deriveddendritic cells resident in the skin ("Thy-1 ÷ dendritic epidermal cells" or Thy-1 ÷ DEC) express predominantly CD3-associated ~/6 receptors (57-60). Subsequent studies have demonstrated that a preponderance of T cells in the murine (61, 62) and chicken (63) intestinal epithelium are y/6 cells. These observations have provokedthe suggestion that ~/6 cells are poised as a line of defense against epidermal or gut infections and/or tumors. Recently, success has been reported in deriving several cloned 7/6 cell lines with clear specificity for allogeneic MHCor MHC-related gene products (64, 65). These reports raise the possibility that 7/6 cells normally react with foreign antigens bound to self-MHCproteins. Overall, the observations of T-cell effector activities of 7/6 cells, the peripheral locations of someof these cells, and the MHC-relatedspecificity of at least some~/6 receptors, represent suggestive evidence that ~/6 cells are a distinct subset of peripheral effector T lymphocytes with specificity for MHCor antigen-modified MHCglycoproteins. This view remains to be confirmed by direct experimentation. The most striking outstanding issues include the nature of ligands recognized by 7/6 cells, and the physiological and evolutionary necessity for a separate subset of T lymphocytes. ~ CHAINS

AND GENES

Murine Murine y genes are located on chromosome13 (67). The seven V genes and four C genes are organized as shownin Figure 1 (22, 29-32, 68). The

Annual Reviews GAMMA AND DELTA TCR

179

MURINE GAMMA GENES V5

V4 V3 Jl C1

V2

VI,3

HUMAN GAMMA GENES


n n n n/,dL.v,,

n n n n

n n,./.L_B__L/,.t

i

i

trot

t I

~

MURINE DELTA GENES

HUMAN DELTA GENES v~, v~ ~z~r~s

D6z /,. I

--5-O--~Z-O-O--~-gJ-S ¢=

DSZ ~6~ ~82 =Zl3 C6 II I I r’--I

I

I

,./

,

~

~~o kb

Fiyure l Genomicorganization of murine and human7 and 6 gene families. For simplicity the exons of the C genes are not depicted. Brackets (//) indicate gaps; the murine 7 genes have been linked by pulsed field electrophoresis (68), although the orientation of the V71.3, J73 and C73genes relative to the other genes is not known.Functional genes are listed above the lines while pseudogenes are listed below the lines. The humanVyl.5 sequence is not determined. The mapsare drawnroughly to scale. The ticks below the lines of the murine ~, and ~ genes correspond to the approximate location of EcoRI sites in the BALB/cstrain.

interspersal of V and J-C gene segmentsis reminiescent of the ,~ light chain locus and differs from most other receptor gene loci, including the human 7 genes (see below). vy GENES ANDREGIONS The seven routine V7 genes isolated thus far fall into five subfamilies based on sequence similarity (Figure 2). The VTI subfamily [according to the nomenclature of Garmanet al (30); see Table 1 for a cross-referencing of murine 7 gene nomenclatures] includes three V genes, V71.1, V71.2, and V71.3, whichare highly related at the nucleotide and amino acid levels (23, 29); V72, V~3, V74, and V75 are each quite distinct in sequence from each other and from the V7I genes (17-50% homologyat the amino acid level) (30-32). The V~ genes, like other V genes, are composed of a major exon, encoding most of the variable region chain, preceded by an exon that encodes the 5’ untranslated portion of the mRNAand the presumed

Annual Reviews 180

RAULET

MOUSE -2 +i

i0

20

30

40

VYl.l MLLLRWPTFCCLWVFG LGQLEQT{LSVTRATDE~AQISCIVSL~YFS NTAIH~YRQK] V~I.2 MLLLRWFTSCCLWVFG LGQLEQTELSVTRETDENVQISCIVYLPYFS NTAIHWYRQK’ VYI.3 G LGQLEQTELSVTREQDESAQISCIVSLPYFS NTAIHWYRQK] ---PLLKVVIFLCLLTFG V?2 HGKLEQPEISISRPRDETAQISCKVFIESFR SVTIHWYRQK[ V?3 MSTSWLFLLSLTCVYG DSWISQDQLSFTRRPNKTVHISCKLSGVPLH NTIVHWYQLK< V?4 MGLLLQVFTLASLRIYSEG SSLTSPLGSYVIKRKGNTAFLKCQIKTSVQKPDAYINWYQEK V?5 MLWALALLLAFLPAGR QTSSNLEERIMSITKLEGSSAIMTCDTHRTGT YIHWYRFQ


HUMAN vTi .2 V?l.3 V?l.8 V?3

MQWALAVLLAFLSP MRWALAVLLAFLSP MQWALAVLLAFLSP MLLALALLLAFLPP MLSLLHTSTLAVLGA MSLLEAFAFSSWA MPLVVAVIFFSLWV

AS QKSSNLEGRTKSVIRQTGSSAEITCDLAEGSNG YIHWYLH AS QKSSNLEGRTKSVTRQTGSSAEITCDLTVTNTF YIHWYLH AS QKSSNLEGRTKSVIRQTGSSAEITCDLAEGSTG YIHWYLH AS QKSSNLEGRTKSVTRPTGSSAVITCDLPVENAV YTHWYLH LCVYGAGHLEQPQISSTKTLSKTARLECVVSGIKIS ATSVYWYRE LG LGLSKVEQFQLSISTEVKKSIDIPCKISSTRFE TDVIHWYRQ FA LGQLEQPEISISRPANKSAHISWKASIQGFS SKIIHWYWQ

Figure2 Deducedamino acid sequences of murine and humanV7 genes.The nomenclatures are those of Garman et al (30) ~r the murinegenes (see Table 1) and Strauss et al (I 15) and Brenneret al (116) ~r the humangenes (an alternative nomenclature~r human7 genes has been proposed; 36). In parentheses to the right of each sequence are re~rences ~r the sequences. Gapswere inserted to align the sequences ~r comparison, and the conserved cysteine residues are indicated by asterisks. The amino-terminiof the matureproteins (+ 1) are estimated. The leader sequence of murineV72is 24 aminoacids longer than shown(30).

cleaved signal peptide of the chain. Heptamer-nonamer recombination signal sequences separated by 23 bp spacers are found just to the 3’ side of each sequencedV7 gene (29, 30). c7 CEYES aYDREC~OYS Of four murine C~ genes (Figure 3; 22, 29, 30), C73is apparently a pseudogene, at least in BALB/cmice, by virtue of a defective 5’ splice site borderingthe secondexon(29). In addition, the J73C~3gene is deleted entirely in several mousestrains, including C57BL/10 (22); it is present in the C57BL/6 and BALB/cstrains. Three of the Cy genes, C71, Cy2, and ~Cy3, are very similar in coding sequence(10, 29, 30). C?I and C~2differ by only six replacementsof 290 aminoacids, anda 5 aminoacid insertion in C71.This insertion is located just to the amino-terminalside of the cysteinc residuc used for disulfide linkage to the 6 chain, in the region connecting the disulfide-linked C domain to the plasma membrane. The C74gene differs significantly in sequencefromthe other Cy genes, with about 66%overall amino acid identity and 76%nucleotide identity in the coding regions (22). In addition, the C74 sequence contains a

Annual Reviews 181

GAMMA AND DELTA TCR

50

60

70

80

90

i00 *

REFERENCES


QQFEYLIYVATNYNQR PLGGKHKKIEASKDFKSSTSTLEINYLKKEDEATYYCAVWM QQFEYLIYVETNYNQR PLGGKNKKIEASKDFQTSTSTLEINYLKKEDEATYYCAVWI QGLZFLLYVLATPTHI FLDKEYKKMEASKNPSASTSILTIYSLEEEDEAIYYCSYG EPLRRIFYGSVKTYKQ DKSHSRLEIDEK DDGTFYLIINNWTSDEATYYCACWD QRLQRMLCSSSKENIV YEKDFSDERYEARTWQSDLSSVLTIHQVTEEDTGTYYCACWD R~EHLLYYNFVSSTTVVDSRFNSEKYHVYEGPDKRYKFVLRNVEES DSALYYCASW

(15) (23) (30,31) (30) (30) (32)

GKAPQRLQYYDSYNSKVVLESGVSPGKYYTYASTRNNLRLILRNLIEN DSGVYYCATWDG (34) GKAPQRLLYYDVSTARDVLESGLSPGKYYTHTPRRWSWILRLQNLIEN DSGVYYCATWDR (24,34) ,GKAPQRLLYYDSYTSSVVLESGISPGKYDTYGSTRKNLRMILRNLIENDSGVYYCATWDG (34) GKAPQRLLYYDSYNSRVVLESGISREKYHTYASTGKSLKFILENLIER DSGVYYCATWDR (34,41) GEVIQFLVSI SYDGTVRKESGIPSGKFEVDRIPETSTSTLTIHNVEKQDIATYYCALWEV (72) NQALEHLIYIVSTKSAARRSMGKTSNKVEARKNSQTLTSILTIKSVEKEDMAVYYCAAWWV (35,36) (36) NKGLEYLLHVFLTISAQ DCSGGKTKKLEVSKNAHTSTSTLKIKFLEKEDEVVYHCAC Figure 2 (Continued).

amino acid (69 bp) insertion (comparedto Cy2) immediately preceding the cysteine residue thoughtto be involved in a disulfide-linkage with the 6 chain. In BALB/c mice, the C71gene contains a single potential site for asparaginc (N)-linkcd glycosylation while the Cy2 gene contains none. In some other strains, such as C57BL/Ka and C57BL/10,the N-glycosylation site in C71is absent (66). The C74gene from C57BL/10also contains a single site for N-glycosylation. Knowledge of the potential glycosylation sites of

Table 1 Correspondencebetween the different mouseVy genenomenclatures

~ VV Vl,l VI.2 Vl.3 V2 V3 V4 V5

Haydayet al (Ref. 29) VI0.8B V10.8A V5,7

Heilig & Tonegawa (Ref. 31)

Trauneckeret al (Ref. 23)

Hucket al (Ref. 36)

V2 VI V3 V4.3 V4.1 V4.2 V4.4

V2 VI V3 V4 V6 V5 V7

V1 V2 V3 V4 V5 V6 V7

Thenomenclature usedin this reviewis fromGarman et al (30).


Annual Reviews

182 RAULET

M

0

Ol

z

t~

0 E~

Annual Reviews


GAMMA AND DELTA

TCR

183

Cy and Vy chains has been useful in some cases in identifying the rearranged ? genes that encode particular ? polypeptides. Each of the four ~ constant regions includes a hydrophobic transmembranedomain of about 20 residues, followed by a short cytoplasmic domain of about 12 amino acids. The cytoplasmic domains of C71 and C72 are identical but differ significantly from the corresponding domain of C74. Like other TCRC genes, all C7 genes encode a lysine residue in the transmembrane domain, which is presumably buried within the membrane.It has been proposed that the lysine residues are involved in intramembrane interactions with other chains, such as those CD3components that display acidic residues within the membrane. C71, C72, and C73 are each comprised of three exons (29, 30). The first (CI) exon encodes most of the extracellular portion of the constant region, while the short CII exon encodes a stretch of variable length (see above) that separates the C domain from the plasma membrane; the CII exon also encodes the cysteine residue presumablyinvolved in a disulfide bridge with the 6 chain. The third exon encodes a short extracellular region, the hydrophobic transmembrane domain, the short intracellular domain, and the 3’ untranslated region of the mRNA. J7 GEYESF~GMEYTS Each of the four C7 genes is preceded by a single knownJ7 gene segment which is numberedaccording to the corresponding C gene (22, 23, 29). In accord with the similarities betweenC7 genes noted above, the sequences of JTl and J72 are extremely similar (identical at the amino acid level), whereas J74 differs from JT1 and J72 at 9 of 19 amino acid residues (Figure 4). Most of the differences occur at the aminoterminal end of the J segments, where they may contribute to the third complementarity determining region (CDR) of the corresponding regions. The J~/3 gene segment,like its correspondingC73gene, is defective, by virtue of two frameshift mutations in the coding regions (23). The genes are flanked on their 5’ sides by heptamer-nonamer recombination signal sequences separated by 12 bp spacers (29). REARRANGEMENT OF MURINE V’~ GENES The murine 7 gene family is often considered as clusters of VT/JTCygenes, based on the genomicorganization and the observations that most V~/rearrangements are to the Jy-C7 gene that is most proximal and in the same transcriptional orientation (Figure 1). Thus VTl.1/J74C74 is one cluster, V71.2/J72C72 is another, V7 t .3/J73C73 is a third, and the fourth cluster is V75,V72,V74,V73/J~ 1C 71. Rare instances of V7 rearrangements to more distal J7C7 genes have been reported, including V75/J74C74(32), VTI.3/J74C74(69), and V72/J74C74 (J. Segal, D. Cohen, personal communication). Table 2 summarizesthe observed rearrangements of murine ~ genes and

Annual Reviews 184

RAULET Murine J.~ sequences

J~l, J~ J~4

SSGFHKVFAEGTKLIVIPS (29) GTSWV-I--K ....V---P (22)

HumanJ.~ sequences


J~l.1 TTGWFKIFAEGTKLIVTSP (73, 73a) J~2.1 SSD-I-T--K--R ...... (73a) JTi.2 GDELGKKI-V-GP ..... I- (34,73a) J~l.3 & NYYK-L-GS--T-V-(33,72)

Murine J~ seouences J81 TDKLVFGQGTQVTVEP J82 SWD-RQMF--T-IELF---

(51) (69,76)

HumanJ~ sequences J~l TDKLIFGKGTRVTVEP J~2 SWD-RQMF--T-IKLF--J63 SWD-RQMF--T-IELF---

(55,86) (82) (82a)

JT2.3 Ftlqure4 Murineand human J7 and J6 amino acid sequences.Identities are indicated by dashes.In parentheses to the right of eachsequence are references. what is currently knownof their expression in y/6 cells in different tissue sites. Remarkably, it appears that different 7 genes are expressed preferentially in different tissues (see V~and V6usagein peripheral 7/6 cells). Human The human7 genes, located on chromosome7 (26, 28), are organized a fashion similar to that of the//chain genes and unlike that of the murine 7 genes: two neighboring JyCy gene clusters are flanked on their 5’ sides by an array of Vygenes (Figure 1; 27, 70). V~ GENES ANDREGIONS Currently 14 humanV7 genes have been identified, of which only seven or eight are potentially functional (Figure 2; 24, 25, 33-36). The functional humanV7 genes fall into four subgroups. Subgroup I consists of at least four functional genes (V1.2, V1.3, V1.4, V1.8) and four nonfunctional genes. The functional subgroup I sequences are identical at 67-87%of their amino acids. Subgroups II, III, and IV each include a single V gene (V2, V3, and V4, respectively) and differ considerably sequence from each other and subgroup IV segments (23~44%amino acid identity between the subgroups). Comparisonof the amino acid sequences to routine V7 sequences reveals similarity between humansubgroup I and murine V75 (up to 49% identity), and between the human subgroup sequence and the murine V71subfamily (up to 51%identity) and V72gene (52%identity). GENESANDREGIONS The sequences of the two human CT regions are very similar overall, although heterogeneity in the membraneproximal connector region of the C~2gene results in significant differences in the structure of the two C~regions (Figure 3). Like murine C7 genes, the humanC71 gene is composedof three exons, C7


185


Table 2 V7/Cv gene rearrangement and expression in the mouse Rearrangement is rare in fetal thymus, commonin thymomas(32). Predominant7 chain of intestinal intraepithelial 3’/(5 cells (62). Rearrangementdetected in fetal thymus, frequent in adult double negative thymocytes,and in e//3 T cells (30, 31); in the latter population is often but not always nonproductively rearranged, and seldom transcribed. Protein is predominant V7of 7 chain in expressed by _adult thymic 7/6 cells (42; Figure 6) and common in splenic "~/6 cells (93). V74-J ~ 1C~1 :

Rearrangementis relatively frequent in fetal thymocytes, and infrequent in adult double negative thymocytes(30). Protein detected hybridomas of adult double negative thymocytes (69).

V?3-J71C]? 1:

Rearrangementis relatively frequent in early fetal thymocytes(30, 31) and DECcells (60, 89), but undetectable in adult double negative 3,/6 cells. Protein is predominant V~, of 7 chains of DECcells and probably in early fetal thymocytes(92, 96). Rearrangementcommonin late fetal and adult double negative cells, and in c~//~ T cells (30); in the latter populationit is often but not always nonproductively rearranged 09-23, 99). Unlike V72-JTIC~l, V1.2J72C72transcripts are detected in activated (but not resting) c~//~ T cells, particularly those of the CD8phenotype (15, 30, 74). There is evidence that 7/6 cells expressing VTI.2-J72C72chains are present in MLR activated spleen cell preparations (85, 114), and indications that an MHC-IE reactive 7/6 T cell hybridomamayexpress the V71.2-J72C72protein (L. Matis and J. Bluestone, unpublished data). Rearrangement of V71.3 to J73C73 has been rarely observed in cloned 7//~ T cells (29); because Cy3is a pseudogene,these rearrangements, if they occur in 7/3 cells, are presumably non-functional. Rearrangementof V71.3to J74C74has been detected in an adult thymic 7/~ T cell hybridoma, and appears to encode the ~ chain expressed on this cell (69). The V71.3 gene (and the J~3-C73gene) is deleted in somemousestrains (e.g. C57BL/10)(22). Rearrangementof V’~ 1.1 to Jy4C74has been observed in some~//3 T cell clones (22), and in somey/fi T cells (69). The rearrangementis detectable but present in a minor fraction of adult thymic ~/~5 cells (Figure 6). The rearrangement of V71.1 to J74C74 is the most commonC74 rearrangement observed. Protein level expression of J74C74-containing chains has been demonstratedin the case of DECcell lines (59a) a small fraction of fetal and adult thymic 7/8 cells (96), and a large fraction of splenic 7/~5 cells (86).

VV1.2-Jy2Cy2:

V71.3:

V71. I-Jy4C?4:

Sites of predominant usageof the indicated7 chain are underlined.

Annual Reviews


186

RAULET

corresponding to the disulfide-linked C domain (exon I), the membrane proximal connector region (exon II), and the transmembrane and cytoplasmic domains(exon III) (70). While the regions of CTl and C72 encoded by the first and third exons are virtually identical in their amino acid sequences, the membraneproximal connector regions differ markedly. First, the C72gene, but not the C71gene, includes two or three homologous tandemcopies of the CII exon (CIIa, b, and c), each of which is similar sequenceto the single CII exon of CT1(41, 70, 71, Figure 3). Becausethese tandem copies of the CII exon are spliced together in the corresponding mRNA,the size of the C~2 connector region is larger than that of C71. Due to genetic polymorphism,different individuals have two (b and c) three (a, b, and c) copies of the CII exon of C72. Secondly, while the CII exon of the C71 encodes a cysteine residue required for disulfide linkage to the 6 chain, none of the three CII exons of C72 encodes a cysteine. Therefore, this chain cannot form a covalent linkage with the 6 chain, and this accounts for the observation that manyhuman~/6 receptors are not covalently associated (71). The heterogeneity in the size and sequence the connector region of murine and human7 chains may have functional as well as structural consequences. J7 GENESEGMENTS While each murine C? gene is preceded by a single J segment, the humanC7 genes are preceded by two (C72) or three (C71) segments(Figure 1; 72, 73, 73a). All five J7 gene segments are potentially functional and rearrangements of each to V~ gene segments have been documented(compiled in 36). Twoof the J7 segments, one in each cluster, encode identical aminoacid sequences, whereasthe others differ markedly, particularly at the aminoterminal ends (Figure 4). Junctional

Diversity

of Murine and Human ~ Genes

No D elements have been identified for 7 genes in mouse or human. However,significant variability exists at manyVT-J7junctions due to the imprecision of the joining process and the addition of short stretches of nucleotides, called N regions, at the V-J junction during joining (25, 36, 73, 74). The latter process, presumably catalyzed by terminal deoxyribonucleotidyl transferase (TdT) in progenitor lymphocytes, adds nucleotides to junctions in a template independent fashion (75). Most of the published sequences are of nonfunctional 7 gene rearrangements from e//3 T cells, and there is currently only limited data concerning ~ junctional diversity in murine 7/3 cells. The sequences of four human 7 cDNAsknown to encode 7 polypeptides at the cell surface include N regions of 3 5 nucleotides (compiled in 36). In contrast several murine dendritic epidermal cell lines lack N regions (see below). Consideringthe limited repertoire


187

Vyand Jy genes, and the apparent lack of D elements in y genes, junctional diversity mayaccount for muchof the overall diversity of y chains. 6

GENES


Murine

AND CHAINS

V6 Genes

The murine C~, J~i, and De5 gene segments are located on chromosome14 between the V~ and J~ gene segments (Figure 1; 50, 51). The C6 gene ~ 75 kb upstream of the Ca gene, but only ~ 8 kb upstream of the most 5’ knownJc~ gene segments. TwoJ6 and two D6 gene segments lie on the 5’ side of C6(51, 69, 76). At present, eight murine V6 gene subfamilies have been identified (Figure 5). As assessed by Southern hybridization, several of these subfamilies appear to include only one member(V61, V62, V63, V65), while the others include two or moremembers(65, 69, 77). It is likely that more V6genes will be identified. The location of D6, J6, and C6 genes between V~ and J~ gene segments raises the possibility that a shared pool of V genes is utilized to produce both ~ chains and 6 chains. The current evidence suggests that there is indeed someoverlap in V-usage between the ~ and 6 chains. Thus, four of the eight V6 gene subfamilies (V63, V66, V67, V68) overlap with or are identical to knownVc~ subfamilies (V~6, V~7, V~4, V~I 1, respectively) (Figure 5); in fact, V63and V~6are probably identical (77). However, of the V6 subfamilies are very different than knownV~ sequences (Figure 6), including V6 genes commonlyexpressed by fetal thymocytes(V61) (76), y/g-dendritic epidermal cells (also V6I, J. Allison, personal communication) and adult thymic double negative 7/c5 cells (V65; 69, 77). That these commonlyused V6 genes have not been observed in surveys of large numbersof c~ cDNA clones indicates that they are preferentially, possibly even exclusively, used to produce 6 chains. At present the mechanismsthat account for preferential usage of certain V gene segments to produce 6 versus ~ chains are not known. The possibility that V~ and V6 genes are distinguished by flanking recombination signals is unlikely, since two sequencedV6genes, like V~genes, are flanked by recombination signal sequences separated by 23 bp (78, 80). The likelihood of rearrangement of a V gene to D6/J6 versus J~ gene segments might be influenced by its genomic location. In line with the observation that J. proximal V, gene segments rearrange preferentially to JH in pre-B cells (79), it is possible that 6 gene rearrangements, which occur earlier in ontogeny than ~ gene rearrangements (see later), might preferentially use D6-Jfi proximal V gene segments. Indeed, V~6, a "singlememberedgene family" probably identical to V63, is closer to J6 than any

Annual Reviews 18 Murine

8 RAULET V~ Sequences


+I VSI V82 V~3 V~4 V~5

(V~6)

V~6

(V~7)

V~7 V~8

(Vd4) (V~l i)

Human V61

i0

20

30

40

5

DVYLEPVTkKTFTVVAGDPASFYCTV TGGDMKNY HMSWYKKNGTNAL ~ TQMLHQSPQSLTIQEGDEVTMSCNLST SL Y ¢ ALLWYRQGDDGSLV. QPDSMEST-EGETVHLPCSHATISGNE Y IYWYRQVPLQGPE~ AQTVSQPQKKKSVQVAESATLDCTYDT SDTN Y LLFWYKQQGGQ Vl C I TLTQSS TDQTVASGTEATLLCTYNADSPN P DLFWYRKRPDRSFQ~ AQRVTQVQPTGSSQWGEEVTLDCSYET SEYF Y CIIWYRQLFSGEMV~ AQKVTQVQSTGSSQWGE VTLHCSYET SEYF Y VILWYKQLFSGEMVI AQRVTQVQPTASSQWGEEVTLDCSYET SEYF Y RIFWYRQLFSGEMV~ AQKVIQVWSTTSRQEGEKLTLDCSYKT SQVL Y HLFWYKHLLSGEMV] QVALSEEDFL TIHCNY SASG YPTLFWYVQYPGEGPQJ ~ RG.

V~ Sequence QKVTQAQSSVSMPVRKAVTLNCLYET

SWWS

Y YIFWYKQLPSKEMI}

Figure 5 Murine and human V6 amino acid sequences. The nomenclaturefor murine V6 genes is an extension of that of Elliott et al (77). Whereapplicable, the corresponding gene subfamily is indicated. In parentheses to the right of each sequence are references for the sequences. V~7is from the DN2.3cell line (69) and V~8is the 3G8sequence in reference 65. Gaps were inserted to align the sequences for comparison, and the conserved cysteine residues are indicated by asterisks. The amino-termini of the mature proteins are estimated.

of 38 other V~ genes tested in a deletion mappingstudy (M. Kronenberg, personal communication). Since V62 is reportedly located on the same EcoRI fragment as V63(77), it is also proximal to D6/J6. In addition, the V35gene is in close proximity to D6/J3 gene segments, though it is located to the 3’ side of C6 (see below). Finally, a study of the extent of V6 gene deletion in thymocytes led to the conclusion that V61~are all relatively proximal to D6-J6 (77). Therefore, segeral of the V6 genes that are.frequently rearranged to D6-J6 in various tissues (V61, V63, V65) are proximal to D6-J6 gene segments in germline DNA,suggesting that proximity mayaccount at least partly for the preferential usage of 6 gene segments. It is unlikely, however,that proximity is a necessary distinguishing feature of V6 genes, since V66 family members are not commonlydeleted by ~ rearrangements and are therefore probably located to the 5’ side of many V~ gene segments (77). Interestingly, the V65geneis located ~ 2.5 kb to the 3’ side of C6 (Figure 1), in the opposite transcriptional orientation, and rearranges by inversion to D6/J6 gene segments (80). Despite its proximity to D6/J6 gene segments, V65is infrequently rearranged in ~/6 cells from the fetal thymus(76, 77); instead V~5-D~J~C~ rearrangements are frequent amongadult thymic y/6 cells (69, 77), accounting for 16-30%of rearranged alleles. Therefore,



60

70

80

90

i00

LVXKLNSNSTDGGKSNLKGKINISKN LVTLQKGGDEK SKDKITAKLD VTHGLQQNTT NSMAFLAIASDR LVILQEAYKQYNATLNRFS%rNFQ ILYRDDTSSHDADFVQGRFSVKHS LIY QTSFDTQNQRNGRYSV~Q LIY QTSFDTQNQRNSRYSVVFQ LIY QPSFDTQNQRSGRYSV-~/FQ LIR QMPSTIAIERSGRYSVVFQ LFR ASKDKEKGSSRGFEATYD LISLFYLASGTKENGRLKSAFDS

QFI LDIQKATMKDAGTYYCGSD KKMQQSSLQIQASQPSHSGTYLCGGK KS ST LILTHVSLRDAAVYHCILR KAAKSFSLEISDSQLGDAATYFCALM KANRTFHLVISPVSLEDSATYYCASG KSLKSISLVISASQPEDSGTYFCALS KSLKSISLVISASQPEDSGTYFCALS KSFKSISLVISASQPEDSGTYFCALS KSRKSISLVISTLQPDDSGKYFCALW KGTTSFHLKKASVQESDSAVYYCALS KERRYSTLHIRDAQLEDSGTYFCAAD

LIK QGSDEQNA.K

KAAKSVALTISALQLEDSAKYFCAL

SGRYSVNFK

189 Ref. (76) (76) (77) (51) (77) (77) (77) (51) (76) (69) (65)

(55,56)

Figure 5 (Cont#~ued).

least in this case, proximity alone does not lead to a high frequency of rearrangement in fetal thymocytes. Human Vc5 Genes The human6 locus, located on chromosome14, is also located between J~ and V~ gene segments (81, 82). Also similar to the murine 6 locus, three J6 and two D6 gene segments are located 5’ to C6 in germline DNA (Figure 1). As in the murine system, the humanV6 gene repertoire appears to be limited, and at least somewhatdistinct from the Vc~gene repertoire (55, 56). Only a single humanV6 sequence, V61, has been reported (Figure 5), which displays 57%amino acid identity with a humanV~ gene segment, and up to 58% amino acid identity with murine V66/V~7members. V61 is expressed by all five of the 7/6-bearing T-cell lines and leukemias examined in several reports, suggesting that it is a commonlyutilized V6 gene segment (55, 56, 78). Recently three additional humanV6 genes have been isolated (M. Krangel, personal communication). One of these genes encodes a V6 which is similar to the murine V65region, displaying 66% aminoacid identity. Interestingly, this V segmentis located to the 3’ side of the C~gene, as is murine V~5. C6 Gene In mice and humansthere is a single C6 gene (51, 81, 82). In the mouse, C6 is encoded by four exons located 75 kb 5’ to the Ca gene and about 8 kb 5’ of the most proximal J~ gene segment, J~l (76).

Annual Reviews


190

RAULET

The deduced amino acid sequence of C6 includes cysteine residues separated by 51 amino acids that presumably form an intrachain disulfide linkage, followed by a connector region that carries a cysteine residue presumedto form a disulfide linkage with the ~ chain (Figure 3; 51, 55, 56). The connector region is followed by a transmembrane domain which includes a charged lysinc residue conserved in all TCRC regions at corresponding positions. The lysine residue is thought to be involved in interactions with CD3components, which carry acidic residues in their transmembrane regions. In addition, the C~ and Ca transmembrane regions each carry a second basic residue (arginine) in the transmembrane region, which may also be important for interactions with other chains. Also like Ca, the C6 intracytoplasmic tail is apparently extremely short ( _< 4 aminoacids). J~ Genes TwoJ6 gene segments, J61 and J62, have been localized 18.5 and 5 kb upstream of the murine C6 gene (69, 76). The deduced Jc~ region sequences match at only 7 of 16 residues (Figure 4). Three humanJ~ gene segments have been identified, two of which are nearly identical (Figure 4; 55, 56, 78, 82, 82a). The murine J6 sequences are very similar to the human counterparts, matchingat 13 of 16 (J61) and 18 of 19 (J62) residues. contrasts with the J7 regions, for which there is no correspondencebetween mouse and human. The J6 gene segments are flanked on their 5’ sides by heptamer-nonamerrecombination signal sequences with 13 bp spacers (69, 76, 78). D6 Genes Two D6 genes, D6I and D62, have been localized approximately 10 kb and 1.2 kb upstream of J61 in both mouse and human(76, 82a). Murine D61and D62are 11 and 16 nucleotides long, respectively, and can be read in all three reading frames. They are each flanked on both sides with heptamer-nonamerrecombination signal sequences; the 5’ signal sequence includes a 12 nucleotide spacer, while the 3’ signal sequence carries a 23 nucleotide spacer. Rearrangement

of ~ Genes

The arrangement of recombination signal sequences flanking 6 gene segments is compatible with V6-J6, V6-D6-J6, and V6-D61-D62-J6rearrangements. The latter type of rearrangement, which is unprecedented in other rearranging gene families, is commonamongadult thymic 6 rearrangements (77). V6-D62-J6 rearrangements are frequent among fetal rearrangements (76). Direct V6-J6 rearrangements may also occur,


191

evidenced by analysis of excised circular DNAfrom thymocytes (83). Interestingly, incomplete V6-D6rearrangements are fairly commonin hybridomasof fetal thymocytes (76). In contrast, V-Drearrangements are rarely if ever observed in heavy chain or TCR-/3gene rearrangements.


DIVERSITY OF THE 7/c~

RECEPTOR

With relatively few V genes (7-8), no D regions, and few distinct J regions (2M), 7 genes are one of the least diverse rearranging receptor gene families. The diversity of the ~ genes is also likely to be less than that of e and/~ gene families with respect to the number of V (~ 10), J (2), and D segments. In addition, neither ? nor 6 chains appear to undergo somatic hypermutations, an important mechanism for diversifying immunoglobulin V regions following antigenic stimulation (74, 76, 77, 80). Although the number of rearranging gene segments is small in both families, there is potential for considerable junctional diversity in both chains. In fact, the junctional diversity of the c5 chain is potentially enormous (77, 78). Unlike other receptor gene rearrangements, assembled genes in the adult thymus often include two D segments (V6-D61-D62J6 rearrangements). Imprecise joining can result in variable inclusion of nucleotides located at all six DNAends involved in these rearrangements, and N regions can be added at all three junctions. Therefore, although the V6 repertoire is relatively limited, the potential diversity of 6 chains may be far greater than that of/3 and c~ chains. Junctional diversity of ? chains due to imprecise joining and N region addition has also been documented (25, 36, 73, 74), although there are no D7 gene segments. Based on these considerations, it has been estimated that there could be 1017 possible ?/6 heterodimers, with almost all the diversity concentrated at the V-J junctions of both ? and 6 chains (77). As discussed below, however, ? and chain diversity maybe muchmore restricted in certain anatomical sites, such as the epidermis.

PERIPHERAL LOCATIONS OF 7/6

CELLS

In the periphery, 0.5-10%of peripheral blood T cells in the human(11, 38, 49, 84) and ~ 3%ofT cells in the murine spleen and lymph nodes (44, 85, 86) are CD4-CD8 ?/6 cells. In the chicken, the proportion of?/6 cells in peripheral blood and spleen is higher than in mammals,approaching 30% of CD3+ T cells (87, 88). Interestingly, most (~70%) chicken ?/3 + cells (88); the remainder are CD4-CD8cells in the spleen are CD4-CD8 cells, as are the ?/3 cells in chickenblood. Strikingly, ?/3 cells are the predominantT-cell type found in the murine

Annual Reviews


192

RAULET

epidermis and the murine and chicken intestinal epithelium. Murine Thyl + dendritic epidermal cells (DEC),the subject of extensive investigations for several years as candidate epidermal T cells, were shownrecently to be largely if not exclusively 7/6 cells (58-60, 89). 7/6-epidermal cells have not been detected in the chicken, and studies in the humanare thus far inconclusive. In the routine and chicken intestinal epithelium, 7/6 cells are also the predominant T-cell type. In both species, most intestinal + cells (61-63). The epithelial intraepithelial 7/6 cells are CD4-CD8 location of some?/6 cells has promptedthe suggestion that they function to combatepidermal and/or intestinal antigens. The findings that 7/6 cells in different peripheral locations exhibit striking differences in V7 and V6 gene usage raise the possibility that 7/6 cells in various locations carry out discrete functions (see below). ONTOGENY

OF

7/6

CELLS

7/6-bearing cells are the first TCRbearing cells to appear in ontogeny, on or before day 14 of gestation in the murine fetal thymus (43, 44, 63, 91). Six to seven percent of day 14 fetal thymocytes, bear 7/6 receptors (92). At this early stage all thymocytes are CD4CD8-, and the proportion of CD4-CD8thymocytes which bear 7/6 receptor cells, 5-10%, is maintained throughout ontogeny (4.2-44, 92); few if any CD4+ + and/or CD8 thymocytes bear 7/6 receptors. Since the absolute number of CD4-CD8thymocytes increases during ontogeny, so does the numberof 7/6-bearing thymocytes, reaching approximately 3 x 10s cells in the young adult thymus. The proportion of 7/6-bearing thymocytes, however, decreases precipitously after day 16 of gestation (to 0.2%of adult thymocytes) + thymocytes. In the parallel with the rapid increase in CD4+ and/or CD8 chicken a similar pattern of 7/6 cell ontogeny has been documented(87). 7/6-bearing splenic T cells first appear sometimebetweenbirth and four weeksof age (93). Thyl + dendritic epidermal cells, most of which are 7/6 cells, also appear between birth and three weeks of age (R. Tigelaar, personal communication).The ontogeny of 7/6 cells of the intestinal epithelium has not been reported. Thymus Dependence

of ~/~ Cells

The early appearance of 7/c~ cells in the thymus, followed later by their appearance in the spleen and epidermis, suggests that splenic and intraepithelial 7/6 cells may be thymus-derived. Direct evidence that many splenic 7/6 cells are thymus-dependentcomes from the demonstration that they are severely deficient in athymic nude mice until at least 11 weeksof age; furthermore, thymus grafting of nude mice results in near normal

Annual Reviews


GAMMA AND DELTA TCR

193

numbersof splenic ~/6 cells several weeks later (93). In the same study, however, it was shown that old (~ 24 weeks) ungrafted nude mice have detectable numbersof splenic 7/6 cells. A similar situation pertains to 7/fl T cells, which are present in small numbersin the spleens of old but not young nude mice (94). Whetherintraepithelial 7/6 cells are thymus-dependenthas been difficult to evaluate. Thyl + dendritic epidermal cells are present in reduced numbers in the skin of eight weekold nude mice, but surprisingly the cells are CD3- and are devoid of full-length TCRtranscripts (95). The obvious conclusion that epidermal 7/6 cells are thymus-dependent is confounded by the fact that nude mice mayexhibit a general epithelial defect. In fact, the Thyl + dendritic epidermal cells of athymic mice type as CD3-even after thymus-grafting (G. Stingl, unpublished data). While these results are consistent with the possibility that (some) 7/6 cells mature in the epidermis and that this process is deficient in nude skin, other explanations are equally plausible. For example, homingof.thymus-derived 7/6 cells to nude skin may be deficient. Another possible explanation is suggested by the observation that the Vy and V6 chains expressed by manyepidermal ~,/6 cells are commonlyexpressed by fetal but not adult ~/6 thymocytes (92, see below): it is possible that epidermal~/6 cells derive from an early wave of fetal thymic ~/6 cells which are not generated in the thymusgrafting experiments performed to date. While extensive additional research will be necessary to resolve this issue, the overall data suggest that manysplenic ~/6 cells, at least, are thymus-dependent. Programmed V Expression

of T/6 Cells

in the Thymus

Emergingdata indicate that during ontogeny, thymocytesexhibit a striking programmedpattern of V7 and V6 gene utilization. Most day 14-day 16 fetal thymocytes express a V73-JTIC71chain, as shownwith a monoclonal anti-V73 antibody and by Southern hybridization studies (30, 31, 92, 96). V6 usage maybe more heterogeneous at these times, but V61 is commonly expressed (76). By day 18 of gestation, V~/3-expressing thymocytes are undetectable (92, 96). Aroundday 16 of gestation, "waves"of cells appear that express distinct ~ chains, including V72-JTICT1, C74-containing chains, and possibly V74-JT1CyI chains (92, 96). In the young adult thymus, the large majority ofCD4-CD8-thymocytes express V72-J71C7I gammachains, although a minority express V74-JylCTI and C74-containing chains (Figure 6; 42, 69, 96). V6 usage by adult CD4CD8 thymocytes appears also to be largely nonoverlapping with that of fetal thymocytes, and predominantly limited to a few V6 genes, including V65 (69, 77). ProgrammedV gene usage has also been documented in pre-B cells,

Annual Reviews

194

RAULET


where the predominantly utilized VH-genes are proximal to the J,-gene segments (79). Similarly, common fetal Vy and Vd gene segments (Vy3, V6 1, and V63) are all relatively J-proximal (Figure I , see Murine V6 Genes, above). Models to account for programmed V-gene usage include: (a) initial preferential rearrangement of J-proximal V gene segments, with subsequent selection for cells expressing other, more distally located V gene segments (79); (h) initial preferential rearrangement of J-proximal V segments, and later replacement of the rearranged V-segments with other V segments by secondary rearrangement events (30,97,98) (two examples

Annual Reviews


GAMMA AND DELTA TCR

195

of such secondary replacement rearrangements in B cell lymphomashave recently been reported; 97, 98) and (c) sequential activation in progenitor cells of distinct subsets of ¥ genes for rearrangement. It is not clear which if any of the above mechanisms operate to program V7 and V6 usage in thymic ontogeny. In addition to distinct V gene usage, fetal and adult thymocytesdiffer in their ~ junctional sequences. Thus most fetal ~ rearrangements involve V6-D62-J6 rearrangements (76), whereas in adult CD4-CD8-thymocytes, V&D61-D62-J61rearrangements are the rule (77). Moreover, N region diversity is extensive in adult but not fetal 6 rearrangements (76, 77). LINEAGE RELATIONSHIP cz/fi CELLS

OF y/6

CELLS

AND

The weight of evidence at present, though indirect, argues persuasively that 7/~ receptor bearing cells do not differentiate to c~/fl receptor bearing cells. First, surveys of cloned e/fl-T cell lines have revealed that manylack functional rearrangements of all known7 genes (21, 99). A comparable examination of 6 rearrangements in most mature e/fi T cells is impossible since 6 genes are deleted from chromosomal DNAby Ve-Je rearrangements (51). However,the DNAdeleted by Ve-J~ rearrangements is present in thymocytes as extrachromosomal circular DNAmolecules, which can be isolated and molecularly cloned. Of more than 400 DNAclones of this type examined, all contained D62 and J61 gene segments in germline

Figure 6 Evidence that most adult thymic 7/6 cells express a ~ chain encoded by V72-JylC71 rearranged genes. 7/6 cells were enriched (to ~50-80%~,/~ cells) from C57BL/6doublenegative thymocytepreparations by short-term (~ 3d) culture in a lymphokinecocktail (42), and DNAsamples were analyzed by Southern hybridization with the indicated probes. The Cy2 probe detects C71, C72 and C~/3. Panels ADare EcoRI digests and panel E is a HINDIII digest. Lane 1 is DNAfrom 7/6 cells and lane 2 is C57BL/6liver DNA.V72-J~/1Cyl(-~Ir in panel B) and Vy1.2-Jy2C~/2(top ~ir in panel A) rearrangementsare frequent in the population. V74-JylC71(’A" in panel C) and V~I. 1-Jy4Cy4(’A" in panel E) rearrangements are detectable on the original autoradiograph, though infrequent. Vy3rearrangements are undetectable. The large majority of 7 chains in the population bind to antiserum reactive with C,/1 and C72 and are glycosylated on asparagine residues (42). Since the V,/1.2-J72C~/2 sequence lacks N-glycosylationsites, it follows that most of the 7/6 cells express a V72-J ~ 1Cy1 chain. Smaller proportions of these cells express Cy4-containing7 chains (96) and V74-J~/ICTIchains (69). Most of the Vy1.2-J~2C72rearrangements appear not to encode a y chain expressed on the cell surface. Experiment performed by D. Pardoll, A. Kruisbeek, B. J. Fowlkes, and D. Raulet.

Annual Reviews


196 RAULI~T configuration, arguing that all or most V~-Ju rearrangements occur on chromosomes that have not previously rearranged 6 genes (A. Winoto, D. Baltimore, unpublished data). Therefore, it appears that neither rearrangements nor productive ,/ rearrangements have occurred in the precursors of most ~//3-T cells. Most likely e/// T cells and ;~/6 T cells represent distinct T-cell lineages. If 7/6 cells and e//~ cells represent separate lineages, at what stage in Tcell differentiation do they separate? Developmentalstudies suggest that in early fetal thymocytes complete rearrangements of 7 and 6 loci and partial D~-J/~ rearrangements occur nearly synchronously in the same cells (76, 100). Rearrangementsof V/~ to D/%J/~are delayed by one or two days, and ~ rearrangements are believed to occur later yet (16, 17, 100, 101). Based on these and other considerations, it was proposed that progenitor thymocytesthat initially fail to produce a functional ~/6 receptor go on to attempt complete ~ and/~ rearrangements (44). This model accounts for the findings that all c¢//~ T cells have ~ rearrangements, manyof which are nonproductive. However, the model also predicts that cells undergoing u-gene rearrangements have previously attempted 6 rearrangements, inconsistent with the aforementioned demonstration that c¢ rearrangements generally occur on chromosomes which have unrearranged Dy2 and JT1 genes. In fact, the latter finding suggests that the cells that rearrange 6 genes early cannot differentiate into c¢//~ cells. Perhaps the progenitors of ~//3 cells are a separate subset that are committed to rearrange ~ genes rather than 6 genes. Alternatively, 6 rearrangements, functional or not, may somehowprevent subsequent e rearrangements in the same cell. In either case, it appears that differentiation ofu//? cells maynot be contingent upon failed attempts to produce a 7/6 receptor. A possible mechanism whereby ~ rearrangements may prevent subsequent c¢ rearrangements on the same chromosome has been recently proposed (102, 103). A novel rearrangement event often deletes the D6, J6, and C6 gene segments in human thymocytes. The rearrangement involves heptamer-nonamer recombination signal sequences located upstream and downstreamof the D6, J6, and C6 genes. The 3’ signal sequence flanks a J-like sequence that is between C&and all known Ju gene segments. However, the upstream signal sequence, which is located 3’ of the known V~ and V6 genes, is not associated with a recognizable gene segment. It was proposed that this rearrangementeven~, is a prerequisite for subsequent c¢ gene rearrangement. Rearrangements of.V6 to D6 should delete the upstream signal sequence, thereby preventing-the deletion of the 6 locus and subsequent c¢ gene rearrangement. Further studies will be necessary to determine whether this deletion event actually plays a role in regulating e gene rearrangements.



Vy AND Vfi y/fi CELLS

197

GENE USAGE IN PERIPHERAL

One of the most striking emerging phenomenaassociated with 7/6 cells is the selective expression of different 7 and 6 V genes in different tissues. As summarizedin Table 2, most epidermal 7/6 cells in the mouse express a receptor composed of a V73-J71C71chain and a V61-D0-J62C6chain (60, 89; D. Asarnow,J. Allison, unpublisheddata). In contrast, 7/6 cells of the intestinal epithelium commonlyexpress a Vy5-J7ICTIchain (62, 104). chains from these cells have not yet been analyzed. In the spleen, most 7/6 cells express V72-JT1C~,Ior J74C74-containingchains (86, 93). Interestingly, allogeneic mixedlymphocytereactions of splenic 7/6 cells lead to an enrichment of 7/6 cells expressing VT1.2-J~2C72 chains (85). If intraepithelial 7/3 cells differentiate in situ, their selective V7and V6 gene usage maybe determined by specific microenvironmentalsignals that influence gene rearrangements or exert a selection for cells with particular receptor structures. If intraepithelial 7/3 cells are instead thymus-derived, differential homingof particular 7/6 cells to different sites maybe operative. Thus, it has been proposedthat the earliest thymic7/6 cells, which express V73-J7ICT1chains, are programmedto hometo the epidermis (92). A later thymic "wave"of ~,/6 cells expressing V75-JT1C~lmayhomespecifically to intestinal epithelium. Most splenic 7/8 cells maybe derived from a later waveofthymic ~/~i cells, which commonlyexpress V~2-J~1C~1 and J74C~4containing chains. Selective expressionof different V segmentsin different peripheral tissues has not been observed for e//%T cells and maysuggest that the different 7//3 receptors are specialized to recognizedistinct and limited sets of ligands at these sites. It is interesting that even the junctional diversity of ~/~ receptors expressed in someperipheral sites maybe highly restricted. A striking example comes from an examination of the sequences of the expressed 6 and 7 genes of several independent Thyl + DECcloned lines (D. Asarnow, J. Allison, unpublished results). Each expresses V61-D62J62C6and V~3-J~1 C71 rearranged genes, and the junctional sequences are virtually identical. Thus, no N region nucleotides are present, and the rearrangement breakpoints are identical. The lack of junctional diversity of ~ and 6 chains expressed by manyepidermal ~,/6 cells fits with the notion that they are derived from early thymic 7/3 cells, whicharise before terminal transferase is detectable (~ day 17). In any case, the observation that many Thyl+ DECdisplay virtually no TCRdiversity suggests that the putative epidermal ligand(s) of Thyl + DECare of extremely limited diversity. A caveat is that several 7/6 Thyl + DEClines analyzed in an independent study express other V3, and V6 genes (57-59, 59a; J. Coligan,

Annual Reviews 198

RAULET

personal communication).Further studies will be necessary to resolve this discrepancy.


SPECIFICITY OF 7/6 CELLS Exposition of the ligands of 3’/6 receptors is crucial for understanding the physiological role of 7/6 cells. A particularly burning issue has been whether 7/6 receptors, like ~//~ receptors, recognize MHC-associatedantigens. Several early reports demonstrated that human7/6 cell lines cause lysis of a variety of tumor target cells regardless of their MHC-antigen expression (39, 40, 45-49). A similar pattern oflysis is displayed by natural killer cells and by MHC-restrictedT-cell clones that have been cultured in high concentrations of lymphokines(105). A critical issue is whether the observed lysis by 7/6 cells involves the CD3-associated ~/6 receptor. Although some studies suggested that anti-CD3 antibodies block such target cell lysis (39, 45, 46, 48, 106), more recent studies with F(ab’)z fragments of anti-CD3 antibodies yielded the opposite conclusion (107). The blockade with intact antibody may therefore be an indirect effect involving the Fc portion of the antibodies. It appears likely that MHCindependenttumor cell lysis by these 7/(5 cell lines worksindependently of their 7/6-TCR. Morerecently, 7/6 cytotoxic cell lines reactive with allogeneic MHClinked antigens were derived from aged nude mice (64, 65; L. Matis, J. Bluestone, unpublished data). Several lines have been established; each recognizes distinct MHC-relatedantigens: a class 1 type antigen encoded in the Tla region, a conventional MHC-class1 antigen (Dk), and a conventional class 2 antigen (Ek). These results are in accord with earlier findings demonstrating enrichment of ~/6 cells in allogeneic mixedlymphocyte cultures (85). The reproducible ability to produce MHC-reactive?,/6 cell lines suggests that 7/8 receptors, like ~/3 receptors, maybe predisposed to recognize MHC-relatedproteins. Amongc~///-T cells, alloreactivity is characteristic of T cells that also react with foreign antigens associated with self-MHCproteins (108, 109). However, as of this writing no selfMHC-restricted ~,/6 cells specific for non-MHCantigens have been identified.

FUNCTIONAL ACTIVITIES

OF 7/6

CELLS

7/6 cell lines exhibit various functional capabilities similar to those of e//~ T cells. As discussed in the previous section, a variety of both humanand murine7/6 cell lines exhibit cytolytic activity, whichin somecases is specific

Annual Reviews


GAMMA AND DELTA TCR

199

for cells bearing a11ogeneic MHC proteins. Cytolytic 7/6 cell lines have been isolated from humanthymus and peripheral blood (37, 39, 40, 4549), nude mousespleen (64, 65), the murine Thyl + DECpopulation (110; W.Havran,J. Allison, unpublisheddata), and the intestinal intraepithelial T cell population (61). Interestingly, freshly isolated intestinal intraepithelial 7/6 cells are cytolytic without induction, suggesting that they maybe activated in situ (61). Production of T cell-derived lymphokinesby a variety of 7/6 cell lines has also been documented. IL-2 production in response to activation with mitogens or anti-T3 antibodies was reported for humanand murine 7/6 cell lines (37, 47, 65), murineThyl + DEClines (110; W.Havran,J. Allison, unpublished data), and T-hybridomas derived from adult murine thymic 7/6 cells (111). In the latter category, one exampleof a hybridomaproducing both IL-2 and ILo4 was documented. Furthermore, early fetal thymic y/6 cells may produce both IL-2 and IL-4 (112). An alloreactive murine 7/6 CTLline produces both GM-CSFand y-interferon (65). Although lymphokineproduction is often considered primarily a function of noncytolytic helper T cells, manyof the 7/6 cell lines examinedexhibit both cytolytic activity and lymphokineproduction. Therefore, the present data do not allow categorization of 7/6 cells into cytotoxic and helper subsets. In sum, 7/6 cell lines exhibit at least someof the typical effector functions described for e//~ T cells, suggesting that they may represent a mature subset of T cells, with a similar or overlapping range of functional activities.

OVERVIEW AND SPECULATIONS Muchearly work on y/6 cells focused on the possibility that these cells represented an immaturestage of T-cell differentiation. However,the evidence cited herein suggests that precursors of ~///cells need not express functional y or 6 chains. Moreover, 7/6 cells exhibit receptor-triggered functional activities similar to those of mature e//3-cells. Despite their "immature" phenotype (they are often CD4-CD8 cells), it is likely that ~/6 cells represent a mature functional lineage of lymphocytes. Attempts to propose a unified model of y/6-cell function are frustrated by the paradoxical properties of these cells. On the one hand, the 7/6 receptor displays enormouspotential junctional diversity, consistent with a role in antigen-recognition akin to the ~//3 receptor. On the other hand, 7/6 cells resident in specific peripheral locations, such as the epidermis, maydisplay highly restricted V gene diversity with little or no junctional

Annual Reviews 200

RAULET

diversity. In these locations the diversity of putative ligands recognizedby 7/6 cells are presumablyhighly restricted. Modelsof?/6 cell function based on each of these points of view will be discussed separately below, followed by an attempt to reconcile them.


A Diverse

~/~ Receptor

As we have seen, 7/6 cells reactive with allogeneic MHCantigens can be demonstrated, raising the possibility that like the ~/fl receptor, the 7/3 receptor has a predilection for binding MHCproteins. MHCproteins function to bind antigenic peptides for recognition by the ~/[~ receptor, and self-MHC-restricted T cells are frequently reactive with allogeneicMHC proteins. It is tempting to suggest, therefore, that 7/3 cells also recognize antigenic peptides bound to MHC-proteins. The potential diversity of the 7/6-receptor is sufficient for a role in MHCrestricted recognition of diverse antigens. However,the concentration of this diversity in the third CDRregion contrasts sharply with the receptor. A recent model suggests that the third CDRof T-cell receptor chains is oriented over the bound peptide of an MHC-peptidecomplex, whereasthe first and second CDRmaybe positioned better for interactions with polymorphic MHCresidues flanking the peptide binding site (113). Extrapolated to the 7/6 receptor, the modelmaypredict a potentially large universe of antigenic peptides but a small universe of restricting elements. The finding of a 7/3 cell line reactive with a Tla-linked class 1 antigen (64, 65) has fueled speculation that the restricting elements of at least some 7/6 cells might include a subset of these MHC-likemolecules, which are relatively nonpolymorphic(69, 104). It must be emphasized however, that ~/6 cell lines that react with "conventional" MHCantigens (Dk k) and E have also been established (65; L. Matis, J. Bluestone, unpublishedresults). A Nondiverse

~/~ Receptor

7/6 cells in distinct anatomical sites express distinct V7 (and possible V6) genes. Moreover,junctional diversity of both 7 and 6 chains of epidermal 7/3 cells (at least) maybe extremely limited. These findings argue for predictable and limited set of ligands for 7/6 cells, at least in the skin. It is possible that most epidermal 7/6 cells recognize one or a few antigens (possibly associated with MHC-likeproteins) which are associated with commonskin pathogens. An alternative view is that the ligands for the receptors on intraepithelial 7/6 cells are unique se/f-antigens which are only expressed following environmental insults such as infections and/or transformation (104). In keeping with the observed MHC-reactivity some7/6 cell lines, the self-antigens maybe encoded by MHC-relatedclass 1 genes that mapin the Tla region and show tissue-specific expression.

Annual Reviews GAMMA AND DELTA

TCR

201


Reconciliation The diverse and nondiverse character of ~,/6 receptors in different anatomical sites constitutes a significant paradox which maybe resolved by recourse to an evolutionary perspective. Perhaps the ~/c5 receptor preceded the ~/fl receptor in evolution and functioned originally to survey tissues for a limited array of predictable antigens (see above). Later, evolution may have superimposed a broader, MHC-restricted antigen recognition capability on the ~/~ receptor, akin to a primitive ~//~ receptor system. Thus fetal thymic ~/fi cells, which may be restricted in diversity, may function in the former role after migration to epithelia. Later wavesof ~/fi cells mayfunction in the latter role, after migration to the secondary lymphoid organs. The two functions may thus be separated phylogenetically as well as ontogenetically. One can imagine that the pressure for yet greater diversity finally forced the separation of these two functions, resulting in the appearanceof the ~/fl receptor system. For the ~ genes the 6 genes were subserved and diversified by creation of manyJ~ gene segments and diversification of V6 genes. The fl genes mayhave evolved from ~ genes following duplication and transposition to another chromosome. The views expressed above are simply a speculative attempt to distill the patchy available information into a unified modelof y/6 cell function and evolution. While the progress to date in describing ~/~i cells and receptors is impressive, it falls short of providing confidencein any modelof 7/6 cell function. Considering our inexperience in determining functional pathways starting with a gene, this is perhaps not surprising. Wehope the experience will provide approaches and perspectives that will facilitate future such endeavors. ACKNOWLEDGMENTS

I thank the many investigators who shared with me their unpublished data; Herman Eisen, Lisa Steiner, and Alan Kormanfor reviewing the manuscript; and Chris Greco for preparing the manuscript. This work was supported by research grant CA28900from the NIH and a Cancer Research Institute/Frances L. and Edwin L. CummingsMemorial Fund Investigator Award. Literature Cited 1. Allison, J. P., Lanier, L. L. 1987. The structure, function and serology of the T cell antigen receptor complex. Ann. Rev. Immunol. 5:503-40 2. Marrack, P., Kappler, J. 1986. The antigen-specific major histocompati-

bility complex-restricted receptor on T cells. Adv. lmmunol. 38:1-30 3. Davis, M. M. 1985. Molecular genetics of T-cell receptor beta chain. Ann. Rev. Immunol. 3:537~50 4. Toyonaga, B., Mak, T. W. 1987. Genes

Annual Reviews


202

RAULET

of the T cell antigen receptor in normal and malignant T cells. Ann. Rev. Immunol. 5:585 620 5. Swain, S. L. 1983. T cell subsets and the recognition of MHCclass, lmmunol. Rev. 74:129 42 6. Dembic, Z., Haas, W., Weiss, S., McCubrey, J., Kiefer, H., yon Boehmer, H., Steinmetz, M. 1986. Transfer of specificity by murine ~ and/~ Tcell receptor genes. Nature 320:232 7. Saito, T., Weiss, A., Miller, J., Norcross, M. A., Germain, R. N. 1987. Specific antigen-Ia activation of transfected humanT cells expressing murine Ti ~/3 humanT3 receptor conrplexes. Nature 325:125-29 8. Kronenberg, M., Sui, G., Hood, L. E., Shastri, N. 1986. The molecular genetics of the T-cell antigen receptor and T-cell antigen recognition. Ann. Rev. Immunol. 4:529-91 9. Garman, R. D., Ko, J.-L., Vulpe, C. D., Raulet, D. H. 1986. T cell receptor variable region gene usage in T cell populations. Proc. Natl. Acad. Sci. USA 83:3987 91 10. Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A. C., Eisen, H. N., Tonegawa, S. 1984. Completeprimary structure of a heterodimeric T-cell receptor deduced from eDNAsequences. Nature 309:757-62 11. Brenner, M. B., McLean, J., Dialynas, D., Strominger, J., Smith, J. A., Owen, F. L., Seidman, J., Ip, S., Rosen, F., Krangel, M. 1986. Identification of a putative second T cell receptor. Nature 322:145-49 12. Hedrick, S. M., Cohen, D. I., Nielsen, E. A., Davis, M. M. 1984. Isolation of eDNAclones encoding T cell-specific membrane-associated proteins. Nalure 308:149-53 13. Yanagi, Y., Yoshikai, Y., Leggett, K., Clark, S. P., Aleksander, I., Mak,T. W. 1984. A human T cell-specific cDNA clone encodes a protein having extensive homology to immunoglobulin chains. Nature 308:145-49 14. Chien, Y.-H., Becker, D. M., Lindsten, T., Okamura,M., Cohen, D. I., Davis, M. M. 1984. A third type of murine Tcell receptor gene. Nature 312:31-35 15. Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A. C., Eisen, H. N., Tonegawa, S. 1984. A third rearranged and expressed gene in a clone of cytotoxic T lymphocytes. Nature 312:36~40 16. Raulet, D. H., Garman, R. D., Saito, H., Tonegawa, S. 1985. Developmental regulation of T-cell receptor gene expression. Nature 314:103-7

17. Snodgrass, H. R., Dembic, Z., Steinmetz, M., von Boehmer, H. 1985. Expression of T-cell antigen receptor genes during fetal development in the thymus. Nature 315:232 33 18. Pernis, B., Axel, R. 1985. A one and a half receptor modelfor MHC-restricted antigen recognition by T lymphocytes. Cell41:13-16 19. Heilig, J. S., Glimcher, L. H., Kranz, D. M., Clayton, L. K., Greenstein, J. L., Saito, H., Maxam,A. M., Burakoff, S. J., Eisen, H. N., Tonegawa,S. 1985. Expression of the T-cell-specific ~/gene is unnecessary in T cells recognizing class 11 MHCdeterminants. Nature 317:68-70 20. Rupp, F., Frech, G., Hengartner, H., Zinkernagel, R. M., Joho, R. 1986. No functional ?,-chain transcripts detected in an alloreactive cytotoxic T-cell clone. Nature 321:876 78 21. Reilly, E. B., Kranz, D. M., Tonegawa, S., Eisen, H. N. 1986. A functional 7 gene formed from known y-gene segments is not necessary for antigenspecific responses of murine cytotoxic T lymphocytes. Nature 321:878-80 22. lwamoto, A., Rupp, F., Ohashi, P. S., Walker, C. L., Pircher, H., Joho, R., Hengartner, H., Mak, T. W. 1986. T cell-specific y genes in C57BL/10mice. J. Exp. Med. 163:1203-12 23. Traunecker, A., Oliveri, F., Allen, N., Karjalainen, K. 1986. Normal T cell development is possible without "functional" ? chain genes. EMBOJ. 5:1589-93 24. Dialynas, D. P., Murre, C., Quertermous, T., Boss, J. M., Leiden, J. M., Seidman,J. G., Strominger, J. L. 1986. Cloning and sequence analysis of complementary DNA encoding an aberrantly rearranged humanT-cell 7 chain. Proe. Natl. Aead. Sei. USA83: 2619-23 25. Quertermous, T., Strauss, W., Murre, C., Dialynas, D. P., Strominger, J. L., Seidmau, J. G. 1986. HumanT-cell ~ genes contain N segments and have marked junctional variability. Nature 322:184-87 26. Rabbitts, T. H., Lefranc, M.-P., Stinson, M. A., Sims, J. E., Schroder, J., Steinmetz, M., Spurr, N. L., Solomon, E., Goodfellow, P. N. 1985. The chromosomallocation of T-cell receptor genes and a T cell rearranging gene: Possible correlation with specific translocations in humanT cell leukaemia. EMBOJ, 4:1461-65 27. Lefranc, M. P., Rabbitts, T. H. 1985. Two tandemly organized human genes

Annual Reviews


GAMMA AND DELTA TCR encoding the T-cell 7 constant-region sequences show multiple rearrangements in different T-cell types. Nature 316:464-66 28. Murre, C., Waldman, R. A., Morton, C. C., Bongiovanni, K. F., Waldmann, T. A., Shows, T. B., Seidman, J. G. 1985. Humany-chain genes are rearranged in leukaemic T cells and map to the short arm of chromosome 7. Nature 316:549-52 29. Hayday,A. C., Saito, H., Gillies, S. D., Kranz, D. M., Tanigawa, G., Eisen, H. N., Tonegawa, S. 1985. Structure, organization, and somatic rearrangement of T cell gammagenes. Cell 40: 259-69 30. Garman,R. 13., Doherty, P. J., Raulet, D. H. 1986. Diversity, rearrangement and expression ofmurine T cell gamma genes. Cell 45:733-42 31. Heilig, J. S., Tonegawa,S. 1986. Diversity of routine gammagenes and expression in fetal and adult T lymphocytes. Nature 322:836~40 32. Pelkonen, J., Traunecker, A., Karja7lainen, K. 1987. A new mouse TCRV gene that shows remarkable evolutionary conservation. EMBOJ. 6: 1941-44 33. Deleted in proof 34. Lefranc, M.-P., Forster, A., Baer, R., Stinson, M. A., Rabbitts, T. H. 1986. Diversity and rearrangement of the humanT cell rearranging 7 genes: Nine germ-line variable genes belonging to two subgroups. Cell 45:237~46 35. Forster, A., Huck, S., Ghanem, N., Lefranc, M.-P., Rabbitts, T. H. 1987. New subgroups in the human T cell rearranging Vygene locus. EMBO J. 6: 1945-50 36. Huck, S., Dariavach, P., Lefranc, M.-P. 1988. Variable region genes in the human T-cell rearranging gamma (TRG) locus: V-J junction and homology with the mouse genes. EMBOJ. 7:71956 37. Bank, I., DePinho, R. A., Brenner, M.B., Cassimeris, J., Alt, F. W., Chess, L. 1986. A functional T3 molecule associated with a novel heterodimer on the surface of immature human thymocytes. Nature 322:179-81 38. Weiss, A., Newton, M., Crommie, D. 1986. Expression of T3 in association with a molecule distinct from the Tcell antigen receptor heterodimer. Proc. Natl. Acad. Sci. USA 83:6998-7002 39. Moingeon,P., Ythier, A., Goubin, G., Faure, F., Nowill, A., Delmon, L., Rainaud, M., Forestier, F., Daffos, F., Bohuon, C., Hercend, T. 1986. A

203

unique T-cell receptor complex expressed on human fetal lymphocytes displaying nature-killer-like activity. Nature 323:638-40 40. Moingeon,P,, Jitsukawa, S., Faure, F., Troalen, F., Triebel, F., Graziani, M., Forestier, F., Bellet, D., Bohuon,C., Hercend, T. 1987. A y-chain complex forms a functional receptor on cloned humanlymphocyteswith natural killerlike activity. Nature 325:723 26 41. Littman, D. R., Newton, M., Crommie, D., Ang, S.-L., Seidman, J. G., Gettner, S. N., Weiss, A. 1987. Characterization of an expressed CD3-associated Ti ~/-chain reveals C? domain polymorphism. Nature 326:85-88 42. Lew, A. M., Pardoll, D. M., Maloy, W. L., Fowlkes, B. J., Kruisbeek, A., Cheng, S.-F., Germain, R. N., Bluestone, J. A., Schwartz, R. H., Coligan, .1. E. 1986. Characterization of T-cell receptor gammachain expression in a subset of murine thymocytes. Science 234:1401-5 43. Nakanishi, N., Maeda, K., Ito, K. I., Heller, M., Tonegawa,S. 1987. T~/protein is expressed on murine fetal thymocytes as a disulfide-linked heterodimer. Nature 325:72(~23 44. Pardoll, D. M., Fowlkes, B. J., Bluestone, J. A., Kruisbeek, A., Maloy, W. L., Coligan, J. E., Schwartz, R. H. 1987. Differential expression of two distinct T-cell receptors during thymocyte development. Nature 326:79-81 45. Borst, J., van de Griend, R. J., van Oostveen, J. W., Ang, S.-L., Melief, C. J., Seidman, J. G., Bolhuis, R. L. H. 1987. A T-cell receptor ?/CD3complex found on cloned functional lymphocytes. Nature 325:683 88 46. Brenner, M. B., McLean,J., Scheft, H., Riberdy, J., Ang, S.-L., Seidman,J. G., Devlin, P., Krangel, M. S. 1987. Two forms of the T-cell receptor gamma protein found on peripheral blood cytotoxic T lymphocytes. Nature 325: 689-94 47. Ferrini, S., Bottino, C., Biassoni, R., Poggi, A., Sekaly, R. P:, Moretta, L., Moretta, A. 1987. Characterization of +, CD4-, CD3 CD8- clones expressing the putative T cell receptor "; gene product. J. Exp. Med. 166:277-82 48. Alarcon, B., DeVries, J., Pettey, C., Boylston, A., Yssel, H., Terhorst, C., Spits, H. 1987. The T-cell receptor ~/ chain-CD3complex: Implication in the -cytotoxic activity of a CD3+CD4 CD8 human natural killer clone, Proc. Natl. Acad. Sci. USA84: 386165

Annual Reviews


204

RAULET

49. Lanier, L. L., Federspiel, N. A., Ruitenberg, J. J., Phillips, J. H., Allison, J. P., Littman, D., Weiss, A. 1987. The T cell antigen receptor complex expressed on normal peripheral blood CD4, CD8- T lymphocytes. J. Exp. Med. 165:1076-94 50. Lindsten, T., Fowlkes, B. J., Samelson, L. E., Davis, M. M., Chien, ¥.-H. 1987. Transient rearrangements of the T cell antigen receptor ~ locus in early thymocytes. J. Exp. Med. 166:761-75 51. Chien, Y., lwashima, M., Kaplan, K. B., Elliott, J. F., Davis, M. M.1987. A newT-cell receptor gene located within the alpha locus and expressed early in T-cell differentiation. Nature 327: 677 82 52. Band, H., Hochstenbach, F., McLean, J., Hata, S., Krangel, M. S., Brenner, M. B. 1987. hnmunochemical proof that a novel rearranging gene encodes the T cell receptor 6 subunit. Science 238:682-84 53. Born, W., Miles, C., White, J., O’Brien, R., Freed, J. H., Marrack, P., Kappler, J., Kubo, R. T. 1987. Peptide sequences of T cell receptor 3 and ~ chains are identical to predicted X and y proteins. Nature 330:573 74 54. Bonyhadi, M., Weiss, A., Tucker, P~ W, Tigelaar, R. E., Allison, J. P. 1987. Delta is the Cx-gene product in the ~/6 antigen receptor of dendritic epidermal cells. Nature 330:574-76 55. Loh, E. Y., Lanier, L. L., Turck, C. W., Littman, D. R., Davis, M. M., Chien, Y.-H., Weiss, A. 1987. Identification and sequence of a fourth humanT cell antigen receptor chain. Nature 330: 569-72 56. Hata, S., Brenner, M. B., Krangel, M. S. 1987. Identification of putative human T cell receptor 6 complementary DNAclones. Science 238:678 82 57. Stingl, G., Gunter, K. C., Tschachler, E., Yamada, H., Lechler, R. 1., Yokoyama, W. M., Steiner, G., Germain, + R. N., Shevach, E. M. 1987. Thy-I dendritic epidermal cells belong to the T-cell lineage. Proc. Natl. Acad. Sci. USA 84:2430-34 58. Stingl, G., Koning, F., Yamada, H., Yokoyama, W. M., Tschachler, E., Bluestone, J. A., Steiner, G., Samelson, L. E., Lew, A. M., Coligan, J. E., Shevach, E. M. 1987. Thy-I + dendritic epidermal cells express T3 antigen and the T-cell receptor y chain. Proc. Natl. Aead. Sei. USA 84:4586-90 59. Koning, F., Stingl, G., Yokoyama, W. M., Yamada, H., Maloy, W. L., Tschachler, E., Shevach, E. M., Coli-

gan, J. E. 1987. Identification of a T3associated ~/6 T cell receptor on Thy-I+ dendritic epidermal cell lines. Science 236:834-37 59a. Koning, F., Yokoyama, W. M., Maloy, W. L., Stingl, G., McConnell, T. J., Cohen, D. 1., Shevach, E. M., Coligan, J. E. 1985. Expression of Cy4 T cell receptors and lack of isotype exclusion by dendritic epidermal T cell lines. J. lmmunol.141: In press 60. Kuziel, W. A., Takashima, A., Bonyhadi, M., Bergstresser, P. R., Allison, J. P., Tigelaar, R. E., Tucker, P. W. 1987. Regulation of T-cell receptor ~,chain RNAexpression in murine Thy1+ dendritic epidermal cclls. Nature 328:263 66 6l. Goodman, T., Lefrancois, L. 1988. Expressionof the ~,/~ T-cell receptor on + intraepithelial lymphointestinal CD8 cytes. Nature 333:855-58 62. Bonneville, M., Janeway, C., Ito, K., Haser, W., Nakanishi, N., Tonegawa, S. 1988. Intestinal intraepithelial lym+ phocytes are a distinct set ofCD8 T cells. Submitted 63. Buoy, R., Chen, C., Cihak, J., Losch, U., Cooper, M. D. 1988. Avian T cells expressing ~/b receptors (TCR-1) home to the splenic sinusoids and the intestinal epithelium. J. lmmunol.In press 64. Matis, L. A., Cron, R., Bluestone, J. A. 1987. Major histocompatibility complex linked specificity of y/6 receptor bearing T lymphocytes. Nature 330: 263-64 65. Bluestone, J., Cron, R., Cotterman, M., Houtden, B., Matis, L. A. 1988. Structure and specificity of TCR receptors on major histocompatibility complex antigen specific CD3+, CD4-, CD8- T lymphocytes. J. Exp. Med. In press 66. Rathbun, G. A., Born, W., Kuziel, W. A., Tucker, P. W. 1988. Diversity of the mouseT cell receptor Cyl gene. Structural analysis in C57BL/Ka.lmmunogenetics 27:121 26 67. Kranz, D. M., Saito, H., Disteche, C. M., Swissheim, K., Pravtcheva, D., Ruddle, F. H., Ei~en, H. N., Tonegawa, S. 1985. Chromosomal locations of the murine T-cell receptor alpha-chain gene and the T-cell gammagene. Science 277:941-45 68. Woolf, T., Lai, E., Kronenberg, M., Hood, L. 1988. Mapping genomic organization by field inversion and two-dimensional gel electrophoresis: application to the murine T cell receptor 7 gene family. Nucleic Acid~ Res. 16:3863-75

Annual Reviews


GAMMA AND DELTA TCR 69. Korman, A., Marusic-Galesic, S., Spencer, D., Kruisbeek, A., Raulet, D. H. 1988. Predominant variable region gene usage by )’/6 T cell receptorbearing cells in the adult thymus. J. Exp. Med. 168: In press 70. Lefranc, M.-P., Forster, A., Rabbitts, T. H. 1986. Genetic polymorphism and exon changes of the constant regions of the humanT-cell rearranging gene 7Proc. Natl. Acad. Sci. USA 83: 95969600 71. Krangel, M. S., Band, H., Hata, S., McLean, J., Brenner, M. B. 1987. Structurally divergent human T cell receptor ~ proteins encodedby distinct C’f genes. Science 237:64-67 72. Lefranc, M.-P., Forster, A., Rabbitts, T. H. 1986. Rearrangement of two distinct T-cell y-chain variable-region genes in humanDNA.Nature 319: 42022 73. Quertermous, T., Strauss, W. M., van Dongen,J. J. M., Seidman,J. G. 1987. HumanT cell y chain joining regions and T cell development. J. lmmunol. 138:268740 73a. Huck, S., Lefranc, M.-P. 1987. Rearrangements to the JPI, JP and JP2 segments in the humanT cell rearranging gene (TRG)’) locus. FEBSLett. 224: 291-96 74. Kranz, D. M., Saito, H., Heller, M., Takagaki, Y., Haas, W., Eisen, H. N., Tonegawa, S. 1985. Limited diversity of the rearranged T-cell y gene. Nature 313:752-55 75. Alt, F. W., Baltimore, D. 1982. Joining of immunoglobulin heavy chain gene segments: Implications from a chromosome with evidence of three D-J, fusions. Proc. Natl. Acad. Sci. USA79: 4118-22 76. Chien, Y., lwashima, M., Wettstein, D. A., Kaplan, K. B., Elliott, J. F., Born, W., Davis, M. M. 1987. Tcell receptor 6 gene rearrangements in early thymocytes, Nature 330:722-27 77. Elliott, J. F., Rock, E. P., Patten, P. A., Davis, M. M., Chien, Y. 1988. The adult T-cell receptor 6 chain is diverse and distinct from that of fetal thymocytes. Nature 331:627-31 78. Hata, S., Satyanarayana, K., Devlin, P., Band, H., McLean,J., Strominger, J. L., Brenner, M. B., Krangel, M. S. 1988. Extensive junctional diversity of rearranged human T cell receptor 6 genes. Science 250:154144 79. Air, F. W., Blackwell, T. K., DePinho, R. A., Reth, M. G., Yancopoulos, G. D. 1986. Regulation of genomerearrangement events during lympho-

205

cyte differentiation. Immunol.Rev. 89: 5-30 80. Korman, A., Maruyama, J., Raulet, D. H. 1988. Rearrangement by inversion of a T cell receptor 6 variable region gene located 3’ of the 6 constant region gene. Submitted 81. Greisser, H., Champagne, E., Tkachuk, D., Takihara, Y., Lalande, M., Baille, E., Minden, M., Mak, T. 1988. The humanT cell receptor c~-6 locus: a physical map of the variable joining and constant region genes. Eur. J. lmmunol. 18:641-44 82. Satyanarayana, K., Hata, S., Devlin, P., Roncarlo, M., deVries, J., Spits, H., Strominger, J., Krangel, M. 1988. Genomic organization of the humanT cell receptor ctfl5 locus. Proc. Natl. Acad. Sci. USA. In press 82a. Takihara, Y., Tkachuk, D., Michalopoulos, E., Champagne, R., Reimann, J., Minden, M., Mak, T. W. 1988. Sequence and organization of the diversity, joining, and constant region genes of the humanT-cell-chain locus. PNAS 85:6097-6101 83. Okazaki, K., Sakano, H. 1988. Thymocyte circular DNAexcised from T cell receptor ~t-t5 gene complex. EMBO J. 7:1669-74 84. Lanier, L. L., Weiss, A. 1986. Presence of Ti (WT31)negative T lymphocytes in normal blood and thymus. Nature 324:268-70 85. Maeda, K., Nakanishi, N., Rogers, B. L., Haser, W. G., Shitara, K., Yoshida, H., Takagaki, Y., Augustin, A. A., Tonegawa, S. 1987. Expression of the T-cell receptor ),-chain gene products on the surface of peripheral T cells and T-cell blasts generatedby allogeneic mixedlymphocytereaction. Proc. Natl. Acad. Set. USA 84:6536-40 86. Cron, R., Koning, F., Maloy, W., Pardoll, D., Coligan, J., Bluestone, J. +, CD4-, 1988. Peripheral murine CD3 CD8- T lymphocytes express novel T cell receptor )’6 receptor structures. J. lmmunol. 141:1074-82 87. Sowder,J. T., Chen,C. H., Ager, L. L., Chan, M. M., Cooper, M. D. 1988. A large subpopulation of avian T cells express a homologueof the mammalian T’;t5 receptor. J. Exp. Med. 167: 31522 88. Chert, C. H., Cihak, J., Losch, U., Cooper, M. D. 1988. Differential expression of two T cell receptors, TCR1 and TCR2, on chicken lymphocytes. Eur. J. lmmunol. 18:539-43 89. Havran, W., Grell, S., Duwe, G., Kimura, J., Wilson, A., Kruisbeek, A.,

Annual Reviews


206

RAULET

O’Brien, R., Born, W., Tigelaar, R., Allison, J. 1988. Limited diversity of TCRy chain expression of murine Thyl + dendritic epidermal cells revealed by a V73 specific monoclonal antibody. Proc. Natl. Acad. Sci. USA. In press 90. Deleted in proof 91. Bluestone, J. A., Pardoll, D., Sharrow, S., Fowlkes, B. J. 1987. Characterization of murine thymocytes with CD3-associated T cell receptor structures. Nature 326:82-84 92. Havran, W., Allison, J. P. 1988. Developmentally ordered appearance of thymocytes expressing different T cell antigen receptors. Nature. In press 93. Pardoll, D., Fowlkes,B. J., Lew, A. M., Maloy, W., Weston, M., Bluestone, J. A., Schwartz, R. H., Coligan, J. E., Kruisbeek, A. M. 1988. Thymus-dependent and thymus-independent developmental pathways for peripheral T cell receptor y/a-bearing lymphocytes. J. Immunol. 140:4091-96 94. MacDonald,H. R., Lees, R. K., Bron, C., Sordat, B., Miescher,G. 1987. T cell antigen receptor expression in athymic (nu/nu) mice. J. Exp. Med. 166:195 209 95. Nixon-Fulton, J. L., Kuziel, W. A., Santerse, B., Bergstresser, P. R., Tucker, P. W., Tigelaar, R. E. 1988. + Thy-I epidermal cells in nude mice are distinct from their counterparts in thymus-bearing mice. A study of morphology,, function and T cell receptor expression. J. lmmunol. In press 96. Houlden, B. A., Cron, R. Q., Coligan, J. E., Bluestone, J. A. 1988. Systematic development of distinct TCR?aT cell subsets during fetal ontogeny. J. Immunol. In press 97. Reth, M., Gehrmann, P., Petrac, E., Wiese, P. 1986. A novel V, to V,DJ, joining mechanism in heavy-chain negative (null) pre-B cells results heavy chain production. Nature 322: 840~43 98. Kleinfield, R., Hardy, R., Tarlinton, D., Dangl, J., Herzenberg, L. A., Weigert, M, 1986. Recombination between an expressed immunoglobulin heavy chain gene and a germ+ line variable gene segment in a Lyl B-cell lymphoma. Nature 322: 84345 99. Heilig, J. S., Tonegawa,S. 1987. T-cell ~ gene is allelically but not isotypically excluded and is not required in known functional T-cell subsets. Proc. Natl. Acad. Sci. USA 84:8070-74 100. Born, W., Rathbun, G., Tucker, P.,

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

Marrack, P., Kappler, J. 1986. Synchronized rearrangement of T-cell y and fl chain genes in fetal thymocyte development. Science 234:479-82 Snodgrass, H. R., Kisielow, P., Kiefer, M., Steinmentz, M., yon Boehmer, H. 1985. Ontogeny of the T-cell antigen receptor within the thymus. Nature 313:592 95 de Villartay, J.-P., Lewis, D., Hockett, R., Waldman,T. A., Korsmeyer, $. J., Cohen, D. I. 1987. Deletional rearrangement in the humanT-cell receptor s-chain locus. Proc. Natl. Acad. Sci. USA 84:8608-12 de Villartay, J.-P., Hockett, R. D., Coran, D., Korsmeyer, S. J., Cohen, D. I. 1988. Deletion of the humana T cell receptor gene by a site-specific recombination. Submitted Janeway, C. A. Jr., Jones, B., Hayday, A. 1988. Specificity and function of T cells bearing ];a receptors. Immunol. Today 9:73-76 Brooks, C. G. 1983. Reversible induction of natural killer cell activity in cloned murine cytotoxic T lymphocytes. Nature 305:155 58 Van de Griend, R. J., Tax, W. J. M., van Krimpen, B. A., Vreugdenhil, R. J., Ronteltap, C. P. M., Bolhuis, R. L. H. 1987. Lysis of tumor cells by CD3+4 8 16+ T cell receptor aftclones, regulated via CD3 and CD16 activation sites, recombinant interleukin 2, and interferon ft. J. lmmunol. 138:1627 33 Phillips, J. H., Weiss, A., Gemlo,B. T., Rayner, A., Lanier, L. 1987. Evidence that the T cell antigen receptor maynot be involved in cytotoxicity mediated by y/6 and c~/fl thymic cell lines. J. Exp. Med. 166:1579-84 Bevan, M.J. 1977. Killer cells reactive to altered-self antigens can also be alloreactive. Proc. Natl. Acad. Sci. USA 74:2094~98 Ashwell, J. D., Chen, C., Schwartz, R. H. 1986. High frequency and nonrandomdistribution of alloreactivity in T cell clones selected for recognition of foreign antigens in association with self class II molecules. J. Immunol. 136: 389 95 Nixon-Fulton, J. L., Hackett, J., Lewis, J., Bergstresser, P. R., Kumar,V., Tigelaar, R. E. 1987. Diversity ofcytotoxic activity mediated by cultured mouse Thy! + epidermalcells. J. Invest. Dermatol. 88: 509A Marusic-Galesic, S., Pardoll, D. M., Saito. T., Leo, O., Fowlkes, B. J., Coligan, J., Germain, R. N., Schwartz,

Annual Reviews


GAMMA AND DELTA TCR R. H., Kruisbeek, A. M. 1988. Activation properties of T cell receptor-Tb hybridomas expressing diversity in both y- and 6-chains. J. Immunol. 140: 411-18 112. Tentori, L., Pardoll, D. M., Zuniga, J. C., Hu-Li, J., Paul, W. E., Bluestone, J. A., Kruisbeek, A. M. 1988. Proliferation and production of IL2 and B cell stimulatory factor I/IL-4 in early fetal thymocytesby activation through Thy-1 and CD3. J. Immunol. 140:1089-94 113. Davis, M. M., Bjorkman, P. J. 1988. T cell antigen receptor genes and T cell

207

recognition. Nature 334:395402 114. Jones, B., Mjolsness, S., Janeway, C. J. Jr., Hayday, A. C. 1986. Transcripts of functionally rearranged gammagenes in primary T cells of adult immunocompetent mice. Nature 323: 635 38 115. Strauss, W. M., Quertermous, T., Seidman, J. G. 1987. Measuring the human T cell receptor ~,-chain locus. Science 237:1217-19 116. Brenner, M. B., Strominger, J. L., Krangel, M. S. 1988. T7~5T cell receptor. Adv. lmmunol. 43:133-91



Annual Reviews www.annualreviews.org/aronline Ann. Rev. Immunol. 1989. 7:209~49 Copyright © 1989 by Annual Reviews Inc. All rights reserved


V-REGION CONNECTIVITY IN T CELL REPERTOIRES P. Pereira,

A. Bandeira and A. Coutinho

Unit6 d’Immunobiologie, Institut Pasteur, 75724Paris C6dex15, France M.-A. Marcos, M. Toribio and C. Martinez-A. Centro de Biologia Molecular, Universidad Aut6nomade Madrid, 28049Madrid, Espafia INTRODUCTION Areviewof the evidenceand ideas concerningT cell repertoire (Tc Rep) selection within an immune networkis both simple, becauselittle has been done, and difficult, becauseof the waythis has beendone. Theproblemis not only conceptual(1). Ourmethodsare excellent tools for clonal analysis but certainly not appropriatefor networkstudies. If a networkis described by its structure (connectivity), dynamics,andmetadynamics (2), it is couragingto realize that wehaveno suitable techniques,evento quantitate the first of those parameters. If weexclude fromnetworkimmunology all studies that address, instead, idiotypic regulation of clonal activities, we are left with astonishinglylittle materialfor a reviewon immune networks, let alone"T cell connectivity."It is, therefore, difficult to identify among the manypaperson idiotypes (Ids), T cell reactivities, Tc Repselection, responsiveness,and tolerance, those that suggest modesof operation and selection compatiblewith a networkorganization. Westart by reviewing experimentsshowingconnectivity amongT cell receptors (TcRs) and continuewith the evidencefor interactions between TcRsand immunoglobulin (Ig) idiotypes in the selection and maintenance of Tc Reps. After briefly discussing antibody (Ab) networksand their relationships to Tc Reps, we describe a frameworkaccommodating these observations and somefindings in tolerance and immuneresponsiveness. 209 0732-0582/89/0410-0209502.00

Annual Reviews 210

PEREIRA ET AL


CONNECTIVITY AMONGT CELL RECEPTORS In a very large and diverse collection of proteins it should always be possible to find (and define) a level of connectivity typical of that set molecules in the conditions as they are. If such molecules are TcR, the interesting question relates to the frequencyof interactions with sufficient affinity to result in functional consequencesfor the T cells expressingthem, and therefore, to the importanceof such connectivity in the biology of the wholeset. Augustin and his colleagues have carried out a series of ingenious experiments demonstrating that direct TcR complementarities can be found in normal individuals (3, 4). Since alloreactive T cells to class-II major histocompatibility complex (MHC)products are abundant in normal animals, the strategy consisted in finding the respective complementarities, which would constitute some sort of "image" of the alloantigen. These authors first isolated T cell lines and hybridomasthat were directly stimulated by monoclonal antibodies (mAbs) to class II determinants and demonstrated the reaction of the mAbwith the respective TcRs. They could then showthat such T cell hybrids stimulated T cell lines and other hybrids with the same class II alloreactivity as the mAbthat had induced them. Specificity controls established the clonal nature of such interactions and thus provided direct evidence for their existence in the normal Tc Rep. Interestingly, fine specificity analysis established that, as expected, the "points of view" of antibodies and TcRthough overlapping are not identical. Furthermore, the TcR shown to be complementary to an allo-class I~reactive TcRwas itself (self) class II-reactive, that is, the twointeracting TcRscould both be described as "images" of class II molecules. Finally, although the interacting T cells were both I-A-reactive helper cells, they could productively interact in the absence of I-A but were nevertheless inhibited by the appropriate anti-I-A mAb,in its quality as anti-TcR. If suggestive and provocative, these results give no indication as to the frequency of such complementarities within the Tc Rep. Experiments by Suciu-Foca et al (5, 6) go further in this respect by showingthat normal peripheral humanT cells proliferate quite vigorously if stimulated in vitro with autologous T lymphoblasts generated in response to allogeneic classII antigens. The specificity of these autologous responses was established by secondary stimulation (primed lymphocytetest) and shownto correlate with the reactivity (and thus probably the TcR) of the stimulator, alloactivated T cells. Similarly suggestive evidence for frequencies of TcR complementarities to class II-reactive TcRs, high enoughto be detected in primary proliferative responses of normal T cells, was reported by the laboratories of Kaplan (7) and Weksler (8). Both used T cell clones

Annual Reviews


T CELLS AND NETWORK

211

hybrids directed to self I-A determinants as stimulator cells and argued for the abundance of such TeR complementarities to anti-self class-II TcRs. These observations were significantly expanded by the work of Quintans and colleagues (9) who showedthat auto (class II)-reaetive well as antigen-specific, self-(class II)-restricted T cell lines and clones were all capable of stimulating considerable proliferative responses in syngeneic normal T cells. As in Augustin’s experiments (2), although these responses occurred in the absence of class II-bearing cells, they could be inhibited by anti-class-II Abs with the same target specificity as the stimulator class lI-reactive T cells. Both CD4+ and CD8+ T cells were shown to participate in these autologous, putative anti-TcR responses; most interestingly, T cell populations responding to a particular TcRcould themselves stimulate second-orderreactions in other T cells, with a clear increase in degeneracy. The most documented example ofa TeR network, however, is provided by the studies on graft-vs-host (GvH) resistance conducted by Darcy Wilson and associates, which opened this area of research and constitute the best evidence for the paramountbiological significance of connectivity amongstTcRs(10-14). If pretreated with small numbersof parental strain A, CD4+ Tc, containing A anti-B (MHC)precursors, (A x B)F1 hybrid rats becomeresistant to local GvHreactions induced by large numbersof parental A T cells, but not to those induced by parental B T cells. Resist+ host T cells which display a clear ance is primarily mediated by CD8 specificity for TcRV-regions, as they suppress and kill A anti-B, but not A anti-C or anti-D T cells. Most interestingly, hosts maderesistant to A anti-B TcRs also suppress and specifically kill all other anti-B T cells prepared in other MHC-haplotypedonors. This finding, which has been interpreted to indicate a limited polymorphism and little somatic modulation among TcRs to MHC-encoded antigens in the species, has other important implications. First of all, and as seen above in other systems, direct interactions between TcRs are not MHC-restricted and thus are likely to result from complementarity between three-dimensional shapes of "unprocessed" receptor proteins. This is important in the context of the repeated observations that alloMHC-reactive TcRs "recognize" conformational determinants (e.g. 15, 16) rather than linear, processed peptide sequences, as is mostly the case for conventional protein antigens (17-20). It can then be argued that TcRs on host "anti-idiotypic" T cells are "images" of self-MHC, as demonstrated in Augustin’s example and suggested in Quintan’s experiments. If this were the case, there would be no reason to postulate a limited polymorphism for anti-MHC TcRs in the species, and Wilson’s observations wouldnot be at variance with previous descriptions of the existence

Annual Reviews


212

PEREIRA ETAL

of a polymorphicidiotypic repertoire of such receptors, as analyzedby Abs(2 l). Thus,all anti-BT cells, regardlessof their idiotypicprofile, share complementarities to B (MHC)and, therefore, to "images" of B in the form of "anti-idiotypic" TcRs.It is remarkable, nevertheless, that GvH resistance is complete and, therefore, that all "aspects" of B (MHC) are fully represented by TcR"images." This interpretation has other implications, becausethe experimentsalso showthat resistance against a very frequent reactivity (A anti-B MHC) is acquired within a very short time after injection of parental cells (D. Wilson,personalcommunication). This leaves little roomfor clonal amplification of the "anti-idiotypic" T cells andsuggestsactivation and/orrelocalization of frequent, pre-existing clones. Sucha high frequencyof TcRsthat "reproduce"the molecularprofiles of self-MHCmaybe explained in two ways. They mayarise by strong positive selection in ontogeny,implying, as arguedbefore (22), that the original modeof reaction in the immunesystem is the generation of mimicriesrather than the production of complementarities.Someevidence for this assumption has been obtained by the abundanceof "natural" serumAbs, particularly in very younganimals, that bind to anti-(self) MHC monoclonalantibodies (23). Someof these are actually allo-MHCspecific (24). Thecentral point here is that sucha selective processmust be driven by other TcRs(these being anti-self MHC) in MHC-unrestricted (but, at the origin, MHC-directed)manners.Thus, since the "points view"of MHC profiles by TcRsand antibodies do not coincide, the "copy" of self-MHC to TcRidiotypes that is seen as faithful by T cells initiating GvHreaction, can only be achieved if such T cells themselves have mediatedthe process of positive selection. Hypothesesthat postulate such basic principles for T cell selection in the thymushavepreviously been proposed (2, 25). Alternatively, if TcRsmimickingself-MHCare not positively selected, it shouldbe expectedthat the idiotypic profiles corresponding to allogenic MHC would also be present at high frequency in normalindividuals and thus wouldco-exist with high frequencies of complementary,anti-allo MHC TcRs. In other words, the Tc Rep would form a very high connectivity network. Wilson’s experiments cannot address this point, but the evidencereviewedabovewouldperhapsindicate the predominanceof self (MHC)images. Thesetwo alternatives are not mutuallyexclusive and positive selection of self-related complementaritiesand mimicriescan well be done from a high connectivitynetworkthat is "species-specific." Thestructure of such a network cannot be expected to consist of two complementarysets of TcRs(complementaritiesand mimicries of MHC) separated in two internally disconnectedhalves. As is already suggestedby Augustin’sexample

Annual Reviews


T CELLS AND NETWORK

213

(the anti-anti-self TcRis itself an anti-allo), a greatly interwovennetwork should apply, where complementarities to antigen should largely be outnumberedby the respective (partial) images. Such a model of Tc Reps has been extensively developed by Eichmannand colleagues (see below) with interesting general properties of suppression and stability. Evolutionarily, the stability of a germ line-encoded, highly connected TcR network has implications for other sets of genes, namely the MHC.As we argued before, to explain some characteristics of MHC polymorphisms(26, 27), this set of genes would only vary within the limits set by evolutionarily stable T cell repertoires. Apart from sporadic but recurrent examples that could be interpreted to indicate direct interactions among TcRs on helper and suppressor cells (e.g. 28-31), and from the work of K6hler (32-34), whoargued functional interactions between complementary TcRs on TH, a wealth of reports have described "cascades" and "circuits" of suppression and contrasuppression involving T cells and their clonally specific products. In manyinstances, the data have been interpreted to indicate direct V-region interactions amongTcRs (see for review, 35-38), and they contribute the notion of TcR connectivity. The discussion of these observations, however, is complicated by the intimate relationship unrevealed in those studies between the participating TcRs and Ig V-region Reps (see below). Furthermore, since TcV-region interactions were not directly shown but were deduced from experiments involving complex cell mixtures and since there are no available data on the frequencies of such T cells and interactions, we cannot use this evidence in quantitative discussions of TcR connectivity. Direct measurements in limiting dilution analysis (LDA)of normal cells and quantitative deductions on connectivity in a T cell network have been performed by Eichmann and colleagues (39-44). The basic experimental design consisted in polyclonally activating normal T cell populations with concanavalin A, letting limited numbersof activated cells expand for a week in the presence of exogeneous growth factors, testing clonal progenies for specific effector functions, and calculating clonal precursor frequencies in the starting T cell population. The approach is similar to one previously used for B lymphocytes(45), and it rests on the assumption that mitogen activation reveals an unbiased sample of the available repertoire. Since total frequencies of responding T cells were not determined, we do not knowto which fraction of all cells the results apply. These approaches, however, offer the considerable advantage of overcomingvariables in the initial activation and growth of cells (antigen concentration, accessory cell activities, etc) whichconstitute serious limitations for antigen-driven specific precursor frequency determinations.

Annual Reviews


214

PEREIRA ET AL

Furthermore, in this particular case, the system provided the opportunity of detecting interactions amongactivated T cells and thus revealed activities that wouldpass undetected with populations of resting T cells. From the shapes of the lines defined by LDAdata points in semilog plots, these authors derived several important conclusions. The first concerns the enormous degeneracy in the functional recognition of antigens by T cells: for every antigen tested (e.g. major or minor histocompatibility antigens, hapten modifications of self, heterologous erythrocytes), the frequency of specific clones approaches 1%of all normal cells. Antigen-specific effector T cells, however,could only be obtained in quantitative conditions that would fit the zero-order term of the Poisson distribution from cultures containing very low numbers of activated responder cells. Higher numbersof such clonally heterogeneous cells drastically suppressedthe generation (or detection?) of specific effector cells. These were again detected within a narrow range of even higher cell concentrations in a second (or third) straight line in the semi-log plots. Specifically primed T cell populations analyzed in the same manner, essentially show clonal frequencies with "appropriate" Poisson’s distribution at the same level as the abundant population in unprimedcells, suggesting to the authors that priming results in resistance to suppression of very frequent, specifi~ clonal precursors, rather than in their expansion. After demonstrating that established T cell clones were also suppressed by hctcrogcneous populations of activated T cells, the authors used such cloned T cells to demonstrate that antigen competes with suppression and restores functional reactivity. This observation provided the major argument for receptor-mediated mechanismsoperating in this system and, together with the demonstration that T cell populations with helper and suppressor phenotypes were both operational suppressors, led to a network model of Tc Reps. A high degeneracy of specific TcR interactions with ligands, if it ensures high specific precursor frequencies, also results in a very high degree of connectivity in Tc Reps. In a randomrepertoire, each complementary TcR to any antigenic pattern is largely (10-100 fold) outnumberedby other TcRs that interact with it with minimal affinities. Such direct TcRinteractions are postulated to be suppressive, in contrast with TcR-antigen contacts that are inductive. It follows that the normal state is dominantly suppressed and that responses can only be obtained by antigen competition with the connected TcRsat the level of high affinity cells. Activation of T cells by antigen would then commanda maturation step in the responding cell such that it would no longer be sensitive to connectivity suppression. As expected, this set of observations and interpretations met with skepticism, so much did they depart from conventional wisdom (biased Tc

Annual Reviews


T CELLS

AND NETWORK

215

Reps to MHC,antigen-dependent clonal expansion, deletion of anti-self T cell reactivities, etc). The experimental observations, however,have been well established, and the modelhas provided a unified view of the question. Unfortunately, it has not explored the phenomenology of allo (MHC)reactivity, Ir-genes, tolerance, thymic education, and it did not include Ab V-regions in a wider view of immunenetworks. The assumption of a symmetrical networkwith "distributed" connectivity can also be criticized, since the data can as well be accommodated with the postulate of a special population of"sticky" TcRs. Furthermore, the whole of the evidence rests on the analysis of activated T cell populations and is applied to normal conditions where most lymphocytes are resting (GvHresistance, for example, which can be considered in the context of a high connectivity T cell network, requires activation of the "anti-idiotypic" T cells). Finally, it is difficult to accept that TcRinteractions with someligands (other TcRson resting T ceils) are suppressive, while being inductive with other ligands (antigen), particularly because the affinity argumentmakeslittle sense the context of a network. Current limitations in the methodsof study, together with the oddity of these approaches in comparison with "fashionable" T cell immunology, mayexplain the little attention currently given to TcRconnectivity. Unfortunately so, for it appears a major characteristic of Tc Reps and a promisingfield for direct clinical applications (see e.g. 46).

EVIDENCE SELECTION IDIOTYPE

FOR B-CELL/ANTIBODY-DEPENDENT OF T-CELL REPERTOIRES: SHARING

Up to the early 1980s, a very productive area of research employedantibodies to IgV-region determinants to probe T cell repertoires. Antibodies to idiotypes, VH-isotypes, or allotypes were repeatedly shownto identify T cell or T-cell products clonally, and to stimulate or suppress specific T cell responses. Extensive reviews (47-49) have dealt quite recently (50) with this topic, which was also addressed in very critical ways (51-53). Before the identification of the TcRproteins (54-58) and structural genes (59, 60), the dominant interpretation of those results has been the expression of Ig V~-genesby T cells. After 1983 a paradoxical development took place: manyof the most active groups deserted the field, although in the light of the new information on TcR structure, those results had extraordinarily important implications for the organization of the immune system and the selection of Tc Rep. Thus, if a minimumof credit is given to that large set of observations, "Id sharing" by structurally different

Annual Reviews


216

PEREIRA ET AL

receptors suggests powerful mechanismsdriving the somatic selection of Tc Rep on the basis of antibody idiotypes, "implying a functional idiotypic network in the most fundamental sense" (50). This would indeed appear to be a very worthwhile topic to pursue. The backlash is understandable after the dominant hypothesis (and some of the biochemical data) had been falsified, and doubts accumulatedconcerning the genetic and cellular basis of other phenomenaintimately connected with "Id sharing" (e.g. specific T cell factors, suppressorT cells, I-J). It is, nevertheless, surprising that, precisely from the momentwhen the experiments could be better controlled and interpreted, manyfewer groups continued, or started, work in this area, and there was little developing of novel experimental approaches or understanding of the phenomenology. Only a few aspects of the older work on "Id sharing" should be discussed here. First of all, howoften does it occur and, consequently, howsignificant is the phenomenonfor Tc Reps? The answer is not simple because "Idsharing" is necessarily detected by an anti-idiotypic antibody (or TcR), and it may, therefore, vary with the reagents used. This is particularly relevant when monoclonal antibodies are used to define idiotypes, and it is probably not only because polyclonal sera are "dirty" reagents that idiotypic mimicries were more often recorded with mixtures of antibodies (see 61, for discussion). The most reasonable way of addressing the question experimentally would be to use as reagents the same combiningsites (Abs or TcRs) that the system itself uses in the putative selection mimicries; in other words, to adopt the "point of view" of the systemitself, in order to understand how it works. Unfortunately, with few exceptions (62, 63) the monoclonalantibodies used as anti-idiotypic are derived after extensive immunizationsand are most likely not present in normal animals. Finding of "Id-sharing," therefore, will be a matter of chance or of extensive screening of anti-idiotype monoclonal antibodies. An operational question here is how and what to screen, since there is no good reason why"Id-sharing" should be found in T cells and B cells participating in the particular immuneresponse that the investigator has chosen to elicit. In practice, and regardless of how fundamental the phenomenonmight be for the immunesystems, it is a priori unlikely that it has evolved for the recognition of, say, TNP. In general, since T cells and B cells "see" molecular patterns in essentially different ways--short peptide sequences together with MHC,versus three dimensional shapes--"Id-sharing" may well exist, but the respective lymphocytesparticipate in immuneactivities that the investigator considers separate. Again, the solution here should perhaps be to consider the globality of the immunesystem operation and to adopt its own "point of view". "Id-sharing" should, therefore, be analyzed in concomitant immuneactivities even if these concern apparently

Annual Reviews


T CELLS

AND NETWORK

217

distinct specificities. A clear suggestion for this is provided by the Oudin and Cazenave phenomenon: "Id-sharing" occurs among Bc/Ig molecules, all stimulated in an immuneresponse, but only someare antigen-specific Abs (64-66). In view of these limitations, we can instead ask howoften "Id-sharing" between T cells and B cells has actually been detected. The phenomenon has been described in a variety of immuneactivities concerning helper, suppressor, and cytolytic lymphocytes(and their products) responding MHCdeterminants (67-69), a number of different haptens (70-72), terial products (73, 74) and conventional protein antigens (75, (reviewedin 47-50). The list is incompleteand unlikely to be representative since with few exceptions (10, 77) failures are usually not reported, and are not aware of any systematic study addressing this point. The closest is a brief description by Martinez-A. et al on screenings of "primary", uncloned Tr~ lines, recently derived from lymph nodes immunizedagainst a variety of very little cross-reactive hapten modifications of "self" (TNP, NP, NIP, FITC, SP), with a number of monoclonal antibodies to idiotopes expressed by several TNP- or NP/NIP-binding Abs from the corresponding mousestrains. They reported that only one of seven antiidiotypic mAbsshowed inhibition of TH activity (71, 78) and immunoprecipitated a heterodimeric, clonotypic molecule from T cells (79). Incidentally, this is actually the only report on "Id-sharing" that shows biochemical evidence of anti-idiotype antibody reactivity with a T cell surface molecule bearing characteristics of TcRs(54-58). Interestingly, the context of the above considerations, these authors also reported that, if instead of conventional anti-Id antibodies they used "natural mAbs" with the same nominalanti-Id specificity, two of three tested were capable of THinhibition (62, 63). Wecould perhaps conclude as others did before (77) that "Id-sharing" between T cells and B cells participating in the response to a nominalantigen is the exception rather than the rule, at least when analyzed with conventional anti-Id mAbs. If the reagents used by the system in its operation are also used in the screenings, however, the situation might be quite different. The next question on "Id-sharing" concerns its significance. In the large diversity of Abs and TcRs it is more than likely, on simple statistical grounds, that similar shapes will be found. It is necessary, therefore, to ascertain if "Id-sharing" is the result of selection or of pure chance. Most of the described exampleshave concerned T cells and B cells with similar nominal specificities, and this has been taken to indicate "meaningful" regulation. This would only be the case, however, if the Id could not be frequently found in the responses to manyother antigens--which cannot be experimentally verified. An example of such aleatory "Id-sharing"

Annual Reviews


218

PEREIRA ETAL

between THand B cells has recently been described (80) and shown be completelyindependentof B cell/Ig-dependentselection of the T cell clonotype(81). There are no indications of the frequencythat should expectedfor this kind of cross-reactivity, but those observationscommand caution on "network interpretations" of Id-sharing, in the absence of formalevidencefor meaningfulId selection. The best argument for the conclusion that "Id-sharing" does represent selection of Tc Repsin a networkthat includes B cell/Ig is the finding that Id-expressionby T cells is controlled by IgVn-linkedgenes (71, 72, 82-88). Since TcRare not encodedby IgH-linked genes, those repeated observations strongly suggest that the Tc Reps analyzed had previously been selected by mechanismsinvolving polymorphicantibody V-regions. This argumenthas recently received a formal proof by the demonstrationthat "idiotype sharing" by TcRsdoes require the presence of B cells/Igs during Tc Repdevelopment(78, 89). Further evidence for this central point was provided by experimentsdirectly comparingIgHlinkage and B cell/Ig-dependenceof the expressionof two distinct T cell clonotypes, shared with the sameB cell idiotype, and both identified by the sameanti-Id mAb:IgH-locuscontrol ofT cell idiotype expression was foundassociated with the requirementfor the presenceof B cell/Igs along with Tc Repdevelopment, but not whenthe "shared Id" was expressed independentlyof the Bc compartment (72, 78, 80, 81). Giventhe variety of experimentalsystems in whichIgH-linked"idiotype-sharing" betweenAbs and TcRshas been documented,this process wouldappear to play a significant role in the establishmentof Tc Reps. Other examplesof mimicrybetweenantibody idiotypes and self structures (90-93) have not shownany IgH-linked polymorphismin expression. expected,these findings indicate that, given a range of alternative possibilities providedby the geneticandstructural diversity of TcRs,a coherent solution for patterns in Tc Repsemergesfrom the operation of an immune network.In other words,"Id-sharing" maywell not represent a particular advantageto the immunesystem and, therefore, maynot be selected for a given purpose. Rather, it maysimply be a consequenceof the wayit functions. EVIDENCE FOR B-CELL/ANTIBODY-DEPENDENT SELECTION OF T-CELL REPERTOIRES: IGH-RESTRICTION OF T CELLS AND T CELL-DERIVED FACTORS In the samemanneras MHC-restrictionof antigen recognition by T cells has beentaken to indicate Tc Repselection uponinteractions with MHC-

Annual Reviews


T CELLS AND NETWORK

219

encoded glycoproteins in T cell ontogeny, IgH-restriction of T cells and their products is suggestive of IgV-region participation in selecting functional Tc Reps. These two types of phenomena, however, are likely to have distinct structural and functional bases and to have very different impacts in Tc Reps. IgH-restriction describes the findings that T cells or their: clonally specific factors fail to interact functionally with lymphocytes from donors that carry a distinct IgH-haplotype. Originally described as limiting interactions between various types of T cells participating in suppressive activities (88), IgH-restriction was later shownto apply also to THactivities (94) and to concern V- rather than C-genes in the locus. The central point in this set of observations, constituting formal proof of B cell/Ig-dependent selection of the Tc Repsanalyzed, is the demonstration that IgH-restricted T cells require the presence of B cells/Ig for the acquisition of the self-restricted phenotype (89). Moreover,such T cells have been shownto "learn" restriction to either self (95) or allogeneic Ig Reps (96) if exposed to the appropriate environments. In contrast with MHC-restriction, peripheral T cells can be re-educated to alternative IgHrestriction patterns, in a process that takes less than two weeks(96). Because some antibody idiotype-related TH activities were shown to proceed independently of MHC-linkedcontrols, in what concerns both Irgeneeffects and restriction in target cell interactions (97, 98) the notion was developed that MHC-and IgH-encoded proteins could serve equivalent functions in the biology of two parallel sets of TH(99, 100), which would differ also in functional capabilities and surface markers (101, 102). discussed below, this division is, in our minds, unwarranted. Thus, the notion of IgH-restriction of TcRs is quite different from that of MHCrestriction. While the latter essentially concerns "presentation" of antigen fragments in the context of MHC-molecules,the former is likely related to direct complementarities between TcRs(some bearing "idiotypic" determinants of antibodies, others constituting the respective "anti-idiotypic" counterparts). Althoughit could in principle be possible, we do not know of evidence suggesting that TcRs or factors recognize antigens "in the context" of other V-regions, be they antibodies or TcRs. Even if some of the systems showing IgH-restriction are antigen-dependent, this may be due to totally different reasons. Furthermore, absence of MHC-restriction (and Ir-gene effects) in these interactions does not have to be interpreted as the property of a distinct class of T cells. Thus, MHC-restrictedT cells can readily be activated by anti-TcR or anti-CD3 antibodies in MHCunrestricted manners(57, 103). In other words, cell activation requires TcRsinteractions with threshold affinities: in somecases, that threshold is only reached if the ligand is a "peptide-MHCcomplex" (and cell interactions stabilized by accessory molecules), but the same TcRcan obviously

Annual Reviews


220

PEREIRAET AL

directly interact with high affinity complementaryantibodies or other TcRs. The above discussion of TcRconnectivity provided several examples of T cells that interact with TcRsof other T cells in MHC-unrestricted manners. Yet, we should perhaps not define a third category of T cells that are TcR-restricted. Let us consider first, that antibody idiotype-specific T cells are likely to be: (a) both IgH-restricted and MHC-restricted, when they recognize polymorphic idiopeptides presented by MHC-molecules; (b) MHC-restricted only, whenthe V-region peptides are not polymorphic; and (c) simply IgH-restricted when they interact with three-dimensional structures on polymorphic antibody V-regions. On this non-network scheme, a second level of complication is brought in by the fact that Tc Reps are selected by B cell/Igs and, therefore, differ according to the IgHhaplotype of the mouse. T cells thus selected can also be said to be restricted by IgH-genes (although we would prefer to say "determined" or "selected"), and the expression is often used in this context. These very cells, however, can showall properties of MHC-restricted or MHC-specific recognition of targets and are also submitted to Ir-gene controls (e.g. 7 i). Provided the appropriate targets, however, these same T cells can also function in MHC-independentmanners by direct interaction with Ab or TcR V-regions. It is our impression that suppressor T cell "cascades" represent Ig-dependent selection of complementaryTcRs, while the observations on MHC-independent Id-specific TH represent Ig-dependent positive selection of TcRswith sufficient affinity to interact directly with unprocessed Ab V-regions. Given that MHC-unrestricted T cells have been very rarely detected in immuneresponses to conventional antigens (104), it is pertinent to ask whether the apparent abundance of MHC-independentT cell reactivities to TcRor AbV-regions should be given a particular significance. Clearly, if Tc Rep selection is an ongoing process throughout life, molecular patterns that are present in the "internal environment" such as Abs and TcRs have more possibilities for selecting complementaryTcRswith sufficient affinity to dispense of MHC presentation. In this case, the difference betweenV-regions and other structural protei.ns available in the organism should be explained (whenarguing for either positive or negative selection). Alternatively, it might be argued that V-regions of TcRsand even Abs are present inside the thymusand thus participate in the primordial selection, in parallel or in conjunction with MHC-proteins(25). Even if correct, however, this hypothesis cannot accommodateall available findings, particularly the B cell/Ig-dependent "education" or "re-education" of peripheral T cells transferred to adoptive hosts. A reasonable model should perhaps consider, in addition to the above possibilities, that the "preoccupation" of Tc Reps with TcRs and Abs (and MHC)may also

Annual Reviews T CELLS

AND NETV~ORK

221


partly due to germ-line gene composition but, most importantly, may simplyhave to do with V-regiondiversity. Since the numberof molecular shapes of V-regionsis several orders of magnitudelarger than that of all other proteins in the vertebrate body,there is obviouslya muchgreater chanceto find the appropriate complementaritiesamongthe former set, whatever the TcRconsidered and the requirements for "appropriate" mightbe. Again, wecomeback to the idea that the immunesystemcannot avoid existing as a networkand that manyof its properties simplyemerge fromthis organization. EVIDENCE FOR B-CELL/IMMUNOGLOBULINDEPENDENT SELECTION OF ALLO(MHC)REACTIVE T-CELL REPERTOIRES Asidefromthe IgH-linkedresponsesof T cells to anti-idiotypic antibodies and the control of suppressor T cell interactions, evidence of quite a different nature has directly implicatedB cell/Ig in the selection of Tc Reps(105-111).Sherman’sexperimentsinvolvedfine-specificity paratopic clonotypingof primaryallo-reactive eytolytic T cells in multipletests of single clones on a panel of target cells expressingdifferent mutationsof the sameclass-I antigen. Analysisof the CTLrepertoire in individual mice revealedthe existenceof"recurrent",strain-specific clonotypicreactivities expressed at high frequency which could, therefore, be used to probe influences of various loci in CTLReps. As expected, there was a marked influence of MHC in the allo-reactive repertoire, mostof whichcould not be explainedbyself-toleranceandcross-reactivities. Interestingly, neonatal mice from MHC-congenic strains expressed a far more similar Tc Rep than did adult individuals, and F1 animals co-dominantlyexpress both parental elonotypic patterns. Repertoire divergenceis, therefore, due to ongoingselection throughout life, as demonstratedby the analysis of MHC-congenic T cells obtained from double parent chimeras. Having established that such selection is MHC-dependent, Shermananalyzed the putative influence of other polymorphicnon-MHC genes and surprisingly found a major effect of IgVH-genes. Theseobservationsare interesting in several regards. First, they constitute the only evidenceto date demonstratingthat available Tc Repsin unprimedindividuals, evenif analyzedin waysthat do not involveB cells or Abreagents, are drastically selected by AbV-regions.Second,they show that B cell/Ig-dependent selection is MHC-determined and established in a time course whichsuggests the requirementfor prolongedco-existence with Abrepertoires. This is a central point as it illuminatesthe discussions

Annual Reviews


222

PEREIRA ET AL

above on "IgH-restricted" T cells (we are dealing with MHC-specific repertoires) and on TcRnetworks. Thus, the observations concernrepertoires of allo-reactive specificities andmodesof selection that cannotbe explainedby deletion. Therefore, they must represent positive selection of complementaritiesto not-self molecularprofiles. Thesedata are not compatiblewith modelsof Tc Repselection solely based on a networkof TcRs(4, 44) and imposethe inclusion of AbV-regionsin such a network, whichapparently concerns the wholeTc Rep(self- and not-self-directed reactivities). Finally, they suggesthowextensiveis the impactof B cell/Igdependentselection on Tc Repsand the interest in extendingthese studies to other aspectsof T cell reactivities, particularlyto Ir-geneeffects. EVIDENCE FOR B-CELL/IMMUNOGLOBULINDEPENDENT SELECTION OF T-CELL REPERTOIRES: B CELL-DEPRIVED ANIMALS Theevidencefor selection of T cell reactivities in a networkthat includes AbIds suggested a direct approach: to compareTc Reps ontogenically establishedin the presenceor absenceof B cell/Ig. Thus,it is possibleto producemicethat are profoundlydepleted of matureB cells andcirculating Ig by chronic treatments frombirth or embryoniclife with anti-isotypic (112) or -allotypic (113) Absto # chains. Since early studies had shown that such animalshaveconservedT cell functions (114), it waspossible compare"Id-sharing" and "IgH-restriction" of T cells in B cell/Igdeprivedmice and control littermates. Results fromthis type of experimentalsystemhavebeen re.ported by several groups, and all agree in the conclusion: T cell clonotypes expressed in B cell-deprived animals are different fromthose available in normalindividuals (78, 89). Martinez-A. and colleagues have shownthat a syngeneic mAbto the BALB/cTNPbinding Id MOPC 460 identifies a TcRclonotype on the majority of TNPBALB/c specific Tn obtained in normal animals of this strain. The mAB specifically inhibits T~proliferation and effector functions on specific targets, stimulates IL-2 production under appropriate conditions, and immunoprecipitates a clonotypic heterodimer from those TH. Although Bc-deprivedBALB/c mice produceTNP-self-specific TH,none of the antiId mAbreactivities could be reproducedwith them(78, 79). Independently,Sy et al (89, 115) and Floodet al (116), studying pressor T cells operating in DTHresponses to azobenzenearsonate(ABA) or Abresponsesto sheeperythrocytes, respectively, describedthe finding that T cells from B cell-deprived donors were both unable to produce suppressorfactors active on cells from normalanimals, and resistant to

Annual Reviews


T CELLS

AND NETWORK

223

the activity of such factors produced by normal donors. The studies of Sy and colleagues are particularly detailed and complete, in a system previously developed by Greene and Benacerraf (reviewed in 37). ABA-specific suppressor T cells from A/J mice produce factors that react with anti-idiotypic Abs directed to the dominantAbId of the strain, and these factors are restricted in their effector activity to "acceptor" T cells from donors carrying the same IgH-haplotype (37, 115). Suppressor factors prepared from T cells of B cell-deprived donors, while competent on cells from these same animals, were no longer active on target cells from normal donors (89, 117), showingthat the repertoires of interacting T cells were altered if established in the absence of B cell/Ig. Interestingly, depletionreconstitution experiments directly support the interpretation that Tc Reps available in normal individuals are positively selected and expandedby B cell/Ig. Thus, interruption of anti-# Ab administration, which is followed by replenishment of the B cell/Ig compartment,results in acquisition of T cell reactivities that are characteristic of normal animals and absent in B cell-deprived donors (95). Moreover,complementaryexperiments to B celldeprivation, namely the transfer of peripheral, mature T cells to IgHcongenic recipients resulted in "re-education" of Tc Reps, which now contained reactivities characteristic of both the donor and the recipient (but not if the latter was anti-#-suppressed) (96). Sy’s experimentspropose another interesting conclusion on the interrelationships betweenT cell and B cell repertoires. The major anti-ABAAb Id of BALB/cis a minor Id of A/J strain Abs and is "shared" by suppressor factors of BALB/cbut not of A/J mice. Such factors prepared in BALB/cor congenic mice carrying the A/J IgH-locus (C.AL-20) are only active on cells from the strain origin. Interestingly, however, cells from anti-#-suppressed animals of either strain, are suitable targets for suppressor factors producedin the other strain, suggesting that clonal dominance(in both T cells and B cells) is established by Id network-positive selection from equivalent potential repertoires. The consequences for the T~ repertoire of autologous B cell reconstitution in anti-p-suppressed mice were also analyzed by Martinez-A. et al (118), with somewhatdifferent results. In this case, B cell-deprivation for more than 4 weeks results in a definitive loss of expression of the TH clonotype studied. This finding, together with the stability OfTHclonotype expression in mice manipulated as adults (to express either large amounts of Ab Id or none) led the authors to conclude that, once established by cell/Ig-dependent mechanisms, Tc Reps were resistant to further alterations in the B cell compartment. These observations are compatible with earlier experiments on "Id-sharing" that had been interpreted to indicate VH-geneexpression by T cells and were difficult to reconcile with the new

Annual Reviews


224

PEREIRA ET AL

interpretations of Ab-dependentselection of T cell clonotypes (119, 120). Theseconsisted in the transfer of T cell populations containing Id-positive cells from adult donors to recipients lacking expression of that Id on Abs. Whenanalyzed a long time after transfer, Id expression was still detectable on T cells. The apparent discrepancy between Sy’s and Martinez-A.’s outcomein autologousB cell reconstitution is likely to result from different rules applying to the expression of the respective Ab ld after interruption of anti-# treatment. If repertoire selection is a circular process (xc ~ B), appears to be the case in Martinez-A.’s system (see below), his results are a reflection of recursivity. Studies on T cell activities in B cell-deprived mice were actually initiated by Janewayet al (121) whoprovided evidence for altered patterns of helper effects, using adoptively transferred, carrier-primed T cells, from normalor anti-/~-suppressed donors, in conventional cooperation systems. Together with Bottomly, Janeway has also produced the original observations on the absence in anti-#-suppressed mice of a particular THspecificity found in normal animals which was indirectly defined by the idiotypic pattern of the Ab responses analyzed (122). This work integrates a series of studies conducted by Bottomly and colleagues on Id-related THcells, which provides another example of the role of B cell/Ig in the establishment of Tc Reps (97, 98, 100, 122-125). Wheninvestigating the dominanceof the T15 Id in the TH-dependent Ab responses to phosphorylcholine, Bottomly described the fact that, in contrast with carrier-primed THfrom normal donors, those obtained from animals that do not express the T15 Id were unable to help a T15-dominated Ab response, while being otherwise competent helpers for specific Ab production (123). Subsequent work showedthat this class of Id-related Tn cells, which were characterized as being simultaneously specific for whatever carrier was used in the experiments, could be obtained by priming (Ir-gene)-nonresponder animals, and these cells displayed no requirements for "linked" collaboration nor MHC-restriction in their effector activity (97, 98). Interestingly, such THpopulations can be specifically "panned" or T15-coated plates (124, 125), showingthat their receptors interact with "unprocessed" Id, and thus explaining their peculiar properties. These observations are reminiscent of those of Woodland & Cantor (126) and Eichmannand colleagues (127), and similar findings in several other systems have also been reported (102, 128-131). They have constituted a major argument for Janeway’s model of Id-restricted (MHCunrestricted) T cells, discussed above, and are interesting here for they represent positive selection of TcRs that are complementaryto an Ab Id present in relatively high concentrations in normal individuals. Indeed, cells of this kind have been found in normal, unprimed animals (130, 131)

Annual Reviews


T CELLS

AND NETWORK

225

and are likely responsible for the stimulation of Id-positive B cells which do not bind antigen after priming with the antigen (66). Their special characteristics--as compared to "conventional" TH--maybe only apparent and result from a commondenominator: expression of TcRs that constitute somesort of "image" of the nominal antigen specific for the Id in question, a "r-type anti-Id." If this were the case, the same TcR can simultaneously be specific for any other antigen, and its interaction with Id-positive target cells would require no further "bridging" and no MHC restriction either. Anyactivation of THin vivo, even if "nonspecific" (as could be the case in genetically nonresponder mice), can be expected to generate cells with these properties. Dependingon the affinity of direct TcR-Id interaction, such "Id-related Try" will require or dispense simultaneous activation of the target B cells by other means--suchas linked, MHC-restricted cooperation with other TH--and/or concomitant activation by the antigen that originally induced them. Selective long-lasting suppression of a given Id can also be achieved by neonatal injection of specific anti-Id Abs(132) or by maternal transmission of such specificities (133-135). Such Id-suppressed animals are equally suppressed for expression of the TcRIds (72, 136). These experiments, however, even if performed in systems where "Id-sharing" was shown to be positively selected, do not have the same demonstrative value as those carried out with anti-p-suppressed mice. Thus, the absence of T cell clonotypes can as well be due to direct effects of the anti-Id Ab on target T cell. Nevertheless, since the effects of neonatal (or sometimesadult) antiId treatment are very long lasting (at least several months)it is likely that, either Id-specific active suppression is established (see e.g. 137-139), else that the absence of Id early in development leads to the positive selection (and consequent clonal dominance)of alternative (but equivalent) clonotypes normally "dominated" by the suppressed Id.

THE ANTIBODY NETWORK AND ITS T CELL-DEPENDENCE Wenow briefly discuss recent advances concerning the Ab network that will be useful for the global consideration of repertoire selection. Direct quantitation of antibody connectivity, performed in Kearney’s laboratory and by Holmberget al (140-142), has demonstrated a very high frequency of V-region interactions amongperinatal antibodies. The number of"connections" in a given set of Abs largely exceeds the number of members and reaches around 20%of all possibilities. High connectivity, however, is not a property of any diverse collection of Abs, as shownby parallel

Annual Reviews


226

PEREIRA ET AL

experiments using Abs randomly selected after polyclonal activation of adult, resting B cells. In this case, connectivity is 10-100 fold lower than that of perinatal Abs (141, 142). Furthermore, neonatal Abs are heterogeneous with respect to connectivity and seem to constitute thi:ee discrete groups, one of whichis very poorly or not at all connected(J. Stewart, F. Varela, A. Coutinho, manuscript in preparation). Interestingly, there is a clear association between connectivity and the use of VH-genesfrom the 7183 and Q 52 homologyfamilies (143), which are the first to be expressed in ontogeny due to their chromosomalposition (144). Since no somatic mutations have been detected in 12 such perinatal, high connectivity Abs (D. Holmberg,~.personal’communication), it appears that the immune system necessarily .starts as a high connectivity, germ-line encoded Ab network, the expression of which is developmentally controlled. Reactivities present in this original network stimulate B cell production by interacting mitogenically with non-Ig surface recep.tors on B cells and their precursors (93, 145; J. Kearney, personal communication). In addition, idiotypic connectivity participates in the (positive) selection of newlyarising Ab clonotypes in apparently T cell-independent manners, as elegantly shown"byVakil et al (146). At this point in development,most B cells are "naturally" activated, even in germ-free (147) and "antigen-free" animals (148). Auto-reactivity of such Abs, however, is only associated with "multispecificity," and further development is necessary, including the appearanceof peripheral T cells in sufficient numbers,before evidence for positive selection of auto-reactive B cells can be detected (see below). The interesting point to consider here is, indeed, the T cell-dependence of the establishment of Bc Reps. This topic has received little attention in the past, and very few examplesare available suggesting (or excluding) the participation of T cells,, before priming with antigens orother "artificial" manipulations. Thus, most observations can be interpreted to indicate helper or suppressor T cell participation in responses (either dealing with Ids as conventional antigens, or in Id-related mannerssuch as those discussed above), and such observations give no indication of T cell-dependent network selection of available B cell Reps. One report on the MHClinked control ofAbIds (149), however,is difficult to explain on this basis. More direct evidence might be sought by the~ analysis of B cell clonal precursor repertoires in unprimed populations, but while an early paper suggests MHC-linkedcontrol of T15 Id frequencies (150), others have provided examples of T cell-independent, IgH-controlled frequencies of other Ids (15l). Recently, Freitas and colleagues have detected very significant differences in the representation of VH-genefamilies in peripheral B cells from normal and athymic mice. Furthermore, they could show that transfer of normal T cells to nude littermates-resulted in profound

Annual Reviews


T CELLS

AND NETWORK

227

alterations in the expressedVH-geneRep. It is unlikely that such modifications result from ongoingresponses to environmentantigens, since they are detected amongsmall, resting B cells. Theseresults provide direct evidence for T cell-dependent selection of Ab Reps (A. A. Freitas, M.-P. Lambezat, B. Rocha, manuscript in preparation). Indirect suggestionsare providedby the T cell-dependence of B cell production discovered in mice carrying the Xid defect, which whenhomozygousfor the nude mutation produce few B cells (152) and very reducedlevels of circulatingIg (152, 153). T cell influencesoperate the level of bone-marrowprecursors (154) and normallevels of B cell productionthat can be reconstitutedby matureT cell or thymusgrafts (154, 155). It is possible that an AbRepdefect is implicatedhere, since Xid mice are reported deficient in the class of perinatal B cells associatedwith the productionof high connectivityantibodies (156). Furthermore,alterations of T cell function during adult bone-marrow reconstitution of Xid mice have also been shownto result in gross modifications of B cell/Ab repertoires (157; unpublishedobservations). Otherresults concerningspecificity repertoires, can also be invokedin this context. For example,mice recovering from neonatal anti-# suppression remain negative for the expressionof a T cell-regulated idiotope (139, 158), suggestingthat the emergentBc Repis in part selected by available Tc Reps.Siskind, Weksler, and colleagues investigating the basis for the qualitatively and quantitatively reducedAbresponsesin agedmice, demonstrateda shift in Id patterns of anti-hapten Absaccompanying reduced responsiveness. Then in doubletransfer experiments,they establishedthat repertoire alterations precede immunizationand are imposedon B cells from younganimals by co-existence with T cells from old mice (159-162; M. Weksler,personal communication). Aseries of experimentson the induction of Id productionuponintravenous injection of nanogramdoses of the same Id in .the absence of antigen are also interesting here as they suggest one level of Bc Rep selection whereT cell influences could be predominant:the induction of "natural" plasma cells and serum Abs(163-167). Such Id-induced responses are T cell-dependent and MHC-(and Igh-) linked; they are quantitatively poor, as comparedwith a conventional immuneresponse, and must represent enhancementof an ongoingactivity following a rise in circulating Id, since responsivenesscorrelates with Id "recurrency,"and the doses of Id injected are muchlower than its serumlevels in normal animals. Other groups carried out similar experiments, but reading out alterations inducedby Id administrationin the Id compositionof antigeninducedresponses (134, 168, 169). Also here, enhancedId expression recorded, and the establishment of "dominance"is T cell-dependent. The

Annual Reviews


228

PEREIRA ET AL

minimal doses of Id injected and the mode of administration make it plausible that similar mechanisms continuously operate in the normal immunesystems. In support of this notion, Primi et al have reported that circulating levels of Ig molecules expressing particular V.-VK combinations are controlled by MHC-linkedgenes (170), and more recent studies have also established the MHC-controlof circulating levels of rheumatoid factors (P. Pereira, manuscript in preparation). Moreinteresting perhaps are the findings on sharing of V-region determinants between circulating Igs and TcRs. These experiments represent the converse of T cell "Id-sharing," as they have used an anti-Vfl mAb (F23.1) (171) to identify circulating antibodies carrying similar idiotopes. SuchAb specificities are positively selected by T cells into the secretory B cell compartment, but only in strains which express such TcR genes (P. Pereira et al, manuscript in preparation). If confirmed by ongoing formal genetic analysis, these studies suggest TcRVfl-linkage of expressed antibody idiotypes and thus the same interpretation as the IgH-linkage of T cell idiotypes. In both cases, the selection of mimicries (and complementarities) are likely to reflect the global operation of a networkthat includes TcRs and antibodies. The network selection of Bc Reps by T cells may well involve other molecules in the "internal environment" which are necessary components of networkconnectivity (1), and selected antibodies showself-related specificities. Evidencefor T cell-dependent selection of auto-Abreactivities has recently been obtained. Huetz et al (172) demonstrated that T cells are necessary for the positive selection--in terms of frequency and "natural" activation--of B cells producting lytic antibodies to bromelain-treated syngeneic erythrocytes. Whilepractically all clonal B cell precursors with this and other auto-Abreactivities are cycling, "naturally" activated lymphocytes in normal animals (173), the spleens of nude mice contain very few such precursors, but their positive selection could be demonstrated upon transfer of T cell populations from normal syngeneic donors (172). Furthermore, transfer of adult T cells to newbornsaccelerates the establishment of this splenic auto-reactive Bc Rep from its normal development at around 6 weeksto 3 weeks of age (F. Huetz, P. Poncet, A. Coutinho, P. Portnoi, submitted). If T cells select Bc Reps and these participate in Tc Repselection, the process is circular and recursive, Recent observations directly address this principle by demonstrating T cell-dependent selection of B cells for the "education" of Tc Reps in secondary hosts (174). The experiments investigated which B cell populations were competent in the positive selection of the T. clonotype discussed above (71, 78). They established the loss T. Id expression in animals that reconstitute the B cell compartmentfrom

Annual Reviews


T CELLS

AND NETWORK

229

adult bone-marrow,that is, either after anti-/~ suppression for the first 4 weeksof life, or after lethal irradiation as adults with bone-marrowprotection. Positive selection of the THclonotype could be recovered, however, if bone-marrowreconstitution was supplemented with either peritoneal Ly-1 + B cells or splenic, Ly-1-, activated B cell blasts from normal, adult donors. Since equivalent B cell populations from nude mice failed to operate in T cell selection, it was possible to demonstrateT cell-dependent Bc Repselection by first transferring normalT cells to nude recipients and, a month later, determining the ability of B cell blasts from such animals to select the TH clonotype upon transfer to adult bone-marrowreconstituted recipients. These experiments not only give indications on the cellular basis of T-Bc Rep selection but have the heuristic value of adopting the "point of view" of the system. Thus, T cell-dependent Bc Rep selection was not directly analyzed on the basis of Abreactivities but simply by the ability of the Bc Rep to secondarily educate and "reproduce" the same Tc Rep that had originally selected them.

CELLULAR AND MOLECULAR MECHANISMS OPERATING IN REPERTOIRE SELECTION: NATURAL IMMUNE ACTIVITIES This review on a multitude of heterogeneous observations would gain coherence if someprinciples on the mechanisms(local rules) of lymphocyte selection could be laid down.Let us start with peripheral, positive selection of Reps. The first principle has to do with lymphocytepopulation dynamics: rates of production in the central lymphoid organs, life spans as competent cells, rates of decay, and peripheral renewal. No significant ongoing repertoire selection can indeed occur in the absence of extensive renewal rates in peripheral immunocompetentcells. These topics have been reviewed elsewhere (175), but it should be underlined that high renewal rates have previously been demonstrated in both the B cell and T cell compartments, though on distinct bases. B cells are constantly renewed, primarily by bone-marrowproduction of new specificities from uncommitted precursors (176, 177), while T cells are also extensively renewed, but mostly by peripheral division of immunocompetentlymphocytes (178, 179). At any time, most B cells in an adult mousehave just been produced in bone-marrow (and will soon disappear), while most cells are the progeny of mature lymphocytes that were stimulated in the periphery and thus have, or have had, available complementarities there. This major difference must be considered, however, with the quantitatively less notorious counterparts in the two lymphocyte compartments: the

Annual Reviews


230

PEREIRA ET AL

persistent B cell populations producedin perinatal life (156) and the recent thymic emigrants produced throughout life (180). Furthermore, special consideration should be given to early developmental periods (embryonic and neonatal). WhenT cell and B cell populations first comeinto contact, high levels of maternal antibodies are present, and peripheral lymphocyte populations do not turn-over but accumulate in numbers. These aspects are all fundamental to understanding "founder" effects in the recursive process of Rep selection. The second set of principles concerns the physiology of lymphocyte activation, because lymphocytesare produced as resting, noncycling cells, and the positive selection of clonal reactivities must rely uponfunctional activation to expansion and/or persistence of the target lymphocytes, by mechanisms that involve interactions with receptor V-regions. Most important, since lymphocyteactivation results in the expression of effector functions (antibody secretion, help, or suppression) positively selected lymphocytesare effector cells that establish the recursivity of the process by selectin# (positively or negatively) the complementarities that induced them. This "linkage" with effector functions of both partners is actually what is so special about V-region connectivity in contrast to receptor interactions with other molecules in the body (a clone stimulated by, say, transferin, has little chance, in turn, of stimulating (or suppressing) transferin production). Documentedexamples of such direct antigen-independent lymphocyte interactions are abundant: TH activation by (activated) B cells or antibodies, and B cell activation by effector TH(see above, and e.g. 181-186); suppressor and cytolytic T cell activation by antibodies or B cells, and suppressionof B cell function by effector T cells (see above, and e.g. 187-197; reviewed in 37, 198); Tr~ inactivation by suppressor cells (e.g. 29, 199) or CTL(13), and suppressor T cell activation by inducer cells (e.g. 88, 200; see 36, 37, 198 for reviews); mutual stimulation of (e.g. 29-34), and of B cells (see above, reviewed in 50). Of particular significance here are the findings that cross-linked Ab binding to TcRs results in stimulation of resting T cells (e.g. 57, reviewed in 201) and -that activated B cells are excellent "stimulator" cells for both helper and suppressor T cells (reviewed in 202;.see also 203-214). Conversely, resting B cells .are, by definition, the "targets" for THactivity (e.g. 71, 185, 186, 215, 216). The third class of principles should consider the kinds of molecular interactions in which TcRs and IgRs .may actually participate. There are at least four types of appropriate ligands for ,productive TcRinteractions with other V-regions: (a) processed lgV-region peptides on (self) molecules; (b) processed TcRV-regionpeptides on (self) MHC molecules; (c) directly complementary (MHC-independent) tridimensional shapes

Annual Reviews


T CELLS AND NETWORK

231

either TcRs,or (d) Ig V-regions.All these possibilities except(b) been documentedand normal immuneReps repeatedly shownto contain specificities towardssyngeneicor autologous V-regions. As originally shownby Janewayet al (217, 218), this is the case for THinteracting with processed MHC-presented peptides of IgV-regions(e.g. 72, 181-184,219221), and for unprocessedIg-V-regionsas well (e.g. 124-126, 128-131, 185, 186). Thesameapplies to suppressorTc-Ig interactions (reviewed 36, 37, 198), and to the direct (MHC-unrestricted)TcR-TcR interactions reviewedabove. In contrast, Ab"sticky ends" excel in interactions with large molecularsurfaces (222) and, therefore, with unprocessedTcRsand other Igs, although Absto peptide-MHC complexesmayactually be quite frequent (223,224; J.-P. Abastadoet al, personal communication). In essence,therefore, the molecularspecificities, the cellular mechanisms, and the population turnovers are all available in the normal immune system to provide conditions for the establishment and operation of a functional V-regionnetwork.Theaboveevidence, on the other hand, does showthat the operation of an immune networkaccounts for the selection of available lymphocyte repertoires before antigenic challenges. It follows that normalanimals should showsigns of this ongoingactivity. Pereira and colleagues have described in normal germ-free and "antigen-free" mice, before any manipulation,the existence of considerable lymphocyte activity: 10-20%of all splenic B cells, CD4+and CD8+T cells are activated, blast cells, manyof whichare in mitotic cycle (148, 225, 226). Importantly, these activated populations, in contrast with small resting lymphocytes isolated fromthe sameindividuals, are effector cells: B cells secrete IgMAbs, CD4÷Tc very efficiently "help" appropriate resting B cells into proliferation and Ig-secretion, whileCD8+ Tc directly suppress B cell responses. Becausethe quantitative levels of lymphocyte activation are quite comparablein conventionallybred and "antigen-free" individuals, this "internal activity" has beeninterpreted to indicate the operation of an immunenetwork, although no evidence has been given for the Vregion specificity of the participating lymphocytes.In turn, these observations have providedthe upper limit for the fraction of normalimmune systemwhich, at any point in time, maybe participating in a functional network. In other words, in a normal immunesystem 80-90%of all lymphocytesare not included in a network, at least in a functionally significant manner.(The above experimentsconcernexclusively splenic and peritoneal cavity lymphocytes, as "antigen-free" mice show very reduced or absent cellularity in mucosal-assoeiatedtissues and lymph nodes. In turn, together with the absenceof non-IgM isotypes, this observation showsthe very profounddeprivation of environmentalstimulation in suchmice, yet with conservedsplenic lymphocyte activity).

Annual Reviews 232

PEREIRA ET AL


A MODEL OF LOCALIZED CONNECTIVITY

LYMPHOCYTE

A simple working hypothesis was derived from the above results and considerations (227). Most immunocompetentlymphocytes produced the periphery display clonal reactivities that find no functionally productive complementarities in the molecular environment where they exist. These remain resting and decay at defined rates. In contrast, other clonal reactivities engagein functionally relevant interactions in the internal environment (with V-regions, other self "somatic" molecules or even external antigens constantly entering the body through mucosal surfaces) are activated and, by their effector functions, recruit into the set of activated lymphocytesother V-regions displaying sufficient interactive affinity (or they limit their expression, in case of CD8+ cells). Activation, on the other hand, results in positive selection and persistence of the activated clonotypes without necessarily involving extensive clonal expansion (175). This model gives rise to a series of predictions, some of which have been supported by experimental evidence and are discussed here. First, "connectivity" should be significantly higher in the "activated cell pool" than amongsmall lymphocytes; this was verified for B cells (141,142) and, indirectly, for TH(63). Connectivity, however, should not be limited to regions but should extend to other structural componentsavailable in the organism. In other words, the repertoire of "activated lymphocytes" should be biased for "self-reactivities." This has been very extensively demonstrated for B cells. While small B cell repertoires appear quite "burnettian" in all analyses of clonal precursors, "large" B cells from normal individuals contain autoreactive specificities in high frequencies (172, 173), and "natural Abs"are essentially a collection ofauto-Abs (228230). Also in the case ofT cells, recent experiments showthe preference of "large" CD4+ and CDS+lymphocytes to functionally interact with "unmodified" target (activated) B cells (A. Bandeira, P. Pereira, Coutinho, C. Martinez-A., manuscript in preparation). In contrast, small T cells are completely deprived of auto-reactivities and very enriched for allo-specificities. Finally, such "localized connectivity" wouldimply that an immunenetwork predominantly (or exclusively) operates in the compartment of activated cells, while the repertoires of small lymphocytesof normal individuals should showlittle evidence of network selection. This point has also been extensively substantiated by experiments. For example, IgH-linked, THclonotypes that are selected by B cell/Ig-dependent mechanisms are found in the activated T cell pool of normal animals, while TH "sharing" antibody idiotypes by aleatory cross-reactivities are small cells

Annual Reviews


T CELLS ANDNETWORK 233 (63); "large" T cells from normal donors positively select auto-reactive cells, while small T cells from the same donors do not (172); B cells that positively select Tu clonotypes are "large" in normal mice, and T cells that "educate" Bc Reps for recursive THRepselection are also in the activated pool (174); the TcRVfl-linked expression of antibody idiotypes is only detected amongactivated B cells and "natural antibody," and all IgHidentical mousestrains tested contain small B cells carrying this Id (P. Pereira, C. Pefia-Rossi, S. Petersson, C. Martinez-A., A. Coutinho, manuscript in preparation). The model, therefore, makes predictions on which Ids will be (or will not be) "shared" by T cells and B cells, and on which kinds of reactivities are (or are not) the result of networkselection. This is compatible with the finding that anti-#-suppressed mice show Tc Reps that, only in some regards, are different from normal animals (e.g. 231234). A second prediction of the model is that immuneresponses to "external" antigens predominantly arise from the small lymphocyte pool, precisely because such molecular profiles are absent from the internal environment. Clonal reactivities in the "large" cell pool, even if complementaryto an immunizingantigen, will treat it as a self-molecule, and consequentlytheir "response" will merely represent a "compensation" in previously ongoing activities. Mostimportantly, because of the different levels of connectivity in either set of lymphocytes,activities in the large cell pool will alwaysbe "self-limited" and rich in "nonspecific" and degenerate components, in contrast with the high-titered specific immuneresponses induced in small lymphocyteswhich, being disconnected, are free from systemic constraints in their clonal expansion. (It is likely, however,that such clonal responses will subsequently fall under "network" control, for the very significant alterations in the composition of the molecular environment brought about by clonal amplifications.) This prediction of the modelhas been verified for B cells in two ways. The kinetics and magnitude of a specific T cell-independent immune response are muchfaster and higher if the responding clones are in the small cell pool (235); the dynamicsof "natural" antibody production, and of its compensations to perturbations in the network they form, is very different from immuneresponse dynamics, actually displaying characteristics of "chaotic" attractors (236). For T cells, however,only indirect evidence is available. Animalsthat are neonatally tolerized to semi-allogeneic cells, maintain high levels of lymphocyteactivity throughout life (237), and their activated T cell pool contains the alloreactive specificities to the tolerized haplotype, which are found amongsmall cells in normal individuals (238; P.-H. Lambert, personal communication). The reported examplesof "split" tolerance (e.g. in mice grafted with allogeneic thymic

Annual Reviews


234

PEREIRA ET AL

epithelium; 239) between in vivo behaviors and specificities revealed in vitro, can also be explained in this context. The same principles could apply to connectivity amongTcRs, a point suggested by the report of Wilson (240) that F1 animals neonatally challenged with parental T cells, cannot mount, as adults, GvH-resistance responses to the same TcRs (see above). This discussion brings us to a few points concerning tolerance (and responsiveness) in the context of this model. Clearly, if a part of normal immunesystem is organized as a high connectivity network which by "trapping" autoreactive cells limits their participation in clonal immune responses, this organization could be responsible for the establishment of some kinds of tolerance and low-responsiveness. This notion provides a satisfactory explanation for self-reactivities in the B cell compartment (241), and it is tempting to apply it to T cells as well. Thus, regardless the extent of intra-thymic clonal deletion (242-244) and/or other forms selection that limit functional responsiveness of T ceils (245), it is inescapable that someform of "tolerance" must operate in post-thymic, peripheral T cells. This modelconsiders peripheral self-tolerance as a systemic, "distributed" property, determined by the particular organization in which auto-reactive T cells are necessarily included by their normal reactivity in the molecular environmentof self. Since recursive repertoire selection is established early in ontogeny, the process predominantly applies to the antigenic universe available at that time (and thus, the ease in inducing "neonatal tolerance"). The process is strongly influenced by maternally derived molecules, but it maintains the "memoryof self" (physiological auto-reactivity) by recursivity. The model, therefore, explains the "learning" of self on the basis of normal lymphocytereactivities, an advantage to clonal models of self-tolerance which often have to invoke ad hoc postulates in this respect. Furthermore, it accommodatesin the same frameworkthe evidence reviewed here on V-region-directed T cell autoreactivities, with thc repeated observations on the existence, in normal individuals, of potentially aggressive T cell reactivities to autologous "somatic" proteins (246-251). Finally, it readily explains the finding that severe mutilations of Tc Reps, by makingit impossible to establish such a self-related network of high connectivity, may well (but paradoxically, for other interpretations) result in pathological autoimmunity(252254). Since inclusion in the network limits clonal responsiveness, somecases of genetically controlled "low responder" phenotypes could have a connectivity explanation. If the appropriate T cell specificities were brought into high connectivity because of receptor properties (such as idiotypic) other than binding to self antigen-MHC,alterations in the network struco

Annual Reviews


T CELLS

AND NETWORK 235

ture, mayrelease those clones from connectivity and shift responder phenotypes. Such an examplehas recently been reported. In contrast with their H-2k congenics, BALB/cmice are nonresponders to low density TNP-self modifications. They predominantly "respond" to highly TNP-derivatized self with Tn that are selected by Bc/Ig-dependent mechanisms to the compartment of activated cells and to "share" an antibody idiotype (63, 71, 78). However, BALB/cmice that had their network altered immunization with anti-Id mAbs,particularly if early in ontogeny, "disconnect" TNP-self-reactive TH, show muchhigher clonal frequencies of THreacting with low TNP-self densities, and acquire a "high-responder" phenotype characteristic of H-2k mice (255; Martinez-A., unpublished). surprising (manipulation of an MHC-linked, low responder idiotype results in a "high responder" phenotype), these results can be readily explained on the basis of the IgH-linked expression of the Tn clonotype in question. Although this separation of two compartments in the normal immune system might have operational and theoretical advantages, it should be pointed out how unclear (at least in time) their frontiers must be. Moreover,while the~view of peripheral "self-tolerance" as a distributed, supraclonal property emergingfrom connectivity is central in this model, it is likely that other (clonal) properties contribute to the functional characteristics of "internally" activated lymphocytes. Werefer to a number of observations from several independent groups (256-263) describing the induction of unresponsiveness (tolerance) in mature, antigen-MHC-activated T cells by exposure to high concentrations of TcR-specific ligands, such as free peptides, peptide-MHCcomplexes on lipid vesicles or nonstimulatory presenting cells, soluble anti-TeR or anti-CD3 antibodies. Interestingly, similar unresponsiveness, which lasts for days or weeks, is induced by "appropriate" restimulation of responding T cells some hours after the responses were initiated. These are all conditions that might apply to immunocompetentT cells reacting to autologous structures. In those experiments, unresponsiveness was assessed by the inability to proliferate and, in somecases, by loss of IL-2 production,~but.effector functions such as B cell-directed help or suppression were not evaluated. It is striking that such "unresponsive" T cells maybe functionally equivalent to the "naturally" activated effector T cells isolated from the spleen of normal individuals. In spite of their excellent performance in functional assays, these cells are resistant to growth induction in vitro, express few high affinity IL-2 receptors, and produce little IL-2 (264). Even if "tolerant," therefore, such cells are perfectly competent to operate in the immunenetwork. Regulation of activated,B cell performance by specific ligand binding has been known

Annual Reviews

236

PEREIRA ET AL

for manyyears (265), and this is likely to play a determinant role in the networkdynamicsof B cells and antibodies (2). It is an exciting possibility that equivalent rules apply to T cells, as it wouldprovide new functional perspectives on lymphocyteconnectivity and further possibilities to address autoreactivity.


CONCLUDING

REMARKS

Since the mid 1970s, immunology has been confronted with two major developments. If the progress in molecular and cell biology continues to solve many questions about the structure and function of isolated components in the immunesystem, the notion that the system owes some of its properties to a network organization has set limits to the kinds of problems that can be approached by component analysis. These two aspects of the discipline are often considered alternatives. Wronglyso. The unique goal is to understand the origin of the cognitive properties of the immunesystem: learning and memory,self-assertion, discrimination betweenprotective responsiveness and tolerance. Andif the structure of a gene and its regulatory elements cannot give us the clues for the process of learning, it is as naive to study "connectivity" without considering its structural basis, the regulation of the genes encodinginteracting molecules, or the physiology of the cells producing them. In short, for the understanding of global behaviors in a viable system, local rules describing individual componentsare as essential as the structure and dynamics of connectivity. Considering its achievements to date, it is clear that the network idea did not quite work in immunology, but the divorce of its practice from quantitation and local rules is only one of the reasons for its failure. More importantly, perhaps, has been the lack of an explicit framework with testable predictions, which introduces the qualitative step from clonal to network thinking and thus departs from some of the notions associated with the clonal selection theory. Thus, most of the so-called networkexperiments have been concerned with (idiotypic) regulation of clonal components in immuneresponses, ignoring the formal incompatibility between those two views of the immunesystemand failing to explore what the network idea can really contribute: the globality of the system’s operation (1, 2, 22, 266). Beyond the dismay to which such "pseudo-network" approaches have brought the network idea, the consequences of these confusions are very apparent. In spite of the marked contributions from molecular and cellular immunology, little progress has been achieved in auto-immune diseases, transplantation tolerance, allergy, and chronic infectious diseases. These represent obvious modifications in systemic behaviors, which cannot

Annual Reviews


T CELLS AND NETWORK

237

be explained and will find no solution on the basis of clonal properties. The recent tendency to study immunephysiology and normal individuals-rather than immuneresponses--will perhaps be the "road to Damascus" for modern immunology. The object of this review is a prime exampleof this state of affairs. Ever since the original observations of Simonsen,Tc Reps have never ceased to display surprising characteristics. Alloreactivity, restriction, Ir-gene effects, self-tolerance, and thymic education are all essential problemsin basic and applied immunologythat remain to be solved. Yet, we know today nearly as much about the genetics and structure of TcRs, as we know of antibodies. Even if further clues might be obtained from TcR crystals, wedo not believe that they will give us the solution to the selection of Tc Reps. First, this is because the unit-target of selection is not a molecule but a whole cell, the physiology of which involves manyother interacting molecules and ligands (as it is already understood for a number of accessory T cell surface structures and hormonereceptors). Second, this is because the process of selecting a diverse repertoire has also to do with multiple interactions that arise within that repertoire as a necessary consequenceof diversity. This simple notion that the targets and effectors of selection belong in one and same set suggests some degree of "autonomy"in the process, and it imposes the consideration of the "whole" on the determination of the "parts." Undoubtedly, the environment (MHC, self and not-self antigens) where the repertoire develops, has a very significant contribution to selection. As in other complexsystems, however, it is likely that idiotypes find their wayof being (select repertoires) within alternatives that are viable for their own composition in the "medium" where they exist. Viable repertoires are then determinant for the manners by which they evolve to alterations in the mediumor in their own composition. In short, if there is some truth in the network idea, lymphocyte repertoires are self-determined within the viability boundaries set by their genetic possibilities, the rules of lymphocytephysiology, and the molecular environment where they develop (267). It follows that the emergent "solutions" for repertoires probably cannot be understood solely by the analysis of single clones and their relationships with environmental molecules but will require the description of the others and of their interactions. ACKNOWLEDGMENTS

Wethank A. Bandella for preparation of the manuscript, our colleagues for discussions and unpublished data, and DRET(France) and CAYCYT (Spain) for support.

Annual Reviews 238

PEREIRA

ET AL


Literature Cited 1. Vaz, N., Martinez-A., C., Coutinho, A. 1984. The uniqueness and boundaries of the idiotypic self. 1984. In ldiotypy in Biology and Medicine, ed. H. K6hler, P.-A. Cazenave, J. Urbain, p. 43. New York: Academic Press 2. Varela, F., Coutinho, A., Dupire, B., Vaz, N. 1988. Cognitive networks: immune, neural and otherwise. In Theoretical Immunology,Vol. III, ed. A. S. Perelson, p. 359. New York: Addison-Wesley 3. Sim, G. K., Augustin, A. A. 1983. Internal images of major histocompatibility antigens on T-cell receptors and their role in the generation of the T-helper cell repertoire. Ann. NY Acad. Sci. 418:272 4. Sim, G. K., MacNeil, I. A., Augustin, A. A. 1986. T helper cell receptors: idiotypes and repertoire. Immunol. Rev. 90:49 5. Suciu-Foca, N., Rohowsky, C., Kung, P., King, D. W. 1982. Idiotype-like determinants on humanT lymphocytes alloaetivated in mixed lymphocyteculture. J. Exp. Med. 156:283 6. Suciu-Foca, S., Rohowsky, C., Kung, P., Lewinson, A., Nicholson, J., Reemtsma, K., King, D. W. 1983. MHC-specifie idiotypes on alloactivated humanT cells: in vivo and in vitro studies. Transplant. Proe. XV:784 7. Nagarkatti, P. S., Nagarkatti, M., Kaplan, A. M. 1985. Normal Lyt1 + 2-T cells have the unique capacity to respond to syngeneic autoreactive T cells. Demonstration of a T cell network. J. Exp. Med. 162:375 8. Kennedy, D. W., Russo, C., Kim, Y. T., Weksler, M. E. 1986. T cell proliferation induced by anti-self-I-Aspecific T cell hybridomas. Evidence of a Tcell network. J. Exp. Med. 164:490 9. Suzuki, H., Evavold, B., Swartz, T. J., Latta, S. L., Quintans, J. 1986. The syngeneic T-T lymphocyte reaction (STTLR). I. Induction of primary anti-T cell proliferative responses in T cell cultures stimulated with self- and antigen-reactive T cells. J. Mol. Cell. Immunol. 2:331 10. Woodland, R. T., Wilson, D. B. 1977. The induction of specific resistance in F~ hybrid rats to local graft-versus-host reactions: nature of the eliciting cell. Eur. J. Immunol. 7:136 I1. Bellgrau, D. L., Wilson, D. B. 1978. Immunologicalstudies on T cell receptors. I. Specifically induced resistance to graft-versus-host disease in rats

12.

13.

14.

15.

16.

17.

18.

19.

20. 21.

22.

mediated by host T cell immunity to alloreactive parental T cells. J. Exp. Med. 148:103 Bellgrau, D., Wilson, D. B. 1979. Immunologicalstudies of T-cell receptors. II. Limited polymorphismof idiotypic determinants on T-cell receptors specific for major histocompatibility complex alloantigens. J. Exp. Med. 149:234 Kimura, H., Wilson, D. B. 1984. Antiidiotypic cytotoxic T cells in rats with graft-versus-host disease. Nature 308: 463 Kimura, H., Pickard, A., Wilson, D. B. 1984. Analysis of T cell populations that induce and mediate specific resistance to graft-versus-host disease in rats. J. Exp. Med. 160:652 De Waal, L. P., Nathenson, S. G., Melief, C. J. M. 1983. Direct demonstration that cytotoxic T lymphocytes recognize conformational determinants and not primary amino acid sequences. J. Exp. Med. 158:1720 Bluestone, J. A., Foo, M., Allen, H., Segal, D., Flavell, R. A. 1985. Allospecific cytolytic T lymphocytesrecognize conformational determinants on hybrid mousetransplantation antigens. J. Exp. Med. 162:268 Berzofsky, J. A., Schechter, A. N., Shearer, G. M., Sachs, D. H. 1977. Genetic control of the immune response to staphylococcal nuclease. III. Time course and correlation between the response to native nuclease and the response to its polypeptide fragments. ,1. Exp. Med. 145:111 Corradin, G., Chiller, J. M. 1979. Lymphocytespecificity to protein3 antigens. II. Fine specificity of T-cell activation with cytochromec and derived peptides as antigenic probes. J. Exp. Med. 149: 436 Towsend, A. R. M, Rothbard, J. B., Gotch, F. M., Bhadur, G., Wraith, D., McMichael,A. J. 1985. The epitopes of influenza nucleoprotein recognized by cytotoxic T lymphocytes can be defined with short synthetic peptides. Cell 44: 959 M611er,G., ed. 1987. Antigenic requirements for activation of MHC restricted responses. Immunol. Rev. 98:1-187 Binz, H., Frischknecht, H., Mercoli, C., Wigzell, H. 1978. Partial characterization of cell surface idiotypes on alloantigen-activated T lymphoblasts. Scand. J. Immunol. 7:481 Coutinho, A., Forni, L., Holmberg,

Annual Reviews


T CELLS AND NETWORK D., Ivars, F., Vaz, N. 1984. Froman antigen-centered, clonal perspective of immuneresponses to an organismcentered, networkperspectiveof autonomousactivity in a self-referential immunesystem. Immunol.Rev. 79:151 23. Holmberg,D., Forsgren, S., Forni, L., Ivars, F., Coutinho,A. 1984.Idiotypic determinants of natural IgM antibodies whichresembleself Ia-antigens. Proc. Natl. Acad. Sci. USA81:3175 24. Hoffmann,G. W., Cooper-Willis, A., Chow, M. 1986. A new symmetry: a anti-B is anti-(B anti-A), and reverse enhancement.J. Immunol.137:61 25. de la Hera, A., Toribio, M. L., Marquez, C., Marcos, M. A. R., Cabrero, E., Martinez-A., C. 1986. A hypothesisfor the selectionof available repertoires: T cell networkearly in the intrathymicdifferentiation. Seand.J. Immunol.24:637 26. Coutinfio, A. 1980. The self-nonself discrimination and the nature and acquisition of the antibodyrepertoire. Ann. Immunol.(Inst. Pasteur) 132C: 131 27. Coutinho,A., Forni, L., Holmberg, D., Ivars, F. 1983. Is the networktheory tautologic? In Nobel Symposium55: Geneticsof the Immune Response,ed. E. M611er,G. M611er,p. 273. NewYork: Plenum 28. Herzenberg,L. A., Okumura,K., Cantor, H., Sato, V.L., Shen,F.-W.,Boyse, E. A., Herzenberg,L. A. 1976. T cell regulation of antibodyresponses: demonstrationof allotype-specifichelper T cells andtheir specific removalby suppressor T cells. J. Exp. Med.144:330 29. Lamb,J. R., Feldmann, M. 1982. A hu~nansuppressor T cell clone which recognizesan autologoushelper T cell clone. Nature 300:456 30. Sun, D., Qin, Y., Chluba,J., Epplen, J. T., Wekerle,H. 1988. Suppression of experimentally induced autoimmune encephalomyelitisby cytolytie T-Tcell interactions. Nature332:843 31. Herzenberg,L. A., Black, S. J., Herzenberg, L. A. 1980. Regulatorycircuits and antibodyresponses. Eur. J. Immunol.10:1 32. McNamara, M., K6hler, H. 1983. Idiotype-recognizing T helper cells that are not idiotype-specific.J. Exp. Med.158: 811 33. McNamara, M., Gleason, K., K6hler, H. 1984. T-cell helper circuits. Immunol. Rev. 79:87 M., K6hler, H., Smyk,S. 34. McNamara, 1984.Idiotype-specificT cells. Rolein regulation. In The Biologies of Idio-

239

types, ed. M. I. Greene,A. Nisonoff, p. 341. NewYork: Plenum 35. Germain,R. N., Benacerraf,B. 1981.A single major pathwayof T lymphocyte interactions in antigenspecific immune suppression. Scand.J. Immunol.13:1 36. Dorf. M. E., Okuda, K., Minami,M. 1982. Dissection of a suppressor cell cascade. Curr. Top. Microbiol. Immunol. 100:61 37. Green, D. R., Flood, P. M., Gershon, R. K. 1983. ImmunoregulatoryT-cell pathways. Ann. Rev. Immunol.1:439 38. Lowy,A., Monroe,J., Royer, H.-D., Greene,M.I. 1984. Therole ofidiotype in T-cell regulatory events. In The Biologyofldiotypes, ed. M. I. Greene, A. Nisonoff, p. 491. NewYork:Plenum 39. Eichmann,K., Falk, I., Melchers,F., Simon,M.M. 1980. Quantitative studies on T cell diversity. I. Determination of the precursor frequencies for two types of StreptococcusA-specific helper cells in nonimmune,polyclonally activated splenic T cells. J. Exp. Med. 152:477 40. Goronzy,J., Schaefer, U., Eichmann, K., Simon,M. M. 1981. Quantitative studies on T cell diversity. II. Determination of the frequencies and Lyt phenotypesof two types of precursor cells for alloreactivecytotoxicT cells in polyclonallyandspecifically activated splenic T cells. J. Exp. Med.153:857 41. Melchers, I., Fey, K., Eichmann,K. 1982. Quantitative studies on T cell diversity.III. Limitingdilutionanalysis of precursorcells for T helpercells.reactive to xenogeneic erythroeytes.J. Exp. Med. 156:1587 42. Fey, K., Melchers, I., Eichmann,K. 1983. Quantitative studies on T cell diversity. IV. Mathematical analysis of multiplelimiting populationsof effector andsuppressorT cells. J. Exp. Med. 158:40 43. Cooper, J., Eichmann,K., Fey, K., Melchers,I., Simon,M.M., Weltzien, H. U. 1984. Networkregulation among T cells: qualitative and quantitative studies on suppression in the nonimmunestate. Immunol.Rev. 79:63 44. Fey, K., Simon,M.M., Melchers,I., Eichmann,K. 1984. Quantitative estimates of diversity, degeneracy, and connectivity in an idiotypic network amongT cells. SeeRef. 38, p. 261 45. Andersson,J., Coutinho,A., Melchers, F. 1977. Frequenciesof mitogen-reactive Bcells in the mouse.II. Frequencies of B cells producingantibodies which lyse sheep or horse erythrocytes, and lrinitrophenylated or nitroiodo-

Annual Reviews


240

PEREIRA ET AL

phenylylated sheep erythrocytes. J. Exp. Med. 145:1520 46. Cohen, I. 1986. Regulation of autoimmunedisease: physiological and therapeutic. Immunol. Rev. 94:5 47. Binz, H., Wigzell, H. 1977. Antigenbinding, idiotypic T lymphocytereceptors. Contemp.Top. lmmunobiol. 7: 113 48. Rajewsky, D., Eichmann, K. 1977. Antigen receptors of T helper cells. Contemp. Top. Immunobiol. 7:69 49. Eichmann, K. 1978. Expression and function of idiotypes on lymphocytes. Adv. Immunol. 26:195 50. Rajewsky, K., Takemori, T. 1983. Genetics, expression and function of idiotypes. Ann. Rev. Immunol. 1:569 51. Jansenius, J. C., Williams, A. F. 1982. The T lymphocyte antigen receptor-paradigm lost. Nature 300:583 52. Miller, J. F. A. P., Morahan, G., Walker,I. D. 1983. T cell antigen receptors: fact and artefact. Immunol. Today 4:141 53. Coutinho, A., Meo, T. 1983. Immunoglobulin gene expression by T lymphocytes. Scand. J. Immunol. 18:79 54. Allison, J. P., McIntyre, B. W., Bloch, D. 1982. Tumor-specific antigen of murine T-lymphoma defined with monoclonal antibody. J. Immunol. 129:2293 55. Meuer, S. C., Fitzgerald, K. A., Hussey, R. E., Hodgdon, J. C., Schlossman, S. F., Reinherz, E. L. 1983. Clonotypic structures involved in antigenspecific human T cell function. Relationship to the T3 molecular complex. J. Exp. Med. 157:705 56. Haskins, K., Kubo, R., White, J., Pigeon, M., Kappler, J., Marrack, P. 1983. The major histocompatibility complex-restricted antigen receptor on T cells. I. Isolation with a monoclonal antibody. J. Exp. Med. 157:1149 57. Kaye, J., Porcelli, S., Tite, J., Jones, B., Janeway, C. A., Jr. 1983. Both a monoclonal antibody and antisera specific for determinants unique to individual cloned helper T cell lines can substitute for antigen and antigen-presenting cells in the activation of T cells. J. Exp. Med. 158:836 58. Samelson, L. E., Germain, R. N., Schwartz, R. H. 1983. Monoclonal antibodies against the antigen receptor on a cloned T-cell hybrid. Proc. Natl. Acad. Sci. USA 80:6972 59. Hedrik, S. M., Cohen, D. I., Nielssen, E. A., Davies, M. M. 1984. Isolation of cDNAclones encoding T cell-specific membrane-associated proteins. Nature 308:149

60. Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A. C., Eisen, H. N., Tonegawa, S. 1984. A third rearranged and expressed gene in a clone of cytotoxic T lymphocytes. Nature 312:36 61. Coutinho, A., Forni, L. 1984. Beware of idiotypes. J. Mol. Cell. Immunol.1: 345 62. Araujo, P. M. F., Holmberg, D., Martinez-A., C., Coutinho, A. 1987. Idiotypic multireactivity of "natural" antibodies: "natural" anti-idiotypes also inhibit helper cells with cross-reactive clonotypes. Scand. J. Imrnunol. 25:497 63. Coutinho, A., Marquez, C., Araujo, P. M. F., Pereira, P., Coutinho,A., Toribio, M. L., Marcos, M. A. R., Martinez-A., C. 1987. A functional idiotypic network of T helper cells and antibodies limited to the compartment of "naturally" activated lymphocytes in normal mice. Eur. J. Immunol. 17: 821 64. Oudin, J., Cazenave, P.-A. 1971. Similar idiotypic specificities in immunoglobulin fractions with different antibody function or even without detectable antibody function. 1"roe. Natl. Acad. Sci. USA 68:2616 65. Cazenave, P.-A., Ternynck, T., Avrameas, S. 1974. Similar idiotypes in antibody-forming cells and in cells synthesizing immunoglobulins without detectable antibody function. Proc. Natl. Acad. Sci. USA71:4500 66. Eichmann, K., Coutinho, A., Melchers, F. 1977. Absolute frequencies of Lipopolysaccharide-reactive B cells producing A5Aidiotype in unprimed, streptococcal A carbohydrate-primed, anti A5Aidiotype-sensitized and antiA5Aidiotype-suppressed A/J mice. J. Exp. Med. 146:1436 67. Ramseier, H., Aguet, M., Lindemann, J. 1977. Similarity of idiotypic determinants of T- and B-lymphocytereceptors for alloantigens. Immunol.Rev. 34: 50 68. Tsuji, M., Yagi, J., Abe, R., Asano, Y., Tada, T. 1987. Shared idiotopes between anti-class II monoclonalantibodies and major histoeompatibility complex recognition J. Immunol. 17:675site of T cells. Eur. 69. Weinberger, J. Z., Germain, R. N., Ju, S.-T., Greene, M. I., Benacerraf, B., Doff, M. E. 1979, Hapten-specific Tcell responses to 4-hydroxy-3-nitrophenyl acetyl. 11. Demonstration of idiotypic determinants on suppressor T cells. J. Exp. Med. 150:761 70. Bach, B. A., Greene, M. I., Benacerraf, B., Nisonoff, A. 1979. Mechanismsof

Annual Reviews


T CELLS AND NETWORK regulation of cell-mediated immunity. IV. Azobenzenearsonate-specific suppressor factor(s) bear cross-reactive idiotypic determinants the expression of which is linked to the heavy-chain allotype linkage group of genes. J. Exp. Med. 149:1084 7 I. Martinez-A.,C., Pereira, P., de la Hera, A., Bandeira, A., Marquez. C., Coutinho, A. 1986. Thebasis for major histocompatibility complex (MHC) and immunoglobulingene control of helper T cell idiotopes. Eur. J. ImmunoL16: 417 72. Cosenza, H., Julius, M. H., Augustin, A. A. 1977. Idiotypes as variable region marker: analysis between receptors on phosphorylcholine-specific T and B lymphocytes. Immunol. Rev. 34:3 73. Eichmann, K., Ben-Neriah, Y., Hetzelberger, D., Polke, C., Girol, D., Lonai, P. 1980. Correlated expression of Vr~ framework and V~ idiotypic determinants on T helper cells and on functionally undefined T cells binding group A streptococcal carbohydrate. Eur. J. lmmunol. 10:105 74. Harvey, M. A., Adorini, L, Miller, A., Sercarz, E. E. 1979. Lysozyme-induced T-suppressor cells and antibodies have a predominant idiotype. Nature 281: 594 75. Miller, G. G. P., Nadler, P. I., Asano, Y., Hodes, R. J., Sachs, D. H. 1981. Induction of idiotype beating, nuclease-specific helper cells by in vivo treatment with anti-idiotype. J. Exp, Med. 154:24 76. Epstein, S. L., Bluestone, J. A., Sachs, D. H. 1984. Expression of anti-MHC idiotypes in immune responses. See Ref. 38, p. 279 77. Bluestone, J. A., Leo, O., Epstein, S. L., Sachs, D. H. 1986. Idiotypic manipulation of the immuneresponse to transplantation antigens. Irnmunol. Rev. 90:5 78. Martinez-A., C., Pereira, P., Bernabr, R. R., Bandeira, A., Larsson, E.-L., Cazenave, P.-A., Coutinho, A. 1984. Internal complementarities in the immune system: regulation of the expression of helper T-cell idiotypes. Proc. Natl. Acad. Sci. USA81:4520 79. Martinez-A., C., Bragado, R., de la Hera, A., Toribio, M. L., Marcos, M. A. R., Bandeira, A., Pereira, P., Coutinho, A. 1986. Functional and biochemical evidence for the recognition of T cell receptors by monoclonal antibodies to an immunoglobulinidiotype. J. Mol. Cell. Immunol. 2:307 80. Martinez-A., C., Coutinho, A., Ban-

241

deira, A., de la Hera, A., Toribio, M. L., Marcos, M. A. R., Pereira, P. 1987. Sporadic idiotypic cross-reactivities between antibodies and T helper cells: one example of aleatory expression ofT cell idiotypes. J. Moll. Cell. lmmunol. 3:21 81. Martinez-A.,C., Pereira, P., de la Hera, A., Toribio, M. L., Marcos, M. A. R., Marquez, C., Coutinho, A. 1987. A commonidiotope on T cell receptors and antibodies expressed in the absence of network selection. Eur. J. Immunol. 17:1391 82. Binz, H., Wigzell, H., Bazin, H. 1976. T cell idiotypes are linked to immunoglobulin heavy chain genes. Nature 264:639 83. H~immerling, G. J., Black, C., Berek, C., Eichmann, K., Rajewsky, K. 1976. Idiotypic analysis of lymphocytes in vitro. II. Genetic control of T helper cell responsiveness to anti-idiotypic antibody. J. Exp. Med. 143:861 84. Krammer, P., Eichmann, K. 1977. T cell receptor idiotypes are controlled by genes in the heavy chain linkage group and the major histocompatibility complex. Nature 270:733 85. Rubin, B., Hertel-Wulff, B., Kumura, A. 1979. Allo-antigen-specific idiotype bearing mouseT cells. I. Specificity characterization and genetic association with the heavy chain IgG allotype. J. Exp. Med. 150:307 86. Ben-Neriah, Y., Girol, D., Lonai, P., Simon, M. M., Eichmann, K. 1980. Allotype-linked genetic control of a polymorphic VH framework determinant on mouse T-helper cell receptors. Nature 285:257 87. Weinberger, J. Z., Greene, M. I., Benacerraf, B., Doff, M. E. 1979. Haptenspecific T cell responses to 4-hydroxy3-nitrophenyl acetyl. Genetic control of delayed-type hypersensitivity by VH and I-A region genes. J. Exp. Med. 149: 1336 88. Eardley, D. D., Shen, F. W., Cantor, H., Gershon, R. K. 1979. Genetic control of immunoregulatory circuits. Genes linked to the Ig locus govern communication between regulatory Tcell sets. J. Exp. Med. 150:44 89. Sy, M.-S., Lowy, A., HayGlass, K., Janeway, C. A. Jr., Gurish, M., Greene, M. I., Benacerraf, B. 1984. Chronic treatment with rabbit anti-mouse /tchain antibody alters the characteristic immunoglobulin heavy-chain restriction of murine suppressor T-cell factors. Proc. Natl. Acad. Sci. USA81: 3846

Annual Reviews


242

PEREIRA ET AL

90. Strosberg, A. D. 1984. Anti-idiotypic antibodies as immunological internal images of" hormones. In Idiotypy in Biology and Medicine, ed. H. K6hler, J. Urbain, P.-A. Cazenave, p. 365. New York: Academic Press 91. Pillemer, E., Weissman,I. L. 1981. A monoclonal antibody detects a VHTEPC15idiotypic determinant crossreactive with a Thy-1 determinant. J. Exp. Med. 153:1068 92. Volanakis, J. E., Kearney, J. F. 1981. Cross-reactivity between C-reactive protein and idiotypic determinants on a phosphocholine-binding murine myelomaprotein. J. Exp. Med. 153: 1604 93. Coutinho, A., Forni, L, Bernab6, R. R. 1980. The polyclonal expression of immunoglobulinvariable region determinants of the membrane of B cells and their precursors. Springer Semin. Immunopathol. 3:171 94. L’Age-Stehr, J. 1980. Priming of T helper cells by antigen-activated B cells. B cell-primed Lyt- 1 + helper cells are restricted to cooperate with B cells expressing the IgVHphenotype of the priming B cells. J. Exp. Med. 153:1236 95. HayGlass,K. T., Benacerraf, B. J., Sy, M.-S. 1986. T cell development in B cell-deficient mice. V. Stopping anti-# treatment results in Igh-restricted expansion of the T suppressor cell repertoire concomitant with the development of normal immunoglobulin levels. J. Exp. Med. 164:36 96. HayGlass, K. T., Naides, S. J., Benacerraf, B., Sy, M.-S. 1985. Suppressor T-cell factor(s) display an altered pattern of Igh (immunoglobulin heavy chain locus) genetic restriction when developed in an Igh-congeneic host. Proc. Natl. Aead. Sei. USA82Z: 2133 97. Bottomly, K., Maurer, P. H. 1980. Antigen-specific helper T cells required for dominant production of an idiotype (ThId) are not under immuneresponse (Ir) gene control. J. Exp. Med. 152: 1571 98. Bottomly, K., Mosier, D. E. 1981. Antigen-specific helper T cells required for dominant idiotype expression are not H-2 restricted. J. Exp. Med. 154: 411 99. Janeway, C. A. Jr. 1984. The role of idiotype and of immunoglobulin in Tcell differentiation and function. See Ref. 38, p. 349 100. Bottomly, D., Dunn, E. 1984. Helper T-lymphocyte influences on idiotypebearing B cells. See ref. 38, p. 365 101. McDougal, J. S., Shen, F. W., Cort,

S. P., Bard, J. 1982. TwoLy-I T helper cell subsets distinguished by Qal phenotype: the priming environment determines whether one or both subsets will be generated. J. Exp. Med. 155:831 102. Rohrer, J. W., Gershon, R. K., Lynch, R. G., Kemp, J. D. 1983. Enhancement of B lymphocyte secretory differentiation by a Lyl +,2-,Qa-I + helper T cell subset that sees both antigen and determinants on immunoglobulin. J. MoLCell. lmmunoL 1:50 ’103. Van Wauwe, F. P., DeMay, J. R., Groosener, J. G. 1980. OKT3:a monoclonal anti-human T lymphocyte antibody with potent mitogenic properties. J. Immunol. 124:2708 104. Schwartz, R. H. 1985. T-lymphocyte recognition of antigen in association with gene products of the major histocompatibility complex. Ann Rev. Immunol. 3:237 105. Sherman, L. A. 1980. Dissection of the B10.D2 anti-H-2k b cytolytic T lymphocyte receptor repertoire. J. Exp. Med. 151:1386 106. Sherman, L. A. 1982. Influence of the major histocompatibility complex on the repertoire of allospecific cytolytic T lymphocytes. J. Exp. Med. 155:380 107. Sherman, L. A. 1982. Genetic and regulatory contributions of the major histocompatibility complex to the developing cytolytic T lymphocyte repertoire. J. Immunol. 128:1849 108. Sherman, L. A. 1985. Cytolytic T lymphocyte receptor repertoire of H-2 disparate cells obtained from double parent chimeras. J. Immunol. 134:1330 109. Sherman, L. A. 1982. Genetic linkage of the cytolytic T lymphocyte repertoire and immunoglobulin heavy chain genes. J. Exp. Med. 156:294 110. Sherman, L. A., Riblet, R. 1985. Comparison of the H-2Kb-specific cytolytic T lymphocyte receptor repertoire in IgH recombinant strains. J. Irnmunol. 134:3569 111. Sherman, L. A., Maleckar, J. R. 1988. Genetic and environmental regulation of the cytolytic T lymphocytereceptor repertoire specific for alloantigen. Immunol. Rev. 101:115 112. Lawton, A. R. III, Asofsky, R., Hylton, M. B., Cooper, M. D. 1972. Suppression of immunoglobulin class synthesis in mice. I. Effects of treatment with antibody to/t-chain. J. Exp. Med. 135:277 113. Cerny, A., Heusser, C., Sutter, S., Huegin, A. W., Bazin, H., Hengartner, H., Zinkernagel,R. M. 1986. Generation of agammaglobulinaemic mice by pre-

Annual Reviews


T CELLS natal and postnatal exposure to polyclonal or monoclonal anti-IgM antibodies. Scand. J. Immunol. 24:437 114. Aden, D. P., Manning, D. D., Reed, N. D. 1974. Exclusion of cooperating T cells as targets for heterologous anti-~ antiserum. Cell. lmmunol. 14: 307 115. HayGlass, K. T., Benacerraf, B., Sy, M.-S. 1986. Theinfluence of B-cell idiotypes on the repertoire of suppressor T ceils. Immunol. Today 7:179 116. Flood, P. M., Janeway, C. A. Jr., Gershon, R. K. 1984. B cell-deprived mice lack functional expression of certain T suppressor cell subsets. J. Mol. Cell. Immunol. 1:167 117. HayGlass, K. T., Naides, S. J., Benacerraf, B., Sy, M.-S. 1985. T cell development in B cell deficient mice. III. Restriction specificity of suppressor T cell factor(s) produced in mice treated chronically with rabbit anti-mouse /~chain antibody. J. Mol. Cell. Immunol. 2:107 118. Martinez-A., C., Bernab6, R. R., de la Hera, A., Pereira, P., Cazenave, P.-A., Coutinho, A. 1985. Establishment of idiotypic helper T-cell repertoires early in life. Nature 317:721 119. Julius, M., Cosenza, H., Augustin, A. 1978. Evidencefor the endogenousproduction of T cell receptor bearing idiotypic determinants. Eur. J. Immunol. 8:484 120. Krawinkel, U., Cramer, M., Kindred, B., Rajcwsky, K. 1979. Isolated hapten-binding receptors of sensitized lymphocytes. V. Cellular origin of receptor molecules. Eur. J. Immunol. 9:815 121. Janeway, C. A. Jr., Murgita, R. A., Weinbaum,F. I., Asofsky, R., Wigzell, H. 1977. Evidence for an immunoglobulin-dependent antigen-specific helper T cell. Proc. Natl. Acad. Sci. USA 74:4582 122. Bottomly, K., Janeway, C. A. Jr., Mathieson, B. J., Mosier, D. E. 1980. Absenceof an antigen-specific helper T cell required for the expression of the T15 idiotype in mice treated with anti/t antibody. Eur. J. lmmunol. 10:159 123. Bottomly, K., Mosier, D. E. 1979. Mice whose B cells cannot produce the T15 idiotype also lack an antigen-specific helper T cell required for T15 expression. J. Exp. Med. 150:1399 124. Becker Dunn, E., Bottomly, K. 1985. T15-specific helper T cells: analysis of idiotype specificity by competitiveinhibition analysis. Eur. J. lmmunol. 15: 728 125. Becker Dunn, E., Kim, J., Bottomly,

AND NETWORK 243

K. 1986. Acloned T cell line that selectively augments antibody .responses of phosphorylcholine-specifi~ : B cells bearing the T15 idiotype. J. Mol. Cell. Immunol. 2:209 126. Woodland, R., Cantor, H. 1978. Idiotype-specific T helper cells are required to induce idiotype-positive B memorycells to secrete antibody. Eur. J. ImmunoL 8:600 127. Hetzelberger, D., Eichmann, K. 1978. Recognition of idiotypes in lymphocyte interactions. I. Idiotypic selectivity in the cooperation between T and B lymphocytes. Eur. J. lmmunol. 8:846 128. Adorini, L, Harvey, M., Sercarz, E. E. 1979. The fine specificity of regulatory T cells. IV. Idiotypic complementarity and antigen-bridging interactions in thc anti-lysozymc response. Eur. J. Immunol. 9:906 129. Kelsoe, G., Isaak, D., Cerny, J. 1980. Thymicrequirement for cyclical idiotypic and reciprocal anti-idiotypic immune responses to a T-independent antigen. J. Exp. Med. 151:289 130. Cerny, J., Caulfield, M. J. 1981. Stimulation of specific antibody-forming cells in antigen-primed nude mice by the adoptive transfer of syngeneic antiidiotypic T cells. J. Immunol.126:2262 131. Jayaraman, S., Swierkosz, J. E., Bellone, C. J. 1982. T cell replacing factor substitutes for an l-J+ idiotypic specific T helper cell. J. Exp. Med.155: 641 132. Strayer, D. S., Cosenza, H., Lee, W., Rowley, D. A., K6hler, H. 1974. Neonatal tolerance induced by antibody against antigen-specific receptors. Science 186:640 133. Weiler, I. J., Weiler, E., Sprenger, R., Cosenza, H. 1977. Idiotype suppression by maternal influence. Eur. J. ImmunoL 7:591 134. Wikler, M., Demens, C., Dewasme,G., Urbain, J. 1980. Immunoregulatory role of maternal idiotypes. Ontogenyof immune networks. J. Exp. Med. 152: 1024 135. Bernab6, R. R., Coutinho, A., Cazenave, P.-A., Forni, L. 1981. Suppression of a "recurrent" idiotype results in profound alterations of the whole B-cell compartment. Proc. Natl. Acad. Sci. USA 78:6416 136. Martinez-A., C., Toribio, M. L., de la Hera, A., Cazenave, P.-A., Coutinho, A. 1986. Maternal transmission of idiotypic network interactions selecting available T cell repertoires. Eur. J. Immunol. 16:1445 137. Eichmann, K. 1975. Idiotypic sup-

Annual Reviews


244

PEREIRA ET AL

pression. II. Amplification of a suppressor T cell with anti-idiotypic activity. Eur. J. Immunol.5:517 138. Owen,F. L., Ju, S.-T., Nisonoff, A. 1977. Binding to idiotypic determinants of large proportions of thymus-derived lymphocytes in idiotypically suppressed mice. Proc. Natl. Acad. Sci. USA 74:2084 139. Takemori, T., Rajewsky, K. 1984. Specificity, duration and mechanismof idiotype suppression induced by neonatal injection of monoclonal antiidiotope antibodies into mice. Eur. J. ImmunoL 14:656 140. Kearney, J. F., Vakil, M., Nicholson, N. 1987. Non-random V~ gene expression and idiotype-anti-idiotype expression on early B cells. In Evolution and Vertebrate Immunity: The Anti#en Receptor and MHCGene Families, ed. G. Kelsoe, D. Schulze. Austin: Texas Univ. Press 141. Holmberg,D., Forsgren, S., Ivars, F., Coutinho, A. 1984. Reactions amongst IgM antibodies derived from normal, neonatal mice. Eur. J. Immunol. 14: 435 G., 142. Holmberg, D., Wennerstrrm, Andrade, L., Coutinho, A. 1986. The high idiotypic connectivity of "natural" newborn antibodies is not found in adult mitogen-reactive B cell repertoires. Eur. J. Immunol. 16:82 143. Holmberg, D. 1987. High connectivity, natural antibodies preferentially use 7183 and QUPC 52 Vr~ families. Eur. J. Immunol. 17:399 144. Yancopoulos, G. D., Desiderio, S., Paskind, N., Kearney, J. F., Baltimore, D., Alt, F. W. 1984. Preferential utilization of the most D-proximal VHgene segmentsin pre-B cell lines. Nature 311: 727 145. Forni, L., Cazenave, P.-A., Cosenza, H., Forsbeck, K., Coutinho, A. 1979. Expression of V-region-like determinants on Ig-negative precursors in murine foetal liver and bone marrow. Nature 280:241 146. Vakil, M., Sauter, H., Paige, C., Kearney, J. F. 1986. In vivo suppression of perinatal multispecific B cells results in distortion of the adult B cell repertoire. Eur. J. Immunol. 16:1159 147. Forni, L, Heusser, C., Coutinho, A. 1988. Natural lymphocyteactivation in postnatal development of germ-free and conventional mice. Ann. Inst. Pasteur/ImmunoL 139:245 148. Pereira, P., Forni, L., Larsson, E.-L., Cooper, M. D., Heusser, C., Coutinho, A. 1986. Autonomous activation of B

and T cells in antigen-free mice. Eur. J. Immunol. 16:685 149. Babu, U. M., Maurer, P. H. 1981. The 6°, expression of anti-poly (LGLU ~°) idiotypic determinants dicLPHE tated by the gene products in the major histocompatibility complex (H-2q). Exp. Med. 154:649 150. Cancro, M. P., Sigal, N. H., Klinman, N. R. 1978. Differential expression of an equivalent clonotype among BALB/c and C57BL/6 mice. J. Exp. Med. 147:1 151. Juy, D., Primi, D., Sanchez, P., Cazenave, P.-A. 1983. The selection and the maintenance of the V-region determinant repertoire is germ-line encoded and T cell independent. Eur. J. Immunol. 13:326 152. Wortis, H. H., Burkly, L., Hughes, D., Roschelle, S., Waneck, G. 1982. Lack of mature B cells in nude mice with Xlinked immunedeficiency. J. Exp. Med. 155:903 153. Mond, J. J., Scher, I., Cossman, J., Kessler, S., Mongini, P. K. A., Hansen, C., Finkelman, F. D., Paul, W. E. 1982. Role of the thymus in directing the development of a subset of B lymphocytes. J. Exp. Med. 155:924 154. Sprent, J., Bruce, J. 1984. Physiology of B cells in mice with X-linked immunodeficiency. II. Influence of the thymusand mature T cells on B cell differentiation. J. Exp. Med. 160:335 155. Karagogeos, D., Wortis, H. H. 1987. Thymus grafts induce B cell development in nude, X-liked immune deficient mice. Eur. J. Immunol.17:141 156. Herzenberg, L. A., Stall, A. M., Lalor, P. A., Sidman, C., Moore, W. A., Parks, D. R., Herzenberg, L. A. 1986. The Ly-1 B cell lineage. Immunol. Rev. 93:81 157. Marcos, M. A. R., de la Hera, A., Gaspar, M. L., Marquez,C., Bellas, C., Mampaso, F., Toribio, M. L., Martinez-A., C. 1986. Modification of emerging repertoires by immunosuppression in immunodeficient mice results in autoimmunity. Immunol. Rev. 94:51 158. Gause, A., Yoshida, N., Kappen, C., Rajewsky, K. 1987. In vivo generation and function of B cells in the presence of a monoclonal anti-IgM antibody: implications B cell tolerance. Eur. J. Immunol. for 17:981 159. Goidl, E. A., Thorbecke, G. J., Weksler, M. E., Siskind, G. W. 1980. Production of auto-anti-idiotypic antibody during the normal immune response: Changes in the auto-antiidiotypic antibody response and the

Annual Reviews


T CELLS

AND NETWORK

245

idiotype repertoire associated with specific helper T cells. J. Exp. Med.156: aging. Proc. Natl. Acad. Sci. USA77: 506 6788 170. Primi, D., Devaux, C., Pierres, M., 160. Goidl, E. A., Choy, J. W., Gibbons, Barbier, E., Cazenave, P.-A. 1986. J. J., Weksler, M. E., Thorbecke, G. Analysis of immunoglobulins expressing the V×21E and Vx2ID gene proL, Siskind, G. W. 1983. Production of auto-antiidiotypic antibody during the ducts. Evidence for multigenic control normal immune response. VII. Analyof the expression of V,-VL combinasis of the cellular basis for the increased tions. Eur. J. lmmunol. 16:518 auto-antiidiotype antibody production 171. Staerz, U. D., Rammensee, H.-G., by aged mice. J. Exp. Med. 157: Benedetto, J. D., Bevan, M. J. 1985. 1635 Characterization of a murine monoclonal antibody specific for an allo161. Kim, Y. T., Goidl, E. A., Samarut, C., typic determinant on T cell antigen Weksler, M. E., Thorbecke, J., Siskind, G. W. 1985. Bone marrow function. I. receptor. J. Immunol. 134:3994 Peripheral T cells are responsible for 172. Huetz, F., Larsson, E.-L., Pereira, P., the increased auto-antiidiotype rePortnoi, D., Coutinho, A. 1988. T cell sponse of older mice. J. Exp. Med. dependence of "natural" auto-reactive 161:1237 B cell activation. Eur. J. ImmunoL18: 162. Tsuda, T., Kim, Y. T., Siskind, G. W., 1615 Weksler, M. E. 1988. Old mice recover 173. Portnoi, D., Freitas, A., Bandeira, A., the ability to produce IgG and high Holmberg, D., Coutinho, A. 1986. avidity antibody following irradiation Immunocompetentautoreactive B lymphocytes are activated cycling cells in with partial bone marrow shielding. Proc. Natl. Acad. Sci. USA85:1169 normal mice. J. Exp. Med. 164:25 163. Forni, L., Coutinho, A., K6hler, G., 174. Marcos, M. A. R., de la Hera, A., Jerne, N. K. 1980. IgM antibodies Pereira, P., Toribio, M. L., Coutinho, induce the production of antibodies of A., Martinez-A., C. 1988. B-cell participation in the recursive selection of the samespecificity. Proc. Natl. Acad. Sci. USA 77:1125 T-cell repertoires. Eur. J. Immunol.18: 164. lvars, F., Holmberg, D., Forni, L., 1015 Cazenave, P.-A., Coutinho, A. 1982. 175. Freitas, A. A., Rocha, B., Coutinho, A. Antigen-independent IgM-inducedanti1986. Lymphocytepopulation kinetics body responses: requirement for "recurin the mouse. Immunol. Rev. 91:5 rent" idiotypes. Eur. J. Imrnunol. 12: 176. Freitas, A. A., Coutinho, A. 1981. Very 146 rapid decay of mature B lymphocytes 165. lvars, F., Holmberg, D., Cazenave, in the spleen. J. Exp. Med. 154:994 P.-A., Coutinho, A. 1983. IgM-in177. Osmond, D. G. 1986. Population dynamics of bone marrow B lymduced specific antibody responses: direct correlation between responsivephocytes. Immunol, Rev. 93:103 ness and natural or induced recurrency 178. Rocha, B., Freitas, A. A., Coutinho, A. of the idiotype. Scand. J. Immunol.17: 1983. Population dynamics of T lym231 phocytes. Renewal rate and expansion 166. Holmberg, D., Ivars, F., Forni, L., in the peripheral lymphoid organs. J. Cazenave, P.-A., Coutinho, A. 1982. Immunol. 131:2158 Idiotypic characterization of antibody179. Miller, R. A., Stutman, O. 1984. T cell induced antibody responses. Immunorepopulation from functionally restricbiology 162:56 ted splenic progenitors: 10,000-fold 167. Holmberg, D., Ivars, F., Coutinho, A. expansion documented by using limit1983. An example of major histoing dilution analyses. J. Immunol.133: compatibility complex-linked control 2925 of idiotypic interactions. Eur. J. lmmu- 180. Scollay, R. G., Butcher, E. C., Weissnol. 13:82 man, I. L. 1980. Thymuscell migration. 168. Kelsoe, G., Reth, M., Rajewsky, K. Quantitative aspects of cellular traffic 1981. Control of idiotype expression by from the thymus to the periphery in monoclonal andi-idiotope and idiomice. Eur. J. Imrnunol. 10:210 tope-bearing antibody. Eur. J. Irnmu- 181. Eichmann, K., Falk, I., Rajewsky, K. nol. 11:418 1978. Recognition of idiotypes in lym169. Rubinstein, L. J., Yeh, M., Bona, C. A. phocyte interaction. II. Antigen-inde1982. Idiotype-anti-idiotype network. pendent cooperation between T and B II. Activation of silent clones by treatlymphocytes that possess similar and ment at birth with idiotypes is associcomplementary idiotypes. Eur. J. ated with the expansion of idiotype Immunol. 8:853

Annual Reviews


246

PEREIRA ET AL

182. Gleason, K., Pierce, S., K6hler, H. 1981. Generation of idiotype-specific T cell help through network perturbation. J. Exp. Med. 153:924 183. Pierce, S. K., Speck, N. A., Gleason, K., Gearhart, P. 3, K6hler, H. 1981. BALB/cT cells have the potential to recognize the TEPC15 prototype antibody and its somatic variants. J. Exp. Med. 154:1178 184. Kawahara, D. J., Miller, A., Sercarz, E. E. 1987. The induction of helper and suppressor cells with secondary antihen egg-white lysozome B hybridoma cells in the absence of antigen. Eur, J. lmmunol. 17:1101 185. Tite, J. P., Kaye, J., Saizawa, K. M., Ming, J., Katz, M. E., Smith, L. A., Janeway, C. A. Jr. 1986. Direct interactions between B and T lymphocytes bearing complementary receptors. J. Exp. Med. 163:189 186. Janeway, C. A. Jr., Broughton, B., Smith, L. A., Marion, T. N., Bottomly, K. 1987. Direct receptor: receptor interactions between T and B lymphocytes: idiotypic restriction in the antibody response to a cloned helper T cell receptor. J. Mol. Cell. lmmunol. 3:83 187. Paul, W. E., Bona, C. 1982. Regulatory idiotypes and immune networks: a hypothesis. Immunol. Today 3:230 188. L’age-Stehr, J., Teichmann, H., Gershon, R. K., Cantor, H. 1980. Stimulation of regulatory T ceil circuits by immunoglobulin-dependent structures on activated B cells. Eur. J. Immunol. 10:21 189. Abbas, A. K., Burakoff, S. J., Gefter, M. L., Greene, M, I. 1980. T lymphocyte-mediated suppression of myelomafunction in vitro. III. Regulation of antibody production in hybrid myeloma cells by T lymphocytes. J. Exp. Med. 152:969 190. Juy, D., Primi, D., Sanchez, P., Cazenave, P.-A. 1982. Idiotype regulation: Evidencefor the revolvement of IgH-C-restricted T cells in the M-460 idiotype suppressive pathway. Eur. J. lmmunol. 12:24 191. Milburn, G. L., Lynch, R. G. 1982. Immunoregulation of murine myeloma in vitro. I1. Suppression of MOPC-315 immunoglobulin secretion and synthesis by idiotype-suppressor T cells. J. Exp. Med. 155:852 192. Lynch, R. G., Millburn, G. L. 1984. Murine plasmaeytoma MOPC315as a tool for the analysis of network regulation. M315idiotopes are inducers and targets of immunoregulatory signals. See Ref. 38, p. 299

193. Kresina, T. F., Baine, Y., Nisonoff, A. 1983. Adoptivetransfer of resistance to growth of an idiotype-secreting hybridoma by T cells from idiotypically suppressed mice. J. lmmunol. 130:1478 194. Monroe, J. G., Gurish, M., Dambrauskas, J., Slaoui, M., Lowy, A., Greene, M. I. 1985. Genetic and biological characterization of a T suppressor cell induced by anti-idiotypic antibody. J. Immunol. 135:1589 195. Slaoui, M., Urbain, J., Lowy, A., Monroe,J. G., Willems, F., Benacerraf, B., Greene, M. I. 1986. Anti-idiotypic treatment of BALB/c mice induces CRib-bearing suppressor cells with altered Igh-restricted function. J. Immunol. 136:1968 196. Hausman, P. B., Sherr, D. H., Doff, M. E. 1986. Ai~ti-idiotypic B cells are required for the induction of suppressor T cells. J. Immunol. 136:48 197. Frischknecht, H., Binz, H., Wigzell, H. 1978. Induction of specific transplantation immune reactions using anti-idiotypic antibodies. J. Exp. Med. 147:500 198. Doff, M. E., Benaceraff, B. 1984. Suppressor cells and immunoregulation. Ann. Rev. Immunol. 2:127 199. Yamananchi, K., Murphy, D. B., Cantor, H., Gershon, R. K. 1981. Analysis of antigen-specific H-2 restricted cell free product(s) made I-J- Ly-2 cells (Ly-2 TsF) that suppresses Ly-2 cell-depleted spleen cell activity. Eur. J. Irnmunol. 11:913 200. Yamananchi, K., Murphy, D. B., Cantor, H., Gershon, R. K. 1981. Analysis of antigen-specific Ig restricted cell material made by I-J+ Ly-1 cells (Ly-I TsiF) that induces Ly-2+ cells to express suppressive activity. Eur. J. lmmunol. 11:905 201. Mrller, G,, ed. 1984. T cell receptors and genes. Immunol. Rev. 81:5 202. Chesnut, R. W., Grey, H. M. 1986. Antigen presentation by B ceils and its significance in T-B interactions. Adv. Immunol. 39:51 203. Frohman, M., Cowing, C. 1985. Presentation of antigen by B cells: functional dependence on radiation dose, interleukins, cellular activation, and differential glycosylation. J. Irnmunol. 134:2269 204. Lanzavecchia, A. 1985. Antigen-specific interaction between T and B cells. Nature 314:537 205. Krieger, J. I., Grammer,S. F., Grey, H. M., Chesnut, R. W. 1985. Antigen presentation by splenic B ceils: resting B cells are ineffective, whereas acti-

Annual Reviews


T CELLS vated B cells are effective accessorycells for T cell responses. J. lmmunoL135: 2937 206. HayGlass, K. T., Naides, S. J., Scott, C. F., Benacerraf, B., Sy, M.-S. 1986. T cell developmentin B cell-deficient mice. IV. Therole of B cells as antigenpresenting cells in vivo. J. Immunol. 136:823 207. Janeway, C. A. Jr., Ron, J., Katz, M.E. 1987. The B cell is the initiating antigen-presenting cell in peripheral lymph nodes. J. Immunol. 138:1051 208. Lichtman, A. H., Tony, H.-P., Parker, D. C., Abbas, A. K. 1987. Antigenpresentation by hapten-specific B lymphocytes. IV. Comparativeability of B cells to present specific antigen and anti-immunoglobulin antibody. J. Immunol. 138: 2822. 209. Ron, Y., Isakov, N., Sprent, J. 1987. Unresponsiveness of MRL/MP-lpr/lpr mice to antigen given subcutaneously in adjuvant: partial restoration of responseafter local injection of B cells. J. Immunol. 139:400 210. Ron, Y., Sprent, J. 1987. T cell priming in vivo: a major role for B cells in presenting antigen to T cells in lymph nodes. J. Immunol. 138:2848 211. Liano, D., Abbas, A. K. 1987. Antigen presentation by hapten-specific B lymphocytes. V. Requirements for activation of antigen-presenting B cells. J. Immunol. 139:2562 212. Krieger, J., Jenis, D. M., Chesnut, R. W., Grey, H. M. 1988. Studies on the capacity of intact cells and purified Ia from different B cell sources to function in antigen presentation to T cells. J. Immunol. 140:388 213. Kurt-Jones, E. A., Liano, D., HayGlass, K. A., Benacerraf, B., Sy, M.-S., Abbas, A. K. 1988. The role of an.tig.en-p.resen.ting B cells in T cell priming ~n wvo. Studies of B celldeficient mice. J. Immunol. 140:3773 214. Gosselin, E. J., Tony, H. P., Parker, D. C. 1988. Characterization of antigen processing and presentation by resting B lymphocytes. J. Immunol. 140:1408 215. Tony, H. P., Parker, D. C. 1985. Major histocompatibility complex-restricted polyclonal B cell responses resulting from helper T cell recognition of antiimmunoglobulin presented by small B lymphocytes. J. Exp. Med. 161:223 216. Tony,H.-P., Phillips, N. E., Parker, D. C. 1985. Role of membrane immunoglobulin (Ig) crosslinking in membrane Ig-mediated, major histocompatibility-restricted T cell-B cell cooperation. J. Exp. Med. 162:1695

AND NETWORK

247

217. Janeway, C. A. Jr., Sakato, N., Eisen, H. N. 1975. Recognition of immunoglobulin idiotypes by thymusderived lymphocytes. J. Itnmunol. 72: 2357 218. Sakato, N., Janeway, C. A. Jr., Eisen, H. N. 1977. Immune responses of BALB/cmice to the idiotype of T15 and other myeloma proteins of BALB/corigin: implications for an immune network and antibody multispecificity. CoMSpriny HarborSyrnp. Quant. Biol. 41:719 219. Jorgensen, T., Hannestad, K. 1982. HelperT cell recognition of the variable domains of a mouse myelomaprotein (315). Effect of the major histocompatibility complex and domain conformation. J. Exp. Med. 155:1587 220. Bogen,B., Malissen, B., Haas, W. 1986. Idiotopeospecific T cell clones that recognize syngeneic immunoglobulin fragments in the context of class II molecules. Eur. J. lmmunol. 16:1373 221. Kawahara, D. J., Marrack, P., Kappler,J. 1982. Helper T cells specific for a VHdeterminant(s) are under H-2linked Ir gene control. Fed. Proc. 41: 366 222. Alzari, P. M., Lascombe, M.-B., Poljak, R. J. 1988. Three-dimensional structure of antibodies. Ann. Rev. Immunol. 6:555 223. Kourilsky, P., Claverie, J.-M. 1986. The peptidic self model: a hypothesis on the molecular nature of the immunological self. Ann. Inst. Pasteur/ Immunol. 137D: 3 224. Kourilsky, P., Chaouat, G., Rabourdin-Combe, C., Claverie, J.-M. 1987. Working principles in the immunesystem implied by the "peptidic self" model. Proc. Natl. Acad. Sci. USA84: 3400 225. Pereira, P., Larsson, E.-L., Forni, L., Bandeira, A., Coutinho, A. 1985. Natural effector T lymphocytesin normal mice. Proc. Natl. Acad. SoL USA 82:7691 226. Bandeira, A., Coutinho, A., MartinezA., C, Pereira, P. 1988. The origin of "natural antibodies" and the internal activity in the immunesystem. Int. Rev. Immunol. 3:47 227. Coutinho, A. 1988. Tolerance to self as a global behaviour of the immune system based upon the connectivity of auto-reactive lymphocytes. In The Tolerance Workshop, ed. P. Matzinger, M. Flajnik, H.-G. Rammensee,G. Stockinger, T. Polink, L. Nicklin. 3: 54. Basel: Editions Roche 228. Guilbert, B., Dighiero, G., Avrameas,

Annual Reviews

248

229.


230.

231.

232.

233.

234.

235.

236.

237.

238.

239.

PEREIRA ET AL S. 1982. Naturally occurring antibodies against nine common antigens in humansera. I. Detection, isolation, and characterization. J. ImmunoL128:2779 Dighiero, G., Lymberi, P., Mazi6, J.C., Rouyre, S., Butler-Browne, G. S., Whalen, R. G., Avrameas, S. 1983. Murine hybridomas secreting natural monoclonal antibodies reacting with self antigens. J. ImmunoL131:2267 Dighiero, G., Lymberi, P., Holmberg, D, Lundkvist, I., Coutinho, A., Avrameas, S. 1985. High frequency of natural autoantibodies in normal newborn mice. J. Immunol. 134; 765 Ron, Y., De Baetselier, P., Gordon,J., Feldman, M., Segal, S. 1981. Defective induction of antigen-reactive proliferating T cells in B cell-deprived mice. Eur. J. Immunol.1 i: 964 Ron, Y., De Baetselier, P., Tzehoval, E., Gordon, J., Feldman, M., Segal, S. 1983. Defective induction of antigenreactive proliferating T cells in B celldeprived mice. II. Anti-# treatment affects the initiation and recruitment of T cells. Eur. J. Immunol. 13:167 Kim, K. J., Rollwagen, F., Asofsky, R., Lelkovits, I. 1984. The abnormal function of T cells in chronically anti#-treated mice with no mature B lymphocytes. Eur. J. Immunol. 14:476 Cerny, A., Huegin, A. W., Sutter, S., Bazin, H., Hengartner, H. H., Zinkernagel, R. 1986. Immunity to lymphocytic choriomeningitis virus in B cell-depleted mice: evidence for B cell and antibody-independent protection by memoryT cells. Eur. J. lmmunol. 16:913 Portnoi, D., Lundkvist, I., Coutinho, A. 1988. Inverse correlation between the utilization of an idiotype in specific immune responses and its representation in pre-immune "natural" antibodies. Eur. J. Immunol. 18:571 Lundkvist, I., Coutinho, A., Varela, F., Holmberg, D. 1988. Evidence for a functional network amongst natural antibodies in normal mice. Submitted Bandeira, A., Coutinho, A., Carnaud, C., Jacquemart, F., Forni, L. 1988. Transplantation tolerance correlates with high levels of lymphocyteactivity. Proc. Natl. Acad. Sci. USA. In press Luzmi,S., Merino, J., Engers, M., Izui, S., Lambert, P.-A. 1986. Autoimmunity after induction of neonatal tolerance to alloantigens: role of B cell chimerism and F~ donor B cell activation. J. ImmunoL136:4420 VonBoehmer, H., Schubiger, K. 1986. Thymocytes appear to ignore Class

I major histocompatibility antigen expressed on thymusepithelial cells. Eur. J. Immunol. 14:1048 240. Wilson, D. 1988. See Re£ 227, p. 88 241. Holmberg, D., Coutinho, A. 1985. Natural antibodies and autoimmunity. hnmunol. Today 6:356 242. Kappler, J. W., Roehm, N., Marrack, P. 1987. T cell tolerance by clonal elimination in the thymus. Cell 49:273 243. Kappler, J. W., Staerz, U., White, J., Marrack, P. C. 1988. Self-tolerance eliminates T ceils specific for Mls-modifled products of the major histocompatibility complex. Nature 332:35 244. MacDonald, H. R., Schneider, R., Lees, R. K., Howe, R. C., Acha-Orbea, H., Fester~stein, H., Zinkernagel, R. M., Hengartner, H. 1988. T-cell receptor Vfl use predicts reactivity and tolerance to Mls"-encoded antigens. Nature 332:40 244a. MacDonald, H. R., Pedrazzini, T., Schneider, R., Louis, J. A., Zinkernagel, R. M., Hengartner, H. 1988. -Intrathymic elimination of Mls~-reae tive (Vfl6+) cells during neonatal tolerance induction to Mlsa-encodedantigens. J. Exp. Med. 167:2005 245. Kisielow, P., Blfithmann, P. Staerz, U. D., Steinmetz, M., Von Boehmer, H. 1988. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4÷ 8 ÷ thymocytes. Nature 333:742 246. Ben-Nun, A., Cohen, I. R. 1982. Spontaneous remission and acquired resistance to autoimmuneencephalomyelitis (EAE) are associated with suppression of T cell reactivity: suppressed EAE effector T cells recoveredat T cell lines. J. Immunol. 128:1450 247. Ben-Nun, A., Eisenstein, S., Cohen, I. R. 1982. Experimental autoimmune encephalomyelitis (EAE)in genetically resistant rats: PVGrats resist active induction of EAEbut are susceptible to and can generate EAEeffector T cell lines, d. Immunol. 129:918 248. Naparstek, Y., Holoshitz, J., Eisenstein, S., Reshef, T., Rappaport, S., Chemke, J., Ben-Nun, A., Cohen, I. R. 1982. Effector T lymphocyte line cells migrate to the thymus and persist there. Nature 300:262 249. Maron, R., Zerubavel, R., Friedman, A., Cohen, I. R. 1983. T lymphocyte line specific for thyroglobulin produces or vaccinates against autoimmtmethyroiditis in mice. J. Irmnunol. 131:2316 250. Hooper, D. C., Taylor, R. B. 1987. Specific helper T cell reactivity against autologous erythrocytes implies that

Annual Reviews


T CELLS AND NETWORK self tolerance need not ..depend on clonal deletion. Eur. J.ilmmunol. 17: 797 251. Hooper, D. C., Young, J. L., Elson, C. J., Taylor, R. B. 1987. Murine T cells reactive against autologous erythrocytes: evidence for in vitro and in vivo priming with mouse and rat red blood cells. Cell. Immunol.106:53 252. Sakaguchi, S., Takahashi, T., Nishizuka, Y. 1982. Study on cellular events in,postt.hymectomy autoimmune ooph, oriti~ in mice. II., Requirements of Lyt-1 c~lls in normal female mice for the prevention of oophoritis. J. Exp. Med. 156:1577 253. Taguchi, O., Takahashi, T., Seto, M., Namikawa, R., Matsuyama, M., Nishizuka, Y. 1986. Development of multiple organ-localized autoimmunediseases in nude-miceafter reconstitution of T cell function by rat fetal thymus graft. J. Exp. Med. 164:60 254. Taguchi, O., Nishizuka, Y. 1987. Self tolerance and localized autoimmunity. Mouse models of autoimmune disease that suggest tissue-specific suppressor T cells are involvedin self tolerance. J. Exp. Med. 165:146 255. Martinez-A., C., Marcos, M. A. R., Pereira, P., Marquez, C., Toribio, M. L., de la Hera, A., Cazenave, P.-A., Coutinho, A. 1987. Turning (Ir-gene) low-responders into high responders by antibody manipulation of the developing im~nunc system. Proc. Natl. Acad. Sci. USA84:3812 256. Lamb, J. R., Skidmore, B. J., Green, N., Chiller, J. M., Feldmann,M. 1983. Induction of tolerance in influenza virus-immune T lymphocyte clones with synthetic peptides of influenza hemagglutinin. J. Exp. Med. 157:1434 257. Jenkins, M. K., Schwartz, R. H. 1987. Antigen presentation by chemically modified splenocytes induces antigenspecific T cell unresponsivenessin vitro and in vivo. J. Exp. Med. 165:302 258. Quill, H., Schwartz, R. H. 1987. Stimulation of normal inducer T cell clones with antigen presented by purified Ia molecules in planar lipid membranes:

249

specific induction of a long-lived state of proliferative nonresponsiveness. J. Immunol. 138:3704 259. Suzuki, G., Kawase, Y., Koyasu, S., Yahara, I., Kobayashi, Y., Schwartz, R. H. 1988. Antigen-induced suppression of the proliferative response of T cell clones. J. lmmunol. 140:1359 260. Jenkins, M. K., Ashwell, J. D., Schwartz, R. H. 1988. Allogeneic nonT spleen cells restore the responsiveness of normal T cell clones stimulated with antigen and chemically modified antigen-presenting cells. J. Immunol. 140: 3324 261. Tomonari, K. 1985. T-cell receptor expressed on an autoreactive T-cell clone, clone 4. Cell. Immunol. 96: 147 262. Webb, S., Sprent, J. 1987. Downregulation of T cell responses by antibodies to the T cell receptor. J. Exp. Med. 165:584 263. Breitmeyer, J. B., Oppenheim, S. O., Daley, J. F., Levine, H. B., Schlossman, _S: F. 1987. Growthinhibition of human ’T’cells-by antibodies recognizing the T cell antigen receptor complex. J. Immunol. 138:726 264. Bandeira, A., Larsson, E.-L., Forni, L., Pereira, P., Coutinho, A. 1987. In vivo activated splenic T cells are refractory to interleukin 2 growthin vitro. Eur. J. Immunol. 17:901 265. Andersson, J., Bullock, W. W., Melchefs, F. 1974. Inhibition of mitogenic stimulation of mouse lymphocytes by anti-mouse immunoglobulin antibodies. I. Modeof action. Eur. ,;J. Immunol. 4:715 266. Jerne, N. K. 1974. Towards a network theory of the immune system. Ann. Immunol. (Inst. Pasteur) 124C: 373 267. Varela, F. J., Sanchez-Leighton, V., Coutinho, A. 1988. Adaptive strategies gleaned from immune networks. Viability theory and comparison with classifier systems. In Evolutionary and Epigenetic Order from Complex Systems: A Waddington Memorial Volume, ed. B. Goodwin, P. Saunders. Edinburgh: Edinburgh Univ. Press. In press



Annual Reviews www.annualreviews.org/aronline Ann. Rev. lmmunol.1989. 7:251-75 Copyright© 1989by AnnualReviewsInc. All rights reserved


THE IMMUNE SYSTEM OF XENO P US Louis Du Pasquier,* Martin F. Flajnik**

Joseph Schwager* and

* Basel Institute for Immunology, Grenzacherstrasse487, CH-4005 Basel, Switzerland

INTRODUCTION Comparativestudies of the immunesystem require that for each key systematicposition at least one animalmodelis investigated in depth. For anuranamphibians,modernrepresentatives of the first vertebrates that achievedthe transition to terrestrial life, this modelis the clawed-toad or SouthAfricanfrog (genusXenopus).Since these frogs are easy to maintain and breed in captivity and are commerciallyavailable, and since they present developmentaland genetic advantagescomparedto other amphibians, they becamethe modelof choice for manyinvestigators. In 1988, after about 20 years of work, weare heading toward a comprehensive view of the Xenopusimmunesystem, thanks to use of a cross-fire of methodologies, rangingfromclassical graft rejection to geneanalysis. This reviewfirst presentsa descriptionof the structural elementsof the Xenopus immunesystem: the lymphoid system, the major histocompatibility complex,and the immunoglobulins. Three functional issues will be then considered:immune responses, tolerance, and antibodydiversity duringontogeny.Finally, a section will be devotedto the impactof polyploidy on the Xenopusimmunesystem. THE LYMPHOID LYMPHOCYTES

ORGANS

AND THE

The Thymus (Table 1) Thethymusarises throughan invagination of the dorsal epitheliumof the second pharyngeal poucharound day 3 after fertilization (stage 40 of **Present address: Department of Immunologyand Microbiology, University of Miami, Miami, Florida, 33101.

251 0732~)582/89/0410~)251 $02.00


Annual Reviews 252 DU PASQUIER ET AL

g

o g

-~++


Annual Reviews IMMUNE SYSTEM OF XENOPUS

ZNN 253

Annual Reviews


254

DU PASQUIER ET AL

Nieuwkoop& Faber) (1-4). It is colonized during the following days cells from the dorsal lateral plate and ventral mesodermthrough the head mesenchyme(5-9). These large lymphoblasts proliferate in situ. One day later, the epitheliumbegins expressing class-II moleculesbut not the classical MHC class-I molecules (10). By day 6-8 the cortex-medulla architecture becomesvisible (11). The cortex contains mainly proliferating lymphocytesintricated in the digitations of the epithelial cells. The medullacontains epithelial cells with cytoplasmic granules and rough endoplasmic reticulum. Unlike the case in mammals,the cortex and medulla are separated by a distinct cellular barrier (12, 13). This area is rich in blood vessels and IgMproducing plasma cells (14). Macrophagesare present from early stages on. After metamorphosis a new type of macrophage presumed by some to be a nurse cell equivalent appears (2). Within the cortex large dendritic cells resembling those described in Rana(15) have been found in Xenopus(16). Myeloidcells, mucouscells and cysts are occasionally seen in the medulla (2). The thymusreaches its peak larval size (about 1-2 x 106 lymphocytes) at stage 58 (3, 17, 18). It involutes during metamorphicclimax (17-19) during which more macrophagesare seen in the shrinking cortex. At this time the thymus is translocated towards the tympanum(3). A second histogenesis follows, with the appearance of myoepithelial cells, aminecontaining cells, and larger cysts (2). The size of the organ reaches about 1-3 x 107 cells about 2-3 months after metamorphosis, and then it undergoes a regression at the time of sexual maturity, when it becomes more and more embeddedin fatty tissue (17, 19). A cell surface glycosylated membraneprotein of 120 kd has been identified with a monoclonal antibody XT-1 on a subpopulation of Xenopus T cells (20-23). Thymectomyseverely impairs allograft and mixed leucocyte reactions (MLRs)(24~26), proliferative responses to classical T-cell mitogens 27), and antibody responses against T-dependent antigens (25, 28). It does not affect response to B-cell mitogens (29), nor antibody response thymus-independentantigens (30, 31), nor to xenografts (32). IgY is absent in the serum of early thymectomized (Tx) animals, whereas IgM is more abundant and IgX not affected (33-35). Late thymectomy(after 8 days) does not impair allo-responses but still interferes with the antibody response to T-dependent antigens, as if someregulatory cells had not yet left the thymus(36). This, together with the stepwise effects ofthymectomy on MLRand phytohemagglutinin (PHA) responsiveness, suggests sequential maturation of the T-cell function (37), The thymus both generates and contains T helper cells (38), although thymic T cells are relatively poor helpers and poor graft-vs-host effectors. ThymicT cells obtained by nylon wool filtration can suppress ongoing T-B

Annual Reviews


IMMUNE SYSTEM OF XENOPUS

255

collaboration in vitro, althoughno quantitative titration of this suppressive activity was ever possible for unknownreasons (38). However, thymic dependent suppression or direct thymus cell suppression is a recurrent finding also under allogeneic conditions (39-43). Thymocytesfrom adults but not from larvae can produce the equivalent of IL-2 after allogenic or mitogen stimulation. Larval thymocytes will respond to mitogens when exogenous IL-2 is added (44). The thymus of Xenopus also contains IgMplasma cells or B cells ready to switch to IgY synthesis (14, 38, 45). In all these experimentscare must be taken of the specificity and affinity of the antibodies. Cross-reactive IgMantibodies can give apparently normal titers in Tx animals although the T-dependent compartmentis indeed affected (46). The Spleen

(Table

1)

The spleen appears about 12-14 days after fertilization as a mesenchymal thickening in the mesogastrium. The mature spleen has delineated regions of red (hematopoietic) and white (lymphopoietic) pulp (1, 47). The nodules of white pulp, with their central arteriola surrounded by lymphocytes, are lined by a boundarylayer of cells. Scattered lymphocytesmainly of the T lineage are present in the perifollicular area (2, 32, 48-50). The larval spleen is not hematopoietic but is richer in B cells than that in the adult. During metamorphosis the spleen cell number reaches a plateau (0.51 x 106), or may even drop. Afterward the organ grows steadily until it contains about 4 x 107 lymphocytesin 300 g adults (17). Large mitotically active dendritic cells are located in the white pulp and can extend pseudopods toward T cells. They are not classical macrophages, and they remain after thymectomy(51). Unlike the case in other species, Xenopusspleen is the only organ to accumulate and retain antigen for several weeksin a specific zone near the periphery of the white pulp (1, 52). Since Tx animals do not showthis type of antigen maintenance it has been proposed that it was due to antigenantibody complexes(53). Particular and soluble antigens have a different distribution (1). During an ongoing immuneresponse, proliferation cells in both the red and the white pulp can be observed in some but not all cases (1, 54). The spleen is a source of antibody forming cells (55). Most of its B cells produce IgM and a very few produce IgY or IgX. Lipopolysaccharideinduced proliferation of Xenopus B cells is poor and may be due to a contaminant of LPSpreparation (56). Lipopolysaccharide (LPS) itself, doses comparable to those used in mammaliansystems, induces Ig synthesis (E. Hsu, T. Leanderson, personal communication;34). Spleen cells

Annual Reviews


256

DU PASQUIER ET AL

respond well to PHAand Con A (57). Pokeweedmitogen (PWA)and antiIg reagents are potent stimulators of proliferation and differentiation (58, 59); phorbol myristateacetate (PMA)induces 80%of splenic leukocytes to synthesize DNA(60). Spleen cells respond muchearlier in ontogeny mitogens than do thymocytes (61). Adult spleens from day-10 thymectomizedanimals do contain IL-2 producing and IL-2 responding cells (44). Splenocytes from Tx animals that have rejected an allograft display a "nonspecific" MLRagainst any MHChaplotype (62). Spleen contains also specific alloreactive T cells capable of MLR and of differentiation into cytotoxic lymphocytes. Helper cells are more abundant or more efficient in the spleen than in the thymus (38). Splenectomy has no major effect immuneresponses (63).

Other Sources of LymphoidCells In the anterior part of the tadpole the ventral and dorsal cavity bodies occupythe central part of the pharynx (11, 64). They are depleted of their lymphocytes after thymectomy(1, 48) and disappear at metamorphosis. Lymphoidnodules are present in the adult but not the tadpole intestine. They are rich in B cells. Numerousplasma cells producing IgM and IgX, but not IgY, are also visible in the mucosa(65). In the liver, lymphopoiesis is associated with the lymphomyeloidperipheral layer (11, 66) whichpersists throughout the life of Xenopus,but not in other anuran species where it disappears at metamorphosis. The liver traps circulating particles by meansof Kupffer-like cells (67) which are class-II positi, ve (10). Free mononuclearliver cells can be stimulated PHAor by the purified protein derivative of tuberculin, but not by Con

A(68). In the kidney, lymphocytes, mainly B cells, accumulate along the whole mesonephros, between the renal tubules (1). Injected antigen has been traced to intertubular areas (67, 69). Mesenteries and gills contain random accumulations of lymphomyeloid cells, and no homologyof lymph glands or jugular bodies found in other anura can be seen in Xenopus(1). Peritoneal cells produce factors with l-like activity (70). The bone marrow of Xenopus develops after metamorphosis. From functional studies with mitogen or from cytological observation it may not be lymphopoietic(68, 71; discussed in 1). Blood contains B and T cells and monocytes: the experiments that led to the demonstration ofT- and B-cell collaboration and the MHCrestriction of this phenomenonin Xenopus were done mainly with peripheral blood cells (72, 73). After immunizationthe numberof plasma cells is only two times lower than in the spleen.

Annual Reviews IMMUNE SYSTEM

OF XENOPU$

257


Embryonic Origin of Lymphoid Stem Cells Twomesodermalstem cell compartmentsexist in the embryo:the ventral bloodisland (VBI)and the dorsal lateral plate mesoderm (5, 74-76). cells associated with early hemopoiesisare derived fromthe VBI.Studies with the X. borealis marker(77) provedthat the thymusrudimentat stage 43 is colonizedonly by cells derived fromthe VBI(78), whereasin adult life, similar experimentswith chimerashave shownthat the thymuswas essentially populatedby cells fromthe dorsal lateral plate (79). Thetwo compartmentsare in fact not completely independent from each other (80). It remainsdifficult to assert the type and function of VBI-derived cells becausethey appear to be a transient population. Morerecently it has been suggestedthat "accessorycells" and thymocytescould arise from a bipotential precursor that diverges into these two lineages after the colonizationof the epithelial thymicrudiment(81, reviewin 82). THE MAJOR HISTOCOMPATIBILITY

COMPLEX

Theexistence of a majorhistocompatibility complexwasoriginally proven by family studies in whichthe cosegregationof classical mammalian MHC functional or serological markers were analyzed. Thus, MLR (83) acute gra.ft rejection occurred, and somecell surface antigens detected with alloantisera segregatedtogether (84); this providedevidencefor a complex called XLAby analogy with HLA. Structural

Aspects (Table 1)

The abovementionedantisera and anti-humanclass-II DRflxenoantisera were used to characterize the XenopusMHC molecules(85, 87). Class I-like molecules have been immunoprecipitatedfrom 125I cell surface labeled leukocytes and erythrocytes. The heavychain has a mol wt ranging from 40-44 kd dependingon the allele examined,of which3 kd consist in N-linked carbohydrate. The light chain, presumablythe homologueof f12 microglobulin, had a mol wt of 13 kd and no N-linked carbohydrates. The family studies confirmedthat the heavy chains are encoded by XLAgenes. Size and charge polymorphismwas obvious from the twodimensional (2d) gel electrophoresisanalysis of 7 alleles (85). Class-II antigens were identified with alloantisera recognizing an approximately 30-kd molecule on the surface of lymphocytesin MHCtypedfamilies (86). Moreover the IgGfraction of a rabbit antihumanclassII, fl chain-inhibited, bidirectional MLR recognizedthe samemoleculesas the alloantisera, and immunoprecipitated the Xenopusclass-II molecules. Theyare composedof two 30- to 35-kd integral membrane glycoproteins.

Annual Reviews


258

DU PASQUIER ET AL

The ~ chains have some N-terminal sequence homology with mammalian ~ chains. Unlike most mammalianclass-II molecules, the deglycosylated fl chains are larger and more acidic than the ~ chains. Each XLAhaplotype seems to carry two ~ and up to 5 fl genes. Both class-II chains are polymorphic.In family studies the 2d gel patterns precisely matched MLRdata. However, for some outbred animals that were attributed a MLRJ haplotype (a poor stimulator in most MLR), the 2d gel pattern failed to display the J pattern. Clearly MLR underestimates the true level of polymorphismin a Xenopus population (87). Until now the following haplotypes of Xenopus MHChave been described a, b, c, d, f, g, r, j, k (88-90). Frequencyoff and r are similar (3.3%) whereas g is rare (0.8%). Linkage disequilibrium has been found between class and class II in a population that had otherwise reached equilibrium (91). A polymorphism based on isoelectric focusing (IEF) pattern has been exploited to study the linkage of the complement component C4 (92) MHCin two strains (J and K) of Xenopus. In a backcross of progeny of 25 individuals there was correlation between MHCtype (defined by MLR and acute graft rejection) and the C4 type. However,the C4 IEF type of one individual did not correlate, perhaps indicating a high frequency of crossover (90). C3 is not MHC-linked;therefore the possible date of separation of C3 and C4 predates the emergenceof amphibians in evolution (93). One Xenopus erythrocyte glycoprotein of 38-43 kd displays some of the hallmarks of chicken BGmolecules. It is only expressed on red cells, does not bind f12 microglobulin, and appears to be MHClinked. It is polymorphic,by size and reactivity with alloantisera, but unlike class I, it has the same V8 protease pattern. Under nonreducing conditions, homoor heterodimers are formed as with BGmolecules (94). It remains determine whether its linkage to MHCis physical or genetic. Functional

Aspects

Cytotoxic T lymphocytes (CTL) can be generated in Xenopus although it is not possible to generate killers after a primary MLR.Anin vivo priming is required, followed by MLR.The CTL specifically recognize MHClinked target molecules on lymphoblasts (95). In vitro assays for T and B collaboration have been set up using T and B separated by nylon wool, from either carrier or hapten primed animals of various MHCtypes (72, 73). Antibodies of isotype IgY, which are highly thymus dependent, and high affinity IgMantibodies (73), could detected only in combinations where T- and B-cells shared at least one MHChaplotype. IgM antibodies of low affinity were produced normally in all cultures, regardless of the MHC type of T- and B-cells. To further determine the homology of MHCfunction in Xenopus,

Annual Reviews


IMMUNE SYSTEM OF XENOPUS

259

several types of experiments were done in which T cells matured in the "wrong" MHCenvironment. Antibody production to a thymus-dependent and graft rejection was examined. Reconstitutions of early Tx Xenopus have been attempted using either genetically undefined, partially defined, MHCtyped, or isogenic thymus donors (96, 97). The best system remains that of LG hybrids (98) where truly isogenic and MHCpredetermined replacements can be done. Tx tadpoles were later reconstituted with larval irradiated thymus of all possible XLAtypes. Under those circumstances one tested whether T cells that differentiated in an allogeneic thymuswere able to collaborate with B cells. Muchless IgY was produced in fully MHC mismatched chimaeras than in chimaeras where one MHChaplotype was matched or in control. IgM antibody production was the same in all groups. To better address the issue of education, the chimaera had to be done. earlier, before the thymus was colonized. In such chimaeras (8), made by sectioning the embryos just behind the gill anlagen, the thymus epithelium is derived from the head, and all the lymphocytes are hematopoietic cells from the body. The results were similar to those of the larval chimaera but even less pronounced (89). MHC fully mismatchedchimaeras made IgY antibodies with low kinetics or in low amounts depending on the MHChaplotypes involved. In summary,although it was not absolute, there was a thymic selection of the T-cell repertoire in Xenopus. Early attempts to generate CTLagainst minor histocompatibility antigens have failed (95). This maynot be surprising since the generation CTLagainst MHC was already difficult. Recently T-cell lines generated to both major and minor histocompatibility antigens have been obtained with growth factors from mitogenstimulated cultures (99). These lines will be useful to study in vitro the MHC restricted recognition of such antigens. Experiments done with chimaeras in vivo have suggested that responses to at least some minor antigens are MHCrestricted. LG15(a/c) and LG6 (a/c) strains have the same MHC genotypes a/c and differ only by minor H antigens. LG3(b/d) differs from these strains at the level of the two MHChaplotypes and rejects a/c skin acutely in 21-25 days. LG3can be made tolerant to LG15(a/c) tissue by grafting an eye anlage during embryoniclife (within 24 hours after fertilization). Such an animal, which b/d lymphocyteshave matured in a b/d thymus, is not able to reject either an LG6or LGI5 graft. Presumably the minor antigens of LG6, to be recognizcd by T cells, have to be recognized in a MHC-restrictedfashion by the host lymphocytes. LG15(a/c head)/LG3 (b/d body) chimaeras which b/d lymphocytes have matured in an a/c thymus, on the other hand, reject LG6(a/c) grafts as rapidly as normal LG15, but tolerate LG15 grafts. This was the first evidence for a positive selection of the T-cell clones in the thymus(89).

Annual Reviews 260

DU PASQUIER ET AL

THE IMMUNOGLOBULIN THEIR GENES

MOLECULES

AND

Heavy Chain Isotypes (Table 1) (reviewed in 100) IgM is the most abundant isotype (2 mg/ml) (101). Its heavychain is heavily glycosylated. IgMexists in a slightly different size in outbred Xenopus (102). All the anti-IgM monoclonal antibodies precipitated from serum a thick diffuse band at mol wt 72,000-73,000 which can be resolved into two distinct bands at mol wt 61,700 and 60,000, whenB cells are treated with tunicamycin (102). Polymeric IgM associated with J chain and forms hexamers (103, 104). The full sequence of an IgM heavy chain has been determined from a cDNAclone (105). It has four domains. The CH3and CH4domains are the most conserved (47% and 42%identities, respectively) whereas CH1 and CH2 show little homology with mammalian IgM (31% and 32%). This # chain is encoded by a four-exon germline gene. The Xenopus~t gene has much larger introns than do those of mammals (106). A putative switch region of about 3 kb in the intron between J, and C/~ consists in multiple repeats of 150 bases rich in AAGCTCAGCT elements (107).


IgM

IgY The Xenopus analog of IgG has been called IgY because its mol wt, most likely due to four constant region domains, is greater than IgG (Table 1) (100). Glycosylated and nonglycosylated forms are present in equimolar amounts in the IgY of all normal serum samples (present at 0.7 mg/ml), as if during biosynthesis only one chain in each dimer could bear the carbohydrate (102). IgY preferentially associates with the heaviest light chain. IgX The existence of a third Ig isotype had been suspected after the immunoprecipitation of what was considered as a second # (58). In fact IgX is an antigenetically distinct isotype but is polymeric like IgM. Its heavy chain (mol wt 63,000) is more glycosylated than/~ (34). ISOTYPE DISTRIBUTION The three isotypes differ by their V8 peptide digest, by several antigenic epitopes, and by their tissue distribution. Splenic IgM+ B cells range from 20%(with monoclonal antibodies) (57, 108) to (with xenoanti-Ig sera) (109). Gut epithelium, liver hematopoietic layers, and thymusmedulla are rich in IgMplasma cells (14). IgY and IgX positive cells are muchrarer in the spleen but manyIgX p~asmacells are found in the gut. All B cells positive for IgY hlso produce IgM (102). The thymus does not contain surface IgM positive cells in large numbers (110) originally described (111).

Annual Reviews IMMUNE SYSTEM OF XENOPUS 261

The Light Chain Three categories of light chains from 25-29 kd have been detected. Two of them, recognizedby monoclonalantibodies, are antigenically distinct (34, 102). L chains from anti-DNP antibodies have been partially sequencedand are related to both VKand V2 sequences (112). A cDNA clone for the constant region of one of the light chain has beenisolated and showscharacters in common with both ~ and 2 (113). Annu. Rev. Immunol. 1989.7:251-275. Downloaded from arjournals.annualreviews.org by HINARI on 08/29/07. For personal use only.

The Variable Region of Heavy Chains The aminoacid heterogeneity of V, peptides is not pronounced,since sequencecould be determinedfor the first frameworkeven in nonimmune samples(112, 114). Thediversity of the variable regions has beenestimated by IEF (llS), maturation of the immuneresponse (ll6), and idiotype analysis (112). All studies revealed a restricted heterogeneitywhich, addition to the inheritance of IEF spectrotypesand idiotypes in isogenic frogs (55, 112), suggested that somatic mutationplayed a minorrole Xenopus(117). TheIg genecluster (Figure 1) and the rearrangementshavebeenstudied to see whetherthe lowheterogeneitywasdueto a special geneorganization. The sequenceof the Vnpart of a completeIgMcDNA clone (105), or the sequence of a genomicVHclone obtained by cross-hybridization (llS), showedgreat conservationof the prototypic V/~structure. Mammalian and Xenopussequences showthe classical frameworksand complementarity determiningregions (CDR).The percentage of identical nucleotides for the entire VHis about 65%.As in most other species the genomicVH element is preceded by a conserved octamer and a TATA box 5’ to the split leader (Figure 1). Someputative DHand the JH elementsare highly conserved because up to 15 of 16 aminoacids of CDR3and framework4 are identical in human,Xenopusand shark. Oneputative D. sequenceis ?

?

?

7

(VHN) (VIlli) n (VHI)f~

(VHI)n

DH?

JH

",,,,,,, 2~

S~t

~

Cp.

C x

Cu

, =l,~, -"- ,,"¯~: ||-"| 1

234

12

34

1 2 3 4

7997 AGTCAAAT TATA

Figure 1 Genomicorganization of Xenopusimmunoglobulingene segments. 7 and 9: heptamersandnonamers.23 and 12, length of spacers in bp. S/t: putative switch region, C#, Cx, Co:genomicconstant regions genes. C/~, representedto scale, correspondto a fragment of 4.2 kb. A single VHelementis representedwith its octamer,TATA box and split leader. Thestructure of the D,uelementis hypothetic.

Annual Reviews


262

DU PASQUIER ET AL

identical (one gap of 2 bp necessary) to one of the shark DHelements (119). To this date, three different VHfamilies have been identified, based on the absence of cross-hybridization under stringent conditions. Southern blot analysis suggests that there are at least 60 VHgenes per haploid locus (119). This numberis large and in principle could generate more antibodies than are detected by IEF. It is possible that not all of them are used. The incidence of pseudogenes in the Vu elements so far analyzed is < 20%. Yet each V~/family seems to have a distinct set of putative regulatory sequences 5’ to the coding region. Within each family, sharing of entire CDRbetween one gene and the other has been noticed. Thus, such a pool of Vuelements mayindeed generate a smaller repertoire than that deduced from the number of its members(120). The cDNAsequences, and the recombination signals and spacers of VH and Jn (Figure 1) suggest the existence of Dnelements. As in other species, presumably the same element can be used in different frames. Genomic Xenopus Dn have not been isolated. Data from cDNAclones suggest that they encode 1 to 11 residues (119). Seven JH have been described. All have the appropriate recombination signal sequences and 22 bp spacers (119). Southern blots with B-cell DNAvisualized the rearrangements in Xenopus. There are multiple rearrangements presumably on both alleles (119). Theseevents result in allelic exclusion (108). The expression of these genes is illustrated in Figure 2 wheretwo typical immuneresponses are represented. Antibody responses are slower in Xenopus than in mammals;the cell division time for the temperature at which they live is about 24 hr. At the beginning only IgMis produced; then it is produced in conjunction with IgY (100). Incomplete switch has been reported, since many lymphocytes keep on producing IgM and IgY antibodies simultaneously (121). A modest affinity maturation of IgM has been noticed (Figure 1) (115). PFCnumbers, kinetics, titers, and spectrotypes are very similar between isogenic individuals (28, 55, 122). A second injection of antigen generated a stronger secondary response (memory). The affinity of the antibodies is immediately that of specific antibodies (115). The contribution of IgX to the immuneresponse unknown. THE

ONTOGENY

Tolerance

OF

THE

IMMUNE

SYSTEM

Induction

Based on transplantation of pituitary, Triplett’s experiments (123) suggested that an early hypophysectomized Hyla would recognize its own pituitary, grownon a transient host, as foreign. In Xenopusclones, better


10o~


100_

100.

263

10-8

10_

Time(weeks after immunization) Figure 2 Antibody responses in Xenopus. Left, plaque forming cell (PFC) response after an injection of sheep red blood cells. The upper and lower curves give the range of the response in a group of four outbred individuals. Right, anti-DNP response. Left Y axis: inactivation constant of DNP-T4bacteriophage. (,~, primary ~ secondary responses.) Right Yaxis: relative affinity estimation. Inhibitor concentration in inhibition of modifiedT4 bacteriophage inactivation assays measuringrelative affinities of antibodies to the DNP(black symbols) and the cross-reacting TNPligand (open symbols) in a primary (Q, ©) or secondary (I, []) response.

experiments contradicted Triplett’s results, A Xenopus grown without its eyes, thyroid, or pituitary will always recognize these isogenic organs as self. Thus, the processes of tolerance generation are not restricted to one phase of the ontogeny (124, 125). Tadpoles of Xenopus develop their allorecognition capacities around stage 49 (126). Whethera tadpole will reject or tolerate an allograft depends on a sensitive balance of parameters. As a rule (review in 132) minor nonMHCdifferences do not promote graft rejection in tadpoles. However, exceptions have been reported (e.g. HDfamily siblings; Ref. 125), and all possible thresholds of graft rejections have been recorded (129-135). Maintenance of tolerance depends on the continuous presence of the tolerizing graft (136). Somehave found a higher frequency of tolerance to MHC-disparate skin graft during metamorphosis (130, 134). Others, working under different genetic conditions, have found that the ability of tadpoles to reject grafts emerges gradually during ontogeny (128, 129). Metamorphosis,nevertheless, represents a crucial period (Figure 3; 127) separating two rather different immunesystems. Although tadpoles

Annual Reviews 264

DU PASQUIER

ET AL

LARVALSTAGES

PROMETAMORPHOSIS METAMORPHOSIS

ADULT AGEIN DAYS

0

10

20

30

14191,5~5’1 1~25~1814 1155

36 45

40

50

60

75

90 105

I I Ill 16364t 661I /,, I ,, I ,l I 5611517 58160611’ 59 62 65 STAGES


CIRCULATION BEGINS - HATCHING

~

PRE8-CELLS,N LIVERT

I

I I PREB-CELLSIN LIVER~" B-CELLSIN LIVER(1)ANDSPLEEN(2)

1 I 1 +2

IgM IN SERUM IgY IN SERUM TADPOLE REPERTOIRE "~’~,

ADULT REPERTOIRE FULLT-CELL HELP

L

[~y

THYMUS INVADEDBY STEMCELLS ~" THYMECTOMY AFFECTSALLOGRAFT RESPONSE AND MLR

MECTOMYAFFECTS ANTIBODYRESPONSE I Tx~TOL.

I

? I

? ’~= i INCOMPLETE

MLR ALLOGRAFT

IMPAIREDMLR REACTIVITY IMLR RESPONSES Ii~,~ ~ ALLO 9¯ ~. TRANSFER OF I SUPPRESSION I

MHCCLASS][ ONTHYMUS EPITHELIUM(1)AND 8-CELLS(2) MHCCLASS][ ONTHYMOCYTES AND T-CELLS SKINCLASS If + CELLS

I

MHCLINKEDAg ON THYMUS EPITHELIUM CLASSICAL MHCCLASS]~ RBCAG~ [___

Ilx103

¯ 3X104

3x105

9x108 1.5xlo s 9x105

lx108 2x106

lx107 I

THYMICLYMPHOCYTES NUMBERS

RELATIVETETRAIODOTHYRONIN CONCENTRATION Figure

3 The ontogeny

of the immune response

in )Senopus

(modified

from Ref.

135).

Abbreviations: Tx, thymectomy; Tol, tolerance. Skin class II÷cells: appearance of class II positive cells in the skin. RBC AG red blood cell antigen linked to MHC. Relative tetraiodothyronin concentration: visualization of level of the T4 hormone. For real values consult specific

chapters in Ref. 168.

Annual Reviews


IMMUNESYSTEMOF XENOPUS 265 express ~z and fl class-II chains identical to those of adults, their tissue distribution is different. Only 50-70%of tadpoles splenic lymphocytes including the B cells are class-II positive, whereasall adult splenic lymphocytes are positive (10). As determined by immunoprecipitation and immunofluorescence, the classical polymorphic membrane-boundclass-I molecules are not expressed before metamorphosis (137-139). The MLR is present during larval and adult life but is impaired at metamorphosis. The larval and adult immunesystems differ to the point that syngeneic larval anti-adult MLRcan be detected (136). It is not knownwhether class-I molecules are the stimulators for this MLR.MHC class-I and adult globin chains follow the same pattern of expression. Within l0 days after climax there is a large influx of erythroblasts producingadult hemoglobulin and class-I molecules 038). Lymphocytes from metamorphosing animals can suppress the response of a young adult to minor H disparate grafts (140). Moreover, cells from larval animals made tolerant to adult skin can inhibit the rejection of semiallogenic graft (133). Moreover, thymectomyat stage 56 decreases the number of tolerance cases in subsequent grafting (141), suggesting that suppression is an important mechanismfor tolerance induction and maintenance. The question of clonal deletion has never really been addressed. Further insight into tolerance came from thymectomies and reconstitution, and from embryonic chimaera experiments. Implantation of MHC-disparatethymusin late larval life of Tx hosts restores the response to MHCdisparate grafts but induces tolerance to skin from the thymus donor (142, 143). However, spleens of reconstituted animals can display MLRtowards the donor stimulator. This situation is analogous to the cases of split tolerance noticed in Tx animals reconstituted in adult life (144) and in head/body chimeras (89). In animals grafted at metamorphosis, tolerance to skin with retained proliferative responses has also been noticed when the animals are challenged in vitro by donor cells. It has been suggested that the cytotoxic effector part of the alloresponse is inhibited (145). These results concur with the observation that tolerated grafts are invaded by host-derived, class-II positive cells (10) and that certain chimaeras are able, upon challenge with head-type red cells, to synthesize antithymus MHC"auto" antibodies (146). So clearly, tolerant animals are not simply unable to recognize the tolerated antigens. Antibody

Responses

The maturation of B cells starts in the liver, at stage 45, where the first pre-B cells (surface IgM- cytoplasmic Ig +, L chain-) and then B cells can be detected (147), followed by the appearance of Ig in the serum (148)

Annual Reviews


266

otJ

PASQUIER ET AL

(Figure 3). There seem to be two wavesof pre-B cell production (149), this would account well for the detection of the two different antibody repertoires (139). The stability of the larval and the adult repertoires suggests that the rearrangements mayoccur only during these two waves. The ability of Xenopus tadpoles to recognize foreign antigens appears around stage 51-52 (148, 150, 151, 153). Tadpoles, probably limited in their T-cell function, produce IgMantibodies and have difficulties switching to IgY, which can be improvedby injection of adult T cells (152). Tx animals, reconstituted with larval thymus, show a normal repertoire (148, 149). "Bridges" exist between the larval and the adult immunesystems, since memoryis transferred through this period, at least in the case of T- and B-cell memoryfor an anti-DNP.KLHresponse (116) and to some extent for graft responses (153). The repertoire abruptly shifts from tadpole to adult at metamorphosis (E. Hsu, unpublished). With respect to the perimetamorphic period, the thymus at stage 57/58--when its cells bind less lectin than in larval or adult life--seems to lose its suppressionactivity on antibody responses (154). THE EFFECT OF POLYPLOIDY IMMUNE SYSTEM

ON THE

In principle the polyploid forms of this genus could express more than one MHCallele. Studies were performed using MLRas a MHCmarker in Xenopus of various ploidy (Table 1, 155). Diploid expression of MLR loci was seen in all species except X. ruwenzoriensis, apparently a recent polyploid. Thus, the need to becometolerant to manyself-MHCantigens appears to have been solved by the silencing of all but one diploid set of loci (156). In contrast to the natural polyploid, all constitutive MHC and other histocompatibility loci were expressed in laboratorymadepolyploids. This suggests that time is involved in the "silencing" phenomenon. With respect to immunogloblin production, allopolyploid species with three genomesof one species and one genomeof the other express specific Ig proportionally to the number of chromosomes.However, a lymphocyte never expresses more than one of its immunoglobulin constant-region Ig genes. This suggests that if a stochastic model for allelic exclusion is correct, the frequency of multiple successful rearrangements has to be very low. The resulting huge waste of lymphocyte precursors would be incompatible with the development of the Xenopus immunesystem characterized by its small numberof lymphocytes(108). Allopolyploidy could


267

a source of extra V genes which the individual might express. Indeed, in contrast to its effects on MHC,polyploidy is accompaniedby an increase in antibody diversity (157) (Table 1). Lab-madeaneuploid Xenopus (158) had their anti-DNP repertoire affected by the presence or the absence of a given chromosomein the genome (159).


CONCLUSION 1. The Xenopus lymphoid system is less complex than in mammals. A smaller numberof lymphocytes and an absence of lymph nodes are associated with a restricted antibody repertoire. This restriction maybe explained by the fact that in spite of an Ig-gene organization similar to that of mammals,the pool of usable VHelements may be smaller, and there do not seem to be somatic mutations. 2. The Xenopus larval immunesystem functions without the classical MHC class-I molecules expressed on cell surface. MHC serves as a marker of adulthood at the time of metamorphosis, in the hematopoietic lineage. The larval immunesystem differs from the adult, and it is not yet clear whichdifferences are due to class-I expression, class-II distribution, endocrine regulations, and cytokine production. 3. The use of thymectomy, embryonic chimaeras, and the natural situation at metamorphosis have deepened our knowledge about tolerance, stressing the role of regulatory interactions versus clonal deletion. 4. Evolutionary studies, madepossible by the polyploidy of someXenopus species, have revealed the trend for the expression of a single MHC and the strong pressure for allelic exclusion at the Ig locus level. For further development, many methods (160) and strains are now available in different species (161), not to mention more revolutionary tools such as interspecies somatic cell hybrids (162) or nuclear transplantation (163) and transgenic Xenopus(164). If nuclear transplantation has failed so far to provide an ideal tool for single cell genetics, it nevertheless has resulted in the production of tadpoles from single lymphocyte nuclei (165). It could be interesting to use it again in conjunction with embryonictransplants and to complementfuture transgenic frog studies. No lymphoid cell lines are available, but chemically induced tumors may provide this precious tool (166, 167). In the future, one can envision that endocrine regulations (154, 168), and the elucidation of the gene organization, and roles of membersof the Ig supergene family and complementwill be areas of predilection where the peculiar genetic, phylogenetic, and ontogenetic characteristics of Xenopuswill be exploited further.

Annual Reviews

268

ou PASQUIER ET AL

ACKNOWLEDGMENTS We thank Ms. S..Kirschbaum The Basel Institute’for F. Hoffman-La

for typing Immunology

Roche & Co. Ltd.,

Basel,

the manuscript. was founded and is

supported

by

Switzerland.


Literature Cited 1. Manning, M. J., Horton, J. D. 1982. RES structure and function of the amphibia. In The Reticuloendothelial system, Vol. 3. Phylogeny and Ontogeny, ed. N. Cohen, M. M. Sigel, pp. 424-59. NewYork: Plenum 2. Clothier, R. J., Balls, M. 1985. Structural changes in the thymus glands of Xenopus laevis during development. In Metamorphosis, ed. M. Balls, M. Bownes, pp. 332-59. Oxford: Clarendon 3. Sterba, G. 1950. ’Uber die morphologischen und histogenetischen Thymusprobleme bei Xenopuslaevis nebst eini~ gen Bemerkungfiber die Morphologie der Kaulquappen. Abh. Sachs. Akad. Wiss. 44:1-54 4. Nieuwkoop,P. D., Faber, J. 1987. Normal Table of Xenopus laevis. Amsterdam: North-Holland 5. Maeno, M., Todate, A., Katagiri, C. 1985. The localization of precursor cells for larval and adult hemopoietic cells in Xenopus laevis in two regions of embryos. Dev. Growth Differ. 27: 137236 6. Tochinai, S. 1978. Thymocytestem cell inflow in Xenopuslaevis after grafting diploid thymic rudiments into triploid tadpoles. Dev. Comp. Immunol. 2:622 35 7. Tompkins, R., Volpe, E. P, Reinschmidt, D. C. 1980. Origin of hemopoietic stem cells in amphibian ontogeny. In Development and Differentiation of Vertebrate Lymphocytes, ed. J. D. Horton, pp. 25-34. Amsterdam: Elsevier/North-Holland 8. Flajnik, M. F., Horan, P. K., Cohen, N. 1984. A flow cytometric analysis of the embryonicorigin of lymphocytesin diploid/triploid chimeric Xenopus laevis. Dev. Biol. 104:247-54 9. Turpcn, J. B., Cohen, N., Deparis, P., Jaylet, A., Tompkins, R., Volpe, E. P. 1982. Ontogeny of amphibian hemopoietic cells. In The Reliculoendothelial System, Vol. 3. Phylogeny and Ontogeny, ed. N. Cohen, M. M. Sigel, pp. 569-88. New York: Plenum

10. DuPasquier, L., Flajnik, M. F. 1987. Xenopus MHCclass II antigens. Ann. Rep. Baxel Inst. Immunol., pp. 47-48 11. Manning, M. J., Horton, J. D. 1969. Histogenesis of lymphoidorgans in larvae of the South African clawed toad Xenopus laevis. J. Embryol. Exp. Mor.ph..22: 265-77 12. ?Nagata,.S. 1976. An electron microscopic study on the thymus of larval and metamorphosed toads, Xenopus laevis (Daudin). J. Fae. Sci., Hokkaido Univ. Ser. VI Zool. 20:236-71 13. Nagata, S. 1977. Electron microscopic study on the early histogenesis of thymusin the toad, Xenopus laevis. Cell. Tiss. Res. 179:87 96 14. Flajnik, M. F., Hsu, E., Kaufman, J. F., Du Pasquier, L. 1988. Biochemistry, tissue distribution and ontogeny of surface molecules detected on Xenopus hemopoietic cells. In Differentiation Antiyens in Lymphohemopoietic Tissues, ed. M. Miyasaka, Z. Trnka, pp. 387-419. NewYork: M. Dekker 15. Bigaj, J., Plytycz, B. 1987. Interdigitating cells in the thymusof the frog Rana temporaria. Fol. Histochem. Cytol. 24:65-68 16. Russ, J. H., Horton, J. D. 1987. Cytoarchitecture of the Xenopusthymus following y-irradiation. Development100: 95-105 17. Du Pasquier, L., Weiss, N. 1973. The thymus during the ontogeny of the toad Xenopus laevis: growth, membrane bound immunoglobulins and mixed lymphocyte reaction. Eur. J. lmmunol. 3:773-77 18. Rollins-Smith, L. A., Parsons, S. C. V., Cohen, N. 1984. During frog.ontogeny, PHA and Con A responsiveness of splenocytes precedes that of thymocytes. Immunology 52:491-500 19. Sterba, G. 1952. Mitteilung fiber die Altersinvolution des Amphibien Thymus. I. Volumetrische Bestimmungen am Thymus des Krallenfrosches Xenopus laevis Daud. Anat. Ariz. 99: 10614

Annual Reviews


IMMUNE SYSTEM OF XENOPUS 20. Nagata, S. 1985. Acell surface marker of thymus dependent lymphocytes in Xenopuslaevis is identifiable by mouse monoclonal antibody. Eur. J. Immunol. 15:83741 21. Nagata, S. 1986. Development of T lymphocytes in Xenopus laevis: Appearance of the antigen recognized by an anti thymocyte mouse monoclonal antibody. Dev. Biol. 114: 38994 22. Nagata, S. 1986. T cell proliferative responses of Xenopus lymphocyte subpopulations separated on anti-thymocyte monoclonal antibody coupled to scpharose beads. Dev. Comp.Immunol. 10:259-64 23. Nagata, S. 1988. T cell-specific antigen in Xenopus identified with a mouse monoclonal antibody: Biochemical characterization and species distribution. Zool. Sci. 5:77-83 24. Horton, J. D., Manning, M. J. 1972. The response to skin allografts in J(enopus laevis following thymectomy at early stages of lymphoid organ maturation. Transplantation 14:141-54 25. Tochinai, S., Katagiri, C. 1975. Complete abrogation of immuneresponse to skin allografts and rabbit erythrocytes in the early thymectomized Xenopus. Devel. Growth. Diff. 17: 383-94 26. Du Pasquier, L., Horton, J. D. 1976. The effect of thymectomyon the mixed leukocyte reaction and phytohemagglutinin responsiveness in the clawed toad Xenopus laevis. Immunogenetics 3:105 12 27. Donnelly, N., Manning, M. J., Cohen, N. 1976. Thymus dependency of lymphocyte subpopulations in Xenopus laevis. In Phylogeny of Thymus and Bone Marrow Bursa Cells, ed. R. K. Wright, E. L. Cooper, pp. 133~41. Amsterdam: Elsevier/North-Holland 28. Du Pasquier, L., Wabl, M. R. 1977. The ontogeny of lymphocyte diversity in anuran amphibians. 9CoM Sprin Harbor Symp. Quant. Biol. 41:771 79 29. Manning, M. J., Donnelly, N., Cohen, N. 1976. Thymus dependent and thymus independent components of the amphibian immune system. In Phylo#eny of Thymus and Bone Marrow Bursa Cells, ed. R. K. Wright, E. L. Cooper, pp. 123 32. Amsterdam: Elsevier/North-Holland 30. Collie, M. H., Turner, R. J., Manning, M. J. 1975. Antibody production to lipopolysaccharide in thymectomized Jgenopus laevis. Eur. J. Immunol. 5: 426-27

269

31. Horton, J. D., Edwards, B. F., Ruben, L. N., Mette, S. 1979. Use of different carriers to demonstrate thymic dependent or thymic independent antitrinitrophenyl reactivity in the amphibian, Xenopus laevis. Dev. Comp. lmmunol. 3:621-33 32. James, H. S., Knowles,K. R., Clothier, R. H., Groves, C. J., Balls, M. 1983. Effect of early thymectomyor exposure to N-Methyl-N-nitrosourea on immuneresponses in Xenopus laevis. In Proc. 1st Int. Symp.on Pathol. Reptiles Amphibians, ed. C. Vago, G. Matz, pp. 157-62. Angers: University of Angers 33. Turner, R. J., Manning, M. J. 1974. Thymicdependence of amphibian antibody response. Eur. J. lmmunol.4: 34346 34. Hsu, E., Flajnik, M. F., Du Pasquier, L. 1985. A third immunoglobulinclass in amphibians. J. Immunol. 135: 19982004 35. Weiss, N., Horton, J. D., DuPasquier, L. 1973. The effect of thymectomyon cell surface associated and serum immunoglobulin in the toad Xenopus laevis: a possible inhibitory role of the thymus on the expression of immunoglobulins. In L’Etude Phylogknique et Ontoybnique de la R~ponse Immunitaire et son apport gtla thborie immunologique, ed. J. Panijel, P. Liacopoulos, pp. 165-74. Paris: INSERM 36. Manning, M. J., Collie, M. H. 1975. Thymic function in amphibians. Adv. Exp. Biol. Med. 64:353 62 37. Horton, J. D., Sherif, N. E. H. S. 1977. Sequential thymectomy in the clawed toad: effect on mixed leucocyte reactivity and PHAresponsiveness. In Developmental Immunobiology, ed. J. B. Solomon, J. D. Horton, pp. 283-90. Amsterdam: Elsevier/North-Holland 38. Hsu, E., Julius, M. H., DuPasquier, L. 1983. Effector and regulator functions of splenic and thymic lymphocytes in the clawed toad Xenopus. Ann. Immunol. (Inst. Pasteur) 134D:277-92 39. Ruben, L. N., Mette, S. A., Edwards, B., Cochran, S. 1980. Thymus dependent suppression of helper function in adult Xenopuslaevis, the South African clawed toad. Thymus 2:19-25 40. Ruben, L. N., Buenafe, A., Oliver, S., Malley, A., Lukas, D. 1985. Suppression in Xenopuslaevis: thymusinducer, spleen effector cells. Immunology54: 65-70 41. Ruben, L. N., James, H. S., Clothier, R. H., Balls, M. 1984. Genetic limits of thymic immunosuppression of antihapten antibody production in Xenopus

Annual Reviews


270

19tJ

PASQUIER ET AL

laevis laevis, the South African clawed toad. J. Immunogenet. 11:97-101 42. Ruben,L. N., Clothier, R. H., Bucnafe, A., Needham,P., James, H. S., Balls, M. 1984. In vitro thymus suppression of hemagglutinin production in Xenopus laevis: location, drug and temperature sensitivity. Thymus6". 143-52 43. Ruben, L. N., Buenafe, A., Seivert, D. 1983. Somecharacteristics of thymus suppression of antibody production in vitro in Xenopus laevis the South African clawed toad. Thymus 5:13-18 44. Cohen, N., Watkins, D., Parsons, S. C. 1987. Interleukins and T-cell ontogeny in Xenopuslaevis. In Developmentaland Comparative Immunology, ed. E. L. Cooper, C. Langlet, J. Bievne, pp. 5368. NewYork: A. R. Liss Inc. 45. Williams, N. H., Cribbin, F. A., Zettergren, L. D., Horton, J. D. 1983. Ontogeny and characterization of mitogen-reactive lymphocytes in the thymus and spleen of the amphibian, Xenopus laevis. Immunology49: 301~ 46. Du Pasquier, L., Horton, J. D. 1982. Restoration of antibody responsiveness in early thymectomized Xenopus by implantation of MHCmismatched larval thymus. Eur. J. Immunol. 12: 546-51 47. Sterba, G. 1951. Untersuchungen an der Milz des Krallenfrosches (Xenopus laevis Daudin). Morphol.Jahr. 90: 22148 48. Manning, M. J. 1971. The effect of early thymectomy on histogenesis of the lymphoid organs in Xenopus laevis. J. Embryol. Exp. Morphol. 26: 21929 49. Horton, J. D., Manning, M. J. 1974. Lymphoid organ development in Xenopus thymectomizedat eight days of age. J. Morphol. 143:385-95 50. Tochinai, S. 1976. Lymphoid changes in Xenopus laevis following thymectomy at the initial stage of its histogenesis. J. Fac. Sci. Hokkaido Univ. Sci. 6, 20:175-82 51. Baldwin, W. M. III, Cohen, N. 1981. A primitive dendritic splenocyte in Xenopus laevis with morphological similarities to Reed-Sternbergcells¯ In Aspects of Developmental and Comparative Immunology, Led. J. B. Solomon, pp. 179-82. Oxford: Pergamon 52. Collie, M.H. 1974. The location of soluble antigen in the spleen of Xenopus laevis. Experientia 30:1205-7 53. Horton, J. D., Manning, M. J. 1974. Effect of early thymectomyon the cellular changes occurring in the spleen of the clawed toad following administra-

tion of soluble antigen. Immunology26: 797-8O7 54. Turner, R. J., Manning, M. J. 1973. Responseof the toad Xenopuslaevis, to circulating antigens. I. Cellular changes in the spleen. J. Exp. Zool. 183:21-33 55. Du Pasquier, L., Wabl, M. R. 1978. Antibody diversity in amphibians, inheritance ofisoelectric focussing antibodypatterns in isogenic frogs. Eur. J. Immunol. 8:428-33 56. Bleichcr, P. A., Rollins-Smith, L. A., Jacobs, D. M., Cohen, N. 1983. Mitogenic responses of frog lymphocytes to crude and purified preparations of bacterial lipopolysaccharide (LPS). Dev. Comp. Immunol. 7:48346 57. Bleicher, P. A., Cohen, N. 1981. Monoclonal anti-IgM can separate T-cell from B-cell proliferative responses in thc frog Xenopus laevis. J. Immunol. 127:1549-55 58. Schwager, J., Hadji-Azimi, I. 1986. Mitogen induced B-cell differentiation in Xenopus laevis. Differentiation 27: 182-88 59. Schwager, J., Hadji-Azimi, I. 1985. Anti-immunoglobulin M induces both B-lymphocyteproliferation and differentiation in Xenopus laevis. Differentiation 30:29-34 60. Hsu, E., Leanderson, T., Franklin, R. M. 1985. Mitogenic effects ofphorbol myristate acetate (PMA) amphibian cells. Ann. Immunol. (Inst. Pasteur) 136D: 105-18 61. Rollins-Smith, L. A., Parsons, S. C., Cohen, N. 1984. During frog ontogeny PHA and Con A responsiveness of splenocytes precedes that of thymocytes. Immunology 52:491-500 62. Nagata, S., Cohen, N. 1983. Specific in vivo and nonspecific in vitro alloreactivities of adult frogs (Xenopus laevis) that were thymectomizedduring ¯ early larval life. Eur. J. lmmunol.13(7): 541-45 63. Turner, R. J. 1973. Response of the toad, Xenopus laevis, to circulating antigens. II. Responses after splenectomy. J. Exp. Zool. 183:35-45 64. Tochinai, S. 1975. Distribution oflympho-epithelial tissues in the larval African clawed toad Xenopus laevis (Daudin). J. Fac. Sci. Hokkaido Univ. Sci. 6, 19:803-11 65. DuPasquier, L., Hsu, E. 1987. Further studies on IgX in Xenopus. Ann. Rep. Basel Inst. Immunol. 1986:38 66. Hadji-Azimi, I., Fischberg, M. 1967. Hematopoi6se p6rih6patique chez le batracien anoure Xenopus laevis. Comparaison entre les individus normaux

Annual Reviews


IMMUNE SYSTEM OF XENOPUS et les porteurs de tumeurs lympho~des. Rev. Suisse. Zool. 74:641-45 67. Turner, R. J. 1970. The influence of colloidal carbon on hemagglutinin production in the toad Xenopus laevis. J. Reticuloendothelial Soc. 8:434-45 68. Green, N., Cohen, N. 1979. Phylogeny of immunocompetentcells. III. Mitogen response characteristics of lymphocyte subpopulations from normal and thymectomized frogs (Xenopus laevis). Cell. Immunol.48:59-70 69. Turner, R. J. 1969. The functiorial developmentof the reticuloendothelial system in the toad Xenopuslaevis (Daudin). J. Exp. Zool. 170:467 80 70. Watkins, D., Parsons, S. C., Cohen, N. 1987. Afactor with interleukin-l-like activity is producedby peritoneal cells from the frog Xenopus laevis. Immunology 62:669-73 71. Hadji-Azimi, I., Coosemans,V., Canicatti, C. 1987. Atlas of adult Xenopus laevis hematology. Dev. Comp. Immunol. 11:807-74 72. Blomberg, B., Bernard, C. C. A., Du Pasquier, O. 1980. In vitro evidence for T-B lymphocyte collaboration in the clawed toad, Xenopus. Eur. J. Immunol. 10:86%76 73. Bernard, C. C. A., Bordmann, G., Blomberg, B., Du Pasquier, L. 1981. Genetic control of T helper function in the clawed toad Xenopuslaevis. Eur. J. lmmunol. 11:151-55 74. Kau, C. L., Turpen, J. B. 1983. Dual contribution of embryonic ventral blood island and dorsal lateral plate mesoderm during ontogeny of hemopoietic cells in Xenopuslaevis. J. Immunol. 131: 2262-66 75. Turpen, J. B., Knudson, C. M., Hoefen, P. S. 1981. The early ontogeny of hematopoietic cells studied by grafting cytogenetically labeled tissue anlagen: localization of a prospective stem cell compartment. Dev. Biol. 85: 99-112 76. Katagiri, C., Ma6no,M., Tochinai, S. 1986. Differential commitment of hemopoieticstem cells localized in distinct compartments of early Xenopus embryos. Curr. Top. Dev. Biol. 20:315 23 77. Thiebaud, C. H. 1983. A reliable cell marker in Xenopus. Dev. Biol. 98: 24549 78. Ma6no,M., Tochinai, S., Katagiri, Ch. 1985. Differential participation of ventral and dorsolateral mesodermsin the hemopoiesis of Xenopus, as revealed in diploid-triploid or interspecific chimeras. Dev. Biol. 110:503-8

271

79. Flajnik, M. F., Horan, P. K., Cohen, N. 1984. A flow cytometric analysis of the embryonic origin of lymphocytes in diploid/triploid chimeric Xenopus laevis. Dev. Biol. 104:247-54 80. Turpen, J. B., Smith, P. B. 1985. Dorsal lateral plate mesoderminfluences proliferation and differentiation of hemopoietic stem cells derived from ventral lateral plate mesoderm during early development of Xenopus laevis embryos. J. Leuk. Biol. 38:415 27 81. Turpen, J. B., Smith, P. B. 1986. Analysis of hemopoietic lineage of accessory cells in the developing thymus of Xenopus laevis. Dev. Biol. 136:412-21 82. Turpen, J. B., Cohen, N., Deparis, P., Jaylet, A., Tompkins, R., Volpe, E. P. 1982. Ontogeny of amphibian hemopoietic cells. In TheReticuloendothelial System, Vol. 3. Phylogeny and Ontogeny, ed. N. Cohen, M. M. Sigel, pp. 569-88. NewYork: Plenum 83. Du Pasquier, L., Miggiano, V. C. 1973. The mixed leukocyte reaction in the toad Xenopus laevis: a family study. Transpl. Proe. 5:1457-61 84. Du Pasquier, L., Chardonnens, X., Miggiano, V. C. 1975. A major histocompatibility complex in the toad Xenopus laevis (Daudin). Immunogenetics 1:482-94 85. Flajnik, M. F., Kaufman, J. F., Riegert, P., Du Pasquier, L. 1986. Identification of class I MHCencoded molecules in the amphibian Xenopus. Immunogenetics 20:433~,2 86. Kaufman, J. F., Flajnik, M. F., Du Pasquier, L., Riegert, P. 1985. Xenopus MHCclass II molecules. I. Identification and structural characterization. J. Immunol. 134:3248-57 87. Kaufman, J. F., Flajnik, M. F., Du Pasquier, L. 1985. Xenopus MHCclass II molecules. II. Polymorphism as determined by two dimensional gel electrophoresis. J. Immunol. 134:3258 64 88. Kobel, H. R., Du Pasquier, L. 1977. Strains and species of Xenopus for immunological research. In Developmental Immunobiology, ed. J. B. Solomon, J. D. Horton, pp. 299 306. Amsterdam: North-Holland 89. Flajnik, M. F., DuPasquier, L., Cohen, N. 1985. Immune responses of thymus/lymphocyte embryonic chimeras: studies on tolerance and MHCrestriction in Xenopus. Eur. J. lmmunol. 15:540-47 90. Nakamura,T., Sekizawa, A., Fujii, T., Katagiri, Ch. 1986. Cosegregation of the polymorphic C4 with the MHCin

Annual Reviews


272

DU PASQUIER ET AL

the frog Xenopus laevis. Immunoglobulin. Ultrastructure and J chain 9enetics 23:18186 isolation. Immunoloyy30:587-91 91. DuPasquier, L., Hsu, E. 1983. Poly- 104. Parkhouse,R. M. E., Askonas,B. A., morphismof histocompatibility anDoormashkin,R. R. 1970. Electron microscopicstudies of mouseimmunotigens and MLRdeterminants in Xenopus. Basel Inst. Immunol. Ann. globulin M. structure and reconRep. 1983:34 struction followingreduction. Immun92. Fujii, T., Sekizawa,A., Katagiri, C. ology 18:575-84 1985. Characterization of the fourth 105. Schwager, J., Mikoryak, C. A., Steiner, componentof complementin the serum L. A. 1988. Aminoacid sequences of heavychains from Xenopuslaevis IgM of the clawed frog Xenopuslaevis. Immunology56:743 50 deduced from eDNAsequence: implications for evolution of immuno93. Grossberger, D., Marcuz, A., Du Pasquier, L., Lambris,J. D. 1988. The globulin domains. Proe. Natl. Aead. conservation of structural and funcSci. USA85:2245-49 tional domainsin the third complement 106. Schwager,J., DuPasquier, L., Grosscomponent(C3) of Xenopusand mamberger, D., Kiefer, H. 1986. Genomic clonesof XenopusIg genes. BaselInst. mals. In press 94. Flajnik, M. F., Kaufman,J. F., Du Immunol.Ann. Rep. 1986:37-38 Pasquier, L. 1984.Studies on a Xenopus 107. Schwager,J., Du Pasquier, L. 1987. Rearrangementand switch regions at erythrocyteantigen. Basel lnst. lmmunol. Ann.Rep. 1984:63 the XenopusIgH locus. Basel Inst. 95. Bernard, C. C. A., Bordmann, G., lmmunol.Ann. Rep. 1987:40 Blomberg,B., Du Pasquier, L. 1979. 108. Du Pasquier, L., Hsu, E. 1983. Immunoglobulin expression in diploid Immunogeneticstudies on the cellmediated cytotoxicity in the clawed and polyploid interspecies hybrids of toad Xenopusluevis. Immunogenetics 9: Xenopus: evidencefor allelic exclusion. Eur. J. Immunol.13:585-90 443-54 96. Du Pasquer, L., Horton, J. D. 1982. 109. Hadji-Azimi,I. 1977. Distribution of Restoration of antibody responsiveimmunoglobulindeterminants on the surface of Xenopuslaevis splenic lymness in early thymectomizedXenopus phocytes. J. Exp. Zool. 201:115-26 by implantation of MHC mismatched larval thymus. Eur. J. Immunol.12: 110. Hadji-Azimi, I., Schwager,J. 19801 Xenopuslaevis larval thymocytesdo 546-51 97. Gearing,A. J., Cribbin,F. A., Horton, not express surface immunoglobulin. J. D. 1984. Restoration of the antiCell Immunol. 53:389-94 11 I. DuPasquier, L., Weiss, N., Loor, F. body response to sheep erythrocytes in thymeetomizedXenopusimplanted 1972. Direct evidence for immunoglobulins on the surface of thymus with MHC-compatible or MHC-inlymphocytesof amphibianlarvae. Eur. compatible thymus. J. Embryol.Exp. Morphol.84:287-302 J. Immunol.2:366-70 98. Kobel, H. R., DuPasquier, L. 1975. 112. Brandt, D. C., Griessen, M., Du Production of large clones of histoPasquier, L., Jaton, J. C. 1980.Antibodydiversity in amphibians:evidence compatible,fully identical clawedtoads for the inheritanceof idiotypiespecifi(Xenopus). lmmunogeneties 2: 8791 cities in isogenic Xenopus. Eur. J. Immunol. 10:731-36 99. Watkins, D., Harding, F., Cohen,N. 1988. In vitro proliferative and cyto- 113. Zezza, D. J., Mikoryak, C. A., toxic responses against Xenopusminor Schwager,J., Steiner, L. A. 1988. histocompatibility antigens. TransAmino acid sequence of constant region of Ig light chains fromXenopus plantation 45:499-501 100. Hadji-Azimi,I. 1979. Anuranimmunolaevis. FASEB J. 2: (Abstr. 2317) globulin, a review. Dev. Comp.Immu- 114. Wang,A. C., Tung, E., Fudenberg, nol. 3:223-43 H. H., Hadji-Azimi, I. 1978. Immunoglobulin evolution: chemical I01. Hadji-Azimi,I. 1971. Studies on Xenostudy of clawedtoad (Xenopuslaevis) pus laevis immunoglobulins.Immunology 21: 463-74 heavyandlight chains. J. Immuno#enet. 5:355-64 ¯ 102: Hsu,E., DuPasquier, L. 1984.Studies 115. Wabl, M. R., DuPasquier, L. 1976. on Xenopus immunoglobulins using Antibody patterns in geneticallyidentimonoclonalantibodies. Mol. Immunol. cal frogs. Nature264:642-44 21:257-70 103. Hadji-Azimi, I., Michea-Hamzehpour, 116. DuPasquier, L., Haimovich,J. 1976. The antibody response during amphiM. 1976. Xenopus laevis immuno-

Annual Reviews


IMMUNE SYSTEM OF XENOPUS bian ontogeny. Immunogeneties 3:38191 117. Du Pasquier, L. 1982. Antibody diversity in lower vertebrates--why is it so restricted? Nature 296:311-13 118. Yamawaki-Kataoka, Y., Honjo, T. 1987. Nucleotide sequences of variable region segments of the immunoglobulin heavy chain of Xenopuslaevis. Nucleic Acids Res. 15:5888 119. Schwager, J., Grossberger, D., Du Pasquier, L. 1988. Organization and rearrangement of immunoglobulin M genes in the amphibian Xenopus. EMBOJ. 7:2409-15 120. Du Pasquier, L., Schwager, J. 1987. Diversity of VHgenes and Vnfamilies in Xenopus. Basel Inst. Immunol. Ann. Rep. 1987:32 121. Hadji-Azimi, I., Parrinello, N., Perrenot, N. 1978. The simultaneous production of two classes of cytoplasmic immunoglobulins by single cells in Xenopuslaevis. Cell. Immunol.39: 31624 122. Du Pasquier, L., Wabl, M. R. 1976. Antibody diversity studied in amphibians. In The Generation of Antibody Diversity: A New Look, ed. A. J. Cunningham, pp. 151-64. London: Academic 123. Triplett, E. 1962. On the mechanismof immunologicself recognition. J. Immunol. 89:505-10 124. Rollins-Smith, L. A., Cohen, N. 1982. Self-pituitary grafts are not rejected by frogs deprived of their pituitary anlagen as embryos. Nature 299: 82021 125. Rollins-Smith, L. A., Cohen, N. 1983. The Triplett phenomenon revisited: self tolerance is not confined to the early developmentalperiod. Transplant. Proc. 15:871-74 126. Horton, J. D. 1969. Ontogeny of the immuneresponses to skin allografts in relation to lymphoid organ development in the amphibian Xenopus laevis Daudin. J. Exp. Zool. 170:449-66 127. Cohen, N., Dimarzo, S., RollinsSmith, L., Barlow, E., VanderschmidtParsons, S. 1985. The ontogeny of allo tolerance and self-tolerance in larval Xenopus laevis. In Metamorphosis, ed. M. Balls, M. Bownes, pp. 388-419. Oxford: Clarendon 128. Obara, N., Kawahara, H., Katagiri, C. 1983. Response to skin grafts exchanged amongsiblings of larval and adult gynogenetic diploids in Xenopus laevis. Transplantation 36:91-95 129. Du Pasquier, L, Chardonnens, X. 1973. Induction of skin allograft toler-

273

ance during metamorphosis of the toad Xenopus laevis: a possible model for studying generation of self-tolerance to histocompafibility antigens. Eur. J. lmmunol. 3:569-73 130. Du Pasquier, L., Chardonnens, X. 1975. Genetic aspects of the tolerance to allografts induced at metamorphosis in the toad Xenopus laevis, lmmuno9enetics 2:431-40 131. Barlow, E. H., Dimarzo, S. J., Cohen, N. 1981. Prolonged survival of major histocompatibility complex disparate skin allografts transplanted to the metamorphosingfrog, Xenopus laevis. Transplantaion 32:51-57 132. Dimarzo, S. J., Cohen, N. 1982. Immunogeneticaspects of in vivo allotolerance induction during the ontogeny of Xenopus laevis. Immunogenetics 16:103-16 133. Nakamura, T., Marno, M., Tochinai, S., Katagiri, C. 1987. Tolerance induced by grafting semi-allogeneic adult skin to larval Xenopus laevis: possible involvement of specific suppressor cell activity, Differentiation 35: 108-14 134. Bernardini, N., Chardonnens, X., Simon, D. 1969. Drveloppement aprrs la m6tamorphose des comp~tences immunologiques envers les homogreffes cutan+es chez Xenopuslaevis (Daudin). C. R. Acad. Sci. Paris 269D: 1011 14 135. Flajnik, M. F., Hsu, E., Kaufman, J. F., Du Pasquier, L. 1987. Changes in the immune system during lnetamorphosis of Xenopus. Immunol. Today 8:58-64 136. Kaye, C., Schermer, J. A., Tompkins, R. 1983. Tolerance maintenance~de~ pends on persistence of the tolerizing antigen: evidence from transplantatiOn’ . on Xenopus laevis. Dev. Comp. lmmunol. 7:497-506 M. F., Kaufman, J. F., 137. Flajnik, Hsu, E., Manes, M., Parisot, R., Du Pasquier, L. 1986. Major histocompatibility complex-encoded class 1 molecules are absent in immunologically competent Xenopus. before metamorphosis. J. Immunol.’q37:38"91-99 138. Flajnik, M. F., DuPasquier, L. 1988. MHCclass I antigens as surface markers of adult erythrocytes during the metamorphosis of Xenopus. Devel. Biol. 128:198-206 139. Du Pasquier, L., Blomberg, B., Bernard, C. C. A. 1979. Ontogeny of immunity in amphibians: changes in antibody repertoire and appearance of adult MHCantigens J. Immunol. 9:900~5in Xenopus. Eur.

Annual Reviews


274

DU PASQUIER ET AL

140. Du Pasquier, L., Bernard, C. C. A. 1980. Active suppression of allogeneic histocompatibility reactions during metamorphosis of the clawed toad Xenopus. Differentiation 16:1-7 141. Barlow, E. H., Cohen, N. 1983. The thymus dependency of transplantation allotolerance in the metamorphosing frog, Xenopus laevis. Transplantation 35:612-19 142. Horton, S. D., Horton, T. 1975. Development of transplantation immunity and restoration experiments in the thymectomized amphibian. Am. Zool. 15:73 143. Nagata, S., Cohen, N. 1984. Induction ofT cell differentiation in early thymectomized Xenopus by grafting adult thymuses from either MHC-matched or from partially or totally MHCmismatched donors. Thymus 6: 89103 144. Arnall, J. C., Horton, J. D. 1986. Impaired rejection of minor histocompatibility antigen-disparate skin grafts and acquisition of tolerance to thymus donor antigens in allothymus implanted, thymectomized Xenopus. Transplantation 41:766-76 145. Arnall, J. C., Horton, J. D. 1987. In vivo studies on allotolerance perimetamorphically induced in control and thymectomized Xenopus. Immunology 62:315-19 146. Flajnik, M. F., Du Pasquier, L. 1986. Anti-MHCauto antibodies in Xenopus head-body chimeras. Basel Inst. Ann. Rep. 1986:96-97 147. Hadji-Azimi, 1., Schwager, J., Thiebaud, C. 1982. B lymphocyte differentiation in Xenopuslaevis larvae. Devel. Biol. 90:253-58 148. Hsu, E., Du Pasquier, L. 1984. Ontogeny of the immunesystem in Xenopus. I. Larval immuneresponse. Differentiation 28:109 15 149. Hadji-Azimi, 1., Coosemans,V., Canieatti, C. 1988. B lymphocyte population in Xenopuslaevis. Submitted 150. Kidder, G. M., Ruben, L. N., Stevens, J. M. 1973. Cytodynamics and ontogeny of the immune response of Xenopus laevis against sheep erythrocytes. J. Embryol. Exp. Morph.29: 7385 151. Hsu, E., Du Pasquier, L. 1984. Ontogeny of the immune system in Xenopus. II. Antibody repertoire differences between larvae and adults. Differentiation 28:116 22 152. Jurd, R. D., Luther-Davies, S. M., Stevenson, G, T. 1975. Humoral antibodies to soluble antigens in larvae

of Xenopus laevis. Comp. Biochem. Physiol. 50B: 65-70 153. Manning, M. J., A1Johari, G. M. 1985. Immunological memory and metamorphosis. In Metamorphosis, ed. M. Balls, M. Bownes,pp. 420-39. Oxford: Clarendon 154. Ruben, L. N., Clothier, R. H., Jones, S. E., Bonyhadi, M. L. 1985. The effect of metamorphosis on the regulation of humoral immunity in Xenopus laevis, the South African clawed toad. In Metamorphosis, ed. M. Balls, M. Bownes, pp. 360-87. Oxford: Clarendon 155. Kobel, H. R., Du Pasquier, L. 1986. Genetics of polyploid Xenopus. Trends Genet. 2:310-15 156. Du Pasquier, L., Miggiano, V. C., Kobel, H. R., Fischberg, M. 1977. The genetic control of histocompatibility reactions in natural and laboratorymade polyploid individuals of the clawed toad Xenopus. Immunogenetics 5:12941 157. Du Pasquier, L., Blomberg, B. B. 1982. The expression of antibody diversity in natural and laboratory-made polyploid individuals of the clawed toad Xenopus. Immunogenetics 15:25140 158. Kobel, H. R., Du Pasquier, L. 1979. Hyperdiploid species hybrids for gene mapping in Xenopus. Nature 279: 15758 159. Du Pasquier, L., Kobel, H. R. 1979. Histocompatibility antigens and immunoglobulins genes in the clawed toad: expression and linkage studies in recombinant and hyperdiploid Xenopus hybrids. Immunogenetics 8: 299310 160. DuPasquier, L., Flajnik, M. F., Guiet, C., Hsu, E, 1985. Methods used to study the immune system of Xenopus (Amphibia, Anura). In Immunological MethodsIII, ed. I. Lefkovits, B. Pernis, pp. 425-65. NewYork: Academic 161. Afifi, A., Picard, J. J., Querinjean, P. 1985. A partially histocompatible family of Xenopus borealis. Lab. Amin. Sci. 35:139~J,1 162. Hengartner, H., Du Pasquier, L. 1981. Somatic cell hybrids from frog lymphocytes and mousemyelomacells. Science 212:I034~35 163. Du Pasquier, L., Wabl, M. R. 1977. Transplantation of nuclei from lymphocytes of adult frogs into enucleated eggs: special focus on technical parameters. Differentiation 8:9-19 164. Etkin, L. D. 1986. Gene expression in transgenic Xenopus laevis. Prog. Clin. BioL Res. 217A: I1 16

Annual Reviews IMMI.YNE SYSTEM OF XENOPUS

275


165. Wabl, M. R., Brun, R., Du Pasquier, treated Xenopuslaevis. In Proc. 1st Int. L. 1975. Lymphocytes of the toad Colloquium on Patholoyy of Reptiles Xenopus laevis have the gene set for and Amphibians, ed. C. Vago, G. Matz, promoting tadpole development. Scipp. 163-72. Angers: Univ. Angers ence 190:1310-12 168. Balls, M., Clothier, R. H., Rowles, 166. Balls, M., Clothier, R. H., Knowles, J. M., Kiteley, N. A., Bennett, G. W. K. R. 1981. The phylogeny of tumour 1985. TRHdistribution, levels and sigimmunity. Dev. Comp. Immunol. 5 nificance during the development of (Suppl.) 1:37-48 Xenopus laevis. In Metamorphosis,ed. 167. Balls, M., Clothier, R. H., Knowles, M. Balls, M. Bownes, pp. 260-70. K. R. 1983. Tumourincidence in NMU Oxford: Clarendon






MOLECULAR GENETICS OF CHRONIC GRANULOMATOUS DISEASE Stuart

H. Orkin

Department of Hematology, Children’s Hospital, Boston, Massachusetts 02115, and Department of Pediatrics, Harvard Medical School, and HowardHughes Medical Institute INTRODUCTION Chronic granulomatous disease (CGD)is an uncommoninherited disorder in which phagocytic cells (neutrophils, monocytes, macrophages, and eosinophils) fail to produce antimicrobial oxidants (1, 2). Affected individuals have traditionally suffered from recurrent and often lifethreatening bacterial and fungal infections. In fact, upon its recognition as a clinical entity nearly 30 years ago, the disorder was termed "fatal granulomatousdisease" (3), which attests to the severity and progressive nature of the infections often witnessed in CGD patients. Although it is a rare condition, CGDhas been the focal point for research efforts to define the biochemical events that participate in the cellular production of the major oxidant, superoxide anion. The underlying theme has been that identification of the specific defect(s) in the pathway to superoxide generation that characterize the disease would provide an understanding of the normal biochemistry of this important host defense system and, perhaps, suggest new approaches to treatment of the inherited condition or modulation of tissue oxidant damagein inflammatory states. While the physiologic deficiencies of phagocytes from CGDpatients have been recognized for nearly two decades (4), it is only of late that some components of the superoxide-generating system of phagocytes have been fractionated and, more recently, shownto be the primarily affected gene products in CGD.Althoughdefinitive biochemical and genetic solutions to some long-standing questions posed by CGDhave recently been attained, 277 07324)582/89/04104)277502.00

Annual Reviews


278 ORKIN fundamental problems remain to be addressed nowwith new tools and perspectives. In this review I summarizethe biochemistryand moleculargenetics of CGDwith particular emphasison those aspects that suggest whichavenues might be pursued in further efforts to understand howthe superoxidegenerating system of phagocytesis assembledand controlled, and howit functions. First, I review the cellular biochemistry of superoxidegeneration, its abnormalities in CGD,and the evident heterogeneity of the condition. Withthis as a background,I then reviewrecent genetic and molecularfindings that haveled to the characterization of specific components, sorted out primary defects from secondary consequences on cellular metabolism,and begunto suggest howparts of the system may be regulated. Finally, I suggest wherenewapproachesthat couple conventional biochemical maneuvers with molecular techniques maybe applied to understand further howthe superoxide generating system of phagocytesis assembledand howit functions. THE CELLULAR BIOCHEMISTRY OF THE RESPIRATORY BURST AND CGD Whenphagocytesencounter bacteria, or other appropriate particulate or soluble stimuli, they undergoa profound metabolic transformation in whichtheir consumptionof oxygenincreases rapidly; soon thereafter, large amounts of superoxide (O~) and hydrogen peroxide (H2Oz) producedand liberated into the surroundingmedium.This process, known as the "respiratory burst," functions to generate the powerfuloxidants that constitute an importantlimb of our antimicrobialhost defense(1, 2). Thegeneration of superoxidereflects the end-productof activation of an otherwise dormantmembrane-bound enzymatic system that catalyzes the one-electron reduction of oxygento O2. In CGDthe failure to produce superoxidecould, in principle, be the consequence of a defect in activation or at a subsequentstep. Theenzymethat catalyzes the electron transfer is generally referred to as the NADPH-oxidase, to reflect its apparentpreference for NADPH, over NADH, as a cofactor. A brief description of various aspects of the superoxide-generatingsystemprovides a framework in whichto consider the componentsthat mightbe defective in CGD. Activation Manydifferent stimuli can activate the respiratory burst. At least two pathwaysof activation appearto exist (5). One,whichis thoughtto involve an interaction with guaninenucleotide-binding(G-) proteins, can be trig-

Annual Reviews MOLECULARGENETICS OF CGD

279


gered by binding of the chemotactic peptide N-fmet-leu-phe to its receptor on the cell surface. Alternatively, stimulation by a soluble mediator, phorbol myristate acetate (PMA),is believed to occur via activation of protein kinase C and subsequent phosphorylation of components of the oxidase. Although considerable work now focuses on the relationship of specific G-proteins to the respiratory burst of phagocytes, at present it is not possible to relate functional abnormalities in CGDto any particular G-protein species.

Protein Phosphorylation Upon stimulation of neutrophils with PMA,some dozen or so proteins, from 11-80 kd in size, exhibit changes in phosphorylation, the majority of which show increased labeling (6-11). Establishing that altered phosphorylation of any of these proteins is intrinsically related to the activation pathwayof the superoxide-generating system has been difficult. However, as described below, several polypeptides in the 44-48 kd size class appear to be reasonable candidates for critical components.

MembraneDepolarization One of the earliest recognized events accompanyingactivation is a depolarization of the neutrophil plasma membrane(12). Largely because depolarization is not observed upon stimulation of neutrophils obtained from CGDpatients, it was hypothesized that defective activation of the superoxide-generating system, rather than a defect in the oxidase itself, might account for the physiologic abnormalities in the disease. More recent findings, however, mitigate against this conclusion. For one, the respiratory burst in normal neutrophils can occur even in the presence of agents that inhibit this membrane depolarization (13, 14). In Vitro Activation One of the more promising developments in neutrophil research has been the development of in vitro assays for activation of the NADPH oxidase (15-19). Nearly simultaneously, several laboratories independently reported that arachidonic acid and somedetergents can activate the oxidase activity in cell-free preparations. The dormant, but activatable, oxidase is located in the plasma membranefraction and requires a cytosolic componentfor activation. The nature of this cytosolic factor is poorly understood. Preliminary data suggest that it is a large protein (~ 240 kd) that may dissociate to ~40 kd subunits (17, 20). It does not appear synonymouswith protein kinase C. Induction of differentiation in cultured promyelocytic leukemia HL60cells is accompanied by the appearance of this cytosolic factor activity (21, 22). Whetherthe cytosolic cofactor is

Annual Reviews 280 ’ ORKIN single protein species or a complexof similarly sized (,,~ 40 kd) subunits currently under investigation in several laboratories.


The Respiratory

Bi~rst

Oxidase

Despite considerable efforts, the enzymatic machinery that performs the one-electron reduction of oxygen to O~ has been one of the components of the superoxide-generating system most refractory to definitive biochemical characterization. Several laboratories have attempted purification of the O7 forming enzymefrom the particulate fraction of neutrophils (1, 23, 24). In some purified preparations major polypeptides 66, 48, and 32 kd were seen under denaturing conditions (24). However, no clear-cut evidence exists, as yet, that these polypeptides are integral components of this enzyme system. Affinity-labelling of a 66-kd polypeptide by NADPH-analogsprovides some tentative supportive evidence for the potential involvementof the similarly sized chain seen in the oxidase preparations (25). Until rigorous purifications of the oxidase are reported and confirmed by several laboratories, it is not possible to discuss the biochemistry of this enzymatic complexin greater detail. Electron

Carriers

and the Neutrophil

Cytochrome

b

Like manyaspects of the cellular biochemistry of the respiratory burst, there has been considerable controversy in the literature regarding the electron carriers involved in catalysis by the oxidase. Althougha detailed discussion of the various possibilities is beyondthe scope of this review, it is sufficient to mentionthat at least three groups--a quinone (ubiquinone50) (26), a flavin (27), and a berne--have been considered as potential carriers. Evidence in favor of involvement of ubiquinone-50 generally has been considered weak (28). An FADrequirement for O7 production particulate preparations in vitro has provided strong support for flavin participation (29). Compellingdata indicates that heine, the prosthetic group for a phagocyte-specific protein, designated cytochrome b55a or b-24~, is involved in superoxide-generation. Although characterization of this protein has proved formidable and controversial, recent data have not only demonstrated its existence but established its structure and critical role in the oxidas~system. The first descriptions of the phagocyte-specific cytochrome appeared in the Japanese literature in the mid-1960s. These reports presented data on horse and rabbit neutrophils (30-32). It was not until more than 10 years later that the observations of Segal & Jones brought this unusual cytochrome to the fore (33, 34). This b-type cytochromehas been observed only in neutrophils, monocytes, macrophages, and eosinophils

Annual Reviews


MOLECULAR GENETICS

OF CGD

281

in humans. During in vitro differentiation of HL60cells and monocytic U937cells the abundance of the cytochrome (as determined by spectral measurements) increases. In normal neutrophils the cytochrome b is particularly abundant, roughly 100 pmol per mgof protein (35). The midpoint potential, the point at which it is balanced between oxidation and reduction, is the lowest for this cytochrome (-245 mV)of any knownb-type cytochrome. This property has suggested that it maybe the terminal electron donor in a short electron transfer chain and maydirectly reduce oxygen to superoxide (36, 37). Since it displays an absorption band at 558 nm, the cytochrome has alternatively been designated cytochrome b558. Hereafter in this review, we refer to this protein as the neutrophil cytochrome b. Absence of the cytochrome spectrum in the majority of patients with CGD(see below) first led Segal and his associates (35) to propose that the cytochrome was important to normal superoxide generation and that abnormalities of the cytochromewere, in fact, likely to reflect primarydefects in this disorder. Purification of the neutrophil cytochromeb has been a formidable task, accomplishedonly recently. Initial estimates of its size ranged from 11127 kd (38~t 1). The disparate sizes reported were in part the consequences of proteolysis, anomalousmigration of some proteins on acrylamide gels, and difficulties involved in distinguishing relevant polypeptides from contaminating species. Even before purification of the cytochrome was attained, Harper and colleagues recognized that the cytochrome was, in fact, a glycoprotein, on the basis of its staining with periodic acid, lectin binding, and reduction in apparent size following treatment with endoglycosidase F (42). Deglycosylated cytochrome appeared to migrate as 50-55 kd species. Within the past two years the laboratories of Segal & Jesaitis independently achieved purification of the neutrophil cytochromeb (43, 44). Surprisingly, the protein was found to be composedof two subunits, a 90kd glycosylated heavy chain (corresponding to the glycosylated cytochrome first noted by Harper et al; 42), and a 22-kd nonglycosylated chain. These subunits are tightly associated. Dissociation is achieved only upon harsh treatment (heat and SDS), conditions under which the heme spectrum is lost. Because of this close association of the componentsof the cytochrome b heterodimer, it has not been possible to determine by biochemical criteria whether the heme prosthetic group is primarily associated with one or the other subunit. Immunoprecipitation and hydrodynamicstudies have suggested that the subunits are associated as heavylight chain heterodimers (45). The structures of the cytochromeb subunits are described in greater detail below.

Annual Reviews

282 ORKIN


INHERITANCE OF CGD

AND GENETIC

HETEROGENEITY

Fromthe earliest descriptions of the disorder it wasapparent that the majorityof affected families display X-linkedtransmission(1). In carrier females, mosaicismof neutrophil function, as revealed by a histochemical stain for oxidase activity (the NBTtest) (46), can be demonstrated, sistent with the Lyonhypothesisof randominactivation of one X-chromosomein each cell (47). DNAlinkage data and the cloning of the gene involved in the X-linked form of CGD have provided definitive evidence, as described below. In the overwhelmingmajority of patients with Xlinked CGD(X-CGD)the spectrum of the neutrophil cytochrome b absent, although rare patients with partial, or even complete, spectral activity have been described. X-CGD in which the spectral activity is absent is generally designated X-, whereasthe variety with cytochrome +. spectrumis X Autosomalrecessive inheritance of CGDis nowwell established and accounts for perhaps 25-35%of affected families (48, and see 1). The occurrenceof CGD in femalesfirst led to the discoveryof this formof the disorder. Themajority of autosomallyinherited CGDis of the cytochrome b positive variety, A+ (49, 50, 51); although a less common subtype in which the cytochromespectrum is absent (A-) (18, 52) is also documented. Althoughbiochemicaland genetic criteria can nowbe used quite effectively to classify formsof CGD,as presentedin moredetail below,elegant demonstrationof distinct subtypes of the disorder wasprovided several years ago by complementationanalysis in somatic cell hybrids derived from monocytesof affected individuals (52, 53). Hybridsformedbetween X- and A+ monocytesexhibit functional complementation, as assayed by the NBTtest, whereasX cells of one patient cannot complementXcells of other individuals. Furthermore,two forms of autosomaldisease could be distinguished, as X- and A cells were also seen to complement each other. However,a more prolonged period of culture was required before functional reconstitution was apparent. The complementationof X- and A- monocyteswas interpreted as evidence that two genes, one of X-chromosome origin and the other located on an autosome,are necessary for ultimate expression of the neutrophil cytochromeb. This conclusion can nowbe refined in light of morerecent biochemicaland moleculardata, as discussed below. In addition to the broad subtypesnoted above, patients with "variant" forms of disease havebeen described. Thesepatients generally havepresented with a milder clinical course, usually evident frompresentationat

Annual Reviews


MOLECULARGENETICS OF CGD

283

a later age, and often with a NADPH oxidase that exhibits kinetic alterations in cell preparations (50, 51, 54, 55). Neutrophils of these patients have generally produced small but measurable amounts of superoxide uponactivation, in contrast to those of"classical" patients that produce no superoxide. "Variant" patients maybe either of the X-linked or autosomal variety. The existence of "variant" CGDpatients underscores the clinical heterogeneity of the disorder. Although other explanations have been entertained in the past, it is increasingly apparent that someof these "variant" patients probably have gene defects that are not as severe in a functional sense as those found in more "classical" patients. Again, additional evidence in favor of this view is presented below.

BIOCHEMICAL DEFECTS IN CGD: CANDIDATES FOR THE PRIMARY GENE PRODUCTS A myriad of biochemical abnormalities have been described in neutrophils from CGDpatients. In a complex cellular system, such as that which generates superoxide, it is difficult to establish conclusively the role of specific constituents without suitable cell-free reconstitution assays and/or information contributed by genetic approaches. The problems inherent in dissecting the componentsof the oxidase system and in proving the functional relevance of any purified (or semi-purified) species, and the disparate abnormalities reported for CGDphagocytes, conspired to render conclusions regarding any one finding controversial and often confusing. Clearly the most prominent candidate for the primary protein product affected in CGDhas been the neutrophil cytochrome b, first shown by Segal and coworkers(35, 56) to be absent at the spectral level in the vast majority of X-CGD patients. Difficulties in achieving purification of the cytochrome (38-41), questions regarding its kinetics, and the apparent absence of heme from enriched preparations of putative respiratory burst oxidase (24) contributed to doubts as to whether the spectral abnormalities in the disorder reflected a primary or secondary defect. The more recent purification of the neutrophil cytochrome b by Segal & Jesaitis and their associates provided evidence that both the 90-kd glycosylated and 22-kd nonglycosylated subunits were lacking at the protein level in X-CGDphagocytes (43, 44). While this finding may have strengthened the association of the cytochrome b with the X-CGDphenotype, quantitative deficiency of two subunits, not apparently related in a precursor-product manner, made it impossible to conclude on biochemical grounds alone where the genetic defect(s) in X-CGDresided. Although the defects might be located within the gene(s) for either (or both) subunits (the latter possibility being extremely unlikely), an equally plausible

Annual Reviews


284

ORKIN

hypothesis might involve deficiency or abnormality of an unrelated protein, whoseintegrity is necessary for proper assembly or stability of the cytochrome b heterodimer. The description of other biochemical defects in CGDphagocytes, either of the X-linked or autosomal variety, presented alternative candidate molecules in virtually all disease classes. Since activation of the dormant NADPH oxidase is a critical step in superoxide generation, manyinvestigators have focused on the possible role of protein phosphorylation in this process and potential disturbances in CGD.Particular attention has been paid to a cluster ofpolypeptides, 44~48kd in size, that becomerapidly phosphorylated upon stimulation of neutrophils with phorbol esters or during activation in cell-free preparations. Segal and associates first reported failure of phosphorylation of a 4448 kd protein in autosomal recessive, but not X-linked, CGDpatients (11, 57, 58). Others, using two-dimensional rather than single dimension acrylamide gel analysis, have described changes in both X- and autosomal CGDneutrophils, but more limited alterations in the X-linked variety (7). Finally, others have failed to observe alterations in either disease class (59). Nonetheless, the emergingconsensus is that changes in protein phosphorylation accompanyactivation and that alterations in such events do occur in CGDneutrophils. However, again, it has not been established whether the observed phosphorylation disturbances in CGDreflect primary defects or secondary consequences of other underlying quantitative or qualitative protein abnormalities. In the case of X-CGD,based on the evidence to be reviewed below, we must conclude that any alterations in phosphorylation of 44-48 kd polypeptide are secondary phenomena. In light of this, we must reserve judgment as to whether the apparent failure of phosphorylation of a 44-48 kd polypeptide in autosomal CGD reflects deficiency or a structural defect of the substrate (the 44 kd polypeptide) or merely indicates a more proximal defect that ultimately results in a change in phosphorylation. Although available data suggest that protein kinase C itself is not abnormalin autosomal CGD,this conclusion also must be considered tentative in light of the increasing complexity of protein kinase C isoforms that have been described (60). Howthese 4448 kd polypeptides are related (if at all) to the cytosolic factor activity defined in cell-free, reconstitution assays is uncertain. Reconstitution assays using purified 44-48 kd protein species have not yet been reported. These could serve to more strongly associate specific protein deficiencies and functional deficits. Recently a protein species of apparent size 40-47 kd that is rapidly phosphorylated by protein kinase C in activated platelets h/is been characterized by cDNAcloning (61). As this protein is also expressed

Annual Reviews



285

promyelocytic HL60cells that are induced to differentiate with retinoic acid, we maywonderwhether it, or a related species, is in actuality one of the neutrophil componentsthat,iszrapidly phosphorylated upon activation. A direct analysis of these issues is nowpossible. Fromthe different isoelectric points observed for the platelet and neutrophil species, however, it is likely that they are not closely related (J. Curnutte, personal communication). Given its postulated role in superoxide-gene~ation,the respiratory burst oxidase (or flavoprotein) has also been considered as the primarily affected product in CGD(62). However, in view of the incomplete biochemical characterization of this enzyme(or enzymecomplex), no investigations protein abnormalities in CGDneutrophils have been reported. In sum, from biochemical data alone the primary protein defects in CGDhave been difficult to assign with confidence. In the X-linked variety of disease, the neutrophil cytochromeb, as proposed by Segal and associates (35), constituted an excellent candidate, although recognition that two subunits associate to form the cytochromedid not immediatelyclarify the situation from a genetic perspective. In the major autosomal form of the disease, the 44-48 kd phosphorylated polypeptides have emerged as candidates for the primarily affected proteins; however, the complexity of the system makesalternative models~plausible.

MOLECULAR APPROACH TO THE X-CGD LOCUS In light of the problems inherent in defining the .critical proteins of a complex cellular system without availability of each purified component and a suitable in vitro reconstitution assay, a molecular genetic approach to unraveling the basis of CGDin its various forms has fundamental advantages. While such an approach has led to specific, definitive conclusions regarding some of the polypeptides described above in their relation to CGD,it has also provided reagents that mayprove useful in addressing someof the manyquestions nowevident in the normal cellular biochemistry of the superoxide-generating system. Chromosomal Location of the Gene for X-CGD As with manyfindings relating to the primary ~basis, of CGD,the initial assignment of the locus for X-CGDto the distal portion of the short arm of the X-chromosomefrom limited family segregation data has been revised with the introduction of newtechniques for analysis. Studies in the pre-DNAlinkage era reported evidence for linkage of CGDto the blood group antigen Xg, a distal Xp marker (situated in Xp22.3-Xpter) (63). addition, the reported association of CGDwith the absence of Kell-related

Annual Reviews


286 ORKIN antigen Kxas well as its linkage to Xg(64, 65) seemedto strengthenthis assignment.Absence of the Kxantigen on red cells results in acanthocytosis and a mild hemolytic anemia, designated the McLeodphenotype. Althoughinitial reports of CGD frequently described association with the McLcodphenotype (63, 66), possibly suggesting a commonmembrane defectin red andwhitebloodcells, it is nowevidentthat only rare patients haveboth phenotypes(67) [a point discussedbelow]. By 1984 it had becomeapparent that the assignment of the X-CGD locus to distal Xpwasinconsistent with cytogeneticdata derived at first from two rare cases studied by Franckeand her associates. In the first instance, a womanwith mild mental retardation and heterozygosity for CGDwas found to be heterozygous for a deletion within chromosomal bandXp21(68). Asecond patient [B.B.] whowasafflicted simultaneously with Duchenne muscular dystrophy (DMD),CGD,McLeodsyndrome, and retinitis pigmentosahad an interstitial deletion of Xp21involving perhaps 5 million basepairs of DNA(3-5 megabases)(69-71). Both these cases, taken together with another patient [N.F.), whohad DMD, McLeod syndrome,and CGDin association with an interstitial deletion of Xp21(72), suggested a more proximal location of X-CGD gene than that suggestedby linkage to Xg. In as muchas complex,but unseen, cytogenetic rearrangementsmight be invokedto explain the apparent Xp21location of the gene in the face of earlier evidencefor linkage to Xg, a formalgeneticlinkage analysis was performedin typical X-CGD families (72), using a collection of cloned probes that recognizerestriction fragmentlength polymorphisms (RFLPs) along Xp. DNA probes within Xp21(designated p754 and PERT84), but not probes moredistal or Xgitself, were found to be linked to the CGD locus (72). Thecalculated Lodscore for linkage to p754was 3.7, which indicates nearly a 10,000: 1 chancethat the X-CGD locus is nearbyrather than elsewherein the genome.In addition, these linkage data, combined with DNAmappingdata of Kunkel & Monacoobtained in DMD patients (73), indicated that the X-CGD locus resided proximal(or centromeric) to the DMD gene (72). Onthe basis of cytogenetic and DNA linkage data therefore, the X-CGD locus was reassigned to Xp21. In addition, no evidence has been obtained that suggests the existence of more than a single locus on the X-chromosome. Identification of the Gene on the X-Chromosome That Is Defective in CGD The assignment of the X-CGD gene to Xp21 within the BBdeletion narroweddownthe area in which the gene resided to about 0.1%of the human genome, roughly 5 megabases, at a minimum.Because major

Annual Reviews



287

questions existed regarding the nature of the protein primarily affected in X-CGDand because adequate biochemical reagents were lacking for any of the candidate proteins, a direct approach to identification of an RNA transcript or DNAsegment of the relevant gene was attempted (74). pool or radioactive cDNA,derived from reverse transcription of mRNA from differentiated HL60cells, was enriched for granulocytic sequences by hybridization subtraction with RNAfrom the B-cell line of a patient NF with an Xp21 deletion. The enriched cDNAwas used directly as a probe in Southern blot hybridization to screen a collection of genomic bacteriophage clones derived from Xp21 which had been isolated by Kunkel and coworkers (75, 76) in their search for the DMDlocus. Two independently isolated, but overlapping, bacteriophage clones reacted with the pooled cDNAprobe and thereby identified a region of Xp21 transcribed in differentiated HL60cells (74). With the portion of the bacteriophage clones that hybridized as a probe, the specific mRNA present in these phagocytecells was identified. The RNAtranscript detected in Northern blot analysis is 4.5 kb in length and is expressed specifically in differentiated HL60cells, normal neutrophils, and monocyte/macrophages, and at a lower level in EBVtransformed B lymphocytes(74). Undifferentiated HL60cells contain little of this mRNA.Cells of nonphagocytic lineages are devoid of the mRNA. Strong evidence in support of the relevance of this RNAtranscript to XCGDwas provided by Northern blot analysis of monocyte RNAisolated from patients with X-CGD. In the first four classical (i.e. cytochromebnegative) X-CGD patients so examined, three revealed markedly deficient mRNA(74). Formal genetic criteria more firmly established the RNAtranscript as that derived from the X-CGDlocus. In the only mRNA-positivepatient amongthe first four studied, a relatively small interstitial deletion (74) (of about 1 kb) removed the coding capacity for the C-terminal 41-amino acids of the protein predicted from the cDNAsequence (S. H. Orkin, unpublished data). That this deletion was entirely contained within the transcribed region of the gene and removed a segment of the predicted protein constitutes a genetic proof that the putative X-CGD transcript is from the relevant locus. Additional biochemical evidence noted below provides confirmation of this conclusion. The Predicted

X-CGD Protein

The complete X-CGDcDNA is 4.27 kb in length and contains a single open reading frame encoding 571 amino acids (including the initiation methionine) (calculated 65-kd polypeptide), which is displayed in Figure 1. The initiator codon in the first published report was


Annual Reviews 288 ORKIN

Annual Reviews



289

not correctly identified due to a DNAsequencing error in the 5’-untranslated region (77). Of particular note and unknown significance, if any, is the short length of the 5’-untranslated region of the X-CGDmRNA (12 nucleotides). What conclusions could be drawn from the predicted sequence of the X-CGD protein? Unfortunately, surprisingly few. As shown in Figure 2, the open-reading frame predicts five potential N-glycosylation sites and at least three substantial hydrophobic domains. Together, these features suggested that the encoded protein was a membraneglycoprotein. Searches of the GenBankand Protein Database failed to reveal significant similarity to other knownproteins or genes. In particular, it has not been possible to discern any similarity to knowncytochrome species or to identify a potential heme-binding motif. On the basis of the computer searches it could only be speculated initially that the X-CGDgene encoded a membrane glycoprotein that might be involved in the assembly of the cytochromeb but was unlikely to be the cytochromeitself.

THE PRODUCT OF THE X-CGD LOCUS AND THE NEUTROPHIL CYTOCHROME B: RELATIONSHIP, STRUCTURE AND EXPRESSION Identification of the X-CGDProtein as the 90-kd Subunit of the Cytochromeb Twoindependent approaches, however, established that the predicted XCGDprotein is synonymouswith the larger subunit of the cytochrome b. First, antisera raised either to a synthetic peptide directed to 20 amino acids near the C-terminusof the predicted protein or to a fl-galactosidase fusion protein reacted with a 90-kd glycoprotein in normal neutrophils, as well as the larger subunit of purified cytochromeb (77). The antisera also reacted with the deglycosylated (roughly 55-kd) form of the large subunit.

5’

N N-glgcosyletion

Sites

Hgdrophobic Domains ¯

ORF e ee e

~

C e

( PolyT-TTTATT

¯

0.5 kb 2 Structureof X-COD cDNA and predictedprotein. Theopenreadingframe(ORF) is shown bythe hatchedbox. FixTure

Annual Reviews

290 ORKIN Second,directly determinedN-terminalaminoacid sequencesof the 90kd chain matchedthe sequence predicted from the corrected cDNA sequence(78). In the vast majorityof X-CGD patients studiedto date, the 90-kdsubunitis virtually absentin Western blot analysis of neutrophils (77).


Structure of the 22-kd Subunit of the Neutrophil Cytochrome b The 90-kd X-chromosome encodedsubunit is tightly associated with a nonglycosylated 22-kdpolypeptide(43, 44). Theprimarystructureof the 22-kdchain has morerecently been determinedby cloning of its cDNA via immunoscreening of a 2gt 11 library (79). DirectN-terminalaminoacid sequencingconfirmedthe authenticityof the cDNA.Thepredictedprimary translationproduct,depictedin Figure3, contains195 aminoacids; the

GCAGTGTCCCAGCCGGGTTCGTGTCGCCATGGGGCAGATCGAGTGGGCCATGTGGGCC~ ~etGlyGlnlleGluTrpAl~e~TrpAl~As CGAGCAGGCGCTGGCGTCCGGCCTGATCCTCATCACCGGGGGCATCGTGGCCACAGCTGG nGluGlnAlaLeuAl~SerGlyLeuIleLeuIleThrGlyGlyIleV~iAl~ThrAlaGl GCGCTTCACCCAGTGGTACTTTGGTGCCTACTCCATTGTGGCGGGCGTGTTTGTGTGCCT yArgPheThrGlnTrpTyrPheGlyAlaTyrSerIleValAlaGlyValPheValCysLe GCTGGAGTACCCCCGGGGGAAGAGGAAGAAGGGCTCCACCATGGAGCGCTGGGGACAGAA uLeuGluTyrProArgGlyLysArgLysLysGlySerThr~etGluAr~TrpGlyGlnLy GCACATGACCGCCGTGGTGAAGCTGTTCGGGCCCTTTACCAGGAATTACTATGTTCGGGC sHisMetThrAlaVaiValLysLeuPheGlyProPheThrArgAsnTyrTyrValArgA1 CGTCCTGCATCTCCTGCTCTCGGTGCCCGCCG~CTTCCTGCTGGCCACCATCCTT~GGAC aValLeuHisLeuLeuLeuSerValProAlaGlyPheLeuLeuAlaThrlleLeuGlyTh CGCCTGCCTGGCCATTGCGAGCGGCATCTACCTACTGGCGGCTGTGCGTGGCGAGCAGTG rAl~CysLeuAl~IleAl~SerGlyIleTyrLeuLeuAl~Al~V~lArgGlyGluGlnTr GACGCCCATCGAGCCCAAGCCCCGGGAGCGGCCGCAGATCGGA~GCACCATCAAGCAGCC pThrProlleGluProLysProArgGluArgProGlnlleGlyGlyThrlleLysGlnPr GCCCAGCAACCCCCCGCCGCGGCCCCCGGCCGAGGCCCGCAAGAAGCCCAGCGAGGAGGA oProSerAs~ProProProArgProProAl~GluAl~ArgLysLysProSerGluGluG1 GGCTGCGGCGGCGGCGGGGGGACCCCCGGGAGGTCCCCAGGTCAACCCCATCCCGGTGAC uAlaAl~Al~AlaAlaGlyGlyProProGlyGlyProGlnValAsnProIleProYalTh CGACGAGGTCGTGTGACCTCGCCCCGGACCTGCCCTCCCACCAGGTGCACCCACCTGCAA rAspGluValValEnd TAAACGCAGCGAAGGCCGGGAAAAAAA Figure3Deducedproteinsequenceofthecytochrome b 22 kd subunit

Annual Reviews


MOLECULAR GENETICSOF CGD

291

sequence is particularly notable for a proline-rich C-terminus (27%of the 63 C-terminal residues). Three or four hydrophobic regions appear to precede the proline-rich tail. Computeranalysis has revealed no overt similarities to other proteins, including a variety of cytochromespecies. Althougha short stretch of similarity (29% identity over a 31-amino acid region) to polypeptide I of mitochrondrial cytochrome c oxidase was identified, its significance is unknown. Therefore, the phagocyte cytochrome b appears to be a unique cytochromespecies in that it is not overtly related to previously studied cytochromes. In light of this circumstance, the location of the hemeprosthetic group(s) in the cytochrome b heterodimer remains unknown. Howthe neutrophil cytochrome b is positioned in the plasma membrane and specific granules is currently poorly understood. On the basis of a cluster of potential N-glycosylation sites near its N-terminus, it has been proposed that an external domainof the 90-kd glycoprotein is likely to exist (80). Consistent with this notion, a monoclonalantibody to the 90kd subunit reacts with the external surface of neutrophils (81). In addition, endoglycosidase F cleaves N-linked carbohydrates from the cytochrome at an extracellular site (82). Preliminary data also suggest that the extreme C-terminus of the 90-kd subunit may lie on the cytoplasmic face of the plasma membrane(82). In as muchas deletion of the C-terminal 41-amino acids of the 90-kd polypeptide in an X-CGD patient leads to characteristic findings of the disease (74) [i.e. absence of both protein subunits], it likely that the C-terminus plays an important role in either cytochromeb assembly or interaction with the 22-kd subunit. Preliminary cell-free assay findings have also been interpreted as consistent with an important functional role for the C-terminusof the 90-kd protein (82). Reyulation Subunits."

of mRNAs Encodin# the Cytochrome Implications.for Assembly

b

As noted above, the mRNA transcript for the 90-kd subunit is expressed in a highly lineage-specific manner (74). In neutrophils, monocyte/ macrophages, and presumably eosinophils, the mRNAis particularly abundant, perhaps accounting for 0.1% or more of cellular mRNA. In addition, the mRNA is inducible during in vitro differentiation of cultured HL60cells along either a granulocytic or monocytic pathway. A low level of mRNA is detectable in normal EBV-transformed B-cells, which are reported to have low levels of NADPH-oxidase activity (83, 84). Although the pattern of expression of the 90-kd subunit mRNA mirrors the distribution of neutrophil cytochromeb and oxidase activity in various cell types, that of the 22-kd subunit does not. Surprisingly, the 22-kd subunit mRNA, which is derived from an autosomal locus (M. C. Dinauer,

Annual Reviews


292

ORKIN

S. H. Orkin, unpublisheddata), is present in a wide variety of cell lines in abundance comparable to that in phagocytic cells. However, although the mRNA is constitutively expressed, non-phagocytic cells contain very little if any stable 22-kd light chain polypeptide. These observations seem most compatible with a model in which the 22-kd subunit may be unstable intracellularly in the absence of the 90-kd glycosylated chain. Overall, the appearance of the cytochromeb heterodimer is temporally correlated with expression of the 90-kd subunit mRNA.Although the rationale for this regulatory scheme, rather than coordinated expression of the individual subunits, is obscure, it implies that the 90-kd glycoprotein may serve a critical role in directing the 22-kd polypeptide into the functional membrane complex (79). Since neither cytochrome b subunit has a cleavable N-terminal signal peptide, targeting to the membranecompartment must involve internal signal sequences. A likely possibility is that the 22-kd subunit is anchored into the membranecomplex via its tight interaction with the 90-kd chain. Modulation of Expression mRNA by Interferon-~

of the 90-kd Subunit

Macrophage functions can be augmented by treatment with the lymphokine interferon-~, the major constituent of macrophage-activating factor (85). Specifically, treated macrophagesacquire the capacity to generate ir~creased amountsof superoxide and to kill microbes more efficiently (8689). Biochemicalcorrelates of this phenomenonof macrophageactivation are largely unknown. Recently it has been observed that recombinant interferon-~ induces the level of 90-kd subunit mRNA in normal cultured monocytes, neutrophils, and promonocytic leukemic cell lines, such as THP-1and U937 (90). In cultured promonocytie THP-1cells the level induction is approximately 10-fold. The abundance of mRNA encoding the 22-kd subunit is virtually unchangedupon interferon-~ treatment; this is consistent with its independent regulation. Nuclear run-on eperiments suggest that transcriptional regulation plays a major role in the induction of the 90-kd subunit mRNA (90). Although interferon-7 undoubtedly has diverse effects on macrophageconstituents, it is reasonable to infer that induction of the 90-kd subunit mRNA may contribute to increased availability of functional cytochromeb and ultimately to augmentedsuperoxide production, mRNA in the 90-kd subunit is the first defined componentof the superoxide-generating system regulated in a physiologically appropriate mannerduring macrophageactivation induced by interferon-~,. Recently Cassatella et al (91) have confirmed the induction of 90-kd subunit mRNA by interferon-7 in cultured HL60, U937, and ML3cells, and they have shown that the induction is insensitive to inhibition of

Annual Reviews MOLECULAR GENETICS

OF CGD

293

protein synthesis. Therefore, it is likely that modification of pre-existing nuclear factors mediate the transcriptional increase in 90-kd subunit mRNA expression. Tumornecrosis factor (TNF) also induces an increase in specific mRNA in cultured HL60,U937, and ML3cells (91). In association with increases in 90-kd subunit mRNA upon interferon- 7 or TNF treatment, the amount of spectrally determined cytochrome b increased (91, 92).


Structure of the X-CGDGene Although the detailed physical organization of the X-CGDgene has not been reported, several general features can be briefly summarized. The transcribed region (i.e. that represented by the cDNA)encompasses approximately 25-30 kb of genomic DNAisolated in several overlapping bacteriophage clones. The last 41 amino acids of the encoded protein and the 2.5 kb T-untranslated region of the mRNA are included in the last exon. The remaining 12 exons are more conventional in size. As noted above, the 5’-untranslated region is unusually short. The cis-acting DNA sequencesthat mediate lineage-specific expression of the gene or its responsiveness to interferon- 7 have yet to be identified. Consensus sequences for interferon-responsive elements (93-96) are not evident in genomic sequences (S. Orkin, unpublished data). No restriction fragment length polymorphismsof the gene or its immediate flanking regions have been identified in a search with more than 40 different enzymesusing several genomic and cDNAfragments as probes (S. C. Goff, S. H. Orkin, unpublished data). Therefore, the locus appears to be highly conserved.

THE MOLECULAR BASIS

OF CGD

On the basis of the data reviewed herein it is nowpossible to begin to discuss the various forms of CGDin molecular genetic terms.

X-CGD(X-, X- Variants,

and +)

The commonest X-linked form of the disorder, designated X- [for Xlinked, cyt b-negative], is caused by mutations in the gene encoding the 90-kd cytochrome b subunit. At present, there is no evidence from DNA linkage data or from analysis of patient material with the cloned cDNA that morethan a simple Xplocus, that within Xp21,exists. Th¢,,d~fe.c, ts ~in the gene are heterogeneous. Deletion of the entire gene has,been, observed in rare patients with complex phenotypes, such as BB .and NF who~had CGD, DMD,McLeodsyndrome, with or without retinitis pigmentosa, and others that presented with CGDin association with McLeodphenotype alone (see below) (69, 72, 74). ,Partial :gene deletion that predicts

Annual Reviews


294

ORKIN

synthesis of a truncated protein has also been described (74). In the majority of X-CGD patients, however, the gene encoding the 90-kd subunit appears grossly normal by conventional Southern blot analysis (74). these instances, mRNA is often, but not always (97), markedly deficient in phagocytes, presumably due to mutations that alter RNAtranscription, mRNA processing, and/or mRNA stability. In the vast majority of patients both cytochromeb protein subunits are absent in phagocytes (43, 44, 79). Variant X-CGDpatients have been described who have either residual detectable superoxide-generation (so-called variant X- disease) or linked disease in which the cytochrome spectrum is present at normal or (X+) (50, 55, 98, 99). Althoughmore studies need to be executed, patients of the former subgroup often appear to have exceedingly low, but detectable, levels of 90-kd subunit mRNA (100). These patients are formally analogous to patients with hemoglobinopathies termed/~+-thalassemia in which a reduced amount of normal/%globin mRNAis produced (101). Individuals with the X+ phenotype will most likely be found to have single amino acid substitutions in the cytochrome b heavy chain that interfere with its electron transfer function, but not with heterodimer assembly, homebinding, or stability. Additional investigation is needed, however,to provide direct evidence for these suppositions. A--CGD Although they are very rare, patients with autosomally inherited cytochrome b-negative CGD(A) have been identified (52, 102). By Western blot X -patients analysis, these patients have a protein phenotype that is indistinguishable from the X--patients: both the 90-kd and 22-kd subunits are lacking in phagocytes (M. C. Dinauer, J. Curnutte, S. H. Orkin, unpublished data). By virtue of its autosomal location and the intimate association of the two subunits of the cytochromeitself, the gene encoding the 22-kd subunit is the likely site of mutations leading to A--CGD. Preliminary data (M. C. Dinauer, S. H. Orkin, unpublished) are consistent with this hypothesis, in that we have studied one A--CGDpatient whose monocytes lack 22-kd mRNA,but not mRNA for the 90-kd chain. Alternatively, A -CGDcould result from mutations in regulatory proteins that primarily affect expression of the 90-kd subunit gene in trans. In such instances, one would anticipate normal levels of 22-kd mRNA and deficiency of 90-kd mRNA in the face of autosomally inherited disease. Currently, the molecular basis of autosomal cytochrome b-- positive disease (A+) is not formally established. Polypeptides of 44-48 kd that are rapidly phosphorylated upon neutrophil activation are potential candidates for the proteins primarily affected. Evidence is already emerging

Annual Reviews MOLECULAR GENETICS OF CGD

295


that autosomal, cytochrome b-positive CGDmaybe a heterogeneous entity. Specifically, cell-free complementation of cytosolicfractions of two different A÷-CGD patients has been reported in preliminary form (103). Purification and cloningofcDNAs for the 44-48kd species will be required to characterizethe relevant proteins. X-CGD in Association with McLeod Syndrome The McLeodSyndromeis rarely seen in association with Duchennemuscular dystrophy (DMD)and/or CGD.Whensuch complex phenotypes are evidentin affectedindividuals,it is inferredthat deletionof the relevant Xp21loci is the molecularbasis. Withthe development of specific probes for the DMD and X-CGD loci and also for anonymoussequences distributed throughout other regions of Xp21, DNAdeletions have been readily established by Southernblot analysis (69, 73, 74). Examination a series of McLeodpatients (with and without DMD or CGD)with specific DNA probes has led to construction of a deletion mapof Xp21 (displayedin Figure4) in whichthe McLeod locus is deducedto lie between the DMD and CGDloci, but within approximately 500 kb of the latter (73). The existence of McLeodpatients whodo not have DMD or CGD or evident DNA deletions demonstratesthat a distinct McLeod locus must be present in Xp21. In two patients with X-CGD and McLeodsyndrome the CGDlocus is completelydeleted (73, 104). It is inferred that the McLeodlocus is also removedby a chromosomal deletion in these individuals. Theprecise nature of the McLeod locus protein product is uncertain, althougha 37-kdprotein (the Kxantigen) is a candidate(73). The molecular bases of various forms of CGDare summarizedin Table i. OTC m

COD l’Ic Leod mmmm~m

OTC

CGD

CX5.7

DI"ID III 754 p84

p87

GENE LOCI L1

cen BBdeletion

I

I

I

OHdeletion (CGD,HcLeod)

Figure 4 Organization of the Xp21region and the relative position of gene loci. The centrometic(cen) andtelomeric(tel) directions are indicated. Thetop line and solid boxes indicate genetic loci. OTC= ornithine transcarbamylaselocus. In the middle, probesfor specific loci and anonymous sequencesare shownwith downward arrows. Thepositions of deletions in two patients BB(Ref. 69) andOM(Ref. 73, 104) are depicted.

Annual Reviews


Clinical genetic subtype

Cytochromeb (spectrumor protein)

X : classical "’variant"

Absent Very reduced

+ X

Normalor reduced

A

Absentor greatly reduced

÷ A

Normal

X-CGD in association Absent with McLeod syndrome ( + DMD)

Primarygene defect 90 kd cyt b subunit 90 kd cyt b subunit (heterogenous:transcription, RNAprocessing, RNA translatability) presumed90 kd cyt b subunit (presumedmissense) probably22 kd cyt b subunit (likely to be heterogenous:? possible existenceof regulatorydefects in trans to genefor 90 kd subunit) ?? 44M8kd polypeptide(s) cytosolicfactor(s) deletion of 90 kd cyt b subunit gene

CLINICAL IMPLICATIONS OF IDENTIFICATION OF THE X-CGD GENE Diagnosis The reduction of NBT(46) by neutrophils or related assays of superoxide production are generally sufficient to diagnose CGDand the carrier state for X-CGDin females. As with many other genetic disorders, availability of molecular reagents could greatly facilitate prenatal diagnosis of XCGD.This would obviate the use of fetal blood sampling (105) for prenatal diagnosis. Either RFLPs or detection of specific gene defects in fetal DNA obtained by chorion villus biopsy or animocentesis could provide the means to a definitive diagnosis for families at risk (106). Where a complete or partial gene deletion underlies X-CGD, prenatal diagnosis can be applied quite simply using Southern blot analysis alone. In the family at risk for CGDdue to the interstitial deletion that removed the terminal 41amino acids of the 90-kd subunit, DNAanalysis has been employed to demonstrate that a subsequent male fetus was unaffected (S. Orkin, J. Curnutte, unpublished data). Unfortunately, however, most families at risk for X-CGDdo not have DNAabnormalities that are detectable in this manner (74). In addition, RFLPs within or flanking the X-CGDgene have not been identified (S. C. Gott, S. H. Orkin, unpublished data). In that X-CGDis essentially an

Annual Reviews MOLECULARGENETICS OF CGD

297


X-linked lethal disorder in genetic terms (until recently whentreatment has improved), one would expect extensive heterogeneity in the specific mutations leading to gene dysfunction, in part based on the hypothesis of Haldane. With new methods for detection of point mutations in genes, coupled with convenient techniques for in vitro amplification of specific gene segments(107), it is reasonable to anticipate that prenatal diagnosis may be accomplished in some instances by identification of a specific family-based mutation. Treatment

of

CGD

The use of chronic antibiotic prophylaxis in the past decade has led to a marked decrease in the incidence of life-threatening infections seen in patients with CGD(108). An improved understanding of the critical components of the NADPH-oxidasesystem and development of molecular reagents raise newpossible strategies for consideration. PHARMACOLOGIC MANAGEMENT-INTERFERON-)Y The recognition that the lymphokine interferon-~ augments superoxide production from phagocytic (25, 109, 110) cells and induces mRNA encoding the 90-kd cytochrome b subunit has led to an evaluation of this agent in various forms of CGD (90, 91). Initially, phagocytic cells harvested from patients were examined for their response to interferon- 7 in vitro (100). In several patients with classical X-CGD,characterized by absent cytochrome b spectrum and superoxide production, no in vitro improvement in NBTreduction in either neutrophils or macrophageswas noted. In patients with "variant" X-CGD,in which baseline superoxide production was measurable (1-10% or normal), interferon-~ treatment has led to relatively consistent increases in NBTreduction and superoxide production. The results obtained with phagocytes of two brothers were particularly informative, as treatment was associated with a restoration of superoxide production to 30-50%of normal (100). Accompanyingthis improved function was induction mRNA for the 90-kd subunit from nearly undetectable levels to about 35% of normal. In light of promising in vitro findings preliminary studies have nowbeen performed in CGDpatients on the effects of interferon-~ administered in vivo. In a Boston study (111) the "variant" X-CGDpatients (who responded in vitro) and a classical X-CGDpatient (who did not) were given subcutaneous injections of interferon-~ at 0.1 mg/mgon two consecutivc days. Dramatic responses were noted in the two brothers with "variant" disease who displayed a partial correction in vitro. In vivo treatment led to restoration of superoxide production to normal levels within’3-5 days. Surprisingly, normal function persisted for nearly one

Annual Reviews


298 monthdespite subcutaneousinjection of interferon-~ limited to the first two days of the trial. In associationwith increasedsuperoxideproduction, in vitro bacterial killing of Staph. aureusreturned towardnormal,cytochromeb spectral measurementsimprovedfrom less than 10%to approximately 20-30%of normal, and the 90-kd subunit (as detected by Western blot analysis) increased in abundance,while still remainingonly a small fraction of normal. Improvements in superoxideproductonwere seen after in vivo treatment of another patient with "variant" disease, and also with a "classical" XCGD patient whohad failed to display an in vitro response(111). Although the augmentationin superoxide production evident following treatment of these patients was muchmore modestthan that seen in the mildly affected brothers, detectable superoxideproduction, cytochrome b spectra, and improvedbacterial killing were demonstrated.Giventhese findings wecan expect that phagocytes of these patients in vivo wouldbe more efficient at disposingof microorganisms than in their baseline state. Preliminary findings from Sechler and colleagues (112) surveying a larger numberof patients, including both X-CGD and A+ subtypes, generally confirmedthe responsiveness of somepatients to interferon- 7 administration. Of particular interest is that, whereasthe phagocytesof only 2520%of X-CGD patients produced more superoxide following lymphokine treatmentin vitro, cells of nearly all A+ patients appearedto respond. Fromthe early clinical experience with interferon-7 in CGD,several points shouldbe addressed.First and foremost,the agent appears to offer promiseas a pharmacolc/gicagent in the management of selected patients with the disorder. Considerableexperience with interferon-~ in cancer chemotherapy trials suggests that it is well tolerated in the dosagesused in the CGD trials. Second,the heterogeneityin response that is already evidentin the limited trials is likely to reflect underlyinggeneticheterogeneity (111, 112). Specifically, amongX-CGD patients we might expect those whohave some residual cytochromeb synthesis in their baseline state (although vastly reduced from normal) to benefit most, particularly if underlying mutations merely impededquantitatively the production of otherwise normal gene product, such as in the case of fl+-thalassemia. Amongpatients with molecular defects that prevent expressionof any cytochromeb (such as partial gene deletions or translation termination mutations), no response might be anticipated. Third, the augmentationof superoxide production in responders has exceeded the increase in cytochromeb measuredspectrally or by Western blot analysis (111), suggesting that very small amountsof cytochromeb are required to permitsubstantial oxidasefunction. This is formallyanalogous to other enzymesystems wherea residual level of activity maylead to

Annual Reviews



299

appreciable conversion of substrate to product. Wemayinfer, then, that the neutrophil cytochrome b is functionally rate-limiting in X-CGD. Fourth, although interferon-~ can induce cytochrome b heavy-chain expression (90, 91), it is quite likely that other important components the oxidase system maybe induced or altered in such a way as to function moreefficiently (11 l, 112). In this regard it is noteworthythat cells of +o CGDpatients, in the experience of Sechler and coworkers (112), have rather consistently responded to interferon-y in vitro. This observation suggests that the gene product deficient in this subtype of CGDmayalso be induced by interferon treatment. Finally, the potential contribution of nonoxidative mechanismsof microbial killing induced by interferon-y may be substantial and mayeventually be found to play a role in any clinical efficacy established for interferon-~ in CGD(113). To imply at this stage that the improved function of phagocytes of treated patients (apart from improved superoxide production) is due solely to induction of cytochrome b expression is not warranted. Genetic management--somatic 9ene therapy Since X-CGDis a disorder of marrow-derivedcells that is nowknownto be due to genetic defects in the gene encoding the 90-kd subunit of the neutrophil cytochrome b, transfer of an expressible copy of its mRNA into pluripotent hematopoietic stem cells, in principle, wouldconstitute definitive therapy (114). Because bone marrow transplantation where applied to X-CGDhas resulted in functional correction (115), somatic genetic therapy has a strong rational basis. In view of the substantial restoration of superoxide production associated with only miniscule amounts of cytochrome b protein, correction of the disease by somatic gene transfer would not necessitate complete correction of protein levels. An approach under investigation is the use of recombinant retroviruses (114) to transfer expressible cDNA for the 90-kd subunit into hematopoietic stem cells and their progenitors. Althoughthis approach has considerable promise, unexpected difficulties encountered in the expression of sequences transferred into primary hematopoietic stem cells and in their efficient infection, as well as the lack of immortalized,deficient phagocyticcells to test for functional reconstitution in culture, makethe practical application of this technology uncertain for the forseeable future (114). The availability of molecular reagents for the X-CGD locus, combined with site-specific gene disruption techniques applicable in murine embryonicstem cells (116-118), mayultimately lead to production of a mouse model for X-CGD.Such an experimental animal wouldgreatly facilitate studies of reconstitution by gene transfer and also provide cultured lines for in vitro cell biologic investigation of cytochrome b assembly and targeting.

Annual Reviews 300

ORKIN


NEW DIRECTIONS AND REMAININGQUESTIONS While it has provided clearcut resolution of several long-standing issues in the field of phagocyticcell biology and suggested newpotential strategies for treatment of CGD,research in the past two years has raised a wealth of questions. AlthoughCGDin its various forms is rare, lessons learned about critical proteins in the superoxide-generating system will undoubtedly lead to general insights into the normal biology of phagocytic cells and their role in host defense. Whathave we learned so far and where are the new challenges? First, biochemical and genetic studies have converged to provide compelling evidence for a critical role for the neutrophil cytochrome b in the pathway of superoxide generation. Although its role may have been considered controversial by someinvestigators until quite recently, little evidence can be mustered now to exclude its participation in superoxide production. This important conclusion supports the insightful research of Segal and his colleagues (35). Nevertheless, even though theprimary amino acid sequences of the two subunits of the cytochrome are now complete from cloned cDNAs(74, 77, 79), manyaspects pertinent to its structure and regulation remain unresolved. The topologic organization of the subunits, the domainsinvolved in their interaction, the identity of the axial ligands of the prosthetic group(s) in the heterodimer, and the structural basis for participation of the cytochrome b in production of superoxide are unknown. Second, the responsiveness of expression of the 90 kD subunit mRNA to differentiation and to interferon-v treatment suggests that it plays an important role in synthesis of the functional membranecomplex and is a critical determinant in different states of macrophageactivation. The manner in which the assembly of the cytochromeb is regulated, a process that is likely to be important in establishing the functional membranecomplex, needs direct analysis. Combinedbiochemical and genetic approaches, perhaps using gene transfer of the subunit cDNAsinto heterologous cells, will be required to address these issues. A better understanding of the organization and modulation of the oxidase may suggest new types of anti-inflammatory agents that may modify damagingeffects of superoxide generation, yet preserve antimicrobial activity. Third, the availability of molecular reagents for both cytochrome b subunits nowpermits direct assessment of the mutations that underlie the X-, X+, and A- forms of CGD. Studies of X-CGDshould clarify the relationship between "variant" disease and residual cytochrome b production and between interferon-v responsiveness and genetic pathology. Identification of presumedpoint mutations that lead,,tcr:X%disease

Annual Reviews



301

mayindicate specific aminoacids that play a role in the electron transfer function of the cytochromeb heterodimer. Studies of X- and A- disease mayprovide insights into the regions of the subunits that specify their tight interaction. Identification of instructive mutationsmaylead the way for directedsite-specific mutationsof the subunitsin the future in an effort to correlate protein structure andfunction. Fourth, studies of the cis-DNAsequences and trans-nuclear protein factors that specify lineage and temporal expression of the X-CGD gene and its inducibility by interferon-y maycontribute morebroadly to an understanding of regulated expression in hematopoietic cells and the molecular basis of interferonaction. It is likely that lineage-specificnuclear factors participate in the regulation of expression of the X-CGD gene. An understanding of their nature maysuggest more generally howthe expressionof genes is coordinatedduring myelomonocytic differentiation. Fifth, recent data on treatment of patients with interferon-~/ suggests that future understandingof various forms of CGDand the regulation of specific components of the oxidase mayprovidethe logical basis for more effective management in the future. In this respect CGD represents one of the fewdisorders in whichbasic studies of genetics haveled to the design of new forms of medical management. Manyoutstanding issues remain untouched.Elucidation of the primary basis of A+-CGD, presumedto be related in somemannerto the rapidly phosphorylated4448kd polypeptidesof neutrophils, will open up investigation of another limb of the NADPH-oxidase system. The structure and role of the elusive "respiratory burst oxidase"merit attention if the various components implicatedby in vitro reconstitution assays are to find their place in the pathway.Althoughin vitro reconstitution assays are being developedto assess the activity of various cytosolic and membrance fractions and purified proteins, progressin understandingthe interactions of specific geneproductswouldbe greatly facilitated if a suitable heterologouscell systemwereestablished. OncecDNAs for a minimalbattery of essential proteins are available, onemightenvisionattemptsto transfer the superoxide-generating system into nonphagocytic cells. If this ambitiousgoal wereaccomplished,the regulatory effects and interactions of the various componentscould be addressed in a direct, systematic manner. ACKNOWLEDGMENTS

The work in the author’s laboratory was supported by a grant from the National Institute of Health and general support from the HowardHughes MedicalInstitute. I amgrateful to numerousindividuals for the developmentof research on CGDin mylaboratory; these include LouKunkel,

Annual Reviews 302

ORKIN

Tony Monaco, Bob Baehner, Brigitte Royer-Pokora, Mary Dinauer, Peter Newburger, Alan Ezekowitz, and A1 Jesaitis. I also very much appreciate Marie Fennell for her work in preparation of the manuscript. The author is an Investigator of the HHMI.


Literature Cited 1. Curnutte, J. T., Babior, B. M. 1987. Chronic granulomatous disease. In Advancesin HumanGenetics, ed. H. Hines, K. Hirschhorn, 10: 229-97. Plenum 2. Tauber,A. I., Borregaard,N., Simons, E., Wright, J. 1983, Chronic granulomatous disease: A syndrome of phagocyteoxidase deficiencies. Medicine 62:286-309 3. Holmes,B., Quie, P. G., Windhorst, D. B., Good,R. A. 1966. Fatal granulomatous disease of childhood: An inborn abnormality of phagocytic function. Lanceti: 1225-28 4. Curnutte, J. T., Whitten, D. M., Babior, B. M. 1974. Defective superoxide productionby granulocytesfrom patients with chronic granulomatous disease. NewEngl. J. Med.290:593-97 5. Sadler, K. L., Badwey,J. A. 1988. Secondmessengersinvolved in superoxide production by neutrophils. In Hematoloyy/Oncology Clinics of North America,PhagocyticDefects II: Abnormalities of the RespiratoryBurst, ed. E. Yeager, J. Curnutte, 2: 185-200. Philadelphia: Saunders 6. Andrews,P. C., Babior, B. M. 1984. Phosphorylationof cytosolic proteins by resting and activated humanneutrophils. Blood64:883-90 7. Hayakawa, T., Suzuiki, K., Suzuki, S., Andrews,P. C., Babior, B. M.1986. A possible role for protein phosphorylation in the activation of the respiratory burst in humanneutrophils. Evidencefrom studies with cells from patients with chronic granulomatous disease. J. Biol. Chem.261:9109-15 8. White,J. R., Huang,C.-K., Hill, J. M. Jr. 1984. Effect of phorbol 12-myristate 13-acetate and its analogue 4alpha-phorbol 12,13-didecanoate on protein phosphorylationand lysosomal enzyme release in rabbit neutrophils.J. Biol. Chem.259:8605-11 9. Okamura,N., Ohashi, S., Nagahisa, N., Ishibashi, S. 1984. C.hanges. in protein phosphorylation In guinea pig polymorphonuclearleukocytes by treatment with membrane-perturbing

agents which stimulate superoxide anion production. Arch. Biochem. Biophys. 228:270-77 I0. Pontremoli,S., Melloni,E., Salamino, F., Sparatore,B., Michetti, M., Sacco, O., Horecker, B. L. 1986. Phosphorylation of proteins in humanneutrophils activated with phorbolmyristate acetate or with chromotactic factor. Arch. Biochem.Biophys. 250: 23-29 11. Segal, W.W., Heyworth,P. G., Cockcroft, S., Barrowman,M. M. 1985. Stimulated neutrophils from patients with autosomal recessive chronic granulomatousdisease fail to phosphorylate a Mr-44,000protein. Nature 316:547-49 C. E., Simons, 12. Whitin,J. C., Chapman, E. R., Chovaniec,M.E., Cohen,H. J. 1980. Correlation between membrane potential changesand superoxideproduction in humangranulocytes stimulated by phorbolmyristate acetate. J. Biol. Chem.255:1874-78 13. Seligrnann,B. E., Ballin, J. I. 1980.Use of lipophilic probes of membrane potential to assess humanneutrophil activation: Abnormality in chronic granulomatous disease. J. Clin. Invest. 66:493-503 C. E., Simons, 14. Whitin,J. C., Chapman, E. R., Chovaniec,M.E., Cohen,H. J. 1980. Correlation between membrane potential changesand superoxideproduction in humangranulocytes stimulated by phorbolmyristate acetate. J. Biol. Chem.255:1874-78 15. Bromberg,Y., Pick, E. 1985. Activation of NADPH-dependent superoxideproductionin a cell-free systemof sodiumdodecylsulfate. J. Biol. Chem. 260:13559~55 16. Curnutte, J. T. 1985. Activation of humanneutrophil nicofinamidcadenine dinucleotide phosphate,reduced(triphosphopridinenucleotide, reduced) oxidase by arachidonic acid in a cell free system.J. Clin. Invest. 75: 174043 17. Curnutte, J. T., Kuver, R., Babior, B. M. 1987. Activationof the respira-

Annual Reviews


MOLECULAR GENETICS OF CGD tory burst oxidase in a fully soluble systemfromhumanneutrophils. J. Biol. Chem.262:6450-52 18. Curnutte, J. T., Kuver, R., Scott, P. J. 1987. Activation of neutrophil NADPH oxidase in a cell-free system. Partial purification of components and characterizationof the activation process. J. Biol. Chem.262:5563-69 19. McPhail,L. C., Shirley, P. S., Clayton, C. C., Snyderman,R. 1985. Activation of the respiratory burst enzymefrom human neutrophilsin a cell-free system: Evidencefor a solublecofactor.J. Clin. Invest. 75:1735-39 20. Gabig, T. G., English, D., Akard, L. P., Scfiell, M.J. 1987.Regulation of neutrophil NADPH oxidase activation in a cell-free systemby guanine nucleotides and fluoride: Evidence for participation of a pertussis and cholera toxin-insensitive G protein. J. Biol. Chem.262:1695-90 21. Nozawa, R., Kato, H., Yokota, T. 1988.Inductionof cytosolic activation factor for NADPH in differentiated HL-60leukemiacells. J. Biochem.103: 43-47 22. Parkinson,J. F., Akard,L. P., Schell, M.J., Gabig,T. G.1987.Cell-freeactivation of phagocyte NADPH-oxidase: tissue anddifferentiation-specific expressionof cytosoliccofactor activity. Biochem.Biophys. Res. Commun. 145: 1198-1204 23. Glass, G. A., DeLisle,D. M., Detogni, P., Gabig, T. G., Magee, B. H., Markert, M., Babior, B. M. 1986. The respiratory burst oxidase of human neutrophils.Furtherstudiesof the purified enzyme.J. Biol. Chem.261: 1324751 24. Markert, M., Glass, G. A., Babior, B. M.1985. Respiratory burst oxidase from humanneutrophils: Purification and someproperties. Proc. Natl. Acad. Sci. USA 82:3144~8 25. Ume,T., Takeshige, K., Minakami,S. 1986. NADPH binding componentof neutrophil superoxide-generatingoxidase. J. Biol. Chem.261:5229-32 26. Bougnoux, P., Vonvini,E., Stevenson, H. C., Markey,S., Zatz, M., Hoffman, T. 1983. Identification of ubiquinone50 as the major methylated nonpolar lipid in humanmonocytes:Regulation of its biosynthesis via methioninedependentpathwaysand relationships to superoxide production. J. Biol. Chem.258:4339-44 27. Gabig, T. G., Kipnes, R. S., Babior, B. M. 1978. Solubilization of the 02formingacivity responsiblefor the res-

303

piratory burst in humanneutrophils.J. Biol. Chem.253:6663-65 28. Lutter, R., VanZwieten,R., Weening, R. S., Hamers,M. N., Roos, D. 1984. Cytochrome b, flavins and ubiquinone50 in nucleated humanneutrophils (PMN cytoplasts). J. Biol. Chem.259: 9603~6 29. Babior, B. M., Kipnes, R. S. 1977. Superoxide-formingenzymefrom humanneutrophils: evidence for a flavin requirement. Blood 50: 517524 30. Hattori, S. 1961.Studiesonthe labile, stable NADH oxidase and peroxidase staining reactions in the isolated particles of horse granulocyte.NagoyaJ. Med. Sci. 23:362-78 31. Ohta, H., Takahashi, H., Hattori, H. 1966. Some oxidative enzymes and cytochromein the specific granules of neutrophil leukocytes. Acta Haematol. Jap. 29:799-808 32. Shinagawa, T., Shinagawa,Y., Tanaka, C. 1966. Electron microscopeand biochemicalstudy of the neutrophilic granules fromleukocytes. J. Electron Micros (Tokyo)15:81-85 33. Segal, A. W., Jones, O. T. G. 1978. Novel cytochromeb system in phagocytic vacuolesof humangranulocytes. Nature 276:515-17 34. Segal, A. W.1988. Cytochromeb-245 and its involvementin the molecular pathology of chronic granulomatous disease. See Ref. 5, 2:213-24 35. Segal, A. W., Cross, A. R., Garcia, R. C., Borregaard, N., Valerius, N. 1983. Absenceof cytochromeb-245in chronicgranulomatous disease: Amulticenter European evaluation of its incidence and relevance. N. En#l. J. Med. 308:245-51 36. Cross, A. R., Jones, O. T. G., Harper, A. M., Segal, A. W.1981. Oxidationreductionproperties of the cytochrome b found in the plasma membrane fraction of humanneutrophils: a possible oxidase in the respiratory burst. Biochem.J. 194:599-606 37. Wood, P. M. 1987. The two redox potential for oxygen reduction to superoxide. TrendsBiochem.Sci. 12: 250-51 38. Harper, A. M., Dunne,M. J., Segal, A. W.1984. Purification ofcytochrome b-245 from humanneutrophils. Biochem. J. 219:519-27 39. Lutter, R., van Schaik, M. L. J., van Zwieten, R., Weverr, R., Roos, D., Hamers,M. N. 1985. Purification and partial characterization of the b-type cytochrome from human polymor-

Annual Reviews


304

ORKIN

phonuclearleukoeytes. J. Biol. Chem. 260:2237~4 40. Pember;S. O., Heyl, B. L., Kinkade, J. M., Jr., Lambeth,J. D. 1984.Cytochrome b~58 from (bovine) granu!ocytes: Partial purificationfromTriton X-114extracts andpropertiesof the isolated cytochrome.J. Biol. Chem.259: : 1.0590-95 4l: Serra, M.C., Bellavite, P., Davoli,A. 1984. Isolation from neutrophil membranes of a complexcontaining active NADPH oxidase and cytochrome b245. Biochim.Biophys.Acta. 788: 13846 42. Harper, A. M., Chaplin, M.F., Segal, A. W. 1985. Cytochromeb-245 from huma,nneutrophils is a glycoprotein. Bioehem.J. 227:783-88 43. Parkos, C. A., Allen, R. A., Cochrane, C. G., Jesaitis, A.J. 1987.Purifiedcytochrome b from human granulocyte plasma membrane is comprisedof two polypeptides with relative molecular weightsof 91,000and 22,000. J. Clin. Invest. 80:732-42 44. Segal, A. W. 1987. Absenceof both cytochromeb-245 subunits from neutrophils in X-linked chronic granulomatous disease. Nature 326: 8891 45. Parkos,C. A., Allen, R. A., Cochrane, C. G., Jesaitis, A. J. 1988. Thequaternay structure of the plasma membrane b-type cytochrome of human granulocytes. Biochim.Biophys. Acta. 932:71-83 46. Baehner, R. L., Nathan, D. G. 1968. Quantitativenitroblue tetrazoliumtest in chronic granulomatousdisease. N. Engl. J. Med.278:971-76 47. Windhorst, D. B., Holmes,B., Good, R. A. 1967. A newlydefined X-linked trait in manwith demonstrationof the Lyoneffect in carrier females.Lanceti: 73239 48. Azimi, P. H., Bodenbender, J. G., Iqintz, R. L., Kontras, S. B. 1968. Chronic granulomatous disease in three female siblings. J. Am. Med. Assoc. 206:286570 49. Borregaard,N., Johansen,K. S., Taudorff,- E., V~andall,J. H. 1979.Cyto¯ chromeb is presentin neutrophilsfrom patients with chronic granulomatous disease. Lanceti: 949-51 50. Bohler, M.-C., Seger, R. A., Mouy,R. 1986. A study of 25 patients with chronic granulomatousdisease: A new classification bycorrelatingrespiratory burst, cytochrome b, andflavorprotein. J. Clin. Invest. 6:136-45 51. Curnutte, J. T., .Berkow, R.: L.,

Roberts, R. L. 1988. Chronicgranulomatousdisease dueto a defect in the cytosolic factor required for NADPH oxidaseactivation. J. Clin. Invest. 81: 606 52. Weening,R. S., Corbeel, L., de Boer, M., Lutter, R., van Zwleten, R., Hamers,M. N., Roos, D. 1985. Cytochromeb deficiency in an autosomal form of chronic granulomatous disease. J. Clin. Invest. 75:915-20 53. Hamers,M.N., de Boer, M., Meerhof, L. J., Weening,R. S., Roos,D. 1984. Complementation in monocytehybrids revealing genetic heterogeneity in chronic granulomatousdisease. Nature 307:553-55 54. Newburger,P. E., Luscinska, F. W., Ryan,T. 1986. Variant chronic granulomatous disease: Modulationof the neutrophil defect by severe infection. Blood 68:91~19 55. Borregaard,N., Cross, A. R., Herlin, T. 1983. A variant form of X-linked chronic granulomatous disease with normalnitroblue tetrazoliumslide test and cytochrome b. Eur. J. Clin. Invest. 13:24348 56. Segal, A. W.,Jones, O. T. (3, Webster, D., Allison, A. C. 1978. Absenceof a newlydescribed cytochromeb from neutrophils of patients with chronic granulomatousdisease. Lancetii: 44649 57. Heyworth,P. G., Segal, A. W. 1986. Further evidencefor the involvement of a phosphoprotein in the respiratory burst oxidase of humanneutrophils. Biochem.J. 239:723-31 58. Caldwell, S. E., McCall, C. E., Hendricks, L., Leone,P. A., Bass, D. A., McPhail, L. C. 1988. Regulation of NADPH oxidase activation and phosphorylation of a 48-kdprotein(s) by cytosolic factor defective in autosomal recessive chronic granulomatousdisease. J. Clin. Invest. 81:1485-96 59. Ishii, E., Irita, K., Fujita, I., Takeshige, K., Kobayashi,M., Usui, T., Ueda, K. 1986. Protein phosphorylationof neutrophils from normal children and patients with chronic granulomatous disease. Eur. J. Pediatr. 145:22-26 60. Ono,Y., Fujii, T., Ogita, K., Kikkawa, U., Igarashi, K., Nishizuka,Y. 1987. Identification of three additional members of rat protein kinase C family; delta; epsilon- and zeta-subspecles. FEBSLett. 226:125-28 61. Tyers, M, Rachubinski, R. A., Stewart,M.I., Varrichio,A. M., Shorr, R. G. Haslam,R. J., Harley,C. B. 1988. Molecular cloning and expression of

Annual Reviews


MOLECULARGENETICS OF CGD the major protein kinase C substrate of platelets. Nature 333:470-73 62. Gabig, T. G., Lefker, B. A. 1984. Deficient flavoprotein component of the NADPH-dependent O2-generating oxidase in the neutrophils from three" patients with chronic granulomatous disease. J. Clin. Invest. 73:701 705 63. Marsh, W. L. 1978. Chronic granulomatous disease, the McLeod syndrome, and the Kell blood groups. Birth Defects 14:9-25 64. Marsh, W. L. 1978. Linkage of the Xg and Kxloci. Cytogenet. Cell Genet. 22: 531-35 65. Marsh, W. L., Oyen, R., Nichols, M. E. 1976. Kx antigen, the McLeod phenotype, and chronic granulomatous disease: Further studies. Vox San9 31:356-62 66. Marsh, W. L., Oyen, R., Nichols, M. E., Allen, F. H. 1975. Chronic granulomatous disease and the Kell blood groups. Br. J. Haematol. 29: 247 62 67. Ito, K., Mukamoto,Y., Konishi, H., Sakura, N., Usui, T. 1979. Kell phenotypes in 15 Japanese patients with chronic granulomatous disease. Vox San9 27:39~40 68. Francke, U. 1984. Random X inactivation resulting in mosaic nullisomy of region Xp2I. 1 -p21.3 associated with heterozygosity for ornithine transcarbamylase deficiency and for chronic granulomatous disease. Cytogenet. Cell Genet. 38:298-307 69. Francke, U., Ochs, H. D., de Martinville, B., Giacalone, J., Lindgren, V., Disteche, C., Pagon, R. A., Hofker, M. H., van Ommen,G.-J. B., Pearson, P. L., Wedgwood,R. J. 1985. Minor Xp21 chromosome deletion in a male associated with expression of Duchenne muscular dystrophy, chronic granulomatous disease, retinitis pigmentosa, and McLeod syndrome. Am. J. Hum. Genet. 37:250-67 70. van Ommen,G.-J. B., Verkerk, J. M. H., Hofker, M. H., Monaco, A. P., Kunkel, L. M., Ray, P., Worton, R., Wieringa, B., Bakker, E., Pearson, P. L. 1986. A physical map of 4 million bp around the Duchenne muscular dystrophy gene on the human X chromosome. Cell 47:49%504 71. Burmeister, M., Monaco,A. P., Gillard, E. F., van Ommen,G.-J. B., Affara, N. A., Feguson-Smith, M. A., Kunkel, L. M., Lehrach, H. 1988. A 10-megabase physical map of humanXp21, including the Duchenne muscular dys.trophy gene. Genomics2:189-202

305

72. Baehner, R. L,, Kunkel, L. M., Monaco, A. P., Haines, J. L., Conneally, P. M., Palmer, C., Heerema, N., Orkin, S. H. 1986. DNAlinkage analy-sis of X-linked chronic granulomatous disease. Proc. Natl. Acad. Sci. USA83: 3398-3401 73. Bertelson, C. V. J., Pogo,.A.O’., Ctiau~ dhuri, A., Marsh, Wi L., Redman, C. M., Banerjee, D., Symmans,W. A., Simon, T., Frey, D., Kunkel, L. M. 1988. Localization of the McLeod locus (XK) within Xp21 by deletion analysis. Am. J. Hum. Genet. 42: 70311 74. Royer-Pokora, B., Kunkel, L. M., Monaco,A. P., Goff, S. C., Newburger, P. E., Baehner, R. L., Cote, F. S., Curnutte, J. T., Orkin, S. H. 1986. Cloning the gene for an inherited humandisorder--chronic granulomatous disease--on the basis of its chromosomal location. Nature 322:3~38 75. Kunkel, L. M., Monaco, A. P., Middlesworth, W., Latt, S. A. 1985. Specific cloning of DNAfragments absent from the DNAof a male patient with an X chromosome deletion. Proc. Natl. Acad. Sci. USA 82:4778-82 76. Monaco,A. P., Bertelson, C. J., Middlesworth, W., Colletti, C. A., Aldridge, J., Fischbeck, K. H., Bartlett, R., Pericak-Vance, M. A., Roses, A. D., Kunkel, L. M. 1985. Detection of deletions spanning the Duchenne muscular dystrophy locus using a tightly linked DNAsegment. 316: 84245 77. Dinauer, M. C., Orkin, S. H., Brown, R., Jesaitis, A. J., Parkos, C. A. 1987. The glycoprotein encoded by the Xlinked chronic granulomatous disease locus is a componentof the neutrophil cytochrome b complex. Nature 327: 717-20 78. Teahan, C., Rowe, P., Parker, P., Totty, N., Segal, A. W. 1987. The Xlinked chronic granulomatous disease gene codes for the B-chain of cytochrome b-245. Nature 327:720-21 79. Parkos, C. A., Dinauer, M. C., Walker, L. E., Allen, R. A., Jesaitis, A. J., Orkin, S. H., 1988. Primary structure and unique expression of the 22-kilodalton light chain of humanneutrophil cytochrome b, Proc. Natl. Acad. Sci. USA 85:3319-23 80. Orkin, S. H. 1987. X-linked chronic granulomatous disease: from chromosomalposition to the in vivo gene product. Trends Genet. 3:149-51 81. Nakamura, M., Murakami, M., Koga, T., Tanaka, Y., Minakami, S. 1987.

Annual Reviews


306

ORKIN

Monoclonal antibody 7D5 raised to cytochrome b558 of human neutrophils: Immunocytochemical detection of the antigen in peripheral phagocytes of normal subject, patients with chronic granulomatous disease, and their carrier mothers. Blood 69: 140408 82. Rotrosen, D., Hunoi, H., Tiffany, H. L., Albert, J., Maloy,W. L., Gallin, J. I., Malech, H. L. 1988. A cytoplasmic carboxyterminal domain of the transmembranous large subunit of neutrophil cytochromeb is involved in activation of the respiratory burst. Clin. Res. 36: 582A 83. Volkman,D. J., Buescher, E. S., Gallin, J. I., Fauci, A. S. 1984. B cell lines as models for inherited phagocytic diseases: Abnormalsuperoxide generation in chronic granulomatous disease and giant granules in Chediak-Higashi syndrome. J. Immunol. 133:3006-3009 84. Pick, E., Gadba, R. 1988. Certain lymphoid cells contain the membraneassociated component of the phagocyte-specific NADPH oxidase. J. Immunol. 140:1611-17 85. Nathan, C. F., Tsunawaki, S. 1986. Secretion of toxic oxygen products by macrophages: Regulatory cytokines and their effects on the oxidase. Ciba Found. Symp. 118:211-30 86. Berton, G., Zeni, L., Cassatella, M. A. 1986. Gammainterferon is able to enhance the oxidative metabolism of humanneutrophils. Biochem. Biophys. Res. Commun. 138:1276-82 87. Cassatella, M. A., Cappeffi, R., Delia Bianca, V. 1988. Interferon gamma activates human neutrophil oxygen metabolism and exocytosis. Immunology 63:499-506 88. Nathan, D. G., Kaplan, G., Levis, W. 1986. Local and systemic effects of intradermal recombinant interferon gammain patients with leptomatous leprosy. N. Engl. J. Med. 315:6 89. Nathan, C. F., Murray, H. W., Wieve, M. E. 1983. Identification of interferon gammaas well as the lymphokine that activates humanmacrophage oxidative metabolism and antimicrobial activity. J. Exp. Med. 158:670~89 90. Newburger,P. E., Ezekowitz, R. A. B., Whitney, C., Wright, J., Orkin, S. H. 1988. Induction of phagocyte tyrochrome b heavy chain gene expression by interferon gamma.Proc. Natl. Aead. Sci. USA 85:5215 19 91. Cassatella, M. A., Hartman, L., Perussia, B., Trinchieri, G. 1988. Tumor necrosis factor and immuneinterferon

synergistically induce cytochrome b245 heavy chain gene expression and NADPHoxidase in human leukemic myeloid cells. Submitted 92. Andrew, P. W., Robertson, A. K., Lowrie, D. B., Cross, A. R., Jones, O. T. 1987. Induction of synthesis of components of the hydrogen peroxidegenerating oxidase during activation of the human monocytic cell line U937 by interferon-gamma. Biochem. J. 248:281-83 93. Reich, N., Evans, B., Levy, D., Fahey, D., Knight, E. Jr., Darnell, J. E. Jr. 1987. Interferon-induced transcription of a gene encoding a 15-kd protein depends on an upstream enhancer element. Proc. Natl. Acad. Sci. USA 84:6394-98 94. Friedman, R. L., Stark, G. R. 1985. Alpha-interferon-induced transcription of HLAand metallothionein genes containing homologous upstream sequences. Nature 314:637-39 95. Korber, B., Mermod, N., Hood, L., Stroynowski, I. 1988. Regulation of gene expression by interferons: control of H-2 promoter responses. Science 239:1302q5 96. Zinn, K., Maniatis, T. 1986. Detection of factors that interact with the human beta-interferon regulatory region in vivo by DNAseI footprinting. Cell 45: 611-18 97. Lomax, K. J., Burch-Whitman, C., Tiffany, H. L., Gallin, J. I., Malech,H. L. 1988. Analysis of chronic granulomatous disease kindreds reveals distinct genetic lesions affecting the same gene product. Clin. Res. 36: 413A 98. Seger, R. A., Tiefenauer, L., Matsunaga, T., Wildfeuer, A., Newburger, P. E. 1983. Chronic granulomatous disease due to granulocytes with abnormal NADPH oxidase activity and deficient cytochrome-b. Blood 61: 423-28 99. Lew, P. D., Southwick, F. S., Stossel, T. P., Whitin, J. C., Simons,E., Cohen, H. J. 1981. A variant of chronic granulomatous disease: Deficient oxidative metabolism due to a low-affinity NADPH oxidase. N. En#l. J. Med. 305: 1329-33 100. Ezekowitz, R. A. B., Orkin, S. H., Newburger, P. E. 1987. Recombinantinterferon gamma augments phagocyte superoxide production and X-chronic granulomatous disease gene expression in X-linked variant chronic granulomatous disease. J. Clin. Invest. 80: 1009-16 101. Orkin, S. H. 1987. Disorders of hemoglobin synthesis: The Thalassemias.

Annual Reviews


MOLECULAR GENETICS OF CGD In The Molecular Basis of Blood Diseases, ed. G. Stamatoyannopoulos, A. Neinhuis, P., Majerus,1: 106-26. Philadelphia: Saunders 102. Ohno,Y., Buescher,E. S., Roberts,R., Metcalf, J. A., Gallin, J. I. 1986. Reevaluationof cytochromeb and flavin adeninedinucleotidein neutrophils from patients with chronic granulomatousdisease and description of a family with probable autosomalrecessive inheritance of cytochromeb deficiency. Blood67:1132-38 103. Nunoi, H., Gallin, J. 1., Malech, H. L. 1988. Different defects in cytosolic factors maybe responsible for autosomalforms of chronic granulomatousdisease (CGD).Clin. Res. 415A 104. Frey, D., Machler, M., Seger, R., Schmid, W., Orkin, S. H. 1988. Genedeletion in a patient with chronic granulomatousdisease and McLeod syndrome: Fine mapping of the Xk genelocus. Blood 71:252-55 105. Newburger, P. E., Cohen, H. J., Rothchild,S. B., Hobbins,J. C., Malawista, S. E., Mahoney, M.J. 1979.Prenatal diagnosis of chronic granulomatousdisease. N. Engl. J. Med.300: 178-81 106. Orkin, S. H. 1984. Prenatal diagnosis of hemoglobin disorders of DNA analysis. Blood63:249-53 107. Saiki, R. K., Scharf, S., Faloona,F., Mullis, K. B., Horn, G. T., Erlich, H. A., Arnheim,N. 1985. Enzymatic amplification of ~-globin genomic sequencesand restriction site analysis for diagnosisof sickle cell anemia.Science 230:1350-54 108. Forrest, C. B., Forehand,J. R., Axtell, R. A., Roberts, R. L., Johnston, R. B. 1988.Clinical features and current managementof chronic granulomatousdisease. See Ref. 5, 2:253-66 109. Babior, B. M., Peters, W.A. 1981.The O2-producing enzymeof humanneutrophils: Further properties. J. Biol. Chem. 256:2321-23

307

110. Babior, B. M., Curnutte,J. T., McMurrich, B. J. 1976.Theparticulate superoxide-forming system from human neutrophils: Properties of the system and further evidence supporting its participation in the respiratory bursts. J. Clin. Invest. 58:989-96 111. Ezekowitz,R. A. B., Dinauer, M. C., Jaffe, H. S., Orkin, S. H., Newburger, P. E. 1988. Partial correction of the phagocytedefect in patients with Xlinked chronic granulomatousdisease by subcutaneousinterferon gamma.N. En#l. J. Med.319:146-51 112. Sechler, J. M. G., Malech, M. L., White,C. J., Gallin, J. I. 1988.Recombinant human interferon-y reconstitutes defectivephagocyefunction in patients with chronic granulomatous diseaseof childhood.Proc.Natl. Acad. Sci. USA85’. 4874-78 113. Lehrer, R. I., Ganz,T., Selsted, M.E. 1988. Oxygen-independent bactericidal system. Mechanismsand disorders. Hematol./Oncol.Clin. N. Am.2: 159269 114. Williams, D. A., Orkin, S. H. 1986. Somaticgene therapy: current status andfuture prospects.J. Clin.Invest. 77: 1053-56 115. Rappeport, J. M., Newburger,P. E., Goldblum, R. M., Goldman, A. S., Nathan, D. G., Parkman, R. 1982. Allogeneic bone marrowtransplantation for chronic granulomatousdisease. J. Pediatr. 101:952-55 116. Jackson, I. J. 1987. Thereal reverse genetics: targeted mutagenesisin the mouse.TrendsGenet. 3:119-20 117. Robertson,E. J., Bradley,A., Kuehn, M., Evans, M. 1986. Germ-linetransmissionof genes introduced into cultured pluripotentcells byretroviral vectors. Nature323:445-48 118. Thomas,K. R., Fogler, K. R., Capecchi, M.R. 1986.Highfrequencyis targeting of genesto specific sites in the mammaliangenome.Cell 44:419-28





Ann.Rev. Immunol.1989. 7.’309-37 Copyright© 1989by AnnualReviewsInc. All rights reserved

CELL BIOLOGY OF CYTOTOXIC AND HELPER T CELL FUNCTIONS"Immunofluorescence MicroscopicStudies of Single Cells and Cell Couples AbrahamKupfer* and S. J. Singer Department of Biology, University of California at San Diego, La Jolla, California 92093

INTRODUCTION This review deals with certain important features of the cell biology of natural killer (NK) cells, cytotoxic T lymphocytes (CTL), and helper (Th) cells. Thesefeatures appearedoriginally in studies of the cell biology of cells other than immunological cells, and their relevance to immune cell interactions was then appreciated and explored. This brief review is therefore meant to be selective rather than exhaustive, and it focuses primarily on problemsthat have been intensively studied in our laboratory over the past decade. At the outset, it should be emphasized that we concentrate on investigations of single (generally cloned) T cells and cell couples, rather than of cell populations. Recent reviews have appeared that deal with other aspects of the cell biology of T cells and their interactions (1-7). The entr+e to this work stemmedfrom our investigations of eukaryotic cell motility; these have been separately reviewed (8). These studies are considered briefly first, since they not only introduce the ideas that have played a prominent part in our immunological work, but they also demon-

*Present Address: National Jewish Center for Immunologyand Respiratory Medicine, Denver, Colorado 80206.

309 07320582/89/0410-0309502.00

Annual Reviews 310

KUPFER& SINGER

strate the generality of someof the phenomenaobserved for understanding cellular interactions in immunology.


THE DIRECTED MIGRATION OF CELLS Whena eukaryotic cell undergoes directed migration in response to a polarized signal, it is generally agreed that its leading edge is propelled forward in the direction of the signal by intracellular forces mediated by the actin cytoskeleton. In addition, however,the cell’s normalrecycling of its surface membranebecomes polarized, so that new membranemass is continually inserted at the leading edge of the cell while other membrane mass is internalized rearwards of the leading edge (8). This polarized recycling of membranemass appears to be essential for directed migration to occur (9), but whether the essential feature is the insertion of new membranemass per se, or the directed secretion of adhesive or other soluble proteins at the leading edge that accompanies such membrane insertion, or both, is not clear. What mechanism determines that membrane recycling becomes polarized? The answer we found (10, 11) is that the Golgi apparatus (GA)inside the cell undergoesa reorientation to face the region of the cell surface that is to becomethe leading edge, within minutes after receipt of a polarized signal to move. The GAis the organelle that generates the membraneboundvesicles containing secretory componentsthat go on to fuse with the plasma membrane. The vesicular membranemass is thereby introduced into the cell membraneand the secretory contents of the vesicles are simultaneously released to the cell exterior (12). In most eukaryotic cells the GAis a compactset of stacked vesicular structures usually present to one side of the cell nucleus. Reorientation of the GAin a particular direction apparently ensures that vesicular traffic derived from the GAis delivered to that region of the plasma membraneapposed to the GA(9, 13). Howthe GAreorientation is achieved is not yet clear, but microtubules probably play an important role. Interphase metazoancells usually contain a single organelle, the microtubule-organizing center (MTOC) that acts the nucleation center for the polymerization of all of the microtubules in the cell (14). Double immunofluorescence observations of the MTOC and the GAin individual cells of widely different types indicate that labeling for the MTOC and GAis always found to be superimposed, and a reorientation of the GAis always accompaniedby the same reorientation of the MTOC (8). This codistribution and coordinated reorientation of the two organelles probably reflects somephysical linkages between them, but this is not established. Oneplausible possibility is that the process that signals

Annual Reviews CELL BIOLOGY OF CELL COUPLES

311

the cell to move may somehowinduce a microtubule-mediated torque on the MTOC which reorients the MTOC along with the GAthat is physically linked to it.

CYTOTOXICITY


NK Interactions Midwaythrough these studies of cell motility in our laboratory, Dr. Gunther Dennert informed us of his recent investigations on the cloning of mouseNKcells and the evidence that an NKcell lysed its bound target cell by secreting cytotoxic componentsto it (15). It occurred to us that, like cell motility, the NKcell : target cell interaction mightbe an instance of a polarized signal (this time from the boundtarget to the NKcell) inducing a polarized secretion (from the NKto the target). The prediction was that in couples formed between cloned NKand target cells, the MTOC/GA inside the NKcell would be found oriented to face the boundtarget. This was then shownto be the case (16). Similar results were found with human NKcells (17, 18).

CTLInteractions In the NK: target cell system, the nature of the surface molecules whose transcellular bonding is responsible for the cell:cell interaction is unknown,and the cytotoxicity has limited specificity. A similar cytotoxic phenomenonoccurs when CTLinteract specifically with allogeneic target cells; in this system, however, it is knownthat the specific binding of a clonotypic T cell receptor (TcR) on the surface of the CTLto a class-I MHCmolecule on the surface of the target cell initiates the cytotoxic process. Wetherefore next examined individual cell couples formed between cloned CTLand their targets. MXO¢/~h REORIEYa’ha’ION With allospecific CTL:target cell couples we found by immunofluorescence observations that, in a way similar to the NKcell couples, the MTOC/GA inside the CTLwas oriented to face its boundallospecific target (19). Earlier electron microscopic studies with CTLpopulations provided evidence that the GAinside the CTLbecomes reoriented to face the contact region with the target (20). Geiger et al (21) had also reported that in CTL: target couples the MTOC within the CTL was oriented towardthe area of cell : cell contact. Theysuggested that the target cell binds to an already polarized CTLat a cell surface region apposed to the MTOC.The cumulative evidence cited below, however, strongly indicates that this interpretation is incorrect; rather, the CTL binds to the target cell at randomregions of the twocell surfaces, following

Annual Reviews 312

KUPFER& SINGER


which the MTOC inside the CTLis rapidly reoriented to face the cell :cell contact region. Onedemonstration of this involves the effects of Ca+2 on the properties of the allospecific CTL: target cell couples. It had been shown(22-24) that such couples could form in the presence of Mg÷2 and in the absence of Ca+2, but that cell killing would no~ occur unless Ca+2 was restored. We found (25) that CTL: target cell couples formed in Mg+2-containing÷2free media exhibited random orientation of the MTOC/GA inside the CTL (Figure 1D); the addition of +2 tothese couples, however, led to n early all couples having the MTOC inside the CTLfacing the cell : cell contact site (Figure 1A). Theseresults strongly suggest that cell : cell binding occurs at cell surface regions unrelated to the position of the MTOC/GA inside the CTL, and that a subsequent Ca+Z-mediatedpolarized signal induces the MTOC/GA reorientation inside the CTLto face the cell : cell contact region. That the binding of specific target cells induces a rapid Ca2+ influx into the CTLhas been demonstrated (26). Another demonstration involved multiple cell conjugates, with several specific target cells attached simultaneously to a single CTL.It is known that in such cases, the targets are lysed sequentially and not simultaneously (27-29). Our results (25) were entirely consistent with the conclusion the MTOC/GA in the CTLfirst orients to face one of the bound target cells which then proceeds to lyse, with its ghost remaining attached to the CTL; the CTLMTOC/GA then reorients to face the second bound target to be lysed; and so on until all the target cells are destroyed. Again,cytolysis appears to be coupled to a signalled reorientation of the MTOC/GA. In addition to the reorientation of the MTOC/GA, cinematographic observations by Nomarskioptics indicated that a rapid redistribution of cytoplasmic granules occurs inside CTLto face toward the boundallospecific target cell and that the fusion of these granules with the CTLmembraneat that region begins within minutes after specific cell: cell contact (30). In these cases of NKand CTLcell couples, the polarization of the MTOC/GA that always accompanies a cytolytic interaction is unidirectional; it occurs inside the effector cell but not the target. A question that then arose was whether such organelle reorientation is a specific characteristic of cytolytic cells that is not shared by non-T-cell targets, or whether instead it reflects the engagementof specific receptor molecules on the effector cells. This question was answered by studies carried out with cell couples formed between two CTL, of the types a anti-b and b anti-c. In such mixedCTLpopulations, it was knownthat only the b antic CTLis lysed (31). Correspondingly, we found (32) that with a two system that showed unidirectional killing, the MTOC/GA in the effector CTL, but not that in its bound target CTL, was reoriented to face the

Annual Reviews


CELL BIOLOGY OF CELL COUPLES

313

Figure 1 The effect of extracellular Cai2 on the orientation of the MTOC and the distribution of talin inside CTL : target cell couples. The allogeneic cloned CTL 18C3 (H2bantiH2d) were mixed with an equal number of the target cells P815 (H2d) in the absence ( A X ) or presence (D-F) of Mg+2/EGTA,as in (25). The cells were plated out, fixed, and double immunofluorescently labeled with affinity purified guinea pig anti-tubulin (A,D) and affinitypurified rabbit anti-talin antibodies (B,E) followed by the appropriate fluorescent-conjugated secondary antibodies. The Nomarski pictures of the same cells that are immunolabeled in A,B, and in D,E are shown in C and F, respectively. The thick arrows point to the cell : cell contact area, and the thin arrows to the position of the MTOC in the CTL. Note that in cell couples formed in the presence of extracellular Ca+*(A-C), the MTOC is oriented to face, and talin is concentrated at, the contact site, but that in the absence of extracellular Ca+Z, the MTOC (D) is not oriented to the contact area but talin (E) is still concentrated at this site. The MTOCs in the target cells in both cases faced away from the contact area. The bar in C represents 10 ,urn (A. Kupfer, S . J. Singer, unpublished).

Annual Reviews


314

KUPFER & SINGER

cell:cell contact region. These results indicate that it is the polarized engagementof the TcRon the bound effector CTLthat is required for the reorientation of the MTOC/GA in that cell, whereas the simultaneous engagement of the class-I MHCmolecules on the target cell does not reorient the MTOC/GA inside the target. In all of the experimentsso far described, and in additional ones reported in the original publications, a perfect correlation was found between the reorientation of the MTOC/GA inside the effector cell and the productive lysis of its bound target cell. This suggests that the MTOC/GA reorien¯ tation is an obligatory accompaniment of the lytic activation of the effector cell. In this connection, it is of interest that certain combinationsof allogeneic CTLand target cells do not require Ca+2 in the mediumfor the targets to be lysed; yet even with such CTL:target cell couples in the ¯ absence of Ca+2, the MTOC/GA reorientation is’observed inside the boundCTL(33). In other studies of cloned CTL: CTLcouples (34), killing was never observed without a reorientation of the MTOC/GA, although in one case with a cloned CTLtarget resistant to lysis, the MTOC/GA reorientation occurred in the effector CTLof the couple, but the target was not lysed. These results support the proposal that the MTOC/GA reorientation in the effector cell is a necessary but not sufficient condition to produce lysis of a boundtarget. CYTOSKELETAL REORGANIZATION In the case of the two CTLsystem a anti-b and b anti-c, where only the b anti-c CTLwas lysed (see, however, 35-37), we showedthat part of the explanation for the unidirectionality of killing is that there is a unidirectional polarization of the MTOC/GA in the boundeffector CTL. The function served by this organelle reorientation is presumablyto direct secretion of cytotoxic componentsfrom the effector to the target where the two cells are bound together (see below). As there is no indication that the two cell surfaces at the cell: cell contact are rendered discontinuous, such putative cytotoxic componentswould be expected to be secreted into the confined intercellular spaces in the region of cell:cell contact. Therefore, even if secretion was unidirectional, both the effector and target CTLmembranes would presumably be equally exposed to the action of the secreted cytotoxic components.Whythen was the effector CTLnot as susceptible to killing as the target CTL? One explanation that occurred to us is that a unidirectional cytoskeletal rearrangement under the membraneof the bound effector cell might occur which somehowprotected the membraneagainst the cytolytic process. Whateverthe merits of this~explanation in retrospect (see below), it led to examine by immunofluorescencemicroscopy the distribution of several important cytoskeletal proteins in couples formed with CTL. These cyto-

Annual Reviews



315

skeletal proteins included ~-actinin (38), vinculin (39, 40), talin (41), 200 kd protein (42), and fimbrin (43). They were chosen for investigation because in a wide variety of different cell types each of these proteins has been proposed to play a role in the linkages of actin microfilaments to one another or to the cell membrane(44). Few studies of these proteins have been made in lymphocytes. In a related connection, there are reports that actin is concentrated inside CTL(45) and NKcells (46) at the sites contact with target cells, but as the targets were not themselves CTL, whether such actin concentration is unidirectional was not clear. In the cell couples formed between two CTLof the type a anti-b and b anti-c, we showed(32) that talin, but not any of the other four cytoskeletal proteins examined, became concentrated under the effector CTLmembrane where it was in contact with the specific target CTL.However,all five proteins remaineduniformly distributed in the boundtarget. The talin redistribution is therefore specific and unidirectional. This talin effect was observed generally inside the CTLin specific CTL-target cell couples whateverthe type of lysable target cell (Figure 1B,E), as well as inside cells boundto lysable targets (32). Interestingly, the redistribution of talin was not dependent upon the presence of Ca+: in the medium(Figure IB,E). +2 The reorientation of the MTOCwhich is dependent on Ca (Figure IA,D), is therefore not tightly coupled to the talin redistribution. The properties of the talin molecule and its interactions are discussed further below. At this stage in our studies of cytotoxic cell interactions, we had discovered two rapid, readily detectable intracellular events in cytolytic effectot cells that invariably accompaniedtheir binding to specific target cells: a remarkable reorientation of the MTOC/GA complex to face the bound target; and a unidirectional redistribution of the cytoskeletal protein talin under the contacting CTLmembrane. In the case of CTL, these effects must be the consequence of the appropriate engagementof the clonotypic TcRmolecules on its surface. Becausesecretion (of lymphokines)by helper T cells was knownto be involved in T cell help, and because the TcR molecule on CTLis closely similar to the TcRon helper T cells (cf 47), then becameinterested in whether,in cell : cell interactions involving helper T cells and antigen-presenting B cells, similar intracellular reorganization phenomena occurred. T CELL HELP: INTRACELLULAR REORGANIZATIONS Cytotoxicity is extremely important physiologically, serving to protect the organism against viral and some bacterial diseases, and it is perhaps

Annual Reviews


316

KUPEER& SINGER

involved in immunesurveillance against cancer (48). However,the cell : cell interaction is a dead end: the target cell is lysed. In the case of Thcells and antigen-presenting cells (APC), the interactions have been implicated the proliferation and ultimate differentiation of B-cell APCinto antibodysecreting ptasmacytes, as well as in the proliferation of, and lymphokine secretion by, the Th cells. The helper T cell system therefore involves cell differentiation rather than cell killing, and as such maybe related mechanistically to other important binary cell:cell interactions in developmental biology. Cell couples of CTLand their allogeneic target cells can be produced and isolated quite easily (49), but the direct binary interaction of syngeneic Th and APChas been muchmore difficult to investigate. This has had to await the cloning of specific Th cells, or the production of hybridomasof specific Thcells (50). In our initial studies in collaboration with S. L. Swain and C. A. Janeway, Jr. (51, 52), cloned mouseTh cell lines were used, and mouse cell hybridomas or lymphomaspulsed with very large concentrations of appropriate antigens served as APC. This allowed us to produce and isolate specific Th : APCcell couples, as well as control nonspecific couples, for immunofluorescencestudies by the same methods that we had earlier used with NKand CTLsystems. The cloned Th cell lines used included D10.G4.1(referred to as D10) and D8 specific for the egg white antigens conalbumin(Con) and ovalbumin(Ova), respectively, in the context of in addition, some experiments were carried out with the T-T hybridoma 2H.10.H1 with a helper response specific for pigeon cytochrome c (Cyt) and Iak. As APC,the B-cell hybridomasLK(Ia d, Iak), LB(Ia d, Iab), BCL1 (Ia d) and CH12Oak) (52) were employed. The Th cells and antigen-pulsed APCwere mixedin equal numbers, the mixtures lightly pelleted, incubated for 10 rain, and then resuspended for plating. They were then fixed and immunofluorescently labeled. Under these conditions, cell couples were formed with both specific and nonspecific combinations of Th and APC. The specific couples could not readily be distinguished from the nonspecific by Nomarski optics. However, in every specific Th:APC combination examined, the Th of most (80-95%) of the cell couples exhibited MTOC/GA oriented to face the bound APC, whereas in the case of nonspecific couples, either presenting a nonspecific antigen or the wrong Ia on the APC, the Th MTOC/GA was randomly oriented. Further studies of such couples showed that talin, but not ~-actinin or vinculin, was concentrated under the Th membranein contact with specific APC, but not with nonspecific APC(52). As predicted, therefore, these results were closely similar to those we had earlier obtained with CTL:target cell couples.

Annual Reviews



317

These two objective criteria--MTOC/GAreorientation and talin redistribution in the boundTh cell--therefore served to establish that a specific binary interaction of Th and APChad occurred. Although prior binding experiments with cell populations (53) provided compelling evidence for specific interactions, these studies together with the cell:cell binding studies of Sanders et al (54) represent the most direct evidence for a binary interaction. In the system consisting of D10Th cells and antigen-pulsed LKcells, evidence was then obtained from chloroquine inhibition and other experiments (52) that antigen processing by the APCwas required in order produce the MTOC/GA reorientation in the Th cell. It was also shown that, as observed with the cytotoxic cell couples, the MTOC/GA reorientation in the Th required extracelllar Ca÷2, whereasthe talin redistribution did not (52). In our studies ofTh : APCcouples up to this point, the B-cell hybridomas and lymphomasthat were used for antigen presentation, lacked the appropriate surface IgM,and therefore were not specific for the antigens recognized by the cloned Th cells we employed. Antigen presentation by such cells required that they be pulsed with very large doses of the antigen, of the order of 500 #g/ml. To achieve conditions that were more nearly physiological, we took advantage of the production of an antigen-specific B-cell hybridoma by Hozumi and colleagues (55). They had stably transfected A20 B hybridoma cells (Ia ~) with the rearranged genomic DNA for the H and L chains of a surface IgMwith binding specificity for the 2,4,6-trinitrophenyl (TNP)hapten. These transfected cells are referred as A20-HL.In our experiments (A. Kupfer, D. R. Wegmann,S. J. Singer, in preparation) the A20-HLcells were pulsed with 2,4-dinitrophenyl (DNP)-modified ovalbumin (DNP-Ova), the DNP and TNP haptens cross-reacting extensively. The DNP-Ovawas used at varying concentrations below I/~g/ml (contrast this with the 500 #g/ml of antigen used with nonspecific APC). Within this concentration interval, the DNP-Ova was endocytosed and processed by the cells via the surface anti-TNP IgM, and peptide fragments of Ova were presented. These APCwere then used to form specific Th : APCcouples, using as Th the cloned cell line D20.36 (referred to as D36) with specificity for the peptide sequence 323-339 Ova in the context of la d (D. R. Wegmann, in preparation). These couples were then examined, as in the previous studies, by double immunofluorescence for their MTOC orientations and talin distributions, as well as for LFA-1on the T-cell surface, as functions of the DNP-Ovaconcentration used. Onaliquots of similar cell mixtures, the activation of the population of D36 Th cells by the APCto undergo proliferation was measured by 3H-thymidine incorporation. The results for MTOC and talin

Annual Reviews


318

KUPFER & SINGER

are discussedin this section andthe LFA-l-datain a later section. At the highest concentrations of DNP-Ova used (0.2 #g/ml), the D36cells were fully activated, and most of the D36: APCcouples showedreorientation of the MTOC/GA inside the Thcell, as well as a collection of talin in the region of the D36cell membrane that was in contact with the APC.With decreasing concentrations of DNP-Ova,below 0.1 #g/ml, there was a progressivedecreasein the fraction of D36cells that wereactivated, and this wasparalleled by a decreasein the fraction of cell couplesexhibiting a reorientation of the MTOC/GA. At DNP-Ova concentrations less than 0.02/~g/ml, little cell activation or MTOC/GA reorientation was found, althoughtalin wasstill collectedin the cell : cell contactregion. Theseresults provide further strong evidencethat a strict correlation exists betweenthe rapid reorientation of the MTOC/GA inside the effector T cells, andthe activationof these cells, inducedby the specific interaction with their congenercells. In addition, they confirmthat the reorientation of the MTOC/GA and redistribution of talin are not tightly coupled. ON THE FUNCTIONS OF INTRACELLULAR REORGANIZATIONS IN EFFECTOR T CELLS Theevidence presented to this point indicates that closely similar and rather massive rearrangements of organelles and the cytoskeleton invariably occur inside NK,CTL,or Th cells whenthey are specifically boundto their congenercells. Whyare these changesso critically associated with two such different effector cell functions as cytotoxicity and T cell help7This questionis addressedin the next two sections. The Possible Functions Served by Reorientation of the MTOC/GA As mentionedearlier, the MTOC/GA reorientation could serve to polarize and direct secretion and insertion of membrane massto the region of the surfaceof an effector cell that is in contactwith its specific congenercell. In the case of motile cells (13), the directed insertion of newmembrane massinto the leading edgeof the cell, oppositeto the perinuclear GA,has beendirectly demonstrated.In the case of cytotoxic cells, proposalshave been madethat one or morecytotoxic componentsare secreted from the effector cell to its boundtarget (cf 4); such componentshave included perforins, possessing complement-likemembrane permeability properties; proteasespresent in cytotoxic cell granulesalong with perforins; and tumor necrosisfactors. Though eachis capableof causingcell lysis in vitro, none of these components,has been definitively shownto be the "magicbullet" of physiological NKand CTLactivities. Furthermore,other lytic mech-

Annual Reviews



319

anisms than secretory ones are not ruled out (56). Nevertheless, the obligatory and specific reorientation of the MTOC/GA that occurs upon stimulating these cells is strongly consistent with a secretory process for cytotoxicity, although, of course, it does not establish it. In the case of Th cell function, a number of soluble T cell-derived lymphokines can serve as B-cell growth and differentiation factors in vitro, apparently depending on the Th subset (57). The T cell is also known to express IL-2 receptors on its surface upon activation (58). The reorientation of the MTOC/GA inside the Th cell ensures that the secretion of these growth and differentiation factors, via secretory vesicles derived from the GA, is specifically directed towards the bound APC. It is of great significance that, by this means, only the specific APC would be stimulated by the soluble, non-antigen-specific secreted factors derived from the Th cell that was bound to it. Similarly, newly expressed membrane molecules, such as TL-2 receptors, would be inserted directly from GA-derived vesicles into the Th membrane where it is in contact with the APC. This would provide confined access of these receptors to their soluble ligands that were also present in the intercellular space. A preliininay experiment that is suggestive of such a directed stimulation of APC in Th : APC couples is shown in Figure 2 (A. Kupfer, S. L. Swain, S. J. Singer, unpublished experiments). The in vivo B-cell tumor line CH12

Figure 2 A demonstration of a possible polarized induction of B-cell proliferation by direct contact with cloned Th cells. Cells of the in vivo B lymphoma CHI2 were pulsed overnight in vitro with Con (500 pg/ml) and mixed with D10 cells at a 2 : 1 ratio. By 10 hr later, the cells were plated out, fixed, and double immunofluorescently labeled with rabbit anti-tubulin antibodies (A) to display the MTOCs, and goat anti-mouse Ig antibodies (B) to identify the CHI2 cells in the aggregates. Panel (C) shows the Nomarski picture of the same aggregate of cells. Note that one D10 cell is bound to two CH12 cells, one of which is larger than the other, and is undergoing mitosis (see small arrows pointing to the asters). Note also that the MTOC in the D10 cell (large arrow) is facing the dividing CH 12 cell. The bar in C represents 10 ,urn (A. Kupfer, S. L. Swain, and S. J . Singer, unpublished).

Annual Reviews


320

KUPFER & SINGER

Oak), which does not normally grow or divide in in vitro culture, was used as APC.After pulsing the CH12cells with 500/~g/ml Con, D10 cells were mixed at a 2:1 ratio with the APC. Compared to 1:1 mixtures, this resulted in an increased production of individual cell clusters containing two specific APCbound to one DI0 cell (Figure 2C). Whensuch cell mixtures were processed 10 hr after their formation and incubation at 37°C, clusters were occasionally observed in which simultaneously one large and one small CH12APCwere attached to the same DI0 cell; the MTOC/GA in the DI0 cell was oriented to face the larger APC(Figure 2A, lower arrow); and the larger APCwas in the process of cell division, exhibiting two aster-like foci of microtubule concentration (Figure 2A, upper arrows). Further studies of this system need to be carried out, but these preliminary results are consistent with the following suggestions: that the cluster remained stable during the 10 hr after its formation; that the MTOC/GA in the DI0 Th cell became reoriented to face one of the two originally identical small APC;that thereafter proliferative signals were transmitted from the D 10 to this APCwhich caused it to undergo enlargement and cell division, while the other bound APCwas not stimulated. The Possible

Functions

Served by Talin Redistributions

The unidirectional redistribution of talin under the membraneof the effector T cell where it contacted the specific congener cell was first observed with CTL: target cell couples (32). The notion then was that the talin redistribution might reflect a mechanismto protect the effector CTL from its own cytotoxic secretions. However, when it became clear that a similar talin redistribution occurred in Th cells bound to specific APC, more general functions were indicated. Several possible additional functions can be suggested, including: (a) the further stabilization of specific cell : cell adhesions, to allow prolongedperiods for the cellular interactions within individual couples as maybe required for the secretion of possibly late-acting T cell-derived B-cell growth differentiation factors, such as IL5 (59); (b) the expedition of the localized fusion of GA-derivedsecretory vesicles with the effector T-cell membrane at the region of cell : cell contact; and (c) the participation of talin in the recruitment of LFA-1into the cell : cell contact region (see below) to promoteintercellular adhesion. have no direct evidence, however,bearing on these or other possible talinmediated functions. T CELL

HELP:

SURFACE

PHENOMENA

After concentrating on intracellular events in effector T cells accompanying their interactions with congener cells, we next turned to investigate

Annual Reviews



321

phenomenaat the T-cell surface. This transition has had some drastic sociobiological consequences, particularly with respect to the population density of cellular immunologists involved. An appropriate metaphor for leaving the interior of the T cell for its surface is that it was like going from central Wyomingin the dead of winter to Coney Island on Labor Day. Finding and making appropriate reference to all of the relevant contributions to T cell surface-properties is about equivalent to naming all the ConeyIslanders wearing bikinis. Webegin with some general remarks about membrane phenomena. In polarizedspecific cell : cell interactions, suchas are characteristic of effector T cell:congener cell couples, everything that happens is initiated at the two cell surfaces. Oneor more types of integral protein receptor molecules in the surface membrane of one cell initially bind to their respective ligand molecules in the other. The early productive consequencesof this transcellular binding include: (a) the induction of one or more polarized signals that ultimately affect the intracellular states of one or both cell partners; and (b) the development of a cell:cell adhesion of some appropriate stability. To discuss how these consequences mayarise, a few aspects of membranedynamics first need to be introduced. Mutual

Cappin9

It is important to realize that the transcellular binding of two integral membraneproteins can often lead to their mutual redistribution in their respective fluid membranes. This phenomenonwas originally pointed out by Singer (60) and has been experimentally (61, 62) and theoretically investigated. Let us suppose that we have two cells P and Q with monomeric cell surface molecules A and a, respectively, that can form a transcellular bond (Figure 3A). Supposethat the intrinsic reaction A + a ~ A has a first-order reverse rate constant kr. If the membraneconcentrations of A and a are sufficiently large, and kr is sufficiently small, the formation of one or a small number of transcellular A-a bonds can maintain a localized cell : cell contact long enough(Figure 3B) for the diffusion in the membraneof other A and a molecules into that contact to form additional A-a bonds and thereby to stabilize the area of cell:cell contact (Figure 3C). Because A and a molecules jointly becomecollected into the contact area, the phenomenonis referred to as "mutual capping," by analogy to the well-knownprocess of antibody-induced capping of manycell surface molecules. It seems likely that mutual capping phenomenaprovide the molecular basis for the formation of cell:cell adhesions in general. This proposal predicts that the cell surface area exhibiting mutual cap formation should alwayscoincide precisely with the area of morphologicalcell : cell contact.

Annual Reviews


322

KUPFER & SINGER

Figure3 Aschematicrepresentationof mutualcappingof a single receptor-ligandpair. Underappropriateconditions,the formationof oneor a fewtranscellularreceptor-ligand bonds(B) results in the diffusionalrecruitmentandbindingof other pairs (C) to make stablecell : cell contact.In the process,the local morphology of the cell surfacesmaychange (C). Reproduced withpermissionfrom(64).

Weshowbelow that this is the case (see also Figure 6 in 60). In addition, however, several other factors may comeinto play. If extensive areas of two interacting cells are "zippered" together by a mutual capping process, the local morphologyof the two cell surfaces may change in a concerted fashion to accommodate the making of the maximumpermissible number of A -a bonds at equilibrium. This maylead to a local flattening together of the two originally curved cell surfaces (as depicted schematically in Figure 3C; cf Figure 5 in 17), or to an extensive interdigitation of villous projections of the two cell surfaces at their contact regions (cf Figure in 17). Furthermore, cytoskeletal-membrane attachments, which often accompanythe antibody-induced capping of membraneintegral proteins, may also be induced during mutual capping processes (see Figure 1B,E and below), Such cytoskeletal involvements mayaffect the morphologyof cell:cell contacts, as well as the stability of the cell:cell adhesion by restricting the diffusibility or turnover of membranecomponents. The situation can be more complicated than that indicated in Figure 3

Annual Reviews CELL BIOLOGY OF CELL COUPLES

323

if cell P expresses independentsurface molecules .4, B .... , that can form transcellular bondsto moleculesa, b ..... respectively on cell Q. Depending on the conditions, "mutual co-capping" of ,4-a, B-b, etc, pairs may occur into the samecell : cell contact region (64). If the properties of‘4 and B molecules are not independent of one another, still other complexities can arise (next section).


Syn-capping There are several possible types of noncovalent interactions of different integral protein molecules ‘4 and B in a membrane. The two monomer molecules mayhave a strong affinity for one another and form a molecular complexin the membrane.In such instances, the antibody-induced capping of A results in the capping of B along with A, and vice versa with antibodies to B. This is referred to as co-capping. This is the case, for example,with the TcR and the T3 complex in T-cell membranes(65). There are other instances, however, where the antibody-induced capping of ‘4 results in the collection of B with the A caps, but the capping of B does not affect the uniform distribution of A in the membrane(66). One explanation for such cases is that when an appropriately induced aggregation of A molecules into small clusters occurs, such clusters of A have a higher affinity for B molecules than does monomericA. This is not necessarily reciprocal, however; clusters of B need not have an increased affinity for A. Alternatively, certain signals maymodifyA so that it acquires an affinity for B, whereas unmodified A and B molecules do not interact. Such apparently nonreciprocal intramembranous interactions, which have been encountered, for example, in the association of viral glycoproteins and other integral proteins with class-I MHC molecules (66, 67), have been called syncapping (66), to distinguish them from co-capping. The evidence cited below suggests that, under appropriate conditions, certain assessory molecules in T-cell membranesmayundergo syn-capping with the T-cell receptor. The Dynamics of the T-Cell

Receptor

in Th ." APC Couples

Our experimental studies of cell surface phenomenahave been largely confined to Th cell interactions. Th cells bear on their surfaces clonotypic TcRwith specificity for a complexligand on the APC;this ligand consists of a fragment of the antigen molecule bound to a class-II MHC molecule (Ag/class-II MHC).With cell couples formed between D10 cells (specific for Con/Ia k) and LKcells pulsed with Con, immunolabeling of the TcR on fixed couples showed that in about 70%of the couples that exhibited a redistribution of talin, the TcRwas significantly (but not completely) concentrated into the cell : cell contact region (68). The area of TcRconcentration corresponded exactly to the morphological cell:cell contact

Annual Reviews 324

KUPFER& SINGER

region as viewed by Nomarski optics. In nonspecific couples that were morphologically similar to the specific couples, no significant TcRredistribution was observed. These results therefore correspond to the prediction of a mutual capping of the TcRon the Th cell with its Ag/class-II MHCligand on the APC. (As the specific ligand on the APCcould not be discriminated from free class-II MHCby immunolabeling, its distribution on the cell couples could not be determined.)


The Dynamics

of

CD4 in Th’APC Couples

In addition to the TcR, several accessory molecules (including CD4, LFA1, CD2) which are integral membraneproteins of the T cell appear to play important roles in most Th : APCinteractions, because monoclonal antibodies (MAb)directed to these molecules often inhibit T-cell activation (cf 69). With most CTL: target cell interactions, CD8is involved rather than CD4. The amino acid sequences, and putative structures in the membrane, of CD4and CD8are known(70). CD4appears in addition to + T lymphocytes the receptor for HIVattachment to, and infection of, CD4 (71). The precise mechanismsby which the accessory molecules participate in the interaction of effector T cells and their congenercells are not clear. Because CD4is closely associated with class-II MHC-restriction, and CD8 with class-I MHC-restriction, it has been suggested (cf 72) that the CD4 molecule forms a transcellular bond to a monomorphicdeterminant on class-II MHCmolecules, and similarly for CD8with class-I MHC.Evidence for such a transcellular interaction between CD4and class-II MHC has been presented (73). On the other hand, other evidence has suggested that CD4interacts with the TcR, or the TcR/T3 complex, in Th cell membranes(74-80). Furthermore, it is not clear what functions would served by either the putative transcellular binding of CD4to class-II MHC, or the suggested intracellular binding of CD4to TcR. In cell couples formed with D10Th cells and Con-pulsed LKcells, it was found (68) that CD4was concentrated, along with the TcR, on the Th cell membranewhere it was in contact with the APC. On the other hand, in nonspecific cell couples formed by the same D10 and LK cells (the latter, however, presenting the nonspecific antigen Ova instead of Con) CD4was not concentrated in the cell : cell contact region. If in the specific couple, CD4had been clustered into the contact region by a simple process of mutual capping with class-II MHCmolecules on the APC, then the CD4clustering would have been expected to occur equally in the nonspecific couple, whose CD4and class-II MHCwere identical to the specific case. The results rather suggest that the concentration of CD4in the contact region of spec.ific Th : APCcouples is somehowdependent on the specific engagementof the TcR.

Annual Reviews CELLBIOLOGYOF CELLCOUPLES

325


The Dynamics of TcR and CD4on Th Cells Treated with Monoclonal Abs to TcR The clustering of TcR and CD4into the contact regions of specific but not nonspecific Th:APCcouples raises several questions: (a) Is such TcRclustering critical in order to generate one or more signals that are transmitted into the Th cell; (b) What is the molecular mechanism responsible for CD4co-clustering with the TcRinto the contact region; and (c) Whatfunctions maybe served by such CD4clustering? To examine these questions, we took advantage of the findings of Janeway and colleagues (81, 82) whoraised a battery of MAbagainst the TcRon D10cells and madea detailed study of their properties and effects on D10cells. All anti-TcR MAbdid activate D10cells ifa cross-linking secondary antibody was additionally employed (81). However, one particular MAb,3D3, directed against a clonotypic determinant on the D10 TcR, activated D10 cells by itself (82), without requiring a secondaryantibody. If TcRclustering on the Th cell were critical for activation, then uniquely amonganti-TcR MAb, 3D3 added alone should induce TcR clustering (capping) on D10 cells. This turned out to be the case (83). Wenext asked whether the 3D3induced clustering of the D10 TcR had any effect on CD4, and found by double immunofluorescence experiments that CD4was co-clustered with the TcR (83), although CD4and TcR molecules are normally independent of one another in Th cell membranes.J. M. Rojo, K. M. Saizawa, and C. A. Janeway, Jr. (personal communication) have obtained similar immunofluorescence results, and a numberof other types of experiments have also suggested an interaction of CD4with the TcR under particular conditions (74-80). Our interpretation of these results with both MAbsand APCsis as follows: (a) The strong correlation between specific clustering of the TcRand Th cell activation, as observed upon stimulation of Th both with soluble antibodies and with specific APC,suggests that appropriate TcRclustering may be required to generate one or more signals for activation of the cell. Manyother types of receptor-mediated cellular stimulations are also thought to require receptor clustering induced by binding of the specific ligand (cf 84-86). That this might be true of T-cell stimulation, therefore, is not unprecedented. TcRclustering has also been proposed to be critical for the activation of antigen-specific B-cell differentiation by antigen-bridged Th cells (86) and also to induce exocytosis from CTL

(88). (b) This clustering of the TcR may be promoted by a conformational change in the TcR, or the TcR-T3 complex, that is proposed generally

Annual Reviews


326

KUPFER & SINGER

to accompany TcR binding to its specific Ag/class-II MHCligand on an APC, or to certain clonotypic MAbsuch as 3D3 (82). In the case of Th:APC couples, the clustering of the TcR is further promoted by a mutual capping process of TcR with its Ag/class-II MHC (Figure 3). (c) This ligand-induced TcR conformational change and promotion TcRclustering results in the syn-capping of CD4with the TcR-T3clusters. That syn-capping is involved is indicated by the facts that anti-TcR MAbs directed to either clonotypic or nonclonotypic determinants on the D10 TcR, when allowed to cap only upon addition of secondary antibodies, produce no significant degree of co-clustering of CD4with the TcRcaps; and capping of CD4on D10 cells produces no co-clustering of the TcR (68, 83). (d) If CD4molecules have only a weakaffinity for class-II MHC(73), the CD4that is co-clustered by syn-capping with the TcRinto the specific Th:APCcell:cell contact region may now be concentrated enough to bind to class-II MHCmolecules on the APC. Such transcellular binding wouldcontribute to cell : cell adhesion, and maybe required under certain conditions to transmit another signal between the two cells (78, 89). Such conjectures, therefore, provide a basis to explain how CD4molecules mayexercise a class-II MHC-mediated critical function in helper T cell interactions. Parallel considerations mayapply to CD8function in cytotoxic T cell interactions (90, 91). The Dynamics

of LFA-1 in Th ."

APC Couples

LFA-1 is a monomorphic member of the superfamily of membrane proteins called integrins (92, 93). The molecule of each memberof this family consists of two chains, ~ and /~. Someof the integrin molecules have binding sites on the externally exposed domains of their c~ chains for componentsof the extracellular matrix, such as fibronectin (93). On the cytoplasmic surface of the membrane,integrin molecules are associated with the actin microfilaments of the cytoskeleton (94, 95). In cells such fibroblasts, integrin molecules therefore serve to mediate transmembrane linkages betweenthe extracellular matrix and the cytoskeleton. In lymphocytes, whichgenerally lack an extensive extracellular matrix, there is strong evidence that LFA-1molecules are involved in cell:cell adhesions or in attachments to solid substrata (5). Furthermore, in cell:cell adhesion, there is evidence (5) that LFA-Ion the T cell maybind transcellularly a ligand on the other cell, called I-CAM-1. Wediscuss next our recent studies (A. Kupfer, D. R. Wegmann,S. J. Singer, in preparation) on immunofluorescencemicroscopy of LFA-I distribution on the Th cells in cell couples formed between DNP-Ova-pulsed

Annual Reviews



327

A20-HLcells, as APC,and the cloned Th cell line D36, introduced in an earlier section. Lacking an appropriate antibody specific for the TcR on D36 cells, we did not simultaneously immunolabel the TcR in these experiments. LFA-1, and talin, were found to becomeconcentrated into the specific cell:cell contact regions over a wide range of concentrations of added DNP-Ova,including concentrations lower than those required to activate the Th cells or to induce a reorientation of its MTOC/GA. Neither LFA-1nor talin was collected, however,into the cell :cell contact regions if no DNP-Ovawas presented. Therefore, the specific engagement of the TcRon the Th cells was required to induce the clustering of LFA-1/talin into the cell : cell contact region, as well as to induce the reorientation of the MTOC of the Th cell. However, clustering of LFA-1and reorientation of the MTOC were not tightly coupled. On the other hand, over the entire range of concentrations of DNP-Ova,LFA-1and talin were always found to be co-distributed into the contact region, and to about the same extent, as judged by the immunofluorescentintensity. Other unpublished studies showedthat LFA-1is also clustered, along with talin, into the cell:cell contact regions of other specific Th:APCcouples, and of productive CTL-target cell and NK-target cell couples, but not in the corresponding nonspecific couples. The co-clustering of LFA-1/talin in these couples is +2. independent of extracellular Ca Clustering of LFA-1into the contact regions of specific effector T cell: congener cell couples, as with CD4, is a striking and unusual result, because LFA-1and TcR molecules are not normally associated in T-cell membranes. This clustering is clearly a reflection of the importance of LFA-I in these cellular interactions. However,it does not appear that LFA-1is involved in generating a simple cell:cell adhesion; the fact that LFA-1does not cluster into nonspecific cell:cell contacts argues against an adhesion resulting from a simple mutual capping of LFA-1 with its putative transcellular ligand I-CAM-1.This is because the LFA-1and ICAM-1 were identical in manyof the specific and nonspecific couples (as with DNP-Ovapresented or not presented, respectively, on the A20-HL cells). On the other hand, we have no evidence that LFA-1is syn-capped with TcR clusters, as we have inferred happens with CD4. Whenthe antiTcR MAb3D3 alone was used to cluster the TcR on D10 Th cells and to induce a co-clustering (syn-capping) of CD4with the TcR clusters (82) (discussed above), LFA-1was generally found to remain uniformly distributed on the D10cell surface. At present, we can only speculate about the mechanismfor the clustering of LFA-Iin specific T cell: congener cell couples, but for reasons discussed next, it maybe related to an induced interaction between the integral membraneprotein LFA-1and the cytoskeletal protein talin in the T cell.

Annual Reviews

328

KUPFER & SINGER


The Association

of Talin

and LFA-1

Amonga large number of cytoskeletal proteins talin uniquely becomes concentrated underneath effector T cell membraneswhere they contact their specific congener cells; this fact deserves an explanation. Talin is a 215-kd protein originally isolated from smooth muscle (41); talin or its isoforms, however,is present in a wide range of different cell types (44). At sites where cells form specialized regions of adhesion to other cells (such as the dense plaques that interconnect smooth muscle cells), or to substratum (such as the focal adhesions that attach cultured fibroblasts to a substratum), talin is also present (44) along with several other cytoskeletal proteins. These are specialized regions where actin microfilaments are knownto be attached to the membrane. In the cell membranesat these same sites, one or another memberof the integrin family of integral proteins (which includes LFA-1)is also often found (93). The possibility therefore arises that talin and LFA-1might interact with one another wherethe latter molecule protrudes from the cytoplasmic surface of a T-cell membrane.In vitro studies (96) have provided evidence for an interaction between talin and chicken smooth muscle integrin, but it appeared to be quite weak. In order to investigate this question in vivo, the antibody-induced capping of LFA-I was carried out on D36 Th cells. By double immunofluorescence we examined whether talin became collected with the LFA-1 caps, as wouldbe expected if the two molecules were associated (A. Kupfer, P. Burn, S. J. Singer, in preparation). Similar experiments have been published on possible integrin/talin associations in chicken lymphocytes (97). Whenthis experiment was carried out on the D36cells, talin was found to remain uniformly distributed on cells that exhibited LFA-1caps. However,if the cells were treated with phorbol myristoyl acetate (PMA), an activator of protein kinase C (98), just prior to the antibody-induced capping of LFA-1,talin was then found to be precisely co-distributed with the LFA-1caps. This redistribution of talin did not occur upon capping membraneproteins unrelated to LFA-I, such as CD4on D36 cells, in the presence of PMA. These experiments indicate that in mouseT cells, LFA-1and talin are not normally associated with one another; when, and only when, an appropriate signal is received by the cell do the two proteins become, directly or indirectly, linked. That PMA delivers such a signal suggests that protein kinase C activation is involved, and that phosphorylation of one or more proteins mediates the LFA-l/talin association, but this has not been established. These findings provide someinsight into the results previously discussed concerning the coordinate concentration of talin under the membrane,and

Annual Reviews


CELL

BIOLOGY

OF CELL

COUPLES

329

LFA-1within the membrane,of an effector T cell where it is in contact with a specific congener cell. The transcellular engagementof the TcRon CTLor on Th cells by a specific ligand on the target cell or APC,respectively, maylead to one or more signals being conveyed into the effector T cell, one consequence of which is the association of talin with LFA1. What factors then induce the recruitment of LFA-1/talin into the contact region are not clear, except that syn-capping of LFA-1with the TcR, which was invoked in the case of CD4, does not appear to be involved. It is not ruled out, however, that the signal that causes LFA-1to become associated with talin also alters the LFA-1molecule so that its affinity for its putative transcellular ligand (I-CAM-I)increases enough to induce their mutual capping into the specific cell:cell contact region. This signal-induced mutual capping of LFA-I and I-CAM-1could make a major and crucial contribution to the overall specific cell : cell adhesion, and thereby account for the important role played by LFA-1 in T-cell functions (5). Because a closely similar redistribution of LFA-1/talin as well as the reorientation of the MTOC/GA accompanies productive NK: target cell binding, the unknownreceptor on NKcells must have signal-induction properties closely similar to those of the TcR of CTLand Th cells as described above.

CONCLUSIONS AND SUGGESTIONS Upto this point we have discussed our experiments largely in the sequence in which they were carried out, in order to indicate the logic behind what was done. In this section, however, the results are summarizedin the sequenceof their probable occurrenceuponinitiation of a specific cell : cell interaction. This requires that cell surface phenomenon be considered first, before their intracellular sequellae. Mostof our studies involving the cell surface were carried out with Th : APCinteractions; it is likely, however, that closely parallel cell surface effects accompany CTLand NKcell interactions with their specific targets. A general commentis that membranedynamics play a key role in the cell : cell interactions studied, being critically involved in the formation of specific cell: cell adhesions and most probably also in the transmission of signals into cells. The global diffusion of integral proteins in lymphocyte membranes, a phenomenonfirst recognized by immunologists nearly 20 years ago, allows these proteins to become clustered and to interact specifically with other molecules within or between cell surfaces. The occurrence of mutual capping and syn-capping phenomena needs to be appreciated, because they provide plausible mechanismsto explain a number of experimental facts that have been difficult to understand.

Annual Reviews 330

KUPFER & SINGER


At Least Two Distinct Signals are Transmitted T Cell in Th : APC Interactions

to the

In the experiments described above (A. Kupfer, D. R. Wegmann,S. J. Singer, in preparation) with cell couples formed between D36Th cells and Ova peptide-presenting A20-HLcells as APC,clear evidence was obtained for two separable signals transmitted to the Th cell, one at very low doses of the presenting antigen, the other at higher doses. At low doses we observed a specific coordinate redistribution of talin under the membrane, and of LFA-1in the membrane,on the Th cell where it was bound to the APC.Only at higher antigen doses, however, were both a reorientation of the MTOC/GA and activation to proliferate also conferred on the Th cell, along with the redistribution of talin/LFA-1. Fromthese and other results discussed and cited above, we propose the following plausible scenario of membraneevents, although many of the details must still be considered speculative. At low antigen doses, only a limited number of Ag/class-II MHCligands are presented on the A20-HL APC,and these engage an equivalently small number of the TcRmolecules on the D36 Th cell. The low concentration of Ag/class-II MHCligand molecules mayor maynot be sufficient to induce a limited degree of mutual capping with TcR. The small degree of TcR : Ag/class-II MHC binding is sufficient, however,to transmit a first signal (possibly involving a ligandinduced conformational change in the TcR) into the D36 cell; this PMAlike signal induces talin to become linked to LFA-1 and must also be responsible for their joint and massive collection into the contact region forming between the two cells. For example, if the first signal, although involving only a small numberof TcRmolecules, caused an alteration of most LFA-1molecules so that they could now effectively bind not only intracellularly to talin but also transcellularly to their ligands (I-CAM-l) on the APC, the mutual capping of LFA-1 and I-CAM-1 could provide the main source of a specific cell : cell adhesion. No mutual capping of LFA-1and I-CAM-1would occur in the nonspecific case, because no first signal via a small numberof ligand-bound TcRwouldarise. In this scheme, therefore, LFA-I would not contribute to nonspecific adhesion. Such a crucial involvement of LFA-1in the formation of signal-induced specific adhesion could explain whyLFA-1is required for productive T cell interactions (5). This first signal wouldbe independent of extracellular +2 (cf Figure 1). However,this first signal alone is insufficient to activate the Th cell to proliferate or to induce its MTOC/GA to reorient. These phenomena require at least another specific signal, which is only realized at larger antigen doses, and which is Ca÷2 dependent. As more antigen is presented by the APC, a larger number of TcR molecules is engaged, above a

Annual Reviews


CELLBIOLOGYOF CELLCOUPLES

331

threshold numberin order for the following effects to occur extensively. The proposed conformational change induced in the TcR by ligand binding, it is suggested, promotesTcRclustering in the Th cell membrane. The TcR would then undergo an extensive mutual capping with its Ag/class-II MHCligand into the contact regions between the two cells, where LFA-1had already been mutually capped via the first signal. These clustering events are considered to be crucial for transmission of a second signal into the Th cell, perhaps functioning directly or indirectly to open a Ca+2 channel in the Th membrane. The CD4that was syn-capped with the TcR would thereby also be collected into the contact region, and perhaps in that concentrated location, CD4could then, and only then, form transcellular bonds to a domain of the class-II MHC molecule. Such transcellular CD4: class-II MHC bonds might be required for still another signal to be transmitted between the Th and its bound APC;this could explain the importance of CD4in manyTh : APCinteractions (and correspondingly CDgin many CTLinteractions). The proposed mutual cocapping of TcR : Ag/class-II MHCand of CD4: class-II MHCpairs into the cell : cell contact region wouldcontribute additionally to the cell : cell adhesion that was primarily due to the first-signal-induced mutual capping of LFA-1 and I-CAM-1. Oneor both of these additional specific signals might then be responsible for inducing the reorientation of the MTOC/GA as well as for turning on a whole cascade of processes leading to activation and proliferation of Th cells. The Reorientation

of the

MTOC/GA

All the evidencepoints to the rapid polarization of the MTOC/GA inside the effector T cell as a prerequisite to the subsequent killing (CTL) stimulation (Th) of the congener cell boundto it. In general, the functions served by such a reorientation wouldbe to direct both the insertion of new membranemass into the T-cell surface and the exocytosis of secretory componentsfrom the T cell within the confined intercellular space generated after cell : cell adhesion.In the case of CTL: target cell interactions, secretion of one or more cytotoxic componentsinto the intercellular space is an attractive modelfor cell lysis, but the unidirectional resistance of the CTLto such toxic componentshas yet to be satisfactorily explained. Talin accumulation under the contacting effector cell membrane,as well as all the integral protein componentscollected into that membraneregion~ may play a role in such resistance. In the case of Th : APCcouples, the directed secretion from the Th cell of a numberof B-cell growth and differentiation factors (57, 99-101) into the intercellular space could act to stimulate only the APCboundto the Th. If the cell : cell adhesion was sufficiently stable and prolonged, sequential delivery of such secretory factors, or the sequen-

Annual Reviews


332

KUPFER & SINGER

tially stimulated appearance of receptors for these factors on the APC, could allow several stages in the differentiation pathwayof the B cell to occur during a single Th : APCencounter. The potentialities presented by a directed insertion of new membrane mass derived from the GAinto the effector T cell membraneat the cell contact region should not be ignored. Newly synthesized or recycled lymphokinereceptors, or other functionally important integral proteins, or proton pumpspresent in GA-derivedvesicle membranes(8), if inserted into the confined cell : cell contact area, could have important consequences for the cellular interaction. ACKNOWLEDGMENTS

Weare grateful to our collaborators in several of the cellular immunological studies discussed in this review: Drs. G. Dennert, C. A. Janeway, Jr., S. L. Swain, and D. R. Wegmann.The outstanding technical assistance of Mrs. Hannah Kupfer, and the invaluable help and advice of Mrs. Margie Adams and Dr. Anne Dutton, are acknowledged. Weare also particularly indebted to Dr. N. Hozumifor the generous gift of the A20H1cell line, and to Drs. M. Bevan, K. Osami, and J. Kayefor gifts of cell lines and reagents. Our work reported here was supported in part by N.I.H. grants AI-23764 to Abraham Kupfer, and AI-06659 and GM15971to S. J. Singer. S. J. Singer is an AmericanCancer Society Research Professor. NOTEADDEDIN PROOF In connection with the possibility that was raised on p. 331 that the engagement of CD4might invoke a separate signal into the APC-bound Th cell, it has recently been shown (102, 103) that CD4molecules (and CD8on CTL)are physically complexedto the lymphocyte-specific protein ~ck. tyrosine kinase, p56 Literature Cited 1. Adkins, B., Mfiller, C., Okada, C. Y., Reichert, R. A., Weissmann, I. L., Spangrude, G. J. 1987. Early events in T-cell maturation. Ann. Rev. Immunol. 5:325-65 2. Boehmer, von H. 1988. The developmental biology of T lymphocytes. Ann. Rev. Immunol. 6:309-26 3. Sitkovsky, M. V. 1988. Mechanistic, functional and immunopharmacological implications of biochemical studies of antigen receptor-triggered cytolytic T-lymphocytes activation, lmmunol. Rev. 103:127-60 4. Young; Y. D.-E., Liu, C.-C., Persechini. P. M.. Cohn. Z. A. 1988. Per-

forin-dependent and -independent pathways of cytotoxicity mediated by lymphocytes. ImmunoLRev. 103: 161202 Springer, T. A., Dustin, M. L., Kishimoto, T. K., Marlin, S. D. 1987. The lymphocyte function-associated LFA-1, CD2, and LFA-3 molecules: cell adhesion receptors of the immunesystem. Ann. Rev. Imrnunol. 5:223-52 Janeway, C. A. Jr., Carding, S., Jones, B., Murray, J., Portoles, P., Rasmussen, R., Rojo, J., Saizawa, K., West, + T cells: J., Bottomly, K. 1988. CD4 specificity and function. Immunol. Rev. I01 : 39-80

Annual Reviews


CELL BIOLOGY OF CELL COUPLES 7, Vitetta, E, S., Bossie, A., Botran, R. F., Myers, C. D., Oliver, K. G., Sanders, V. M., Stevens, T. L. 1987. Interaction and activation of antigen-specific T and B cells, lmmunol. Rev. 99:193-240 8. Singer, S. J., Kupfer, A, 1986. The directed migration of eukaryotic cells. Ann. Rev. Cell Biol. 2:337-65 9. Kupfer, A., Kronebusch, P. J., Rose, J. K., Singer, S. J. 1987. Acritical role for the polarization of membranerecycling in cell motility. Cell Motil. Cytoskel. 8:182-89 10. Kupfer, A., Louvard, D., Singer, S. J. 1982. The polarization of the Golgi apparatus and the mi¢rotubule-organizing center in cultured fibrobtasts at the edge of an experimental wound. Proc. Natl. Acad. Sci. USA79:2603-7 11. Nemere, I., Kupfer, A., Singer, S. J. 1985. Reorientation of the Golgi apparatus and the microtubule-organizing center inside macrophages subjected to a chemotactic gradient. Cell Motil. 5:17-29 12. Farquhar, M. G., Palade, G. E. 1981. The Golgi apparatus (complex)-(19541981)-fromartifact to center stage. J. Cell Biol. 91: 77s-103s 13. Bergmann, J. E., Kupfer, A., Singer, S. J. 1983, Membraneinsertion at the leading edge of motile fibroblasts. Proc. Natl. Acad. Sci. USA80:1367-71 14. Brinkley, B. R. 1985. Microtubule organizing centers. Ann. Rev. Cell Biol. 1:197-224 15. Podack, E. R., Dennert, G. 1983. Cellmediated cytolysis: assembly of two types of tubules with putative cytolytic functions by cloned natural killer cells. Nature 302:442-45 16. Kupfer, A., Dennert, G., Singer, S. J. 1983. Polarization of the Golgi apparatus and the microtubule-organizing center within cloned natural killer cells bound to their targets. Proc. Natl. Acad. Sci. USA80:722~1~28 17. Carp6n, O., Virtanen, I., Saksela, E. 1982. Ultrastructure of humannatural killer cells: Natureof the cytolytic contacts in relation to cellular secretion. J, Immunol. 128:2691-97 18. Carp~n, O. 1987. The role of microtubules in humannatural killer cellmediated cytocoxicity. Cell. Immunol. 106:376-86 19. Kupfer, A., Dennert, G. 1984. Reorientation of the microtubule-organizing center and the Golgi apparatus in cloned cytotoxic lymphocytestriggered by binding to lysable target cells. J. Immunol. 133:2762-66 20. Bykovskja, S. N., Rytenko, A. N., Rauschenbach, M. O., Bykovsky, A.

333

F. 1978. Ultrastructural alteration of cytolytic T lymphocytesfollowing their interaction with target cells. I. Hypertrophy and change of orientation of the Golgi apparatus. Cell. Immunol. 40: 164-74 21. Geiger, B., Rosen, D., Berke, G. 1982. Spatial relationships of microtubuleorganizing centers and the contact area of cytotoxic T lymphocytes and target cells. J. Cell Biol. 95:137-43 22. Goldstein, P. A., Smith, E. T. 1976. The lethal hit stage of mouse T and non-T celt-mediated cytolysis: differences in cation requirements and characterization of an analytical "cation pulse" method. Eur. J. Immunol. 6: 31-37 23. Plaut, M., Bubbers, J. E., Henney, C. S. 1976. Studies on the mechanism of lymphocyte-mediatedcytolysis. VII. Twostages in the T cell-mediated lytic cycle with distinct cation require~nents. J. Immunol. 116:150-55 24. Martz, E. 1977. Mechanismof specific tumor-cell lysis alloimmune T-lymphocytes: resolution and characterization of discrete steps in the cellular interaction. Contemp. Top. Immunobiol. 7: 301 61 25. Kupfer, A., Dennert, G., Singer, S. J. 1985. The reorientation of the Golgi apparatus and the microtubule-organizing center in the cytotoxiceffector cell is a prerequisite in the tysis of bound target cells. J. Mol. Cell Immunol.2: 37~49 26. Gray, L. S., Gnarra, J. R., Engelhard, V. H. 1987. Demonstration of a calcium influx in cytolytic T lymphocytes in response to target cell binding. J. Immunol. 138:63-69 27. Zagury, D., Bernard, J., Thiernesse, N., Feldmann, M., Berke, G. 1975. Isolation of individual functionally reactive cytotoxic T-lymphocytes: conjugation, killing and recycling at the single cell-level. Eur. J. Immunol. 5: 818-22 28. Rothstein, T. L., Mage, M., Jones, G., McHugh,L. L. 1978, Cytotoxic T-lymphocytes sequential killing of immobilized allogeneic tumor target cells measured by time-lapse micro-cinematography. J. Immunol. 121:1652-56 29. Zagury, D., Bernard, J., Jeannesson, P., Thiernesse, N., Cerottini, J.-C. 1979. Studies on the mechanism of T-cellmediatedlysis at the single effector cell level. I. Kinetic analysis of lethal hits and target cell lysis in multicellular conjugates. J. Immunol. 123:1604-9 30. Yannelli, J. R., Sullivan, J. A., Mandell, G. L., Engelhard, V. H. 1986.

Annual Reviews


334

KUPFER & SINGER

Reorientation and fusion of cytotoxic T lymphocytes granules after interaction with target cells as determined by high resolution cinemicrography. J. Immunol. 136:377-82 31. Kuppers, R. C., Henney, C. S. 1977. Studies on the mechanism of lymphocyte-mediated cytolysis. IX. Relationships between antigen recognition and lytic expression in killer T cells. J. Immunol. 118:71-76 32. Kupfer, A., Singer, S. J., Dennert, G. 1986. On the mechanism of unidirectional killing in mixtures of two cytotoxic lymphocytes. Unidirectional polarization of cytoplasmic organelles and the membrane-associated cytoskeleton in the effector cell. J. Exp. Med. 163:489-98 33. Ostergaard, H., Clark, W. R. 1987. The role of Ca+z in activation of mature cytotoxic T lymphocytes for lysis. J. Immunol. 139:3573-79 34. Ostergaard, H., Gorman, K., Clark, W. R. 1988. Cloned cytotoxic T lymphocytestarget cells fail to induce early activation events in effector cytotoxic T lymphocytes. Cell. Immunol. 114:18847 35. Blakely, A., Gorman, K., Ostergaard, H., Svoboda, K., Liu, C.-C., Young, J. D.-E., Clark, W. R. 1987. Resistance of cloned cytotoxic T lymphocytes to cell-mediated cytotoxicity. J. Exp. Med. 166:1070-83 36. Kranz, D. M., Eisen, H. N. 1987. Resistance of cytotoxic T lymphocytes to lysis by a clone of cytotoxic T lymphocytes. Proe. Natl. Acad. Sci. USA 84:3375 79 37. Skinner, M., Marbrook, J. 1987. The most efficient cytotoxic T lymphocytes are the least susceptible to lysis. J. Immunol. 139:985-87 38. Lazarides, E., Burridge, K. 1975. Actinin: immunofluorescent localization of a muscle structural protein in non-muscle cells. Cell 6:289-98 39. Geiger, B. 1979. A 130Kprotein from chicken gizzard: its localization at the termini of microfilament bundles in cultured chicken cells. Cell 18:193-205 40. Geiger, B., Tokuyasu, K. T., Dutton, A. H., Singer, S. J. 1980. Vinculin, an intracellular protein localized at specialized sites where microfilament bundles terminate at cell membranes. Proc. Natl. Acad. Sci. USA77: 412731 41. Burridge, K., Connell, L. J. 1983. A new protein of adhesion plaques and ruffling membranes.J. Cell Biol. 97: 359 67

42. Maher, P., Singer, S. J. 1983. A 200-kd protein isolated from the fascia adherens membrane domains of chicken cardiac muscle cells is detected immunologically in fibroblast focal adhesions. Cell Motil. 3:419-29 43. Bretscher, A., Weber, K. 1980. Fimbrin, a new microfilament-associated protein present in microvilli and other cell surface structures. J. Cell Biol. 86:335~40 44. Burridge, K. 1987. Substrate adhesions in normal and transformed fibroblasts: organization and regulation of cytoskeletal, membrane,and extracellular matrix components at focal contacts. Cancer Rev. 4:18-78 45. Ryscr, J.-E., Rungger-Br/indle, E., Chaponnier, C., Gabbiani, G., Vassalli, P. 1982. The area of attachment of cytotoxic T lymphocytes to their target cells showshigh motility and polarization of actin but not myosin. J. Immunol. 128:1159-62 46. Carp6n, O., Virtanen, I., Lehto, V.-P., Saksela, E. 1983. Polarization of NK cell cytoskeleton upon conjugation with sensitive target cells. J. lmmunol. 131:2695-98 47. Rupp, F., Brecher, J., Giedlin, M. A., Mosmann, T., Zinkernagel, R. M., Hengartner, H., Joho, R. H. 1987. Tcell antigen receptors with identical variable regions but different diversity and joining region gene segments have distinct specificities but cross-reactive idiotypes. Proc. Natl. Acad. Sci. USA 84:219-22 48. Herberman, R. B., Reynolds, C. W., Ortaldo, J. R. 1986. Mechanism of cytotoxicity by natural killer cells. Ann. Rev. Immunol. 4:651 80 49. Berke, G. 1983. Cytotoxic T-lymphocytes. Howdo they function? Immunol. Rev. 75:5~42 50. Srendi, B., Schwartz, R. H. 1981. Antigen-specific, proliferating T lymphocyte clones. Methodology,specificity, MHCrestriction and alloreactlvity. Immunol. Rev. 54:184-223 51. Kupfer, A., Swain, S. L., Janeway, C. A. Jr., Singer, S. J. 1986.Thespecific direct interaction of helper T cells and antigen-presenting B cells. Proc. Natl. Aead. Sci. USA 83:608~83 52. Kupfer, A., Swain, S. L., Singer, S. J. 1987. The specific direct interaction of helper T cells and antigen-presenting B cells. II. Reorientation of the microtubule organizing center and reorganization of the membrane-associated cytoskeleton inside the bound helper T cells. J. Exp. Med. 165:1565-80

Annual Reviews


CELL BIOLOGY OF CELL COUPLES 53. Marrack, P., Skidmore, B., Kappler, J. W. 1983. Bindingof antigen-specific, H-2-restricted T cell hybridomas to antigen-pulsed adherent cell monolayers. J. Immunol. 130:2088-92 54. Sanders, V. M., Snyder, J. M., Uhr, J. W., Vitetta, E. S. 1986. Characterization of the physical interaction between antigen-specific B and T cells. J. Immunol. 137:2395-2404 55. Watanabe, M., Wegmann,D. R., Ochi, A., Hozumi,N. 1986. Antigen presentation by a B-cell line transfected with cloned immunoglobulin heavy- and light-chain genes specific for a defined hapten. Proc. Natl. Acad. ScL USA83: 5247-51 56. Berke, G., Rosen, D. 1987. Are lytic granules, and perforin 1 thereof, involvedin lysis inducedby in vivo primed, peritoneal exudate CTL? Transpl. Proc. 19:412-16 57. Coffman, R. L., Seymour, B. W. P., Lebman, D. A., Hiraki, D. D., Christiansen, J. A., Shrader, B., Cherwinski, H. M., Savelkoul, H. F. J., Finkelman, F. D., Bond, M. W., Mosmann,T. R. 1988. The role of helper T cell products in mouseB cell differentiation and isotype regulation, lmmunol. Rev. 102: 528 58. Takatsu, K,, Kikuchi, Y., Takahashi, T., Honjo, T., Matsumoto, M. Harada, N., Yamaguchi, N., Tominaga, A. 1987. Interleukin 5, a T-cell-derived Bcell differentiation factor also induces cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. USA 84:4234-38 59. Swain, S. L., McKenzie,D. T., Dutton, R. W,, Tonkonogy, S. L., English, M. 1988. The role oflL4 and IL5: characterization of a distinct helper T cell subset that makes IL4 and IL5 (Th2) and requires priming before induction of lymphokine secretion. Immunol. Rev. 102:77-105 60. Singer, S. J. 1976. The fluid mosaic model of membrane structure: some applications to ligand-receptor and cell-cell interactions. In Surface Membrane Receptors, Interface between Cells and their Environment, ed. R. A. Bradshaw, W. A. Frazier, R. C. Merrell, D. I. Gottlieb, R. A. HogueAngeletti, pp. 1-24. NewYork: Plenum 6l. Weis, R. M., Balakristman, K., Smith, B. A., McConnell, H. M. 1982. Stimulation of fluorescence in a small contact region between rat basophil leukemia cells and planar lipid membranetargets by coherent evanescent radiation. J. Biol. Chem. 257:6440-45 62. McCloskey, M. A., Poo, M.-M. 1986.

335

Contact-induced redistribution of specific membranecomponents: local accumulation and development of adhesion. J. Cell Biol. 102:2185-96 63. Bell, G. I., Dembo, M., Bongrand, P. 1984. Cell adhesion. Competition between nonspecific repulsion and specific bonding. Biophys. J. 45: 105164 64. Singer, S. J., Kupfer, A. 1988. Membrane dynamicsin cell-cell interactions, with application to immunology. In The T Cell Receptor, ed. M. Davis, J. Kappler, pp. 361-76. NewYork: Liss K. A., 65. Meuer, S. C., Fitzgerald, Hussey, R. E., Hodgdon,J. C., Schlossman, S. F., Reinherz, E. L. 1983. Clonotypic structures involved in antigen-specific humanT cell function. J. Exp. Med. 157:705-19 66. Geiger, B., Rosenthal, K. L., Klein, J., Zinkernagel, R. M., Singer, S. J. 1979. Selective and unidirectional membrane redistribution of an H-2 antigen with an antibody-clustered viral antigen: relationship to mechanisms of cytotoxic T cell interactions. Proc. Natl. Acad. Sci. USA 76:4603-7 67. Bourguignon, L. Y. W., Hyman, R., Trowbridge,I., Singer, S. J. 1978. Participation of histocompatibility antigens in capping of molecularly independent cell surface components by their specific antibodies. Proc. Natl. Acad. Sci. USA 75:2406-10 68. Kupfer, A., Singer, S. J., Janeway, C. A. Jr., Swain, S. L. 1987. Coclustering ofCD4(L3T4) molecule with the T-cell receptor is induced by specific direct interaction of helper T cells and antigen-presenting cells. Proc. Natl. Acad. Sci. USA 84:5888-92 69. Krensky, A. M., Robbins, E., Springer, T. A., Burakoff, S. J. 1984. LFA-1, LFA-2, and LFA-3 antigens are involved in CTL-target conjugation. J. ImmunoL 132:2180-82 70. Littman, D. R. 1987. The structure of the CD4 and CD8 genes. Ann. Rev. Immunol. 5:561-84 71. McDougal, J. S., Kennedy, M. S., Sligh, J. M., Cort, S. P., Mawle, A., Nicholson, J. K. 1986. Binding of HTLV/LAVto T4÷ T cells by a complex of the 110Kviral protein and the T4 molecule. Science 231:382-85 72. Swain, S. L, 1983. T cell subsets and the recognition of MHCclass. Immunol. Rev. 74:12942 73. Doyle, C., Strominger, J. L. 1987. Interaction between CD4and class II MHCmolecules mediates cell adhesion. Nature 330:256-59

Annual Reviews


336

KUPFER & SINGER

74. Anderson, P., Blue, M.-L., Schlossman,S. F. 1988. Evidencefor a specific association between CD4and approximately 5% of the CD3: T cell receptor complexes on Helper T lymphocytes. J. lmmunol. 140:1732-37 75. Weyand, C. M., Goronzy, M. J., Fathman, C. G. 1987. Modulation of CD4 by antigenic activation. J. Immunol. 138:1351-54 76. O’Neil, H. C., McGrath,M. S., Allison, J. P., Weissman,I. L. 1987. A subset of T cell receptors associated with L3T4 molecules mediates C6VLleukemia cell binding of its cognate retrovirus. Cell 49:143-51 77. Saizawa, K., Rojo, J., Janeway, C. A. Jr. 1987. Evidencefor a physical association of CD4and the CD3:c~ : fl T-cell receptor. Nature 328:260-63 78. Rivas, A., Takada, S., Koide, J., McDevitt, G. S., Engleman,E. G. 1988. CD4molecules are associated with the antigen receptor complex on activated but not resting T cells. J. Immunol.140: 2912-18 79. Ledbetter, J. A., June, C. A., Rabinovitch, P. S., Grossmann,A., Tsu, T. T., Imboden, J. B. 1988. Signal transduction-through CD4receptors: stimulatory versus inhibitory activity is regulated by CD4proximity to the CD3/T cell receptor. Eur. J. Immunol.18: 52532 80. Owens, T., Fazekas, De St. Groth, B., Miller, J. F. A. P. 1987. Coaggregation of the T-cell receptor with CD4and other T-cell surface molecules enhances T-cell activation. Proe. Natl. Acad. Sci. USA 84:9209-13 81. Tite, J. P., Kaye, J., Saizawa, K. M., King, J., Katz, M. E., Smith, L. A., Janeway, C. A. Jr. 1986. Direct interactions between B + T lymphocytesbearing complementary receptors. J. Exp. Med. 163:189 202 82. Rojo, J. M., Janeway, C. A. Jr. 1988. The biologic activity of anti-T cell receptor V region monoclonal antibodies is determined by the epitope recognized. J. Immunol. 140:1081-88 83. Kupfer, A., Singer, S. J. 1988. Molecular dynamics in the membranes of helper T cells. Proc. Natl. Acad. Sci. USA. In press 84. McClain, D. A., Maegawa,H., Lee, J., Dull, T. J., Ulrich, A., Olefsky, J. M. 1987. A mutant insulin receptor with defective tyrosine kinase displays no biologic activity and does not undergo endocytosis. J. Biol. Chem. 262: 1466371 85. Russell, D. S., Gherzi, R., Johnson,

E. L., Chou, C.-K., Rosen, O. M. 1987. The protein-tyrosine kinase activity of the insulin receptor is necessary for insulin-mediated receptor down-regulation. J. Biol. Chem. 262: 11833M0 86. Glenney, J. R. Jr., Chen, W., Lazar, C. S., Walton, G. M., Zokas, L. M., Rosenfeld, M. G., Gill, G. N. 1988. Ligand-induced endocytosis of the EGFreceptor is blocked by mutational inactivation and by microinjection of anti-phosphotyrosine antibodies. Cell 52:675-84 87. Julius, M. H., Rammensee,H.-G., Ratcliffe, M. J. H., Lamers, M. C. Langhorne, J., K6hler, G. 1988. The molecular interactions with helper T cells whichlimit antigen-specific B cell differentiation. Eur. J. Immunol. 18: 38186 88. Takayama, H., Sitkovsky, M. V. 1987. Antigen receptor-regulated exocytosis in cytotoxic T lymphocytes. J. Exp. Med. 166:725M3 89. Gay, D., Maddon,P., Sekaly, R., Talle, M. A., Godfrey, M., Long, E., Goldstein, G., Chess, L., Axel, R., Kappler, J., Marrack, P. 1987. Functional interaction between human Tcell protein CD4and the major histocompatibility complex HLA-DRantigen. Nature 328:626-29 90. Emmrich, F., Strittmatter, U., Eichmann, K. 1986. Synergism in the activation of humanCD8T cells by crosslinking the T-cell receptor complex with the CD8differentiation antigen. Proc. Natl. Acad. Sci. USA 83: 82988302 91. Goldstein, S. A. N., Mescher, M. F. 1987. Cytotoxic T cell activation by class I protein on cell-size artificial membranes: antigen density and LYT2/3 function. J. Immunol. 138: 203443 92. Hynes, R. O. 1987. Integrins: a family of cell surface receptors. Cell 48: 54954 93. Buck, C. A., Horwitz, A. F. 1987. Cell surface receptors for extracellular matrix molecules. Ann. Rev. Cell Biol. 3: 179-205 94. Rogalski, A. A., Singer, S. J. 1985. An integral glycoprotein associated with membrane attachment sites of actin microfilaments. J. Cell Biol. 10l: 785801 95. Maher, P. A., Singer, S. J. 1988. An integral membrane protein antigen associated with the membraneattachment sites of actin microfilaments is identified as an integrin/~-chain. Mol. Cell. Biol. 8:564-70

Annual Reviews


CELL BIOLOGY OF CELL COUPLES 96. Horwitz, A., Duggan, K., Buck, C., Beckerle, M. C., Burridge, K. 1986. Interaction of plasma membranefibronectin receptor with talin--a transmembrane linkage. Nature 320: 53133 97. Burn, P., Kupfer, A., Singer, S. J. 1988. Dynamicmembrane-cytoskeletal interactions: specific association of integrin and talin arises in vivo after phorhol ester treatment of peripheral blood lymphocytes. Proc. Natl. Acad. Sci. USA 85:497-501 98. Nishizuka, Y. 1986. Studies and perspectives of protein kinase C. Science 233:305-12 99. DeFranco, A. L. 1987. Molecular aspects of B-lymphocyteactivation. Ann. Rev. Cell Biol. 3:143-78 100, Sideras, P., Noma,T., Honjo, T. 1988.

337

Structure and function of interleukins 4 and 5. Immunol. Rev. 102:189-212 101. Cambier, J. C., Ransom, J. T. 1987. Molecular mechanisms of transmembrane signalling in B lymphocytes.Ann. Rev. Immunol. 5:175-99 J. M., 102. Rudd, C. E., Trevillyan, Dasgupta, J. D., Wong,L. L., Schlossman, S. F. 1988. The CD~, receptor is complexedin detergent lysates to a protein-tyrosine kinase (pp58) from human T lymphocytes. Proc. Natl. Acad. Sci. USA 85:5190-94 103. Veillette, A., Bookman,M. A., Horak, E. M., Bolen, J. B. 1988. The CD4and CD8T cell surface antigens are associated with the internal ~ck. membranetyrosine-protein kinase p56 Cell 55:3018



Annual Reviews www.annualreviews.org/aronline ,,Inn. Rev. ImmunoL1989. 7:339-69 Copyright © 1989 by Annual Reviews Inc. All rights reserved


THE LEUKOCYTE COMMON ANTIGEN FAMILY Matthew

L. Thomas

Department of Pathology, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, Missouri 63110 INTRODUCTION The leukocyte-commonantigen (L-CA) family is a group of high molecular weight glycoproteins uniquely expressed on the surface of all leukocytes and their hemopoietic progenitors (1-14). Membersof this family differ both protein sequence and carbohydrate structures and are expressed by leukocyte populations in specific patterns (15-27). An example of the differential expression is shownfor the rat L-CAfamily in Figure 1 (19). Thymocytesexpress the lowest apparent molecular weight form of 180 kd; B lymphocytes express the highest form, 220 kd; and T lymphocytes express multiple forms. Differences also exist between T-cell subsets (Figure 1C); CD8T cells (Tc/~) express the higher molecular weight forms more abundantly than do CD4T cells (TH). The cell-type-specific patterns of expression are conserved throughout mammalianevolution (1, 2, 9-11, 19, 28), and there appear to be similar patterns of expression in chicken lymphocytes(29). L-CAis referred to in the literature by different names, including T200 (30), B220 for the B cell form (12), the mouseallotypic marker Lyo5 (31) and more recently CD45(32). L-CAis the most accurate descriptive nameand is used for the purpose of this review. The L-CAfamily is a major cell surface componentof lymphocytes and carries muchof the carbohydrate of these cells. It has been estimated that 10%of the lymphocyte surface is occupied by one or more L-CAmembers (33). Because of this abundance, L-CAwas easily detected on SDS-polyacrylamide gels of lymphocytemembranes(36-39) (Figure 1A). It was initially characterized as the major specificity of antilymphocytesera (34) and as an allotypic marker (31, 35). The primary protein structure has been determined from the analysis of cDNAclones, and this information, 339 0732-0582/89/04104?339502.00

Annual Reviews 340

THOMAS


(A) (B)"THYMOCYTES TOTAL APPARENT r PASSED T CELLS .3 Mrx10 / IELUTED I[B [ELLS

200 150 100 -

TOTALT

(c)

i rTc/~

50IBO" Fi#ure1 SDS-PAGE of NaB[3H]4 labeled L-CA purified by monoclonal antibodyaffinity chromatography fromrat lymphoid cell surfaces. A. Thetracks showtotal cell extract (TOTAL), the unretainedmaterialpassedthroughthe affinity column(PASSED), and materialeluted fromthe column(ELUTED). B. Comparison of affinity purified L-CA from thymoeytes, T cells andB cells. C. Comparison of affinity purifiedL-CA fromtotal T cells, CD4T cells (TH)andCD8T cells (Tc/s). Reproduced fromRef. (19) withpermission.

along with studies on the genomicorganization, has delineated the molecular basis for the L-CAfamily (40-48). The function of L-CAhas been enigma, but studies with antibodies to L-CAhave implicated this family in lymphocyteactivation (49-74). This review examinesthese observations, and the discussion is divided into four sections. The first section describes the genomieorganization and the molecular basis of the family. The second section details the glycoprotein structure, and the third section examines the differential expression of family members.In the last section, functional studies are correlated with family memberexpression and the glycoprotein structure. I use these observations to argue that the carbohydrate structures on L-CAare of functional importance and may well be involved in

Annual Reviews LEUKOCYTE-COMMON

341

ANTIGEN

determining cell-cell interactions. Theseinteractions will likely result in transmembrane signaling to the cytoplasmic domain.


GENOMIC

STRUCTURE

The L-CAgene maps to chromosome1 in both mice (75) and humans0q32; 44) and is part of a syntenic region between humansand mice (76). This syntenic group has several genes of immunologicalrelevance. Besides the L-CAgene, the genes for Factor H and C4 binding protein map to this region for both species (76-78). Additionally, in mice this region contains the genes for the IgG Fc receptor, and the genes responsible for the GLD lymphoproliferative disorder and MLSantigens (76, 79). In humans, the genes encoding the complement receptors CR1and CR2as well as decay accelerating factor and membranecofactor protein map to this region (76,79-82). The conserved linkage group appears to be approximately centimorgans in length (76). The significance of the conserved linkage group is not yet understood. The mouse L-CAgene is composed of 34 exons encoding a protein of 1291 amino acids (Figure 2) (N. A. Johnson, M. L. Thomas,unpublished). Transcription appears to be initiated at either of two exons, la or lb, that encode most of the 5’ untranslated region, although the precise start site has not yet been mapped(47). These two exons are not used in a cell-typespecific fashion and it is not clear whethertheir differential use results in any functional difference in the mRNA.Exon 2 encodes the remainder of the 5’ untranslated region and all of the leader sequence. Betweenexons 2 and 3 is an intron of approximately 50 kb (47). Exons 3-33 are located within a 60 kb stretch of DNA,and therefore, the mouse L-CAgene is approximately 110 kb in length. Exons 3-15 encode amino acid residues 1-537 of the external domain; exon 16 encodes the membranespanning region residues 538-574; and exons 17-32 encode residues 576-1178 of the cytoplasmic domain. Exon33 is the largest exon and encodes the remaining

17 19 20 21 1alb2

leader

79

11

34

s 618\1ol 12

I

I I O-linked Cys1

Figure 2 Genomic structure

16 18~ 122 1314

I Cys2

15]

~1

II

11

I repeat1 transmembrane

for the mouse L-CA gene.

25

28 30 31

232,128

Exons that

~71~, ~13z33

~I

I repeat2

encode possible

domains are grouped and labeled below the line. The size of each exon is not to scale. data is from Refs (46, 47) and our unpublished observations.

subThe

Annual Reviews 342

THOMAS

portion of the cytoplasmic domain, residues 1179-1268and the 3’ untranslated region of 1.1 kb.


From One Gene, Multiple Messayes The L-CAfamily is generated by alternative splicing of three exons. Exons 4, 5, and 6 are used differentially to generate potentially eight different mRNAs (46-48). Six of the eight possible transcripts have been isolated as cDNAsfrom three different species: humans, mice, and rats (42, 4448). 13 lymphocytesuse all three exons resulting in the largest molecular weight form thymocytes splice between exon 3 and 7. Multiple transcripts occur in peripheral T lymphocyte populations (27, 42-48). Other than coding region

mRNA

7 1) L-CA456

cDNA h,m,r

~

Expression gp Mr (kD) B Thy T4 230-240 +

7 2) L-CA4s 7 3) L-CA4s

210-220

+

+

210-220

?

?

210-220

+

+

190-200

?

?

190-200

+

+

190-200

+

+

7 4) L-CAs6 7 5) 4 L-CA 3

6) 5 L-CA

~ 3

7

5

h,r 6

7

7) L-CAe 3 8) L-CA37

7 170-180

+

+

+

Figure 3 Schematic diagram of the differential exon usage for L-CA mRNA.Exons 3-7 are indicated by open rectangles. The exons used within each form are indicated by the LCAsubscript and the splicing events by the V. The forms that have been cloned are indicated in the cDNAcolumn. The species from which each form was isolated are indicated by: h; human;m: mouse; r: rat; and no indicates that no cDNAshave been isolated for that form. The label gp Mrindicates the approximate molecular weights of the glycoprotein for each type of L-CA.The expression columnindicates the possible expression patterns for different L-CAmembers.Definitive proof for the expression patterns within the T cells subsets is not available and the ? indicates only a possible expression based on SDS-PAGE. Data is from Refs (44-48).

Annual Reviews LEUKOCYTE-COMMON

ANTIGEN

343

thymocytesand B cells, the exact use of the variable exons in leukocyte populations is not known.Decipheringwhenin normal cell development each of the various mRNAs is expressedis an importantissue that remains to be resolved.


Regulation of Alternative

Splicing

Manyexamples exist of differential exonsplicing, but little is knownabout the regulation of these events. Cell-type-specificalternative exonusage, however,tends to be the exceptionrather than the rule (83). Splicing mammalianpre-mRNA involves the use of small nuclear ribonucleoprotein particles (snRNPs) (reviewedin 84, 85). It is likely that snRNPs, or associatedfactors, are involvedin regulating the differential splicing events, andthe mostlikely sites of interaction wouldbe 5’ to the 3’ splice site. A comparisonof humanand mousesequences 5’ and 3’ to exons 49 is shownin Figure 4A. There are stretches of significant homologyin the sequence5’ to the 3" splice site of the differentially splicedexons4-6. Of particular note, is a similar nanomersequencewhichis found 5’ to exons 4-6 in both humansand mice (Figure 4B) and which contains conserved TGAT sequence. The importance of these conserved sequences in the regulation of the differential splicing remainsto be determined. In summary,L-CAis transcribed from a single gene, and the family is generatedby the alternative splicing of three exons,4, 5, and6, to generate potentially eight mRNAs. Thepatterns of expression are controlled in a cell type-specificfashion. L-CA GLYCOPROTEIN: CONSIDERATIONS

STRUCTURAL

The completeprimary sequence for human,mouse,and rat L-CAhas been determinedfrom the analysis of cDNA clones (Figure 5) (40, 42, 44-48). The glycoprotein consists of an amino-terminal,external domainthat is (depending upon the family member)391-552 amino acids in length, membrane spanningregion of 22 aminoacids and a very large cytoplasmic domainof 705 aminoacids. It is apparent from the consensus sequence (Figure 5) that the cytoplasmicdomainis highly conservedbetweenspecies, 85%identical residues over 705 aminoacids, while the external domainis muchless so, 35%over 538 aminoacids. The disparity in conservation betweenthe external and internal domainsindicates that the structural requirements,presumably in interacting with other molecules,are different. The External Domain Theprimarysequencehas yielded insights into the architecture of the LCAmolecule. Biochemicalanalysis combinedwith protein sequence data


Annual Reviews 344 THOMAS

0

0

0

Annual Reviews LEUKOCYTE-COMMON ANTIGEN

345

divides the external domain into at least three subdomains: an O-linked region and two separate cysteine rich regions. The rat thymocyteform (LCA37; Figure 3) is the only memberthat has been extensively studied at the glycoprotein level (3, 40, 45, 86). It is heavily glycosylated (25% weight carbohydrate; 86) with 14 potential N-linked carbohydrate sites, all but one of which are glycosylated (45). Other L-CAfamily members contain between 1 t and 18 N-linked carbohydrate sites.


SEQUENCE ENCODEDBY THE VARIABLE EXONS The

immediate mouse aminoterminal sequence, encoded by exons 3-8, and the intron/exon boundaries are shown in Figure 6. The sequence is rich in serines and threonines (34%). There is 14%proline content and no cysteines, indicating a random structure, a sequence characteristic of O-linked carbohydrate sites (87). Therefore, the sequence encoded by exons 3-8 are potential sites for Olinked carbohydrate attachment, and the variable use of exons 4, 5, and 6 changes the numberof potential site~. For the rat thymocyteform, it has been demonstrated that all the O-linked carbohydrates are found within the first 32 amino acids (45). (This corresponds to sequence encoded mouseexons 3, 7, and 8; Figures 5 and 6.) O-LINKED CARBOHYDRATE STRUCTURES The use of the variable exons in a cell type-specific mannerimplies that the sequence encoded by these exons will be of functional consequence.There are two reasons for believing that it is the carbohydratesstructures that are of importance. First, the protein sequencesencodedby exons 3-8 are very similar, rich in serines, threonines, and prolines and are characteristic of O-linked glycosylation sites. Second, a comparison of the human, mouse, and rat protein sequence for this region (Figure 5) shows that the overall homology is low, only 40%. This wouldindicate that it is not the precise protein structure that is of importance. Little is knownabout the types of carbohydrate structures found in this region, with one exception. Lefrancois & Bevan described two monoclonal antibodies, CT1 and CT2, that inhibit cytolysis ofcytotoxic T lymphocytes (CTLs) (88). The CT1and CT2antigens are expressed on few cells types; amongleukocytes, only fetal thymocytes, intestinal intraepithelial CD8T lymphocytes and CTLsexpress these antigens (89-91). These antibodies

Figure 4 A. Comparisonof humanand mouse 3’ and 5’ intron sequences adjacent to exons 4-9. The variable exons are exons 4~6. Sequencesof significant homologyare overlined for humanand underlined for mouse. The amino acids encoded by the exons are given in the single letter code. B. Acomparison of a nanomersequence located 5’ to the 3’ splice site, Humansequence data is from Ref, (48) and the mouse sequence is from Refs (46, 47) our unpublished data.


Annual Reviews 346 ~r~or~as



347

Annual Reviews 348

THOMAS T--3--T

4

GQT~ PT~ P S~ DELS~,~ENALLLPQ~, DP LPARTTE~S TPP s, i s~ ERGNGS~S~ ETTYH49


PGVLS, ,TLLPHLS, PQPDS, Q,TPS, AGGADT, QT.FS..SQADNP ,TLT. eA~GGG~, DPP 99 T6__ T-GT~G ~.R’~VPGT 1 P/kD T.~kFPVD"I’P $ L/~RN S S/S,k,.S P ~B~ SNVS T ~ ~ SS ~S ~z19 --7 -T-- 8 --T LT,T,LT,PS,T,LGL~S, ,TDPP S,,T,TIAT, T,,TKQT. CA 1’79 Figured The mouseamino-terminal sequenceencodedby exons3-8. The slartof the malure protein is indicaled by ~ andthesequence encoded by eache×onis indicated by the

lines abovethe seqnence. Adot is beloweachserineandthreonineresidue,andpotentialNlinedcarbohydrates sites are overlined.Thesequence is fromRefs(42, 46).

recognize an O-linked car.bohydrate structure found predominantly on the L-CAof these cell types (92). Although the inhibitory effects of the CT1/CT2antibodies on CTLs are not necessarily through L-CA, since there are two other cell-surface glycoproteins that express the antigen. The carbohydrate structure recognized by CT1and CT2 is unusual, involving N-acetylgalactosamine in a ill,4 linkage to galactose and sialic acid in an c~2,3 linkage (GalNAcfll~,(NeuAce2,3)Gal)(93). Since all O-linked carbohydrate structures appear to be found at the amino-terminal end of L-CA,this region most likely contains the CT1/CT2epitopes. It may, however,be located at multiple sites within the O-linked region. Another example of a monoclonal antibody that recognizes a restricted carbohydrate epitope on L-CAis the NK-9antibody (84). The antigen expressed by virtually all humanT cells and NKcells. Whilelittle is known about the structure of this epitope, it is of muchinterest to determine whether it is another unusual O-linked carbohydrate structure. CYSTEINE RICHREGION The exterior domain is approximately 540 amino acids in length for the largest molecular weight form (L-CA456;Figure 3), the amino-terminal serine/threonine rich region is 177 amino acids in length, and therefore there are approximately 360 amino acids between the O-linked region and the membranespanning region. The homology between species for this region is 33%over 352 residues. This level of homologyis unusually low for the same molecule amongdifferent species and is more similar to that found amongmembersof a supergene family, such as the immunoglobulinfamily (95). This supports the argument that the carbohydrates in this region are of functional importance, and indeed this region contains manyN-linked carbohydrate sites. However,similar to membersof the immunoglobulinfamily, there are key conserved residues such as the cysteines, tyrosines, prolines, and tryptophans (Figure 5),

Annual Reviews


LEUKOCYTE-COMMON ANTIGEN

349

and therefore the exterior domain of human, mouse, and rat L-CAmay have a conserved three-dimensional structure. The region carboxy-terminal to the O-linked subdomain contains 16 cysteine residues in humanand rat L-CAand 18 in mouse, and it is divided into two subdomains. All the cysteine residues are presumably disulfide bonded. However,the position of the disulfide bonds is unknown,and the only evidence to show disulfide linkage is the separation upon reduction of two 50-kd tryptic fragments of the exterior domain(40). The positions of the cysteine residues, with the exception of the two extra cysteine residues of mouse, are conserved (Figure 5) and are clustered into two groups, of eight cysteines each. Thefirst eight cysteines are within a stretch of approximately 100 amino acids (ca. 180-280; Figure 5), and the next eight (disregarding the two nonconserved cysteines in mouse) are within the following 220 aminoacids (ca. 280-500). The potential N-linked carbohydrate groups follow this pattern. One third of the potential sites are found in the O-linked region, one third in the first cysteine-rich region, and the last one third in the second cysteine region. After the first cysteine of the second cluster (humanresidue 319; Figure 5) there is a short stretch of about 20 aminoacids that is very divergent betweenthe three species. This region contains the two extra cysteines in mouse L-CA. The exterior domain of rat L-CAcan be isolated as a 100kd fragment (40). Further digestion with trypsin cleaves in this region lysinc residue 289, yielding two fragments of 50 kd each that are disulfide bonded(40). This is consistent with the theory that this stretch of 20 amino acids is a loop structure linking the two cysteine clusters. In summary, amino acid sequence and protein chemistry data suggest that the exterior domainis divided into at least three subdomains: an Olinked region, which varies between the different forms, and two cysteine clusters, one of approximately 100 amino acids and the other of 220 amino acids. Cytoplasmic

Domain

The cytoplasmic domain (see note added in proof) is remarkable its size, 705 aminoacids. It is the largest reported to date, corresponding to an approximate molecular weight of 83 kd. An internal duplication of about 300 aminoacids (Figure 7) indicates that the cytoplasmic domain consists of at least two subdomains. In marked contrast to the exterior domain, comparison of the human, mouse, and rat sequence shows that the cytoplasmic domain is highly conserved (Figure 4). The level of homology is 85% over 705 amino acids. Furthermore, if conservative amino acid substitutions are accounted for, the level rises to 95%.The large size of the cytoplasmic domainsuggests that it will interact


Annual Reviews

350 THOMAS

Annual Reviews



351

with cytoplasmic components,and this is likely to be important in transmembranesignaling. The L-CAcytoplasmic domain in mice is known to be phosphorylated at serine residues (96). The mouse thymomacell line, BW5147,L-CA constitutively phosphorylated, and two-dimensional tryptic peptide maps indicate that there are several sites ofphosphorylation,all at serine residues (96). In recent studies by Autero & Gahmberg,L-CAof peripheral human T cells becamephosphorylated after treating cells with tumor promoting phorbol esters (97). This would be through the protein kinase C pathway, and L-CAis a knownsubstrate for this enzyme(98). There are multiple potential sites for phosphorylation by protein kinase C in the cytoplasmic domain. Most notable of these is the conserved lys-lys-arg-ser sequence immediately past the membrane spanning region. This sequence is in relation to the membranesimilar to that of the protein kinase C phosphorylation sites of the IL-2 receptor and the epidermal growth factor receptor (98). There are other potential phosphorylation sites of interest: a proline/ serine rich region at approximateposition 900 (914 for humans,Figure 4) and a glutamic acid/aspartic acid/serine rich region at approximate position 950 (964 for humans; Figure 4). These are potential sites for glycogen synthase kinase 3 and 5, respectively (99). The glutamic acid/ aspartic acid region is of particular interest because if the serines at this site are phosphorylated, this area of the molecule becomes extremely acidic. Whetherany or all of these sites are phosphorylated in a physiological situation and how this relates to L-CAfunction remain to be determined. L-CAhas been reported to associate with other molecules; however, only one has been characterized. Bourguignon and her colleagues have recently presented evidence that L-CAcan associate with the cytoskeletal protein fodrin (100, 101), and this interaction wouldbe through the cytoplasmic domain. It is knownthat L-CA will form caps on lymphocytes whencoupled with antibodies and that other surface glycoproteins will cocap (102). This phenomenonmaybe due to a cytoskeletal association with L-CA.The physiological significance is of course uncertain, but perhaps this is an indication that L-CAmaybe involved in somefunction involving the cytoskeleton such as cell motility or membraneorganization. L-CA, therefore, has a large, highly conserved cytoplasmic domain which may interact with cytoskeletal components. Cytoplasmic interactions of L-CAare likely to be important in signal transduction. Electron Microscopy Rat thymocyte L-CA, as viewed by electron microscopy after low-angle shadowing, consists of a rod-like structure of 18 nmand a globular head

Annual Reviews


352

THOMAS

of 12 nm(103). The electron micrographs ofWoollett et al (103) are shown in Figure 8. The rod-like structure is the external domain(Figure 8C). This is demonstrated by the isolation and electron microscopy of a 100-kd tryptic fragment that contains the external portion of the molecule and the membranespanning region but very little of the cytoplasmic domain. It is apparent that the globular domain is missing. WhenL-CAis viewed in the absence of detergent, multimers are formed clustered around the head group with the tail structure pointing outwards (Figure 8B). Remarkably, as seen by electron microscopy, multimers of L-CAexist in the presence of detergents. This suggests that L-CAexists as a multimer on the cell surface. There are several lines of evidence to support this conclusion. Multimers of L-CAexist in gel filtration chromatographyin the presence of detergents, suggesting that the hydrophobic interactions of L-CAare strong (103). Second, L-CAimmunoprecipitated from cells treated with a reducible cross-linking reagent are completely cross-linked to one another (61, 96). Upon reduction of the cross-link, only L-CA molecules are found; no other molecules are released. This indicates that only and all L-CAmolecules are within the distance of the cross-linking reagent to one another. Finally, it should be noted that there is high homology in the membranespanning region of the L-CAmolecule between humans, mice, and rats (Figure 5), and this maybe important in determining multimer associations. The L-CAglycoprotein consists of an amino-terminal, heavily glycosylated exterior domain, composedof at least three subdomains; a membrane spanning region; and a large cytoplasmic domain, composed of at least two subdomains. L-CAfamily membersdiffer by the size of the Olinked carbohydrate region in the exterior domain.

L-CA FAMILY MEMBER EXPRESSION Understanding the pattern of expression for each of the L-CAfamily membersis a critical question that needs to be resolved. In cells where only one form is expressed, as in B cells, the pattern is clear. However,in cells, such as T cells, wherea single cell expresses multiple forms and the pattern changes in differentiation and activation, elucidating the pattern of expression is difficult, although some inference can be made through studies with monoclonal antibodies. There are two types of antibodies to L-CA:antibodies that recognize commonepitopes (epitopes found on all members) and antibodies restricted epitopes (epitopes found on some but not all members;CD45R, 32). Antibodies to restricted epitopes discussed in this review are listed in Table 1. Restricted epitopes can be generated by either of two means:


Annual Reviews


Annual Reviews 354 THOMAS

I1~.~

I1~1AA

~A

Annual Reviews


LEUKOCYTE-COMMON

ANTIGEN

355

sequences encoded by the variable exons (including novel epitopes created by the joining of two exons) or posttranslational modifications, such as glycosylation. Whilethe sites of the restricted antigenic epitopes have been dctcrmincd for only a fcw monoclonal antibodies, manyof the restricted epitopes have been shownto be protein in nature (19, 112, 114). However, there are somecarbohydrate epitopes such as the CT1carbohydrate antigen described above (93). Three monoclonalantibodies that recognize restricted epitopes and have been useful in examining L-CAfamily memberexpression are 2H4, OX22 and UCHLI(104, 109, 115). The antigenic epitope recognized by the monoclonal antibody 2H4 was localized to the sequence encoded by humanvariable exon 4 (or junctions with exon 4) by in vitro translation (106). The antigenic determinent of the monoclonal antibody MRC 22 was localized by isolating and sequencing from rat L-CA,a tryptic peptide containing the antigenic site (45). The partial sequence of the OX22 antigenic peptide is Gly-Ala-Asp-Thr-Gln-Xaa-Leu-Ser-Ser-Gln-AlaAsp-Leu-, which corresponds to a sequence encoded by variable exon 5 (Figure 4). However,since the complete sequence of the tryptic peptide not known, the tryptic sequence could extend into sequence encoded by variable exon 6 or, depending upon the family member, constant exon 7 (45). The UCHL1antibody recognizes a restricted site on human L-CA (109). This antibody only recognizes the low molecular weight form of CA. Therefore, this interesting epitope is most likely contained in the sequence encoded by the junction of exon 3 with 7 (110). While these and other antibodies have been useful in examining L-CAfamily member expression, they do not, with the exception of UCHLI,determine which family memberis expressed. For example, cells that stain with 2H4 are positive for the expression of exon 4, but whether exon 4 is expressed in conjunction with other variable exons cannot be determined. Macropha#es and Granulocytes Most human macrophages and granulocytes express one to two forms of L-CAand bind UCHL1,indicating that at least the low molecular weight form is expressed (61, 108, 116). However, some macrophages bind the 2H4antibody and, in certain circumstances, mayexpress higher molecular weight forms (104). It should be noted that while L-CAis an abundant cell surface glycoprotein on tymphocytes, the expression is lower on macrophages and granulocytes (117). However, stimulation of granulocytes rapidly increases the amountof L-CAexpressed from intracellular stores (118). It will be int~esting to determine whether the increased level expression is also followed by a concomitant change in the family members expressed.

Annual Reviews 356

T~xorvtAs


Variable

Exon Expression

in Lymphoid Populations

B L¥~PHOCV’rES All B cells express the highest molecular weight form of L-CAand, therefore, use all three variable exons. The appearance of the high molecular weight form occurs very early in B-cell differentiation, probably preceding immunoglobulin rearrangement (119). Whether there are changes in L-CAfamily memberexpression when B cells are activated to becomeantibody secreting cells is less clear. The forms expressed by virally transformed B cells, plasmacytomas,and leukemic cell lines are not always the high molecular weight form (L-CA456)(119, 120), yet it is clear whether this reflects a physiological situation. Mouseplasma cells and germinal centers are positive with the monoclonal antibody RA3-2C2 (111), and rat memoryB cells are OX-22positive (115). However, RA3-2C2and OX-22recognize restricted epitope found on several different forms of L-CA,and therefore whether these cells express the highest molecular weight form, L-CA456, cannot be determined with the use of these antibodies. Differences in the carbohydrate structures between L-CAon T cells and B ceils has been well documentedby differential lectin binding (20, 22, 24), analysis of carbohydrate structures, and reactivity with antibodies directed against carbohydrate determinants (20~26). The differences in carbohydrate structures between L-CAon T and B cells is mainly due to the O-linked carbohydrate structures. For example, soybean lectin, which recognizes terminal N-acetyl galactosamine residues, and to a lesser extent galactose, binds selectively to B lymphocytes through O-linked structures on L-CA(22). The carbohydrate structures on B-cell L-CAchange upon activation (121). The addition of IL-2 to LPS/dextran sulfate-stimulated mouse cells causes a five- to fifteen-fold increase in the numberof peanut agglutinin and soybean lectin sites on L-CA.The total level of L-CAon these cells stays relatively constant and, therefore, these changes are through desialation or the addition of sugar moities. The increase in the number of peanut agglutinin sites on L-CAof B cells is of particular interest. Mouse B lymphocytes from germinal centers of lymph nodes and Peyer’s patches are in various stages of activation and can be selectively labeled with peanut agglutinin (122). It is likely that the specific labeling of mouse germinal centers with peanut agglutinin is through the L-CAcarbohydrate structures on these cells. T LYMPHOCYTES Amongleukocytes, T cells express the most intriguing, yet most difficult, L-CApatterns to interpret. There are two reasons for this. First, individual T cells express more than one form and, second, it is not possible to distinguish between some forms by SDS-PAGE.For

Annual Reviews



357

example, the form that uses exons 4 and 5--L-CAas--will be a similar molecular weight to the form that uses exons 5 and 6, L-CA56.It is clear, however, that there is a developmentally regulated expression of L-CA family membersin differentiation from prothymocytes to peripheral T-cell subsets. Immature thymocytes express the lowest molecular weight form, L-CA37. As thymocytes mature, higher molecular weight forms of LCAare expressed (112, 115). This is shown by the small percentage thymocytes that express higher molecular weight forms, the observation that thymocytes expressing higher molecular weight forms are located primarily in the medulla (112, 115, 118) and that the thymocytesexpressing higher molecular weight forms are positive for either CD4or CD8(112). The forms of L-CA expressed by mature thymocytes, although similar when visualized by SDS-PAGE, are not identical to the forms exprcsscd by lymphnode T cells. This is shownby differences in restricted epitopes (112). Likewise, the L-CAforms expressed by CD4and CD8T cells are similar when examined by SDS-PAGE(see Figure 1) (19, 112); however, react differently with antibodies to restricted L-CAepitopes. For example, the 14.8 antibody to mouseL-CAreacts with all CD8T cells but not with CD4T cells (112, 113). The OX-22antibody to rat L-CAreacts with all CD8T cells but with only two thirds of the CD4T cells (115). The 2H4antibody has a similar distribution profile for humanT cells (104). Therefore, while these antibodies are not recognizing equivalent epitopes, they do show that the L-CAfamily membersexpressed by CD4cells are not identical to those expressed by CD8cells. This observation can be extended to the clonal level; CD4T-cell clones are dissimilar to CD8Tcell clones (27, 46). Both CD4and CD8T cells change in their expression of L-CA forms upon activation; however, activated CD4T cells tend to express lower molecular weight forms of L-CAwhile activated CD8T cells tend to express higher molecular weight forms. Recent studies have shown that humanand rat CD4T-cell populations can be divided into functional subsets on the basis of antibodies that bind to restricted L-CAepitopes. For example, the OX-22phenotype subdivides the rat CD4T-cell population (115, 123). The OX-22÷, CD4T cells mediate graft-vs-host reactions, suppress antibody synthesis in animals undergoing these reactions, respond well in both mixed lymphocyte reactions and Concanavalin A (Con A) stimulated cultures, and produce IL2 on activation (115, 123). The OX-22-, CD4T-cell population provides help for B cells in an antigen specific driven assay. The 2H4, WR16, and UCHLIantibodies separate human CD4T cells into homologous populations (71,104, 107, 109, 116). The 2H4÷, CD4T cells induce suppression of IgG secretion in a pokeweedmitogen (PWM)driven B cell response, and they proliferate well to Con A, while the 2H4-,

Annual Reviews


358

THOMAS

CD4T-cell populations respond poorly to mitogens and provide helper signals for PWM-inducedIgG synthesis (104). The WR16antibody seems to divide the CD4T cells into similar populations (7 l, 107). The UCHLI antibody recognizes the reciprocal population to the 2H4antibody. That is, the CD4T cells that are 2H4- are UCHLl+ and have the corresponding functional phenotype (109, 116). Interestingly, the 2H4 antibody recognizes an epitope in exon 4 and immunoprecipitates two higher molecular weight forms of L-CAfrom CD4T cells (105), while the UCHL1 antibody appears to recognize the sequence encoded by the join of exon 3 with exon 7 and recognizes the low molecular weight form (110). Therefore, the 2H4+, CD4T cells express the higher molecular weight forms of L-CA while the UCHL1 +, CD4T cells express the lower molecular weight forms, and this correlates exon usage with functional differences in CD4T-cell subsets. There are manyantibodies to mouserestricted L-CAepitopes, although none have been described that will functionally split the mouseCD4T cell population. However, mouseCD4T-cell clones can be separated into two groups based on their lymphokine production (124, 125). The two groups appear to be remarkably similar to the rat and human CD4T-cell subpopulations separated on the basis of L-CAphenotype. Most mouse CD4 T-cell clones express the lowest molecular weight form of L-CA(L-CA37) regardless of their lymphokineproduction, although there are exceptions (C. T. Weaver, M. L. Thomas, unpublished). Therefore, mouse CD4 cell clones express L-CAforms similar to the rat OX-22-, CD4T cells and the human2H4-, CD4T cells, even though some mouse T-cell clones + or 2H4+, CD4T cells. have a functional phenotype similar to the OX-22 Clearly, the forms of L-CAexpressed by T cells change upon activation. +, CD4T cells and the rat OX-22+, CD4T cells Both the human WRI6 becomenegative for these epitopes upon activation (7 l, 123). Unstimulated mouseCD4T cells, similar to rat CD4T cells, express multiple forms of L-CA.It is likely, therefore, that mouseCD4T-cell clones express the low molecular weight form of L-CAsince they are continuously being activated by in vitro maintenance and that as for humansand rats, there exist for mice in vivo, functional subsets of CD4T cells. Whether the two CD4 subpopulations represent separate or linked lineages is disputed (126128, 128a). However,. this problem is difficult to solve based on L-CA phenotype, since the forms of L-CAexpressed change upon both differentiation and activation. The use of monoclonal antibodies to restricted human L-CAepitopes in immunohistochemicalanalysis reveals interesting patterns of reactivity (108). One antibody, PD7, that immunoprecipitates 220-, 205-, and 190kd forms, reacts with most lymphocytes but fails to react to a subset of

Annual Reviews



359

lymph node T cells and sinusoidal macrophages. In the thymus, PD7stains many medullary thymocytes but fewer cortical thymocytes. The RP1/11 and UCHL1antibodies immunoprecipitate single forms of 220- and 180kd, respectively, and react with a smaller number of leukocytes. RP1/11 reacts with a subpopulation of T cells in the paracortical zone of both tonsils and lymph nodes and strongly with B cells in germinal centers and mantel zones. A small number of medullary thymocytes react with RP 1/11 and in this regard are similar to the reactivity of OX-22in the rat (115). UCHL1 reacts with a small number ofT cells in both the germinal centers and paracortical regions of lymph nodes. In the thymus, UCHL1 reacts with all but a few medullary thymocytes. CT1/CT2 carbohydrate antigens are localized to CD4, CD8thymocytes in the subcapsular region of the thymus of the newbornmouse(89). The differential staining patterns antibodies to restricted L-CAepitopes mayreflect the importance of variable exon usage by leukocytes in tissues. In summary,L-CAfamily memberexpression is precisely controlled in leukocyte differentiation and activation: This suggests that each L-CA family memberwill have distinct interactions. As variable exon usage results in changing the numberof potential O-linked carbohydrate sites and since there are differences between L-CAcarbohydrate structures between leukocyte populations, the distinct interactions for each L-CA family memberwill most likely be mediated through carbohydrate structures. This maybe important for cellular location and for interactions betweenleukocytes and other cells in tissues.

FUNCTIONAL ASPECTS The use of antibodies to disrupt cellular function has provided evidence that L-CAis involved in early lymphocyte activation and is important in transmitting signals across the membrane.These results have been intriguing, and will be even more interesting whenthe precise interactions of L-CAwith other molecules are defined.

B Lymphocytes Antibodies to both commonand restricted epitopes of humanL-CAwill inhibit B-cell proliferation induced by anti-IgM and T cell-replacing factors (65). This inhibition is more effective on small resting B cells and appears to interfere with an early stage of activation as the antibody needs to be addedwithin the first 24 hours of culture. The addition of Ly-5 monoclonal antibodies to Mishell-Dutton cultures inhibits the generation of plaqueforming cells to T cell~lependent antigens, but not T cell-independent antigens (59). Ly-5 will also inhibit IgG responses, but not IgMresponses

Annual Reviews

360 THOMAS


or proliferation, whenB cells are inducedwith lipopolysaccharide(63). This effect is dueto a decreasein the number of cells that switchto become IgG-secretingcells rather than to a decreasein clonal expansionof IgGsecreting cells (129). Theseresults implicateL-CAin B-cell activation and differentiation. However,since the antibodies recognizecommon epitopes, these experimentsleave unanswered the question of whetherthe antibodies are inhibiting an L-CAfunction uniqueto B cells. Natural Killer Cells NKcells expressseveral different tbrmsof L-CAand are similar to T cells in this regard (61, 94). Antisera to the mouseL-CAallotypic marker,Ly5, is a potent inhibitor to NKcell cytolysis, as are somemonoclonal antibody to common L-CAepitopes (49-51, 53). Carbohydratestructures havebeen implicated in mouseNK-cellcytolysis, and it is of interest, therefore, that the poly-N-acetyllactosaminestructures on NKcell L-CA are involvedin target binding and that binding is required for cytolysis (72). Purified L-CAincorporated into liposomes can inhibit conjugate formationbetweenNKcells and target cells. However,no inhibition occurs if the liposomesare treated with Ly-5 or endo-/~-galactosidase. These experimentsare importantin that they are the first demonstrationlinking L-CAfunction with the carbohydrate structures and in coupling L-CA functionwith cell-cell interactions. The monoclonalantibody 13.1 to humanL-CAinhibits NKcell cytolysis. Theinhibition appears to occur at a stage post NKcell binding to target cell (55-57). Whilethe 13.1 antibody recognizes an L-CAepitope on all peripheral blood lymphocytes,the epitope is different from other commonepitopes and maps more distal to the membrane(32, 61). The nature of the 13.1 epitope, whetherit be protein or carbohydrate,has not been determined. Together, the data suggest that L-CAcarbohydrate is importantin binding to target cells and that L-CAis importantin mediating post-binding events, perhapsin the proper positioning of membranes andintracellular signaling. T Lymphocytes L-CAhas been implicated in CTLcytolysis. The mouseallotypic sera Ly5 can inhibit the generation and effector function of CTLsin a mixed lymphocytereaction (54, 60), and at least one monoclonalantibody to commonL-CAdeterminant appears to inhibit CTLfunction. However, not all investigatorsfind this inhibition(50), andothersfind that antibodies to commonL-CAepitope do not inhibit CTLcytolysis (58, 130). The variation betweenexperimentalresults maybe due to differences between

Annual Reviews



361

L-CAepitopes, and it is seen in the effects of antibodies to L-CAon lectin stimulation discussed below. Antibodies to L-CAwill also inhibit NKlike activity of CTLswhen they are stimulated with growth factors (64). It is possible that whenantigen specificity is lost, the interactions of molecules such as L-CAmay become more critical. The correlation betweenfunctional differences in T-cell subsets with the differential expression of L-CAfamily memberssuggests that L-CAis important in mediating these functions. Further evidence is provided by the observation that the 2H4 antibody, which subdivides the humanCD4 T-cell population (discussed above), will also modulate function (66, 73). 2H4÷, CD4T cells mediate the induction of suppression. Whenthese cells are treated with 2H4 antibody, they no longer are able to induce suppression (66). The antibody, however, when included during the culture period, enhances suppression (73). Althoughit is not clear whythe antibody has opposite effects under different conditions, the point is that LCAis involved in activating T cells in this system, perhaps in mediating cell-cell interactions. This is not inconsistent with the effects seen in an NKcell system. Recent observations have shown that the ratio of T-cell subsets, as defined by the expression of various restricted L-CAepitopes, changes in diseases such as multiple sclerosis (131,132), multiple myeloma(133) rheumatoid arthritis (134). Lymphnode T cells from mice with recessive lymphoproliferative disorder, either lpr or yld, aberrantly express the high molecular weight form of L-CA (135). These mice have an abnormal proliferation of T cells resulting in a 50-fold increase in lymph node weight. Whether the forms of L-CAexpressed by T cells contribute to the manifestation or are a result of the disease is an intriguing question that remains to be determined. The WR16antibody, which identifies the same subset as the 2H4 antibody, inhibits pokeweed mitogen-induced proliferation (71), again suggesting an effect in intracellular signaling and T-cell activation. This observation is supported by data that show that antibodies to commonLCAepitopes can inhibit phytohemagglutinin (PHA)-induced proliferation (69). Whenperipheral blood monocytesare cultured with suboptimal doses of PHAin the presence of antibody to a restricted L-CAepitope, the proliferative response is augmentedwith a concomitant increase in the IL2 receptor expression (62). This has also been found with mouse spleen cells (74). Later in the response, the antibody inhibits proliferation (62). This suggests that L-CAis involved in triggering the PHA-inducedmitogenic response, but depending upon the system employed and the antibody used, proliferation is either inhibited or augmented.This provides evidence

Annual Reviews 362

THOMAS

that the carbohydrate structures on L-CAfunction by interacti..ng with lectins, and these:interactions are important for cellular signaling. Further support for-L,GAin T-cell transmembranesignaling is provided by the observations that’antibodies to commonhumanL-CAepitopes can function to replace the signals given by monocytesin triggering lymphocyte activation with anti-CD3 sepharose (68), and antibodies to commonmouse L-CAepitopes can inhibit proliferation induced with anti-Thy-1 antibodies


(52). The critical feature in the above experiments is that antibodies to L-CA inhibit activation and proliferation of lymphocytesand that the effects are dependent upon the antibodies used. This implies that the cause of inhibition is not steric hindrance~0f antibodies binding to an abundant surface glycoprotein. In somecases, such as in NKcell-target cell bifiding and in modulation of lectin-induced responses, the carbohydrate structures are clearly important in L-CAfunction. Finally, it should be noted that while the differential use of the variable exons implies distinct interactions for each L-CAfamily member, leukocyte populations that express the same protein forms of L-CAmayhave distinct L-CAinteractions due to different glycosylation patterns..Thenext step forward is defining the interactions for each L-CAfamily member.

CONCLUSION I have presented several lines of evidence to argue that L-CAfunctions by interacting through its various carbohydrate structures with other cell surfaces, and this then results in transmembrane signaling to the cytoplasmic domain, The interactions..of the L-CAare likely to be important in lymphocyteactivation. The data can be summarizedas follows: (a) There is precise differential use of three exons within leukocyte populations; (b) the variable exons encode potential O-linked carbohydrate sites; (c) there is a lack of protein sequence conservation betweenspecies for the external portion of the molecule; (d) the glycosylation patterns of L-CAvary between leukocytes; (e) antibodies to L-CAmodulate lymphocyte activation and proliferation; and (f) the large cytoplasmic domainis highly conserved betweenspecies. A critical question that needs to be resolved is what other molecules interact with L-CA. ACKNOWLEDGMENTS

I thank Paul Allen, Nancy Johnson, Jeff Milbrandt, and Emil Unanuefor careful reading of the manuscript and critical comments;NancyJohnson, Chris Meyers, and Jeanette Pingle for the genomic data in Figures 2 and


363

4; Matt Haffner for preparing the figures; and Alan Williams for the ph6tographs in Figures 1 and 8 and for permission to reproduce them. M. L~ Thomas is recipient of an Established Investigator Award from the American Heart Association.


NOTE ADDED IN PROOF

Two,recent reports have identified molecules with significant homologyto the cytoplasmic domain of L-CA. Charbonneauet al (136) have isolated several different protein tyrosine phosphotases. Sequenceanalysis of one of these, PTP 1B, shows homology to each of the L-CA cytoplasmic subdomains. Interestingly, the homology between the subdomains is approximately the same as to PTP 1B with the areas of highest homology (Figure 7) also being the area of highest homologyto PTP1B. This suggests that the cytoplasmic domain of L-CAwill have tyrosine phosphotase activity. Streuli et al (137) have isolated cDNAs for a cell surface molecule in which the external domain has homologyto the external domainof NCAMand the cytoplasmic domain has homology to the cytoplasmic domain of L-CA. It is intriguing to speculate that L-CAbelongs to a family of tyrosine phosphotases in which the interactions of the external domain regulate the activity of tyrosine phosphotase domains. This, of course, maybe important in the regulation of cell growth. Literature Cited 1. Trowbridge, I. S. 1978. Interspecies spleen-myeloma hybrid producing monoclonal antibodies against mouse lymphocytesurface glycoprotein, T200. J. Exp. Med. 148:313-23 2. Standring, R., McMaster, W. R., Sunderland, C. A., Williams, A. F. 1978. The predominant heavily glycosylated glycoproteins at the surface of rat lymphoid cells are differentiation antigens. Eur. J. Immunol. 8: 83239 3. Sunderland, C. A., McMaster, W. R., Williams, A. F. 1979. Purification with monoclonal antibody of a predominant leukocyte-common antigen and glycoprotein from rat thymocytes. Eur. J. Immunol. 9:155-59 4. Michaelson, J., Scheid, M., Boyse, E. A. 1979. Biochemicalfeatures of Ly5 alloantigen. Irnmunogenetics 9: 19397 5. Scheid, M. P., Triglia, D. 1979. Further description of the Ly-5 system. Immunogenetics 9:423-33 6. Hoessli, D. C., Vassalli, P. 1980. High molecular weight surface glycoproteins

of murine lymphocytes. J. Immunol. 125:1758-63 7. Dunlap, B., Mixter, P. F., Koller, B., Watson, A., Widmer, M. B., Bach, F. H. 1980. Molecular relationships between large membrane proteins (LMP) expressed on T and B lymphocytes. J. Immunol. 125:1829-31 8. Triglia, D. 1980. Expression of Ly-5 on yolk sac and fetal liver cells of the mouse. Immuno#enetics 11:303-7 9. Omary, M. B., Trowbridge, I. S., Battifora, H. A. 1980. Humanhomologue of murine T200glycoprotein. J. Exp. Med. 152:842-52 10. Andersson, L. C., Karhi, K. K., Gahmberg, C. G., Rodt, H. 1980. Molecularidentification of T cell-specific antigens on humanT lymphocytes and thymocytes. Eur. J. Immunol. 10: 359-62 1 I. Dalchau, R., Kirkley, J., Fabre, J. W. 1980. Monoclonal antibody to a human leukocyte-specific membrane glycoprotein probably homologous to the leukocyte-common (L-C) antigen of the rat. Eur. J. Immunol. 10:737-44

Annual Reviews


364

THOMAS

12. Coffman, g. L., Weissman,I. L. 1981. B220: a B cell-specific memberof the T200 glycoprotein family. Nature 289: 681-83 13. Dalchau, R., Fabre, J. W. 1981. Identification with a monoelonalantibody of a predominantly B lymphocyte-specific determinant of the human leukocyte commonantigen. J. Exp. Med. 153: 753~55 14. Tung, J. S., Scheid, M. P., Pierotti, M. A., Hammerling, U., Boyse, E. A. 1981. Structural features and selective expression of three Ly-5+ cellsurface molecules. Immunogenetics 14: 101~5 15. Watson, A., Dunlap, B., Bach, F. H. 1981. The biosynthesis of Ly-5 in T and B cells. J. Immunol. 127:38~12 16. Sarmiento, M., Loken, M. R., Trowbridge, I. S., Coffman,R. L., Fitch, F. W. 1982. High molecular weight lymphocytesurface proteins are structurally related and are expressed on different cell populations at different times during lymphocyte maturation and differentiation. J. Immunol. 128: 1676-84 17. Tung, J.-S., Scheid, M. P., Palladino, M. A. 1983. Different forms of Ly-5 within the T-cell lineage. Immunogenetics 17:649-54 18. Tung, J.-S., Deere, M. C., Boyse, E. A. 1984. Evidence that Lyo5 product ofT and B cells differ in protein structure. Immunogenetics 19:149-54 19. Woollett, G. R., Barclay, A. N., Puklavec, M., Williams, A. F. 1985. Molecular and antigenic heterogeneity of the rat leukocyte-common antigen from thymocytes and T and B lymphocytes. Eur. J. Immunol. 15:168-73 20. Axelsson, B., Kimura, A., Hammarstrom, S., Wigzell, H., Nilsson, K., Mellstedt, H. 1978. Helix pomatia A hemagglutinin: selectivity of binding to lymphocyte surface glycoproteins on T cells and certain B cells. Eur. J. lmmunol. 8:757~54 21. Childs, R. A., Feizi, T. 1981. Differences in carbohydrate moieties of high molecular weight glycoproteins of human lymphocytes of T and B origins revealed by monoclonal autoantibodies with antMand anti-i specificities. Biochem. Biophys. Res. Commun. 102:1158 64 22. Brown,W. R. A., Williams, A. F. 1982. Lymphocytecell surface glycoproteins which bind to soybean and peanut lectins. Immunology46:713 26 23. Morishima, Y., Ogata, S.-I., Collins, N. H., Dupont. B., Lloyd. K. O. 1982.

Carbohydrate differences in human high molecular weight antigens of Band T-cell lines. Immunogenetics 15: 529 35 24. De Petris, S., Takacs, B. 1983. Relationship between mouse lymphocyte receptors for peanut agglutinin (PNA) and Helix pomatia agglutinin (HPA). Eur. J. lmmunol. 13:831M0 25. Childs, R. A., Dalchau, R., Scudder, P., Hounsell, E. F., Fabre, J. W., Feizi, T. 1983. Evidence for the occurrence of O-glycosidically linked oligosaccharides of poly-N-acetyllactosamine type on the human leucocyte common antigen. Biochem. Biophys. Res. Commun. 110:424-31 26. Ewald, S. J., Refling, P. H. 1985. Analysis of structural diff6rences between Ly-5 molecules of T- and Bcells. Mol. Immunol. 22:581-88 27. Lefrancois, L., Thomas,M. L., Bevan, M. J., Trowbridge,I. S. 1986. Different classes of T lymphocyteshave different mRNAsfor the leukocyte-common antigen, T200. J. Exp. Med. 163: 133% 42 28. Maddox, J. F., Mackay, C. R., Brandon, R. 1985. The sheep analogue of leucocyte commonantigen (LCA). Immunology 55:347-53 29. Houssaint, E., Tobin, S., Cihak, J., Losch, U. 1987. A chicken leukocyte commonantigen: biochemical characterization and ontogenetic study. Eur. J. Immunol. 17:287-90 30. Trowbridge, I. A., Mazauskas, C. 1976. Immunological properties of murine thymus-dependent lymphocyte surface glycoproteins. Eur. J. Immunol. 6: 55762 31. Komuro,K., Itakura, K., Boyse, E. A., John, M. 1975. Ly-5: a new T-lymphocyte antigen system. Immunogenetics1: 452-56 32. Cobbold, S., Hale, G., Waldmann,H. 1987. Non-lineage, LFA-I family, and leucocyte commonantigens: newly and previously defined clusters. In Leucocyte Typing HI, ed. A. J. McMichaelet al, pp. 788-803. Oxford: Oxford Univ. Press 33. Williams, A. F., Barclay, A. N. 1985. Glycoprotein antigens of the lymphocyte surthce and their purification by antibody affinity chromatography. In Handbook of Experimental Immunology, ed. D. M. Weir, L. A. Herzenberg, pp. 22.1-22.24. Oxford: Blackwell Sci. 34. Fabre, J. W., Williams, A. F. 1977. Quantitative serological analysis of a rabbit anti-rat lymphocyte serum and

Annual Reviews



365

splicing. Proc. Natl. Acad. Sci. USA preliminary biochemical characteriza84:5360q53 tion of the major antigen recognized. Transplantation 23:34~59 47. Saga, Y., Tung, J.-S., Shen, F.-W., 35. Lubaroff, D. M. 1973. Analloantigenic Boyse, E. A. 1987. Alternative use of marker on rat thymus and thymus5’ exons in the specification of Ly-5 derived cells. Transpl. Proc. 1:115 isoforms distinguishing hematopoietic cell lineages. Proc. Natl. Acad. Sci. 18 36. Trowbridge, I. S., Ralph, P., Bevan, USA 84:5364-68 M. J. 1975. Differences in the surface 48. Streuli, M., Hall, L. R., Saga, Y., proteins of mouseB and T cells. Proc. Schlossman, S. F., Saito, H. 1987. Natl. Acad. Sci. USA 72:157-61 Differential usage of three exons generates at least five different mRNAs 37. Trowbridge, I. S., Hyman, R., Mazauskas, C. 1976. Surface molecules encoding human leukocyte common of cultured humanlymphoid cells. J. antigens. J. Exp. Med. 166:1548-66 49. Kasai, M., Leclerc, J. C., Shen, F.-W., lmmunol. 6:777-82 Cantor, H. 1979. Identification of Ly 5 38. Andersson, L. C., Wasastjerna, C., Gahmberg, C. G. 1976. Different suron the surface of "natural killer" cells face glycoprotein patterns on human in normal and athymic inbred mouse T-, B- and leukemic-lymphocytes. Int. strains. Immunogenetics 8:153-59 J. Cancer 17:40-46 50. Minato, N., Reid, L., Cantor, H., Lengyel, P., Bloom, B. R. 1980. Mode 39. Gahmberg, C. G., Hayry, P., Andersof regulation of natural killer cell son, L. C. 1976. Characterization of surface glycoproteins of mouse lymactivity by interferon. J. Exp. Med.152: phoid cells. J. Cell Biol. 68:642-53 124-37 40. Thomas, M. L., Barclay, A. N., 51. Seaman, W. E., Talal, N., Herzenberg, Gagnon,J., Williams, A. F. 1985. EviL. A., Herzenberg, L. A., Ledbetter, dence from cDNAclones that the J. A. 1981. Surface antigens on mouse rat leukocyte-common antigen (T200) natural killer cells: use of monoclonal spans the lipid bilayer and contains a antibodies to inhibit or to enrich cytocytoplasmic domain of 80,000 Mr. Cell toxic activity. J. Immunol. 127:982-86 41:83-93 52. Maino, V. C., Norcross, M. A., 41. Shen, F.-W., Saga, Y., Litman, G., Perkins, M. S., Smith, R. T. 1981. Freeman, G., Tung, J.-S., Cantor, H., Mechanism of the Thy-l-mediated T Boyse, E. A. 1985. Cloning of Ly-5 cell activation: roles of Fc receptors, cDNA.Proc. Natl. Acad. Sci. USA 82: T200, Ia, and H-2 glycoproteins in 7360q53 accessory cell function. J. lmmunol. 42. Saga, Y., Tung, J.-S., Shen, F.-W., 126:1829-36 Boyse, E. A. 1986. Sequences of Ly-5 53. Brooks, C. G., Kuribayashi, K., Sale, cDNA:isoform-related diversity of LyG. E., Henney, C. S. 1982. Charac5 mRNA.Proc. Natl. Acad. Sci. USA terization of five cloned murine cell 83: 6940-44, and correction (1987) 84: lines showing high cytolytic activity against YAC-Icells. J. Immunol. 128: 1991 43. Raschke, W. C. 1987. Cloned murine 2326-35 T200 (Ly-5) cDNAreveals multiple 54. Nakayama,E. 1982. Blocking ofeffectranscripts within B- and T-lymphocyte tor cell cytotoxicity and T-cell prolineages. Proc. Natl. Acad. Sci. USA liferation by Lyt antisera, lmmunol. 84:161-65 Rev. 68:117-34 44. Ralph, S. J., Thomas,M. L., Morton, 55. Targen, S. R., Newman, W. 1983. C. C., Trowbridge, I. S. 1987. StrucDefinition of a "trigger" stage in the tural variants of humanT200 glycoNKcytolytic reaction sequence by a protein (leukocyte-common antigen). monoclonal antibody to the glycoprotein T-200. J. Immunol. 131:114~ EMBOJ. 6:1251-57 45. Barclay, A. N., Jackson, D. I., Willis, 53 A. C., Williams, A. F. 1987. Lympho56. Fast, L. D., Beatty, P., Hansen, J. A., cyte specific heterogeneity in the rat Newman,W. 1983. T cell nature and leucocyte commonantigen (T200) heterogeneity of recognition structures due to differences in polypeptide of humannatural killer (NK)cells. sequences near the NHz-terminus. lmmunol. 131:2404-10 EMBOJ. 6:1259-64 57. Newman,W., Fast, L. D., Rose, L. M. 46. Thomas,M. L., Reynolds, P. J., Chain, 1983. Blockade of NKcell lysis is a A., Ben-Neriah, Y., Trowbridge, I. S. property ofmonoclonal antibodies that 1987. B-cell variant of mouseT200(Lybind to distinct regions of T-200. J. 5): Evidence for alternative mRNA Immunol. 131: 1742-47

Annual Reviews


366 THOMAS 58. Harp, J. A, Ewald, S. J. 1983. Modulation of in vicro immuneresponses by monoclonal antibody to T200 antigen. Cell. Immunol. 81:71-80 59. Yakura, H., Shen, F.-W., Bourcet, E., Boyse, E. A, 1983. Onthe function of Ly-5 in the regulation of antigen-driven B cell differentiation. Comparisonand contrast with Lyb-2. J. Exp. Med. 157: 1077-88 60, Harp, J. A., Davis, B. S., Ewald, S. J. 1984. Inhibition of T cell responses to alloantigens and polyclonal mitogens by Ly-5 antisera. J. lmrnunol. 133: 1015 61. Newman, W., Targen, S. R., Fast, L. D. 1984. Immunobiological and immunochemical aspects of the T-200 family of glycoproteins. Mol. Immunol. 21:1113 21 62. Ledbetter, J. A., Rose, L. M., Spooner, C. E., Beatty, P. G., Martin, P. J., Clark, E. A. 1985. Antibodies to commonleukocyte antigen p220 influence human T cell proliferation by modifying IL 2 receptor expression. J. Immunol. 135:1819 25 63. Yakura, H., Kawabata, I,, Shen, F.-W., Katagiri, M. 1986. Selective inhibition of lipopolysaccharide-induced polyclonal IgG response by monoclonal Ly-5 antibody. J. Immunol. 136: 2729-33 64. Brooks, C. G., Holscher, M. 1987. Cell surface molecules involved in NK recognition by cloned cytotoxic T lymphocytes, J. Immunol. 138:1331 37 65. Mittler, R. S., Greenfield, R. S., Shacter, B. Z., Richard, N. F., Hoffman, M. K. 1987. Antibodies to the common leukocyte antigen (T200) inhibit an early phase in the activation of resting humanB cells. J. Immunol. 138:3159 66 66. Takeuchi, T., Schlossman, S. F., Morimoto, C. 1987. The 2H4 molecule but not the T3-receptor complexis involved in suppressor inducer signals in the AMLRsystem. Cell. Immunol. 107: 107 I~1 67. Takeuchi, T., Rudd, C. E., Schlossman, S. F., Morimoto, C. 1987. Induction of suppression following autologous mixed lymphocytereaction: role of a novel 2H4 antigen. Eur. J. Immunol. 17:97-103 68. Matorell, J., Vilella, R., Borche, L., Rojo, I., Vives, J. 1987. A secondsignal for T cell mitogenesis provided by monoclonal antibodies CD45(T200). Eur. J. Immunol. 17:1447 51 69. Bernabeu, C., Carrera, A. C., De Landazuri, M. O., Sanchez-Madrid, F.

1987. Interaction between the CD45 antigen and phytohemagglutinin. Inhibitory effect on the lectin-induced T cell proliferation by anti-CD45 monoclonal antibody. Eur. J. Immunol. 17:1461-66 70. Small, R. M., Walden, S. M., Ewald, S. J. 1987. Effects of Ly-5 antibodies on antibody-dependent cell-mediated cytotoxicity (ADCC). Immunology 60:15965 71. Moore, K., Nesbitt, A. M. 1987. Func+ T lymtional heterogeneity of CD4 phocytes: two subpopulations with counteracting immunoregulatory functions identified with the monoclonal antibodies WR16 and WR19. Immunology 61:159-65 72. Gilbert, C. W., Zaroukain, M. H., Esselman, W. J. 1988. Poly-N-acet.yllactosamine structures on murme cell surface T200glycoprotein participate in natural killer cell binding to YAC-1targets. J. Imrnunol. 140: 282l 28 73. Morimoto, C., Matsuyama, T., Rudd, C. E., Forsgren, A., Letvin, N. L., Schlossman, S. F. 1988. Role of the 2H4molecule in the activation of suppressor inducer function. Eur. J. ImrnunoL 18:731-37 74. Marvel, J., Mayer, A. 1988. CD45R gives immunofluorescence and transduces signals on mouseT cells. Eur. J. Immunol. 18:825-28 75. Shen, F.-W., Tung, J.-S., Boyse, E. A. 1986. Further definition of the Ly-5system. Immunogenetics 24:14(:~49 76. Seldin, M.F., Morse, H. C. III, Reeves, J. P., Scribner, C. L., LeBoeuf,R. C., Steinberg, A. D. 1988. Genetic analysis of autoimmune gld mice I. Identification of a restriction fragmentlength polymorphismclosely linked to the gld mutation within a conserved linkage group. J. Exp. Med. 167:688-93 77. de Cordoba, S. R., Lublin, D. M., Rubinstein, P., Atkinson, J. P. 1985. Human genes for three complement components that regulate the activation of C3 are tightly linked. J. Exp. Med. 161:1189-95 78. D’Eustachio, P., Kristensen, T., Wetsel, R. A., Riblet, R., Taylor, B. A., Tack, B. F. 1986. Chromosomallocation of the genes encoding complement components C5 and factor H in the mouse. J. lmmunol. 137:3990-95 79. O’Brien, S. J., ed. 1987. Genetic Maps, A compilation of linkage and restriction mapsof genetically studied organisms. Volume 4. New York: Cold Spring Harbor Lab.

Annual Reviews


LEUKOCYTE-COMMON ANTIGEN 80. Weis, J. Y. H., Morton, C. C., Bruns, G. A. P., Weis, J. J., Klickstein, L. B., Wong, W. W., Fearon, D. T. 1987. A complement receptor locus: genes encoding C3b/C4b receptor and C3d/ Epstein-Barr virus receptor map to 1q32. J. Immunol. 138:312-15 81. Lublin, D. M., Lemons,R. S., Le Beau, M. M., Holers, V. M., Tykocinski, M. L., Medof, M. E., Atkinson, J. P. 1987. The gene encoding decay-accelerating factor (DAF)is located in the complement-regulatory locus on the long arm of chromosome 1. J. Exp. Med. 165:1731 36 82. Lublin, D. M., Liszewski, M. K., Post, T. W., Arce, M. A., Le Beau, M. M., Rebentisch, M. B., Lemons, R. S., Seya, T., Atkinson, J. P. 1988. Molecular cloning and chromosomal localization of human membranecofactor protein (MCP). J. Exp. Med. 168: 18194 83. Left, S. E., Rosenfeld, M. G. 1986. Complextranscriptional units: diversity in gene expression by alternative RNAprocessing. Ann. Rev. Biochem. 55:1091-1117 84. Sharp, P. A. 1987. Splicing of messenger RNAprecursors. Science 235: 766-71 85. Maniatis, T., Reed, R. 1987. The role of small nuclear ribonucleoprotein particles in pre-mRNAsplicing. Nature 325:673-78 86. Brown, W. R. A., Barclay, A. N., Sunderland, C. A., Williams, A. F. 1981. Identification of a glycophorinlike moleculesat the cell surface of rat thymocytes. Nature 289:456-60 87. Davis, G. C., Elhammer, A., Russell, D. W., Schneider, W. J., Kornfeld, S., Brown, M. S., Goldstein, J. L. 1986. Deletion of clustered O-linked carbohydrates does not impair function of low density lipoprotein receptor in transfected fibroblast. J. Biol. Chem. 261:2828 38 88. Lefrancois, L., Bevan, M. J. 1985. Functional modifications of cytotoxic T-lymphocyte T200 glycoprotein recognized by monoclonal antibodies. Nature 314:449-51 89. Lefrancois, L. 1987. Expression of carbohydrate differentiation antigens during ontogeny of the murine thymus. J. Immunol. 139:2220-29 90. Lefrancois, L. 1987. Carbohydrate differentiation antigens of murine T cells: expression on intestinal lymphocytes and intestinal epithelium. J. lmmunol. 138:3375-84 91. Lefrancois, L., Bevan, M. J. 1984.

367

Novel antigenic determinants of the T200 glycoprotein expressed preferentially by activated cytotoxic T lymphocytes. J. Immunol. 135:374-83 92. Lefrancois, L., Puddington, L., Machamer, C. E., Bevan, M. J. 1985. Acquisition of cytotoxic T lymphocytespecific carbohydrate differentiation antigens. J. Exp. Med. 162:1275-93 93. Conzelmann, A., Lefrancois, L. 1988. Monoclonal antibodies specific for T cell-associated carbohydrate determinants react with humanblood group antigens Cad and Sda. J. Exp. Med. 167:119-31 94. Nieminen, P., Saksela, E. 1986. NK-9, a distinct sialylated antigen of the T200 family. Eur. J. Immunol. 16:513-18 95. Williams, A. F., Barclay, A. N. 1988. The immunoglobulin superfamily-domains for cell surface recognition. Ann. Rev. Immunol. 6:381-405 96. Omary, M. B., Trowbridge, I. S. 1980. Disposition ofT200 glycoprotein in the plasma membrane of a murine lymphomacell line. J. Biol. Chem. 255: 1662-69 97. Autero, M., Gahmberg, C. G. 1987. Phorbol diesters increase the phosphorylation of the leukocyte common antigen CD45in humanT cells. Eur. J. Immunol. 17:1503-6 98. Shackelford, D. A., Trowbridge, I. S. 1986. Identification of lymphocyteintegral membraneproteins as substrates for protein kinase C. J. BioL Chem. 261:8334~41 99. Hemmings,B. A., Aitken, A., Cohen, P., Rymond, M., Hofmann, F. 1982. Phosphorylation of the type-ll regulatory subunit of cyclic-AMP dependent protein kinase by glycogen synthase kinase 3 and glycogen synthase kinase 5. Eur. J. Biochem. 127:473-81 100. Bourguignon, L. Y. W., Suchard, S. J., Nagpal, M. D., Glenny, J. R. 1985. A T-lymphoma transmembrane glycoprotein (gp 180) is linked to the cytoskeletal protein, fodrin. J. Cell Biol. 101:477-87 101. Suchard, S. J., Bourguignon, L. Y. W. 1987. Further characterization ofa fodrin-containing transmembrane complex from mouse T-lymphoma cells. Biochim. Biophys. Acta 896:35-46 102. Bourguignon, L. Y. W., Hyman, R., Trowbridge,I., Singer, S. J. 1978. Participation of histocompatibility antigens in capping of molecularly independent cell surface components by their specific antibodies. Proc. Nail. Acad. Sci. USA 75:2406-10 103. Woollett, G. R., Williams, A. F.,

Annual Reviews


368 THOMAS Shotton, O. M. 1985. Visualisation by low-angle shadowing of the leucocytecommonantigen. A major cell surface glycoprotein of lymphocytes. EMBO J. 4:2827-30 104. Morimoto, C., Letvin, N. L., Distaso, J. A., Aldrich, W. R., Schlossman, S. F. 1985. The isolation and characterization of the human suppressor inducer T cell subset. J. Immunol.134: 1508-15 105. Rudd, C. E., Morimoto, C., Wong, L. L., Schlossman,S. F. 1987. The subdivision of the T4 (CD4) subset on the basis of the differential expressionof LC/T200 antigens. J. Exp. Med. 166: 1758-73 106. Streuli, M., Matsuyama, T., Morimoto, C., Schlossman, S. F., Saito, H. 1987. Identification of the sequence required for expression of the 2H4 epitope on the humanleukocyte commonantigens. J. Exp~ Med. 166: 156772 107. Moore,K., Nesbitt, A. M. 1986. Identi+ supfication and isolation of OKT4 pressor cells with monoclonalantibody WRI6. Immunology 58:659~64 108. Pulido, R., Cebrian, M., Acevedo, A., de Landazuri, M. O., Sanchez-Madrid, F. 1988. Comparative biochemical and tissue distribution study of four distinct CD45antigen specificities. J. Immunol. 140:3851-57 109. Smith, S. H., Brown,M. H., Rowe, D., Callard, R. E., Beverley, P. C. L. 1986. Functional subsets of human helperinducer cells defined by a new monoclonal antibody, UCHL1.Immunology 58:63-70 110. Terry, L. A., Brown, M. H., Beverley, P. C. L. 1988. The monoclonal antibody, UCHL1,recognizes a 180,000 MWcomponent of the humanleucocytecommonantigen, CD45. Immunology 64:331-36 111. Coffman, R. L. 1982. Surface antigen expression and immunoglobulin gene rearrangement during mouse pre-B cell development, lmmunol. Rev. 69:5 23 112. Lefrancois, L., Goodman, T. 1987. Developmental sequence of T200 antigen modifications in murine T cells. J. Immunol. 139:3718-24 113. Scheid, M. P., Landreth, K. S., Tung, J.-S., Kincade, P. W. 1982. Preferential but nonexclusive expression of macromolecular antigens on B-lineage cells. Immunol. Rev. 69:143-59 114. Dalchau, R., Flanagan, B. F., Fabre, J. W. 1986. Structural implications of the location and stability to proteolytic enzymes of immunodominant deter-

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

minants of the human leukocyte common molecule. Eur. J. Immunol. 16: 993-99 Spickett, G. P., Brandon, M. R., Mason, D. W., Williams, A. F., Woollett, G. R. 1983. MRCOX-22, a monoclonal antibody that labels a newsubset of T |ymphocytes and reacts with the high molecular weight form of the leukocyte-common antigen. J. Exp. Med. 158:795-810 Akbar, A. N., Terry, L., Timms, A., Beverley, P. C. L., Janossy, G. 1988. Loss of CD45R and Gain of UCHL1 reactivity is a feature of PrimedT cells. J. lmmunol. 140:2171-78 Shah, V. O., Civin, C. I., Loken, M. R. 1988. Flow cytometric analysis of human bone marrow. IV. Differential quantitative expression of T-200 commonleukocyte antigen during normal hemopoiesis. J. lmmunol. 140: 1861-67 Lacal, P., Pulido, R., Sanchez-Madrid, F., Mollinedo, F. 1988. Intracellular location of T200 and Mol glycoproreins in humanneutrophils. J. Biol. Chem. 90:9946-51 Kincade, P. W. 1987. Experimental models for understanding B lymphocyte formation. Adv. lmmunol. 41: 181267 Raschke, W. C. 1980. Transformation by Abelson murine leukemia virus: properties of the transformed cells. Cold Spring HarborSyrup. Quant. Biol. 44:1187-94 Cook, R. G., Landolfi, N. F., Mehta, V., Leone, J., Hoyland, D. 1987. Interleukin 2 mediates an alteration in the T200 antigen expressed on activated B lymphocytes. J. lmmunol. 139:991 97 Butcher, E. C., Rouse, R. V., Coffman, R. L., Nottenburg, C. N., Hardy, R. R., Weissman, I. L. 1982. Surface phenotype of Peyer’s patch germinal center cells: implications for the role of germinal centers in B cell differentiation. J. Immunol. 129:2698-2707 Arthur, R. P.~ Mason,D. 1986. T cells that help B cell responses to soluble antigens are distinguishable from those producing interleukin 2 on mitogenic or allogeneic stimulation. J. Exp. Med. 163:774-86 Mosmann,T. R., Coffman, R. L. 1987. Two types of mouse T helper T-cell clone. Implication for immune regulation. Immunol. Today 8:233~7 Janeway, C. A. Jr., Carding, S., Jones, B., Murray, J., Portoles, P., Rasmussen, R., Rojo, J., Saizawa, K., West, ÷ T cells: J., Bottomly, K. 1988. CD4

Annual Reviews


LEUKOCYTE-COMMON ANTIGEN specificity and function, lmmunol.Rev. 101:39-80 126. Sanders, M. E., Makgoba, M. W., Sharrow, S. O., Stephany, D., Springer, T. A., Young, H. A., Shaw, S. 1988. Human memory T lymphocytes express increased levels of three cell adhesion molecules (LFA-3, CD2, and LFA-I) and three other molecules (UCHL1, CDw29, and Pgp-1) and have enhanced IFN-~, production. J. Immunol. 140:1401-7 127. Serra, H. M., Krowka,J. F., Ledbetter, J. A., Pilarski, L. M. 1988. Loss of CD45R(Lp220) represents a postthymic T cell differentiation event. J. lmmunol. 140:1435-41 128. Sanders, M. E., Makgoba, M. W., Shaw, S. 1988. Humannaive and memory T cells: reinterpretation of helperinducer and suppressor-inducer subsets. ImmunoLToday 9:195 99 128a. Mason, D. W. 1988. Subpopulations of T lymphocytes. Immunol. Lett. 14: 169-70 129. Yakura, H., Kawabata, I., Ashida, T., Katagiri, M. 1988. Differential regulation by Ly-5 and Lyb-2 of lgG production induced by lipopolysaccharide and B cell stimulatory factor-1 (IL-4). J. Immunol. 141:875-80 130. Davignon, D., Martz, E., Reynolds, T., Kurzinger, K., Springer, T. A. 1981. Lymphocytefunction-associated antigen I (LFA-1): a surface antigen distinct from Lyt-2,3 that participates in T lymphocyte-mediated killing. Proc. Natl. Acad. Sci. USA 78:4535-39 131. Rose, L. M., Ginsberg, A. H., Rothstein. T. L.. Ledbetter, J. A.,

369

Clark, E. A. 1985. Selective loss of a subset of T helper cells in active multiple sclerosis. Proc. Natl. Acad. Sci. USA 82:7379-93 132. Sobel, R. A., Hailer, D. A., Castro, E. E., Morimoto, C., Weiner, H. L. 1988. The 2H4 (CD45R) antigen selectively decreased in multiple sclerosis lesions. J. lmmunol. 140:2210-14 133. Serra, H. M., Mant, M. J., Ruether, B. A., Ledbetter, J. A., Pilarski, L. M. 1988. Selective loss of CD4+ CD45R+T cells in peripheral blood of multiple myelomapatients. J. Clin. Immunol. 8:259-65 134. llowite, N. T., Wedgewood, R. J., Rose, L. M., Clark, E. A., Lindgren, C. G., Ochs, H. D. 1987. Impaired in vivo and in vitro antibody responses to bacteriophage (I)X 174 in juvenile rheumatoid arthritis. J. Rheumatol. 14: 957~3 135. Tung, J.-S., Saga, Y., Boyse, E. A. 1987. The incongruous Ly-5 phenotype of Ipr/lpr and gld/gld T cells. Immuno,qenetics 25:126-29 136. Charbonneau, H., Tonks, N. K., Walsh, K. A., Fisher, E. H. 1988. The leukocyte commonantigen (CD45): putative receptor-linked protein tyrosine phosphotase. Proc. Natl. Acad. Sci. USA 85:7182-86 137. Streuli, M., Krueger, N. X., Hall, L. R., Schlossman, S. F., Saito, H. 1988. A new member of the immunoglobulin superfamily that has a cytoplasmic region homologous to the leukocyte commonantigen. J. Exp. Med. 168: 1553-62





Ann. Rev. Immunol. 1989. 7:371-405 Copyright © 1989 by Annual Reviews Inc. All rights reserved

T CELL RECEPTORS IN MURINE AUTOIMMUNE DISEASES Hans Acha-Orbea,

L. Steinman,

and H. O. McDevitt

Departments of Medical Microbiology/Immunology and Pediatrics/Neurology, Stanford University, Stanford, California 94305 INTRODUCTION A critical site in the regulation of the immuneresponse is the formation of the trimolecular complexbetween T cell receptor (TCR), class-II molecules of the major histocompatibility complex (MHC),and antigen. Susceptibility for manyautoimmunediseases is associated with genes in the class-II region of the major histocompatibility complex(MHC)(1, 1 a). The gene products of this region are expressed on antigen presenting cells (APC), and their function is to bind fragments of proteins (peptides) present them to T cell receptors, primarily on CD4+ T cells (2-5). These class-II antigens are very polymorphic (6, 7). While any given class-II molecule can bind a wide variety of peptides, different class-II molecules showdistinct broad specificity patterns. This is reflected in the variation between alleles of MHC molecules in their capacity to present different peptides to the immunesystem (8-10). After successful activation of the T cell, the other functions of the immunesystem are initiated whichinclude production of lymphokines (factors influencing other cells of the immune system), activation of antibody production, and maturation and proliferation of cells (11). TCRsrecognize peptides primarily in the context of the appropriate "self" MHCmolecules. This corecognition is termed MHC-restriction (12). The repertoire of TCRsis generated by a "random" somatic recombination event (13-15). In the normal individual a fine balance is maintained to allow the immunesystem to distinguish between "self" and "nonself." Because the MHC molecules cannot make this distinction (16), 371 0732-0582/89/0410~371502.00

Annual Reviews


372

ACHA-ORBEA,

STEINMAN & McDEVITT

the T cell provides some of the regulatory mechanismsthat prevent selfdestruction. Such mechanismsare associated with tolerance induction. Both positive selection for MHC restriction and negative selection against self-reactive T cells upon maturation in the thymus have been suggested as possible mechanisms(17-19). Yet another control might be suppression of self-destructive T cells in the periphery (20, 21). Furthermore, the expression of class-II molecules on only a fraction of an individual’s cells maylimit the frequency of T cells that encounter self plus class-II MHC. Whenthis balance of self-tolerance is perturbed, several additional mechanisms may contribute to the autoimmuneresponse: Aberrant expression of class-II antigens on cells normally devoid of class-II antigen (which then present self-peptides against which tolerance had not been induced), inappropriate lymphokine production, or induction of an immune response against a microbe which cross-reacts with "self" components. This may then induce a cascade of reactions that culminate in an autoimmunedisease. The aim of specific immuneintervention in autoimmuneconditions is to prevent or reverse such self-destruction. Experimental approaches have focused on addition or removal of specific lymphokinesand administration of anti-class I! or anti-CD4 monoclonal antibodies. Recent efforts have been directed at the depletion of T-cell subsets expressing particular TCR variable genes. This review includes a short introduction on the structure and function of the TCRand a brief review of a particularly clear model of T cellmediated murine autoimmunedisease, experimental allergic encephalomyelitis (EAE). A summaryon various approaches for immuneintervention emphasizes the role of anti-TCR antibodies in treatment of autoimmunedisease. The review concentrates on the role of CD4+T cells in autoimmunedisease. Although it is well knownthat this T-cell subset is not responsible for all the steps observed in pathogenesis, these cells play an important role in most of the autoimmunediseases analyzed. T CELL

RECEPTOR

One of the most striking features of the immunesystem is its ability to mount a highly specific immuneresponse toward an indefinitely large numberof foreign antigens. This specificity is mediated by two types of receptor molecules, the TCRexpressed on the cell surface ofT lymphocytes (22-24) and the immunoglobulins (antibodies) expressed on B lymphocytes, and which can be secreted upon activation. These two antigen recognition systems are functionally connected. Most B lymphocytes require help from T lymphocytes for activation (25).

Annual Reviews


TCR IN AUTOIMMUNE DISEASE

373

Antibodies and TCRsperform their tasks using different strategies: The immunoglobulinsusually have a high affinity with dissociation constants in the range of 10-s_ 10-14 (26) and tend to recognize the three-dimensional (conformational) structures of antigens. TCRs,in contrast, generally not recognize native antigen alone but recognize for the most part linear peptides, associated with the highly polymorphic MHCmolecules, and they require accessory molecules such as CD2, CD4, CD8, and LFA-1to enhancetheir ability to bind strongly enoughfor activation (27-29). These antigenic peptides are degraded componentsof proteins processed by the antigen-presenting cells and presented on the cell surface in association with MHCmolecules. Only the MHCmolecules present during the maturation of the T cells are recognized as restriction elements (12). These peptides apparently bind to a single site on the MHCmolecule (30). This observation is supported by the crystallographic structure of a functionally similar class-I MHCmolecule in which a noncovalently associated molecule was found embeddedbetween two s-helices. Most of the polymorphic residues pointed into this grove (3 I, 32). A structure similar to class I has been proposed for class II molecules (33). Generation

of Diversity

Both the B and T cell antigen recognition systems use the same basic mechanismto generate vast diversity. These molecules require a highly diverse antigen-binding region and a constant anchor region. The constant region mediates functions such as complementfixation and B-cell activation for antibodies (34), while the constant region anchor of the TCR is involved in T-cell activation uponantigen contact (35). In the germline, pools of variable (V), junction (J), diversity (D), and constant (C) gene segments are encoded. Unique receptor structures are generated by somatic recombination of these gene segments (13-15, 36). In the TCR 20-30 Vr, 12 Jr, 2 Dr, an estimated 100 V~ and 50-100 J~-gene-segments are encoded in the germline. For the ~-chain the functional molecules are composed of VJCand for the fl-chain of VDJC-segments. Diversity is achieved by the joining of these elements in all the distinct permutations possible and is diagrammaticallyshownin Figure 1. On each clone of T cells only one type of TCRheterodimer is expressed. This is achieved by allelic exclusion (14, 37). In addition to these somatic recombination events, further heterogeneity can be introduced by the following mechanisms: JUNCTIONAL DIVERSITYThe mechanism which joins the different elements is not precise. Nucleotides at either side of the VrDr and DaJr or the V,J~-junctions can be deleted upon recombination. In immunoglobulins deletion of nucleotides have been observed in the J-region only.

Annual Reviews 374

ACHA-ORBEA,

STEINMAN & McDEVITT


V

V

rC~ 3. germ[l~.

O1

I1II I I I Ill I II I II II Hill I~

J 1.1-17 C D 2 d 2,1-2.7 C V Illllll-I

,

--V

! IlJllll

J 2,4-7C v II II I II IIIII

V

D J

IlaHI !_/

I

C

H TM CY

TCR ~ OrOlein

Figure 1 Germline organization

and rearrangement

of TCR c~ and/~ chains.

N-REGION DIVERSITY In these junctional regions random sequences of nucleotides can be inserted. This mechanismis found in both the ce- and the fl-chain of the TCRbut only in the heavy chain of immunoglobulin molecules. O REGION The D region can be translated in all three reading frames in TCR/?-chains, but in the immunoglobulin heavy chain usually only one of the three possible reading frames is found. sOMATIC MUTATIONHigh frequencies

of point mutations are found in antibody molecules. In TCRsno convincing evidence for such a mechanism had been found in sequence analysis of manydifferent TCR-molecules. J-REGIONELEMENTS The TCRlocus encodes more J-region elements than either immunoglobulin locus (50-100 J~, 12 J~ in the TCR,vs 4 in the heavy and 4 in each light chain of immunoglobulins). In immunoglobulinvariable region sequences, three clear hypervariable

Annual Reviews TCP,

IN AUTOIMMUNE DISEASE

375


regions were defined that were found to interact with antigen; these were termed complementarity determining regions CD1, CD2, and CD3(3840). The third coincides with the VJ-junctional region. In TCRmolecules the whole V region seems to be hypervariable, although the major peak of hypervariability is found at the same location as CD3in antibody molecules (41, 91). Similarities in amino acid sequence between TCRand antibody molecules have been found especially in the V- and J-regions. Manyof the amino acids important for immunoglobulin heavy and light chain pairing are conserved (41). Recognition

of Peptides

by the TCR

T cells recognize antigens as peptide fragments associated with MHC molecules. Antigen processing can be mimickedby proteolytic cleavage of proteins or by addition of short synthetic peptides (4, 42). The shortest peptides that can initiate an immuneresponse are 7-9 amino acids long, but usually longer peptides elicit better responses (27). The length peptides bound to MHCantigens in vivo is unknown. Whenit became possible to measure the binding of peptides to solubilized MHCmolecules it becameclear that different MHC molecules select different peptides to be presented to the TCR(43, 44). This is due to the polymorphism MHC molecules. Nevertheless, there is not a complete correlation between binding or particular peptides and initiation of an immuneresponse (44). Using peptides with single amino acid substitutions it was possible to define amino acid residues that interact with MHCmolecules, the TCR, or possibly both (45, 46). Competitionexperiments with different peptides showed that, with the peptide combinations tested, probably only one major peptide binding site exists on the class-II molecule. This agrees with crystallographic structure analysis of a class-I molecule(31, 32). In relating the observations from the competition experiments to a structural model, two possibilities have been suggested: (a) there is only one binding site the MHC molecule, or (b) peptide binding to the MHC induces an allosteric modification in the protein which precludes binding of a second peptide. In addition, structural motifs are present in those peptides which bind MHCmolecules (4749). In several experimental systems, it has been observed that certain mouse strains can elicit an immuneresponse to peptides but not to the proteins containing these peptide sequences (5053). In these instances, nonresponsivenessto particular protein is not due to the inability of the synthetic peptide to bind the MHC molecule or holes in the repertoire of the TCR.One possible interpretation is that naturally proccsscd peptidcs arc bigger than the synthetic peptides used in these studies and that these longer peptides cannot bind strongly enough to certain MHCmolecules or cannot be recognized by TCR(52). Other

Annual Reviews


376

ACHA-ORBEA,

STEINMAN & McDEVITT

explanations are the presence of distinct suppressor epitopes within the protein (53) or deletion of the epitope by antigen processing. A peptide must bind to a MHC molecule to elicit a T-cell response (43, 44). Analysis of T-cell epitopes in different antigens revealed that MHC disparate mousestrains often recognize different peptides of the same protein. Even small epitopes on antigens can elicit a heterogeneousimmune response as determinedby peptide fine specificity for cytochromes(28), repressor (55), staphylococcal nuclease (54), myoglobin(27, 56), simplex glycoprotein (57), ovalbumin (45), and lysozyme (53, 58-60). showthe heterogeneity of the response in one mousestrain, the following example is given: Betweenamino acids 111-118 there are at least three sperm whale myoglobinepitopes seen in the context of I-Ed, whenanalyzed on peptides with single aminoacid differences, peptides of different length, or allo cross-reactivity patterns. Betweenresidues 108 and 117 at least three epitopes could be mappedin the context of I-Ad when analyzed with peptides of different length (27, 56). In experimentally induced autoimmunediseases, antigens involved in pathogenesis are easily defined, and several have been characterized: 1. Arthritis in rat and mouse:Native type-II collagen in collagen-induced arthritis and mycobacterial heat shock protein in adjuvant arthritis (61, 62). 2. Experimental allergic thyroiditis (EAT)in mouse:Thyroglobulin(63). 3. Experimental allergic myasthenia gravis (EAMG):Acetylcholine receptor (AChR)(64). 4. Experimental allergic encephalomyelitis (EAE) in mouse and rat: Myelin basic protein (MBP)and proteolipid protein (PLP) (65-67). It is muchmore difficult to find the initiating antigen in spontaneous autoimmunediseases. Finding a structure in the target organ that can elicit an autoimmuneresponse upon injection in adjuvant does not prove it is the autoantigen involved in spontaneous pathogenesis. Nevertheless, target antigens have been defined in humanrheumatoid arthritis (type-II collagen) (68) and myasthenia gravis (acetylcholine receptor) (64). In the AChRpeptides p195-212 and p257-269 have been partially characterized as epitopes in myasthenia gravis patients of the HLA-DR5and HLA-DR3,DQw2MHChaplotypes, respectively (69). TCR Structure-Function Since it has been possible to obtain cDNAsequences of TCRs, attempts have been made to correlate the complex antigen/MHCrecognition with TCRstructural components. These studies have sought to correlate the TCRcDNAsequences from T-cell clones or hybridomas with patterns of

Annual Reviews


TCR IN AUTOIMMUNEDISEASE

377

fine specificity and cross-reactivity. It is clear that the TCRis the sole structure on the T-cell surface that determinesthe fine specificity in recognition, but in most cases accessory molecules are needed to stabilize or initiate the interaction (70-72). The results obtained in a variety of antigen recognition systems are summarizedin Table 1. T-cell clones reactive to pigeon cytochromec provide the best analyzed antigen-recognition system. It has been selected because of its manypatterns of allo-cross-reactivity, different fine specificities towardspecies variants of cytochromemolecules, and various allo restriction patterns (28). Nevertheless, a limited heterogeneity of TCR-elementshas been detected in this response (73-77). The majority of analyzed T-cell clones utilize memberof the V~I 1 family. Most of the differences in TCRgene-segment usage correlated with the different fine specificity patterns. In T-cell hybridomas with a specific allo-cross-reactivity, selection of TCRelements was observed (76). At least three out of the four V and J elements were found shared between the T-cell clones which exhibited a specific allocross-reactivity pattern. Differences in the VJ~junction can changepeptide Table 1 TCRelements used preferentially

Antigen k’b bCyt.cI-E

Responder mouse strain BI0.A

in antigen/MHCrecognition

Most frequently used aTCR~elements V~ll (100%), J~84 (57%)

bCyt.c I-E k’sorS B10.A

V~I 1 (40%), J~84 (60%)

MBPpl-9

PL

V~4 (100%), J~TA31

MBPpl-ll

B10.PL

Most frequently used TCR~ elements a V~3 (100%), J~l.2 (78%), J~2.5 (22%) Val (71%), Jal.2 (58%), Ja2.1 (43%) Va8.2 (>78%), Ja2.7 (50%)

(75%) V,2.3 (58%), J~39

Va8.2 (79%), Ja2.7 (79%)

(100%) V~4.2 (42%) MBPp89-100 aMls + I-E Arsonate

SJL -SJL, SWR, C57L, C57Br a I-Aa, I-Ak, L

C57BL10

bml2

None None V~3 (80%), J~TA20’ (40%) None

Val3 (21%), Ja2.2 (21%) Val7 (100%) cV~8.1, V~6 cV~17 Va2 (40%), Ja2.5 (40%) None

~Forreferencessee text. ThemostfrequentVand J regionsare listed withthe percentof clonesanalyzed expressingthis element. ~Pigeoncytochrome c peptide 81-104-specificT cell clones. Allorestriction on the indicated class II antigens. ~Themajorityof T cell clones expressingVal7are specific for I-E+antigen;the majorityof Va8.1or Va6expressingT cell clones react with Mls". Thefrequencyof T cell clonesreacting with these antigen/ MHC complexes is unknown.

Annual Reviews


378

ACHA-ORBEA,

STEINMAN & McDEVITT

fine specificity, and T-cell clones distinct in one aminoacid in the N-region of the//-chain and in the usage of a different V~-chainshowdifferences in allo-cross-reactivity (73-75). Comparisonof TCRsequences with specificity for cytochrome c and restriction to I-Ek and b revealed that one amino acid, asparagine at position 100, of the TCRfl-junctional region is found in the majority of T-cell clones. In those T-cell clones restricted to I-Ek and s, another amino acid, aspartic acid, is found at the same position in the majority of TCRsequences (77). All the T-cell clones with restriction to I-Ek and b express V~3 and one of two J~-regions, and a memberof the V~I 1-family with one out of three J~ regions. Similar patterns are found in the T-cell clones restricted to I-Ek and s. The length of the VaJa-junctionalregion seemsto be critical. In contrast, the junctional region of the e-chain showsneither specifically selected amino acids nor specific length. In the response to arsonate (a hapten that covalently binds to free amino groups) four out of five T cells utilize V~3, independent of the MHC restriction element used. After 4 weeks, in vitro culture of Ars-reactive Tcell lines V~3 was enriched. Transfer of the e-chain from an arsonatespecific T cell clone to cells that utilized the samerestriction element but exhibited different antigen reactivity, resulted in specificity for arsonate and the patterns observed in the recipient T-cell clone (78). In contrast to pigeon cytochrome e and arsonate reactive TCR, the response to modified APCs with the hapten N-iodoacetyl-sulfonicnaphtyl-ethylene-diamine (AED), responses show greater heterogeneity TCRusage. Twoout of four T-cell clones use the same J~ and J~ elements, two other ones use a memberof the same V, family (79). Even greater heterogeneity was found for beef-insulin and leukocytic choriomeningitis virus (LCMV)responses (80~82). In all three systems, the exact antigen structures have not been identified, and the greater diversity in TCRusage mayreflect the heterogeneity of peptides recognized. In the response to AED,a hapten reacting with free sulfhydril groups on the APCsurface, the TCRrecognition could be dependent on peptide sequences in addition to the hapten structure. dSperm whale myoglobin-specific T-cell clones in mice expressing I-A and I-Ed showa correlation between the restriction element and usage of Va8(83). Ten out of eleven I-E~ restricted T-cell clones utilized V~8. With alloreactive T cells different results were obtained in different systems. In short-term culture in vitro, two- to four-fold enrichment for specific V regions was observed. Analysis of a large number of T-cell hybridomas in C57BL/10anti-bm 12 responses did not reveal any significant increases in specific TCRelements (84-86; E. Palmer, personal communication).

Annual Reviews



379

Strong selection for TCRelements has been found in specific antigen/ MHCrecognition systems, but use of these elements was not confined to any one antigen or MHCrestriction pattern. In addition analysis of TCRsof T cells with manydifferent specificities and restriction elements has shownthat there is no correlation in the usage of specific TCRelements and the recognition of different classes of MHCmolecules (88, 89). For example, the same V~ and Va regions were found in a class I-restricted cytotoxic T cell and a class II-restricted helper T cell hybridoma.The cells did not showcross-reactivities (89). In addition a T-cell clone specific for sperm whale myoglobinpeptide in association with I-Ed used the same V~, Jt~ and a memberof the same V~ family as an I-Au restricted myelin basic protein (MBP)-specific T-cell clone (A. M. Livingstone, personal communication; 90). The primary difference between TCRsequences of those two clones was found in the D region, although the J,-region usage in the sperm whale myoglobin-specific T-cell clones has not yet been determined. Despite the considerable sequence similarity, these T-cell clones showno cross-reactions (90). In antibody molecules three hypervariable regions in both the heavy and the light chain are found. These hypervariable regions are called complementarity determining regions and are designated CD1, CD2, and CD3(3840). The third coincides with the V J-junction. In TCRmolecules the VJ-junction represents the only significant hypervariable region (91). Assuming that TCRand antibodies have a similar three-dimensional structure, a model has been proposed in which the third hypervariable region in TCRs, which shows the greatest variability, interacts with antigen, whereas the others interact primarily with MHCmolecules (41). The selection of a particular amino acid in the N-region of cytochromespecific T-cell clones with a defined allorestriction pattern agrees with such a model(77). Clonal deletion, selection of mutants, cell fusion and transfer experiments implicated the r-chain more often in MHCrecognition than the a-chain (18, 19, 71, 73). Careful analysis of the results showthat the structure of the r-chain is important for MHCrecognition but does not by itself determine specificity for a particular MHC epitope. In transfer experiments,/?-chain genes were introduced in cells which had a similar antigen specificity but a different MHC restriction pattern. After transfection of the r-chain both patterns were observed. Because both T cells expressed a similar a-chain, a contribution of the a-chain to MHC recognition could not be excluded (71). Whenthe thymomacell line BW5147 is used as a fusion partner to generate T-cell hybridomas, the resulting cells are specific for the primary antigen and often are cross-reactive for allo MHCantigens. Because BW5147contains functional TCRc~ and fl chains, and variants of BW5147 which lack expressed ~ and/3 chains do

Annual Reviews


380

ACHA-ORBEA;,STEINMAN

& McDEVITT

not showthis pattern of cross-reactivity (92), it is possible to conclude tentatively that the ~ chain contributes significantly to antigen and allo MHC recognition in the T cell. A strong correlation between V~17-usageand I-E recognition has been demonstrated (18, 19). In addition Val7 does not recognize I-E per because macrophagesor transfected fibroblasts which express high levels of I-E on the surface cannot stimulate T cell hybridomasspecific for I-E, whereasI-E on the surface of a B cell can (93). The TCR V regions show hypervariability throughout the entire sequence with the exception of conserved amino acids which are probably involved in chain-chain pairing (41). It could be argued that different parts of the V regions can take part in antigen/MHCrecognition. This would explain whycertain V regions such as V~8 are found in manyantigen/MHC recognition systems. Furthermore, the apparent requirement for accessory molecules in the formation of a stable T cell-target cell interaction may suggest a requirement for a conformational change before a stable interaction can occur. It is possible to conclude from the studies with cytochromec that more than one region of the TCRis selected in a specific recognition pattern, that several elements are selected in concert. It appears that a specific three-dimensional structure of TCR,in a response to a single antigenic epitope in association with MHCmolecules, is required, not merely a particular element. It is possible that a single element plays a major role in recognition of the target antigen or MHC;however, the complete combining site is formed by contributions from all the variable parts of the TCR.A more careful analysis on T cells with identical fine specificity should help to clarify thi~ issue. Clonal Deletion The class-II molecules determine which peptides will be presented to the TCR.Because it has been shown that APCscan present self-antigens in vivo (16), additional mechanisms must operate to protect an organism from self-destruction. Goodevidence for one of several possible mechanisms has recently been presented. With the availability of monoclonal antibodies specific for certain TCRV~regions, analysis of expression levels of various TCRV genes in different mouse strains became possible. Surprising results were obtained: It was found that in I-E-expressing mouse strains and FI hybrids between I-E+ and I-E- mice, the majority of V~17and V~ 11-expressing T cells were detectable in thethymus, but absent from the peripheral T-cell population (18, 19; E. Palmer, personal communication). This was not due to TCRrecognition of cells expressing native I-E molecules but was correlated with antigen presented by I-E

Annual Reviews



381

molecules (93). In another series of experiments, it was shown that mouse strains expressing at least one copy of Mlsa (new nomenclature system: Mls-la), a strong mixed lymphocyte stimulatory determinant, both V~6and V~8.1expressing T cells were depleted in the periphery (94, 95). As seen with Val7, levels of expression in the thymus were normal, and furthermore Va6-positive cells were found in the cortex but not in the medulla of the thymus (H. Hengartner, personal communication). From the strikingly similar results in these two systems, it is possible to conclude that under some circumstances particular V~l-bearing T cells are depleted in the thymus. It is possible that clonal deletion does not alwaysresult in such complete removal of V regions in the periphery but that in other antigen-recognition systems, certain ~-heterodimeric structures will be eliminated. Becauseof the lower frequency of specific heterodimeric TCRs, these instances are muchharder to detect.

TCR IN EXPERIMENTAL ALLERGIC ENCEPHALOMYELITIS Characteristics of EAE Experimental allergic encephalomyelitis (EAE) is an induced autoimmune disease of the central nervous system (65) and mimics in manyrespects the humandisease of multiple sclerosis (MS) (96-98). The disease can induced in manyspecies including mice and rats and, accidentally, in humans(through contamination of a vaccine preparation with spinal cord material). The disease is characterized by the acute onset of paralysis. Perivascular infiltration by mononuclearcells in the CNSis observed in both models. Demyelination is found in mice (99-102, 108, 116). An inflammation similar to a delayed type hypersensitivity (DTH)reaction is found in the lesions. Onlya small proportionof the participating cells is actually specific for the target antigen(s). In rats, irradiation of recipient rats before transfer of disease with activated spleen cells reveals no such DTH-likereaction, even though paralysis is observed (105). If sublethal doses of antigens are administered, rats recover several days after the onset of clinical signs. After one weekrats often showone relapse and thereafter are protected for the rest of their lives fromfurther induction of EAE(102, 103, 106). Mice develop acute, chronic and chronic-relapsing forms of the disease, depending on the methodof induction (65, 96, 97, 107-109). The disease is more difficult to induce in mice than in other species. Coinjection of pertussis vaccine is usually required for induction of disease (107-109).

Annual Reviews


382

ACHA-ORBEA,

STEINMAN & McDEVITT

EAE is clearly mediated by CD4÷ T lymphocytes. This was shown originally in adoptive transfer studies with lymph node cells (110), cells (111-113), selected T-cell subsets (114), and later by isolation encephalitogenic T-cell lines or clones (115, 116). One of the genes mediating susceptibility is localized in the MHC classII region (50a). The best analyzed encephalitogenic protein is myelin basic protein (MBP),but other encephalitogenic antigens are found in the brain (65-67). Using synthetic peptides it was possible to map the epitopes accurately without the problems with contamination involved in puriu) fication of proteolytic peptide fragments. In the PL mousestrains (H-2 two encephalitogenic peptides in MBPhave been characterized: The MBPpeptide p35-47 and the acetylated N-terminal nonapeptide MBPpl-9 (MBPis naturally acetylated at its N-terminal in vivo) (50, 117, 118). N-terminal nonapeptide is also encephalitogenic in the B10.PL mouse strain. The MBPp35-47 peptide has not been tested in B10.PL. In the SJL mouse(H-2s) at least three overlapping encephalitogenic peptides at positions MBPp96-109, p89-100, and p92-101 have been identified (119121). In the Lewis rat the major encephalitogenic peptide is MBPp68-88, and another unmappedepitope lies outside this region (12~124). Using autologous MBPas an immunogenin rats results in about 50%of hybridomasdirected against either epitope. It has been observed that induction of EAEis easier with xenogenic MBP.The encephalitogenic T-cell clones elicited against these xeno-determinants also react with self-determinants. Therefore it was argued that suppression of self-reactive immuneresponses is responsible for this effect (124). The

PL-Mouse

To analyze the heterogeneity of TCRin MBP-specific T-cell clones from PL or (PL x SJL)F1 mice, a panel of 21 MBP-pl-9-specific T-cell clones was analyzed for expression of TCRscontaining V~8with the monoclonal anti-Va8 antibodies KJ-16 and F23.1 (125, 126). These T-cell clones were derived from individual mice primed and stimulated with different forms of antigens, or if pairs of clones were taken from the same cloning experiment, their TCRcomposition was distinct by anti-V~8 antibody reactivity. At least 78%of the PL and (PL x SJL) derived MBPpl-9-specific T-cell clones were Va8+. This represents a minimumestimate because more than one clone per experiment was taken only whenit was distinguishable with the Ve8 marker (127). In most mouse strains V~8 is the predominant TCRV region expressed on 10-25%of peripheral T cells. Therefore it was important to see whether the increased frequency of V~8 expressed by MBPp 1-9 reactive T cells represents the situation in vivo or whether it is a cloning artifact. To

Annual Reviews



383

÷T + cells were sorted into Va8 address this issue, MBPpl-11-primed, CD4 and Va8- cells and then directly stimulated in vitro. Becausevery little selection occurs under these conditions, the response of the V~8+ but not + T cells to MBPpl-11 reflects the finding with the T-cell the Va8-, CD4 clones. In the response to a second encephalitogenic MBP-peptide, MBP p35-47, the majority of clones are Va8-, and in the same experiment the ÷ T-cell population (90). bulk of the response was found in the Va8-, CD4 To further characterize the heterogeneity of TCRin MBP-p1-9-specific, I-Au restricted T-cell clones, cDNA clones of TCR~- and r-chains of eight representative, independent T-cell clones were sequenced. The summary of these experiments is shownin Table 2. T-cell clones from PL mice with this fine specificity pattern use a very limited set of V regions for both chains. All utilize the same V,-chain and as mentioned above over 78% the same Va-chain. There seems to be no stringent usage of specific Ja and junctional Da regions. Amongthe eight analyzed T-cell clones, four different VaJa and three different J~ patterns were observed, but the repertoire is very limited considering all the available possibilities. The predominant pattern was found in 50%of T-cell clones for the J~-chain and in r 75% for the J,-chain. As shown in Figure 2 the region in the VaJ junction revealed a unique amino acid sequence for every VaJ/~ combination (90). Analysis of fine specificity differences with MBP-pl-9peptides with single amino acid substitutions surprisingly revealed very similar dose response curves in all the T-cell clones. The only difference betweenclones that share only the V~ region was lack of reactivity with one very weakly stimulatory peptide (90). TwoT-cell clones that shared the amino acid sequence for both TCR Table 2 TCRgene elements used by MBP-p1-9-specific, I-Au restricted T cell clones, derived from PL mice T cell clone PJB-20 PJPR-2.2 PJPR-6.2 F1-21 P JR-25 PJB-18 PJPR-7.5 Fl-12

Origin

Va

PL PL PL (PL x SJL)F1 PL PL PL (PL x SJL)F1

8.2 8.2 8.2 8.2 8.2 8.2 8.2 4

Ja 2.7 2.7 2.7 2.7 2.3 2.3 2.5 2.5

V~

J~

aP JR-25 P JR-25 P JR-25 PJR-25 P JR-25 PJR-25 PJR-25 PJR-25

TA31 TA31 TA31 TA31 TA31 TA31 TT11 FI-12

Encephalitogenicity

ND + +

~ V~PJR-25 is a memberof the V,4 family with 95%homologyto V,TA65.J,FI-I 2 has 94%homology u. in nucleotidesequenceto J,5CC7(90). All the clonesare restricted to I-A

Annual Reviews

384

ACHA-ORBEA,STEINMAN~ McDEVITT

~ I : GGGACAGGGGGC GROUP

CLONE

V~

~ 2 : GGGACTGGGGGGGC

Junctionalregion TGC TCC TAT ...

P~-20

ASG 8.2 GCC AGC GGT G

GLGE ~ CTG ~G GAG


A s G

S 5 G

A S G D 8.2 GeeAAGC$GGTGGATDG ~’~% PJR-25 PJB-18

P~R-2.2 P~-20

A

~ AC-C AC,C

P~-25 P~-25

TGC C~LR G CALR ~ G~ CALR

$

~ G

CTG AG

~

~

GG

AG

~ P~R-7.5

4 FI-12

~-25

P~-25

GG CALS

CALS T~ GCT ~G AG CAL$ T~ S~ ~G AG

D T AT A ER T ~G A~

d~

AGT GCA ... A G GCAA "’" GT GCA ...

~

0

CAA GAC Cc-,r._,TX,~;._.C~ AAC CAA

CALR

2P~-25~-25~T~GAG

G

ACG T~S CT TCC G

8.2 GCC AGC GGT GAT G

A s

-4 F~-~

A

Y TAT ...

A Y

d~ 2.3

J~ 25

d~"=~’"

PNYGNE CC ~C TAT G~ ~T ~G ANYGNE CC ~C ~AT GGA ~T GAG PNYGNE

0 ~ T~

ANYGNE ~CTATG~T~G PNYGNE

J=T~J

NGGSG AT ~G ~C A~ GGC NTDK ~r Ace ~C ~

J~11

J u ~

Figure 2 TCRnucleotide and anaino acid sequences of MBP1-9 specific T cell clones. In Figure 2A the TCRanucleotide and aminoacid sequencesof the junctional region are shown. The groupswere defined by the different VJcombinations.In Figure 2B the c~-chain sequences are shown. Reproducedwith permission from Cell.

chains differed in their ability to transfer disease. Thereforethe TCR on a CD4÷T cell does not by itself determine encephalitogenicity. Other factors such as production of lymphokinesand homingpatterns in vivo could determineencephalitogenicity.

Annual Reviews



385

Based on these findings it was possible to analyze whether treatment of autoimmunedisease was possible with anti-V~-specific monoclonal antibodies. The monoclonal antibodies used efficiently deplete V~-expressing T cells in vivo. As shownin Figure 3, treatment with anti-V~8 results in remission of an ongoing autoimmunedisease which was induced with an encephalitogenic Va8-expressing T clone (90). This clearly demonstrates that the autoimmuneresponse is not just triggered by the encephalitogenic T cell and then is self perpetuating; rather, removal of the critical T cell blocks an ongoing disease and the health of the experimental animal is regained. In addition, EAEinduced with MBPpl-ll peptide could be prevented with anti-V~8 treatment (see Table 3), and ongoing disease induced with whole MBPcould significantly reverted (see Table 4) (90). This is remarkable because least.two encephalitogenic peptides mapto this protein in the PL mouse. Although it is not knownwhich of the two epitopes is immunodominant in vivo, the MBPpeptide p35-47 elicits mostly Va8 T cells. The remission of MBPinduced disease with anti-V~8 implicates a dominantrole for these cells in EAE. The BIO.PL

Mouse

The B10.PL mouse strain differs from the PL mouse strain in that the genes outside the MHCare derived from B10. In this strain MBPpl-9

3

0

10

20

30

40

Daysafter injection o! F23.1 Fit~ure 3 Reversal of T cell clone-induced disease with the monoclonal anti-V8 antibody F23.1. (PL x SJL)F1 mice were injected with the encephalitogenic T cell clone P Jr-25, 5 × 106 cells intraveneously (i.v.). After paralysis was apparent the mice were randomizedand treated with monoclonalantibody F23.1 or as a negative control an isotype matched antibody $5.2 twice on day one and three after paralysis was observed. Severity scale: 3, paraplegia; 2, paraparesis; 1, limp tail; 0 normal. Darkboxes represent the control experimentwith antibody $5.2, open boxes F23.1-treatment. Reproduced with permission from Cell.

Annual Reviews 386

ACHA-ORBEA, STEINMAN& McDEVITT

Table3 Preventionof MBP-peptide p 1-11 inducedEAE withV/38-specific monoclonal antibodyF23.1. Monoclonal ~ antibody


F23.1 $5.2

Incidence of bdisease

Mean onset of clinical disease (day)

1/19 9/20

20 15

~(PL × SJL)FI mice were immunizedon day 0 with MBPpl11 in completeFreund’sadjuvant and pertussis vaccine. They received500~g of the purified antibodieson days - 1, ~ and 9. Antibody$5.2 represents an isotype matchedcontrol antibody whichdoes not bind to mouseT cells. bNumber of animalssick/numberof total animals Reproducedwith permissionfrom Cell.

also is encephalitogenic, MBPp35-47 has not been tested. In a study analyzing the TCRheterogeneity of MBPp 1-9-specific T-cell hybridomas, similarities to as well as differences from the above results described for PL mice were obtained (128). The majority of MBPpl-9-specific T-cell clones utilized V~8(79%)and J~2.7. (This region is called either J~2.6 J~2.7 by various authors, depending on whether the sixth J region of the second cluster, a pseudogene, is counted or not). The remaining hybridomasutilize Vfl3 and J~2.2. Fifty-eight percent of the clones use V~2.3; the rest use V,4.2. Both types of T-cell clones rearrange the same J~ gene segment, J~39 (128). In a way similar to the above results, different V~J~ rearrangements use strikingly different junctional regions, whereas rearrangements including the same V and J regions show a conservation

Table 4 Reversal of guinea pig MBPinduced disease with Vfl8 specific monoclonalantibody F23.1. Number of mice with clinical symptoms72 hours after treatment Treatment F23.1 KJ23a

Number of mice with clinical symptoms14 days after treatment

None

Mild

Severe

None

Mild

Severe

Deaths

12 1

5 12

2 9

14 9

3 2

1 7

1 4

aDiseasewasinducedas described in Table3 but with guinea pig MBP.Treatmentwas started 24 hours after the first clinical signs of EAEweredetected. Theexperimentwasdone in a doubleblind fashion. The micereceived 200#g of the antibodies i,p, Monoclonal antibodyF23,1depletes the Va8positive T cells and KJ23adepletesthe Vgl7 positive T cells in vivo, Bothantibodiesdeplete similar amountsof T cells in the (PLx SJL)F1miceused in these experiments.Reproduced with permissionfromCell.

Annual Reviews TCR IN AUTOIMMUNE DISEASE

387


of aminoacid sequencesimilarity. Contraryto the results with PLmice, all the ~-chain junctional sequencesof the V~2.3-expressingclones were identical at the nucleotidelevel, and both, V~2.3and V=4.2,use the same J, region. In addition, the most predominant 3-chains in both mouse strains use the sameV andJ elementsbut different D-regions. Howcan these differences be explained? 1. The major differences between the PL and B10.PLmousestrains are the genes outside the MHC. Thevariation in the results betweenthese twostrains mayreflect strain-to-strain variations in V-regionsequences. In support of this, the V,4-sequencefromthe PLmicerepresents a new memberof the V~4-familywhichhas not yet been shownto be present on the B10background.A similar situation could be found for the V~2.3sequence.If there are strain-related differences in the V gene segments,critical residues could be altered, and this wouldexplain why PLmice use only one V-regionin the response to the MBP 1-9 peptide. 2. T cells expressing the V=2.3-sequence maybe clonally deleted due to peptides present in the PLbut not the B 10 background. 3. T-cell cloningcould select for specific T-cell populations.Becausethe T-cell clones analyzed comefromdifferent cloning experiments, such a selection uponcloning probably wouldreflect differential growth requirementsof different types of clones. SuchT-cell clones of the two groupswouldprobablyalso have a different function in vivo. u in hy4. The V~2.3 chain could have a higher affinity for MBP/I-A bridomasthan in T cell clones due to higher levels of accessorymolecules or the contribution of the BWTCRelementsin binding. Further experimentsare neededto clarify this issue. The SJL Mouse SJL mice are not susceptible to induction of EAEwith the N-terminal MBP-peptides. In this mousestrain, a peptide in the C-terminalregion of the MBP protein is responsible for encephalitogenicity(50). It has been shownthat at least two encephalitogenic peptides are located between position p89-101(120). About50%of the T-cell clones recognizing MBP p89-101in association with I-As expressVal7and require proline at position 101 to be stimulated. The T-cell clones reactive to MBP 89-100do not require position 101 of the peptide for recognition (120). A third encephalitogenicpeptide, MBP p96-109,has beenidentified in this region of MBP(121). Althoughthe TCRusage of the Val7- T cell clones has not yet been + T cell clones is blockedby the characterized, disease inducedwith Va17 Val 7 specific monoclonal antibodyK J-23. In contrast, anti-Val7 does not

Annual Reviews 388

ACHA-ORBEA,

STEINMAN & McDEVITT


block disease induced with MBPp89-101or intact MBP(120). Until the TCRusage of the Val7-negativepopulationis analyzedin moredetail, it is not clear whether the SJL as well as the PLand B10.PLmice use a limited set of TCRelements. The response in SJL mice is clearly more complexthan in either PL or BI0.PL mice and maybe similar to the responsepattern seen with the modelpeptides. EAE in Rats In the Lewisrat one of the major encephalitogenic epitopes is located between positions MBPp68-88 (122-124). About 50%of hybridomas raised against rat MBPrecognize a different undefined epitope (124). Independent groups determined the sequences of TCRsfrom a total of three encephalitogenicT-cell clones. Bothfound that the V-regionsused by these MBPp68-88-reactiveT cell clones were strikingly homologous to the mouseV~8.2region (80 and 93 %in aminoacid sequence).Furthermore the TCRof one T-cell clone sequencedclones used a V~2like sequence (70%homology)(129; E. Heber-Katz,submitted). Analysis of Southern blot andNorthernblot data revealed that these V regions are found in the majority of the MBP specific T-cell clones analyzed.Becauseof the usage of different restriction elementsin mouseand rat and the recognition of different non-cross-reactivepeptides, these results cannot be explained. Thesefindings in the rat modelof EAEstrongly underscorethe importance of TCRV-regionsequencesto disease. An anti-TCR monoclonal antibody raised against a MBPp68-88specific T-cell clone, detecting less than 1%of peripheral blood cells, decreased the intensity of or prevented MBP-induced autoimmune disease in 67%of the experimental Lewisrats (130). Dueto the short course the disease in rats, reversal of ongoingdiseasehas not beenreported. Summary of EAE Similar to describedmodelproteins, there are multiplediscrete epitopes on the autoantigen MBP,associated with different MHC restriction elements. However,in two of three systemsanalyzedthus far, the encephalitogenic epitope(s) elicited a T-cell responsewith very little heterogeneity.In the MBP pl-9 response, no majordifferences in fine specificity towarda series of peptides with single aminoacid substitutions could be detected among the T-cell clones analyzed. Surprisingly, in both Lewisrat and H-2u mousestrains, whichrecognize different peptides and different MHC molecules, very similar TCRVregion elementswereused. Differencesbetweenthe two H-2u mousestrains, PLand B10.PL,were observedin the usage of one or two V,-regions, in

Annual Reviews



389

the V~J~-junctional diversity, and the heterogeneity of the J~-regions. Tcell clones with different TCRcomposition possess nearly identical fine specificity patterns, when tested on MBP-peptideswith single amino acid substitutions or cross-reactivities to allo MHC antigens. This suggests that the overall structure of the TCRmolecule is highly selected for and is not simply correlated with the primary sequence of someof its structural components. This raises the possibility of obtaining more specific antibodies that wouldallow depletion of specific idiotypes and lead to highly specific immuneintervention. Treatment of ongoing autoimmune disease with anti-TCR monoclonal antibodies clearly showed that removal of the responsible autoimmuneT cell is sufficient to reverse a clinical condition. Theseantibodies, however, recognize about 15%of all peripheral T cells, and these results should be extended with more specific reagents. EAEcould theoretically recur since in mice treated with anti-V~8, encephalitogenic T lymphocytesexist which are V~8-negative and recognize the encephalitogenic epitope MBPpl-9. Of MBPpl-9-specific T-cell clones, 5-20% are V~8-negative. Moreover V~8-negative T cells recognizing other encephalitogenic epitopes such as MBPp35-47 exist in the PL mouse. Yet in the mice given anti-V~8 antibodies, EAE is reduced even when the immunogen is whole MBP, an autoantigen with at least two discrete encephalitogenic epitopes, including MBPpl-9 and p35-47 (see Tables 3 and 4). In SJL mice, a group of nested epitopes can induce encephalomyelitis. About 50%of the T-cell clones share V~17, and these T-cell clones are specific for one particular peptide, MBPp89-101. Depletion of V~17positive T cells in vivo does not influence susceptibility to peptide-induced disease. The complexity of TCRused in this response remains to be determined. Overall the response of SJL mice is more similar to the above described examples of model peptides like myoglobin or lysozyme. From experiments in different strains of mice it appears that the MHC alleles and not TCRrepertoire determine whichepitope is encephalitogenic. SJL (/-A s) and SWR(I-A q) mice have deleted about 50%of their TCRV~regions, including Ve8, but express an additional Ve gene element, Ve17 (18, 131). B10.T(6R) mice (I-A q) and A.SWmice (1-As) express Ve8 but not V~17. The peptide MBPp89-101 is encephalitogenic in all four strains of mice. These observations indicate that, in mice sharing MHC genes, the T-cell repertoire does not influence susceptibility towards N- or C-terminal MBP-peptides. The bias toward susceptibility to the N-terminal MBP,fragments in the (PL x SJL)Fl-mice is therefore not due to depletion V~17-positive T cells but could be explained either by low levels of expression (50, 132) and/or dominance of the epitopes where Vo8expression is involved.

Annual Reviews

390

ACHA-ORBEA, STEINMAN & McDEVITT

OTHER INDICATIONS FOR LIMITED HETEROGENEITY OF TCR IN AUTOIMMUNE DISEASES


Collagen-InducedArthritis Native type-II collagen can be used to induce arthritis in mice (61). Mice expressing H-2q or H-2r are prone to disease (133). Of the H-2q expressing strains of mice, SWRis unique in that the disease cannot be induced (134); this suggests that other genes outside the MHC influence susceptibility. One of the characteristics of the SWRmouse strains is a deletion of about 50%of the TCRV~ gene elements (131). Genetic backcrosses with C57BL/6mice have shown that (C57BL/6 × SWR)F1mice arc susceptible, and the majority of the F1 × C57BL/6 or F1 × SWRbackcross mice which are susceptible had inherited one or two copies of the TCR~genes from the C57BL/6 parent (134). The two diseased mice which did not express V~8(one of the deleted V~-regions in SWRmice) inherited part the C57BL/6TCR~genes only, including Va6 as determined by restriction fragment length polymorphism (RFLP) at the V~ locus between the SWR and C57BL/6mice (135). In crosses between M1 s"-positive and -negative mice, the V~6expressing T cells are climinated during maturation in the thymus (95). In agreement with the above results, crosses between SWR (M l sa-negative) and CBA(Mlsa-positive) showed reduced incidence disease whereas crosses with M1 sa-negative mousestrains such as C3Hor C57BL/6resulted in susceptibility (136). These results implicate specific TCRgenes in the susceptibility to collagen-induced arthritis. Alternative explanations for these observations are possible however. One of the problems not addressed in the study is that SWRmice express normal amounts of mRNAfor Va6 in lymphatic organs (131). Whether the RFLPpattern represents coding sequence differences between the two alleles is not known. Because autoimmunediseases are end products of very complex mechanisms, other genes in the mousestrains could explain the outcome of these experiments as well. With the availability of V~6monoclonalantibodies, the issue can be clarified (137). Nonobese

Diabetic

Mouse

The nonobese diabetic (NOD)mouse is a model for spontaneous-onset, insulin-dependent diabetes mellitus (138). At least three recessive genes are involved in the disease process, one of which maps to the MHC,probably the class-lI l-A-region (139-142). NODmice do not express I-E molecules; no mRNA for the/-E~-gene is detectable (139). The disease is mediated by T cells, and both CD4+ and CD8+ cells are required for induction of disease by adoptive transfcr into irradiated healthy animals (143, 144).

Annual Reviews



391

Recent experiments have shown that crossing NODmice with transgenic mice expressing functional I-E~ genes and backcrossing these mice to NOD results in protection from invasion of immunecells into the pancreatic islet cells (insulitis) (145). It was argued that I-E is involved in mediating suppression of the immuneresponse. This result was surprising because backcross studies of NODmice with C57.BL6mice, which do not express I-E~, also show a recessive gene mappingto the MHCwhich is needed for disease to occur. Therefore the expression of I-E must operate at a different level. At least two explanations are possible: (a) The transgene was inserted close to one of the other recessive susceptibility genes of the single I-Etransgenic C57.BL6mouseused in the crosses, and (b) expression of I-E results in clonal elimination of TCRmolecules involved in the disease process. Amongthe six V~-genes analyzed V~I 1 and V~17are depleted in I-E expressing mousestrains (18, 19; Ed Palmer, personal communication). Therefore, it is possible that one or several of the TCRelements whichare eliminated in I-E-expressing mice are involved in pathogenesis.

T-Cell Vaccination It was shownby Cohenand colleagues that it is possible to use autoimmune T-cell clones or lines as immunogensto prevent or revert autoimmune diseases (see Table 5). Either doses of T-cell clones which are below the threshold for inducing disease, or irradiated, fixed or pressure-treated T cells exerted this effect (146-149). These animals were protected for prolonged times, and disease could not be induced with these T-cell lines or T-cell clones, or with antigen. This protection is specific in that animals protected from adjuvant arthritis are still susceptible to EAE.This protection can be transferred to unprotected animals with lymph node cells. T-cell clones specific for autoimmuneT cell clones have been isolated from rats that recovered from EAE. These T-cell clones are either CD4+or CD8÷ T cells. The CD8÷ T cells lyse their target cells specifically, and this cytotoxicity is not blockable with anti-CD4, anti-CD8,anti-class I, or anti-class II antibodies. Whentransferred together with the disease-causing clones, they protect from disease (150, 151). It remains to be determined whether the protective T-cell clones recognize TCRor another specific Tcell surface marker expressed on EAE-causing T-cell clones but not on adjuvant arthritis-causing T-cell clones.

OTHER FORMS OF SPECIFIC INTERVENTION

IMMUNE

The interaction between T cells, antigen-presenting cells and antigen initiates normal immuneresponses as well as autoimmune responses.

Annual Reviews

392

ACHA-ORBEA,STEINMAN8, McDEVITT

Therefore several approaches involving immuneintervention in autoimmunediseases have focused on molecules that play a role in this interaction. The results are summarizedin Table 5.


Class-H

Molecules

Treatment protocols with anti-class II-specific antibodies have been shown to partially deplete class II--expressing ce~ls in vivo (152). Successful prevention of autoimmune diseases was achieved in many murine autoimmunediseases (see Table 5) (153-161, Boitard et al, submitted). form of treatment results in a general immunesuppression. In the human population most individuals are heterozygousfor class-II alleles. In studies using F1 mice expressing different class-II antigens, it was shownthat immuneresponses linked to the first allele, despite the depletion of the majority of the cells expressing the second class-II allele, remainedintact (162). Utilization of noncytotoxic antibodies should reveal whether this approach can be used to prevent or revert autoimmunediseases without depletion of cells. These approaches not only interfere with antigen presentation but also result in a suppression of the autoimmuneresponse (163166).

Table5 Attempts in prevention andtreatment of autoimmune diseases Autoimmune disease EAEmouse(PL/J,

T cell Anti-class II Anti-CD4Anti-TCRvaccination Peptides TNF-c¢ Anti-inf-7 + +

+ +

+ +

ND

+

ND

ND

+ + + + + + +

+ + + + + + + R +

+ ND ND ND ND

ND + ND ND ND

+ + ND ND ND ND

ND ND ND ND ND ND

ND ND ND ND ND ND

ND

ND

ND

÷

+

ND

ND

++

++

ND

ND

ND

+

+ +

+

ND

ND

+

ND

ND

ND

B10.PI~) EAEmouse(SJL) EAE(rat) EAMG BBrat NODmouse Collagen-induced arthritis (mouse) Adjuvantarthritis (rat) (NZB x NZW)FI Lupusnephritis EAT(mouse)

Forreferencessee text. +, preventionof autoimmune disease. + +, preventionand reversal of clinically detectable autoimmune disease. R, preventionand reversal of histologically detectable autoimmune disease. -, no effect on either preventionor reversal of autoimmune disease. ND,not done.

Annual Reviews TCR IN AUTOIMMUNE DISEASE


T-Cell

393

Subsets

Treatment of mice with the rat monoclonal antibody GKI.5 results in rapid depletion of CD4+T cells (167). CD4+T cells represent about 60%of the T-cell population which primarily interact with peptides bound to class-II molecules. Depletion of these T cells results in protection from or reversal of many autoimmunediseases (see Table 5) (167-178). several antigen-recognition systems it was shownthat administration of antigen shortly after depletion of CD4+T cells can result in induction of tolerance toward this antigen for a prolonged time. This includes tolerance to the xenogeneic rat antibody GK1.5(179-181). The protocols used for prevention of autoimmune disease used antibody treatment for limited time periods and nevertheless resulted in prolonged protection from autoimmunedisease in autoimmunediabetes (178). It is not clear yet whether the long lasting effects are simply due to the tolerance induction toward the target antigens or whether other mechanismsare involved. Recently it was shownthat application of anti-CD4 antibodies which do not deplete the CD4+ population had similar effects (182, 183). Interleukins After initiation of immuneresponses different interleukins are produced. If an individual produceslowlevels of a specific interleukin, administration of this factor could prevent autoimmunedisease. This was shownin the case of lupus nephritis in (NZB × NZW)F1mice using TNF-a (184). Applicationof interferon-3, resulted in precipitation of the disease in several autoimmune diseases: (185-188). Depletion of interferon-~ with monoclonal antibodies has been shown to have a protective effect in lupus nephritis in (NZB× NZW)F1mice (189). Peptides Application of low doses of peptides or application of the antigen intradermally (which at higher doses can induce an autoimmune disease) resulted in lifelong protection from future induction of disease in adjuvantinduced arthritis and EAEin rats (62, 113, 190-192). Analysis of similar approaches in other systems in which the target antigens are well characterized is necessary. Because the reagents used for prevention have the potential to trigger an autoimmuneresponse, care must be taken in applying this approach in humans. Careful analysis of the mechanismof this effect is needed. In EAEand arthritis, induction of tolerance towards the autoantigen prevented future induction of the autoimmunedisease (193195). Anotherapproach is based on finding specific altered peptides which bind very strongly to specific MHC proteins and prevent presentation of self antigens by competitive interactions or induction of tolerance.

Annual Reviews

394

ACHA-ORBEA, STEINMAN ~ McDEVIT~


CONCLUSIONS The new approaches with anti-TCR-antibodies use more specific reagents for interfering with autoimmunediseases than do those employed previously. EAErepresents the first disease modelin which the heterogeneity of TCRin autoimmunedisease has been analyzed in detail. In mice of the H-2U-haplotype and in Lewis rats, a very limited heterogeneity of TCRs was found to be involved in the autoimmune process. In both models depletion of these cells prevented autoimmunedisease (90, 128-130). SJL mice (H-2~) a more complexpattern was observed. A series of at least three encephalitogenic nested epitopes are located in a region encompassing 20 amino acids of the MBPmolecule (119-121). This situation reminiscent of results obtained with model peptides (27, 45, 53-60a). remains to be determined whether the autoantigen MBPis unusual in this respect or whether such simple patterns are found in other autoimmune diseases as well. Indirect evidence for such limited heterogeneity in other spontaneous or induced autoimmunediseases has been found in collageninduced arthritis, experimental allergic thyroiditis (EAT), diabetes (NOD mouse), and adjuvant arthritis (134-136, 145-151). What triggers an autoimmunereaction is not fully understood. Is autoimmunitythe result of exposure of a very specific autoantigenic peptide to the immunesystem or a more general breakdownof tolerance mechanisms? Furthermore, it remains to be determined whether in the original autoimmunereaction a similar complexity of antigens and TCRsis involved at different stages of the disease, e.g. whenit becomesclinically apparent or chronic. It is possible that a specific T cell starts a chain reaction and that at a later stage the autoimmunecells do not reflect the composition of TCRof the original T-cell population. After destruction of target cells, newautoantigen-reactive T cells could be activated due to the presentation of autoantigens by APCs. The results from treatment of EAEwith antiTCRantibodies are encouraging in that removal of specific T cells during ongoing disease reverses clinically apparent symptoms.According to these results, the immunesystem does not escape when the majority of autoreactive cells is depleted. While leaving 5-20%of the encephalitogenic T cells in PL mice intact after antibody treatment, mice remain protected (90). On the other hand, removal of only 50%of autoreactive T cells the SJL mousehad no effect on the course and intensity of the disease (120). Using different combinations of monoclonal antibodies which can removespecific subsets of T cells will allow examination of whether more than one population of T cells bearing certain TCRV genes must be targeted.

Annual Reviews



395

Treatment of autoimmunediseases with monoclonal antibodies is not without problems. Most of the available antibodies are isolated from mice, rats, rabbits, or hamsters. These antibodies can elicit an immuneresponse in a xenogenic host, and repeated application can lead to anaphylactic shock or rapid elimination of the antibody (196). Recent approaches tailored the CD-regions of the murine antibodies into human immunoglobulin molecules or made chimeric antibodies with constant regions from human and variable regions from mice (197-201). A better understanding of the role of various constant regions will allow utilization of the optimal reagents. The optimal duration of treatment is unknown.If such treatment can be stopped without reoccurrence of the autoimmune response, fewer problems might ensue. Lessons from previous depletion experiments with anti-I-A or anti-CD4 antibodies showed that removal of specific populations of lymphocytes influences the entire immunesystem. Althoughit is clear that mousestrains lacking certain TCRV regions are perfectly healthy, it has to be determined whether depletion of specific V-regions alter the immuneresponse of an individual. This can be easily addressed in mice by depletion of specific V regions in adult mice and analysis of the immuneresponse and tolerance patterns in vivo and in vitro. As shown in the comparison of the I-Au and I-As mouse strains for EAE,different peptides and different TCRmolecules are involved in the autoimmune response. How could anti-TCR treatment be valuable for autoimmune diseases in humans, given the heterogeneity in MHCantigens? Autoimmunediseases often are tightly MHC-linked. In insulindependent diabetes mellitus, for example, about 50%of patients express only one type of susceptibility mediating class-II molecules (for review, see 202). In pemphigusvulgaris, close to 100%of patients have at least one of two class-II antigens responsible for susceptibility (203). In these diseases and many others, the target antigens are not yet known, and further research is needed to evaluate whether such a treatment procedure with a single anti-TCR antibody or pools of anti-TCR V-region antibodies could be useful. On the other hand, if limited TCR-heterogeneityis found in the response to certain autoantigens, treatment could be tried without knowledgeof the autoantigen. Thus "reverse genetics" might be employed for treatment of autoimmunedisease. Due to the clonal deletion process genetic correlations cannot always be expected between certain autoimmune diseases and TCRV regions when the germ-like repertoire is examined. Additional experiments addressing the repertoire of T cells in the periphery are needed. Oneof the possible explanations for dominant protection by certain MHCantigens

Annual Reviews


396


such as DR2 from insulin-dependent diabetes mellitus (IDDM)(for review, see 202) could be due to tolerance to cross-reactive determinants or depletion of self-reactive T-cell clones. In humans, analysis of the knowntarget antigens in myasthenia gravis, rheumatoid arthritis, and thyroiditis will help understand these patterns. In a recent report, TCRrearrangements were analyzed from T cells isolated from joints of rheumatoid arthritis patients. Using constant region probes of the TCR,one specific band was detected in southern blot analysis for each patient (204). Specific bands under these conditions are indicative usage of the same V and J region. The same rearrangements were found in different joints of the same patient but not in the periphery. In a study in multiple sclerosis, restricted heterogeneity of TCRgenes was found in the spinal fluids of patients (205). In another study certain TCRV and RFLPswere found in increased frequencies in multiple sclerosis patients (206). By utilizing antibodies to TCR,highly specific reagents could be used which would impair only a minor part of the T-cell repertoire. What is neededin the near future is the careful analysis of T-cell receptors involved in different aut~immunediseases. The results of these studies will be valuable in deciding which form of therapy is appropriate for a particular autoimmunedisease. ACKNOWLEDGMENTS

Wethank Ed Palmer, Ellen Heber-Katz, and Hans Hengartner for communication of results prior to publication. The critical reading of the manuscript by Theresa Lopez, Alexandra M. Livingstone, and David C. Wraith is highly appreciated. The work was supported by National Institute of Health grants RO1NS18235 and AI,07757, the National Multiple Sclerosis Society, the SwimFoundation, and the Rosenthal Foundation. H. Acha-Orbea was supported by the Juvenile Diabetes Foundation and the Swiss National Science Foundation. Literature Cited 1. Moeller, G. Ed. 1983. HLAand disease susceptibility, lmmunol.Rev. 70 la. Nepom, G. T. 1988. Immunogenetics of the HLA-associated diseases. Concepts Immunopathol. 5:80-105 2. Unanue, E. R., Allen, P. M. 1987. The basis for the immunoregulatoryrole of macrophagesand other accessory cells. Science 236:551-57 3. Dialynas, D. P., Wilde, D. B., Marrack, P., Pierres, A., Wall, K. A., .Havran,

W., Otten, G., Loken, M. R., Pierres, M., Kappler, J., Fitch, F. W. 1983. Characterization of the routine antigenic determinant, designated L3T4a, recognized by monoclonal antibody GK1.5: Expression of L3T4a by functional T cell clones appears to correlate primarily with class II MHCantigenreactivity. Immunol. Rev. 74:29-56 4. Shimonkevitz, R., Kappler, J., Marrack, P., Grey, H. 1983. Antigen recog-

Annual Reviews


TCR IN AUTOIMMUNEDISEASE nition by H-2-restricted T cells. I. Cellfree antigen processing. J. Exp. Med. 158:303-16 5. Watts, T. H., Brian, A. A., Kappler, J. W., Marrack, P., McConnell, H. M. 1984. Antigen presentation by supported planar membranes containing affinity purified I-Aa. Proc. Natl. Acad. Aci. USA 81:7564-68 6. Klein, J. 1982. Immunology: The Science of Self-Nonself Discrimination. NewYork: Wiley 7. Moeller, G. 1985. Molecular geneticsof class I and II MHC antigens. Immunol. Rev. 84, 85 8. Rosenthal, A. S. 1978. Determinant selection and macrophage function in genetic control of the immune response. Immunol. Rev. 40: 13652 9. Allen, P. M., Babbitt, B., Unanue, E. 1987. T cell recognition of lysozyme: The biochemical basis of presentation. Immunol. Rev. 98:171-87 10. Buus, S., Sette, A., Grey, H. M. 1987. The interaction between protein derived immunogenic peptides and la. Immunol. Rev. 98:115 41 11. Jelinek, D. F., Lipsky, P. E. 1987. Regulation of human B lymphocyte activation, proliferation and differentiation. Adv. Immunol. 40:1-59 12. Zinkernagel, R. M., Doherty, P. C. 1979. MHC-restrictedcytotoxic T cells: studies on the biological role of polymorphic major transplantation antigens determining T cell restriction specificity, function and responsiveness. Adv. lmmunol. 27:52-177 13. Davis, M. M. 1985. Molecular genetics of the T cell-receptor beta chain. Ann. Rev. Immunol. 3:537-60 14. Kronenberg, M., Siu, G., Hood, L. E. 1986. The molecular genetics of the Tcell antigen receptor and T-cell antigen recognition. Ann. Rev. Immunol. 4: 529 91 15. Toyonaga, B., Mak, T. W. 1987. Genes of the T cell antigen receptor in normal and malignant T cells. Ann. Rev. Immunol. 5:585-620 16. Lorenz, R. G., Allen, P. 1988. Direct evidence for functional self-protein/Ia molecule complexesin vivo. Proc. Natl. Acad. Sci. USA 85:5220-23 17. Sprent, J., Webb,S. R. 1987. Function and specificity of T cell subsets in the mouse. Adv. Immunol. 41:39-133 17a. Teh, H. S., Kisielow, P., Scott, B., Kishi, H., Uematsu, Y., Bluethmann, H., von Boehmer, H. 1988. Thymic major histocompatibility complexantigens and the ~/~ T-cell receptor deter-

397

mine the CD4/CD8 phenotype of T cells. Nature 335:229-33 18. Kappler, J. W., Wade, T., White, J., Kushnir, E., Blackman, M., Bill, J., Roehm,N., Marrack, P. 1987. A T cell receptor V~ segment that imparts reactivity to a class II major histocompatibility complex product. Cell 49: 263-71 18a. Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz, D. M., Russell, J. H., Loh, D. Y. 1988. Selective expression of an antigen receptor on CD8-bearing T lymphocytes in transgenic mice. Nature 335: 271-74. 19. Kappler, J. W., Roehm, N., Marrack, P. 1987. T cell tolerance by clonal elimination in the thymus. Cell 49: 27380 20. Green, D. R., Flood, P. M., Gershon, R. K. 1983. Immunoregulatory T-cell pathways. Ann. Rev. Irnmunol. 1: 43963 21. Asherson, G. L., Colizzi, V., Zembala, M. 1986. An overview of T-suppressor cell circuits. Ann. Rev. Immunol.4: 3768 22. Haskins, K., Kappler, J., Marrack, P. 1984. The major histocompatibility complex-restricted antigen receptor on T cells. Ann. Rev. Immunol. 2:51-66 23. Meuer, S., Acuto, O., Hercend, T., Schlossman, S., Reinherz, E. L. 1984. The humanT-cell receptor. Ann. Rev. Immunol. 2:23-50 24. Allison, J. P., Lanier, L. L. 1987.Structure, function, and serology of the Tcell antigen receptor complex. Ann. Rev. Immunol. 5:503-40 25. Singer, A., Hodes, R. J. 1983. Mechanisms ofT-cell B-cell interaction. Ann. Rev. lmmunol. 1:211-41 26. Steward, M. W., Streensgaard, J. 1983. Antibody Affinity: Thermodynamic Aspects and Bioloy&al Significance. Boca Raton, FI: CRC 27. Livingstone, A. M., Fathman, C. G. 1986. The structure of T cell epitopes. Ann. Rev. Immunol. 5:477-501 28. Schwartz, R. H. 1985. T-lymphocyte recognition of antigen in association with gene products of the major histocompatibility complex. Ann. Rev. lmmunol. 3:237-61 29. Springer, T. A., Dustin, M. L., Kisb_imoto, T. K., Marlin, S. D. 1987. The lymphocyte function-associated LFA1, CD2, and LFA-3 molecules: Cell adhesion receptors of the immune system. Ann. Rev. Immunol. 5:223-52 30. Guillet, J.-G., Lai, M.-Z., Briner, T. J., Buus, S., Sette, A., Grey, H. M., Smith, J. A., Gefter, M. L. 1987. Immu-

Annual Reviews


398


nological self, non-self discrimination. Science 235:865-70 31. Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S., Strominger, J. L., Wiley, D. C. 1987. The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens. Nature 329:512-18 32. Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S., Strominger, J. L., Wiley, D. C. 1987. Structure of the humanclass I histocompatibility antigen, HLA-A2.Nature 329:506-12 33. Brown, J. H., Jardetzky, T., Saper, M. A., Samraoui, B., Bjorkman, P. J., Wiley, D. C. 1988. A hypothetical model of the foreign antigen binding site of class II histocompatibility molecules. Nature 332:845-850 34. Morgan, E. V., Weigle, W. O. 1987. Biologic activity residing in the Fc region of immunoglobulins. Adv. Immunol. 40:61-134 35. Weiss, A., Imboden, J., Hardy, K., Manger, B., Terhorst, C., Stobo, J. 1986. The role of the T3/antigen receptor complexin T-cell activation. Ann. Bey. Irnmunol. 4:593~519 36. Tonegawa,S. 1983. Somatic generation of antibody diversity. Nature 302: 57581 37. Pernis, B., Chioppino,G., Kelus, A. S., Gess, P. G. H. 1965. Cellular localization of immunoglobulinswith different allotypic specificities in rabbit lymphoid tissues. J. Exp. Med. 122: 85376 38. Wu, T. T., Kabat, E. A. 1970. Analysis of the sequences of the variable regions of Bence Jones protein and myeloma light chain for antibody complementarity. J. Exp. Med. 132:211-50 39. Davies, D. R., Metzger, H. 1983. Structural basis of antibody function. Ann. Rev. ImmunoL 1:87-117 40. Alzari, P. M., Lasambe,M.-B., Poljak, R. J. 1988. Three dimensional structure of antibodies. Ann. Rev. Immunol. 6: 555-80 41. Davis, M. M., Bjorkman, P. J. 1988. Tcell antigen receptor genes and T-cell recognition. Nature 334:395-402 42. Shimonkevitz, R., Colon, S., Kappler, J. W., Marrack, P., Grey, H. M. 1984. Antigen-recognition by H-2 restricted T cells. II. A tryptie ovalbuminpeptide that substitutes for processed antigen. J. Immunol. 133:2067-74 43. Babbitt, B. P., Allen, P. M., Matsueda, G., Haber, E., Unanue, E. R. 1985. Binding of immunogenicpeptides to Ia histocompatibility molecules. Nature 317:359~1

44. Buus, S., Sette, A., Colon, S. M., Miles, C., Grey, H. M. 1987. The relation between major histocompatibility complex (MHC) restriction and the capacity of la to bind immunogenic peptides. Science 235:1353-58 45. Sette, A., Buus, S., Colon, S., Smith, J. A., Miles, C., Grey, H. M. 1987. Structural characteristics of an antigen required for its interaction with Ia and recognition by T cells. Nature 328:395-99 46. Allen, P. M., Matsueda, G. R., Evans, R. J., Dunbar, J. B. Jr., Marshall, G. R., Unanue, E. R. 1987. Identification of the T-cell and Ia contact residues of a T-cell antigenic epitope. Nature 327:713-15 47. DeLisi, C., Berzofsky, J. A. 1985. T-cell antigenic sites tend to be amphipatic structures. Prec. Acad. Natl. Sci. USA 82:7048-52 48. Rothbard, J. B., Taylor, W. R. 1988. A sequence pattern commonto T cell epitopes. EMBOJ 7:93-100 49. Sette, A., Buus, S., Colon, S., Miles, C., Grey, H. M. 1988. I-Ad-binding peptides derived from unrelated protein antigens share a commonstructural motive. J. Immunol. 141:45-48 50. Fritz, R. B., Skeen, M. J., Jen Chou, C.-H., Garcia, M., Egorov, I. K. 1985. Major histocompatibility complexlinked control of the murine immune response to myelin basic protein. J. Immunol. 134:2328-32 50a. Moore,M. J., Singer, D. E., Williams, R. M. 1980. Linkage of severity of experimental allergic encephalomyelitis to the rat major histocompatibility locus. J. Immunol. 124:1815-20 51. Langton, B. C., Mackewicz, C. E., Wan, A. M., Andria, M., Benjamini, E. 1988. Structural features of an antigen required for cellular interactions and for T cell activation in a MHCrestricted response. J. Immunol. 141: 447-56 52. Brett, S. J., Cease, K. B., Berzofsky, J. A. 1988. Influences of antigen processing on the expression of the T cell repertoire. Evidence for MHC-specific hindering structures on the products of processing. J. Exp. Med. 168:357-73 53. Manca, F., Clarke, J. A., Miller, A., Sercarz, E. E., Shastri, N. 1984. Alimited region within hen egg-white lysozymeserves as the focus for a diversity of T cell clones. J. Immunol.133: 207578 54. Finnegan, A., Smith, M. A., Smith, J. A., Berzofsky, J., Sachs, D. H., Hodes, R. J. 1986. The T cell reper-

Annual Reviews


TCR IN AUTOIMMUNEDISEASE toire for recognition of a phylogenetically distant protein antigen. Peptide specificity and MHC restriction of staphylococcal nuclease-specific T cell clones. J. Exp. Med. 164:897410 55. Lai, M.-Z., Ross, D. T., Guillet, J.-G., Briner, T. J., Gefter, M. L., Smith, J. A. 1987. T lymphocyte response to bacteriophage ). repressor cl protein. Recognition of the same peptide presented by Ia molecules of different haplotypes. J. Immunol. 139:3973-80 56. Cease, K. B., Berkover,I., York-Jolley, J., Berzofsky, J. A. 1986. T cell clones specific for an amphipatic or-helical region of sperm whale myoglobin show differing fine specificities for synthetic peptides. Amultiview/single structure interpretation of immunodominance. J. Exp. Med. 164:1779-84 57. Heber-Katz, E., Valentine, S., Dietzschold, B., Burns-Purzycki, C. 1988. Overlapping T cell antigenic sites on a synthetic peptide fragment from herpes simplex virus glycoprotcin D, the degenerate MHC restriction elicited, and functional evidence for antigen-Ia interaction. J. Exp. Med. 167:275-87 58. Allen, P. M., Matsueda, G. R., Haber, E., Unanue,E. R. 1985. Specificity of the T cell receptor: Twodifferent determinants are generated by the same peptide and the I-Ak molecule. J. Immunol. 135:368-73 59. Allen, P. A., McKean, D. J., Beck, B. N., Sheffield, J., Glimcher, L. H. 1985. Direct evidence that a class II molecule and a simple globular protein generate multiple determinants. J. Exp. Med. 162:1264-74 60. Shastri, N., Oki, A., Miller, A., Sercarz, E. E. 1985. Distinct recognition phenotypes exist for T cell clones specific for small peptide regions of proteins. J. Exp. Med. 162:332-45 60a. Ogasawara, K., Maloy, W. L., Schwartz, R. H. 1987. Failure to find holes in the repertoire. Nature 325:450-52 61. Stuart, J. M., Townes, A. S., Kang, A. H. 1984. Collagen autoimmunearthritis. Ann. Rev. Immunol. 2:199-218 62. van Eden, W., Thole, J. E. R., van der Zee, R., Noordzij, A., van Embden,J. D. A., Hensen,E. J., Cohen,I. R. 1988. Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331:171-73 63. Maron, R., Cohen, I. R. 1980. H2K mutation controls immune response phenotype of autoimmunethyroiditis. Critical expression of mutant gene product in both thymus and thyroid glands. J. Exp. Med. 152:1115-20

399

64. Lindstrom, J., Shelton, D., Fuji, Y. 1988. Myasthenia gravis. Adv. ImrnunoL 42:233-84 65. Paterson, P. Y. 1976. Experimental autoimmune (allergic) encephalomyelitis: Induction, pathogenesis, and suppression. In Textbook of lmmunopathology, P. A. Miescher, H. J. MuellerEberhard, pp. 179-213. New York: Grune & Stratton 66. McFarlin, D. E., Blank, S. E., Kibler, R. F., McKneally,S., Shapira, R. 1973. Experimental allergic encephalomyelitis in the rat: Responseto encephalitogenie proteins and peptides. Science 179:478-80 67. Satoh, J., Sakai, K., Endoh,M., Koike, F., Kunishita, T., Namikawa, T., Yamamura,T., Tabira, T. 1987. Experimental allergic encephalomyelitis mediated by murine encephalitogenic T cell lines specific for myelinproteolipid apoprotein. J. Immunol. 138: 17984 68. Holoshitz, J., Druckcr, I., Yaretzky, A., van Eden, W., Klajman, A., Zapidot, Z., Frenkel, A., Cohen,I. R. 1986. T lymphocytes of rheumatoid arthritis patients showaugmentedreactivity to a fraction of mycobacteriacross-relative with cartilage. Lancet ii: 305-9 69. Brocke, S., Brautbar, L., Steinman, L., Abromsky,O., Rothbard, J., Mozes, E. 1988. In vitro proliferative responses and antibody titres specific to human acetylcholine receptor synthetic peptides in patients with myastheniagravis and relation to MHC class II genes. J. Clin. Invest. In press 70. Dembic, Z., Haas, W., Weiss, S., McCubrey, J., Kiefer, H., von Boehmer, H., Steinmetz, M. 1986. Transfection of specificity by murine ct and fl T-cell receptor genes. Nature 320:232-38 71. Saito, T., Germain, R. N., 1987. Predictable acquisition of a new MHC recognition specificity following expression of a transfected T-cell receptor flchain gene. Nature 329:256-59 72. Gabert, J., Langlet, C., Zamoyska,R., Parnes, J. R., Schmitt-Verhulst, A.-M., Malissen, B. 1987. Reconstruction of MHC class I specificity by transfer of the T cell receptor and Lyt-2 genes. Cell 50:545-54 73. Fink, P. J., Matis, L. A., McElligott, D. L., Bookman, M., Hedrick, S. M. 1986. Correlation between T-cell specificity and the structure of the antigen receptor. Nature 321:219-26 74. Winoto, A., Urban, J. L., Lan, N, C., Goverman,J., Hood, L., Hansburg, D. 1986. Predominant use of a V~ gene

Annual Reviews


400


seg,mentin mouseT-cell receptors for cytochromee. Nature 324:679-82 75. Sorger, S. B., Hedrick, S. M., Fink, P. J., Bookman,M. A., Matis, L. A. 1987.Generationof diversity in T cell receptor repertoire specific for pigeon cytochromec. J. Exp. Med.165:279 301 76. Matis,L. A., Sorger,S. B., McElligott, D. L., Fink, P. J., Hedrick,S. M.1987. Themolecularbasis of alloreactivity in antigen-specific, majorhistocompatibility complex-restricted T cell clones. Cell 51:59-69 77. Hedrick,S. M., Engel, I., McElligott, D. L., Fink, P., Hsu,M.-L., Hansburg, D., Matis, L. A. 1988. Selection of aminoacid sequencesin the beta chain of the T cell antigen receptor. Science 239:1541-44 78. Tan, K.-N., Datlot, B. M., Gilmore, J. A., Kronman,A. C., Lee, J. H., Maxam,M. M., Rao, A. 1988. The T cell receptorV,3genesegmentis associated with reactivity to p-azobenzenearsonate. Cell 54:247-61 79. Iwamoto,A., Ohashi,P. S., Pircher, H., Walker, C. L., Michalopoulos,E. E., Rupp,F., Hengartner, H., Mak,T. W. 1987. T cell receptor variable gene usage in a specific cytotoxic T cell response.Primarystructure of the antigen-MHCreceptor of four haptenspecific cytotoxicT cell clones.J. Exp. Med. 165:591-600 80~Spinella, D. G., Hansen;~T.H., Walsh, W. D., Behlke, M. A., Tillinghast, J. P., Chou,H. S., Whiteley, P. J., Kapp,J. A., Pierce, C. W., Shevach, E. M., Loh, D. Y. 1987. Receptor diversity of insulin-specific T cell lines from C57B.L(H-2b) mice. J. Immunol. 138:3991=95 81. Sherman,D. H., Hochman, P. S., Dick, R., Tizard, R., Ramachandran, K. L., Flavell, R. A., Huber, B. T. 1987. Molecularanalysis of antigen recognition by insulin-specific T-cell hybridomas from B6 wild-type and bm12 mutantmice. Mol. Cell. Biol. 7: 186572 82. Pircher, H., Michalopoulos, E. E., Iwamoto,A., Ohashi,P. S., Baenziger,J., Hengartner, H., Zinkernagel, R. M., Mak,T. W. 1987. Molecular analysis of the antigenreceptorof virus-specific cytotoxicT cells and,identificationof a newV~family. Eur./J. Immunol.17: 1843~[6 83. Morel,P. A., Livingstone,A. M., Fathman,C. G. 1987. Correlationof T cell receptor V gene family with MHC restriction. J. Exp. Med.166:583-88

84. Reimann,J., Bellan, A. 1986. Use of Va8genes in splenic Lyt-2+ cytotoxic lymphocyteprecursors reactive to bm 1 or bm14 alloantigen in individual C57BL/6mice. Eur. J. lmmunol. 16: 1597-1602 85. Garman, R. D., Ko, J.-L., Vulpe, C. D., Raulet,D. H.i986. T-cell receptor variable region usage in T-cell populations. Proc. Natl. Acad. Sci. USA 83:3987-91 86. Schilham, M. W., Lang, R., AchaOrbea, H., Benner,R., Joho, R., Hengartner, H. 1987. Fine specificity and T-cell receptor fl-chain generearrangementsof five H-2b-specificcytotoxic T cell clones. Immunogenetics25: 17178 87. Deleted in proof 88. Acuto, O., Campen,T. ,J., Royer, H. D., Hussey, R. E., Poole, C. B., Reinherz,E. L. 1985. Molecularanalysis of T cell receptor (Ti) variable (V) geneexpression.Evidencethat a single Ti Va gene family can be used in formation of V domains on phenotypically andfunctionallydiverseT cell populations. J. Exp. Med.161:1326-43 89. Rupp,F., Brecher,J., Giedlin, M.A., Mosmann,T., Zinkernagel, R. M., Hengartner,H., Joho, R. H.. 1987. Tcell antigen receptors with identical variable regionsbut different diversity and joining region genesegmentshave distinct speeificities but cross-reactive idiotypes. Proc. Natl. Acad.Sci. USA 84:219-22 90. Acha-Orbea,H., Mitch_ell, D. J., Timmermann,L., Wraith, D. C., Tausch, G. S., Waldor,M. K., Zamvil, S. S., McDevitt, H. O.,Steinman, L. 1988. Limitedheterogeneityof T cell receptors fromlymphocytesmediatingautoimmuneencephalomyelitisallows specific immuneintervention. Cell 54: 263-73 91. Kabat, E. A., Wu,T. T., Reid-Miller, ’M., Perry, H. M., Gottesman, K. S. ! 1987.:Sequencesof proteins of immunological interest. Public Health Service, NIH,Bethesda, Md. 92. Blackman;,M. A., Kappler, J. W., Marrack, P. 1988. T-cell specificity and repertoire. Immunol.Rev. 101:5-19 93. Marrack,P., Kappler,J. 1988. T cells can distinguish between allogeneic major histocompatibility complex productson different.tell types. Nature 332:840 43 94. Kappler,J. W., Staerz, U., White,J., Marrack, P. C. 1988. Self-tolerance eliminates T cells specific for Mlsmodifiedproducts of the majorhisto-

Annual Reviews


TCR IN AUTOIMMUNEDISEASE compatibility complex. Nature 332: 354O 95. MacDonald, H. R., Schneider, R., Lees, R. K., Howe, R. C., Acha-Orbea, H., Festenstein, H., Zinkernagel, R. M., Hengartner, H. 1988. T-cell receptor Va use predicts reactivity and tolerance-to Mlsa-encoded antigens. Nature 332:40~15 96. Lassman, H., Wisniewsky, H. M. 1979. Chronic relapsing experimental allergic encephalomyelitis. Clinicopathological comparison with multiple sclerosis. Arch. Neurol. 36:490-97 97. Wisniewsky, H. M., Keith, A. B. 1977. Chronic relapsing experimental allergic encephalomyelitis: An experimental model for multiple sclerosis. Ann. Neurol. 1:144-48 98. Raine, C. S. 1983. Multiple SclerosisPatholoyy, Diaynosis and Manayement, ed. J. Hallpike, C. W. M. Adams, W. W. Tourtellotte. London: Chapman & Hall 99. Zamvil, S. S., Nelson, P. A., Mitchell, D. J., Knobler, R. L., Fritz, R. B., Steinman, L. 1985. Encephalitogenic T cell clones specific for myelin basic protein. An unusual bias in antigen recognition. J. Exp. Med. 162: 210724 100. Mokhtarian, F., McFarlin, D. E., Raine, C. S. 1984. Adoptive transfer of myelin basic protein-sensitized T cells produces chronic relapsing demyelinating disease in mice. Nature 309: 356-58 101. Traugott, U., McFarlin, D. E., Raine, C. S. 1986. Immunopathology of the lesion in chronic relapsing autoimmune encephalomyelitis in the mouse. Cell. lmmunol. 99:395-410 102. McFarlin, D. E., Blank, S., Kibler, R. F. 1974. Recurrent experimental allergic encephalomyelitis in the Lewis rat. J. Irnmunol. 113:712-15 103. Willenborg, D. O. 1979. Experimental allergic encephalomyelitis in the Lewis rat: Studies on the mechanismof recovery from disease and acquired resistance to reinduction. J. lmmunol. 123: 1145-50 104. Deleted in proof 105. Sedgwick, J., Brostoff, S., Mason, D. 1987. Experimental allergic encephalomyelitis in the absence of a classical delayed-type hypersensitivity reaction. J. Exp. Med. 165:1058 75 106. Ben-Nun, A., Cohen, I. R. 1982. Spontaneous remission and acquired resistance to autoimmuneencephalomyelitis (EAE)are associated with suppression of T cell reactivity: Suppressed EAE

401

effector cells recovered at T cell lines. J. Immunol. 128:1450-57 107. Brown, A. M., McFarlin, D. E. 1981. Relapsing experimental allergic encephalomyelitis in the SCLmouse. Lab. InvestitTation 45:278-84 108. Lublin, F. D., Maurer, P. H., Berry, R. G., Tippett, D. 1981. Delayed, relapsing experimental allergic encephalomyelitis in mice. J. lmmunol.126: 819-22 109. Fritz, R. B., Jen Chou, C.-H., McFarlin, D. E. 1983. Relapsing murine experimental allergic encephalomyelitis induced by myelin basic protein. J. Immunol. 130:1024~26 110. Paterson, P. Y. 1960. Transfer of allergic encephalomyelitis in rats by means of lymph node cells. J. Exp. Med. 111:119 35 111. Jankovic, B. D., Waksman, B. H., Arnason, B. G. 1962. Role of the thymus in immunereactions in rats. J. Exp. Med. 116:177-206 112. Gonatas, N. K., Howard, J. C. 1974. Inhibition of experimental allergic encephalomyelitis in rats severely depleted ofT cells. Science 186:839 113. Ortiz-Ortiz, L., Weigle, W. O. 1976. Cellular events in the induction of experimental allergic encephalomyelitis in rats. J. Exp. Med. 144:604~16 114. Pettinelli, C. B., McFarlin, D. E. 1981. Adoptive transfer of experimental allergic encephalomyelitis in SJL/J mice after in vitro activaton of lymph nodecells, by myelin basic protein: Requirement for Lyt 1+2- T lymphocytes. J. Immunol. 127:1420-23 115. Ben-Nun, A., Wekerle, H., Cohen, J. R. 1981. The rapid isolation ofclonable antigen-specific T lymphocyte lines capable of mediating autoimmuneencephalomyelitis. Eur. J. Immunol. I1:195-99 116. Zamvil,S., Nelson, P., Trotter, J., Mitchell, D., Knobler, R., Fritz, R., Steinman,L. 1985. T-cell clones splecific for myelin basic protein inducechronic relapsing paralysis and demyelination. Nature 317:355-58 117. Zamvil, S. S., Mitchell, D. J., Moore; A. C., Kitamura, K., Steinman, L., Rothbard, J. B. 1986. T cell epitope of the autoantigen myelin basic protein that induces encephalomyelitis. Nature 324:258-60 118. Zamvil, S. S., Mitchell, D. J., Powell, M. B., Sakai, K., Rothbard, J. B., Steinman, L. 1988. Homozygous and hybrid class II (I-E) restricted T cell recognition in autoimmune encephalomyelitis. J. Exp. Med. In press

Annual Reviews


402


119. Sakai, K., Zamvil, S. S., Mitchell, D. J., Lim, M., Rothbard, J. B,, Steinman, L. 1988. Characterization of a m~jor encephalitogenic T cell epitope in SJL/J mice with synthetic oligopeptides of myelin basic protein. J. Neuroimmunol. 19:21-32 120. Sakai, K., Sinha, A. A., Mitchell, D. J., Zamvil, S. S., Rothbard, J. B., McDevitt, H. O., Steinman, L. 1988. Involvement of distinct T cell receptors in the autoimmuneencephalitogenic response to nested epitopes of myelin basic protein. Proc. NatL Acad. Sci. USAIn press 121. Kono,D. H., Urban, J. L., Horvath, S. J., Ando, D. G., Saavedra, R. A., Hood, L. 1988. Two minor determinants of myelin basic protein induce experimental allergic encephalomyelitis in SJL/J mice. J. Exp. Med. 168: 213-27 122. Kibler, R. F., Fritz, R. B., Chou, F. C.-H., Chou, C.-H., Peacocke, N. Y., Brown, N. M., McFarlin, D. E. 1977. Immuneresponse of Lewis rats to peptide CI (residues 68-88) of guinea pig and rat myelin basic protein. J. Exp. Med. 146:1323-31 123. Chou, C.-H. J., Fritz, R. B., Chou, F. C.-H., Kibler, R. F. 1979. The immuneresponse of Lewis rats to peptide 68-88 of guinea pig myelin basic protein. I. T cell determinants. J. Immunol. 123:1540-43 124. Happ, M. P., Heber-Katz, E. 1988. Differences in the repertoire of the Lewis rat T cell response to self and non-self myelin basic proteins. J. Exp. Med. 167:502-13 125. Haskins, K., Hannum, C., White, J., Roehm, N., Kubo, R., Kappler, J., Marrack, P. 1984. The antigen-specific, major histocompatibility comple~ restricted receptor on T cells. VI. An antibody to a receptor allotype. J. Exp. Med. 160:452-71 126. Staerz, U. D., Rammensee, H.-G., Benedetto, J. D., Bevan, M. J. 1985. Characterization of a murine monoclonal antibody specific for an allotype determinant on T cell antigen receptor. J. Immunol. 134:3994~4000 127. Zamvil, S. S., Mitchell, D. J., Lee, N. E., Moore, A. C., Waldor, M. K., Sakai, K., Rothbard, J. B., McDevitt, H. O., Steinman, L., Acha-Orbea, H. 1988. Predominant expression of a T cell receptor Vpgene subfamily in autoimmune encephalomyelitis. J. Exp. Med. 167:1586-96 128. Urban, J. L., Kumar, V., Kono, D. H., Gomez,C., Horvath, S. J., Clayton, J.,

Ando, D. G., Sercarz, E. E., Hood, L. 1988. Restricted use of T cell receptor V genes in murine autoimmune encephalomyelitis raises possibilities for antibody therapy. Cell 54:577-92 129. Epplen, J. T., Chluba, J., Steimle, V., Hinkkanen, A. 1988. T cell receptors in autoimmunedisease. J. Cell. Biol. Suppl. 12E: 18 (Abstract R204) 130. Owhashi, M., Heber-Katz, E. 1988. Protection from experimental allergic encephalomyelitis conferred by a monoclonal antibody directed against a shared idiotope on rat T cell receptors specific for myelinbasic protein. J. Exp. Med. In press 131. Behlke, M. A., Chou, H. S., Huppi, K., Loh, D. Y. 1986. Murine T-cell receptor mutants with deletions of//-chain variable region genes. Proc. Natl. Acad. Sci. USA 83:767-71 132. Fritz, R. B., Skeen, M. J. 1987. Influence of the H-2u haplotype on immune function in F1 hybrid mice. II. F1 antiparent mixed lymphocyte reactivity. Immunogenetics 25:161q56 133. Wooley, P. H., Luthra, H. S., Stuart, J. M., David, C. S. 1981. TypeII collagen-induced arthritis in mice. I. Major histocompatibility complex (I-region) linkage and antibody correlates. J. Exp. Med. 154:688-700 134. Banerjee, S., Haqqi, T. M., Luthra, H. S., Stuart, J. M., David, C. S. 1988. Possible role ofVt~T cell receptor genes in susceptibility to collagen-induced arthritis in mice. J. Exp. Med. 167:832-39 135. Banerjee, S., Behlke, M. A., Dungeon, G., Loh, D. Y., Stuart, J., Luthra, H. S., David, C. S. 1988. Va6 gene of T cell receptor maybe involved in type II collagen induced arthritis in mice. FASEBJ. 2(4) (Abstr. 2120) 136. Anderson, G. D., Banerjee, S., Stuart, J. M., Luthra, H. S., David, C. S. 1988. Interaction of MHCand TCRloci in susceptibility to collagen-induced arthritis in mice. FASEBJ. 2(4) (Abstr. 2121) 137. Payne, J., Huber, B. T., Cannon,N. A., Schneider, R., Schilham, M. W., AchaOrbea, H., MacDonald, R. H., Hengartner, H. 1988. Twomonoclonal rat antibodies with specificity for the V~6 region of the routine T cell receptor. Proc. Natl. Acad. Sci. USA 85:7695 138. Makino, S., Kunimoto, K., Muraoka, Y., Mizushima,Y., Katagiri, K., Tochino, Y. 1980. Breeding of a non-obese, diabetic strain of mice. Exp. Anim. 29: 1-13 139. Hattori, M., Buse, J. B., Jackson,

Annual Reviews


TCR IN AUTOIMMUNE DISEASE R. A., Glimcher,L., Dorf, M.E., Minami, M., Makino, S., Moriwaki,K., Kuzuya,H., Imura,H., Strauss, W.M., Seidman,J. G., Eisenbarth,G. S. 1986. The NODmouse: Recessive diabetogenic gene in the majorhistocompatibility complex.Sciene23 I: 733-35 140. Prochazka,M., Leiter, E. H., Serreze, D. V., Coleman,D. L. 1987. Threerecessiveloci requiredfor insulin-dependent diabetes in nonobesediabetic mice. Science 237:286-89 141. Wicker,L. S., Miller, B. J., Coker, L. Z., McNally,S. E., Scott, S., Mullen, Y., Appel,M.C. 1987.Geneticcontrol of diabetesandinsulitis in the nonobese diabetic (NOD)mouse.J. Exp. Med. 165:1639-54 142.Acha-Orbea, H., McDevitt, H. O. 1987. Thefirst external domainof the non-obesediabetic mouseclass II Aa chain is unique Proc.Natl. Acad.Sci. USA84:2435-39 143. Miller, J. B., Appel,M.C., O’Neill,J. J., Wicker,L. S. 1988. Both the Lyt2+ and L3T4+ T cell subsets are requiredfor the transfer of diabetes in nonobesediabetic mice. J. Immunol. 140:52-58 144 Bendelac,A., Carnaud,C., Boitard, C., Bach, J. F. 1987. Syngeneictransfer of autoimmune diabetes from diabetic NODmice to healthy neonates. Requirement for both L3T4÷and Lyt2+ T cells. J. Exp. Med.166:823-32 145. Nishimoto, H., Kikutani, H., Yamamura,K.-I., Kishimoto,T. 1987. Prevention of autoimmuneinsulitis by expression of I-E molecules in NOD mice. Nature 328:432-34 146. Ben-Nun,A., Wekerle,H., Cohen,I. R. 1981. Vaccination against autoimmune encephalomyelitis with T-lymphocyte line reactive against myelinbasic protein. Nature292:60--61 147. Holoshitz, J., Naparstek,Y., Ben-Nun, A., Cohen,I. R. 1983. Lines ofT lymphocytes induce or vaccinate against autoimmune arthritis. Science219: 5658 148. Lider, O., Karin, N., Shinitzky, M., Cohen,I. R. 1987. Therapeutic vaccination against adjuvantarthritis using autoimmuneT cells treated with hydrostatic pressure. Proc.Acad.Sci. Natl. USA84:4577-80 149. Maron,R., Zerubavel, R., Friedman, A., Cohen,I. R. 1983. T lymphocyte line specific for thyroglobulinproduces or vaccinates against autoimmune thyroiditis in mice.J. Immunol.131:2316 150. Lider, O., Reshef,T., Beraud,E., BenNun, A., Cohen, I. R. 1988. Anti-

403

idiotype network induced by T cell vaccination against experimental allergic encephalomyelitis. Science239: 181-83 151. Sun, D., Qin, Y., Chluba,J., Epplen, J. T., Wekerle,H. 1988. Suppression of experimentally induced autoimmune encephalomyelitisby cytolytic T-Tcell interaction. Nature332:843-45 152. Waldor, M. K., Hardy, R. R., Hayakawa, K., Steinman,L., Herzenberg,L. A., Herzenberg, L. A. 1984. Disappearance and reappearanceof B cells after in vivo treatment with monoclonal anti-I-A antibodies. Proc.Natl. Acad. Sci. USA81:2855-58 153. Steinman, L., Rosenbaum, J. T., Sriram, S., McDevitt,H. O. 1981. In vivo effects of antibodies to immune response geneproducts: Preventionof experimental allergic encephalomyelitis. Proc. Natl. Acad.Sci. USA78: 7111-14 154. Adelman, N. E., Watling, D. L., McDevitt, H. O. 1983. Treatment of (NZBx NZW)F1 disease with anti-I-A monoclonalantibodies. J. Exp. Med. 158:1350-55 155. Waldor,M. K., Sriram, S., McDevitt, H. O., Steinman,L. 1983.In vivo therapy with monoclonalanti-I-A antibody suppressesimmune responsesto acetylcholine receptor. Proc. Natl. Acad. Sci. USA80:2713-17 156. Sriram, S., Steinman,L. 1983. Anti-1A antibody suppresses active encephalomyelitis: Treatment model for diseases linked to IR genes. J. Exp. Med. 158:136247 157. Wooley,P. H., Luthra, H. S., Lafuse, W.P., Huse,A., Stuart, J. M., David, C. S. 1985. TypeII collagen-induced arthritis in mice. III. Suppressionof arthritis by using monoclonal andpolyclonalanti-Ia antisera. J. Immunol. 134: 2366-74 158. Boitard, C., Michie,S., Serrurier, P., Butcher, G. W., Larkins, A. P., McDevitt, H. O. 1985. In vivo prevention of thyroid and pancreatic autoimmunity in the BBrat by antibody to class II majorhistocompatibilitycomplex gene products. Proc. Natl. Acad. Sci. USA82:6627-31 159. Sriram, S., Topham,D. J., Carroll, L. 1987.Haplotypespecific suppressionof experimental allergic encephalomyelitis withanti-IAantibodies.J. Immunol. 139:1485-89 160. Vladutiu, A. U., Steinman, L. 1987. Inhibition of experimental allergic thyroiditis in micebyanti-I-A antibodies. Cell. Immunol.109:169-80

Annual Reviews


404


161. Jonker, M., Mitchell, D., Steinman,L. Treatment of EAEby anti-MHCclass II specific monoclonalantibodies in Rhesus monkeys. In Immunologyof HLA,ed. B. Dupond.In press 162. Rosenbaum,J. T., Adelman, N. E., McDevitt,H. O. 1981. In vivo effects of antibodies to immune responsegene products. I. Haplotypespecific suppression ofhumoral immuneresponses with a monoclonalanti-l-A. J. Exp. Med. 154:1696702 163. Perry, L. L., Greene,M. I. 1982. Conversion of immunityto suppression by in vivo administration of I-A subregion-specific antibodies. J. Exp. Med. 156:480-91 164. Williams, I. R., Perry, L. L. 1985. Regulation of transplant immunity in vivo by monoelonalantibodies recognizinghost class II restriction elements. II. Effects of anti-Ia immunotherapy on host T cell responses to graft alloantigens. J. Immunol.134: 2942-47 165. Aoki,I., Idhii, N., Minami,M., Nagashima,Y., Misugi,K., Okuda,K. 1987. Induction of suppressor T cells by intraveneous administration of monoclonal anti-I-A antibodies. Transplantation 44:421-25 166. Broder, S., Mann,D. L., Waldman, T. A. 1980.Participation of suppressor T cells in the immunosuppressive activity of a heteroantiserum to human Ia-like antigens(p23, 30). J. Exp.Med. 151:257-62 t67. Waldor,M. K., Sriram, S., Hardy,R., Herzenberg,L. A., Herzenberg,L. A., Lanier, L., Lim,M., Steinman,L. 1985. Reversal of experimentalallergic encephalomyelitis with a monoclonal antibody to a T-cell subset marker (L3T4). Science 227:415-17 168. Deleted in proof 169. Brostoff, S. W., Mason,D. W.1984. Experimentalallergic encephalomyelitis: successfultreatmentin vivo with a monoclonalantibody that recognizes T helper cells. J. Immunol.133:1938 42 170. Wofsy,D., Ledbetter, J. A., Hendler, P. L., Seaman,W.E. 1985. Treatment of murinelupus with monoclonalantiT cell antibody. J. Immunol.134: 85257 171. Ranges, G. E., Sriram, S., Cooper, S. M.1985. Preventionof type II collagen-induced arthritis by in vivotreatmentwith anti-L3T4.~LExp. Med.162: 1105-10 172. Wofsy,D., Seaman,W. E. 1985. Successful treatment of autoimmunityin

NZB/NZW F1 mice with monoclonal antibody to L3T4.J. Exp. Med. 161: 378-91 173. Christadoss,P., Dauphinee,M.J. 1986. Immunotherapy for myastenia gravis: A murine model. J. Immunol. 136: 2437~40 174. Like, A. A., Biron, C. A., Weringer, E. J., Byman,K., Sroczynski, E., Guberski, D. L. 1986. Prevention of diabetes in Biobreeding/Worcester rats with monoclonalantibodies that recognizeT lymphocytes or natural killer cells. J. Exp. Med.164:1145-59 175. Wang, Y., Hao, L., Gill, R. G., Lafferty, K. J. 1987. Autoimmune diabetes in NODmouse is L3T4T-lymphocyte dependent. Diabetes 36: 53538 176. Koike,T., Itoh, Y., Ishii, T., Ito, I., Takabayashi, K., Maruyama, N., Tomioka,H., Yoshida, S. 1987. Preventive effect of monoclonalanti-L3T4 antibody on developmentof diabetes in NODmice. Diabetes 36:539-41 177. Wofsy,D., Seaman,W.E. 1987. Reversal of advanced lupus in MZB/NZW F1 mice by treatment with monoclonal antibody to L3T4. J. Immunol.138: 324~53 178. Shizuru, J. A., Taylor-Edwards,C., Banks, B. A., Gregory, A. K., Fathman, C, G. 1988. Immunotherapyof the nonobesediabetic mouse:Treatmentwith an antibodyto T-helperlymphocytes. Science 240:659~62 179. Gutstein, N. L., Seaman,W.E., Scott, J. H., Wofsy,D., 1986. Induction of immune tolerance by administration of monoclonal antibody to L3T4. J. Immunol. 137:1127 32 180. Benjamin, R. J., Waldman,H. 1986. Induction of tolerance by monoclonal antibody therapy. Nature 320:449-51 181. Goronzy,J., Weyand,C. M., Fathman, C. G. 1986. Long-term humoral unresponsivenessin vivo, induced by treatment with monoclonal antibody against L3T4.J. Exp. Med.164: 91125 182. Waldor, M. K., Mitchell, D., Kipps, T. J., Herzenberg,L. A., Steinman,L. 1987. Importance of immunoglobulin isotype in therapy of experimental allergic encephalomyelitiswith monoclonal anti-CD4 antibody. J. lmmunol. 139:3660-64 183. Charlton, B., Mandel,T. E. 1988. Progressionfrominsulitis to fl-cell destruc+T tion in NODmouserequires L3T4 lymphocytes.Diabetes 37:1108-12 184. Jacob, C. O., McDevitt, H. O. 1988. Tumornecrosis factor-~ in murine

Annual Reviews


TCR IN AUTOIMMUNE DISEASE autoimmune "lupus" nephritis. Nature 331:356-58 185. Engelman,E. G., Sonnenfeld,G., Dauphinee,M., Greenspan, J. S., Talal, N., McDevitt,H. O., Merigan,T. C. 1981. Treatment of NZB/NZW F 1 mice with mycobacteriumbovis strain BCGor type II interferon preparations accelerates autoimmune disease. Arthritis Rheum.24:1396-1402 186. Heremans,H., Billan, A., Colombatti, A., Hilders, J., DeSomer,P. 1978. Interferon treatment of NZBmice: Ac~ celerated progression of autoimmune disease. Infect. Immunol.21:925 187. Segiescu,D., Cerutti, I., Efthymiou, E., Kahan, A., Chany, C. 1979. Adverse effects of interferon treatmenton the life span of NZBmice. Biomed.Exp. 31:48 188. Parsitol, H.S., Hirsch,R. L., Haley,A. S., Johnson,K. P. 1986.Exacerbations of multiplesclerosis in patientstreated with gamma interferon. Lanceti: 89394 189. Jacob, C. O., van der Meide, P. H., McDevitt,H. O. 1987. In vivo treatment of (NZB×NZW)F1lupus-like nephritis with monoclonalantibody to y-interferon. J. Exp. Med.166:798-803 190. Alvord,E. C., Shaw,C. M., Huby,S., Kies, M. W. 1965. Encephalitogeninduced inhibition of experimental allergic encephalomyelitis: Prevention, suppression and therapy. Ann. NY Acad. Sci. 71:142230 191. Einstein,E. R., Csejty,J., Davis,W.J., Ravch,H. C. 1968.Protective action of encephalitogen and other basic proteins in experimental allergic encephalomyelitis. Immunochemistry5: 567-75 192. Higgins,P. J., Weiner,H. L. 1988.Suppressionof experimentalallergic encephalomyelitisby oral administrationof myelinbasic protein andfragments.J. lmmunol. 140:440-45 193. Sriram,S., Schwartz,G., Steinman,L. 1983. Administrationof myelin basic protein-coupledspleen cells prevents experimental allergic encephalitis.Cell. Immunol. 75:378-82 194. McKenna, R. M., Carter, B. G., Paterson, J. A., Sehon,A. H. 1983.Thesuppressionof experimentalallergic encephalomyelitisin Lewisrats bytreatment with myelinbasic protein-cell conjugates. Cell. Immunol.81:391-402 195. Schoen, R. T., Greene, M. I., Trentham,D. E. 1982.Antigen-specificsuppression of type II collagen-induced arthritis by collagen-coupledspleen cells. J. Immunol.128:717-19

405

196. Chatterjee,S., Bernoco,D., Billing, R. 1982.Treatmentwith anti-Ia and antiblast/monocytemonoclonalantibodies can prolongskin allograft survival in nonhumanprimates. Hybridoma1:69 197. Morrison,S. L., Johnson,M. J., Herzenberg,L. A., Oi, V. T. 1984.Chimeric human antibody molecules. Mouse antigen binding domain with human constantregion. Proc.Natl. Acad.Sci. USA 81:6851-55 198. Boulianne, G. L., Hozumi,N., Shulman,M.J. 1984.Productionof functional chimeric mouse/human antibodies. Nature 312:643-46 199. Roberts, S., Cheetham, J. C., Rees,A. R. 1987. Generation of an antibody with enhancedaffinity and specificity for its bindingby protein engineering. Nature 328:731-34 200. Verhoeyen,M., Milstein, C., Winter, G. 1988. Reshapinghumanantibodies: Graftingan antilysozymeactivity. Science 239:153635 201. Riechman,L., Clark, M., Waldman, H., Winter, G. 1988. Reshapinghumanantibodies for therapy. Nature332:323-27 202. Todd, J. A., Acha-Orbea,H., Bell, J. I., Chao, N., Fronek, Z., Jacob, C. O., McDermott,M., Sinha, A. A., Timmerman,L., Steinman, L., McDevitt, H. O. 1988. A molecularbasis for MHC class II-associated autoimmunity. Science 240:10034 203. Szafer, F., Brautbar, C., Tzfoni, E., Frankek G., Sherman,L., Cohen,I., Hacham-Zadeh, S., Aberer, W., Tappeiner, G., Holubar, K., Steinman, L., Friedman,A. 1987. Detection of disease-specific restriction fragment length polymorphismsin pemphigus vulgaris linked to the Dqw1 and DQw 3 alleles of the HLA-D region. Proc. Natl. Acad. Sci. USA84:6542-45 204. Stamenkovic, I., Stegagno,M., Wright, K. A., Krane, S. M., Amento,E. P., Colvin, R. B., Duquesnoy,R. J., Kurnick, J. T. 1988. Clonal dominance amongT-lymphocyteinfiltrates in arthritis. Proe. Natl. Aead.Sei. USA 85:1179-83 205. Hailer, D. A., Duby,A. D., Lee,S. J., Benjamin,D., Seidman,J. G., Weiner, H. L. 1988. Oligoclonal T lymphocytes in the cerebrospinalfluid of patients with multiple sclerosis. J. Exp. Med. 167:1313-22 206. Oksenberg,J., Bernard, C., King, M. C., Erlich, H., Cavalli-Sforza, L., Steinman,L. 1988.V~andC, alleles of the T cell receptorlinkedto multiplesclerosis and myastheniagravis. Proc.Natl. Acad.Sci. USA.In press






MANIPULATION OF T-CELL RESPONSES WITH MONOCLONAL ANTIBODIES Herman Waldmann ImmunologyDivision, Department of Pathology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QP, United Kingdom

INTRODUCTION T cells are central to all immuneresponses and are therefore a prime target for any therapeutic intervention designed to control immunefunction. For many years polyclonal antilymphocyte and anti T-cell sera have been investigated as immunosuppressive agents. The recent advent of monoclonal antibody technology (1, 2) and the development of a wide range of xenogeneic monoclonal antibodies (Mabs) to T-cell surface molecules have together rekindled interest in using antibodies to regulate the immune response. In animal models therapy with Mabs to T cells has provided new knowledgeon the roles of T cells and their subsets in immunityand immunopathology.In clinical practice we have witnessed the evolution of a newgeneration of therapeutic agents of substantial promise. In theory Mabsmaybe used to enhance, suppress, or alter the quality of an immuneresponse. The outcome maybe restricted to the response to a specific antigen, or to sets of antigens, or it could be directed to the immuneresponse in general. The field of monoclonal antibody therapy is still in its early stages, and mostavailable informationdeals with the global effects of therapy. In time, more knowledgeof the structure and function of the immunesystem maymake it routinely possible to achieve antigen° specific or function-specific regulation. This review, although not comprehensive, summarizes what has been learned about Mabtherapy in animal models, reflects on the impact of Mabs in human transplantation, and speculates on how Mabs may be 407 0732-0582/89/0410-0407502.00

Annual Reviews 408

WALDMANN

used to establish antigen-specific immunologicaltolerancefor therapeutic purposes.


IMMUNOSUPPRESSION Mabs may immpnosuppress by eliminating vital cells or by perturbing the function of critical receptors, adhesion molecules or their respective ligands. The degree to which Mabscan eliminate cells will depend on their capacity to recruit the natural effector systems either humoral (complement) or cellular (e.g. K-cells or other cell types with a range of receptors). The rules by which Mabscan harness these systems are best studied in vitro, and the present state of knowledgeas relevant to therapy is summarized below. The Complement

System

The complementcascade is initiated by the binding of C1 to the antigenantibody complex. This binding occurs through the Clq subcomponent. The bulk component of the cascade is C3 which when bound to cells can result in cell elimination by at least two mechanisms. One involves the assembly of a series of "terminal" components in the membrane,and the other involves clearance by cells of the macrophage/granulocyte series which possess receptors (CD1la/CD18) for bound C3. To exploit or avoid complementactivation, it is essential to knowwhat factors govern the ability of Mabsto activate the process. Information, albeit incomplete, is available for three categories of Mabs: mouse, rat, and human. The major variables seem to be antigen density, antibody isotype, and someas yet ill-defined property of the target antigen itself. ~SOT,~’EThe impact of antibody isotype is ideally demonstrated by the use of class-switch variants or chimeric recombinant antibodies although there is abundant supporting data from myelomaproteins and Mabs. For mouseIgG the subclass hierarchy for lysis seems to run in parallel with Clq binding as mIgG2a= or > IgG2b > IgG3 > IgG1 (3, 4, 5). For human Igs the hierarchy is hu’IgG1 > IgG3 > IgG2 > IgG4 for lysis (6) and IgG3 > IgG1 > IgG2 > IgG4 for Clq binding (6,7). IgG3 less effective than IgGl in the activation of C4 for reasons whichare as yet unclear (8). In the case of rat immunoglobulins,Fust et al (9) demonstrated consumption of hemolytic complement by aggregated myeloma proteins in the order rIgG2b > IgG2a > IgG2c > IgGl. Hughes-Jones et al (10) showedthat rat Mabsof the ’Ig2b subclass were more efficient than other IgG subclasses for Clq binding and lysis. Recently Bruggemannet al (11)

Annual Reviews T-CELLFUNCTIONS .IN VIVO

409


using a set of matchedchimeric molecules have demonstrated a hierarchy for lysis of rIgG2b > IgG1 > IgG2c or IgG2a compared to IgG2b > IgG2c > IgG1 > IgG2a for Clq binding. The reasons for variation in C1 q binding betweenisotypes with identical specificity is unknown as all share a commonClq-binding motif (12). It maybe that interactions between antibody Fc regions mayhelp create a favorable juxtaposition of antibody pairs to create Clq binding sites and also that the different isotypes mayvary for this property. THETARGET ANTIGEN The importance of antigen-specificity in complement activation is documentedby Bindon et al (13). The special feature of "good" antigens was that they permitted Mabs to bind and activate more C1 than did "poor" antigens. Again the reasons are unclear but are consistent with notions of Fc-Fc interactions. The proportion of human lymphocyte-surface molecules that are "good" (i.e. support lysis with humancomplementis small). A particularly effective set of antigens is family of glycoproteins defined by the rat IgM MabCAMPATH-I (14). The ability of CAMPATH-1 antibodies to lyse lymphocytes with human complement has been particularly valuable for purging of T cells from donor marrowin clinical bone marrowtransplantation (BMT)(see later). SYNERGISTIC ANDMONOVALENT MABSIt has been possible to improve complement lysis for some of the "poor" antigens. One approach has involved the use of synergistic pairs of Mabsdirected to different epitopes on the same molecule (10, 15, 16). Synergy works best with pairs of rat IgG2bs or mouseIgG2as or a combination of both (10, 15). One possible therapeutic application where trials are underwayis the use of synergistic pairs of CD45Mabs for eliminating passenger leukocytes from organ grafts. The leukocyte commonantigen defined by CD45Mabs is present on all leukocytes. Passenger leukocytes have been implicated as the major immunogenicelements of tissue grafts (17), and their removal should reduce graft immunogenicity. SomeMabsrapidly redistribute their target antigen on the cell surface (18). This mayproduce resistance to complementlysis. Sometimesthe rate of modulation can be reduced and lysis achieved by use of monovalent Mabs with one active and one inactive F(ab) arm (19). This has shown (Figure 1) for a rlgG2b CD3 Mab (CAMPATH 3) and may widely applicable. The monovalent CD3Mablike its wild-type or OKT3 is active in vivo and can, like its parent, reverse acute rejection episodes in renal transplantation (20). It is not clear howimportant complementactivation is, nor which parts of the pathwayare relevant to the elimination of cells in vivo. A detailed assessment of the in vitro complementactivating properties of particular

Annual Reviews 410

WALDMANN

Y

¥

Biv(~lent

Honov~lent ¯


"Inactive"

¯

~o

100

10

1 0"1 -1 Antibodyconcentration I pg.m[

Fi#ure1 Complement lysis by a monovalentCD3antibody. A hybridomaproducinga rlgG2bCD3Mabwasfusedto a cell line producingan irrelevant light chain. Thethree possiblechaincombinations wereseparatedbyHPLC, andfractionstestedfor lysis of human T-cellblasts withhuman complement by a chromium releaseassay (fromRef. 20).

isotypes and their mutants to defined target antigens will allow assignment of in vivo function with defined residues and will permit selection or creation of the appropriate antibodies for the desired therapeutic effect.

Elimination of Cells by Mechanisms Involving Fc Receptors Knowledgeof the diversity and function of Fc receptors in mouse and humanis accumulating rapidly. However, it is still not clear which Fc receptors either alone or in concert may participate in destruction of cells in vivo. Again the availability of class-switch variants, recombinant antibodies and the mutants thereof, as well as of cloned Fc receptor genes, should make it possible to define which Fc receptors and which accessory cell types are responsible for depletion of lymphocytes and other immunerelated hemopoieticcells in vivo. MOUSE FC RECEPTORS Three distinct

Fc receptors for IgG-Fc have been

Annual Reviews


T-CELL

FUNCTIONS

IN VIVO

411

described (21-23): a high affinity trypsin-sensitive receptor for IgG2a (moFcyRI);a low-affinity trypsin-resistant receptor for mIgG1,mIgG2b, and mIgG2a(mo FCyRII), and a receptor for IgG3 (moFCyRIII). moFC~RI and III are found on cells of the monocyte/macrophagelineage, while Fc~RIIis widely distributed on manycells types. Recently, two genes have been identified as encoding the Fc),RII receptor (21). These encode transmembrane proteins that have near-identical extracellular domains but that differ in the transmembrane and cytoplasmic domains. The FcR~ transcript is expressed only in macrophages,while the FCR/~transcript is found in both macrophages and lymphocytes. Both forms of receptor participate in phagocytosis and are inducible with ~IFN. Although it is uncertain which Fc receptors are responsible for eliminating Mab-coatedcells in vivo, it is widely accepted that antibody-dependent cellular cyotoxicity (ADCC)is probably a good in vitro correlate for those in vivo FcR-dependent effector systems. Where mouse Mabs have been studied in conjunction with mouse effector systems, both moFcRyI and moFcRylIhave been implicated in ADCC (24, 25). There is too little information on interactions of rat lg isotypes with mouseFc receptors to explain whythe rat isotypes vary in their ability to deplete cells in vivo. HUMAN FC RECEPTORS There are at least three receptors on human leukocytes for humanIgGs. The huFc),RI is a high affinity receptor found on cells of the monocyte/macrophageseries and is also inducible on macrophages. HumanIgs are thought to bind in the rank order IgGl = or > than IgG3 >IgG4 >> IgG2 (26, 23). Residue 235 (Leu) on huIgG3 been implicated as crucial to the interaction with huFcTRI(27). A role for huFc3,RI in ADCC has been suggested both from the use of conventional assays and from use of heteroconjugates (i.e. antibodies with dual specificity, whereone of the specificities is directed to the Fc receptor itself). huFcyRIis also 7IFN inducible. The situation for huFcTRII (CDw32)is similar to the mouse. The huFcTRIIis a widely expressed low affinity receptor with a well-defined polymorphism. Three huFc),RII cDNAshave been isolated (28, 29), again with extensive homologyin the extracellular domainsbut with differences in the intraeytoplasmic regions. The binding of humanIgGs ranks huIgG 1 = IgG3 >> IgG2 and IgG4, although there is some uncertainty about the ranking of IgG3 (29, 30). This pattern of reactivity of the different human isotypes with the different Fc receptors suggests that the exact site on humanIg which interacts with huFcTRII may differ from that bound by huFcyRI (30). Like moFcyRII, the huFcTRII binds mIgG2band mIgG1, and it has been suggested that this receptor maybe involved in the T-cell triggering generated by mIgG1 Mabs to human CD3, polymorphism in

Annual Reviews


412

WALDMANN

this receptor accounting for triggering failures in 30%of normal individuals. There is insufficient data concerning interactions with rat Mabs; but as all the rat Ig genes have been cloned (31, 32) and expressed, this information should soon be available. The huFcTRIII(CDI 6) is a low affinity receptor found on neutrophils, eosinophils, natural killer cells, and a subset of T cells. Recently Simmons & Seed (33) have described a eDNAclone encoding this receptor, and observed a close relationship with.the form of the moFC?,RII.The binding hierarchy of hulgGs with, transfected receptor was huIgG3 -- IgG1 >> IgG4 and IgG2, and mIgG3=IgG2a > IgGl >> IgG2b. Several groups have studied ADCCusing human K-cells and murine Mabswith therapeutic relevance in mind. The ranking observed was essentially mIgG3> IgG2a > IgG2b with IgG1 apparently inactive (34, 35). Recent studies with heteroconjugates with one specificity for huFc7RIII showed that neutrophils could also be triggered to ADCCactivity. Hale et al (36) using a wide selection of rat Mabshave shownthat rlgG2b are far more potent than rIgG2a and rIgG2c for ADCC.Recent work with recombinant (chimeric) Mabsconfirms an isotype hierarchy for rIgGs rlgG2b;. IgG1. >> IgG2a > IgG2c (11). Lymphocyte

Depletion

In Vivo

RODENa" MODELS The injection of Mabsin vivo can affect immunefunction by depletion of cells or perturbation of functional molecules. It is desirable to determine what features of Mabsallow cell depletion and ~ whether, depletion is relevant to an observed influence on the immuneresponse. Most data in this area have accumulated from rat Mabsinjected into mice, and this is the major topic of the section. There are four rat I~Gsubclasses; the order of their CHgenes is 72c ~2a ~1 ~,2b (31). Class-switch variants (37) and recombinant antibodies nowbe isolated and should allow detailed analysis of differences of subclass function in vivo without variations due to differences.in fine-specificity. However,the available published data deal with collections of Mabsreactive with the same antigen, but differing in their variable regions, Cobbold(38) and Ledbetter &Seaman(39) first observed that rat IgG2b Mabsdirected to the Thy-1 antigen were able to deplete T cells in vivo. The rIgG2b Mabs were more effective than rIgG2a, IgG2c, or IgM both for.depletion and immunosuppression. Ledbetter & Seaman noticed that T cell depletion with rIgG2a anti-Thy 1 was delayed compared to the IgG2b. Both groups were careful to~ avoid measurement artefacts due to modulation. Le Gros et al (40) described a rIgG2a Mab, a single intraperitoneal injection of which could remove(surprisingly) all detectable T cells from lymphoid organs but not from the thymus. Thierfelder

Annual Reviews


T-CELL FUNCTIONS IN VIVO

413

et al (41) precoated allogeneic marrowand spleen cells with nine different anti-Thy-1. Mabsprior to transplantation, into irradiated recipients. Only two of these Mabs were able to prevent GVHD.Both were rlgG2b, while the ineffective Mabs were IgM (5.Mabs) and IgG2a (2 Mabs). recently Cobboldet al (42) compareda series of class-switch variants and three different IgG2cMabsdirected to Thy-1. Of the class-switch variants, IgG1 and IgG2b could deplete whereas the IgG2a could only modulate the surface antigen. Of the three IgG2cMabs,one could deplete while the others modulated. Three rat IgG2b Mabs to CD5(Lytl), CD4(L3T4), and CD8(Lyt2) have been compared to IgG2a Mabs for acute depletion (43, ,~4, and Qin et al, unpublished data), rlgG2b were always superior. The issue is by no means trivial because, in a number of publications, rlgG2a Mabshave been used with intent to "deplete" lymphocytes, and in somecases (e.g. for Mab53.6 widely used as a CD8Mab)contradictory claims (depletion or not) have been made. It is conceivable that the antiimmunoglobinresponse that manycell-binding Mabselicit may actually "develop" depletion and explain the delayed cell clearance seen by Ledbetter. Resolution of the contradictions is important because eventually one would like to use a Mabto block function without resorting to monotonous preparation of antibody fragments. What can probably be concluded is that rlgG2b and possibly rlgG1 (for which too little information isavailable) are the most effective of rat Mabsfor in vivo depletion of mouse lymphocytes. Synergistic pairs of Mabs may be more effective than single antibodies in vivo. Qin et al (45) demonstrated that pairs IgG2b Mabsto non-overlapping epitopes of the CD4molecule were more lytic in vitro and in vivo than either one was alone. Comparablelevels of lymphocytedepletion are also possible in the rat (46), but the role of mlgGisotypes has been less clearly defined. A clearer picture of the rules for depletion in these,-.rodent modelswill emerge with use of panels of class-switch variants or recombinant Mabs and more knowledgeof the effector, mechanismsresponsible for destruction of antibody coated cells in vivo. SUBHUMAN PRIMATES ANDHUMAN Surprisingly there is little data available to demonstrate effective depletion of lymphocytes by Mabs in humanor other primates. V.ery often lymphocytes are reported coated but not depleted (47), and other times, modulation is-observed. Antigenic modulation is due to a redistribution of antigen on and from the cell surface brought about by interaction with bi- or multi-va!ent antibody .as,,for example, after therapy wj~h CD3and CD5therapeutic Mabs(48). Perhaps the most lytic~..am.~!ymph0cyte.~an.tibodyso far described is the panlymphocyte rIgG2bMab’ C~AMPA~H:IG (37). It is one of a panel

Annual Reviews


414

WALDMANN

rat Mabsof different isotypes but with similar specificity, all knownto lyse lymphocytes with homologous (human) complement in vitro. These exhibit a hierarchy of complement lysis rIgM > IgG2b = IgG2a > IgG~ > IgG2c. The rIgG1 and IgG2b versions were derived as classswitch variants from the rIgG2a form. In a recent study (Figure 2) (49) a patient with an advanced prolymphocytic leukemia was treated sequentially with the IgM; IgG2a, and then the rIgG2b. The IgM Mab was shown to consume hemolytic complement but could produce only a transient clearance of lymphocytes from blood with no long-term depletion, while the rIgG2a was totally ineffective even at very high doses. However the rIgG2b variant produced profound depletion of lymphocytes in blood, spleen and bone marrow with no detectable change in the hemolytic complementlevels. The extent of depletion in subsequent patients was equally profound. Clearly CAMPATH IG (rIgG2b) is unusually lymphotoxic reagent with "debulking" properties that predict a broad application in serotherapy of leukemia, and in immunosuppression. Recently a huIgG1 (CAMPATH-IH)form has been obtained by the

100

25

25

25

25 25 25 25 25

1

2

3

4 5 6

De 198

1

2

3

4

7 8 14

21

36

Day Number Fiyure 2 The effects of Mabs of different isotypes against the CAMPATH1 antigen expressed on lymphocytes. This patient with BCLLwas treated unsuccessfully in 1985 with rlgG2a and the rlgM CAMPATH1. By 1987 the disease had progressed into a prolymphocytic transformation, and further therapy was instituted. Substantial clearance of the blood lymphocytes was obtained. The arrows refer to occasions when the stated number of mgs of Mab were injected (from Ref. 49).

64

Annual Reviews


T-CELL

FUNCTIONS

IN VIVO

415

recombinant DNAapproach (50), and this too has exquisite lymphocytotoxic function in vivo (160). The unusual potency of CAMPATH IG and CAMPATH 1H may depend on features of both the antigen and antibodies. The antigen density on lymphocytes is high (about 5 x 105 molecules per cell); and the antigen is not modulated by antibodies. This and the strong interaction between the Fc regions of rlgG2b or hulgG1, and humaneffector systems seem to guarantee potency. Of course the prospect of debulking the lymphocyte pool in a large animal like humanis a different proposition than for the rodent. In the rodent, short courses of Mabscan achieve more than 2 logs of depletion with striking effects on immunity.The samelevel of depletion in the human would probably only have a small-term impact, given the remaining large pool of cells that could easily be recruited into an immuneresponse. For "good" antigens (like the CAMPATH-1 target) that do not modulate with prolonged therapy, it is very likely that extended therapy possible with humanized Mabs (160) will guarantee maximal use of effector systems with the prospect of high level depletion.

MonoclonalAntibodies as Ablative Ayents to Establish the Functions of T-Cell Subsets The immuneresponse to any antigen has particular cell requirements for its induction and others for its effector phase. Research into the cellular basis of these two phases of a response has traditionally depended upon adoptive transfer and/or in-vitro culture systems. The availability of Mabs that could ablate T cells or subsets has made it possible to use "immunological surgery" to study T-cell function in vivo. In the same way that endocrinologists had been able to infer function from ablation of particular endocrine tissues, it becamepossible to achieve this for T cells. StmSEa’-OEI~LETEO ~CE In 1984 Cobbold et al (43) described a way constructing subset-depleted mice (long-term depleted of CD8, CD4, or all T-cells) by injection of high doses of rlgG2b Mabsinto adult thymectomized mice. Adult thymectomy precluded any de novo reconstitution ofT cells from primary lymphoid sources. Both CD4- and CD8depleted animals remained depleted for at least 100 days, with very little restoration of the function associated with the depleted subset (43, and Qin unpublished). A la carte subset-depleted animals have been used to provide a ready source of CD4-and CD8- cells for in vitro or adoptive transfer studies, and for determining the functions of each subset in the induction of responses to simple antigens, grafts, and microorganisms, and in autoimmunity and immunopathology.It has also been possible to produce rats long-term depleted of CD8cells, but it has been harder to

Annual Reviews


416

WALDMANN

construct rats long-term depleted of CD4cells because the Mab-treated animals seem gradually to reconstitute from the small residual pool of CD4cells (D. Mason, personal communication). The difference between mouseand rat maysimply be a question of the size of the lymphocytepool that Mabsneed to deplete. Probably too few CD4cells are left in a welldepleted mouseto be able to reconstitute the animal. CD4- mice were shown to be unable to make IgG responses to Tdependent antigens (Figure 3) and to viral antigens even after boosting (51, 52); CD4- mice were unable to generate DTHto herpes simplex virus (52) and were less effective at rejecting skin grafts betweenboth whole "major" and multiple "minor" mismatch combinations; they could not easily be primedto exhibit secondset rejection (53). A striking finding with these mice is the variable dependence of CD8 cells for CD4"help." There are manyinstances where CD8T-cell functions could be elicited without CD4T-cell help. For example, it was easier to generate class I-restricted cytotoxic cells to herpes simplex virus in CD4animals than in controls (51). Similarly, the absence of CD4cells did not

LOG 2 Agg[utingtiontitee p~eo,m~

CD5 ~ ~

i~

"-

CD 8 L~

Figure 3 The anti-SRBCantibody responseof thymectomizedmice that have beentreated with various monoclonalantibodies and then challenged with SRBC someweekslater, The antibodies and their isotype are displayed on the ordinate, The total and mercaptoethanolresistant (hatched) antibody titres are shownon the abscissa. Apart from the CD8Maball the r]gG2b Mabswere able to immunosuppress(adapted from Ref. 43). The symbol* indicates significant difference to the positive control; while ** indicates no significant difference from the preimmune serum.

Annual Reviews



417

affect the development of CD8lymphocyte-dependent choriomeningitis with LCMvirus (52). In experimental transplantation, CD4-mice can often reject grafts. This has been shownfor grafts across complete MHC barriers (53), second set rejection of minors(53), resistance to bone marrowgrafts (67, 83), in the rejection of allogeneic tumors (42). The participation of either both of CD4and CD8cells in tumor allograft rejection is consistent with other model systems where both subsets have been separately implicated in rejection of tissue allografts (38, 54). Kauffmannhas highlighted a numberof models ofintracellular bacterial infection where CD4- mice exhibited hitherto unsuspected protective functions for CD8cells in the antibacterial response (55, 56, 57). CD8- mice survive in a remarkably healthy state--a somewhat disconcerting fact for those whohave argued so forcefully for a critical role of lyt2 + cells in immunoregulationand self-tolerance. I-Iumoral responses in such animals were equivalent to controls (58) both in quality and quantity, although very little has been done to examinethe fine details of the response (e.g. the isotype pattern). Sequential antigenic competition, once suggested as mediated by CD8+ cells (59) was as active in CD8mice as in controls (59). Similarly, one cannot persuade nonresponder H-2 (C57BL/10) CD8- mice to become responders to hen egg lysozyme (HEL) after ablation of CD8cells (H. Waldmann,unpublished data). had been suggested that the N-C"suppressogenic" peptide of HELmaintains unresponsiveness to remaining parts of the molecule through suppressive CD8cells (60). Given the large number of complex regulatory circuits involving CD8+ cells that have colored the literature, it wouldbe helpful to establish their importance and relevance in CD8-mice. It is sad that suppressor enthusiasts have not taken the bait. One report from Sedgwick (46) has shown that CD8- animals acquire and can recover from acquired allergic encephalomyelitis(active or passive)just like normal animals. Following spontaneous recovery of CD8-- animals, like their normal counterparts, showed no signs of relapse and remained resistant to further disease induction, ruling out a role for CD8cells in immunoregulation in this disease model. By using mice that were CD4-, CD8-, and both CD4- and CDS-, Cobbold et al (53) were able to demonstrate that both CD4and CD8 cells could contribute collaboratively and also independently to allograft rejection and to GVHD (see also below). ANTIBODY ABLATION IN EUTHYMIC MICEEuthymic mice have the capacity to regenerate new T cells from the thymus. Although parenterally administered Mabsenter the thymus the effector mechanismsnecessary for cell

Annual Reviews 418

WALDMANN


ablation are lacking in that organ. Despite this, functional blockade is still possible. However,acute experinaents that examinethe function of peripheral cells are probably not prone to any changes that might be induced in thymic T cell development. In contrast, long-term administration of Mabsmayhave an effect on differentiation in the thymus, which may in turn influence the quality and repertoire of exported T cells. Indeed, this maybe a useful way of analyzing the stages of T-cell development(61, 62). SUBSETS REJECTING SKINALLOGRAFTS In the early 1980s a great deal of discussion occurred as to whichT-cell subsets were responsible for allograft rejection (63, 64, 65). The work of Mason and Loveland & McKenzie had highlighted an important role for CD4(rat) and lytl + CD8- cells (mouse). In contrast counterarguments maintained that the experimental systems used (adoptive transfer models) were never free of CD8+ cells that could have been the real effectors, especially as studies with cloned CD8+ cells indicated that these could bring about "rejection" if injected directly into the graft (66). The Mabablative approach provided an ecumenical solution--that both T-cell subsets could effect rejection either independently or in collaboration, and that the fine details of strain combinations determined which had the dominant role (53, 67). Cobbold al (67) examined skin graft rejection across complete major and minor differences (BALB/cskin to CBA/Ca)and multiple minors (B10. BRskin to CBA/Ca).Animals received multiple injections of depleting cocktails of CD4, CD8, or both Mabs. In the full mismatch CD8Mabs did not delay rejection while CD4Mabsproduced a significant delay of about 9 days. The combination of both Mabs delayed the rejection process by a further 30 days (Figure 4). This was a clear demonstration that both major subsets of T cells could reject skin grafts independently. In this particular strain combination the natural rate of rejection of the CD4subset was faster than the CD8subset. Similar studies for second set rejection showed that CD8cells played a dominantrole following priming, and that priming was dependent on the presence of CD4cells. Similar experiments in "minor" mismatch combinations failed to uncover an independent role for CD8cells in first set rejection but showedthat both subsets contributed to rapid secondset rejection (53). Graft survival in the primedanimals that had received a short course of both CD4and CD8Mabs was remarkably prolonged, extending far beyond the time needed to reconstitute the depleted T cells. If in the complete mismatchcombinations, Mabswere injected at different stages of the rejection process, it was possible to showthat CO8Mabs alone could delay rejection over a whole MHC difference. It was concluded

Annual Reviews 419

T-CELL FUNCTIONS IN VIVO BALBIcskin --* CBA(HHC÷minors) I/’//.HAb"

//’,/> EL ]

E 8O

Ll_

.-~. I~0 LL

L

~


NoI4A

_0.]z++CD8

o

I

I

i

Figure 4 Both CD4and CD8cells can participate independently in graft rejection. Adult CBA/Camice received BALB/cgrafts and regular injections of antibodies. The CD4therapy significantly delayed the rejection process, but CD8Mabtherapy alone did not. The combination of both Mabswas the most effective producing a substantial delay in graft survival (adapted from Ref. 67).

that once CD4cells had begun to respond to the graft they would nowbe able to help CD8cells enter the rejection process faster than the CD8cells might have done independently. These studies highlighted the opportunistic nature of the rejection process and demonstrated that CD8cells could be triggered without help from CD4cells, but that they could also accept help from CD4cells if this was available. Woodcocket al (68) independently reached similar conclusions for the role of CD4cells in rejection and also showed that CD8cells could be triggered to cytotoxic function without CD4help. Subsequently, Wheelan & McKenzie(69) performed similar ablative experiments although they used rlgG2a Mab53.6.72 as a CD8reagent, and a rat IgG2a MabH129.19 as a CD4reagent. The CD4Mabdespite its isotype was claimed to be extraordinarly effective at depletion (the effect lasting 45 days) as was the particular CD8Mabin their hands. They were able to confirm that both CD4and CD8cells could mediate rejection. For minor mismatches or a class-I MHCmismatch, they demonstrated a dominant and independent role of CD8cells. In the combinationBml2to B6 (class-II difference) they showed that both CD4and CD8Mabscould delay rejection, and that the best immunosuppression could only be obtained with a combination of both Mabs. They failed to achieve any effects on graft survival in mismatches involving both class I and class II. This probably reflected the quality of the particular Mabs used. More recently Auchincloss et al

Annual Reviews 420

WALDMANN

(70) using a pair of rIgG2b Mabswere able to confirm the ability of combination of CD4 and CD8 Mabs to delay MHCmismatched grafts (BALB/c skin onto thymectomized C57/BL6). They too concluded that help for rejection, could be elicited without CD4cells. comparison~w4th ¯ skin grafts, vascularized heart grafts and allogeneic or xenogeneic isletg have been relatively easy to immunosuppress with CD4and CD8Mabs (71-74). Even ifthe mismatch were whole MHCeither of CD4or CD8 Mabscould produce:prolonged graft survival (71)! Bone-marrowgrafts across MHC barriers tend to engraft with difficulty even in lethally irradiated recipients. However,therapy of recipients with CD4and CD8Mabs prevented rejection, even in animals subjected to lower levels of irradiation (600 rads; see Figure 5) (67). Mabsto both the subsets were required. Neither alone was effective. In a multipleminor mismatched combination, engraftment of marrow could be achieved without any irradiation if recipients were pretreated with both CD4and CD8 Mabs but not when treated with any one Mab (75). In marrow rejection, unlike that of skin, CD8cells reactive with minor antigens seem to be able to reject without CD4"help" (75), In general the .results ffom Mabablation complementthe recent data


SURSETS INVOINED IN REJECTION OF OTHER TISSUES ]~y

--donor

marrow rejected--

100-

~.-1,/ / 60- ///

.///

/L//

),~,o ¯

/..// /7./ /L.// /~.-/

~]onor

//I/-~’.gr~ftm~

///, /.~’/

.~...

//1/ //1/

///

369 [ontrol

time from / / transp[ont 369 369 369 {weeks) ~DStr~ted CDktr~led[O4,¢.OStr~ted r~pient recipi~t recipient

/I/

Figure 5 Marrowgraft rejection in irradiated, T-cell subset~lepleted mice. Adult CBA mice were given CD4, CD8, or both CD4and CD8Mabs, irradiated with 600 rads and then given allogeneic T cell~zlepleted BALB/cmarrowand spleen cells. Chimerismwas established at the time periods sh3q~iaS D3norchimerism was (a) virtually complete in recipients that had been conditioned Witfi b0th~antibodies, (b) hardly detectable in CD8treated,-and,,(e) mixed in CD4treated recipidnts (adapted from Ref. 67).

Annual Reviews



421

arising fromadoptive transfer of purified subsets of T cells in defined strain combinations (76, 77). CD4cells are biased to recognize MHC class IIassociated determinants, and CD8cells of class I-associated determinants. The rejection potential of CD4-CD8-T cells is still uncertain, but in the systems discussed above they appeared to have little if any impact. An effective rejection response requires T-cell collaboration. It wouldseem that CD4cells have a special propensity to. give help both to themselves and to CD8cells. However, CD8cells are not always dependent on CD4 help, and there clearly are opportunities for collaborative ventures between CD8cells themselves, which depend upon the particular tissue and the antigens concerned. The immunological issues in resolving the cellular basis of GVHD were similar to those in skin grafting, and in part based on the confusion from use of anti-lyt 1 antibodies as markers for cells that were later defined by CD4Mabs. In 1980 Korngold & Sprent (78, 79) showed that GVHDover certain minor-mismatched combinations was mediated by CD8cells. Mason (80) demonstrated that CD4cells and also, to a lesser extent, CD8cells could mediate GVHD across barriers in the rat. Vallera et al (81) and Korngold&Sprent (82) suggested that lethal GVHD across MHCbarriers involved predominantly lyt 1 + cells and not CD8cells. As all T cells express somelyt 1 (CD5) was possible tht CD8cells were damaged in the complement-dependent lytic procedures adopted. Using opsonizing rIgG2b Mabs to coat the donor marrow, or simply taking donor marrowfrom subset depleted mice, Cobbold et al (83) showedthat CD8cells could be the dominant effectors of GVHDin the combination CBAinto (CBAx BALB/c)F1. They also showed that both subsets could mediate GVHD in a fully allogeneic combination BALB/cinto CBA/Cawhere the recipients had been p.retreated with Mabsto permit engraftment. These results therefore emphasized that both the major T-cell subsets could mediate acute lethal GVHD. In addition, experiments with subset-depleted marrow donors clearly excluded an obligatory need for CD4cells to help CD8cells mediate lethal GVHD.Using purified CD4and CD8cells Korngold & Sprent (84) were also.able to demonstrate differential participation of either CD4or CD8 cells, depending upon t, he.strain combinations used. Hamilton and Korngo!d~ & Sprent-(8.5; 86) have since documenteda variable participation ¯ C-D4cells, in GVHD to non-H2differences, and have also established that (’CD4,.cells mediating lethal GVHD react with MHCclass-II, and CD8cells :- react with class I-associated determinants.The,fagtors ’that.,di.ct.ate which subset dominates in minor or whole MHCmismatched-~ombinations remain unclear. SUBSETS INVOLVEDIN GRAFT-VS-HOSTDISEASE

Annual Reviews 422

WALDMANN


The Interaction of Host-versus-graft Reactions in Marrow Transplantation

and Graft-versus-host

Although the alloreactions (GVHand HVG)that confound BMTtend be studied separately, a successful outcometo therapeutic transplantation requires prevention of GVHD and graft rejection in the same patient. Mousemodels have emphasized the interplay between donor T cells reacting against the host and host T cells that have survived irradiation reacting against the donor marrow. The phenomenonhas sometimes been referred to as "reciprocal interference" because alleviation of GVHD enhances rejection and abolition of rejection enhances susceptibility to GVHD. Cobbold et al (67, 83) showed that failures of engraftment across MHC barriers in irradiated recipients could be overcomeby eliminating residual recipient CD4and CD8T cells. This capacity of donor marrowto damage host and of host T cells to damage donor marrow can have dramatic consequences on the balance of donor vs host chimerism that can be achieved (Figure 5). An even more striking interaction was described Harper et al (87) who demonstrated a requirement for donor CD4cells to interact with recipient CD8cells to produce an SLE-like syndrome resembling chronic GVHD in unirradiated recipients. Ablation of recipient CD8cells abrogated disease and indeed improved donor chimerism. T-CELLSUBSETSIN MICROBIAL INFECTIONS The use of Mab ablation to study the roles of the major T cell subsets in microbial infections was initiated by Nash et al (51) in their studies of herpes virus infections in mice. They observed that CD8ablation abolished class I-restricted cytotoxicity but spared DTHand the capacity to produce HSV-specific antibodies. In contrast CD4-depleted animals were unable to make antiviral antibodies and to mount a primary DTHresponse, but could generate augmentedlevels of class I-restricted cytotoxic cells. The ability to clear virus was greatly compromised by CD4depletion. CD8-depleted mice cleared virus normally in the periphery but could not clear it from the nervous system; this suggests that CD4and CD8T cells mayexert different antiviral effects in different tissues. Moskophidiset al (88) determinedthe nature of effector cells that eliminate LCMvirus. During the effector phase of viral clearance the levels both of cytotoxic activity (CTL)and of cells producing antiviral antibodies were high. Treatment with rlgG2b antiThy-1 and rlgG2b CD4Mabsin this effector phase obliterated cytotoxic function and prevented viral clearance. In contrast CD4ablation abolished antibody production but did not affect viral clearance. Unlike the situation with herpes simplex CD8cells were critical to the effector arm of immunity but CD4cells or antibody were not. Bullet et al (89) observed that CD4 ablation did not prevent mice mounting a CTLresponse and recovering

Annual Reviews


T-CELL

FUNCTIONS

IN VIVO 423

from infection with ectromelia. They were however unable to confirm that CD8cells were responsible for protection because they used a CD8Mab incapable of depleting CD8cells (the rlgG2a Mab53-6.7). The issue of CD4dependence of CD8responses in this context is probably no different from that discussed previously for GVHD and graftrejection. The dependence on CD4help will predictably vary from virus to virus. Sometimesthe damageelicited by viral infection requires host participation. Mabablation can help define the harmful as well as the protective properties of T cells. This is exemplifiedin the case of Theiler’s virus encephalomyelitis produced by intracerebral inoculation. Mice depleted of CD4cells fail to clear virus and die of an acute encephalitis. Welshet al (90) found that depletion of T cells just before the demyelinating phase of the disease substantially lowered incidence of paralysis. Rodriguez & Sriram implicated CD8cells in the demyelination process following infection (91). Traditional dogmawould not have predicted any protective role of CD8 cells in bacterial and protozoal infections. Recently there have been data supporting such a role for CD8cells, much of this data based on the ablative approach (reviewed by Kauffmann;55). Perhaps the most clearcut situation is that reported for immunityfollowing vaccination with malarial sporozoites; the immunitycould be co~npletely eliminated by therapy with CD8Mab but not with CD4Mabs (92). A role for CD8cells has also been claimed for resistance to mycobacteria (56, 57), listeria (93) leishmaniasis (94). Their role in these infections is uncertain. CD4Mab therapy in a mouse model of cutaneous leishmaniasis has revealed a schizophrenic role for CD4cells in genetically resistant and susceptible strains. CD4Mabtreatment of resistant CBAmice rendered them partially susceptible to leishmania major infection whereas treatment of susceptible BALB/cmice rendered them more resistant (95). Liew et have noted an inverse correlation between IL-3 production and resistance, and they have interpreted the data in the context of a balance between two types of CD4T-cells, one that can promote disease and one that is protective. MONOCLONAL ANTIBODY THERAPY OF AUTOIMMUNITY

Mab therapy has

been helpful in understanding the cellular basis of a range of autoimmune disease models in the mouseand rat, and in the derivation of strategies for treatment of humandisease. Steinmanet al (96) demonstrated that anti-I-A antibodies could be used to prevent and reverse EAEin SJL mice, and this work highlighted the potential value of allele-specific anti-I-A therapy. Theyand others showed

Annual Reviews


424

WALDMANN

similar benefit of anti-I-A Mabsin experimental myasthenia gravis and the lupus-like nephritis of NZBmice (97-103). In some of these models the B-cell masswas clearly depleted which mayin part explain the level of immunosuppression. Seaman et al (104) demonstrated that a rIgG2b anti-Thy 1 Mab was able to ameliorate renal disease and lymphoproliferation in MRL/lprmice although the same antibody had little effect on lupus-like disease in NZB/NZW F1 mice (105). In 1984 Brostoff & Mason (106) were able successfully treat a rat model of EAEwith a CD4Mab known to have little depleting ability. Simultaneously Wofsy& Seaman(~07, 108) were able to treat lupus in the NZB/NZW mouse by weekly injections of high doses of rIgG2b CD4Mab. These two publications were followed ~by several that have confirmed the benefit of CD4therapy in models-of collagen-induced arthritis (109), mouse EAE(110), diabetes (111), thyroiditis (112). Clearly the value of CD4Mabtherapy in these disorders is so impressive that it has becomeimperative to establish mechanisms. Carteron, Schiementi, and Wofsy (personal communication), Brostoff al (113), and Steinmann (l 14) have shown that similar effects may obtained with F(ab)2 fragments of CD4Mabs. This shows that. _depletion maynot always be needed to achieve a therapeutic effect. The concept of immunosuppression without depletion is considered further in a subsequent section. A most encouraging outcome to CD4therapy in autoimmunity is the fact that in .two instances of a ."natural disease" prolonged CD4therapy could eventually be stopped without disease recurrence (108, 111). This suggests either that therapy held animals over somecritical risk period or that the immunesystem had somehowchanged to become specifically unresponsive to the culpable antigens. Specific unresponsiveness with CD4 Mabsis considered later in somedept.h. The above results argue for only a minor role of CD8cells in most autoimmunedisorders or alternatively for a critical role of CD4cells alone in initiating the disease. However, Kantwerk et al (115) and Kong al (112) have documentedthe improved therapeutic effect of short-term therapy with both CD4and CD8Mabs, compared to CD4mab alone, in diabetes that develops following low dose streptozotocin or experimental thyroiditis. SELECTIVE STRATEGIES FOR MAB THERAPY IN AUTOIMMUNITYThus

far I

have discussed therapy directed at all CD4cells. In an attempt to focus therapy specifically to the autoimmuneresponse, several laboratories have adopted alternative approaches..One strategy has involved the use of antiIL-2 receptor antibodies expressed somewhatselectively on activated T cells

Annual Reviews



425

(116, 117). These and other activation markers woulddefine T cells that have showntheir cards by participating in the response. If activated cells could be selectively eliminated, then the hope is that newly developing T cells wouldbe subject to natural tolerance mechanismsthat wouldpreclude their involvement in autoimmunity. Although this has been strongly advocated for transplant prophylaxis (117, 118), preliminary data in some models of organ-specific disease are also encouraging (116). Another way to selectively control the reactive T cells has recently been reported by Zamvil et al (119). They show a dominant role of T cells bearing V/~8 the pathogenesis of EAEin the PL/J mouse. Therapy with a V/38 specific Mabwas able to prevent induction of the disease. Recently Acha-Orbeaet al (120) have also been able to reverse ongoingdisease with the same Mab. These data indicate that a T cell receptor-specific approach to therapy can be used with a minimal impairment to overall immunity. These exercises in therapy of autoimmunity have laid downa framework for future research. It isn’t obvious what the advantages of CD4therapy over anti-iA therapy maybe. Advocatesof one argue that the distribution of class-II molecules in vivo is too diverse to be safe; proponents of CD4 therapy voice concern about the risks of infection that may follow ablation of blockade of CD4cells, and the fact that sometimes one may be interfering with a CD4cell with suppressive properties. (CD4Mab ablation of suppressive cells has indeed been demonstrated in mouse modelsof thyroiditis; 121 and 125). As more information becomes available on T cell receptor usage and lymphokine participation in autoimmunity it may be possible to devise some forms of combination therapy that maximize the benefit of inactivating the dominantT cell clones while controlling the others. The ability to achieve immunological tolerance therapeutically will of course be a major goal in the evolution of monoclonal anitbody therapy. In the following section this subject is exploredin somedepth.

MONOCLONAL ANTIBODIES FOR THE INDUCTION OF TOLERANCE Tolerance Induction with Depleting and Nondepleting Antibodies In the course of analyzing the immunosuppressive effects of CD4Mabs, Benjamin& Waldmann(123, 124), and Gutstein et al (124) observed rIgG2b CD4Mabsgiven at higl~ doses produced immunological tolerance to irrelevant rIgG2b immufioglobulifis:’Benjamin et al (122, 126) showed that CD4therapy would also all6w tolerance to humanand rabbit immu-

Annual Reviews


426

WALDMANN

noglobulins if these were given simultaneously (Figure 6). Goronczyet (127) noticed that CD4therapy produced a state of long-term unresponsiveness to sperm whale myoglobin, where even primary challenge 4 weeksafter CD4therapy could not elicit a response. This finding, a point the authors discussed, was not obviously compatible with an interpretation of tolerance, and the nature of the unresponsible state they reported has not been resolved. Qin et al (45) observed that synergistic pairs of CD4 Mabsinjected at doses too low to deplete cells could also induce tolerance to HGG.Gutstein & Wofsy (128) and Benjamin et al (126) demonstrated that F(ab)2 fragments of CD4Mabs were immunosuppressive and also capable of permitting tolerance (126, 129). Coullie et al (130), Charlton al (131), and Qin et al (75) have all shown three independent rIgG2a mabs to be immunosuppressive, and recently S. Qin, H. Waldmann,S. P. Cobbold & Y. N. Kong (unpublished) have shown that a rIgG2a CD4 mabmay also permit tolerance without cell-depletion. Recently Benjamin et al (126) have established the cellular basis of this form of tolerance. Tolerance was induced in T helper cells and not in B cells; CD8cells were unnecessary; nor could "suppressor cells" be detected on adoptive transfer. Most interestingly each challenge with the immunogenic form of test antigen reinforced the state of tolerance so that it could be maintained indefinitely (Figure 7). Finally, they observed that "Mabfacilitated tolerance" was not simply a property of CD4Mabs but could also be induced by a non-depleting LFA-I antibody. The conclusion seems to be that CD4cells may be tolerized when confronted with certain antigens in circumstances where the CD4and

1:102~0

aggregafedHGG

(b)

~ 1:2560 >, 1:6/+0 c 1:160

(Weeks) 6

7

Figure 6 CD4Mab facilitates tolerance induction to HGG.Mice were pretreated with 2 mg CD4Mab (small arrows) or saline and HGG(thick arrow). Mice pretreated with (squares) were unresponsive on rechallenge with aggregated HGG,compared to animals that had had no CD4Mab (diamonds) and those that had received CD4therapy only. panel (b) are shownthe titres (on d56) of anti-FGG from HGGtolerant animals and controls (adapted from Ref. 122).

Annual Reviews T-CELL FUNCTIONS IN VIVO J,

$

30

60

4,

,1,

427

4’

1:102/.0 =1:2560

c 1:160


E

90

120

150

Days aff~r Tolerance Induction Fiyure 7 Antigen challenge reinforces HGGtolerance. All mice were pretreated with 1 mg of CD4Mab on days 0, 1 and 2 and were given HGGor saline as well on day 2. Groups were challenged with aggregated HGGon one of days 30, 60, 90, or 120. After the first challenge the same mice received monthly challenge doses (arrows). The titres were assessed 14 days following each challenge. Exposure to antigen at day 60 elicited no response, but the duration of the animal’s unresponsive state was extended (from Ref. 126).

LFA-1 surface "adhesion" molecules have been perturbed with appropriate Mabs. Similar strategies have recently been applied in attempts to produce transplantation tolerance. Qin et al (132) have used CD4and CD8Mab therapy to avoid GVHDand marrow rejections and have succeeded in achieving something akin to classical transplantation tolerance (133) the adult. Judicious use of appropriate rIgG2b Mabsenabled the establishment ofchimerism across "multiple minor" and class I ÷ minors without the need for any irradiation or chemotherapy (Figure 8). Chimeric animals were able to maintain donor skin-grafts indefinitely. All components of the experiment could be administered together (i.e. antibodies, skin and marrow). In this same model F(ab)2 fragments of CD8Mabswere sufficient to permit tolerance of CD8cells if the CD4cells were ablated. Recently we have also shown (75) that tolerance can be induced in CD4 T cells without any need for their depletion. This was accomplished with high doses of the rIgG2a Mab.If this form of "tolerance therapy" could be extended over stronger antigenic barriers then there would be obvious clinical application in transplantation and correction of autoimmunity. More recently R. J. Benjamin, S. Qin, S. P. Cobbold, H. Waldmann (manuscript in preparation) have succeeded in establishing hemopoietic chimerism and tolerance across complete H-2 differences by using a combination of CD4, CD8, and LFA-1 Mabs together with low dose (300 rads) irradiation. A major goal of future research will be to understand

Annual Reviews 428

WALDMANN Mab+BIO.A BM

[ Mabonly

1


BIO. BM

I

I

I

I

I

Figure 8 "Classical transplantation tolerance" in the adult animal. CBA/CA mice (n = 6) were prepared by injection with 3 mg/mouse of rlgG2b CD4and CD8. B10.A marrow was transplanted on the first day of treatment and CD4and CD8Mabsover 5 days. Three weeks later donor skin was transplanted. The group that had received both marrow and Mabs becametolerant to B 10.A skin. Animalsreceiving either one alone did not (taken from Ref.

132).

and obviate the need for irradiation to be able to use BMTto achieve transplantation tolerance in the adult. This work demonstrates that CD4/CD8Mabtolerance therapy is not only confined to special protein antigens but is potentially exploitable in transplantation. A more complete understanding of the cellular mechanisms underlying tolerance and immunity may permit the evolution of rationale "tolerance therapies" using Mabs together with other immunosuppressive drugs. In trying to moveto such rational therapies we have proposed a hypothesis of how tolerance may arise in normal development and how monoclonal antibodies maybe used to simulate this process in the adult (134). There are two underlying assumptions to the hypothesis. The first requires that T cell receptor occupancyand the signal provided by recognition of peptide-MHCbe equivalent for both tolerance and immunity. The second assumption is that the critical decision (ONor OFF) is decided by secondary signals in a Bretscher-Cohn manner (135). Wewould suggest that these costimulatory signals arise through multiple cell interactions with each cell contributing some level of cumulative "help" which ultimately reaches a threshold which pushes the collaborative unit and each of its individual cellular componentsto "ON"(i.e. activation). By inference a cell that has adequately bound antigen but is isolated from collaborating

Annual Reviews


T-CELL

FUNCTIONS

IN VIVO

429

partners cannot receive costimulatory signals and so defaults to "OFF". In other words the choice "ON" or "OFF"is dictated by the frequencies of antigen-reactive cells, and the chance that these can find each other and form a collaborative unit. For a potent antigen with manyforeign epitopes the chance of such an event would be high. However, any circumstances where antigen-specific T cells might be isolated in space or time would predispose to tolerance. By this token we can explain the ease of tolerance induction in the neonatal mouse, the irradiated animal, and the animal depleted of lymphocytcs by thoracic duct drainage or by administration of ALG. What of normal ’self-tolerance’? Here we have made the analogy of lemmingsfalling over a cliff of self-antigen. In essence we argue that the thymus, the environment where T-cell tolerance is largely generated, is a relatively helpless one. Nosooner can a T cell express an anti-self receptor than it comesto the antigen-cliff and "jumps" or deletes. In this waythere will be a perpetual loss of self-reactive cells one by one. As a result no helpful cells could accumulate to provide costimulator activity to the incoming self-reactive set, and consequently these newcomersdie. The situation would be different in the peripheral lymphoidtissues following exposure to a foreign antigen. MatureT cells are continuously recirculating from blood to lymph. Antigens with multiple epitopes are likely to successfully recruit sufficient specific T cells to generatean active collaborative unit where all the necessary requirements for triggering and growth are met. As a result tolerance is prevented. To someextent bystander responses to other antigens in the vicinity wouldalso tend to prevent tolerance. It can be seen that this hypothesis views tolerance as a process that is normally "prevented" in the encounter with a foreign antigen. Its relevance to Mab therapy is based on the philosophy that Mabs may be used to ALLOW natural tolerance to just happen. Let us consider how tolerance is facilitated by CD4, CD8, or LFA-1 Mabs. If therapy has depleted the CD4cells, then the numberof residual antigen-specific T cells maybe too small to form a collaborative unit allowing tolerance to occur in remaining cells. If antigen persists or is repeatedly injected, then newly formed T cells would be tolerized as they appear. Bone marrowis a self-replicating form of antigen and therefore donor antigen would persist. Protein antigens like HGGwould eventually be cleared to a subtolerogenic level. AnyT cells emergingfrom the thymus at this point wouldnot be tolerized but wouldbe tolerizable if any reexposure to antigen occurred before sufficient T cells had been replaced. This would adequately explain the reinforcement phenomenon(126) described before, without recourse to implicating "suppressor cells" (whatever they

Annual Reviews


430

WALDMANN

are). Let me develop the argument further to explain whyit should be Mabs like CD4, CD8, LFA-1, and not many others, that can facilitate tolerance. It is important to realize that blockadeof T cell function is also a subtle wayof isolating T cells from potential collaborative units. While blockaded by CD4and CD8, Mabs T cells simply ignore antigen, because both these two molecules are required to guarantee proper T cell receptor occupancy by antigen (peptide-MHC). Any T cell that had evaded depletion wouldstill be subject to Mabblockade. It is only whenT cells emerge from Mabblockade (e.g. after therapy is stopped) that tolerance may occur. This exodus from blockade will be a gradual process which might result in a staggered exposure of T cells to persisting antigen. The otherwise competent T cell would find itself isolated and "helpless" and consequentlytolerance susceptible. Tolerance wouldarise as a result of too few T cells being available to establish a collaborative unit. For complex antigens or antigens that are rapidly cleared it maybe hard to achieve this staggered exit from blockade, and so immunity would not be prevented. On this model, depletion would not be essential if the blockade and exit therefrom were carefully orchestrated. It is likely that the sameprinciples might apply to tolerance with LFA-1and anti-IL2 receptor Mabs. In the latter case one source of help would be compromised. This model would argue that interference with other growth factor receptors and other adhesion molecules would also increase the chance of isolating antigenspecific T cells from "help" and thus enable the induction of tolerance. Future Prospects for Tolerance Therapy The basis of the abovehypothesis is that a T cell is tolerizable if it can be isolated from other helpful cells while contacting antigen on appropriate APCs.The ease of such an isolation will depend upon a numberof factors. For example memorycells may be more frequent, may be upregulated for desirable adhesion molecules, and maybe poised to express other critical growth factor receptors in a way that increases chances of effective collaborations. It maybe possible to use combinations of methods to try to protect or isolate memorycells from sources of help to determine if they too are tolerizable. If memory cells are tolerizable then "tolerance therapy" with minimal cell depletion may be a therapeutic goal in autoimmunity. In translating the rodent work to human we should remember that the starting T cell pool is many orders of magnitude greater in humans. Consequently, the task of preventing collaborative events will be that much harder. Controlled debulking of T lymphocytes with antibodies like the humanized form of CAMPATH-1 may establish a platform onto which nondepleting antibodies directed to CD4, CD8, LFA-I and IL-2 receptor maybe effectively exploited for tolerance induction.

Annual Reviews T-CELL

FUNCTIONS

IN VIVO

431


THE ANTIGLOBULIN RESPONSE TO THERAPEUTIC ANTIBODIES It is a little disconcertingthat experimentalanimalsandpatients can make such good antiglobulin responses towards the very same immunosuppressiveMabswith whichthey are treated. These responses can be fast andcan be directed to both constant and idiotypic determinants(136). Clinically they are probablyundesirable,if oneis to avoidsensitization to future Mabtherapy, let alone to guaranteea prolongedperiod of treatment. Benjaminet al (123) showedthat a large numberof rat Mabs mousehemopoietic cells were very immunogenic.However,high doses of rIgG2b CD4Mabs were completely non-immunogenicand induced tolerance to rIgG2bimmunoglobulins (123, 124). This state of tolerance to a therapeutic antibody undoubtedlybenefited the Wofsy(108) and Shizuru(111) studies on long-term therapy of mouselupus and diabetes. Micetolerant to rIgG2bwereable to generate antiidiotypic responsesto other rIgG2bMabsdirected to cell surfaces but not to non-cell binding Mabs(123) (Figure 9). There are somefundamentalquestions to be sweredhere. Whatis the cellular basis for the anti-idiotypic response? Whatdeterminantsare recognizedby helper T cells? Whyis it so hard to tolerize to idiotypic determinantson cell-binding antibodies but not to constant region determinants? More recently Benjamin et al (126) have shown that an immunosuppressive rat LFA-1Mabis also non-immunogenic in mice, so "silence" is not a property unique to CD4Mabs. 1:102k0 ,~, 1:2560 °.2 1:6~0

~

~ 2~

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

Recommend Documents