Advances in Cancer Research, Volume 27

ADVANCES IN CANCER RESEARCH VOLUME 27 Contributors to This Volume J. I. Brewer Martin Lipkin C. Dean Buckner Paul ...

Author: George Klein | Sidney Weinhouse

11 downloads 732 Views 22MB 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

ADVANCES IN CANCER RESEARCH VOLUME 27

Contributors to This Volume J. I. Brewer

Martin Lipkin

C. Dean Buckner

Paul E. Neiman

Sushilkumar

G. Devare

Fred H. Reynolds, Jr.

Alexander Fefer

John Roboz

B. Halpern

C. R. Stanhope

Ronald B. Herberman

John R. Stephenson

Harold B. Hewitt

Rainer Storb

Howard T. Holden

E. Donna11 Thomas

B. D. Kahan

E. E. Torok Alice S. Whittemore

ADVANCES IN CANCER RESEARCH Edited by

GEORGE KLEIN Department of Tumor Biology Karolinska lnstitutet Stockholm, Sweden

SIDNEY WEINHOUSE Fels Research Institute Temple University Medical School Philadelphia, Pennsylvania

Volume 27-1978 ACADEMIC PRESS

New York San Francisco London

A Subsidiary of Harcourt Brace Jovanovich. Publishers

0

COPYRIGHT 1978, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART O F THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY. RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM T H E PUBLISHER.

ACADEMIC PRESS, INC.

111 Fifth Avenue, New York, N e w York 10003

United Kingdom Edition published by ACADEMIC PRESS, MC. (LONDON) LTD. 24/28 Oval Road, London N W 1 7 D X

LIBRARY OF CONGRESS CATALOG CARD NUMBER:52- 13360 ISBN 0-12-006627-0 PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS CONTRIBUTORS TO VOLUME27 ..............................................

ix

Translational Products of Type-C RNA Tumor Viruses JOHNR. STEPHENSON. SUSHILKUMAR G. DEVARE. AND FREDH . REYNOLDS. JR. I . Introduction ......................................................... I1. Type-C Viral Genome Structure and Complexity ........................ 111. Proteins of Type-C Tumor Viruses ..................................... IV. Genetic Mapping of the Type-C Tumor Virus Genome .................. V. Relatedness of Structural Proteins Coded for by Leukemia and Sarcoma Type-C Viral Genomes ......................... VI. Type-B and T4pe-D Oncomavirus Structural Proteins ................... VII . Summary and Conclusions ............................................ References ...........................................................

1 5 6 27

35 39 42 43

Quantitative Theories of Oncogenesis ALICE S . WHITTEMORE

I . Introduction .......................................................... I1. Expected Rates of Tumor Appearance .................................. 111. The Single Stage Theory of Iversion and Arley ........................ IV. The Multicell Theory of Fisher and Holloman .......................... V. The Multistage Theory with Negligible Cell Loss ....................... VI. The Multistage Theory with Non-Negligible Cell Loss .................. VII . The Multistage Theory with Proliferative Advantage of Intermediate Cells .................................................... VIII. Single Stage of Multistage Theory with Variation in Transformed Cell Types ............................................... IX . Implications for Dose-Response Relationships .......................... X . Conclusion ........................................................... References ........................................................... V

55 56 57 62 65 68 73 78 83 86 87

vi

CONTENTS

Gestational Trophoblastic Disease: Origin of Choriocarcinoma, lnvasive Mole and Choriocarcinoma Associated with Hydatidiform Mole, and Some Immunologic Aspects J . 1. BREWER,E. E. TOROK, 8. D. KAHAN, C. R. STANHOPE,AND B. HALPERN I. The Origin of Gestational Choriocarcinoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. lnvasive Mole and Choriocarcinoma Associated with Hydatidiform Mole ................................................... I I I. Immunobiology of Trophobl astic Disease ............................... References ...........................................................

89 125 138 145

The Choice of Animal Tumors for Experimental Studies of Cancer Therapy HAROLDB. HEWITT I. Introduction

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

11. Analysis of Species Used in Current Cancer Research . . . . . . . . . . . . . . . . . . . 111. Origin and ?ifaintenance of Animal Tumor Systems i n

149 153

Relation to Their Validity as Xlodels of Clinical Cancer. ................. 157 192 I\'. Reflections and Conclusions ........................................... 196 References ...........................................................

Mass Spectrometry in Cancer Research JOHN

ROBOZ

I. Scope of Applications and Analytical Techniques ....................... 11. Identification, Quantification, and Metabolism of Carcinogens . . . . . . . . . . . . 111. Metabolism and Monitoring of Antineoplastic Agents . . . . . . . . . . . . . . . . . . . IV. Biological Markers .................................................... References ...........................................................

202 2 15 233 253 260

Marrow Transplantation in the Treatment of Acute Leukemia E. DONNALL THOMAS,C. DEAVBUCKNER,ALEXANDERFEFER, PAUL E. NEIMAN,AND W N E R STORB I. introduction

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

............ 111. Analysis of Survival ................................................... 11. Patient Selection, Methods. and Surnniary of Clinical Results

269 270 271

CONTENTS

I V. Nature of Recurrent Leukemia ......................................... V. Efforts to Prevent Leukemia Relapse ................................... VI. Graft versus Leukemia ................................................ VII . Transplantation in Remission .......................................... VIII . Conclusions .......................................................... References ...........................................................

vi i 273 275 276 277 278 278

Susceptibility of Human Population Groups to Coion Cancer MARTINLIPKIN

I . Introduction ......................................................... I1 . Role of Environment in Increasing the Susceptibility of

Individuals to Colon Cancer ........................................... 111. Inherited Diseases that Increase Susceptibility to Colon Cancer ......... I V. Proliferative Abnormalities and Susceptibility to Colon Cancer .......... V. Newer Immunologic Studies .......................................... VI . Nuclear Protein and Enzyme Alterations ............................... VII . Studies of Cutaneous Cells ............................................ VIII . Examination of Fecal Contents ........................................ IX . Conclusion ........................................................... References ............................................................

281 282 287 293 296 296 299 300 300 301

Natural Cell-Mediated Immunity RONALD B . HERBERMAN AND HOWARD T. HOLDEN I . Introduction .......................................................... I1 . Characteristics of Natural Cytotoxicity ................................. 111. Specificity of Natural Cell-Mediated Cytotoxicity ....................... IV. Nature of Effector Cells ............................................... V. Relationship of Natural Cell-Mediated Cytotoxicity to Antibody-Dependent Cell-Mediated Cytotoxicity ....................... VI. Model for Placement of NK and K Cells in Pathway of Differentiation of T Cells .............................................. VII . Discrimination between Natural Cell-Mediated Cytotoxicity and Cytotoxicity by Other Effector Cells ............................... VIII . In Viuo Relevance of Natural Cytotoxicity .............................. References ...........................................................

305 307 324 333

SUBJECTINDEX ............................................................. CONTENTSOF PREVIOUSVOLUMES ..........................................

379 383

345 351 354 361 370

This Page Intentionally Left Blank

CONTRIBUTORS TO VOLUME 27 Numbers in parentheses indicate the pages on which the authors’ contributions begin.

J . I. BREWER,Department of Obstetrics and Gynecology and the Cancer Center, Northwestern University Medical School, Chicago, Illinois 60611 (89) C. DEANBUCKNER,The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 and The Department of Medicine, Division of Oncology, University of Washington School of Medicine, Seattle, Washington 98195 (269) SUSHILKUMAR G. DEVARE, Laboratory of RNA Tumor Viruses, National Cancer Institute, Bethesda, Maryland 20014 (1) ALEXANDER FEFER,The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 and The Department of Medicine, Division of Oncology, University of Washington School of Medicine, Seattle, Washington 98195 (269) B. HALPERN, Department of Obstetrics and Gynecology and the Cancer Center, Northwestern University Medical School, Chicago, Illinois 60611 (89) RONALDB. HERBERMAN,Laboratory of Immunodiagnosis, National Cancer Institute, Bethesda, Maryland 20014 (305) HAROLDB. HEWITT,Department of Morbid Anatomy, King’s College Hospital Medical School, London S.E.S., England (149) HOWARDT. HOLDEN,Laboratory of Immunodiagnosis, National Cancer Institute, Bethesda, Maryland 20014 (305) B. D. KAHAN,* Department of Surgery, Northwestern University Medical School, Chicago, Illinois 60611 (89) MARTIN LIPKIN,Memorial Sloan-Kettering Cancer Center, New York, New York 10021 (281) PAUL E. NEIMAN,The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 and The Department of Medicine, Division of Oncology, University of Washington School of Medicine, Seattle, Washington 98195 (269) * Present address: Divisions of Organ Transplantation and Immunology, Departments of Surgery and Biochemistry, University of Texas Medical School at Houston, Houston, Texas 77030. ix

X

CONTRIBUTORS TO VOLUME

27

FREDH. REYNOLDS, JR., Viral Oncology Program, Frederick Cancer Research Center, Frederick, Maryland 21 701 ( 1) JOHN ROBOZ, Department of Neoplastic Diseases, Mount Sinai School of Medicine, The City University of New York, New York, New York 10029 (201) C . R. STANHOPE,Department of Obstetrics and Gynecology and the Cancer Center, Northwestern University Medical School, Chicago, lllirwis 60611 (89) JOHN R. STEPHENSON,Laboratory of RNA Tumor Viruses, National Cancer Institute, Bethesdu, Maryland 20014 ( 1 ) RAINER STORB,The Fred Hutchinson Cancer Research Center, Seattle, Wushington 98104 and The Department of Medicine, Division of Oncology, University of Washington, Seattle, Washington 98195 (269) E. DONNALL THOMAS, The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 and The Department of Medicine, Dioision of Oncology, University of Washington, Seattle, Washington 98195 (269) E . E. TOROK,Department of Obstetrics and Gynecology and the Cancer Center, Northwestern University Medical School, Chicago, lllinois 60611 (89) ALICE S. WHITTEMORE,Department of Environmental Medicine, New York University Medical Center, New York, New York (55)

ADVANCES IN CANCER RESEARCH, VOL. 27

TRANSLATIONAL PRODUCTS OF TYPE-C RNA TUMOR VIRUSES

John R. Stephenson, Sushilkurnar G. Devare, and Fred H. Reynolds, Jr. Laboratory of RNA Tumor Viruses, National Cancer Institute, Bethesda, Maryland and Viral Oncology Program, Frederick Cancer Research Center, Frederick, Maryland

I. Introduction .................................................... 11. Type-C Viral Genome Structure and Complexity.. ................ 111. Proteins of Type-C RNA Tumor Viruses ................................. A. RNA-Dependent DNA-Polymerase .................................. B. Structural Proteins .......................... .... ...... C. Src Gene-Coded Transforming Protein(s) ............................ IV. Genetic Mapping of the Type-C RNA Tumor Virus Genome . . . . . . . . . . . . . . A. Location of gag, pol, and env within the Viral Genome. ............... B. Intracistronic Mapping of the gag Gene of a Prototype Mammalian ............... Type-C Virus .................................. C. Identification of Functionally Analogous gag GeneMammalian Type-C Virus Isolates of Diverse Origin .................. V. Relatedness of Structural Proteins Coded for by Leukemia and Sarcoma ........... .......... Type-C Viral Genomes ....... VI. Type-B and Type-D Oncornavirus Structural Proteins ............. VII. Summary and Conclusions .... ........... References . . . . . . . . . . . . . . . . . . . ...........

6 7 8 24 27 27 30 31 35

I. Introduction

The existence of oncornavirus genetic sequences in a naturally integrated state within the cellular genome of a broad spectrum of vertebrate species is well established (Lieber and Todaro, 1975; Aaronson and Stephenson, 1976). Release of infectious virus particles, while generally repressed, can occur both spontaneously (Hartley et al., 1969; Aaronson et al., 1969; Stephenson and Aaronson, 1972b; Lieber et al., 1973) and following treatment with chemicals (Lowy et al., 1971; Aaronson et al., 1971b). Following activation, infectious virus may be transmitted horizontally both between individual animals of the same species (Hardy et al., 1973; Jarrett et al., 1973) as well a5 to other species (Benveniste and Todaro, 1976).The association of infectious oncornavirus particles with lymphoid tumors of many species has also been demonstrated (Gross, 1959; Lilly et al., 1975; Essex, 1975). Moreover, there is accumulating evidence that elevated endogenous virus expression may be an important determinant of host susceptibility to neoplastic transformation (Niman et al., 1977). 1 Copyright @ 1978 b y 4rademic €‘re\\,Inc

All light\ nf ieproduction in any toim re\ewed l5HN 0-12-006627-0

2

JOHN R. STEPHENSON ET AL.

Studies of the translational products of oncornaviruses have been initiated in efforts to gain insight into the role that these viruses may have both in normal cellular processes and in the etiology of tumors of their natural hosts. As early as 1958, Bernhard proposed a classification scheme for the diverse group of RNA viruses now included under the general term " oncornavirus". According to this system, RNA tumor viruses are designated as type-A, type-B, type-C (Bernhard, 1958, 1960), or type-D (Dalton et al., 1974) primarily on the basis of morphologic criteria. Intracellular virus-like particles occurring in a variety of mouse tumors have been designated type-A (Dalton et al., 1961). These are distinguished from other oncornaviruses mainly by virtue of their association with the endoplasmic reticulum rather than plasma membrane (Dalton, 1962). The second class of oncornaviruses, designated as type-B, have eccentrically located nucleoids and their envelope possesses characteristic projections or spikes (Sarkar et al., 1972). While mouse mammary tumor virus (MMTV), the prototype virus of this group, has been studied extensively, much less information is currently available regarding type-B viruses of other species of origin, such as the guinea pig (Opler, 1967; Nadel et al., 1967) and domestic cow (Miller et al., 1969; Van Der Maaten et al., 1974). The possibility that type-B particles may represent maturational products of intracytoplasmic type-A particles has been suggested on the basis of apparent similarities in the immunologic properties of their major structural proteins (Sarkar and Dion, 1975; Tanaka, 1977). The most extensively studied class of oncornaviruses are the type-C RNA tumor viruses. This group of viruses is characterized by their centrally located nucleoid and a pattern of virion assembly which occurs as a budding process at the plasma membrane (Sarkar et al., 1972). Type-C oncornaviruses can be distinguished from either type-B and type-D viruses on the basis of both morphologic criteria (Bernhard, 1958; Dalton et al., 1974) and the divalent cation preference of their RNA-dependent DNA-polymerase (Scolnick et al., 1970; Howk et al., 1973; Abrell and Gallo, 1973).In addition, many type-C virus structural proteins can be readily distinguished from those of type-B and type-D viruses. Moreover, a number of structural proteins of all type-C oncornavirus isolates examined to date have been found to share crossreactive interspecies antigenic determinants (Gilden, 1975; Stephenson et al., 1977b), and the major structural proteins of type-C isolates of several species have been shown to exhibit extensive regions of primary structure homoIogy (Oroszlan et al., 1975, 1976). Another characteristic property of type-C RNA tumor viruses is their unique

RNA TUMOR VIRUS TRANSLATIONAL PRODUCTS

3

ability to provide helper functions for replication-defective sarcoma viruses (Hartley and Rowe, 1966; Huebner, 1967; Sarma et al., 1970; Aaronson and Rowe, 1970). In view of the fact that type-C viruses represent the major emphasis of the present review, the origins of many of the presently available isolates are summarized in Table I. It should be noted that in several instances endogenous type-C viruses of one species were transmitted to and became stably associated with the germ line of a second species (Benveniste and Todaro, 1974, 197513). In fact, the majority of type-C virus isolates can be traced back to two main lineages of ancestral viruses, one of rodent origin and the second, endogenous to primates. An understanding of the relatedness of different type-C virus isolates is important in the evaluation of much of the currently available information regarding properties of their structural proteins. Type-C viruses of a number of mammalian species, including endogenous viruses that have existed within the pig genome for millions of years (Benveniste and Todaro, 1975b), as well as a group of infectious horizontally transmitted isolates of gibbon apes (Kawakami et al., 1972) and a woolly monkey isolate (Theilen et al., 1971) are all related to known endogenous mouse type-C virus isolates and appear to be TABLE I MAMMALIAN TYPE-C ONCORNAVIRUSES Species of origin Rodent Mouse Mus musculus Mus caroli Mus cervicolw Rat Rattus norvegicus Hamster Cricetulus griseus Carnivores Cat Felis catus

Felis sylvestris ArtiodactyIs Pig Sus scrofa Deer Odocoileus hemionus Primates Baboon Papio cynocephalus Papio hamadyas Gelada Theropithecus gelada Woolly monkey Lugothrix spp. Gibbon ape Hylobates lar

Prototype virus isolate

Ancestral origin

R-MuLV, AKR-MuLV, etc. CERO CI CERV CI, CII RT 21C, SF-1, RMTDV CCL 14.1

Rodent Rodent Rodent Rodent Rodent

RD 114 FeLV FS-1, WCV-1

Primate Rodent Primate

CCL-33, PK(15) DKV

Rodent Unknown

M7, M28, BAB8-K BILN TG-1-K SSAV-1 GALV

Primate Primate Primate Rodent Rodent

4


evolutionarily related to ancestral mouse viruses (Lieber et al., 197513). While other endogenous rodent viruses, such as those of hamster (Graffi et al., 1968; Kelloff et al., 1970) and rat (Bergs et al., 1970) origin, have not been as well studied, these also appear to constitute a highly related group (Benveniste and Todaro, 1975a). In addition, feline leukemia virus (FeLV), a horizontally transmitted type-C virus of cats, has been shown to possess significant nucleic acid sequence homology with, and was apparently derived from an endogenous rodent virus (Benveniste and Todaro, 1975a). Endogenous type-C viruses of Old World monkeys, apes, and possibly man constitute the second major lineage of mammalian type-C viruses. While isolation of infectious viruses of this group have been limited to baboon species of the genus Papio (Todaro et al., 1976; Stephenson and Aaronson, 1977), the presence (Benveniste and Todaro, 1976) and partial expression (Stephenson and Aaronson, 1977)of related nucleic acid sequences within the genomes of a much broader range of Old World primates has also been demonstrated. A class of endogenous feline viruses, the prototype of which is designated RD114 (McAIlister et al., 1972), are apparently of primate origin, having entered the germ line of an ancestral cat 20-30 million years ago (Benveniste and Todaro, 1974). In addition, there is suggestive evidence that a less well-characterized group of type-C viruses, endogenous to ungulates may be somewhat more closely related to primate than to rodent viruses (Aaronson et al., 1976; Tronick et al., 1977). The fourth major group of oncornaviruses, designated as type-D (Dalton et al., 1974)were described subsequent to the original oncornavirus classification proposed by Bernhard. These particles are somewhat larger in size than type-B or type-C viruses and have pleomorphic bullet-shaped nucleoids. The prototype isolates of this cIass include the Mason-Pfizer monkey virus (MPMV) (Chopra and Mason, 1970; Kramarsky et al., 1971) and a recently reported endogenous virus of the langur (Todaro et al., 1977a). A number of oncornavirus isolates of squirrel monkey origin have also been tentatively classified as type-D viruses (Heberling et al., 1977; Todaro et al., 1978). In addition to the four classes of oncornaviruses summarized above, there is a category of RNA tumor viruses generally known as RNA sarcoma viruses. These are replication-defective, transforming viruses which appear to have arisen as a result of genetic recombination between type-C viral and host cell genetic sequences (Scolnick et d., 1973, 1975; Frankel and Fischinger, 1977). Mammalian sarcoma isolates studied to date, while competent for transformation, have invariably been found to require type-C leukemia helper viruses for their


5

replication (Hartley and Rowe, 1966; Huebner, 1967; Aaronson and Rowe, 1970). Isolates of this group of viruses have been restricted to four mammalian species; these include two rodents, mouse (Moloney, 1966; Levy et al., 1973) and rat (Harvey, 1964; Kirsten and Mayer, 1967), one carnivore species, cat (Snyder and Theilen, 1969; Gardner et at., 1971),and one primate, woolly monkey (Wolfe et al., 1971). II. Type-C Viral Genome Structure and Complexity

The single-stranded type-C viral genomic RNA has a sedimentation coefficient of about 70 S, corresponding to an estimated molecular weight of approximately 1.2 x lo7 (Robinson et al., 1965; Duesberg, 1968; Montagnier et al., 1969). In addition, smaller RNA species with sedimentation values of 4 S and 7 S are found within the virion (Bishop et al., 1970a,b). Denaturation of the 70 S genomic RNA leads to production of two 35 S RNA subunits (Duesberg, 1968). On the basis of electrophoretic mobility (Duesberg and Vogt, 1973) and end-group analysis (Keith et al., 1974), a molecular weight of about 3 x lo6 was derived for each 35 S subunit. That the viral genome is polyploid and all 35 S subunits are similar in their sequence has been demonstrated by oligonucleotide fingerprinting analysis using ribonuclease TI (Billeter et al., 1974; Duesberg et al., 1974; Coffin and Billeter, 1976), size measurements of infectious DNA (Hill and Hillova, 1974) as well as by molecular hybridization (Baluda et al., 1974). Moreover, the recent application of heteroduplex mapping techniques to studies of type-C viruses have indicated the viral RNA to consist of two 35 S monomeric subunits, attached near their 5’ ends in a dimer linkage structure (Kung et al., 1975, 1976; Bender and Davidson, 1976). The 4 S virionassociated RNA species has been shown to represent tRNA (Erikson and Erikson, 1971; Bonar et al., 1967; Travnicek, 1968) and has been identified as tRNAnp for avian type-C viruses (Dahlberg et al., 1974b; Harada et al., 1975) and tRNAP” in the case of some but not all mammalian type-C virus isolates (Peters et al., 1977). Studies on Rous sarcoma virus (Furuichi et al., 1975; Keith and Fraenkel-Conrat, 1975) and the Moloney strain of murine leukemia virus (MuLV) (Stoltzhs and Dimock, 1976; Bondurant et al., 1976; Rose et al., 1976) have shown that the 5’ end of the viral RNA is capped by the structure m7G(5’)ppp(5‘)NmpNp.Such capping structures are common among eukaryotic mRNAs and may act to protect the RNA from attack by phosphatases and other nucleases and in addition may promote initiation of translation (Shatkin, 1976). The 3’ terminus of each 30-35 S RNA has a poly(A) sequence of about 200

6


nucleotides (Lai and Duesberg, 1972; Ross et al., Ic172; Keith et al., 1974; Wang et aE., 1975). A tRNA molecule is associated with the viral 35 S RNA and functions as the primer for RNA-directed DNApolymerase (RDDP), initiating synthesis of complementary DNA at a unique site located within 150-200-nucleotide residues from the 5’ terminus of the viral genome (Faras et al., 1974; Taylor and Illmensee, 1975; Cashion et al., 1976; Haseltine et al., 1976). Recently, the sequence of the first 101 bases beginning at the 5’ end of the Prague RSV-C genome has been determined (Haseltine et al., 1977; Shine et al., 1977).These studies have resulted in the identification of a possible initiation triplet (AUG) for protein synthesis located 85 bases from the 5’ cap structure. Moreover, a sequence of 21 nucleotides immediately preceding the 3’ poly(A) of a prototype avian type-C virus, PrRSV-C, has been identified as: 5’GCCAUUUUACCAUUCACCA poly(A) 3’ (Schwartz et al., 1977). The fact that this sequence is identical to that of the first 21 nucleotides located at the 5’ end of the 35 S RNA indicates that the viral genome is terminally redundant. This possibility has recently been confirmed (Coffin and Haseltine, 1977). Independent evidence for terminal redundancy was derived !?om restriction endonuclease mapping of DNA sequences complementary to the Moloney sarcoma virus genome (Canaani et ul., 1977). This terminal redundancy provides for the possibility of circularization of the viral genome prior to integration into host cellular DNA. In fact, circular structures have been visualized by electron microscopy heteroduplex analysis and a replication mechanism involving a circular intermediate has been proposed (Junghans et at., 1977). 111. Proteins of Type-C RNA Tumor Viruses

In view of the above findings indicating the complexity of the type-C viral genome to be of the order of 2-3 x 106, the maximum size of the translational product for which it can code is about 300,000. Studies on type-C virus-coded proteins have now led to identification and characterization of a sufficient number of proteins to essentially account for this entire coding capacity. These consist of a protein with RNA-dependent DNA-polymerase enzymatic activity as well as structural components of the virion, including a 70,000 molecular weight envelope glycoprotein and several low molecular weight nonglycosylated proteins. In addition, a number of type-C virus isolates are known to have acquired transformation-specific sequences by recombination with host cell genes. Such recombinant viruses, which in general are replication-defective, apparently code for one or more pro-


7

teins associated with malignant transformation. In the following sections, currently available information regarding the properties of type-C viral translational products, with emphasis on possible functions, is reviewed. In addition, an attempt has been made to define and genetically map regions of the viral genome coding for individual translational products. A. RNA-DEPENDENTDNA-POLYMERASE

The RNA-dependent DNA-polymerase (RDDP), also known as “reverse transcriptase,” has the capacity to use both, polyribonucleotides and polydeoxyribonucleotides as template to synthesize complementary DNA (Baltimore, 1970; Temin and Mizutani, 1970; Baltimore and Smoler, 1971; Spiegelman et al., 1970a,b; Temin and Baltimore, 1972; Verma, 1977). The purified RDDP also exhibits ribonuclease activity “RNase H” which can selectively degrade the RNA moiety of RNADNA hybrids (Moelling et al., 1971; Baltimore and Smoler, 1972; Keller and Crouch, 1972; Leis et al., 1973). Analysis of mutants of avian and mammalian type-C viruses, characterized by temperaturesensitive lesions in their RNA-dependent DNA-polymerase, DNAdependent DNA-polymerase and RNase H activities convincingly demonstrated these activities to be virus-coded and essential for integration of the viral genome into the cellular DNA (Linial and Mason, 1973; Mason et al., 1974; Verma et al., 1974, 1976; Tronick et al., 1975). Most of the viral RDDP requires a primer such as transfer RNA and some metal ions for activity (Dahlberg et al., 197413; Hasteline and Baltimore, 1976; Grandgenett, 1976b). Thus, the type-C viral enzyme prefers Mn2+ ions while type-B and type-D viruses prefer Mg2+ions for their activity (Scolnick et al., 1970; Howk et al., 1973; Abrell and Gallo, 1973; Michalides et al., 1975). In addition to transcription of their natural template, all viral polymerases faithfully copy synthetic template-primers, such as poly(A) - oligo(dT), poly(C) oligo(dG), to various extents (Spiegelman et al., 1970a,b; Mizutani et al., 1970; Riman and Beaudreau, 1970). Recently, optimal conditions for reverse transcription of complete copy of the viral genome in vitro have been described (Rothenberg and Baltimore, 1977). Under conditions of limiting Mg2+ ion concentration, full length, apparently infectious (Rothenberg et al., 1977) complementary DNA copies of the viral RNA can be synthesized. The RDDP from the murine leukemia viruses has been shown to consist of a single polypoptide of about 70,000 molecular weight (Ross et al., 1971; Tronick et aZ., 1972; Gerwin and Milstein, 1972;

8


Hurwitz and Leis, 1972). In contrast, the avian type-C viral reverse transcriptase contains two subunits, a (70,000) and /3 (110,000) (Temin and Baltimore, 1972; Verma et al., 1974; Gibson and Verma, 1974; Kacian et al., 1971; Grandgenett et al., 1973). The a subunit exhibits both polymerase and nuclease activities while the p subunit apparently enhances the binding of a to the template or substrate (Verma et al., 1974; Gibson and Verma, 1974; Grandgenett and Green, 1974; Moelling, 1974; Grandgenett, 1976a). Pulse-labeling of the Rauscher (R)-MuLV infected mouse cells has indicated that RDDP is initially synthesized in the form of a large precursor protein of about 200,000 molecular weight (Naso et al., 1975; Arlinghaus et al., 1976). Posttranslational cleavage of this high molecular weight precursor gives rise to an 80,000gag gene-coded precursor and the viral RDDP ofabout 75,000 molecular weight (Naso et al., 1975; Arlinghaus et al., 1976). The reverse transcriptase also provides a useful antigentic marker for the identification and characterization of type-C viruses of diverse origin (Aaronson et al., 1971a; Scolnick et al., 1972a). Antisera prepared against the enzyme of a given mammalian type-C virus most strongly inhibits the activity of the homologous enzyme and to a lesser degree enzymes of type-C virus isolates of other mammalian species (Aaronson et al., 1971a; Scolnick et al., 1972a; Parks et al., 1972). However, antisera to mammalian type-C viral enzymes do not inhibit the reverse transcriptases of avian type-C viruses or of mammalian oncornaviruses that are not type-C viral in origin (Aaronson et al., 1971a; Scolnick et a1., 1972a). Recently, radioimmunoassays for the RNA-dependent DNA-polymerases of avian (Panet et al., 1975; Reynolds and Stephenson, 1977) and mammalian (Krakower et al., 1977) type-C viruses have been described. By use of homologous competition assay for R-MuLV it was possible to distinguish R-MuLV enzyme from that of other murine viruses while in heterologous more broadly reactive assays, a number of mammalian type-C viruses showed immunologic cross-reactivity (Krakower et al., 1977). Application of competition immunoassays for the viral reverse transcriptase to studies of intracellular RDDP expression have led to the demonstration that translation of the type-C viral genome must involve more than one initiation site (Reynolds and Stephenson, 1977). B. STRUCTURALPROTEINS

Mammalian type-C viruses of diverse species of origin exhibit marked similarities in their structural components. Thus, it has been


9

possible to identify functionally analogous structural proteins of different type-C RNA viruses on the basis of their biochemical and immunologic properties. Type-C viral structural proteins can be separated into two groups on the basis of the map positions at which they are coded within the viral genome. One group, which includes the major envelope glycoprotein (gp70) and a nonglycosylated 15,000 molecular weight protein (p15E) are synthesized in the form of a common precursor coded for by a viral gene generally referred to as enu. The remaining viral proteins are characterized by molecular weights in the 10,000 to 30,000 range and are synthesized as a 65,000 molecular weight precursor protein coded for by a region of the viral genome designated gag. These latter proteins are nonglycosylated and are generally thought to be located in the nucleoid or core of the virion. The biochemical and immunologic properties and posttranslational processing of enu and gag coded structural proteins are considered below.

1. Env Gene-Coded Proteins There is accumulating evidence from a number of laboratories consistent with the possibility that the major envelope glycoprotein, gp70, and a lower molecular weight, nonglycosylated virion structural protein, p15E, are initially synthesized in the form of a common precursor. This is indicated b y the demonstration of an 85,000-90,000 molecular weight glycoprotein in R-MuLV infected cells which is precipitable by anti-gp70 and anti-p15(E) sera and which by pulse chase experiments gives rise to cleavage products of around 70,000 and 15,000 molecular weights, respectively. Methionine-labeled peptide sequences analogous to those of gp70 and p15(E) within this precursor have been identified b y tryptic digest analysis (Arcement et al., 1976; Shapiro et al., 1976; Van Zaane et al., 1976; Famulari et nl., 1976). Inhibition of glycosylation of the primary en0 gene product by use of 2-deoxy-D-glucose or cytocbalasin B leads to formation of a 70,000 molecular weight nonglycosylated protein (Shapiro et a1., 1976; W. J. M. van de Ven, personal communication) which presumably represents enu gene translational product prior to glycosylation. The product-precursor relationships between these various enu gent coded proteins have been confirmed by in uitro protein synthesis studies (Gielkens et al., 1974; Van Zaane et al., 1977). For instance, in a rabbit reticulocyte cell-free system, 22 S mRNA, isolated from R-MuLV infected cells, was shown to code for synthesis of a 70,000 nonglycosylated protein containing antigenic determinants in common with gp70 (Gielkens et al., 1974). Injection of the same mRNA

10


into axenopus laevis oocyte translation system resulted in synthesis of a glycosylated 82,000 molecular weight protein as well as significant amounts of the two env gene cleavage products, gp70 and p15(E) (Van Zaane et al., 1977). In addition to the major 70,000 molecular weight envelope glycoproteins and the nonglycosylated envelope protein, p15(E), there have been reports of the presence of a 45,000 molecular weight glycoprotein constituent of type-C viruses (August et al., 1974; Fleissner et al., 1974; Ikeda et al., 1975; Moroni, 1972; Moennig et al., 1974). However, on the basis of amino acid composition, immunologic crossreactivity (Marquardt et al., 1977; Krantz et al., 1977; Charman et al., 1977) and peptide map (Elder et al., 1977), this latter glycoprotein appears to represent a breakdown product of gp70. a. Major 70,000 Molecular Weight Envelope Glycoprotein (gp70). The 70,000 molecular weight envelope glycoproteins of mammalian type-C virus isolates of mouse (Strand and August, 1974a; Hino et al., 1976), woolly monkey (Hino et al., 1975), baboon (Stephenson et al., 1976a), and feline (Stephenson et al., 1977a) origin, have been isolated and studied in detail. Immunologic characterization of gp7Os of these diverse viruses have indicated the presence of type, group, and interspecies-specific antigenic determinants (Strand and August, 1973, 1974a; Hino et al., 1975, 1976; Stephenson et al., 1976a). That these antigenic specificities reside in the protein moiety and not the carbohydrate residues was demonstrated by the use of glycosidase enzymes to remove selectively the carbohydrate portions of the molecule (Bolognesi et al., 1975). Competition immunoassays which measure type-specific antigenic determinants of these viral envelope glycoproteins have proven useful in discriminating closely related type-C virus isolates (Strand and August, 1974a; Hino et al., 1975, 1976; Stephenson et a1., 1976a), while interspecies immunoassays have been primarily utilized for detection and characterization of type-C virus isolates of diverse mammalian species (Strand and August, 1973; Stephenson et al., 1976a).These viral-coded glycoproteins have been shown to represent the major constituents of the viral envelope and to be associated with the spikes or surface projections on the virion surface (Nermut et al., 1972; Witte et al., 1973; Kennel et al., 1973; McLellan and August, 1976). As a consequence of their location within the virion envelope, viral glycoproteins represent the major target of neutralizing antibody (Ikeda et al., 1974; Steeves et al., 1974; Hunsmann et al., 1974). Studies of endogenous type-C viruses of mouse cells have led to the demonstration of multiple classes of biologically distinguishable vir-


11

uses (Stephenson et al., 1975a). Structural proteins, including gp70, coded for b y one such virus, designated Class 111, are expressed in the mouse throughout embryonic life (Stephenson et al., 1974; Hino et al., 1976). As a result, mice appear to develop immunologic tolerance to these proteins (Huebner et al., 1970, 1971; Stephenson et al., 1976b). In contrast, expression of translational products of other endogenous mouse type-C viruses, including a mouse cell tropic virus, designated Class I, is more tightly regulated (Stephenson et al., 1975a). Occasional spontaneous or chemical activation of virus of this class leads to its spread throughout the host (Stephenson and Aaronson, 1972b; Kawashima et al., 1976). Since proteins of Class I virus are not expressed during embryonic life, tolerance to antigenic determinants unique to this gp70 species fails to develop and virus replication is thus subject to regulation by host immune surveillance mechanisms (Stephenson et al., 1976b). The results of recent in vivo studies indicate that spontaneously activated Class I mouse cell-tropic virus can, in certain instances, recombine with and acquire genetic sequences coding for immunologic determinants of an endogenous viral envelope glycoprotein (Hartley et al., 1977; Elder et at., 1977a). These findings are consistent with early in vitro studies demonstrating genetic recombination between endogenous and exogenous mouse type-C viral genomes (Stephenson et al., 1974) and may provide one means by which exogenous virus can partially circumvent its recognition by host immune surveillance. There is evidence that a type-D oncornavirus of primates, MPMV, may have acquired glycoprotein antigenic determinants from its host, the rhesus monkey, by similar means to that described above for Class I ecotropic mouse type-C viruses (Stephenson et al., 1976a). This is suggested b y the fact that the envelope glycoproteins of MPMV and those of type-C viruses of the baboon-RD114 group exhibit immunologic cross-reactivity while other structural proteins of these viruses lack shared determinants. The phenomenon of interference between closely related oncornaviruses has been well established. Thus, infection of cells by a given oncornavirus isolate renders them resistant to superinfection b y the same or highly related viruses. Moreover, there has long been known to be a close correlation between host range, interference, and neutralization properties of related oncornavirus isolates. While these phenomena were initially demonstrated in the case of different subgroups of avian type-C viruses (Vogt and Ishizaki, 1965), such studies have been more recently extended to mammalian type-C viruses (Todaro et al., 1973; Levy, 1973). The correlaton between these various

12


properties indicated the involvement of cellular receptor sites for virus infection. The availability of 1251-labeledgp70 has led to the identification and development of techniques for quantitation of such receptors (DeLarco and Todaro, 1976). Thus, 1Z51-labeledR-MuLV gp70 has been shown to bind and form a stable complex with receptor sites on murine but not other mammalian cells. Purified gp70-binding to mouse cells is prevented if the cells are actively producing related ecotropic virus, presumably because the receptors are occupied and are not available to bind exogenously applied gp70 (DeLarco and Todaro, 1976). Moreover, the presence or absence of appropriate receptors for different viruses may be an important host range determinant. For instance, mouse xenotropic viruses appear to use a different family of receptors from mouse ecotropic viruses, since they neither interfere with viral infectivity (Todaro et al., 1973; Levy, 1973; Fischinger et al., 1975) nor with gp7O-binding to mouse cells (DeLarco and Todaro, 1976). Endogenous type-C viral translational products, including gp70 (Hino et al., 1975), appear, in general, to be expressed throughout the life of their host. Thus, the resistance that cells of most species exhibit to infection b y such endogenous viruses may simply be due to the ability of endogenous gp70 to occupy available receptor sites. A major difficulty, however, is that expression of endogenous type-C viral glycoprotein, especially during embryonic life, leads to the development of immunologic tolerance (Stephenson et al., 1976b).Thus, strong evolutionary selective pressure may exist for an exogenous virus to recombine with genetic sequences coding for host glycoprotein determinants in order to circumvent the hosts immune surveillance, but retain the ability to utilize cell receptors other than those occupied by the host glycoprotein. Such a situation could account for much of the recent data obtained from studies of recombinant mouse (Stephenson et al., 1974; Fischinger et al., 1975; Hartley et al., 1977) and possibly primate (Stephenson et al., 1976a) oncornavirus isolates. There is accumulating evidence indicating that gp70 expression in the mouse may in some way be linked to differentiation and development. By use of immunofluorescence techniques, expression of type-C viral gp70 was demonstrated to be restricted to certain anatomical sites and to b e at much higher levels in lymphoid and epithelial cells than in other cells (Lerner et al., 1976). In fact, the major site of gp70 expression was reported to be the male genital tract (Lerner et al., 1976) and a protein immunologically and biochemically related to gp70 was found at high levels in secretions of the epididymis and ductus deferens (Del Villano and Lerner, 1976). In another study,


13

McClintock et al. (1977) reported constitutively high levels of AKRMuLV gp70 expression in the absence of overt virus in bone marrow cells of all strains of mice examined, including several which possess, at most, only a portion of the Class I type-C viral genome (Chattopadhyay et al., 1974). Moreover, the results of independent reports by several laboratories have established gp70 to be a constituent of the surface of normal thymocytes and to share immunologic and biochemical properties with the thymocyte differentiation marker Gix (Tung et al., 1975; Obata et al., 1975; Del Villano et al., 1975). b. Nongl ycosylated Envelope Protein ( p l 5 E ) . A nonglycosylated mouse type-C viral envelope protein of about 15,000 molecular weight (p15E) has also been described (Ikeda et al., 1975; Schafer et al., 1975). This protein has a marked tendency to aggregate in the absence of detergent and generally chromatographs in the void volume of agarose gel filtration columns even in the presence of 6 M guanidine hydrochloride (GuHC1) (Ikeda et al., 1975). Biochemical and immunologic characterization of p15E has shown it to be distinct from the 15,000 molecular weight gag gene-coded protein (p15) (Ikeda et al., 1975; Schafer et al., 1975), although both p15 (Strand et al., 1974; Barbacid et al., 1977) and p15E (Schafer et al., 1975) have been shown to possess group and interspecies antigenic determinants. Moreover, p15E has been shown to be a surface component of the virion and appears to represent the 15,000-17,000 molecular weight protein immunoprecipitable by normal mouse sera (Ihle et al., 1974, 1975; NOwinski and Koehler, 1974). There is evidence that in some instances p15E may be cleaved, giving rise to a low molecular weight polypeptide of about 12,000 (p12E) (Arcement et al., 1976).This possibility is supported by the results of pulse chase experiments as well as by peptide mapping data (Naso et al., 1976; Van Zaane et al., 1976). Recently, two groups of investigators have independently reported that in type-C viral particles, the envelope glycoprotein, gp70, is frequently found to be linked by disulfide bonds to a nonglycosylated protein of about 15,000 molecular weight. This was shown by a comparison of the electrophoretic mobilites of Moloney leukemia virus proteins under both reducing and nonreducing conditions (Leamnson et al., 1977; Witte et al., 1977). While the relationship of this 15,000 molecular weight protein to p 15E remains speculative, the possibility that they represent the same protein is supported by their similar electrophoretic mobilities and the fact that both have been shown to be constituents of the virion envelope (Leamnson et al., 1977; Witte et al., 1977). Moreover, there is evidence that both the protein disulfide bonded to gp70 and p15E have relatively acidic isoelectric points

14


(Witte et al., 1977), while the 15,000 molecular weight gag gene coded proteins of mouse type-C viruses are more basic (Stephenson et al., 197%). 2. The gag Gene-Coded Virion Structural Proteins The nonglycosylated structural proteins coded by the avian type-C viral gag gene were shown in early studies to be synthesized in the form of a high molecular weight precursor polypeptide (Vogt et al., 1975). An analogous situation has more recently been described for a prototype mammalian type-C virus, R-MuLV (Naso et al., 1975; Stephenson et al., 197513; Van Zaane et al., 1976; Shapiro et al., 1976). Posttranslational cleavage of the 65,000-70,000 molecular weight R-MuLV precursor results in proteins of molecular weights of about 30,000 (p30), 15,000 (p15), 12,000 (p12), and 10,000 (p10). At present, the most widely adapted system for identification of the virus coded proteins is based on molecular weights as determined by gel filtration in the presence of 6 M GuHCl (Nowinski et al., 1972; August et al., 1974). Although this convention is useful in identification of proteins of murine type-C virus isolates, differences in the molecular weights of analogous proteins of virus isolates of evolutionarily diverse origin frequently lead to considerable confusion. This difficulty can be avoided, however, if gag gene-coded proteins are identified by a more generalized nomenclature system taking into account both immunologic and biochemical properties. For these reasons, mammalian type-C viral gag-coded proteins are designated in the following sections as: (a) major virion group-specific antigen, (b) hydrophobic structural protein, (c) type-specific RNA binding protein and (d) basic ribonucleoprotein. a. Major Virion Group-Specijc Antigen. The most extensively studied of the type-C viral gag gene proteins, to date, has been the major group-specific virion antigen (p30). This viral-coded structural protein is generally identified on the basis of its molecular weight of about 27,000-30,000 as determined by agarose gel filtration in the presence of 6 M GuHCl and its broadly reactive immunologic properties (Fig. 1).The initial identification of p30 was based on the results of complement-fixation tests indicating the expression of a virus-specific antigen in various mouse tissues (Hartley et al., 1965, 1969; Geering et al., 1966). Subsequent to these studies a number of groups successfully purified the protein associated with this immunologic reactivity by the application of gel filtration and preparative isoelectric focusing techniques (Oroszlan et al., 1970, 1971; Gilden and Oroszlan, 1972). An independent approach to purification of p30 involved agarose gel filtration in the presence of 6 M GuHCl (Nowinski et al., 1972).


30

50 70 FRACTION NUMBER

15

90

FIG. 1. Molecular weight analysis of type-C virus gag gene-coded structural proteins of a prototype RNA tumor virus, FeLV, by agarose gel filtration in the presence of 6 M guanidine hydrochloride. Individual column fractions were tested in competition immunoassays for FeLVp30 (A),p15 (B), p12 (C),andplO(D).[From Khan and Stephenson (1977) with permission of American Society for Microbiology.]

Immunologic analysis of type-C viral p30 by use of complementfixation (Huebner et al., 1964; Huebner, 1967; Hartley et al., 1969) and gel immunodiffusion analysis (Geering et al., 1966; Schafer et al., 1969; Gilden and Oroszlan, 1972) led to the demonstration of pronounced group-specific antigenic determinants. Immunologic assays recognizing such determinants provided one of the first means of discriminating type-C virus isolates of different species of origin. Subsequent independent studies b y several groups of investigators showed additional antigenic determinants shared by the major internal antigens of type-C RNA tumor viruses of diverse species of origin (Geering et al., 1970; Gilden et al., 1971; Schaferet al., 1972).With the development of sensitive and specific competition immunoassays, direct confirmation of these findings was obtained (Parks and Scolnick, 1972; Tronick et al., 1973). Moreover, in addition to these group- and interspecies-specific antigenic determinants, the major structural proteins of murine type-C viruses were shown to possess less pronounced type-specific determinants (Stephenson et al., 1974). The immunologic relatedness of the major structural proteins of type-C virus isolates of diverse species of origin appears to be reflected in their primary structure. Partial amino acid sequences have been determined for internal antigens of type-C virus isolates of a number of mammalian species including the cat, baboon, mouse, rat, gibbon ape, and woolly monkey (Oroszlan et al., 1975, 1977). The amino terminus of p30 proteins of each of these virus isolates have

TABLE I1 EXPRESSIONOF ANTIGEN RELATED TO THAT OF THE MAJOR STRUCTURAL OF THE P. cynocephalus BABOONVIRUSa

PROTEIN

(p30)

Viral p30 antigenic r e a c t i v i v

Family

Genus

Common name

Cercocebus

C . atys

Sooty white-crowned

Cynopithecus Papio

C . niger P. cynocephalus P. anubis P. hamadryas P. papio P. sphinx C . neglectus C . aethiops C . patas M. mulatta M . arctoides M. nernestrina M. irus P. pygmaeus P. troglodytes G. gorilla H. lar H . sapiens

mangabey Celebes (black) ape Yellow baboon Olive baboon Hamadryas baboon West African baboon Mandrill De Brazzas guenon Grivet monkey Patas monkey Rhesus macaque Stump-tailed macaque Pig-tailed macaque Crab-eating macaque Orangutan Chimpanzee Gorilla Gibbon ape Man

Cercopithecidae Cercopithecus Macaca

Pongo Pongidae < Z i , l a Homoinidae

Species

Hylobates Homo

Number tested

1 1 1 1 1 2 1

1 1 2 10 2 2 3 4

14 1 3 119

Number positive 0 1 1 1

0 0 0 1 0 0 10

Range (ng ~ 3 0 1 ~ cellular protein) 50% in the incidence of mammary tumors arising in their colony of BALBkCrgl mice between the time they imported a breeding nucleus from the Cancer Research Genetics Laboratory at Berkeley in 1963, and 1967; the high incidence persisted. All of 6 of the tumors they examined con-

CHOICE OF ANIMAL TUMORS FOR THERAPY STUDIES

175

tained B-type particles. After confirming that their BALB/c mice had not become confused with their MMTV-infected strain A mice, they concluded that the BALB/c mice had become contaminated with milk-borne MMTV and that the infection had subsequently been disseminated through their colony. Collins and Parker (1972) found viral contamination of 69% of the 465 specimens of leukemia virus suspensions or transplanted tumors which they examined from various sources; half of the 23 specimens of leukemia L1210 were contaminated, 5 with polyoma virus; the one specimen of P 388 Leukemia and all specimen of Sarcoma 180 and Lewis Lung Tumor were contaminated. Although most contaminations were b y the nononcogenic lactic dehydrogenase and mouse hepatitis viruses, it is clear that the presence of these viruses, especially if they can bring about allogenization of the tissues they infect, could invalidate immunological and other findings obtained using infected materials. These findings of Collins and Parker (1972) are a shocking revelation of the poor quality control of biological materials widely used in cancer research. It is clear from these examples that the immunological status of a spontaneous tumor is particularly in question when it arises and is maintained in a laboratory in which artificial induction of tumors is commonly practiced or to which there is free importation of oncogenic viruses or tumors induced b y them or of tumors which may have been transplanted and maintained within several previous laboratories. It may be significant that in my own laboratory, in which no immunogenicity has ever been demonstrated in any of the many spontaneous tumors that have been exclusively used (Hewitt et aE., 1976), importation of tumors or viruses (with the recent exception of MMTV intrinsic in a small C3H mouse colony) has been rigidly prohibited during the 20-year history of the two low cancer strain colonies maintained. To summarize this section on virus-induced tumors: Such tumors almost invariably exhibit evident or potential immunogenicity attributable to the virus or to virus-associated antigens, and their use for experimental studies intended to have relevance for spontaneous tumors is imprudent, unnecessary, and commonly misleading. In the case of MMTV-induced mammary tumors isotransplanted to MMTVinfected mice, immunity is unlikely to be manifest in a way that affects the results of assays of therapeutic agencies, but their use for study of immunotherapeutic procedures is more questionable. Evidence for implication of oncogenic virus in the generation of clinical malignant tumors is insufficient to justify use of virus-induced animal tumors as

176

HAROLD B. HEWITT

models. Finally, artifactual tumor immunogenicity may be associated with a nominally spontaneous tumor by unrealized contamination with oncogenic virus or by errors in tumor signification.

E. TUMORSTRANSPLANTED UNDER CONDITIONS ISOGENICITY

OF INCOMPLETE

The basic genetic principles governing the success of transplantations of normal or tumor tissue from one individual to another were established as long as 60 years ago using mice (Little and Tyzzer, 1916) and have since been refined and extended by further study. It is now evident that a multiplicity of alleles may exist at each of at least 15 different histocompatibility loci (Snell and Stimpfling, 1966); using tail skin exchange grafts, Bailey and Mobraaten (1964) estimated that the number of loci may be as high as 30. Although these loci vary considerably in their “strength,” as indicated by the ease or difficulty with which their associated histocompatibility barriers can be transgressed, it is known that all such barriers can be fortified b y preimmunization; indeed, the weaker loci have only been identified with the aid of immunization. The multiplicity of loci is an important consideration in respect of the liability of a strictly inbred mouse colony to undergo “genetic drift,” whereby a degree of histoincompatibility may arise between a transplanted tumor derived from an ancestral mouse and a descendant recipient of the same colony. Whatever the overall frequency of spontaneous mutation at any one locus may be, the risk of incompatibility arising must increase in proportion to the number of loci able to contribute to it. It is probably only at the H-2 locus that allelic transformations can result in absolute resistance to the growth of a transplanted tumor, and it appears that the supervention of such solid resistance to a tumor transplanted within its own substrain is very rare. I n over 20,000 intrastrain tumor transplantations, we have only encountered two absolutely resistant mice (resistant, that is, to all tumors which had arisen in that strain) in a colony continuously inbred without deliberate selection of sublines for over 20 years. Since no further resistant mouse has been encountered in the several years since then, it is concluded that progeny of the resistant mice did not contribute significantly to the breeding colony. However, allelic transformations at minor histocompatibility loci can be expected to introduce subtle artifacts into a transplanted tumor system being used to measure the potential of a therapeutic agent: The primary implant may grow readily, but the resistance induced by its growth may attain to a level sufficient to restrain its regrowth (or the


177

emergence of metastases) following regression of the primary induced b y a cytotoxic effect of the agent under test. That is, the artifact associated with minor histoincompatibility may contribute substantially to the end-point of “cure,” and so invalidate extrapolation of the data obtained to a system in which the tumorhost relationship is free from such artifact. It seems that the risk of artifactual immunity being introduced into a tumor transplant system was at one time overrated and led to idealistic recommendations that transplantation should be confined within a litter; such a severe restriction clearly prohibits all assay procedures and cannot provide for any progressive studies. In our experience, using multiple sublines of an inbred strain, some of which are likely to have been divergent over as many as 30 generations, we have encountered no evidence of tumor immunogenicity even using (spontaneous) tumors which have been serially transplanted for over 8 years; our tests for immunogenicity have included comparative challenge assays of viable tumor cells in normal and putatively preimmunized mice. A similar prolonged freedom from evidence of “genetic drift” within an inbred mouse strain has been reported by Godfrey and Searle (1963), who used exchange skin grafting to seek evidence of histoincompatibility. Dr. M. F. W. Festing (personal communication) has calculated that there should still be a 96%chance that a tumor would be accepted by recipients of the inbred strain of origin after about 13 years of propagation of the strain b y brotherkister mating, assuming 3 generations per year and that mutation was the exclusive source of genetic variation. H e used Silvers and Gasser’s (1973) formula for calculating the probability that two sublines will be histocompatible after specified numbers of generations of brothedsister breeding, but modified their data by employing a preferred lower estimate of the mutation rate-that of J. Klein (1975), of 9.3 x l W 4 per generation for all histocompatibility loci. Thus, although the risk of genetic drift as a source of transplantation immunity cannot be eliminated from any transplant system, it certainly appears not to be great using an inbred strain of mouse maintained in the laboratory in which it is used and a tumor arising in it. It is unfortunate that many investigators secure these confined conditions for minimizing the risk of genetic drift by chemically inducing tumors in mice of their own substrain; they thus invite what is almost a certainty of artifactual tumor immunogenicity, however genetically homogeneous their inbred strain may be (see Section 111,C). “ Considerable emphasis has been given to progressive antigenic simplification” in serially transplanted tumors, whereby they acquire increasing ability to transgress a range of histocompatibility barriers.

178

HAROLD B. HEWITT

The phenomenon is referred to in two contexts: Either it is welcomed as an advantage because it widens the utility of a tumor by increasing the range of strains which can be employed to provide graft recipients, or it is advanced to “explain” an inability to demonstrate immunogenicity in a tumor which has undergone prolonged serial transplantation, as is required to sustain the hypothesis that autochthonous tumors are usually immunogenic (M. Woodruff, personal communication). However, there is little doubt that increasing transgression of histocompatibility barriers is attributable more to increasing tumor growth rate and “escape” from host immune influences than to total loss of antigenic elements from the tumor. Certainly, there are a great many tumors whose early immunogenicity was imposed by chemical induction or by their origin in an animal of unspecified genetic constitution which have retained their immunogenicity after continual serial passage for over 50 years; and in many cases, such as the Ehrlich tumor of mice, the transplantation history includes passage through a very wide variety of strains, giving every opportunity for eliminative immunoselection. There is no doubt that transplantation immunogenicity is commonly imposed on a tumor system b y transference of a tumor from one laboratory to another. When a tumor is of origin in a specified animal strain and mice of the same nominal strain are used as recipients in the receiving laboratory, a difference between the substrains used in the two laboratories is often overlooked as a possible source of relative incompatability between tumor and hosts. Substrains may often have been separately propagated for over 20 years, during which period some genetic dissociation is very likely to have taken place. Moreover, the long history of both substrains may imply propagation of each in a sequence of laboratories and/or commercial breeding establishments in any of which breeding errors may have been made. Thus, in many instances the description of a tumor as “isogeneically transplanted” is purely nominal and has not been validated b y histocompatibility tests done between mice of the substrains held by the donating and receiving laboratories. Smith and Scott (1972) obtained evidence of immunostimulation by C. pamum vaccine using a murine ascites leukemia of CBA mice which was transferred from our laboratory to theirs some years previously, although this tumor had not displayed evidence of immunogenicity in our hands. It was disclosed that this tumor had been transplanted at the receiving institution to a substrain of CBA mouse different from that in which it had arisen and with which our own experiments had been done. Graff et al. (1975) have given a striking example of the way in which


179

subline differences between mice to which a tumor is transplanted a potentially serious problem for experimental present chemotherapists.” The distinction of their study is that they were alerted to the hazard by an incidence of spontaneous regressions and investigated it by quantitative transplantation assays. They assayed a transplantable leukemia (which had arisen in an AKR/J mouse in 1962) in 5 sublines of AKR obtained from different sources. Using AKWJ mice obtained directly from the Jackson Laboratory or mice of their own AKWJ colony derived 4 years previously, they found that very few mice survived the injection of 10 leukemia cells; a proportion of AKR mice from 2 substrains commercially bred but obtained through the U.S. National Cancer Institute rejected as many as lO3-10’ cells; and 75% of AKR mice obtained directly from a commercial breeder rejected lo7 cells. They demonstrated by exchange skin grafts a considerable degree of histoincompatibility between substrains and within some of them and concluded that the most likely source of the genetic inconsistencies was errors of breeding. Leukemia L1210, perhaps the leukemia most commonly used in experimental chemotherapy studies, has been shown to be immunogenic under practically all conditions of transplantation. It is likely that this property was partly conferred at its inception, for it arose in an MCA-treated mouse of the DBAI2 strain almost 30 years ago. Mihich (1969) refers to its immunogenicity in DBN2, (BALB/c x DBAI2) F1, and (C57B1/6 x DBAI2) F1mice. Although mice of the three DBAI2 substrains he examined all gave 100% takes with 10 cells, and evidently shared the H-2d histocompatibility factor, differences between the substrains at minor histocompatibility loci were associated with widely divergent results from chemotherapy studies: The percentage of leukemic mice surviving at 50 days after a standard combined chemotherapy treatment with 4, 4’-diacetyl-diphenylurea-bis-guanylhydrazone and arabinosylcytosine was 57% using DBAI2 HaDD, 16% using DBAIBJ, and 5 % using DBN2 Cr. Since “cured” mice resisted challenge with between 103 and lo6 L1210 cells, survivors evidently owed their cure to a contribution, varying with the substrain of mouse used, of host resistance mustered during temporary arrest of the disease by the drugs. Mihich (1969) demonstrated also that the presence or absence of immune exertion by DBAI2 HaDD leukemia recipients was associated with quite different evaluations of the combined chemotherapy schedule employed. Using whole-body irradiated mice, synergism between the two drugs was observed; using unirradiated mice, only an additive effect was obtained. A further commonly used MCA-induced ascites leukemia of “

180

HAROLD B. HEWITT

DBAI2 mice, P-388, also displays a wide range of differential transplantability to different substrains of DBAI2. Berry and Andrews (1961) required only 2 leukemia cells for 50% successful intraperitoneal transplantation to their DBNBJN mice and were unable to raise this value by “pre-immunization” of their mice using lethally irradiated leukemia cells. A corresponding value of -400 cells was obtained b y Birch et al. (1975) using DBAIBJ mice; and in their transplant system mice could be readily immunized against the leukemia. There can be no doubt that studies similar to those made by Graff et al. (1975) or Mihich (1969) would be equally revealing of inconsistencies between substrains if they were made of other transplantable animal tumor systems which are commonly used to assess therapeutic agents. In respect to immunity studies, it is to be noted that an immune response against a transplanted tumor which is, in fact, attributable to a histocompatibility difference between the mouse substrain in which the tumor arose and that to which it is transplanted is ostensibly deserving of the description “tumor specific,” although the term is not intended to have this connotation. A tumor-specific immune reaction can be identified as such only in the autochthonous host or in transplantees which have been shown to be isogenic with the host of tumor origin with respect to the transplantation of normal tissues. It is clear that the widespread use of familiar veteran tumors having a specified animal strain of origin does not at all achieve the consistency and comparability of data from different laboratories which is claimed for the practice. Substrain differences between recipients may very well be associated with a wider range of response to therapeutic agents than would be encountered for comparisons made between the responses of different nonimmunogenic tumors. It should be added that the older named or coded tumors are subject also to divergence between different lines of the tumor itself, which can have had different transplantation histories over several decades; some of this divergence may involve antigenic modifications. Atassi and Taignon (1974) reported the results of treatment of L1210 leukemic mice with adriamycin or adriamycin-DNA complex. The prognosis of treated mice bearing their own line of L1210 leukemia was significantly different from that of mice bearing a line of L1210 leukemia obtained from another laboratory. For logistic convenience, an increasing number of researchers now use as tumor recipients F1mice of a cross between two inbred strains of which one is the strain of tumor origin. While the laws governing the genetics of transplantation permit the assumption of histocompati-


181

bility for such transplantation to F, hybrids, it should be appreciated that, in the case in which some residual heterozygosity remains in the relevant inbred strain of the F1 cross, the risk of histoincompatibility is substantially increased by using F, mice as recipients instead of mice of the original strain (M. F. W. Festing, personal communication). A further hazard of the use of F, mice as tumor recipients was referred to in Section 111,D: Transplantation of an MMTV-induced tumor to F, hybrid mice of a cross in which the strain of tumor origin is represented by the male, the female being of a MMTV-fi-ee strain, is associated with transplantation resistance against the tumor evoked by MMTV-associated antigens in it. Before concluding this subsection, reference must be made to the continued use of frankly allografted tumors, those of origin in animals of unspecified and technically unreproducible genetic constitution. These are all tumors which arise in exotic species from which homozygous inbred strains have not yet been derived, or tumors which arose in familiar laboratory species at a time before inbred strains were developed. It is remarkable that Ehrlich’s tumor and Crocker tumor 180 of mice, and Walker Carcinosarcoma 256 of rats, are still among the commonest tumors used in cancer research. Since the biological features of these tumors are adequately represented in any number of more recently arisen tumors, allowing isografting to the inbred strains of origin, their continued usage can be ascribed only to sentimental regard for them as memorials to the enterprising men who pioneered our present research facilities; the veteran tumors referred to arose respectively in 1907, 1914, and 1928 (Stewart et al., 1959). Use of the Ehrlich tumor was reported in 11%of the papers describing experimental studies with mouse tumors which were published in 1975 (Roberts and Drobycz, 1975),a drop of only 3% from the usage in 1970. All that need be said of frankly allografted tumors in the context of this essay is that results of therapy studies done with them can have no quantitative relevance for nonimmunogenic tumors, and that demonstration of their invariable antigenicity has no significance for the immunology of naturally occurring cancer. The large extent to which they are still employed is a poignant revelation of the difficulty encountered b y many contemporary cancer researchers in obtaining appropriate experimental tumor systems. In conclusion, it must be stated that the conference of partial allogenicity upon a transplanted tumor system by unrealized genetic contamination of the inbred strain of animal used, by failure to take account of possible genetic differences between the substrains of recipients during interlaboratory transfer of tumors, by contamination of

182

HAROLD B. HEWITT

tumors with possibly allogenizing viruses, or by errors of labeling resulting in confusion between tumors are among the most subtle and frequent sources of artifactual tumor immunogenicity. Because these influences are unrealized, a description of the tumors as “isogeneically transplanted” is nominally justified, and the deficiencies of the animal model are thus concealed and perpetuated. When confronted by a transplanted tumor system having a history of serial transplantation in one or more previous laboratories, it is rarely feasible to determine whether any tumor immunogenicity revealed in it has been superimposed as an artifact or represents the continuance of tumor specific antigenicity active in the original autochthonous host. It is certainly impossible to assess the relative contributions from the two sources when both may be exerting an influence on the effectiveness of a therapeutic agent under trial. These uncertainties deserve paramount consideration in deciding the criteria to be met by a transplanted animal tumor system which is to be adopted for use in programmed studies of either tumor immunology or experimental cancer therapy when the results obtained are to be given the status of clinical relevance. It is difficult to escape a conclusion that systems should be rejected for such studies if tumor immunogenicity can be demonstrated in them b y any of .the accepted immunization techniques available for revealing it.

F. RELATIVE USAGE OF DIFFERENT CATEGORIES OF ANIMAL TUMORSUSED IN RECENT EXPERIMENTAL STUDIES OF CANCERTHERAPY To assess current practice in the selection of animal tumor systems for therapy research I have sampled the recent literature in a way that should have avoided bias in reaching the conclusions to be drawn. I have perused the titles of all articles published in the British Journal of Cancer over the period 1974 through 1975 (Volumes 29-32) and in Cancer Research during 1975 (Volume 35) and have selected out for analysis all articles which report examinations of therapeutic or potentially therapeutic agents using transplanted animal tumor systems. The sample comprised 45 articles referring to the use of 71 tumors. Of the tumors used, 38 were employed in chemotherapy studies and 33 in immunotherapy studies (in the case of three articles in which a combination of two modalities was tested, the tumors used were allocated to the modality which appeared to be of predominant interest). Table I1 shows separately, for the chemotherapy and immunotherapy series, the distribution of the tumors used among the


183

TABLE I1 ANIMALTUMORSUSED IN CHEMOTHERAPY OR IMMUNOTHERAPYEXPERIMENTS CATEGORIZED BY NATUREOF ORIGIN" Origin

Chemotherapy

Immunotherapy

Total

Chemically induced Virus-induced Radiation-induced Allografted Spontaneous

16 (42)a 5 (13) 2 (5) 3 (8) 12 (32)

18 (55) 5 (15) 2 (6) 1(3) 7 (21)

34 (48) 10 (14) 4 (6) 4 (6) 19 (27)

" Compilation from

relevant articles published in Cancer Res., Vol. 35 (1975) and

Br. J . Cancer, Volumes 29-32 (1974-1975). Numbers in parentheses indicate percentage of total in each column.

different categories of tumor origin: chemically induced, virus induced, radiation induced, allografted, or spontaneous. Allografted tumors are those arising in animals of unspecified genetic constitution or transplanted to animals of a strain different from that of the animal in which the tumor arose. Of the 4 radiation-induced tumors used, 3 were carried in congenic resistant mice, so that these systems embodied artifactual immunity deliberately introduced. There is no significant difference in the distributions for chemotherapy and immunotherapy experiments. This indicates neglect or unawareness of the need for particular care to exclude artifactual immunity from tumor systems to be used in immunotherapy studies. In the case of chemotherapy studies, the intromission of artifactual immunity may be expected only to augment the effectiveness of an agent under test, whereas in the case of immunotherapy studies, artifactual immunity in a system may be entirely responsible for any effect of immunostimulation demonstrated. Table I1 indicates that almost half the tumors used had been chemically induced, a condition that is almost invariably associated with artifactual tumor immunogenicity (Section 111,C). Only 27% of the tumors were of spontaneous origin, which is the category of tumor least likely to b e associated with immunogenicity and the one to which all but a very few clinical cancers belong. However, as discussed in Section III,E, artifactual immunogenicity may become superimposed upon a transplant system derived from a spontaneous tumor, and the risk of such contamination increases with the length of time elapsing since its origin. I have therefore analyzed the histories of the 19 tumors of spontaneous origin which appear in the sample for the purpose of assessing their status as models of autochthonous

184

HAROLD B. HEWITT

tumors of spontaneous origin with reference to an ideal that is easily attainable within a laboratory bred colony of mice. That is, a tumor arising spontaneously, preferably within the colony providing graft recipients, and used within 15 years of its origin. Table I11 lists the 19 tumors of spontaneous origin which appear in the sample, from which it is seen that in just over one half of the investigations, tumors were used which arose over 20 years previously [Lewis Lung Tumor (LLT), B16 Melanoma, Ridgeway Osteogenic Sarcoma and Neuroblastoma C 13001 and which have been heavily exposed to the numerous conditions which can superimpose artifactual tumor immunogenicity on a system, as discussed in Section II1,E. The predominant usage of LLT deserves discussion because it provides an understanding of the influences conducing to selection of experimental tumor systems. The advantage apparently distinguishing LLT is its large potential for pulmonary metastasis; it thus provides a convenient system for therapy studies directed to control of secondary disease following ablation of the primary implant (e.g., Mayo et al., 1972). However, LLT is not unique in its metastasizing potential, which is possessed by several of our own spontaneous carcinomas. As might be expected of a tumor which has undergone a great many serial passages and has been very extensively transferred from one laboratory to another since its origin, evidence concerning its immunogenicity after transplantation is inconsistent. DeWys (1972) associated its Gompertzian growth curve with a tumor-related systemic growth retarding factor, whereby the presence of a larger tumor in a nominally TABLE 111

TUMORS OF SFONTANEOUS ORIGINUSED IN CHEMOTHERAPY AND IMMUNOTHERAPY

Tumor Lewis Lung Tumor B1G Melanoma Ridgeway Osteosarcoma Neuroblastoma C 1300

EMTG Line 1 Lung Carcinoma Epithelioma Sp. 1 Mammary Cancer D7T4S SBI Lymphoma WHT Squamous Cancer 'D' a

EXPERIMENTS (1974-1975)" Strain of origin

C57B 116 C57B 116 AKm A/ Jax BALBIc BALBlc Wistar (rat) BALBic BALBIc WHTiHt

From same sample as used for Table 11.

Years since origin

24 21 27 35 ? ? 10 >5 ? 10

-

Instances in sample

5 3 1 1 2 1 2 2 1 1


185

isogeneic mouse was associated with loss of carcass weight and depression of the growth rate of smaller tumors in the same animal. Although these findings are suggestive of the exertion of concomitant immunity, DeWys showed that the number of tumor cells required to initiate tumors in mice which had had primary tumor implants excised 8 or 15 days previously was not greater than that required in mice with no previous experience of the tumor; thus, in his system, tumor immunogenicity was excluded. However, Carnaud et al. (1974) revealed “weak but consistent” immunity in their LLT system by successful immunization of recipients using repeated small inoculations of viable cells; they admit that their finding could represent an artifact associated with prolonged serial passage of the tumor. The results of some chemotherapy studies using LLT have been so remarkably good, relative to the expectations for chemotherapy of solid clinical tumors, that a contribution from host resistance factors in the experimental system appears very likely. Mayo et at., (1972) obtained -40% cures of mice bearing 400-mg tumors using cyclophosphamide or methyl CCNU alone, or the two in combination, and 50% cures of mice whose primary tumors had been previously excised at a “quite advanced” stage of tumor growth. Variation of the level of artifactual immunogenicity in different transplant systems of LLT is likely to be responsible for inconsistent results obtained for the effectiveness of the nontoxic immunostimulant levamisole. Neither Spreafico et al. (1975)nor Johnson et al. (1975)could demonstrate any effect of levamisole on the growth of LLT, although the former obtained 90% cures of mice bearing the chemically induced leukemia L1210 and the latter obtained cures of MSV-induced rhabdomyosarcoma. On the other hand, Renoux and Renoux (1972) cured 3/12 mice bearing LLT by early treatment with a single subcutaneous dose of 0.5 mg/kg of levamisole. The findings oE Renoux and Renoux (1972) may very well exemplify the consequences of a change of mouse substrain associated with interlaboratory transfer of LLT. They used as recipients mice designated as C57BL/Rho, and acknowledge Dr. I. Gesser for his provision of the tumor; however, reference to a previous paper in the same journal by Gesser and Bourali-Maury (1972) reveals that their studies of LLT were made using mice designated as C57BL/6. A suggestion that LLT may have exhibited some immunogenicity very early in its history of interlaboratory transfer is gleaned from the report of Sugiura and Stock (1955), who obtained 100% successful transplantation of LLT to their nominally isogeneic mice, yet 4 % of the tumors spontaneously regressed; these regressions are an indication, at least, of some genetic heterogeneity among the recipients. It may b e assumed that the above incon-

186

HAROLD B. HEWITT

sistencies and uncertainties associated with the use of LLT would be equally displayed by the 21-year-old B 16 Melanoma. Certainly, circumstantial evidence of its immunogenicity has been revealed in some of its transplanted systems. For example, Chalmer et al. (1975) found that B 16 Melanoma grew faster and produced more lung metastases in mice chronically exposed to cigarette smoke than in untreated mice; they attributed this finding to an immunosuppressive effect of the exposure to smoke, and this interpretation is supported by their finding that the effect was greater using a more strongly immunogenic tumor induced by murine sarcoma virus. Although B 16 Melanoma provides a very special facility for metabolic studies of melanin production, this feature does not explain its very wide usage in therapy studies, which generally have no reference to it. It is concluded that LLT and B 16 Melanoma have become prominent not on account of unique features intrinsic in either tumor (LLT is described as an anaplastic derivative of a squamous cell carcinoma) but to thetr adoption b y the Division of Cancer Treatment of the U.S. National Cancer Institute for screening of chemotherapeutic agents. It is probable that this employment constitutes a recommendation which discountenances reservations invited b y the long transplantation history of both tumors. The Ridgeway Osteogenic Sarcoma, now almost 30 years from its year of origin, although it gives 100% takes in AKR mice and is known to recur after failed therapy (Laster, 1975), has given various indications suggesting a distinctive hosthumor relationship. The tumor is described as having arisen spontaneously in mice of the A h strain (Laster, 1975), but is currently transplanted in AKR mice, a condition which would appear to imply some lack of nominal isogeneity; it is described by Dunham and Stewart (1953)as being subject to 3% spontaneous regressions in AKR mice. It is remarkably susceptible to cure by actinomycin D; Schwartz et al. (1966) obtained 90-100% cures following a single intravenous dose of 800 pg/kg, but these authors rejected a contribution of host resistance to cure b y demonstrating that tumors were able to grow from secondary implants into mice during drug-induced regression of primary implants. Certain observations of Biano et al. (1971) suggest that the hosthumor relationship may be complex; a considerable splenomegaly accompanies tumor growth, but the number of antibody-forming plaques in the spleen was progressively reduced as the tumor increased in size. It appears possible that some immunosuppresive influence accompanies tumor growth, and that the system deserves a more exhaustive examination for features which could explain the apparent absence of immunogenicity despite the large opportunities for intromission of artifactual immunity


187

to which the tumor has been exposed during its very long history of transplantation. Table I11 shows 2 instances of use of the EMTG tumor in therapy studies. This tumor has a complex history (Medina and DeOme, 1970; Rockwell et al., 1972): During early transplantation of a hyperplastic mammary nodule arising in a hormonally stimulated female mouse of strain BALBlc Crgl, the nodule transformed spontaneously to a histologically dimorphic malignant tumor, at which time it was transferred from the laboratory of origin (Cancer Research Genetics Laboratory, Berkeley) to Dr. R. F. Kallman of the Department of Radiology, Stanford University School of Medicine. Kallman transplanted the tumor (KHJ J) serially in BALB/c Ka mice (originally propagated from BALB/c Crgl), and over a period of at least 109 passages it remained free from any evidence of immunogenicity. The TD50 was small (2-16 cells) and could not be raised by preimmunization of recipients with lethally irradiated homologous cells or by addition of such cells to the inocula of viable cells. This is in accordance with expectations for a tumor whose origin involved no implication of chemical carcinogen or oncogenic virus, At the 25th serial passage of KHJJ, the tumor was adapted to growth in tissue culture and was propagated in culture for 33 serial passages as EMT; at the 33rd passage it was cloned as EMT6, and this derivative was henceforth alternately cultivated in vitro and in vivo in BALB/c Ka mice. However, EMTG was found to evoke “a decided immunologic response;” the TD50 was substantially higher in preimmunized mice and was boosted by addition of lethally irradiated cells to the inocula. Subsequent use of EMTG in other laboratories has revealed that the immunogenicity in BALB/c mice has been sustained at a level which can seriously complicate chemotherapy studies (A. Begg, personal communication). Since the immunogenicity of EMTG has run parallel with the nonimmunogenicity of KHJJ, from which it was derived, it is clear that the immunity is an artifact superimposed upon KHJJ by its adaptation to tissue culture; it is conceivable that the antigenic modification has been isolated and perpetuated by removal of the cell line from an influence of immunoselection. The history of EMTG is of interest because it illustrates another possible source of superimposed artifactual tumor immunogenici ty . Line 1 Lung Carcinoma (Table 111) is a tissue culture adapted cell line derived from a spontaneous alveolar carcinoma of the lung of a BALB/c mouse at its 12th serial passage in viuo; the tissue culture line was the source of cells used as inocula for subsequent in vivo studies of tumor growth (Yuhas et al., 1974). Although the original tumor

188

HAROLD B. HEWITT

maintained by serial passage in vivo has exhibited no failures to take or spontaneous regressions, more detailed study of the Line 1 Carcinoma transplanted in vivo has revealed a complex hostftumor relationship which Yuhas and his colleagues have interpreted immunologically. Yuhas et al. (1974) showed that subcutaneous tumors grew more slowly in mice immunosuppressed by whole-body irradiation (WBR) or by treatment with cortisone and that the growth depression induced b y WBR could be cancelled by the administration of syngeneic normal lymphoid cells to the mice after their exposure, Yuhas et al. (1975a) demonstrated later that there was mutual growth inhibition between the primary tumor and pulmonary metastases naturally disseminated from it or superimposed on the tumor-bearing mice by intravenous injection of tumor cells; this they interpreted as a manifestation of concomitant immunity in the system. Yuhas et al. (1975b) subsequently showed that growth of primary implanted tumors was significantly depressed by treatment of the mice with a combination of C. parvum vaccine with intravenous tumor cells. Thus, the Line 1 Lung Carcinoma must be regarded as a weakly inimunogenic system. Unfortunately the information in the relevant papers is insufficient to determine whether the immunogenicity is a consequence of the adaptation to tissue culture or was exhibited by the original unadapted tumor. Without the required information, and in the light of the clearer evidence for the EMT6 tumor, it seems wise to conclude that the immunity in the Line 1system is an artifact imposed by exposure to in vitro conditions. Epithelioma Sp. 1 (rat) arose spontaneously in a highly inbred Wistar rat and was used within the laboratory of origin in both the instances of its use which appear in the sample (Table 111); it was probably under 10 years old when used. Although this tumor has full eligibility as a spontaneous tumor and has been serially transplanted under conditions which minimize the risk of superimposed histoincompatibility, it is described as “weakly immunogenic” ( Baldwin, 1966).Rats in which primary implants had been destroyed by induction of ischemia in situ resisted challenge with viable cells of the tumor; they nevertheless went on to develop pulmonary metastases; rats could not be immunized using radiation-killed homologous cells. Thus, the evidence for immunogenicity is indeed weak. In the two investigations contributing to our sample, no immunotherapeutic effect on the growth of this tumor was obtained using either orally administered BCG or levamisole (Baldwin et al., 1975; Hopper et al., 1975). WHT Squamous Carcinoma ‘D’ arose spontaneously in my own laboratory and has been shown to be nonimmunogenic by repeated im-


189

munization studies and numerous incidental observations (Hewitt et al., 1973, 1976; Hewitt and Blake, 1975). The investigation contributing this tumor to the sample (Table 111) was carried out in another laboratory, but the tumor and a breeding nucleus of WHT mice were supplied together &om my laboratory, so that no change in the substrain of recipients was associated with transfer of the system. Mammary Carcinoma D7T4S (Table 111) arose from a hyperplastic alveolar nodule appearing in a hormonally hyperstimulated MMTVfree BALB/c female mouse and had been serially transplanted in the strain of origin for several years. Eligible as a tumor model in respect of its origin and conditions of maintenance, it is nevertheless described as being slightly immunogenic (Dao et al., 1972). The SBI Lymphoma (Table 111)is strictly of spontaneous origin and was transplanted within the BALB/c colony in which it arose; over 80% successful transplantation is achieved with 5 cells, but no formal tests for its immunogenicity have been reported. The investigation reporting its use, in our sample, did not put to test, or require the exertion of, its immunogenicity. Summarizing the content of Table 111, I conclude that, out of the 19 tumors of spontaneous origin used in the investigations contributing to the sample, 10 have been serially transplanted over such long periods that they have been heavily exposed to conditions which could have imposed upon them artifactual features affecting the hosthumor relationship (LLT, B 16 Melanoma, Ridgeway Osteosarcoma and Neuroblastoma C1300), and 3 (EMT6 and Line 1 Carcinoma) exhibit tumor immunogenicity which is likely to be associated with their maintenance under in vitro conditions. Only the remaining 6 tumors, represented by Epithelioma Sp. 1, Mammary Carcinoma D7T4S, SBI Lymphoma, and WHT Squamous Carcinoma “D” conform to the ideal specifications I have suggested: spontaneous origin, less than 15 years transplantation history, and transplanted within the substrain of origin. Referring these 6 eligible tumors to the total sample of 71 tumor usages (Table 11), including tumors which were artificially induced or allografted, w e find that over 90% of the tumors currently used in experimental therapy research are of very questionable status as models of spontaneous autochthonous cancer. This does not, of course, imply that data obtained using tumor systems of relatively low status as models do not have validity within a limited context. But in as much as results obtained using questionable systems are subject to significant influence by artifacts, it is not allowable to extrapolate findings quantitatively to systems free of such artifact. That this limitation is very little regarded is evident from the frequency and im-

190

HAROLD B. HEWITT

prudence with which clinical implications are attached to reports of experimental studies. A similar analysis to that recorded in Table I11 has been made of the animal tumor systems selected for experimental studies of the action of C . pamum vaccine on tumor growth. This was undertaken for two reasons: I t employs a rather different method of sampling from the recent literature, covering a longer period (1972 through 1975) and a relatively unrestricted range of journals, and it confines attention to a sphere of interest in which clinical application of laboratory studies is being intensively promoted. (The Compendium of Tumor Immunotherapy Trials, No. 4, August 1976, records 48 clinical trials of the immunotherapy potential of C . parvum vaccine, either in progress or projected.) It is clearly an area of research in which a specially high standard of care is to be expected in the selection of animal models of human cancer. The method of sampling was by reference to all the articles indexed under “Corynebacterium paruum” in Volumes IX through XI1 (1972-1975) of the annual publication of Roberts or Roberts et al. entitled “Research Using Transplanted Tumors of Laboratory Animals.” The sample comprised 28 articles referring to the use of a total of 46 tumors. Table IV, recording the results of the analysis from C. parvum experiments, shows that the majority (76%)of tumors were chemically induced or virus-induced. The one radiation-induced tumor was, in fact, a partial allograft in the experiment describing its use (Smith and Scott, 1972). The two uses of a tumor specified as “ex-culture” were of CBAT-3 fibrosarcoma, a tumor resulting from malignant transformation in uitro of a culture of embryonic CBA fibroblasts; it is immunogenic. Thus, only 5 (11%) of the tumors used were of spontaneous origin in viuo and had been transplanted under moninally TABLE IV

c.

CATEGORIZATION OF ANIMAL TUMORSU S E D IN STUDIES OF pU?WU?Tl VACCINE (RANDOM SAMPLE OF PUBLICATIONS 1972-1975)

Tumor origin

Number

% of total

Chemically induced Virus-induced Radiation-induced Ex-culture Allografted Spontaneous

27 8 1 2 3 5

59 17 2 4 7 11


191

isogeneic conditions. All of these 5 spontaneous tumors appeared in the previous sample, were included in the analysis recorded in Table 111, and have been discussed above. (Three were representatives of the 24-year-old Lewis Lung Tumor, 1was B 16 Melanoma, and 1 was the Line 1 Lung Carcinoma.) It is remarkable that, in this series of experiments devoted to a clinically oriented topic, there was not a single example of the use of an ideal model of spontaneous autochthonous cancer, that is, a tumor of fairly recent spontaneous origin in the laboratory in which the experiments were done and having no history of maintenance in vitro. The descriptions of the experiments contributing to the sample commonly included the results of tests for the immunogenicity of the tumors used. These were almost invariably positive, as reported in the relevant papers or by reference to previous papers. For example, Smith and Scott (1972) employed 5 different tumors in their studies of C . parvum: 3 induced tumors (Hepatoma 129, Plasmacytoma PCGA, Leukemia L5178), a radiation-induced leukemia (R-I1)transplanted to a relatively histoincompatible mouse substrain, and fibrosarcoma CBAT-3 transformed from normal fibroblasts after maintenance in tissue culture. All of these tumors were immunogenic as shown by a rise in TD50 (challenge) after immunization of recipients with lethally irradiated homologous cells. This finding for the 5 tumor systems used by Smith and Scott (1972) is in striking contrast to ours (Hewitt et al., 1976) using a range of 7 spontaneous murine tumors having unexceptionable status as models and the same immunization procedure. We found that the TD50 (challenge) was less in immunized than in normal mice in the case of all 7 tumors. Detailed study of the results of the experiments by Smith and Scott (1972) reveals that the effectiveness of C. pamum vaccine as an immunostimulant was directly proportional to the strength of artifactual immunogenicity exerted in a system; for example, the vaccine was considerably more effective against allografts of Hepatoma 129 in BALB/c mice than against grafts in mice of the strain of origin (CBA). A high proportion of the authors who have demonstrated a restraining effect of C. parvum vaccine on the growth of artifactually immunogenic tumors induced by chemicals or viruses suggest by implication, or by forthright recommendation of clinical trial, that their results have significance for clinical therapy. This is so even when the animal model used is highly immunogenic and when the mode or time of administration of the vaccine in the experiments could have no realistic representation in the clinical situation. In many cases the vaccine has been adminstered before the tumor is grafted or has been mixed with the grafted cells at the time of grafting. Rarely have the

192

HAROLD B. HEWITT

authors explicitly entertained the possibility that their positive results may merely reflect the artifactual peculiarities of the experimental system they have used, as did Pimm and Hopper (1975). IV. Reflections and Conclusions

The discipline of cancer research can be roughly divided into two areas which are distinguished by a difference of orientation in relation to clinical cancer. “Basic” or “fundamental” investigations are addressed to the general biology of growth and encompass a broad spectrum of inquiries and approaches; the question whether the products of such research will conduce to better management of clinical malignant disease is left open, as indeed it must be if the research is to progress by internal instigation. The other area of research, to which this contribution is relevant, has been called “applied,” although the teasing phrase “mission oriented” nicely takes account of the sociological influences through which “applied” research seeks approval and gains favor. It must be acknowledged that there is growing impatience with the pace of progress in clinical treatment of cancer in relation to the volume of research afforded, to which research funding bodies have responded b y pretending to clairvoyant recognition of “relevance” in the evaluation of projects for which support is sought. Whatever may be the foundation or merits of this emphasis on relevance, there is no doubt that it has influenced the general direction of cancer research in a way that underlines the need to review and revise the criteria to be regarded in the selection of animal tumor models. Pressure on researchers to claim clinical relevance for their projects has had untoward effects on the style of presentation of fundamental and apparently esoteric investigations, whereby tortuous and highly speculative projections have been made from intimate studies of cell structure or function to the prevention or treatment of human cancer. Yet this tendency is in defiance of a more sober understanding that the utility of knowledge is an eventuality not susceptible to reliable prediction (Flowers, 1972). With respect to research for which more brash claims to relevance are made, by their description as “applied,” association between those who conduct animal experiments and those who treat patients has been fostered by placing the two in ever closer proximity. Indeed, there are now several situations in which large animal laboratory facilities and clinical undertakings fall under a single director. The danger of such intimacy is that it encourages a rather direct translation of animal data to clinical prescription in circumstances where knowledge and experience in one


193

sphere may be unmatched by expertise in the other. Spectacular cures of animal tumors by treatments of negligible toxicity are regularly reported, and it is understandable that a clinical oncologist confronted b y such findings, and not made aware of the limitations of the animal model with which they were obtained, may feel an ethical obligation to confirm them clinically. Since most clinical oncologists are not equipped by training or experience to assess the status of an animal tumor as a model of clinical cancer, the responsibility for evaluating the clinical relevance of experimental data should reside with the experimenter. Tables 11, 111, and IV reveal that this responsibility is very far from being met by those on whom it rests: Only a small proportion of the animal tumor systems used in current therapy research can be claimed to be free from manifest or latent artifacts of a kind that can be highly influential on the data obtained by their use. Yet it is the rule, rather than the exception, for reports of such experimental studies to include a discussion of clinical implications if not actual recommendation of clinical application. Unrestrained enthusiasm for the general theory that spontaneously arising tumors are (at least potentially) immunogenic, based on a conviction that the creation of neoantigens is an invariable accompaniment of malignant transformation, has encouraged permissive acceptance of tumor antigenicity in an animal tumor system. It may be argued that artifactual introduction of this property serves merely to exaggerate, for experimental convenience, a feature of the natural disease. This is not an argument which can be rationally sustained. A belief in the natural immunogenic potentiality of an autochthonous tumor of spontaneous origin is adequately catered to by use of an animal tumor system which, by a priori reference to its history, can be considered to be free from a high risk of imposed artifacts. Any expression in the transplanted animal tumors of a primordial immunogenicity conceived to have been imposed b y spontaneous malignant transformation will then be exerted at a realistic level of effectiveness. Exaggeration b y artifact serves only to distort the model either quantitatively or absolutely. Since tumor immunogenicity is not detectable in many experimental systems derived from spontaneous tumors (Hewitt et al., 1976), and since distinction of natural from artifactual immunogenicity is rarely feasible, a strong case can be made for accepting for therapy studies only animal tumor models which have been shown to be nonimmunogenic. In circumstances in which the immune resistance of tumor-bearers does make a contribution to cure of the tumors by a therapeutic agent, it is often not appreciated how large the contribution may be. Mea-

194

HAROLD B. HEWITT

surement of the single dose of radiation required to cure half of a group of uniform tumors of specified size (the TCD5O) presents a specially favorable opportunity for assessing this contribution. This is so because of three features peculiar to the action of radiation on tumors. First, the dose of radiation actually delivered to every cell of the tumor, exposed under appropriate conditions, can be specified with great accuracy. Second, the cells of different tumors have closely similar radiosensitivities. And, finally, the approach to cure with increasing dose is described over the greater part of the dose/effect curve by an exponential relating the dose to the proportion of clonogenic cells reproductively killed. Thus, the TCD50 can be predicted from a generally applicable dose/effect curve in relation to an estimate of the average size of the clonogenic cell population in the tumors. For nonimmunogenic tumors of -8 mm diameter, the .TCD5O is predicted to be approximately 5000 rads, and this value has been verified in numerous radiobiological studies of such tumors (e.g., Howes, 1969). For their study of the adjuvant affect of C. purvum vaccine on the TCD50, Milas et al. (1975) employed an MCA-induced fibrosarconia displaying the immunogenicity characteristically associated with tumors so induced. Their TCD5O (without administration of C. paruum) was only 3400 rads. Thus, the TCD50 was less, b y 1600 rads, than the value of 5000 rads expected for nonimmunogenic tumors (1600 rads would reduce the clonogenic cell population of a nonimmunogenic tumor to less than 1%). Milas et ul. (1975) found that the injection of C. purvum vaccine into tumor-bearing mice on the day of irradiation reduced the TCD50 by 1000 rads. In terms of cell killing, therefore, stimulation of resistance by C. purvum added a contribution to tumor cure which was rather less than that already contributed by the unstimulated resistance intrinsic in the system. It is of interest to consider what may be the influences which are responsible for the predominant usage for therapy studies of tumor systems whose histories imply a high risk of the inclusion of artifacts. The question is most difficult to answer in respect of large, long established centers of oncological research which maintain their own breeding colonies of mice and therefore have ideal conditions for the isolation and maintenance of tumors of spontaneous origin. It is evident from a review of the relevant papers from several such large centers that the animal tumors used in them are induced or are veteran tumors imported from other laboratories. It must be conlcuded that regular surveillance of colonies of ex-breeder animals is not undertaken and that the value of accumulating a panel of internally evolved tumors of different types is not appreciated. Yet it is only by maintenance of such a facility that a need to import tumors, with inevitable


195

risk to the hygiene of resident colonies, can be avoided. That the internal resources of some large centers are not adequate to meet the demands of workers within them is evident from requests made to our own laboratory for the provision of nonimmunogenic systems. In the case of small research facilities, the lack of intrinsically evolved tumor systems is more easily understandable, especially where a breeding unit is not maintained, and mice for experiment are obtained from a commerical source as required. Perhaps the most common encouragement to hasty importation of tumors having questionable status as models arises out of the situation confronting researchers embarking on short term limited projects; in such cases, there will be too little time to await the appearance of a spontaneous tumor having the features required for the study contemplated. The risk of superimposing immunological artifact on a tumor system is specially high in the case of facilities which maintain breeding colonies of inbred animal strains but do not have tumors of origin in the resident substrains. In these circumstances, a tumor that has arisen in the same nominal strain elsewhere is imported from another laboratory and is liable to acquire artifactual immunogenicity by its transplantation across a possible minor histocompatibility barrier between the substrain of origin and that in the receiving laboratory. It has been our practice in meeting requests for transplanted tumors from our own laboratory to advise the breeding up of a colony of the relevant mouse substrain from a nucleus we supply in advance of dispatch of the tumor, and it is our experience that the advice is appreciated and acted on. We have had no reports of tumors which w e have supplied under these conditions having developed immunogenicity after transfer. In the one case in which our advice proved to have been overlooked and the CBA leukemia line we supplied was transplanted to a different CBA strain, the leukemia exhibited immunogenicity (Smith and Scott, 1972). The important question arises whether the selection of tumor systems for immunotherapy studies has been subject to bias in favor of systems which are more likely to entaiI artifactual immunogenicity. From the chemotherapy column of Table 11, it is seen that 12/38 (32%) of the tumors used were of spontaneous origin. Combining the data of the immunotherapy column of Table I1 with the data of Table IV for studies of C. pamum vaccine shows that only 12/79 (15%) of the tumors used were of spontaneous origin; the difference approaches significance (0.1> p > 0.05). It is my belief that this evident bias reflects the results of excessive promotion of the topic of immunotherapy and betrays a response to pressures on researchers to obtain results which favor the promotion; it could imply that negative results obtained with tumors of spontaneous origin have gone unreported. My

196

HAROLD B. HEWITT

analysis confirms the previous assertion by Klein and Klein (1977) that bias has perverted this area of research. These are serious considerations which raise ethical issues concerning the clinical application of experimental findings. It is clear that the quality of experimental studies undertaken within limited research facilities could be much improved, and the high incidence of laboratory artifacts reduced, if the larger institutes of oncological research undertook to collect, maintain, and monitor transplanted animal tumor systems and made these available to researchers in smaller laboratories. The cost of maintaining such resources could be largely offset by realistic charges for material supplied. There are several national bodies in existence which presume to “coordinate” cancer research and to arrange facilities for intercommunication. It would be a valuable aid to researchers if such bodies undertook to organize facilities which would encourage the use and availability of experimental tumor systems which can be authenticated as acceptable models of naturally occurring cancer. It appears that such a service might have a more telling effect on the output of relevant research than arbitrary specification of the lines of research to be followed. I suggest also that referees of scientific papers and the editors of journals assume a responsibility to disallow the drawing of clinical implications from findings made by the use of animal tumor systems which entail obvious artifacts. It is appropriate to end this contribution by a quotation from a previous article in this series: “We should abandon the common models in animals (i.e., highly antigenic chemically induced tumors or those produced by ‘laboratory’ viruses) in the same way as we abandoned the use of allogeneic models for the study of relevant immunologic events , . . and concentrate on spontaneous tumors appearing in experimental animals” (Stutman, 1975).

ACKNOWLEDGMENTS The work of the author referred to in this chapter was supported exclusively by the Cancer Research Campaign. I gratefully acknowledge the skilled technical assistance of Miss Eileen Blake, A.I.M.L.T., and Miss Angela Walder, A.I.A.T., whose work has included the isolation and maintenance of over thirty murine tumors of spontaneous origin. I am grateful also to Mrs. Karen Jepson for assistance in preparing the manuscript.

REFERENCES Altman, P. L., and Dittmer, D . S.(1964).“Biology Data Book.” Fed. Am. SOC.Exp. Biol., Washington, D.C.


197

Attia, M. A. M., and Weiss, D. W. (1966). Cancer Res. 26, 1787. Atassi, G., and Taignon, H. J. (1974). Eur. J . Cancer 10, 399. Bailey, D. W., and Mobraaten, L. E. (1964).Genetics 50, 233 (abstr.). Baldwin, R. W. (1966). Znt. J . Cancer 1, 257. Baldwin, R. W., and Glaves, D. (1968).Br. Emp. Cancer Campaign Annu. Rep. 46,236. Baldwin, R. W., Hopper, D. G., and Pimm, M. V. (1975). Br. J . Cancer 31, 124. Berry, R. J., and Andrews, J. R. (1961). Am. N.Y. Acad. Sci. 95, 1001. Biano, G., Brown, B. L., Jones, E. E., and Rosender, V. M. (1971). Proc. S O C . E x p . Biol. Med. 136, 507. Bierman, H. R., Aggeler, P. M., Thelander, H., Kelly, K. H., and Cordes, F. L. (1956).]. Am. Med. Assoc. 161,220. Birch, J. M., Moore, M., and Craig, A. W. (1975). Br. J . Cancer 31,630. Bland Sutton, J. (1890). “Evolution and Disease.” Walter Scott, London. Breyere, E. J. (1972). Natl. Cancer Znst., Monogr. 35, 289. Bross, I. D. J., and Blumenson, L. E. (1971). Cancer (Philadelphia) 28, 1637. Bystryn, J.-C., Bart, R. S., Livingston, A., and Kopf, A. W. (1974).J.Znoest. Dermatol. 63,369. Carnaud, C., Hoch, B., and Trainin, N . (1974).J . Natl. Cancer Znst. 52, 395. Carter, S. K. (1973). Eur. J . Cancer 9, 833. Chalmer, J., Holt, P. G., and Keast, D. (1975).J . Natl. Cancer Inst. 55, 1129. Ciuffo, G. (1907). G. Ztal. Mal. Vener. 42, 12. Collins, M. J., and Parker, J. C. (1972).]. Natl. Cancer Znst. 49, 1139. Corbett, T. H., Griswold, D. P., Roberts, B. J., Peckham, J. C., and Schabel, F. M. (1975). Cancer Res. 35, 2434. Dao, T. L., Varela, R., and Morreal, C. (1972).In “Estrogen Target Tissues and Neoplasia” (T. L. Dao, ed.), pp. 163-179. Univ. of Chicago Press, Chicago, Illinois. Del Regato, J. A. (1972). Cancer (Philadelphia) 29, 1443. De ThC, G., Geser, A., and Munoz, N. (1976). In “Scientific Foundations of Oncology” (T. Symington and R. L. Carter, eds.), pp. 414-425. Heinemann, London. DeWys, W. D. (1972). Cancer Res. 32,374. Dunham, L. J., and Stewart, H. L. (1953).J.Natl. Cancer Znst. 13, 1299. Easson, E. C. (1968).In “Prognostic Factors in Breast Cancer” (A. P. M. Forrest and P. B. Kunkler, eds.), p. 118. Livingstone, Edinburgh. Editorial. (1942). Br. Med. J . 1,644. Everson, T.C. (1964). Ann. N.Y. Acad. Sci. 114, 721. Fieldsteel, A. H., Dawson, P.J., and Kurahara, C. (1973). Cancer Res. 33, 551. Fieldsteel, A. H., Dawson, P. J., Kurahara, C., and Brooks, R. E. (1975).Br.]. Cancer 32, 741. Fisher, B. (1973). Cancer (Philadetphia)31, 1271. Flowers, B. (1972). Br. J . Radiol. 45,803. Foley, E. J. (1953). Cancer Res. 13,835. Gaugas, J. M., Chesterman, F, C., Hirsch, M. S., Rees, R. J. W., Harvey, J. J., and Gilschrist, C. (1969).Nature (London) 221, 1033. Gesser, I., and Bourak-Maury, C. (1972). Nature (London),New Biol. 236, 78. Godfrey, J., and Searle, A. G. (1963). Genet. Res. 4, 21. Graff, R. J., Valeriote, F., and Medoff, G. (1975).J . Natl. Cancer Znst. 55, 1055. Gross, L. (1943). Cancer Res. 3, 326. Gullino, P. M., Grantham, F. H., and Courtney, A. H. (1967). Cancer Res. 27, 1010. Haddow, A., and Alexander, P. (1964). Lancet 1,452. Hageman, P. C., and Miihlbock, 0. (1976). In “Scientific Foundations of Oncology” (T. Symington and R. L. Carter, eds.), pp. 380-388. Heinemann, London.

198

HAROLD B. HEWITT

Hawkes, M. J., Hill, R. P., Lindop, P. J., Ellis, R. E., and Rotblat, J. (1968).B r . ] . Radiol. 41, 134. Henle, G., and Henle, W. (1970).J . Infect. Dis. 121, 303. Hewitt, H. B. (1962). Int. Congr. Radiol., Trans., l o t h , 1962 Abstract No. 980. Hewitt, H. B. (1976). In “Fundamental Aspects of Metastasis” (L. Weiss, ed.), p. 343. North-Holland Publ., Amsterdam. Hewitt, H. B., and Blake, E. (1975). Br.]. Cancer 31,25. Hewitt, H. B., Blake, E., and Porter, E. H. (1973). Br. J . Cancer 28, 123. Hewitt, H. B., Blake, E. R., and Walder, A. S. (1976). Br. J . Cancer 33, 241. Hill, M. J., Hawksworth, G., and Tattersall, J. (1973). Br. J . Cancer 28, 562. Hopper, D. G., Pimm, M. V., and Baldwin, R. W. (1975). Br. J . Cancer 32, 345. Howes, A. E. (1969). Br. J . Radiol. 42, 441. Hummel, K. P., Richardson, F. L., and Fekete, E. (1966).I n “Biology of the Laboratory Mouse” (E. L. Green, ed.), 2nd ed., p. 247. McGraw-Hill, New York. Johnson, R. K., Houchens, D. P., Gaston, M. R., and Goldin, A. (1975). Cancer Chemother. Rep., Part 1 59,697. Kessler, I. I., and Lilienfeld, A. M. (1969).Adu. Cancer Res. 12, 225. Klein, G. (1968). Cancer Res. 28,625. Klein, G. (1975). N. Engl. J . Med. 293, 1353. Klein, G., and Klein, E. (1977). Transplant. Proc. 9, 1095. Klein, G., Sjogren, H. O., Klein, E., and Hellstrom, K. E. (1960). Cancer Res. 20, 1561. Klein, J. (1975). “Biology of the Mouse Histocompatibility-2 Complex.” SpringerVerlag, Berlin and New York. Kobayashi, H., and Kodama, T. (1966). Nature (London) 212,939. Kobayashi, H., Sendo, F., Shirai, T., Kaji, H., Kodama, T., and Saito, H. (1969).J. Natl. Cancer Inst. 42, 413. Kreel, L., and Tavill, A. (1973). Br. J . Radiol. 46, 43. Laster, W.-R., Jr. (1975). Cancer Chemother. Rep., Part 2 5, 151. Lavrin, D. H. (1970). Cancer Res. 30, 1156. Lavrin, D. H., Blair, P. B., and Weiss, D. W. (1966). Cancer Res. 26, 929. Leading Article. (1977). Br. Med. J . 1, 187. Lemon, P. G. (1967). In “Pathology of Laboratory Rats and Mice” (E. Cotchin and F. J. C. Roe, eds.), p. 25. Blackwell, Oxford. Lindgren, M., Borgstrom, S., and Landberg, T. (1968). In “Prognostic Factors in Breast Cancer” (A. P. M. Forrest and P. B. Kunkler, eds.), p. 103. Livingstone, Edinburgh. Little, C. C., and Tyzzer, E. E. (1916).J. Med. Res. 33, 393. Maderna, C. (1935). R i f o m a Med. 51,93. Magrath, I. T., and Ziegler, J. L. (1976). Br. Med. J . 1, 615. Malmgren, R. A., Bennison, B. E., and McKinley, T. W. (1952).Proc. S O C .E x p . Biol. Med. 79, 484. Maruyama, Y. (1968). Znt. J . Cancer 3,593. Math& G. (1972). Ann. Inst. Pasteur, Paris 122,855. Mayo, J. G., Laster, W. R., Jr., Andrews, C. M., and Schabel, F. M., Jr. (1972). Cancer Chemother. Rep., Part 1 56, 183. Medina, D., and DeOme, K. B. (1970). Cancer Res. 30, 1055. MBnard, S., Colnaghi, M. I., and Cornalba, G. (1973). Br. J . Cancer 27, 345. h.leyer, J. A. (1973). Cancer (Philadelphia) 31, 1468. Mihich, E. (1969). Cancer Res. 29, 2345. Milas, L., Hunter, N., and Withers, H. R. (1975). Cancer Res. 35, 1274. Murphy, E. D. (1966).I n “Biology of the Laboratory Mouse” (E. L. Green, ed.), 2nd ed., p. 521. McGraw-Hill, New York.

CHOICE O F ANIMAL TUMORS FOR THERAPY STUDIES

199

Niederman, J. C., Evans, A. S., Subrahmanyan, L., and McCoIlum, R. W. (1970).N . Engl. ]. Med. 282, 361. Odashima, S. (1964). Natl. Cancer Inst., Monogr. 16,51. Outzen, H. C., and Prehn, R. T. (1973).Cancer Res. 33, 408. Parker, J. C., Tennant, R. W., and Ward, T. C. (1966).Natl. Cancer Inst., Monogr. 20,25. Parmiani, G . (1970). I n t . ] . Cancer 5, 260. Pearlman, A. W. (1976). Cancer (Philadelphia) 38, 1816. Pimm, M. V., and Hopper, D. G. (1975). Lancet 1,806. Prehn, R. T. (1963).]. Natl. Cancer Inst. 31,791. Prehn, R. T. (1969).Ann. N.Y. Acad. Sci. 164, 449. Prehn, R. T. (1975).J . Natl. Cancer Inst. 55, 189. Prehn, R. T., and Main, J. M. (1957).J.Natl. Cancer Inst. 18, 769. Rapp, H. J. (1972).Natl. Cancer Inst., Monogr. 39, 1. Reinhold, H. S., and Buisman, G. H. (1973).Br. 1.Radiol. 46, 54. Renoux, G., and Renoux, M. (1972). Nature (London)New Biol. 240, 217. R&Csz, L. (1960).Cancer Res. 20, 443. Rice, J. M. (1972).Natl. Cancer Inst., Monogr. 35, 197. Roberts, D. C. (1970). “Research Using Translational Tumors of Laboratory Animals. A Cross-referenced Bibliography,” Vol. VII. Imp. Cancer Res. Fund, London. Roberts, D. C., and Drobycz, B. (1975).“Research Using Transplanted Tumors of Laboratory Animals. A Cross-referenced Bibliography,” Vol. XII. Imp. Cancer Res. Fund, London. Rockwell, S. C., Kallman, R. F., and Fajardo, L. F. (1972).]. Natl. Cancerlnst. 49,735. Rous, P. (1965).Nature (London) 207,457. Schwartz, H. S., Sodergren, J. F., Stjernberg, S. S., and Philips, S. S. (1966).Cancer Res. 26, Part 1, 1873. Schwartz, R. S. (1968).Cancer Res. 28, 1452. Silvers, W. K., and Gasser, D. L. (1973). Genetics 75,671. Smith, S. E., and Scott, M. T. (1972).Br. J . Cancer 26, 361. Snell, G. D. (1958).I. Natl. Cancer Inst. 20, 787. Snell, G. D., and Stimpfling, J. H. (1966). I n “Biology of the Laboratory Mouse” (E. L. Green, ed.), 2nd ed., p. 457. McGraw-Hill, New York. Spratt, J. S., and Spratt, T. L. (1964).Ann. Surg. 159, 161. Spreafico, F., Vecchia, A., Mantovani, A., Poggi, A., Franchi, G., Anaclerio, A., and Garattini, S. (1975). Eur. J . Cancer 11, 555. Stewart, H. L., Snell, K. C., Dunham, L. J., and Schlyens, S. M. (1959). “Transplantable and Transmissible Tumors of Animals.” Armed Forces Inst. Pathol., Washington, D.C. Stutman, 0. (1975).Adu. Cancer Res. 22, 261. Sugiura, K., and Stock, C. C. (1955). Cancer Res. 15, 38. Suit, H. D., and Silobrcic, V. (1967).]. Natl. Cancer Inst. 39, 1121. Suit, H. D., Shalek, R. J., and Wette, R. (1965).I n “Cellular Radiation Biology,” p. 25. Williams & Wilkins, Baltimore, Maryland. Svet-Moldavsky, G. L., Liozner, A. L., Mkheidze, D. M., Sokolov, P. P., and Bykovsky, A. P, (1970).I . Natl. Cancer Inst. 45, 475. Vaage, J. (1968). Cancer Res. 28, 2477. Vaage, J., Kalinovsky, T., and Olson, R. (1969). Cancer Res. 29, 1920. Variot, G. (1894).1. Clin. Ther. Inj., Paris 2, 529. Warwick, G. P. (1976). I n “Scientific Foundations of Oncology” (T. Symington and R. L. Carter, eds.), p. 302. Heinemann, London. Waynforth, H. B., and Magee, P. N. (1974). B r . ] . Cancer 30,512.

200

HAROLD B. HEWITT

Weiss, D. W. (1967).In “Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability” (L. Lecam and J. Neyman, eds.), p. 657. Univ. of California Press, Berkeley. Weiss, D. W. (1977). Cancer Immunol. Immunother. 2, 11. Yuhas, J. M., Pazmino, N. H., Proctor, J. O., and Toya, R. E. (1974).Cancer Res. 34,722. Yuhas, J. M., Pazmino, N. H., and Wagner, E. (1975a). Cancer Res. 35, 237. Yuhas, J. M., Toya, R. E., and Wagner, E. (1975b). Cancer Res. 35, 242.

.

ADVANCES IN CANCER RESEARCH VOL . 27

MASS SPECTROMETRY IN CANCER RESEARCH'

John Roboz Department of Neoplastic Diseases. Mount Sinai School of Medicine. New York. New York

I . Scope of Applications and Analytical Techniques ......................... A . Advantages and Limitations ......................................... B . Nature of Mass Spectra .............................................. C . Components of Mass Spectrometer Systems .......................... D . Analytical Techniques .............................................. ..... I1 Identification. Quantification. and Metabolism of Carcinogens . A . Survey Analyses .................................................... B . Polycyclic Aromatic Hydrocarbons ................................... C Polychlorinated Biphenyls ................................... D . Bis-Chloromethyl Ether ............................................. E . Vinyl Chloride .................. F. Nitrosamines ................ ........................... G . Miscellaneous ............... H . Trace Elements . . ............................................. 111 . Metabolism and Monit of Antineoplastic Agents ..................... A . Cyclophosphamide ........................................... B . Nitrosoureas ............................ ......................... C . Purines. Pyrimidines. and Their Nucleosides ......................... D . Daunorubicin and Adriamycin ....................................... E . Antitumor Hormones ................................................ F. Hexamethylmelamine ............................................... G . Platinum Coordination Complexes ....................... H . Miscellaneous ...................................................... I V. Biological Markers ......................................... A . General ................................................... B. Polyamines ......................................................... C . Steroids .................................. ....................... D . Miscellaneous ...................................................... E . Conclusions ........................................................ References ......................................... ........

.

.

202 202 204 206 212 215 215 217 224 225 227 228 231 232 233 233 241 243 245 249 250 250 251 253 253 254 256 257 259 260

'This work was partially supported by Grant # 5PO1-CA15936-03from the National Cancer Institute. National Institutes of Health . 20 1

.

Coplright @ 1978 b y Acsdeniir Prrc. Inc . A l l rights of reproduction in ;my fanil reserved . ISBN (I-12-006627-(I

202

JOHN ROBOZ

I. Scope of Applications and Analytical Techniques

A. ADVANTAGESAND LIMITATIONS Computerized GC/MS2 is probably the most comprehensive and versatile of all instrumental methods of analysis. A mass spectrometric analysis provides both unequivocal identification and quantification of the constituent(s) of interest, often in the presence of a host of similarly behaving native biological or artificial chemical materials. Recent developments in MS and computer instrumentation and analytical techniques elevated mass spectrometry to a unique status: high sensitivity and high selectivity are provided simultaneously, and are combined with general applicability and small sample requirements. Virtually any material, organic or inorganic, can be analyzed by MS as long as the molecular weight of the compound, or a suitable derivative of it, is less than about 1200. (Compounds with molecular weight up to 3000 may be analyzed when conditions are favorable). This limitation is imposed b y the requirement that the samples must be volatilized without decomposition, so that a vapor pressure of at least lo-' torr is obtained in the ion source of the MS. Although compounds of low rnolecular weight account for no more than 25% of the normal constituents of a cell, they include the end products of the endogenous metabolic processes in the body, volatile effluents in breath, and most drugs and carcinogens and their metabolites; all elements in the periodic table are also included. An important feature of MS is the small sample requirement: 1-5 p g for identification of an unknown, 100 ng for confirmation of expected identity, and 0.05-5 ng for quantification of known contituents in biological materials. Another advantage of MS is speed. Confirmation of identity and/or quantification can often be accomplished in minutes or hours. Identification of an unknown may, of course, take days or weeks. A disadvantage is that materials cannot be recovered after analysis. When a sample is ionized, ions are formed simultaneously from all constituents present, and the mass spectrum does not readily reveal *Abbreuiations: MS, GC, HPLC, TLC refer to the technique or instrument, depending upon context, of mass spectrometry, gas chromatography, high pressure liquid chromatography, or thin layer chromatography, respectively; G U M S and HPLC/MS refer to the combined techniques or instruments; Ion sources or ionization techniques (Section 1,G); EI = electron (bombardment) ionization, CI = chemical ionization, FD = field desorption ionization; pg, ng, and pg refer to microgram, nanogram, and picogram, respectively.

MASS SPECTROMETRY IN CANCER RESEARCH

203

whether the original sample was a mixture or a pure compound. Thus, a pure sample must be provided when the complete mass spectrum of an unknown is needed for identification. Samples may be purified by chemical or chromatographic methods and presented for MS analysis in solution; removal of solvents presents no problem. In another approach, a GC or LC is combined with the MS, and the pure sample is provided when the chromatographically separated component of interest arrives on-line into the ion source of the MS. There are four areas in which MS can be applied to problems in cancer research and therapy: (1) confirmation of identity, (2)identification of unknowns, (3) quantification of selected trace constituents, and (4)“profile” surveying, in a single analysis, of a large number of endogenous or drug metabolites, closely related groups of carcinogenic chemicals, or all elements in the periodic table. Mass spectrometry is also an important tool in the elucidation of the structure of complex molecules; however, the technique alone is usually not adequate. For example, NMR, IR, optical rotary dispersion, as well as MS were needed to establish the structure of daunomycin (Arcamone et al., 1968). Montgomery and Struck (1976) synthetized some 80 cyclic and acyclic anlogs of cyclophosphamide, isophosphamide, triphosphamide, and phosphoramide mustard; 19 were exclusively and 40 primarily characterized by mass spectrometry. To confirm the presence of a compound, one needs enough material to obtain a reasonably complete mass spectrum which is compared to that of the authentic compound. Mass spectra of more than 30,000 compounds are available in data libraries, and computerized matching may be made by telephone in minutes. “Unknowns” may be biologically active constituents isolated b y some means, peaks in the gas or liquid chromatograms obtained in profile searching, metabolites of an antineoplastic drug or a carcinogen, or impurities present in the drugs or carcinogens studied. The structure of these compounds is usually known once their identity is established. They may be of significance or they may be artifacts. Unfortunately, it takes just as much work to identify an unimportant constituent as it takes for a significant one. Identification is simple when the mass spectrum obtained is similar to one available in a data library. When the mass spectrum of the compound encountered is not available, the best approach is to determine the exact molecular weight of the molecular ion (accurate mass determination using high-resolution mass spectra) and calculate possible molecular compositions. The use of MS to quantify trace amounts, down to picomole levels of

204

JOHN ROBOZ

selected known components in relatively crude mixtures is based upon a new technique called selected ion monitoring (Section 1,D). In profile analysis, constituents of similar chemical nature are sepa: rated and quantified in a single analysis. Separation is accomplished by GC or LC followed by characterization and quantification using MS. Areas of application include searching for qualitative or quantitative changes in endogenous metabolic constituents caused by the disease (biological markers), or b y an antineoplastic drug or carcinogen, characterization and quantification of sets of isomeric carcinogens, and exploratory surveying of all inorganic elements in a single analysis. After a brief discussion of the basic principle of MS, three areas of application are reviewed: ( 1) identification, quantification, and metabolism of carcinogens, (2) metabolism and monitoring of antineoplastic agents, and ( 3 )searching for biological markers. The literature is covered for the period 1970 to early 1978. No attempt was made for completeness. Instead, a variety of approaches and techniques are described, and applications are illustrated in as many different areas as possible. A list of recommended texts and sources of current information is given at the end of the chapter.

B. NATUREOF MASS SPECTRA The process whereby a neutral molecule or atom becomes electrically charged in the vapor phase is called ionization. There is a minimum amount of energy (6-15 eV) that must be provided to remove an electron from the lowest energy orbital from a molecule. The unbroken molecule which thus becomes a positive ion is called a molecular ion (M+'). When excess energy is provided, it is transformed into vibrational energy by radiationless transitions, followed b y distribution over all integral degrees of freedom. When the vibrational energy concentrated in a particular bond becomes equal to the dissociation energy of that bond (2-5 eV for single bonds), the molecular ion dissociates and a fragment ion is formed. When there is still more excitation energy available, additional cleavages may take place and still smaller fragment ions will form. The fragmentation processes of polyatomic molecules are considered to be a series of competing and consecutive unimolecular reactions, similar to the rate processes characterizing ordinary chemical reactions. Mass spectrometers are ion optical devices which produce a beam of gaseous ions &om an evaporated sample, separate the resulting mix-


205

ture of ions according to their mass-to-charge ( d e ) ratios, and provide output signals from which the nominal or exact mass and abundance of each detected species may be determined. The mass spectrum of a compound reveals in a graphical, pictorial, or tabular form the masses and intensities of those ionic species for which output signals were obtained. When the intensities of the masses are normalized with respect to the highest intensity (base peak), the array obtained is called the mass spectral fragmentation pattern. The nature and degree of the fragmentation of an organic molecule greatly depends upon the total energy received during ionization. Depending upon the type of ion source used (Section 1,C) and also upon experimental conditions for any particular ion source, a mass spectrum may contain only one or a few masses, or possibly as many as 100 or more masses. Thus, mass spectra are not well-defined properties of molecules, and there is no such thing as the “right” or “correct” spectrum of a compound. For example, the EI mass spectrum of the methyl derivative of 5fluoro-2’-deoxyuridine (FUdR) is so vastly different from the CI mass spectrum (Fig. 1) that one would not even suspect that both represent the same original compound. The fact that several kinds of mass spectra may be obtained for most materials is not a disadvantage. On the contrary, judicious selection of the technique of ionization and/or experimental conditions greatly expands the art. s of application of mass spectrometry. Although the fragmentation pattern of a compound may change significantly when experimental conditions (e.g., ion source temperature) are changed, the patterns are remarkably constant for a set of given conditions. In addition, most organic compounds are characterized by individual fragmentation patterns (“fingerprints”). These two features form the bases of the use of MS to confirm identity. Mass spectra provide two basic kinds of information: the “masses” of all ionic species detected and their corresponding intensities. “Mass” (mle)refers to mass-to-charge ratio, where mass is measured in atomic mass units (W = 12.00000 by definition) and “charge” refers to the number of positive charges acquired in the course of ionization. Usually one deals only with singly charged positive ions, i.e., e = 1. Negative ion MS, in spite of current advances (Hunt et al., 1977),is of only moderate interest at this time. When masses are determined to an accuracy of 0.5-LOU (atomic mass unit) in a low-resolution MS, the terms nominal or unit mass are used. When the accuracy of mass measurement is 0 . 0 0 1 ~ in high-resolution instruments, the terms exact or precise mass are employed. 8

JOHN ROBOZ

A

r75

H3CX5F

0

I

HEOW OCH3

I/

60 -

100

0

150

250

0 200

300

I ! , , 350

n/r

,

,

, I El

FIG. 1. (A) Electron impact ionization mass spectrum of trimethyl-5-fluoro-2’deoxyuridine. The molecular ion atmle = 288 is present but of very low abundance. The peak at mle = 113 originates from the fragment of mle = 145 after further loss of -CH,OH. The peak at mle = 87 originates from the breakup of the pyrimidine ring. (B) Chemical ionization (methane as reagent gas) of methylated FUdR. The peak of highest intensity is the quasimolecular ion (M + 1).The small peaks at masses higher than the molecular weight originate from typical association reactions with the reagent gas and can usually be ignored.

c. COMPONENTS OF MASS SPECTROMETER SYSTEMS Most mass spectrometers consist of five functional elements: sample inlet system, ion source, mass analyzer, ion detector and recorder, and vacuum system. Dedicated minicomputers are employed so

MASS SPECTROMETRY I N CANCER RESEARCH

207

frequently that they soon may be considered integral parts of MS systems.

1. Inlet Systems Solid samples or compounds dissolved in a volatile solvent are placed in a small, one-end-closed capillary which is introduced into the middle of the ion source via a vacuum lock. Here the sample is vaporized by heating the capillary tube. These “direct insertion probes” may be initially cooled to prevent evaporation in the warm ion source before desired operational conditions are attained. The sample may be a pure solid, an isolated peak from GC, HPLC, or TLC, or even a crude biological mixture. Although the direct probe will yield mixed mass spectra when several components with similar vapor pressure are present, selected components may still be detected and/or quantified with selected ion monitoring (Section 1,D). The technique is sensitive: 50 ng material is often adequate to take several mass spectra. GCs are now routinely interfaced with both low- and high-resolution MS of all types. Since GC utilizes a carrier gas (usually He) at atmospheric pressure while MS must operate under vacuum torr in the ion source, in the analyzer) an interface (separator, enricher) is needed to remove the carrier gas without removing the sample being analyzed. In the jet separators, the GC effluent passes a small orifice as an expanding supersonic jet stream. The light helium atoms diffuse to the area around the heated jet and are pumped away, while heavier organic molecules continue on a straight line and enter the ion source. In the effusion-type separators the GC effluent passes cm. Helium passes through a heated glass tube with pores of through the pores and is pumped away while the organic molecules continue on a straight line. In the semipermeable membrane separators the organic materials d i h s e through a thin (0.0025 cm) membrane made of dimethyl-silicone polymer in which the carrier gas has poor solubility. With these separator$ 40-99% of the carrier gas is removed, and usually no more than 10-20% of the sample is lost. When all GC effluent enters the C I source, the same gas may be used as GC carrier and CI reagent gas (see next section), and the interface system, with its inherent problems of plumbing, may be eliminated. Combined HPLCiMS is currently under development; this technique is likely to become a major tool in biomedical analysis (McFadden and Schwartz, 1976; McFadden et al., 1977). Pilot studies for the separation of benzo(a)pyrene and pyrene with direct on-line combination of thin layer chromatography and mass spectrometry were reported by Issaq et al. (1977).

208

JOHN ROBOZ

2. lon Sources Of the 13 distinct methods available for ionization (Milne and Lacey, 1974), four are important in biomedical applications: electron (electron impact, electron bombardment), chemical, field, and spark ionization. The first three are used for the analysis of organic compounds, the last one for the elements. In the EI source the vapors of the sample are bombarded by energetic (70 eV) electrons emanating from a heated rhenium wire. All ions formed are repelled from the area of ionization by a weak positive voltage and leave the ion source via a small exit slit to be accelerated on their way into the mass analyzer. Because the amount of energy available is much more than needed for ion formation, fragment ions form in great abundance. In fact, the M+ ions often cannot he detected because all ions formed undergo additional fragmentation. In the CI source, a large amount of a reagent gas (1torr pressure) is cointroduced with the sample. The reagent gas is ionized with an electron beam in a conventional manner. For example, CH4+,CH3+, etc., ions form when methane is used as the reagent gas. Next, these primary ions enter into secondary ion-molecule reaction with unreacted methane molecules and form secondary ions such as CH5+and CzHs+,and even tertiary products such as C3H5+and C3H,+. These reagent gas ions collide with the neutral molecules of the sample, present only in very small quantity, and ionize them by ion-molecule collisions. If the sample is a good proton acceptor, a proton is added and an (M 1)+ion (quasimolecular ion) forms. Because the ions of the reagent gas do not have much excess energy to transfer, CI yields simple spectra with a high abundance of the (M 1)+ ions and relatively little fragmentation (Fig. 1B). In an FD ion source, the sample is absorbed from a solution onto an activated ion emitter which is a thin wire on which millions of microneedles of carbon are “grown.” In the presence of a strong electrostatic field ( lo8V/cm) created around the wire or a sharp blade, there is a finite probability for electrons to tunnel from the outside orbitals of molecules into unoccupied orbitals of the metal. The field desorption takes place on the multiple fine edges of the whiskers which also serve as sample reservoir. Since little energy is transferred, the spectra contain an abundance of M+‘ and (M 1)+or (M - l)+(hydrogen abstraction) ions. In the radiofrequency spark source, pulses of an ac potential of 20-100 kV are generated for a few microseconds between two electrodes formed by the sample material. Sparks may be made one at a time, or at a repetition rate as high 10“ per second. In the electrical

+

+

+


209

discharge the sample material evaporates and a great abundance of ions of all types is formed. Most elements in the periodic table can be analyzed with similar sensitivity in a single analysis. Response is linear with concentration, and samples as small as a few micrograms may be analyzed. Nonconducting samples, such as water, are deposited onto a conducting electrode material of ultrahigh purity.

3. Mass Analyzers The objectives of mass analyzers are: first, to resolve an ion beam of mass m &om another beam of nearly the same mass m + Am (dispersive or prism action) and second, to maximize the resolved ion intensities (focusing or lens action). Most mass analyzers are based on either magnetic deflection or quadrupole filtration of ions. In magnetic analyzers, ions traveling through a magnetic field experience a force that is perpendicular to both the magnetic field and their direction of travel. The radius of curvature of the path of an ion depends on its mle value, the voltage through which the ion was accelerated prior to entering the magnetic field, and the strength of the magnetic field. For example, in a 90” sector magnetic field lighter ions describe a shorter path than heavier ions (Fig. 2). Under given conditions of accelerating voltage and magnetic field strength, only ions of a particular mle value will hit the electrical detector placed in a preselected place; all other ions will hit the walls. For successive focusing of ions of smaller or larger mle on the collector, the accelerating voltage or the magnetic field strength must be “scanned.” Singlefocusing mass analyzers focus ions homogeneous as to their mass and velocity but of slightly different initial directions. These analyzers have a quadratic mass scale due to the square functions of both magnetic field strength and radii of curvature. Single-focusing magnetic mass analyzers are normally used, alone or in GUMS combinations, to determine nominal masses. In double-focusing instruments an electrostatic field is added to the magnetic field to provide velocity focusing, i.e., focusing ions homogeneous as to their mass and initial direction but of different initial velocity, in addition to direction focusing. Of course, the magnetic field will also provide mass dispersion. There are several ways to combine electrostatic and magnetic fields. In the Mattauch-Herzog design an electrostatic field with a rotation angle of 30” is combined with a 90” magnetic field. When both magnetic field strength and acceleration voltage are kept constant (no scanning), all ions describe a different path length and are detected simultaneously on an ionsensitive photoplate. In another popular double-focusing arrangement

2 10

JOHN ROBOZ

FIG. 2. Schematic drawing of 90" sector field magnetic mass spectrometer. (HVP = high vacuum pumping line; EID = electron impact ionization detector used for the recording of a total ion chromatogram which resembles gas chromatograms obtained with flame ionization detectors; SEM = secondary electron multiplier ion detector.) (Courtesy of Varian Associates.)

(Nier-Johnson geometry), a 90" electrostatic analyzer is followed by a 60" magnetic analyzer; only electrical detection with scanning is available. Double-focusing instruments are used for exact mass determinations either when an unknown must be identified or when interferences of nearly identical masses must be resolved. In the quadrupole mass filters (Fig. 3) the accelerated ions enter along the long axis of four precisely aligned parallel cylindrical rods. A programmed combination of radiofrequency (rf) and dc voltages is applied to diagonal pairs of the electrodes. At any given combination of the fields, only ions of one particular mle value can maintain a stable trajectory and reach the detector at the other end; all other ions have unstable trajectories and eventually hit one of the rods. Mass scanning is by varying the dc voltage and the rf voltage in unison to maintain a constant ratio. The mass scale of quadrupole analyzers is linear and they provide only nominal masses. Quadrupole instruments have become the choice for many applications because of their adaptability


211

FIG.3. Schematics of the operation of a quadrupole type mass spectrometer.(Courtesy of Finnigan Corporation.)

for combinations with GC, and recently LC, and relative ease of computerization. The most important measure of performance of an MS is resolution. Resolution of 600 means, for practical purposes, that masses of 600 and 601 can be distinguished. Low resolution instruments, including combined G U M S systems of both magnetic and quadrupole type, provide resolution of 600-2500. Double-focusing high-resolution magnetic instruments offer resolution in the 10,000-100,000 range. The importance of high resolution lies in the ability to separate ions of the same nominal mass that have differing elemental compositions and therefore different exact masses resulting from the packing fraction of the monoisotopic atomic weights of the elements making u p the compositions of the ion. When masses are determined to an accuracy of 0.003 u or better, unequivocal empirical formulas may be calculated. 4. Ion Current Detectors and Recorders Most MS systems employ secondary electron multipliers to detect and amplify the separated ion currents. Separated ion beams emerging from the mass analyzer are accelerated to 2-5 kV and impinge on a

212

JOHN ROBOZ

“conversion dynode” from which secondary electrons are released (Fig. 2). These, in turn, are focused on a second dynode from which the number of electrons released is greater than the number impinging upon it. In a 20-stage multiplier structure the amplification is about 10’. The final collector is connected to a high-speed electrometer the output of which is fed into a high-speed oscillographic recorder, an oscilloscope, a magnetic tape recorder, or a dedicated minicomputer, Ion currents measured are normally in the 10-8-10-16 A range, but even the arrival of 1 iodsec (1.6 x A) may be measured. Ion-sensitive photoplates are composite collector-transducer detectors. Exact masses are determined from the position of the lines in comparison with coanalyzed mass standards (perfluorokerosene) with the aid of a microdensitometer. Ion intensities are determined from the blackness of the lines.

5. Computerization Dedicated minicomputers, with continually increasing perfonnancelcost ratios, are becoming the rule rather than the exception for the acquisition and handing of data. They maximize the productivity of both tfie instrument system and operator. In interactive data systems, computers control the operation of both the MS and GC, acquire all data regardless of their significance, display data during acquisition so that immediate decisions may be made to change operational conditions if needed, and permit interrogation of all data at the convenience of the operator so that unwanted data can be discarded but useful data will not be missed (Roboz, 1977). As data processors, computers calculate peak intensities, background corrections, nominal and exact masses, chromatographic retention times, peak areas, and concentrations. They normalize spectra in several ways so that highspeed comparisons with libraries becomes possible, calculate possible molecular compositions corresponding to a determined exact mass value, and have many other uses of convenience.

D. ANALYTICALTECHNIQUES 1. Nominal and Exact Mass Measurements For confirmation of identity, a reIatively complete mass spectrum determined to an accuracy of 0.5-1.0 mass unit is obtained and compared to spectra in a data library (Milne and Heller, 1976). For quan-


213

tification at trace levels, nominal masses of only one or two selected ions are monitored (see below). Masses determined to an accuracy of 0.001 mass unit are used for the identification of unknowns and in cases in which interfering constituents of nearly identical mass cannot be resolved chromatographically. Nominal or exact mass measurement techniques may be used in straight mass Spectrometry with all types of ion sources and in GUMS.

2. Selection of Zonization Technique The mode of ion production determines if the ions will indeed be representative of the sample, and also if the mass spectra will provide the kind of information desired. Current tendency is for combined operation of sources, e.g., EI/CI or EI/FD (Gierlich et al., 1974).When structural information is desired, EI spectra is the first choice because of the considerable fragmentation, particularly involving the skeleton of the compound. Thousands of spectra are available in libraries and much is known about correlations between structure and spectra; one disadvantage is that the molecular ion is often not detectable. The main advantage of CI is the great abundance of the protonated ions of the intact molecule. C I provides high sensitivity and specificity for assaying known compounds. There are also additional advantages such as the possibility of using different reagent gases to induce various chemical reactions in the ion source and the possibility of eliminating interfacing in GUMS b y using the same gas for GC carrier and reagent gas. Both E I and C I require the samples to be vaporized, thus derivatization is frequently needed. FD spectra provide little formation on structure. The M+' or (M 1)+ ions are abundant, derivatization is not required, samples may be applied in aqueous solution, polar and thermally unstable materials may be studied, and the molecular weight limit may be extended to several thousands. The disadvantages are the experimental difficulties with biological materials; current methodology is not adequate for quantification. After technological developments this technique is expected to assume major importance (Schulten, 1977). For the survey analysis of elements in multicomponent systems the obvious choice is spark source mass spectrometry.

+

3. Selected Zon Monitoring When the mass spectrum is scanned in conventional MS, most of the scan time is spent recording the intensities of masses containing little or no useful information. For example, if the mass range 100-600 is

2 14

JOHN R O B 0 2

scanned in 5 seconds, only 0.01 second is spent on each mass. In selected ion monitoring (mass fragmentography, multiple ion detection), intensities or one, two, or more preselected masses are monitored simultaneously or in rapid repetitive succession. For example, the (M 1)’ ions in the chemical ionization mass spectra of 5-fluorouracil and an internal standard (6-mercaptopurine) are monitored to assay the former in blood (Section 11,C). The use of this technique is obvious when “gentle” ionization techniques, such as chemical ionization, is used. I n this technique the MS is used as a specific detector. It may be made even more specific if not the nominal but the exact mass of the compound of interest is monitored (Section 11,D). The technique is eminently suited for analysis of partially purified biological samples. In addition to specificity, the technique is about 103 times more sensitive than conventional G U M S or GC with flame ionization detection. A mass fragmentogram is similar in appearance to a gas chromatogram, except that there are several traces, each corresponding to the monitoring of a preselected mass. Identity is confirmed by checking GC retention times and the known relative intensities of the preselected peaks. For quantification, peak areas of the compound and internal standard are integrated. Stable-isotope enriched internal standard may be used to eliminate inaccuracies due to losses in sample preparation or on the GC column. Selected ion monitoring is perhaps the single most useful analytical technique currently available for trace quantification. It is sensitive (pmole quantities), specific, and may be employed to virtually all materials. Many applications are reviewed in later sections.

+

4. Stable Isotope Dilution In stable isotope dilution the concentration of a component in a matrix is determined horn the change produced in its natural isotopic composition by the addition of a known quantity of the same component the isotopic composition of which has been changed b y incorporation of stable isotopes. A potentially important use of stable isotopes is the administration of labeled, but not radioactive, drugs to humans followed by a search for metabolites in wiwo. When the degree of labeling with 15N, 13C, or 2H is so high that the Mf’ and (M 1)+ions are of comparable abundance, both the molecular and several fragment ions of the metabolites will appear as “twin ions” with characteristic intensities. A search for such ion pairs will reveal metabolites. When selected ion monitoring is used in searching for doublets, the isotopic method actually augments specificity.

+


2 15

It. Identification, Quantification, and Metabolism of Carcinogens

A. SURVEYANALYSES 1. General Of some 6000 compounds tested for carcinogenic activity, only 1000 showed positive response (Ember, 1975). The 9 published volumes of the continuing series “Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Man” (IARC, 1972-1975) are an indispensable source of information on the carcinogenicity of chemicals. Of the 222 chemicals evaluated in detail, 111 are unquestionably carcinogenic to experimental animals and 19 to man. Occupational exposure is known for 95 compounds, medicinical exposure for 29, and general environmental exposure for 52, including 15 polycyclic aromatic hydrocarbons. While it is clear that these carcinogens are hazardous when one is exposed to them as an “occupational hazard,” there is virtually nothing known about the hazards of long-term exposure at trace or ultratrace levels as “environmental hazard.” In “survey analysis” the main objective i s to detect, identify, and perhaps quantify as many constituents as possible so that their potential health hazards may be evaluated. Often no particular class of compounds is surveyed. Instead, such physical properties as volatility or polarity determine what kind of compound will be analyzed. G U M S has become the method of choice for survey analysis since dozens or possibly hundreds of compounds may be detected and identified in a single collected sample. In what follows, a few representative techniques are reviewed for surveying water and air. In the rest of this section, advances are discussed on those carcinogens known to be hazardous to man, and examples are shown for the study of the metabolism and mode of action of some carcinogens. In two outstanding general reviews on the environmental applications of MS, Alford (1975b, 1977) lists 668 references. In some publications reviewed, hundreds of pollutants were identified, in others only one or a few. In another fine review, Oswald et al. (1974) d‘iscusses factors affecting methodology as well as several compound classes. These reviews provide an overview of the all-important sampling techniques, analytical methodology developments, and applications in virtually all areas of environmental research, but contain relatively few references specifically dealing with carcinogens.

2. Water In a review of the analysis of organic compounds in water to support health effects studies, Garrison ( 1976) quotes some interesting statis-

2 16

JOHN ROBOZ

tics. To date, 1296 compounds have been identified in water and more than 90% were identified by combined GC-MS. About one-half of these were found only once, the majority less than five times. All these compounds are volatile or were made volatile by derivatization. It is estimated that 95% of all organic contaminants in water are nonvolatile. Progress in combined LC/MS will soon permit the systematic identification of nonvolatiles so that their possible health hazards may be evaluated. A comprehensive review of techniques for the analysis of organic pollutants in water together with applications to waters of various origin was published in a book (Keith, 1976a). A detailed description of sampling, separating, and identifying hundreds of organic substances in potable water (Grob, 1973) and the application of the technique to waters in the vicinity of Zurich, Switzerland (Grob and Grob, 1974) is typical of the many publications available in this field. Average sensitivity is 1 part in lOI3 (w/w). Once pollutants are identified, the next step is to trace the orgin of the pollution or, when that is known, trace the efficiency of various cleanup treatments. An instructive example is the identification and quantification of organic compounds in unbleached, treated kraft paper mill wastewaters after various treatments (Keith, 197613). Most analytical techniques employ charcoal, macroreticular resins, polyurethane foams, or solvent extraction for the collection and concentration of organics (Webb, 1975). Direct injection of aqueous samples into GClMS has obvious limitations; however, it is a fast and convenient tool, particularly for highly volatile constituents which might be missed in conventional techniques (Harris et al., 1974).

3. Air The general survey analysis of pollutants of air has received much attention in recent years and dozens of publications appeared suggesting techniques to improve the efficiency of sample collection, to increase resolution of constituents, and to provide specificity. Grob and Grob (1971) used cigarette-filter charcoal to trap volatile organics in the air of Zurich, Switzerland. After extraction with carbon disulfide, components were separated by capillary column GC and identified by MS. Over 100 components were identified, most of them aliphatic and aromatic hydrocarbons. Ciccioli et al. (1976) compared Tenax GC and Carbopack B in terms of sample recovery in “personal” samplers and found their performance comparable in many respects. Versino et al. (1976) collected organic micropollutants on a porous polymer glass absorption column and thermally eluted them with the aid of helium


217

into a GUMS. In still another approach to sampling, Pellizzari et al. (1976) collected pollutants on a bed of Tenax GC, thermally eluted the vapors into a capillary trap cooled to -196"C, metered the samples into a capillary column for separation followed by MS identification. Typically, 80 to 150 compounds were found in air samples and several carcinogenic compounds were detected. A problem in cryogenic preconcentration of air samples is the presence of water. Swofford et al. (1976) exploited the presence of water and used it as the reagent gas in the CI mode. Lao et al. (1976a) and Thomas and Lao (1976, 1977) developed MS methodology for the detection and quantification of at least 16 known carcinogens which may be present in both urban and industrial atmospheres. These are: 4,4'-dinitrobiphenyl, 6-9-dimethyl- and 5,7-dimethyl benzacridine, 3,4-benzacridine, 2-aminochrysene, 6-aminocry6,12-diazaanthrene, 2-azafluoranthrene, sene, 1,2,7,8-dibenzcarbazole, ISazafluoranthrene, 1-azapyrene, 2-aminofluorene, 2-aminoanthracene, 4-aminofluoranthrene, and Saminopyrene.

B. POLYCYCLIC AROMATIC HYDROCARBONS

1. Isolation, Fractionation, Analysis Since polycyclic aromatic hydrocarbons (PAH) are usually present in environmental samples only as traces, and also accompanied by a variety of other extractable materials, isolation and fractionation of varying sophistication are needed. Furthermore, since individual polycyclic hydrocarbons vary considerably in their toxicities and even more in their carcinogenic properties, identification and quantification of specific compounds must be made in the presence of several compounds of very similar composition. For example, chrysene and the 1-,2-,4-, and 6-methylchrysenes exhibit moderate carcinogenicity, whereas the 5-methyl (also 3-methyl) chrysenes are strong tumor initiators. Caustland et al. (1976) synthetized many derivatives of benzo[alanthracene, dibenz[a,h]anthacene, 3-methylcholanthrene, and 7,12dimethylbenz[a]anthracene. The E I mass spectra of these compounds exhibit features characteristic of aromatic compounds, such as relatively intense molecular ions, abundant doubly charged ions, and other similarities. The phenols show M - 1,M - 28, and M - 31 peaks in addition to the intense and characteristic M - 29fragment. The epoxides exhibit small but characteristic M - 16 ions in addition to those ions common in phenols. The ortho-quinones of K-region have their

2 18

JOHN ROBOZ

most intense peak (base peak) at M - CO with less intense molecular ion, while with the non-K-region quinones just the opposite is true. PAH can also be analyzed by CI (Janini et al., 1976b). As expected CI spectra are simple; however, because the molecular ions of PAH are very abundant even in the EI mode, CI does not offer significant increase in sensitivity. Still another approach is to utilize EI at low ionization voltage (Blumer, 1975). Here the E I ion source is operated at 12 eV, at which point there is enough energy to form the molecular ions only and there is virtually no fragmentation. This is an old technique in the petroleum industry; however, sensitivity is rather poor. Inspection of the mass spectra of PAH reveals that MS alone does not readily distinguish between isomeric PAH because very few fragment ions form and the molecular ions tends to dominate the spectra regardless of the method of ionization employed. Thus, to analyze mixtures of isomeric ring systems or to distinguish among alkyl isomers, components must first be separated b y GC or’HPLC. Lao e t al. (1973) developed a GC method, using packed Dexsil-300 columns, for the separation and subsequent MS identification of some 70 PAH (from 3 to 7 rings), present in nanogram quantities. In subsequent developments, Lao et aE. (1976b) utilized Dexsil-400 and Dexsil-410 packing materials and succeeded in separating benz [alanthracene, a carcinogen, from chrysene and triphenylene. These components usually appear unresolved in gas chromatograms. Because benz[a]anthracene is often used as an “indicator” of PAH in environmental samples, this separation is an important advance in methodology. The same columns also separate benzo[a]pyrene and benzo[e]pyrene. In another approach, Lee et al. (1976a,b) used prior LC separation, followed by capillary column G U M S (Novotny et al., 1974) to resolve toxicologically important alkyl derivatives. The initial separation step by HPLC is particularly important in the analysis of airborne particulates which have relatively low concentrations of alkylated PAH with respect to the parent compounds. This technique was utilized to identify some 150 PAH in smoke condensates of tobacco and marijuana, and some 100 PAH in airborne particulates. I n a novel approach, Janini et aE. (1975, 1976a,b) utilized nematic liquid crystals to separate PAH prior to MS identification and quantification. Nematic crystals are unique compounds which do not melt sharply into a disordered liquid upon heating. Instead, at well-defined temperatures they pass through one or more intermediate stages in which some structural order still exists. For example, N,”-bis(p-methoxybenzy1idene)-cup’-bi-p-toluidine (BMBT) has a nematic phase ( 181”-320“C range) where the molecules are oriented with their long


2 19

axes parallel. These liquid crystal will attract more, and retain longer, those structural isomers which are elongated than isomers which are more compact in shape. Accordingly, the order of elution is not in order of boiling points. With the development of two improved liquid crystals (Janini et aE., 1976b), N,N’-bis(p-phenylbenzylidene)a,a’-bip-toluidine (BPhBT) and N,”-bis(p-hexyloxybenzylidene)a,a’-bip-toulidine (BHxBT), several previously impossible separations were accomplished (Figs. 4,5, and 6 ) . Separation on liquid crystals is followed by analysis with CIiMS operated in the selected ion monitoring mode. For example, the base peak (highest intensity) in the mass spectra of all components shown in Fig. 4 is the same (mle 252); without the GC separation these compounds could not be distinguished by MS. Using this technique, as little as 4 ng ofbenzo[a]pyrene could be detected and quantified. 2. Air

A11 techniques reviewed in the previous section have been applied to analyze PAH in air at various locations. The references given contain numerous examples of survey analysis of PAH, including the detection of carcinogenic PAH. 3. Tobacco Smoke The widely publicized health effects of cigarette smoking have prompted numerous investigations concerning carcinogenic materials in tobacco smoke. These studies can be divided into three major areas:

I

5

7 9 Time (minutes)

11

FIG.4. Total ion chromatogram of 20-carbonpentacyclic arenes on liquid crystal phase (BPhBT) column at 275°C. (Reprinted from Janini et al., 197613, with permission.)

220

JOHN ROBOZ

FIG. 5. Total ion chromatogram of five 22-carbon arenes on liquid crystal phase (BPhBT) column at 275°C. Compounds in order of elution: 1,2,3,4-dibenzanthracene, 1,lZ-benzoperylene, 1,2,5,6-dibenzanthracene, 1,2,7,8-dibenzophenanthrene (picene), and 2,3,6,7-dibenzanthracene (pentacene). Reprinted from Janini et d., 1976b with permission.

In

I

1

0

b

8

12

16

20

24

28

Time (minutes)

FIG.6. Chromatogram of five 24 = carbon arenes on BPhBT liquid crystal column at 290°C. Compounds in order of elution: 4,5,6,7-dibenzopyrene, 4,5,7,8dibenzopyrene, coronene, 2,3,6,7-dibenzopyrene, and 1,2,6,7-dibenzopyrene. Reprinted from Janini et d., 197613 with permission.


22 1

(a) problems in sampling and fractionation, (b) identifying tumorpromoting subfractions, and (c) identifying constituents. Collection techniques differ according to the objectives: In cigarette smoking PAH are inhaled in undiluted form, whereas in indoor pollution PAH are inhaled in highly diluted form. Among the many subfractions isolated, the neutral and weakly acidic fractions appear to contain most carcinogens (Severson et at., 1976). Tumor promoting subfractions are identified by bioassay. When individual identifications are needed, combined GC-MS is used. Hechtet al. (1975)obtained various subfractions of the weakly acidic fraction of cigarette smoke particulate matter, tested them for carcinogenicity, and identified by GUMS a variety of compounds in the active subfractions ranging from fatty acids to PAH. Holzer et al. (1976) identified 133 components in cigarette smoke, including many PAH. Hecht et al. (1978b)studied the reaction of nicotine and sodium nitrate: several nitrosamines and fragmentation products of the pyrrolidine ring were identified by MS. The methodology for analyzing PAH in smoke is now available, and more than 500 PAH, ranging from indene to dimethylcoronene, have been identified (Severson et al., 1976). These techniques may now be applied to the various types of tobaccos and cigarette preparations. For example, air-cured tobacco contains about one-half as much PAH than flue-cured tobacco, and the former is significantly less carcinogenic (Brunnemann and Hoffmann, 1976).

4. Carbon Black Recent suggestions connecting excess stomach cancer among workers in the tire industry focused attention to identify and quantify carcinogenic PAH in carbon blacks. Gold (1975) separated CH,Cl,extracted oil furnace black into neutral and hydrocarbon fractions. The latter was further fractionated, yielding a concentrated PAH fraction in which at least a dozen compounds, including naphthalene, acenaphthalene, pyrene, and benzo[g,h,i]fluoranthene, were identified. One peak, present in considerable quantity, could not be identified by low resolution MS or UV analysis. Precise mass measurement yielded the formula C18Hlowhich agreed with that of a previously isolated but not yet characterized carcinogen. Additional work resulted in the identification of the carcinogen as 3,4-dihydrocyclopenta[c,d]pyrene. Lee and Hites (1976) and Hase et al. (1976) utilized capillary column GUMS to identify some 28 compounds in carbon black. Seven were sulfur-containing polycyclics, including dibenzothiophene, benzo[a]dibenzothiophene, benzo[d,efldibenzothiophene, and ben-

222

JOHN ROBOZ

zo[d,efJnaphtobenzothiophene.There is evidence that substitution of a sulfur for an ethylene group in a ring may increase or decrease the carcinogenic potency of the particular compound. Thus, the methodology developed should be applicable to future work concerning the carcinogenic components of carbon black, and also air particulate matter in which sulfur-containing carcinogens have also been detected (Lee et al., 197613). Soot. Some 40 polycyclic aromatic hydrocarbons (concs. > l%), including benzo[a]pyrene (0.2%), pyrene, fluoroanthene, benz[a]anthracene, and phenanthrene, were identified in soot of an A 1 electrolysis furnace (Tausch and Stehlik, 1977).

5. Metabolism of Benzo [alpyrene Polycyclic aromatic hydrocarbons are metabolized into organic solvent-soluble phenols, trans-dihydrodiols, quinones, and watersoluble glutathione conjugates. Both the detoxification and metabolic activation to toxic derivatives or to active carcinogens are catalyzed by aryl hydrocarbon hydroxylase which is a microsomal enzyme system containing several P450 cytochrome enzymes. Because of its ubiquitous appearance, benzo[a]pyrene (BP), has been used as the prototype for the PAH. Gelboin and co-workers (Selkirk et al., 1974a,b, 1975; Yang and Gelboin, 1976) studied the metabolism of BP and utilized MS, TLC, and UV to identify metabolites. Benzo[a]pyrene was incubated with rat liver microsomes prepared from rats treated with methylcholanthrene. After stopping the reaction with acetone, the mixture was extracted with ethyl acetate, the solvent-soluble metabolites concentrated by evaporation to dryness, dissolved in a few drops of methanol, and metabolites were separated 700

65

50

75

100

12s

FRACTION

FIG. 7 . High pressure liquid chromatographic profile of the metabolites of benzo[ulpyrene (BP) after incubation with liver microsomes prepared from Bmethylcholanthrene treated rats. Fractions of 0.2 ml were collected at 20-second intervals. Q = quinone, OH = hydroxy. Metabolites were identified by mass spectrometry. Reprinted from Selkirk et ul., 1974b with permission.

223


using HPLC (Fig. 7). Next, the HPLC peaks were collected and analyzed individually in a MS using the solid probe introduction system. Among the eight metabolites detected and identified, there were three dihydrodiols(9,lO-, 7,8-, and 4,5-dihydrodihydroxybenzo[a]pyrene), three quinones (benzo[u]pyrene-1,6-dione,and -6,12-dione), and two phenols ( 8 ,and 3-hydroxybenzo[a]pyrene).The three dihydrodihydroxy metabolites have the same molecular weight so that MS determination of the molecular weight alone, even by high-resolution MS, would not be adequate for characterization. Although several fragmentation peaks are the same in all three compounds, e.g., loss of water, loss of water plus a CHO fragment, and the doubly charged ion, there are different fragmentations that permit differC2H2in the 9, 10-diol, loss of HOH CO entiation: loss of HOH in the 7,8-diol, and the absence of both these ions in the 4,Sdiol metabolite. This is a good illustration of how fragmentation by E I may be utilized for identification and characterization. The areas of the HPLC peaks were utilized to quantify the various metabolites, and also to study how enzyme inhibitors affect the formation of each metabolite. For example, the inhibitors 7,8-benzoflavone and 1,2-epoxy-3,3,3-trichloropropane selectively affect the hydrase rather than the oxidase activity of the enzyme system resulting in both qualitative and quantitative effects on several metabolites (Selkirk et al., 1974a,b). In a subsequent study (Selkirk et al., 1975), the pH of the eluting solvent in the HPLC analysis was changed resulting in the detection of still another metabolite in the presence of the epoxide hydrase The unknown was ideninhibitor, 1,2-epoxy-3,3,3-trichloropropane. tified by MS as benzo[a]pyrene-4,5-epoxide.This identification provides considerable support to the suggested role of epoxides in the intermediary metabolism of BP. In still further studies by the same group (Yang et al., 1976; Yang and Gelboin, 1976) on BP diol-epoxides by nonenzymatic reduction, trihydroxypentahydrobenzo[alpyrenes were detected and characterized by obtaining the mass spectra of the isolated HPLC fractions. The activity of epoxide hydratase toward epoxides of different PAH, including benzo[a]pyrene-4,5-oxide,benzo[a]anthracene-5,6-oxide, and 3-methylcholanthrene-11, E-oxide, can be determined with specificity and sensitivity (1 ngiml incubation mixture) by GCiMS quantification (in selected ion monitoring mode) of the TMS derivatives of the corresponding trans-diols formed during incubation (Bettencourt et al., 1977).

+

+

224

JOHN R O B 0 2

C. POLYCHLORINATED BIPHENYLS There are 210 possible polychlorinated biphenyls (PCBs) and many are present in commercial preparations. Although PCBs are not known to be carcinogenic to humans, they induce pathological changes upon long-range exposure, are highly toxic, even lethal, when high concentrations are inhaled, and several of them cause cancer in experimental animals. Since PCBs usually occur together with common chlorinated insecticides and related compounds which have very similar chromatographic retention times, GUMS is essential for separation and positive identification in environmental samples. The chromatographic (and also biological) aspects of PCB have been reviewed by Fischbein

(1972). The EI fragmentation of PCB was discussed by Safe and Hutzinger (1972),while the use of isotopic abundance ratios in the identification of PCB compounds in the presence of related substances was developed by Rote and Morris (1973).These two papers, together with a text on the MS of pesticides and related pollutants (Safe and Hutzinger, 1973), provide basic information needed for the identification of individual PCBs even in complex mixtures. To increase sensitivity, so-called ion abundance chromatograms may be used (Eichelberger et al., 1974). Starting with an idealized set of mass spectra for PCB, these authors evaluated several sets of candidate masses for monitoring PCB in the presence of those pesticides likely to be in the solvent extract of environmental samples. Quantification is possible using area measurements on the ion abundance chromatograms obtained in subset mass scanning. An alternative approach is to employ capillary column chromatography for high-resolution separation followed by MS identification, either on the basis of complete mass spectra for characterization or selected ion monitoring quantification at highest sensitivity (Schulte and Acker, 1974). In recent years there has been an abundance of publications on the metabolism of PCB in humans and experimental animals. Jensen et al. (1974) studied PCB in human adipose tissues and concluded that two adjacent unsubstituted carbon atoms are required for rapid metabolism. Curley et al. (1975) identified tetrachlorodibenzofuran in the urine of rats following dietary exposure to Aroclor 1254. Burse et al. (1976) identified (or at least characterized) six hydroxylated biphenyl metabolites in rats following prolonged dietary exposure to Archlor 1242 and 1016. In both studies, the metabolites were ex-


225

tracted from urine with organic solvents, purified by some form of chromatography, methylated with diazomethane, and analyzed by GC/M S . In a different approach, Safe et al. (1975) utilized high-resolution MS with ion-sensitive photoplate detection to detect and identify several oxygenated metabolites of PCB in crude goat urine extract (also in other biological materials, see references in Safe et al., 1975). This technique utilizes the fact that chlorine-containing xenobiotics exhibit a large negative mass deficiency, while natural compounds containing C,N,O, and H have positive mass deficiency. When molecular formulas are determined with a precision of a few millimass units, oxygenated pentachloro-, tetrachloro-, and monochloro-biphenyl metabolites can easily be identified in crude mixtures,

D. BiS-CHLOROMETHYL ETHER

A high-resolution MS method for the detection of bis-chloromethyl ether (BCME) in air with a detection limit of 0.1 ppb was developed by Collier (1972). A 15,000-fold increase in concentration was provided when 15 liters of air was forced through Poropak Q (bead form) adsorber at a rate of 1.5 liter/min. Next, the adsorber was attached to the inlet system of the MS, the system pumped to lo+ torr, and the adsorbed organic components eluted into a reservoir b y heating (18OOC for 5 minutes). The collected sample was leaked into the ion source and the narrow mass range of 78.9-79.1 scanned (EI)at a resolving power of 1/3500. After mass calibration, the most abundant ion, at mle = 78.9950 (C1CH2-O-CH2)was quantified. The high resolution used provided separation from most of the ions that might be present from other contaminants, e.g., CH,SiCl+, CH3S2+,C2HSN+,and C,H,+. Two possible contaminants were not resolved: CH,O,P+ (mfe= 78.9949) would require a resolving power of 1/79,000; however, this ion does not contain chlorine so evaluation of BCME may be made on the basis of the chlorine-containing isotope at mle = 80.9921. The other possible unresolved interference is the C,HClF+ ion (mle = 78.9751), which may be problematic in the absence of BCME. When this is suspected, a small amount of BCME is added to see if the two peaks are identical or not. An alternative technique (Shadoff et al., 1973) utilized lowresolution GC/MS, which is fast and convenient to separate all compounds present so there was no need for high resolution to surmount interferences. Simultaneous monitoring of the mle = 79 peak (base

226

JOHN ROBOZ

peak) and also the mle = 81 peak (chlorine isotope peak) provided selectivity combined with high sensitivity (fractional ppb level, accuracy within 10%). The advantages of both of these techniques were combined by Evans et al. (1975) who first passed the desorbed gas mixture through a GC and then monitored the C,H,OCl+ peak at high resolution (resolving power 3800) so that single ion monitoring at high resolution was achieved and the highest degree of specificity maintained (Fig. 8). With this method BCME is detectable to 0.01 ppb (10-liter air sample). The concern over the carcinogenicity of BCME increased considerably when Frankel et al. (1974) reported that it could be formed spontaneously in air or solution from hydrochloric acid and formaldehyde. The gas-phase reactions were carried out either in glass flasks or Saran bags under controlled conditions and BCME was both identified and quantified by high-resolution MS or GUMS. Reactant concentrations strongly affected BCME yield, while temperature or humidity were less important variables. Although the amount of BCME formed was only about 0.01 mol%, the strong carcinogenicity of BCME requires alertness about this hazardous chemical. Since BCME is an impurity in chlormethyl methyl ether (CMME), a commonly employed chlormethylating agent, a technique was developed (Tou and Kallos, 1974) to study the kinetics of the decomposition of both compounds in humid air. Both the nature of the reactor surface (e.g., glass, Teflon, etc.) and experimental conditions (e.g., humidity, temperature) were varied and preselected fragment ions were monitored for each compound. The upper limit of the rate of hydrolysis was only 0.00047 per minute for BCME at 2.6% water concentration; thus routine monitoring presents no problem.

FIG.8. Appearance of the mle = 79 region of the mass spectrum at resolving power 3800 in the analysis of bis-chloromethyl ether. The ions at mle = 78.9950 correspond to (ClC2H40)+.Reprinted from Evans et al., 1975, with permission.


227

Two homologs of BCME, bis-chloroethyl ether and bis-chloroisopropyl ether, are suspected carcinogens and have been found in the environment. These compounds may be analyzed in water, after extraction with petroleum ether and concentration by evaporation, b y utilizing mass fragments formed by the loss of the CH&1 group (M+ - 49). Selection of these ions assures specificity and a detection limit of 10 ppb (Bonelli et al., 1975).

E. VINYLCHLORIDE Recent concern about the carcinogenic effects of vinyl chloride prompted analytical activities in three areas: (a) detection in various environments, (b) routine monitoring of the level of vinyl chloride, and (c) understanding action by exploring the metabolism. For the positive identification of vinyl chloride in various environments, MS is the ideal tool because of both high specificity and sensitivity. For example, Pelizzari et al. (1976) obtained profiles of organic vapor pollutants in the atmosphere in the Houston and Los Angeles areas. Some 120 peaks were detected in the total ion current chromatogram of the thermally desorbed (from Tenax cartridges) vapors after separation in a capillary column in GUMS. Vinyl chloride was present in the atmosphere in the Houston area but not in the area sampled near Los Angeles. Incidentally, a total of 21 halogenated hydrocarbons were detected, including chloroform, carbon tetrachloride, and the strongly carcinogenic trichloroethylene. For routine monitoring, as required by the Occupational Safety and Health Administration, the problems of sample collection for analysis by GC (flame ionization) were investigated (Ahlstrom et al., 1975; Levine et al., 1975). Vinyl chloride can be analyzed with adequate sensitivity and reproducibility by monitoring the mle = 62 peak (molecular ion) by GUMS (Bonelli et al., 1975; Taylor and Springer,

1975). There are three GC/MS techniques for the analysis of vinyl chloride in water (Alford, 1975a). When 2 pl of water is injected directly, detection limit is 0.05 mg/liter. For lower levels, extraction with CCl, is needed and the limit of detection is 2 pg/liter (0.14 ng injected). The vinyl chloride content of wastewater from a plastics manufacturing plant was in the 0.2-9.3 mg/liter range. Using simultaneous recording of the mle = 62 and mle = 64 peaks, and direct injection of 1ml water, Fujii (1977) was able to detect 0.1 ppb of vinyl chloride. Because proposed limits ofvinyl chloride in both food and drinks are at the ppb level, VanLierop and Stek (1976) used selected ion monitoring and

228

JOHN R O B 0 2

integrated the areas for quantification. As little as 0.3 ppb of vinyl chloride was determined in salad oil. Preliminary work on the metabolism of vinyl chloride suggested that, similar to aromatic hydrocarbons, epoxidation may be responsible for the formation of the carcinogenic metabolite. The suspected epoxide, chloroethyleneoxide, in the pure state slowly rearranges to chloroacetaldehyde (ClCH,CHO). Gothe et al. (1974) used prepared rat liver homogenate in a conventional manner but, in addition to adding the necessary cofactors for the mixed-function exygenase system, 3,4-dichlorobenzenethiol(a nucleophile) was also added to trap metabolites. A mixture of air and vinyl chloride was exposed to the fortified liver homogenate for 2 hours, the reaction mixture extracted with hexane, the extract divided into bicarbonate soluble (acid) and neutral fractions, and both fractions were further fractionated by TLC. The zones of the postulated metabolites (determined with authenic compounds) were collected and analyzed by GUMS. The expected 3,4dichlorophenylthioacetaldehydewas found in the neutral fraction. The bicarbonate-soluble fraction yielded, after methylation, a peak corresponding to 3,Cdichlorophenylthioaceticacid methyl ester. In the second set of experiments, the trapping agent was not added to the liver homogenate but instead the vapors leaving the area of incubation were passed over the reagent in a second tube. The same metabolic products were found as in the first type of experiment. Watanabe et al. (1976a,b) detected three metabolites in the urine of rats using 14C-labeled vinyl chloride. One metabolite, corresponding to one-third of the total urinary radioactivity, was identified b y MS as N-acetyl-S-(2-hydroxyethy1)cysteine. The second metabolite (onefourth of radioactivity) was shown by MS as thiodiglycolic acid; the third metabolite could not be identified. It was concluded that the major metabolic pathway of vinyl chloride in the rat is conjugation with glutathione. Mueller et al. (1976) used GUMS to identify thiodiacetic acid and S-(carboxymethy1)cysteine in the urine of rats after a 48-hour exposure to 1000 ppm vinyl chloride.

F. NITROSAMINES Pensabene et al. (1972) synthetized 25 nitrosamines and provided basic data on their IR, GC, and MS properties. Chemical ionization mass spectra of nitrosamines was obtained by Gaffield et al. (1976). Gough and Webb (1973) developed a GUMS technique for the analysis of trace nitrosamines (1mg/kg) in foodstuffs. This technique employs a carrier gas pressure programming and peak-cutting device


229

which utilizes a silicone membrane separator between GC and MS to remove excess carrier gas and enrich nitrosamine concentration, permits analysis of dimethyl- and diethylnitrosamines which would elute on the tail of the solvent in conventional analysis. To eliminate possible interferences, high resolution (e.g., 12,000) was employed. Gough et aZ. (1977) applied the technique for the quantification of N-nitrosodimethylamine and N-nitrosopyrrolidine in a variety of food products and their vapors, urine, and vegetation. Hecht et al. (1978a) studied the metabolic a-hydroxylation of N-nitrosopyrrolidine and identified a-hydroxynitrosopyrrolidine and 3-formyl- l-propanediazonhydroxide as unstable intermediates which decompose to Zhydroxytetrahydrofuran. For routine analysis at high sensitivity Brooks et al. (1972) converted several volatile N-nitrosamines into electron capturing derivatives with heptafluorobutyric anhydride (HFBA) using pyridine as catalyst. The structure of these derivatives was determined by obtaining both nominal and exact mass measurements on the molecular ions and important fragmentation products such as (M - F - H,O)+, (M - F C,H,)+, and ( M - C,F, - H,O)+ (Gough et al., 1975). The molecular weights of the HFBA-derivatives of N-nitrosodiethyl-, dipropyl-, and dibutylamines, N-nitrosopiperidine, and N-nitrosopyrrolidine correspond to the expected molecular ions; also the difference in molecular weight of any two of the derivatives is the same as that ofthe corresponding nitrosamines, indicating that all of them reacted with HFBA in a similar mode. In the case of N-nitrosodimethylamine the M+' ion was observed at mle = 522 which could only be explained by assuming two molecules of NDMA reacting with HFBA to form a stable structure containing a saturated ring with two nitrogen atoms. Sen (1972) and Sen et al. (1972) analyzed 59 samples of various meat products for the presence of dimethynitrosamine (DMN). DMN was present in five samples of uncooked salami and dry sausage in quantities in the 0.01-0.08 ppm range and in three samples at the 0.005 ppm (detection limit) level. The presence of DMN was confirmed using the mle = 74 (M+') and the mle = 30 (NO+)peaks in G U M S at low resolution and also b y high-precision mass measurement of the molecular ion (M+ = 74.0480 k 0.0004) of underivatized DMN. Sen et al. (1973) also investigated if nitrosopyrrolidine (NPy) might form during the cooking of nitrate-treated cured bacon. NPy probably originates from naturally occurring nitrosoproline (noncarcinogenic) which loses the carboxyl group upon heating. The level of NPy found in commercial bacon products upon frying was in the 0.004-0.02 ppm level; no NPy was detected in uncooked bacon. Considering that 50

230

JOHN ROBOZ

ppm of a nitrosamine in the diet of rats induces a large incidence of tumors (Lijinsky, 1976), 100 ppb of NPy in bacon may be hazardous. Since 3-hydroxyproline and 3-hydroxypyrrolidine are also known to be present in pork, Sen et al. (1976) developed methodology to identify S-hydroxy-l-nitrosopyrrolidine (HNPy) which might also be present in fi-ied bacon. Since HNPy is nonvolatile, the 3-methoxy derivative was formed by methylating with sodium hydride and methyl iodide in ethyl acetate at room temperature. To eliminate interferences and maintain specificity, two ions were monitored at a MS resolution of 5000: the M+' ion (mle = 130.0742) and NO+ (m/e = 29.9980). No data have yet been reported on HNPy in bacon. The question of the presence, and importance, of N-nitrosodimethylaniine (DMN) in air is a subject of current debate. Fine and co-workers (1975) used GC and HPLC for separation, and MS for identification to confirm the presence of DMN in the air over Baltimore, Maryland and in Belle, West Virginia, at various locations. In Baltimore, D M N levels were found in the range of 1000-36,000 ng/m3 in the vicinity of a chemical factory where DMN is used as an intermediate. In other areas, less DMN was found; however, levels appeared higher in areas where chemical factories were located. The authors concluded with a thought-provoking comparison of data on DMN uptake: 4 slices of fried bacon = 500 ng, 20 cigarettes = 1000 ng, breathing air containing 1000 ng/m3 for 1day = 14,000 ng. These numbers are to be considered in light of the facts that 50,000 ng/kg/day of DMN fed twice weekly is clearly carcinogenic in the mink, and 200,000 ng/m3 DMN in air is carcinogenic to mice and rats. A cause for further concern are recent reports that the N-nitrosodimethylamine content of some commercial weed killer products is as high as 640,000 mg/liter (Rawls, 1976), and that N-nitrosodiethanolamine was separated by HPLC and identified by MS in toiletry products (Editorial, 1977b). Preussmann et a l . ( 1976) utilized 3,4dichlorothiophenol (Section I1,E) to trap indirectly alkylating electrophiles formed in the presence of rat liver microsomes. Ten nitrosamines, l,%dimethylhydrazine, and 3,3dimethylphenyltriazene (all known carcinogens) were tested and all gave the appropriate 3,4-dichlorophenyl alkyl thioethers as confirmed by MS. The noncarcinogenic methyl-t-butylnitrosamine did not alkylate. Positive relationship was established between carcinogenicity and alkylating capacity of nitrosamines. This technique may be of value for rapid screening of suspected indirect alkylating agents. In studying the conditions for the formation of nitrosamines in the


23 1

digestive system, Wishnok and Tannenbaum ( 1977) reacted morpholine with saliva. I n addition to the expected N-nitrosomorpholine, a morpholinocyanamide was identified by high-resolution mass spectrometry, suggesting a hitherto unknown metabolic pathway for secondary amines.

G. MISCELLANEOUS

1. Phorbol Myristate Acetate Mouse skin tumorigenesis has been conceptualized to proceed in two steps: first, normal cells are initiated to a potential cancer state in an irreversible process by a chemical or physical agent; next, a promoting agent, when applied repeatedly, produces a reversible process leading to the cancerous state. The active principles of croton oil, a known tumor promoter in mouse skin, are the diesters of the parent diterpene alcohol, phorbol. Phorbol is a nonpromoter; phorbol myristate acetate ( 12-0-tetredecanoyl-phorol- 13-acetate) is the most biologically potent ester isolated from croton oil, Crombie (1968)investigated the structure and EI mass spectra of the parent diterpene alcohol and some derivatives. Segal et al. (1975) identified phorborol myristate acetate as a metabolite in mouse skin. The metabolite was also synthetized by reducing the c-5 carbonyl group of phorbol myristate acetate to a secondary alcohol with NaBH,, and its structure confirmed by MS. The parent compound exhibited a weak M+. ion, and major peaks corresponding to the loss of water, acetic acid, and myristic acid. The metabolite had no M + ion but showed intense fragmentation peaks corresponding to the loss of water, water + acetic acid, myristic acid, and myristic acetic acids. The FD spectrum of intact phorbol shows intense ( M = l)+and M+’ ions, and little fragmentation (Schulten, 1973).

+

2. AfEatoxins Aflatoxins in food and feed are usually assayed by employing TLC or LC for separation, densitometry, UV, or fluorescence for quantification, and relative chromatographic retention time or a separate microbiological assay for confirmation. Haddon et al. (1977)developed a high-resolution selected ion monitoring technique for the positive confirmation of Aflatoxin B1 in complex mixtures. Surface bonding between aflatoxins and glass sample containers increases sensitivity to such an extent that 0.1 ng can be positively identified.

232

JOHN ROBOZ

3. Dieth ylstilbesterol This synthetic estrogen is no longer used to prevent miscarriage, but quantification is still needed in foodstuff. The EI spectrum of the dimethylated compound exhibits a strong M+’ ion which may be monitored so that 10 pg injected material can be detected b y G U M S (Abramson, 1972). The dihydroxy and monomethoxydiethyl metabolites were identified in rat liver (Engel et ul., 1976), and the monoglucuronide metabolite was identified in the urine of meat producing animals (Aschbacher, 1976) by mass spectrometric analysis of chromatographically separated products.

4. Stilbenes Microsomal N-hydroxylation is believed to be essential, though not necessarily sufficient, for the metabolic activation of carcinogenic aromatic amides Mass spectrometric identification of metabolites of truns-4’-halo-4-acetamidostilbenes was reported by Gammans et al. (1977); the analytical approach developed may be employed in other studies of this type.

H. TRACEELEMENTS Results of investigations on the role of trace metals and other elements in cancer are contradictory and conflicting (Schwartz, 1975; Editorial, 1977a). Selenium appears to have a protective effect which may be offset b y zinc, iron, and nickel which are suspected carcinogens, while manganese possibly reduces the effects of nickel, and copper might inactivate carcinogenic viruses. Traces of chromium, lead, and nickel found in fibrous silicates may be involved in asbestos carcinogenesis. The “elemental fingerprints” obtained by spark source MS (Section 1,C) may be used to study subtractive, additive, or synergistic interactions of metals in a comprehensive survey of all elements at ultratrace levels in a single analysis. Biological samples must first be ashed or digested in oxidizing acids to eliminate interferences from hydrocarbons formed in the spark. In wet ashing at low temperature with perchloric or nitric acid (in quartz dishes) losses are low and group separation and/or isotope dilution may be employed. Direct analysis of serum, urine, and homogenized tissues may be made after lyophilization (Fitchett et al., 1974). After pretreatment, the sample must be imbedded into a conductor so that a spark can be created: Ultrapure graphite is often used. Evans and Morrison (1968)and Morrison (1967, 1969) provided experimental details and data for some 30 metals in


233

serum, bone, lung, and kidney tumor. Sensitivity is 10 ppb for most elements in the periodic table, reproducibility is ~fr25%.Applications of surveying cancer-related biological materials include the analysis of iron, copper, and zinc in the serum of cancer patients (Brown and Jacobs, 1972) and the determination of some 72 elements in coal dust and in the ashed lung tissue of coal miners (Brown et al., 1972). A thought-provoking finding of the latter work was the positive correlation in trends found when the relative concentrations of a variety of elements, ranging fi-om major constituents such as calcium to trace contaminants such as scandium, in coal dust and lung tissue were plotted together. Survey analyses were also made of human hair (Yurachek et al., 1969) and human fingernails (Harrison and Clemena,

1972). Numerous publications appeared on the use of spark source MS to survey elements in air, domestic tap water, industrial waste waters, sediments, and fossil fuels. For references consult the reviews by Alford (1975b, 1977). Because of the complexity of the technique, spark source MS is the method of choice only in survey analysis. Once the significance of a particular element is established, other techniques (e.g., atomic absorption) should be employed for routine monitoring. Ill. Metabolism and Monitoring of Antineoplastic Agents

A. CYCLOPHOSPHAMIDE

1. Metabolism It is now well established that the effectiveness of cyclophosphamide (CP) arises via activation b y hepatic microsomal enzymes; the active metabolite(s) reach the target sites through the systemic circulation. Of the 32 papers presented at a symposium on the metabolism and mechanism of action of CP [Cancer Treatment Reports, 60, 299-525 (1976)] six were devoted exclusively to mass spectral studies and nearly 50% reported results obtained, at least in part, b y MS. In addition, at least 15 publications appeared during the last 3 years on the identification of major and minor metabolites of CP by MS. Evaluation of the cytotoxicity, both in uivo and in uitro, of the metabolites resulted in conflicting opinions about the reasons for the comparative selectivity and effectiveness of CP. The most important steps in the metabolism of CP (Fig. 9, Struck et al., 1975) are the ring-nitrogen oxidation to 4-hydroxycyclophosphamide or its tautomer aldophosphamide, followed by p-elimi-

234

JOHN ROBOZ

[ClCH,!=O] Chloroacetaldehydc

1 .I CH,CH,CI

\

Phasphocamidc murldrd

\

-t Enrymal8cally

demonstrated

FIG. 9. The metabolism of cyclophosphamide. Reprinted from Struck et al., 1975, with permission.

nation of acrolein to yield phosphoramide mustard. Phosphoramide mustard[N,N-bis( 2-chloroethyl)phosphorodiamidic acid] was isolated both from liver microsomes incubated with CP and from patients treated with the drug. It is a potent alkylating agent with strong antitumor activity and is now thought of as the chemotherapeutically active metabolite of CP. Whether the intermediate aldophosphamide breaks down in the liver and releases phosphoramide mustard into the circulation, or is metabolized intracellularly after reaching tumor cells is the subject of current studies. Carboxyphosphamide and 4-ketocyclophosphamide are major metabolites found in urine, but have no cytotoxic or alkylating activity and are probably products of detoxification. Nornitrogen mustard [N,N-bis(chloroethyl)amine] and 4-hydroxycyclophosphamide are cytotoxic, both in vitro and in vivo; their importance is not known.


235

Colvin et al. (1973) incubated mouse liver microsomes with C P (labeled with I4C universally in the side chain), separated the metabolic products on Sephadex LH-20 columns, isolated all radioactive fractions, derivatized them with diazomethane, and identified unreacted CP, phosphoramide mustard, and nornitrogen mustard by comparing their mass spectra to those of authentic synthetic compounds and also by confirming the molecular composition of both molecular ions and important fragment ions using exact mass measurements. The same group (Fenselau et d.,1975) identified circulating phosphoramide mustard in the blood and urine of patients receiving CP. Metabolites were separated in an Amberlite XAD-2 resin column, the methylated derivatives separated on silica plates, eluted with methanol, and analyzed by GUMS. Methylation of phosphoramide mustard led to a mixture of the mono-, di-, and trirnethylated parent compound, the monomethyl derivative comprising about 80% of the total. The EI mass spectra exhibited ions formed by cleavage in the C-C bond once removed from nitrogen and contained one chlorine atom so that ion pairs of predictable relative intensities ( 3 : 1) were obtained. The CI spectra (isobutane) yielded the (M + 1)+ions as the most abundant species with two chlorines resulting in a cluster of three peaks with intensity ratios of 9 : 6 : 1. Specificity was increased by monitoring all three masses (Fig. 10). Connors et al. (1974b) reported the identification of several metabolites of C P both in vivo and in vitro and offered detailed discussion of their possible significance. They have also investigated the metabolic activation of isophosphamide which appears to follow a pathway similar to that of CP. The important metabolite 4hydroxyphosphamide, which may act as “transport” from parent drug to either toxic or nontoxic final metabolites, appeared too unstable for direct identification. It was “trapped” with ethanol and identified by MS (Connors et al., 1974a). In the same study, 5 more metabolites, phosphoramide mustard, 4-keto-cyclophosphamide,carboxyphosphamide, 2-(2-chloroethylamino)tetrahydro-2H- 1,3,2oxyazophosphorine-2-oxide, and 3hydroxypropyl-N,N-bis(2-chloroethyl)phosphorodiamide,were also identified. This paper provides a great deal of information on in vitro methodology with rat liver microsomes as well as the implications of ) the results. This work was continued by Connors et aZ. ( 1 9 7 4 ~using cyclophosphamide-4d2to study the hydroxylation of CP by rat liver microsomes. I n still further studies (Cox et aE., 1976a), three alkylated analogs of C P (&methyl-, 4-methyl-, and 5,5-dimethyl-) were synthesized and incubated with various fractions of liver microsomes. Several products were isolated, identified, and their possible

236

JOHN ROBOZ Synthetic Siondord

d e 199 u)

w

m/e 201 U

Urine E x t r a c t

B

m/e 199

1600

li00

TEMPERATURE

FIG. 10. (A) Selected ion profiles of synthetic N,NN-bis(2-chloroethyl)phosphorodiamidic acid (phosphoramide mustard) treated with diazomethane. (B) Selected ion profiles of urine extract treated with diazomethane. Reprinted from Fenselau et d., 1975, with permission.

significance elucidated. When deuterium-labeled analogs of most known metabolites of CP were synthetized (Griggs and Jarman, 1975), it became possible to investigate the influence of deuterium substitution on the rates of metabolic pathways involving oxidation at the C-4 position as well as on the rate of elimination of acrolein from aldophosphamine. Isotope effects involving the 4-d, and 4,6-d4 analogs caused no change in antitumor activity compared to CP; however, 5,5-dideuteration caused a marked drop in potency. This appears important with respect to the proposed mechanism of


237

activation involving the elimination pathway leading to the formation of the cytotoxic phosphoramide mustard (Cox et al., 197613). In a search for aldophosphamide in blood, Struck (1974, 1976; Struck and Hill, 1972) has isolated an intermediate product in a model oxygenase system in the presence of semicarbazide (NH,NHCONH,) utilizing the technique developed by Sladek (1973) and characterized it in a semicarbazone form. Aldophosphamide was considered to be unstable, particularly after unsuccessful attempts to detect it in the blood of CP-treated mice (Struck et al., 1975). Using a different approach, Fenselau et al. (1977) succeeded in isolating and identifying aldophosphamide as a cyanohydrine derivative both in vitro as well as in vioo in humans. First aldophosphamide was synthesized and characterized by mass spectrometry utilizing the trimethylsilyl derivative. The cyanohydrine derivative was prepared by treating aldophosphamide with sodium bisulfite and sodium cyanide; the product was characterized by chemical ionization mass spectrometry. In the in vitro studies, CP was incubated with rat liver microsomes, the metabolite extracted, derivatized, and characterized by mass spectrometry. In the in vivo studies, the metabolite was isolated fi-oni the blood of patients receiving CP, the derivative formed, and aldophosphamide identified by mass spectrometry. Administration of synthetic 4-hydroxy-CP (Struck et al., 1975) followed b y extraction with chloroform permitted in vivo recovery of this metabolite. Administration of phosphoramide mustard led to the identification of nornitrogen mustard and a new metabolite, 3(2-chloroethyl)1,2-oxazolidine-2-one. Investigating the alkylating capability of phosphoramide mustard and nornitrogen mustard, in terms of reacting with nitrobenzylphyridine, it was established that at physiological pH, phosphoramide mustard retained significant alkylating activity whereas nornitrogen mustard did not. On the bases of these experimental findings, the authors postulated that phosphoramide mustard is the most important active metabolite of CP, and it is formed from intermediates outside the target cells. In subsequent study, Colvin et al. (1976) confirmed the differences in alkylating capability between phosphoramide mustard and nornitrogen mustard. Both compounds were incubated with ethanethiol (C,H,SH) and the products were identified as N,N-bis(2-(ethylthio)ethyl)phosphorodiamidic acid and 2,2’-bis(ethy1thio)diethylamine. It was concluded that phosphoramide mustard alkylates as an intact molecule. Further experiments with deuterated analogs of both metabolites confirmed that CP alkylation proceeds via an aziridinium intermediate.

238

JOHN ROBOZ

In most studies of the metabolism of CP, EI/MS was employed; EI spectra often lack molecular ions of appreciable intensity. Przybylski et al. (1976) used both E I and FD ionization to characterize the metabolites of CP. The F D spectrum of 4-hydroxycyclophosphamide exhibits a strong (M 1)+ ion, permitting full characterization, whereas the molecular ion in the electron impact spectrum is not detectable. The E I and F D spectra of phosphoramide mustard complemented each other in establishing the structure of the free acid. Attempts to characterize aldophosphamide b y FD/MS failed, supporting the thought that this compound is too unstable to play an active role in the cytotoxic action of CP. FD/MS analysis of “activated” sulfur-containing C P derivatives confirmed several postulates concerning these compounds (Draeger et al., 1976).

+

2. Quanti5cation

The observation that patients could “taste” C P after intravenous infusion prompted Duncan et al. (1973) to develop a method for the identification of the drug in saliva, synovial fluid, milk, sweat, and cerebrospinal fluid. C P was extracted with chloroform, the layers separated by centrifugation, the organic layer concentrated by evaporation, and introduced into the ion source via a direct probe. Here the crude extract was evaporated and the presence of C P confirmed either by obtaining the complete mass spectrum at low resolution, or by measuring exact masses of selected fragments at high resolution. The most abundant ion, in the E I spectrum of C P (mle = 211) corresponds to the loss of the CH&l radical from the molecular ion. This peak is accompanied by another two mass units higher in the proper abundance ratio (3: 1) for one chlorine atom. The molecular ion of C P is weak but detectable. When the concentration of CP was 1 mg/ml body fluid, high-precision mass measurements were made on the peaks at mass 211 and 213, respectively. This was selected ion monitoring at high resolution and resulted in the combination of high sensitivity and complete specificity even in the presence of endogenous lipids of the same nominal mass. C P was identified in all body fluids studied. The presence of C P in milk contraindicates administration of the drug to nursing mothers. Its identification in cerebrospinal fluid is of significance in terms of the drug crossing the blood-brain barrier; and its appearance in saliva and sweat helps to explain observed incomplete recovery of the radioactively labeled drug in urine, feces, and respiratory gases. Griggs and Jarman ( 1975) synthesized cyclophosphamide-4,4,6,6-d4 and added a known quantity of it to blood and urine samples from


239

patients. Blood was diluted with citrate-saline, red cells removed b y centrifugation, the drug extracted with chloroform, and concentrated to a small final volume. Urine was also extracted with chloroform. MS analyses were made by evaporating the crude chlorofomi extract inside the ion source from a direct probe and determining the relative heights of the peaks corresponding to the ( M - CH,Cl)- ion which are the most abundant peaks in the EI mass spectra of C P and the deuterated analog, respectively. For quantification, normal blood was spiked with the drug and a calibration curve obtained down to I pg/ml level. The technique was extended to include quantification of 4-ketocyclophosphamide in blood after extraction with ethyl acetate and methylation with diazomethane (Cox et al., 1976a). Analogs tetradeuterated in the bis-(2-chloroethy1)amino function were used as internal standard. The limit of detection was 1pg/ml, which was poorer than that of those methods involving prior gas chromatographic separation and selective ion monitoring (order of 10 ng/ml). The technique reported by Pantarotto et a2. (1974a) used the stable trifluoroacetyl derivative of C P with isophosphamide as internal standard. As confirmed b y MS, only one trifluoroacetyl group is picked u p b y the molecule. When CP is to be quantified in tissues, intereferences from endogenous substances require either laborious separation processes or high specificity. In a still further improved technique (Pantarotto et aE., 1976), the drug was determined by monitoring 4 selected masses in the CI mass spectra of the N-trifluoroacetyl derivatives of CP and the internal standard. The (M 1)+ions of both the drug and the internal standard are of the same mass, (mle = 357); however, GC retention times are different, thus monitoring presents no problem (Fig. 11).The other masses monitored are the isotope ion at mle = 357, and the two isotope ions corresponding to the loss of HC1 (mle = 321 and m/e = 323). The technique is linear in the 150 ng/ml to 100pg/ml range (serum). Incidentally, the technique can be used in reverse, i.e., for quantification of isophosphamide with C P as internal standard. The method is particularly u s e h l for the quantification of C P in tissues. For example, tissue concentrations determined in lung, kidney, heart, liver, spleen, and brain of C57BL16J mice bearing Lewis lung tumor were used to determine pharmacokinetic parameters of the 85 mg/kg C P injected intravenously 25 days after tumor implantation. Turning to the quantification of metabolites, Jardine et al. (1976) used GUMS with CI and selected ion monitoring to assay phosphoramide mustard and nornitrogen mustard in human serum and urine. Deuterium labeled analogs served as internal standards, and

+

240

JOHN ROBOZ

w 0 3

t _1

a E a

50

FIG. 11. Mass hagmentogram of brain extract, trifluoroacetate derivatives. The first set of peaks represents isophosphamide (internal standard), the second set represents cyclophosphamide. Reprinted from Pantarotto et al., 1976, with permission.

trifluoroacetylation used to form derivatives suitable for GC separation. An alternative derivatization technique, using ethereal diazomethane, yields mono-, di-, and trimethyl derivatives. The dimethyl derivatives are most suitable for selected ion monitoring; however, when interferences occur due to endogenous constituents, the trimethyl derivatives may be utilized. (Fenselau et aZ., 1975; Jardine et al., 1976). In a typical analysis, 22 pg/ml phosphoramide mustard was found in the serum of a patient 75 minutes after receiving a l-hour infusion of C P (50mgkg). In another experiment, 10,30, and 15pg/ml of CP, nornitrogen mustard, and phosphoramide mustard, respectively, were found in the O-4hour urine collection from a patient who received a dose of 15 mg/kg C P intravenously.


24 1

Jardine et al. (1978) extended the technique for the quantification of CP, phosphoramide mustard, and nornitrogen mustard in the plasma and urine of patients receiving C P therapy (60 or 70 mg/kg i.v.). All 3 compounds were available in deuterated form as internal standards. Peak plasma levels of C P and phosphoramide mustard (50-100 nmole/ ml) occurred at 3 hours after C P administration. Plasma half-lives, the origin of nornitrogen mustard, and the mechanism of action of CP are discussed. Both low- and high-resolution FD spectra have been obtained and interpreted for C P and almost all known metabolites (Schulten, 1974, 1976a,b). Quantification in tissues has been in the presence of interfering endogenous constituents. Once developed, this technique is likely to offer significant advantages over techniques requiring evaporation and derivatization, particularly for the analysis of nonvolatile (phosphorodiamidic acid) and thermally unstable (4-hydroxycyclophosphoramide, carboxyphosphoramide) metabolites.

B. NITROSOUREAS BCNU [ 1,3-bis(2-chloroethyl)-nitrosourea] was quantified (to 0.4 p M ) in plasma (Weinkam et al., 1978) by monitoring the (M + 1)+ ions of BCNU and its octadeutero analog (internal standard) upon evaporating the residue of a hexane-ether extract in a direct probe. May et al. (1974) investigated the role of hepatic mixed function oxidase in the metabolism of CCNU [ 1-(2chloroethyl)-3-cyclohexyl- 1nitrosoureal. Liver microsomes were prepared from both normal and phenobarbital treated rats and incubated with 14C-labeled CCNU. After removing excess CCNU with hexane, a metabolite was extracted with ether and purified by HPLC. EI mass spectra revealed that all fragments containing the cyclohexyl ring were 15 mass units higher in the metabolite than in intact CCNU. In addition, fragment ions of the metabolite indicated an unsaturated ring while the corresponding fragments of intact CCNU suggested a saturated ring. At the same time, the chloroethyl and nitroso groups were present in both compounds. These observations led to the identification of the metabolite as cis-4hydroxy CCNU, i.e., the cyclohexyl ring of the intact CCNU was stereospecifically hydroxylated by the liver microsomes. Additional evidence was obtained using cochromatography (thin-layer and liquid) with cyclohexyl (14C)-and2-chloroethyl (14C)-labeled CCNU, and NMR data on cis- and trans-Chydroxy CCNU. It is noted that, as expected, the hydroxylating activity of the NADPH- and 0,-fortified microsomes

242

JOHN ROBOZ

was induced by pretreatment with phenobarbital and inhibited by heating. In further studies on the metabolism of CCNU, and also methylCCNU ( 1-(2-chloroethyl)-3(trans-4-methyl cyclohexy1)-1-nitrosourea), Reed et al. (1975) identified 2-chloroethanol as the major product of nonenzymatic degradation in buffers, under physiological conditions, resulting from the deprotonation of the 3-nitrogen. The 2-chloroethyl moiety was trapped with C1-, Br-, and I- ions and the dichloroethane that formed identified by GC/MS. Other degradation products identified were acetaldehyde, vinyl chloride, ethylene, and cyclohexylamine. For the in v i m study of the metabolism of CCNU, Reed and May (1975) injected I4C-labeled CCNU into rats and isolated a metabolite from urine by HPLC which was identified by GUMS as a thiodiacetic acid. In addition, rapid microsomal hydroxylation of the cyclohexyl ring yielded five metabolites (cis or trans-2-hydroxy, trans-3-hydroxy, cis-3-hydroxy, cis-4-hydroxy, and trans-4-hydroxy-CCNU), all identified by MS. Essentially the same results were obtained by Hilton and Walker (1975) using HPLC for separation and MS for identification of the metabolites of CCNU both in vivo and in vitro. They concluded that although the hydroxylation of the cyclohexyl moiety of CCNU probably has no effect on the formation of the cytoxic intermediate, it certainly does effect protein binding, tissue distribution, and rate of excretion. The decomposition of BCNU ( l-3-bis(2-chloroethy1)-1-nitrosourea) in aqueous solution was investigated b y Colvin et al. (1974). A solution of 14C-labeled(in the chloroethyl group) BCNU was incubated in a buffer under various experimental conditions and volatile materials formed during decomposition were swept away with nitrogen and collected in ether at -70°C. Vinyl chloride, acetaldehyde, and 1,2dichloroethane were identified as volatile metabolites of BCNU by MS. When BCNU was allowed to decompose in aqueous solution in a gastight vial, followed by extraction of the product with ether, all the listed decomposition products were found again, plus another decomposition product (63% of the total), identified as chloroethanol. When chloroethylamine was reacted with nitrous acid in aqueous solution, the reaction products and their relative quantities were the same as observed in the decomposition of BCNU. The authors concluded that chloroethylcarbonium ion (or chloroethyldiazonium precursor) is generated during the decomposition of BCNU at physiologic pH and ternperature, and the reaction of this moiety with nucleic acids may explain the alkylating-like activity of BCNU in vivo. Hill et al. (1975)


243

employed MS to characterize the microsomal metabolism of the nitrosoureas. The metabolic product of BCNU was 1,3bis(2-chloroethyl) urea, the products of CCNU and methyl-CCNU were ringhydroxylated derivatives. Numerical values obtained on the rates of the microsomal metabolism led to the conclusion that microsomal metabolism may take place before chemical decomposition of the administered drugs occur. Ludlum et d.(1975) incubated BCNU with the synthetic polynucleotides poly(C) and poly(G). After chemical and enzymatic digestion and separation from unreacted nucleotides b y column chromatography, 3 major derivatized (TMSi) nucleotides were identified b y MS (also by UV): 3-P-hydroxyethyl) cytidine, 3,N4-ethano-cytidine, and 7-(P-hydroxyethyl)-8-oxoguanosine. The authors concluded that BCNU generates active two-carbon fragments, probably chloroethyl carbonium ions, which confer alkylating activity when attached to CMP or GMP with a nucleic acid strand.

c. PURINES, PYRIMIDINES, AND THEIRNUCLEOSIDES Pantarotto et aZ. (1974a) developed a GUMS-CI-selected ion monitoring technique for the quantification of 5-fluorouracil (5FU), 5-fluoso-2‘-deoxyuridine (5-FUdR), 6-mercaptopurine (6MP), 6-mercaptopurine ribonucleoside (GMPR), and cytosine arabinoside (AM-C) in fluids down to a level of 100 ng/ml; 0.1 ml serum was diluted with water and extracted with n-butanol. An aliquot of the butanol phase was evaporated to dryness and the derivatizing reagent trimethylbenzylammonium hydroxide added. Methylation occurs “on-column” in the sample injection port of the GC. The (M + 1)+ ions of the drugs and the internal standard (imipramine) were monitored in the CI mode. An application of this technique (with minor modifications) is shown in Figs. 12 and 13. Several peaks appeared when the masses mle = 181 (5FU) and mle = 159 (6MP, used here as internal standard) were monitored in the GC effluent, but these did not interfere with those representing 5FU and 6MP. Peak areas were measured in terms of counts, and quantification was made with the aid of a calibration curve obtained by plotting area ratios of 5FU/6MP for known quantities of 5FU added to pooled normal serum. The same technique can also be used to quantify 6MP using 5FU as internal standard. Finn and Sadee (1975) quantified 5FU utilizing isotope dilution combined with selected ion monitoring. Highly enriched 14C-labeled 5FU (

140

r= v,

B c

,

l t r

4331M’)

282 (M-CF3CO) 223

73

50

364(H-Cf31

II

W

IT

307 293 3 20

126

IN HUMAN URINE

5

255

310(M-CF3)

86.100 112

I h

I.

.I

1111

I

-

1

I

, I

379(H*I ?Q6364,

drolyzed urines of patients with leukemia; free putrescine and spermidine were identified in unhydrolyzed urine. A large unknown peak appearing in the nonhydrolyzed urine of a patient with acute myelocytic leukemia was identified as monoacetylsperminidine (Fig. 16), a metabolic product of spermidine. Another unknown was identified as 1,3-diaminopropane, also related to spermidine metabolism. Significant increase was obtained in the acetylspermidine content of urine of leukemic patients under therapy with vincristine and prednisone after treatment with vincristine; the ratio of acetylspermidine to spermidine excretion appeared constant (Denton et al., 1973). Smith and Daves (1977) added deuterated putrescine-d,, cadaverine-d,, spermidine-d,, spermine-d,, and acetylspermidine-A-d2 to the samples prior to separation and derivatization. Thus, all possible losses during manipulations were the same for the compounds and internal standards. All nondeuterated trifluoroacetylpolyamines have their base peak at mle = 126 ( F,CCONH = CH,+); deuterated, derivatives appear 2 mass units higher. To increase sensitivity, ions resulting from the loss of CF, or CF,CO were monitored. This permitted the use of a large excess (a100 pmole) of deuterated standard without changing the peak height for the mass of the unlabeled polyamine, resulting in a “carrier effect.” Linear response curves were obtained in the 1-1000 pmole range.

256

JOHN ROBOZ

C. STEROIDS Zumoff et nl. (1975) reviewed the relationship of urinary excretion of androgen metabolites and estrogens to the natural history of breast cancer and concluded that the reported relationships of low urinary excretion of androgen metabolites to increased risk of breast cancer, and the protective value of a high ratio of urinary estriol to estrone estradiol are dubious. According to their hypothesis, the lack of active estrogen prior to age 30 is of protective value. MS has not been used extensively for the monitoring of steroids in large scale profile studies, although analytical techniques are readily available. Adlercreutz et nl. (1975) developed a technique for the determination of at least l l estrogens in body fluids or tissues with a sensitivity as low as 10-14M. Luyten and Rutten (1974) utilized highresolution capillary columns in the GC to separate urinary steroids so that a rather complete urinary steroid profile may be obtained in a single analysis. In what follows, several pilot studies are reviewed on the use of MS to identify potential markers in neoplasis. Reimendal and Sjovall ( 1973) developed computerized procedures for the rapid routine identification of a large number of steroids in profiles. For example, it took only 5 minutes to obtain a complete profile of steroids in the disulfate fraction from the plasma of a patient with choriocarcinoma. The concentration of the pregnanediol isomers was considerably higher in this patient than in normals; the significance of this finding is not known. Millington et al. (1974) and Millington (1975) developed a technique for the identification and quantification of endogenous steroids in human breast tumors by high-resolution mass fragmentography . Monitoring exact masses resulted in specificity and elimination of all interferences so that crude tissue extracts could be analyzed. Estradiol-l7/3 was the major classical estrogen in the primary breast tumors studied, dehydroepiandosterone was found in all tumors, and testosterone could not be found in any tumor. Most other steroids found were present in the range of 0.5-10 ng/gm tissue. Similar techniques were employed to the analysis of CI9 steroids in human benign hyperplastic prostate tissue (Millington et al., 1975).Sensitivity limit was at the feintomole level and quantification improved with the use of epimers as internal standards. The most predominant steroid found was 5a-dihydrotestosterone (0.8-40 ng/gm tissue). Other steroids of interest included the 3a, 17p and 3p, 17p isomers ofthe 5a-androstanediols (0.01-2.0 ng/gm range for most samples), and testosterone (0.07-5.0 nglgm).

+


257

High-resolution mass fragmentography was used by Snedden and Parker (1976) to identify and quantify oestrone, oestradiol, oestriol, and progesterone directly in ovarian tissues (follicle corpus and various cysts). The specificity of the technique permitted the use of crude extracts. Different levels of progesterone were found in normal and diseased tissues of the same ovary; also, steroid levels changed parallel to the mentrual cycle. Maynard ( 1977) and Maynard et al. (1977) used GUMS to quantify silylated steroid derivatives in malignant and benign breast tissues. Levels of dehydroepiandrosterone varied significantly (malignant: 20-200 ng/gm tissue, benign: 200-1000 ng/gm), while 5-androstane3/3,17/3-diol levels were constant ( 10-200 ng/gm). Some 5 mg/day of an unknown steroid was detected in the urine of a patient with inoperable adrenal carcinoma (Fantl, 1973). After isolation, glucoronidase hydrolysis, chloroform extraction and TLC purification, the compound was identified by MS as &,16a-dihydroxyandrost-5-en-17one. This metabolite has a hydroxyl group in the C-16 position and apparently inhibits the action of steroid A4-5isomerase leading to the formation of a 5-ene-3a-hydroxy group. The only other known urinary metabolite with the same effect is pregn-5en-3a716a,20triol, identified by MS as a trimethylsilyl derivative in the urine of normal adult men and women, patients with various disorders, and patients with adrenal carcinoma (Fantl et al., 1973). The latter group excreted the compound u p to 15 mg/24 hour compared to 50-220 pg/24 hours for all other subjects. No other steroid hormones were excreted in excess amounts in the urine of the patients with carcinoma. Possible explanations for the appearance of these metabolites, are given in the references cited. D. MISCELLANEOUS

1. Catecholamines Abnormal secretion and metabolism of dopamine, norepinephrine, and epinephrine have repeatedly been shown to b e related to pheochromocytoma and neuroblastoma. Fluorimetry, GC, and GUMS have been used to quantify catecholamines. In a technique developed by Wang et al. (1975a) blood is deproteinized with perchloric acid and free catecholamines are derivatized with trifluoroacetic anhydride after several steps of purification. For the determination of total catecholamine content, an additional step of hydrolysis is included. To avoid interferences, selected ions may be

258

JOHN ROBOZ

monitored with the MS. Both the free and total catecholamine content was found considerably higher in patients with pheochromocytoma and neuroblastoma than in normal subjects. The more specific detection resulted in values lower than previously reported indicating that interfering materials might lead to false positives. This work was extended (Wang et al., 1975b) to include 3-methoxytyramine, normetanephrine, and metanephrine. All three 3-0-methylated catecholamines were found in much larger quantities in the plasma of patients with neuroblastoma and, to a lesser extent, pheochromocytoma, than in that of normal subjects, e.g., 510 pmole/ml vs 4-10 pmole/ml for 3-methoxytyramine. 2. Bile Acids and Sterols Methodology for the profiling and quantification of bile acids, such as cholic, deaxycholic, and lithocholic acids, as well as sterols, such as cholesterol arid coprostanol, has been available (Sjovall et al., 1971)for the investigation of the suggested association between concentrations of fecal bile acids and cholesterol metabolites and colon cancer. The fecal excretion of cholesterol, coprostanol, and cholestane-3/3,5a76/3trio1 was higher in patients with ulcerative colitis than in other groups (Reddy et al., 1977). The main role of mass spectrometry in such investigations is the positive identification of all components included in the profiling.

3. Volatiles G U M S has been extensively used to analyze underivatized volatile components in breath, body fluids, and tissues, both to detect and quantify preselected individual constituents and to obtain profiles in searching for aberrations associated with disease states. The analytical methodology is well developed [Politzer et al. (1976) reviews all aspects] and is ready for applications. Lovett et al. (1976) conducted a pilot study to detect differences in the volatiles from urine and ascitic fluid of normal and tumor bearing (Ehrlich ascites) mice. Some 20 compounds were identified and about one-half of these are also known to exist in human urine. The relative amounts of several components in urine changed as the disease progressed. The profiles of ascitic fluids contained several unidentified peaks not present in the salinewashed abdominal cavities of normal mice.

4. Lymphocytes In linear programmed thermal degradation MS, a small sample is placed at the tip of a direct insertion probe, the entire sample is slowly evaporated by programmed temperature, and certain preselected mas-


259

ses are monitored so that individual profiles are obtained. Some 10- 100 p g of lyophilized lymphocytes from various leukemia patients were placed on the solid probe with drops of methanol, and a number of degradation profiles were obtained upon heating. The profiles showed similarities between patients with the same disease, while considerable differences were found between cells of chronic lymphocytic and hairy cell leukemia, It was concluded that future work was warranted and the technique should also be applicable to isolates of other cellular fractions (Yergey et al., 1978).

5. Carbohydrates

The carbohydrate moiety of carcinoembryonic antigen samples from four different tumors were investigated by Coligan et a,?. (1976). Fucose, mannose, galactose, N-acetylneuraminic acid, and N-acetyglucosamine were released, permethylated, and analyzed as alditol acetates by gas chromatography-mass spectrometry. No qualitative differences were found when different antigen preparations were compared. On the bases of the quantitative data, certain constrains were proposed on the structure of the carbohydrate moiety of carcinoembryonic antigen. A mass spectrometric technique was developed for the quantification of neuraminidase-susceptible and total N-acetylneuraminic acid (NANA) in biological materials (Roboz et al., 1978b).NANA is determined as the fully silylated compound by monitoring the protonated molecular ion produced by isobutane chemical ionization. Unlike the commonly employed thiobarbituric method (based on colorimetric detection), this technique provides positive identification of NANA, eliminates all interferences b y 2-deoxyribose and other carbohydrates, and is capable of detecting as little as 400 pg (0.6 x mole) of NANA. The first application of the technique was the determination of neuraminidase-susceptible NANA in leukemic cells employed in chemoimmunotherapy of acute myelocytic leukemia. 6. Thymidine Zncorporation Heck et al. (1977) studied in wiwo synthesis of DNA in the rat b y incorporating multilabeled, nonradioactive thymidine into the DNA of cells undergoing replication. The isotopic composition of the thymine products were determined b y both field desoprion mass spectrometry and scintillation counting in various organs.

E. CONCLUSIONS Combined gas chromatography-mass spectrometry is the best tool available currently for the multicomponent analysis of environmental

260

JOHN ROBOZ

pollutants and exogenous and/or endogenous metabolites in body fluids. The technique has been used to search for (i) environmental carcinogens samples, (ii) metabolites of both carcinogens and antineoplastic agents both in vivo and in vitro, and (iii) biological markers of cancer. Once detected, suspected identities of individual compounds are confirmed by comparing complete mass spectra to those of authentic compounds. Unknowns are identified by determining exact masses to a precision of 1millimass unit, and calculating all possible mol.ecular compositions. The computerized technique of selected ion monitoring, particularly in combination with chemical ionization, provides a unique combination of high sensitivity (picomole range), specificity, and general applicability for the quantification of selected carcinogens, antineoplastic agents, metabolites, and endogenous contituents in crude environmental or biological samples .

SELECTEDGENERALBIBLIOGRAPHY Budzikiewicz, H., Djerassi, C., and Williams, D. (1967). “Mass Spectrometry of Organic Compounds.” Holden-Day, San Francisco, California. McFadden, W. ( 1973). “Techniques of Combined Gas Chromatography/Mass Spectrometry.” Wiley (Interscience), New York. McLaffetry, F. (1973). “Interpretation of Mass Spectra,” 2nd ed. Benjamin, New York. Roboz, J. ( 1968). “Introduction to Mass Spectrometry.” Wiley (Interscience), New York. Waller, G., ed. (1972). “Biochemical Applications of Mass Spectrometry.” Wiley, New York. Watson, J. (1976:).“Introduction to Mass Spectrometry: Biomedical, Environmental, and Forensic Applications.” Raven Press, New York.

REFERENCES Abramson, F. P. (1972).Anal. Chem. 44,28A-35A. Adlercreutz, H. ( 1974). In “Mass Spectrometry in Biochemistry and Medicine” (A. Frigerio and N. Castagnoli, eds.), pp. 165-181. Raven, New York. Adlercreutz, A. ( 1977). I n “Quantitative Mass Spectrometry in Life Sciences” (A. P. De Leenheer and R. R. Roncucci, eds., pp. 15-28. Elsevier, Amsterdam. Adlercreutz, H., Nieminen, U., and Ervast, H. S . (1974)./. Steroid Biochem. 5,619-626. Adlercreutz, H., Martin, F., Wahlroos, O., and Soini, E. (1975)./. Steroid Biochem. 6, 247-259. AEI (1976). Application Note No. MS30819. AEI Scientific Apparatus, Manchester, England. Ahlstrom, D. H., Kilour, R. J., and Liebman, S. A. (1975). Anal. C h m . 47, 114-1412. Alford, A. (1975a).In “Environmental Applications of Advanced Instrumental Analysis,” EPA-660/4-75-004 (Environ. Monit. Ser.). U.S. Environ. Protect. Agency, Corvalhs, Oregon. Alford, A. (1975b). Biomed. Mass Spectrom. 2 , 229-253. Alford, A. (1977). Biomed. Mass Spectrom. 4, (in press).


26 1

Arcamone, F., Cassinelli, G., Franceschi, G., Orezzi, P., and Mondelli, R. (1968). Tetrahedron Lett. 30, 3349-3352. Arcamone, F., Cassinelli, G., Franceschi, G., Penco, S., Pol, C., Redalli, S., and Selva, A. (1972).Int. Symp. Adriamycin [Proc.],1971 pp. 9-22. Arcamone, F., Penco, S., and Vigevani, A. (1975). Cancer Chemother. Rep. 6, 123-129. Asbell, M., Schwartzbach, F., Bullock, F., and Yesair, D. (1972). J . Pharmacol. E x p . Ther. 182, 63-67. Aschbacher, P. (1976).J . Toxicol. Enoiron. Health 1, Suppl. 45-59. Bachur, N. R. (1971).J.Pharmacol. Exp. Ther. 177,573-578. Bachur, N. R. (1975).Cancer Chemother. Rep. 6, 153-158. Bachur, N. R., Hilderbrand, R. C., and Jaenke, R. S. (1974).J.Pharmacol. E x p . Ther. 191, 33-1340. Bartos, F., Bartos, D., Grettie, D., Campbell, R., Marton, L., and Smith, R. (1977). Biochem. Biophys. Res. Commun. 75,915-919. Bettencourt, A., Lhoest, G., Roberhoid, M., and Mercier, M. (1977).J.Chromatogr. 134, 323-330. Blumer, M. (1975). Finnigan Spectra 5, 8-10. Bonelli, E . J., Taylor, P. A,, and Morris, W. J. (1975).Am. Lab., July, 29-35. Brooks, J. B., Alley, C. C., and Jones, R. (1972).Anal. Chem. 44, 1881-1884. Brown, R., and Jacobs, M. (1972).Pittsburgh Conf. Anal. Chem. Paper No. 205. Brown, R., Jacobs, M. L., and Taylor, H. E. (1972).Am. Lab., Nov., 29-40. Brunnemann, K. D., and Hoffmann, D. (1976).In “Carcinogenesis” (R. Freudenthal and P. Jones, eds.), Vol. I, pp. 283-297. Raven, New York. Bullock, F., Bruni, R., and Asbell, M. (1972).J.Pharmacol. E x p . Ther. 182,70-76. Burse, V. W., Moseman, R. F., Sovocool, G. W., and Villanueva, E. C. (1976). Bull. Enoiron. Contam. Toxicol. 15, 122-128. Caustland, D., Fischer, D., Kolwyck, K., Duncan, W., Wiley, J., Menon, C., Engl, J., Selkirk, J., and Roller, P. (1976).In “Carcinogenesis” (R. Freudenthal and P. Jones, eds.), Vol. 1, pp. 349-381. Raven, New York. Ciccioli, P., Bertoni, G . , Brancaleoni, E., and Bruner, F. (1976).J. Chromatogr. 126, 757-770. Coligan, J., Pritchard, D., Schnute, W., andTodd, C. (1976).Cancer Res. 36,1915-1917. Collier, L. (1972). Enoiron. Sci. Technol. 6, 930-932. Colvin, M., Padgett, C. A., and Fenselau, C. (1973). Cancer Res. 33, 91.5918. Colvin, M., Coewns, J. W., Brundrett, R. B., Kramer, B. S., and Ludlum, D. B. (1974). Biochem. Biophys. Res. Commun. 60,515-520. Colvin, M., Brundrett, R. B., Kan, M. N., Jardine, I., and Feneslau, C. (1976).Cancer Res. 36, 1121-1126. Connors, T. A., Foster, A. B., Gilsenan, A. M., Jarman, M., and Tisdale, M. J. (1972). Biochem. Pharmacol. 21, 1309-1316. Connors, T. A., Cox, P. J., Farmer, P. B., Foster, A. B., and Jarman, M. (1974a).Biochem. Pharmacol. 23, 115-129. Connors, T. A,, Cox, P. J., Farmer, P. B., Foster, A. B., and Jarman, M. (1974b).In “Mass Spectrometry in Biochemistry and Medicine” (A. Frigerio and N. Castagnoli, eds.), pp. 19-27. Raven, New York. Connors, T. A., Cox, P. J., Farmer, P. B., Foster, A. B., Jarman, M., and McLeod, J. K. ( 1 9 7 4 ~ )Biomed. . Mass Spectrom. 1, 130-136. Connors, T., Hickman, J., Jarman, M., Melzack, D., and Ross, W. (1975). Biochem. PharmacoZ24, 1665-1670. Cox, P. J., Farmer P. B., Foster, A. B., Gilby, E. D., and Jarman, M. (1976a). Cancer Treat. Rep. 60, 483-493.

262

JOHN R O B 0 2

Cox, P. J., Farmer, P. B., and Jarman, M. (1976b).Adu.Mass Spectrom. Biochem. Med. 1, 59-7 1. Cradock, J., Egorin, M., and Bachur, N. R. (1973). Arch. Znt. Pharacodyn. Ther. 202, 48-61. Crombie, L., Gaines, M., and Pointer, D. (1968).J.Chem. SOC., Sect. C , 1347-1362. Curley, A., Burse, V. W., Jennings, R. W., Villanueva, E. C., and Kimbrough, R. D. (1975). Bull. Enuiron. Contam. Toxicol. 14, 153-158. Davidson, J. P., Faber, P. J., Fischer, R. G., Mansy, S., Peresie, H. J., Rosenberg, B., and VanCamp, LA.(1975). Cancer Chemother. Rep. 59, 287-300. Denton, M., Glazer, H., Walle, T., Zellner, D., and Smith, F. (1973). In “Polyamines in Normal and Neoplastic Growth” (D. Russel, ed.), pp. 373-380. Raven, New York. Draeger, U., Peter, G., and Hohorst, H. J. (1976). Cancer Treat. Rep. 60, 355-361. Duncan, J. H., Colvin, M., and Fenselau, C. (1973). Toxicol. Appl. Pharmacol. 24, 3 17-322. Durie, B. G., Salmon, S. E., and Russell, D. H. (1977).Cancer Res. 37, 214-221. Editorial. (19774. Chem. G Eng. News Jan. 17, pp. 35-37. Editorial. (19771)).Chem. G Eng. News Mar. 28, pp. 7-8. Eichelberger, J. W., Harris, L. E., and Budde, W. L. (1974).Anal.Chem. 46,227-232. Ember, L. (1975). Enuiron. Sci. Technol. 9, 1116-1121. Engel, P., Weidenfeld, J., and Merriam, G. (1976).J.Toxicol. Enuiron. Health 1, Suppl. 31-44. Evans, C., and Morrison, G. (1968). Anal. Chem. 40,869-875. Evans, K., Mathias, A., Mellor, N., Silvester, R., and Williams, A. (1975).Anal. Chem. 47, 821-824. Fantl, V. (1973).J. Endocrinol. 56, 615-616. Fantl, V., Booth, M., and Gray, C. H. (1973).J.Endocrinol. 57, 135-142. Fenselau, C., Kan, M. N., Billets, S., and Colvin, M. (1975).CancerRes. 35,1453-1457. Fenselau, C., Kan, M. N., Subba Rao, S., Myles, A., Friedman, O., and Colvin, M. (1977). Cancer Res. 37, 2538-2543. Fine, D. H., Rounbehler, D. P., Pelizzari, E. D., Bunch, J. E., Berkley, R. W., McCrae, J., Bursey, J. T., Sawicki, E., Krost, K., and DeMarrais, G. A. (1975). Bull. Enuiron. Contam. Toxicol. 15,739-746. Finn, C., and Sadee, W. (1975).Cancer Chemother. Rep. 59, 279-286. Fischbein, L. (1972).]. Chromatogr. 68, 343-426. Fitchett, A. W., Buck, R. P., and Mushak, P. (1974).Anal. Chem. 46,710-713. Frankel, L. S., McCallum, K. S., and Collier, L. (1974). Enuiron. Sci. Technol. 8, 356359. Fujii, T. (1977). Anal. Chem. 49, 1985-1987. Gaffield,W., Fish, R., Holmstead, R., Poppiti, J., and Yergey, A. (1976).ZARC Sci. Publ. 14, 11-20, Gammans, R., Sehon, R., Anders, M., and Hanna, P. (1977). Drug Metab. Dispos. 5, 310-316. Garrison, A. (1976).Ann. N . Y.Acad. Sci. 298, 2-19. Genrke, C. W., and Patel, A. (1977).J. Chromatogr. 130, 103-114. Gierlich, H. H., Heindrichs, A., and Beckey, H. D. (1974).Reu. Sci. Instrum. 45, 12081211. Gold, A. (1975).Anal. Chem. 47, 1469-1472. Gothe, R., Calleman, C. J., Ehrenberg, L., and Wachtmeister, C. A. (1974).Ambio 3, 234-236. Gough, T. A., and Webb, K. S. (1973).J . Chromatogr. 79, 57-63. Gough, T. A., Sugden, K., and Webb, K. S. (1975).Anal. Chem. 47,509-512.


263

Cough, T. A., Webb, K. S., Pringuer, M., and Wood, B. ( 1 9 7 7 ) ~Agric. . Food Chem. 25, 663-667. Griggs, L. J., and Jarman, M. (1975).J. Med. Chem. 18, 1102-1106. Grob, K. (1973).J. Chromatogr. 84,255-273. Grob, K., and Grob, G. (1971).J.Chromatogr. 62, 1-13. Grob, K., and Grob, G. (1974).I. Chromatogr. 90, 30-313. Haake, P., and Mastin, S. H. (1971).J.Am. Chem. SOC.93,6823-6828. Haddon, W., Masri, M., Randall, V., Elsken, R., and Meneghelly, B. (1977).J.Assoc. On. Anal. Chem. 60, 107-113. Harris, L. E., Budde, W. L., and Eichelberg, J. W. (1974).Anal. Chem. 46, 1912-1917. Harrison, W. W., and Clemena, G. G. (1972).Clin. Chim. Acta 36,485-492. Hase, A., Lin, P. H., and Hites, R. A. (1976). In “Carcinogenesis” (R. Freudenthal and P. Jones, eds.), Vol. 1, pp. 435-442. Raven, New York. Hecht, S., Throne, R., Maronpot, R., and Hoffman, D. (1975).J . Natl. Cancer lnst. 55, 1329-1336. Hecht, S., Chen, C., and Hoffmann, D. (1978a). Cancer Res. 38, 215-218. Hecht, S., Chen, C., Ornaf, R., Jacobs, E., Adams, J., and Hoffmann, D., (1978b).J. Org. Chem. 43, 72-76. Heck, H., McReynolds, J., and Anbar, M. (1977). Cell Tissue Kinet. 10, 111-119. Hignite, C., Shen, D., and Azarnoff, D. (1978).Cancer Treat. Rep. 62, 13-18. Hill, D. L., Kirk, M. C., and Struck, R. F. (1975). Cancer Res. 35, 296-301. Hillcoat, B. L., Kawai, M., McCulloch, P. B., Rosenfeld, J., and Williams, C. K. (1976). Br. J. Clin. Pharmacol. 3, 135-143. Hilton, J., and Walker, M. D. (1975). Biochem. Pharmacol. 24, 2153-2158. Holzer, G., Oro, J., and Bertsch, W. (1976).J.Chromatogr. 126, 771-785. Hunt, D. F., Stafford, G. C., Shabanowitz, J., and Crow, F. W. (1977).Anal. Chem. 49, 1884. Issaq, H., Schroer, J., and Barr, E. (1977). Chem. Instrum. 8, 51-53. Janini, G., Johnston, K., and Zielinski, W. (1975).Anal. Chem. 47,670-674. Janini, G., Muschik, G., and Zielinski, W. (1976a).Anal. Chem. 48,809-813. Janini, G., Muschik, G., Schroer, J., and Zielinski, W. (1976b).Anal. Chem. 48, 18791883. Jardine, I., Brundrett, R., Colvin, M., and Fenselau, C. (1976). Cancer Treat. Rep. 60, 403-408. Jardine, J., Fenselau, C., Appler, M., Kan, M., Brundrett, R., and Colvin, M. (1978). Cancer Res. 38, 408-415. Jarman, M., Gilby, E. D., Foster, A. B., and Bondy, P. K. (1975).Clin. Chim. Acta 58, 61-69. Jensen, S., and Sundstrom (1974).Ambio. 3, 70-76. Johnson, B. M., Worzalla, J. F., and Bryan, G. T. (1973). 21st Annu. Conf. Mass Spectrom., 1973 pp. 261-265. Johnson, B. M., Wonalla, J. F., and Kaiman, B. D. (1974). 22nd Annu. Con$ Mass Spectrom., 1974 pp. 176-179. Johnson, B. M., Nakamura, S., Johnson, C., and Citrin, D. (1977).25th Annu. Conf. Mass Spectrom., 257-259. Keith, L. H., ed. (1976a). “Identification and Analysis of Organic Pollutants in Water.” Ann Arbor Sci. Publ., Ann Arbor, Michigan. Keith, L. H. (1976b). Enoiron. Sci. Technol. 10,555-564. Kirk, M. C., Coburn, W. C., and Piper, J. R. (1976).Biomed. Mass Spectrom. 3,245-247. Lao, R., Thomas, R., and Dubois, L. (1973).Anal. Chem. 45,908-915. Lao, R., Thomas, R., and Monkman, J. (1976a). Carcinog.4ompr. Suru. 1,271-281.

264

JOHN ROBOZ

Lao, R., Thomas, R., and Monkman, J. (1976b).I n “Carcinogenesis” (R. Freudenthal and P. Jones, eds.), Vol. 1, pp. 271-281. Raven, New York. Lee, M. L., and Hites, R. A. (1976).Anal. Chem. 48, 1890-1893. Lee, M. L., Novotny, M., and Bartle, K. D. (1976a).Anal. Chem. 48, 405-416. Lee, M. L., Novotny, M., and B a d e , K. D. (1976b).Anal. Chem. 48, 1566-1572. Levine, S. P., Hebel, K. G., Bolton, J., and Kugel, R. E. (1975).AnaL Chem. 47, 1075A1080A. Lijinsky, W. (1976).Ambio 5,67-72. Lovett, A., Sherwood, E., and Oro, J. (1976). Pap., 171st Meet. Am. Chem. SOC. (to be published). Ludlum, D. B., Kramer, B. S., Wang, J., and Fenselau, C. (1975). Biochemistry 14, 5480-5485. Luyten, J., and Rutten, G. (1974).J . Chromatogr. 91, 393-406. McFadden, W., and Schwartz, H. L. (1976).J . Chromatogr. 122, 389-396. McFadden, W., Bradford, D., Games, D., and Gower, J. (1977).Am. Lab., 55-64 (Oct. issue). Macfarlane, R. D., and Torgeson, D. F. (1976). Science 191, 920-925. Maurer, K., and Rapp, U. (1976).Adu. Mass Spectrom. Biochem. Med. 1,541-551. May, H. E., Boose, R., and Reed, D. J. (1974). Biochem. Biophys. Res. Commun. 57, 426-433. Maynard, P. (1977).J . Endocrinol. 73, 16P. Maynard, P., Pike, A., Weston, A., and Griffiths, K. (1977).Eur. J. Cancer 13, 971-975. Miller, J,, Kigwana, L., Streeter, D., Robins, R., Simon, L., and Roboz, J. (1977).Ann. N.Y. Acad. Sci. 284,221-229. Millington, D. S. (1975).J . Steroid Biochem. 6, 239-245. Millington, D. S., Jenner, D. A., Jones, T., and Griffiths, K. (1974).Biochem.]., 139,473475. Millington, D. S., Buoy, M. E., Brooks, G., Harper, M. E., and Griffiths, K. (1975). Biomed. Mass Spectrom. 2, 219-224. Milne, G. W., and Heller, S. R. (1976).Am. Lab., Sept., 43-54. Milne, G., and Lacey, M. (1974). Crit. Rev. Anal. Chem. pp. 45-104. Miyazaki, M., Ishibashi, M., Itoh, M., and Yarnashita, K. (1977).J . Chromatogr. 133, 311-318. Montgomery, J. A., and Struck, R. F. (1976). Cancer Treat. Rep. 60, 381-395. Morrison, G. (1967).I n “Nuclear Activation Techniques in the Life Sciences,” pp. 211228. IAEA, Vienna. Morrison, G. (1969). Trace Subst. Enuiron. Health-2, Proc. Uniu. Mo. Annu. Conf., 2nd, 1968 pp. 307-318. Mueller, G., Norpoth, K., and Eckard, R. (1976).Int. Arch. Occup. Environ. Health 38, 69-75. Myers, C. E., Lippman, M. E., EIiot, H. M., and Chabner, B. A. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 3683-3686. Nadel, E. (1969).Ann. N.Y. Acad. Sci. 161,581-585. Novotny, M., Lee, M., and Bartle, D. (1974).J . Chromatogr. Sci. 12, 606-609. Oswald, E. O., Albro, P. W., and McKinney, J. D. (1974).J. Chromatogr. 98, 363-448. Pantarotto, C., Bossi, A., Belvedere, G., Martini, A,, Donelli, M. G., and Frigerio, A. (1974a).J. Pharm. Sci. 63, 1554-1558. Pantarotto, C., Martini, A., Belvedere, G., Bossi, A,, Donelli, M. G., and Frigerio, A. (1974b).J . Chromatogr. 99,519-527. Pantarotto, C., Martini, A., Belvedere, G., Donelli, M. G., and Frigerio, A. (1976). Cancer Treat. Rep. 60,493-500.


265

Patel, A. B., and Gerhke, C. W. (1977).j . Chromatogr. 130, 115-128. Pellizzari, E. D., Bunch, J. E., Bekley, R. E., and McRee, J. (1976). Anal. Chem. 48, 803-807. Pensabene, J. W., Fiddler, W., Dooley, C. J., Doerr, R. C., and Wasserman, A. B. (1972). 1.Agric. Food Chem. 20, 274-277. Politzer, I. R., Dowty, B. J., and Laseter, J. (1976). Clin. Chem. 22; 1775-1788. Preussmann, R., Arjungi, K. S., and Ebers, G. (1976). Cancer Res. 36,2459-2462. Przybylski, M., Ringsdorf, H., Voelcker, G., Draeger, U., Peter, G., and Hohorst, H. J. (1976). Cancer Treat. Rep. 60,501-509. Rawls, R. (1976).Chem. 6 Eng. News, Sept. 20, 33-34. Reddy, B. S., Martin, C. W., and Wynder, E. (1977).Cancer Res. 37, 1697-1701. Reed, D. J., and May, H. E. (1975). L$e Sci. 16,1263-1270. Reed, D. J., May, H. E., Boose, R. B., Gregory, K. M., and Beilstein, M. A. (1975).Cancer Res. 35, 568-575. Reimendal, R., and Sjovall, J. B. (1973).Anal. Chem. 45, 1083-1089. Reinhold, V., and Bruni, R. (1976). Biomed Mass Spectrom. 3, 335-339. Roboz, J. (1974). 22nd Annu. Conf. Mass Spectrom. p. 184. Roboz, J. (1975).Ado. Clin. Chem. 17, 109-193. Roboz, J. (1977). In “Mass Spectrometry in Drug Metabolism” (A. Frigerio and E. Ghisalberti, eds.), pp. 343-364. Plenum, New York. Roboz, J., and Ohnuma, T. (1977). In “Mass Spectrometry in Drug Metabolism,” (A. Frigerio and E. Ghisalberti, eds.), pp. 151-166. Plenum, New York. Roboz, J., and Suzuki, R. (1978).]. Chrom. (in press). Roboz, J., Suzuki, R., Bekesi, G., and Hunt, R. (1977). Biomed. Mass Spectrom. 4, 291-296. Roboz, J., Flor, R., Alessandro, R., and Ohnuma, T. (1978a).Proc. Am. Assoc. Cancer Res. 19, 129. Roboz, J., Suzuki, R., and Bekesi, G. (197813).Anal. Biochem. (in press). Roller, P. P., Sutphin, M., and Aszalos, A. A. (1976).Biomed. Mass Spectrom. 2,166-171. Roos, I. A., Thomas, A. J., and Eagles, J. (1974). Chem.-Biol.Interact. 8, 421-427. Rosenberg, B. (1975). Cancer Chemother. Rep. 59,589-598. Rosenfeld, J., Taguchi, V., Hillcoat, B., and Kawai, M. (1977).Anal. Chem. 49,725-727. Rote, J. W., and Morris, W. J. (1973).]. Assoc. 08. Anal. Chem. 56, 188-199. Russell, D. H. (1971). Nature (London) 233, 144-145. Russell, D. H. (1977). Clin. Chem. 23, 22-27. Russell, D. H., and Russell, S. D. (1975).Clin. Chem. 21, 860-863. Sadee, W., Finn, C., and Staroscik, J. (1976). Adv. Mass Spectrom. Biochem. Med. 1, 509-515. Safe, S., and Hutzinger, 0. (1972).J . Chem. Soc., Perkin Trans. I 5, 686-691. Safe, S., and Hutzinger, 0. (1973).“Mass Spectrometry of Pesticides and Pollutants.” CRC Press, Cleveland, Ohio. Safe, S . , Platonow, N., Hutzinger, O., and Jamieson, W. D. (1975). Biomed. Mass Spectrom. 2, 201-203. Schulte, E., and Acker, L. (1974).Z. Anal. Chem. 268,260-264. Schulten, H. R. (1973).In “New Methods in Environmental Chemistry and Toxicology” (F. Coulston, F. Korte, and M. Goto, eds.), pp. 31-42. Int. Acad. Printing, Tokyo. Schulten, H. R. (1974). Biomed. Mass Spectrom. 1, 223-230. Schulten, H. R. (1976a).Cancer Treat. Rep. 60,501-507. Schulten, H. R. (1976b).Adv. Mass Spectrom. Biochem. Med. 1, 289-298. Schulten, H. R. (1977). In “Methods of Biochemical Analysis” (D. Click, ed.), Vol. 24, pp. 313-448. Wiley, New York.

266

JOHN ROBOZ

Schwartz, M. K. (1975). Cancer Res. 35, 3481-3487. Schwartz, M. K. (1976). Cancer 37,542-548. Segal, A., VanDuuren, B., and Mate, U. (1975).Cancer Res. 35, 2154-2159. Selkirk, J., Croy, R., and Gelboin, H. (1974a).Science 184, 169-171. Selkirk, J., Croy, R. G., Roller, P., and Gelboin, H. (1974b).Cancer Res. 34,3474-3480. Selkirk, J., Croy, R., and Gelboin, H. (1975).Arch. Biochem. Biophys. 168, 322-326. Sen, N. P. (1972). Food Cosmet. Toxicol. 10, 219-223. Sen, N. P., Schwinghamer, L. A., Donaldson, B. A., and Miles, D. F. (1972). J. Agric. Food Chem. 20, 1280-1281. Sen, N. P., Donaldson, B., Iyengar, J. R., and Panalaks, T. (1973).Nature (London) 241, 473-474. Sen, N. P., Miles, W. F., Seaman, S., and Lawrence, J. F. (1976).J. Chromatogr. 128, 169-173. Severson, R., Snook, M., Higman, H., Chortyk, O., and Akin, F. (1976). In “Carcinogenesis” (R. Freudenthal and P. Jones, eds.), Vol. l, pp. 253-270. Raven Press. Shadoff, L. A., Kallos, G. H., and Woods, J. S. (1973).A n d . Chem. 45,2341-2344. Sjovall, J., Eneroth, P., and Ryhage, R. (1971). In “The Bile Acids” (P. Nair and D. Kritchevsky, eds.), pp. 209-247. Plenum, New York. Sladek, N. (1973).Cancer Res. 33, 651-658. Smith, R. G., and Daves, G. D. (1977). Biomed. Mass Spectrom. 4, 146-150. Snedden, W., and Parker, R. B. (1976). Biomed. Mass Spectrom. 3,295-298. Struck, R. F. (1974). Cancer Res. 34,2933-2935. Struck, R. F. (1976). Cancer Treat. Rep. 60, 317-319. Struck, R. F., and Hill, D. L. (1972). Proc. Am. Assoc. Cancer Res. 13, 50. Struck, R. F., Kirk, M. C., Witt, M. H., and Laster, W. R. (1975).Biomed. Mass Spectrom. 2, 46-52. Swofford, H., Wang, I., Martinsen, D., and Buttrill, S. (1976). 23rd Annu. Con$ Mass Spectrom. pp. 322-324. Takanashi, S., and Bachur, N. R. (1975).J . Pharmacol. Exp. Ther. 195, 41-49. Takanashi, S., and Bachur, N. R. (1976). Drug Metab. Dispos. 4,79-87. Tausch, H., and Stehlik, G. (1977). Chromatographia 10, 350-357. Taylor, P. A., and Springer, D. (1975).Finnigan Spectra 5,6-8. Thomas, R., and Lao, R. (1976). 24th Annu. Con$ Mass Spectrom. p. 125. Thomas, R., and Lao, R. (1977).Int. J . Sci. Enuiron. (in press). Tou, J. C., and Kallos, G. J. (1974).Anal. Chem. 46, 1866-1869. VanLierop, J. B., and Stek, W. (1976).J . Chromatogr. 128, 183-187. Versino, B., Knoppel, H., Groot, M., Peil, A., Poelman, J., Schauenburg, H., and Geiss, F. (1976).J . Clzromatogr. 122, 373-388. Vigevani, A,, Gioia, B., and Cassinelli, G. (1974).Carbohydr. Res. 32, 321-330. Walle, T. (1973).In “Polyamines in Normal and Neoplastic Growth” (D. Russell, ed.), pp. 355-365. Raven, New York. Wang, M., Imai, K., Yoshioka, M., and Tamura, Z. (1975).Clin. Chim. Acta 63, 13-19. Wang, M., Yoshioka, M., Imai, K., and Tamura, Z. (1975). Clin. Chim. Acta 63, 21-27. Watanabe, P. G., McGowan, G. R., and Gehring, P. J. (1976a).Toxicol. Appl. Pharmacol. 36, 339-362. Watanake, P. G., McGavan, G. R., and Gehring, P. J. (197613).Toxicol. Appl. Pharmacol. 37,49-53. Webb, R. G. (1975). I n “Environmental Applied Advances Institute Analysis,” EPA660/4-75-003 (Environ. Monit. Ser.). U.S. Environ. Protect. Agency, Corvallis, Oregon. Weinkam, R., Wen, J., Furst, D., and Levin, V. (1978).Clin. Chem. 24, 45-49.


267

Wishnok, J., and Tannenbaum, S. (1977).Anal. Chem. 49,715A-718A. Worzalla, J. F.,Johnson, B. M., Ramirez, G., and Bryan, G. T. (1973). Cancer Res. 34, 2669-2674. Worzalla, J. F., Kaiman, B. D., Johnson, B. M., Ramirez, G., and Br)ian, G. T. (1974). Cancer Res. 34, 2669-2674. Wu, A., Au, Jo, and SadBe, W. (1978).Cancer Res. 38, 210-214. Yang, S. K., and Gelboin, H. V. (1976). Cancer Res. 36, 4185-4189. Yang, S. R., McCourt, M. D., Roller, P. P., and Gelboin, H. V. (1976). Proc. Natl. Acad. S C ~ U.S.A. . 73, 2594-2598. Yergey, A., Risby, T., and Golomb, H. (1978). Biomed. Mass Spectrom. 5, 47-51. Yurachek, J. P., Clemena, G. G., and Harrison, W. W. (1969).Anal. Chem. 41,1666-1668. Zumoff, B., Fischman, J., Bradlow, H. L., and Hellman, L. (1975). Cancer Res. 35, 3365-3373.


ADVANCES I N CANCER RESEARCH, VOL. 27

MARROW TRANSPLANTATION IN THE TREATMENT OF ACUTE LEUKEMIA1s2 E. Donnall Thornas: C.Dean Buckner, Alexander F e f e ~Paul . ~ E. Neirnat~,~ and Rainer Storb The Fred Hutchinson Cancer Research Center, Seattle, Washington and the Department of Medicine, Division of Oncology, University of Washington School of Medicine, Seattle, Washington

I. Introduction .......................................................... 11. Patient Selection, Methods, and Summary of Clinical Results ............ 111. Analysis of Survival ................................................... IV. Nature of Recurrent Leukemia.. ....................................... V. Efforts to Prevent Leukemic Relapse ................................... VI. Graft versus Leukemia ................................................ VII. Transplantation in Remission .......................................... VIII. Conclusions .............................. References ...............................

269 270 271 273 275 276 277 278 278

I. Introduction

Almost 30 years have elapsed since the experiments of Jacobsen, Lorenz, and their colleagues demonstrated that mice could be protected against otherwise lethal irradiation b y shielding of the spleen or by intravenous infusion of marrow. By the mid 1950s, convincing evidence had accumulated indicating that the protective effect was due to the colonization of the recipient marrow b y donor cells. The implications of these laboratory studies for clinical application in man were 1957). A review of the early immediately apparent (Thomas et a?., studies on human marrow transplantation for the treatment of hematological disorders showed little or no evidence of success (Bortin, 1970). Work in the past decade, however, has produced a number of long-term survivors of marrow transplantation. The basic principles

* This investigation was supported by Grants CA 18029 and CA 15704, awarded by the National Cancer Institute, DHEW. Abbreviations used in this chapter are as follows: ALL, acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CML-BC, chronic myelogenous leukemia in blast crisis; CY, cyclophosphamide; GVHD, graft-versus-host disease; and TBI, total body irradiation. Dr. Thomas is a recipient of Research Career Award A1 02425 from the National Institute of Allergy and Infectious Diseases. Dr. Fefer is an American Cancer Society Professor of Clinical Oncology. Dr. Neiman is a Scholar of the Leukemia Society of America. 269 Copyright @ 1978 h y Academic Press, Inc. All rights of reproduction in any form rererved.

ISBN 0-12-006627-0

270

E. DONNALL THOMAS ET AL.

of marrow transplantation biology and its clinical application have been summarized in recent reviews (Thomas et al., 1975, 1977b). The present review is concerned with the clinical results of marrow transplantation in the treatment of patients with acute leukemia and the observations that have been made concerning the pathophysiology of acute leukemia. Chemotherapy of human leukemia is usually based on the premise that the therapeutic effect is due to the killing of leukemic cells. Since normal marrow cells are also killed by these drugs, the physician and the patient must tread a narrow path between therapeutic benefit and preservation of marrow function (Holland et al., 1976). I n theory, marrow transplantation should make it possible to kill the greatest number of leukemic cells with unusually large amounts of irradiation or chemotherapeutic drugs without regard to damage to normal marrow cells since the marrow transfused after therapy will restore normal marrow function. In addition, the transplanted marrow, since it is immunologically competent, might help in killing any residual leukemic cells by an immune attack directed at transplantation antigens or leukemia associated antigens on the leukemic cells. II. Patient Selection, Methods, and Summary of Clinical Results

Table I summarizes the results of marrow transplantation for 153 cases of acute leukemia carried out in Seattle between February of 1969 and October of 1975. The patients had either acute lymphoblastic leukemia (ALL), acute myelogenous leukemia (AML), or chronic myelogenous leukemia in blast crisis (CML-BC). Donors were either syngeneic (identical twin) or allogeneic (genotypically HLA matched sibling confirmed by nonreactivity in mixed leukocyte culture). All patients were given 1000 rad total body irradiation (TBI) from opposing 6oCosources shortly before the marrow infusion. Fourteen patients were given TBI only. Almost all the rest received cyclophosphamide [(CY), 60 mg/kg on each of 2 days] 4 and 3 days before TBI. Many patients in addition received other forms of antileukemic therapy either in an unsuccessful attempt to induce remission or to destroy leukemic cells prior to transplantation. Details of the preparation for engraftment and the transplant procedure have been presented elsewhere (Thomas and Storb, 1970; Thomas et al., 1975). Following engraftment, recipients of allogeneic marrow were given methotrexate for u p to 100 days postgrafting in an effort to reduce the incidence and severity of graft-versus-host disease (GVHD) (Storb et at., 1970). Recipients of syngeneic marrow were not given methotrexate but many

27 1

MARROW TRANSPLANTATION

TABLE I SEATTLEMARROWTRANSPLANT SUMMARY IN PATIENTS WITH ACUTE LEUKEMIAAS OF JUNE, 1977 Survival

Disease ALL” AML” CML-BCe a

Number

Donor Twinb HLA matched Twind HLA matched Twin HLA matched

15 52 14 58 5

10

Number with Number recurrent alive at leukemia 1 year 8 21 8 15 3 0

8 16

4 10 2 2

Number now alive

Longest living survivor (years)

5 9“ 3 6 1 2

7 7 5 5 1 1

Patients transplanted through October, 1975. Includes 2 patients with lymphosarcoma-leukemia. Includes 3 patients with recurrent leukemia after transplantation now in remission. Includes 1 patient with erythroleukemia evolving into AML. Patients transplanted through April, 1976.

received “immunotherapy” in the form of autologous killed leukemic cells plus added donor lymphocytes (Fefer et al., 1974). It should be emphasized that all of these patients had had extensive prior chemotherapy for leukemia, many had had prophylactic or therapeutic central nervous system irradiation, and almost all were in relapse at the time of transplantation (Thomas et al., 1977a). Ill. Analysis of Survival

Figure 1 shows a (Kaplan and Meier, 1958) plot of the survival of patients with ALL or AML who received either syngeneic (29 patients) or allogeneic (110 patients) marrow grafts. In the first 4 months after grafting, there was a significant mortality rate due to the complications of advanced illness at the time of grafting to recurrent leukemia, and, in the case of the allogeneic recipients, to complications of GVHD. Thereafter, in the first 2 years, the slope of the survival curve represents primarily loss of patients due to recurrent leukemia. The one death shown at 4 years is that of a patient with a central nervous system relapse 27 months after grafting. Most important, after approximately 2 years, the survival curve becomes almost flat. The flat survival curve extending from 2 years to 7 years constitutes an opera-

272

E. DONNALL THOMAS ET AL. 100

a

3

4

12.5

- 8

1 0

\ ALLOGENEIC

1

2

3

4 YEARS

5

6

7

FIG. 1. Kaplan-Meier product limit estimates of the percentage surviving for 110 patients as of June, 1977. Day 0 is the day of marrow transplantation. The circles indicate surviving patients.

tional definition of cure in these patients, particularly since no maintenance chemotherapy was given after marrow grafting (Thomas et al., 1977c; Fefer et al., 1977). An analysis of a wide range of factors that might predict whether or not a patient will survive for the first 100 days after marrow grafting is underway. As yet, in this group of endstage patients, critical factors determining survival have not been identified. The most significant factor is an overall assessment of the general clinical condition of the patient at the time of grafting. Patients in good clinical condition in early relapse have a much better chance of survival than do patients in advanced relapse with lack of platelets and granulocytes and associated problems of infection and bleeding (Thomas et al., 1977a). Figure 2 shows a plot on a semilogarithmic scale of the probability that a patient will be in remission on a given day provided that he has lived to reach that particular day. This method of presenting the data permits the elimination of deaths from competing causes (such as GVHD and infection), The slopes of the curves thus reflect the rate of leukemic relapse. The results show that the rate of leukemic relapse is quite constant during the first lY2 years postgrafting. However, the curves are flat after approximately 2 years indicating no further relapses, The curve for the syngeneic graft recipients is not statistically significantly different from the curve for recipients of allogeneic grafts. The hypothesis that syngeneic marrow might be just as susceptible to a leukemogenic factor as the patient's own original marrow and the

273


ALLOCENEIC

a

I/12.5 0

1

2

3 4 YEARS

5

6

1

FIG.2. Kaplan-Meier product limit estimates of probability of being in remission. The circles indicate living patients in remission. The longest time to relapse with a syngeneic graft was I4 months and with an allogeneic graft 27 months.

related hypothesis that an allogeneic marrow graft might have a lesser susceptibility are not supported by these data. Rather, it would appear that the apparent eradication of leukemia in some patients, and the failure of eradication in others, is a reflection of the sensitivity of the leukemic cell population to the chemotherapy and TBI used in preparation for engraftment. At the present time there is no information to indicate whether sensitivity or resistance is a characteristic of the original leukemic cell population or is acquired during the course of therapy.

IV. Nature of Recurrent Leukemia

With a syngeneic marrow graft, donor and recipient are identical with regard to all blood genetic markers. Accordingly it is impossible to tell whether the recurrent leukemia represents regrowth from an original clone or represents reinduction. With an allogeneic graft, especially with a donor of opposite sex, it is possible to identify the origin of both the normal and abnormal cells following marrow grafting. In patients given 1000 rad TBI, the normal marrow cells seen after grafting have invariably been entirely of the donor type. During our early experience with marrow transplantation we observed a mechanism of leukemic relapse which was completely unanticipated, and which has significant implications for the etiology and pathogenesis of this disease (Fialkow et al., 1971;Thomas et al., 1972). Among six patients with ALL who were prepared with 1000 rad TBI alone, 2 were females who redeveloped active leukemia 62 and 135

274


days following transplantation. Both of these individuals received allogeneic marrow &om histocompatible male siblings within 24 hours of irradiation. The normal marrow cells present after grafting and subsequently the leukemic cells after relapse were shown to have metaphases containing a Y chromosome and, in the second case, fluorescent staining of the leukemic cells showed the presence of a Y body. Apparently the leukemic cell population represented a malignant transformation of lymphoid cells in the donor marrow. Alternatively, it is possible that such an event represented a transfer of the Y chromosome to a surviving remnant of the original leukemic cell population following a cell fusion event. The occurrence of such an event in two of six comparable individuals, however, requires one to propose a fairly specific and complex series of steps for which no experimental evidence presently exists. Experimental model systems do suggest other possible explanations involving transformation of donor stem cells, one of which would be the existence of a leukemogenic virus resident in the host, and perhaps induced by the TBI. Unfortunately, there is at present no unambiguous evidence for the existence of infectious leukemogenic viruses in man, inducible or otherwise. While these apparent marrow graft transformation events may provide encouragement to those pursuing an infectious viral theory of leukemogenesis, recent research in tumor virology suggests alternative mechanisms. Several investigators have demonstrated that DNA proviruses or oncogenic retroviruses can transform target cells when such cells are exposed to free DNA from virus-infected cells (Hill and Hillova, 1972; Cooper and Temin, 1974). Further, this transforming potential of viral “oncogene”-containing DNA may be preserved in the absence of overt viral replication in the cells bearing the infectious provirus. It is intriguing, in this connection, that marrow graft transformation has not occurred in the 11 episodes of leukemic relapse in patients with donor of opposite sex that we have had the opportunity to study since the two original episodes. It may be significant that the protocols for preparation of engraftment in the later patients all incorporate significant cytoreduction prior to irradiation. Thus, the level and duration of exposure of newly infused marrow stem cells to material released from the dying leukemic cell mass may have been substantially reduced. The transfer of genes by free DNA is called “transfection,” and has, so far, only been demonstrated for viral genes. It remains possible that somatic cellular genes responsible for malignant transformation might have been transfected as well. One attractive aspect of the transfection theory is that, unlike the competing ideas mentioned above, it may prove susceptible to direct experimental testing.


275

V. Efforts to Prevent Leukemic Relapse

In the initial group of 10 patients in the Seattle series prepared with TBI only, 4 patients with AML died too early to be evaluated for leukemic relapse while 5 of 6 patients with ALL suffered a relapse of leukemia. Therefore, in an effort to destroy more leukemic cells, high dose CY was added to the regimen before irradiation. In a group of 43 patients given CY plus TBI only, 14 relapsed. Other antileukemic drugs have been given before CY and TBI in an effort to reduce the leukemic cell population further. Depending on past drug exposure, 31 patients were given a variety of cytoreductive regimens, usually including cytosine arabinoside and rubidomycin prior to CY and TBI, and 11 relapsed. Nineteen patients were given 8-14 mg/kg 1,3bis(2-chloroethy1)-l-nitrosourea prior to CY and TBI. Three have relapsed but only 1 is a long-term survivor. In a more recent unpublished study, 7 patients were given a 5-day continuous infusion of cytosine arabinoside, 600 mg/m2/day, before CY and TBI. One of 7 patients survived over 100 days and he relapsed 1 year after transplantation. In an effort to avoid the problems of GVHD, Math6 et al. (1970) treated 11leukemic patients with antilymphocytic serum before infusing marrow from related donors, only 4 of whom were HLA identical with the recipient. Although GVHD was not observed, the grafts were transient and 10 surviving patients showed recurrent leukemia. This study points out the need for some form of leukemic cell cytoreduction in the preparative regimen. Other investigators have utilized vigorous chemotherapy instead of TBI before transplantation. In studies of a series of acute leukemia patients, Santos et al. (1976) used 50 mg/kg of CY given on each of 4 days in 25 patients and Graw et al. (1972)used 45 mg/kg of CY x 4 in 9 patients. All patients who survived developed recurrent leukemia. Graw et al. (1974) developed a BACT (bis-chloroethyl-nitrosourea, cytosine arabinoside, CY, 6-thioguanine) regimen. Ten patients were treated with this regimen, which proved to be very toxic with 5 deaths before day 35 (Buja et al., 1976). Four patients died with recurrent leukemia at 30-437 days postgrafting. However, one patient who had endstage AML is alive and well on no maintenance chemotherapy now 5 years after transplantation (Bleyer et al., 1975). The UCLA marrow transplant group has developed a regimen known as SCARI (6-thioguanine, cytosine arabinoside, rubidomycin, CY, TBI) which employs high dose chemotherapy prior to CY and TBI. Fifteen patients were treated with SCARI (UCLA Bone Marrow Transplantation Group, 1977). The regimen proved to b e quite toxic

276


especially for adult patients. However, the preliminary results show 4 patients living at approximately 1 year after grafting and only 2 patients have suffered a relapse of leukemia. From these observations it appears that none of the regimens designed to increase the killing of leukemic cells has resulted in any overall improvement in survival in comparison to CY and TBI. Efforts to kill more leukemic cells in these endstage patients in relapse carry the price of increased toxicity. Since most of these patients have had a variety of chemotherapeutic agents before coming to marrow transplantation, it is difficult to find an effective drug to include in the regimen that has not already been employed. Dimethyl myleran is a drug that may show some promise in this regard (Floersheim, 1969; Buckner et al., 1975; Einstein et al., 1976; Kolb et al., 1974). It is an effective antileukemic agent which was abandoned because of excessive marrow toxicity (Bierman et al., 1958; Clifford et al., 1964). With the marrow transplantation regimen, excessive marrow toxicity can be ignored and preliminary studies to evaluate dimethyl myleran given before CY plus TBI are underway. VI. Graft versus Leukemia

Table I1 shows the incidence and seventy of GVHD in the 100 patients with acute leukemia transplanted after CY and TBI. One-half of the patients had significant GVHD, a continuing major problem in marrow transplantation biology which has been reviewed elsewhere (Thomas et al., 1975). Of particular interest here is the possible antileukemic effect of GVHD. Barnes et al. (1956) pointed out that a small residual population of leukemic cells might be killed off by a reaction of the engrafted marrow against the leukemic cells. Subsequently Math6 et al. (1965) enlarged upon this concept and coined

GRADEOF GVHD

AND

TABLE I1 RELATION TO RELAPSE OF LEUKEMIA"

Grade of GVHDb

Number of patients without relapse

Number of patients with relapse

0 I 11-IV

7 11 44

12 13 6

Ninety-three of 100 patients with acute leukemia who lived long enough to be evaluated after CY, TBI, and allogeneic marrow grafting. Grade 0, no CVHD; I, skin rash only; 11-IV GVHD in 2 or more target organs.


277

the term “adoptive immunotherapy.” More recently Bortin (1974) and Fefer et al. (1976) have presented evidence in support of a graftversus-leukemia effect in murine systems. Table I1 shows the incidence of leukemic relapse in relation to GVHD in our patients given allogeneic grafts. It is apparent that some patients who had no identifiable GVHD have not had a relapse of leukemia. Many patients with severe GVHD do not survive long enough to have an opportunity to suffer a leukemic relapse. However, as indicated in Table I1,6 patients have suffered a relapse of leukemia in the face of severe active GVHD. Certainly, in these recipients of marrow from a donor matched at the major histocompatibility complex, if a graft-verus-leukemia reaction occurred it was not sufficient to prevent leukemic relapse. Figure 2 demonstrates a comparison of the leukemic relapse rate in patients given syngeneic marrow as compared to patients given allogeneic marrow and does not show a significant difference between these two groups. These data do not suggest that an allogeneic marrow graft had an antileukemic effect. This conclusion is only tentative, however, in view of the increased death rate from other causes associated with GVHD in the recipients of allogeneic marrow and of the unknown role of immunotherapy in the recipients of syngeneic marrow. VII. Transplantation in Remission

Since marrow transplantation appears to be curative for some patients with endstage acute leukemia, it is now ethically acceptable to consider transplantation earlier in the course of the disease. Marrow transplantation for patients in remission who are known to have a poor prognosis would entail the following advantages : ( 1) treatment before the leukemic cell population becomes resistant to therapeutic modalities, (2)treatment at a time when the body burden of leukemic cells is minimal, and (3) treatment when the patient is in excellent clinical condition and, therefore, better able to tolerate the antileukemic and transplantation regimes. In 1976 the Seattle marrow transplant team began to perform allogeneic marrow grafts in patients in remission after preparation with CY plus TBI. Almost all of the patients with ALL have had more than 1 relapse either in the marrow or in extramedullary sites. The majority of the patients with AML have been in the first remission. As expected, these patients tolerated the transplant regimen very well. The support requirements for platelet and white blood cell transfusions and time in the hospital were strik-

278

E. DONNALL THOMAS ET AL

ingly reduced. The earliest death was on day 62 following interstitial pneumonia due to cytomegalovirus. It is clearly too early to evaluate the results of this continuing study since the initial patients are now just beyond 1year after grafting. The crucial question of whether or not the increased risks of marrow grafting and GVHD and its complications can be balanced b y long-term remission without continued chemotherapy, and even “cure” in some patients, will take several years to answer. VIII. Conclusions

Marrow transplantation has now become established as a therapeutic option of definite benefit to some patients with acute leukemia. For the patient with an identical twin donor, marrow transplantation appears to be the treatment of choice since the complete remission rate even in adults with acute leukemia is approximately 90%. Some progress is being made toward resolution of the problems of transplantation biology that occur even though the sibling donor-recipient pairs are matched at the major histocompatibility complex. Efforts to reduce the incidence of recurrent leukemia by more intensive antileukemic therapy or b y transplantation in remission are under way. Marrow transplantation makes it possible to ignore the complications of marrow toxicity which usually limit the chemotherapy of leukemia. This will permit the exploration of other drug regimens and of drugs considered to be too toxic for ordinary use. Marrow transplantation remains a complex undertaking that requires a skilled team of physicians of many disciplines, nurses, and technologists. Pending development of better methods of specific treatment or prevention of hematologic malignancies, marrow transplantation is likely to be undertaken for an ever increasing number of patients in the coming decade.

REFERENCES Barnes, D. W. H., Corp, M. J., Loutit, J. F., and Neal, F. E. (1956).Br. A4ed.J.2,626-627. Bierman, H. R., Kelly, K. H., Knudson, A. G., Jr., Maekawa, T., and Timmis, G. M. (1958).Ann.N . Y.Acad. Sci. 68, 1211-1222. Bleyer, W. A., Blaese, R. M., Bujak, J. S., Herzig, G. P., and Graw, R. G., Jr. (1975).Blood 45. 171-181. Bortin, M. M. (1970). Transplantation 9, 571-587. Bortin, M. M. (1974). Clin. Immunobiol. 2, 287-306. Buckner, C. D., Ilillingham, L. A,, Giddens, W. E., Jr., and Thomas, E. D. (1975).Erp. Hematol. (Copenhagen) 3, 275-288. Buja, L. M., Ferrans, V. J., and Graw, R. G., Jr. (1976). Hum. Pathol. 7, 17-45.


279

Clifford, P., Clift, R. A., Kahn, A. C., and Temmis, G. M. (1964). Br. J. Cancer 13, 435-438. Cooper, G. M., andTemin, H. M. (1974).J.Virol. 14, 1132-1141. Einstein, A. B., Jr., Cheever, M. A., and Fefer, A. (1976).J. Natl. Cancer Inst. 56, 609-613. Fefer, A., Einstein, A. B., Thomas, E. D., Buckner, C. D., Clift, R. A., Glucksberg, H., Neiman, P. E., and Storb, R. (1974).N . Engl. J. Med. 290, 1389-1393. Fefer, A., Einstein, A. B., Jr., and Cheever, M. A. (1976).Ann. N.Y. Acad. Sci. 277, 492-504. Fefer, A., Buckner, C. D., Thomas, E. D., Cheever, M. A., Clift, R. A., Glucksberg, H., Neiman, P. E., and Storb, R. (1977). N. E n g l . J. Med. 297, 146-148. Fialkow, P. J., Thomas, E. D., Bryant, J. I., and Neiman, P. E. ( 197l).Lancet 1,251-255. Floersheim, G. L. (1969). Lancet 1, 228-233. Craw, R. G., Jr., Yankee, R. A., Rogentine, G. N., Leventhal, B. G., Herzig, G. P., Halterman, R. H., Merritt, C. B., McGinniss, M. H., Krueger, G. R. D., Whang-Peng, J., Carolla, R. L., Gullion, D. S., Lippman, M. E., Gralnick, H. R., Berard, C. W., Terasaki, P. I., and Henderson, E. S. (1972).Transplantation 14, 79-90. Craw, R. G., Jr., Lohrmann, H.-P., Bull, M . I., Decter, J., Herzig, G. P., Bull, J. M., Leventhal, B. G., Yankee, R. A., Herzig, R. H., Krueger, G. R. F., Bleyer, W. A,, Buja, hl. L., McGinniss, M. H., Alter, H. J., Whang-Peng, J., Gralnick, H . R., Kirkpatrick, C. H., and Henderson, E. S. (1974).Transplont. Proc. 6, ,349-354. Hill, M., and Hillova, J. (1972).Nature (London), New Biol. 237,35-39. Holland, J. F., Glidewell, O., Ellison, R. R., Corey, R. W., Schwartz, J., Wallace, H. J., Hoagland, H. C., Wiernik, P., Rai, K., Bekesi, J. G., and Cuttner, J. (1976). Arch. Intern. Med. 136, 1377-1381. Kaplan, E. L., and Meier, P. (1958).J.A m . Stat. Assoc. 53, 457-481. Kolb, H. J., Storb, R., Weiden, P. L., Ochs, H. D., Kolb, H., Graham, T. C., Floersheim, G. L., and Thomas, E. D. (1974). Biomedicine 20, 341-351. MathC, G., Amiel, J. L., Schwanenberg, L., Cattan, A., and Schneider, M. (1965). Cancer Res. 25, 1525-1531. MathC, G., Amiel, J. L., Schwarzenberg, L., Choay, J., Trolard, P., Schneider, M., Hayat, M.. Schlumberger, J. R., and Jasmin, C. (1970). Br. Med.1. 2, 131-136. Santos, G. W., Sensenbrenner, L. L., Anderson, P. N., Burke, P. J., Klein, D. L., Slavin, R. E., Schacter, B., and Borgaonkar, D. S . (1976). Transplant. Proc. 8, 607-610. Storb, R., Epstein, R. B., Graham, T. C., and Thomas, E. D. (1970). Transplantation 9, 240-246. Thomas, E. D., and Storb, R. (1970). Blood 36, 507-515. Thomas, E. D., Lochte, H. L., Jr., Lu, W. C., and Ferrebee, J. W. ( 1957).N. EngZ.J. Med. 257,491-496. Thomas, E. D., Bryant, J. I., Buckner, C. D., Clift, R. A., Fefer, A., Johnson, F. L., Neiman, P., Ramberg, R. E., and Storb, R. (1972). Lancet 1, 1310-1313. Thomas, E. D., Storb, R., Clift, R. A., Fefer, A., Johnson, F. L., Neiman, P. E., Lerner, K. G., Glucksberg, H., and Buckner, C. D. (1975). N. Engl. J . Med. 292, 832-843 and 895-902. Thomas, E. D., Buckner, C. D., Banaji, M., Clift, R. A., Fefer, A., Floumoy, N., Goodell, B. W., Hickman, R. O., Lerner, K. G., Neiman, P. E., Sale, G . E., Sanders, J. E., Singer, J., Stevens, M., Storb, R., and Weiden, P. L. (1977a).Blood 49,511-533. Thomas, E. D., Fefer, A., Buckner, C. D., and Storb, R. (1977b). Blood 49,671-681. Thomas, E. D., Flournoy, N., Buckner, C. D., Clift, R. A., Fefer, A,, Neiman, P. E., and Storb, R. ( 1 9 7 7 ~ )L. e d . Res. 1,67-70. UCLA Bone Marrow Transplantation Group. (1977).Ann. Intern. Med. 86, 155-161.


ADVANCES IN CANCER RESEARCH, VOL. 27

SUSCEPTIBILITY OF HUMAN POPULATION GROUPS TO COLON CANCER Martin Lipkin Memorial Sloan-Kettering Cancer Center, New York. New York

I. Introduction .......................................................... 11. The Role of Environment in Increasing the Susceptibility of Individuals to Colon Cancer. .................................. 111. Inherited Diseases That Increase Susceptibility to Colon Cancer.. . . . . . . . Inherited Adenomatosis and Related Disorders ......... IV. Proliferative Abnormalities and Susceptibility to Colon Cancer .......... V. Newer Immunologic Studies .......................................... VI. Nuclear Protein and Enzyme Alterations ............................... VII. Studies of Cutaneous Cells ............................................ VIII. Examination of Fecal Contents ........................ IX. Conclusion ........................................................... References ..................................

281 287 293 296 296 299 301

I. Introduction

Colorectal cancer poses a major problem in the United States and other countries. It is currently responsible for a high proportion of the malignant neoplasms found in the United States. Recent figures have indicated 99,000 new cases in the United States in 1975 and 49,000 resultant deaths (Silverberg and Holleb, 1975).The relative frequency of successful treatment of individuals varies with the stage at which the disease is detected; a higher rate of success is related to detection of disease at an earlier stage (Gilbertsen, 1974). Earlier detection of disease in the population aggregates having increased risk (listed in Table I) could reduce the burden denoted by the above figures. Recent studies indicate that improvements have been made in the identification of factors that are associated with the predisposition and development of colorectal cancer in several of these population groups. The early identification of abnormalities in cell development and related parameters, in individuals having high susceptibility to colon cancer can be carried out with greater precision than was formerly possible. Recent approaches to the identification of these abnormalities are enumerated in Table 11. This chapter will discuss current findings in this area, their contribution to our under28 1 Copyright Q 1978 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 012-006627-0

282

MARTIN LIPKIN

TABLE I POPULATION GROUPS AT INCREASED RISK FOR COLON CANCER Familial Polyposis Syndromes Inherited adenomatosis of colon and rectum Gardner’s syndrome (Oldfield) Turcots syndrome Diffuse gastrointestinal polyposis Colonic adenomas Previous colon, breast, endometrial, bladder cancer Site-specific colon cancer Cancer family syndrome Inflammatory bowel disease Common variable immune deficiency Residence in geographic areas having high frequencies of colon cancer

standing of the development of neoplasia in man, and their potential role in early detection and prevention of the disease. I I . The Role of Environment in Increasing the Susceptibility of Individuals to Colon Cancer

It is now well established that environment plays a role in the development of colorectal cancer. Numerous studies have shown a strong correlation between geographic, economic, and dietary exposure and colon cancer, and programs of surveillance are desirable for high-risk groups. Epidemiological studies of populations at high and low risk for colon cancer have shown striking differences in the composition of the diet in these population groups. In industrialized countries including those of northwest Europe and North America a great deal of animal TABLE I1 RECENT L4PPROACHESTO IDENTIFICATION OF HIGH RISK GROUPS Prior to clinically detectable lesions Study of environmental factors Proliferative abnormalities Immunologic measurements Study of cutaneous cells Nuclear protein and enzyme alterations Examination of fecal contents After clinically detectable lesions Hemoccult, radiologic, and endoscopic studies

SUSCEPTIBILITY TO COLON CANCER

283

fat, protein, and refined carbohydrates are consumed; in these geographic regions the incidence of colon cancer is much higher than in the developing countries of Africa, South America (except meat-eating Argentina and Uruguay), rural India, and Japan, where much less meat is consumed and the diet is higher in vegetable fiber (Berg, 1973; Doll, 1969; Haenszel and Correa, 1971; Haenszel et al., 1973;Wynder and Shigematsu, 1967). These variations in incidence do not appear to be related to genetic differences, because migrant groups tend to assume colon cancer incidence rates of their adopted countries (Haenszel and Kurihara, 1968). In studies of Japanese Issei (first generation) and Nisei (subsequent generation) migrants in Hawaii, a higher incidence of large bowel cancer was observed in individuals who no longer continued the practice of eating at least one Japanese style meal daily (Haenszel et al., 1973). Colon cancer patients appeared to consume more beef and legumes than did controls. A rise in the consumption of meat accounted for the major difference in dietary habit between residents of Japan and Hawaii, and the increase in beef consumption paralleled the higher risk of bowel cancer among Japanese migrants. A correlation between the daily consumption of meat and the incidence of colon cancer also has been noted in individuals from many countries, suggesting an etiological role (Armstrong and Doll, 1975; Wynder and Shigematsu 1967). Observations of this type have led investigators to postulate that high animal protein and fat, characteristic of high-risk populations, are responsible for colon cancer. In one study of 28 countries positive correlations were found between the incidence of colon cancer and the amount of meat the various populations consumed (Gregor et al., 1969). On the basis of findings of this type, it has been postulated that the high incidence of colon cancer is due to the nature of the intestinal flora, which might possibly synthesize carcinogenic agents from food, and intestinal secretions such as bile acids. Accordingly, diet not only would be expected to influence the composition of the intestinal flora, but also the quantity of substrates available for the production of carcinogens. Burkett and his colleagues have also postulated that fiber content of diet may be associated with the development of colon cancer (Burkitt, 1975). Studies correlating the selenium levels with incidence of colon cancer are of recent interest (Andrews et UI?., 1968; Jacobs et al., 1976; Jansson et al., 1976).Previous epidemiologic studies had shown a high correlation between geographic regions having deficiency of selenium and increased incidence of colon cancer (Andrews et al., 1968; Jacobs et al., 1976; Jansson et al., 1976). Experimental carcinogenesis also has

284

MARTIN LIPMN

been inhibited by selenium administration (Harr et al., 1976; Schrauzer and Ishmael, 1974). Recently, addition of 4 ppm of selenium in drinking water reduced the number of rats developing dimethylhydrazine-induced colon tumors, and the total number of methylazoxymethanol-induced rat tumors (Jacobs et al., 1976). It is also true that other factors in these environments, for example the prevalence of infectious disease as well as other chronic illnesses, vary greatly among these populations. In addition, it is known that other factors not related to food also change simultaneously with dietary habits, in relation to industrial and economic development. The epidemiological evidence available can therefore suggest further studies that should be carried out, but does not supply definitive proof that any given dietary factors are causally related to variations in the incidence of colon cancer. In attempting to explore these leads to elucidate mechanisms related to pathogenesis, studies have been carried out to examine the fecal flora obtained from individuals in different parts of the world. Fecal samples were obtained from individuals residing in England, Scotland, and the United States, regions having high incidence of colon cancer, and from Uganda, India, and Japan where low incidence of colon cancer occurs. Although the same broad classifications of bacteria were found in feces from all the populations studied, differences in relative numbers of several of the bacterial groups were observed. The British and American subjects had more gram-negative anaerobes than did the Ugandans, Indians, or Japanese; in contrast, the latter had larger amounts of aerobic bacteria. The ratio of anaerobes to aerobes was therefore higher in individuals consuming a Western style diet than in those consuming largely vegetarian diets (Hill et al., 1971). The concentrations of acid and neutral steroids in the feces of those subjects also differed when individuals on high and low meat diets were compared. Fecal specimens from British and Americans on high meat diets contained higher amounts of steroids than feces of Ugandans, Indians, and Japanese, whose diets contained little or no animal fat and protein. Neutral steroid amounts were low in the feces of Ugandans and Indians, intermediate in the feces of Japanese, and high in the feces of British and Americans. I n addition, microbial conversion products of cholesterol, coprostanol and coprostanone, contributed a smaller amount to the total neutral steroid content of the feces of the Ugandans, Indians, and Japanese than the feces of the Western group. Interesting differences also were noted in the fecal composition ofacid steroids. The acid steroid concentrations were greater in the feces of British and Americans than in the Ugandans, Indians, and Japanese. The extent of con-


285

version of acid steroids also appeared to be higher in the British and Americans than in other groups. Further observations were made in which daily fecal excretion of cholesterol metabolites was believed to be grater in Americans who consume a diet containing meat, than in Americans on a meatless diet (Reddy and Wynder, 1973). Despite these observations, other studies have suggested no significant differences in bacterial counts or species isolated from the feces of volunteers on high meat and meatless diets. When individuals at high risk were compared with low-risk populations, no major differences in varieties of intestinal bacteria were found. Minor quantitative variations were present, but characteristic organisms in high- or lowrisk groups were not described, and organisms present in individuals on high-risk diets were also found in individuals on low-risk diets. It was suggested that taxonomic grouping of bacteria is not important in analyzing the effects of diet on intestinal flora, but that the effect of altered diet on bacterial metabolic activity might be of interest, and in fact more useful in trying to assess significant factors present (Feingold et al., 1974; Moore and Holdeman, 1975). High-risk populations had a more heterogenous and variable microflora. Alterations in the flora were also believed to be effected by situations which produced anger or stress in the host. Similarly, concentrations of acid or neutral steroids and their metabolites were minimally effected during high meat compared to meatless diets (Hentges et al., 1976). However, some individuals excreted significantly higher amounts of cholesterol than the others. It has thus been suggested that ingestion of high animal protein itself does not result in major change in bacterial and chemical composition of feces. On the other hand, fat content may be more important since acid and neutral steroid concentrations were observed to decline when fat megt in the diet was replaced by lean meat (Hill, 1971; Reddy and Wynder, 1973). Secondary bile acids in the feces were also higher during fat meat ingestion when the concentrations of total bile acids were high, leading to the possibility that secondary bile acids are products of the dehydroxylation of primary fecal bile acids by intestinal bacteria. In further support of the above, high fat, high meat mixed Western diet, and nonmeat diet were compared in human volunteers for steroid content of feces (Reddy et al., 1975a), while protein content of both diets was similar. Findings indicated that total anaerobic microflora count, and fecal excretion of secondary bile acids and cholesterol metabolites, were greater during consumption of the mixed Western diet than the nonmeat diet, supporting a role of dietary fat on composition of intestinal flora and level of steroid conversion products in feces. Thus, individuals with colon cancer have been reported to have

286

MARTIN LIPKIN

higher amounts of steroids in their feces than controls without cancer, and steroid conversion products such as deoxycholic acid, lithocholic acid, and cholesterol metabolities. Of further interest is the observation that the activity of fecal 7 a-dehydroxylase is higher in patients with colon cancer compared to controls, in association with conversion of cholic and chenodeoxycholic acids to deoxycholic and lithocholic acids. In individuals with cancer of the large bowel a high frequency were reported to have concentrations of bile acids in their feces greater than individuals with other diseases Hill et al., 1975; Reddy et al., 1975a). The colon cancer patients also had greater amounts of acid steroids in the form of secondary bile acids than did the other patients. These findings continue to support the view that an association of fecal steroids and the production of colon cancer is significant. In addition, it has been reported (Wilkins and Hackman, 1974) that normal North Americans shows two patterns of neutral sterol conversion as measured by fecal analysis: “High converters” appear to have a stable pattern of extensive conversion of cholesterol, sitosterol, and campesterol by the intestinal flora to degradation products, while “low converters” have little or no such conversign. This finding further points to the possibility that genetic, as well as environmental factors influence neutral sterol conversion. Other recent work has indicated that bile salt excretion can be affected differently by various dietary sources, and feeding wheat bran decreased the cholesterol saturation of bile, apparently b y increasing hepatic synthesis of chenodeoxycholic acid (Pomare et al., 1974). Related animal studies have been of interest in that rats fed high fat diets have been reported to be more susceptible to colon tumor induction by 1,2-dimethylhydrazine (DMH) (and more recently by azoxymethane) than animals fed a diet containing a normal amount of fat (Reddy et al., 1975b).Fecal excretion of acid and neutral steroids was higher in animals fed high fat diets than in animals on low fat diets, similar to human populations on high and low fat diets. Bile acids have a potentiating effect in inducing colon carcinoma in laboratory animals, and are believed to be promoting agents. The development of colonic tumors in rats exposed to the carcinogen N-methyl-N’-nitro-N-nitrosoguanidine (MNNG) was increased by instilling lithocholic or taurodeoxycholic acids intrarectally (Narisawaet al., 1974); the carcinogenic effect of azoxymethane in rats was increased by increasing the concentration of bile reaching the colon, by feeding cholestyramine and by diverting the bile flow to the lower section ofthe small intestine (Chomchai et al., 1974; Nigro et al., 1973).Cholestyramine has been reported to increase the frequency of intestinal neoplasms induced with 1,2dimethylhydrazine in germfree rats (Asanoet al., 1975). It also has been


287

reported that vitamin A deficient rats were more susceptible to dimethylhydrazine induced colon cancer (Rodgers and Nernberne, 1973; Rodgers et al., 1973). The possibility that significant bacterial conversion of bile acids and cholesterol may take place has been referred to. It is believed that the fecal concentrations of total acid steroids, deoxycholic acid and fecal bacteria containing the enzymes 7-hydroxycholanoyldehydroxylase and 3-oxo-cholanoyl A 4 dehydrogenase are of importance. The enzymes convert primary bile acids to secondary, and produce double bond formation on the bile acid nucleus. Thus, the possibility has been considered that both unsaturated and saturated bile acids contribute to the etiology of colon cancer, fulfilling roles of cocarcinogen and carcinogen (Hill, 1975). On the other hand, it was postulated (Mastromarino et al., 1976)that cholesterol metabolites may be important in the pathogenesis of colon cancer since fecal microbial 7-hydroxycholanoyl dehydroxylase and cholesterol dehydrogenase activities were noted to be higher in cancer patients compared to controls. Anaerobic intestinal bacteria contain enzymes that might induce the production of secondary bile acids and cholesterol metabolites. Thus, it is believed by some investigators that diets having high fat content, and perhaps a corollary low fiber content typical of the Western diet, are key factors in increasing colon cancer risk. These diets lead to changes in the lumenal contents of the large bowel as noted above. In attempting to understand the pathogenesis of colon cancer, key questions now concern the degree to which high-risk subgroups within the general population respond to dietary changes, and the degree and mode of interaction that may occur between dietary elements and colonic cells genotypically predisposed to neoplasia. Ill. Inherited Diseases That increase Susceptibility to Colon Cancer

As noted above, current findings suggest that the development of colorectal cancer is largely influenced by environmental factors, although attempts to identify specific elements, and in particular their mode of action, have not been definitive. At present, inheritance is believed to have a minor role in the genesis of colorectal cancer. This view is based largely on the frequency of identification of the clinically definable dominant inherited syndromes leading to colon cancer. Recent findings, however, have indicated again that familial associations in colon cancer are higher than in control groups, suggesting that inherited factors could have a greater role in the genesis of colorectal cancer than has been generally believed. Thus, it was recently noted

288

MARTIN LIPKIN

that there was a significant increase in the number of deaths due to bowel cancer among first degree relatives of index cases compared with the expected incidence. Early age of onset, the presence of adenomas or other carcinomas in the operation specimens, and a history of previous carcinoma were found to be associated with an increased risk that the index case would have a positive family history (Lovett, 1974). It had been recognized that a significant percentage of colorectal cancer patients have family members who also have colon cancer, and that relatives of colorectal cancer patients were at an increased risk compared to the general population (Moertel et al., 1958). Familial associations could be influenced by genetic as well as environmental factors, and inherited predispositions could influence the development of colon cancer on the basis of cellular and related physiological abnormalities. Patients under age 40 who develop colorectal cancer have been reported more likely to have a family history of colon cancer than those over age 40 (Wynder and Shigematsu, 1967). In relatives of patients with multiple colonic malignancies the age of onset of colonic cancer has been reported to be significantly earlier than in the general population (Moertel et a1., 1958). The probability of developing a new colon cancer is increased in individuals who have had prior colon cancer or polyps, or female genital cancers, breast cancer, or bladder cancer. The degree to which inherited or environmental factors influence this predisposition remains to be determined. In individuals with previous colon cancer, the probability of developing a second or metachronous lesion is about three times higher than in the general population (Morson and Bussey, 1970). When the initial cancer was located in the cecum, a second primary colon cancer occurred most frequently. Associations of colonic and other neoplasms have similarly suggested that inherited factors are important. Female genital cancers, primarily endometrial, are twice as frequent as expected in colon cancer patients. Associations of breast cancer, endometrial, and ovarian cancer have been noted in women with colon cancer (Fraumeni, 1973; Lynch, 1967). A disease entity of hereditary adenocarcinomatosis has also been noted, an autosomal dominant with 90% penetrance (Anderson, 1970), and cancer families showing a striking incidence of primary malignancies at multiple anatomic sites including the colon, with an early age of onset have been identified in families highly predisposed to cancer (Lynch and Krush, 1967). Cancers that show patterns of inheritance that appear to be dominant can show tumor specificity, oc-


289

curring in the same tissue and in the same organs in family members, earlier age of onset than in the general population, and a tendency to develop at multiple sites within an organ (Anderson, 1970). The concept of cancer families has been broadened recently to include neoplasms of different types. The neoplasms in these families, which appear to b e influenced by genetic predisposition, affect diverse organs especially the colon and endometrium. They tend to develop earlier in life than usual and may occur separately or as multiple cancers in family members. Since adenocarcinomas of the colon and reproductive sites are known to coexist excessively in studies of multiple primary cancers, the familial syndrome may represent a scattering over the family tree of tumors that share etiologic influences (Fraumeni, 1973). Efforts to indicate a single-gene difference in colonic carcinomas with a genetic basis have largely been unsuccessful, except for the inherited disease adenomatosis of the colon and rectum (familial polyposis) (Alm and Licznerski, 1973; Gardner, 1962; Morson and Bussey, 1970). Ulcerative colitis, solitary polyps, and regional enteritis can also occur in familial aggregates and can develop into carcinomas. Although these syndromes suggest a multifactorial basis for possible genetic determination, the involvement of genetic mechanisms is still not clear. INHERITED

ADENOMATOSIS AND RELATEDDISORDERS

Several varieties of precancerous colorectal disease of genetic origin involving polyp formation have been well described. Znherited adenomatosis of the colon and rectum (A.C.R.; familial polyposis) is associated with innumerable colonic adenomas; it has been possible to estimate population frequency, relative fitness, and mutation rate. This is one of the few genetic disorders associated with malignancy offering such a quantity of these types of data, which enabled a fairly accurate description of the disease when clinically apparent. Various studies have estimated the disease frequency in the population based on clinically observable manifestations. These estimated frequencies range from 1 in 8300 in Michigan and 1 in 7150 in Kentucky, to 1 in 23,750 in St. Mark’s Hospital in London (Pierce, 1968; Reed and Neel, 1955; Veale, 1965).A penetrance rate of approximately 80% with an autosomal dominant mode of inheritance occurs. Fewer cases than expected have resulted in some families due to this incomplete penetrance. In A.C.R., the presence of hundreds to thousands of adenomatous

290

MARTIN LIPKIN

polyps characterizes the classical expression of this disease. These may carpet the entire colon and rectum, with denser polyposis shown in the rectum. Sessile and pedunculated adenomatous and villous tumors are usually seen. It is of note that, in A.C.R., carcinomas develop with highest frequency in the lower colon, as occurs in the general population (Alm and Licznerski, 1973). To date no chromosomal entity can be cited as a useful diagnostic or prognostic marker for colon carcinoma or its precursor states, comparable to the Philadelphia chromosome in chronic granulocytic leukemia. Linkage between Duffy blood groups and the gene for A.C.R. susceptibility has been suggested (Veale, 1965). Earlier studies of chromosome pattern in adenomas and adenocarcinomas of the colon had shown hyperploidy in adenomas, and a degree of epithelial dysplasia related to a higher incidence of structural and numerical chromosome abnormalities. In well-differentiated adenocarcinomas, the degree of chromosomal abnormalities was pronounced (Enterline and Arvan, 1967). Abnormalities have more recently been reported involving chromosome numbers 8 and 14 in human colonic adenomas (Market aE., 1973; Mitelman et al., 1974). The changes preceded histologic evidence of invasiveness and were found with or without a hereditary basis. In A.C.R., excessive heteroploidy has been found in the cells of adenomas of A.C.R; while in sporadic polyps, the chromosome numbers were more normal. Cells with abnormal karyotype also showed involvement of C or D group chromosomes, both in A.C.R. and sporadic polyps. Histological changes in colonic mucosa of A.C.R. also have been found in association with the above findings. Minute mammillations of surface epithelium and localized hyperplasia were earlier described together with chronic inflammatory reaction and enlarged lymphoid follicles. Differentiation was made between adenomatous polyps having malignant potential, when arising in A.C.R. and sporadically in the general population, and hyperplastic polyps found in the general population. The proliferative abnormalities described in the next section occur in adenomatous and not in hyperplastic polyps. Figure 1 shows the cumulative percentage of incident cases having A.C.R. at various ages compared to the incident cases of colon cancer in other population groups. These data are useful in defining population groups at increased risk for colon cancer on the basis of age of onset of disease. The distributions of incident cases of (1) cancer in A.C.R., (2) familial colon cancer, (3) multiple primary cancers including the colon with family history of colon cancer, and (4) colon cancer

SUSCEPTIBILITY TO COLON CANCER *--a

Onsetof polyposisinA C R

291

Onset of colon and rectolconcei in general population

{ o---o Onset of concer in A C R

Onset of concer in familial colon cancer &---a Multipleprimary cancers including colon,mthfamily history of colon cancer n--q

u)

8 100 0 L

C

rl E I

0 0 0' 0 c

5 50

Y a 0)

-0 c

;20

(3

10

0 Age in years

FIG.1. Cumulative percentage of incident cases having inherited adenomatosis of the colon and rectum at various ages, compared to incident cases of colon cancer in other population aggregates. Data of A.C.R. are from Bussey (1975). Data of familial sitespecific colon cancer are from Dr. H. Lynch, and multiple primary cancers are from Memorial Hospital registry. The range of data of colon and rectal cancer in general population is from the Third National Cancer Survey, National Cancer Institute, and includes white and black males and females.

in the general population were subjected to statistical analysis. Pairs of distributions were compared by the Kolmogorov-Smirnov 2 sample test. All pairs of distributions shown in Fig. 1 were found to be significantly different ( p < .05, two sided test) indicating that the first three population groups are at increased risk for colon cancer compared to the general population. In A.C.R., cancer follows the development of adenomas with great regularity decades before the appearance of colon carcinoma in the general population; no disease leads to cancer with greater certainty. A delay between the onset of adenomatosis and cancer can also be noted, which is most pronounced in the younger age groups. In the older age groups, the development of adenomatosis and cancer appear in closer temporal sequence, in this respect resembling the onset of colon cancer that can be observed in the general population, in patients having one or several adenomas. Adenomatosis can also be detected near birth in a small fraction of individuals. In the past, estimates have indicated that when first seen, two-thirds of patients with adenomatosis will show evidence of cancer; by 40 years of age, over 50% will develop adenocarcinoma.

292

MARTIN LIPKIN

The usual age in A.C.R. for diagnosis has been 25 years, while cancer in polyposis patients is usually diagnosed at later periods. This is about 20 years earlier than average figures for cancer without adenomatosis. Patients usually manifest multiple cancer: Approximately 50% of colectomy specimens show two or more cancers (Morson and Bussey, 1970). As observed, cancers have been reported in patients with adenomatosis at early ages, and after puberty adenoma formation and cancers increase. Individuals who are examined because of a known familial association are often successfully identified at a much earlier age as having A.C.R. without cancer, compared to those who seek medical advice due to symptomatic presentation of disease. As noted, most patients will die of colon carcinoma by age 50 unless a colectomy is performed. Gardner’s syndrome, a variant of A.C.R., is an autosomal dominant disorder showing a high degree of penetrance. Adenomatous polyps of the colon, and occasionally of the small intestine, are formed and there is propensity for adenocarcinoma development within the polyps. The syndrome has a lower incidence than A.C.R., extimated at 1 in 14,000 births. The disease is characterized b y colonic polyposis and soft tissue abnormalities such as sebaceous cysts, desmoid tumors, epidermal cysts, lipomas, subcutaneous nodules, and fibromas. In addition, bone abnormalities including osteomas of the skull, exostoses of the skeleton, cortical thickening of the long bones, and abnormal dentition can be found. Similar to A.C.R., the polyps are adenomas. They are distributed almost completely in the colon and rectum, although they can occur in the small intestine, especially in the region of the ampulla of Vater. Gene expression is variable, with some individuals having polyposis alone and other individuals having the various manifestations of the major triad. The extracolonic manifestations also can develop prior to the colonic adenomas. The age distributions of onset of colonic polyps and progression to colon cancer are similar to A.C.R. A particular problem can arise following surgery. Postoperative fibromas and low grade fibroblast or desmoid malignancies may arise in healing surgical wounds, and can result in invasion of mesenteric root or bowel. Other associated tumors can also occur, including carcinoma of the duodenum and papilla of Vater. Variants of the syndrome can include Turcot syndrome (polyposis coli associated with tumors of the central nervous system) and the Oldfield syndrome (extensive familial sebaceous cysts, polyposis coli, and adenocarcinoma). In Gardner’s syndrome, the adenomatosis and other features of the disease may be due to a single gene although it remains possible that additional genes are involved.


293

A third autosomal dominant inherited disease with variable expression is the Peutz-Jeghers syndrome, characterized by melanin pigmentation of the buccal mucosa, lips, face, fingers, toes, vagina, and anus. Polyps of the gastrointestinal tract, specifically the small intestine are found; about one-third are in the colon and rectum. However, the polyps are hamartomas rather than adenomas (Jeghers et al., 1949). This disorder appears to have very little malignant potential when compared to A.C.R. or Gardner’s syndrome. Nevertheless, some associated stomach and duodenal associated carcinomas have been reported (Dodds et al., 1972). The presence of one or more adenomas occurs in 5 to 10%of individuals in the general population, and can be associated with the development of adenocarcinoma. Kindreds have also been reported showing an association of single and multiple adenomas with adenocarcinoma, a link that appears to be genetically influenced. One study, by Woolf et al., (1975) showed that 45% of the adult members of one generation had solitary adenomas as well as the occurrence of adenomas in multiple generations. This family also had a high incidence of colon carcinoma. An autosomal dominant mode of inheritance is reinforced by these observations. Another inherited disorder is juvenile polyposis of the colon. These polyps are hamartomas and are not viewed as potentially malignant. Relatives of these juveniles do, however, express an above normal occurrence of adenomas and colorectal adenocarcinomas. IV. Proliferative Abnormalities and Susceptibility to Colon Cancer

Studies of cell proliferation have aided our understanding of events that develop during neoplastic transformation of colonic cells in A.C.R., and in individuals in the general population who have developed colon cancer. In individuals having A.C.R., colonic epithelial cells predestined to develop neoplasia show characteristic proliferative changes. During progressive stages of abnormal development, cell phenotypes appear in which epithelial cells gain an increased ability to proliferate and to accumulate in the mucosa. In the normal colon of man, the major proliferative activity occurs in the lower and midregions of the crypts adjacent to the base, occupying about three-quarters of the crypt columns. Approximately 15 to 20% of the proliferating cells are in DNA synthesis. The number of cells in the proliferative cycle diminishes as they advance to the mouth region of the crypts; within hours cells undergo further differentiation and proliferative activity ceases as they approach the crypt surface (De-

294

MARTIN LIPKIN

schner et al., 1963; Lipkin and Deschner, 1968; Lipkin and Quastler, 1962; Lipkin et al., 1963; Messier and Leblond, 1960). In subjects with A.C.R, patches of flat mucosa can be detected having colonic epithelial cells that fail to repress DNA synthesis during migration to the surface of the mucosa (Deschner and Lipkin, 1970; Deschner et al., 1966). This finding has been observed in normal appearing colonic epithelial cells of subjects with A.C.R. before they develop adenomatous changes and before the cells begin to accumulate as polyps. It has been noted in 80% of random biopsy specimens (Fig. 2). It also was recently observed with high frequency in flat mucosa of individuals in the general population who developed primary colon cancer (Maskens and Deschner, 1977). In inherited adenomatosis, colonic epithelial cells develop additional properties. Cells that fail to undergo normal maturation with repression of proliferative activity also acquire altered morphological characteristics identified pathologically as “adenomatous.” These cells accumulate in the colonic mucosa and form tubular or villous structures initiating the formation of adenomas. Carcinomas develop with increasing frequency as these adenomatous excrescences enlarge (Morson, 1976). Figure 3 shows a sequence of events believed to lead to malignancy in inherited A.C.R. Proliferating colonic cells that have inherited the germinal mutation fail to repress DNA synthesis. Additional events then occur giving rise to new clones from the original cell population. The first leads to the development of the well-known

FIG.2. Quantitation with modified Venn diagram of a set of cases having two simultaneous phenotypic attributes. The diagram shows in 35 individuals in A.C.R. familial aggregates, the observed relationship between the presence of adenomas (n’)and failure of colonic epithelial cells to repress DNA synthesis (s’) in biopsies of flat mucosa. (s’) is present in 80% of random biopsy specimens of flat mucosa and (n’) is present in 54%of A.C.R. population, in contrast to controls without known gastrointestinal disease or history compatible with A.C.R. where (s‘) is present in 18%of random biopsy specimens and (n’) is zero. Numerical values are continually modified as new cases are added. (p) indicates fraction of cases where neither (s’) nor (n’)have been detected in colonic mucosa.


295

FIG. 3. A sequence of events believed to lead to malignancy in inherited A.C.R. Notations s' and n' as in Fig. 2 (n'denoting development of adenomatous cells that proliferate and accumulate, initiating neoplastic growth) m' denoting malignancy. Proliferating colonic cells that have inherited the germinal mutationfail to repress DNA synthesis. Additional events are then believed to occur giving rise to new clones from the original cell population and leading to the development of adenomatous and malignant cells.

adenomatous cells that proliferate and accumulate on the surface of the mucosa. I n A.C.R., according to this concept, an additional event then occurs in the cells giving rise to invasive malignancy. Numerical estimates of the events that could lead to colon cancer have recently been made (Knudsen, 1976). following the approach of Fisher (1958). I n that analysis as well as our own, there is high agreement between the age-specific cumulative incidence rate of cancer in A.C.R. and numerical estimates when based on the occurrence of a low number of sequential events after the germinal mutation. A failure of colonic epithelial cells to repress D N A synthesis also occurs in ulcerative colitis (Eastwood and Trier, 1973). In ways similar to diseases of the colon, in atrophic gastritis, a condition associated with the development of gastric malignancy, epithelial cells also fail to repress D N A synthesis and undergo abnormal muturation as they migrate through the gastric mucosa (Deschner et al., 1972; Winawer and Lipkin, 1969). In precancerous disease of the cervical epithelium in humans (Wilbanks et al., 1967) and in the cervix of rodents after a chemical carcinogen (Hasegawa et al., 1976), a similar event occurs. Thus, during the development of neoplasms in other organs as well as in colon, persistent D N A synthesis has been observed in cells that would be normally expected to be terminal or end cells prior to invasive carcinoma. The findings support the possibility that common defects develop in the regulatory control of cell proliferation during neoplastic transformation in the epithelial lining of these various organs.

296

MARTIN LIPKIN

V. Newer Immunologic Studies

Immunologic factors that may be related to increased susceptibility of familial aggregates to colon cancer are under investigation. An abnormality has recently been detected in some individuals at increased risk of colon cancer. When cancer free individuals from families highly predisposed to colon cancer (without classic A.C.R.) were studied to determine the nature of their cell-mediated immune capacities, 44% demonstrated an apparent perturbation of adherent cell function, manifesting itself as an inappropriate suppression of a potentially normal lymphocyte ability to respond to an allogenic stimulus. The defect in recognitive immunity appeared to be the same defect that was demonstrated in individuals with established malignancies (Table 111). Patients with recognized Gardner's syndrome also showed the deficit of recognitive immunity (Berlinger et al., 1977). These studies, which are being extended to additional disorders leading to colon cancer, offer the possibility of new means for the early detection of susceptible population groups. VI. Nuclear Protein and Enzyme Alterations

The enzyme composition of colonic cells has been shown to change in association with the development of adenomas and carcinomas in the colon of man. Activity of thymidine kinase is higher in neoplastic TABLE I11 MIXED LEUKOCYTE CULTURE RESPONSE OF UNAFFECTED MEMBERS OF COLON CANCERFAMILIAL AGGREGATES" Relative response of family members as percentage of controls

Relative response after filtration of cells through G I0 beads

P

38 48

88 77

M

36 66 53

70 78 60

C

63

68

J

61

75

V

39

71

Family

From Berlinger et al. (1977).

297


than in mature colonic cells (Salser and Balis, 1973; Troncale et al., 1971). More recent studies have indicated that after administration of the carcinogen 1,S-dimethylhydrazine (DMH), thymidine kinase was altered both quantitatively and qualitatively, with properties resembling fetal enzyme after long term DMH treatment. I n human tissue, the enzyme from malignant cells was antigenically more like surface cell enzyme than from crypts. The enzyme from placental extracts and from adenomatous polyps was similar to tumor enzyme (Salser and Balis, 1973) (Table IV). Other cellular enzymes including ornithine decarboxylase increased after DMH in the colon but not in the liver, while the liver carcinogen acetylaminofluorene induced a marked increase in liver enzyme but not colonic (Ball et al., 1976).These findings have indicated specificity in enzyme complement associated with carcinogen-induced neoplastic transformation of the cells, and the possibility of indices denoting tumorigenesis in human populations highly susceptible to cancer. Recent studies have also shown that nuclei isolated from DMHinduced tumors contained characteristic complements of nonhistone nuclear proteins. These were not prominent in normal colonic epithelial nuclei nor in epithelial cells surrounding the tumors. Characteristic abnormal nuclear proteins were also detected in human colonic carcinomas and in a human colon carcinoma cell line (Fig. 4)(Boffa and Allfrey, 1976). Although these were not found in nonmalignant adenomatous polyps in A.C.R., the identification and selective accumulation of such proteins in colonic tumor nuclei and the developTABLE I V OF TK IN VARIOUS TISSUEEXTRACTS AFTER PREINCUBATION WITH ANTIHUMAN COLONIC T K FROM RABBITS~

RESIDUAL ACTIVITY

Source of TK

Partially purijied by a single precipitation with ammonium sulfate Normal colon Adenocarcinoma colonic Placenta I Crude extract Ovarian carcinoma PHA-stimulated lymphocytes Myeloblastic leukemic leukocytes From Salser and Balis (1973).

Control TK activity (in pica moles per incubation sample)

Undiluted

402 2 662 3 68 T 2

54 60

80 2 4 762 2 85? 4

35 95 95

8

298

MARTIN LIPKIN

HN

HTI

HTn

C

4 Molecular weight x

lo3

FIG.4. Electrophoretic analysis of nuclei isolated from normal human colonic epithelial cells, from two different adenocarcinomas of the colon, and from a cell line HT-29 derived from an adenocarcinoma. Electrophoretic profiles of all tumor cells show prominent peaks (indicated by vertical lines) while normal colonic epithelial nuclei do not. (From Boffa and Allfrey, 1976.)

ment of analytical procedures for their detection in single cells offers a new approach to the early detection in single cells of molecular events associated with malignancy in man. In mice, morphologic and proliferative changes have been induced by carcinogens similar to those observed in man. After DMH was ad-


299

ministered to mice, the distal colon was the main site of tumor nodule formation, a distribution common to that found in man. Multifocal tumors ranging from adenomatous polyps to metaplasias and carcinomas grew from the mucosa and then protruded into the lumen of the rectosigmoid. Early focal atypias and hyperplasia, located mainly on the folds, and adenomatous polyps and carcinomas appeared to be part of the progressive pathologic changes in mice and rats, and developed following administration. These changes were accompanied by an increased proliferative activity in the cells. Both DMH and MNNG induced an extension of the proliferative region of the flat mucosa toward the surface of the colonic crypt, as observed in man, with development of cells that continue to incorporate thymidine into DNA throughout their life span (Deschner, 1974; Kikkawa, 1974; Thurnherr et al., 1973; Wiebecke et al., 1973) (Fig. 5). VII. Studies of Cutaneous Cells

Recent studies have suggested that phenotypic expressions of the inherited disease ACR may extend to cutaneous cells. Increased heteroploidy in cutaneous epithelial cells derived from individuals

FIG. 5. Increase in proliferative compartment in mice after repeated injections of DMH (A, upper diagram, from Thurnherr et al., 1973) and after rectal installation in MNNG in rat (B, lower diagram, modified from Kikkawa, 1974).

300

MARTIN LIPKIN

with Gardner’s syndrome was observed (Danes, 1976). It has also been noted that cutaneous fibroblasts derived from normal subjects grow in well-organized monolayers, in contrast to those from individuals with A.C.R., which have shown larger regions of criss-cross arrays and random orientation forming multilayers. A susceptibility to neoplastic transformation when exposed to Kirsten murine sarcoma virus has been reported (Pfeffer and Kopelovich, 1977).Of particular interest is the recent report of differences in actin distribution within cultured cells from individuals with A.C.R. compared to normals (Kopelovich e t al., 1977).In order to determine the specificity of findings of this type, the observations are presently being extended to include analysis of cutaneous cells from larger control groups and from additional families with various patterns of inherited polyposis and colon cancer. VIII. Examination of Fecal Contents

Additional studies are in progress to identify abnormal constituents of fecal contents and to examine their potential carcinogenic activity in colon cells. Studies of potential bacterial markers were referred to previously. The bile acids and their bacterial conversion products are a group of compounds currently under examination. Several recent reports compared the fecal neutral steroids and bile acids in patients with A.C.R. and controls other than relatives (Drasar et d., 1975; Reddy et al., 1976; Hackman et al., 1976; Watne and Core, 1975). Individuals with A.C.R. excreted higher amounts of cholesterol and lower levels of coprostanol and coprostanone. Nondegradation of cholesterol also was found in about one-quarter of individuals in the general population (Hackman et al., 1976). Further studies are also in progress to assess the utility of these variations in cholesterol and its metabolites in screening A.C.R. family siblings for disease. A further difference in fecal neutral steroid concentration of North American white and South African black populations was recently found, with most of the neutral steroids free (nonesterified) in North Americans; and a high proportion of plant and animal steroids esterified to long-chain fatty acids in Africans (Salyers, 1977). Recently, an additional and potentially important lead to the identification of increased susceptibility to colon cancer, was provided with detection of mutagenic activity in the feces of humans. It was suspected that N-nitrosa compounds might be associated with this mutagenic activity (Vargheseet al., 1977). Further work has suggested that anitroso group exchange reaction can occur with transfer from nitrosamine to amide moiety, resulting in the generation ofhighly reactive nitrosamide


301

compounds in feces (Mandel, 1977). Several laboratories are now actively pursuing the question of whether fecal mutagens may interact with cells of the colon to induce the sequence of proliferative and morphologic changes that lead to the evolution of colon cancer. IX. Conclusion

The findings described have identified abnormal stages of development of colonic epithelial cells and related physiological and environmental factors that may influence the development of neoplasia, as well as individuals and familial aggregates at increased risk. Individuals and population groups are now classified on the basis of abnormal cell phenotype and related physiological and fecal content abnormalities. These classifications are leading to new indices that attempt to identify heightened degrees of susceptibility of individuals at increased risk of colon cancer, and the stages of development of their disease. With precise quantitative information available, elements that modify or accelerate the progression of disease in man can be studied. Future programs designed to identify high-risk population groups will require systematic analyses of mode and time of appearance of abnormal phenotypic expressions in these individuals and their families. Analysis of the risk factors together with corresponding study of interactions that take place with relevant carcinogens can lead to new means to identify individuals and population groups at increased risk of colon cancer, and studies are under way for that purpose. New programs designed to prevent the evolution of the cellular changes leading to malignancy are being considered at the present time utilizing the measurements described herein. ACKNOWLEDGMENTS Studies reported here from the author’s laboratory were aided by Contract 1-CP-43366 and Grant CA98748, from the National Cancer Institute, Department of Health, Education and Welfare.

REFERENCES A h , T., and Licznerski, G. (1973). Clin. Gastroenterol. 2,577-602. Anderson, D. E. (1970). Collect. Pap. Annu. Symp. Fundam. Cancer Res. 23,85-109. Andrews, E. D., Hartley, W. J., and Grant, A. B. (1968). N . 2. Vet.J. 16, 3-10. Aries, V., Crowther, J. S., Drasar, B. S., Hill, M. J., and Williams, R. E. 0.(1969).Gut 10, 334-335. Armstrong, B., and Doll, R. (1975). Int. J. Cancer 15,617-631. Asano, T., Pollard, M., and Madsen, D. (1975).Proc. Sac. E x p . Biol. Med. 150, 78G785. Ball, W. J., Salser, J. S., and Balis, M. E. (1976). Cancer Res. 36, 2686-2689.

302

MARTIN LIPKIN

Berg, J. S., Haenszel, W., and Devesa, S. (1973). Proc. 7th Natl. Conf. Cancer Colon Rectum, 1973, pp. 459-464. Berlinger, N. T., Lopez, C., Vogel, J., Lipkin, M., and Good, R. A. (1977).]. Clin. lnuest. 59,761-769. Boffa, L. C., and Allfrey, V. G. (1976).Cancer Res. 36, 2678-2685. Burkitt, D. P. (1975).J.Natl. Cancer lnst. 54, 3-6. Bussey, H. J. R. (1975). “Familial Polyposis Coli.” Johns Hopkins Univ. Press, Baltimore, Maryland. Chomchai, C., Bhadrachari, H., and Nigro, H. (1974).Dis. Colon Rectum 17,310-312. Danes, B. S. (1976). Cancer 38, 1983-1988. Deschner, E. E . (1974). Cancer 34,824-828. Deschner, E. E., and Lipkin, M. (1970).J.Natl. Cancer Inst. 44, 175-185. Deschner, E. E., Lewis, C. M., and Lipkin, M. (1963).J.Clin. Inuest. 42, 1922-1928. Deschner, E. E., Lipkin, M., and Solomon, C. (1966).J . Natl. Cancerlnst. 36,849-857. Deschner, E. E., Winawer, S. J., and Lipkin, M. (1972).J.Natl. Cancer Inst. 48, 15671574. Dodds, W. J., Schulte, W. J., Henley, G. T., and Hogan, W. J. (1972).Am.J.Roentgenol. Radium Ther. Nucl. Med. [N.S.] 115, 374-377. Doll, R. (1969). Br. J. Cancer 23, 1-8. Drasar, B. S., Bone, E. S., Hill, M. J., and Marks, C. G. (1975). Gut 16, 824-825. Eastwood, G . L., and Trier, J. S. (1973). Gastroenterology 64, 383-390. Enterline, H. T., and Arvan, D. A. (1967). Cancer 20, 1747-1759. Feingold, S., Alteberg, H., and Sutter, V. (1974).Am. J. Clin Nutr. 27, 1456-1469. Fisher, J. C. (1958).Nature (London) 181,651-652. Fraumeni, J. F., Jr. (1973).In “Cancer Medicine” (J. F. Holland and E. Frei, eds.), pp. 7-15. Lea & Febijer, Philadelphia, Pennsylvania. Gardner, E. J. (1962).Am. J. Hum. Genet. 14, 376-390. Gilbertsen, V. (1974). Cancer 34, 939-939. Gregor, O., Tomon, R., and Prusova, F. (1969). Gut 10, 1031-1034. Hackman, A. S., Wilkins, T. D., Finegold, S. M., and Sutter, V. C. (1976). Lancet, April 3, 752. Haenszel, W., and Correa, P. (1971). Cancer 28, 14. Haenszel, W., and Kurihara, M. (1968).J.Natl. Cancer lnst. 40, 43-68. Haenszel, W., Berg, J,, Sezi, M., Kurihara, M., and Locke, P. (1973)J. Natl. Cancer Inst. 51, 1765-1779. Harr, J. R., Exon, J. H., Weswig, P. H., and Whanger, .P. D. (1976). Clin. T O X ~ C6, O~. 287-293. Hasegawa, I. Matsuura, Y., and Tojo, S. (1976). Cancer Res. 36, 359-364. Hentges, D., Flynn, M., Burton, G., Franz, J., Gehrke, C., Gerhardt, K., Maier, B., Tsutakawa, R., and Wixom, R. (1976).Proc. Int. Symp. Detect. Preu. Cancer, 3rd, Symposium 17, no. 363. Hill, M. (1971).J . Pathol. 104, 239-245. Hill, M. (1975). Cancer 36,2387-2400. Hill, M., Drasar, B., Aries, V., Crowther, J., Hawksworth, G., and Williams, R. (1971). Lancet 1,95-100. Hill, M., Drasar, B., and Williams, R. (1975).Lancet 1,535-538. Hill, M. J. (1974). Cancer 34, Suppl., 815-818. Jacobs, M. M., Jansson, B., and Griffin, A. C. (1976).Cancer Lett. (in press). Jannsson, B., Malahy, M. A., and Seibert, G. B. (1976). Proc. lnt. Symp. Detect. Preo. Cancer, 3rd, 1976 (in press).


303

Jeghers, H., McKusick, V. A,, and Katz, K. (1949).N . E n g . ] . Med. 241,1021-1036,1949, Kikkawa, N. (1974). Med. J . Osaka Uniu. 24, 293-314. Knudsen, A. G., Jr. (1976). Proc. Can. Cancer Res. Conf. 11,93-103. Kopelovich, L., Conlon, S., and Pollack, R. (1977).Proc. Natl. Acad. Sci. U.S.A. 74,3019. Lipkin, M., and Deschner, E. (1968). E x p . Cell Res. 49, 1-12. Lipkin, M., and Quastler, H. (1962).Nature (London) 194, 1198-1199. Lipkin, M., Bell, B., and Sherlock, P. (1963).J . Clin. Inuest. 42, 767-776. Lovett, E. (1974).Proc. R. SOC.Med. 67,751-752. Lynch, H. T. (1967). Recent Results Cancer Res. 12,67-85. Lynch, H. T., and Gush, A. J. (1967).Gastroenterology 53,517-527. Mandel, M., Ichinotsubo, D., and Mower, H. (1977). Nature (London) 267, 248-249. Mark, J., Mitelman, F., Dencker, H., Norryd, C., nnd Tranberg, K. G. (1973).Acta Pathol. Microbiol. Scand., Sect. A 81,85-90. Maskens, A. P., and Deschner, E. E. (1977).J . Natl. Cancer Inst. 58, 1221-1224. Mastromarino, A., Sharma, C., Molless, A,, Reddy, B., and Wynder, E. (1976).Am. S O C . Microbiol. p. 129. Messier, B., and Leblond, C. P. (1960).Am. J . Anat. 106,247-254. Mitelman, F., Mark, J., Nilsson, D. G., Deucker, H., Norryd, C., and Tranberg, K. G. (1974). Hereditas 78,63-68. Moertel, C. G . , Bargen, J. A., and Dockerty, M. B. (1958). Gastroenterology 34, 385. Moore, W., and Holdeman, L. (1975). Cancer Res. 35, 3418-3420. Morson, B. C. (1976). Clin. Gastroenterol. 5, 505-525. Morson, B. C., and Bussey, H. J. R. (1970). I n “Current Problems in Surgery” (M. Ravitch, ed.), pp. 1-50. Yearbook Publ., Chicago, Illinois. Narisawa, T., Magadia, N., Weisburger, J., and Wynder, E. L. (1974).J . Natl. Cancer Inst. 55, 1093-1097. Nigro, N., Bhadrachari, N., and Chomchai, C. (1973). Dis. Colon Rectum 16, 438-443. Pfeffer, L. M., and Kopelovich, L. (1977). Cell 10, 313-320. Pierce, E. R. (1968). Dis.Colon Rectum 11, 321. Pomare, E., Heaton, K., Lowbeer, T., and White, C. (1974).Gut 15,8244325. Reddy, B. S., and Wynder, E. L. (1973).J.Natl. Cancer Inst. 50, 1437-1442. Reddy, B. S., Weisburger, J., and Wynder, E. L. (1975a).J . Nutr. 105,878-884. Reddy, B. S., Narisawa, T., Maronpot, R., Weisburger, J., and Wynder, E. L. (197513). Cancer Res. 35, 3421-3426. Reddy, B. S.,Mastromarino, A., Gustafson, C., Lipkin, M., and Wynder, E. L. (1976). Cancer 38, 1694-1698. Reed, T. E., and Neel, J. V. (1955).A m . J . Genet. 7, 236263. Rodgers, A., and Newberne, P. (1973). Nature (London) 246,491-492. Rodgers, A., Herndon, B., and Newberne, P. (1973). Cancer Res. 33, 1003-1009. Salser, J. S.,and Balis, M. E. (1973). Cancer Res. 33, 1889-1897. Schrauzer, G. N., and Ishmael, D. (1974).Ann.Clin. Lab. Sci. 2,441-447. Silverberg, E., and Holleb, A. (1975). Ca 25,8-22. Thurnherr, N., Deschner, E. E., Stonehill, E. H., and Lipkin, M. (1973). Cancer Res. 33, 940-945. Troncale, F., Hertz, R., and Lipkin, M. (1971). Cancer Res. 31, 463-467. Varghese, A. J., Land, P., Furrer, R. and Bruce, W. R. (1977).Proc. 68th Annu. Meet. Am. Assoc. Cancer Res. 18, 317. Veale, A. M. 0.(1965). “Intestinal Polyposis.” Cambridge Univ. Press, London and New York. Watne, A. L., and Core, S. (1975).J. Sztrg. Res. 19, 157-161.

304

MARTIN LIPKIN

Wiebecke, U., Krey, U., Lohrs, U., andEder, M. (1973).VirchowsArch.A 360,179-193. Wilbanks, G. D., Richart, R. M., and Terner, J. Y. (1967).Am. J. Obstet, Gynecol. 98, 792-799. Wilkins, T.,and Hackman, A. (1974). Cancer Res. 34, 2250-2254. Winawer, S. J., and Lipkin, M. (1969).J. Natl. Cancer Znst. 42, 9-17. Woolf, C. M., Richards, R. C , and Gardner, E. J . (1955).Cancer 8, 403-408. Wynder, E . L., and Shigematsu, T. (1967). Cancer 20, 1520.

ADVANCES IN CANCER RESEARCH. VOL. 27

NATURAL CE LL-MEDlATE D I M MUNlTY Ronald 6. Herberman and Howard T. Holden

.

Laboratory of Immunodiagnosis.National Cancer Institute: Bethesda Maryland

I . Introduction .......................................................... I1. Characteristics of Natural Cytotoxicity ................................. A . Influence of Age . . . . . . . . . . . . B. Influence of Genetic Backgrot C . Influence of Environmental Factors and Disease ..................... D . I n Viuo Augmentation of Reactivity .................... E . Effects of in Vitro Cultivation of Lymphoid Cells . . . . . . . . . . I11. Specificity of Natural Cell-Mediated Cytotoxicity ................ A . Mice .............................................................. B. Rats . . . . . . . . . . . . . . . ............................................ C . Human ........................................................... IV. Nature of Effector Cells . . . . . . . . . . . . . . . . . ........................... A . Organ Distribution ................................................ .... B. Cell Surface Markers and Other Characteristics of N K Cells . C . Effect of Thymosin in Vitro ........................................ D . Effects of in Viuo Manipulations of Thymus Function ................ E Effect of Immunosuppression on NK Activity . . . . . . . . . . . . . . . V. Relationship of Natural Cell-Mediated Cytotoxicity to AntihodyDependent Cell-Mediated Cytotoxicity ...... A . Correlation of NK Activity with ADCC .............................. B. Comparison of Effector Cells Mediating NK Activity and ADCC . . . . . . C . Possible Mechanisms of Action of NK Cells ......................... VI . Model for Placement of NK and I; Cells in Pathway of Differentiation of T Cells .............................................. VII . Discrimination Between Natural Cell-Mediated Cytotoxicity and Cytotoxicity by Other Effector Cells ................................... A . Mice .......................................... B. Human ....... ............................................ a1 Cytotoxicity .............................. VIII . I n Viuo Relevance A . Correlation of Decreased Tumor Growth with Natural Cytotoxicity . . . . B . I n Viuo Relevance of N K Activity Against Nonmalignant Cells . . . . . . . C . Implications of Natural Cytotoxicity for Immune Surveillance ... Addendum ........................................................... References ...................... ..... ............

.

305 307 310 312 316 319 321 324 324 329 331 333 333 334 340 341 343 345 345 ,347 349 351 354 355 359 361 362 363 365 366 370

I . Introduction

In vivo resistance of the host against progressive tumor growth has mainly been attributed to mature ~Tcells. specifically immune against 305 Copyright 0 1978 by Academir Press. Inc . .411 iights of reproduction 111 iiny fwm reserved.

ISBN 0-12-(X)6627-0

306

RON,4LD B. HERBERMAN AND HOWARD T. HOLDEN

tumor-associated antigens. Indeed, in adoptive transfer experiments in mice and rats, in which the recipients were successfully protected against tumor cell challenge, such specifically immune T cells have been shown to b e necessary (Collavo et al., 1974; Gorczynski and Norbury, 1974; Berenson et al., 1975; Glaser et al., 1976b). Because of this important role of immune T cells, it was anticipated that nude mice and other T cell-deficient individuals would completely lack resistance against spontaneous or induced tumors, and that the incidence of rapidly progressive malignancies would be much higher than in individuals with competent T cell immunity. The failure to find such explosive tumor growth in nude mice, or a complete spectrum of malignant diseases in patients with immunodeficiency diseases, has been taken as strong evidence against the immunosurveillance theory (Moller and Moller, 1975; Outzen et al., 1975; Schwartz, 1975; Rygaard and Povlsen, 1976). In in vitro studies of cell-mediated immunity against tumors, the central role of mature T cells has also been stressed. Cell-mediated cytotoxicity against tumor cells has generally been thought to be mediated by three types of effector cells (Cerottini and Brunner, 1974); (a) specifically immune, mature T cells, (b) antibody-dependent cytotoxic cells, which bear surface receptors for the Fc portion of IgG molecules and thereby interact with IgG antibodies bound to target cells, and ( c ) activated macrophages or macrophages armed with specific antibodies. The recent recognition of the existence of natural cell-mediated immunity, particularly natural cell-mediated cytotoxicity (Oldham et al., 1973; Herberman et al., 1973, 1974a; Rosenberg et al., 1974), has substantially altered our concepts concerning the potential mechanisms for in vivo resistance against tumor growth and for in vitro cell-mediated immune reactions. Natural cell-mediated cytotoxicity has been found to be a general phenomenon in several species, in rats (Holtermann et al., 1973; Nunn et al., 1976; Shellam and Hogg, 1977) and in man (Oldham et al., 1973; Takasugi et al., 1973; Rosenberg et al., 1974) as well as in mice (Herberman et al., 1973, 1974a, 1975a; Kiessling et id., 1975a). With mice, almost all studies have involved the use of a short-term 51Crrelease cytotoxicity assay (CRA). With rats and humans, the phenomenon has been studied with both the CFL4 and also with long-term visual microcytotoxicity and radioisotopic cytotoxicity assays. Some studies with lymphoproliferative assays (Kanner et al., 1970; Shoji and McKhann, 1971; Forbes et al., 1973; Burk et al., 1975; Boyer and Fahey, 1976; Lopez et al., 1976; Krueger et al., 1977; Lee and Ihle, 1977) or migration inhibition assays (Lopez

NATURAL CELL-MEDIATED IMMUNITY

307

et al., 1976) have also shown that normal individuals may have cellmediated reactivity against tumor-associated or oncogenic virusassociated antigens. However, the characteristics of natural immunity detected in these noncytotoxic assays, and the nature of the effector cells mediating these reactions, have not been studied in detail. The findings of ubiquitous natural cell-mediated immunity must be considered in studies of cell-mediated immunity in tumor-bearing individuals and in discussions of immune surveillance. The occurrence of natural cell-mediated cytotoxicity against human tumor-derived cell lines has created particular problems in evaluating the reactivity of cancer patients (Herberman and Oldham, 1975).It is now clear that one must carefully evaluate the role of natural killer (NK) cells, as well as other, more well-known mechanisms of cytotoxicity, when cytotoxic reactions are measured. In this review, w e will concentrate on natural cell-mediated cytotoxicity, since most of the information in natural cellular immunity has been gathered with this assay. We will summarize the known information on the expression of natural cytotoxicity in rodents and in man, its specificity, the nature of the effector cells and their relationship to other immune mechanisms, and the possible in vivo relevance of natural cytotoxicity, particularly in regard to resistance against tumor growth. Since a large amount of information in each of these areas has been obtained and there are a number of differences among the different species studied, Table I represents an attempt to summarize many of the important points, which can be referred to for orientation by the reader while going through the details below.

II. Characteristics of Natural Cytotoxicity

The basic observation which initiated studies of natural cellmediated cytotoxicity was that lymphoid cells from some normal mice, rats, and human donors, which were not inoculated with tumor cells or other sources of antigen, had significant levels of cytotoxic reactivity against certain syngeneic or allogeneic tumor cells (Herberman et al., 1973, 1974a; Oldham et al., 1973; Rosenberg et al., 1974; Nunn et al., 1976). At first, only low levels of reactivity were observed. However, upon further investigations it was found that a large number of factors influenced the expression of NK activity and that with certain donors and tumor target cells, very high levels consistently could be detected. Such high natural cytotoxicity may exceed the levels of reactivity by specifically immune, mature T cells against highly antigenic

TABLE I SUMMARY OF CHARACTERISTICS OF NATURAL CELL-MEDIATED CYTOTOXICITY Mice LRvels of NK activity Age

Genetic background

Environmental factors and disease

In viuo boosting

I n vitro

Specijici t y

Absent at birth in most strains; peak levels at 5-8 weeks, low after 12 weeks Low activity in A strain against several target cells; high levels in CBA and nude mice; high reactivity dominant in F, Activity present in gem-free; some evidence for inhibitory effects of environmental factors, tumors, and other diseases Specific boosting (peak at 3 days) by some tumor cells, normal thymocytes, and bone marrow cells; apparently nonspecific boosting by murine viruses, BCG, C . parvurn Decrease in activity at 37°C within 2-3 hours and almost complete loss b y 24 hours At least 3 specific antigens detected with possible association of antigens with endogenous type-C viruses; antigens present on lymphomas, other mouse tumors, some nontransformed culture lines, normal thymocytes, bone marrow cells, PHA blasts; most xenogeneic cell lines negative but some human lines positive

Rats

Human

Absent at birth; peak levels at 5-8 weeks; low after 10-12 weeks Low activity in BN rats

No clear age relationship; some activity i n cord blood

Activity present in gem-fiee

Lab workers = other normal donors; possible increase with some viral infections; low activity in some cancer patients Boosting by influenza vaccine, peak at 2-3 days

No information yet available

No lability at 37°C; increased activity after 3-24 hours Antigens with possible association with endogenous type-C viruses mainly limited to rat cells; reactivity against some mouse tumor cells and human cultured cells

Males > female; low activity in males with HLA-A3, B7 haplotype

No lability at 37°C; increased activity after 5 days in medium with fetal bovine serum or other stimulants Probability of multiple antigens; broad specificity on human cultured cell lines, some human tumor cells, and some rodent tumor lines; not due to fetal bovine serum antigens

Eflector cells Organ distribution

Cell surface markers

Other characteristics

Effect of human thymosin in oitro I n uioo manipulations of thymus function

Relation of NK activity to ADCC

In viuo relevance

Present in spleen, lymph nodes, peripheral blood, peritoneal cavity, bone marrow; absent in thymus Weak expression of 0 antigen on NK cells after boosting and from nudes; Fc receptors present but difficult to detect; no complement receptor Nonadherent, nonphagocytic; inhibited by trypsin but recovers in 18 hours at 37"; moderately resistant to irradiation; sensitive to cyclophosphamide; not inhibited by ammonium chloride solution Small decrease in activity High in nude, nude-asplenic mice and neonatally thymectomized mice; low in thymus-grafted nude mice Good correlation between levels of activity with age, strain; effector cells have similar characteristics

Some tumors sensitive to NK grow poorly in nude mice, greater tumor resistance in young mice; may mediate bone marrow resistance and possibly anti-microbial resistance

Present in spleen, lymph nodes, peripheral blood, peritoneal cavity, bone marrow

Present in spleen and peripheral blood; absent in lymph nodes, tonsil, thymus

Activity not inhibited by anti-T cell serum; no easily detectable F c receptor or complement receptor

Low affinity receptor for sheep erythrocytes; easily detectable Fc receptor; no complement receptor

Nonadherent, nonphagocytic; inhibited by papain; moderately resistant to irradiation; not inhibited by ammonium chloride solution

Nonadherent, nonphagocytic; inhibited by trypsin and chymotrypsin; moderately resistant to irradiation; markedly inhibited by ammonium chloride solution but recovers after 18 hours at 37" Small decrease in activity

No information yet High in neonatally thymectomized and adult thyrnectomized, irradiated, bone marrow reconstituted rats No information yet

No information yet

No information yet

2 k

2 ir2 0

M

r

c:

s

U

GM

U

Same organ distribution, correlation between levels of activity among individuals; effector cells have same characteristics but difference in effects of trypsin and protein A No information yet

d

2

' w 0 co

310

RONALD B . HERBERMAN AND HOWARD T. HOLDEN

syngeneic tumor cells (e.g., in the mouse sarcoma virus (MSV) system; cf. Lavrin et al., 1973; Herberman et al., 1975a). In this section, we will discuss the major factors which have been found to affect the expression of natural cell-mediated cytotoxicity.

A.

INFLUENCE OF

AGE 1. Mice

In nude mice, as well as in conventional thymus-bearing mice, age has been shown to have an important and consistent effect on expression of natural cytotoxic reactivity. Our studies (Herberman et al., 1975a) and those of Kiessling et al. (1975a) indicated that lymphoid cells from mice less than 3 weeks of age lacked detectable cytotoxic reactivity, and that cytotoxicity appeared at about 4 or 5 weeks of age and reached peak levels between 5 and 8 weeks of age. Thereafter, there was a decline in activity to low levels. In conventional mice, there was almost no detectable activity in mice greater than 12 weeks of age. However, in nude mice, the decline has appeared to be more gradual, with persistence of some reactivity for longer periods of time. In some other studies, age was also found to be an important factor, but the kinetics of appearance and persistence of NK activity were substantially different. Gomard et al. (1974), in studies of AKR mice, found no natural reactivity against Gross virus-induced lymphoma cells in mice less than 3 months old, and detected considerable reactivity in 3- to 5-month-old mice and even in some mice which were more than 5 months of age. In contrast, we have found considerable levels of natural reactivity in 2-month-old AKR mice. Greenberg and Playfair (1974) only found natural reactivity in older NZB mice, mainly at 9 months of age. A small proportion of mice had very high levels of cytotoxicity against a subline of P815 ascites tumor cells. However, they did not test their mice at 2 months of age, and it is possible that a higher incidence of reactivity would have been found at that time. Indeed, Glimcher et al. (1977) and we have found that young NZB mice are strongly reactive against a different lymphoma target cell. It seems possible that the kinetics of natural cell-mediated cytotoxicity might vary with the target cells used. However, we found that there was a good correlation between the levels of reactivity against different target cells (Herberman et at., 1975a). Development of strong cytotoxic reactivity against one tumor line was always associated with high reactivity against other susceptible tumor cells. There


31 1

were no instances of high reactivity against one target and low or absent reactivity against another target cell which was sensitive to cytotoxicity by effector cells from that strain.

2. Rats Age was found to play a similar important role in the expression of natural cell-mediated cytotoxicity in rats. Nunn et al. (1976) found that reactivity in CRA was maximal in 5- to 8-week-old rats. Lymphoid cells from 1-week-old rats never gave positive results, but, b y 3 weeks of age, significant reactivity was seen in some rats. The lymphoid cells from rats older than 10 weeks of age were either unreactive or produced levels of cytotoxicity only slightly higher than the baseline control. With the rats it was possible to perform serial studies with peripheral blood lymphocytes from individual animals. Reactivity was found to change abruptly, with a shift from high levels to unreactivity within 1week. Shellam and Hogg (1977) obtained a similar pattern of results, but the time course for appreciable levels of reactivity was more prolonged. Cytotoxic activity could be detected by 2 weeks of age, plateau levels were first reached at about 10 weeks, and the levels first declined at about 36 weeks. In longer term cytotoxicity assays, somewhat different results have been obtained (Oldham et al., 1977). In a [3H]-proline release assay, the highest activity was seen at 3-4 weeks of age, but activity then remained at about the level up to 16 weeks of age. In an [1251]iododeo~~ridine release assay involving a 48-hour period of incubation, there was no detectable 3- to 4-week peak but rather a similar incidence of high reactivity from 3 weeks to greater than 16 weeks of age. As discussed below (Section II,E), this persistence of high levels of reactivity of older rats in the assays with longer periods of incubation may be related to in vitro activation of cytotoxic reactivity.

3. Human In contrast to the marked age dependence of NK reactivity in rodents, age has not been found to have a major effect on human NK reactivity. Although some differences in reactivity have been associated with age, the general observation has been that most normal donors, over a wide age range, have easily detectable NK reactivity. The exceptions to this have been the finding of Rosenberg et al. (1972) that donors under the age of 16 had a considerably lower incidence of cytotoxic reactivity against leukemic target cells and the observation of Campbell et al. (1974) that cord blood lymphocytes had much lower

312

RONALD B. HERBERMAN AND HOWARD T. HOLDEN

reactivity than that of adults. However, several investigators have found that cord blood lymphocytes had moderate levels of reactivity (Levin et al., 1975; W. H. West and R. B. Herberman, unpublished observations) or levels indistinguishable from adults (Jondal and Pross, 1975). Takasugi et al. (1973) divided their adult donors by age into three groups, less than 29 years, 30-49 years, and greater than 50, and found no differences in the incidence of reactivity against most target cells. Similarly, Oldham et al. (1975) found no significant relationship between cytotoxicity and the age of normal donors. In contrast to the transient NK reactivity in young rodents, high levels of NK reactivity have been found to persist in some adult human donors for long periods of time (Rosenberg et al., 1974). B. INFLUENCE OF GENETICBACKGROUND

1. Mice

The levels of natural cytotoxic reactivity have been shown to vary widely among different strains of mice. The findings with lymphoid cells from a particular strain seem to be strongly influenced by the target cell used for the studies. In experiments with YAC and RBL-5 lymphoma cells (Kiessling et al., 1975a; Herberman et al., 1975a), CBA mice were found to have strong reactivity and A mice had low levels or undetectable activity. Based on such findings with YAC target cells, Kiessling and his associates have performed a detailed series of experiments to determine the genetic factors influencing levels of natural reactivity. Certain F1 hybrids with A mice as one parent showed good reactivity, whereas others had low reactivity (Petrinyi et al., 1975). From such studies, these investigators concluded that high natural reactivity was dominant. In further experiments, the levels of NK activity appeared to be influenced b y multiple genes, and the H-2 genotype appeared to be one important factor ( P e t r h y i et al., 1976). It seemed from such experiments that one of the genes affecting the levels of reactivity was present on the 17th chromosome, linked to H-2. From recent studies of several congenic resistant strains, the non-H-2 background of the mice appeared to have a considerable influence on reactivity and it has been concluded that the gene(s) must be situated outside of the H-2 locus (R. Kiessling, personal communication). From some recent studies in our laboratory, it seems unlikely that the genetic factors described in the above paragraph determine the overall levels of NK activity. Using RLd 1 lymphoma cells as targets,


313

we have observed appreciable levels of cytotoxic reactivity in A mice (Herberman et al., 1977a).The median cytotoxicity values of A strain mice and some other strains against RLd1 are shown in Table 11. A strain mice and some other strains found by Petrcinyi et al. (1975) to be low reactors had levels of reactivity similar to those of C57BL/6 and BALB/c mice. It thus appears that the genetic factors described by Petr6nyi et al. (1975, 1976) affect the recognition of some target cells by NK cells and other factors may affect expression of natural cytotoxic reactivity. As will be discussed in Section 111, NK cells appear to react with a variety of antigenic specificities, and it seems likely that genetic factors help to determine which antigens are recognized. This would also explain a number of other apparent discrepancies among investigators as to the levels of natural cytoxicity in various strains. Even sublines of the same tumor cells may show consistent differences in sensitivity to lysis by NK cells from different strains. Sendo et al. (1975) and Glimcher et a2. (1977) found that spleen cells from BALB/c mice had no appreciable cytotoxic reactivity against cultured RLd 1 TABLE I1 NATURAL CELL-MEDIATEDCYTOTOXIC REACTIVITYOF SPLEEN CELLS FROM VARlOUS CONVENTIONAL STRAlNS OF MICE AGAINST RL61 TISSUECULTURECELLS Strain

A A.TL A.TH A.CA A.SW A.BY A.AL ATFR-1 ATFR-2 BALBJc C57BW6 AKR 129 SJL CBNN CBNH CBNCAHN CBNTG N ZB a

Effector cell :target cell ratio of 200: 1.

Median percentage of cytotoxici ty"

10 8 8 6

7 5 2

5 7 10

11 18 21 5 30 29 15 19 15

314

RONAL~D B. HERBERMAN AND HOWARD T. HOLDEN

cells. In contrast, we have found that BALB/c mice react against an established culture line of RLd 1to about the same extent as C57BL/6 mice (Herberman et al., 1975a, 1977a). WP did find that C57BIJ6 and not BALB/c were reactive against RLd 1 ascites tumor cells. To examine directly the differences in reactivity of BALB/c mice against RLd 1 cells, we tested the same effector cells against our established RLd 1 tissue culture line and also against RLd 1 cultured cells from Dr. Cantor (Glimcher et al., 1977) and a short-term RLd 1culture that we had initiated 1 month before. The BALB/c spleen cells reacted well only with our established line, whereas effector cells from some other strains reacted well against all three target cells. Similar to the above differences noted for BALB/c, we and Gomard et al. (1974) observed appreciable reactivity in AKR mice, whereas Zarling et al. (1975) and Sendo et al. (1975) found AKR mice to be unreactive. Zarling et al. (1975) also found C58 mice to be unreactive, whereas Sendo et al. (1975) observed considerable reactivity. In contrast to the considerable variation among strains of conventional thymus-bearing mice in their cytotoxic reactivity against a given target cell, nude mice with a variety of different genetic backgrounds have been found to be highly reactive against several different target cells (Herberman et al., 1975a; Kiessling et al., 197513). Table I11 summarizes the reactivity of several types of nude mice against RLd 1 tissue culture target cells. Young nude mice with random-bred genetic backgrounds (NIH and Swiss), the type most frequently used b y investigators, have had quite high levels of reactivity. Nude mice which were obtained from the Animal Production Section of NIH after several backcross generations with inbred mice also were TABLE 111 NATURALCELL-MEDIATED CYTOTOXIC REACTIVITY OF SPLEEN CELLS FROM VARIOUS STRAINS OF NUDE MICE AND OF LYMPHNODE CELLS FROM LASAT

MICE

Strain of nude mice

Median percentage of cytotoxicitya

NIH Swiss BALBIc CBNN C3H C57BIJ6 Lasat

30 35 25 50 25 20 57

Effector cell :target cell ratio of 50 : 1 against RL8 1 tissue culture.


315

highly reactive. Nude mice on a CBA background were somewhat more reactive than other nude mice, but the differences were not as striking as that seen with conventional CBA mice versus some other strains. Lymph node cells from lasat mice (Lozzio, 1976; Lozzio et al., 1976; Machado et al., 1976), which have the gene for asplenia as well as nu/nu genotype and are therefore asplenic as well as nude and have a more profound immunologic deficit, involving B cell as well as T cell functions, have been shown to have very high reactivity. The findings of high reactivity in nude and lasat mice with profound immunologic deficiencies raised the question of whether the observed high reactivity of CBNN might also be related to their defective B cell responsiveness to some antigens (Scher et al., 1975). We therefore compared the levels of NK activity in CBA/N mice with CBA/H and other CBA lines which do not carry this genetically determined defect (Table 11).All of the CBA sublines were found to be highly reactive, and it therefore seems unlikely that the B cell competence of the mice has a major influence on this phenomenon. 2. Rats In contrast to the extensive studies in mice of the influence of genetic background on NK activity, only a small amount of information is available on the variation in reactivity among strains of rats. Most of the studies have been confined to W/Fu rats (Nunn et al., 1976; Shellam and Hogg, 1977; Oldham et al., 1977). Random-bred SpragueDawley rats were shown to have activity against the (C58NT)D tumor which was comparable to that of W/Fu rats (Nunn et al., 1976). Shellam and Hogg (1977) tested a variety of strains, with different Ag-B genotypes, against another target cell, WIFuG-1, and found that most strains had comparable levels of reactivity. However, BDIX and (BDIX x W/Fu)F1 rats had lower reactivity, and BN rats were completely unreactive. Dr. J. R. Oehler in our laboratory has also observed low or absent reactivity in BN rats. It is of interest to note that the (BDIX x WlFu) hybrid rats had reactivity similar to that of the low reacting parent rather than that of WlFu, since, in analogous studies in mice, high reactivity appeared to be a dominant trait (see Section II,B,l). In contrast to the findings with the BDIX x W/Fu)F, hybrids, Williams et al. (1977) found that F1 hybrids of low reacting BN rats and WIFu rats had high NK activity against several clones of SV40transformed BN fibroblasts. Hybrids of BN with two other strains also had higher activity than BN. It is not clear whether these disparate results are due to the different strains used or to the different target cells.

316


3. Human There have been several reports of an association between reactivity and HLA phenotype. Petrinyi et al. (1974) initially reported that individuals with the HLA-A3, B7 haplotype had decreased cytotoxic activity. This has been confirmed and extended by Trinchieri and his associates (Santoli et al., 1976; Trinchieri et al., 1977). Male donors with HLA-B7 had significantly depressed activity and those with the HLA-A3, B7 haplotype had quite low activity. Significantly elevated reactivity was associated with HLA-B12. For no apparent reason these correlations were only significant for male donors, and female donors had activity only Vz to 2/3 that of male donors. It was also of interest, in regard to the possible in vivo role of natural cytotoxicity (see Section VIII), that the HLA-AS, B7 haplotype has been associated with an increased incidence of multiple sclerosis. The only other suggestion for a role of genetic factors in expression of natural cytotoxicity has come from the studies of Hellstrom et al. (1973), which indicated that adult black donors had significantly more reactivity than white donors in a visual microcytotoxicity assay against cultured melanoma cells. However, this report was based on tests of a small number of donors, and other investigators have failed to find a significant race-related difference in the reactivity of normal donors (e.g., Oldham et al., 1975).

c. INFLUENCE

OF

ENVIRONMENTAL FACTORS AND DISEASE

1. Mice The time course of spontaneous appearance of cytotoxic reactivity in mice within 8 weeks of age, persistence for only 2 to 4 weeks and then decline to low levels, is similar to that seen after immunization with alloantigens (Canty and Wunderlich, 1970)and tumor antigens (Lavrin et al., 1973). This suggested that most mice are exposed to antigen(s) within a few weeks after birth, which induced cytotoxicity shortly thereafter. Initially, environmental factors did not appear to play a major role in the kinetics or levels of natural cytotoxicity (Herberman et al., 1975a). Conventional and nude mice, whether raised conventionally or pathogen-free, seemed to have comparable reactivity. Furthermore, mice coming fiom a wide variety of sources developed similar levels of activity at approximately the same age. Because of these observations, it was suggested that endogenous factors in the mice, probably activation of endogenous type-C viruses, were responsible for producing sensitization. Recently we have noted considerable fluctuations in levels of activity in nude mice, partially associated with the source of the mice, the place for housing prior to testing, and


317

the timing of transport to the testing laboratory. In addition, the general health of the nude mice when tested has sometimes been found to influence reactivity, with sick mice usually being less reactive than healthy mice. When differences have been noted, they have generally been in the direction of low activity at a time when high, peak reactivity was expected. The exact nature of the environmental factors producing these variations in nude mice has not yet been determined. However, these recent observations indicate that environmental factors may play an important role in either depressing or augmenting the innate cytotoxic reactivity of nude mice, and perhaps also of conventional mice. Mice bearing tumors have been found to have depressed levels of NK activity. Herberman et al. (19754 found that mice bearing murine sarcoma virus (MSV)-induced primary tumors had lower reactivity than uninoculated mice, Becker and Klein (1976) have confirmed this observation, and they obtained similar findings with conventional mice bearing a syngeneic methylcholanthrene-induced sarcoma and with nude mice bearing tumors induced by human lymphoblastoid cell lines. The finding of some NK activity within MSV tumors (Becker and Klein, 1976) indicates that this depressed reactivity in the spleen or other peripheral lymphoid organs could be due, at least in part, to a shift in distribution of NK cells.

2. Rats There have only been limited studies of the effects of environment on N K activity in rats and no information is available on the levels of reactivity in rats with tumors or with other diseases. Thus far, environmental factors have not been shown to play an important role. W/Fu rats in the United States (Nunn et al., 1976) and in England (Shellam and Hogg, 1977)have been shown to have appreciable levels of cytotoxicity. It is not possible to compare the levels seen in these studies directly since different target cells and other conditions were used. Shellam and Hogg (1977) found similar levels of reactivity in rats tested on the day of arrival in their laboratory and in rats of the same age housed for weeks or months in their animal house. Nunn et al. (1976) compared young germ-free Sprague-Dawley rats with conventionally raised rats and found similar levels of reactivity. Shellam and Hogg (1977) compared conventional and germ-free rats of another strain and also detected no differences.

3. Human It has been difficult to identify environmental factors which influence human NK reactivity. However, some data suggest that environ-

318


mental factors may affect the levels of cytotoxicity. In contrast to the good reproducibility in results when the same donor is tested repeatedly over a short time period, several investigators have noted that some normal donors have considerable changes in levels of NK activity when followed for a long period of time (e.g., Rosenberg et al., 1974; Heppner et al., 1975). Usually there has been no obvious cause for such decreases or increases in reactivity. We have noted that some donors who usually have low levels of reactivity became highly reactive for a period of time after developing viral respiratory infections. This may be related to the augmenting effects of viruses, as discussed below (Section 11,D). The issue of environmental factors has also come up in studies of laboratory workers and of relatives and associates of cancer patients. Takasugi et al. (1973) noted that laboratory workers have a higher incidence of reactivity against various target cells than did other normal donors. We initially thought there was such an association (Oldham et al., 1973) but upon further testing this relationship was not seen (Oldham et al., 1975). Pross and Jondal (1975) also found no difference between laboratory workers and other normal donors in cytotoxic reactivity against the mouse tumor, P815X2. Rosenberg et al. (1972) detected cytotoxic reactivity of relatives of leukemia patients against leukemic target cells. However, other, unrelated normal donors were also reactive and there was no clear difference among donors in the same age range. Byers et al. (1975) described a significantly higher incidence of cytotoxic reactivity in household contacts of patients with osteogenic sarcomas or breast cancer against tumor-derived cell lines, which they concluded was specific for the type of cancer of the contact patient. However, the ALAb cell line was the main breast cancerderived cell line to show this reactivity, and Levin et al. ( 1976a) have found this to be an excellent target cell for more generalized natural cytotoxic reactivity. Although the possibility of increased cytotoxic reactivity in workers exposed to cancer cells and related materials or relatives exposed to cancer patients is intriguing, the data obtained thus far are equivocal or conflicting. Many investigators have compared the incidence of cytotoxic reactivity against tumor cell lines of cancer patients with that of normal donors. In most cases, the studies have focused on cell lines derived from the same type of cancer as that of the patients studied, and the results among studies have been conflicting and difficult to evaluate (Herberman and Oldham, 1975; Stevenson and Laurence, 1975). This may be partly due to the likelihood that several mechanisms of cytotoxicity are measured in such studies, including NK reactivity as


319

well as disease-specific reactivity. To circumvent this problem and attempt to examine NK reactivity of carcinoma patients and of normal donors directly, some investigators have utilized lymphoblastoid, myeloid, or similar target cells (Rosenberg et al., 1974; Pross and Jondal, 1975). McCoy et al. (19734 and Pross and Jondal(l975) have suggested that such assays might be a useful measure of a form of lymphocyte function. Analogous to the decreased NK reactivity in tumor-bearing mice, McCoy et al. (19734 found that many patients with lymphomas, colon cancer, lung cancer, and melanoma had depressed reactivity against F265, a lymphoblastoid cell line, and that the reactivity of breast cancer patients was in the normal range. Similarly, Takasugi et al. (1973) reported that many normal donors were more reactive than cancer patients against a variety of cancer-derived target cells. Takasugi et al. (19774 have further reported that reactivity of cancer patients declined with tumor progression. In a recent study, Cannon et al. (1977) tested lymphocytes from normal donors and breast cancer patients against the myeloid cell line K562, as an indicator of NK reactivity, and against breast cancer cell lines. The reactivity against K562 was similar for each population, whereas (as will be discussed in more detail in Section VII) the breast cancer patients had higher levels of reactivity against the breast cancerderived cell lines. Heppner et at?.(1975) also found that the incidence of reactivity of cancer patients versus normal donors varied with the target cell line. With breast cancer-derived lines, cancer patients were usually more active than normal donors, but with some melanoma lines, the normal donors were frequently more reactive than the cancer patients. Oldham et al. (1975) also studied cytotoxicity against breast cancer and melanoma target cells, and observed that the incidence of reactivity of cancer patients and normal donors against all of the targets was similar.

D. In Vivo AUGMENTATION OF REALITY

1. Mice

The possibility for environmental factors to influence the levels of natural cell-mediated reactivity led to a series of experiments in mice to determine the effects of in vivo challenge by a variety of materials (Herberman et d., 1977a). Using the analogy to the kinetics of cytotoxicity induced by deliberate immunization, we were particularly interested in whether a “secondary” response could b e elicited in mice after decline in their spontaneous levels of cytotoxicity. We

320


found that reactivity in nude BALB/c mice, as well as in normal BALB/c mice, could be substantially augmented by inoculation of a variety of tumor cells. In young mice, the levels of reactivity were increased and in older mice, cytotoxicity rapidly reappeared. This augmented cytotoxicity reached a peak at 3 days after inoculation, and then declined to baseline levels b y 6 to 7 days. Only cells which appeared to bear antigens recognized by the NK cells were found to augment reactivity. Initially, only mouse cells were found to be active. However, after detecting cytotoxic reactivity against the human Chang liver tissue culture cell line (see Section III,A), we tested the ability of these xenogeneic cells to boost and found them to be active. Similarly, after recent indications that allogeneic bone marrow might contain antigens recognized by NK cells, we found that allogeneic bone marrow cells but not allogeneic spleen cells or syngeneic bone marrow cells could significantly augment cytotoxic reactivity. In addition to this apparently specific boosting, a variety of murine viruses, including murine sarcoma virus and lymphocytic choriomeningitis virus (LCMV), were able to cause marked increases in the levels of cytotoxicity in either nude or conventional mice (Herberman et al., 1977a). This effect appears to b e dependent on the ability of the viruses to replicate in the recipients since inactivated virus or viruses unable to infect mouse cells have been inactive. The viruses appeared to induce nonspecific, polyclonal activation of the natural cytotoxic mechanism, and the bacterial adjuvants, BCG and Corynebacteriurn purvum, has similar effects (Herberman et ul., 1977a). Wolfe et al. (1976) have also recently noted the boosting of natural cell-mediated cytotoxicity by BCG. In all of these studies, the characteristics of the effector cells after boosting were the same as those seen with NK cells (see Section IV). Pfizenmaier et al. (1975) observed a similar phenomenon, the rapid and temporary appearance of cytotoxicity against some cultured syngeneic target cells after inoculation of LCMV. On the basis of the effects of pretreatment of the effector cells with anti-8 plus complement, these investigators concluded that T cells were responsible for this activity. Good sensitivity to anti-8 would seem to be at variance with the experience with NK cells (see Section IV), but, in fact, depression of reactivity by such treatment was difficult to achieve (H. Wagner, personal communication). The ability of microorganisms to strongly augment the levels of NK activity may be related to some of the observed environmental effects on cytotoxicity. If mice were tested shortly after environmental exposure to a virus capable of augmenting, high levels might be seen. However, if they were tested several days later, low levels, possibly even below their baseline spontaneous levels, might be observed.


32 1

The augmentation of NK activity by tumor cells bearing the relevant antigens might appear to be at variance with the observation of depressed reactivity in mice bearing primary MSV tumors (discussed in Section I1,C). However, as noted above, the timing may be critical. When mice are tested at 14 days or more after tumor cell or MSV inoculation, this is considerably beyond the period of augmentation. This would probably also account for the failure of Greenberg and Playfair (1974) to detect augmentation of natural reactivity in young NZB mice after inoculation of tumor cells, since they only looked at 12 days post inoculation. 2. Human There is very little information as yet on the ability to boost NK activity in rats or in man. It will be important to evaluate carehlly the effects on NK reactivity of inoculation of tumor cells, BCG, C. parvurn, and other agents which are being used for immunotherapy of cancer patients. Some evaluation of NK reactivity in patients receiving immunotherapy has been performed (e.g., Oldham et al., 1976a), but no studies have been reported on sequential daily testing of cytotoxicity after inoculation. This type of detailed kinetic study would probably be needed to detect augmentation analogous to that observed in mice. Based on the ability of several viruses to augment NK reactivity and the anecdotal observations of increased cytotoxicity in some donors with upper respiratory viral infections, we have recently performed a study on the effects of inoculation with swine influenza vaccine (W. H. West and R. B. Herberman, unpublished observations). A series of normal adult volunteers were tested for cytotoxicity against K562 target cells prior to vaccination and at frequent intervals thereafter. In the majority of donors, a significant increase in reactivity was seen 1to 3 days after inoculation. Thus, it appears that human NK activity can be augmented in a fashion quite analogous to that seen in mice.

E. EFFECTSOF in Vitro CULTIVATION

OF

LYMPHOID CELLS

In many studies, it has been of interest to test for NK reactivity after lymphoid cells were placed in culture for several days. In addition, many of the cytotoxicity assays involve incubation periods of 1 to 2 days. When reactive cells are cultured in the presence of tumor antigens or other possible stimulants, one might anticipate that in vitro augmentation of NK activity would result, The results of such studies have been complex, with different effects being observed in each species studied.

322


1. Mice One of the distinguishing characteristics of NK activity in mice has been its lability at 37°C. Incubation of reactive cells for 2 hours or more at 37°C has resulted in a substantial fall in cytotoxic activity (Herberman et al., 1975b). In contrast, such incubation of cytotoxic immune T cells has had no inhibitory effect (Herbermanet al., 1975b). The augmented reactivity seen after inoculation of tumor cells or microorganisms has also exhibited this lability (Herberman et al., 1977a; Wolfe et al., 1976). The loss in NK reactivity has not been accompanied by a large decrease in viable cells, and addition of 2-mercaptoethanol to the culture medium, which enhances in vitro survival of mouse lymphoid cells, has had no effect. The mechanism for this loss of functional activity, which has not been seen with rat or human cells, remains unclear. However, this lability of the effector cells probably accounts for our failure thus far to observe any evidence for in vitro augmentation of NK reactivity when we have used any of the procedures which work well with rat or human cells. Because of the lability of mouse NK activity, it has been essential to control carefully all in vitro manipulations with reactive cells to distinguish the lability from the effect of the particular treatment. Shustik et nl. (1976) were successful in generating cytotoxicity in normal cells cultured for 5 days, and the levels achieved were comparable to those achieved by sensitization against allogeneic stimulating cells. Gorczynski (1976a,b) had similar results, with cultured male spleen cells having considerably more cytotoxicity than those of female donors.

2. Rats Incubation of normal rat lymphoid cells at 37°C prior to testing has not resulted in a loss of reactivity but rather has been associated with significant increases in cytotoxicity. Shellam and Hogg (1977) observed that preincubation of cells for 3 hours resulted in consistent augmentation of cytotoxicity. Glaser et al. (1976a) measured at daily intervals cytotoxic reactivity of normal lymphoid cells cultured for up to 10 days. After 1 day of culture, normal spleen cells were significantly more cytotoxic than before culture; this cytotoxic reactivity then decreased to negligible levels with further time in culture. Culture of normal cells with (C58NT)D lymphoma cells resulted in persistent reactivity against that tumor, with a further rise at days 6 and 7. However, the characteristics of this augmented reactivity, and its relationship to NK reactivity versus specifically immune antitumor


323

cytotoxic reactivity were not evaluated. Also the similarity of the augmented reactivity after several hours to 1 day of culture in medium alone to the in vitro augmentation of immune rat spleen cell cytotoxicity (Ortiz de Landazuri and Herberman, 1972a) has not been thoroughly expolored. Glaser et al. (1976~)found that incubation of rat spleen cells for 1 day in the presence of phytohemagglutinin (PHA) or endotoxin (LPS) resulted in a considerable augmentation in cytotoxic activity against a syngeneic tumor, (C58NT)D, and also against other syngeneic and xenogeneic target cells. These results are similar to the augmentation of cytotoxic reactivity of human lymphocytes by in vitro incubation with mitogens or other stimuli (see Section II,E,4). However, there were some observations with the mitogen-treated rat cells which complicate interpretation of the mechanism. The stimulated spleen cells reacted against a broader array of target cells than did unstimulated normal spleen cells, and different lymphoid cells appeared to be involved in generation of augmented cytotoxicity b y PHA as compared to LPS (see Section IV,B for more detailed discussion of the nature of effector cells).

3. Guinea Pigs There have been no reports, to our knowledge, of natural cellmediated cytotoxicity in guinea pigs. However, some recent studies of Dr. A. Altman (unpublished observations) with cultured normal guinea pig lymph node cells seem quite similar to those described in this section for cultured lymphoid cells of other species. After culture in medium containing fetal bovine serum (FBS), high levels of cytotoxicity against a syngeneic hepatoma cell and some other target cells have been observed. As described below for human PBL, this activity was dependent on culturing of the cells in heterologous serum (either FBS or horse serum) and was not seen in medium containing only serum from the same species (i.e., guinea pig serum).

4. Human As with rat lymphoid cells and in contrast to mouse lymphocytes, incubation of human peripheral blood lymphocytes (PBL) at 37°C has not led to loss of cytotoxic reactivity. Considerable increases in cytotoxicity upon in vitro culture under a variety of conditions have been noted. The kinetics of these increased levels of cytotoxicity have differed from those seen with rat cells, in that activity has usually remained high for 7 days or more of culture. The augmented levels of cytotoxicity have many characteristics which are similar or identical to NK reactivity, but, until the phenomena are shown to be the same,

324


Bonnard (1978) has proposed the term CCC (cytotoxicity from cultured cells). Several investigators have reported that incubation of human PBL at 37°C in medium containing FBS resulted in a rapid increase in cytotoxic reactivity against tumor target cells (Stejskal et al., 1973; Levin et al., 1976b; Zielske and Golub, 1976; Ortaldo et al., 1977a,b). This phenomenon was found to b e highly dependent on the presence of FBS in the culture medium and did not occur in medium containing human serum (Zielske and Golub, 1976; Ortaldo et al., 1977a,b). However, cultures of PBL in the presence of mitogens, antigens, allogeneic PBL, or autologous or allogeneic lymphoblastoid, myeloid or tumor cell lines have also been associated with the development of high levels of cytotoxicity (Stejskal et al., 1973; Svedmyr et al., 1974; Martin-Chandon et al., 1975; Stejskal and Perlmann, 1976; Ortaldo et al., 1977a,b; Morales et al., 1977). With the mitogens and soluble antigens, the PBL were washed free of the stimuli prior to assay and therefore the effector phase was not dependent on the presence of the stimulant. There has generally been a good correlation between the development of a lymphoproliferative response and augmentation of cytotoxicity, and, since FBS as well as mitogens and antigens can stimulate proliferation, the same mechanism may underlie all of these observations. 111. Specificity of Natural Cell-Mediated Cytotoxicity

A. MICE

In virtually all of the studies of natural cell-mediated cytotoxicity, the investigators detected some degree of specificity in the phenomenon. However, the conclusions reached regarding the nature of the antigens have varied considerably depending on the tumor cells tested. In several studies, the detected antigens appeared to b e associated with type-C viruses (Herberman et al., 1975a; Kiessling et al., 1975a; Sendo et al., 1975; Zarling et al., 1975; Lee and Ihle, 1977). Kiessling et al. (1975a) tested for NK activity against a limited number of target cells and concluded that susceptibility to lysis was restricted to lymphomas induced by Moloney leukemia virus. However, in more recent studies in this laboratory (Becker et al., 1977; Kiessling, personal communication), this restriction was not found to hold. Becker et al. (1977) found no correlation between susceptibility to NK activity and any serologically defined group of type-specific antigen associated


325

with type-C viruses. Furthermore, we (Herberman et al., 1975a) and others (e.g., Becker et al., 1976) have not seen any restriction of susceptibility to homologous H-2 type. Zarling et al. (1975) suggested that Gross leukemia virus antigens or embryonic antigens were being dected. Sendo et al. (1975) suggested that NK activity was directed against X. 1’ antigen, associated with expression of an endogenous type-C virus. Similarly, we have suggested (Herberman et al., 1975a) that NK activity is directed against several antigens associated with murine endogenous type-C viruses. Blair and Lane (1975a,b) and Gillette and Lowery (1976)have described reactivity in microcytotoxicity or cytostasis assays against mouse mammary tumor virus-infected target cells. In contrast to these reports of reactivity against virus-associated antigens, Small and Trainin (1975) and Pfizenmaier et al. (1975) have observed reactivity against autoantigens which bears some resemblance to NK activity. To determine the specificity of NK activity, two approaches have been taken. The first has been to test directly a variety of target cells for susceptibility to lysis. The other, which has provided more detailed information on the heterogeneity of the detected antigens, has been to test a wide variety of cells for thier ability to inhibit release of 51Cr from labeled target cells.

1. Direct Testing In almost all of the studies referred to above, normal lymphoid cells were tested against lymphoma target cells, and it has been widely assumed that only lymphoid target cells are susceptible to rapid lysis b y NK cells. However, we have recently found that many nonlymphoid target cells, harvested directly from in vivo tumors or from cultured cell lines, are susceptible to NK activity in a 4-hour 51Crrelease cytotoxicity assay (Table IV). Some sublines of 3T3, including the MSV transformed nonproducer line KA31 and SV40 transformed lines, were good target cells for nude spleen cells. Although most of the studies have been performed with established tumor cell lines, we have found that primary spontaneous thymomas of AKR mice are quite susceptible to NK activity. The low but significant levels of reactivity against the untransformed 3T3 cell line indicated that sensitivity to NK activity was not limited to transformed cultured cells. Similarly, Shustik et al. ( 1976) found that cultured lymphocytes developed high reactivity against cultures of syngeneic and allogeneic embryo fibroblasts. As discussed earlier (Section II,D), Pfizenmaier et al. (1975) observed transient cytotoxicity, against short-term cultures of syngeneic normal

326

RONALD B. HERBERMAN AND HOWARD T. HOLDEN TABLE IV NATURALCYTOTOXICITY O F NUDE SPLEEN CELLS AGAINST NONLYMPHOID CELL LINES Target cell

3T12 (embryo fibroblasts, transformed) 3T3 (embryo fibrobIasts) Type I CI 6 (spontaneously transformed) KA31 (MSV transformed, nonproducer) SV40 transformed: SV3T3 SVA 31C 14 E4 T-AUN lines (fibroblasts, spontaneously transformed) T-AUN C1, T-S-1 (SV40+) T-S-5 (SV40-) B16 (melanoma) LM5F-22 3LL (Lewis lung tumor) Mammary cancer cell lines TA3 L8a a

Percentage of cytotoxicitya 45

5 3.6 10 24 14 9

31 22 40 10 8 10 2.7 0.5

Effector cell :target cell ratio of 100: 1.

macrophages, after inoculation of mice with LCMV. Since most of the characteristics of their observations were compatible with augmented NK activity, it was of interest to examine directly the susceptibility to NK cells of overnight cultures of peritoneal macrophages. We found that spleen cells from NIH nude mice, and from CBA mice, 4 days after inoculation with LCMV, had significant levels of activity against these target cells. We have recently also observed low levels of NK activity against mouse PHA blasts. As already discussed earlier (Section II,B), the susceptibility to NK activity of a given tumor cell has been found to vary considerably with the growth conditions. In uiuo transplanted tumors have been more resistant to NK activity of most strains than in uitro cultures of the same tumors (Kiessling et al., 1975a; Herberman et al., 1975a), and some cultured cells have been much more sensitive to lysis (Kiessling et al., 1975a). This is consistent with the recent observations of Aoki et al. (1977) that several virus-associated antigens varied widely in expression, depending on whether the cells were grown in uitro or in uiuo. It initially appeared that mouse NK activity was restricted to mouse

327


target cells. However, recently we have detected considerable reactivity of nude spleen cells against some human target cells but not against others (Herbermanet al., 1977a).The human Chang liver tissue culture cell line has been particularly sensitive to lysis. Similarly, R. Kiessling and his associates (personal communication) have observed lysis of the MOLT-4 cell line derived from a patient with acute lymphocytic leukemia. We have also observed some reactivity of nude and other mouse spleen cells against the rat lymphoma, (C58NT)D,which is also a susceptible target cell for rat NK activity (Nunn et al., 1976). We have also observed recently that mouse NK activity is not completely restricted to tumor or cultured target cells. The finding that normal mouse thymocytes could boost cytotoxic reactivity led us to test thymocytes as target cells (Herberman et al., 1977a). Table V shows the results obtained in tests of BALB/c nude spleen cells against syngeneic and allogeneic thymocytes and other normal target cells. Significant reactivity was observed against thymocytes and bone marrow cells but not against spleen or lymph node cells.

2. Inhibition of Cytotoricity Assays

The above results of direct tests with a variety of target cells indicate that many tumor cells and some normal cells are susceptible to N K activity. The resistance of some target cells, including some known TABLE V NATURALCYTOTOXICITY OF BALBlc NUDE SPLEENCELLS AGAINST NORMALTARGETCELLS Target cells

Tissue Thymus Thymus Thymus Thymus Thymus

Bone marrow Bone marrow Spleen Spleen Spleen Lymph nodes Lymph nodes

Strain BALBlc BALBlc C57 B L16 AKR NZB BALBlc WIFu rat BALBlc C57BLl6 NZB BALBlc C57BU6

Effector cell:target cell ratio of 200: 1.

Age of donor (weeks) 0 8 8 12 6

8 6 8 8 6 8 8

Percentage of cytotoxicity"

9 12 13 20 9 6 4 -1 1 1 -1 0

328


to be susceptible to immune cytotoxic T cells, to lysis by normal lymphoid cells suggests that the reactions are specific. However, these data do not allow one to determine whether all susceptible target cells share the same antigen or whether several antigens are involved. To analyze the specificity of NK activity in more detail, we (Herberman et al., 1975a) and others (Kiessling et al., 1975a; Sendo et al., 1975; Zarling et al., 1975) have utilized the assay of inhibition of W r release cytotoxicity (Ortiz d e Landazuri and Herberman, 197213; Herberman et al., 197613). With each labeled target cell used in the assays, different antigenic specificities were determined. Because most of the tumor cells studied had some expression of endogenous type C viruses, it was postulated that the antigens detected by NK cells were associated with murine endogenous type-C viruses and these were designated MEV-SA2, 3 and 4 (Herberman et al., 1975a). In the inhibition studies of Sendo et al. (1975), an association of the antigens with endogenous type-C virus was also suggested. Consistent with this hypothesis, Lee and Ihle (1977) have found that NK activity of (B6C3)F1 mice could be specifically inhibited by gp69/71, the major envelope glycoprotein, of AKR endogenous leukemia virus. It therefore seems likely that murine endogenous type-C virus-associated antigens account for at least some of the NK activity. However, with the recent findings of NK activity against some xenogeneic tumor cells and some normal cells, it seems likely that additional, nonviral specificities are also involved. More inhibition studies, using a range of different target cells, are needed to sort out the distribution and nature of each antigen recognized by NK cells.

3. Examination of Tumors in Nude Mice for Antigenicity We are currently studying tumors which arise in nude mice for their susceptibility to NK activity and for the presence of the relevant antigens of these tumors. We received from Dr. 0. Stutman, under code, three tumors induced by methylcholanthrene. Two tumors were induced in CBA/H nude mice and the third in a normal CBA/H mouse. By direct testing and by the inhibition assay, the two tumors from nude mice lacked NK-related antigens, whereas the other tumor was positive. We have also received from Dr. H. Outzen five lymphocytic leukemias which arose spontaneously in BALB/c nude mice and three sarcomas of BALB/c nude mice induced by methylcholanthrene. All of these tumors also lacked NK-related antigenicity. The significance of this observation is somewhat confounded by the lack of antigenicity on several of Outzen’s spontaneous and induced tumors of normal BALB/c mice. It remains of interest, however, that none of the lym-


329

phomas or sarcomas of nude mice studied so far has been susceptible to NK activity. As will be discussed later (Section VIII), this has important implications regarding the possible role of NK cells in immune surveillance. B. RATS

The studies of natural cell-mediated cytotoxicity in rats have also supported the specificity of the phenomenon. The approaches to analysis of specificity, i.e., direct testing and inhibition assays, and the pattern of results have been very similar to those in mice. However, in contrast to the studies in mice, little direct testing for specificity has been done, but rather the inhibition assays have been mainly relied upon. 1. Direct Testing

In the studies of Nunn et al. (1976), only the syngeneic (C58NT)D ascites tumor cell was used. Oldham et al. (1977) tested (C58NT)D ascites and tissue culture cells as targets and also the ERTh/V-G tissue culture line. In the short-term CRA, good reactivity was seen against the ascites (C58NT)D line but not against the cultured (CS8NT)D cells. The resistance of the cultured cells to NK activity could not be attributed to a general resistance to lysis since immune spleen cells reacted well against this line (Ortaldo et al., 1976; Oldham et al., 1977) and natural cytotoxicity could be detected against it in assays with longer incubation periods (Oldham et al., 1977). It is of interest to note, however, that the increased susceptibility of an ascites target compared to cultured cells is the opposite of that usually observed with mouse target cells (as discussed above, Section III,A,l). With another target cell, W/FuG-1, Shellam and Hogg (1977) found that the cultured cells were considerably more susceptible to rapid lysis than were the ascites cells. These investigators also tested a variety of other rat tumor cells and mouse cell lines for susceptibility to NK activity. Only rat tumors induced b y Gross or Moloney leukemia viruses were susceptible, and other rat tumor cells, normal rat lymphoid cells, and mouse tumor cells, including some induced by leukemia viruses, were resistant. Glaser et al. (1976~)tested the direct specificity of spleen cells cultured for 1day in medium alone or in the presence of PHA or LPS. The cells cultured alone reacted mainly against (C58NT)D, had low reactivity against another syngeneic line, LW-6, and no significant reactivity against mouse tumor cells. In contrast, the mitogenstimulated cells reacted well against all of the target cells tested. The

330


extent of reactivity of rat NK cells against mouse tumor target cells needs to be more thoroughly investigated. We have recently found that RL81 mouse cultured cells are very sensitive to cytotoxicity by normal rat spleen cells. The human Chang and K562 cell lines were also sensitive to lysis b y rat NK cells. 2. Inhibition of Cytotoxicity Assays Nunn et al. (1976) tested a large number of rat and mouse cell lines for their ability to inhibit cytotoxicity of 51Cr-labeled(C58NT)D target cells. These studies demonstrated that the NK reactivity had specificity, with some cells inhibiting well and others giving little or no inhibition. The specificities of the immune reactivity against (C58NT)D and NK reactivity were found to be similar, but differed from any known serologic antigens. A variety of rat cells shared the antigens, but virusinduced mouse leukemias, including RBL-5 which is susceptible to mouse NK activity, were negative. This pattern of species-restricted specificity was analogous to MEV-SA-1,2,3, and 4 detected in studies of mice (Herberman et al., 1975a). Similarly, since there was a correlation between ability to inhibit and evidence for expression of rat endogenous type-C virus, it was suggested that the antigens detected were associated with rat endogenous viruses. Shellam and Hogg ( 1977)confirmed and extended these observations. They found inhibition with rat lymphomas induced b y murine leukemia viruses, with rat fibroblasts deliberately infected with these viruses, and with rat cell lines known to express rat endogenous virus or virus-associated antigens. Other nonvirus-induced tumors, tumors induced by pol yoma virus, and normal adult or fetal cells were negative. The mouse tumor cells tested were also negative. In addition to the inhibition assay, Shellam and Hogg (1977) performed adsorptions of effector cells on different monolayers and then tested for residual cytotoxicity against W/FuG-1. The same pattern of specificity was seen in these tests. They also tested murine sarcoma virus, feline leukemia virus, and Rous sarcoma virus for their ability to inhibit NK reactivity and found that only the murine virus gave a dose-responsive inhibition. However, this inhibition did not fit the species restriction seen with intact cells, and Shellam and Hogg cautioned that disrupted virus preparations may cause nonspecific inhibition. It should be noted that the evidence for species restriction of antigens was obtained when rat tumor cells were used as labeled targets. The recent finding of direct cytotoxicity of rat cells against mouse RL81 target cells requires that this issue be examined further to determine if RL81 cells can inhibit cytotoxicity of rat target cells or


33 1

whether a separate antigenic specificity is being recognized on the heterologous cells.

C. HUMAN 1. Direct Testing Many investigators have tested lymphocytes from normal donors against a variety of human cell lines derived from tumors. In early studies, histologic type specific cytotoxicity of cancer patients was described, and reactivity b y normal donors was not noted (e.g., Hellstrom et at., 1971; Fossati et al., 1971; Sinkovics et al., 1971; O’Toole et al., 1972; Levy et d., 1972; Heppner et a l . , 1973). When investigators began to take notice of natural cytotoxicity, most described it as nonspecific (e.g., DeVries et al., 1975; Bukowski et al., 1976; Jondal and Pross, 1975; Hersey et al., 1975; Pierce and DeVald, 1975; Peter et al., 1976a,b). However, several investigators have recognized that lack of histologic type specific cytotoxicity should not be equated with complete lack of specificity, but rather that specificity or lack thereof needs to be carefully evaluated b y large “checkerboard” experiments (Herberman and Oldham, 1975). Klein has proposed (Bean et al., 1975) a useful categorization of the results of such tests and we will use this nomenclature in our discussion. She defined diseaserelated cytotoxicity as cytotoxicity solely against specific tunior target cells b y cells from patients with that type of cancer, selective cytotoxicity as cytotoxicity for some target cells of other histologic types but not for all target cells tested, and nonselective (presumably nonspecific) cytotoxicity as killing of all target cells tested, regardless‘ of histologic type. Most recent studies have described a lack of disease-related specificity, mainly because of natural cytotoxicity (e.g., Oldham et al., 1973, 1975; Takasugi et al., 1974; DeVries et al., 1975; Berkelhammer et al., 1975; Pavie-Fischer et al., 1975; Heppner et al., 1975; Canevari et at., 1976; Bukowski et al., 1976). However, several investigators have found selectivity in their cytotoxicity results (Matthews and MacLaurin, 1974; Rosenberg et aZ., 1974; Oldham et at., 1975 Pross and Jondal, 1975; Heppner et al., 1975). Others, on the basis of direct testing against many cell lines, have characterized their results as nonselective (Kiuchi and Takasugi, 1976; Takasugi et al., 1977b; Bakacs et al., 1977). In a workshop in which a number of groups of investigators performed tests with the same sources of effector cells and target cells, virtually all found some evidence for selec-

332


tive reactivity by some donors (Bean et al., 1975). Takasugi’s group, while not finding selective cytotoxicity when analyzing their data by conventional means, have described selectivity revealed by a complicated procedure, termed interaction analysis (Kiuchi and Takasugi, 1976; Takasugi et d.,1977b). From those studies, they concluded that reactions were directed against at least two, widely distributed common antigens, TA1 and TA2 (Takasugi et al., 197713). Using a different procedure, Cannon et al. (1977) have also qualified the distinction between selectivity and nonselectivity b y demonstrating quantitative selectivity or relatively high cytotoxicity of some donors against breast cancer-derived cell lines. From all of these studies, it can be concluded that human natural cytotoxic reactivity, like that of rodents, is not nonspecific but rather appears to be directed against antigens which are represented on many different cell lines of various histologic types. An important observation has been that human PHA blasts are resistant to NK activity (Ortaldo and Bonnard, 1977), which probably accounts for the ability to detect specific histocompatibility antigen associated cytotoxicity with these targets. In one study (Santoli et d., 1976), reactivity seemed to be restricted to human and monkey cell lines, with little or no reactivity against mouse, rabbit, or hamster cell lines. However, others (PetrLnyi et al., 1974; Pross and Jondal, 1975) have detected strong reactivity against some mouse target cells. The detected antigens have been shown not to be due to fetal bovine serum, since target cells grown for some time in medium with human serum were also susceptible to cytotoxicity (McCoy et al., 1973b; Rosenberg et al., 1974). 2. Inhibition of Cytotoxicity Assays To analyze further the specificity of the antigens recognized by human NK cells, a few studies have been done with inhibition of cytotoxicity assays. In initial studies with the F-265 lymphoid cell line as target cells (McCoy et a1., 1973b; Rosenberg et a1., 1974), inhibition was seen with F-265 and another lymphoid cell line, whereas with a third lymphoid cell line, human PBL, thymus cells, and erythrocytes were negative. Ortaldo et al. (1977~)performed a more extensive study with the K562 target cell. A broad but reproducible pattern of specificity was demonstrated, with most but not all human established tissue culture cell lines being inhibitory. Some fresh single cell suspensions from human malignant tissues also were positive. Some other human established cell lines, leukemic blast cells, and normal PBL were negative. Of interest, in contrast to the species restriction generally found in the inhibition assays with rodent NK and labeled target


3.33

cells, positive results were not limited to human cells; some mouse and rat lymphoma cells inhibited the cytotoxicity against K562. Takasugi et al. (1977b) have performed more extensive inhibition studies using several different target cells. They have also concluded that specific antigens are being detected and that their results support the expression of either one or both of two antigens, TA1 and TA2, on many cell lines. IV. Nature of Effector Cells

Much attention has been directed toward an understanding of the nature of NK cells. The findings of high levels of cytotoxicity reactivity in nude mice, the resistance of rat NK activity to anti-T cell sera plus complement, and the presence of cytotoxic reactivity in human PBL not readily forming rosettes with SRBC indicated that natural cellmediated cytotoxicity was mediated by effector cells different from the well-studied cytotoxic T lymphocyte, produced b y deliberate immunization in uiuo or in uitro (Cerottini and Brunner, 1974). Extensive investigations have further indicated that NK cells are not macrophages or B cells, but may be a subpopulation of T cells or pre-T cells. In this section, w e will review the large body of information which has been obtained in this area.

A. ORGANDISTRIBUTION 1. Mice

NK cells have been detected in most of the lymphoid organs of mice, with particularly high activity in the spleen, lymph nodes, and peripheral blood and low activity in the bone marrow and peritoneal cavity (Herberman et al., 1975a; Kiessling et al., 1975a). Only thymocytes were found to lack activity consistently. In our experience, the levels of reactivity in spleen and lymph nodes of individual mice were similar. In contrast, Greenberg and Playfair (1974) found that activities in these two organs did not correlate, and that the levels in peritoneal exudate cells were similar to those of lymph node cells. The reason for the discrepancy is not clear, but it may be related to the older age of the NZB mice studied by Greenberg and Playfair (1974). After boosting with tumor cells or microorganisms, high levels of augmented reactivity have been detected in spleen, lymph nodes, and peritoneal exudates (Herberman et al., 1977a; Wolfe et al., 1976). In a recent study, we have found that inoculation with LCMV also produced very

334


high levels of reactivity in the bone marrow. This has important implications regarding the possible relationship of N K activity to bone marrow transplantation resistance (Section VII1,B) and the possible origin of NK cells from bone marrow stem cells (Section VI). Because of the latter possibility, we have examined the time of appearance of augmented reactivity in the bone marrow relative to that in the spleen. The kinetics for both organs were the same, with only a small increase after 1 day, and high levels on days 2 and 3. Similarly, we have found that reactivity can be first detected in the bone marrow of young mice at the same time as in the spleen.

2. Rats NK cells have also been found to be widely distributed among the lymphoid organs of rats. Nunn et al. (1976) found the highest levels of reactivity in spleen and lymph nodes, but also detected significant reactivity with peripheral blood lymphocytes, thymus cells, and peritoneal cells. This reactivity of thymus cells was not observed with mouse or human thymus cells and also has not been confirmed by other studies of natural cytotoxicity in rats (Shellam and Hogg, 1977; Oldham et al., 1977). This discrepancy needs to be reexamined. I t is possible that the observed cytotoxicity with rat thymus was due to contamination by adjacent mediastinal lymph nodes. Shellam and Hogg (1977) also tested bone marrow cells and thoracic duct lymphocytes and found no significant reactivity, even after preincubation at 37°C.

3. Human In almost all studies of human natural cell-mediated cytotoxicity, only PBL were tested. In collaboration with Drs. W. H. West and J. L. Weese, we have recently examined the distribution of NK cells in various lymphoid organs. Reactivity was readily detected with spleen cells, but thymus, lymph nodes, and tonsil had little or no cytotoxic activity. B. CELL SURFACE MARKERS AND OTHER CHARACTERISTICS OF

NK CELLS 1. Mice

The mouse NK cell generally has been thought to b e a non-T cell, particularly because of the high levels of reactivity in nude mice. The most important experiments to examine this question directly have


335

involved the pretreatment of lymphoid cells from conventional and/or nude mice with anti-8 serum plus complement (Herberman et al., 1973, 197513; Gomard et al., 1974; Kiessling et al., 197513; Sendo et al., 1975).In none of the experiments with conventional, thymus-bearing mice was reactivity specifically decreased b y such treatment. Under the same conditions, immune cytolytic T lymphocyte activity was completely eliminated (Herberman et al., 1973). With spleen cells from nude mice, treatment with anti-8 plus complement resulted in a partial reduction in activity (Herberman et al., 197513). This effect appeared to be dependent on complement and on the presence of anti-BC3H (or Thy 1.2) antibody, since absorption with Thy 1.2 positive brain removed the inhibitory effect. Inhibitory activity of such treatment has been dependent on the use of high concentrations of antiserum, optimal amounts of rabbit serum as complement source, and optimal length of incubation. Kiessling et al. (1975b) failed to affect nude reactivity by anti-8 treatment, but they used diluted guinea pig serum as the complement source, We have repeatedly confirmed the partial inhibitory effects of anti-8 serum plus complement on nude spleen cell reactivity, and have seen as much as a 90% reduction in activity. We have also observed a similar effect on the reactivity of spleen cells from both nude and conventional mice after boosting (Herberman et al., 1977a).The cytotoxicity of normal spleen cells after culturing in vitro for 5 days was also partially susceptible to anti-8 serum plus complement (Shustik et al., 1976). It thus appears that at least some of the NK cells in nude mice and in mice after boosting have low but detectable expression of 8 antigen. The only concern relative to this conclusion is that the effects may have been due to another antibody in the antisera. Our absorption studies with brain tissues (Herberman et al., 1973, 1975b) have ruled out the possible effect of autoantibodies which may be present in such antisera (Dennert and Lennox, 1972). However, these antisera may also contain antibodies to other lymphocyte antigens (Frelinger and Murphy, 1976) and to virus-associated antigens. We have therefore recently examined the effects of an anti-Thy 1.2 serum, prepared in congenic mice and kindly provided by Drs. Boyse and Shen. Even after absorption to remove antiviral reactivity, this antiserum produced a 50% decrease in the reactivity of BALB/c nude spleen cells. Treatment of NK cells with a variety of other antisera plus complement has not produced any decrease in reactivity. Greenberg and Playfair (1974) observed no effect by a rabbit anti-T cell serum. Similarly, we have noted no inhibition by two different heterologous antisera to mouse brain, which have considerable reactivity against mouse

336


T cells. We have also obtained no evidence for the presence of T L antigen (Boyse and Old, 1969), Th-B antigen (Yutoku et al., 1976), kappa light chain surface immunoglobulin, or Ia antigen (Niederhuber et al., 1976) on the NK cells. Gorczynski (197613) has reported that a heterologous anti-mouse brain serum plus complement could inhibit the cytotoxicity of cultured normal spleen cells of male mice but had much less effect on cells from female donors. However, treatment of female spleen cells prior to culture led to a considerable increase in the level of cytotoxicity generated in vitro (Gorczynski,

1976~). Very recently, Glimcher et al. (1977) treated NK cells with anti-Ly antisera plus complement to determine whether the reactive cells bore these T cell antigens. The anti-Ly2 antiserum had no effect, but an anti-Ly 1.2 serum abolished NK activity. However, the antigen recognized on the NK cells by this antiserum was shown not to be Ly 1.2, but rather was a unique specificity, provisionally designated NK. It is of interest to note that one of the antisera with potent anti-NK activity was made b y immunization with thymocytes. This is consistent with the NK cell being in the T cell lineage, as was suggested above by the piesence of a low density of 6 antigen. Most studies have indicated that the NK cells are nonadherent and nonphagocytic (Herberman et al., 197513; Kiessling et al., 1975b; Sendo et al., 1975; Zarling et at., 1975). After passage of reactive spleen cells over an adherence column, relative activity has usually been increased. By calculating recovery of total lytic units, we have found almost complete recovery of NK activity from such columns (Herberman et al., 197713). In contrast, in the study of Gomard et al. (1974), activity was much decreased or eliminated by column passage or by treatment with carbonyl iron and magnet. Therefore, the reactivity observed by those investigators appeared to be due to macrophages and is much different &om the NK activity observed by others. Consistent with our observation of no inhibitory effect by pretreatment of NK cells with anti-kappa serum plus complement, Kiessling et al. (1975b) found that passage of reactive cells over an antiimmunoglobulin column resulted in an increase in relative activity. Depletion of cells bearing receptors for complement also had no inhibitory effect on NK activity (Herberman et al., 197513). The initial experiments performed by us (Herberman et al., 1975b) and by Kiessling et al. (1975b)to deplete Fc receptor-bearing lymphocytes were interpreted as having no effect on NK activity. However, after studies in this laboratory (West et al., 1977a) and in others (Peter et al., 1975a) indicated that human NK cells possess Fc receptors, we


337

reinvestigated this issue (Herberman et al., 197%). By absorption of cytotoxic spleen cells from nude or conventional mice on monolayers of sheep erythrocytes plus IgG antibodies to sheep erythrocytes, 50-90% of the total cytotoxic reactivity could b e removed. Parallel adsorption of cells on monolayers of sheep erythrocytes alone or on erythrocyte-antibody monolayers coated with protein A, to block the Fc portion of IgG, resulted in little or no depletion of NK activity. The presence of Fc receptors on the NK cells was confirmed by forming rosettes with IgG antibody coated sheep erythrocytes and then showing that this caused the NK cells to sediment more rapidly in a velocity sedimentation separation procedure. NK cells appear to have trypsin-sensitive sites on their surface which are required for cytotoxic reactivity. Brief exposure of NK cells to low concentrations of trypsin had no inhibitory effect on reactivity (Herberman et al., 1975b). However, Kiessling et al. (1976a) originally observed and we have confirmed that more prolonged treatment (30 to 45 minutes) with higher concentrations of trypsin resulted in a substantial decrease in NK activity as measured in a 4 hour assay. In contrast, cells treated in an identical manner continued to have substantial levels of cytotoxicity in an 18hour assay (A. Santoni, H. T. Hojden, and R. B. Herberman, unpublished observations). Kiessling et al. (1975b) directly studied the morphologic appearance of a highly reactive subpopulation of cells after depletion of most of the phagocytic cells, T cells, and B cells. Almost all of the remaining, highly cytotoxic cells had the morphologic characteristics of small lymphocytes. We have essentially confirmed these observations. In addition, with the kind help of Dr. T. Aoki, highly reactive subpopulations of nude spleen cells and lymph nodes were examined by immunoelectron microscopy. Almost all of the cells were small or medium sized lymphocytes, and a small proportion had detectable patches of 6 antigen.

2. Rats NK cells in rats have been shown to be nonadherent, nonphagocytic cells (Nunn et al., 1976; Shellam, 1977; Oldham et al., 1977). The studies performed thus far have also indicated that the NK cell lacks cell surface markers of mature T cells. Treatment with heterologous anti-T cell antisera plus complement, which eliminated immune cytotoxicity against tumor cells, had no significant effect on the cytotoxicity by normal lymphoid cells (Nunn et al., 1976; Shellam, 1977; Oldham et al., 1977). However, these data do not definitely rule out the T cell nature of NK cells. In the studies with mouse NK cells,

338


only anti-theta serum and not heterologous anti-T cell sera affected cytotoxic reactivity, and then only under certain circumstances. Nude spleen cells and cells obtained after boosting were sensitive to antitheta treatment, whereas lymphoid cells from young normal mice were resistant. The ability to inhibit, with anti-T cell serum plus complement, cytotoxic reactivity of normal spleen cells after stimulation in vitro with PHA (Glaser et al., 1976c) may be an analogous finding. This result is also similar to the sensitivity of cytotoxicity of cultured mouse lymphocytes to anti-8 plus complement (Shustik et al., 1976). Examination of rat NK cells for complement and Fc receptors has also yielded negative results (Nunn et al., 1976; Oldham et al., 1977). However, we have not yet had the opportunity to reexamine the possible expression of Fc receptors on these cells, using the same approach as that described for mouse NK cells. Shellam (1977) has described some other characteristics for rat NK cells: They had no detectable surface immunoglobulin, had the size of typical small lymphocytes in separations by velocity sedimentation at unit gravity, and were moderately sensitive to irradiation. Treatment of NK cells with papain resulted in loss of reactivity, which recovered when the cells were incubated for 4 to 5 hours at 37°C.

3. Human The characteristics of human NK cells have been extensively studied by many investigators. However, the results obtained regarding most of the cell surface markers have been conflicting. Almost the only point of agreement has been that most of the human NK activity is due to nonadherent, nonphagocytic cells (e.g., Peter et al., 1975b; Levin et al., 1975; Hersey et al., 1975; West et al., 1977a). Even in this respect, there has been one report of activity in a visual microcytotoxicity assay by granulocytes (Takasugi et d., 1975). Most investigators have concluded that human NK cells are non-T cells (DeVries et al., 1974; Jondal and Pross, 1975; Pross and Jondal, 1975; Peter et al., 1975b; Levin et al., 1975; Kiuchi and Takasugi, 1976; Bakacs et al., 1977). In all of these studies, cells forming rosettes with sheep erythrocytes (E-RFC) were separated from nonrosette forming cells, and considerable NK activity was seen with the non-E-RFC fractions. However, some of these investigators (DeVries et al., 1974; Levin et al., 1975) and others (Dean et al., 1975) have observed a significant amount of NK reactivity b y E-RFC. West et al. (1977a) found that most of the cytotoxic reactivity against K-562, more than 80% of the total lytic units (Kay et al., 1977), was associated with E-RFC. There are several likely explanations for these divergent re-


339

sults. It has been shown that treatment of E-RFC with ammonium chloride solutions to remove SRBC contamination can result in loss of most NK activity for several hours (Kay et al., 1977). For studies using this procedure, this could have caused a falsely low estimation of reactivity in the E-rosetting population (e.g., DeVries et al., 1974; Jondal and Pross, 1975; Peter et al., 1975b; Kiuchi and Takasugi, 1976). Another, more fundamental explanation for the differing results is that NK cells appear to have low affinity receptors for SRBC (West et al., 1977a). If rosetting is not performed under optimal conditions of time, temperature, erythrocyte to PBL ratio, then many of the NK cells failed to rosette. However, by optimizing the conditions for rosetting or by using neuraminidase-treated SRBC, most NK cells were shown to reside in the E-RFC fractions (West et al., 1977a; Kay et al., 1977). The finding that human NK cells have a characteristic T cell marker but in lower density or lower affinity than the majority of T cells is quite analogous to the demonstration that mouse NK cells have a low density of e antigen. Hersey et al. (1975) found most of their cytotoxic reactivity in non-E-RFC, but they observed that many cells in that fraction reacted with an anti-T cell serum. On the basis of this, and other circumstantial evidence, they concluded that NK cells were activated T cells. Svedmyr et al. (1974, 1975) found that cultured PBL, which became cytotoxic in the presence of autologous lymphoid cell lines (see earlier discussion, Section II,E), lacked detectable receptors for SRBC. However, such activity could be generated from purified E-RFC (Svedmyr et a1 ., 1974) and it was shown that at the peak of stimulation in culture, the blast cells lost E receptors as measured by their assay (Svedmyr et al., 1975). The authors concluded that the cultured cytotoxic cells were of T cell origin. This would fit our earlier suggestion that this form of cytotoxicity is the same as NK activity and that NK cells are in the T cell lineage. The change in ability to rosette could represent a shift in differentiation from mature T cells to pre-T cells with concomitant acquisition of NK activity (see further discussion in Section VI) . There have also been many conflicting reports as to whether human NK cells have either Fc receptors or complement receptors. Some groups reported that the complement receptor was a characteristic marker for these cells (Pross and Jondal, 1975; DeVries et al., 1974). However, others have failed to detect this receptor on a substantial proportion of NK cells (Hersey et al., 1975; Kiuchi and Takasugi, 1976; West et al., 1977a). West et al. (1977a) have shown that rosette formation solely through the complement receptor requires the use of IgM

340


antibodies to SRBC, without contamination by IgG antibodies which can mediate rosette formation through the Fc receptor. It seems likely that the studies reporting the presence of complement receptors used such mixed reagents. The presence of Fc receptors on human N K cells has been readily detected by many investigators (Peter et al., 1975b, 1976a, 1976b; Hersey et al., 1975; Kiuchi and Takasugi, 1976; West et al., 1977a; Kay et al., 1977; Bakacs et al., 1977) using several different separation techniques. By analysis of total lytic units, Kay et al. (1977) found that more than 80% of NK activity was associated with Fc receptor-bearing cells. Ortaldo et al. (1977a,b) have found that cytotoxicity of cultured PBL is also mediated by cells with Fc receptors. Of particular interest was the observation that cytotoxicity associated with Fc receptor-bearing cells was generated even when cells with Fc receptors were removed from PBL before culturing. These data indicated that some cells acquired Fc receptors in culture at about the same time they became cytotoxic. Human NK activity could be inhibited substantially by gamma globulin or by EA antigen-antibody complexes (Peter et al., 1975b; West et al., 1977a). However, soluble antigen-antibody complexes have not been inhibitory (H. D. Kay, unpublished observations). It seems likely that large aggregates or complexes could bind to the Fc receptors on NK cells and sterically interfere with their cytotoxic activity. Alternatively, if natural cytotoxicity is actually a form of ADCC (discussed below, Section V), then blocking of Fc receptors by some types of complexes may even more directly inhibit activity. Some other characteristics of NK cells have been examined. Treatment with cyclic AMP was markedly inhibitory (Rosenberg et al., 1974). Pretreatment of NK cells with trypsin or chymotrypsin has resulted in a loss of most or all NK activity (Kay et al., 1977). in Vitro C. EFFECTOF THYMOSIN

1. Mice

The presence of high levels of NK activity in nude mice and the detection of some 8 antigen on the NK cells suggested that the cytotoxic reactivity might be mediated by prethymic T cells (Pritchard and Micklem, 1973; Scheid et al., 1973; Loor and Roelants, 1974). Incubation of such cells with thymopoietin and other thymic humoral factors has been found to induce detectable quantities of 8 and TL antigens (Scheid et al., 1973; Goldstein et al., 1975). Incubation of nude spleen cells at 37°C with thymopoietin or with calf thymosin did


34 1

not have a detectable effect on NK activity. It is somewhat difficult to relate our negative results with the effects described by others, since, in the one reported study on the effects of thymopoietin on functional activity of lymphocytes (Basch and Goldstein, 1975), only marginal effects on lymphocyte stimulation by mitogens were observed. In addition to the other thymic humoral factors, we have performed experiments with a human thymic hormone preparation of Dr. A. White. Incubation of nude spleen cells with this material at 37°C for 2 hours has produced a small decrease in cytotoxic activity relative to that seen with spleen cells incubated in medium without the factor. Incubation with Thymic Humoral Factor (THF, Umiel and Trainin, 1975) gave similar results.

2. Human Incubation of human PBL whith the human thymic hormone preparation has also resulted in a small but quite consistent decrease in cytotoxic activity against the K562 cell line (W. H. West and R. B. Herberman, unpublished observations). In contrast, incubation of cells under the same conditions with heated factor or with a control protein fraction had no inhibitory effect.

D. EFFECTSOF in Vivo MANIPULATIONSOF THYMUSFUNCTION 1. Mice The high levels of NK activity in athymic nude and lasat mice, and the inhibitory effects of in vitro incubation of nude spleen cells with human thymosin, indicated that the expression of natural cellmediated cytotoxicity might be inversely related to the degree of thymic function. To explore this further, we have performed a series of investigations in vivo to determine the effect of increased or decreased thymic function. Grafting of fetal BALBlc thymuses into young BALB/c nude mice resulted in a considerable restoration of mature T cell function, as evidenced by good lymphoproliferative responses to the mitogens, phytohemagglutinin (PHA), and concanavalin A (Con A). The grafted nude mice, at 6 weeks of age, had similar levels of NK activity as thymus-bearing nu/+ BALB/c littermates, and much lower activity than sham operated nude littermates. In contrast to the strong effects of thymus grafts, daily administration of calf or human thymosin for 3 to 4 weeks, according to the protocol of Ikehara et al. (1975), had no detectable effect on NK activity. These mice also had very little detectable restoration of lymphoproliferative responses to PHA and

342

RONALD B. HERBERMAN A N D HOWARD T. HOLDEN

Con A. This and the weak or undetectable in vitro effects of thymic hormone preparations indicate that these materials lack the potent effects of intact thymus on functional activity. We have also performed thymectomies on newborn BALB/c mice to determine whether early removal of the thymus would mimic the athymic state of nude mice. In mice with partial thymectomy, some increase in NK.activity above that of sham operated was seen, and with total thymectomy, very high levels resulted. Nude mice are usually the offspring of heterozygous, nu/+, mothers and therefore may be affected by humoral thymic factors via the placenta before birth and through the milk thereafter. Hale et al. (1976) showed that homozygous nude females could produce offspring, and that such nude mice had a more profound deficit in T cell helper activity for antibody production. It was therefore of considerable interest to examine NK activity in such mice. Virtually all of the nude mice from homozygous nude mothers, when tested at 6 to 8 weeks of age, have displayed high levels of NK activity, very similar to that of nude mice from heterozygous mothers. These results indicate that NK activity can develop in the absence of any maternal thymic influence. Both types of nude mice may be at maximal levels of reactivity and therefore the further thymus-related immunologic deficit in the offspring of nude mothers may not result in even higher levels of reactivity. Consistent with this is the observation of Loor et al. (1975) that levels of &positive pre-T cells were similar in nude mice from homozygous and heterozygous nude mothers.

2. Rats Shellam (1977) examined the effects on NK activity of administration of an immunosuppressive antilymphocyte serum and found that it had no effect. H e then tested rats after neonatal thymectomy or after adult thymectomy, irradiation, and bone marrow reconstitution. Similar to the results in mice, these manipulations resulted in higher levels of NK activity.

3. Human Studies of the effects of altered in uivo thymic function, comparable to those in mice and rats, have not yet been reported. It will be of interest to study NK activity in patients with various types of T cell deficiencies, and particularly patients with failure of thymus development. In addition, patients receiving thymic hormone preparations can be studied to see if these cause some in vivo decrease in NK activity.


343

E. EFFECT OF IMMUNOSUPPRESSION ON NK ACTIVITY In contrast to the extensive information on the effects of athymia and thymectomy on NK activity, there is little information available regarding the effects of other forms of immunosuppression. Such information could be quite useful in dissociating NK from other types of cytotoxicity and in evaluating the in vivo role of NK cells. Since there is such a paucity of data and since NK cells and K cells share many characteristics (see Section V), we will discuss together information on the effects on NK and K cells by immunosuppressive treatments. We recognize the obvious objection that even if the cells mediating both functions are similar the mechanisms of killing may be different and the treatments may act at different phases in the development of these functions. NK and ADCC activity in mice has been shown to be moderately resistant to the effects of irradiation in vitro.,Low doses of irradiation (350-700 R) did not appreciably decrease NK or ADCC activity while higher doses (>1000 R) produced a substantial reduction in cytotoxicity (Dr. A. Santoni, unpublished observations). Whole body irradiation with 350 R had no appreciable effect on NK activity when measured after 0 to 3 days, whereas 800 R caused a decrease of greater than 70% in NK and ADCC activity at 1 to 3 days (J. Y. Djeu and A. Santoni, unpublished observations). Shellam (1977) examined the effects of irradiation in rats. NK cells were somewhat more resistant to in vitro radiation than were immune T cells, with 50% of activity remaining after 1000 R and 30% after 5000 R. In contrast, whole body radiation had a greater effect, with a substantial loss of activity at 3 days after 500 R and complete elimination of reactivity with 900 R. Treatment of mice with cyclophosphamide resulted in a considerable decrease in NK reactivity, with a reduction of greater than 70% at 1 to 3 days after in vivo administration of 110 mg/kg (Dr. J. Y. Djeu, unpublished observations). Some limited information has been obtained on the effects of immunosuppression on human NK and K cell cytotoxicity. Rosenberg et al. (1974) found that 500 R x-irradiation, given 24 hours prior to assay, did not affect reactivity, whereas 1000 R or more caused a marked reduction in cytotoxicity. If the x-irradiation was performed just prior to assay, much of the activity persisted even after 15,000 R. Campbell et al. (1976) examined the effects of irradiation on various lymphocyte functions and found that antibody dependent cell-mediated cytotoxicity as well as some T cell functions and the number of B cells in the peripheral blood were substantially reduced after 3000-3500 R irradiation administered over a period of 1 month. Performance of the

344


cytotoxicity assay in the presence of some immunosuppressive drugs, actinomycin D or methotrexate, reduced the activity of normal lymphocytes (Rosenberg et al., 1974). A few patients with collagen diseases who were receiving corticosteroid therapy were also found to have very low levels of NK activity (Rosenberg et al., 1974). This was confirmed in a more systematic study by Parrillo and Fauci (1977a), which showed that intravenous administration of hydrocortisone caused a profound inhibition of NK activity after 24 hours. It can be seen from the above summary that only fragmentary information has been gathered thus far on the effects of immunosuppressive agents on natural cytotoxicity or on ADCC. It will be important to perform more systematic studies and examine the kinetics of depression after various in vivo treatments. Furthermore, in order to assess the relative in vivo importance of NK activity, information must also be obtained at the same time on the effects of the treatment regimen on other immune functions, including immune T cell cytotoxicity, ADCC, and macrophage mediated effects. In previous studies, usually only one function was examined. Thus far, very little information of this type is available. Shellam (1977)found that the profound depression of NK activity in rats at 3 days after irradiation was followed at 10 days b y levels of activity substantially higher than that of untreated controls. Similarly, studies in our laboratory have shown that administration of cyclophosphamide (110 or 330 mg/kg) caused a marked suppression of NK activity in mice at 1to 3 days but that this was followed by a rebound to levels above those of the untreated controls (J. Y. Djeu, unpublished observations). In contrast to the full recovery of NK activity by 7 to 10 days after cyclophosphamide treatment, induction of immune T cell cytotoxicity by MSV was inhibited b y administration of the drug 7 days earlier (D. H. Lavrin, unpublished observations). However, NK activity was not examined in this study and such direct comparisons are needed. After therapeutic radiotherapy of patients, Campbell et al. (1976) examined the kinetics of the return to normal of various lymphocyte subpopulations, including K cells. K cell activity and B cell levels became normal by 3 to 6 months after treatment, whereas T cell activity (lymphoproliferative response to PHA) remained abnormal for over 1 year. We have also observed divergent temporal effects of cyclic combination chemotherapy on NK activity and lymphoproliferative responses in mixed leukocyte cultures (Herberman et al., 1972). Parrillo and Fauci (1977a) performed a kinetic study of the effects of corticosteroids on NK activity and on lymphocyte subpopulations. At 4 hours, NK activity was unchanged, E-RFC were decreased, and cells


345

with Fc receptors and K cell activity (Parrillo and Fauci, 197713) was increased. These limited results suggest that the effects of immunosuppressive agents on NK activity can vary considerably with the type of treatment, dose and frequency of administration, and time of assay in relation to treatment. Because of these complexities and the differential effects on various immune functions, caution must be exercised in attributing the in vivo effects of a particular regimen of immunosuppression to inhibition of NK activity. V. Relationship of Natural Cell-Mediated Cytotoxicity to Antibody-Dependent Cell-Mediated Cytotoxicity

It has recently become apparent that the expression of NK activity in mice and in human donors is correlated closely with the expression of antibody-dependent cell-mediated cytotoxicity (ADCC). The effector cells mediating both forms of cytotoxicity also have very similar characteristics, and these observations raise the possibility that natural cell-mediated cytotoxicity is actually a form of ADCC. A. CORRELATION OF NK ACTIVITY WITH ADCC

1. Mice

Natural antibodies against mouse tumor cells and type-C virusassociated antigens have been detected by many investigators (Aoki et al., 1966; Mellors et al., 1969; Herberman and Aoki, 1972; Ihle et al., 1973; Sato et al., 1973; Nowinski and Kaehler, 1974; Aaronson and Stephenson, 1974; Martin and Martin, 1975) and this had led to consideration of the possible relationship of NK activity to ADCC. ADCC has been found to account for a major portion of the natural reactivity observed in microcytotoxicity assays with mouse mammary tumor virus-infected target cells (Blair and Lane, 1975a,b; Lane et al., 1975; Blair et al., 1976). Greenberg and Playfair (1974) and Kiessling et al. (1976a) failed to find any correlation between NK activity and ADCC against chicken erythrocyte target cells. However, the ADCC in their systems was primarily mediated by adherent phagocytic cells, and other types of effector cells, which are nonadherent and nonphagocytic, have been shown to play a role in ADCC of mouse tumor target cells (Blair and Lane, 1975b; Lamon et al., 1975). Therefore, it would seem more appropriate to compare NK activity with the latter type of ADCC against mouse tumor target cells. In collaboration with Dr. A.

346


Santoni (unpublished observations), we have directly compared the expression of NK reactivity with the levels of ADCC against mouse tumor target cells coated with alloantiserum. The effect of age on reactivity in ADCC was very similar to that described (Section I1,A) for NK activity. In C57BL16, BALB/c and nude mice, little ADCC reactivity was detected prior to 4 weeks of age, and activity reached peak levels at 5 to 9 weeks of age and was usually low or undetectable thereafter. In addition to the overall correlation among groups of mice of different ages, there was a fairly good correlation in levels of NK activity and ADCC among individual mice. As observed by 0. Stutman (personal communication), C3HfBALB/c mice had detectable levels of ADCC activity at 1 to 2 weeks of age. In this strain, NK activity was also detectable in such young mice. As with NK reactivity, young nude mice with random-bred, BALB/c and C57BL/6 genetic backgrounds had high levels of ADCC activity. Among different strains of conventional, thymus-bearing mice, a good correlation was seen in tests for NK activity and ADCC activity against the same target cell (RBL-5). C3H/HeN (both high and low mammary tumor incidence sublines) and NZB had high levels of reactivity; C57BL/6 and BALB/c had intermediate levels; and A, AKR, and SJL had low levels. As a further correlation, inoculation of older mice with LCMV or C. parvum, which had been found to boost NK activity, also led to a rapid increase in levels of ADCC activity. Therefore, in contrast to the previous reports of a lack of correlation between NK activity and ADCC (Greenberg and Playfair, 1974; Kiessling et a1 ., 1976a),we have found the two types of cell-mediated cytotoxicity, against mouse tumor target cells, correlated very well for a variety of different characteristics. 2. Rats There have been very few efforts thus far to relate NK reactivity to ADCC. However, in view of the correlation between augmentation of human CCC and ADCC (see Section V,A,3), it is of interest to note that ) that incubation of rat spleen cells at 37°C Glaser et al. ( 1 9 7 6 ~found with PHA or LPS led to increased ADCC activity as well as to increased cytotoxicity against tumor cells (as discussed in Section II,E,2). Another observation of possible relevance, in view of the persistence of NK activity in thymectomized rats (Shellam, 1977), is that ADCC activity was also detectable in such animals (Harding et al., 1971).

3. Human Several investigators have recently examined the correlation of levels of NK reactivity with levels of ADCC reactivity against tumor


347

target cells. In each of these studies, of normal donors and/or cancer patients, good correlations between NK and ADCC activities have been seen (Peter et al., 1975a,b; Santoli et al., 1976; Trinchieri et al., 1977; W. H. West and R. B. Herberman, unpublished observations). In addition, the organ distribution of effector cells for both forms of cytotoxicity were found to be the same, with considerable reactivity with PBL and spleen cells and little reactivity with thymus, tonsil, or lymph node cells (West et al., 1977b, and unpublished observations). When human PBL have been cultured at 37" in medium containing FBS (see Section 11,E73),additional ADCC reactivity was generated along with CCC (Ortaldo et al., 1977a,b). Stimulation of PBL in vitro with mitogens, soluble antigens, or allogeneic cells has also produced augmented ADCC reactivity (Connolly et al., 1975; MacDonald and Bonnard, 1975; Ortaldo et al., 1977a,b) in addition to CCC.

B. COMPARISON OF EFFECTORCELLS MEDIATING NK ACTIVITY AND ADCC 1. Mice

Previous studies of the relationship between the NK cells and eEector cells mediating ADCC in mice have indicated several major differences (Greenberg and Playfair, 1974; Kiessling et al., 197513, 1976a). ADCC activity was attributed to a monocytic cell (Greenberg and Playfair, 1974; Kiessling et al., 1975b), an adherent, phagocytic cell with a receptor for complement (Kiessling et al., 1975b, 1976a). However, as pointed out previously (Section V,A,l), ADCC was tested against chicken erythrocyte target cells, and some studies have indicated that other types of effector cells may also be responsible &r ADCC against mouse tumor target cells (Blair and Lane, 1975b; Lamon et al., 1975). Therefore, in collaboration with Dr. A. Santoni (unpublished observations), we have compared the characteristics of the NK cell with the effector cells (which we will call K cells) mediating ADCC against alloantibody coated mouse tumor cells in an overnight 51Crrelease cytotoxicity assay. In this system, like the NK cells, the K cells were shown to be nonphagocytic and nonadherent. Adsorption of spleen cells on EA monolayers to remove Fc receptor-bearing cells resulted in a moderate decrease in relative K cell activity, but as has been described for NK cell activity, the depletion was only partial. These experiments have indicated a relative difficulty, in comparison to human cells (West et al., 1977a; Kay et al., 1977), in removing mouse Fc receptor-bearing cells by such procedures. Pretreatment of spleen cells with anti4 serum plus complement has resulted in a substantial loss in ADCC activity as well as a partial decrease in NK

348


activity. However, such treatment with anti4 without complement had similar effects on K cell activity, and it remains unclear as to whether the inhibition is due to selective action on K cells bearing 6 antigen or to nonspecific inhibition by antigen-antibody complexes. The sensitivity of N K cells and K cells to irradiation in vitro was quite similar, with little effect of 500R and then a progressive decline in activity with increasing dose. Both types of activity were found to be sensitive to in vivo cyclophosphamide treatment. Taken together with the close correlation in levels of N K and K cell activity in different mice, these data point to a very close similarity between N K and K cells. However, there have been some consistent differences between the characteristics of N K and K cells. In contrast to the lability of N K activity at 37"C, ADCC activity was unaffected by incubation for up to 3 hours at 37°C. Kiessling et al. (1976a) and we have obsetved than N K activity in a short-term assay is trypsin sensitive, in contrast to the lack of inhibitory effect of this enzyme on ADCC activity. Also, addition of aggregated gammaglobulin (Herberman et al., 1975b; Kiessling et al., 1976a) or an anti-immunoglobulin reagent (Herberman et al., 1975b), which can effectively inhibit ADCC, had no inhibitory effect on N K activity. It should be noted, however, that the inhibitory effects on ADCC of gammaglobulin and anti-immunoglobulin have been seen in experiments with antibody coated target cells, in which these agents would be expected either to block the free Fc receptors of the effector cells or to mask the Fc portion of the immunoglobulins on the target cells. These experiments tend to rule out ADCC against antibody coated target cells as a mechanism for N K activity but do not bear on the possibility of ADCC by K cells already armed in vivo (see Section V,C).

2. Human A series of recent studies have indicated that human N K cells and K cells have very similar cell surface markers and other characteristics. Lymphocytes appear to be responsible for most of the observed reactivities, but normal granulocytes have shown to be capable of mediating both direct cytotoxicity (Takasugi et al., 1975) and ADCC (Zighelboim et al., 1974). Peter et al. (1975b, 1976a) performed a series of separation procedures on human PBL and showed that N K and K cells were present in the same fractions. They found that both types of effector cells were nonadherent and nonphagocytic and had Fc receptors for IgG. These investigators, as well as many others (Wisloff et al., 1974; Zighelboim et al., 1974; Brier et al., 1975; MacDermott et al., 1975; Cordier et al., 1976), reported that the K cell was a non-T cell,


349

which did not form rosettes with SRBC. However, this did not represent a discrepancy from the characteristics of NK cells, since Peter et al. (1975b) also concluded that NK cells lacked receptors for SRBC. Perlmann et al. (1975) initially reported that K cells could form rosettes with SRBC if the latter were pretreated with neuraminidase. By using the procedures previously described (Section IV, B), West et al. (1977b) have demonstrated that K cells, as well as NK cells, have low affinityreceptors for SRBC. In a direct comparison of the characteristics of NK and K cells, Kay et al. (1977) found that up the 80% of the total lytic units of both NK and ADCC activities were associated with E-RFC. As described by Peter et al. (1975b),both NK and K cells had Fc receptors. In studies of lymphocytes in various lymphoid organs, there has been an excellent correlation between the presence of low affinity E-RFC with Fc receptors (West et al., 1977c) and the expression of NK and ADCC (West et al., 1977b) activities. All of these were present in PBL and spleen, and virtually undetectable in thymus, tonsil, and lymph nodes. In addition, both types of cytotoxic reactivity were inhibited by pretreatment of the PBL with an ammonium chloride solution (Kay et al., 1977). As with the studies of mouse lymphocytes, some clear differences between NK reactivity and ADCC have been observed. NK reactivity is markedly inhibited after treatment of PBL with trypsin or chymotrypsin, whereas such treatments have no effect on ADCC (Kay et al., 1977). Further, when protein A was added to the cytotoxicity assays, ADCC, but not NK, was significantly inhibited.

c. POSSIBLE

MECHANISMS OF ACTION

OF

NK CELLS

All of the data gathered to date indicate that NK cells are directly cytotoxic and that their action is dependent on contact with the target cells. Although we and others have assumed that NK activity was a form of cell-mediated cytotoxicity in which the information for reactivity and specificity was an integral part of the lymphocyte, the possible role of ADCC must now be considered. The close correlation between levels of NK activity and K cell activity may be more than coincidental. Some of the natural cell-mediated cytotoxicity observed in a microctyotoxicity assay against mouse mammary tumors has been shown to be mediated by ADCC, with coating of target cells by antibody during the incubation period (Blair and Lane, 1975a,b; Blair et al., 1976). As discussed (Section V,B), it seems unlikely that coating of target cells by antibodies is involved in NK cell activity. However, since NK cells appear to possess Fc receptors, it is possible that these cells are armed

350


in vivo with natural antibodies or with antigen-antibody complexes (Perlmann et al., 1972; Greenberg and Shen, 1973; Sakselaet al., 1975). The reduction in NK activity by trypsinization or by incubation at 37.T would be consistent with this since such treatments might remove the immunoglobulins needed for reactivity. Based on this possibility, we have performed a series of experiments in which untreated and trypsinized lymphocytes were incubated with autologous serum and with culture fluids from explanted lymphocytes. Thus far the cells incubated with serum or culture fluids have not been consistently found to have increased cytotoxic reactivity against the tumor target cells. However, there has been very little experience with armed lymphocytes. It is possible that some unidentified variables are critical to the success of such experiments. Pollack and Nelson (1974) described arming of lymphocytes for reactivity in a microcytotoxicity assay using serum from mice 1 to 2 days after inoculation of tumor cells or oncogenic virus. The arming factor in their system was shown not to be assocaited with either IgG or IgM (Pollack and Nelson, 1975),and it could be produced in lethally irradiated mice (Pollack and Nelson, 1976). Similarly, Peter et aZ. (1976b) have obtained evidence for an augmenting soluble factor, which may be produced by B cells, but which could not be removed b y anti-IgG. It will be necessary to rule out the role of K cells with such unusual arming or augmenting factors as well as with more conventional antibodies before it can be concluded that NK activity is a mechanism entirely distinct from ADCC. The diversity of natural antibodies against mouse tumor cells, type-C viruses, and other antigens would provide a ready explanation for the heterogenous specificities detected by NK cells. Another possible mechanism has been suggested for natural cytotoxicity by effector cells with Fc receptors (Rager-Zisman et al., 1976). In their study, Herpes simplex virus-infected target cells were shown to express Fc receptors and nonspecific killing appeared to occur b y cross-linking of effector and target cells by aggregated immunoglobulins through their F c receptors. Although this mechanism needs to be considered in each new system, it does not appear to explain most of the observations of NK activity discussed above. In both our mouse and human systems, we have found that some target cells without detectable Fc receptors to be quite susceptible to cytotoxicity. Furthermore, performance of the cytotoxicity assays in medium without serum or with agamma FBS had no effect on results. Rager-Zisman et al. (1976) also found that the effector cells for their cross-linking phenomenon were adherent cells with the characteristics of macrophages, which are different from the characteristics of NK cells described earlier (Section IV,B).


35 1

VI. Model for Placement of N K and K Cells in Pathway of Differentiation of T

Cells The considerable evidence presented herein for an inverse correlation between levels of NK activity and thymic function might be explained by two alternative mechanisms: (a) NK cells are prethymic T cells and further differentiation under the influence of the thymus results in thymic and post-thymic T cells in which NK activity is no longer expressed or expressed only at low levels; and (b) NK cells are non-T cells, whose activity is modulated b y T suppressor cells. Most of the data presented thus far are consistent with either mechanism. However, the results in the mouse of treatment with anti-8 serum plus complement tend to rule out the suppressor cell model. With that mechanism, one would predict that removal of T cells from a lymphoid cell preparation of conventional mice by such treatment would lead to higher levels of activity and that such treatment of nude cells would either have no effect or would cause some increase. However, the data indicate that anti-8 treatment of lymphoid cells of normal mice had no effect and such treatment of nude cells resulted in partial depression of NK activity. We have further tested the suppressor cell hypothesis by mixing BALB/c nude spleen cells with spleen cells from syngeneic conventional mice. N o reduction of the high level of NK activity of the nude cells was detected. Similarly, the association of human NK cells with E-RFC does not support the suppressor model and is consistent with the T cell nature of NK cells. There is considerable evidence for the existence of prethymic T cells, with high levels of such cells in nude mice and in neonatally thymectomized mice. Raff (1973) originally described a low percentage (1-2%) of 8 antigen positive cells in nude mice, detectable by immunofluorescence. Somewhat higher values ( 5 4 % )were detected b y cytotoxicity assays (Raff and Wortis, 1970) and nude lymph node cells were shown to be capable of absorbing out anti-8 antibodies (Raff, 1971). Loor and Roelants (1974) subsequently reported that u p to 20% of nude mouse spleen cells contained a low density of e antigen, detectable by a rabbit antiserum to brain associated e antigen. The expression of low density 8 antigen on lymphocytes seemed to be independent of humoral thymic factors, since cells from nude mice born from homozygous nude mothers had comparable characteristics (Loor et al., 1975). This group also reported that T L antigen could be detected by immunofluorescence on as many as 16% of nude mouse spleen cells (Roelants et al., 1976). However, T L antigens have not been detected by cytotoxicity on lymph node cells of healthy nude mice (Scheid et at., 1975). Further support for the presence of prethymic T cells in nude mice has come from studies in which in v i t r o

352


incubation of cells with thymic factors or other agents (Scheid et al., 1973; Goldstein et al., 1975) or in uiuo inoculation of some thymic factors (Scheid et al., 1975) led to a considerable increase in the proportion of lymphocytes expressing the T cell associated markers, 8 and TL. Environmental infection of nude mice with hepatitis virus produced similar effects (Scheid et al., 1975). In most of these studies, only antigenic markers of T cells were analyzed and little direct evidence for induction by humoral factors of functionally active T cells has been presented (Ikehara et al., 1975; Basch and Goldstein, 1975). Nonetheless, Pritchard and Micklem (1973) suggested that nude mice have precursors of functionally active T cells, and since several studies have either indicated the presence of some functionally active T cells in nude mice (Kirov, 1974; Hale et al., 1976) or the ability to induce functional reactivity in lymphocytes of nude mice (Ramseier, 1975). Based on the above information and the results obtained in our studies of the nature of the NK cells, William West and I have developed a model for placement of N K cells in the pathway of differentiation of T cells (Fig. 1). This model is similar to that proposed by Loor et al. (1976) for changes in cell surface antigens during early T cell differentiation. Prethymic T cells contain a low density of 8 antigen. Within this population are included cells with Fc receptors and cells with NK activity. The cells with NK activity and those with Fc receptor seem to b e overlapping if not identical subpopulations.

Mphrr

?mbymic T Wh

I I

I

I I

?

I

I

?

I I

I

I

I

L------+

I I L

Thymoun

I

- -- - - - - - - - - - - - -- - - - - -- E mtfmilp?

FIG. 1. Model for placement of NK and K cells in the pathway of differentiation of T cells.


353

Human NK cells also fit this formulation, having easily detectable Fc receptors and low affinity receptors for SRBC, a marker as characteristic for human T cells as 8 antigen is for mouse T cells (West et al., 1977a). West et al. (1977~)have recently obtained considerable evidence that there is a separable subpopulation of human T cells, characterized by the presence of Fc receptors for IgG and low affinity receptors for SRBC. This popularion bears considerable similarity to the mouse pre-T cell subpopulation, and these may be analogous. It also seems likely that human K cells mediating ADCC are also included in this same population of lymphocytes (West et al., 1977b). In the studies with human effector cells, NK and K cells had virtually identical cell surface characteristics (Kay et aZ., 1977). In the mouse, the placement of K cells in this scheme is somewhat less certain. As discussed previously (Section V,B), we are not yet certain about the presence of 8 antigen on K cells. However, the correlation between NK cell activity and K cell activity in relation to age, strain, and nude mice (as discussed in Section V,A) has been striking, and both NK and K cells in mice have Fc receptors. In analogy with the human, it seems likely that mouse K cells will reside in the same subpopulations as NK cells. Prethymic T cells appear to originate in the bone marrow (Scheid et al., 1973; Loor et al., 1976). Similarly, Haller, Kiessling, and their associates have obtained evidence that the bone marrow contains precursors of NK cells (R. Kiessling, personal communication). Bone marrow cells from high responder mice were able to transfer NK reactivity to low responder recipients. We have failed to detect NK activity earlier, with age or after boosting, in bone marrow than in the spleen, but this may simply indicate that transfer of NK cells between the bone marrow and peripheral lymphoid tissues occurs rapidly. According to our model, when pre-T cells come under thymic influence, by passing through the thymus in the normal state, some differentiation occurs, with loss of Fc receptors and NK and K reactivity, and expression in the mouse of increased amounts of 8 antigen and increase in, or appearance of, TL antigen, and in the human expression of higher affinity receptors for SRBC. In the absence of the thymus in nude mice or neonatally thymectomized mice, high levels of pre-T cells with NK and K reactivity persist. Exposure of such cells to humoral thymic factors or development of some diseases in nude mice (Scheid et al., 1975)could also lead to some differentiation of the cells, with loss of their associated NK functional activities. After T cells leave the thymus, they further differentiate into mature T cells which in the mouse, continue to express moderate amounts of 8 antigen, lose T L antigen, and in the human, express receptors with

354


moderate affinity for SRBC. These cells remain inactive in NK and K cell reactivity and usually continue to lack Fc receptors. Human effector cells which are generated in mixed lymphocyte cultures and have specific cytotoxic reactivity against allogeneic target cells appear to reside in this subpopulation (Ortaldo and Bonnard, 1977), and in general, cytotoxic specifically immune mature T cells may be generated from this population. This leaves us to account for the expression of NK and K cell activities in peripheral lymphoid organs. In the absence of the thymus in nude mice, this would simply involve a movement from the bone marrow into other organs. This process, bypassing the thymus, may also occur to some extent in conventional mice. Alternatively, or in addition, some cells leaving the thymus and some mature T cells in the periphery may be activated or de-differentiate back into cells with pre-T cell characteristics. Ortaldo et al. (1977a,b) have obtained a considerable amount of evidence to support this possibility, When human PBL were depleted of Fc receptor-bearing cells and then placed in culture, within 4 to 5 days cells with detectable Fc receptors appeared and these had high levels of direct cytotoxic activity and ADCC. Treatment of peripheral lymphocytes with human thymosin has been shown to decrease NK and K cell activities and to change some human PBL from low affinity E-RFC to high affinity E-RFC (W. H. West, A. White, and R. B. Herberinan, unpublished observations). Taken together, these observations support the possibility that lymphocytes can reversibly move from one subpopulation to the other. VII. Discrimination Between Natural Cell-Mediated Cytotoxicity and

Cytotoxicity by Other Effector Cells

The contribution of NK activity to the cytotoxicity measured in any in vitro assay needs to be carefully considered. NK cells may be present in immune lymphoid populations, and in fact, the immunizing procedure may have augmented NK activity as well as produced more specific reactivity. Exactly what percentage of the total reactivity is represented by this effector function is dependent on variables such as choice of target, species, age (in rodents), the organ used as the source of effector cells, previous exposure to modulating agents (e.g., viruses, bacteria, chemotherapeutic drugs, irradiation, tumor cells), and the nature of the in vitro assay. Therefore, in order to determine accurately the role of other cytotoxic effector mechanisms (e.g., specifically cytotoxic T cells, and macrophages), it is necessary to define, eliminate, or control for natural cytotoxicity in all in vitro cytotoxicity as-


355

says. At the moment there are no readily available methods for specifically eliminating NK activity. However, b y utilizing the information we have at the present concerning the nature of the NK cell, it is possible to approach the problem in a systematic manner, keeping in mind that as we enrich, deplete, or control for NK activity we might be influencing other effector cell activities as well.

A. MICE Until recently, many investigators used lymphoid cells from normal mice as baseline controls for studies of cytotoxicity produced b y immunization. However, the observations of NK activity and particularly the considerable variability in levels of NK activity among individual mice have raised some serious concerns regarding the proper baseline to be used in cytotoxicity assays. Some aspects of this have been previously discussed (Herberman et al., 1976a). In order to assess the levels of NK activity, several possible baseline controls have been used, and the baseline of choice depends on the type of assay. The most commonly used control is the medium control, but as previously discussed (Herberman et al., 1976a), this is often an artificial control which has little relationship to the results obtained in any groups with lymphoid cells present. For short-term assays, e.g., 4-hour W r release assay, an autologous control has been quite satisfactory. This consists of the use of unlabeled target cells in place of, and at the same concentration as, the lymphoid effector cells. This has provided a reproducible baseline (Herberman et al., 1976a) and has been especially useful for studying low levels of NK reactivity. Another baseline control that has been useful in long-term assays (incubation periods of 12 to 18 hours or more) as well as in short-term assays is a thymus cell control in which normal thymus cells or thymus cells from immune animals are used in place of effector cells (Herberman et al., 1976a; Oldham et d., 1977; H . T. Holden and A. Santoni, unpublished observations). This has been suitable because, as discussed herein (Section IV,A), thymus cells from either normal or immune animals lack NK activity and they also lack detectable immune cytotoxic T cell activity. However, when thymus cells are used as controls, it is essential to avoid contamination by adjacent lymph nodes. When one is interested in studying non-NK immune cytotoxic activity, a baseline control of normal lymphoid cells may be quite useful and, in some cases, may be the best available baseline (e.g., Tinget al., 1977). At first glance, this seems very logical since it should reflect the

356


amount of immune cytotoxicity above the natural levels. However, there are a few problems which need to be considered before this baseline is adopted. The most important is the susceptibility to NK activity of the particular target cell and the amount of variation among mice being studied in their levels of NK activity. If the target cell is reIativeIy resistant to NK activity and older mice are studied, with low levels of NK activity, a pool of normal lymphoid cells from several mice would seem reasonable. However, if much NK activity was seen against the target cell, the normal baseline controls would b e high and it would be important to know how much of that was due to N K activity versus poor condition of the target cells. In addition to these considerations, it should be realized that the immunization procedure itself could cause a change in the level of NK activity, and it would be very helpful to have a means to assess the contribution of this to the apparent level of immune cytotoxicity achieved. Any of the above methods could provide a stable baseline control for a given test. Nevertheless, it is still difficult to make a comparison between data obtained in tests performed on separate days. However, by using the autologous control in conjunction with cryopreserved target cells and a cryopreserved lymphocyte internal standard, assay variation can be monitored and controlled (Holden et al., 1976, 1977). These techniques keep assay to assay fluctuations to a minimum so that results obtained on different days can be evaluated and compared. An entirely different approach to the study of non-NK cytotoxic effector cells would be to deplete the NK cells selectively. Unfortunately, selective removal of NK cells is difficult in the mouse. There are several methods that can be employed to decrease the NK activity in mouse spleen cells but in most cases they give variable results and only remove part of that cytotoxicity. As discussed earlier (Section IV,B), mouse NK cells have an Fc receptor; however, adsorption ofcells on EA monolayers does not completely deplete NK cells even though a large percentage of the activity is lost after this treatment (Herberman et al., 197713).This may be explained by the absence of Fc receptors on some NK cells or by the inability of this technique to remove all cells with low affinity Fc receptors. A further problem with this approach is that at least some mouse immune cytotoxic T cells have Fc receptors and therefore this would not be a reliable discriminant (Stout et al., 1976). Two different antisera have been employed to decrease NK cytotoxicity, anti-8 (Herberman et al., 1975b) and anti-NK (Glimcher et al., 1977). However, treatment with anti-8 plus complement presents several problems. Although this will partially deplete NK activ-


357

ity in some cases, it also will eliminate mature T cells and hence cannot be used to study residual immune T cell cytotoxicity. Furthermore, such treatment cannot be used to deplete NK activity reliably because it only causes up to 90% decrease in cytotoxic reactivity in nude mice (Herberman et al., 1975b) and in boosted animals (Herberman et d., 1977a) but it does not affect natural cytotoxic reactivity in uninoculated, conventional mice. However, since anti-8 plus complement can substantially reduce the levels of boosted NK activity, caution should be exercised in attributing cytotoxic reactivity to mature T cells as opposed to NK cells if the activity can only be removed after vigorous treatment with anti-8 plus complement or if the activity is only partially inhibitied by such treatment. Another problem which is associated with this antiserum is that treatment of spleen cells with anti-8 alone, without the addition of complement, often increases the cytotoxicity against some tumor targets, especially in long-term cytotoxicity assays (A. Santoni, unpublished observations). The mechanism of this effect is unknown and is currently under investigation. Anti-NK antiserum appears to be much more specific in its action, with selective and fairly complete elimination of NK effector cells, and therefore could potentially be the most useful method of depleting NK function from effector cell populations. The main limitation at present is that this antiserum only affects NK activity of some mouse strains. NK activity in the mouse can also be altered by incubation at 37°C (Herberman et al., 197513) or b y extended trypsin treatment (Kiessling et al., 1976a).Our laboratory has shown that NK activity is sensitive to incubations at 37°C for 2 hours or longer (Herberman et al., 1975b), and, in fact, overnight incubation at 37°C reduces the level of activity to almost baseline levels. Assuming that other effector functions are not so unstable, this technique can be very useful for depleting NK activity. However, its value might be limited in long-term assays since culture of normal cells for several days can actually lead to the generation of cytotoxicity (Shustik et al., 1976; Gorczynski, 1976a). Treatment with 1.0% trypsin for 30 to 45 minutes can decrease the level of NK activity to low levels if the cytotoxicity assay is short (4-6 hours). However, the technique is not effective in reducing NK activity in long-term (18 hour) cytotoxicity assays since the levels of activity are siniilar with or without treatment (A. Santoni, unpublished observations) and this appears to represent a regeneration of the cytotoxicity during the prolonged incubation period. NK activity may also be distinguished from specific immune activity by the specificity of the reactions. As previously stated (Section 111) most of the studies on NK activity in the mouse have demonstrated

358


some degree of specificity. After defining the specificity that is being recognized by the immune cytotoxic cells, it should be possible to identify target cells which do not carry the relevant antigens (and hence are not lysed by immune cytotoxic effector cells) but are still sensitive to cytotoxicity by normal lymphoid cells from that strain of mouse. However, in many cases the cells which carry the relevant antigens for immune cytotoxicity also are sensitive to NK activity; nevertheless, the antigens which are recognized by the different effector populations do not appear to be the same, at least in the system that we have examined (Herberman et al., 1975a). Therefore by examination of the specificity by the inhibition of 51Crrelease cytotoxicity assay (Herberman et al., 1976b), one might be able to distinguish NK from immune cytotoxicity. A further approach along this line, which has already been found to work with human PBL, is to remove NK cells b y adsorption to monolayers of cells bearing NK-associated antigens and then test for residual activity against specific target cells. NK activity in uninocuIated mice has a consistent relationship to age, with the peak of natural killer activity at about 6 to 8 weeks of age with low or undetectable levels after 12 weeks of age. Therefore, the role of NK activity can be minimized in vitro by employing older mice for study. However, this is complicated by the discovery that many different agents are capable of boosting NK activity in mice. Immunization with a particular antigen or a particular tumor cell does not mean that the cytotoxicity measured in vitro is specifically against that tumor cell, especially when tests are performed within a few days after inoculation. Several investigators have reported the detection of cell-mediated cytotoxicity 2 to 3 days after inoculation of syngeneic or allogeneic tumor cells (Lamon et al., 1972; Forman and Britton, 1973; Pollack and Nelson, 1974). The observed reactions were thought to b e specific but the nature of the specificity at these earIy times was not extensively evaluated. Forman and Britton (1973)found that the effector cells harvested within a few days after inoculation were resistant to treatment with anti-8 plus complement, whereas later on they were quite sensitive. It is quite possible that at least part of the reactions observed at 2 to 3 days represented a boost in natural reactivity rather than early primary immunization to the specific antigen used for immunization. Pfizenmaier et al. (1975) found that early after immunization with LCM virus, autoreactivity separate from specific anti-LCM reactivity was seen. Again, as discussed above (Section II,D), it seems quite possible that their observations were related to the boosting of natural reactivity. Finally, many investigators have shown cytotoxic reactivity in peritoneal exudate cells after the animals have been in-


359

jected intraperitoneally with BCG or C. paruum. We (Herberman et al., 1977a) with BCG and C. pamum and Wolfe et al. (1976) with BCG have shown that NK cytotoxicity is greatly enhanced after these treatments. Hence, at least some of the cytotoxic reactivity detected after these injections could be related to the NK activity which has been stimulated. We have recently observed that macrophages from C. parvum injected animals, purified by adherence on petri dishes, are contaminated by considerable numbers of NK cells (P. Puccetti and H. T. Holden unpublished observations). Hence, considerable caution must be exercised in attributing observed cytotoxicity to either immune, mature T cells or to macrophages, and in many cases, NK cells may play a significant role. B. HUMAN Discriminating specific immune cytotoxic activity from NK activity is considerably more difficult in studies of human lymphoid cells than it is in mice. The history of exposure to possible immunogens is less certain, and there are considerably more variable, and less predictable, levels of NK activity in human subjects. This has made it very difficult to study cell-mediated cytotoxicity of cancer patients against their own tumor cells or against allogeneic tumor cells of the same histologic type (Herberman and Oldham, 1975). Since the problems of natural cytotoxicity have been recognized, an assortment of possible baseline controls, similar to those discussed for studies in mice, have been considered. In order to measure the variation in reactivity among normal donors, it has been suggested that several normal individuals be tested in each experiment and that the least active normal (Oldham et al., 1975) or the median reactive normal (Herberman et al., 1976a) be used as the baseline. Although these approaches provide a reasonably good reflection of normal reactivity on a population basis, they are not adequate for following the reactivity of a given individual over time, since the baseline is not completely stable. The use of a standard cryopreserved population of normal PBL should be helpful for this purpose (Oldham et al., 1976b; Holden et al., 1977).However, it is even more important to develop methods for clearly discriminating between NK activity and other forms of immune cytotoxic reactivity. One approach, developed by Cannon et a l . (1977), has been to test simultaneously the cytotoxic reactivity of PBL against a good indicator cell line for NKactivity and against a target cell with antigens relevant to the system under study. By using a cell line derived from breast cancer

360


and the K562 cell line, it was possible to identify some patients with breast cancer who had relative hyperreactivity against the breast cancerderived line. These data suggested that these patients had another type of effector cell activity in addition to NK activity. Another approach, based on differences in specificity of the various types of effector cells, has demonstrated the presence of cytotoxic activity mediated by cells with characteristics of NK cells in mixed lymphocyte cultures (Ortaldo and Bonnard, 1977).PHA blasts, which are good targets for specifically immune T cells and are resistant to NK activity, reflected one type of activity, whereas K562 target cells mainly reflected NK activity. More definitive discrimination would be expected to result from selective removal of NK cells from a mixture of lymphoid cells. By removal of cells forming rosettes with EAC (SRBC; anti-SRBC; and complement complexes), it was possible to identify effector cells specific for Epstein-Barr virus-associated antigens in the blood in patients with infectious mononucleosis (Svedmyr and Jondal, 1975)and in a biopsy of Burkitt’s lymphoma (Jondal et al., 1975). As discussed previously (Section IV,B,3), this depletion was probably through the Fc receptors on human NK cells. By adsorption of NK cells on monolayers of SRBC coated with IgG antibody (West et al., 1977a) or of immobilized soluble antigen-antibody complexes (Kay et al., 1977), it has also been possible to deplete NK activity almost completely. Since immune cytotoxic T cells do not appear to have detectable Fc receptors (Ortaldo and Bonnard, 1977), this approach would appear to b e a very useful one for making the needed discrimination. It has already been shown to identify effector cells in the peripheral blood of some breast cancer patients, which lack Fc receptors and are cytotoxic against breast cancer-derived target cells (W. West, G. 13. Bonnard, and R. B. Herberman, unpublished observations). The other marker on human NK cells which provides the basis for separation is the receptor for SRBC. As discussed above (Section IV,B,3), more than 80% of the NK activity can be removed by rosetting with SRBC (Kay et al., 1977).This b y itself would not be expected to separate NK cells from immune, cytotoxic T cells. However, it is possible at least partially to separate mature T cells with high affinity receptors for SRBC from NK cells in the low affinity E-RFC population b y rosette formation at 29°C (West et al., 1977a,b,c). Adsorption of NK cells on monolayers of target cells which have the relevant antigens offers yet another method for selective depletion of natural cell-mediated cytotoxicity. Studies in progress in our laboratory (W. West and R. B. Herberman, unpublished observations) indicate that this procedure can deplete most or all of the activity against


36 1

K562 cells and not affect the reactivity of some individuals against a breast cancer-derived cell line. In summary then, it seems likely that reliable and effective methods for separating NK activity from other forms of human cytotoxic effector cells are available. In the near future, it should be possible to assess the distribution and specificity of these other types of cytotoxic cells which react with tumor cell lines and/or other target cells. VIII. In Vivo Relevance of Natural Cytotoxicity

Understanding of the in vivo relevance of natural cell-mediated cytotoxicity is the ultimate objective in this area of research. With the multiplicity of other possibly relevant effector mechanisms, it is very difficult to define the role of NK cells in resistance to tumor growth and other in vivo phenomena. It is particularly difficult to answer these questions in human studies since the opportunities for in vivo manipulations are obviously quite limited. Most of the information that has been, and will probably continue to be, gathered in this important area has come from studies in mice, as discussed below. Several investigators have noted that nude mice develop spontaneous tu-mors or carcinogen-induced tumors with relatively low incidence (Rygaard and Povlsen, 1976; Stutman, 1974; Outzen et al., 1975). In addition, transplanted tumors of mice and of heterologous species have not produced progressive tumor growth in nude mice with the consistency that one might have anticipated from the ready growth of skin allografts and xenografts. Bonmassar et al. (1975) have noted the impaired growth of an allogeneic lymphoma in nude mice and in lethally irradiated mice. Rotter and Trainin (1975) found that the Lewis lung carcinoma, 3LL, grew poorly in lethally irradiated, bone marrow reconstituted mice. More than one million 3LL tumor cells and B16 melanoma cells failed to produce tumors in some nude mice (Giovanella et al., 1974). Similarly, Gillette and Fox (1975) showed that several tumors grew less well in thymectomized, lethally irradiated, bone marrow reconstituted mice than in normal mice. Intravenously inoculated syngeneic or allogeneic tumor cells were found to produce fewer tumor colonies in the lungs of nude mice than in normal, nu/+, littermates (Skov et al., 1976; Fidler et al., 1976). Shin et a l . (1975) failed to produce tumors in nude mice by several clonal isolates of virus transformed 3T3 cells. Stutman ( 1975) observed that 120-day-old nude mice had some resistance to tumor induction by polyoma virus, and spleen cells from these mice were able to transfer partial resistance to this virus. Nude mice have become widely used for growth of human tumor

362


cells. In the course of these studies, some resistance to tumor growth has been noted. Although metastatic spread of some transplanted tumors has been observed (Giovanella et al., 1973, 1974) many investigators have noted the rarity of metastasis in nude mice, even when metastatic deposits of human tumors were transplanted (Rygaard and Povlsen, 1969; Castro, 1972; Ozzello et al., 1974; Schmidt and Good, 1975, 1976; Maguire et al., 1976). Some human tumors only produced tumors when inoculated intracranially (Epstein et al., 1976) and others, especially gastric carcinomas, produced no tumors in nude mice (Schmidt and Good, 1975).Maguire et al. (1976) have even noted occasional regression of a highly malignant hamster tumor in nude mice. These many observations of some resistance to growth of syngeneic, allogeneic, and xenogeneic tumors in nude mice might be attributed to their high levels of natural cell-mediated cytotoxicity. The findings that some human cell lines are sensitive to NK activity and that inoculation of these cells can boost reactivity are consistent with this possibility. As more direct evidence, a series of experiments have already been performed which show a correlation of resistance to tumor growth with sensitivity of a tumor to cytotoxicity by NK cells.

A.

CORRELATION OF CYTOTOXICITY

DECREASED TUMORGROWTH WITH NATURAL

Kiessling and his associates (Kiessling et al., 1975c; Petranyi et al., 1976) have performed an extensive series of experiments which demonstrate a correlation between the levels of NK activity in different strains of mice and the resistance of F1 hybrids between each strain and A mice to the A strain lymphoma, YAC, the cultured line of which is very sensitive to natural cytotoxicity. Mice which were thymectomized, irradiated, and fetal liver reconstituted also showed this resistance (Kiessling et al., 1976b).The main problem in relating the in vitro and in vivo observations in the above studies is that the ascitic form of YAC was used for the in vivo studies, and Kiessling et al. (1975a) found this to be rather resistant to NK activity. Another type of correlation has been observed when some tumor cells have been transplanted into mice of different ages. Sendo et al. (1975)observed that young (B6 x BALB/c) F1mice, at the time of peak levels of NK activity, were more resistant to growth of R L d l than were older mice. Similarly, we have found that 6- to 8-week-old nude mice were more resistant to growth of a low dose of MCDV-12 lymphoma cells than were 12- to 14-week-old nude mice. The R L 8 1 tumor also grew poorly in 8-week-old nude mice, even after 350 R irradiation of the recipients, but it produced a significantly higher


363

tumor incidence in nude mice pretreated with cyclophosphamide, which has been found to depress NK activity. Because of our interest in NK activity and its possible relevance to in viuo resistance to tumor growth, we have been quite interested in examining the characteristics of tumor cell lines which grew poorly in nude or thymectomized mice, to determine whether the resistance to such tumors could be explained by sensitivity to NK activity. All of the tumors in this category, the Lewis lung carcinoma (Rotter and Trainin, 1975), the B16 melanoma (Skov et al., 1976; Fidler et al., 1976), E4, an SV40 transformed 3T3 line (Gillette and FOX,1975), and the L5MF-22 lymphoma (Bonmassar et d.,1975) have been shown to be sensitive to natural cell-mediated cytotoxicity by nude spleen cells (see Table IV in Section 11).In collaboration with C. Chang, we have performed more extensive studies on some clones of the spontaneously transformed T-AL/N cell line. It was initially noted that these cells grew less well in allogeneic nude mice than in syngeneic AWN mice. With the T-S-5 clone, the TDS0was 200 to 400 times higher in nude mice than in conventional mice. I n addition, some of the tumors produced in nude mice were found to regress. When these cells were tested as target cells, they were observed to be highly sensitive to cytotoxicity by spleen cells from nude mice (Table IV) but resistant to, lysis by spleen cells from AL/N mice. As a further correlation with the ability of antigenic tumors to boost the levels of NK activity in nude mice, the inoculation of T-S-5 into nude mice increased the resistance to subsequent challenge with the same tumor. Based on these observations, we have formulated the hypothesis that the failure to observe a very high incidence of spontaneous carcinogen-induced tumors in nude mice might be due to their high levels of NK activity. Only tumors with low or no sensitivity to natural cytotoxicity would then be likely to be detected in nude mice. Consistent with this hypothesis, all of the tumors which arose in nude mice that we have studied thus far have been resistant to NK activity and, by inhibition studies, have lacked detectable antigens (see Section III,A,3). A further prediction from our hypothesis would b e that chronic suppression of NK activity in nude mice would result in a higher spontaneous tumor incidence, and we are setting u p experiments to examine this possibility.

B. In Vivo RELEVANCEOF NK ACTIVITYAGAINST CELLS NONMALIGNANT

The findings that NK activity is not completely restricted to tumor cells provide a possible explanation for some in vivo phenomena, par-

364


ticularly that of resistance to bone marrow transplantation. When Bonmassar et al. (1975) observed the impaired growth of an Hhincompatible tumor in nude mice, they suggested that the same mechanism might be responsible for resistance to tumor cells and to bone marrow. Gallagher et al. (1976) have also suggested that the two phenomena might be related. In a recent workshop (Trentin and Bennett, 1977), the information on correlations between NK activity and bone marrow resistance was discussed. Despite some apparent differences in characteristics, a number of similarities was noted. Both activities develop at about 3 weeks of age, are relatively radioresistant and sensitive to cyclophosphamide, and are present in nude mice. In collaboration with J. J. Trentin and G. Cudcowicz, R. Kiessling (personal communication) performed some comparative experiments. The bone-seeking isotope, which depresses bone marrow resistance (Bennett, 1973), also was found to suppress NK activity. One discrepant finding has been the slight effects of silica on NK activity, in contrast to the depressive effects of this agent on bone marrow resistance (Lotzovi et al., 1975). Our finding that NK cells have some cytotoxic activity against bone marrow cells (Section II1,A) provides more direct evidence for the possible relationship between these phenomena. The ability to boost NK activity with allogeneic bone marrow cells (Section II,D) and the presence of high levels of NK reactivity in the bone marrow after boosting (Section IV,A) further support this possibility. However, in contrast to the results described in the previous paragraph, we have found an appreciable amount of NK reactivity persisting in mice after treatment with 89Sr,and inoculation of LCMV into such mice caused some, albeit lower than normal, augmentation of NK activity. The reactivity of NK cells against normal thymocytes could also explain the difficulties in transplanting thymus grafts from donors older than 3 weeks into nude mice (Radov et al., 1975). Also the higher levels of autoreactivity in neonatally thymectomized mice and the inhibition of this reactivity by T H F (Small and Trainin, 1975) raise the possibility that natural cell-mediated cytotoxicity is also involved in this phenomenon. It is also possible that NK activity is involved in resistance against infections by viruses and other microorganisms. Several studies have indicated that nude mice are relatively resistant to infection by some agents, e.g., myxoviruses (Haller and Lindenman, 1974), Listeria rnonocytogenes (Emmerling et al., 1975), Brucella abortus (Cheers and Waller, 1975), and Candida albicans (Cutler, 1976; Rogers et al., 1976). The observation that many microbial agents can induce a rapid


365

augmentation of NK activity suggests that the microorganisms themselves or infected host cells may be targets for this reactivity.

C. IMPLICATIONS OF NATURALCYTOTOXICITY FOR IMMUNE SURVEILLANCE

The original formulations of the theory of immune surveillance (Burnet, 1957; Thomas, 1959) focused on the central role of the immune response as a natural defense against neoplasia. Only more recently has the theory been modified to stress the relationship of thymus-dependent immunity to immune surveillance (Burnet, 1970). It has been this modification of the theory which has aroused a series of criticisms of the concept of immune surveillance (Moller and Moller, 1975; Outzen et al., 1975; Schwartz, 1975; Rygaard and Povlsen, 1976), and which even led to a counter theory of immune stimulation (Prehn and Lappe, 1971). Much attention has been directed toward two apparent contradictions to the theory of immune surveillance, the relatively low incidence of tumors in nude mice and the failure of some tumors to develop in thymectomized mice. Although these data do challenge the modified concept of immune surveillance, in which thymus-dependent immunological reactions are required for effective antitumor resistance, they do not really bear on the basic theory itself. The discovery that nude mice and neonatally thymectomized mice and rats have high levels of NK activity, a potentially very effective alternative mechanism for immune surveillance, provides a good explanation for most of the available in vim data. The available information on the incidence of tumors in immunodeficient or immunosuppressed humans has also engendered controversy regarding the role of immune surveillance. With some forms of depressed immunity, the incidence of some types of tumors, especially those of the reticuloendothelial system, have been clearly increased. However, in other diseases associated with immune depression, e.g., leprosy, an increased incidence of cancer has not been noted (see review b y Melief and Schwartz, 1975). As discussed earlier (Section IV,E), this variable association of immune depression with elevated tumor incidence might be related to different effects of disease or immunosuppressive regimen on NK cell activity and other possible defense mechanisms. It will b e very important to evaluate the levels of these effector functions in the various conditions carefully to determine whether any correlate with the incidence of tumors in these patients. The other principal challenge to the concept of immune surveil-

366


lance has been that, in contrast to the antigenicity of virus-induced tumors, spontaneous tumors frequently lack detectable antigenicity and therefore might not be susceptible to control by the immune system (reviewed very recently by Klein and Klein, 1977). Much has been made of the findings that tumors arising in uitro are not more antigenic than those arising in uiuo where the immune system might have been expected to select for tumors with weak or absent tumor associated antigens (Prehn, 1971; Heidelberger, 1973). However, almost all of the negative evidence has been obtained by procedures designed to detect transplantation resistance and other immune responses which have generally been associated with immune T cell activity. If, as w e suggest, there is a role for NK activity in immune surveillance, then the question of antigenicity and resistance to tumor growth needs to be asked by protocols designed to detect this function as well as that of immune T cell-mediated cytotoxicity. For example, to detect increased resistance to challenge by tumor cells which might be induced by augmentation of NK activity, the time of challenge would probably have to be much sooner after immunization than the 1 to 2 week interval usually employed, and attempts at hyperimmunization by repeated inoculation of tumor cells might be counterproductive. In addition, the antigens associated with NK activity may be quite distinct from those detected by immune T cells. We have found that the antigens recognized on RBL-5 tumor cells by the T cell immune response induced by MSV (Herberman et al., 197413) are different from those recognized by NK cells ( H e r b e m a n et ul., 1975a). If this is true for a wide variety of tumors, then the entire question of antigenicity, or lack thereof, with regard to the role of immune surveillance will need to be reexamined. Addendurn

In the last year, since this review was originally prepared, there have been several recent developments on the nature of NK cells and on factors influencing the levels of natural cell-mediated cytotoxicity, which add substantially to the understanding of this phenomenon. Therefore, the following additions should be related to the appropriate sections of the review.

To SECTION I I , D , l Ojo et al. (1978) have confirmed that Corynebucterium parvum can boost when given intraperitoneally, but they and Savary and Lotzovi


367

(1978) found that intravenous inoculation led to depressed NK activity. Welsh and Zinkernagel (1977) have confirmed that LCMV induces rapid augmentation of NK activity. MacFarlan et a l . (1977) showed that Semliki Forest virus could also rapidly boost NK activity. In all of these studies, the characteristics of the effector cells after boosting were the same as those seen with NK cells. After Oehler et a l , ( 1978b) found that poly I : C would strongly boost NK reactivity in rats, we performed extensive studies in mice to determine the role of interferon in augmenting NK activity. Poly I : C and a variety of other interferon-inducing agents were shown to be able to induce considerable increases in cytotoxic reactivity, and the peak responses occurred at around the time of previously described rises in interferon levels (Djeu et al., 1978). Administration of crude or purified virus-induced interferon also induced an increase in NK activity and this could be detected within 2 hours. Gidlund et a l . (1978) have also shown that interferon inducers and interferon could induce augmentation of NK activity and that simultaneous administration of anti-interferon could efficiently block the effects of the interferon inducers and could partially block the effects of C. parvum. It therefore appears likely that interferon plays a central role in the boosting of NK activity and it will be of interest to determine what role it has in inducing or maintaining the spontaneous levels of natural cytotoxicity. The mechanism of the effects of interferon on NK cells remains to be determined, but these may be related in some way to the ability of interferon to also augment cytotoxic reactivity of immune T cells (Lindahl et al., 1972) and of macrophages (Schultz et al., 1977).

To SECTION II,D,2 Oehler et al. (197813) recently studied the effects of a variety of agents on the cytotoxic reactivity of normal rats. The results were quite similar to those in mice. C. paruum, LCMV, and Kilham rat virus all strongly boosted reactivity, with a peak in most lymphoid organs at around 3 days. The specificity of the augmented cytotoxicity, and the cell surface characteristics of the effector cells were indistinguishable from the specificity and characteristics of rat NK cells. Inoculation of BN rats, which have low levels of natural cytotoxicity, with C. parvum induced cytotoxic reactivity in the peritoneal cavity as high as that obtained with WIFu rats. In contrast, the boosted activity in BN spleens was substantially below that observed in WIFu spleens after C. paruum. A new finding which was made in the rats was that poly I : C was able

368


to strongly boost NK reactivity, and the peak effects occurred within 1 day. Other polynucleotides, including poly A : U which has strong adjuvant properties, had no such effect on NK. Since poly I : C, and not the other polynucleotides tested, is a potent inducer of interferon, this led to the hypothesis that boosting of NK is mediated by interferon. As discussed above, this hypothesis has been correct for boosting of NK activity in mice.

To SECTION II,E,l With the recent findings that in uiuo administration of poly I : C or interferon could rapidly boost NK reactivity, we have also examined their effects on cultured NK cells (Djeu et al., 1978). Exposure of normal spleen cells to poly I : C for as little as 2 hours resulted in a substantial increase in cytotoxic reactivity. Crude and purified interferon preparations had similar effects and these could be seen after preincubation for only 45 minutes. These in uitro boosting effects were completely inhibited by the addition of small amounts of a specific anti-mouse interferon. Thus, as described above in the section on in uiuo augmentation of NK activity, interferon may also play a central role in the augmentation, and perhaps even the maintenance, of NK activity in uitro. We have recently observed that macrophages may influence the levels of NK activity in uitro. Overnight cultures of normal spleen cells in the presence of adherent splenic or peritoneal macrophages resulted in higher cytotoxic reactivity. At least some of this effect may be attributable to the production of interferon by these macrophages. The in uitro boosting of NK activity by poly I : C was found to be dependent on the presence of macrophages, whereas interferon was able to boost cultures of spleen cells depleted of adherent cells.

To SECTION IV,B,2 Initial studies of rat NK cells for complement and Fc receptors yielded negative results (Nunn et al., 1976; Oldham et al., 1977). However, using the same more sensitive approach as that described above for mouse NK cells, rat NK cells were also shown to have detectable Fc receptors (Oehler et al., 1978a). T o SECTION IV,B,3 Recent studies by Ortaldo and Robbins (1978) have shown that the in uitro generation of cytotoxic cells from Fc negative precursors in-


369

volves the cooperation of radioresistant E-rosetting cells with radiosensitive null cells which lack detectable surface immunoglobulins. It will be of interest to determine what role, if any, interferon plays in the generation of this reactivity. T o SECTION IV,E Some interesting differences in the effects of immunosuppressive treatments have been noted when their effects on boosting of rat NK activity b y poly I : C were studied (Oehler and Herberman, 1978). Hydrocortisone, x-irradiation, and cyclophosphamide, at doses which had suppressive effects on NK, had little or no effect on the ability of rats to be boosted by poly I : C . To explain these findings, it was postulated that the precursors of NK cells were resistant to these treatments whereas the NK cells themselves were sensitive. T o SECTION V,B,2 Studies on the possible role of immunoglobulins in NK have continued. When protein A or F(ab’), anti-human F(ab’), was added to the cytotdxicity says, ADCC but not NK was significantly inhibited (Kay, 1978). However, Troye et a l . (1977) found that Fab anti-human Ig could inhibit the cytotoxicity of normal lymphocytes against bladder carcinoma-derived cell lines. In recent experiments, we have also examined the effects of a F(ab’), anti-mouse F(ab’), reagent (generously provided by Dr. T. Chusid), which would be expected to interact with cytophilic antibody. This had no inhibitory effect on mouse NK activity. T o SECTION V,C Recently, Koide and Takasugi (1977) have suggested that human NK is mediated b y arming of K by natural antibodies. The reduction in NK activity by trypsinization or by incubation at 37°C would be consistent with this, since such treatments might remove the immunoglobulins needed for reactivity. Based on this possibility, Kay (1978) has performed a series of experiments in which untreated and trypsinized lymphocytes were incubated with autologous or allogeneic serum and with culture fluids from explanted lymphocytes. Thus far he has been unable to confirm the results of Koide and Takasugi. The cells incubated with serum or culture fluids have not been consistently found to have increased cytotoxic reactivity against the tumor target cells. It is possible that some unidentified variables are critical to the success of

370


such experiments. However, the failure to inhibit NK with anti F(ab’), reagents tends to rule out the central role of arming antibodies.

To SECTIONVIII,A Several investigators have recently noted a good correlation between sensitivity of tumor cells to cytotoxicity by normal lymphocytes and inhibition of growth in uivo. Warner et aE. (1977) showed that tumors sensitive to NK grew poorly in nude mice. Harmon et a l . (1977) found that rat methylcholanthrene-induced sarcomas that were sensitive to cytotoxicity by normal spleen cells were also inhibited in their in vivo growth when admixed with spleen cells.

REFERENCES Aaronson, S. A., and Stephenson, J. R. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 19571961. Aoki, T., Boyse, E. A., and Old, L. J. (1966). Cancer Res. 26, 1415-1419. Aoki, T., Herberman, R. B., Hartley, J. W., Liu, M., Walling, M. J., and Nunn, M. (1977). J. Natl. Cancer Inst. 58, 1069-1078. Bakacs, T., Gergely, P., Cornain, S., and Klein, E. (1977). Submitted for publication. Basch, R. S., and Goldstein, G. (1975). Cell. Immunol. 20, 218-228. Bean, M. A., Bloom, B. R., Herberman, R. B., Old, L. J., Oettgen, H. F., Klein, G., and Terry, W. D. (1975). Cancer Res. 35,2902-2913. Becker, S., and Klein, E. (1976). Eur. J . Immunol. 6, 892-898. Becker, S., Feny6, E. M., and Klein, E. (1976).Eur. J. Immunol. 6, 882-885. Bennett, M. (1973).]. Immunol. 110,510-516. Berenson, J. R., Einstein, A. B., Jr., and Fefer, A. (1975).J.Immunol. 115, 234-238. Berkelhammer, J., Mastrangelo, M. J., Laucius, J. F., Bodurtha, A. J., and Prehn, R. T. (1975). Int. J. Cancer 16, 571-578. Blair, P. B., and Lane, M. A. (1975a).J. Immunol. 114, 17-23. Blair, P. B., and Lane, M. A. (197513).J. Immunol. 115, 184-189. Blair, P. B., Lane, M. A., and Mar, P. (1976).J.Immunol. 116, 610-614. Bonmassar, E., Campanile, F., Houchens, D., Crino, L., and Goldin, A. (1975). Transplantation 20, 343-346. Bonnard, G. D. (1978). In “Immunodiagnosis of Cancer” (R. B. Herberman and K. R. McIntire, eds.). Dekker, New York (in press). Boyer, P. J. J., and Fahey, J. L. (1976).J. Immunol. 116,202-209. Boyse, E. A., and Old, L. J. (1969).Annu. Reu. Genet. 3, 269-290. Brier, A. M., Chess, L., and Schlossman, S. F. (1975).J.Clin. Inuest. 56, 1580-1586. Bukowski, R. M., Barna, B., Deodhar, S. D., and Hewlett, J. S. (1976). Cancer 38, 1962-1967. Burk, M. W., Yu, S., Ristow, S. S., and McKhann, C. F. (1975).Int.]. Cancer 15,99-108. Burnet, F. M. (1957). Br. Med. /. 1, 779-786 and 841-847. Burnet, F. M. (1970). Prog. E x p . Tumor Res. 13, 1-27. Byers, V. S., Levin, A. S., Hackett, A. J., and Fudenberg, H. H. (1975).J. Clin. Inuest. 55, 500-512. Campbell, A. C., Waller, C., Wood, J., Aynsley-Green, A., and Yu, V. (1974).Clin. E x p . Immunol. 18, 469-482.


37 1

Campbell, A. C., Wiernik, G., Wood, J., Hersey, P., Waller, C. A., and MacLennan, J . C. M. (1976).Clin. E x p . Immunol. 23, 200-208. Canevari, S., Fossati, G., and Dellaporta, G. (1976).J.Natl. Cancer Inst. 56,705-709. Cannon, G. B., Bonnard, G. D., Djeu, J., West, W. H., and Herberman, R. B. (1977).Int. J. Cancer 19,487-497. Canty, T. G., and Wunderlich, J. R. (1970).J.Natl. Cancer Inst. 45,761-772. Castro, J. E. (1972). Nature (London), New Biol. 239,83-84. Cerottini, J.-C., and Brunner, K. T. (1974).Adu. Immunol. 18,67-132. Cheers, C., and Waller, R. (1975).J.Immunol. 115,844-847. Collavo, D., Colombatti, A., Chieco-Bianchi, L., and Davies, A. J. S. (1974). Nature (London)249, 169-170. Connolly, J. M., Schwartz, R. H., Hanwerger, B. S., and Wunderlich, J. R. (1975). Transplantation 20, 186-192. Cordier, G., Samarut, C., and Revillard, J. P. (1976). In “Leukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roos, and W. P. Zeijlemaker, eds.), pp. 619-626. Academic Press, New York. Cutler, J. E. (1976).J.Reticuloendothel. S O C . 19, 121-124. Dean, J. H., Silva, J . S., McCoy, J . L., Leonard, C. M., Cannon, G. B., and Herberman, R. B. (1975).J.Immunol. 115, 1449-1455. Dennert, G., and Lennox, E. (1972). Nature (London),New Biol. 238, 114-115. DeVries, J. E., Cornain, S., and Rumke, P. (1974). Int. J. Cancer 14,427-434. DeVries, J. E., Meyerung, M., Van Dongren, A., and Rumke, P. (1975).1nt.J.Cancer 15, 30 1-306. Djeu, J. Y., Heinbaugh, J. A., Holden, H. T., and Herberman, R. B. (1978)J. Imrnunol. Emnierling, P., Finger, H., and Bockemuhl, J. (1975).Infect. Immun. 12, 437-439. Epstein, A. L., Herman, M. M., Kim, H., Dorfman, R. F., and Kaplan, H. S. (1976h Cancer 37, 2158-2176. Fidler, I. J., Caines, S., and Dolan, Z. (1976). Transplantation 22, 208-212. Forbes, J., Konda, S., Schimpff, R. D., and Smith, R. T. (1973).Fed. Proc., Fed. Am. S O C . E x p . Biol. 32, 867. Forman, J., and Britton, S. (1973).]. E x p . Med. 137, 369-386. Fossati, G., Colnaghi, M. I., Dellaporta, G., Cascinelli, N., and Veronesi, U. (1971).Int. J. Cancer 8, 344-350. Frelinger, J. A., and Murphy, D. B. (1976). Immunogenetics 3,481-487. Gallagher, M . T., Lotzovi, E., and Trentin, J. J. (1976).Biomedicine 25, 1-3. Gidlund, M., Orn, A., Wigzell, H., Senik, A., and Gresser, I. (1978).Nature, submitted. Gillette, R. W., and Fox, A. (1975). Cell. Immunol. 19, 328-335. Gillette, R. W., and Lowery, L. T. (1976). Cancer Res. 36, 4008-4014. Giovanella, B. C., Yim, S. O., Morgan, A. C., Stehlin, J. S., and Williams, L. J., Jr. (1973). J. Natl. Cancer Inst. 50, 1051-1053. Giovanella, B. C., Stehlin, J. S., and Williams, L. J., Jr. (1974).J.Natl. Cancer Inst. 52, 921-930. Glaser, M., Bonnard, G. D., and Herberman, R. B. (1976a).J.Immunol. 116, 430-436. Glaser, M., Lavrin, D. H., and Herberman, R. B. (1976b).J.Immunol. 116, 1507-1511. Glaser, M., Djeu, J. Y., Kirchner, H., and Herberman, R. B. ( 1 9 7 6 ~J. ) . Zmmunol. 116, 1512-1519. Glimcher, L., Shen, F. W., and Cantor, H. (1977).J. E x p . Med. 145, 1-9. Goldstein, G., Scheid, M., Hammerling, U., Boyse, E. A., Schlesinger, D. H., and Niall, H. D. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 11-15. Gomard, E., Leclerc, J. C., and Levy, J. (1974).Nature (London) 250,671-673. Gorczynski, R. M. (1976a).Immunology 31,607-614. Gorczynski, R. M. (1976b).Immunology 31,615-623.

372


Gorczynski, R. M. (1976~). Zmmunology 31,625-630. Gorczynski, R. M., and Norbury, C. (1974). Br. J. Cancer 30, 118-128. Greenberg, A. H., and Playfair, J. H. L. (1974). Clin. E x p . Immunol. 16, 99-110. Greenberg, A. H., and Shen, L. (1973).Nature (London), New Biol. 245,282-285. Hale, M. L., Hanna, E. E., and Hansen, C. T. (1976).Nature (London) 260,44-45. Haller, O., and Lindenmann, J. (1974).Nature (London) 250,679-680. Harding, B., Pudifin, D. J., Gotch, F., and MacLennan, I. C. M. (1971).Nature (London), New Biol. 232, 80-81. Harmon, R.C., Clark, E. A., Red&, A. L., Hildemann, W. H., and Mullen, Y. (1977).Int. J. Cancer 20, 748-758. Heidelberger, C. (1973).Adu. Cancer Res. 18, 317-366. Hellstrom, I,, Hellstrom, K. E., Sjogren, H. O., and Warner, G. A. (1971).Int. J . Cancer 7, 1-16. Hellstrom, I., Hellstrom, K. E., Sjogren, H. O., and Warner, G. A. (1973).Znt. J. Cuncer 11, 116-122. Heppner, G. H., Stolbach, L., Byrne, M., Cummings, F. J., McDonough, E., and Calabresi, P. (1973. Znt. J. Cancer 11, 245-260. Heppner, G. H., Henry, E., Stolbach, L., Cummings, F. J., McDonough, E., and Calabresi, P. (1975). Cancer Res. 35, 1931-1937. Herberman, R. B., and Aoki, T. (1972).J.E x p . Med. 136,94-111. Herberman, R. B., and Oldham, R. K. (1975).J. Natl. Cancer Inst. 55,749-753. Herberman, R. B., Rosenberg, E. B., Halterman, R. H., McCoy, J. L., and Leventhal, B. G. (1972).Natl. Cancer Inst., Monogr. 35, 259-266. Herberman, R. B., Nunn, M. E., Lavrin, D. H., and Asofsky, R. (1973).J.Natl. Cancer Inst. 51, 1509-1512. Herberman, R. B., Ting, C. C., Kirchner, H., Holden, H., Glaser, M., Bonnard, G. D., and Lavrin, D. (1974a). Prog. Immunol., Int. Congr. Zmmunol., 2nd, 1974 Vol. 11, pp. 285-295. Herberman, R. B., Aoki, T., Nunn, M., Lavrin, D. H., Soares, N., Gazdar, A., Holden, H., and Chang, K. S. S. (1974b).J.Natl. Cancer Inst. 53, 1103-1111. Herberman, R. B., Nunn, M. E., and Lavrin, D. H. (1975a).Znt.J. Cancer 16,216-229. Herberman, R. B., Nunn, M. E., Holden, H. T., and Lavrin, D. H. (197513).Znt.J.Cancer 16,230-239. Herberman, R. B., Oldham, R. K., and Connor, R. J. (1976a). In “In Vitro Methods in Cell Mediated and Tumor Immunity” (B. R. Bloom and J. R. David, eds.), pp. 481-488. Academic Press, New York. Herberman, R. B., Nunn, M. E., and Holden, H. T. (1976b).In “In Vitro Methods in Cell Mediated and Tumor Immunity” (B. R. Bloom and J. R. David, eds.), pp. 489-495. Academic Press, New York. Herberman, R. B., Nunn, M. E., Holden, H. T., Staal, S., and Djeu, J. Y. (1977a).Znt.J. Cancer 19, 555-564. Herberman, R. B., Bartram, S., Haskill, J. S., Nunn, M. E., Holden, H. T., and West, W. H. (1977~4. j . rtntnunoi. 119,322-326. Hersey, P., Edwards, A,, Edwards, J., Adams, A. E., Milton, G. W., and Nelson, D. S. (1975). Int. J. Cancer 16, 173-183. Holden, H. T., Oldham, R. K., Ortaldo, J. R., and Herberman, R. B. (1976).In “In Vitro Methods in Cell Mediated and Tumor Immunity” (B. R. Bloom and J. R. David, eds.), pp. 723-729. Academic Press, New York. Holden, H. T., Oldham, R. K., Ortaldo, J. R., and Herberman, R. B. (1977). J. Natl. Cancer Inst. 58, 611-622. Holtermann, 0. A., Klein, E., and Casale, G. P. (1973). Cell. Immunol. 9, -139-352.


373

Ihle, J. N., Yurconic, M., Jr., and Hanna, M. G., Jr. (1973).]. E x p . Med. 138, 194-208. Ikehara, S., Hamashima, Y., and Masuda, T. (1975).Nature (London) 258, 335-337. Jondal, M., and Pross. H. (1975).Znt. J. Cancer 15,596-605. Jondal, M., Svedmyr, E., Klein, E., and Singh, S. (1975).Nature(London) 225,405-407. Kanner, S. P., Mardiney, M. R., Jr., and Mangi, R. J. (1970).J.Zmmunol. 105,1052-1057. Kay, H. D. (1978).Fed. Proc., in press. Kay, H. D., Bonnard, G. D., West, W. H., and Herberman, R. B. (1977).J. Inamunol. 118,2058-2066. Kieding, R., Klein, E., and Wigzell, H. ( 1 9 7 5 ~ )Eur.J. . Immunol. 5, 112-117. Kiessling, R., Klein, E., Pross, H., and Wigzell, H. (197513).J . Immunol. 5, 117-121. Kiessling, R., Petrinyi, G., Klein, G., and Wigzell, H. ( 1 9 7 5 ~ )Znt. . J . Cancer 15, 933940. Kiessling, R., Petranyi, G., Karre, K., Jondal, M., Tracey, D., and Wigzell, H. (1976a).J. E x p . Med. 143,772-785. Kiessling, R., Petrinyi, G., Klein, G., and Wigzell, H. (1976b).Znt. J. Cancer 17, 1-7. Kirov, S. M. (1974).Eur. J . Zmmunol. 4, 739-745. Kiuchi, M., and Takasugi, M. (1976).J.Natl. Cancer Inst. 56,575-582. Klein, G., and Klein, E. (1977). Transplant. Proc. 9, 1095-1 104. Koide, Y., and Takasugi, M. (1977).J.Nat/. Cancer Inst. 59, 1099-1106. Krueger, J. G., Segal, R. A., and Moyer, R. C. (1977).Cancer Res. 37, 320-322. Lamon, E. W., Skurzak, H. M., and Klein, E. (1972). Znt.J. Cancer 10, 581-588. Lamon, E. W., Skurzak, H. M., Anderson, B., Whitten, H. D., and Klein, E. (1975).J. Zmmunol. 114, 1171-1176. Lane, M. A., Roubinian, J., Slomich, M., Trefts, P., and Blair, P. B. (1975).]. Ztnmunol. 114, 24-29. Lavrin, D. H., Herberman, R. B., Nunn, M., and Soares, N. (1973).]. Natl. Cancer Inst. 51, 1497-1508. Lee, J. C., and Ihle, J. N. (1977).J.Zmmunol. 118, 928-934. Levin, A. C., Massey, R. J., Deinhardt, F., Schauf, V., and Wolter, J. (1975).In “Neoplasm Immunity: Theory and Application” (R. G. Crispen, ed.), pp. 107-126. ITR, Chicago, Illinois. Levin, A. C., Massey, R. J., Wolter, J., Schauf, V., and Deinhardt, F. (1976a).Proc. Am. Assoc. Cancer Res. 17,65. Levin, A. C., Massey, R. J., and Deinhardt, F. (1976b). Fed. Proc., Fed. Am. SOC. E x p . Biol. 35, 472. Levy, N. L., Mahaley, M. S., and Day, E. D. (1972).Cancer Res. 32, 477-482. Lindahl, P., Leary, P., and Gresser, I. (1972).Proc. Nat. Acad. Sci. USA 69, 721-725. Loor, F., and Roelants, G. E. (1974). Nature (London) 251,229-230. Loor, F., Hagg, L.-B., Mayor, K. S., and Roelants, G. E. (1975).Nature (London) 255, 657-658. Loor, F., Roelants, G. E., Kindred, B., Mayor, K. S., and Hagg, L.-B. (1976).In “Progress in Differentiation Research” (N. Muller-%rat, ed.), pp. 559-566. North-Holland Publ., Amsterdam. Lopez, D. M., Ortiz-Muniz, G., and Sigel, M. M. (1976). Proc. SOC. E x p . Biol. Med. 151, 225-230. Lotzovi, M., Gallagher, M. T., and Trentin, J. J. (1975). Biomedicine 22, 387-392. Lozzio, B. B. (1976). Biomedicine 24, 144-147. Lozzio, B. B., Machado, E. A., Lozzio, C. B., and Lair, S. (1976)./. E x p . Med. 143, 225-231. McCoy, J. L., Herberman, R. B., Perlin, E., Levine, P. H., and Alford, C. (1973a). Proc. Am. Assoc. Cancer Res. 14, 107.

374


McCoy, J. L., Herberman, R. B., Rosenberg, E. B., Donnelly, F. C., Levine, P. H., and Alford, C. (1973b). Natl. Cancer lnst., Monogr. 37,59-67. MacDermott, R. P., Chess, L., and Schlossman, S. F. (1975). Clin. lmrnunol. lmmunopathol. 4, 415-422. MacDonald, H. R., and Bonnard, G. D. (1975). Scand. J. lmrnunol. 4, 129-138. MacFarlan, R. I., Burns, W. H., and White, D. 0. (1977).J.lmrnunol. 119, 1569-1575. Machado, E. A,, Lozzio, B. B., and Lair, S. V. (1976).In “Immuno-Aspects of the Spleen” (J. Battisto, ed.). (in press). North-Holland Publ., Amsterdam Maguire, H., Jr., Outzen, H. C., Custer, R. P., and Prehn, R. T. (1976).J. Natl. Cancer lnst. 57, 439-442. Martin, S . , and Martin, J. (1975). Nature (London) 256, 498-499. Martin-Chandon, M., Vanky, F., Carnaud, C., and Klein, E. (1975). lnt. J. Cancer 15, 342-350. Matthews, N., and MacLaurin, B. P. (1974).Aust. J. E x p . B i d . 52, 655-661. Melief, C. J. M., and Schwartz, R. S. (1975). In “Cancer: A Comprehensive Treatise” (F. F. Becker, ed.), Vol. I, pp. 121-160. Plenum, New York. Mellors, R. C., Aoki, T., and Huebner, R. J. (1969).J.E x p . Med. 129, 1045-1062. Moller, G., and Moller, E. (1975).J.Natl. Cancer lnst. 55,755-759. Morales, A., Bonnard, G. D., Dean, J. H., and Herberman, R. B. (1977). Fed. Proc., Fed. Am. SOC. E x p . Biol. 36, 1325. Niederhuber, J. E., Frelinger, J. A., Dine, M. S., Shoffner, P., Dugan, E., and Shreffler, D. C. (1976).J. E x p . Med. 143, 372-381. Nowinski, R. C., and Kaehler, S. L. (1974). Science 185,869-871. Nunn, M. E., Djeu, J. Y., Glaser, M., Lavrin, D. H., and Herberman, R B. (1876)J. Natl. Cancer lnst. 56, 393-399. Oehler, J. R., and Herberman, R. B. (1978). lnt. J. Cancer, 21,221-229. Oehler, J. R., Lindsay, L. R., Nunn, M. E., and Herberman, R. B. (1978a).1nt.J. Cancer 21,204-209. Oehler, J. R., Lindsay, L. R., Nunn, M. E., Holden, H. T., and Herberman, R. B. (1978b). Int. J. Cancer, 21, 210-220. Ojo, E., Haller, O., Kimura, A., and Wigzell, H. (1978).Submitted for publication. Oldham, R. K., Siwarski, D., McCoy, J. L., Plata, E. J., and Herberman, R. B. (1973). Natl. Cancer lnst., Monogr. 37, 49-58. Oldham, R. K., Djeu, J. Y., Cannon, G. B., Siwarski, D., and Herberman, R. B. (1975).J. Natl. Cancer lnst. 55, 1305-1318. Oldham, R. K., Weiner, R. S., Math&, G., BrBard, J., Simmler, M. C., Carde, P., and Herberman, R. B. (1976a).lnt. J. Cancer 17, 326-377. Oldham, R. K., Dean, J. H., Cannon, G. B., Ortaldo, J. R., Dunston, G., Applebaum, F., McCoy, J. L., Djeu,J., and Herberman, R. B. (197613).1nt.J.Cancer 18, 145-155. Oldham, R. K., Ortaldo, J. R., and Herberman, R. B. (1977). Cancer Res. 37,4467-4474. Ortaldo, J. R., and Bonnard, G. D. (1977). Fed Proc., Fed. Am. SOC. E x p . Biol. 36, 1325. Ortaldo, J. R., Oldham, R. K., Holden, H. T., and Herberman, R. B. (1976). Cell. lmmunol. 25,60-73. Ortaldo, J. R., Kay, H. D., and Bonnard, G. D. (1977a).Proc.Leukocyte Cult. Conf. l l t h , 1976 (D. D. Lucas, ed.), pp. 542-544. Academic Press, New York. Ortaldo, J. R., Bonnard, G. D., and Herberman, R. B. (1977b).J.lmmunol. 119, 13511357. Ortaldo, J. R., Oldham, R. K., Cannon, G. C., and Herberman, R. B. (1977~). J. Natl. Cancer lnst. 59, 77-82. Ortaldo, J. R., and Robbins, M. L. (1978). Fed. Proc., in press.


375

Ortiz de Landazuri, M. O., and Herberman, R. B. (1972a).J. E x p . Med. 136,969-983. Ortiz de Landazuri, M., and Herberman, R. B. (1972b).Nature (London),New Biol. 238, 18- 19. O’Toole, C., Perlmann, P., Unsgaard, B., Moberger, G., and Edsmyr, F. (1972).Int. J. Cancer 10,77-91. Outzen, H. C., Custer, R. P., Eaton, G. J., and Prehn, R. T. (1975).J. Reticuloendothel. S O C . 17, 1-9. Ozzello, L., Sordat, B., Merenda, C., Carrel, S., Hurlimann, J., and Mach, J. P. (1974)J Natl. Cancer Znst. 52, 1669-1672. Parrillo, J. E., and Fauci, A. S. (1977a). Submitted for publication. Parrillo, J. E., and Fauci, A. S. (1977b). Submitted for publication. Pavie-Fischer, J,, Kourilsky, F. M., Picard, F., Banzet, P., and Puissant, A. (1975). CEin. E x p . Zmmunol. 21,430-441. Perlmann, P., Perlmann, H., and Biberfeld, P. (1972).J. Zmmunol. 108,558-561. Perlmann, P., Biberfeld, P., Larsson, A., Perlmann, H., and Wihlin, B. (1975).In “Membrane Receptors of Lymphocytes” (M. Seligman, ed.), pp. 161-169. North-Holland Publ., Amsterdam. Peter, H. H., Pavie-Fischer, J., Fridman, W. H., Aubert, C., Cesarini, J., Roubin, R., and Kourilsky, F. M. (19754. J. Immunol. 115, 539-548. Peter, H. H., Kalden, J. R., Seeland, P., Diehl, V., and Eckert, G. (197513) Clin. E x p . Zmmunol. 20, 193-207. Peter, H. H., Knoofp, F., and Kalden, J. R. (1976a).Z . Zmmun. (itaets)forsch.,E x p . Klin. Zmmunol. 151, 263-281. Peter, H. H., Eife, R. F., and Kalden, J. R. (1976b).J.Immunol. 116, 342-348. Petrinyi, G. G., Benczur, M., Onody, C. E., and Hollan, S. R. (1974).Lancet 1,736. PebAnyi, G. G., Kiessling, R., and Klein, G. (1975). Zmmunogenetics 2, 53-61. PetrAnyi, G. G., Kiessling, R., Povey, S., Klein, G., Herzenberg, E., and Wigzell, H. (1976). Immunogenetics 3, 15-28. Pfizenmaier, K., Trostmann, H., Rollinghoff, M., and Wagner, H. (1975). Nature (London) 258,238-240. Pierce, G. E., and DeVald, B. L. (1975).Cancer Res. 35, 1830-1839. Pollack, S. B., and Nelson, K. (1974).Znt. J . Cancer 14, 522-529. Pollack, S. B., and Nelson, K. (1975). Znt. J. Cancer 16, 339-346. Pollack, S. B., and Nelson, K. (1976). 1nt.J. Cancer 18, 250-254. Prehn, R. T. (1971). In “Immune Surveillance” (R. T. Smith and M. Landy, eds.). pp. 451-462. Academic Press, New York. Prehn, R. T., and LappB, M. A. (1971).Transplant. Reu. 7, 26-54. Pritchard, H., and Micklem, H. S. (1973). Immunology 14, 597-607. Pross, H. F., and Jondal, M. (1975). Clin. E x p . Zmmunol. 21, 226-235. Radov, L. A,, Sussdorf, D. H., and McCann, R. L. (1975).lmmunology 29, 977-988. Raff, M. C. (1971). Transplant. Reu. 6,52-80. Raff, M. C. (1973).Nature (London) 246, 350-351. Raff, M. C., and Wortis, H. H. (1970). Zmmunology 18,931-942. Rager-Zisman, B., Grose, C., and Bloom, B. R. (1976).Nature (London) 260, 369-370. Ramseier, H. ( 1975).Zmmunogenetics 1, 507-510. Roelants, G. E., Mayor, K. S., Hagg, L.-B., and Loor, F. (1976).Eur. J . Zmmunol. 6, 75-81. Rogers, T. J., Balish, E., and Manning, D. D. (1976).J.Reticuloendothel. Soc. 20, 291298. Rosenberg, E. B., Herberman, R. B., Levine, P. H., Halterman, R. H., McCoy, J. L., and Wunderlich, J. R. (1972).Znt. J . Cancer 9, 648-658.

376


Rosenberg, E. B., McCoy, J. L., Green, S. S., Donnelly, F. C., Siwarski, D. F., Levine, P. H., and Herberman, R. B. (1974).J.Natl. Cancer Inst. 52, ,345-352. Rotter, V., and Trainin, N. (1975).Transplantation 20, 68-74. Rygaard, J., and Povlsen, C. 0. (1969).Acta Pathol. Microbiol. Scand. 77, 758-760. Rygaard, J., and Povlsen, C. 0. (1976). Transplant. Reu. 28, 43-61. Snkseli, E., Imir, T.,and Mikell, 0. (1975).J. Immunol. 115, 1488-1492. Santoli, D., Trinchieri, G., Zmijewski, C. M., and Koprowski, H. (1976).J . Immunol. 117,765-770. Sato, H., Boyse, E. A., Aoki, T., Iritani, C., and Old, L. J. (1973).J . Exp. Med. 138, 593-606. Savary, C. A., and Lotzovi, E. (1978).J.Immunol. 120,239-242. Scheid, M. P., Hoffmann, M. K., Komuro, K., Hammerling, U., Abbott, J., Boyse, E. A., Cohen, G. H., Hooper, J. A., Schulof, R. S., and Goldstein, A. L. (1973).J.E x p . Med. 138, 1027-1032. Scheid, M. P., Goldstein, G., and Boyse, E. A. (1975).Science 190, 1211-1213. Scher, I., Ahmed, A., Strong, D. M., Steinberg, A. D., and Paul, W. E. (1975).J . E x p . Med. 141,788-802. Schmidt, M., and Good, R. A. (1975).J.N u t / . Cancer Inst. 55,81437, Schmidt, M., and Good, R. A. (1976).Lancet 1, 39. Schultz, R. M., Papamatheakis, J. D., and Chirigos, M. A. (1977).Science 197,674-676. Schwartz, R. S. (1975). N . Engl. J . Med. 293, 181-184. Sendo, F., Aoki, T., Boyse, E. A., and Buofo, C. K. (1975)./. Natl. Cancer Inst. 55, 603-609. Shellam, G. R. (1977).Inst. J. Cancer 19, 225-235. Shellam, G. R., and Hogg, N. (1977).Int. J . Cancer 19,212, 224. Shin, S.-I., Freedman, V. H., Risser, R., and Pollack, R. (1975). Proc. Natl. Acad. Sci. U.S.A. 72,4435-4439. Shoji, M., and McKhann, C.F. (1971).Proc. Am. Assoc. Cancer Res. 12,99. Shustik, C., Cohen, I. R., Schwartz, R. S., Latham-Griffin, E., and Waksal, S. D. (1976). Nature (London) 263,699-701. Sinkovics, J. G., Dreyer, D. A., Shirato, E., Cabiness, J. R., and Schullenberger, C. C. (1971).Tex. Rep. Biol. Med. 29, 227-242. Skov, C. B., Holland, J. M., and Perkins, E. H. (1976).J.Natl. CancerInst. 56, 193-195. Small, M., and Trainin, N. (1975). Cell. Immunol. 20, 1-11. Stejskal, V., and Perlmann, P. (1976).Eur. J . Immunol. 6, 347-352. Stejskal, V., Holm, G., and Perlmann, P. (1973). Cell. Immunol 8, 71-81. Stevenson, G. T., and Laurence, D. J. R. (1975).Int. J . Cancer 16,887-896. Stout, R. D., Waksal, S. D., and Herzenberg, L. A. (1976).J. E x p . Med. 144, 54-68. Stutman, 0. (1974). Science 183,534-536. Stutman, 0.(1975).J. Immunol. 114, 1213-1217. Svedmyr, E., and Jondal, M. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 1622-1626. Svedmyr, E., Wigzell, H., and Jondal, M. (1974). Scand. J . Immunol. 3, 499-508. Svedmyr, E., Jondal, M., and Leibold, W. (1975).Scand. J . Irnrnunol. 4, 721-734. Takasugi, M., and Mickey, M. R. (1976).]. Natl. Cancer Inst. 57,255-261. Takasugi, M., Mickey, M. R., and Terasaki, P. I. (1973).Cancer Res. 33, 2898-2902. Takasugi, M., Mickey, M. R., and Terasaki, P. I. (1974).J . Natl. Cancer Inst. 53, 15271538. Takasugi, M., Akira, D., and Kinoshita, K. (1975). Cancer Res. 35,2169-2176. Takasugi, M., Raniseyer, A., and Takasugi, J. (1977a). Cancer Res. 37, 413-418. Takasugi, M., Akira, D., Mullen, Y., Takasugi, J., and Ivler, R. (1977b). Proc. Int. Symp. Detection Preu. Cancer, 3rd, 1977 (in press).


377

Thomas, L. (1959). In “Cellular and Humoral Aspects of the Hypersensitive State” (H. S. Lawrence, ed.), pp. 529-530. Harper, New York. Ting, C. C., Park, J. Y., Nunn, M. E., and Herberman, R. B. (1977)J. Natl. Cancer Inst. 58, 323-330. Trentin, J. J., and Bennett, M. (1977). Transplant. Proc. 9, 1303-1306. Trinchieri, G . , Santoli, D., Zmijewski, C. M., and Koprowski, H . (1977). TranspZant. Proc. 9, 881-884. Troye, M., Perlmann, P., Pape, G. R., Spiegelberg, H. L., Naslund, I., and Gidlof, A. (1977).J. Zmmunol. 119, 1061-1067. Umiel, T., and Trainin, N. (1975). Eur. J . Zmmunol. 5,85-88. Warner, N. L., Woodruff, M. F. A., and Burton, R. C. (1977).Znt.J. Cancer 20, 146-155. Welsh, R. M., Jr., and Zinkernagel, R. M. (1977).Nature 268, 646-648. West, W. H., Cannon, G . B., Kay, H. D., Bonnard, G . D., and Herberman, R. B. (1977a).J. Zmmunol. 118,355-361. West, W. H., Boozer, R. B., and Herberman, R. B. (1977b).J.Zmmunol. 120,90-95. West, W. H., Payne, S. M.,Weese, J. L., and Herberman, R. B. ( 1 9 7 7 ~ J. ) . Immunol. 119,548-554. Williams, R. M., Leifer, J., and Moore, M. J. (1977). Transplantation 23, 283-286. Wisloff, F., Frbland, S. S., and Michaelsen, T. E. (1974). Int. Arch. Allergy Appl. Zmmunol. 47, 139. Wolfe, S . A., Tracey, D. E., and Henney, C. S. (1976).Nature (London) 262,584-586. Yutoku, M., Grossberg, A. L., Stout, R., Herzenberg, L. A., and Pressman, D. (1976). Cell. Immunol. 23, 140-157. Zarling, J. M., Nowinski, R. C., and Bach, F. H. (1975).Proc. Natl. Acad. Sci. U.SA. 72, 2780-2784. Zielske, T. V., and Golub, S. H. (1976). Cancer Res. 36, 3842-3846. Zighelboim, J., Gale, R. P., Chiv, A., Bonavida, B., Ossorio, R. C., and Fahey, J. L. (1974). Clin. Zmmunol. Zmmunopathol. 3, 193-200.


SUBJECT INDEX A ABO blood group antigens, in trophoblastic disease, 140-141 Adenomatosis, inherited colonic cancer and, 289-293 Adriamycin, mass spectometric determination of, 245-249 Aflatoxins, mass spectrometric detection of, 231 Air, mass spectrometry of pollutants in, 216-217 polycyclic aromatic hydrocarbons, 219 Animal tumors for use in cancer studies, 149-200 analysis of species of, 153-157 chemically induced, 163-167 origin of, 157-192 spontaneous tumors, 159-162 tumor transplants, 176-182 virus-induced, 167-176 Antiviral drugs, mass spectrometry of, 253 Armitage and Doll theory of oncogenesis, 74-75 B B-type oncornaviruses, structural proteins of, 39-43 Benzo[u]pyrene, metabolism of, mass spectrometric studies of, 222-223 Bile acids and sterols, mass spectrometry of, 258 Biological markers, mass spectrometry of, 253-259

C Cancer research, mass spectrometry in, 201-267 Carbohydrates, mass spectrometry of, 259 Carbon black, polycyclic aromatic hydrocarbons in, 221-222 Carcinogens, mass spectrometric studies of, 215-233 Catecholamines, mass spectrometry of, 257-258

Cell-mediated immunity, 305-377 age effects on, 310-312 augmentation of, 319-321 characteristics of, 307-324 decreased tumor growth in, 362-363 effector cells in, 333-345 effects of in vitro cultivation of lymphoid cells, 321-324 environmental factors in, 316-319 genetic factors in, 312-316 in immune surveillance, 365-366 immunosuppression effects on, 343-345 i n uiuo relevance of, 361-366 natural, 354-361 specificity of, 324-333 N K cells and K cells in, 351-361 relationship to antibody-dependent cell-mediated cytotoxicity, 345-350 his-Chloromethyl ether, mass spectrometric studies of, 225-227 Choriocarcinoma (gestational) with hydatidiform mole, 125-138 immunotherapy trials of, 143-145 native host reactions toward, 142-143 origin of, 89-125 from seemingly normal pregnancy, 97-125 Colon cancer cutaneous cells in, 299-300 environmental factors in, 282-287 fecal contents in, 300 human susceptibility to, 281-303 immunologic studies on, 296 inherited diseases and, 287-293 nuclear protein and enzyme alterations in, 296-299 proliferative abnormalities and, 293296 Computers, use in mass spectrometry, 212 C-type RNA tumor viruses enu gene-coded proteins of, 9-14 gug gene-coded proteins of, 14-24 genome structure and complexity of, 5-6 genetic mapping, 27-30 proteins of, 6-27

379

380

SUBJECT INDEX

SRC gene-coded transforming proteins of, 24-27 structural proteins of, 8-24 coded by leukemia and sarcoma genomes, 35-39 translational products of, 1-53 properties, 35 Cutaneous cells, in colonic cancer, 299300 Cyclophosphamide mass spectrometric determination of, 238-241 metabolism of, 233-238 Cytotoxicity, natural, characteristics of, 307-324

Gardner’s syndrome, colonic cancer incidence and, 292 Genetics of cell-mediated immunity, 312-316 of colonic cancer, 287-293 Gestational trophoblastic disease, 89-147 ABO blood group antigens in, 140-141 choriocarcinoma, 89- 125 HLA antigens and, 138-139 immunology of, 138-145 invasive mole, 125-138 Guinea pigs, cell-mediated immunity in, 323

D

Hexamethylamine, mass spectrometric determination of, 250 HLA antigens, trophoblastic disease and, 138- 143 Hoimones, antitumor type, mass spectrometry of, 249-250 Human chorionic gonadotropin (hCG), in trophoblastic choriocarcinomn detection, 92 Humans cell-mediated immunity in, 31 1-312, 316, 317-319, 321, 32Z324, 348 effector cells, 334, 338-342 specificity, 331-333 Hydatidiform mole, invasive mole and choriocarcinoma associated with, 125-138 follow-up studies, 132-138

Daunorubicin, mass spectrometric determination of, 245-249 Diethylstilbestrol, mass spectrometric detecton of, 232 Dinitrobenzyl aziridines, mass spectrometry of, 252 D-type oncornaviruses, structural proteins of, 39-43 E Ellipticine, mass spectrometry of, 252253 Enu gene-coded proteins, of C-type RNA tumor viruses, 9-14 Environment egects on cell-mediated immunity, 316319 role in colonic cancer, 282-287 Enzymes, in colonic cancer, 296-299

H

I

Feces, abnormal chemicals in, in colonic cancer, 300 Fisher theory of oncogenesis, 75-76 application to terminated exposures, 76-77

Immune surveillance, cell-metliated immunity in, 365-366 Iminunobiology, of trophoblastic disease, 138-145 Initiation-promotion phenomenon, of oncogenesis, 78 Inlet systems, for mass spectrometry, 207 Invasive mole, incidence of, 125-138 Ion sources, for mass spectrometry, 208-209

G

J

F

Gag gene-coded proteins, of C-type RNA

tumor viruses, 14-24

Juvenile polyposis of colon, colonic cancer incidence and, 293

38 1

SUBJECT INDEX K

of vinyl chloride, 227-228 of water pollutants, 215-216

K cells, in cell-mediated immunity, 351-354 L

Leukemia, marrow transplantation in therapy of, 269-279 Leukemia virus, structural proteins coded by, 35-39 Lung tumors, urethane-induced, multistage theory of oncogenesis in, 68-71 Lymphocytes, mass spectrometry of, 258-259 Lymphoid cells, in vitro cultivation of, effects on cell-mediated immunity, 321-324

M Marrow transplantation in leukemia therapy, 269-279 during remission, 277-278 graft vs. leukemia, 276-277 patient selection and clinical results, 270-271 recurrence data, 273-274 relapse prevention, 275-276 survival data, 271-273 Mass analyzers, for mass spectrometry, 209-2 11 Mass spectrometry advantages and limitations of, 202204 analytical techniques for, 212-215 of antineoplastic agents, 233-253 of antitumor hormones, 249-250 of biological markers, 253-259 of his-chloromethyl ether, 225-227 in cancer research, 201-267 components of systems for, 206-212 computer use in, 212 of nitrosamines, 228-231 of polychlorinated biphenyls, 224-225 of polycyclic aromatic hydrocarbons, 217-223 stable isotope dilution for, 214 theory of, 204-205 of trace elements, 232-233

Methotrexate, mass spectrometric determination of, 251-252 Mice cell-mediated immunity in, 310-311, 312-315, 316-317, 319-321, 322, 340-341, 345-346,347-348 effector cells, 333-337 specificity, 3 2 4 3 2 9 Multicell theory, of oncogenesis, 62-65 Multistage theory of oncogenesis, 65-68 with proliferative advantage of intermediate cells, 73-74 of radiation-induced tumors, 68-71 single stage of, 78-79 of urethane-induced tumors, 71-73 N

Nitrogen mustards, mass spectrometry of, 252 Nitrosamines, mass spectrometric studies of, 228-231 Nitrosoureas, mass spectrometric determination of, 241-243 NK cells, in cell-mediated immunity, 351-354 Nuclear proteins, role in colonic cancer, 296-299

0 Oldfield syndrome, colonic cancer incidence and, 292 Oncogenesis Armitage and Doll two-stage theory of, 74-76 clone growth of transformed cells and, 78-79 Fisher theory of, 75-76 implications for dose-response relationships, 83-86 initiation-promotion phenomenon of, 78 experiments involving, 79-83 multicell theory of, 62-65 multistage theory of, 65-68 single stage of, 78-83 quantitative theories of, 55-88 single stage theory of, 57-62

382

SUBJECT INDEX P

Peutz-Jeghers syndrome, colonic cancer incidence and, 293 Phorbol myristate acetate, mass spectrometric detection of, 231 Platinum coordination complexes, mass spectrometric determination of,

250-251 Polyamines, mass spectrometry of, 254-

255 Polychlorinated biphenyls, mass spectrometric analysis of, 224-225 Polycyclic aromatic hydrocarbons (PAH), mass spectrometric analysis of, 217-

223 Purines, mass spectrometric determination of, 243-245 Putrescine, mass spectrometry of, 254-

255 Pyrimidines, mass spectrometric determination of, 243-245 R

Radiation tumors induced by multistage theory of, 68-71 theories, 86 Rats cell-mediated immunity in, 310, 315,

317, 322-323, 342, 346 effector cells, 334, 337-338 specificity, 329-331 RNA-dependent DNA-polymerase, properties of, 7-8 RNA tumor viruses, C-type, translational products of, 1-53

Spermidine, mass spectrometry of, 254 Spermine, mass spectrometry of, 254-255 SRC gene-coded transforming proteins, of C-type RNA tumor viruses, 24-27 Steroids, mass spectrometry of, 256-257 Stilbenes, mass spectrometric determination of, 232 Structural proteins of B- and D-type oncornaviruses, 39-43 of C-type RNA tumor viruses, 8-24 of RNA-dependent DNA-polymerase, 8 T Therapy of cancer, animal tumors for study of, 149-200 Thymidine, incorporation of, mass spectrometry, 259 Thymosin, effects on cell-mediated immunity, 340-341 Tobacco smoke, polycyclic aromatic hydrocarbons in, 219-221 Trace elements, mass spectrometric determination of, 232-233 Translational products, of C-type RNA tumor viruses, 1-53 Tumors (See also Oncogenesis) in animals for therapy studies, 149-200 expected rates of appearances of, 56-57

v Vinyl chloride, mass spectrometric studies of 227-228 Viruses, tumors induced by, cancer studies using, 167-176 Volatiles, mass spectrometry of, 258

S W Sarcoma virus, structural proteins coded by, 35-39 Single stage theory, of oncogenesis, 57-62

Water, mass spectrometric analysis of pollutants in, 215216

CONTENTS OF PREVIOUS VOLUMES

Volume 1 Electronic Configuration and Carcinogenesis C. A . Coulson Epidermal Carcinogenesis E . V. C o w d y The Milk Agent in the Origin of Mammary Tumors in Mice L. Dmochowski Hormonal Aspects of Experimental Tumorigenesis T. U . Gardner Properties of the Agent of Rous No. 1 Sarcoma R. J . C. Harris Applications of Radioisotopes to Studies of Carcinogenesis and Tumor Metabolism Charles Heidelberger The Carcinogenic Aminoazo Dyes James A . Miller and Elizabeth C. Miller The Chemistry of Cytotoxic Alkylating Agents M . C. J . Ross Nutrition in Relation to Cancer Albert Tannenbaum and Herbert Siluerstone Plasma Proteins in Cancer Richard J . Winzler AUTHOR INDEX-SUBJECT INDEX

Volume 2 The Reactions of Carcinogens with Macromolecules Peter Alexander Chemical Constitution and Carcinogenic Activity G. M . Badger 383

Carcinogenesis and Tumor Pathogenesis 1. Berenblum Ionizing Radiations and Cancer Austin M . Brues Survival and Preservation of Tumors in the Frozen State James Craigie Energy and Nitrogen Metabolism in Cancer Leonard I). Fenninger and C. Burroughs Mider Some Aspects of the Clinical Use of Nitrogen Mustards Calvin T. K l o p p and Jeanne C . Bateman Genetic Studies in Experimental Cancer L. W. Law The Role o f Viruses in the Production of Cancer C . Oberling and M . Cuerin Experimental Cancer Chemotherapy C . Chester Stock AUTHOR INDEX-SUBJECT INDEX

Volume 3 Etiology of Lung Cancer Richard Doll The Experimental Development and Metabolism of Thyroid Gland Tumors Harold P. Morris Electronic Structure and Carcinogenic Activity and Arom'atic Molecules: New Developments A . Pullman and B. Pullman Some Aspects o f Carcinogenesis P. Rondoni Pulmonary Tumors in Experimental Animals Michael B. Shimkin

384

CONTENTS O F PREVIOUS VOLUMES

Oxidative Metabolism of Neoplastic Tissues Sidney Weinhouse AUTHOR INDEX-SUBJECT INDEX

Volume 4 Advances in Chemotherapy of Cancer in Man Sidney Farber, Rudolf Toch, Edward Manning Sears, and Donald Pinkel The Use o f Myleran and Similar Agents in Chronic Leukemias D. A. G. Galton The Employment of Methods of Inhibition Analysis in the Normal and Tumor-Bearing Mammalian Organism Abraham Goldin Some Recent Work on Tumor Immunity P. A. Gorer Inductive Tissue Interaction in Development Clifford Grobstein Lipids in Cancer Frances L. Haven and W. R. Bloor The Relation between Carcinogenic Activity and the Physical and Chemical Properties of Angular Benzacridines A. Lacassagne,N. P. BuuHoi, R. Daudd, and F. Zajdela The Hormonal Genesis of Mammary Cancer 0. Muhlbock AUTHOR INDEX-SUBJECT INDEX

Volume 5 Tumor-Host Relations R. W. Begg Primary Carcinoma of the Liver Charles Berman Protein Synthesis with Special Reference to Growth Processes both Normal and Abnormal P. N . Campbell

The Newer Concept of Cancer Toxin War0 Nakahara and Fumiko Fukuoka Chemically Induced Tumors of Fowls P. R. Peacock Anemia in Cancer Vincent E . Price and Robert E. Greenfield Specific Tumor Antigens L. A. Zilber Chemistry, Carcinogenicity, and Metabolism of BFluorenamine and Related Compounds Elizabeth K . Weisburger and John H . Weisburger AUTHOR INDEX-SUBJECT INDEX

Volume 6 Blood Enzymes in Cancer and Other Diseases Oscar Bodansky The Plant Tumor Problem Armin C . Braun and Henry N . Wood Cancer Chemotherapy by Perfusion Oscar Creech, Jr. and Edward T. Krementz Viral Etiology o f Mouse Leukemia Ludwick Gross Radiation Chimeras P. C. Koller, A. J . S . Daoies, and Sheila M . A. Doak Etiology and Pathogenesis of Mouse Leukemia 1.F. A. P. Miller Antagonists of Purine and Pyrimidine Metabolites and of Folic Acid G. M . Timmis Behavior of Liver Enzymes in Hepatocarcinogenesis George Weber AUTHOR INDEX-SUBJECT INDEX

Volume 7 Avian Virus Growths and Their Etiologic Agents 1. W. Beard

CONTENTS OF PREVIOUS VOLUMES Mechanisms of Resistance to Anticancer Agents R. W. Brockman Cross Resistance and Collateral Sensitivity Studies in Cancer and Chemotherapy Dorris J . Hutchison Cytogenic Studies in Chronic Myeloid Leukemia W. M . Court Brown and lshbel M . Tough Ethionine Carcinogenesis Emmanuel Farber Atmospheric Factors in Pathogenesis of Lung Cancer Paul Kotin and Hans L. Falk Progress with Some Tumor Viruses of Chickens and Mammals: The Problem of Passenger Viruses G. Negroni AUTHOR INDEX-SUBJECT INDEX

Volume 8 The Structure of Tumor Viruses and Its Bearing on Their Relation to Viruses in General A. F. Howatson Nuclear Proteins o f Neoplastic Cells Harris Busch and William J. Steele Nucleolar Chromosomes: Structures, Interactions, and Perspectives M . J . Kopac and Gladys M . Mateyko Carcinogenesis Related to Foods Contaminated by Processing and Fungal Metabolites H . F. Kraybill and M . B. Shimkin Experimental Tobacco Carcinogenesis Ernest L. Wynder and Dietrich Hoffman AUTHOR INDEX-SUBJECT INDEX

Volume 9 Urinary Enzymes and Their Diagnostic Value in Human Cancer Richard Stambuugh and Sidney Weinhou se

385

The Relation of the Immune Reaction to Cancer Louis V. Caso Amino Acid Transport in Tumor Cells R. M . Johnstone and P. G. Scholefield Studies on the Development, Biochemistry, and Biology of Experimental Hepatomas Harold P. Morris Biochemistry of Normal and Leukemic Leucocytes, Thrombocytes, and Bone Marrow Cells 1. F. Seitz AUTHOR INDEX-SUBJECT INDEX

Volume 10 Carcinogens, Enzyme Induction, and Gene Action H . V. Gelboin In Vitro Studies on Protein Synthesis by Malignant Cells A. Clark G r i . n The Enzymatic Pattern of Neoplastic Tissue W. Eugene Knor Carcinogenic Nitroso Compounds P. N . Magee and J . M . Barnes The Sulfhydryl Group and Carcinogenesis 1.S . Harrington The Treatment o f Plasma Cell Myeloma Daniel E. Bergsagel, K . M . Grifith, A. Haut, and W. J . Stuckley, Jr. AUTHOR INDEX-SUBJECT INDEX

Volume 11 The Carcinogenic Action and Metabolism of Urethan and N-Hydroxyurethan Sidney S . Miruish Runting Syndromes, Autoimmunity, and Neoplasia D. Keast Viral-Induced Enzymes and the Problem of Viral Oncogenesis Saul Kit

386


The

Growth-Regulating Activity of Polyanions: A Theoretical Discussion of Their Place in the Intercellular Environment and Their Role in Cell Physiology William Regelson Molecular Geometry and Carcinogenic Activity of Aromatic Compounds. New Perspectives Joseph C . Arcos and Mary F. Argus

AUTHOR INDEX-SUBJECT INDEX CUMULATIVE INDEX

Volume 12 Antigens Induced by the Mouse Leukemia Viruses G. Pasternak Immunological Aspects of Carcinogenesis by Deoxyribonucleic Acid Tumor Viruses C. 1. Deichman Replication of Oncogenic Viruses in Virus-Induced Tumor Cells-Their Persistence and Interaction with Other Viruses H . Hanafusa Cellular Immunity against Tumor Antigens Karl Erik Hellstrom and lngegerd Hellstrom Perspectives in the Epidemiology of Leukemia Irving L. Kessler and Abraham M . Lilienfeld AUTHOR INDEX-SUBJECT INDEX

Volume 13 The Role of Immunoblasts in Host Resistance and Immunotherapy of Primary Sarcomata P. Alexander and J. G . Hall Evidence for the Viral Etiology of Leukemia in the Domestic Mammals Oswald Jarrett

The Function of the Delayed Sensitivity Reaction as Revealed in the Graft Reaction Culture Haim Ginsburg Epigenetic Processes and Their Relevance to the Study of Neoplasia Gajanan V. Sherbet The Characteristics of Animal Cells Transformed in Vitro lan Macpherson Role of Cell Association in Virus Infection and Virus Rescue J . Svoboda and 1. Hloidnek Cancer of the Urinary Tract D . B . Clayson and E . H . Cooper Aspects of the EB Virus M . A . Epstein AUTHOR INDEX-SUBJECT INDEX

Volume 14 Active Immunotherapy Georges Math6 The Investigation of Oncogenic Viral Genomes in Transformed Cells by Nucleic Acid Hybridization Ernest Winocour Viral Genome and Oncogenic Transfonnation: Nuclear and Plasma Membrane Events George Meyer Passive Immunotherapy of Leukemia and Other Cancer Roland Motta Humoral Regulators in the Development and Progression of Leukemia Donald Metcalf Complement and Tumor Immunology Kusuya Nishioka Alpha-Fetoprotein in Ontogenesis and Its Association with Malignant Tumors G . 1. Abeler Low Dose Radiation Cancers in Man Alice Stewart AUTHOR INDEX-SUBJECT INDEX


Volume 15 Oncogenicity and Cell Transformation by Papovavirus SV40: The Role of the Viral Genome J . S . Butel, S . S . Tevethia, and J . L. Melnick Nasopharyngeal Carcinoma (NPC) 1.H . C. H o Transcriptional Regulation in Eukaryotic Cells A. 1. MacGilZiuray, 1. Paul, and C. Threlfall Atypical Transfer RNA's and Their Origin in Neoplastic Cells Ernest Borek and Sylvia]. Kerr Use of Genetic Markers to Study Cellular Origin and Development of Tumors in Human Females Philip J . Fialkow Electron Spin Resonance Studies of Carcinogenesis Harold M . Swartz Some Biochemical Aspects of the Relationship between the Tumor and the Host V. S . Shapot Nuclear Proteins and the Cell Cycle Gary Stein and Renato Baserga AUTHOR INDEX-SUBJECT INDEX

387

l,%Bis(Bchloroethy1)- 1-nitrosourea (BCNU) and Other Nitrosoureas in Cancer Treatment: A Review Stephen K . Carter, Frank M . Schabel, Jr., Lawrence E. Broder, and Thomas P. Johnston AUTHOR INDEX-SUBJECT INDEX

Volume 17 Polysaccharides in Cancer: Glycoproteins and Glycolipids Vijai N . Nigam and Antonio Cantero Some Aspects of the Epidemiology and Etiology of Esophageal Cancer with Particular Emphasis on the Transkei, South Africa Gerald P. Warwick and John S . Harington Genetic Control of Murine Viral Leukemogenesis Frank Lilly and Theodore Pincus Marek's Disease: A Neoplastic Disease of Chickens Caused by a Herpesvirus K . Nazerian Mutation and Human Cancer Alfred G. Knudson, Jr. Mammary Neoplasia in Mice S . Nandi and Charles M . McCrath AUTHOR INDEX-SUBJECT INDEX

Volume 16 Polysaccharides in Cancer Vijai N. Nigam and Antonio Cantero Antitumor Effects of Interferon Ion Gresser Transformation by Polyoma Virus and Simian Virus 40 Joe Sambrook Molecular Repair, Wound Healing, and Carcinogenesis: Tumor Production a Possible Overhealing? Sir Alexander Haddow The Expression of Normal Histocompatibility Antigens in Tumor Cells Alena Lengerovd

Volume 18 Immunological Aspects of Chemical Carcinogenesis R. W. Baldwin Isozymes and Cancer Fanny Schapira Physiological and Biochemical Reviews of Sex Differences and Carcinogenesis with Particular Reference to the Liver Yee Chu Toh Immunodeficiency and Cancer John H . Kersey, Beatrice D. Spector, and Robert A. Good

388


Recent Observations Related to the Chemotherapy and Immunology of Gestational Choriocarcinoma K . D. Bagshave Glycolipids of Tumor Cell Membrane Sett-itiroh Hakomori Chemical Oncogenesis in Culture Charles Heidelberger AUTHOR INDEX-SUBJECT INDEX

Volume 19 Comparative Aspects of Mammary Tumors J . M. Hamilton The Cellular and Molecular Biology of RNA Tumor Viruses, Especially Avian Leukosis-Sarcoma Viruses, and Their Relatives Howard M. Temin Cancer, Differentiation, and Embryonic Antigens: Some Central Problems J . H. Coggin, Jr. and N . G. Anderson Simian Herpesviruses and Neoplasia Fredrich W. Deinhardt, Lawrence A . Falk, and Lauren G. W d f e Cell-Mediated Immunity to Tumor Cells Ronald B. Herberman Herpesviruses and Cancer Fred Rapp Cyclic AMP and the Transformation of Fibroblasts Ira Pastan and George S . Johnson Tumor Angiogenesis Judah Folkman SUBJECT INDEX

Volume 20 Tumor Cell Surfaces: General Alterations Detected by Agglutinins Annette M . C . Rapin and Max M. Burger

Principles of Immunological Tolerance and Immunocyte Receptor Blockade G. J . V. Nossal The Role of Macrophages in Defense against Neoplstic Disease Michael H . Levy and E . Frederick Wheelock Epoxides in Polycyclic Aromatic Hydrocarbon Metabolism and Carcinogenesis P. Sims and P. L. Grover Virion and Tumor Cell Antigens of C-Type RNA Tumor Viruses Heinz Bauer Addendum to “Molecular Repair, Wound Healing, and Carcinogenesis: Tumor Production a Possible Overhealing?” Sir Alexander Haddow SUBJECT INDEX

Volume 21 Lung Tumors in Mice: Application to Carcinogenesis Bioassay Michael B. Shimkin and Gary D. Stoner Cell Death in Normal and Malignant Tissues E. H. Cooper, A. J . Bedford, and T. E. Kenny The Histocompatibility-Linked Immune Response Genes Bamj Benacerrafand Daoid H . Katz Horizontally and Vertically Transmitted Oncomaviruses of Cats M . Essex Epithelial Cells: Growth in Culture of Normal and Neoplastic Forms Keen A. Rafferty, Jr. Selection of Biochemically Variant, in Some Cases Mutant, Mammalian Cells in Culture G. B. Clements The Role of DNA Repair and Somatic Mutation in Carcinogenesis James E. Trosko and Ernest H. Y. Chu SUBJECT INDEX


Volume 22

Volume 24

Renal Carcinogenesis J . M. Hamilton Toxicity of Antineoplastic Agents in Man: Chromosomal Aberrations, Antifertility Effects, Congenital Malformations, and Carcinogenic Potential Susan M . Sieber and Richard H . Adamson Interrelationships among RNA Tumor Viruses and Host Cells Raymond V. Gilden Proteolytic Enzymes, Cell Surface Changes, and Viral Transformation Richard Roblin, lih-Nan Chou, and Paul H . Black Immunodepression and Malignancy Osias Stutman

The

SUBJECT lNDEX

Volume 23 The Genetic Aspects of Human Cancer W. E . Heston The Structure and Function of Intercellular Junctions in Cancer Ronald S. Weinstein,Frederick B. Merk, and Joseph Alroy Genetics of Adenoviruses Harold S. Ginsberg and C . S . H . Young Molecular Biology of the Carcinogen, 4-Nitroquinoline 1-Oxide Minako Nagao and Takashi Sugimura Epstein-Barr Virus and Nonhuman Primates: Natural and Experimental Infection A . Frank, W. A. Andiman,and G. Miller Tumor Progression and Homeostasis Richmond T. Prehn Genetic Transformation of Animal Cells with Viral DNA or RNA Tumor Viruses Mirosluv Hill and Juna Hillooa SUBJECT INDEX

389

Murine Sarcoma Virus-Induced Tumor: Exception or General Model in Tumor Immunology? J . P. Levy and J. C. Leclerc Organization of the Genomes of Polyoma Virus and SV40 Mike Fried and Beverly E. Grifin &-Microglobulin and the Major Histocompatibility Complex Per A. Peterson, Lars Rask, and Lars Ostberg Chromosomal Abnormalities and Their Specificity in Human Neoplasms: An Assessment of Recent Observations by Banding Techniques Joachim Mark Temperature-Sensitive Mutations in Animal Cells Claudio Basilico Current Concepts of the Biology of Human Cutaneous Malignant Melanoma Wallace H . Clark, Jr.. Michael J . Mastrangelo, Ann M . Ainsworth, David Berd, Robert E. Bellet, and Evefinu A. Bernardino SUBJECT INDEX

Volume 25 Biological Activity of Tumor Virus DNA F. L. Graham Malignancy and Transformation: Expression in Somatic Cell Hybrids and Variants Harvey L. Ozer and Krishna K . ]ha Tumor-Bound Immunoglobulins: In Situ Expressions of Humoral Immunity lsaac P. Witz The Ah Locus and the Metabolism of Chemical Carcinogens and Other Foreign Compounds Snorri S . Thorgeirsson and Daniel W. Nebert Formation and Metabolism of Alkylated Nucleosides: Possible Role in Car-

390


cinogenesis by Nitroso Compounds and Alkylating Agents Anthony E . Pegg Immunosuppression and the Role of Suppressive Factors in Cancer lsao Kamo and Herman Friedman Passive Immunotherapy of Cancer in Animals and Man Steuen A . Rosenberg and William D .

Terry SUBJECT INDEX

Volume 26 The

Epidemiology Cancer

of

Pelayo Correa and William Haensrel Interaction between Viral and Genetic Factors in Murine Mammary Cancer 1.Hilgers and P. Bentuelzen Inhibitors of Chemical Carcinogenesis Lee W. Wattenberg Latent Characteristics of Selected Herpesviruses lack (3. Stevens Antitumor Activity of Corynebacteriun parvum Luka Milas and Martin T. Scott

Large-Bowel SUBJECT INDEX

A

B C 8

D 9

E D F 1 6 2 H 3

1 4 J

S

Advances in Cancer Research, Volume 27

Advances in Ecological Research, Volume 27

Advances in Food Research, Volume 27

Advances in Food Research, Volume 27