No title

HIGHLIGHTS HIGHLIGHTS ADVISORS CEZMI AKDIS SWISS INSTITUTE OF ALLERGY AND ASTHMA RESEARCH, SWITZERLAND MARCO BAGGIOLINI ...

Author: Nature Reviews Immunology

7 downloads 747 Views 15MB 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

HIGHLIGHTS HIGHLIGHTS ADVISORS CEZMI AKDIS SWISS INSTITUTE OF ALLERGY AND ASTHMA RESEARCH, SWITZERLAND MARCO BAGGIOLINI UNIVERSITA DELLA SVIZZERA ITALIANA, SWITZERLAND BRUCE BEUTLER SCRIPPS RESEARCH INSTITUTE, USA ANDREW CHAN GENENTECH, INC., USA ANNE COOKE UNIVERSITY OF CAMBRIDGE, UK JAMES DI SANTO PASTEUR INSTITUTE, FRANCE TASUKU HONJO KYOTO UNIVERSITY, JAPAN GARY KORETZKY UNIVERSITY OF PENNSYLVANIA, USA CHARLES MACKAY GARVAN INSTITUTE OF MEDICAL RESEARCH, AUSTRALIA FIONA POWRIE UNIVERSITY OF OXFORD, UK CAETANO REIS E SOUSA IMPERIAL CANCER RESEARCH FUND, UK ALAN SHER NATIONAL INSTITUTE OF ALLERGY AND INFECTIOUS DISEASES, USA ANDREAS STRASSER THE WALTER AND ELIZA HALL INSTITUTE, MELBOURNE, AUSTRALIA ERIC VIVIER CENTRE D’IMMUNOLOGIE DE MARSEILLE-LUMINY, FRANCE

T U M O U R I M M U N O LO G Y

Dealing with tumours Tumours can prevent the immune system mounting an effective immune response in several ways. Production of the immunosuppressive cytokine transforming growth factor-β (TGF-β) by tumour cells, or by other cells at the tumour site, can contribute to the suppression of anti-tumour immune responses. Reporting in Nature Medicine, Leonid Gorelik and Richard Flavell provide evidence to show that blockade of TGF-β signalling in T cells enables the immune system to protect mice against tumours. TGF-β is a cytokine that mediates immunosuppression in several ways, including inhibition of T-cell activation by antigen-presenting cells, and inhibition of T-cell differentiation into cytotoxic T lymphocytes (CTLs) and T helper 1 cells (TH1). In this study, Gorelik and Flavell investigated the contribution of TGF-β signalling in anti-tumour responses by T cells. They analysed mice expressing a dominant-negative form of TGF-β receptor type II, whose CD4+ and CD8+ T cells are insensitive to TGF-β signalling. The effect on tumour growth was assessed using the mouse thymoma line EL-4 and the mouse metastatic melanoma line B16-F10, both of which produce TGF-β. Transgenic mice were able to resist challenge by both types of tumour cell, but non-transgenic littermates developed progressive tumours. Although rejection was associated with the development of tumour-specific CTLs, cell-depletion

experiments showed that CD4+ T cells were also required for effective tumour eradication. But does tumour eradication depend on blockade of TGF-β signalling in both CD4+ and CD8+ T cells? To investigate this, the authors used an adoptive-transfer system with Rag1–/– mice — which lack endogenous T and B cells — as hosts. Transfer of transgenic CD8+ T cells and non-transgenic CD4+ T cells prevented the development of tumours, but the opposite combination of cells was unable to prevent tumour growth, showing that blockade of TGF-β signalling in CD8+T cells is crucial for tumour eradication. So, TGF-β-signalling blockade can protect mice against tumour challenge, but can it also combat established tumours? Rag1–/– mice

NATURE REVIEWS | IMMUNOLOGY

were injected with EL-4 tumour cells followed by adoptive transfer of transgenic T cells. T-cell transfer on day 3 after tumour innoculation did prevent tumour growth, but cell transfer on day 7 had no effect. The results from this study have important implications for cancer therapy. The data support the idea that although tumour cells use several tricks to confound the immune system, one way to enhance antitumour responses might be to block TGF-β signalling in T cells. Elaine Bell References and links ORIGINAL RESEARCH PAPER Gorelik, L. &

Flavell, R. Immune-mediated eradication of tumors through the blockade of transforming growth factor-β signaling in T cells. Nature Med. 7, 1118–1122 (2001) WEB SITE Richard Flavell’s lab: http://www.biology. yale.edu/FacultyResearch/Flavell.html

VOLUME 1 | NOVEMBER 2001 | 8 7

© 2001 Macmillan Magazines Ltd

HIGHLIGHTS

VIRAL IMMUNITY

Squatters’ rights! As any housing officer will tell you, squatters (people who take unauthorized possession of unoccupied premises) use many tricks to avoid detection and eviction. Viruses use similarly cunning mechanisms to ensure that the cells they reside in are not targeted for destruction, via apoptosis, by the immune system. Tumournecrosis factor (TNF)-related apoptosisinducing ligand (TRAIL), a member of the TNF superfamily, is emerging as an important molecule used by cells of the immune system to kill virus-infected cells, but how, or if, viruses can inhibit TRAIL-induced apoptosis is unknown. Tollefson and colleagues now report, in the Journal of Virology, that adenoviruses have evolved proteins that inhibit TRAILinduced apoptosis, so enabling persistance of the viral infection. Previous studies have established that proteins encoded by the E1B and E3 transcription units of adenovirus, including E1B-19K, E314.7K and the E3 protein RID (receptor internal-

MUCOSAL IMMU N ITY

Defensive position The body’s mucosal surfaces are defended by a coating of immunoglobulin A. In the gut, IgA is secreted by antibody-forming cells (AFCs) that are positioned in a crucial immune-effector site — the villus lamina propria (LP). These IgA+ AFCs were thought to originate in germinal centres within descrete inductive immune sites known as Peyer’s patches (PP). But a recent report in Nature from Tasuku Honjo’s group shows that this is not the whole story. An initial clue that the origins of IgA+ AFCs might be different came from studies of mice deficient in activation-induced cytidine deaminase (AID). B cells from AID –/– mice cannot switch their immunoglobulin class, and there is a striking accumulation of IgM+ B cells and AFCs in the gut LP. This led the authors to propose that the LP IgM+ B cells might be the precursors of both IgM+ LP AFCs in AID –/– mice and IgA+ AFCs in wild-type mice.

88

Uninfected

Wild-type adenovirus

An indirect immunofluorescence image using a mouse monoclonal antibody specific for TRAIL receptor 1. Courtesy of Ann Tollefson, St Louis University, USA.

ization and degradation), protect infected cells from apoptosis induced by TNF-α and Fas ligand (FasL). In the present study, the authors carried out apoptosis assays on A549 human lung carcinoma cells infected with wild-type adenovirus or with mutants that lack one or more E3 or E1B protein in the presence of TRAIL and cycloheximide (which increases the sensitivity of cells to TRAIL-induced apoptosis). Mock-infected cells underwent apoptosis, as did cells infected with mutants that lacked the expression of RID, E3-14.7K and E1B-19K. However, cells infected with wild-type adenovirus or mutants expressing RID, but not E1B-19K or E3-14.7K, remained viable. Therefore, the adenoviral protein RID can block TRAIL-induced apoptosis. How does RID inhibit the TRAIL pathway? The authors have previously shown that RID

If LP IgA+ AFCs are generated in situ, then actively switching B cells should be present in the LP. But how can this transient event be detected? The authors used three molecular indicators: AID, which is expressed only in B cells undergoing class switching; germ-line transcripts of the α-chain gene, which are produced just prior to switching; and circular transcripts, which are short-lived by-products of class-switch recombination. These indicators show that both LP and PP IgA+ B cells have recently class-switched. In vitro and in vivo experiments showed that LP IgM+ B cells have a greater tendancy than PP IgM+ B cells to differentiate into IgA+ AFCs. But what is it about the micorenvironment of the LP that supports switching from IgM to IgA, an event previously thought to be restricted to germinal centres? The authors show that stromal cells isolated from the LP enhance the switching of splenic B cells to IgA, and suggest that factors secreted by the stromal cells, particularly transforming growth factor-β, might promote differentiation to IgA+ AFCs. This study indicates that, in addition to being a crucial effector site, the gut LP is an important inductive site of the gut mucosal immune system.

proteins protect against Fas-induced apoptosis by causing internalization and degradation of cell-surface Fas. Here, they show that the same applies for TRAIL; TRAIL-receptor 1 is cleared from the surface of cells infected by wild-type adenovirus, or any mutant expressing RID, and is transported to lysosomes for degradation. This study therefore provides an insight into how adenoviruses inactivate TRAIL-induced apoptotic pathways and so avoid eviction. Jenny Buckland References and links ORIGINAL RESEARCH PAPER Tollefson, A. E. et al. Inhibition

of TRAIL-induced apoptosis and forced internalization of TRAIL receptor 1 by adenovirus proteins. J. Virol. 75, 8875–8887 (2001) FURTHER READING Tollefson, A. E. et al. Forced degradation of Fas inhibits apoptosis in adenovirus-infected cells. Nature 392, 726–730 (1998)

References and links ORIGINAL RESEARCH PAPER Fagarasan, S., Kinoshita, K.,

Muramatsu, M., Ikuta, K. & Honjo, T. In situ class switching and differentiation to IgA-producing cells in the gut lamina propria. Nature 413, 639–644 (2001) FURTHER READING Nagler-Anderson, C.A. Man the barrier! Strategic defences in the intestinal mucosa. Nature Rev. Immunol. 1, 59–67 (2001) WEB SITE Taksuku Honjo’s lab: http://www2.mfour.med.kyoto-u.ac.jp

Jen Bell

| NOVEMBER 2001 | VOLUME 1

www.nature.com/reviews/immunol


HIGHLIGHTS

IN BRIEF

GENE THERAPY

It’s all in the timing Everyone knows that time is of the essence when you have a train to catch. Recent work by Cohen and colleagues in the journal Blood shows that timing is also essential when it comes to the application of suicide gene therapy of graft-versus-host disease (GVHD), a life-threatening complication of allogeneic haematopoietic stem-cell transplantation (HSCT). Mature donor T cells present in an allograft after HSCT improve engraftment and T-cell reconstitution, and provide a graft-versus-leukaemia (GVL) effect. In addition to these beneficial effects, activation of donor T cells that are specific for recipient alloantigens (termed alloreactive) result in GVHD. Is it possible to selectively eliminate the alloreactive donor T cells, but spare the T cells which mediate immune reconstitution and the GVL effect? A strategy has been developed to eliminate donor T cells. Before transplantation, donor T cells are transduced with the herpes simplex type I thymidine kinase (TK) suicide gene. Treatment with gancyclovir (GCV; a thymidine analogue that is toxic to dividing cells expressing TK) allows these donor T cells to be eliminated. However, as the TK–GCV system is based on the cell-cycle status of donor TK-expressing T cells, and not on their alloreactivity, its therapeutic usefulness was thought to be limited. Cohen et al. assessed the kinetics of T-cell expansion after semiallogeneic bone marrow transplantation (BMT) (when both alloreactive and homeostatic T-cell expansion occur) and syngeneic BMT (when only homeostatic expansion occurs). T cells from double-transgenic mice expressing TK in CD4 + and CD8+ T cells and a marker (human CD4) were stained with carboxyfluorescein diacetate succinimide ester (CFSE) and infused with wild-type BM into lethally irradiated recipients. Spleen cells were then collected

A U TO I M M U N I T Y

Granulocyte macrophage colony-stimulating factor: a new putative therapeutic target in multiple sclerosis. McQualter, J.L. et al. J. Exp. Med. 194, 873–882 (2001)

from the recipient mice at different time points after BMT, and donor T-cell division was assessed on the basis of CFSE fluorescence. In semiallogenic hosts, the donor T cells proliferated rapidly, and after 88 hours most donor T cells had divided at least once. In syngeneic hosts, T-cell divisions were significantly delayed. These results indicate that alloreactive T cells divide earlier than non-alloreactive T cells. The authors then investigated the persistence and expansion of donor T cells in the semiallogeneic BMT setting after GCV treatment. This treatment resulted in the death of most donor T cells, with surviving T cells being those that had not divided. After a 7-day GCV course, a pool of donor T cells persist, which significantly expands after GCV treatment is stopped. Finally, the authors show that the surviving T cells contribute to the replenishment of the T-cell compartment and provide a diversified T-cell receptor repertoire. This study has identified a therapeutic window when GCV treatment can be administered to specifically kill alloreactive donor T cells (and so control GVHD), but spare nondividing, non-alloreative T cells, which enables T-cell reconstitution to be maintained. Jenny Buckland References and links ORIGINAL RESEARCH PAPER Cohen, J.L.,

Boyer, O. & Klatzmann, D. Suicide gene therapy of graft-versus-host disease: immune reconstitution with transplanted mature T cells. Blood 98, 2071–2076 (2001) FURTHER READING Cohen, J. L., Boyer, O. & Klatzmann, D. Would suicide gene therapy solve the ‘T-cell dilemma’ of allogeneic bone marrow transplantation? Immunol. Today 20, 172–176 (1999)

The chronic autoimmune demyelinating disease mutliple sclerosis (MS) is characterized by an inflammatory infiltration of immune cells into the central nervous system (CNS). Immunization of mice with myelin proteins results in a similar disease, experimental autoimmune encephalomyelytis (EAE), and provides a useful animal model. The cytokine granulocyte macrophage colony-stimulating factor is implicated in chronic inflammation, leading the authors to investigate whether this might be a therapeutic target in MS. GM-CSF –/– mice were found to be resistant to the induction of EAE and, importantly, immune cells failed to infiltrate the CNS. T- C E L L S I G N A L L I N G

Single-cell analysis of signal transduction in CD4 T cells stimulated by antigen in vivo. Zell, T. et al. Proc. Natl Acad. Sci. USA 98, 10805–10810 (2001)

Lymphocyte signal transduction is commonly analysed in vitro, in bulk populations of transformed cells. However, results are often contradictory. Here, the early signalling events following activation of naive CD4+ T cells was studied in a more physiological ex vivo system. Traceable numbers of naive CD4+ T-cell receptor transgenic T cells were transferred into normal mice, which were then challenged with specific antigen. At various time points, lymphoid tissues were removed, fixed immediately and then phosphorylation of key signalling molecules was measured by flow cytometry. In contrast with previous in vitro studies, this analysis indicates that phosphorylation of c-jun and p38 mitogen-activated protein kinase does not depend on co-stimulatory signals from CD28. ALLE RGY

Histamine regulates T-cell and antibody responses by differential expression of H1 and H2 receptors. Jutel, M. et al. Nature 413, 420–425 (2001)

Specialized subsets of T helper cells (TH1 and TH2) are present at sites of inflammation where effector cells, including mast cells and basophils, produce mediators such as histamine. In this study, Cezmi Akdis’ group show that histamine can regulate inflammatory reactions by enhancing TH1-type responses through the histamine receptor type 1 (H1R). Conversely, TH1and TH2-type responses are both negatively regulated through the H2R. Secretion of interferon-γ was suppressed in H1R knockout mice, and TH2 cytokines (IL-4 and IL-13) were predominantly expressed. TH1 and TH2 cytokine expression was upregulated in H2R knockout mice. Interestingly, mice lacking the H1R display increased antibody responses, expressing higher levels of immunoglobulin E (IgE), IgG1, IgG2b and IgG3 in comparison with the H2R-deficient mice. These results show a new immunoregulatory role for the effector mediator histamine.




HIGHLIGHTS

H A E M ATO P O I E S I S

Stem cell origins

Haematopoiesis is the process whereby mature blood cells of distinct lineages are produced from mulitpotent haematopoietic stem cells (HSCs). Despite decades of study, the origin of these HSCs during mammalian embryogenesis has remained unknown. Reporting in Immunity, Cumano and colleagues show that, in mice, HSCs are generated in an intraembryonic region, termed the splanchnopleura (Sp), and not from the extraembryonic yolk sac (YS) blood islands. During vertebrate embryogenesis, haematopoietic cells first appear within the extraembryonic YS. As the bone marrow and fetal liver (the main haematopoietic organs in mammals) require an input of exogenous haematopoietic precursors to generate differentiated progeny, it was suggested that HSCs originate in the YS and later migrate to the fetal liver. However, the intraembryonic Sp region has also been shown to have haematopoietic activity. The differentiation potential of haematopoietic precursors from the YS and Sp, separated from mouse

N AT U R A L K I L L E R C E L L S

Natural killer selection Natural killer (NK) cells are often called ‘innate lymphocytes’ as they combine innate recognition with the effector mechanisms of T lymphocytes — cytokine secretion (interferon-γ; IFN-γ) and cytotoxic killing. However, in terms of responses to infection, NK cells might be more similar to lymphocytes than previously thought. A report from Dokun et al. in Nature Immunology indicates that, similar to T and B cells, antigen-specific NK cells selectively expand in response to infection. NK cells identify their targets by integrating signals from their inhibitory and activation receptors. But not all NK cells are created equal. Although their antigen receptors are invariant, NK cell subsets carry distinct complements of receptors. Previously, this group showed that, in mice, an activation receptor of the Ly49 family, Ly49H, confers specific protection against

90

murine cytomegalovirus (MCMV). This implies that NK cells are capable of virusspecific recognition. So, are Ly49H+ NK cells preferentially activated by MCMV? The present study assessed this in two ways — by scoring for IFN-γ production through intracellular cytokine staining and by measuring NK cell proliferation using a bromodeoxyuridine (BrdU)-incorporation assay. In the early phase of the response to MCMV there is a burst of IFN-γ production by NK cells, which peaks at 36 hours. However, at 2 days post-infection, there are comparable percentages of Ly49H+ and Ly49H– NK cells producing IFN-γ, and there was no difference in the proliferation of the two subsets. By contrast, the authors found that by day 6 post-infection, there has been an outgrowth of Ly49H+ NK cells, and these cells proliferate to a much greater degree than their Ly49H– counterparts.


embryos before circulation, has been investigated previously by the authors. These experiments established that lymphoid potential is restricted to precursors of intraembryonic origin. In transplantation experiments using normal, irradiated mice as recipients, neither precursors from the YS or from Sp were able to reconstitute an adult haematopoietic compartment. As this might have been due to technical limitations of the system used, the authors developed a new organ culture and cell-transfer system to investigate this further. The protocol involved culturing the Sp and YS, again separated before the onset of blood circulation, for 4 days (to ensure a sufficient number of HSCs were available), before transplanting the precursors into recombinase activating gene (Rag)–/– or Rag –/– crossed with common γchain (Rag γc –/–) mice. Rag –/– mice lack T and B cells (so reducing possible competition between donor and recipient precursors) and Ragγc–/– mice additionally lack natural killer cells (which might kill haematopoietic precursors).

But is this preferential proliferation actually driven by Ly49H recognition of MCMV? Vaccinia virus infection, which induces NK cell activation, did not cause the selective proliferation of Ly49H+ cells, indicating this might indeed be an MCMVspecific response. In addition, administration of Ly49H antibodies was shown to inhibit bulk NK cell expansion in response to MCMV, which shows that Ly49H has a direct role in MCMV-triggered NK cell activation. This study provides a new model of NK cell activity in viral infection. Yet it remains to be seen whether this delayed, preferential proliferation of NK cells is itself important for antiviral immunity. Jen Bell References and links ORIGINAL RESEARCH PAPER Dokun, A. O. et al. Specific and nonspecific NK cell activation during virus infection. Nature Immunol. 2, 951–956 (2001) FURTHER READING Cerwenka, A. & Lanier, L. L. Natural killer cells, viruses and cancer. Nature Rev. Immunol. 1, 41–49 (2001) WEB SITE Wayne Yokoyama’s lab: http://www.hhmi.org/ research/investigators/yokoyama.htm



Ragγc–/– mice injected with YS cells generated donor-derived myeloid cells, but this reconstitution was transient, and after 3 months donorderived progeny were no longer observed. No lymphocytes were ever detected in these mice. Therefore, YS cells cannot generate lymphocytes, but can provide short-term myeloid reconstitution. By contrast, after injection of Sp-derived cells into Ragγc–/– recipients, donor-derived myeloid cells, as well as B and T cells were generated. These myeloid and lymphoid cells were still present 8 months after injection, indicating that Sp-derived percursors can provide long-term reconstitution of recipient mice. The authors conclude that the only cells capable of adult long-term haematopoiesis are from the Sp region. So, during mouse embryogenesis, HSCs are generated intraembryonically and do not derive from the YS. Jenny Buckland References and links ORIGINAL RESEARCH PAPER Cumano, A. et al.

Intraembryonic, but not yolk sac hematopoietic precursors, isolated before circulation, provide long-term multilineage reconstitution. Immunity 15, 477–485 (2001)

N AT U R A L K I L L E R C E L L S

Targeting tumour cells The ‘missing-self ’ hypothesis, formulated by Klaus Kärre and colleagues in 1986, proposed that natural killer (NK) cells seek out and destroy cells that have lost expression of major histocompatibility complex (MHC) class I antigens. Two recent papers, published in Nature and in Proceedings of the National Academy of Sciences, now show that NK cells can reject tumour cells that express ligands for the activating NK receptor NKG2D, despite the expression of MHC class I molecules by the tumour cells. The formulation of the missing-self hypothesis predicted the existence of receptors on NK cells that inhibit their activity and which recognize MHC class I molecules. Many of these receptors have now been identified. Recent work has also identified several activating NK receptors, including the lectin-like molecule NKG2D, whose engagement provides dominant activating signals to the NK cell. Previous work by Tom Spies’ group showed that NK cells can kill NKG2D ligandexpressing cells in vitro. The mouse ligands for NKG2D are retinoic acid early inducible-1 (Rae-1) and H60, which are expressed by some tumour cells, but not by normal adult cells. Both groups looked for direct evidence to support the idea that tumour cells ectopically expressing ligands for NKG2D could stimulate antitumour responses by NK cells. The Raulet group used a retroviral expression system to express Rae1β and H60 in three mouse tumour cell lines that express MHC class I molecules — EL4 thymoma cells, RMA T-cell lymphoma cells (which were used in the original Kärre study) and B16-BL6 melanoma cells. Transduced EL4 and B16-BL6 cells that were injected subcutaneously into recipient mice were rapidly and completely

rejected. Tumour cells were rejected in wild-type mice depleted of CD8+ T cells and in Rag –/– mice (which lack T and B cells), but grew in Rag –/– mice depleted of NK cells, indicating that conventional NK cells are responsible for the rejection. Rae1βor H60-transduced RMA cells were also rejected in wild-type mice, but rejection required both CD8+ T cells and NK cells. The Lanier group also used the RMA cells to investigate NK cell responses. RMA cells stably transfected with Rae1γ or Rae1δ were injected intraperitoneally into recipient mice. Mice injected with mock-transfected cells developed tumours and died, whereas mice challenged with transfected cells rejected tumour cells in an NK-cell dependent manner. So, NK cells can reject tumour cells expressing NKG2D ligands, despite MHC class I expression. But do mice primed with NKG2D ligandexpressing tumour cells develop an adaptive T-cell response against subsequent challenge by non-transduced parental tumour cells? This is where the results from the two groups differ. Raulet and colleagues found that NKG2D ligandnegative tumour cells were rejected by mice previously challenged with transduced tumour cells, and that this was a CD8+ T-cell-dependent process. By contrast, Lanier and colleagues found that mice that had rejected Rae1γ-transfected RMA cells were unable to reject parental tumours on re-challenge. These results show that NK cells can, and do, participate in rejection of MHC class-I bearing tumour cells and, although there are some discrepancies in the results, this approach might be effective for tumour vaccine development. Elaine Bell References and links ORIGINAL RESEARCH PAPERS Diefenbach, A., Jensen, E. R., Jamieson, A. M. & Raulet, D. H. Rae1 and H60 ligands of the NKG2D receptor stimulate tumour immunity. Nature 413, 165–171 (2001) | Cerwenka, A., Baron, J. L. & Lanier L. L. Ectopic expression of retinoic acid early inducible-1 gene (RAE-1) permits natural killer cell-mediated rejection of a MHC class I-bearing tumor in vivo. Proc. Natl Acad. Sci. USA 98, 11521–11526 (2001) WEB SITES Lewis Lanier’s lab: http://cc.ucsf.edu/people/lanier_lewis.html David Raulet’s lab: http://mcb.berkeley.edu/labs/raulet/




HIGHLIGHTS

WEB WATCH • (http://vaccines.org/) Vaccines on the Web

Vaccines and vaccine safety are hot topics in the news these days (see the article on p160 of this issue, by Wilson and Marcuse). But where do you go to find up-to-date vaccine news and web resources? The Vaccine Page is a regularly updated website that is partly supported by the Bill and Melinda Gates Children’s Vaccine Program. By clicking through to the news site, readers can access timely articles from various international news sources. But The Vaccine Page is more than just a news service. The site also provides an annotated database of vaccine resources on the web, assembled and checked by the editors of UniScience News Net. Vaccine resources are listed by country and by category, for ease of navigation. Category sections are handily organized into those directed at parents, researchers and medical practitioners, and there are also sections focusing on journal access and vaccine organizations. The Vaccine Page is a member of the Allied Vaccine Group (AVG), which consists of websites “dedicated to presenting valid scientific information about the sometimes confusing subject of vaccines” and which clearly promotes the benefits of vaccination. AVG members believe that the benefits of immunization far outweigh the risks.The publishers of The Vaccine Page also maintain the website of the AVG, and The Vaccine Page uses the AVG search engine VACSearch PLUS, which returns results from the websites of all members of the AVG. So, if you want to access vaccine news and links to web resources which are pro-vaccination, check out The Vaccine Page today.

Elaine Bell

92

T- C E L L M E M O R Y

Learning to remember Memory is a complex business — no less so in the immune system than in the nervous system. Typically, exposure to antigen during infection induces an immune response in which effector T cells develop that fight the infection. Memory T cells also develop, conferring on the host the ability to mount a rapid and augmented response on subsequent exposure. The development of immunological memory is not yet fully understood, but many immunologists subscribe to the view that memory T cells are quiescent cells that develop from fully differentiated effector T cells. Now, two papers from von Andrian’s group describe the cytokine-stimulated development of memory T cells in the absence of prior effector T-cell differentiation, and the migratory properties of these cells. Reporting in the Journal of Clinical Investigation, Manjunath and colleagues made use of T-GFP mice, in which naive and early-activated T cells uniformly express green fluorescent protein, but terminally differentiated cytotoxic T lymphocytes (CTLs) lose this expression. T-GFP mice were crossed with transgenic P14 mice, which express a T-cell receptor specific for the lymphocytic choriomeningitis virus glycoprotein peptide (gp33–41) epitope. Naive T cells from these doubly transgenic mice were stimulated with gp33–41 peptide for 2 days, then washed and cultured with either interleukin (IL)-2 or IL-15 for a further 5 days, and their phenotype analysed. T cells exposed to high doses of IL-2 developed phenotypic and functional characteristics of effector cells — the cells were large, they had lost GFP expression, they did not express CCR7 (a chemokine receptor whose expression is retained on a subset of memory T cells termed central memory cells), they produced interferon-γ and exhibited peptide-specific

cytotoxicity. By contrast, naive cells exposed to IL-15 acquired various features associated with memory T cells — the cells were small, they retained GFP and CCR7 expression, they lost expression of various activation markers, and were not cytotoxic. The memory-like cells were assessed for their ability to respond to secondary antigen exposure in adoptivetransfer experiments — the exposed T cells survived for several weeks after transfer and were able to respond to subsequent antigen challenge. The second paper, by Weninger and colleagues published in the Journal of Experimental Medicine, examined the migratory properties of these cells. IL-2stimulated effector T cells preferentially accumulated in inflammed tissues and were excluded from most lymphoid organs. Central memory-like, IL-15exposed cells homed to lymphoid organs, such as lymph nodes and Peyer’s patches, were only moderately capable of homing to inflammed tissues, and rapidly responded to recall antigen. So, these results indicate that differentiation of CD8+ T cells into effector cells is not necessary for the development of memory T cells, and support the concept proposed by Lanzavecchia’s group of the existence of distinct subsets of memory cells. Elaine Bell References and links ORIGINAL RESEARCH PAPERS Manjunath, N. et al. Effector

differentiation is not prerequisite for generation of cytotoxic T lymphocytes. J. Clin. Invest. 108, 871–878 (2001) | Weninger, W., Crowley, M. A., Manjunath, N. & von Andrian, U. H. Migratory properties of naive, effector, and memory CD8+ T cells. J. Exp. Med. 194, 953–966 (2001) FURTHER READING Sallusto, F., Mackay, C. R. & Lanzavecchia, A. The role of chemokine receptors in primary, effector, and memory immune responses. Annu. Rev. Immunol. 18, 593–620 (2000) WEB SITE Ulrich von Andrian’s lab: http://cbrweb.med.harvard.edu/~uva/




HIGHLIGHTS

I N N AT E I M M U N I T Y

imd unveiled Although they lack an adaptive immune system, Drosphila are armed with a battery of highly effective antimicrobial peptides. In response to infection, antibacterial and antifungal peptides are induced by independent recognition pathways, initially defined by imd and Toll mutants, respectively. Remarkably, these basic signalling pathways are highly conserved in humans, making Drosophila immunity an extremely useful model of mamalian innate immunity. Although imd mutants were described 6 years ago, the gene had not been characterized. But now, Georgel et al., reporting in Developmental Cell, have identified the elusive imd. Previous studies mapped the imd gene to a defined interval containing many genes. To pinpoint imd, the authors fine-mapped this region using a panel of mutants generated by transposase-induced male recombination — a relatively new technique which introduces traceable deletions. The mutants were then screened for complementation of the

imd mutation. This approach showed that imd is a single gene encoding a 30-kDa protein. Importantly, imd has a carboxy-terminal death domain, a protein–protein interaction module that is involved in apoptotic and immune signalling pathways. Although there is no human homologue of imd, its death domain is remarkably similar to that of mamalian RIP (receptor-interacting protein), an adaptor molecule involved in NF-κB activation and apoptosis. imd was known to act upstream of an NF-κB-like molecule, but it had not been implicated in apoptosis. Surprisingly, overexpression of imd during development is lethal, and several clues indicated that imd might activate apoptosis, including the resistance of imd mutant flies to UV-induced apoptosis. This study, then, fills a key gap in the Imd pathway and indicates for the first time that antibacterial and apoptotic pathways in Drosophila might intersect. Jen Bell References and links ORIGINAL RESEARCH PAPER Georgel, P. et al. Drosophila immune deficiency (IMD) is a death domain protein that activates antibacterial defense and can promote apoptosis. Dev. Cell 1, 503–514 (2001) FURTHER READING Medzhitov, R. Toll-like receptors and innate immunity. Nature Rev. Immunol. 1, 135–145 (2001)

IN BRIEF T- C E L L A C T I V AT I O N

OX40 promotes Bcl-xL and Bcl-2 expression and is essential for long-term survival of CD4 cells. Rogers, P.R., Song, J., Gramaglia, I., Killeen, N. & Croft, M. Immunity 15, 445–455 (2001)

Optimal T-cell activation occurs when a T cell receives a signal from the T-cell receptor and a signal from a co-stimulatory receptor, for example CD28. CD28 signalling enhances T-cell proliferation, cytokine secretion and expression of anti-apoptotic proteins. This paper provides direct evidence that OX40 acts synergistically and at a later stage than CD28, and promotes T-cell survival by increasing expression of the anti-apopototic molecules Bcl-xL and Bcl-2. T- C E L L D E V E LO P M E N T

Epigenetic silencing of CD4 in T cells committed to the cytotoxic lineage. Zou, Y. et al. Nature Genet. 29, 332–336 (2001)

The development of immature double-positive thymocytes into mature single-positive CD4+ and CD8+ T cells requires the termination of expression of either the CD4 or CD8 co-receptor. The first intron of the CD4 gene contains a silencer element that represses CD4 transcription. Zou et al. used the Cre/loxP system to show that the CD4 silencer is only required at distinct stages of development. Once a cell is committed to the CD8+ lineage, the CD4 locus remains silent even if the silencer element is removed. T- C E L L S I G N A L L I N G

Deficiency of small GTPase Rac2 affects T cell activation. Yu, H., Leitenberg, D., Li, B. & Flavell, R.A. J. Exp. Med. 194, 915–925 (2001)

Yu et al. investigated the function of Rac2, a haematopoieticspecific Rho GTPase, in T-cell signalling. Rac2–/– T cells responded poorly to T-cell-receptor stimulation, showing reduced proliferation, Ca2+ mobilization and activation of ERK1/2 and p38. Actin polymerization and cap formation were also decreased in these cells in comparison with wild-type cells. These results show that Rac2 mediates both transcriptional and cytoskeletal changes during T-cell activation. I N N AT E I M M U N I T Y

Subsets of human dendritic cell precursors express different Toll-like receptors and respond to different microbial antigens. Kadowaki, N. et al. J. Exp. Med. 194, 863–869 (2001)

Dendritic cells (DCs) can prime naive T cells and direct the development of immune responses. Yong-Jun Liu’s group investigated the expression of Toll-like receptors (TLRs) — which recognize specific molecular patterns on microbial pathogens — on human DC subsets. The results show that DC subsets express distinct sets of TLRs, supporting the view that DC subsets have evolved to recognize different microbial pathogens.




REVIEWS POSITIVE AND NEGATIVE REGULATION OF T-CELL ACTIVATION BY ADAPTOR PROTEINS Gary A. Koretzky and Peggy S. Myung Adaptor proteins, molecules that mediate intermolecular interactions, are now known to be as crucial for lymphocyte activation as are receptors and effectors. Extensive work from numerous laboratories has identified and characterized many of these adaptors, demonstrating their roles as both positive and negative regulators. Studies into the molecular basis for the actions of these molecules shows that they function in various ways, including: recruitment of positive or negative regulators into signalling networks, modulation of effector function by allosteric regulation of enzymatic activity, and by targeting other proteins for degradation. This review will focus on a number of adaptors that are important for lymphocyte function and emphasize the various ways in which these proteins carry out their essential roles. ADAPTOR MOLECULES

Molecules that lack any known intrinsic enzymatic, DNA binding or receptor functions, but mediate protein–protein or protein–lipid interactions. Most function as flexible molecular scaffolds by regulating the spatio-temporal dynamics of specific effector molecules.

Abramson Family Cancer Research Institute and Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, 415 BRB II/III, 421 Curie Boulevard, Philadelphia, Pennsylvania 19104, USA. Correspondence to G.A.K. e-mail: [email protected]. edu

Cell-surface receptors sample their environment and, when engaged, elicit an array of biochemical signals. The ultimate biological response of the cell requires that these signals be integrated appropriately. ADAPTOR MOLECULES, proteins which contain modular domains that mediate protein–protein or protein–lipid interactions (BOX 1), are vital in this integration process, and allow cells to respond to a broad range of environmental and developmental cues. Intermolecular complexes nucleated by adaptor proteins permit the selective partitioning of specific molecules into discrete subcellular locations. This physical inclusion or exclusion of signalling molecules from particular sites within the cell governs the magnitude, duration and type of downstream signalling pathways engaged. In addition to operating as ‘molecular bridging’ proteins that can nucleate the formation of multimolecular signalling complexes, adaptors can also induce intramolecular conformational changes in their binding partners, thereby regulating their intrinsic enzymatic activities. Furthermore, it is clear that adaptor proteins can function either as positive or negative regulators on the basis of the signalling complexes they create and by promoting the stabilization or destabilization of protein expression. Additional functions for these molecules in

coupling receptor stimulation to polarized actin rearrangements required for cellular activation, adhesion and migration have also been shown. Importantly, the function of adaptor proteins is not only influenced by their primary structure, but also by the location of the adaptors in the cell. In this regard, it has become clear that adaptor proteins move within the cell, taking with them associated molecules, as the cell responds to its environment. The cellular responses to T-cell antigen receptor (TCR) stimulation are diverse, driving thymocyte development and lineage commitment, antigen-specific T-cell activation and apoptosis, as well as T-cell homeostasis. Developing thymocytes express clonotypic TCRs and are subjected to a rigorous selection process designed to ensure self-major histocompatibility complex (MHC) restriction, as well as preventing auto-reactivity to selfantigens. It is thought that this highly sensitive process is contingent on the ability of a thymocyte to discriminate between differences in the strength or half-life of TCR engagement by antigen. How these distinctions are made throughout development relies vitally on the signalling pathways that are initiated following engagement of surface antigen receptors and co-receptors. To this end, adaptor proteins can function as molecular gauges of




REVIEWS

PROTEIN TYROSINE KINASES

(PTKs). Enzymes that catalyse the phosphorylation of proteins on tyrosine residues within the context of specific peptide motifs. PTKs can be generally categorized as either receptor PTKs or as cytosolic PTKs (e.g. ZAP70). IMMUNORECEPTOR TYROSINEBASED ACTIVATION MOTIFS

(ITAMs). Regions found within the CD3 chains of the TCR and other immunoreceptors characterized by tyrosine and leucine or isoleucine residues with discrete spacing. Following receptor engagement, the tyrosines are inducibly phosphorylated and become docking sites for SH2 domain containing proteins including SYK-family protein tyrosine kinases.

receptor stimulation, which measure the quantitative and qualitative nature of extracellular signals by acting as molecular scaffolds between signalling mediators. Dysregulation of these control mechanisms can lead to marked shifts in the TCR repertoire, as well as peripheral T-cell responsiveness, resulting in either states of immune compromise or autoimmunity. Using in vivo and cell-line approaches, several studies have identified and examined the role of adaptor proteins in lymphocytes, which have led to important insights into how early signalling events following antigen-receptor engagement are integrated and relayed to downstream mediators. Several of the adaptor molecules crucial for these processes are expressed exclusively within haematopoietic lineages and, consequently, much information about these molecules derives from studies of their role in haematopoietic-specific functions. Other adaptor molecules are expressed more widely.

This review will provide an overview of just a few of the crucial adaptor proteins that have provided insights into the molecular mechanisms of T-cell development and activation (TABLE 1). Examples have been chosen to include cytosolic proteins, as well as adaptors which are found at the membrane. Proteins with a mainly positive regulatory function are presented as well as those which interfere with T-cell activation. These examples will provide a partial framework to describe examples of how adaptor molecules might modulate a diverse range of effector outcomes following antigen-receptor engagement. Positive regulators

In contrast to receptor PROTEIN TYROSINE KINASES (PTKs), the TCR complex lacks intrinsic tyrosine kinase activity, but rather couples to cytosolic PTKs belonging to the SRC, SYK and TEC families. The most proximal

Box 1 | Adaptor domains Src-homology 2 (SH2): binds specific phosphotyrosine (pY)-containing motifs in the context of three to six amino acids located carboxy-terminal to the pY, providing specificity. An invariant arginine in the SH2 domain is required for pY binding. Src-homology 3 (SH3): binds proline-rich sequences in a left-handed polyproline type II helix. Proline residues of the ligand are usually preceeded by an aliphatic residue. Additional interactions between the ligand and the SH3 domain influence both specificity and affinity. Phosphotyrosine-binding (PTB): binds pY-containing peptide motifs (Asp–Pro–Xaa–pY) in the context of aminoterminal sequences, lending high affinity and specificity to the interaction. Pleckstrin homology (PH): recognize specific phosphoinositides, which allow PH-containing proteins to respond to the generation of lipid second messengers. For some proteins, PH domains mediate their translocation to the plasma membrane. WW: derive their designation from two conserved Trp residues intervened by 20–22 amino acids. They bind prolinerich sequences and might share overlapping binding sites with SH3 domains. C1: Cys-rich sequence that binds diacylglycerol (DAG). On association with DAG, the C1 domain exhibits increased affinity for the lipid membrane, promoting membrane recruitment of C1-containing proteins. PDZ: binds to four to five amino-acid residues of the carboxyl terminus of transmembrane receptors or ion channels. The consensus binding motif consists of a Val or Ile at the C-terminal residue and two to three upstream residues that provide specificity. PDZ domains can also heterodimerize with other PDZ domains on different proteins. ENA/VASP homology 1 (EVH1): module present in ENA/VASP-family members as well as other proteins that regulate the actin cytoskeleton. Binds proteins containing a proline-rich D/EFPPPPXD motif, which are present in some components of the actin cytoskeleton, such as vinculin and zyxin, as well as the Listeria ActA. Thought to be responsible for the translocation of ENA/VASP proteins to focal contacts. Crystal structures indicate that the overall structure of the EVH1 domain is similar to PH domains91,92, implicating the potential for targeting of ENA/VASP proteins to the membrane through interactions with phospholipids. ENA/VASP homology 2 (EVH2): a domain 160–190 amino acids in length, which contains a repetitive mixed charge cluster at the C-terminal end. Present in the C terminus of ENA/VASP-family proteins and is responsible for both Factin-binding and tetramerization activities of ENA/VASP proteins. Both of these properties are required for the actin-crosslinking functions of ENA/VASP proteins. Tyrosine kinase-binding (TKB): a phosphotyrosine-binding domain divergent from typical SH2 and PTB domains. Consists of three structural motifs (a four-helix bundle, an EF hand and a divergent SH2 domain) that together form an integrated phosphoprotein-recognition domain. RING finger: motif containing eight Cys or His residues with conserved spacing. Involved in protein–protein interactions and recently shown to associate with, and to activate, E2 uibiquitin-conjugating enzymes, thereby lending E3 ubiquitin ligase function to RING finger-containing proteins. Proline-rich: amino-acid sequence stretches that are rich in proline residues able to bind various modular units including SH3 and WW domains.

96




REVIEWS

JURKAT T CELL

Human leukaemic T-cell line used to study several aspects of T-cell biology and signalling — in particular, signaltransduction events initiated by the TCR.

biochemical event known to occur following engagement of the TCR is activation of the SRC-family PTKs, LCK and FYN1,2. This results in phosphorylation of tyrosine residues contained within IMMUNORECEPTOR TYROSINE-BASED ACTIVATION MOTIFS (ITAMs) of the TCRassociated CD3 modules3. Phosphorylation of these residues generates binding sites for the tandem SH2 (Src-homology domain 2) domains of the SYK-family PTK, ZAP70 (REFS 4–6). Recruitment of ZAP70 to the TCR permits ZAP70 phosphorylation and activation, also by SRC-family PTKs. In turn, ZAP70 phosphorylates several downstream substrates, initiating a cascade of signalling pathways that results in nuclear transcriptional changes as well as cellular morphological changes. Understanding how early tyrosine kinase events are coupled to downstream mediators has been

an important area of investigation. The cloning and characterization of key adaptor molecules in T cells has identified several vital players underlying the molecular mechanism of T-cell activation following PTK activation. Interactions of SLP76 and LAT. SH2-domain containing leukocyte-specific phosphoprotein (SLP76) is a cytosolic adaptor protein which consists of three protein-binding motifs7. Initial studies of SLP76 in the JURKAT T-CELL leukaemic line showed that overexpression of this adaptor augments TCR-induced interleukin-2 (IL2) promoter activity8. Subsequent to these gain-of-function studies, Jurkat T cells and mice deficient in Slp76 expression (BOX 2) were established9–11. Experiments examining these loss-of-function mutants indicated an essential

Table 1 | Adaptor proteins in lymphocytes Adaptor Mr (kDa)

Expression

Binding proteins

Features

SLP76

76

Platelets, mast cells, T cells/thymocytes, monocytes/macrophages and NK cells

VAV1, NCK, GADS, SLAP 130/FYB, PLCγ1 ITK and HPK1

Cytosolic adaptor tyrosine-phosphorylated by ZAP70 following TCR stimulation. SLP76 phosphorylation induces association with NCK and VAV1, implicating SLP76 in regulating actin rearrangements. Central proline-rich region binds the GRB2 homologue, GADS. Carboxy-terminal SH2 domain associates with tyrosine phosphorylated SLAP-130/FYB. Similar to LAT, SLP76 functions upstream of TCR-mediated PLCγ1 phosphorylation, Ca2+ influx and ERK activation.

LAT

36–38 Platelets, mast cells, T cell/thymocytes and NK cells

GRB2, GRAP, GADS, p85 subunit of PI3K, PLCγ1 and ITK (also indirectly associates with SLP76, VAV and c-CBL)

Transmembrane, constitutively associated with lipid rafts. Cytoplasmic domain contains several tyrosine residues that are phosphorylated by ZAP70 following TCR stimulation, which induces recruitment of several SH2-domain containing proteins. Required for TCR-induced PLCγ1 phosphorylation, Ca2+ influx and ERK activation.

GADS

40

SLP76, LAT, SHC and BCR-ABL

Cytosolic adaptor homologous with GRB2, but contains a proline/ glutamine-rich region between the SH2 and C-terminal SH3 domains. C-terminal SH3 domain constitutively binds proline-rich sequence in SLP76, SH2 domain associates with phosphorylated LAT following TCR stimulation. In this way, GADS recruits SLP76 to LAT following TCR stimulation.

PAG/ CBP

80–85 Ubiquitous

CSK and FYN

Palmitoylated type III transmembrane adaptor constitutively targeted to lipid rafts. Cytoplasmic domain contains several tyrosine residues phosphorylated by SRC-family PTKs. In resting cells, PAG/CBP is constitutively tyrosine phosphorylated and bound to the inhibitory kinase, CSK. On TCR stimulation, PAG/CBP is dephosphorylated, liberating CSK, therefore releasing SRC-family PTKs from CSKmediated inhibition. PAG/CBP might also allosterically activate CSK. PAG/CBP also contains two proline-rich motifs and several serine/ threonine residues (putative substrates for CK2 and/or PKC), indicating other potential modes of regulation.

Cbl-b

110

Ubiquitous

p85α, ZAP70, LCK, PLCγ1 and VAV1

Cytosolic adaptor, contains TKB domain (composed of an EF-hand, a four-helix bundle and a divergent domain) and a RING-finger domain in its amino-terminus. Like other members of the CBL/SLI family, Cbl-b is thought to function as a ubiquitin ligase, coupling TKB-bound substrates to E2 ubiquitin-conjugating enzymes, thereby targeting signalling molecules for proteasome-mediated degradation.

VASP

~46

Ubiquitous

Zyxin, vinculin, ActA, SLAP-130/FYB, profilin, ABL, SRC and F-actin

Associated with regulation of actin cytoskeleton in cells such as platelets, neurons, fibroblasts and leukocytes. VASP molecules share a conserved N-terminal EVH1 domain, which binds to proline-rich sequences present in components of the actin cytoskeleton, and a C-terminal EVH2 domain that binds F-actin and mediates oligomerization of ENA/VASP members. Sequences intervening the EVH1 and EVH2 domains are divergent among ENA/VASP proteins. VASP contains a central proline-rich region that associates with SH3 and WW domain-containing proteins, such as profilin.

Platelets, thymocytes/ T cells, macrophages, mast cells and NK cells

Description of several adaptor proteins characterized in lymphocytes. Palmitolyation modifications are present in both LAT and PAG/CBP, which mediate their constitutive localization to lipid rafts. ERK, extracellular signal-regulated kinase; EVH1; ENA/VASP-homology 1; GEF, guanine nucleotide-exchange factor; ITK, IL-2inducible T-cell kinase; LAT, linker for activated T cells; NK cells, natural killer cells; PI3K, phosphatidylinositol 3-kinase; PLC-γ1; phospholipase C-γ1; PTKs, protein tyrosine kinases; SH2, SRC homology 2; TCR, T-cell receptor; TKB, tyrosine kinase-binding.




REVIEWS

LIPID RAFTS

Micro-aggregates of cholesterol and sphingomyelin thought to occur in the plasma membrane. Also described as glycolipidenriched membrane microdomains (GEMs) or detergent-insoluble glycosphingolipid-enriched membrane microdomains (DIGs).

98

role for SLP76 in thymocyte development, as well as TCR-mediated signalling pathways required for IL2 promoter activity. Potential mechanisms underlying SLP76 function were suggested after the identification of another new adaptor molecule, the linker for activated T cells (LAT). LAT is a transmembrane adaptor protein that contains numerous tyrosine residues in its cytoplasmic domain (TABLE 1), which become phosphorylated by SYK-family PTKs in response to TCR engagement12. In addition, LAT contains palmitoylation sites that are responsible for its constitutive localization to LIPID RAFTS and are required for its TCR-dependent tyrosine phosphorylation13–15. Similar to SLP76, the expression of LAT is restricted to thymocytes, T cells, mast cells, natural killer (NK) cells and platelets. Following TCR engagement, LAT becomes tyrosine phosphorylated and associates with several SH2 domain-containing proteins, including growth factor receptor-bound protein 2 (GRB2), phospholipase C γ1 (PLCγ1), IL-2-inducible T-cell kinase (ITK) and the p85 subunit of the lipid kinase, phosphatidylinositol 3-kinase (PI3K)12. In addition, LAT contains at least two binding sites for the SH2 domains of the adaptor GADS16–18 (TABLE 1). Through its constitutive association with GADS, SLP76 indirectly binds to LAT following TCR ligation16–19, thereby coupling its associated proteins to regions of active signal complex formation. So, one model has emerged proposing that SLP76 and LAT function as mutually dependent intermolecular scaffolds, together recruiting crucial signalling regulators to sites of raft aggregation and T-cell/antigen-presenting cell (APC) contact (FIG. 1). Evidence to support the role of SLP76 and LAT in common pathways downstream of the TCR has been derived from both cell line and in vivo studies (BOX 2). Thymocytes deficient in either SLP76 or LAT expression exhibit a complete block in thymopoiesis at the pro-T3 stage in which the rearranged TCRβ chain is coupled to the invariant pre-Tα chain10,11,20. Progression through this developmental checkpoint requires effective signalling initiated by the pre-TCR21–23, therefore implicating SLP76 and LAT in this process. Furthermore, the generation of mutant Jurkat T-cell lines lacking either SLP76 or LAT expression show a vital role for both molecules in mediating TCR-induced PLCγ1 phosphorylation, extracellular signal-regulated kinase (ERK) activation, Ca2+ influx and IL2 promoter activity9,13,24. However, although both LAT and SLP76 cooperate in similar TCR signalling pathways, the precise molecular mechanism for their functions remains unclear. Biochemical studies have indicated a trimolecular complex, including SLP76, LAT and GADS. Evidence to support the hypothesis that recruitment of SLP76 to LAT (and hence lipid rafts) following TCR ligation is important for transducing TCR signals derives from reconstitution studies using LAT-deficient Jurkat T cells25. In these experiments, expression of a chimeric molecule in which the extracellular and transmembrane domains, in addition to the palmitoylation sites of LAT, are fused to full-length SLP76 results in rescue of TCRinduced IL-2 NFAT (nuclear factor of activated T cells)

promoter activity in the absence of endogenous LAT. A variant of the LAT/SLP76 chimaera in which the tyrosines of SLP76 were mutated to phenylalanine fails to rescue signalling in LAT-deficient cells. This indicates that a crucial function of SLP76 might be to recruit signalling molecules to lipid rafts in which the engaged TCR and activated PTKs can effectively couple to downstream mediators. Although the identity of the key SLP76-interacting protein is not yet known, candidates include VAV and ITK and/or other TEC-family PTKs. In this regard, it is important to note that activation of PLCγ1 seems to require both SYK- and TEC-family kinases. Therefore, one important function of SLP76 might be to recruit ITK to lipid rafts through the interaction between SLP76 and LAT, therefore allowing for optimal phosphorylation and activation of PLCγ1. It is clear, however, that this model is also overly simplistic as recent work has shown that PLCγ1 itself might have a vital role as a molecular bridge26. In this report, evidence was provided, which indicated that, in addition to binding LAT, PLCγ1 associates with phosphorylated ZAP70 (by means of a tyrosine between the two ZAP70 SH2 domains). This results in a transient recruitment of ZAP70 to LAT, leading to full phosphorylation of LAT and other associated molecules. This putative ‘adaptor’ function of PLCγ1 underscores the notion that, in addition to having important roles as enzymes, many effector molecules are crucial by virtue of their ability to create large intermolecular complexes. It is likely that the functions of SLP76 and LAT go beyond the mere recruitment of molecules to lipid rafts to increase the local concentration of substrates to upstream effector molecules. Support for the idea that SLP76 and LAT might function independently of their interaction has been shown in mutations which interfere with the association between SLP76 and LAT but do not completely abrogate function of the TCR. So, reconstitution of the LAT-deficient Jurkat T-cell line with a LAT molecule that cannot bind to GADS, and therefore fails to associate with SLP76, still rescues TCR-mediated Ca2+ influx to wild-type levels, but fails to reconstitute ERK activation18,27. By contrast, mutation of a tyrosine residue of LAT that abolishes its interaction with the SH2 domain of PLCγ1, but not GADS, results in a loss of PLCγ1 phosphorylation and defective NFAT promoter activity. In addition, reconstitution studies in SLP76-deficient Jurkat T cells show that the GADS-binding domain of SLP76 is not absolutely required for its activity, but rather contributes to optimal TCR-mediated PLCγ1 and ERK activation28. Furthermore, mice deficient in Gads expression (TABLE 2) do not show a complete block in thymopoiesis29. Rather, Gads-deficient thymocytes show an arrest in thymocyte development at the pro-T3 stage, but with a significant proportion of thymocytes maturing to the CD4+CD8+ double-positive and CD4+ or CD8+ single-positive stages. Interestingly, in the context of either MHC-class-II- or class-I-restricted TCR transgenes, Gads-deficient thymocytes do show severe defects in both positive and negative selection. Further evidence




REVIEWS to support the function of GADS as a vehicle for delivering important signalling molecules to LAT are indicated by studies in which expression of a Gads mutant that lacks a functional SH2 domain results in a dominant inhibition of TCR-mediated signalling and interferes with thymocyte maturation through the pro-T3 stage as well as thymocyte selection processes30.

Further evidence to challenge the model that SLP76 largely functions to recruit important effector molecules to LAT, thereby localizing them to lipid rafts, comes from genetic-complementation studies in cell lines. Experiments utilizing LAT-deficient Jurkat T cells reconstituted with different tyrosine-to-phenylalanine mutants of LAT indicate that these mutants cannot

Box 2 | Techniques to assess adaptor function

Biochemical Various methodologies to analyse protein–protein interactions, including co-precipitation of proteins by specific antibodies or by immobilized recombinant fusion proteins (for example, His-tagged proteins coupled to nickel substrates or glutathione S-transferase fusion proteins immobilized to glutathione substrates). In addition, biochemical analyses have allowed examination of signalling pathways regulated by adaptor proteins by assessing enzymatic activities of effector molecules (for example, in vitro kinase assays) or indicators of their activation state (for example, detection of their phosphorylation by phospho-specific antibodies). Subcellular localization of some adaptor proteins have also been elucidated by biochemical techniques. Separation of total plasma membrane fractions from cytosolic fractions can determine the association of proteins with membranes under different stimulation conditions. Similarly, lipid rafts can be resolved from total plasma membrane fractions, based on relative resistance of rafts to Triton extraction and their low density on a sucrose gradient.

Genetic Yeast two-hybrid: used to identify and characterize new adaptor proteins or their binding partners. Takes advantage of the modular nature of transcription factors in which the DNA-binding domain of a transcription factor (for example, GAL4 or LexA) is fused to a specific protein, while another known protein or library to be screened is fused to the transactivation module. Neither module alone can function independently; however, if the two proteins interact, the transcriptional modules cooperate in mediating transcription of a reporter gene. Cell-line genetic assays: random mutagenesis of T-cell lines has led to the identification of several mutant clones deficient in their ability to induce T-cell receptor (TCR)-mediated signals. Characterization of these mutant cell lines by reverse genetics has shown requirements for several adaptors in TCR-mediated signalling events. The DT40 chicken B-cell line is unique owing to its ability to undergo a high rate of homologous recombination, which allows for targeting of specific genes for disruption. Other approaches include expression of dominant-interfering or dominant-activating mutants of specific molecules in cell lines, as well as use of antisense to inhibit translation of specific transcripts. In vivo genetic assays in mice: has led to much of our understanding of how proteins function in thymocyte and T-cell biology in vivo. Specific genes can be targeted for disruption or for expression of specific mutations by homologous recombination to generate knockout and knock-in mice, respectively. For transgenic mice, genes of interest, or their mutant variants, can be expressed by a variety of promoters designed to direct expression in specific lineages and can be used to modulate expression at specific times. Use of the Cre recombinase/lox system has led to the generation of inducible knockouts, allowing for the analysis of genes that result in embryonic lethality when disrupted by standard techniques. Retroviral-mediated transduction of haematopoietic cells can also be used to asses the effects of expressing specific genes in transduced primary cells in vivo and in vitro.

Imaging Confocal microscopy: Uses fluorochrome-conjugated antibodies or fusion proteins tagged with specific fluorescent proteins to visualize molecules within cells. Can be used to characterize co-localization and/or clustering of proteins in or on cells, as well as to define their subcellular localization. Fluorescently conjugated reagents can resolve distinct structural components of cells, such as actin polarization and raft aggregation (for example, phalloidin or green fluorescent protein (GFP)–actin to examine actin polymerization and cholera toxin to study lipid rafts on cells). Fluorescence-activated cell sorting (FACS): Used to characterize surface phenotypic markers expressed on specific cell types within a heterogeneous population. Can also be used to examine expression levels of intracellular proteins when used with permeablized cells. Commonly used to define developmental subsets as well as to detect cellular responses to stimuli that can be detected with fluorescently labelled reagents. Fluorescence resonance energy transfer (FRET): Used to measure protein–protein interactions microscopically or by FACS-based methods. Cyan fluorescent protein (CFP), yellow fluorescent protein (YFP) or red fluorescent protein (RFP) fused proteins are expressed and assessed for interaction by measuring energy transfer between fluors which can only occur if proteins physically interact. Emission spectra indicating the interaction can be measured by flow cytometric analysis or visualized microscopically to discern subcellular locations in which FRET occurs within cells. FRET can also be used to examine the activation state of certain proteins if their activation results in a specific protein–protein interaction; for example, FLAIR (fluorescence activation indicator for RHO proteins) is a modification of FRET based on the observation that RAC associates with p21-activated kinase (PAK) only when activated.




REVIEWS

TCR GEMs γ ε

α

β

τ CD4

-76

Cγ

RAS

P P

LAT 1P

LCK

P GRB2

P

P

VAV

PL

P

ZAP70

SLP

GDP

S1 CDγ GLA P

NC

K

P

K PA

DAG

G

RasGRP

TP

RH O

[Ca2+]i

Actin reorganization

PKCθ

MAPK

Activation of transcription factors e.g. NFAT, NF-κB Nucleus

Figure 1 | Positive regulation of TCR signalling by adaptor proteins. A model for SLP76 and LAT function in T-cell receptor (TCR) signalling is illustrated. Following TCR engagement, SRC-family protein tyrosine kinases (PTKs; e.g. LCK) are activated, resulting in phosphorylation of CD3 modules of the TCR complex and activation of SYK-family PTKs (e.g. ZAP70). Activated ZAP70 phosphorylates LAT and SLP76. Tyrosine-phosphorylated LAT then recruits several Src-homology 2 (SH2) domain-containing proteins, including GRB2, GADS and phospholipase Cγ1 (PLCγ1) to lipid rafts. Through its constitutive association with GADS, SLP76 is also recruited to LAT following TCR stimulation. Evidence indicates that SLP76 also constitutively associates with the SH3 domain of PLCγ1, and that formation of a multimolecular complex between LAT, GADS, SLP76 and PLCγ1 is required for optimal PLCγ1 activation. Activation of PLCγ1 results in the hydrolysis of phosphatidylinositol 4,5bisphosphate to inositol 3,4,5-triphosphate (IP3) and diacylglycerol (DAG). IP3 production leads to increases of cytosolic free Ca2+ [Ca2+]i, whereas DAG can activate both protein kinase Cθ (PKCθ) and Ras guanyl nucleotide-releasing protein (RasGRP). Phosphorylated LAT also recruits the SH2 domain of GRB2 to lipid rafts, and therefore, the GRB2-associated RasGEF, SOS, thereby providing an additional possible mechanism of Ras activation through LAT. Tyrosine-phosphorylated SLP76 also associates with the RHO-family GEF, VAV, and the adaptor protein, NCK. A trimolecular complex between SLP76, VAV and NCK-associated p21activated kinase 1 (PAK1) has been proposed as a potential mechanism for SLP76 regulation of actin cytoskeletal rearrangements following TCR stimulation. GEMs, glycolipid-enriched membrane microdomains; NFAT, nuclear factor of activated T cells.

GUANINE NUCLEOTIDEEXCHANGE FACTOR

(GEF). Proteins that activate low-molecular-mass GTPases, such as RHO-family GTPases and RAS by stimulating the dissociation of GDP, and therefore promoting formation of the active GTP-bound state of these GTPases.

100

functionally complement each other when coexpressed within the same cell. This finding indicates that the function of LAT might be partly explained by combinatorial cis-acting properties, which operate to form a specific multimolecular complex between several proteins27. Recent evidence has shown a constitutive interaction between SLP76 and PLCγ1 that is vital for SLP76 function in TCR-mediated signalling pathways28. In addition, reconstitution of LAT-deficient Jurkat T cells with a LAT mutant that cannot bind to GADS, and hence SLP76, results in a secondary inhibition of the LAT–PLCγ1 association18. From these data, one model indicates that the association between SLP76 and LAT, as well as PLCγ1, is required for stable

recruitment of PLCγ1 to LAT. Additionally, it is also possible that formation of a defined complex involving LAT, PLCγ1, SLP76 and GADS functions to regulate the conformational state of PLCγ1 or other effectors. Collectively, these studies show that, although crucial for optimal TCR-signalling efficiency, the recruitment of SLP76 to LAT is not absolutely required for function. It is likely that the induced translocation of SLP76 to LAT regulates the efficiency of TCR-signalling pathways by modulating the amount of SLP76 recruited into lipid rafts. However, it seems that TCR signalling occurs, albeit at a suboptimal level, in the absence of a LAT–SLP76 association, resulting in shifts in the TCR repertoire of selected thymocytes observed in Gads-deficient mice29,30. This is reminiscent, but not identical to the abnormalities in selection observed when levels of the GADS family member, GRB2, are modulated. In Grb2+/– thymocytes, GRB2 protein levels are decreased, which results in reduced RAS activation during selection processes. This results in defects in negative selection, although positiveselecting events seem to occur normally31. It remains unclear, however, if the observed effect is owing to decreased RAS activation through the GRB2–SOS pathway, or if the GRB2 deficiency affects associations between LAT and SLP76. An alternative explanation for the results described above is that an association between LAT and SLP76 is required for TCR-signalling pathways, and that SLP76 can bind to LAT independently of GADS, although inefficiently. It is also possible that SLP76 might still localize to lipid rafts independently of LAT in such a manner that permits the two adaptors to cooperate functionally. When comparing results from cell line and in vivo studies, it is notable that certain functions (for example, proliferation) can be uncoupled from Ca2+ influx and IL-2 production32. Therefore, studies which use IL2 promoter activity in Jurkat T-cell lines as a model for T-cell activation might only show requirements for a subset of readouts that do not necessarily reflect the full repertoire of thymocyte and T-cell functions. In addition, there exist unique characteristics of the Jurkat T-cell model, such as mutation of the lipid phosphatase PTEN (phosphatase and tensin homologue)33. This mutation results in the accumulation of lipid second messengers which might alter some of the signalling properties of Jurkat cells, making it difficult to extrapolate to primary T cells. So, in vivo studies in combination with comprehensive measurements of T-cell function will be important for generating more definitive information on the structural prerequisites of adaptor proteins in thymocyte and T-cell biology. Interactions with the cytoskeleton. Adaptor molecules also serve important roles in coupling TCR engagement to cortical actin rearrangements that are required for T-cell activation. SLP76 contains tyrosine residues that, when phosphorylated, associate with the SH2 domains of VAV and NCK34–36. The GUANINENUCLEOTIDE-EXCHANGE-FACTOR (GEF) activity of VAV functions to activate RAC/RHO GTPases that are known to regulate actin reorganization. As NCK associates with




REVIEWS

Table 2 | Gene knockouts of some adaptor proteins and their phenotypes Adaptor

Phenotype

Slp76

10,11,93,94 Complete block in thymocyte development at the pro-T3 stage (CD4–CD8–, CD25+CD44–), indicating a requirement for SLP76 in pre-T-cell-receptor (TCR) signalling. Deficient mice also show defects in haemostasis, presumably due to a requirement for SLP76 in receptor signalling in platelets. Slp76–/– bone marrow-derived mast cells exhibit defects in FcεR functions, including mast-cell degranulation and interleukin (IL)-6 secretion, resulting in in vivo resistance to IgE-mediated passive anaphylaxis.

References

Lat

Complete block in thymocyte development at the pro-T3 stage. Defects in mast-cell function, also similar to Slp76–/–. In contrast with Slp76-deficient mice, Lat–/– mice show no gross defects in haemostasis.

20,95

Slp65/Blnk

Block in B-cell development at the pro-B/pre-B-cell transition. In addition, humans deficient in SLP65/BLNK expression also lack mature B cells.

96–98

Gads

Relative arrest in thymocyte development at the pro-T3 stage. Maturation to the CD4+CD8+ double-positive (DP) and CD4+ and CD8+ single-positive stages occurs; however, both positive and negative selection is defective. Mature T cells are present in Gads-deficient mice, but do not indicate TCR-induced Ca2+ influx.

29

Grb2 (haploinsufficiency)

Grb2+/– DP thymocytes show reduced Grb2 expression compared with Grb2+/+ thymocytes, with defects in negative selection attributed to decreased Ras signalling. In Grb2+/– thymocytes, extracellular signal-regulated kinase (ERK) activation is intact, although Ras-dependent activation of Jun N-terminal kinase (Jnk) and p38 is deficient, indicating that Jnk and p38 maintain a higher threshold for activation through Ras than that required for Erk.

31

Cbl-b

Diffuse spontaneous autoimmune disease with infiltration of activated polyclonal T and B cells into multiple tissues. T cells exhibit enhanced TCR-mediated proliferation, IL-2 production, antigen-receptor clustering and lipid-raft aggregation, which occur independently of a requirement for CD28 co-stimulation.

32,69,70

c-Cbl

Lymphoid hyperplasia. c-Cbl-deficient thymocytes show increased TCR surface expression, as well as enhanced TCR-induced phosphorylation of ZAP70, LAT and SLP76.

66,67

Slap-130/ Fyb

Grossly normal T-cell development and proximal signalling through the TCR, but profound defect in proliferation of purified T cells ex vivo. T cells stimulated through their TCR fail to upregulate integrin function, indicating a defect in ‘inside-out’ signalling.

99,100

the RAC effector molecule p21-activated kinase 1 (PAK1) (REFS 37–39), it has been suggested that SLP76 functions as a molecular bridge to couple GTP-bound RAC to NCK-associated PAK1 (FIG. 1). In support of this hypothesis, overexpression of a SLP76 mutant that cannot be tyrosine phosphorylated in Jurkat T cells results in an inhibition of PAK1 activation, as well as TCR-induced actin rearrangements36. A similar model has been proposed for the fibroblast adaptor protein, IRSp53, which has been shown to function as a molecular link between activated RAC and the RAC effector molecule, WASP-family Verprolin-homologous protein (WAVE)40. Therefore, SLP76 might be vital for positioning activated RHO-family GTPases with specific effector molecules in distinct spatial patterns. However, it should be noted that studies utilizing SLP76-deficient Jurkat T cells indicate that SLP76 is not required for either RAC1 activation or PAK1 activity in response to TCR stimulation41. Additional studies will be required to examine the activities of these effector molecules at a subcellular level in addition to downstream measurements of RHO GTPase activation, such as TCR-mediated actin reorganization. Recently, LAT has been shown to be required for actin-dependent spreading of Jurkat T cells on antiTCR-coated coverslips42. This spreading is thought to be vitally dependent on TCR-mediated Ca2+ influx. The requirements for spreading seem more complex, however, as a LAT mutant that cannot associate with SLP76

is not sufficient to induce spreading in the LAT-deficient Jurkat T-cell line, although it is sufficient to mediate a substantial Ca2+ response following TCR engagement18,27. It is interesting to speculate that the failure of this LAT mutant to function in this assay might relate to a possible role for SLP76 in coupling VAV GEF activity for RHO GTPases to specific subcellular regions, including regions of T-cell–APC contact. In another study, using a similar assay with immobilized agonist antibodies, Jurkat T-cell spreading has been shown to be inhibited by a dominant-negative mutant of RHOA and the RHO-specific inhibitor, C3 exotoxin, but is not affected by inhibition of RAC or CDC42 (REF. 43). However, although these studies examine an important TCRmediated morphological response that probably represents a required component of T-cell–APC synapse formation, it is unclear what the relative requirements for these various signalling molecules are under more physiologic conditions in which TCR engagement occurs on the surface of an APC in the presence of other receptors, such as adhesion molecules. Several lines of evidence indicate that the subcellular localization of active RAC is crucial for its function in mediating specific morphological changes. This observation was first made using fluorescence-activation indicator for RHO proteins (FLAIR), a technique based on fluorescence resonance energy transfer (FRET) and developed to visualize and quantify the spatio-temporal dynamics of active RAC1 GTPase in live cells44 (BOX 2).


VOLUME 1 | NOVEMBER 2001 | 1 0 1


REVIEWS Subsequently, similar techniques have been developed to examine the spatio-temporal pattern of active RAS and RAP1 localization in response to growth-factor stimulation of various cell lines45. In this study, the local activities of the GTPase regulators, including the GEFs and/or GTPase-activating proteins (GAPs) for RAS and RAP1, were implicated in regulation of the local activities of both GTPases. Additional studies in both yeast and neurons provide supporting evidence to indicate that the spatial distribution of active RHOfamily GTPases or the RHO-family GEFs is tightly coupled to specific morphological responses46–48. It is interesting to speculate that SLP76 might function to spatially orient VAV or RHO-family effector molecules to a particular subcellular ‘landmark’, perhaps marked by phosphorylated LAT, within sites of raft aggregation. Negative regulators

Inhibition of enzymatic function by SRC family PTKs. In addition to possessing enzymatic function, SRC family PTKs also contain protein-binding domains that enable them to act as adaptors. All SRC family members contain one SH3 and one SH2 domain, which is located amino-terminal to their kinase region. Extensive study, including biochemical as well as structural analyses, of the adaptor domains has shown their role as regulators of kinase activity. So, an intramolecular interaction between SRC family SH2 domains and a phosphorylated carboxy-terminal tyrosine results in a conformation that limits substrate accessibility, that renders the kinase domain catalytically inactive. The SH3 domain collaborates in this process as recent studies show that the SH2 and SH3 domains are functionally coupled by a linker region49. This coupling creates an ‘inducible snap lock’ that clamps the SH2 and SH3 domains on the phosphorylated C-terminal tyrosine (FIG. 2a). Mutations of residues that induce conformational flexibility in the connector region result in a loss of coupling between the SH2 and SH3 domains and generates a kinase that is constitutively active, despite the association between the SH2 domain and the C-terminal phosphotyrosine. These observations emphasize that proteinbinding domains are vital in the allosteric regulation of enzymatic activity. Furthermore, the function of protein-binding domains are thought to be intimately dependent on the context of the entire protein. On the basis of this model for SRC kinase regulation, it is clear that the interaction of kinases and phosphatases that are specific for the C-terminal tyrosine should be a crucial determinant of TCR-signalling efficiency. Studies in cell lines and genetically altered mice show the expected phenotype. Therefore, in the absence of CSK (which phosphorylates the negative regulatory C-terminal tyrosine of SRC-family kinases), the TCR becomes hyperactivated50, and when CSK is forced to the membrane, TCR signalling is downregulated51. By contrast, CD45 (which removes the phosphate from the C-terminal tyrosine) deficiency in cells52 or targeted deletion of CD45 (REF. 53) results in a failure of the TCR to couple with its downstream signalling machinery.

102

Until recently, however, little was known about how CSK and CD45 themselves are regulated. It is now appreciated that subcellular localization of both CSK and CD45 are essential features of their function. Although the intermolecular interactions responsible for directing CD45 within the cell are just now being defined, recent studies indicate that the subcellular localization of CSK is largely dependent on a recently identified adaptor known as phosphoprotein associated with glycolipid-enriched microdomains (PAG)54 or CSK-binding protein (CBP)55 (TABLE 1). Control of subcellular localization, PAG/CBP. One mode of negative regulation used by adaptor proteins involves the sequestration of signalling molecules into discrete subcellular locations, thereby governing the proximity to their substrates or upstream regulators. So, to interfere with activation events, adaptor molecules can either recruit the activities of negative effector molecules to the vicinity of their substrates or sequester positive regulators away from sites of activating signalling complexes. PAG/CBP is an adaptor protein that seems to utilize the former mechanism of action. PAG/CBP is a ubiquitously expressed transmembrane molecule, which is constitutively localized to lipid rafts. PAG/CBP contains several phosphorylatable tyrosine residues and two potential binding sites for SH3 domain-containing proteins in its cytoplasmic domain, in addition to a putative palmitoylation site54,55. In unstimulated cells, PAG/CBP is tyrosine phosphorylated and associates with the SH2 domain of CSK bringing the kinase to lipid rafts. Co-localization of CSK with raft-associated SRC-family kinases provides a mechanism by which CSK can maintain these enzymes in the inactive state. Following TCR engagement, PAG/CBP becomes dephosphorylated by an, as yet, unidentified phosphatase, and releases CSK, thereby liberating SRC-family kinases from phosphorylationmediated inhibition (FIG. 2b). It has also been shown that PAG/CBP might allosterically regulate CSK activity by binding to and increasing CSK substrate accessibility56. One study indicated that addition of a phosphopeptide of PAG/CBP, known to mediate its association with the SH2 domain of CSK, significantly increases CSK activity in a cell-free in vitro system, indicating that PAG/CBP might directly regulate CSK activity, as well as its subcellular localization. Expression of PAG/CBP in COS cells inhibits SRC kinase activity54,55, and overexpression of PAG/CBP in Jurkat T cells inhibits TCR-mediated NFAT promoter activity54, supporting the notion that PAG/CBP functions as a negative regulator of SRC kinase activity in T cells. In vivo studies examining PAG/CBP function will likely yield additional information on the role of this adaptor in the immune system, as well as in other cell types. Regulation of protein stabilization, c-CBL. The protooncogene c-CBL is an adaptor protein which has provided additional insights into the molecular mechanisms of how lymphocyte function might be negatively regulated. c-CBL belongs to the conserved CBL/SLI family




REVIEWS of proteins and contains a highly conserved N-terminal region consisting of a tyrosine kinase-binding (TKB) and RING-finger domains. The C-terminal region possesses numerous binding sites for SH3- and SH2-containing proteins. Crystal structures of the N-terminal domain complexed to a phosphorylated ZAP70 peptide determined that three structural motifs (a four-helix bundle, an EF hand and a divergent SH2 domain) together form an integrated phosphoproteinrecognition domain57. This region, designated the TKB domain, is structurally divergent from the typical phosphotyrosine-binding SH2 and PTB domains. The c-CBL RING-finger domain binds E2 ubiquitinconjugating enzymes, increasing their enzymatic function58,59. The active E2s are brought into proximity a

b SH3

GEMs α

SH2 P

α

β

to the TKB binders with consequent ubiquitylation of these target proteins. The ubiquitylated proteins are then degraded by the proteasome complex (FIG. 2c). The function of c-CBL as a ubiquitin ligase has been shown for several substrate ROS receptor PTKs, which associate with the TKB domain of c-CBL following their tyrosine phosphorylation58–62. In addition, the TKB domain of c-CBL has been shown to associate with SYK and ZAP70 tyrosine kinases in activated T cells57,63–65, although definitive evidence supporting c-CBL-mediated ZAP70 degradation remains less conclusive. c-Cbl clearly does have a role, however, in TCR signalling events as mice deficient in this protein exhibit lymphoid hyperplasia and enhanced thymocyte signalling through ZAP70 (REFS 66,67). Mechanistically, the importance of

GEMs

β α

β α γ ε

Kinase domain Inactive

P

PAG/ CPB

LCK

α

TCR β

β

τ

PAG/ CPB

P

P P

Y LCK

CSK

T-cell activation

Y

CSK Stimulated

Unstimulated SH3

c

Ubiquitylated protein

SH2 TKB P

ZAP70

Proteasome-mediated degradation of ubiquitylated protein

Y CBL Kinase domain

E2 Proteasome

RING

Active Ubiquitin

Degraded protein

Figure 2 | Negative regulation of TCR signalling by adaptors. Examples of negative regulation by adaptor molecules and adaptor domains are depicted. a | Allosteric inhibition by the adaptor domains of SRC-family kinases. The SRC-homology 2 (SH2) domain of SRC-family kinases binds to the carboxy-terminal phosphotyrosine residue, thereby restricting substrate accessibility and kinase activity. The SH3 domain has also been shown to regulate SRC kinase activity through intramolecular interactions that create an inducible ‘snap lock’, which is dependent on interdomain hinge regions as well. On dephosphorylation of the C-terminal tyrosine by the CD45 phosphatase, the adaptor domains are released and result in activation of the kinase. b | Recruitment of negative effector molecules to their substrates. In unstimulated T cells, raft-associated PAG/CBP is constitutively tyrosine phosphorylated and associates with the SH2 domain of CSK, bringing CSK into close proximity to its substrates (SRC-family PTKs) at the plasma membrane. Following TCR stimulation, PAG/CBP is dephosphorylated, resulting in the release of CSK from the membrane and relieving SRC-family kinases from CSK phosphorylation-mediated inhibition. Evidence also indicates that PAG/CBP might also regulate CSK activity independently of its ability to recruit CSK to lipid rafts. c | Regulation of protein stability. CBL family members regulate protein ubiquitylation and protesome-dependent degradation by functioning as an E3 ligase. The RING-finger domain associates with E2 ubiquitin-conjugating enzymes, whereas the aminoterminal tyrosine kinase-binding (TKB) domain recognizes tyrosine-phosphorylated substrates, such as ZAP70 or RPTKs. So, one model for CBL function indicates that CBL functions as a negative regulator of signalling pathways by promoting the ubiquitylation of important signalling molecules and their subsequent degradation by the proteasome complex. Note, evidence supporting this model for CBL-mediated PTK degradation is clear for receptor PTKs, but remains less conclusive for ZAP70 in T cells. GEMs, glycolipid-enriched membrane microdomains; Y, phosphorylatable tyrosine residue.




REVIEWS the RING domain for c-CBL function in T cells has been substantiated by studies of oncogenic forms of this protein, including the retrovirus-encoded v-Cbl and the 70Z/3 Cbl mutants, which carry disruptions in the RING-finger domain and fail to negatively regulate TCR signal transduction. Crystallographic analysis of ternary complexes involving the N-terminal domain of c-CBL, the recognition sequence of ZAP70, and the ubiquitin-conjugating enzyme E2L 3 (UBCH7) show that c-CBL might function to link E2s to their substrate PTKs and promote their ubiquitylation simply by increasing the effective local concentration of the substrate to the E2 enzyme68. However, the crystal structure also shows that the RING and TKB domains are tightly associated, providing precise relative orientations of the two domains that might be vital for functionally coupling E2 activity to its substrate. The conserved linker region between the TKB and RING domains contributes to this rigid arrangement. Furthermore, mutation of a residue in the linker domain, which disrupts the relative positions of the E2 and TKB domains, abrogates c-CBL function without affecting its ability to associate with either the substrate or E2s57,58. This is, therefore, another example of adaptor domains that do not always function independently of each other, but might intimately communicate to regulate enzyme activity as well as substrate specificity.

LAMELLIPODIA

Thin sheet-like processes, which extend at the leading edge of moving cells or neuronal growth cones in an actin-dependent fashion; promoted by the RHO-family GTPase, RAC1.

104

Relationship with low-molecular-mass GTPases, CBL-b. CBL-b (TABLE 1), another member of the CBL/SLI family of negative regulators, shares the conserved TKB and RING-finger domains with c-CBL and has been shown to be a crucial modulator of lymphocyte activation. Mice lacking Cbl-b expression exhibit spontaneous diffuse autoimmunity69 (TABLE 2). In these mice, both B and T cells display lower thresholds for activation, showing enhanced proliferation in response to B-cell receptor (BCR) or TCR stimulation. In addition, CBL-b-deficient T cells are relieved of the requirement for CD28 co-stimulation to induce IL-2 production and T-cell proliferation, attributed to an observed hyperphosphorylation of VAV1 following TCR stimulation69,70. Evidence indicates that the hyperresponsiveness displayed by CBL-b-deficient T cells might be caused by dysregulated membrane-receptor clustering and lipidraft aggregation32. This correlates with sustained tyrosine phosphorylation of several proteins following TCR stimulation, indicating that enhanced TCR clustering and/or actin reorganization prolongs the duration of TCR signalling and obviates the need for co-stimulatory signals to induce T-cell activation. CBL-b deficiency fails to complement defects in antigen receptor-induced TCR clustering or T-cell proliferation observed in T cells lacking the Wiskott–Aldrich syndrome protein (WASP), supporting the notion that Cbl-b negatively regulates VAV1 GEF activity for CDC42/WASP-dependent actin rearrangements in response to TCR engagement. It is notable that mice deficient in both Cbl-b and Vav1 expression still show autoimmune disease. This is thought to be due to hyperphosphorylated Vav2 that is present in Cbl-b/Vav1 double-deficient

T cells. Additionally, although genetic evidence indicates that augmentation of the VAV/CDC42/WASP signalling pathway is responsible for the hyperproliferative phenotype of the Cbl-b-deficient mice, it remains unclear how the function of Cbl-b as an E3 ubiquitin ligase leads to Vav1 hyperactivity. One model indicates that Cbl-b can serve as an E3 ligase for the p85 subunit of PI3K71. Failure to degrade this protein in Cbl-b-deficient animals is thought to lead to enhanced PI3K function, which in turn leads to production of lipid second messengers which interact with the VAV pleckstrin homology (PH) domain stimulating its guanine nucleotide exchange activity. Additional biochemical and genetic studies will be required to rigorously test this model. Adaptors and cell migration

Another group of adaptor proteins that have a wellcharacterized function in the regulation of neuronal and fibroblast migration, as well as axon guidance, are the ENA/VASP (vasodilator-stimulated phosphoprotein) proteins72 (TABLE 1). This family includes the Drosophila Enb (Enabled)73 and the mammalian ENA-related proteins, VASP74, EVL and Mena75. Members of this family share an N-terminal ENA/VASP-homology 1 (EVH1) domain, which binds to proteins containing a D/EFPPPPXD motif and is responsible for its localization to focal adhesions75–77. In addition, these proteins contain a central proline-rich domain followed by a C-terminal EVH2 domain, which associates with F-actin75,78,79. Studies which use Listeria monocytogenes as a cell-motility model, indicate that ENA/VASP proteins enhance actin-dependent translocation of the bacterium76,80–82. By virtue of the association between ENA/VASP proteins and F-actin, it has been speculated that their function in actin processes might be to couple nascent actin networks to the membrane. VASP has been shown to be enriched at leading edges of cells and correlates with a role in LAMELLIPODIAL protrusion83. However, evidence on the basis of studies of axon-guidance regulation indicates that ENA interacts with and functions downstream of the neuronal repulsive guidance receptor, Robo84. In addition, overexpression of ENA/VASP slows cellular migration, whereas displacement of these proteins from the membrane enhances cellular translocation85. These and other studies support a function for ENA/VASP proteins in serving as ‘molecular brakes’. The apparent contradiction in findings of the role for ENA/VASP in Listeria versus eukaryotic cell translocation can be explained by a potential requirement for actin polymerization in mediating active cellular repulsion or retraction. It has been suggested that in order for a cell to ‘pause’ and retract, multiple pseudopodia and lamellipodia are assembled in all directions, rather than forming a polarized lamellipodium at the leading edge as observed in moving cells72. In fact, it has been postulated that ENA/VASP proteins might function in both repulsive and attractive processes in which the outcome is dictated by the subcellular localization of these molecules. It has also been proposed that ENA/VASP proteins might promote cell retraction by stabilizing or




REVIEWS moulding actin filaments in a manner that enhances cell–cell or cell–substratum adhesion, thereby inhibiting cell translocation72. The recent finding of Robo receptors and the secreted Robo receptor ligand, SLIT3, in cells of the immune system presents the potential for conservation between molecular mechanisms governing neuronal guidance cues and those regulating leukocyte migration86. In this study, SLIT inhibited stromal cell-derived factor 1 (SDF1)-induced lymphocyte chemotaxis as well as N-formyl peptide f-Met-Leu-Phe (fMLP)-induced neutrophil migration. The inhibition was abolished by the addition of a soluble form of the Robo extracellular domain. These findings implicate a role for Robo and, potentially, other neuronal guidance cues in regulating chemokine-dependent migration of leukocytes. It will be interesting to determine if this or similar guidance mechanisms can also modulate T-cell–APC interactions by promoting cell–cell adhesion and formation of the immunological synapse. Understanding how ENA/VASP molecules function in these pathways to regulate lymphocyte functions will be essential. It is plausible that members of the ENA/VASP family might have other important functions in lymphocyte activation. The haematopoietic-specific adaptor protein, SLP76-associated phosphoprotein (SLAP-130; also known as FYN-binding protein or FYB), binds to the SH2 domain of SLP76 following its tyrosine phosphorylation by SRC-family PTKs stimulated by TCR engagement87,88. Although its function in T cells is unclear, evidence indicates a role for SLAP130/FYB in regulating cytoskeletal reorganization, which might be attributed to its potential functional association with ENA/VASP members. SLAP-130/FYB contains a putative recognition sequence for the EVH1 domain of ENA/VASP proteins, and has been shown to bind ENA/VASP proteins in T cells and platelets89. In addition, SLAP-130/FYB colocalizes with the ENA/VASP member EVL at the interface between T cells and anti-CD3-coated beads. In platelets, SLAP130/FYB has also been shown to colocalize with VASP

1.

2.

3. 4.

5.

6.

Straus, D. B. & Weiss, A. Genetic evidence for the involvement of the lck tyrosine kinase in signal transduction through the T cell antigen receptor. Cell 70, 585–593 (1992). June, C. H., Fletcher, M. C., Ledbetter, J. A. & Samelson, L. E. Increases in tyrosine phosphorylation are detectable before phospholipase C activation after T cell receptor stimulation. J. Immunol. 144, 1591–1599 (1990). Reth, M. Antigen receptor tail clue. Nature 338, 383–384 (1989). Chan, A. C., Irving, B. A., Fraser, J. D. & Weiss, A. The ζ chain is associated with a tyrosine kinase and upon T-cell antigen receptor stimulation associates with ZAP-70, a 70kDa tyrosine phosphoprotein. Proc. Natl Acad. Sci. USA 88, 9166–9170 (1991). Chan, A. C., Iwashima, M., Turck, C. W. & Weiss, A. ZAP-70: a 70 kd protein-tyrosine kinase that associates with the TCR ζ chain. Cell 71, 649–662 (1992). Iwashima, M., Irving, B. A., van Oers, N. S., Chan, A. C. & Weiss, A. Sequential interactions of the TCR with two distinct cytoplasmic tyrosine kinases. Science 263, 1136–1139 (1994).

at the periphery of spread platelets adhering to fibrinogen90. Further studies are needed to determine the function of SLAP-130/FYB in mediating actin reorganization in T cells, as well as other cell types. Conclusions

The identification of GRB2 as the prototypic adaptor protein approximately 10 years ago showed how disparate signalling pathways could be coupled by creating a multimolecular complex. As our knowledge of adaptor proteins has evolved, it has become apparent that these molecules serve complex functions in addition to recruiting molecules to specific subcellular regions. To understand more completely the diverse roles of adaptors which impact on complex cellular processes (such as thymocyte development and T-cell function), it will be necessary to integrate genetic, biochemical, imaging and cell biologic techniques to their study (BOX 2). Advances in each of these areas have led to and will certainly continue to provide new insights into adaptor protein biology. The implementation of ‘knockout’ technology to target genes encoding various adaptors has shown the importance of these proteins as both positive and negative regulators of T-cell function (TABLE 2). Using such mice as hosts, it is now possible to perform a comprehensive, in vivo structure/function analysis of the targeted adaptor to determine domains vital for various functions. These studies will be aided by crystallographic data indicating how primary sequences of adaptor proteins can be translated into functional protein-interaction domains. Additional biochemical analyses characterizing the binding partners for the various domains will help guide future genetic reconstitution experiments. Perhaps most importantly will be the application of real-time image analysis in living cells to provide essential information about the temporal and spatial organization of signalling complexes created as T cells are stimulated through various surface receptors. Only through the utilization of all of these approaches will we reach a thorough understanding of the role of adaptor molecules in T-cell development and activation.

7.

Jackman, J. K. et al. Molecular cloning of SLP-76, a 76-kDa tyrosine phosphoprotein associated with Grb2 in T cells. J. Biol. Chem. 270, 7029–7032 (1995). 8. Motto, D. G., Ross, S. E., Wu, J., Hendricks-Taylor, L. R. & Koretzky, G. A. Implication of the GRB2-associated phosphoprotein SLP-76 in T cell receptor-mediated interleukin 2 production. J. Exp. Med. 183, 1937–1943 (1996). 9. Yablonski, D., Kuhne, M. R., Kadlecek, T. & Weiss, A. Uncoupling of nonreceptor tyrosine kinases from PLC-γ1 in an SLP-76-deficient T cell. Science 281, 413–416 (1998). 10. Pivniouk, V. et al. Impaired viability and profound block in thymocyte development in mice lacking the adaptor protein SLP-76. Cell 94, 229–238 (1998). 11. Clements, J. L. et al. Requirement for the leukocytespecific adapter protein SLP-76 for normal T cell development. Science 281, 416–419 (1998). References 10 and 11 describe mice with targeted disruption of the Slp76 gene. These mice have a complete block in thymocyte development, presumably due to impaired signalling through the pre-TCR, emphasizing the vital role played by SLP76 as a positive regulator of T-cell function.


12. Zhang, W., Sloan-Lancaster, J., Kitchen, J., Trible, R. P. & Samelson, L. E. LAT: the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular activation. Cell 92, 83–92 (1998). 13. Zhang, W., Irvin, B. J., Trible, R. P., Abraham, R. T. & Samelson, L. E. Functional analysis of LAT in TCRmediated signaling pathways using a LAT-deficient Jurkat cell line. Int. Immunol. 11, 943–950 (1999). 14. Zhang, W., Trible, R. P. & Samelson, L. E. LAT palmitoylation: its essential role in membrane microdomain targeting and tyrosine phosphorylation during T cell activation. Immunity 9, 239–246 (1998). 15. Lin, J., Weiss, A. & Finco, T. S. Localization of LAT in glycolipid-enriched microdomains is required for T cell activation. J. Biol. Chem. 274, 28861–28864 (1999). 16. Liu, S. K., Fang, N., Koretzky, G. A. & McGlade, C. J. The hematopoietic-specific adaptor protein Gads functions in T-cell signaling via interactions with the SLP-76 and LAT adaptors. Curr. Biol. 9, 67–75 (1999). 17. Asada, H. et al. Grf40, a novel Grb2 family member, is involved in T cell signaling through interaction with SLP76 and LAT. J. Exp. Med. 189, 1383–1390 (1999).



REVIEWS 18. Zhang, W. et al. Association of Grb2, Gads, and phospholipase C-γ 1 with phosphorylated LAT tyrosine residues. Effect of LAT tyrosine mutations on T cell angigen receptor-mediated signaling. J. Biol. Chem. 275, 23355–23361 (2000). 19. Law, C. L. et al. GrpL, a Grb2-related adaptor protein, interacts with SLP-76 to regulate nuclear factor of activated T cell activation. J. Exp. Med. 189, 1243–1253 (1999). 20. Zhang, W. et al. Essential role of LAT in T cell development. Immunity 10, 323–332 (1999). Similar to Slp76-deficient animals, Lat–/– mice exhibit a complete arrest at the pro-T3 stage of thymocyte development. This result and studies in mutant variants of the Jurkat T-cell line indicate that Slp76 and LAT might function together as regulators of TCR signalling events. 21. Cheng, A. M. & Chan, A. C. Protein tyrosine kinases in thymocyte development. Curr. Opin. Immunol. 9, 528–533 (1997). 22. van Oers, N. S. T cell receptor-mediated signs and signals governing T cell development. Semin. Immunol. 11, 227–237 (1999). 23. Kruisbeek, A. M. et al. Branching out to gain control: how the pre-TCR is linked to multiple functions. Immunol. Today 21, 637–644 (2000). 24. Finco, T. S., Kadlecek, T., Zhang, W., Samelson, L. E. & Weiss, A. LAT is required for TCR-mediated activation of PLCγ1 and the Ras pathway. Immunity 9, 617–626 (1998). 25. Boerth, N. J. et al. Recruitment of SLP-76 to the membrane and glycolipid-enriched membrane microdomains replaces the requirement for linker for activation of T cells in T cell receptor signaling. J. Exp. Med. 192, 1047–1058 (2000). 26. Williams, B. L. et al. Phosphorylation of Tyr319 in ZAP-70 is required for T-cell antigen receptor-dependent phospholipase C-γ1 and Ras activation. EMBO J. 18, 1832–1844 (1999). 27. Lin, J. & Weiss, A. Identification of the minimal tyrosine residues required for LAT function. J. Biol. Chem. 6, 29588–29595 (2001). 28. Yablonski, D., Kadlecek, T. & Weiss, A. Identification of a phospholipase c-γ1 (Plc-γ1) SH3 domain-binding site in SLP-76 required for T-cell receptor-mediated activation of PLC-γ1 and NFAT. Mol. Cell. Biol. 21, 4208–4218 (2001). 29. Yoder, J. et al. Requirement for the SLP-76 adaptor GADS in T cell development. Science 291, 1987–1991 (2001). Describes mice made deficient in Gads. Unlike Slp76–/– or Lat–/– mice, thymocyte development is normal in Gads-deficient animals and mature T cells populate the periphery, but severe defects in thymocyte selection and function of the mature T cells which develop. 30. Kikuchi, K. et al. Suppression of thymic development by the dominant-negative form of Gads. Int. Immunol. 13, 777–783 (2001). 31. Gong, Q. et al. Disruption of T cell signaling networks and development by Grb2 haploid insufficiency. Nat. Immunol. 2, 29–36 (2001). 32. Krawczyk, C. et al. Cbl-b is a negative regulator of receptor clustering and raft aggregation in T cells. Immunity 13, 463–473 (2000). 33. Shan, X. et al. Deficiency of PTEN in Jurkat T cells causes constitutive localization of Itk to the plasma membrane and hyperresponsiveness to CD3 stimulation. Mol. Cell. Biol. 20, 6945–6957 (2000). 34. Fang, N. & Koretzky, G. A. SLP-76 and Vav function in separate, but overlapping pathways to augment interleukin-2 promoter activity. J. Biol. Chem. 274, 16206–16212 (1999). 35. Raab, M., da Silva, A. J., Findell, P. R. & Rudd, C. E. Regulation of Vav–SLP-76 binding by ZAP-70 and its relevance to TCR ζ/CD3 induction of interleukin-2. Immunity 6, 155–164 (1997). 36. Bubeck Wardenburg, J. et al. Regulation of PAK activation and the T cell cytoskeleton by the linker protein SLP-76. Immunity 9, 607–616 (1998). 37. Bokoch, G. M. et al. Interaction of the Nck adapter protein with p21-activated kinase (PAK1). J. Biol. Chem. 271, 25746–25749 (1996). 38. Galisteo, M. L., Chernoff, J., Su, Y. C., Skolnik, E. Y. & Schlessinger, J. The adaptor protein Nck links receptor tyrosine kinases with the serine-threonine kinase Pak1. J. Biol. Chem. 271, 20997–21000 (1996). 39. Lu, W., Katz, S., Gupta, R. & Mayer, B. J. Activation of Pak by membrane localization mediated by an SH3 domain from the adaptor protein Nck. Curr. Biol. 7, 85–94 (1997). 40. Miki, H., Yamaguchi, H., Suetsugu, S. & Takenawa, T. IRSp53 is an essential intermediate between Rac and WAVE in the regulation of membrane ruffling. Nature 408, 732–735 (2000).

106

41. Ku, G. M., Yablonski, D., Manser, E., Lim, L. & Weiss, A. A PAK1–PIX–PKL complex is activated by the T-cell receptor independent of Nck, Slp-76 and LAT. EMBO J. 20, 457–465 (2001). 42. Bunnell, S. C., Kapoor, V., Trible, R. P., Zhang, W. & Samelson, L. E. Dynamic actin polymerization drives T cell receptor-induced spreading: a role for the signal transduction adaptor LAT. Immunity 14, 315–329 (2001). Using real-time imaging, this paper shows that, unlike wild-type Jurkat cells, LAT-deficient Jurkat cells fail to spread on anti-TCR-coated coverslips. Reconstitution of LAT expression rescues this signalling defect; however, the structural features of LAT required for this function and the precise signalling pathways required remain to be determined. 43. Borroto, A. et al. Rho regulates T cell receptor ITAMinduced lymphocyte spreading in an integrin-independent manner. Eur. J. Immunol. 30, 3403–3410 (2000). 44. Kraynov, V. S. et al. Localized Rac activation dynamics visualized in living cells. Science 290, 333–337 (2000). Highlights the use of a method based on FRET modified to examine the spatio-temporal pattern of RAC1 activation in living cells. 45. Mochizuki, N. et al. Spatio-temporal images of growthfactor-induced activation of Ras and Rap1. Nature 411, 1065–1068 (2001). 46. Kang, P. J., Sanson, A., Lee, B. & Park, H. O. A GDP/GTP exchange factor involved in linking a spatial landmark to cell polarity. Science 292, 1376–1378 (2001). 47. Marston, A. L., Chen, T., Yang, M. C., Belhumeur, P. & Chant, J. A localized GTPase exchange factor, Bud5, determines the orientation of division axes in yeast. Curr. Biol. 11, 803–807 (2001). 48. Shamah, S. M. et al. EphA receptors regulate growth cone dynamics through the novel guanine nucleotide exchange factor ephexin. Cell 105, 233–244 (2001). 49. Young, M. A., Gonfloni, S., Superti-Furga, G., Roux, B. & Kuriyan, J. Dynamic coupling between the SH2 and SH3 domains of c-Src and Hck underlies their inactivation by C-terminal tyrosine phosphorylation. Cell 105, 115–126 (2001). Uses a combination of genetic and simulated crystallographic studies to show a unique role for the region linking the SH3 and SH2 domains of c-SRC and HCK. 50. Schmedt, C. et al. Csk controls antigen receptor-mediated development and selection of T-lineage cells. Nature 394, 901–904 (1998). 51. Chow, L. M., Fournel, M., Davidson, D. & Veillette, A. Negative regulation of T-cell receptor signalling by tyrosine protein kinase p50csk. Nature 365, 156–160 (1993). 52. Koretzky, G. A., Picus, J., Thomas, M. L. & Weiss, A. Tyrosine phosphatase CD45 is essential for coupling T-cell antigen receptor to the phosphatidyl inositol pathway. Nature 346, 66–68 (1990). 53. Kishihara, K. et al. Normal B lymphocyte development but impaired T cell maturation in CD45-exon6 protein tyrosine phosphatase-deficient mice. Cell 74, 143–156 (1993). 54. Brdicka, T. et al. Phosphoprotein associated with glycosphingolipid-enriched microdomains (PAG), a novel ubiquitously expressed transmembrane adaptor protein, binds the protein tyrosine kinase csk and is involved in regulation of T cell activation. J. Exp. Med. 191, 1591–1604 (2000). References 53 and 54 describe the initial characterization of PAG/CBP, indicating its localization in lipid rafts and its association with CSK. These papers provide evidence for how an adaptor protein might regulate an enzyme by directing its subcellular localization. 55. Kawabuchi, M. et al. Transmembrane phosphoprotein Cbp regulates the activities of Src-family tyrosine kinases. Nature 404, 999–1003 (2000). 56. Takeuchi, S., Takayama, Y., Ogawa, A., Tamura, K. & Okada, M. Transmembrane phosphoprotein Cbp positively regulates the activity of the carboxyl-terminal Src kinase, Csk. J. Biol. Chem. 275, 29183–29186 (2000). 57. Meng, W., Sawasdikosol, S., Burakoff, S. J. & Eck, M. J. Structure of the amino-terminal domain of Cbl complexed to its binding site on ZAP70 kinase. Nature 398, 84–90 (1999). 58. Joazeiro, C. A. et al. The tyrosine kinase negative regulator c-Cbl as a RING-type, E2-dependent ubiquitin-protein ligase. Science 286, 309–312 (1999) This study shows a function for c-Cbl as an E3 ubiquitin-protein ligase with direct in vitro evidence of a role for the TKB and RING domains. 59. Yokouchi, M. et al. Ligand-induced ubiquitination of the epidermal growth factor receptor involves the interaction of the c-Cbl RING finger and UbcH7. J. Biol. Chem. 274, 31707–31712 (1999).


60. Lee, P. S. et al. The Cbl protooncoprotein stimulates CSF-1 receptor multiubiquitination and endocytosis, and attenuates macrophage proliferation. EMBO J. 18, 3616–3628 (1999). 61. Miyake, S., Lupher, M. L. Jr, Druker, B. & Band, H. The tyrosine kinase regulator Cbl enhances the ubiquitination and degradation of the platelet-derived growth factor receptor alpha. Proc. Natl Acad. Sci. USA 95, 7927–7932 (1998). 62. Waterman, H., Levkowitz, G., Alroy, I. & Yarden, Y. The RING finger of c-Cbl mediates desensitization of the epidermal growth factor receptor. J. Biol. Chem. 274, 22151–22154 (1999). 63. Lupher, M. L. Jr et al. Cbl-mediated negative regulation of the Syk tyrosine kinase. A critical role for Cbl phosphotyrosine-binding domain binding to Syk phosphotyrosine 323. J. Biol. Chem. 273, 35273–35281 (1998). 64. Lupher, M. L. Jr, Reedquist, K. A., Miyake, S., Langdon, W. Y. & Band, H. A novel phosphotyrosine-binding domain in the N-terminal transforming region of Cbl interacts directly and selectively with ZAP-70 in T cells. J. Biol. Chem. 271, 24063–24068 (1996). 65. Lupher, M. L. Jr, Songyang, Z., Shoelson, S. E., Cantley, L. C. & Band, H. The Cbl phosphotyrosinebinding domain selects a D(N/D)XpY motif and binds to the Tyr292 negative regulatory phosphorylation site of ZAP-70. J. Biol. Chem. 272, 33140–33144 (1997). 66. Murphy, M. A. et al. Tissue hyperplasia and enhanced T-cell signalling via ZAP70 in c-Cbl-deficient mice. Mol. Cell. Biol. 18, 4872–4882 (1998). 67. Thien, C. B., Bowtell, D. D. & Langdon, W. Y. Perturbed regulation of ZAP-70 and sustained tyrosine phosphorylation of LAT and SLP76 in c-Cbl-deficient thymocytes. J. Immunol. 162, 7133–7139 (1999). 68. Zheng, N., Wang, P., Jeffrey, P. D. & Pavletich, N. P. Structure of a c-Cbl–UbcH7 complex: RING domain function in ubiquitin-protein ligases. Cell 102, 533–539 (2000). 69. Bachmaier, K. et al. Negative regulation of lymphocyte activation and autoimmunity by the molecular adaptor Cbl-b. Nature 403, 211–216 (2000). References 68 and 69 describe that Cbl-b–/– mice develop spontaneous autoimmunity after 6 months of age characterized by autoantibody production and generalized lymphocytic infiltration into multiple tissues. 70. Chiang, Y. J. et al. Cbl-b regulates the CD28 dependence of T-cell activation. Nature 403, 216–220 (2000). 71. Fang, D. et al. Cbl-b, a RING-type E3 ubiquitin ligase, targets phosphatidylinositol 3-kinase for ubiquitination in T cells. J. Biol. Chem. 276, 4872–4878 (2001). 72. Machesky, L. M. Putting on the brakes: a negative regulatory function for Ena/VASP proteins in cell migration. Cell 101, 685–688 (2000). 73. Gertler, F. B., Doctor, J. S. & Hoffmann, F. M. Genetic suppression of mutations in the Drosophila abl protooncogene homolog. Science 248, 857–860 (1990). 74. Halbrugge, M. & Walter, U. Analysis, purification and properties of a 50,000-dalton membrane-associated phosphoprotein from human platelets. J. Chromatogr. 521, 335–343 (1990). 75. Gertler, F. B., Niebuhr, K., Reinhard, M., Wehland, J. & Soriano, P. Mena, a relative of VASP and Drosophila Enabled, is implicated in the control of microfilament dynamics. Cell 87, 227–239 (1996). 76. Niebuhr, K. et al. A novel proline-rich motif present in ActA of Listeria monocytogenes and cytoskeletal proteins is the ligand for the EVH1 domain, a protein module present in the ENA/VASP family. EMBO J. 16, 5433–5444 (1997). 77. Carl, U. D. et al. Aromatic and basic residues within the EVH1 domain of VASP specify its interaction with prolinerich ligands. Curr. Biol. 9, 715–718 (1999). 78. Huttelmaier, S. et al. Characterization of the actin binding properties of the vasodilator-stimulated phosphoprotein VASP. FEBS Lett. 451, 68–74 (1999). 79. Bachmann, C., Fischer, L., Walter, U. & Reinhard, M. The EVH2 domain of the vasodilator-stimulated phosphoprotein mediates tetramerization, F-actin binding, and actin bundle formation. J. Biol. Chem. 274, 23549–23557 (1999). 80. Chakraborty, T. et al. A focal adhesion factor directly linking intracellularly motile Listeria monocytogenes and Listeria ivanovii to the actin-based cytoskeleton of mammalian cells. EMBO J. 14, 1314–1321 (1995). 81. Laurent, V. et al. Role of proteins of the Ena/VASP family in actin-based motility of Listeria monocytogenes. J. Cell Biol. 144, 1245–1258 (1999). 82. Loisel, T. P., Boujemaa, R., Pantaloni, D. & Carlier, M. F. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature 401, 613–616 (1999).



REVIEWS 83. Rottner, K., Behrendt, B., Small, J. V. & Wehland, J. VASP dynamics during lamellipodia protrusion. Nature Cell Biol. 1, 321–322 (1999). 84. Bashaw, G. J., Kidd, T., Murray, D., Pawson, T. & Goodman, C. S. Repulsive axon guidance: Abelson and Enabled play opposing roles downstream of the roundabout receptor. Cell 101, 703–715 (2000). 85. Bear, J. E. et al. Negative regulation of fibroblast motility by ENA/VASP proteins. Cell 101, 717–728 (2000). 86. Wu, J. Y. et al. The neuronal repellent Slit inhibits leukocyte chemotaxis induced by chemotactic factors. Nature 410, 948–952 (2001). 87. Musci, M. A. et al. Molecular cloning of SLAP130, an SLP76-associated substrate of the T cell antigen receptorstimulated protein tyrosine kinases. J. Biol. Chem. 272, 11674–11677 (1997). 88. da Silva, A. J. et al. Cloning of a novel T-cell protein FYB that binds FYN and SH2-domain-containing leukocyte protein 76 and modulates interleukin 2 production. Proc. Natl Acad. Sci. USA 94, 7493–7498 (1997). 89. Krause, M. et al. Fyn-binding protein (Fyb)/SLP-76associated protein (SLAP), ENA/vasodilator-stimulated phosphoprotein (VASP) proteins and the ARP2/3 complex link T cell receptor (TCR) signaling to the actin cytoskeleton. J. Cell Biol. 149, 181–194 (2000).

90. Obergfell, A. et al. The molecular adapter SLP-76 relays signals from platelet integrin αIIbβ3 to the actin cytoskeleton. J. Biol. Chem. 276, 5916–5923 (2001). 91. Fedorov, A. A., Fedorov, E., Gertler, F. & Almo, S. C. Structure of EVH1, a novel proline-rich ligand-binding module involved in cytoskeletal dynamics and neural function. Nature Struct. Biol. 6, 661–665 (1999). 92. Prehoda, K. E., Lee, D. J. & Lim, W. A. Structure of the enabled/VASP homology 1 domain-peptide complex: a key component in the spatial control of actin assembly. Cell 97, 471–480 (1999). 93. Pivniouk, V. I. et al. SLP-76 deficiency impairs signaling via the high-affinity IgE receptor in mast cells. J. Clin. Invest. 103, 1737–1743 (1999). 94. Clements, J. L. et al. Fetal hemorrhage and platelet dysfunction in Slp-76-deficient mice. J. Clin. Invest. 103, 19–25 (1999). 95. Saitoh, S. et al. LAT is essential for FcεRI-mediated mast cell activation. Immunity 12, 525–535 (2000). 96. Xu, S. et al. B cell development and activation defects resulting in xid-like immunodeficiency in Blink/Slp-65deficient mice. Int. Immunol. 12, 397–404 (2000). 97. Pappu, R. et al. Requirement for B cell linker protein (BLNK) in B cell development. Science 286, 1949–1954 (1999).


98. Jumaa, H. et al. Abnormal development and function of B lymphocytes in mice deficient for the signaling adaptor protein SLP-65. Immunity 11, 547–554 (1999). 99. Peterson E. J. et al. Coupling of the TCR to integrin activation by Slap-130/Fyb. Science 293, 2263–2265 (2001). 100. Griffiths, E. K. et al. Regulation of T cell activation and intregrin adhesion adapter Fyb/Slap. Science 293, 2260–2263 (2001).

Online links DATABASES The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ c-Cbl | CD3 | CD28 | CD45 | CDC42 | CSK | Enb | EVL | FYB | FYN | GADS | GAL4 | GRB2 | Il2 | IRSp53 | ITK | LAT | LCK | Mena | NCK | NFAT | PAG | PAK1 | PLCγ1 | PTEN | RAC1 | RAP1 | Robo | SDF1 | SLIT3 | SLP76 | UBCH7 | VASP | VAV1 | Vav2 | vinculin | WASP | WAVE | ZAP70 | zyxin FURTHER INFORMATION Gary Koretzky’s lab: http://www.uphs.upenn.edu/~immun/koretzky.html Access to this interactive links box is free online.



REVIEWS

ROLE OF CHEMOKINES IN THE PATHOGENESIS OF ASTHMA Nicholas W. Lukacs The prevalence of asthma has risen drastically in the last two decades, with a worldwide impact on health care systems. Although several factors contribute to the development of asthma, inflammation seems to be a common factor that leads to the most severe asthmatic responses. In the past decade, researchers have characterized a large group of chemotactic cytokines, also known as chemokines, which are implicated in asthmatic inflammation. These chemokines control and direct the migration and activation of various leukocyte populations. Targeting chemokines should lead to new ways of controlling the inflammatory asthmatic response.

University of Michigan Medical School, Department of Pathology, 1301 Catherine Road, Ann Arbor, Michigan 48109-0602, USA. e-mail: [email protected]

108

The incidence of asthma has increased twofold in ‘westernized’ societies in the past 20 years1. Recent estimates indicate that there has been an increase in asthma in children and young adults at a rate of 5–6% per year worldwide. Not only has there been an increase in overall frequency, but also an alarming increase in fatal asthma, especially in children2,3. The number of emergency-room visits and hospital admissions has also risen drastically, putting a strain on health care systems. It is estimated that there are billions of dollars lost in health care costs, as well as lost work and school time which lead to a reduction in overall productivity. Asthma is a heterogeneous disease that is broadly defined as a clinical syndrome characterized by altered lung function, peribronchial inflammation and airway responsiveness4,5. The precise definition and diagnosis of asthma are problematic, owing to the difficulty in accurately measuring the inflammatory response. Most asthmatics are identified by their adverse reaction to allergens and/or diminished lung function. A clear diagnosis of asthma is also hampered by the number of different forms of the condition, including allergic asthma, non-allergic intrinsic asthma, nocturnal asthma and occupational asthma, all of which have different phenotypes and disease patterns1. For example, nocturnal asthmatics seem to be affected predominantly during sleep, whereas occupational asthmatics exposed to particles, such as latex, have episodes primarily in the workplace6. So, although asthma is thought to have a genetic predisposition, the increased incidence might be

due to compounding factors, including: changes in environmental antigens, exposure to infectious agents, the increase in indoor lifestyles, as well as a general recognition by physicians of the asthmatic condition. The asthmatic response can be divided into two phases — early and late (FIG. 1). The early phase is an immediate response, usually to an antigenic or an environmental stimulus that initiates local mast-cell activation and mediator release, and causes an acute asthmatic attack. This initial phase will often resolve, but several hours later a second and more severe latephase response will occur. The late-phase response is induced by inflammatory cell infiltration and can last for prolonged periods7. The constant exposure to inducing stimuli, such as allergens, continues to promote inflammation and long-term damage to the airway, leading to severe consequences in chronic asthmatics8. The cellular infiltrate that characterizes asthma consists primarily of mononuclear cells (lymphocytes and macrophages) as well as eosinophils5,9. Because of the presence of eosinophils in most chronic asthmatics, these cells have long been thought to be the causative population that induces the detrimental effects of the disease. The ability to control leukocyte infiltration into the lungs is viewed as the key to regulating disease severity. This review will concentrate primarily on findings in animal models of asthma that might not fully reflect the human disease, but are likely to indicate a number of important mechanisms that should be considered.




REVIEWS

Early phase (5–60 mins)

Late phase (4–24 hours)

Eosinophils

IgE

Allergen Macrophages

Mast cell Airway damage/ reactivity Leukocyte accumulation

Degranulation Acute mediators

Mucus

T cells

Figure 1 | Phases of asthmatic responses. Asthma can be divided into two distinct phases. The early phase occurs quite rapidly on exposure to allergen, inducing mast cell degranulation, which causes mediator release and subsequent changes in airway function. Mediators released during the early phase and allergen-specific immune responses cause subsequent progression into the late-phase response. The late-phase response is often more severe and is characterized by the accumulation of mononuclear cells (monocytes and lymphocytes) as well as eosinophils. The cell populations that accumulate during the late-phase response are associated with prolonged airway dysfunction and damage. IgE, immunoglobulin E.

Overview of chemokines and asthma

A large family of peptide chemotactic cytokine mediators (also known as chemokines) have been identified that control leukocyte migration and activation. Chemokines bind to G-protein-coupled serpentine receptors and have been divided primarily into two main groups on the basis of their sequence homology and the position of the first two cysteine residues, C-x-C (alpha) and C-C (beta)10. There are six known CxC chemokine receptors and ten known functional CC chemokine receptors. At present, 16 CxC ligands (CxCL1–16) and 28 CC ligands (CCL1–28) have been identified (TABLE 1). Many of these factors originally had multiple names, but a recent consensus within the field has standardized their nomenclature10,11 (TABLE 2). The confusion related to chemokine biology stems mainly from the promiscuous binding of a single chemokine to multiple receptors, although individual receptors can bind multiple chemokines. Interestingly, evidence has begun to accumulate that chemokines can have diverse functions during an inflammatory/immune response, which relate to cellular recruitment, activation and differentiation. The pleuripotent nature of some chemokines stems from the complex and multi-directional G-protein signalling pathways that can occur after receptor coupling12. Identifying the function of each chemokine during an asthmatic response, and the nature of the signals transduced, will be essential to effectively target a specific chemokine or receptor.

The expression of chemokines during asthmatic disease has been well established13. The expression of distinct chemokines within the airway has led to the realization that there might be specific profiles of chemokines that mediate various stages of asthmatic disease. Cellular sources of chemokines within the asthmatic airway include airway epithelial cells and alveolar macrophages. Given that these cells are resident, the chemokines they release have immediate effects on the environment of the airway and surrounding lung tissue. As eosinophils are implicated in the pathophysiology of asthma, research has focused primarily on chemokines that have chemotactic activity for eosinophils. A number of chemokines, such as RANTES (regulated on activation, normal T-cell expressed and secreted)/CCL5, MCP-3 (monocyte chemoattractant protein 3)/CCL7 and MCP-4/CCL13, have been identified in the airways of asthmatics. These induce eosinophil recruitment through the CCR3 receptor, which is highly expressed on eosinophils14,15. In addition, two other CCR3 ligands, eotaxin-2/CCL24 and eotaxin-3/CCL26, have also been identified, but their expression in asthmatics has not been reported. Whether all of these chemokines function as eosinophil chemoattractants or whether a specific chemokine mediates the bulk of the chemotactic activity is unknown. In human asthma, CCL5, eotaxin/CCL11 and CCL13 are produced at high levels within the airway epithelium15–19. This concentrated




REVIEWS

Table 1 | Chemokines and their receptors Chemokine receptor

Ligands (CCL or CxCL)

Cellular expression of receptor

CC family CCR1 CCR2 CCR3 CCR4 CCR5 CCR6 CCR7 CCR8 CCR9 CCR10

CCL3, 7, 9, 15, 16, 23 CCL2, 7, 8, 12, 13 CCL5, 7, 8, 11, 13, 24, 26 CCL17, 22 CCL3, 4, 5 CCL20 CL19, 21 CCL1, 16, 17 CCL25 CCL27, 28

B, DC, E, M, N, NK, MC, T BA, DC, EC, F, M, MC, NK, N, SM, T BA, E, EP, T BA, DC, M, T DC, M, NK, T B, DC, N, NK, T B, NK, T N, NK, T T B, M, T

CxC family CxCR1 CxCR2 CxCR3 CxCR4 CxCR5 CxCR6

CxCL1, 7, 8 CxCL1, 2, 3, 5, 6, 7, 8 CxCL9, 10, 11 CxCL12 CxCL13 CxCL16

E, EC, MC, N, NK E, EC, MC, N E, EC, NK, T DC, NK, M, T B, M, T T

results highlight the importance of IL-5 in eosinophil recruitment and activation28,29. Our understanding of leukocyte recruitment and activation in a compartmentalized tissue, such as the lung, remains rudimentary. As outlined below, there are likely to be several chemokines involved in the detrimental processes in asthma, which might differ depending on the characteristic of the specific disease process in a particular patient. The identification of appropriate chemokine and/or receptor targets will be best evaluated using a combined approach of identifying levels in human specimens and testing these hypotheses in appropriate animal models. Eosinophil accumulation and chemokines

B, B cell; BA, basophil; DC, dendritic cell; E, eosinophil; EC, endothelial cell; EP, epithelial cell; F, fibroblast; M, monocyte/macrophage; MC, mast cell; N, neutrophil; NK, natural killer cell; SM, smooth muscle; T, T cell.

expression might preferentially target eosinophils to the airway. DEGRANULATION of these cells leads to the release of epithelium-damaging proteins. In addition to recruiting and activating eosinophils, these same chemokines can affect other asthma-related leukocyte populations, such as basophils and T helper (TH)2-type lymphocytes20–24. Other chemokines and their receptors might also be important. MCP-1/CCL2 was one of the first chemokines to be identified by immunohistochemical staining of airway epithelial cells in asthmatics25,26. Recent studies27 have found that the development of status asthmaticus was associated with significantly higher levels of CCL2, MIP-1α/ (macrophage inflammatory protein-1α) CCL3 and CCL5 in the bronchoalveolar lavage (BAL) fluid, along with increased interleukin-5 (IL-5), when compared with samples from patients with milder forms of asthma. These latter

Perhaps the most characteristic phenotype found in chronic asthmatics has been the accumulation of eosinophils in and around the airways. These cells have the ability, when properly activated, to degranulate and release a number of tissue-damaging products that have been shown to exacerbate the asthmatic condition30. Chemokines include CCL5, CCL7, CCL11 and CCL13. Early studies showed that CCL5 was a potent eosinophil chemoattractant, and when injected in vivo could elicit an eosinophil-rich exudate31,32. However, of these chemokines, only CCL11 binds specifically to CCR3 and is the most potent chemokine for movement of these cells33–38. CCL11 was first identified in the BAL fluid of allergic guinea-pigs39, and has subsequently been identified in humans and mice40,41. However, studies examining asthma-type responses in CCL11–/– mice showed no significant defect in eosinophil accumulation42,43, indicating that CCL11 is not essential for eosinophil accumulation and activation. The CCR3 receptor seems to be involved in the activation and degranulation of eosinophils, as well as with the primary migration of the cells, as a number of CCR3 ligands have been reported to induce degranulation of eosinophils. A recent study44, however, indicates that CCL11 primarily induces eosinophil degranulation

Table 2 | Standardized nomenclature for chemokines and their receptors New nomenclature

DEGRANULATION

Release of mediators, proteases and so on from stored granules in the cytoplasm of mast cells, neutrophils, eosinophils or other leukocyte populations.

110

Abbreviation

Common (full) name

Receptors

CCL1

TCA3

T-cell activation-3

CCR8

CCL2

MCP-1

Monocyte chemotactic protein-1

CCR2

CCL3

MIP-1α

Macrophage inflammatory protein-1α

CCR1,CCR5

CCL5

RANTES

Regulated on activation, normal T cells expressed and secreted

CCR1, CCR3, CCR5

CCL7

MCP-3


CCR1, CCR2, CCR3

CCL11

–

Eotaxin

CCR3

CCL13

MCP-4


CCR2, CCR3

CCL17

TARC

Thymus and activation-regulated chemokine

CCR4, CCR8(?)

CCL22

MDC

Macrophage-derived chemoattractant

CCR4

CCL24

–

Eotaxin-2

CCR3

CCL26

–

Eotaxin-3

CCR3

CxCL8

IL-8

Interleukin-8

CxCR1, CxCR2

CxCL12

SDF-1

Stromal-derived factor-1

CxCR4




REVIEWS

Vessel

Interstitium

Airway epithelium

Macrophage CCL3 Mucus

CCL22 CCL7

CCL5

CCL11

Degranulation

Smooth muscle Endothelium

Figure 2 | Chemokine-induced migration of eosinophils in lungs. Eosinophil migration from the vascular supply to the airway is probably dependent on sequential chemokine gradients, which allow the cells to extravasate through the various compartments of the lung. There are several eosinophil-active chemokines that are produced primarily by activated macrophages in the interstitium. These include macrophage inflammatory protein-1α (MIP-1α)/CCL3, monocyte chemoattractant protein 3 (MCP-3)/CCL7 and macrophage-derived chemokine (MDC)/CCL22. Other chemokines, such as eotaxin/CCL11 and RANTES (regulated on activation, normal T-cell expressed and secreted)/CCL5, are highly expressed in epithelial cells of the airway. The localized responses of these particular chemokines would allow the eosinophils to migrate to the airway using different chemokine receptors, therefore avoiding desensitization of migratory responses.

AIRWAY HYPERREACTIVITY

Physiologic measurement of the changes in airway resistance induced by a pharmacologic stimuli, such as methacholine. OVA-INDUCED EOSINOPHILIA

Ovalbumin-induced animal model of allergic airway inflammation.

and leukotriene C4 (LTC4) release, whereas other chemokines, including CCL2, CCL5, CCL7, CxCL8 and stromal cell-derived factor-1 (SDF-1)/CxCL12, can induce receptor-mediated calcium flux and chemotaxis. Therefore, even though CCL5 and CCL7 can bind to CCR3, they might be comparably poor inducers of degranulation, indicating that ligation of a specific receptor might yield differential activation. The degranulation of eosinophils by chemokines seems to be dependent on the extracellular-signal-regulated kinase (ERK) activation pathways45,46, indicating the activation potential of chemokine receptor signalling pathways. Additional studies have shown that other nonCCR3-binding chemokines can cause the migration and activation of eosinophils. Early studies using peripheral human eosinophils indicated that CCL3, a CCR1/CCR5 ligand, could also induce chemotaxis of eosinophils47. A role for CCL3 as an eosinophil recruitment factor both in vitro and in vivo has also been shown in mouse systems48,49. In addition, inhibition of CCL5 during an allergic airway response can greatly inhibit peribronchial eosinophil accumulation48,50,51. Both CCL3 and CCL5 bind to similar receptors and might, therefore, function through analogous mechanisms. The activation of other CCR3-independent eosinophil-recruitment pathways has also been described52. A recent study53 has shown that macrophage-derived chemokine (MDC)/CCL22

mediates peripheral human blood eosinophil recruitment through a CCR3 and CCR4-independent mechanism. CCL22 was originally described as binding specifically to CCR4. Other studies have shown that neutralization of CCL22 significantly inhibited eosinophil accumulation and AIRWAY HYPERREACTIVITY in an animal model of asthma54. Interestingly, CCR4 –/– mice have no defect in OVA-INDUCED EOSINOPHILIA55. So, CCL22 might bind to another receptor. To understand better the relationship and function of the different chemokines during eosinophil migration, the expression pattern and tissue localization needs to be examined. The principal cell types that produce the various eosinophil-active chemokines are distinct. The transmigrating eosinophils must emigrate from the vascular supply through the interstitial compartment and finally through and into the airways. This movement of cells must make use of at least two different chemotactic gradients, which probably involve separate chemokine receptors. The source of the chemokines can differ significantly and might determine at which step in the process each is involved. For example, CCL3, CCL7 and CCL22 are highly expressed in macrophages, whereas CCL5, CCL11 and CCL13 seem to be more highly expressed by airway epithelial cells56,57. Therefore, it might be possible to predict how interstitial macrophage-derived chemokines might provide the chemotactic gradient for eosinophils to move from




REVIEWS

CCL1 CCR8

CCL11

CCR3 CCR4 CCL22

IL-4 IL-13

STAT6 Pathways

Lymphocyte recruitment

Eosinophil recruitment TH2 lymphocytes Mucus production IL-5 peribronchial thickening fibrosis

Eosinophil recruitment

Eosinophilia

Figure 3 | TH2 cytokine-induced responses. The T helper (TH) type 2 cytokines interleukin (IL)-4 and IL-13 induce the production of specific chemokines through STAT6 signalactivated pathways. This group of chemokines, which includes macrophage-derived chemokine (MDC)/CCL22, T-cell activation 3 (TCA3)/CCL1 and eotaxin/CCL11, induces migration of TH2 lymphocytes through specific receptors (CCR4, CCR8 and CCR3, respectively). Some of these chemokines, for example, eotaxin, have also been shown to be potent eosinophil chemoattractants. The continued activation of these cell populations promotes the chronic pathophysiological dysfunction observed during asthma, including mucous production, peribronchial thickening and fibrosis. These pathways might be initiated and maintained through chronic allergen exposure.

the vascular compartment into the interstitium, whereas the airway-epithelium-derived chemokines might provide a second gradient that localizes the cells around and into the airways (FIG. 2). This would also fit with the fact that eosinophils do not degranulate until they reach the final destination of the airway, the site of the most potent degranulator, CCL11. This issue has been specifically addressed using an animal model of asthma and multiple allergen challenges. The results showed that initial eosinophil localization to the lung was dependent on CCL3, whereas secondary recruitment and degranulation was dependent on CCL11, but not CCL3 (REF. 58). These concepts, which are especially relevant in the highly compartmentalized lung, could be applied to all leukocytes as they move from one tissue compartment into another during disease progression. Lymphocyte responses and chemokines

As previously described, the airway inflammation process that occurs in the development of asthma is characterized by tissue infiltration of eosinophils. In addition, antigen-specific lymphocytes that are activated in the lymph node are recruited into the inflamed

112

lung. These CD4+ TH cells produce an array of TH2 cytokines, including IL-4, IL-5 and IL-13. These TH2mediated cell responses result in the production of a characteristic subset of chemokines, including CCL11, CCL13, CCL22, T-cell activation 3 (TCA3)/CCL1 and thymus and activation-regulated chemokine (TARC)/CCL17 (REFS 54,59–62; FIG. 3). In a recent study63, the regulation of chemokine production from TH1 or TH2 lymphocyte populations indicated that there was a differential chemokine response that depended on regulation by the STAT6 (signal transducers and activators of transcription 6)-mediated pathways. These studies confirmed and extended earlier results indicating that the TH1-type cells produce CCL5, whereas TH2-type cells produce CCL1 and CCL22. Because STAT6 is activated primarily by the TH2-type cytokines, IL-4 and IL-13, it is clear that the chemokine phenotype from T cells during an allergic response can be controlled almost exclusively by the level of expression of the TH2-type cytokines. A more recent study64, however, has indicated that transfer of Stat6-positive lymphocytes into a Stat6 –/– mouse was not sufficient to increase chemokine production or reconstitute the asthma-like physiological responses. These data indicated that local structural cells might be the significant source of chemokines in an allergic lung. This latter concept is supported by previous studies that show that STAT6 is also active in structural cells and can control chemokine production65–67. In vitro analysis of T-lymphocyte subsets, TH1 versus TH2, has also indicated a dichotomy of chemokine receptor expression. The expression of chemokine receptors on various subsets of naive and activated T lymphocytes indicates a progression toward distinct subsets10,68,69 (FIG. 4). Naive T lymphocytes express a number of chemokine receptors, including CCR7 and CxCR4, which are subsequently lost after initial activation of the lymphocyte21,70–72. These receptors seem to be required for entry into secondary lymph nodes through the high endothelial venule, which express the specific ligands CCL21 and CxCL12 (REFS 73–75). Subsequently, lymphocytes begin to alter their chemokine receptor expression on the basis of the immune environment that develops, TH1 or TH2. Lymphocytes that are skewed toward TH1-type responses (IL-12 and interferon-γ) differentially express CxCR3 and CCR5, whereas those skewed toward TH2-type cytokine production (IL-4, IL-5 and IL-13) differentially express CCR3, CCR4, CxCR4 and CCR8 (REFS 21,72,76,77). However, these assumptions, taken from in vitro analyses, should be used only as a guide, as it is unclear how stable these phenotypes are in vivo 78. The receptors expressed on TH2-skewed lymphocytes correspond with the expression of chemokine ligands that have been implicated in allergic asthma responses. Specifically, the expression of CCL11, CCL13 and CCL7 can be identified in both human asthma and in animal models, and might be involved in the recruitment of TH2-type cells through CCR3. Studies have also shown that the CCR4 ligands, CCL22 and CCL17, are both involved in the allergic responses. In an exquisite study performed in a mouse model of asthma, the specific recruitment of TH2-type




REVIEWS

Naive

CCR7 CxCR4

TH cell

Ag-activation

CCR7 CCR5

TH0

CxCR3

IL-12

IL-4 Ag-activation CCR3

CCR5 TH1

CxCR3

CCR4 TH2

CCR8 CxCR4

Figure 4 | Differential expression of CCRs on TH subsets. The expression of chemokine receptors on T helper (TH) lymphocytes seems to be related to the differentiation state of cells. Naive cells express CCR7 and CxCR4, two receptors that participate in entry of the lymph node through the high endothelial venule (HEV). Short-term or newly activated cells seem to maintain CCR7 expression, but also express CCR5 and CxCR3. Continued antigenic (Ag) stimulation in vitro in the presence of TH type 1- (IL-12) or TH type 2- (IL-4) cytokines significantly alters the chemokine receptor patterns. CCR7, the HEV-related receptor, is downregulated and the differential expression of CCR5 and CxCR3 on TH1-type cells and CCR3, CCR4, CCR8 and CxCR4 on TH 2-type cells are induced. Although the in vivo relevance of these preferential patterns is unknown, these in vitro analyses allow a better understanding of chemokine receptor regulation in specific cytokine environments.

cells seemed to depend initially on CCR3 ligands and subsequently on CCR4 ligands79. In addition, a recent publication has begun to identify specific receptor expression on CD4+ T lymphocytes in the airways of asthmatics. In this study, which included a total of 24 patients, virtually all IL-4 (TH2-type)-expressing lymphocytes were CCR4+ in endobronchial biopsies taken 24 hours after allergen challenge, whereas only 28% were CCR8+ (REF. 80). These data support the concept that certain chemokine receptors might be indicative of specific lymphocyte responses. The nature of chemokine production, as well as the location and cell types producing the chemokines, will determine the receptors used by lymphocytes migrating into the inflammed lung. Chemokines and structural cell activation

As discussed above, the structural cells of the lung might be a significant source of a number chemokine mediators, allowing localization of specific leukocyte populations into and around the airway. Studies have clearly shown that certain chemokines are induced during specific types of immune response16,18,59,81–85. Recent evidence indicates that the structural cells themselves can

express chemokine receptors and might be specifically activated during the asthmatic response86–93. In particular, airway epithelial cells from asthmatics and patients with other respiratory diseases can express functional surface CCR3 (REFS 91,93). It is not clear what purpose these receptors have during an inflammatory response. However, because CCR3 ligands were able to induce a Ca2+ flux and phosphorylation responses in these cells, the event might have a significant impact on disease progression and severity. This is especially intriguing because the same cells produce significant levels of the CCR3 ligands, CCL5 and CCL11, within the asthmatic lung. Future experiments will probably yield interesting data on the relationship between chemokine receptor expression, cell survival, epithelial cell activation and disease pathogenesis. Other structural cells that might have an impact on asthmatic disease progression also express chemokine receptors. Initial data indicated that certain chemokines had the ability to induce proliferation of endothelial and smooth muscle cells94–96. These initial findings have now been extended to show that both stromal and vascular cell populations specifically express a number of chemokines receptors. Chemokines have clearly been identified in the regulation of angiogenesis97,98, but more central to asthmatic conditions, other studies have shown that certain chemokines can have an immediate impact on airway function. In particular, CCL1 was shown to directly induce smooth-muscle cell activation, indicating that CCR8 ligation might have a direct role in inducing airway reactivity95. This becomes even more important as the peribronchial smooth-muscle cell hypertrophy continues to expand in chronic asthmatic disease. The impact of chemokine receptor expression on smooth muscle cells is unclear, but considering the pro-proliferative role that chemokines have on these and other structural cell populations, they might alter a potential disease-contributing population. One of the recently developed concepts in asthmatic disease is that there is significant contribution of airway remodelling in the progression and intensity of chronic airways diseases99. Studies of pulmonary fibroblast populations from various disease states clearly indicate a difference in the chemokine profile in fibroblasts taken from non-fibrotic versus fibrotic lesions. In particular, a clear difference can be found in the level of expression of CCL2 and its receptor CCR2 (REF. 83). Initial studies on fibroblast biology showed that CCL2 was a competence factor for fibroblast activation100. Subsequent studies have shown that CCL2 might have a causal role in the development of fibrotic disease, especially by directly inducing transforming growth factor-β (TGF-β) production, which leads to collagen expression101. In animal models of fibrosis, neutralization of CCL2 significantly reduces the development of fibrosis and can alter end-stage disease102,103. The exact role of chemokines in the fibrotic process is not completely clear. The chemokines might simply direct the inflammatory response leading to end-stage disease or, as suggested above, directly contribute to fibroblast proliferation, activation and collagen deposition.




REVIEWS

Table 3 | Present prospective chemokine targets in asthma Target

Described function and criteria for targeting

CCL2

Basophil/mast-cell activation T-cell chemoattractant Macrophage activator Skews T cells toward TH2 Highly expressed in asthmatics Directly involved in airway hyperreactivity Associated with fibrotic responses

CCL11

Major eosinpohil recruiter and activator Highly expressed during allergic inflammation Basophil recruitment factor

CCL22

Induces eosinophil recruitment in vitro Neutralization blocks airway changes in animals

CCR3

Highly expressed on eosinophils Expressed on basophils Differentially expressed on TH2 cells Expressed on epithelium of asthmatics

CCR8

Expressed on TH2 but not TH1 cells CCR8 –/– mice have decreased TH2 responses Direct activation of airway contraction by ligand, CCL1

CxCR4

Expressed on naive and TH2 type cells Blocking results in decreased allergen-induced TH2 cytokines

Chemokine receptor targets for treatment

The identification of particular profiles of chemokines and their receptors during asthmatic inflammation is an important aspect of current research into this complex disease. However, as discussed above, the promiscuous nature of chemokine binding and the diverse functions of chemokines during an inflammatory response means that choosing the correct chemokine or receptor to target remains problematic. The issue of promiscuity of chemokine binding to receptors has led to considerable concerns about whether to target the receptor or a specific chemokine during disease. In addition, asthma is clearly a multifactorial disease and it is unlikely that a single target will alleviate most symptoms for a large proportion of patients. At present, the most successful therapy has been the compliant use of steroids that nonspecifically block the inflammatory component of asthma9,104. Complementary therapies have also begun to target individual components of the response, such as leukotrienes105. The development of therapies to specifically target inflammatory mechanisms will allow the use of high-dose steroids to be avoided. Multiple chemokine targets have been identified in animal models (TABLE 3). Optimistically, we might expect the choice of the correct chemokine pathways to alter several aspects of the disease, including both acute and chronic phenotypes. Initially, the choices seemed to be fairly straightforward: abrogate the main eosinophil-specific chemokine pathways (that is, CCL11/CCR3) and the disease will be attenuated. With the discovery of multiple CCR3 ligands and the results of studies that found little effect by neutralizing or deleting CCL11, it became clear that merely blocking the CCL11/CCR3 interaction might not be effective. In addition, multiple studies in animal models have indicated that physiological changes in the lungs do not necessarily correlate with eosinophil accumulation, but rather correlate with TH2-lymphocyte responses that are required and/or sufficient for the

114

development of airway hyperreactivity. A more reasonable strategy, therefore, might be to alter the entire lung immune environment by blocking the recruitment of TH2-type lymphocytes that express distinct patterns of chemokine receptors. Although studies are limited, there are some promising results both in vitro and in vivo indicating that this latter strategy might be successful. Recent studies have already indicated that by either blocking ligands for particular receptors (CCL22 and CxCL12) or using chemokine receptor doublenegative mice (CCR8), the lung immune environment can be altered, reducing the pro-asthmatic TH2 cytokines54,106,107. The ligand for CCR8 — CCL1 — has yet to be described in asthma or allergic inflammation, but CCR8 is found exclusively on TH2-type cells63,72,76,108. A recent study using CCR8–/– mice indicated attenuation of several asthma-related features106. Although the development of a peripheral TH2-type response was normal in the CCR8–/– mice, there was an alteration in the response after a local lung allergen challenge. The changes in the pulmonary immune response centred on reduced production of TH2-type cytokines, IL-4, IL-5 and IL-13, and a significant reduction in eosinophil accumulation. Although the conclusions were not definitive, it seemed that the CCR8–/– mice have a defect in their ability to properly recruit TH2-type lymphocytes to the lung on allergen provocation. This might define CCR8 as an excellent target for altering the long-term consequences of chronic TH2-type cytokine production and the associated eosinophil recruitment. Another receptor that seems to have a single, defined ligand and is also preferentially associated with TH2-type but not TH1-skewed lymphocytes is CxCR4 (REF. 77). The neutralization of its only known ligand, CxCL12, showed a significant alteration of the allergic lung environment, including the attenuation of airway hyperreactivity and decreased airway eosinophilia107. Studies in our laboratory have extended these latter studies by targeting CxCR4 and showing a reduction in TH2 cytokines during asthma-type responses in mice (Lukacs et al., unpublished observations). These lymphocyte-targeting strategies might be the most beneficial in the long run for altering the phenotype of chronic allergic asthmatic responses that rely on the local activation of TH2-type cytokines. An issue that has confronted researchers is whether to target a particular chemokine or a relevant receptor. The dilemma in this area is exemplified by research in mice, in which CCL2 and its primary receptor CCR2 have been investigated. Data from multiple groups have identified CCL2 as a key chemokine that is directly involved in the progression of airway hyperreactivity in asthmatics25–27,51,102,109,110. The potential mechanism of action of CCL2 is related to several aspects of the allergen-induced responses, including recruitment of T lymphocytes, activation of basophil/mast cell populations and generation of arachidonic mediators, but has little to do with the accumulation of eosinophils in and around the airway. Interestingly, in mice deleted of CCR2, the only known CCL2 receptor, there was no attenuation in the airway hyperreactive responses, but rather increases in airway remodelling in chronic models




REVIEWS of allergic disease111,112. This issue has recently been reviewed113 and is probably related to one of three mechanisms: the differential use of CCR2 by multiple ligands, the existence of a second CCL2 receptor, or the unforeseen role of CCR2 in regulating chronic immune responses. Because of complicated compensatory pathways, studies using receptor-deficient mice should be supported by data from other systems. Overall, these data might indicate at least one instance when targeting a specific chemokine is more beneficial than targeting the chemokine receptor (which, in this case, has multiple ligands). With these and other studies, it is becoming clear that altering asthmatic responses by blocking specific chemokine pathways is not as straightforward as was originally hoped.

1. 2. 3.

4. 5. 6.

7. 8. 9.

10. 11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Hartert, T. V. & Peebles, R. S. Epidemiology of asthma: the year in review. Curr. Opin. Pulm. Med. 6, 4–9 (2000). Marwick, C. Helping city children control asthma. J. Am. Med. Assoc. 277, 1503–1504 (1997). Schuh, S., Johnson, D., Stephens, D., Callahan, S. & Canny, G. Hospitalization patterns in severe acute asthma in children. Pediatr. Pulmonol. 23, 184–192 (1997). Kon, O. M. & Kay, A. B. T cells and chronic asthma. Int. Arch. Allergy Immunol. 118, 133–135 (1999). Kay, A. B. Pathology of mild, severe, and fatal asthma. Am. J. Respir. Crit. Care Med. 154, S66–S69 (1996). Mapp, C., Boschetto, P., Miotto, D., De Rosa, E. & Fabbri, L. M. Mechanisms of occupational asthma. Annu. Allergy Asthma Immunol. 83, 645–664 (1999). Bentley, A. M., Kay, A. B. & Durham, S. R. Human late asthmatic reactions. Clin. Exp. Allergy 27, 71–86 (1997). Muro, S., Minshall, E. M. & Hamid, Q. A. The pathology of chronic asthma. Clin. Chest Med. 21, 225–244 (2000). Busse, W. W., Banks-Schlegel, S. & Wenzel, S. E. Pathophysiology of severe asthma. J. Allergy Clin. Immunol. 106, 1033–1042 (2000). Rossi, D. & Zlotnik, A. The biology of chemokines and their receptors. Annu. Rev. Immunol. 18, 217–242 (2000). Zlotnik, A. & Yoshie, O. Chemokines: a new classification system and their role in immunity. Immunity 12, 121–127 (2000). Outlines the new nomenclature for the chemokines, which clarifies this area of confusion. The table and charts clearly identify the order and specific receptor interactions. Sotsios, Y. & Ward, S. G. Phosphoinositide 3-kinase: a key biochemical signal for cell migration in response to chemokines. Immunol. Rev. 177, 217–235 (2000). Lukacs, N. W., Oliveira, S. H. & Hogaboam, C. M. Chemokines and asthma: redundancy of function or a coordinated effort? J. Clin. Invest. 104, 995–999 (1999). Powell, N. et al. Increased expression of mRNA encoding RANTES and MCP-3 in the bronchial mucosa in atopic asthma. Eur. Respir. J. 9, 2454–2460 (1996). Stellato, C. et al. Production of the novel C-C chemokine MCP-4 by airway cells and comparison of its biological activity to other C-C chemokines. J. Clin. Invest. 99, 926–936 (1997). Lamkhioued, B. et al. Monocyte chemoattractant protein (MCP)-4 expression in the airways of patients with asthma. Induction in epithelial cells and mononuclear cells by proinflammatory cytokines. Am. J. Respir. Crit. Care Med. 162, 723–732 (2000). Taha, R. A. et al. Eotaxin and monocyte chemotactic protein-4 mRNA expression in small airways of asthmatic and nonasthmatic individuals. J. Allergy Clin. Immunol. 103, 476–483 (1999). Berkman, N. et al. Expression of RANTES mRNA and protein in airways of patients with mild asthma. Am. J. Respir. Crit. Care Med. 154, 1804–1811 (1996). Wang, J. H., Devalia, J. L., Xia, C., Sapsford, R. J. & Davies, R. J. Expression of RANTES by human bronchial epithelial cells in vitro and in vivo and the effect of corticosteroids. Am. J. Respir. Cell. Mol. Biol. 14, 27–35 (1996). Sallusto, F., Lanzavecchia, A. & Mackay, C. R. Chemokines and chemokine receptors in T-cell priming and TH1/TH2-mediated responses. Immunol. Today 19, 568–574 (1998). Sallusto, F., Lenig, D., Mackay, C. R. & Lanzavecchia, A. Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes. J. Exp. Med. 187, 875–883 (1998).

Conclusion

The role of chemokines in asthmatic disease progression is evident at multiple levels, including differential leukocyte recruitment and local cellular activation. In view of the diversity of chemokine production and the promiscuous binding pattern of multiple receptors, it might be difficult to identify a specific target. Only after determining the function of specific chemokine and chemokine receptors during disease progression can specific molecules be targeted. Future insight into the function of chemokines will be best achieved by considering the role of chemokines in immune and non-immune cells, and by coordinating results from animal models with the responses observed in patient populations.

22. Bonecchi, R. et al. Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (TH1s) and TH2s. J. Exp. Med. 187, 129–134 (1998). References 21 and 22 are the initial studies that identified the preferential expression of specific chemokine receptors on polarized T lymphocytes. They outline the basic principle of preferential expression on the basis of a specific cytokine environment. 23. Sallusto, F., Mackay, C. R. & Lanzavecchia, A. Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Science 277, 2005–2007 (1997). 24. Uguccioni, M. et al. High expression of the chemokine receptor CCR3 in human blood basophils. Role in activation by eotaxin, MCP-4, and other chemokines. J. Clin. Invest. 100, 1137–1143 (1997). 25. Alam, R. et al. Increased MCP-1, RANTES, and MIP-1α in bronchoalveolar lavage fluid of allergic asthmatic patients. Am. J. Respir. Crit. Care Med. 153, 1398–1404 (1996). 26. Sousa, A. R. et al. Increased expression of the monocyte chemoattractant protein-1 in bronchial tissue from asthmatic subjects. Am. J. Respir. Cell Mol. Biol. 10, 142–147 (1994). 27. Tillie-Leblond, I. et al. CC chemokines and interleukin-5 in bronchial lavage fluid from patients with status asthmaticus. Potential implication in eosinophil recruitment. Am. J. Respir. Crit. Care Med. 162, 586–592 (2000). 28. Collins, P. D., Marleau, S., Griffiths-Johnson, D. A., Jose, P. J. & Williams, T. J. Cooperation between interleukin-5 and the chemokine eotaxin to induce eosinophil accumulation in vivo. J. Exp. Med. 182, 1169–1174 (1995). 29. Warringa, R. A. et al. Modulation of eosinophil chemotaxis by interleukin-5. Am. J. Respir. Cell Mol. Biol. 7, 631–636 (1992). 30. Corrigan, C. J. & Kay, A. B. T cells and eosinophils in the pathogenesis of asthma. Immunol. Today 13, 501–507 (1992). 31. Kapp, A., Zeck-Kapp, G., Czech, W. & Schopf, E. The chemokine RANTES is more than a chemoattractant: characterization of its effect on human eosinophil oxidative metabolism and morphology in comparison with IL-5 and GM-CSF. J. Invest. Dermatol. 102, 906–914 (1994). 32. Meurer, R. et al. Formation of eosinophilic and monocytic intradermal inflammatory sites in the dog by injection of human RANTES but not human monocyte chemoattractant protein 1, human macrophage inflammatory protein 1α, or human interleukin 8. J. Exp. Med. 178, 1913–1921 (1993). 33. Heath, H. et al. Chemokine receptor usage by human eosinophils. The importance of CCR3 demonstrated using an antagonistic monoclonal antibody. J. Clin. Invest. 99, 178–184 (1997). 34. Gao, J. L. et al. Identification of a mouse eosinophil receptor for the CC chemokine eotaxin. Biochem. Biophys. Res. Commun. 223, 679–684 (1996). 35. Ponath, P. D. et al. Molecular cloning and characterization of a human eotaxin receptor expressed selectively on eosinophils. J. Exp. Med. 183, 2437–2448 (1996). 36. Combadiere, C., Ahuja, S. K. & Murphy, P. M. Cloning and functional expression of a human eosinophil CC chemokine receptor. J. Biol. Chem. 271, 11034 (1996). 37. Daugherty, B. L. et al. Cloning, expression, and characterization of the human eosinophil eotaxin receptor. J. Exp. Med. 183, 2349–2354 (1996).


38. Ponath, P. D. et al. Cloning of the human eosinophil chemoattractant, eotaxin. Expression, receptor binding, and functional properties suggest a mechanism for the selective recruitment of eosinophils. J. Clin. Invest. 97, 604–612 (1996). 39. Jose, P. J. et al. Eotaxin: a potent eosinophil chemoattractant cytokine detected in a guinea pig model of allergic airways inflammation. J. Exp. Med. 179, 881–887 (1994). A protein mediator that preferentially induced eosinophil accumulation in the airway was first discovered in guinea-pig airways, but the importance for human disease was quickly realized. Much of the early therapy in the chemokine field has been geared towards inhibiting the action of eotaxin with its specific receptor, CCR3. 40. Kitaura, M. et al. Molecular cloning of human eotaxin, an eosinophil-selective CC chemokine, and identification of a specific eosinophil eotaxin receptor, CC chemokine receptor 3. J. Biol. Chem. 271, 7725–7730 (1996). 41. Rothenberg, M. E., Luster, A. D. & Leder, P. Murine eotaxin: an eosinophil chemoattractant inducible in endothelial cells and in interleukin 4-induced tumor suppression. Proc. Natl Acad. Sci. USA 92, 8960–8964 (1995). 42. Yang, Y., Loy, J., Ryseck, R. P., Carrasco, D. & Bravo, R. Antigen-induced eosinophilic lung inflammation develops in mice deficient in chemokine eotaxin. Blood 92, 3912–3923 (1998). 43. Rothenberg, M. E., MacLean, J. A., Pearlman, E., Luster, A. D. & Leder, P. Targeted disruption of the chemokine eotaxin partially reduces antigen-induced tissue eosinophilia. J. Exp. Med. 185, 785–790 (1997). The finding that eotaxin-deficient mice showed virtually normal allergic airway responses allowed investigators to realize that targeting the receptor, CCR3, might be more efficacious than an individual chemokine ligand. 44. Fujisawa, T. et al. Chemokines induce eosinophil degranulation through CCR-3. J. Allergy Clin. Immunol. 106, 507–513 (2000). 45. Bates, M. E., Green, V. L. & Bertics, P. J. ERK1 and ERK2 activation by chemotactic factors in human eosinophils is interleukin 5-dependent and contributes to leukotriene C4 biosynthesis. J. Biol. Chem. 275, 10968–10975 (2000). 46. Kampen, G. T. et al. Eotaxin induces degranulation and chemotaxis of eosinophils through the activation of ERK2 and p38 mitogen-activated protein kinases. Blood 95, 1911–1917 (2000). 47. Rot, A. et al. RANTES and macrophage inflammatory protein 1α induce the migration and activation of normal human eosinophil granulocytes. J. Exp. Med. 176, 1489–1495 (1992). 48. Lukacs, N. W. et al. Differential recruitment of leukocyte populations and alteration of airway hyperreactivity by C-C family chemokines in allergic airway inflammation. J. Immunol. 158, 4398–4404 (1997). 49. Lukacs, N. W., Strieter, R. M., Shaklee, C. L., Chensue, S. W. & Kunkel, S. L. Macrophage inflammatory protein-1α influences eosinophil recruitment in antigen-specific airway inflammation. Eur. J. Immunol. 25, 245–251 (1995). 50. Lukacs, N. W. et al. C-C chemokine-induced eosinophil chemotaxis during allergic airway inflammation. J. Leukoc. Biol. 60, 573–578 (1996). 51. Gonzalo, J. A. et al. The coordinated action of CC chemokines in the lung orchestrates allergic inflammation and airway hyperresponsiveness. J. Exp. Med. 188, 157–167 (1998).



REVIEWS

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

116

One of the earliest studies showing that chemokines were not just a family of cytokines with overlapping functions, but rather induced factors that participated in a response in a coordinated manner. Multiple chemokines have a significant role in different aspects of the allergen-induced responses in the lung. Sabroe, I. et al. Differential regulation of eosinophil chemokine signaling via CCR3 and non-CCR3 pathways. J. Immunol. 162, 2946–2955 (1999). Bochner, B. S. et al. Macrophage-derived chemokine induces human eosinophil chemotaxis in a CC chemokine receptor 3- and CC chemokine receptor 4-independent manner. J. Allergy Clin. Immunol. 103, 527–532 (1999). Gonzalo, J. A. et al. Mouse monocyte-derived chemokine is involved in airway hyperreactivity and lung inflammation. J. Immunol. 163, 403–411 (1999). Chvatchko, Y. et al. A key role for CC chemokine receptor 4 in lipopolysaccharide-induced endotoxic shock. J. Exp. Med. 191, 1755–1764 (2000). Mantovani, A., Gray, P. A., Van Damme, J. & Sozzani, S. Macrophage-derived chemokine (MDC). J. Leukoc. Biol. 68, 400–404 (2000). Matsukawa, A. et al. Pivotal role of the CC chemokine, macrophage-derived chemokine, in the innate immune response. J. Immunol. 164, 5362–5368 (2000). Campbell, E. M., Kunkel, S. L., Strieter, R. M. & Lukacs, N. W. Temporal role of chemokines in a murine model of cockroach allergen-induced airway hyperreactivity and eosinophilia. J. Immunol. 161, 7047–7053 (1998). Sekiya, T. et al. Inducible expression of a TH2-type CC chemokine thymus- and activation-regulated chemokine by human bronchial epithelial cells. J. Immunol. 165, 2205–2213 (2000). Andrew, D. P. et al. STCP-1 (MDC) CC chemokine acts specifically on chronically activated TH2 lymphocytes and is produced by monocytes on stimulation with TH2 cytokines IL-4 and IL-13. J. Immunol. 161, 5027–5038 (1998). Mochizuki, M., Bartels, J., Mallet, A. I., Christophers, E. & Schroder, J. M. IL-4 induces eotaxin: a possible mechanism of selective eosinophil recruitment in helminth infection and atopy. J. Immunol. 160, 60–68 (1998). Li, L., Xia, Y., Nguyen, A., Feng, L. & Lo, D. TH2-induced eotaxin expression and eosinophilia coexist with TH1 responses at the effector stage of lung inflammation. J. Immunol. 161, 3128–3135 (1998). Zhang, S., Lukacs, N. W., Lawless, V. A., Kunkel, S. L. & Kaplan, M. H. Cutting edge: differential expression of chemokines in TH1 and TH2 cells is dependent on Stat6 but not Stat4. J. Immunol. 165, 10–14 (2000). Identifies a primary role for TH2 cytokine-induced STAT6 regulation of chemokines and their receptors on T lymphocytes, indicating that their expression might be dependent on gene regulation and not on differentiation state. Mathew, A. et al. Signal transducer and activator of transcription 6 controls chemokine production and T helper cell type 2 cell trafficking in allergic pulmonary inflammation. J. Exp. Med. 193, 1087–1096 (2001). Goebeler, M. et al. Interleukin-13 selectively induces monocyte chemoattractant protein-1 synthesis and secretion by human endothelial cells. Involvement of IL-4Rα and Stat6 phosphorylation. Immunology 91, 450–457 (1997). Matsukura, S. et al. Interleukin-13 upregulates eotaxin expression in airway epithelial cells by a Stat6-dependent mechanism. Am. J. Respir. Cell Mol. Biol. 24, 755–761 (2001). Hoeck, J. & Woisetschlager, M. STAT6 mediates eotaxin-1 expression in IL-4 or TNF-α-induced fibroblasts. J. Immunol. 166, 4507–4515 (2001). Campbell, J. D. & HayGlass, K. T. T cell chemokine receptor expression in human TH1- and TH2-associated diseases. Arch. Immunol. Ther. Exp. (Warsz.) 48, 451–456 (2000). Yoshie, O. Role of chemokines in trafficking of lymphocytes and dendritic cells. Int. J. Hematol. 72, 399–407 (2000). Sallusto, F. & Lanzavecchia, A. Understanding dendritic cell and T-lymphocyte traffic through the analysis of chemokine receptor expression. Immunol. Rev. 177, 134–140 (2000). This review provides an excellent focus on the coordination of T-lymphocyte and dendritic cell traffic from tissue to lymph node. Sallusto, F., Lenig, D., Forster, R., Lipp, M. & Lanzavecchia, A. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature 401, 708–712 (1999). Sallusto, F. et al. Switch in chemokine receptor expression upon TCR stimulation reveals novel homing potential for recently activated T cells. Eur. J. Immunol. 29, 2037–2045 (1999). Warnock, R. A. et al. The role of chemokines in the microenvironmental control of T versus B cell arrest in

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

Peyer’s patch high endothelial venules. J. Exp. Med. 191, 77–88 (2000). Tangemann, K., Gunn, M. D., Giblin, P. & Rosen, S. D. A high endothelial cell-derived chemokine induces rapid, efficient, and subset-selective arrest of rolling T lymphocytes on a reconstituted endothelial substrate. J. Immunol. 161, 6330–6337 (1998). Stein, J. V. et al. The CC chemokine thymus-derived chemotactic agent 4 (TCA-4, secondary lymphoid tissue chemokine, 6Ckine, exodus-2) triggers lymphocyte function-associated antigen 1-mediated arrest of rolling T lymphocytes in peripheral lymph node high endothelial venules. J. Exp. Med. 191, 61–76 (2000). D’Ambrosio, D. et al. Selective up-regulation of chemokine receptors CCR4 and CCR8 upon activation of polarized human type 2 TH cells. J. Immunol. 161, 5111–5115 (1998). Jagodzinski, P. P. & Trzeciak, W. H. Differential expression of chemokine receptor CXCR4 in TH1 and TH2 subtypes of CD4+ lymphocytes. Folia Histochem. Cytobiol. 38, 21–23 (2000). Nanki, T. & Lipsky, P. E. Lack of correlation between chemokine receptor and TH1/TH2 cytokine expression by individual memory T cells. Int. Immunol. 12, 1659–1667 (2000). Lloyd, C. M. et al. CC chemokine receptor (CCR)3/eotaxin is followed by CCR4/monocyte-derived chemokine in mediating pulmonary T helper lymphocyte type 2 recruitment after serial antigen challenge in vivo. J. Exp. Med. 191, 265–274 (2000). The first study to show that different chemokine receptors are used at distinct phases of disease progression to promote T-lymphocyte accumulation. Complements the earlier studies by this group that show coordinated chemokine production. Panina-Bordignon, P. et al. The C-C chemokine receptors CCR4 and CCR8 identify airway T cells of allergenchallenged atopic asthmatics. J. Clin. Invest. 107, 1357–1364 (2001). Li, L. et al. Effects of TH2 cytokines on chemokine expression in the lung: IL-13 potently induces eotaxin expression by airway epithelial cells. J. Immunol. 162, 2477–2487 (1999). Saito, T. et al. Respiratory syncytial virus induces selective production of the chemokine RANTES by upper airway epithelial cells. J. Infect. Dis. 175, 497–504 (1997). Hogaboam, C. M. et al. Differential monocyte chemoattractant protein-1 and chemokine receptor 2 expression by murine lung fibroblasts derived from TH1and TH2-type pulmonary granuloma models. J. Immunol. 163, 2193–2201 (1999). Petering, H. et al. Detection of MCP-4 in dermal fibroblasts and its activation of the respiratory burst in human eosinophils. J. Immunol. 160, 555–558 (1998). Ying, S. et al. Eosinophil chemotactic chemokines (eotaxin, eotaxin-2, RANTES, monocyte chemoattractant protein-3 (MCP-3), and MCP-4), and C-C chemokine receptor 3 expression in bronchial biopsies from atopic and nonatopic (Intrinsic) asthmatics. J. Immunol. 163, 6321–6329 (1999). Gupta, S. K., Lysko, P. G., Pillarisetti, K., Ohlstein, E. & Stadel, J. M. Chemokine receptors in human endothelial cells. Functional expression of CXCR4 and its transcriptional regulation by inflammatory cytokines. J. Biol. Chem. 273, 4282–4287 (1998). Hayes, I. M. et al. Human vascular smooth muscle cells express receptors for CC chemokines. Arterioscler. Thromb. Vasc. Biol. 18, 397–403 (1998). Crane, I. J., Wallace, C. A., McKillop-Smith, S. & Forrester, J. V. CXCR4 receptor expression on human retinal pigment epithelial cells from the blood-retina barrier leads to chemokine secretion and migration in response to stromal cell-derived factor 1α. J. Immunol. 165, 4372–4378 (2000). Jordan, N. J. et al. Expression of functional CXCR4 chemokine receptors on human colonic epithelial cells. J. Clin. Invest. 104, 1061–1069 (1999). Dwinell, M. B., Eckmann, L., Leopard, J. D., Varki, N. M. & Kagnoff, M. F. Chemokine receptor expression by human intestinal epithelial cells. Gastroenterology 117, 359–367 (1999). Stellato, C. et al. Expression of the C-C chemokine receptor CCR3 in human airway epithelial cells. J. Immunol. 166, 1457–1461 (2001). One of the first studies to show that the chemokine receptor CCR3 can be expressed on airway epithelial cells during asthmatic responses. Interestingly, these same cells produce an abundant amount of CCR3-specific ligands during asthmatic responses. The consequences and mechanisms associated with this finding will probably have a significant impact on airway biology and physiology during disease. Zhang, J., Lathbury, L. J. & Salamonsen, L. A. Expression of the chemokine eotaxin and its receptor, CCR3, in human endometrium. Biol. Reprod. 62, 404–411 (2000).


93. Oyamada, H. et al. CCR3 mRNA expression in bronchial epithelial cells and various cells in allergic inflammation. Int. Arch. Allergy Immunol. 120 (Suppl. 1), 45–47 (1999). 94. Porreca, E. et al. Monocyte chemotactic protein 1 (MCP-1) is a mitogen for cultured rat vascular smooth muscle cells. J. Vasc. Res. 34, 58–65 (1997). 95. Luo, Y., D’Amore, P. A. & Dorf, M. E. β-chemokine TCA3 binds to and activates rat vascular smooth muscle cells. J. Immunol. 157, 2143–2148 (1996). 96. Yue, T. L. et al. Interleukin-8. A mitogen and chemoattractant for vascular smooth muscle cells. Circ. Res. 75, 1–7 (1994). 97. Belperio, J. A. et al. CXC chemokines in angiogenesis. J. Leukoc. Biol. 68, 1–8 (2000). 98. Rossi, D. & Zlotnik, A. The biology of chemokines and their receptors. Annu. Rev. Immunol. 18, 217–242 (2000). 99. Bento, A. M. & Hershenson, M. B. Airway remodeling: potential contributions of subepithelial fibrosis and airway smooth muscle hypertrophy/hyperplasia to airway narrowing in asthma. Allergy Asthma Proc. 19, 353–358. (1998). 100. Rollins, B. J. JE/MCP-1: an early-response gene encodes a monocyte-specific cytokine. Cancer Cells 3, 517–524 (1991). 101. Gharaee-Kermani, M., Denholm, E. M. & Phan, S. H. Costimulation of fibroblast collagen and transforming growth factor β1 gene expression by monocyte chemoattractant protein-1 via specific receptors. J. Biol. Chem. 271, 17779–17784 (1996). 102. Blease, K. et al. Antifungal and airway remodeling roles for murine monocyte chemoattractant protein-1/CCL2 during pulmonary exposure to Asperigillus fumigatus conidia. J. Immunol. 166, 1832–1842 (2001). 103. Lloyd, C. M. et al. RANTES and monocyte chemoattractant protein-1 (MCP-1) play an important role in the inflammatory phase of crescentic nephritis, but only MCP-1 is involved in crescent formation and interstitial fibrosis. J. Exp. Med. 185, 1371–1380 (1997). 104. Redington, A. E. Fibrosis and airway remodelling. Clin. Exp. Allergy 30 (Suppl. 1), 42–45 (2000). 105. Donohue, J. F. & Ohar, J. A. New combination therapies for asthma. Curr. Opin. Pulm. Med. 7, 62–68 (2001). 106. Chensue, S. W. et al. Aberrant in vivo T helper type 2 cell response and impaired eosinophil recruitment in CC chemokine receptor 8 knockout mice. J. Exp. Med. 193, 573–584 (2001). The first study to indicate that chemokine receptors can be targeted for therapy to alter the TH2-specific asthmatic responses. The preferential expression of CCR8 on TH2 cells and reduced TH2 cytokine phenotype within the lung of CCR8-deficient mice is a vital clue for therapeutic targeting of a specific receptor. 107. Gonzalo, J. A. et al. Critical involvement of the chemotactic axis CXCR4/stromal cell-derived factor-1α in the inflammatory component of allergic airway disease. J. Immunol. 165, 499–508 (2000). 108. Zingoni, A. et al. The chemokine receptor CCR8 is preferentially expressed in TH2 but not TH1 cells. J. Immunol. 161, 547–551 (1998). 109. Jahnz-Rozyk, K. M., Kuna, P. & Pirozynska, E. Monocyte chemotactic and activating factor/monocyte chemoattractant protein (MCAF/MCP-1) in bronchoalveolar lavage fluid from patients with atopic asthma and chronic bronchitis. J. Investing. Allergol. Clin. Immunol. 7, 254–259 (1997). 110. Campbell, E. M. et al. Monocyte chemoattractant protein-1 mediates cockroach allergen-induced bronchial hyperreactivity in normal but not CCR2–/– mice: the role of mast cells. J. Immunol. 163, 2160–2167 (1999). 111. MacLean, J. A. et al. CC chemokine receptor-2 is not essential for the development of antigen-induced pulmonary eosinophilia and airway hyperresponsiveness. J. Immunol. 165, 6568–6575 (2000). 112. Blease, K. et al. Enhanced pulmonary allergic responses to Aspergillus in CCR2–/– mice. J. Immunol. 165, 2603–2611 (2000). 113. Luther, S. A. & Cyster, J. G. Chemokines as regulators of T cell differentiation. Nature Immunol. 2, 102–107 (2001).

Acknowledgements I thank R. Kunkel for design of the original figures.

Online links DATABASES The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ CCL1 | CCL2 | CCL3 | CCL5 | CCL7 | CCL11 | CCL13 | CCL17 | CCL21 | CCL22 | CCL24 | CCL26 | CCR1 | CCR2 | CCR3 | CCR4 | CCR5 | CCR7 | CCR8 | CxCL1 | CxCL8 |CxCR4 | CxCL12 | ERK | IL-4 | IL-5 | IL-12 | IL-13 | STAT6 | TGF-β Access to this interactive links box is free online.



REVIEWS

TOWARDS A BLOOD-STAGE VACCINE FOR MALARIA: ARE WE FOLLOWING ALL THE LEADS? Michael F. Good Although the malaria parasite was discovered more than 120 years ago, it is only during the past 20 years, following the cloning of malaria genes, that we have been able to think rationally about vaccine design and development. Effective vaccines for malaria could interrupt the life cycle of the parasite at different stages in the human host or in the mosquito. The purpose of this review is to outline the challenges we face in developing a vaccine that will limit growth of the parasite during the stage within red blood cells — the stage responsible for all the symptoms and pathology of malaria. More than 15 vaccine trials have either been completed or are in progress, and many more are planned. Success in current trials could lead to a vaccine capable of saving more than 2 million lives per year.

Cooperative Research Centre for Vaccine Technology, The Queensland Institute of Medical Research, The Bancroft Centre, 300 Herston Road, Herston QLD 4006, Australia. e-mail: [email protected]

The malaria parasite remains a scourge on human civilization and in recent years the incidence of the disease has been increasing. It is estimated that 1.5–2.5 million people die each year from malaria — mostly young children and pregnant women. Although most of these deaths occur in sub-Saharan Africa, no country is without malaria — either through endemic transmission or through importation of cases from endemic regions of the world. It is now more than 120 years since the French physician Charles Louis Alphonse Laveran first observed malaria parasites under the microscope, and more than 100 years since Ronald Ross and Giovanni Grassi identified the vector of malaria transmission. Malaria is not, therefore, a newly described disease. Since then, there have been many significant developments in malaria research, which have included: unravelling the complex life cycle of the parasite; the development of anti-malarial drugs and insecticides; the discovery of ‘malaria therapy’ for tertiary syphilis (in which the fever associated with malaria killed the heat-sensitive Treponema pallidum organism responsible for syphilis); the development of drug resistance; the cloning of malaria genes (by Kemp and Anders and colleagues in Melbourne, and by Ellis, Godson and

the Nussenzweig’s group in New York); and now earlyphase vaccine trials (BOX 1). Although there are dozens of species of malaria parasites, those that infect humans are limited to Plasmodium falciparum, P. vivax, P. ovale and P. malariae. P. falciparum and P. vivax are the most common, and P. falciparum is responsible for most of the malaria deaths. Multiple strains of each species exist, differing at crucial antigenic determinants. Although some degree of immunity can develop between strains, there is generally thought to be no inter-species immunity. Malaria vaccine development

Vaccination is a successful method of disease control and there have been numerous success stories — the most notable being the eradication of smallpox and the virtual elimination of polio. Many factors conspire to make the development of a malaria vaccine an incredibly difficult challenge. An important factor impeding vaccine development is the complex biology of the life cycle of the parasite, which exists in different forms (and each form with a different pattern of antigen expression) in different tissues of the body and the mosquito (FIG. 1). In these various forms the parasite is susceptible to immune attack, although




REVIEWS

Box 1 | Dates in malaria research • 1880 Discovery of the malaria parasite by Charles Louis Alphonse Laveran108. • 1897–1898 Discovery of the malaria insect vector and life cycle by Ronald Ross109 and Giovanni Grassi110. • 1922 Malaria therapy for syphilis by Julius Wagner von Jauregg111. • 1939 DDT (1,1,1-trichloro-2,2-bis-(p-chlorophenyl)-ethane) discovered as an insecticide by Paul Muller112. • 1943 Chloroquine discovered. • 1947 Exoerythrocytic stage of parasite life cycle described by Fairley113. • 1957 Chloroquine resistance first developed. • 1983 Malaria antigens cloned by Kemp et al.114 and Nussenzweig and co-workers115. • 1986 The first human vaccine trials underway. • 2001 Sequence of malaria genome (near completion).

the type of immune response required is very different for each form. Several vaccine strategies, therefore, need to be used. A second factor that impedes vaccine development is the ability of the malaria parasite to alter itself. ANTIGENIC VARIATION and ALLELIC POLYMORPHISM are important obstacles to SUBUNIT VACCINE development, especially given that many of the sequence alterations in malaria proteins occur in regions that are crucial to immunity. Other factors impeding malaria vaccine development include: immunological non-responsiveness of certain individuals (depending on their human leukocyte antigen and other antigens) to proteins that might comprise a vaccine1; CLONAL IMPRINTING or ORIGINAL ANTIGENIC SIN influenced by stochastic events and prior exposure to other, perhaps crossreactive antigens2–4; the difficulties encountered in properly folding recombinant subunit vaccines so as to maintain their immunogenic properties; the lack of suitably potent adjuvants necessary to induce high-titre antibody responses; and the lack of animal/parasite systems which adequately model the situation with humans and malaria parasites in terms of disease pathogenesis and immunological responses.

Most of the vaccines that have been developed for various diseases have relied on the ‘simple’ approach of presenting the entire antigenic compartment of the organism in the form of killed or living, but attenuated, organisms5. Such an approach is not possible at present for malaria as the organisms grow within red blood cells (RBCs) in the mammalian host (FIG. 1), and although in vitro culture is possible (for example, for P. falciparum, the most virulent of the human plasmodia), a source of RBCs is required. However, it is simply impractical and potentially unsafe to consider growing a vaccine in human RBCs for a disease for which 40% of the world’s population is potentially at risk. Malaria vaccine development has therefore focused on the development of a subunit vaccine. The purpose of this review is to discuss the different strategies for developing a vaccine to the blood stage of malaria. Why is it important to target the blood stage of the parasite? There are different phases in the life cycle of the parasite (FIG. 1): the stage inside the Anopheline mosquito vector, the ‘pre-erythrocytic stage’ and the erythrocytic or blood stage. The preerythrocytic stage, during which the sporozoites travel in the blood after inoculation, and then invade and develop within hepatocytes, is perhaps the best understood in terms of immunity6,7, and vaccine development is further advanced for this stage than others. The mechanism of immunity required to block parasite transmission by the mosquito is also well understood — antibodies taken up by the mosquito during its blood meal can prevent gamete fertilization or zygote development8. This type of vaccine protects only at the population level and does not protect the vaccinee from malaria, and it will be suitable for only certain populations. The blood stage is the stage for which immunity is least well understood and which arguably presents the greatest challenges in terms of vaccine development. It is also the stage that is responsible for all the symptoms and pathology of malaria, the most serious of which are anaemia and cerebral malaria9, and is therefore an important target for vaccine development. In spite of much effort, a blood-stage malaria

ANTIGENIC VARIATION

The antigenic changes that can occur within a parasite clone through switching the expression of different ‘variant’ genes. ALLELIC POLYMORPHISM

Multiple forms of a gene at a single genetic locus. SUBUNIT VACCINES

Vaccines comprising only a small part of the entire organism, typically a recombinant protein. CLONAL IMPRINITNG/ORIGINAL ANTIGENIC SIN

Prior exposure to one strain diverts the antibody response following exposure to a second strain to shared epitopes.

118

Table 1 | Blood-stage malaria vaccines trialled and in development Name

Type

Status

SPf66

Synthetic peptide

Multiple trials completed, not progressing

NYVAC-Pf7 antigen

Viral-vector-encoding parasite

Not progressing

CSP.MSP2

Recombinant protein

Not progressing

MSP1/MSP2/RESA

Multicomponent recombinant protein

Phase I and II trials in Australia and PNG adults and children (continuing)

AMA1

Recombinant protein

Phase I complete (continuing)

MuStDO15

DNA, multi-gene

Under development

FALVAC-1,-2

Recombinant protein, multi-epitope

Under development

MSP119

Recombinant protein

Under development

MSP142

Recombinant protein

Under development

MSP3–5

Recombinant protein

Under development

RAP2

Recombinant protein

Under development

AMA1, apical membrane antigen 1; MSP, merozoite surface-protein; PNG, Papua New Guinea; RAP2, rhoptry-associated protein 2; RESA, ring-infected erythrocyte surface antigen. See REFS 33,75,76.




REVIEWS

Figure 1 | Life cycle of the malaria parasite. The life cycle in the mammalian host commences with the inoculation of sporozoites by an infected anopheline mosquito that travel by the circulation to the liver. After about 1 week (depending on the species of malaria) parasites have multiplied intracellularly and merozoites rupture from infected hepatocytes to invade red blood cells (RBCs). For Plasmodium falciparum and P. vivax there is a 48-hour period inside the RBCs during which merozoites multiply and approximately 16 fresh merozoites are released from ruptured RBCs to invade fresh cells. Sexual forms (gametocytes) develop within RBCs and are taken up by the mosquito. These emerge in the gut of the mosquito as gametes, which fuse to form an oocyst, and sporozoites develop. Sporozoites are released and travel to the salivary gland of the mosquito.

vaccine remains elusive, although several vaccines have been trialled or are now in clinical trials (TABLE 1). What do we want a blood-stage vaccine to do? Because the level of parasitaemia is in general proportional to the severity of disease9, a vaccine must limit parasite growth. Most would argue, however, that it is not necessary to induce sterile immunity after vaccination (that is, no parasites present within a vaccinated individual). It is well recognized that most adults in malaria-endemic settings are clinically immune (that is, they do not suffer symptoms associated with malaria, but they nevertheless have parasites at low density in their blood). The challenge for developing a malaria vaccine might be less if sterile immunity is not required, but until the immunological and other factors that control parasite growth are properly understood, and until their potential is tested in vaccine trials, it is too early to say what type of immunity (sterile or otherwise) should be the goal. Certainly, if sterile immunity could be achieved, then transmission of malaria between individuals would cease. Immunity to malaria

AGGLUTINATING ANTIBODIES

Antibodies directed to parasiteencoded antigens that are expressed on the surface of the red blood cell and lead to clumping of infected red cells.

Before considering the current approaches for development of a subunit vaccine, it is instructive to consider what we know (or think we know) about immunity to malaria. When considering what we know, it can be helpful to distinguish between ‘natural’ immunity and ‘non-natural’ immunity. Natural immunity is induced by multiple exposure to parasites and takes many years of endemic exposure to develop, although in some populations this period can be shortened10.

Theoretically, a vaccine could mimic, but accelerate, this process. Non-natural immunity refers to immune mechanisms that might not be induced to any great extent by natural exposure, but that could theoretically be induced by a vaccine, and that could be highly effective. Examples are given below in which these different approaches are being taken. Variant antigens

Natural immunity and variant antigens. Several years ago Marsh and Howard observed that the sera of convalescent children were able to agglutinate the strains of P. falciparum parasite to which the child was recently exposed, but not other strains circulating in the same village11. By contrast, sera from immune adults in the same village were mostly able to agglutinate all strains. It was subsequently shown that only one strain was present within each agglutinate12. These data indicated that ‘natural’ immunity (as held by adults in a malaria-endemic setting) was due to the development of AGGLUTINATING ANTIBODIES that could recognize each of the different strains in a community. Because each agglutinate contained only parasites of the one strain, it seemed that the immune antibodies within an individual consisted of multiple specificities, as opposed to a monospecific antibody recognizing an antigenic determinant shared by all strains. It is not surprising, therefore, that ‘natural’ immunity takes several years of endemic parasite exposure to develop13. During infection of humans with P. falciparum there is a cyclic recrudescence of parasites to peak levels approximately every 20 days (FIG. 2; W. E. Collins and




% Parasitaemia

REVIEWS

0

40

80

120

160

200

240

Day of infection

Figure 2 | Fluctuation of Plasmodium falciparum in the blood after infection by mosquito inoculation. The sequential peaks in parasite density over the course of an infection are shown. The waning of each peak is thought to be largely due to antibodies to variant antigens expressed on the red-cell surface. However, other factors might contribute to this and the gradual diminution in the height of the parasite peaks over time. Adapted with permission from REF. 9 © (1994) American Association for the Advancement of Science.

defined as ‘cryptic’19. Although crytpic epitopes are poorly immunogenic after infection, they can be recognized by antibodies (or T cells) raised artificially to the isolated epitope. Examples of such epitopes are found in malaria20,21 and other organisms22–25. Because cryptic epitopes are only immunogenic when presented out of context in a non-native form (for example, as peptides), variants would not be selected by immune pressure, and they often have conserved sequences. Of interest is the observation that immune responses to cryptic epitopes can often recognize the native protein or organism21. Recently described PfEMP1-specific monoclonal antibodies that can agglutinate multiple strains26 are likely to be targeting cryptic epitopes. Such epitopes could be the basis for an exciting approach to malaria vaccine development, which could induce a form of non-natural immunity. Merozoite surface proteins

G. M. Jeffrey, unpublished observations referred to in REF 9). This waxing and waning of parasite density in the blood is likely to be due to the selective pressure on the surface of infected RBCs by antibodies to ‘variant’ antigens that arise when a single clone changes phenotype as a result of gene switching14. The principal antigen is P. falciparum erythrocyte membrane protein (PfEMP1) (REFS 15–17), for which there are ~50 variant copies represented in the genome. It is generally believed, but technically unproven, that after one of the variant antigens is clonally expressed, antibodies develop against the expanding clone, which is then eliminated. However, a different variant, not recognized by the antibodies, emerges and the cycling continues. This provides a partial explanation at the molecular level for the above-mentioned observations11. However, the peaks of parasitaemia with each wave gradually diminish, indicating that other immunological responses (other antigens or other types of immune response) also contribute to the development of natural immunity. The relative contribution of each of these is yet to be determined. The fact that natural immunity to malaria (as acquired by individuals living in malaria-endemic regions) takes several years to develop is thought to depend largely on the time taken to acquire antibodies to the multiple PfEMP1 variants. Antibodies to PfEMP1 prevent adherence of the more mature forms of the parasite to the small blood vessels of various organs and tissues, and promote agglutination of parasitized RBCs. Many of the circulating parasites are thought to be removed from the blood in the spleen.

MEROZOITE

The form of the parasite that emerges from infected liver cells and red blood cells (RBCs) and invades fresh RBCs.

120

Non-natural immunity and variant antigens. One approach to vaccine development would be to identify and combine multiple variant epitopes of PfEMP1, analogous with one of the approaches being taken to develop a vaccine for Streptococcus pyogenes18. A related, but distinct, approach is to define a conserved epitope on PfEMP1 that is not normally recognized after infection, but which nevertheless could be a target for agglutinating antibodies. Such an epitope would be

Most blood-stage vaccine research is focused on antigens that are expressed not on the surface of the infected RBC, but on the surface of the MEROZOITE27. Merozoites are released from a rupturing RBC and quickly invade other RBCs. Merozoite surface-protein (MSP)-specific antibodies therefore have only a brief period of time in which to be active. The most studied MSPs are MSP1 (REFS 28,29) and apical membrane antigen 1 (AMA1)30–32. Both antigens are highly polymorphic and have complex folding patterns. Immunization with antigens from P. falciparum or their homologues from monkey or rodent malaria parasites have been shown to protect animals from challenge with a defined strain of the parasite33–39. Both MSP1 and AMA1, or parts of them, are being used in clinical trials (TABLE 1). AMA1 appears on the merozoite surface after its release from organelles of the parasite, referred to as rhoptries. There is extensive allelic diversity of the protein as a result of point mutations, although some regions are conserved. The external amino-terminal domain of the protein has a complex folding pattern. Purified native and recombinant proteins from simian and rodent malaria parasites have been shown to protect against challenge with homologous strains. An extensive review of AMA1 studies has been published33. P. falciparum MSP1 is a large protein of ~200 kDa28,29,40. During release from infected RBCs, the protein undergoes a series of proteolytic digestions, such that only the small 19-kDa carboxy-terminal tail, MSP119 (with a conserved sequence) is carried on the surface of the merozoite into fresh RBCs41. MSP119 and a precursor, MSP142, are the principal MSP1-derived vaccine targets, although a larger amino-terminal fragment, 190L, has shown promise in monkey trials42 and early-stage human vaccine trials43,44. An amino-terminal peptide epitope from MSP1 (not MSP119 or MSP142) was also a component in a multideterminant synthetic vaccine (SPf66), which underwent extensive human trials, resulting overall in negative results — some trials showing efficacy and the more convincing showing none45–49. It has been shown that invasion-inhibitory antibodies can prevent MSP119 processing and invasion, and that blocking




REVIEWS antibodies can inhibit these antibodies28,29,50. Epitopes for inhibitory and blocking monoclonal antibodies could be defined on MSP1, and mutagenized proteins that bound inhibitory but not blocking antibodies could be designed, providing a novel, non-natural approach to malaria vaccines based on MSP1 (REF. 51). An additional region of MSP1 (block 2) has also been identified as a potential target of protective antibodies52. Although MSP119 is considered highly conserved in nature, there are several defined point mutations in the molecule (reviewed in REF. 53), and it is not known what effect, if any, these will have on the efficacy of an MSP119 vaccine based on a single strain. The molecule has a complex folding pattern40,54 with two epidermal growth factor (EGF)-like domains. Proper folding is essential to the ability of a recombinant MSP119 to induce immunity36. Rodent studies indicate that for MSP119, high-titre antibodies present at the time of challenge are required for immunity37,39,55. This is not surprising given that such antibodies have to act quickly during the interRBC phase in the life cycle. Anti-MSP119-specific antibodies do not require Fc function for activity, as shown by the ability of transferred antibodies to reduce parasite density in Fc receptor-deficient mice56,57 and by the ability of recombinant single light-chain antibodies to delay parasite growth (P. Vukovic, M. Foley and M.F.G., unpublished observations). Although Fc function might be important for some malaria-specific antibodies58,59, this does not seem to be the case for MSP119. Although earlier studies with the native MSP1 molecule indicated that antibody-independent cell-mediated immunity is important in protection (that is, immunity mediated by T cells in the absence of antibodies)60, it has been shown that vaccination with recombinant MSP119 cannot protect B-cell-deficient mice, nor can defined synthetic helper CD4+ T-cell epitopes from the molecule induce any protection in normal mice39,61. It is curious and unexplained why MSP119 does not seem to be a target of protective T cells, whereas immunity induced by the

% Parasitaemia

Immunodeficient mice with antibodies at time of challenge

Mice without antibodies at time of challenge

Normal mice with antibodies at time of challenge

Challenge Period of passive immunity

Period of active immunity

Figure 3 | Schematic representation of the effect of antibodies specific for MSP119 in normal and immunodeficient mice. The main effect of MSP119-specific antibodies is to delay the prepatent period of immunity (the period after infection when parasitaemia is so low that parasites cannot be detected by microscopy). In normal mice, that were passively transfused with MSP119-specific antibodies and later challenged with Plasmodium yoelii parasites, the parasitaemia rises and then eventually declines, indicating the induction of an active immune response at a time post-challenge. MSP119, 19-kDa carboxy-terminal tail of merozoite surface-protein 1.

larger MSP1 molecule seems to be mediated, at least in part, by T cells. Furthermore, the ability of cultured T-cell clones and polyclonal populations (cell lines) to adoptively transfer protection is well known (see below), although the antigenic targets of these protective T cells have not been defined. MSP119-specific antibodies at the time of challenge are required to reduce parasite burden; however, in the mouse at least, such antibodies seem incapable of eradicating all parasites. Passively transferred antibodies, even at very high titre, seem incapable of eradicating all parasites in recipient mice that lack either CD4+ T cells, B cells or both62. The observation that such antibodies can both reduce parasite burden and clear all parasites from immunocompetent mice indicates that an active immune response involving both B cells and CD4+ T cells is required for protection even in mice that contain high-titre MSP119-specific antibodies at the time of challenge. The principal effect that MSP119 vaccination has is likely to be the induction of antibodies that significantly reduce, but do not eliminate, parasite burden post-challenge (FIG. 3) This would provide the vaccinee with the additional time necessary to mount a protective immune response to the parasite independent of the vaccine. The specificity of the immune response that develops post-challenge is not known, but it need not be to MSP1 (REF. 62). Recent data indicate that it might target multiple antigens (W. Zhang and M.F.G., unpublished observations). It is not yet known whether the requirements for immunity observed in the mouse will also be relevant in humans. Although some population-based studies have argued for an association between natural immunity and immune responses to MSP1 (REFS 63–65), others have not66. The association would seem less than that between natural immunity and the presence of antibodies to several variants67. As mentioned above, however, there is evidence for inhibitory and blocking antibodies with specificity for MSP1. The different studies that have examined the correlation of protection with the presence of MSP1-specific antibodies have not looked at the fine specificity of these antibodies, and the discrepancies in the data might possibly be explained by the presence of antibodies of different fine specificities in the different populations. It is of interest, also, that antibodies to MSP119 seem to be an important component of the invasion-inhibitory repertoire of malaria parasitespecific antibodies. Elegant studies with transgenic parasites have shown that immune sera (human or rodent) were significantly less capable of blocking invasion of parasites that expressed heterologous versus homologous MSP119 sequences68. Although a strong correlation between invasion-inhibitory antibodies and clinical immunity has not been proven, there is reason to think that such antibodies must contribute to a reduction in parasite load, and as such this study highlights the potential importance of MSP119-specific antibodies. As mentioned earlier, data from mouse models for MSP1 19 indicate that very high antibody levels might be required39. If that applies to humans, then novel adjuvants, such as SBAS2 (REFS 69,70), or novel




REVIEWS

Endocytosed parasite

Although this section has focused primarily on MSP1 and AMA1, several other merozoite vaccine candidates are progressing towards clinical trial. Many of these are referred to in TABLE 1 and in other references33,75,76.

IL-12 MHC II

Spleen

TCR

Immature dentritic cell

Cellular immunity and vaccine strategies

CD28 B7

Death of parasites intracellularly

Maturation inhibited by pRBC

Mature dendritic cell

CD4+ T cell

CD40

CD40L IFN-γ TNF-α

TNF-α 02•–

Phagocytosis of pRBC

N0•

Macrophage

Figure 4 | Schematic of a possible mechanism of action of antibody-independent cellmediated immunity. Some of the factors that are thought to give rise to antibodyindependent (T-cell) immunity are shown. This commences with the activation of CD4+ T cells by mature dendritic cells leading to macrophage activation, phagocytosis of parasitized red blood cells (pRBC), and elaboration of cytokines and small inflammatory molecules (such as nitric oxide and oxygen redicals). T-cell immunity is thought to occur largely, but not entirely, in the spleen. IFN-γ, interferon-γ; IL-12, interleukin-12; MHC II, major histocompatibility complex class II; TCR, T-cell receptor; TNF-α, tumour-necrosis factor-α.

vaccine-delivery schedules, such as those targeting DNA-based vaccines to antigen-presenting cells71, might enable human vaccinees to generate significantly higher titre antibodies following vaccination and be protected from challenge. Human clinical trials with MSP119 have not yet been undertaken. However, a recent trial of a trivalent falciparum vaccine that combines MSP1 (190L), RESA (ring-infected erythrocyte surface antigen) and MSP2 (REF. 33) has recently been completed. There was a reported 62% reduction in parasite densities in the blood of vaccinated children33,44. Although the correlates of immunity have not been defined, it is entirely possible that the type of immunity induced by MSP119 will be different to that induced by the large amino-terminal fragment of MSP1, 190L. Other multicomponent vaccines are also under development (for example, see REFS 72,73). A multicomponent vaccine would offer theoretical advantages over a single-component vaccine: the percentage of immunological non-responders would be reduced and the effect of antigenic polymorphism would also be reduced. However, a drawback to a multicomponent recombinant protein vaccine will be the difficulty in producing and properly folding each of the component parts. The cost of such a vaccine would obviously be significantly greater than for a single-component vaccine. However, multicomponent DNA vaccines for malaria74 would be far less expensive. A challenge with DNAbased vaccines will be to induce a sufficiently high-titre antibody response to the encoded antigens71. Another challenge relates to whether the encoded polypeptide folds appropriately, which might be affected by the ordering of the epitopes on the DNA vaccine.

122

A third general strategy for developing a vaccine for malaria is to identify and use antigens that are targets for antibody-independent cell-mediated immunity. This term refers to the ability of CD4+ T cells that express αβ T-cell receptors to act, in the absence of antibody, and limit parasite growth and is referred to from here on as ‘T-cell immunity’. Early studies elegantly showed the importance of T cells in malaria immunity and indicated that T-cell responses can be effective in controlling parasite densities in the absence of B cells (in mice treated with anti-µ-chain antibodies77). These experiments have now been repeated and extended using B-cell-deficient mice78,79. Data from several papers show that CD4+ T cells, in the apparent absence of antibody, can significantly limit parasite growth; although, in some situations at least, parasite eradication was not achieved79. Adoptive-transfer studies with CD4+ T-cell lines and clones also show that effector T cells of limited antigenic specificity are able to reduce parasite density and, in some cases, to eradicate parasites80–82. Research into vaccines that stimulate T-cell immunity to malaria has been impeded, however, by two factors: a lack of clear understanding of how T cells could kill parasites that are hidden inside RBCs that lack expression of major histocompatibility complex (MHC) molecules; and lack of information of the target antigens of T cells. The most generally accepted model of how parasites might be killed is outlined in FIG. 4, commencing with activation of CD4+ T cells in the spleen, after antigen presentation by dendritic cells (DCs) and ending with the death of parasites (probably intracellularly) after phagocytosis by macrophages and by small inflammatory molecules (oxygen and nitric oxide radicals) in the spleen81,83,84. Evidence from elegant rodent studies indicates that T-cell immunity is regulated by interleukin-12 (IL-12), involves further cytokines (IFN-γ and TNF-α), and might operate finally through nitric oxide83,85. Human studies also support a role for IFN-γ in resistance to malaria86. Immunity to malaria is largely species specific, even though there is obviously no specificity in the action of molecules such as nitric oxide. It would seem that memory CD4+ T cells are reactivated specifically after reinfection, but the parasites are then killed non-specifically. The activation of T cells by parasites, however, might be far more complex. A recent study87 showed that parasitized RBCs might inhibit the maturation of immature DCs, possibly by interaction of CD36 and PfEMP1.The size of the population of already matured DCs at the time of infection might be crucial to the outcome. A further population of γδ T cells (those expressing γδ T-cell receptors) might also be important in immunity, possibly independently of αβ CD4+ T cells and antibody88,89.




REVIEWS When considering the role of CD4+ T cells in human malaria immunity it is instructive to consider the impact of human immunodeficiency virus (HIV)/AIDS on malaria prevalence. Data from several earlier studies90–97 indicated that HIV infection was not associated with a loss of malaria immunity. A recent comprehensive study by Whitworth et al.98, however, found a statistically significant association between the HIV status of adults (mean age 35 years) and the presence of parasites in the blood, and between HIV status and clinical malaria. However, the reported parasite densities were low throughout the study and there were no deaths due to malaria. The overall effects, although statistically significant, would seem small. Furthermore, as noted by Verhoef et al.99, the regression lines of parasite density versus CD4+ T-cell count were not parallel in the HIVpositive cohort and the HIV-negative cohort, as would be expected if HIV was acting directly through a mechanism that involves CD4+ T cells. An earlier study in Malawi100, however, found a significant association between HIV status and the presence of parasitaemia in multigravida women during pregnancy and at delivery. However, little or no difference was observed between parasite burdens of primigravidae with and without HIV. Furthermore, and in contrast with the Whitworth et al.98 study, the prevalence and density of malaria parasitaemia in women more than 60 days after delivery did not differ significantly by HIV status or gravidity. The pregnancy data indicate that HIV specifically interferes with the development of pregnancy-specific immunity, which is thought in large part to be due to the development of antibodies to PfEMP1 epitopes that bind receptors on the placenta101–103. The impact of HIV on non-pregnancy-associated malaria immunity is, however, still a controversial area, and further studies are required to clarify the situation and provide insight into the role of T cells in human malaria immunity. It is difficult to reconcile the existing epidemiological data with the data from animal models that indicate that T cells can be highly efficacious in their anti-parasite effect. A possible explanation is that effector T cells are not activated during infection with P. falciparum, or that, if activated, they are quickly deleted. There is evidence to support this in humans with data showing apoptosis of mononuclear cells taken from patients with acute malaria and placed in vitro104. The specificity of these T cells, however, could not be determined. An alternative hypothesis is that lymphopaenia observed in malaria patients might be due to reallocation of CD4+ T cells to other sites, particularly during severe malaria105. However, we have examined the effect of malaria infection on adoptively transferred CD4+ T cells labelled with a fluorescent dye and specific for whole rodent parasites (P. berghei, P. yoelii or P. chabaudi). This dye labels cytoplasmic proteins and enables one to track and enumerate the labelled cells in recipient mice and to estimate the number of cellular divisions that have occurred (judged by the sequential halving of the fluorescent signal at each division). This technique, in conjunction with other methodologies to quantify cells and cell death, enables one to determine

whether cells are expanded in number or deleted from the recipient as a result of infection. We found that infection specifically deletes CD4+ T cells that are specific for the parasite, although it spares CD4+ T cells specific for an irrelevant antigen, ovalbumin (REF. 106 and H. Xu et al., unpublished observations). The epidemiological data with respect to HIV/AIDS and malaria might reflect a possible deletion of malariaspecific effector CD4+ T cells after exposure to malaria, which would not be further affected by infection with HIV. An alternative explanation for the lack of a pronounced worsening of malaria resistance (in the nonpregnant human) in the face of HIV exposure is that severe disease in malaria might, in part, be T-cell mediated107. However, this is a controversial area, and it is difficult at this stage to reconcile all the epidemiological and experimental data. If effector CD4+ T cells do not contribute greatly to natural immunity to malaria then antigens that would otherwise be targeted by these cells might not be under immune pressure. Antigenic variation and allelic polymorphism are the hallmarks of target antigens on the surface of infected RBCs or merozoites recognized by protective antibodies (see above). If target antigens of T cells could be defined they could be ideal vaccine candidates for a vaccine designed to stimulate a nonnatural form of immunity. By boosting the number of parasite-specific CD4+ effector T cells, the ability of the parasite to multiply might be severely impeded. CD4+ T cells might get the ‘upper hand’ before the parasite leads to their apoptotic deletion. Evidence to support this concept comes from studies in rodents in which it has been shown that it is possible to passively transfer immunity to malaria (that is, lower parasite densities in the blood after challenge) with clones or lines of cultured T cells80–82. The future

So the question that is posed in the title of this article remains — are we following all the leads? We are, but progress is slow. When malaria antigens were first cloned, there was a belief that it would be a relatively short time until a vaccine was produced. The ‘average’ vaccine takes 10–15 years to develop. Against the many challenges, there have been some very optimistic signs. The resilience and dedication of the scientific community to pursue this vaccine are noteworthy. Over the years we have therefore accumulated a lot of knowledge about the pathogenesis of malaria and the nature of immunological responses. The P. falciparum genome sequence is now nearly complete (see link to PlasmoDB: The Plasmodium Genome Resource) and this will provide new insights into pathogenesis and vaccine development. Governments are continuing to invest in the more basic aspects of malaria vaccine research and important philanthropic organizations (such as the Bill & Melinda Gates Foundation) have pledged significant funding to further study vaccine development and to enable trials of the main vaccine candidates to proceed.With continuing perseverance, an effective vaccine will be developed.




REVIEWS 1.

2.

3.

4.

5.

6. 7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

124

Good, M. F., Kumar, S. & Miller, L. H. The real difficulties for malaria sporozoite vaccine development. Nonresponsiveness and antigenic variation. Immunol. Today 9, 351–355 (1988). Good, M. F., Zevering, Y., Currier, J. & Bilsborough, J. ‘Original antigenic sin’, T cell memory, and malaria sporozoite immunity: an hypothesis for immune evasion. Parasite Immunol. 15, 187–193 (1993). Taylor, R. R. et al. Selective recognition of malaria antigens by human serum antibodies is not genetically determined but demonstrates some features of clonal imprinting. Int. Immunol. 8, 905–915 (1996). Riley, E. M. The role of MHC- and non-MHC-associated genes in determining the human immune response to malaria antigens. Parasitology 112, S39–S51 (1996). Plotkin, S. L. & Plotkin, S. A. in Vaccines (eds Plotkin, S. A. & Mortimer, E. A.) 1–7 (W. B. Saunders Co., Philadelphia, 1988). Good, M. F. & Doolan, D. L. Immune effector mechanisms in malaria. Curr. Opin. Immunol. 11, 412–419 (1999). Doolan, D. L. & Hoffman, S. L. The complexity of protective immunity against liver-stage malaria. J. Immunol. 165, 1453–1462 (2000). Kaslow, D. C. Transmission-blocking vaccines: uses and current status of development. Int. J. Parasitol. 27, 183–189 (1997). Miller, L. H., Good, M. F. & Milon, G. Disease pathogenesis in malaria. Science 264, 1878–1883 (1994). This paper provides a concise overview of the pathogenesis of malaria. Baird, J. K. Age-dependent characteristics of protection v. susceptibility to Plasmodium falciparum. Ann. Trop. Med. Parasitol. 92, 367–390 (1998). Marsh, K. & Howard, R. J. Antigens induced on erythrocytes by P. falciparum: expression of diverse and conserved determinants. Science 231, 150–153 (1986). This paper provides some of the best early insights into the nature of natural immunity to malaria and provides an explanation for why prolonged exposure to the parasite is required for this to develop. Newbold, C. I., Pinches, R., Roberts, D. J. & Marsh, K. Plasmodium falciparum: the human agglutinating antibody response to the infected red cell surface is predominantly variant specific. Exp. Parasitol. 75, 281–292 (1992). Greenwood, B. M. et al. Mortality and morbidity from malaria among children in a rural area of The Gambia, West Africa. Trans. R. Soc. Trop. Med. Hyg. 81, 478–486 (1987). Biggs, B. A. et al. Antigenic variation in Plasmodium falciparum. Proc. Natl Acad. Sci. USA 88, 9171–9174 (1991). Baruch, D. I. et al. Cloning the P. falciparum gene encoding PfEMP1, a malarial variant antigen and adherence receptor on the surface of parasitized human erythrocytes. Cell 82, 77–87 (1995). Su, X. Z. et al. The large diverse gene family var encodes proteins involved in cytoadherence and antigenic variation of Plasmodium falciparum-infected erythrocytes. Cell 82, 89–100 (1995). Smith, J. D. et al. Switches in expression of Plasmodium falciparum var genes correlate with changes in antigenic and cytoadherent phenotypes of infected erythrocytes. Cell 82, 101–110 (1995). References 15–17 describe the cloning of the variant antigens of P. falciparum, which enables us to develop rational molecular approaches to understanding pathogenesis and natural immunity to malaria. Brandt, E. R. et al. New multi-determinant strategy for a group A streptococcal vaccine designed for the Australian Aboriginal population. Nature Med. 6, 455–459 (2000). Sercarz, E. E. et al. Dominance and crypticity of T cell antigenic determinants. Annu. Rev. Immunol. 11, 729–766 (1993). Good, M. F., Branigan, J., Smith, G. & Houghten, R. A. Peptide immunization can elicit malaria protein-specific memory helper but not proliferative T cells. Importance of cryptic epitopes in consideration of vaccine design. Pept. Res. 3, 110–115 (1990). Amante, F. H., Crewther, P. E., Anders, R. F. & Good M. F. A cryptic T cell epitope on the apical membrane antigen 1 of Plasmodium chabaudi adami can prime for an anamnestic antibody response: implications for malaria vaccine design. J. Immunol. 159, 5535–5544 (1997). Lo-Man, R. et al. A fully synthetic immunogen carrying a carcinoma-associated carbohydrate for active specific immunotherapy. Cancer Res. 59, 1520–1524 (1999). Raju, R., Diethelm-Okita, B., Okita, D. & Conti-Fine, B. M. Epitope repertoire of human CD4+ lines propagated with tetanus toxoid or with synthetic tetanus toxin sequences. J. Autoimmun. 9, 79–88 (1996).

24. Bou-Habib, D. C. et al. Cryptic nature of envelope V3 region epitopes protects primary monocytotropic human immunodeficiency virus type 1 from antibody neutralization. J. Virol. 68, 6006–6013 (1994). 25. Robert-Guroff, M. et al. Alteration of V3 loop context within the envelope of human immunodeficiency virus type 1 enhances neutralization. J. Virol. 68, 3459–3466 (1994). 26. Gamain, B., Miller, L. H. & Baruch, D. I. The surface variant antigens of Plasmodium falciparum contain cross-reactive epitopes. Proc. Natl Acad. Sci. USA 98, 2664–2669 (2001). 27. Good, M. F., Kaslow, D. C. & Miller, L. H. Pathways and strategies for developing a malaria blood-stage vaccine. Annu. Rev. Immunol. 16, 57–87 (1998). 28. Holder, A. A. in Malaria Vaccine Development: A Multiimmune Response Approach (ed. Hoffman, S. L.) 77–104 (American Society for Microbiology Press, Washington, DC, 1996). 29. Holder, A. A. Malaria vaccines. Proc. Natl Acad. Sci. USA 96, 1167–1169 (1999). 30. Deans, J. A. et al. Vaccination trials in rhesus monkeys with a minor, invariant, Plasmodium knowlesi 66 kD merozoite antigen. Parasite Immunol. 10, 535–552 (1988). 31. Collins, W. E. et al. Protective immunity induced in squirrel monkeys with recombinant apical membrane antigen-1 of Plasmodium fragile. Am. J. Trop. Med. Hyg. 51, 711–719 (1994). 32. Crewther, P. E., Matthew, M. L., Flegg, R. H. & Anders, R. F. Protective immune responses to apical membrane antigen 1 of Plasmodium chabaudi involve recognition of strain-specific epitopes. Infect. Immun. 64, 3310–3317 (1996). 33. Anders, R. F. & Saul, A. Malaria vaccines. Parasitol. Today 16, 444–447 (2000). 34. Kumar, S. et al. Immunogenicity and efficacy in aotus monkeys of four recombinant Plasmodium falciparum vaccines in multiple adjuvant formulations based on the 19-kilodalton C terminus of merozoite surface protein 1. Infect. Immun. 68, 2215–2223 (2000). 35. Egan, A. F., Blackman, M. J. & Kaslow, D. C. Vaccine efficacy of recombinant Plasmodium falciparum merozoite surface protein 1 in malaria-naive,-exposed, and/or-rechallenged Aotus vociferans monkeys. Infect. Immun. 68, 1418–1427 (2000). 36. Ling, I. T., Ogun, S. A. & Holder, A. A. Immunization against malaria with a recombinant protein. Parasite Immunol. 16, 63–67 (1994). 37. Daly, T. M. & Long, C. A. Humoral response to a carboxylterminal region of the merozoite surface protein-1 plays a predominant role in controlling blood-stage infection in rodent malaria. J. Immunol. 155, 236–243 (1995). 38. Tian, J.-H. et al. Genetic regulation of protective immune response in congenic strains of mice vaccinated with a subunit malaria vaccine. J. Immunol. 157, 1176–1183 (1996). 39. Hirunpetcharat, C. et al. Complete protective immunity induced in mice by immunization with the 19-kilodalton carboxyl terminal fragment of the merozoite surface protein-1 (MSP119) of Plasmodium yoelii expressed in Saccharomyves cerevisiae: Correlation of protection with antigen-specific antibody titer, but not with effector CD4+ T cells. J. Immunol. 159, 3400–3411 (1997). 40. Chitarra, V., Holm, I., Bentley, G. A., Petres, S. & Longacre, S. The crystal structure of C-terminal merozoite surface protein 1 at 1.8 Å resolution, a highly protective malaria vaccine candidate. Mol. Cell 3, 457–464 (1999). 41. Blackman, M. J., Heidrich, H. G., Donachie, S., McBride, J. S. & Holder, A. A. A single fragment of a malaria merozoite surface protein remains on the parasite during red cell invasion and is the target of invasion-inhibiting antibodies. J. Exp. Med. 172, 379–382 (1990). 42. Herrera, M. A. et al. Protection against malaria in Aotus monkeys immunized with a recombinant blood-stage antigen fused to a universal T-cell epitope: correlation of serum γ-interferon levels with protection. Infect. Immun. 60, 154–158 (1992). 43. Lawrence, G. et al. Effect of vaccination with 3 recombinant asexual-stage malaria antigens on initial growth rates of Plasmodium falciparum in non-immune volunteers. Vaccine 18, 1925–1931 (2000). 44. Genton, B. et al. Safety and immunogenicity of a three-component blood-stage malaria vaccine in adults living in an endemic area of Papua New Guinea. Vaccine 18, 2504–2511 (2000). 45. Alonso, P. L. et al. Randomised trial of efficacy of SPf66 vaccine against Plasmodium falciparum malaria in children in southern Tanzania. Lancet 344, 1175–1181 (1994). 46. D’Alessandro, U. et al. Efficacy trial of malaria vaccine SPf66 in Gambian infants. Lancet 346, 462–467 (1995). 47. Nosten, F. et al. Randomised double-blind placebo-controlled trial of SPf66 malaria vaccine in


48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

children in northwestern Thailand. Shoklo SPf66 Malaria Vaccine Trial Group. Lancet 348, 701–707 (1996). Bojang, K. A. et al. An efficacy trial of the malaria vaccine SPf66 in Gambian infants – second year of follow-up. Vaccine 16, 62–67 (1998). Metzger, W. G. et al. Serological responses of Gambian children to immunization with the malaria vaccine SPf66. Parasite Immunol. 21, 335–340 (1999). Guevara Patino, J. A., Holder, A. A., McBride, J. S. & Blackman, M. J. Antibodies that inhibit malaria merozoite surface protein-1 processing and erythrocyte invasion are blocked by naturally acquired human antibodies. J. Exp. Med. 186, 1689–1699 (1997). Uthaipibull, C. et al. Inhibitory and blocking monoclonal antibody epitopes on merozoite surface protein 1 of the malaria parasite Plasmodium falciparum. J. Mol. Biol. 307, 1381–1394 (2001). Conway, D. J. et al. A principal target of human immunity to malaria identified by molecular population genetic and immunological analyses. Nature Med. 6, 689–692 (2000). Miller, L. H., Roberts, T., Shahabuddin, M. & McCuthchan, T. F. Analysis of sequence diversity in the Plasmodium falciparum merozoite surface protein-1 (MSP-1). Mol. Biochem. Parasitol. 59, 1–14 (1993). Blackman, M. J., Ling, I. T., Nicholls, S. C. & Holder, A. A. Proteolytic processing of the Plasmodium falciparum merozoite surface protein-1 produces a membrane-bound fragment containing two epidermal growth factor-like domains. Mol. Biochem. Parasitol. 49, 29–33 (1991). Hirunpetcharat, C. et al. Intranasal immunization with yeast-expressed 19 kD carboxyl terminal fragment of Plasmodium yoelii merozoite surface protein-1 (yMSP119 ) induces protective immunity to blood stage malaria infection in mice. Parasite Immunol. 20, 413–420 (1998). Rotman, H. L., Daly, T. M., Clynes, R. & Long, C. A. Fc receptors are not required for antibody-mediated protection against lethal malaria challenge in a mouse model. J. Immunol. 161, 1908–1912 (1998). Vukovic, P., Hogarth, P. M., Barnes, N., Kaslow, D. C. & Good, M. F. Immunoglobulin G3 antibodies specific for the 19-kilodalton carboxyl-terminal fragment of the Plasmodium yoelii merozoite surface protein 1 transfer protection to mice deficient in Fc–γRI receptors. Infect. Immun. 68, 3019–3022 (2000). Oeuvray, C. et al. Cytophilic immunoglobulin responses to Plasmodium falciparum glutamate-rich protein are correlated with protection against clinical malaria in Dielmo, Senegal. Infect. Immun. 68, 2617–2620 (2000). Druilhe, P. & Perignon, J. L. Mechanisms of defence against P. falciparum asexual blood stages in humans. Immunol. Lett. 41, 115–120 (1994). Freeman, R. R. & Holder, A. A. Characteristics of the protective response of BALB/c mice immunized with a purified Plasmodium yoelii schizont antigen. Clin. Exp. Immunol. 54, 609–616 (1983). Tian, J. H. et al. Definition of T cell epitopes within the 19 kDa carboxyl terminal fragment of Plasmodium yoelii merozoite surface protein 1 (MSP1) and their role in immunity to malaria. Parasite Immunol. 20, 263–278 (1998). Hirunpetcharat, C. et al. Absolute requirement for an active immune response involving B cells and TH cells in immunity to Plasmodium yoelii passively acquired with antibodies to the 19 kDa carboxyl terminal fragment of merozoite surface protein-1. J. Immunol. 162, 7309–7314 (1999). Riley, E. M. et al. Naturally acquired cellular and humoral immune responses to the major merozoite surface antigen (PfMSP1) of Plasmodium falciparum are associated with reduced malaria morbidity. Parasite Immunol. 14, 321–337 (1992). Branch, O. H. et al. A longitudinal investigation of IgG and IgM antibody responses to the merozoite surface protein-1 19-kiloDalton domain of Plasmodium falciparum in pregnant women and infants: associations with febrile illness, parasitemia, and anemia. Am. J. Trop. Med. Hyg. 58, 211–219 (1998). Egan, A. F., Burghaus, P., Druilhe, P., Holder, A. A. & Riley, E. M. Human antibodies to the 19 kDa C-terminal fragment of Plasmodium falciparum merozoite surface protein 1 inhibit parasite growth in vitro. Parasite Immunol. 21, 133–139 (1999). Dodoo, D. et al. Levels of antibody to conserved parts of Plasmodium falciparum merozoite surface protein 1 in Ghanaian children are not associated with protection from clinical malaria. Infect. Immun. 67, 2131–2137 (1999). Marsh, K., Otoo, L., Hayes, R. J., Carson, D. C. & Greenwood, B. M. Antibodies to blood stage antigens of Plasmodium falciparum in rural Gambians and their relation to protection against infection. Trans. R. Soc. Trop. Med. Hyg. 83, 293–303 (1989).



REVIEWS 68. O’Donnell, R. A. et al. Antibodies against merozoite surface protein (MSP)-119 are a major component of the invasion-inhibitory response in individuals immune to malaria. J. Exp. Med. 193, 1403–1412 (2001). 69. Stoute, J. A. et al. A preliminary evaluation of a recombinant circumsporozoite protein vaccine against Plasmodium falciparum malaria. N. Engl. J. Med. 336, 86–91 (1997). 70. Stoute, J. A. et al. Long-term efficacy and immune responses following immunization with the RTS,S malaria vaccine. J. Infect. Dis. 178, 1139–1144 (1998). 71. Boyle, J. S., Barr, I. G. & Lew, A. M. Strategies for improving responses to DNA vaccines. Mol. Med. 5, 1–8 (1999). By targeting antigen-presenting cells, DNA vaccines can induce far greater antibody responses. This might be crucial to developing successful vaccines against organisms such as malaria. 72. Shi, Y. P. et al. Immunogenicity and in vitro protective efficacy of a recombinant multistage Plasmodium falciparum candidate vaccine. Proc. Natl Acad. Sci. USA 96, 1615–1620 (1999). 73. Shi, Y. P. et al. Development, expression, and murine testing of a multistage Plasmodium falciparum malaria vaccine candidate. Vaccine 18, 2902–2914 (2000). 74. Wang, R. et al. Induction of antigen-specific cytotoxic T lymphocytes in humans by a malaria DNA vaccine. Science 282, 476–480 (1998). 75. Engers, H. D. & Godal, T. Malaria vaccine development: current status. Parasitol. Today 14, 56–64 (1998). 76. James, S. & Miller, L. Malaria vaccine development: Status report, 2000. http://www.nature.com/nm/special_focus/ malaria/commentaries/malcom 77. Grun, J. L. & Weidanz, W. P. Antibody-independent immunity to reinfection malaria in B-cell-deficient mice. Infect. Immun. 41, 1197–1204 (1983). 78. van der Heyde, H. C., Huszar, D., Woodhouse, C., Manning, D. D. & Weidanz, W. P. The resolution of acute malaria in a definitive model of B cell deficiency, the JHD mouse. J. Immunol. 152, 4557–4562 (1994). 79. von der Weid, T., Honarvar, N. & Langhorne, J. Gene-targeted mice lacking B cells are unable to eliminate a blood stage malaria infection. J. Immunol. 156, 2510–2516 (1996). 80. Brake, D. A., Long, C. A. & Weidanz, W. P. Adoptive protection against Plasmodium chabaudi adami malaria in athymic nude mice by a cloned T cell line. J. Immunol. 140, 1989–1993 (1988). 81. Taylor-Robinson, A. W., Phillips, R. S., Severn, A., Moncada, S. & Liew, F. Y. The role of TH1 and TH2 cells in a rodent malaria infection. Science 260, 1931–1934 (1993). Indicates that TH1- and TH2-type T cells can control malaria infections through different mechanisms. 82. Amante, F. H. & Good, M. F. Prolonged TH1-like response generated by a Plasmodium yoelii-specific T cell clone allows complete clearance of infection in reconstituted mice. Parasite Immunol. 19, 111–126 (1997). 83. Stevenson, M. M., Tam, M. F., Wolf, S. F. & Sher, A. IL-12-induced protection against blood-stage Plasmodium chabaudi AS requires IFN-γ and TNF-α and occurs via a nitric oxide-dependent mechanism. J. Immunol. 155, 2545–2556 (1995). 84. Favila-Castillo, L. et al. Protection of rats against malaria by a transplanted immune spleen. Parasite Immunol. 18, 325–331 (1996). 85. Su, Z. & Stevenson, M. M. Central role of endogenous γ-interferon in protective immunity against blood-stage

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99. 100.

101.

102.

Plasmodium chabaudi AS infection. Infect. Immun. 68, 4399–4406 (2000). Luty, A. J. et al. Interferon-γ responses are associated with resistance to reinfection with Plasmodium falciparum in young African children. J. Infect. Dis. 179, 980–988 (1999). Urban, B. C. et al. Plasmodium falciparum-infected erythrocytes modulate the maturation of dendritic cells. Nature 400, 73–77 (1999). Indicates that the maturation of dendritic cells might be impeded by parasitized RBCs, therefore dampening a developing T-cell response to the parasite. van der Heyde, H. C., Elloso, M. M., Chang, W. L., Kaplan, M., Manning, D. D. &Weidanz, W. P. γδ T cells function in cell-mediated immunity to acute blood-stage Plasmodium chabaudi adami malaria. J. Immunol. 154, 3985–3990 (1995). Seixas, E. M. & Langhorne, J. γδ T cells contribute to control of chronic parasitemia in Plasmodium chabaudi infections in mice. J. Immunol. 162, 2837–2841 (1999). Nguyen-Dinh, P. et al. Absence of association between Plasmodium falciparum malaria and human immunodeficiency virus infection in children in Kinshasa, Zaire. Bull. World Health Organ. 65, 607–613 (1987). Simooya, O. O., Mwendapole, R. M., Siziya, S. & Fleming, A. F. Relation between falciparum malaria and HIV seropositivity in Ndola, Zambia. Br. Med. J. 297, 30–31 (1988). Muller, O. & Moser, R. The clinical and parasitological presentation of Plasmodium falciparum malaria in Uganda is unaffected by HIV-1 infection. Trans. R. Soc. Trop. Med. Hyg. 84, 336–338 (1990). Allen, S. et al. Human immunodeficiency virus and malaria in a representative sample of childbearing women in Kigali, Rwanda. J. Infect. Dis. 164, 67–71 (1991). Greenberg, A. E. et al. Plasmodium falciparum malaria and perinatally acquired human immunodeficiency virus type 1 infection in Kinshasa, Zaire. A prospective, longitudinal cohort study of 587 children. N. Engl. J. Med. 325, 105–109 (1991). Colebunders, R. et al. Incidence of malaria and efficacy of oral quinine in patients recently infected with human immunodeficiency virus in Kinshasa, Zaire. J. Infect. 21, 167–173 (1990). Kalyesubula, I. et al. Effects of malaria infection in human immunodeficiency virus type 1-infected Ugandan children. Pediatr. Infect. Dis. J. 16, 876–881 (1997). Chandramohan, D. & Greenwood, B. M. Is there an interaction between human immunodeficiency virus and Plasmodium falciparum? Int. J. Epidemiol. 27, 296–301 (1998). Whitworth, J. et al. Effect of HIV-1 and increasing immunosuppression on malaria parasitaemia and clinical episodes in adults in rural Uganda: a cohort study. Lancet 356, 1051–1056 (2000). Verhoef, H., Veenemans, J. & West, C. E. HIV-1 infection and malaria parasitaemia. Lancet 357, 232–233 (2001). Steketee, R. W. et al. Impairment of a pregnant woman’s acquired ability to limit Plasmodium falciparum by infection with human immunodeficiency virus type-1. Am. J. Trop. Med. Hyg. 55, 42–49 (1996). Fried, M., Nosten, F., Brockman, A., Brabin, B. J. & Duffy, P. E. Maternal antibodies block malaria. Nature 395, 851–852 (1998). Reeder, J. C. et al. The adhesion of Plasmodium falciparum-infected erythrocytes to chondroitin sulfate A is mediated by P. falciparum erythrocyte membrane protein 1. Proc. Natl Acad. Sci. USA 96, 5198–5202 (1999).


103. Buffet, P. A. et al. Plasmodium falciparum domain mediating adhesion to chondroitin sulfate A: a receptor for human placental infection. Proc. Natl Acad. Sci. USA 96, 12743–12748 (1999). 104. Toure-Balde, A. et al. Plasmodium falciparum induces apoptosis in human mononuclear cells. Infect. Immun. 64, 744–750 (1996). 105. Hviid, L. et al. Rapid reemergence of T cells into peripheral circulation following treatment of severe and uncomplicated Plasmodium falciparum malaria. Infect. Immun. 65, 4090–4093 (1997). 106. Hirunpetcharat, C. & Good, M. F. Deletion of Plasmodium berghei-specific CD4+ T cells adoptively transferred into recipient mice after challenge with homologous parasite. Proc. Natl Acad. Sci. USA 95, 1715–1720 (1998). Indicates that malaria infection can cause deletion of parasite-specific CD4+ T cells. This might prove to be a useful parasite-defence mechanism. 107. Hirunpetcharat, C., Finkelman, F., Clark, I. A. & Good, M. F. Malaria parasite-specific TH1-like T cells simultaneously reduce parasitemia and promote disease. Parasite Immunol. 21, 319–329 (1999). 108. Laveran, A. A new parasite found in the blood of malarial patients. Parasitic origin of malarial attacks. Bull. Mem. Soc. Med. Hop. Paris 17, 158–164 (1880). 109. Ross, R. On some peculiar pigmented cells found in two mosquitos fed on malaria blood. Br. Med. J. 2, 1786–1788 (1897). 110. Russel, P. F., West, L. R., Manwell, R. D. & MacDonald, G. Practical Malariology 13 (Oxford Univ. Press, 1963). 111. Russel, P. F., West, L. R., Manwell, R. D. & MacDonald, G. Practical Malariology 631 (Oxford Univ. Press, 1963). 112. Bruce-Chwatt, L. J. Essential Malariology 301–302 (Alden Press, Oxford, 1985). 113. Fairley, N. H. et al. Sidelights on malaria in man obtained by subinoculation experiments. Trans. R. Soc. Trop. Med. Hyg. 40, 621–676 (1947). 114. Kemp, D. J. et al. Expression of Plasmodium falciparum blood-stage antigens in Escherichia coli: detection with antibodies from immune humans. Proc. Natl Acad. Sci. USA 80, 3787–3791 (1983). 115. Ellis, J. et al. Cloning and expression in E. coli of the malarial sporozoite surface antigen gene from Plasmodium knowlesi. Nature 302, 536–538 (1983).

Acknowledgments I am very grateful to L. Miller, D. Kemp, R. Anders, L. Martin, M. Wykes, S. Elliott and H. Xu for critically reviewing this manuscript. M.F.G. receives research funding support from the National Health and Medical Research Council (Australia) and the United Nations Development Programme/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases.

Online links DATABASES The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ IFN-γ | TNF-α FURTHER INFORMATION Bill & Melinda Gates Foundation: http://www.gatesfoundation.org/ Michael Good’s lab: http://www.qimr.edu.au/research/labs/michaelg/index.html PlasmoDB: The Plasmodium Genome Resource: http://www.plasmodb.org/ Access to this interactive links box is free online.



REVIEWS

CROSS-PRESENTATION IN VIRAL IMMUNITY AND SELF-TOLERANCE William R. Heath* and Francis R. Carbone‡ T lymphocytes recognize peptide antigens presented by class I and class II molecules encoded by the major histocompatibility complex (MHC). Classical antigen-presentation studies showed that MHC class I molecules present peptides derived from proteins synthesized within the cell, whereas MHC class II molecules present exogenous proteins captured from the environment. Emerging evidence indicates, however, that dendritic cells have a specialized capacity to process exogenous antigens into the MHC class I pathway. This function, known as cross-presentation, provides the immune system with an important mechanism for generating immunity to viruses and tolerance to self. ISOTYPE SWITCHING

When B cells change their class of antibody (immunoglobulin) production from one isotype to another, for example from IgM to IgG. THYMIC SELECTION

The process of choosing which thymocytes develop into mature T cells on the basis of the specificity of their T-cell receptors. PERIPHERAL TOLERANCE

The generation of tolerance to self for mature T cells that have left the thymus and are recirculating in the periphery.

*Immunology Division, The Walter and Eliza Hall Institute, P.O. Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. ‡Department of Microbiology and Immunology, University of Melbourne, Parkville, Victoria 3052, Australia. Correspondence to W.R.H. e-mail: [email protected]

126

T lymphocytes can be separated into two subpopulations on the basis of their expression of the cell-surface markers CD4 and CD8. The CD4+ subset is primarily responsible for providing help to other immune cells through direct cell–cell interactions or the secretion of cytokines. Collaboration with B cells, for example, leads to ISOTYPE SWITCHING and enhanced antibody production. CD4+ T cells also have an important role in the induction of inflammatory responses and the generation of CD8+ T-cell immunity. Effective priming of CD8+ T cells leads to their development into mature cytotoxic T lymphocytes (CTLs), which are best known for their capacity to kill virus-infected cells. In this review, for simplicity, we will refer to CD4+ T cells as ‘helper T cells’ and CD8+ T cells as ‘CTLs’. Helper T cells and CTLs use their T-cell receptors to recognize peptide antigens presented by molecules encoded by the MHC. Helper T cells recognize peptides presented by MHC class II molecules, whereas CTLs are restricted to MHC class I molecules. This preference for different classes of MHC molecules relates to a demarcation in the antigen-processing pathways that supply peptides. MHC class II molecules generally present peptides derived from exogenous antigens that enter the cell by the endocytic route, whereas MHC class I molecules present endogenously derived antigens, usually synthesized within the cell presenting the antigen (FIG. 1a,b). The targeting of CTLs to endogenously synthesized

antigens is important as it ensures that virus-specific CTLs only kill cells that are directly infected with virus. Bystander cells that simply endocytose viral debris from infected neighbouring cells will not process this antigen into the MHC class I pathway and will therefore not be targeted by CTLs. Although CTLs perform the very important function of killing cells infected with viruses or intracellular bacteria, their ability to destroy target tissues comes at a price. CTLs with specificity for self-antigens can sometimes attack normal host tissues and cause autoimmunity1,2. For this reason, it is very important to maintain tight control over the generation of effector CTLs, maximizing their pathogen-fighting capacity, while minimizing their autoimmune potential. Although this is primarily achieved during THYMIC SELECTION, in which most selfreactive T cells are deleted, other controls are important. Self-reactive CTLs can, for example, be regulated by 3–7 PERIPHERAL-TOLERANCE mechanisms , and also by checkpoints that prevent their maturation in the absence of signals from helper T cells8–10 (FIG. 2). In this review, we will examine the role of cross-presentation in the generation of CTL immunity (cross-priming) and in the maintenance of self-tolerance (cross-tolerance). What is cross-presentation?

Naive T cells recirculate throughout the secondary lymphoid compartment, moving between the lymph


www.nature.com/reviews/immunology


REVIEWS

CD8 CD4 MHC l-peptide Endocytosed antigens

Plasma membrane Cytosolic antigens

MHC ll-peptide

c Endosome

Cytosolic diversion of endocytosed antigen

Proteasome

a

b Antigenic peptides MllC/CllV

Pathogen

TAP Golgi MHC ll-li

Antigen peptides

MHC ll-CLIP

MHC ll-peptide

Endoplasmic reticulum MHC l

MHC l-peptide

MHC ll-li

Figure 1 | Different antigen-processing pathways for the MHC class I and class II molecules. a | MHC class I molecules present peptides that are primarily derived from endogenously synthesized proteins of either self or pathogen origin. These proteins are degraded into peptides by the proteasome and then transported through the transporters of antigen-processing (TAP) molecules into the endoplasmic reticulum for loading on MHC class I molecules. b | By contrast, MHC class II molecules present proteins that enter the cell through the endocytic route. During maturation of MHC class II molecules, they are prevented from binding to endogenous antigens in the endoplasmic reticulum by association with the invariant chain (Ii). Invariant chain–MHC class II complexes (MHC II–Ii) move through the Golgi to the MIIC/CIIV compartment where the invariant chain is degraded to CLIP (for class II-associated invariant-chain peptide). CLIP is then removed from the CLIP–MHC class II (MHC–CLIP) complexes and exchanged for antigenic peptide. c | Dendritic cells can endocytose antigens from other cells and cross-present them to CD8+ cytotoxic T lymphocytes. The TAP-dependence of such cross-presentation44,52, indicates that it involves diversion of the cellular antigens into the conventional MHC class I pathway, although the mechanism(s) for this diversion are as yet undefined. In most cases, these antigens will also be processed into the MHC class II presentation pathway for recognition by CD4+ helper T cells. (MIIC, MHC II loading compartment; CIIV, MHC II vesicles.)

nodes, blood and spleen. This limited recirculation pattern means that many pathogens enter the body at sites where they will not directly encounter naive T cells. For their initial encounter with antigen, T cells rely on dendritic cells (DCs) to capture pathogen products from the site of infection and transport them to the draining lymph nodes. In this way, naive T cells can scan the entire body for the presence of pathogens simply by scanning antigens presented on DCs that migrate to the secondary lymphoid compartment. As well as transporting antigen, DCs express co-stimulatory molecules that allow them to activate naive T cells, classifying them as professional antigen-presenting cells (APCs). So, once a specific encounter occurs between a T cell and a DC, T cells are activated, proliferate and differentiate, and are then able to enter peripheral tissues to fight the invading pathogen. For MHC class II-restricted responses, which are directed at exogenous antigens, it is easy to imagine how DCs can capture pathogen products and present them to MHC class II-restricted helper T cells in the draining lymph nodes. For MHC class I-restricted responses, which are generally thought to target antigens that are

synthesized within the cell presenting the antigen, it becomes a little more complicated to describe the role of the DC. In the simplest case, DCs could themselves be infected with the pathogen, so allowing MHC class Irestricted presentation of pathogen-derived antigens. Not all pathogens are known to infect DCs, however, and pathogen-infected DCs are often functionally compromised11–17. Therefore, as suggested by Bevan some years ago18, an exogenous pathway for processing MHC class I-restricted antigens within DCs might be necessary. In fact, Bevan’s hypothesis came from his discovery of such a pathway for the priming of CTL immunity19. He showed that protein antigens (in this case minor histocompatibility antigens) that were synthesized in one cell could be captured as exogenous antigens by APCs, processed into the MHC class I antigen-presentation pathway, and used to prime CTL immunity. Bevan termed this ‘cross-priming’, and we have subsequently defined the antigen-processing associated with cross-priming as ‘cross-presentation’. Cross-presentation has been previously used in two contexts in the literature. In the first case, it simply meant processing of exogenous antigens into the MHC class I pathway.




REVIEWS

a

Dendritic cell

Co-stimulatory signal

Licensing signal CD40

CD8

CD4 CD40L

Naive CTL priming

b

Co-stimulatory signal

Licensing signal Pathogen

CD8

Naive CTL priming

c

No co-stimulatory signal

CD8

No licensing signal

Naive CTL not primed

Figure 2 | Licensing of dendritic cells is required for the generation of CTL immunity. a | To prime naive cytotoxic T lymphocytes (CTLs), dendritic cells (DCs) first require a helper T cell-dependent signal via CD40/CD40L (CD154) (REFS 8–10). b | Such CD40-dependent licensing is, however, not always necessary, since pathogen-derived signals, such as viral products, can also license DCs8. In either case, the DCs must be licensed before they can prime naive CTLs. In the absence of licensing, naive CTLs cannot be primed by DCs (c).

The second definition referred to the capture and representation of cell-associated antigen in either the MHC class I or MHC class II pathways. Both definitions have their merits, but it is time to choose which should be used. The most common view seems to be that crosspresentation describes the processing of exogenous antigen into the MHC class I pathway. In this case, crosspriming and cross-tolerance can only really be used in reference to the response of CTL and not helper T cells, which is somewhat of a limitation, but acceptable. As discussed below, this is a property that is primarily limited to a subset of DCs20,21. This makes sense since the indiscriminate capacity of all cells to present exogenous antigens in the MHC class I pathway

128

(cross-presentation) would potentially target noninfected tissue cells that endocytosed viral debris, for destruction by virus-specific CTLs. Therefore, DCs have been bestowed with a specialized property that allows them to cross-present antigens derived from other cells, for the stimulation of naive CTLs (FIG. 1c). Identity of the cross-priming APC

Despite the discovery of cross-priming in the mid1970s19, the phenotype of the APC responsible for this process remained elusive for a quarter of a century. Although this cell was clearly of bone marrow origin22,23, and a professional APC (that is, able to activate naive T cells) 18, researchers had been unable to directly identify the specific subset responsible for cross-presentation. Several groups provided evidence that DCs, macrophages and even B cells were able to cross-present antigens in vitro under specific circumstances24–29, but little success was derived from in vivo attempts to isolate the cross-presenting APC30. Recently, however, Bevan and co-workers21 examined the three known splenic DC subsets (TABLE 1) and provided the first evidence that CD8+ DCs are responsible for cross-priming in vivo. In these studies, mice were injected with ovalbumin (OVA)-bearing cells (known to induce CTLs by crosspriming31), left for 14 hours to allow their DCs to capture and process antigen, and then DC subsets were isolated from the spleen and examined in vitro for their capacity to activate OVA-specific CTLs. Only the CD8+ DC subset cross-presented OVA under these conditions, despite evidence of antigen capture by all three DC subsets. Shortman and colleagues20 later reported that soluble OVA injected intravenously was also crosspresented by CD8+ DCs. Again, all three subsets of splenic DCs captured OVA, but only the CD8+ subset cross-presented it to CTLs. Interestingly, double-negative DCs (CD8–CD4–) could be induced to cross-present OVA if exposed to the bacterial product, lipopolysaccharide, although their cross-presenting capacity was poor compared with that of the CD8+ DC subset. So, why has it taken so long to identify CD8+ DCs as the cross-priming APC? First, CD8+ DCs were themselves only identified as a DC subset in 1992 (REF. 32), they were undetected prior to this because of their sensitivity to isolation procedures. Second, the capacity of this subset to capture antigens in vivo and then crosspresent them in vitro is very inefficient21. Therefore, fragility combined with in vitro presentation inefficiency frustrated early attempts at identifying the cross-presenting APC. Although CD8+ DCs now seem to be the predominant cross-presenting subset, the presence of lipopolysaccharide did allow cross-presentation by double-negative DCs20. So, perhaps different DC subsets will cross-present under different conditions. As discussed in detail below, cross-presentation is associated with both immunity and tolerance. So far, identification of the tolerogenic cross-presenting APC has not been achieved, although Kurts and colleagues33 recently reported that this cell is CD11c+, supporting the idea that it is of DC origin. More precise phenotypic definition of this DC subset must await future studies.




REVIEWS

Table 1 | Phenotype of the three subsets of DCs in the spleen Phenotype CD4/CD8 Mac-1 DEC205

References CD8+ – +

CD8–CD4– + –

CD4+ + –

80

+ +

+ +

+ +

81 20

+

– (+ with LPS signal)

–

20

+

–

–

21

T-cell area T-cell area

Marginal zone T-cell area

Marginal zone T-cell area

Function Phagocytic Pinocytic Cross-presentation of soluble OVA Cross-presentation of cell-associated OVA Location Location (resting) Location (after LPS)

82,83 82

LPS, lipopolysaccharide; OVA, ovalbumin.

Antigens and cross-presentation

How exogenous antigens enter the MHC class I pathway has been extensively reviewed elsewhere34, but three general mechanisms have been defined. The first involves direct ‘injection’ of pathogen-derived antigenic material into the cytosol of host APCs (mediated by viruses35 and some bacteria, such as Listeria monocytogenes36), which allows processing of pathogen proteins by the normal cytosolic machinery for MHC class I. The second mechanism involves endosomal processing. This consists of either direct endosomal loading of preformed MHC class I molecules with peptide determinants that are generated in the endosomal

Figure 3 | Monkey dendritic cells (DCs) acquire labelled plasma membrane from other live monkey DCs. DCs, produced from monkey CD14+ monocytes by culture in granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-4 for 4 days, were labelled with either 1,1′dioctadecyl-3,3,3′,3′-tetramethylindodi-carbocyanine perchlorate (falsely coloured green) or 5chloromethylfluorescein diacetate (falsely coloured red) and co-cultured in a live cell microscopy chamber56. This image was collected at 160 minutes. As single labelled cells interact, small amounts of label are transferred for donor cells to recipient DCs. The arrow indicates an example of transferred material. We thank Dr S. Barratt-Boyes and Dr L. Harshyne for providing this image.

compartment37,38 or ‘regurgitation’ of peptide antigen from the endosomal compartment onto the cell surface for association with preformed MHC class I molecules39. The third mechanism of cross-presentation involves ‘cytosolic diversion’ by as yet undefined pathways. In this case, exogenous proteins are diverted from either the endosomal compartment or directly from the extracellular fluid into the cytosol for processing in the conventional MHC class I pathway. This third situation might be represented by a single mechanism or several different processes, but includes cross-presentation of heat-shock proteins40, antibody complexes25, exosomes41,42, apoptotic cells29, necrotic cells43 and macropinocytosis24. Although there are several pathways for cross-presentation, our current understanding of which pathway(s) operate in vivo for cross-presentation of cell-associated antigens derived from virus-infected cells or self tissues is minimal. Many types of protein antigens have been reported to be cross-presented, including nuclear44, cytoplasmic and cell surface45, foreign31 and self 45, as well as viral46, bacterial31 and eukaryotic19. The level of expression by the donor cell seems to be very important for successful cross-presentation47. Expression levels crucially dictate whether sufficient antigen will be cross-presented to stimulate CTLs. Under normal circumstances, cross-presentation is probably less efficient than direct presentation, since cross-presentation requires the additional step of transfer from one cell to another. So, to detect cross-presentation in a model system, it is important that the donor tissue expresses sufficient antigen (for an extended discussion of this important issue see REF. 30). Apoptotic cells have been reported to be a good antigen source for cross-presentation in vitro29,48,49, and whereas necrotic cells were initially thought to be excluded from cross-presentation29, recent studies show that this is not the case43. Although there is some evidence that necrotic cells can cross-prime in vivo, as illustrated by the capacity of sonicated cell debris to induce CTLs50, there is no direct in vivo evidence that apoptotic cells are cross-presented. McPherson and colleagues51 report that a subset of rat DCs constitutively carry apoptotic gut epithelial cells to the mesenteric lymph




REVIEWS

Pancreas

with no evidence for nibbling by non-DC cell types, such as macrophages. These observations support the possibility that DCs resident in the tissues might move from cell to cell, nibbling pieces of tissue cells without causing damage. This material could then be transported to the draining lymph node for cross-presentation to naive T cells.

Draining lymph node

Dendritic cell Islet β cell

Processing in the class ll pathway

Autoantigen

Conventional class l pathway

Cross-presentation into the class l pathway

DC captures antigen from β cell

Role of cross-presentation in self-tolerance

Proliferation

Islet β cell

CD8 CD8

CD8

CD8

Deletion CD8

CD8

CD8 CD8

CD8 CD8

Figure 4 | Cross-presentation of self-antigens leads to induction of CTL tolerance to peripheral tissues. Dendritic cells (DCs) capture antigen from peripheral tissues, such as pancreatic β-cells, and then cross-present them to autoreactive cytotoxic T lymphocytes (CTLs) in the draining lymph node. This leads to proliferation followed by deletion of the autoreactive CTL, resulting in tolerance to the self-antigens. At present, it is unclear whether the DC captures antigen directly from the tissue cells or simply resides in the draining lymph node where it captures shed antigens.

CENTRAL TOLERANCE

The generation of tolerance to self during T-cell development in the thymus.

130

nodes. They have yet to examine antigen presentation, but their studies formally show that DCs can capture and transport apoptotic cells in vivo. In our own studies47, pancreatic islets that transgenically express low amounts of OVA failed to supply antigen for crosspresentation in the draining lymph node, unless they were first exposed to activated CTLs that destroyed islet tissue. Therefore, killing islet cells enhanced crosspresentation. Whether this was due to the generation of apoptotic cells or simply due to the release of antigen has yet to be resolved52. Despite the evidence that cellular destruction can facilitate cross-presentation, there are situations in which neither apoptosis nor necrosis seem necessary. For example, several lines of transgenic mice express antigens in the pancreatic islets that are cross-presented in the draining lymph nodes in the apparent absence of apoptosis47,53,54. So, how can cross-presentation occur without cellular destruction? One simple possibility is that antigens are secreted and then captured in the draining nodes by DCs that are capable of cross-presentation. However, careful analysis of the response to soluble versus cell-associated OVA has led us to conclude that even when islets express secreted OVA, cross-presentation requires that the antigen is captured in a cellassociated form, perhaps directly from the islet cells55. Support for a mechanism in which DCs can directly capture cellular antigens from live cells comes from a recently described process we refer to as ‘nibbling’56. In these studies, DCs were cultured with various other cell types labelled with dyes. In all cases, material was captured by unlabelled DCs through the nibbling of small vesicles from the donor cells. An example of this transfer is shown in FIG. 3. This was found to be DC-specific,

As well as providing a mechanism for generating immunity to intracellular infections, cross-presentation has been reported to participate in tolerance induction. Von Boehmer’s laboratory first reported a role for cross-presentation in CENTRAL TOLERANCE57, showing that minor histocompatibility antigens were cross-presented within the thymus, where they tolerized CTLs; a process later shown to occur by deletion58. They referred to this process as ‘cross-tolerance’. More recently, cross-tolerance was observed for antigens that were expressed extra-thymically in organs, such as the pancreas and kidney 3,59. These studies were initiated by the discovery that pancreatic and renal antigens can be constitutively cross-presented in draining lymph nodes by a bone marrow-derived APC45, most probably a DC33. Such cross-presentation induced proliferation of naive CD8 T cells, but ultimately led to their deletion3 (FIG. 4). In related studies, helper T cells were shown to be tolerized by MHC class II-restricted presentation of tissuederived antigens59,60. Although in this case it is not strictly cross-presentation, since antigens enter the conventional MHC class II pathway, the tolerance process is most probably mediated by the same DCs through the same tolerogenic signals. Antigen expression levels significantly determine whether self-antigens are cross-presented in sufficient amounts to cause CTL deletion53. So, antigen dose strictly determines the state of tolerance to self, with high-dose antigens inducing deletional tolerance by cross-presentation, and low-dose antigens being ignored. Even when sufficient antigen is expressed to cause cross-presentation, the rate of deletion might be affected by dose. Sherman and colleagues54 reported the more rapid deletion of haemagglutinin-specific CTL in homozygous compared with heterozygous transgenic mice, due to increased expression of haemagglutinin in the pancreas. Interestingly, they showed the deletional process to be rather slow, with deletion of 103 adoptively transferred CTLs requiring 2–4 weeks. This rate would, however, be more than adequate for dealing with the few self-reactive CTLs expected to be generated in a normal repertoire. In addition to antigen dose, the site of expression and age of the host influence cross-presentation of selfantigens61, and hence cross-tolerance62. This understanding began with the observation that presentation of an islet antigen in non-obese diabetic mice did not begin until about the third week of life61. Similarly, both OVA61 and haemagglutinin62, transgenically expressed in the islet β-cells, failed to be cross-presented until this time. This did not seem to be due to a lack of antigen expression62, supporting the idea that cross-presentation




REVIEWS

ALLOGENEIC

Individuals within a species that express allelically variant genes that lead to rejection of transplanted tissue.

of pancreatic antigens is suppressed during the juvenile period. However, this is not a general phenomenon for all tissues since OVA expressed in the kidney was crosspresented at the earliest time examined (day 10) (REF. 61). An important consequence of the late onset of crosspresentation of pancreatic antigens is that there is no deletion of islet-specific CTLs in young mice. In mice expressing haemagglutinin as a transgenic antigen in the pancreatic islets, for example, infection of juvenile mice with influenza virus induced haemagglutinin-specific CTLs that caused autoimmune diabetes62. This was not the case for adult mice, which had deleted their haemagglutinin-specific CTLs by cross-presentation and were resistant to diabetes induction by influenza virus infection. These observations raise the important question of whether juvenile diabetes in humans is associated with a similar failure to cross-present islet antigens during our early years of life. These data indicate that cross-presentation of tissue antigens, and hence cross-tolerance, can be regulated both temporally and regionally. How this regulation occurs must now be addressed. Overall, several studies clearly show that cross-presentation can lead to peripheral self-tolerance3,54,59,60,62. The important unanswered questions are: (i) how are T cells tolerized, (ii) what is the DC subset responsible for this tolerance induction and, most importantly, (iii) does this form of tolerance have a significant role in the natural maintenance of self-tolerance? With respect to the last question, most studies have used transgenically expressed antigens to monitor peripheral tolerance, and it will be important to verify the role of cross-presentation in self-tolerance to natural antigens. Role of cross-presentation in viral immunity

Cross-presentation was first discovered because of its role in generating CTL immunity (cross-priming) to minor histocompatibility antigens expressed by transplanted ALLOGENEIC cells19. Other cell-associated antigens, including those expressed by tumours22 and viruses46, have been reported to cross-prime. However, there are

Box 1 | Cross-priming and immunity to tumours For generating cytotoxic T lymphocyte (CTL) immunity to tumours there are really only two possibilities: either the CTL are primed by direct recognition of antigen on the tumour cells, or else the tumour antigens are cross-presented on host dendritic cells. At present, there is a growing body of evidence that indicates cross-priming can induce immunity to tumours22,41,44,63,64,76,77, and an equally convincing set of data showing it cannot19,78,79. In some cases, in which crosspriming is absent, CTL immunity might be generated by allowing access of the tumour to the secondary lymphoid compartment for direct encounter with naive T cells78,79. Whether cross-priming is involved in the natural immunosurveillance of spontaneous tumours remains unresolved. Certainly, some tumours are induced by viruses, raising the possibility that, at least in these cases, virus-specific cross-priming could limit tumorigenesis. However, it is difficult to assess the significance of cross-priming in natural tumour immunity, since detectable tumours must have escaped immunosurveillance to arise. So, although failure of some tumours to cross-prime could be taken as evidence that cross-priming is not involved in natural tumour immunity, it could equally well mean that tumours only arise if they subvert cross-priming. At present, we cannot distinguish between these two alternatives.

few studies that unequivocally show cross-priming to be vital for natural, protective, CTL immunity. This does not mean that cross-priming has no role in natural immunity, only that it remains difficult to discriminate between the role of cross-presentation and direct presentation in natural CTL priming. In this review, we have focused on the role of cross-priming in viral immunity, although there is good evidence that this mechanism also participates in tumour immunosurveillance (BOX 1). Over the years, it has been shown that many viral antigens can be cross-presented22,43,44,62–64, although in these cases viral proteins were not expressed during virus infections but by transformed or transfected cells, or within transgenic mice. Virus-specific CTL immunity has been shown to depend on bone marrow-derived cells (presumably DCs) for several infections, including influenza virus65, vaccinia virus46,66, poliovirus46 and lymphocytic choriomeningitis virus65, consistent with a role for cross-priming in viral immunity. In these cases, however, it is difficult to exclude direct presentation by virus-infected DCs. Irrespective of their preferred cell tropism for productive replication, many viruses have a broader range of infectivity, in which even partial or abortive infections can give rise to perfectly good CTL determinants. Consequently, active measures are required to unambiguously exclude the possibility of direct presentation, even when a virus is not formally known to infect DCs. In one attempt to exclude direct presentation, Norbury and colleagues67 used a combination of chemical and UV treatment to confine infection to an introduced non-haemopoietic cell line, and thus demonstrate virus-specific CTL cross-priming. Probably the most convincing evidence that crosspriming is sufficient to prime CTL immunity during virus infection, however, comes from Rock and colleagues46. They provided direct evidence for CTL cross-priming to poliovirus in a murine model, in which direct infection of DCs was made impossible. Mice are not a natural host of poliovirus as they lack the receptor found on human cells necessary for virus infection. But by transgenically expressing this receptor on only the non-bone marrow-derived compartment, Sigal and co-workers46 were able to ensure that bone marrow-derived APCs could not be infected. The demonstration that CTL immunity required antigen presentation by a bone marrow-derived cell, despite the inability of poliovirus to infect such cells, implicated cross-priming in the induction of CTL immunity to this virus. Although this example directly shows that virus immunity can be mediated by cross-priming, it does not address the contribution of cross-priming versus direct priming when both pathways are available. Therefore, it remains a formal possibility that crosspriming was observed only because other pathways were inaccessible. Overall, these studies illustrate that cross-priming is observed extensively in experimental systems, supporting the idea that this process exists and is likely to be important for natural infections. One can envisage




REVIEWS

Infected dendritic cell

Infected tissue cell

Pathogen Viral antigen

No migration No direct presentation to CD8+ T cells in the draining lymph node

Processing in the class ll pathway

Cross-presentation into the class l pathway

Shuts down conventional class l pathway

Shuts down

Cross-presenting dendritic cell

Peripheral tissue

Migrates Cross-presentation to CD8+ T cells in the draining lymph node

Lymph node

CD8

CD8

Figure 5 | Viral subversion of dendritic cell function: cross-priming is required to generate CTL immunity. Some viruses impair dendritic cell (DC) function during infection, for example by inhibition of MHC class I-restricted presentation or blocking migration of DCs. In these circumstances, viral inhibition of infected DCs prevents cytotoxic T lymphocyte (CTL) stimulation by the direct MHC class I pathway. However, cross-presentation of antigens that are derived from infected cells by uninfected DCs is likely to result in CTL immunity in the face of inhibitory mechanisms. Although virus-mediated inhibition of target-cell expression of MHC class I will also impair the effector phase of responses induced by cross-priming, this is unlikely to completely block recogniton by effector CTLs.

that cross-priming would have an important role in cases in which a virus infection is truly localized to a peripheral non-lymphoid compartment, such as for papilloma virus, in which the infection is confined to the epithelial cells of the skin 68. In addition, crosspriming could be important in instances where viruses have evolved mechanisms that specifically inhibit conventional MHC class I-restricted antigen processing and presentation69–75. In these situations, crosspresentation would result in successful CTL priming in the face of inhibitory mechanisms that would otherwise prevent direct presentation by infected DCs. In addition to the specific targeting of antigen presentation, there are emerging examples of viruses that inhibit various aspects of DC function. Recently, several viruses, including herpes simplex virus11, measles virus12,13, retrovirus14, canarypox virus15, vaccinia virus16 and lymphocytic choriomeningitis virus 17, have all been shown to have detrimental effects on DCs. Given that professional APCs, such as DCs, are essential for priming viral responses, but that some of these same viruses (for example, vaccinia virus) can ‘deactivate’ DCs during infection, the logical conclusion is that immunity is induced by cross-priming (FIG. 5). By cross-presenting viral antigens, there is no requirement for DCs to be infected (and exposed to the associated hazards) in order to prime naive CTLs to MHC class I-restricted viral antigens. It therefore seems that two different reasons could justify the maintenance of cross-priming by the immune system: first, as suggested by Bevan18, it might be very important for generating CTL immunity to tissue-specific viruses that do not infect DCs; and additionally, it might be vital for generating immunity to viruses that infect DCs, but then inhibit their function.

132

The current verdict on cross-presentation

Over the past decade, a number of significant steps have been taken towards understanding cross-presentation and its consequences — cross-priming and cross-tolerance. Building on the earlier efforts of Bevan and many others, Pardoll and co-workers clearly showed a role for cross-priming in immunity to experimental tumours, whereas Rock’s lab generated the first convincing evidence for virus-specific cross-priming. Extending the studies of von Boehmer’s lab on thymic cross-tolerance, several groups provided strong evidence that peripheral tissue antigens could be cross-presented in the draining lymph nodes for induction of peripheral self-tolerance. In addition, important observations have been made about the nature of antigenic material targeted for crosspresentation, including evidence that cross-presentation can be targeted by heat-shock proteins, apoptotic and necrotic cells, exosomes or even immune complexes. Finally, with the advent of methods for the purification of DC subsets, largely pioneered by Shortman and colleagues, Bevan’s lab was again able to make a significant impact in the field of cross-priming 25 years after his original discovery, by reporting that it is primarily the CD8+ DC subset that is responsible for this process. Despite all this new and exciting information, our understanding of cross-presentation is still only in its infancy. Our new knowledge of which DC subsets are responsible for this process will help us enormously in future efforts to gain a detailed understanding of the mechanism of cross-presentation. Finally, it will be the design of experimental models that can address the nature of tolerance to natural self-antigens, and the contribution of direct versus cross-presentation to immunity, that will be important if we are to define the extent to which we depend on cross-presentation for immunity and tolerance.




REVIEWS

1.

2.

3.

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18. 19.

20.

21.

22.

23.

Miller, B. J., Appel, M. C., O’Neil, J. J. & Wicker, L. S. Both the Lyt-2+ and L3T4+ T cell subsets are required for the transfer of diabetes in nonobese diabetic mice. J. Immunol. 140, 52–58 (1988). Bendelac, A., Carnaud, C., Boitard, C. & Bach, J. F. Syngeneic transfer of autoimmune diabetes from diabetic NOD mice to healthy neonates. Requirement for both L3T4+ and Lyt-2+ T cells. J. Exp. Med. 166, 823–832 (1987). Kurts, C., Kosaka, H., Carbone, F. R., Miller, J. F. & Heath, W. R. Class I-restricted cross-presentation of exogenous self-antigens leads to deletion of autoreactive CD8+ T cells. J. Exp. Med. 186, 239–245 (1997). Provides evidence that self-antigens can be crosspresented and that this leads to deletional tolerance. Schonrich, G. et al. Down-regulation of T cell receptors on self-reactive T cells as a novel mechanism for extrathymic tolerance induction. Cell 65, 293–304 (1991). Schonrich, G. et al. Tolerance induction as a multi-step process. Eur. J. Immunol. 24, 285–293 (1994). Oldstone, M. B., Nerenberg, M., Southern, P., Price, J. & Lewicki, H. Virus infection triggers insulin-dependent diabetes mellitus in a transgenic model: role of anti-self (virus) immune response. Cell 65, 319–331 (1991). Ohashi, P. S. et al. Ablation of ‘tolerance’ and induction of diabetes by virus infection in viral antigen transgenic mice. Cell 65, 305–317 (1991). Ridge, J. P., Di Rosa, F. & Matzinger, P. A conditioned dendritic cell can be a temporal bridge between a CD4+ T-helper and a T-killer cell. Nature 393, 474–478 (1998). Bennett, S. R. et al. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393, 478–480 (1998). Schoenberger, S. P., Toes, R. E., van der Voort, E. I., Offringa, R. & Melief, C. J. T-cell help for cytotoxic T lymphocytes is mediated by CD40–CD40L interactions. Nature 393, 480–483 (1998). Salio, M., Cella, M., Suter, M. & Lanzavecchia, A. Inhibition of dendritic cell maturation by herpes simplex virus. Eur. J. Immunol. 29, 3245–3253 (1999). Servet-Delprat, C. et al. Measles virus induces abnormal differentiation of CD40 ligand-activated human dendritic cells. J. Immunol. 164, 1753–1760 (2000). Fugier-Vivier, I. et al. Measles virus suppresses cellmediated immunity by interfering with the survival and functions of dendritic and T cells. J. Exp. Med. 186, 813–823 (1997). Gabrilovich, D. I. et al. Murine retrovirus induces defects in the function of dendritic cells at early stages of infection. Cell Immunol. 158, 167–181 (1994). Ignatius, R. et al. Canarypox virus-induced maturation of dendritic cells is mediated by apoptotic cell death and tumor necrosis factor alpha secretion. J. Virol. 74, 11329–11338 (2000). Engelmayer, J. et al. Vaccinia virus inhibits the maturation of human dendritic cells: a novel mechanism of immune evasion. J. Immunol. 163, 6762–6768 (1999). Sevilla, N. et al. Immunosuppression and resultant viral persistence by specific viral targeting of dendritic cells. J. Exp. Med. 192, 1249–1260 (2000). Bevan, M. J. Antigen recognition. Class discrimination in the world of immunology. Nature 325, 192–194 (1987). Bevan, M. J. Cross-priming for a secondary cytotoxic response to minor H antigens with H-2 congenic cells which do not cross-react in the cytotoxic assay. J. Exp. Med. 143, 1283–1288 (1976). Pooley, J. L., Heath, W. R. & Shortman, K. Cutting edge: intravenous soluble antigen is presented to CD4 T cells by CD8– dendritic cells, but cross-presented to CD8 T cells by CD8+ dendritic cells. J. Immunol. 166, 5327–5330 (2001). Provides evidence that CD8+ dendritic cells (DCs) are responsible for cross-presentation of soluble ovalbumin. It also shows that CD8–CD4– DCs can cross-present when exposed to lipopolysaccharide. den Haan, J. M., Lehar, S. M. & Bevan, M. J. CD8+ but not CD8– dendritic cells cross-prime cytotoxic T cells in vivo. J. Exp. Med. 192, 1685–1696 (2000). The first report to show that CD8+ dendritic cells cross-present cell-associated antigens. Huang, A. Y. et al. Role of bone marrow-derived cells in presenting MHC class I-restricted tumor antigens. Science 264, 961–965 (1994). Bennett, S. R., Carbone, F. R., Karamalis, F., Miller, J. F. & Heath, W. R. Induction of a CD8+ cytotoxic T

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

lymphocyte response by cross-priming requires cognate CD4+ T cell help. J. Exp. Med. 186, 65–70 (1997). Norbury, C. C., Chambers, B. J., Prescott, A. R., Ljunggren, H. G. & Watts, C. Constitutive macropinocytosis allows TAP-dependent major histocompatibility complex class I presentation of exogenous soluble antigen by bone marrow-derived dendritic cells. Eur. J. Immunol. 27, 280–288 (1997). Regnault, A. et al. Fcγ receptor-mediated induction of dendritic cell maturation and major histocompatibility complex class I-restricted antigen presentation after immune complex internalization. J. Exp. Med. 189, 371–380 (1999). Ke, Y. & Kapp, J. A. Exogenous antigens gain access to the major histocompatibility complex class I processing pathway in B cells by receptor-mediated uptake. J. Exp. Med. 184, 1179–1184 (1996). Rock, K. L., Gamble, S. & Rothstein, L. Presentation of exogenous antigen with class I major histocompatibility complex molecules. Science 249, 918–921 (1990). Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B. & Rock, K. L. Efficient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl Acad. Sci. USA 90, 4942–4946 (1993). Albert, M. L., Sauter, B. & Bhardwaj, N. Dendritic cells acquire antigen from apoptotic cells and induce class Irestricted CTLs. Nature 392, 86–89 (1998). Heath, W. R. & Carbone, F. R. Cross-presentation, dendritic cells, tolerance and immunity. Annu. Rev. Immunol. 19, 47–64 (2001). Carbone, F. R. & Bevan, M. J. Class I-restricted processing and presentation of exogenous cellassociated antigen in vivo. J. Exp. Med. 171, 377–387 (1990). Vremec, D. et al. The surface phenotype of dendritic cells purified from mouse thymus and spleen: investigation of the CD8 expression by a subpopulation of dendritic cells. J. Exp. Med. 176, 47–58 (1992). Kurts, C., Cannarile, M., Klebba, I. & Brocker, T. Dendritic cells are sufficient to cross-present selfantigens to CD8 T cells in vivo. J. Immunol. 166, 1439–1442 (2001). Yewdell, J. W., Norbury, C. C. & Bennink, J. R. Mechanisms of exogenous antigen presentation by MHC class I molecules in vitro and in vivo: implications for generating CD8+ T cell responses to infectious agents, tumors, transplants, and vaccines. Adv. Immunol. 73, 1–77 (1999). Yewdell, J. W., Bennink, J. R. & Hosaka, Y. Cells process exogenous proteins for recognition by cytotoxic T lymphocytes. Science 239, 637–640 (1988). Finelli, A. et al. MHC class I restricted T cell responses to Listeria monocytogenes, an intracellular bacterial pathogen. Immunol. Res. 19, 211–223 (1999). Schirmbeck, R., Melber, K. & Reimann, J. Hepatitis B virus small surface antigen particles are processed in a novel endosomal pathway for major histocompatibility complex class I-restricted epitope presentation. Eur. J. Immunol. 25, 1063–1070 (1995). Bachmann, M. F. et al. TAP1-independent loading of class I molecules by exogenous viral proteins. Eur. J. Immunol. 25, 1739–1743 (1995). Harding, C. V. & Song, R. Phagocytic processing of exogenous particulate antigens by macrophages for presentation by class I MHC molecules. J. Immunol. 153, 4925–4933 (1994). Srivastava, P. K., Udono, H., Blachere, N. E. & Li, Z. Heat shock proteins transfer peptides during antigen processing and CTL priming. Immunogenetics 39, 93–98 (1994). Wolfers, J. et al. Tumor-derived exosomes are a source of shared tumor rejection antigens for CTL crosspriming. Nature Med. 7, 297–303 (2001). Zitvogel, L. et al. Eradication of established murine tumors using a novel cell-free vaccine: dendritic cellderived exosomes. Nature Med. 4, 594–600 (1998). Lu, Z. et al. CD40-independent pathways of T cell help for priming of CD8+ cytotoxic T lymphocytes. J. Exp. Med. 191, 541–550 (2000). Huang, A. Y., Bruce, A. T., Pardoll, D. M. & Levitsky, H. I. In vivo cross-priming of MHC class I-restricted antigens requires the TAP transporter. Immunity 4, 349–355 (1996). Kurts, C. et al. Constitutive class I-restricted exogenous presentation of self antigens in vivo. J. Exp. Med. 184, 923–930 (1996). First paper to report cross-presentation of tissue antigens and showed this process is constitutive.


46. Sigal, L. J., Crotty, S., Andino, R. & Rock, K. L. Cytotoxic T-cell immunity to virus-infected non-haematopoietic cells requires presentation of exogenous antigen. Nature 398, 77–80 (1999). First paper to provide direct evidence that CTL immunity to viruses could be induced by crosspriming. 47. Kurts, C., Miller, J. F., Subramaniam, R. M., Carbone, F. R. & Heath, W. R. Major histocompatibility complex class Irestricted cross-presentation is biased towards high dose antigens and those released during cellular destruction. J. Exp. Med. 188, 409–414 (1998). Provides the important observation that antigen dose is vital to whether a tissue antigen will be cross-presented. Also shows that tissue damage enhances cross-presentaton. 48. Bellone, M. et al. Processing of engulfed apoptotic bodies yields T cell epitopes. J. Immunol. 159, 5391–5399 (1997). 49. Arrode, G. et al. Incoming human cytomegalovirus pp65 (UL83) contained in apoptotic infected fibroblasts is cross-presented to CD8+ T cells by dendritic cells. J. Virol. 74, 10018–10024 (2000). 50. Debrick, J. E., Campbell, P. A. & Staerz, U. D. Macrophages as accessory cells for class I MHCrestricted immune responses. J. Immunol. 147, 2846–2851 (1991). 51. Huang, F. P. et al. A discrete subpopulation of dendritic cells transports apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes. J. Exp. Med. 191, 435–444 (2000). Shows that gut-associated dendritic cells constitutively capture apoptotic epithelial cells and transport them to the mesenteric lymph node. 52. Miller, J. F. et al. Induction of peripheral CD8+ T-cell tolerance by cross-presentation of self antigens. Immunol. Rev. 165, 267–277 (1998). 53. Kurts, C. et al. CD8 T cell ignorance or tolerance to islet antigens depends on antigen dose. Proc. Natl Acad. Sci. USA 96, 12703–12707 (1999). 54. Morgan, D. J., Kreuwel, H. T. & Sherman, L. A. Antigen concentration and precursor frequency determine the rate of CD8+ T cell tolerance to peripherally expressed antigens. J. Immunol. 163, 723–727 (1999). 55. Li, M. et al. Cell-associated ovalbumin is crosspresented much more efficiently than soluble ovalbumin in vivo. J. Immunol. 166, 6099–6103 (2001). 56. Harshyne, L. A., Watkins, S. C., Gambotto, A. & BarrattBoyes, S. M. Dendritic cells acquire antigens from live cells for cross-presentation to CTL. J. Immunol. 166, 3717–3723 (2001). First paper to show that dendritic cells might capture and cross-present antigens from other cells without killing the donor cells. 57. von Boehmer, H. & Hafen, K. Minor but not major histocompatibility antigens of thymus epithelium tolerize precursors of cytolytic T cells. Nature 320, 626–628 (1986). 58. Merkenschlager, M., Power, M. O., Pircher, H. & Fisher, A. G. Intrathymic deletion of MHC class I-restricted cytotoxic T cell precursors by constitutive crosspresentation of exogenous antigen. Eur. J. Immunol. 29, 1477–1486 (1999). 59. Adler, A. J. et al. CD4+ T cell tolerance to parenchymal self antigens requires presentation by bone marrow derived antigen presenting cells. J. Exp. Med. 187, 1555–1564 (1998). 60. Forster, I. & Lieberam, I. Peripheral tolerance of CD4 T cells following local activation in adolescent mice. Eur. J. Immunol. 26, 3194–3202 (1996). 61. Hoglund, P. et al. Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes. J. Exp. Med. 189, 331–339 (1999). 62. Morgan, D. J. et al. Ontogeny of T cell tolerance to peripherally expressed antigens. Proc. Natl Acad. Sci. USA 96, 3854–3858 (1999). 63. Gooding, L. R. & Edwards, C. B. H-2 antigen requirements in the in vitro induction of SV40-specific cytotoxic T lymphocytes. J. Immunol. 124, 1258–1262 (1980). 64. Schoenberger, S. P. et al. Cross-priming of CTL responses in vivo does not require antigenic peptides in the endoplasmic reticulum of immunizing cells. J. Immunol. 161, 3808–3812 (1998). 65. Sigal, L. J. & Rock, K. L. Bone marrow-derived antigenpresenting cells are required for the generation of cytotoxic T lymphocyte responses to viruses and use transporter associated with antigen presentation



REVIEWS

66.

67.

68.

69.

70. 71.

72. 73.

134

(TAP)-dependent and -independent pathways of antigen presentation. J. Exp. Med. 192, 1143–1150 (2000). Lenz, L. L., Butz, E. A. & Bevan, M. J. Requirements for bone marrow-derived antigen-presenting cells in priming cytotoxic T cell responses to intracellular pathogens. J. Exp. Med. 192, 1135–1142 (2000). Norbury, C. C. et al. Multiple antigen-specific processing pathways for activating naive CD8+ T cells in vivo. J. Immunol. 166, 4355–4362 (2001). Tindle, R. W. & Frazer, I. H. Immune response to human papillomaviruses and the prospects for human papillomavirus-specific immunisation. Curr. Top. Microbiol. Immunol. 186, 217–253 (1994). Tortorella, D., Gewurz, B. E., Furman, M. H., Schust, D. J. & Ploegh, H. L. Viral subversion of the immune system. Annu. Rev. Immunol. 18, 861–926 (2000). Hill, A. et al. Herpes simplex virus turns off the TAP to evade host immunity. Nature 375, 411–415 (1995). Ahn, K. et al. Human cytomegalovirus inhibits antigen presentation by a sequential multistep process. Proc. Natl Acad. Sci. USA 93, 10990–10995 (1996). Fruh, K. et al. A viral inhibitor of peptide transporters for antigen presentation. Nature 375, 415–418 (1995). Gilbert, M. J., Riddell, S. R., Plachter, B. & Greenberg, P. D. Cytomegalovirus selectively blocks antigen processing and presentation of its immediate-early gene product. Nature 383, 720–722 (1996).

74. Levitskaya, J., Sharipo, A., Leonchiks, A., Ciechanover, A. & Masucci, M. G. Inhibition of ubiquitin/proteasomedependent protein degradation by the Gly-Ala repeat domain of the Epstein-Barr virus nuclear antigen 1. Proc. Natl Acad. Sci. USA 94, 12616–12621 (1997). 75. Paabo, S. et al. Adenovirus proteins and MHC expression. Adv. Cancer Res. 52, 151–163 (1989). 76. Ronchetti, A. et al. Immunogenicity of apoptotic cells in vivo: role of antigen load, antigen-presenting cells, and cytokines. J. Immunol. 163, 130–136 (1999). 77. Chiodoni, C. et al. Dendritic cells infiltrating tumors cotransduced with granulocyte/macrophage colonystimulating factor (GM-CSF) and CD40 ligand genes take up and present endogenous tumor-associated antigens, and prime naive mice for a cytotoxic T lymphocyte response. J. Exp. Med. 190, 125–133 (1999). 78. Kundig, T. M. et al. Fibroblasts as efficient antigenpresenting cells in lymphoid organs. Science 268, 1343–1347 (1995). 79. Ochsenbein, A. F. et al. Roles of tumour localization, second signals and cross priming in cytotoxic T-cell induction. Nature 411, 1058–1064 (2001). 80. Vremec, D., Pooley, J., Hochrein, H., Wu, L. & Shortman, K. CD4 and CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Immunol. 164, 2978–2986 (2000).


81. Kamath, A. T. et al. The development, maturation, and turnover rate of mouse spleen dendritic cell populations. J. Immunol. 165, 6762–6770 (2000). 82. De Smedt, T. et al. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184, 1413–1424 (1996). 83. Pulendran, B. et al. Developmental pathways of dendritic cells in vivo: distinct function, phenotype, and localization of dendritic cell subsets in FLT3 ligandtreated mice. J. Immunol. 159, 2222–2231 (1997).

Acknowledgements The authors thank several people for their suggestions upon reading drafts of this manuscript, including Dr G. Davey, Dr G. Belz, Dr J. Villadangos, Dr. G. Behrens, Ms J. Mintern, Ms M. Li and Dr M. Bevan.

Online links DATABASE LINKS The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ CD4 | CD8 | CD40 | CD40L | CLIP | GM-CSF | Herpes simplex virus | Measles virus |TAP | Vaccinia virus Access to this interactive links box is free online.



REVIEWS

O N L I N E O N LY Biogs William R. Heath received his Ph.D. from the University of Melboune in 1988 and then spent several years at Scripps Research Foundation in La Jolla before moving to the Walter and Eliza Hall Institute in 1990. Since then, his studies, in collaboration with Dr Carbone, have focused on understanding the mechanisms of peripheral tolerance, particularly with respect to cytotoxic T lymphocytes (CTL). This led to the observation that tissue antigens are cross-presented and induce deletional tolerance. He has also been responsible for dissecting the requirements for cross-priming, providing the first evidence that CD40 licensing of dendritic cells is important for their capacity to prime CTL. Francis R. Carbone received his Ph.D. from the University of Melboune in 1985. He spent a postdoctoral period in the US at Scripps Research Foundation where, in Dr Michael Bevan’s lab, he described the cytosolic class I processing pathway and the capacity of exogenous protein antigens to cross-prime in vivo. He returned to Australia in 1990 and is currently an Associate Professor at the Department of Microbiology and Immunology, University of Melboune. Dr Carbone has a long-standing interest in cross-presentation and is currently applying this to the study of herpes simplex virus infection. At a glance summary T lymphocytes can be separated into two subpopulations, based on their expression of CD4 and CD8. The CD4+ subset is primarily responsible for providing help to other immune cells, whereas CD8+ T cells are best known for their capacity to kill virus-infected cells. Cross-presentation is defined as the processing of exogenous antigens into the major histocompatibility complex (MHC) class I pathway. Cross-priming and cross-tolerance refer to the induction of cytotoxic T lymphocyte (CTL) immunity or tolerance, respectively, that is induced by cross-presented antigens. Despite the discovery of cross-priming in the mid1970s, the antigen-presenting cell responsible for this process has only recently been identified. Bevan and coworkers provided evidence that it is the CD8+ dendritic cell (DC). Although there are several pathways for cross-presentation, our current understanding of which pathway(s) operate in vivo for cross-presentation of cell-associated antigens that are derived from virus-infected cells or self tissues is minimal. As well as providing a mechanism for generating immunity to intracellular infections, cross-presentation has been reported to participate in tolerance induction. Such cross-tolerance is most probably mediated by DCs and leads to the deletion of self-reactive CTLs. Antigen expression levels, the site of expression, the time of expression and the availability of help, crucially determine whether self-antigens cause cross-tolerance. There are few studies that unequivocally show crosspriming to be crucial for natural, protective, CTL immunity. This does not mean that cross-priming has

no role in natural immunity, only that it remains difficult to discriminate between the role of cross-presentation and direct presentation in natural CTL priming. Virus-specific CTL immunity has been shown to depend on bone marrow-derived cells (presumably DCs) for several infections, including influenza virus, vaccinia virus, poliovirus and lymphocytic choriomeningitis virus, consistent with a role for cross-priming in viral immunity. One can envisage that cross-priming has an important role in cases where a virus infection is localized to a peripheral non-lymphoid compartment, such as for papilloma virus infection of the epithelial cells of the skin. In addition, cross-priming could be important where viruses have evolved mechanisms that specifically disrupt the immune functions of DCs. Links CD4 http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l =920 CD8 http://www.ncbi.nlm.nih.gov/LocusLink/list.cgi?Q=c d8%20not%20cd1a%20not%20il10%20not%20Igsf6 &ORG=Hs CD40 http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l =958 CD40L http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l =959 CLIP http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l =16149 GM-CSF http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l =1437 Herpes simplex virus http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd= Retrieve&db=Nucleotide&list_uids=9629378&dopt= GenBank Measles virus http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd= Retrieve&db=Nucleotide&list_uids=9626945&dopt= GenBank TAP http://www.ncbi.nlm.nih.gov/LocusLink/list.cgi?Q=ta p%20not%20tapasin%20not%20export%20not%20d ocking&ORG=Hs Vaccinia virus http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd= Retrieve&db=Nucleotide&list_uids=9790357&dopt= GenBank




REVIEWS

TOLL-LIKE RECEPTORS AND INNATE IMMUNITY Ruslan Medzhitov Toll-like receptors have a crucial role in the detection of microbial infection in mammals and insects. In mammals, these receptors have evolved to recognize conserved products unique to microbial metabolism. This specificity allows the Toll proteins to detect the presence of infection and to induce activation of inflammatory and antimicrobial innate immune responses. Recognition of microbial products by Toll-like receptors expressed on dendritic cells triggers functional maturation of dendritic cells and leads to initiation of antigen-specific adaptive immune responses.

Section of Immunobiology, Yale University School of Medicine, and Howard Hughes Medical Institute, New Haven, Connecticut 06520, USA. e-mail: [email protected]

One of the most fascinating problems in immunology is understanding how the host organism detects the presence of infectious agents and disposes of the invader without destroying self tissues. This problem is not trivial given the enormous molecular diversity of pathogens and their high replication and mutation rates. In response to this challenge, multicellular organisms have evolved several distinct immune-recognition systems. In vertebrate animals, these systems can be broadly categorized as ‘innate’ and ‘adaptive’. Adaptive immune recognition relies on the generation of a random and highly diverse repertoire of antigen receptors — the T- and B-cell receptors (TCR and BCR) — followed by clonal selection and expansion of receptors with relevant specificities. This mechanism accounts for the generation of immunological memory, which provides a significant adaptive fitness to vertebrate animals. However, the adaptive immune response has two main limitations. First, randomly generated antigen receptors are unable to determine the source and the biological context of the antigen for which they are specific. Second, a clonal distribution of antigen receptors requires that specific clones expand and differentiate into effector cells before they can contribute to host defence. As a result, primary adaptive immune responses are delayed, typically for 4–7 days, which is too much of a delay to combat quickly replicating microbial invaders. However, the adaptive immune system does not function independently. Indeed, almost every aspect of the adaptive immune response is controlled by a combination of

permissive and instructive signals, which are provided by the evolutionarily ancient and more universal innate immune system. As will be discussed, the innate immune system detects the presence and the nature of infection, provides the first line of host defence, and controls the initiation and determination of the effector class of the adaptive immune response. Although the innate immune system was first described by Elie Metchnikoff over a century ago, progress in its analysis has been largely overshadowed by the fascinating intricacies of adaptive immunity. Nevertheless, the discoveries of antimicrobial peptides, complement and dendritic cells (DCs), as well as studies in plant and invertebrate immunity, have all greatly contributed to our current understanding of the innate immune system. The recent discovery and characterization of the Toll-like receptor (TLR) family have incited new interest in the field of innate immunity. It is already clear that these receptors have a vital role in microbial recognition, induction of antimicrobial genes and the control of adaptive immune responses. Indeed, recent studies have shown that TLRs have a crucial role in the recognition of ‘molecular signatures’ of microbial infection, in engaging differential signalling pathways, and in controling DC maturation and differentiation of T helper (TH) cells. Innate immune recognition

The strategy of innate immune recognition is based on the detection of constitutive and conserved products of




REVIEWS

Box 1 | PAMPs and virulence factors Pathogen-associated molecular patterns (PAMPs) and virulence factors are not equivalent. PAMPs did not evolve to interact with the host immune system; they evolved to perform essential physiological functions. Pattern-recognition receptors evolved to recognize PAMPs, and therefore to detect the presence of infection. Virulence factors, by contrast, developed as a microbial adaptation to the unique environment within the host. As PAMPs are essential for microbial survival, they are incapable of sustaining mutations. As a result, they are conserved within a class of microbes. Virulence factors are produced by pathogens in order to interact with the host: to invade host cells, to form colonies, to avoid host immune responses, or to adjust to new nutrient sources. Because each group of pathogens has developed a unique strategy for survival within the host, there are multiple virulence factors that can vary between different strains and species of pathogens. Virulence factors are typically encoded by ‘pathogenicity islands’, which are associated with several features characteristic of mobile DNA and can be acquired by, or deleted from, the microbial genome. Furthermore, unlike PAMPs, which in most cases are expressed constitutively, the genes encoding virulence factors are turned on and off depending on the stage of the infection cycle. The lack of conservation and the inducible expression of virulence factors are two probable reasons why, at least in animals, they were not selected during evolution as targets for innate immune recognition. It should be noted, however, that in plants, the distinction between PAMPs and virulence factors might not hold. In addition to PAMP recognition, some plant hostdefence receptors are thought to interact with virulence factors, in particular with the effectors of the TYPE III SECRETION SYSTEM.

TYPE III SECRETION SYSTEM

A specialized multisubunit secretion apparatus found in many Gram-negative bacterial pathogens. It allows the bacteria to secrete various effector proteins directly into the cytosol of the host cells, where they have several functions, such as induction of apoptosis and stimulation of phagocytosis.

136

microbial metabolism. Many metabolic pathways and individual gene products are unique to microorganisms and absent from host cells. Some of these pathways are involved in housekeeping functions and their products are conserved among microorganisms of a given class and are essential for their survival. For example, lipopolysaccharide (LPS), lipoproteins, peptidoglycan and lipoteichoic acids (LTAs) are all molecules made by bacteria, but not by eukaryotic cells. Therefore, these products can be viewed as molecular signatures of microbial invaders, and their recognition by the innate immune system can signal the presence of infection1,2. One important aspect of innate recognition is that its targets are not absolutely identical between different species of microbes. However, although there are several strain- and species-specific variations of the fine chemical structure, these are always found in the context of a common molecular pattern, which is highly conserved and invariant among microbes of a given class. For example, the lipid-A portion of LPS represents the invariant pattern found in all Gram-negative bacteria and is responsible for the pro-inflammatory effects of LPS, whereas the O-antigen portion is variable in LPS from different species of bacteria and is not recognized by the innate immune system. Because the targets of innate immune recognition are conserved molecular patterns, they are called pathogen-associated molecular patterns (PAMPs). Accordingly, the receptors of the innate immune system that recognize PAMPs are called pattern-recognition receptors (PRR)1. PAMPs have three common features that make them ideal targets for innate immune recognition. First, PAMPs are produced only by microbes, and not by host cells. Therefore, recognition of PAMPs by the innate immune system allows the distinction between ‘self’ and ‘microbial non-self ’. Second, PAMPs are invariant

between microorganisms of a given class. This allows a limited number of germ-line-encoded PRRs to detect the presence of any microbial infection. So, recognition of the conserved lipid-A pattern in LPS, for example, allows a single PRR to detect the presence of almost any Gram-negative bacterial infection. Third, PAMPs are essential for microbial survival. Mutations or loss of PAMPs are either lethal for that class of microorganisms, or they greatly reduce their adaptive fitness. Therefore,‘escape mutants’ are not generated. These properties of PAMPs indicate that their recognition must have emerged very early in the evolution of host-defence systems. Indeed, many PAMPs are recognized by the innate immune systems not only of mammals, but also of invertebrates and plants. It is important to note that PAMPs are actually not unique to pathogens and are produced by both pathogenic and non-pathogenic microorganisms. In fact, none of the gene products that are unique to pathogens — the so-called ‘virulence factors’ — are known to be recognized by the mammalian innate immune system (BOX 1). This means that PRRs cannot distinguish between pathogenic and commensal microorganisms. This distinction, however, is vitally important. We live in constant contact with commensal microflora, and continuous activation of inflammatory responses by commensals would have potentially lethal consequences for the host. This, however, does not occur under normal physiological conditions. The exact mechanisms that allow the host to ‘tolerate’ non-pathogenic microorganisms are largely unknown. Presumably, compartmentalization (for example, confinement of microflora to the luminal side of intestinal epithelium), as well as antiinflammatory cytokines, such as transforming growth factor-β (TGF-β) and interleukin (IL)-10, have an important role in this process. The innate immune system uses various PRRs that are expressed on the cell surface, in intracellular compartments, or secreted into the blood stream and tissue fluids. The principal functions of PRRs include: opsonization, activation of complement and coagulation cascades, phagocytosis, activation of proinflammatory signalling pathways and induction of apoptosis1,3 (TABLE 1). Toll-like receptors

The Toll-like receptors are PRRs that have a unique and essential function in animal immunity. TLRs comprise a family of type I transmembrane receptors, which are characterized by an extracellular leucine-rich repeat (LRR) domain and an intracellular Toll/IL-1 receptor (TIR) domain4–6. LRRs are found in a diverse set of proteins in which they are involved in ligand recognition and signal transduction7. The characteristic feature of the LRRs is the consensus sequence motif, L(X2)LXL(X2)NXL(X2)L(X7)L(X2), in which X is any amino acid. The LRR region in the TLRs is separated from the transmembrane region by a so-called ‘LRR carboxy-terminal domain’, which is characterized by the consensus motif CXC(X23)C(X17)C.




REVIEWS The TIR domain of Toll proteins is a conserved protein–protein interaction module, which is also found in a number of transmembrane and cytoplasmic proteins in animals and plants8. Interestingly, most of the TIR domain-containing proteins in animals and plants have a role in host defence (FIG. 1). In transmembrane proteins, the TIR domain is also present in the cytoplasmic portions of members of the IL-1 receptor (IL-1R) family, including IL-1R and IL-18R. Instead of LRR domains, IL-1R and IL-18R have three immunoglobulin domains in their extracellular portions. In mammals, the TIR domain is also present in several cytoplasmic proteins, including two signalling adaptors, MyD88 (REFS 9–12) and TIRAP13 (TIR domain-containing adaptor protein), both of which function in TLR signal transduction (FIG. 1; and see below). TLRs in Drosophila immunity

The first identified member of the Toll family, Drosophila Toll, was discovered as a maternal-effect gene that functions in a pathway that controls dorsoventral axis formation in fruitfly embryos4,14. Other genes in this pathway encode the Toll ligand Spätzle, the adaptor protein Tube, the protein kinase Pelle, the nuclear factor-κB (NF-κB)-family transcription factor Dorsal, and the Dorsal inhibitor and mammalian inhibitor of κB (IκB) homologue Cactus15. Spätzle is secreted as a precursor protein that has to be processed by serine proteases before it can activate Toll15. It should be noted that although genetic studies clearly show that Spätzle functions upstream of Toll, direct binding of Spätzle to Toll has yet to be shown.

The similarity between the Drosophila Toll pathway and the mammalian IL-1R pathway indicated that the Toll pathway might function in fruitfly immunity, as well as in developmental patterning. This was shown in Toll mutant Drosophila, which rapidly succumb to fungal infection, due to a failure to induce the antifungal peptide Drosomycin16. Similarly, fruitflies with loss-offunction mutations in spätzle, tube or pelle were also highly susceptible to fungal infection16. Therefore, the Toll pathway controls not only dorsoventral patterning in embryos, but also the antifungal immune defence in adult fruitflies. One difference between the two pathways is that a different member of the Drosophila NF-kB family, Dif (Drosophila immunity factor), rather than Dorsal, is involved in the antifungal response in adult fruitflies17,18. Interestingly, Drosophila Toll does not function as a PRR, in that it does not seem to recognize pathogens directly. Instead, the processing of Spätzle into a biologically active form is induced on infection and leads, in turn, to the activation of the Toll pathway19. This is shown both by the requirement for Spätzle for antifungal responses, and by the analysis of mutations in the necrotic gene. necrotic encodes a serine protease inhibitor of the serpin family. Mutations in this gene result in the spontaneous activation of the Toll pathway and constitutive induction of the Drosomycin gene19. These results indicate that in Drosophila, the patternrecognition event occurs upstream of Toll and triggers a protease cascade, analagous to complement activation by the lectin pathway in mammals. Interestingly, the Toll pathway can also be activated in response to Gram-positive infection, indicating that several pattern-recognition molecules might function upstream

Table 1 | Pattern-recognition receptors PRR

Protein/domain family

Ligands

Function

References

MBL

C-type lectin

CRP, SAP

Pentraxins

LBP

Lipid-transfer protein family

Terminal mannose residues Phosphorylcholine on microbial membranes LPS

Activation of the lectin pathway of complement Opsonization, activation of classical complement pathway LPS recognition

LPS, peptidoglycan Terminal mannose residues LPS, dsRNA, oxidized LDL, anionic polymers Bacterial cell walls

Co-receptor for TLRs Phagocytosis

42 105

Phagocytosis, LPS clearance, and lipid homeostasis Phagocytosis

106

Secreted PRRs 102 103,104 41

Cell-surface PRRs CD14 Macrophage mannose receptor Macrophage scavenger receptor MARCO

Leucine-rich repeats C-type lectin Scavenger receptor cysteine-rich domain Scavenger receptor cysteine-rich domain

107

Intracellular PRRs PKR

dsRNA-binding domain, protein kinase domain

dsRNA

NODs

Leucine-rich repeats, Nucleotide-binding domain, CARD domain

Ligands for most NOD proteins are unknown. NOD1 and NOD2 were shown to recognize LPS

Activation NF-κB and MAP kinases; inhibition of translation and induction of apoptosis in virally infected and stressed cells Activates NF-κB and MAP kinases; some family members may be involved in the induction of apoptosis. The exact function is unknown.

74

108,109

CARD, caspase-recruitment domain; CRP, C-reactive protein; LBP, lipopolysaccharide (LPS)-binding protein; LDL, low-density lipoprotein; MAP, mitogen-activated protein; MARCO, macrophage receptor with collagenous structure; MBL, mannan-binding lectin; NF-κB, nuclear factor-κB; PKR, double-stranded RNA (dsRNA)activated protein kinase; PRR, pattern-recognition receptor; SAP, serum amyloid protein; TLRs, Toll-like receptors.




REVIEWS

Plants

Drosophila

Mammals

PAMP Protease PAMP Spätzle

IL-1

Toll

TLR4

IL-1R

Extracellular Pathogen or PAMP

Cytoplasm

MyD88

TIRAP

MyD88

MyD88

RPP5, N, L6

Immune response

Immune response

Immune response

Figure 1| TIR domain in host-defence pathways. The Toll/interelukin-1 (IL-1) receptor (TIR) domain is a protein-interaction module found in transmembrane and cytoplasmic proteins involved in animal and plant immunity. RPP5, N and L6 are prototypic examples of intracellular plant-disease-resistance proteins that contain an amino-terminal TIR domain as well as a nucleotide (ATP or GTP)-binding domain and leucine-rich repeat (LRR) domains. Drosophila has two types of protein with TIR domains: Tolls and MyD88. At least one out of nine Tolls in Drosophila, as well as MyD88, are involved in host defence. Toll is activated by a proteolytically processed form of the Spätzle protein. The cleavage of Spätzle is triggered by an unknown pattern-recognition molecule responsive for fungal and Gram-positive bacterial pathogens (see text for details). Mammals have at least four types of proteins with TIR domains: members of the Toll-like receptor (TLR) and IL-1 receptor (IL-1R) families, MyD88 and TIRAP (TIR domain-containing adaptor protein). In TLRs and IL-1Rs, the TIR domain is carboxy-terminal to LRRs and immunoglobulin domains, respectively. Both mammalian and Drosophila MyD88 contain carboxy-terminal TIR domains and amino-terminal death domains and function as adaptor proteins. TIRAP is another adaptor protein that does not have a Drosophila homologue. TIR has a carboxy-terminal TIR domain, but lacks a death domain. The amino-terminal region of TIRAP does not share similarity with any known protein. PAMP, pathogen-associated molecular pattern.

DEATH DOMAIN

A protein–protein interaction domain found in many proteins that are involved in signalling and apoptosis.

138

of the protease cascade that controls cleavage of Spätzle20. The upstream cascade that generates active Spätzle in response to infection has not yet been identified. Despite their profound defect in antifungal immunity, fruitflies harbouring mutations in Toll and the other components of the Toll pathway show normal resistance to infection by Gram-negative bacteria16. Similar to wildtype fruitflies, they produce antimicrobial peptides specific for Gram-negative bacteria, such as Diptericin16. Drosophila therefore discriminates between different classes of pathogens, such that the antifungal peptide Drosomycin is selectively produced on fungal infection, whereas Diptericin is made in response to Gram-negative bacteria21. Furthermore, although the Toll pathway regulates antifungal defence, resistance to Gram-negative infection is conferred by a distinct pathway, which was defined by a mutation in the imd (immune-deficient) gene22. imd mutants fail to induce the antibacterial peptide Diptericin and, therefore, have a profound defect in resistance to Gram-negative bacterial pathogens, although remaining essentially normal with regard to fungal and Gram-positive infection23. The imd gene has recently been identified and shown to encode an adaptor protein with a DEATH DOMAIN24. So, Imd presumably functions downstream of a putative receptor responsible for sensing Gram-negative bacteria24. Genetic analyses led to the identification of five additional Drosophila genes that function in the Imd pathway: Dredd 25,26, dIKK-β 27,28 (IκB kinase-β),

dIKK-γ 29, dTAK1 (a homologue of TGF-β-activated kinase 1)30 and Relish31. Mutations in any of these genes yield phenotypes very similar to imd mutants — susceptibility to Gram-negative bacterial infection due to impaired induction of antibacterial peptides, such as Diptericin23. Dredd is a Drosophila caspase that was previously implicated in the control of apoptosis during fruitfly development32. Drosophila IKK-γ and IKK-β are homologues of human IKK-γ —also known as NEMO (NF-κB essential modulator) — and IKK-β. In human cells, IKK-β and NEMO are essential regulators of NF-κB activation33. Relish is a Drosophila homologue of the mammalian Rel/NF-κB family members, p100 and p105 (REF. 34). Interestingly, the Toll and Imd pathways use different NF-κB transactivators that are activated by distinct mechanisms20. Dif, similar to its mammalian homologues p50 and p65, is activated on stimulus-dependent degradation of its inhibitor Cactus17,18. Relish, in contrast, is homologous to mammalian p105, and is activated by a proteolytic processing event that removes its autoinhibitory ankyrin repeats28. Dredd was shown to function downstream of Drosophila IKKγ and IKKβ, but its involvement in Relish processing has not yet been shown26. Interestingly, the Imd pathway lacks an IκBlike molecule, an obvious target of Drosophila IKK-β phosphorylation that would be analogous with the mammalian NF-κB pathway, as Cactus seems to function exclusively in the Toll pathway20. Conversely, how




REVIEWS Toll activation results in Cactus degradation is not yet clear, as no Cactus kinase has yet been identified. Therefore, although there are similarities in the Toll/NF-κB pathways in Drosophila and mammals, there are also intriguing differences. One of the main questions in Drosophila immunity that remains unresolved is the identities of the patternrecognition molecules that trigger processing of Spätzle in response to fungal and Gram-positive infection (FIG. 2). Another very important question is the identity of the receptor that controls activation of the Imd pathway in response to Gram-negative bacterial infection. As there are nine TLRs in Drosophila35, an attractive possibility is that one of them might be responsible for the activation of the Imd pathway. A mutation in 18 Wheeler, another Toll family member, was shown to affect expression of several antibacterial peptides36. However, 18 Wheeler does not seem to function in the Imd pathway23,35. Moreover, none of the Drosophila Tolls could induce activation of the Diptericin promoter in Drosophila cell lines, and only Toll and Toll-5 were able to activate Drosomycin35. Therefore, it is likely that a receptor unrelated to Toll might control the Imd pathway and function as a sensor for Gram-negative PAMPs such as LPS.

be secreted into serum, or expressed as a glycophosphoinositol (GPI)-linked protein on the surface of macrophages42. CD14-deficient mice have a profound defect in responsiveness to LPS, showing the importance of CD14 in LPS recognition43. Another component of the LPS receptor complex is MD-2 (REF. 44). MD-2 is a small protein that lacks a transmembrane region and is expressed on the cell surface in association with the ectodomain of TLR4 (REF. 44). Although its precise function is not known, MD-2 is required for LPS recognition by TLR4 (REF. 45). The molecular mechanism of TLR-mediated recognition is one of the most challenging issues in Toll biology. Several lines of evidence indicate that TLR4 might, in fact, interact with LPS directly46,47; however, this interaction is clearly aided by CD14 and MD-2 (REF. 48).

Fungal infection Gram-positive bacterial infection

Gram-negative bacterial infection

Toll 1

?

TLRs in mammalian immunity

In mammalian species there are at least ten TLRs, and each seems to have a distinct function in innate immune recognition. In the past few years, dozens of TLR ligands have been identified37. Many more ligands are yet to be identified, both for those TLRs that already have assigned ligands and those with no known ligands. TLR ligands are quite diverse in structure and origin. However, several common themes are emerging based on the available information. First, most TLR ligands are conserved microbial products (PAMPs) that signal the presence of infection (FIG. 3). Second, many, and perhaps all, individual TLRs can recognize several, structurally unrelated ligands. Third, some TLRs require accessory proteins to recognize their ligands. Finally, although the actual mechanism of ligand recognition is still not known, available evidence indicates that mammalian TLRs recognize their ligands by direct binding and therefore function as PRRs. TLR4. Human TLR4 was the first characterized mammalian Toll5. It is expressed in a variety of cell types, most predominantly in the cells of the immune system, including macrophages and DCs5. TLR4 functions as the signal-transducing receptor for LPS38–40. This discovery was made by positional cloning of the Lps gene in the LPS-non-responsive C3H/HeJ mouse strain38,39, and was confirmed in Tlr4 knockout mice40. C3H/HeJ mice are unresponsive to LPS due to a point mutation in the TIR domain of Tlr4, which abrogates downstream signalling38,39. Recognition of LPS by TLR4 is complex and requires several accessory molecules. LPS is first bound to a serum protein, LBP (LPS-binding protein), which functions by transferring LPS monomers to CD14 (REF. 41). CD14 is a high-affinity LPS receptor that can either

Tube

MyD88

Imd

TAK1

Pelle

?

IKKβ

IKKγ Dredd

Cactus

Dif

Dif

Drosomycin

Rel

Rel

Diptericin

Figure 2 | Drosophila Toll and Imd pathways. The Drosophila Toll pathway is activated by fungal and Grampositive bacterial pathogens and induces production of antifungal peptides, such as Drosomycin. Toll signals through two adaptor proteins, Tube and MyD88, which function upstream of the protein kinase Pelle. The signalling components immediately downstream of Pelle are not known. Activation of this pathway leads to degradation of Cactus and release of the nuclear factor-κB (NF-κB) family transcription factor Dif (Drosophila immunity factor). Dif, in turn, activates the transcription of Drosomycin and other antimicrobial peptides. The Imd pathway is triggered in response to Gram-negative bacterial infection through an unknown receptor. In addition to Imd, this pathway involves the Drosophila homologue of the protein kinase TAK1(TGF-β-activated kinase), the IKK-γ/IKK-β protein kinase complex, the caspase Dredd and the NF-κB family transcription factor Relish. This pathway is responsible for the induction of antibacterial peptides, such as Diptericin and Drosocin, in response to bacterial infection. Rel, Relish.




REVIEWS

Peptidoglycan (G+) Lipoprotein Lipoarabinomannan (Mycobacteria) LPS (Leptospira) LPS (Porphyromonas) GPI (Trypanosoma cruzi) Zymosan (Yeast)

TLR6

TLR2

LPS (G–) Lipoteichoic acids (G+) RSV F protein

TLR1

CD14

dsRNA

TLR4

TLR3

Flagellin

TLR5

Unmethylated CpG DNA

TLR9

MD-2

Figure 3 | Ligand specificities of TLRs. Toll-like receptors (TLRs) recognize a variety of pathogen-associated molecular patterns (PAMPs). Recognition of lipopolysaccharide (LPS) by TLR4 is aided by two accessory proteins: CD14 and MD-2. TLR2 recognizes a broad range of structurally unrelated ligands and functions in combination with several (but not all) other TLRs, including TLR1 and TLR6. TLR3 is involved in recognition of double-stranded (dsRNA). TLR5 is specific for bacterial flagellin, whereas TLR9 is a receptor for unmethylated CpG motifs, which are abundant in bacterial DNA. G+, Gram-positive; G–, Gram negative; GPI, glycophosphoinositol; RSV, respiratory syncytial virus.

Another protein that seems to cooperate with TLR4 in LPS recognition is RP105. RP105 is an LRR-containing protein expressed almost exclusively on the surface of B cells49. The extracellular region of RP105 is related to the ectodomain of TLR4; however, RP105 lacks the TIR domain and instead has a short cytoplasmic region with a tyrosine-phosphorylation motif49. Ligation of RP105 leads to activation of SRC-family tyrosine kinases, including LYN50. Similar to TLR4, RP105 is associated with an accessory protein, MD-1, which is a homologue of MD-2 (REF. 51). Deletion of the RP105 gene results in reduced responsiveness of B cells to LPS52. As TLR4deficient mouse B cells are completely unresponsive to LPS, RP105 and TLR4 presumably cooperate in LPS recognition and signalling in B cells, although the exact nature of this cooperation remains unknown. In addition to LPS, TLR4 is involved in the recognition of several other ligands, including LTA53, and a heat-sensitive cell-associated factor derived from Mycobacterium tuberculosis54. TLR4 is also implicated in the recognition of the heat-shock protein HSP60 (REF. 55). HSP60 is a molecular chaperone that is conserved from bacteria to mammals. It is normally not available for recognition by cell-surface receptors, but presumably can be released from necrotic cells during tissue injury or lysis of virally infected cells. The physiological significance of HSP60 recognition by a TLR is not yet understood, but the inflammatory response induced by necrotic cells (which might be mediated by HSPs and other ligands released from dying cells) might have a role in tissue remodelling and wound healing56. Interestingly, TLR4 and CD14 were also shown to trigger a response to the fusion (F) protein of respiratory syncytial virus (RSV)57. It is not clear yet whether the F protein of RSV represents an example of a viral PAMP, in that some conserved feature of the F protein is shared with fusion proteins of other viruses. An alternative

140

possibility is that the RSV evolved the ability to stimulate TLR4 for its own benefit. More examples of viral interactions with TLRs are likely to be discovered in the near future. Not surprisingly, some viruses evolved the ability to interfere with TLR function. For example, the vaccinia virus encodes two cytoplasmic proteins that block TLR and IL-1R signal transduction58. TLR2. TLR2 has been shown to be involved in the recognition of a broad range of microbial products, including: peptidoglycan from Gram-positive bacteria53,59, bacterial lipoproteins60–62, mycobacterial cell-wall lipoarabinomannan63,64, glycosylphosphatidylinositol lipid from Trypanosoma Cruzi65, a phenol-soluble modulin produced by Staphylococcus epidermidis66, and yeast cell walls67 (FIG. 3). In addition, TLR2 functions as a receptor for atypical LPS produced by Leptospira interrogans68 and Porphyromonas gingivitis69, both of which are structurally different from Gram-negative LPS. This unusually broad range of ligands recognized by TLR2 is explained, in part, by cooperation between TLR2 and at least two other TLRs: TLR1 and TLR6 (REFS 70,71). So, the formation of heterodimers between TLR2 and either TLR1 or TLR6 dictates the specificity of ligand recognition70,71. For example, TLR2 cooperates with TLR6 for the recognition of mycoplasmal macrophage-activating lipopeptide 2 kDa (MALP-2)71. Interestingly, it is TLR6 that discriminates between bacterial lipoproteins, which are triacylated at the amino-terminal cysteine residue, and the diacylated mycoplasmal lipoprotein MALP-2. This conclusion is based on the finding that TLR2-deficient macrophages are unresponsive to both bacterial and mycoplasmal lipoproteins, whereas TLR6-deficient cells are unresponsive to MALP-2, but respond normally to bacterial lipoproteins71. Therefore, TLR2 cooperates with TLR6 for recognition of MALP-2, but presumably with another TLR for the recognition of triacylated bacterial




REVIEWS lipoproteins. It is not known yet whether TLR2 heterodimerization is induced by appropriate ligands or occurs prior to ligand interaction. It is interesting to note that both TLR1 and TLR6 are expressed constitutively on many cell types, whereas expression of TLR2 is regulated and seems to be restricted to antigen-presenting cells and endothelial cells72. The combinatorial recognition by TLR2 and regulation of its expression might provide an important mechanism to control cellular responsiveness to microbial products. The full repertoire of possible TLR heterodimers is not yet known, but TLR4 and TLR5, at least, are likely to function as homodimers70. TLR3. TLR3 has two interesting features that distinguish it from other mammalian TLRs. First, cloning of human and mouse TLR3 immediately showed that, unlike all other TLRs, TLR3 does not contain the conserved proline residue in the position equivalent to proline-712 of mouse TLR4. Substitution of this proline residue for histidine in the Tlr4 gene in the C3H/HeJ mouse strain results in unresponsiveness to LPS. Equivalent substitutions in some other TLRs abolish their signalling activities67,70. Therefore, the fact that TLR3 lacks the conserved proline at this crucial position indicated that the TLR3 signalling mechanism might differ from that of other TLRs. The second interesting feature of TLR3 is that it is expressed predominantly, albeit not exclusively, in dendritic cells72. Recent studies have shown that TLR3 functions as a cell-surface receptor for double-stranded RNA (dsRNA) (FIG. 3)73. dsRNA is a molecular pattern produced by most viruses at some point of their infection cycle. It has long been known to have immunostimulatory activity, partly because of its ability to activate the dsRNA-dependent protein kinase, PKR74. However, PKR-deficient cells are still able to respond to both dsRNA and its synthetic analogue, polyinosine-polycytosine (polyIC)75, indicating the existence of another receptor for dsRNA. This receptor seems to be TLR3, as cells deficient for TLR3 have a profound defect in their responsiveness to polyIC, as well as to viral dsRNA73. Although a contribution of TLR3 to antiviral defence remains to be shown, the fact that dsRNA — an important viral PAMP — is recognized by a TLR, significantly broadens the range of pathogens that can be detected by the TLRs.

Interestingly, TLR5 is expressed on the basolateral side of the intestinal epithelium, where it can sense flagellin from pathogenic bacteria, such as Salmonella78. Polarized expression of TLR5 (and presumably other TLRs) on surface epithelia might provide an important mechanism of discrimination between commensal and pathogenic bacteria, as pathogenic, but not commensal microbes, can cross the epithelial barriers. TLR9. Perhaps the most enigmatic example of pattern recognition is the recognition of unmethylated CpG motifs in bacterial DNA by TLR9 (REF. 79) (FIG. 3). Unmethylated DNA in a particular sequence context (the so-called ‘CpG motif ’) has long been known for its potent immunostimulatory activity80. A single nucleotide substitution or methylation of a cytosine residue within the CpG motif completely abrogates the immunostimulatory property of bacterial DNA80. Because bacteria lack cytosine methylation, and most CpG is methylated in the mammalian genome, CpG motifs might signal the presence of microbial infection. The essential role of TLR9 in CpG DNA recognition was shown using Tlr9 knockout mice79. Interestingly, signalling by CpG DNA requires its internalization into late endosomal or lysosomal compartments81. The reason for this is not yet known, and it will be important to determine the subcellular localization of TLR9. It is not yet known whether any other TLR ligands need to be internalized in order to activate TLRs. Notably, TLR2 is expressed on the cell surface and is recruited to phagosomes on interaction with yeast cell walls (zymosan)67. Additionally, some available data indicate that signalling by LPS might require its internalization82. Another enigmatic aspect of CpG DNA recognition is that the optimal response of mouse versus human cells requires slightly different sequence motifs flanking CpG dinucleotides83. It has recently been shown that CpG DNA that optimally stimulates mouse cells is also a much stronger activator of transfected mouse TLR9 compared with human TLR9; the opposite is true of CpG DNA that preferentially stimulates human cells84. These results indicate that TLR9 itself can distinguish between the two immunostimulatory CpG motifs, and therefore can presumably recognize CpG DNA directly84. TLR signalling pathways

TLR5. TLR5 is involved in recognition of flagellin — a conserved protein that forms bacterial flagella76 (FIG. 3). An unusual aspect of this TLR ligand is that, unlike most other PAMPs, flagellin is a protein, and it does not undergo any posttranslational modification that would distinguish it from host cellular proteins. However, the amino- and carboxy-termini of flagellin are extremely conserved, presumably because they form a hydrophobic core of the flagella and have significant structural constraint on variability77. This extreme structural conservation and the vitally important function of flagellin for bacterial mobility explain why it was selected as a target for recognition by Toll.

Activation of signal transduction pathways by TLRs leads to the induction of various genes that function in host defence, including inflammatory cytokines, chemokines, major histocompatibility complex (MHC) and co-stimulatory molecules. Mammalian TLRs also induce multiple effector molecules such as inducible nitric oxide synthase and antimicrobial peptides, which can directly destroy microbial pathogens85. Although both TLRs and IL-1Rs rely on TIR domains to activate NF-κB and MAP (mitogenactivated protein) kinases and can induce some of the same target genes, a growing body of evidence points to several differences in signalling pathways activated




REVIEWS

TLR2, TLR9

TLR4

TIRAP

MyD88 TOLLIP

IRAK

MyD88 TOLLIP

IRAK

TRAF6 TRAF6

TRAF6

IL-1R

MyD88 TOLLIP

IRAK

TRAF6 TRAF6

PKR

NF-κB, JNK,p38

TAK1

MKK6

TAK1

MKK6

TAK1

MKK6

NF-κB

JNK, p38

NF-κB

JNK, p38

NF-κB

JNK, p38

Figure 4 | Toll signalling pathways. The Toll-like receptor (TLR) and interleukin-1 receptor (IL-1R)-family members share several signalling components, including the adaptor MyD88, Toll-interacting protein (TOLLIP), the protein kinase IRAK (IL-1R-associated kinase) and TRAF6 (TNF receptor-associated factor 6). TRAF6 can activate nuclear factor-κB (NF-κB) through TAK1 (TGF-βactivated kinase), and JNK (c-Jun N-terminal kinase) and p38 MAP kinases through MKK6 (mitogen-activated protein kinase kinase 6). TLR4 signals through another adaptor in addition to MyD88–TIRAP (Toll/interelukin-1 (IL-1) receptor domain-containing adaptor protein), which activates MyD88-independent signalling downstream of TLR4. The protein kinase PKR functions downstream of TIRAP, but its importance in this pathway has not yet been established.

ENDOTOXIC SHOCK

A clinical condition induced by hyper-reaction of the innate immune system to bacterial LPS. It is mediated by the inflammatory cytokines IL-1 and TNF-α, which are produced in high amounts due to sustained stimulation of TLR4 by LPS.

142

by individual TLRs. Signalling pathways activated by TLRs can be divided into ‘shared’ and ‘specific’. A shared signalling pathway is induced by all TLRs as well as by the IL-1R family. The specific pathways are activated by some, but not other TLRs, and might also account for differences in signalling between TLRs and IL-1Rs. The signalling pathway that seems to be shared by all members of the Toll and IL-1R families includes four essential components: the adaptor proteins, MyD88 (REFS 9–12) and TOLLIP (Toll-interacting protein)86,87; a protein kinase, IRAK (IL-1R-associated kinase)9,10,88; and another adaptor, TRAF6 (TNFreceptor-associated factor 6)9,10,89 (FIG. 4). The essential roles of MyD88 and TRAF6 in TLR and IL-1R signalling have been confirmed by targeted deletion of their genes90–92. MyD88 contains two protein-interaction domains: an amino-terminal death domain and a carboxy-terminal TIR domain. The TIR domain of MyD88 associates with the TIR domain of TLR and the IL-1R, whereas the death domain interacts with the amino-terminal death domain of IRAK and recruits IRAK to the receptor complex9–12. TOLLIP lacks a TIR domain, but contains a C2 domain, which in other proteins is known to interact with membrane lipids86. TOLLIP can also associate with IRAK and the TIR domains of the receptors, and recruits IRAK to the receptor complex, although with different kinetics86. The functional differences between MyD88 and TOLLIP are not yet understood. On recruitment to the receptor complex, IRAK is autophosphorylated and associates with TRAF6 (REF. 88). TRAF6 induces activation of TAK1 and MKK6 (MAP kinase kinase 6), which, in turn, activate NF-κB, JNK (c-Jun N-terminal kinase) and p38 MAP kinase, respectively93.

In addition to MyD88-dependent signalling, TLR2 has been shown to engage a signalling pathway that involves protein kinase B (PKB)94. The cytoplasmic domain of TLR2 was shown to interact with a RHO family GTPase, RAC1, and phosphatidylinositol 3-kinase (PI3K), which functions upstream of PKB94. PI3K and PKB are activated by a wide variety of cell-surface receptors and have several roles in cellular physiology. In the context of TLR2 signalling, PKB was shown to be involved in a pathway that leads to the phosphorylation of NF-κB, which is required for its transactivation activity94. It is not yet known whether this pathway is unique to TLR2, but as NF-κB phosphorylation is a necessary step in transactivation, it is likely that this pathway might be activated by all TLRs. Analysis of MyD88-deficient mice showed several unexpected features of the signalling downstream of TLR4 and TLR3. Macrophages and DCs derived from MyD88 knockout mice do not produce the cytokines IL-1β, TNF-α, IL-6 and IL-12 when stimulated with LPS, polyIC, MALP-2 or CpG, which signal through TLR4, TLR3, TLR2 and TLR9, respectively60,73,92,95. Consequently, MyD88-deficient mice are completely resistant to ENDOTOXIC SHOCK92. However, a detailed analysis of the NF-κB and MAP kinase signalling pathways has shown that LPS and polyIC, but not CpG or MALP-2, could induce activation of NF-κB, JNK and p38 in MyD88-deficient cells60,73,92,95,96. Activation of these signalling pathways through TLR4 occurred with delayed kinetics and, importantly, was insufficient for the induction of cytokine gene expression92. These unexpected findings indicated that TLR4 and TLR3 use at least two signal-transduction pathways for activation of NF-κB and MAP kinases. One of the signalling pathways is




REVIEWS

COMPLETE FREUND’S ADJUVANT

(CFA). A mixture of mycobacterial lysate with mineral oil. When animals are immunized with antigen mixed with CFA, they induce strong immune responses to the antigen.

MyD88-dependent and is used by all TLRs, whereas the other pathway is MyD88-independent and is triggered by TLR4, and possibly by TLR3, but not by TLR2 or TLR9. The IL-1 and IL-18 receptors also fail to induce signalling in the absence of MyD88, indicating that these receptors also lack the MyD88-independent signalling pathway91. Another interesting aspect of MyD88-independent signalling is that it can induce DC maturation73,96. When immature bone-marrow-derived DCs (BMDCs) are stimulated with LPS, polyIC or CpG, they produce large amounts of IL-12 and upregulate cell-surface expression of MHC and co-stimulatory molecules. MyD88-deficient BMDCs stimulated with LPS, polyIC or CpG fail to produce IL-12 or IL-6 (REFS 73,95,96). However, they can still be induced to upregulate expression of MHC and co-stimulatory molecules, such as CD80 and CD86, when treated with LPS or polyIC, but not when stimulated with CpG73,95,96. These results show that the MyD88-independent signalling pathway(s) stimulated by TLR4 and TLR3 is sufficient for DC maturation, whereas the MyD88dependent signalling pathway is required for the induction of IL-6 and IL-12 (REFS 73,96). In addition to the transcriptional events that can be induced through the MyD88-independent pathway, it has been shown that caspase-1 processing of IL-18 into its biologically active form can also be induced by TLR4 independently of MyD88 (REF. 97).

PAMP or Pathogen Pathogen

Toll

TH1

Endocytic PRR

IL-12 IFN-γ

CD28 CD80/86

MHC-II Dendritic cell

Naive T cell TCR

TH2 IL-4 IL-5 IL-13 IL-10

Figure 5 | Role of TLRs in the control of adaptive immunity. TLRs sense the presence of infection through recognition of PAMPs (pathogen-associated molecular patterns). Recognition of PAMPs by Toll-like receptors (TLRs) expressed on antigen-presenting cells (APC), such as dendritic cells, upregulates cell-surface expression of co-stimulatory (CD80 and CD86) molecules and major histocompatibility complex class II (MHC II) molecules. TLRs also induce expression of cytokines, such as interleukin-12 (IL)-12, and chemokines and their receptors, and trigger many other events associated with dendritic cell maturation. Induction of CD80/86 on APCs by TLRs leads to the activation of T cells specific for pathogens that trigger TLR signalling. IL-12 induced by TLRs also contributes to the differentiation of activated T cells into T helper (TH)1 effector cells. It is not yet known whether TLRs have any role in the induction of TH2 responses. IFN-γ; interferon-γ; PRR, pattern-recognition receptor.

Recently, a new adaptor protein TIRAP (also called MAL, for MyD88 adaptor-like) was identified and shown to function downstream of TLR4 (REFS 13,98). TIRAP has a carboxy-terminal TIR domain, but unlike MyD88, TIRAP does not have a death domain, and instead has a serine/proline-rich region of unknown function at the amino-terminus. TIRAP associates with the TIR domain of TLR4, and a dominant-negative form of TIRAP inhibits TLR4, but not TLR9 or IL-1R signalling, indicating that TIRAP controls activation of the MyD88-independent pathway13. Interestingly, TIRAP also associates with the protein kinase PKR and two PKR-regulatory proteins, PACT (PKR-activating protein) and p58, indicating that PKR functions downstream of TIRAP13. Indeed, PKR can be activated by LPS even in the absence of MyD88, indicating its involvement in the MyD88-independent pathway13. These results are consistent with a report showing impaired LPS signalling in PKR-deficient cells99. Taken together, this indicates that TLR4 uses two adaptors with TIR domains — MyD88 and TIRAP — which control activation of distinct signal-transduction pathways. TLR2 and TLR9, as well as IL-1R, use only MyD88, which accounts for differences in signalling by these receptors and TLR4 (FIG. 4)13. Tolls and control of adaptive immunity

Specificity of the TLRs for products of microbial origin allows them to signal the presence of infection and to direct the adaptive immune responses against antigens of microbial origin. DCs have a key role in coupling innate and adaptive immune-recognition systems. Immature DCs are located in peripheral tissues, including the potential pathogen-entry sites, where they can detect and capture microbial invaders 100. Not surprisingly, immature BMDCs express a full set of TLRs, which, on recognition of their ligands, induce DC maturation. Mature DCs express high levels of MHC and co-stimulatory molecules (CD80 and CD86) and migrate to draining lymph nodes where they present pathogen-derived antigens to naive T cells100. TLRs also induce expression by DCs of various cytokines, including IL-12, which directs TH cell differentiation into TH1 effector cells (FIG. 5)37. The role of Toll-mediated recognition in the control of adaptive immune responses was studied using MyD88-deficient mice. When these mice are immunized with ovalbumin mixed with COMPLETE FREUND’S ADJUVANT (CFA), they show a profound block in antigen-specific T-cell proliferation, the production of interferon-γ (IFN-γ) and the generation of ovalbumin-specific IgG2a antibodies101. These results clearly show a crucial requirement for Toll-mediated recognition in the generation of antigen-specific T H1 responses. Surprisingly, however, T H2 responses in these mice are largely unaffected under the same conditions101. So, B cells in these mice produce the same amounts of antigen-specific IgG1 and IgE as do B cells in wild-type mice, whereas T cells produce even higher amounts of IL-13 on re-stimulation with




REVIEWS antigen101. These results indicate that TLR-mediated recognition is vital for the generation of TH1, but not TH2 effector responses101. One possible explanation of these observations is that all of the known TLR ligands are products of either prokaryotic, viral or protozoan metabolism, and TH1 responses are required for protection against pathogens of these classes. TH2 responses, by contrast, are protective against multicellular eukaryotic parasites, such as helminths. These pathogens might not produce any ligands for Tolls, and perhaps are recognized by a distinct set of PRRs that could be specific for glycoproteins and glycolipids produced by worms, but not by the host or prokaryotic pathogens. Allergens also lack PAMPs that are recognized by TLRs and might initiate adaptive immune responses by a TLR-independent mechanism. It is also possible that TH2 responses might be TLR dependent, but MyD88 independent. This is less likely, however, as MyD88 is expressed constitutively in most cell types. Whichever is the case, the complete block of TH1 responses to antigen administered with CFA clearly shows that adjuvants function by triggering TLRs on DCs and other antigen-presenting cells. Indeed, MyD88-deficient DCs fail to mature and to activate naive T cells when stimulated by mycobacterial lysate, which is the active ingredient of CFA101.

1.

2.

3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

13.

144

Janeway, C. A. Jr. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb. Symp. Quant. Biol. 54, 1–13 (1989). This is a landmark paper that introduced the concepts of pattern recognition and the role of innate immune recognition in the control of adaptive immunity. Janeway, C. A. Jr. The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol. Today 13, 11–16 (1992). Medzhitov, R. & Janeway, C. A. Jr. Innate immunity: impact on the adaptive immune response. Curr. Opin. Immunol. 9, 4–9 (1997). Hashimoto, C., Hudson, K. L. & Anderson, K. V. The Toll gene of Drosophila, required for dorsal–ventral embryonic polarity, appears to encode a transmembrane protein. Cell 52, 269–279 (1988). Medzhitov, R., Preston-Hurlburt, P. & Janeway, C. A. Jr. A human homologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 388, 394–397 (1997). Rock, F. L., Hardiman, G., Timans, J. C., Kastelein, R. A. & Bazan, J. F. A family of human receptors structurally related to Drosophila Toll. Proc. Natl Acad. Sci. USA 95, 588–593 (1998). Kobe, B. & Deisenhofer, J. Proteins with leucine-rich repeats. Curr. Opin. Struct. Biol. 5, 409–416 (1995). Aravind, L., Dixit, V. M. & Koonin, E. V. Apoptotic molecular machinery: vastly increased complexity in vertebrates revealed by genome comparisons. Science 291, 1279–1284 (2001). Muzio, M., Ni, J., Feng, P. & Dixit, V. M. IRAK (Pelle) family member IRAK-2 and MyD88 as proximal mediators of IL-1 signaling. Science 278, 1612–1615 (1997). Medzhitov, R. et al. MyD88 is an adaptor protein in the hToll/IL-1 receptor family signaling pathways. Mol. Cell 2, 253–258 (1998). Wesche, H., Henzel, W. J., Shillinglaw, W., Li, S. & Cao, Z. MyD88: an adapter that recruits IRAK to the IL-1 receptor complex. Immunity 7, 837–847 (1997). Burns, K. et al. MyD88, an adapter protein involved in interleukin-1 signaling. J. Biol. Chem. 273, 12203–12209 (1998). Horng, T., Barton, G. M. & Medzhitov, R. TIRAP: an adapter molecule in the Toll signaling pathway. Nature Immunol. 2, 835–841 (2001).

Perspectives

The identification and functional characterization of TLRs in Drosophila and mammals have brought our understanding of the innate immune system to a new level. The role of the TLRs in host defence is so fundamental, it is likely that their function affects most aspects of the mammalian immune system. Loss-offunction mutations in TLRs are likely to result in immunodeficiencies, whereas gain-of-function mutations might predispose an individual to inflammatory or autoimmune disorders. The importance of the TLRs in the control of adaptive immune responses also makes them crucial targets for immune intervention. Although much progress has been made in the characterization of individual TLRs, there are many more fundamental questions to address: what are the full compliment of PAMPs and other ligands recognized by TLRs? What are the differences between individual TLRs in the induction of cellular and immune responses? What is the mechanism of ligand recognition by TLRs? Can TLRs detect any features of pathogens that are important for the choice of effector responses? What is the biological significance of differential TLR expression? And why are TLRs not continuously activated by commensal microflora? The answers to these questions will greatly expand our understanding of the complex interactions between pathogens and the host immune response.

14. Anderson, K. V. Toll signaling pathways in the innate immune response. Curr. Opin. Immunol. 12, 13–19 (2000). 15. Belvin, M. P. & Anderson, K. V. A conserved signaling pathway: the Drosophila Toll-dorsal pathway. Annu. Rev. Cell Dev. Biol. 12, 393–416 (1996). 16. Lemaitre, B., Nicolas, E., Michaut, L., Reichhart, J. M. & Hoffmann, J. A. The dorsoventral regulatory gene cassette spätzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 86, 973–983 (1996). This is a seminal study that showed the role of the Toll pathway in Drosophila immunity. 17. Meng, X., Khanuja, B. S. & Ip, Y. T. Toll receptor-mediated Drosophila immune response requires Dif, an NF-κB factor. Genes Dev. 13, 792–797 (1999). 18. Rutschmann, S. et al. The Rel protein DIF mediates the antifungal but not the antibacterial host defense in Drosophila. Immunity 12, 569–580 (2000). 19. Levashina, E. A. et al. Constitutive activation of Tollmediated antifungal defense in serpin-deficient Drosophila. Science 285, 1917–1919 (1999). 20. Khush, R. S., Leulier, F. & Lemaitre, B. Drosophila immunity: two paths to NF-κB. Trends Immunol. 22, 260–264 (2001). 21. Lemaitre, B., Reichhart, J. M. & Hoffmann, J. A. Drosophila host defense: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms. Proc. Natl Acad. Sci. USA 94, 14614–14619 (1997). This study indicated that Drosophila can discriminate between different pathogen classes and induce an appropriate set of antimicrobial peptides. 22. Lemaitre, B. et al. A recessive mutation, immune deficiency (imd), defines two distinct control pathways in the Drosophila host defense. Proc. Natl Acad. Sci. USA 92, 9465–9469 (1995). 23. Imler, J. L. & Hoffmann, J. A. Signaling mechanisms in the antimicrobial host defense of Drosophila. Curr. Opin. Microbiol. 3, 16–22 (2000). 24. Georgel, P. et al. Drosophila immune deficiency (IMD) is a death domain protein that activates antibacterial defence and can promote apoptosis. Dev. Cell 1, 503–514 (2001) 25. Elrod-Erickson, M., Mishra, S. & Schneider, D. Interactions between the cellular and humoral immune responses in Drosophila. Curr. Biol. 10, 781–784 (2000).


26. Leulier, F., Rodriguez, A., Khush, R. S., Abrams, J. M. & Lemaitre, B. The Drosophila caspase Dredd is required to resist Gram-negative bacterial infection. EMBO Rep. 1, 353–358 (2000). 27. Lu, Y., Wu, L. P. & Anderson, K. V. The antibacterial arm of the Drosophila innate immune response requires an IκB kinase. Genes Dev. 15, 104–110 (2001). 28. Silverman, N. et al. A Drosophila IκB kinase complex required for relish cleavage and antibacterial immunity. Genes Dev. 14, 2461–2471 (2000). 29. Rutschmann, S. et al. Role of Drosophila IKK-γ in a Tollindependent antibacterial immune response. Nature Immunol. 1, 342–347 (2000). 30. Vidal, S. et al. Mutations in the Drosophila dTAK1 gene reveal a conserved function for MAPKKKs in the control of rel/NF-κB-dependent innate immune responses. Genes Dev. 15, 1900–1912 (2001). 31. Hedengren, M. et al. Relish, a central factor in the control of humoral but not cellular immunity in Drosophila. Mol. Cell 4, 827–837 (1999). 32. Chen, P., Rodriguez, A., Erskine, R., Thach, T. & Abrams, J. M. Dredd, a novel effector of the apoptosis activators reaper, grim, and hid in Drosophila. Dev. Biol. 201, 202–216 (1998). 33. Karin, M. & Delhase, M. The IκB kinase (IKK) and NF-κB: key elements of proinflammatory signalling. Semin. Immunol. 12, 85–98 (2000). 34. Dushay, M. S., Asling, B. & Hultmark, D. Origins of immunity: Relish, a compound Rel-like gene in the antibacterial defense of Drosophila. Proc. Natl Acad. Sci. USA 93, 10343–10347 (1996). 35. Tauszig, S., Jouanguy, E., Hoffmann, J. A. & Imler, J. L. From the cover: Toll-related receptors and the control of antimicrobial peptide expression in Drosophila. Proc. Natl Acad. Sci. USA 97, 10520–10525 (2000). 36. Williams, M. J., Rodriguez, A., Kimbrell, D. A. & Eldon, E. D. The 18-wheeler mutation reveals complex antibacterial gene regulation in Drosophila host defense. EMBO J. 16, 6120–6130 (1997). 37. Akira, S., Takeda, K. & Kaisho, T. Toll-like receptors: critical proteins linking innate and acquired immunity. Nature Immunol. 2, 675–680 (2001). 38. Poltorak, A. et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282, 2085–2088 (1998).



REVIEWS 39. Qureshi, S. T. et al. Endotoxin-tolerant mice have mutations in Toll-like receptor 4 (Tlr4). J. Exp. Med. 189, 615–625 (1999); erratum 189, 1518 (1999). 40. Hoshino, K. et al. Cutting edge: Toll-like receptor 4 (TLR4)deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product. J. Immunol. 162, 3749–3752 (1999). References 38–40 describe the first indication of TLR4 function in vivo. 41. Wright, S. D., Tobias, P. S., Ulevitch, R. J. & Ramos, R. A. Lipopolysaccharide (LPS) binding protein opsonizes LPSbearing particles for recognition by a novel receptor on macrophages. J. Exp. Med. 170, 1231–1241 (1989). 42. Wright, S. D., Ramos, R. A., Tobias, P. S., Ulevitch, R. J. & Mathison, J. C. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249, 1431–1433 (1990). 43. Haziot, A. et al. Resistance to endotoxin shock and reduced dissemination of Gram-negative bacteria in CD14-deficient mice. Immunity 4, 407–414 (1996). 44. Shimazu, R. et al. MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like receptor 4. J. Exp. Med. 189, 1777–1782 (1999). 45. Schromm, A. B. et al. Molecular genetic analysis of an endotoxin nonresponder mutant cell line. A point mutation in a conserved region of md-2 abolishes endotoxininduced signaling. J. Exp. Med. 194, 79–88 (2001). 46. Lien, E. et al. Toll-like receptor 4 imparts ligand-specific recognition of bacterial lipopolysaccharide. J. Clin. Invest. 105, 497–504 (2000). 47. Poltorak, A., Ricciardi-Castagnoli, P., Citterio, S. & Beutler, B. Physical contact between lipopolysaccharide and Toll-like receptor 4 revealed by genetic complementation. Proc. Natl Acad. Sci. USA 97, 2163–2167 (2000). 48. Da Silva Correia, J., Soldau, K., Christen, U., Tobias, P. S. & Ulevitch, R. J. Lipopolysaccharide is in close proximity to each of the proteins in its membrane receptor complex. Transfer from CD14 to TLR4 and MD-2. J. Biol. Chem. 276, 21129–21135 (2001). 49. Miyake, K., Yamashita, Y., Ogata, M., Sudo, T. & Kimoto, M. RP105, a novel B cell surface molecule implicated in B cell activation, is a member of the leucine-rich repeat protein family. J. Immunol. 154, 3333–3340 (1995). 50. Chan, V. W. et al. The molecular mechanism of B cell activation by Toll-like receptor protein RP-105. J. Exp. Med. 188, 93–101 (1998). 51. Miyake, K. et al. Mouse MD-1, a molecule that is physically associated with RP105 and positively regulates its expression. J. Immunol. 161, 1348–1353 (1998). 52. Ogata, H. et al. The Toll-like receptor protein RP105 regulates lipopolysaccharide signaling in B cells. J. Exp. Med. 192, 23–29 (2000). 53. Takeuchi, O. et al. Differential roles of TLR2 and TLR4 in recognition of Gram-negative and Gram-positive bacterial cel wall components. Immunity 11, 443–451 (1999). 54. Means, T. K. et al. Human Toll-like receptors mediate cellular activation by Mycobacterium tuberculosis. J. Immunol. 163, 3920–3927 (1999). 55. Ohashi, K., Burkart, V., Flohe, S. & Kolb, H. Cutting edge: heat shock protein 60 is a putative endogenous ligand of the Toll-like receptor-4 complex. J. Immunol. 164, 558–561 (2000). 56. Li, M. et al. An essential role of the NF-κB/Toll-like receptor pathway in induction of inflammatory and tissue-repair gene expression by necrotic cells. J. Immunol. 166, 7128–7135 (2001). 57. Kurt-Jones, E. A. et al. Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus. Nature Immunol. 1, 398–401 (2000). 58. Bowie, A. et al. A46R and A52R from vaccinia virus are antagonists of host IL-1 and Toll-like receptor signaling. Proc. Natl Acad. Sci. USA 97, 10162–10167 (2000). 59. Schwandner, R., Dziarski, R., Wesche, H., Rothe, M. & Kirschning, C. J. Peptidoglycan- and lipoteichoic acidinduced cell activation is mediated by Toll-like receptor 2. J. Biol. Chem. 274, 17406–17409 (1999). 60. Takeuchi, O. et al. Cutting edge: preferentially the Rstereoisomer of the mycoplasmal lipopeptide macrophage-activating lipopeptide-2 activates immune cells through a Toll-like receptor 2- and MyD88dependent signaling pathway. J. Immunol. 164, 554–557 (2000). 61. Aliprantis, A. O. et al. Cell activation and apoptosis by bacterial lipoproteins through Toll-like receptor-2. Science 285, 736–739 (1999).

62. Brightbill, H. D. et al. Host defense mechanisms triggered by microbial lipoproteins through Toll-like receptors. Science 285, 732–736 (1999). 63. Means, T. K. et al. The CD14 ligands lipoarabinomannan and lipopolysaccharide differ in their requirement for Toll-like receptors. J. Immunol. 163, 6748–6755 (1999). 64. Underhill, D. M., Ozinsky, A., Smith, K. D. & Aderem, A. Toll-like receptor-2 mediates mycobacteria-induced proinflammatory signaling in macrophages. Proc. Natl Acad. Sci. USA 96, 14459–14463 (1999). 65. Campos, M. A. et al. Activation of Toll-like receptor-2 by glycosylphosphatidylinositol anchors from a protozoan parasite. J. Immunol. 167, 416–423 (2001). 66. Hajjar, A. M. et al. Cutting edge: functional interactions between Toll-like receptor (TLR) 2 and TLR1 or TLR6 in response to phenol-soluble modulin. J. Immunol. 166, 15–19 (2001). 67. Underhill, D. et al. The Toll-like receptor 2 is recruited to macrophage phagosomes and discriminates between pathogens. Nature 401, 811–815 (1999). 68. Werts, C. et al. Leptospiral lipopolysaccharide activates cells through a TLR2-dependent mechanism. Nature Immunol. 2, 346–352 (2001). 69. Hirschfeld, M. et al. Signaling by Toll-like receptor 2 and 4 agonists results in differential gene expression in murine macrophages. Infect. Immun. 69, 1477–1482 (2001). 70. Ozinsky, A. et al. The repertoire for pattern recognition of pathogens by the innate immune system is defined by cooperation between Toll-like receptors. Proc. Natl Acad. Sci. USA 97, 13766–13771 (2000). 71. Takeuchi, O. et al. Discrimination of bacterial lipoproteins by Toll-like receptor 6. Int. Immunol. 13, 933–940 (2001). 72. Muzio, M. et al. Differential expression and regulation of Toll-like receptors (TLR) in human leukocytes: selective expression of TLR3 in dendritic cells. J. Immunol. 164, 5998–6004 (2000). 73. Alexopoulou, L., Czopik-Holt, A., Medzhitov, R. & Flavell, R. Recognition of double stranded RNA and activation of NFκB by Toll-like receptor 3. Nature 413, 696–712 (2001). 74. Williams, B. R. PKR; a sentinel kinase for cellular stress. Oncogene 18, 6112–6120 (1999). 75. Chu, W. M. et al. JNK2 and IKKβ are required for activating the innate response to viral infection. Immunity 11, 721–731 (1999). 76. Hayashi, F. et al. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410, 1099–1103 (2001). 77. Samatey, F. A. et al. Structure of the bacterial flagellar protofilament and implications for a switch for supercoiling. Nature 410, 331–337 (2001). 78. Gewirtz, A. T., Navas, T. A., Lyons, S., Godowski, P. J. & Madara, J. L. Cutting edge: bacterial flagellin activates basolaterally expressed TLR5 to induce epithelial proinflammatory gene expression. J. Immunol. 167, 1882–1885 (2001). 79. Hemmi, H. et al. A Toll-like receptor recognizes bacterial DNA. Nature 408, 740–745 (2000). 80. Krieg, A. M. et al. CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374, 546–549 (1995). 81. Hacker, H. et al. CpG-DNA-specific activation of antigen-presenting cells requires stress kinase activity and is preceded by non-specific endocytosis and endosomal maturation. EMBO J. 17, 6230–6240 (1998). 82. Thieblemont, N. & Wright, S. D. Transport of bacterial lipopolysaccharide to the Golgi apparatus. J. Exp. Med. 190, 523–534 (1999). 83. Krieg, A. M. The role of CpG motifs in innate immunity. Curr. Opin. Immunol. 12, 35–43 (2000). 84. Bauer, S. et al. Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proc. Natl Acad. Sci. USA 98, 9237–9242 (2001). 85. Thoma-Uszynski, S. et al. Induction of direct antimicrobial activity through mammalian Toll-like receptors. Science 291, 1544–1547 (2001). 86. Burns, K. et al. Tollip, a new component of the IL-1RI pathway, links IRAK to the IL-1 receptor. Nature Cell Biol. 2, 346–351 (2000). 87. Bulut, Y., Faure, E., Thomas, L., Equils, O. & Arditi, M. Cooperation of Toll-Like receptor 2 and 6 for cellular activation by soluble tuberculosis factor and Borrelia burgdorferi outer surface protein A lipoprotein: role of Toll-interacting protein and IL-1 receptor signaling molecules in Toll-like receptor 2 signaling. J. Immunol. 167, 987–994 (2001).


88. Cao, Z., Henzel, W. J. & Gao, X. IRAK: a kinase associated with the interleukin-1 receptor. Science 271, 1128–1131 (1996). 89. Cao, Z., Xiong, J., Takeuchi, M., Kurama, T. & Goeddel, D. V. TRAF6 is a signal transducer for interleukin-1. Nature 383, 443–446 (1996). 90. Lomaga, M. A. et al. TRAF6 deficiency results in osteopetrosis and defective interleukin-1, CD40, and LPS signaling. Genes Dev. 13, 1015–1024 (1999). 91. Adachi, O. et al. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18-mediated function. Immunity 9, 143–150 (1998). 92. Kawai, T., Adachi, O., Ogawa, T., Takeda, K. & Akira, S. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity 11, 115–122 (1999). 93. Wang, C. et al. TAK1 is a ubiquitin-dependent kinase of MKK and IKK. Nature 412, 346–351 (2001). 94. Arbibe, L. et al. Toll-like receptor 2-mediated NF-κB activation requires a Rac1-dependent pathway. Nature Immunol. 1, 533–540 (2000). 95. Schnare, M., Holt, A. C., Takeda, K., Akira, S. & Medzhitov, R. Recognition of CpG DNA is mediated by signaling pathways dependent on the adaptor protein MyD88. Curr. Biol. 10, 1139–1142 (2000). 96. Kaisho, T., Takeuchi, O., Kawai, T., Hoshino, K. & Akira, S. Endotoxin-induced maturation of Myd88-deficient dendritic cells. J. Immunol. 166, 5688–5694 (2001). 97. Seki, E. et al. Lipopolysaccharide-induced IL-18 secretion from murine Kupffer cells independently of myeloid differentiation factor 88 that is critically involved in induction of production of IL-12 and IL-1β. J. Immunol. 166, 2651–2657 (2001). 98. Fitzgerald, K. A. et al. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction. Nature 413, 78–83 (2001). 99. Goh, K. C., deVeer, M. J. & Williams, B. R. The protein kinase PKR is required for p38 MAPK activation and the innate immune response to bacterial endotoxin. EMBO J. 19, 4292–4297 (2000). 100. Banchereau, J. & Steinman, R. M. Dendritic cells and the control of immunity. Nature 392, 245–252 (1998). 101. Schnare, M. et al. Toll-like receptors control activation of adaptive immune responses. Nature Immunol. 2, 947–950 (2001). 102. Holmskov, U. L. Collectins and collectin receptors in innate immunity. APMIS Suppl. 100, 1–59 (2000). 103. Gewurz, H., Mold, C., Siegel, J. & Fiedel, B. C-reactive protein and the acute phase response. Adv. Intern. Med. 27, 345–372 (1982). 104. Schwalbe, R. A., Dahlback, B., Coe, J. E. & Nelsestuen, G. L. Pentraxin family of proteins interact specifically with phosphorylcholine and/or phosphorylethanolamine. Biochemistry 31, 4907–4915 (1992). 105. Fraser, I. P., Koziel, H. & Ezekowitz, R. A. The serum mannose-binding protein and the macrophage mannose receptor are pattern recognition molecules that link innate and adaptive immunity. Semin. Immunol. 10, 363–372 (1998). 106. Pearson, A. M. Scavenger receptors in innate immunity. Curr. Opin. Immunol. 8, 20–28 (1996). 107. Elomaa, O. et al. Cloning of a novel bacteria-binding receptor structurally related to scavenger receptors and expressed in a subset of macrophages. Cell 80, 603–609. (1995). 108. Inohara, N. et al. Nod1, an Apaf-1-like activator of caspase-9 and nuclear factor-κB. J. Biol. Chem. 274, 14560–14567 (1999). 109. Bertin, J. et al. Human CARD4 protein is a novel CED4/Apaf-1 cell death family member that activates NF–κB. J. Biol. Chem. 274, 12955–12958 (1999).

Online links DATABASES The following terms in this article are linked online to: Flybase: http://flybase.bio.indiana.edu/ 18 Wheeler | Cactus | Dif | Diptericin | Dorsal | Dredd | Drosomycin | dIKK-β | dIKK-γ | imd | necrotic | Pelle | Relish | Spätzle | dTAK1 | Toll | Toll-5 | Tube LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ CD14 | CD80 | CD86 | HSP60 | IFN-γ | IL-1β | IL-1R | IL-6 | IL-10 | IL-13 | IL-18R | IRAK | LBP | LYN | MD-1 | MD-2 | MKK6 | MyD88 | p38 MAP kinase | p50 | p58 | p65 | p100 | p105 | PACT | PKR | RAC1 | RP105 | TAK1 | TGF-β | TIRAP | TLR1 | TLR2 | TLR3 | TLR4 | TLR5 | TLR6 | TLR9 | TNF-α | TOLLIP | TRAF6 Access to this interactive links box is free online.



PERSPECTIVES OPINION

From T to B and back again: positive feedback in systemic autoimmune disease

the amplification of activated autoreactive lymphocytes, details of which will be discussed in turn. Finally, we examine how these autoreactive lymphocytes contribute to end-organ damage, emphasizing the new idea that T cells contribute directly to pathogenesis, in addition to their roles in helping B cells make autoantibodies (FIG. 1d,i). Central and peripheral tolerance

Mark J. Shlomchik, Joseph E. Craft and Mark J. Mamula Systemic lupus erythematosus, a prototypical systemic autoimmune disease, is the result of a series of interactions within the immune system that ultimately lead to the loss of self-tolerance to nuclear autoantigens. Here, we present an integrated model that explains how self-tolerance is initially lost and how the loss of tolerance is then amplified and maintained as a chronic autoimmune state. Key to this model are the selfreinforcing interactions of T and B cells, which we suggest lead to perpetuation of autoimmunity as well as its spread to multiple autoantigen targets.

In systemic lupus erythematosus (SLE), T- and B-cell autoimmune responses result in the generation of autoantibodies and immune complexes, along with autoreactive T cells, which together cause pathology in several target organs, including skin, blood vessels, lung and kidney. Disease is the result of a cascade of events that occur on the background of an appropriate genetic predisposition1–4. The importance of stochastic or environmentally driven events that are subsequently amplified is emphasized by the fact that even inbred lupus-prone mice vary considerably in the time of onset, the severity and the pathology elicited by the autoimmune response5. Clinical disease is accompanied by a diverse B- and T-cell autoimmune response that targets specific autoantigens, such as nuclear constituents6,7. As SLE is a

multigenic and age-dependent disease, the final disease phenotype is probably the result of many interactions arising from an initial loss of peripheral tolerance followed by the amplification of specific autoimmune responses. Although individual elements of these interactions are often discussed and studied in isolation2,8,9, we have integrated these concepts into a model that explains the onset and development of chronic lupus autoimmunity. Here, we will focus on the interactions within the immune system that lead to a loss of self-tolerance to lupus autoantigens, ultimately leading to the pathologic processes of systemic autoimmune disease. As the genetic loci that contribute to systemic autoimmunity in mouse models are being elucidated2, it will be important to have a cellular and biological framework from which to understand their roles in promoting autoimmune disease. Evolution of systemic autoimmunity

We present a model outlining a series of cellular interactions, guided by autoantigen recognition, that accounts for both the initial loss of tolerance to intracellular macromolecules and subsequent events leading to full-blown autoimmune disease (FIG. 1). The model emphasizes the loss of tolerance in the periphery and events thereafter, as we believe that a loss of peripheral tolerance, rather than of central tolerance, initiates autoimmunity. A key feature of this model is the role of the T-cell/B-cell interaction in


The presence of self-reactive lymphocytes is a prerequisite for autoimmunity. Early efforts to identify the origins of self-reactive cells focused on defects that might arise in central mechanisms of lymphocyte selection in the bone marrow and thymus. This was a logical first step because the targets of lupus autoimmunity are ubiquitous in nature, as opposed to sequestered in peripheral tissues. However, it is now clear that in many situations central lymphocyte selection is similar in normal mice and in mouse strains that develop spontaneous autoimmunity10–15. Although this issue is still controversial — central-tolerance defects could have a role in some situations16, particularly Fas (CD95) deficiency17,18 — we emphasize here the loss of tolerance in the periphery, which we believe is the more crucial step. Despite intact central tolerance that eliminates high-affinity anti-self lymphocytes, lower-affinity self-reactive B and T cells, including those specific for lupus autoantigens, develop, and are easily identified in the normal peripheral immune repertoire19–22. These self-reactive B- and T-cell responses have also been found by immunization of

“...even inbred lupus-prone mice vary considerably in the time of onset, the severity and the pathology elicited by the autoimmune response.” VOLUME 1 | NOVEMBER 2001 | 1 4 7


PERSPECTIVES

Normal

SLE

c

d Affinity maturation APC

Organ damage

b a Tolerance

Activation by self-peptide or molecular mimic Auto-antigen release Cytokines and cytolysis

e g

i f

h

Epitope spreading

Figure 1 | Mechanisms for the induction and amplification of lupus autoimmunity. a | Although normal T cells exposed to self-antigen in the periphery become tolerized, lupus-prone T cells are sensitive to lower thresholds of activation by agonist or weak-agonist peptides. b | Once activated, T cells can provide primary stimulation to genetically hyper-responsive B cells. c |These autoantigen-stimulated B cells undergo somatic hypermutation and affinity maturation. d | On the synthesis of pathogenic autoantibodies, tissue damage results in the release of self-antigen, e,f | which is also taken up and presented by specific antigen-presenting B cells in a second round of T-cell activation, g | therefore leading to a positive-feedback cycle. h | Autoimmune T- and B-cell responses are diversified, which results in epitope spreading. This continuing and cyclic process of B cell–T cell cognate interaction serves to amplify the ensuing autoimmune processes. i | Activated T cells can also directly cause tissue pathology by migrating to the target organ and releasing cytokines and by mediating direct cytotoxicity. APC, antigen-presenting cell. T cells are shown orange; B cells are red.

animals with random peptide sequences of self-antigens, an approach that has identified a variety of cryptic self-peptides to which central selection of lymphocytes has failed23,24. Other studies have shown that the naive peripheral repertoire of both normal humans and animal models can be stimulated with immunodominant autoantigens25,26. A recent study27 has shown that nearly 40% of transgenic self-reactive T cells can escape negative selection in the thymus, even in the presence of the tolerance-inducing ligand. These observations illustrate that central tolerance is incomplete, and probably affects only the highest affinity self-reactive lymphocytes. They also highlight that potentially pathogenic lymphocytes are actively regulated

148

and prevented from effector functions in the periphery. As described below, the altered threshold for T-cell activation in SLE might allow normally benign (cryptic) self-peptides to initiate autoimmunity. Similarly, self-reactive B cells are easily found among the circulating repertoire of normal humans and animals19–21,28–32. These B cells are either quiescent owing to the lack of appropriate T-cell help or are subject to regulatory mechanisms, including the induction of anergy, activation-induced cell death (AICD) and signalling for death by Fas and the tumour-necrosis factor (TNF) receptor30–34. Regulatory T cells, including the CD25+ subset, have been shown to keep organ-specific autoimmunity in check. They might also have


a role in preventing or regulating systemic autoimmunity, although there is no direct evidence on this point as yet. There is no doubt that these mechanisms are important in regulating self-reactive B cells, either directly or indirectly through regulation of T-cell help, as genetic defects in molecules that function in these peripheral tolerance mechanisms lead to autoimmunity, for example in Fas, TNF, cytotoxic T-lymphocyte antigen 4 (CTLA4/CD152) and transforming growth factor-β (TGF-β)35–39 (TABLE 1). In addition, overexpression of positive regulators of the TNF family (for example, B-lymphocyte stimulator; BLyS) and the B7 family (B7related protein 1; B7RP1) also lead to autoimmune-like syndromes40–42.



PERSPECTIVES Although such global perturbations lead to autoimmunity, the loss of self-tolerance is not generalized, but instead is selective, targeting several key autoantigens. Autoimmunity in SLE is focused on highly conserved intracellular macromolecules, including nucleosomes and ribonucleoproteins7. Evidence from B-cell-receptor (BCR) transgenic-mouse systems, as well as analysis of isolated autoreactive B cells, indicates that the initial loss of selftolerance occurs in cells with low, but relevant, affinities for self 43–45. This further supports the idea that central tolerance remains intact and continues to prevent the development of high-affinity self-reactive cells. In the B-cell compartment, there is direct evidence from transgenic models that the cells that initiate autoimmunity are selfreactive, although they might generally be of low affinity43,46. Such B cells might be stimulated directly by self-antigen 43,46, or by a crossreactive epitope on an infecting or environmental antigen47, and then sustained by self-antigen. Somatic hypermutation and selection can result in the generation of high-affinity B cells from low-affinity precursors, provided that T-cell help is available48–53 (FIG. 1c). Interestingly, the T cells that help B cells need not initially be selfreactive. Certain autoreactive B cells might be able to present foreign antigens that complex with self-antigens, and thereby get help from foreign-antigen-reactive T cells54. This could occur, for example, when a viral DNA-binding protein is taken up and presented by an anti-DNA B cell47. Yet autoreactive T cells are also activated in systemic autoimmunity. These cells are vital for the process, in particular to sustain clones of autoreactive B cells55–59 (FIG. 1b,e). The role of self-antigen

As self-antigens are required to drive autoreactive T and B cells at some stage of disease, the question is raised: how can the same selfantigens cause active tolerance or ignorance in some circumstances, and autoimmunity in others? A lymphocyte’s interpretation of self-antigen can, in fact, be modulated in many ways. The presence of inflammation, the release of cytokines and the consequent activation of antigen-presenting cells (APCs) could modulate the sensitivity of peripheral T or B cells to a given concentration of autoantigen60. In addition, defects in mechanisms of clearing apoptotic cells or immune complexes, such as complement, Fc receptor or phagocytic defects, serve to elevate the absolute concentration of available autoantigen61–63. Indeed, such mutations are linked to higher incidence of lupus in

humans64. Evidence from knockout mice65,66 and screens for lupus-predisposing genes67–69 also implicate this mechanism. By aggregating, either as an immune complex or autonomously, a self-antigen would also increase its ability to crosslink BCRs. Similarly, self-antigen in the form of immune complexes would be more efficiently taken up by dendritic cells (DCs) and presented to T cells70. In addition, altered forms of selfantigen might trigger the activation of previously quiescent self-reactive lymphocytes. For example, two post-translational modifications that arise spontaneously within cells — dimethylarginine and isoaspartyl additions to self-proteins — can initiate both autoantibodies and autoreactive T-cell

responses71–73. Overall, it is clear that a source of self-antigen that fails to drive negative selection is nonetheless required to stimulate autoimmunity. Although the exact source of these autoantigens is not known, it has been proposed that apoptotic cells might display aberrant forms of nuclear autoantigens74. Intrinsic defects in signal interpretation

Autoreactive lymphocytes that are found in the periphery are usually quiescent in normal individuals. In humans and mice with lupus, genetic defects might result in heightened sensitivity of lymphocytes to stimulation by self-antigen and/or a heightened or prolonged response subsequent to activation (FIG. 1a)75–82.

Table 1 | Lupus-prone natural, transgenic and knockout mouse strains Mouse strain

Phenotype

MRL/Faslpr

Systemic autoimmunity and lymphoproliferation with autoantibodies resembling human SLE; Fas gene is mutated (lpr allele)

References

(NZBxNZW) F1

A lupus-prone animal that is an F1 hybrid of two strains, each with mild symptoms of immunological hyperactivity or autoimmunity; prominent glomerulonephritis and anti-nuclear antibodies

NZM2410

NZM inbred strain of mouse including genetic contributions from NZB and NZW; has a penetrant lupus-like syndrome with glomerulonephritis; has been used extensively to map the genes responsible

1,2

BXSB

A lupus-prone strain in which only the males have disease, owing to the presence of the Yaa gene on the Y chromosome, which confers B-cell hyperactivity

77

5,10–12,14

9,13,107

MRL/Faslpr/Tnfr1–/–

Accelerated systemic autoimmunity and lympho-proliferation

36

CD152–/–

Early-onset massive lymphoproliferation and infiltrative tissue destruction

37

TGF-β–/–

Early-onset infiltrative tissue destruction and autoantibodies

PD-1–/–

Glomerulonephritis, arthritis, hyper-responsive B cells

BLyS/transgenic (overexpressor)

B-cell hyperplasia, hypergammaglobulinaemia, autoantibodies

40,41

B7RP-Fc transgenic

B-cell hyperplasia, hypergammaglobulinaemia (overexpressor)

42

MER–/–

Knockout of a protein tyrosine kinase gene expressed in phagocytes; defect in clearing apoptotic cells by phagocytosis; elevated anti-nuclear antibodies

63

C1q–/–

Glomerulonephritis, elevated autoantibodies

C4–/– and CD21/35–/–

Increased autoantibodies and nephritis

lpr

–/–

lpr

MRL/Fas /Jh

B-cell deficient version of MRL/Fas ; no spontaneous lupus phenotype; no T-cell activation or infiltration

MRL/Faslpr/mIgM transgenic

Version of MRL/Faslpr that has B cells but no circulating antibodies; restores T-cell activation and infiltration

MRL/Faslpr/B7-1–/– or B7-2–/–

Version of MRL/Faslpr that lacks either B7-1 (CD80) or B7-2 (CD86); disease is only slightly, if at all, reduced

MRL/Faslpr 2–12 (anti-Sm) Ig transgenic

Version of MRL/Faslpr that harbours a transgene that vastly increases the frequency of B cells specific for the Sm autoantigen; used to study regulation of B cells and how they affect T-cell tolerance

38,39 97

61,65 66 85,86 87 94,95 55

*This is not meant to be an exhaustive list, and only those models mentioned in the text are highligted. B7RP, B7-related protein; BLyS, B-lymphocyte stimulator; C1q, complement component 1, q subcomponent; C4, complement component 4; mIgM, membrane immunoglobulin M; NZB, New Zealand Black; NZM, New Zealand Mixed; NZW, New Zealand White; PD-1, programmed cell death 1; SLE, systemic lupus erythematosus; Sm, Smith; TGF-β; transforming growth factor-β; Tnfr1, tumour-necrosis factor receptor 1.




PERSPECTIVES Indeed, mapping studies of the lupus-prone New Zealand Mixed (NZM) mouse have identified chromosomal loci that segregate with enhanced T-cell activation and/or proliferation and reduced AICD among CD4+ T cells78 as well as B-cell hyperactivity76. Recent studies by Craft and co-workers79 have shown differences in TCR signal interpretation between T cells from lupus-prone and control mice. By expressing cytochrome c-specific TCR transgenes in lupus-prone and normal mice, and stimulating with highor low-affinity peptide ligands, it was found that peripheral T cells in lupus-prone mice have a significantly lower threshold for T-cell activation as compared with normal mice. Low-affinity ligands for TCR stimulated proliferation and enhanced interleukin-2 (IL-2) production and activation-marker expression in T cells of lupus-prone mice, but not control mice. Indeed, this intrinsic T-cell abnormality might ultimately lead to enhanced helper functions for B-cell autoimmunity (FIG. 1b). Similar abnormalities might exist for B cells75,76,80, although this is less well understood. Exactly how intrinsic lymphocyte signalling abnormalities predispose to SLE is unclear. However, we speculate that endogenous TCR ligands that normally signal homeostatic maintenance of peripheral T cells provide activation signals in a lupusprone environment. This could help explain the source of initial T-cell activation. In the context of the positive-feedback model, we propose that even small differences in initial T- or B-cell activation could, over a period of time, be amplified and have important consequences in determining clinical disease versus sustained self-tolerance . T-cell/B-cell interactions

Once self-reactive lymphocytes are aberrantly activated in SLE, this process must be expanded and amplified in order to generate chronic, sustained autoimmune disease (FIG. 1b–h). Although DCs and perhaps macrophages might have a role, particularly in the initial activation of autoreactive T cells, much evidence has implicated cognate T-cell/B-cell interactions as vital in this process. Early studies that eliminated T-cell populations using cytotoxic antibodies or thymectomy illustrated the importance of cellular immunity in diseases marked by autoantibody synthesis, at a time when B-cell disregulation was thought to be the key factor53,83. Subsequently, genetic knockouts of T or B cells showed a requirement for both cell subsets for the development of SLE51,84–86. Advancing the concept further, Chan et al.87 constructed MRL/Fas lpr mice (TABLE 1) in which B cells cannot secrete

150

“T-cell/B-cell interactions constitute a positivefeedback loop that enables diversification and continuous amplification of the autoimmune response.” immunoglobulin, and so have virtually no circulating autoantibodies. This model addressed the possibility that T-cell autoimmunity was a secondary response to autoantibody-mediated tissue pathology. These mice experienced a spontaneous activation of T cells, similar to that found in non-transgenic MRL/Fas lpr autoimmune mice. Therefore, soluble antibody is not required for T-cell activation. By contrast, neither MRL/Fas lpr nor MRL/Fas+/+ mice made deficient in B cells develop activated T-cell subsets86,88. Intriguingly, in the NZM2410 model, Mohan et al.78 have identified a locus on chromosome 7 (Sle3) that affects several parameters of T-cell activation, yet leads to the accumulation of activated B cells as well. Similarly, the same group showed that lupus-associated genes from another locus, Sle1, are expressed in B cells and mediate loss of B-cell tolerance to chromatin antigens, but can also act non-cell-autonomously to cause loss of T-cell tolerance to chromatin antigens89. Taken together, these studies support an active role for B cells in the initiation and diversification of autoreactive T cells in SLE, as earlier proposed by Mamula and Janeway90,91 and by others92,93. Defining APC–T-cell interactions will be crucial to understanding the early events in lupus autoimmunity. For example, activated autoreactive B cells are potent APCs for their cognate auto-antigen24,55,56; we hypothesize that they are capable of breaking peripheral T-cell tolerance by causing the initial activation of ignorant or even anergic autoreactive T cells. In addition, these B cells could further stimulate autoreactive T cells that had already been activated by DCs. In any event, T cells that provide help for autoantibody synthesis in SLE might not be subject to the same threshold of antigen stimulation or indeed the same necessity for co-stimulation as is typically required for immune responses to foreign antigen. Lupus-prone MRL/Faslpr mice deficient in either B7-1 (CD80) or B7-2 (CD86) resemble B7-sufficient autoimmune mice with regard to their levels of of autoantibody production, spontaneous activation of T cells and kidney pathology94,95.


These observations indicate that naive autoreactive T cells in lupus-prone mice might be sensitive to stimulation by B cells or other APCs not normally considered to be ‘professional’ APCs. Given the importance of co-stimulation in T-cell/B-cell interactions, it will be crucial to further investigate the roles of new co-stimulatory pairs, such as inducible T-cell co-stimulator (Icos)–B7RP1, B7h and its ligand, and PD-1 (programmed cell death 1)–PD-1L in autoimmunity96,97. These molecules are particularly intriguing in the context of our model as they are generally implicated in sustaining or regulating the later, rather than initial, phases of lymphocyte activation. T-cell/B-cell interactions constitute a positive-feedback loop that enables diversification and continuous amplification of the autoimmune response (FIG. 1). After a few turns of the cycle, this type of feedback loop would not require the presence of any exogenous antigens or stimuli that might have initiated it. Small differences in the activation threshold of lymphocytes; their extent of proliferation once activated; as well as pathways that control activated cells, for example by apoptosis — all of which could distinguish lupus-prone individuals from normal — would be magnified by this mechanism and would make it more probable that the positive-feedback cycle would become self-perpetuating. This model predicts that if the B-cell compartment is enriched in particular self-reactive specificities, then this should be reflected in the T-cell compartment. However, this prediction needs to be tested. Emerging data from Mamula and colleagues, who are studying the T-cell response to a prototypical B-cell lupus antigen, Sm (for Smith, a component of the RNA spliceosome), indicate that in lupusprone MRL mice, peripheral T cells are activated by autoantigen-specific B cells. In mice transgenic for an immunglobulin heavy chain that predisposes B cells to Sm reactivity, autoreactive T cells can be elicited readily, in contrast with non-transgenic mice (REF. 55 and J. Yan and M.J.M., unpublished observations). By contrast, the same B cells fail to activate T cells in normal strains of mice. These observations might reflect both the APC function of lupus-prone B cells as well as the lowered activation threshold of lupus T cells. Establishing this point and the mechanistic details would be an important advance in our understanding of how autoreactive T cells are generated and maintained. If our hypothesis is correct, then autoreactive B cells, perhaps selectively loaded with auto-antigens, ought to be useful tools for the isolation and characterization of autoreactive T cells.



PERSPECTIVES

B cells promote epitope spreading

B cells are very efficient APCs for antigens that are taken up specifically through the BCR. Because a B cell can present any antigen that it takes up as part of a macromolecular complex98, it can in principle activate and get help from a broad range of T cells (FIG. 2). Notably, two dominant autoantibody specificities in human systemic autoimmunity and also in mouse lupus — anti-DNA and anti-immunoglobin IgG (known as rheumatoid factor; RF) — have this capability. AntiDNA B cells can present peptides derived from a myriad of self- and even non-selfproteins that can bind nucleic acids, including histones and spliceosomal and ribosomal proteins58. In addition, these B cells could get initial help from T cells that recognize foreign (for example, viral) DNA-binding proteins47,99. Subsequently, these B cells could present self-antigens in an immunogenic way to naive self-reactive T cells, therefore driving the amplification process described in FIG. 1. This would also have the consequence of spreading activation to T cells of different specificities. For example, T cells that are specific for histones, could, in turn, then activate histone-specific B cells. RFs are similar in this regard, as they efficiently present any antigens that are contained in immune complexes, whether foreign or self 29,98. This could even include complexes of anti-DNA with chromatin components, therefore amplifying a pre-existing anti-self response. Indeed, epitope spreading has been shown in a variety of spontaneous and induced organ-specific autoimmune diseases, such as experimental autoimmune encephalitis100, diabetes in NOD mice101 and organ rejection102. It has also been observed following immunization with certain lupus autoantigens, such as from small nuclear ribonucleoprotein (snRNP). Nonetheless, it has been difficult to show directly in mouse lupus models or SLE patients because epitope spreading implies a cascade of events in which autoreactivity to a single initial epitope is followed by broader autoreactivity, a sequence that is hard to establish in a spontaneous disease such as SLE. However, there is evidence that autoantibodies to complex autoantigens, such as Sm, first seem to react with one or a few epitopes and evolve in predictable ways to include more epitopes. The role of epitope spreading in systemic autoimmunity remains controversial, with some workers favouring the notion that crossreactivity explains many of the observations. Importantly, our model of B-cell presentation to T cells does have direct support and provides a clear mechanism for epitope spreading.

b

c

a

d

B-cell activated by self-antigen or molecular mimic

b a

c d

Diversification of T-cell autoimmunity

Diversification of B-cell autoimmunity

Figure 2 | Mechanism of epitope spreading. A B cell, specific for determinant ‘a’, takes up through its B-cell receptor a multideterminant antigen that consists of multiple proteins or protein–nucleic acid complexes (‘b’–‘d’). These multiple T-cell-antigenic determinants, all arising from one initial complex antigen, are processed by the B cell and presented in the context of major histocompatiblily complex (MHC) class II. The single anti-a-specific B cell can thereby activate, and receive help from, a diverse set of T cells. These, in turn, can provide help to a diverse group of B cells that can recognize separate B-cell antigenic determinants on any part of the complex antigen, resulting in the synthesis of anti-b, c and d antibodies. T cells are shown orange; B cells are red.

Pathogenesis by both B and T cells

Classically, tissue damage in SLE has been attributed to effector functions of high-affinity autoantibodies, which is undoubtedly an important mechanism103 (FIG. 1d). However, we propose additionally that effector T cells can have a direct role in tissue damage. In this way, the autoimmune response in SLE would be no different than the immune response to most pathogens, which also engages multiple effector arms of the immune system. For example, a subset of patients with nephritis104,105, as well as MRL lupus-prone mice106, have interstitial nephritis, consisting of T-cell infiltrates in the interstitium, indicating that T cells might directly cause tissue damage.


More definitive proof of this mechanism comes from the fact that elimination of B cells in lupus-prone mice resulted in a complete abrogation of pathology, including T-cell interstitial infiltration85,88 as well as infiltrative dermatitis (O.T. Chan, J. M., McNiff and M.J.S, unpublished observations). Autoimmune-prone mice with B cells that cannot secrete antibodies also had interstitial nephritis and vasculitis, as well as glomerular disease, therefore establishing a direct role for the B cell in promoting T-cell activation and subsequent pathogenesis87. These data provide very strong evidence that T cells, the optimal activation and expansion of which is dependent on B cells, have a direct pathogenic role (see FIG. 1i).



PERSPECTIVES

Conclusion

We have presented a model for the development, maintenance and expansion of autoreactive lymphocytes in SLE, emphasizing T-cell/B-cell interactions in the periphery. The engagement of positive-feedback amplification cycles could well explain the stochastic or ‘environmental’ contributions to systemic autoimmunity5,103. The amplification and momentum inherent in such a mechanism could also explain why it is difficult to gain complete remissions while treating autoimmunity. It also suggests potential targets and strategies for interrupting autoimmunity. We predict that simultaneous T-cell- and B-cell-directed therapies should be synergistic. Co-stimulatory blockade could work by affecting both B- and T-cell responses, and, as both pro- and antiinflammatory stimuli would be amplified, the feedback model could explain why in NZB/W F1 mice, early blockade of CD40–CD40L (CD154) interactions has long-lasting effects107. B-cell deletion might be particularly effective, not just because it would block autoantibody production, but also because it would deplete autoreactive B cells participating in cycles of autoantigen presentation to autoreactive T cells. For example, the B-celldepleting anti-CD20-based Rituxan was recently reported to be possibly effective in treating a small number of patients with rheumatoid arthritis108. Several important questions remain. What are the initiating stimuli for autoimunity? What are the sources of autoantigens and how are they presented in an immunogenic form by APCs? How do these T cells initially become activated and how are they maintained? And what are the pathogenic consequences of these T cells? Our model predicts that the specificity of autoreactive B cells should influence the T-cell compartment as autoimmunity evolves. Systems to explore this are beginning to evolve in our laboratories and elsewhere. We propose that intrinsic lymphocyte signalling defects comprise one important type of genetic contribution, with the result that autoreactive lymphocytes interpret what are normally survival or even negative signals as activating ones 79. Other junctures in our model at which genetic predisposition could be involved include mutations that increase co-stimulation or decrease negative signalling during the T-cell/B-cell interaction, as well as mutations that affect the handling of auto-antigens. Answers to these questions, along with further progress in elucidating the genetic basis of SLE, should lead to an integrated picture of

152

lupus. This should, in turn, lead to intelligent targeting of therapies, that might even have to be individualized for the immunlogical stage of disease (FIG. 1) or combined therapies that target individual elements of the immune disregulation. Mark J. Shlomchik is in the Department of Laboratory Medicine, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06520-8035, USA. Mark J. Shlomchik and Joseph E. Craft are in the Section of Immunobiology, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06520-8035, USA. Joseph E. Craft and Mark J. Mamula are in the Department of Medicine, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06520-8035, USA. Correspondence to: M.J.S. e-mail: [email protected] 1.

2.

3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

Morel, L., Rudofsky, U. H., Longmate, J. A., Schiffenbauer, J. & Wakeland, E. K. Polygenic control of susceptibility to murine systemic lupus erythematosus. Immunity 1, 219–229 (1994). Wakeland, E. K., Wandstrat, A. E., Liu, K. & Morel, L. Genetic dissection of systemic lupus erythematosus. Curr. Opin. Immunol. 11, 701–707 (1999). Vyse, T. J. & Kotzin, B. L. Genetic basis of systemic lupus erythematosus. Curr. Opin. Immunol. 8, 843–851 (1996). Harley, J. B., Moser, K. L., Gaffney, P. M. & Behrens, T. W. The genetics of human systemic lupus erythematosus. Curr. Opin. Immunol. 10, 690–696 (1998). Eisenberg, R. A., Craven, S. Y., Warren, R. W. & Cohen, P. L. Stochastic control of anti-Sm autoantibodies in MRL/Mp-lpr/lpr mice. J. Clin. Invest. 80, 691–697 (1987). Hardin, J. A. The lupus autoantigens and the pathogenesis of systemic lupus erythematosus. Arth. Rheum. 29, 457–460 (1986). Tan, E. M. Antinuclear antibodies: diagnostic markers for autoimmune diseases and probes for cell biology. Adv. Immunol. 44, 93–151 (1989). Mamula, M. J. Epitope spreading: the role of self peptides and autoantigen processing by B lymphocytes. Immunol. Rev. 164, 231–239 (1998). Singh, R. R. & Hahn, B. H. Reciprocal T–B determinant spreading develops spontaneously in murine lupus: implications for pathogenesis. Immunol. Rev. 164, 201–208 (1998). Fatenejad, S., Peng, S., Disorbo, O. & Craft, J. Central T cell tolerance in lupus-prone mice: influence of autoimmune background and the lpr mutation. J. Immunol. 161, 6427–6432 (1998). Singer, P. A., Balderas, R. S., McEvilly, R. J., Bobardt, M. & Theofilopoulos, A. N. Tolerance-related Vβ clonal deletions in normal CD4–8–, TCR– α/β+ and abnormal lpr and gld cell populations. J. Exp. Med. 170, 1869–1877 (1989). Herron, L. R. et al. Selection of the T cell receptor repertoire in Lpr mice. J. Immunol. 151, 3450–3459 (1993). Rozzo, S. J., Drake, C. G., Chiang, B.-L., Gershwin, M. E. & Kotzin, B. L. Evidence for polyclonal T cell activation in murine models of systemic lupus erythematosus. J. Immunol. 153, 1340–1351 (1994). Rubio, C. F., Kench, J., Russell, D. M., Yawger, R. & Nemazee, D. Analysis of central B cell tolerance in autoimmune-prone MRL/lpr mice bearing autoantibody transgenes. J. Immunol. 157, 65–61 (1996). Singer, G. G. & Abbas, A. K. The Fas antigen is involved in peripheral but not thymic deletion of T lymphocytes in T cell receptor transgenic mice. Immunity 1, 365–371 (1994). Smyth, L., Howell, M. & Crispe, I. N. Self-reactivity and the expression of memory markers vary independently in MRL-Mp+/+ and MRL-Mp-lpr/lpr mice. Dev. Immunol. 2, 309–318 (1992). Zhou, T., Bluethmann, H., Zhang, J., Edwards, C. D. & Mountz, J. D. Defective maintenance of T cell tolerance to a superantigen in MRL-lpr/lpr mice. J. Exp. Med. 176, 1063–1067 (1992).


18. Kishimoto, H., Surh, C. D. & Sprent, J. A role for Fas in negative selection of thymocytes in vivo. J. Exp. Med. 187, 1427–1438 (1998). 19. Conger, J. D., Pike, B. L. & Nossal, G. J. V. Clonal analysis of the anti-DNA repertoire of murine B lymphocytes. Proc. Natl Acad. Sci. USA 84, 2931–2935 (1987). 20. Nemazee, D. & Sato, V. Induction of rheumatoid factors in the mouse. Regulated production of autoantibody in secondary humoral response. J. Exp. Med. 158, 529–545 (1983). 21. Welch, M. J., Fong, S., Vaughan, J. & Carson, D. Increased frequency of rheumatoid factor precursor B lymphocytes after immunization of normal adults with tetanus toxoid. Clin. Exp. Immunol. 51, 200–304 (1983). 22. Decker, P., Le Moal, A., Briand, J. P. & Muller, S. Identification of a minimal T cell epitope recognized by antinucleosome TH cells in the C-terminal region of histone H4. J. Immunol. 165, 654–662 (2000). 23. Lehmann, P. V., Forsthuber, T., Miller, A. & Sercarz, E. E. Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358, 155–157 (1992). 24. Mamula, M. J., Lin, R.-H., Janeway, C. A. Jr. & Hardin, J. A. Breaking T cell tolerance with foreign and self coimmunogens: a study of autoimmune B and T cell epitopes of cytochrome c. J. Immunol. 149, 789–795 (1992). 25. Anderson, A. C. et al. High frequency of autoreactive myelin proteolipid protein-specific T cells in the periphery of naive mice: mechanisms of selection of the self-reactive repertoire. J. Exp. Med. 191, 761–770 (2000). 26. Zhang, J. et al. Increased frequency of interleukin 2responsive T cells specific for myelin basic protein and proteolipid protein in peripheral blood and cerebrospinal fluid of patients with multiple sclerosis. J. Exp. Med. 179, 973–984 (1994). 27. Bouneaud, C., Kourilsky, P. & Bousso, P. Impact of negative selection on the T cell repertoire reactive to a self-peptide: a large fraction of T cell clones escapes clonal deletion. Immunity 13, 829–840 (2000). 28. Shlomchik, M. J., Zharhary, D., Camper, S., Saunders, T. & Weigert, M. A Rheumatoid factor transgenic mouse model of autoantibody regulation. Int. Immunol. 5, 1329–1341 (1993). 29. Hannum, L. G., Ni, D., Haberman, A. M., Weigert, M. G. & Shlomchik, M. J. A disease-related RF autoantibody is not tolerized in a normal mouse: implications for the origins of autoantibodies in autoimmune disease. J. Exp. Med. 184, 1269–1278 (1996). 30. Erikson, J., Radic, M. Z., Camper, S. A., Hardy, R. R. & Weigert, M. G. Expression of anti-DNA immunoglobulin transgenes in non-autoimmune mice. Nature 349, 331–334 (1991). 31. Mandik-Nayak, L., Bui, A., Noorchashm, H., Eaton, A. & Erikson, J. Regulation of anti-double-stranded DNA B cells in nonautoimmune mice: localization to the T–B interface of the splenic follicle. J. Exp. Med. 186, 1257–1267 (1997). 32. Nguyen, K. A. et al. Characterization of anti-singlestranded DNA B cells in a non-autoimmune background. J. Immunol. 159, 2633–2644 (1997). 33. Cyster, J. G. & Goodnow, C. C. Antigen-induced exclusion from follicles and anergy are separate and complementary processes that influence peripheral B cell fate. Immunity 3, 691–701 (1995). 34. Rathmell, J. C. et al. CD95 (Fas)-dependent elimination of self-reactive B cells upon interaction with CD4+ T cells. Nature 376, 181–184 (1995). 35. Watanabe-Fukunaga, R., Brannan, C. I., Copeland, N. G., Jenkins, N. A. & Nagata, S. Lymphoproliferation disorder in mice explained by defects of Fas antigen that mediates apoptosis. Nature 356, 314–317 (1992). 36. Zhou, T. et al. Greatly accelerated lymphadenopathy and autoimmune disease in lpr mice lacking tumor necrosis factor receptor 1. J. Immunol. 156, 2661–2665 (1996). 37. Tivol, E. A. et al. Loss of CTLA-4 leads to massive lymphoproliferation and fatal multiorgan tissue destruction, revealing a critical negative regulatory role of CTLA-4. Immunity 3, 541–547 (1995). 38. Yaswen, L. et al. Autoimmune manifestations in the transforming growth factor-β1 knockout mouse. Blood 87, 1439–1445 (1996). 39. Gorelik, L. & Flavell, R. A. Abrogation of TGFβ signaling in T cells leads to spontaneous T cell differentiation and autoimmune disease. Immunity 12, 171–181 (2000).



PERSPECTIVES 40. Mackay, F. et al. Mice transgenic for BAFF develop lymphocytic disorders along with autoimmune manifestations. J. Exp. Med. 190, 1697–1710 (1999). 41. Moore, P. A. et al. BLyS: member of the tumor necrosis factor family and B lymphocyte stimulator. Science 285, 260–263 (1999). 42. Yoshinaga, S. K. et al. Characterization of a new human B7-related protein: B7RP-1 is the ligand to the costimulatory protein ICOS. Int. Immunol. 12, 1439–1447 (2000). 43. Wang, H. & Shlomchik, M. J. Autoantigen-specific B cell activation in Fas-deficient rheumatoid factor immunoglobulin transgenic mice. J. Exp. Med. 190, 639–649 (1999). 44. Jacobson, B. A. et al. An isotype switched and somatically mutated rheumatoid factor clone isolated from a MRL-lpr/lpr mouse exhibits limited intraclonal affinity maturation. J. Immunol. 152, 4489–4499 (1994). 45. Mandik-Nayak, L. et al. MRL-lpr/lpr mice exhibit a defect in maintaining developmental arrest and follicular exclusion of anti-double-stranded DNA B cells. J. Exp. Med. 189, 1799–1814 (1999). 46. Okamoto, M. et al. A transgenic model of autoimmune hemolytic anemia. J. Exp. Med. 175, 71–79 (1992). 47. Moens, U. et al. In vivo expression of a single viral DNAbinding protein generates systemic lupus erythematosusrelated autoimmunity to double-stranded DNA and histones. Proc. Natl Acad. Sci. USA 92, 12393–12397 (1995). 48. Shlomchik, M. J., Marshak-Rothstein, A., Wolfowicz, C. B., Rothstein, T. L. & Weigert, M. G. The role of clonal selection and somatic mutation in autoimmunity. Nature 328, 805–811 (1987). 49. Shlomchik, M. J. et al. Anti-DNA antibodies from autoimmune mice arise by clonal expansion and somatic mutation. J. Exp. Med. 171, 265–292 (1990). 50. Radic, M. Z. & Weigert, M. Genetic and structural evidence for antigen selection of anti-DNA antibodies. Annu. Rev. Immunol. 12, 487–520 (1994). 51. Peng, S. L. et al. Murine lupus in the absence of αβ T cells. J. Immunol. 156, 4041–4049 (1996). 52. Peng, S. L., Fatenejad, S. & Craft, J. Induction of nonpathologic, humoral autoimmunity in lupus-prone mice by a class II-restricted, transgenic αβ T cell. Separation of autoantigen-specific and-nonspecific help. J. Immunol. 157, 5225–5230 (1996). 53. Wofsky, D., Ledbetter, J. A., Hendler, P. L. & Seaman, W. E. Treatment of murine lupus with monoclonal anti-T cell antibody. J. Immunol. 134, 852–857 (1985). 54. Desai, D. D. & Marion, T. N. Induction of anti-DNA antibody with DNA-peptide complexes. Int. Immunol. 12, 1569–1578 (2000). 55. Shinde, S. et al. T cell autoimmunity in Ig transgenic mice. J. Immunol. 162, 7519–7524 (1999). 56. Mamula, M. J., Fatenejad, S. & Craft, J. B cells process and present lupus autoantigens that initiate autoimmune T cell responses. J. Immunol. 152, 1453–1461 (1994). 57. Monneaux, F., Briand, J. P. & Muller, S. B and T cell immune response to small nuclear ribonucleoprotein particles in lupus mice: autoreactive CD4+ T cells recognize a T cell epitope located within the RNP80 motif of the 70K protein. Eur. J. Immunol. 30, 2191–2200 (2000). 58. Mohan, C., Adams, S., Stanik, V. & Datta, S. Nucleosome: a major immunogen for pathogenic autoantibody-inducing T cells of lupus. J. Exp. Med. 177, 1367–1381 (1993). 59. Kaliyaperumal, A., Mohan, C., Wu, W. & Datta, S. K. Nucleosomal peptide epitopes for nephritis-inducing T helper cells of murine lupus. J. Exp. Med. 183, 2459–2469 (1996). 60. Janeway, C. A. Jr. & Bottomly, K. Signals and signs for lymphocyte responses. Cell 76, 275–285 (1994). 61. Nash, J. T. et al. Immune complex processing in C1qdeficient mice. Clin. Exp. Immunol. 123, 196–202 (2001). 62. Taylor, P. R. et al. A hierarchical role for classical pathway complement proteins in the clearance of apoptotic cells in vivo. J. Exp. Med. 192, 359–366 (2000). 63. Scott, R. S. et al. Phagocytosis and clearance of apoptotic cells is mediated by MER. Nature 411, 207–211 (2001). 64. Pickering, M. C. & Walport, M. J. Links between complement abnormalities and systemic lupus erythematosus. Rheumatology (Oxford) 39, 133–141 (2000). 65. Botto, M. et al. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nature Genet. 19, 56–59 (1998).

66. Prodeus, A. P. et al. A critical role for complement in maintenance of self-tolerance. Immunity 9, 721–731 (1998). 67. Morel, L., Blenman, K. R., Croker, B. P. & Wakeland, E. K. The major murine systemic lupus erythematosus susceptibility locus, Sle1, is a cluster of functionally related genes. Proc. Natl Acad. Sci. USA 98, 1787–1792 (2001). 68. Topaloglu, R. et al. Survey of Turkish systemic lupus erythematosus patients for a particular mutation of C1Q deficiency. Clin. Exp. Rheumatol. 18, 75–77 (2000). 69. Salmon, J. E. et al. Fcγ RIIA alleles are heritable risk factors for lupus nephritis in African Americans. J. Clin. Invest. 97, 1348–1354 (1996). 70. Lanzavecchia, A. Mechanisms of antigen uptake for presentation. Curr. Opin. Immunol. 8, 348–354 (1996). 71. Mamula, M. J. et al. Isoaspartyl post-translational modification triggers autoimmune responses to selfproteins. J. Biol. Chem. 274, 22321–22327 (1999). 72. Brahms, H. et al. The C-terminal RG dipeptide repeats of the spliceosomal Sm proteins D1 and D3 contain symmetrical dimethylarginines, which form a major B-cell epitope for anti-Sm autoantibodies. J. Biol. Chem. 275, 17122–17129 (2000). 73. Doyle, H. A. & Mamula, M. J. Post-translational protein modifications in antigen recognition and autoimmunity. Trends Immunol. 22, 443–449 (2001). 74. Rosen, A., Casciola, R. L. & Ahearn, J. Novel packages of viral and self-antigens are generated during apoptosis. J. Exp. Med. 181, 1557–1561 (1995). 75. Kozono, Y., Kotzin, B. L. & Holers, V. M. Resting B cells from New Zealand Black mice demonstrate a defect in apoptosis induction following surface IgM ligation. J. Immunol. 156, 4498–4503 (1996). 76. Mohan, C., Morel, L., Yang, P. & Wakeland, E. K. Genetic dissection of systemic lupus erythematosus pathogenesis: Sle2 on murine chromosome 4 leads to B cell hyperactivity. J. Immunol. 159, 454–465 (1997). 77. DesJardin, L. E., Butfiloski, E. J., Sobel, E. S. & Schiffenbauer, J. Hyperproliferation of BXSB B cells is linked to the Yaa allele. Clin. Immunol. Immunopathol. 81, 145–152 (1996). 78. Mohan, C., Yu, Y., Morel, L., Yang, P. & Wakeland, E. K. Genetic dissection of Sle pathogenesis: Sle3 on murine chromosome 7 impacts T cell activation, differentiation, and cell death. J. Immunol. 162, 6492–6502 (1999). 79. Vratsanos, G., Jung, S., Park, Y. & Craft, J. CD4+ T Cells from lupus-prone mice are hyperresponsive to T cell receptor engagement with low and high affinity peptide antigens. A model to explain spontaneous T cell activation in lupus. J. Exp. Med. 193, 329–338 (2001). 80. Jongstra-Bilen, J., Vukusic, B., Boras, K. & Wither, J. E. Resting B cells from autoimmune lupus-prone New Zealand Black and (New Zealand Black x New Zealand White) F1 mice are hyper-responsive to T cell-derived stimuli. J. Immunol. 159, 5810–5820 (1997). 81. Liossis, S.-N. C., Kovacs, B., Dennis, G., Kammer, G. M. & Tsokos, G. C. B cells from patients with systemic lupus erythematosus display abnormal antigen receptormediated early signal transduction events. J. Clin. Invest. 98, 2549–2557 (1996). 82. Liossis, S.-N. C., Ding, X. Z., Dennis, G. J. & Tsokos, G. C. Altered pattern of TCR/CD3-mediated protein-tyrosyl phosphorylation in T cells from patients with systemic lupus erythematosus. J. Clin. Invest. 101, 1448–1457 (1998). 83. Hang, L., Theofilopoulos, A. N., Balderas, R. S., Francis, S. J. & Dixon, F. J. The effect of thymectomy on lupusprone mice. J. Immunol. 132, 1809–1813 (1984). 84. Peng, S. L., Madaio, M. P., Hayday, A. C. & Craft, J. Propagation and regulation of systemic autoimmunity by γδ T cells. J. Immunol. 157, 5689–5698 (1996). 85. Shlomchik, M. J., Madaio, M. P., Ni, D., Trounstine, M. & Huszar, D. The role of B cells in lpr/lpr-induced autoimmunity. J. Exp. Med. 180, 1295–1306 (1994). 86. Chan, O. & Shlomchik, M. J. A new role for B cells in systemic autoimmunity: B cells promote spontaneous T cell activation in MRL-lpr/lpr mice. J. Immunol. 160, 51–59 (1998). 87. Chan, O. T., Hannum, L. G., Haberman, A. M., Madaio, M. P. & Shlomchik, M. J. A novel mouse with B cells but lacking serum antibody reveals an antibody-independent role for B cells in murine lupus. J. Exp. Med. 189, 1639–1648 (1999). 88. Chan, O. T., Madaio, M. P. & Shlomchik, M. J. B cells are required for lupus nephritis in the polygenic, Fas-intact MRL model of systemic autoimmunity. J. Immunol. 163, 3592–3596 (1999).


89. Sobel, E. S., Mohan, C., Morel, L., Schiffenbauer, J. & Wakeland, E. K. Genetic dissection of SLE pathogenesis: adoptive transfer of Sle1 mediates the loss of tolerance by bone marrow-derived B Cells. J. Immunol. 162, 2415–2421 (1999). 90. Mamula, M. J. & Janeway, C. A. Jr. Do B cells drive the diversification of immune responses? Immunol. Today 14, 151–152 (1993). 91. Lin, R.-H., Mamula, M. J., Hardin, J. A. & Janeway, C. A. Jr. Induction of autoreactive B cells allows priming of autoreactive T cells. J. Exp. Med. 173, 1433–1439 (1991). 92. James, J. A. & Harley, J. B. B-cell epitope spreading in autoimmunity. Immunol. Rev. 164, 185–200 (1998). 93. McCluskey, J. et al. Determinant spreading: lessons from animal models and human disease. Immunol. Rev. 164, 209–229 (1998). 94. Liang, B., Gee, R. J., Kashgarian, M. J., Sharpe, A. H. & Mamula, M. J. B7 costimulation in the development of lupus: autoimmunity arises either in the absence of B7.1/B7.2 or in the presence of anti-b7.1/B7.2 blocking antibodies. J. Immunol. 163, 2322–2329 (1999). 95. Liang, B., Kashgarian, M. J., Sharpe, A. H. & Mamula, M. J. Autoantibody responses and pathology regulated by B7-1 and B7-2 costimulation in MRL/lpr lupus. J. Immunol. 165, 3436–3443 (2000). 96. Coyle, A. J. & Gutierrez-Ramos, J. C. The expanding B7 superfamily: increasing complexity in costimulatory signals regulating T cell function. Nature Immunol. 2, 203–209 (2001). 97. Nishimura, H. & Honjo, T. PD-1: an inhibitory immunoreceptor involved in peripheral tolerance. Trends Immunol. 22, 265–268 (2001). 98. Roosnek, E. & Lanzavecchia, A. Efficient and selective presentation of antigen–antibody complexes by rheumatoid factor B cells. J. Exp. Med. 173, 487–489 (1991). 99. Andreassen, K. et al. T cell lines specific for polyomavirus T-antigen recognize T-antigen complexed with nucleosomes: a molecular basis for anti-DNA antibody production. Eur. J. Immunol. 29, 2715–2728 (1999). 100. Tuohy, V. K. et al. The epitope spreading cascade during progression of experimental autoimmune encephalomyelitis and multiple sclerosis. Immunol. Rev. 164, 93–100 (1998). 101. Zechel, M. A., Krawetz, M. D. & Singh, B. Epitope dominance: evidence for reciprocal determinant spreading to glutamic acid decarboxylase in non-obese diabetic mice. Immunol. Rev. 164, 111–118 (1998). 102. Suciu-Foca, N., Harris, P. E. & Cortesini, R. Intramolecular and intermolecular spreading during the course of organ allograft rejection. Immunol. Rev. 164, 241–246 (1998). 103. Kotzin, B. Systemic lupus erythematosus. Cell 85, 303–306 (1996). 104. Alexopoulos, E., Seron, D., Hartley, R. B. & Cameron, J. S. Lupus nephritis: correlation of interstitial cells with glomerular function. Kidney Int. 37, 100–109 (1990). 105. O’Dell, J. R., Hays, R. C., Guggenheim, S. J. & Steigerwald, J. C. Tubulointerstitial renal disease in systemic lupus erythematosus. Arch. Intern. Med. 145, 1996–1999 (1985). 106. Hewicker, M. & Trautwein, G. Sequential study of vasculitis in MRL mice. Lab. Anim. 21, 335–341 (1987). 107. Mohan, C., Shi, Y., Laman, J. D. & Datta, S. K. Interaction between CD40 and its ligand gp39 in the development of murine lupus nephritis. J. Immunol. 154, 1470–1480 (1995). 108. Edwards, J. C. & Cambridge, G. Sustained improvement in rheumatoid arthritis following a protocol designed to deplete B lymphocytes. Rheumatology (Oxford) 40, 205–211 (2001).

Online Links DATABASES The following terms are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ B7RP1 | TNF-β MGI: http: //www.informatics.jax.org/ B7h | BLyS | CD40 | CD80 | CD86 | CD95 | CD152 | CD154 | Icos | IL-2 | PD-1 | TNF OMIM: http://www.ncbi.nlm.nih.gov/Omim/ systemic lupus erythematosus Access to this interactive links box is free online.



PERSPECTIVES

OPINION

Xenotransplantation and other means of organ replacement Marilia Cascalho and Jeffrey L. Platt Exciting new technologies, such as cellular transplantation, organogenesis and xenotransplantation, are thought to be promising approaches for the treatment of human disease. The feasibility of applying these technologies, however, might be limited by biological and immunological hurdles. Here, we consider whether, and how, xenotransplantation and various other technologies might be applied in future efforts to replace or supplement the function of human organs and tissues.

Few fields of medicine have engendered more excitement and controversy than those focusing on the replacement of organs. Allotransplantation — the transplantation of cells, tissues or organs between individuals of the same species — is now the preferred treatment for organ failure, but its application is limited because human organs are in short supply. Xenotransplantation — the transplantation of cells, tissues or organs between individuals of different species — offers the possibility of overcoming these organ shortages, and is a potential avenue for the application of new technologies, such as genetic engineering, cloning and rational design of therapeutics. However, xenotransplantation provokes controversy because successful application would require overcoming severe immunological hurdles and because the transplanted organs might carry with them organisms that could give rise to new infections in human populations. Emerging technologies, such as artificial organs, stemcell biology and organogenesis, might offer a way around some of the biological and societal hurdles, but these technologies might have limitations of their own. In this review, we shall consider some of the new opportunities and limitations of xenotransplantation and other technologies. Approaches to replacement of organs

The importance of xenotransplantation, and other approaches for the replacement of organs, varies considerably between different organ systems. For example, in the case of cardiac failure, xenotransplantation was once

154

seen as the best alternative to allotransplantation, which can only address a small fraction of the need because there are severe shortages of human hearts available for transplantation. However, there has been substantial progress in both the development of mechanical devices that can be used to supplement cardiac function and in the development of a totally artificial heart1. Furthermore, cellular therapies might offer an alternative approach to augment cardiac function and avoid cardiac replacement2–4. So, the need for wholeorgan cardiac transplants might conceivably diminish over time. By contrast, in the case of the kidney and the lung, no fully implantable devices exist, and there is little prospect for cellular therapy because the structure and function of these organs are too complex. Although renal failure can be treated by dialysis, pulmonary function cannot be replaced by any means other than transplantation. In the case of the liver, whole-organ xenotransplantation and implantable devices seem unfeasible because of the complex metabolic processes that occur; however, ALLOGENEIC and even XENOGENEIC hepatocyte transplantation offer promise. To properly assess how the

various alternatives to transplantation, such as cellular therapies and organogenesis, might be applied in the future, it is important to consider how these technologies might develop in the next few years. Cellular therapies

One promising approach to replace or augment the function of an organ is cellular transplantation, which involves the injection of cells that have the potential for replacing cells damaged or destroyed by disease. For example, recent studies have shown that skeletal myoblasts (primitive muscle cells) or stem cells of various types can be transplanted into the damaged heart, and on healing, the transplanted cells assume the function of cardiac myocytes and can augment cardiac function replacement2–4. As another example, isolated hepatocytes or stem cells can be transplanted into the liver to address genetic defects5,6. One advantage of using stem cells is the possibility that the cells might be taken from the affected patient, therefore obviating immune response to foreign cells. Reports indicating that such cells can be derived from the bone marrow7, central nervous system8 or fat9 of individuals are encouraging. Cellular transplantation does have limitations, however. For example, it might not be possible to improve the function of structurally complex organs, such as the kidney or lung. In addition, cellular transplantation might prove ineffective in diffuse diseases, such as myocarditis, amyloidosis and portal hypertension. Another limitation is that differentiated cells and stem cells from mature individuals have a limited proliferative potential and might therefore not be capable

Box 1 | Zoonosis One potential complication of transplantation is the conveying of infectious organisms from the transplant to the recipient; in xenotransplants, such an infection would be a zoonosis. The risk of the spread of pathogens should be less in xenotransplantation than in allotransplantation, because most pathogens of pigs do not infect humans, and because pigs can be raised to be free of known human pathogens. However, attention has been drawn to the possibility that the xenotransplant might serve as the source of a ‘new’ infectious agent generated by spread of an endogenous pig virus, or mutation or recombination of pig viruses, and that such an agent might spread widely among humans. This concern has been widely discussed in the medical literature. To date, no ‘new’ organism, including the porcine endogenous retrovirus54, has been found to be transmitted to humans55,56. Although concern about this subject will surely continue, that concern might be balanced by two further considerations. First, it is clear that whether or not xenotransplantation becomes available, epidemics caused by new infectious agents will occur, and some, like hepatitis C, might be associated with organ failure. If larger numbers of individuals should experience organ failure, then xenotransplantation might be seen as a solution to the problem, rather than a potential cause. Second, if a new and dangerous zoonotic organism should emerge in the course of xenotransplantation, the carefully monitored recipients might alert society to the risk and enable the development of approaches to prevent introduction and spread of the organism by other means.




PERSPECTIVES of repairing large masses of defective tissue. For these purposes, an organ transplant would be optimal.

a

Cell xenotransplant, e.g. hepatocytes Recipient blood vessel

Organogenesis

One potential approach to replacing the function of structurally complex organs, such as the kidney or lung, is organogenesis, or the growing of organs de novo from primitive cells or tissues or stem cells8,10. If feasible, organogenesis would avoid the main limitation of cellular transplants. Whether organogenesis can produce physiologically competent organs is not yet clear. Fetal mouse metanephric kidney tissue can be grown from primitive mesenchyme in culture11,12. However, the nephrons in these organs lack the blood vessels needed for function and can be grown to a size of only a few millimeters in vitro12. To overcome these problems, organogenesis might be carried out in vivo. Hammerman recently reported that fetal kidneys transplanted in the renal capsule or omentum of rats can undergo vascularization and might even exhibit some function13. If the technical capability to allow organogenesis in a human existed, the growth process would presumably require a period of months, if not years, and therefore a temporary measure, such as xenotransplantation, would be needed for vital organs. As an alternative, an animal could be used as a temporary host for the developing organ.

Xenotransplant successful

Free tissue xenotransplant, e.g. skin

Donor blood vessel

b

Primary non-function

Recipient blood vessel

Cellular rejection

Xenotransplant failure

Organ xenotransplant, e.g. kidneys

Anastomosis

Recipient blood vessel

Xenotransplant successful

Hyperacute rejection

Acute vascular rejection

Cellular rejection

Chronic rejection

Donor blood vessel Xenotransplant failure

Figure 1 | Biological responses to xenotransplantation. Biological responses to transplantation depend on the means by which a xenotransplant receives its vascular supply. a | Cell and tissue xenotransplants are vascularized, for the most part, by recipient blood vessels. These grafts are subject to failure of engraftment (primary non-function) or cellular rejection, but not to vascular rejection. b | Organ xenografts are vascularized by donor blood vessels. The grafts are subject to a series of vascular responses beginning with hyperacute rejection (minutes to hours), acute vascular rejection (days to weeks), cellular rejection and chronic rejection.

Xenotransplantation

Until organogenesis becomes feasible, xenotransplantation might be the best approach for the replacement of the kidney and lung, and possibly for other organs, and for conditions not amenable to cellular therapies. Xenotransplantation might also contribute to the development of organogenesis. As discussed above, genetically modified xenotransplants might be used to deliver specific gene products for such purposes as reconstituting defective pathways or promoting tissue growth. We have recently discussed other applications of xenotransplantation14. However, several obstacles remain before xenotransplantation can be widely used. These include the immune response of the recipient against the transplant and the physiological limitations of the transplant in the foreign host. In addition, there is a possibility that infectious agents might be transferred from the transplant to the recipient (BOX 1). Xenotransplantation has been attempted on a number of occasions during the past 100 years, so more is known about the hurdles to xenotransplantation than the hurdles to other approaches to replace organ function.

In considering these obstacles, it is important to distinguish between grafted cells and tissues on the one hand and grafted organs on the other hand, as we believe the biological barriers to xenotransplantation depend, to a significant extent, on the way in which the graft is connected to the recipient. As mentioned above, the immune response to a xenotransplant is a difficult obstacle to overcome. The elements of the immune system involved in xenograft recognition have been recently reviewed by us14. The immune responses to xenotransplantation are much more severe than the immune responses to allotransplantation for at least three reasons. First, all individuals have innate immunity against xenogeneic cells, and this innate immune response, which in humans includes xenoreactive antibodies (XA), complement and natural killer cells, recruits adaptive immune responses against the graft14. Second, xenogeneic transplants carry a diverse set of foreign antigens against which cellular and humoral immune responses can be elicited (in allotransplants, by contrast, the main foreign antigens are major histocompatibility


complex antigens)15. Third, immune regulation, which might partially control responses to allografts, might, in our view, fail to do so in responses to xenografts. Xenotransplantation of cells and tissues. Grafts of isolated cells, such as hepatocytes, are nourished and maintained by the microenvironment, growth factors and ingrowth of capillaries of the recipient (FIG. 1). Transplants consisting of tissues, such as the skin, are maintained by both donor and recipient growth factors, and have a mixed vascular supply, consisting of in-grown blood vessels of recipient origin and blood vessels formed by the spontaneous anastomosis of donor and recipient capillaries. Cell and tissue transplants, especially transplants of bone marrow cells and pancreatic islets, might be subject to a condition known as ‘primary non-function’. We believe that primary nonfunction of xenogeneic transplants is caused by one or more of three factors: first, the inability of growth factors of the recipient to support newly implanted cells and/or failure of graft factors to support angiogenesis by



PERSPECTIVES

Table 1 | Approaches to prevention of hyperacute rejection Approach

Method

Depletion of xenoreactive antibodies

Column absorption

References 67,68

Inhibition of complement

Cobra venom factor, sCR1

Genetic engineering for the expression of complement-regulatory proteins

DAF, CD59, MCP

69,70

Genetic engineering to decrease antigen expression

Knock-out α-1,3-galactosyltransferase gene and possibly other genes in pigs

33,71,72 40–42

DAF, decay-accelerating factor; MCP, membrane co-factor protein; sCR1, soluble complement receptor type 1.

host vessels16; second, the action of natural killer cells or recently activated T cells on the newly implanted graft; and third, the action of complement on xenogeneic cells and tissues introduced into the blood (for example, pancreatic islets injected into the portal vein)17. The main hurdle to xenotransplantation of cells and tissues is cellular rejection. As mentioned, cell-mediated immune responses to xenotransplantation are thought to be

especially severe15,18,19 and might, in our view, be further amplified by the humoral immune reactions and by failure of immune regulation between species14,20. Some fundamental aspects of the cellular immune response to xenotransplantation have been reviewed by us14 and others21. Although the cellular immune response to xenotransplantation is severe, that response seems to be subject to control by immunosuppressive agents that are currently available6,22,23.

TXA2

IL-1β E-selectin

Tissue factor

Complement

Xenotransplantation of vascularized organs. Whole-organ grafts are connected to the recipient by anastomosis of large blood vessels of the donor and recipient. Aside from this connection, the graft remains entirely of donor origin. Hence, organ xenografts are not generally compromised by incompatibility of the local environment in which they are placed. Conversely, the blood vessels of organ xenografts are directly exposed to components of the immune system of the recipient, and it is the interaction of the immune system with donor blood vessels that gives rise to distinct types of vascular disease which have to this point prevented the clinical transplantation of xenogeneic organs (FIG. 1). Vascularized organs are first subject to hyperacute rejection, a devastating condition that destroys a xenograft within minutes to a few hours24. Hyperacute rejection of pig organs transplanted into primates is triggered by the binding of xenoreactive natural antibodies to Galα1-3Gal, a saccharide expressed by pigs and other lower mammals25.

IL-1α Endothelial cell activation Xenoreactive antibodies

Membraneattack complex

Thrombosis

Vasoconstriction Matrix

Acute vascular rejection

Endothelium of blood vessels in xenograft Inflammation

Apoptosis Matrix exposed

Loss of NO

Figure 2 | The pathogenesis of acute vascular rejection. Acute vascular rejection is induced by xenoreactive antibodies directed against the endothelial lining of blood vessels in the graft, and possibly by complement. Whereas the endothelium of normal blood vessels promotes blood flow and inhibits thrombosis and inflammation, the endothelium of xenografts promotes vasoconstriction, thrombosis and inflammation, giving rise to the picture of ischaemia and thrombosis that is characteristic of acute vascular rejection of xenografts64,65. These pathophysiological changes in endothelium are due, at least in part, to coordinate elaboration of tissue factor, plasminogen-activator inhibitor type 1 (PAI-1), E-selectin and thromboxane A2 (TXA2), and other products of genes induced by the action of xenoreactive antibodies, as well as small amounts of complement or platelets29,35,45,65,66. These coordinate changes induce thrombosis, inflammation and vasoconstriction and are, in turn, induced through two pathways. One pathway leads to the production of interleukin (IL)-1α, which acts as an autocrine factor inducing the production of various proteins, such as tissue factor and E-selectin, resulting in the release of thromboxane A2 and IL-1β. The other pathway involves apoptosis, which leads to loss of endothelial cells, which exposes the matrix and decreases the availability of nitric oxide (NO).

156




PERSPECTIVES

Table 2 | Approaches to prevention of acute vascular rejection Method

Result

Pre-transplant infusion with donor haematopoietic cells

Tolerance to Galα1,3Gal and other xenospecific antigens

37,73

Knock-out α-1,3-galactosyltransferase and possibly other genes in pigs

Decreased antigen expression

40–42

Suppression of pro-coagulant or pro-inflammatory genes

Inhibition of endothelial cell activation

Transient depletion of xenoreactive antibodies

Induction of accommodation

Binding of these antibodies activates complement, which, in turn, causes graft destruction. The mechanisms underlying susceptibility to hyperacute rejection have been the subject of controversy, as binding of antidonor antibodies occurs in a variety of conditions, including some xenografts in which hyperacute rejection is not observed26. We believe that hyperacute rejection is caused by the rapid insertion of terminal complement complexes in the cell membranes of the endothelial lining of blood vessels in the donor organ24, and anything that modifies the kinetics of complex formation modifies susceptibility to rejection (TABLE 1). Among the factors that might influence the rate of complement reactions is the availability and function of complement-regulatory proteins27,28. We had postulated that activation of complement in xenografts is amplified because complement-regulatory proteins, such as decay-accelerating factor (CD55), CD59 and membrane cofactor protein, which function more effectively against homologous than against heterologous complement, fail to protect the xenograft against complementmediated injury29. However, on the basis of studies using isolated cells, some have questioned whether complement-regulatory proteins, particularly CD59, do indeed function in a species-specific fashion30. We believe this controversy is addressed by xenotransplantation. First, expression of CD55 (REF. 26), but not CD59, prevents hyperacute rejection31, indicating that terminal complement components not controlled by CD59, that is C5b67, might be sufficient to induce changes in the endothelium underlying hyperacute rejection32. Second, pig organs expressing human complement-regulatory proteins at very low levels are protected from hyperacute rejection, so establishing the idea that failure of complement control is an important obstacle to xenotransplantation33. These observations further establish that the safest, and perhaps the most clinically applicable, approach to preventing hyperacute rejection is expression in the graft of complement-

References

45 29,44,48

regulatory proteins compatible with the complement system of the recipient. If hyperacute rejection is prevented, an organ xenograft becomes susceptible to a condition we have called ‘acute vascular rejection’34. Acute vascular rejection seems to be caused by xenoreactive antibodies, which bind to the xenograft causing ‘activation’ of endothelium in the graft 34,35, and possibly apoptosis36 (FIG. 2).

“…accommodation might be vital to the success of xenotransplantation… there is much interest in how it can be reliably induced…” Acute vascular rejection is thought by many in the field to be the main biological obstacle to xenotransplantation of organs; accordingly, much effort is now directed at developing the means to prevent or treat this disorder (TABLE 2). One way to prevent acute vascular rejection might be to induce immunological tolerance to the xenotransplant donor. Xenogeneic tolerance might be induced by engraftment of donor bone marrow or stem cells37,38, but the biological hurdles to engraftment of xenogeneic bone marrow cells, which include the action of antibodies and complement on the cells and the incompatibility of host growth factors16,39, might make induction of tolerance to organ xenografts difficult to achieve. Another way to prevent acute vascular rejection might be to eliminate the antigens targeted by xenoreactive antibodies. Recent progress in the cloning of pigs40–42, and in gene targeting43, makes it possible for the first time to consider knocking out pig antigens targeted by xenoreactive antibodies. Most of the efforts towards this end have focused on knocking out α1,3-galactosyltransferase, which catalyses the synthesis of


Galα1-3Gal, which has been shown to be the target of some of the antibodies that cause acute vascular rejection44. However, although it might be possible to eliminate this antigen from xenograft donors, it might not be possible to eliminate what we fear might be a myriad of other xenogeneic antigens that could be targeted by xenoreactive antibodies, and eliminating an antigen by gene targeting might uncover new epitopes. Another approach to preventing acute vascular rejection might involve inhibition of expression of genes associated with activation of endothelium45. A fourth approach involves the induction of ‘accommodation’ (BOX 2). First described in organs allografted across ABO blood-group barriers46,47, accommodation is an acquired resistance of an organ to immune-mediated injury29. As it might prove difficult or impossible to prevent humoral responses to xenotransplants, there is much interest in the possibility that accommodation can be used to avert the consequences of humoral rejection 20,29,45. Accommodation has been used to prevent acute vascular rejection in rodent and, arguably, in pig-to-primate xenografts44,48. If acute vascular rejection of a xenograft is averted, the graft might be subject to chronic rejection. Whether, and to what extent, organ xenografts are susceptible to chronic rejection is, as yet, unknown. If chronic rejection is caused by an immune response to the graft, as some experimental evidence indicates49, then it should be common and severe in xenotransplants. If chronic rejection is caused by qualities of the graft, such as preservation time, ISCHAEMIA and donor age, then it should not be much of a problem. In any case, because xenotransplantation offers an unlimited supply of organs, the impact of chronic rejection might be less serious as the chronically rejected organ can be replaced. Glossary ALLOGENEIC

Of, or relating to, the same species; for example, allogeneic transplants are transplants between individuals of the same species. ISCHAEMIA

A condition in which the flow of blood to a tissue or organ is less than normal, and which results in injury to that tissue or organ. ISLETS OF LANGERHANS

The tissue of the pancreas that contains endocrine cells, including the β-cells that secrete insulin. SUBSTANTIA NIGRA

A part of the brain affected by Parkinson’s disease. XENOGENEIC

Of, or relating to, a foreign species.



PERSPECTIVES Physiological hurdles to xenotransplantation. Whether a xenogeneic organ or tissue would function adequately in a human patient is an important consideration for the clinical application of xenotransplantation. Studies in which pig organs have been transplanted into nonhuman primates indicate that the kidneys, hearts and lungs of pigs would function sufficiently well in a human to sustain life50–52. In fact, the main functional impairment of these xenogeneic organ grafts is from rejection. By contrast, presumed incompatibilities between the complex metabolic systems of the pig and human liver

would seem to preclude ready application of hepatic xenotransplantation. However, Ramirez and colleagues53 recently reported that porcine liver xenografts can function adequately in baboons, and pig hepatocytes have been observed to sustain the life of rats with cirrhosis (I. Fox et al., unpublished observations). Even if physiological hurdles were found to be a barrier to xenotransplantation, genetic engineering could be applied to the problem. For example, if an organ or tissue were missing a functional protein, the gene for that protein might be introduced by transgenic techniques.

Xenotransplantation in the clinic

How close is xenotransplantation to clinical application, and how might it be applied to the treatment of human disease in the future? It is our view that xenotransplantation of isolated cells and tissues between species could be undertaken today. This view is based on accumulating evidence that the rejection of cell and tissue xenografts can be controlled with conventional regimens of immunosuppression6 and preliminary success in engraftment of porcine SUBSTANTIA NIGRA cells in human subjects23. Less obvious is how cell and tissue xenotransplants might be applied.

Box 2 | Mechanisms of accommodation

Complement

Agonist (IL-1α) Receptor (IL-1R)

The antigens on the graft change

Effector

Matrix

Xenoreactive antibodies

Accommodation

Xenograft cell

The xenoreactive antibodies change Xenograft cells develop resistance to injury

Receptor desensitization

Negative regulation of receptor–effector pathway

BCL2

Inhibition of effector function

HO-1

Tissue injury Loss of IL-1R

Expression of IκB or BCL2

Expression of CD59 or HO-1

Because accommodation might be vital to the success of xenotransplantation and might be exploited for treatment or prevention of vascular disease, there is much interest in understanding how it can be reliably induced and what mechanisms underlie it. Accommodation of xenotransplants has been induced by temporary depletion of xenoreactive antibodies followed by the return of those antibodies without causing humoral rejection44. In this setting, accommodation might be brought about by a change in xenoreactive antibodies or a change in the antigens in the graft57 (see figure). Another possibility is that the binding of xenoreactive antibodies or the action of inflammatory agonists, in subtoxic amounts induces changes in the graft, which make the graft inured to humoral injury. Resistance to injury might result from one or more of three changes in the graft: first, desensitization or loss of receptors for inflammatory agonists; second, interruption of cell activation or effector pathways — for example, by inhibitory κB (IκB) or BCL2; and third, production of proteins, such as CD59 or haem oxygenase-1 (HO-1), that repair or block the detrimental effects of the agonists that would otherwise induce tissue injury. Consistent with the latter possibility are experiments showing that endothelial cells exposed to xenoreactive antibodies acquire resistance to complement-mediated injury owing to increased expression of CD59 (REF. 58) and other inhibitors of injury59. Experiments in rodents have shown that accommodation is associated with expression of Bcl-2 and HO-1 (REF. 60). Moreover, organ grafts deficient in HO-1 or in functional complement-regulatory proteins seem to be subject to severe vascular injury61. However, efforts to prevent vascular injury by expression of these genes might not be sufficient to induce a state of accommodation, as grafts with increased expression of HO-1 and/or CD59 might still undergo acute vascular rejection (REFS 31,62; and Z.E. Holzknecht and J.L. Platt, unpublished observations). This indicates that accommodation is multifactorial and still incompletely understood. Although we discuss accommodation here in the context of acute vascular rejection, it is possible that accommodation will be found to mediate resistance to other forms of tissue injury63. IL-1α; interleukin-1α; IL-1R, IL-1 receptor.

158




PERSPECTIVES

“It is our view that xenotransplantation of isolated cells and tissues between species could be undertaken today.” One application might be in the treatment of cirrhosis caused by hepatitis viruses. In such cases, animal hepatocytes might be preferred over human hepatocytes to avoid reinfection by human viruses. Another potential application is the transplantation of xenogeneic ISLETS OF LANGERHANS for the treatment of diabetes. Xenogeneic islet transplants might be less subject to destruction by the autoimmune response that underlies type 1 diabetes. In the future, cellular xenotransplants might, in conjunction with genetic engineering, be used to deliver genes or other cellular products that are absent or deficient in expression. The application of organ xenotransplantation as a primary approach to replacement of organs clearly depends on controlling the immune response of the recipient so that the xenografted organ can endure. The challenge is to reliably prevent or treat acute vascular rejection. How can one know this challenge has been met? Nonhuman primate-model systems that have been used to explain fundamental aspects of the immune response to xenotransplantation might not be optimal for testing therapeutic strategies designed for optimal effect in humans. For example, the human complement-regulatory proteins expressed in transgenic pigs might fail to fully control the complement system of the baboon. So, evaluation of the feasibility of xenotransplantation might be more effectively undertaken in human subjects who could be recipients of bridge or temporary transplants of pig organs. Such transplants might be used to keep a person alive until a device or an allograft can be inserted, or until an engineered organ can be fashioned from the patient’s own cells. The use of animals for treating organ failure in humans might acquire a broader application than solely as a source of organs. For example, as in vitro tissue culture is unlikely to yield fully developed functional organs, perhaps pigs or other animals could be used as surrogate recipients to allow completion of organ development. Human organs and tissues grown and maintained in animals might then be available for transplantation. One advantage of this method would be the possibility of genetically modifying stem cells (for

example, to introduce antiviral genes) before implantation in the animal host and transfer into the patient. A new application of xenotransplantation might be in the cloning of human cells, tissues or organs. This possibility is raised by recent successes in the cloning of animals. For example, nuclei from a human patient might be transferred to enucleated stem cells of an animal, and the cells might then be grown in vitro or in an animal to generate differentiated human tissue that is autologous with the patient. Therefore, future applications of xenotransplantation might call for ‘human’-to-animal transplants, and genetic modification of animals might be undertaken to sustain such transplants. With the use of animals as biological reactors, xenotransplantation might acquire broader meaning and impact in the treatment of human disease. Marilia Cascalho and Jeffrey L. Platt are in Transplantation Biology and the Departments of Surgery and Immunology, Medical Sciences Building 2–66, 200 1st Street SW, Mayo Clinic, Rochester, Minnesota 55905, USA. Jeffrey L. Platt is also in the Department of Pediatrics, Medical Sciences Building 2–66, 200 1st Street SW, Mayo Clinic, Rochester, Minnesota 55905, USA. Correspondence to J.L.P. e-mail: [email protected] 1. 2.

3.

4. 5.

6.

7.

8. 9.

10.

11.

12.

13. 14. 15.

SoRelle, R. Totally contained AbioCor artificial heart implanted. Circulation 104, E9005–E9006 (2001). Murry, C. E., Wiseman, R. W., Schwartz, S. M. & Hauschka, S. D. Skeletal myoblast transplantation for repair of myocardial necrosis. J. Clin. Invest. 98, 2512–2523 (1996). Taylor, D. A. et al. Regenerating functional myocardium: improved performance after skeletal myoblast transplantation. Nature Med. 4, 929–933 (1998). Menasche, P. et al. Myoblast transplantation for heart failure. Lancet 357, 279–280 (2001). Fox, I. et al. Treatment of the Crigler–Najjar syndrome type I with hepatocyte transplantation. N. Engl. J. Med. 338, 1422–1426 (1998). Gunsalus, J. R., Brady, D. A., Coulter, S. M., Gray, B. M. & Edge, A. Reduction of serum cholesterol in watanabe rabbits by xenogeneic hepatocellular transplantation. Nature Med. 3, 48–53 (1997). Lagasse, E. et al. Purified hematopoietic stem cells can differentiate into hepatocytes in vivo. Nature Med. 6, 1229–1234 (2000). Clarke, D. L. et al. Generalized potential of adult neural stem cells. Science 288, 1660–1663 (2000). Zuk, P. A. et al. Multilineage cells from human adipose tissue: implications for cell-based therapies. Tissue Eng. 7, 211–228 (2001). Thomson, J. A. et al. Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147 (1998). Grobstein, C. Inductive epithelio–mesenchymal interaction in cultured organ rudiments of the mouse. Science 118, 52–55 (1953). Ekblom, P. Formation of basement membranes in the embryonic kidney: an immunohistological study. J. Cell Biol. 91, 1–10 (1981). Hammerman, M. R. Growing kidneys. Curr. Opin. Nephrol. Hypertens. 10, 13–17 (2001). Cascalho, M. & Platt, J. L. The immunological barrier to xenotransplantation. Immunity 14, 437–446 (2001). Dorling, A., Lombardi, G., Binns, R. & Lechler, R. I. Detection of primary direct and indirect human anti-porcine T cell responses using a porcine dendritic cell population. Eur. J. Immunol. 26, 1378–1387 (1996).


16. Gritsch, H. A. et al. The importance of nonimmune factors in reconstitution by discordant xenogeneic hematopoietic cells. Transplantation 57, 906–917 (1994). 17. Bennet, W. et al. Expression of complement regulatory proteins on islets of Langerhans: a comparison between human islets and islets isolated from normal and hDAF transgenic pigs. Transplantation 72, 312–319 (2001). 18. Murray, A. G., Khodadoust, M. M., Pober, J. S. & Bothwell, A. L. M. Porcine aortic endothelial cells activate human T cells: direct presentation of MHC antigens and costimulation by ligands for human CD2 and CD28. Immunity 1, 57–63 (1994). 19. Yamada, K., Sachs, D. H. & DerSimonian, H. Human anti-porcine xenogeneic T cell response. Evidence for allelic specificity of mixed leukocyte reaction and for both direct and indirect pathways of recognition. J. Immunol. 155, 5249–5256 (1995). 20. Platt, J. L. New directions for organ transplantation. Nature 392 (Suppl.), 11–17 (1998). 21. Auchincloss, H. Jr & Sachs, D. H. Xenogeneic transplantation. Annu. Rev. Immunol. 16, 433–470 (1998). 22. Marchetti, P. et al. Prolonged survival of discordant porcine islet xenografts. Transplantation 61, 1100–1102 (1996). 23. Deacon, T. et al. Histological evidence of fetal pig neural cell survival after transplantation into a patient with Parkinson’s disease. Nature Med. 3, 350–353 (1997). 24. Platt, J. L. Hyperacute xenograft rejection (RG Landes Co., Austin, Texas, 1995). 25. Galili, U., Clark, M. R., Shohet, S. B., Buehler, J. & Macher, B. A. Evolutionary relationship between the natural anti-Gal antibody and the Gal α1-3Gal epitope in primates. Proc. Natl Acad. Sci. USA 84, 1369–1373 (1987). 26. Zaidi, A. et al. Life-supporting pig-to-primate renal xenotransplantation using genetically modified donors. Transplantation 65, 1584–1590 (1998). 27. Lachmann, P. J. The control of homologous lysis. Immunol. Today 12, 312–315 (1991). 28. Hourcade, D., Holers, V. M. & Atkinson, J. P. The regulators of complement activation (RCA) gene cluster. Adv. Immunol. 45, 381–416 (1989). 29. Platt, J. L. et al. Transplantation of discordant xenografts: a review of progress. Immunol. Today 11, 450–456 (1990). 30. Van den Berg, C. W. & Morgan, B. P. Complementinhibiting activities of human CD59 and analogues from rat, sheep, and pig are not homologously restricted. J. Immunol. 152, 4095–4101 (1994). 31. Diamond, L. E. et al. Characterization of transgenic pigs expressing functionally active human CD59 on cardiac endothelium. Transplantation 61, 1241–1249 (1996). 32. Saadi, S. & Platt, J. L. Transient perturbation of endothelial integrity induced by natural antibodies and complement. J. Exp. Med. 181, 21–31 (1995). 33. McCurry, K. R. et al. Human complement regulatory proteins protect swine-to-primate cardiac xenografts from humoral injury. Nature Med. 1, 423–427 (1995). 34. Leventhal, J. R. et al. The immunopathology of cardiac xenograft rejection in the guinea pig-to-rat model. Transplantation 56, 1–8 (1993). 35. Blakely, M. L. et al. Activation of intragraft endothelial and mononuclear cells during discordant xenograft rejection. Transplantation 58, 1059–1066 (1994). 36. Shimizu, A. et al. Acute humoral xenograft rejection: destruction of the microvascular capillary endothelium in pig-to-nonhuman primate renal grafts. Lab. Invest. 80, 815–830 (2000). 37. Sachs, D. H. & Sablinski, T. Tolerance across discordant xenogeneic barriers. Xenotransplantation 2, 234–239 (1995). 38. Sablinski, T. et al. Long-term discordant xenogeneic (porcine-to-primate) bone marrow engraftment in a monkey treated with porcine-specific growth factors. Transplantation 67, 972–977 (1999). 39. Ierino, F. L. et al. Disseminated intravascular coagulation in association with the delayed rejection of pig-to-baboon renal xenografts. Transplantation 66, 1439–1450 (1998). 40. Onishi, A. et al. Pig cloning by microinjection of fetal fibroblast nuclei. Science 289, 1188–1190 (2000). 41. Polejaeva, I. A. et al. Cloned piglets produced by nuclear transfer from adult somatic cells. Nature 407, 86–90 (2000). 42. Betthauser, J. et al. Production of cloned pigs from in vitro systems. Nature Biotechnol. 18, 1055–1059 (2000).



PERSPECTIVES 43. Denning, C. et al. Deletion of the α(1,3) galactosyl transferase (GGTA1) gene and the prion protein (PrP) gene in sheep. Nature Biotechnology 19, 559–562 (2001). 44. Lin, S. S. et al. The role of anti-Galα1-3Gal antibodies in acute vascular rejection and accommodation of xenografts. Transplantation 70, 1667–1674 (2000). 45. Bach, F. H., Winkler, H., Ferran, C., Hancock, W. W. & Robson, S. C. Delayed xenograft rejection. Immunol. Today 17, 379–384 (1996). 46. Chopek, M. W., Simmons, R. L. & Platt, J. L. ABOincompatible renal transplantation: initial immunopathologic evaluation. Transplant. Proc. 19, 4553–4557 (1987). 47. Bannett, A. D., McAlack, R. F., Morris, M., Chopek, M. & Platt, J. L. ABO incompatible renal transplantation: a qualitative analysis of native endothelial tissue ABO antigens after transplant. Transplant. Proc. 21, 783–785 (1989). 48. Lin, Y., Vandeputte, M. & Waer, M. Accommodation and T-independent B cell tolerance in rats with long term surviving hamster heart xenografts. J. Immunol. 160, 369–375 (1998). 49. Hancock, W. W., Buelow, R., Sayegh, M. H. & Turka, L. A. Antibody-induced transplant arteriosclerosis is prevented by graft expression of anti-oxidant and anti-apoptotic genes. Nature Med. 4, 1392–1396 (1998). 50. Schmoeckel, M. et al. Orthotopic heart transplantation in a transgenic pig-to-primate model. Transplantation 65, 1570–1577 (1998). 51. Sablinski, T. et al. Xenotransplantation of pig kidneys to nonhuman primates: I. Development of the model. Xenotransplantation 2, 264–270 (1995). 52. Daggett, C. W. et al. Total respiratory support from swine lungs in primate recipients. J. Thorac. Cardiovasc. Surg. 115, 19–27 (1998). 53. Ramirez, P. et al. Life-supporting human complement regulator decay accelerating factor transgenic pig liver xenograft maintains the metabolic function and coagulation in the nonhuman primate for up to 8 days. Transplantation 70, 989–998 (2000). 54. Patience, C., Takeuchi, Y. & Weiss, R. A. Infection of human cells by an endogenous retrovirus of pigs. Nature Med. 3, 282–286 (1997). 55. Patience, C. et al. No evidence of pig DNA or retroviral infection in patients with short-term extracorporeal connection to pig kidneys. Lancet 352, 699–701 (1998). 56. Paradis, K. et al. Search for cross-species transmission of porcine endogenous retrovirus in patients treated with living pig tissue. Science 285, 1236–1241 (1999). 57. Andres, G. et al. Cellular mechanisms of adaptation of grafts to antibody. Transplant. Immunol. 4, 1–17 (1996). 58. Dalmasso, A. P., Benson, B. A., Johnson, J. S., Lancto, C. & Abrahamsen, M. S. Resistance against the membrane attack complex of complement induced in porcine endothelial cells with a Gal α(1-3)Gal binding lectin: up-regulation of CD59 expression. J. Immunol. 164, 3764–3773 (2000). 59. Delikouras, A., Hayes, M., Malde, P., Lechler, R. I. & Dorling, A. Nitric oxide-mediated expression of Bcl-2 and Bcl-xl and protection from TNFα–mediated apoptosis in porcine endothelial cells after exposure to low concentrations of xenoreactive natural antibody. Transplantation 71, 599–605 (2001). 60. Bach, F. H. et al. Accommodation of vascularized xenografts: expression of ‘protective genes’ by donor endothelial cells in a host TH2 cytokine environment. Nature Med. 3, 196–204 (1997). 61. Soares, M. P. et al. Expression of heme oxygenase-1 can determine cardiac xenograft survival. Nature Med. 4, 1073–1077 (1998). 62. Imai, T. et al. Vascular smooth muscle cell-directed overexpression of heme oxygenase-1 elevates blood pressure through attenuation of nitric oxide-induced vasodilation in mice. Circ. Res. 89, 55–62 (2001). 63. Holzknecht, Z. E. & Platt, J. L. Accommodation and the reversibility of biological systems. Transplantation 71, 594–595 (2001). 64. Magee, J. C. et al. Immunoglobulin prevents complement mediated hyperacute rejection in swine-to-primate xenotransplantation. J. Clin. Invest. 96, 2404–2412 (1995). 65. Nagayasu, T. et al. Expression of tissue factor mRNA in cardiac xenografts: clues to the pathogenesis of acute vascular rejection. Transplantation 69, 475–482 (2000). 66. Parker, W. et al. Transplantation of discordant xenografts: a challenge revisited. Immunol. Today 17, 373–378 (1996). 67. Cooper, D. K. C. et al. Effects of cyclosporine and antibody adsorption on pig cardiac xenograft survival in the baboon. J. Heart Transplant. 7, 238–246 (1988).

160

68. Lin, S. S. et al. The role of natural anti-Galα1-3Gal antibodies in hyperacute rejection of pig-to-baboon cardiac xenotransplants. Transplant Immunol. 5, 212–218 (1997). 69. Leventhal, J. R. et al. Prolongation of cardiac xenograft survival by depletion of complement. Transplantation 55, 857–866 (1993). 70. Pruitt, S. K. et al. The effect of soluble complement receptor type 1 on hyperacute rejection of porcine xenografts. Transplantation 57, 363–370 (1994). 71. Cozzi, E. et al. in Xenotransplantation: The Transplantation of Organs and Tissues Between Species (eds Cooper, D. K. C., Kemp, E., Platt, J. L. & White, D. J. G.) 665–682 (Springer, Berlin, 1997). 72. Diamond, L. E. et al. A human CD46 transgenic pig model system for the study of discordant xenotransplantation. Transplantation 71, 132–142 (2001). 73. Bracy, J. L., Sachs, D. H. & Iacomini, J. Inhibition of xenoreactive natural antibody production by retroviral gene therapy. Science 281, 1845–1847 (1998).

Acknowledgments Work in the laboratories of the authors is supported by grants from the National Institutes of Health and by the Von Liebig Foundation.

Online links DATABASES The following terms in this article are linked online to: LocusLink: http: //www.ncbi.nlm.nih.gov/LocusLink/ BCL2 | CD55 | CD59 | E-selectin | HO-1 | IL-1α | IL-1β | IL-1R | PAI-1 FURTHER INFORMATION Jeffrey Platt’s lab: http://www.mayo.edu/research/people/3/34611_platt/ Transplantation Society: http: //www.transplantationsoc.org/xeno.htm Access to this interactive links box is free online.

SCIENCE AND SOCIETY

Vaccine safety–vaccine benefits: science and the public’s perception Christopher B. Wilson and Edgar K. Marcuse The development of cowpox vaccination by Jenner led to the development of immunology as a scientific discipline. The subsequent eradication of smallpox and the remarkable effects of other vaccines are among the most important contributions of biomedical science to human health. Today, the need for new vaccines has never been greater. However, in developed countries, the public’s fear of vaccine-preventable diseases has waned, and awareness of potential adverse effects has increased, which is threatening vaccine acceptance. To further the control of disease by vaccination, we must develop safe and effective new vaccines to combat infectious diseases, and address the public’s concerns.

The discipline of immunology developed from observations in the fields of public health and clinical medicine. In the fifth century BC, Thucydides noted that individuals who recovered from plague did not develop disease again, and similar observations of ‘immunity’ to plague were made in Europe in the fourteenth century1,2. The observation that mild smallpox infection protected against disease on subsequent exposure led to the practice of variolation — the inoculation of dried pus from smallpox pustules into the skin or nose. This was first practised in India and China, and then introduced in 1721


in England by Lady Montague, and in New England by Cotton Mather3. Jenner’s clinical trial of cowpox virus vaccination, and publication of Variolae Vacciniae in 1798, gave birth to the field of immunology, but neither an understanding of the basis for its efficacy nor universal acceptance of this practice were soon to follow3,4. Instead, scientific and public scepticism and alarm were common early responses (FIG. 1). The subsequent formulation of the germ theory of disease by Koch and Pasteur, and von Berhing’s identification of neutralizing factors for toxins, provided a foundation for the mechanistic understanding of protective immunity5. In the ensuing years, vaccines for more than 20 infectious diseases have been developed, and in 1977, Jenner’s original experiment was brought to full fruition when smallpox was eradicated worldwide6. Immunization is one of the most stunning and economically effective contributions of biomedical science to human health. So, immunologists can be proud of the fundamental biomedical insights that have arisen from the field, and of the practical application of these insights in the prevention of disease. Advances of the last century allow us to better understand the successes (and failures) of past vaccines, and enable a more rational and diverse approach to new vaccine development. An example is the development of polysaccharide–protein conjugate vaccines



PERSPECTIVES

Figure 1 | The Cow-Pock. The Cow-Pock or the Wonderful Effects of the New Inoculation! by James Gillray was published in England in 1802 by the Anti-Vaccine Society. The etching, which shows Edward Jenner among patients in the Small Pox and Inoculation Hospital at St Pancras (London), suggests the transformation into cows of individuals vaccinated by Jenner. Reproduced with permission from The Wellcome Library, London.

against Haemophilus influenzae type b. These arose from observations showing: that antibodies to type b capsular polysaccharide are protective; that polysaccharides do not induce T-cell help and are not immunogenic in early life; and that linkage of polysaccharide to protein results in a T-cell-dependent antibody response to both components. Routine use of these vaccines has nearly eliminated meningitis and other diseases caused by H. influenzae type b6. Advances in our understanding of the determinants of protective immunity and immunological memory, of the mechanisms by which adjuvants affect the quality and magnitude of immunological responses, and of microbial genomics, offer the promise for new and more effective vaccines in the near future. Public concerns and vaccine safety

In recent years, a vocal minority in the developed world has questioned the safety and net benefits of vaccines. Vaccines are unique among medical interventions in that they are given to healthy individuals to prevent diseases that often do not pose an immediate threat to the recipient. Many vaccine-preventable diseases are now so infrequent that the only context in which many individuals have heard of these diseases is when hypothetical adverse effects of the relevant vaccine are presented by the

media as fact in an emotionally gripping story. We illustrate, through examples of real and falsely attributed adverse reactions to vaccines, the effects on vaccine use and development.

“…a perceived risk of a vaccine might outweigh concerns about the disease it is designed to prevent. Perceptions, be they true or false, drive behaviour.” Whooping cough. Whooping cough vaccines were developed in the late 1940s, by formalin inactivation of whole Bordetella pertussis, and were later combined with diphtheria and tetanus toxoids to create the DTP vaccine. Widespread early childhood immunization led to a marked reduction in the incidence of whooping cough. Because encephalopathy and encephalitis were well-recognized complications of smallpox vaccine7, encephalopathy presenting soon after immunization with a pertussis-containing vaccine was attributed to the vaccine8. As awareness of the severity of this disease faded, public concern emerged


regarding potential adverse reactions to the vaccine, including seizures, infantile spasms, encephalopathy and sudden infant death syndrome (SIDS)9,10. As a result, DTP vaccine usage declined in several countries, which was followed by a resurgence of the disease11 (FIG. 2). It was later determined that DTP does not cause SIDS12,13, infantile spasms14 or epilepsy15. In some children, DTP does cause transient fever, hypotonic-hyporesponsive episodes, protracted inconsolable crying and seizures, and might rarely cause acute encephalopathy (

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

Recommend Documents