commentary: tcr–mhc/peptide interactions: kissing-cousins or a shotgun wedding?

8
Commentary: TCR–MHC/peptide interactions: kissing-cousins or a shotgun wedding? Eric Huseby 1 , John Kappler 1,2,3 and Philippa Marrack 1,2,4 1 Howard Hughes Medical Institute and Integrated Department of Immunology, National Jewish Medical and Research Center, Denver, USA 2 Departments of Immunology and Medicine, University of Colorado Health Sciences Center, Denver, USA 3 Department of Pharmacology, University of Colorado Health Sciences Center, Denver, USA 4 Departments of Biochemistry and Molecular Genetics, University of Colorado Health Sciences Center, Denver, USA The purpose of this Commentary is to put into modern-day perspective Jerne’s hypothesis that antigen receptors encoded in the genome have been evolutionarily selected for their ability to react with major histocompatibility proteins and that the process of eliminating self reactivity is the catalyst for the generation of diversity of antigen receptors. In writing his hypothesis Jerne was trying to deal with the obsession of the immune system with the MHC, an obsession that was manifest in his days by the strong reactions of the immune system with allogeneic MHC proteins. However, Jerne’s hypothesis also took on other issues that were not understood at the time — issues that included lymphocyte selection and tolerance, the generation of somatic diversity and the ability of the MHC to control responses to other antigens. In so doing, Jerne generated a hypothesis that accounted remarkably satisfactorily for what was known in 1971. Whilst the details of much of the hypothesis have since turned out to be incorrect, in his ideas Jerne did anticipate many of the most interesting and surpris- ing findings of the subsequent 33 years. See accompanying article http://dx.doi.org/10.1002/eji.200425132 Key words: T cell / MHC / Positive selection / Negative selection / Alloreactivity Received 1/3/04 Accepted 9/3/04 [DOI 10.1002/eji.200425000] Abbreviation: Ir gene: Immune response gene 1 Historical setting Jerne postulated that antigen receptors that are encoded in the germline were evolutionarily selected for reactivity with MHC molecules of the particular species [1]. Jerne predicted that the germline bias towards species-specific MHC molecules would be the catalyst for antigen receptor diversification and expected that it would also explain many of the immunological phenom- ena of his time. At the time Jerne proposed his theory of the somatic generation of the immune response, a number of hypotheses and experimental data had been generated describing the workings of biology and the immune sys- tem. For example, many people had shown that muta- tion could occur, in both bacteria and eukaryotes, and that mutants could be selected for or against by various conditions [2]. The clonal selection theory was well established and evidence supporting its basic tenets was accumulating [3–5]. Thus it was appreciated that antigen receptors, which Jerne conceived in his hypoth- esis to be antibodies, but which have later turned out also to include and TCR molecules, varied in sequence and specificity from one cell to another [6]. It was guessed that the total number of antigen receptors in any individual was very large [7]. Evidence suggested that each lymphocyte bore receptors of only one sequence [8, 9]. It was also known that engagement of these receptors would stimulate the cell bearing them to divide [10]. The fact that the immune system is tolerant to self, and that, moreover, this tolerance is learned — “acquired” — by each individual as its immune system develops had also been established [11, 12]. The idea that tolerance involves the death of self-reactive cells had been pro- posed [13]. It was known that lymphocytes develop in Eur. J. Immunol. 2004. 34: 1243–1250 Generating diverse antigen receptors 1243 © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

Upload: eric-huseby

Post on 11-Jun-2016

217 views

Category:

Documents


4 download

TRANSCRIPT

Commentary:

TCR–MHC/peptide interactions: kissing-cousins ora shotgun wedding?

Eric Huseby1, John Kappler1,2,3 and Philippa Marrack1,2,4

1 Howard Hughes Medical Institute and Integrated Department of Immunology, National JewishMedical and Research Center, Denver, USA

2 Departments of Immunology and Medicine, University of Colorado Health Sciences Center,Denver, USA

3 Department of Pharmacology, University of Colorado Health Sciences Center, Denver, USA4 Departments of Biochemistry and Molecular Genetics, University of Colorado Health Sciences

Center, Denver, USA

The purpose of this Commentary is to put into modern-day perspective Jerne’s hypothesisthat antigen receptors encoded in the genome have been evolutionarily selected for theirability to react with major histocompatibility proteins and that the process of eliminating selfreactivity is the catalyst for the generation of diversity of antigen receptors. In writing hishypothesis Jerne was trying to deal with the obsession of the immune system with the MHC,an obsession that was manifest in his days by the strong reactions of the immune systemwith allogeneic MHC proteins. However, Jerne’s hypothesis also took on other issues thatwere not understood at the time — issues that included lymphocyte selection and tolerance,the generation of somatic diversity and the ability of the MHC to control responses to otherantigens. In so doing, Jerne generated a hypothesis that accounted remarkably satisfactorilyfor what was known in 1971. Whilst the details of much of the hypothesis have since turnedout to be incorrect, in his ideas Jerne did anticipate many of the most interesting and surpris-ing findings of the subsequent 33 years.See accompanying article http://dx.doi.org/10.1002/eji.200425132

Key words: T cell / MHC / Positive selection / Negative selection / Alloreactivity

Received 1/3/04Accepted 9/3/04

[DOI 10.1002/eji.200425000]

Abbreviation: Ir gene: Immune response gene

1 Historical setting

Jerne postulated that antigen receptors that areencoded in the germline were evolutionarily selected forreactivity with MHC molecules of the particular species[1]. Jerne predicted that the germline bias towardsspecies-specific MHC molecules would be the catalystfor antigen receptor diversification and expected that itwould also explain many of the immunological phenom-ena of his time.

At the time Jerne proposed his theory of the somaticgeneration of the immune response, a number ofhypotheses and experimental data had been generateddescribing the workings of biology and the immune sys-tem. For example, many people had shown that muta-

tion could occur, in both bacteria and eukaryotes, andthat mutants could be selected for or against by variousconditions [2]. The clonal selection theory was wellestablished and evidence supporting its basic tenetswas accumulating [3–5]. Thus it was appreciated thatantigen receptors, which Jerne conceived in his hypoth-esis to be antibodies, but which have later turned outalso to include § g and + ˇ TCR molecules, varied insequence and specificity from one cell to another [6]. Itwas guessed that the total number of antigen receptorsin any individual was very large [7]. Evidence suggestedthat each lymphocyte bore receptors of only onesequence [8, 9]. It was also known that engagement ofthese receptors would stimulate the cell bearing them todivide [10].

The fact that the immune system is tolerant to self, andthat, moreover, this tolerance is learned — “acquired” —by each individual as its immune system develops hadalso been established [11, 12]. The idea that toleranceinvolves the death of self-reactive cells had been pro-posed [13]. It was known that lymphocytes develop in

Eur. J. Immunol. 2004. 34: 1243–1250 Generating diverse antigen receptors 1243

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

particular organs — the bone marrow, bursa of Fabriciusand thymus are the organs quoted by Jerne — and that,at least in the case of the thymus, many more cells aregenerated in the organ than are exported to the rest ofthe animal [14, 15].

Finally Jerne was well aware of the high frequency oflymphocytes reactive with MHC proteins of the species[16–18]. He also knew that the allele of MHC proteinsexpressed in any given animal controlled its ability torespond to other, apparently unrelated, proteins [10].

In thinking about an all-encompassing hypothesis oflymphocyte development and specificity, Jerne wishedto account for all these phenomena and he had a tool toaccount for them — the ideas of somatic variation andselection derived from earlier very elegant studies onbacteria [2].

2 Jerne’s theory of antigen receptordiversification

Jerne based his hypothesis on the assumptions thatstem cells differentiate into mature lymphocytes withincertain primary lymphoid compartments — the thymus,bursa of Fabricius and bone marrow — and that there areonly a small number of antigen receptor V-genesencoded in the germline. He then hypothesized thatthese antigen receptor V-genes had been under a strongevolutionary selection to react with the MHC proteins ofthe particular species. He argued that because an indi-vidual inherits only a few of many possible alleles ofMHC, and because antigen receptor V-genes are not co-inherited with MHC, then the antigen receptor V-genescould not anticipate which MHC proteins would beexpressed by the individual. Thus the germline of thespecies must include genes coding for antigen receptorV-genes that recognize all possible MHC of the species.Because of this, V-genes within any individual would fallinto one of two groups: one group coding for antigenreceptors that would bind to the MHC proteinsexpressed by the individual (self or subset S) andanother group directed at all other relevant MHC proteinsof the species (allo or subset A).

His thoughts were that if a developing lymphocyteexpressed an antigen receptor V-gene that was reactivewith a self-MHC protein (subset S), the lymphocytewould recognize the expressed MHC and be stimulatedto proliferate at this immature stage of development.During these rounds of proliferation, the antigen receptorV-gene would be able to undergo mutational diversifica-tion through spontaneous somatic mutation. After thisproliferation, two outcomes were possible. If the self-

reactive antigen receptor V-gene failed to incorporate amutation or the mutation failed to eliminate self-MHCreactivity, the developing lymphocyte would be killed;however, if the mutation caused the antigen receptor tolose its self-MHC reactivity, the lymphocyte bearing itwould be allowed to survive and exit into the secondarylymphoid compartment. Thus, the immune systemwould use the evolutionarily selected reactivity betweenantigen receptor and self-MHC as the catalyst to initiatethe process of V-gene diversification.

If a developing lymphocyte expressed an antigen recep-tor V-gene that was reactive with an allo-MHC protein(subset A), there would be no issue of self-tolerance toovercome and the lymphocyte would be allowed todirectly exit into the periphery without going through theprocess of receptor diversification and with an undeter-mined amount of proliferation. Lymphocytes expressingthese non-mutated germline-encoded allo-MHC-reactive antigen receptors would thus be present in theperiphery of all individuals, probably at fairly high fre-quency. These cells would thus account for the findingsthat about 2% of lymphocytes could respond in a mixedlymphocyte culture with MHC-mismatched cells of thesame species or participate in allogeneic graft rejection[16, 18].

In his hypothesis Jerne also had to account for immuneresponse genes (Ir genes) — the fact that some individu-als are unable to respond to certain antigens and thatthis lack of response maps genetically to the MHC locus[19–21]. Jerne explained this phenomenon by suggest-ing that antigens that are affected by Ir genes requirevery particular antigen receptors to recognize them.These particular receptors could not be generated bysomatic mutation of receptors specific for certain MHCalleles — the alleles that fail to allow responses to thoseantigens.

Jerne’s ideas fit nicely with the requirements of self-tolerance and clonal selection, the phenomena of allore-activity and the massive proliferation of lymphocyteswithin the thymus as well as providing a method for thegeneration of antigen receptor V-gene diversity.

3 How right was he?

In spite of the intrinsic beauty and internal consistency ofJerne’s ideas we now know that most of them haveturned out to be incorrect. These points are summarizedin Table 1 and Fig. 1. Because a single receptor is gener-ated from at least four variable parts, each of which canbe selected from a number of components, the totalnumber of receptors encoded in the germline is greater

1244 E. Huseby et al. Eur. J. Immunol. 2004. 34: 1243–1250

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

Table 1. Modern explanations of the questions answered by Jerne

Question Jerne Modern immunology

Germline lymphocyte antigen receptorspecificity.

MHC of species. § / g TCR MHC of species?B cells Random/pathogen.

+ / ˇ TCR ?

Size of germline antigen receptorrepertoire.

Relatively small,estimate = 100.

Relatively large,estimate G 107.

Immature lymphocyte proliferation. Reaction with self-MHC. Cytokines and pre-TCR/BCR.

The catalyst for antigen receptorsomatic diversity.

Reaction with self-MHC. T cells None.Immature B cells None.Mature B cells Reaction

with antigen.

The mechanisms for antigen receptorsomatic diversity.

Somatic mutations arising during anti-self proliferation.

Somatic recombination prior to expressionof functional antigen receptor for T cells

and immature B cells.Somatic hypermutation of mature B cells

after interaction with pathogen.

Death of lymphocytes in the generatingorgan.

Deletion of selfreactive cells.

Predominately death by neglectSome deletion of self reactive cells.

High frequency of allo reactivity. Germline receptor repertoire specific forMHC of the species.

Increased avidity of self-MHC + foreignantigen specific T cells for

allo-MHC + peptide.

Basis of receptor specificity for foreignantigens.

Elimination of self reactive antigenreceptors.

The ability of TCR to interact with lowaffinity for self-MHC + self-peptide.

The elimination of self reactiveBCR and TCR.

than 107 [22]. There is no evidence that antibody geneshave been selected to code for proteins that react withthe MHC of the species; however, some experimentshave suggested that the genes for § g TCR moleculeshave evolved do this [23–25]. Evidence on this point is farfrom complete, and is clouded by the fact that positiveselection in the thymus for weak reactivity with self MHCbound to self peptides certainly contributes to the highfrequency with which § g TCR+ T cells react with alloge-neic MHC bound to self or foreign peptides [26–29].

Though there is extensive proliferation within the thymus,reaction of immature lymphocytes with self does not leadto proliferation of the cells. On the contrary we now knowthat this event either causes the self-reactive lymphocyteto die [30, 31] or, in the case of B cells, change its recep-tor genes by the process of editing [32–34]. Jerne cor-rectly predicted, as had Lederberg before him [13], thatimmature lymphocytes that react too well with self die.However, they (at least T cells) do not usually avoid thisdeath by mutating their receptor genes [35, 36].

One might allow that receptor editing is a version of suchmutation, and if this is included then Jerne’s prediction of

avoidance-of-death-by-receptor-mutation holds true, forB cells but probably not for T cells [32, 37, 38]. Even inthis case, mutation is not caused by spontaneous eventsand rapid division in response to antigen stimulation butrather by specific reinduction of RAG enzymes and theircohorts [39, 40]. Jerne’s hypotheses about recognition ofMHC and clonal deletion in the thymus were driven inpart by the knowledge that the thymus creates far morecells than are exported [14, 15]. However, we now knowthat most of this death is caused by neglect — the deathof nonselected cells — rather than being caused by rec-ognition of self [41, 42].

In the mature lymphocyte repertoire, hypermutation ofBCR genes does occur, and it is induced, as Jerne pre-dicted, by engagement of the BCR by antigen. However,in this case the antigen is usually foreign, and the hyper-mutation is caused by specific enzymes and not sponta-neous mutational events [43].

Finally, at the time Jerne wrote this article many peoplebelieved that Ir genes, linked to the MHC locus, codedfor TCR molecules themselves. The failure of animalswith certain MHC alleles to respond to certain antigens

Eur. J. Immunol. 2004. 34: 1243–1250 Generating diverse antigen receptors 1245

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

Fig. 1. A comparison of Jerne’s ideas with modern ideas about the development of the lymphocyte repertoire.

1246 E. Huseby et al. Eur. J. Immunol. 2004. 34: 1243–1250

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

was therefore due to MHC-linked absence of certainreceptor genes [19–21]. Jerne correctly realized that Irgenes might not code for lymphocyte receptors butrather reflect some impact of MHC alleles on antigen rec-ognition by lymphocytes. He incorrectly suggested thatthis may be caused by a hole in the antigen receptor rep-ertoire due to inability to generate certain specificities bymutation of a particular set of lymphocyte receptorgenes. However, he was closer to the truth than mostother people at the time. Now we know that failure torespond to antigen is usually due to failure of the antigento give rise to a peptide that can be presented by theMHC of the individual [44–46]. Ir genes are thus linked toMHC because certain MHC alleles cannot bind certainpeptides.

4 Why were some of Jerne’s ideas wrong?

The main reason why Jerne could not guess what wenow believe to be the truth is that he was missing essen-tial information. The most crucial of this was knowledgeof the structure and properties of MHC [47–49], in partic-ular the ability of single MHC molecules to bind many,but not all, peptides [50], and the requirement by T cellsto react with antigen only when bound to MHC [51]. Like-wise although he and others correctly anticipated thephenomenon of central tolerance, he could not have pre-dicted the idea of positive selection, the notion that T cellprecursors are selected to survive based on weak reac-tivity with self MHC and peptide [52, 53]. Although it waswell known that many more lymphocytes are producedin than exported from the thymus, the notion of death-by-neglect had certainly not come up [41, 42]. Thus Jer-ne’s explanation that most cells are killed in the thymusby reaction with self was reasonable at the time and con-sistent with the notions of central tolerance.

On the other hand, Jerne failed to include some factsthat were known at the time he was writing. In particularit was at the time well recognized that the specificimmune system included both B and T cells [54]. Thesetwo groups of lymphocytes were known to be producedin different organs [15]. It was also known that B cellssecrete antibody and that T cells drive what was called atthe time cell-mediated immunity — a term that encom-passed, among other responses, delayed-type hyper-sensitivity reactions and mixed lymphocyte reactions[55]. It was also known that mixed lymphocyte reactionsare directed against allogeneic MHC [10]. Jerne couldhave made more of these known differences between Band T cells but he chose not to, probably for the sake ofsimplicity and elegance.

5 Second-generation models ofalloreactivity

5.1 The role of peptides

The problem of T cell recognition of allogeneic MHC con-tinues to be a subject of interest and controversy even tothis day. One of the major issues has arisen from the real-ization that nearly all MHC proteins are expressed boundto a ligand, usually a peptide. Thus in uninfected animalsany given MHC protein is bound to a collection of pep-tides derived from self proteins [44–46]. Recognition ofallogeneic MHC could therefore also include recognitionof these peptides. Consequently, two types of hypothe-ses about allogeneic MHC reaction have been put for-ward.

One theory, derived from the finding that cytotoxic T cellscan recognize minor histocompatibility antigens in asso-ciation with self-MHC proteins, suggests that because ofsequence polymorphisms allogeneic MHC moleculeswill present a diverse set of peptides which the TCR rep-ertoire has not been tolerized to [56, 57]. This hypothesisgives rise to the idea that alloreactive T cells recognizeallo-MHC + peptide (self or foreign) in a manner that ispeptide specific, just like conventional T cells recogniz-ing self-MHC + foreign peptides.

An alternative hypothesis suggests that because MHCmolecules may have polymorphisms that are present atthe interface between the TCR and peptide/MHC com-plex or because of peptide-induced conformationalchanges in the MHC proteins, alloreactivity may arisedue to the acquisition of additional binding energy fromhigh density, low affinity interactions with these new con-formations [58–60]. According to these hypotheses,bound peptides may not themselves be directly involvedin the interaction between T cells and foreign MHC.

Both of these theories can and have been tested byexperiments and evidence suggests that both ideas arecorrect. Some alloreactive T cells are highly peptide spe-cific whereas others show very peptide-degenerate rec-ognition of MHC proteins [29, 60]. Further evidence hassuggested that the role of peptides in the CTL responseto allogeneic class I MHC is determined by the similarityof self-MHC to the allo-MHC in question [61]. Interest-ingly, the peptide specificity based on homologous prox-imity to self is not necessarily true for CD4 T cells recog-nizing MHC class II proteins [62].

On the other hand, T cells have been found that recog-nize allogeneic MHC and are completely peptide-independent. Interestingly here again there may be a dis-tinction between class I and class II MHC. Peptide-

Eur. J. Immunol. 2004. 34: 1243–1250 Generating diverse antigen receptors 1247

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

independent recognition of class I has been demon-strated, whereas the analogous situation for class II rec-ognition has yet to be documented [61–64]. Perhaps thisdistinction reflects the longer, more exposed configura-tion of peptides bound to MHC class II versus class I.

5.2 The role of positive and negative selection

Clearly the predilection of § g TCR molecules to reactwith peptide/self-MHC complexes is imposed to someextent by positive selection. Since MHC proteins,regardless of their allele, share many structural features,a byproduct of positive selection could be relatively fre-quent selection of T cells that recognize self MHC + selfpeptides weakly, but allogeneic MHC ± peptidesstrongly. Less appreciated, however, is the role of toler-ance. Evidence that clonal deletion in the thymus playssome role in the appearance of allo-reactive T cellscomes from mice in which tolerance is limited, usually byexpression of MHC proteins bound to only one or only afew peptides (rather than the usual thousands). Forexample, 50–60% of the CD4+ T cells in mice expressingMHC class II molecules presenting only a single peptide(MHC-SP) react with allogeneic MHC, whereas CD4+ Tcells in normal mice have only a 0.5–5.0% chance ofsuch reactivity [18, 65, 66].

One interpretation of this result is that MHC normallyexists in many conformations that are common to allMHC. Each conformation is induced by a particular setof peptides. In MHC-SP mice only one of these confor-mations is present, hence T cells specific for the otherconformations escape central tolerance and emerge,able to react with the other conformations of self andallogeneic MHC. Such T cells should be relativelypeptide-independent in their reaction with MHC, andsuch, indeed, has been shown by us to be the case ([62,67] and unpublished data).

5.3 The role of germline receptors

In spite of the powerful effects of positive and negativeselection it is still possible that the bias of § g TCR mole-cules to react with MHC is predetermined by the struc-ture of germline-encoded TCR § and g chains, as pre-dicted by Jerne. There is evidence that this is so. Ran-dom pairs of TCR § and g chains give rise to TCR mole-cules that can react with MHC [23]. Thymocytes thathave not undergone positive and negative selection rec-ognize, with unexpectedly high frequency, MHC proteins[24, 25]. T cells have been found that cross-reactbetween class I and class II MHC proteins, suggestingsome overall intrinsic reactivity with the shape of the

MHC protein ([68, 69] and unpublished data). Of coursethese phenomena could all be helped along by the factthat there are germline-encoded proteins that do reactwith class I and class II of any allele — these proteins arethe coreceptors, CD4 and CD8. Nevertheless it is stillpossible that the germline-encoded geometry of the Vregion loops that contact the MHC — the CDR1 andCDR2 loops — may be the kissing-cousins of MHC pro-teins, imparting MHC reactivity on the TCR. Theextremely random CDR3 regions will then impart peptidediscrimination on the molecules.

6 Jerne’s legacy

Much of Jerne’s 1971 hypothesis has turned out to bewrong. Jerne was right about some things, though. In thearticle he correctly identified most of the immunologicalconundrums of the time. The solutions to these puzzleshave turned out to be the keys to the huge advances inour understanding of immune recognition that haveoccurred in the intervening 33 years. Thus Jerne may nothave been right but with his courageous insights he cer-tainly showed everyone else the way.

Acknowledgements: We thank are our senior colleagues,Drs Richard Dutton, Jan Klein, Reinhard Obst, Claes Ohlen,Darcy Wilson, Ethan Shevach and Hugh McDevitt, whohelped us reconstruct the knowledge base of 33 years agoand apologize for the many thousands of studies on thenature of TCR/MHC interactions we were unable to cite. Thiswork was supported by USPHS grant AI 18785.

References

1 Jerne, N. K., The somatic generation of immune regulation. Eur.J. Immunol. 1971. 1: 1–9.

2 Luria, S. E. and Delbruck, M., Mutations of bacteria from virussensitivity to virus resistance. Genetics 1943. 28: 491–511.

3 Talmage, D. W., Allergy and immunology. Annu. Rev. Med. 1957.8: 239–256.

4 Burnet, F. M., The Clonal Selection Theory of Aquired Immunity.Cambridge University Press, Cambridge 1959.

5 Sell, S. and Gell, P. G., Studies on rabbit lymphocytes in vitro. I.Stimulation of blast transformation with an antiallotype serum.J. Exp. Med. 1965. 122: 423–440.

6 Pernis, B., Forni, L. and Amante, L., Immunoglobulin spots onthe surface of rabbit lymphocytes. J. Exp. Med. 1970. 132:1001–1018.

7 Ada, G. L. and Byrt, P., Specific inactivation of antigen-reactivecells with 125I-labelled antigen. Nature 1969. 222: 1291–1292.

8 Pernis, B., Chiappino, G., Kelus, A. S. and Gell, P. G., Cellularlocalization of immunoglobulins with different allotypic specifici-ties in rabbit lymphoid tissues. J. Exp. Med. 1965. 122: 853–876.

1248 E. Huseby et al. Eur. J. Immunol. 2004. 34: 1243–1250

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

9 Cebra, J. J., Colberg, J. E. and Dray, S., Rabbit lymphoid cellsdifferentiated with respect to alpha-, gamma-, and mu-heavypolypeptide chains and to allotypic markers Aa1 and Aa2. J. Exp.Med. 1966. 123: 547–558.

10 Dutton, R. W., Spleen cell proliferation in response to homolo-gous antigens studied in congenic resistant strains of mice. J.Exp. Med. 1966. 123: 665–671.

11 Owen, R. D., Immunogenetic consequences of vascular anasto-moses between bovine twins. Science 1945. 102: 400–401.

12 Billingham, R. E., Brent, L. and Medawar, P. B., Activityacquired tolerance of foreign cells. Nature 1953. 172: 603–606.

13 Lederberg, J., Genes and antibodies. Science 1959. 129:1649–1653.

14 Matsuyama, M., Wiadrowski, M. N. and Metcalf, D., Autoradio-graphic analysis of lymphopoiesis and lymphocyte migration inmice bearing multiple thymus grafts. J. Exp. Med. 1966. 123:559–576.

15 Miller, J. F. and Mitchell, G. F., Thymus and antigen-reactivecells. Transplant. Rev. 1969. 1: 3–42.

16 Nisbet, N. W., Simonsen, M. and Zaleski, M., The frequency ofantigen-sensitive cells in tissue transplantation. A commentaryon clonal selection. J. Exp. Med. 1969. 129: 459–467.

17 Lafferty, K. J. and Jones, M. A., Reactions of the graft versushost (GVH) type. Aust. J. Exp. Biol. Med. Sci. 1969. 47: 17–54.

18 Wilson, D. B. and Nowell, P. C., Quantitative studies on themixed lymphocyte interaction in rats. IV. Immunologic potentialityof the responding cells. J. Exp. Med. 1970. 131: 391–407.

19 Levine, B. B. and Benacerraf, B., Genetic control in guinea pigsof immune response to conjugates of haptens and poly-L-lysine.Science 1965. 147: 517–518.

20 Levine, B. B., Ojeda, A. and Benacerraf, B., Studies on artificialantigens. Iii. The genetic control of the immune response tohapten-poly-L-lysine conjugates in guinea pigs. J. Exp. Med.1963. 118: 953–957.

21 McDevitt, H. O. and Tyan, M. L., Genetic control of the antibodyresponse in inbred mice. Transfer of response by spleen cells andlinkage to the major histocompatibility (H-2) locus. J. Exp. Med.1968. 128: 1–11.

22 Davis, M. M. and Bjorkman, P. J., T-cell antigen receptor genesand T-cell recognition. Nature 1988. 334: 395–402.

23 Blackman, M., Yague, J., Kubo, R., Gay, D., Coleclough, C.,Palmer, E., Kappler, J. and Marrack, P., The T cell repertoiremay be biased in favor of MHC recognition. Cell 1986. 47:349–357.

24 Zerrahn, J., Held, W. and Raulet, D. H., The MHC reactivity ofthe T cell repertoire prior to positive and negative selection. Cell1997. 88: 627–636.

25 Merkenschlager, M., Graf, D., Lovatt, M., Bommhardt, U.,Zamoyska, R. and Fisher, A. G., How many thymocytes auditionfor selection? J. Exp. Med. 1997. 186: 1149–1158.

26 Zinkernagel, R. M., Callahan, G. N., Althage, A., Cooper, S.,Klein, P. A. and Klein, J., On the thymus in the differentiation of“H-2 self-recognition” by T cells: evidence for dual recognition?J. Exp. Med. 1978. 147: 882–896.

27 Fink, P. J. and Bevan, M. J., H-2 antigens of the thymus deter-mine lymphocyte specificity. J. Exp. Med. 1978. 148: 766–775.

28 Marrack, P., Bender, J., Jordan, M., Rees, W., Robertson, J.,Schaefer, B. C. and Kappler, J., Major histocompatibility com-

plex proteins and TCRs: do they really go together like a horseand carriage? J. Immunol. 2001. 167: 617–621.

29 Housset, D. and Malissen, B., What do TCR-pMHC crystalstructures teach us about MHC restriction and alloreactivity?Trends Immunol. 2003. 24: 429–437.

30 Nossal, G. J. and Pike, B. L., Evidence for the clonal abortiontheory of B-lymphocyte tolerance. J. Exp Med. 1975. 141:904–917.

31 Kappler, J. W., Roehm, N. and Marrack, P., T cell tolerance byclonal elimination in the thymus. Cell 1987. 49: 273–280.

32 Tiegs, S. L., Russell, D. M. and Nemazee, D., Receptor editingin self-reactive bone marrow B cells. J. Exp. Med. 1993. 177:1009–1020.

33 Goodnow, C. C., Balancing immunity and tolerance: deletingand tuning lymphocyte repertoires. Proc. Natl. Acad. Sci. U S A1996. 93: 2264–2271.

34 Casellas, R., Shih, T. A., Kleinewietfeld, M., Rakonjac, J.,Nemazee, D., Rajewsky, K. and Nussenzweig, M. C., Contri-bution of receptor editing to the antibody repertoire. Science2001. 291: 1541–1544.

35 Marrack, P., Shimonkevitz, R., Hannum, C., Haskins, K. andKappler, J., The major histocompatibility complex-restrictedantigen receptor on T cells. IV. An antiidiotypic antibody predictsboth antigen and I-specificity. J. Exp. Med. 1983. 158:1635–1646.

36 Jorgensen, J. L., Esser, U., Fazekas de St Groth, B., Reay, P.A. and Davis, M. M., Mapping T-cell receptor-peptide contactsby variant peptide immunization of single-chain transgenics.Nature 1992. 355: 224–230.

37 Buch, T., Rieux-Laucat, F., Forster, I. and Rajewsky, K., Failureof HY-specific thymocytes to escape negative selection byreceptor editing. Immunity 2002. 16: 707–718.

38 Holman, P. O., Walsh, E. R. and Hogquist, K. A., The central tol-erance response to male antigen in normal mice is deletion andnot receptor editing. J. Immunol. 2003. 171: 4048–4053.

39 Fugmann, S. D., Lee, A. I., Shockett, P. E., Villey, I. J. andSchatz, D. G., The RAG proteins and V(D)J recombination: com-plexes, ends, and transposition. Annu. Rev. Immunol. 2000. 18:495–527.

40 Jung, D. and Alt, F. W., Unraveling V(D)J recombination; insightsinto gene regulation. Cell 2004. 116: 299–311.

41 Berg, L. J., Pullen, A. M., Fazekas de St Groth, B., Mathis, D.,Benoist, C. and Davis, M. M., Antigen/MHC-specific T cells arepreferentially exported from the thymus in the presence of theirMHC ligand. Cell 1989. 58: 1035–1046.

42 Scott, B., Bluthmann, H., Teh, H. S. and von Boehmer, H., Thegeneration of mature T cells requires interaction of the alpha betaT-cell receptor with major histocompatibility antigens. Nature1989. 338: 591–593.

43 Papavasiliou, F., Casellas, R., Suh, H., Qin, X. F., Besmer, E.,Pelanda, R., Nemazee, D., Rajewsky, K. and Nussenzweig, M.C., V(D)J recombination in mature B cells: a mechanism for alter-ing antibody responses. Science 1997. 278: 298–301.

44 Hunt, D. F., Michel, H., Dickinson, T. A., Shabanowitz, J., Cox,A. L., Sakaguchi, K., Appella, E., Grey, H. M. and Sette, A.,Peptides presented to the immune system by the murine class IImajor histocompatibility complex molecule I-Ad. Science 1992.256: 1817–1820.

45 Rudensky, A., Preston-Hurlburt, P., Hong, S. C., Barlow, A.and Janeway, C. A. Jr., Sequence analysis of peptides bound toMHC class II molecules. Nature 1991. 353: 622–627.

Eur. J. Immunol. 2004. 34: 1243–1250 Generating diverse antigen receptors 1249

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de

46 Falk, K., Rotzschke, O., Stevanovic, S., Jung, G. and Ram-mensee, H. G., Allele-specific motifs revealed by sequencing ofself-peptides eluted from MHC molecules. Nature 1991. 351:290–296.

47 Unanue, E. R., Antigen-presenting function of the macrophage.Annu. Rev. Immunol. 1984. 2: 395–428.

48 Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S.,Strominger, J. L. and Wiley, D. C., The foreign antigen bindingsite and T cell recognition regions of class I histocompatibilityantigens. Nature 1987. 329: 512–518.

49 Townsend, A. and Bodmer, H., Antigen recognition by class I-restricted T lymphocytes. Annu. Rev. Immunol. 1989. 7: 601–624.

50 Babbitt, B. P., Allen, P. M., Matsueda, G., Haber, E. andUnanue, E. R., Binding of immunogenic peptides to Ia histocom-patibility molecules. Nature 1985. 317: 359–361.

51 Zinkernagel, R. M. and Doherty, P. C., Restriction of in vitro Tcell-mediated cytotoxicity in lymphocytic choriomeningitis withina syngeneic or semiallogeneic system. Nature 1974. 248:701–702.

52 Bevan, M. J., In a radiation chimaera, host H-2 antigens deter-mine immune responsiveness of donor cytotoxic cells. Nature1977. 269: 417–418.

53 Sebzda, E., Mariathasan, S., Ohteki, T., Jones, R., Bachmann,M. F. and Ohashi, P. S., Selection of the T cell repertoire. Annu.Rev. Immunol. 1999. 17: 829–874.

54 Claman, H. N., Chaperon, E. A. and Triplett, R. F., Thymus-marrow cell combinations. Synergism in antibody production.Proc. Soc. Exp. Biol. Med. 1966. 122: 1167–1171.

55 Johnston, J. M. and Wilson, D. B., Origin of immunoreactivelymphocytes in rats. Cell. Immunol. 1970. 1: 430–444.

56 Bevan, M. J., Interaction antigens detected by cytotoxic T cellswith the major histocompatibility complex as modifier. Nature1975. 256: 419–421.

57 Matzinger, P. and Bevan, M. J., Hypothesis: why do so manylymphocytes respond to major histocompatibility antigens? Cell.Immunol. 1977. 29: 1–5.

58 Sherman, L. A., Recognition of conformational determinants onH-2 by cytolytic T lymphocytes. Nature 1982. 297: 511–513.

59 Bevan, M. J., High determinant density may explain the phenom-enon of alloreactivity. Immunol. Today 1984. 5: 128–130.

60 Sherman, L. A. and Chattopadhyay, S., The molecular basis ofallorecognition. Annu. Rev. Immunol. 1993. 11: 385–402.

61 Obst, R., Netuschil, N., Klopfer, K., Stevanovic, S. and Ram-mensee, H. G., The role of peptides in T cell alloreactivity isdetermined by self-major histocompatibility complex molecules.J. Exp. Med. 2000. 191: 805–812.

62 Huseby, E. S., Crawford, F., White, J., Kappler, J. and Mar-rack, P., Negative selection imparts peptide specificity to themature T cell repertoire. Proc. Natl. Acad. Sci. U S A 2003. 100:11565–11570.

63 Smith, P. A., Brunmark, A., Jackson, M. R. and Potter, T. A.,Peptide-independent recognition by alloreactive cytotoxic T lym-phocytes (CTL). J. Exp. Med. 1997. 185: 1023–1033.

64 Elliott, T. J. and Eisen, H. N., Cytotoxic T lymphocytes recognizea reconstituted class I histocompatibility antigen (HLA-A2) as anallogeneic target molecule. Proc. Natl. Acad. Sci. U S A 1990. 87:5213–5217.

65 Ignatowicz, L., Kappler, J. and Marrack, P., The repertoire of Tcells shaped by a single MHC/peptide ligand. Cell 1996. 84:521–529.

66 Suchin, E. J., Langmuir, P. B., Palmer, E., Sayegh, M. H., Wells,A. D. and Turka, L. A., Quantifying the frequency of alloreactiveT cells in vivo: new answers to an old question. J. Immunol. 2001.166: 973–981.

67 Crawford, F., Huseby, E., White, J., Marrack, P. and Kappler,J., Identification of antigen mimetopes for broadly reactive andconventional T cells using a baculovirus peptide/MHC displaylibrary. PLOS:Biology 2004. In press.

68 Arsov, I. and Vukmanovic, S., Dual MHC class I and class IIrestriction of a single T cell receptor: distinct modes of toleranceinduction by two classes of autoantigens. J. Immunol. 1999. 162:2008–2015.

69 Morgan, D. J., Nugent, C. T., Raveney, B. J. and Sherman, L.A., In a transgenic model of spontaneous autoimmune diabetes,expression of a protective class II MHC molecule results in thy-mic deletion of diabetogenic CD8+ T cells. J. Immunol. 2004.172: 1000–1008.

Correspondence: Eric Huseby or Philippa Marrack, HowardHughes Medical Institute and Integrated Department ofImmunology, National Jewish Medical and Research Center,Denver, CO 80206, USAFax: +1-303-398-1396e-mail: husebye — njc.org or marrackp — njc.org

1250 E. Huseby et al. Eur. J. Immunol. 2004. 34: 1243–1250

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji.de