Why is the adaptive immune system so diverse and specific?
What will you learn today?
Self tolerance is more complicated than what the textbook says
Repertoires are very diverse because lymphocytes have to be specific
Simple mathematical models help us thinking clearly
Polymorphism versus diversity of MHC First short refresh on tolerance Death by neglect 0.3
R/R0
0.2
0.1
0 1 2 3 4 MHC diversity (log M )
Figure 2: Positive and negative selection according to the avidity model [13]. The curve in (a) depicts the distribution of thymocyte avidities for self peptide–MHC complexes. In our model, the chance p to be positively selected by a single MHC type is the chance that the avidity between the thymocyte T cell receptor and any of the self peptide– MHC complexes exceeds threshold T1. Thymocytes with avidities for self peptide–MHC complexes exceeding the upper threshold T2 are negatively selected (with chance n per MHC type). Panel (b) depicts the size of the T cell repertoire as a function of MHC diversity. The number of clones in the functional repertoire R is plotted as a fraction of the total initial lymphocyte repertoire R0. Parameters are: p =0.01, and n =0.005.
2 MHC diversity within the individual
Since individual MHC diversity increases the presentation of pathogens to the immune system, one may wonder why the number of MHC genes is not much higher than it is. The argument that is mostly invoked is that more MHC diversity within the individual would lead to T cell repertoire depletion during self tolerance induction. This argument is incomplete, however, because more MHC diversity could also increase the number of clones in the T cell repertoire through positive selection. In order to be rescued in the thymus, lymphocytes need to recognize MHC–self peptide complexes with su⇤cient avidity. A high MHC diversity thus increases both the number of lymphocyte clones that are positively selected and the number of clones that are negatively selected. To calculate the net e ect of these two opposing processes we develop a simple mathematical model [5].
Consider an individual with M di erent MHC molecules and an initial T lymphocyte repertoire consisting of R0 di erent clones. Let p and n denote the (unconditional) chances that a clone is positively selected by a single MHC type, because its avidity is higher than a threshold T1, or negatively selected because its avidity exceeds a higher threshold T2, respectively (see Fig. 2). By this definition, thymocytes can only be negatively selected by MHC molecules by which they are also positively selected, i.e. n
R = R (1 n)M (1 p)M , (7) 0 [5]. The functional repertoire R thus contains all T cell clones that⇥ fail to be negatively selected, minus the ones that also fail to be positively selected by any of the M di erent MHC molecules of the host.
Experimental estimates for the parameters of this model have recently become available. In mice, around 3% of the T cells produced in the thymus end up in the mature T cell repertoire, and at least 50% of all positively selected T cells have been shown to undergo negative selection in the thymus [17]. Thus, 94% of all thymic T cells fail to be positively selected by any of the MHC molecules in the host [17]. These estimates can be used to calculate the chances p and n of a T cell clone to be positively or negatively selected by a single type of MHC molecule. Taking into account that inbred mice are homozygous and therefore express 3 types of class I MHC and 3 types of class II MHC molecules, p and n follow from: (1 p)6 =0.94 and n = p/2. This yields p =0.01 and n =0.005.
Using these experimental estimates, the number of clones in the functional T cell repertoire R increases with
3 What is the diversity of the T cell repertoire? Parham cites estimate from Arstila [Science, 1999]: 2.5x107 Nowadays next generation sequencing (NGS):
The current study was designed explicitly to improve estima- A CD4 naive T cells CD8 naive T cells tion of the TCR repertoire richness of stringently defined naïve p=0.008 p=0.008 and memory T cells. By sequencing multiple replicate TCRB li- 2x108 braries of cells from each T-cell subset and applying nonparametric 1x108 statistical analysis, we find that human TCR repertoires are an order of magnitude more diverse than previously estimated. 5x107 Despite significant age-related decreases in richness, humans 7 maintain high diversity during healthy aging. Strikingly, we find 2x10 age-associated changes in the distribution of clonal sizes, in 1x107 particular in the naïve compartment, which may reflect un- Distinct TCRB sequences evenness of homeostatic expansion with clonal expansion of 20 - 35 70 - 85 20 - 35 70 - 85 2014 PNAS Qial. et some naïve T cells that equal or exceed clonal sizes of memory T B p=0.008Age p=0.008Age cells. The inequalities in clonal sizes could compromise the im- 4x107 mune response to the majority of antigenic epitopes while Note that this is β-chains only (nucleotides & Chao2)! causing an increased responsiveness to selected few epitopes. 2x107
sequences In mice almost every naive T cell unique TCR β Results 1x107 Gene Segment Use and CDR3 Features in Young and Elderly TCRB Rearrangements. In initially evaluating TCRB repertoires in 5x106 young and elderly subjects we compared the composition of TCRB gene rearrangements at the level of TCRBV and TCRBJ Distinct TCR 20 - 35 70 - 85 20 - 35 70 - 85 gene segment use and the features of the CDR3-encoding Age (years) Age (years) junctional nucleotides. Apheresis lymphocytes were obtained C CD4 memory T cells CD8 memory T cells from four 20- to 35- and five 70- to 85-y-old healthy adults who p=0.06 p=0.11 were regular platelet donors. Naïve CD4 and CD8 T cells were purified by cell sorting. We used a stringent definition of naïve 2x106 cells (CD3+CD4+ or CD8+CCR7+CD45RAhighCD28+) and very restrictive gating to ensure purity. Approximately 1.5–3 million sequence reads were obtained for each T-cell subset of each 1x106 donor (Table S1). TCRBV and TCRBJ gene segments were used at comparable frequencies in the repertoires of naïve CD4 and 0.0 CD8 T cells in young and old individuals (Fig. S1A), whereas the sequences TCRB Distinct 20 - 35 70 - 85 20 - 35 70 - 85 memory repertoires of CD4 and, particularly, CD8 T cells show variable and individual-specific gene segment use frequencies 6 p=0.06 p=0.10 (Fig. S1B). CDR3 sequences in the young and elderly were D 2x10 comparable in length and did not show any definitive age-related features (Fig. S1 C and D). 6
sequences 1x10 High Richness of Naïve CD4 and CD8 TCRB Repertoires in Young and β Elderly Adults. To obtain sequence data to estimate global TCRB repertoire richness we used the experimental design of analyzing 0.0 aseriesofreplicatelibrariesfromindependentcellaliquotsfrom 20 - 35 70 - 85 20 - 35 70 - 85 each T-cell subset in each individual. These replicates allowed us to Distinct TCR Age (years) Age (years) calculate repertoire richness by applying the “Chao2” estimator, anonparametricestimatorofunseenspecies(14).Theapproach Fig. 1. Age is associated with a modest decrease in diversity of the TCRB reper- allows estimation of the extent to which the full repertoire is cov- toire. TCRB sequences were obtained from replicate samples of naïve (A and B)and ered and use of this information to determine a lower bound of the memory (C and D)CD4andCD8Tcells.AlowerboundofTCRBrichnesswasesti- total number of species in the repertoire. Because the Chao2 esti- mated by applying nonparametric statistics using the Chao2 estimator. Results are mator requires only a binary characterization of presence or ab- shown for nucleotide (A and C)andderivedaminoacidsequences(B and D). – – sence of each clone in each replicate library, it circumvents the Estimates were compared by Wilcoxon Mann Whitney test. Increase in age is as- sociated with a decline in richness of naïve CD4 and CD8 T cells; however, the challenges that arise in experimental designs using only a single repertoire in the elderly remains highly diverse. Richness in CD4 and CD8 memory library and avoids confusing the effects of PCR amplification with Tcellsmarkedlydiffered,whereastheimpactofagewasnegligiblysmall. the presence of expanded T-cell clones. To not count possible se- quencing errors as independent sequences, we rejected single se- quences as erroneous if a highly similar clone of greater frequency (15). Second, we estimated the confidence intervals using the ap- was identified in the same library (see Materials and Methods for the proach originally developed by Chao (16). The 95% confidence definition of similarity). intervals with both methods were very narrow (Table S2). The lower bounds on TCRB gene richness obtained with this Naïve TCRB repertoire richness declined significantly in the approach yielded higher estimates than previous studies (Fig. 1). 70- to 85-y-old adults to a lower bound richness of 8–57 million Young adults carried an estimated 60–120 million different different nucleotide sequences encoding ∼5–15 million TCR TCRB genes, both in the CD4 and CD8 naïve T-cell repertoires. β-chain amino acid sequences (P = 0.008, Fig. 1 A and B). In- This high diversity in nucleotide sequences was reflected in a terestingly, the estimates in elderly CD4 and CD8 naïve T cells large functional repertoire of TCR β chains with a lower boundary were similar despite the greater decline in CD8 compared with of ∼20 million different amino acid sequences. To determine the CD4 naïve T-cell numbers with aging (17, 18). robustness of our estimates, we used two approaches to estimate confidence intervals. We applied the BCa variant of bootstrapping CD4 and CD8 Memory T Cells Differ in TCRB Richness Independent of that is designed for obtaining confidence intervals when the un- Age. Memory cells have been selected from the naïve repertoire derlying bootstrap distributionisnotsymmetricaboutitscenter and clonally expanded in response to antigen and are therefore
13140 | www.pnas.org/cgi/doi/10.1073/pnas.1409155111 Qi et al. Immunity CD4+ Memory-Phenotype T Cells in Unexposed Adults
No evidence for clonal deletion?
A Figure 2. Frequency Analysis of the Human 100 + self antigens Antigen-Specific CD4 Lymphocytes
CD4) All blood samples were obtained from individuals
6 seronegative for HIV and CMV exposures. HSV 10 exposure is as indicated (naı¨ve versus exposed). (A) Frequencies of tetramer-tagged cells per million CD4+ T cells. Each symbol represents an 1 antigen-specific population from one individual, and the bar indicates the mean of experiments cells (per 10
+ performed independently with blood obtained at 0.1 different times. Number of Ag specific cellsspecific Ag of Number Tet (B) Comparison between T cells recognizing Antigen HIV CMV Fib FLU-HA FLU-PAtetanus gp100PPins a novel foreign antigen (pre: HIV-1, CMV, and Specificity: HSV-naive FLU-PB1 HSV-exposed HSV-naı¨ve), exposed foreign antigen (post: Flu, tetanus, and HSV-exposed), and self-antigen (self: BCFrequencies of tetramer-tagged cells per million CD4 T cells: 1-10D per 106 Fib, gp100, and PPins). The figure summarizes Each symbol representsns an antigen-specific populationns from one individualp .= 0.0192 data from all 26 individuals. <0.0001Fib (fibrogen), <0.0001 gp100, PPins (preproinsulin):16 self antigens. 16 (C) Precursor frequencies of self-specific lym- CD4) CD4) CD4) 8 90 8 6 phocytes that recognized gp100, Fib, or PPins in 6 6 60 4 Su et al. Immunity 20134 people ages <40 years (n = 7), 40–60 years (n = 13), 20 2 2 and >60 years (n = 6). 15 1 1 (D) Precursor frequency of self-specific lympho- 10 0.5 0.5 cytes that recognized gp100, Fib, or PPins in cells (per 10
cells (per 10 0.25 males (n = 17) and females (n = 9) (see also Fig-
cells (per 10 5 0.25 + +
+ ure S2). 0 0.125 0.125 Tet Tet <40 40-60 >60 MF
Tet pre post self Antigen Exposure Age Gender
The Unprimed CD4+ T Cell Repertoire Contains epitopes (HA and PB1), highlight the heterogeneity of a T cell Pre-existing Memory-Phenotype T Cells response in humans. Another notable feature in these data is To broadly survey antigen-specific CD4+ T cells, we selected the absence of a statistically significant difference between the self-peptides from gp100, fibrinogen (Fib), or preproinsulin frequencies of cells that recognize self-antigens (PPins, Fib, or (PPins) for their relevance in antitumor responses and autoim- gp100) and those that bind to a naı¨ve foreign antigen from munity (Nielen et al., 2005; Yuan et al., 2009; Zhang et al., HIV-1, CMV, or HSV in uninfected individuals (Figure 2B). These 2008). Foreign antigens from gag p24 from HIV-1, pp65 from self-antigen-specific T cells were examined for potential age- or CMV, and VP16 from HSV were selected because highly sensi- gender-dependent change, particularly because the risk of tive serologic tests are available for clearly distinguishing autoimmunity is generally higher in the elderly and in females. exposed individuals from unexposed individuals. The sensitivity Although we did not detect significantly more gp100-, Fib-, or of the HSV, HIV, and CMV serologic tests are 97%–100%, and PPins-specific T cells in older individuals (Figure 2C), we did with respect to HIV, the risk of a false negative is estimated to note a gender-specific difference—there were 1.7-fold more be less than 1 in 2.6 million blood donations nationwide (Lack- gp100-, PPins-, and Fib-specific lymphocytes in females than ritz et al., 1995). We also examined several epitopes from the in males (Figure 2D). influenza virus and tetanus toxin, to which most people have The absence of an obvious frequency difference between been exposed. Specific epitopes from each protein were self- and foreign-antigen-specific populations suggests that chosen from the literature on the basis of T cell stimulation negative selection is incomplete for the particular self-specific- and evidence that they are the product of natural processing ities examined. Alternatively, postthymic antigen stimulation and presentation (Alexander et al., 2010; Arif et al., 2004; Auger might have resulted in expansion of certain populations that et al., 2005; Bronke et al., 2005; Diethelm-Okita et al., 1997; masked a difference in their initial frequencies. To assess the Lamb et al., 1982; Norris et al., 2004; Novak et al., 2001; latter, we analyzed the memory versus naı¨ve phenotype of these Phan et al., 2003). tetramer-positive cells by using the classical memory marker We made tetramers by using peptide exchange to examine CD45RO (Appay et al., 2008). Unexpectedly, we found an abun- antigen-specific CD4+ T cells from 26 HLA-DR4+ healthy blood dance of HIV-1-, CMV-, and HSV-specific memory-phenotype donors (Figure S2). We found that the frequency of precursor T cells in individuals who were seronegative for those viruses cells recognizing a self-antigen or an unexposed viral epitope (Figure 3A). The fraction of these memory-phenotype T cells generally ranged from one to ten cells per million CD4+ T cells ranged from 12% to 93% between different individuals, and (Figure 2A). Notably, we found that previous exposure did not the average exceeded 50% (Figures 3B and 3C). These necessarily result in a detectable expansion of cells specific to memory-phenotype T cells positively correlate with T cell that antigen, as shown by the similarity in the frequencies of frequency (Figure 3D) and have similar frequencies of central HSV-specific T cells between seropositive individuals and those memory-phenotype cells by chemokine receptor CCR7 ex- who are HSV negative. These data, as well as large interindi- pression as influenza-specific T cells (Appay et al., 2008) vidual differences in T cells recognizing the two dominant flu (Figure S3).
Immunity 38, 373–383, February 21, 2013 ª2013 Elsevier Inc. 375 Michael B.A. Oldstone papers
Ohashi et al Cell 1991 and Oldstone et al Cell 1991
Induction of diabetes by viral infection of transgenic mice expressing LCMV protein in pancreas o mice are normal, no evidence for tolerance o infect mice with LCMV: virus is rejected
→ but later mice develop diabetes Polly Matzinger papers
Anderson et al J Immunol 2001
Graft male tissue onto female mouse lacking an immune system o graft is rejected when immune system reconstitutes o graft is accepted when female receives a male thymus
→ how to interpret this? Testing Time-, Ignorance-, and Danger-Based Models of Tolerance
Colin C. Anderson,1* Joseph M. Carroll,† Stefania Gallucci,* John P. Ridge,* Allen W. Cheever,‡ and Polly Matzinger*
In this study, we present data showing that tolerance to Ags in the periphery is not determined by the time at which the Ag appears, or by special properties of tissues in newborn mice or newly developing immune systems. We placed male grafts onto immunoincompetent female mice, allowed the grafts to heal for up to 5 mo, and then repopulated the recipients with fetal liver stem cells. We found that the newly arising T cells were neither tolerant nor ignorant of the grafts, but promptly rejected them, though they did not reject female grafts, nor show any signs of autoimmunity. We also found that the H-Y Ag was continuously cross-presented on host APCs, that this presentation was immunogenic, not tolerogenic, and that it depended on the continuous presence of the graft. In searching for the stimulus that might activate the host APCs, we analyzed mRNA expression with a highly sensitive real-time quantitative PCR assay. By using two different “housekeeping” molecules for comparison, we analyzed the message levels for several stress and/or inflammatory molecules in the healed grafts. We found that the long-healed grafts were not equivalent to “normal” skin because the healed grafts expressed lower levels of GAPDH. Altogether, these data suggest that acceptance vs rejection of peripheral tissues is not attributable to ignorance, timing-based tolerance, or special circulation prop- erties of naive T cells in neonatal tissues. It is more likely attributable to an aspect of the context of Ag presentation that remains to be identified. The Journal of Immunology, 2001, 166: 3663–3671.
© 2001 Nature Publishing Group http://medicine.nature.com his report addresses two currently held ideas. The first is riod.AR Therefore,TICLES tests for a tolerogenic period must be done earlier, the long-standing concept that there is an early period in at the inception of immune competence. T ontogeny during which the developing immune system is A numberIm ofm experimentsunity or to fittingleranc thesee: Op testpos criteriaite outc haveome showns of particularly susceptible to tolerance induction, such that it becomes that MHC or multiplemi minorcroch histocompatability-mismatchedimerism from skin grafts grafts tolerant of Ags that are present at that time. The second is the given before the development of immunocompetence can be re- newer conclusion that establishment of tolerance to skin cannot jected by a newly developingCOLIN C immune. ANDERSON & systemPOLLY MATZI (10–15).NGER However, occur in adult mice because T cell trafficking into the skin occurs it could be reasonablyGhost Lab, Sectio arguedn on T-cell Toler thatance andthese Memory, La experimentsboratory of Cellular and Mo arelecular I notmmunol aogy valid, NIAID/NIH Building 4 Room 111, 9000 Rockville Pike, Bethesda, Maryland 20892-0420, USA only during the neonatal period (1). We have tested both of these test of time-basedCorresp models.ondence should be Aaddres numbersed to C.C.A.; e-m ofail: can models,[email protected] includingov concepts and our data fit with neither. Bretscher’s, are based on the hypothesis that antigenic exposure Although evidence that newly developing immune systems are Solid organ transplants contain small numbers of leukocytes that can migrate into the host and early in ontogenyestablish lon isg-las tolerogenicting microchimerism. becauseAlthough such ofmicro thechimer lowism is of frequencyten associated with of tolerance-prone dates back to the experiments in newborn mice graft acceptance and tolerance, it has been difficult to demonstrate a true causal link. Using skin effector Thfr cellsom mutan int mic thee defici peripheryent for leukocyte s atubset thiss, we fou timend that d (9,onor T 16–18).-cell chimerism Theis a from Medawar and colleagues (2), recent evidence revealed that ‘double-edged sword’ that can result in very different outcomes depending on the host’s im- number of Agsmunol inogic MHCal maturity orand multiplethe antigenic dis minorparities inv mismatchedolved. In immunologica graftslly mature couldhosts, these experiments may have been misinterpreted. Newborn mice chimerism resulted in immunity and stronger graft rejection. In immature hosts, it resulted in be greater thantoleran thece to t numberhe chimeric T c ofells, tissue-specificbut not to graft antigens n selfot expr Agsessed by to the whichchimeric cell thes. are perfectly able to respond if given Ags in appropriate doses, Clinical efforts aimed at augmenting chimerism to induce tolerance must take into account the immune systemmaturati muston state of normally host T cells, the ty establishpe of chimerism p peripheralroduced by each or tolerancegan and the antige (19,nic with appropriate adjuvants, or on appropriate APCs (3–6). Fur- 20) and consequently,disparities involved,the lest th frequencye result be increased of reje Thction r cellsather tha againstn tolerance. these Ags thermore, adult mice can be rendered tolerant if given large doses would‘Passenge ber’ leu highkocytes, m enoughigrating from thatsolid org anyan tran tolerance-pronesplants, which the H-Y ant periodigen is on b wouldoth donor s bekin and T cells), of Ag-bearing cells (4) or smaller doses of cells from which pro- often enter the host circulation. In 1957, it was suggested that chimerism-induced tolerance resulted in graft acceptance. verythese short. migrants, Therefore,rather than the tis Bretschersues themselves, m suggestedight be Howeve anr, in experimentcases in which the sk inin d whichiffered from the host by fessional APCs have been removed (7, 57, 58). One interpretation the primary immunogenic cells1. It is now accepted that graft- multiple minor or major histocompatibility antigens, the toler- femalederived a recipientsntigen-presenting wouldcells (APCs) be can givenactivate ho APCsst T ance presenting was only specific fo ther the ch singleimeric cells mi-, and the skin grafts from these results is that newborn and adult immune systems are cells, and that depletion of donor APCs leads to enhanced ac- were rejected. These data establish a direct causal link between 2,3 nor© 2001 Nature Publishing Group http://medicine.nature.com cep Agtance o H-Yf solid org beforeans . More r theecently recipient, there has also b haseen c developedhimerism and immu immunocompe-ne responses to graft antigens. They affect both tolerizable and immunizable and that the decision to respond intriguing albeit indirect evidence that ‘passenger’ leukocytes future protocols for tolerance induction in transplantation, and tencemay a (9);lso indu Brestcher’sce tolerance4. For exa predictionmple, long-term o beingrgan re- i thatndicate sucha mecha earlynism by w presentationhich viruses (such as human im- or become tolerant must therefore be governed by features other cipients often have low levels of circulating donor hematopoi- munodeficiency virus) and other antigens produced in T cells ofet aic singlecells5–8, and minor-histocompatabilitypartial depletion of these donor cells by Agmay co wouldntrol immun induceity. tolerance than age or timing of Ag expression (8). However, another inter- ‘parking’, irradiation or antibody treatment is associated with ratherincrease thand rejection the9–11. Th immunityese findings led to r thatecent su wasggestions seenPassen whenger T cells e neonatalstablish chimerism females in immunodeficient hosts pretation, recently elaborated by Bretscher (9), is that the newborn that transplantation protocols should seek to enhance donor To determine whether ‘passenger’ leukocytes from skin can mi- werechime challengedrism to improve graf witht acceptan H-Yce, and s expressingtudies in mice gra APCste into the (4).host and Therefore, establish chimerism we, we grafted skin is too old and has already passed through the tolerance-prone pe- have shown some success. If such protocols are to be used in from male B6.Ly5a mice onto T cell-deficient B6 nude female graftedhumans, h maleowever, it andis impor femaletant to unders skintand the tocondi immunodeficienttions mice, which express Ly5 femaleb, a marker tha recipi-t enables the discrimina- in which donor chimerism leads to acceptance in contrast to tion of donor and host cells. Then, 11 months later (ample time ents,those allowedin which it lead thes to rap graftsid rejection to. heal for severalfor chime months,rism to develop then), we mea reconsti-sured chimerism in the blood To elucidate the ‘rules of the game’, we studied donor (Fig. 1a and Table 1). Unexpectedly, there were no donor-de- *Ghost Lab, Section on T cell Tolerance and Memory, Laboratory of Cellular and tutedchime therism fr miceom skin g withrafts in m fetalice. We c liverhose skin cells because orac- ariv fetaled B cell thymuss and seldom a andny APC followeds; the chimeric cells were al- ceptance is difficult to achieve with skin, and it has thus be- most entirely T-cell receptor (TCR)αβ+, CD4 or CD8 T cells Molecular Immunology, National Institute of Allergy and Infectious Diseases/Na- thecom fatee the of‘gold the stand grafts.ard’ of experimental transplantation. (called ‘chimeric T cells’ here) (Fig. 1b). The lack of donor-de- tional Institutes of Health, Bethesda, MD 20892; †Genetics Institute, Andover, MA Because irradiation, antibodies and drugs can have complicat- rived B cells was not simply due to competition with resident B ‡ iDifferentng effects on the models transplantedpredict organ, we to differentok a genetic ap outcomes- cells in the nud ofe hos thists, as rec experiment.ombination activation gene knock- 01810; and Biomedical Research Institute, Rockville, MD 20852 proach, using donors of skin grafts deficient in T cells or T and B out recipient mice (lacking both T and B cells; called RAG-KO Somecells. G time-basedraft-derived chimerism models, led to either inimmu whichnity or tole tolerancer- here) gave nea isrly i baseddentical res onults (d theata not low shown). Thus, even Received for publication May 15, 2000. Accepted for publication January 4, 2001. ance depending on the immunological maturity of the host. If in completely lymphocyte-deficient hosts, chimerism derived frequencythe immune sy ofstem helper was mature T, don cellsor cell ch earlyimerism i inndu ontogeny,ced from a skin g predictraft seemed to that be mai thenly the male result of migrating T- The costs of publication of this article were defrayed in part by the payment of page strong cytotoxic T-lymphocyte (CTL) responses and increased cell populations. charges. This article must therefore be hereby marked advertisement in accordance skin,the fr becauseequency of rejec ittion preexists,. However, when shouldthe immune besystem toleratedAll mice w byith he thealthy, newlylong-standin arisingg grafts showed some de- was allowed to regenerate from an immature state, the chimeric gree of chimerism, which increased with time. Their TCR reper- with 18 U.S.C. Section 1734 solely to indicate this fact. Tc cellsells migra (9,ted to t 16–18).he host thymus, Another where they indu recentced deletion time-basedtoires varied random model,ly and were u innexp whichectedly oligoclonal, with of CTLs. In cases in which the skin graft shared antigens with many mice lacking several Vβ families (Fig. 1c). For example, 1 Address correspondence and reprint requests to Dr. Colin C. Anderson, Ghost Lab, tolerancethe chimeric tocells skin (such a iss ma basedle grafts on onto fe ama particularle mice, in one propertymouse (Fig.1c, ci ofrcles neonatal) had no CD4 ce skinlls expressing Vβ2, 4, Laboratory of Cellular and Molecular Immunology, National Institute of Allergy and 80 NATURE MEDICINE • V OLU ME 7 • N U MBER 1 • JA N UARY 2001 Infectious Diseases/National Institutes of Health, Building 4, Room 111, 9000 Rock- that allows traffic of naive T cells, would predict that both the male ville Pike, Bethesda, MD 20892-0420. E-mail address: [email protected] graft and the normal adult female skin graft should be rejected (1).
Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00 The Journal of Immunology 3665
Results reported that nude mice have a small number of mature T cells, Newly generated T cells reject healed male skin grafts which may have contributed to the rejection process. Therefore, we reperformed the grafting experiments in B10-RAG recipi- In our first series of experiments, we asked whether a healed graft ents by reconstituting with male or female fetal liver cells. In carrying a single, weak, minor histocompatability Ag, would in- addition, to test whether the maturity of the thymus itself has an duce tolerance, be ignored, or induce immunity in T cell popula- influence, we thymectomized a cohort of female B10-RAG tions that were new thymic emigrants. We assessed the response of mice and reconstituted them with a day-15 fetal B6 thymus plus newly generated T cells by giving female B6 nude mice a nude an injection of day-15 fetal liver cells (Fig. 1, C and D). Once male skin graft and 9 wk later, a female thymus. These thymuses again, in both groups, we found that the female mice reconsti- were from normal B6 donors (Fig. 1A), or from B10-RAG donors tuted with female fetal tissues rejected the long-standing male (Fig. 1B) that have no functional T cells and must be repopulated grafts. Thus, T cells generated from fetal precursors, in a fetal by stem cells from the nude host before any T cell development thymus, in hosts that (unlike nudes) have no preexisting T cells, can occur. The expectation was that newly maturing T cells from went on to reject a healed preexisting male skin graft, a sur- the female thymus should slowly seed the periphery in low num- prising result that was not predicted by any current model of bers, encounter the H-Y Ag either in the male skin graft or in its immunological tolerance. The mice did not reject grafts of fe- draining lymph node (25, 30), and become tolerant. As a control to male skin, showing that some aspects of peripheral tolerance ensure that central thymic tolerance was functioning properly in were intact, and those given male thymuses did not reject male these animals, we also grafted one cohort of nude females with a skin, showing that central tolerance was also intact. Rejection male B10-RAG thymus. of preexisting grafts was not peculiar to B6 or B10 mice, as Fig. 1, A and B, shows that newly generated T cell populations C.B-17 SCID mice reconstituted with day-12 BALB/c fetal rejected the long-standing male grafts, though the grafts had healed liver cells also rejected a preexisting B10.D2 minor-histocom- for 9 wk (or 14 wk, data not shown) and appeared as healthy as the patibility different skin graft (not shown and Ref. 59). host’s own skin. The rejection did not require the transfer of ma- ture T cells with the thymus graft because the B10-RAG thymus grafts, like the normal B6 thymuses, lead to rejection. Recipients Lack of autoimmune responses in reconstituted mice reconstituted with a male thymus did not reject the male skin (Fig. The finding that the long-standing male grafts were rejected by 1B) showing that central tolerance mechanisms were functionally newly generated recent thymic emigrants counter the expectations intact and able to induce tolerance in the newly developing T cells, of most time-based models because these models hold that a pre- but peripheral tolerance to the grafts did not occur. Thus it ap- existing Ag should induce tolerance. They counter the ignorance peared that mature naive T cells and, more surprisingly, newly model because the graft Ags are peripheral Ags that should be developing T cells, were immunized rather than tolerized by the ignored. They counter the Danger model because the grafts had peripheral Ag on the healed grafts, despite the low frequency of healed. Healthy tissues should not immunize. responding cells to H-Y and the lack of cross-reactive environ- However, one recently elaborated time-based model did offer a mental Ags to drive the response (7, 24). possible explanation for the rejection. This model suggests that In the experiments above, the bone-marrow-derived precur- tolerance to peripheral Ags cannot be established in adult mice. It Testingsors were from adult Time-, marrow, and Ignorance-, the thymuses in which they andwas proposed Danger-Based that establishment of tolerance Models to skin, for example, of developed were from neonatal animals. It remained possible requires that naive T cells traffic into the skin, which occurs only Tolerancethat peripheral tolerance depends on some property intrinsic to in the neonatal period (1). If this were the case, we would expect fetal but not adult precursor cells or thymuses. It has also been to find general signs of peripheral autoimmunity, as has sometimes
Colin C. Anderson,1* Joseph M. Carroll,† Stefania Gallucci,* John P. Ridge,* Allen W. Cheever,‡ and Polly Matzinger*
In this study, we present data showing that tolerance to Ags in the periphery is not determined by the time at which the Ag appears, or by special properties of tissues in newborn mice or newly developing immune systems. We placed male grafts onto immunoincompetent female mice, allowed the grafts to heal for up to 5 mo, and then repopulated the recipients with fetal liver stem cells. We found that the newly arising T cells were neither tolerant nor ignorant of the grafts, but promptly rejected them, though they did not reject female grafts, nor show any signs of autoimmunity. We also found that the H-Y Ag was continuously cross-presented on host APCs, that this presentation was immunogenic, not tolerogenic, and that it depended on the continuous presence of the graft. In searching for the stimulus that might activate the host APCs, we analyzed mRNA expression with a highly sensitive real-time quantitative PCR assay. By using two different “housekeeping” molecules for comparison, we analyzed the message levels for several stress and/or inflammatory molecules in the healed grafts. We found that the long-healed grafts were not equivalent to “normal” skin because the healed grafts expressed lower levels of GAPDH. Altogether, these data suggest that acceptance vs rejection of peripheral tissues is not attributable to ignorance, timing-based tolerance, or special circulation prop- erties of naive T cells in neonatal tissues. It is more likely attributable to an aspect of the context of Ag presentation that remains toFIGURE be identified. 1. Well-healedThe skin Journal grafts induce of Immunology, immunity. A, B62001, nude female 166: mice 3663–3671. (n ϭ 6) were given a B6 nude male skin graft and 9 wk later were given a neonatal B6 female thymus graft. B, B6 nude female recipients received a B10-RAG male skin graft and, 9 wk later, a neonatal male (dashed line; n ϭ 4) or femalehis (solid report line; addressesn ϭ 4) B10-RAG two currently thymus. C,held B10-RAG ideas. female The recipients first is (n ϭ 4)riod. received Therefore, two skin tests grafts, for a B10-RAG a tolerogenic female period(dashed line) must and be done earlier, male skinthe graft long-standing (solid line). Eleven concept weeks later, that therethey were is an reconstituted early period with day-15 in B6at female the inception fetal liver of cells. immuneD, Thymectomized competence. B10-RAG female recipients receivedontogeny a B10-RAG during whichmale skin the graft developing and 7 wk later immune were reconstituted system is with day-15A B6 number female fetalof experiments liver cells and thymusfitting graft these (solid test line; criterian ϭ have shown T6). Controls (dashed line) were similarly treated except they received a male fetal thymus graft (n ϭ 1) or a female skin graft (n ϭ 2). particularly susceptible to tolerance induction, such that it becomes that MHC or multiple minor histocompatability-mismatched grafts tolerant of Ags that are present at that time. The second is the given before the development of immunocompetence can be re- newer conclusion that establishment of tolerance to skin cannot jected by a newly developing immune system (10–15). However, occur in adult mice because T cell trafficking into the skin occurs it could be reasonably argued that these experiments are not a valid only during the neonatal period (1). We have tested both of these test of time-based models. A number of models, including concepts and our data fit with neither. Bretscher’s, are based on the hypothesis that antigenic exposure Although evidence that newly developing immune systems are early in ontogeny is tolerogenic because of the low frequency of tolerance-prone dates back to the experiments in newborn mice effector Th cells in the periphery at this time (9, 16–18). The from Medawar and colleagues (2), recent evidence revealed that number of Ags in MHC or multiple minor mismatched grafts could these experiments may have been misinterpreted. Newborn mice be greater than the number of tissue-specific self Ags to which the are perfectly able to respond if given Ags in appropriate doses, immune system must normally establish peripheral tolerance (19, with appropriate adjuvants, or on appropriate APCs (3–6). Fur- 20) and consequently, the frequency of Th cells against these Ags thermore, adult mice can be rendered tolerant if given large doses would be high enough that any tolerance-prone period would be of Ag-bearing cells (4) or smaller doses of cells from which pro- very short. Therefore, Bretscher suggested an experiment in which fessional APCs have been removed (7, 57, 58). One interpretation female recipients would be given APCs presenting the single mi- from these results is that newborn and adult immune systems are nor Ag H-Y before the recipient has developed immunocompe- both tolerizable and immunizable and that the decision to respond tence (9); Brestcher’s prediction being that such early presentation or become tolerant must therefore be governed by features other of a single minor-histocompatability Ag would induce tolerance than age or timing of Ag expression (8). However, another inter- rather than the immunity that was seen when neonatal females pretation, recently elaborated by Bretscher (9), is that the newborn were challenged with H-Y expressing APCs (4). Therefore, we is too old and has already passed through the tolerance-prone pe- grafted male and female skin to immunodeficient female recipi- ents, allowed the grafts to heal for several months, then reconsti- *Ghost Lab, Section on T cell Tolerance and Memory, Laboratory of Cellular and tuted the mice with fetal liver cells or a fetal thymus and followed Molecular Immunology, National Institute of Allergy and Infectious Diseases/Na- the fate of the grafts. tional Institutes of Health, Bethesda, MD 20892; †Genetics Institute, Andover, MA 01810; and ‡Biomedical Research Institute, Rockville, MD 20852 Different models predict different outcomes of this experiment. Received for publication May 15, 2000. Accepted for publication January 4, 2001. Some time-based models, in which tolerance is based on the low frequency of helper T cells early in ontogeny, predict that the male The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance skin, because it preexists, should be tolerated by the newly arising with 18 U.S.C. Section 1734 solely to indicate this fact. T cells (9, 16–18). Another recent time-based model, in which 1 Address correspondence and reprint requests to Dr. Colin C. Anderson, Ghost Lab, tolerance to skin is based on a particular property of neonatal skin Laboratory of Cellular and Molecular Immunology, National Institute of Allergy and Infectious Diseases/National Institutes of Health, Building 4, Room 111, 9000 Rock- that allows traffic of naive T cells, would predict that both the male ville Pike, Bethesda, MD 20892-0420. E-mail address: [email protected] graft and the normal adult female skin graft should be rejected (1).
Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00 ARTICLES
A need therefore exists for a systematic study of tolerance within designated eGFPp), which was incorporated into an I-Ab tetramer used normal polyclonal CD4+ T cell repertoires specific for self peptides for the detection of CD4+ T cells with TCRs specific for this epitope16 with defined patterns of expression. Here we carried out such a study (Fig. 1a,b). This experimental system allowed us to determine the effects using cellular enrichment based on peptide–MHC class II tetramers. of different patterns of self-antigen expression on the same epitope- Intrathymic deletion was the mechanism used for T cells specific for specific T cell repertoire. Another advantage of this approach was that peptides from proteins expressed uniformly by thymic dendritic cells wild-type C57BL/6 (B6) mice, which express the I-Ab molecule but not and/or medullary thymic epithelial cells, whereas ‘ignorance’ was the eGFP, eYFP or CFP, could be used as controls with which to study the
mechanism- used for tolerance to peptides from cytosolic or nuclear relevant T cell repertoire in the absence of the self peptide. 187 , 3 proteins expressed only outside the thymus. Partial intrathymic We assessed tolerance in the eGFPp–I-Ab–specific CD4+ T cell .These 5 . 1 , 9 ,
deletion and impaired effector5 T cell potential but enhanced Treg cell repertoire in 14 different strains with transgenic expression of eGFP, , 1
potential among remaining 4 clones was the mechanism of tolerance for eYFP or CFP (Supplementary Table 1) by immunizing the mice with & . Other tissue- T cells was cells T 8 2 + , many peptides from tissue-restricted proteins that were also expressed eGFPp in complete Freund’s adjuvant (CFA). We obtained single- 7 ARTICLES
ARTICLES + of Department sparsely in the thymus. Thus, CD4 T cells2 tolerate self peptides with cell suspensions of secondary lymphoid organs from the immunized . This issue has come
4 b different expression1 patterns via different mechanisms. mice and stained these with the eGFPp–I-A tetramer labeled with T cells T 12– allophycocyanin, then with magnetic beads coated with antibody Diabetes Center, University of University Center, Diabetes
+ b 3 , Whitney E Purtha E Whitney , A need therefore exists for a systematic study of tolerance within designated eGFPp), which was incorporated into an I-A tetramer used 2 RESULTS to allophycocyanin. The suspensions underwent enrichment for + + 16
normal polyclonal CD4 T cell repertoiresrevealmicecansuchAnalysis of. specific for self peptides for+ the detection of CD4 T cells with TCRs specific for this epitope
A0 transgenic model of CD4 T cell tolerance to self tetramer-bound cells on magnetized columns, then were eluted and 1 , with defined patterns of expression. Here we8 carried out such a study (Fig. 1a,b). This experimental system allowed us to determine the effects , or may becomemay orlackanergic a to due 7 2 We used enhanced green fluorescent protein (eGFP), enhanced yellow flu- stained with antibodies designed to identify T cells. We assessed the using cellular enrichment based on peptide–MHC class II tetramers. of different patterns of self-antigen expression on the same epitope- b + , Kristin A Hogquist A Kristin , orescent protein (eYFP), or cyan fluorescent protein (CFP) as model self differentiation of eGFPp–I-A –specific CD4 T cells into the TH1, 3 Intrathymic deletion was the mechanism cells used for T cells specific for specific TARTICLES cell repertoire. Another advantage of this approach was that antigens in mouse strains with transgenic expression of these proteins and T 17, follicular helper T cell (T cell) and T cell subsets by intra- peptides from proteins expressedreg uniformly by thymic dendritic cells wild-type C57BL/6 (B6) mice, which express the HI-Ab molecule but not FH reg 7 T cells specific for green fluorescentgreen for specificcells T b + the MHC class II molecule I-A . These fluorescent proteins all contain an cellular staining of the lineage-defining transcription factors T-bet,
, You Jeong Lee Jeong You , and/or medullary thymic epithelial cells, whereas ‘ignorance’ was the eGFP, eYFP or CFP, could be used as controls with which to study the 1 I-Ab-binding peptide first identified in eGFP (HDFFKSAMPEGYVQE; RORGt, Bcl-6 and Foxp3, respectively (Fig. 1c).
mechanism used for tolerance T cells that encounter and recognize these epitopes to peptides from cytosolic or nuclear relevant T cell repertoire in the absence of the self peptide. proteins expressed only outside+ the thymus. Partial intrathymic We assessed tolerance in the eGFPp–I-Ab–specific CD4+ T cell
deletion and impaired effector T cell potential but enhanced6. Further, it is clear Tthat an abundance of T cellsreg with the cell repertoire in 14 different strains with transgenic expression of eGFP, 1 a 101 b c potential among remaining clones was the mechanism of tolerance*** for*** eYFP ****** or CFP (Supplementary*** NS Table 1) by immunizing the mice with doi:10.1038/ni.332 5 5 13.1 1.1 5 10 10 10 22.0 , Mark S Anderson S Mark , TH17 Treg many peptides from tissue-restricted proteinsMuch of the evidence cited above has been provided by studies of that were also expressed eGFPp in completeNS Freund’s adjuvant (CFA). We56 obtained single- 5 5 ****** *** + *** *** ** 4 4 T 1 Tolerance is established+ in polyclonal10 CD4 T cells 10 10 104 H mice with transgenic expression of monoclonalspecifictransgenicselfTCRswithmiceexpressionof for epitopes II class peptide–MHC ofcostimulation derived from the innate immune system secondarylymphoidorgans, specific cellsT would beretained thein repertoire in a fully responsive state of ignorance restrictedproteins that aresecreted arepresentor indying cells may reach secondary lymphoid organs, where they would be processed by antigen-presenting cells toCD4 epitopes. form self peptide–MHC class II to the fore in several studies as showingprominent a thatmechanism clonalin polyclonaldeletion T cell is repertoires,beennot reported asfor certain monoclonal has T cell populations tolerance mechanisms because of the large numbers of T cells avail Because limitations. several has approach this However, study. for able mousewithtransgenica expressionTCRsinglecontains froma TCR a larger repertoire, it can reveal only those thatapply to TCRs tolerancewith that affinity for a self mechanismspeptide–MHC class II epitope same TCR can create abnormal outcomes may differentiate into T into differentiate may studies, however, have been limited patternsofexpression. to epitopes with ubiquitous sparsely in the thymus. Thus, CD4 T cells tolerate self peptides with cell suspensions of secondary). lymphoid organs from the immunized 57.0 3 WT 10 WT 3 3 T
T cells b 10 FH
by distinct mechanisms, according to self-peptide 10 + . - - - different expression- patterns via different mechanisms. mice and stained these with the eGFPp–I-A tetramer labeled with
4
1 2 10 10 t 0 0 expression patterns allophycocyanin, then with magnetic0 beads coated with antibody CD4 CD44 85.8 +
RESULTS to allophycocyanin. The suspensions underwent enrichment forROR T-bet 1 1,6 3 1 2 3 3 4 5 3 3 4 5 3 4 5 . Some. T Deepali Malhotra , Jonathan +L Linehan , Thamotharampillai Dileepan Tet 10 , You Jeong Lee , Whitney E Purtha , 0 10 10 10 –10 0 10 10 10 0 10 10 10 6 A transgenic model of CD4 hi Jennifer V Lu3, Ryan W Nelson 1T, Brian cell T Fifetolerance4, Harry T Orrto 5self, Mark S Anderson3, Kristintetramer-bound A Hogquist2 & cells on magnetized columns, then were eluted and , Harry T Orr T Harry , 2– Tet Foxp3 Bcl-6 4 We used enhancedMarc K Jenkins green1 fluorescent protein (eGFP), enhanced yellow flu- stained with antibodies designed to identify T cells. We assessed the 2 [email protected] CD44 T cells have revealed several mechanisms of self-tolerance; however, which self-tolerance;however, of mechanisms severalrevealed have cells T 10 b + + , Thamotharampillai Dileepan Thamotharampillai , orescent Studiesprotein of repertoires (eYFP), of ormouse cyan monoclonal fluorescent CD4+ T cells protein have revealed (CFP) several as mechanisms model self of self-tolerance; differentiation however, which of eGFPp–I-A –specific2016 FEBRUARY CD4 T cells into the TH1, 6 ,
+ Minneapolis,School,Medical Minnesota of University Immunology, for CenterMedicine, of Department e
1 mechanisms operate in normal repertoires is unclear. Here we studied polyclonal CD4 T cells specific for green fluorescent 4 antigens in mouse strains.Low-affinity inter with transgenic expression of these proteins and T 17, follicular helper T cell (T cell) and T cell subsets by intra- 9.8 2.3 protein expressed1 in various organs, which allowed us to determine the effects of 1specific expression patternsH on the same FH reg 33 18.8 10 TH17 Treg ) thymic epithelial cells cells epithelial thymic ) epitope-specific T cells. Peptides T cells in the thymus b presented uniformly by thymic antigen-presenting cells were tolerated by clonal deletion, T 1
+ H the MHC class II molecule I-A+ . These fluorescent proteins all contain an cellular staining of the lineage-defining transcription factors T-bet, whereas peptides excluded from the thymus were ignored. Peptides with limited thymic expressionCFP induced partial clonal WT Cluster 1: Cluster 1: b eGFPeGFPeYFPeGFP eGFPeGFPeGFPG eGFPeYFPeGFPeGFPeGFPeGFP 53.1 I-A -bindingdeletion peptide and impaired first effector identified T cell potential in eGFPbut enhanced (HDFFKSAMPEGYVQE; regulatory T cell potential. These mechanisms ROR weret, alsoBcl-6 active and for Foxp3, respectivelyeGFP (Fig. 1c). eGFP TFH AireCdh1 Ins2 + Ins1 Ins1 Ins1 , Brian T Fife T Brian , T cell populations specific for endogenously expressed self antigens. Thus, the immunotoleranceUBC Itgax ofChat polyclonal CD4 T cells was ACTB Pcp2 Foxp3Foxd1 2 NUMBER 1 Myh11 CD207Zfp423
maintained by distinct mechanisms, according to self-peptide expression patterns. cells in the thymus might not
+ 87.9 Self-reactive CD46 + T cells are generated in the normal course of rear- secondary lymphoid organs, specific T cells would be retained in the 10 1 17 rangementa of the T cell antigen receptor (TCR)1. Low-affinity interNS - repertoirebd in a fullyT reg TresponsiveFH T H T H state of ignorance7,8. cOther tissue- *** *** ****** *** WT 1 1.5 57.5 actions between the TCR and ligands composed of self peptide and restricted proteins that 5are secretedeGFP or are present in dying cells may 5 13.1 1.1 5 11.2 10 10 10 22.0 TH17 Treg major histocompatibility (MHC) class II (self peptide–MHC class II) reach secondary lymphoid organs,Ins1 where they would be processed TH17 Treg 4.5 5 ****** *** *** *** ** NS eGFP56 T 1 TH1 10 + 1 104 Foxd1 4 4 H
are essential for the positive selection of CD4 T cells in the thymus . by antigen-presenting cells to form self peptide–MHC class II 10 VOLUME 17 eGFP Cluster 2: 10 Cluster 2: 39.4 + 57.0 However, high-affinity interactions between the TCR and self epitopes. CD4 T cells 3that encounterIns2 and recognize2 these epitopes WT eYFP WT 3 eGFP eGFP TFH , Jonathan L Linehan L Jonathan , 10 3 T T cells CD2072 10 Ins2 10 FH Ins2
1 peptide–MHC class II can cause autoimmunity. Therefore, tolerance may differentiate into T cells or may become anergic due to a lack + reg 4 cell) if its TCR binds strongly to a self peptide– 2 eGFP mechanisms10 have evolved to limit T cells with TCRs with high affinity of costimulation derived10 fromMyh11 the innate immune system1,9. t eGFP 0 reg 1 0 for self peptide–MHCrear normalthecoursegeneratedof in cells are T class II ligands. Much of the evidence0 citedFoxp3 above has been provided by studies of CD4
, Ryan W Nelson W Ryan , eGFP CD44 + 85.8 41.0
+ Zfp423 3 Different tolerance mechanisms may be engaged depending on mice with transgenic expression of monoclonal TCRs specific for self eGFP ROR T-bet 3 Pcp27,8,10 Minneapolis,School,Medical Minnesota of Neuroscience, University Translational for Institute Pathology, and MedicineLaboratory of Department Bethesda,Health, of InstitutesNationalDiseases, Infectious and Allergy of InstituteNationalDiseases, Parasitic of Laboratoryaddress:Present where a self peptide–MHC class II epitope is displayed. A develop- peptide–MHC class II epitopes . Analysis3 4 of5 such mice can reveal5 6 3 3 4 5 3 4 5
Tet 10 eGFP 0UBC 10 10 10 3 –10 0 10 10 10 0 10 10 10 ing thymocytehi may die by apoptosis, or it may differentiate into a tolerance mechanisms because of theeYFP large numbers of T cells avail- 105 5 7.1 64.3 5 Itgax Foxp3 Bcl-6 10 T 17 T 10 0 regulatory T cell (Treg cell) if its TCR binds strongly to a self peptide– able for study. However, thisTet approacheGFP has several limitations. Because 0.06 H reg Chat 4 4 4 TH1 MHC class II2 epitope displayed on particular antigen-presenting a mouse with transgenic TCR expressionCFP contains a single TCR from 10 10 10 CD44 10 Cdh1 Cluster 3: Cluster 3: 0 cells1,2. Epitopes in this class may be derived from proteins expressed a larger repertoire, it can reveal only those tolerance mechanisms 14 strains expressedderivedproteinsbefrom thismayclass in Epitopes . eGFP 3 3 3 T + 2 Aire eGFP 10 eGFP FH or taken up by MHC class II–positive (MHCII ) thymic epithelial cells that apply, to TCRs with that affinity eGFPfor a self peptide–MHC class II UBC UBC 10 10 e 1 ACTB or dendritic cells. However, MHCII+ cells in the thymus might not epitope11. Further, it is clear that an abundance of T cells with the 9.8 2.3 t 0 1 Different tolerance mechanisms may be engaged depending on 33 T 17 T 18.8 display epitopes10 from all host proteins. Thus, T cells expressing TCRs same TCR can create abnormal outcomes12–14. This issue has come H reg 0 0 Department of Microbiology and Immunology, Center for Immunology, University of Minnesota Medical School, Minneapolis, Minnesota, USA. Minnesota, Minneapolis, School, Medical Minnesota of University Immunology, for Center Immunology, and Microbiology of Department T 1 display epitopes from all host proteins. Thus, T cells expressing TCRsexpressingcells T Thus, proteins. hostall from epitopes display with high affinity for certain self epitopes may not be deleted in the thymusandthusexitmay secondaryto lymphoid organs Minnesota, USA. Minnesota, rangementof the T cell antigen receptor (TCR) mechanisms operate in normal repertoires is unclear. Here we studied polyclonal CD4 polyclonal studied we Here unclear. is repertoires normal in operate mechanisms USA. Minnesota, ( M.K.J. to addressed be should Correspondence USA. Maryland, 2016; January 4 online published 2015; October 21 accepted 2015; August 21 Received Tolerance is established in polyclonal CD4 Malhotra Deepali CD4 monoclonal mouse of repertoires of Studies CD4 Self-reactive 1 IMMUNOLOGYNATURE protein expressed in various organs, which allowed us to determine the effects of specific expression patterns on the same the on patternsexpressionspecific of effects the determine to us allowed which organs,various in expressedprotein deletion,clonal by toleratedwere antigen-presenting cellsthymic by uniformlypresentedPeptides cells. epitope-specificT clonal partialinduced expression thymic limited with Peptidesignored. were thymus the from excludedpeptides whereas for active also weremechanisms Thesepotential. cell T regulatoryenhanced but potentialcell T effectorimpaired and deletion CD4 polyclonal of immunotolerance the Thus, antigens. self expressed endogenously for specific populations cell T MHC class II epitope displayed on cells particular antigen-presenting (MHCII II–positive class MHC by up taken or Laboratory Medicine and Pathology, Center for Immunology, University of Minnesota Medical School, Minneapolis, Minnesota, USA. Minnesota, Minneapolis, School, Medical Minnesota of University Immunology, for Center Pathology, and Medicine Laboratory California San Francisco, San Francisco, California, USA. California, Francisco, San Francisco, San California by distinct mechanisms, according to self-peptideto accordingmechanisms,distinct by patterns expression However, high-affinity interactions peptide–MHCTherefore,classcausecanIIautoimmunity.tolerance between the affinity high with TCRs with cells T TCRlimit to evolved have mechanisms and self forself peptide–MHC class II ligands. where a self peptide–MHC class II epitope is displayed. A develop ing thymocyte may die by apoptosis, or it may differentiateregulatorycellT(T into a or dendritic cells. However, MHCII maintained by distinct mechanisms, according to self-peptide expression patterns. expression self-peptide to according mechanisms, distinct by maintained cells in this class may be specific for epitopes from proteins that are expressedlocationssecretedinpoorarewithdrain lymphatic notor age.Since these proteins would notreach antigen-presenting cells in Jennifer V Lu MarcKJenkins actions between the TCR and ligands composed of self peptide and histocompatibilitymajor (self(MHC)classIIpeptide–MHC classII) essentialpositivetheselectionare forCD4 of H 28.6 with high affinity for certain self epitopes may not be deleted in the to the fore in several studies showing that clonal deletion is not 10–3
CFP WT 0 CD44 ROR T-bet eGFPeGFPeYFPeGFP eGFPeGFPeGFPeGFPeYFPeGFPeGFP2–eGFP6 eGFP Cluster 1: 96 Cluster 1: 53.1 thymus and thus may exit to secondary lymphoid organs . Some T as prominent a mechanism192288 in383 polyclonal479575670 T cell repertoires, as has 3 4 5 T 3 4 5 3 4 5 eGFP 1,8383,0054,1735,3406,5087,6764,5,15eGFP 0 10 10 10 FH 0 10 10 10 0 10 10 10 cells in this classUBC mayItgax beAire specificCdh1Chat for epitopesIns2 from proteinsIns1 that are beenIns1 reported for certain monoclonal T cell populationsIns1 . These ACTB Pcp2 not secreted or are expressed in locationsMyh11 withCD207 poorZfp423Foxp3 lymphaticFoxd1 drain- studies, however, have beenCD44 limitedhi toTet epitopes+ cells with ubiquitous Tet Foxp3 Bcl-6 age. Since these proteins would not reach antigen-presenting cells in patterns of expression. 87.9
1 17 reg FH H H hi + b + d T T T T Figure 1 Three main patterns of tolerance to self antigens. (a) Total CD44 Tet (eGFPp–I-A -tetramer (Tet)-binding) CD4 T cells from pooled 1Department of Microbiology andWT Immunology, Center for Immunology,1 University of Minnesota Medical School, Minneapolis, Minnesota, USA. 2Department of 1.5 57.5 3 11.2 Laboratory Medicine and Pathology, CentereGFP for Immunology, University of Minnesota Medical School, Minneapolis, Minnesota, USA. Diabetes Center, University of TH17 Treg Ins1 4 secondary lymphoid organs of wild-type4.5 (WT) and transgenic mice 13–14 d after immunization with 100 Mg eGFPp emulsified in CFA. Each symbol California San Francisco, San Francisco, eGFPCalifornia, USA. Department of Medicine, Center for Immunology, University of Minnesota Medical School, Minneapolis, TH1 Minnesota, USA. 5Department ofFoxd1 Laboratory Medicine and Pathology, Institute for Translational Neuroscience, University of Minnesota Medical School, Minneapolis, eGFP represents Clusteran individual 2: mouse (colors matchCluster those 2: of clusters identified in d); small horizontal39.4 lines indicate the geometric mean. (b) Flow cytometry Minnesota, USA. 6Present address:Ins2 Laboratory of Parasitic2 Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, eGFP eGFP T Maryland, USA. CorrespondenceCD207 should beeYFP addressed to M.K.J. ([email protected]). Ins2 Ins2 + FH eGFP analyzing CD44 and tetramer (Tet) staining in tetramer-enriched CD4 T cells from pooled secondary lymphoid organs of wild-type (WT) mice primed Received 21 August 2015; acceptedMyh11 21 October 2015; published online 4 January 2016; doi:10.1038/ni.3327 + − − + Foxp3eGFP 14 d earlier with eGFPp in CFA. (c) Expression of RORGt and Foxp3 by Tet cells (left) and of T-bet and Bcl-6 by Foxp3 RORGt Tet cells (right) among eGFP 41.0 Zfp423 hi + + hi + + eGFP the CD44 Tet CD4 T cells identified in b. (d) Unsupervised hierarchical clustering of CD44 Tet CD4 T cells (geometric mean cell number and NATURE IMMUNOLOGY VOLUMEPcp2 17 NUMBER 2 FEBRUARY 2016 187 UBCeGFP 3 eYFP phenotype) in 15 mouse5 strains (right margin), grouped5 into7.1 three64.3 main clusters5 (far left and right). (e) Flow cytometry analyzing CD44 and tetramer Itgax 10 10 T 17 T 10 0 eGFP 0.06+ H reg T 1 Chat staining of tetramer-enriched4 CD4 T cells from pooled4 secondary lymphoid4 organsH of three eGFP-expressing mouse strains in clusters 1–3 identified CFP 10 10 10 Cdh1 Cluster 3: Cluster 3: + hi + 0 eGFP in (d) (left), and transcription3 factor expression as in c by the CD4 CD44 Tet cells identified at left (right). Numbers adjacent to outlined areas Aire eGFP 10 eGFP 3 3 TFH eGFP UBC 10 10 ACTB UBC (b,c,e) indicate percent cells in each. NS, not significantt (P > 0.05); *P < 0.05, **P < 0.005 and ***P < 0.0001, versus wild-type B6 mice (one-way 0 0 0 analysis of variance (ANOVA) of log -transformed values). 28.6Data are pooled from 12 independent experiments with 3–23 mice per group (a,d) or are 10–3 10
0 CD44 ROR T-bet 96 192288383479575670 3 4 5 3 4 5 3 4 5 1,8383,0054,1735,3406,5087,676 representative of 12 experiments0 10 10 (b10,c) or eight experiments0 (10e). 10 10 0 10 10 10 CD44hi Tet+ cells Tet Foxp3 Bcl-6
hi + b + Figure 1 Three main patterns of tolerance to self188 antigens. (a) Total CD44 Tet (eGFPp–I-A -tetramer (Tet)-binding) CD4 T cellsVOLUME from pooled 17 NUMBER 2 FEBRUARY 2016 NATURE IMMUNOLOGY secondary lymphoid organs of wild-type (WT) and transgenic mice 13–14 d after immunization with 100 Mg eGFPp emulsified in CFA. Each symbol represents an individual mouse (colors match those of clusters identified in d); small horizontal lines indicate the geometric mean. (b) Flow cytometry analyzing CD44 and tetramer (Tet) staining in tetramer-enriched CD4+ T cells from pooled secondary lymphoid organs of wild-type (WT) mice primed 14 d earlier with eGFPp in CFA. (c) Expression of RORGt and Foxp3 by Tet+ cells (left) and of T-bet and Bcl-6 by Foxp3−RORGt−Tet+ cells (right) among the CD44hiTet+ CD4+ T cells identified in b. (d) Unsupervised hierarchical clustering of CD44hiTet+ CD4+ T cells (geometric mean cell number and phenotype) in 15 mouse strains (right margin), grouped into three main clusters (far left and right). (e) Flow cytometry analyzing CD44 and tetramer staining of tetramer-enriched CD4+ T cells from pooled secondary lymphoid organs of three eGFP-expressing mouse strains in clusters 1–3 identified in (d) (left), and transcription factor expression as in c by the CD4+CD44hiTet+ cells identified at left (right). Numbers adjacent to outlined areas (b,c,e) indicate percent cells in each. NS, not significant (P > 0.05); *P < 0.05, **P < 0.005 and ***P < 0.0001, versus wild-type B6 mice (one-way analysis of variance (ANOVA) of log10-transformed values). Data are pooled from 12 independent experiments with 3–23 mice per group (a,d) or are representative of 12 experiments (b,c) or eight experiments (e).
188 VOLUME 17 NUMBER 2 FEBRUARY 2016 NATURE IMMUNOLOGY A simple mathematical model A first model (DeBoer.prsl93)
S number of self epitopes evoking tolerance (105) R0 potential repertoire (before tolerance) R “functional”A repertoire first model after(DeBoer.prsl93) tolerance (> 108) p lymphocyte recognition probability (precursor frequency)
S number of self epitopes evoking tolerance (105) Probability of responding to foreign epitope: R0 potential repertoireA simple (before mathematical tolerance) model S number of self epitopes evoking tolerance (10R5) [Burroughs.i04] 8 “functional” repertoireP =1 after(1 tolerancep) 12 ( 10 ) R R0 Thepotential “optimal”repertoirei (before tolerance) specificity (huge: 10 cells)> R “functional” repertoire after tolerance (> 108) [Qi.pnas14] wherep lymphocyte recognitionp recognition probability probability (precursor freq. 10-5) (precursor[Blattman.jem02, Su.i13] frequency) Size of functional repertoire: We had S R = R0(1 p) S Probability of responding toR foreign0(1 p) epitope: Pi =1 (1 p) Probability of responding to a foreign epitope: R Pi =1 (1 p) AccommodatesThis probability two of problems:responding, Pi, recognition has its maximum and at deletion. Taking the derivative of Pi to p gives the optimum where 1 5 13 p = 10 S ⇥ R = R (1 p)S 1993 Soc Royal Proc Perelson & Boer De 0 Thus specificity seems determined by number of self epitopes (DeBoer.prsl93,Whitaker.jtb93,Nemazee.it96,Borghans.ji99). Accommodates two problems: recognition and deletion.
14 13 A simple mathematical model
1 Too Too much Pi specific deletion 0.8 1/S
0.6
0.4
0.2
0 -14 -12 -10 -8 -6 -4 -2 0 SPECIFIC DEGENERATE Cross-reactivity (log(p))
Wide range of cross-reactivities for which functional reper- toire is complete. True optimum at p =1/S.
15 Lymphocytes are specific AND recognize many peptides
Repertoire: 108 clones Peptides: 1013
5 p = 10
5 p = 10
5 13 8 10 10 = 10 peptides per clone ⇥
Because there are many more peptides (2010 ≃ 1013) than clones, we need sufficient cross-reactivity [Mason Imm. Today 98].
But each peptide is recognized by very few clones: p ≃ 10−5 1 1 1
1 1 1
1 0.8 1 0.8 1 0.8
0.8 0.8 0.8
0.8 0.6 0.8 0.6 0.8 0.6 1 1 1 Pi Pi Pi 0.6 0.6 0.6 0.6 0.6 0.6 Pi 0.8 0.4 Pi0.8 0.4 0.8 Pi 0.4 Pi Pi Pi
0.4 0.4 0.4 0.4 0.4 0.4 0.6 0.6 0.6 0.2 0.2 0.2
Pi Pi Pi 0.2 0.2 0.2 0.2 0.2 0.2 0.4 0.4 0.4
0 –10 –8 –6 –4 –2 0 –10 –8 –6 –4 –2 0 –10 –8 –6 –4 –2
0 log[p] 0 log[p] 0 log[p] 0 –10 –8 0.2 –6–10 –4–8 –2 –6 0 –4 –10 –2 –80.2 –6 –10 –4 –8–2 –60 –4–10 –2–80.2 –6 –4–10 –2 –8 –6 –4 –2 log[p] log[p] log[p] log[p] log[p] log[p]
0Figure 1: The probability of mounting0 an immune response Pi from0 Eq. (2) as a function of the specificity p of the –10 –8 –6 –4 –2 5 –10 –8 9–6 –4 –2 –10 –8 –6 –4 –2 Figure 1: The probabilitylymphocytes. of mounting Parameters an immuneS response= 10 Pandi fromR0 Eq.= 10 (2). as Panel a function (b) depicts of the specificity the e ectp ofof decreasing the the initial repertoire size Figure 1: The probabilitylog[p] of mounting an immune responselog[p] Pi from Eq. (2) as a functionlog[p] of the specificity p of the 5 9 95 8 97 lymphocytes. ParametersfromS R=0 10= 10and, RR00== 10 10., Panel to R0 (b)= depicts 10 . Panel the e (c)ect depicts of decreasing the e the ect initial of incomplete repertoire size tolerance induction, i.e., f = 1 and lymphocytes.9 8 Parameters7 S = 10 and R0 = 10 . Panel (b) depicts the e ect of decreasing the initial repertoire size from R0 = 10 , R0 = 10 ,9 to R0 = 10 8. Panel (c) depicts7 the e ect of incomplete tolerance induction, i.e., f = 1 and from R0 =f 10=0,.8R in0 = Eq. 10 (6)., to R0 = 10 . Panel (c) depicts the e ect of incomplete tolerance induction, i.e., f = 1 and f =0.8 in Eq. (6).Figure 1: The probability of mounting an immune response Pi from Eq. (2) as a function of the specificity p of the f =0.8 in Eq. (6). 5 9 lymphocytes. Parameters S = 10 and R0 = 10 . Panel (b) depicts the e ect of decreasing the initial repertoire size 9 8 7 from R0 = 10 , R0 = 10 , to R0 = 10 . Panel (c) depicts the e ect of incomplete tolerance induction, i.e., f = 1 and ⇧pPi of Eq. (3)f and=0.8 solving⇧p inP Eq.i of⇧ (6).p Eq.Pi = (3) 0 to and find solving that that⇧pP thei = maximum 0 to find is that at Pi that=1/S the. This maximum optimum is suggests at Pi =1/S. This optimum suggests that the lymphocyte⇧pPi of Eq.that specificity (3) the and lymphocyte is solving largely⇧ determinedpP specificityi = 0 to by find is the largely that number that determined of the self maximum epitopes by the the is number immune at Pi =1 system of/S self. This has epitopes optimum the immune suggests system has to be tolerantthat to. the Thus,to lymphocyte be the tolerant specificity specificity to. is Thus, not determined is the largely specificity determined by the recognitionis not by determined the of number pathogens, by of the self but recognition by epitopes the demand the of immune pathogens, system but by has the demand to remain tolerantto be⇧pP toleranti toof a Eq. large to. (3) number and Thus, solving theof self specificity⇧pP epitopes.i = 0 to is find Once not that lymphocytes determined that the maximum by are the specific recognition is at thePi repertoire=1 of/S. pathogens, This has optimumto be but suggests by the demand that theto lymphocyte remain tolerant specificity to a is large largely number determined of self by epitopes. the number Once of self lymphocytes epitopes the immune are specific system the has repertoire has to be su⇤cientlyto diverse remain tosu guarantee tolerant⇤ciently to recognition diverse a large to number of guarantee foreign of epitopes self recognition epitopes. (Fig. 1b). of Once foreign lymphocytes epitopes (Fig. are specific 1b). the repertoire has to be su⇤tociently be tolerant diverse to. to Thus, guarantee the specificity recognition is not of determined foreign epitopes by the recognition (Fig. 1b). of pathogens, but by the demand to remain tolerant to a large number of self epitopes. Once lymphocytes are specific the repertoire has to be su⇤ciently diverse to guarantee recognition of foreign epitopes (Fig. 1b). 1.1 Incomplete tolerance 1.1 Incomplete1.1 Incomplete tolerance tolerance Although there1.1 is promiscuous Incomplete expression tolerance of self antigens in the thymus, it remains unlikely that self tolerance is complete.Although HealthyAlthough there individuals is promiscuous there do harbor is promiscuous lymphocytes expression expression of that self can antigens recognize of self in antigens the self thymus, epitopes. in the it To thymus, remains study how unlikelyit remains the that unlikely self tolerance that self tolerance results changeis complete. whenAlthoughis tolerance complete. there Healthy is is promiscuous incomplete individuals Healthy we expression individuals define do harbor a of new self do lymphocytes parameter antigensharbor lymphocytes inf thefor that thymus, the can fraction recognize thatit remains of can self self unlikelyrecognize epitopes epitopes. that that self self To epitopes. tolerance study how To studythe how the manage to induce tolerance. For f = 1 the new model should be identical to the previous one. For a foreign resultsis complete. changeresults when Healthy change tolerance individuals when tolerance is incomplete do harbor is incomplete lymphocytes we define a we that new define can parameter recognize a new parameterf selffor epitopes. the fractionf for To the study of fraction self how epitopes the of self that epitopes that epitope we now require that it is recognized, but that none of the clonotypes recognizing the foreign epitope manageresults tomanage change induce when to tolerance. induce tolerance tolerance. For is incompletef = 1 For the wef new= define 1 model the a new new should parameter model be should identicalf for the be to fraction identical the previous of self to the epitopes one. previous For that a one. foreign For a foreign also recognize onemanage of the to (1 inducef)S tolerance.self epitopes For f that= 1 fail the to new induce model tolerance. shouldIncomplete be Otherwise identical tolerance the to the clone previous will be one. held For a foreign responsibleepitope forepitope auto-immunity. weepitope we now now require we require Following now that that require Borghans it it is is that recognized,et it al. is[4] recognized, but we but thatlet that nonebe none but the of fraction the that of the clonotypes none clonotypesof clonotypes of recognizing the clonotypes recognizing recognizing the foreign recognizing the foreign epitope the epitope foreign epitope also recognizealso recognize one of the one (1 off the)S self (1 epitopesf)fS S numberself that epitopes of self fail epitopes to that induce evoking fail tolerance. to tolerance induce (0 tolerance.< Otherwise f <= 1) the Otherwise clone will the be clone held will be held at least one ignoredalso recognize self epitope, one i.e., of the (1 f)S self epitopesR0 potential that failrepertoire, to induce R “functional” tolerance. repertoire, Otherwise p precursor the clone freq. will be held (1 f)S responsibleresponsibleresponsible for for auto-immunity. auto-immunity. for auto-immunity. Following Following=1 (1 Borghans Following Borghansp) et. Borghanset al. al.[4][4] we we letet let al.be [4] thebe we fraction the let fraction ofbe clonotypes the of fraction clonotypes(4) recognizing of clonotypes recognizing recognizing at leastat least oneat one least ignored ignored one self ignored self epitope, epitope, self i.e., i.e., epitope,Fraction i.e., of clones recognizing at least one ignored self epitope: The chance that the system remains tolerant when stimulated with a foreign(1 f(1)S f epitope)S (1 is thef)S probability that =1=1(1(1 p)=1p ) .(1 . p) . (4) (4) (4) none of the clones in the functional repertoire R will respond (with chance p ) and is potentially auto-reactive (with chanceThe ),The chance i.e., chance that that the the system system remains remains tolerant tolerantSize whenof when functional stimulated stimulated repertoire: with with a foreign a foreign epitope epitope is the probabilityis the probability that that The chance that the systemR remains tolerantfS when stimulated with a foreign epitope is the probability that nonenone of the of the clones clones inP in thet the= (1 functional functionalp ) repertoirewhere repertoireRR=RwillRwill0(1 respond respondp) (with. (with chance chancep) andp) is and potentially is potentially(5) auto-reactive auto-reactive (withnone chance of ), the i.e., clones in the functional repertoire R will respond (with chance p) and is potentially auto-reactive Now the chance(with of chance a “successful ), i.e., immune response” is theProbability probabilityR to remain that tolerant: the systemfS remains tolerant and (with chance ), i.e., Pt = (1 p ) R where R = R0(1 p) . fS (5) responds to the foreign epitope, which is the chancePt = to (1 remain p ) tolerantwhere minusR R the= R chance 0(1 top) not. respondfS at all: (5) Pt = (1 p ) where R = R0(1 p) . (5) Now the chance of a “successful immune response” is the probability that the system remains tolerant and Now the chance of a “successful immune response”R is the probability that the system remains tolerant and responds to the foreign epitope,P whichs = Pt is the(1Probability chancep) , to of remaina successful tolerant immune minus response: the chance to not(6) respond at all: respondsNow to the the foreign chance epitope, of a “successful which is the immune chance to response” remain tolerant is the probabilityminus the chance that the to not system respond remains at all: tolerant and responds to the foreign epitope, whichP = isP the(1 chancep)R , to remain tolerant minus the chance(6) to not respond at all: where the functional repertoire R is given in Eq. (5). To studys howt incomplete tolerance a ects the results we P = P (1 p)R , Borghans et al JI 1999 (6) plot Eq. (6) for f =0.8 and f = 1 in Fig. 1c. s t R where the functional repertoire R is given in Eq. (5). To studyPs = howPt incomplete(1 p) tolerance, a ects the results we (6) plot Eq. (6) for f =0.8 and f = 1 in Fig. 1c. Fig. 1c demonstrateswhere the that functional the e ect repertoire of incompleteR is tolerance given in is Eq. enormous. (5). To The study region how of incomplete specificity values tolerance where a ects the results we the chanceplot of a Eq. successful (6)where for response thef =0 functional.8P andapproachesf = repertoire 1 in one Fig. isR 1c. muchis given narrower. in Eq. Moreover (5). To study the optimum how incomplete has shifted tolerance a ects the results we Fig. 1c demonstrates thats the e ect of incomplete tolerance is enormous. The region of specificity values where leftwards, i.e., towardsplot a specificity Eq. (6) for muchf =0 smaller.8 and thanf =p 1=1 in/S Fig.. Thus 1c. the p =1/S estimate [8] is an upper the chance of a successful response Ps approaches one is much narrower. Moreover the optimum has shifted bound for theFig. lymphocyteleftwards, 1c demonstrates i.e., crossreactivity: towards that a the specificity when e ect the of much incomplete initial smaller repertoire than tolerancep is=1 su/S⇤ isciently. enormous. Thus large the p the The=1 immune/S regionestimate systemof specificity [8] is an upper values where operates eventhe betterbound chanceFig. when for of the 1c a lymphocytes successful demonstrates lymphocyte response are crossreactivity: that more theP specifics approaches e ect when [4]. of the The incomplete initialone conclusion is repertoire much tolerance remains narrower. is su is⇤ that enormous.ciently Moreover lymphocytes large The the the areimmune region optimum of system specificity has shifted values where specific to avoid auto-immunity, and not to recognize many pathogens. leftwards,operatesthe i.e., even chance towards better of when a a specificity successful lymphocytes much response are smaller morePs specificapproaches than p [4].=1 The one/S. conclusion is Thus much the remains narrower.p =1/S thatestimate Moreover lymphocytes [8] the is are optimum an upper has shifted boundspecific forleftwards, theto avoid lymphocyte auto-immunity, i.e., towards crossreactivity: and a specificity not to recognize when much the many smaller initial pathogens. repertoire than p =1 is/S su.⇤ Thusciently the largep =1 the/S immuneestimate system [8] is an upper operatesbound even better for the when lymphocyte lymphocytes crossreactivity: are more specific when [4]. the The initial conclusion repertoire remains is su⇤ciently that lymphocytes large the immune are system specific tooperates avoid auto-immunity, even better when and not lymphocytes2 to recognize are many more pathogens. specific [4]. The conclusion remains that lymphocytes are 2 specific to avoid auto-immunity, and not to recognize many pathogens.
2 2 Regulatory T cells K. Saeki et al. / Journal of Theoretical Biology 375 (2015) 4–12 7
A T cell scans a DC
The DC activates a TN cell The DC activates a TS cell A TR cell deactivates the DC No interaction p = y (1 - (1 - p)n ) p = y (1 - (1 - p)n + m (1- f ) ) p = y (1 - (1 - p)n + m (1- f (1 - α)) ) N N S S R R pU = 1- pN - pS - pR
Fig. 3. The four possible scenarios when a DC is scanned by a random T cell: (i) the encounter leads to the activation of a non-self-reactive T cell, (ii) the encounter leads to the activation of a self-reactive T cell, (iii) the DC is recognized by a regulatory T cell and deactivated, or (iv) the T cell does not recognize any of the presented epitopes and remains inactive. The occurrence of events (i)–(iii) requires both that the randomly selected T cell is of the appropriate type and that it recognizes at least one of the epitopes presented on the DC. This is reflected in the two terms constituting the probability of these events. Saeki et al JTB 2015
The occurrence of event (i) requires that a randomly encoun- tolerant to self, can be calculated as tered T cell is non-self-reactive and that it recognizes at least one k 1 of the n foreign epitopes on the DC. Therefore, the probability that k À i Pt pN pU pR ∑ pN pU a single T cell-DC encounter leads to activation of a non-self- ¼ð þ Þ þ i 0 ð þ Þ ¼ reactive T cell (and therefore an immune response to the patho- k gen), pN, is given by the product of these two probabilities, 1 p p p p k p Àð N þ U Þ ; 4 n N U R p y 1 1 p : 3a ¼ð þ Þ þ 1 pN pU ! ð Þ N ¼ Nð Àð À Þ Þ ð Þ Àð þ Þ To consider the probability of event (ii), we assume that all self- where the first term indicates the probability that the DC does not epitopes are presented on a DC with equal probability. Thus, a DC activate any Tregs or self-reactive T cells during its k encounters, is expected to present 1 f m different self-epitopes not pre- ð À Þ while the second term provides the probability that the DC is shut sented in the thymus, and the number of potential ligands of a down by a regulatory T cell before mounting an autoimmune self-reactive T cell on a DC is 1 f m n. Because the probability ð À Þ þ response. In the same manner, the probability that none of the of recognition for one epitope is p, the probability that a random T encounters leads to any immune response can be calculated as cell-DC encounter leads to activation of a self-reactive T cell is given by k 1 pU k P0 p p À ; 5 1 f m n U R p y 1 1 p ð À Þ þ : 3b ¼ þ 1 pU ð Þ S ¼ Sð Àð À Þ Þ ð Þ À According to the same assumption, the number of potential again including the probability that no recognition takes place in ligands of a regulatory T cell is expected to be n foreign epitopes all k encounters (first term), and the probability that none of the and 1 1 α f m self-epitopes, since Tregs potentially recognize ð Àð À Þ Þ encounters leads to activation of the T cell until the DC is shut any self-epitope that does not induce apoptosis in the thymus. down by a Treg (second term). Hence, pR, the probability that a randomly encountered T cell is Now, suppose that there are q activated DCs present in the regulatory and recognizes at least one epitope on the DC, which draining LN. We define an immune response as “successful” if leads to deactivation of the DC, is none of these DCs elicits an autoimmune response, while at least 1 1 α f m n one DC evokes an immune response to at least one foreign p y 1 1 p ð Àð À Þ Þ þ : 3c R ¼ Rð Àð À Þ Þ ð Þ epitope. Then the probability of a successful immune response Finally, the probability of no cognate recognition in an encoun- can be written as ter between a DC and a randomly selected T cell in the draining LN, p , is P Pq Pq: 6 U suc ¼ t À 0 ð Þ p 1 p p p : 3d U ¼ À R À S À N ð Þ i.e. the probability of staying tolerant to self, minus the probability As soon as a DC is recognized by a Treg, the DC will be shut that no reaction at all is mounted. To investigate the influence of down and lose its capacity to activate naïve T cells. Let k be the regulatory T cells on the immune system, this probability Psuc was average number of different T cell clones that scan a single studied as a function of T cell specificity p and Treg induction activated DC. In this model we set k R T T T , assuming likelihood parameter α. Other parameter values were set to ¼ ¼ R þ S þ N that scanning is quick and a DC is scanned by the whole peripheral S 105 self-epitopes (Burroughs et al., 2004), R 1010 initial T ¼ 0 ¼ repertoire of T cell clones. The probability that no self-reactive T cell clones (Borghans et al., 1999), f 0.9, m 90 self-epitopes, ¼ ¼ cell is activated in all these encounters, i.e. that the animal stays n 10 self-epitopes and q 1 DC. ¼ ¼ MHC: very polymorphic but not diverse.
Hundreds of alleles in every population Almost everyone heterozygous at every locus 6 major loci: 12 different MHC molecules per host
Why not many more? 0.3
R/R0
0.2
0.1
0 1 2 3 4 MHC diversity (log M )
Figure 2: Positive and negative selection according to the avidity model [13]. The curve in (a) depicts the distribution of thymocyte avidities for self peptide–MHC complexes. In our model, the chance p to be positively selected by a single MHC type is the chance that the avidity between the thymocyte T cell receptor and any of the self peptide– MHC complexes exceeds threshold T1. Thymocytes with avidities for self peptide–MHC complexes exceeding the upper threshold T2 are negatively selected (with chance n per MHC type). Panel (b) depicts the size of the T cell repertoire as a function of MHC diversity. The number of clones in the functional repertoire R is plotted as a fraction of the total initial lymphocyte repertoire R0. Parameters are: p =0.01, and n =0.005.
2 MHC diversity within the individual
Since individual MHC diversity increases the presentation of pathogens to the immune system, one may wonder why the number of MHC genes is not much higher than it is. The argument that is mostly invoked is that more MHC diversity within the individual would lead to T cell repertoire depletion during self tolerance induction. This argument is incomplete, however, because more MHC diversity could also increase the number of clones in the T cell repertoire through positive selection. In order to be rescued in the thymus, lymphocytes need to recognize MHC–self peptide complexes with su⇤cient avidity. A high MHC diversity thus increases both the number of lymphocyte clones that are positively selected and the number of clones that are negatively selected. To calculate the net e ect of these two opposing processes we develop a simple mathematical model [5].
Consider an individual with M di erent MHC molecules and an initial T lymphocyte repertoire consisting of R0 di erent clones. Let p and n denote theMHC (unconditional) diversity within an individual chances that a clone is positively selected by a single MHC type, because its avidity is higher than a threshold T1, or negatively selected because its avidity
p0.3 probability of positive selection exceeds a higher threshold T2, respectively (see Fig.R/R 2).n0 probability By of thisnegative selection definition, thymocytes can only be negatively Note0.2 p > n selected by MHC molecules by which they are also positivelyM number of MHC selected,molecules per hosti.e. n