Myelin Oligodendrocyte Glycoprotein induces ... - Wiley Online Library

Eur. J. Immunol. 2016. 46: 2247–2259

DOI: 10.1002/eji.201646416

L. E. Lucca et al.

Research Article

Myelin oligodendrocyte glycoprotein induces incomplete tolerance of CD4+ T cells specific for both a myelin and a neuronal self-antigen in mice Liliana E. Lucca ∗1,2,3 , Pierre-Paul Axisa ∗1,2,3 , Meryem Aloulou1,2,3 , Corine Perals1,2,3 , Abdulraouf Ramadan1,2,3 , Pierre Rufas1,2,3 , Bruno Kyewski4 , Jens Derbinski4 , Nicolas Fazilleau1,2,3 , Lennart T. Mars1,2,3 and Roland S. Liblau1,2,3,5 1

INSERM, U1043, Toulouse, France Centre National de la Recherche Scientifique, U5282, Toulouse, France 3 Centre de Physiopathologie Toulouse-Purpan, Université Toulouse 3, Toulouse, France 4 Developmental Immunobiology, Tumor Immunology Program, German Cancer Research Center, Heidelberg, Germany 5 CHU Toulouse, Département d’Immunologie, Toulouse, France 2

T-cell polyspecificity, predicting that individual T cells recognize a continuum of related ligands, implies that multiple antigens can tolerize T cells specific for a given self-antigen. We previously showed in C57BL/6 mice that part of the CD4+ T-cell repertoire specific for myelin oligodendrocyte glycoprotein (MOG) 35–55 also recognizes the neuronal antigen neurofilament medium (NF-M) 15–35. Such bi-specific CD4+ T cells are frequent and produce inflammatory cytokines after stimulation. Since T cells recognizing two selfantigens would be expected to be tolerized more efficiently, this finding prompted us to study how polyspecificity impacts tolerance. We found that similar to MOG, NF-M is expressed in the thymus by medullary thymic epithelial cells, a tolerogenic population. Nevertheless, the frequency, phenotype, and capacity to transfer experimental autoimmune encephalomyelitis (EAE) of MOG35-55 -reactive CD4+ T cells were increased in MOGdeficient but not in NF-M-deficient mice. We found that presentation of NF-M15-35 by I-Ab on dendritic cells is of short duration, suggesting unstable MHC class II binding. Consistently, introducing an MHC-anchoring residue into NF-M15-35 (NF-M15-35 T20Y) increased its immunogenicity, activating a repertoire able to induce EAE. Our results show that in C57BL/6 mice bi-specific encephalitogenic T cells manage to escape tolerization due to inefficient exposure to two self-antigens.

Keywords: Autoimmunity

r

Central nervous system

r

Polyspecificity

r

T cell

r

Tolerance

Additional supporting information may be found in the online version of this article at the publisher’s web-site

Correspondence: Prof. Roland S. Liblau e-mail: [email protected]

C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

∗

These authors contributed equally to this work.

www.eji-journal.eu

Basic

Immunomodulation and immune therapies

2247

2248

L. E. Lucca et al.

Introduction By discriminating between self and non-self, the adaptive immune system maintains self-tolerance, thus keeping at bay potentially deleterious autoreactive lymphocytes. For T cells, the induction of self-tolerance is accomplished through deletion or deviation into a regulatory phenotype of developing T cells with high affinity for self-antigens in the thymus (central tolerance) and further silencing of autoreactive T cells in the periphery (peripheral tolerance) [1]. Both central and peripheral tolerances require that self-antigens are processed and presented in the context of MHC molecules. In the thymus, medullary thymic epithelial cells (mTECs) have the unique ability to express a vast array of tissue-specific antigens (TSAs). This projection of self-antigens is achieved by promiscuous gene-transcription in part promoted by the transcription factor AIRE [2]. While cross-presentation of TSAs and peripheral antigens by resident and migratory dendritic cells plays a major role in establishing central tolerance, mTECs are now recognized as both suppliers and presenters of TSAs, through unconventional pathways of antigen processing and MHC class II loading [3–6]. Tolerance to self-antigen has been elegantly studied by comparing the magnitude of the antigen-specific T cell response between mice that express or do not express the self-antigen in question [7]. For example, in B.10PL mice, the immunodominant myelin basic protein (MBP) epitope is MBPAc1-11 . MBP−/− mice on the same background display unchanged responses to MBPAc1-11 , but respond strongly to MBP121-150 , a novel immunodominant peptide [8]. Indeed, while MBP121-150 is a stable binder of I-Au , MBPAc1-11 forms very unstable MHC:peptide complexes and thus fails to elicit tolerance [9]. Notwithstanding the importance of these studies in defining mechanisms of tolerance induction, it could be argued that they do not take into account the implications of T-cell polyspecificity on tolerance. This concept, whereby each TCR can recognize several distinct related antigens [10], is supported by a growing body of evidence. Initial work on altered peptide ligands showed that analogs of an immunogenic peptide harbouring mutated TCR contact positions retained a stimulatory potential [11]. Moreover, comparison of the estimated size of the TCR repertoire and the repertoire of ligands that can be presented by the MHC molecules implies that each TCR has to recognize a broad array of different antigens [12, 13]. Recently, identification of ligands bound by given TCRs by deep-sequencing revealed a striking polyspecificity relying on conservation of the TCR binding motifs, hence indicating that each TCR displays specificity for a continuum of structurally related antigens [14]. An important implication of T-cell polyspecificity is molecular mimicry, defined as cross-activation of autoreactive lymphocytes by microbe-derived molecules due to structural similarities between foreign and selfantigens [15–19] and considered an important liability in development of autoimmunity [20]. Given the vast spectrum of peptides that can be recognized by a single TCR, it is likely that several selfantigens could trigger a given TCR, leading to an extension of the


Eur. J. Immunol. 2016. 46: 2247–2259

concept of molecular mimicry to two or more self-antigens. This also poses the question of the impact of self-polyspecificity on tolerance and autoimmunity. Arguably, T cells that recognize multiple self-antigens could have an increased chance of being tolerized in the thymus and/or the periphery. Indeed, a few cases of molecular mimicry between self-antigens with a tolerogenic outcome have been reported. In the SJL mouse a very high frequency of PLP139-151 -reactive CD4+ T cells display an antigen-experienced phenotype. While these cells persist in PLP−/− mice, their incapacity of producing effector cytokines upon in vitro antigenic stimulation implies an encounter with a tolerogenic self-antigen other than PLP [21]. In a mouse model of herpetic stromal keratitis, in which HSV-1 infection induces a T-cell-mediated reaction against a corneal self-antigen, disease resistance across mouse strains is found to be associated with immunoglobulin gene allotypes encoding for a peptide cross-reactive with the corneal target of the autoimmune attack. The authors proposed that bi-specificity for two self-antigens, one tissue-restricted (corneal) and one readily available (immunoglobulin), could be a way to strengthen protection from autoimmunity against sequestered selfantigens [22]. In C57BL/6 mice, we showed that CD4+ T cells expressing the 2D2-TCR exhibited specificity for both the myelin oligodendrocyte glycoprotein (MOG)35-55 peptide and a neurofilament medium (NF-M)15-35 peptide in the context of I-Ab [23]. While MOG is exclusively present in central nervous system (CNS) myelin, NF-M is a neuronal intermediate filament localized in axons and neurons of both CNS and peripheral nervous system (PNS). The 2D2 T cells exhibited increased pathogenic capacity upon recognition of the two self-antigens, demonstrating that both are relevant targets in vivo, a feature we coined cumulative autoimmunity [24]. NF-M18-30 shares sequence homology with MOG38-50 , including four amino acids that are TCR contact residues for MOG recognition [23, 25, 26]. These residues are crucial for TCR triggering of NF-M18-30 -specific clones, proving that bi-specificity arises indeed from molecular mimicry [27]. Moreover, bi-specificity for MOG35-55 and NF-M15-35 is a feature shared by different MOG35-55 -specific CD4+ T cells of C57BL/6 mice. As a consequence, NF-M15-35 is a relevant target for the polyclonal CD4+ T cell response triggered during MOG35-55 -induced experimental autoimmune encephalomyelitis (EAE) [28]. Given that bi-specific T cells could be expected to be more efficiently tolerized, and thus prevented from partaking into the autoimmune response of MOG35-55 -induced EAE, the goal of the current study was to understand how these bi-specific T cells escape tolerance induction. While we and others have reported the effect of genetic MOG deletion on the magnitude of the MOG35-55 -specific response [29–31], these studies did not take into consideration the impact of NF-M on tolerance of the bi-specific CD4+ T cells. In this work, by analyzing the frequency and functionality of these cells in mice lacking MOG and NF-M, alone or in combination, we evaluate the implications of TCR polyspecificity to self on tolerance induction.

www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259


transcripts was also occasionally detected in thymocytes (4/12 samples). Notably, using a semi-quantitative approach, the signal for MOG was detected in mature (CD80hi ), but not in immature, mTECs and the signal for NF-M was stronger in mature than in immature mTECs (Fig. 1B). We did not assess MOG protein expression in mTECs because of the limited number of cells obtained after cell sorting and overall low sensitivity.

Frequency of MOG35-55 -specific conventional CD4+ T cells is increased in MOG−/− NF-M−/− mice

Figure 1. NF-M and MOG are expressed in medullary thymic epithelial cells (mTECs) of C57BL/6 mice. (A) RNA purified from brain or thymus tissue from C57BL/6 mice or MOG−/− NF-M−/− mice was used in RT-PCR reactions to assess expression of NF-M and HPRT. In the RT- samples, no retrotranscription of the mRNA was perfomed prior to PCR. (B) Expression of NF-M, MOG and HPRT was assessed in medullary thymic epithelial cells (mTECs), thymic dendritic cells (DCs), and thymic macrophages (Mφ). Black wedges indicate serial threefold dilutions of the cDNA. Data representative of two independent experiments for NF-M expression and three independent experiments for MOG expression are shown.

Results NF-M and MOG are expressed in the thymus by medullary thymic epithelial cells A vast array of organ-specific self-antigens is expressed in the thymus, allowing induction of a broad scope of central tolerance. In order to study the impact of the expression of MOG and NFM on the maturation of the bi-specific CD4+ T cell population, we initially assessed the thymic expression of these two antigens. Our group and others have previously described presence of MOG transcripts in the thymus [30, 32], and more specifically in mTECs [33] of a wild-type (WT) mice. By performing RT-PCR with a primer pair spanning exons 1 to 3 of the nefm gene, we detected NF-M mRNA in the brain but also in the thymus of C57BL/6 mice. This amplicon was absent in age-matched MOG−/− NF-M−/− mice, proving specificity of the signal (Fig. 1A). In order to identify the cellular source of the NF-M transcript, we sorted thymic populations obtained as previously described [33]. As depicted in Fig. 1B, NF-M and MOG expression was detected in mTECs but not in macrophages and DCs. Low expression of MOG C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

To evaluate the impact of MOG and NF-M expression on the bispecific repertoire, we first compared the thymic development of MOG/NF-M-bi-specific 2D2 T cells maturing in WT versus the MOG−/− NF-M−/− background. No difference was observed in terms of deletion and Treg induction of the 2D2-expressing thymocytes (Supporting Information Fig. 1). Since the T cell clone harboring the 2D2 TCR was isolated in the periphery of a MOG35-55 -immunized C57BL/6 mouse [34], it is possible that the affinity of this TCR for MOG and NF-M is below the threshold for thymic tolerance induction. Hence, we next analyzed the polyclonal MOG-specific T cell repertoire. We first compared the immunogenicity of MOG35-55 between WT C57BL/6 mice and mice lacking both MOG and NF-M expression. If part of the polyclonal MOG35-55 -specific response is indeed tolerized by encounter with MOG or NF-M, antigen-deficient mice would exhibit quantitative and/or qualitative changes in the response to MOG35-55 immunization. In order to identify MOG35-55 -specific and MOG35-55 /NF-M15-35 bi-specific CD4+ T cells, we used MHC-tetramers. Identification of MOG-specific CD4+ T cells by flow cytometry using MOG38-49 I-Ab tetramers was described previously [35, 36]. In addition, we obtained NF-M18-30 -I-Ab tetramers but, in our hands, this tool failed to identify NF-M15-35 -specific CD4 T cells. Hence, we focused on the MOG35-55 -specific pool using MOG38-49 -I-Ab tetramers. After MOG35-55 immunization, an increased frequency of MOG38-49 -I-Ab tetramer+ CD44hi CD4+ T cells was observed in the draining lymph nodes of MOG−/− NF-M−/− mice as compared to WT mice (Fig. 2A and B). Taking into account that induction of Treg cells is a major mechanism of tolerization, we distinguished conventional T cells (Tconv) from Foxp3+ Treg cells among the tetramer+ CD4+ T cells of WT and MOG−/− NF-M−/− mice. While the absolute number of tetramer+ Foxp3+ Treg cells was unaffected by the presence or absence of MOG and NF-M, the total number of tetramer+ Foxp3− Tconv was threefold higher in MOG−/− NF-M−/− as compared to WT mice, resulting in a greater Tconv/Treg cell ratio (Fig. 2C and D). The mean fluorescence intensity (MFI) of the tetramer staining is considered a reliable measure of the affinity of the TCR for the respective peptide-MHC complex [37]. We therefore compared the tetramer MFI of the MOG35-55 -specific CD4+ T cells in WT and MOG−/− NF-M−/− mice and observed that in the absence of the two self-antigens, the overall TCR affinity of the MOG-specific repertoire was increased both in the Tconv and the Treg subsets (Fig. 2E). Treg cells from both WT and MOG−/− NF-M−/− mice www.eji-journal.eu

2249

2250

L. E. Lucca et al.

Eur. J. Immunol. 2016. 46: 2247–2259

Figure 2. Increased frequency of MOG35-55 -specific conventional CD4+ T cells in MOG−/− NF-M−/− mice. (A) Representative FACS plots for MOG38-49 I-Ab tetramer versus CD44 or Foxp3 staining on draining lymph node cells from MOG−/− NF-M−/− or WT mice, 9 days after MOG35-55 immunization. Gating on CD4+ CD44high cells (top) and CD4+ CD44hi tetramer+ cells (bottom) are shown. Gates for the MOG38-49 -I-Ab tetramer staining were positionned based on the control CLIP-I-Ab tetramer. (B) Percentage of MOG-specific CD4+ T cells as evaluted by flow cytometry. (C) T conv / Treg cell ratio among tetramer+ CD4+ T cells in the indicated genotypes. (D) The absolute numbers and (E) MFI of MOG38-49 -I-Ab tetramer staining of MOG-specific conventional (Foxp3− ) and regulatory (Foxp3+ ) CD4+ T cells are shown. (F) Histograms showing Nrpl-1 surface expression among MOG-specific regulatory (Foxp3+ ) CD4+ T cells from indicated genotypes or isotype control. All analyses are gated on CD19− CD8− CD4+ CD44hi cells. Symbols represent individual samples/mice and bars represent the mean ± SEM of eight mice per group from one experiment representative of three experiments performed. *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t-test.

expressed neuropilin-1, at similar levels, indicative of their thymic origin (Fig. 2F). Therefore, both the number of MOG-specific CD4+ T cells and their affinity for I-Ab :MOG35-55 are increased following immunization of MOG−/− NF-M−/− mice.

MOG but not NF-M expression blunts accumulation of MOG35-55 -specific and bi-specific CD4+ T cells Since the MOG35-55 -specific CD4+ T cell compartment exhibits enhanced responses in MOG−/− NF-M−/− mice following MOG35-55 -immunization, we analyzed the individual contribution of MOG and NF-M expression on the MOG35-55 -specific and the MOG35-55 /NF-M15-35 bi-specific T cell responses. To study the bispecific response, we measured antigen-induced expression of CD40L as a marker of Ag-specific CD4+ T cells [38], since NFM tetramer was not usable. We enumerated splenic CD4+ T cells that up-regulate CD40L upon in vitro restimulation either with MOG35-55 or NF-M15-35 from MOG35-55 -immunized mice of different genotypes. As depicted in Fig. 3A and B, upon MOG35-55 restimulation about 3% of total live splenic CD44hi CD4+ T cells from WT mice express CD40L, compared to 0.5-0.8% with the control OVA323-339 peptide or without antigen. NF-M15-35 -restimulation induced CD40L in 2% of CD4 T cells. CD40L expression was higher in MOG−/− NF-M−/− and MOG−/− animals as compared to WT or NF-M−/- mice, indicating that absence of MOG expression increased the frequency of MOG35-55 -specific CD4+ T cells. Similar results were observed upon NF-M15-35 restimulation (Fig. 3A and B), indicating that MOG expression also affects the frequency of bi-specific CD4+ T cells. The absence of NF-M expression had no additional effect on T-cell responses in either MOG-deficient or MOG-sufficient mice. C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

We next asked whether the higher frequency and affinity of MOG35-55 -specific CD4+ T cells in MOG−/− NF-M−/− mice could be linked to a different composition of the TCR repertoire. To this end, we sorted CD40L+ CD4+ T cells of mice from the 4 genotypes (WT; MOG−/− ; NF-M−/− ; MOG−/− NF-M−/− ) after MOG35-55 immunization and in vitro restimulation and measured the relative usage of each Vβ gene by RT-qPCR (Fig. 3C). Consistent with previous reports [39, 40], Vβ1, Vβ6 and Vβ8.2 usage was increased among MOG35-55 -reactive T cells, and equally across the four genotypes (Fig. 3C). The usage of other Vβ was also similar between groups (Supporting information Fig. 2A). Therefore, the absence of MOG and/or NF-M did not have a significant impact on the overall Vβ usage among MOG-specific T cells. We then assessed whether MOG or NF-M expression affected the use of the GETGGNYAEQ CDR3 motif, resulting from a public Vβ8.2-Jβ2.1 rearrangement found at high frequency on CNS-infiltrating CD4+ T cells during MOG35-55 -induced EAE [29]. This rearrangement was detectable at a frequency of 3% in WT mice (Fig. 3D). Strikingly, the GETGGNYAEQ CDR3 motif was increased fivefold in MOG−/− NF-M−/− and MOG−/- mice as compared to WT, reaching approximately 15% of total Vβ8.2 transcripts. Conversely, NF-M-deficiency did not have an impact on the use of this specific rearrangement as the frequency was similar in NF-M−/− versus WT mice, and in MOG−/− NF-M−/− versus MOG−/− mice. Of note, measurement of the frequency of the GETGGNYAEQ rearrangement among sorted Vβ8.2+ CD4+ T cells from non-immunized mice deficient for either MOG or NF-M or in combination, or in WT mice did not reveal any difference (Supporting information Fig. 2B). Collectively, these data indicate that expression of MOG blunts the selection and/or expansion of a public rearrangement associated with EAE, indicating an effect on the composition of the specific repertoire following immunization. www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259


Figure 3. MOG expression limits the accumulation of MOG35-55 -specific and MOG35-55 /NF-M15-35 bi-specific CD4+ T cells and usage of a public TCRβ rearrangment after MOG35-55 immunization. (A) Splenocytes of MOG35-55 immunized mice were stained with anti-CD40L mAb 6 hours after stimulation ex vivo with irradiated splenocytes and MOG35-55 , NF-M15-55 or OVA323-339 peptides. Events were gated on live Thy1.2+ CD4+ CD44hi cells. (B) Quantification of CD40L+ cells among total live CD4+ CD44hi T cells under the indicated restimulation conditions. Data are shown as mean ± SEM of four pools (of two mice each) from four independent experiments. *p < 0.05, one-way ANOVA. (C) Mice of the indicated genotypes were immunized with MOG35-55 and at day 9 cells from spleen and daining inguinal lymph nodes were restimulated with MOG35-55 (naive WT mouse served as a control). Percentage of indicated Vβ usage analyzed by RT-qPCR on FACS-sorted live Thy1.2+ CD4+ CD44hi CD40L+ cells. (D) Percentage of the GETGGNYAEQ public Vβ8.2Jβ2.1 rearrangement among Vβ8.2 RTqPCR products. (C, D) Data are shown as the mean ± SEM of 3–4 mice per group pooled from four independent experiments. ***p < 0.001, ****p < 0.0001, one-way ANOVA.

MOG, but not NF-M, expression restrains the immunogenicity and encephalitogenicity of MOG35-55 Having established that MOG limits the accumulation of MOG35-55 specific CD4+ T cells, we explored their functional capacity. To this end, WT mice and MOG−/− NF-M−/− mice were immunized with MOG35-55 and the production of IFNγ and IL-17 by purified CD4+ T cells from the spleen was assessed upon ex vivo restimulation with MOG35-55 or NF-M15-35 . We observed a marked and reproducible increase in the production of these proinflammatory cytokines by MOG−/− NF-M−/− compared to WT CD4+ T cells (Fig. 4A–D). This observation indicates that the expression of either or both antigens is essential in limiting MOG35-55 specific Th1 and Th17 responses. Moreover, increased cytokine secretion was observed regardless of the peptide used for restimulation (MOG35-55 or NF-M15-35 ), indicating that both the MOG35-55 monospecific and the MOG35-55 /NF-M15-35 bi-specific compartments are affected (Fig. 4A–D). We next investigated the relative contribution of MOG and NF-M expression in this process. Strikingly, in immunized MOG−/− mice, the recall production of IFNγ C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

and IL-17 by CD4+ T cells was comparable to the levels observed in MOG−/− NF-M−/− mice, while the recall responses from NF-M−/− mice did not differ from those of WT mice. These results indicate that, while MOG expression blunts the functional responses of MOG-specific CD4+ T cells, NF-M expression does not impact the bi-specific compartment. To assess whether increased immunogenicity translates into enhanced pathogenicity, passive EAE was induced by transfer of activated CD4+ T cells from MOG35-55 -immunized WT or MOG−/− NF-M−/− mice in WT recipients. Recipients of MOG−/− NF-M−/− CD4+ T cells developed a more severe EAE (Supporting information Fig. 3). Therefore, as expected, the increased frequency of MOG-specific CD4+ T cells in MOG−/− NF-M−/− mice and their increased production of proinflammatory cytokines translated into enhanced pathogenicity. To evaluate the T-cell intrinsic properties following priming under the same antigenic environment, we transferred 2 × 106 T cells from unmanipulated mice of the 4 genotypes in RAG2−/− recipients and induced EAE by MOG35-55 -immunization 24 hours later. As presented in Fig. 4E, this suboptimal protocol is www.eji-journal.eu

2251

2252

L. E. Lucca et al.

Eur. J. Immunol. 2016. 46: 2247–2259

Figure 4. Increased encephalitogenicity of MOG35-55 -specific CD4+ T cells in MOG-deficient mice. (A-D) Purified CD4+ T cells from the spleen of MOG35-55 -immunized mice of the indicated genotypes were restimulated in vitro with (A, C) MOG35-55 or (B, D) NF-M15-35 . The production of (A, B) IFN-γ and (C, D) IL-17 was determined by ELISA. Data are shown as the mean ± SEM of 4 pools of 4–10 WT, MOG−/− ,MOG−/− NF-M−/− mice and 3 pools of 4–10 NF-M−/− mice, from 4 independent experiments. (E-G) 2 × 106 unfractionated T cells from unimmunized mice of the indicated genotypes were adoptively transferred into RAG2−/− recipients, followed by immunization with MOG35-55 . (E) Daily EAE score of recipients of T cells from mice of the indicated genotypes was monitored. (F) Kaplan-Meier curve for the percentage of disease-free mice is shown. *p < 0.05, **p < 0.01, Log-rank test. (G) The mean cumulative EAE score of recipients of T cells from the indicated genotypes is shown. *p < 0.05, **p < 0.01, Kruskall–Wallis test. Data are shown as mean ± SEM of 13 MOG−/− NF-M−/− mice, 9 MOG−/− mice, 7 NF-M−/− mice and 10 WT mice pooled from 4 independent experiments.

ineffective for the induction of full-blown EAE in recipients of WT or NF-M−/− donor T cells. Nevertheless, recipients transferred with T cells from MOG−/− NF-M−/− or MOG−/− mice developed classical EAE at a rate of 77% and 56%, respectively (Fig. 4F). Moreover, the cumulative score of mice receiving MOG−/− NF-M−/ T cells was significantly increased compared to the MOG-expressing WT and NF-M−/− groups (Fig. 4G). Taken together, these data indicate that the CD4+ T-cell compartment from MOG-deficient mice displays enhanced encephalitogenic property and enhanced ability to transfer disease.

NF-M expression does not result in tolerance of the NF-M15-35 -specific T cells We next asked whether immune tolerance to NF-M could be revealed by studying the immunogenicity of NF-M15-35 . However, since we previously reported that NF-M15-35 binds poorly to I-Ab [27], we reckoned that the NF-M15-35 peptide is an inefficient tool. Indeed, immunization of WT or even MOG−/− NF-M−/− mice with NF-M15-35 did not elicit any reliable response (data not shown). To circumvent this limitation, we mutated the NF-M15-35 peptide at position 20 (NF-M15-35 T20Y peptide) to increase its affinity for C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

I-Ab . Indeed, using a biochemical competition assay [27], the IC50 of NF-M15-35 T20Y was reduced by 20-fold (37 nM) as compared to that of the native NF-M15-35 (685 nM), indicating its increased affinity for I-Ab . To test the impact of the NF-M15-35 T20Y substitution on antigen presentation, we used a protocol in which BMDCs were briefly pulsed with soluble peptides, left to rest for variable times after thorough washing, and then used to stimulate purified 2D2 CD4+ T cells labeled with CellTace violet. As depicted in Supporting information Fig. 4, NF-M15-35 T20Y allowed a more stable T-cell stimulation as compared to NF-M15-35 . These results indicate that presentation of NF-M15-35 by professional APCs is short-lived but can be efficiently prolonged by substitution at an anchoring residue. The NF-M15-35 T20Y peptide was then used to probe the size and phenotype of the NF-M15-35 -monospecific and MOG35-55 /NFM15-35 bi-specific T cell pools. We immunized WT and NFM−/− mice with NF-M15-35 T20Y and enumerated splenic CD4+ T cells that up-regulated CD40L upon in vitro restimulation with MOG35-55 , NF-M15-35 or NF-M15-35 T20Y, nine days after immunization. As depicted in Fig. 5A and B, upon NF-M15-35 T20Y restimulation 2.5%±0.3 of splenic CD44hi CD4+ T cells from WT mice upregulated CD40L, compared to 0.24%±0.1 without www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259


Figure 5. NF-M15-35 T20Y-immunization does not reveal tolerance to NF-M and induces MOG-dependent EAE. (A) Splenocytes of NF-M15-35 T20Yimmunized mice were stained with anti-CD40L mAb during 6 h of restimulation in vitro with MOG35-55 , NF-M15-35 , NF- M15-35 T20Y, OVA323-339 or no antigen. The population displayed is gated on live Thy1.2+ CD4+ CD44hi cells. (B) Frequency of CD40L+ cells among total live Thy1.2+ CD4+ CD44hi T cells in the indicated restimulation conditions. Data are shown as the mean ± SEM of nine mice pooled from three independent experiments. Statistical analyses were performed to compare each experimental condition to the absence of antigen for each genotype (*p < 0.05, **p < 0.01,***p < 0.001, ****p < 0.0001, one-way ANOVA). (C) Purified CD4+ T cells from the spleen of NF-M T20Y-immunized mice of the indicated genotypes were restimulated with NF-M15-35 T20Y, NF-M15-35 or MOG35-55 in vitro. The production of IFN-γ and IL-17 was measured by ELISA. Data are shown as the mean ± SEM of six mice pooled from two independent experiments. (D) Daily EAE score of the indicated mice immunized with NF-M15-35 T20Y. Data are shown as mean ± SEM of nine mice per group pooled from two independent experiments.

restimulation and 0.29%±0.1 with the control OVA323-339 peptide. NF-M15-35 T20Y immunization also allowed activation of a T-cell repertoire recognizing the native NF-M15-35 as CD40L induction was observed on 1.8%±0.3 of WT CD4+ T cells after restimulation with NF-M15-35 , demonstrating that the T20Y mutation did not alter the orientation of key TCR-contact residues. Importantly, the proportion of NF-M15-35 -reactive T cells was not increased in NF-M−/− mice (Fig. 5B). Of note, NF-M15-35 T20Y-immunization was able to activate bi-specific CD4+ T cells that also recognize MOG35-55, as around 1% of CD4+ T cells responded to MOG35-55 restimulation in both WT and NF-M−/− mice. The cytokine production of purified splenic CD4+ T cells to NF-M15-35 T20Y, NF-M15-35 and MOG35-55 was then assessed nine days after immunization (Fig. 5C). Using NF-M15-35 T20Yimmunization we were able to prime in vivo NF-M15-35 -specific and bi-specific CD4+ T cells but did not observe any differences in the production of IFNγ or IL-17 between WT and NF-M−/− mice in response to either peptide. This result confirms that expression of NF-M does not tolerize the NF-M15-35 -specific repertoire, including the MOG35-55 /NF-M15-35 bi-specific T cells. In order to evaluate the pathogenicity of these cells, we used NF-M15-35 T20Y as the immunogen to induce EAE. Remarkably, WT mice developed signs of ascending paralysis typical of EAE (Fig. 5D). To investigate the relative contribution of MOG and NF-M as targets of the encephalitogenic T-cell response evoked by NF-M15-35 C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

T20Y-immunization, we also immunized MOG−/− , NF-M−/− and MOG−/− NF-M−/− mice with the NF-M T20Y peptide and monitored disease development (Fig. 5D). While double-deficient mice remained free of disease, NF-M−/− mice developed EAE with a similar frequency and severity as WT mice. In the MOG−/− group, only three mice (33%), displayed clear hind limb paresis, and the disease only lasted for a few days. These results indicate that the NF-M15-35 -specific repertoire activated by NF-M15-35 T20Yimmunization establishes moderate EAE mostly through in vivo cross-recognition of MOG35-55 . In contrast, the targetting of NFM15-35 alone, as occurs in MOG−/− mice, is much less effective at inducing disease.

Discussion The functional consequences of T-cell polyspecificity for distinct self-antigens are only starting to be explored. A T cell with such a property, when activated, could mediate an autoimmune attack directed at multiple targets and could therefore represent a particular threat for the organism [23]. Conversely, T cells recognizing two self-antigens could have a “double chance” to be tolerized by the expression of their cognate ligands, provided the two antigens are presented in a tolerogenic context. It is therefore surprising that bi-specific T cells, able to produce inflammatory cytokines www.eji-journal.eu

2253

2254

L. E. Lucca et al.

upon activation and drive autoimmune disease, are detected at such high frequency in C57BL/6 mice [27, 28]. MOG represents only 0.1% of CNS myelin proteins and is absent from PNS myelin. As such, MOG epitopes are scarce in the periphery [30]. In this work, using a previously described semiquantitative approach [7, 33], we report thymic expression of MOG in mature mTECs at the mRNA level. Several groups have investigated ectopic expression of self-antigens in the thymus, some succeding in detecting MOG [30, 32, 33, 41], while others reported its absence from the array of thymus-expressed TSA [42]. This lack of consistency in previous findings could be due to the fact that individual TSAs are expressed by only a low proportion of mTECs (1–2%) [3] and with a rapid turnover. Moreover, such low level of expression challenges the dection of TSAs at the protein level. NF-M is a structural protein of CNS and PNS neurons, where it integrates the axonal cytoskeleton. Hence, NF-M epitopes could be presented in the periphery. Moreover, we show here that NF-M transcripts are detected in the thymus of C57BL/6 mice, specifically in mTECs. Nevertheless, the analysis of the impact of MOG and NF-M expression on the development of the MOG35-55 -specific and MOG35-55 /NFM15-35 -bi-specific T cells revealed unexpected findings. When we analyzed the immunogenicity of MOG35-55 with respect to the expression of MOG and NF-M, we observed exacerbated CD4+ T-cell responses in the absence of MOG, but not in the absence of NF-M alone. These observations suggest that MOG tolerizes the MOG35-55 -specific CD4+ T cell response. A wide heterogeneity of mechanisms and efficay of tolerance induction between autoantigens has been described over the years. Recently, Malhotra and colleagues took advantage of the numerous tissue-specific GFP transgenic mouse models to study systematically T-cell tolerance towards a GFP peptide, using tetramers [43]. They identified three clusters of tolerance induction depending on GFP expression patterns: (i) ignorance, (ii) partial clonal deletion with concomitant Treg induction, and (iii) profound clonal deletion. According to our results MOG35-55 falls into the second cluster. Hence, C57BL/6 mice remain highly susceptible to MOG35-55 induced EAE [44], and it has been reported that overexpression of MOG in the thymus of C57Bl/6 mice results in partial resistance to EAE induction [45]. Although the high MOG35-55 -specific Treg frequency observed, notably in immunized WT mice, seems counter-intuitive, our data are congruent with previous studies using the same tetramer tools [46–48]. Instability of Treg phenotype in the context of MOG35-55 immunization may enable the mounting of a pathogenic autoimmune response. In light of the increased immunogenicity of MOG35-55 in MOG−/− mice, we interpret the lack of impact of MOG expression on 2D2 T-cell development as a reflection of this TCR originating from a WT mouse and having therefore escaped tolerance induction. Isolating autoreactive TCRs from a MOG−/− NF-M−/− background and reintroducing them in retrogenic mice expressing MOG and NF-M alone or in combination will provide a more sensitive tool for assessing the effect of MOG and NF-M on thymic development and peripheral maturation of trackable T cell clones. The increased frequency


Eur. J. Immunol. 2016. 46: 2247–2259

of MOG35-55 -specific CD4+ T cells in the polyclonal repertoire of MOG-deficient mice was accompanied by an increased affinity of their TCR. This result is consistent with the notion that negative selection allows the elimination of autoreactive T cell clones harbouring a high affinity TCR while those carrying a TCR of lower affinity are spared [37, 49, 50]. The persistence of MOG35-55 specific CD4+ T cells with high affinity TCRs in MOG-deficient animals could result in a repertoire with a lower threshold for activation following MOG35-55 immunization, thus leading to the observed high frequency of autoreactive cells. This increase favors the generation of a more pathogenic repertoire as observed by the enhanced production of inflammatory cytokines and enhanced encephalitogenicity of CD4+ T cells from MOG-deficient animals. We propose that such increased ability to transfer disease could be due to both higher frequency of MOG35-55 -specific cells in the pre-immune repertoire and better expansion after immunization, but we did not directly compare the relative contribution of these two mechanisms. As such, our current study conflicts with previous data from our group and others that failed to detect increased MOG35-55 -specific CD4+ T-cell responses or greater pathogenic potential of primed CD4+ T cells in a different MOG−/− mouse model [29, 30, 51]. The difference could possibly be attributed to the fact that, in the aforementioned MOG−/− mouse strain, the targeting cassette had been inserted downstream of the region encoding MOG35-55 , and by consequence in case of abortive transcription and translation of the targeted mog gene, a residual amount of MOG35-55 epitopes could be generated in the thymus [52], possibly involving Aire in facilitating mRNA processing [53]. Of note, study of tolerance to MOG35-55 in other H-2b strains demonstrated increased inflammatory profiles of CD4+ T cells from MOG-deficient animals [31]. It appears surprising that, despite using different approaches to detect tolerance induction by NF-M, we did not observe any qualitative and quantitative differences in the response to NF-M15-35 in the presence or absence of NF-M. Nevertheless, we previously reported that NF-M15-35 is non-immunogenic in C57BL/6 mice and that NF-M15-35 binds to I-Ab without a clearly defined binding motif [27]. This feature could also explain why MHC tetramers loaded with NF-M15-35 failed to detect NF-M15-35 -specific T cells. Moreover, using an in vitro approach to assess duration of antigen presentation by DCs, NF-M15-35 -presentation was found to be short-lived, but can be prolonged by mutation of a residue at an MHC-anchoring position that increases affinity for I-Ab . Previous reports demonstrated that MHC-binding residues of a given peptide can be modified without affecting the 3-dimensional structure determining TCR interaction [8]. Taken together, these observations suggest that NF-M expression does not impact tolerance due to suboptimal presentation of the NF-M15-35 peptide by the I-Ab . Our observation is reminiscent of the case of T-cell responses against MBP in H-2u mice, where instability of the binding between I-Au and MBPAc1-11 prevented tolerization of the repertoire specific for this self-antigen [9]. Moreover, spontaneous development of diabetes in NOD mice has been linked to the poor binding of self-antigens to I-Ag7 molecules. Introducing a copy of a

www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259

different MHC II allele (I-Ak ) into the NOD background resulted in deletion of autoreactive thymocytes and protection from diabetes. As such, the I-Ag7 allele has a recessive effect on development of autoimmunity probably resulting from inefficient central tolerance [54]. In the case of NF-M15-35 , we have shown that pulsing BMDCs with antigen is sufficient to elicit sustained activation of 2D2 T cells with MOG35-55 but not with NF-M15-35 , revealing an intrinsic difference in the efficacy with which NF-M15-35 can be presented. Indeed, it was not expected that a poor immunogenic antigen would be a relevant target of the autoimmune response during EAE as previously shown in the 2D2 model [23] and in MOG-induced EAE in C57BL/6 mice [28]. Different mechanisms could explain these seemingly contradictory features of NF-M15-35 as a poor tolerogen and an important autoimmune target. It is likely that differences in antigen processing, relative abundance of peptide and MHC class II molecules, and presence of co-stimulatory molecules on steady-state thymic and peripheral APCs versus activated APCs participating in EAE could account for this discrepancy. However, further work is needed to elucidate the contribution of bi-specific T cells in different models of CNS autoimmunity. Specifically, in the context of MOG35-55 induced EAE, bi-specific T cells will be activated by the immunodominant MOG35-55 antigen and will likely be exposed to the subdominant NF-M15-35 antigen in the immunogenic context of an inflamed CNS. Since NF-M15-35 does not tolerize the bi-specific T cells, we reason that, in an inflammatory context, this antigen would now become a target of the cumulative autoimmune response. Interestingly, the MBP-induced EAE model revealing a link of suboptimal antigen-presentation to a tolerization defect, has served more recently for the development of antigen-specific immunotherapies. In this model, intranasaly or subcutaneous instillation of MBPAc1-9 K4Y stabilized self-antigen allowed skewing of the pathogenic T-cell response towards a tolerogenic phenotype resulting in a less severe disease [55, 56]. Antigen-specific immunotherapies receive increasing attention in the context of autoimmunity, allergy or cancer because they allow fine-tuning of the immune response of interest without jeopardizing the general immune function [55–59]. These new approaches stress the need for the identification of antigenic targets involved in disease development. In this context, it has to be kept in mind that TCR polyspecificity can alter the efficiency and safety of such treatment due to potential off-target antigens. For example, immunotherapy directed against MAGE-A3 antigen in cancer patients resulted in fatal cardiac toxicity due to recognition of a peptide derived from the protein Titin [60]. This clinical situation is reminiscent of the phenotype we observed following NF-M15-35 T20Y-immunization, where priming of the NF-M15-35 -specific repertoire resulted in EAE through MOG35-55 -targeting. Taken together, our results add to the current notion that polyspecificity is an inherent property of T cells, and that TCRs recognize a spectrum of different ligands with different affinities shaping the repertoire during both the tolerization and the effector phase.



Experimental Procedures Mice C57BL/6 mice were obtained from Charles River laboratories (L’Arbresle, France). RAG2−/− mice were obtained from CREFRE (Toulouse, France). C57BL/6 NF-M−/− mice [61] and MOG−/− [62] were intercrossed to obtain MOG−/− NF-M−/− double knockout mice. MOG-specific 2D2-TCR transgenic mice [34] were bred into the MOG−/− NF-M−/− strain. All mice were housed under specific pathogen-free conditions at the UMS-006 animal facility, which is accredited by the French Ministry of Agriculture to perform experiments on live mice in appliance to the French and European regulations on care and protection of the Laboratory Animals (EC Directive 2010/63). All experimental protocols were approved by the local ethics committee and are in compliance with European Union guidelines.

NF-M and MOG expression analysis Organs were snap-frozen in dry ice and reduced to suspension R in a Precellys tubes (Bertin Technologies). RNA was extracted R using Trizol (Thermofisher) and retro-transcribed into cDNA using the Blue script III kit (Invitrogen). Super Taq TP05aa enzyme (Enzyme Technologies Ltd) was used for PCR amplification. For NF-M transcript detection, a primer pair spanning nefm exons 1 to 3 was designed (5 -GTGCGAGGCACTAAGGAGTC3 and 5 -CCTCTTCTGCCTGGTCTGAC-3 ), amplifying a 633bp amplicon. The forward primer anneals to a sequence in exon 1 which is replaced by a neoR cassette in NF-M−/− mice [61]. HPRT mRNA expression was assessed using the following primers: 5 -TGACACTGGTAAAACAATGCAAACT-3 and 5 AACAAAGTCTGGCCTGTATCCAA-3 ; amplifying a 152bp fragment. PCR amplification products were then run on a 1.5% agarose gel. For MOG and NF-M transcript detection in purified thymic subpopulations, semi-quantitative RT-PCR was done using the following primers for MOG: 5- ACCTGCTTCTTCAGAGACCACT-3 and 5 -GGGGTTGACCCAATAGAAGG-3 . After the PCR run, the PCR products were run on a 2.3% agarose gel. Primers were designed to be intron spanning (spanned intron: 2769nt). Amplicon size was 84nt. Purification of thymic stromal cells and subsequent RTPCR were performed as described previously [7].

Peptides MOG35-55 (MEVGWYRSPFSRVVHLYRNGK), NF-M15-35 (RRVTETRSSFSRVSGSPSSGF) and OVA323-339 (ISQAVHAAHAEINEAGR) peptides with over 95% purity were purchased from Polypeptide Laboratories (San Diego, CA). NF-M15-35 T20Y (RRVTEYRSSFSRVSGSPSSGF) peptide was purchased from Genecust (Dudelange, Luxembourg).

www.eji-journal.eu

2255

2256

L. E. Lucca et al.

Flow cytometry R Cells were stained with viability dye (Fixable Viability eFluor 780 or 506, eBioscience) and CD16/CD32 Fc-blocked (clone 2.4G2). MOG38-49 -I-Ab and CLIP-I-Ab tetramers were obtained from the NIH tetramer core facility (Emory University, Atlanta, USA) and staining was performed for 2 h at room temperature at a concentration of 0.03 mg/mL. Surface staining was performed with the following antibodies from Becton-Dickinson and eBioscience specific for CD19 (1D3), CD8 (53-6.7), CD90.2 (53-2.1), CD4 (RM4-5 or GK1.5), CD44 (IM7), CD40L (MR1). For intra-cellular staining cells were permeabilized using Fixation/Permeabilization kit (Becton-Dickinson) and stained with anti-Foxp3 (FJK-16s) from ebioscience. Samples were acquired on a LSR II or Fortessa flow-cytometer. Cell sorting was performed with an FACS ARIA II. Data were analyzed with the FlowJo software version 7.6.5.

Immunization and EAE induction 8–12 week-old mice were immunized subcutaneously at the base of the tail with 100 μg of indicated peptides dissolved in 100 μL of PBS, emulsified in 100 μL CFA (Difco) supplemented with 500 μg M. tuberculosis H37RA (Difco). For active EAE, pertussis toxin (List Biological Laboratories, Campbell, CA) was injected i.v. at days 0 and 2 (400 and 200 ng respectively). Alternatively, total T cells from unimmunized mice of different genotypes were magnetically sorted by negative selection using anti-B220 (RA3-6-B2), anti-CD11b (M1/70), anti-MHC II (m5-114) et anti-FcR (2.4G2) antibodies followed by incubation with Dynabeads (Dynal), and 2 × 106 purified T cells were injected i.v. into 8–12 week-old RAG2−/− recipients, in which active EAE was induced 24 h later. Clinical scores were recorded daily on the following scale: 0: no clinical signs, 0.5: partially limp tail, 1: complete tail paralysis, 2: loss in coordinated movement, hind limb paresis, 2.5: one hind limb paralyzed, 3: both hind limbs paralyzed, 4: forelimbs paralyzed, 5: moribund state. For Kaplan-Meier survival curves, animals were considered sick when they displayed score 1 or above for at least 2 consecutive days.

Passive EAE Nine days after MOG35-55 -immunization, spleen and draining lymph nodes from a WT or MOG−/− NF-M−/− mice were harvested and single-cell suspensions were restimulated in vitro with 10 μg/mL MOG35-55 , in the presence of IL-23 (10 ng/mL, R&D) and anti-IFNγ mAb (clone XMG1.2, 10 μg/mL). After 72h, CD4+ T cells were magnetically sorted using Dynabeads untouched CD4 cells (Dynal) and 2 × 106 purified cells were transferred into lightly irradiated (3Gy) WT recipients.


Eur. J. Immunol. 2016. 46: 2247–2259

Culture and pulsing of bone marrow derived dendritic cells (BMDCs) Bone-marrow cells from C57BL/6 mice were isolated from tibia and femur and cultured in GM-CSF supplemented medium [63]. After 8 days BMDCs were incubated with 10 μg/mL of indicated antigens for 4h at 37°C and washed.

In vitro recall response and cytokine measurement CD4+ T cells were magnetically sorted by negative selection (Dynabeads Mouse untouched CD4 cells, Dynal). For proliferation, sorted CD4+ T cells were stained with CellTraceTM Violet (LifeTechnologies) and stimulated with antigen-pulsed BMDCs. For cytokine release, they were cultured with irradiated splenocytes in 96-well plates and the indicated peptides. Supernatants were collected after 72 h. IFN-γ and IL-17A concentrations in supernatants were measured by a DuoSet sandwich ELISA (R&D Systems, Minneapolis, MN). Plates were developed with a horseradish peroxidase colorimetric reaction and absorbance was measured at 450 and 540 nm using a Versamax spectrophotometer (MTX Lab Systems).

Antigen-induced CD40L up-regulation Magnetically sorted CD4+ T cells (Dynabeads untouched CD4 cells, Dynal) were rested overnight with IL-2 (1 ng/mL) and then cultured with irradiated splenocytes together with the indicated peptides. After 1 h, 2 μg/mL of anti-CD40L mAb (MR1) was added for 5 h. Cells were washed and stained for flow cytometry acquisition as described [38].

TCR repertoire and clonotype analysis R RNA was extracted from sorted T cells using RNeasy kit (Qiagen) and retro-transcribed into cDNA using SuperScriptTM II enzyme (Invitrogen) according to manufacturers’ instructions. Vβ usage was performed by quantitative PCR using Cβ antisense and 24 different Vβ sense primers as described [64]. In parallel, R Taq DNA polymerase the cDNA was amplified with Platinium (Invitrogen) using Vβ8.2 sense and Cβ antisense primers. The PCR products were purified using Qiaquick PCR purification kit (Qiagen) and used as DNA matrix for quantitative PCR using a Cβ probe (5 (6FAM)AAATGTGACTCCACCCAAGGTCTCCTTGTT (TAM)3 ), the Cβ antisense primer and either the Vβ8.2 sense primer or a Vβ clonotypic consensus sense primer (5 GGTGARACTGGGGNAACTA3 ). Percentage of clonotype was then estimated as 2-(DCt consensus-DCt Vβ8.2).

www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259

Statistical analysis


9 Fairchild, P. J., Wildgoose, R., Atherton, E., Webb, S. and Wraith, D. C., An autoantigenic T cell epitope forms unstable complexes with class II

Statistical analyses were performed using PRISM 6 (Graphpad software Inc.). ****: p< 0.0001, ***: p < 0.001, **: p < 0.01, *: p < 0.05.

MHC: a novel route for escape from tolerance induction. Int. Immunol. 1993. 5: 1151–1158. 10 Wucherpfennig, K. W., Allen, P. M., Celada, F., Cohen, I. R., De Boer, R., Garcia, K. C., Goldstein, B. et al., Polyspecificity of T cell and B cell receptor recognition. Semin. Immunol. 2007. 19: 216–224. 11 Sloan-Lancaster, J. and Allen, P. M., Altered peptide ligand-induced partial T cell activation: molecular mechanisms and role in T cell biology. Annu. Rev. Immunol. 1996. 14: 1–27. 12 Mason, D., A very high level of crossreactivity is an essential feature of

Acknowledgements: The authors would like to thank Drs. A. Saoudi and A. Dejean for their input on the project and critical reading of the manuscript, the staff of the UMS006 animal facility for care of the mice, J. Sidney and A. Sette for evaluation of peptide:I-Ab affinity, F. L’Faqihi and V. Duplan for running the core flow cytometry facility, the NIH tetramer core facility for providing the MOG38-49 -I-Ab tetramer and Fabian Brunk and Sheena Pinto, DKFZ, Heidelberg for performing MOG specific PCR. This work was supported by the Institut National de la Santé et de la Recherche Médicale, the Centre National de la Recherche Scientifique, the Midi-Pyrénées Région, the French MS society (ARSEP), and the Medical Research Foundation (FRM, DEQ20090515409) to R.S.L., and the Italian Federation for Multiple Sclerosis (FISM, 2012/B/6) to L.E.L.

the T-cell receptor. Immunol. Today 1998. 19: 395–404. 13 Sewell, A. K., Why must T cells be cross-reactive? Nat. Rev. Immunol. 2012. 12: 669–677. 14 Birnbaum, M. E., Mendoza, J. L., Sethi, D. K., Dong, S., Glanville, J., Dobbins, J., Ozkan, E. et al., Deconstructing the peptide-MHC specificity of T cell recognition. Cell 2014. 157: 1073–1087. 15 Damian, R. T., Molecular mimicry in biological adaptation. Science 1965. 147: 824. 16 Fujinami, R. S. and Oldstone, M. B., Amino acid homology between the encephalitogenic site of myelin basic protein and virus: mechanism for autoimmunity. Science 1985. 230: 1043–1045. 17 Wucherpfennig, K. W., Sette, A., Southwood, S., Oseroff, C., Matsui, M., Strominger, J. L. and Hafler, D. A., Structural requirements for binding of an immunodominant myelin basic protein peptide to DR2 isotypes and for its recognition by human T cell clones. J. Exp. Med. 1994. 179: 279–290. 18 Harkiolaki, M., Holmes, S. L., Svendsen, P., Gregersen, J. W., Jensen, L. T., McMahon, R., Friese, M. A. et al., T cell-mediated autoimmune

Conflict of interest: The authors declare no financial or commercial conflict of interest.

disease due to low-affinity crossreactivity to common microbial peptides. Immunity 2009. 30: 348–357. 19 Gautam, A. M., Liblau, R., Chelvanayagam, G., Steinman, L. and Boston, T., A viral peptide with limited homology to a self peptide can induce clin-

References

ical signs of experimental autoimmune encephalomyelitis. J. Immunol. 1998. 161: 60–64. 20 Albert, L. J. and Inman, R. D., Molecular mimicry and autoimmunity. N.

1 Anderton, S. M. and Wraith, D. C., Selection and fine-tuning of the autoimmune T-cell repertoire. Nat. Rev. Immunol. 2002. 2: 487–498.

Engl. J. Med. 1999. 341: 2068–2074. 21 Anderson, A. C., Nicholson, L. B., Legge, K. L., Turchin, V., Zaghouani,

2 Anderson, M. S., Venanzi, E. S., Klein, L., Chen, Z., Berzins, S. P., Turley,

H. and Kuchroo, V. K., High frequency of autoreactive myelin proteolipid

S. J., von Boehmer, H. et al., Projection of an immunological self shadow

protein-specific T cells in the periphery of naive mice: mechanisms of

within the thymus by the aire protein. Science 2002. 298: 1395–1401.

selection of the self-reactive repertoire. J. Exp. Med. 2000. 191: 761–770.

3 Derbinski, J., Pinto, S., Rosch, S., Hexel, K. and Kyewski, B., Promiscu-

22 Avery, A. C., Zhao, Z. S., Rodriguez, A., Bikoff, E. K., Soheilian, M., Foster,

ous gene expression patterns in single medullary thymic epithelial cells

C. S. and Cantor, H., Resistance to herpes stromal keratitis conferred by

argue for a stochastic mechanism. Proc. Natl. Acad. Sci. U. S. A. 2008. 105:

an IgG2a-derived peptide. Nature 1995. 376: 431–434.

657–662. 23 Krishnamoorthy, G., Saxena, A., Mars, L. T., Domingues, H. S., Mentele, 4 Nedjic, J., Aichinger, M., Emmerich, J., Mizushima, N. and Klein, L.,

R., Ben-Nun, A., Lassmann, H. et al., Myelin-specific T cells also recognize

Autophagy in thymic epithelium shapes the T-cell repertoire and is

neuronal autoantigen in a transgenic mouse model of multiple sclerosis.

essential for tolerance. Nature 2008. 455: 396-400. 5 Koble, C. and Kyewski, B., The thymic medulla: a unique microenvironment for intercellular self-antigen transfer. J. Exp. Med. 2009. 206: 1505-1513.

Nat. Med. 2009. 15: 626–632. 24 Liblau, R. S., Wekerle, H. and Tisch, R. M., Cumulative autoimmunity: T cell clones recognizing several self-epitopes exhibit enhanced pathogenicity. Front Immunol. 2011. 2: 1–6.

6 Klein, L., Kyewski, B., Allen, P. M. and Hogquist, K. A., Positive and neg-

25 Ben-Nun, A., Kerlero de Rosbo, N., Kaushansky, N., Eisenstein, M.,

ative selection of the T cell repertoire: what thymocytes see (and don’t

Cohen, L., Kaye, J. F. and Mendel, I., Anatomy of T cell autoimmunity

see). Nat. Rev. Immunol. 2014. 14: 377–391. 7 Klein, L., Klugmann, M., Nave, K. A., Tuohy, V. K. and Kyewski, B., Shaping of the autoreactive T-cell repertoire by a splice variant of self protein expressed in thymic epithelial cells. Nat. Med. 2000. 6: 56–61.

to myelin oligodendrocyte glycoprotein (MOG): prime role of MOG44F in selection and control of MOG-reactive T cells in H-2b mice. Eur. J. Immunol. 2006. 36: 478–493. 26 Petersen, T. R., Bettelli, E., Sidney, J., Sette, A., Kuchroo, V. and Back-

8 Huseby, E. S., Sather, B., Huseby, P. G. and Goverman, J., Age-dependent

strom, B. T., Characterization of MHC- and TCR-binding residues of the

T cell tolerance and autoimmunity to myelin basic protein. Immunity

myelin oligodendrocyte glycoprotein 38-51 peptide. Eur. J. Immunol. 2004.

2001. 14: 471–481.

34: 165–173.


www.eji-journal.eu

2257

2258

L. E. Lucca et al.

27 Lucca, L. E., Desbois, S., Ramadan, A., Ben-Nun, A., Eisenstein, M., Carrie,

Eur. J. Immunol. 2016. 46: 2247–2259

42 Bruno, R., Sabater, L., Sospedra, M., Ferrer-Francesch, X., Escudero,

N., Guery, J. C. et al., Bispecificity for myelin and neuronal self-antigens

D., Martinez-Caceres, E. and Pujol-Borrell, R., Multiple sclerosis candi-

is a common feature of CD4 T cells in C57BL/6 mice. J. Immunol. 2014. 193:

date autoantigens except myelin oligodendrocyte glycoprotein are tran-

3267–3277.

scribed in human thymus. Eur. J. Immunol. 2002. 32: 2737–2747.

28 Ramadan, A., Lucca, L. E., Carrie, ´ N., Desbois, S., Axisa, P. P., Hayder, M.,

43 Malhotra, D., Linehan, J. L., Dileepan, T., Lee, Y. J., Purtha, W. E., Lu, J. V.,

Bauer, J. et al., In situ expansion of T cells that recognize distinct self-

Nelson, R. W. et al., Tolerance is established in polyclonal CD4(+) T cells

antigens sustains autoimmunity in the central nervous system. Brain

by distinct mechanisms, according to self-peptide expression patterns.

2016. 139: 1433–1346.

Nat. Immunol. 2016. 17: 187–195.

29 Fazilleau, N., Delarasse, C., Sweenie, C. H., Anderton, S. M., Fillatreau,

44 Mendel, I., Kerlero de Rosbo, N. and Ben-Nun, A., A myelin oligodendro-

S., Lemonnier, F. A., Pham-Dinh, D. et al., Persistence of autoreactive

cyte glycoprotein peptide induces typical chronic experimental autoim-

myelin oligodendrocyte glycoprotein (MOG)-specific T cell repertoires in

mune encephalomyelitis in H-2b mice: fine specificity and T cell receptor

MOG-expressing mice. Eur. J. Immunol. 2006. 36: 533–543.

V beta expression of encephalitogenic T cells. Eur. J. Immunol. 1995. 25:

30 Delarasse, C., Daubas, P., Mars, L. T., Vizler, C., Litzenburger, T., Igle-

1951–1959.

sias, A., Bauer, J. et al., Myelin/oligodendrocyte glycoprotein-deficient

45 Siatskas, C., Seach, N., Sun, G., Emerson-Webber, A., Silvain, A., Toh, B.

(MOG-deficient) mice reveal lack of immune tolerance to MOG in wild-

H., Alderuccio, F. et al., Thymic gene transfer of myelin oligodendrocyte

type mice. J. Clin. Invest. 2003. 112: 544–553.

glycoprotein ameliorates the onset but not the progression of autoim-

31 Linares, D., Mana, P., Goodyear, M., Chow, A. M., Clavarino, C., Hunting-

mune demyelination. Mol. Ther. 2012. 20: 1349-1359.

ton, N. D., Barnett, L. et al., The magnitude and encephalogenic potential

46 Bailey-Bucktrout, S. L., Martinez-Llordella, M., Zhou, X., Anthony, B.,

of autoimmune response to MOG is enhanced in MOG deficient mice. J.

Rosenthal, W., Luche, H., Fehling, H. J. et al., Self-antigen-driven acti-

Autoimmun. 2003. 21: 339–351.

vation induces instability of regulatory T cells during an inflammatory

32 Akirav, E. M., Xu, Y. and Ruddle, N. H., Resident B cells regulate thymic expression of myelin oligodendrocyte glycoprotein. J. Neuroimmunol. 2011. 235: 33–39.

autoimmune response. Immunity 2013. 39: 949–962. 47 Oukka, M., Interplay between pathogenic Th17 and regulatory T cells. Ann. Rheum. Dis. 2007. 66(S3): iii87–iii90.

33 Derbinski, J., Schulte, A., Kyewski, B. and Klein, L., Promiscuous gene

48 Korn, T., Reddy, J., Gao, W., Bettelli, E., Awasthi, A., Petersen, T. R.,

expression in medullary thymic epithelial cells mirrors the peripheral

Backstrom, B. T. et al., Myelin-specific regulatory T cells accumulate in

self. Nat. Immunol. 2001. 2: 1032–1039.

the CNS but fail to control autoimmune inflammation. Nat. Med. 2007.

34 Bettelli, E., Pagany, M., Weiner, H. L., Linington, C., Sobel, R. A. and

13: 423–431.

Kuchroo, V. K., Myelin oligodendrocyte glycoprotein-specific T cell recep-

49 Liu, G. Y., Fairchild, P. J., Smith, R. M., Prowle, J. R., Kioussis, D. and

tor transgenic mice develop spontaneous autoimmune optic neuritis. J.

Wraith, D. C., Low avidity recognition of self-antigen by T cells permits

Exp. Med. 2003. 197: 1073–1081.

escape from central tolerance. Immunity 1995. 3: 407–415.

35 Altman, J. D., Moss, P. A., Goulder, P. J., Barouch, D. H., McHeyzer-

50 Zehn, D. and Bevan, M. J., T cells with low avidity for a tissue-restricted

Williams, M. G., Bell, J. I., McMichael, A. J. et al., Phenotypic analysis

antigen routinely evade central and peripheral tolerance and cause

of antigen-specific T lymphocytes. Science 1996. 274: 94–96.

autoimmunity. Immunity 2006. 25: 261–270.

36 Sabatino, J. J., Jr., Shires, J., Altman, J. D., Ford, M. L. and Evavold, B. D.,

51 Leech, M. D., Carrillo-Vico, A., Liblau, R. S. and Anderton, S. M., Recog-

Loss of IFN-gamma enables the expansion of autoreactive CD4+ T cells to

nition of a high affinity MHC class I-restricted epitope of myelin oligo-

induce experimental autoimmune encephalomyelitis by a nonencephal-

dendrocyte glycoprotein by CD8(+) T cells derived from autoantigen-

itogenic myelin variant antigen. J. Immunol. 2008. 180: 4451–4457.

deficient mice. Front Immunol. 2011. 2: 17.

37 Nelson, R. W., Beisang, D., Tubo, N. J., Dileepan, T., Wiesner, D. L.,

52 Pinto, S., Sommermeyer, D., Michel, C., Wilde, S., Schendel, D., Uckert,

Nielsen, K., Wuthrich, M. et al., T cell receptor cross-reactivity between

W., Blankenstein, T. et al., Misinitiation of intrathymic MART-1 tran-

similar foreign and self peptides influences naive cell population size

scription and biased TCR usage explain the high frequency of MART-1-

and autoimmunity. Immunity 2015. 42: 95–107. 38 Chattopadhyay, P. K., Yu, J. and Roederer, M., A live-cell assay to detect

specific T cells. Eur. J. Immunol. 2014. 44: 2811–2821. 53 Giraud, M., Jmari, N., Du, L., Carallis, F., Nieland, T. J., Perez-Campo, F.

antigen-specific CD4+ T cells with diverse cytokine profiles. Nat. Med.

M., Bensaude, O. et al., An RNAi screen for Aire cofactors reveals a role for

2005. 11: 1113–1117.

Hnrnpl in polymerase release and Aire-activated ectopic transcription.

39 Mendel Kerlero de Rosbo, N. and Ben-Nun, A., Delineation of the minimal encephalitogenic epitope within the immunodominant region of myelin oligodendrocyte glycoprotein: diverse V beta gene usage by T cells recognizing the core epitope encephalitogenic for T cell receptor V beta b and T cell receptor V beta a H-2b mice. Eur. J. Immunol. 1996. 26: 2470– 2479. 40 Mendel, I., Katz, A., Kozak, N., Ben-Nun, A. and Revel, M., Interleukin-6 functions in autoimmune encephalomyelitis: a study in gene-targeted mice. Eur. J. Immunol. 1998. 28: 1727–1737. 41 Sakuma, H., Park, I. K., Kohyama, K., Feng, D. and Matsumoto, Y., Quantitation of myelin oligodendrocyte glycoprotein and myelin basic protein in the thymus and central nervous system and its relationship to the clinicopathologic features of autoimmune encephalomyelitis. J. Neurosci. Res. 2006. 84: 606–613.


Proc. Natl. Acad. Sci. U. S. A. 2014. 111: 1491–1496. 54 Ridgway, W. M., Fasso, M. and Fathman, C. G., A new look at MHC and autoimmune disease. Science 1999. 284: 749–751. 55 Gabrysova, L., Nicolson, K. S., Streeter, H. B., Verhagen, J., SabatosPeyton, C. A., Morgan, D. J. and Wraith, D. C., Negative feedback control of the autoimmune response through antigen-induced differentiation of IL-10-secreting Th1 cells. J. Exp. Med. 2009. 206: 1755–1767. 56 Burton, B. R., Britton, G. J., Fang, H., Verhagen, J., Smithers, B., SabatosPeyton, C. A., Carney, L. J. et al., Sequential transcriptional changes dictate safe and effective antigen-specific immunotherapy. Nat. Commun. 2014. 5: 4741. 57 Meiler, F., Zumkehr, J., Klunker, S., Ruckert, B., Akdis, C. A. and Akdis, M., In vivo switch to IL-10-secreting T regulatory cells in high dose allergen exposure. J. Exp. Med. 2008. 205: 2887–2898.

www.eji-journal.eu

Eur. J. Immunol. 2016. 46: 2247–2259

58 Steinman, L., The re-emergence of antigen-specific tolerance as a potential therapy for MS. Mult. Scler. 2015. 21: 1223–1238. 59 Stauss, H. J., Morris, E. C. and Abken, H., Cancer gene therapy with T cell receptors and chimeric antigen receptors. Curr. Opin. Pharmacol. 2015. 24: 113–118. 60 Linette, G. P., Stadtmauer, E. A., Maus, M. V., Rapoport, A. P., Levine, B. L., Emery, L., Litzky, L. et al., Cardiovascular toxicity and titin crossreactivity of affinity-enhanced T cells in myeloma and melanoma. Blood 2013. 122: 863–871. 61 Jacomy, H., Zhu, Q., Couillard-Despres, S., Beaulieu, J. M. and Julien, J. P., Disruption of type IV intermediate filament network in mice lacking the neurofilament medium and heavy subunits. J. Neurochem. 1999. 73: 972–984. 62 Hovelmeyer, N., Hao, Z., Kranidioti, K., Kassiotis, G., Buch, T., Frommer, F., von Hoch, L. et al., Apoptosis of oligodendrocytes via Fas and TNF-R1 is a key event in the induction of experimental autoimmune encephalomyelitis. J. Immunol. 2005. 175: 5875–5884.



63 Inaba, K., Inaba, M., Romani, N., Aya, H., Deguchi, M., Ikehara, S., Muramatsu, S. et al., Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 1992. 176: 1693– 1702. 64 Serre, L., Fazilleau, N. and Guerder, S., Central tolerance spares the private high-avidity CD4(+) T-cell repertoire specific for an islet antigen in NOD mice. Eur. J. Immunol. 2015. 45: 1946–1956.

ˆ Full correspondence: Prof. Roland S. Liblau, INSERM-U1043, Hopital Purpan, BP 3028, 31024, Toulouse, France Fax: +33 5 62 74 45 58 e-mail: [email protected] Received: 15/3/2016 Revised: 9/5/2016 Accepted: 17/6/2016 Accepted article online: 23/6/2016

www.eji-journal.eu

2259