T Cell Repertoire Diversity Is Required for Relapses in Myelin Oligodendrocyte Glycoprotein-Induced Experimental Autoimmune Encephalomyelitis

T Cell Repertoire Diversity Is Required for Relapses in MyelinOligodendrocyte Glycoprotein-Induced ExperimentalAutoimmune Encephalomyelitis1

Nicolas Fazilleau,2,3* Cecile Delarasse,2† Iris Motta,† Simon Fillatreau,‡ Marie-Lise Gougeon,§

Philippe Kourilsky,*§ Danielle Pham-Dinh,¶ and Jean M. Kanellopoulos4†

Comparison of TCR�� repertoires of myelin oligodendrocyte glycoprotein (MOG)-specific T lymphocytes in C57BL/6 and TdT-deficient littermates (TdT�/�) generated during experimental autoimmune encephalomyelitis (EAE) highlights a link between adiversified TCR�� repertoire and EAE relapses. At the onset of the disease, the EAE-severity is identical in TdT�/� and TdT�/�

mice and the neuropathologic public MOG-specific T cell repertoires express closely similar public V�-J� and V�-J� rearrange-ments in both strains. However, whereas TdT�/� and TdT�/� mice undergo successive EAE relapses, TdT�/� mice recoverdefinitively and the lack of relapses does not stem from dominant regulatory mechanisms. During the first relapse of the diseasein TdT�/� mice, new public V�-J� and V�-J� rearrangements emerge that are distinct from those detected at the onset of thedisease. Most of these rearrangements contain N additions and are found in CNS-infiltrating T lymphocytes. Furthermore, CD4�

T splenocytes bearing these rearrangements proliferate to the immunodominant epitope of MOG and not to other immunodom-inant epitopes of proteolipid protein and myelin basic protein autoantigens, excluding epitope spreading to these myelin proteins.Thus, in addition to epitope spreading, a novel mechanism involving TCR�� repertoire diversification contributes to autoimmuneprogression. The Journal of Immunology, 2007, 178: 4865–4875.

E ncephalitogenic T cells play a major role in the develop-ment of autoimmune diseases including multiple sclerosisand its animal model, experimental autoimmune encephalo-

myelitis (EAE)5 (1). EAE is a disease of the CNS mediated by IL-17-producing effector CD4� T lymphocytes that acquire the Th17phenotype in the presence of TGF-� and IL-6 (reviewed in Ref. 2).EAE can be induced in genetically susceptible animals by immuni-zation with myelin proteins such as the myelin basic protein (MBP)(3), the proteolipid protein (PLP) (4), or the myelin oligodendrocyte

glycoprotein (MOG) (5). Notably, MOG is the only CNS myelin au-toantigen known to initiate both a demyelinating autoantibody re-sponse and an encephalitogenic T cell response in EAE. Using MOG-deficient C57BL/6 mice, it was recently shown that MOG, albeit avery minor CNS myelin protein, is a major self-Ag target (6).

Numerous studies have shown that epitope spreading is in-volved in the disease progression of T cell-mediated chronic au-toimmune diseases (7). In (SJL � B10.PL)F1 mice immunizedonly with the immunodominant (ID) peptide of MBP and sufferingfrom chronic EAE, determinant spreading to three subdominant(SD) peptides of MBP occurred (8). Immunization of SJL micewith the ID epitope of PLP also induced relapsing EAE. AlthoughCD4� T cells specific for the ID epitope of PLP persisted throughthe disease, T cells recognizing MBP and SD PLP epitopesemerged sequentially during the relapses (9, 10). Furthermore,tolerization with these epitopes decreased the incidence of re-lapses, suggesting that the spreading of the immune response tointramolecular and intermolecular determinants was required forthe disease progression (9–11). Epitope spreading seems to be ageneral mechanism aggravating autoimmune diseases, because itparticipates in the pathogenesis of autoimmune diabetes in NODmice (12, 13). However, epitope spreading is not an absolute re-quirement for EAE relapses (14, 15). Indeed, in Biozzi ABH miceMOG-induced EAE relapses were not abolished by tolerizationwith the well-defined spreading epitopes such as PLP 56–70 andMBP 12–26. In contrast, tolerization after the first peak of thedisease with the initiating MOG epitope resulted in protectionfrom relapses (16).

Unraveling the mechanism(s) involved in EAE relapses, otherthan epitope spreading, should be of high interest for understand-ing autoimmune pathology. Whether or not a diminished repertoireaffects EAE onset and EAE relapses is presently unknown. A diverserepertoire of heterodimeric �� TCRs for Ag is produced by the so-matic recombination of separate DNA segments, V and J for the

*Institut National de la Sante et de la Recherche Medicale, Unite 277, Institut Pasteur,Paris, France; †University of Paris Sud, Institut de Biochimie et Biophysique Mo-leculaire et Cellulaire, Unite Mixte de Recherche 8619, Centre National de la Re-cherche Scientifique, Orsay, France; ‡Immune Regulation Group, Deutsches Rheuma-Forschungszentrum, Berlin, Germany; §Antiviral Immunity, Biotherapy and VaccineUnit, Institut Pasteur, Paris, France; and ¶Unite Mixte de Recherche 546, InstitutNational de la Sante et de la Recherche Medicale, Paris, France

Received for publication November 23, 2005. Accepted for publication February 1, 2007.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by grants from the Institut National de la Sante et de laRecherche Medicale, Centre National de la Recherche Scientifique, the Pasteur In-stitute, l’Association de Recherche contre le Cancer (to N.F. and J.M.K.), Pasteur-Weizmann (to N.F.), Association de Recherche sur la Sclerose en Plaques (to C.D.,D.P.-D., and J.M.K.), the European Leukodystrophy Association (to D.P.-D.), andEuropean grant for AUTOROME STREP (to M.-L.G.).2 N.F. and C.D. have contributed equally to this work.3 Current address: Institute for Childhood and Neglected Diseases-120, The ScrippsResearch Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037.4 Address correspondence and reprint requests to Dr. Jean M. Kanellopoulos, UniteMixte de Recherche, 8619 Centre National de la Recherche Scientifique, UniversiteParis 11, Orsay, France. E-mail address: [email protected] Abbreviations used in this paper: EAE, experimental autoimmune encephalomyeli-tis; ID, immunodominant; LNC, lymph node cell; MBP, myelin basic protein; MOG,myelin oligodendrocyte glycoprotein; PLP, proteolipid protein; PPD, purified proteinderivative; SD, subdominant; wt, wild type; SI, stimulation index.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00

The Journal of Immunology

www.jimmunol.org

�-chain and V, D, and J for the �-chain, that yields the CDR3 of theTCR (17). During this process, TdT catalyzes the addition of tem-plate-independent nucleotides at the coding ends of V, (D), and J genesegments (18, 19). The repertoire of TdT-deficient mice (TdT�/�) ischaracterized by shorter CDR3s (20), and the �� TCR diversity isdiminished by a factor of 10–20 with no compensatory mechanismcounterbalancing this decrease (21). Nonetheless, TdT�/� mice arenot immunodeficient (22) and respond to viruses and protein Ags (22,23). Moreover, epitope immunodominance is unaltered in TdT�/�

animals (22).To delineate the potential role of TCR diversity in autoimmune

diseases, we compared the T cell repertoires of C57BL/6 TdT�/�

and TdT�/� littermates in the MOG-induced EAE model for sev-eral reasons: 1) this Th17-mediated disease is reproducible inC57BL/6 mice (2, 6, 24); 2) the autoantigen is well-defined com-pared with those involved in autoimmune-prone animals; 3) thereis one ID peptide in C57BL/6 mice (MOG 35–55); and 4) a singleMHC class II molecule, I-Ab, presents MOG 35–55.

We have previously established that MOG 35–55-specific Tlymphocytes express public rearrangements (25), i.e., common toall mice of the same MHC haplotype (26). These rearrangements,V�9-J�23, V�9-J�31, and V�8.2-J�2.1, have characteristicCDR3� and CDR3� sequences and are found in CNS-infiltratingcells during the disease and in MOG-stimulated lymph nodes cells(LNC) (25). In the CNS, these public MOG-specific T cells havean inflammatory phenotype, i.e., secreting IFN-� and TNF-� (25).

In this work, we show that TdT�/� and wild-type (wt) micereproducibly develop EAE after immunization with the recombi-nant MOG protein or MOG 35–55 peptide. However, whereasTdT�/� animals recover definitively after the first peak of EAE,TdT� littermates undergo successive EAE relapses. The lack ofEAE relapses in TdT�/� is not due to regulatory mechanisms pe-culiar to this strain. T cell repertoire analyses at the onset of thedisease show that the V� repertoire of MOG-specific T lympho-cytes in TdT�/� mice is identical with its wt counterpart while V�rearrangements are closely similar. Interestingly, in the first EAErelapse the public V� and V� rearrangements identified at the first

peak of the disease disappear in wt mice and new public MOG-specific V� and V� rearrangements, the majority of which con-taining N additions, emerge. Thus, in the absence of obviousepitope spreading, MOG EAE relapse is linked to the highly di-versified T cell repertoire of wt mice and results from the emer-gence of MOG-specific T cells bearing novel TCR rearrangements.

Materials and MethodsMice

All mice used in this study were 8 wk old and were obtained from PasteurInstitute (Paris, France) housing facilities. Mice were housed under con-ventional conditions. TdT�/�, TdT�/�, and TdT�/� littermates were usedin EAE and immunological experiments. Mice were used in accordance tothe Pasteur Institute’s guidelines.

FIGURE 1. Encephalomyelitis in TdT-deficient mice. C57BL/6TdT�/� and TdT�/� littermates (eight per group) were immunized witheither the rMOG protein (A) or the MOG 35–55 peptide (B). Results fromone representative experiment of two are shown for A and one of 10 for B.The same experiment was performed on TdT�/� littermates and the resultsare similar to those observed with TdT�/� mice (data not shown).

FIGURE 2. Encephalomyelitis in TdT-deficient mice previously in-jected with CD4� T splenocytes from TdT�/� and TdT�/� mice. A and B,CD4� T splenocytes from naive C57BL/6 CD45.1 animals were isolated asdescribed in Materials and Methods. CD4� CD45.1 T splenocytes (15 �106) were injected i.v. in nonirradiated TdT�/� CD45.2 recipient mice,EAE was induced in eight mice 14 days later, and the severity of thedisease was monitored daily (A). Forty-five days after the induction ofEAE, CNS-infiltrating cells were isolated by a Percoll gradient and ana-lyzed by FACS. Shown is a histogram of CD45.1 staining of CD4-positiveT cells for one representative mouse of five (B). C, CD4� T splenocytesfrom TdT�/� and TdT�/� mice animals were isolated as described in Ma-terials and Methods. CD4� T splenocytes (15 � 106) were injected i.v. intononirradiated TdT�/� mice, EAE was induced in eight mice 14 days later,and the severity of the disease was monitored daily.

4866 EAE-RELAPSES AS A CONSEQUENCE OF TCR�� DIVERSITY

Peptide and protein

The MOG 35–55 (MEVGWYRSPFSRVVHLYRNGK), PLP 178–191 (NTWTTCQSIAFPSK) (27), and MBP 54–72 (SHHAARTTHYGSLPQKSQR) (28) are ID peptides in C57BL/6 mice. The MOG 113–127 (LKVEDPFYWVSPGVL), MOG 120–134 (YWVSPGVLTLIALVP), MOG 183–197 (FVIVPVLGPLVALII) have been described as potential SD peptidesof MOG (6). All peptides were produced by NEOSYSTEM. Their puritywas tested by HPLC. Murine recombinant MOG 1–116 was prepared asdescribed in (6).

Induction and assessment of EAE

TdT�/�, TdT�/�, and TdT�/� littermates were immunized subcutaneouslywith 200 �g of rMOG 1–118 or 100 �g of murine MOG 35–55 in CFAcontaining 600 �g of Mycobacterium tuberculosis H37RA (Difco Labo-ratories) according to a previously described protocol (6). For EAE induc-tion, animals received additional i.v. injections of 200 ng of pertussis toxin(List Biological Laboratories) on days 0 and 2. Disease severity was mon-itored daily according to the following scale: 0, no disease; 1, flaccid tail;2, hind limb weakness; 3, hind limb paralysis; 4, forelimb weakness; and5, moribund.

Immunization and cell cultures

Mice were immunized in the hind footpads with 10 nmol of MOG 35–55peptide in CFA. Nine days later, LNC were cultured at 5 � 105 cells perwell with different concentrations of the MOG 35–55 peptide for 4 days.All cultures were done in HL-1 medium (Cambrex) supplemented withglutamine (2 mM). The cultures were then pulsed with 1 �Ci of [3H]thy-midine (ICN Biomedicals) for the last 18 h of the 4-day culture.

Isolation of CNS-infiltrating cells

Preparation of CNS-derived lymphocytes was performed using Percoll sep-aration. Briefly, murine CNS was dissected out and homogenized in 4.5 mlof RPMI 1640 medium (Invitrogen Life Technologies) with collagenase at400 U/ml and DNase at 100 U/ml (both from Sigma-Aldrich) for 45 minat 37°C and 5% CO2. The CNS homogenate was washed with RPMI 1640and centrifuged in a 40–80% Percoll gradient. Lymphocytes were isolatedand washed three times with RPMI 1640.

CFSE labeling, cell transfer, Abs, and cell sorting

EAE-induced TdT�/� mice were sacrificed 45 days after EAE inductionand cells were collected from the CNS and the spleen. The splenocyteswere labeled with 5 �M CFSE. Cells were washed with ice-cold PBS andtheir proliferation was tested as described above using the MOG 35–55peptide, another peptide of MOG, or the ID peptide of MBP or PLP. Usinga FITC-CD4 mAb from BD Pharmingen, stimulated CFSElow CD4�

splenocytes were then sorted out on a FACStarPlus (BD Biosciences) at theflow cytometry unit of the Pasteur Institute. Cell purity after sorting wasanalyzed by flow cytometry and was �96% in all samples.

CD4� T cell splenocytes from naive TdT�/� CD45.1 mice were isolatedusing beads from Miltenyi Biotec. CD4�CD45.1� T cells (15 � 106) wereinjected i.v. into nonirradiated TdT�/� CD45.2 recipient mice. Fourteen

days later, EAE was induced using MOG 35–55 peptide as describedabove. Mice were sacrificed 45 days later and cells were collected from theCNS and the spleen. FACS analyses of CNS-infiltrating cells were per-formed using FITC-CD45.1 and PE-CD4 mAbs from BD Pharmingen.Splenocytes were labeled with 5 �M CFSE and their proliferation wastested as described above.

Immunoscope analysis, cloning, and sequencing

Total RNA from splenocytes, LNC, or CNS-infiltrating cells was extractedusing an RNeasy mini kit from Qiagen and reverse transcribed into cDNAusing oligo(dT) and SuperScript II (Invitrogen Life Technologies). Immu-noscope analyses were performed as described elsewhere (29). Sequencingof CDR3 sequences was done using a TOPO Blunt cloning kit (InvitrogenLife Technologies). Briefly, a PCR amplification was performed on clonedbacteria followed by a second step of elongation using an ABI PRISMBigDye Terminator kit (Applied Biosystems). Reaction mixtures were thenanalyzed on a 48-capillary 3730 DNA Analyzer (Applied Biosystems).

FIGURE 3. Encephalomyelitis in TdT�/� and TdT�/� mice after adop-tive transfer. EAE was induced in both TdT�/� and TdT�/� littermates.Total splenocytes of some of these animals were recovered 30 days afterthe induction of EAE. These cells were depleted of CD11c� and B220�

cells. Then, 15 � 106 B220� CD11c� splenocytes were injected in non-irradiated animals suffering form EAE (day 30). The mean of three differ-ent experiments (at least five mice per group) is presented.

FIGURE 4. CDR3 size distribution among MOG 35–55-stimulatedLNC and CNS-infiltrating T cells during EAE in TdT-deficient mice. A,TdT�/� and wt mice were immunized in the hind footpads with 10 nmolof MOG 35–55 peptide in CFA. Nine days later, LNC were cultured withincreasing concentrations (�M) of peptide or with 10 �g/ml PPD for 4days. The cultures were then pulsed with 1 �Ci of [3H]thymidine for thelast 8 h. Data represent the mean of proliferative responses for fiveC57BL/6 mice and five TdT�/� mice. SI Stimulation index (SI � [3H]thy-midine cpm incorporated in lymphocytes stimulated by MOG or PPD/[3H]thymidine cpm incorporated in unstimulated cells). B–D, TdT�/� micewere immunized in the hind footpads with 10 nmol of MOG 35–55 peptidein CFA. Nine days later, LNC were cultured with either 30 �M peptide or20 �g/ml PPD for 4 days. In addition, EAE was induced in TdT�/� micewith MOG 35–55 peptide in CFA and pertussis toxin. Fifteen days later,mice were killed and CNS-infiltrating cells were recovered by a Percollgradient as described in Materials and Methods. cDNA from MOG stim-ulated-LNC (B), PPD-recalled LNC (C), and CNS-infiltrating cells (D) wassubjected to immunoscope analyses. Data represent CDR3� and � sizedistributions for two representative TdT�/� mice. Rearrangements shownare V�8.2-J�2.1 and V�9-C�.

4867The Journal of Immunology

ResultsEAE relapses occur in wt mice only

We have induced EAE in TdT�/�, TdT�/�, and TdT�/� litter-mates by immunization with rMOG or the MOG ID peptide 35–55.Approximately 85% of the animals developed EAE in 17 daysfollowing immunization (88.3% of incidence and mean day onsetof 17.12 � 2.27 in wt mice, 86.95% of incidence and mean dayonset of 17.34 � 1.76 in TdT�/� mice). We established a meanmaximal clinical score of 3.21 � 1.01 and 3.36 � 1.15 in wt andTdT�/� animals, respectively. Importantly, whereas all mice suf-fered from EAE, relapses occurred only in TdT�/� (Fig. 1) andTdT�/� animals (data not shown). Similar EAE-profiles have beenobtained twice with rMOG (Fig. 1A) and 10 times with MOG35–55 (Fig. 1B). Relapses never occurred in TdT�/� mice.

Absence of EAE relapses is not due to regulatory mechanismsspecific of the TdT�/� strain

The lack of relapse in TdT�/� animals may be due to dominantregulatory mechanisms such as regulatory T cells. Alternatively, itcould reflect holes in the T cell repertoires of TdT�/� mice. Todetermine whether dominant regulatory mechanisms prevent re-lapse in TdT�/� mice, we transferred CD4� T lymphocytes fromnaive C57BL/6 animals bearing CD45.1 as a marker for graftedcells into TdT�/� mice. After 2 wk, EAE was induced in theseanimals. As shown in Fig. 2, all mice developed the disease and

presented a relapse. At the peak of the relapse, five animals weresacrificed and CNS-infiltrating cells were analyzed by flow cytom-etry. The vast majority of CNS-infiltrating CD4� cells bearCD45.1 (mean of 80.71 � 3.14%), indicating that they are of do-nor origin (Fig. 2). Therefore, the lack of relapse in TdT�/� miceis recessive and can be complemented by the transfer of wt CD4�

T cells.However, one could argue that regulatory T cells need to be

amplified in TdT�/� mice during the course of the disease to in-hibit the relapses. Thus, we collected the splenocytes fromTdT�/�, TdT�/�, and TdT�/� littermates 30 days after the induc-tion of EAE. B220� and CD11c� cells were eliminated and theremaining cells were injected in nonirradiated TdT�/�, TdT�/�,and TdT�/� animals suffering from EAE (day 30). The cells fromTdT�/� mice did not inhibit the relapses in either TdT�/� andTdT�/� animals (Fig. 3).

In summary, the absence of EAE relapses cannot be attributed toT cell-mediated regulatory mechanisms specific to the TdT�/�

strain. We thus investigated whether relapses were due to the di-minished T cell repertoire of TdT�/� animals.

T cell repertoires against the ID peptide of the murine MOG inTdT�/� mice

LNC derived from rMOG or MOG 35–55-immunized TdT�/�,TdT�/� and TdT�/� littermates proliferated equally well upon

Table I. CDR3� sequences from MOG 35–55-specific T LNC (1–5) and from CNS-infiltrating cells during EAE (6 –10) in TdT�/� mice (V�8.2-J�2.1 rearrangement)

CDR3Length 3� V� End P/D� 5� J� Beginning

Deduced Amino AcidSequences a

LNC 1 8 tgtgccagcggtg ggactgggg atgctgagcagttcttcgga GGTGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

LNC 2 8 tgtgccagcggtgatg cagggg atgctgagcagttcttcgga GDAGDAEQ10 tgtgccagcggtgatg caggactgggg tatgctgagcagttcttcgga GDAGLGYAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

LNC 3 8 tgtgccagcg ccgggactgggg atgctgagcagttcttcgga AGTGDAEQ8 tgtgccagcggtgatg ctgggg atgctgagcagttcttcgga GDAGDAEQ

10 tgtgccagcggtgatg cactggggg actatgctgagcagttcttcgga GDALGDYAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

LNC 4 8 tgtgccagcg ccgggactgggg atgctgagcagttcttcgga AGTGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ10 tgtgccagcggtgatg cactggggggg tatgctgagcagttcttcgga GDALGGYAEQ

LNC 5 8 tgtgccagcggtg ggactgggg atgctgagcagttcttcgga GGTGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

CNS 6 8 tgtgccagcggtg ggacagggg atgctgagcagttcttcgga GGTGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

CNS 7 8 tgtgccagcggtgatg ctgggg atgctgagcagttcttcgga GDAGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ10 tgtgccagcggtgatg cactggat aactatgctgagcagttcttcgga GDALDNYAEQ

CNS 8 8 tgtgccagcg ccgggactgggg atgctgagcagttcttcgga AGTGDAEQ8 tgtgccagcggt ggacagggg atgctgagcagttcttcgga GGTGDAEQ

10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

CNS 9 8 tgtgccagcg ccgggactgggg atgctgagcagttcttcgga AGTGDAEQ10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ

CNS 10 10 tgtgccagcggtga gactgggggg aactatgctgagcagttcttcgga GETGGNYAEQ10 tgtgccagcggtgatg cactggggg actatgctgagcagttcttcgga GDALGDYAEQ

a Boldface letters indicate the 10-aa-long GETGGNYAEQ CDR3� sequence that is found in all TdT�/� animals, in agreement with the public CDR3� previously identifiedin C57BL/6 mice.


challenge with MOG 35–55 or with a purified protein derivative(PPD) (Fig. 4A).

We have previously determined that C57BL/6 mice express V�and V� public repertoires in response to the MOG 35–55 (25). Wethen analyzed the repertoire of TdT�/� MOG-specific T lympho-cytes using the immunoscope method (29). LNC from TdT�/�

mice immunized with MOG 35–55 were recalled in vitro withMOG 35–55 or PPD and collected. Their RNAs were extractedand reverse transcribed into cDNA. Aliquots were amplified byPCR with the 24 V�- and C�-specific primers or the 21 V� andC�-specific primers. A single peak corresponding to a CDR3� of8 aa was amplified with V�9-C� specific primers in all TdT�/�

animals (Fig. 4B). For the V�, only the V�8.2 segment was am-plified in all mice and MOG 35–55-stimulated T cells used pref-erentially the V�8.2-J�2.1 rearrangement with CDR3�s of 8 and10 aa (Fig. 4B). Immunoscope analyses performed on CFA-stim-ulated LNC recalled in vitro with PPD revealed bell-shaped curvesfor V�8.2-J�2.1 and V�9-C� rearrangements characteristic ofpolyclonal repertoires without oligoclonal expansions (Fig. 4C).Therefore, the V�8.2-J�2.1 and V�9-C� amplifications reflect theexpansion of MOG-specific T cells rather than cells reactive toAgs of the CFA. These findings were corroborated by immuno-scope analyses of CNS-infiltrating cells from five differentTdT�/� mice on day 18 after immunization with MOG 35–55.Oligoclonal expansions are observed for V�9-C� and V�8.2-J�2.1 rearrangements with CDR3 lengths identical with thosefound in LNC (Fig. 4D).

To further define the recurrent CDR3 identified in TdT�/� mice,the V�9-C� and V�8.2-J�2.1 PCR products were cloned and se-

quenced. A CDR3� of 10 aa with the GETGGNYAEQ sequenceis found in all TdT�/� animals (Table I, boldface letters), in agree-ment with the public CDR3� previously identified in C57BL/6mice (25). In C57BL/6, eight different nucleotide sequences en-code this public CDR3� among which seven contain N diversity(25). One sequence lacking N addition was found in both strains.Shorter CDR3� sequences are also found in LNC and cells infil-trating the CNS of TdT�/� mice: GGTGDAEQ (two of five forLNC and two of five for CNS-infiltrating cells), GDAGDAEQ(two of five and one of five) and AGTGDAEQ (two of five andtwo of five). These shorter CDR3� sequences are characteristic ofTdT�/� rearrangements as reported for naive T lymphocytes (21,30) and T cells specific for various protein epitopes (22, 23, 31).

As previously observed for wt mice, two rearrangements, V�9-J�23 (10 animals of 10) and V�9-J�31 (7 of 10 animals), arepreferentially used in TdT�/� animals (Table II). Three differentCDR3� sequences are encoded by the V�9-J�23. The SNYNQGKL sequence is found in 7 animals of 10 (Table II, boldfaceletters) and the SAYNQGKL sequence is found in four mice(TdT�/� mice nos. 1, 4, 7, and 10) while the SDYNQGKL se-quence is observed in mice nos. 1, 2, and 5. Thus, SxYNQGKL isthe V�9-J�23 CDR3� consensus sequence. For the V�9-J�31combination, two sequences are observed: SRNSNNRI in fourmice (Table II, italic letters) and SANSNNRI in TdT�/� mice nos.1 and 3.

In summary, MOG-reactive T cells from C57BL/6 and TdT�/�

mice carry the same public CDR3� sequence (GETGGNYAEQ)and their repertoires for the TCR�-chain are very similar. Indeed,the V�9-J�31 CDR3� sequence SRNSNNRI, which is public in

Table II. CDR3� sequences from MOG 35–55-specific T LNC (1–5) and CNS-infiltrating lymphocytes during EAE (6–10) in TdT�/� animals

V� J� 3� V� End P 5� J� BeginningDeduced Amino Acid

Sequencesa

LNC 1 9 23 tgtgctgtgagcg c ttataaccaggggaagcttatctttgga SAYNQGKL9 23 tgtgctgtgagcg attataaccaggggaagcttatctttgga SDYNQGKL9 31 tgtgctgtgagc c ggaatagcaataacagaatcttctttggt SRNSNNRI9 31 tgtgctgtgagcg c gaatagcaataacagaatcttctttggt SANSNNRI

LNC 2 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 23 tgtgctgtgagcg attataaccaggggaagcttatctttgga SDYNQGKL9 31 tgtgctgtgagc c ggaatagcaataacagaatcttctttggt SRNSNNRI

LNC 3 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 31 tgtgctgtgagcg ggaatagcaataacagaatcttctttggt SGNSNNRI9 31 tgtgctgtgagcg c gaatagcaataacagaatcttctttggt SANSNNRI

LNC 4 9 23 tgtgctgtgag gaattataaccaggggaagcttatctttgga RNYNQGKL9 23 tgtgctgtgagcg c ttataaccaggggaagcttatctttgga SAYNQGKL

LNC 5 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 23 tgtgctgtgagcg attataaccaggggaagcttatctttgga SDYNQGKL9 31 tgtgctgtgagc c ggaatagcaataacagaatcttctttggt SRNSNNRI

CNS 6 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 23 tgtgctgtgag gaattataaccaggggaagcttatctttgga RNYNQGKL

CNS 7 9 23 tgtgctgtgagcg c ttataaccaggggaagcttatctttgga SAYNQGKL9 31 tgtgctgtgagcg ggaatagcaataacagaatcttctttggt SGNSNNRI

CNS 8 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL

CNS 9 9 23 tgtgctgtgag gaattataaccaggggaagcttatctttgga RNYNQGKL9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 31 tgtgctgtgagc c ggaatagcaataacagaatcttctttggt SRNSNNRI

CNS 10 9 23 tgtgctgtgagc aattataaccaggggaagcttatctttgga SNYNQGKL9 23 tgtgctgtgagcg c Ttataaccaggggaagcttatctttgga SAYNQGKL

a The SNYNQGKL sequence, set in boldface, is found in seven animals of 10.


wt mice, was found in 4 of 10 TdT�/� animals. The public V�9-J�23 rearrangement generating the CDR3� (RSYNQGKL) in wtmice is replaced in TdT�/� animals by CDR3� sharing the con-sensus sequence SxYNQGKL. Altogether, these data show thatTdT-deficiency does not impair the development of the encepha-litogenic MOG-specific public T cell repertoires, consistent withthe similar disease course observed in the two strains. Thus, TdTdeficiency does not impact on the first stage of the disease.

T cell repertoires in first EAE relapse in wt mice

We have excluded the possibility that the lack of relapses inTdT�/� mice was due to an increase in T-regulatory cells or to adistinctive first phase of the disease. Altogether, our results suggestthat it may reflect the limited T cell repertoire of TdT�/� mice,resulting in an inability to activate an encephalitogenic TCRrepertoire specifically responsible for the relapses. Should that be

Table III. CDR3� sequences from CNS-infiltrating cells during EAE-relapse in C57BL/6 TdT�/� mice (V�8.3-C� rearrangement)

J�CDR3Length 3� V� End N/P/D�a 5� J� Beginning

Deduced Amino AcidSequences

CNS 11 1.1 8 tgtgccagcagtgatg gtcagggg acagaagtcttctttggt SDGQGTEV1.1 8 tgtgccagcagtga ccgcagggg cacagaagtcttctttggt SDRRGTEV2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtga cggctggggggtg gcagaaacgctgtattttggc SDGWGVAETL

CNS 12 1.1 8 tgtgccagcagt ttcagggag aacacagaagtcttctttggt SFRENTEV1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV2.3 10 tgtgccagcagtga aggggggggc agtgcagaaacgctgtattttggc SEGGGSAETL2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL

CNS 13 1.1 8 tgtgccagcagtgatg ggggcgg cacagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt2.3 10 tgtgccagcagtga cggctggggggtg gcagaaacgctgtattttggc SDGWGVAETL2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtgatg tactgggt agtgcagaaacgctgtattttggc SDVLGSAETL

CNS 14 1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL

CNS 15 1.1 8 tgtgccagcagtgatg gtcagggg acagaagtcttctttggt SDGQGTEV2.3 10 tgtgccagcagtgatg tactgggt agtgcagaaacgctgtattttggc SDVLGSAETL

a Underlined nucleotides are due by TdT activity.

Table IV. CDR3� sequences of CNS-infiltrating T cells from EAE-induced C57BL/6 TdT�/� mice during EAE relapse (V�4-C� rearrangement)

V� J� 3� V� End P/Na 5� J� BeginningDeduced Amino Acid

Sequencesb

CNS 11 4 18 tgtgctgctgag agg ggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctgagg gt ggttcagccttagggaggctgcattttgga EGGSALGRL4 18 tgtgctgctgagg at ggttcagccttagggaggctgcattttgga EDGSALGRL4 42 tgtgctgctgagg ca ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL

CNS 12 4 18 tgtgctgctgag agaggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 42 tgtgctgctgagg cg ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg cgagg ggaagcaatgcaaagctaaccttcggg EARGSNAKL

CNS 13 4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 42 tgtgctgctgagg cgcgg ggaagcaatgcaaagctaaccttcggg EARGSNAKL4 42 tgtgctgctgagg ca ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL

CNS 14 4 18 tgtgctgctgag agg ggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg ca ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL

CNS 15 4 18 tgtgctgctgagg at ggttcagccttagggaggctgcattttgga EDGSALGRL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg cg ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL

a Underlined nucleotides are due to TdT activity.b Three 9-aa-long public CDR3� sequences are set in special type. ERGSALGRL, which is encoded by two different nucleotide sequences containing or not containing N

additions, is set in boldface type and italics. The DRGSALGRL encoding sequence has no N additions and is set in italics with underlining. EAGGSNAKL is produced by twodifferent sequences containing N or P additions and is set in boldface type only.


the case, the repertoire involved in the relapse in C57BL/6 miceshould be different from the one identified in the first disease phaseand should display some clear features of TdT dependency. Todetermine the repertoire responsible for EAE relapses, we inducedEAE with MOG 35–55 and collected the splenocytes and CNS-infiltrating cells from TdT�/� mice at the maximum disease se-verity of the relapse. In all animals, CNS-infiltrating cells exhibitoligoclonal expansions using V�8.3 and J�1.1 or J�2.3 segments(CDR3� lengths of 8 and 10 aa, respectively) and V�4-C�

(CDR3� length of 9 aa) (data not shown). Most CDR3� are codedby two consensus sequences: SDGxGTEV and SDxxGxAETL(Table III). Most of the CDR3� sequences contain N additions (16of 19). Importantly, V�8.2-C� and V�8.2-J�2.1 PCR productswere not found in the brains of the relapsing mice. Therefore, therearrangements used during the first peak of the disease are notinvolved in the relapse. Concerning V� usage, oligoclonal expan-sions of V�4 and J�18 or J�42 segments but not V�9-C� oneswere found in CNS-infiltrating cells of all mice (data not shown).Three 9-aa-long public CDR3� sequences were identified: ERGSALGRL, DRGSALGRL, and EAGGSNAKL (Table IV). ERGSALGRL is encoded by two different nucleotide sequences con-taining or not containing N additions whereas the DRGSALGRLencoding sequence has no N additions. EAGGSNAKL is producedby two different sequences containing N or P additions.

To determine whether these V� and V� public rearrangementswere specific for one of the myelin epitopes, splenocytes frommice nos. 11, 12, and 13 were isolated, labeled with CFSE, andrecalled in vitro with the ID peptides of MOG, MBP, PLP, andthree different potential SD peptides of MOG (6). Splenocytesfrom EAE-suffering mice responded strongly to MOG 35–55only (Fig. 5A). Proliferating CD4� splenocytes (CFSElow) werepurified (Fig. 5B) and their RNA extracted for immunoscopeanalyses and sequencing. The public V�4-J�18, V�4-J�42,

FIGURE 5. Ag specificity of splenocytes from EAE-induced C57BL/6TdT�/� mice during EAE relapse. A, Forty-five days after the induction ofEAE in TdT�/� mice, the animals were killed and cells were collectedfrom either the spleen or the CNS. Splenocytes were labeled with CFSE,cultured for 4 days with various epitopes of the myelin, and their prolif-eration was tested. Data represent SI of TdR incorporation (cpm) fromtriplicate cultures of three different mice. SI (SI � [3H]thymidine cpmincorporated in lymphocytes stimulated by MOG, MBP, or PLP peptides/[3H]thymidine cpm incorporated in unstimulated cells). B, StimulatedCFSE-labeled CD4� splenocytes were sorted out and immunoscope anal-yses were performed.

Table V. CDR3� sequences of CD4� CFSE-labeled MOG35–55-stimulated T splenocytes from EAE-induced C57BL/6 TdT�/� mice during EAErelapse (V�8.3-C� rearrangement)



11 1.1 8 tgtgccagcagtgatg gtcagggg acagaagtcttctttggt SDGQGTEV1.1 8 tgtgccagcagt acagggt caaacacagaagtcttctttggt STGSNTEV1.1 8 tgtgccagcagtg ggacag caaacacagaagtcttctttggt SGTANTEV2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtgatg cgggacagggg gcagaaacgctgtattttggc SDAGQGAETL2.3 10 tgtgccagcagtga cggctggggggtg gcagaaacgctgtattttggc SDGWGVAETL

12 1.1 8 tgtgccagcagt ttcagggag aacacagaagtcttctttggt SFRENTEV1.1 8 tgtgccagcagtg ggactg caaacacagaagtcttctttggt SGTANTEV1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtgatg ggggcgg cacagaagtcttctttggt SDGGGTEV2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtga aggggggggc agtgcagaaacgctgtattttggc SEGGGSAETL2.3 10 tgtgccagcagtgat ccggggggcgcac cagaaacgctgtattttggc SDPGGAPETL2.3 10 tgtgccagcagtgatg cgggacagggg gcagaaacgctgtattttggc SDAGQGAETL

13 1.1 8 tgtgccagcagt acagggt caaacacagaagtcttctttggt STGSNTEV1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtga ccgcagggg cacagaagtcttctttggt SDRRGTEV2.3 10 tgtgccagcagtga aggggggggc agtgcagaaacgctgtattttggc SEGGGSAETL2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtga cggctggggggtg gcagaaacgctgtattttggc SDGWGVAETL2.3 10 tgtgccagcagt cgactggggcct agtgcagaaacgctgtattttggc SRLGPSAETL2.3 10 tgtgccagcag gggactggggggc gtgcagaaacgctgtattttggc RGLGRRAETL2.3 10 tgtgccagcagtgat ccggggggcgcac cagaaacgctgtattttggc SDPGGAPETL

a Underlined nucleotides are due to TdT activity.


V�8.3-J�1.1, and V�8.3-J�2.3 oligoclonal expansions werefound in the sorted CD4� population of the three mice, whereasV�8.2-C�, V�8.2-J�2.1, and V�9-C� PCR products were unde-tectable. All CDR3� and � sequences found in the CNS were alsoidentified in CFSElowCD4� splenocytes (Tables V and VI). Theconsensus CDR3� sequence SDGxGTEV, found in the brain ofmice nos. 11, 12, and 13, was also present in MOG 35–55-prolif-erating T cells (Table V). Three sequences, SDAWGGAETL,SDGWGVAETL, and SDALGGAETL, identified in CNS, werealso found in spleens from at least two mice. However, 9 of 24splenocyte CDR3� sequences were absent in the CNS. Impor-tantly, unique CDR3� sequences found in the CNS of mice nos. 11

(SDRRGTEV) and 12 (SFRENTEV) were present in the spleno-cytes of animals nos. 13 and 12, respectively. Concerning the threepublic CDR3� sequences observed in CNS-infiltrating T cells,they were also found in CFSElowCD4� T splenocytes from at leasttwo of three animals (Table VI). These results strongly suggest thatthe CNS-public rearrangements are borne by MOG 35–55-specificT lymphocytes.

We also determined which repertoire was expressed by CNS-infiltrating lymphocytes during the first relapse occurring inTdT�/� mice (Fig. 2) that have received CD4� T lymphocytesfrom naive C57BL/6 animals before EAE-induction. PublicV�8.3-J�1.1 and V�8.3-J�2.3 rearrangements, previously found

Table VI. CDR3� sequences of CD4� CFSE-labeled MOG 35–55-stimulated T splenocytes from EAE-induced C57BL/6 TdT�/� mice during EAErelapse (V�4-C� rearrangement)

V� J� 3� V� End P/Na 5� J� BeginningDeduced Amino Acid

Sequencesb

11 4 18 tgtgctgctgag agg ggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 42 tgtgctgctgagg ca ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 22 tgtgctgc ccccg catcttctggcagctggcaactcatctttgga PASSGSWQL

12 4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 18 tgtgctgctgagg at ggttcagccttagggaggctgcattttgga EDGSALGRL4 42 tgtgctgctgagg ca ggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 42 tgtgctgctgagg c tggaggaagcaatgcaaagctaaccttcggg EAGGSNAKL4 17 tgtgctgctgag agggg cagtgcagggaacaagctaacttttgga ERGSAGNKL4 58 tgtgctgctgagg ccac aggcactgggtctaagctgtcatttggg EATGTGSKL

13 4 18 tgtgctgctgag agg ggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctgag agaggttcagccttagggaggctgcattttgga ERGSALGRL4 18 tgtgctgctga tagaggttcagccttagggaggctgcattttgga DRGSALGRL4 42 tgtgctgctgagg cgcgg ggaagcaatgcaaagctaaccttcggg EARGSNAKL4 27 tgtgctgctgagg cgacgg ccaatacaggcaaattaacctttggg EATANTGKL4 37 tgtgctgctga cggg acaggcaataccggaaaactcatctttgga DGTGNTGKL

a Nucleotides underlined are due to TdT activity.b Three 9-aa-long public CDR3� sequences are set in special type. ERGSALGRL, which is encoded by two different nucleotide sequences containing or not containing N

additions, is set in boldface type and italics. The DRGSALGRL encoding sequence has no N additions and is set in italics with underlining. EAGGSNAKL is produced by twodifferent sequences containing N or P additions and is set in boldface type only.

Table VII. CDR3� sequences of CNS-infiltrating cells and of CD4� CFSE-labeled, MOG35–55-stimulated T splenocytes during EAE relapse fromtwo various TdT�/� mice in which CD4� T cells from wt animals were previously transferred (V�8.3-C� rearrangement)



CNS 16 1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtga aggggggggc agtgcagaaacgctgtattttggc SEGGGSAETL

CNS 17 1.1 8 tgtgccagcagtgatg ggggcgg cacagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtgatg gtcagggg acagaagtcttctttggt SDGQGTEV2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtgat ccggggggcgcac cagaaacgctgtattttggc SDPGGAPETL

Spleen 16 1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtga ccgcagggg cacagaagtcttctttggt SDRRGTEV2.3 10 tgtgccagcag gggactggggggc gtgcagaaacgctgtattttggc RGLGRRAETL2.3 10 tgtgccagcagtga aggggggggc agtgcagaaacgctgtattttggc SEGGGSAETL2.3 10 tgtgccagcagtgatg cttgggggg gtgcagaaacgctgtattttggc SDAWGGAETL

Spleen 17 1.1 8 tgtgccagcagtgatg gggggggg acagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagtgatg ggggcgg cacagaagtcttctttggt SDGGGTEV1.1 8 tgtgccagcagt ttcagggag aacacagaagtcttctttggt SFRENTEV1.1 8 tgtgccagcagtgatg gtcagggg acagaagtcttctttggt SDGQGTEV2.3 10 tgtgccagcagtgatg cactggggg gtgcagaaacgctgtattttggc SDALGGAETL2.3 10 tgtgccagcagtga cggctggggggtg gcagaaacgctgtattttggc SDGWGVAETL2.3 10 tgtgccagcagtgat ccggggggcgcac cagaaacgctgtattttggc SDPGGAPETL

a Underlined nucleotides are due to TdT activity.


in the relapsing C57BL/6 mice, were strongly expressed in thebrains of individual recipient animals (Table VII) and were similarto those described in Table III. To ascertain that in this experi-mental setting EAE relapse is not due to epitope spreading, thesplenocytes from these mice were stimulated with the same pep-tides as those used in Fig. 5A. T cells responded to MOG 35–55only (data not shown). Additionally, splenocytes from two mice(nos. 16 and 17) were labeled with CFSE before stimulation andthe CFSElowCD4� T lymphocytes were isolated after 4 days inculture. TCR rearrangements were sequenced and revealedCDR3� sequences identical with those found previously (TableVII). Taken together, our results indicate that the first and seconddisease peaks involve two independent waves of MOG-reactive Tcells. This suggests that the lack of relapse in TdT�/� mice isprimarily caused by an inability of these animals to mount thesecond autoreactive wave and that no regulatory mechanisms pre-clude the appearance of MOG-specific T cells expressing theV�8.3-J�1.1 and V�8.3-J�2.3 public rearrangements.

Do T lymphocytes found in the brain of relapsing mice expandduring the first peak of the disease?

We determined whether the public rearrangements (V�8.3-J�1.1and V�8.3-J�2.3) found in the first relapse were generated early inthe immune response against MOG 35–55. LNC from wt micewere collected 9 days after immunization and recalled in vitro for4 days with MOG 35–55. The public V�8.2-J�2.1 CDR3� se-quence GETGGNYAEQ was found in all mice, whereas theV�8.3-C� amplification was productive in five of nine mice (datanot shown). After sequencing, the public CDR3� sequencesSDAWGGAETL and SDGGGTEV, characteristic of the first re-lapse, were observed in two mice only (data not shown). Similaramplifications were performed on cDNA from CNS-infiltratingcells during the first peak of the disease in eight wt mice. We foundthat one mouse expressed the public CDR3� SDGGGTEV onlyand another one expressed SDGGGTEV and the private SDAGQGAETL, also found in the splenocytes of two mice (nos. 11 and12) (Table V and data not shown). This suggests that the publicrearrangements found in the first relapse are expressed by T lym-phocytes that emerge progressively during the initiation of the dis-ease and become predominant while T cells bearing the V�8.2-J�2.1 rearrangement disappear.

DiscussionIn the present study, we have used TdT� and TdT�/� littermatesto investigate the impact of TCR�� diversity on the occurrence ofMOG-induced EAE relapse. At the onset of the disease, the se-verity of MOG-induced EAE is identical in both strains and theirpublic T cell repertoire is highly homologous. However, the courseof the disease is dramatically different depending on the strainbecause EAE relapses occur in TdT� mice only. The lack of EAErelapses in TdT�/� mice is in agreement with the decrease in theincidence of clinical symptoms in TdT-deficient, autoimmune-prone animals. Indeed, when the TdT-deficiency was backcrossedonto autoimmune disease-prone backgrounds, these mice were lesssusceptible to autoimmune nephritis (32), insulitis and diabetes(20, 33), and lupus disease (33, 34). It was suggested that the lackof N additions and/or the decrease in diversity of the T and B cellrepertoires due to TdT deficiency might be responsible for thedelayed onsets of different autoimmune diseases and for the lesssevere clinical and biological symptoms at the end stages of thesediseases. In addition, it was recently shown that in C57BL/6 miceimmunization with recombinant murine MOG does not generatepathogenic B cells or demyelinating Abs (35). Thus, anti-MOG Bcells do not play a major pathogenic role in this EAE model. In this

report, we show that failure to relapse in TdT�/� mice does notstem from dominant regulatory mechanisms. Importantly, relapsein wt animals is not due to epitope spreading but to MOG-specificT lymphocytes expressing new public V� and V� rearrangementscontaining N additions.

As reported for various ID epitopes (22, 23), MOG 35–55-spe-cific public T cell repertoires of TdT� and TdT�/� littermattes areclosely similar. Wt and TdT�/� mice share the public CDR3�sequence GETGGNYAEQ (Ref. 25) and this work, suggesting astrong bias for this amino acid sequence encoded by eight differentnucleotide sequences. Concerning the V�-chain, the public V�9-J�23 rearrangement generating the CDR3� (RSYNQGKL) in wtmice is replaced in TdT�/� animals by CDR3� sharing the con-sensus sequence SxYNQGKL. Moreover, the V�9-J�31 CDR3�SRNSNNRI is public in wt mice and is found in four of 10 TdT�/�

animals. It is worth noticing that MOG-specific public T cell rep-ertoires from CNS-infiltrating cells and stimulated LNC in TdT�/�

mice are identical. Thus, in the two strains the first peak of EAEinvolves encephalitogenic MOG-specific CD4� T cells that ex-press highly homologous public TCR rearrangements.

In contrast, new V� and V� public rearrangements, most ofwhich contain N additions, are found in CNS-infiltrating T cells ofwt mice undergoing the first EAE relapse. CD4� T splenocytesbearing these rearrangements and derived from these mice prolif-erate to MOG 35–55 and not to other ID peptides of PLP and MBPautoantigens. ID epitopes of MBP and PLP are immunogenic inTdT� and TdT�/� animals and the ID epitope of PLP is enceph-alitogenic in both strains (data not shown), whereas C57BL/6 miceare known to be resistant to MBP-induced EAE. Thus, the absenceof relapse in TdT�/� mice is not related to holes in their T cellrepertoires, and in wt mice the first EAE relapse is not due toepitope spreading of the T cell response even though we cannotrule out the implication of other myelin autoantigens. Thus, EAErelapse predominantly involves MOG-specific T cells bearing pub-lic N diversified rearrangements.

Strikingly, public T lymphocytes emerging during the first peakof the disease are undetectable by PCR in wt mice during therelapse and thus do not participate in it. This suggests that thesepublic T lymphocytes are not anergic or do not acquire a protectiveTh2 phenotype and are probably eliminated. Many hypothesescould explain this phenomenon: 1) autoreactive T cells can bedeleted through activation induced cell death; 2) Fas-FasL and/orTNF-TNF-R1 interactions are responsible for EAE remissions asshown in different antigenic models (36–38); and 3) various reg-ulatory T cells inhibit encephalitogenic responses in EAE. Indeed,in B10.PL, MBP-induced EAE results from the expansion ofCD4� T cells expressing predominantly the public V�8.2 CDR3�DAGGGY. These encephalitogenic lymphocytes are deleted ortheir activity is suppressed by regulatory CD4� T cell clones spe-cific for framework 3 of the V�8.2 gene segment (39) and CD8�

T cell clones specific for the V�8.2 peptide 42–50 (40). Otherstudies have provided evidence that Qa1-restricted CD8� T cellscan modulate EAE (41–43). These regulatory T cells specificallyinteract through their �� TCR with Qa-1/V� peptides complexespresent on autoimmune CD4� T cells (44–46). Recently, the im-portance of this regulatory pathway was evidenced in Qa-1-defi-cient mice (47). In the absence of Qa-1-restricted CD8� T cellscapable of inhibiting PLP-specific CD4 T cells, these Qa-1-deficient mice were susceptible to EAE relapses after secondaryand tertiary immunization with PLP while wt animals becameresistant (47).

Transfer experiments using CD4� T splenocytes from naive wtmice as donor cells and naive TdT�/� mice as recipients clearlyestablish that no regulatory mechanisms in TdT�/� animals inhibit


the expansion of autoreactive T cells. Indeed, wt CD4� T cellsinduce EAE relapse in TdT�/� mice and express the characteristicpublic V�8.3-J�1.1/V�8.3-J�2.3 rearrangements. Nevertheless,regulatory T cells could emerge in TdT�/� mice during the courseof the disease and inhibit relapses. Because various subsets of Tcells endowed with regulatory properties have been described, wetransferred purified T splenocytes from EAE-recovered TdT�/�

and TdT�/� littermates into EAE-recovered TdT�/� and TdT�/�

mice. As shown in Fig. 3, no suppressive activity peculiar to theTdT�/� strain could explain the absence of relapses.

The emergence of MOG-specific T lymphocytes bearing thenew public V� rearrangements during the first EAE relapse led usto investigate whether these T cells were present at the onset of thedisease. The public CDR3� sequences were only present in stim-ulated-LNC from two of nine mice and in CNS-infiltrating cellsfrom two of eight mice. This suggests that T cells bearing theseCDR3� sequences expand progressively and become predominantas the public V�8.2-J�2.1 T lymphocytes from the initial EAEpeak disappear. The sequential emergence of these public reper-toires against a single epitope during EAE may be explained by thefollowing: 1) higher avidity and/or higher precursors frequency ofthe public V�8.2-J�2.1-bearing T lymphocytes as compared withpublic V�8.3-J�1.1/V�8.3-J�2.3 T cells; and 2) specific elimina-tion of V�8.2-J�2.1 T cells by yet undefined mechanisms. Inter-estingly, in PLP 139–151-induced EAE, Targoni et al. (48)showed that the vast majority of PLP 139–151-specific CD4 Tcells that appear at the onset of the disease and persist duringclinical EAE are eliminated without acquiring a Th2 phenotype.The first relapse is due to intramolecular epitope spreading of theimmune response characterized by the expansion of PLP 78–191-pecific T lymphocytes (48). In this model, the preferential expan-sion of PLP 178–191 CD4 T cells in the CNS during the first EAErelapse may be explained by a repertoire of PLP 139–151-specificT lymphocytes less diversified than the MOG 35–55 T cell reper-toire or to higher affinity of the TCRs specific for PLP 178–191when compared to the low-affinity PLP 139–151-specific T lym-phocytes remaining after remission.

Determinant spreading occurring during ongoing autoimmunediseases involves the sequential emergence(s) of autoreactive Tcells recognizing self-epitopes different from the disease-inducingself-determinant (7). The public repertoire-bearing encephalito-genic T lymphocytes from the second peak of the disease are spe-cific for the MOG 35–55 peptide like those of the first peak, in-dicating that no epitope spreading happens in our MOG-inducedEAE model. Thus, the absence of spreading to the ID determinantof MPB and PLP could be due to the following nonexclusive rea-sons: 1) predominant Ag processing and presentation of MOG35–55 vs ID epitopes of MBP and PLP; 2) persistence of a largeT cell repertoire against MOG 35–55 (indeed, comparingMOG�/� and MOG�/� mice, we have previously shown (25) thatMOG is unable to impact on the public MOG-specific T cell rep-ertoire); and 3) lower avidity and/or lower frequencies of PLP- andMBP-specific T cells remaining after tolerance induction.

Our studies in TdT�/� and TdT� mice have unraveled majormodifications of the T cell repertoire that impact upon the evolu-tion of MOG-induced EAE. If these observations can be general-ized to other organ-specific autoimmune diseases, one can specu-late that successive waves of pathogenic T lymphocytes expressingdifferent repertoires lead to progressive destruction of tissues.Thus, one wave of autoimmune T lymphocytes that arises is even-tually eliminated or down-regulated but, in the highly diversified Tcell repertoire of TdT� individual, another subset of specific Tcells expands and leads to relapse. Our works show that during thenatural course of an organ-specific autoimmune disease a highly

adaptable T cell repertoire is able to induce relapses against asingle autoantigen without obvious epitope spreading. Thus, itwould be of high interest to determine the relative role played byautoantigen spreading vs the fine tuning of pathogenic T cell rep-ertoires in organ-specific autoimmune diseases.

AcknowledgmentsWe thank Drs. M. L. Albert, S. Anderton, P. Bousso, F. Lemmonier, andN. A. Mitchison for their comments on the manuscript.

DisclosuresThe authors have no financial conflict of interest.

References1. Lassmann, H. 2004. Experimental autoimmune encephalomyelitis. In Myelin Bi-

ology and Disorders. R. A. Lazzarini, ed. Elsevier Academic Press, San Diego,pp. 1039–1071.

2. Weaver, C. T., L. E. Harrington, P. R. Mangan, M. Gavrieli, and K. M. Murphy.2006. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity24: 677–688.

3. Zamvil, S. S., D. J. Mitchell, A. C. Moore, K. Kitamura, L. Steinman, andJ. B. Rothbard. 1986. T-cell epitope of the autoantigen myelin basic protein thatinduces encephalomyelitis. Nature 324: 258–260.

4. Tuohy, V. K., Z. Lu, R. A. Sobel, R. A. Laursen, and M. B. Lees. 1989. Iden-tification of an encephalitogenic determinant of myelin proteolipid protein forSJL mice. J. Immunol. 142: 1523–1527.

5. Kerlero de Rosbo, N., I. Mendel, and A. Ben-Nun. 1995. Chronic relapsing ex-perimental autoimmune encephalomyelitis with a delayed onset and an atypicalclinical course, induced in PL/J mice by myelin oligodendrocyte glycoprotein(MOG)-derived peptide: preliminary analysis of MOG T cell epitopes. Eur. J. Im-munol. 25: 985–993.

6. Delarasse, C., P. Daubas, L. T. Mars, C. Vizler, T. Litzenburger, A. Iglesias,J. Bauer, B. Della Gaspera, A. Schubart, L. Decker, et al. 2003. Myelin/oligo-dendrocyte glycoprotein-deficient (MOG-deficient) mice reveal lack of immunetolerance to MOG in wild-type mice. J. Clin. Invest. 112: 544–553.

7. Vanderlugt, C. L., and S. D. Miller. 2002. Epitope spreading in immune-mediateddiseases: implications for immunotherapy. Nat. Rev. Immunol. 2: 85–95.

8. Lehmann, P. V., T. Forsthurber, A. Miller, and E. E. Sercarz. 1992. Spreading ofT-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358:155–157.

9. McRae, B. L., C. L. Vanderlugt, M. C. Dal Canto, and S. D. Miller. 1995.Functional evidence for epitope spreading in the relapsing pathology of experi-mental autoimmune encephalomyelitis. J. Exp. Med. 182: 75–85.

10. Vanderlugt, C. L., K. L. Neville, K. M. Nikcevich, T. N. Eagar, J. A. Bluestone,and S. D. Miller. 2000. Pathologic role and temporal appearance of newly emerg-ing autoepitopes in relapsing experimental autoimmune encephalomyelitis. J. Im-munol. 164: 670–678.

11. Yu, M., J. M. Johnson, and V. K. Tuohy. 1996. A predictable sequential deter-minant spreading cascade invariably accompanies progression of experimentalautoimmune encephalomyelitis: a basis for peptide-specific therapy after onset ofclinical disease. J. Exp. Med. 183: 1777–1788.

12. Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting,P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann.1993. Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase inmurine insulin-dependent diabetes. Nature 366: 69–72.

13. Tisch, R., X.-D. Yang, S. M. Singer, R. S. Liblau, L. Fugger, and H. O. Mc Devitt.1993. Immune response to glutamic acid decarboxylase correlates with insulitis innon-obese diabetic mice. Nature 366: 72–75.

14. Takacs, K., and D. M. Altmann. 1998. The case against epitope spread in ex-perimental allergic encephalomyelitis. Immunol. Rev. 164: 101–110.

15. Jones, R. E., D. Bourdette, N. Moes, A. Vandenbark, A. Zamora, and H. Offner.2003. Epitope spreading is not required for relapses in experimental autoimmuneencephalomyelitis. J. Immunol. 170: 1690–1698.

16. Smith, P. A., M. Morris-Downes, N. Heijmans, G. Pryce, E. Arter, K. O’Neill,J. B. t Hart, D. Baker, and S. Amor. 2005. Epitope spread is not critical for therelapse and progression of MOG 8–21 induced EAE in Biozzi ABH mice. J. Neu-roimmunol. 164: 76–84.

17. Wilson, R. K., E. Lai, P. Concannon, R. K. Barth, and L. E. Hood. 1988. Struc-ture, organization and polymorphism of murine and human T-cell receptor � and� chain gene families. Immunol. Rev. 101: 149–172.

18. Gilfillan, S., A. Dierich, M. Lemeur, C. Benoist, and D. Mathis. 1993. Micelacking TdT: mature animals with an immature lymphocyte repertoire. Science261: 1175–1178.

19. Komori, T., A. Okada, V. Stewart, and F. W. Alt. 1993. Lack of N regions inantigen receptor variable region genes of TdT-deficient lymphocytes. Science261: 1171–1175.

20. Gilfillan, S., C. Benoist, and D. Mathis. 1995. Mice lacking terminal deoxynu-cleotidyl transferase: adult mice with a fetal antigen receptor repertoire. Immunol.Rev. 148: 201–219.

21. Cabaniols, J. P., N. Fazilleau, A. Casrouge, P. Kourilsky, and J. M. Kanellopoulos.2001. Most �/� T cell receptor diversity is due to terminal deoxynucleotidyl trans-ferase. J. Exp. Med. 194: 1385–1390.


22. Gilfillan, S., M. Bachmann, S. Trembleau, L. Adorini, U. Kalinke,R. Zinkernagel, C. Benoist, and D. Mathis. 1995. Efficient immune responses inmice lacking N-region diversity. Eur. J. Immunol. 25: 3115–3122.

23. Fazilleau, N., J. P. Cabaniols, F. Lemaitre, I. Motta, P. Kourilsky, andJ. M. Kanellopoulos. 2005. V� and V� public repertoires are highly conserved interminal deoxynucleotidyl transferase-deficient mice. J. Immunol. 174: 345–355.

24. Malipiero, U., K. Frei, K. S. Spanaus, C. Agresti, H. Lassmann, M. Hahne,J. Tschopp, H. P. Eugster, and A. Fontana. 1997. Myelin oligodendrocyte gly-coprotein-induced autoimmune encephalomyelitis is chronic/relapsing in perforinknockout mice, but monophasic in Fas- and Fas ligand-deficient lpr and gld mice.Eur. J. Immunol. 27: 3151–3160.

25. Fazilleau, N., C. Delarasse, C. H. Sweenie, S. M. Anderton, S. Fillatreau,F. A. Lemonnier, D. Pham-Dinh, and J. M. Kanellopoulos. 2006. Persistence ofautoreactive myelin oligodendrocyte glycoprotein (MOG)-specific T cell reper-toires in MOG-expressing mice. Eur. J. Immunol. 36: 533–543.

26. Cibotti, R., J. P. Cabaniols, C. Pannetier, C. Delarbre, I. Vergnon, J. M.Kanellopoulos, and P. Kourilsky. 1994. Public and private V� T cell receptor rep-ertoires against hen egg white lysozyme (HEL) in nontransgenic versus HEL trans-genic mice. J. Exp. Med. 180: 861–872.

27. Tompkins, S. M., J. Padilla, M. C. Dal Canto, J. P. Ting, L. Van Kaer, andS. D. Miller. 2002. De novo central nervous system processing of myelin antigenis required for the initiation of experimental autoimmune encephalomyelitis.J. Immunol. 168: 4173–4183.

28. Shaw, M. K., C. Kim, H. W. Hao, F. Chen, and H. Y. Tse. 1996. Induction ofmyelin basic protein-specific experimental autoimmune encephalomyelitis inC57BL/6 mice: mapping of T cell epitopes and T cell receptor V� gene segmentusage. J. Neurosci. Res. 45: 690–699.

29. Pannetier, C., M. Cochet, S. Darche, A. Casrouge, M. Zoller, and P. Kourilsky.1993. The sizes of the CDR3 hypervariable regions of the murine T-cell receptor� chains vary as a function of the recombined germ-line segments. Proc. Natl.Acad. Sci. USA 90: 4319–4323.

30. Feeney, A. J. 1991. Junctional sequences of fetal T cell receptor � chains havefew N regions. J. Exp. Med. 174: 115–124.

31. Gavin, M. A., and M. J. Bevan. 1995. Increased peptide promiscuity provides arationale for the lack of N regions in the neonatal T cell repertoire. Immunity 3:793–800.

32. Conde, C., S. Weller, S. Gilfillan, L. Marcellin, T. Martin, and J. L. Pasquali.1998. Terminal deoxynucleotidyl transferase deficiency reduces the incidence ofautoimmune nephritis in (New Zealand Black � New Zealand White)F1 mice.J. Immunol. 161: 7023–7030.

33. Robey, I. F., M. Peterson, M. S. Horwitz, D. H. Kono, T. Stratmann,A. N. Theofilopoulos, N. Sarvetnick, L. Teyton, and A. J. Feeney. 2004. Terminaldeoxynucleotidyltransferase deficiency decreases autoimmune disease in diabe-tes-prone nonobese diabetic mice and lupus-prone MRL-Faslpr mice. J. Immunol.172: 4624–4629.

34. Feeney, A. J., B. R. Lawson, D. H. Kono, and A. N. Theofilopoulos. 2001.Terminal deoxynucleotidyl transferase deficiency decreases autoimmune diseasein MRL-Faslpr mice. J. Immunol. 167: 3486–3493.

35. Marta, C. B., A. R. Oliver, R. A. Sweet, S. E. Pfeiffer, and N. H. Ruddle. 2005.Pathogenic myelin oligodendrocyte glycoprotein antibodies recognize glycosy-lated epitopes and perturb oligodendrocyte physiology. Proc. Natl. Acad. Sci.USA 102: 13992–13997.

36. Sabelko-Downes, K. A., A. H. Cross, and J. H. Russell. 1999. Dual role for Fasligand in the initiation of and recovery from experimental allergic encephalomy-elitis. J. Exp. Med. 189: 1195–1205.

37. Dittel, B. N., R. M. Merchant, and C. A. Janeway, Jr. 1999. Evidence for Fas-dependent and Fas-independent mechanisms in the pathogenesis of experimentalautoimmune encephalomyelitis. J. Immunol. 162: 6392–6400.

38. Bachmann, R., H. P. Eugster, K. Frei, A. Fontana, and H. Lassmann. 1999.Impairment of TNF-receptor-1 signaling but not Fas signaling diminishes T-cellapoptosis in myelin oligodendrocyte glycoprotein peptide-induced chronic de-myelinating autoimmune encephalomyelitis in mice. Am. J. Pathol. 154:1417–1422.

39. Kumar, V., and E. E. Sercarz. 1993. The involvement of T cell receptor peptide-specific regulatory CD4� T cells in recovery from antigen-induced autoimmunedisease. J. Exp. Med. 178: 909–916.

40. Madakamutil, L. T., I. Maricic, E. Sercarz, and V. Kumar. 2003. Regulatory Tcells control autoimmunity in vivo by inducing apoptotic depletion of activatedpathogenic lymphocytes. J. Immunol. 170: 2985–2992.

41. Koh, D. R., W. P. Fung-Leung, A. Ho, D. Gray, H. Acha-Orbea, and T. W. Mak.1992. Less mortality but more relapses in experimental allergic encephalomyeli-tis in CD8�/� mice. Science 256: 1210–1213.

42. Jiang, H., S. I. Zhang, and B. Pernis. 1992. Role of CD8� T cells in murineexperimental allergic encephalomyelitis. Science 256: 1213–1215.

43. Jiang, H., R. Ware, A. Stall, L. Flaherty, L. Chess, and B. Pernis. 1995. MurineCD8� T cells that specifically delete autologous CD4� T cells expressing V�8TCR: a role of the Qa-1 molecule. Immunity 2: 185–194.

44. Jiang, H., H. Kashleva, L. X. Xu, J. Forman, L. Flaherty, B. Pernis,N. S. Braunstein, and L. Chess. 1998. T cell vaccination induces T cell receptorV�-specific Qa-1-restricted regulatory CD8� T cells. Proc. Natl. Acad. Sci. USA95: 4533–4537.

45. Jiang, H., N. S. Braunstein, B. Yu, R. Winchester, and L. Chess. 2001. CD8� Tcells control the TH phenotype of MBP-reactive CD4� T cells in EAE mice.Proc. Natl. Acad. Sci. USA 98: 6301–6306.

46. Jiang, H., S. Curran, E. Ruiz-Vazquez, B. Liang, R. Winchester, and L. Chess.2003. Regulatory CD8� T cells fine-tune the myelin basic protein-reactive T cellreceptor V � repertoire during experimental autoimmune encephalomyelitis.Proc. Natl. Acad. Sci. USA 100: 8378–8383.

47. Hu, D., K. Ikizawa, L. Lu, M. E. Sanchirico, M. L. Shinohara, and H. Cantor.2004. Analysis of regulatory CD8 T cells in Qa-1-deficient mice. Nat. Immunol.5: 516–523.

48. Targoni, O. S., J. Baus, H. H. Hofstetter, M. D. Hesse, A. Y. Karulin,B. O. Boehm, T. G. Forsthuber, and P. V. Lehmann. 2001. Frequencies of neu-roantigen-specific T cells in the central nervous system versus the immune pe-riphery during the course of experimental allergic encephalomyelitis. J. Immunol.166: 4757–4764.


Documents

T Cell Repertoire Diversity Is Required for Relapses in Myelin Oligodendrocyte Glycoprotein-Induced Experimental Autoimmune Encephalomyelitis