Abstract
Multiple sclerosis (MS) is an inflammatory and demyelinating disease of the CNS with associated axonal loss. There is strong evidence for an autoimmune pathogenesis driven by myelin-specific T cells. Myelin oligodendrocyte glycoprotein (MOG) induces a type of experimental autoimmune encephalomyelitis in animals which is very MS-like since there are demyelinating CNS lesions and axonal loss. This underscores the potential role of MOG in MS pathogenesis. We performed a T cell reactivity pattern analysis of MS patients at the onset of relapse or progression of neurological deficits and controls that were stratified for the genetic risk factor HLA-DRB1*1501. For the first time, we show that there is an HLA-DR-restricted promiscuous dominant epitope for CD4+ T cells within the transmembrane/intracellular part of MOG comprising aa 146–154 (FLCLQYRLR). Surprisingly, controls had broader T cell reactivity patterns toward MOG peptides compared with MS patients, and the transmembrane and intracellular parts of MOG were much more immunogenic compared with the extracellular part. Measurements of in vitro binding affinities revealed that HLA-DRB1*1501 molecules bound MOG 146–154 with intermediate and HLA-DRB1*0401 molecules with weak affinities. The binding of MOG 146–154 was comparable or better than myelin basic protein 85–99, which is the dominant myelin basic protein epitope in context with HLA-DRB1*1501 molecules in MS patients. This is the first study in which the data underscore the need to investigate the pathogenic or regulatory role of the transmembrane and intracellular part of MOG for MS in more detail.
Multiple sclerosis (MS)3 is an inflammatory and demyelinating disease of the CNS with major socioeconomic impact (1). Besides environmental stimuli, genetic factors determine disease outbreak. There is an increased risk for patients being HLA-DRB1*1501, DRB5*0101, and DQB1*0602, which are contained in the Dw2 haplotype to develop MS, and HLA-DRB1*1501 is associated with an earlier age of disease manifestation (2). Myelin-specific autoreactive T cells pass the blood brain barrier and initiate an autoimmune attack against myelin sheaths in the CNS (3). Depending on the degree of the myelin sheath-specific autoimmune attack, axons can also be harmed, leading to irreversible axonal loss and long-term global brain atrophy (4). Several myelin proteins are thought to be targets of tissue damage in MS (5). Much attention has been given to myelin-basic protein (MBP) and proteolipid protein. More recently, myelin oligodendrocyte glycoprotein (MOG) has gained considerable attention with regard to the autoimmune attack in MS (6). This is due to the high histopathological similarity between MOG-induced animal models and biopsies from MS patients, both showing large demyelinating lesions and axonal loss (7, 8). MOG only composes ∼0.01% of the protein content of the myelin sheath and is exposed on its outer surface (6). It has an extracellular part including aa 1–122 with an Ig-like domain, a transmembrane part, and an intracellular part comprising aa 123–218 (Fig. 7⇓; Refs. 9, 10, 11). Up to now, T cell responses against the extracellular part of MOG have been mainly looked at. Using the ELISPOT methodology, Wallström et al. (12) demonstrated increased IFN-γ-secreting cells in HLA-DR2(15) positive MS patients compared with controls toward several MOG peptides among which MOG 63–87 was immunodominant. Others investigated T cell responses to MOG of non-HLA-stratified MS patients and controls indicating increased T cell reactivity to MOG in MS patients (13, 14, 15). Also, MOG-specific B cell responses are up-regulated in MS, and MOG-specific Abs contribute to lesion formation (16, 17). Because MOG can be considered an important autoantigen in MS and a potential target for therapy, we investigated T cell responses by ELISPOT assay for IFN-γ-secreting cells to peptides of the complete human MOG 1–218 sequence. We used the ELISPOT assay to detect IFN-γ-secreting cells, because this is more sensitive than proliferation as assessed by [3H]TdR-uptake, and adds a functional outread. We demonstrate a previously unknown dominant T cell epitope that is recognized in context of HLA-DR molecules by CD4+ T cells within the transmembrane and intracellular sequence 146–154 of MOG in MS patients and controls.
Materials and Methods
Subjects
MS patients and controls were stratified for HLA-DRB1*1501. This resulted in four groups with 11 MS patients and 10 controls being HLA-DRB1*1501 positive, and 7 MS patients and 10 controls being HLA-DRB1*1501 negative. A total of 13 patients had a relapsing-remitting MS disease course with an average disease duration at a sample collection of 1 year and 6 mo (mean expanded disability status scale (EDSS) 2.0). Four patients had a secondary progressive disease course at sample collection with an average disease duration of 8 years (mean EDSS 5.0), and one patient had a primary progressive disease course with a duration of 7 years at sample collection (EDSS 6.5). Sixteen of 18 patients had not been treated at all or had not received any immunosuppressive or immunomodulatory treatment for at least 1 year before sample collection. Two of 18 patients had not received immunosuppressive or immunomodulatory treatment for 1 mo before sample collection. All patients were admitted to Department of Neurology of the University of Tubingen (Tubingen, Germany) due to acute relapse (relapsing remitting disease course) or fast worsening of neurological deficits (primary or secondary progressive disease course). The mean age of MS patients was 34 ± 9 years and of controls 38 ± 11 years, with a male to female ratio of 50:50% in MS patients and 55:45% in controls. The age distribution and female:male ratio was equal in the HLA-DRB1*1501 positive and negative groups. The studies had been approved by the ethical review board of the University of Tübingen.
HLA-typing
Genomic DNA for HLA-genotyping was prepared by the QIAampBlood kit (Qiagen, Hilden, Germany). Low resolution pregenotyping was performed for the 116 major HLA-DRB1, -DRB3, -DRB4, -DRB5, and 29 HLA-DQB alleles, and HLA-DR or HLA-DQ subtyping was done specifically for the pretyped HLA-DRB1/DQB1 alleles by group-specific amplification and subsequent direct sequencing in patients and controls.
Cloning and bacterial expression of recombinant human (rh)MOG 1–125
The cDNA of rhMOG was obtained by reverse transcription of total RNA from a human glioma cell line. RNA was prepared with TRIzol reagent (Life Technologies, Gaithersburg, MD) and first strand cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI) and the gene-specific primer R785. The amplified PCR product that contains the whole coding sequence of MOG mRNA was separated by agarose gel electrophoresis and purified with a QIAEXII-kit (Qiagen). The cDNA fragment was cloned into Bluescript II KS+ vector (Stratagene, La Jolla, CA). The DNA sequence encoding the extracellular domain of the mature protein (including four N-terminal amino acids of the transmembrane domain) was PCR-amplified and subcloned in pQ60 (Qiagen). The Escherichia coli-expressed His-tagged fusion protein (rhMOG 1–125) was purified under denaturing conditions by metal chelate affinity chromatography on Ni-NTA agarose columns (Qiagen) according to the manufacturer’s guidelines.
Synthesis and analysis of peptides
The 16 aa-long peptides (Table I⇓) and the N-acetylated C-amidated 9 aa-long peptides (Table II⇓) were prepared by solid phase peptide synthesis using F-moc/tBu chemistry. The peptides were purified by preparative HPLC (Abimed, Langenfeld, Germany). The identity of the purified peptides was confirmed by electrospray mass spectrometry. The purity of peptides was >95% as determined by analytical HPLC (Abimed). MBP 85–99 (ENPVVHFFKNIVTPR) and influenza A peptide (YRNLVWFIKKNTRYP) (18) were synthesized and purified in the same way. The peptides were used at a concentration of 10 μg/ml in all experiments. This concentration had given the highest number of spots in ELISPOT analyses in pilot experiments (data not shown).
List of peptides spanning the complete MOG sequence
List of truncated N-acetylated C-amidated MOG peptides of sequence MOG 142–160
Mitogen
PHA was purchased from Sigma-Aldrich (Deisenhofen, Germany) and used in a concentration of 3 μg/ml in all experiments. This concentration had given optimal results in preceding experiments.
Purification of PBMC
PBMCs were isolated from heparinized blood samples on Lymphoprep density gradients (Nyegaard, Oslo, Norway; 200 × g, 25 min, room temperature). Cells were frozen at a density of 5 × 106 cells/ml in freezing medium containing 40% FCS Gold (PAA, Linz, Austria), 10% DMSO (Serva, München, Germany), and 50% complete medium (CM) consisting of RPMI 1640 (Life Technology, Eggenstein, Germany) supplemented with 2 mM glutamine (Life Technology), 100 U/ml of penicillin, 100 μg/ml of streptomycin (Biochrom, Berlin, Germany), and 3% heat-inactivated human AB serum in liquid nitrogen. The cells were recovered by thawing in a 37°C water bath until the cells reached the melting point and then by adding 1 ml CM per 5 × 106 cells three times every 5 min at room temperature. Thereafter, the cells were washed three times in CM and resuspended at a density of 106/ml in CM. The viability was 90–95%.
ELISPOT analysis for IFN-γ-secreting cells and restriction analysis
Ninety six-well nitrocellulose plates (Millipore, Molsheim, France) were coated with 10 μg/ml capture mAb 1-D1K (Mabtech, Stockholm, Sweden) overnight at 4°C. After washing, the membrane was blocked with culture medium containing 10% FCS (FCS Gold, PAA) for 1 h. A total of 2 × 105 PBMCs/well were cultured for 40 h in 37°C, 5% CO2. For each Ag or mitogen, triplicates were used. Each plate contained a positive and negative control. All peptides and mitogens were tested for one patient or control at the same time point with exactly the same procedures. After 40 h, the cells were discarded and the membranes were thoroughly washed by immersing the plates six times in PBS. To visualize areas of the membrane that had bound secreted IFN-γ, biotinylated detector mAb 7-6B-1 (1 μg/ml, Mabtech) was added for 3 h and staining performed with avidin-biotin peroxidase complex (Vectastatin Elite ABC kit; Vector Laboratories, Burlingame, CA) and chromogen solution containing carbazole (Sigma-Aldrich). Areas of the membrane where a specific color reaction had occurred appeared as dark brown-red spots and were both counted by an automated ELISPOT counter (Autoimmun-Diagnostika, Albstadt, Germany) and manually cross-checked. The average number of spots in triplicates secreted after exposure with Ag or mitogen were expressed as numbers of IFN-γ-secreting cells per 2 × 105 cells added initially to the wells. For the restriction analysis, the HLA typing Abs Genox with the specificity anti-HLA-DQ and/or Tü 36 with the specificity anti-HLA-DR were added at a concentration of 10 μg/ml to the cultures. These concentrations of Abs had given optimal results in pilot experiments (data not shown).
Enrichment of IFN-γ-secreting cells and FACS analyses
Enrichment of cells secreting IFN-γ after contact with Ag was performed with magnetic cell separation as described by the manufacturer (IFN-γ secretion assay; Miltenyi Biotec, Bergisch Gladbach, Germany). PBMC from MS patients and controls were incubated with no Ag, MOG 146–154 (10 μg/ml), or as a positive control staphylococcal enterotoxin B (Sigma-Aldrich, 10 μg/ml) for 12 h at a concentration of 5 × 106 cells/ml at 37°C, 5% CO2. Thereafter, cells were exposed to a bivalent Ab to IFN-γ and leukocyte surface Ag (Miltenyi Biotec) at 4°C resulting in an activity matrix for IFN-γ on the cell surface. Subsequently, cells were reinduced to secrete IFN-γ at 37°C, 5% CO2. Secreted IFN-γ is captured by the activity matrix on the cell surface. Next, cells were incubated with a PE-labeled Ab to IFN-γ (Miltenyi Biotec). Finally, cells with bound IFN-γ on their cell surface were separated with a PE-specific microbead (Miltenyi Biotec) in a magnetic field. Cells were analyzed for expression of IFN-γ (mouse PE-labeled anti-human IFN-γ; Miltenyi Biotec) and CD4 (mouse FITC-labeled anti-human CD4; BD Biosciences, Lincoln Park, NJ) and/or CD8 (mouse FITC-labeled anti-human CD8) expression following standard procedures by FACS (BD Biosciences).
Peptide binding assay
Relative affinities of MOG peptides for purified HLA-DRB1*1501 and HLA-DRB1*0401 molecules were measured by an inhibition ELISA based on a dissociation-enhanced lanthanide fluoroimmunoassay (Wallac, Turku, Finland). In the inhibition ELISA, HLA-DR (50 nM) molecules were incubated with fixed amounts of respective tracer peptides (10–50 nM) in the presence of a range of dilutions of the unlabeled MOG-peptides (10-fold dilutions between 1 nM and 100 μM). The binding buffer consisted of a carbonate buffer titrated to pH 5 containing 2 mM EDTA, 0.01% azide, 0.1 mM PMSF, and 0.1% Nonidet P-40 (Boehringer Mannheim, Indianapolis, IN). After an incubation of 48 h at 37°C, the peptide-MHC complexes were transferred to Ab-coated (L243) ELISA plates (FluoroNunc; Nunc, Roskilde, Denmark) to remove the excess of nonbound peptides. Europium-labeled streptavidin (Wallac) was added to the plates and incubated for 1 h at room temperature. Finally, the plates were treated with an enhancement solution (Wallac), which releases chelated europium from streptavidin and forms a highly fluorescent solution that can be measured by using a dissociation-enhanced lanthanide fluoroimmunoassay fluorometer (Wallac). The peptide concentration yielding 50% inhibition of binding of the tracer peptide (IC50) was determined by plotting the percentage of inhibition vs the concentration of added MOG peptide. Peptides were tested in two to three independent experiments.
Statistical analysis
Repeated measures ANOVA on signed ranks were used with the Dunnett’s test (Sigma Stat; Jandel Scientific, San Rafael, CA). The Friedman repeated measures analysis of variance on ranks (repeated measures ANOVA on signed ranks) is a parametric test that compares effects of a series of different experimental treatments on a single group. Each subject’s responses are ranked from smallest to largest without regard to other subjects, then the rank sums for the subjects are compared. Dunnett’s test is the analog of the Student-Newman-Keuls test for the case of multiple comparisons against a single control group. It is conducted similarly to the Bonferroni t test, but with a more sophisticated mathematical model of the way the error accumulates to derive the associated table of critical values for hypothesis testing. This test is less conservative than the Bonferroni test, and is only available for multiple comparisons vs a control. This analysis was performed independently for each of the four groups (DRB1*1501 positive MS patients, DRB1*1501 positive controls, DRB1*1501 negative MS patients, DRB1*1501 negative controls). As controls, the individual background reactivities were used (T cell responses without Ag).
Results
T cell repertoire analysis to the extracellular part of MOG
We stimulated PBMCs from MS patients and controls with 52 overlapping peptides covering the complete human MOG sequence (Table I⇑) or with the extracellular rhMOG 1–125 (mature peptide, aa 1–125) and enumerated IFN-γ-secreting cells with the ELISPOT assay (Figs. 1⇓ and 2⇓). The T cell responses as assessed by IFN-γ secretion toward peptides of the extracellular part of MOG were weak and heterogeneous. HLA-DRB1*1501 positive controls showed responses to peptide MOG 81–96 (p < 0.05) (Fig. 2⇓), while MS patients and HLA-DRB1*1501 negative controls reacted weakly (NS). Some controls reacted with MOG 73–88 (NS), but not MS patients (NS). Only a few individuals had IFN-γ-producing cells that reacted with rhMOG 1–125 (NS). Background values for nonstimulated cultures did not differ significantly between MS patients and controls (NS) (data not shown).
Determinant mapping by ELISPOT assay for IFN-γ-secreting cells in a DRB1*1501 DRB1*04011 DQB1*0302 DQB1*0602 MS patient. IFN-γ secreting mononuclear cells from peripheral blood reactive with 16 aa-long peptides covering the complete MOG sequence and toward rhMOG 1–125 are indicated. The number of spots indicating cells having secreted IFN-γ after antigenic stimulus was enumerated with an automated analysis system. The procedures were performed as described in Materials and Methods. M indicates rhMOG 1–125.
Reactivity profiles of MS patients and controls toward overlapping peptides of the complete human MOG sequence. Reactivity profiles of PBMCs secreting IFN-γ in ELISPOT assay toward overlapping MOG peptides and rhMOG 1–125 in 11 DRB1*1501 positive MS patients, 10 DRB1*1501 positive healthy controls, 7 DRB1*1501 negative MS patients, and 10 DRB1*1501 negative healthy controls are indicated. Data represent percentage of reactivity per group toward the respective Ags. Reactivity per patient and per Ag was scored positive if the number of spots per determination was higher than twice the individual background reactivity. Repeated measures ANOVA on signed ranks were used with the Dunnett’s test for each of the four groups independently (DRB1*1501 positive MS patients, DRB1*1501 positive controls, DRB1*1501 negative MS patients, DRB1*1501 negative controls). A value of p < 0.05 indicates statistical significance after multiple comparisons within each group. In general, the intracellular part of MOG 123–218 was more immunogenic than the extracellular part (MOG 1–122) and controls showed a broadened reactivity profile toward MOG compared with MS patients. The numbering of peptides refers to the MOG sequence. M indicates rhMOG 1–125. The procedures were performed as described in Materials and Methods.
T cell repertoire analysis to the transmembrane and intracellular part of MOG
Compared with the extracellular part of MOG, more reactive T cells were present to the transmembrane and intracellular part of MOG in MS patients and controls (Table I⇑, Figs. 1⇑ and 2⇑). DRB1*1501 positive controls reacted strongly to several peptides predominantly in the region MOG 125–160 (peptides MOG 125–140, p < 0.05; MOG 129–144, p < 0.05; MOG 133–148, p < 0.05; MOG 137–152, p < 0.05; MOG 141–156, p < 0.05; and MOG 145–160, p < 0.05), region MOG 169–188 (peptide MOG 169–184, p < 0.05; MOG 173–188, p < 0.05), peptide MOG 181–196 (p < 0.05), peptide MOG 189–204 (p < 0.05), and peptide MOG 201–216 (p < 0.05). In contrast, HLA-DRB1*1501 positive patients showed a focused response toward MOG region 141–160 (peptides MOG 141–156, p < 0.05; MOG 145–160, p < 0.05). HLA-DRB1*1501 negative controls showed responses to MOG peptide 145–160 (p < 0.05), MOG region 169–188 (peptides MOG 169–184, p < 0.05; MOG 173–188, p < 0.05), MOG peptide 189–204 (p < 0.05), and MOG peptide 201–216 (p < 0.05). Like HLA-DRB1*1501 positive MS patients, HLA-DRB1*1501 negative patients had a focused response to MOG region 141–160 (peptides MOG 141–156, p < 0.05; MOG 145–160, p < 0.05). MS patients and controls showed quantitatively dominant reactivities to MOG region 141–160 (peptides MOG 141–156 and MOG 145–160) with up to 423 IFN-γ-secreting cells per 2 × 105 PBMC (0.2%) (Fig. 1⇑, individual quantitative data shown for one HLA-DRB1*1501 positive patient). All investigated subjects showed T cell responses to PHA of at least 1000 spots per 2 × 105 PBMC (data not shown).
MOG 146–154 is the dominant T cell determinant
We analyzed the region within MOG 141–160 more in detail with overlapping nine amino acid-long N-acetylated C-amidated peptides (Table II⇑, Fig. 3⇓). The N-acetylated nine amino acid-long peptides bind to MHC class II molecules, but not to MHC class I molecules due to the N-acetyl group. This analysis revealed a dominant T cell response toward peptide MOG 146–154 in HLA-DRB1*1501 positive MS patients and controls (both p < 0.05). There were broader T cell reactivities in HLA-DRB1*1501 negative patients (peptides MOG 144–152, p < 0.05; MOG 145–153, p < 0.05; and MOG 146–154, p < 0.05) and controls (peptides MOG 144–152, p < 0.05; MOG 145–153, p < 0.05; MOG 146–154, p < 0.05; and MOG 148–156, p < 0.05), indicating that the T cell epitopes might be slightly shifted within MOG 141–160 depending on the HLA molecules present. To exclude an influence of disulfide bonds or posttranslational modifications of cysteine (C), we investigated the T cell responses with a modified peptide containing valine (V) instead of cysteine (C) in position 148 (Ac-LVFLCLQYR-NH2→Ac-LVFLVLQYR-NH2). There was no decreased reactivity to this peptide in MS patients or controls (data not shown).
Determination of the main T cell determinant within MOG 142–160. Reactivities of PBMCs secreting IFN-γ in ELISPOT assay toward nine amino acid-long peptides of the main immunogenic MOG region MOG 142–160 are shown in seven HLA-DRB1*1501 positive MS patients, nine HLA-DRB1*1501 positive controls, eight HLA-DRB1*1501 negative MS controls, and nine HLA-DRB1*1501 negative controls. Data represent reactivity per patient toward the respective Ags. Reactivity per patient and per peptide was scored positive if the number of spots per determination was higher than twice individual background reactivity (light gray boxes), triple individual background reactivity (gray boxes), four times individual background reactivity (dark gray boxes), or five times individual background reactivity (black boxes). Repeated measures ANOVA on signed ranks were used with the Dunnett’s test for each of the four groups independently (DRB1*1501 positive MS patients, DRB1*1501 positive controls, DRB1*1501 negative MS patients, DRB1*1501 negative controls). A value of p < 0.05 indicates statistical significance after multiple comparisons within each group. MOG 146–154 was the dominant T cell epitope in DRB1*1501 positive MS patients and controls. Non-HLA-DRB1*1501 MS patients and controls showed a more promiscuous reactivity pattern. MS patients reacted to three peptides and controls to four peptides. The numbering of peptides refers to the MOG sequence. The procedures were performed as described in Materials and Methods.
T cell reactivity to influenza A peptide
To exclude a generally decreased immune reactivity in MS patients, we measured T cell reactivities to an influenza A peptide (YRNLVWFIKKNTRYP) (18), with ELISPOT assay for IFN-γ secreting cells. MS patients and controls reacted similarly to this peptide arguing against a generally compromised immune status in MS patients (NS) (data not shown).
Phenotypic analysis of MOG 146–154 reactive cells
To determine the phenotype of IFN-γ-secreting MOG 146–154 reactive T cells, we performed enrichment of these cells and subsequent FACS analysis for expression of CD4 or CD8 and IFN-γ. MOG 146–154-specific T cells expressing IFN-γ were enriched by the IFN-γ enrichment assay in three MS patients (two HLA-DRB1*1501 positive, one HLA-DRB1*1501 negative) and one control (HLA-DRB1*1501 positive). The MOG 146–154-reactive IFN-γ-secreting T cells were of the CD4 phenotype (Fig. 4⇓).
MOG 146–154 IFN-γ-secreting T cells are of the CD4 phenotype. Cells secreting IFN-γ after contact with MOG 146–154 on MHC class II molecules of APC were enriched by IFN-γ secretion assay based on an affinity matrix on the cell surface that binds secreted IFN-γ. Subsequently, cells were detected by a secondary PE-labeled Ab against IFN-γ and then separated from cells not having secreted IFN-γ by PE-labeled microbeads in a magnetic field. Isolated cells were stained for IFN-γ and CD4 or CD8 expression and analyzed by FACS. MOG 146–154 reactive IFN-γ-secreting cells of all investigated MS patients (n = 3) and a control (n = 1) were of the CD4 phenotype. The procedures were performed as described in Materials and Methods.
Restriction analysis of MOG 146–154-reactive cells
The restriction pattern of MOG 146–154 reactive T cells was assessed in a HLA-DRB1*1501 positive and a negative MS patient and a HLA-DRB1*1501 positive control. The T cell response was HLA-DR restricted in the investigated HLA-DRB1*0701 DRB1*15011 DQB1*02× DQB1*0602 MS patient and in the HLA-DRB1*1103 DRB1*15011 DQB1*0301 DQB1*0602 control as well as in a HLA-DRB1*0701 DRB1*0407 DQB1*02× DQB1*0301 MS patient, because only anti-HLA-DR Abs (Tü 36) resulted in decreased T cell responses as compared with addition of anti-HLA-DQ Abs (Genox) (Fig. 5⇓).
Restriction analysis by ELISPOT for IFN-γ-secreting cells toward MOG 146–154. The restriction of the T cell response toward MOG 146–154 is shown for three individuals. The T cell response toward MOG 146–154 was HLA-DR restricted in a HLA-DRB1*0701 DRB1*15011 DQB1*02× DQB1*0602 MS patient (upper panel) and in the HLA-DRB1*1103 DRB1*15011 DQB1*0301 DQB1*0602 control (middle panel), as well as in a HLA-DRB1*0701 DRB1*0407 DQB1*02× DQB1*0301 MS patient (lower panel). Blocking studies were performed with the mAbs Genox (anti-HLA-DQ) and/or Tü 36 (anti-HLA-DR). Collection of cells, blocking studies, and evaluation of data were performed as described in Materials and Methods.
Binding of MOG 141–160-derived peptides to purified HLA-DRB1*1501 and HLA-DRB1*0401 molecules
We assessed binding affinities of MOG 141–156, MOG 145–160, the complete set of single stepped MOG peptides covering the region MOG 141–160, MBP 85–99, and influenza A peptide to purified HLA-DRB1*1501 and HLA-DRB1*0401 molecules. MOG 141–156 and MOG 145–160, as well as influenza A peptide and MBP 85–99 bound to both alleles. The strength of the binding for these peptides differed in both alleles with HLA-DRB1*1501 binding all four peptides with a higher affinity as indicated by a lower IC50 and a higher 1/IC50 (Fig. 6⇓). Large differences emerged in the binding patterns of the shorter N-acetylated C-amidated nine amino acid-long stepped MOG peptides. HLA-DRB1*1501 molecules bound all these peptides with low to high affinity, depending on the peptide sequence (Fig. 6⇓). In contrast, HLA-DRB1*0401 only bound peptides MOG 144–152, MOG 145–153, and MOG 146–154 with a weak affinity (Fig. 6⇓). Within the set of overlapping single stepped nine amino acid-long MOG peptides, intermediate to high affinity binding values were obtained for MOG 145–153, MOG 146–154, MOG 148–156, MOG 150–158, and MOG 151–159 for HLA-DRB1*1501 molecules. The highest affinity values for HLA-DRB1*0401 molecules, representing weak affinities, were obtained for MOG 144–152, MOG 145–153, and MOG 146–154. Taken together, this data indicates allele-specific differences in the binding of MOG peptides to HLA-DRB1*1501 and HLA-DRB1*0401 molecules with more peptides binding with a higher affinity to HLA-DRB1*1501 molecules. MOG 146–154 had similar to slightly better affinity than MBP 85–99 for HLA-DRB1*1501 or DRB1*0401 molecules. MBP 85–99 had been shown to be the major HLA-DRB1*1501-restricted T cell epitope in MS patients, but is also recognized by T cells in context with several other HLA-DR molecules (19, 20, 21, 22, 23).
In vitro peptide binding profiles of MOG 141–160-derived peptides to HLA-DRB1*1501 and HLA-DRB*0401 molecules. Binding affinities of MOG 141–156 and MOG 145–160, MBP 85–99, influenza A peptide, and nine amino acid-long MOG peptides of MOG sequence MOG 142–160 were measured to purified HLA-DRB1*1501 and HLA-DRB1*0401 molecules with a competitive binding assay and detection by the fluorochrome europium. The number of binding MOG peptides and the strength of binding differed for HLA-DRB1*1501 and HLA-DRB1*0401 molecules. MOG 141–156, MOG 145–160, MBP 85–99, and influenza A peptide bound to both molecules with varying affinities. For all nine amino acid-long peptides covering sequence MOG 142–160, binding was detected to HLA-DRB1*1501 molecules (lower panel). In contrast, HLA-DRB1*0401 molecules bound MOG 144–152, MOG 145–153, and MOG 146–154 (upper panel). MOG 146–154 bound equal or slightly better to HLA-DRB1*1501 or HLA-DRB1*0401 molecules compared with MBP 85–99. The procedures were performed as described in Materials and Methods.
Discussion
This study demonstrates for the first time that 1) there is a dominant MOG epitope recognized by CD4+ T cells within the intracellular part of MOG comprising aa 146–154 in MS patients and controls; 2) MS patients show a more focused T cell reactivity pattern toward MOG compared with controls; 3) the dominant MOG T cell epitope is promiscuously recognized in context of several MHC class II molecules; and 4) the intracellular part of MOG is much more immunogenic compared with the extracellular part.
In animal models with marmosets, rats, and mice, the extracellular part of MOG leads to a very MS-like disease (7, 24, 25, 26) (Fig. 7⇓). In addition, Amor et al. (25) investigated the encephalitogenic potential of peptides derived from the transmembrane and intracellular part of MOG in Biozzi AB/H and SJL mice. Our study is the first systematic study investigating T cell responses in MS patients and controls to peptides of the complete MOG sequence. Interestingly, besides the T cell responses to MOG 81–96, only minor T cell responses to the extracellular part of MOG were detectable. In contrast, several stretches of the transmembrane/intracellular domain of MOG proved to be highly immunogenic. We defined an immunodominant T cell epitope encompassing MOG 146–154 with N-acetylated and C-amidated nine amino acid-long peptides that can bind to MHC class II molecules. There was no major HLA-guided influence on the selection of T cells arguing for promiscuous presentation of this MOG peptide on MHC class II molecules. In the tested MS patients and controls, the MOG 146–154-specific T cell response was HLA-DR restricted. We assessed in vitro binding affinities of MOG peptides spanning MOG sequence 141–160 to two purified HLA-DR molecules. Both HLA-DRB1*1501 as well as HLA-DRB1*0401 molecules bound MOG 146–154 to a varying degree allowing presentation of this peptide to T cells. Interestingly, HLA-DRB1*1501 molecules bound more of the nine amino acid-long MOG peptides within the sequence MOG 142–160 compared with HLA-DRB1*0401 molecules. Both alleles bound MOG 146–154 slightly better than MBP 85–99 that is the major HLA-DRB1*1501-restricted MBP stretch in MS patients (19, 20, 21, 22, 23).
Schematic view of the MOG molecule. MOG is exposed on the extracellular surface of the myelin sheath. The extracellular part has an Ig-like structure and composes aa 1–122. It is glycosylated at position 31. The transmembrane part is composed of aa 122–150 and the cytoplasmatic domain of aa 150–218. Major T cell reactivity was detected against the transmembrane and cytoplasmatic domain in MS patients and controls. MOG 146–154 was dominantly recognized in MS patients and controls. MOG model according to Della Gaspera et al. (11 ).
Our results of promiscuous recognition of MOG 146–154 by T cells in context of several MHC class II molecules are partly in line with T cell reactivity to MBP in MS patients and controls. The immunodominant MBP 84–102 and MBP 87–106 peptides and their core sequence MBP 89–99 are recognized in the context of several MHC molecules in man as well as in mice and rats (19, 20, 21, 22, 23, 27). This MBP stretch induces experimental autoimmune encephalomyelitis (EAE) in HLA-DR2 transgenic mice (28). In future studies, our laboratories will assess the immunogenic and encephalitogenic potential of intracellular MOG determinants in animal models.
Epitope mapping studies in humans with regard to myelin-Ags have shown conflicting results as far as reactivity patterns of T cell responses in diseased individuals compared with controls are concerned. Some studies have shown that there is an increased frequency of autoreactive T cells from blood in diseased individuals and that T cell reactivity is increased in MS patients, while others recorded negative findings in regard with disease-associated differences in blood (12, 23, 29). Especially IFN-γ-secreting cells have been shown to be elevated in diseased individuals in blood and cerebrospinal fluid (30). Our study is partly contradictory to these observations because controls showed even more T cell reactivities in blood ex vivo compared with MS patients. Importantly, we did not find a generally compromised immune status in regard to PBMC reactivity in MS patients that would have explained this finding, because T cells from blood reacted similar to an influenza A peptide in MS patients and controls (18).
We did not observe the same reactivity profile to the extracellular part of MOG as Wallström et al. (12) had demonstrated with MOG 63–87 being immunodominant. Reasons might lie in differences in the genetic background of the patients and the preceding Ag exposure with effects on the T cell repertoire in the Swedish compared with the German MS population (31). Additionally, compared with the explicit strength of the T cell responses to MOG 81–96 in some MS patients and controls and intracellular MOG in our study, in this preceding study the T cell responses as assessed by IFN-γ ELISPOT were very weak with 3–4 spots in mean per 1 × 105 PBMC to the immunodominant MOG 63–87. In this study, beside MOG 63–87 also MOG 76–100 was slightly more recognized by MS patients compared with controls (1–2 spots per 1 × 105 PBMC) (12). The peptide MOG 79–96 has been shown to induce severe disease in DBA/1 (H-2q) mice (32) and mild disease in LEW.1AV1 (RT1av1) and LEW.1N (RT1n) rats (26). Interestingly, we did not find T cell responses in humans to peptides MOG 89–104, MOG 93–108, MOG 97–108, and MOG 97–112. These peptides contain the major encephalitogenic MOG stretches MOG 92–106 in SJL/J (H-2s) mice (25), MOG 97–108 in HLA-DRB1*0401 transgenic mice (33), and MOG 91–108 in different inbred rat strains, bearing the RT1a, RT1av1, and RT1n haplotypes, respectively (26). Nor did we detect strong T cell responses to peptides MOG 33–48 and MOG 37–52 that contain the encephalitogenic MOG determinant MOG 35–55 of C57BL/6 (H-2b) and NOD/LT (H-2g7) mice (6). In Biozzi AB/H (H-2dq1) mice, MOG 134–148 induced mild signs of EAE in one of five mice (25). The same authors did not detect encephalitogenic sequences in Biozzi AB/H (H-2dq1) and SJL (H-2s) mice to overlapping peptides covering MOG 141–218 (25). The data underscores the need to evaluate encephalitogenic MOG responses in humanized animal models additionally to inbred mouse and rat strains (34).
In MOG-induced EAE in LEW.1N rats, we have recently shown that MHC class II-regulated CNS autoaggression and T cell responses in peripheral lymphoid tissues are dissociated (26). The major encephalitogenic MOG 91–108 peptide in LEW.1N rats did not induce a detectable proliferative response and T1 or T2 T cell response in lymph nodes or spleen ex vivo after active immunization with MOG 91–108 or MOG 1–125. Instead, in the target tissue, the CNS, strong cellular, and cytokine responses were present after immunization with this peptide (26). In contrast, there were several determinants within MOG 1–125 that did raise strong T cell responses in peripheral lymphoid tissue, but did not induce disease (26). These data point to the need to investigate disease-associated cellular responses within the target organ (35). Unfortunately, investigations of the intra-CNS immune response in humans are difficult to perform because the availability of cerebrospinal fluid is limited and there are a number of ethical problems.
There was a focusing of the T cell reactivity patterns toward MOG 141–160 in MS patients. The decreased number of T cell determinants in MS patients could reflect alteration of the T cell repertoire by preceding Ag exposure after damage of the blood brain barrier leading to a state of peripheral T cell tolerance to certain MOG determinants (36). Moreover, migration of encephalitogenic T cells in MS patients to the target organ and reduction of the size of the T cell pool in blood reactive with MOG cannot be excluded at present (3, 37). Such a scenario would argue for a pathogenic role of the T cell reactivities that are absent or reduced in the blood of MS patients compared with controls in the disease process and again forces investigations at the target organ site. Alternatively, other functional differences might exist between the MOG reactive T cells secreting IFN-γ of MS patients and controls. Moreover, regulatory mechanisms could be up-regulated in MS patients after establishment of disease reducing the reactivity profile of T cells to MOG determinants in blood and lymphoid organs like up-regulation of TGF-β (38) or increased NK cell reactivity (39).
Our data indicate that during thymic selection, MOG 146–154 cross-reactive T cells must be positively selected on several HLA class II molecules and not deleted. In EAE, it has been recently shown that intrathymic expression of myelin components can lead to tolerance (40). Compared with other proteins of the myelin sheath, the expression of MOG is low because it comprises only 0.01% of the protein content of the myelin sheath (6). To date, no expression of human MOG on non-CNS tissue including the thymus has been reported, but it is not excluded that MOG is expressed at a very low level in the thymus. Potentially, this absent or low level expression would allow escape of MOG-specific T cells from negative selection in the thymus (40). The extracellular part of MOG shows homology to butyrophilins and B7 family members that are expressed in the thymus (41). In rats, butyrophilins have been shown to be cross-reactive with the extracellular part of MOG (42). In contrast, structural similarities between self-proteins and the transmembrane or intracellular domain of MOG have not been described to the same degree as to the extracellular part of MOG. As a consequence, central tolerance mechanisms might result in more effective negative selection of T cells reactive to the extracellular part of MOG compared with the intracellular part. This would explain the high number of IFN-γ-secreting cells against the intracellular part of MOG present in the periphery.
Footnotes
-
↵1 This study was supported by the Interdisciplinary Center of Clinical Research Tübingen (Bundesministerium für Bildung und Forschung Fö. 01KS9602) and Sonderforschungsbereich 510 (project D6 to R.W.).
-
↵2 Address correspondence and reprint requests to Dr. Robert Weissert, Department of Neurology, University of Tübingen, Hoppe-Seyler-Strasse 3, 72076 Tübingen, Germany. E-mail address: robert.weissert{at}uni-tuebingen.de
-
↵3 Abbreviations used in this paper: MS, multiple sclerosis; CM, complete medium; EAE, experimental autoimmune encephalomyelitis; EDSS, expanded disability status scale; MBP, myelin basic protein; MOG, myelin oligodendrocyte glycoprotein; rhMOG, recombinant human MOG.
- Received June 18, 2001.
- Accepted April 22, 2002.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
