| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Neuroimmunology, Max-Planck-Institute for Neurobiology, Martinsried, Germany;
Department of Animal and Avian Sciences, University of Maryland, College Park, MD 20742;
Institut National de la Santé et de la Recherche Médicale Unité 546, Hôpital Pitié-Salpêtrière, Paris, France; and
Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
The MOG-specific Ab response is complex and recognizes both linear and conformation-dependent epitopes. However, the demyelinating component of this response is restricted to conformation-dependent epitopes present on the extracellular Ig-like domain of the protein (MOGIgd), whereas Abs recognizing linear MOGIgd peptides are unable to bind to the native protein and are unable to initiate demyelination in vivo (4, 5). The ability of MOG to trigger a demyelinating autoantibody response in EAE indicates that autoaggressive MOGIgd-specific B cell clones are not eliminated from the immune repertoire. This lack of tolerance is attributed to the localization of MOG within the immunologically privileged environment of the CNS, where it is sequestered from normal lymphocyte trafficking and therefore unable to trigger Ag-specific B cell tolerance (6). However, while MOG itself may not induce tolerance, a recent study identified a MOG-independent tolerogenic effect that influences the composition and pathogenicity of the MOG-specific B cell repertoire (7).
The MOG-specific hybridoma 8.18C5 secretes a conformation-dependent MOGIgd-specific mAb that can mediate demyelination in vivo in animals with EAE (4, 8). This Ab was derived from BALB/c (H-2d) mice immunized with rat CNS glycoproteins, and its particular combination of H and L chains is clearly permissible on this parental background (9). In transgenic C57BL/6 (H-2b) mice carrying the IgH chain of mAb 8.18C5, the transgenic H chain can pair with endogenous L chains to provide a functional MOG-specific B cell repertoire that differentiates without tolerogenic censure (6). However, attempts to replicate the fine specificity of 8.18C5 in C57BL/6 mice by simultaneously expressing both H and L chains were thwarted by the elimination of the transgenic L chain by receptor editing during development (7). This tolerogenic effect was also observed when the transgenic H and L chains were introduced into MOG-deficient mice, indicating that receptor editing was MOG independent and must therefore be mediated by some other self Ag.
We speculated that this effect might result in strain-specific differences in the pathogenicity of the MOG-specific Ab response. To test this hypothesis, we compared the B cell response to MOGIgd in different mouse strains using a combination of ELISA, FACS, and complement-dependent cytotoxicity assays to differentiate between effects on Ab responses to linear (nonpathogenic) epitopes and conformation-dependent (pathogenic) responses to the native protein. We now report that one or more genes encoded within the MHC selectively censor the ability of H-2b mice to mount a pathogenic conformation-dependent autoantibody response to MOG. This effect does not influence the ability of H-2b mice to generate T cell and Ab responses to linear MOG epitopes and is not mediated by MOG polymorphisms that might influence expression of the target autoantigen in immune tissues. This novel form of selective tolerance discriminates between conformational and linear epitopes and determines the ability of mice to mount a pathogenic autoantibody response to MOG.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
Female SJL (H-2s) and C57BL/6 (H-2b) mice were purchased from Charles River (Sulzfeld, Germany), and MHC congenic female mice C57BL/10 (H-2b) and B10.S (H-2s), BALB/c (H-2d), and BALB/b (H-2b) from Harlan Winkelmann (Borchen, Germany). F1 (SJL x C57BL/6) mice were bred in the animal facility at the Max-Planck-Institute for Neurobiology (Martinsried, Germany). C57BL/6 (H-2b) MOG knockout mice and sex-matched wild-type controls were bred in the Salpêtrière Hospital animal facility (Paris, France).
Antigens
The expression and purification of rat rMOGIgd were performed, as described previously (10). The exoplasmic domain of butyrophilin (BTNexo, aa 1–216) was cloned and expressed as a soluble C-terminal hexahistidine-tagged recombinant baculovirus product (Invitrogen, San Diego, CA) and purified from the supernatant of infected High Five insect cells (BD PharMingen, San Diego CA) by metal chelate chromatography. Peptides were purchased from Sigma/Genosys (Cambridge, U.K.).
Immunization protocols
Mice were immunized into the base of the tail with a single s.c. injection of an emulsion containing 100 µg Ag (OVA or rMOGIgd) in IFA, or with 100 µg MOGIgd in CFA (Life Technologies, Rockville, MD) supplemented with 4 mg/ml inactivated Mycobacterium tuberculosis (H37 RA; Difco Laboratories, Detroit, MI). Animals were monitored regularly for clinical signs of EAE (0, no clinical disease; 1, tail weakness; 2, paraparesis; 3, paraplegia; 4, paraplegia with forelimb weakness or paralysis; 5, moribund state or death). Moribund animals were sacrificed within 24 h in accordance with federal care regulations.
Construction of expression vectors and DNA vaccination
cDNA of murine MOG was obtained, as described previously, by reverse transcription and PCR amplification of RNA isolated from mouse brain using primers that generated a fragment containing the full-length MOG coding sequence together with its signal sequence (11). A chimeric BTN-MOG cDNA construct was generated by sequential PCR amplification using the murine MOG template described above and a bovine BTN cDNA. This chimeric construct encodes an optimized Kozak sequence, followed by aa 1–242 of bovine BTN, including the signal sequence and entire exoplasmic domain, fused in-frame to the transmembrane sequence and cytoplasmic tail of mouse MOG (aa 118–218). The PCR products were cloned into the expression vector pcDNA3.1- (Invitrogen, San Diego, CA) and plasmid DNA purified from the transformed Escherichia coli strain DH5
by miniprep (Qiagen, Hilden, Germany) and sequenced to verify integrity. Large-scale preparation of plasmid DNA was conducted using Qiagen Endotoxin-Free Plasmid Kits (Qiagen). Each preparation was checked by agarose gel electrophoresis, and DNA concentration was determined from its OD at 260 nm. DNA vaccination was performed, as described previously, with 100 µg plasmid DNA in PBS (1 mg/ml) injected in the tibialis anterior muscle (11). The expression vector pcDNA3.1 was used as control.
T cell proliferation and ELISA
A single cell suspension was prepared from the draining lymph nodes 10 days postinfection (p.i.) and cultured in DMEM supplemented with glutamine, sodium pyruvate, penicillin, streptomycin, nonessential amino acids, and 2-ME (complete DMEM) containing 1% mouse serum in the presence of rMOGIgd or 2.5 µg/ml Con A. Proliferation assays were performed using 2 x 105 cells/well in 96-well plates in a total volume of 200 µl. Ag-specific proliferation was assessed by [3H]thymidine incorporation for the last 16–18 h of a 3-day culture period using a Packard Matrix 96 Direct beta counter (Meriden, CT). For cytokine assays, 106 cells/well were cultured in 1 ml complete DMEM in 24-well plates, and the supernatant was collected after 48 h and stored at -80°C. Cytokine ELISA kits were purchased from Endogen (Biozol, Eching, Germany), and assay was performed according to the manufacturer’s instructions. All assays were performed in triplicate. Ab levels were determined by ELISA using serum obtained from peripheral blood and stored at -20°C. Ninety-six-well vinyl assay plates were coated with 10 µg/ml of rMOGIgd, BTNexo, or peptide in PBS. After blocking with 1% BSA, the assay plates were incubated with test serum. Alkaline phosphatase-conjugated goat anti-mouse IgG (Southern Biotechnology Associates, Birmingham, AL) was used to detect specific binding. Color was developed by the addition of p-nitrophenyl phosphate, and the OD was read at 405 nm.
FACS analysis
The expression vector pcDNA3.1 (Invitrogen) containing the complete mouse MOG cDNA was transfected into a mouse myeloma cell line (Ag8) to generate a MOG-expressing target cell line (Ag8/MOG), as described (11). Cells were maintained in complete DMEM with 10% FCS and 1 mg/ml G418. To detect Ab responses to native MOGIgd, transfected and untransfected cells were washed with PBS/1% FCS and incubated with test serum (1/30) for 1 h on ice. After washing, cells were stained with Cy2-labeled anti-mouse IgG (Dianova, Hamburg, Germany) for 1 h and washed again; propidium iodide was added to exclude dead cells; and cells were analyzed immediately by FACS (BD Biosciences).
Complement-dependent Ab-mediated cytotoxicity assay
Anti-MOG Ab-mediated cytotoxicity was determined by incubating in triplicate 2 x 105 Ag8 or Ag8/MOG cells/well in complete DMEM/1% FCS containing test serum at a dilution of 1/30 in the presence or absence of freshly reconstituted rabbit complement (Behring, Marburg, Germany) at a final dilution of 1/30. After incubation for 45 min at 37°C with 10% CO2, cell viability was determined using the tetrazolium salt WST-1 (Roche, Germany), and absorbance was read at 450 nm.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
To compare B cell responses to MOGIgd in H-2b (C57BL/6 and C57BL/10) and H-2s (SJL/J and B10.S) mouse strains, mice were immunized with rat rMOGIgd in IFA. Blood was collected 4 wk later, and the serum Ab response was analyzed by FACS to identify responses to the native murine protein expressed at the surface of a MOG-transfected mouse myeloma cell line. In addition, sera were analyzed by ELISA to determine the response to the recombinant Ag and to map its specificity using a panel of overlapping synthetic MOGIgd peptides.
FACS analysis revealed that H-2s mice mounted an Ab response to native, glycosylated MOGIgd determinants expressed on the surface of MOG-transfected cells in vitro, whereas this response was absent in MOGIgd-immunized H-2b mice (Fig. 1). In contrast, all four mouse strains were found to have mounted an Ab response to bacterially derived rat MOGIgd when assayed by ELISA, irrespective of their genetic background or MHC haplotype (Fig. 2A). Furthermore, epitope mapping using synthetic MOG peptides demonstrated that the Ab response to the recombinant protein recognized specific linear MOGIgd epitopes in all four strains (Fig. 2, B and C) (11). C57BL/6 and C57BL/10 mice both mounted a MOGIgd-specific Ab response to aa sequences 1–26 and 50–74, and in addition C57BL/6 mice also recognized aa 27–50. In B10.S mice, the specificity of the Ab response was as previously described for SJL mice and recognized epitopes present in peptides 1–26 and 50–87 (Fig. 2, B and C) (11). These results demonstrate that the absence of a response to the native protein in H-2b mice is not associated with a generalized lack of responsiveness to MOGIgd epitopes, but that H-2b mice are selectively unresponsive to those MOGIgd conformation-dependent epitopes implicated in the pathogenesis of Ab-dependent demyelination (4, 5).
|
|
The inability of MOGIgd-specific Abs to bind to the native extracellular domain of MOGIgd in H-2b mice indicates that these Abs are unlikely to be cytotoxic. This was confirmed using sera from MOGIgd-immunized C57BL/10 and B10.S mice in an in vitro complement-dependent cytotoxicity assay. As previously described for SJL mice, sera from MOGIgd-immunized B10.S mice mediate the Ag-specific, complement-dependent lysis of MOG-transfected Ag8 cells in vitro (Fig. 3A) (11). In contrast, sera obtained from MOGIgd-immunized C57BL/10 mice did not exhibit Ag-specific cytotoxicity (Fig. 3A). No significant difference in cell survival was observed with any sera when assay was performed with untransfected target cells or in the absence of complement (data not shown). Absence of complement-dependent cytotoxicity in H-2b mice was not due to the absence of a complement-fixing IgG2a Ab response, as immunization with MOGIgd in IFA induced MOGIgd-specific IgG1 and IgG2a Ab responses in both mouse strains. High levels of both isotypes were detected in B10.S mice, whereas in C57BL/10 mice we observed a bias toward Abs of the IgG1 isotype (Fig. 3B).
|
Selective loss of responsiveness to conformation-dependent B cell epitopes is not associated with a defect in the Th1 T cell response
Previous studies demonstrated that both H-2s and H-2b mice can mount an encephalitogenic response to MOG T cell epitopes, indicating that the inability of H-2b mice to mount a conformation-dependent and pathogenic Ab response was not associated with T cell tolerance (14). This was confirmed following analysis of the T cell response in the draining lymph nodes of C57BL/10 and B10.S mice 10 days after immunization with MOGIgd in CFA. In accordance with other studies, both strains supported a strong and quantitatively similar Th1 MOGIgd-specific T cell response as determined by both T cell proliferation and production of IFN-
(Fig. 4, A and B), but not of the Th2 cytokine IL-4 (data not shown). The effect of the H-2b haplotype on the conformational B cell response is therefore not associated with a haplotype-specific effect that influences the Th1 response to MOGIgd.
|
To rule out that immunization with bacterially derived nonglycosylated rat MOG rather than the native murine protein was not responsible for the failure of H-2b mice to mount a conformation-dependent response to MOGIgd, we investigated the Ab response to murine MOGIgd following vaccination with a DNA construct encoding the full-length mouse autoantigen. DNA vaccination leads to expression of Ag by host cells in its native conformation and with appropriate posttranslational modifications, and elicits Abs directed to native epitopes (15). As described previously, MOG-DNA vaccination induced a strong autoantibody response to MOGIgd in SJL mice, which was shown to be purely conformation dependent and cytolytic (Fig. 5A) (11). In contrast, MOG-DNA vaccination did not induce any detectable MOGIgd-specific Ab response in C57BL/6 mice (Fig. 5A), although vaccination with DNA encoding a control protein, bovine BTN, induced a strong Ab response in both C57BL/6 and SJL/J mice (Fig. 5B). Bovine BTN exhibits both structural and sequence homologies to MOGIgd (Fig. 6), indicating that the failure of C57BL/6 to mount a B cell response to MOGIgd is highly discriminating and Ag specific.
|
|
|
MOG is itself located within the telomeric region of the MHC, raising the possibility that unresponsiveness to MOG-DNA vaccination in H-2b mice was due to an H-2b-specific MOG polymorphism (16). We therefore investigated the Ab response induced by MOG-DNA vaccination in MOG-deficient MOG-/- H-2b mice and wild-type littermate controls. Abolition of MOG expression failed to rescue the Ab response to MOG-DNA vaccination, demonstrating that MOG itself is not responsible for this effect in H-2b mice (Table I).
MOG-DNA vaccination exacerbates EAE in SJL mice, but not in C57BL/6 mice
We previously demonstrated that the autoantibody response induced by MOG-DNA vaccination in SJL mice enhances the severity of EAE irrespective of the identity of the encephalitogen (11). We predicted that because C57BL/6 mice are nonresponders to MOG-DNA vaccination, we would not see a similar potentiation of disease in this strain. In SJL mice vaccinated with MOG-DNA, the onset of EAE induced by immunization with the proteolipid protein peptide 139–154 was 3–4 days earlier than in mice vaccinated with control DNA, and disease severity was increased (Fig. 7A). In contrast, prior vaccination with MOG-DNA had no significant effect on either onset or severity of EAE in C57BL/6 mice immunized with the encephalitogenic peptide MOG35–55 (Fig. 7B).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
Our analysis of H-2b mice actively immunized with MOGIgd supports this interpretation. The MOG-specific Ab response in the H-2b strains C57BL/6 and C57BL/10 only recognizes linear MOGIgd epitopes that are not accessible to Ab when the protein is in its native conformation. Consequently, H-2b-derived MOGIgd-specific Abs are incapable of activating complement at the membrane surface and lysing MOG-expressing target cells in vitro. This is in stark contrast to the MOGIgd-specific Ab response in H-2s (SJL/J, B10.S) and H-2d (BALB/c) mice that has a significant conformation-dependent and potentially cytopathic component.
The inability of C57BL/10 mice to mount a cytolytic Ab response to MOG may be one factor that contributes to their resistance to EAE induced by immunization with MOGIgd in IFA. This protocol is generally considered only weakly encephalitogenic, as it induces a bias toward a counterinflammatory Th2 T cell response, but nonetheless, this protocol will induce severe EAE in the context of an appropriate genotype (12, 13). In the current study, immunization with MOGIgd in IFA skewed Ig isotype usage in favor of a dominant Th2-associated MOGIgd-specific IgG1 response in C57BL/10 mice, suggesting a marked difference in the ability of B10.S and C57BL/10 mice to develop Th2 T cell responses to this autoantigen. Although complement fixation by MOG-specific Abs is isotype dependent, it has been demonstrated using a panel of mAbs that IgG1 responses to MOG can fix complement and mediate demyelination in vivo (8, 17). This suggests that the shift in Ig isotype usage in H-2b mice is unlikely to be responsible for the inability of this strain to mount a cytolytic Ab response to MOG. This MHC-dependent effect on the balance of Th1 to Th2 T cell responses to MOG is currently under investigation in our laboratories, and may provide a murine parallel of MHC-dependent effects on susceptibility to MOG-induced EAE in the rat (18, 19).
The mechanism that censors the conformation-dependent Ab response to MOG in H-2b mice is still a matter of speculation. The responsible genes are encoded within the MHC, the sequence, organization, and function of which are now understood in great detail. MOG was itself considered a candidate gene, as it is located within the MHC, and low levels of MOG transcription are reported to occur within primary and secondary lymphoid organs4 (3, 16, 20). These observations raised the possibility that strain-specific differences in MOG expression outside the CNS could influence the composition of the B cell repertoire. However, abolition of MOG expression in H-2b mice fails to restore the conformation-dependent Ab response, indicating that loss of responsiveness is not mediated by MOG itself. The MHC also encodes multiple gene products involved in Ag degradation, processing, and presentation. Polymorphic variation affecting these pathways could either result in the degradation of the target structure on the protein or generate a MOG-specific T cell repertoire that fails to provide efficient T cell help for the conformation-dependent Ab response in H-2b mice. Although these possibilities cannot be excluded at present, lack of T cell help appears unlikely, as the MHC congenic strains C57BL/10 and B10.S mount quantitatively and qualitatively similar T cell responses to MOG.
An alternative mechanism is that the presence of a conformational mimic of MOG encoded within the MHC mediates this tolerogenic event. B cell tolerance relies heavily upon the elimination of self-reactive B cells with defined specificities during development in the bone marrow and subsequently in the periphery during B cell differentiation. Negative selection is normally discussed in terms of the induction of tolerance by either clonal deletion, anergy, or receptor editing mediated through contact with the nominal autoantigen (21). However, in analogy to molecular mimicry in which structural and sequence homologies between foreign and self Ags can trigger autoimmunity and in some cases tolerance, it is possible that tolerance to one self Ag may be induced by cross-reactivity of the B cell receptor with another component of self (22, 23, 24). The demonstration that the transgenic L chain of a demyelinating MOG-specific mAb is eliminated by receptor editing in H and L chain double-transgenic B cells in MOG-deficient C57BL/6 mice suggests that cross-reactivity does occur (7).
Although the identity of this putative cross-reactive autoantigen is unknown, candidate Ags may be encoded within a cluster of BTN family genes located between the classical MHC class II and III regions of the murine MHC (25). The N-terminal Ig-like domain of these BTN-like proteins has a very high level of sequence and structural homology to MOGIgd. Moreover, the genes encoding these proteins are polymorphic, and preliminary data indicate that there are strain-specific differences in their expression patterns between H-2b and H-2s mice (M. Pagany, unpublished observations). Active immunization with bovine BTN itself induces an encephalomyelitis in the Dark Agouti rat due to molecular mimicry with MOG (26). This immunological cross-reactivity is, however, restricted to the T cell response to these Ags. BTN itself is not encoded within the murine MHC, indicating that this milk protein is not responsible for the effects described in this study. Whatever the mechanism responsible for censoring the pathogenic Ab response to murine MOG in H-2b mice, a recent study suggests that it is highly sensitive to amino acid substitutions that may subtly alter the conformation of the extracellular domain, as C57BL/6 mice can mount a pathogenic Ab response to the human protein (27). Murine and human MOGIgd differ at only 11 aa positions, including a S > P substitution at position 42 (Fig. 6). Which of these amino acid substitutions is responsible for the ability of the human protein to induce this pathogenic response has still to be determined. Although negative selection is an essential step in selecting the B cell repertoire, there is also evidence that immature B cells undergo positive selection dictated by the specificity of their Ig receptor (28). Using IgH chain transgenic mice, these authors find evidence of positive selection within a small fraction of the mature B cell receptor repertoire (28), a situation analogous to the selection of the T cell repertoire that also depends on a combination of positive and negative selection steps. Intriguingly, F1 (SJL/J x C57BL/6) mice develop MOG-specific Ab titers following MOG-DNA vaccination that are intermediate to those obtained in the parental responder (H-2s) and nonresponder (H-2b) strains, suggesting a gene dosage effect that could be compatible with either partial negative or enhanced positive selection.
In summary, using MOG as a model for B cell-mediated autoaggression in the CNS, we demonstrate the existence of a genetically determined mechanism that renders H-2b mice selectively unresponsive to pathogenic conformation-dependent B cell epitopes of MOG. This MHC-dependent effect is not associated with an effect on either the response to nonpathogenic linear MOG B cell epitopes or the MOGIgd-specific T cell response. This mechanism may account in part for the observation that conformation-dependent Ab responses to MOGIgd are only detectable in a small percentage of multiple sclerosis patients5 (29).
|
Note: |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Christopher Linington, Department of Neuroimmunology, Max-Planck-Institute for Neurobiology, Am Klopferspitz 18a, 82152 Martinsried, Germany. E-mail address: lining{at}neuro.mpg.de
3 Abbreviations used in this paper: EAE, experimental autoimmune encephalomyelitis; BTN, butyrophilin; BTNexo, butyrophilin exoplasmic domain; MOG, myelin oligodendrocyte glycoprotein; MOGIgd, MOG Ig domain; p.i., postinfection.
4 M. Pagany, M. Jagodic, C. Bourquin, T. Olsson, and C. Linington. Genetic variation in myelin oligodendrocyte glycoprotein expression and susceptibility to experimental autoimmune encephalomyelitis. Submitted for publication.
Received for publication December 19, 2002. Accepted for publication April 28, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote:References |
|---|
expression of encephalitogenic T cells. Eur. J. Immunol. 25:1951.[Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
