Myelin oligodendrocyte glycoprotein
Recombinant mouse MOG
Recombinant human MOG
Spinal cord homogenate
Central nervous system
Normal mouse serum
Myelin oligodendrocyte glycoprotein (MOG) is a powerful encephalitogen for experimental autoimmune demyelination. However, the use of MOG peptides or recombinant proteins representing part of the protein fails to fully address the possible pathogenic role of the full-length myelin-derived protein expressing post-translational modifications. Immunization of mice with central nervous system tissues from wild-type (WT) and MOG-deficient (MOG–/–) mice demonstrates that MOG in myelin is necessary for the development of chronic demyelinating experimental autoimmune encephalomyelitis (EAE) in mice. While immunization with WT spinal cord homogenate (SCH) resulted in a progressive EAE phenotype, MOG–/– SCH induced a mild self-limiting acute disease. Following acute EAE with MOG–/– SCH, mice developed T cell responses to recombinant mouse MOG (rmMOG), indicating that MOG released from myelin is antigenic; however, the lack of chronic disease indicates that such responses were not pathogenic. Chronic demyelinating EAE was observed when MOG–/– SCH was reconstituted with a dose of rmMOG comparable to MOG in myelin (2.5% of total white matter-derived protein). These data reveal that while immunization with the full-length post-translational modified form of MOG in myelin promotes the development of a more chronic autoimmune demyelinating neurological disease, MOG (and/or other myelin proteins) released from myelin during ongoing disease do not induce destructive autoimmunity.
Immunization of susceptible mouse strains with the major myelin proteins proteolipid protein (PLP) and myelin basic protein (MBP) or their encephalitogenic peptide sequences induces an experimental neurological disease of the central nervous system (CNS), experimental autoimmune encephalomyelitis (EAE), an animal model of the human disease multiple sclerosis (MS). More recently, minor myelin antigens such as myelin oligodendrocyte glycoprotein (MOG) 1, myelin-associated glycoprotein 2, and oligodendrocyte-specific protein 2, as well as myelin-associated antigens αB-crystallin 3 and other CNS antigens such as amyloid-β 4, have also been shown to be encephalitogenic with varying degrees of neurological severity and histopathology.
MOG is a highly conserved, quantitatively minor constituent of CNS myelin 5 located on the outermost lamellae of the myelin sheath 6 and on the surface of mature oligodendrocytes 7, thereby representing a highly susceptible target for autoimmune attack. The failure to detect MOG protein within the thymus has been described to result in an absence of self-tolerance and, as a result, release of a repertoire of MOG-reactive lymphocytes into the periphery 8, 9.
Immunization with MOG induces inflammatory T cells in both susceptible rodent strains 1, 10, 11 and out-bred non-human primates 12, 13, resulting in an experimental model that resembles many of the clinical and histological features of MS. Furthermore, while antibodies specific for several myelin components are able to enhance myelin uptake by macrophages, only antibodies to MOG enhance EAE and demyelination in mice 14, 15.
While the encephalitogenic potential of synthetic preparations of MOG, i.e. MOG peptides and recombinant forms of MOG, has been defined, the pathological significance relative to the myelin proteins expressed within whole spinal cord homogenate (SCH) has remained unclear. Recently, the generation of MOG null (MOG–/–) mice that express no overt phenotype has been described 9. These mice are susceptible to myelin-induced EAE but are resistant to disease following immunization with recombinant rat MOG 9.
In the current study, we have used CNS tissues derived from MOG–/– and WT mice to investigate the role of full-length myelin-derived ‘native’ MOG, incorporating post-translational modifications, in chronic neurological demyelinating disease. The clear difference between the disease induced with MOG–/– compared to WT tissue demonstrates for the first time that the relatively low levels of MOG in myelin are sufficient to have a major impact on the course of disease and resultant pathology. Moreover, the fact that Biozzi ABH mice immunized with MOG–/– CNS tissues develop T cell responses to MOG after acute neurological disease demonstrates broadening of the myelin-specific immune repertoire to MOG during ongoing neurological disease, yet the failure of the mice to develop chronic disease indicates that such responses are not pathogenic.
MOG in myelin is necessary for chronic relapsing EAE
To examine the role of the full-length myelin-derived MOG for EAE induction, Biozzi ABH (H-2dq1) mice were immunized with SCH from C57BL/6 WT mice or from mice lacking the MOG gene (MOG–/–). Immunization with SCH from WT mice consistently induced more severe clinical signs of acute EAE on post-sensitization day (PSD) 16–17 compared to mice immunized with MOG–/– SCH (p<0.05). Furthermore, while mice immunized with WT SCH developed chronic disease (Fig. 1), those immunized with MOG–/– SCH recovered from the acute phase and did not show any further signs of clinical disease (PSD 21–45, p<0.05). Similarly, SJL and C57BL/6 mice immunized with WT SCH developed more severe disease in the acute phase compared to those immunized with MOG-deficient SCH (data not shown).
To determine whether development of the chronic disease was merely a function of exhibiting severe disease during the acute phase, we compared WT SCH- and MOG–/– SCH-immunized Biozzi mice that reached a score ⩾3 for neurological symptoms during the acute phase of EAE (see Materials and methods). It was observed that, while this grade of paralysis was observed in both groups, only those in the WT SCH-immunized group developed chronic disease, suggesting that the significant differences observed between PSD 21 and PSD 45 were not a direct consequence of the earlier EAE severity.
MOG in myelin is essential for demyelination in chronic relapsing EAE
In the spinal cords of Biozzi ABH mice immunized with WT SCH, the histological features during the acute (PSD 10–17), remission (PSD 26–29), and relapse/chronic (PSD 51–56) phases of EAE were essentially similar to that previously reported 16. Briefly, in acute disease severe inflammation and minimal demyelination was observed (Fig. 2A, B), which resolved during the remission phase. During the relapse phase, in which several WT SCH-immunized mice developed chronic disease, significantly higher levels of inflammation (Fig. 2A, C; p<0.05) and extensive demyelination (Fig. 2B, D; p<0.01) were observed as compared to MOG–/– SCH-immunized mice (Fig. 2A, B, E, F).
In a further study, Biozzi ABH mice were immunized with myelin purified from the spinal cords of WT (n=5) or MOG–/– (n=5) mice (data not shown). Disease induced by WT myelin was similar to that seen with WT SCH (mean score 3.7±0.2, mean onset 17.0±2.1 days). In contrast, MOG–/– myelin induced less severe disease (mean score 1.7±0.5; p<0.02). As with the SCH-immunized mice, inflammation and demyelination were prominent in the chronic phase of WT myelin-immunized mice (inflammation score 1.3±0.4, demyelination score 0.9±0.4), while mice immunized with MOG–/– myelin exhibited significantly less severe pathology (inflammatory score 0.1±0.1, demyelination score 0.0±0.0; p<0.05 and p<0.02, respectively).
Epitope spreading to rmMOG following MOG–/– SCH-induced EAE
As MOG is a quantitatively minor myelin component, we evaluated whether the MOG in myelin, or that released as a consequence of myelin damage, is sufficient to elicit T cell reactivity following acute EAE. Biozzi ABH mice were immunized with WT SCH or MOG–/– SCH, and T cell responses to rmMOG were measured on PSD 10 (prior to acute disease) or PSD 26 (after recovery from acute disease). As expected, mice immunized with MOG–/– SCH failed to generate MOG-specific T cell responses prior to the onset of neurological disease (Fig. 3), while three out of four mice immunized with WT SCH developed T cell responses to rmMOG at PSD 10. On PSD 26 (after acute EAE), T cell responses to rmMOG were detectable in both WT and MOG–/– SCH-immunized mice (Fig. 3). Clearly, these T cell reactivities are directed toward endogenous MOG released from myelin during neurological disease.
MOG constitutes 2.5% of the total protein content of myelin
To examine whether the amount of MOG in myelin is immunogenic and/or pathogenic, we first calculated the level of MOG in myelin using Western blot analysis (Fig. 4). These studies were performed on human myelin due to the quantities of myelin required. To quantify MOG as a component of myelin, increasing amounts of recombinant human MOG (rhMOG) were titrated into a fixed amount of whole myelin protein, and the mixtures were subjected to Western blot analysis using a monoclonal antibody that recognizes an epitope found on both the recombinant protein and MOG in myelin. In each sample the resulting two major bands representing the shorter recombinant version and full-length MOG were quantified by densitometric scanning. The amount of recombinant protein required to produce a signal equal to that of full-length MOG was used to define the absolute amount of the latter. Using the known amount of total myelin protein used to prepare each sample as a reference, the weight percentage of MOG in myelin was calculated.
‘Physiological’ doses of rmMOG induce EAE
Based on the level of MOG in myelin, it was estimated that 1 mg SCH, the dose used to immunize mice, contains between 3 and 4 μg MOG. This dose was calculated from a) the amount of myelin isolated from spinal cord tissue (weight/weight), i.e. 40–50%, b) amino acid analysis of myelin showing that it contains 20–25% protein 17, and c) data in the present study whereby 2.5% of the protein in myelin is MOG. To determine if this level of MOG is encephalitogenic, Biozzi ABH mice were immunized with between 1 and 500 μg rmMOG (the higher doses represent doses generally used in EAE studies). It was observed that at doses higher than 10 μg rmMOG, the majority of mice exhibited chronic relapsing EAE (Table 1). Immunization with 5 μg rmMOG also induced EAE in 3/5 mice, of which 3/3 developed relapses. Likewise, while only 1/5 mice immunized with 2 μg rmMOG developed EAE, this mouse also exhibited a relapse at a comparable time point to mice immunized with higher doses (Table 1). Histological examination of mice immunized with 2 or 5 μg rmMOG that developed EAE revealed inflammation and demyelination in the spinal cord. As controls, immunization with varying doses (3–500 μg) of an unrelated recombinant protein, rMRF-4, did not induce disease or pathology in the CNS of Biozzi mice. This is the first report showing that low doses of MOG, equivalent to the levels in myelin, induce both clinical and histopathological signs of EAE.
|rmMOG dose (μg)||Acute||Relapse|
|No. EAE||Group score||Mean EAE scoreb)||Onset dayc)||No. EAE||Group score||Mean EAE scoreb)||Onset dayc)|
Antibodies to MOG peptides and rmMOG are absent in WT and MOG–/– SCH-immunized mice
To determine if mice immunized with SCH or low doses of rmMOG develop serum antibody responses to rmMOG or overlapping MOG peptides, ELISA were performed. While antibody responses to MOG peptides and rmMOG were observed in mice immunized with high doses (100 μg) of rmMOG (16/23), only 1/10 mice immunized with 2–5 μg developed antibody responses. However, immunization with either WT (n=22) or MOG–/– (n=24) myelin or SCH did not evoke antibody responses to either rmMOG or MOG peptides in any mouse.
Reconstitution of MOG–/– SCH with 5 μg rmMOG induces chronic EAE and demyelination
To investigate whether reconstitution of MOG–/– SCH with rmMOG at a dose equivalent to that in myelin could restore the chronic neurological deficits seen in WT SCH-induced EAE, mice were immunized with MOG–/– SCH to which either 3 μg or 5 μg rmMOG was added. As before, WT SCH-immunized mice developed severe chronic EAE, while mice immunized with MOG–/– SCH developed a significantly milder disease between PSD 13 and 17 (p<0.05). Addition of 3 μg rmMOG to MOG–/– SCH induced disease similar to MOG–/– SCH, but the peak severity was comparable to EAE induced by WT SCH. When MOG–/– SCH was reconstituted with 5 μg rmMOG, the disease was similar to that of WT SCH-immunized Biozzi ABH mice, in which the severity of the clinical disease in the acute phase was significantly higher (p<0.05) on PSD 13 and 17 as compared to immunization with MOG–/– SCH alone (Fig. 5A). During chronic EAE (PSD 22–30), the severity of disease following immunization with WT SCH or MOG–/– SCH plus 5 μg rmMOG was significantly higher than that observed with either MOG–/– SCH alone or MOG–/– SCH supplemented with 3 μg rmMOG (p<0.05). Histological signs of EAE following immunization with MOG–/– SCH plus 5 μg rmMOG were similar to those in mice immunized with WT SCH (Fig. 5B), while the CNS pathology in mice immunized with MOG–/– SCH plus 3 μg rmMOG was similar to that in mice immunized with MOG–/– SCH only (Fig. 5B). T cell proliferative responses to rmMOG were absent on PSD 10 following immunization with MOG–/– SCH alone or MOG–/– SCH plus 3 μg rmMOG (Fig. 5C). In contrast, such responses were observed in mice immunized with WT SCH or MOG–/– SCH plus 5 μg rmMOG. Following recovery from acute EAE (PSD 30), mice immunized with either MOG–/– SCH alone or MOG–/– SCH plus 3 μg rmMOG developed proliferative response to rmMOG, while the responses in the other two groups increased compared to PSD 10.
Phagocytosis is affected by the absence of MOG
MOG-specific monoclonal antibodies have previously been shown to enhance myelin phagocytosis 14 and augment EAE severity 15. To determine if the chronicity of disease observed in mice immunized with WT SCH or myelin is a function of enhanced macrophage phagocytosis of myelin, we performed a macrophage phagocytosis assay using the macrophage cell line J774.2. It was observed that in the presence of either normal mouse serum (NMS) or heat-inactivated serum as well as under serum-free conditions, the phagocytosis of WT myelin was consistently higher as compared to phagocytosis of MOG–/– myelin (***p<0.001) (Fig. 6). Similar data were obtained using bone marrow-derived macrophages (data not shown).
Using CNS tissues from MOG-deficient and WT mice for the induction of EAE, we demonstrate an important role of myelin-derived full-length ‘native’ MOG during the development of chronic disease and demyelination in mice. Biozzi ABH (H-2dq1) mice immunized with SCH or myelin developed less severe neurological signs of acute EAE and CNS pathology when MOG was absent from the complex CNS antigen mixture. Furthermore, the mice immunized with MOG–/– SCH developed T cell responses to rmMOG following acute EAE, indicating that MOG released from myelin was antigenic but failed to induce chronic disease. This is the first report that definitively demonstrates broadening of the myelin-specific immune repertoire to MOG during ongoing neurological disease.
The heterogeneity of the clinical and pathological features in chronic relapsing EAE and MS may result from determinant spreading, where CNS antigens released through myelin damage induce a broadening of the (auto)immune repertoire. While determinant spreading has yet to be convincingly demonstrated in MS, spreading of the responses has been reported in chronic relapsing EAE 18–25. Such demonstration is crucially important to examine the impact of determinant spreading on strategies for antigen-specific intervention in chronic demyelinating disease. The generation of MOG-specific T cell responses following MOG–/– SCH-induced neurological damage provides definitive evidence that myelin damage leads to release of MOG and that this ‘native’ MOG is immunogenic. However, in this model, it appears that the immunogenicity is not pathogenic, since mice do not develop progressive disease. In line with this finding, Biozzi ABH mice infected with Semliki Forest virus develop T cell responses to myelin proteins known to induce chronic relapsing EAE 1, 2, 26, 27, but these animals fail to develop clinical disease (S. Amor, unpublished data). Furthermore, while Biozzi ABH mice with chronic relapsing EAE induced by MOG peptide 8–21 developed T cell responses to other encephalitogenic myelin epitopes, ‘tolerance’ to MOG 8–21 prevented further relapses (P. Smith, unpublished data), suggesting that immune responses to other myelin epitopes do not significantly contribute to further episodes of disease. It is probable that while myelin-specific T cells are generated, they are kept under control either by the CNS micro-environment or by conditions in local lymph nodes 28, 29.
In contrast to the study by Linares et al. 30, our data demonstrate that Biozzi ABH mice show a marked difference in EAE severity following immunization with WT compared to MOG-deficient myelin. This discrepancy is likely due to differences in the immunization protocol and adjuvant composition. The significant differences observed between WT and MOG–/– SCH-induced EAE in Biozzi ABH mice indicate that immune responses to myelin-derived MOG play a major role not only in disease onset but also in determining the severity of chronic neurological deficits.
While robust T cell responses to rmMOG were observed following immunization with WT SCH and, after acute EAE, with MOG-deficient SCH, we were unable to demonstrate humoral responses to rmMOG. The fact that we previously detected antibody responses to recombinant rat MOG in Biozzi ABH mice 1 following immunization with SCH but could not detect antibodies specific for rmMOG (this study) emphasizes the importance of using syngeneic proteins as autoantigens. Moreover, glycosylation of a MOG peptide improved the detection of specific autoantibodies in the sera of MS patients 31, 32; thus, the failure to detect antibodies to rmMOG or MOG peptides in our study may be due to influences of post-translational modifications present on MOG in myelin within our spinal cord preparations. The role of the glycosylation of MOG in EAE is currently being addressed using glycosylated preparations of MOG. These data suggest a return to more immunologically complex antigens such as SCH and myelin to understand the role of myelin proteins in the MS and/or EAE process of myelin damage.
We have recalculated MOG to constitute approximately 2.5% of total white matter-derived protein, indicating that mice immunized with 1 mg SCH received 3–5 μg MOG. At this level rmMOG alone or added to MOG-deficient SCH induced chronic relapsing EAE and generated a similar T cell response to rmMOG as seen in WT SCH-immunized mice. Although the encephalitogenic potential of recombinant MOG has been established in several EAE models 1, 33–37, this report is the first to describe that rmMOG at equivalent levels to that in myelin is sufficient to induce chronic relapsing EAE.
We previously described how an antibody directed to MOG, but not antibodies specific for other myelin antigens, augmented phagocytosis of myelin as well as exacerbated clinical EAE, possibly by complement-dependent mechanisms 14. In this study macrophage uptake of MOG–/– myelin was clearly reduced, suggesting that the low clinical and histological signs of EAE in MOG–/– SCH- or myelin-induced EAE may be due in part to decreased macrophage activity, possibly via complement-dependent mechanisms. However, the discrepancy between in vivo and in vitro data indicates that other processes may also be relevant.
Taken together, these data confirm a disproportionate pathogenic role for MOG in the initiation and chronicity of EAE due to an augmented inflammatory and demyelinating immune response. Importantly, the role of MOG in inducing chronic relapsing disease does not appear to result from determinant spreading but is likely to involve other mechanisms yet to be unraveled.
Materials and methods
Biozzi ABH (H-2dq1) and C57BL/6 (H-2b) were bred from stock at the BPRC, The Netherlands and Imperial College London, UK. Mice with a null mutation in the MOG gene (MOG–/–) on the C57BL/6 background were obtained from Equipé de Neurogénétique Moléculaire, Université de Paris 9 and bred at Imperial College London or BPRC. According to laws on animal experimentation in both the UK and the Netherlands, the procedures of this study have been reviewed and approved by the respective committees. The housing, care, and biotechnical handlings were in conformity with guidelines set by the committees.
Recombinant MOG production
E. coli strain JM109 was transfected with the cDNA encoding N-terminal amino acids 1–116 of mouse MOG 38 or N-terminal amino acids 1–125 of human MOG and a 31-amino acid fusion protein sequence ligated to the pRSET A (Invitrogen, UK) expression vector, as previously described 39. A 56-amino acid recombinant protein, Xenopus myogenic regulatory factor-4 (MRF-4), expressed in an identical bacterial vector system, was used as a control protein 40.
SCH was lyophilized and reconstituted in PBS prior to use as previously described 1, 16. Myelin was purified from the spinal cords of WT and MOG–/– mice as previously described , and protein concentrations were determined using the Bradford technique. Overlapping MOG 22-mer peptides with 7-amino acid overlaps (based on the entire mouse sequence) were purchased from ABC Biotechnology (UK).
Calculation of MOG in white matter
The concentration of MOG in normal appearing white matter from human myelin isolated from the brains of MS donors was determined by Western blot analysis. Total protein was obtained by solubilization of the samples in 80% tetrahydrofurane: 20% water: 0.1% trifluoroacetic acid, and subsequent delipidation was performed by repeated ether precipitation. Defined amounts of rhMOG (1–125) were titrated into a fixed sample of white matter-derived total protein and subjected to SDS-PAGE. Western blotting was performed using monoclonal antibody Z12, which recognizes both normal full-length MOG and rhMOG. Since rhMOG differs in molecular weight from the natural forms of MOG, the different versions of MOG appear as separate bands on a Western blot. Protein signals were evaluated using densitometric scanning, and the point of signal equivalence for each version of MOG was used to calculate the amount of full-length MOG in myelin.
Induction of EAE
Mice were injected s.c. into two sites on the flanks with a sonicated emulsion consisting of either 1 mg SCH, 500 μg myelin, or doses of 1 μg to 500 μg rmMOG, dissolved in PBS and emulsified in complete Freund's adjuvant (CFA) on PSD 0 and 7 1. Pertussis toxin (200 ng) derived from Bordetella pertussis (Sigma-Aldrich, UK) was dissolved in PBS was administered i.p. on PSD 0, 1, 7, and 8 as previously described 26. As controls for the EAE and in vitro studies, mice were immunized with various doses of the MRF-4 protein or with CFA only. Mice were weighed and graded daily for neurological symptoms: 0, normal; 1, limp tail; 2, impaired righting reflex; 3, partial hind limb paralysis; 4, complete hind limb paralysis; 5, complete forelimb paralysis/moribund. A grade of 0.5 was given for signs of lesser severity. Statistical analysis of clinical scores was carried out using the Mann-Whitney Rank Sum test (SigmaStat).
T cell proliferation assays
Spleens were removed and a single-cell suspension prepared using Lympholyte-M (Cedar Lane Laboratories, USA). Cells (1×106/ml) were cultured in RPMI medium supplemented with 2% NMS, 2 mM L-glutamine, 100 IU/ml penicillin, 100 μg/mL streptomycin, 5 mM Hepes, and 5×10–5 M 2-mercaptoethanol with 20 μg/well mouse MOG peptides or rmMOG for 72 h. Proliferation was measured by the incorporation of [3H]-thymidine (Amersham Biosciences) following addition of 1 μCi/well during the last 18 h of culture and is expressed as mean counts per minute (cpm) ± SD of triplicate cultures. Stimulation indices (SI) were calculated as the proliferative response in the presence of antigen divided by the proliferative response in the absence of antigen.
Microlon plates (Greiner Bio-one, Germany) were coated overnight at 4°C with 10 μg/ml mouse MOG peptides, rmMOG, or MRF-4 protein in PBS. Plates were washed twice in PBS-Tween (PBS-T) and blocked for 1 h at 37°C with 2% BSA/PBS. After blocking, 100 μL diluted plasma (1:100) in 1% BSA/PBS were added and incubated for 2 h at 37°C. Plasma from naive mice (NMS) was used as a negative control. After washing in PBS-T, the plates were incubated for 1 h at 37°C with alkaline phosphatase-conjugated rabbit anti-mouse Ig (Dako, Denmark). The reaction product was visualized using p-nitrophenyl phosphate-Tris buffer (Sigma-Aldrich, UK) and the absorbance read at 405 nm. An absorbance above the mean plus three SD of the NMS reactivity against the peptides was taken as positive.
Flow cytometry analysis of myelin phagocytosis
Myelin prepared from the spinal cords of WT and MOG–/– mice was labeled with DiI 14 and washed three times in PBS. Fresh NMS was prepared from clotted peripheral blood, and heat-inactivated (Hi) serum was prepared by incubating it for 30 min at 56°C to inactivate the complement system. The mouse macrophage cell line J774.2 (ECACC, UK) was suspended in DMEM containing 5% Hi fetal calf serum, and 5×105 cells were incubated in 24-well plates overnight to allow adherence. Cells were incubated with DMEM only or DMEM containing either 5% non-Hi or Hi NMS. To each well, 20 μg myelin was added, and the macrophages were allowed to phagocytose for 1.5 h at 37°C. After incubation, cells were washed three times with DMEM to remove free myelin. Macrophages were detached from the plates using 4 mg/ml lidocaine (Sigma) in PBS (10 min, 37°C). Fluorescence intensity (FL2 channel), a measurement for binding and uptake of DiI-labeled myelin, was determined using a FACScan cytometer. The mean fluorescence intensity of phagocytosis in serum-free medium was set to 100%, and the effects of NMS and Hi NMS were calculated from this baseline. Significance was determined by the Student's t-Test (SigmaStat).
Brain and spinal cords fixed in 5% formal saline were processed for routine histology and 5-μm sections cut. The sections were stained with hematoxylin and eosin (H&E) to evaluate inflammatory infiltrates or Luxol fast blue/cresyl fast violet (LFB/CFV) to assess the degree of demyelination 27. An observer unaware of the treatment groups performed light microscopy to qualitatively examine the degree of inflammation and demyelination 15. Statistical analysis of histological scores was carried out using the Mann-Whitney Rank Sum test (SigmaStat).
This research was supported by grants from the Multiple Sclerosis Society of Great Britain and Northern Ireland and Stichting MS Research, The Netherlands.