Antibodies to native myelin oligodendrocyte glycoprotein are serologic markers of early inflammation in multiple sclerosis

Chinese Hamster Ovary (CHO)-MOG Assay (MOG_cme) Validation.

Fig. 1 shows high levels of MOG_cme expression, as demonstrated by staining of MOG-transfected CHO cells with the monoclonal anti-MOG Ab 8-18C5. Detection of hMOG_cme-specific Abs with this cell-based assay was sensitive because a concentration of <1 ng/ml of 8-18C5 produced a binding ratio (BR) >1.5 (data not shown). Staining with a positive control serum (patient 1158) is shown in Fig. 1C. This control was used in each assay to normalize for interassay variability and minimize experimental errors such as variation in surface expression of MOG. The mean (± SEM) BR to hMOG_cme of this control from nine independent experiments was 1.96 ± 0.145 (Fig. 1D), and the interassay coefficient of variation was 22%. The intraassay coefficient of variation (quadruplicate) was 3.2%. In each assay analyzing human serum, the binding against nontransfected CHO (ntCHO) cells was used as background control.

Characterization of Exposed Epitopes of hMOG_cme.

We analyzed the binding properties of monoclonal, recombinant marmoset Fab Ab fragments produced against the nonglycosylated extracellular domain of recombinant rat MOG_aa1–125 (rMOG₁₂₅). Four Fabs, designated M3-8, M26, M3-24, and M3-31, were selected by their ability to recognize rMOG₁₂₅ in ELISA, and because they recognize distinct conformationally defined epitopes (12). The M3-31 and M26 Fabs strongly stained the MOG-transfected CHO cells identical to 8-18C5 (0.5 μg/ml) (Fig. 2 A and B). On the contrary, no binding was observed for the two other Fabs, M3-24 and M3-8 (Fig. 2 C and D). These results indicate that very specific epitopes of MOG are expressed on the MOG-transfected CHO cells and do not overlap with the other epitopes displayed by rMOG₁₂₅ on solid ELISA support. The observation that only three of five monoclonal reagents tested bind to the transfected cells also renders a nonspecific binding effect unlikely.

IgG Reactivity in Human Serum.

Compared with age-matched healthy controls (HCs), the titers of IgG directed against membrane-bound hMOG_cme were most significantly increased in CIS (P < 0.001). Increased titers were also present in RRMS and secondary progressive MS (SPMS) subtypes, compared with HC (P < 0.01 and P < 0.05, respectively) (Fig. 3). The differences were also significant when comparing primary progressive MS (PPMS) with all other subtypes [P < 0.001 (CIS), P < 0.01 (RRMS), and P < 0.05 (SPMS)]. No statistical difference was found between PPMS and HC (P not significant) or between the CIS, RRMS, or SPMS subtypes when paired comparisons were made. No treatment-related difference was found.

IgG Reactivity in Marmoset EAE.

Eleven C. jacchus marmosets immunized with human white matter were tested for plasma reactivity against hMOG_cme on MOG-transfected CHO cells. For these serial studies, the first time point of Ab detection (i.e., serum conversion) was compared with the appearance of the first clinical signs of EAE. Three animals (U30-00, UO61-02, and UO53-01) were killed before onset of neurological deficits (preclinical disease), but exhibited CNS inflammation and blood-brain-barrier breakdown as demonstrated by cerebrospinal fluid pleocytosis (mean cerebro-spinal fluid mononuclear cells = 173 per μl; range 80–340 per μl). Serum reactivity against hMOG_cme was consistently detected in the earliest blood sample obtained after immunization (mean ± SD = 14 ± 2 days; range = 13–18 days) (Table 1) and was clearly present in each animal before the appearance of any clinical sign (mean ± SD = 21 ± 9 days; range = 16–43 days). The difference between time of appearance of serum IgG reactivity to hMOG_cme and appearance of clinical signs was highly significant (P < 0.0001) (Fig. 4). Reactivity was not detected in preimmune plasma.

Comparison of Human IgG Binding on hMOG₁₂₅ and hMOG_cme.

The serum binding characteristics of the CIS cohort (n = 36) were tested by ELISA using recombinant human MOG_aa1–125 (hMOG₁₂₅) and compared with hMOG_cme reactivity by FACS on the MOG-transfected CHO cells. Although some CIS patients displayed high reactivity against hMOG₁₂₅, unlike for hMOG_cme reactivity, there was no statistical significant difference between CIS and HCs (data not shown). By linear regression and comparison test, results from these two assays showed no correlation (P not significant), indeed suggesting that different epitopes are detected by both assays (Fig. 5A).

Specificity of MOG and Differential MOG-Epitope Binding in Human Serum.

To discriminate the epitopes displayed by hMOG₁₂₅ from those displayed on hMOG_cme on MOG-transfected CHO cells, we performed a series of preabsorption experiments with two sera, both representative of early and inflammatory forms of MS: the positive control used in our cell-based assay (RRMS 1158) and a CIS patient displaying a high reactivity to both hMOG_cme and hMOG₁₂₅ (CIS 008). Preabsorption against ntCHO cells served as control in the hMOG_cme assay (FACS of MOG-transfected CHO cells), and preabsorption against 1% BSA served as a control for the hMOG₁₂₅ ELISA assay. Compared with these controls, preabsorption of either sample on hMOG₁₂₅ did not affect hMOG_cme reactivity. Preabsorption on hMOG_cme resulted in a decrease in BR (31% for CIS 008 and 24% for RRMS 1158), when tested with hMOG_cme expressed on the MOG-transfected cells (Fig. 5B). Similarly, when tested on hMOG₁₂₅, no change in reactivity occurred when samples were preabsorbed on the MOG-transfected cells (hMOG_cme). On the contrary, the samples preabsorbed on hMOG₁₂₅ displayed a decrease in BR, by 21% and 57%, respectively, when tested on hMOG₁₂₅ (Fig. 5C). These experiments unequivocally demonstrate that hMOG_cme and hMOG₁₂₅ display separate epitopes of the MOG protein.

Patients.

Ninety-two patients with clinically definite MS (Poser criteria) (24), 36 patients with CIS, and 37 HCs were recruited from the University of California, San Francisco MS Center, and the MS Center of Pamplona, Spain (CIS). All investigations were conducted according to the Declaration of Helsinki. Blood was obtained by venipuncture after informed consent in full compliance with the Institutional Review Board, and clotted serum was stored at −40°C until use. Patients were classified as RRMS (n = 35), SPMS (n = 33), and PPMS (n = 24) MS by clinical history (24). A CIS was defined by a first clinical event indicative of demyelination with no history of previous neurological symptom. Age, gender, disease duration, and disability state [Expanded Disability Status Score (EDSS) (25)] were recorded at time of sampling. HCs were chosen to match sex and age of the CIS group. The median age, disease duration, and EDSS were higher for the SPMS group than for the RRMS and CIS groups (Table 2). All RRMS and 19/33 SPMS patients were treated with IFN-β. Two of the PPMS patients were treated with mitoxantrone and monthly pulsed steroids.

Animals.

C. jacchus marmosets were cared for in accordance with the guidelines of the Institutional Animal Care and Usage Committee. EAE was induced by immunization with 100 mg of human white matter, which contains native, membrane-embedded MOG homogenized in complete Freund’s adjuvant as described (26). Plasma was obtained from EDTA-anticoagulated blood at baseline and 2- to 4-week intervals and stored at −40°C. The animals were scored every other day for the development of clinical signs by using a published scale (27).

Preparation of MOG-Transfected Cells.

CHO cells were transfected with a full-length construct corresponding to the major α-1 form of human MOG as described (28). CHO cells were cultured in T-225 flasks (Costar), in RPMI medium 1640 supplemented with 10% FCS, 1× Glutamax, 1 mM sodium pyruvate, and 50 μg/ml gentamycine. G418 (500 μg/ml, GIBCO) was added to the medium of transfected cells. Stable surface expression of MOG was verified by immunofluorescence and FACS on transfected cells after multiple passes and washes; cells were used for flow cytometry (FACS) assay when a confluence of 80–90% was reached. To check for surface expression of hMOG_cme, 3 × 10⁴ MOG-transfected CHO cells were deposited on a slide and fixed with 100% methanol for 5 min at −20°C. After blocking with PBS containing 2% BSA and 2% FCS for 30 min, cells were incubated 1 h at 37°C with the mouse monoclonal anti-MOG Ab 8-18C5 (5 μg/ml, gift of C. Linington, University of Aberdeen, Aberdeen, Scotland). Fluorescence was revealed after 1-h incubation at 37°C by a goat anti-mouse IgG FITC Ab (Sigma) and examined under a fluorescence microscope (Nikon Eclipse E600). Negative controls were done with secondary Ab alone.

Serum IgG Reactivity.

Cells were trypsinized, diluted in FACS buffer (PBS, 0.1% Na azide, and 2% FCS), and plated in a 96-well plate (Costar) at a density of 200,000 per well. After blocking in FACS buffer containing 10% FCS for 15 min at 4°C, cells were washed and human serum (1:10) was added for 1-h incubation at 4°C. After washing, cells were incubated with a goat anti-human IgG FITC (Caltag, South San Francisco, CA) at the recommended concentration for 30 min at 4°C. After a final wash, cells were resuspended in FACS buffer containing propidium iodide (Molecular Probes) at 2 μg/ml and gently shaken. Samples were kept on ice and analyzed by gating the selected live cell population (10⁴ cells) within 1 h of harvesting by trypsinization. Each FACS experiment included an internal positive control consisting of the monoclonal anti-MOG 8-18C5 Ab (0.5 μg/ml) and rabbit anti-mouse FITC (DAKO) as secondary Ab. For each sample, the geometric mean intensity (Gmean) of FITC (winmdi 2.8 software) was measured for MOG-transfected CHO cells and compared with that of ntCHO cells. The BR was calculated as the Gmean for MOG-transfected CHO cells divided by the Gmean for ntCHO. To compare different assays, for each sample the BR was normalized to that of a human positive control (RRMS 1158) included in each experiment.

For studies in marmosets, MOG-transfected CHO cells were incubated for 1 h at 4°C with marmoset serum diluted 1:100. FITC-conjugated Ab against whole monkey IgG (Sigma) diluted at 1:100 was used as secondary Ab and incubated 30 min at 4°C. FACS analysis was performed as described above. IgG binding against hMOG_cme was considered positive when the BR (Gmean preimmunization/Gmean time point tested) was >1.5.

Differential Reactivity of Monoclonal Fab Fragments.

Recombinant Fabs were derived from a C. jacchus marmoset immunized with rMOG_1–125 produced in Escherichia coli (rMOG₁₂₅) (12). Fabs were diluted in FACS buffer at 0.5 μg/ml and added to ntCHO or MOG-transfected CHO cells. FITC-conjugated Ab against whole monkey IgG (Sigma) diluted at 1:100 was used as secondary Ab for 30 min at 4°C. FACS was performed as described above.

ELISA Assay.

hMOG₁₂₅ expressed in E. coli was coated overnight on polystyrene microtiter plates at 0.5 μg per well (Maxisorb, Nunc). After washing and blocking with 1% BSA in PBS + 0.05% Tween (BSA-PBS-T) for 2 h at room temperature, sera (1:200) were diluted in BSA-PBS-T and added to the plate. Ab binding was detected by an alkaline phosphatase AP-labeled goat-anti-human IgG (Sigma) for 1 h at room temperature. Plates were developed with para-nitrophenyl phosphate (Moss, Pasadena, MD) for 30 min in the dark at room temperature and read at 405 nm in a microplate reader (SpectraMax, Molecular Devices). Results were expressed as BR, i.e., signal over BSA background.

Preabsorption of Sera on hMOG_cme and hMOG₁₂₅.

To further validate the cell-based assay and eliminate the possibility of nonspecific binding effects from either MS serum or transfection procedures, we conducted binding experiments after preincubation of serum with the respective antigens. For hMOG_cme preabsorption, 5 × 10⁶ MOG-transfected CHO cells and ntCHO cells were separately incubated with serum diluted 1:10 for 1 h at room temperature with gentle agitation. After centrifugation at 900 × g for 2 min, supernatant were collected and preabsorption was repeated three times in total with fresh cells. After the final preabsorption, supernatants were centrifuged at 3,600 × g for 5 min and collected for subsequent experiments.

For hMOG₁₂₅ preabsorption, ELISA plates were coated with either 1 μg BSA or 0.5 μg hMOG₁₂₅ overnight and blocked in 1% BSA in PBS plus 0.05% Tween for 2 h, then sera were incubated 1 h. Supernatants were collected and preabsorption was repeated eight times in total with fresh hMOG₁₂₅. After the final preabsorption step, supernatants were collected as above.

Data Analysis.

Statistical analyses were performed by using Kruskal–Wallis with Dunn’s post hoc test for multiple comparisons for interpatient group differences. Interassay correlation was analyzed with Spearman r correlation. Survival analysis for time-dependent variables was assessed by Kaplan–Meyer analysis and the Cox proportional hazard model (prism 3.0).