Characterization of the human myelin oligodendrocyte glycoprotein antibody response in demyelination

Abstract

Over recent years, human autoantibodies targeting myelin oligodendrocyte glycoprotein (MOG Ab) have been associated with monophasic and relapsing central nervous system demyelination involving the optic nerves, spinal cord, and brain. While the clinical relevance of MOG Ab detection is becoming increasingly clear as therapeutic and prognostic differences from multiple sclerosis are acknowledged, an in-depth characterization of human MOG Ab is required to answer key challenges in patient diagnosis, treatment, and prognosis. Herein, we investigated the epitope, binding sensitivity, and affinity of MOG Ab in a cohort of 139 and 148 MOG antibody-seropositive children and adults (n = 287 patients at baseline, 130 longitudinal samples, and 22 cerebrospinal fluid samples). MOG extracellular domain was also immobilized to determine the affinity of MOG Ab. MOG Ab response was of immunoglobulin G1 isotype, and was of peripheral rather than intrathecal origin. High affinity MOG Ab were detected in 15% paediatric and 18% adult sera. More than 75% of paediatric and adult MOG Ab targeted a dominant extracellular antigenic region around Proline42. MOG Ab titers fluctuated over the progression of disease, but affinity and reactivity to Proline42 remained stable. Adults with a relapsing course intrinsically presented with a reduced immunoreactivity to Proline42 and had a more diverse MOG Ab response, a feature that may be harnessed for predicting relapse. Higher titers of MOG Ab were observed in more severe phenotypes and during active disease, supporting the pathogenic role of MOG Ab. Loss of MOG Ab seropositivity was observed upon conformational changes to MOG, and this greatly impacted the sensitivity of the detection of relapsing disorders, largely considered as more severe. Careful consideration of the binding characteristics of autoantigens should be taken into account when detecting disease-relevant autoantibodies.

Keywords: Antibody; Diagnosis; Epitope, antigen conformation; Multiple sclerosis; Myelin oligodendrocyte glycoprotein; Optic neuritis.

Fig. 1

Human native-MOG Ab response in demyelinating disorders. a Paediatric (n = 139) and adult (n = 148) sera from patients with demyelinating disorders were positive for native MOG Ab by live flow assay (upper row). Most MOG Ab were of the IgG1 isotype (lower row). Based on level of native-MOG Ab (dotted blue line), patients had high (top 33% of all patients), mid (middle 33% of all patients), or low (lowest 33% of all patients) native-MOG Ab titers. All controls were negative (grey). Dotted-black line indicates positivity threshold (mean of control sera +3SD). Native-MOG Ab positivity is shown between brackets. b There was a strong positive correlation between the ΔMFI of serum diluted at 1:50 and MOG Ab titers represented by dilution end-point (DEP) in children (n = 58, P < 0.0001, R² = 0.852) and adults (n = 89, P < 0.0001, R² = 0.793). c CSF native-MOG Ab were detected in 10 seropositive children (12 samples) and 9 adults (10 samples). 10/12 paediatric and 8/10 adult CSF MOG Ab titers were lower than in matched serum (P = 0.111 and P = 0.429, respectively) when ΔMFI values were normalized. Two children had slightly elevated levels in CSF compared to serum, and two adults had significantly higher native-MOG Ab titers in CSF than serum (filled-blue). d Distribution of seropositive native-MOG Ab among paediatric and adult clinical phenotypes. Relapsing ADEM is multiphasic ADEM according to [30]. e Regardless of disease course, patients with BON had significantly higher titers of native-MOG Ab than UON patients (P = 0.01). BON patients also exhibited higher MOG Ab titers than UON patients when sera were collected at disease onset (monophasic and relapsing UON, n = 27; monophasic and relapsing BON n = 21) (P = 0.040). Ab- = antibody negative, Ab+ = antibody positive, ADEM = acute disseminated encephalomyelitis, relapsing ADEM* = multiphasic ADEM, BON = bilateral optic neuritis, CSF = cerebrospinal fluid, CTL = controls, DEP = dilution end-point, LETM = longitudinally-extensive transverse myelitis, MFI = median fluorescence intensity, sTM = short transverse myelitis, UON = unilateral optic neuritis

Fig. 2

Human MOG Ab binding is influenced by conformational changes to native MOG. a and b High surface expression of native-MOG was comparable to formaldehyde-fixed MOG by immunocytochemistry (a, scale bar = 20 μm) and flow cytometry (b). c 78/139 (56%) native-MOG Ab+ children and 79/148 (53%) adult sera were positive for fixed-MOG Ab, whereas fixed-MOG Ab were not detected in all paediatric (n = 24) and adult (n = 24) control sera, and all native-MOG Ab- paediatric (n = 24) and adult (n = 24) sera (grey). Dotted line represents the positivity threshold (mean of controls +3SD). Fixed-MOG Ab positivity is shown between brackets. d In the fixed flow assay, 59% of children (36/61) and adults (41/69) who failed to bind fixed-MOG (filled-red) had low native-MOG Ab titers. A conformational insensitivity was observed in 10/46 children (22%) and 8/49 adults (16%) who could bind fixed-MOG despite low native-MOG Ab titers (empty-blue in low category). However, a sensitivity to conformation was seen in 25/61 (41%) children and 28/69 (41%) adults had mid to high native-MOG Ab titers. Number and percentage of patients negative in the fixed flow assay are shown in brackets. e Fixed-MOG Ab titers were higher in BON than UON patients (P = 0.01). f Adults seronegative by fixed flow assay predominantly presented with monophasic and relapsing UON (52%, 25/48), and BON (27%, 13/48), and 51% of non-binders had a relapsing course. Relapsing ADEM is multiphasic ADEM according to [30]. Unknown or other phenotypes (n = 16) and phenotypes with less than two patients are not represented (n = 5). g In the fixed biochip assay, reduced assay sensitivity was observed in 11/19 children (58%) and 13/24 adults (54%) who could not bind fixed-MOG in the fixed biochip assay (filled- red) and had low titers of native-MOG Ab. Conformational sensitivity was observed in 8/19 (42%) children and 11/24 adults (46%) who were negative by fixed biochip assay despite mid to high native-MOG Ab titers, and conformational insensitivity in 13/24 children (34%) and 7/20 (35%) adults positive in the fixed biochip assay but with low native-MOG Ab (empty-blue in low category). h Most adults seronegative by fixed biochip assay presented with monophasic and relapsing UON (57%), followed by BON (28%), and 64% had a relapsing course. Unknown phenotypes (n = 5) and phenotypes with less than two patients are not represented (n = 5). Relapsing ADEM is multiphasic ADEM according to [30]. θ_MRW = mean residue weight ellipticity, Ab = antibody, native-MOG^1–117 = extracellular MOG, fixed-MOG = fixed MOG, fixed-CTL = fixed transduced control cells, MFI = median fluorescence intensity, UON = unilateral optic neuritis, BON = bilateral optic neuritis, ON mixed = combination of BON and UON, ON/TM = simultaneous ON and TM, LETM = longitudinally extensive transverse myelitis, relapsing ADEM* = multiphasic ADEM

Fig. 3

Proline42-binding MOG Ab are the dominant IgG in serum of children and adults with demyelinating disorders. a Circular dichroism spectra of extracellular MOG (native-MOG^1–117) and extracellular P42S (native-P42S^1–117) domains showed characteristic minima at 215 nm, indicating β-sheet folding in both antigens. b and c High surface expression of native-P42S was comparable to native-MOG by immunocytochemistry (b) and flow cytometry (c). Scale bar = 20 μm. d 118/139 children (85%) and 113/148 adults (76%) harbored native-P42 Ab (orange), whereas 11/139 children (8%) and 21/148 (14%) adults had native-S42 Ab (blue). 10/139 paediatric (7%) and 14/149 (9%) adult sera recognized an unknown epitope on native-MOG (grey). Dotted line represents the control reference range determined by age-matched controls (n = 24 children, n = 24 adults). Number and percentage of patients in each epitope category are shown. e There was a positive correlation between native-P42 Ab titers in CSF and matched-serum in six children (diamonds) and six adults (triangles) (P = 0.006, R² = 0.545). f Most patients with monophasic UON and BON recognized native-P42 (100%, 20/20 and 83%, 19/23, respectively, orange), whereas relapsing ON patients had lower titers of native-P42 Ab and were more likely to recognize an epitope outside P42 (blue or grey, P = 0.03, left). Percentage and number of native-P42 binders in each phenotype group are shown. 75% of adult patients with an epitope outside P42 (non-P42, n = 28) presented with a relapsing course (right). g Among patients with native-P42 and fixed-MOG Ab, most paediatric and adult sera recognized the P42 epitope in native-P42 and fixed-P42 conditions (77%, 53/69 children; 80%, 52/65 adults), whilst a group lost their binding to the P42 epitope when MOG was fixed (23%, 16/69 children; 20%, 13/65 adults). θ_MRW = mean residue weight ellipticity, Ab = antibody, BON = bilateral optic neuritis, CSF = cerebrospinal fluid, MFI = median fluorescence intensity, native-MOG^1–117 = extracellular MOG, native-MOG = native human MOG, native-CTL = native transduced control cell, native-P42 Ab = native Proline42 binding MOG Ab, ON mixed = combination of BON and UON, S42 = Serine42, UON = unilateral optic neuritis

Fig. 4

The titer, but not epitope, of human MOG Ab changes over time. a Compared to the sample at baseline, MOG Ab titers, represented by DEP, decreased by more than 30% (blue-filled) in 13/19 paediatric (68%) and 14/32 adult sera (44%), and increased by more than 30% (green-filled) in 4/19 children (21%) and 6/32 adults (19%) for up to 4.7 years (median 17.6 months, IQR 3.6–28.9) and 9.1 years (median 4.1 months, IQR 1.7–13.9), respectively. Dotted lines represent native-MOG Ab titer (100%) at the baseline sample of each patient. b Native-MOG Ab immunoreactivity toward P42 did not change regardless of initial native-P42 Ab titers in all 17 paediatric and 24/27 adult (89%) native-P42 Ab-seropositive patients for up to 4.7 and 9.1 years, respectively. Two adults with high titers of native-P42 Ab decreased (blue), and one adult developed high immunoreactivity to native-P42 (green). c and d Sera collected during active disease (n = 13 paediatric and n = 12 adult samples, red) and disease remission (n = 15 paediatric and n = 12 adult samples, blue) were available in 10 children and 10 adults. c Native-MOG Ab titers were higher during acute samples than remission samples (P = 0.006, children; P = 0.004, adults). d Among 9 children and 7 adults recognizing native-P42, 8/9 children and 6/7 adults had slightly weaker immunoreactivity to native-P42 during remission (children, P = 0.317; adults, P = 0.226). Ab = antibody, DEP = dilution end-point, native-MOG = native human MOG, native-P42 Ab = native Proline42 binding MOG Ab

Fig. 5

A small population of children and adults harbor high affinity MOG Ab. a 20/134 children (15%) and 26/145 adult sera (18%) bound to native-MOG^1–117 and had high affinity MOG Ab, whereas no controls (grey), and one native-MOG Ab- child and one adult were positive for native-MOG^1–117 Ab (grey). Dotted line represents the positivity threshold and native-MOG^1–117 Ab positivity is shown between brackets. b Presence of high affinity native-MOG^1–117 Ab (purple) were not associated with native-MOG Ab titers, with similar detection of high affinity MOG Ab (purple) in children and adults across low, mid, and high native-MOG Ab titers. c High affinity MOG Ab was sustained for up to 1.6 years in 5/6 children (median disease duration 3.4 months, IQR 1.6–5.8) and 4.9 years in all five adults (median disease duration 5.1 months, IQR 4.2–13.7), whereas high affinity MOG Ab disappeared in one child at 20.8 months. Ab- = antibody negative, Ab+ = antibody positive, CTL = controls, native-MOG^1–117 Ab = high affinity MOG Ab binding to extracellular MOG, native-MOG = native human MOG, OD = optical density

Fig. 6

Major MOG Ab binding Patterns in children and adults with demyelinating disorders. a Native-MOG Ab binding was divided into eight patterns characterized by Proline42 epitope (orange), conformational sensitivity (red), and affinity (purple). MOG Ab responses from paediatric (42%) and adult (34%) sera targeted P42 and included conformation insensitive low affinity Ab (Pattern 1), followed by 30% of children and 28% of adults with similar low affinity response against Proline42, but were conformation-sensitive (Pattern 3). Responses targeting an epitope outside of Proline42 (light-orange) were of low affinity but were more sensitive to conformational changes to MOG (lighter red). Color intensity decreases with lower frequency of patients in each category (5 children and 3 adults not tested for affinity). b Association between clinical course and MOG Ab binding Patterns. Adults with a monophasic disease course were likely to present Pattern 1 (P = 0.028). Patterns 5 and 7 were more likely to include adults with a relapsing course (P = 0.035). Scale bar denotes frequency of patients in each binding Pattern. Affinity was not assessed in 2 monophasic (1 child, 1 adult), 5 relapsing patients (4 children, 1 adult), and one adult with unknown phenotype. Ab = antibody