Association between B-cell depletion and attack risk in neuromyelitis optica spectrum disorder: An exploratory analysis from N-MOmentum, a double-blind, randomised, placebo-controlled, multicentre phase 2/3 trial

Abstract

Background: Inebilizumab is an anti-CD19 antibody approved for the treatment of neuromyelitis optica spectrum disorder (NMOSD) in adults with aquaporin-4 autoantibodies. The relationship between B-cell, plasma-cell (PC), and immunoglobulin depletion with longitudinal reductions in NMOSD activity after inebilizumab treatment was characterised post hoc in an exploratory analysis from the N-MOmentum study (NCT02200770).

Methods: Peripheral blood CD20+ B cells, PC gene signature, and immunoglobulin levels were assessed throughout N-MOmentum (follow-up ≥2.5 years); correlations with clinical metrics and magnetic resonance imaging (MRI) lesion activity were assessed.

Findings: Inebilizumab induced durable B-cell and PC depletion within 1 week versus placebo. Although no association was observed between B-cell counts at time of attack and NMOSD activity, depth of B-cell depletion after the first dosing period correlated with clinical outcomes. All participants receiving inebilizumab demonstrated a robust long-term therapeutic response, and participants with ≤4 cells/μL after the first 6-month dosing interval had persistently deeper B-cell depletion, lower annualised attack rates (estimated rate [95% CI]: 0.034 [0.024-0.04] vs 0.086 [0.056-0.12]; p = 0.045), fewer new/enlarging T2 MRI lesions (0.49 [0.43-0.56] vs 1.36 [1.12-1.61]; p < 0.0001), and a trend towards decreased Expanded Disability Status Scale worsening (0.076 [0.06-0.10] vs 0.14 [0.10-0.18]; p = 0.093). Antibodies to inebilizumab, although present in a proportion of treated participants, did not alter outcomes.

Interpretation: This analysis suggests that compared with placebo, inebilizumab can provide specific, rapid, and durable depletion of B cells in participants with NMOSD. Although deep and persistent CD20+ B-cell depletion correlates with long-term clinical stability, early, deep B-cell depletion correlates with improved disease activity metrics in the first 2 years.

Funding: Horizon Therapeutics (formerly from Viela Bio/MedImmune).

Keywords: Anti-CD19 monoclonal antibody; Aquaporin-4 antibody-positive neuromyelitis optica spectrum disorder; B-cell suppression; Devic disease.

Fig. 1

Participant flow in N-MOmentum. For analysis purposes, participants were divided into three groups: “RCP inebilizumab” (participants randomised to inebilizumab during the RCP), “RCP placebo” (participants randomised to placebo during the RCP), and “any inebilizumab” (participants who received inebilizumab at any point during the study, including those from the RCP inebilizumab group and all participants in the OLP; shaded in pink). OLP = open-label period, RCP = randomised controlled period.

Fig. 2

Kinetics of B-cell and immunoglobulin depletion during the study. (a) Total CD20+ B-cell counts. (b) Plasma-cell gene signature normalised to a pool of 10 HDs using the ddCT method. (c) Total immunoglobulin, (d) immunoglobulin G, (e) immunoglobulin M, (f) immunoglobulin A, and (g) immunoglobulin E concentrations over time during the RCP in placebo-treated (in grey) and inebilizumab-treated (in red) participants. Points show medians and shaded regions show IQRs. Dashed black lines show LLOQ of 0.2 cells/μL for FACS assays. Samples with counts below the LLOQ were imputed to 0.05 cells/μL. Mann–Whitney U test between dose groups; p > 0.05 where p not noted. Vertical dashed lines show when participants were transitioned to the OLP and the placebo cohort was administered two inebilizumab doses; the inebilizumab cohort was given an additional maintenance dose. ddCT = delta delta cycle threshold, FACS = fluorescence-activated cell sorting, FC = fold change, HD = healthy donor, IQR = interquartile range, LLOQ = lower limit of quantification, OLP = open-label period, RCP = randomised controlled period.

Fig. 2

Kinetics of B-cell and immunoglobulin depletion during the study. (a) Total CD20+ B-cell counts. (b) Plasma-cell gene signature normalised to a pool of 10 HDs using the ddCT method. (c) Total immunoglobulin, (d) immunoglobulin G, (e) immunoglobulin M, (f) immunoglobulin A, and (g) immunoglobulin E concentrations over time during the RCP in placebo-treated (in grey) and inebilizumab-treated (in red) participants. Points show medians and shaded regions show IQRs. Dashed black lines show LLOQ of 0.2 cells/μL for FACS assays. Samples with counts below the LLOQ were imputed to 0.05 cells/μL. Mann–Whitney U test between dose groups; p > 0.05 where p not noted. Vertical dashed lines show when participants were transitioned to the OLP and the placebo cohort was administered two inebilizumab doses; the inebilizumab cohort was given an additional maintenance dose. ddCT = delta delta cycle threshold, FACS = fluorescence-activated cell sorting, FC = fold change, HD = healthy donor, IQR = interquartile range, LLOQ = lower limit of quantification, OLP = open-label period, RCP = randomised controlled period.

Fig. 2

Kinetics of B-cell and immunoglobulin depletion during the study. (a) Total CD20+ B-cell counts. (b) Plasma-cell gene signature normalised to a pool of 10 HDs using the ddCT method. (c) Total immunoglobulin, (d) immunoglobulin G, (e) immunoglobulin M, (f) immunoglobulin A, and (g) immunoglobulin E concentrations over time during the RCP in placebo-treated (in grey) and inebilizumab-treated (in red) participants. Points show medians and shaded regions show IQRs. Dashed black lines show LLOQ of 0.2 cells/μL for FACS assays. Samples with counts below the LLOQ were imputed to 0.05 cells/μL. Mann–Whitney U test between dose groups; p > 0.05 where p not noted. Vertical dashed lines show when participants were transitioned to the OLP and the placebo cohort was administered two inebilizumab doses; the inebilizumab cohort was given an additional maintenance dose. ddCT = delta delta cycle threshold, FACS = fluorescence-activated cell sorting, FC = fold change, HD = healthy donor, IQR = interquartile range, LLOQ = lower limit of quantification, OLP = open-label period, RCP = randomised controlled period.

Fig. 3

Decreases in NMOSD progression observed with long-term inebilizumab treatment.(a) AAR. (b) Annualised rate of EDSS worsening. (c) Annualised rate of new/enlarging MRI T2 lesions (only participants randomised to inebilizumab are included). (d) Annualised NMOSD-related inpatient hospitalisation rate. Plots show rates in placebo-treated participants during RCP (in grey) and in all participants during the first dosing period of inebilizumab treatment (first dosing interval), and yearly afterwards (in blue). The final bar in each plot displays the combined rate of each endpoint after 2.5 years or more of continued inebilizumab treatment. Mean data with error bars showing 95% CI estimated by negative binomial regression. AAR = annualised attack rate, CI = confidence interval, EDSS = Expanded Disability Status Scale, MRI = magnetic resonance imaging, NMOSD = neuromyelitis optica spectrum disorder, RCP = randomised controlled period.

Fig. 3

Decreases in NMOSD progression observed with long-term inebilizumab treatment.(a) AAR. (b) Annualised rate of EDSS worsening. (c) Annualised rate of new/enlarging MRI T2 lesions (only participants randomised to inebilizumab are included). (d) Annualised NMOSD-related inpatient hospitalisation rate. Plots show rates in placebo-treated participants during RCP (in grey) and in all participants during the first dosing period of inebilizumab treatment (first dosing interval), and yearly afterwards (in blue). The final bar in each plot displays the combined rate of each endpoint after 2.5 years or more of continued inebilizumab treatment. Mean data with error bars showing 95% CI estimated by negative binomial regression. AAR = annualised attack rate, CI = confidence interval, EDSS = Expanded Disability Status Scale, MRI = magnetic resonance imaging, NMOSD = neuromyelitis optica spectrum disorder, RCP = randomised controlled period.

Fig. 4

Temporal relationship between CD20+ B-cell counts and attacks. Box plot of CD20+ B-cell counts in all samples drawn from participants who never experienced attacks, in samples drawn preceding attacks (sample draws 15–115 days preceding attacks), during attack assessment, and after attacks (sample draws 8–200 days after attack). “Other samples” refers to samples that were drawn from participants who experienced attacks, but not directly preceding the attack, directly after the attack, or during the attack. The bold bar shows the median, the box shows the IQR, and the whiskers show the range of data. IQR = interquartile range, LLOQ = lower limit of quantification, WB = whole blood.

Fig. 5

B-cell kinetics in the first dosing interval of inebilizumab predict long-term deep B-cell depletion. (a) Profile plot of CD20+ B-cell counts in individual patients throughout the first dosing interval with inebilizumab. (b) Total CD20+ B-cell counts over 3 years of inebilizumab treatment in patients with B-cell counts >4 cells/μL and ≤4 cells/μL at the end of the first dosing period. Points show medians ± IQR. Mann–Whitney U test was used to calculate p values. IQR = interquartile range, LLOQ = lower limit of quantification.

Fig. 6

Improvement occurs in all participant subgroups over time.(a) AAR. (b) Annualised rate of EDSS worsening. (c) Annualised rate of new/enlarging T2 MRI lesions (only participants randomised to inebilizumab are included). (d) Annualised inpatient hospitalisation rate split by patients with ≤4 cells/μL or >4 cells/μL after the first inebilizumab dosing interval (6 months after treatment). Plot shows rates in placebo-treated participants during the RCP (in grey), then in both inebilizumab-treated participant groups during the first dosing period of inebilizumab treatment (first dosing interval), and yearly afterwards (in blue and red). The final bars in each plot display the combined rate of each endpoint after 2.5 years or more of continued inebilizumab treatment. Mean data with error bars showing 95% CI estimated by negative binomial regression. AAR = annualised attack rate, CI = confidence interval, EDSS = Expanded Disability Status Scale, MRI = magnetic resonance imaging, RCP = randomised controlled period.

Fig. 6

Improvement occurs in all participant subgroups over time.(a) AAR. (b) Annualised rate of EDSS worsening. (c) Annualised rate of new/enlarging T2 MRI lesions (only participants randomised to inebilizumab are included). (d) Annualised inpatient hospitalisation rate split by patients with ≤4 cells/μL or >4 cells/μL after the first inebilizumab dosing interval (6 months after treatment). Plot shows rates in placebo-treated participants during the RCP (in grey), then in both inebilizumab-treated participant groups during the first dosing period of inebilizumab treatment (first dosing interval), and yearly afterwards (in blue and red). The final bars in each plot display the combined rate of each endpoint after 2.5 years or more of continued inebilizumab treatment. Mean data with error bars showing 95% CI estimated by negative binomial regression. AAR = annualised attack rate, CI = confidence interval, EDSS = Expanded Disability Status Scale, MRI = magnetic resonance imaging, RCP = randomised controlled period.