Comprehensive analysis of B cell repopulation in ocrelizumab-treated patients with multiple sclerosis by mass cytometry and proteomics

Abstract

Ocrelizumab, an anti-CD20 antibody, depletes CD20⁺ B cells, which subsequently repopulate over months. Little is known about changes in other immune cell populations and molecular markers associated with B cell repopulation. Here, we performed a comprehensive characterization of immune cells from ocrelizumab-treated patients with multiple sclerosis (MS) using mass cytometry. About 50% of patients showed naive B cell repopulation after 6 months mainly with a transitional phenotype, whereas CD27⁺ memory B cells only rarely repopulated. This repopulation was associated with a reduction of memory T cells and activated myeloid cells, as well as reduced expression of activation/migration markers in both cell types. A plasma proteomics analysis identified proteins including TNFRSF13C, associated with B cell depletion and repopulation. Plasma levels of neurofilament light-chain protein declined after ocrelizumab treatment was not linked with B cell repopulation. These findings identify potential soluble markers for monitoring of ocrelizumab treatment in MS.

Keywords: Immune response; Immunology; Proteomics; Treatment.

Graphical abstract

Figure 1

B cell depletion and repopulation after ocrelizumab treatment (A) Schematic overview of two non-overlapping longitudinal cohorts of ocrelizumab-treated patients with MS. In cohort 1, whole blood samples were analyzed at three time points, i.e., before the first ocrelizumab infusion (Baseline, n = 31), 2 weeks (2 weeks, n = 30) and 6 months (6 months, n = 29) after the first infusion. In cohort 2, long-term effects were determined approximately 1.5 (1.5 years, n = 50), 2 (2.0 years, n = 45), and 2.5 (2.5 years, n = 25) years after the first infusion. Whole blood samples were analyzed using streamlined CyTOF analysis workflow. (B) UMAP projection, coloring indicates 1–18 clusters. The phenotype of each cluster is shown based on the median expression of selected markers. (C) UMAP plots showing the depletion of CD20⁺ B cell subpopulations (red circle) at 2 weeks and 6 months as well as 1.5 years, 2.0 years, and 2.5 years after first ocrelizumab infusion. (D and E) Dot plots demonstrating the depletion and repopulation of five CD20⁺ B cell clusters (D) and four CD20⁻ B cell subsets (E) over five time points. Each dot represents the value of one patient. The lines connect longitudinal data points from the same patient. Statistical significance was assessed using a linear mixed model with random effects (Patient_id) and fixed effects (time point). The Bonferroni method was used to control FDR. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 2

Phenotypic alterations of repopulating B cells after ocrelizumab treatment (A) Schematic overview of patients with and without B cell repopulation following ocrelizumab treatment (6 months: 16 out of 29 [55%]; 1.5 years: 31 out of 50 [62%]; 2.0 years: 30 out of 45 [67%], and 2.5 years: 14 out of 25 [56%]). (B) Histogram and pie charts showing changes in the proportion of 18 B cell clusters in patients with B cell recovery at different time points. Dots indicate mean, and error bars indicate standard error of mean. Coloring indicates 1–18 clusters. (C) Differentially expressed markers (with arcsinh transformation) of repopulating CD27⁻IgD⁺ naive (B12) and CD27⁻IgD⁻ double-negative (B14) B cell clusters at 6 months (n = 16), 1.5 years (n = 31), 2.0 years (n = 30), and 2.5 years (n = 14) compared to Baseline (n = 16). (D and E) showing significantly differential levels of distinct plasma proteins (analyzed by NULISA technology) between patients at baseline (n = 31) and 6 months (n = 29) (D) and between patients with and without B cell repopulation at 6 months (E). Volcano plots (left) illustrate the differentially expressed proteins, generated through Limma analysis. Proteins with significant difference after correction for multiple testing (Benjamini & Hochberg) are colored in red; proteins with a significant p value but a non-significant adjusted p value are shown in yellow, whereas all others are in black. Each dot represents one protein, and horizontal dashed line represents a p value threshold of 0.05. Bar plots (right) showing differentially expressed proteins with name. (F) Significant correlation between the proportion of B cell sub-cluster and NPQ of TNFRSF13C at Baseline (upper panel) and 6 months (lower panel). Nonparametric Spearman correlation test (r), two-sided. (G and H) Box plots showing levels of NfL in plasma of patients at baseline (n = 31) and 6 months (n = 29) (G) and between patients with and without B cell repopulation at 6 months (H). Each dot represents one patient. Statistical significance was determined using Mann-Whitney U tests.

Figure 3

B cell repopulation associates with compositional and phenotypic changes of T cell subsets (A) Phenotypic heatmap and tables of 18 defined cluster (left panel) identities depicting the proportion and median expression levels of selected markers for CD3⁺ T cells in (A). Heat colors of expression levels have been scaled for each marker individually (to the 1st and 5th quintiles) (black, high expression; white, no expression). (B) Differences in proportional changes of T cell sub-clusters in patients with and without B cell recovery at different time points, as compared with those without recovery. (C) Boxplots showing differences in marker expression of T cell subpopulations between patients with and without B cell recovery at different time points. Each dot represents one patient. Whisker plots show the min (smallest) and max (largest) values. The line in the box denotes the median. Statistical significance was determined using a two-stage step-up method of Benjamini, Krieger, and Yekutieli correction for cluster proportion (B) and Mann-Whitney U test for marker expression (C). ∗p < 0.05, ∗∗p < 0.01.

Figure 4

B cell repopulation associated with compositional and phenotypic changes in MNK cell and granulocyte subsets (A) Phenotypic heatmap of cluster identities depicting the median expression levels of selected markers for MNK cells. Heat colors of expression levels have been scaled for each marker individually (to the 1st and 5th quintiles) (black, high expression; white, no expression). (B) Boxplots showing differentially abundant NK cell clusters in patients with and without B cell repopulation at different time points as compared with patients without B cell repopulation. (C) Boxplots showing significant differences in marker expression of NK cell subpopulations between patients with and without B cell recovery at different time points. (D) Proportional differences in myeloid cell proportions between patients with and without B cell recovery at different time points. (E) Boxplots showing altered marker expression of myeloid cell subpopulations in patients with and without B cell repopulation. (F) Phenotypic heatmap of cluster identities depicting the median expression levels of selected markers for granulocytes. (G) Differences in proportion of granulocyte clusters in patients with B cell recovery at different time points as compared with patients without B cell repopulation. (H) Boxplots showing differences in marker expression of granulocyte subpopulations between patients with and without B cell recovery at different time points. Each dot represents one patient. Whisker plots show the min (smallest) and max (largest) values. The line in the box denotes the median. Statistical significance was determined using two-stage step-up method of Benjamini, Krieger, and Yekutieli correction for cluster proportion (B, D, and G) and Mann-Whitney U test for marker expression (C, E, and H). ∗p < 0.05 and ∗∗p < 0.01.