Highly Sensitive Flow Cytometric Detection of Residual B-Cells After Rituximab in Anti-Neutrophil Cytoplasmic Antibodies-Associated Vasculitis Patients

Abstract

Background: B-cell depletion with rituximab (RTX) is an effective treatment for anti-neutrophil cytoplasmic antibodies (ANCA)-associated vasculitis (AAV) patients. Nevertheless, relapses are frequent after RTX, often preceded by B-cell repopulation suggesting that residual autoreactive B-cells persist despite therapy. Therefore, this study aimed to identify minimal residual autoimmunity (MRA) in the B-cell compartment of AAV patients treated with RTX.

Methods: EuroFlow-based highly-sensitive flow cytometry (HSFC) was employed to study B-cell and plasma cell (PC) subsets in-depth in AAV patients before and after RTX treatment. Additionally, peripheral blood mononuclear cells (PBMCs) of these RTX-treated AAV patients were cultured and in vitro stimulated with CpG, IL-2, and IL-21 to induce antibody-secreting cells (ASC). (ANCA)-IgG was measured in these supernatants by ELISA.

Results: By employing EuroFlow-based HSFC, we detected circulating CD19⁺ B-cells at all timepoints after RTX treatment, in contrast to conventional low-sensitive flow cytometry. Pre-germinal center (Pre-GC) B-cells, memory B-cells and CD20⁺CD138^- plasmablasts (PBs) were rapidly and strongly reduced, while CD20^-CD138^- PrePC and CD20^-CD138⁺ mature (m)PCs were reduced slower and remained detectable. Both memory B-cells and CD20^- PCs remained detectable after RTX. Serum ANCA-IgG decreased significantly upon RTX. Changes in ANCA levels strongly correlated with changes in naive, switched CD27⁺ and CD27^- (double-negative) memory B-cells, but not with plasma cells. Lastly, we demonstrated in vitro ANCA production by AAV PBMCs, 24 and 48 weeks after RTX treatment reflecting MRA in the memory compartment of AAV patients.

Conclusion: We demonstrated that RTX induced strong reductions in circulating B-cells, but never resulted in complete B-cell depletion. Despite strongly reduced B-cell numbers after RTX, ANCA-specific memory B-cells were still detectable in AAV patients. Thus, MRA is identifiable in AAV and can provide a potential novel approach in personalizing RTX treatment in AAV patients.

Keywords: ANCA antibodies; ANCA-associated vasculitis; B-cells; glomerulonephritis; highly sensitive flow cytometry; immunomonitoring; minimal residual autoimmunity; rituximab.

Figure 1

Comparing the detection of circulating CD19⁺ B-cells with LSFC versus HSFC (A). Absolute numbers of circulating CD19⁺ B-cells were analyzed with HSFC (x-axis) and correlated to LSFC (y-axis) with, respectively, a detection limit of 1*10³ B-cells/L and 1*10⁶ B-cells/L. Each dot represents one timepoint for one patient before and 2, 4, 12, 24, or 48 weeks after RTX treatment (B). Each line represents the absolute CD19⁺ B-cell count for an individual patient during RTX treatment measured with LSFC (C). Each line represents the absolute CD19⁺ B-cell count for an individual patient during RTX treatment measured with HSFC. In (B, C), the dotted line indicated the detection limit for LSFC (10⁶ cells/L). Gray area indicates 1–20 analyzed events. Red dotted line indicated the detection limit for HSFC (10³ cells/L).

Figure 2

Residual memory and plasma cells after RTX. Absolute counts of (A) total CD19⁺ B-cells (B), Pre-GC B-cells (C), memory B-cells, and (D) plasma cells are shown for each individual patient that received remission-induction therapy with RTX (n = 6). The median is indicated by the thick black line. Red dashed line indicates the mean detection limit for HSFC. Gray area indicates 1–20 analyzed events (E). Median ± IQR fold change as compared to baseline for each B-cell subset per timepoint is shown (F). Mean distribution of B-cell subsets per timepoint during RTX treatment is shown.

Figure 3

Phenotyping residual memory and plasma cells after RTX (A). Median absolute counts of unswitched memory (CD27⁺IgM⁺ and/or IgD⁺), switched memory (CD27⁺IgG⁺ or IgA⁺) and double negative (DN, CD27⁻, IgG⁺, or IgA⁺) B-cells are shown for AAV patients (n = 6) after remission-induction treatment with RTX (B). Median ± IQR fold change from baseline for different memory B-cell subsets are shown for each timepoint after RTX (C). Median absolute counts of PBs (CD20⁺CD138⁻, red), PrePCs (CD20⁻CD138⁻, purple) and mPCs (CD20⁻CD138⁺, blue) are shown for AAV patients (n = 6) after RTX as remission-induction therapy (D). Median [IQR] fold change from baseline for the subsets in A are shown for each timepoint after RTX. Red dashed line indicated the detection limit for HSFC. Gray area indicates 1–20 analyzed events.

Figure 4

Changes of circulating immunoglobulins were not reflected by plasma cell kinetics after RTX (A). Median percentage change in total immunoglobulin -M (red line), G (blue line) and -A (orange line) levels after RTX as remission-induction treatment over time (n = 6) (B). Heatmap of spearman correlations of the change in IgG, IgM, and IgA with the change in different B-cell subsets as compared to baseline for all timepoints (n = 6). Gradients indicates Spearman’s R. *p < 0.05, **p < 0.01. ***p < 0.001.

Figure 5

Changes of circulating ANCAs associated with changes of naive and memory B-cells but not with plasma cells after RTX (A). Individual percentage change as compared to baseline of anti-PR3 and anti-MPO IgG serum levels after remission-induction with RTX treatment (n = 5) (B). Individual percentage change as compared to baseline of anti-PR3 and anti-MPO IgG serum levels is shown for each patient that received RTX as maintenance treatment (n = 8) (C). Heatmap of the Spearman correlations of changes in anti-MPO IgG (n = 6) and anti-PR3 IgG (n = 8) with changes in B-cell subsets as compared to baseline for all timepoints (n = 14). Gradients indicates Spearman’s R. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6

Minimal residual autoimmunity after RTX: presence of ANCA-specific memory B-cells. 1*10⁶ PBMCs/well from healthy controls (HCs) and AAV patients before, 24 and 48 weeks after RTX treatment, were stimulated for 10 days with CpG ODN class B, IL-2, and IL-21 to induce antibody-secreting cells (ASCs) in a 48-well plate (A). Representative bivariate dot plots of ASCs at day 0 and day 7 of PBMC cultures demonstrated the induction of CD27⁺⁺CD38⁺⁺ ASCs 7 days after polyclonal stimulation of PBMCs from a HC and an AAV patient (MPO-ANCA) (B). Absolute counts of total CD19⁺ B-cells were shown for each individual at baseline of the cultures (day 0) (C). Absolute counts of induced ASCs per well were shown for each individual after 7 days of culturing (D). Total IgG production was measured in the supernatants of each well after 10 days of culturing. Here the median of 5 wells is shown per individual (E). Total ANCA-IgG production was measured in the supernatants of each well after 10 days of culturing. Anti-PR3 IgG and anti-MPO IgG are, respectively, shown on the left and right y-axis. Each dot represents the median of 5 wells for each individual sample. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.