CD20 as a gatekeeper of the resting state of human B cells

Abstract

CD20 is a B cell-specific membrane protein and represents an attractive target for therapeutic antibodies. Despite widespread usage of anti-CD20 antibodies for B cell depletion therapies, the biological function of their target remains unclear. Here, we demonstrate that CD20 controls the nanoscale organization of receptors on the surface of resting B lymphocytes. CRISPR/Cas9-mediated ablation of CD20 in resting B cells resulted in relocalization and interaction of the IgM-class B cell antigen receptor with the coreceptor CD19. This receptor rearrangement led to a transient activation of B cells, accompanied by the internalization of many B cell surface marker proteins. Reexpression of CD20 restored the expression of the B cell surface proteins and the resting state of Ramos B cells. Similarly, treatment of Ramos or naive human B cells with the anti-CD20 antibody rituximab induced nanoscale receptor rearrangements and transient B cell activation in vitro and in vivo. A departure from the resting B cell state followed by the loss of B cell identity of CD20-deficient Ramos B cells was accompanied by a PAX5 to BLIMP-1 transcriptional switch, metabolic reprogramming toward oxidative phosphorylation, and a shift toward plasma cell development. Thus, anti-CD20 engagement or the loss of CD20 disrupts membrane organization, profoundly altering the fate of human B cells.

Keywords: B lymphocyte; CD20; plasma cell; therapeutic antibody.

Fig. 1.

CD20 expression and localization on WT Ramos B cells. (A) Flow cytometry analysis showing the expression of IgM-BCR, IgD-BCR, CD19, and CD20 on Ramos WT cells compared to unstained control (CTRL); n = 18. (B) Fab-PLA study of the proximity of CD20 to either the IgD-BCR or to the IgM-BCR on resting (R), or 5-min pervanadate-activated (A) WT Ramos B cells. Representative microscope images (Left) of PLA signals are shown in red and nuclei in blue. (Scale bar, 10 µM.) Scatter dot plot represents the mean (red bar) and SD of PLA signals (Signal Counts); n = 3. (C) Schematic drawing of the proposed localization of coreceptors and surface marker within IgM and IgD-class protein islands on the surface of human resting B cells.

Fig. 2.

Loss of CD20 results in altered expression of B cell surface markers and transient activation. (A) Timeline of developmental states in CD20KO Ramos cell line generation. MS1A4 gene targeting with CRISPR/Cas9 at day 0, developmental stage at days 3 to 6 shown as CD20KO-new (KO-N, orange), at days 10 to 17 shown as CD20KO-intermediates (KO-I, light blue), and after 1 mo shown as CD20KO-late (KO-L, dark blue). (B) Flow cytometry analysis showing expression of surface molecules of KO-N, KO-I, and KO-L Ramos cells over time compared to WT and unstained control (gray); n = 18. (C) Representative Western blot analysis of KO-L Ramos B cell lysates (Right) compared to WT (Left). Lysates were taken 20 d after MS4A1 gene targeting; n = 14 (D) Fab-PLA analysis of IgM-BCR proximity to CD19 in resting WT Ramos B cells compared to the unstimulated KO-N cells 3 d after MS4A1 gene targeting. (Scale bar: Left, 10 µm.) PLA microscope images were quantified as scatter dot plot with mean and SD (Right); n = 3. (E) Intracellular flow cytometry analysis of phosphorylated Akt (pAkt-Ser473) or phosphorylated Syk (pSyk-Tyr525,526) of KO-N cells compared to WT. Gating used to analyze high IgM-BCR expressing KO-N cell population for pSyk or pAkt levels, respectively. (F) Flow cytometry analysis showing the over-time expression of the B cell-specific surface activation markers CD86, and CD69 of KO-N compared to WT.

Fig. 3.

Gain-of-function studies with a conditional CD20KO (cKO). (A) Schematic of CRISPR-mediated insertion of an aptamer-controlled exon (c-exon) in between exon 3 and exon 4 of the human MS4A1 gene generating a CD20 conditional KO (cKO-top). Tetracyclin (Tet) induces c-exon skipping and restores the ORF of the MS4A1 gene (cKO-Tet; Bottom). (B) Flow cytometry analysis of the expression of the surface markers CD20, CD19, IgM, CD22, CD81, and CD40 on cKO (blue) or cKO-Tet Ramos cells (red). The untreated (WT) and Tet-treated (Ramos-Tet) Ramos cells are shown as control (gray). The cKO-Tet cells are derived from 30 d posttransfection cKO Ramos B cells treated for 12 h with 6 µM Tet to restore CD20 expression, n = 4.

Fig. 4.

In vitro and in vivo treatment with rituximab (RTX) leads to CD20KO phenotype. (A) Flow cytometry analysis after treatment with RTX for 3 d showing the loss of CD19, IgM-BCR, and CD22 on Ramos-RTX compared to CD20KO-L, untreated Ramos-WT, and unstained control (Ctrl); n = 3 (B) Flow cytometry analysis of negatively selected naive B cells from peripheral blood of a healthy donor were treated with RTX for 60 min (HD-RTX) and compared to untreated (HD), to plasma cells (PCs) of the same donor, or left untreated and unstained (Ctrl); n = 3. (C) Example of B cells taken from EDTA whole-blood samples of RA patient undergoing RTX treatment after 0, 15, 30, 60, and 120 min. RTX (1 mg/mL) flow rate, 50 mL/h. Flow cytometry staining of CD19⁺/CD27⁻/IgM⁺ selected B cells shows internalization of IgM and CD19; n = 3.

Fig. 5.

Increased PC differentiation of CD20KO Ramos B cells and primary naive B cells and of CLL B cell relapse after RTX therapy. (A) Flow cytometry analysis showing size (FSC-A) and expression of PC markers TACI, CD138, and AID on KO-N, KO-I, and KO-L Ramos B cells compared to WT and unstained control (Ctrl); n = 6. (B) Expression of surface IgA-BCR (Left) and ELISA of IgA secretion (Right) of KO-L compared to Ramos WT, PBS, and RPMI control. (C) Summarized intracellular flow cytometry analysis of six independently generated CD20KO Ramos B cell lines showing the fold change (FC) of Pax5 and Blimp1 expression of KO-L Ramos B cell lines compared to WT. (D) Representative examples of Western blot analysis for B cell differentiation markers of KO-L Ramos B cells compared to WT. Lysates were taken 20 d after induction of CD20KO; n = 3. (E) Flow cytometry analysis 24 h after CD20CRISPR/Cas9 CD20KO of isolated naive HD B cells. CD20^low B cells were gated and shown in blue, compared to transfection control (empty plasmid) in black; n = 3. (F) PLASMA UP GENES Heatmap, expression of PC differentiation up-regulated genes. The color code indicates the row-wise scaled intensity across the samples. Genes are ranked according to their log2 fold FC in WT (Left) vs. KO-L (Right); n ≥ 3. (G) Example of enrichment plot (curve) of PC differentiation up-regulated genes (ticks) in one patient. (H) Enrichment barcode illustrating the distribution of PC differentiation up-regulated genes (colored segments) in every individual patient. From GSE37168. Treated patients were ordered from Left to Right based on their enrichment score, from high to low. Significant enrichment scores are depicted by an asterisk (*).

Fig. 6.

Metabolic switch of CD20KO Ramos B cells. (A) Oxygen consumption rate (OCR) of KO-L compared to WT Ramos cells. Metabolite flux analysis was performed in triplicate and is displayed as mean ± SD. One, out of three experiments, is shown. Used inhibitors: A = oligomycin; B = FCCP; C = rotenone plus antimycin. (B) Mitochondrial mass staining with Mito Tracker Red-CMXRos in KO-L compared to WT Ramos cells and normalized to volume of WT Ramos cells; n = 4. (C) NAD⁺ levels of KO-L compared to WT Ramos cells determined with untargeted metabolomic profiling; n = 3. (D) Mitochondrial superoxide levels stained with MitoSox Red. KO-L cells, WT Ramos cells, and unstained control (Ctrl) are shown; n = 4. (E) Survival rates (propidium iodide [PI]-negative cells) of KO-L and WT Ramos cells after buthionine sulfoximine (BSO) titration. The experiment was performed three times and mean MFI values were normalized to day 0. (F) Extracellular acidification rate (ECAR) of KO-L and WT Ramos cells as a measure of glycolysis is shown. The measurement was performed in technical triplicates and is displayed as mean ± SD. One, out of three independent experiments, is shown. D = glucose, A = oligomycin, and E = 2DG. (G) Representative Western blot analysis of hexokinase II (HK II) protein levels of KO-L and WT Ramos cells. Lysates were taken 20 d after induction of CD20 KO, and data are representative for at least three independently generated CD20 KO Ramos B cell lines. (H) Levels of the glycolytic intermediates glucose 6-phosphate (G6P), fructose 6-phosphate (F1,6BP), and pyruvate as determined by untargeted metabolomic profiling are shown; n = 3. (I) Cell death as determined by PI staining after treatment with BPTES; n = 3. (J) Proliferation index determined with MFI of CellTrace staining per day of KO-L cells cultivated in galactose compared to WT Ramos cells; n = 3. (K) Levels of UDP-galactose, galactose-1-phosphate (G 1-P), and galactonate as determined by untargeted metabolomic; n = 4.