Abnormalities of bone marrow B cells and plasma cells in primary immune thrombocytopenia

Abstract

Primary immune thrombocytopenia (ITP) is an autoantibody-mediated hemorrhagic disorder in which B cells play an essential role. Previous studies have focused on peripheral blood (PB), but B cells in bone marrow (BM) have not been well characterized. We aimed to explore the profile of B-cell subsets and their cytokine environments in the BM of patients with ITP to further clarify the pathogenesis of the disease. B-cell subpopulations and their cytokine/chemokine receptors were detected by using flow cytometry. Plasma concentrations of cytokines/chemokines were measured by using enzyme-linked immunosorbent assay. Messenger RNA levels of B cell-related transcription factors were determined by using quantitative polymerase chain reaction. Regulatory B cell (Breg) function was assessed by quantifying their inhibitory effects on monocytes and T cells in vitro. Decreased proportions of total B cells, naive B cells, and defective Bregs were observed in patients with ITP compared with healthy controls (HCs), whereas an elevated frequency of long-lived plasma cells was found in BM of autoantibody-positive patients. No statistical difference was observed in plasmablasts or in short-lived plasma cells between patients with ITP and HCs. The immunosuppressive capacity of BM Bregs from patients with ITP was considerably weaker than HCs. An in vivo study using an active ITP murine model revealed that Breg transfusion could significantly alleviate thrombocytopenia. Moreover, overactivation of CXCL13-CXCR5 and BAFF/APRIL systems were found in ITP patient BM. Taken together, B-cell subsets in BM were skewed toward a proinflammatory profile in patients with ITP, suggesting the involvement of dysregulated BM B cells in the development of the disease.

Graphical abstract

Figure 1.

Total B cells, immature B cells, naive B cells, and memory B cells of patients with ITP and HCs. (A) Elimination of adherent cells and fragments, and gate settings for lymphocytes from BM and PB. (B, C) patients with ITP had fewer CD19⁺ B cells in BM (7.81 ± 0.84% vs 15.26 ± 1.84%; P < .001) but more in PB (6.94 ± 0.63% vs 4.83 ± 0.41%; P = .013) compared with HCs. (D, E) No statistical difference was found in the proportion of CD19⁺ B cells between BM and PB in patients with ITP (P = .893), whereas the proportion of BM CD19⁺ B cells was significantly higher than its PB counterparts in HCs (15.26 ± 1.84% vs 4.83 ± 0.41%; P = .002). (F, G) No difference was found in CD19⁺CD127⁺ immature B-cell compartments in BM CD19⁺ cells between patients with ITP and HCs (4.91 ± 0.94% vs 4.01 ± 0.58%; P = .504), whereas a lower frequency of CD19⁺CD127⁺ B cells in lymphocytes was observed in patients with ITP (0.22 ± 0.07% vs 0.53 ± 0.12%; P = .029). (H) Density plots of CD27 and CD38 double staining from CD19⁺ cells in flow cytometry. CD19⁺CD27^– cells (Q1 + Q3) were naive B cells, and CD19⁺CD27⁺CD38^–/low cells (Q4) were memory B cells. (I, J) The percentage of naive B cells was significantly lower in BM B cells from patients with ITP compared with HCs (64.04 ± 3.04% vs 77.33 ± 2.33%; P = .011), whereas no statistical difference was found in PB naive B cells between patients with ITP and HCs (median [range], 69.36% [47.66%-93.53%] vs 67.69% [49.45%-85.50%]; P = .817). (K, L) No statistical difference was observed in naive B cells between BM and PB from patients with ITP (median [range], 69.82% [12.78%-89.30%] vs 69.36% [47.66%-93.53%]; P = .731), whereas more naive B cells existed in BM B cells than in PB from HCs (78.10 ± 2.38% vs 64.53 ± 3.24%; P < .001). (M, N) The frequency of memory B cells was higher in BM B cells of patients with ITP than that of HCs (26.23 ± 3.20% vs 12.61 ± 2.57%; P = .002), but no difference was observed in PB memory B cells between patients with ITP and HCs (25.17 ± 2.98% vs 22.84 ± 4.42%; P = .654). (O, P) No statistical difference was found in percentage of memory B cells in CD19⁺ cells between BM and PB in patients with ITP (26.23 ± 3.20% vs 25.17 ± 2.98%; P = .813). HCs had remarkably decreased frequency of memory B cells in BM compared with PB (12.61 ± 2.57% vs 22.84 ± 4.42%; P = .021). *P < .05; **P < .01; ***P < .001. APC, allophycocyanin; FITC, fluorescein isothiocyanate; FSC-A, forward scatter pulse area; SSC-A, side scatter pulse area.

Figure 2.

Plasmablasts, SLPCs, and LLPCs in patients with ITP and HCs. (A) CD19⁺CD27⁺CD38^hi cells were gated for analysis of CD19⁺CD27⁺CD38^hiKi67⁺CD138⁺ plasmablasts (Q2) and CD19⁺CD27⁺CD38^hiKi67^–CD138⁺ SLPCs (Q1). (B-E) No statistical difference in frequency of BM or PB plasmablasts in CD19⁺ cells was found between patients with ITP and HCs. Moreover, no significant difference in percentage of plasmablasts in CD19⁺ cells was observed between BM and PB from patients with ITP or HCs. (F, G) There was no remarkable difference in percentage of SLPCs in CD19⁺ cells between patients with ITP and HCs. (H, I) patients with ITP (median [range], 4.32% [1.32%-34.93%] vs 0.34% [0.06%-1.07%]; P < .001) and HCs (median [range], 10.70% [0.72%-20.39%] vs 0.10% [0.01%-0.93%]; P = .016) have more SLPCs in BM B cells than in PB B cells. (J) CD19^–CD38^hiCD138⁺ LLPCs were analyzed from autoantibody-negative or antibody-positive patients and HCs, respectively. (K) No statistical difference was found between all patients with ITP and HCs (median [range], 0.16% [0.01%-1.38%] vs 0.10% [0.02%-0.42%]; P = .273). The proportion of BM LLPCs from autoantibody-positive patients was significantly higher than that from HCs (median [range], 0.27% [0.13%-1.38%] vs 0.10% [0.02%-0.42%]; P = .021) and autoantibody-negative patients (median [range], 0.27% [0.13%-1.38%] vs 0.08% [0.01%-0.15%]; P < .001), whereas no significant difference was found between autoantibody-negative patients and HCs (P = .485). *P < .05; **P < .01; ***P < .001. APC, allophycocyanin; PE, phycoerythrin; SSC, side scatter.

Figure 3.

Decreased number and impaired ability of BM Bregs to secrete inhibitory cytokines in patients with ITP. (A) Representative density plots of BM and PB CD19⁺CD24^hiCD38^hi Bregs in patients with ITP and HCs. (B, C) The percentages of BM (median [range], 5.50% [0.03%-16.77%] vs 18.67% [4.58%-36.72%]; P < .001) and PB (median [range], 0.31% [0.02%-2.27%] vs 3.05% [0.77%-15.19%]; P < .001) Bregs in CD19⁺ cells were significantly lower in patients with ITP compared with their respective counterparts in HCs. (D, E) Both patients with ITP and HCs had more Bregs in BM B cells than in PB B cells (patients with ITP: median [range], 5.50% [0.03%-16.77%] vs 0.31% [0.02%-2.27%], P < .001; HCs: median [range], 18.67% [4.58%-36.72%] vs 3.05% [0.77%-15.19%], P = .001). (F, G) Quantification of IL-10⁺ or TGF-β⁺ B cells in BM CD19⁺ cells from patients with ITP and HCs, respectively. (H, I) BM B cells from patients with ITP produced less IL-10 (2.35 ± 0.51% vs 7.15 ± 0.86%; P < .001) and TGF-β (median [range], 1.32% [0.39%-4.11%] vs 4.10% [1.34%-5.91%]; P = .028) compared with HCs after stimulation with CPG-ODN. *P < .05 , **P < .01; ***P < .001. APC, allophycocyanin; CPG-ODN, CPG oligonucleotide; FITC, fluorescein isothiocyanate; PE, phycoerythrin; SSC, side scatter.

Figure 4.

Compromised immunoregulatory capacity of BM Bregs from patients with ITP. (A) Quantification of CD14⁺TNF-α⁺ monocytes cultured alone or with activated BM B cells that contained Bregs from patients with ITP and HCs, respectively. (B, C) Percentage of TNF-α⁺ monocytes cocultured with BM B cells was significantly lower than that cultured alone, both in patients with ITP (21.24 ± 4.90% vs 36.61 ± 7.82%; P = .001) and in HCs (15.70 ± 5.44% vs 41.52 ± 11.90%; P = .014). (D) BM B cells from patients with ITP exhibited decreased capacity to suppress monocyte TNF-α production compared with HCs (39.95 ± 5.45% vs 61.55 ± 6.89%; P = .029). The percentage of inhibition was calculated as [1 – (percentage of TNF-α⁺ monocytes cocultured with B cells/percentage of TNF-α⁺ monocytes cultured alone)] × 100%. (E) IFN-γ secreted by CD4⁺ T cells cultured alone or with BM B cells from patients with ITP and HCs, respectively. (F, G) BM B cells from the ITP (16.31 ± 4.89% vs 14.46 ± 4.09%; P = .283) and HC (26.22 ± 3.47% vs 15.85 ± 2.31%; P < .001) groups both inhibited IFN-γ secretion by CD4⁺ T cells, but statistical significance was not reached in the ITP group. (H) Inhibition of CD4⁺IFN-γ⁺ T cells by BM B cells from patients with ITP was weaker than that from HCs (median [range], 22.78% [6.84%-35.80%] vs 42.10% [30.16%-48.72%]; P = .009). Percentage of inhibition was calculated as [1 – (percentage of CD3⁺IFN-γ⁺ T cells cocultured with B cells/percentage of CD3⁺IFN-γ⁺ T cells cultured alone)] × 100%. (I) CD4⁺ T cells cultured alone or with BM B cells for 72 hours, and frequencies of CD4⁺CD25⁺Foxp3⁺ Tregs were determined by using flow cytometry. (J, K) BM B cells from patients with ITP and HCs both enhanced the percentage of Tregs in CD4⁺ T cells (patients with ITP, 4.04 ± 1.31% vs 15.09 ± 3.40% [P = .007]; HCs, 1.32 ± 0.38% vs 21.93 ± 1.89% [P < .001]). (L) BM B cells from patients with ITP had weaker ability to promote Treg differentiation compared with HCs (median [range], 3.74 [1.58-14.55] vs 20.70 [8.82-25.82]; P = .009). Promotion of Tregs was calculated as Tregs % cocultured with B cells/Tregs % cultured alone. (M) Platelet counts of active ITP mice transfused with BM Bregs from immunized CD61-KO mice were higher than that of control group mice on days 14, 21, and 28 after splenocyte transfusion. *P < .05; **P < .01; ***P < .001. APC, allophycocyanin; PE, phycoerythrin; SSC, side scatter.

Figure 5.

Aberrant expression of B-lineage transcription factors and abnormal B cell–related cytokines/chemokines in patients with ITP. (A) Patients with ITP had decreased mRNA levels of Pax5 in BMMNCs compared with HCs (0.0049 ± 0.0012 vs 0.0100 ± 0.0017; P = .022). No statistical difference was found in mRNA expression of XBP1 or IRF4 in BMMNCs between patients with ITP and HCs (all P > .05). (B) mRNA levels of Pax5 (0.0017 ± 0.0005 vs 0.0108 ± 0.0016; P < .001), XBP1 (0.0201 ± 0.0033 vs 0.0631 ± 0.0093; P < .001), and IRF4 (0.0037 ± 0.0007 vs 0.0182 ± 0.0037; P < .001) in PBMCs from patients with ITP were much lower compared with HCs. (C, D) Plasma concentrations of BM and PB CXCL13 in patients with ITP were much higher than in HCs (BM: median [range], 94.08 pg/mL [34.46-423.50 pg/mL] vs 55.33 pg/mL [13.31-97.54 pg/mL], P = .003; PB: median [range], 83.21 pg/mL [53.12-291.96 pg/mL] vs 59.27 pg/mL [27.73-116.58 pg/mL], P = .029). (E, F) No difference was found in plasma CXCL13 between BM and PB from the ITP or HC group (all P > .05). (G, H) Plasma levels of BAFF in BM and PB of patients with ITP were higher compared with that in HCs (BM: median [range], 1338.0 pg/mL [989.6-3419.6 pg/mL] vs 1123.0 pg/mL [459.6-1499.6 pg/mL], P = .011; PB: median [range], 1388.0 pg/mL [993.6-3451.6 pg/mL] vs 1022.0 pg/mL [559.6-1255.6 pg/mL], P < .001). (I, J) No divergence was found in BAFF between BM and PB from patients with ITP or HCs (all P > .05). (K, L) Higher plasma concentration of APRIL was observed in BM from patients with ITP than in HCs (median [range], 3.79 ng/mL [1.12-9.34 ng/mL] vs 2.34 ng/mL [1.56-3.79 ng/mL]; P = .005), whereas no difference was found in PB between patients with ITP and HCs (median [range], 1.49 ng/mL [0.47-4.53 ng/mL] vs 1.96 ng/mL [1.42-4.36 ng/mL]; P = .230). (M, N) Plasma APRIL was lower in PB than in BM in patients with ITP (median [range], 3.79 ng/mL [1.12-9.34 ng/mL] vs 1.49 ng/mL [0.47-4.53 ng/mL]; P < .001); no such tendency was found in HCs (median [range], 2.52 ng/mL [1.35-3.18 ng/mL] vs 1.96 ng/mL [1.42-4.53 ng/mL]; P = .770). *P < .05; **P < .01; ***P < .001.

Figure 6.

CXCR5, BAFF-R, BCMA, and TACI expression in BM and PB B cells of patients with ITP and HCs. (A) Representative density plots of CXCR5 on CD19⁺ cells in BM and PB of patients with ITP and HCs. (B, C) BM (69.25 ± 4.27% vs 26.28 ± 2.64%; P < .001) and PB (median [range], 90.78% [80.86%-96.06%] vs 85.13% [56.76%-91.61%]; P = .046) B cells from patients with ITP had elevated levels of CXCR5 compared with those from HCs. (D, E) Surface CXCR5 level on B cells in PB was significantly higher than in BM, both from patients with ITP (median [range], 72.36% [57.54%-92.65%] vs 90.37% [59.93%-96.06%]; P = .006) and HCs (30.12 ± 3.82% vs 76.86 ± 4.86%; P = .003). (F) Representative density plots of BAFF-R on CD19⁺ cells in the BM and PB of patients with ITP and HCs. (G, H) BM and PB B cells from patients with ITP had elevated levels of BAFF-R compared with HCs (BM: median [range], 84.10% [60.30%-97.54%] vs 37.08% [16.56%-52.58%], P < .001; PB: median [range], 93.97% [87.20%-97.76%] vs 90.97% [76.97%-95.09%], P = .016). (I, J) BAFF-R expression on B cells was remarkably higher in PB than in BM, both from patients with ITP and from HCs (patients with ITP: median [range], 85.75% [60.30%-97.54%] vs 93.83% [87.20%-97.38%], P = .002; HCs: median [range], 37.76% [16.56%-58.27%] vs 89.00% [76.97%-95.09%], P = .031). (K) Representative density plots of BCMA on CD19⁺ cells in BM and PB of patients with ITP and HCs. (L, M) BM B cells from patients with ITP exhibited elevated BCMA levels compared with HCs (median [range], 6.04% [0.77%-22.76%] vs 3.77% [1.24%-6.55%]; P = .018), whereas there was no statistical difference in PB B-cell BCMA levels between patients with ITP and HCs. (N, O) No statistical difference was found in B-cell BCMA levels between BM and PB in patients with ITP (P = 0.454), whereas the level of B-cell BCMA in PB was considerably higher than in BM in HCs (2.91 ± 0.52% vs 4.47 ± 0.67%; P = .009). (P-S) No statistical difference was found in BM or PB B-cell TACI levels between patients with ITP and HCs. Furthermore, no statistical difference in B-cell TACI level between BM and PB was observed in patients with ITP or in HCs (all P > .05). *P < .05; **P < .01; ***P < .001. PE, phycoerythrin; SSC, side scatter.

Figure 7.

mRNA levels of CXCR5, BAFF-R, BCMA, and TACI of B cells. (A, B) The mRNA level of B-cell CXCR5 from patients with ITP was also higher than that from HCs, both in BM (0.070 ± 0.021 vs 0.016 ± 0.007; P = .034) and PB (median [range], 0.071 [0.015 to 0.338] vs 0.035 [0.017-0.089]; P = .017). (C, D) Increased mRNA levels of BAFF-R in BM (0.066 ± 0.020 vs 0.011 ± 0.003; P = .024) and PB (median [range], 0.090 [0.010-0.443] vs 0.030 [0.012-0.115]; P = .023) B cells from patients with ITP was found compared with HCs. (E, F) Elevated mRNA levels of BCMA were observed in BM B cells of patients with ITP compared with HCs (0.049 ± 0.011 vs 0.012 ± 0.002; P = .011), whereas no statistical difference was found in PB B cells between patients with ITP and HCs. (G, H) TACI mRNA levels of BM B cells in patients with ITP were also higher than those in HCs (0.064 ± 0.018 vs 0.016 ± 0.004; P = .034); no statistical difference was reached in PB B cells between patients with ITP and HCs. *P < .05; **P < .01; ***P < .001.