Patient backgrounds
The mean age of all patients was 70.6 ± 8.3 years, and there were roughly the same number of men and women (male : female = 54:46). All patients had new-onset disease. Microscopic polyangiitis was diagnosed in 34 patients (63.0%) and granulomatosis with polyangiitis in 20 patients (37.0%). For remission induction therapy, the patients were divided into those receiving high-dose GC therapy combined with RTX (RTX group) and those receiving GC therapy combined with intravenous CY pulse (IV-CY group). The baseline patient characteristics in these groups are compared in Table 1. In both groups, the common organ dysfunctions were respiratory dysfunction and nephropathy. BVAS was 17.4 ± 6.5 in the RTX group and 15.1 ±4.6 in the IV-CY group, showing that disease activity was high in the majority of the patients. Between the RTX and IV-CY groups, no significant differences were observed in age, sex, disease duration, vasculitis type, BVAS, organ dysfunction, or GC dose. Concomitant AZA treatment tended to be used more frequently in the IV-CY group than in the RTX group
No difference in the rate of remission in BVAS between groups
First, the therapeutic effects of RTX and conventional therapy on AAV were assessed. At 6 months after treatment initiation, BVAS was significantly improved in both the RTX and IV-CY groups (mean change: -14.2 ± 6.3, p < 0.01 in the RTX group; -15.2 ± 5.1, p < 0.01 in the IV-CY group) (Figure 1A). The rate of remission in BVAS at 6 months after treatment initiation tended to be higher in the RTX group than in the IV-CY group, but this difference was not statistically significant [RTX group = 61.8%, IV-CY group: 40.0%, p = 0.16] (Figure 1B). In addition, no significant differences in the incidences of adverse events, severe infection, leukopenia, or thrombocytopenia or in mortality were observed between the two groups (Table 2). The most common adverse event was infection in both groups, and most deaths were caused by severe infection.
Increased proportion of IgD–CD27– double-negative memory B cells in patients with AAV
Phenotypes of peripheral T and B cells before treatment initiation were analyzed by 8-color flow cytometry and compared between AAV patients and 15 HCs matched for age and sex (Figure 2, Supplementary Table S2). Among cluster of differentiation (CD)-4 T cells, the proportion of naïve CD4 T cells was significantly higher in AAV patients than in HCs, whereas the proportion of central memory CD4 T cells was significantly lower. As for CD8 T cells, no differences in phenotype were observed between AAV patients and HCs. By contrast, in B cells, the proportion of immunoglobulin (Ig)-M unswitched memory B cells was significantly lower in AAV patients than in HCs (HC = 19.0 ± 6.9, AAV = 12.1 ± 6.7, p < 0.01), whereas the proportion of IgD-CD27- double-negative B cells was significantly higher (HC = 5.4±2.7, AAV = 9.8±7.9, p = 0.04]. The absolute number of IgM+ unswitched memory B cells was significantly smaller in patients with AAV than in healthy controls (HC:AAV = 16.8 ± 7.5/μl:7.6 ± 7.9/μl, p < 0.001). On the other hand, the absolute number of double-negative B cells did not differ significantly between healthy controls and patients with AAV (HC:AAV = 4.9 ± 3.0/μl:6.0 ± 5.8/μl, p = 0.48). These data are shown in Supplementary Figure 2. There was a positive correlation between the proportion of class-switched memory B cells and the proportion of IgD-CD27- B cells in patients with AAV (n = 54), as shown in Supplementary Figure 3 (r = 0.50, p < 0.001, Pearson product-moment correlation coefficient).
Next, the association between peripheral CD4+ T and B cell phenotypes and clinical signs before treatment initiation was analyzed (Figure 3, Supplementary Table S3). Proportions of peripheral T and B cell phenotypes did not correlate with BVAS at baseline. However, among peripheral B cells, the proportion of naive B cells positively correlated with the rate of improvement in BVAS at 6 months after treatment initiation (r = 0.35, p < 0.01), whereas the proportion of class-switched memory B cells negatively correlated with this rate (r = -0.28, p = 0.04). There was no correlation between the proportion of plasmablasts and the rate of BVAS improvement (Supplementary Figure 5A).
Association of excessive B cell differentiation with treatment resistance
Next, among the AAV patients, we compared the response to treatment between patients with and without excessive B cell differentiation. Patients with excessive B cell differentiation were defined as those in whom the proportion of class-switched memory B cells or IgD-CD- B cells among all B cells (CD19+CD20+) was >2 SDs higher than the mean in the HCs (class-switched memory B cells/B cells > 23% or IgD-CD27- B cells/B cells > 11.3%) (Figure 4A). Based on this definition, 24 of the 54 AAV patients (44.4%) showed excessive B cell differentiation (Figure 4B).
Between the patients with and without excessive B cell differentiation, no notable differences were observed concerning patient characteristics or T cell phenotypes before administration of remission induction therapy (Supplementary Tables S4 and S5). However, after 6 months therapy, BVAS was significantly higher in patients with excessive B cell differentiation than in those without (p < 0.01, Figure 4C), and the rate of remission in BVAS was significantly lower in those with excessive B cell differentiation at 6 months [presence of excessive B cell differentiation = 37.5%, absence of excessive B cell differentiation = 66.7%, p < 0.01] (Figures 4D). On the other hand, the rate of BVAS remission at 6 months was compared between a group of patients with AAV (n = 39) with a proportion of plasmablasts ≥ HC mean + 2SD (plasmablasts/CD19+ cells ≥ 0.6%) and another group (n = 15) with the same parameter < HC mean + 2SD. This comparison revealed no significant inter-group difference (48.7%:66.7%, p = 0.36) (Supplementary Figure 5B).
MPA is more frequently therapy-resistant than is GPA, and it is possible that in the present study, the higher percentage of MPA cases in the CY-treated group (although this difference was not statistically significantly) affected the results. However, among patients with MPA free of excessive B cell differentiation, the rate of BVAS remission did not differ between the CY-treated group (75%, n = 8) and the RTX-treated group (55%, n = 11) (p = 0.63). On the other hand, among patients with MPA showing excessive B cell differentiation, the rate of BVAS remission was significantly lower in the CY-treated group (0%, n = 7) than in the RTX-treated group (75%, n = 8) (p < 0.01). Few patients with GPA were studied, and there was no significant difference between the rates of BVAS remission among those patients in the two treatment groups. These results, shown in Supplementary Figure 4 (A,B), suggest that excessive B cell differentiation is probably involved more closely in the resistance to treatment than the type of vasculitis. These results reveal that patients with excessive B cell differentiation are resistant to treatment.
Favorable effectiveness of RTX in improving BVAS, reducing GC use, and increasing survival among patients with circulating B cell abnormalities
Lastly, the two treatment groups were further divided according to the presence or absence of excessive B cell differentiation, and responses to treatment were compared among the four resulting treatment groups. No significant differences were observed in disease duration, age, type of vasculitis, BVAS, or GC dose (Table 3). After 6 months of remission induction therapy, patients with excessive B cell differentiation in the IV-CY group had significantly higher BVAS than patients in the other three groups (Figures 5A and Figure 5C). Meanwhile, patients with excessive B cell differentiation in the RTX group had significantly lower BVAS at 6 months after treatment initiation compared with those with excessive B cell differentiation in the IV-CY group. BVAS in patients with excessive B cell differentiation in the RTX group were not significantly different from those in patients without excessive B cell differentiation, in either the RTX or IV-CY groups (Figures 5A and 5C).
Furthermore, patients with excessive B cell differentiation who achieved BVAS remission were compared. Although all patients receiving RTX achieved BVAS remission, none of those receiving IV-CY did. There were significantly more patients responding to treatment in the RTX group (RTX group = 56.3%, IV-CY group = 0.0%, p < 0.01) (Figure 5B).
Next, in patients with excessive B cell differentiation, rates of GC dose reduction were compared. The RTX group showed higher rates of dose reduction at 3 months [RTX group = 67.6±13.1%, IV-CY group = 47.2±21.6%, p < 0.01] and 6 months [RTX group = 84.3±7.0%, IV-CY group = 69.1±21.9%, p = 0.03] of remission induction therapy than the IV-CY group (Supplementary Figure 6).
Regarding the incidence rates of adverse events in patients with excessive B cell differentiation, no significant difference was observed between the RTX and IV-CY groups. However, the incidence of severe infection was lower in the RTX group (RTX group = 4/16 patients, conventional therapy group = 6/8 patients, p = 0.03, Fisher's exact test). Moreover, the 6-month survival rate after administration of high-dose GC therapy was significantly higher in the RTX group than in the IV-CY group (RTX group = 93.8%, IV-CY group = 62.5%, p = 0.0486) (Figure 5D).