Introduction

Rituximab, an anti-CD20 antibody targeting B-lymphocytes (B-cells), is a promising agent for treating complicated steroid-dependent nephrotic syndrome (SDNS) in children [1,2,3,4]. Rituximab significantly reduces the number of relapses and can ameliorate critical adverse effects of corticosteroids such as growth retardation, obesity, glaucoma, hypertension, osteoporosis, and severe infection [5]. Rituximab was approved for child-onset refractory nephrotic syndrome by the Ministry of Health, Labour and Welfare in Japan in 2013 on the basis of the findings of a randomized clinical trial [6]. Generally, B-cell depletion is strongly correlated with persistent remission [7, 8], relapse rarely occurring during peripheral blood B-cell depletion. However, patients are likely to have relapses a few months after B-cell recovery. Therefore, maintenance therapy with immunosuppressive agents, including mycophenolate mofetil (MMF), is a good option for extending the relapse-free period after the recovery of B cells [3]. Surprisingly, we have encountered some patients whose disease has relapsed during B-cell depletion after receiving rituximab; thus, rituximab was relatively ineffective in these patients. To the best of our knowledge, the characteristics of such patients have not yet been well described. We have, therefore, investigated their characteristics and clinical course (relapsed group) and compared them with those of patients who did not have relapses during B-cell depletion (non-relapsed group).

Materials and methods

Study design and patient cohort

We retrospectively analyzed 82 patients with SDNS (58 male, 24 female, median age 12.3 years; range 3.1–21.8 years) treated with rituximab from January 2007 to December 2012 in the National Center for Child Health and Development, Tokyo, Japan. Rituximab was administered to patients with complicated SDNS who were unresponsive to conventional immunosuppressive agents including cyclosporine A (CsA; n = 80), cyclophosphamide (CPM; n = 39), mizoribine (MZR; n = 12), mycophenolate mofetil (MMF; n = 12), and tacrolimus (Tac; n = 5) (with some overlap).

Steroid dependency was defined as the occurrence of two consecutive relapses during the tapering-off period, or within 2 weeks of discontinuation of steroids. Steroid treatment was based on the International Study of Kidney Disease in Children (ISKDC) protocol [9]. At primary onset, patients were initially treated with daily prednisolone (PSL) 60 mg/m2/day for four weeks. Thereafter, PSL was reduced to 40 mg/m2 on alternate days for 4 weeks. Relapses were treated with PSL 60 mg/m2/day until remission (no or trace proteinuria for three consecutive days) had been achieved. The dosage of PSL was then reduced to the same dose given on alternate days for 2 weeks and subsequently further reduced to 60 mg/m2/2 days, followed by 40 and 20 mg/m2/2 days every 2 weeks.

Having received approval from the ethics committee of our institution for off-label use of rituximab and obtained parental consent, rituximab (375 mg/m2, maximum 500 mg) was administered to patients during their proteinuria-free periods. Seventy-nine patients received a single dose of rituximab and three received four doses of rituximab in 4 consecutive weeks. Median age at first rituximab infusion was 12.3 years (range 3.1–21.8 years). The median duration of follow-up after the first rituximab infusion was 25 months (range 1–71 months). Two weeks after rituximab, PSL was switched to alternate day treatment and its dosage gradually tapered by 5–10 mg every 2–4 weeks. Most patients successfully discontinued PSL; however, a few patients required maintenance with low-dose PSL. Eight patients discontinued immunosuppressive agents after rituximab, whereas the other 74 continued them, including CsA (n = 49), Tac (n = 5), CPM (n = 2), MZR (n = 20), and MMF (n = 66) (with some overlap) to prevent further relapse. CD19 depletion was assessed repeatedly by flow cytometry assay. B-cell depletion was defined as peripheral blood CD19 positive cells ≤1% of leukocytes, or 0–5 cells/mm3.

Statistical analysis

Data are expressed as medians and range. Statistical analysis was performed with JMP software version 9.0 (SAS Institute Japan, Tokyo, Japan). All p values are two sided and p < 0.05 was considered to denote statistical significance. Comparisons were performed using the Wilcoxon non-parametric test.

Results

Patients who had relapses during B-cell depletion

All patients achieved B-cell depletion. Glomerular filtration rate and that estimated from serum creatinine were within the normal range at the time of the first rituximab in all patients. Forty-two patients had a history of steroid-resistant nephrotic syndrome and had achieved complete remission with intensive therapy such as CsA and methylprednisolone pulse therapy.

Six of 82 patients (7.3%, five male, one female; median age at first rituximab infusion 16.0 years; range 4.0–19.9 years) had relapses during B-cell depletion after receiving rituximab (relapsed group). The remaining 76 patients (53 male, 23 female; median age at first rituximab infusion 12.2 years; range 3.1–21.8 years) had no relapses during B-cell depletion (non-relapsed group). The six patients with relapses had received multiple rituximab treatments (23 times in all) because of their refractory clinical courses (Table 1). As shown in Table 1, eleven of the 21 initial relapses after receiving rituximab occurred during B-cell depletion. Three patients had repeated relapses during B-cell depletion (patients 2, 3, 4). In particular, Patient 3 developed relapses during B-cell depletion after every rituximab treatment. Five of these six patients had relapses in spite of treatment with immunosuppressive agents after rituximab infusion. Summaries of these patients’ clinical courses are presented in Fig. 1.

Table 1 Characteristics of the first relapses after each rituximab therapy in the six patients who experienced relapse during B-cell depletion
Full size table
Fig. 1
figure 1

Clinical courses of patients who had relapses during B-cell depletion. Stars indicate relapses and filled stars relapses during B-cell depletion. Half of these patients (Patients 2, 3, 4) had repeated relapses during B-cell depletion and five of them had relapses while receiving immunosuppressive agents after RTX infusion

Full size image

Characteristics of patients who had relapses before the first administration of rituximab

To elucidate variables associated with relapse during B-cell depletion, we compared selected characteristics between the relapsed and non-relapsed groups. Table 2 shows characteristics of each patient in the relapsed group before the first rituximab and summarized data for the non-relapsed group. As shown in Table 2, age of onset, age at first rituximab, history of steroid resistance, histology on renal biopsy, type of immunosuppressive agent prior to rituximab, and number of relapses in the year before the first administration of rituximab did not differ significantly between the two groups.

Table 2 Patient characteristics before the first administration of rituximab
Full size table

Characteristics of patients who had relapses after receiving rituximab

We also compared the clinical courses of the two groups after rituximab. In the relapsed group, the median time to initial relapse during B-cell depletion was 85 days after rituximab, which is significantly shorter than that in the non-relapsed group (410 days, p = 0.0003) (Table 3). The median annual numbers of relapses after rituximab in the relapsed and non-relapsed groups were 2.5 and 0.9, respectively (p < 0.0001).

Table 3 Characteristics of initial relapses in patients who did and did not have relapses during B-cell depletion
Full size table

As shown in Table 1, five patients in the relapsed group also had 10 relapses after B-cell recovery; their median time from B-cell recovery to initial relapse was significantly shorter than that in the non-relapsed group (31 vs. 161 days, p = 0.014). However, time to ceasing steroids or commencing a maintenance dose after rituximab, duration of B-cell depletion after rituximab, and use of immunosuppressive agents after rituximab did not differ between the two groups.

Adverse effects

There were no life-threatening adverse effects in any of the 82 patients. Approximately half of them had mild acute infusion reactions to rituximab; there were no severe reactions. Agranulocytopenia occurred in one patient and was successfully treated with granulocyte colony-stimulating factor. No other adverse events were observed.

Discussion

Many studies have demonstrated that rituximab has significant efficacy in patients with refractory SDNS; however, there have been no studies regarding relapses during B-cell depletion after administration of rituximab. Six patients (7.3%) in our cohort, some of whom were receiving ongoing maintenance therapy with immunosuppressive agents, had relapses during B-cell depletion. Interestingly, five patients who had relapses during B-cell depletion also had 10 episodes of relapse after B-cell recovery; their median time from B-cell recovery to initial relapse was significantly shorter than that in the non-relapsed group (31 vs. 161 days, p = 0.014). Thus, these patients’ relapses after rituximab were less B-cell dependent than those of the remaining patients. Rituximab appeared to be relatively ineffective in these patients.

Our findings suggest that various pathophysiologic mechanisms play a part in childhood nephrotic syndrome. While many studies have shown that rituximab can accomplish sustained long remissions of idiopathic nephrotic syndrome, its mechanisms of action are still unclear, unlike with autoantibody-mediated disease such as antineutrophil cytoplasmic autoantibody vasculitis [10]. It has been assumed that proteinuria in nephrotic syndrome is mediated by unrecognized nephrotic factors, perhaps from T cells [11]. Crosstalk between B cells and T-helper T cells may be involved in the pathophysiology. Therefore, both calcineurin inhibitors targeting helper T cells and RTX targeting B cells can be of clinical benefit; rituximab may alter the status of T cells, including regulatory T cells (Treg) [12].

Patients with IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X-linked) syndrome associated with genetic mutation of the FOXP3 gene, which is characterized by dysfunction of Treg, have a propensity to develop nephrotic syndrome [13]. Additionally, in idiopathic nephrotic syndrome, decreases in Treg numbers are significantly associated with relapse and increases are characteristically observed at the time of remission [14, 15]. Rituximab is known to contribute to restoration of Treg numbers and up-regulation of their functions in patients with idiopathic thrombocytopenic purpura, systemic lupus erythematosus, and cryoglobulinemia vasculitis [16,17,18,19,20,21]. Although we did not examine T-cell status or Treg in this study, we believe that further investigation is necessary to elucidate the mechanism of relapses that are relatively B-cell independent. A variety of responses to medications results in some patients developing T-cell skewed nephrotic syndrome, whereas others develop B-cell skewed nephrotic syndrome.

This study had some limitations: a small number of patients, a heterogeneous patient cohort, a relatively short follow-up, and the possibility of residual B cells in the body after rituximab. Because CD20 is internalized after binding to rituximab, we used CD19 (another B-cell antigen) as an established reliable marker for monitoring peripheral B-cell numbers after rituximab treatment.

Because most patients in this study had received only a single dose of rituximab (375 mg/m2, maximum 500 mg), we have to consider the possibility of failure to completely eliminate all CD20-positive cells. Pharmacokinetic studies in patients with rheumatoid arthritis have shown that rituximab follows a two-compartment model with a terminal half-life of 19–22 days [22, 23]. The terminal half-life of rituximab is 2.7-fold longer after the fourth than after the first dose [24]. Prospective studies in patients with nephrotic syndrome have shown that one dose of rituximab results in profound B-cell depletion (<1% of leukocytes, or 0–5 cells/mm3) in 83.3–90.0% of patients and that B-cell depletion lasts for 4–6 months after a single dose of rituximab [3, 25]. In this study, all patients in both relapsed and non-relapsed groups did achieve peripheral blood B-cell depletion and the duration of that depletion was consistent with these reports [3, 25]. Ramos et al. reported that one infusion of 150 mg/m2 of RTX is enough to deplete CD20-positive cells from the spleens of patients who have received ABO-mismatched renal transplants [26]. Additionally, there is some evidence that lack of response to rituximab is associated with persistence of memory B-cells in specific body compartments, including bone marrow, even when B-cells cannot be detected in peripheral blood [27]. In patients with acute idiopathic thrombotic thrombocytopenic purpura, relapses can occur despite peripheral blood B-cell depletion [28]. Therefore, peripheral blood monitoring may not always accurately reflect B-cell status. Means of assessing overall B-cell depletion require investigation.

In conclusion, 7.3% of our cohort had relapses during B-cell depletion despite ongoing maintenance therapy with immunosuppressive agents. Our findings suggest that some patients with childhood-onset nephrotic syndrome have relapses that are relatively B-cell independent. These findings shed light on the heterogeneity of the pathophysiology of childhood-onset nephrotic syndrome. Further immunological investigation of T cells, especially Treg, is necessary to elucidate the mechanism of B-cell independent relapses. Development of novel treatments for patients with intractable disease is also necessary.