Introduction

To avoid or reduce steroid toxicity in steroid-sensitive childhood idiopathic nephrotic syndrome (SSNS), there is convincing data for the use of oral cyclophosphamide in patients with frequent relapses [1]. Evidence, however, is less stable for the treatment of steroid-dependent nephrotic syndrome (SDNS), i.e. recurrence of nephrotic syndrome during or within 2 weeks after cessation of steroid treatment [2, 3]. One of the major concerns with regard to the use of alkylating agents such as cyclophosphamide or chlorambucil in children and adolescents is gonadotoxicity [4, 5]. Therefore, cyclosporine A (CSA) has been advocated when toxic effects of prednisone and cyclophosphamide are expected. CSA results in a remission rate of 85% in children with SDNS but bears, however, the risk of calcineurin-inhibitor-induced nephrotoxicity [6, 7]. Thus, alternative immunosuppressive drugs are under investigation, such as mycophenolate mofetil (MMF) [8, 9], tacrolimus [10], and sirolimus [11]. Levamisole has been found to be of some benefit in frequent relapsing nephrotic syndrome, has limited toxicity, but is difficult to obtain [12] (Fig. 1)

Fig. 1
figure 1

Course of treatment in childhood steroid-sensitive nephrotic syndrome (modified according to [10]). Evidence levels [3] are shown as follows: Level 1 randomized clinical studies with robust clinical endpoint. Level 2 randomized clinical studies with surrogate parameter as endpoint. Level 3 nonrandomized, controlled studies. Level 4 therapy studies with before–after comparison or case series with historical control group. Level 5 case series with >10 patients. Level 6 case reports with < 10 patients

Full size image

In the last 3 years, a number of case reports have been published suggesting the use of the anit-CD20 monoclonal antibody rituximab in SDNS. It is the objective of this editorial commentary to discuss current data on the use of rituximab and the structure of a potential clinical trial in this field.

Role of B- and T-lymphocyte function in steroid-sensitive nephrotic syndrome

Figure  2 presents a schematic overview of the role of B- and T-lymphocyte function in steroid-sensitive nephrotic syndrome (SSNS).

Fig. 2
figure 2

Schematic overview of immunological mechanisms in steroid-sensitive nephrotic syndrome (SSNS) and potential mechanisms of action of established immunosuppression. Lymphocytes could in initiate podocyte damage by a “toxic” substance (cytokines, direct toxicity, interaction of lymphocytes and cytoskeleton). Whereas steroids suppress macrophages and T cells, including their cytokine production, cyclosporine A (CSA) specifically inhibits calcineurin-mediated activation of CD4+ helper T cells. Cyclophosphamide induces generalized depletion of lymphocytes, including T and B cells. Rituximab specifically depletes B cells only, although T-cell involvement seems possible

Full size image

It is beyond the scope of this editorial to summarize immunological findings in SSNS. However, for more than 30 years SSNS was thought to be primarily a disorder of T-cell function. This was based on clinical findings such as recurrence of nephrotic syndrome after measles infection, association with Hodgkin’s disease, and the relapsing course often triggered by allergies or infection [13]. The efficacy of calcineurin inhibition in the treatment of SSNS supported the T-cell-specific hypothesis as well as numerous in vivo and vitro experimental data showing alterations and dysregulation of T-cell populations, activation, and cytokines in SSNS. The induction of nephrotic syndrome in rats secondary to the infusion of supernatants of T cells is one of the most convincing arguments for the role of T cells in NS [14]. It was speculated that secretion of a toxic substance by lymphocytes could be the cause the increase in glomerular permeability [13] (Fig. 2).

More recently, however, data have accumulated that strongly suggest a role of B-cell immunity in SSNS. It was suggested that the permeability factor mentioned above could be an immunoglobulin and therefore reinforces the concept that B cells play a role in SSNS [15]. A reduction of total B cells in remission of NS has been found, but much more interestingly, elevated numbers of CD19+ cells were present in steroid-treated patients during remission. This is in sharp contrast to T cells, which are profoundly suppressed by steroids [16]. Furthermore, studies of B-cell activation in SSNS have shown an increased activation of T and B cells in relapse of SSNS and, again, patients with steroid dependency showed the highest rate of B-cell activation [17]. Taken together, these immunological data of B-cell dysfunction, especially in steroid dependency, indicate that B cells, also, could secrete the toxic substance suggested by Shalhoub, either directly or indirectly through communication with T cells. A direct interaction with podocytes might also be possible (Fig. 2). Steroid treatment alone seems to be unable to suppress B cells (and B-cell activation) sufficiently in patients who developed steroid dependency [16, 17].

Current clinical experience of rituximab

Remuzzi and coworkers [18] reported the first successful rituximab treatment in nephrotic syndrome in adults with idiopathic membranous nephropathy. In the last 3 years, several published case reports have also indicated the efficacy of rituximab in idiopathic nephrotic syndrome, in recurrent FSGS in renal transplants but even more so in steroid-sensitive minimal-change nephrotic syndrome (MCNS) (Table 1). In all steroid-sensitive patients, rituximab was used as rescue treatment for very complicated patients [19–31].

Table 1 Details of eight cases of rituximab (RTX) treatment in patients with childhood-onset severe nephrotic syndrome
Full size table

Primary nephrotic syndrome

One case of successful treatment has been reported in a young adult with multirelapsing MCNS with the first nephrotic episode in childhood, who had the features of steroid dependency and continued to relapse into adulthood [24]. In addition to detailed published data, even more patients have been described in abstracts [25, 26] (not listed in Table 1).

In none of the reports on rituximab use have significant adverse events been recognized so far. However, with the small number of patients reported, no conclusive data are available. Nonetheless, more than 500,000 patients worldwide now have received rituximab, and serious adverse reactions have occurred only in a small minority of patients, whereas in the great majority, rituximab was safe and well tolerated [32]. It should be noted, however, that the current antibody is not humanized, so that a potential risk for allergic reactions exists. In addition, there are recent reports on progressive multifocal leukoencephalopathy in systemic lupus erythematosus (SLE) treated with rituximab (https://doi.org/www.fda.gov/cder/drug/advisory/rituximab.htm). However, this disease has been reported quite frequently in patients with SLE and might therefore not be inflicted by rituximab.

The duration of remission in patients with nephrotic syndrome lasted between 9 and 28 months after the application of rituximab. For the 9-month period, this is consistent with the duration of B-cell suppression by the anti-CD20 antibody. Longer-lasting remissions, however, may not merely been explained by the abolition of B cells. In most patients, immunosuppressive treatment was continued (Table 1). Also in most cases, rituximab was given in a dose recommended for patients with B-cell lymphoma, i.e. 375 mg/m2 per dose at weekly intervals over a period of 4 weeks. Only in one report was a single dose applied that, nonetheless, resulted in the same duration of remission compared with the complete treatment course with anti-CD20 antibodies [23]. This patient relapsed 13 months after rituximab while on prednisolone 10 mg on alternate days and tacrolimus 2 mg b.d. (weight 85 kg). His CD19+ count was 0.43 × 109/l prior to his first dose and was first detected in the peripheral blood 4 months after rituximab, at a concentration of 0.14 × 109/l. At 9 months, the level was 0.2 × 109/l, but a month before the latest relapse, it had risen to 1.19 × 109/l. This relapse was again resistant to high-dose steroids (80 mg/day) and an increased dose of tacrolimus (3 mg b.d. and trough level 9.1 μg/l). He was given a single dose of rituximab (375 mg/m2) and within a week his proteinuria had resolved (personal communication G. C. Smith (2007), Cardiff).

One patient was primarily treated for acute idiopathic thrombocytopenia and experienced a relief of SDNS at the same time [19].

One major problem, however, is that so far, only positive outcomes have been reported with regard to the use of rituximab in childhood nephrotic syndrome. Therefore, publication bias cannot be excluded at the moment. In that context, two children were communicated to the authors (personal communication N. Webb (2007), Manchester) who did not respond to rituximab (4 × 375 mg/m2), although the biological effect could be proven by an absence of CD19 count after the first course of treatment. The first patient with SDNS was 13 years old and had received prednisolone, alkylating substances, levamisole, vincristine, CSA, and MMF. The patient had a history of nephrotic syndrome of 6 years and 4 months before receiving rituximab. Renal biopsy had shown minimal changes and global sclerosis in 2/30 glomeruli. The second patient not responding to rituximab was 7 years old, of black UK ethnic origin, and was secondary steroid resistant after having responded initially. Renal biopsy had shown minimal changes. He had received prednisolone, methylprednisolone, alkalyting substances, levamisole, CSA, tacrolimus, and MMF previously. After rituximab, there was no effect on proteinuria whatsoever. These two cases illustrate that nonresponders to rituximab must be expected in the group of patients with primary nephrotic syndrome.

Recurrent FSGS after renal transplantation

Three children showed remission of nephrotic syndrome due to recurrence of FSGS in the renal graft [20, 21, 27]. Two of these patients went into remission while being treated for posttransplant lymphoproliferative disorder (PTLD) (Table 1). Similar results were seen in two adults with recurrent FSGS after renal transplantation [29, 30; not shown in Table 1] Only very recently, two cases of treatment failure after recurrent FSGS in renal transplants were reported [31].

Necessity of a clinical trial

Despite the encouraging case reports, systematic clinical trials using anti-CD20 antibodies for SDNS are currently lacking. Many clinical issues need to be resolved by a prospective study, such as the optimal dose of rituximab. Probably, dose reduction is feasible in nephrotic patients, as complete and long-lasting depletion of B cells has been demonstrated after one single dose only [21]. A pilot study could include patients with treatment-refractory SDNS, e.g. those with persistent relapses on continuous treatment with CSA. However, it should be considered to extend the pilot phase to patients with less complicated courses, e.g. those with frequent relapses and/or steroid dependency after unsuccessful cytotoxic treatment with cyclophosphamide. If rituximab has low toxicity and no long-term side effects, it could be a drug to rationally replace the toxic and not infrequently ineffective cyclophosphamide, as well as CSA therapy with its unpredictable long-term side effects. Until a clinical study is performed, a registry might help in gathering unbiased information on the use of rituximab.

Conclusions

In conclusion, the use of the anti-CD20 antibody rituximab provides intriguing perspectives for the treatment of SDNS in childhood, especially with regard to reducing the adverse effects inflicted by the current use of cyclophosphamide and CSA. The route of application and long-term effect is advantageous in situations of noncompliance. However, whether and at what time the currently used immunosuppressants should be substituted by rituximab can only be answered by randomized clinical studies.