Rituximab in adult minimal change disease and focal segmental glomerulosclerosis-What is known and what is still unknown?

Primary forms of minimal change disease and focal segmental glomerulosclerosis are rare podocytopathies and clinically characterized by nephrotic syndrome. Glucocorticoids are the cornerstone of the initial immunosuppressive treatment in these two entities. Especially among adults with minimal change disease or focal segmental glomerulosclerosis, relapses, steroid dependence or resistance are common and necessitate re-initiation of steroids and other immunosuppressants. Effective steroid-sparing therapies and introduction of less toxic immunosuppressive agents are urgently needed to reduce undesirable side effects, in particular for patients whose disease course is complex. Rituximab, a B cell depleting monoclonal antibody, is increasingly used offlabel in these circumstances, despite a low level of evidence for adult patients. Hence, critical questions concerning drug-safety, long-term efficacy and the optimal regimen for rituximab-treatment remain unanswered. Evidence in the form of large, multicenter studies and randomized controlled trials are urgently needed to overcome these limitations.


Keywords:
Nephrotic syndrome Rituximab Minimal change disease Focal segmental glomerulosclerosis Long-term remission Infections A B S T R A C T Primary forms of minimal change disease and focal segmental glomerulosclerosis are rare podocytopathies and clinically characterized by nephrotic syndrome. Glucocorticoids are the cornerstone of the initial immunosuppressive treatment in these two entities. Especially among adults with minimal change disease or focal segmental glomerulosclerosis, relapses, steroid dependence or resistance are common and necessitate re-initiation of steroids and other immunosuppressants. Effective steroid-sparing therapies and introduction of less toxic immunosuppressive agents are urgently needed to reduce undesirable side effects, in particular for patients whose disease course is complex. Rituximab, a B cell depleting monoclonal antibody, is increasingly used offlabel in these circumstances, despite a low level of evidence for adult patients. Hence, critical questions concerning drug-safety, long-term efficacy and the optimal regimen for rituximab-treatment remain unanswered. Evidence in the form of large, multicenter studies and randomized controlled trials are urgently needed to overcome these limitations.

Background
Nephrotic syndrome (NS) describes a clinical condition with heavy proteinuria, hypoalbuminemia, edema, hyperlipidemia and other associated features such as coagulation disorders and acute kidney injury. Minimal change disease (MCD) and primary focal segmental glomerulosclerosis (FSGS) are among the leading causes of primary NS in adults. The overall incidence of these glomerulopathies is very low (0.6-1.8 per 100,000 adults for MCD and FSGS) [1]. Uncertainty exists about the pathogenesis and optimal treatment strategies for MCD and FSGS, especially among adult patients [2]. In this review, we focus on the potential role of rituximab in adult MCD and primary FSGS.
Especially in childhood, the umbrella term idiopathic NS is still often used collectively for primary MCD and FSGS. MCD is the most common cause of idiopathic NS in childhood, accounting for up to 90% of cases. Because of its favorable course and high response rates to oral corticosteroids, renal biopsy is not performed routinely and the disease is thus termed 'steroid-sensitive nephrotic syndrome' in pediatric patients [3]. Prevalence and steroid responsiveness however decrease with age. Consequently, MCD accounts for only 10%-15% of adult patients with idiopathic NS and is associated with a longer time to remission under corticosteroid treatment and a higher risk of acute kidney injury [3,4]. Nonetheless, overall long-term prognosis for adults with MCD is still excellent with remission rates of 75%-90% and a low risk for end stage kidney disease (ESKD) < 5% [5]. The diagnosis MCD implies the absence of glomerular lesions on light microscopy (LM).
Immunofluorescence for deposits is usually negative and extensive podocyte foot process effacement is detected by electron microscopy (EM) [3].
Primary FSGS is thought to be an immune-mediated podocytopathy accompanied by a characteristic focal pattern of glomerulosclerosis with five different histologic variants on LM. Similar histological findings on LM are observed in patients with secondary forms of FSGS which are a sign of adaptive glomerular changes due to excessive nephron workload and hyperfiltration [6][7][8]. In addition to adaptive glomerular changes, the term secondary FSGS encompasses various other conditions such as genetic, virus-associated or drug-induced forms, that often are difficult to distinguish from true idiopathic/primary FSGS and should be managed conservatively, aiming to control blood pressure and proteinuria [6,9]. Hence, establishing the correct diagnosis is crucial for the management of patients and an important hurdle in the recruitment of respective clinical trials. Interpretation of FSGS-studies is complex as mixed cohorts with primary and secondary disease forms are a frequent limitation [10,11]. In the following sections, we focus on primary FSGS (from this point abbreviated by 'FSGS' only) and the potential responsiveness to B cell-depleting therapies. In contrast to MCD, FSGS is generally a progressive disease with a tendency to develop ESKD. The strongest predictor of ESKD is resistance to immunosuppressive therapy and failure to achieve remission, with 5and 10-year kidney survival rates averaging 65% and 30% among these patients, respectively [12]. Even achieving partial remission is associated with significantly improved kidney survival rates compared to no remission [13,14].

Pathophysiology of disease
Historically, both terms are meant to describe two separate disease entities but there is an ongoing debate whether MCD and FSGS may represent one disease continuum with a common pathogenesis and a histological development from MCD to FSGS due to progressive podocyte injury [15]. The underlying mechanism of MCD and FSGS is not fully understood. Initial podocyte injury is regarded to be crucial in the pathogenesis of both MCD and FSGS but several causes are being discussed that might lead to podocyte depletion [6]. As depicted in Fig. 1, a dysregulation of adaptive immunity appears to be central in the pathogenesis of MCD and FSGS. The following section will outline different immunological processes involved.

Circulating permeability factors
There is some evidence suggesting that the presence of one or more circulating permeability factors that disrupt the podocyte integrity in primary FSGS [16,17]. The strongest clinical support for this hypothesis in FSGS is the high recurrence rate after kidney transplantation of about 30%-40% of patients, which may occur within hours of transplantation [18]. Circulating factors such as soluble urokinase plasminogen activator receptor, cardiotrophin-like cytokine factor-1 and anti-CD40 antibodies have been proposed as potential causative for FSGS but none have been validated and proven for their pathogenic role yet [19,20]. For MCD there is generally less evidence indicating the presence of circulating factors. Nonetheless, clinical observations made in 1974, such as remission induced by measles infection, occurrence of MCD in Hodgkin's disease and clinical response to immunosuppressive therapy, suggest a systemic abnormality of T cell function resulting in the secretion of a circulating factor [21]. The presence of a vascular permeability factor produced by lymphocytes in MCD patients was first described by Lagrue et al. in 1975 [17,22]. Since then various candidates such as the cytokines, interleukin (IL)-8 and IL-13, circulating hemopexin or podocyte-secreted, hyposialated angiopoietin-like 4 have been identified [23][24][25][26], but existing data remains inconsistent and unvalidated [17].

Atopy and eosinophils in the nephrotic syndrome
Involvement of T helper 2 (T H 2) cell-mediated immunity in the pathogenesis of idiopathic NS has been suggested [27]. This hypothesis matches with the frequently described associations between MCD and atopy, as the latter disposition is well-known to be driven by T H 2 cell responses and elevated levels of IgE, IL-4 and IL-13 [28][29][30][31][32]. In children with idiopathic NS, serum levels of IgE and IL-13 correlated with disease activity, independently of atopic co-morbidities suggesting a coincidence of atopy in children with idiopathic NS rather than a causal linkage [31][32][33][34]. Peripheral blood eosinophils are another classical diagnostic marker implicated in the pathogenesis of atopic diseases but its role in idiopathic NS is still unclear [35]. We suggest further research should aim at investigating the potential link of eosinophilia with idiopathic NS and the impact on clinically meaningful endpoints.

Possible contact points with rituximab -B cells and beyond
The role of B cells in the pathogenesis of MCD and FSGS has gained attention, mainly due to the successful use of B cell depleting agents, as further discussed below. This sheds new light on pathophysiological processes involved in these two glomerular diseases. Recently, an IgGantibody directed against Ubiquitin Carboxyl-Terminal Hydrolase L1 (UCHL1) was shown to cause podocyte detachment and associated with relapses of idiopathic NS in mice [36]. In a murine model, B cells activated locally in the kidney were able to induce glomerular injury and proteinuria by production of IL-4. In contrast, IL-4-deficient B cells did not induce proteinuria, whereas overexpression of IL-4 alone was sufficient to cause foot process effacement and proteinuria. Increased B cell numbers were present in the kidneys of children with FSGS compared with mesangio-capillary glomerulonephritis [37]. In kidney biopsies of patients with MCD, STAT6 activation (induced by IL-4) was increased, suggesting IL-4 exposure in these patients [38]. These observations suggest that B cells are involved in the pathogenesis of MCD and FSGS, possibly by production of cytokines, such as IL-4 [38]. Interestingly, IL-4 is the critical cytokine contributing to the risk of atopy and allergic diseases, triggering Ig class-switch to produce IgE and T cells to CD4 + T H 2 cells [39]. Recently, Shimada et al. postulated a "two-hit" theory to describe the pathogenesis of MCD. Accordingly, cytokines (e.g. IL-13), microbial products or allergens induce a direct stimulation of podocytes leading to the induction of CD80 (also named B7.1) as the initial hit. This causes an alteration of the podocyte structure and increases permeability. Under normal conditions, CD80 expression on podocytes is controlled by T regulatory (Treg) cytokines or production of cytotoxic T-lymphocyte antigen 4 (CTLA-4) and IL-10 by the podocyte itself. Hence, a second hit due to Treg dysregulation or impaired podocyte autoregulation is crucial and leads to sustained podocyte injury and MCD [40]. Modifications of T cell subsets were observed in patients enrolled in a multicenter, double-blind, and randomized trial assessing the efficacy of rituximab compared to a control group in childhood MCD. A decrease of CD4 + CD25 high FoxP3 high Treg cells was associated with disease relapses, while rituximab-treated patients had a low relapse rate accompanied by a low frequency of CD4 CD8 double negative (CD3 + CD4 − CD8 − ) invariant natural killer T (iNKT) cells expressing an invariant T cell receptor α-chain (Vα24). Rituximab was shown to specifically reduce the frequency of CD4 + follicular T cells (TFH cells) that drive naïve and memory B cells to differentiate into antibody-secreting cells (see Fig. 1). As relapses are associated with rapid reconstitution of switched memory B cells, the authors conclude that B cell depletion might induce qualitative alterations of TFH cells and thus inhibit reconstitution of switched memory B cells [41]. Recently, genome-wide association studies identified different HLA and non-HLA risk loci involved in adaptive immunity for childhood-onset steroid-sensitive NS, supporting the concept of an underlying immune dysregulation in the disease pathogenesis [42][43][44].

A hypothetical mode of action of rituximab in MCD and FSGS
B cells are involved in the pathogenesis of primary MCD and FSGS as illustrated in Fig. 2, and presumably, play more an indirect role, either through production of a circulating factor as proposed mainly for FSGS or by interaction with T cells as described in MCD.
Rituximab, a chimeric, monoclonal IgG1 antibody, exerts its B celldepleting effects via binding to CD20. Besides direct, cytotoxic effects (complement-and antibody-dependent), a diversity of indirect effects, including lipid raft modifications, kinase and caspase activation and effects on apoptotic/antiapoptotic molecules appear to play a crucial role for the observed variability in response to rituximab treatment [45]. In addition, other B cell-independent mechanisms might be responsible for the antiproteinuric effects of rituximab seen in patients with renal disease. Besides modifying effects on T cell subsets, RTX might also lead to a depletion of a small number of T cells that coexpress both CD3 and CD20 (approximately 3%-5% of all circulating T cells), as reported in patients with multiple sclerosis [46]. A direct regulation of podocyte function has been postulated and sphingomyelin phosphodiesterase acid-like 3b (SMPDL-3b) proposed as such target [47]. Notably, while binding of rituximab to SMPDL-3b was shown in previous studies using fixed sample assays, Kim et al. could not reproduce this in an experimental in vivo setting with nonfixed cells [38]. In patients at high risk for recurrent FSGS after kidney transplantation, rituximab treatment was associated with a lower frequency of posttransplant nephrotic range proteinuria compared to historical controls. Additionally, rituximab partially prevented SMPDL-3b downregulation observed in podocytes treated with sera of patients with recurrent FSGS after transplantation [48]. In contrast, a benefit in preventing disease recurrence after transplantation was not confirmed in an observational study [49]. Krüppel-like factor 15, a zinc-finger transcription factor expressed in human podocytes, contributes to mediate the beneficial effects of glucocorticoid therapy via stabilization of the actin cytoskeleton of podocytes in both MCD and FSGS [50]. To date, the relevance of this finding in the discussion of rituximab efficacy is unclear.

Immunosuppression in MCD and FSGS
Existing treatment strategies are generally supported by a weak level of evidence as reflected in the current KDIGO Clinical Practice Guidelines for Glomerulonephritis 2012. Evidence is particularly scarce for adult MCD and FSGS, as most available data is derived from clinical trials in children and from observational studies. Supportive measures, such as antihypertensive, antiproteinuric and dietary approaches, are pivotal for all patients with proteinuric glomerular diseases [2]. MCD and FSGS generally show poor response to these measures, thus initial immunosuppressive therapy with corticosteroids is recommended for all patients with clinical features of nephrotic syndrome (evidence grade 1C) [12,51]. Current definitions for NS, complete (CR) and partial remission (PR) and corresponding disease courses are given in Table 1.

Autoimmunity Reviews xxx (xxxx) xxxx
Considering the toxicity of such high steroid doses, such recommendations are still controversial [59]. Conversely, a short-term (2 months) steroid regimen may be an effective treatment option for adult steroid-sensitive MCD patients [60]. Calcineurin inhibitors (CNI), alkylating agents and antimetabolites are commonly used steroid sparing agents in MCD and FSGS. Tacrolimus monotherapy [61] or in combination with low-dose steroids [62] may be an option to a conventional glucocorticoid regimen as initial therapy for adult MCD. Nonetheless, these alternative agents are so far only recommended as second-or third-line agents in adults with FR or SD MCD and FSGS [12,51,59].
Cyclophosphamide, cyclosporine alone or cyclosporin in a steroidcombined regimen have all shown efficacy as first-line therapy in adult patients with SD or FR MCD [63][64][65]. In a Chinese study including adult patients with SR/SD MCD, tacrolimus was equally effective compared with cyclophosphamide in inducing and maintaining remission and might induce remission more rapidly than intravenous pulse cyclophosphamide [66,67]. Whereas CNIs block T cell activation and stabilize actin directly in podocytes, mycophenolate mofetil has an antiproliferative effect on both B and T cells. In a retrospective review of adult MCD cases by Waldman et al., mycophenolate mofetil was studied as second-line agent in 14 patients (7 SD, 4 SR, 3 partial steroid responders) showing remission in 64% of patients (6/7 SD, 1/4 SR) [4]. A recent multicenter clinical trial found no superiority of a low-dose oral steroid therapy plus mycophenolate mofetil compared with a conventional oral steroid therapy alone after 4 weeks of treatment, at least for the first episode of MCD [68].
In FSGS, one randomized controlled trial (RCT) showed that cyclosporin in combination with low-dose prednisone is superior to prednisone alone in preserving renal function and achieving CR or PR [69]. However, the active treatment period with cyclosporin was only six months in this study and relapses were common after remission, occurring in 60% of remitters by week 78 [69]. Another five-year followup study showed higher remission rates of cyclosporin with prednisolone (85.7%) compared to azathioprine with prednisolone (80%) and prednisolone alone (62.5%) [54]. Tacrolimus monotherapy might be another option as first-line therapy for adult FSGS [70]. The efficacy of mycophenolate mofetil alone compared to steroids has never been tested. One RCT compared a 12-months course of cyclosporin to a combination of oral pulse dexamethasone and mycophenolate mofetil in 93 children and 45 young adults with SR FSGS and used six primary outcome levels of proteinuria response. No difference was seen in achieving PR or CR in the first year (primary outcome) in both, adults and children. Nevertheless, good outcome levels (level one to three) were achieved in 45.9% of cyclosporin treated and 33.3% in the combined dexamethasone/ mycophenolate mofetil group. Additionally, there was no significant difference in maintaining remission for 26 weeks after cessation of treatment (main secondary outcome) among the two subgroups [71]. As the study was underpowered and secondary forms of FSGS were likely included, results may be interpreted cautiously [72]. One open label randomized pilot study from India compared the efficacy (measured end point was change in urinary protein/ creatinine ratio) of mycophenolate mofetil in combination with prednisolone to prednisolone alone. After 6-months, the mycophenolate mofetil-based regimen was as effective as steroids alone but remission was induced faster and steroid exposure could be reduced [73]. Additional treatment with chlorambucil in combination with steroids showed a lack of efficacy in a RCT [74]. In complicated disease courses, including SD, SR or FR FSGS, the same agents have been evaluated in a few randomized and observational studies [55]. One RCT compared tacrolimus and cyclophosphamide, both in combination with prednisone, in adult patients with SR or SD FSGS and showed similar efficacy of both regimens. Six-and twelve-month remission rates were 67% and 73% in the tacrolimus-group, compared with 56% and 67% in the cyclophosphamide-group [75]. Other immunosuppressive agents have been tested in small trials or case series such as levamisole, mizoribine, tumor necrosis factor-α inhibitors, pirfenidone, fresolimumab, saquinivir and sirolimus [6] and the results have been reviewed elsewhere [57].
Each of the described steroid sparing agents has major disadvantages. CNI and antimetabolites show frequent relapse rates after withdrawal in both MCD and FSGS [76]. CNI lead to nephrotoxicity with prolonged use and serum levels need to be monitored tightly whereas mycophenolate mofetil has dose-dependent effects on the gastrointestinal system and bone marrow [3]. Alkylating drugs like cyclophosphamide are very efficient but have major toxic side effects including infertility, urotoxicity and oncogenicity [77]. Neoplastic complications of alkylating drugs may occur years or even decades after treatment exposure [78]. Thus, most trials may have underestimated the true number of malignancies attributable to the treatment and thus the malignancy risk of these agents. In view of the availability of modern, effective and well tolerated drugs these limitations argue for alternatives in the management of MCD and FSGS.

Rituximab
Rituximab is approved for the treatment of non-Hodgkin's lymphoma, rheumatoid arthritis (RA) and ANCA-associated vasculitis (AAV) by the U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA). Beyond that, rituximab is increasingly used off-label in various autoimmune diseases, including myasthenia gravis, systemic lupus erythematosus, idiopathic thrombocytopenic purpura and Sjögren syndrome [79].

Safety
In these indications, rituximab showed a favorable long-term safety profile [80]. Infusion-related reactions are frequent, manifesting as an immediate hypersensitivity reaction. A pooled analysis of patients with RA reported infusion-related reactions in about 25% of patients during the first infusion with less than 1% of cases considered as serious [81]. Thus, pretreatment with paracetamol, antihistamines and corticosteroids is recommended [80]. Hepatitis B virus (HBV) reactivation in HBsAg-negative/ HBcAb-positive patients following rituximab has been reported and can be prevented by HBV-screening and antiviral prophylaxis [80]. Severe infectious complications, such as Pneumocystis jirovecii pneumonia (PCP) and progressive multifocal leukencephalopathy (PML) due to reactivation of JC virus are rare but potentially fatal complications, mainly described among patients with lymphoproliferative disorders and concomitant or previous exposure to other immunosuppressive agents [80,82]. To our knowledge, no case of PML has been reported in patients treated with rituximab monotherapy. Given the fact that rituximab treatment often allows tapering or withdrawal of other immunosuppressants, the true risk of severe infections attributable to rituximab is unclear [83]. Infectious complications are typically reported in patients with pre-existing immune defects, significant comorbidities or concomitant intense immunosuppression [84]. Among patients with RA, the overall serious infection rate was 4.31 per 100 patient-years [81]. A slightly higher risk was found among 370 patients with various autoimmune diseases, with a rate of serious infections of 5.3 per 100 patient-years during rituximab therapy [85]. Importantly, infection rates vary between indications, an observation that may be at least partly explained by different steroid-regimes. A higher infection rate is seen in patients with AAV, as shown in the landmark trials RAVE and RITUXVAS [86][87][88]. In RITUXVAS, 21.2 serious infections per 100 patient-years were seen [87]. Reported rates in observational studies are even higher and trimethoprim-sulfamethoxazole prophylaxis should be considered in these patients [89,90]. In contrast, low infection rates are reported in patients with NS. In a retrospective analysis of 24 adult patients who received rituximab for membranous nephropathy (n = 11), MCD (n = 7), FSGS (n = 4) and membranoproliferative glomerulonephritis (n = 2) only one single serious infection (1.6 per P. Gauckler, et al. Autoimmunity Reviews xxx (xxxx) xxxx 100 patient-years) was reported (bronchopneumonia), but fully recovered after a course of antibiotics [91]. NEMO, a prospective off-on trial, evaluated the effects of rituximab in 10 children and 20 adults with MCD, mesangial proliferative glomerulonephritis and FSGS. After one year of follow-up 5 serious infections were observed in the adult subgroup (25 per 100 patient-years). At the time of infection, all patients were still receiving concomitant immunosuppressive treatment and all fully recovered [92]. In contrast, no serious infectious complications were observed in a number of studies [93][94][95][96][97][98]. Non-uniform reporting of serious infectious complications may explain these differences and information from RCTs are expected to inform about the true infectious risk of rituximab in these diseases. A detailed overview of infectious complications is provided in supplementary Table 1 (see  supplementary appendix). Hypogammaglobulinemia and late-onset neutropenia are two other side effects associated with rituximab [80]. Patients with nephrotic syndrome might be at particular risk for persistent hypogammaglobulinemia and, in children, low IgG levels at baseline appear to increase the risk [99,100]. Thus, monitoring levels of immunoglobulins after rituximab application may be informative, especially in patients with recurrent infections, to identify the actual risk of hypogammaglobulinemia in these patients and potentially predict the infection risk.
Long-term data evaluating the malignancy risk after rituximab in renal disease is scarce. A retrospective study analyzing the malignancy risk of 323 patients with AAV over a mean follow-up of 5.6 years showed that rituximab-treated patients had a lower malignancy risk than cyclophosphamide-treated patients. Notably, rituximab was not associated with an increased malignancy risk in patients with AAV compared with the general population [101]. In the above-mentioned safety-study by van den Brand et al., three blood malignancies and five solid cancers (two of them fatal) were observed and possibly related to the combined therapy of an alkylating agent (cyclophosphamide or chlorambucil) with corticosteroids during a period of 40 months followup of patients with membranous nephropathy. In comparison, 2 solid cancers were observed in the rituximab-group and assessed as unrelated to treatment by physicians directly in charge of the patients [102]. Corresponding long-term follow-up data for rituximab in MCD and FSGS is not available. To address this appropriately, studies with a follow-up period of 60 months or longer are needed.

Initial dosing
The optimal initial dosing of rituximab for off-label indications remains unknown. A classic four-dose protocol consisting of 4 weekly doses of 375 mg/m 2 of body surface area is used for hematologic indications, whereas a regimen of 2 applications of 1 g fixed-dose two weeks apart is used in RA. For adults with MCD or FSGS the optimal initial dosing remains unknown. Different dosing protocols (375 mg/m 2 of one to four weekly doses or 1 g once or on day 1 and 15) were used in subgroups of a few retrospective studies [93][94][95]97,103,104] and 3 prospective trials [92,96,105]. Here, either no correlation was found between the different treatment protocols [93,94] or no conclusions could be made because of the small size of the treated subgroups [95]. There is some evidence that even a single-dose of rituximab (375 mg/ m 2 ) may be effective to induce remission and reduce relapse rates in adult MCD and FSGS [92,105]. Larger, controlled trials are urgently needed to allow comparisons between different dosing protocols in MCD and FSGS.

Subsequent dosing
Rituximab usually persists in the circulation for 3 to 6 months, followed by recovery of B cells to pretreatment levels by 12 months [106]. In patients with nephrotic range proteinuria however, drug halflife is shortened due to urinary loss of rituximab [80]. Correspondingly, lower rituximab levels were reported in proteinuric patients with membranous nephropathy compared to non-proteinuric patients with RA who received the same dosing regimen [107]. Hence, a shorter halflife of rituximab among patients with NS may require respective adaptions in rituximab-dosing [108]. In MCD and FSGS, the optimal timing and dosing of rituximab-reapplication to either maintain remission or in disease relapse are not known. After B cell depletion, reemergence begins with CD19 + transitional B cells, thus measuring CD19 + B cells was proposed for monitoring rituximab-treatment in RA [109]. Another rationale supporting CD19-monitoring was a maskingeffect of CD20 antigen upon rituximab-binding. However, this could not be confirmed ultimately [110] and additional evidence exists that rituximab might interfere with CD19 expression as well [111]. Although B cell reconstitution after rituximab is associated with a certain risk for disease recurrence, remission may last despite complete B cell recovery and relapses can occur in the presence of sustained depletion of B cells [112][113][114][115]. Hence, B cell recovery after rituximab appears to be an unreliable marker to predict relapses in patients with NS [116]. Delayed reconstitution of switched memory B cells may be useful to predict relapses in those patients [116]. Additionally, an increase in the CD4 + /CD8 + T cell ratio after rituximab was observed but no association with the individual relapse risk could be drawn [116].
Currently, no guideline exists for the use of rituximab in MCD and FSGS. A possible treatment algorithm how and when rituximab may be applied in these entities is provided in Fig. 3 for MCD and Fig. 4 for FSGS.

Efficacy
Treatment with rituximab, despite the paucity of efficacy data, changed the therapeutic landscape in both diseases. While reduction/ withdrawal of steroids and concomitant immunosuppression was possible in most cases, relapse rates remained low and were significantly reduced compared to time periods before rituximab administration. Retrospective studies with long term follow-up provided evidence that relapse free survival can be achieved in the long run, which argues for an immunologic switch at least in some patients. Table 2 provides an overview of trials investigating the efficacy and safety of rituximab in MCD and FSGS. Respective RCTs for this indication are limited to pediatric studies.
Less evidence supports the use of rituximab in adult patients with MCD or FSGS. A long-term follow-up of 16 adult patients with SD/SR/ FR MCD treated with rituximab together with steroids showed CR in 13 patients enabling discontinuation or tapering of steroids. At a median follow-up of 44 months, 8 patients remained in remission and relapses occurred in 7 patients with re-application of rituximab in 4 of these, while one patient had no response to therapy [103]. In another retrospective analysis of 17 adult patients with SD/FR MCD, rituximab treatment led to a significant reduction of relapses per year (1.32 +/− 0.85 to 0.16 +/− 0.21, p < 0.05) and tapering of concomitant steroid therapy [93]. In this study, relapse rates were lower in patients who received rituximab in remission compared to those not in remission (20% and 57%, respectively). In NEMO, total relapses decreased from 88 to 22 during one year of follow-up, compared with the year before RTX application. The steroid maintenance dose per patient decreased from a median of 0.27 mg/kg to 0 mg/kg (p < 0.001). The effects were significant across all subgroups (children, adults, MCD and FSGS) [92].
A retrospective study of the Spanish GLOSEN registry compared 50 adult patients with SD/FR NS treated with steroids rituximab in combination with another immunosuppressant. While 28 patients received additional rituximab treatment, 22 patients served as control group. CR was achieved in 82% in the rituximab group versus 63% in the control group. The relapse rate per year before and after rituximab application was significantly reduced (p < 0.001) and additional immunosuppressants to achieve sustained remission were lower in the rituximab group than in the control group. The baseline relapse rate before rituximab treatment was significantly higher in the rituximab group than in the control group, thus a direct comparison between the groups was not possible [94]. Guitard et al. evaluated 41 adults with MCD receiving rituximab in a multicenter retrospective study. Overall response, defined as remission of NS and withdrawal of at least one immunosuppressant, was achieved in 78%. In the median follow-up period of 39 months, relapses occurred in 18 responders (56%). Of these, 17 received a second course of rituximab leading to clinical response. 9 patients had sustained remission even after B cell recovery [95]. In 2014, we analyzed available study data (14 studies) of adult patients (n = 86) with SD/FR MCD (n = 77) or FSGS (n = 9) regarding relapses before and after rituximab treatment. Treatment with rituximab reduced the number of relapses per year from 1.3 (0-9) before  Table 2 Overview of studies of rituximab in minimal change disease and focal segmental glomerulosclerosis in adults. rituximab to 0 (0-2) after rituximab (p < 0.001), while proteinuria decreased from 2.43 (0-15) g/day to 0 (0-4.89) g/day (p < 0.001), respectively [117]. Of note, the number of patients with FSGS included in this study was small (n = 9) and all were steroid-responsive. Roccatello et al. reported results of eight patients with FSGS having major risk factors precluding corticosteroids or conventional immunosuppression who received high-dose rituximab (8 weekly doses of 375 mg/m 2 ) and were prospectively followed up for a minimum of 2 years. Only one patient showed an improvement of proteinuria and renal function while 7 patients did not respond and had persistent proteinuria in nephrotic range [96]. The low response rate might be attributed at least partially to a high number of secondary forms of FSGS among the cohort, as the included patients were comparably old with a high rate of co-morbidities and heterogenous biopsy characteristics (only 2 out of 8 patients showed widespread podocyte foot process effacement). Similar negative findings were reported from the GLOSEN registry. Of eight patients with SR FSGS and previous failure of at least one additional immunosuppressant, only three showed a positive influence of rituximab [97]. In a recent retrospective multicenter study of older patients (≥ 60 years) receiving immunosuppressive therapy for MCD/FSGS, rituximab induced CR or PR in 18 out of 23 individuals [104]. In a recent prospective single-centre study from India, 24 patients (mean age 24.29 years, 17-48) with SD/SR MCD (n = 11) or FSGS (n = 13) and CNI-dependence, rituximab was used to maintain remission. All patients were in CR or PR at the time of rituximab -administration. After 12 months, patients with MCD and SD NS had excellent outcomes with CR occurring in 100% and 93% respectively, while patients with FSGS or SR NS could maintain CR in only 38% and 22% respectively [98]. Considering the complexity of the study population (dependence to CS and CNI) and the study design (rituximab use for maintenance treatment), comparisons of these results with other studies are limited. It should be emphasized that several of the presented studies have significant limitations, mainly due to mixed cohorts of patients with MCD and FSGS or patients diagnosed as idiopathic NS without further differentiation. Furthermore, reported patients had varying previous treatment modalities and response definitions, declared as SD, SR, or FR. Presumably, a proportion of patients diagnosed as SD/SR/FR FSGS actually exhibit an underlying secondary cause of disease and therefore show no response to any immunosuppressive agent.
Currently, three controlled trials have been initiated independently in order to assess the efficacy and safety of rituximab in adult MCD and FSGS. TURING (EudraCT: 2018-004611-50) is a randomized, doubleblind, placebo-controlled trial investigating safety and efficacy of rituximab in combination with steroids to steroids alone in 112 patients with relapsing or newly diagnosed MCD or FSGS. RIFIREINS (NCT03970577) is a French multicenter trial of 98 adult patients with a first episode of MCD comparing the efficacy of rituximab versus steroids to maintain remission. Lastly, a smaller open-label RCT (NCT03298698) will compare continued treatment with high-dose prednisone to treatment with rituximab in adult patients with MCD/ FSGS unresponsive to 8 weeks of high-dose prednisone.
Taken together, rituximab appears to be a promising strategy for FR/SD MCD to reduce relapse rates and corticosteroid exposure, although the evidence is still limited. The efficacy data in patients with FSGS are conflicting so far.

Conclusion
NS in adults due to MCD or FSGS is a heterogenous disease complex with a highly variable clinical course. Current treatment strategies are based on supportive, antiproteinuric measures and immunosuppressive agents such as steroids, alkylating agents and CNIs. Steroid-dependence/resistance or frequent relapses require prolonged drug exposure resulting in high rates of adverse events attributable to these drugs [118]. B cells are involved in the pathogenesis of MCD and FSGS, giving a rationale for a targeted therapy in both entities. While efficacy and safety of rituximab are shown for several autoimmune-mediated diseases, clinical data supporting its use in MCD and FSGS is still limited. Consequently, rituximab is still a second line treatment option for these entities, reserved for patients with a complex disease course or with contraindication for first-line agents. Several ongoing RCTs will hopefully provide evidence-based data in the upcoming years. Until then, several questions related to the use of rituximab in these patients remain unanswered [119]. Drug-safety and long-term efficacy, the optimal regimen for rituximab-application, prediction of relapses after rituximab and the role of peripheral blood eosinophils in adult MCD and FSGS are central issues that we aim to address with RITERM, a multicenter, international, retrospective study.

Author contributions
PG and AK designed the study. PG wrote the first draft of the manuscript. All other co-authors critically revised the manuscript, carried out major modifications and approved the final version.

Declaration of Competing Interest
VA reports grants from ADDMEDICA, outside the submitted work; AB reports consulting fees from Chemocentryx, from Merck/MSD and from Astra Zeneca, outside the submitted work; MB reports lecture fees from Boehringer Ingelheim, from Bristol-Myers Squibb, from Novartis, travel support from Hexal, lecture fees and travel support from Vifor Fresenius and lecture fee from Servier, outside the submitted work; CKC reports grants from GlaxoSmithKline, grants and consultancy fees from Retrophin, outside the submitted work; SF reports lecture fees from VIFOR Pharma, outside the submitted work; CK reports being funded by National Institute for Health Research (NIHR) as academic clinical fellow (ACF), during the conduct of the study. All the other authors declared no competing interests.