Lupus nephritis and ANCA-associated vasculitis: towards precision medicine?

Abstract

Historically the treatment of lupus nephritis (LN) and anti-neutrophil cytoplasmic antibody (ANCA) vasculitis was ‘one size fits all’; however, with the emergence of precision medicine initiatives, the field is moving towards more personalized treatment approaches. The recent development of a more accurate and reproducible histopathological classification system for LN could lead to better disease categorization and therefore more targeted therapies. A better understanding of the pathophysiology of LN has provided evidence that not only T but also B cells play an important role, opening new opportunities for individualized treatment approaches. Recent trials have shown calcineurin inhibitors and the anti-CD20 antibodies rituximab and ofatumumab to be effective in the treatment of LN, adding new treatment options. State-of-the-art targeted therapy in ANCA-associated vasculitis (AAV) takes interindividual heterogeneity in disease severity, type of ANCA antibody [myeloperoxidase versus proteinase 3 (PR3)] and the risk for side effects of therapy into consideration. In addition, within an individual, induction therapy differs from maintenance therapy, the same holding true in incident and relapsing disease. Rituximab is now widely used in AAV and it has become clear that prolonged B cell depletion, as in LN, must be achieved to obtain a long-lasting clinical response, especially in anti-PR3-associated disease. Still, despite these advances, molecular and genetic markers are rarely incorporated into diagnostic and treatment algorithms and true precision medicine remains an aspiration that hopefully can be achieved.

INTRODUCTION

The Precision Medicine Initiative defines precision medicine as an emerging approach for disease treatment and prevention that takes the individual variability in genes, environment and lifestyle into account. The basic principle is that the same treatment may not be effective in all patients and, especially when side effects are considered, a better characterization of patients will decrease the number needed to treat while increasing the number needed to harm [1].

Towards precision medicine in the diagnosis of lupus nephritis (LN)

The pathogenesis and manifestations of LN are heterogeneous and variable between and within individuals. This complexity probably contributed to the failure of multiple clinical trials during the last decades. About 10–30% of patients still progress to end-stage kidney disease within 15 years of diagnosis and patients with systemic lupus erythematosus (SLE), without or with kidney damage, have a >14-fold and >60-fold increased risk for premature death, respectively [2]. As the clinical presentation of different manifestations of LN is very unspecific, the diagnostic workup often includes a kidney biopsy. As the histopathological diagnosis determines the choice of treatment, we need to optimize the classification and increase the interobserver agreement among nephropathologists [3]. In 2016, a working group for LN classification met at Leiden University Medical Center with the aim of reaching a new consensus [3, 4] and proposed several modifications [3, 5]. Several areas of problematic definitions were identified and adapted. Still, the aim of this new improved classification system to better define subpopulations, which will lead to more tailored therapy, has so far not been followed by appropriate studies and therapeutic interventions. Au contraire, recent years have surfaced new attempts to overcome the challenge of disease prediction and therapeutic success.

Since kidney biopsies are invasive and have some challenges and restrictions besides the diagnostic evaluation, some authors ask the question of whether biomarkers would lead to precision medicine in LN [6]. It has been discussed whether a panel of urine, serum, tissue and genetic biomarkers could identify patients in whom LN will develop and also which inflammatory pathways are likely to be active and thus tailored therapy could be applied [6]. Recently Mejia-Vilet et al. [7] demonstrated that the urinary soluble CD163 nicely discriminates LN activity and therefore could be used as a non-invasive marker in the decision process. Other working groups have followed a more systemic approach to overcome this issue. Brant et al. [8] used a combination of system biology (protein–protein interaction network analysis via Betti number calculations) and statistical thermodynamics (via Gibbs homology with energy threshold filtering) approaches in a population of 291 lupus patients to identify ∼300 proteins that might play a substantial role. The most important class consists of ribosomal proteins and ribosomal-related proteins, followed by ubiquitin and proteins belonging to the ubiquitination pathways. Based on their findings, the authors concluded that lupus is related to ribosomal stress and it is likely that epigenetic factors such as nutritional status and exposure to sunlight may be associated aspects in the response [8].

Another approach towards precision medicine in lupus has been recently reviewed by Nagafuchi et al. [9] on the issue of immune profiling. Transcriptome analysis of whole blood or targeted immune cells discovered gene modules closely related to disease activity, disease subtype or future relapse (reviewed in Nagafuchi et al. [9]). The bulk transcriptome reflects the average expression of individual cells, but a small cell fraction, e.g. CD8 T cells, may be responsible for the transcriptome difference [10]. To fully characterize a disease-relevant small cell subset that is responsible for the transcriptome change, deeper immune phenotyping, probably by mass cytometry or single-cell RNA sequencing in combination with bulk transcriptome analysis, is needed. A recent multicentre study (the SLE Prospective Observational Cohort Study) has been launched that aims to describe the disease course of SLE and its association with type I interferon gene signature status. Data from trials like this to better define the inflammatory pathways are eagerly awaited [11].

As already described for renal biopsies, reproducibility, as well as comparable methods and evaluation, is necessary to clinically apply these new techniques as a further step in the direction of precision medicine. There is the dilemma between standardized and exploratory research protocols, so that exploratory immune profiling ‘omics’ analysis needs to be followed by standardized multicentre validation studies, as has been followed by the consensus on International Society of Nephrology (ISN)/Renal Pathology Society (RPN) guidelines for kidney biopsies [3]. A good overview of recent papers on the current state of this field is given by the homepage of the Accelerating Medicine Partnerships (AMP) group of the Lupus Foundation of America (www.lupus.org).

Towards precision medicine in LN therapy

The standard care for induction therapy in severe proliferative LN comprises prednisone at 1 mg/kg/day, followed by a taper over weeks, in combination with cyclophosphamide (National Institutes of Health or Euro-Lupus trial regimen) or mycophenolate mofetil (MMF; reviewed in Parikh and Rovin [2]). Since the 1980s, calcineurin inhibitors (CNIs) have been evaluated based on the assumption that these agents block T cell–mediated immune responses and stabilize podocytes in the kidney, thus protecting against podocytopathy and proteinuria [3]. A recent trial comparing the combination of tacrolimus and MMF with intravenous cyclophosphamide over 24 weeks showed superiority of the CNI-based regimen [12]. The extension of this trial over a period of 18 months showed no difference in the long-term results in renal relapse, but there was a significant reduction in adverse events in the tacrolimus arm [13].

Another alternative to the current standard induction therapy is a new member of the CNI family, voclosporin, a modified cyclosporine molecule, which in vitro is more potent than cyclosporin A and shows improved pharmacokinetic and pharmacodynamic properties. A large randomized controlled multicentre study (AURA-LV) including 265 patients with LN investigated the effectiveness of an induction therapy using low- and high-dose voclosporin in combination with MMF and corticosteroids. Compared with the placebo arm (MMF and corticosteroids only), the rate of complete remission in the voclosporin group was significantly increased, but an increased rate of adverse events, mainly due to infections, as well as fatal events was seen in the voclosporin arms (high and low dose) [14]. Ten fatalities occurred in the low-dose group, 2 in the high-dose group and 1 in the placebo arm, 9 of 13 occurring within the first 2 months and only 1 between Months 6 and 12 of the study period in the voclosporin group [14]. This was not confirmed by the following Phase III AURORA trial [15]. In addition, the voclosporin arm demonstrated significantly improved 1-year outcomes in two populations that have been historically difficult to treat: African American and Hispanic patients [15]. Finally, few studies evaluated the advantage or disadvantage of CNIs based on immune profiling. There is evidence that patients with a high percentage of follicular helper T cells were more resistant to treatment with CNIs [16]. This is of importance since those cells are essential for B cell maturation and autoantibody production [17].

In the last years, B cell–depleting therapeutic regimens (rituximab and recently the humanized antibody ofatumumab) have been tested in LN [18]. Experimental evidence in several mouse models has shown that after depletion of B cells, nephritis cannot develop, although the animals showed immune complex deposition in the kidney. As nephritis still develops in mouse models with B cells incapable of producing antibodies, B cells seem to function as antigen-presenting cells [19]. The first large study evaluating the efficacy of rituximab was the Lupus Nephritis Assessment with Rituximab trial [20]. Although the study showed negative results based on the primary clinical endpoint, a post hoc analysis demonstrated that in patients with a complete depletion of peripheral B cells, remission, defined by an albumin:creatinine ratio <0.5 g/g creatinine, was achieved in 47% of the patients compared with only 13% if depletion was incomplete [21]. There is also some evidence that regulatory T cells could be involved in the mode of action since, in a small trial, rituximab led to a 3.5-fold increase of those cells over 9 months [22]. Taken together, rituximab has a recognized role in LN patients, especially those who are refractory to current therapy modalities. There is a need to achieve long-lasting (>71 days) and complete depletion of peripheral B cells in order to obtain a complete clinical response.

Another substance, the monoclonal anti-CD20 antibody obinutuzumab, with superior peripheral and tissue B cell–depleting capacity compared with rituximab, showed promising outcomes in a Phase II trial (NOBILITY; reviewed in Lei et al. [23]). Besides data on the completeness of B cell depletion in patients with rituximab, only a few studies have evaluated biomarkers as predictors for therapeutic success of rituximab. One of them, the MRC MASTERPLANS Consortium, targets patient stratification and individualized treatment in LN [24]. They showed that urine levels of lipocalin-like prostaglandin D synthase, transferrin, alpha-1-acid glycoprotein, ceruloplasmin, monocyte chemoattractant peptide-1 (MCP-1) and soluble vascular cell adhesion molecule-1 correlate with active LN. More importantly, a combined model of urinary proteins may predict response to rituximab treatment at 12 months [24]. In addition, single-cell RNA sequencing of keratinocytes and urine cells in patients with LN is promising as a means to assess intrarenal gene signatures without the need for a kidney biopsy [25].

In summary, since LN definitely is a heterogeneous disease, diagnostic as well as therapeutic ‘precision medicine’ is of importance in affected patients. Recently the international Nephropathology Working Group has proposed new definitions for the histopathologic classification in line with the ISN/RPS guidelines [3]. A multitargeted induction therapy including CNIs like tacrolimus or the new formulation voclosporin have led to higher remission rates in combination with MMF and a reduced dose of steroids, which may be an option in younger individuals or female patients without prior therapy with MMF. But based on available data, diagnosis and treatment of LN as compared with the field of oncology is far from fulfilling the definition of precision medicine. The future of kidney disease classification will likely combine histological and molecular approaches, a strategy that is gaining traction in the world of kidney transplantation, where tissue gene expression demonstrated equal or sometimes better ability to discriminate between rejection and non-rejection than histological evaluation [26].

Towards precision medicine in the diagnosis of ANCA-associated vasculitis (AAV)

Diagnosis of AAV may be easy in patients with typical clinical features and ANCA positivity. In severe cases (e.g. with rapidly declining kidney function), treatment should not be delayed waiting on a kidney biopsy and should be started immediately. On the other hand, renal biopsy not only confirms the diagnosis, but also provides assessment of the activity and chronicity of the disease. Based on recent classification, four classes (focal, crescentic, mixed and sclerotic) of ANCA-associated glomerulonephritis are distinguished, with apparent consequences not only for the assessment of renal outcome, but also for considerations concerning treatment [27]. The benefits of induction treatment of AAV in elderly patients with isolated ANCA-associated glomerulonephritis of the sclerotic class should be thoroughly considered.

Until recently, to diagnose renal relapse of AAV without renal biopsy was very difficult and some renal relapses may have passed unrecognized, resulting in gradual progression to end-stage kidney disease. CD163⁺M2 macrophages infiltrate the kidney at sites of fibrinoid necrosis and seem to be indispensable for the formation of glomerular crescents [28]. Urinary excretion of soluble CD163 is increased in patients with active renal AAV compared with patients in remission [29] and is highly predictive of active renal vasculitis [30]. Prediction of only subtle renal flares was further improved by the combination of urinary CD163 and MCP-1 measurement [31] or the combination of urinary CD163 and urinary CD25 (T cell activation marker) [32] or urinary gremlin (member of the family of bone morphogenetic protein antagonists) [33].

Early non-invasive diagnosis and treatment of renal flares should help to prevent the deterioration of renal function in patients with AAV.

Towards precision medicine in the treatment of AAV by inter- and intra-individual stratification of treatment approaches

Induction versus maintenance treatment

The outcome of untreated granulomatosis with polyangiitis (GPA) is very poor, with a median survival of only 5 months [34]. With the introduction of cyclophosphamide (combined with high-dose corticosteroids), the prognosis dramatically improved and 1-year mortality decreased from 82% to 13% [35]. As a consequence, some patients were treated with continuous oral cyclophosphamide for many years. Once the side effects became obvious, the concept of aggressive, but time-limited induction treatment to suppress disease activity followed by long-term less-aggressive maintenance treatment to prevent relapses of the disease gradually evolved (Table 1).

Severity of the disease

From 1980 to 1990 , all patients (even those with localized disease) were treated with oral cyclophosphamide (and corticosteroids) and the induction treatment lasted usually at least 1 year [35]. With this treatment, active vasculitis was no longer the main cause of death, but was surpassed by infections, cardiovascular disease and malignancy [36].

The concept to stratify patients and therapy according to the severity of the disease (localized, early systemic, generalized and severe renal vasculitis [37]) was developed to balance the relatively mild to life-threatening or (vital) organ-threatening disease and effective, but toxic treatment. It was suggested to use cyclophosphamide only in generalized or severe renal vasculitis and to replace it in patients with early systemic vasculitis with either methotrexate or MMF. Even in patients with generalized vasculitis, the cumulative dose of cyclophosphamide should be minimized either by an early switch (not after 1 year, but after 3 months) from oral cyclophosphamide to azathioprine or by replacing oral continuous cyclophosphamide with cyclophosphamide pulses. Moreover, doses of pulsed cyclophosphamide should be reduced according to the patient’s age and renal function.

On the other hand, in patients with severe renal vasculitis and/or patients with alveolar haemorrhage and dual positivity of ANCA and anti-glomerular basement membrane (GBM) antibodies, plasma exchange was recommended as add-on treatment to cyclophosphamide and high-dose corticosteroids. However, recent data from the Plasma Exchange and Glucocorticoids for Treatment of Anti-Neutrophil Cytoplasm Antibody-Associated Vasculitis (PEXIVAS) study [38] did not confirm the benefit of plasma exchange in a broad population of patients with AAV and plasma exchange should now be only considered in patients with severe renal vasculitis and/or alveolar haemorrhage.

All cyclophosphamide-sparing approaches described above were associated with higher relapse rates during long-term follow-up. Despite that, most centres prefer intravenous pulsed cyclophosphamide to oral continuous cyclophosphamide, which should be limited to patients with difficult access to the infusion centre [39].

The concept of a differentiated approach to patients based on the severity of the disease remains in recent European League Against Rheumatism/ERA-EDTA guidelines [40], although the classification of patients was changed slightly to no organ-threatening disease (formerly localized or early systemic vasculitis), organ- or life-threatening disease (formerly generalized vasculitis) and rapidly progressive renal failure (formerly severe renal vasculitis) and alveolar haemorrhage.

Newly diagnosed versus relapsing disease

Major relapses of AAV were recommended to be treated in the same way as the initial presentation of the disease, i.e. with induction treatment with either oral continuous or intravenous pulsed cyclophosphamide. More recently, rituximab has been shown to be similarly effective as cyclophosphamide in newly diagnosed AAV but significantly better than cyclophosphamide in patients with major (most often pulmonary) relapses [41].

Nonetheless, the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines recommended treating patients with major relapses in a similar way as those with newly diagnosed diseases (i.e. with cyclophosphamide), with rituximab being only an alternative in individuals not tolerating or not responding to cyclophosphamide [42].

The high efficacy of rituximab in inducing remission in patients with relapsing AAV was recently demonstrated in the Rituximab Vasculitis Maintenance Study [43]. According to the draft of the new 2020 KDIGO guidelines, relapses of AAV should be treated with rituximab, and this agent can be also used as a first-line treatment in newly diagnosed patients with AAV and should be especially considered in young patients, women with child-bearing potential, frail elderly patients and in PR3-ANCA patients.

Anti-PR3 versus anti-myeloperoxidase (MPO) disease

Positivity of anti-PR3 antibodies is associated with a significantly higher rate of relapses than anti-MPO positivity. Patients with PR3-ANCA also usually have more active disease and granulomatous lesions and more widespread extrarenal involvement. Induction treatment with rituximab was shown to be more effective in patients with PR3-ANCA compared with those with MPO-ANCA. Because of a higher risk of relapses, maintenance treatment of patients with anti-PR3 disease should be considered for ≥24 months, especially if the patients remain ANCA positive.

In the Rituximab for ANCA-Associated Vasculitis (RAVE) trial, a single course of rituximab without any maintenance therapy was more effective than cyclophosphamide in terms of inducing complete remission and achieving ANCA negativity in patients with PR3-ANCA, but not in MPO-ANCA, and the higher rate of complete remission persisted even after 18 months [44]. Rituximab-treated PR3-ANCA patients also experienced a lower rate of early relapses, suggesting that rituximab should be preferred as the induction treatment even in newly diagnosed PR3-ANCA patients [45].

As maintenance treatment with rituximab was shown in the Maintenance of Remission using Rituximab in Systemic ANCA-associated Vasculitis (MAINRITSAN) study to be more effective than maintenance treatment with azathioprine [46] and relapses are more frequent in patients with anti-PR3 antibodies, rituximab should be the preferred maintenance treatment in this population [47]. In another randomized controlled trial (MAINRITSAN2 [29]), individually tailored rituximab administration (initial dose with reinfusions only with CD19⁺ repletion, or ANCA reappearance or a sharp increase in its titre) was similarly effective as a fixed-dose schedule of 500 mg of rituximab every 6 months, suggesting that in most patients pre-emptive treatment may be safely replaced by more a personalized approach.

Recently published data from the MAINRITSAN3 study clearly demonstrated the superiority of long-term rituximab administration (up to 46 months) compared with placebo (last dose of rituximab at 18 months) [48]. Relapse-free survival at the end of follow-up (56 months from the beginning of the original MAIRITSAN2 study) was 96% in patients treated with long-term rituximab compared with 74% in patients treated with short-term rituximab followed by placebo. As 83% of relapses in the placebo arm occurred in anti-PR3, the authors concluded that prolonged rituximab administration should be considered especially in PR3-ANCA patients. In MAINRITSAN3, prolonged treatment with rituximab resulted in slightly more frequent hypogammaglobulinaemia compared with placebo, but with no clinical consequences in terms of increased infection rate. The risk of hypogammaglobulinaemia should always be considered, especially in patients with previous exposure to cyclophosphamide [48].

Some earlier studies suggested that rituximab is less effective in patients with granulomatous versus vasculitic manifestations of AAV [49]. In fact, orbital masses and pachymeningitis may respond more poorly, but pulmonary involvement responds well to rituximab even in patients who are refractory to other treatment [50]. Although relapses are associated with the type of autoantibody (PR3 versus MPO), mortality and the risk of end-stage kidney disease are also associated with the clinical diagnosis (GPA versus MPA), ear–nose–throat and lung and the severity of kidney involvement. Based on the outcome, patients may be subdivided into several clusters, taking into consideration both ANCA subtype, organ involvement and granulomatous or vasculitic features [51]. Cluster analysis of the large cohort of 673 patients with AAV distinguished five clusters (renal AAV PR3-ANCA, renal AAV without PR3-ANCA, non-renal AAV, cardiovascular AAV and gastrointestinal AAV) with different mortality and relapse rates [52].

How to avoid toxicity of cyclophosphamide and high-dose corticosteroids?

Replacement of cyclophosphamide with rituximab eliminated the increased risk of secondary malignancy in cyclophosphamide-treated patients [53]. However, serious adverse events and infectious complications were not different (e.g. in the RAVE trial in patients treated with either cyclophosphamide or rituximab induction, probably because of high-dose corticosteroids in both arms) [41]. Patients in the PEXIVAS trial randomized to rapid tapering of high-dose corticosteroids had similar efficacy but significantly lower risk of serious infections compared with patients with slower tapering, despite a 40% reduction in the cumulative dose of corticosteroids [38].

Recently, complete replacement of corticosteroids with the complement 5a receptor inhibitor avacopan resulted in a higher rate of sustained remission at Week 52 and a significantly greater increase of estimated glomerular filtration rate [54]. Thus, especially for patients prone to steroid-induced side effects, avacopan could become a valuable treatment option.

A combination of 3-month low-dose cyclophosphamide with two pulses of rituximab and only a very short course of oral corticosteroids (up to 2 weeks [55]) may represent another effective strategy for early withdrawal of corticosteroids to avoid corticosteroid-associated adverse events (e.g. steroid-induced diabetes), with comparable efficacy to historical controls treated with higher doses of cyclophosphamide used in, for example, the CYCLOPS trial [56], and substantially lower exposure to both cyclophosphamide and corticosteroids and should be tested in randomized controlled trials.

Genetic determinants of response to treatment and its safety

Genetic factors may determine the response to treatment of AAV, its durability and safety. Gene variants of glucocorticoid receptor were associated with a higher 5-year mortality risk and higher risk of end-stage renal disease [57]. The single nucleotide polymorphism (519AA) of Fcγ receptor IIA was associated in AAV patients treated with rituximab or cyclophosphamide with a shorter time to complete remission [58]. Risk or relapse in patients with AAV treated with rituximab was associated with the regulatory region single nucleotide polymorphism of B cell activating factor [59]. As cyclophosphamide is a prodrug that requires activation by cytochrome P450 enzymes, the risk of cyclophosphamide-induced leukopaenia was related in patients with AAV to the CYP450 2C9 genotype [60].

Personalized treatment of AAV on the horizon

The clinical presentation of AAV is heterogeneous and outcome is very variable based on age, organ involvement, type of ANCA and renal function. Initially all patients were treated with similar lifesaving but toxic long-term treatment with cyclophosphamide and slowly tapered high-dose corticosteroids. A reduction of the cumulative dose of cyclophosphamide and its replacement with rituximab and a reduction of the cumulative dose of corticosteroids in the induction phase of therapy or even putative replacement of corticosteroids with avacopan should result in at least similarly effective but much better-tolerated treatment with lower acute and chronic toxicity. This new (but also more expensive) treatment should be preferred especially in young patients to avoid gonadotoxicity and preserve fertility, in frail elderly patients, in relapsing patients and in PR3-ANCA-positive patients. In the future it is hoped that genetic and biomarker data, along with clinical and histopathologic data, might be incorporated into treatment decisions, thus paving the way to a new era of precision medicine in the treatment of AAV.

FUNDING

This article is part of a supplement supported by a sponsorship from Amicus Therapeutics UK, a research grant from Boehringer Ingelheim RCV GmbH & Co KG, an educational sponsorship agreement from Astellas Pharma and a restricted research grant from Vifor Pharma Österreich GmbH. This supplement is part of the project DC-ren that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement 848011.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES