Article Text
Abstract
The current treatment paradigm in lupus nephritis consists of an initial phase aimed at inducing remission and a subsequent remission maintenance phase. With this so-called sequential treatment approach, complete renal response is achieved in a disappointing proportion of 20–30% of the patients within 6–12 months, and 5–20% develop end-stage kidney disease within 10 years. Treat-to-target approaches are detained owing to uncertainty as to whether the target should be determined based on clinical, histopathological, or immunopathological features. Until reliable non-invasive biomarkers exist, tissue-based evaluation remains the gold standard, necessitating repeat kidney biopsies for treatment evaluation and therapeutic decision-making. In this viewpoint, we discuss the pros and cons of voclosporin and belimumab as add-on agents to standard therapy, the first drugs to be licenced for lupus nephritis after recent successful randomised phase III clinical trials. We also discuss the prospect of obinutuzumab and anifrolumab, also on top of standard immunosuppression, currently tested in phase III trials after initial auspicious signals. Undoubtably, the treatment landscape in lupus nephritis is changing, with combination treatment regimens challenging the sequential concept. Meanwhile, the enrichment of the treatment armamentarium shifts the need from lack of therapies to the challenge of how to select the right treatment for the right patient. This has to be addressed in biomarker surveys along with tissue-level mapping of inflammatory phenotypes, which will ultimately lead to person-centred therapeutic approaches. After many years of trial failures, we may now anticipate a heartening future for patients with lupus nephritis.
- lupus nephritis
- lupus erythematosus
- systemic
- biological therapy
- therapeutics
- B-lymphocytes
Statistics from Altmetric.com
Introduction
In different cohorts, 35%–60% of patients with systemic lupus erythematosus (SLE) develop lupus nephritis (LN) during their disease course.1 2 Once established, LN should always be considered a severe condition. This is because each LN flare results in irreversible nephron loss corresponding to shortening of the kidney lifespan by decades. Persistently active LN increases the rate of nephron loss, leading to earlier onset of chronic kidney disease (CKD) and end-stage kidney disease (ESKD).1 Therefore, flares and residual disease activity are essential predictors of long-term renal function impairment, and the management of LN resembles a race against nephron loss.
The sequential treatment paradigm
The current treatment paradigm is so-called ‘sequential’, comprising an initial phase aimed at inducing disease control and a subsequent phase aimed at maintaining the achieved response. First-line options for the initial or induction phase include oral mycophenolate mofetil (MMF; most commonly 2–3 g/day), or equivalent doses of mycophenolate sodium, and low-dose intravenous cyclophosphamide (CYC) 500 mg every second week for a total of six pulses; high-dose intravenous CYC (given monthly for 6 months, as per the NIH protocol) is still considered another option, however, with potentially more side effects, especially regarding fertility. Alongside, methylprednisolone at a total dose of 0.5–3 g is given in most cases, followed by oral prednisone at a dose of 0.3–0.5 mg/kg/day, tapered to ≤7.5 mg/day within 3–6 months. The subsequent or maintenance phase consists of either MMF or azathioprine (AZA). Calcineurin inhibitors (CNI) are used as an add-on option to optimise therapy, mainly in membranous LN, and are given at the lowest possible therapeutic dose, which necessitates measurement of drug concentrations. Hydroxychloroquine is recommended for all patients unless a contraindication exists.3 In refractory cases, the treatment should be intensified, which may include therapy switch, addition of a CNI, or rituximab as an off-label option. Importantly, response to initial therapy should not lead to therapy discontinuation; it is recommended to continue for at least 3 years from the flare onset.3
With the current sequential treatment strategies, complete renal response (CRR) is achieved in a disappointing proportion of 20%–30% of the patients within 6–12 months from the onset of LN,4 5 and 20%–35% of the patients who achieve adequate disease control show relapses within 3–5 years.1 6 At least 20% of LN patients develop CKD and 5%–20% reach ESKD within 10 years from the LN onset,7–9 not to mention the side effects of current immunosuppressants which contribute to increased morbidity and reduced quality of life.1 On the whole, it is apparent that improvement of LN therapy and prognosis remains on the agenda. To address this need, the LN research community has focused on two main trajectories, that is, implementation of a treat-to-target approach and a paradigm shift from sequential to combination and personalised therapy.
Treat-to-target approach in the management of LN
With regard to the former trajectory, a key question is whether the target should be determined based on clinical, histopathological or immunopathological features. The Euro-Lupus8 and MAINTAIN10 Nephritis Trials provided robust evidence that early decrease in proteinuria predicts good long-term renal outcome, and data from three independent study populations indicated a cut-off of 0.7–0.8 g/day at 1 year to be the best predictor of creatinine values ≤1 mg/dL 7 years after the LN onset.11–13 However, while the positive predictive value of this target was excellent, the negative predictive value was poor,11 12 highlighting the need for predictors of poor long-term prognosis. In other words, the remaining challenge is to determine patients at high risk for renal function worsening irrespective of clinical response and might, therefore, benefit from early adjustments in therapy.
The concept of a histopathological target emerged from observations that clinical outcome based on proteinuria and/or urinalysis and histopathological outcome based on repeat kidney biopsies are discordant.14–16 We recently showed that National Institutes of Health (NIH) activity and chronicity index scores >3 in per-protocol repeat kidney biopsies were associated with subsequent renal relapse and renal function impairment, respectively, while the baseline biopsy exhibited no predictive value.16 Notably, active lesions in glomeruli mainly accounted for the association with renal relapses, whereas chronic damage in the tubulointerstitial compartment was an important contributor to the association with poor long-term renal function,16 observations of particular importance in light of evidence that, although underemphasised in current LN classification,17 the tubulointerstitial compartment is important in the inflammatory process in LN18 19 and in renal prognosis.20–22 Knowledge of these associations might help avert the poor renal prognosis in patients at risk by adjusting the treatment, addressed in the imminent multicentre randomised study entitled ‘Per-protocol repeat kidney biopsy in incident cases of LN’, abbreviated ReBioLup (http://rebiolup.com).23 Systematic tissue-based survey within ReBioLup or similar endeavours may be foreseen to facilitate drug development through microarchitecture-level and single cell-level understanding of the role of the different compartments of the inflamed kidney towards more effective resolution of the inflammation and deceleration, or, desirably, reverse of fibrosis.
Lastly, recent implications that the degree of resorption of immune deposits in electron microscopy after induction therapy are associated with the clinical response to therapy24 warrant further survey in relation to long-term prognosis, and support the notion that immunopathological targets may provide additive prognostic value to activity and chronicity scores.
From sequential to combination therapy
Results of two recent randomised phase III clinical trials portend a new era in the management of LN, with combination regimens gradually substituting the sequential treatment paradigm. In fact, these trials led to labelling voclosporin and belimumab for LN.
Voclosporin
The potential of add-on CNIs emerged from multiple studies, particularly from Asia, examining the effect of tacrolimus as a monotherapy or combined with MMF.25 The effect of a more modern CNI, that is, voclosporin, an analogue of ciclosporin with enhanced action against calcineurin, greater metabolic stability and a quicker elimination of metabolites minimising the need for monitoring, was recently studied in the phase II Aurinia Urinary Protein Reduction Active - Lupus With Voclosporin (AURA-LV) and phase III Aurinia Renal Response in Active Lupus With Voclosporin (AURORA 1) clinical trials of LN. AURA-LV comprised 265 class III–V LN patients and demonstrated higher proportions of CRR at week 48 among patients treated with low-dose (23.7 mg two times per day) or high-dose (39.5 mg two times per day) voclosporin (49% or 40%, respectively) versus placebo (24%) in combination with MMF (2 g/day) and glucocorticoids.26 The difference at week 24 was statistically significant only for the low-dose voclosporin group (33% vs 19%). More serious adverse events were reported in the voclosporin groups, and a higher number of deaths in the low-dose group compared with the placebo and high-dose voclosporin groups (11%, 1% and 2%, respectively).26 AURORA 1 comprised 357 patients with class III–V LN exposed to voclosporin (23.7 mg two times per day) or placebo (1:1) on top of MMF and low-dose oral glucocorticoids, commencing at a prednisone dose of 20–25 mg/day with a subsequent forced quick tapering schedule. This combination regimen was given for 52 weeks, with 41% of the patients in the voclosporin group achieving CRR (including urinary protein to creatinine ratio (uPCR)) ≤0.5 and estimated glomerular filtration rate (eGFR) >60 mL/min/1.73 m2), compared with 23% in the placebo group. The separation was also significant at week 24 and remained significant after stratification into patients of Hispanic and non-Hispanic origin but did not reach significance in the subgroup of white patients (36% of the study population). Moreover, the separation between voclosporin and placebo was significant only in patients who were on MMF at baseline (55%), and not in patients who commenced at MMF at the study baseline (45%).27 In AURORA 1, the adverse event profile was balanced between the voclosporin and placebo groups.27
Importantly, the voclosporin trials demonstrated that a low-dose glucocorticoid regimen is feasible, since the forced tapering schedule requiring that the prednisone dose was decreased to 2.5 mg/day by week 16 was achieved in ≥75% of the patients in each arm, in both trials.26 27 Questions that remain to be answered include the long-term toxicity of voclosporin, its efficacy regarding prevention of renal flares and extrarenal activity, and the optimal time of withdrawal. As a matter of fact, some scepticism may be raised about proteinuria reduction at least partially being explained by the stabilisation of the podocyte cytoskeleton, thus overestimating the effect of voclosporin on inflammatory renal activity, and the potential tissue-level drug toxicity. Regarding the latter, repeat biopsy data would be desirable to appositely address this concern. Finally, while nephrologists have experience from CNI use and may therefore be expected to embrace a modern CNI with ease, the threshold for rheumatologists is anticipated to be higher.
Belimumab
B cell hyperactivity has a key role in SLE pathogenesis, and blockade of B cell activating factor with belimumab has proven successful in several phase III trials of SLE,28 which however had excluded patients with severe active biopsy-proven LN. However, a post hoc analysis revealed significantly greater proteinuria reductions in belimumab-treated versus placebo-treated patients,29 constituting the main rationale for the LN-specific phase III Belimumab International Study in Lupus Nephritis (BLISS-LN) trial.30 BLISS-LN assigned 448 patients with biopsy-proven class III–V LN at a ratio 1:1 to intravenous belimumab or placebo as add-on to oral glucocorticoids and either oral MMF or intravenous CYC followed by AZA for a total of 104 weeks. The primary endpoint was the so-called primary efficacy renal response (PERR) at week 104, a composite measure requiring uPCR ≤0.7 and eGFR no worse than 20% below the preflare value or ≥60 mL/min/1.73 m2. PERR at week 104 was achieved in 43% of patients who received belimumab and 32% of patients who received placebo, yielding a statistically significant separation of 11%, that was also significant at week 52 (47% vs 35%). Consistent separations were seen regarding CRR, with a more stringent uPCR cut-off set at <0.5 (30% vs 20% at week 104), and no safety signals emerged.30
While these results may indeed be considered auspicious, worth noting observations from subgroup analyses include that a benefit from belimumab was documented for patients who received belimumab on top of MMF but not for patients who received it on top of CYC/AZA, and in non-black patients but not in the black patient population.30 Nevertheless, patients in the CYC/AZA group had a more severe disease profile, evidenced by higher levels of proteinuria, lower eGFR, lower complement levels and longer disease duration at baseline. Moreover, African-Americans are known to have a more severe disease course.1 Among the pros, BLISS-LN is the largest to date clinical trial of LN. Some scepticism is raised about the magnitude of response not exceeding 30% for CRR, along with some recently reported de novo LN cases during belimumab therapy.31 This said, since belimumab already is an established biological for SLE, the threshold for clinicians to also use it in active LN may be expected to be low.
Other promising add-on agents: obinutuzumab and anifrolumab
B cell depletion is not a new concept for LN. Following encouraging observations from off-label use of the chimeric anti-CD20 monoclonal antibody rituximab, the Lupus Nephritis Assessment with Rituximab (LUNAR) trial tested its efficacy on top of MMF and high-dose glucocorticoids but failed to demonstrate superiority over placebo.32 Despite the failure of LUNAR and no further investment from the industry, its off-label use in refractory cases is now even recommended.3 Moreover, the combination of belimumab and rituximab is currently trialled in investigator-initiated settings (SYNBIoSe, NCT0228498433; SynBioSe-2, NCT03747159). Among more modern B cell depleting agents, obinutuzumab is a fully humanised anti-CD20 monoclonal antibody that has a partially dissimilar to rituximab binding specificity on CD20, resulting in greater and more sustained depletion. It was recently tested for LN in the phase II NOBILITY trial, that randomised 126 patients with class III/IV±V to receive two pulses of intravenous obinutuzumab or placebo on top of MMF and glucocorticoids, and two more pulses after 6 months. CRR requiring uPCR <0.5 yielded a separation of 12% between the obinutuzumab and placebo group at week 52 in favour of obinutuzumab (35% vs 23%) that became greater and statistically significant at week 104 (41% vs 23%).34 While NOBILITY was a rather small phase II trial comprising a vast majority of Hispanic participants, the favourable results along with the absence of safety signals warranted the phase III REGENCY trial (NCT04221477) that is currently recruiting patients from a more diverse population.
Anifrolumab is a fully human monoclonal antibody against the type I interferon receptor that has been tested with encouraging results in two phase III clinical trials of non-renal SLE, the Treatment of Uncontrolled Lupus via the Interferon Pathway (TULIP)-135 and TULIP-236 trials. The phase II TULIP-LN trial randomised 145 patients with biopsy-proven LN to anifrolumab basic regimen (BR; 300 mg), anifrolumab intensified regimen (IR; 900 mg for three doses, 300 mg thereafter) or placebo every fourth week, on top of MMF and glucocorticoids.37 The trial did not meet the primary endpoint, but IR anifrolumab was numerically superior to placebo across multiple clinical outcomes, including CRR (requiring uPCR ≤0.7) at week 52 (46% vs 31%) and steroid sparing effects. An important observation was that anifrolumab clearance was higher in patients with LN compared with previous experience from the non-renal TULIP trials owing to proteinuria, resulting in suboptimal serum concentrations in the BR group. Apart from the expected higher incidence of herpes zoster in anifrolumab-exposed patients, adverse events were non-serious and similar across groups. Collectively, despite the failure to meet the primary endpoint, the CRR frequency of 46% with the IR justifies the imminent phase III trial.
Lastly, several other pharmaceuticals are currently tested, for example, add-on interleukin-17A inhibition with secukinumab in the phase III SELUNE trial (NCT04181762).
How to choose? Towards precision medicine
Undoubtedly, the treatment landscape within LN is changing. Until recently, there was no drug approved for LN, with first-line therapy relying on off-label use of mainly MMF and CYC. We are currently witnessing a rapid enrichment of the therapeutic armamentarium with new-generation drugs, shifting the challenge from the need for new treatments to the need to determine how to use them wisely. Indeed, how will clinicians choose among add-on voclosporin, belimumab, obinutuzumab, anifrolumab and future agents?
Initially, drug availability, costs, pregnancy issues and the clinician’s experience and gut feeling will steer the choice of therapy. Thus, patients with extrarenal manifestations such as arthritis or skin disease will likely be given a B cell targeting add-on agent rather than a CNI. Hopefully, biomarker mining will soon shed more light on which patient subgroups are expected to respond better to each one of these new regimens, based on tissue-based approaches,38 39 urinary markers40 and molecular/cellular signatures. Borrowing examples from rheumatoid arthritis, for example, the recently suggested RNA sequencing-based treatment selection,41 application of such technologies in tissue-level investigations can be anticipated to entail a vertical take-off towards precision medicine in LN in the future. Moreover, associations between transcriptional signatures in renal tissue and urine samples may engender the prospect of the readily available and non-invasive ‘liquid biopsy’ in treatment selection and evaluation, as recently contemplated.39
Concluding remarks
Although direct comparison across trials should be avoided to reckon definition dissimilarities, CRR attainment was ~30% in the placebo arms, illustrating that the current treatment paradigm does not meet the expectations. However, even in the active agent arms, CRR did not exceed—at best—46% which abates the enthusiasm and points to remaining need for improvement also after the introduction of those new therapeutic modalities. This probably addresses the concern about the risk for overtreatment in the abundance of new options. In fact, how many patients do we need to treat differently to avoid one case of ESKD with current sequential and how many with future combination regimens? Importantly, however, even if the benefit from addition of the newly approved agents voclosporin and belimumab to standard therapy may be considered moderate, preserving renal function for an additional small number of patients is an achievement in the right direction.
Also, should add-on agents be part of the initial phase of treatment or be initiated after a first period of inadequate response? Frankly, are we able to address this question when treatment evaluation relies on an indirect outcome, that is, proteinuria, especially when agents with particular antiproteinuric potency come into play? Until reliable non-invasive biomarkers exist, tissue-level approaches and repeat biopsies will be essential towards treatment optimisation through precision medicine, with the ultimate goal being nephron loss deceleration and preservation of the renal function. For the moment, the research agenda is long, and includes biomarkers in ‘liquid biopsy’ reflecting kidney histopathology, exploration of the value of repeat kidney biopsies for treatment finetuning (for example, in ReBioLup) or withdrawal, and early markers of long-term prognosis. In light of the recent drug approvals, it also includes survey on biomarkers for person-centred treatment selection, optimal time for add-on agent and optimal treatment duration. It is important to bear in mind that trial settings never resemble reality; in fact, both AURORA 1 and BLISS-LN excluded patients with severely impaired renal function and failed to show efficacy of add-on voclosporin in MMF naïve patients or add-on belimumab in black patients in subgroup analyses. Moreover, long-term data on renal relapses are awaited. Nevertheless, the recent trial successes may be an omen of a paradigm shift from sequential to combination regimens for the treatment of LN. In the fullness of time, we foresee a heartening future for patients living with LN.
Ethics statements
Patient consent for publication
References
Footnotes
Handling editor Josef S Smolen
Contributors IP and FAH conceived of and drafted the manuscript, revised it critically for important intellectual content, and approved the final version prior to submission.
Funding IP is supported by the Swedish Rheumatism Association (R-932236), King Gustaf V’s 80-year Foundation (FAI-2019-0635), Professor Nanna Svartz Foundation (2019-00290), Ulla and Roland Gustafsson Foundation (2019-12), Region Stockholm and Karolinska Institutet. FAH is supported by Fondation Saint-Luc and Fonds National de la Recherche Scientifique.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.