Sirolimus versus mycophenolate mofetil for the treatment of lupus nephritis: Results from a real-world CSTAR cohort study

Introduction

Systemic lupus erythematosus (SLE) is a complex autoimmune disease characterized by variable manifestations affecting multiple organ systems, often leading to organ damage and increased mortality.^[1,2] Lupus nephritis (LN) occurs in approximately 40% of SLE patients ^[3,4] and is associated with a heightened risk of end-stage renal disease (ESRD) and elevated mortality rates.^[2,3,5,6]

While biological agents such as belimumab ^[7,8] and rituximab^[9] have shown promise in the treatment of SLE and LN, conventional immunosuppressive therapies, such as cyclophosphamide (CYC), mycophenolate (MMF), methotrexate (MTX), azathioprine (AZA), and calcineurin inhibitors (CNIs) remain the standard of care (SoC) for managing SLE and LN.^[4,10,11] Furthermore, the treat-to-target (T2T) strategy has been recommended for SLE in recent years.^[12] Achieving remission or a state of lupus low disease activity state (LLDAS) with effective immunosuppressants or biologics has emerged as a key treatment goal and a predictor of outcomes in SLE.^{[13, 14, 15]} Regarding LN, the treatment target focuses on clinical response, particularly primary efficacy renal response, which has been linked to improved prognosis and is widely used as an endpoint in recent clinical trials.^[8] Given the severity of LN and the adverse effects associated with current medications, there is a pressing need for novel treatment options and further evidence supporting the T2T approach to improve the prognosis of LN.

Sirolimus, an inhibitor of the mammalian target of rapamycin (mTOR), has been implicated in the pathogenesis of SLE.^[16] Emerging studies have demonstrated the efficacy and safety of sirolimus as a novel immunosuppressant for SLE and LN.^{[17, 18, 19]} However, no comparative studies have directly evaluated sirolimus against the SoC for LN patients. Leveraging the Chinese SLE Treatment and Research (CSTAR) group, the largest registry of SLE in China,^[20] we conducted a prospective real-world cohort study to compare the effectiveness and safety of sirolimus versus MMF in the treatment of LN.

Methods

Patients

As of March 2022, the CSTAR registry included 338 rheumatology centers across 31 provinces in China. Based on the CSTAR registry, this prospective real-world study enrolled LN patients who were prescribed either sirolimus or MMF between March 2012 and March 2022. This study was approved by the Medical Ethics Committee of Peking Union Medical College Hospital (PUMCH)(approval number: S-478), and written informed consent was obtained from all recruited patients. Patients were diagnosed with LN based on their fulfilment of the corresponding items for lupus nephritis from the SLE classification criteria established by American College of Rheumatology (ACR) in 1997,^[21] the Systemic Lupus International Collaborating Clinics (SLICC)/ACR in 2012,^[22] or the European League Against Rheumatism/ACR in 2019.^[1] Within the CSTAR cohort, 108 LN patients were treated with sirolimus, and 1124 LN patients were treated with MMF, all of whom had completed baseline and 3- or 6-month follow-up data. Inclusion criteria required patients to have active lupus status and to have been switched to or initiated on sirolimus or MMF alone as immunosuppressant at baseline. The dosage of background immunosuppressants in patients who have been initiated on sirolimus or MMF at baseline remained stable for at least one month before baseline. Patients receiving combined immunosuppressants, including CYC, MTX, AZA, tacrolimus (TAC) or cyclosporin A, in addition to sirolimus or MMF at baseline, or those who had already achieved LLDAS/remission at baseline, were excluded. The screening flowchart is illustrated in Figure 1. After applying the inclusion and exclusion criteria, a total of 63 patients treated with sirolimus and 622 patients treated with MMF were included in this study.

Figure 1

The screening flow chart. LLDAS: lupus low disease activity state.

Results

Effectiveness of Sirolimus Versus MMF

Data on the effectiveness of sirolimus versus MMF at 3, 6, and 12 months were presented in Table 2 and Figure 2. Both sirolimus and MMF demonstrated good efficacy in treating LN. The proportions of patients achieving complete or partial renal remission was similar between the two groups (65.9% vs. 68.4% at 3 months, 56.3% vs. 72.4% at 6 months, and 76.9% vs. 85.7% at 12 months, P = 0.809, 0.189, and 0.557, respectively), as well as the prevalence of hematuria (27.3% vs. 31.6% at 3 months, 31.3% vs. 13.8% at 6 months, and 15.4% vs. 7.1% at 12 months, P = 0.669, 0.105, and 0.496, respectively), and the change in 24hUP level (-0.960 (-2.083, 0.115) vs. -0.287 (-0.900, -0.002) at 3 months, -0.220 (-1.540, 0.270) vs. -0.669 (-2.085, -0.475) at 6 months, and-1.470 (-4.410, 0.150) vs. -0.675 (-1.433, 1.750) at 12 months, P = 0.284, 0.121, and 0.440, respectively).

Figure 2

Clinical effectiveness of sirolimus versus MMF in treating lupus nephritis (LN) at 3-, 6-, and 12-month. MMF: mycophenolate mofetil, LLDAS: lupus low disease activity state, GC: glucocorticoid, CR: complete remission, PR: partial remission.*P < 0.05; **P < 0.01; ***P < 0.001.

Table 2

Efficacy of sirolimus vs. mycophenolate mofetil (MMF) in treating patients with lupus nephritis at 3-, 6-, and 12-month

	3 months			6 months			12 months
	Sirolimus	MMF	P value	Sirolimus	MMF	P value	Sirolimus	MMF	P value
SLEDAI-2K change	-2.00 (-6.00, 0.00)	-4.00 (-8.00, 0.00)	0.228	-2.00 (-6.00, 0.00)	-4.00 (-8.00, -1.00)	0.141	-2.00 (-6.00,-0.50)	-4.00 (-6.00, -1.50)	0.532
	-0.200	-0.250		-0.200	-0.200		-0.400	-0.350
PhGA change	(-0.600, 0.000)	(-0.650, 0.200)	0.928	(-0.800, 0.100)	(-0.750, 0.200)	0.743	(-0.650,-0.025)	(-0.775, 0.150)	0.908
	47.4%	51.4%		50.0%	64.0%
Clinical response	(18/38)	(18/35)	0.729	(13/26)	(16/25)	0.313	50.0% (5/10)	66.7% (8/12)	0.429
LLDAS/remission	15.9% (7/44)	10.5% (4/38)	0.476	18.8% (6/32)	13.8% (4/29)	0.602	61.5% (8/13)	50.0% (7/14)	0.547
Lupus nephritis
Remission (CR+PR)	65.9%	68.4%	0.809	56.3%	72.4%	0.189	76.9%	85.7%	0.557
	(21+8/44)	(24+2/38)		(17+1/32)	(20+1/29)		(8+2/13)	(12+0/14)
	-0.960	-0.287		-0.220	-0.669		-1.470	-0.675
24hUP level change (g)	(-2.083, 0.115)	(-0.900, -0.002)	0.284	(-1.540, 0.270)	(-2.085, -0.475)	0.121	(-4.410, 0.150)	(-1.433, 1.750)	0.440
	27.3%	31.6%		31.3%	13.8%
Hematuria	(12/44)	(12/38)	0.669	(10/32)	(4/29)	0.105	15.4% (2/13)	7.1% (1/14)	0.496
Laboratory parameters
Normalised anti-dsDNA	41.7%	66.7%	0.082	28.6%	60.0%	0.059	50.0% (3/6)	42.9% (3/7)	0.797
	(10/24)	(16/24)		(6/21)	(9/15)
Recovered hypocomplementaemia	74.1%	41.7%	0.019	75.0%	30.0%	0.003	71.4% (5/7)	0% (0/8)	0.003
	(20/27)	(10/24)		(18/24)	(6/20)
	0.206 (0.084, 0.321)	0.075		0.219	0.001		0.259	-0.007
Change in C3 (g/L)		(-0.063, 0.196)	0.008	(0.062, 0.378)	(-0.072, 0.130)	0.001	(-0.003, 0.570)	(-0.074, 0.125)	0.063
		0.014		0.054	-0.001		0.020	-0.003
Change in C4 (g/L)	0.033 (0.009, 0.071)	(-0.001, 0.060)	0.250	(0.014, 0.103)	(-0.038, 0.060)	0.011	(-0.008, 0.101)	(-0.028, 0.026)	0.121
Use of GC
							-5.00	-6.25
Change in GC dosage (mg/d)	0.00 (-13.75, 0.00)	0.00 (-5.00, 0.00)	0.174	0.00 (-10.63, 0.00)	0.00 (-6.25, 0.00)	0.771	(-22.50, 0.00)	(-12.50,-1.88)	0.961
GC dosage≤7.5 mg/d prednisone	31.8%	34.2%		46.9%	37.9%		84.6%
	(14/44)	(13/38)	0.818	(15/32)	(11/29)	0.481	(11/13)	64.3% (9/14)	0.228

1. SLEDAI-2K: Systemic Lupus Erythematosus Disease Activity Index 2000, PhGA: physician’s global assessment, LLDAS: lupus low disease activity state, CR: complete remission, PR: partial remission, 24hUP: 24-hour urine protein, anti-dsDNA: anti-double-stranded DNA, GC: glucocorticoid.

Improvement in SLE disease activity were comparable between the two groups. Changes in SLEDAI-2K score (-2.00 (-6.00, 0.00) vs. -4.00 (-8.00, 0.00) at 3 months, -2.00 (-6.00, 0.00) vs. -4.00 (-8.00, -1.00) at 6 months, and-2.00 (-6.00, -0.50) vs. -4.00 (-6.00, -1.50) at 12 months, P = 0.228, 0.141, and 0.532, respectively) and PhGA score (-0.200 (-0.600, 0.000) vs. -0.250 (-0.650, 0.200) at 3 months, -0.200 (-0.800, 0.100) vs. -0.200 (-0.750, 0.200) at 6 months, and-0.400 (-0.650, -0.025) vs. -0.350 (-0.775, 0.150) at 12 months, P= 0.928, 0.743, and 0.908, respectively) showed no significant differences. Additionally, the proportions of patients achieving clinical response (47.4% vs. 51.4% at 3 months, 50.0% vs. 64.0% at 6 months, and 50.0% vs. 66.7% at 12 months, P = 0.729, 0.313, and 0.429, respectively), LLDAS or remission (15.9% vs. 10.5% at 3 months, 18.8% vs. 13.8% at 6 months, and 61.5% vs. 50.0% at 12 months, P = 0.476, 0.602, and 0.547, respectively) were similar between the two groups.

Regarding serological activity, the sirolimus group demonstrated significantly greater improvement in hypocomplementemia compared to MMF group, with higher rates of normalized complement levels at all follow-up intervals (74.1% vs. 41.7% at 3 months, 75.0% vs. 30.0% at 6 months, and 71.4% vs. 0% at 12 months, P = 0.019, 0.003 and 0.003, respectively). Significant differences were also observed in change of serum C3 level (0.206 (0.084, 0.321) vs. 0.075 (-0.063, 0.196) at 3 months, 0.219 (0.062, 0.378) vs. 0.001 (-0.072, 0.130) at 6 months, and 0.259 (-0.003, 0.570) vs. -0.007 (-0.074, 0.125) at 12 months, P = 0.008, 0.001, and 0.063, respectively) and C4 level (0.033 (0.009, 0.071) vs. 0.014 (-0.001, 0.060) at 3 months, 0.054 (0.014, 0.103) vs. -0.001 (-0.038, 0.060) at 6 months, and 0.020 (-0.008, 0.101) vs. -0.003 (-0.028, 0.026) at 12 months, P = 0.250, 0.011, and 0.121, respectively). No significant differences were observed in the proportion of patients achieving normalized anti-dsDNA antibody between the two groups (41.7% vs. 66.7% at 3 months, 28.6% vs. 60.0% at 6 months, and 50.0% vs. 42.9% at 12 months, P = 0.082, 0.059, and 0.797, respectively).

The change in glucocorticoid (GC) dosage [0.00 (-13.75, 0.00) vs. 0.00 (-5.00, 0.00) at 3 months, 0.00 (-10.63, 0.00) vs. 0.00 (-6.25, 0.00) at 6 months, and -5.00 (-22.50, 0.00) vs. -6.25 (-12.50, -1.88) at 12 months, mg/d, P = 0.174, 0.771, and 0.961, respectively] and the proportions of patients achieving prednisone doses ≤7.5 mg/day (31.8% vs. 34.2% at 3 months, 46.9% vs. 37.9% at 6 months, and 84.6% vs. 64.3% at 12 months, P = 0.818, 0.481 and 0.228, respectively) were not significantly different between the two groups.

During the follow-up period, none of the 106 patients in either group switched therapies or required additional immunosuppressants or biologic agents such as belimumab or rituximab.

Discussion

This real-world, prospective, open-label cohort study compared the effectiveness and safety of sirolimus with MMF for the treatment of LN. To the best of our knowledge, this is the first prospective study to compare sirolimus against SoC therapies and provide treat-to-target (T2T) evidence in LN management. Sirolimus demonstrated similar clinical effectiveness and steroid-sparing effect to MMF, and better effectiveness in serological improvement. In comparison to the MMF group, the sirolimus group exhibited no significant differences in rates of LN remission (65.9% vs. 68.4%, 56.3% vs. 72.4%, and 76.9% vs. 85.7% at three follow-up visits) and LLDAS/remission (15.9% vs. 10.9%, 18.8% vs. 13.8%, and 61.5% vs. 50.0% at three follow-up visits). Significantly greater improvements in complement levels were observed in sirolimus group, especially at 6-month follow-up (75% vs. 30%, P < 0.01). These outcomes are consistent with prior CSTAR studies^[24,25] and other cohorts,^[17,18] reinforcing sirolimus as a well-tolerated and effective option for LN.

Sirolimus, also named as rapamycin, is a mTOR inhibitor which has been successfully used for the prevention of graft-versus-host disease. Activation of mTOR signaling has been observed in SLE patients,^[16] and this pathway may contribute to various pathogenesis mechanisms of SLE, including induction of interleukin (IL)-4 production and necrotic death of double-negative T cells,^[26] increasing interferon α (IFNα) production,^[27] promotion of BAFF-stimulated cell proliferation and survival,^[28] promotion of Th17 cell differentiation,^[29] and inhibition of CD4⁺ T cell differentiation into Treg cells.^[30]

The efficacy of sirolimus in treating autoimmune diseases and SLE has been reported in several studies.^[24,25,31] A multicenter, prospective clinical trial involving 30 patients demonstrated the effectiveness of sirolimus in relapsed/refractory autoimmune cytopenia.^[32] A single-center, single-arm, phase II study conducted with 20 patients from the CSTAR registry highlighted the efficacy and safety of sirolimus in treating refractory connective tissue disease-related thrombocytopenia. ^[33] A real-world study enrolled 49 SLE patients from CSTAR registry found that sirolimus significantly improved skin rash, arthritis, and thrombocytopenia.^[25] Eriksson et al. had reported that sirolimus was effective in SLE patients with musculoskeletal manifestations, such as arthritis and tendinitis, in a retrospective Swedish cohort involving 27 SLE patients with long-term follow-up (mean time: 47.1 months).^[34] Despite these promising findings, the role of sirolimus in LN still warrants further investigation.

Regarding LN, animal experiments and in vitro evidence have demonstrated that mTOR inhibition by sirolimus can ameliorate the manifestation of LN.^{[35, 36, 37]} In LN patients, renal mTOR complex 1 activation has been proposed as a biomarker for disease activity and clinical prognosis.^[38] The mTOR activation might affect the pathogenesis of LN through various mechanisms, including the activation of Th17 cells and IL-17 signaling pathway,^[37,39] activation of Th1 cells and suppression of dendritic cells (DC),^[35] promotion of intra-renal expression of monocyte chemoattractant protein-1 (MCP-1),^[36] and the dysregulation of autophagy not only in T and B cells,^[40] but also in renal resident cells such as podocytes.^[41]

Despite these mechanistic insights, clinical evidence supporting sirolimus use in LN remains limited. Case reports indicated that sirolimus may be effective in patients with LN refractory to conventional immunosuppressants, such as CYC, MMF, tacrolimus or cyclosporin A.^[42,43] Retrospective studies have also highlighted its potential in LN. A study of 16 LN patients treated with sirolimus for an average duration of 45.3 months^[18] reported significant improvement in proteinuria, with only one patient experiencing renal flare. Another cohort of 12 LN patients revealed the improvements in proteinuria and hematuria following long-term sirolimus treatment.^[17] Previous CSTAR studies involving SLE patients with various manifestations have also suggested the potential effectiveness of sirolimus in achieving LN remission.^[24,25] However, no prospective controlled studies have specifically evaluated the effectiveness of sirolimus in LN. The first CSTAR study comparing sirolimus with conventional immunosuppressants in SLE found sirolimus to be as effective as tacrolimus.^[24] Consistent with these findings, our study provides further evidence by comparing sirolimus with SoCs treatments in LN.

Our study demonstrated that sirolimus exhibits comparable efficacy to MMF, the SoCs for LN according to guideline.^[4] No significant differences were observed between the two groups in terms of the primary endpoint and main clinical secondary endpoints related to SLE and LN activity. In recent years, the implementation of T2T strategy for SLE and LN, which was proven in improving clinical outcomes, has become a clinical reality.^[44] To provide robust T2T evidence, our study specifically excluded patients who had already achieved LLDAS or remission at baseline. The results were encouraging, as sirolimus and MMF showed no significant differences in clinical response, LLDAS/remission rates, or renal remission. The equivalence in efficacy, particularly for T2T endpoints not previously reported, positions sirolimus as a promising novel treatment option for LN. The 2023 updated European Alliance of Associations for Rheumatology (EULAR) recommendation for SLE highlighted the combination therapy with belimumab (either with CYC or MMF) or calcineurin inhibitors (especially voclosporin or TAC, combined with MMF), in the initial therapy of active LN.^[45] Notably, in our study, 16 patients received sirolimus as an add-on therapy with background immunosuppressants (primarily MMF), suggesting that sirolimus may also serve as a novel treatment option for combination therapy in LN.

Sirolimus has demonstrated good effect on the improvement of serological parameters in SLE and LN. Previous studies from the CSTAR registry and other cohorts have shown that sirolimus significantly improves hypocomplementemia in SLE and LN patients.^[17,18,25] Consistent with these findings, our study also observed notable improvements in hypocomplementemia. The CSTAR cohort ^[25] and US-based SLE cohort^[17] reported significant reductions in anti-dsDNA titers following sirolimus treatment. However, in our study, no significant differences in anti-dsDNA changes were observed between the two groups among LN patients, aligning with the results from other cohorts involving LN patients.^[17,18]

Sirolimus has also shown promising effect on steroid-sparing in a single-arm, open-label, phase 1/2 trial involving 43 SLE patients.^[19] In this study, the mean daily dose of prednisone required to control disease activity decreased from 23.7 mg to 7.2 mg after 12 months of sirolimus treatment. Similarly, our study observed a trend toward steroid dose reduction, particularly at 12-month follow-up. However, no significant difference was noted between the two groups. This may be attributed to the relatively low initial glucocorticoid dosage in our study and the fact that MMF, a classic immunosuppressant, is also known to have steroid-sparing effects in LN.

There were several limitations in our study. Firstly, as a real-world study with retrospective analysis of prospectively collected data, there were inherent differences in baseline characteristics, disease conditions, and previous background immunosuppressants between the two groups. Although we utilized PSM method to minimize the selection bias, some residual confounding might still have influenced the results. Secondly, the proportion of LN patients who underwent renal biopsy in CSTAR cohort was relatively low (16.6% in previous report,^[46] 25.5% in our study). It might be attributed to the fact that the CSTAR cohort was primarily established by rheumatologist rather than nephrologists, who have more experience with renal biopsy. Future efforts should focus on improving the implementation of renal biopsy for active LN patients. Third, our study lacked longterm follow-up data for sirolimus in LN, as reported in earlier retrospective study.^[18] The number of patients with 12-month or longer follow-up data was limited, which may introduce bias in the data analyses. Therefore, further high-quality, randomized, double-blind, controlled clinical trials involving more patients with renal biopsy data and extended follow-up duration are needed to establish the efficacy and safety of sirolimus in LN patients.

In conclusion, this study represented the first comparison of the effectiveness and safety of sirolimus versus MMF, the SoC recommended by LN guidelines, in the treatment of LN. It also provides T2T evidence supporting the use of sirolimus in LN. Sirolimus and MMF demonstrated similar effectiveness in achieving disease activity remission and glucocorticoid dosage tapering. Moreover, sirolimus exhibited greater effectiveness in improving serological parameters compared to MMF. Importantly, sirolimus was well tolerated among LN patients. These findings suggest that sirolimus is a promising novel therapeutic option for LN. However, further high-quality studies are needed to confirm the efficacy and safety of sirolimus for LN treatment.