Lupus nephritis (LN) leads to chronic kidney disease (CKD) through progressive fibrosis. Mycophenolate inhibits inosine monophosphate dehydrogenase and is a standard treatment for LN. The mammalian or mechanistic target of rapamycin (mTOR) pathway is activated in LN. Rapamycin inhibits mTOR and is effective in preventing kidney transplant rejection, with the additional merits of reduced incidence of malignancies and viral infections. The effect of mycophenolate or rapamycin on kidney fibrosis in LN has not been investigated. We investigated the effects of mycophenolate and rapamycin in New Zealand Black and White first generation (NZB/W F1) murine LN and human mesangial cells (HMCs), focusing on mechanisms leading to kidney fibrosis. Treatment of mice with mycophenolate or rapamycin improved nephritis manifestations, decreased anti-double stranded (ds) DNA antibody titer and reduced immunoglobulin G (IgG) deposition in the kidney. Both mycophenolate and rapamycin, especially the latter, decreased glomerular mTOR Ser2448 phosphorylation. Renal histology in untreated mice showed mesangial proliferation and progressive glomerulosclerosis with tubular atrophy, and increased expression of transforming growth factor β1 (TGF-β1), monocyte chemoattractant protein-1 (MCP-1), α-smooth muscle actin (α-SMA), fibronectin (FN) and collagen. Both mycophenolate and rapamycin ameliorated the histopathological changes. Results from in vitro experiments showed that both mycophenolate and rapamycin decreased mesangial cell proliferation and their binding with anti-dsDNA antibodies. Mycophenolate and rapamycin also down-regulated mTOR and extracellular signal-regulated kinase (ERK) phosphorylation and inhibited fibrotic responses in mesangial cells that were induced by anti-dsDNA antibodies or TGF-β1. Our findings suggest that, in addition to immunosuppression, mycophenolate and rapamycin may reduce fibrosis in LN, which has important implications in preventing CKD in patients with LN.
Lupus nephritis (LN) is a severe and common manifestation of systemic lupus erythematosus (SLE), a debilitating autoimmune disease characterized by a loss of self-tolerance, anti-double-stranded (ds) DNA antibody production and immune-mediated kidney injury. A significant proportion of LN patients develop chronic kidney disease (CKD) and end-stage renal disease (ESRD). Abatement of acute immune-mediated kidney injury and prevention of disease flares are critical to ensure optimal long-term kidney and patient survival . Current standard-of-care treatment for LN entails high-dose corticosteroids given together with mycophenolate or cyclophosphamide to induce remission, followed by long-term maintenance with low-dose corticosteroids and either mycophenolate or azathioprine to prevent disease flares [2–6]. Mycophenolate inhibits lymphocyte proliferation through its effect on inosine monophosphate dehydrogenase .
Dysregulated activation of the mammalian target of rapamycin (mTOR) pathway has been demonstrated to contribute to the pathogenesis of LN [8–10]. The mTOR pathway is an evolutionarily conserved serine-threonine kinase signaling pathway . Data to date show that the mTOR pathway plays important roles in cell proliferation, growth, metabolism and survival, and the biological functions are mediated through two different multi-protein complexes, namely mTOR complex 1 (mTORC1) and mTORC2 [11,12]. mTOR activation through mTORC1 or mTORC2 can be distinguished by distinct phosphorylation at serine residues at Ser2448 and Ser2481, respectively . We have previously shown that mTOR inhibition with rapamycin delayed the onset of nephritis, and also reduced the progression of established nephritis, in New Zealand Black and White first generation (NZB/W F1) murine LN [8,9]. Our results also showed that rapamycin decreased glomerular immune deposition and the expression of monocyte chemoattractant protein-1 (MCP-1) and regulated on activation normal T cell expressed and secreted (RANTES), accompanied by improved renal histology [8,9]. Other investigators have also reported that rapamycin treatment prolonged survival in murine lupus models [10,14]. mTOR inhibition is an established treatment to prevent kidney transplant rejection [15–17]. Additional advantages of this class of drugs include a lower incidence of malignancies and their antiviral effect [18–20]. These properties could potentially benefit patients with SLE, who show an increased incidence rate of malignancies, attributed to immune dysregulation and also prior exposure to immunosuppressive medications, especially cyclophosphamide [21,22]. We have reported original preliminary clinical evidence both on the short-term efficacy and long-term tolerability of mTOR inhibitor treatment in LN patients [23,24], suggesting that mTOR inhibitors might have a role in the clinical management of LN patients.
Immune-mediated inflammation and immunosuppressive interventions have been the major research emphases in LN. Yet, progression of CKD due to progressive renal fibrosis affects most patients, and is the cause of ESRD . Currently there is no therapy for renal fibrosis. There are data, including some from our group, to show that mycophenolate and rapamycin, in addition to their immunosuppressive effects, also possess anti-proliferative effects and may reduce the pro-fibrotic responses in fibroblasts, mesangial cells and renal tubular epithelial cells [26–32]. In the present study, we investigated the effects of mycophenolate and rapamycin in NZB/W F1 murine LN and in human mesangial cells (HMCs), focusing on processes relevant to renal fibrosis.
Materials and methods
Chemicals, assays and drugs
All chemicals were of the highest purity and were purchased from Sigma–Aldrich (Tin Hang Technology Ltd, Hong Kong) unless otherwise stated. Primary HMCs were purchased from Lonza Cologne GmbH (Gene Company Limited, Hong Kong). Tissue culture flasks were purchased from Falcon (Becton-Dickinson, Gene Company Limited, Hong Kong). RPMI 1640 culture medium, fetal bovine serum (FBS), penicillin/streptomycin, l-glutamine, donkey anti-mouse Alexa Fluor 448, donkey anti-rabbit Alexa Fluor 594 and DAPI were purchased from Life Technologies Ltd (Thermo Fisher Scientific, Hong Kong). Antibodies to human and mouse fibronectin (FN), α-smooth muscle actin (α-SMA), podocin and β-actin were purchased from Sigma–Aldrich (Tin Hang Technology Ltd, Hong Kong). Anti-mouse transforming growth factor β1 (TGF-β1) antibody was purchased from Santa Cruz Biotechnology Inc. (Genetimes Technology International Holding Limited, Hong Kong). Antibodies to phosphorylated (phospho) (Ser2448 and Ser2481) and total mTOR and phospho and total phosphoinositide 3-kinase (PI3K) were purchased from Cell Signaling Technology (Gene Company Limited, Hong Kong). Donkey anti-human Alexa Fluor 488 was purchased from Jackson Immunoresearch Laboratories, Inc (Genetimes ExCell International Holdings Ltd, Hong Kong). Antibodies to human vascular endothelial growth factor receptor 1 (VEGFR1) and platelet-derived growth factor receptor β1 (PDGFRβ1) were purchased from Abcam (HK) Limited, Hong Kong. Mowiol® 4-88 immunofluorescence mounting medium was purchased from Polysciences, Inc. (Genetimes ExCell International Holdings Ltd, Hong Kong). Serum and/or urine creatinine, albumin and urea levels were measured in NZB/W F1 mice using QuantiChrom™ Creatinine, Albumin and Urea Assay Kits respectively (BioAssay Systems, California, U.S.A.). Anti-dsDNA antibody level was determined in serum samples using Mouse Anti-dsDNA Antibody Quantitative enzyme-linked immunosorbent assay (ELISA) Kits according to the manufacturer’s instructions (Alpha Diagnostic Inc., Onwon Trading Ltd, Hong Kong). Mycophenolate was provided by Roche Diagnostics (Palo Alto, California, U.S.A.), and Rapamycin (Rapamune) was purchased from Wyeth Hong Kong Limited.
All animal experiments were approved by the Institutional Committee on the Use of Live Animals in Teaching and Research. Female NZB/W F1 mice were housed in a specific pathogen-free animal facility at the University of Hong Kong, and kept under normal housing conditions in a 12-h night and day cycle. Water and standard chow were available ad libitum. Treatment commenced when mice were 22–25 weeks of age when they developed proteinuria, defined as >300 mg/dl detected on two separate occasions at least 2 days apart. Mice were randomized into three groups to receive (1) vehicle (Control group), (2) mycophenolate (100 mg/kg/day) and (3) rapamycin (3 mg/kg/day) for periods up to 12 weeks. The effect and clinical relevance of the drug dosages used were established in previous studies [8,9,33]. Treatment was administered by daily oral gavage for 2, 6 and 12 weeks, after which time mice were killed, sera and urine collected and kidneys harvested (Figure 1). After 12 weeks of treatment some mice had follow-up 2, 6 or 12 weeks post-therapy to determine treatment sustainability (n=6 mice per time point per treatment). Mice were initially randomized equally between the groups but since some mice in each group died before the experimental time point, an additional 12, 4 and 5 mice were allocated into Control, Mycophenolate and Rapamycin groups, respectively, to ensure that six mice were available for each time point per group. Six mice with established proteinuria as defined above, were killed at the beginning of the study for baseline clinical, serologic and histological data (T = 0).
Schematic diagram showing treatment, sample collection and studies conducted in NZB/W F1 mice
White blood cell count
To ensure that the dose of mycophenolate and rapamycin administered to NZB/W F1 mice had comparable immunosuppressive effect, the number of peripheral blood lymphocytes in Control and treated mice was determined at the time of killing as previously described . The cell number of white blood cell populations was determined by two independent observers without knowledge of the samples, and expressed as number of cells/ml.
All samples were measured in duplicate for all assays. Anti-dsDNA antibody level was determined in serum samples from NZB/W F1 mice using anti-dsDNA immunoglobulin G (IgG) quantitative ELISA kits according to the manufacturer’s instructions (Alpha Diagnostic Inc., Onwon Trading Ltd, Hong Kong). Lower and upper limits of detection were 50 and 1000 IU/ml, respectively, and values greater than mean + 2 SD of anti-dsDNA antibody level detected in 36-week-old C57BL/6N mice, i.e. 15.84 IU/ml [12.04 + (2 × 1.90)] was considered seropositive. Serum urea and creatinine levels were measured using QuantiChrom™ Urea and Creatinine assay kits, respectively. Urine albumin-to-creatinine ratio was measured weekly in spot urine using QuantiChrom™ Albumin and Creatinine assay kits.
Assessment of renal histopathology
Paraffin-embedded kidney sections (5 μm) from Control and treated mice were stained with Hematoxylin and Eosin (H&E), Periodic acid–Schiff (PAS) and Masson’s trichome. Renal histology and matrix protein deposition were scored by independent observers in a blinded manner. Renal lesions relating to inflammation and fibrosis in the glomerular and tubulo-interstitial compartments were graded 0–3 (for normal, mild, moderate and severe) and expressed as mean glomerular and tubulo-interstitial lesion scores for each group. Approximately 15 glomeruli, tubular, interstitial and vascular areas were evaluated for glomerular hypercellularity, mesangial matrix expansion, crescent formation, leukocyte and mononuclear cell infiltration, hyaline deposits, fibrinoid necrosis, tubular atrophy, protein cast deposition and vasculopathy for each mouse [33,34].
For semi-quantitative assessment of PAS (mesangial matrix expansion and enlargement of glomeruli) and Masson’s trichrome staining, images of approximately 15 glomeruli and tubules per mouse kidney were captured and graded as follows: 0 = 0–5% staining; 1 = >5–25% staining; 2 = >25–50% staining; 3 = >50–75% staining; 4 = >75% staining .
Cytochemical and immunohistochemical staining
Paraffin-embedded kidney sections (5 μm) from Control and treated mice were stained for TGF-β1, MCP-1, α-SMA and FN as previously described [27,33]. Signal detection and visualization was by the peroxidase-antiperoxidase method and specimens were counterstained with Hematoxylin . Staining of fibrotic mediators in the capillary loops, mesangium and tubulo-interstitium was assessed semi-quantitatively in a blinded manner in approximately 15 glomeruli and tubules per mouse kidney, and graded as described above for PAS and Masson’s trichrome staining . IgG deposition and/or mTOR staining in normal kidney specimens from patients undergoing nephrectomy, kidneys of patients with active proliferative LN (n=10) or mice with active disease was assessed by indirect immunofluorescence using snap-frozen renal specimens (8 μm) as previously described [27,36]. To determine glomerular mTOR localization, renal biopsies from LN patients were incubated with rabbit anti-human phospho-mTOR, followed by Alexa Fluor 448, and mouse anti-human PDGFRβ1 (mesangial marker) or VEGFR1 (endothelial marker), following by Alexa Fluor 594. Localization of mTOR in podocytes was assessed by separate staining with rabbit anti-human phospho-mTOR and rabbit anti-human podocin in sequential kidney sections. Sections were mounted in fluorescent mountant and epifluorescence viewed using a Nikon 80i upright fluorescent microscope and Spot RT3 slider digital camera system (Chintek Scientific (China) Ltd, Hong Kong). In mouse and human sections, 15 and 4–6 glomeruli per kidney section respectively, were analyzed, and florescence in the glomerular capillary walls, mesangium and podocytes were scored blindly on a scale of 0–3 (0 = no staining; 1 = weak staining; 2 = moderate staining, 3 = strong staining) [36,37].
Human polyclonal anti-dsDNA antibodies
Polyclonal IgG anti-dsDNA antibodies were isolated from the sera of 11 patients with active LN using sequential affinity chromatography as previously described [38,39]. Anti-dsDNA activity and IgG concentration in each preparation was determined using a commercial ELISA according to the manufacturer’s instructions (Bio-Rad, Hong Kong) and with an in-house ELISA, respectively . All anti-dsDNA antibody preparations were pre-treated with deoxyribonuclease (DNase) I prior to use .
Primary HMCs were maintained in RPMI 1640 medium supplemented with insulin (5 μg/ml), transferrin (5 μg/ml), l-glutamine (2 μM), penicillin (100 U/ml), streptomycin (100 μg/ml) and 10% FBS. All experiments were performed on growth-arrested cells of the fifth to seventh passages. The effect of mycophenolate and rapamycin on cell proliferation was investigated in HMCs cultured in 96-well microtiter plates. Sixty percent confluent, growth-arrested HMCs were incubated with 10% FBS in the presence or absence of mycophenolate (0–5 μg/ml) or rapamycin (0–3 ng/ml) for up to 72 h, and cell proliferation assessed using the MTT assay according to the manufacturer’s instructions (Roche Diagnostics (Hong Kong) Limited, Hong Kong). For assessment of anti-dsDNA antibody binding, 80% confluent growth-arrested cells were incubated with mycophenolate (5 μg/ml) or rapamycin (3 ng/ml) for 24 h at 37°C prior to incubation with Control IgG or anti-dsDNA antibodies (10 μg/ml, final concentration) for 30 min at 4°C, then cells were washed with PBS and fixed with 1% paraformaldehyde and the amount of bound IgG assessed as previously described . To determine the effect of mycophenolate and rapamycin on fibrotic processes, HMCs were incubated with mycophenolate (5 μg/ml) or rapamycin (3 ng/ml) for 1 h at 37°C prior to incubation with serum-free medium (SFM), Control IgG or anti-dsDNA antibodies (10 μg/ml, final concentration) for up to 72 h as previously described [38,39]. In separate studies, growth-arrested HMCs were incubated with mycophenolate (5 μg/ml), rapamycin (3 ng/ml) and inhibitors to extracellular signal-regulated kinase (ERK) (PD98059, 10 μM) and PI3K (LY294002, 25 μM) for 1 h prior to incubation with SFM, Control IgG or anti-dsDNA antibodies for 24 h. To confirm that the anti-fibrotic properties of mycophenolate and rapamycin were independent of their actions on anti-dsDNA antibody binding, in further studies, growth-arrested HMCs were incubated with mycophenolate (1 and 5 μg/ml) or rapamycin (1 and 3 ng/ml) for 1 h prior to incubation with SFM or exogenous TGF-β1 (10 ng/ml) for up to 48 h. The concentrations of mycophenolate and rapamycin used in experiments were previously demonstrated to be clinically relevant levels since they represent blood trough levels in renal transplant recipients and LN patients when given a daily dose of 2–5 mg rapamycin and 2–3 g mycophenolate, respectively [24,40–43]. The concentration of anti-dsDNA antibody and Control IgG used to stimulate HMCs was the optimum IgG concentration for maximum cellular binding, and which also demonstrated the effect of anti-dsDNA antibodies on cellular functions pertaining to inflammatory and fibrotic processes [36,38,39]. Following stimulation, supernatants were decanted, cells were washed with PBS and lysed with 20 mM sodium acetate, pH 6.0, containing 4 M urea, 1% Triton X-100 and a cocktail of proteinase inhibitors (200 μl) [31,36].
Western blot analysis
Renal cortical tissue from Control and treated mice were homogenized in 50 mM Tris/HCl, pH 7.5, containing 1% Triton X-100 and proteinase inhibitors (10 μg/ml pepstatin A, 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 mM N-ethylmaleimide (NEM), 5 mM benzamidine-HCl, 0.01 mM ethylenediaminetetraacetic acid (EDTA) and 0.05 mM ε-amino-n-caproic acid). Samples were centrifuged at 12000×g at 4°C for 30 min and the supernatants collected.
Aliquots of renal cortical tissue or cell lysates (20 μg total protein content) were electrophoresed under denaturing conditions on 8% polyacrylamide gels to determine FN and collagen III expression, and on 12% polyacrylamide gels to determine phospho-mTOR, total mTOR, phospho-ERK, total ERK, phospho-AKT, total AKT, α-SMA and β-actin. Samples were transferred on to nitrocellulose membranes and immunoblotted with the relevant primary antibodies followed by the addition of secondary antibodies as previously described [27,36]. Bands were visualized with enhanced chemiluminescence (ECL), semi-quantitated by densitometry using ImageJ (NIH, U.S.A.) and expressed as arbitrary densitometric units (DU). Expression of FN and α-SMA was normalized to β-actin, and phospho-mTOR, phospho-ERK and phospho-AKT was normalized to total mTOR, ERK and AKT, respectively.
Total messenger ribonucleic acid (mRNA) was extracted from cortical renal specimens from Control and treated mice using RNeasy mini kits according to the manufacturer’s instructions (Qiagen Hong Kong Pte Ltd, Hong Kong). One microgram of total mRNA was reverse-transcribed into complementary deoxyribonucleic acid (cDNA) with moloney murine leukemia virus (M-MLV) reverse transcriptase using the random hexamer method . TGF-β1 gene expression was assessed by quantitative real-time PCR using Taqman gene expression assay (Assay-on-Demand ID code: Mm00441726_m1) on a LightCycler 480 II real time PCR system (Roche Diagnostics, DKSH Hong Kong Limited, Hong Kong). All samples were analyzed in triplicate, and TGF-β1 mRNA expression calculated using the ΔΔCt (2−ΔΔCt) method, normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). MCP-1 transcript in renal cortical tissue was determined in Control and treated mice using reverse transcription polymerase chain reaction (RT-PCR) as previously described . Gene transcripts were normalized to GAPDH. PCR primer sequences used were:
Measurement of MCP-1 in culture supernatant
HMCs were incubated with SFM, Control IgG or anti-dsDNA antibodies in the presence or absence of mycophenolate, rapamycin or specific inhibitors to ERK or PI3K, for up to 72 h, after which time the supernatant was decanted and centrifuged at 2000 rpm for 10 min to remove any cell debris. Secreted MCP-1 was measured using a commercial MCP-1 OptEIA™ ELISA kit according to the manufacturer’s instructions (BD Biosciences Pharmingen, Genetimes Technology International Holding Limited, Hong Kong). Lower and upper limits of detection were 15 and 1000 pg/ml respectively. Samples were measured in duplicate in serial dilution and normalized to their cellular protein content.
Animal results are expressed as mean ± SEM. All in vitro studies were repeated three times and expressed as mean ± SD. Statistical analysis was performed using Prism 6.0 for Windows (GraphPad Software, Inc., California, U.S.A.). Inter-group and intra-group comparisons for animal studies were assessed by repeated measures ANOVA followed by Bonferroni’s multiple comparison post-test. Inter-group comparison for in vitro studies was assessed by ordinary ANOVA followed by Bonferroni’s multiple comparison post-test or Mann–Whitney U test, as appropriate. Survival rates were determined using Fisher’s exact test. Two-tailed P<0.05 was considered statistically significant.
Effect of mycophenolate or rapamycin treatment on proteinuria and kidney function in NZB/W F1 mice
Mice treated with mycophenolate or rapamycin for 12 weeks showed similar peripheral blood lymphocyte counts, which were significantly lower than untreated Controls (Figure 2A). There was no effect on body weight, which did not differ between the treatment groups (data not shown). Survival rate was 89.9, 88.6 and 89.9% in mycophenolate-treated, rapamycin- treated and Control mice, respectively, at this time-point (P=NS, mycophenolate or rapamycin vs Control). At 12 weeks after cessation of treatment, the survival rates were 77.03, 75.92 and 44.48%, respectively (P=0.021, mycophenolate or rapamycin vs Control) (Figure 2B). Proteinuria, serum creatinine and urea, and anti-dsDNA antibody levels increased with progressive disease in untreated Controls (Figure 2C–F). Treatment with mycophenolate or rapamycin resulted in significantly lower level of proteinuria compared with Controls after 6 weeks, which persisted for 2 weeks after cessation of treatment. Proteinuria reduction was ≥50% in 71.43 and 33.33% of mycophenolate- and rapamycin-treated mice, respectively, after 12 weeks of treatment (P=0.02, mycophenolate vs Control, P=NS, rapamycin vs Control). Serum creatinine and urea levels were significantly reduced to comparable levels between mycophenolate- and rapamycin-treated mice, compared with Control mice, and the better kidney function in treated groups lasted for 6–12 weeks after cessation of treatment (Figure 2D,E). Anti-dsDNA antibody level was also lower in the treated groups (Figure 2F).
The effect of mycophenolate or rapamycin on survival and clinical and serological parameters in NZB/W F1 mice
Glomerular mTOR phosphorylation and IgG deposition
Increased mTOR phosphorylation at Ser2448, but not at Ser2481, was observed in the glomeruli of NZB/W F1 mice with active nephritis, and it co-localized with IgG deposition. mTOR activation was also noted in the tubulo-interstitium, but to a lesser extent (Figure 3A). Glomerular IgG deposition increased with progression of nephritis (Figure 3B). Treatment with mycophenolate or rapamycin reduced IgG deposition. Glomerular IgG deposition gradually increased after treatment was stopped (Figure 3B). Treatment with rapamycin for 12 weeks resulted in near complete suppression of mTOR activation. Mycophenolate treatment also suppressed mTOR activation, but this occurred later than the rapamycin group and to a lesser extent. mTOR activation reappeared after treatment was stopped (Figure 3C).
Effect of mycophenolate or rapamycin on IgG deposition and mTOR phosphorylation in NZB/W F1 mice
Renal histology and fibrosis
Nephritis was accompanied by progressive glomerular hypertrophy, basement membrane thickening, increased matrix protein deposition and mononuclear cell infiltration in the periglomerular and tubulo-interstitial area. Tubular atrophy, cast formation, glomerular synechiae, glomerulosclerosis and interstitial fibrosis were noted after approximately 12 weeks. Treatment with mycophenolate or rapamycin attenuated these abnormalities to a similar extent. The abnormal findings reappeared approximately 6 weeks after stopping treatment (Figures 4–6). The mesangial matrix and tubulo-interstitium showed increased collagen deposition with active nephritis, and this was reduced by both treatments (Figure 6).
Effect of mycophenolate or rapamycin on kidney histology in NZB/W F1 mice
Effect of mycophenolate or rapamycin on matrix protein deposition in NZB/W F1 mice
Effect of mycophenolate or rapamycin on collagen deposition in NZB/W F1 mice
Real-time PCR and RT-PCR was used to assess the gene expression and cytochemistry was used to assess protein expression of fibrosis mediators TGF-β1 and MCP-1 [44,45]. In Control mice, TGF-β1 mRNA expression increased progressively. Treatment with mycophenolate or rapamycin for 12 weeks significantly reduced TGF-β1 mRNA expression, which was sustained for 6 weeks after treatment was stopped (Figure 7A). TGF-β1 protein expression was detected in the glomerulus and tubulo-interstitium after 12 and 18 weeks, respectively, and treatment with either drug for 12 weeks reduced renal TGF-β1 expression, which was sustained for 2–12 weeks after stopping treatment (Figure 7C). MCP-1 gene expression increased progressively in Control mice, with maximum transcript level at 18 weeks. Increased MCP-1 protein expression was predominantly localized to proximal tubular epithelial cells and interstitium. Treatment with either drug for 2 weeks significantly decreased MCP-1 transcript (Figure 7B), and decreased tubulo-interstitial MCP-1 protein expression was noted after 12 weeks of treatment, reappearing after treatment was stopped (Figure 7D).
Effect of mycophenolate or rapamycin on TGF-β1 and MCP-1 gene and protein expression in NZB/W F1 mice
α-SMA expression increased in Control mice as nephritis progressed, predominantly in interstitial cells with weaker staining in podocytes and glomerular endothelial and mesangial cells. Treatment with mycophenolate or rapamycin for 12 weeks resulted in similar reduction in α-SMA expression (Figure 8A,C). Increased FN expression in Control mice was localized to both glomeruli and tubulo-interstitium. Treatment with either drug for 12 weeks resulted in reduced FN expression (Figure 8B,D).
Effect of mycophenolate or rapamycin on α-SMA and FN protein expression in NZB/W F1 mice
Effect of mycophenolate and rapamycin on proliferation and anti-dsDNA IgG binding in HMCs
Treatment with mycophenolate (1 and 5 μg/ml) for 24 h significantly reduced FBS-induced HMC proliferation by 49.61 and 53.86%, respectively (P<0.001 for both concentrations), and the anti-proliferative effect was sustained for 72 h. Treatment with rapamycin (1 and 3 ng/ml) for 24 h reduced FBS-mediated cell proliferation by 46.43 and 48.57%, respectively (P<0.001 for both concentrations). The anti-proliferative effect was similar between mycophenolate and rapamycin (Figure 9A,B). Data from cellular ELISA using HMC as substrate showed significantly more cellular binding by anti-dsDNA IgG antibody than normal IgG (2.86 ± 0.22 vs 6.30 ± 0.38 μg IgG binding/μg cell protein, P<0.001) (Figure 9C). Prior incubation of HMC with either drug for 24 h did not affect their binding by normal IgG, but the binding by anti-dsDNA was reduced by 45.24 and 41.27% respectively in cells treated with mycophenolate or rapamycin (P<0.001 for both drugs) (Figure 9C).
Effect of mycophenolate and rapamycin on cell proliferation and anti-dsDNA antibody binding in HMCs
Effect of mycophenolate and rapamycin on mTOR, PI3K and ERK phosphorylation, and the expression of α-SMA, FN and collagen III in HMCs
We and others have reported that increased mTOR, PI3K and mitogen-activated protein kinase (MAPK) activation contributes to fibrogenesis in mesangial cells [36,46–48]. Under basal conditions, HMCs showed weak expression of Ser2448 phosphorylated mTOR, AKT or ERK. Control IgG had no effect on mTOR, AKT or ERK phosphorylation. Anti-dsDNA antibodies isolated from LN patients significantly increased mTOR Ser2448 phosphorylation after 24 h, which was sustained for 72 h (Figures 10A and 11A). Anti-dsDNA antibodies increased ERK, but not PI3K activation (Figures 10B and 11B,C). Rapamycin, LY294002 (PI3K inhibitor), and mycophenolate significantly decreased anti-dsDNA antibody-induced mTOR activation, while PD98059 (ERK inhibitor) had no effect (Figures 10 and 11). Mycophenolate, rapamycin, PD98059 and LY294002 decreased ERK phosphorylation (Figures 10B and 11C). Both drugs did not affect PI3K/AKT phosphorylation (Figure 11B).
Effect of mycophenolate and rapamycin on mTOR and ERK activation in HMCs
Effect of mycophenolate, rapamycin, PD98059 and LY294002 on mTOR, AKT and ERK activation in HMCs
Anti-dsDNA antibodies significantly increased α-SMA, FN and collagen III expression after 24 h, when compared with SFM and Control IgG. PI3K, but not mTOR or ERK phosphorylation, contributed to increased α-SMA expression induced by anti-dsDNA antibodies, since this was reduced by pre-incubation with LY294002 but not rapamycin or PD98059 (44.86% reduction, P<0.01). Mycophenolate reduced anti-dsDNA antibody-induced α-SMA (44.29% reduction, P<0.01) (Figure 12A). Constitutive and anti-dsDNA antibody-induced FN expression was mediated through mTOR, PI3K and ERK phosphorylation. Mycophenolate, rapamycin, LY294002 and PD98059 all reduced FN expression (Figure 12A). Pre-incubation of HMC with mycophenolate, rapamycin or LY294002, but not PD98059 significantly decreased anti-dsDNA antibody-induced collagen III expression (Figure 12A). Compared with SFM and normal IgG, anti-dsDNA antibodies significantly increased MCP-1 secretion in HMCs, which was mediated through PI3K signaling (Figure 12B). Inhibition of PI3K, mTOR or ERK signaling with LY294002, rapamycin or PD98059 resulted in 95.14, 39.25 and 33.08% reduction in MCP-1 secretion, respectively. Mycophenolate exposure reduced MCP-1 secretion by 37.83% (Figure 12B). In separate experiments, we found that pre-incubation of HMCs with mycophenolate or rapamycin significantly decreased TGF-β-induced α-SMA and FN expression in HMCs (Figure 12C).
Effect of mycophenolate, rapamycin, PD98059 and LY294002 on expression of mediators of fibrosis in HMCs
Glomerular mTOR phosphorylation in renal biopsies from LN patients
Renal biopsies were obtained from ten patients with active diffuse proliferative LN (nine females, one male; 40.30 ± 12.22 years of age; anti-dsDNA antibody titer 215.40 ± 109.60 IU/ml; serum creatinine level 95.90 ± 47.27 µmol/l and proteinuria 2.51 ± 2.25 g/D). LN kidney biopsies showed intense glomerular staining for mTOR phosphorylation at Ser2448 compared with little staining in healthy subjects (Figure 13A). mTOR phosphorylation at Ser2481 was not detected in both LN patients and normal kidney tissue (data not shown). Increased mTOR phosphorylation in LN co-localized with IgG deposition (Figure 13B) and was predominantly localized to mesangial cells, glomerular endothelial cells and to a lesser extent podocytes (Figure 13C,D).
mTOR expression and glomerular localization in renal specimens from patients with active LN
Despite improvements in immunosuppressive management over the past few decades, a significant proportion of LN patients develop CKD on long-term follow-up, the latter characterized by progressive accumulation of matrix proteins in the kidney parenchyma and replacement of normal nephrons by fibrous matter, resulting in progressive loss of renal function. CKD is irreversible and progresses to ESRD. Reducing acute immune-mediated kidney injury and prevention of progressive renal fibrosis are both important for long-term kidney and patient survival .
Against this background of the common occurrence of CKD in LN, we compared the effect of mycophenolate and rapamycin in NZB/W F1 mice and cultured HMCs focusing on processes relevant to kidney fibrosis. The doses of mycophenolate and rapamycin chosen in the experiments were based on literature data [8,9,27,30,33], and the data showed that the exposure for the two drugs in the experiments resulted in comparable immunosuppressive effects. While there had been concern that mTOR inhibitor treatment might induce proteinuria in some kidney transplant patients, this was not observed in our experiments. Rapamycin-treated mice showed a marked reduction in glomerular mTOR activation after 2 weeks, with almost complete inhibition after 12 weeks. Mycophenolate also reduced mTOR activation, although the decrease was not as rapid. In line with our findings, other investigators have reported that mycophenolate suppressed PI3K/AKT/mTOR pathway in a rat model of temporal lobe epilepsy .
Increased TGF-β1, MCP-1, α-SMA, FN and collagen expression were observed in the glomeruli and tubulo-interstitium of NZB/W F1 mice with active nephritis. TGF-β1 is an important mediator of kidney fibrosis, and increased MCP-1 expression has been associated with crescent formation, renal inflammatory cell infiltration and tubulo-interstitial fibrosis . Mycophenolate and rapamycin showed similar efficacy in reducing MCP-1, α-SMA, FN and collagen expression in NZB/W F1 mice. While both mycophenolate and rapamycin decreased TGF-β1 expression, the suppressive effect of rapamycin on TGF-β1 expression appeared more sustained even after cessation of treatment. We and others have shown that renal fibrosis is attributed, at least in part, to increased activation of protein kinase C (PKC), TGF-β/SMAD, mTOR and MAPK signaling pathways [27,30,31,51]. Mycophenolate decreased fibrogenic processes in the kidney through down-regulation of PKC, ERK, p38 MAPK and c-Jun N-terminal kinase (JNK) activation [27,30]. In contrast, the anti-fibrotic effect of rapamycin through signaling pathways other than mTOR is less well defined. The anti-fibrotic effect of rapamycin in patients with chronic allograft nephropathy appears to be mediated, at least in part, through a reduction in plasminogen activator inhibitor-1 (PAI-1) expression . In cultured AKR-2B-derived fibroblasts, rapamycin inhibited TGF-β1 mediated anchorage-independent growth but had no effect on TGF-β1 induced FN and collagen expression . This discrepancy may be cell-specific.
Mesangial immune complex deposition is characteristic in LN, and is accompanied by mesangial hypercellularity and local expression of chemokines that contributes to influx of immune and inflammatory cells into the glomerulus. Progression of LN is characterized by increased inflammatory cell infiltration, increasing renal damage and increased matrix protein and fibrous tissue expression . We previously reported on the anti-proliferative effect of mycophenolate on mesangial cell proliferation . The present data from in vitro studies in HMCs showed comparable effect of mycophenolate and rapamycin in reducing FBS-induced mesangial cell proliferation, suggesting a direct anti-proliferative property. Rapamycin has been reported to inhibit renal epithelial cell proliferation by inhibiting G1 to S phase transition . Also, both mycophenolate and rapamycin significantly decreased anti-dsDNA antibody binding to HMCs. Anti-dsDNA antibodies can bind to the surface of HMCs through annexin II, α-actinin or ribosomal P protein, triggering downstream inflammatory and fibrotic processes [36,56–59]. The detailed mechanisms through which mycophenolate and rapamycin might interfere with the interaction between anti-dsDNA antibody and cell surface antigens remain to be investigated. Altered glycosylation of membrane proteins, or protein catabolism through mTORC1 are possibilities [60,61].
Anti-dsDNA antibodies induced mTOR and ERK activation in HMCs, which was accompanied by downstream induction of MCP-1 secretion, and α-SMA, FN and collagen III expression. Suppression of PI3K or mTOR activation with LY294002 and rapamycin, respectively, attenuated anti-dsDNA antibody-induced inflammatory and fibrotic processes in HMCs and highlights the importance of the PI3K/AKT/mTOR signaling pathway in LN pathogenesis. Mycophenolate also decreased mTOR activation, MCP-1 secretion and fibrogenesis induced by anti-dsDNA antibodies in HMC, although the suppressive effect on mTOR activation was less rapid and less effective than that of rapamycin. Neither mycophenolate nor rapamycin has any effect on PI3K phosphorylation suggesting that PI3K activation is upstream of the actions of these immunosuppressive agents. Mycophenolate and rapamycin decreased ERK activation induced by anti-dsDNA antibodies, and the inhibitory effect was comparable between both drugs. A recent study demonstrated that inhibition of ERK signaling pathway by trametinib ameliorated mTORC1 activation in murine models of renal fibrosis , suggesting that the ERK signaling pathway was upstream of mTORC1. We previously reported that mycophenolate could decrease anti-dsDNA antibody-induced ERK phosphorylation in proximal tubular epithelial cells, and also attenuated ERK activation in the tubulo-interstitium of NZB/W F1 mice with active disease [30,31]. It is possible that mycophenolate reduced mTOR phosphorylation through down-regulation of ERK phosphorylation [30,31], although further studies are necessary. Inhibition of ERK activation was also accompanied by a significant reduction in MCP-1 secretion and FN expression but not collagen III expression. From our experiments the suppressive effect of mycophenolate on collagen III expression appeared more apparent than that of rapamycin. Mycophenolate, but not rapamycin or PD98059, reduced α-SMA expression in HMCs, suggesting that the inhibitory effect of mycophenolate on mesangial cell activation was not mediated through mTOR or ERK activation. In rat mesangial cells, elevated glucose concentration increased α-SMA expression through up-regulation of PKC signaling . We previously demonstrated that mycophenolate attenuated anti-dsDNA antibody induced PKC-α, -βI and -βII activation and downstream TGF-β1 secretion in HMC . It is therefore possible that mycophenolate may reduce anti-dsDNA antibody induction of α-SMA expression through inhibition of PKC activation. Figure 14 summarizes the effect of mycophenolate and rapamycin on signaling pathways and mediators of fibrosis induced by anti-dsDNA antibodies in HMCs. Both mycophenolate and rapamycin reduced constitutive and TGF-β1-induced α-SMA and FN expression, suggesting that they directly reduce the pro-fibrotic cellular response in HMC.
Schematic diagram summarizing the effect of mycophenolate and rapamycin on signaling pathway activation and mediators of fibrosis induced by anti-dsDNA antibodies in HMCs
Patients and mice with LN show increased glomerular mTOR activation at Ser2448 but not Ser2481, suggesting that mTORC1 and not mTORC2 is induced and may be involved in the pathogenesis leading to kidney injury or damage in LN. mTOR activation was observed in mesangial cells and glomerular endothelial cells, and to a lesser extent in podocytes. mTOR activation co-localizes with IgG deposition, highlighting an association between autoantibody deposition and mTOR activation. Our in vitro studies also demonstrated mTOR activation in HMCs following stimulation with anti-dsDNA antibodies. Our preliminary clinical evidence suggest efficacy and safety of using mTOR inhibitors in the treatment of patients with LN even as long-term therapy [23,24]. Based on the results from in vitro and animal experiments, and also the preliminary clinical data, further clinical investigations are warranted to define the role of mTOR inhibition in the clinical management of LN, which should be valuable in tailoring treatment according to the distinct characteristics of each patient with the objective of maximizing benefit while minimizing the potential untoward outcomes. In this context, the reduced incidence of malignancies and viral infections associated with mTOR inhibitor treatment, and the lack of nephrotoxicity, are potential benefits in the setting of LN.
In conclusion, data from this series of experiments show that mycophenolate and rapamycin reduce fibrosis in murine LN and the suppressive effect on pro-fibrotic cellular responses in mesangial cells suggest a direct effect on resident kidney cells that may be independent of their immunosuppressive actions, and the direct cellular effects by these agents are mediated through signaling pathways some of which are applicable to both.
LN leads to CKD through progressive fibrosis. mTOR activation is implicated in fibrogenesis, but its role in renal fibrosis of LN remains to be defined.
The present study compared the effect of mTOR inhibitor rapamycin with mycophenolate, the current standard-of-care treatment for LN, in NZB/W F1 mice, and the results showed comparable efficacy of the two drugs on improving renal histopathology including fibrosis and kidney function.
The data provide original evidence of anti-fibrotic effect of mycophenolate and rapamycin in murine LN, mediated in part through their actions on mesangial cells.
We thank Dr Qing Zhang for technical assistance.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by the Hong Kong Research Grant Council General Research Fund [grant number HKU 7550/06M]; the UGC Matching Grant Scheme (Phases V and VI); the Department of Medicine Academic Activities Fund; donations from Mr C.S. Yung, Mr S. Ho, and the Hui Hoy and Chow Sin Lan Charity Fund and the Family of Mr Hui Ming; the Endowment Fund established for the ‘Yu Chiu Kwong Professorship in Medicine’ (awarded to T.M.C.); and the Wai Hung Charitable Foundation Limited.
Study conception and design: S.Y. and T.M.C. Acquisition of data: C.Z.Z., C.C.Y.C., K.F.C. and M.K.M.C. Acquisition of clinical samples and retrieval of clinical data: D.Y.H.Y., M.K.M.M., K.W.C. and T.M.C. Analysis and interpretation of the data: S.Y., C.Z.Z., C.C.Y.C. and T.M.C. Drafting the article: S.Y. and T.M.C. Approval of the final version for submission: C.Z.Z., C.C.Y.C., K.F.C., M.K.M.C., D.Y.H.Y., M.K.M.M., K.W.C., S.Y. and T.M.C.
chronic kidney disease
enzyme-linked immunosorbent assay
extracellular signal-regulated kinase
end-stage renal disease
fetal bovine serum
glyceraldehyde 3-phosphate dehydrogenase
human mesangial cell
mitogen-activated protein kinase
monocyte chemoattractant protein-1
messenger ribonucleic acid
mammalian or mechanistic target of rapamycin
- NZB/W F1
New Zealand Black and White first generation
platelet-derived growth factor receptor β1
protein kinase C
reverse transcription polymerase chain reaction
systemic lupus erythematosus
α-smooth muscle actin
transforming growth factor β1
vascular endothelial growth factor receptor 1