The Mineralocorticoid Receptor Antagonist Eplerenone Suppresses Interstitial Fibrosis in Subcutaneous Adipose Tissue in Patients With Type 2 Diabetes

In type 2 diabetes, modifications of adipocytes alter the expression of various collagens, which are increasingly appreciated as key contributors to the development of pathological processes (1,2). The mineralocorticoid receptor (MR) could play a role since primarily mice models have shown that its expression is amplified in obesity and associated with impaired glucose metabolism (3–5) and that treatment with an MR antagonist (MRA) reverses the expression of several metabolic-related changes in adipocytes (3,6). Furthermore, the MRA eplerenone has been shown to reduce collagen gene expression and inhibit myocardial fibrosis in rodent models (7,8). An upregulation of extracellular matrix (ECM) components with several collagens, primarily collagen types I (Col-1) and VI (Col-6), is found in the adipose tissue of diabetic animals (1) in which knockout models of Col-6 improved metabolism and reduced adipose fibrosis (1). More recently, we and others have examined important postreceptor molecular targets of the MR, such as galectin-3 (Gal-3), whose increased expression has been associated with cardiorenal fibrosis (9,10). Furthermore, we identified neutrophil gelatinase–associated lipocalin (NGAL) as an important target of MR-induced cardiac and vascular remodeling (11,12) and lipocalin-like prostaglandin D2 synthase (PTGDS), which is involved in metabolism, as an MR target in adipose tissue in obese mice (5).

Importantly, translation of the evidence from experimental studies in mice to human adipose tissue is lacking. Therefore, several major questions remain unanswered, especially regarding the expression and function of the MR in human adipose tissue, the impact of blocking the MR on pericellular fibrosis, and the role of these MR target molecules in human adipocytes. In this prespecified substudy of the Mineralocorticoid Receptor Antagonist in type 2 Diabetes (MIRAD) trial, the primary objective was to investigate the effects of high doses of the selective MRA eplerenone on fibrosis, expression of collagens, and MR signaling target molecules in subcutaneous adipose tissue (ScAT) in patients with type 2 diabetes.

The study population consisted of patients with type 2 diabetes enrolled in the MIRAD study, which was a 26-week, randomized, double-blind, placebo-controlled trial that investigated the effects of high-dose eplerenone on liver fat content in patients with type 2 diabetes (13). All patients were proposed to participate, with an aim of including 30 in this prespecified substudy. The study design was previously described (14). In the fasting state, subcutaneous fat biopsies were taken lateral to the umbilicus during a sterile procedure using a Bergström needle and snap frozen in liquid nitrogen. All adipose tissue analyses were performed blinded.

Plasma adiponectin was measured by radioimmunoassay following manufacturer’s instructions (Millipore). Serum leptin, Gal-3, and free NGAL were quantified by ELISA (R&D Systems).

Adipose tissue expression of collagen type I, metalloproteinase-1 (MMP-1), tissue inhibitors of metalloproteinase-1 (TIMP-1), and TIMP-2 was measured using 5–10 μg of proteins by ELISA following the manufacturer’s instructions (R&D Systems). All samples were run in triplicate, and the ELISA method was previously validated in our laboratory by a model of human myocardial aortic valve (15).

Aliquots of 20 μg of proteins were electrophoresed and transferred to nitrocellulose membranes (Bio-Rad). Membranes were incubated with primary antibodies for Col-3 (1:100; Santa Cruz Biotechnology), Col-4 (1:100; Abcam), Col-6 (1:100; Abcam), α-smooth muscle actin (SMA) (1:50; Sigma), and transforming growth factor-β (TGF-β) (1:100; Abcam) followed by secondary antibody and enhanced chemiluminescence reagents. Stain-free detection was performed as control (16). After densitometric analyses, optical density values were expressed as arbitrary units.

Total RNA was extracted with TRIzol reagent (Euromedex) and purified using the RNeasy kit (QIAGEN). Quantitative PCR analysis was performed with SYBR Green PCR technology (ABgene). Relative quantification was achieved with MyiQ software (Bio-Rad), according to manufacturer’s instructions. Data were normalized by hypoxanthine-guanine phosphoribosyltransferase, β-actin, and 18S and expressed as percentages relative to controls. All PCRs were performed at least in triplicate for each experimental condition. Specific primer sequences are presented in Supplementary Table 1.

Aliquots of adipose tissue samples containing 15 μg of proteins were resolved on a 10% SDS polyacrylamide gel containing 0.3% gelatin. The gel was rinsed with a solution of 2.5% Triton X-100 followed by incubation for 48 h at 37°C in 1,000 mmol/L Tris-HCl (pH 7.5) with 1,000 mmol/L CaCl₂ and 5,000 mmol/L NaCl. Gels were fixed in 40% methanol and 10% acetic acid and then stained for 30 min in 0.25% Coomassie blue R-250. MMP-2 and MMP-9 are expressed as the ratio between active and total MMP.

Paraffin-embedded adipose tissue sections (5 μm) were used. For sirius red staining, the slides were hydrated and incubated for 2 h with hematoxylin-eosin or for 1 h with picrosirius red. The immunochemistry was performed following the protocol of the BOND Polymer Refine Detection automatic immunostainer (Leica Biosystems) and registered on a computer using the Leica Biosystems program. The immunostaining program protocol included fixative solution, bond wash solution, immunohistochemistry blocker, and primary antibody incubation for Col-1 (1:500; Sigma) and Col-6 (1:500; Abcam). After primary antibody incubation, slides were incubated with post–primary poly-horseradish peroxidase-IgG. The signal was revealed by using 3,3'-diaminobenzidine substrate. As negative controls, samples followed the same procedure described above but in the absence of primary antibodies. Adipose tissue fibrosis analyses were performed by histomorphometry using Nikon NIS software with content color thresholds. The quantification of total fibrosis was expressed as the ratio of fibrous tissue area stained with sirius red/total tissue surfaces as described by Henegar et al. (17). At least three to five random fields at ×20 magnification for each biopsy specimen were measured, with a minimum of 35–40 adipocytes analyzed per individual. Moreover, sirius red staining was expressed as area stained with sirius red/mean area of adipocytes and as area stained with sirius red/number of adipocytes. All measurements and quantifications were performed blinded in the automated image analysis system (Nikon).

Clinical and demographic characteristics are presented as mean ± SD or numbers (percentages) as appropriate. Data from Western blotting, ELISA, RT-PCR, zymography, and histology are presented as mean ± SEM. Using paired t test, changes in mean values between baseline and week 26 within the eplerenone and placebo groups were analyzed. Independent t tests were used to compare differences between the eplerenone and placebo group. P < 0.05 was considered statistically significant. All patients signed informed consent before enrollment.

The data sets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

A total of 30 patients with type 2 diabetes were enrolled. Baseline characteristics were well balanced between groups (Table 1). Six patients dropped out of the study; thus, complete paired fat biopsies were obtained in 24 patients, 12 from each group. At baseline, no differences between groups were present in the expression of any mRNA or protein factors measured in ScAT (Figs. 1–3).

A decrease in mRNA levels of 46% of the MR was observed from baseline to week 26 in the eplerenone group (P < 0.0035) and reduction compared with placebo (P = 0.0286), with no change in the placebo group (Fig. 1). mRNA expression of the downstream molecule NGAL decreased from baseline with eplerenone but did not reach statistical significance, although reduced levels compared with placebo at week 26 were found (P = 0.039). Treatment with eplerenone reduced expression of Gal-3 from baseline at week 26 by 41% (P = 0.0269), with no difference compared with placebo, and no changes within the placebo group were observed (Fig. 1). Finally, mRNA expression of PTGDS decreased by 42% in the eplerenone group (P = 0.0478) (Fig. 1). There was a correlation between the reduced mRNA expression of the MR and Gal-3 mRNA expression (r = 0.46; P < 0.01) (Supplementary Fig. 2).

With regard to circulating biomarkers, a treatment effect of eplerenone on reduced plasma Gal-3 at week 26 was found (−2.8 ng/mL [95% CI, −0.8; −4.9]; P = 0.010), whereas there was no effect of eplerenone on concentrations of serum NGAL (Supplementary Table 2). Circulating adiponectin and leptin did not change after intervention with eplerenone (Supplementary Table 2).

At baseline, sirius red staining showed that collagen fibers were mainly organized in fibrotic bundles surrounding the adipocytes. Quantitative immunohistological analysis demonstrated a marked reduction, by ∼40% (P = 0.0367), in the abnormal pericellular fibrosis in eplerenone-treated patients at week 26 compared with placebo (Fig. 2). Furthermore, sirius red–marked fibrosis decrease was adjusted for adipocyte number (P = 0.0142) and for adipocyte area (P = 0.0236) in the eplerenone group at week 26 compared with baseline. The Col-1 and Col-6 immunostainings clearly showed reduction of these central collagens after 26 weeks of treatment with the MRA eplerenone (Fig. 2). Adipocyte size did not change with MR inhibition.

We investigated the mechanisms by which eplerenone reduces pericellular fibrosis by measuring the mRNA expression of the main collagen subtypes—Col-1a1, Col-3, Col-4, Col-5a1, and Col-6a1, Col-6a2, and Col-6a3—and the corresponding protein expression. At the mRNA level, Col-1a1 decreased by 39% in the eplerenone group (P = 0.0414) and compared with placebo at week 26 (P = 0.0003) (Fig. 3). Protein levels decreased accordingly, although not statistically significantly (Fig. 3). Col-6a1 mRNA levels decreased by 29% compared with baseline (P = 0.0148) and with placebo (P = 0.0491) (Fig. 3). Furthermore, subtypes Col-6a2 and Col-6a3 mRNA levels were reduced by 56% and 35%, respectively, compared with baseline levels (P = 0.0060 vs. P = 0.0227, respectively). In line with the decrease in the mRNA expression of the Col-6 α-subunits with eplerenone, we observed a 35% reduction in the protein expression of Col-6 and a decrease compared with placebo (P = 0.0152) (Fig. 3). In addition, gene expression of Col-4 was substantially reduced by 55% with eplerenone compared with baseline (P = 0.0339) (Fig. 3). We observed no difference in mRNA expression of Col-3 and Col-5a1 or in protein expression of Col-3 or Col-4. In accordance with the reduced collagen expressions, a decrease in the mRNA expression of the profibrotic factor α-SMA within the eplerenone group (P = 0.0429) and compared with placebo (P = 0.0202) was found (Fig. 3), with no significant change in TGF-β. Inhibition of the MR had no effect on the expression of MMP-1, TIMP-1, MMP-2 ratio, or MMP-9 ratio, while we observed a reduction in TIMP-2 protein expression of 81% (Supplementary Fig. 1).

The present randomized clinical trial provides the first evidence that inhibition of the MR reduces ScAT pericellular fibrosis in patients with type 2 diabetes. This reduction in histologically confirmed fibrosis was further demonstrated by a significant decrease in the expression of Col-1 and Col-6 at the mRNA and protein levels. The MR expression was reduced in the eplerenone group followed by a corresponding reduction in the MR target molecule NGAL, Gal-3, and PTGDS mRNA. These findings suggest a profound effect of the MRA eplerenone on ScAT of the present patients with type 2 diabetes.

Various independent research groups have shown in rodent models that MRAs revert the dysmetabolic processes found in adipose tissue in the obese and diabetic states (3,18,19). We demonstrate a profound impact on Col-1 and Col-6 both on protein and on mRNA levels. The beneficial impact of the MRA-induced reduction in Col-6 expression in the current study is supported by an experimental study that reported that elimination of Col-6 is linked with normalized adipose tissue structure and improved metabolism (1). The clinical implications of the reduced fibrosis in ScAT in patients with type 2 diabetes are yet to be determined since we did not demonstrate an effect of eplerenone on liver fat in the MIRAD trial (13). A human study reported an impact of increased pericellular fibrosis in ScAT on weight loss after bariatric surgery, thus suggesting a link with obesity (2). The antifibrotic effect of MRAs is well known in the cardiorenal setting (10). Accordingly, eplerenone reduced collagen gene expression and inhibited myocardial fibrosis in rat models and after myocardial infarction (7,8). We have demonstrated that the downstream molecules NGAL and Gal-3 are involved in the deleterious effects of aldosterone-mediating fibrosis in the cardiovascular system (9,20,21). However, Gal-3 expression is also increased in adipose tissue in obese rats (22). In the current study, we observed a downregulation of the MR and expression of NGAL (although not significant) and Gal-3 in adipocytes, with corresponding reduced serum Gal-3 concentrations in patients treated with eplerenone. Therefore, we demonstrate for the first time novel mechanistic insights of aldosterone activation of the MR in adipocytes, with a link to pericellular fibrosis in ScAT in patients with type 2 diabetes. These findings are in line with an experimental study that showed that inhibition of Gal-3 prevents an increase in ECM components, including Col-1a1, Col-6a1, and Col-6a2 (22). Furthermore, we show that the decrease in mRNA levels of the MR was followed by a decrease in the enzyme PTGDS, confirming that expression of MR and PTGDS is in parallel in human ScAT (5,23). Thus, our study indicates that treatment with an MRA is associated with Gal-3 and NGAL and may be protective of aldosterone-induced fibrosis in ScAT in patients with type 2 diabetes.

Some limitations need to be addressed in the current study. Since we investigated changes in ScAT, it is unknown whether our beneficial findings apply to other fat depots, especially visceral adipose tissue. ScAT is an important fat depot, considering the hypothesis that fibrosis in ScAT limits its ability to expand and affects adipocyte function, which could result in ectopic fat accumulation. Thus, reducing fibrosis in ScAT might positively affect adipocyte function and revert the dysfunctional state. A strength of the current study is the paired biopsies in 24 patients with type 2 diabetes before and after a randomized and double-blind, placebo-controlled intervention, providing novel data regarding the antifibrotic effect in human ScAT.

In conclusion, we observed that inhibition of the MR changes the function of adipose tissue by reducing fibrosis with the major collagen subunits in patients with type 2 diabetes. The present human data suggest a potential association with the downregulation molecules of the MR Gal-3 and NGAL in adipocytes. Future experimental studies should address the causality between the MR and regulation of fibrosis and the expression of Gal-3 and NGAL in adipose tissue.

Acknowledgments. The authors thank project nurse Tina Dreyer and biotechnicians Ulla Kjerulf, Helle Reinikka Christensen, and Syela Azemovski at the endocrine research laboratory at University Hospital of Herlev for qualified assistance.

Funding. This study was supported by Foundation de France grant 00086498; Instituto de Salud Carlos III-FEDER, Fondo de Investigaciones Sanitarias grant PI18/01875; the Danish Medical Association Research Foundation; the Danish Diabetes Academy; Danish Heart Foundation grant 15-R99-A5855; Miguel Servet contract CP13/00221; A.P. Møller og Hustru Chastine Mc-Kinney Møllers Fond for the promotion of medicine; and the Herlev Hospital Research Foundation.

Duality of Interest. F.J. reports grants and personal fees from AstraZeneca and Bayer. J.F. has served on scientific advisory boards at Merck Sharp & Dohme, AstraZeneca, Novo Nordisk, and Otsuka. P.R. reports grants and personal fees from AstraZeneca, Bayer, and CVRx; personal fees from Fresenius; grants and personal fees from Novartis; and personal fees from Grunenthal, Servier, Stealth Peptides, Vifor Fresenius Medical Care Renal Pharma, Idorsia, Novo Nordisk, Ablative Solutions, G3P, Corvidia, and Relypsa outside of the submitted work; P.R. is a cofounder of CardioRenal, a company that develops a telemonitoring loop in heart failure. C.K. has served on scientific advisory panels and received speaker fees from Boehringer Ingelheim, Merck Sharp & Dohme, AstraZeneca, Amgen, Novartis, Novo Nordisk, and Shire. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. M.L.J. collected and analyzed data, conducted the statistical analysis, interpreted data, wrote the manuscript, and approved the final version. J.I., A.F.-C., and P.R. analyzed and interpreted data, critically revised the manuscript, and approved the final version. M.S. and J.F. critically revised the manuscript and approved the final version. M.P.S., M.R.H., and J.R. collected data, critically revised the manuscript, and approved the final version. F.D. and F.J. analyzed data, critically revised the manuscript, and approved the final version. N.L.-A. analyzed data, conducted the statistical analysis, interpreted data, critically revised the manuscript, and approved the final version. C.K. initiated the study, interpreted data, wrote and revised the manuscript, and approved the final version. C.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.