Serelaxin alleviates cardiac fibrosis through inhibiting endothelial-to-mesenchymal transition via RXFP1

Abstract

Rationale: Cardiac fibrosis is an integral constituent of every form of chronic heart disease, and persistence of fibrosis reduces tissue compliance and accelerates the progression to heart failure. Relaxin-2 is a human hormone, which has various physiological functions such as mediating renal vasodilation in pregnancy. Its recombinant form Serelaxin has recently been tested in clinical trials as a therapy for acute heart failure but did not meet its primary endpoints. The aim of this study is to examine whether Serelaxin has an anti-fibrotic effect in the heart and therefore could be beneficial in chronic heart failure. Methods: We utilized two different cardiac fibrosis mouse models (ascending aortic constriction (AAC) and Angiotensin II (ATII) administration via osmotic minipumps) to assess the anti-fibrotic potential of Serelaxin. Histological analysis, immunofluorescence staining and molecular analysis were performed to assess the fibrosis level and indicate endothelial cells which are undergoing EndMT. In vitro TGFβ1-induced endothelial-to-mesenchymal transition (EndMT) assays were performed in human coronary artery endothelial cells and mouse cardiac endothelial cells (MCECs) and were examined using molecular methods. Chromatin immunoprecipitation-qPCR assay was utilized to identify the Serelaxin effect on chromatin remodeling in the Rxfp1 promoter region in MCECs. Results: Our results demonstrate a significant and dose-dependent anti-fibrotic effect of Serelaxin in the heart in both models. We further show that Serelaxin mediates this effect, at least in part, through inhibition of EndMT through the endothelial Relaxin family peptide receptor 1 (RXFP1). We further demonstrate that Serelaxin administration is able to increase its own receptor expression (RXFP1) through epigenetic regulation in form of histone modifications by attenuating TGFβ-pSMAD2/3 signaling in endothelial cells. Conclusions: This study is the first to identify that Serelaxin increases the expression of its own receptor RXFP1 and that this mediates the inhibition of EndMT and cardiac fibrosis, suggesting that Serelaxin may have a beneficial effect as anti-fibrotic therapy in chronic heart failure.

Keywords: EndMT; Notch; Serelaxin; fibrosis; histone methylation.

Figure 1

Serelaxin ameliorates cardiac fibrosis. (A) HE and Masson's Trichrome Staining microphotography of sham and AAC-operated mouse heart showing the reduction of interstitial and perivascular fibrosis in Serelaxin-treated (500 µg/kg/day) hearts compared to sham and vehicle-treated hearts. (B-D) Graphs showing a significant reduction of fibrosis in Serelaxin-treated hearts. The dot plots represent the percentage of overall, interstitial and perivascular fibrotic area in the sham (n=6), vehicle-treated (n=10) and low (n=4), middle (n=6) and high dose (n=13) Serelaxin-treated AAC mouse hearts. (E) Kaplan-Meier survival curve summarizes survival rates of mice from the AAC operation date until 4 weeks after operation in vehicle-treated (n=28), low (13), middle (13) and high dose Serelaxin-treated (n=29) groups. Administration of high dose Serelaxin substantially increased survival. (F)-(G) Bar graphs showing heart weight related to body weight and tibia length. AAC-operation increased the heart weight compared to sham group, while administration of high but not low and middle dose Serelaxin prevented the increase in heart weight. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Overall survival was analyzed by using a Kaplan-Meier survival method with a log rank test to determine statistical differences. Error bars represent mean ± SEM, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001.

Figure 2

Serelaxin blocks EndMT in a mouse model of pressure overload. (A) Immunofluorescence staining of endothelial cell marker CD31, mesenchymal marker α-SMA and DAPI. (B) Quantifications of CD31 and α-SMA positive area and ratio of CD31⁺α-SMA⁺ / CD31⁺ cells. Compared to sham, vehicle-treated hearts showed an increased expression of α-SMA but a decreased protein expression of CD31, while treatment with Serelaxin showed reduced α-SMA and restored CD31 protein expression. Double positive cells significantly increased in AAC-operated animals and were reduced by Serelaxin administration. (C) Immunofluorescence staining of endothelial cell marker CD31, fibrotic marker alpha-1 type I collagen (Col1α(I)) and DAPI. (D) Quantifications of CD31 and Col1α(I) positive area: Col1α(I) expression was upregulated in AAC hearts and is inhibited by Serelaxin. Compared to sham, vehicle-treated hearts showed a decreased protein expression of CD31, while treatment with Serelaxin again restored CD31 expression. (E) qPCR analysis showing the relative mRNA expression level of CD31 and EndMT key regulators Snail, Slug, and Twist in AAC hearts treated with vehicle or Serelaxin. Vehicle-treated AAC-operated hearts showed an increased expression of Snail, Slug and Twist but a decreased expression of CD31, while Serelaxin treatment reduced Snail, Slug and Twist expression and restored CD31 expression. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars represent mean ± SEM, n≥3, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001.

Figure 3

Serelaxin partially inhibits TGFβ1-induced EndMT in human coronary artery endothelial cells (HCAECs) and mouse cardiac endothelial cells (MCECs) via RXFP1. (A) qPCR analysis showing the expression of endothelial cells marker CD31 and expression of EndMT key regulators SNAIL, SLUG, and TWIST in TGFβ1-treated HCAECs supplemented with different doses of Serelaxin after 4 days. Cells without any treatment were used as control. Serelaxin treatment significantly rescued expression of CD31 (100 and 200 ng/ml) and decreased expression of SNAIL, SLUG and TWIST (100 and 200 ng/ml). (B) qPCR analysis showing the mRNA expression of EndMT transcriptional factors Snail, Slug and Twist and of endothelial cell marker CD31 in TGFβ1-treated MCECs. Upon TGFβ1 treatment Snail, Slug and Twist expression increased and CD31 expression decreased. Treatment of Serelaxin showed a reduced expression of Snail, Slug and Twist and a restored expression of CD31. (C) qPCR analysis showing the relative mRNA expression level of RXFP1-4 in HCAECs. Among all four genes, only RXFP1 and RXFP4 expression was detectable. (D) qPCR analysis showing the expression of EndMT key regulators SNAIL, SLUG and TWIST in TGFβ1 and Serelaxin-treated HCAECs in combination with knockdown of RXFP1 or RXFP4. Cells without any treatment were used as control. Serelaxin treatment showed a reversal effect on TGFβ1-induced EndMT but not in RXFP1 knockdown cells. (E) qPCR analysis showing the relative mRNA expression level of Rxfp1-4 in MCECs. Among all 4 genes, only Rxfp1 expression was detectable but not the others. (F) qPCR analysis showing the relative mRNA level of Rxfp1 in scrambled and siRNA-mediated Rxfp1 knockdown cells. (G) qPCR analysis showing the expression of EndMT key regulators Snail, Slug and Twist in TGFβ1 and Serelaxin-treated MCECs upon Rxfp1 knockdown. Cells without any treatment were used as control. Serelaxin treatment showed a reversal effect on TGFβ1-induced EndMT but not upon Rxfp1 knockdown. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars represent mean ± SEM, n≥3, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001.

Figure 4

Serelaxin inhibits EndMT via RXFP1 in vitro and in vivo. (A) Immunofluorescence staining of WGA, RXFP1 and DAPI in sham and AAC-operated hearts treated with vehicle or high dose Serelaxin. (B) Quantification of RXFP1 and WGA positive area showed a decreased RXFP1 and increased WGA protein expression in AAC-operated hearts compared to sham. Upon treatment with Serelaxin, RXFP1 expression was restored and WGA expression decreased. (C) Immunofluorescence staining of endothelial cell marker CD31, RXFP1 and DAPI. (D) Quantifications of CD31 (left panel) and RXFP1 (right panel) positive area. Compared to sham, AAC-operated vehicle-treated hearts showed a decreased expression of RXFP1 and CD31, while treatment with Serelaxin showed increased protein expressions of both CD31 and RXFP1. Each bar shows both the double positive area (streak lines) as well as single positive area (gray). The information on significance refers to the total CD31 (left panel) or RXFP1 (right panel) positive area. (E) Western blot analysis showing protein levels of RXFP1 in sham, AAC-operated and AAC-operated+Serelaxin-treated mouse hearts. Compared to sham, AAC operation reduced RXFP1 expression. Upon administration of Serelaxin, the protein level of RXFP1 was increased. (F) qPCR analysis showing the relative mRNA expression level of Rxfp1 in sham and AAC-operated hearts treated with vehicle or Serelaxin. Rxfp1 expression was decreased in AAC-operated hearts but restored upon Serelaxin treatment. (G) qPCR analysis showing the relative mRNA expression level of RXFP1-4 in human hearts. RXFP1 and RXFP4 expression were detected whereas RXFP1 was mainly expressed. (H) qPCR analysis showing a reduced relative mRNA expression level of RXFP1 in diseased human hearts compared to healthy control hearts. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars, representing mean ± SEM, n≥3, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001.

Figure 5

Serelaxin promotes the preservation of endothelial cell properties in both in vivo and in vitro by regulating candidate genes. (A) Heat map and (B) scatter plot show the genes with altered mRNA expression level in only TGFβ1-treated (x-axis) or in both TGFβ1- and Serelaxin-treated (y-axis) HCAECs. STEAP1, NOTCH1, JAGGED1, RAC1, DSP, GSC, ERBB3, FZD7, GSK3B and TGFB2 were significantly regulated by Serelaxin treatment, which is shown by separation from dot lines (cut-off by 4 folds). (C-D) qPCR analysis showing the expression of candidate genes, which were identified from qPCR array in TGFβ1- and Serelaxin-treated MCECs. Cells without any treatment were used as control. (E) qPCR analysis of the panel of candidate genes in vivo in sham and AAC-operated mice treated with vehicle or Serelaxin. Serelaxin significantly rescued the effects of AAC on all genes except for Gsc and Steap1. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars, representing mean ± SEM n≥3, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001.

Figure 6

Serelaxin rescues the Notch1 pathway in AAC-operated hearts. Immunofluorescence staining of (A) Jagged1, (B) Notch1 and (C) NICD in combination with WGA and DAPI in sham and AAC-operated hearts treated with vehicle or Serelaxin. Quantifications showed that protein expressions of (D) Jagged1, (E) Notch1 and (F) NICD were downregulated in AAC-operated hearts and could be restored by Serelaxin. (G) Western blot analysis shows the amount of the soluble form of Jagged1 in the medium of TGFβ1-treated MCECs. Ponceau-S stained membrane picture indicates that an equal amount of total precipitated protein was loaded for both the control and TGFβ1-treated MCECs. Soluble Jagged1 was reduced in TGFβ1-treated MCECs as quantified by densitometry analysis (right panel). Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars, representing mean ± SEM, n≥3, n.s. no significance, * p<0.05, ** p<0.01, *** p<0.001.

Figure 7

Serelaxin reactivates Rxfp1 expression via histone modifications. (A) Schematic representing the mouse Rxfp1 locus with locations of CHIP-qPCR primers. (B) qPCR analysis showing the enrichment of 4 different histone modification marks (active: H3K4me3 and H3K27ac, repressive: H3K9me3, H3K27me3) in the Rxfp1 promoter region (Amplicon 1-3) and intron 10 (Amplicon 4). The enrichment of TGFβ1-influenced H3K4me3, H3K9me3 and H3K27me3 marks were significantly compromised by Serelaxin if performed with amplicons 1-3 targeting the promoter region but not with amplicon 4 targeting intron10. (C) ChIP-qPCR analysis showing the enrichment of histone modification marks in the Rxfp1 promoter region after RXFP1 knockdown. Upon RXFP1 knockdown, treatment with Serelaxin did not significantly increase the activating modifications or decrease the repressive modification marks. (D) Western blot analysis showing protein levels of pSMAD2 and pSMAD3 in sham, AAC-operated and AAC-operated+Serelaxin-treated mouse hearts, total SMAD2 and SMAD3 were used as protein loading controls. AAC-operated hearts showed increased levels of pSMAD2 and pSMAD3 compared to sham. Upon treatment with Serelaxin these protein levels were significantly reduced. (E) qPCR analysis showing the enrichment of pSMAD2/3 in the Rxfp1 promoter region (Amplicon 2) and intron 10 (Amplicon 4). The enrichment of TGFβ1-induced pSMAD2/3 was significantly compromised by Serelaxin. The qPCR analysis with primer targeting intron10 did not show any significant differences between these groups. Student t-test was used for single comparison and one-way ANOVA with Bonferroni post-hoc analysis was used for multiple group comparisons. Gene expression and associated error bars, representing mean ± SEM, n≥3, n.s. no significance, ** p<0.01, *** p<0.001, **** p<0.0001.

Figure 8

A proposed model depicting the molecular mechanism by which Serelaxin activates Notch signaling pathway to inhibit TGFβ-induced SMAD2/3 phosphorylation to restore the RXFP1 expression via histone modifications.