Foetal calf serum (FCS) was obtained from HyClone (Shanghai, China). The antibodies and their commercial sources are listed below: Cell Signaling Technology (Beverly, MA): U0126 (#9903), anti-mitogen-activated protein kinase 1/2 (MEK1/2) (#9122), anti-phospho-MEK1/2 (#9154), anti-extracellular signal-regulated protein kinase 1/2 (ERK1/2) (#4695), anti-phospho-ERK1/2 (#4370), anti-c-Jun N-terminal kinase 1/2 (JNK1/2) (#9258), anti-phospho-JNK1/2 (#4668), anti-p38 (#9212), and anti-phospho-p38 (#4511); Santa Cruz Biotechnology, Inc.: anti-ANP (#sc20158) and anti-β-myosin heavy chain (β-MHC) (#sc53090); Aviva Systems Biology: anti-RGS14 (#OAAF04168); and Bioworld Technology: anti-GAPDH (#MB001). The bicinchoninic acid (BCA) protein assay kit was obtained from Pierce (Rockford, IL, USA). All other reagents, including the cell culture reagents, were purchased from Sigma.

The failing human heart samples were obtained from the left ventricle (LV) of dilated cardiomyopathy (DCM) patients after heart transplantation. The control samples were collected from the LV of normal heart donors who died because of an accident. The Institutional Review Board (IRB) of the Third Xiangya Hospital, Central South University approved the study. The relatives of the heart donors signed informed consent.

The Animal Care and Use Committee affiliated with the IRB of the Third Xiangya Hospital, Central South University approved all animal experimental protocols. All animals were housed in a light—(12 h light/12 h dark), temperature-controlled environment, and humidity-controlled environment. Food and water were available ad libitum. The animal models used in this study are described below.

Full-length mouse RGS14 Complementary DNA (cDNA) (OriGene, MC204443) was ligated into the chicken β-actin gene (CAG) promoter expression vector, which was linearized and purified using the QIAquick Gel Extraction Kit (Qiagen, 28704). This DNA construct was microinjected into fertilized mouse embryos (C57BL/6J background). Founder transgenic mice were identified by tail DNA amplification and then bred with C57BL/6J mice. Tail genomic DNA was identified using polymerase chain reaction (PCR). The following primers were used for the PCR amplification of the CAG gene promoter: forward, 5′-CCCCCTGAACCTGAAACATA-3′; reverse, 5′-CTGCGCTGAATTCCTTCTTC-3′. The expected size for the amplification product was 579 bp. The RGS14 flox mice were crossed with α-MHC-MerCreMer transgenic mice (Jackson Laboratory, 005650) to generate cardiac-specific RGS14-TG mice. Four independent transgenic lines were established. To induce RGS14 expression specifically in the heart, 6-week-old double transgenic mice were injected intraperitoneally with tamoxifen (80 mg/kg per day; Sigma, T-5648) for 5 consecutive days to cause Cre-mediated CAT gene excision. CAG-CAT-RGS14/MHC-Cre mice without tamoxifen administration (CRMC) served as the control group.

Directive sequences of the target site for the RGS14 gene in the mouse were predicted by the online CRISPR design system (http://​crispr.​mit.​edu) (Fig. 3a). A pair of oligomers (oligo1, TAGGGGCCTGGGAACCTGCAGTGC; oligo2, AAACGCACTGCAGGTTCCCAGGCC) was cloned into the BsaI restriction site of the pUC57-single guide RNA (sgRNA) expression vector (Addgene, 51132). DNA was amplified by PCR with primers spanning the T7 promoter and sgRNA regions (forward primer, GATCCCTAATACGACTCACTATAG; reverse primer, AAAAAAAGCACCGACTCGGT). The sgRNA was transcribed by the MEGAshortscript kit (Ambion, AM1354) and purified by the miRNeasy Micro kit (Qiagen, 217084). The Cas9 expression plasmid (Addgene 44758) was linearized with PmeI and used as the template for in vitro transcription using the T7 Ultra Kit (Ambion, AM1345). Cas9 and sgRNA mRNA injections of single-cell embryos were performed by the FemtoJet 5247 microinjection system. Genomic DNA was extracted, and a 405 bp DNA fragment overlapping the sgRNA target site was amplified by PCR with the following primers: RGS14-F, 5′-CTGTGTGGACACTCCCATCC-3′; and RGS14-R, 5′-ACCACAGAGAGAAGCAGCAC-3′. The purified PCR product was denatured and reannealed in NEB Buffer 2 to form heteroduplex DNA that was digested with T7EN (NEB, M0302L) for 45 min and analyzed by 3.0 % agarose gel (Fig. 3b). These mice were sequenced to select for frameshift mutations (Fig. 3c). The following primers were used to screen F1 and F2 offspring: RGS14-F, 5′-CTGTGTGGACACTCCCATCC-3′; and RGS14-R, 5′-ACCACAGAGAGAAGCAGCAC-3′. Finally, RGS14 knockout (RGS14-KO or RGS14 ^−/−) mice were generated and identified as shown in Fig. 3d, e. Littermate controls of the RGS14-KO mice were wild-type mice (WT or RGS14 ^+/+).

To obtain CaMEK1 flox mice, the coding sequence of mouse MEK1 S218D/S222D cDNA was ligated into the CAG promoter expression vector. This DNA construct was microinjected into fertilized mouse embryos (C57BL/6J background), and the resulting TG mice were PCR-genotyped using tail genomic DNA and the following primers: forward, 5′-CCCCCTGAACCTGAAACATA-3′; and reverse, 5′-CTGCGCTGAATTCCTTCTTC-3′. The expected size for the amplification product was 515 bp. The CaMEK1 flox mice were crossed with α-MHC-MerCreMer transgenic mice, which were obtained from the Jackson Laboratory (No. 005650), to generate CAG-MEK1/MHC-Cre mice. Tamoxifen (80 mg/kg per day; Sigma, T-5648) was then injected into the CAG-MEK1/MHC-Cre mice containing the CaMEK1 gene at 6 weeks of age for 5 consecutive days. CAG-MEK1/MHC-Cre mice without tamoxifen administration (CMMC) served as the control group. Finally, the CRMC mice were crossed with the CMMC mice and treated with tamoxifen to generate CaMEK1/RGS14 double transgenic (DTG) mice.

Aortic banding (AB) was performed as described previously [22, 37]. Briefly, mice were anesthetized using an intraperitoneal injection of sodium pentobarbital (50 mg/kg, Sigma) and ventilated with room air using a small animal ventilator (model VFA-23-BV, Kent Scientific, USA). The mice were kept warm on a heating pad until they regained consciousness. The left chest was opened after blunt dissection at the second intercostal space, and the thoracic aorta was identified. We tied the thoracic aorta to a 27 G or 26 G needle with a 7-0 silk suture depending on the body weight. The needle was removed quickly after the ligation, and the thoracic cavity was closed. Finally, the adequacy of aortic constriction was determined by the Doppler analysis. The mice in the control group were subjected to the same procedure without ligation of the aorta.

After the indicated times, the surviving mice were anesthetized using 1.5–2 % isoflurane and then subjected to echocardiography to examine cardiac function and structure, as previously described [21]. Briefly, a Mylab30CV ultrasound system switched to M-mode tracings with a 15 MHz probe was used to determine echocardiography. The LV end-diastolic dimension (LVEDd), LV end-systolic dimension (LVESd), and LV fractional shortening [FS (%) = (LVEDd-LVESd)/LVEDd × 100 %] were measured from the short axis of the LV at the level of the papillary muscles.

The animals were sacrificed 4–8 weeks after the AB or sham surgery. The hearts were harvested, arrested in diastole with 10 % potassium chloride solution, fixed with 10 % formalin, dehydrated, and embedded in paraffin. Paraffin-embedded hearts were cut transversely into 4–5 μm sections. Sections at the mid-papillary muscle level were stained with hematoxylin and eosin (H&E) and picrosirius red (PSR) to calculate the cardiomyocyte cross-sectional area (CSA) and collagen deposition volume, respectively. Fluorescein isothiocyanate-conjugated wheat germ agglutinin (WGA) was used to visualize the size of the cardiomyocytes. The immunofluorescence analysis was performed using the standard immunocytochemical techniques. Cardiomyocyte CSA, interstitial collagen deposition and perivascular collagen deposition were measured using the Image-Pro Plus 6.0 software.

Total mRNA was isolated from heart tissues or neonatal rat cardiomyocytes (NRCMs) using TRIzol reagent (Invitrogen). cDNA, which was obtained by reverse transcription of RNA, was synthesized using the Transcriptor First Strand cDNA Synthesis kit (Roche). Quantitative real-time PCR was performed using SYBR Green (Roche), and the relative expression of the target genes was calculated. GAPDH was measured and used for normalization. Cardiac tissue and cultured cardiomyocytes were lysed in RIPA lysis buffer, and the protein concentration was determined with a BCA protein assay kit. The proteins (50 μg) were resolved via SDS-PAGE (Invitrogen) and transferred to a PVDF membrane (Millipore), which was then subsequently blocked with milk. After overnight incubation with the indicated primary antibodies at 4 °C, the membranes were washed at least three times and then incubated with a secondary antibody for 1 h at room temperature. Finally, enhanced chemiluminescence-treated membranes were visualized using a FluorChem E imager (ProteinSimple, FluorChem E). The results were normalized to GAPDH.

The heart ventricles of 1- to 2-day-old Sprague–Dawley rats were enzymatically dissociated into individual cardiomyocytes in PBS containing 0.03 % trypsin and 0.04 % type II collagenase. Fibroblasts were then removed by a differential attachment technique, and the NRCMs were plated at a density of 1 × 10⁶ cells/well in six-well plates and cultivated in DMEM/F12 medium containing 20 % FCS, penicillin/streptomycin, and bromodeoxyuridine to inhibit fibroblast proliferation. The cardiomyocytes were maintained in serum-free DMEM/F12 for 12 h and then treated with angiotensin II (Ang II, 1 μM) for 24 or 48 h to induce hypertrophy.

To obtain cardiac fibroblasts, the adherent non-myocyte fractions obtained during pre-plating were grown in DMEM containing 10 % FCS to confluence and passaged with trypsin-EDTA. All experiments were performed on cells from the first or second passages. Cardiac fibroblasts were placed in DMEM medium containing 0.1 % FCS for 24 h before the stimulation by Ang II for 24 h, and the expression of RGS14 was investigated.

Finally, cardiomyocytes were infected with adenoviral RGS14 (AdRGS14) to overexpress RGS14, and an adenoviral vector encoding the green fluorescent protein gene (AdGFP) was infected into cardiomyocytes as a control group. Adenoviral short hairpin RGS14 (AdshRGS14) constructs were obtained and infected into cardiomyocytes to knockdown RGS14 expression, and an adenoviral short hairpin RNA (AdshRNA) was used as the non-targeting control. NRCMs were infected with different adenoviruses in diluted medium for 12 h.

U0126, an inhibitor of MEK1/2, was dissolved in dimethyl sulfoxide (DMSO) at a volume of 1 ml per 100 g of body weight and it was injected intraperitoneally into mice every 3 days (1 mg/kg) after AB. The control group was injected with a similar volume of DMSO.

The results are expressed as the mean ± standard deviation (SD). All data were analyzed using Student’s two-tailed t test and analysis of variance (ANOVA) to compare the means of two groups of samples and multiple groups, respectively. The T approximation test was used for the analysis when the sample was less than 7. All statistical analyses were performed with the SPSS software (version 17.0). P < 0.05 was considered statistically significant.

To investigate if the expression of RGS14 was altered during the process of pathological remodelling, we first measured RGS14 expression in human failing hearts, murine hypertrophic hearts, and NRCMs. According to the western blotting results, the RGS14 protein levels in human failing hearts were reduced to approximately 40 % of the levels in normal donor hearts. This decrease was accompanied by an enhancement of the foetal gene profile of ANP and β-MHC (Fig. 1a). Similar results were found in murine hypertrophic hearts and NRCMs. As shown in Fig. 1b, the expression levels of RGS14 were approximately two-fold and four-fold lower in the experimental mouse hearts after 4 or 8 weeks of AB, respectively, compared with the sham-operated control group. In addition, the expression levels of ANP and β-MHC were dramatically increased at week 4 and more pronounced at week 8 (Fig. 1b). Furthermore, stimulation of NRCMs with Ang II (1 μmol/L) for 24 or 48 h led to RGS14 downregulation and β-MHC and ANP up-regulation (Fig. 1c). The expression of RGS14 was not significantly changed in response to AngII stimulation for 24 h in the isolated cardiac fibroblasts (Figure S1). These results indicated that RGS14 in cardiomyocytes was markedly altered by hypertrophic stress in vivo and in vitro, which demonstrated that RGS14 might be involved in cardiac remodelling.

To explore the functional contribution of RGS14 to the hypertrophy of cardiomyocytes in vitro, we first performed studies using primary cultured NRCMs infected with AdshRGS14, AdRGS14, or control vectors, and further stimulated with Ang II (1 μmol/L) for 48 h. The protein expression of RGS14 is shown in Fig. 2a and S2A. Immunostaining with the α-actin antibody suggested that neither the NRCMs infected with AdRGS14 nor the NRCMs infected with AdshRGS14 exhibited altered cell size compared with the NRCMs infected with AdGFP or AdshRNA under basal conditions. However, AdshRGS14 treatment enhanced the Ang II-induced increase in cell size (Fig. 2b, c). Furthermore, the mRNA levels of ANP and β-MHC were approximately 1.5- and 2.4-fold higher, respectively, in the NRCMs infected with AdshRGS14 after exposure to Ang II than in the controls (Fig. 2e). The cells with overexpression of RGS14 (AdRGS14) exhibited significantly reduced cell-surface areas (approximately 40 %) compared with the AdGFP-infected cells, and the mRNA expression of ANP and β-MHC was significantly decreased (approximately 25 %) in AdRGS14-infected cells compared with the cells infected with AdGFP (Fig. 2b, d, f). Together, these observations indicate that RGS14 protected against the hypertrophic response in cardiomyocytes.

The potential function of RGS14 during cardiac remodelling in vivo was investigated. RGS14-KO mice were generated using the CRISPR/Cas9 methods (Fig. 3a–c). Gene sequencing and western blotting of RGS14 expression in heart tissue from RGS14-KO and littermate control mice were performed (Fig. 3d, e). At baseline, the RGS14-KO mice displayed normal cardiac morphology and contractile function (Table S1). The levels of RGS2, 3, 4, and 5 were not significantly changed in the RGS14-KO mice compared with the WT mice (Figure S3). As shown in Fig. 4a, aortic banding induced a 58 % increase in the ratio of heart weight to body weight (HW/BW), indicating the development of cardiac hypertrophy in wild-type mice. The RGS14 ^−/− mice exhibited a significantly aggravated hypertrophic effect with an increase of approximately 27 % in the ratio of HW/BW compared with the WT mice subjected to AB surgery. Similar effects were observed in the ratio of lung weight to body weight (LW/BW) and in the ratio of heart weight to tibia length (HW/TL) (Fig. 4b, c). No comparable differences were observed in the sham-treated RGS14 ^−/− and the WT mice (Fig. 4a–c). Cardiac function was also measured by echocardiography. The parameters of LVEDd, LVESd, and FS% indicated that myocardial contraction in the AB-treated RGS14 ^−/− mice was reduced compared with the WT group subjected to AB (Fig. 4d–f). H&E and WGA staining showed a greater ventricular CSA in the RGS14 ^−/− mice than in the control mice subjected to AB surgery (Fig. 4g, h). Because fibrosis is a classical feature of pathological cardiac remodelling and is characterized by the accumulation of collagen in the heart [48], we evaluated the effects of RGS14 deletion on cardiac fibrosis in pressure-overloaded hearts. Fibrosis was determined by visualizing the extent of collagen staining and calculating the total collagen volume. Both perivascular and interstitial fibrosis analyses consistently demonstrated an increased fibrotic response in the AB-treated RGS14 ^−/− mice compared with the AB-treated WT mice (Fig. 4g, i). We measured the synthesis of collagen by analyzing the mRNA expression of hypertrophic markers (ANP, BNP, and β-MHC) and fibrotic markers (collagen I, collagen III, and fibronectin) (Fig. 4j). Our results consistently revealed an increased fibrotic response in RGS14 ^−/− hearts. Collectively, these findings reveal that ablation of RGS14 exacerbates hypertrophy and fibrosis in response to chronic pressure overload.

To further confirm the protective effect of RGS14 on cardiac remodelling, cardiac-specific transgenic mice overexpressing murine RGS14 under the control of the CAG promoter were generated, and four independent RGS14-TG mice were generated (Fig. 5a). Cardiac RGS14 expression in these TG mice was approximately two- to eight-fold higher than that in their CRMC littermates. TG line 2, carrying the highest levels of RGS14 expression, was selected for further research. At baseline, the RGS14-TG2 mice displayed normal cardiac morphology and contractile function (Table S1). The levels of RGS2, 3, 4, and 5 were not significantly changed in the RGS14-TG2 mice compared with the CRMC mice (Figure S3). A morphological disparity occurred when comparing RGS14-TG2 to CRMC mice 4 weeks after AB. As shown in Fig. 5b, AB induced a 40 % increase in the HW/BW compared with the sham control, suggesting the development of cardiac hypertrophy in the CRMC mice, whereas the TG2 mice exhibited a significant protective effect against hypertrophy, with a decrease of 23 % in the HW/BW. A similar protective effect against hypertrophy was observed in the LW/BW and HW/TL. In contrast to the CRMC mice, the TG2 mice exhibited significant decreases in the LW/BW and HW/TL (Fig. 5c, d). These changes were consistent with the protective role of RGS14 in cardiac function in response to hypertrophic stimulation, as measured by cardiac function parameters (LVEDd, LVESd and FS%) by echocardiography (Fig. 5e–g). Compared with the CRMC mice, the RGS14-TG2 mice had smaller heart and cardiomyocyte sizes (Fig. 5h, i) and reduced fibrosis volumes (Fig. 5h, j). Furthermore, AB increased the mRNA levels of ANP, BNP, β-MHC, collagen I, collagen III, and fibronectin in the hearts of the CRMC mice, and these expression levels were significantly suppressed in the RGS14-TG2 mice (Fig. 5k). We also examined the hypertrophic response of RGS14-TG1 mice (RGS14 expression is 2.5-fold higher than that of CRMC) to evaluate the relevance of RGS14 expression in cardiac hypertrophy. As shown in Figure S4, the heart weight/body weight ratio and the cross-sectional area were slightly, but significantly reduced in the RGS14-TG1 group compared with the CRMC controls, 4 weeks after AB surgery. The protective effect of RGS14 on pressure overload-induced cardiac hypertrophy in the RGS14-TG1 mice was significantly weaker than in the RGS14-TG2 mice, suggesting a possible gene doses effect. Together, these results indicate that overexpression of RGS14 might suppress cardiac remodelling induced by AB in vivo.

To determine the underlying mechanism of the anti-hypertrophy effect of RGS14, the expression and activity of MAPK signalling molecules were detected. RGS14-TG2 mice were selected for the current research. As shown in Fig. 6a, b, the phosphorylation levels of MEK1/2, ERK1/2, JNK1/2, and p38 were significantly elevated after AB surgery, whereas the total protein expression remained unchanged. Deletion of RGS14 further increased the activation of MEK1/2 and ERK1/2 by 1.6- and 2.6-fold, respectively, compared with the WT mouse hearts subjected to AB. Overexpression of RGS14 restored the phosphorylation level of MEK1/2 and ERK1/2 to approximately normal. Neither p38 nor JNK phosphorylation was altered in the AB-induced RGS14 ^−/− and RGS14-TG hearts.

To exclude potential compensatory mechanisms in vivo, further NRCM experiments were conducted. AdshRGS14 and AdRGS14 were used to knock down or overexpress RGS14 expression in NRCMs, respectively. The protein expression of RGS14 in different groups is shown in Figure S2. The expression and activity of MAPK signalling molecules were detected compared with AdshRNA and AdGFP controls. Western blot analysis showed that RGS14 down-regulation enhanced the expression of phosphorylated MEK1/2 and ERK1/2 compared with AdshRNA-infected cells under Ang II treatment (Fig. 6c), whereas RGS14 up-regulation strongly suppressed the levels of MEK1/2 and ERK1/2 phosphorylation in AdRGS14-infected NRCMs compared with the AdGFP-infected group (Fig. 6d). Our results demonstrate that the RGS14-elicited anti-hypertrophic effect is largely associated with the inhibition of MEK–ERK1/2 signalling in hearts.

To determine if the MEK–ERK1/2 pathway plays an essential role in RGS14-induced protection in AB-induced cardiac hypertrophy, U0126 (an inhibitor of MEK) was infused into WT and RGS14-KO mice before AB surgery. Western blot analysis indicated lower levels of phosphorylated MEK1/2 and ERK1/2 in U0126-treated RGS14-KO mice compared with control mice (Fig. 7a). As shown in Fig. 7b–j, U0126 significantly restricted the deteriorative cardiac remodelling in RGS14-KO mice in response to AB. In U0126-treated RGS14-KO mice, the HW/BW, LW/BW, and HW/TL ratios were decreased compared with the DMSO-control group (Fig. 7b–d), and heart function according to LVEDd, LVESd, and FS% values was improved compared with the DMSO-control group (Fig. 7e–g). Moreover, smaller cross-sectional areas of cardiomyocytes (Fig. 7h, i) and lower collagen volumes (Fig. 7h, j) were observed in the RGS14-KO mice treated with U0126 compared with the DMSO-treated controls. These parameters were equal in both the U0126-treated RGS14-KO mice and the WT mice (Fig. 7b–j). These findings suggest that pre-inhibition of MEK–ERK1/2 signalling protects against AB-induced cardiac remodelling in RGS14-KO mice.

Transgenic mice with floxed CaMEK1 were crossed with transgenic αMHC-MerCre-Mer mice to generate cardiac-specific CaMEK1 transgenic mice (CaMEK1-TG), and transgenes were identified by the western blot analysis (Fig. 8a, b). The level of MEK was significantly increased in the CaMEK1-TG mice compared with the CMMC group. RGS14/CaMEK1 DTG mice were generated by crossing CRMC mice with CMMC mice and treating the mice with tamoxifen (Fig. 8c). At baseline, the CaMEK1-TG and DTG mice displayed normal cardiac morphology and contractile function (Table S1). The RGS14 and phospho-ERK levels were determined in the CaMEK1/RGS14 DTG mice, as shown in Figure S5. As expected, the overexpression of CaMEK1 produced more AB-induced cardiac remodelling compared with the CRMC group (Fig. 8d–l). Four weeks after AB surgery, the HW/BW, LW/BW, and HW/TL ratios were significantly increased in the DTG group compared with the RGS14-TG group (Fig. 8d–f). In addition, the LVEDd, LVESd, and FS% values according to the echocardiograph indicated deteriorated cardiac function in the DTG group compared with the RGS14-TG group (Fig. 8g–i). Moreover, the heart areas, cardiomyocyte cross-sectional areas, and cardiac fibrosis volumes were significantly up-regulated in the DTG mice, as shown in Fig. 8j–l, compared with the RGS14-TG mice. These parameters were equal in the DTG mice and the CaMEK1 mice (Fig. 8d–l). Our results demonstrate that targeted MEK1 activation abolishes the protective effects of RGS14 on cardiac remodelling after AB surgery. Therefore, our findings indicate that the protective role of RGS14 in pathological cardiac remodelling is at least in part due to the inhibition of MEK1 signalling.