Rosmarinic acid attenuates cardiac fibrosis following long-term pressure overload via AMPKα/Smad3 signaling

Abstract

Agonists of peroxisome proliferator-activated receptor gamma (PPAR-γ) can activate 5' AMP-activated protein kinase alpha (AMPKα) and exert cardioprotective effects. A previous study has demonstrated that rosmarinic acid (RA) can activate PPAR-γ, but its effect on cardiac remodeling remains largely unknown. Our study aimed to investigate the effect of RA on cardiac remodeling and to clarify the underlying mechanism. Mice were subjected to aortic banding to generate pressure overload induced cardiac remodeling and then were orally administered RA (100 mg/kg/day) for 7 weeks beginning 1 week after surgery. The morphological examination, echocardiography, and molecular markers were used to evaluate the effects of RA. To ascertain whether the beneficial effect of RA on cardiac fibrosis was mediated by AMPKα, AMPKα2 knockout mice were used. Neonatal rat cardiomyocytes and fibroblasts were separated and cultured to validate the protective effect of RA in vitro. RA-treated mice exhibited a similar hypertrophic response as mice without RA treatment, but had an attenuated fibrotic response and improved cardiac function after pressure overload. Activated AMPKα was essential for the anti-fibrotic effect of RA via inhibiting the phosphorylation and nuclear translocation of Smad3 in vivo and in vitro, and AMPKα deficiency abolished RA-mediated protective effects. Small interfering RNA against Ppar-γ (siPpar-γ) and GW9662, a specific antagonist of PPAR-γ, abolished RA-mediated AMPKα phosphorylation and alleviation of fibrotic response in vitro. RA attenuated cardiac fibrosis following long-term pressure overload via AMPKα/Smad3 signaling and PPAR-γ was required for the activation of AMPKα. RA might be a promising therapeutic agent against cardiac fibrosis.

Fig. 1. Rosmarinic acid (RA) attenuated cardiac dysfunction in mice following long-term pressure overload

a The maximum carotid pressure of mice in the four groups (n = 13–14). b Fractional shortening (FS) of mice as determined via echocardiography at 8 weeks after the AB surgery (n = 15). c, d Hemodynamic analysis of mice with or without RA protection (n = 13–14). e, f Statistical results of the heart weight (HW)/body weight (BW) and HW/tibia length (TL) (n = 15). g, h Interventricular septal thickness at systole or diastole (IVSs or IVSd) (n = 15). i, j HE staining and statistical results of the cross sectional area (n = 6). k The relative mRNA level of Bnp normalized to Gapdh in mice (n = 6). Values represent the mean ± SEM. *P < 0.05 vs. the corresponding Sham group, ^#P < 0.05 vs. AB+Veh

Fig. 2. RA protected against cardiac fibrosis in vivo

a Representative images of PSR staining and immunohistochemical images of α-SMA (n = 6). b Statistical results of average collagen volume and α-SMA density (n = 6). c-h The relative mRNA levels of fibrotic markers normalized to Gapdh in mice (n = 6). Values represent the mean ± SEM. *P < 0.05 vs. the corresponding Sham group, ^#P < 0.05 vs. AB+Veh

Fig. 3. RAactivated AMPKα in vivo and knockout of AMPKα abrogated its protective effects

a-f Representative western blots and quantitative results in wild type (WT) mice (n = 6). g, h PSR staining and statistical results of average collagen volume in WT and AMPKα2 knockout (KO) hearts (n = 6). i FS in WT and KO mice (n = 8–10). Values represent the mean ± SEM. *P < 0.05 vs. the corresponding Sham group within WT mice, ^#P < 0.05 vs. AB+Veh in WT mice, ^&P < 0.05 vs. the corresponding Sham group in KO mice

Fig. 4. RA suppressed Smad3 phosphorylation and nuclear translocation through AMPKα activation

a-c Representative western blots and statistical results (n = 6). Values represent the mean ± SEM. *P < 0.05 vs. the corresponding Sham group within WT mice, ^#P < 0.05 vs. AB+Veh in WT mice, ^&P < 0.05 vs. the corresponding Sham group in KO mice

Fig. 5. RA blocked transdifferentiation of neonatal rat cardiac fibroblasts (CFs) via AMPKα in vitro

a, b The relative mRNA levels of collagen I (Col I) and collagen III (Col III) in CFs (n = 3). c Statistical results on cell viability evaluated by cell count kit 8 (n = 5). d Representative western blots and statistical results (n = 6). e, f Representative immunofluorescence staining of α-SMA in the presence or absence of shAmpkα2 and statistical results (n = 6). Green represented α-SMA, nuclei was stained with DAPI (blue). g Representative images of wound scratch assay at 0 and 12 h (n = 3). Values represent the mean ± SEM of three independent experiments. *P < 0.05 vs. the corresponding PBS-treated group within shRNA-treated CFs, ^#P < 0.05 vs. TGF-β-treated CFs within shRNA group, ^&P < 0.05 vs. the corresponding PBS-treated CFs in shAmpkα2 group. In Figs. 5a, b and 5d, *P < 0.05 vs. the matched group, NS no significance

Fig. 6. RA suppressed Smad3 phosphorylation and nuclear translocation through activating AMPK in vitro

a, b Representative western blots and statistical results. CFs were pretreated with shAmpkα2 or shRNA for 4 h at a MOI of 50, incubated with RA for 30 min and then stimulated with TGF-β (10 ng/ml) for additional 1 h (n = 6). c Characteristic immunofluorescence images of T-Smad3 distribution (n = 6). CFs were pretreated with RA for 30 min and then stimulated with TGF-β (10 ng/ml) for additional 1 h. Green represented T-Smad3, nuclei was stained with DAPI (blue). Values represent the mean ± SEM of three independent experiments. *P < 0.05 vs. the corresponding PBS-treated CFs within shRNA group, ^#P < 0.05 vs. TGF-β-treated CFs within shRNA group, ^&P < 0.05 vs. TGF-β-treated CFs in shAmpkα2 group

Fig. 7. RAactivated AMPKα via the activation of PPAR-γ

a The expression of PPAR-γ and the quantitative results in vivo (n = 6). b Representative western blots and statistical results of PPAR-γ in CFs (n = 6). c Activation of AMPKα was blocked in the presence of GW9662 in CFs (n = 6). d The relative mRNA level of α-Sma normalized to Gapdh in the presence or absence of GW9662 in CFs (n = 6). e, f Representative immunofluorescence staining of α-SMA and statistical results (n = 6). Green represented α-SMA, nuclei was stained with DAPI (blue). g Activation of AMPKα was blocked after knockdown of PPAR-γ with siPpar-γ in CFs (n = 6). h The relative mRNA level of α-Sma normalized to Gapdh with or without siPpar-γ treatment in CFs (n = 6).Values represent the mean ± SEM. Representative images are shown in b-h from three independent experiments. *P < 0.05 vs. the corresponding control CFs within vehicle or siRNA group, ^#P < 0.05 vs. TGF-β-treated CFs within vehicle or shRNA group, ^&P < 0.05 vs. the corresponding control CFs within GW9662 or siPpar-γ group. In Figs. 7a, *P < 0.05 vs. the corresponding Sham group, ^#P < 0.05 vs. AB+Veh. In Figs. 7b, *P < 0.05 vs. the matched group, NS no significance