Protocatechuic acid attenuates isoproterenol-induced cardiac hypertrophy via downregulation of ROCK1–Sp1–PKCγ axis

Cardiac hypertrophy is an adaptive response of the myocardium to pressure overload or adrenergic agonists. Here, we investigated the protective effects and the regulatory mechanism of protocatechuic acid, a phenolic compound, using a mouse model of isoproterenol-induced cardiac hypertrophy. Our results demonstrated that protocatechuic acid treatment significantly downregulated the expression of cardiac hypertrophic markers (Nppa, Nppb, and Myh7), cardiomyocyte size, heart weight to body weight ratio, cross-sectional area, and thickness of left ventricular septum and posterior wall. This treatment also reduced the expression of isoproterenol-induced ROCK1, Sp1, and PKCγ both in vivo and in vitro. To investigate the mechanism, we performed knockdown and overexpression experiments. The knockdown of ROCK1, Sp1, or PKCγ decreased the isoproterenol-induced cell area and the expression of hypertrophic markers, while the overexpression of Sp1 or PKCγ increased the levels of hypertrophic markers. Protocatechuic acid treatment reversed these effects. Interestingly, the overexpression of Sp1 increased cell area and induced PKCγ expression. Furthermore, experiments using transcription inhibitor actinomycin D showed that ROCK1 and Sp1 suppression by protocatechuic acid was not regulated at the transcriptional level. Our results indicate that protocatechuic acid acts via the ROCK1/Sp1/PKCγ axis and therefore has promising therapeutic potential as a treatment for cardiac hypertrophy.


Isoproterenol-induced cardiomyocyte hypertrophy was reversed by ROCK1 downregulation.
To elucidate the mechanism of protocatechuic acid function, we decided to focus on ROCK1, the kinase involved in the regulation of cell shape in cardiac hypertrophy 4,33 . Protocatechuic acid treatment reduced ROCK1 protein expression levels in the heart tissues of isoproterenol-treated mice (Fig. 3A,B) and reversed the isoproterenol-induced Rock1 mRNA and protein expression levels in H9c2 cells (Fig. 3C − E). To investigate the role of ROCK1 in cardiac hypertrophy, we transfected H9c2 cells with ROCK1 siRNA. As expected, ROCK1 siRNA reduced the levels of endogenous Rock1 mRNA (Fig. 3F). Furthermore, both mRNA and protein expression levels of isoproterenol-induced Rock1, Nppa (ANP), and Nppb (BNP) were also significantly decreased by ROCK1 siRNA (Fig. 3G,H, Supplementary Fig. 2).
Next, we investigated the effect of ROCK1 downregulation on cardiomyocyte hypertrophy. In the absence of isoproterenol stimulation, there were no significant differences in cell size between control and ROCK1 siRNA groups ( Supplementary Fig. 3); however, ROCK1 knockdown significantly decreased the isoproterenol-induced cell hypertrophy (Fig. 3I,J).
Knockdown of Sp1 reduced the isoproterenol-induced cardiomyocyte size and the expression of hypertrophic markers. It has been reported that the expression of cardiac hypertrophy marker ANP is regulated by transcription factor Sp1 23,34 . Therefore, we evaluated the effect of protocatechuic acid treatment on Sp1 expression in vivo and in vitro. As expected, the isoproterenol stimulation increased the expression of Sp1 mRNA; however, this effect was decreased by protocatechuic acid treatment both in H9c2 cells and in mouse heart tissues (Fig. 4A,B). Western blot analysis of mouse heart tissues showed similar results (Fig. 4C,D) www.nature.com/scientificreports/ evaluate the effect of Sp1 knockdown on cardiac hypertrophy, H9c2 cells were transfected with Sp1 siRNA and then treated with isoproterenol. The knockdown of Sp1 successfully reduced the mRNA levels of Sp1 (Fig. 4E), as well as Nppa and Nppb, in both control and isoproterenol-treated groups (Fig. 4F,G). In order to investigate whether the silencing of Sp1 affected the cell size, cells were stained with Alexa Fluor 488 phalloidin. In the    www.nature.com/scientificreports/ absence of isoproterenol stimulation, Sp1 knockdown did not have any effect on the cell size ( Fig. 4H,I); however, in response to the stimulation, it completely reversed isoproterenol-induced cell hypertrophy (Fig. 4H,I).

Protocatechuic acid reduced Sp1 overexpression-induced cardiac hypertrophy.
To decipher the role of Sp1 in cardiac hypertrophy, we transfected H9c2 cells with pCMV-Sp1. As expected, the overexpression of Sp1 significantly enhanced mouse Sp1 mRNA levels (Fig. 5A). Cardiac hypertrophy markers, such as Nppa, Nppb, and Myh7, were also upregulated by Sp1 overexpression (Fig. 5B). Western blotting confirmed that the protein levels of BNP were significantly increased in H9c2 cells (Fig. 5C,D). Furthermore, Sp1 overexpression induced Sp1 and Nppb expression levels, while protocatechuic acid treatment reversed this effect (Fig. 5E,F). Western blot analysis showed similar results (Fig. 5G). Next, to evaluate the effect of Sp1 overexpression on cardiomyocyte phenotype, pCMV-Sp1-transfected cells were double-stained using anti-Sp1 antibody and Alexa Fluor phalloidin 488, and the cell size of Sp1-positive cells (red) was measured. As shown in Fig. 5H,I, Sp1 overexpression significantly increased the area of the cells; however, protocatechuic acid treatment reversed this effect.
Protocatechuic acid reduced isoproterenol-or Sp1 overexpression-induced PKCγ expression. It has been shown that the kinases that belong to the PKC family are involved in the regulation of cardiac hypertrophy 11 . Furthermore, PKC isoforms are reported to have distinct localization and function 35 .
To evaluate the role of PKC family in our model system, we measured the expression of PKC isoforms in H9c2 cells. The mRNA levels of Prkca and Prkcg were increased in response to isoproterenol and decreased by protocatechuic acid treatment (Fig. 6A); Prkcd and Prkce levels were not significantly affected by isoproterenol. To further evaluate whether the expression of PKC isoforms could be affected by Sp1 overexpression, H9c2 cells were transfected with pCMV or pCMV-Sp1 and then treated with vehicle or protocatechuic acid. Sp1 overexpression significantly increased the expression levels of Prkcg mRNA; however, the levels were reduced by protocatechuic acid treatment (Fig. 6B). Other PKC isoforms were not affected by Sp1 overexpression (Fig. 6B). Similar to gene expression results, the protein levels of PKCγ were also increased ( Fig. 6C,D). To determine whether PKCγ is regulated by ROCK1 and Sp1, siRNA experiments were performed. The knockdown of ROCK1 or Sp1 significantly suppressed isoproterenol-induced PKCγ mRNA and protein levels ( Fig. 6E-G). At the same time, Sp1 siRNA did not affect isoproterenol-induced ROCK1 protein levels ( Supplementary Fig. 4A), while ROCK1 siRNA decreased isoproterenol-induced Sp1 protein levels ( Supplementary Fig. 4B). In the absence of isoproterenol stimulation, Sp1 siRNA transfection did not alter Rock1 mRNA levels ( Supplementary Fig. 5A), ROCK1 siRNA reduced Sp1 mRNA levels ( Supplementary Fig. 5B), while Prkcg mRNA levels were significantly downregulated by both ROCK1 siRNA and Sp1 siRNA ( Supplementary Fig. 5C). Next, we evaluated the role of PKCγ in cardiac hypertrophy. As expected, PKCγ overexpression increased Prkcg mRNA levels (Fig. 6H); however, it also upregulated the mRNA levels of Nppa and Nppb (Fig. 6I,J), but had no effect on Rock1 and Sp1 expression levels ( Supplementary Fig. 6).

Downregulation of ROCK1 and Sp1 expression by protocatechuic acid was not regulated at the transcriptional level.
To investigate the mechanism of ROCK1 and Sp1 regulation by protocatechuic acid, H9c2 cells were treated with isoproterenol and then incubated with vehicle or protocatechuic acid in the presence or absence of actinomycin D, a well-characterized transcription inhibitor. In the absence of isoproterenol stimulation, actinomycin D treatment did not affect the expression of Rock1 and Sp1 mRNA. As expected, the expression of Rock1 and Sp1 induced by isoproterenol was decreased by protocatechuic acid. Furthermore, actinomycin D treatment also significantly reduced isoproterenol-induced expression of Rock1 and Sp1; however, there was no additive inhibitory effect on Rock1 and Sp1 expression in cells treated with protocatechuic acid and actinomycin D (Fig. 8A,B).

Discussion
Here, we showed, both in vitro and in vivo, that protocatechuic acid could be used as a potential treatment for cardiac hypertrophy. We demonstrated that protocatechuic acid attenuated cardiac hypertrophy by inhibiting the ROCK1-Sp1-PKCγ axis (Fig. 8C). Protocatechuic acid is a phenolic compound, structurally similar to gentisic acid 31 . We recently reported that gentisic acid inhibited cardiac hypertrophy in the mouse model of transverse aortic constriction (TAC) 31 . Gallic acid (3,4,5-trihydroxybenzoic acid), another phenolic compound with one more hydroxyl group than protocatechuic acid, also reduced isoproterenol-induced cardiac hypertrophy in mice 36 . Here we characterized protocatechuic acid-mediated inhibition of gross heart hypertrophy by evaluating echocardiographic parameters and cross-sectional area. Protocatechuic acid treatment reduced the isoproterenol-induced increase of cardiomyocyte surface area. Moreover, the isoproterenol-induced increase of left ventricular wall thickness was attenuated by protocatechuic acid. www.nature.com/scientificreports/ www.nature.com/scientificreports/ and Myh7, is characteristic of cardiac hypertrophy 37 . Here, we demonstrated that protocatechuic acid treatment decreased fetal gene expression both in vivo and in vitro. Numerous vegetables and fruits, including green chicory, red chicory, black olives, and black raspberry, are rich in protocatechuic acid 38 . Protocatechuic acid is also a main metabolite of complex polyphenols, such as anthocyanins, which are widely distributed in the human diet. Protocatechuic acid is transformed to protocatechuic aldehyde by an aromatic carboxylic acid reductase 39 , another compound that was reported to protect against isoproterenol-induced cardiac hypertrophy in rats 40 .
Here we showed that the expression of ROCK1, Sp1, PKCα, and PKCγ was increased in response to isoproterenol stimulation in vitro. We hypothesized that the upregulation of these genes was linked to the development of cardiac hypertrophy. RhoA-ROCK pathway is involved in a wide range of biological functions, including contraction, migration, proliferation, differentiation, and apoptosis 41,42 . Activated Rho induces myofibrillar organization and ANP expression in myocytes 43 . The observations describing the role of ROCK1 in cardiac hypertrophy are contradictory. For example, TAC-induced cardiac hypertrophy and fibrosis were promoted in cardiac-specific ROCK1-deficient mice 44 . Global ROCK1 deletion mice, as well as hemizygous ROCK1 +/− mice, exhibited reduced cardiac fibrosis, and the development of cardiac hypertrophy was not affected after TAC or in response to angiotensin II treatment 45,46 . Shi et al. 47 reported that in the transgenic model of Gαq overexpression-induced hypertrophy, ROCK1 deficiency improved the contractile function without reducing cardiac growth. However, our results clearly demonstrated that the knockdown of ROCK1 reduced the size of cardiomyocytes increased by isoproterenol treatment, suggesting that ROCK1 was involved in the pathogenesis of cardiac hypertrophy.
It has been reported that Sp1 levels are increased in right ventricular hypertrophy and in isoproterenol-or angiotensin II-stimulated cardiomyocytes 23,48,49 . In cardiomyocytes, Sp1 interacts with cardiac-specific transcription factor GATA4 and activates the ANP promoter 34 . Sp1 overexpression and knockdown experiments showed that Sp1 acted as an important mediator in the regulation of cardiac hypertrophy. Notably, Sp1 knockdown completely inhibited the isoproterenol-induced increase of cell surface area, as well as the expression of cardiac hypertrophic markers, in H9c2 cells, while Sp1 overexpression increased the expression of these genes. This is the first report demonstrating that Sp1 directly contributes to cardiac hypertrophy as determined by the measurement of cell size. Moreover, protocatechuic acid treatment reduced the Sp1 overexpression-induced www.nature.com/scientificreports/ increase of cell size and hypertrophic marker genes. These findings suggest that Sp1 plays a critical role in the regulation of cardiac growth. The PKC family includes many different isoforms and is implicated in pathogenesis of cardiac hypertrophy 10 . Several studies reported conflicting results regarding the involvement of PKCβ in cardiac hypertrophy 15,16 . Furthermore, it was shown that the PKCα isoform induced cardiac hypertrophy in part through extracellular signalregulated kinase 1/2, while PKCδ, PKCε, and PKCζ did not stimulate hypertrophic growth 12 . In the present study, we discovered that PKCγ played a role in cardiac hypertrophy. PKCγ induced the expression of hypertrophic genes, such as ANP and BNP, which were also increased by Sp1 overexpression. Moreover, PKCγ expression levels were significantly reduced by the knockdowns of ROCK1 and Sp1, either in the presence or absence of cardiac hypertrophic stimulation. These findings indicated that PKCγ was a new downstream target of Sp1 and ROCK1. Although PKCγ does not induce cardiac hypertrophy directly, our PKCγ siRNA experiments showed that it was indirectly involved in isoproterenol-induced cardiac hypertrophy. Furthermore, the isoproterenolinduced expression of ROCK1 and Sp1 was upregulated at the transcriptional level, whereas the downregulation of ROCK1 and Sp1 by protocatechuic acid was not mediated at the transcription level as determined by transcription inhibitor actinomycin D. However, the exact mechanisms involved in ROCK1 and Sp1 downregulation by protocatechuic acid still need to be elucidated.
So far, we have demonstrated that gallic acid, gentisic acid, and protocatechuic acid inhibited cardiac hypertrophy regardless of the number or location of hydroxyl group on benzoic acid. However, there was no direct comparative study of all three phenolic compounds.
In conclusion, we identified protocatechuic acid as an anti-hypertrophic phytochemical. Furthermore, ROCK1 and Sp1 are involved in the regulation the pathological hypertrophy state against beta-adrenergic agonist stimulus. Here we also show that Sp1 is a new pro-hypertrophic regulator. Our data indicate that ROCK1, Sp1, and PKCγ are involved in the development of cardiac hypertrophy and, therefore, could be used as new therapeutic targets for cardiac hypertrophy. Animal model of cardiac hypertrophy. Male CD-1 mice (age, 7 weeks; average weight ~ 33 g) were anesthetized by an intraperitoneal injection of ketamine (120 mg/kg) and xylazine (6.2 mg/kg), and cardiac hypertrophy was induced by isoproterenol infusion (25 mg/kg/day) using the osmotic minipump (Alzet). Mice were randomly divided into three following groups (n = 8/group): vehicle-treated sham group, isoproterenol-infused group, and isoproterenol-infused group with protocatechuic acid (100 mg/kg/day). Isoproterenol was dissolved in 0.1% ascorbic acid and 0.9% saline, while protocatechuic acid was dissolved in dimethyl sulfoxide (DMSO) and diluted using 0.9% saline. Both drugs were administered for 5 days: isoproterenol was continuously infused using an osmotic minipump (24 h/day), while protocatechuic acid was administered daily via intraperitoneal injection (total volume 400 μl). The mice were euthanized after 5 days.

Reagents
Echocardiography. Echocardiography was performed using a Vivid S5 echocardiography system (GE Healthcare, Chicago, IL, USA) with a 13-MHz linear array transducer. Mice were anesthetized by an intraperitoneal injection of tribromoethanol (Avertin; 114 mg/kg) before the procedure. M-mode (2-D guided) images and recordings were acquired from the long-axis view of the left ventricle at the level of the papillary muscles. The thickness of left ventricular posterior and interventricular septa was measured from the images, whereas the left ventricular end-diastolic diameter (LVEDd) and left ventricular end-systolic diameter (LVESd) were measured from the M-mode recordings.
Histology. Mice were euthanized using a 100% grade CO 2 for approximately 2-3 min. The hearts of mice were fixed with 4% paraformaldehyde and embedded in paraffin. The paraffin-embedded tissues were then cut into 3-µm sections, deparaffinized with xylene, and rehydrated in the series of graded ethanols. To measure the cross-sectional cardiomyocyte area, tissue sections were stained with hematoxylin and eosin (H&E) as previously described 50 . To evaluate cell morphology, sections were also stained with wheat germ agglutinin conjugated to Alexa Fluor 488 (1:200) as previously described 36  www.nature.com/scientificreports/ Digital images were obtained using a microscope (Nikon Eclipse 80i microscope, Tokyo, Japan) at a 400 × magnification. The cross-sectional area was quantified using the NIS Elements Software Version AR 3.0 (https:// www. nikon metro logy. com/ images/ broch ures/ nis-eleme nts-en. pdf,Nikon, Tokyo, Japan).

Reverse transcription polymerase chain reaction (RT-PCR).
Total RNA was isolated from heart tissues using TRIzol reagent (Invitrogen/Life Technologies, Carlsbad, CA, USA) and 1 μg was reverse transcribed with TOPscript RT DryMIX (Enzynomics, Daejeon, South Korea). qRT-PCR was performed with a SYBR Green PCR kit (Enzynomics) and gene expression was quantified using the 2 -∆∆Ct method. The PCR primers used in this study are listed in Table 1.