Liganded Peroxisome Proliferator-Activated Receptors (PPARs) Preserve Nuclear Histone Deacetylase 5 Levels in Endothelin-Treated Sprague-Dawley Rat Cardiac Myocytes

Ligand activation of peroxisome proliferator-activated receptors (PPARs) prevents cardiac myocyte hypertrophy, and we previously reported that diacylglycerol kinase zeta (DGKζ) is critically involved. DGKζ is an intracellular lipid kinase that catalyzes phosphorylation of diacylglycerol; by attenuating DAG signaling, DGKζ suppresses protein kinase C (PKC) and G-protein signaling. Here, we investigated how PPAR-DGKζ signaling blocks activation of the hypertrophic gene program. We focused on export of histone deacetylase 5 (HDAC5) from the nucleus, a key event during hypertrophy, since crosstalk occurs between PPARs and other members of the HDAC family. Using cardiac myocytes isolated from Sprague-Dawley rats, we determined that liganded PPARs disrupt endothelin-1 (ET1)-induced nuclear export of HDAC5 in a manner that is dependent on DGKζ. When DGKζ-mediated PKC inhibition was circumvented using a constitutively-active PKCε mutant, PPARs failed to block ET1-induced nuclear retention of HDAC5. Liganded PPARs also prevented (i) activation of protein kinase D (the downstream effector of PKC), (ii) HDAC5 phosphorylation at 14-3-3 protein chaperone binding sites (serines 259 and 498), and (iii) physical interaction between HDAC5 and 14-3-3, all of which are consistent with blockade of nucleo-cytoplasmic shuttling of HDAC5. Finally, the ability of PPARs to prevent neutralization of HDAC5 activity was associated with transcriptional repression of hypertrophic genes. This occurred by first, reduced MEF2 transcriptional activity and second, augmented deacetylation of histone H3 associated with hypertrophic genes expressing brain natriuretic peptide, β-myosin heavy chain, skeletal muscle α-actin, and cardiac muscle α-actin. Our findings identify spatial regulation of HDAC5 as a target for liganded PPARs, and to our knowledge, are the first to describe a mechanistic role for nuclear DGKζ in cardiac myocytes. In conclusion, these results implicate modulation of HDAC5 as a mechanism by which liganded PPARs suppress the hypertrophic gene program.


Introduction
Cardiac hypertrophy is the increase in myocardial mass provoked by hemodynamic stress or myocardial injury, and is a convergence point for many risk factors leading to heart failure. If left unchecked, prolonged hypertrophy is maladaptive and gives rise to cardiac arrest and/or failure [1,2]. Thus, attenuation of hypertrophy is a promising therapeutic target to prevent heart failure.
At the cardiac myocyte level, hypertrophy is characterized by increases in cell size, protein synthesis, and changes in gene expression [3]. The latter includes sequential activation of immediate early response genes (e.g. proto-oncogenes such as c-fos and c-jun), a fetal gene program (e.g. atrial natriuretic peptide, skeletal muscle a-actin, and b-myosin heavy chain), and sarcomeric genes (e.g. cardiac muscle a-actin) [3]. In particular, re-induction of fetal genes such as brain natriuretic peptide (BNP) is one of the most consistent markers of transcriptional activation in hypertrophy [4].
It is therefore well-established that a key event in the hypertrophic gene response is phosphorylation of HDAC5 [11]. In particular, ET1 activates protein kinase C (PKC) and its downstream effector, protein kinase D (PKD). PKD in turn phosphorylates HDAC5 to initiate the signaling cascade which culminates in its exclusion from the nucleus [16]. Indeed, phosphorylation of HDAC5 promotes binding of 14-3-3 chaperone proteins, which then allows export of HDAC5 from the nucleus, de-repression of MEF2, and activation of fetal cardiac genes [12]. In addition, nuclear export of HDAC5 would allow histone acetylation and chromatin relaxation, thereby contributing to the activation of fetal genes [6].
Peroxisome proliferator-activated receptors (PPARs) have beneficial effects in the cardiovascular system. Activation of PPARs a [22][23][24], b/d [24][25][26][27], and c [22,24,28,29] prevents fetal gene activation in response to mechanical or neurohumoral stimuli. A possible intersection between PPAR-and HDAC5mediated repression of cardiac fetal genes lies in the ability of liganded PPARs to activate diacylglycerol kinase zeta (DGKf) [24]. DGKs are intracellular lipid kinases that phosphorylate diacylglycerol (DAG) to produce phosphatidic acid. The resultant decrease in availability of DAG attenuates activation of PKC, thereby terminating an arm of pro-hypertrophic G protein signaling. We previously reported that DGKf contributes to the ability of PPARs to suppress hypertrophy, but the mechanism by which fetal gene expression was attenuated remained unclear.
The aim of this study was to determine the mechanisms by which PPAR-DGKf signaling inhibits hypertrophic gene expression. Because PKC/PKD activity opposes the anti-hypertrophic actions of HDAC5 [16], and liganded PPARs suppress myocyte hypertrophy by interfering with PKC signaling [24], we considered whether PPAR activation might preserve HDAC5 function. Then we sought to elucidate the mechanism by which liganded PPARs promote nuclear retention of HDAC5. Finally, we verified the role of DGKf.

Isolation of rat cardiac myocytes
The study is in full compliance with the Canadian Council on Animal Care, and approval was granted by the University of Manitoba Animal Care Committee (protocol 09-064). 1-day-old Sprague-Dawley rats were sacrificed by decapitation, ventricular myocytes were isolated by alternate cycles of 0.05% trypsin plus mechanical disruption as described [24], and cultured on gelatin-coated plates in DMEM containing 10% cosmic calf serum (CCS) (Hyclone) for 18-24 h prior to experimentation.

Constitutively active PKCe construct
Constitutively active (ca)PKCe was constructed by first cloning rat PKCe cDNA into a pCDH lentivirus vector (MJS BioLynx Inc., Canada). The enzyme was then made constitutively-active by deletion of residues 154-163 of its inhibitory pseudosubstrate domain [30], and verified by DNA sequence analysis (University of Calgary Core DNA Services; Calgary, Canada).

Lentiviral preparation and infection
Lentiviruses expressing small hairpin RNAs (shRNA) against DGKf using shRNAexpressing plasmids (TRCN 0000025394, 0000025395, 0000025398; Open Biosystems; Ottawa, Canada) or caPKCe were prepared as previously described [24]. Scrambled sequences served as non-silencing controls. Lentivirus vector plasmids were co-transfected with psPAX2 (packaging) and pMD2.G (enveloping) vectors using FuGENE6 Reagent (Roche; Indianapolis, Indiana). High-titer lentiviral stock was produced in HEK-293T cells 48 h after transfection. Myocytes were infected for 24 h by application of the lentivirus to the culture medium, and then cultured for a further 72 h (to achieve shRNA knockdown or caPKCe expression) prior to treatments and further experimentation. Knockdown was confirmed by western blotting.
Following the 1 h pre-treatment, hypertrophy was stimulated by addition of ET1 (0.1 mM).

Cell Size
Myocyte size was assessed by immunofluorescent microscopy and computerassisted planimetry, as previously described [31]. Each experiment consisted of 30 cells analyzed per experimental group.

Isolation of cellular nuclear extracts and total protein quantification
Myocyte nuclear isolations were performed from 107 cells using a method adopted and modified from Cayman Chemical's nuclear extraction kit (Ann Arbor, Michigan, USA, CN: 10009277). After treatments, myocytes were washed twice and suspended in phosphate-buffered saline (PBS) and centrifuged at 5006g for 5 minutes at 4˚C. Cell pellets were re-suspended in a phosphatase inhibitor solution in PBS (20 mM NaF, 1 mM b-glycerophosphate, 1 mM Na3VO4). After centrifugation at 5006g for 5 minutes at 4˚C, cell lysis was performed by incubation with hypotonic buffer (5 mM NaF, 10 mM Na2MoO4, 0.1 mM EDTA, 20 mM HEPES, pH 7.5) for 15 minutes on ice. Supernatants containing the cytosolic fraction were removed after the addition of 10% Igepal and subsequent centrifugation at 10006g for 10 minutes at 4˚C. The remaining pellets were re-suspended by vortex in extraction buffer (0.1 mM EDTA, 1.5 mM MgCl2, 420 mM NaCl, 20 mM NaF, 1 mM b-glycerophosphate, 10 mM Na3VO4, 25% glycerol, 0.5 mM phenylmethylsulfonyl fluoride, protease inhibitor tablets, 10 mM HEPES, pH 7.9) and rocked at 4˚C for 15 minutes twice with mixing by vortex in between. Supernatants containing the nuclear fraction were collected by pelleting at 14, 0006g for 10 minutes at 4˚C. Total protein content was quantified from the resulting nuclear extracts using the bicinchoninic acid (BCA) method.

Diacylglycerol (DAG) assay
Nuclear DAG levels were quantified from myocyte nuclear isolates by competitive enzyme-linked immunosorbant assay (ELISA) using a Rat DAG ELISA kit (BlueGene, Shanghai, China, CN: E2D0010). In accordance with manufacturer's instructions, 50 mg of total protein was incubated with DAG-HRP conjugate in a total volume of 160 ml in plates pre-coated with monoclonal anti-DAG for 1 hour at 37˚C. Plates were washed five times and incubated with a proprietary HRPsubstrate at 37˚C for 15 minutes. After the addition of a stop solution the amount of DAG was measured at 450 nm.

Immunoblotting and immunoprecipitation
Cell lysates were prepared in RIPA buffer and clarified by centrifugation. Conventional western blotting of cytosolic and nuclear extracts was used to assess levels of DGKf, native and phosphorylated HDAC5, and native and phosphorylated PKD. To assess physical interaction between HDAC5 and 14-3-3 proteins, HDAC5 precipitates were generated from equal amounts of cardiac myocyte nuclear extracts, and 14-3-3 was detected by western blotting.

DGKf immunostaining
Cells were processed as previously described [32]. Briefly, myocytes were cultured in 4-chamber slides, serum-deprived, and then subjected to the indicated treatments. Cells were then washed with PBS and fixed with 3.7% paraformaldehyde, followed by PBS containing 0.2% Triton X-100 for 2 min. Slides were blocked with PBS containing 0.2% bovine serum albumin and normal horse IgG for 1 h. Nuclei were stained with DAPI (blue; Invitrogen, Grand Island, NY), and immunofluorescent staining was performed using a primary DGKf antibody followed by a secondary antibody conjugated to Alexa-488 (green). Cells were imaged by fluorescence microscopy.

PKCe activity
PKCe was immunoprecipitated from equal amounts of cardiac myocyte nuclear extracts. PKCe activity was assessed using a commercially-available ELISA-based PKC activity assay kit according to the manufacturer's protocol.

MEF2 luciferase assay
Myocytes were infected with lentivirus that expresses a MEF2-luciferase reporter (i.e. firefly luciferase reporter gene driven by a basic promoter element joined to tandem repeats of MEF2; SABiosciences). Myocytes were maintained in DMEM/ 10% CSC for 24 h, then serum-deprived for 24 h. Following treatments (as described in figure legends), luciferase activity was measured from lysates using a Luciferase Reporter Assay System (Promega).
Statistics n-Values are indicated in legends. Error bars represent SEMs. Data were subjected to unpaired t-test or one-way ANOVA and, as appropriate, post-hoc analyses (Newman-Keuls, Bonferroni, or Holm-Sidak multiple comparison tests).

DGK contributes to the anti-hypertrophic actions of PPARs
We first verified the effects of liganded PPARs, and the role of DGK, using the DGK inhibitor, R59022. ET1 increased myocyte size, and this was attenuated by troglitazone, fenofibrate, and GW501516. The anti-hypertrophic actions of PPAR ligands were abolished by R59022 (Fig. 1).

Liganded PPARs attenuate nuclear PKCe in a manner dependent on DGKf
As the ability of PPARs to preserve nuclear HDAC5 levels relies on DGKf activity and suppression of PKCe activity (Fig. 2), we determined that PPAR-DGKf signaling attenuates nuclear PKCe activity. ET1 increased PKCe activity in nuclear extracts (240¡8% vs. control, p,0.01), and this was attenuated by troglitazone, fenofibrate, and GW501516 (Fig. 4). However, the ability of liganded PPARs to prevent PKCe activation was abolished by shRNA knockdown of DGKf.

Suppression of MEF2 by liganded PPARs involves DGKf
Repression of fetal cardiac genes by HDAC5 is achieved, at least in part, through association with MEF2 [17]. In the absence of nuclear HDAC5, MEF2 is free to promote transcription of hypertrophic genes. We found that ET1 increased MEF2 transcriptional activity, as shown by activation of a 3xMEF2-luciferase reporter ( Fig. 7; 194¡33% vs. control, p,0.01), and this was prevented by PPAR ligands. However, the ability of liganded PPARs to prevent MEF2 activation was abolished by shRNA knockdown of DGKf.

Liganded PPARs suppress histone H3 acetylation by a DGKf-dependent mechanism
Acetylation of histone H3 is a response to hypertrophic stimuli in vitro [33] and in vivo [34]. Indeed, ET1 induced H3 acetylation of hypertrophic genes, which favors their transcription (Table 1). Liganded PPARs suppressed ET1-dependent histone acetylation, and again, these PPAR actions were abolished by shRNA knockdown of DGKf.

Discussion
These findings provide insight into the mechanisms by which liganded PPARs suppress hypertrophic gene activation in cardiac myocytes. As depicted in Fig. 8, liganded PPARs increase nuclear DGKf, which in turn prevents ET1-induced activation of nuclear PKCe and PKD, phosphorylation of HDAC5 at conserved 14-3-3-binding sites, and association between HDAC5 and 14-3-3. Collectively, these events lead to nuclear retention of HDAC5. Rescue of HDAC5 activity in the nucleus favors repression of hypertrophic genes vis-à-vis suppression of MEF2 transcriptional activity and augmented deacetylation of histone H3 associated with hypertrophic genes.
Our finding that liganded PPARs increase nuclear DGKf levels may be attributable to either their inhibitory effects on PKCe or their function as nuclear receptor transcription factors [35]. Topham et al. reported that PKC regulates the nuclear localization of DGKf via phosphorylation of the MARCKS-homology domain, a motif which is similar to some nuclear localization sequences. However, this regulatory mechanism was restricted to only the PKC a and c isoforms [36]. It is more likely, therefore, that PPARs modulate nuclear DGKf levels by influencing transcription of the DGKf gene, either directly or indirectly. Admittedly, the PPRE consensus sequence does not appear in the DGKf gene promoter, but this does not preclude direct binding of PPARs. Nakachi et al. detected 167 gene promoters with strict PPARc binding during adipocyte differentiation in 3T3-L1 cells, yet PPRE consensus sequences were present in only 15 of these target genes [37]. Three of these genes were activated by PPARc, lack consensus PPREs, and instead contain imperfect or atypical PPREs that vary in the binding site and/or spacer sequences [38]. Indirect regulation of the DGKf promoter might also be achieved through the formation of a complex with other  1 mM). Subsets of cells were also pre-treated with vehicle or PPAR ligands (troglitazone, 1 mM; fenofibrate, 10 mM; GW501516, 1 mM; 1 h), shRNA knockdown of DGKf, or expression of a constitutively active (ca) PKCe mutant. A, ET1 treatment significantly reduced nuclear HDAC5, and this was attenuated by liganded PPARs. B/C, DGKf knockdown or D, circumvention of DGKf-mediated PKC inhibition using caPKCe blocked PPAR-dependent retention of HDAC5 in the nucleus (Figure 2A) transcription factors such as Sp1. The DGKf promoter contains Sp1 sites, and PPAR-dependent modulation of numerous genes involves Sp1 [39][40][41][42]. In fact, we have data to support the notion that PPAR-mediated regulation of the DGKf promoter occurs through a complex comprised of PPAR and Sp1 (see S1 Figure). Troglitazone-induced DGKf expression was blocked by mithramycin, an Sp1  1 mM). Subsets of cells were also pre-treated with vehicle, R59022 (DGK inhibitor, 10 mM), or PPAR ligands (troglitazone, 1 mM; fenofibrate, 10 mM; GW501516, 1 mM; 1 h). A, Nuclear DAG levels were assessed by ELISA of myocyte nuclear isolates. As would be expected as an output of GPCR-signaling, ET1 increased nuclear DAG levels. Likewise, inhibition of DGK using R59022 (i.e. inhibition of phosphorylative conversion of DAG to phosphatidic acid) also led to augmented nuclear DAG concentrations. n54-6. To determine how ET1 might be modulating nuclear DAG concentrations, nuclear DGKf protein was assessed by B, western blotting of nuclear extracts and total lysates, and C, immunostaining and is presented as percent of normalized DGKf protein vs. vehicle-treated controls. The ability of ET1 to reduce nuclear DGKf was abolished by liganded PPARs. n54. *p,0.05 and **p,0.01 vs. vehicle-treated controls. {p,0.05 and {p,0.01 vs. ET1-treated cells. inhibitor [39]. Also, troglitazone promoted binding of Sp1 to the DGKf promoter and promoted binding between PPARc and Sp1. These data suggest that PPARs and Sp1 are recruited to the DGKf promoter, perhaps as a PPARNSp1-containing complex. Whatever the proximal signaling mechanism, the mechanism by which liganded PPAR modulate nuclear DGKf, at least in part vis-à-vis the nature of the putative interaction between PPARs and the DGKf promoter, requires further study.
DGKf overexpression suppresses cardiac hypertrophy through inhibition of PKCe [43] [44]. Here, we extend this by demonstrating that, specifically in the nucleus, PPARs terminate PKCe and its downstream effector PKD [16] via DGKf. PKD phosphorylates HDAC5 at serines 259 and 498, which serve as docking sites for 14-3-3 [12]; in turn, 14-3-3 binding masks the nuclear localization sequence situated between the phosphorylation sites and reveals a nuclear export sequence at the C terminus. The end result is HDAC5 export from the nucleus, chromatin relaxation and de-repression of HDAC target hypertrophic genes [9,[11][12][13][14]. Thus, the ability of PPAR/DGKf signaling to terminate nuclear PKCe/PKD activity serves to rescue HDAC5 function, thereby conferring resistance to ET1 as a hypertrophic stimulus. These findings are consistent with a previous report that mutation of serines 259 and 498 results in loss 14-3-3 protein binding and a significant increase in HDAC5 deacetylation activity [45].
Our notion, that preserved HDAC function is responsible for PPAR effects, is consistent with a previous report in which PPARc-induced adipogenesis was attenuated by a chemical inhibitor of class II HDACs, MC1568 [46]. Furthermore, using PPRE-luciferase-expressing reporter mice, Nebbioso et al showed that MC1568, impairs PPARc signaling mostly in heart and adipose tissues [46]. In hypertrophied myocytes, our findings suggest liganded PPARs suppress hypertrophic gene expression, at least in part, by tethering of MEF2 to HDAC5. This inhibits MEF2 transcriptional activity (Fig. 7) by abolishing binding to other pro-hypertrophic transcription factors such as GATA and nuclear factor of activated T-cells (NFAT) [18][19][20].
In addition to MEF2 sequestration, our data suggest a secondary antihypertrophic action of HDAC5. In myocytes, induction of fetal genes such as the atrial natriuretic peptide (ANP) and b-myosin heavy chain genes is associated with acetylation of histone H3 [17,47] that is preferentially localized to the promoter regions of the gene [33]. Further, histone H3 acetylation is exaggerated in the hearts of pressure overloaded mice [34]. We found that PPARs suppress acetylation of histone H3 on the promoters of genes encoding BNP, b-myosin heavy chain, skeletal muscle a-actin, and cardiac muscle a-actin ( Table 1), suggesting that histone H3 deacetylation may be a secondary mechanism by which nuclear PPAR/DGKf signaling represses the hypertrophic gene program.
doi:10.1371/journal.pone.0115258.g005 Fig. 6. Interaction of nuclear HDAC5 with 14-3-3 is modulated by liganded PPARs. Myocytes deprived of serum for 24 h were stimulated with ET1 (0.1 mM). Subsets of cells were also pre-treated with vehicle or PPAR ligands (troglitazone, 1 mM; fenofibrate, 10 mM; GW501516, 1 mM; 1 h) in the presence and absence of shRNA knockdown of DGKf. HDAC5 was immunoprecipitated, and interaction between HDAC5 and 14-3-3 was assessed biochemically by immunoblotting. Data are presented as percent of normalized 14-3-3 vs. vehicle-treated controls. ET1 treatment significantly increased binding between HDAC5 and 14-3-3. n54. The ability of ET1 to increase HDAC5N14-3-3 interaction was A, abolished by liganded PPARs (n56), and B, restored by shRNA knockdown of DGKf (n54). **p,0.01 vs. vehicle-treated controls. {p,0.01 vs. ET1-treated cells. PPARa are related to suppression of AP-1 DNA-binding activity [23]. ET1induced PKCe leads to activation of the AP-1 transcription factor [43,48], and forced overexpression of DGKf in cardiac myocytes blocks not only ET1dependent hypertrophy, but also AP-1 signaling [43]. AP-1 is regulated by p300, a transcriptional co-activator and HAT. In fact, p300-induced acetylation of dimeric AP-1 complexes enhances AP-1 binding to DNA [49]. Thus, prohypertrophic ET1/PKCe-dependent activation of AP-1 signaling might also be  attenuated by PPAR-DGKf crosstalk and depletion of DAG, though this remains to be determined. The present study shows for the first time that the ability of PPAR ligands to block activation of hypertrophic genes in response to ET1 involves PPAR/DGKf crosstalk in the cardiac myocyte nucleus. The mechanistic contribution of DGKf is likely due to attenuation of PKCe signaling by depleting its substrate, DAG. Retention of HDAC5 in the nucleus appears to be a consequence of attenuated PKCe and PKD signaling, leading to suppression of pro-hypertrophic events such as histone acetylation of hypertrophic genes and MEF2 activation. Taken in context with other studies on DGKf, interventions that target PPAR/DGKf signaling might represent a novel therapeutic approach to address the problem of cardiac hypertrophy.
Supporting Information S1 Figure. Troglitazone promotes recruitment of PPARc and Sp1 to the DGKf promoter. doi:10.1371/journal.pone.0115258.s001 (PDF) Fig. 8. Proposed mechanism by which liganded PPARs abrogate ET1-induced hypertrophic gene expression. ET1 induces accumulation of DAG in the plasma membrane, which causes translocation and activation of PKCe, an early signaling mediator of hypertrophy. Downstream events including phosphorylation of PKD and HDAC5 give rise to 14-3-3-facilitated export of HDAC5 from the nucleus and subsequently de-repression of hypertrophic genes. Liganded PPARs stimulate nuclear DGKf to phosphorylate DAG. The resulting decreased availability of DAG attenuates PKCe and downstream signaling effectors, resulting in nuclear retention of HDAC5. doi:10.1371/journal.pone.0115258.g008