HDAC5 integrates ER stress and fasting signals to regulate hepatic fatty acid oxidation[S]

Disregulation of fatty acid oxidation, one of the major mechanisms for maintaining hepatic lipid homeostasis under fasting conditions, leads to hepatic steatosis. Although obesity and type 2 diabetes-induced endoplasmic reticulum (ER) stress contribute to hepatic steatosis, it is largely unknown how ER stress regulates fatty acid oxidation. Here we show that fasting glucagon stimulates the dephosphorylation and nuclear translocation of histone deacetylase 5 (HDAC5), where it interacts with PPARα and promotes transcriptional activity of PPARα. As a result, overexpression of HDAC5 but not PPARα binding-deficient HDAC5 in liver improves lipid homeostasis, whereas RNAi-mediated knockdown of HDAC5 deteriorates hepatic steatosis. ER stress inhibits fatty acid oxidation gene expression via calcium/calmodulin-dependent protein kinase II-mediated phosphorylation of HDAC5. Most important, hepatic overexpression of a phosphorylation-deficient mutant HDAC5 2SA promotes hepatic fatty acid oxidation gene expression and protects against hepatic steatosis in mice fed a high-fat diet. We have identified HDAC5 as a novel mediator of hepatic fatty acid oxidation by fasting and ER stress signals, and strategies to promote HDAC5 dephosphorylation could serve as new tools for the treatment of obesity-associated hepatic steatosis.

Endoplasmic reticulum (ER) is responsible for protein folding, lipid and sterol biosynthesis, and calcium storage. Accumulation of unfolded proteins in ER leads to ER stress, and the unfolded protein response (UPR) through PKR-like endoplasmic reticulum kinase, inositol-requiring enzyme 1 (IRE1), and activating transcription factor 6 (ATF6) pathways serve as major mechanisms for restoring ER homeostasis under ER stress conditions (18). However, unresolved or prolonged ER stress influences cellular calcium metabolism (19), and the release of ER calcium stores into the cytosol activates calcium/calmodulin-dependent protein kinase II (CaMKII), which is critical for ER stress-induced apoptosis (20,21). Hepatic ER stress is closely associated with obesity-induced steatosis (22)(23)(24)(25). Obesity and type 2 diabetes directly induce hepatic ER stress (26,27), which leads to steatosis (23). However, it is not completely understood how elevated ER stress in the liver contributes to steatosis.
In this study, we identify HDAC5, a major component of the fasting glucagon signaling pathway, as a key mediator of hepatic fatty acid oxidation gene expression. We demonstrate that fasting-induced dephosphorylation of HDAC5, which is suppressed by ER stress, promotes its binding to and activation of PPAR. As a result, ER stress-dependent hepatic steatosis is greatly attenuated in mice expressing a phosphorylation-defective HDAC5. Our data thus provide new evidence demonstrating the effect of HDAC5 on hepatic lipid homeostasis under physiological and pathological conditions. Furthermore, we demonstrate a potential, novel therapeutic strategy for treatment of obesity-associated hepatic steatosis.

Animals and adenovirus
Male C57BL/6J mice were purchased from Shanghai Laboratory Animal Center (Shanghai, China) and were adapted to colony cages with 12 h light/dark cycle in a temperature-controlled environment with free access to water and standard irradiated rodent diet (5% fat; Research Diet D12450, New Brunswick, NJ). For high-fat diet (HFD) studies, 6-week-old mice were maintained on HFD (60% fat; Research Diets D12492) for 12 weeks. For adenovirus injection, 1×10 8 pfu Ad-HDAC5, Ad-HDAC5 2SA, Ad-HDAC5300-480, Ad-GFP, Ad-unspecific RNAi (USi), and Ad-HDAC5 RNAi (HDAC5i) were delivered by tail-vein injection. Six days after injection, mice were fasted for 24 h before sacrifice. All animal studies were approved by the animal experiment committee of Tongji University and in accordance with the guidelines of the School of Medicine, Tongji University.

In vitro analysis
Mouse tissues were frozen in liquid nitrogen and kept at -80°C until further use. Livers were homogenized by using tissue homogenizer at 4°C in lysis buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 5 mM EDTA, 30 mM sodium pyrophosphate, 30 mM sodium fluoride, 1% Triton-X 100, and protease inhibitor cocktail). Lysates were reserved for immunoblot and immunoprecipitation. Liver triglyceride levels were determined as previously reported (15).

Quantitative real-time PCR and immunoblot
Real-time PCR was performed as previously (31). Briefly, total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA), and reverse transcription was done using FastQuant RT kit (Tiangen, Shanghai, China). Real-time PCR was carried out using SuperReal SYBR Green kit (Tiangen, Shanghai, China) and Lightcycler 96 (Roche, Penzberg, Germany). All reactions were performed in duplicate. The amplification efficiency for each primer pair and the cycle threshold (Ct) were determined automatically by Lightcycler software (Roche, Penzberg, Germany). The fold-change was calculated by the comparative CT (2 CT ) method against -actin (32). Immunoblot and immunoprecipitation were performed as described (33). Briefly, cells were washed with PBS and then resuspended in lysis buffer (150 mM NaCl, 1% Triton-X 100, 1 mM EDTA, 50 mM Tris pH7.5, and protease inhibitor cocktail). For immunoblot, protein content in the supernatant was determined by using the Micro BCA protein assay kit (Pierce, Rockford, IL) and suspended in sample buffer (100 mM Tris, PH 6.8, 4% SDS, 20% glycerol, 0.1% bromophenol blue). The samples were separated on SDS-PAGE gels, transferred, probed with antibodies, and visualized using ECL reagents. For immunoprecipitation, the supernatant was precleaned with protein A/G agarose for 30 min and then incubated overnight on a rocker with primary antibodies at 4°C, followed by incubation with Protein A/G agarose beads for another 2 h. The immunoprecipitates were extensively washed with lysis buffer and suspended in sample buffer for SDS-PAGE analysis.

Luciferase reporter assay
Human embryonic kidney (HEK) 293T cells were transfected with Fgf21-PPRE-luc and respiratory syncytial virus (RSV) -gal, together with PPAR/retinoid X receptor  (RXR) plasmid and indicated constructs for 24 h, and luciferase assays were performed by using the Promega GloMax96 system according to the manufacturer's instructions (33). We used -gal assay to normalize the expression levels.

Oil Red O staining
Livers embedded in optimal cutting temperature compound (Tissue-Tek, Laborimpex) were used for Oil Red O staining to assess hepatic steatosis.

Statistical analysis
All studies were performed on at least three independent occasions. Results are reported as mean ± SEM. Differences between two groups were assessed with unpaired Student's t test. Data involving more than two groups were assessed by ANOVA with Bonferroni post hoc test. A P value of <.05 was considered statistically significant.
Although HDAC5 is phosphorylated at consensus SIK recognition sites and sequestered in the cytoplasm under ad lib conditions, fasting triggered HDAC5 dephosphorylation at Ser256 and Ser498 and nuclear translocation (16).
Changes in hepatic fatty acid oxidation in AD-HDAC5i-infected mice under fasted conditions prompt us to further investigate the effect of HDAC5 and the phosphorylationdefective HDAC5 mutant (HDAC5 S259/498A, HDAC5 2SA), which exhibits a permanent nuclear localization identical to wild-type HDAC5 localization upon glucagon or FSK treatment (16), on PPAR transcriptional activity. Although HDAC5 expression caused a significant increase of PPAR/RXR-induced activation of the Fgf21-luciferase reporter in HEK293T cells, HDAC5 2SA expression further boosted the effect (Fig. 1E). This effect seemed to be PPAR ligand-independent, because PPAR activation function 2 (AF2) mutant (lacking the ligand-dependent activation) (36) and RXR-induced activation of the Fgf21luciferase reporter were still able to be promoted by HDAC5 (supplemental Fig. S2A). Consistently, Ad-HDAC5 expression promoted the expression of PPAR target genes known to regulate fatty acid oxidation, including Cpt1a, Hmgcs2, Lcad, and Mcad; this effect was further enhanced when primary hepatocytes were expressed with Ad-HDAC5 2SA (Fig. 1F). Together, these data indicate that HDAC5 plays an important role in regulating hepatic fatty acid oxidation gene expression under fasted state.

HDAC5 promotes fatty acid oxidation gene expression via its interaction with PPAR
HDAC5 has been reported to interact with transcription factors, such as FOXO1 (16), MEF2 (37), and p65 (33), and modulate their transcriptional activity. On the basis of the effect of HDAC5 on PPAR activity, we tested whether HDAC5 associates with PPAR. Indeed, we recovered endogenous PPAR via immunoprecipitation with endogenous HDAC5 in primary hepatocytes ( Fig. 2A). Consistent with this association, hemagglutinin (HA)-tagged PPAR but not HA-tagged RXR could be pulled down by Flagtagged HDAC5 in HEK293T cells (Fig. 2B, C). Interestingly, exposure to FSK greatly increased the interaction between HDAC5 and PPAR (Fig. 2D) and HDAC5 2SA showed a higher affinity to interacting with PPAR ( Fig.  2E) in HEK293T cells. HDAC5 contains an adaptor domain in the N-terminal region and a conserved catalytic domain (HDAC domain) in the C-terminal region. To further establish the interaction domain of HDAC5 with PPAR, we tested various truncated forms of Flag-tagged HDAC5 for the ability to bind PPAR in HEK293T cells (Fig. 2F). HA-tagged PPAR was found to associate with Flag-tagged HDAC5 as well as HDAC5 1-661, HDAC5 300-661 mutants, and HDAC5 480-661 mutant to a lesser extent; however, 300-480 truncated mutation of HDAC5 disrupted the HDAC5-PPAR interaction (Fig. 2G-I).
To explore whether HDAC5 interaction with PPAR directly modulates PPAR activity, we determined the effect of HDAC5 and HDAC5 300-480 mutant on PPARinduced activation of the Fgf21-luciferase reporter in HEK293T cells. In a manner consistent with the interaction data, HDAC5 but not HDAC5 300-480 mutant significantly increased PPAR/RXR-induced activation of the Fgf21-luciferase reporter (Fig. 3A). Furthermore, WY14643 stimulated expression of PPAR target genes (Cpt1a, Hmgcs2, Lcad, and Mcad) and was dramatically increased by Ad-HDAC5 but not Ad-HDAC5 300-480 mutant infection in primary hepatocytes (Fig. 3B). We further tested the influence of HDAC5 300-480 mutant on hepatic lipid homeostasis in vivo; mice were injected intravenously with adenovirus encoding Ad-GFP, Ad-HDAC5, or Ad-HDAC5 300-480 mutant (supplemental Fig. S3A). After 24 h of fasting, while body weight, liver weight, and plasma glucagon levels remained unchanged, mice injected with Ad-HDAC5 but not Ad-HDAC5 300-480 mutant showed decreased plasma NEFA levels, lipid accumulation, and triglyceride levels in liver and increased ketone bodies (Fig. 3C, D, and supplemental Fig. S3B, C). Consistently, expression of PPAR target genes known to regulate fatty acid oxidation (Cpt1a, Mcad, Lcad, Hmgcl, Hmgcs2, PPARa and Fgf21) was significantly increased in the livers of mice injected with Ad-HDAC5 but not Ad-HDAC5 300-480 mutant (Fig. 3E). Taken together, these data suggest that HDAC5 promotes PPAR transcriptional activity through their interaction.

Obesity-induced ER stress suppresses fatty acid oxidation gene expression through phosphorylation of HDAC5
Considering that obesity is characterized by hyperglucagonemia (38) and that HDAC5 undergoes glucagonstimulated dephosphorylation, we tested whether HDAC5 phosphorylation is altered in this setting. Surprisingly, contrary to what we expected, HFD-fed mice exhibited increased hepatic amounts of phosphorylated HDAC5 in relation to RD controls under fasted state (Fig. 4A), suggesting that another mechanism besides glucagon might contribute to its phosphorylation. It is well known that obesity-induced ER stress contributes to hepatic steatosis (27), and we further confirmed the induction of ER stress in HFD-fed  mouse livers by showing that hepatic mRNA level of GRP78, an ER chaperone (27), was elevated in HFD-fed mice in comparison with controls (Fig. 4B). ER stress stimulates the release of ER calcium stores into the cytosol, which activates CaMKII (20,21). On the basis of the fact that HDAC5 has been implicated as a substrate of CaMKII (39)(40)(41), we tested whether obesity-induced ER stress accounted for the increased HDAC5 phosphorylation through calcium-CaMKII pathway. Indeed, HDAC5 phosphorylation was significantly upregulated in a time-dependent manner by the ER stress inducer, THA treatment in primary hepatocytes (Fig. 4C). Interestingly, FSK-stimulated dephsphorylation of HDAC5 was also inhibited by THA treatment in primary hepatocytes (Fig. 4D). Pretreatment of primary hepatocytes with the intracellular calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetate or CaMKII inhibitor KN62 blocked THA-induced HDAC5 phosphorylation (Fig. 4E, F). We next directly tested the requirement of HDAC5 phosphorylation for the effect of ER stress on fatty acid oxidation gene expression. While exposure of primary hepatocytes to THA strongly inhibited fatty acid oxidation gene expression, expression of HDAC5 2SA, but not HDAC5, fully restored fatty acid oxidation gene expression (Cpt1a, Hmgcs2, Lcad, and Mcad) (Fig. 4G). These data suggest that ER stress inhibits PPAR activity via CaMKIIinduced HDAC5 phosphorylation.
Aberrant increase in ER stress-induced HDAC5 phosphorylation levels in HFD-fed mice indicate that strategies to promote HDAC5 dephosphorylation could serve as potential new tools to ameliorate obesity-associated hepatic steatosis. Indeed, Ad-HDAC5 2SA injection in HFD-fed mice (supplemental Fig. S4A) greatly decreased plasma NEFA levels, hepatic lipid accumulation, triglyceride levels in liver, and increased ketone bodies compared with controls, whereas body weight, liver weight, and plasma glucagon levels remained unchanged (Fig. 4H, I, and supplemental S4B, C). Consistently, expression of PPAR target genes known to regulate fatty acid oxidation (Cpt1a, Mcad, Lcad, Hmgcs2, Ppara, and Fgf21) were significantly increased in the livers of mice injected with Ad-HDAC5 2SA (Fig. 4J).

DISCUSSION
The liver is a major organ that controls glucose and lipid metabolism in response to hormonal signals. In the past decade, the ER stress-induced UPR pathway has emerged as an important modulator of hepatic glucose and lipid metabolism. ATF6 reduces hepatic glucose output by disrupting the CREB-CRTC2 interaction (42) and increases fatty acid oxidation to attenuate hepatic steatosis through PPAR (43). IRE1 promotes glucagon-stimulated gluconeogenesis (44) and prevents hepatic steatosis through repressing expression of key metabolic transcriptional regulators such as PPAR (25). XBP1s inhibits hepatic gluconeogenesis by targeting FOXO1 for proteasomal degradation (45) and meanwhile promotes lipogenesis (22). Beside the UPR pathways, ER stress also leads to the release of Ca 2+ from the ER lumen to the cytosol to activate CaMKII (20,21). Here, we report that the calcium-CaMKII-HDAC5 pathway mediates ER stress-induced suppression of fatty acid oxidation gene expression. Meanwhile, it has been reported that HDAC5 could also interact with LXR to impact lipogenesis (46), and we were also able to detect the interaction of HDAC5 with LXR when expressed in HEK293T cells (supplemental Fig. S5A). Thus, it is tempting to speculate that this pathway may contribute to broader hepatic metabolic pathways, which will need further investigation.
Glucagon levels are elevated in subjects with type 2 diabetes and contribute to the development of excessive hepatic glucose production and hyperglycemia (47). Although glucagon is known to induce hepatic fatty acid oxidation and suppress lipogenesis in liver, excessive triacylglycerol deposits cause steatosis in subjects with type 2 diabetes. The detailed mechanism for this paradox still remains unsolved and greatly limits the use of glucagon antagonism as a potential strategy for type 2 diabetes in human. Our previous work showed that impaired glucagon-stimulated cAMP efflux from liver by obesity accounts for the excessive triacylglycerol deposits in the pathophysiology of type 2 diabetes (15). Here, we provide a new insight into the mechanism by showing that ER stress induced by obesity and type 2 diabetes suppresses glucagon-stimulated HDAC5 dephosphorylation and HDAC5mediated PPAR activity, which lead to hepatic steatosis. Hence, hyperglucagonemia with defective glucagon signaling defines a new glucagon resistance status in obesity and type 2 diabetes, and ER stress functions as an important inducer of glucagon resistance together with insulin resistance (48).
Taken together, we show that regulation of HDAC5 phosphorylation status by fasting glucagon and ER stress serves as an important mechanism for modulating PPAR activity and hepatic fatty acid oxidation. This mechanism has an important role in the development of hepatic steatosis under both physiological and pathological conditions. Thus, a systematic investigation into the role of the glucagon signal pathway and the ER stress pathway in fatty acid oxidation would likely lead to novel therapeutic strategies for manipulating obesity-associated hepatic steatosis.