Enhancement of Inflammatory Protein Expression and Nuclear Factor Κb (NF-Κb) Activity by Trichostatin A (TSA) in OP9 Preadipocytes

The production of inflammatory proteins such as interleukin-6 (IL-6) by preadipocytes and mature adipocytes is closely associated with the impairment of systemic glucose homeostasis. However, precisely how the production is regulated and the roles of histone deacetylases (HDACs) remain largely unknown. The aim of this study was to establish whether HDAC inhibitors affect the expression of inflammatory proteins in pre/mature adipocytes, and, if so, to determine the mechanism involved. Trichostatin A (TSA), an HDAC inhibitor, enhanced lipopolysaccharide (LPS)-induced production of IL-6 in OP9 preadipocytes but not the mature adipocytes. Moreover, TSA also enhanced palmitic acid-induced IL-6 production and the expression of inflammatory genes induced by LPS in preadipocytes. Although TSA did not affect TLR4 mRNA expression or the activation of MAPKs, a reporter gene assay revealed that the LPS-induced increase in nuclear factor κB (NF-κB) activity was enhanced by TSA. Moreover, TSA increased the level of NF-κB p65 acetylation at lysine 310 and duration of its translocation into the nucleus, which leads to enhancement of NF-κB activity and subsequently expression of inflammatory genes. These findings shed new light on the regulatory roles of HDACs in preadipocytes in the production of inflammatory proteins.

Lipopolysaccharide (LPS), an abundant glycolipid of the outer membrane of gram-negative bacteria, is recognized by TLR4 and induces inflammatory responses in various cells [9,10]. A small amount of LPS is constitutively absorbed through the gut even in a healthy state. This triggers weak systemic inflammation including in adipose tissue, which, in turn, leads to impairment of glucose homeostasis, which is called metabolic endotoxemia [11]. Moreover, saturated free fatty acids including palmitic acid (PA) also induce inflammatory responses as ligands of TLR4 [12,13]. TLR4 signaling triggers the activation of transcription factors including nuclear factor kB (NF-kB) and of mitogen-activated protein kinases (MAPKs) such as extracellular signal-regulated kinase (ERK), p38 MAPK and c-Jun N-terminal kinase (JNK) [9]. NF-kB is activated via phosphorylation and degradation of IkB, an inhibitory factor of NF-kB, by the IkB kinase (IKK) enzyme complex [14,15], which leads to the nuclear translocation of NF-kB and the subsequent transcription of NF-kB-dependent genes, such as the IL-6, MCP-1, cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS) genes.
It has been shown that HDACs are closely involved in inflammatory responses and HDAC inhibitors affect the expression of inflammatory proteins in various cell lines. For example, TSA enhanced inflammatory gene transcription in microglial cells, epithelial cells, fibroblasts, thymocytes and splenocytes [21][22][23][24]. It also did the same in peripheral blood mononuclear cells from diabetics [25]. Additionally, knockdown of HDAC2 by siRNA enhanced LPS-induced granulocyte macrophage-colony stimulating factor expression in epithelial cells and primary human lung macrophages [26]. On the other hand, TSA repressed IL-1b/ LPS/IFNc-induced iNOS expression in murine macrophages [27]. However, the roles of HDACs in the inflammatory responses of preadipocytes and mature adipocytes, which are important in the maintenance of systemic glucose homeostasis, remain largely unknown. We therefore examined the effects of HDAC inhibitors on the expression of inflammatory proteins in pre/mature adipocytes.
Cell culture and drug treatments OP9 cells (obtained from Riken Bio Resource Center Cell Bank (RCB1124), Tsukuba, Japan), a line of bone marrow-derived mouse stromal cells, were cultured in minimum essential medium alpha (MEM-a, Invitrogen, Carlsbad, CA) supplemented with 20% (v/v) heat-inactivated fetal bovine serum (FBS, Biowest, Miami, FL), 18 mg/ml penicillin G potassium (Meiji Seika, Tokyo, Japan) and 50 mg/ml streptomycin sulfate (Meiji Seika) at 37uC in a humidified atmosphere of 95% (v/v) air and 5% (v/v) CO 2 . The cells were seeded in each well of a multi well plate (Becton, Dickinson and Company, Franklin Lakes, NJ) at 5,000 cells/cm 2 . 3T3-L1 cells (obtained from Health Science Research Resources Bank, Japanese Collection of Research Bioresources Cell Bank (JCRB9014), Osaka, Japan) were cultured in Dulbecco's modified Eagle's medium (Nissui Seiyaku, Tokyo, Japan) supplemented with 10% (v/v) FBS, 18 mg/ml penicillin G potassium and 50 mg/ml streptomycin sulfate at 37uC in a humidified atmosphere of 95% (v/v) air and 5% (v/v) CO 2 . The cells were seeded in each well of a multi well plate at 15,000 cells/cm 2 . Both OP9 cells and 3T3-L1 cells were incubated at 37uC for 3 days until confluent, and then used for experiments. They were preincubated for 1 hour in the presence or absence of HDAC inhibitors (TSA and SAHA), MAPK inhibitors (U0126, SB203580, and SP600125) or an IKK-2 inhibitor (TPCA-1), and stimulated with LPS, PA or TNFa for a specified period of time.

Differentiation into mature adipocytes
Adipocyte differentiation induced by insulin and oleic acid (IO) was performed according to the method described by Wolins et al. [28] with minor modifications. Briefly, OP9 cells were seeded at 5,000 cells/cm 2 and grown in MEM-a supplemented with 20% (v/v) FBS, 18 mg/ml penicillin G potassium and 50 mg/ml streptomycin sulfate for 3 days until confluent (day 0, preadipocytes) and then cultured for specified periods in MEM-a with 0.2% (v/v) FBS, 175 nM insulin and 900 mM oleic acid bound to albumin (5.5:1 molar ratio). Adipocyte differentiation was induced by TGZ as follows: OP9 cells were seeded at 5,000 cells/cm 2 , grown for 3 days until confluent (day 0, preadipocytes), and treated for specified periods with MEM-a supplemented with 20% (v/v) FBS and 3 mM TGZ. Typical morphological aspects of OP9 preadipocytes and differentiated adipocytes are represented in Figure 1A.

Oil red O staining
Cells were washed with phosphate-buffered saline (PBS), then fixed in 10% (v/v) neutralized formalin (pH 7.2) for 10 minutes at room temperature. After a wash with PBS, 60% (v/v) isopropanol was added for 1 minute. Then, cells were stained in a 0.18% (w/v) Oil red O solution (Sigma-Aldrich) for 15 minutes at room temperature. The cells were washed with 60% (v/v) isopropanol and PBS, and observed under a phase-contrast microscope (Nikon, Tokyo, Japan). For quantification of lipid contents, the Oil red O in the cells was extracted with 100% isopropanol for 15 minutes at room temperature and the absorbance of extracts at 540 nm was determined.
ELISA IL-6 and MCP-1 levels were measured by ELISA according to the manufacturer's instructions (eBioscience, San Diego, CA, and R&D systems, respectively).

Quantitative real-time PCR
Total RNA was extracted using a GenElute mammalian total RNA kit (Sigma-Aldrich). cDNA was synthesized from 500 ng of the total RNA by reverse transcription using PrimeScript RT master mix (Takara, Shiga, Japan) according to the manufacturer's instructions, and diluted 10 fold with Tris-EDTA buffer. Real-time PCR was performed using SYBR premix Ex Taq II (Takara) in a Thermal cycler dice real time system (Takara). As an internal control, levels of glyceraldehyde-3-phosphate dehydrogenase

Plasmid constructs and site-directed mutagenesis
The NF-kB-responsive luciferase reporter gene was obtained from Stratagene (La Jolla, CA), and the Renilla luciferase gene (control, pGL4.75[hRluc/CMV] Vector), from Promega. The murine IL-6 promoter (21232 to +39) was supplied by Dr. T. Kishimoto, Osaka University, Japan and Dr. A. Kimura, Keio University, Japan, and integrated into the pGL3-basic vector (Promega). Mutations in the putative NF-kB, AP-1 and CCAAT/ enhancer binding protein (C/EBP)-binding sites or cAMP response element (CRE) in the murine IL-6 promoter were generated using the QuikChange lightning site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Luciferase assay OP9 cells were transfected in 24-well plates with the target (500 ng) and control (20 ng) luciferase reporter plasmid in 48 ml of Opti-MEM (Invitrogen), per well, using X-tremeGENE HP DNA transfection reagent (Roche, Basel, Switzerland) according to the manufacturer's instructions. After 24 hours, the culture medium was changed and the cells were stimulated. Luciferase activity in the cells after 8 hours of LPS stimulation was determined using a Dual-luciferase reporter assay system (Promega).

Immunostaining
After stimulation, OP9 cells on 96-well plates (Becton, Dickinson and Company) were fixed in 4% neutral buffered paraformaldehyde for 10 minutes and then 220uC methanol for 10 minutes. The cells were next permeated in 0.5% (v/v) Triton X-100-PBS for 10 minutes. They were blocked in 4% (w/v) Block Ace (DS Pharma Biomedical, Osaka, Japan) for 1 hour at room temperature, and incubated overnight at 4uC with the rabbit anti-NF-kB p65 antibody. After being washed with PBS containing 0.1% (v/v) Tween20, the cells were incubated at room temperature with the Alexa Fluor 488-conjugated anti-rabbit IgG antibody (Invitrogen). Then, they were stained with 300 nM 49,6-diamidino-2-phenylindole (DAPI) for 5 minutes at room temperature, and fluorescence images of 4 randomly chosen fields were taken using an IN Cell Analyzer 2000 (GE Healthcare, Buckinghamshire, UK). NF-kB nuclear translocation was calculated with an established protocol using IN Cell WorkStation software (GE Healthcare).

Fractionation of nucleus and cytoplasm
Fractionation of the nucleus and cytoplasm was performed using a nuclear extract kit (Active Motif, Carlsbad, CA) according to the manufacturer's instructions.

Statistical analysis
The statistical significance of the results was analyzed using Student's t-test or a one-way ANOVA with the Bonferroni/ dunnett post hoc test for multiple comparisons.

Results
OP9 preadipocytes were made to differentiate into mature adipocytes using both methods As shown in Fig. 1A, treatments both with IO and with TGZ caused the formation of many lipid droplets in the cytosol of OP9 cells. The lipid contents reached a maximum at 4 days with the IO-based method and at 6 days with the TGZ-based method (Fig. 1B). Moreover, the mRNA expression of adipocyte differentiation markers such as PPARc (Fig. 1C), adiponectin (Fig. 1D) and C/EBPa (Fig. 1E) was increased by the treatments.
TSA did not affect TLR4-mediated IL-6 production in mature adipocytes IO and TGZ-treated mature OP9 adipocytes were incubated for 1 hour in the presence or absence of TSA (3 and 10 nM), and then stimulated with LPS (1 mg/ml). The acetylation levels of histone H4 were apparently increased by TSA (10 nM) during the pretreatment for 1 hour (at time 0) and lasted 24 hours in both differentiated cells ( Fig. 2A and B). On the other hand, the stimulation by LPS caused the production of IL-6 in mature adipocytes but was not affected by pretreatment with TSA (3 and 10 nM) ( Fig. 2C and D).
TSA enhanced TLR4-mediated IL-6 production and inflammatory gene expression in preadipocytes Next we carried out similar experiments in OP9 preadipocytes. As in the differentiated adipocytes, TSA increased acetylation levels of histone H4 during 1 hour of pretreatment (at time 0) and the effect lasted 24 hours (Fig. 3A). However, unlike in the mature adipocytes, TSA significantly enhanced the LPS-induced IL-6 production although TSA alone did not induce IL-6 production (Fig. 3B). This enhancing effect of TSA appeared from 12 hours after LPS stimulation (Fig. 3C). Levels of IL-6 mRNA in LPSstimulated OP9 cells increased within 2 hours and remained high for 24 hours. TSA also significantly increased the levels of IL-6 mRNA from 2 to 24 hours after LPS stimulation (Fig. 3D). Moreover, treatment with PA (300 mM) induced IL-6 production, which was also significantly enhanced by TSA pretreatment (Fig. 3E). SAHA (100 nM), another pan-HDAC inhibitor, enhanced LPS-induced IL-6 production as well (Fig. 3F). Similar results were obtained with 3T3-L1 cells, another preadipocyte cell line. Namely, TSA enhanced LPS-induced IL-6 production significantly at both the mRNA (Fig. 3G) and protein (Fig. 3H) levels.
Then, we examined whether TSA enhanced the expression of other inflammatory genes such as the COX-2, iNOS and MCP-1 genes. The expression levels of these genes were increased by LPS and all were enhanced by TSA although the effect varied (Fig. 4A, B and C). Corresponding to the level of MCP-1 mRNA, the concentration of MCP-1 in culture supernatants was increased by LPS stimulation, which was enhanced slightly but significantly by TSA pretreatment (Fig. 4D).  The effect on inflammatory protein expression by TSA was independent of TLR4 To clarify the mechanism by which HDAC inhibitors enhanced the expression of inflammatory proteins in preadipocytes, we firstly examined whether TSA affected the expression of TLR4 and TLR4-independent responses. OP9 cells were incubated for 2224 hours in medium containing LPS with or without TSA, and then mRNA levels of TLR4 were determined. The levels were slightly changed by LPS but the changes were less affected by TSA (Fig. 5A). Furthermore, TNF-a-induced IL-6 production was also significantly enhanced by TSA in OP9 preadipocytes (Fig. 5B).
TSA did not affect LPS-induced phosphorylation of ERK1/ 2 and p38 MAPK Secondly, we examined whether TSA augmented the LPSinduced activation of MAPKs. As shown in Fig. 6A, LPS-induced IL-6 production was significantly inhibited by U0126, an inhibitor of MAPK kinase, and by SB203580, an inhibitor of p38 MAPK, whereas it was not reduced by SP600125, an inhibitor of JNK. Therefore, we focused on LPS-induced phosphorylation of ERK1/2, downstream of MAPK kinase, and p38 MAPK. LPS induced the phosphorylation of ERK1/2 and p38 MAPK within 5 minutes and this lasted at least 30 minutes (Fig. 6B), whereas TSA affected the LPS-induced phosphorylation of neither ERK1/2 ( Fig. 6C) nor p38 MAPK (Fig. 6D) at 15 minutes.

TSA enhanced LPS-induced NF-kB-dependent transcriptional activation
Thirdly, we examined whether TSA affected the LPS-induced activation of transcription factors. To clarify which transcription factors contributed to IL-6 production in preadipocytes, a reporter gene assay was carried out using a wild-type or mutated IL-6 promoter construct, which has AP-1, NF-kB and C/EBP-binding sites and CRE. As shown in Fig. 7A, the transcriptional activity of the wild-type IL-6 promoter construct was increased by LPS or PA stimulation, and completely abolished by mutation of the NF-kBbinding site. In contrast, the mutation of the AP-1 and C/EBPbinding sites and CRE in the IL-6 promoter less affected  transcriptional activity. These results were confirmed by using TPCA-1, an IKK-2 inhibitor. TPCA-1 pretreatment completely suppressed LPS-or PA-induced IL-6 production in a concentration-dependent manner ( Fig. 7B and C). Therefore, we focused on LPS-induced transcriptional activation of NF-kB. As shown in Fig. 7D, stimulation with LPS increased NF-kB-dependent transcriptional activity and the effect was enhanced by TSA pretreatment.

TSA increased acetylation of NF-kB p65 at lysine 310 and the duration of its nuclear translocation
Because acetylation at lysine 310 increases the transcriptional activity of p65, we next examined the effects of TSA on this acetylation. As shown in Fig. 8A, B and C, acetylation of p65 at lysine 310 was increased 1 hour after the treatment with TSA in preadipocytes but not in mature adipocytes. The nuclear translocation of p65 was temporarily induced by LPS (Fig. 8D) and reached a peak at 30 minutes after LPS stimulation (Fig. 8E). Moreover, although the nuclear translocation of p65 at 0230 minutes was less affected by TSA, that from 60 to 120 minutes was significantly enhanced (Fig. 8E). The enhancement was confirmed by the determination of levels of p65 in the nuclear and cytoplasmic fractions. At 60 minutes, LPS markedly increased the levels of p65 in the nuclear fraction and this was enhanced by TSA although TSA alone had no effect (Fig. 8F). These results suggested that the efflux of p65 was reduced by the treatment with TSA.

Discussion
In the present study, we clarified that TSA, a HDAC inhibitor, enhanced the expression of inflammatory proteins and NF-kBdependent transcriptional activity in OP9 preadipocytes. These results might be caused by increases in the acetylation of NF-kB p65 at lysine 310 and duration of the nuclear translocation of NF-kB. On the other hand, TSA did not affect LPS-induced IL-6 production and acetylation of NF-kB p65 at lysine 310 in mature adipocytes.
HDAC inhibitors did not themselves induce IL-6 production, but they increased LPS-induced IL-6 production in OP9 preadipocytes (Fig.3). The enhancement of TSA was not dependent on TLR4. Although TSA treatment increased the mRNA levels of TLR4 in embryonic stem cells and intestinal epithelial cells [35,36], it did not do so in OP9 preadipocytes (Fig. 5A). In addition, TNFa-induced IL-6 production was also enhanced by TSA (Fig. 5B), indicating that TSA affected the expression and activities of signaling molecules and/or transcription factors.
Inflammatory stimulation can induce the phosphorylation of MAPKs, increasing their activity and the subsequent expression of inflammatory proteins. We therefore focused on MAPKs as candidates for proteins whose activity is affected by TSA. HDAC inhibitors affect the phosphorylation of MAPKs in several cell lines. For example, valproic acid induced the phosphorylation of p38 MAPK but not ERK1/2 or JNK in microglial cells and lymphocytes [37,38]. However, we found that, in OP9 preadipocytes, the LPS-induced phosphorylation of ERK1/2 and p38 MAPK were not affected by TSA treatment whereas the production of IL-6 was reduced by a MEK inhibitor and a p38 MAPK inhibitor (Fig. 6).
We next focused on transcription factors including NF-kB, which is a key regulator for inflammatory protein expression. In fact, the experiments using reporter gene constructs containing a wild-type or mutated IL-6 promoter indicated that NF-kB strongly contributed to LPS-and PA-induced IL-6 production in preadipocytes (Fig. 7A). These results were confirmed by using an IKK-2 inhibitor (Fig. 7B and C). More importantly, we found that TSA enhanced LPS-induced NF-kB-dependent transcriptional activation in preadipocytes (Fig. 7D). Acetylation of the p65 subunit of NF-kB was shown to be critical to the regulation of NF- kB's activity by affecting transcriptional activity and the duration of nuclear translocation. The acetylation of p65 is regulated by HATs such as p300 and CBP and by HDACs, especially HDAC3 [39][40][41][42]. For example, acetylation of p65 at lysine 310 is associated with an increase in transcriptional activity without affecting the translocation into the nucleus. On the other hand, acetylation of p65 at lysine 218 or 221 leads to an increase in the duration of nuclear translocation of NF-kB by attenuating its interaction with IkB. We clarified that both the acetylation of p65 at lysine 310 and the LPS-induced translocation of NF-kB into the nucleus were enhanced by treatment with TSA (Fig. 8). These results suggested that the increase in the nuclear translocation and transcriptional activity of NF-kB contribute to a rise in NF-kB-dependent transcriptional activity and subsequent expression of inflammatory proteins.
The expression of COX-2, iNOS and MCP-1, all enhanced by TSA as described in Fig. 4, was regulated by NF-kB but other transcription factors also regulate them separately. For example, the expression of MCP-1 is regulated by Sp1 in addition to NF-kB [43]. Because acetylation of Sp1 is associated with a decrease in its DNA-binding activity [44], TSA might induce hyperacetylation of Sp1 and act negatively on LPS-induced expression. Thus, modulation of the acetylation of transcription factors other than NF-kB was also involved in the HDAC inhibitor-induced enhancement of gene expression. In the future, there is a need to investigate the effects of HDAC inhibitors on the activity of transcription factors other than NF-kB in pre/mature adipocytes. The OP9 cell line was established from the calvaria of newborn mice genetically deficient in functional macrophage colonystimulating factor [45], and is used as a pre/mature adipocyte model [28]. It is known that, during adipocyte differentiation, the expression of various proteins including transcription factors changes dynamically [28,46]. Notably, an increase in PPARc induces the gene expression of several adipocyte-specific proteins (ex. adiponectin) and other transcription factors (ex. C/EBPa) thus transforming the cell into the characteristic lipid-rich mature adipocyte [47]. Based on the results in Fig. 1, we selected an incubation period of 4 days for the IO-based method and 6 days for the TGZ-based method, as appropriate for differentiation. Interestingly, the enhancing effects of TSA on LPS-induced IL-6 production were observed in preadipocytes but not mature adipocytes ( Fig. 2 and 3). Treatment with TSA increased acetylation of p65 at lysine 310 in preadipocytes but not mature adipocytes (Fig. 8A, B and C), which might be one cause of the difference in the effects of HDAC inhibitors. In adult humans, preadipocytes account for 20240% of cells in white adipose tissue [48] and produce inflammatory proteins via TLR4 signaling as well as infiltrated macrophages [49][50][51]. Moreover, LPS-induced IL-6 production was much lower in mature adipocytes than preadipocytes [50,51]. These findings indicate that preadipocytes play a greater role in chronic inflammation in adipose tissue than do mature adipocytes.
Some reports have shown that HDAC inhibitors suppressed inflammatory responses in macrophages at relatively high concentrations [27,52,53]. We also confirmed that TSA and Figure 8. Effects of TSA on acetylation and LPS-induced nuclear translocation of NF-kB p65. OP9 preadipocytes (A) and mature adipocytes induced to differentiate by IO for 4 days (B) or TGZ for 6 days (C) were incubated for 1 hour with TSA (10 nM). The acetylated p65 was detected by Western blotting. The mean value at time 0 (vehicle) is set to 1.0. Data represent the mean 6 S.E.M. (n = 325). OP9 preadipocytes were preincubated for 1 hour in the presence or absence of TSA (10 nM), and then stimulated with LPS (1 mg/ml) for the period indicated. p65 in the cells was immunostained (green) and the nucleus was labeled with DAPI (blue). A representative image of immunostaining of the cells stimulated for 60 minutes with LPS (D). The nuclear and cytoplasmic ratio of p65-associated fluorescence intensity is indicated as the average for 4 randomly selected fields (E). The values at time 0 (vehicle) and time 30 (LPS only) are set to 0 and 100, respectively. Data represent the mean 6 S.E.M. (n = 4). The cells stimulated for 60 minutes with LPS were fractionated to cytoplasm and nucleus, and then p65 in these fractions was detected by Western blotting (F). The value for the control is set to 1.0. Data represent the mean 6 S.E.M. (n = 4). ***P,0.001, **P,0.01, *P,0.05 between the indicated groups. doi:10.1371/journal.pone.0059702.g008 SAHA decreased LPS-induced IL-6 production in RAW264 macrophages (data not shown). The proposed mechanism by which HDAC inhibitors exert their effects is interference with RNA polymerase II's recruitment to inflammatory genes [54]. In contrast, in microglial cells, HDAC inhibitors enhanced the production of inflammatory proteins [24] as consistent with our findings in preadipocytes. Thus HDAC inhibitors exert negative and positive effects on the expression of inflammatory proteins probably dependent on types of cells and on concentrations. The precise mechanisms remained to be elucidated.
The enhancement of LPS-and PA-induced inflammatory protein expression in preadipocytes is important to the maintenance of glucose homeostasis. Chronically elevated IL-6 levels in obesity have been considered to contribute to systemic insulin resistance by inducing the expression of suppressor of cytokine signaling 3 in adipose tissue and liver [7,55,56]. In contrast, it was recently reported that IL-6, especially produced temporarily in response to acute stimuli such as exercise, improved glucose tolerance in skeletal muscle, liver and intestinal L cells [57][58][59].
Thus the roles of IL-6 in glucose metabolism might differ with the duration and extent of IL-6 production.
In conclusion, we clarified that HDAC inhibitors enhanced the production of inflammatory cytokines in preadipocytes but not in adipocytes by affecting NF-kB through acetylation of the p65 subunit. These findings shed new light on the roles of HDACs in preadipocytes in the production of inflammatory proteins and on the regulation of systemic glucose homeostasis.