Long Non-coding RNA H19 Inhibits Adipocyte Differentiation of Bone Marrow Mesenchymal Stem Cells through Epigenetic Modulation of Histone Deacetylases

Bone marrow mesenchymal stem cells (BMSCs) exhibit an increased propensity toward adipocyte differentiation accompanied by a reduction in osteogenesis in osteoporotic bone marrow. However, limited knowledge is available concerning the role of long non-coding RNAs (lncRNAs) in the differentiation of BMSCs into adipocytes. In this study, we demonstrated that lncRNA H19 and microRNA-675 (miR-675) derived from H19 were significantly downregulated in BMSCs that were differentiating into adipocytes. Overexpression of H19 and miR-675 inhibited adipogenesis, while knockdown of their endogenous expression accelerated adipogenic differentiation. Mechanistically, we found that miR-675 targeted the 3′ untranslated regions of the histone deacetylase (HDAC) 4–6 transcripts and resulted in deregulation of HDACs 4–6, essential molecules in adipogenesis. In turn, trichostatin A, an HDAC inhibitor, significantly reduced CCCTC-binding factor (CTCF) occupancy in the imprinting control region upstream of the H19 gene locus and subsequently downregulated the expression of H19. These results show that the CTCF/H19/miR-675/HDAC regulatory pathway plays an important role in the commitment of BMSCs into adipocytes.

Overexpression of H19 and miR-675 inhibits adipocyte differentiation. LncRNA H19 was located both in nucleus and cytoplasm, and it was enriched in the cytoplasmic fraction (Supplementary Figure S1A). To determine whether H19 and miR-675 directly affect adipocyte differentiation, lentiviruses were used to overexpress H19 and miR-675 in BMSCs. The efficiency of lentiviral transduction was > 90% and the expression of H19 and miR-675 was significantly upregulated by > 8-fold (Supplementary Figure S1B), as described previously 15 . After induction to the adipogenic lineage, H19 and miR-675 significantly inhibited intracellular lipid accumulation as indicated by Oil red O staining ( Fig. 2A). The mRNA and protein expression of the adipocyte-specific factors PPARγ, C/EBPα, and FABP4 was significantly reduced by the overexpression of H19 and miR-675 (Fig. 2B,C). To determine whether H19 acts on adipogenesis via miR-675, we designed a mutant H19 that carried mutation in the sequences of miR-675. The inhibition of adipocyte differentiation induced by H19 was abrogated through mutation of miR-675, as indicated by Oil red O staining and the expression of adipocyte-specific genes ( Fig. 2A-C).
Knockdown of H19 and miR-675 promotes adipocyte differentiation. To further confirm the effects of H19 and miR-675 on adipocyte differentiation, we knocked down the endogenous H19 and miR-675 in BMSCs using lentivirus transfection. To control for potential off-target shRNA effects, we used two different shRNA sequences targeting H19. The expression of H19 and miR-675 was significantly reduced by ~70% (Supplementary Figure S1B), as described previously 15 . The inhibition of H19 and miR-675 promoted adipocyte formation as shown by Oil red O staining (Fig. 3A), and knockdown of H19 and miR-675 also significantly upregulated the mRNA and protein expression of the adipogenic marker genes PPARγ, C/EBPα, and FABP4 (Fig. 3B,C). HDACs 4, 5, and 6 are directly targeted by miR-675. HDACs respond to signals that regulate a broad and complex array of physiological processes, including adipocyte differentiation and metabolism 18 . Our previous study showed that H19 and miR-675 reduce the expression of HDACs 4 and 5 15 , but the mechanism remains unclear. So, we further assessed the expression of HDACs 1-6 in BMSCs overexpressing miR-675. Compared with the negative control, miR-675 substantially reduced the mRNA and protein levels of class II HDACs 4, 5, and 6, and slightly inhibited the expression of HDAC 1 (by ~20%), whereas no effect on the expression of HDACs 2 and 3 was found (Fig. 4A,B). Consistently, the ectopic overexpression of H19 downregulated the expression of HDACs 4, 5, and 6 in BMSCs, and the repression was relieved through mutation of miR-675 sequences (Fig. 4C).

Knockdown of HDACs 4, 5, and 6 inhibits adipocyte differentiation.
To determine the role of HDACs 4, 5, and 6 in adipogenesis, we first measured their expression patterns during adipocyte differentiation. The expression of HDACs 4-6 was gradually upregulated with the highest expression on day 11 (Fig. 6A,B), which was inversely correlated with H19 expression during adipogenesis. We then used specific small-interfering RNAs (siRNAs) to suppress their endogenous expression. The specific siRNAs were transfected into BMSCs in growth medium, which was changed to adipogenic-differentiation medium on day 1. On day 3, the siRNAs were transfected again, and the cells were harvested on day 7. Successful knockdown of HDACs 4, 5, and 6 ( Fig. 6D) inhibited the adipocyte differentiation of BMSCs as indicated by Oil red O staining (Fig. 6C), and the mRNA and protein expression of adipogenic genes PPARγ, C/EBPα, and FABP4 was also significantly downregulated ( Fig. 6E,F).
Trichostatin A (TSA) reduces CTCF occupancy in the H19 imprinting control region (ICR) and downregulates H19 expression. H19 is an imprinted gene, and its expression is regulated by chromatin structure and epigenetic mechanisms 20,21 . Thus, we sought to determine the epigenetic effect of HDACs on the expression of H19 and miR-675. We treated BMSCs with TSA (400 nM), an HDAC inhibitor, for 3, 7, and 14 days. This treatment significantly suppressed H19 expression in a time-dependent manner. The expression of H19 was significantly reduced after 3 days, while further treatment caused further reduction and H19 expression was suppressed to a low level after 7 days (Fig. 7A). The expression of miR-675 was also downregulated after TSA treatment, and displayed a pattern similar to H19 expression (Fig. 7A).
H19 expression is controlled by CTCF binding to the ICR upstream of the H19 gene. The CTCF protein promotes enhancer function at the H19 promoter ( Fig. 7B) 22,23 . To determine whether TSA treatment leads to changes in the CTCF-binding status at the H19 ICR, we analyzed the levels of CTCF occupancy in this region using chromatin immunoprecipitation (ChIP). Indeed, endogenous CTCF directly interacted with the H19 ICR in BMSCs. And following TSA treatment, the occupancy of CTCF protein in the H19 ICR was reduced (Fig. 7B).
TSA inhibits the enzymatic HDAC activity without significant change of their expression (Supplementary Figure S2). To confirm the role of HDACs in H19 transcription, we applied siRNA gene knockdown studies and found that knockdown of HDACs 4-6 resulted in a phenotype similar to that observed after TSA treatment. H19 expression was reduced by HDAC 4-6 knockdown, and CTCF occupancy in H19 ICR was also moderately reduced with significant change in HDAC6 knockdown group (Fig. 7C).

Discussion
In this study, we demonstrated that lncRNA H19 and H19-derived miR-675 were significantly downregulated in BMSCs that were differentiating into adipocytes. Overexpression of H19 and miR-675 in BMSCs inhibited adipogenesis, while knockdown of their expression accelerated adipogenic differentiation. These phenomena indicate that H19 plays key roles in the process of BMSC adipogenesis. A few lncRNAs, such as ADINR 24 , PU.1-as 25 , and NEAT1 26 , have been found to directly participate in the genetic control of adipogenic differentiation in adipose-derived stem cells. However, the stem cells from different tissues differ in their metabolic activity and ability to differentiate 27,28 . The roles of lncRNAs in BMSC adipogenesis are still largely unknown. It has been reported that when the balance between adipogenesis and osteogenesis is disturbed and BMSCs tend to differentiate into adipocytes rather than osteoblasts, marrow fat progressively accumulates and bone loss occurs [4][5][6] . Several studies have identified regulators of the switch between osteogenesis and adipogenesis in BMSCs, such as miR-188 8 , Maf 29 , Ezh2, and Kdm6a 30 . Here, we demonstrated that H19 plays inhibitory roles in adipogenesis, while our previous study showed that H19 promotes osteoblast differentiation 15 . Thus, H19 seems to function in the switch of osteoblast and adipocyte differentiation of BMSCs. In a preclinical mouse model, BMSC-specific inhibition of miR-188 by intra-bone marrow injection of aptamer-antagomiR-188 increases bone formation and decreases bone marrow fat accumulation 8 . Our studies indicating a role of lncRNA H19 in the shift from adipogenesis to osteogenesis in BMSCs may also provides a potential target for treating fatty marrow and bone loss.
LncRNAs exert their functions via diverse mechanisms, including co-transcriptional regulation, modulation of gene expression, scaffolding of nuclear or cytoplasmic complexes, and pairing with other RNAs 11 . Another means by which lncRNAs acquire functionality is by acting as precursors of miRNAs 31 or as sinks for pools of active miRNAs 32 to regulate transcripts targeted by that set of miRNAs. Several previous studies have shown that miR-675 confers functionality on H19 14,16,17,33 . Our results addressing the inhibitory effect of miR-675 on adipogenic differentiation and lost of function in H19-mut group suggest that miR-675 is at least partially responsible for the inhibition of adipogenesis induced by H19. However, as lncRNAs act via diverse mechanisms, H19 may interact with other molecules involved in the complex biology of adipogenic differentiation. In this study, we used PPARγ , FABP4, and C/EBPα as adipocyte markers, all of which were significantly upregulated during adipocyte differentiation. However, a previous study has showed that FABP4 is induced by PPARγ but negatively regulates PPARγ activity in macrophages and adipocytes 34 . There is no consensus on the interrelationship between FABP4 and PPARγ , and our results were consistent with some previous studies concerning adipocyte differentiation 24,[35][36][37] . The discrepancy of FABP4 expression in previous studies may be attributable to the different cells and tissues.
We found that miR-675 bound directly to the 3′ UTRs of class II HDACs 4-6 and downregulated their mRNA and protein levels, while minor effects were found on the expression of class I HDACs 1, 2, and 3. The expression of HDACs 4-6 was inversely correlated with H19 and miR-675 expression during adipocyte differentiation. The downregulation of class II HDACs 4, 5, and 6 inhibited the adipogenic differentiation of BMSCs. Consistent with our results, previous studies have demonstrated the essential role of class II HDACs in adipocyte differentiation. Growing evidence shows that class I is directly involved in regulation of cell growth and apoptosis, whereas class II members regulate differentiation processes 38 . Treatment of mesenchymal stem cells with pan-HDAC inhibitors, class II-specific inhibitors, or specific siRNAs targeting HDACs, attenuates adipogenesis and reduces the expression of adipocyte markers following the induction of differentiation [38][39][40][41][42] . H19 and miR-675 inhibited the adipogenic differentiation of BMSCs at least partially through the downregulation of HDACs.
HDAC inhibition reduced CTCF enrichment in the H19 ICR and downregulated H19 expression. H19 has been reported to be regulated by chromatin structure and epigenetic mechanisms, including DNA methylation, CTCF insulator, and enhancer activity 20,21 . Correct positioning of nucleosomes within the ICR is required for CTCF stably binding, which promotes enhancer function at the H19 promoter 22,23 . TSA treatment reduced the occupancy of the CTCF protein in the H19 ICR and abolished the boundary activity of the ICR, thereby downregulating H19 expression. There is evidence that regional changes in acetylation within the promoter of H19 occur after TSA treatment, and H19 expression is significantly reduced 43 . And there is also evidence that HDACs from nuclear extracts are bound by the CTCF zinc-finger domain, and CTCF may function by associating with the HDAC complex 44 . From these results, it is not clear whether the reduced CTCF occupancy is causally related to the histone acetylation status of H19 promoter, but it implies the role of HDACs in this lncRNA transcription. Also, knockdown of HDACs 4-6 resulted in the similar but weaker phenotype compared to TSA treatment, indicating that other histone or non-histone acetylation may be involved. The precise molecular mechanism that underlies the regulation of specific histone variations needs further investigation.
In conclusion, H19 and H19-derived miR-675 inhibits the adipocyte differentiation of BMSCs through the epigenetic modulation of HDACs; miR-675 directly targets HDACs 4, 5, and 6; and the inhibition of HDACs reduces the levels of CTCF occupancy in the H19 ICR and reduces H19 expression (Fig. 8). Further research may elucidate whether H19 and miR-675 modulate the shift of cell lineage commitment of BMSCs in vivo and provide a potential therapeutic target for bone marrow adiposity.  Methods Cell cultures and adipocyte differentiation. Primary BMSC lines from three donors were obtained from ScienCell (San Diego, CA, USA) and cultured at sub-confluent density in growth medium consisting of α -minimum essential medium supplemented with 10% fetal bovine serum and 1% antibiotics. All cell-based in vitro experiments were repeated in triplicate. For the adipocyte differentiation experiment, cells were allowed to become confluent for 1 day, and then cultured in standard growth medium supplemented with 50 nM insulin (Sigma-Aldrich, Saint Louis, MO, USA), 100 nM dexamethasone (Sigma-Aldrich), 0.5 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich), and 200 μ M indomethacin (Sigma-Aldrich). The adipogenic medium was changed every 2 days, and cells were harvested at the indicated times. The 293T cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and 1% antibiotics. H19, mutant H19 (H19-mut), or H19-targeting sequences (shH19-1 and shH19-2) were constructed as described previously 15 . Site-directed mutagenesis of the H19 sequences was performed using the Site-Directed Mutagenesis Kit (SBS Genetech, Beijing, China). It carried mutation in the sequences of miR-675, as follows: TGG TGC GGA GAG GGC CCA CAG TG was changed to TCC ACG CGA GAG GGC CCA CAG TG. A vacant lentiviral vector (NC) and a scrambled non-targeting vector (shNC) were used as control group. Recombinant lentiviruses harboring miR-675 or miR-675 inhibitor sequences (anti-miR-675) and the control vector (miR-NC) were obtained from Integrated Biotech Solutions Co. (Shanghai, China). Transfection of the BMSCs was performed by exposing them to dilutions of the viral supernatant in the presence of polybrene (5 μ g/ml) for 24 h.

Lentivirus infection. Recombinant lentiviruses harboring full-length
Oil red O staining. The cells were washed with phosphate-buffered saline (PBS) and fixed in 10% formalin for 30 min. The cells were then rinsed with 60% isopropanol. Oil red O (0.3%, Sigma-Aldrich) was added and incubated for 10 min with gentle agitation. After staining, the cells were washed with distilled water to remove unbound dye, visualized by light microscopy, and photographed. For quantitative assessment, the Oil red O was eluted by 100% isopropanol and quantified by spectrophotometric absorbance at 520 nm against a blank (100% isopropanol).

RNA oligoribonucleotides and chemicals.
A chemically-modified double-stranded miR-675 mimic and the corresponding miRNA mimic control (mimic NC) were obtained from RiboBio Co. (Guangzhou, China). siRNAs targeting HDAC 4, 5, and 6 transcripts (siHDAC4-6) and the corresponding siRNA control (siNC) were from Integrated Biotech Solutions Co. The sequences are listed in Supplementary Table S4. TSA (Sigma-Aldrich) was dissolved in dimethyl sulfoxide. Transient transfection. Cells at 70-80% confluence were transfected with plasmids, miRNA mimics, or siRNAs using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions as described previously 45 . Reporter vectors. Predicted miR-675 target genes and their target binding sites were investigated using RNA22 software 19 . The reporter vectors were constructed by Integrated Biotech Solutions Co. Briefly, the 3′ UTRs of HDAC 4, 5, and 6 mRNA, containing the predicted miR-675 binding sites, were synthesized and cloned into a modified version of pcDNA3.1(+ ) containing a firefly luciferase reporter gene (a gift from Professor Brigid L.M. Hogan, Duke University, Durham, NC, USA) 46 , at a position downstream of the luciferase reporter gene as Scientific RepoRts | 6:28897 | DOI: 10.1038/srep28897 described previously 45 . Site-directed mutagenesis of selected putative seeding-sequence regions was performed using the Site-Directed Mutagenesis Kit from SBS Genetech. All constructs were confirmed by DNA sequencing.
Dual luciferase reporter assay. Luciferase assays were performed as described previously 45 . Briefly, 293T cells grown in 48-well plates were transfected with 100 nM miR-675 mimic or control, 40 ng luciferase reporter, and 4 ng pRL-TK, a plasmid expressing Renilla luciferase (Promega, Madison, WI, USA) using Lipofectamine 2000 (Invitrogen). The Renilla and firefly luciferase activities were measured 24 h after transfection using the Dual Luciferase Reporter Assay System (Promega). All luciferase values were normalized to those of Renilla luciferase and expressed as fold-induction relative to the basal activity.
RNA isolation and qRT-PCR. Total RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer's instructions and then reverse-transcribed into cDNA using the cDNA Reverse Transcription Kit from Applied Biosystems (Foster City, CA, USA). qRT-PCR was conducted using SYBR Green Master Mix on the ABI Prism 7500 real-time PCR System (Applied Biosystems) as described 45 . The following thermal settings were used: 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. The primers used for H19, miR-675, HDACs 1-6, PPARγ, C/EBPα, FABP4, U6 (internal control for miRNAs), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH, internal control for mRNAs and lncRNAs) are listed in Supplementary Table S4. The data were analyzed using the 2 −ΔΔCt relative expression method as described previously 45 . Cell fractionation. Cytoplasmic and nuclear RNAs were fractionated using a Nuclei Isolation Kit (KeyGEN, Nanjing, China). Briefly, cells were harvested in lysis buffer, treated with Reagent A, incubated on ice for 15 min, followed by centrifugation at 4 °C. The pellet was then resuspended in lysis buffer followed by centrifugation. The supernatant was transferred to a new tube as the cytoplasmic fraction; the pellet was resuspended in Medium Buffer A and then added to a new tube with Medium Buffer B, followed by centrifugation at 4 °C. The supernatant was saved as the cytoplasmic fraction. The pellet was used as the nuclear fraction. RNA was extracted from both fractions using TRIzol.
Western blot analysis. Western blot analysis was performed as described previously 45 . Briefly, cells were harvested, washed with PBS, and lysed in RIPA buffer. Proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Primary antibodies against PPARγ (Cell Signaling Technology, Beverly, MA, USA), HDACs 1-6 (Cell Signaling Technology), C/ EBPα (HuaxingBio Science, Beijing, China), FABP4 (HuaxingBio Science), and GAPDH (Abcam, Cambridge, UK) were diluted 1:1,000. The intensities of the bands obtained by Western blot analysis were quantified using ImageJ software (http://rsb.info.nih.gov/ij/). The background was subtracted, and the signal of each target band was normalized to that of the GAPDH band.
ChIP assay. ChIP assays were performed using the EZ-Magna ChIP assay kit (Merck Millipore, Darmstadt, Germany) according to the manufacturer's instructions. Briefly, cells were washed with PBS and cross-linked with 1% formaldehyde for 10 min. Chromatin was sonicated on ice to generate chromatin fragments of 500-2000 bp. Then the DNA-protein complexes were isolated using antibodies against CTCF (Cell Signaling Technology) or isotype IgG (Cell Signaling Technology). The protein/DNA complexes were then eluted and reverse cross-linked. Input control DNA or immunoprecipitated DNA was quantified by qRT-PCR, using SimpleChIP Human H19/IGF2 ICR Primers (Cell Signaling Technology). Relative enrichment was calculated as the amount of amplified DNA normalized to the input and relative to values obtained after IgG immunoprecipitation. Statistical analysis. Statistical analyses were performed using SPSS version 16.0 (SPSS, Chicago, IL, USA).
All data are expressed as mean ± standard deviation (SD). Differences between groups were analyzed using Student's t-test. In cases of multiple-group testing, one-way analysis of variance was conducted. A two-tailed value of P < 0.05 was considered statistically significant.