Epigenetic Regulation of PLIN1 in Obese Women and its Relation to Lipolysis

Increased adipocyte lipolysis links obesity to insulin resistance. The lipid droplet coating-protein Perilipin participates in regulation of lipolysis and is implicated in obesity. In the present study we investigate epigenetic regulation of the PLIN1 gene by correlating PLIN1 CpG methylation to gene expression and lipolysis, and functionally evaluating PLIN1 promoter methylation. PLIN1 CpG methylation in adipocytes and gene expression in white adipose tissue (WAT) was quantified in two cohorts by array. Basal lipolysis in WAT explants and adipocytes was quantified by measuring glycerol release. CpG-methylation of the PLIN1 promoter in adipocytes from obese women was higher as compared to never-obese women. PLIN1 promoter methylation was inversely correlated with PLIN1 mRNA expression and the lipolytic activity. Human mesenchymal stem cells (hMSCs) differentiated in vitro into adipocytes and harboring methylated PLIN1 promoter displayed decreased reporter gene activity as compared to hMSCs harboring unmethylated promoter. Treatment of hMSCs differentiated in vitro into adipocytes with a DNA methyltransferase inhibitor increased levels of PLIN1 mRNA and protein. In conclusion, the PLIN1 gene is epigenetically regulated and promoter methylation is inversely correlated with basal lipolysis in women suggesting that epigenetic regulation of PLIN1 is important for increased adipocyte lipolysis in insulin resistance states.

PLIN1 in relation to lipolysis. We further studied adipocyte CpG-methylation of PLIN1 in relation to lipolysis in more detail. There were inverse correlations between PLIN1 mRNA and basal lipolysis in isolated adipocytes (Fig. 2a,c) and WAT explants (Fig. 2b,d) in both the explorative and validation cohorts. Methylation of all examined CpG-sites in the PLIN1 gene promoter and 5′ region in adipocytes was inversely related to PLIN1 mRNA, as well as WAT ex vivo and adipocyte basal lipolysis (Table 3). By contrast, only one of five examined CpG-sites in the PLIN1 gene body and 3′ UTR was nominally associated with these phenotypes.
To assess whether CpG-methylation of PLIN1 affects promoter activity, a luciferase reporter gene assay was used. hMSCs were transfected with a plasmid containing the PLIN1 promoter region cloned into a CpG free backbone vector (pCpGL-PLIN). Cells transfected with methylated pCpGL-PLIN plasmid displayed a significant and marked decrease (>60%) of luciferase activity in comparison to those transfected with unmethylated pCpGL-PLIN plasmid (P < 0.001). The methylated and unmethylated control plasmid showed no difference in reporter gene activity (Fig. 3). To further evaluate epigenetic regulation of PLIN1, hMSCs differentiated in vitro to adipocytes were treated with 50 or 200 µM of DNA methyltransferase inhibitor RG108 for 24 h. The higher concentration of RG108 (200 µM) was determined to be cytotoxic (results not shown), therefore all subsequent experiments were performed using 50 µM RG108. As a consequence of RG108 treatment, levels of PLIN1 mRNA (Fig. 4a) and Perilipin protein (Fig. 4b,c) were upregulated by about 140% and 30%, respectively. Demethylation activity of RG108 was confirmed by Global DNA methylation Imprint ® Methylated DNA Quantification Kit, and showed that treatment of adipocytes with RG108 decreased global DNA methylation by 4% and 24%, respectively (Fig. 4d). To confirm that RG108 demethylated PLIN1, two CpG sites in the promoter were selected for analysis by Pyrosequencing (cg08749443 and cg04998447). The results showed that 50 µM RG108 decreased methylation at both CpGs by 5% (Fig. 4e).

Discussion
We herein report that PLIN1, a key regulator of basal lipolysis, is subject to functional regulation by epigenetic modifications. It is demonstrated that CpG methylation of the PLIN1 gene promoter inhibits promoter activity in vitro, and that adipocyte promoter methylation is inversely correlated with PLIN1 mRNA levels in clinical cohorts. Furthermore, a positive correlation between PLIN1 CpG methylation and adipose basal lipolysis ex vivo was found. This indicates that epigenetic regulation is important for the regulation of lipolysis in human WAT.
Differential CpG methylation of PLIN1 was not evident in our original published global methylome analysis of obese versus never-obese women. This is most likely due to the use of a stringent threshold to adjust for multiple testing 16 . However, the validation of DMS in the PLIN1 promoter in the present study using two cohorts with array data, as well as an independent method (Pyrosequencing) clearly establishes the presence of differential methylation of PLIN1 in obesity. Furthermore, there is a positive correlation between PLIN1 CpG-methylation in intact WAT and BMI (see supplementary tables in Rönn T et al. 17 ). The reported relationship is weaker than the one observed here, which might be due to the fact that we study isolated adipocytes and hereby avoid the confounding effects of other cell types in intact WAT. By contrast CpG-methylation of PLIN1 has not been associated with BMI in reported epigenome-wide association studies on whole blood or leukocytes suggesting that the effect is specific to adipocytes or WAT where PLIN1 influences lipolysis 18,19 . We observed some correlation between CpG methylation of the PLIN1 promoter and age in the validation, but not in the explorative cohort. This is in agreement with reported finding for WAT that, although the average methylation of CpG-sites covering the genome is positively correlated with age, only a minor proportion of individual CpG-sites display significant correlation between methylation and age 17 . Specifically, CpG-methylation of PLIN1 did not correlate with age in the large cohorts of men and women studied by Rönn et al.
Increased lipolysis is implicated in insulin resistance, as well as in a number of other conditions including cachexia, hepatosteatosis, and cardiovascular disease 2, 20-22 . In addition, there is evidence for epigenetic dysregulation in these disease states [23][24][25] . Despite the importance of PLIN1 for lipolytic regulation 5 , the PLIN1 gene locus has not come out as a susceptibility locus in genome-wide association studies for e.g. obesity or insulin resistance [26][27][28] . The finding that CpG methylation controls the activity of the PLIN1 promoter thus shed new light on the regulation of adipocyte lipolysis and potentially why lipolysis is altered in metabolic diseases. The degree of global DNA demethylation, 4-24%, reported here is in the same range as has been reported previously when treating non-dividing cells with a global methyltransferase inhibitor 29 . Future work is needed to define how CpG methylation interacts with other known regulators of lipolysis 4 . In the present study methylation of all examined CpG-sites in the PLIN1 promoter were positively associated with obesity and lipolysis in the clinical cohorts. Transcriptional regulation of PLIN1 is incompletely defined; the transcription factors PPARG, NFkappaB and LXRA have been shown to regulate PLIN1 transcription [13][14][15] . The reported binding motifs for PPARG and LXRA do not contain any CpG-sites, nor the predicted binding site for NFkappaB. As regards the studied CpG-sites in the PLIN1 promoter, cg08749443 overlaps a predicted binding site for Sp1, whereas the other CpG-sites show no overlap with predicted transcription factor binding sites according to AliBaba using the public Transfac database to predict transcription factor binding matrices (http://www.gene-regulation.com/pub/databases.html). The results, thus, do not permit us to draw any conclusion if methylation of some part(s) of the promoter is more important than others.
What might in turn regulate CpG-methylation of the PLIN1 promoter? It is established that external factors such as diet and physical exercise influence the methylome [30][31][32] ; however, in the few available human intervention studies that have the DNA methylome as one outcome neither physical exercise nor high-fat overfeeding are associated with differential CpG methylation of PLIN1 30,32 . Thus, it remain to establish if and which behavioral factors influence CpG-methylation of the PLIN1 promoter. In addition, it is possible that metabolic factors such as fat mass (obesity) and enlarged fat cells are more important.
There are some limitations with the present study. We only examined women and one adipose region. Perilipin protein in WAT has been reported to be higher in men than in women 9 33 .
In conclusion, the PLIN1 gene is epigenetically regulated and promoter methylation is inversely correlated with basal lipolysis in women suggesting that epigenetic regulation of PLIN1 is important for increased adipocyte lipolysis in insulin resistant states such as obesity.

Methods
Subjects. The subjects in this study have been described before 16 . Briefly, the discovery cohort comprised fifteen obese women (body mass index (BMI) >30 kg/m 2 ) and fourteen never-obese healthy control women (BMI <30 kg/m 2 ) who were recruited in association with planned visits to our surgical units for gastric by-pass surgery because of obesity and through local advertisement for the purpose of studying WAT factors regulating body weight. Clinical data are presented in Table 1. All 14 never-obese women were healthy. Three of the obese women had type 2 diabetes, out of which two were treated with diet and metformin, and one subject with diet alone. CpG methylation of examined CpG-sites did not differ significantly between metformin treated and other women. For six CpG-sites methylation beta-values in the metformin treated women were within the interval defined by the average + SD for the women not treated with metformin; for remaining three CpG-sites the methylation beta-value of 1≥ subject was outside this region, but not in a consistent direction. Overall, this suggests that metformin does not influence methylation of PLIN1. Nine of the obese individuals were treated for hypertension and one patient had stable multiple sclerosis and did not receive any drugs. The women undergoing gastric by-pass surgery participated in a trial on the effect of bariatric surgery (NCT01785134 at www.clinicaltrials.gov) and were investigated before surgery. The study was approved by the regional ethics board in Stockholm and written Transcriptome analysis on WAT specimens was conducted for 18 of the above individuals (9 obese, and 9 never-obese). For the remaining subjects included in this study we did not have sufficient amount of WAT for transcriptome analysis.   . PLIN1 promoter methylation inhibits promoter activity. PLIN1 promoter activity is decreased after (hatched bar) versus without (black bar) DNA methylation by SssI methyltransferase. hMSCs were transfected with methylated and unmethylated pCpGL-PLIN1 plasmid. As negative control, cells were transfected with empty vector, pCpGL-basic. Each sample was prepared in quadruplicates and the experiment was repeated three times. Y axis is the ratio between firefly and renilla luciferase. Renilla luciferase is expressed from a second plasmid as an internal control. RLU = Relative luciferase units. ***P < 0.001.
For validation of PLIN1 results we studied an independent group of women with a wide variation in BMI (n = 69, age 40 ± 12 years, BMI 30.9 + 9.3 kg/m 2 ) who were examined in the same manner as the explorative cohort. Global transcriptome profiles was available on subcutaneous WAT from 57 of these women as reported 34 . Clinical examination. Participants were investigated at 8 AM after an overnight fast. All subjects had been weight stable (< ± 2 kg body weight change) during at least 6 months prior to investigation according to self-report Anthropometric measurements were performed followed by venous blood sampling. Plasma and serum were used for analysis of NEFA and other clinical chemistry variables as described 35 . An abdominal subcutaneous WAT biopsy was obtained by fine needle aspiration as described 36 .
WAT handling. The adipose tissue was brought to the laboratory, rinsed repeatedly in saline and visual blood vessels and cell debris were removed. Adipose tissue specimen (about two grams) were divided into portions, one of which was subjected to collagenase treatment to obtain isolated adipocytes as previously reported 37 . The mean weight and volume of these cells were determined as previously described 38,39 . 200 µl of packed isolated adipocytes and 300 mg unfractionated WAT pieces were frozen in liquid nitrogen and kept at −70 °C for subsequent DNA (cells) or RNA (tissue) preparation, whereas remaining tissue was used immediately for cell culture experiments.
Lipolysis assay. Basal lipolytic activity was determined in adipose tissue explants as described 40 . In brief, pieces of adipose tissue (200 or 300 mg) were incubated for 2 h (100 mg/ml) at 37 °C with air as the gas phase in Krebs-Ringer phosphate buffer (pH 7.4) supplemented with glucose (8.6 mmol/l), ascorbic acid (0.1 mg/ml) and bovine serum albumin (20 mg/ml). Glycerol release into the medium was measured using a sensitive bioluminescence method and expressed as amount of glycerol release per 2 h and 10 7 adipocytes. Adipocytes are the only adipose source of glycerol, which is an end product of lipolysis and only metabolized by adipose tissue to a minimal extent. Methodological experiment revealed that glycerol release in these type of experiments is linear with (e) The methyltransferase inhibitor RG108 decreased methylation of specific CpG-sites in the PLIN1 promoter in adipocytes. Methylation of cg08749443 and cg04998447 was determined by Pyrosequencing after adipocytes were treated with 50 µM RG108 (hatched bar) compared to non-treated control cells (black bar). The analysis was repeated twice, n > 3. ***P < 0.001, **P < 0.01, *P < 0.05.
incubation time for at least 4 h. Cross-sectional and longitudinal studies have demonstrated that basal lipolytic activity is strongly and negatively related to in vivo insulin sensitivity.
Basal lipolysis in isolated adipocytes was investigated using the same protocol as described above in diluted adipocyte suspensions (2% v/v).
Quantification of PLIN1 CpG methylation by array. PLIN1 CpG methylation in the explorative cohort was analyzed in DNA extracted from adipocytes in a previously reported dataset applying the Infinium Human Methylation 450 BeadChip assay (Illumina, San Diego, CA, USA) 16 . The beta value (β), which represents the ratio of intensities between methylated and unmethylated alleles, was used to quantify methylation at specific CpG loci. The β-values vary from 0 (no methylation) to 1 (100% methylation). Of 11 probes mapping to the PLIN1 gene, we excluded two containing SNPs with MAF >10% within the probes according to Illumina annotation.
CpG-methylation in adipocytes from the validation cohort was analyzed by EPIC arrays. Briefly, 500 ng of genomic DNA was bisulfate converted with EZ-96 DNA Methylation kit (Zymo Research, Irvine, CA, USA) and genome wide DNA methylation analysis was performed using the Infinium Human Methylation EPIC BeadChip (Illumina, San Diego, CA, USA). The laboratory procedures were performed according to the manufacturer's protocol. For analysis, visualization and extraction of methylation data, the GenomeStudio software version 2011.1 (Illumina Inc.) was used. The analysis reported here was limited to CpG-sites in the PLIN1 gene region that should be confirmed from the explorative cohort. A comprehensive analysis of the full EPIC data is ongoing.
Pyrosequencing assay. Five of the obese and nine of the never-obese women were used for validation of DMS by pyrosequencing; from remaining women we did not have any cells left. Genomic DNA was prepared from adipocytes using the QiAamp DNA Mini kit (Qiagen, Hilden, Germany). The DNA purity and quality was confirmed by A260/280 ratio >1. 8  Quantification of PLIN1 expression. PLIN1 mRNA in the clinical samples was quantified by microarray using Gene 1.1 ST arrays (Affymetrix, Santa Clara, CA, US) in 20 of the individuals in the explorative cohort, and by Gene 1.0 or 1.1 ST arrays in 57 of the women in the validation cohort; both groups formed parts of larger studies as described 16,34 . For remaining subjects we did not have sufficient amount of WAT for transcriptome analysis. Total RNA and DNA was extracted from cell culture samples using the AllPrep DNA/RNA Kit (Qiagen). Concentration and purity of nucleic acids was measured using a Nanodrop ND-1000 Spectrophotometer (Thermo Fisher Scientific, Lafayette, CO). cDNA synthesis was performed using iScript cDNA Synthesis kit (Bio-Rad Laboratories, Hercules, CA) together with random hexamer primers (Invitrogen, Carlsbad, CA). Assessment of PLIN1 mRNA levels was performed using SYBR Green Mix (Bio-Rad Laboratories) and primers: PLIN1 forward 5′-TGGAAACTGAGGAGAACAAG-3′ and reverse 5′-ATGTCACAGCCGAGATGG-3′. Expression was normalized to the internal reference gene 18S forward 5′-TGACTCAACACGGGAAACC-3′ and reverse 5′-TCGCTCCACCAACTAAGAAC-3′ using the ΔΔCt method 41 .
Construction of reporter vector. PLIN1 promoter was PCR amplified using KAPA HotStart ReadyMix (Kapa Biosystems, Wilmington, MA) from human genomic DNA (Roche, Basel, Switzerland) using primers, forward 5′-TATTGGATCCGTACAGCCCAGCACATTCACAACT-3′ and reverse 5′-TATTAAGCTTGCCCCAGGACCCCAACAC-3′ (Sigma Aldrich, Dorset, UK) and cloned into the pCpGL-basic vector (kindly provided by prof. M. Rehli, Regensburg, Germany) via BamHI and HindIII (Thermo Scientific) sites, which yielded the pCpGL-PLIN vector. The correct insertion of a construct was controlled by sequencing. The cloned genomic sequence covered 1,731 base pairs upstream of the PLIN1 transcription start site, and included all CpG-sites in the promoter region of PLIN1 whose methylation status was assayed by microarray.
In vitro methylation of plasmid DNA. Both plasmids, pCpGL-PLIN and pCpGL-basic (no insert), were methylated using SssI methyltransferase (New England Biolabs, Hitchin, UK) according to the manufacturer's recommendation. In brief, 10-15 µg of plasmid DNA was incubated with or without SssI methyltransferase (20 U/ µl; 2 U/µg DNA) in the presence of 640 µM S-Adenosylmethionine (SAM) (New England Biolabs) for four hours at 37 °C, with another 640 µM SAM being added after the first two hours of incubation. Plasmid DNA was purified using QIAquick PCR purification kit (Qiagen). Methylation of plasmid DNA was controlled by digestion using methylation sensitive restriction enzyme HpaII (Thermo Fisher Scientific) for four hours at 37 °C.