LB-OR-5 PPARγ Binding Sites in Human White Adipose Tissue

Program: Late-Breaking Abstracts
Session: LB-OR-Late-Breaking Oral Session
Saturday, June 15, 2013: 11:30 AM-1:00 PM
Presentation Start Time: 12:30 PM
Room 122 (Moscone Center)
Raymond E Soccio*1, Eric R Chen1, Fenfen Wang1, Joanna R DiSpirito1, Heewoong Lim2, Kyoung-Jae Won2 and Mitchell A Lazar2
1Perelman School of Medicine at the University of Pennsylvania, 2Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA
The nuclear receptor PPARγ is the master regulator of fat cell development, and is implicated in type 2 diabetes both as the target of anti-diabetic thiazolidinedione drugs and in genome-wide association studies.  Using chromatin immunoprecipitation followed by deep sequencing (ChIP-Seq) to describe genome-wide binding regions (cistromes) for PPARγ in cultured human SGBS adipocytes, we previously found that only ~10% of binding sites from mouse 3T3-L1 adipocytes were retained at syntenic genomic locations in the human cells.  Deeper sequencing tripled the number of high confidence binding sites in each species but confirmed the low degree of binding site retention across species.  However, these cell culture models may not reflect the binding of PPARγ in intact adipose tissue.  To address this, PPARγ ChIP-seq was performed in subcutaneous white adipose tissue (WAT) from obese humans and lean C57Bl/6 mice.  Similar to the findings in cultured adipocytes, ~14% of the human adipose PPARγ cistrome was retained in mouse adipose tissue.  Comparing the WAT cistromes to cultured adipocytes, within each species ~40% of WAT binding regions were present in the respective cell culture model.  These data strongly indicated that the most clinically relevant PPARγ cistrome is not from animal or cell culture models, but from actual human fat.  Therefore, ChIP-seq was performed on four additional WAT samples, revealing marked variability among subjects.   The PPARγ cistrome of each subject overlapped by only ~30% with that of the initial subject, and in each cistrome more than half of identified binding sites were unique to that subject.  This variability is much greater than we observe in replicate WAT cistromes from inbred lean C57BL/6 male mice, likely due to technical challenges with obese human adipose tissue as well as heterogeneity in age, sex, body weight, genotype, and other clinical factors.  Analysis of the PPARγ cistromes from multiple subjects revealed three classes of human adipose PPARγ binding sites: “core” sites identified in nearly all subjects, which typically show strongest binding; “variable” binding sites that are found in multiple subjects, but are typically weaker than the core sites and not detected in every subject; and “sporadic” subject-specific binding sites, typically the weakest and of uncertain relevance.  Preliminary analysis of single nucleotide polymorphisms (SNPs) falling in PPARγ binding sites suggests that genotype may account for some of the observed variability, and these genetic differences have the potential to affect metabolic phenotypes.

Disclosure: MAL: Ad Hoc Consultant, Pfizer, Inc., Ad Hoc Consultant, BioU, LLC, Ad Hoc Consultant, Merck & Co., Ad Hoc Consultant, Madrigal Pharmaceuticals, Ad Hoc Consultant, Lycera, Corp., Board Member, Novartis Pharmaceuticals, Board Member, Eli Lilly & Company, Ad Hoc Consultant, Vanda Pharmaceuticals. Nothing to Disclose: RES, ERC, FW, JRD, HL, KJW

*Please take note of The Endocrine Society's News Embargo Policy at

Sources of Research Support: NIH K08 grant DK094968 awarded to RES; NIH R01 grant DK49780 award to MAL