Epigenetic regulation of adipogenesis by PHF2 histone demethylase

Abstract

PHF2 is a JmjC family histone demethylase that removes the methyl group from H3K9me2 and works as a coactivator for several metabolism-related transcription factors. In this study, we examined the in vivo role of PHF2 in mice. We generated Phf2 floxed mice, systemic Phf2 null mice by crossing Phf2 floxed mice with CMV-Cre transgenic mice, and tamoxifen-inducible Phf2 knockout mice by crossing Phf2 floxed mice with Cre-ERT2 transgenic mice. Systemic Phf2 null mice had partial neonatal death and growth retardation and exhibited less adipose tissue and reduced adipocyte numbers compared with control littermates. Tamoxifen-induced conditional knockout of PHF2 resulted in impaired adipogenesis in stromal vascular cells from the adipose tissue of tamoxifen-inducible Phf2 knockout mice as well as of Phf2 knocked-down 3T3-L1 cells. PHF2 interacts with CEBPA and demethylates H3K9me2 in the promoters of CEBPA-regulated adipogenic genes. These findings suggest that PHF2 histone demethylase potentiates adipogenesis through interaction with CEBPA in vivo. Taken together, PHF2 may be a novel therapeutic target in the treatment of obesity and the metabolic syndrome.

FIG. 1.

Strategic scheme for targeted disruption of mouse Phf2. A: Targeting strategy with positive/negative selection. Strategy of genomic Southern blotting in the screening for homologous recombinant embryonic stem cell clones is also included. E5, E6, and E7 represent exon 5, exon 6, and exon 7 of Phf2, respectively. B, H, E, and N represent BglII, HindIII, EcoRI, and NsiI cut sites, respectively. P1, P2, P3, P4, and P5 represent locations of primers used in C and D. , the LoxP sites; , Frt sites. B: Southern blotting analysis of targeted embryonic stem cell clones. Restriction enzymes used for screening recombination events with probe A were BglII and HindIII. An 8.3-kb fragment in WT and a 5.3-kb fragment after homologous recombination were expected with probe A. Restriction enzymes used for screening recombination events with probe B were EcoRI and NsiI. A 9.0-kb fragment in WT and a 7.2-kb fragment after homologous recombination were expected with probe B. C: To detect the presence of the LacZ allele (Z) and the WT allele (+), primers P1, P2, and P3 were used. The PCR bands of the WT allele (242 bp) and the LacZ allele (495 bp) are indicated. D: To detect the presence of the floxed allele (fl) and the WT allele (+), primers P4 and P5 were used. The PCR bands of the WT allele (162 bp) and the floxed allele (245 bp) are indicated. E: Western blot analysis of PHF2 protein expression in Phf2^Z/Z mice. Extracts of mouse embryonic fibroblasts from WT or Phf2^Z/Z were immunoprecipitated and detected with anti-PHF2 antibody. WT, wild type.

FIG. 2.

Physiological features of systemic Phf2 knockout mice. A: Genotypes of progeny of crosses between Phf2^Z/+ at birth. B: Genotypes of progeny of crosses between Phf2^Z/+ at 2 weeks of age. C: Survival rate of progeny of crosses between Phf2^Z/+. D: Growth curves of wild-type, heterozygous, and homozygous Phf2 knockout mice until 10 days after birth. E: Growth curves of male mice and female mice of indicated genotypes between 2 and 10 weeks of age. ◇, wild-type mice (+/+); ■, heterozygous Phf2 knockout mice (Z/+); ▲, homozygous Phf2 knockout mice (Z/Z). *P < 0.05; **P < 0.01 compared with wild type.

FIG. 3.

Assessment of adipose tissue of systemic Phf2 knockout mice. A: Ratio of body weight and nasoanal length or normalized tissue weight of male Phf2^Z/Z to WT littermates at 5 weeks of age. Tissue weights were normalized to body weights (n = 6). B: Weights of epididymal (epi) WAT, subcutaneous (sub) WAT, and mesenteric (mes) WAT of male Phf2^Z/Z KO mice and WT littermates at 5 weeks of age (n = 3). C: Weights of gonadal WAT of female Phf2^Z/Z KO mice and WT littermates at 5 weeks of age (n = 4). D: Mean adipocyte areas of epididymal WAT from Phf2^Z/Z KO mice and WT littermates (n = 5). High-magnification micrographs of WAT are shown. E: Adipocyte number in epididymal fat pads of Phf2^Z/Z KO mice and WT littermates (n = 7). F: Real-time qPCR analysis of adipocyte marker genes and Phf2 of Phf2^Z/Z KO mice and WT littermates (n = 3). *P < 0.05; **P < 0.01 compared with WT. KO, knockout; WT, wild type. (A high-quality color representation of this figure is available in the online issue.)

FIG. 4.

Effects of PHF2 knock down on adipogenesis. A–C: SVCs from Cre-ERT2; Phf2^fl/fl (fl/fl) or Cre-ERT2 (+/+) were treated with or without 4-OHT and differentiated into adipocytes by treatment with insulin, dexamethasone, and isobutylmethylxanthine. Image (A) and quantification (B) of oil red O staining and the results of real-time qPCR analysis of adipocyte marker genes (C) are shown (n = 3). D–F: 3T3-L1 cells infected with retroviruses containing either pSuper-retro-shLacZ or pSuper-retro-shPHF2 were differentiated into adipocytes. Image (D) and quantification (E) of oil red O staining and the results of real-time qPCR analysis of adipocyte marker genes (F) are shown (n = 3). *P < 0.05; **P < 0.01 compared with control.

FIG. 5.

Association of PHF2 with CEBPA. A: FLAG-CEBPA or FLAG-PPARG was transfected into HEK293T cells. Cells were harvested and immunoprecipitated with anti-FLAG antibody and detected with anti-PHF2 antibody or anti-FLAG antibody. B: FLAG-PHF2 was transfected into 3T3-L1 adipocytes. Cells were harvested and immunoprecipitated with anti-FLAG antibody and detected with anti-CEBPA antibody or anti-FLAG antibody. C: 3T3-L1 cells were fixed in formaldehyde on day 4 after differentiation, after which chromatin samples were subjected to ChIP analysis with indicated antibodies and amplified with primers toward indicated loci (n = 3). D and E: 3T3-L1 cells stably transfected with pSuper-retro-shLacZ or pSuper-retro-shPHF2 were differentiated into adipocytes and subjected to ChIP analysis with anti-CEBPA (D) or anti-H3K9me2 antibody (E), with primers toward the indicated loci (n = 3). *P < 0.05; **P < 0.01 compared with control. NC, negative control region.