HMGA1 is a novel transcriptional regulator of the FoxO1 gene

Abstract

Purpose: The forkhead transcription factor (FoxO1) is a master transcriptional regulator of fundamental cellular processes ranging from cell proliferation and differentiation to inflammation and metabolism. However, despite its relevance, the mechanism(s) underlying FoxO1 gene regulation are largely unknown. We have previously shown that the chromatin factor high-mobility group A1 (HMGA1) plays a key role in the transcriptional regulation of glucose-responsive genes, including some that are involved in FoxO1-mediated glucose metabolism. Here we investigated the impact of HMGA1 on FoxO1 gene expression.

Methods: FoxO1 protein and gene expression studies were performed by Western blot analysis combined with qRT-PCR of material from human cultured cells and EBV-transformed lymphoblasts, and from primary cultured hepatocytes from wild-type and Hmga1 ^-/- mice. Reporter gene assays and chromatin immunoprecipitation for binding of HMGA1 to the endogenous FoxoO1 locus were performed in cells overexpressing HMGA1 and in cells pretreated with siRNA targeting HMGA1.

Results: HMGA1 increased FoxO1 mRNA and protein expression in vitro, in cultured HepG2 and HEK-293 cells by binding FoxO1 gene promoter, thereby activating FoxO1 gene transcription. Forced expression of HMGA1 in primary cultured hepatocytes from Hmga1 ^-/- mice and in EBV-transformed lymphoblasts from subjects with reduced expression of endogenous HMGA1 increased FoxO1 mRNA and protein levels.

Conclusion: These findings may contribute to the understanding of FoxO1 gene regulation and its role in metabolism.

Keywords: DNA/chromatin interaction; FoxO1; Gene transcription; HMGA1; Insulin signaling.

Fig. 1

FoxO1 gene expression is induced by HMGA1. a pcDNA3-HMGA1 effector vector was transfected into HepG2 cells, either in the absence or presence of siRNA against HMGA1, or a non-targeting control siRNA, and endogenous FoxO1 mRNA expression was measured thereafter. b Increasing amounts (0, 0.5, 1 µg) of pcDNA3-HMGA1 effector vector were transfected into HEK-293 cells, and FoxO1 mRNA was measured as in a. Results are means ± S.E.M. of three independent experiments, each in triplicate. *P < 0.05 vs. control (first column in each condition). Representative Western blots (WB) of FoxO1 and HMGA1 out of three independent experiments for each condition are shown in the autoradiograms. Bar graphs above the gel panels are derived from densitometric scanning of WBs, using the ImageJ software program. *P < 0.05 vs. controls (first columns in each condition). Sp1, control of nuclear protein loading

Fig. 2

HMGA1 regulates FoxO1 gene transcription. a FoxO1-Luc reporter vector was transfected into HepG2 cells, either in the absence or presence of siRNA against HMGA1, or a nontargeting control siRNA, and Luc activity was measured 72 to 96 h later. b FoxO1-Luc reporter vector was transfected into HEK-293 cells, with increasing amounts (0, 0.5, 1 µg) of pcDNA3-HMGA1 effector vector, in the absence or presence of 10 µM distamycin A, and Luc activity was measured as in a. In both experimental conditions, Luc activity from the reporter plasmid was normalized by the renilla Luc activity produced from a pRL Renilla-Luc control vector cotransfected as an internal control. White bar, mock (no DNA); black bar, (pGL3 basic vector, without the FoxO1 promoter). Results are means ± S.E.M. of three independent experiments, each in triplicate. *P < 0.05 compared with nontargeting (control) siRNA; **P < 0.05 vs. the FoxO1-Luc, in the absence of a HMGA1 effector vector, which is assigned an arbitrary value of 1. Representative WBs of HMGA1 out of three independent experiments for each condition are shown in the autoradiograms. Bar graphs above the gel panels are derived from densitometric scanning of WBs, using the ImageJ software program. *P < 0.05 vs. control (first column in each condition). Sp1, control of nuclear protein loading

Fig. 3

HMGA1 binds FoxO1 promoter in living cells. ChIP of the FoxO1 promoter gene in HepG2 cells untreated or pretreated with either distamycin A or HMGA1 siRNA. ChIP was carried out using an anti-HMGA1 specific antibody (Ab). qRT-PCRs of ChIP-ed samples are shown in each condition. A representative ChIP assay out of at least three independent experiments is presented. *P < 0.05 vs. nontargeting control siRNA (slashed bar)

Fig. 4

FoxO1 gene expression in primary culture cells. FoxO1 mRNA abundance was measured in primary cultured hepatocytes from both Hmga1 ^+/+ (black bar) and Hmga1 ^–/– (gray bar) mice, before and after transfection of the cells with the pcDNA3-HMGA1 expression vector. Results are means ± S.E.M. of five independent experiments. *P < 0.05 vs. Hmga1 ^+/+ hepatocytes. Representative WBs are shown in each condition. Bar graph above the gel panel is derived from densitometric scanning of anti-HMGA1 WB, using the ImageJ software program. Sp1, control of nuclear protein loading

Fig. 5

FoxO1 expression in EBV-transformed lymphoblasts from subjects expressing variable amounts of HMGA1. mRNA and protein expression (WB) for both HMGA1 and FoxO1 were concomitantly measured in cultured EBV-transformed lymphoblasts from normal subjects (WT, wild-type, n = 6) and subjects carrying HMGA1 gene variants (c.*369del and rs139876191; n = 3 and 4, respectively) that decrease HMGA1 protein expression [30, 34], either in the absence (−) or presence (+) of the effector vector for HMGA1. qRT-PCR of HMGA1 and FoxO1 mRNA levels are shown as percent of maximal value (WT, 100%). Results are means ± S.E.M. for three separate assays. *P < 0.05 vs. controls (WT). **P < 0.05 vs. untransfected cells in each variant group. Representative WBs for HMGA1 and FoxO1 are shown. Bar graphs above the gel panels are derived from densitometric scanning of WBs, using the ImageJ software program. *P < 0.05 vs. controls (WT); **P < 0.05 vs. untransfected cells in each variant group. Sp1, control of nuclear protein loading

Fig. 6

Binding of HMGA1 to FoxO1 promoter in subjects’ cells. The occupancy of the FoxO1 gene promoter by HMGA1 was measured by ChIP in EBV-transformed lymphoblasts from normal WT subjects (slashed bar, n = 3) and subjects carrying the HMGA1 c.*369del variant (gray bars, n = 3), using an anti-HMGA1 specific antibody (Ab), either before or after transfecting cells with the c.*369del variant with an HMGA1 effector vector. qRT-PCRs of ChIP-ed samples are shown in each condition. *P < 0.05 vs. WT subjects; **P < 0.05 vs. untransfected cells carrying the HMGA1 c.*369del variant. A representative ChIP assay is presented