Differential roles of hypoxia-inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hypoxic gene regulation

Abstract

Transcriptional responses to hypoxia are primarily mediated by hypoxia-inducible factor (HIF), a heterodimer of HIF-alpha and the aryl hydrocarbon receptor nuclear translocator subunits. The HIF-1alpha and HIF-2alpha subunits are structurally similar in their DNA binding and dimerization domains but differ in their transactivation domains, implying they may have unique target genes. Previous studies using Hif-1alpha(-/-) embryonic stem and mouse embryonic fibroblast cells show that loss of HIF-1alpha eliminates all oxygen-regulated transcriptional responses analyzed, suggesting that HIF-2alpha is dispensable for hypoxic gene regulation. In contrast, HIF-2alpha has been shown to regulate some hypoxia-inducible genes in transient transfection assays and during embryonic development in the lung and other tissues. To address this discrepancy, and to identify specific HIF-2alpha target genes, we used DNA microarray analysis to evaluate hypoxic gene induction in cells expressing HIF-2alpha but not HIF-1alpha. In addition, we engineered HEK293 cells to express stabilized forms of HIF-1alpha or HIF-2alpha via a tetracycline-regulated promoter. In this first comparative study of HIF-1alpha and HIF-2alpha target genes, we demonstrate that HIF-2alpha does regulate a variety of broadly expressed hypoxia-inducible genes, suggesting that its function is not restricted, as initially thought, to endothelial cell-specific gene expression. Importantly, HIF-1alpha (and not HIF-2alpha) stimulates glycolytic gene expression in both types of cells, clearly showing for the first time that HIF-1alpha and HIF-2alpha have unique targets.

FIG. 1.

786-O cells express HIF-2α but not HIF-1α. 786-O (PRC-3) is a VHL-defective renal carcinoma cell line, whereas 786-O (WT-8) cells have been stably transfected with functional VHL (38). RCC-4 is another VHL-mutant renal carcinoma cell line, expressing both HIF-1α and HIF-2α. The RCC-4 (T3-14) cells have been stably transfected with functional VHL. Enhanced expression of HIF-1α mRNA in VHL-transfected RCC-4 cells has previously been observed (55). Hypoxia (H) treatment (1.5% O₂ for 16 h) stabilized HIF-2α protein in 786-O (WT-8) cells and stabilized both HIF-1α and HIF-2α in RCC-4 (T3-14) cells as shown by HIF-1α and HIF-2α Western blots. The ARNT Western blot served as a protein loading control.

FIG. 2.

Confirmation of DNA microarray data by Northern blot analyses. (A) Hypoxia (H) (1.5% O₂ for 16 h) stimulates expression of ADRP, NDRG-1, ADM, and VEGF in both 786-O (WT-8) and RCC-4 (T3-14) cells and is likely to be HIF-dependent since these genes are overexpressed in the parental 786-O (PRC-3) and RCC-4 cells independent of O₂. Severe hypoxia (SH) (0.1% O₂ for 16 h) further enhances hypoxic gene expression in comparison to 1.5% O₂. (B) GADD45A and CHOP are hypoxia-responsive genes but are independent of HIF activity. (C) PGK-1, LDHA, PGM-1, and PKM mRNAs, which encode four glycolytic enzymes, are not stimulated by hypoxic treatment (1.5 or 0.1% O₂) of 786-O cells but are induced in RCC-4 cells.

FIG. 3.

HIF-1α is associated with hypoxic induction of glycolytic genes in multiple cell types. (A) Northern blot analysis of HIF-1α and HIF-2α expression in two primary endothelial cell lines and three transformed cell lines. HIF-1α is expressed in all five lines with similar levels, whereas HIF-2α is expressed highly in two endothelial cells. (B) Western blot analysis of HIF-1α and HIF-2α protein in cells treated with hypoxia-mimetic DFX (100 μM) for 6 h. The levels of HIF-1α and HIF-2α protein expression are in general agreement with their mRNA levels, except for the levels of HIF-1α in HMVEC-L cells and HIF-2α in Hep3B cells. (C) Hypoxic treatment (1.5% O₂ for 16 h) increases the expression of the glycolytic genes (PGK and LDHA) and HIF-2α inducible genes (NDRG-1 and VEGF) in all five cell lines.

FIG. 4.

HIF-1α expression in 786-O cells restores the glycolytic gene response to hypoxia treatment. (A) Establishment and characterization of 786-O (WT-8) clones stably transfected with HIF-1α. WT-8L cells express low levels of HIF-1α, whereas WT-8H cells express higher levels of HIF-1α as demonstrated by HIF-1α Northern blot and Western blot assays. The pcDNA3 vector-transfected control WT-8V cells and parental WT-8 cells express no HIF-1α RNA or protein. HEK293 cells, whose endogenous HIF-1α mRNA (asterisk) exhibits a larger size than transfected HIF-1α mRNA (double asterisk), served as positive control for HIF-1α protein. (B) Northern blot analysis of two glycolytic genes in HIF-1α-rescued WT-8 cells, showing that LDHA and PGK-1 are induced in WT-8H cells after treatment with 1.5% O₂ for 16 h. The level of induction in WT-8H is comparable to that in HEK293 cells by endogenous HIF-1α.

FIG. 5.

Mutagenesis of hydroxylated prolines confers HIF-1α and HIF-2α function to normoxic cells. (A) Schematic representation of HIF-1α and HIF-2α single and double proline mutants. (B) Double proline HIF-1α and HIF-2α mutants are stabilized under normal O₂ tension. HEK293-T cells transfected with wild-type single-proline mutant and double-proline mutant HIF-α subunits were split into two dishes, one of which was treated with DFX for 6 h before nuclear extraction preparation. The amount of HIF-α protein was measured by Western blot analysis with specific antibodies and a nonspecific (NS) band is shown for protein loading control. (C) HIF-α double proline mutants are highly active in wild-type (WT)-HRE-Luc reporter transactivation assays in normoxic cells. The mutant (MT)-HRE-Luc reporters are not regulated by HIF-1α or HIF-2α.

FIG. 6.

Establishment of HEK293 TET-on HIF-1αDPA and HIF-2αDPA clones. (A) HIF-α mRNA expression is tightly regulated by doxycycline. Northern blot analysis of HIF-1α and HIF-2α mRNA expression in HEK293 TET-on HIF-1αDPA clone 130 (left) and HEK293 TET-on HIF-2αDPA clone 63 (right) treated with different amounts of doxycycline for 20 h. (B) Doxycycline (1 μg/ml for 20 h) induced HIF-1α mRNA and protein expression in two independent HEK293 TET-on HIF-1αDPA clones (left) and HIF-2α mRNA and protein expression in two independent HEK293 TET-on HIF-2αDPA clones (right). Dox. conc., concentration of doxycycline.

FIG. 7.

HIF-1α and not HIF-2α stimulates genes encoding glycolytic enzymes in HEK293 TET-on cells. (A) Northern blot analysis of two glycolytic genes and two HIF-2α-responsive genes in HEK293 TET-on HIF-1αDPA clones treated with 1 μg of doxycycline/ml for 20 h (+). −, no doxycycline. HIF-1α increases the transcription of all four genes. (B) Northern blot analysis of two glycolytic genes and two HIF-2α-responsive genes in HEK293 TET-on HIF-2αDPA clones treated with 1 μg of doxycycline/ml for 20 h. HIF-2α induced NDRG-1 and VEGF expression but not glycolytic gene expression. (C) Addition of doxycycline (1 μg/ml) increases the glucose usage of a HEK293 TET-on HIF-1DPA clone but not a HEK293 TET-on HIF-2DPA clone or HEK293 TET-on parental cells. The percent change of glucose usage was defined as [glucose (−Dox) − glucose (+DOX)]/glucose (−DOX). (D) Doxycycline treatment (1 μg/ml for 48 h) increases VEGF protein secretion in both HEK293 TET-on HIF-1DPA and HIF-2DPA clones.