MicroRNA-155 promotes resolution of hypoxia-inducible factor 1alpha activity during prolonged hypoxia

Abstract

The hypoxia-inducible factor (HIF) is a key regulator of the transcriptional response to hypoxia. While the mechanism underpinning HIF activation is well understood, little is known about its resolution. Both the protein and the mRNA levels of HIF-1α (but not HIF-2α) were decreased in intestinal epithelial cells exposed to prolonged hypoxia. Coincident with this, microRNA (miRNA) array analysis revealed multiple hypoxia-inducible miRNAs. Among these was miRNA-155 (miR-155), which is predicted to target HIF-1α mRNA. We confirmed the hypoxic upregulation of miR-155 in cultured cells and intestinal tissue from mice exposed to hypoxia. Furthermore, a role for HIF-1α in the induction of miR-155 in hypoxia was suggested by the identification of hypoxia response elements in the miR-155 promoter and confirmed experimentally. Application of miR-155 decreased the HIF-1α mRNA, protein, and transcriptional activity in hypoxia, and neutralization of endogenous miR-155 reversed the resolution of HIF-1α stabilization and activity. Based on these data and a mathematical model of HIF-1α suppression by miR-155, we propose that miR-155 induction contributes to an isoform-specific negative-feedback loop for the resolution of HIF-1α activity in cells exposed to prolonged hypoxia, leading to oscillatory behavior of HIF-1α-dependent transcription.

Fig. 1.

Differential expression of HIFα subunits in hypoxia and potential role for hypoxia-induced miRNAs. (A) Caco-2 colonic epithelial cells exposed to 0 to 48 h of hypoxia (1% O₂) demonstrate a transient increase in cytosolic and nuclear HIF-1α and a sustained increase in HIF-2α protein levels. α-Tubulin and TATA-box binding protein (TBP) were used as loading controls. (B) Real-time PCR analysis normalized to the corresponding normoxic control was used to determine HIF-1α and HIF-2 α mRNA levels in Caco-2 cells exposed to hypoxia. (C) miRNA array analysis identified a cohort of hypoxia-induced miRNAs in Caco-2 cells. Those miRNAs previously identified as hypoxia-inducible are outlined in bold. The last three miRNAs are included as sample controls that were not induced in hypoxia. (D) HIF-1α is a predicted target of miR-155. Watson-Crick base pairing is indicated by solid lines, while non-Watson-Crick base pairing is indicated by dashed lines. In all cases, n = 3 to 4 independent experiments. The data are shown as representative blots or means ± the standard errors of the mean (SEM) (*, P < 0.05; nsd, no significant difference).

Fig. 2.

Transcriptional upregulation of miR-155 in response to hypoxia. (A) Real-time PCR analysis demonstrates mature miR-155 expression in Caco-2 cells exposed to hypoxia. (B and C) Mir-155 expression levels from colon and liver tissue, respectively, harvested from mice exposed to 10% atmospheric O₂ for the indicated periods. (D and E) Caco-2 and HeLa cells were treated with (or without) actinomycin D and exposed to 48 h of hypoxia. Real-time PCR analysis was used to measure primary transcript pri-miR-155 expression in response to hypoxia. All data are shown as the fold over basal miR-155 or pri-miR-155 expression, with 18S rRNA being used as an endogenous control. n = 3 independent experiments for in vitro experiments. n = 6 independent mice were used for in vivo measurements. Values are presented as means ± the SEM (*, P < 0.05).

Fig. 3.

HIF-1-dependent miR-155 expression in hypoxia. (A) Bioinformatic analysis using the Genomatix software tool MatInspector reveals the presence of three HRE consensus motifs and two NRE consensus motifs in the promoter of the miR-155 gene. (B) Effective siRNA-dependent knockdown of HIF-1α, HIF-2α, and p65 was confirmed by immunoblot analysis. (C) Caco-2 cells were transfected with HIF-1α, HIF-2α, or p65-siRNA prior to hypoxic exposure (48 h), and mature miR-155 expression was analyzed by real-time PCR. (D) Protein levels of HIFα subunits in HIF-1α or HIF-2α knockdown HepG2 cells were analyzed by immunoblotting. (E) HepG2 cells stably expressing shRNA directed against HIF-1α or HIF-2α were exposed to hypoxia, and pri-miR-155 expression was measured by using real-time PCR. n = four independent experiments throughout. Values are presented as means ± the SEM (*, P < 0.05; **, P < 0.01).

Fig. 4.

miR-155 reduces HIF signaling. (A) Caco-2 cells cotransfected with HRE-luciferase and either control siRNA (si-NT), HIF-1α siRNA (si-HIF-1α), or miR-155 with or without anti-miR-155 and exposed to 6 h of 1% O₂. A luciferase assay was carried out, and the values were normalized to cotransfected β-Gal. (B) Real-time PCR analysis of HIF-1α mRNA levels in Caco-2 cells in normoxia following transfection with miR-155. HIF-1α siRNA was used as positive control. (C) Immunoblot analysis reveals HIF-1α and HIF-2α protein levels after transfection with miR-155 for 48 h prior to hypoxic exposure for 8 h. (D) Caco-2 cells were transfected with miR-155, and hypoxia-induced PHD2 mRNA expression was determined by real-time PCR analysis. In all cases n = 3 to 4 independent experiments. The data are shown as representative blots or as means ± the SEM (*, P < 0.05; **, P < 0.01).

Fig. 5.

HIF-1α is a direct target of miR-155. (A) Wild-type (WT) and mutant (MT) versions of the predicted miR-155 binding site in the 3′UTR of HIF-1α mRNA were cloned into a luciferase vector. (B) Caco-2 and HeLa cells were transfected with miR-155 (155) or miR-control (ctl) 24 h prior to transfection with the reporter luciferase construct. After 24 h, a luciferase assay was carried out and normalized to the cotransfected β-Gal. (C and D) HeLa and Caco-2 cells were transfected with increasing doses of either miR-155 or miR-control (miR-ctl), and the HIFα protein or mRNA levels were measured. For the HIF-1α mRNA statistical data, comparisons were made to the corresponding concentration of the control RNA. n = 3 to 4 independent experiments throughout. The data shown are representative blots or means ± the SEM (*, P < 0.05; **, P < 0.01).

Fig. 6.

Inhibition of miR-155 sustains HIF-1α activation in hypoxia. (A) Caco-2 and HeLa cells were transfected with nonspecific anti-miR control or anti-miR-155 oligonucleotides 24 h prior to hypoxic exposure over time (0 to 48 h). The HIF-1α and HIF-2α levels were analyzed in whole-cell protein extracts by immunoblotting. (B) HIFα/β-actin ratios of densitometric analysis are shown from n = 3 independent experiments in each case.

Fig. 7.

PHD2/3 and miR-155 negative-feedback loops are part of the regulatory network of HIF-1α in prolonged hypoxia. A computational model based on experimental data was generated to estimate the relative contributions of miR-155 and PHD2/3 in determining temporal HIF-1α activity during prolonged periods of hypoxia. (A) Temporal deviation from normal (transient) HIF-1α activity predicted by removal of the PHD2/3 feedback loop, the miR-155 feedback loop, or both feedback loops from the model are shown in red, green, and blue, respectively. (B) Schematic representation of the proposed network of known and predicted negative-feedback loops in the HIF-1a pathway. The “?” indicates currently unidentified direct or indirect negative-feedback mechanism(s). (C) Caco-2 cells were transfected with a vector containing an HRE promoter controlling expression of a secreted form of luciferase derived from Gaussia princeps and cultured under normoxia (21% O₂; red) or hypoxia (1% O₂; black). The medium was sampled every 3 h, and the HRE-luciferase activity assessed for each time period and normalized to the cell number. Representative traces are shown for HIF activity in arbitrary units (AU).