The hypoxia-inducible miR-429 regulates hypoxia-inducible factor-1α expression in human endothelial cells through a negative feedback loop

Abstract

Hypoxia-inducible factors (HIFs) 1 and 2 are dimeric α/β transcription factors that regulate cellular responses to low oxygen. HIF-1 is induced first, whereas HIF-2 is associated with chronic hypoxia. To determine how HIF1A mRNA, the inducible subunit of HIF-1, is regulated during hypoxia, we followed HIF1A mRNA levels in primary HUVECs over 24 hours using quantitative PCR. HIF1A and VEGF A (VEGFA) mRNA, a transcriptional target of HIF-1, increased ∼ 2.5- and 8-fold at 2-4 hours, respectively. To determine how the mRNAs were regulated, we identified a microRNA (miRNA), miR-429, that destabilized HIF1A message and decreased VEGFA mRNA by inhibiting HIF1A. Target protector analysis, which interferes with miRNA-mRNA complex formation, confirmed that miR-429 targeted HIF1A message. Desferoxamine treatment, which inhibits the hydroxylases that promote HIF-1α protein degradation, stabilized HIF-1 activity during normoxic conditions and elevated miR-429 levels, demonstrating that HIF-1 promotes miR-429 expression. RNA-sequencing-based transcriptome analysis indicated that inhibition of miRNA-429 in HUVECs up-regulated 209 mRNAs, a number of which regulate angiogenesis. The results demonstrate that HIF-1 is in a negative regulatory loop with miR-429, that miR-429 attenuates HIF-1 activity by decreasing HIF1A message during the early stages of hypoxia before HIF-2 is activated, and this regulatory network helps explain the HIF-1 transition to HIF-2 during chronic hypoxia in endothelial cells.

Keywords: HIF-2; VEGF A; angiogenesis; hypoxamiR; miRNA.

Figure 1.

Hypoxia induces dynamic changes of expression profiles of HIF1A (A), FIH1 (B), PHD2 (C), and VEGFA (D) in HUVECs. The mRNA levels were monitored in quantitative real-time PCR experiments. The results from 3 independent experiments (n = 32) are plotted normalized to 18S rRNA levels and expressed as a fold change over the normoxic control. Error bars represent sd. *P < 0.05.

Figure 2.

Hypoxia induces dynamic changes of protein levels of HIF-1α (A), FIH-1 (B), PHD2 (C), and VEGFA (D) in HUVECs. The bar graphs below show the relative protein amounts at each time point. The protein levels were detected with SDS-PAGE and Western blot and related to β-actin levels. There were 2 individual samples (4 µg total protein per lane) tested for each time point, and the experiments were repeated twice. The protein levels (bar graphs) are normalized to normoxic control. *P < 0.05.

Figure 3.

Hypoxia-induced changes in expression profiles of miR-200b (A), miR-200a (B), and miR-429 (C) in HUVECs. The miRNA levels were monitored in quantitative real-time PCR experiments. The results from 3 independent experiments (n = 32) are plotted normalized to RNU44/RNU48 and expressed as a fold change over the normoxic control. *P < 0.05.

Figure 4.

HIF-1 induces miR-429 expression. A) The localization of putative HRE sequences in miR200b-miR200a-miR429 cluster (numbering is based on human National Center for Biotechnology Information genomic sequence). B) HUVECs were treated with hypoxia mimetics [100 µM DFO (white) for 6 hours, and 200 µM CoCl₂ (gray) for 12 hours], and miRNA levels were monitored in quantitative real-time PCR experiments. ctrl, control. The miRNA level results from 3 independent experiments (n = 10) are plotted normalized to RNU44 levels and expressed as a fold change over the untreated control. *P < 0.05.

Figure 5.

miR-429 does not alter the expression of PHD2 (A) and FIH1 (B). HUVECs were transfected with miR-429 mimic or antagomir, and the mRNA levels were monitored in quantitative real-time PCR experiments in normoxic conditions as well as after 4 and 8 h of exposure to hypoxia. PHD2 mRNA and FIH1 levels from 3 independent experiments (n = 12) are plotted normalized to 18S rRNA levels and expressed as a fold change over the transfection control. The predicted target sites of miR-429 in PHD2 and FIH1 3′ UTRs are shown on the right. miR-429 levels for mimic or antagomir transfection were monitored for each experiment, as shown in (C). *P < 0.05.

Figure 6.

A) miR-429 specifically modulates expression of HIF1A mRNA under normoxia and hypoxia. HUVECs were transfected with miR-429 mimic or antagomir, and the mRNA levels were monitored in quantitative real-time PCR experiments in normoxic conditions and after 4 and 8 hours of hypoxia. HIF1A mRNA levels from 3 independent experiments (n = 12) are plotted normalized to 18S rRNA levels and expressed as a fold change over the transfection control. *P < 0.05. B) The corresponding changes of HIF-1α protein levels were detected with SDS-PAGE and Western blot and related to total protein levels. There were 2 individual samples (3 µg total protein per lane) tested for each treatment, and the experiments were repeated twice. Mimic 429 (m429) and antagomir 429 (ant429) for each condition are shown. miR-429 levels were monitored for each experiment. C) The predicted target site of miR-429 in HIF1A 3′ UTRs is shown, and the minimum free energy (mfe) for miR-429-HIF1A mRNA was calculated with RNAhybrid (33) (miR-429 is shown in red; HIF1A target mRNA sequence is in green).

Figure 7.

A) miR-429 binds to predicted target sequence at HIF1A 3′ UTR. HUVECs were transfected with HIF1A target sequence-specific TP and/or miR-429 analog (Mimic 429). The HIF1A mRNA levels were monitored in quantitative real-time PCR experiments. The HIF1A mRNA level results from 3 independent experiments (n = 12) are plotted normalized to 18S rRNA levels and expressed as a fold change over the transfection control. *P < 0.05. B) The corresponding changes of HIF-1α protein levels were detected with SDS-PAGE and Western blot and related to β-actin levels. There were 2 individual samples (3 µg total protein per lane) tested for each treatment, and the experiments were repeated twice.

Figure 8.

A) miR-429 modulates VEGFA expression indirectly through changing HIF1A levels. HUVECs were transfected with miR-429 analog or inhibitor, and the mRNA levels were monitored in quantitative real-time PCR experiments in normoxic conditions and after 4 and 8 hours of hypoxia. The mRNA level results from 3 independent experiments (n = 12) are normalized to 18S rRNA levels and expressed as a fold change over the transfection control. The predicted target site of miR-429 in VEGFA 3′ UTRs is shown on the right. The VEGFA (B) and HIF1A (C) mRNA levels were monitored in quantitative real-time PCR experiments following transfection of HUVECs with VEGFA TP or HIF1A TP and/or miR-429 analog (Mimic 429). The mRNA level results from 3 independent experiments (n = 12) are normalized to 18S rRNA levels and expressed as a fold change over the transfection control. *P < 0.05.

Figure 9.

Model of negative feedback loop between HIF1A and miR-429. During hypoxia, HIF1A mRNA is induced, whereas HIF-1α protein is stabilized, allowing for the for-mation and accumulation of active HIF-1 complexes. The HIF-1 is translocated to the nucleus in order to regulate expression of its target genes. HIF-1 binds to the HRE sequence located in the miR200b-miR200a-miR429 cluster and induces miR-429 expression and, thus, increases the cytosolic levels of this miRNA. miR-429 binds in cytosol to the target sequence located at the 3′ UTR of HIF1A mRNA, which leads to decreased HIF-1 activity.