Biphasic regulation of miR-17∼92 transcription during hypoxia: roles of HIF1 and p53 hyperphosphorylation at ser15

Abstract

We have reported previously that during hypoxia exposure, the expression of mature miR-17∼92 was first upregulated and then downregulated in pulmonary artery smooth muscle cells (PASMC) and in mouse lungs in vitro and in vivo. Here, we investigated the mechanisms regulating this biphasic expression of miR-17∼92 in PASMC in hypoxia. We measured the level of primary miR-17∼92 in PASMC during hypoxia exposure and found that short-term hypoxia exposure (3% O₂, 6 h) induced the level of primary miR-17∼92, whereas long-term hypoxia exposure (3% O₂, 24 h) decreased its level, suggesting a biphasic regulation of miR-17∼92 expression at the transcriptional level. We found that short-term hypoxia-induced upregulation of miR-17∼92 was hypoxia-inducible factor 1α (HIF1α) and E2F1 dependent. Two HIF1α binding sites on miR-17∼92 promoter were identified. We also found that long-term hypoxia-induced suppression of miR-17∼92 expression could be restored by silencing of p53. Mutation of the p53-binding sites in the miR-17∼92 promoter increased miR-17∼92 promoter activity in both normoxia and hypoxia. Our findings suggest that the biphasic transcriptional regulation of miR-17∼92 during hypoxia is controlled by HIF1/E2F1 and p53 in PASMC: during short-term hypoxia exposure, stabilization of HIF1 and induction of E2F1 induce the transcription of miR-17∼92, whereas during long-term hypoxia exposure, hyperphosphorylation of p53 suppresses the expression of miR-17∼92.

NEW & NOTEWORTHY We showed that the biphasic transcriptional regulation of miR-17∼92 during hypoxia is controlled by two distinct mechanisms: during short-term hypoxia exposure, induction of HIF1 and E2F1 upregulates miR-17∼92. Longer hypoxia exposure induces hyperphosphorylation of p53 at ser15, which leads to its binding to miR-17∼92 promoter and inhibition of its expression. Our findings provide novel insights into the spatiotemporal regulation of miR-17∼92 that may play a role in the development of human lung diseases including pulmonary hypertension (PH).

INTRODUCTION

There is increasing evidence that microRNAs (miRNAs) are important players in the pathogenesis of disease development. A majority of the research efforts have been dedicated to the identification of the downstream targets of miRNAs and the role of identified miRNA and its target(s) in the given biological systems (1). Bioinformatics tools and databases as well as global proteomics offer powerful tools for target prediction and confirmation (2). However, only limited studies attempt to investigate the molecular mechanisms underlying the regulation of miRNA genes, either by transcriptional activators or repressors for miRNA genes, due to the lack of strategies to define regulatory elements other than traditional promoter analysis (3).

The miR-17∼92 cluster includes six mature miRNAs (miR-17, miR-18a, miR-19a, miR-19b, miR-20a, and miR-92a) that are organized in a polycistronic cluster and transcribed by a common promoter region as one common primary miRNA until it is processed into mature miRNAs (4, 5). Previously, we found that miR-17∼92 induces PASMC proliferation in vitro, and that smooth muscle cell (SMC)-specific knockout of miR-17∼92 significantly attenuates hypoxia-induced pulmonary vascular remodeling and experimental pulmonary hypertension (PH) in mice, suggesting a critical role of miR-17∼92 in PH via its regulation of PASMC proliferation (2). Other studies have identified the role of miR-17∼92 in tumorigenesis, lung development, and angiogenesis (3, 4, 6, 7). However, less is known about how miR-17∼92 expression is regulated. We previously reported that the expression pattern of miR-17∼92 in PASMC and mouse lungs exposed to hypoxia is complex: an early transient increase followed by a decrease in the later stages of exposure to hypoxia in both PASMC in vitro and mouse lungs in vivo (2). However, the molecular mechanisms underlying this biphasic expression of miR-17∼92 during hypoxia exposure are not known.

In this study, we showed that the biphasic transcriptional regulation of miR-17∼92 during hypoxia is controlled by two distinct mechanisms: during short-term hypoxia exposure, induction of HIF1α and E2F1 upregulates miR-17∼92; and after longer hypoxia exposure, hyperphosphorylation of p53 at ser15 leads to the binding of p53 to the miR-17∼92 promoter and inhibition of its expression. We have also identified two functional HIF1α binding sites in the proximal promoter of the miR-17∼92 gene. This biphasic regulation of miR-17∼92 provides novel insights into the spatiotemporal regulation of miR-17∼92 in PASMC during hypoxia exposure.

MATERIALS AND METHODS

Cell Culture

Normal human PASMC were purchased from Lonza (Walkersville, MD; Cat. No. CC-2581. Lot No. 0000419239: 35-yr-old Hispanic male donor; Lot No. 0000669096: 51-yr-old White male donor; Lot No. 22TL221791: 23-yr-old White female donor) and were maintained in SmGM-2 medium (Lonza) containing 5% fetal bovine serum (FBS), growth factors, and 1% penicillin-streptomycin. These primary human PASMC were tested negative for human immunodeficiency virus, hepatitis B, hepatitis C, mycoplasma, bacteria, yeast, or fungi and were authenticated by α-smooth muscle actin (α-SMA) staining by the provider. Purchased human PASMC were further stained in our laboratory with α-SMA and smooth muscle myosin heavy chain (SMMHC) antibodies (Supplemental Fig. S1A), and smooth muscle protein 22-α (SM22α) antibody (Supplemental Fig. S1B), to confirm their smooth muscle cell characteristics. Passages 5–7 of PASMC were used for experiments to minimize the potential alteration of cell behavior during passaging.

Primary mouse PASMC were isolated from lungs of 6- to 8-wk old male C57BL/6 mice (at least three donor mice were included) as we described before (2, 8) and were also maintained in SmGM-2 medium (Lonza) containing 5% fetal bovine serum (FBS), growth factors, and 1% penicillin-streptomycin. Briefly, after the trachea was exposed and the lung was perfused with PBS, Solution 1 [5 mL M199 medium (Sigma) + 25 mg agarose + 25 mg iron filings] was injected into blood circulation and Solution 2 [5 mL M199 medium (Sigma) + 25 mg agarose] into the trachea. Then the whole lung was collected, cut into small pieces, and then digested with collagenase A (∼1 h at 37°C). Digested lung tissue was dispensed, and lung vessels (containing iron fillings) were precipitated with a magnet column (Invitrogen), followed by culturing and maintaining in SmGM-2 medium. Typically, mouse PASMC were ready for experiments in about 2 wk of culture. The purity of isolated mouse PASMC was confirmed by immunostaining of α-SMA and SMMHC (Supplemental Fig. S2A), and SM22α (Supplemental Fig. S2B), with respective antibodies following our previously published procedure (8). Passage 1∼3 of the cells were used for the studies. All cells were maintained in a humidified incubator with a constant supply of 5% CO₂ at 37°C. Hypoxia (3% O₂) was achieved in INVIVO2 300 (Ruskinn Technology Limited, UK) with a continuous O₂ sensor and a constant 5% CO₂ was supplied.

Immunostaining of α-SMA and SMMHC, or SM22α

Briefly, PASMC were fixed with 4% paraformaldehyde at room temperature for 10 min and blocked with PBS that contains 3% bovine serum albumin (BSA), 3% goat serum, and 0.1% Triton X-100, at room temperature for 30 min. Cells were then incubated with α-SMA (Sigma-Aldrich) and SMMHC (Proteintech Group, Rosemont, IL), or SM22α (Proteintech Group) primary antibody, respectively, at 4°C overnight. After washing with PBS that contains 0.1% Tween-20, cells were incubated with anti-mouse and anti-rabbit antibodies (Invitrogen) at room temperature for 30 min. ProLong Gold Antifade Reagent with DAPI (Invitrogen) was applied after washing and before cells were imaged on a Zeiss LSM 710 Confocal Microscope.

Western Blot Analysis

After a specific treatment or exposure to hypoxia, we washed cells twice with cold phosphate-buffered saline (PBS), and lysed cells with modified RIPA (mRIPA) buffer (50 mM Tris, pH 7.4, 1% NP-40, 0.25% deoxycholate, 150 mM NaCl, and protease inhibitors) on ice for 30 min. The cell lysates were centrifuged at 13,000 g for 10 min, and the supernatants were collected to determine the protein concentrations with Bio-Rad protein assay solution (Bio-Rad, Hercules, CA), followed by separation by SDS-polyacrylamide gel electrophoresis. After transferring to BA85 nitrocellulose membranes (PROTRAN; Whatman, Dassel, Germany), the membrane was probed with primary and secondary antibodies, and the abundance of proteins was detected with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Waltham, MA). The following primary antibodies were used in this study: α-tubulin and β-actin (Sigma-Aldrich and Proteintech Group, Inc.), HIF1α (BD; Franklin Lakes, NJ), HIF2α (Novus Biologicals, Centennial, CO), E2F1 (Sigma, St. Louis, MO), c-Myc (Sigma), and phosphor-p53 antibody sampler kit (including antibodies to total p53 and phspho-p53 at ser6, ser9, ser15, ser20, ser37, ser46, thr81, and ser392; Cell Signaling Technology, Danvers, MA). Anti-mouse and anti-rabbit IgG-HRP conjugates were purchased from Bio-Rad. Primary antibodies used in this study have been validated either in this study or as previously demonstrated by providers or others. Representative full-length blots and details of antibody validation are presented in Supplemental Material.

Chromatin Immunoprecipitation

Chromatin immunoprecipitation (ChIP) assay was carried out with the Magna ChIP A/G kit (EMD-Millipore, Kankakee, IL) following the manufacturer’s protocol. Briefly, after exposure to normoxia or hypoxia (3% O₂, 6 or 24 h), PASMC were fixed in 1% formaldehyde in the culture media for 10 min at room temperature, washed with cold PBS three times, scraped with cold PBS supplemented with protease inhibitor cocktail II, and collected into microfuge tubes. Cells were pelleted by centrifugation at 800 g at 4°C for 5 min and lysed in cell lysis buffer with protease inhibitor cocktail II. Cell lysates were centrifuged at 800 g at 4°C for 5 min. Pellets were resuspended in nuclear lysis buffer with protease inhibitor cocktail II, sonicated with MISONIX Ultrasonic liquid processors (Farmingdale, NY), and centrifuged at 12,000 g at 4°C for 10 min to remove insoluble material. We removed 5 µL of the supernatant as “Input.” The remaining supernatant was mixed with dilution buffer containing protease inhibitor cocktail II and incubated with immunoprecipitating antibody and fully resuspended protein A/G magnetic beads for 1 h at 4°C. Preimmune IgG was used as a negative control. Protein A/G magnetic beads were pelleted and washed with low salt wash buffer, high salt wash buffer, LiCl wash buffer, and TE buffer sequentially. Beads were mixed with the ChIP elution buffer with proteinase K and incubated at 62°C for 2 h with shaking, and at 95°C for 10 min, before they were cooled down to room temperature. Beads were removed, the supernatants were mixed with binding buffer, and the mixtures were applied to a spin filter, followed by centrifugation for 30 s at 12,000 g. The spin filters were washed with the wash buffer and centrifuged for 30 s at 12,000 g. DNA was eluted with elution buffer from the spin filter membranes and collected by centrifugation for 30 s at 12,000 g. Eluted DNA samples were subjected to PCR with site-specific primer sets [DNA 2.0 µL, H₂O 12.6 µL, 10× PCR buffer (w/o MgCl₂) 2.0 µL, MgCl₂ (50 mM) 0.6 µL, 2.5 mM dNTP 1.6 µL, primers 0.8 µL, Taq (5 U/mL) 0.4 µL] with the following program: 94°C 3 min, 32 times (94°C 20 s, 59°C 30 s), 72°C 30 s, 72°C 2 min. The PCR products were analyzed by 2% agarose gel electrophoresis. To confirm the binding of phosphor-P53 (ser15) to miR-17∼92 promoter, quantitative real-time reverse transcription PCR (qRT-PCR) were performed using SYBR Green PCR Master Mix (Applied Biosystems) and site-specific primers.

Quantitative Real-Time Reverse Transcription PCR

Total RNA was isolated from cultured cells using a miRNeasy Mini Kit (Qiagen, Valencia, CA), treated with an RNase-Free DNase Set (Qiagen), and quantified with Nanodrop 2000 spectrophotometer (Thermo Scientific, Rockford, IL). miRNA expression was determined by quantitative real-time reverse transcription PCR (qRT-PCR) analysis as we described before (2). Briefly, a poly(A) tail was added to the 3′-end of miRNAs with a Poly(A) Polymerase Tailing Kit (Epicentre Biotechnologies, Madison, WI). Poly(A) tailed-miRNAs were then reverse transcribed using M-MLV reverse transcriptase (Invitrogen, Grand Island, NY) with a poly(T) adaptor, which is composed of a poly(T) sequence and a sequence complementary to the universal primer used in the following qRT-PCR analysis. SNORD44, SNORD47, and SNORD48 were used as internal controls. For qRT-PCR analysis of mRNA expression, total RNA was reversely transcribed using M-MLV reverse transcriptase with oligo (dT)12–18 primer (Invitrogen). Ribosomal protein L19 (RPL19) was used as an internal control. qRT-PCR was performed using SYBR Green PCR master mix (Applied Biosystems, Foster City, CA) on StepOnePlus or ViiA 7 Real-Time PCR system (Applied Biosystems). Primer sequences are provided in Supplemental Table S1.

miR-17∼92 Promoter Luciferase Reporter Assay

A full-length miR-17∼92 promoter was constructed to the pGL3-Promoter vector (Promega) as described before (9). A construct with mutated p53 binding sites was also constructed as described (9). All constructs were sequenced to confirm the DNA sequences. PASMC were transfected with these reporter plasmids together with Renilla luciferase control plasmid and were then exposed to normoxia or hypoxia (3% O₂) for 24 h before the measurement of luciferase activity. Lipofectamine 2000 reagent (Invitrogen) was used for transfection. The luciferase activity was measured using Dual Luciferase Assay System (Promega) on GloMax-96 Microplate luminometer (Promega). Relative promoter activity was calculated by normalizing the firefly luciferase activity to that of Renilla luciferase.

Small Interfering RNA Suppression of HIF1α, E2F1, and p53

Normal human PASMC were plated in 60 mm dishes at 60%–70% confluence and transfected with gene-specific small interfering RNA (siRNAs) or negative control siRNA (siNeg, used as control) (Ambion) using Lipofectamine 2000 in Opti-MEM (Invitrogen), as recommended by the manufacturer’s protocol. Six hours after transfection, cells were changed to complete medium and incubated overnight, followed by the exposure to hypoxia (3% O₂) or normoxia as specified in individual experiments.

Statistical Analysis

In experiments that were repeated at least three times independently, we performed a nonparametric Mann–Whitney test, one-way ANOVA test (when more than 3 groups), or two-way ANOVA test (when variable = 2) using GraphPad Prism 9 (GraphPad, San Diego, CA) when applicable. Bonferroni posttests were carried out after one-way ANOVA. The significant difference values were set at 0.05, 0.01, and 0.001.

RESULTS

Biphasic Transcriptional Regulation of miR-17∼92 in Mouse and Human PASMC during Hypoxia: Early Transient Induction and Late Suppression

Previously, we have shown that in both human and mouse PASMC, hypoxia induces expression of the mature miRNAs in the miR-17∼92 cluster in the early phase and inhibits their expression in the late phase of hypoxia exposure (2), suggesting a biphasic regulation of miR-17∼92 by hypoxia. Since these miRNAs are transcribed together as a cluster, we examined whether this biphasic expression was regulated at the transcriptional level. We measured the levels of primary miR-17∼92 (Pri-miR-17∼92) in mouse and human PASMC exposed to hypoxia (3% O₂) for various time periods (3, 6, 12, 24, and 48 h) and found that, in both mouse and human PASMC, a short-term hypoxic exposure (3% O₂, 3 and 6 h) induced the expression of Pri-miR-17∼92, whereas long-term hypoxia exposure (3% O₂, 24–48 h) decreased the Pri-miR-17∼92 levels (Fig. 1). These results suggest that hypoxia regulates the biphasic expression of miR-17∼92 at the transcriptional level.

Figure 1.

Biphasic regulation of miR-17∼92 transcription in PASMC by hypoxia. Isolated mouse PASMC (mPASMC) (A) and human PASMC (B) were exposed to normoxia (21% O₂, N) or hypoxia (3% O₂, H) for 0, 3, 6, 12, 24, and 48 h and then the expression levels of pri-miR-17∼92 were determined by qPCR analysis. The ratio of pri-miR-17∼92 levels under hypoxic and normoxic conditions at each time point (H/N expression ratio) was calculated and then compared with that at time 0 of exposure (n = 3–5). Nonparametric Mann–Whitney test was performed. Data are presented as means ± SE. *P < 0.05; **P < 0.01. qPCR, quantitative PCR.

Short-Term Exposure to Hypoxia Induces HIF1α and E2F1 but Not HIF2α or c-Myc in Human PASMC

A handful of transcription factors are known to upregulate the miR-17∼92 cluster, including c-Myc and E2F, which are induced by HIF (10–12). Among the nine E2F transcription factors, E2F1, E2F2, and E2F3a are transcriptional activators (13). We found that short-term hypoxia exposure induced mRNA levels of HIF1α (starting 1 h) and E2F1 (at 3 and 6 h), but not HIF2α or c-Myc, in human PASMC (Fig. 2A). Similarly, short-term hypoxia exposure also induced protein levels of HIF1α (Fig. 2B) and E2F1 (Fig. 2C), whereas protein levels of HIF2α and c-Myc remained unchanged (Fig. 2, D and E). These data suggest that the expression level of Pri-miR-17∼92 may correlate with that of HIF1α and E2F1 in short-term hypoxia exposure.

Figure 2.

Short-term exposure to hypoxia induces HIF1α and E2F1 but not HIF2α or c-Myc in human PASMC. A: human PASMC were exposed to normoxia (21% O₂, N) or hypoxia (3% O₂, H) for 0, 1, 3, and 6 h and then the mRNA levels of HIF1α, HIF2α, E2F1, and c-Myc were determined by qPCR analysis. The ratio of these mRNA levels under hypoxic and normoxic conditions at each time point was calculated and compared with that at time 0 of exposure (n = 3–5). Data are presented as means ± SE. *P < 0.05. B–E: PASMC lysates were also collected after the cells were exposed to normoxia and hypoxia for 3 or 6 h. The protein levels of HIF1α (B), E2F1 (C), HIF2α (D), and c-Myc (E) were determined by Western blot analysis. Representative blots and respective quantification (means ± SD) are presented. n = 3–6 as indicated in each scatter plot. Nonparametric Mann–Whitney test was used for statistical analysis. *P < 0.05; **P < 0.01. HIF1α, hypoxia-inducible factor 1α; ns, no significance; PASMC, pulmonary artery smooth muscle cell; qPCR, quantitative PCR.

Induction of miR-17∼92 in Short-Term Hypoxia Exposure Is HIF1/E2F1 Dependent

To address whether induction of miR-17∼92 is dependent on HIF1a and/or E2F1, we silenced HIF1α and E2F1 in human PASMC using the respective siRNA and examined the expression of the miR-17∼92 cluster in response to hypoxia exposure (3% O₂, 6 h). We confirmed the silencing of HIF1α or E2F1 in PASMC by Western blot analysis (Fig. 3, A and B) and qRT-PCR analysis (Fig. 3, C and D) and also showed that the HIF2α level was unchanged (Fig. 3E). We found that silencing of HIF1α and E2F1 suppressed hypoxia-induced expression of mature miR-17∼92 and Pri-miR-17∼92 (Fig. 3, F–J), suggesting that hypoxia upregulates miR-17∼92 via induction of HIF1α and E2F1.

Figure 3.

Short-term hypoxic induction of miR-17∼92 is HIF1α/E2F1 dependent. Human PASMC were transfected with siRNA against HIF1α or E2F1 and then exposed to normoxia (21% O₂, N) or hypoxia (3% O_2, H) for 6 h. A and B: the silencing of HIF1 (A) and E2F1 (B) was confirmed by Western blot analysis (n = 3). Representative blots and respective quantification (means ± SD) are presented. Two-way ANOVA analysis was performed. C–E: the mRNA levels of HIF1α, E2F1, and HIF2α were measured by qPCR analysis (n = 3). One-way ANOVA analysis was performed. F–J: the ratio of the expression levels of pri-miR-17∼92, as well as each mature miR-17∼92, at hypoxia over normoxia was calculated. n = 4–5 as indicated in each scatter plot. One-way ANOVA analysis was performed. *P < 0.05; **P < 0.01; ***P < 0.001. HIF1α, hypoxia-inducible factor 1α; ns, no significance; PASMC, pulmonary artery smooth muscle cell; qPCR, quantitative PCR.

HIF1α Is Associated with the miR-17∼92 Promoter

E2F1 has been reported to directly bind to the promoter of miR-17∼92. The miR-17∼92 promoter contains two functional E2F binding sites (TTTSSCGC sequence, where S = C or G) 3 kb upstream of the transcription start site (14). However, it is unclear whether HIF1α binds to the miR-17∼92 promoter. By in silico analysis of miR-17∼92 promoter, we identified six putative HIF binding sites (RCGTG) (Fig. 4A). Since both HIF1α and HIF2α can bind to HIF response elements (HRE), we exposed PASMC to normoxia or hypoxia for 6 h and performed chromatin immunoprecipitation (ChIP) assays using HIF1α and HIF2α antibodies. Immunoprecipitated DNA fragments were purified and PCR with human miR-17∼92 promoter-specific primers covering six HRE was performed. We confirmed sites 3 and 4 to be HREs on the miR-17∼92 promoter (Fig. 4B). We found that site 3 was responsible for the binding of HIF1α at basal levels, whereas site 4 was specific for the hypoxia-induced HIF1α binding (Fig. 4B). Interestingly, the binding of HIF2α to sites 3 and 4 remained unchanged in normoxia or hypoxia conditions (Fig. 4A), confirming the HIF1α-specific regulation of miR-17∼92. These data suggest a differential HIF1α binding in miR-17∼92 between normoxia and hypoxia.

Figure 4.

Hypoxia promotes HIF1α binding to the HRE site (4) on the miR-17∼92 promoter. A: a diagram of the miR-17∼92 proximal promoter showing putative HIF response elements (HRE). B: human PASMC were exposed to normoxia or hypoxia (3% O₂, H) for 6 h and their nuclear extract was collected. DNA fragments were precipitated with HIF1α or HIF2α antibodies. After purification, the DNA fragments were subject to PCR with human miR-17∼92 promoter-specific primers. The DNA fragments from IgG treated (IgG) or untreated (Input) nuclear extracts were used as negative control and positive control, respectively. Representative blots were presented (n = 3). Respective quantification (means ± SD) for each putative site was calculated as follows: in each independent experiment, the signal (quantification of PCR product band intensity by ImageJ) in IgG control group in normoxia was set as 1, and other groups were compared with the respective control by calculating the band intensity ratio (vs. IgG). Then if there was a HIF antibody pull-down and if the pull-down efficiency is different in normoxia and hypoxia were analyzed using two-way ANOVA analysis. *HIF antibody pulled-down signal vs. IgG pull-down signal; #H vs. N signal. *P < 0.05; ** or ##, P < 0.01. HIF1α, hypoxia-inducible factor 1α; PASMC, pulmonary artery smooth muscle cell.

p53 Is Responsible for the Suppression of miR-17∼92 Expression in Hypoxia in PASMC

We found that the Pri-miR-17∼92 level was decreased in PASMC exposed to prolonged hypoxia (3% O₂, 24 h; Fig. 1), which is consistent with our previous report that prolonged hypoxia decreases mature miR-17∼92 expression in vivo (2). These observations suggest a mechanism of repression of miR-17∼92 in prolonged hypoxia. A previous report identified two putative p53 binding sites in the proximal region of the miR-17∼92 promoter (BS1: –691 to –716 and BS2: –20 to –44) (Fig. 5A) and showed that p53 directly binds to the BS2 to suppress miR-17∼92 expression in colon cancer cells in hypoxia (9). To address the role of p53 in the hypoxia-mediated expression of miR-17∼92 in PASMC, we first transfected PASMC with a luciferase reporter construct containing a full-length miR-17∼92 promoter region (Full) or a full-length construct with mutated p53 binding site (MBS1 or MBS2), followed by the exposure to hypoxia (3% O₂) for 24 h. We found that hypoxia suppressed the promoter activity of the full construct, whereas mutation of both p53 binding sites abolished the hypoxia-mediated suppression of luciferase activity (Fig. 5B). These results suggest that p53 is responsible for the suppressed expression of miR-17∼92 in prolonged hypoxia.

Figure 5.

Chronic hypoxia suppresses miR-17∼92 via p53 in PASMC. A: a diagram of the miR-17∼92 proximal promoter luciferase reporter constructs containing putative p53 binding sites (Full) or the mutated p53 binding sites (MBS1 and MBS2). B: human PASMC were transfected with the wildtype (Full) or mutated (MBS1 or MBS2) miR-17∼92 promoter constructs, together with the Renilla luciferase control plasmid, and then exposed to normoxia (N) or hypoxia (H, 3% O₂) for 24 h. Luciferase activity was measured by dual-luciferase reporter assay and compared with that of Full construct in normoxia (n = 3–4). Two-way ANOVA analysis was performed. Data are presented as means ± SD. *P < 0.05; ***P < 0.001. HIF1α, hypoxia-inducible factor 1α; PASMC, pulmonary artery smooth muscle cell.

To further confirm these findings, we silenced p53, as well as HIF1α, and/or HIF2α, or E2F1, in PASMC and exposed them to normoxia or hypoxia for 24 h and then measured the expression of miR-17∼92 (hypoxia over normoxia ratio) in these cells. We confirmed the knockdown efficiency of p53 (Fig. 6A and Supplemental Fig. S3) and found that silencing of p53 increased Pri-miR-17∼92 and mature miR17∼92 levels in hypoxia exposure, whereas silencing of HIF1α, and/or HIF2α, or E2F1 (Fig. 6, B–D) did not significantly change the suppressed expression of Pri-miR-17∼92 and mature miR-17∼92 in long-term hypoxia exposure (Fig. 6E).

Figure 6.

Silencing of p53 restores hypoxia-repressed expression of miR-17∼92 at 24-h exposure. Human PASMC were transfected with siRNA against p53, HIF1α, HIF2α, or E2F1, and then exposed to normoxia or hypoxia (3% O₂) for 24 h. A–D: knockdown efficiency of each siRNA was confirmed by Western blot analysis (n = 3). Representative blots and respective quantification (means ± SD) are presented. Two-way ANOVA analysis was performed. E: expression of Pri-miR-17∼92 and each mature miRNA were measured by qPCR analysis (n = 3–6 as indicated in each scatter plot). The ratio of their expression levels under hypoxia and normoxia was calculated (H/N ratio) and compared with that in control cells [PASMC transfected with negative siRNA (siNeg)]. One-way ANOVA analysis was performed. Data are presented as means ± SD. *P < 0.05; **P < 0.01, *** P < 0.001. HIF1α, hypoxia-inducible factor 1α; PASMC, pulmonary artery smooth muscle cell.

Under stress conditions, p53 is hyper-phosphorylated at multiple sites. To address whether the p53-mediated suppression of miR-17∼92 is due to the hyperphosphorylation of p53, we exposed PASMC cells to hypoxia for different periods of time (30 min, 60 min, 24 h, and 48 h) and measured the phosphorylation of p53 at multiple sites. We found that hypoxia exposure did not change the phosphorylated or total p53 protein levels by 30 or 60 min (Supplemental Fig. S4), while it significantly induced ser15 phosphorylation of p53 by 24 h (Fig. 7). Ser6, ser9, ser20, ser46, or ser392 phosphorylation levels were unchanged by either 24 or 48 h hypoxia exposure (Fig. 7). We were unable to detect phosphorylation of p53 at ser37 or Thr81 (data not shown). These data suggest that the ser15 phosphorylation is likely to be responsible for p53-mediated repression of miR-17∼92.

Figure 7.

Hypoxia induces phosphorylation of p53. Human PASMC were exposed to normoxia or hypoxia (3% O₂) for 24 or 48 h. The total and site-specific phosphorylation of p53 were detected with specific antibodies, with Tubulin as loading control for equal loading (n = 3–4). Representative blots and quantification (shown as means ± SD) were presented. Nonparametric Mann–Whitney test was used for statistical analysis. *P ≤ 0.05. PASMC, pulmonary artery smooth muscle cell.

To further confirm the critical role of ser15 phosphorylation in p53/miR-17∼92 binding, we performed ChIP using a ser15 phospho-p53 antibody and qPCR with the immunoprecipitated DNA fragments using primers covering the identified p53 binding site (BS1 and BS2). As shown in Fig. 8, we found that DNA fragments that covered both p53 binding sites (BS1 and BS2) were significantly enriched when ser phosphor-p53 antibody was applied, confirming the binding of phosphorylated p53 (ser15) to the promoter of miR-17∼92 in prolonged hypoxia exposure.

Figure 8.

Phospho-P53 (ser15) binds to miR-17∼92 promoter in prolonged hypoxia exposure. Human PASMC were exposed to normoxia or hypoxia (3% O₂, H) for 24 h and their nuclear extract was collected. DNA fragments were precipitated with phosphor-P53 (ser15) antibody. After purification, the DNA fragments were subjected to qPCR analysis with human miR-17∼92 promoter-specific primers. The DNA fragments from IgG-treated (IgG) or untreated (Input) nuclear extracts were used as negative control and positive control, respectively. Input was set as 1 and phosphor-P53 or IgG groups were compared with respective input (n = 3). Data are shown as % of Input (means ± SD). Signal in hypoxia samples was compared with that in normoxia samples. Two-way ANOVA analysis was performed. *P < 0.05; **P < 0.01. PASMC, pulmonary artery smooth muscle cell.

DISCUSSION

Our current study demonstrates that hypoxia regulates the expression of miR-17∼92 in PASMC in a temporal manner: an early induction of miR-17∼92 via HIF1/E2F1 and an inhibition at a later stage via p53 phosphorylation. These data are consistent with the biphasic expression of miR-17∼92 we have observed in lungs of mice exposed to hypoxia and developed PH (2). However, we are lacking in vivo data to confirm if this biphasic expression of miR-17∼92 in PASMC reflects a biphasic function of miR-17∼92 in hypoxia-mediated PH: if upregulation of miR-17∼92 in PASMC in vivo in the early phase of hypoxia exposure promotes cell proliferation, leading to pulmonary vascular remodeling and PH, whereas reduction of miR-17∼92 in the late phase reflects an adaptive response of the body to inhibit further progression of pulmonary vascular remodeling and PH by inhibition of proliferation of PASMC in vivo (2). Further studies are warranted to support this conclusion.

The biphasic expression of miR-17∼92 cluster during hypoxia exposure also cautions us to be careful when interpreting clinical biomarker data without knowledge of the time frame of the disease progression when the clinical data are collected. During human development, genes are turned on and off at different stages of development, and in some diseases, these developmental processes are recapitulated. The molecular mechanisms underlying these temporal changes in gene expression both during normal development and during disease development are not well understood. We have shown that HIF and p53 phosphorylation controls the expression of miR-17∼92 in a time-dependent manner (Figs. 3, 4, 5, 6, and 7), shedding some novel insights into the temporal regulation of this miRNA cluster during the development and progression of PH.

Although HIF is generally thought to be a prohypertensive factor (15–17), a recent report suggests that loss of HIF1α in PASMC increases hypoxia-induced vasoconstriction (18), suggesting a complex role for HIF1α in PH. This seeming inconsistency may be explained by the role of p53, as there is a negative correlation between p53 and HIF activity (19–21) and p53 can interact with HIF and decrease HIF activity (22). We show in this paper that p53 phosphorylation can inhibit HIF downstream targets (Figs. 5, 6, and 7) in the presence of HIF1, providing a mechanism by which a HIF switch occurs.

Although here we show that HIF1α is an upstream regulator of miR-17∼92 (Figs. 2, 3, and 4), previously we also reported that miR-17/20a of the miR-17∼92 cluster directly targets prolyl hydroxylases 2 (PHD2) in PASMC, and that inhibition of miR-17∼92 led to decreased HIF1α levels in PASMC in hypoxia (3% O₂, 6 h) (23). These findings suggest a positive feedback loop of HIF1α/miR-17∼92 regulation in PASMC: although induction of HIF1α in hypoxia exposure (3% O₂, 6 h) upregulates the expression of miR-17∼92 in PASMC, miR-17∼92 may also induce HIF1α level via suppression of PHD2. However, others also found that miR-17∼92 can directly target and inhibit HIF1α in lung cancer cells and macrophages (24, 25), suggesting a possible negative feedback loop between HIF1 and miR-17∼92 in the specific cell types tested. Thus, the biological consequence of this HIF1α/miR-17∼92 feedback is context-dependent (Figs. 2, 3, and 4) (23–25).

We showed that reduction of miR-17∼92 requires the phosphorylation of p53 at ser15 (Fig. 7). In the absence of p53, there is no phosphorylation of p53 at ser15 to inhibit miR-17∼92, leading to the maintenance of miR-17∼92 induction and PASMC proliferation. Our results suggest that p53 is required for the reduction of miR-17∼92 in the late stage of PH and prolonged hypoxia; thus, p53 may be playing a protective role in PH, consistent with recent reports by Mizuno et al. (21) and Mouraret et al. (26). This suggests that the ultimate function of any given gene may be confounded by the type of posttranslational modification of the protein.

Taken together, our studies provide new evidence regarding the underlying mechanism of the biphasic expression of miR-17∼92 in PASMC during continuous hypoxia. Although we did observe similar biphasic expression of miR-17∼92 in mouse lung during hypoxia exposure (2), further studies are warranted to determine if similar regulatory mechanisms are playing a role in vivo and how this biphasic expression of miR-17∼92 contributes to the pathogenesis of PH.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL MATERIAL

Supplemental Figs. S1–S4, Supplemental Blots (Figs. S2, S3, S6, S7), and Supplemental Table S1: https://doi.org/10.6084/m9.figshare.25166309.

GRANTS

This work was partly supported by the NIH HL123804 (to J. U. Raj and G. Zhou), a Gilead Sciences Research Scholars Program in Pulmonary Arterial Hypertension Award (to T. Chen), and an American Heart Association Career Development Award (AHA Award No. 18CDA34110301, to T. Chen).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.-H.S., G.Z., J.U.R., and T.C. conceived and designed research; M.R.S., S.G., J.B.H., Q.Z., A.C., R.A. and H.-L.Y., and T.C. performed experiments; M.R.S., S.G., J.B.H., G.Z., J.U.R., and T.C. analyzed data; G.Z., J.U.R., and T.C. interpreted results of experiments; G.Z. and T.C. prepared figures; G.Z. and T.C. drafted manuscript; M.R.S., S.G., J.B.H., H.-L.Y., G.Z., J.U.R., and T.C. edited and revised manuscript; G.Z., J.U.R., and T.C. approved final version of manuscript.