The HIF‐1α/p53/miRNA‐34a/Klotho axis in retinal pigment epithelial cells promotes subretinal fibrosis and exacerbates choroidal neovascularization

Abstract Wet age‐related macular degeneration (wAMD), characterized by choroidal neovascularization (CNV), is a leading cause of irreversible vision loss among elderly people in developed nations. Subretinal fibrosis, mediated by epithelial‐mesenchymal transition (EMT) of retinal pigment epithelium (RPE) cells, leads to unsuccessful anti‐vascular endothelial growth factor (VEGF) agent treatments in CNV patients. Under hypoxic conditions, hypoxia‐inducible factor‐1α (HIF‐1α) increases the stability and activation of p53, which activates microRNA‐34a (miRNA‐34a) transcription to promote fibrosis. Additionally, Klotho is a target gene of miRNA‐34a that inhibits fibrosis. This study aimed to explore the role of the HIF‐1α/p53/miRNA‐34a/Klotho axis in subretinal fibrosis and CNV. Hypoxia‐induced HIF‐1α promoted p53 stability, phosphorylation and nuclear translocation in ARPE‐19 cells (a human RPE cell line). HIF‐1α‐dependent p53 activation up‐regulated miRNA‐34a expression in ARPE‐19 cells following hypoxia. Moreover, hypoxia‐induced p53‐dependent miRNA‐34a inhibited the expression of Klotho in ARPE‐19 cells. Additionally, the HIF‐1α/p53/miRNA‐34a/Klotho axis facilitated hypoxia‐induced EMT in ARPE‐19 cells. In vivo, blockade of the HIF‐1α/p53/miRNA‐34a/Klotho axis alleviated the formation of mouse laser‐induced CNV and subretinal fibrosis. In short, the HIF‐1α/p53/miRNA‐34a/Klotho axis in RPE cells promoted subretinal fibrosis, thus aggravating the formation of CNV.


| INTRODUC TI ON
Age-related macular degeneration (AMD) is the main cause of irreversible vision loss among elderly individuals in developed countries.
Late AMD is categorized into two types, wet and dry, by the presence of choroidal neovascularization (CNV) or geographic atrophy (GA) involving the macular centre, respectively. Although wet AMD accounts for only 10% of AMD cases, this type is responsible for 90% of AMD-induced vision loss. 1 The reason for the high blindness ratio of wet AMD is that CNV may progress to end-stage subretinal fibrosis.
Subretinal fibrosis is characterized by complex interactions between cellular components and local inflammatory factors within subretinal lesions, which ultimately leads to the reconstruction of the extracellular matrix (ECM) and subretinal scar formation. Among these interactions, epithelial-mesenchymal transition (EMT) of retinal pigment epithelium (RPE) cells is the critical contributor. 2 In the context of AMD, RPE cells lose their cell-cell adhesions and apical-basal polarity, transforming into mesenchymal cells via EMT. 3 Multiple extracellular ligands, such as galectin-1 4 and interleukin-2 (IL-2), 5 are involved in the initiation and development of the EMT programme in RPE cells. 6 Thoroughly studied for its central role in control of EMT, the ligand transforming growth factor-beta (TGF-β) is identified as the master regulator of EMT process. 7,8 At present, anti-vascular endothelial growth factor (VEGF) agents have become the first-line drugs for the treatment of CNV. Although anti-VEGF agents usually stabilize or improve visual acuity, subretinal fibrosis develops in approximately 50% of treated eyes within 2 years following anti-VEGF therapy. 9 The formation of subretinal fibrosis can lead to local damage to the RPE, photoreceptors and choroidal vessels, resulting in continuous malfunction of the macular visual system. Thus, targeting subretinal fibrosis is a novel strategy for the treatment of CNV.
Until now, the pathogenesis of CNV is still vague. However, accumulating studies have revealed that hypoxia facilitates the progression of CNV. Hence, hypoxia-inducible factor-1α (HIF-1α) is the main regulator of oxygen homeostasis that participates in wet AMD. HIF-1α acts as a transcription factor of numerous target genes, among which p53 can promote fibrosis in the kidney 10 and lung. 11 P53 is also a transcription factor that is induced by hypoxia. HIF-1α not only binds to the hypoxia response element 3 (HRE3) region on the p53 promoter to activate the transcription of p53, 12 but also enhances the stability of p53 13 and promotes the phosphorylation and nuclear translocation of p53. 14 Therefore, we investigated whether HIF-1α up-regulates p53 expression, stability and activation under hypoxic conditions to promote subretinal fibrosis during CNV.
MicroRNA (miRNA) is a type of small non-coding RNA that post-transcriptionally modulates gene expression. MiRNA-34a (miR-NA-34a) is a well-known miRNA regulated by p53. Similar to p53, miRNA-34a promotes lung 15 and liver 16 fibrosis. An Italian group determines that many miRNAs, including miRNA-34a, are up-regulated in the serum of neovascular AMD patients. 17 Additionally, miRNA-34a is significantly up-regulated, while tolerance to oxidative stress is reduced, in hydrogen peroxide-induced prematurely senescent ARPE-19 cells (a human RPE cell line), 18 indicating that miRNA-34a is involved in CNV.
Klotho (KL) is a membrane-bound protein that plays an anti-ageing role because Kl-null mice exhibit phenotypes similar to human premature ageing syndromes. 19 The level of circulating Klotho decreases with age and thereby increases the risk for age-associated diseases. 20 In cultured human RPE, KL up-regulates the expression of stress-related genes and reduces the production of reactive oxygen species (ROS), thus protecting RPE from oxidative stress-induced injury, 21 suggesting that KL plays a protective role in CNV.
MiRNA-34a down-regulates Klotho protein levels by directly binding to the three-prime untranslated region (3' UTR) of KL. Moreover, miRNA-34a promotes renal fibrosis by down-regulating KL in tubular epithelial cells. 22 Herein, we sought to determine the role of the HIF-1α/p53/miR-NA-34a/Klotho axis in subretinal fibrosis in CNV. Our data provide a novel approach for the treatment of wet AMD.

| Isolation of nuclear and cytoplasmic proteins
Nuclear and cytoplasmic proteins of ARPE-19 cells were isolated via NE-PER ® nuclear and cytoplasmic extraction reagents (#78833, Thermo Fisher Scientific).

| Quantitative real-time PCR (qRT-PCR)
Quantitative real-time PCR was performed as previously described. 23 U6 small nuclear RNA (snoRNA) was used as the internal control for miRNA-34a and pri-miRNA-34a. GAPDH mRNA was used as the internal control for Klotho. The relative levels of miRNA-34a, pri-miRNA-34a and Klotho were
In the luciferase reporter assay, ARPE-19 cells were cultured in 24well plates, and each well was co-transfected with firefly luciferase reporter plasmid (500 ng) and with miRNA-34a mimic or negative control RNA with Lipofectamine 3000 (#L3000008, Thermo Fisher Scientific).
An internal control reporter plasmid (10 ng) expressing Renilla reniformis luciferase was co-transfected to normalize the transfection efficiency.
Luciferase activities were measured 24 hours after transfection using a Renilla luciferase assay system (#E2810, Promega). Relative luciferase activity (arbitrary units) was expressed as fold changes over the control group after normalizing for the transfection efficiency.

| Chromatin immunoprecipitation (ChIP)
Four micrograms of DO-1 anti-p53 monoclonal antibody (#ab1101, Abcam) was used. Isotype-matched pre-immune mouse IgG  5 minutes and 10 minutes after fluorescein injection. In addition, fluorescein leakage was graded by two independent blinded observers using previously established criteria. 26 The total CNV area was analysed from ICGA images using ImageJ software.

| Statistical analyses
Experimental results are represented as the mean ± SEM. All data were analysed by Student's t test or one-way ANOVA with Tukey's post hoc test via GraphPad Prism software. A P value <.05 indicated significance.

| Hypoxia-induced HIF-1α promotes p53 stability, phosphorylation and nuclear translocation in ARPE-19 cells
First, p53 stabilization under hypoxic conditions regulated by HIF-1α was detected by CHX treatment, which showed that compared with the normal treatment, hypoxia increased p53 protein stability,
The relative protein level of HIF-1α compared with the GAPDH level (D) and the ratio of p-p53 (S15)/p53 (E), p-p53 (S20)/p53 (F), p-p53 (S46)/p53 (G) and p53/GAPDH (H) were analysed. *** P < .001, hypoxia group vs normal group. ## P < .01, hypoxia + digoxin group vs hypoxia group. I, Nuclear and cytoplasmic separation samples were prepared, and Western blot was performed to measure p53 protein levels. Histone 3 (H3) and GAPDH were used as the nuclear and cytoplasmic markers, respectively while the HIF-1α inhibitor digoxin decreased p53 protein stability ( Figure 1A,B). The activation of p53 is dependent on its phosphorylation at S15, S20 and S46 sites. 27 Western blot showed that HIF-1α, p-p53 (S15), p-p53 (S20) and p-p53 (S46) increased following hypoxia compared with their levels in the normal group, while digoxin down-regulated HIF-1α expression and phosphorylation of p53 ( Figure 1C-H). As a transcription factor, p53 enters the nucleus after its activation. 28 Nuclear and cytoplasmic separation showed that hypoxia induced the nuclear translocation of p53, which was inhibited by digoxin ( Figure 1I). Collectively, the above data indicated that hypoxia-induced HIF-1α promoted p53 stability, phosphorylation and nuclear translocation in ARPE-19 cells.

| HIF-1α-dependent p53 activation up-regulates miRNA-34a expression in ARPE-19 cells after hypoxia
A previous study revealed that p53 induces microRNA-34a (miRNA-34a) expression in multiple cancer cells, including osteosarcoma and breast cancer cells, as well as in irradiated mice, by binding to a specific p53-binding site in the gene that encodes miRNA-34a. 29 Therefore, miRNA-34a in ARPE-19 cells was detected, which showed that HIF-1α-dependent p53 activation promotes miRNA-34a expression ( Figure 2A). The discovery that modification of p53 affected miRNA-34a expression at the transcript level (pri-miRNA; Figure 2B) further supported the speculation that p53 enhanced miRNA-34a transcription in ARPE-19 cells. As shown in Figure 2C, In addition, the luciferase activity of p53-overexpressing cells was F I G U R E 2 Hypoxia-inducible factor-1α -dependent p53 activation promotes miRNA-34a expression in ARPE-19 cells following hypoxia. Human RPE cells were divided into the following groups: normal, hypoxia, hypoxia + AAV vector, hypoxia + digoxin, hypoxia + digoxin + nutlin-3a (MDM2 inhibitor and p53 agonist; 5 μmol/L for 24 h) and hypoxia + AAV-p53 mutant (S15A, S20A and S46A) infection. A, qRT-PCR was performed to measure miRNA-34a levels in ARPE-19 cells. B, qRT-PCR was performed to measure pri-miRNA-34a levels in ARPE-19 cells. In Figure 2A,B, ** P < .01, hypoxia group vs normal group. ## P < .01, # P < .05, compared with the hypoxia group. % P < .05, hypoxia + digoxin + nutlin-3a group vs hypoxia + digoxin group. C, The p53-binding sites in the miRNA-34 promoter. D, P53 regulated the miRNA-34a promoter. The p53 mutant resulted in a decrease in miRNA-34a promoter activity (miRNA-34a WT). The mutagenesis of the three p53-binding sites, singularly (miRNA-34a mutant1, miRNA-34a mutant2 and miRNA-34a mutant3) or in combination (miRNA-34a mutant1/2/3), abrogated this effect. E, ChIP was conducted with an anti-p53 antibody on ARPE-19 genomic DNA. The immunoprecipitated chromatin was found to be enriched with the target miRNA-34a promoter (miRNA-34a-1, miRNA-34a-2 and miRNA-34a-3, the regions encompassing each of three p53-binding sites) by qPCR. Data are reported as fold enrichment over control samples (immunoprecipitation with IgG) reduced by the p53BS mutant reporters compared with WT reporter ( Figure 2D comparison between white columns). This difference was abolished by p53 mutant infection ( Figure 2D comparison between black columns). These data suggest that miRNA-34a is a direct transcriptional target of p53. To further confirm that p53 bound to the miRNA-34a promoter, we performed ChIP. ARPE-19 cell lysates were immunoprecipitated with a p53 antibody, and the regions surrounding the three p53BS elements in the miRNA-34a promoter (miRNA-34a-1, miRNA-34a-2 and miRNA-34a-3) were amplified and quantified by quantitative PCR (qPCR). The qPCR-ChIP assay verified that the amplicons surrounding three p53BS elements were highly enriched compared with those in the negative control ( Figure 2E). The evaluation of p53 functional binding sites in the miRNA-34a promoter region strongly demonstrated that miRNA-34a is a direct transcriptional target of p53 in ARPE-19 cells.

| Hypoxia-induced p53-dependent miRNA-34a inhibits the expression of Klotho in ARPE-19 cells
Next, we aimed to identify the target mRNA of  Figure 3A,B). As shown in Figure 3C
In summary, our study demonstrates that the hypoxia-induced

CO N FLI C T O F I NTE R E S T
All authors declare that they have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.