Introduction

Age-related macular degeneration (AMD) is a chronic degenerative disorder of the retina that affects millions of elderly people globally, resulting in severe visual loss and blindness1. AMD is primarily divided into two subtypes: non-neovascular AMD (“dry” AMD) and neovascular AMD (“wet” AMD). Although AMD is primarily categorized into these two major subtypes—non-neovascular AMD (“dry” AMD) and neovascular AMD (“wet” AMD)—the clinical course and prognosis can vary widely between them2. While wet AMD is notorious for its potential to cause rapid and severe vision loss, dry AMD, often perceived as having a more benign course, can progress to geographic atrophy (GA), a late-stage manifestation characterized by loss of retinal pigment epithelium (RPE) and photoreceptor cells. This condition significantly impacts central vision and quality of life, thus challenging the assumption of a universally favorable prognosis for all cases of dry AMD. Dry AMD, accounting for 80%-85% of AMD cases, may progress to wet AMD when vascular hyper-proliferation or enhanced angiogenesis occurs due to the significant production of vascular endothelial growth factor (VEGF)1. At present, anti-angiogenesis therapies targeting VEGF have become the standard of care for wet AMD3. Nonetheless, some patients are insensitive to anti-VEGF therapy or experience reduced efficacy of VEGF inhibitors, and resistance to this therapy frequently occurs after repeated administration4,5; hence, novel anti-VEGF agents are needed. Recent research has been conducted to study the molecular factors that affect VEGF expression during AMD progression. For example, numerous studies have documented several key proteins that can regulate the production and release of VEGF in AMD6,7.

Arachidonate 5-lipoxygenase (ALOX5), a member of the lipoxygenase family of enzymes, is a pivotal enzyme in the synthesis of leukotrienes and has been found to be abnormally expressed in human inflammatory disorders8,9,10. Owing to its non-heme iron-containing property, ALOX5 has been established as an inducer of cell ferroptosis through catalyzing lipid peroxidation11,12. There is ample evidence that ALOX5 can contribute to the pathogenesis of various diseases. For example, polymorphism of the ALOX5 gene can contribute to a high susceptibility to atherothrombosis and thus enhances atherosclerosis development in middle-aged individuals13. ALOX5 is present at high levels in breast cancer and has a promoting function in cell growth and migration via the AKT pathway14. Enhanced expression of ALOX5 has been observed in patients with diabetes and is associated with the development of severe COVID-1915. Additionally, knocking down ALOX5 can exert a cardioprotective effect in ischemia–reperfusion injury16.

As a critical player in the inflammatory cascade, ALOX5 is capable of inducing angiogenesis in diverse tumor types, such as breast cancer and oral cancer17,18. Abnormal expression of ALOX5 is associated with the development of AMD19. Furthermore, inhibition of ALOX5 exhibits mitigatory effects on sodium iodate-induced lipid peroxidation in retinal pigment epithelium (RPE) cells, neuroretina degeneration, and cell death in the retina photoreceptor layer, thereby ameliorating AMD progression20. However, no studies have proven the causal role of ALOX5 in angiogenesis during AMD.

Hypoxia can induce many changes, such as VEGF expression in RPE cells, and thus contributes to vascular dysfunctions and neovascularization, promoting the progression of AMD21. Hypoxia-induced RPE cells have been widely used for investigation of neovascular AMD pathogenesis22,23. In the current study, we used the RPE cell line ARPE-19 with hypoxia exposure to elucidate the role of ALOX5 in angiogenesis underlying AMD pathogenesis.

Results

Hypoxia promotes cell growth and increases VEGF expression in ARPE-19 cells

To validate the influence of hypoxia in growth and VEGF expression of ARPE-19 cells, we first exposed ARPE-19 cells to hypoxia for 24 h. Through CCK-8 assays, we found that exposure of hypoxia led to a remarkable promotion in cell viability compared with normoxia control (Fig. 1A). However, considering the subsequent results of PCNA, this enhancement in viability is likely attributed to cell proliferation. To explore this further, we conducted an EdU assay to evaluate cell proliferation status, which indicated that hypoxia indeed significantly promoted cell proliferation, as evidenced by the increased number of EdU-positive cells (Fig. 1B). To validate these findings, we employed immunofluorescence and western blotting to analyze the expression of the cell cycle marker PCNA in ARPE-19 cells under hypoxic conditions. ARPE-19 cells exposed to hypoxia exhibited a higher fluorescence intensity of PCNA compared with controls (Fig. 1C), demonstrating the promotion of hypoxia exposure in ARPE-19 cell growth. Furthermore, hypoxia-exposed ARPE-19 cells showed a high expression of VEGF, measured through immunofluorescence (Fig. 1D). Additionally, the secretion level of VEGF was strongly augmented by hypoxia (Fig. 1E).

Fig. 1
figure 1

Hypoxia enhances cell proliferation and VEGF expression in ARPE-19 cells. (A) CCK-8 cell viability assays in ARPE-19 cells exposed to hypoxia for 24 h or normoxia conditions. (B) EdU cell proliferation assays in hypoxia- or normoxia-exposed ARPE-19 cells. Shown were representative images. (C and D) Immunofluorescence assays (C) and western blotting (D) for expression of PCNA in hypoxia- or normoxia-exposed ARPE-19 cells. Shown were representative images. (E) Immunofluorescence assays for expression of VEGF in hypoxia- or normoxia-exposed ARPE-19 cells. (F) The secretion level of VEGF in ARPE-19 cells treated as indicated by ELISA. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

To validate the influence of hypoxia on the growth and VEGF expression of ARPE-19 cells, we first exposed ARPE-19 cells to hypoxia for 24 h. Through CCK-8 assays, we found that exposure to hypoxia led to a remarkable increase in cell viability compared with the normoxia control (Fig. 1A). However, considering the subsequent results for PCNA, this enhancement in viability is likely attributed to cell proliferation. To explore this further, we conducted an EdU assay to evaluate the cell proliferation status, which indicated that hypoxia indeed significantly promoted cell proliferation, as evidenced by the increased number of EdU-positive cells (Fig. 1B). To validate these findings, we employed immunofluorescence and western blotting to analyze the expression of the cell cycle marker PCNA in ARPE-19 cells under hypoxic conditions. ARPE-19 cells exposed to hypoxia exhibited a higher fluorescence intensity (Fig. 1C) and increased protein expression (Fig. 1D) for PCNA compared with controls, demonstrating that hypoxia exposure promotes ARPE-19 cell growth. Furthermore, hypoxia-exposed ARPE-19 cells showed a high expression of VEGF, as measured by immunofluorescence (Fig. 1E). Additionally, the secretion level of VEGF was strongly augmented by hypoxia (Fig. 1F).

Hypoxia upregulates the expression of ALOX5 in ARPE-19 cells

To identify important genes implicated in AMD pathogenesis, we downloaded the DEGs (differentially expressed genes) from AMD individuals in the GSE115828 dataset, based on the RNA-seq profile available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE115828. Subsequently, we constructed modules using the WGCNA (Weighted Gene Co-expression Network Analysis) method. Through dynamicMods, eight gene co-expression modules were identified among these DEGs, based on a scale-free network with an optimum soft threshold power value of 24 and a scale-free R2 of 0.85 (Fig. 2A and B, and Supplementary Table 1). Notably, the genes in the grey module could not be assigned to any of the other modules because they were not co-expressed with other genes. The Minnesota Grading System (MGS) is a method used to analyze the histopathology at each stage of AMD and to elucidate pathophysiologic changes in AMD, with MGS1 to MGS4 representing progressively more severe disease stages24. The tan module was significantly and positively correlated with the MGS scores of AMD patients (Fig. 2B), and we therefore selected this module for further analyses. By intersecting the 153 genes in the tan co-expression module with the 1776 DEGs identified in MGS3 and MGS4 samples versus MGS1 and MGS2 samples from the GSE115828 dataset, we found ten genes associated with AMD severity, as illustrated by the Venn diagram (Fig. 2C and Supplementary Table 2).

Fig. 2
figure 2

ALOX5 is associated with MGS of AMD and hypoxia increases ALOX5 mRNA expression in ARPE-19 cells. (A) WGCNA analysis of the DEGs in AMD individuals from GSE115828 dataset. (B) Association analysis of the gene expression levels of eight modules with the five clinical features (MGS, gender, age, RIN: RNA integrality, PMI: postmortem interval) in AMD samples. Red: positive correlation; green: negative correlation; numerical value in the module: Pearson’s correlation coefficient (r); numerical value within parentheses: P value. (C) Venn diagram showing genes associated with MGS of AMD. Cyan: 153 genes in tan co-expression module by WGCNA; Red: 1776 DEGs in MGS3-4 samples versus MGS1-2 samples from GSE115828 dataset. (D) Relative expression of ALOX5 mRNA by qRT-PCR in ARPE-19 cells exposed to hypoxia for 24 h or normoxia conditions, with β-actin as a reference gene. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

Among the ten genes, we focused on ALOX5 in this study, considering its crucial role in regulating angiogenesis17,25. Additionally, in the public dataset used for AMD research, we observed a slight downregulation in the expression of ALOX5 (log2 fold change = − 0.173), however, the adjusted p-value did not indicate any significant difference. We then utilized qRT-PCR to analyze the effect of hypoxia on ALOX5 expression. ARPE-19 cells exposed to hypoxia exhibited a higher expression of ALOX5 mRNA compared to normoxia-treated cells (Fig. 2D). These data suggest that hypoxia in ARPE-19 cells enhances the expression of ALOX5, a protein associated with AMD severity as measured by the Minnesota Grading System (MGS).

Silencing ALOX5 reduces VEGF expression induced by hypoxia in ARPE-19 cells

Given our data indicating that hypoxia induces ALOX5 upregulation, we next aimed to evaluate the influence of ALOX5 on VEGF expression. The lentiviral vector targeting ALOX5 (ALOX5-shRNA1, ALOX5-shRNA2, or ALOX5-shRNA3) was used to knock down ALOX5 expression. Transfection with the ALOX5-shRNA lentiviral vector, but not the shNC control, significantly downregulated the level of ALOX5 mRNA in ARPE-19 cells (Fig. 3A). For subsequent experiments, a mixture of the lentiviral vectors ALOX5-shRNA1 and ALOX5-shRNA2 (also referred to as shALOX5) was applied due to their high efficacy in downregulating ALOX5 (Fig. 3A). Depletion of ALOX5 through shALOX5 transfection in ARPE-19 cells reduced VEGF expression induced by hypoxia (Fig. 3B). Furthermore, ALOX5 silencing inhibited hypoxia-induced VEGF secretion in ARPE-19 cells (Fig. 3C). Consequently, we conclude that hypoxia induces VEGF expression in ARPE-19 cells, at least partially, by upregulating ALOX5.

Fig. 3
figure 3

ALOX5 depletion decreases VEGF expression induced by hypoxia in ARPE-19 cells. (A) qRT-PCR of ALOX5 mRNA level in ARPE-19 cells after transfection by ALOX5-shRNA1, ALOX5-shRNA2, ALOX5-shRNA3, or shNC lentiviral vector. (B and C) ARPE-19 cells were introduced with shALOX5 or shNC lentiviral vector before hypoxia exposure or normoxia culture and assayed for VEGF expression by immunofluorescence (B) and VEGF secretion level by ELISA (C). *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

ALOX5 regulates the viability and migration of HUVECs

To evaluate whether ALOX5 expression in hypoxia-exposed ARPE-19 cells can affect the viability and migration of HUVECs, we incubated HUVECs with the relevant conditioned media (CM). Remarkably, incubation with hypoxia-ARPE-19-CM resulted in increased viability (Fig. 4A) and migration (Fig. 4B) of HUVECs compared with the normoxia-ARPE-19-CM control. Interestingly, shALOX5-ARPE-19-CM significantly reduced the viability (Fig. 4A) and migration (Fig. 4B) of HUVECs. These data collectively suggest that hypoxia-ARPE-19-CM promotes the viability and migration of HUVECs, at least partly, by upregulating ALOX5.

Fig. 4
figure 4

ALOX5 regulates viability and migration of HUVECs. (A and B) ARPE-19 cells were transfected by shALOX5 or shNC lentiviral vector before hypoxia exposure and then the CM was collected and used to culture HUVECs. After 24 h, CCK-8 cell viability assay (A) and transwell migration assay (B) were performed in treated HUVECs. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

ALOX5 affects angiogenesis of CAM in vivo

CAM angiogenesis assays have been widely used to investigate changes in angiogenesis26. To determine the role of ALOX5 in angiogenesis, we performed CAM model experiments using shALOX5-ARPE-19-CM or shNC-ARPE-19-CM. Administration of shALOX5-ARPE-19-CM significantly reduced the total length of blood vessels formed in the CAM compared with the shNC control (Fig. 5), demonstrating that hypoxia-ARPE-19-CM enhances angiogenesis in the CAM, at least partly, via ALOX5 upregulation.

Fig. 5
figure 5

ALOX5 affects angiogenesis of CAM. On embryonic day 7, the CM derived from shNC- (shNC-ARPE-19-CM) or shALOX5-transfected (shALOX5-ARPE-19-CM) ARPE-19 cells under hypoxia was carefully added on the CAM surface and incubation was then performed for 48 h. Images were captured and total length of formed vessels was gauged. *P < 0.05.

Full size image

YTHDF1 binds to ALOX5 3′UTR and regulates ALOX5 expression by elevating its mRNA stability

To identify the RNA binding proteins (RBPs) of the ALOX5 3′UTR, we adopted the RAP-MS technique using total extracts of hypoxia-exposed ARPE-19 cells in biotinylated 3′UTR probes or a Bio-NC control. The precipitated proteins were subjected to qualitative proteomic analysis using an HPLC–MS/MS method. Following qualitative proteomic analysis, we employed GO enrichment analysis to observe the biological processes associated with these ALOX5 3′UTR pull-down proteins. Results showed that these pull-down proteins have close associations with mRNA-based processes, including mRNA stability, transport, and splicing (Fig. 6A). RBPs bind to the 3′UTR of mRNA to regulate diverse mRNA fates, including mRNA stability27. By intersecting the proteins associated with mRNA stability (List 1) (Supplementary Table 3) among these pull-down proteins with the putative RBPs of ALOX5 predicted by the online program starBase (List 2) (https://pubmed.ncbi.nlm.nih.gov/32452512/), a total of five proteins (FXR1, FMR1, YTHDF1, FXR2, and KHSRP) were identified (Fig. 6B). Among these five proteins, we found that the abundance level of the m6A reader YTHDF1 was higher in the Bio-3′UTR group than in the Bio-NC group, which was validated by western blotting using the corresponding pull-down proteins (Fig. 6C).

Fig. 6
figure 6

YTHDF1 depletion downregulates the expression of ALOX5 protein in hypoxia-stimulated ARPE-19 cells. (A) GO enrichment analysis of these pull-down proteins of the mix of Bio-3′UTR. (B) Venn diagram revealing the putative RBPs pulled down by ALOX5 3′UTR. List 1: the 14 proteins associated with mRNA stability in these pull-down proteins; List 2: the putative RBPs of ALOX5 predicted by online program starbase. (C) Representative western blotting images depicting YTHDF1 abundance in the pull-down proteins of Bio-3′UTR and Bio-NC groups. (D) qRT-PCR of YTHDF1 mRNA expression in ARPE-19 cells after transfection by vector of shYTHDF1#1, shYTHDF1#2, shYTHDF1#3, or shNC. (E) Relative ALOX5 mRNA expression by qRT-PCR in shNC- or shYTHDF1-transfected ARPE-19 cells under hypoxia or normoxia, with β-actin as a reference gene. (F) ARPE-19 cells transfected with shYTHDF1 vector or shNC control were exposed to normoxia or hypoxia conditions for 24 h and then treated with Actinomycin D for various time frames, followed by the detection of ALOX5 mRNA level. (G) Relative ALOX5 protein level in shNC- or shYTHDF1-transfected ARPE-19 cells under hypoxia or normoxia. (H) VEGF secretion levels in ARPE-19 cells transfected with shNC- or shYTHDF1 were detected by ELISA under hypoxia or normoxia conditions. Representative pictures were shown. GAPDH served as a loading control. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

To determine the regulation of YTHDF1 in ALOX5, we performed silencing experiments using a lentiviral vector containing an shRNA specific for YTHDF1 in ARPE-19 cells under hypoxia. The efficacy of shYTHDF1#1, shYTHDF1#2, and shYTHDF1#3 in downregulating YTHDF1 mRNA was confirmed by qRT-PCR (Fig. 6D). Consequently, a mixture of the three shRNA lentiviral vectors (shYTHDF1) was used to silence YTHDF1 in subsequent assays. In ARPE-19 cells, YTHDF1 depletion significantly reduced ALOX5 mRNA expression induced by hypoxia (Fig. 6E). We then employed Actinomycin D treatment to evaluate the effect on ALOX5 mRNA stability. Hypoxia exposure notably enhanced ALOX5 mRNA stability, whereas knockdown of YTHDF1 strongly reversed this effect (Fig. 6F). Furthermore, YTHDF1 knockdown abolished hypoxia-driven upregulation of ALOX5 protein in ARPE-19 cells (Fig. 6G). Collectively, these findings establish that YTHDF1 binds to the ALOX5 3′UTR to upregulate ALOX5 expression by increasing its mRNA stability. Additionally, to provide further insights into the research results of YTHDF1, we verified the effect of knocking down YTHDF1 on VEGF secretion in hypoxia-induced ARPE-19 cells. The results showed that knocking down YTHDF1 also inhibited VEGF secretion in hypoxia-induced ARPE-19 cells (Fig. 6H).

Discussion

The adverse progression of AMD is tightly associated with upregulated VEGF expression and enhanced angiogenesis1. Hypoxia can enhance VEGF expression and release from RPE-19 cells, thereby promoting the pro-angiogenic activity of endothelial cells (ECs) to migrate and proliferate, driving neovascularization and AMD progression21,28. In this paper, using ARPE-19 cells under hypoxic conditions, our data confirmed that hypoxia increases VEGF expression and release from ARPE-19 cells and promotes cell growth. Incubation of HUVECs with the conditioned medium from hypoxia-exposed ARPE-19 cells led to enhanced cell viability and migration. Based on the anti-angiogenic mechanism, anti-VEGF therapies for AMD have been developed29. Furthermore, anti-angiogenic genomic therapeutics have recently been proposed as potential treatment strategies30. Hence, defining molecular factors that affect VEGF expression and angiogenesis is of great significance. Here, for the first time, we identify ALOX5 as a strong inducer of VEGF expression in ARPE-19 cells under hypoxia and a potent regulator of angiogenesis. Moreover, we provide an upstream regulatory mechanism for ALOX5 expression through an RBP, YTHDF1.

ALOX5 participates in driving the development of various diseases due to its pro-inflammatory and pro-ferroptotic properties8,10. Moreover, ALOX5 has been implicated in AMD pathogenesis19, and inhibition of ALOX5 can protect RPE cells from oxidative stress-induced cell death and injury20,31. In the current work, bioinformatics analysis predicted a close association between ALOX5 expression and AMD progression. Intriguingly, we first demonstrate that hypoxia can upregulate ALOX5 expression in ARPE-19 cells. Emerging evidence also indicates that dysregulated ALOX5 might be crucial in enhancing angiogenesis25,32. We thus hypothesize that hypoxia-induced ALOX5 might contribute to angiogenesis in AMD. Through shRNA silencing experiments, we first establish that hypoxia-induced VEGF expression in ARPE-19 cells is partly dependent on ALOX5 upregulation. More interestingly, we first uncover that shALOX5-ARPE-19-CM can suppress the viability and migration of HUVECs and diminish angiogenesis in CAM. These findings suggest that hypoxia-induced ALOX5 might promote angiogenesis by upregulating VEGF expression and enhancing the migration and growth of vascular ECs.

YTHDF1, as the most abundant and potent m6A reader, can regulate mRNA stability and degradation, as well as direct mRNA translation, through interactions with mRNA 3′UTRs, thereby regulating target gene expression33,34. Emerging evidence implicates dysregulated YTHDF1 as a critical player in the pathogenesis of various diseases by influencing its downstream targets33. For instance, YTHDF1 can promote the development of ovarian cancer and colorectal cancer by modulating the translation of target genes35,36. YTHDF1 elevates MAGED1 translation to exert a regulatory function in pulmonary hypertension pathogenesis through m6A methylation modification37. YTHDF1 also affects inflammatory responses by enhancing NLRP3 translation, thereby regulating inflammatory diseases38. Furthermore, YTHDF1 weakens sepsis pathogenesis by promoting NLRP3 ubiquitination and diminishing pyroptosis through enhanced WWP1 expression39. Additionally, YTHDF1 is implicated in hypoxia adaptation and tumor pathogenesis40,41. Moreover, YTHDF1 exhibits regulatory activity in tumor angiogenesis42,43. Nonetheless, the role of YTHDF1 in regulating angiogenesis and AMD pathogenesis remains to be fully explored. In this study, we first demonstrate that YTHDF1 can bind to the ALOX5 3′UTR. Importantly, we uncover, for the first time, that YTHDF1 regulates ALOX5 expression by enhancing the stability of ALOX5 mRNA in ARPE-19 cells under hypoxia. YTHDF1 typically regulates target gene expression in an m6A-dependent manner33,44. Future work will investigate whether the YTHDF1-mediated enhancement of ALOX5 mRNA stability is due to modifications in m6A methylation. However, such research is currently hampered by the lack of in vivo analyses using AMD animal models.

Taken together, our findings indicate that YTHDF1-regulated ALOX5 functions as a contributor to VEGF expression in hypoxia-exposed ARPE-19 cells and promotes the viability, migration, and angiogenesis of vascular ECs. Our present study suggests that strategies aimed at inhibiting ALOX5 might have therapeutic value for AMD due to their potential anti-angiogenic activity.

Materials and methods

Cell line and hypoxia exposure

A retinal pigment epithelial ARPE-19 cell line (CL-0026, Procell, Wuhan, China) was cultivated in a 5% CO2 humidified incubator at 37 °C in growth DMEM/F12 (1:1) (Procell) with the addition of 1% penicillin–streptomycin (Servicebio, Wuhan, China) and 10% FBS (Life Technologies, Darmstadt, Germany). Human umbilical vein endothelial cells (HUVECs, PCS-100-013, ATCC, Manassas, VA, USA) were propagated under standard protocols (Endothelial Cell Growth Medium 2, PromoCell, Heidelberg, Germany). For induction of hypoxia, ARPE-19 cells were maintained in an incubator of 1% O2 content45 for 24 h.

Online data analysis

We used the online dataset GSE115828 from GEO database to observe the differentially expressed genes (DEGs) in individuals with AMD at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE115828. The association between the DEGs and clinical features (Minnesota Grading System (MGS), gender, age, RNA integrality (RIN), and postmortem interval (PMI)) of AMD patients was analyzed by the weighted gene co-expression network analysis (WGCNA) with WGCNA R package under standard protocols46.

Gene knockdown with shRNA

ShRNA lentiviral vectors used in this study were procured from Miaolingbio (Wuhan, China) including pLV3-U6-ALOX5(human)-shRNA1-EGFP-Puro (ALOX5-shRNA1), pLV3-U6-ALOX5(human)-shRNA2-EGFP-Puro (ALOX5-shRNA2), pLV3-U6-ALOX5(human)-shRNA3-EGFP-Puro (ALOX5-shRNA3), pLV3-U6-YTHDF1(human)-shRNA1-mCherry-Puro (shYTHDF1#1), pLV3-U6-YTHDF1(human)-shRNA2-mCherry-Puro (shYTHDF1#2), pLV3-U6-YTHDF1(human)-shRNA3-mCherry-Puro (shYTHDF1#3), and pLV3-U6-Scrambled-shRNA control (shNC). A vector mix of ALOX5-shRNA1 and ALOX5-shRNA2 (ALOX5-shRNA1:ALOX5-shRNA2 = 1:1, called shALOX5) was used to establish the stable ALOX5 knockdown cell line. To generate an stable YTH N6-methyladenosine RNA-binding protein 1 (YTHDF1) silencing cell line, the lentiviral vector mix of shYTHDF1#1, shYTHDF1#2 and shYTHDF1#3 (shYTHDF1#1:shYTHDF1#2:shYTHDF1#3 = 1:1:1, called shYTHDF1) was applied. ARPE-19 cells were transfected with these lentiviral vectors using RFect Plasmid DNA Transfection Reagent as suggested by the manufacturer (Baidai, Changzhou, China) and subsequently the knockdown efficacy was analyzed by quantitative real-time PCR (qRT-PCR).

RNA expression by qRT-PCR

Total RNA from ARPE-19 cells after hypoxia exposure or/and transfection as indicated was prepared using Direct-zol RNA MiniPrep Kit as per the manufacturing suggestion (Zymo Research, Irvine, CA, USA), followed by the quantification by NanoDrop 1000 (Thermo Fisher Scientific, Runcorn, UK). RNA (0.5 μg) was converted to cDNA with ReverTra Ace RT Master Mix as described by the manufacturer (Toyobo, Osaka, Japan). Amplification by qRT-PCR was done using the THUNDERBIRD qPCR Mix (Toyobo) with 200 nM specific primers (ALOX5-forward: 5′-TGGCGCGGTGGATTCATAC-3′, ALOX5-reverse: 5′-TGTGCAGGGGTCTGTTTTGT-3′, YTHDF1-forward: 5′-CGTGGACACCCAGAGAACAA-3′, YTHDF1-reverse: 5′-GTACAAATAAATGCAGATCCATC-3′) in a 25 μl final reaction mix. Relative expression, where applicable normalized to β-actin (β-actin-forward: 5′-GCGCTCGTCGTCGACAACG-3′, β-actin-reverse: 5′-GGGGTACTTCAGGGTGAGGA-3′), was determined by the 2−ΔΔCt method.

Cell proliferation assay

ARPE-19 cells were plated on a 96-microwell dish 12 h before hypoxia exposure and were assayed for proliferation with Click-iT EdU-555 Cell Proliferation Kit from Servicebio. Following a 2-h incubation with EdU reagent (10 μM), the cells were treated with IF555 fluorochrome for EdU staining and subsequently incubated with Hoechst 33,342 solution for nuclear staining, based on the manufacturer’s recommendations. The EdU-positive cells (red) were determined relative to total nuclei (blue) under fluorescence microscopy with high-power fields (× 200, Leica, Wetzlar, Germany).

Immunofluorescence

ARPE-19 cells transfected with or without shALOX5 or shNC lentiviral vector were seeded into 24-well culture dishes and then carried out hypoxia exposure or normoxia culture. After PBS washing, the cells were fixed for 15 min with 4% paraformaldehyde. PBS containing 0.5% Triton X-100 was applied for permeabilization. Following the blocking with 3% BSA for 20 min at room temperature, the cells were probed with anti-PCNA (mouse monoclonal, Cat#60097-1-Ig, 1:500, Proteintech, Wuhan, China) or anti-VEGF (mouse monoclonal, Cat#66828-1-Ig, 1:300, Proteintech) antibody overnight at 4 °C. The cells were subsequently exposed to goat anti-mouse IgG cross-adsorbed Alexa Fluor 488 (Cat#GB25301, 1:500, Servicebio) for 1 h in darkness. After nuclear staining by DAPI, images were acquired, and data analysis was done using ImageJ (NIH, Bethesda, MD, USA).

Enzyme-linked immunosorbent assay (ELISA)

Un-transfected, shNC-, or shALOX5-transfected ARPE-19 cells were performed hypoxia exposure or normoxia treatment. Then, the growth medium was collected to measure the secretion level of VEGF using Human VEGF ELISA Kit (Liankebio, Hangzhou, China) based on the manufacturer’s protocols. Plates were read at 450 nm, and VEGF level was calculated using the standard curve generated with the standard sample.

Preparation of the conditioned medium (CM)

ARPE-19 cells were transfected with or without shALOX5 or shNC lentiviral vector for 24 h and then subjected to hypoxia exposure or normoxia culture for 24 h, followed by the acquirement of the CM (named normoxia-ARPE-19-CM, hypoxia-ARPE-19-CM, shNC-ARPE-19-CM, and shALOX5-ARPE-19-CM).

Cell viability assay

For evaluation of cell viability, Cell Counting Kit-8 (CCK-8) was applied as per the accompanying instructions (SAB Biotech, Nanjing, China). Before viability assay, ARPE-19 cells (2 × 104 cells/well) grown in 96-well culture dishes were subjected to hypoxia exposure, or HUVECs were maintained for 24 h in the corresponding CM. After that, the above cells were incubated with 10% CCK-8 for 2 h. Cell viability was proportional to the absorbance at 450 nm.

Cell migration assay

For migration assay, 24-well transwell plates with 8 μm pore size membranes separating 2 chambers were purchased from BD Bioscience (Tokyo, Japan). HUVECs were re-suspended in the collected CM and loaded on the insert (upper chamber). The complete growth media were loaded in the lower chamber. Plates were subsequently incubated at 37 °C for 24 h. After staining by crystal violet (0.5%), the HUVECs migrating through the insert membranes into the lower chamber was visualized with high-power fields (× 100) on a microscope. The migratory cells were scored using ImageJ.

In vivo angiogenesis assay

The chicken embryo chorioallantoic membrane (CAM) is widely used as an in vivo model in investigating angiogenesis26. For CAM experiment, twelve fertilized chicken eggs were procured from Chaoda Poultry Co., LTD (Shenyang, China) and incubated in a brooder (HuiDaFuHuaJi, Dezhou, China) at 37.5 ± 0.5 °C and 60%-65% relative humidity. To observe the effect of ALOX5 on angiogenesis, the eggs were administered with shNC-ARPE-19-CM or shALOX5-ARPE-19-CM. Each group included six eggs. On embryonic day 7, after being disinfected, the eggs were cracked and the chorioallantoic membrane was carefully exposed. A sterile silicone ring with inside diameter of 4 mm was placed on the chorioallantoic membrane, and then 100 μl of the conditioned medium was added within the ring. After 48 h incubation at 37 °C, pictures were obtained and total length of formed vessels was analyzed by ImageJ. All animal experiments were performed with the approval of the Animal Ethics Committee of the First Affiliated Hospital of Zhengzhou University (No. 2015-LW-121) and the procedures for Care and Use of Laboratory Animals in fundamental research.

RNA affinity antisense purification-mass spectrometry (RAP-MS)

Two biotinylated oligonucleotides antisense to ALOX5 3′untranslated region (3′UTR) were synthesized by GenePharma (Shanghai, China). Biotinylated scrambled sequence as the control mock (Bio-NC) and positive control were also obtained from GenePharma. Bio-3′UTR was made and used for RAP as described elsewhere47, with minor modifications. In brief, after being subjected to hypoxia exposure for 24 h, ARPE-19 cells were washed with cold PBS, treated with 1% methanal and Glycine, as well as incubated with 1 mM PMSF. For lysate generation, 1.5 × 107 cells were re-suspended in 1 ml Total Cell Lysis Buffer (10 mM Tris–HCl pH 7.5, 500 mM LiCl, 0.2% SDS, 0.1% sodium deoxycholate, 0.2% NP-40) containing PMSF and RNase, followed by the addition of DNase and 1 ml Hybridization Buffer (20 mM Tris–HCl pH 7.5, 5 mM TCEP, 6 M guanidine thiocyanate, 0.2% sodium deoxycholate, 0.4% SDS, 0.4% NP-40, 1 M LiCl, 10 mm EDTA). Total exactions denatured at 65 °C were then incubated with about 15 μl denatured probe mix as follows: 37 °C 30 min, 50 °C 5 min, and 37 °C 90–180 min, before addition of 35 μl Streptavidin-coated beads (Solarbio, Beijing, China) for 30 min at 45 °C. Beads were harvested, washed in Hybridization Buffer and bound proteins were eluted using Protein Acid Elution reagent for MS analysis.

MS analysis was carried out by Qinglianbio Biotechnology Co., Ltd. (Beijing, China) using RIGOL L-3000 HPLC System (RIGOL, Beijing, China). Data analysis was done with Proteome Discoverer2.4 software (Thermo Fisher Scientific). GO enrichment analysis of these pull-down proteins was done using DAVID database at https://david.ncifcrf.gov/.

Western blotting

ARPE-19 cells transfected by shYTHDF1 or shNC lentiviral vector were performed hypoxia exposure or normoxia treatment and re-suspended for protein extraction in Extraction Buffer consisted of 2% SDS, 100 mM Tris–HCl pH 7.5, 0.4% NP-40, 1 mM EDTA, and 1 × protease inhibitor cocktail (Life Technologies). For protein concentration analysis, the Rradford Protein Assay Reagent was applied as recommended by the manufacturers (Bio-Rad, Hemel Hempstead, UK). About 15 μg protein was resolved on 10% polyacrylamide gels, and the resulting gels were then blotted on nitrocellulose (Bio-Rad) by semidry blotting. Primary antibodies used for immunoblotting included YTHDF1 (Cat#17479-1-AP, 1:3000), ALOX5 (Cat#10021-1-Ig, 1:1000), and GAPDH (Cat#10494-1-AP, 1:10000), and secondary antibody was goat anti-rabbit IgG coupled by HRP (Cat#SA00001-2, 1:5000), all of which were procured from Proteintech. For blot development, the Chemiolominescent Substrate (EuroClone, Milan, Italy) was employed, and densitometry analysis was done with Image Lab Software (Bio-Rad) after visualization by Amershem Imager 680 (GE Healthcare, Madison, WI, USA).

Actinomycin D assay

To elucidate the influence of YTHDF1 in stability of ALOX5 mRNA, ARPE-19 cells were transfected by shYTHDF1 or shNC lentiviral vector before hypoxia exposure and subsequently incubated with 2 μg/ml Actinomycin D (Sigma-Aldrich, Steinheim, Germany) to block transcription. At 0, 1, 2, 3, or 4 h after Actinomycin D treatment, whose cellular RNA was processed by qRT-PCR for quantification of ALOX5 mRNA level.

Statistical analysis

All assays were repeated at least three independent biological repeats. Unless otherwise noted, mean ± SD were presented in the graphs. Quantitative data were analyzed either t-test (unpaired, two groups) or ANOVA (one-way, three and four groups). Significance was set at p < 0.05.