USP14 Regulates ATF2/PIK3CD Axis to Promote Microvascular Endothelial Cell Proliferation, Migration, and Angiogenesis in Diabetic Retinopathy

Diabetic retinopathy (DR) is one of the leading causes of blindness in diabetic patients. However, the pathogenesis of DR is complex, and no firm conclusions have been drawn so far. It has become a hot spot in ophthalmology research to deeply study the mechanism of DR pathological changes and find effective treatment options. Human retinal microvascular endothelial cells (HRMECs) were induced by high glucose (HG) to construct DR cell model. CCK-8 assay was used to detect the viability of HRMECs. Transwell assay was used to detect the migration ability of HRMECs. Tube formation assay was used to identify the tube formation ability of HRMECs. The expressions of USP14, ATF2 and PIK3CD were detected by Western blot analysis and qRT-PCR assay. Immunoprecipitation (IP) was used to ascertain the relationship of USP14 and ATF2. To explore the regulatory relationship between ATF2 and PIK3CD by dual-luciferase reporter gene assay and Chromatin immunoprecipitation (ChIP) assay. High glucose treatment promoted the proliferation, migration, and tube formation of HRMEC, and the expressions of USP14, ATF2 and PIK3CD were significantly up-regulated. USP14 or ATF2 knockdown inhibited HG-induced HRMECs proliferation, migration, and tube formation. USP14 regulated the expression of ATF2, and ATF2 promoted PIK3CD expression. PIK3CD overexpression attenuated the inhibitory effectiveness of USP14 knockdown on proliferation, migration and tube formation of DR cell model. Here, we revealed that USP14 regulated the ATF2/PIK3CD axis to promote proliferation, migration, and tube formation in HG-induced HRMECs.


Introduction
Diabetes is a disease in which the disorder of glucose metabolism leads to microvascular lesions in various organs in the body. The main manifestations of the ocular region are the formation of microhemangiomas, neovascularization, bleeding, institutionalization, fibrous tissue hyperplasia and spasm, and finally, retinal traction leading to retinal holes and detachment (Bek 2018). Therefore, diabetic retinopathy (DR) is the most serious blindness ophthalmopathy. Recently, the number of patients with DR has increased dramatically, and the problems caused by DR have attracted more and more attention (Yao et al. 2020). Currently, anti-vascular endothelial growth factor (VEGF) therapy is the mainstream treatment for DR, but it requires repeated intravitreal injections and is only effective for middle-advanced DR, so it has limitations (Singer et al. 2016). In diabetic patients, pathophysiological changes affecting visual function have already occurred in the retina before retinal microvascular damage occurs, but there are few therapeutic measures for these early-stage damages (Dumitrescu et al. 2017;Niestrata-Ortiz et al. 2019). Thence, exploring the molecular biology alteration mechanism of DR and finding potential therapeutic methods have invaluable significance.
As an important part of the ubiquitin-proteasome system (UPS), deubiquitinating enzymes (DUB) regulate the ubiquitination process of target proteins and have a regulatory role in the function and degradation of proteins (Mines et al. 2009). Ubiquitin-specific protease 14 (USP14) is an important deubiquitinating enzyme that regulates protein stability (Kim and Goldberg 2017). USP14 has been implicated in diseases such as signal transduction, neurological diseases, glucose and lipid metabolism, and tumorigenesis, making it a promising drug target (Liu et al. , 2018Lundgren and Odrzywol 2018). According to Fu et al. reported that USP14 expression was significantly up-regulated in retina of diabetic rats and HG-stimulated Müller cells, and the lncRNA OGRU/miR-320/USP14 axis was involved in the progression of DR by regulating inflammation and oxidative stress (Fu et al. 2021). However, little is known about the role of USP14 in angiogenesis and its related mechanisms in the development of DR, and further investigation is needed.
Activating transcription factor 2 (ATF2) belongs to the basic leucine zipper transcription factor family, which binds to DNA in the form of homodimers or heterodimers and regulates the transcriptional activity of target genes (Jiang et al. 2017). Lin et al. confirmed that the p38 MAPK/ATF2 signaling pathway played an important role in regulating cognitive function and apoptosis in diabetes (Lin et al. 2020). Furthermore, Shao et al. found that ATF2 expression was up-regulated in diabetic mouse retina, and miR-451a/ATF2 was involved in retinal pigment epithelial cell proliferation and migration by regulating mitochondrial function (Shao et al. 2019). However, there is a lack of research on the role of ATF2 in promoting angiogenesis in DR. Interestingly, Geng et al. demonstrated that USP14 deubiquitinated and activated ATF2 in prostate cancer, stabilizing its expression (Geng et al. 2019). Therefore, we can not help but speculate that USP14 is involved in DR by regulating ATF2 expression.
Phosphoinositide-3 kinase catalytic subunit δ (PI3Kδ, or PIK3CD) is an important regulator of cell growth (Foukas et al. 2010). Research on PI3Kδ and diabetes-related diseases have long been reported (Colomiere et al. 2009). Moreover, PI3Kδ expression was up-regulated in human retinal microvascular endothelial cells (HRMECs) induced by HG, and promoted cell proliferation, migration , and angiogenesis by regulating the activation of the AKT pathway in vascular endothelial cells, and inhibiting PI3Kδ expression significantly attenuated pathological retinal angiogenesis . According to Jaspar prediction, ATF2 had a binding site for PIK3CD, which has not yet been reported. Therefore, we speculated that ATF2 regulated PIK3CD expression, thus affecting the development of DR.
Therefore, this study aims to confirm that USP14 promoted the proliferation, migration, and tube formation of HG-induced HRMECs by regulating the ATF2/ PIK3CD axis, providing a new perspective for effective treatment of diabetic retinopathy.

DR Cells Model Construction
Human retinal microvascular endothelial cells (HRMECs) were purchased from Procell Life Science Technology Co,.Ltd. (Wuhan, China). HRMECs were cultured in Endothelial medium (ECM, Solarbio, Beijing, China) with 5% fetal bovine serum (FBS, Beyotime, Shanghai, China), 1% endothelial cell growth supplement (Beyotime), and 1% penicillin/streptomycin solution (Beyotime) at 37 °C and 5% CO 2 . HRMECs in the logarithmic growth phase were taken for subsequent experimentations. HRMECs in high glucose (HG) group were cultured in 30 mmol/L glucose medium for 48 h, and the cells in the Normal group were cultured in a 5.5 mmol/L glucose medium. In the Mannitol group, cells were cultured in 5.5 mmol/L glucose + 24.5 mmol/L mannitol medium.

Western Blot
HRMECs were lysed by RIPA lysis solution (Beyotime) to obtain the total protein, and protein concentration was measured by BCA protein quantitative kit (Beyotime). Appropriate protein dosage was disunited by sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) electrophoresis and moved to PVDF membrane. 5% skim milk was sealed for 2 h, and primary antibody diluent of corresponding protein was incubated in 4 °C overnight. Thenceforth, the secondary antibody dilution was nurtured at ambient temperature for 2 h. Thenceforth the bands were nurtured with ECL luminescence solution (Thermo Fisher, MA, USA) and placed in chemiluminescence imaging system for exposure development. The relative expression of each protein was deconstructed by ImageJ software. The following primary antibodies were used: USP14 (ab246010, 1/2000), ATF2 (ab32160, 1/1000), PIK3CD (ab32401, 1/500), and GADPH (ab8245, 1/1000).

Quantitative Real-Time PCR
Total RNA of HRMECs was obtained on the basis of the standard procedure described in the instruction of TRIzol reagent (Invitrogen, MA, USA). RNA concentration was confirmed by ultraviolet spectrophotometer. Subsequently, the first strand cDNA was synthesized by reverse transcription accordant with the guidelines of the reverse transcription kit (Biosystems, Shanghai, China). Real-time PCR was manipulated utilizing PCR SYBR (TaKaRa, Tokyo, Japan) Green method, the amplification efficiency for the primers used for qRT-PCR was 94%. The gene expression was evaluated by 2 −∆∆Ct method. The primers' information is as demonstrated: USP14

Cell Counting Kit-8 (CCK-8) Assay
HRMECs with logarithmic growth phase were cultured into 96-well culture plates with a cell density of 5 × 10 3 cells per well. The 96-well plates were incubated in a 37 °C incubator with 5% CO 2 for 24 h, then 10 µL of CCK-8 solution (Beyotime) was added to each well and cultured for 4 h, and the absorbance value was detected by enzyme labeling instrument (BioTek, VT, USA) at 450 nm wavelength.

Immunoprecipitation (IP)
HRMECs were lysed with RIPA lysis solution (Beyotime). The cleavage products were centrifuged at 4 °C, and the cleavage products were collected. Appropriate amount of cleavage products was directly detected by Western blot to detect the expression of USP14 and ATF2 proteins as input, and the rest of the cleavage products were incubated overnight with anti-USP14 or anti-ATF2 antibodies. After incubation, protein G beads (Santa Cruz) were added to incubate for 4 h, and then the precipitated protein G beads were collected by centrifugation. Proteins on the beads were eluted, and the ATF2 and USP14 proteins were detected by western blot.

Immunofluorescent Staining
HRMECs were immunofluorescent labeled according to the manufacturer's instructions (Botai Biotechnology, Nantong, China). The cells were incubated with blocking buffers for 1 h to inhibit nonspecific binding. Next, the cells were incubated with the primary medium for ATF2 (ab32160, 1/1000) and USP14 (ab246010, 1/2000) antibodies for 1 h, followed by secondary antibody for 1 h, and DAPI for 5 min before observation under a fluorescence microscope.

Transwell Assay
HRMECs (2 × 10 5 cells) in logarithmic growth stage were suspended in 200 μL serumfree medium, and the cell suspension was added to the upper chamber, and the lower chamber with 600 μL medium with FBS. After being cultured for 24 h, the HRMECs were fixed with 4% paraformaldehyde and stained with crystal violet solution. Randomly select 5 visible points under the microscope to take pictures and count. The average number of cells that penetrated into lower chamber was used as the experimental result.

Tube Formation Assay
200 μL of Matrigel matrix glue (BD, NJ, USA) was added to the 24-well plate and put into 37 °C, 5% CO 2 incubator for solidification. The cells cultured for 48 h in each group were prepared into 2 × 10 5 /mL suspension with ECM medium and added to the 24-well plate containing Matrigel matrix glue according to 1 mL/well. The cells were cultured in 37 °C, 5% CO 2 incubator. After 8 h, the capillary structure was photographed. ImageJ software was applied to analyze the capillary-like structures.

Dual-Luciferase Reporter Gene Assay
Jaspar (https:// jaspar. gener eg. net/) was used to speculate the binding of ATF2 to PIK3CD. The sequence of ATF2 binding site of PIK3CD promoter (ACC TGA GGT CAG G, wild type sequence) was mutated to (CGG ATC AAG TCA T, mutant type sequence), and then cloned into the pGL3 luciferase reporter gene (Geneharma, Shanghai, China). The corresponding luciferase reporter gene vector was co-transfected with oe-ATF2 or oe-NC. Luciferase activity was determined according to the procedures of luciferase report Analysis Kit (Promega, WI, USA) after 48 h.

Statistical Analysis
GraphPad prism 7 software (IBM SPSS software, IL, USA) was used to examine the data. The statistic was expressed as mean ± standard deviation (all experiments were repeated for 3 times). The comparison between two groups was analyzed by student's t-test. Furthermore, one-way ANOVA was used to test the discrepancy between multiple groups. P < 0.05 was esteemed statistically noteworthy.

USP14 Knockdown Inhibited Proliferation, Migration, and Tube Formation of HG-Induced HRMECs
In order to explore the role of USP14 in microvascular endothelial cells, HRMECs were treated with high glucose or mannitol. We noticed that compared with the normal group or the mannitol group, the expression of USP14 mRNA and protein 1 3 was significantly up-regulated in the DR cell model (Fig. 1A, B). Then, we knocked down USP14 in HRMECs, and USP14 expression decreased significantly (Fig. 1C). We observed that USP14 expression was up-regulated in HG-induced HRMECs, while USP14 expression in DR cells with USP14 knockdown was down-regulated (Fig. 1D). CCK-8 assay showed that the viability of HG-induced HRMECs was significantly enhanced, but USP14 knockdown in HG-induced HRMECs could effectively inhibit the viability of HRMECs. Additionally, USP14 knockdown also inhibited the viability of non-HG-induced HRMECs (Fig. 1E). Compared with control group, the migration ability of DR cell model was significantly enhanced, and USP14 knockdown significantly inhibited the migration of HRMECs. It was worth mentioning that USP14 knockdown also significantly inhibited the migration ability of normal HRMECs (Fig. 1F). Tube formation assay revealed that compared with the Control group, the tube formation rate in the HG group increased significantly, and this phenomenon was significantly inhibited in HG-induced HRMECs with USP14 knockdown. Moreover, USP14 knockdown significantly inhibited tube formation in normal HRMECs (Fig. 1G). EdU assay showed similar results, indicating that viability of HRMECs with HG treatment was abnormally upregulated, while USP14 knockdown significantly inhibited that of HG-induced HRMECs. Furthermore, it should be noted that USP14 knockdown significantly inhibited the viability of non-HG treated HRMECs (Fig. 1H). These results suggested that USP14 expression was abnormally up-regulated in DR cell model, while USP14 knockdown significantly inhibited the migration, tube formation, and proliferation of HG-induced HRMECs.

USP14 Regulated ATF2 Expression
To study the regulation of USP14 on ATF2 in the DR cell model, ATF2 expression in HG-induced HRMECs was detected by qRT-PCR assay and Western Blot, respectively. These results showed that there was no obvious differentiation in ATF2 expression between the normal group and mannitol group, but both ATF2 mRNA and protein expressions were significantly up-regulated in DR cell model ( Fig. 2A, B). Subsequently, oe-USP14 was transfected in HRMECs, and the results showed that USP14 was successfully overexpressed in HRMECs (Fig. 2C). Consistent with the above results, USP14 expression in HG-induced HRMECs was up-regulated, while USP14 expression in DR cells with oe-USP14 transfection was further up-regulated (Fig. 2D). Then, we knocked down or overexpressed USP14 in DR cell model. The results showed that USP14 knockdown down-regulated the abnormally overexpression of ATF2 protein in HG-induced HRMECs, while USP14 overexpression further up-regulated ATF2 protein expression in DR cell model (Fig. 2E). IP assay revealed that USP14 could be combined with ATF2 (Fig. 2F). Additionally, we further confirmed that USP14 and ATF2 were co-located in the cytoplasm (Fig. 2G). Furthermore, we found that ubiquitin levels were down-regulated and ATF2 expression was up-regulated in HG-induced HRMECs, and USP14 knockdown could effectively up-regulate ubiquitin levels and down-regulate ATF2 expression (Fig. 2H). After HRMECs

Fig. 1 USP14 knockdown inhibited proliferation, migration, and tube formation of HG-induced
HRMECs. DR cell model was constructed using HRMECs treated with 30 mmol/L glucose medium, and sh-USP14 was transfected into HG-induced HRMECs. A-D USP14 expression was ascertained by qRT-PCR assay and western blot. E HRMECs viability was determined by CCK-8 assay. F HRMECs migration ability was ascertained by Transwell assay. G HRMECs angiopoiesis was ascertained by Tube formation assay. H HRMECs viability was determined by EdU assay. Each group performed 3 independent experiments leastwise. *P < 0.05, **P < 0.01, ***P<0.001 was treated with ubiquitin-proteasome inhibitor MG132 (10 μM), the protein level of ATF2 in cells was significantly up-regulated (Fig. 2I). After that, we knocked down USP14 and added protein synthesis inhibitor cycloheximide (CHX, 10 μM) in HRMECs to detect the half-life of ATF2 at different time points. We found that the degradation of ATF2 in cells with USP14 knockdown was accelerated (Fig. 2J). The above results revealed that USP14 regulated ATF2 expression in DR cell model.

ATF2 Knockdown Inhibited Proliferation, Migration, and Tube Formation of HG-Induced HRMECs
In order to explore the role of ATF2 in HRMECs, ATF2 was knocked down in HRMECs. Compared with the control group, ATF2 protein level in HRMECs was down-regulated (Fig. 3A). In HG-induced HRMECs, we noticed that ATF2 expression was up-regulated, which was significantly suppressed after ATF2 knockdown (Fig. 3B). Subsequently, viability of HG-induced HRMECs was significantly increased, and ATF2 knockdown could effectively inhibit viability of DR cell Fig. 3 ATF2 knockdown inhibited proliferation, migration, and tube formation of HG-induced HRMECs. Sh-ATF2 was transfected into HG-induced HRMECs. A, B ATF2 expression was detected by western blot. C The HRMECs viability was determined by CCK-8 assay. D HRMECs migration ability was ascertained by Transwell assay. E The HRMECs angiopoiesis was ascertained by Tube formation assay. Each group performed 3 independent experiments leastwise. *P < 0.05, **P < 0.01, ***P < 0.001 model. Additionally, ATF2 knockdown also inhibited the viability of non-HGinduced HRMECs (Fig. 3C). Additionally, although the migration ability of DR cell model was significantly enhanced compared with the control group, ATF2 knockdown significantly inhibited the migration ability of HRMECs. It was worth mentioning that ATF2 knockdown also significantly inhibited the migration ability of normal HRMECs (Fig. 3D). Tube formation assay revealed that tube formation was increased in HG group, and ATF2 knockdown could significantly inhibit this trend. Moreover, ATF2 knockdown significantly inhibited tube formation in normal HRMECs (Fig. 3E). These results indicated that ATF2 knockdown inhibited the migration, tube formation, and proliferation of HG-induced HRMECs.

ATF2 Promoted PIK3CD Expression
In order to confirm that ATF2 promoted PIK3CD expression and affected the development of DR, we first analyzed PIK3CD expression in the DR cell model by qRT-PCR assay and Western blot analysis. The expression levels of PIK3CD mRNA and protein were significantly up-regulated in HG-induced HRMECs (Fig. 4A, B). After HRMECs were transfected with oe-ATF2 plasmid, ATF2 protein level was significantly up-regulated (Fig. 4C). Similarly, ATF2 protein was significantly upregulated in DR cell model, and ATF2 overexpression could be further enhanced the Fig. 4 ATF2 promoted PIK3CD expression. Sh-ATF2 or oe-ATF2 was transfected into HG-induced HRMECs. A, B PIK3CD expression was ascertained by qRT-PCR assay and western blot. C, D ATF2 expression was detected by western blot. E, F PIK3CD expression was ascertained by qRT-PCR assay and western blot. G, H Dual-luciferase reporter assay and ChIP assay were manipulated to confirm the relationship of ATF2 and PIK3CD. Each group performed 3 independent experiments leastwise. *P < 0.05, **P < 0.01, ***P < 0.001 trend (Fig. 4D). The expression levels of PIK3CD mRNA and protein were up-regulated in DR cell model. The expression levels of PIK3CD mRNA and protein were down-regulated in HG-induced HRMECs with ATF2 knockdown, while the expressions of mRNA and protein were further up-regulated in HG-induced HRMECs with ATF2 overexpression (Fig. 4E, F). Additionally, ATF2 knockdown in normal HRMECs also had the same effect, which significantly inhibited the expression of PIK3CD mRNA and protein ( Fig. S1A and S1B). After that, dual-luciferase reporter gene assay revealed that ATF2 overexpression could significantly enhance the luciferase activity of PIK3CD-WT, but had no obvious influence on PIK3CD-MUT (Fig. 4G). Furthermore, ChIP assay showed that anti-ATF2 could enrich more PIK3CD than IgG group (Fig. 4H). These results indicated that ATF2 promoted PIK3CD expression.

PIK3CD Overexpression Reversed the Function of USP14 Knockdown on Proliferation, Migration, and Tube Formation of HG-Induced HRMECs
In order to explore the role of PIK3CD in HRMECs, we transfected HRMECs with oe-PIK3CD plasmid and found that PIK3CD expression was significantly up-regulated (Fig. 5A). Similarly, in the DR cell model, we noticed that PIK3CD expression was up-regulated, and the trend was more pronounced with PIK3CD overexpression (Fig. 5B). Moreover, we noted that USP14 knockdown inhibited PIK3CD protein expression in HG-induced HRMECs, while PIK3CD overexpression significantly attenuated the effect of USP14 knockdown (Fig. 5C). As previously mentioned, viability, migration, and tube formation were enhanced in HG-induced HRMECs, whereas USP14 knockdown suppressed proliferation, migration, and tube formation in the DR cell model. Furthermore, PIK3CD overexpression enhanced viability, migration, and tube formation of DR cell model with USP14 knockdown (Fig. 5D-F). These results revealed that PIK3CD overexpression attenuated the inhibitory effect of USP14 knockdown on the proliferation, migration, and tube formation of HG-induced HRMECs.

Discussion
DR is a very complex pathological process. Retinal neovascularization is the most important process in the development of this disease (Arrigo et al. 2022;Du et al. 2018). Therefore, to explore the mechanism of diabetic retinal neovascularization and how to safely and inhibit the neovascularization has become an important topic in the research and clinical work of diabetic retinopathy. The UPS family plays an important role in maintaining protein homeostasis and has been associated with diabetes-related complications (Lu et al. 2021;Mao et al. 2021). In a study of type II diabetes, Forand et al. found that inhibiting the dissociation of USP7/Insulinreceptorsubstrate1 (IRS1) and preventing the ubiquitination of IRS1 had a protective effect on diabetic mice (Forand et al. 2016). Additionally, USP9X was observed to play a protective role in diabetic 1 3 nephropathy by stabilizing de-ubiquitin connexin43 (Cx43) expression . High glucose induced USP22 overexpression in podocytes, and silencing USP22 attenuated the cytotoxicity of podocytes (Shi et al. 2016). Moreover, USP14 overexpression was involved in the development of diabetic retinopathy by inducing deubiquitination of transforming growth factor-β-1 (TβR1) and HG promoted USP14 expression by lncRNA-OGRU/miR-320 axis (Fu et al. 2021). In this study, USP14 and its downstream protein ATF2 were abnormally overexpressed in HG-induced HRMECs, and USP14 directly interacted with ATF2 to Fig. 5 PIK3CD overexpression reversed the function of USP14 knockdown on proliferation, migration, and tube formation of HG-induced HRMECs. Sh-USP14 and/or oe-PIK3CD were transfected into HGinduced HRMECs. A, B PIK3CD expression was ascertained by western blot. C The HRMECs viability was determined by CCK-8 assay. D HRMECs migration ability was ascertained by Transwell assay. E The HRMECs angiopoiesis was ascertained by Tube formation assay. F The expression of PIK3CD and USP14 was ascertained by western blot. Each group performed 3 independent experiments leastwise. *P < 0.05, **P < 0.01, ***P < 0.001. prevent the degradation of ATF2. USP14 knockdown inhibited the proliferation, migration, and tube formation of HRMECs upon high glucose.
ATF2 is a member of the ATF/CREB bZIP transcription factor family. In a study of ovarian cancer, Yi et al. found that ATF2 might be involved in the exocrine function of angiogenesis (Yi et al. 2015). Furthermore, Zhao et al. found that 15 (S)-HETE activated ATF2 in HRMECs and led to angiogenesis and differentiation (Zhao et al. 2009). Recently, Wang et al. claimed that ATF2 played an important role in VEGF-induced retinal neovascularization, and ATF2 knockdown alleviated VEGF-induced angiogenesis and attenuated HRECs migration ability (Wang et al. 2021). In this study, a similar phenomenon was observed in the DR cell model, where ATF2 knockdown inhibited the proliferation, migration, and tube formation of HRMECs. Additionally, it is worth noting that ATF2 overexpression up-regulated PIK3CD expression in DR cell model. Our study shows that ATF2 promoted PIK3CD expression by transcription regulation in DR cell model.
In the occurrence and development of diabetes, continuous hyperglycemia promotes PI3K pathway activation, and ultimately accelerates the development of DR (Tang et al. 2022). PI3K/Akt signaling pathway is a key signaling pathway that induces endothelial cell proliferation and angiogenesis (Di et al. 2015). In an in vitro study of diabetic retinopathy, Wang et al. found that up-regulation of miR-199a-3p inhibited PI3K/Akt signaling pathway, thereby improving HG-stimulated HRMECs angiogenesis ). Qiu et al. inhibited PI3K/Akt pathway by recombinant human maspin, and attenuated proliferation, oxidative stress, and angiogenesis in HRMECs upon high glucose (Qiu et al. 2018). PI3Kδ, an important member of the PI3K family, is encoded and synthesized by PIK3CD gene (Fransson et al. 2012). Wu et al. confirmed that PI3Kδ expression was upregulated in DR cell model, which activated Akt signal pathway and aggravated the progress of DR (Arrigo et al. 2022). Our results confirmed this finding which PIK3CD expression was up-regulated in DR cell model and intervening PIK3CD expression affected HG-induced cell behavior. Furthermore, PIK3CD overexpression attenuated the inhibitory effect of USP14 knockdown on the proliferation, migration, and tube formation of HG-induced HRMECs. These results indicated that USP14 participated in the regulation of ubiquitination level of ATF2 and ATF2 further promoted the expression of PIK3CD by transcription regulation. Direct regulation of PIK3CD would reverse the effect of USP14.
In brief, these evidences suggested that USP14 regulated the ATF2/PIK3CD axis, promoting HRMECs proliferation, migration, and angiogenesis under high glucose conditions. Although the present study focused on the molecular mechanism, it provided a new idea and experimental basis for the clinical treatment of diabetic retinopathy.