Circ-ADAM9 Promotes High Glucose-Induced Retinal Pigment Epithelial Cell Injury in DR via Regulating miR-338-3p/CARM1 Axis

Background Circular RNAs (circRNAs) have been reported to be involved in the regulation of retinal pigment epithelial (RPE) cell injury and are closely related to the development of diabetic retinopathy (DR). More research is needed to confirm the role and mechanism of circ-ADAM9 in DR progression. Methods High glucose (HG)-induced RPE cells (ARPE-19) were used to mimic the hyperglycemia condition. The expression of circ-ADAM9, microRNA (miR)-338-3p, and coactivator-associated arginine methyltransferase 1 (CARM1) was measured using quantitative real-time PCR. Cell proliferation and apoptosis were determined using MTT assay, EdU assay, and flow cytometry. The protein expression of apoptosis markers and CARM1 was examined by the western blot analysis. Also, MDA level and SOD activity were determined to assess cell oxidative stress. In addition, the interaction between miR-338-3p and circ-ADAM9 or CARM1 was confirmed by dual-luciferase reporter assay and RIP assay. Results The expression of circ-ADAM9 was upregulated in DR patients and HG-induced ARPE-19 cells. Silenced circ-ADAM9 could promote proliferation and inhibit inflammation, apoptosis, and oxidative stress in HG-induced ARPE9 cells. In terms of mechanism, circ-ADAM9 could sponge miR-338-3p to upregulate CARM1. The inhibitory effect of circ-ADAM9 knockdown on HG-induced ARPE9 cell injury could be reversed by an miR-338-3p inhibitor. As a target of miR-338-3p, CARM1 knockdown could alleviate HG-induced ARPE9 cells' injury, and its overexpression also could reverse the negatively regulation of miR-338-3p on HG-induced ARPE9 cell injury. Conclusion Circ-ADAM9 contributed to HG-induced ARPE9 cell injury by regulating miR-338-3p/CARM1 axis, which provided effective targets for DR treatment.


Introduction
Diabetic retinopathy (DR) is the most important manifestation of diabetic microangiopathy, which is one of the serious and syndromic manifestations of diabetes [1,2]. e main reason is that the retinal tissues and vascular microcirculation of DR patients have changed, resulting in the damage of eye nutrition and visual function [3,4]. e dysfunction of the blood-retinal barrier (BRB) induced by hyperglycemia is considered to be one of the earliest changes in DR [5,6]. Retinal pigment epithelial (RPE) cells are a key part of the external retinal barrier, and their dysfunction is considered to be a vital reason for the dysfunction of BRB and the progression of DR [7][8][9]. It is important to elucidate the molecular mechanisms of RPE cell injury induced by hyperglycemia for revealing the pathogenesis of DR.
As a noncoding RNA with circular structure, the important role of circular RNA (circRNA) in human diseases has been confirmed by more and more studies [10,11]. CircRNA has a wealth of functional microRNA (miRNA) binding sites to act as miRNA sponge, thereby indirectly regulating target gene expression [12,13]. A lot of evidence shows that circRNA regulates the progression of diseases by interacting with disease-related miRNAs [14]. Circ_RUSC2 had been discovered to be a potential target for cardiovascular diseases, which could promote the proliferation and migration of vascular smooth muscle cells through the miR-661/SYK axis [15]. In DR-related studies, circDNMT3B was found to alleviate vascular dysfunction in diabetic retinas by upregulating BAMB1 via sponging miR-20b-5p [16]. Also, circ_0041795 had been discovered to promote high glucose (HG)-induced RPE cell apoptosis and inflammation to aggravate DR progression by the miR-646/VEGFC axis [17]. Moreover, studies had shown that circ-ITCH could inhibit the expression of neovascularization-related markers and the secretion of inflammation factors in RPE cells to inhibit DR progression through targeting miR-22 [18]. erefore, circRNA is a key regulator for DR progression.
Circ_0084043 is located at chr8: 38883321-38959449 and is derived from the ADAM9 gene, also called circ-ADAM9. A recent study showed that circ-ADAM9 promoted HG-induced apoptosis, inflammation, and oxidative stress in RPE cells, suggesting that circ-ADAM9 might contribute to DR progression [19]. However, the current evidence is still limited, and more evidence is needed to further confirm the potential of circ-ADAM9 as a therapeutic target for DR. Our study aimed to reveal the role and new molecular mechanism of circ-ADAM9 in HG-induced RPE cell injury, hoping to provide a reliable theoretical basis for circ-ADAM9 to become a target of DR treatment.

Serum Samples.
A total of 48 participants were recruited from Tianjin Eye Hospital, including 24 DR patients and 24 age-matched normal control volunteers (undergoing physical examination). e characteristics of DR patients and normal control volunteers are shown in Table 1. After centrifuged the blood samples, the serums were collected and stored at −80°C. Each participant signed an informed consent. Our study was approved by the Ethics Committee of Tianjin Eye Hospital. -19) were obtained from ATCC (Manassas, VA, USA) and were cultured in the DMEM/F12 medium (Sigma-Aldrich, St. Louis, MO, USA) containing 10% FBS (Sigma-Aldrich) and penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA) at 37°C with 5% CO 2 . To mimic HG environment, ARPE-19 cells were cultured in 30 mM D-glucose (Sigma-Aldrich) culture medium for 24 h. e medium containing 5.5 mM of D-glucose was considered as a normal glucose (NG) condition. For cell transfection, it could be performed when the cell reached 50% confluences using Lipofectamine 3000 (Invitrogen). All oligonucleotides and vectors were synthesized from GenePharma (Shanghai, China), including circ-ADAM9 small interference RNA (si-circ-ADAM9) or pCD5 overexpression vector, miR-338-3p mimic or inhibitor (anti-miR-338-3p), coactivator-associated arginine methyltransferase 1 (CARM1) siRNA (si-CARM1) or pcDNA overexpression vector, and their negative controls. After transfection for 24 h, cells were exposed to HG or NG conditions for 24 h.

Quantitative Real-Time PCR (qRT-PCR).
e TRIzol reagent (TaKaRa, Dalian, China) was used for the isolation of total RNA. e cDNA was synthesized by a PrimeScript RT reagent kit (TaKaRa). e PCR reaction was conducted with SYBR Green (Invitrogen) on the PCR System. e fold change was calculated using 2 −ΔΔCT method with GAPDH or U6 as internal control. Primers are listed in Table 2. For a subcellular localization analysis, a PARIS kit was used to extract the cytoplasm and nuclear RNA from ARPE-19 cells, and the RNA was used for detecting circ-ADAM9, U6, and GAPDH expression. For RNase R assay, the extracted RNA was treated with RNase R and then the RNA was used for measuring circ-ADAM9 and linear RNA ADAM9 expression.

ELISA.
e concentrations of IL-6 and TNF-α in the culture medium were analyzed using commercially Human IL-6 ELISA Kit and TNF-α ELISA Kit (all from Sigma-Aldrich), respectively.  en, membranes were incubated with secondary antibody (1 : 20,000, ab205718, Abcam), and the blot was visualized using HRP substrate ECL luminescent solution (Millipore, Billerica, MA, USA). Protein expression was analyzed using ImageJ software with GAPDH as internal control.

Determination of MDA Level and SOD Activity.
According to the manufacturer's instructions, MDA level and SOD activity were assessed using corresponding commercial kits (Nanjing Jiancheng Technology, Nanjing, China).

Dual-Luciferase Reporter Assay.
e wild-type (WT) and mutant-type (MUT) sequences of circ-ADAM9 or CARM1 3'UTR with binding sites and mutant sites for miR-338-3p were inserted into the pmirGLO luciferase vector. 293T cells were transfected with the WT/MUT vectors and miR-338-3p mimic or miR-NC using Lipofectamine 3000. Luciferase activity was analyzed by a Dual-Luciferase Reporter Gene Assay Kit (Beyotime).
2.11. RIP Assay. ARPE-19 cells were lysed in RIP lysis solution and cell lysates were incubated with immunoprecipitation buffer containing magnetic beads conjugated with anti-Ago2 or anti-IgG according to the instructions of RIP Kit (Millipore). e immunoprecipitated RNAs were extracted to detect the RNA level using qRT-PCR.

Statistical Analysis.
All data were obtained from 3 independent experiments and expressed as mean ± SD. GraphPad Prism 7.0 was used for the data analysis. Linear correlation was determined by the Pearson's correlation coefficient analysis. e comparison among groups was performed using Student's t-test or one-way ANOVA. P < 0.05 was considered as a significant difference.

e Level of Circ-ADAM9 Was Increased in DR Patients
and HG-Induced ARPE-19 Cells. In the serum of DR patients and normal control volunteers, we discovered that circ-ADAM9 was upregulated in DR patients (Figure 1(a)). Moreover, circ-ADAM9 also was highly expressed in HGinduced ARPE-19 cells compared to the cells under NG condition (Figure 1(b)). Subcellular localization analysis showed that circ-ADAM9 was mainly distributed in the cytoplasm of cells (Figure 1(c)). RNase R assay results revealed that linear RNA ADAM9 mRNA expression could be decreased after RNase R treatment, while circ-ADAM9 could resist the digestion of RNase R (Figure 1(d)). ese results indicated that circ-ADAM9 was indeed a circRNA, which might be involved in the regulation of DR progression.

Knockdown of Circ-ADAM9 Suppressed Inflammation, Apoptosis, and Oxidative Stress in HG-
en, the role of circ-ADAM9 was investigated in HG-induced ARPE-19 cell injury. After ARPE-19 cells were transfected with si-circ-ADAM9 and treated with HG, we found that the elevated circ-ADAM9 expression induced by HG could be inhibited by si-circ-ADAM9 (Figure 2(a)). HG could promote the secretions of inflammatory factors IL-6 and TNF-α and inhibit the viability and the EdU-positive cell rate in ARPE-19 cells, while these effects could be abolished by knockdown of circ-ADAM9 (Figure 2(b)-2(d)). Silenced circ-ADAM9 also inhibited cell apoptosis rate and apoptosis markers Bax and cleaved-caspase-3 protein expression in HG-induced ARPE-19 cells (Figures 2(e) and 2(f )). Furthermore, HG treatment enhanced MDA level and suppressed SOD activity in ARPE-19 cells, and knockdown of circ-ADAM9 also could eliminate these effects (Figures 2(g) and 2(h)).

Silenced CARM1 Alleviated HG-Induced ARPE-19 Cell
Injury. To explore the function of CARM1 in DR progression, we silenced CARM1 in HG-induced ARPE-19 cells using si-CARM1 (Figure 6(a)). rough measuring cell inflammation and proliferation, we found that CARM1 knockdown could inhibit IL-6 and TNF-α concentrations, while promoting the viability and the EdU-positive cell rate in HG-induced ARPE-19 cells (Figures 6(b)-6(d)). Also, silencing of CARM1 inhibited cell apoptosis rate, apoptosis marker expression, MDA level, and increased SOD activity in HG-induced ARPE-19 cells (Figures 6(e)-6(h)). ese data illuminated that CARM1 might play an active role in DR progression.

CARM1 Reversed the Inhibition Effect of miR-338-3p on HG-Induced ARPE-19 Cell Injury.
To further verify that miR-338-3p targeted CARM1 to regulate DR progression, e pcDNA CARM1 overexpression vector was confirmed to significantly increase CARM1 expression in ARPE-19 cells (Figure 7(a)). In HG-induced ARPE-19 cells cotransfected with miR-338-3p mimic and pcDNA CARM1 overexpression vector, we found that the decreasing effect of miR-338-3p mimic on miR-338-3p expression could be abolished by pcDNA CARM1 overexpression vector (Figure 7(b)). e inhibition effect of miR-338-3p on the concentrations of IL-6 and TNFα and the promotion effect on the viability and the EdUpositive cell rate could be eliminated by overexpressing CARM1 in HG-induced ARPE-19 cells (Figures 7(c)-7(e)).
rough analyzing cell apoptosis rate, apoptosis markers' expression, and oxidative stress markers' levels, we discovered that overexpression of CARM1 also reversed the suppressive effect of miR-338-3p on the apoptosis and oxidative stress in HG-induced ARPE-19 cells (Figures 7(f )-7(j)). All data revealed that miR-338-3p mediated DR progression by targeting CARM1.

Discussion
Many studies have shown that circRNA plays a functional regulatory role in DR progression and is expected to become a biomarker of DR [20,21]. Here, we investigated circ-ADAM9 roles in HG-induced DR models in vitro. In previous studies, circ-ADAM9 was discovered to be involved in regulating melanoma malignant progression. Chen et al. reported that circ-ADAM9 could regulate the miR-429/ TRIB2 axis to facilitate melanoma cells' proliferation and metastasis [22]. Moreover, circ-ADAM9 was found to enhance the proliferation and glycolysis of melanoma cells by upregulating KLF3 through sponging miR-31 [23]. Consistent with the results of the previous study [19], our research revealed that circ-ADAM9 played an active role in DR progression. In this, we found that circ-ADAM9 knockdown inhibited the apoptosis, inflammation, and oxidative stress in HG-induced RPE cells and significantly improved cell proliferation. Our study once again confirmed the role of circ-ADAM9 in DR progression, providing new evidence for circ-ADAM9 as a potential target of DR treatment.
Li et al. showed that circ-ADAM9 mediated DR progression by sponging miR-140-3p [19]. In order to identify a novel mechanism by which circ-ADAM9 regulated DR progression, we performed a bioinformatics analysis and suggested that miR-338-3p might be a new target of circ-ADAM9. miR-338-3p had been found to mediate malignant progression as a tumor suppressor in cancers, such as colorectal cancer [24], prostate cancer [25], and renal cell carcinoma [26]. Studies had shown that miR-338-3p was a significantly low expressed miRNA in the serum of DR patients, and it might be used as a biomarker for the early risk prediction of DR [27]. Not only that the study of Wu et al. revealed that miR-338 in the retina of diabetic rats was significantly downregulated, and its abnormal expression was closely related to DR development [28]. Here, we demonstrated that miR-338-3p was underexpressed in DR patients and HG-induced ARPE-19 cells and revealed that overexpressed miR-338-  3p had an inhibitory effect on HG-induced RPE cell injury. Further analysis indicated that miR-338-3p inhibitor reversed the negatively regulation of si-circ-ADAM9 on HG-induced RPE cell injury, which confirmed that circ-ADAM9 indeed sponged miR-338-3p to promote DR progression. CARM1, a member of the protein arginine methyltransferase family, is a coactivator of many tumor-associated transcription factors and is abnormally expressed in many cancers [29,30]. A study had shown that CARM1 expression was significantly increased in type 2 diabetes, which might play a vital role in diabetes-related diseases [31]. Under the condition of HG, CARM1 was discovered to be upregulated in RPE cells and promote cell apoptosis [32]. Moreover, Guo et al. suggested that miR-542-5p could target CARM1 to inhibit HG-induced RPE cell apoptosis [33]. In this, the high expression of CARM1 also was found in DR patients and HG-induced RPE cells. Silencing CARM1 could effectively relieve HG-induced RPE cell injury, which was consistent with the results of previous studies [32,33]. In addition, we confirmed that CARM1 was targeted by miR-338-3p, and it could reverse the inhibitory effect of miR-338-3p on HG-induced RPE cell injury. Furthermore, the positive regulation of circ-ADAM9 on CARM1 expression also confirmed the presence of circ-ADAM9/miR-338-3p/CARM1 axis, which ELISA was used to analyze the concentrations of IL-6 and TNF-α. Cell proliferation and apoptosis were determined using MTT assay (c), EdU assay (d), and flow cytometry (e). (f ) e WB analysis was performed to examine the protein expression of Bax and cleaved-caspase-3. (g, h) Cell oxidative stress was analyzed by detecting MDA level and SOD activity. * * P < 0.01, * * * P < 0.001, and * * * * P < 0.0001. improved the molecular mechanism by which circ-ADAM9 regulated DR progression. According to our research results, we pointed out that the increase in CARM1 expression was partly due to the targeting of miR-338-3p by circ-ADAM9, which removed the inhibition of CARM1 expression by miR-338-3p. Of course, because DR development involves complex pathological mechanisms and a regulatory network, the expression of CARM1 is bound to be affected by many mechanisms. Our research only reveals one of the potential molecular mechanisms that mediate the progression of DR.
In conclusion, our study showed that circ-ADAM9 promoted HG-induced inflammation, apoptosis, and oxidative stress in RPE cells through regulating miR-338-3p/ CARM1 axis. Our study showed that circ-ADAM9 might be a target of DR treatment, which provided a new way of thinking for DR treatment strategies.

Data Availability
No data were used to support this study.

Conflicts of Interest
e authors declare that they have no conflicts of interest.