Circular Noncoding RNA HIPK3 Mediates Retinal Vascular Dysfunction in Diabetes Mellitus
Background: The vascular complications of diabetes mellitus are the major causes of morbidity and mortality among people with diabetes. Circular RNAs are a class of endogenous noncoding RNAs that regulate gene expression in eukaryotes. In this study, we investigated the role of circular RNA in retinal vascular dysfunction induced by diabetes mellitus.
Methods: Quantitative polymerase chain reactions, Sanger sequencing, and Northern blots were conducted to detect circular HIPK3 (circHIPK3) expression pattern on diabetes mellitus–related stresses. MTT (3-[4,5-dimethythiazol-2-yl]-2,5-diphenyl tetrazolium bromide) assays, EdU (5-ethynyl-2′-deoxyuridine) incorporation assays, Transwell migration assays, and Matrigel assays were conducted to detect the role of circHIPK3 in retinal endothelial cell function in vitro. Retinal trypsin digestion, vascular permeability assays, and ELISA assays were conducted to detect the role of circHIPK3 in retinal vascular dysfunction in vivo. Bioinformatics analysis, luciferase activity assays, RNA pull-down assays, and in vitro studies were conducted to reveal the mechanism of circHIPK3-mediated retinal vascular dysfunction.
Results: circHIPK3 expression was significantly upregulated in diabetic retinas and retinal endothelial cells following stressors related to diabetes mellitus. circHIPK3 silencing or overexpressing circHIPK3 changed retinal endothelial cell viability, proliferation, migration, and tube formation in vitro. circHIPK3 silencing in vivo alleviated retinal vascular dysfunction, as shown by decreased retinal acellular capillaries, vascular leakage, and inflammation. circHIPK3 acted as an endogenous miR-30a-3p sponge to sequester and inhibit miR-30a-3p activity, which led to increased vascular endothelial growth factor-C, FZD4, and WNT2 expression. Ectopic expression of miR-30a-3p mimicked the effect of circHIPK3 silencing on vascular endothelial phenotypes in vivo and in vitro.
Conclusions: The circular RNA circHIPK3 plays a role in diabetic retinopathy by blocking miR-30a function, leading to increased endothelial proliferation and vascular dysfunction. These data suggest that circular RNA is a potential target to control diabetic proliferative retinopathy.
What Is New?
circHIPK3 expression is significantly upregulated following stressors related to diabetes mellitus.
circHIPK3 maintains retinal endothelial cell function in vitro.
cicHIPK3 regulates diabetes mellitus–related retinal vascular dysfunction in vivo.
circHIPK3 acts as an endogenous miR-30a-3p sponge to sequester and inhibit miR-30a-3p activity.
What Are the Clinical Implications?
The easily accessible vasculature of the eye provides a window to investigate the mechanism of deep vasculature dysfunction.
circHIPK3 is a novel circular RNA that positively affects endothelial angiogenic function.
Altering circHIPK3 expression independently or in concert with current interventions (such as vascular endothelial growth factor and platelet-derived growth factor) would prevent and reduce vascular complications.
Diabetic vascular complications are the leading causes of end-stage renal failure, blindness, several neuropathies, and atherosclerosis. They account for the disabilities and high mortality in patients with diabetes mellitus.1,2 Retinal vessels are the early and prevalent targets of diabetic damage. Diabetes mellitus–induced retinal vascular dysfunction has become a major cause of blindness, which is characterized by vascular leakage, inflammation, and angiogenesis.3 Moreover, diabetic microvascular dysfunction in the retina, glomerulus, and vasa nervorum has similar pathological features.4 In addition, the vasculature of the eye and the heart shares several common characteristics.5 Retinal vessels can be observed and tested by noninvasive methods. Thus, clarifying the underlying mechanism of retinal vascular dysfunction could inform study of the development of retinopathy and other diabetes mellitus–related vascular complications.
Abnormal vascular endothelial growth factor expression is recognized as a leading cause of diabetes mellitus–related retinal vascular dysfunction.6 Other signaling pathways have also been proposed to underlie diabetic retinal vascular dysfunction, including advanced glycation end product formation, protein kinase C activation, increased proinflammatory signaling, and abnormal phosphatidylinositol-3 kinase/Akt signaling.7,8 These factors may be affected mutually. However, the exact mechanism of retinal vascular dysfunction is still not fully understood.
Circular RNAs (circRNAs) are a novel class of noncoding RNAs, characterized by covalently closed loop structures with neither 5′ to 3′ polarity nor a polyadenylated tail. They are expressed in a tissue-specific and developmental stage–specific manner.9 circRNAs regulate gene expression by acting as microRNA (miRNA) sponges, RNA-binding protein sequestering agents, or nuclear transcriptional regulators.10,11 Several lines of evidence indicate that circRNAs are aberrantly expressed in several vascular diseases, neurological disorders, and cancers.12–14 However, the role of circRNAs in diabetes mellitus–induced vascular dysfunction is still unknown.
In this study, we characterized the expression and regulation of circHIPK3 in retinal endothelial cells and diabetes mellitus–induced retinal vascular dysfunction. The results show that circHIPK3 expression is significantly upregulated on high glucose stress in vivo and in vitro. circHIPK3 regulates retinal endothelial cell function and vascular dysfunction by acting as miRNA sponges.
Animals were housed in a specific pathogen-free facility and maintained according to the guidelines of the Care and Use of Laboratory Animals (published by the National Institutes of Health, NIH publication no. 86-23, revised 1996). They were also handled according to the guidelines of (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research. The experiments were approved by the Animal Care and Use Committee of Eye and ENT Hospital and Nanjing Medical University. The surgical specimens were handled according to the Declaration of Helsinki. All patients gave informed consent before inclusion.
Streptozotocin-Induced Diabetic Mice
C57BL/6 mice (8 wk old, male) were fasted for 6 hours before streptozotocin (Sigma) injection. They received an intraperitoneal injection of streptozotocin (50 mg/kg) or vehicle (citrate buffer control) for 5 consecutive days. The fasting blood glucose was determined by using a glucometer (Precision PC; Medic) at 7 days after the last streptozotocin injection. Mice with glucose levels >15 mmol/L were considered hyperglycemic (diabetic).
At the beginning of diabetes induction, C57BL/6 mice (8 wk old, male) were anesthetized by intraperitoneal injection of ketamine (80 mg/kg) and xylazine (4 mg/kg). Approximately 1.5 μL (1×1012 μg/mL) of adeno-associated virus containing circHIPK3-small hairpin RNA (shRNA) or scrambled-shRNA was delivered into the vitreous using a 33-gauge needle. There was a 1- to 2-week delay in the conversion of recombinant adeno-associated virus DNA to a transcriptionally active double-stranded form. Thus, adeno-associated virus vectors were injected 2 weeks before diabetes induction.15 To maximize virus delivery, these mice were administered an intravitreal injection once a month.
Retinal Trypsin Digestion Assay
The eyes were enucleated and fixed in 4% paraformaldehyde for 24 hours. They were then equatorially bisected, and the retinas were removed. The retinas were incubated with 3% trypsin at 37°C for 3 hours, and then gently shaken to free vessel network, washed, and mounted on the glass slides to dry. Retinal vasculature was then stained with periodic acid-Schiff and hematoxylin.
Evans Blue Dye Leak
Diabetic mice and age-matched nondiabetic controls were anesthetized with ketamine (80 mg/kg) and xylazine (4 mg/kg). The right jugular vein and right iliac artery were cannulated and then filled with heparinized saline. Evans blue (45 mg/kg) was injected through the jugular vein over 10 seconds. Two hours later, ≈0.2 mL blood was obtained from the anesthetized mice. The animals were perfused via the left ventricle with PBS followed by 1% paraformaldehyde. The cornea, lens, and vitreous humor were removed. The remaining retina and sclera were fixed in 4% paraformaldehyde in PBS for 30 minutes at room temperature. The retina was treated with formamide (Sigma-Aldrich) overnight at 78°C, and then centrifuged at 12 000g for 15 minutes. The supernatant was detected spectrophotometrically at 620 nm and 740 nm and compared with terminal plasma samples collected from the same animals without formamide treatment. The concentration of dye in the plasma was calculated from a standard curve of Evans blue in formamide.
Data were tested for the normality by using the D’Agostino-Pearson omnibus normality test and homogeneity of variances by using the Levene test. Continuous data were expressed as mean±SEM. For normally distributed data with equal variance, the difference was evaluated by 2-tailed Student t test (2-group comparisons) or 1-way, 2-way, or repeated-measures ANOVA followed by the post hoc Bonferroni test (multigroup comparisons) as appropriate. For nonnormally distributed data or data with unequal variances, the difference was evaluated by a nonparametric Mann-Whitney U test (2-group comparisons) or the Kruskal-Wallis test followed by the post hoc Bonferroni test (multigroup comparisons). Categorical data were expressed as percentages and compared with the Fisher exact test. A value of P<0.05 was considered significant. All statistical analysis was performed using the SPSS software, version 13.0.
circHIPK3 Expression Pattern in Endothelial Cells
Many endothelial circRNAs have been identified in human umbilical vein endothelial cells. Of them, circHIPK3 was found to be abundantly expressed in human umbilical vein endothelial cells.16 circBase retrieval revealed that the HIPK3 host gene might produce 20 circRNAs in the human genome and 3 circRNAs in the mouse genome. The expression of these circRNA candidates was detected in human and mouse retinal endothelial cells. Four and 2 circRNAs from the HIPK3 host gene were identified in human and mouse genomes, respectively (Figure IA in the online-only Data Supplement). circRNAs are expected to be resistant to exonuclease ribonuclease R, an exonuclease that degrades linear RNA molecules.10 Two circRNAs (hsa_circ_0000284 and hsa_circ_0000285) and 1 circRNA (mmu_circ_0001052) from the HIPK3 host gene were identified in endothelial cells followed by ribonuclease R treatment (Figure IA in the online-only Data Supplement). Circular HIPK3 isoform located at chr11:33307958 to 33309057 in the human genome (hsa_circ_0000284) and chr2:104310905 to 104312004 in the mouse genome (mmu_circ_0001052) were highly conserved (Figure II in the online-only Data Supplement). We focused on this circRNA isoform for detailed functional characterization and named it circHIPK3.
Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) assays revealed that circHIPK3 was expressed in several mouse tissues, including brain, retina, heart, liver, artery, and vein (Figure 1A). circHIPK3 expression was also detected in human retinal vascular endothelial cells (HRVECs), and other endothelial cells, as well, including EA.hy.926, human coronary artery endothelial cells, and human umbilical vein endothelial cells (Figure 1B). The amplified product of circHIPK3 was sent for sequencing. The sequencing result was completely in accordance with circHIPK3 the sequence as shown in circBase (Figure 1C). We also verified circHIPK3 expression in several endothelial cells by Northern blots (Figure IIIA in the online-only Data Supplement).
circHIPK3 stability in HRVECs was examined after actinomycin D (an inhibitor of transcription) treatment. Quantitative reverse-transcription polymerase chain reaction assays revealed that circHIPK3 was highly stable with a half life >24 hours. By contrast, the linear transcript, HIPK3 mRNA, was easily degraded with a half life of <5 hours (Figure 1D). circHIPK3 was resistant to ribonuclease R digestion, whereas linear HIPK3 mRNA was easily degraded (Figure 1E). Quantitative reverse-transcription polymerase chain reaction analysis of nuclear and cytoplasmic RNA and fluorescence in situ hybridization assays showed that circHIPK3 was mainly expressed in the cytoplasm of HRVECs (Figure 1F and 1G).
We next determined whether circHIPK3 expression was altered under diabetic conditions in vivo. Quantitative reverse-transcription polymerase chain reaction assays showed that retinal circHIPK3 expression in diabetic mice was significantly higher than that in the nondiabetic controls (Figure 1H). HRVECs were cultured in high-glucose medium to mimic diabetic conditions in vitro. High-glucose treatment could upregulate circHIPK3 expression in a time-dependent manner (Figure 1I). Northern blots further verified that high glucose upregulated circHIPK3 expression in vivo and in vitro (Figure IIIB in the online-only Data Supplement). Moreover, diabetes mellitus–related pathological factors, such as oxidative stress and inflammatory stimulus, could significantly upregulate circHIPK3 expression in vitro (Figure IV in the online-only Data Supplement).
Bioinformatics analysis using the TRANSFAC program revealed the enrichment for circHIPK3 transcribed by c-myb, a transcription factor involved in the pathogenesis of diabetes mellitus.17 Both high-glucose treatment and oxidative stress could significantly upregulate c-myb expression (Figure VA in the online-only Data Supplement). c-myb silencing led to a marked reduction in circHIPK3 expression (Figure VB in the online-only Data Supplement). Chromatin immunoprecipitation assays revealed increased binding of c-myb to the promoter of circHIPK3 under high-glucose and oxidative stress conditions (Figure VC in the online-only Data Supplement). Moreover, high glucose and oxidative stress increased the luciferase activity of the vector-containing c-myb site within circHIPK3 promoter, but did not affect the luciferase activity of the vector with the mutant c-myb–binding site (Figure VD in the online-only Data Supplement). Collectively, these results suggest that c-myb is an upstream regulator of circHIPK3 expression through binding to the circHIPK3 promoter.
circHIPK3 Regulates Endothelial Cell Function Under Basal Conditions and High-Glucose Conditions In Vitro
We next investigated the role of circHIPK3 in retinal endothelial cells. Three different small interfering RNAs (siRNAs) were designed for circHIPK3 silencing. siRNA1 or siRNA3 transfection significantly downregulated circHIPK3 expression (Figure 2A).
We first investigated the role of circHIPK3 in HRVECs under basal conditions. MTT (3-[4,5-Dimethythiazol-2-yl]-2,5-diphenyl tetrazolium bromide) assays revealed that circHIPK3 siRNA1 transfection significantly decreased HRVEC viability (Figure 2B). EdU (5-Ethynyl-2′-deoxyuridine) incorporation assays revealed that siRNA1 transfection reduced HRVEC proliferation (Figure 2C). Transwell and Matrigel tube formation assays revealed that circHIPK3 silencing by siRNA1 significantly inhibited the migration and tube formation of HRVECs (Figure 2D and 2E). circHIPK3 silencing by siRNA3 but not siRNA2 also significantly reduced the viability, proliferation, migration, and tube formation of HRVECs (Figure VI in the online-only Data Supplement). We also revealed that circHIPK3 silencing by siRNA1 or siRNA3 significantly decreased the viability, proliferation, migration, and tube formation of HRVECs under high-glucose conditions (Figure VII in the online-only Data Supplement).
We then conducted the gain-of-function analysis of circHIPK3 and determined whether circHIPK3 was sufficient alone to drive endothelial phenotypes. circHIPK3 overexpression significantly increased the viability and proliferation, and accelerated the migration and tube formation of HRVECs under both basal conditions and high-glucose conditions (Figure VIII in the online-only Data Supplement).
circHIPK3 Serves as a miRNA Sponge in Retinal Endothelial Cells In Vitro
Stable transcripts with many miRNA-binding sites may function as miRNA sponges.18 circHIPK3 was mainly expressed in the cytoplasm of HRVECs. We speculated that circHIPK3 might act as a miRNA sponge to regulate gene expression. Ago2 protein is a core component of RNA-induced silencing complex that binds miRNA complexes to target mRNAs.19 RNA immunoprecipitation assays showed that endogenous circHIPK3 was specifically enriched in the immunoprecipitate pulled down by Ago2 antibody but not IgG (Figure 3A).
Bioinformatics prediction analysis showed that miR-338-3p, miR-30d-3p, miR-30a-3p, miR-30e-3p, and miR-335-3p potentially interacted with circHIPK3.20 The entire circHIPK3 sequence was inserted into pGL3 luciferase reporter to create LUC-circHIPK3 vector. Three miRNAs (miR-30a-3p, miR-30d-3p, and miR-30e-3p) significantly reduced the activity of LUC-circHIPK3 by at least 37% (Figure 3B). miR-30–binding sites on circHIPK3 were shown (Figure IXA in the online-only Data Supplement). LUC-circHIPK3 mutant without miR-30–binding sites was also constructed. miR-30 mimic transfection had no effect on the activity of LUC-circHIPK3 mutant (Figure 3C). If circHIPK3 indeed interacted with miR-30, they should be coexpressed. RNA-fluorescence in situhybridization assays revealed a large degree of overlap between circHIPK3 and miR-30a-3p, miR-30d-3p, or miR-30e-3p (Figure IXB in the online-only Data Supplement). Furthermore, using a biotin-coupled miR-30a-3p, miR-30d-3p, and miR-30e-3p, we observed greater enrichment of circHIPK3 in miR-30–captured fraction in comparison with the negative control, biotinylated miR-335 (Figure 3D). These results suggest that circHIPK3 serves as a binding platform for Ago2 and miRNAs, and may act as a miRNA sponge.
Next, we used TargetScan to predict the potential target genes of miR-30a-3p, miR-30d-3p, and miR-30e-3p. The sequence of miR-30a-3p, miR-30d-3p, and miR-30e-3p was highly conserved between human genome and mouse genome (Figure X in the online-only Data Supplement). Three candidates, including vascular endothelial growth factor-C, FZD4, and WNT2, were identified as the common target genes, which have been reported to be involved in angiogenesis and retinopathy.21,22 The seed sequence is highly conserved between human and murine genome (Figure XI in the online-only Data Supplement). miR-30a-3p, miR-30d-3p, and miR-30e-3p mimic transfection significantly downregulated VEGFC, FZD4, and WNT2 expression (Figure XII in the online-only Data Supplement). By contrast, miR-30 mimic transfection did not affect the expression of other angiogenic factors, including fibroblast growth factor 2, NOTCH1, CXCR4, LRP (low-density lipoprotein receptor–related protein) 5, and LRP6 (Figure XIII in the online-only Data Supplement).23
The 3′-untranslated region of VEGFC, FZD4, and WNT2 gene was cloned into the luciferase vector and cotransfected with miR-30 mimic into HEK293T cells. A significant reduction in luciferase activity was detected in the presence of miR-30 mimic, whereas the mutation of the miR-30 target site completely abolished this repression (Figure XIV in the online-only Data Supplement). We also revealed that circHIPK3 silencing significantly reduced the expression of VEGFC, FZD4, and WNT2 in vitro (Figure XV in the online-only Data Supplement).
circHIPK3-miR-30a-3p-VEGFC/WNT2/FZD4 Network Regulates Retinal Endothelial Cell Function In Vitro
We determined the relative expression abundance of circHIPK3, miR-30a-3p, miR-30d-3p, and miR-30e-3p in HRVECs and in mouse retinas. circHIPK3 had an expression abundance similar to that of miR-30a-3p, and had higher expression levels than miR-30d-3p and miR-30e-3p (Figure XVI in the online-only Data Supplement). Because of its high abundance, we mainly investigated the role of miR-30a-3p in HRVECs. miR-30a-3p mimic transfection significantly decreased the viability (Figure 4A), proliferation (Figure 4B), migration (Figure 4C), and tube formation ability (Figure 4D) of HRVECs, which could mimic the effects of circHIPK3 silencing. circHIPK3 overexpression abrogated miR-30a-3p–mediated repressive effects under both basal conditions and high-glucose conditions in vitro (Figure 4A through 4D and Figure XVII in the online-only Data Supplement). By contrast, miR-30a-3p inhibitor transfection significantly increased the viability and proliferation and accelerated the migration and tube formation of HRVECs (Figures XVIII and XIX in the online-only Data Supplement).
We also showed that exogenous VEGFC addition could partially reverse the repressive effect of circHIPK3 silencing on cell viability, proliferation, migration, and tube formation of HRVECs (Figure XX in the online-only Data Supplement). In addition, WNT2 or FZD4 overexpression partially reversed the repressive effect of circHIPK3 silencing on HRVEC function (Figure XXI in the online-only Data Supplement). Collectively, these results suggest that circHIPK3-VEGFC/WNT2/FZD4 cross talk regulates endothelial cell function in vitro.
circHIPK3 Regulates Diabetes Mellitus–Induced Retinal Vascular Dysfunction In Vivo
We then determined whether circHIPK3 was involved in retinal vascular dysfunction in vivo. We designed 3 different adeno-associated viral shRNAs for circHIPK3 silencing, including 1 shRNA targeting the backsplice sequence of circHIPK3 (shRNA1), 1 shRNA targeting the sequence only existing in HIPK3 linear transcript (siRNA2), and 1 shRNA targeting the sequence shared by both linear and circular HIPK3 transcript (shRNA3). shRNA1 or shRNA3 transfection significantly downregulated circHIPK3 expression. shRNA1 showed higher circHIPK3 silencing efficiency than shRNA3 (Figure XXIIA in the online-only Data Supplement). We thus selected shRNA1 for circHIPK3 silencing because of its silencing efficiency and specificity.
circHIPK3 shRNA1 injection significantly reduced retinal circHIPK3 but not linear HIPK3 mRNA expression throughout the experiment (Figure 5A). circHIPK3 shRNA1 injection did not alter blood glucose level and body weight of diabetic mice (Table I in the online-only Data Supplement). Moreover, viral shRNA injection did not induce a detectable immune response. Interleukin-6 and MCP-1 (monocyte chemoattractant protein 1) levels in the serum and vitreous of mice injected with scrambled shRNA or circHIPK3 shRNA did not differ from that of PBS-injected mice (Figure XXIIB and XXIIC in the online-only Data Supplement).
Diabetic retinopathy is the most common microvascular complication of diabetes mellitus.3 Long-term hyperglycemia led to severe capillary degeneration. circHIPK3 silencing partially reduced the injurious effects of hyperglycemia on retinal vascular function as shown by decreased acellular capillary number (Figure 5B). Evans blue leakage assays revealed that circHIPK3 silencing alleviated diabetes mellitus–induced retinal vascular leakage (Figure 5C). Retinal inflammation plays an important role in the pathogenesis of diabetic microvascular complication.24 ELISA assays showed that, in comparison with scrambled shRNA-injected retinas, circHIPK3-silencing retinas had lower expression levels of interleukin-2, interleukin-3, MCP-1, vascular endothelial growth factor, and tumor necrosis factor-α (Figure 5D). We further determined the effect of circHIPK3 overexpression on diabetes mellitus–induced retinal vascular dysfunction. circHIPK3 overexpression significantly increased acellular capillary number and retinal vascular leakage (Figure XXIII in the online-only Data Supplement). Collectively, these results indicate that circHIPK3 regulates diabetes mellitus–related retinal vascular dysfunction in vivo.
circHIPK3-miR-30a-3p-VEGFC/WNT2/FZD4 Network Regulates Diabetes Mellitus–Induced Retinal Vascular Dysfunction In Vivo
We mainly investigated the role of miR-30a-3p diabetes mellitus–induced retinal vascular dysfunction in vivo. miR-30a-3p silencing by antagomir injection led to increased VEGFC, FZD4, and WNT2 expression (Figure XXIVA in the online-only Data Supplement). Using a biotin-coupled miR-30a-3p, we observed greater enrichment of VEGFC, FZD4, and WNT2 in miR-30a-3p–captured fraction in comparison with the negative control, biotinylated miR-335, in mouse retinal homogenate (Figure XXIVB in the online-only Data Supplement). We also observed greater enrichment of circHIPK3 in the miR-30a-3p–captured fraction (Figure XXIVC in the online-only Data Supplement). In diabetic retinas, we detected increased expression of VEGFC, WNT2, or FZD4. circHIPK3 silencing significantly decreased the expression of VEGFC, FZD4, and WNT2 (Figure XXIVD and XXIVE in the online-only Data Supplement).
We further investigated the role of miR-30a-3p in retinal vascular dysfunction in vivo. Retinal trypsin digestion assays revealed that, in comparison with the diabetic group (diabetic retinopathy), miR-30a-3p upregulation by agomir injection significantly decreased acellular capillary number. By contrast, miR-30a-3p downregulation by antagomir injection had an opposite effect (Figure 6A). Evans blue assays showed that miR-30a-3p upregulation significantly decreased diabetes mellitus–induced retinal vascular leakage, whereas miR-30a-3p downregulation increased retinal vascular leakage (Figure 6B).
We also determined whether exogenous miR-30 addition could overwhelm the sponge function of circHIPK3. In comparison with the diabetic retinopathy group, circHIPK3 silencing led to decreased acellular capillary number and reduced diabetes mellitus–induced retinal vascular leakage. circHIPK3 silencing could significantly increase the release of sponged miR-30a-3p. miR-30a-3p downregulation by exogenous antagomir injection could rescue the phenotype of circHIPK3 silencing on retinal vascular dysfunction in vivo (Figure 6C and 6D). Thus, we conclude that exogenous miR-30a-3p intervention could overwhelm the miRNA sponge function of circHIPK3 in retinal vascular dysfunction.
Translation of circHIPK3 Molecular Effect to Human Disease
We then investigated whether circHIPK3 expression was altered in a diabetes complication, diabetic retinopathy. circHIPK3 expression in the fibrovascular membranes of diabetic patients was significantly higher than that in the idiopathic epiretinal membranes of nondiabetic controls (Figure 7A and Table II in the online-only Data Supplement). Because circHIPK3 was upregulated in the pathological tissues of the patients with diabetic retinopathy, we hypothesized that its circulating level was also upregulated. circHIPK3 was significantly upregulated in the plasma fraction but not the cellular fraction of peripheral blood in diabetic patients (Figure 7B and 7C, and Table III in the online-only Data Supplement). Aqueous humor is an important body fluid in the eye, which is known to be related with various ocular diseases.25 circHIPK3 expression was significantly upregulated in the aqueous humor of diabetic patients, but not in other patients with glaucoma, cataract, or trauma (Figure 7D and Table IV in the online-only Data Supplement). Collectively, these results suggest that circHIPK3 is potentially involved in the pathogenesis of diabetes complications.
circRNAs have recently gained attention because of their roles in gene expression regulation and human diseases.26 In this study, we investigated the role of circHIPK3 in maintaining vascular endothelial function. circHIPK3 could regulate retinal microvascular dysfunction in vivo and retinal endothelial cell function in vitro. Mechanistically, circHIPK3 acts as an endogenous miR-30a-3p sponge to inhibit miR-30a-3p activity, thereby leading to increased VEGFC, FZD4, and WNT2 expression.
Diabetes mellitus–induced retinal vascular complication is the major cause of blindness, which is characterized by inflammation, loss of capillaries, increased vascular permeability, and neovascularization.18 Hyperglycemia significantly upregulates circHIPK3 expression in retinal endothelial cells and mouse retinas. circHIPK3 silencing could reduce microvascular leakage and inflammation response, and decrease diabetes mellitus–induced acellular capillary number. Endothelial cells are the main targets of hyperglycemic damage. In the diseased condition, unceasing and excessive proliferation and migration of endothelial cells occur in the retinal vascular system.27 Previous studies have shown that circHIPK3 silencing relieves the proliferation speed of tumor cells.13 We show a similar role of circHIPK3 in retinal endothelial cells. circHIPK3 silencing decreases abnormal proliferation, mobility, and tube formation of retinal endothelial cells in vitro. We thus speculate that circHIPK3 deregulation is responsible for diabetes mellitus–induced retinal microvascular dysfunction.
circHIPK3 upregulation is shown as a stress response on high-glucose stress. At the early stage, hyperglycemia causes blood flow disruption, basement membrane thickening, pericyte loss, and vascular leak.28,29 Capillary dropout further leads to a hypoxic condition, increased inflammatory response, and eventually neovascularization.18 This stage is accompanied by increased circHIPK3 expression. circHIPK3 silencing could retard the progression of microvascular dysfunction. Endothelial cells would undergo growth, increased permeability, remodeling, and phenotypic alterations in diabetes mellitus. Increased circHIPK3 expression could lead to abnormal endothelial cell proliferation and migration. circHIPK3 silencing could alleviate diabetes mellitus–induced retinal endothelial dysfunction.
circRNAs have been shown as promising candidates for additional layers of gene expression control. They can regulate gene expression by competing for miRNA or protein binding.13,30 They also can compete with linear RNA via affecting the accumulation of full-length mRNA.31 A recent study reveals a very novel mechanism of circular RNA action. Circ-ZNF609 can be translated and functions in myogenesis, providing an example of a protein-coding circRNA in eukaryotes.32 circHIPK3 is mainly expressed in the cytoplasm of HRVECs. circHIPK3 serves as a binding platform for Ago2 and miRNAs, and functions as a miRNA sponge for miR-30a-3p, miR-30d-3p, and miR-30e-3p. Previous study shows that circHIPK3 binds to 18 sites of 9 miRNAs. circHIPK3 functions as the sponge of miR-124 and affects the proliferation of tumor cells.13 However, we did not detect a significant reduction of circHIPK3 in miR-124 inhibition HRVECs. We speculate that circHIPK3 may play its biological roles in tissue- or cell-specific context.
Disorders of retinal vascular growth and function are responsible for vision loss in ocular neovascular diseases, including diabetic retinopathy, age-related macular degeneration, retinal artery/vein occlusion, and retinopathy of prematurity.33,34 VEGFC, FZD4, and WNT2 expression has been reported to be upregulated in these diseases.35,36 circHIPK3 functions as a miRNA sponge to inhibit miR-30a-3p activity, thereby upregulating VEGFC, FZD4, and WNT2 expression. circHIPK3 overexpression becomes a sink for miR-30a-3p, and releases the repressive effect of miR-30a-3p on VEGFC, FZD4, and WNT2. This regulatory mechanism would provide a novel insight into microvascular dysfunction.
We show that, in retinal endothelial cells, high-glucose treatment upregulates circHIPK3 expression. This upregulation is responsible for altered retinal endothelial cell function and microvascular dysfunction. The regulatory effect of circHIPK3 on retinal vascular function is mediated by acting as an endogenous miR-30a-3p sponge. This study shed light on a new potentially targeted method to prevent diabetes mellitus–induced vascular complications by using a noncoding RNA-based approach.
We thank Q. Jiang, J. Yao, X.-M. Li, and H.-M. Ge (Nanjing Medical University, China) for the excellent technical and clinical assistance; Z.-H. Wang, W. Song, T.-G. Liu (College of Information, Shanghai Ocean University, China) for statistical discussion and the support of bioinformatics analysis; and S.-L. Huang (Fudan University, China) for technical assistance for Northern blot and helpful discussion.
Sources of Funding
This work was generously supported by the grants from the National Natural Science Foundation of China (no.81770945 to Dr Yan; no. 81730025 and 81525006 to Dr Zhao), a grant from the Shanghai Youth Talent Support Program (to Dr Yan), a grant from the Scientific Research Start-up Funding for Advanced Talents (to Dr Yan), and a grant from the Young Scientists Program from Eye and ENT Hospital (to Dr Yan).
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.117.029004/-/DC1.
Circulation is available at http://circ.ahajournals.org.
- Received April 23, 2017.
- Accepted August 28, 2017.
- © 2017 American Heart Association, Inc.
- You X,
- Vlatkovic I,
- Babic A,
- Will T,
- Epstein I,
- Tushev G,
- Akbalik G,
- Wang M,
- Glock C,
- Quedenau C,
- Wang X,
- Hou J,
- Liu H,
- Sun W,
- Sambandan S,
- Chen T,
- Schuman EM,
- Chen W
- Holdt LM,
- Stahringer A,
- Sass K,
- Pichler G,
- Kulak NA,
- Wilfert W,
- Kohlmaier A,
- Herbst A,
- Northoff BH,
- Nicolaou A,
- Gäbel G,
- Beutner F,
- Scholz M,
- Thiery J,
- Musunuru K,
- Krohn K,
- Mann M,
- Teupser D
- Boeckel JN,
- Jaé N,
- Heumüller AW,
- Chen W,
- Boon RA,
- Stellos K,
- Zeiher AM,
- John D,
- Uchida S,
- Dimmeler S
- Robitaille J,
- MacDonald ML,
- Kaykas A,
- Sheldahl LC,
- Zeisler J,
- Dubé MP,
- Zhang LH,
- Singaraja RR,
- Guernsey DL,
- Zheng B,
- Siebert LF,
- Hoskin-Mott A,
- Trese MT,
- Pimstone SN,
- Shastry BS,
- Moon RT,
- Hayden MR,
- Goldberg YP,
- Samuels ME
- Liu CH,
- Sun Y,
- Li J,
- Gong Y,
- Tian KT,
- Evans LP,
- Morss PC,
- Fredrick TW,
- Saba NJ,
- Chen J
- Yan B,
- Yao J,
- Liu JY,
- Li XM,
- Wang XQ,
- Li YJ,
- Tao ZF,
- Song YC,
- Chen Q,
- Jiang Q
- Shen Y,
- Dong LF,
- Zhou RM,
- Yao J,
- Song YC,
- Yang H,
- Jiang Q,
- Yan B
- Barrett SP,
- Salzman J
- Cai H,
- Harrison DG
- Legnini I,
- Di Timoteo G,
- Rossi F,
- Morlando M,
- Briganti F,
- Sthandier O,
- Fatica A,
- Santini T,
- Andronache A,
- Wade M,
- Laneve P,
- Rajewsky N,
- Bozzoni I