siRNA targeting CCN1 (CCN1 siRNA) and scrambled siRNA were purchased from GenePharma Co., Ltd. (Shanghai, China), and used according to previously published methods (23,24). The sequences were as follows: CCN1 (Cyr61-homo-553) sense, 5′-GGGAAAGUUUCCAGCCCAACUTT-3′ and antisense, 5′-AGUUGGGCUGGAAACUUUCCCTT-3′; CCN1 (Cyr61-homo-789) sense, 5′-GAGGUGGAGUUGACGAGAAACTT-3′ and antisense, 5′-GUUUCUCGUCAACUCCACCUCTT-3′; CCN1 (Cyr61-homo-1072) sense, 5′-GCAAGAAAUGCAGCAAGACCATT-3′ and antisense, 5′-UGGUCUUGCUGCAUUUCUUGCTT-3′; CCN1 (Cyr61-homo-1268) sense, 5′-GAUGAUCCAGUCCUGCAAAUGTT-3′ and antisense, 5′-CAUUUGCAGGACUGGAUCAUCTT-3′; and scrambled siRNA sense, 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense, 5′-ACGUGACACGUUCGGAGAATT-3′. The siRNA were cloned into the pGPU6/GFP/Neo siRNA expression vector kit (GenePharma, Ltd.) to create the pGPU6/GFP/Neo-Cyr61 siRNA and the pGPU6/GFP/Neo-scrambled siRNA plasmids. The plasmids contained the Bbs1 and BamH1 restriction sites. All plasmids contained green fluorescent protein (GFP). The Cyr61-homo-1072 construct yielded the optimal results (data not shown), and was thus used in the following experiments.

HUVECs were purchased from Cell Systems (Kirkland, WA, USA) and cultured in Dulbecco's minimum essential medium (DMEM; HyClone, Logan, UT, USA) with 10% fetal bovine serum (FBS; HyClone, Thermo Fisher Scientific, Waltham, MA, USA) in a 37°C humidified atmosphere containing 95% air and 5% CO₂. Subconfluent monolayers of HUVECs from passages 3 to 10 were used in the following experiments.

The HUVECs were subsequently divided into the normoxia (20% O₂/5% CO₂/75% N₂) and hypoxia groups (1% O₂/5% CO₂/94% N₂). The hypoxia group was further subdivided into the hypoxia group, the hypoxia-scrambled siRNA group (transiently transfected with scrambled siRNA) and the hypoxia-CCN1 siRNA group (transiently transfected with CCN1 siRNA). The plasmids (500 ng/µl) were transiently transfected into the HUVECs using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA).

The PI3K/AKT inhibitor, LY294002, was used to treat the cells exposed to hypoxia to determine the effects of inhibiting this pathway. The cells were cultured under hypoxic conditions in the presence of LY294002 (Sigma, St. Louis, MI, USA) (40 µmol/l, dissolved in DMSO; the final concentration of DMSO in the cell culture was 0.1%). The cells were treated with LY294002 for 30 min before being placed in the incubator. An apoptosis assay, western blot analysis and reverse-transcription quantitative PCR (RT-qPCR) were performed after 24 h of exposure to hypoxia.

The cell counting kit-8 (CCK-8) assay (C0038; Beyotime, Jiangsu, China) was used to evaluate cell proliferation. The HUVECs were plated in 96-well plates at a density of 2,000 cells/well. Cell proliferation was evaluated each day for 4 days following transfection. CCK-8 (10 µl) was added to each well followed by incubation for 2 h at 37°C. After 10 min of vortexing, the absorbance was measured in a micro-plate reader (Sunrise RC; Tecan, Mannedorf, Switzerland) at 450 nm.

Apoptosis was measured using a fluorescein isothiocyanate (FITC) Annexin V Apoptosis Detection kit (BD Biosciences, San Diego, CA, USA) according to the manufacturer's instructions. The HUVECs were collected, washed in cold phosphate-buffered saline (PBS), and labeled with 5 µl Annexin V-FITC and 5 µl propidium iodide (PI), and then incubated for 15 min at room temperature in the dark. Flow cytometry was immediately performed and the cells were analyzed using CellQuest software (BD Biosciences, Franklin Lakes, NJ, USA). Annexin V was set as the horizontal axis and PI was set as the vertical axis. Mechanically damaged cells were located in the upper left quadrant, late apoptotic or necrotic cells in the upper right quadrant, dual-negative and normal cells in the lower left quadrant, and early apoptotic cells in the lower right quadrant of the flow cytometry dot plot. The apoptotic rate was calculated as the ratio of early and late apoptotic cells to total cells, as previously described (25,26).

The HUVECs were washed twice with PBS and fixed with 4% paraformaldehyde for 30 min. The cells were then washed with 0.1% Triton X-100 for 10 min and twice with PBS, and were incubated with 1% bovine serum albumin (BSA) at room temperature for 1 h. The cells were incubated overnight with the following commercially available primary antibodies: rabbit anti-Cyr61 polyclonal antibody (ab24448; Abcam, Cambridge, UK), mouse anti-p-PI3K monoclonal antibody (sc-12929), or rabbit anti-p-AKT1/2/3 (Ser473) polyclonal antibody (sc-101629) (both from Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). The cells were washed thrice with PBS. The primary antibodies were replaced by isotype controls, which were used as negative controls. The cells were then treated with fluorescence-conjugated secondary antibody [FITC-conjugated AffiniPure rabbit anti-goat IgG (H+L; ZF-0314), or tetramethylrhodamine (TRITC)]-conjugated AffiniPure goat anti-mouse IgG (H+L; ZF-0313) (Zhongshan Jinqiao Biotechnology Co., Ltd., Beijing, China) for 1 h and counter-stained with 0.5 µg/ml 4′,6-diamidino-2-phenylindole (DAPI; Beyotime Institute of Biotechnology, Jiangsu, China) for 5 min. Images were digitally captured using a confocal laser scanning microscope (FV1000; Olympus Corp., Tokyo, Japan).

Specific pathogen-free healthy C57BL/6J newborn mice (female or male; n=240; and their mothers) were obtained from the Animal Laboratory of China Medical University, Shenyang, China and were housed with their mothers. The room temperature was maintained at 23±2°C. Light did not exceed 300 lux on a 12 h light-dark cycle.

All procedures and animal experiments were approved by the Animal Care and Use Committee of the China Medical University.

OIR was induced in the C57BL/6J mice as previously described in the study by Smith et al (27). Briefly, on postnatal day (P)7, the pups and their mothers were placed in homemade glass containers coupled to an RSS-5100 oxygen analyzer (Rex Xinjing Instrument Co., Ltd., Shanghai, China). The mice were exposed to hyperoxia (75±2% O₂) for 5 days (P7-P12), and were then re-exposed to normoxia (room air) for 5 days. The rationale of exposing mice to hyperoxia and then to normoxia was to emulate a state of relative hypoxia. Neovascularization occurred when the mice re-exposed to normoxia and peaked at P17, as previously observed (27). The mice were randomly divided into 4 groups: the normoxia, hyperoxia, hyperoxia-scrambled siRNA and hyperoxia-CCN1 siRNA groups (n=60/group).

In the normoxia group, the newborn mice were maintained in room air from P0 to P17. In the hyperoxia group, OIR was induced by the mice being exposed to hyperoxia (75±2% O₂) for 5 days (P7–P12) and then re-exposed to normoxia (room air) for 5 days (P12–P17). The same OIR induction protocol was used in the hyperoxia-scrambled siRNA and hyperoxia-CCN1 siRNA groups. The mice were administered an intravitreal injection of 1 µl (500 ng/µl) of the scrambled siRNA plasmid or the CCN1 siRNA plasmid on P11 using a 33-gauge needle attached to a Hamilton syringe, and were returned to room air on P12, as previously described (28,29). The mice in all 4 groups were anesthetized by an intraperitoneal injection of ketamine hydrochloride (100 mg/kg body weight), and were then sacrificed by decapitation on P17 in order to collect the retinas for morphological and pathological examinations, as well as for mRNA and protein expression analyses. In the present study, the percentage of GFP-positive cells in the retina was used to assess the transfection rate 1 day after the intravitreal injection of siRNA. The transfection rate was approximately 80%. In addition, after examination it was clear that transfection did not cause endophthalmitis or retinal detachment (data not shown), suggesting that the gene transfer was successful and did not affect the eyes.

Retinal vascular patterns were assessed on P17, as previously described (30). The eyes were enucleated and fixed with 4% paraformaldehyde for 3 h. The retinas were then dissected, flat-mounted through 4 incisions in the center of the disc to divide them into 4 quadrants, as previously described (31) and processed for magnesium-activated adenosine diphosphatease (ADPase) staining. The ADPase-stained retinas were then flat-mounted on microscope slides with a gelatin-coated cover slip and carefully examined using an Olympus B201 optical microscope (Olympus Corp.). For neovascularization grading evaluation under the microscope, each retina was divided into 12-h clocks in order to assess the neovascularization clock hour scores, as previously described (32–34). Three independent reviewers blinded to grouping assessed the clock hour scores in order to assess the severity of neovascularization.

To quantify preretinal neovascularization, the retinal structures were analyzed on 6-µm hematoxylin and eosin (H&E)-stained sections, as previously described (35). Briefly, the eyes were enucleated, fixed and embedded in paraffin. Serial sections (6-µm-thick) of whole eyes were cut sagittally, through the cornea and parallel to the optic nerve, and stained with H&E. Vascular cell nuclei, identified under a light microscope (Olympus B201, Olympus Corp.), were considered to be associated with new vessels if they were found on the vitreal side of the internal limiting membrane (ILM), as previously described (35). Three independent reviewers blinded to grouping counted the cells.

Immunohistochemistry was performed using an SABC immunohistochemistry kit (Boster Bioengineering Co., Wuhan, China). Formalin-fixed, paraffin-embedded eye tissue sections (6-µm-thick) were placed on slides, deparaffinized in xylene and rehydrated by incubation in graded ethanol baths in PBS. Endogenous peroxidase was blocked with 3% hydrogen peroxide in methanol. The sections were then treated with 10% normal goat serum and incubated overnight with the following commercially available primary antibodies: rabbit anti-Cyr61 polyclonal antibody (ab24448; Abcam), mouse anti-p-PI3K monoclonal antibody (sc-12929), or rabbit anti-p-AKT1/2/3 (Ser473) polyclonal antibody (sc-101629) (both from Santa Cruz Biotechnology, Inc.) at 4°C. The sections were incubated with biotinylated horse secondary antibody againts mouse IgG (ZB-2020; Zhongshan Jinqiao Biotechnology Co., Ltd.) and reacted with the avidin-biotinylated peroxidase complex. The primary antibody was replaced with PBS for the negative controls, and 3,3′-diaminobenzidine (DAB) was used as the chromogen. The sections were counterstained with hematoxylin, dehydrated and mounted. Images were digitally captured using an Olympus B201 optical microscope (Olympus).

The HUVECS and retinal tissue homogenates were lysed using 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.5% Nonidet P-40, 0.5% sodium deoxycholate and phenylmethylsulfonyl fluoride (PMSF) (all from Sigma). Cell debris were removed by centrifugation at 12,000 rpm, 4°C, for 20 min. Protein levels were determined by bicinchoninic acid (BCA) assay (Beyotime Institute of Biotechnology).

For western blot analysis, protein samples (30 µg) were run on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). Non-specific sites on each blot were blocked with 5% BSA diluted in TBS with 0.05% Tween-20 (TBST) for 2 h. The membranes were incubated with specific primary antibodies overnight at 4°C [rabbit anti-Cyr61 polyclonal antibody (ab24448; Abcam), or mouse anti-p-PI3K monoclonal antibody (sc-12929) or rabbit anti-p-AKT1/2/3 (Ser473) polyclonal antibody (sc-101629) (Santa Cruz Biotechnology, Inc.)], followed by incubation with horseradish peroxidase-conjugated secondary antibody (ZB-5305; Zhongshan Jinqiao Biotechnology Co., Ltd.) for 60 min. Intensive washing was performed between incubations. Signals were detected by enhanced chemiluminescence (Pierce Biotechnology, Rockford, IL, USA). Protein levels were determined by densitometric scanning of the protein bands using a GIS-2020 image-processing system (TechNew Group Co., Ltd., Shanghai, China) and normalized to the intensity of the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) band (using a rabbit anti-mouse GAPDH polyclonal antibody, bsm-0978M; Biosynthesis Biotechnology Co., Ltd., Beijing, China).

Total RNA was extracted from the HUVECs and retinas using TRIzol reagent (Invitrogen). The RNA purity was determined using absorbance at 260 and 280 nm (A260/280). cDNA synthesis was performed using a reverse transcriptase kit (PrimeScript™ RT Reagent kit-Perfect Real-Time; Takara Bio, Otsu, Japan), according to the manufacturer's instructions. The primer sequences used in the HUVECs and retinas are presented in Table I. Quantitative (real-time) PCR was performed using the SYBR-Green PCR Master Mix (Premix Ex Taq™-Perfect Real-Time; Takara Bio) in a total volume of 20 µl on a 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). All reactions involved an initial denaturation at 95°C for 30 sec followed by 50 cycles of 95°C for 5 sec and 60°C for 31 sec. GAPDH was used as the reference gene. The Ct value was defined as the number of PCR cycles in which the fluorescence signal exceeded the detection threshold value. First, the ΔCt value was calculated as follows: Ct gene - Ct GAPDH. Subsequently, the ΔΔCt value was calculated as follows: ΔCt treated - ΔCt control. Lastly, the 2^−ΔΔCt value was calculated to represent the relative mRNA expression of target genes, as previously described (36).

SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Data are expressed as the means ± standard deviation (SD) of 3 independent experiments. Statistical significance was evaluated by one-way analysis of variance (ANOVA) with Fisher's least significant difference (LSD) test for post hoc analysis. A P-value <0.05 was considered to indicate a statistically significant difference.

The first step of angiogenesis is endothelial cell proliferation. The cell growth curves of 24, 48, 72 and 96 h, as assessed by the CCK-8 method, indicated that the growth rate was slower in the hypoxia-CCN1 siRNA group than in the hypoxia and hypoxia-scrambled siRNA groups Fig. 1A). An FITC Annexin V apoptosis detection kit was used to evaluate the effects of CCN1 siRNA on early and late apoptosis in the HUVECs. As shown in Fig. 1B, the early apoptotic rate was decreased, but the late apoptotic rate was significantly increased in the hypoxia-CCN1 siRNA group compared with the hypoxia-scrambled siRNA group (total apoptotic rate, 69.1±1.1 vs. 40.4±1.0%, P<0.01; Fig. 1B). These results indicated that transfection of the cells with CCN1 siRNA exerted more prominent anti-proliferative and pro-apoptotic effects on the HUVECs.

Transfection with CCN1 siRNA decreased the mRNA expression levels of all 3 factors (CCN1, PI3K and AKT). Compared with the hypoxia-scrambled siRNA group, the mRNA levels of CCN1, PI3K and AKT in the hypoxia-CCN1 siRNA group were decreased by 78.2, 50.0 and 62.7%, respectively (Fig. 2A). Immunofluorescence staining (Fig. 2B) and western blot analysis (Fig. 2C) revealed that the protein levels of CCN1, p-PI3K and p-AKT were higher in the hypoxia (protein, 0.53±0.02, 0.36±0.01 and 0.37±0.01, respectively) and hypoxia-scrambled siRNA (protein, 0.49±0.07, 0.42±0.03 and 0.42±0.03, respectively) groups compared with the normoxia group (protein, 0.22±0.03, 0.23±0.02 and 0.12±0.01, respectively; all P<0.05); however, no significant differences were observed between the hypoxia and hypoxia-scrambled siRNA groups (all P>0.05). In addition, the CCN1, p-PI3K and p-AKT protein expression levels were decreased in the hypoxia-CCN1 siRNA group compared with the hypoxia and hypoxia-scrambled siRNA groups (all P<0.05). Compared with the hypoxia-scrambled siRNA group, the CCN1, p-PI3K and p-AKT protein levels were decreased by 42.9, 26.2 and 50.0%, respectively (all P<0.05; Fig. 2C).

We performed an experiment using LY294002, an inhibitor of the PI3K/AKT pathway. The results revealed that the early apoptotic rate was decreased, but that the late apoptotic rate was significantly increased in the hypoxia-LY294002 group (total apoptotic rate, 58.1±1.2 vs. 37.9±1.5%, P<0.05) (Fig. 3A). Compared with the hypoxia group, the mRNA expression of CCN1 in the hypoxia-LY294002 group was downregulated by 84.1% (P<0.05; Fig. 3B). Compared with the hypoxia group, treatment with LY294002 decreased CCN1 protein expression in the cells exposed to hypoxia (P<0.05; Fig. 3C). These results suggest that a PI3K/AKT inhibitor may be used to decrease CCN1 expression, and that this process involves an autocrine loop.

To determine whether the silencing of CCN1 using CCN1 siRNA suppresses oxygen-induced ischemic RNV, we examined the retinal vasculature using an ADPase assay in retinal flat-mounts on P17. In our model of OIR, in the mice treated with CCN1 siRNA, alterations in vessel morphology and distribution were observed (in the flat mount image; Fig. 4A). Compared with the hyperoxia group (5.60±0.73), the retinas from the hyperoxia-CCN1 siRNA group had less severe neovascular tufts and regions of non-perfusion, vascular tortuosity and less irregular expansion (1.53±0.72, P<0.05); these values were still slightly higher than in the normoxia group (1.23±0.49, P<0.05), but much lower than in the hyperoxia-scrambled siRNA group (4.76±1.04, P<0.05) (Fig. 4A).

To further confirm the effects of CCN1 siRNA on RNV, we quantified the number of preretinal neovascular cells, a characteristic of OIR (37). Preretinal neovascular cells growing in the vitreous humor were counted on 10 non-continuous cross-sections from each eye, according to a previously established method (35). As shown in Fig. 4B, the numbers of preretinal neovascular cells in the retinas from the hyperoxia group (32.5±1.8) and the hyperoxia-scrambled siRNA group (31.4±2.6) were significantly higher than those in the retinas from the normoxia group (1.3±0.2) (both P<0.05; Fig. 4B). Moreover, the numbers of preretinal neovascular cells in the hyperoxia-CCN1 siRNA group (12.0±2.8) were significantly lower than those in the retinas from the hyperoxia and hyperoxia-scrambled siRNA groups (both P<0.05), confirming the anti-neovascularization effects of the silencing of CCN1 (by CCN1 siRNA) on the retina.

RT-qPCR was used to measure the CCN1, PI3K and AKT mRNA expression levels in the retinal samples. In the hyperoxia and hyperoxia-scrambled siRNA groups, the CCN1 (+244 and +122%, respectively), PI3K (+404 and +215%, respectively) and AKT (+202 and +140%, respectively) expression levels were increased compared with the normoxia group (all P<0.05; Fig. 5A). Compared with the hyperoxia-scrambled siRNA group, the administration of CCN1 siRNA decreased the CCN1, PI3K and AKT mRNA expression levels (−43.7, −58.7 and −42.9%, respectively, all P<0.05; Fig. 5A).

Western blot analysis revealed similar results in the retinal samples. In the hyperoxia and hyperoxia-scrambled siRNA groups, the CCN1 (+429 and +406%, respectively), p-PI3K (+124 and +115%, respectively) and p-AKT (+153 and +147%, respectively) protein expression levels were increased compared with those in the normoxia group (all P<0.05; Fig. 5B). Compared with the hyperoxia-scrambled siRNA group, the silencing of CCN1 by CCN1 siRNA decreased the CCN1, PI3K and AKT protein expression levels (−45.3, −22.5 and −28.4%, respectively, all P<0.05; Fig. 5B).

Immunohistochemistry was also performed to investigate the localization and expression levels of CCN1, p-PI3K and p-AKT (Fig. 6). Immunohistochemistry of the retinal sections revealed that the CCN1, p-PI3K and p-AKT expression levels were weakly detected only in the ganglion cell layer (GCL) and inner plexiform layer (IPL) of the normoxia group, whereas in the hyperoxia and hyperoxia-scrambled siRNA groups, they were strongly detected in the GCL, IPL, inner nuclear layer (INL) and outer plexiform layer (OPL), with neovascularization breaking through the ILM. However, the expression of CCN1, p-PI3K and p-AKT in the hyperoxia-CCN1 siRNA group was low in the GCL, IPL with less neovascularization breaking through the ILM compared with the hyperoxia and hyperoxia-scrambled siRNA groups.