Ethics
All procedures were performed in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the use of Animals in Ophthalmic and Vision Research. In addition, the Institutional Animal Care and Ethics Committee of Zhejiang University (Hangzhou, China) approved all animal experiments. All manipulations were performed following the rules outlined in our previous paper [32].
EPC culture
Prior to the experimental procedures, Sprague-Dawley (SD) rats weighing 200–300 g (2–3 months old) were anesthetized with an intraperitoneal injection of sodium pentobarbital (30 mg/kg body weight; Merck KGaA, Darmstadt, Germany). EPCs were isolated, cultured, and identified as described previously 34. Briefly, blood samples were obtained from the SD rats. After centrifugation in Ficoll-Paque Plus (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) peripheral blood mononuclear cells were isolated; the cells were re-suspended in endothelial growth medium (EGM–2-MV; Lonza, Basel, Switzerland) and then plated into 6-well culture plates coated with fibronectin. The cultured medium was replaced every 2 d and non-adherent cells were removed thereafter.
Recombinant lentivirus construction and transfection of EPC
The endostatin fragments were cloned by PCR. The recombinant lentiviral vector (endostatin-lentivirus-GFP) was produced by co-transfection of 293T human embryonic kidney cells using Lipofectamine® 2000 reagent (Invitrogen, Grand Island, NY, USA). The primary EPCs were transferred into 6-well plates at 106 cells/well for lentiviral transduction. The medium containing the recombinant lentiviral vector (endostatin-lentivirus-GFP) and Polybrene (5μg/ml; Sigma Aldrich Corp.; St. Louis, MO, USA) was added at a multiplicity of infection of 100 to improve infection efficiency and was mixed with the cells. After incubation for 24 h, the cell-culture medium was removed and replaced with Dulbecco’s minimum essential medium supplemented with 10% fetal bovine serum. Cells were cultured for 96 h. Then, the relative expression levels of the endostatin gene were quantified by qRT-PCR. Non-transduced cells were used as the blank control, and cells transduced with GFP alone were used as the NC group.
For the qRT-PCR analysis, total RNA was extracted from the transfected EPCs using TRIzol® reagent (Invitrogen, USA). The relative expression levels of endostatin genes were quantified as previously described 34. Briefly, using the 2-ΔΔCt method, each sample was subjected to triplicate experiments: PCR products were incubated at 95˚C (3 min) and then run for 40 cycles at 95˚C (12 s) and 62˚C (40 s). The following primers were used: endostatin, forward: 5’-TCTCCCAAGTCGAAGACCCT–3’ and reverse: 5’-GAACAGCAGCGAAAAGTCCC–3’; GAPDH, forward: 5’-TCTCTGCTCCTCCCTGTTCT–3’ and reverse: 5’- ATCCGTTCACACCGACCTTC–3’. Results were normalized against GAPDH as a housekeeping gene control.
Groups and animal treatment
On postnatal day (P) 7, the SD rats were randomly divided into experimental and control groups. Litters in the experimental group were exposed to hyperoxia (70% O2), and were fed by their mothers for 5 days (P7-P12) in a chamber and then returned to normal room air (20% O2) to induce oxygen-induced retinopathy (OIR). Oxygen concentration and room temperature were monitored and recorded three times per day. The control group was kept in normoxia (room air) under a normal diet and a 12 h light and dark cycle. Following euthanasia with pentobarbital sodium at different time points (P14, P15, P17, and P19), fundus fluorescein angiography (FFA) was performed for the experimental and control groups to confirm the established OIR rat model.
The age-matched rats under the normoxia condition with non-intravitreal injections were considered the blank control group. The OIR rats were used to compare the efficacy of intravitreal injections of endostatin-lentivirus, endostatin-lentivirus-EPC, and simple EPC groups. The OIR rats were randomized into four groups, as follows:
Group Ⅰ: Blank control group (normoxia and non-injection, n = 10)
Group Ⅱ: OIR+NC group (empty-lentivirus injection, 1μg/μL, 1.5 μL, n = 10)
Group Ⅲ: OIR+endostatin-lentivirus group (endostatin-lentivirus injection, 1μg/μL, 1.5 μL, n = 10)
Group Ⅳ: OIR+EPC group (EPCs injection, 5×106/mL, 1.5 μL, n = 10)
Group Ⅴ: OIR+endostatin-lentivirus-EPC group (endostatin-lentivirus-EPCs injection, 5×106/mL, 1.5 μL, n = 10)
On P14, the right eye of each anaesthetized rat received the same volume (1.5 μL) of intravitreal injection using a Hamilton syringe (#87900, Bonaduz, Switzerland) through the pars plana under a dissecting microscope; after injection, the animals were kept in normoxia until they were analyzed. Then, neovascularization leakage areas were compared using fundus fluorescein angiography at 1 h, 1 d, 3 d, and 5 d after intravitreal injections. Pups were sacrificed with an intraperitoneal overdose injection of pentobarbital on P19. Then the eyes were rapidly collected for histology analysis (HE and IHC). The Hematoxylin-eosin (HE) staining method was used to observe and count the number of nuclear cells in the ECs outside the retinal inner limiting membrane of the area affected by retinal neovascularization. Immunohistochemistry (IHC) was used to measure the retinal expression of endostatin, vascular endothelial growth factor (VEGF), and CD31. Figure 10 presents a detailed flow chart of animal treatment.
Fundus fluorescein angiography: Fundus fluorescein angiography was performed at 1 h, 1 d, 3 d, and 5 d after the intravitreal injections. Prior to fundusfluorescein angiography treatment, neonatal SD rats were anesthetized with a pentobarbital injection, and pupils were dilated with tropicamide phenylephrine eye drops (Santen Pharmaceutical Co. Ltd., Japan). Fluorescein sodium (10%, 75 mg/kg) was injected into the rats’ tail veins, and fundus fluorescein angiography images were immediately captured by a Heidelberg fundus fluorescein angiography instrument (Spectralis HRA, Germany). The retinal neovascularization leakage area was assessed by outlining the border of each lesion using Heidelberg software and analyzed statistically.
HE staining: Animals were sacrificed at P19 after the last intravitreal injection; one eye of each rat was enucleated and fixed in 4% paraformaldehyde at 4℃ overnight. The orientation of the corneal limbus was marked at 12 o’clock of the corneal limbus. Then, the eyes were dehydrated and embedded in paraffin. Five-micrometer serial sections of whole eyes were cut sagittally through the cornea parallel to the optic nerve (without cutting the optic nerve). Each section was 30 μm apart and was stained with hematoxylin and eosin. The images were obtained and analyzed under a light microscope (Olympus, Japan). Three randomly selected areas in the sample coverslips were examined at a high magnification view (×200). The number of nuclear cells in the ECs outside the retinal inner limiting membrane of the retinal neovascularization was counted and calculated, and the average value of each magnification view was used for statistical analysis.
Assessment of endostatin, VEGF, and CD31 expression: Retinal expression of endostatin, VEGF, and CD31 was evaluated by the IHC method. The IHC staining procedure was performed following the manufacturer’s instructions. Unless otherwise stated, all washes were conducted for 3 times for 5 min in PBS at PH 7.4 and performed at room temperature, while incubations were at 37°C. Eye sections were deparaffinized and dehydrated, and then the antigens were repaired using a heated citric acid repair liquid (P0083; Beyotime Institute of Biotechnology, Shanghai, China).
Endogenous peroxidase activity was blocked by incubating the sections in 3% H2O2 for 20 min; then, the sections were washed and placed in goat serum for 10–15 min to block non-specific labeling. The sections were incubated at 4°C overnight with one of the following primary antibodies (50 μL each): polyclonal rabbit anti-ES (1:200; ab202973; Abcam, UK), polyclonal rabbit anti-VEGF (1:200; 19003–1-AP; Proteintech, USA), or polyclonal rabbit anti-CD31 (1:200; AF6191; Affinity, USA).
Negative controls were incubated on slides without any primary antibodies. Slides were washed and incubated for 15 min with 50 μL biotin-conjugated secondary antibody (1:1000; ab6802; Abcam, UK). Subsequently, a tertiary layer of streptavidin peroxidase was applied according to the manufacturer’s instructions (SABC-POD kit, Boster Biological Technology, Pleasanton, CA).
Antigen–antibody complexes were detected by incubation with diaminobenzidine (P0203; Beyotime Institute of Biotechnology, Shanghai, China) at room temperature for 3–30 min. Slides were counterstained with hematoxylin (Beyotime Institute of Biotechnology, Shanghai, China) for 3 min. Finally, the sections were washed, dehydrated, embedded in paraffin, and photographed. Positive cells were brown-stained in the cytoplasm and nucleus of the ganglion cell layer or inner nuclear layer, but no brown-stained cells were negative.
For immunocytochemical analysis, sections were coded and counted in a blind fashion on a light microscope (Olympus, Japan). A total of six visual fields from randomly selected areas in the sample coverslips were examined. Scopes were chosen as the percentage of positive cells that colocalized with endostatin, VEGF, or CD31. IHC staining gray scale was analyzed using Image-ProPlus (IPP) software (version 6.0, Media Cybernetics Inc., Rockville, MD, USA) and expressed as integral optical density (IOD). The average value of the IOD area was used for statistical analysis.
Statistical analysis
The software used was SPSS 19.0 (IBM Corp., Armonk, NY, USA). A one-way analysis of variance was used to analyze all data. Values of P<0.05 were considered statistically significant differences.