Therapeutic effect against retinal neovascularization in a mouse model of oxygen-induced retinopathy: bone marrow-derived mesenchymal stem cells versus Conbercept

Commercially available red fluorescent protein-labeled BMSC of C57BL/6 mouse (RFP-BMSC, Catalog No. MUBMX-01201, Cyagen, Guangzhou, China) were cultured to determine the expressions of CD44, CD29, Sca-1, CD31 and CD117, and tested for osteogenic and adipogenic differentiation. The cells were harvested and diluted in Dulbecco’s modified Eagle’s medium (DMEM, Cyagen, Guangzhou, China) at different concentrations for injection (ranged from 5 × 10⁶ cells/ml to 1 × 10⁸ cells/ml).

All animal experiments were approved by the animal ethical committee of the Fujian Medical University, and performed in accordance with the National Institutes of Health Guide for the Use of Laboratory Animals. Pregnant C57BL/6 mice (Subline J, Specific pathogen free class) were purchased from Slaccas Animals (Shanghai, China) and housed in room air (21% oxygen) at 25 °C with free access to food and water. After delivery, the neonates together with their mother were exposed to 75% oxygen (hyperoxygen) from postnatal day 7 (P7) to P12 and returned to room air to develop oxygen-induced retinopathy (OIR). Litters at P12 (5.8 ± 0.2 g in weight) were anesthetized and randomly grouped according to the treatments: intravitreal injection of DMEM (1 μl) (OIR-DMEM group), intravitreal injection of RFP-BMSC (1 μl, OIR-BMSC group) and intravitreal injection of Conbercept (1 μl, Kanghong, Inc., Chengdu, China) (OIR-CON group). The litters without hyperoxygen exposure (healthy group) and litters without treatment after hyperoxygen exposure (OIR-blank group) were set as controls.

Mice were anesthetized with an intraperitoneal injection of ketamine (90 mg/kg) and xylazine (8 mg/kg), and administered intraventricular injection of 0.3 ml of FITC-dextran (2000 kDa, 50 mg/ml, Sigma-Aldrich, MO, USA). After 5 minutes, the mice were euthanized by intraperitoneal injection of sodium pentobarbital (200 mg/kg) and the eyes were enucleated and then fixed with 4% paraformaldehyde for 2 h. The corneas and the lenses were removed under a stereo microscope (66 VisionTech, Suzhou, China). The retinas were carefully dissected and flat-mounted on a glass slide with anti-fluorescent quenching solution. The retinal vessels were viewed by fluorescence microscopy (Zeiss Axiophot, NY, USA). The avascular areas and neovascular tuft areas were assessed to evaluate the aspects of angiogenesis and treatment outcomes. The retinas were digested in 3% trypsin (Gibco, CA, USA) for 2–3 h at 37 °C to isolate the retinal vasculature. Subsequently, the retinal vasculature was stained with periodic acid–Schiff reagent and hematoxylin to investigate the capillary network.

Mice were euthanized by intraperitoneal injection of sodium pentobarbital (200 mg/kg) and eyes were enucleated and immediately placed in 4% paraformaldehyde for 24 h, dehydrated using a graded ethanol series, and embedded in paraffin. Sagittal sections were cut from each eye. Serial sections of 5 μm thickness were cut, deparaffinized in xylene and then hydrated. Hematoxylin and eosin staining was used to quantify neovascularization by counting the vascular nuclei that are extended anteriorly from the internal limiting membrane into the vitreous. The average number of neovascular nuclei per section was calculated in each group. The retinal vascular nuclei were counted using a masked protocol by light microscopy (magnification, 100–250). A terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay kit was used to test apoptosis in the retinal sections according to the manufacturer’s instructions (Roche, IN, USA). Images were captured under a Zeiss fluorescence microscope. The number of TUNEL-positive cells was calculated by Image-Pro plus 6.0. Images from random quadrant of posterior retinas were selected and masked to investigators for statistical analysis.

After hyperoxygen exposure from postnatal day 7 (P7) to P12, the neonatal mice developed OIR in the room air. OIR reached a peak severity at P17 during observation. The retinas were flat-mounted and examined at P17 to evaluate the vascular morphology in different groups. The superficial and deep vascular layers that are extended from the optic nerve to the periphery were observed in the retinas of healthy group, which showed the existence of fewer neovascular complexes. In contrast, multiple neovascular tufts and central avascular areas were observed in the retinas of OIR-blank and OIR-DMEM groups (red and yellow zone, Fig. 1b). The retinas from hyperoxygen-exposed mice treated with either BMSC or Conbercept showed fine radial branching pattern in the vessels and significantly reduced the neovascular tufts and the avascular areas. The avascular areas and the neovascular tufts in OIR-BMSC and OIR-CON groups were significantly smaller than those in OIR-blank and OIR-DMEM groups (P < 0.05), while no significant difference existed between OIR-BMSC group and OIR-CON group (Fig. 1c, d).

In the whole mount retinal digests, the endothelial nuclei and pericytes nuclei were elongated and spherical, respectively. The endothelial nuclei decreased in the capillaries of OIR-blank group and OIR-DMEM group, as observed that fusiform nuclei relatively increased in OIR-blank and OIR-DMEM groups. Acellular capillaries were calculated per mm². Compared with hyperoxygen exposure without treatment, both BMSC and Conbercept treatment significantly reduced the number of acellular capillaries in the retinas (P < 0.05, Fig. 2). Further comparison revealed no significant difference between BMSC and Conbercept treatment (P>0.05).

To examine whether BMSC was specifically recruited to the sites of retinal neovascularization, we detected the expression of red fluorescent protein (RFP) in the retinas of 3 days (P15), 5 days (P17) and 10 days (P22) after BMSC injection. The results showed few RFP-positive cells in the retinal vascular areas at P15. The cells were increased and dispersed around the retinal vascular sites at P17, indicating that the BMSC migrate closer to the vessels. Retinal flat-mount at P22 revealed that the cells participated in the formation of vascular structure (Fig. 3).

The severity of proliferation in the vitreous after injection is proportional to the concentration of BMSCs (Fig. 4). Therapeutic dose was thereby tapered to 5 × 10⁶ cells/ml. The extent of neovascularization was quantified in the retinal sections by counting the average number of vascular nuclei that extend beyond the inner limiting membrane into the vitreous body. There were no neovascular nuclei in the healthy group. Both OIR-blank group and OIR-DMEM group showed active neovascularization, and are characterized by lumen-like structure accompanied with nuclei located above the inner limiting membrane. In contrast, fewer blood vessels were observed beyond the inner limiting membrane in OIR-BMSC group and OIR-CON group (Fig. 5). Neovascular nuclei in OIR-BMSC group and OIR-CON group were significantly lower than those in OIR-blank group (P < 0.05). Intravitreal injection of either BMSC or Conbercept significantly attenuated the neovascularization response with no difference between BMSC and Conbercept (P>0.05).

We also performed the TUNEL assay to determine the apoptotic response in the retina during vascular tuft regression. BMSC injection significantly reduced the number of TUNEL-positive cells in contrast to OIR-blank group (P < 0.05), while Conbercept injection showed no significant difference (Fig. 6). Further comparison in regard to the apoptosis of retinal cell between BMSC and Conbercept treatment also revealed significant difference (P < 0.05). This suggested that the BMSC have anti-apoptotic effects in hypoxic environment.