Profiling of genes associated with the murine model of oxygen-induced retinopathy

Abstract

Purpose: To compare the clinical features and gene expression patterns of the physiologic development of retinal vessels and oxygen-induced retinopathy (OIR) in a mouse model, with the aim of identifying differential regulators of physiologic and pathological angiogenesis in the retina.

Methods: C57BL/6J mice were used. Seven-day-old pups were subjected to OIR induction following the standard protocols of entering a hyperoxic chamber on day 7 (P7) and returning to a normoxic condition (relative hypoxia) on day 12 (P12). The retinal vasculatures in the OIR model 24 h (P8-O) or 5 days (P12-O) after switching to the hyperoxic environment and 24 h (P13-O) after returning to normoxic conditions were evaluated with retinal flat mounts and compared with those of age-matched controls (i.e., P8-N, P12-N, P13-N). Gene expression profiling was performed using Phalanx Mouse Whole Genome OneArray microarrays. Normal 9-day-old mice were considered representative of physiologic angiogenesis and compared with 30-day-old mice. A bioinformatics analysis was performed on differentially expressed genes using various comparisons, and real-time reverse-transcription PCR was used to confirm the changes in the genes of interest.

Results: The sequential orders and patterns of vasculature development in normal mice and the OIR models were significantly different. In brief, in the early days (P1 to P7) for normal mice, retinal vessels grew from the optic disc into the non-vascularized retina in a radial fashion. In the hyperoxic stage of the OIR model, the main central retina became devoid of a vascular network, and when the mice returned to the normoxic room, the vessels grew from peripheral perfused areas toward the center of the retina, but the development of intermediate and deep layers of vasculature was significantly delayed. Gene profiling at three critical time points (P8, P12, and P13) showed that 162 probes were upregulated to ≥1.5-fold or downregulated to ≤0.67-fold at one or more time points in the OIR model compared to the controls. In the 45 upregulated genes for the P8-O/P8-N group, enriched genes were mainly related to cytoskeleton formation, whereas the 62 upregulated genes for P13-O/P13-N participated in various pathological processes. In the physiologic conditions on P9, however, 135 genes were upregulated compared with P30; the gap junction and Fc gamma R-mediated phagocytosis were the two main enriched pathways for these genes. Fifty-three probes, including vascular endothelium growth factor A, annexin A2, and endothelin 2, changed at P13-O but not at P9-N, and these changed genes might reflect the modulation of pathological neovascularization.

Conclusions: Angiogenesis in physiologic and pathological conditions is characterized by the differential presentation of vasculature and gene expression patterns. Investigation of those genes unique to the OIR model may help develop new strategies and therapies for intervening in retinal neovascularization.

Figure 1

Normal development of the vascular plexuses of C57BL/6J mouse retinas. Following FD-2000S perfusion, retinal whole mounts were examined to assess the distribution of vessels along the width and depth. Great caution was needed to differentiate the vessels at different levels; this differentiation was achieved by continuous slow adjustments of the microscope focus knob. The specified vasculatures (superficial, intermediate, and deep) in the pictures were highlighted by tracing the vessels in red using Adobe Photoshop software. A: Observation of retinal vessels as wholemount. From P0 to P7, only superficial retinal vessels were seen to originate from the optic disc, extending rapidly from the inner retina to the peripheral retina during the first week of postnatal development. B: Observation of vessels focusing on different layers of the retinal vasculature. From P8, a deep, intermediate vascular plexus started to develop. The deep vascular plexus began forming from vertical vessels diving down from the superficial vascular plexus, expanding from the inner retina into the peripheral retina. In addition, between P8 and P10 the intermediate vascular plexus had not begun to form. At P12, the intermediate vascular plexus was observed, and between P15 and P17, the superficial, intermediate, and deep vascular plexuses were seen all over the retina. After P21, the retinal vessel network was obviously remodeled, especially the intermediate vessel layer, and no obvious changes in the retinal vessels were observed after P25.

Figure 2

Central non-perfusion area and vasoconstriction under hyperoxic conditions in C57BL/6J mouse retinas. After the mice spent 24 h in a hyperoxic condition, a central area of vasoobliteration (VO) developed rapidly and increased in size until P11, extending further into the peripheral retina, while the size of the central vasoobliteration area decreased slightly from P12. From P8 to P12, only the superficial retinal vascular network was observed, but initiated vascular sprouts were sometimes observed at P12. The development and differentiation of the deep and intermediate vascular plexuses were obviously suppressed. The clear, sharp appearance of all vessels at the single superficial level in all panels also implied the lack of the two other layers of vasculature observed in Figure 1.

Figure 3

Neovascularization under relative hypoxic conditions in C57BL/6J mouse retinas. After the mice were relocated to room air for 24 h, the size of the vasoobliteration (VO) area decreased slightly, but the deep vascular plexus was observed in the peripheral area, in which area superficial vessels were preserved during hyperoxia. At P15, the size of the VO area continued to decrease, and the deep—but not the intermediate—vascular plexus was observed in the area where superficial vessels were preserved during hyperoxia. The retinal arteries and veins had a dilated and tortuous appearance, and vascular branch sprouts emerged from existing vessels. From P17 to P19, retinal neovascularization progressed rapidly, and numerous vascular sprouts appeared in the transition zone between the vascular and avascular retina. The deep and intermediate plexuses can be observed in almost the same area as the superficial plexus. From P21 to P23, the size of the VO area was small, and neovascularization began to regress. At P25, the VO area was fully revascularized in three layers, neovascularization had completely resolved, and the retinal vessels began to remodel.

Figure 4

Statistical analysis of microarray (SAM) plots of four comparison pairs. Three arrays were included in each group. The two parallel dashed lines are the cutoff threshold specified by the actual false discovery rate, and the total number of upregulated (red dots) and downregulated (green dots) genes were given for each plot.

Figure 5

Clustering of the regulated genes that were changed at P8-O/P8-N, P12-O/P12-N, or P13-O/P13-N during the oxygen-induced retinopathy model induction process. Three arrays were included in each group, and the average folds-over-age controls were used. The color bar stands for the log₂ values of the probe change folds, with red for upregulation and green for downregulation.