A direct method for imaging gradient levels of retinal hypoxia in a model of retinopathy of prematurity (ROP)

Abstract

Background: Retinal hypoxia may contribute to the development of preretinal neovascularization (NV) in patients with retinopathy of prematurity (ROP). We hypothesized that levels of NV may be associated with levels of retinal hypoxia. Imaging retinal hypoxia could be an important diagnostic tool to predict levels of disease severity in ROP patients. In this study, we have investigated a direct method for imaging gradient levels of retinal hypoxia using a model of ROP. We believe that this discovery will help understand the ROP pathogenesis in premature infants.

Methods: The rat 50/10 oxygen-induced retinopathy (OIR) model was generated by exposing the newly born Brown-Norway rat pups to a 24 hours alternate cycles of 50% and 10% oxygen for 14 days. HYPOX4 was used as a direct method for imaging gradient levels of retinal hypoxia at the peripheral avascular retina. A separate group of rat OIR pups were used to confirm gradient levels of retinal hypoxia using pimonidazole immunostaining. Gradient levels of retinal hypoxia was analyzed using ImageJ software from fluorescence intensities of HYPOX-4 and Pimonidazole immunostaining. We also confirmed the development of neovascularization in this model.

Results: Retinal hypoxia was observed in the peripheral avascular retinas in rat OIR. Based on fluorescence intensity measurements, retinal hypoxia was at minimal levels near the ciliary bodies. Retinal hypoxia was at its maximum levels towards the avascular-vascular transition zones. Interestingly, we observed hemiretinal avascular retina temporal to the optic nerve in this OIR model, similar to human ROP retinas. In the retinal cross-section, hypoxia was not detectable near the ora serrata in rat OIR may be due to oxygen delivery by the ciliary bodies. Both pimonidazole and HYPOX-4 showed similar patterns of retinal hypoxia at the peripheral avascular retina in this model. As expected, preretinal neovascularization was observed at the avascular-vascular transition zones arising from the existing retinal vascular structures in this OIR model in Brown-Norway rats.

Conclusions: In this study, we have characterized gradient levels of retinal hypoxia in the rat model of 50/10 OIR using a direct method from HYPOX-4 fluorescence. We observed minimal levels of retinal hypoxia near the ciliary bodies in this model and increased towards the avascular-vascular transition zones. In addition, we observed that the central vascularized retina remains gradient hypoxic in this model which could be detected using HYPOX-4. This study may clarify our understanding of persistent mild hypoxia in the vascularized central, mid-peripheral, and increased gradient levels of hypoxia at the avascular retina in the ROP patients.

Keywords: Fluorescence imaging; HYPOX-4; Molecular imaging; Optical imaging; ROP; Retinal hypoxia; Retinopathy of prematurity.

Fig. 1

Direct method for imaging gradient levels of hypoxia in rat 50/10 oxygen-induced retinopathy (OIR) model. (A) Chemical structure of HYPOX-4. This in vivo imaging probe contains hypoxia sensitive compound, pimonidazole conjugated via an amide linkage to dye compatible with clinically used fluorescence imaging equipment. HYPOX-4 is water soluble and has no residual toxicity to the retinal cells. (B) Graph of inspired oxygen treatment to develop 50/10 OIR model in Brown Norway rat pups. In this model, newborn rat pups experience alternating episodes of 50% oxygen for 24 h then 10% oxygen for 24 h for a total of 14 days. At postnatal day-14 (P14) pups are returned to room air. Retinal hypoxia was monitored at P14. This model recapitulates features of neovascularization and intensifies at P19, similar to infants with severe ROP

Fig. 2

In vivo fluorescence imaging of retinal hypoxia in 50/10 OIR pups (P15). HYPOX-4 was administered intraperitoneally to OIR pups 2 h after return to room air on P14. Imaging was performed 18 h post-injection of HYPOX-4. Oregon green dye that lacks the nitroimidazole-moiety, was used as dye control. (A-B) It was difficult to visualize the peripheral avascular retina in this 50/10 OIR retina using the Micron-IV imaging system. However, hypoxia was detected using HYPOX-4-dependent fluorescence within the peripheral avascular retina as shown in B (yellow arrow). Mild HYPOX-4 fluorescence was detectable in the central and middle retinal. (C-D) Oregon green dye was leaked outside the blood vessels and trapped in the vitreous and was unable to discriminate between hypoxic and normoxic tissues

Fig. 3

Imaging of retinal hypoxia using HYPOX-4 in 50/10 rat OIR. Isolectin B4 (IB4) was used to counter stain the vascular structures. (A, B) HYPOX-4 was localized largely in the avascular OIR retina. (C, D) Magnification of A and B respectively. (E, F) ImageJ software was used to analyze gradient levels of retinal hypoxia from HYPOX-4 fluorescence intensities across the 50/10 OIR retina as shown in F. Total of 12 eyes were analyzed for this study

Fig. 4

Retinal hypoxia was characterized in peripheral avascular retina in rat 50/10 OIR model. Pimonidazole hydrochloride was injected intraperitoneally at a dose of 60 mg/Kg on day-14, 2 h after removal to room air. Rat pups were sacrificed two hours after the pimonidazole injection (which is four hours after removal to room air). Retinas were dissected, flat-mounted and immunoassayed for pimonidazole-adducts. (A, B) Gradient levels of retinal hypoxia were analyzed from fluorescence intensity measurements. ImageJ software was used to analyze the fluorescence intensities across the peripheral avascular 50/10 OIR retina. (C, D) Intense retinal hypoxia was observed at the peripheral avascular retina in this model. In addition, mild hypoxia was also observed at the central retina as shown in D. Total of 12 eyes were analyzed for this study. Scale bar in image C-D, 100 μm

Fig. 5

Imaging retinal hypoxia in 50/10 OIR retinal cross sections. Pimonidazole immunostaining (HP-1) was used to visualize the retinal hypoxia and ICAM-2 was used to counterstain retinal vascular structures. (A-E) Retinal hypoxia was observed at different depth of the 50/10 OIR retina, mostly in the inner retina. (F-G) Retinal hypoxia was also monitored in 50/10 rat retinal cross sections. Retinal hypoxia was observed at different layers including retinal ganglion cell layer (RGC), inner plexiform layer (IPL) and inner nuclear layer (INL). DAPI was used to counterstain the nuclei to localize HP-1 fluorescence in retinal layers. Scale bar in image A 1000 μm, B-E 100 μm

Fig. 6

Characterization of preretinal neovascularization in Brown Norway 50/10 OIR model. (A) IB4 staining showing the neovascularization at the border of vascular/avascular area. Hypoxia may contribute to the development of neovascularization observed at P19. (B) Magnification view of A. (C) Preretinal neovascular tufts were localized in retinal cross section at P19. Scale bar in image A-B 1000 μm