Effect of Light on Oxygen-Induced Retinopathy in the Mouse

Transcription

1 Effect of Light on Oxygen-Induced Retinopathy in the Mouse Eva Wesolowski and Lois E. H. Smith Purpose. To examine the effect, of light on retinal neovascularization and vasculogenesis in a reproducible and quantifiable model of oxygen-induced proliferative retinopathy in the mouse. Methods. C57B1/6J mice were reared in room air, 68% oxygen, or 75% oxygen and were exposed to darkness, low cyclical light ( lux), or high-intensity continuous light ( lux). The entire retinal vascular pattern was visualized in lluorcscein-dextran perfused flat-mount preparations. Proliferativc retinopathy was quantified by counting neovascular nuclei in 6 ^m cross-sections of whole eyes. Results. Light exposure did not exacerbate the proliferative relinopaihy that was seen after 68% oxygen exposure, which induced a meager proliierative response, nor after 75% oxygen exposure, which induced an exuberant proliferative response. In room air, retinas from all three illumination groups had normal vascular patterns. Conclusions. In this model of oxygen-induced retiiiopathy, under the conditions tested, light neither exacerbated the hyperoxia-induced ncovascularization nor affected normal retinal vascular development. Invest Ophihalmol Vis Sci. 1994;35: X he incidence of retinopathy of prematurity (ROP) is increasing, 1 and the factors influencing its pathogenesis need further exploration. Retinopathy of prematurity is correlated with the degree of retinal vascular immaturity and postconceptional age, 2 as well as with the level and duration of hyperoxia exposure. 3 However, many other factors, including the infant's clinical status, 4 may affect the incidence and severity of the disease. The effect of light exposure on ROP in newborns in nurseries has been debated for 50 years and has yet to be resolved. 5 Clinical studies have yielded conflicting results. There is a need for prospective, randomized, tightly controlled clinical trials, but such trials are difficult to perform. The use of animal models with oxygen-induced retinopathy, in which variables can be more easily controlled, may help to evaluate the contribution of light to the pathogenesis of ROP. From the Department of Ophthalmology. Children's Hospital. Boston, Massachusetts. Supported in part Iry Lions Fniiiultilion Inc. of Massachusetts. National Institutes oj Health grant F.YOH67O (LF.HS). and grants from Knights Templar Eye Foundation Inc. and Fight for Sight/National Society to f'revenl lilindness (F.W). Submitted for publication March 3, 1993; mused May 25, 1993; accepted June 17, Proprietary interest category: :V. lieprint requests: Lois F.. H. Smith, Ml), I'hl). Department of Ophthalmology. Children's Hospital, 300 Longiuood Avenue, Boston. MA We examined the effect of light on oxygen-induced proliferative retinopathy in a reproducible and quantifiable mouse model that is described in the preceding article by Smith el al in this issue. 6 In this study, we found that light neither aggravated the oxygen-induced proliferative retinopathy nor altered the normal retinal vascular development. MATERIALS AND METHODS Light Exposure This study adhered to the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. In total, 132 newborn mice (C57BL/6J) from 48 litters were randomly assigned to one of three groups until sacrifice on postnatal day 17 (PI 7) or 21 (P21): group 1, light deprivation from birth (P0); group 2, incandescent light, lux (lumens/meter 2 ), 12 hours on/12 hours oft", from P0; group 3, broad-spectrum fluorescent light, lux, 24 hours/day, from P2. Before P2, pups in group 3 received low cyclical light to avoid cannibalism by the dams. Illumination 112 Investigative Ophthalmology & Visual Science. January 1 <)<>!. Vol. *ir>. No. 1 Copyright Association lor Research in Vision and Ophthalmology

2 Effect of Light on Oxygen-Induced Retinopathy in the Mouse 113 was measured near the pups' eyes with a General Electric (Cleveland, OH) light meter, model # 213. Group ] was kept in a ventilated, light-excluding plywood chamber. When cages were inspected or changed twice a week, a red light (< 50 lux) lit the interior of the chamber for several seconds. Group 3 was exposed to four 20-watt lamps (Designer 830, Royal White, OSRAM Sylvania Inc., Danvers, MA) with spectral characteristics similar to the Cool White lamp often used in nurseries. Clear polycarbonate cages without bedding allowed light penetration from all sides. Clear acrylic shields separated the lamps from the cages. Heads of the group 3 pups were observed to be exposed to light >90% of the time. Ambient temperature was 23 ± 2 C to control for the possible influence of body temperature on the threshold for retinal light damage. 7 Hyperoxic Exposure Proliferative retinopathy was induced as described in the preceding article by Smith et al, 6 except that a clear acrylic hyperoxia chamber was used for maximum light penetration and oxygen levels were at 68% ± 2% or 75% ± 3%. Carbon dioxide, measured with a Bacharach Fyrite gas analyzer (Bacharach Inc., Pittsburgh, PA), was undetectable in the chamber. Control groups were exposed to room air from P0 until sacrifice. Preparation of Retinas Mice (PI7 or P21) were anesthetized and perfused with fluorescein-dextran or paraformaldehyde. The eyes were removed and processed for retinal flat mounts, paraffin cross- sections, and glial fibrillary acidic protein (GFAP) staining as described by Smith et al. 6 Quantification of Proliferative Retinopathy This method is described by Smith et al. 6 Six sections per eye were counted for neovascular nuclei. Groups were compared by analysis of variance. For groups that were different, adjusted Student's Mests were performed with the Instat program (GraphPad Software, San Diego, CA). To assess the fidelity of the counting method, double-masked determinations were made by two observers, and linear regression showed high correlation (r = 0.95, n = 60 cross-sections from 20 eyes, P < 0.001). Morphometric and Light Microscopic Analysis of Retinal Light Damage The retinal outer nuclear layer (ONL) thickness was measured with an ocular micrometer in 6 nm paraffin cross-sections under light microscopy (X400). Values were compared for groups 1 and 3 with the unpaired two-tailed Student's Mest. The total ONL area per section was calculated 8 and normalized for retinal length. Cross-sections were examined by light microscopy for signs of degenerative changes in photoreceptor cells with continuous light exposure. RESULTS Histology of the Retinal Vasculature After Room Air and Light Exposure In room air, the vascular pattern delineated in retinal flat mounts with fluorescein-dextran perfusion was the same in the absence (Fig. 1 A) or in the presence of intense continuous light (Fig. IB). A superficial radial branching plexus was seen connected to a deeper polygonal plexus. Low cyclical light exposure of lux (data not shown) produced the same pattern as exposure to darkness or lux. Similar patterns were observed in PI 7 mice (Fig. I A, 1 B) and P21 mice (data not shown). In room air, the retinal vascular pattern seen in paraffin cross-sections was the same without (Fig. 1C) and with light exposure (Fig. ID). Small PAS-positive blood vessels were found in the nerve fiber and in the inner and outer plexiform layers. No vascular nuclei extended from the nerve fiber layer into the vitreous. Histology of the Retinal Vasculature After Hyperoxia and Light Exposure After exposure to hyperoxia, there was virtually no fluorescein-dextran perfusion of the central capillary network in the posterior pole except for a few tortuous larger radial vessels. In the midperiphery of all the retinal quadrants (without predilection for any quadrant), there were clusters of abnormal new vessels or "tufts." The same pattern was seen in low cyclical light, lux (data not shown), darkness (Fig. 2A), and lux (Fig. 2B). There was less central hypoperfusion after 68% (data not shown) than after 75% oxygen. Similar patterns were observed in PI7 mice (Fig. 2, A and B) and P21 mice (data not shown). In paraffin cross-sections, the vascular pattern following hyperoxic exposure was the same without (Fig. 2C) and with light exposure (Fig. 2D). Tufts of vascular nuclei extended from the internal limiting membrane into the vitreous. After exposure to 68% oxygen (data not shown), there were fewer neovascular tufts and smaller intraretinal vessel profiles than after 75% oxygen (Fig. 2 C and D). Quantification of Proliferative Retinopathy Mice reared in room air had no proliferative retinopathy after exposure to either lux of light or to darkness (Fig. 3). There were 0.6 ± 0.9 (SD; SEM = 0.1; n = 9) nuclei in the light and 0.7 ± 1 (SD; SEM = 0.1; n = 10) nuclei in the dark (P > 0.05, dark versus intense light). Light exposure ( lux) did not

3 114 Investigative Ophthalmology & Visual Science, January 1994, Vol. 35, No. 1 V - ONL- FIGURE l. Comparison of retinas from dark and light-exposed mice reared in room air. Flatmount preparations of fluoresceiii-dextran perfused retinas from PI 7 mice exposed to (A) darkness or to (B) lux of continuous light. Paraffin cross-sections, 6 jum, stained with PAS and hematoxylin, from P21 mice exposed to (C) darkness or to (D) lux of continuous light. Original magnification X6 for A and B, XI00 for C and D. V, vitreous; ILM, internal limiting membrane; GCL, ganglion cell layer; IPL, inner plexiform layer; ONL, outer nuclear layer. Arrows indicate intraretinal blood vessels. affect the proliferative response to 68% oxygen (P > 0.05), even though 68% oxygen induced minimal proliferative retinopathy (Fig. 3). After exposure to 75% oxygen, a treatment that induced more robust proliferative retinopathy, there was less neovascularization with exposure to lux of light than with exposure to darkness (P < 0.001; Fig. 3). The SEMs (not shown) for the hyperoxia-treated groups were 4% to 11% of the mean value. Light, hyperoxia exposure, or both did not affect the number of intraretinal blood vessel profiles (P > 0.05; Fig. 4). There was no relationship between the number of new vessels and the weight or sex of the mice. All treatment groups weighed 15% to 36% less than did the mice exposed to darkness and room air. Intense light or hyperoxia may have caused maternal stress that interfered with nursing because the weight gain of the litters improved when the dams were rotated with surrogates every few days. Morphometric and Light Microscopic Analysis of Retinal Light Damage Mice reared under hyperoxic conditions and exposed to either darkness or to lux of light had the same ONL thickness (except for one zone in the lightexposed group in which the ONL was significantly thicker, P < 0.04; Fig. 5). The ONL area per retinal length was not significantly different ( ± mm 2 /mm in the dark versus ± mm 2 /mm in intense light, P> 0.05). ONL areas were normalized for retinal length to offset disparity in area due to differences in length among groups (21%). Within the groups, the measured lengths were within 10% of the mean length. No degenerative changes in photoreceptor cells were noted after exposure of the mice to lux of continuous light for 15 to 19 days. In PI 7 and P21 mice reared in room air, GFAP localized to the astrocytes in the inner layer of

4 Effect of Light on Oxygen-Induced Retinopathy in the Mouse 115 FIGURE 2. Comparison of retinas from dark- and light-exposed mice reared in 75% hyperoxia. Flat-mount preparations of fluorescein-dextran perfused retinas from PI 7 mice exposed to (A) darkness or to (B) lux of continuous light. Paraffin cross-sections, 6 /im, stained with PAS and hematoxylin, from P21 mice exposed to C darkness or to D lux of continuous light. Original magnification X6 for A and B, XI00 for C and D. V, vitreous; ILM, internal limiting membrane; GCL, ganglion cell layer; IPL, inner plexiform layer; ONL, outer nuclear layer; H, extravasated red blood cells in the vitreous. In A and B, arrows indicate neovascular tufts. In C and D, thin arrows indicate neovascular tufts, and intraretinal blood vessels are indicated by thick arrows. the retina regardless of light history (data not shown). After hyperoxic exposure, the increased GFAP staining in astrocyte and Muiler's cell processes was the same in the absence or the presence of lux of light (data not shown). DISCUSSION In 1943, Terry 9 first suggested that exposure of premature infants to light might be one of the factors causing ROP. Since then, a number of clinical trials have yielded conflicting reports on the influence of light on ROP. 10 " 14 Many of these studies were nonrandomized, lacked controls for pertinent variables, and had small numbers of patients. Such studies are difficult to interpret because gestational age, medical problems, genetic characteristics, and a myriad of other variables, including hyperoxic exposure 3 and intensity of light exposure, 5 differ from infant to infant and nursery to nursery. We examined the effect of light exposure on a reproducible, quantifiable, and genetically defined mouse model of oxygen-induced retinopathy, described by Smith et al. 6 No adverse effect of light on oxygen-induced proliferative retinopathy was seen under conditions producing either scant neovascularization (68% oxygen) or more exuberant neovascularization (75% oxygen). Interestingly, the retinas had more neovascular nuclei after 75% oxygen exposure in the dark than in the light. The reason for this is unclear. Perhaps the mouse retina, like cat 15 and human retinas, 16 has a higher metabolic demand in the

5 116 Investigative Ophthalmology & Visual Science, January 1994, Vol. 35, No n 60^ Dark Lux Previous studies in the rat 18 and kitten 19 revealed no effect of light on oxygen-induced retinopathy. However, these studies reported nonquantitative evaluations, and the degree of retinopathy was not defined. Therefore, we cannot determine if the retinopathy was mild enough to allow detection of small changes that might be induced by light. We attempted Z u 5 o 80- Dark Lux Room Air 68% 0. 75% 0, FIGURE 3. Quantification of hyperoxia-induced neovascularization in 6 juni paraffin cross-sections of P2I mouse eyes. * P < for exposure to 68% oxygen versus room air; ** P < for exposure to 75% oxygen versus room air; in 75% oxygen, P < for dark exposure versus intense light exposure. Error bars indicate SD, n = 9, 20 mice per group. o "a> S 40 c 0) 1 Room Air 75% 0. Nerve Fiber + Ganglion Cell Layers dark than in the light. The vasoprolifcrative response might reflect the retina's attempt to meet its metabolic needs. However, the new vessels extended away from the phot.orecept.ors into the vitreous, and the number of vascular profiles in the deeper retina was unchanged by light. We observed biologic variability in the neovascular response among mice within a group and among littennates from the same group, accounting for the high standard deviation. Despite the variability, however, a sufficient number of animals were assessed to minimize the error and to establish statistical significance. The large change in proliferative response seen with a small change in hyperoxia levels may reflect a "threshold" for this model and highlights the difficulty of designing and executing meaningful clinical trials in which variables are difficult to control. 3 A preliminary study with a small sample of animals appeared to show more neovascularization with light exposure than dark exposure. 17 In these initial studies, the data from mice exposed to 68% and to 75% oxygen were pooled. In this report, when the data obtained from 68% and 75% oxygen were assessed separately and the sample size was tripled, the difference between light and dark was insignificant. Additionally, by increasing the sample size, an initial finding of light-induced neovascularization in room air was shown to be invalid. B Dark Room Air 75% 0. Inner Plexiform Layer FIGURE 4. Mean number of intraretinal blood vessel profiles per 6 fiin paraffin cross-sections of P21 mouse eyes (A) in the nerve fiber and ganglion cell layers and (B) in the inner plexiform layer. In A and B, P > 0.05 for exposure to darkness versus intense light; P > 0.05 for exposure to room air versus hyperoxia. Error bars indicate SD, n = 7', 11 mice per group.

6 Effect of Light on Oxygen-Induced Retinopathy in the Mouse Inferior Superior Distance (mm) From Optic Nerve FIGURE 5. Distribution of retinal ONL thickness of P2 1 mice reared in darkness or in lux of light and exposed to (58% oxygen, lives were oriented in the superior-inferior axis. KITOI bars indicate SD, n = 6, mice per group (* P < 0.04). to address this problem by examining the effect of light on two degrees of retinopathy, one with mild and the other with more exuberant vasoproliferaiion. Despite the differences in experimental design and choice of animal species, our findings concur with the two original reports. The light intensity of lux used in this study is comparable to that experienced intermittently by premature infants from heat lamps and phototherapy lights in the nursery. 20 However, the mouse has fused lids until PI4, decreasing to some degree the light reaching the retina. In this study, the eyes were open for 48 hours before the onset of proliferative retinopathy at PI 6, and yet light did not aggravate the vascular proliferation. In premature infants, the closed eyelids act as a red-pass filter, 21 and, unlike neonatal rodents, infants can open their eyes for significant periods of time. 22 Therefore, the premature infant retina, despite protective eye shields during phototherapy, might receive a higher light dose and at shorter wavelengths than can be simulated in rodent models. We did not examine the effect of short cycles of intense light (for example, 2 hours on/2 hours off) on our model, though it has been reported that intermittent light exposure produces more damage to photoreceptors than does continuous light. 23 Our study did not show an effect of intense light exposure on three parameters that reflect light toxicity to the rodent retina: ONL thinning, 8 retinal histology, 24 and increased GFAP staining. 2 ' The ONL was slightly thicker in one zone of the light-exposed retinas, but the ONL area, related to the number of photoreceptors present, 8 was the same after exposure to light or darkness. This apparent ONL thickening might have been clue to differences in retinal length rather than to light exposure. Although no histopathologic changes were seen in the retina after light exposure, the damage may be subtle and may require ultrastructural analyses. Another study showed increased GFAP staining with intense light exposure in rats. 25 These results differ from ours, possibly because of the difference in animal species used, the duration and intensity of light exposure, or both. A link between light-induced retinal damage and proliferative retinopathy has not been established. It has been postulated that light of sufficient energy might generate oxygen radicals 26 that have also been implicated in the pathogenesis of ROP. 27 Intense light exposure is known to induce oxidation of rod outer segment membrane lipids. 28 However, there is circumstantial evidence that light-induced photoreceptor damage may be unrelated to ROP. We see oxygen-induced retinopathy in mice without evidence of photoreceptor injury after exposure to hyperoxia and darkness or intense light. Despite light-induced photoreceptor injury in rats, the same pattern of oxygeninduced retinopathy is seen with dark and intense light exposure. 18 Furthermore, in room air, intense light can damage photoreceptors in the rat, 18 pig, 29 monkey, 30 and in the human 31 without inducing retinal vasoprolife ration. Although light damage to photoreceptors may be unrelated to ROP, shorter wavelengths of light might be phototoxic to immature blood vessels. Blue light inhibits the growth of bovine aortic endothelial cells, RPF cells, and fibroblasts in vitro. 32 This effect is prevented by adding antioxidant enzymes, suggesting a photooxidative mechanism of light damage. However, it is unclear whether these cells contribute to the pathogenesis of ROP. Many phototherapy lamps emit nearly all their energy between 400 and 500 nm, possibly adding to oxidative stress on the retina. In our study, of the total energy emitted by the fluorescent lamps, only 5% was ultraviolet (<380 nm) and 1 5% was violet and blue light. Perhaps a greater intensity of blue light might have affected the degree of proliferative retinopathy in our model. Genetic factors may affect the susceptibility to retinal phototoxicity. Adult C57BI/6J mice are more resistant to light-induced retinal damage than other mouse strains. 33 This may explain, at least in part, why we saw no evidence of ONL thinning with intense light exposure. In addition, neonatal mice, like young rats, 34 might be more resistant to light damage than older animals. Infants also differ in their genetic susceptibility to ROP. Black people are less susceptible than is the general population to ROP, 35 and it has been postulated that ocular melanin may have a protective role.

7 118 Investigative Ophthalmology & Visual Science, January 1994, Vol. 35, No. 1 In conclusion, light exposure did not affect normal retinal vascular development or exacerbate oxygen-induced retinopathy in our mouse model under controlled experimental conditions. However, one should be cautious in extrapolating from animal models of oxygen-induced retinopathy to human ROP. At present, little is known about, the mechanisms of the developing retinal vasculature, and further basic and clinical research are needed to address this important issue. Key Words light, oxygen, mice, oxygen-induced retinopathy, retinopaihy of prematurity Acknowledgments The authors thank Richard Sullivan for technical assistance with the paraffin cross-sections and for printing the photomicrographs; Dr. Sandra Kostyk for helpful discussions; Angela McLcllan for assisting in quantifying the neovascularization; Erica Wheeler for performing the GFAP staining; and Diane Medeiros for measuring the CO 2 concentrations. References 1. Gibson DL, Sheps SB, Uh SH, Schechter MT, McCormick AQ. Rctinopathy of prematurity-induced blindness: Birth weight-specific survival and the new epidemic. Pediatrics ; 80: Phelps D. Retinopathy of prematurity. A' EnglJ Med. 1992; 326: Flynn JT, Bancalari E, Sim-Snydcr E et al. A cohort study of transcutaneous oxygen tension and the incidence and severity of retinopathy of prematurity. A' EnglJ Med ; 326: Lucey JF, Dangman B. A reexainination of the role of oxygen in retrolental fibroplasia. Pediatrics. 1984; 73: Robinson J, Fielder AR. Light and the immature visual system. Eye. 1992;6: (v Smith I.EH, Wesolowski E, Mcl.ellan A, el al. Oxygcninduced retinopathy in the mouse. Invest Ophthalmol Vis Sri. J 994;35: delint PJ, van Norren D, Toebosch AMVV. Effect of body temperature on threshold for retinal light damage. Invest. Ophlhalmol Vis Sci. 1992;33: Bush RA, Williams 'IT. The effect of unilateral optic nerve section on retinal light damage in rats. Exp Eye Res. 1991; 52: Terry TL. Fibroblastic overgrowth of persistent tunica vasculosa Icntis in premature infants: IV: Etiologic factors. Arch Ophthalmol. 1943; 29: Hepner WR, Krause AC, Davis ME. Retrolental fibroplasia and light. Pediatrics. 1949; 3: Locke JC, Reese AB. Retrolental fibroplasia: The negative role of light, mydriatics, and the ophthalmoscopic examination in its etiology. Arch Ophthalmol. 1952;48: Glass P, Avery GB, Subramanian KNS, Keys MP, Sostek AM, Friendly DS. Effect of bright light in the hospital nursery on the incidence of retinopathy of prematurity. A' EnglJ Med. 1985;313: ' 13. Hommura S, Usuki Y, Takei K et al. Ophthalmic care of very low birthweight infants, report 4: Clinical studies of the influence of light on the incidence of ROP. Nippon Ganka Gakkaii Zasshi. 1988;92: Ackerman B, Sherwonit E, Williams J. Reduced incidental light exposure: Effect on the development of retinopathy of prematurity in low birth weight infants. Pediatrics. 1989; 83: Yamamoto F, Borgula GA, Steinberg RH. Effects of light and darkness on ph outside rod photoreceptors in the cat retina. Exp. Eye Res. 1992;54: Fcke GT, Zukerman R, Green GJ, Weiter JJ. Response of human retinal blood flow to light and dark. Invest Ophthalmol, Vis Sci. 1983; 24: Wesolowski E, Smith LEH, D'Amato R. Effect of light on a murine model of retinopathy of prematurity. Invest Ophthalmol Vis Sci. 1992;33: Ricci B, Lepore D, lossa M, Santo A, D'Urso M, Maggiano N. Effect of light on oxygen-induced retinopathy in the rat model. Doc Ophthalmol. 1990;74: Krerner I, Levitt A, Goldberg G, Wilunsky E, Merlob P, Nissenkorn I. The effect of light on oxygen-induced vasoproliferative retinopathy in newborn kittens. Invest Ophthalmol Vis Sci. 1992; 33: Glass P. Light and the developing retina. Doc Ophthalmol. 1990; 74: Robinson J, Bayliss SC, Fielder AR. Transmission of light across the adult and neonatal eyelid in vivo. Vis Res. 1991; 31: Robinson J, Moseley MJ, Thompson JR, Fielder AR. Eyelid opening in prcterm neonates. Arch Dis Child. 1989; 64: Organisciak DT, Jiang YL, Wang HM, Pickford M, Blanks JC. Retinal light, damage in rats exposed to intermittent light: Comparison with continuous light exposure. Invest Ophlhalmol Vis Sci. 1989;30: Kuwabara T, Corn RA. Retinal damage by visible light. Arch Ophthalmol. 1968;79: Burns MS, Robles M. Miiller cell GFAP expression exhibits gradient from focus of photoreceptor light damage. Cnrr Eye Res. 1990;9: Ham WT Jr, Mueller HA, Ruffolo JJ Jr, et al. Basic mechanisms underlying the production of photochemical lesions in the mammalian retina. Curr Eye Res. 1984; 3: Warner BB, Wispe JR. Free radical-mediated diseases in pediatrics. Semin Perinatal. 1992; 16: Organisciak DT, Darrow RM, Jiang YL, Marak GE, Blanks JC. Protection by dimcthykhiourca against retinal light damage in rats. Invest Ophthalmol Vis Sci. 1992; 33:

8 Effect of Light on Oxygen-Induced Retinopathy in the Mouse Sisson TRC, Glauser SC, Glauser EM, Tasman W, Kuwabara T. Retinal changes produced by phototherapy. J Pediatr. 1970;77: Messner KH, Maisels J, Leure-du Pree AE. Phototoxicity to the newborn primate retina. Invest Ophthalmol VisSci. 1978; 17: Tso MOM. Experiments on visual cells by nature and man: In search of treatment for photoreceptor degeneration. Invest Ophthalmol Vis Sci. 1989; 30: Dorey CK, Delori FC, Akeo K. Growth of cultured RPE and endothelial cells is inhibited by blue light but not green or red light. Curr Eye Res. 1990; 9: Ginsberg HM, La Vail MM. Light-induced retinal degeneration in the mouse: Analysis of pigmentation mutants. In: LaVail MM, HollyfieldJG, Anderson RE, eds. Retinal Degeneration: Experimental and Clinical Studies. New York: Alan R Liss; 1985: O'Steen WK. Hormonal influences on retinal photodamage. In: Williams TP, Baker BN, eds. The Effects of Constant Light on Visual Processes. New York: Plenum Press;.1980: Palmer EA, Flynn JT, Hardy RJ, et al. Incidence and early course of retinopathy of prematurity. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Ophthalmology. 1991;98: