Translational Science Review Pathophysiology and Mechanisms of Severe Retinopathy of Prematurity PDF

Translational Science Review The goal is to provide authoritative and cutting-edge reviews of topical type, however, if you have suggestions for topics, please contact Jayak- state-of-the-art basic research that is expected to have broad clinical impac...

Translational Science Review The goal is to provide authoritative and cutting-edge reviews of topical type, however, if you have suggestions for topics, please contact Jayak- state-of-the-art basic research that is expected to have broad clinical impact rishna Ambati ([email protected]) the Editor for this section. in the next few years. This is primarily a “by invitation only” submission Pathophysiology and Mechanisms of Severe Retinopathy of Prematurity M. Elizabeth Hartnett, MD Retinopathy of prematurity (ROP) affects only premature infants, but as premature births increase in many areas of the world, ROP has become a leading cause of childhood blindness. Blindness can occur from aberrant developmental angiogenesis that leads to fibrovascular retinal detachment. To treat severe ROP, it is important to study normal developmental angiogenesis and the stresses that activate pathologic signaling events and aberrant angiogenesis in ROP. Vascular endothelial growth factor (VEGF) signaling is important in both physiologic and pathologic developmental angiogenesis. Based on studies in animal models of oxygen-induced retinopathy (OIR), exogenous factors such as oxygen levels, oxidative stress, inflammation, and nutritional capacity have been linked to severe ROP through dysregulated signaling pathways involving hypoxia-inducible factors and angio- genic factors like VEGF, oxidative species, and neuroprotective growth factors to cause phases of ROP. This translational science review focuses on studies performed in animal models of OIR representative of human ROP and highlights several areas: mechanisms for aberrant growth of blood vessels into the vitreous rather than into the retina through over-activation of VEGF receptor 2 signaling, the importance of targeting different cells in the retina to inhibit aberrant angiogenesis and promote physiologic retinal vascular development, toxicity from broad and targeted inhibition of VEGF bioactivity, and the role of VEGF in neuroprotection in retinal development. Several future translational treatments are discussed, including considerations for targeted inhibition of VEGF signaling instead of broad intravitreal anti-VEGF treatment. Ophthalmology 2015;122:200-210 ª 2015 by the American Academy of Ophthalmology. Retinopathy of prematurity (ROP) was described in 1942 by in preterm infants before the development of stage 5 ROP, Terry1 as “retrolental fibroplasia,” which likely represents several changes in our understanding of ROP occurred. the most severe form of ROP, stage 5. Earlier stages of ROP First, the hypothesis of ROP has been revised in that there is were not well described because the Schepens/Pomerantzeff a delay in physiologic retinal vascular development and binocular indirect ophthalmoscope2 had not been adopted some hyperoxia-induced, vasoattenuation in phase 1, fol- universally to examine the peripheral retina. To understand lowed by vasoproliferation into the vitreous as intravitreal potential causes of ROP, investigators exposed newborn neovascularization (IVNV) in phase 2 (Fig 1).3 Second, it is animals, which vascularize their retinas postnatally, to recognized that the phenotype of ROP differs throughout the conditions similar to what human premature infants then world in association with resources for prenatal care and experienced. From early studies in animals and later a oxygen regulation. Preterm infants of older gestational ages clinical trial in human infants by Arnall Patz3, it became and larger birth weights than those screened in the United recognized that high oxygen at birth damaged fragile, newly States now are demonstrating severe ROP in some regions formed retinal capillaries, causing “vaso-obliteration.” After with insufficient nutrition and neonatal or prenatal re- animals were removed from supplemental oxygen to sources and care, and where high, unregulated oxygen ambient air, “vasoproliferation” occurred at junctions of is used.4,5 Finally, heritable causes are recognized as vascular and avascular retina. Thus, the 2-phase hypothesis important,6 but candidate gene studies often have been was developed, almost 30 years before the classification small and have not replicated findings potentially because of human ROP into zones and stages. With advances in of phenotypic variability. neonatal care, including the ability to monitor and regulate The International Classification of ROP describes oxygen, and in funduscopic imaging of the peripheral retina stages and zones of ROP severity.7 Because human retinal 200  2015 by the American Academy of Ophthalmology http://dx.doi.org/10.1016/j.ophtha.2014.07.050 Published by Elsevier Inc. ISSN 0161-6420/14  Hartnett
Translational Science Review Figure 1. Human retinopathy of prematurity (ROP) classified by zone, stage, and the presence of plus disease. To facilitate comparing phases of ROP with experimental studies, ROP can be divided into Early Phase ROP, which comprises delayed physiologic retinal vascular development, and stages 1 and 2 ROP; Vascular Phase ROP, which comprises stage 3 ROP and, in severe ROP, plus disease; and Fibrovascular Phase ROP, which comprises stages 4 or 5 ROP with partial or total retinal detachment, respectively. Drawing by James Gilman, CRA, FOPS. vasculature is not complete until term birth, an infant born growth factor (VEGF). Ensuing endothelial cells proliferate prematurely initially has incomplete vascular coverage of and migrate toward the gradient of VEGF and thereby extend the retina. The zones of ROP define the area of retina the inner vascular plexus toward the ora serrata. Angiogenesis covered by physiologic retinal vascularization. The stages also is important in the development of the deep retinal plexi. often progress sequentially and describe the severity of Besides astrocytes, glia, like Müller cells, and neurons, such disease. Stages 1 and 2 represent early ROP, and stage 3 as ganglion cells, are also important.13e15 The process is represents the vascular phase in which IVNV occurs (Fig 1). complex and requires interactions between different cell types Stages 4 and 5 ROP represent the fibrovascular phase with and regulation of signaling pathways through several family retinal detachment.8 Laser treatment for severe ROP, now members of VEGF and other pathways, including delta-like 4/ described as type 1 ROP in the Early Treatment for Reti- notch and robo/slit, as examples, which regulate interactions nopathy of Prematurity Study,9 can reduce the risk of a poor between the sensing, endothelial tip cells and the proliferating outcome in approximately 90% of eyes. In some infants, stalk cells.16 Of all the factors involved in physiologic retinal aggressive posterior ROP occurs, in which stage 3 and se- vascular development, it is clear that VEGF is essential. vere plus disease developsdwithout prior stages 1 or 2din zone 1 or posterior zone 2. It is important to consider human retinal vascular devel- Animal Models to Study Retinopathy of opment when studying what goes awry in ROP. Because of Prematurity the difficulty in obtaining intact human preterm infant eyes, studies on human retinal vascular development have been It is not safe to experiment on human preterm infant eyes limited, but reports indicate that the initial retinal vasculature because of risks of bleeding and retinal detachment. develops through vasculogenesis in the posterior pole from Therefore, models of oxygen-induced retinopathy (OIR) precursor cells that migrate out of the deep retina and into are performed in animals that vascularize their retinas inner layers.10,11 At approximately 15 weeks of gestation11 postnatally to study disease mechanisms. Most OIR until at least 22 weeks of gestation, these precursors become models recreate only some aspects of human ROP. All angioblasts and form the inner vascular plexus that extends to models have limitations because they use newborn, instead approximately zone 1. After 22 weeks of gestation, when it is of premature, animals. Newborn animals are healthy and difficult to obtain fetal human tissue, the ensuing development do not have the comorbidities of human preterm infants, of the vascular plexi is based on studies carried out in other such as necrotizing enterocolitis, sepsis, bronchopulmo- species and believed to occur through budding angiogenesis, nary dysplasia, shunting of oxygenated and deoxygenated that is, the proliferation and growth of blood vessels from blood, and immature lung development. Animals experi- existing blood vessels. In several species, astrocytes sense a ence much higher arterial oxygen levels when given physiologic hypoxia12 and upregulate vascular endothelial similar inspired oxygen levels as premature infants with 201  Ophthalmology Volume 122, Number 1, January 2015 these comorbidities. Neonatologists strive to avoid high with 75% inspired oxygen, oxygen levels that are avoided oxygen in the perinatal period, but most animal models use in preterm infants. Day 7 pups placed into 75% constant high oxygen, making them less representative of human inspired oxygen experience vaso-obliteration of newly ROP. These are important considerations when choosing formed capillaries in the central retina and then are placed an OIR model to study a scientific question. The 2 most into room air and form intravitreal vascular buds at the commonly used OIR models are in the mouse and rat. Also junctions of vascular and avascular retina (Fig 2). Thus, important is the beagle OIR model. None of these species the model is not similar to the phases of human ROP in is premature; rather, they complete retinal vascular devel- that complete inner plexus vascularization has occurred opment after term birth. already when the pups are placed into high oxygen, unlike the preterm infant whose retina is incompletely vascular- Mouse Oxygen-Induced Retinopathy Model ized. Nonetheless, several signaling pathways important in human ROP have been identified using the mouse model. The use of transgenic mice makes the mouse OIR model The model also may reflect aspects of ROP in the United most helpful to study molecular mechanisms of high States and the United Kingdom in the 1940s or in places oxygen-induced vascular loss followed by regrowth of currently that lack resources to regulate oxygen and pro- vessels either into the retina or into the vitreous during vide prenatal and perinatal care.5 relative hypoxia.17 However, there are a few ways in which the model does not represent human ROP. First, Rat Oxygen-Induced Retinopathy Model oxygen levels used do not recreate what human preterm infants experience. The arterial oxygen (PaO2) in healthy The most representative model of human ROP in the era of newborn mice can approach very high levels (500 mmHg) oxygen regulation is the rat OIR model, which has aspects Figure 2. Models of mouse and rat oxygen-induced retinopathy (OIR) showing oxygen profiles, phases 1 and 2 OIR, and retinal flat mounts stained with lectin to visualize the vasculature. ROP ¼ retinopathy of prematurity. 202  Hartnett
Translational Science Review of both vasoattenuation centrally and delayed physiologic definitions.3,8 Phase 1 in the rat OIR reflects the Early Phase retinal vascularization peripherally18 (Fig 2). Shortly after of human ROP, that is, delayed physiologic retinal vascular birth, pups and dams are placed into a controlled oxygen development. Phase 2 in rat and mouse models of OIR environment that changes inspired oxygen levels from 50% reflect vasoproliferative IVNV, similar to the Vascular to 10% every 24 hours for 14 days. This oxygen profile Phase of human stage 3 ROP with plus disease. However, recreates transcutaneous arterial oxygen extremes similar to human ROP also has a third Fibrovascular Phase, in which those in a human preterm infant with severe ROP.19 The retinal detachment occurs in stages 4 and 5 ROP, and few notion of oxygen fluctuations, including intermittent epi- animal models demonstrate this form of human ROP. sodes of hypoxia, has been associated with ROP.20 How- However, the beagle OIR model shares some features seen ever, the duration of the fluctuations in oxygenation in the in stage 4 ROP, such as retinal folding and dragging of rat model is 24 hours, whereas in the human preterm infant, vessels.23 minute-to-minute fluctuations occur. The rat pups experi- For clarity, the phases of OIR are described by the animal ence extrauterine growth restriction, a factor associated with and phase. Phase 1 in the mouse is vaso-obliteration and in severe ROP. The appearance of first delayed physiologic the rat is delayed physiologic retinal vascular development, retinal vascular development followed by IVNV at the and phase 2 is IVNV in both models. The 3 phases of human junction of vascular and avascular retina at day 18 is similar ROP are described as Early (delayed physiologic retinal in appearance to type 1 severe ROP.9 Thus, the rat OIR vascular development and some vasoattenuation),3 Vascular model closely represents human preterm infants with severe (IVNV) and Fibrovascular (retinal detachment; Fig 1). ROP. The study of molecular mechanisms or potential treatments had been limited to pharmacologic manipulations in the rat, because the availability of transgenic rats is Pathophysiology of Human Severe limited. Now, other techniques have been developed to Retinopathy of Prematurity permit study of molecular mechanisms in the rat. One example is the use of gene therapy to introduce short-hairpin Most early investigations sought to understand causes of the RNAs (shRNAs) or genetic mutations to silence or knock vascular phase of human ROP by studying phase 2 OIR out certain genes. Different viruses or viral vectors are used with IVNV, but several investigators24,25 strove to under- in gene therapy and include adeno-associated virus, stand the early phase of human ROP by studying phase 1 adenovirus, or lentivirus, as examples. Several valuable OIR. The thinking was that in facilitating vascularization of aspects of a lentiviral vector are that it is not infectious (is avascular retina, there would be less hypoxia-induced self-inactivating) and has a large cargo-carrying capacity. IVNV, and this line of thought aligned with clinical obser- The lentiviral vector cassette contains the only genetic cargo vations that infants with zone 1 ROP, compared with zone 2 delivered into the genome to allow stable transgene ROP, were at greater risk of severe ROP developing and expression. Using lentivirus, cell-specific promoters have having poor outcomes.26 been linked with shRNAs to target certain cell types in the Several exogenous stresses implicated in ROPdsuch as retina and to knock-down specific gene products in those fluctuations in oxygenation, oxidative stress, nutritional cells only. This has been a novel and useful technique to factors, and poor infant growthdactivate inflammatory, determine the effects of angiogenic signaling in pathologic oxidative, and hypoxic signaling pathways.3 Studies of and physiologic retinal angiogenesis from knockdown of phase 2 OIR focused on induced angiogenic factors from genes in specific retinal cells and to assess safety on trans- these activated signaling pathways. As with most biologic duced and other cells within the retina.14,21,22 In addition, processes, it has become recognized that interactions and techniques to delete genes have been developed that will crosstalk exist within different signaling pathways. permit gene knockout in many species besides mice to study molecular events in various models. Hypoxia-Inducible Factors Beagle Oxygen-Induced Retinopathy Model Hypoxia-inducible factors (HIFs) are transcription factors that bind DNA at the hypoxia-responsive element and The beagle OIR model23 is especially useful to translate drug enable transcription of a number of downstream genes that doses from the puppy eye to the human preterm infant eye are angiogenic, including VEGF, angiopoietins, and eryth- because of greater similarity in size between eyes of the puppy ropoietin, as examples. The classic mechanism involves and preterm infant than between those of the preterm infant hypoxia, which occurs in avascular retina as soon as a and newborn rodent. The newborn beagle retina initially newborn pup is removed from supplemental oxygen and vascularizes through a process of vasculogenesis that is fol- placed in ambient air. Hypoxia prevents HIFs from degra- lowed by angiogenesis similar to what occurs in premature dation by prolyl hydroxylases and thus allows them to human infant retinas. However, the model uses high oxygen translocate to the nucleus to cause angiogenic gene tran- to cause OIR, which differs from the pathogenesis of ROP in scription.3 Hypoxia-inducible factors also can be stabilized most premature infants. In the beagle model, newborn post- through oxidative compounds or inflammatory cytokines, natal day 1 pups are placed into 100% oxygen for 4 days and mediated through NFkB, which can lead to downstream then into ambient air to recreate the phases of OIR.23 angiogenic effector compounds, including succinate or When comparing the phases of OIR (Fig 2) with what RTP801. Using mouse and rat OIR models, investigators occurs in human ROP (Fig 1), it is helpful to clarify studied prolyl hydroxylase inhibitors to stabilize HIF and 203  Ophthalmology Volume 122, Number 1, January 2015 promote physiologic retinal vascular development in phase angiogenic and angiostatic effects from activation of STAT3 1 OIR models.27 Others found that administration of HIF- in the two cell types counter one another. Broad inhibition of induced growth factors, including erythropoietin or VEGF, STAT3 with an intravitreal agent, then, does not seem to reduced avascular retina in phase 1 OIR.3 However, a po- have an effect on phase 2 IVNV. These studies highlight the tential concern with these strategies is that early and complexity of oxidative signaling pathways and subsequent vascular phases of human ROP may not be sufficiently biologic events, including angiogenesis, and also point to distinct in the individual preterm infant to determine a safe the importance in identifying signaling events in specific cells window of time to administer an angiogenic agonist to treat (Fig 3). early ROP without causing vascular ROP. Extrauterine Growth Restriction and Nutritional Oxidative Stress Effects Oxidative stress has been proposed in ROP because of the The roles of birth weight and postnatal growth in preterm susceptibility of the phospholipid-rich retina to reactive infants have been recognized as important factors associated oxygen species that can be generated in high or low oxygen. with ROP and in animal models of OIR.32 In human preterm Repeated oxygen fluctuations in the rat OIR model also lead infants, low insulin-like growth factor-1 (IGF-1) was asso- to the generation of oxidative compounds. Although use of ciated with extrauterine growth restriction, poor retinal antioxidants, such as superoxide dismutase in liposomes25 vascular growth, and later vasoproliferation.33 Omega-3 or apocynin,24 reduced avascular retina in phase 1 of the rat fatty acids also were found to be important in reducing OIR model, these substances did not reduce IVNV in phase vasoproliferation in the mouse OIR in part by inhibiting 2 in the rat OIR model. In addition, human clinical trials tumor necrosis factor a and facilitating neuroprotection.32 that tested n-acetyl cysteine, vitamin E, or lutein have not inhibited severe ROP successfully or safely to date.15,28 Genetic Variation These findings may reflect the complexities in oxidative Besides environmental factors, 70% of the variance in ROP signaling and that reactive oxygen species can be damaging was reported secondary to heritable factors in a study of or beneficial to the retina. Besides direct interaction with the monozygotic and dizygotic preterm twins.6 Small candidate phospholipids in retina, some species act as signaling ef- gene studies found several gene variants, including those in fectors that promote physiologic or pathologic events. Nitric the wnt pathway (FZD4, LRP5, NDP). Variants of genes in oxide can be activated by nitric oxide synthetases, including the wnt pathway cause familial exudative vitreoretinopathy, endothelial nitric oxide synthetase, and can act as an which shares features of ROP but occurs in full-term infants. endothelial relaxing agent in blood vessels, but in high Other investigators found variants in EPAS1 that transcribes oxygen, nitric oxide can form nitro-oxidative forms like erythropoietin, SOD that transcribes the antioxidant enzyme peroxynitrite that lead to microvascular degeneration in superoxide dismutase, or VEGF. However, most studies phase 1 OIR. Oxidative stress can activate VEGF receptor 2 involved small samples of infants with broad ranges in (VEGFR2) signaling that is needed in physiologic angio- birth weights and gestational ages and did not control for genesis or overactivate VEGFR2 signaling in phase 2 OIR. multiple comparisons. In addition, interactions between In the immunocompromised preterm infant, nicotinamide genes and their function may be affected by other factors adenine dinucleotide phosphate (NADPH) oxidase can that are linked with ROP. A recent study performed in 817 generate reactive oxygen species that defend against invading samples from extremely lowebirthweight infants and that micro-organisms. However, NADPH oxidaseegenerated controlled for multiple comparisons found variants in the reactive oxygen species also can cause endothelial cell gene encoding brain-derived neurotrophic factor (BDNF) injury and avascular retina in phase 1 OIR through activa- associated with severe ROP.34 tion of isoforms NOX1 or NOX2 or can increase vaso- proliferation in phase 2 OIR through activation of isoforms Vascular Endothelial Growth Factor Signaling NOX1 or NOX2 or through NOX4-induced activation of Pathway the transcription factor, signal transducer and activator of transcription 3 (STAT3), in endothelial cells.28,29 In contrast, Many laboratories have studied ROP using the mouse OIR activation of STAT3 in Müller cells inhibits the expression model. This review focuses on the effects of stresses similar of erythropoietin and thus reduces angiogenesis in phase 1 to what human preterm infants experience in the early and OIR.30 Although exogenous erythropoietin improves phys- vascular phases of ROP and, therefore, reports mainly on iologic retinal vascularization in phase 1 OIR, it does not studies that used the rat OIR model adapted to study mo- reduce phase 2 IVNV in the rat OIR model. Following along lecular mechanisms. with this line of evidence, an intravitreal injection of Vascular endothelial growth factor is important in physi- a STAT3 inhibitor reduces only phase 2 IVNV compared ologic retinal vascular development and pathologic angio- with vehicle under conditions of supplemental oxygen. In genesis, and both processes occur in the preterm infant retina. the rat OIR model with supplemental oxygen, VEGF is not Therefore, it is important first to determine the differences in increased, and Müller cell STAT3, therefore, is not activated, VEGF signaling that lead to IVNV instead of physiologic but endothelial cell STAT3 is activated to mediate IVNV.31 In retinal vascular development. It is helpful to review aspects the rat OIR model without supplemental oxygen, endothelial of VEGF signaling. Vascular endothelial growth factor has cell and Müller cell STAT3 proteins are activated, and the different family members, but much of the work on 204  Hartnett
Translational Science Review Figure 3. Diagram showing activated signaling pathways leading to the phases of human retinopathy of prematurity based on experimental methods in the rat oxygen- induced retinopathy (OIR) model. Overactivation of vascular endothelial growth factor (VEGF) receptor 2 can cause both phases of OIR and differential effects from STAT3 signaling based on the cell activated. Also, targeted inhibition of VEGF in Müller cells can cause cell death and thinning of the outer nuclear layer. EC ¼ endothelial cell; EPOR ¼ erythropoietin receptor; NAPDH ¼ nicotinamide adenine dinucleotide phosphate; PRVD ¼ physiologic retinal vascular development. angiogenesis has involved VEGFA, henceforth referred to as than into the retina. Evidence was not found to support this VEGF for this review. Vascular endothelial growth factor prediction. Vascular endothelial growth factor measured in activates different receptors. Vascular endothelial growth the vitreous was more than 10-fold lower than in the retina factor receptor 2 (VEGFR2) is activated in pathologic at the time point when IVNV occurred in the rat OIR model. angiogenesis. Vascular endothelial growth factor receptor 1 A limitation may have been the inability to measure local (VEGFR1) also can be angiogenic, but in development binds vitreous VEGF overlying IVNV compared with retinal VEGF VEGF with higher affinity than does VEGFR2 and can act as anterior to developing intraretinal vascularization. a decoy, preventing binding with VEGFR2. Vascular endo- Fluctuations in oxygenation are associated with ROP. thelial growth factor receptor 3 is important in lymphangio- Therefore, another study was carried out to determine genesis and to some extent in the regulation of angiogenesis. whether repeated oxygen fluctuations, compared with hyp- Vascular endothelial growth factor has different mRNA oxia alone, altered the expression of VEGF splice variants to splice variants. Some of the translated forms are secreted, lead to different biologic outcomes. In the rat OIR model, and therefore have access to the vitreous, whereas others are repeated oxygen fluctuations increased the expression of cell-associated proteins that impact signaling locally and retinal VEGF164, an analog to human VEGF165, whereas create a gradient for intraretinal angiogenesis.16 hypoxia increased VEGF120.35 This finding suggested that A critical question in ROP, which involves both physi- VEGF164 was more associated with pathologic features in ologic retinal vascular development and aberrant IVNV, phase 2 OIR. Another study reported that increased is why does hypoxic retina in phase 1 ROP activate an expression levels of VEGF164 and VEGFR2 were associated angiogenic signaling pathway that lead to blood vessel temporally with pathologic features in both phases 1 and 2 growth into the vitreous as IVNV rather than into the of the rat OIR model, whereas the other VEGF splice var- avascular retina to provide physiologic intraretinal vascular iants (VEGF120 and VEGF188) and VEGFR1 were associ- support? Several studies investigated why blood vessels ated with the control situation, physiologic retinal vascular grow into the vitreous rather than into the retina in phase 2 development under ambient oxygen conditions. These studies OIR.8 One possibility examined was whether VEGF con- support the thinking that VEGF164 and VEGR2 may have centration was greater in the vitreous than in the retina, roles in the features of both phases 1 and 2 OIR and poten- thereby drawing vascular growth toward the vitreous rather tially the early and vascular phases of human ROP.8 205  Ophthalmology Volume 122, Number 1, January 2015 To study VEGF164eVEGFR2 signaling on OIR phases, retina in a pattern similar to IVNV. Inhibition of VEGFR2 different approaches were used to inhibit VEGFR2 signaling: signaling then permits ordered, intraretinal vascularization a neutralizing antibody to rat VEGF164 or a VEGFR2 kinase (Fig 4). The investigators also tested an intravitreal neutral- inhibitor.36 Because VEGF is an angiogenic factor, inhibition izing antibody to VEGF164 against a control immunoglobulin of VEGFR2 activation was predicted to reduce not only G antibody and found the anti-VEGF164 antibody reduced IVNV, but also physiologic retinal vascular development and, tortuosity of arterioles compared with control antibody.38 therefore, to cause persistent avascular retina. At certain This study provides evidence that VEGF signaling also plays doses, each intervention reduced phase 2 IVNV, but sur- a role in arteriolar tortuosity, as seen in human plus disease. prisingly no dose inhibited physiologic retinal vascular Subsequently, the Efficacy of Intravitreal Bevacizumab development. These findings suggested that overactivation of Treatment for Stage 3þ ROP study found that inhibition VEGFR2 signaling may both inhibit physiologic retinal of VEGF with an antibody reduced IVNV and permitted vascular development and cause IVNV. Studying the VEGF ongoing physiologic retinal vascular development in some signaling pathway in vivo is problematic, because a single infants,39 providing clinical evidence for the experimental allele knockout of VEGF or one of its splice variants or re- findings that regulation of VEGF signaling orders disoriented ceptors is lethal. The investigators, therefore, used an em- developmental angiogenesis and may have a role in treating bryonic stem cell model in which a knockout of VEGFR1 both phases 1 and 2 ROP. However, other infants treated with (flt1) caused VEGF to bind and overactivate VEGFR2 and bevacizumab demonstrated persistent avascular retina and thus increased angiogenesis. Compared with control, over- later IVNV, sometimes at 60 weeks postgestational age,40 activation of VEGFR2 disordered angiogenesis and caused a suggesting that individual doses of an anti-VEGF agent and pattern of growth similar to IVNV.37 Physiologic vasculari- other factors, are involved in physiologic retinal vascular zation was restored with a transgene of VEGFR1 containing a development in some phenotypes of ROP, and that it is promoter (CD31) specific to endothelial cells. The endothelial important to study long-term effects from VEGF inhibition. VEGFR1 thus trapped excessive VEGF and reduced its Therefore, a study was performed to test a later time binding and activation of VEGFR2. This work demonstrated point in the rat OIR model. An intravitreal neutralizing that overactivated VEGFR2 in endothelial cells caused antibody to rat VEGF164 at a dose that inhibited phase 2 aberrant angiogenesis in vitro. IVNV was compared with an isotype goat immunoglobulin Investigators then determined whether VEGFR2 activa- G control. The anti-VEGF164 antibody led to later recurrent tion affected IVNV in the rat OIR model. Retinal flat mounts IVNV in association with increased expression of angio- from the rat OIR model were colabeled with lectin to visu- genic compounds in addition to and independent of VEGF, alize the vasculature, and an antiphosphohistone H3 label was including erythropoietin.41 This study suggests that broad used to identify mitoses of dividing vascular cells in anaphase inhibition of VEGF may lead to rebound angiogenic effects. (Fig 4). Two lines were drawn onto each mitotic figure in In addition, the anti-VEGF164 intravitreal antibody inhibited imaged retinal flat mounts. One line was between each pair of pup weight gain, raising systemic safety concerns from antiphosphohistone, H3-labeled chromosomes at the cleav- broad intravitreal inhibition of VEGF bioactivity. age plane set up by the dividing vascular cells. The other line To knock down VEGF specifically in Müller cells that was drawn along the long axis of the developing vessel. The had been shown to over-express it, a lentivector gene ther- angles between the 2 lines of all mitotic figures were apy approach was used to introduce a cell-specific promoter measured. Angles at 90 predicted elongation of developing and an shRNA to VEGFA in the rat OIR model.21,42 vessels, whereas those 180 apart predicted widening of the Compared with a control lentivector, the shRNA to Müller vessels. Mitotic cleavage planes having multiple different celleVEGFA reduced retinal VEGF to levels in retinas of angles with the long axes of vessels predicted disordered pups of the same developmental ages raised in room air and angiogenesis. Two approaches to inhibit VEGF then were inhibited VEGFR2 signaling in endothelial cells. Targeted compared: a neutralizing antibody to rat VEGF164 and a gene knockdown of VEGFA was compared with its control len- therapy approach using a lentivector specifically to target and tivector and then with the experimental approach using an knock down overexpressed VEGFA in Müller cells.21,22 intravitreal antibody to VEGF164 compared with its intra- (Knockdown in Müller cells was chosen because VEGF vitreal immunoglobulin G control. Both the VEGFA lenti- splice variant expression levels had been localized to the vector and VEGF164 antibody caused the same fold reduction inner nuclear layer, corresponding to the location of Müller in IVNV areas compared with respective controls, but did not cells,21 at time points preceding the development of phase 1 affect the extent of physiologic retinal vascular development and 2 OIR in the rat.8) With each method to reduce VEGF measured as vascularized to total retinal areas. However, the bioactivity, doses were chosen that reduced VEGFR2 intravitreal anti-VEGF164 antibody reduced capillary den- signaling and phase 2 IVNV compared with respective con- sities in the inner and deep retinal plexi, whereas the VEGFA trols, but did not reduce physiologic retinal vascular devel- lentivector did not.22 Thus, targeted knockdown of VEGFA in opment. In retinas treated with the targeted knockdown of Müller cells following repeated fluctuations in oxygenation Müller cell VEGFA22 or the intravitreal VEGF164 antibody,37 seemed safer than broad intravitreal anti-VEGF164 antibody. cleavage angles predicted more ordered angiogenesis than in These studies support a line of thinking that intravitreal anti- each respective control condition. VEGF164 antibody reduces capillary support in the retinal Together, these studies support the hypothesis that over- plexi and leads to activation of angiogenic pathways that activation of VEGFR2 disorders dividing endothelial cells, cause recurrent IVNV. The studies also support a cell-tar- potentially allowing them to grow outside the plane of the geted approach to inhibit VEGF in ROP. 206  Hartnett
Translational Science Review Figure 4. A, Drawing depicting intravitreal neovascularization (IVNV) growing aberrantly into the vitreous instead of into the retina. B, Drawing depicting endothelial cells growing into the vitreous as IVNV rather than into the retina as intraretinal blood vessels. The angle between the cleavage plane of dividing daughter cells and the long axis of the vessel predicts whether the vessel will be elongated or widened. Arrows point to cleavage planes (i.e., line between mitotic figures) at 180 degree predicting vessel widening (left) and 90 degrees predicting vessel elongation (right). One line of evidence shows that overactivated vascular endothelial growth factor receptor 2 signaling disorders divisions of endothelial cells, permits their access to the vitreous cavity, and diverts them from growing into the retina. Drawings by James Gilman, CRA, FOPS. 207  Ophthalmology Volume 122, Number 1, January 2015 Because VEGF is neuroprotective and VEGF164 was asso- reduced systemic VEGF for at least 2 weeks.40 The reduction ciated with pathologic features in the rat OIR model,8 in- of systemic VEGF may have implications in developing or- vestigators used the lentivector gene therapy approach gans, including lung, brain, and kidney. A preterm infant’s to knockdown Müller cell VEGFA or the splice variant, vitreous volume is approximately 1 ml and blood volume is VEGF164, compared with a control lentivector containing an approximately 120 ml, whereas an adult’s is approximately 4 shRNA to the nonmammalian gene luciferase.14 Pup weight ml and blood volume is more than 5000 ml. Therefore, there is gain was not adversely affected by either experimental or con- less dilution of an intravitreal drug that enters the preterm trol condition. Both the VEGFA and VEGF164 lentivector infant blood stream compared with the adult. In the United shRNAs significantly reduced IVNV compared with control, States, an infant with severe ROP is often younger and smaller but only the VEGF164 knockdown maintained inhibition at a (with less blood volume) than an infant with severe ROP in later time point in the OIR model. Also, targeted Müller cell countries lacking optimal resources for prenatal care. There- knockdown of VEGFA, but not of VEGF164, increased cell fore, the safety profiles from studies in the United States and death and thinned the outer nuclear layer, suggesting that tar- throughout the world may not be comparable. Antiangiogenic geting Müller cell VEGF164 may be safer than targeting treatment may need to be individualized based on eye and VEGFA. However, longer-term studies on structure and func- infant size. It is important to monitor body weight gain and not tion are needed. Taken together, these studies raise concern only birth weight, vascular coverage, persistent avascular about the safety of even targeted knockdown of VEGFA and retina, and recurrence of IVNV as safety parameters in infants. support the investigation of other treatment strategies. These outcomes alone may be insufficient based on outer nuclear layer thinning and cell death found after experimental Clinical or Translational Implications targeted knockdown of VEGFA.14 The Efficacy of Intravitreal Bevacizumab Treatment for Stage 3þ ROP study suggests that VEGF inhibition with New Study Directions bevacizumab may alter angiogenic pathophysiology, but follow-up studies also suggest that the treatment has a Erythropoietin Derivatives broader effect on the overall biochemistry of ROP and may account for some of the late failures. Experimental studies Besides anti-VEGF agents, there is renewed interest in show that inhibition of VEGF using intravitreal antibodies, erythropoietin derivatives for their neuroprotective effects. VEGFR2 inhibitors, or targeted knock down of overex- However, some experimental evidence suggests erythro- pressed VEGF or VEGF164 in Müller cells may reduce poietin may increase the risk of severe ROP.45 Darbe- IVNV and permit physiologic retinal vascular development. poetin, a form of erythropoietin, neither increased nor Broad inhibition of VEGF signaling in multiple cell types reduced the risk of severe ROP in one trial, although such as what occurs with an intravitreal anti-VEGF antibody numbers were low.46 Erythropoietin binds the erythro- can lead to recurrent IVNV and systemic toxicity shown by poietin receptor (EPOR), which forms a homodimer and reduced body weight gain.41 Even targeted knockdown of activates the Janus-activated Kinase/Signal Transducer and Müller cell-derived VEGF experimentally may lead to Activation of Transcription (JAK/STAT) pathway in hema- retinal neuronal death. Other studies have shown the VEGF- topoiesis. Activated EPOR also can bind the b common re- Trap to inhibit retinal vascularization23 and retinal neural ceptor to form a tissue protective factor, which is protective function.43 The intravitreal aptamer pegaptanib did not in models of stroke and inflammation. Certain forms of inhibit severe ROP (Trese M, personal communication, erythropoietin preferentially bind the tissue protective factor 2014), but many questions exist, including the mechanism and are being studied in ROP. In the rat OIR model, the b of action of the aptamer, the timing when delivered, the common receptor was not highly expressed, whereas VEGF dose used and lack of specificity in targeting Müller cells. increased the expression and activation of EPOR and led to Pegaptanib is being studied for ROP; therefore, experimental an interaction between activated EPOR and VEGFR2, which studies are needed to determine long-term safety as well as overactivated STAT3 in endothelial cells to cause phase 2 efficacy of VEGF164 knockdown. Although gene therapy or IVNV. This report supports the idea that EPOR is important subretinal injections are not recommended in premature in- in phase 2 IVNV and can be activated by erythropoietin fant eyes, studies to regulate VEGFR2 signaling in endo- or VEGF.47 thelial cells and to preserve the neuroprotective effects of VEGFR2 signaling in the developing retina seem warranted IGF-1/IGF-1BP3 and Omega 3 Fatty Acids based on experimental evidence. The American Academy of Ophthalmology and the A clinical trial examining the insulin-like growth factor-1/ American Academy of Pediatrics provided guidelines for the insulin-like growth factor-binding protein 3 (IGF-1/IGFBP3) use of anti-VEGF agents in ROP.44 Still, more information is underway in Europe to test its role in infant growth, on appropriate dose, type of agent, and long-term safety is increasing physiologic retinal vascular development to needed. It is not reasonable to make assumptions that the reduce early avascular retina and to prevent the vascular systemic effects from an intravitreal drug in the preterm infant phase of ROP. To reduce the potential of causing vaso- will be similar to those in adults. A single intravitreal injection proliferation, attempts are being made only to replenish of anti-VEGF treatment seems to change the natural history of IGF-1 to levels that would be normal in preterm infants at ROP, with occurrences reported almost 4 months later and low risk of severe ROP developing. 208  Hartnett
Translational Science Review Nutrition Prematurity randomized trial. Arch Ophthalmol 2003;121: 1684–94. Omega-3 fatty acids can reduce phase 1 and phase 2 in the 10. Chan-Ling T, McLeod DS, Hughes S, et al. Astrocyte- mouse OIR model and are being investigated in human infants. endothelial cell relationships during human retinal vascular Weight, IGF, neonatal ROP (WINROP), an algorithm origi- development. Invest Ophthalmol Vis Sci 2004;45:2020–32. nally based on IGF-1 but simplified to include only weight 11. McLeod DS, Hasegawa T, Prow T, et al. The initial fetal gain, is being studied to identify infants at the greatest risk of human retinal vasculature develops by vasculogenesis. Dev severe ROP.48 The hope is that this strategy will reduce the Dyn 2006;235:3336–47. 12. Chan-Ling T, Gock B, Stone J. The effect of oxygen on vaso- burden in screening for ROP, which is increasing worldwide. formative cell division: evidence that “physiological hypoxia” Antioxidants is the stimulus for normal retinal vasculogenesis. Invest Oph- thalmol Vis Sci 1995;36:1201–14. Oxidative signaling is important to ROP phases, but as 13. Bai Y, Ma JX, Guo J, et al. Müller cell-derived VEGF is a experimental models demonstrate, the picture is complex significant contributor to retinal neovascularization. J Pathol and outcomes after activation of factors can depend on the 2009;219:446–54. cell type within the retina. Clinical trials testing certain 14. Jiang Y, Wang H, Culp D, et al. Targeting Muller cellederived antioxidants (lutein, vitamin E, n-acetyl cysteine) have not VEGF164 to reduce intravitreal neovascularization in the rat model of retinopathy of prematurity. Invest Ophthalmol Vis Sci vascularized ROP safely or effectively. 2014;55:824–31. Other Avenues of Study 15. Rivera JC, Sapieha P, Joyal JS, et al. Understanding retinop- athy of prematurity: update on pathogenesis. Neonatology Experimental evidence suggests that neural guidance mol- 2011;100:343–53. ecules, such as the semaphorins that repel neurons during 16. Gerhardt H, Golding M, Fruttiger M, et al. VEGF guides development, also may guide capillaries in physiologic angiogenic sprouting utilizing endothelial tip cell filopodia. retinal vascular development and pathologic conditions such J Cell Biol 2003;161:1163–77. as IVNV.15 Inflammatory mediators and the prostaglandin 17. Smith LE, Wesolowski E, McLellan A, et al. Oxygen-induced pathways have been studied in experimental models of retinopathy in the mouse. Invest Ophthalmol Vis Sci 1994;35: 101–11. ROP.49 Plasmin is being tested for stage 4 or 5 ROP in a 18. Penn JS, Henry MM, Tolman BL. Exposure to alternating clinical trial. However, for early and vascular phases of hypoxia and hyperoxia causes severe proliferative retinopathy human ROP, plasmin breaks down the extracellular matrix in the newborn rat. Pediatr Res 1994;36:724–31. into components important for physiologic retinal vascular 19. Cunningham S, Fleck BW, Elton RA, Mclntosh N. Trans- development.50 b-Adrenergic inhibition has been suggested cutaneous oxygen levels in retinopathy of prematurity. Lancet to reduce severe ROP, but b-adrenergic agonism also can be 1995;346:1464-5. antiangiogenic.51 More study is needed. 20. Di Fiore JM, Kaffashi F, Loparo K, et al. The relationship between patterns of intermittent hypoxia and retinopathy References of prematurity in preterm infants. Pediatr Res 2012;72: 606–12. 21. Wang H, Smith GW, Yang Z, et al. Short hairpin RNA- 1. Terry TL. Extreme prematurity and fibroblastic overgrowth of mediated knockdown of VEGFA in Muller cells reduces persistent vascular sheath behind each crystalline lens: (1) intravitreal neovascularization in a rat model of retinopathy of preliminary report. Am J Ophthalmol 1942;25:203–4. prematurity. Am J Pathol 2013;183:964–74. 2. Schepens CL. A new ophthalmoscope demonstration. Trans 22. Wang H, Yang Z, Jiang Y, et al. Quantitative analyses of Am Acad Ophthalmol Otolaryngol 1947;51:298–301. retinal vascular area and density after different methods to 3. Hartnett ME, Penn JS. Mechanisms and management of reti- reduce VEGF in a rat model of retinopathy of prematurity. nopathy of prematurity. N Engl J Med 2012;367:2515–26. Invest Ophthalmol Vis Sci 2014;55:737–44. 4. Gilbert C. Retinopathy of prematurity: a global perspective of 23. Lutty GA, McLeod DS, Bhutto I, Wiegand SJ. Effect of VEGF the epidemics, population of babies at risk and implications for Trap on normal retinal vascular development and oxygen- control. Early Hum Dev 2008;84:77–82. induced retinopathy in the dog. Invest Ophthalmol Vis Sci 5. Shah PK, Narendran V, Kalpana N. Aggressive posterior 2011;52:4039–47. retinopathy of prematurity in large preterm babies in South 24. Saito Y, Geisen P, Uppal A, Hartnett ME. Inhibition of India. Arch Dis Child Fetal Neonatal Ed 2012;97:F371–5. NAD(P)H oxidase reduces apoptosis and avascular retina in an 6. Bizzarro MJ, Hussain N, Jonsson B, et al. Genetic susceptibility animal model of retinopathy of prematurity. Mol Vis [serial to retinopathy of prematurity. Pediatrics 2006;118:1858–63. online] 2007;13:840–53. Available at: http://www.molvis. 7. International Committee for the Classification of Retinopathy org/molvis/v13/a92/. Accessed August 2, 2014. of Prematurity. The International Classification of Retinopathy 25. Niesman MR, Johnson KA, Penn JS. Therapeutic effect of of Prematurity revisited. Arch Ophthalmol 2005;123:991–9. liposomal superoxide dismutase in an animal model of reti- 8. Hartnett ME. Studies on the pathogenesis of avascular nopathy of prematurity. Neurochem Res 1997;22:597–605. retina and neovascularization into the vitreous in peripheral 26. Schaffer DB, Palmer EA, Plotsky DF, et al. Prognostic factors severe retinopathy of prematurity (an American Ophthal- in the natural course of retinopathy of prematurity. Ophthal- mological Society thesis). Trans Am Ophthalmol Soc mology 1993;100:230–7. 2010;108:96–119. 27. Sears JE, Hoppe G, Ebrahem Q, Anand-Apte B. Prolyl hy- 9. Early Treatment for Retinopathy of Prematurity Cooperative droxylase inhibition during hyperoxia prevents oxygen- Group. Revised indications for the treatment of retinopathy of induced retinopathy. Proc Natl Acad Sci U S A 2008;105: prematurity: results of the Early Treatment for Retinopathy of 19898–903. 209  Ophthalmology Volume 122, Number 1, January 2015 28. Wang H, Zhang SX, Hartnett ME. Signaling pathways trig- 41. McCloskey M, Wang H, Jiang Y, et al. Anti-VEGF antibody gered by oxidative stress that mediate features of severe reti- leads to later atypical intravitreous neovascularization and nopathy of prematurity. JAMA Ophthalmol 2013;131:80–5. activation of angiogenic pathways in a rat model of retinop- 29. Wang H, Yang Z, Jiang Y, Hartnett ME. Endothelial NADPH athy of prematurity. Invest Ophthalmol Vis Sci 2013;54: oxidase 4 mediates vascular endothelial growth factor receptor 2020–6. 2-induced intravitreal neovascularization in a rat model of 42. Greenberg KP, Geller SF, Schaffer DV, Flannery JG. Targeted retinopathy of prematurity. Mol Vis [serial online] 2014;20: transgene expression in Muller glia of normal and diseased 231–41. Available at: http://www.molvis.org/molvis/v20/231/. retinas using lentiviral vectors. Invest Ophthalmol Vis Sci Accessed August 2, 2014. 2007;48:1844–52. 30. Wang H, Byfield G, Jiang Y, et al. VEGF-mediated STAT3 43. Tokunaga CC, Mitton KP, Dailey W, et al. Effects of anti- activation inhibits retinal vascularization by down-regulating VEGF treatment on the recovery of the developing retina local erythropoietin expression. Am J Pathol 2012;180:1243–53. following oxygen-induced retinopathy. Invest Ophthalmol Vis 31. Byfield G, Budd S, Hartnett ME. The role of supplemental Sci 2014;55:1884–92. oxygen and JAK/STAT signaling in intravitreous neo- 44. American Academy of Pediatrics Section on Ophthalmology, vascularization in a ROP rat model. Invest Ophthalmol Vis Sci American Academy of Ophthalmology, American Association 2009;50:3360–5. for Pediatric Ophthalmology and Strabismus, American 32. Holmes JM, Duffner LA. The effect of postnatal growth Association of Certified Orthoptists. Screening examination of retardation on abnormal neovascularization in the oxygen premature infants for retinopathy of prematurity. Pediatrics exposed neonatal rat. Curr Eye Res 1996;15:403–9. 2013;131:189–95. 33. Hellström A, Smith LE, Dammann O. Retinopathy of pre- 45. Chen J, Connor KM, Aderman CM, et al. Suppression of maturity. Lancet 2013;382:1445–57. retinal neovascularization by erythropoietin siRNA in a mouse 34. Hartnett ME, Morrison MA, Smith S, et al. Genetic variants model of proliferative retinopathy. Invest Ophthalmol Vis Sci associated with severe retinopathy of prematurity in extremely 2009;50:1329–35. low birth weight infants. Invest Ophthalmol Vis Sci 2014 Aug 46. Ohls RK, Christensen RD, Kamath-Rayne BD, et al. 12. pii: IOVS-14-14841. doi: 10.1167/iovs.14-14841. A randomized, masked, placebo-controlled study of darbe- 35. McColm JR, Geisen P, Hartnett ME. VEGF isoforms and their poetin alfa in preterm infants [report online]. Pediatrics expression after a single episode of hypoxia or repeated fluctua- 2013;132:e119–27. tions between hyperoxia and hypoxia: relevance to clinical ROP. 47. Yang Z, Wang H, Jiang Y, Hartnett ME. VEGFA acti- Mol Vis [serial online] 2004;10:512–20. Available at: http: vates erythropoietin receptor and enhances VEGFR2- //www.molvis.org/molvis/v10/a63/. Accessed August 2, 2014. mediated pathological angiogenesis. Am J Pathol 36. Budd S, Byfield G, Martiniuk D, et al. Reduction in endo- 2014;184:1230–9. thelial tip cell filopodia corresponds to reduced intravitreous 48. Wu C, Vanderveen DK, Hellstrom A, et al. Longitudinal but not intraretinal vascularization in a model of ROP. Exp postnatal weight measurements for the prediction of Eye Res 2009;89:718–27. retinopathy of prematurity. Arch Ophthalmol 2010;128: 37. Zeng G, Taylor SM, McColm JR, et al. Orientation of endo- 443–7. thelial cell division is regulated by VEGF signaling during 49. Capozzi ME, McCollum GW, Penn JS. The role of cyto- blood vessel formation. Blood 2007;109:1345–52. chrome p450 epoxygenases in retinal angiogenesis. Invest 38. Hartnett ME, Martiniuk D, Byfield G, et al. Neutralizing VEGF Ophthalmol Vis Sci 2014;55:4253–60. decreases tortuosity and alters endothelial cell division orien- 50. Penn JS, Rajaratnam VS. Inhibition of retinal neo- tation in arterioles and veins in a rat model of ROP: relevance to vascularization by intravitreal injection of human rPAI-1 in a plus disease. Invest Ophthalmol Vis Sci 2008;49:3107–14. rat model of retinopathy of prematurity. Invest Ophthalmol Vis 39. Mintz-Hittner HA, Kennedy KA, Chuang AZ. BEAT-ROP Sci 2003;44:5423–9. Cooperative Group. Efficacy of intravitreal bevacizumab for stage 51. Chen J, Hellstrom A, Smith LE. Author response: different 3þ retinopathy of prematurity. N Engl J Med 2011;364:603–15. efficacy of propranolol in mice with oxygen-induced retinop- 40. Patel RD, Blair MP, Shapiro MJ, Lichtenstein SJ. Significant athy: could differential effects of propranolol be related to treatment failure with intravitreous bevacizumab for retinop- differences in mouse strains [letter]? Invest Ophthalmol Vis athy of prematurity [letter]. Arch Ophthalmol 2012;130:801–2. Sci 2012;53:7728–9. Footnotes and Financial Disclosures Originally received: June 30, 2014. Abbreviations and Acronyms: Final revision: July 21, 2014. EPOR ¼ erythropoietin receptor; HIF ¼ hypoxia-inducible factors; IGF- Accepted: July 29, 2014. 1 ¼ insulin growth factor-1; IGFBP3 ¼ insulin-like growth factor binding Available online: October 14, 2014. Manuscript no. 2014-1032. protein 3; IVNV ¼ intravitreal neovascularization; JAK ¼ Janus-activated John A. Moran Eye Center, University of Utah, Salt Lake City, Utah. kinase; OIR ¼ oxygen-induced retinopathy; PRVD ¼ physiologic retinal Financial Disclosure(s): vascular development; ROP ¼ retinopathy of prematurity; shRNA ¼ short- hairpin RNA; STAT3 ¼ signal transducer and activator of transcription 3; The author(s) have no proprietary or commercial interest in any materials VEGF ¼ vascular endothelial growth factor; VEGFR1 ¼ vascular endo- discussed in this article. thelial growth factor receptor 1; VEGFR2 ¼ vascular endothelial growth Supported by the National Eye Institute, National Institutes of Health, factor receptor 2. Bethesda, Maryland (grant nos.: EY015130 [M.E.H.] and EY017011 [M.E.H.]); March of Dimes, White Plains, NY (grant no.: 6-FY13-75 Correspondence: [M.E.H.]); and Departmental Support from Research to Prevent Blindness. M. Elizabeth Hartnett, MD, John A. Moran Eye Center, University of Utah, 65 Mario Capecchi Drive, Salt Lake City, UT 84132. E-mail: me.hartnett@ The sponsor or funding organization had no role in the design or conduct of this research. hsc.utah.edu. 210