approaches. One medical treatment that has advanced furthest from basic science into clinical practice is the inhibition of vascular endothelial growth factor (VEGF). AntiVEGF compounds were initially developed for treatment of wet age-related macular degeneration (AMD) but have recently also found their way into clinical trials for PDR (reviewed in [3]) and are considered for treatment of ROP [4–9].

VEGF has been extensively studied and is rightfully considered a “master switch” for angiogenesis [10]. It is unquestionably one of the major players in proliferative retinopathies and a valid target for anti-angioproliferative treatment approaches. However, both ROP as well as PDR have underlying pathomechanisms that are regulated by extensive and intricate metabolic pathways both locally in the retina as well as on a systemic level. It is therefore not only legitimate but rather essential to further investigate the underlying pathomechanisms of ROP and PDR to unveil angiogenic mediators that function upstream of VEGF expression. In proliferative retinopathies as well as in other angiogenesis-related diseases, VEGF can be viewed as possibly the most important mediator of a final common angiogenic pathway that is, however, activated through a variety of upstream mechanisms that can be very disease-specific [11]. Instead of targeting VEGF at the end of the angiogenic cascade, altering these disease-specific mediators upstream of VEGF might be a more effective approach to treating PDR and ROP. By summarizing our current knowledge about IGFBP-3 in regard to proliferative retinopathies, this chapter aims at evaluating the pathogenetic relevance as well as the potential therapeutic potential of one of the factors that might alter disease mechanisms upstream of VEGF expression in proliferative retinopathies.

THE GROWTH-HORMONE/INSULIN-LIKE GROWTH FACTOR PATHWAY IN PROLIFERATIVE RETINOPATHIES

Proliferative Diabetic Retinopathy (PDR)

Various systemic factors have been identified in diabetic patients that affect the severity of PDR: Obesity, smoking, and unstable control of blood glucose have all been found to be associated with increased severity of PDR. A potential role of growth hormone (GH) in PDR has been first suggested in the 1950s after anecdotal observations of attenuated diabetic retinopathy in women with postpartum hemorrhagic necrosis of the pituitary gland (Sheehan syndrome) [12]. Numerous studies thereafter have found that pituitary dysfunction can prevent or reverse proliferative retinopathy in diabetes patients [13–20]. Additionally, it was reported that GH replacement therapy for patients with GH deficiency can induce a diabetic-like retinopathy, which is attenuated after discontinuation of GH treatment [21].

These early observations about the role of GH in PDR have led to intense research into the downstream mediators of GH signaling. In this respect, insulin-like growth factor 1 (IGF-1) appears not only interesting as one of GH’s prime downstream effectors but also because IGF-1 shares receptor-binding affinities with insulin, the disease-defining hormone in diabetes. Clinical studies have found increased levels of IGF-1 in serum and vitreous of patients with PDR [22–31]. However, a clear correlation between disease stage or progression and IGF-1 levels could not be confirmed in all studies [32–34]. These differing results may in part be attributed to differing methodologies for measuring IGF-1. Some studies did not distinguish between free IGF-1 and IGF-1 bound

to binding proteins (IGFBPs; reviewed in [2]). The role of IGFBPs in regulating IGF bioavailability and action will be the focus of Section “IGFBP-3 as a Regulator of the Growth-Hormone/Insulin-Like Growth Factor Pathway” of this chapter.

Retinopathy of Prematurity (ROP)

Early investigations in humans have found that the severity of ROP is mainly determined by (1) postnatal oxygen exposure, (2) low gestational age/birth weight, and (3) slow postpartum weight gain [35–43]. The fact that prematurity is the most significant risk factor for ROP suggests that factors involved in growth and development are critical. Hellstrom et al. were the first to describe a direct link between low growth hormone levels and reduced retinal vascularization in children with congenital GH deficiency [44]. Consequent studies focused on one of the prime downstream mediators of GH function: IGF-1. IGF-1 is expressed in liver cells when they are exposed to GH stimulation [45, 46] and plays an important role in fetal growth and development during all stages of pregnancy but particularly in the third trimester [47]. The serum concentration of IGF-1, but not IGF-2, increases with gestational age and correlates with fetal size [48, 49]. IGF-1 levels rise significantly in the third trimester of pregnancy, but after birth decrease due to the loss of IGF-1 provided by the placenta [47]. Intriguingly, low levels of IGF-1 in preterm infants postpartum have been found to prevent normal retinal vascular growth [50] and correlated directly with the severity of clinical ROP [51–54].

The role of IGF-1 in ROP, however, becomes more complex when later disease stages are considered: While physiologic IGF-1 levels might be necessary during early retinal development to prevent ROP, IGF-1 might play a detrimental role during the proliferative stages of ROP. If during the course of postnatal retinal development in the preterm infant the retinal vascular development fails to keep up with the increased retinal demand for oxygen, the peripheral avascular parts of the developing retina will eventually respond to this oxygen shortage by expressing pathologically high levels of pro-angiogenic mediators like VEGF to boost retinal vessel growth. Due to this pro-angiogenic overstimulation retinal vessel growth becomes erratic and abnormal vessels begin to sprout from the retina into the vitreous. These disorganized neovascular tufts eventually lead to severe complications like intravitreal bleeding or retinal detachment caused by traction of the abnormal vessels on the underlying retina. In this second phase of ROP, IGF-1 can act as a permissive factor for retinal neovascularization amplifying VEGF-stimulated pathological vessel growth in the hypoxic retina. The detrimental role of IGF-1 during this phase of proliferative retinopathy is illustrated by the observation that inhibition of IGF-1 prevents hypoxia-induced retinal neovascularization despite high levels of intraocular VEGF [55]. Targeting IGF-1 in ROP infants therefore needs to be carefully timed and correlated to the clinical stage of the disease: During the early stages, when normal vascularization of the retina can still be achieved, IGF-1 levels should be monitored and increased to physiologic levels if needed. This first phase of ROP occurs from birth to approximately postmenstrual age 30–32 weeks. If by this time the retinal vasculature has not developed sufficiently to meet the demands of the maturing retina, high growth factor concentrations from the avascular parts of the retina will induce pathological neovascularization. This marks the second phase of ROP. During the second phase of ROP, IGF-1 supplementation can have detrimental effects by augmenting the growth of pathologic neovessels (reviewed in [50]).

Animal Models of Proliferative Retinopathies

Most of our understanding regarding the underlying mechanisms of proliferative retinopathies comes from the use of animal models of oxygen-induced retinopathy (OIR) that closely mimic the disease process of ROP in humans. In contrast to humans, many animals such as mice, rats, kittens, and beagle pups have incompletely vascularized retinas at birth and therefore resemble the immature retinal state of premature infants. The model that is most widely used to study disease mechanisms and possible interventions is a mouse model of OIR that was first described in 1994 [56]. In this model, neonatal mice are exposed to 75% oxygen from postnatal day 7–12. During this 5-day exposure to hyperoxia, vessel regression and the cessation of normal radial vessel growth occurs, mimicking the first phase of ROP. Other animal models also mimic this early phase of oxygen-induced vessel regression [57, 58]. The second phase of ROP that is characterized by abnormal vessel formation can also be studied in the OIR mouse model: When mice are returned to room air on postnatal day 12, the non-perfused parts of the retina become hypoxic and induce the expression of angiogenic growth factors. As a consequence, formation of abnormal retinal vascular tufts can be observed that closely resemble the erratic neovascularizations seen during the second phase of ROP in human preterm infants. Diabetic retinopathy shows a similar pattern with a first phase characterized by slow loss of retinal capillaries and a second phase of retinal neovascularization. The OIR model can therefore also be used as a tool to investigate some aspects of PDR. This is important as the currently established diabetic animal models do not develop proliferative retinopathy.

The OIR model has greatly promoted our understanding of the growth-hormone/ insulin-like growth factor pathway in ROP. Early animal studies have found that normal retinal blood vessels grow more slowly in IGF-1 knockout mouse than in wild-type controls, a pattern very similar to that seen in premature babies with ROP [51]. Subsequent studies using the OIR model have found that mice with low IGF-1 levels and transgenic mice expressing a GH receptor antagonist are resistant to hypoxia-induced retinopathy [59]. Direct proof of the pro-angiogenic role of IGF-1 in the second phase of ROP was established using an IGF-1 receptor antagonist, which was found to suppress retinal neovascularization without altering retinal VEGF levels [55]. Additionally, mice with vascular endothelial cell-specific knockout of either the IGF-1 receptor or insulin receptor show a substantial reduction in retinal neovascularization compared to control mice [60]. Mechanistically, it was suggested that IGF-1 regulates retinal neovascularization at least in part through control of VEGF activation of p44/42 MAPK, establishing a hierarchical relationship between IGF-1 and VEGF receptors [51, 55].

As outlined earlier in this chapter, no good animal models for PDR exist to date. However, an animal study of normoglycemic/normoinsulinemic transgenic mice overexpressing IGF-1 through an insulin promoter at supraphysiological levels in the retina showed loss of pericytes and thickening of basement membrane of retinal capillaries [61]. In older transgenic mice overexpressing IGF-1, neovascularization of the retina and vitreous cavity was observed which was consistent with increased IGF-1 induction of VEGF expression in retinal cells [62]. These accumulated findings suggest that once proliferative neovascular (and therefore leaky) vessels occur in the retina, leaked serum IGF-1 may further promote the proliferation of retinal vessels through stimulation