Ocular Neovascularization: a Valuable Model System

Ocular neovascularization: a valuable model system

Peter Anthony Campochiaro*,1 and Sean Francis Hackett1

1The Departments of Ophthalmology and Neuroscience, The Johns Hopkins University School of Medicine, Maumenee 719, 600 N. Wolfe Street, Baltimore, MD 21287-9277, USA

There is no unique formula for angiogenesis. Instead there related proteins, and/or their induced expression, which is a large group of potential participating proteins that may alter the angiogenesis cascade and may even alter interact in complex ways. Depending upon the surround- the effects of particular proteins. Some proteins promote ing cell types and the relative expression levels of angiogenesis in one setting and suppress it in others. angiogenesis-related proteins,the ‘angiogenesis cascade’ Proteins depend upon other proteins for their actions can vary. Therefore,it is valuable to study and compare and lack of expression of a binding partner in a the role of proteins in several well-characterized vascular particular locale may have a profound effect on the beds. The eye provides a useful model system,because it action of a protein. contains several vascular beds sandwiched between In order to define the potential actions of a protein avascular tissue. This allows for unequivocal identification and its interactions with other proteins in angiogenesis, and quantitation of new vessels. Retina-specific promoters it is useful to study its effects in several well-character- combined with inducible promoter systems provide a ized vascular beds and in different types of angiogenesis. means to regulate the expression of proteins of interest. The eye provides a valuable model system for such As a relatively isolated compartment,the eye also investigations, because it contains several vascular beds provides advantages for gene transfer. By gaining insight separated by avascular tissue. The vascular beds can be regarding the molecular signals involved in various types visualized in vivo and the presence of neovascularization of ocular angiogenesis,general concepts can emerge that can be unequivocally identified and quantified because may apply to other settings,including tumor angiogenesis. of the surrounding avascular tissue. Also, retina-specific One concept that has emerged is that despite participation promoters combined with inducible promoter systems of multiple stimulatory factors for ocular neovasculariza- provide a useful way to control expression of proteins of tion,VEGF plays an essential role and interruption of interest. By observing the different effects of proteins in VEGF signaling is an important therapeutic strategy. different ocular vascular beds, at different stages of Another concept is that while most studies have focused on development, and in different disease models, a more prevention of ocular neovascularization,regression of new complete picture of the protein’s actions and interac- vessels is desirable and is achievable with at least three tions with other proteins can emerge. agents,combretastatin A-4 phosphate,pigment epithelium-derived factor,and angiopoietin-2. Finally,endostatin and angiostatin,which have been sources of Vascular beds within the eye controversy because of inconsistent results in tumor In the eye, there are three vascular beds that have been models,have been shown to have good efficacy when intensively studied, the hyaloidal, retinal, and choroidal delivered by gene transfer in models of ocular neovascu- circulations. The hyaloidal vasculature develops in the larization. These results provide leads for new ocular embryonic eye and extends from the optic nerve through treatments and perspective for evaluation of studies of the vitreous to surround the developing lens. Near the neovascularization in extraocular tissues. end of development, when other circulatory systems in Oncogene (2003) 22, 6537–6548. doi:10.1038/sj.onc.1206773 the eye are completed, the hyaloidal vessels regress. This provides an opportunity to study molecular signals Keywords: VEGF; angiogenesis; angiopoietins; involved in vascular regression. retinopathies; age-related macular degeneration The retinal circulation, unlike the hyaloidal circulation, can be studied in adult animals as well as during its development. In rodents, retinal vascular development Introduction occurs postnatally, facilitating the study of developmental or physiologic angiogenesis. On the day of birth, Numerous gene products participate in angiogenesis, postnatal day 0 (P0), there are no retinal vessels and flat but that participation may vary somewhat in different mounts of the retina after perfusion with fluorescein- settings. Vascular beds differ with regard to surrounding labeled dextran show hyaloidal vessels, which have not cell types, constitutive expression of angiogenesis- yet regressed, but no retinal vessels (Figure 1a). The central retinal artery enters the eye through the optic *Correspondence: PA Campochiaro; E-mail: [email protected] nerve and retinal vessels develop from the optic nerve to Ocular neovascularization PA Campochiaro and SF Hackett 6538

Figure 1 Retinal vascular development. Mice were perfused with fluorescein-labeled dextran on the day of birth, postnatal day (P) 0 (a, b), P4 (c), P7( d), P10 (e), or P18 (f), and retinal flat mounts were examined by fluorescence microscopy. The large vessels seen in flat mounts from P0 mice are hyaloidal vessels; retinal vessels are just beginning to develop at the border of the optic nerve (b, arrowheads). By P4, the margin of the developing blood vessels on the surface of the retina is located halfway between the optic nerve and the peripheral edge of the retina (c, arrowheads). Large hyaloidal vessels are still seen overlying the avascular peripheral retina. At P7, superficial vessels cover about 80% of the retina (d, arrowheads). The hyaloidal vessels have started to regress. At P10, the superficial vessels have reached the peripheral edge of the retina and the deep capillary bed has begun to develop posteriorly (e). By P18, retinal vascular development and remodeling is completed (f)

the periphery of the retina. At P4, superficial retinal make up the outer avascular half of the retina. It vessels extend from the optic nerve halfway to the receives its oxygen and nutrients from the choroidal peripheral edge of the retina (Figure 1b). At P7, vessels circulation, a high-flow system supplied by multiple long cover the entire surface of the retina (Figure 1c) and and short posterior ciliary arteries, which all feed into an sprouts from superficial vessels begin to grow into the extensive network of fenestrated capillaries, the chor- retina to form the intermediate and deep capillary beds. iocapillaris. The choriocapillaris allows plasma to pool By P18, all the three vascular beds of the retina have beneath the retinal pigmented epithelium (RPE), which been formed and remodeled, resulting in an adult retinal has tight junctions and specialized transport systems, circulation, which is relatively stable thereafter and it constitutes the outer blood retinal barrier. (Figure 1d). Retinal vascular endothelial cells have tight junctions and modified vesicular transport, and are the Developmental retinal neovascularization sites of the inner blood–retinal barrier. This is in part due to their interactions with pericytes and retinal glia Retinal astrocytes enter the retina from the optic nerve (Janzer and Raff, 1987). (Watanabe and Raff, 1988). As they migrate away from The retinal circulation supplies the inner half of the the optic nerve into avascular retina, they become retina (Figure 2a and b). The outer nuclear layer, which hypoxic and express increased levels of vascular is made up of the cell bodies of the photoreceptors, and endothelial growth factor A (VEGF) (Stone et al., the inner and outer segments of the photoreceptors, 1995). VEGF stimulates the growth of blood vessels

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6539

Figure 2 Mice with oxygen-induced ischemic retinopathy develop severe retinal neovascularization. (a, b) Paraffin-embedded ocular sections stained with Griffonia simplicifolia lectin (GSA), which selectively stains vascular cells, from a normal postnatal day (P) 17 mouse shows normal superficial, intermediate, and deep capillaries. There are no vessels extending above the internal limiting membrane (ILM). Hematoxylin counterstain shows retinal neurons. (c, d) Sections from P17mice with ischemic retinopathy stained with GSA show prominent neovascularization that has broken through the ILM and is on the surface of the retina (arrows). Dropout of the normal retinal blood vessels throughout the posterior retina is seen in the low magnification view in (c). This is also illustrated by the black (nonperfused) area in the retinal flat mount from a fluorescein dextran-perfused P17mouse with ischemic retinopathy ( e) from the optic nerve to the periphery of the retina and as provide an alternative source of survival factors. the blood vessels become functional; they alleviate Platelet-derived growth factor-B (PDGF-B) secreted by hypoxia, resulting in decreased expression of VEGF. endothelial cells is critical for the recruitment of The relative hypoxia of the deeper layers of the retina pericytes, and PDGF-deficient mice lack pericyte results in a VEGF gradient that favors sprouting from envelopment of retinal vessels (Lindahl et al., 1997). the superficial vessels, resulting in penetrating branches that form the intermediate and deep capillary beds. Pathologic retinal neovascularization VEGF also acts as a survival factor for the endothelial cells of the newly formed retinal vessels (Alon et al., During the period of retinal vascular development, 1995). Dependence on VEGF is transient and is alteration of VEGF expression in the retina causes probably eliminated by the development of cell–cell problems. This is achieved experimentally by placing contacts, particularly association with pericytes that neonatal mice in high oxygen for several days, which

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6540 causes downregulation of VEGF, stopping further drugs on choroidal neovascularization (Tobe et al., vascular development and causing vascular regression. 1998; Seo et al., 1999; Kwak et al., 2000; Ozaki et al., This results in large areas of nonperfused retina and 2000). when the mice are placed in room air, there is severe retinal ischemia resulting in high expression of VEGF and retinal neovascularization (Figure 2c–e). This VEGF and ocular neovascularization constitutes a murine model of ischemic retinopathy The correlation of increased expression of VEGF and (Smith et al., 1994), which is most analogous to retinal neovascularization suggested a role for VEGF as retinopathy of prematurity, but also mimics important a stimulatory factor. Subsequent studies have confirmed aspects of proliferative diabetic retinopathy, the most that VEGF plays a central role, because retinal common ischemic retinopathy in patients (Klein et al., neovascularization is suppressed by agents that bind 1984). VEGF (Aiello et al., 1995; Adamis et al., 1996; Saishin et al., 2003) or block VEGF receptors (Seo et al., 1999; Choroidal neovascularization Ozaki et al., 2000). In fact, VEGF receptor kinase inhibitors, which are very effective at decreasing VEGF Choroidal neovascularization occurs in diseases in signaling, completely block retinal neovascularization, which there are abnormalities of Bruch’s membrane indicating that VEGF is a necessary stimulator (Seo and/or the RPE, the most common of which is age- et al., 1999; Ozaki et al., 2000). To determine if increased related macular degeneration (The Macular Photocoa- expression of VEGF is sufficient to cause neovascular- gulation Study Group, 1991). Bruch’s membrane is a ization, the rhodopsin promoter was coupled to a full- five-layered extracellular membrane structure that sepa- length cDNA for human VEGF165 and transgenic mice rates the choriocapillaris from the RPE; it seems to (rho/VEGF mice) were generated that express VEGF in provide a physical and biochemical barrier to vascular photoreceptors (Okamoto et al., 1997). Rho/VEGF invasion of the subretinal space. Choroidal neovascular- mice develop neovascularization that originates from ization can reliably be produced in experimental animals the deep capillary bed of the retina (Figure 4a, b), but including monkeys and mice, by rupturing Bruch’s not from the superficial retinal capillaries or choroidal membrane with laser photocoagulation (Miller et al., vessels. The deep capillary bed is the last vascular bed to 1986; Tobe et al., 1998). The murine model of choroidal develop starting around P7, the same time that neovascularization due to laser-induced rupture of expression of the VEGF transgene begins, raising the Bruch’s membrane (Figure 3) has been particularly possibility that expression of a developmentally regu- useful for exploring the effect of gene products and lated permissive factor is responsible for the differential

a b

c d

Figure 3 Rho/VEGF transgenic mice with expression of VEGF in photoreceptors develop subretinal neovascularization. The rhodopsin/VEGF transgene begins expression at about postnatal day (P) 7. By P10, endothelial cells are seen migrating into the outer nuclear layer (a). At P14, well-formed blood vessels are seen extending from the deep capillary bed of the retina into the subretinal space (b), and at P21 networks of vessels are seen in the subretinal space (c). At P21, retinal ﬂat mounts of ﬂuorescein dextran-perfused transgenics show numerous tufts of new vessels partially surrounded by retinal pigmented epithelial cells located along the outer surface of the retina. Feeder vessels are seen extending from retinal vessels that are out of focus in the background

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6541 It is not known if hypoxia plays any role in choroidal a neovascularization due to age-related macular degeneration (AMD), but it is very unlikely to play a role in choroidal neovascularization that occurs in young patients with high myopia or angioid streaks in which the major problem is rupture of Bruch’s membrane. As noted above, this clinical situation is modeled by laser- induced rupture of Bruch’s membrane in mice, which results in choroidal neovascularization (Tobe et al., 1998). Despite the absence of hypoxia in this model, VEGF is a necessary stimulator, because VEGF receptor kinase inhibitors or other agents that specifi- cally and efficiently bind VEGF, essentially eliminate the choroidal neovascularization (Seo et al., 1999; Kwak et al., 2000; Saishin et al., 2003). While VEGF is b necessary, it is not sufficient, because increased expression of VEGF in photoreceptors or RPE cells does not cause choroidal neovascularization (Okamoto et al., 1997; Ohno-Matsui et al., 2002). Clinical trials using VEGF inhibitors have been initiated in patients with subfoveal choroidal neovascularization due to AMD. In an uncontrolled phase IA trial, in which 15 patients received intraocular injection of an aptamer that binds VEGF, 80% of patients had stabilization or improvement in vision after 3 months (The EyeTech Study Group, 2002). Visual acuity was improved by three lines or more in 27% of patients. A humanized rhuMAb VEGF Fab antibody is also being tested. Intravitreous injections of 125I-labeled VEGF Fab antibody showed penetration Figure 4 Choroidal neovascularization at Bruch’s membrane throughout the retina and localization in the RPE rupture sites is almost completely prevented by oral administration for up to 7days (Mordenti et al., 1999). The half-life of PKC412. (a) At 2 weeks after rupture of Bruch’s membrane with in the vitreous was 3.2 days. Preliminary results laser photocoagulation, there is prominent choroidal neovascularization stained red with GSA and HistoMark Red. (b) Oral in a phase I trial testing intravitreous injection of the administration of PKC412 during the 2-week period after rupture VEGF Fab antibody in patients with subfoveal of Bruch’s membrane, markedly suppresses the development of choroidal neovascularization due to AMD, showed a choroidal neovascularization substantial number of patients with visual improvement during the treatment phase (Schwartz et al., 2001). sensitivity of the deep capillary bed. This hypothesis was Although judgment of efficacy for each of these tested by using the rhodopsin promoter in combination VEGF antagonists must be reserved until results of with the reverse tetracycline-inducible promoter system ongoing randomized, controlled trials are obtained, (Gossen et al., 1995); double transgenic mice with these preliminary results are encouraging because doxycycline-inducible expression of VEGF in photo- spontaneous improvement in vision is very unusual in receptors were generated (Ohno-Matsui et al., 2002). patients with subfoveal choroidal neovascularization Administration of doxycycline to adult double trans- due to AMD. genic mice resulted in sprouting of neovascularization VEGF plays a major role as a stimulator of from the deep capillary bed, but not the superficial angiogenesis in some tumors such as renal cell carcino- retinal capillaries or the choroidal vessels just as was the ma and glioblastoma, and it contributes to angiogenesis case in rho/VEGF mice. This suggests that even in adult in almost all tumors (Plate et al., 1992; Shweiki et al., retina, there is likely to be a gene product that is 1992; Brown et al., 1993; Takahashi et al., 1994; Takano expressed in close proximity to the deep capillary bed et al., 1996; Nicol et al., 1997; Edgren et al., 1999). that permits VEGF to induce neovascularization. Outcomes in early stage oncology trials can be difficult Therefore, VEGF is necessary, but not sufficient for to assess, because patients with advanced stages of retinal neovascularization, which originates from the multiple different tumor types that may respond superficial capillary bed. This is supported by the differently are included. However, occasional partial demonstration in primates that repeated intraocular or complete responses have been noted in patients with injections of VEGF (Tolentino et al., 1996) or sustained advanced disease treated with inhibitors of VEGF or intravitreous release of VEGF (Ozaki et al., 1997) VEGF receptors. This and the encouraging preliminary results in several changes to retinal vessels including results in patients with choroidal neovascularization, dilation, leakage, and microaneurysms, but not retinal support the strategy of targeting tumor angiogenesis neovascularization. with VEGF antagonists.

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6542 VEGF isoforms support of this hypothesis, antagonists of GH or IGF- 1 decrease retinal neovascularization by 30–50% (Smith There are three isoforms of VEGF-A due to alternative et al., 1997). Intraocular injection of gutless adenoviral splicing of pre-mRNA (Park et al., 1993). In mice, the vectors expressing IGF-1 does not cause retinal or shortest isoform is VEGF120. It lacks exons 6 and 7, anterior segment neovascularization (unpublished data). which promote binding to extracellular matrix, and Also, coinjection with vectors expressing VEGF and therefore is the most soluble isoform. The longest IGF-1 does not cause retinal neovascularization. There- isoform is VEGF188, which is poorly soluble because it fore, IGF-1 may contribute to retinal neovasculariza- binds avidly to extracellular matrix and cell surfaces. An tion, but increased expression of IGF-1 even when intermediate isoform, VEGF , lacks exon 6, and is 164 combined with increased expression of VEGF is not more soluble than VEGF , but does some binding to 188 sufficient to initiate retinal neovascularization. extracellular matrix unlike VEGF . Mice engineered 120 Basic fibroblast growth factor (FGF2) was one of the to selectively express VEGF had normal retinal 164 first angiogenesis factors to be identified and has been vascular development and were healthy, while mice that suspected to participate in ocular neovascularization. expressed only VEGF had severely impaired out- 120 However, FGF2-deficient mice exhibit normal retinal growth and patterning of developing retinal vessels and vascular development and compared to wildtype mice, died early due to impaired myocardial angiogenesis show comparable amounts of choroidal neovasculariza- et al resulting in heart failure (Carmeliet ., 1999; tion at Bruch’s membrane rupture sites (Ozaki et al., Stalmans et al., 2002). Mice that selectively expressed 1998; Tobe et al., 1998). Transgenic mice that express VEGF , showed normal outgrowth of developing 188 high levels of FGF2 in the retina do not develop retinal vessels, but im-paired development of arterioles. neovascularization, nor do they show increased amounts Therefore, VEGF appears to be a critical isoform for 164 of ischemia-induced retinal neovascularization (Ozaki developmental retinal angiogenesis. et al., 1998). However, if photoreceptors expressing high levels of FGF2 are damaged providing the FGF2 access Other members of the VEGF family to the extracellular compartment, then increased chor- Other members of the VEGF family include VEGF-B oidal neovascularization occurs compared to wild type (Olofsson et al., 1995), VEGF-C (Joukov et al., 1996; mice (Yamada et al., 2000). Thus, increased expression Lee et al., 1996), VEGF-D (Achen et al., 1998), VEGF- of FGF2 in the retina does not stimulate angiogenesis, E (Meyer et al., 1999), and placental growth factor but rather confers a latent angiogenic potential that is (PlGF) (Maglione et al., 1991). VEGF-B does not play not manifested unless there is cell injury. an important role in retinal neovascularization, because mice deficient in VEGF-B have normal retinal vascular Tie receptors development and no difference in hypoxia-induced retinal neovascularization compared to wild type mice Tie1 and Tie2 receptors are selectively expressed on (Reichelt et al., 2003). VEGF-C and -D have their vascular endothelial cells and are required for embryo- primary action on lymphatic vessel development and nic vascular development (Dumont et al., 1994; Sato maintenance through VEGF receptor-3 (Kukk et al., et al., 1995). Mice with targeted disruption of Tie1, die 1996; Achen et al., 1998). VEGF-E is a viral homolog of between E14.5 and birth with severe edema and VEGF-A that has similar angiogenic activity despite hemorrhage, suggesting that Tie1 signaling promotes activation of VEGF receptor-2, but not VEGF receptor- vascular integrity (Puri et al., 1995; Sato et al., 1995). 1 (Meyer et al., 1999). It is not known if VEGF-C or -D Mice with targeted disruption of Tie2 or mice expressing play any role in the eye, but PlGF acts synergistically a dominant-negative Tie2 mutant die between E9.5 and with VEGF to promote retinal neovascularization, E10.5, showing lack of development of the endothelial because PlGF-deficient mice have significantly less lining of the heart, lack of remodeling of the primary- ischemia-induced retinal neovascularization than wild- capillary plexus to form more complex higher order type mice (Carmeliet et al., 2001). Since PlGF binds branching vessels, and failure of vascular invasion of VEGFR-1, and not VEGFR-2, this implies that neuroectoderm (Dumont et al., 1994). The ligand for VEGFR-1 plays an important role in the development Tie1 has not been identified. The first binding partner of retinal neovascularization, and this has been identified for Tie2, angiopoietin-1 (Ang1), binds with confirmed by the demonstration that anti-VEGFR-1 high affinity and initiates Tie2 phosphorylation and antibodies suppress ischemia-induced retinal neovascu- downstream signaling (Davis et al., 1996). Ang1 is a larization (Luttun et al., 2002). critical Tie2 agonist, because mice deficient in Ang1 die around E12.5 and show vascular defects similar to, but Potential participation of other peptide growth factors in less severe than those seen in Tie2-deficient mice (Suri ocular neovascularization et al., 1996). The second Tie2-binding partner identified, Ang2, binds with high affinity, but does not stimulate Ever since the demonstration that hypophosectomy has phosphorylation of Tie2 in cultured endothelial cells an ameliorative effect on proliferative diabetic retino- (Maisonpierre et al., 1997). In vitro, Ang2 acts as a pathy, it has been suspected that growth hormone (GH) competitive inhibitor of Ang1, because it decreases and/or insulin-like growth factor-1 (IGF1) contribute to Ang1 binding to Tie2 and Ang1-induced phosphoryla- retinal neovascularization (Wright et al., 1969). In tion. With regard to embryonic vascular development,

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6543 Ang2 also acts as an Ang1 antagonist, because transgenic mice, which caused a moderate increase transgenic mice overexpressing Ang2 have a phenotype in number and a large increase in diameter of the similar to Ang1-deficient mice (Maisonpierre et al., dermal vessels (Suri et al., 1998). Transgenic mice with 1997). increased expression of VEGF in skin showed a large The functions of Ang1 and Ang2 and the manner in increase in leaky dermal vessels, and double transgenic which Ang/Tie2 signaling modulates VEGF signaling mice with coexpression of Ang1 and VEGF had an after embryonic development are less clear. In the additive effect on angiogenesis, but the vessels did not female reproductive system, coexpression of Ang2 and leak spontaneously and were resistant to inflammation- VEGF occurs in developing follicles in association with induced leakage (Thurston et al., 1999). Increased neovascularization, while high expression of Ang2 alone expression of Ang1 in adult mice by intravascular occurs in atretic follicles in association with vascular injection of adenoviral vectors expressing Ang1 did regression (Maisonpierre et al., 1997). This observation not cause increased vascularity, but conferred leakage led to the hypothesis that Ang2 blockade of Tie2 resistance to existing vessels (Thurston et al., 2000). signaling disrupts stabilizing input to endothelial cells Overexpression of Ang1 in the retina of transgenic mice from the extracellular matrix and surrounding cells, causes suppress of retinal or choroidal neovasculariza- making endothelial cells more responsive to VEGF and tion, and suppresses VEGF-induced vascular perme- thereby stimulating NV, but Ang2 also disrupts survival ability (unpublished data). Therefore, increasing levels signals from the surrounding environment, so that in the of Ang1, even without inhibition of VEGF, may be a absence of VEGF. endothelial cells undergo apoptosis, good strategy for treatment of neovascularization. resulting in vessel regression (Maisonpierre et al., 1997). Subsequent studies in the female reproductive system, Prostaglandins and cyclooxygenase inhibitors tumor vasculature, and brain vasculature have been supportive of this hypothesis (Beck et al., 2000; Hazzard Patients who regularly take nonsteroidal anti-inflamma- et al., 2000; Pichiule and LaManna, 2002), but it does tory drugs (NSAIDs) have a 40–50% reduction in not apply to all vascular beds. In adult heart, retina, or mortality from colorectal cancer (Smalley and DuBois, choroid, high expression of Ang2 when there is low 1997). This and the demonstration that cyclooxygenase- expression of VEGF, does not cause vascular regression 2 is upregulated in colorectal and other cancers has (Visconti et al., 2002; Oshima et al., manuscript suggested that prostaglandins (PGs) may act as tumor submitted). In the retina and choroid, new vessels are promoters and that inhibition of cyclooxygenases may sensitive to Ang2, but established blood vessels are not. be chemoprotective (Kawamori et al., 1998; Williams When Ang2 expression is increased in the setting of et al., 1999, 2000). At least part of the tumor-promoting ischemic retinopathy, there is a striking increase in effect of PGs appears to be through stimulation of retinal neovascularization, but in several other situa- angiogenesis and a proangiogenic PG effect has been tions, even with moderately increased expression of noted in other systems (Form and Auerbach, 1982; VEGF, high expression of Ang2 causes regression of Ziche et al., 1982; Diaz-Flores et al., 1994). Conversely, neovascularization. These data suggest that increasing cyclooxygenase inhibitors suppress some types of Ang2 levels, particularly when combined with inhibition angiogenesis (Deutsch and Hughes, 1979; Davel et al., of VEGF, may be a useful way to achieve regression of 1985; Tsujii et al., 1998; Jones et al., 1999; Sawaoka pathologic neovascularization. et al., 1999; Yamada et al., 1999; Masferrer et al., 2000). It has been postulated that Ang1 counters the effects Nepafenac, a potent cyclooxygenase-1 (COX-1) and of Ang2 and acts a vascular stabilizing factor. Testing of COX-2 inhibitor, readily penetrates the cornea when this hypothesis has been difficult because, Ang1 forms applied topically, and suppresses retinal and choroidal multimers complicating its isolation and use for in vitro neovascularization (Takahashi et al., 2003b). Nepafenac studies (Procopio et al., 1999). A genetically engineered also blunted the ischemia-induced increase in VEGF chimeric protein, Ang1*, that does not form large mRNA in the retina, suggesting that this is how it multimers, but still activates Tie2 receptors, has allowed achieves its antiangiogenic effects. Therefore, COX investigation of the effects of activating Tie2 in various inhibitors are modest suppressors of tumor and ocular in vitro systems (Koblizek et al., 1998). Ang1* promotes angiogenesis, and are likely to provide benefits with survival of cultured endothelial cells and stabilizes regard to prophylaxis, but may be too weak to benefit tubular networks (Kwak et al., 1999; Papapetropoulos established disease. et al., 1999; Kim et al., 2000b). Since Ang1* does not behave exactly like Ang1, it cannot be concluded that Tubulin-binding agents Ang1 has the same functions, but such studies at least provide some support for a potential Ang1 vascular Tubulin-binding agents, such as vincristine, vinblastine, maintenance function. But there is also evidence and colchicine, cause tumor necrosis from damage to suggesting that Ang1 may stimulate chemotaxis, tubule tumor blood vessels, but at doses that are too toxic for formation, and vascular sprouting and hence act as a patients to tolerate (Seed et al., 1940; Hill et al., 1993). proangiogenic agent (Koblizek et al., 1998; Witzenbich- Combretastatin A-4 is a naturally occurring structural ler et al., 1998; Hayes et al., 1999; Kim et al., 2000a). analog of colchicine that binds tubulin at the same site This is supported by increased expression of Ang1 in as colchicine, but with different characteristics (Pettit skin under control of the keratin-14 promoter in et al., 1989; Woods et al., 1995) that impart selective

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6544 toxicity to tumor vasculature (Dark et al., 1997). but showed no difference in ischemia-induced retinal Combretastatin A-4-phosphate (CA-4-P) is a more neovascularization compared to wild type mice. These soluble, inactive prodrug that is converted to CA-4 by data indicate that NO is an important stimulator of endogenous nonspecific phosphatases (Pettit et al., choroidal neovascularization and that reduction of NO 1995). The selectivity of CA-4 for tumor vasculature by pharmacologic means is a good treatment strategy. allows for antitumor effects at doses of CA-4-P that are However, the situation is more complex for ischemia- well tolerated (Chaplin et al., 1999). A single dose of induced retinal neovascularization, for which NO 100 mg/kg is well tolerated in adult mice and has produced in endothelial cells by eNOS is stimulatory, beneficial effects on several types of tumors (Dark but NO produced in other retinal cells by iNOS and/or et al., 1997; Horsman et al., 1998; Grosios et al., 1999; nNOS is inhibitory. Selective inhibitors of eNOS may be Tozer et al., 1999). This led to a Phase I study in patients needed for treatment of retinal neovascularization. with advanced cancer, which showed a reasonable safety profile for 10 min intravascular infusions of doses Integrin antagonists r60 mg/m2 given every 3 weeks, and some preliminary evidence of possible efficacy in that a patient with Integrins avb3 and avb5 are not detectable on normal anaplastic thyroid cancer had a complete response endothelial cells in skin, but are highly expressed on (Dowlati et al., 2002). Daily intraperitoneal injections endothelial cells participating in angiogenesis (Brooks of CA-4-P significantly suppress neovascularization in et al., 1994a). Treatment of tumors or chick chorioal- transgenic mice with ectopic expression of VEGF in lantoic membrane with LM609, a monoclonal antibody photoreceptors and mice with laser-induced rupture of that binds avb3, causes endothelial cell apoptosis, Bruch’s membrane (Nambu et al., 2003). Administra- vascular regression, and suppression of tumor growth tion of CA-4-P to mice with established choroidal (Brooks et al., 1994b). A humanized version of neovascularization results in significant regression of the LM609, Vitaxin, is being tested in clinical trials for its neovascularization. Therefore, CA-4-P shows promise effects on tumors (Gutheil et al., 2000; Posey et al., as a novel treatment for tumors and for ocular 2001). neovascularization. With regard to avb3, there is good correlation between neovascularization in tumors and retinal Inhibitors of nitric oxide synthetase neovascularization. There is upregulation of avb3on endothelial cells participating in ischemia-induced Nitric oxide (NO) has been shown to have proangio- retinal neovascularization, and agents that bind avb3 genic or antiangiogenic effects depending upon the and/or avb5 partially suppress retinal neovasculariza- setting. Mice with targeted deletion of one of the three tion (Friedlander et al., 1996; Hammes et al., 1996; Luna isoforms of nitric oxide synthase (NOS) were selected to et al., 1996). The effect is modest (about 50% inhibition) investigate the effects of NO in ocular neovasculariza- compared to some other agents, and agents that bind tion (Ando et al., 2001, 2002). In transgenic mice with avb3 and avb5 have no significant effect in models increased expression of VEGF in photoreceptors, of choroidal neovascularization (unpublished data); deficiency of any of the three isoforms caused a therefore, there has been hesitancy to move into clinical significant decrease in subretinal neovascularization, trials for ocular neovascularization with these agents. but no alteration of VEGF expression. In mice with Also, their mechanism of action is uncertain. While laser-induced rupture of Bruch’s membrane, deficiency it was initially thought that the antibodies and of iNOS or nNOS, but not eNOS, caused a significant cyclic peptides that bind avb3 act as antagonists that decrease in choroidal neovascularization. In mice with promote apoptosis by perturbing endothelial cell adhe- oxygen-induced ischemic retinopathy, deficiency of sion, there is evidence suggesting that they may act by eNOS, but not iNOS or nNOS, caused a significant stimulating intracellular signaling through avb3, result- decrease in retinal neovascularization and decreased ing in activation of antiangiogenic pathways (Hynes, expression of VEGF. These data suggest that NO 2002). A better understanding of the mechanism of contributes to both retinal and choroidal neovascular- action of antibodies and cyclic peptides that bind avb3 ization, but that different isoforms of NOS are involved. and/or avb5 may provide insights as to how to improve Oral administration of NG-monomethyl-l-arginine efficacy. (l-NMMA), a broad-spectrum NOS inhibitor, caused significant inhibition of choroidal neovascularization in Proteinase inhibitors mice with laser-induced rupture of Bruch’s membrane, and significantly inhibited subretinal neovascularization Since neovascularization involves degradation of the in transgenic mice with expression of VEGF in extracellular matrix by invading endothelial cells, it has photoreceptors (rho/VEGF mice), but did not inhibit been postulated that proteinase inhibitors will inhibit retinal neovascularization in mice with ischemic retino- neovascularization. This hypothesis has been supported pathy. Triple homozygous mutant mice deficient in all by studies demonstrating that systemic administration three NOS isoforms had marked suppression of of a nonspecific matrix metalloproteinase (MMP) choroidal neovascularization at sites of rupture of inhibitor suppresses retinal neovascularization (Das Bruch’s membrane and near-complete suppression et al., 1999) and mice deficient in MMP-9 develop less of subretinal neovascularization in rho/VEGF mice, choroidal neovascularization at Bruch’s membrane

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6545 rupture sites than wild type mice (Lambert et al., 2002). Other agents that have been demonstrated to suppress However, genetic manipulation of plasminogen activa- ocular neovascularization after intraocular gene transfer tor inhibitor-1 or tissue inhibitor of metalloproteinase-1 include soluble VEGF receptor-1 (Flt-1) (Honda et al., does not suggest that these endogenous proteinase 2000; Lai et al., 2001b; Gehlbach et al., 2003b), angio- inhibitors have antiangiogenic activity (Lambert et al., statin (Lai et al., 2001a; Raisler et al., 2002), tissue 2001; Yamada et al., 2001). Since several endogenous inhibitor of metalloproteinases-3 (Takahashi et al., 2000), inhibitors of angiogenesis are generated by proteolytic and soluble Tie2 (Hangai et al., 2001). While intraocular cleavage of proteins, inhibition of proteinase activity can injection of vectors maximizes delivery of gene product to have proangiogenic effects that compromise antiangio- the retina, there is still some concern regarding the genic activity and thereby complicate the use of possibility of vector-related toxicity or toxicity from proteinase inhibitors to treat tumor or ocular neovascu- ectopic expression of proteins in specialized cells such as larization. photoreceptors. Periocular injection of adenoviral vectors encoding Flt-1 (Gehlbach et al., 2003b) or PEDF Endogenous inhibitors of angiogenesis (Gehlbach et al., 2003a) results in significant inhibition of choroidal neovascularization and provides an Several endogenous inhibitors of angiogenesis have been alternative, potentially safer, means of delivery. described, including angiostatin (O’Reilly et al., 1994), endostatin (O’Reilly et al., 1997), antithrombin III (O’Reilly et al., 1999), platelet factor 4 (Maione et al., 1990), thrombospondin (Good et al., 1990), pigment Summary and conclusions epithelium-derived factor (PEDF) (Dawson et al., 1999), and several others. Adenoviral vectors encoding endo- There are compelling data suggesting that treatment statin were injected into the tail veins of mice in a model with antiangiogenic agents can have beneficial effects on of choroidal neovascularization (Mori et al., 2001a). tumors, but complexities in animal models and particu- Constructs utilizing a CMV promoter resulted in about larly in tumor patients make outcomes difficult to assess. 10-fold higher endostatin serum levels on average than The eye provides a model system with several vascular constructs utilizing a RSV promoter. There was a strong beds surrounded by avascular tissue, which facilitates positive correlation between endostatin serum levels and identification and quantitation of neovascularization. inhibition of choroidal neovascularization, indicating Inducible, retina-specific expression of transgenes that endostatin is a good candidate for antiangiogenic provides a powerful tool to explore the effects of ocular gene therapy. Subsequent studies have demon- angiogenic and antiangiogenic proteins. Data obtained strated that intraocular injection of gutless adenoviral from animal models and ocular clinical trials provide vectors or bovine immune deficiency viral vectors important correlative information that can help put encoding endostatin inhibit VEGF-induced ocular results in tumor models and clinical trials in context. neovascularization and vascular leakage (Takahashi One concept that has emerged from studies in the eye is et al., 2003a). As the name implies, PEDF is produced that VEGF is an extremely good target for treatment of by pigmented epithelial cells in the eye (Tombran-Tink neovascular diseases. Despite contributions of other et al., 1991; Steele et al., 1993), making it a particularly angiogenic factors, signaling through VEGF receptors is good candidate for a negative regulator of ocular necessary for ocular neovascularization and blockade of neovascularization. Intraocular injection of adenoviral VEGF receptors is a very efficient approach. Another vectors encoding PEDF resulted in high levels of PEDF concept that is emerging is that Ang1 and Ang2 in the retina and significant suppression of retinal or modulate the effects of VEGF, but differentially on choroidal neovascularization compared to null vector- newly developed vessels, with no identifiable effect on injected eyes (Mori et al., 2001b). Intraocular gene mature retinal and choroidal vessels. This is good news transfer of PEDF in eyes with established choroidal for treatment development, because a combination of neovascularization, resulted in regression of neovascu- Ang2 and VEGF antagonists can promote regression of larization (Mori et al., 2001c). Adeno-associated established neovascularization without adverse effects viral (AAV) vectors provide the advantages of mini- on normal blood vessels. Results in eye models have mal immune response and long-term expression. At supported results in tumor models, suggesting that the 4 or 6 weeks after intraocular injection of AAV vectors vascular targeting agent Combretastain A-4-P, is a encoding PEDF, there was significant suppression promising small molecule that can also promote of choroidal neovascularization (Mori et al., 2002). regression of established neovascularization. And final- These results are encouraging, because they suggest ly, effects in tumor models of purported endogenous that intraocular PEDF gene transfer may be useful inhibitors of neovascularization such as endostatin and for both prevention of neovascularization and treat- angiostatin have been confusing and controversial. ment of established neovascularization, and have led Ocular gene transfer of endostatin, angiostatin, and to a phase I clinical trial investigating intraocular PEDF, have shown unequivocally that these proteins injection of adenoviral vectors encoding PEDF in are antiangiogenic and provide promising therapeutic patients with age-related macular degeneration and agents. These results suggest perseverance in studies subfoveal choroidal neovascularization (Rasmussen investigating their potential application to treatment of et al., 2001). tumors.

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6546 Acknowledgements Wasserman Merit Award (PAC)); Dr and Mrs William Lake. Supported by EY05951, EY12609, and core grant P30EY1765 PAC is the George S and Dolores Dore Eccles Professor of from the NEI; Research to Prevent Blindness, Inc (a Lew R Ophthalmology.

References

Achen MG, Jeltsch M, Kukk E, Makinen T, Vitali A, Wilks Dumont DJ, Gradwohl G, Fong G-H, Puri MC, Gerstenstein AF, Alitalo K and Stacker SA. (1998). Proc. Natl. Acad. Sci. M, Auerbach A and Breitman ML. (1994). Genes Dev, 8, USA, 95, 548–553. 1897–1909. Adamis AP, Shima DT, Tolentino MJ, Gragoudas ES, Ferrara Edgren M., Lennernas B, Larrsson A and Nilsson S. (1999). N, Folkman J, D’Amore PA and Miller JW. (1996). Arch. Anticancer Res, 19, 869–873. Ophthalmol., 114, 66–71. EyeTech Study Group (2002). Retina, 22, 143–152. Aiello LP, Pierce EA, Foley ED, Takagi H, Chen H, Riddle L, Form DM and Auerbach R. (1982). Proc. Soc. Exp. Biol. Ferrara N, King GL and Smith LEH. (1995). Proc. Natl. Med., 172, 214–218. Acad. Sci. USA, 92, 10457–10461. Friedlander M, Theesfeld CL, Sugita M, Fruttiger M, Thomas Alon T, Hemo I, Itin A, Pe’er J, Stone J and Keshet E. (1995). MA, Chang S and Cheresh DA. (1996). Proc. Natl. Acad. Nat. Med., 1, 1024–1028. Sci. USA, 93, 9764–9769. Ando A, Mori K, Yamada H, Yamada E, Takahashi K, Saikia Gehlbach P, Demetriades AM, Yamamoto S, Deering T, Duh J, Yang A, Kim M and Campochiaro PA. (2001). J. Cell. EJ, Yang HS, Cingolani C, Lai H, Wei L and Campochiaro Physiol., 191, 116–124. PA. (2003a). Gene Therapy, 10, 637–646. Ando A, Yang A, Nambu H and Campochiaro PA. (2002). Gehlbach P, Demetriades AM, Yamamoto S, Deering T, Xiao Mol. Pharmacol., 62, 539–544. WH, Duh EJ, Yang HS, Lai H, Kovesdi I, Wei L Beck H, Acker T, Wiessner C, Allegrini PR and Plate KH. and Campochiaro PA. (2003b). Hum. Gene Ther., 14, (2000). Am. J. Pathol., 157, 1473–1483. 129–141. Brooks P, Clark R and Cheresh D. (1994a). Science, 264, Good DJ, Polverini PJ and Rastinejad F. et al. (1990). Proc. 569–571. Natl. Acad. Sci. USA, 87, 6624–6628. Brooks PC, Montgomery AM, Rosenfeld M, Reisfeld RA, Gossen M, Freundlieb S, Bender G, Muller G, Hillen W and Hu T, Klier G and Cheresh D. (1994b). Cell, 79, Bujard H. (1995). Science, 268, 1766–1769. 1157–1164. Grosios K, Holwell SE, McGown AT, Pettit GR and Bibby Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, MC. (1999). Br. J. Canc., 81, 1318–1327. Dvorak JF and Senger DR. (1993). Am. J. Pathol., 143, Macular Photocoagulation Study Group (1991). Arch. 1255–1262. Ophthalmol., 109, 1109–1114. Carmeliet P, Moons L, Luttun A, Vincenti V, Compernolle V, Gutheil JC, Campbell TN, Pierce PR, Watkins JD, Huse WD, De Mol M, Wu Y, Bono F, Devy L, Beck H, Scholz D, Bodkin DJ and Cheresh DA. (2000). Clin. Cancer Res., 6, Acker T, DiPalma T, Dewechin M, Noel A, Stalmans I, 3056–3061. Barra A, Blacher S, Vandendriessche T, Ponten A, Eriksson Hammes H, Brownlee M, Jonczyk A, Sutter A and Preissner U, Plate KH, Foidart JM, Schaper W, Charnock-Jones DS, K. (1996). Nat. Med., 2, 529–533. Hicklin DJ, Herbert JM, Collen D and Persico MG. (2001). Hangai M, Moon YS, Kitaya N, Chan CK, Wu D-Y, Peters Nat. Med., 7, 575–583. KG, Ryan SJ and Hinton DR. (2001). Hum. Gene Ther., 12, Carmeliet P, Ng Y-S, Nuyens D, Theilmeier G, Brusselmans 1311–1321. K, Cornelissen I, Ehler E, Kakkar VV, Stalmans I, Mattot Hayes AJ, Huang WQ, Mallah J, Yang D, Lippman ME and V, Perriard J-C, Dewerchin M, Flameng W, Nagy A, Lupu Li LY. (1999). Microvasc Res., 58, 224–237. F, Moons L, Collen D, D’Amore PA and Shima DT. (1999). Hazzard TM, Christenson LK and Stouffer RL. (2000). Mol. Nat. Med., 5, 495–502. Hum. Reprod., 6, 993–998. Chaplin DJ, Pettit GR and Hill SA. (1999). Anticancer Res, 19, Hill SA, Sonergan SJ, Denekamp J and Chaplin DJ. (1993). 189–196. Eur. J. Cancer, 29A, 1320–1324. Dark GD, Hill SA, Prise VE, Tozer GM, Pettit GR and Honda M, Sakamoto T, Ishibashi T, Inomata H and Ueno H. Chaplin DJ. (1997). Cancer Res, 57, 1829–1834. (2000). Gene Therapy, 7, 978–985. Das A, McLamore A, Song W and McGuire PO. (1999). Arch. Horsman M, Ehrnrooth E, Ladekarl M and Overgaard J. Ophthalmol., 117, 498–503. (1998). Int. J. Radial. Oncol. Biol. Phys., 42, 895–898. Davel LE, Miguez MM and de Lustig ES. (1985). Transplan- Hynes RO. (2002). Nat. Med., 8, 918–921. tation, 39, 564–565. Janzer RC and Raff MC. (1987). Nature, 325, 253–257. Davis S, Aldrich TH, Jones P, Acheson A, Ryan TE, Bruno J, Jones MK, Wang H, Peskar BM, Levin E, Itani RM, Radziejewski C, Maisonpierre PC and Yancopoulos GD. Sarfeh IJ and Tarnawski AS. (1999). Nat. Med., 5, (1996). Cell, 87, 1161–1169. 1418–1423. Dawson DW, Volpert OV, Gillis P, Crawford SE, Joukov V, Pajusola K, Kaipainen K, Chilov D, Lahtinen I, Xu H-J, Benedict W and Bouck NP. (1999). Science, 285, Kukk E, Saksela O, Kalkkinen N and Alitalo K. (1996). 245–248. EMBO J, 15, 290–298. Deutsch TA and Hughes WF. (1979). Am. J. Ophthalmol., 87, Kawamori T, Rao CV, Seibert K and Reddy BS. (1998). 536–540. Cancer Res, 58, 409–412. Diaz-Flores L, Gutierrez R, Valladares F, Varela H and Perez Kim I, Kim HG, Moon SO, Chae SW, So JN, Koh KN, Ahn M. (1994). Anat. Rec., 238, 68–76. BC and Koh GY. (2000a). Circ. Res., 86, 952–959. Dowlati A, Robertson K, Cooney M, Petros WP, Stratford M, Kim I, Kim HG, So J-N, Kim JH, Kwak HJ and Koh GY. Jesberger J, Raﬁe N, Overnoyer B, Makkar V, Stambler B, (2000b). Circ. Res., 86, 24–29. Taylor A, Waas J, Lewin JS, McCrae KR and Remick SC. Klein R, Klein BEK, Moss SE, Davis MD and DeMets DL. (2002). Cancer Res, 62, 3408–3416. (1984). Arch. Ophthalmol., 102, 520–526.

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6547 Koblizek TI, Weiss C, Yancopoulos GD, Deutsch U and Mori K, Gehlbach P, Yamamoto S, Duh E, Zack DJ, Li Q, Risau W. (1998). Curr. Biol., 8, 529–532. Berns KI, Raisler BJ, Hauswirth WW and Campochiaro PA. Kukk E, Lymboussaki A, Taira S, Kaipainen A, Jeltsch M, (2002). Invest. Ophthalmol. Vis. Sci., 43, 1994–2000. Joukov V and Alitalo K. (1996). Development, 122, 3829– Nambu H, Nambu R, Melia M and Campochiaro PA. (2003). 3837. Invest. Ophthalmol. Vis. Sci. (in press). Kwak HJ, So J-H, Lee SJ, Kim I and Koh GY. (1999). FEBS Nicol D, Hii SI, Walsh M, Teh B, Thompson L, Kennett C Lett, 448, 249–253. and Gotley D. (1997). J. Urol., 157, 1482–1486. Kwak N, Okamoto N, Wood JM and Campochiaro PA. Ohno-Matsui K, Hirose A, Yamamoto S, Saikia J, Okamoto (2000). Onvest. Ophthalmol. Vis. Sci., 41, 3158–3164. N, Gehlbach P, Duh EJ, Hackett SF, Chang M, Bok D, Lai C-C, Wu W-C, Chen S-L, Xiao X, Tsai T-C, Huan S-J, Zack DJ and Campochiaro PA. (2002). Am. J. Pathol., 160, Chen T-L, Tsai RJ-F and Tsao Y-P. (2001a). Invest. 711–719. Ophthalmol. Vis. Sci., 42, 2401–2407. Okamoto N, Tobe T, Hackett SF, Ozaki H, Vinores MA, Lai C-M, Brankov M, Zaknich T, Lai YK-Y, Shen W-Y, LaRochelle W, Zack DJ and Campochiaro PA. (1997). Am. Constable IJ, Kovesdi I and Rakoczy PE. (2001b). Hum. J. Pathol., 151, 281–291. Gene Ther., 12, 1299–1310. Olofsson B, Pajusola K, Kaipainen A, Von Euler G, Joukov V, Lambert V, Munaut C, Jost M, Noel A, Werb Z, Foidart JM Saksel O, Orpana A, Pettersson RF, Alitalo K and and Rakic J-M. (2002). Am. J. Pathol., 161, 1247–1253. Eriksson U. (1995). Proc. Natil. Acad. Sci. USA, 93, Lambert V, Munaut C, Noel A, Frankenne F, Bajou K, 2576–2581. Gerard R, Carmeliet P, Defresne MP, Foidart J-M and O’Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane Rakic J-M. (2001). FASEB J, 15, 1021–1027. WS, Flynn E, Birknead JR, Olsen BR and Folkman J. Lee J, Gray A, Yuan J, Luoh SM, Avraham H and Wood WI. (1997). Cell, 88, 277–285. (1996). Proc. Patil. Acad. Sci. USA, 93, 1988–1992. O’Reilly MS, Holmgren S, Shing Y, Chen C, Rosenthal RA, Lindahl P, Johansson BR, Leveen P and Betsholtz C. (1997). Moses M, Lane WS, Cao Y, Sage HE and Folkman J. Science, 277, 242–245. (1994). Cell, 79, 315–328. Luna J, Tobe T, Mousa SA, Reilly TM and Campochiaro PA. O’Reilly MS, Pirie-Sheherd S, Lane WS and Folkman J. (1996). Lab. Invest., 75, 563–573. (1999). Science, 285, 1926–1928. Luttun A, Tjwa M, Moons L, Wu Y, Angelillo-Scherrer A, Oshima Y, Oshima S, Takahashi K, Nambu H, Apte RS, Duh Liao F, Nagy JA, Hooper A, Priller J, De Klerck B, E, Hackett SF, Okoye G, Melia M, Wiegard S, Yancopoulos Compernolle V, Daci E, Bohlen P, Dewerchin M, Herbert G, Zack DJ and Campochiaro PA. (2003). Different effects JM, Fava R, Matthys P, Carmeliet G, Collen D, Dvorak of angiopoietin 2 in different vascular beds; new vessels are HF, Hicklin DJ and Carmeliet PA. (2002). Nat. Med., 8, most sensitive. Manuscript submitted. 831–839. Ozaki H, Hayashi H, Vinores SA, Moromizato Y, Campo- Maglione D, Guerriero V, Viglietto G, Delli-Bovi P and chiaro PA and Oshima K. (1997). Exp. Eye Res., 64, 505– Persico MG. (1991). Proc. Natl. Acad. Sci. USA, 88, 9267– 517. 9271. Ozaki H, Okamoto N, Ortega S, Chang M, Ozaki K, Sadda S, Maione TE, Gray GS, Petro J, Hunt AJ, Donner AL, Vinores MA, Derevjanik N, Zack DJ, Basilico C and Bauer SI, Carson HF and Sharpe RJ. (1990). Science, 247, Campochiaro PA. (1998). Am. J. Pathol., 153, 757–765. 77–79. Ozaki H, Seo M-S, Ozaki K, Yamada H, Yamada E, Maisonpierre PC, Suri C, Jones PF, Bartunkova S, Wiegand Hofmann F, Wood J and Campochiaro PA. (2000). Am. J. SJ, Radziejewski C, Compton D, McClain J, Aldrich TH, Pathol., 156, 679–707. Papadopoulos N, Daly TJ, Davis S, Sato TN and Papapetropoulos A, Garcia-Cardena G, Dengler TJ, Maison- Yancopoulos GD. (1997). Science, 277, 55–60. pierre PC, Yancopoulos GD and Sessa WC. (1999). Lab. Masferrer JL, Leahy KM, Koki AT, Zweifel BS, Invest., 79, 213–223. Settle SL, Woerner BM, Edwards DA, Flickinger AG, Park JE, Keller G-A and Ferrara N. (1993). Mol. Biol. Cell, 4, Moore RJ and Seibert K. (2000). Cancer Res, 60, 1317–1326. 1306–1311. Pettit GR, Singh SB, Hamel E, Lin CM, Alberts DS and Meyer M, Causs M, Lepple-Wienhues A, Waltengerger J, Garia-Kendall D. (1989). Experientia, 45, 205–211. Augustin HG, Ziche M, Lanz C, Buttner M, Rziha H-J and Pettit GR, Temple C, Narayanan VL, Varma R, Simpson MJ, Dehio C. (1999). EMBO J, 18, 363–374. Boyd MR, Rener GA and Banasal N. (1995). Anti-Cancer Miller H, Miller B and Ryan SJ. (1986). Invest. Ophthalmol. Drug Des, 10, 299–309. Vis. Sci., 27, 1644–1652. Pichiule P and LaManna JC. (2002). J. Appl. Physiol., 93, Mordenti J, Cuthbertson RA, Ferrara N, Thomsen K, Berleau 1131–1139. LT, Licko V, Allen PC, Valverde CR, Meng YG, Fei DT, Plate KH, Breier G, Welch HA and Risau W. (1992). Nature, Fourre KM and Ryan AM. (1999). Toxicol. Pathol., 27, 359, 845–848. 536–544. Posey JA, Khazaeli MB, DelGrosso A, Saleh MN, Lin CY, Mori K, Ando A, Gehlbach P, Nesbitt D, Takahashi K, Huse WD and LoBuglio AF. (2001). Canc. Biother. Radio- Goldsteen D, Penn M, Chen T, Mori K, Melia M, Phipps S, pharm., 16, 125–132. Moffat D, Brazzell K, Liau G, Dixon KH and Campochiaro Procopio WN, Pelavin PI, Lee WM and Yeilding NM. (1999). PA. (2001a). Am. J. Pathol., 159, 313–320. J. Biol. Chem., 274, 30196–30201. Mori K, Duh E, Gehlbach P, Ando A, Takahashi K, Pearlman Puri MC, Rossant J, Alitalo K, Bernstein A and Partanen J. J, Mori K, Yang HS, Zack DJ, Ettyreddy D, Brough DE, (1995). EMBO J., 14, 5884–5891. Wei LL and Campochiaro PA. (2001b). J. Cell. Physiol., Raisler BJ, Berns KI, Grant MB, Beliaev D and Hauswirth 188, 253–263. WW. (2002). Proc. Natl. Acad. Sci. USA, 99, 8909–8914. Mori K, Gehlbach P, Ando A, McVey D, Wei L and Rasmussen HS, Chu KW, Campochiaro P, Gehlbach P, Haller Campochiaro PA. (2001c). Invest. Ophthalmol. Vis. Sci., JA, Handa JT, Nguyen QD and Sung JU. (2001). Hum. Gene 43, 2428–2434. Ther., 12, 2029–2032.

Oncogene Ocular neovascularization PA Campochiaro and SF Hackett 6548 Reichelt M, Shi S, Hayes M, Kay G, Batch J, Gole GA and Takahashi K, Saishin Y, Saishin Y, Mori K, Ando A, Browning J. (2003). Clin. Exp. Ophthalmol., 31, 61–65. Yamamoto S, Oshima Y, Nambu H, Melia MB, Bingaman Saishin Y, Saishin Y, Takahashi K, Lima Silva R, Hylton D, DP and Campochiaro PA. (2003b). Invest. Ophthalmol. Vis. Rudge JS, Wiegard SJ and Campochiaro PA. (2003). J. Cell. Sci., 44, 409–415. Physiol., 195, 241–248. Takahashi T, Nakamura T, Hayashi A, Kamei M, Nakabaya- Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, shi M, Okada AA, Tomita N, Daneda Y and Tano Y. Fujiwara Y, Gendron-Maguire M, Gridley T, Wolburg H, (2000). Am. J. Ophthalmol., 130, 774–781. Risau W and Qin Y. (1995). Nature, 376, 70–74. Takano S, Toshii Y, Kondo S, Suzuki H, Maruno T, Shirai S Sawaoka H, Tsujii S, Tsujii M, Gunawan ES, and Nose T. (1996). Cancer Res, 56, 2185–2190. Sasaki Y, Kawano S and Hori M. (1999). Lab. Invest., 79, Thurston G, Rudge JS, Ioffe E, Zhou H, Ross L, Croll SD, 1469–1477. Glazer N, Holash J, McDonald DM and Yancopoulos GD. Schwartz SD, Blumenkranz M, Rosenfeld PJ, Miller JW, (2000). Nat. Med., 6, 460–463. Haller J, Fish G, Lobes L, Singerman L, Green WL and Thurston G, Suri C, Smith K, McClain J, Sato TN, Reimann J. (2001). Invest. Ophthalmol. Vis. Sci., 42, Suppl Yancopoulos GD and McDonald DM. (1999). Science, S522. 286, 2511–2515. Seed S, Slaughter DP and Limarzi LR. (1940). Surgery, 7, 696– Tobe T, Ortega S, Luna L, Ozaki H, Okamoto N, Derevjanik 709. NL, Vinores SA, Basilico C and Campochiaro PA. (1998). Seo M-S, Kwak N, Ozaki H, Yamada H, Okamoto N, Fabbro Am. J. Pathol., 153, 1641–1646. D, Hofmann F, Wood JM and Campochiaro PA. (1999). Tolentino MJ, Miller JW, Gragoudas ES, Jakobiec FA, Flynn Am. J. Pathol., 154, 1743–1753. E, Chatzistefanou K, Ferrara N and Adamis AP. (1996). Shweiki D, Itin A, Soffer D and Keshet E. (1992). Nature, 359, Ophthalmology, 103, 1820–1828. 843–845. Tombran-Tink J, Chader GG and Johnson SV. (1991). Exp. Smalley W and DuBois RN. (1997). Adv. Pharmacol., 39, 1–20. Eye Res., 53, 411–414. Smith LEH, Kopchick JJ, Chen W, Knapp J, Kinose F, Daley Tozer GM, Prise VE, Wilson J, Locke RJ, Vojnovic B, D, Foley E, Smith RG and Schaeffer JM. (1997). Science, Stratford MRL, Dennis MF and Chaplin DJ. (1999). Cancer 276, 1706–1709. Res, 59, 1626–1634. Smith LEH, Wesolowski E, McLellan A, Kostyk SK, Tsujii M, Kawano S, Tsuji S, Sawaoka H, Hori M and DuBois D’Amato R, Sullivan R and D’Amore PA. (1994). Invest. RN. (1998). Cell, 93, 705–716. Ophthalmol. Vis. Sci., 35, 101–111. Visconti RP, Richardson CD and Sato TN. (2002). Proc. Stalmans I, Ng Y-S, Rohan R, Fruttiger M, Bouche A, Yuce Natil. Acad. Sci. USA, 99, 8219–8224. A, Fujisawa H, Hermans B, Sahni M, Jansen S, Hicklin D, Watanabe T and Raff MC. (1988). Nature, 332, 834–837. Anderson DJ, Gardiner T, Hammes H-P, Moons L, Williams CS, Mann M and DuBois RN. (1999). Oncogene, 18, Dewerchin M, Collen D, Carmeliet P and D’Amore PA. 7908–7916. (2002). J. Clin. Invest., 109, 327–336. Williams CS, Tsujii M, Reese J, Dey SK and DuBois RN. Steele FR, Chader GJ, Johnson LV and Tombran-Tink J. (2000). J. Clin. Invest., 105, 1589–1594. (1993). Proc. Natl. Acad. Sci. USA, 90, 1526–1530. Witzenbichler B, Maisonpierre PC, Jones P, Yancopoulos GD Stone J, Itin A, Alon T, Pe’er J, Gnessin H, Chan-Ling T and and Isner JM. (1998). J. Biol. Chem., 213, 18514–18521. Keshet E. (1995). J. Neurosci., 15, 4738–4747. Woods JA, Hatﬁeld JA, Pettit GR, Fox BW and McGown Suri C, Jones PP, Patan S, Bartunkova S, Maisonpierre PC, AT. (1995). Br. J. Cancer, 71, 705–711. Davis S, Sato TN and Yancopoulos GD. (1996). Cell, 87, Wright AD, Kohner EM, Oakley NW, Hartog M, Joplin GF 1171–1180. and Fraser TR. (1969). Br. Med. J., 2, 346–348. Suri C, McClain J, Thurston G, McDonald DM, Zhou H, Yamada E, Tobe T, Yamada H, Okamoto N, Zack DJ, Werb Oldmixon EH, Sato TN and Yancopoulos GD. (1998). Z, Soloway P and Campochiaro PA. (2001). Histol. Science, 282, 468–471. Histopathol., 16, 87–97. Takahashi A, Sasaki H, Kim SJ, Tobisu D, Kakizoe T, Yamada H, Yamada E, Kwak N, Ando A, Suzuki A, Esumi Tsukamoto T, Kumamoto Y, Sugimura T and Terada M. N, Zack DJ and Campochiaro PA. (2000). J. Cell. Physiol., (1994). Cancer Res, 54, 4233–4237. 185, 135–142. Takahashi K, Saishin Y, Saishin Y, Lima Silva R, Oshima Y, Yamada M, Kawai M, Kawai Y and Mashima Y. (1999). Oshima S, Melia M, Paszkiet B, Zerby D, Kadan MJ, Liau Curr. Eye Res., 19, 300–304. G, Kaleko M, Connelly S, Luo T and Campochiaro PA. Ziche M, Jones J and Gullino PM. (1982). J. Natl. Cancer (2003a). FASEB J., 17, 876–898. Inst., 69, 475–482.

Oncogene