Introduction

Vascular endothelial growth factor (VEGF) is not only an important regulator of physiological angiogenesis but also involved in pathological neovascularization.¹ Formation of new blood vessels caused by the overproduction of growth factors such as VEGF is a key component of diseases such as wet age-related macular degeneration (AMD), proliferative diabetic retinopathy (PDR) and tumor growth.^{2, 3, 4} Blocking of VEGF with antibodies, soluble VEGF receptors (sVEGFR) or inhibition of VEGF receptor (VEGFR) tyrosine kinase activity are useful strategies that have shown promising preclinical and clinical results.^{4, 5, 6} Ranibizumab, an anti-VEGF drug given intravitreally to wet-AMD patients results in substantially improved visual acuity.⁷ Intraocular gene delivery of VEGF antagonists could have theoretical advantages over the current treatment, which requires monthly intravitreal injections (for years) by a retinal specialist.

The angiogenic effect of VEGF is mediated predominantly by its binding to VEGFR KDR.⁸ Another high-affinity VEGFR, Flt-1, binds VEGF about 10 times stronger than KDR, however Flt-1 activation does not significantly stimulate angiogenesis.^{9, 10} Both receptors, Flt-1 and KDR, have similar structure with extracellular regions consisting of seven domains, 1–7.^{10, 11} There exists another naturally occurring soluble form of Flt-1 (sFlt-1) that contains only the extracellular domains and has the same VEGF-binding affinity as full-length Flt-1.^{11, 12} The VEGF-binding function of Flt-1 has been mapped to the second domain.^{13, 14, 15, 16} There have been previous studies with two truncated soluble receptor hybrids, Flt(1–3)-IgG and Flt(1–7)-IgG, consisting of either the first three domains or all seven domains fused to human IgG1-Fc region.¹³ The molecule Flt(1–3)-IgG was reported to have the same VEGF-binding affinity as Flt(1–7)-IgG, however Flt(2)-IgG that contains only second domain was not capable of inhibiting VEGF.^{13, 17} Another molecule called VEGF-Trap, generated by the second domain of Flt-1 fused to the third domain of KDR and human IgG1-Fc region, has been shown to be a very potent VEGF binder.¹⁸

Adeno-associated virus (AAV) vectors offer an attractive tool for intraocular gene delivery because of their nonpathogenic nature, low toxicity, minimal immunogenicity and long-term persistence.^{19, 20, 21, 22} Intravitreal administration of AAV serotype 2 (AAV2) vector in mice results mostly in transduction of ganglion cells and few cells in the inner nuclear layer.²¹ Subretinal delivery of AAV2 vector encoding full-length sFlt-1 (transduction of photoreceptors and retinal pigmented epithelium) prevented development of laser-induced choroidal neovascularization in all treated monkeys.²²

In this study we have designed and constructed small novel soluble hybrid molecules that have strong anti-VEGF activity in vitro, incorporated these molecules into AAV2 vectors and then tested in vivo persistence of expression and efficacy and safety following intravitreal delivery of these AAV2-based vectors in the oxygen-induced retinopathy of prematurity (OIR) mouse model.

Results

sFLT01 molecule inhibits VEGF-mediated cell proliferation

We first constructed molecules D1-3, sFLT01 and D2-Fc (Figure 1a). The molecule D1-3 has served a positive control for VEGF binding and contains first three Flt-1 domains. The second molecule, sFLT01, contains only one Flt-1 domain 2 linked by 9Gly to the human IgG1 heavy-chain Fc region that resulted in the generation of a forced homodimer. The formation of sFLT01 dimers was confirmed by western blot analysis (Figure 2b). The third molecule, D2-Fc, is identical to sFLT01, except it does not contain 9Gly linker. Plasmids encoding these molecules under control of a cytomegalovirus (CMV) promoter were used for transfection of 293 cells and the conditioned media (CM) were evaluated for their ability to block VEGF-stimulated human umbilical vein endothelial cells (HUVECs) proliferation (Figure 1b). The molecule D2-Fc was not able to neutralize VEGF and inhibit HUVECs proliferation whereas sFLT01 demonstrated an efficient inhibition of VEGF-dependent HUVECs proliferation (Figure 1b).

Figure 1.

Generation of novel soluble vascular endothelial growth factor (VEGF) receptor hybrid molecules. (a) Schematics of molecules D1-3, D2-Fc and sFLT01. The white blocks indicate signal peptides (sp) and Flt-1 domains (D1, D2, D3); black block represents 9Gly linker; grey blocks represent domains CH2 of human IgG1-Fc region. (b) Inhibitory effect of conditioned media containing molecules D1-3, D2-Fc and sFLT01 on VEGF-induced human umbilical vein endothelial cells (HUVECs) proliferation. Recombinant full-size rhFlt-1 protein (50 ng ml^-1) was used as a positive control. The same lot of HUVECs was used in this set of assays. Data are expressed as means.d. where 'n' represents the number of independent proliferation experiments (for sFLT01 and D2-Fc, n=3; for D1-3, n=2). Student's non-paired t-test; ^**P<0.005 for difference between '+VEGF control' and 'VEGF binder'; D1-3 (n=2) was not included in statistics.

Full figure and legend (60K)

Figure 2.

Novel hybrid receptor molecules containing the human IgG1 CH3 domain. (a) Schematics of molecules sFLT02 and D2-CH3 as compared to sFLT01. The white blocks indicate signal peptides (sp) and Flt-1 domain 2 (D2); black block represents 9Gly linker; shaded blocks represent domain CH3 of human IgG1-Fc region. (b) Western Blot comparing migration of sFLT01, sFLT02 and D2-CH3 (c) Inhibitory effect of molecules sFLT01 and sFLT02 on human umbilical vein endothelial cells (HUVECs) proliferation. A different lot of HUVECs was used for this set of proliferation assays. Data are expressed as means.d. of three independent proliferation experiments (n=3). Student's non-paired t-test; ^*P<0.05 for difference between '+VEGF control' and 'VEGF binder'.

Full figure and legend (72K)

IgG1 CH3 domain and peptide linker are important in VEGF binding

Using the sFLT01 as a parental molecule, we generated several mutant derivatives with various deletions along the IgG1-Fc region. When comparing several constructs with various deletions either in CH2 or in CH3 domains of IgG1, we have observed that the CH3 domain is involved in preserving the VEGF-binding function (data not shown). The molecule sFLT02, where the second domain of Flt-1 was linked to the IgG1 CH3 domain through 9Gly (Figures 2a and b), retained the VEGF-binding function (Figure 2c). This set of proliferation assays was performed using a different lot of HUVECs (not the same lot as used in Figure 1b) and it demonstrates that sFLT01 has comparable VEGF blocking potency as in the first set of proliferation assays. The molecule sFLT02 was similar to the parental molecule sFLT01 in neutralizing VEGF (Figure 2c). On the other hand, the molecule D2-CH3 (Figures 2a and b) without 9Gly linker appeared to be a very weak inhibitor of VEGF-induced HUVECs proliferation (Figure 2c), even though VEGFR1 (Flt-1) enzyme-linked immunosorbent assay (ELISA) assay of CM from transfected 293 cells showed high concentrations of D2-CH3 protein (150 ng ml^-1) compared to sFLT02 (70–90 ng ml^-1). Hence we conclude that the presence of both the CH3 domain and the peptide linker facilitates VEGF binding. Western blot analyses of sFLT02 and D2-CH3 (Figure 2b) show a prevalence of the dimeric forms for both proteins under nonreducing conditions ruling out the possibility that the lack of efficacy of D2-CH3 is due to its inability to dimerize.

For the next construct we used another type of linker, the 15-mer (Gly₄Ser)₃.²³ D2-(Gly₄Ser)₃-Fc protein was generated and it contained Flt-1 domain 2, (Gly₄Ser)₃ linker and the IgG1 Fc. The molecule D2-(Gly₄Ser)₃-Fc was further characterized in HUVECs proliferation assay. Biological activity of D2-(Gly₄Ser)₃-Fc as measured by inhibition of HUVEC proliferation was similar to that of sFLT01 (data not shown).

sFLT01 binds VEGF better than the other novel constructs

The relative binding affinity between VEGF and our novel molecules was determined using a cell-free assay system. CM containing known concentrations of sVEGFR (ranging from 0.29 to 150 pM) were serially diluted and mixed with VEGF (10 pM final concentration). The amount of unbound VEGF was then measured by a human VEGF-specific ELISA. sFLT01 binds VEGF with higher affinity than D1-3, sFLT02 and D2-CH3, where the latter showed a minimal VEGF-binding affinity (Figure 3). These data are in agreement with our HUVECs proliferation assays shown in Figures 1 and 2.

Figure 3.

Vascular endothelial growth factor (VEGF)-binding affinities of VEGF soluble receptors. Conditioned media from independent transfections of 293 cells were used in this cell-free binding assay. Increasing concentrations of soluble VEGF receptors molecules (x axis) were incubated overnight with 10 pM of human VEGF165. The amount of unbound VEGF (y axis) was measured by a human VEGF-specific enzyme-linked immunosorbent assay (ELISA) in triplicate. The number of independent transfections (n) for each molecule are: sFLT01 (n=5), D1-3 (n=3), sFLT02 (n=2) and D2-CH3 (n=1). Data are expressed as means.d.; Student's non-paired t-test; ^***P<0.0001, ^**P<0.001 for differences between sFLT01 and D1-3. LOQ, limit of quantitation=0.371pM.

Full figure and legend (48K)

AAV2.sFLT01 transgene and mRNA localization

AAV2-based vectors were constructed encoding sFLT01 and adult C57BL/6 mice were intravitreally injected with 2.2 10⁹ drps (DNase resistant particles) of AAV2.sFLT01. The animals were killed 28 days following injection and the eyes were processed for histological sectioning. The presence of sFLT01 protein and mRNA was confirmed by immunofluorescence and by in situ hybridization. sFLT01 protein was predominantly observed in the retinal ganglion cells of mouse retinas (Figure 4a). Infrequently, cells deeper in the retina (presumably Müller cells) were also shown to contain the sFLT01 protein. Another set of retinas was examined using in situ hybridization to detect the mRNA of the sFLT01 transgene. The retinal ganglion cells were the predominant cell type transduced in the mouse retina (Figure 4c). No sFLT01 protein was detected in the uninjected control eyes. No sFLT01 mRNA message was detected in the AAV2.sFLT01 injected eyes when sense probe was used.

Figure 4.

Detection of sFLT01 protein and mRNA in the retina. sFLT01 was detected in retina 28 day after intravitreal injection with 2.2 10⁹ drps of AAV2.sFLT01 as (a) a protein by immunofluorescence or (c) as mRNA by in situ hybridization. Control in (b) represents the uninjected control eyes. Control in (d) represents the AAV2.sFLT01 injected eyes hybridized with sense probe. RGC: retinal ganglion cells; IPL, OPL: inner and outer plexiform layers; INL, ONL: inner and outer nuclear layers; PR: photoreceptors.

Full figure and legend (511K)

Longevity of transgene expression

Adult C57BL/6 mice were intravitreally injected with 1 10⁹ drps of AAV2.sFLT01. Animals were killed at predetermined time points and the level of sFLT01 was measured in individual eyes (homogenates of retina and vitreous humor) by ELISA against human Flt-1. The samples from each time point were assayed separately for animals treated with AAV2.sFLT01 (Figure 5). The data suggest that AAV2-mediated delivery of sFLT01 results in long-term stable protein expression in the murine eye with no transgene-related toxicities observed. The apparent rise in sFLT01 in the 12-month AAV2.sFLT01 cohort was due to a change in the tissue homogenization protocol.

Figure 5.

Longevity of sFLT01 expression. Persistent transgene expression in excess of 1 year was observed in adult C57BL/6 mice treated with a single intravitreal injection of AAV2.sFLT01 (1 10⁹ drps). The amount of sFLT01 was quantified in individual retinal homogenates by enzyme-linked immunosorbent assay (ELISA) against human Flt-1. LOQ, limit of quantitation.

Full figure and legend (38K)

AAV2.sFLT01 intravitreal delivery inhibits angiogenesis in OIR model

The efficacy of AAV2.sFLT01 was examined in vivo using OIR mouse model.²⁴ AAV2.sFLT01 was intravitreally administered into left eyes of neonatal C57BL/6 mice in a dose of 1 10⁹ drps per animal. Data shown in Figure 6 were expressed as percentage of neovascularization of untreated eyes in the AAV2.sFLT01 group (n=43), or between left and right eyes in the control group (n=26). The occurrence of neovascularization was significantly reduced to 5447% in AAV2.sFLT01-treated eyes (AAV1.sFLT01 group) as compared to 10266% between left and right eyes in the control group (means.d.). Treatment with AAV2.sFLT01 significantly reduced ocular neovascularization (Student's t-test; P<0.0009) compared to the untreated contralateral eyes.

Figure 6.

Adeno-associated virus (AAV)-mediated delivery of sFLT01 inhibits retinal neovascularization in the murine oxygen-induced retinopathy (OIR) model. The eyes of newborn mice were treated with a single intravitreal injection of 1.1 10⁹ drps of AAV2.sFLT01 vector into the left eye. The number of endothelial cell nuclei internal to the inner limiting membrane in the treated (left) eye was compared to the contralateral (right) eye. Data in both groups, AAV.sFLT01 (n=43) and control (n=26), are expressed as percentage of neovascularization of untreated (right) eyes (means.d.). The difference was found to be statistically significant by Student's t-test (P<0.0009; as compared to the untreated contralateral eye).

Full figure and legend (10K)

Discussion

Many reports from preclinical and clinical studies demonstrate that antagonizing VEGF is a potentially useful strategy for treatment of pathological neovascularization which is a key component of ocular diseases like wet-AMD or PDR. In this study, we have developed several novel hybrid molecules that inhibit VEGF in vitro and are capable of inhibiting ocular neovascularization in vivo. It has been reported that domain 2 of VEGFR Flt-1 requires the presence of flanking sequences from Flt-1 or other VEGFR domains for efficient ligand binding and neutralization.^{12, 17} Our results show that domains other than second domain of Flt-1 were not necessary to preserve VEGFR/ligand binding. To obtain dimerization of the soluble receptor, we constructed Flt-1 domain 2 linked directly to IgG1-Fc fragment (molecule D2-Fc), but such a strategy did not enhance its binding to VEGF, in agreement with Davis-Smyth et al.¹³ We then have linked the second domain of Flt-1 to Fc through a 9Gly linker, and thus generated a novel molecule, sFLT01, with potent VEGF binding. Our results show, for the first time, that Flt-1 domain 2 does not require presence of other VEGFR domains for high-affinity VEGF binding. Several other constructs where 9Gly was replaced with other short linkers like 15-mer (Gly₄Ser)₃, polyglycine pentamer (5Gly) or other random linkers showed VEGF binding comparable to sFLT01.

We also have constructed several other hybrid proteins by deleting selected regions of IgG1 Fc. Although deletion of the entire Fc CH2 domain did not have a significant impact, we find that the CH3 domain is important for VEGF binding of our molecules. The molecule sFLT02, which retains VEGF binding, was generated by the complete deletion of the Fc CH2 domain, such that Flt-1 domain 2 is linked directly to the Fc CH3 domain through the 9Gly linker.

From the variety of gene therapy vectors currently available we first decided to use the AAV vector because of its versatility, safety and long-term in vivo persistence.^{25, 26} There are several routes to deliver gene therapy vectors into the eye, however the most common ones are intravitreal and subretinal.^{19, 21} We first decided to investigate an intravitreal route of delivery because of its potential ease of use in the retinal clinic. The AAV2.sFLT01 vector predominantly transduced retinal ganglion cells in agreement with previous observation by other investigators.²¹ Subretinal delivery of AAV2 encoding full-length sFlt-1 gene has been successfully tested in mouse and in primate models of ocular neovascularization.^{22, 27, 28} An adenoviral vector encoding full-length sFlt-1, injected both intravitreously or periocularly suppressed choroidal neovascularization at rupture sites in Bruch's membrane.²⁹ Our AAV2.sFLT01 vector administered intravitreally to neonatal mice significantly reduced the occurrence of neovascularization in the OIR model. To our knowledge this is the first demonstration of long-term persistent expression of a secreted protein following intravitreal delivery of AAV2 to retinal ganglion cells. There were no gross histological observations of AAV2.sFLT01-related toxicity for the duration of this study (1 year). These results have significant implications for a potential human therapeutic that could be very infrequently administered by a simple intravitreal injection.

Materials and methods

Soluble VEGF receptor hybrids construction

DNAs encoding hybrid sVEGFR molecules were synthesized by DNA 2.0 Inc. (Menlo Park, CA, USA). The construct D1-3 contains the Flt-1 signal peptide sequence and the first three domains of human Flt-1. In all other constructs, the Flt-1 signal peptide sequence was fused directly to Flt-1 domain 2. In molecules D2-Fc, the Flt-1 domain 2 is fused directly to human IgG1-Fc region. The molecule sFLT01 has Flt-1 domain 2 fused to human IgG1-Fc region through the polyglycine 9-mer (9Gly) linker. In molecules sFLT02 and D2-CH3, a shorter fragment of IgG1, domain CH3, was used instead of using the full size of IgG1-Fc region. For initial testing, all hybrid molecules were produced by plasmid transfection of 293 cells followed by harvesting of CM, where the transgenes were under control of the CMV promoter.

HUVEC proliferation assay

Pooled HUVECs were purchased from Cambrex—Lonza (East Rutherford, NJ, USA) and expanded through four passages in EGM-2-MV media (EGM basal media supplemented with bovine brain extract, hEGF, hydrocortisone, gentamicin, amphotericin-B, 5% fetal bovine serum (FBS), VEGF, hFGF-B, R3-IGF-1 and ascorbic acid) according to the manufacturer's instructions. For proliferation assays, HUVECs were seeded at 2 10³ cells per well in a 96-well culture plate and incubated overnight in M199 starvation media (M199, 5% FBS). The following day, fresh M199 media supplemented with 10 ng ml^-1 recombinant human VEGF (R&D Systems, Minneapolis, MN, USA) and CM (5 l, approximately 1 ng ml^-1 of each) from transfected 293 cells containing sVEGFR molecules were added. HUVECs were incubated 3–4 days followed by the addition of the MTS reagent CellTiter 96 AQueous One Solution (Promega, Madison, WI, USA) and incubated for another 2–4 h. Absorbance was measured at OD490 on a VersaMax plate reader using SOFTmax PRO v4.5 (Molecular Devices, Sunnyvale, CA, USA). Data represent the means of independent proliferation experiments (means.d.) each assayed in triplicate. Each independent proliferation experiment used supernatants from an independent transfection.

Western blot analysis

CM from transfected 293 cells (15 l), containing proteins of hybrid sVEGFR molecules were analyzed by western blot under nonreduced and reduced conditions. Briefly, samples were separated by SDS-electrophoresis and transferred to polyvinylidene difluoride membrane. Blots were then probed with goat anti-human VEGFR1 horseradish peroxidase antibody conjugate (R&D Systems) followed by detection using ECL Western Blotting Detection Reagent (GE Healthcare Biosciences, Piscataway, NJ, USA).

VEGF-binding assay

Human VEGF (R&D Systems) adjusted to 20 pM in phosphate-buffered saline was mixed with an equal volume of increasing concentrations of sVEGFR molecules (0.001–10 000 pM; final VEGF concentration=10 pM) overnight at room temperature with gentle shaking. To determine relative binding affinities, samples were assayed for residual unbound VEGF using the Human Quantikine VEGF ELISA kit (R&D Systems), which detects only unbound (noncomplexed) VEGF. VEGF concentration (pM) was plotted as a function of increasing sVEGFR concentration (pM).

AAV vector

Synthetic sFLT01 chimeric transgene was cloned into a plasmid pCBA(2)-int-BGH, obtained from Mark Sands (Washington University Medical School, St Louis, MO, USA), which contains hybrid chicken -actin (CBA) promoter and bovine growth hormone polyadenylation signal sequence (BGH poly A).³⁰ The whole sFLT01 expression cassette was then cloned into a previral plasmid vector pAAVSP70 containing AAV2 inverted terminal repeats (ITRs).³¹ Total size of the resulting AAV genome in plasmid sp70.BR/sFLT01 including the region flanked by ITR was 4.6 kb.

The recombinant vector AAV2.sFLT01 was produced by triple transfection of 293 cells protocol using helper plasmids p5rep--CMVcap and pHelper (Stratagene, La Jolla, CA, USA), and purified according to the protocol using an iodixanol step gradient and HiTrap Heparin column (GE Healthcare Life Sciences, Piscataway, NJ, USA) on an ÄKTA FPLC system (GE Healthcare Life Sciences, Piscataway, NJ).^{32, 33} The AAV2.sFLT01 viral preparation had a titer of 2.2 10¹² drps (DNase resistant particles) per ml. Viral titers were determined using a real-time TaqMan PCR assay (ABI Prism 7700; Applied Biosystems, Foster City, CA, USA) with primers that were specific for the BGH poly A sequence.

Animals

Adult C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). The animals were maintained in Genzyme's AAALAC-accredited vivarium and given free access to food and water throughout the study. All procedures were performed under a protocol approved by the Institutional Animal Care and Use Committee.

Intravitreal injection and OIR model

For the longevity experiment, 9-week-old C57BL/6 mice were intravitreally injected with 1 l of AAV2.sFLT01 containing 1 10⁹ drps of vector and killed 12 months later. For the immunofluorescence and in situ hybridization experiments, adult C57BL/6 mice were intravitreally injected with 1 l of AAV2.sFLT01 containing 2.2 10⁹ drps of AAV2.sFLT01 and killed 28 days later. For the OIR model, neonatal C57BL/6 mouse pups were intravitreally injected on the day of birth (0) with 0.5 l of AAV2.sFLT01 containing 1.1 10⁹ drps of AAV2.sFLT01. The pups and their nursing dam were placed in an isobaric chamber and were exposed to a hyperoxic environment (75% oxygen) from day 7 to 12.²⁶ The animals from the OIR model were killed on day 17. The eyes of all animals were processed for paraffin embedding and were serially sectioned. A single 5-m section was taken at each 100-m level through the entire eye resulting in 10–20 sections to evaluate per eye. The sections were stained with hematoxylin and eosin and the degree of neovascularization was determined by counting the number of endothelial cell nuclei internal to and contiguous with the inner limiting membrane ignoring the region of the regressing hyloid vessels. The number of nuclei in the treated eye was compared to the number of nuclei in the contralateral control eye and the data were expressed as percentage of neovascularization of untreated eyes (means.d.).

Detection of transgene by immunofluorescence

The sFLT01 transgene was detected in eyes that were fixed in 10% neutral buffered formalin and embedded in paraffin. Tissue sections (5 m) were prepared and sFLT01 was detected using a goat anti-VEGFR1 primary antibody (R&D Systems) and a fluorescein isothiocyanate-conjugated rabbit anti-goat immunoglobulin G (IgG) secondary antibody (Invitrogen Corp., Carlsbad, CA, USA).

Detection of sFLT01 mRNA by in situ hybridization

The sFLT01 mRNA was detected in paraffin-embedded tissue sections by overnight hybridization of a DIG-labeled RNA probe specific for sFLT01. Signal was detected using peroxidase-labeled anti-DIG (Roche Applied Science, Indianapolis, IN, USA) and amplified using both biotinyl tyramide (Dako North America Inc., Carpinteria, CA) and alkaline phosphatase-labeled anti-biotin (Alpha Diagnostics International Inc., San Antonio, TX, USA). Fast Red (Dako North America Inc.) was then used to detect the complex, and the sections were counterstained using Mayer's hematoxylin (Dako North America Inc.).

References

1. Ferrara N, Henzel WJ. Pituitary follicular cells secrete a novel heparin-binding growth factor specific for vascular endothelial cells. Biochem Biophys Res Commun 1989; 161: 851–858. | Article | PubMed | ISI | ChemPort |
2. Kvanta A, Algvere PV, Berglin L, Seregard S. Subfoveal fibrovascular membranes in age-related macular degeneration express vascular endothelial growth factor. Invest Ophthalmol Visual Sci 1996; 37: 1929–1934. | ChemPort |
3. Adamis AP, Miller JW, Bernal MT, D'Amico DJ, Folkman J, Yeo TK et al. Increased vascular endothelial growth factor levels in the vitreous of eyes with proliferative diabetic retinopathy. Am J Ophthalmol 1994; 118: 445–450. | PubMed | ISI | ChemPort |
4. Ferrara N. Vascular endothelial growth factor: basic science and clinical progress. Endocr Rev 2004; 25: 581–611. | Article | PubMed | ISI | ChemPort |
5. Aiello LP, Pierce EA, Foley ED, Takagi H, Chen H, Riddle L et al. Suppression of retinal neovascularization in vivo by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF-receptor chimeric proteins. Proc Natl Acad Sci USA 1995; 92: 10457–10461. | Article | PubMed | ChemPort |
6. Willet CG, Boucher Y, di Tomaso E, Duda DG, Munn LL, Tong RT et al. Direct evidence that the VEGF-specific antibody bevicizumab has antivascular effects in human rectal cancer. Nat Med 2004; 10: 145–147. | Article | PubMed | ISI | ChemPort |
7. Rosenfeld PJ, Brown DM, Heier JS, Boyer DS, Kaiser PK, Chung CY et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 2006; 355: 1419–1431. | Article | PubMed | ISI | ChemPort |
8. Terman BI, Dougher-Vermazen M, Carrion ME, Dimitrov D, Armellino DC, Gospodarowicz D et al. Identification of the KDR tyrosine kinase as a receptor for vascular endothelial cell growth factor. Biochem Biophys Res Commun 1992; 187: 1579–1586. | Article | PubMed | ISI | ChemPort |
9. de Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N, Williams LT. The fms-like tyrosine kinase a receptor for vascular endothelial growth factor. Science 1992; 255: 989–991. | Article | PubMed | ISI | ChemPort |
10. Quinn TP, Peters KG, De Vries C, Ferrara N, Williams LT. Fetal liver kinase 1 is a receptor for vascular endothelial growth factor and is selectively expressed in vascular endothelium. Proc Natl Acad Sci USA 1993; 90: 7533–7537. | Article | PubMed | ChemPort |
11. Shibuya M, Yamaguchi S, Yamane A, Ikeda T, Tojo A, Matsushime H et al. Nucleotide sequence and expression of a novel human receptor-type tyrosine kinase gene (flt) closely related to the fms family. Oncogene 1990; 5: 519–524. | PubMed | ISI | ChemPort |
12. Kendall RL, Thomas KA. Inhibition of vascular endothelial cell growth factor activity by an endogenously encoded soluble receptor. Proc Natl Acad Sci USA 1993; 90: 10705–10709. | Article | PubMed | ChemPort |
13. Davis-Smyth T, Chen H, Park J, Presta LG, Ferrara N. The second immunoglobulin-like domain of the VEGF tyrosine kinase receptor Flt-1 determines ligand binding and may initiate a signal transduction cascade. EMBO J 1996; 15: 4919–4927. | PubMed | ChemPort |
14. Barleon B, Totzke F, Herzog C, Blanke S, Kremmer E, Siemeister G et al. Mapping of the sites for ligand binding and receptor dimerization at the extracellular domain of the vascular endothelial growth factor receptor Flt-1. J Biol Chem 1997; 272: 10382–10388. | Article | PubMed | ISI | ChemPort |
15. Wiesmann C, Fuh G, Christinger HW, Eigenbrot C, Wells JA, de Vos AM. Crystal structure at 1.7 A resolution of VEGF in complex with domain 2 of the Flt-1 receptor. Cell 1997; 91: 695–704. | Article | PubMed | ISI | ChemPort |
16. Davis-Smyth T, Presta LG, Ferrara N. Mapping the charged residues in the second immunoglobulin-like domain of the vascular endothelial growth factor/placenta growth factor receptor Flt-1 required for binding and structural stability. J Biol Chem 1998; 273: 3216–3222. | Article | PubMed | ISI | ChemPort |
17. Cunningham SA, Stephan CC, Arrate MP, Ayer KG, Brock TA. Identification of the extracellular domains of Flt-1 that mediate ligand interactions. Biochem Biophys Res Commun 1997; 231: 596–599. | Article | PubMed | ChemPort |
18. Holash J, Davis S, Papadopoulos N, Croll SD, Ho L, Russell M et al. VEGF-Trap: a VEGF blocker with potent antitumor effects. Proc Natl Acad Sci USA 2002; 99: 11393–11398. | Article | PubMed | ChemPort |
19. Ali RR, Reichel MB, Thrasher AJ, Levinsky RJ, Kinnon C, Kanuga N et al. Gene transfer into the mouse retina mediated by an adeno-associated viral vector. Hum Mol Genet 1996; 5: 591–594. | Article | PubMed | ISI | ChemPort |
20. Ali RR, Reichel MB, Hunt DM, Bhattacharya SS. Gene therapy for inherited retinal degeneration. Br J Ophthalmol 1997; 81: 795–801. | Article | PubMed | ChemPort |
21. Ali RR, Reichel MB, De Alwis M, Kanuga N, Kinnon C, Levinsky RJ et al. Adeno-associated virus gene transfer to mouse retina. Hum Gene Ther 1998; 9: 81–86. | Article | PubMed | ISI | ChemPort |
22. Lai CM, Shen WY, Brankov M, Lai YK, Barnett NL, Lee SY et al. Long-term evaluation of AAV-mediated sFlt-1 gene therapy for ocular neovascularization in mice and monkeys. Mol Ther 2005; 12: 659–668. | Article | PubMed | ChemPort |
23. Huston JS, Levinson D, Mudgett-Hunter M, Tai MS, Novotny J, Margolies MN et al. Protein engineering of antibody binding sites: recovery of specific activity in an anti-digoxin single-chain Fv analogue produced in Escherichia coli. Proc Natl Acad Sci USA 1988; 85: 5879–5883. | Article | PubMed | ChemPort |
24. Smith LE, Wesolowski E, McLellan A, Kostyk SK, D'Amato R, Sullivan R et al. Oxygen-induced retinopathy in the mouse. Invest Ophthalmol Vis Sci 1994; 35: 101–111. | PubMed | ISI | ChemPort |
25. Bennett J. Commentary: an aye for eye gene therapy. Hum Gene Ther 2006; 17: 177–179. | Article | PubMed | ChemPort |
26. Jacobson SG, Acland GM, Aguirre GD, Aleman TS, Schwartz SB, Cideciyan AV et al. Safety of recombinant adeno-associated virus type 2-RPE65 vector delivered by ocular subretinal injection. Mol Ther 2006; 13: 1074–1084. | Article | PubMed | ISI | ChemPort |
27. Bainbridge JW, Mistry A, De Alwis M, Paleolog E, Baker A, Thrasher AJ et al. Inhibition of retinal neovascularisation by gene transfer of soluble VEGF receptor sFlt-1. Gene Therapy 2002; 9: 320–326. | Article | PubMed | ISI | ChemPort |
28. Lai YK, Shen WY, Brankov M, Lai CM, Constable IJ, Rakoczy PE. Potential long-term inhibition of ocular neovascularisation by recombinant adeno-associated virus-mediated secretion gene therapy. Gene Therapy 2002; 9: 804–813. | Article | PubMed | ISI | ChemPort |
29. Gehlbach P, Demetriades AM, Yamamoto S, Deering T, Xiao WH, Duh EJ et al. Periocular gene transfer of sFlt-1 suppresses ocular neovascularization and vascular endothelial growth factor-induced breakdown of the blood–retinal barrier. Hum Gene Ther 2003; 14: 129–141. | Article | PubMed | ChemPort |
30. Xu L, Daly T, Gao C, Flotte TR, Song S, Byrne BJ et al. CMV-beta-actin promoter directs higher expression from an adeno-associated viral vector in the liver than the cytomegalovirus or elongation factor 1 alpha promoter and results in therapeutic levels of human factor X in mice. Hum Gene Ther 2001; 12: 563–573. | Article | PubMed | ISI | ChemPort |
31. Ziegler RJ, Lonning SM, Armentano D, Li C, Souza DW, Cherry M et al. AAV2 vector harboring a liver-restricted promoter facilitates sustained expression of therapeutic levels of alpha-galactosidase A and the induction of immune tolerance in Fabry mice. Mol Ther 2004; 9: 231–240. | Article | PubMed | ChemPort |
32. Vincent KA, Piraino ST, Wadsworth SC. Analysis of recombinant adeno-associated virus packaging and requirements for rep and cap gene products. J Virol 1997; 71: 1897–1905. | PubMed | ISI | ChemPort |
33. Zolotukhin S, Potter M, Zolotukhin I, Sakai Y, Loiler S, Fraites TJ et al. Production and purification of serotype 1, 2, and 5 recombinant adeno-associated viral vectors. Methods 2002; 28: 158–167. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Sirkka Kyostio-Moore and for helpful discussions; Michelle Deng for valuable comments on the paper; Shelley Nass and Denise Woodcock for virus production and Bob Brown for support with illustrations.