|Molecular Vision 2006;
Received 22 November 2005 | Accepted 27 April 2006 | Published 26 May 2006
VEGF165b, an endogenous C-terminal splice variant of VEGF, inhibits retinal neovascularization in mice
A. J. Churchill,2
S. J. Harper,1
David O. Bates,1
T. A. Gardiner3
1Microvascular Research Laboratories, Preclinical Veterinary School, Department of Physiology, University of Bristol, Bristol, UK; 2Department of Ophthalmology, Bristol Eye Hospital, Bristol, UK; 3Ophthalmology and Vision Science, Queen's University Belfast, Royal Victoria Hospital, Grosvenor Road, Belfast, Northern Ireland, UK
Correspondence to: Dr. David Bates, Microvascular Research Laboratories, Department of Physiology, Preclinical Veterinary School, University of Bristol, Southwell Street, Bristol BS2 8EJ, UK; Phone: 0117 928 9818; FAX: 0117 928 8151; email: Dave.Bates@bristol.ac.uk
Purpose: Hypoxia driven ocular angiogenesis occurs in a range of ischemic retinopathies including proliferative diabetic retinopathy and retinopathy of prematurity. These conditions are initiated and sustained by hypoxia dependent vascular endothelial growth factor (VEGF) expression in the eye. There are two families of VEGF isoforms formed by differential splicing, the pro-angiogenic VEGF family, known to contribute to ocular neovascularization, and the anti-angiogenic VEGF family, which are downregulated in diabetic retinopathy in humans. The first member of the VEGF family to be isolated was VEGF165b. To determine whether VEGF165b could inhibit hypoxia driven angiogenesis in the eye, the oxygen induced retinopathy mouse model of ocular neovascularization was used.
Methods: 1 ng of recombinant human VEGF165b peptide was injected intraocularly upon return to normoxia after 5 days exposure to 95% oxygen, and neovascularization assessed.
Results: VEGF165b significantly inhibited the percentage area of retinal neovascularization from 23±3% to 12±3.3%, and significantly increased normal vascular areas from 62±4% to 74±4%. The percentage area of residual ischemic retina was not affected.
Conclusions: These results show that a single injection of VEGF165b can significantly reduce preretinal neovascularization without inhibition of physiological intraretinal angiogenesis. Controlling the balance of VEGFxxx to VEGFxxx isoforms may therefore be therapeutically valuable in the treatment of proliferative eye diseases such as diabetic retinopathy and age related macular degeneration. The regulation of splicing between these two families of isoforms may provide a novel therapeutic strategy for proliferative eye disease.
Vascular endothelial growth factor (VEGF) has been identified in a number of studies as a key component in the abnormal proliferation of retinal blood vessels after a hypoxic insult, such as in proliferative diabetic retinopathy, or retinopathy of prematurity. Hypoxic ischemia of the inner retina leads to increased expression of VEGF-A, abnormal vascular permeability and often to retinal neovascularization (RNV). RNV introduces fibrovascular tissue to the preretinal area and vitreous body and may progress to vitreous hemorrhage, traction retinal detachment, and loss of vision. Recently it has been shown that differential splicing of the VEGF-A gene results in two families of isoforms, expressed using either a proximal splice site in the terminal exon (exon 8a), or an alternate splice acceptor site, 66 bases further downstream. This alternate splicing results in a mRNA containing 18 bases coded for by exon 8b, in place of the 18 bases of exon 8a , and hence produce proteins of the same length as other forms, but with a different C terminal amino acid sequence. This family of isoforms has been termed the VEGFxxxb family, where xxx is the number of amino acids that they encode, and the proximal splice family VEGFxxx. VEGF165b, VEGF189b, VEGF145b, VEGF183b and VEGF121b have been identified in human eye tissues . The only one of these isoforms for which there is any functional information is VEGF165b, the first identified member of this family. Interestingly, the receptor binding domains and dimerization domains are intact in these new isoforms, but evidence so far shows that it does not elicit full signalling from at least one of its cognate receptors, VEGFR2 (KDR/FLK1) (i.e., pre-formed VEGF165b homodimers bind to the receptor but do not stimulate endothelial cell growth, migration, endothelium dependent vasodilatation, or VEGFR2 phosphorylation in heterologous expression systems). These effects were specific to VEGF and no effect was seen on fibroblast growth factor (FGF) induced endothelial cell growth. Further experiments showed that VEGF165b inhibited VEGF165 induced-angiogenesis in the rabbit cornea and the rat mesentery, and inhibited tumor growth in xenotransplanted tumors in mice .
Therefore VEGFxxxb appears to be able to inhibit VEGFxxx dependent angiogenesis, or at least the VEGF165b isoform can inhibit its own counterpart, VEGF165. Since hypoxia induced angiogenesis in the retina leads to neovascularization predominantly through a VEGF mediated mechanism, we sought to determine the effect of VEGF165b on hypoxia dependent angiogenesis in a mouse model of retinal neovascularization.
Oxygen induced retinopathy (OIR)
All animal experiments were carried out in accordance with UK legislation under the Animal (scientific procedures) Act. OIR was induced in C57/BL6 as described by Smith et al. . In short, neonatal mice and their nursing dams were exposed to 75% oxygen (PRO-OX 110 chamber oxygen controller; Biospherix Ltd., Redfield, NY) between postnatal day 7 (P7) and P12 producing vaso-obliteration and cessation of vascular development in the capillary beds of the central retina. Return of the animals to room air at P12 produces hypoxia in the ischemic central retina, and results in preretinal neovascularization between P15 and P21 with maximal neovascularization at P17. As the C57/BL6 dams showed inconsistent care and nutrition of the pups when returned to room air, at P12 all C57/BL6 pups were cross-fostered to nursing Swiss dams that had not been exposed to hyperoxia.
Multiple timed matings were prepared for the mice. Neonates were divided in 2 groups of 16 and 17 (total 33). At P13 one group received an intravitreal injection of VEGF165b (about 25 fmol recombinant human VEGF165b, R&D Systems, 1 μl at 1 ng/μl), in the right eye only, and a control group was injected with 1 μl of vehicle (Hank's solution, 140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1.3 mM MgCl2, 5 mM D-glucose, 10 mM HEPES (free acid), pH 7.4). The injection was performed under isoflurane anesthesia using 10 μl beveltip Flexifil syringe, fitted with a 33 gauge needle (WPI, Sarasota, FL). The injection was made via the ciliary body and at an oblique angle to avoid contact with the lens. However, animals with eyes showing evidence of lens damage (retrolental adhesions or vitreous strands) or retinal detachment were not included in the analysis (3 animals). There was no evidence of endophthalmitis in any eyes (cellular infiltrates and cloudy vitreous). At P17 the animals were terminally anesthetized, and both eyes enucleated. The eyes were immersion fixed in 4% paraformaldehyde for 4 h at 4 °C and washed extensively for a further 4 h period. After this, the anterior segment/lens complexes were removed and the posterior eye cups relaxed by placing four radial cuts from the retinal periphery to points within 1 mm of the optic disk. The specimens were placed in 96 well plates and permeabilized by an overnight soak at 4 °C in PBS containing 0.5% Triton X-100, 1% normal goat serum and 0.1 mM CaCl2. The retinal vasculature was visualized by reaction with biotinylated GS isolectin B4 at 20 μg/ml (Sigma, St. Louis, MO) followed by Alexa488-streptavidin (Molecular Probes, Eugene, OR). The retinas were washed extensively and flat-mounted in Vectashield (Vector Laboratories, Peterborough, UK).
Quantitation of neovascular, ischemic and normally vascularized areas
Flat-mounted retinas were assessed using NIH Image J version 1.31 for quantification of the areas of avascular ischemic retina (I), normal intraretinal vascularization (N) and preretinal neovascularization (NV), expressed as percentages of total analyzed retinal area. Coded and randomized retinal flat mounts were analyzed in a single-blind fashion by a trained observer. In lectin-stained flat-mounts neovascular tufts appear as amorphous densely stained conglomerates in comparison to the discrete chicken-wire arrangement of the normal intraretinal vasculature ; the ischemic areas remain unstained apart from aggregates of microglial cells, which are easily distinguished from the vasculature.
Statistical comparisons of neovascularized, ischemic and normally vascularized retinal areas in VEGF165b-injected and vehicle-injected mice was carried out using the unpaired t-test (2 tailed). Comparison of injected right eyes to uninjected left eyes employed the paired t-test (2 tailed). Linear regression was used to compare the influences of residual ischemia and normally vascularized retina on the area of neovascularization at P17. The study was powered to show a 95% certainty of detecting a 5% reduction in neovascularization from 23.3±4% compared to control (calculated using G Power, version 2.1.2. Universitat Trier, Trier, Germany), and animal numbers calculated to achieve this level of significance, and more than enough to overcome inherent variability. The results fulfilled all criteria for parametric analysis: normal distribution and no significant difference in the standard deviations.
Figure 1A shows a flat-mounted P17 mouse retina stained with isolectin B4 and Figure 1B the effect of 5 days of hyperoxia from P7 to P12 followed by four days of normoxia (i.e., relative hypoxia). The ischemic insult results in large areas of hypoxia with associated preretinal neovascularization. Figure 1C shows a higher power magnification of growing tips of blood vessels showing an endothelial cell at the tip of a sprout with multiple filopodia growing at multiple angles including both along and above the plane of the section. In comparison, Figure 2 shows the effect of intravitreal injection of VEGF165b 1 ng in 1 μl of Hank's saline on day P13 on the vasculature at day P17. Although the areas of ischemia are still present, there is evidence of physiological angiogenesis proceeding at the edge of the ischemic area. Figure 3 shows quantitation of the neovascular areas. Figure 3A shows a vehicle-injected mouse retina with 16% ischemic area. Figure 3B shows the retina of a mouse with similar ischemic area in the injected eye, but exposed to VEGF165b. This retina exhibited less neovascularization, and more normal vasculature. A comparison with the contralateral eye is given for the control injection in Figure 3C, and for the VEGF165b injected mouse in Figure 3D. It can be seen that while in control animals exposure to hyperoxia for 5 days followed by normoxia resulted in significant ischemic (15±1.4%; Figure 4A) and neovascular areas (23±3%; Figure 4B), VEGF165b treatment significantly reduced the neovascular area to 12±3.3% (p<0.05; Figure 4B), and significantly increased the normally vascularized area from 61.8±4.4 to 74.4±4.1% (p<0.05; Figure 4C). To compare matched eyes from the same mice, VEGF165b injection reduced the neovascular area to 22±6.0% of that in the matched eye, (from 59±2.8% to 12±3.5%, Figure 5A) which was a significantly greater reduction than vehicle injection (36.7±7.4% of control, a reduction of neovascular area from 56±2.7% to 23±3%, Figure 5B). Therefore, both the VEGF165b and the vehicle injected eyes had significantly smaller neovascular area than the untreated left eyes of both groups showing that, intravitreal injection by itself significantly inhibits neovascularization but VEGF165b injection induces a significant additional reduction in neovascularization, over and above that induced by vehicle injection. This is consistent with previous reports of ocular wounding inhibiting retinal neovascularization . To determine whether VEGF165b injection had any effect on neovascularization in the contralateral eye, vascular areas were also compared between right and left eyes of mice that had been injected with VEGF165b or vehicle control in their right eyes. The uninjected left eyes showed no change in either the normal, ischemic or neovascular areas (Figure 5C).
Linear regression analysis with the area of neovascularized retina as the dependent outcome variable (Y-axis) demonstrated a strong correlation between the residual ischemic retina and the area of NV at P17 (R=0.6, p value of slope=0.0012). However, the analysis showed a highly significant negative correlation between the area of NV and that of normally vascularized retina (R=-0.98, p value of slope<0.0001).
The present study has shown that 1 ng (about 25 fmol) VEGF165b, a member of the recently discovered family of inhibitory VEGF isoforms (VEGFxxxb family), can significantly reduce the aggressive preretinal neovascularization in oxygen-induced retinopathy in the mouse. Interestingly, the antineovascular effect of VEGF165b produced no exacerbation of inner retinal ischemia since it permitted physiological angiogenesis to proceed normally. Quantitatively, VEGF165b inhibition of RNV resulted in the recovery of a morphologically normal intraretinal vasculature in areas that would otherwise have been covered by intravitreal neovascular tufts. This study has further shown that the degree to which normal perfusion is re-established in formerly ischemic retina is predictive of the corresponding reduction in preretinal neovascularization, as suggested by previous studies [5,7].
VEGF-A isoforms have been shown to be massively upregulated in diabetic retinopathy, particularly in retinal pigmented epithelial cells, glial cells, and vitreal fibroblasts [8-10]. There is increasing functional evidence that inhibition of angiogenesis, both nonspecifically with pharmacological antiangiogenic agents and specifically with anti-VEGF agents ameliorates diabetic retinal signs in well-characterized animal models of retinopathy [11-13]. Moreover, there is evidence that inhibition of VEGF may be effective in human disease, as anti-VEGF therapies such as Pegaptinib® and Lucentis® are now in phase II clinical trials for diabetic macular edema and proliferative diabetic retinopathy .
We recently showed that the VEGFxxxb family of isoforms is expressed in human retina, vitreous and iris, and that VEGFxxxb isoforms are the most abundant species in normal vitreous . However, these isoforms are relatively downregulated in diabetic vitreous resulting in a switch to an angiogenic phenotype . There is mounting evidence that subtle polymorphisms in the 5' untranslated region of VEGF can predispose to diabetic retinopathy, suggesting that regulation of transcription or splicing of the VEGF gene is important in this condition . VEGFxxxb isoforms have an almost identical sequence to their VEGFxxx isoforms and the mRNA and protein are the same size , so these families of isoforms are not distinguished from each other by the majority of methods previously used to detect VEGF165 (including western blotting, most PCR protocols, immunohistochemistry, in situ hybridization, ribonuclease protection assay, etc.) . This may explain the elusiveness of exon 8b isoforms until recently. However, it also means that publications describing VEGF expression in retinal tissue, or vitreous fluid and plasma from normal, diabetic patients would not distinguish the pro- and anti-angiogenic VEGF isoforms. Our understanding of VEGF biology therefore has to be radically re-evaluated since expression of what had previously been thought to be VEGF (particularly, in the normal retina) is to a large extent VEGF165b.
The results shown here indicate that VEGF165b can inhibit ocular neovascularization due to relative hypoxia, a situation that characterizes diabetic retinopathy and other vaso-occlusive retinopathies [16-18]. It did not appear to inhibit the normal neovascularization along the plane of the retina (the ischemic area was not altered and filopodia were still apparently normal, but in the plane of the retina), possibly because the location of injections was superficial to the retina, and therefore the effect was most pronounced superficial to the retina where the abnormal neovascularization was more extensive. Previous studies have shown that inhibiting all VEGF isoforms can reduce retinal neovascularization in animal models of retinal hypoxia , and in human ocular neovascular pathologies with variable effectiveness . However, it is not yet clear whether the effectiveness of this inhibition depends upon the relative ratios of the pro- and anti-angiogenic families of isoforms, or whether inhibitors such as VEGF-TRAP, sFlt-1 or Ranibizumab have equal affinities for the two families of isoforms. It is likely that agents that preferentially inhibit the angiogenic family will be more effective in those patients with less downregulation of the anti-angiogenic families. Furthermore, it has now been shown that anti-VEGF treatment can induce vessel regression in some tissues .
The identification of an endogenous inhibitor of the pro-angiogenic isoforms of VEGF suggests that VEGF165b may be a potential therapeutic candidate for RNV, particularly as it shows no inhibition of physiological angiogenesis, a process already compromised by the diabetic state . The effect of other VEGFxxxb isoforms on RNV has not yet been identified, and it will be of interest to determine whether other isoforms with different heparin binding domains, such as VEGF189b or VEGF121b will have additional anti-angiogenic capacity. Perhaps even more interesting is the concept that switching from proximal splice site selection (PSS) to distal splice site selection (DSS) in diabetic patients may prevent the progression of RNV by restoring the balance of pro- and anti-angiogenic isoforms. This splicing switch has been shown to occur in renal glomerular epithelial cells during differentiation in experimental conditions , but the mechanism regulating this is unknown. These mechanisms are likely to be key to this process and, although understanding how these mechanisms are regulated is an area of research that is very poorly understood, it may be of potential enormous therapeutic interest.
In summary, terminal exon distal splice site selection in the VEGF gene results in an endogenous protein, VEGF165b, which we have shown here inhibits retinal neovascularization in oxygen induced retinopathy in the neonatal mouse. This inhibition of hypoxia dependent angiogenesis suggests that administration of VEGF165b is a potential therapeutic strategy for retinal neovascularization, and that control of VEGF splicing between pro- and anti-angiogenic families may provide a novel mechanism to control this neovascularization.
This work was supported by Diabetes UK (RD02/CS02520) and the Wellcome Trust (69029). Dr Bates is supported by the British Heart Foundation BB2000003.
1. Bates DO, Cui TG, Doughty JM, Winkler M, Sugiono M, Shields JD, Peat D, Gillatt D, Harper SJ. VEGF165b, an inhibitory splice variant of vascular endothelial growth factor, is down-regulated in renal cell carcinoma. Cancer Res 2002; 62:4123-31.
2. Perrin RM, Konopatskaya O, Qiu Y, Harper S, Bates DO, Churchill AJ. Diabetic retinopathy is associated with a switch in splicing from anti- to pro-angiogenic isoforms of vascular endothelial growth factor. Diabetologia 2005; 48:2422-7.
3. Woolard J, Wang WY, Bevan HS, Qiu Y, Morbidelli L, Pritchard-Jones RO, Cui TG, Sugiono M, Waine E, Perrin R, Foster R, Digby-Bell J, Shields JD, Whittles CE, Mushens RE, Gillatt DA, Ziche M, Harper SJ, Bates DO. VEGF165b, an inhibitory vascular endothelial growth factor splice variant: mechanism of action, in vivo effect on angiogenesis and endogenous protein expression. Cancer Res 2004; 64:7822-35.
4. Smith LE, Wesolowski E, McLellan A, Kostyk SK, D'Amato R, Sullivan R, D'Amore PA. Oxygen-induced retinopathy in the mouse. Invest Ophthalmol Vis Sci 1994; 35:101-11.
5. Gardiner TA, Gibson DS, de Gooyer TE, de la Cruz VF, McDonald DM, Stitt AW. Inhibition of tumor necrosis factor-alpha improves physiological angiogenesis and reduces pathological neovascularization in ischemic retinopathy. Am J Pathol 2005; 166:637-44.
6. Stitt AW, Graham D, Gardiner TA. Ocular wounding prevents pre-retinal neovascularization and upregulates PEDF expression in the inner retina. Mol Vis 2004; 10:432-8 <http://www.molvis.org/molvis/v10/a54/>.
7. Sennlaub F, Courtois Y, Goureau O. Inducible nitric oxide synthase mediates the change from retinal to vitreal neovascularization in ischemic retinopathy. J Clin Invest 2001; 107:717-25.
8. Malecaze F, Clamens S, Simorre-Pinatel V, Mathis A, Chollet P, Favard C, Bayard F, Plouet J. Detection of vascular endothelial growth factor messenger RNA and vascular endothelial growth factor-like activity in proliferative diabetic retinopathy. Arch Ophthalmol 1994; 112:1476-82.
9. Sone H, Kawakami Y, Okuda Y, Kondo S, Hanatani M, Suzuki H, Yamashita K. Vascular endothelial growth factor is induced by long-term high glucose concentration and up-regulated by acute glucose deprivation in cultured bovine retinal pigmented epithelial cells. Biochem Biophys Res Commun 1996; 221:193-8.
10. Amin RH, Frank RN, Kennedy A, Eliott D, Puklin JE, Abrams GW. Vascular endothelial growth factor is present in glial cells of the retina and optic nerve of human subjects with nonproliferative diabetic retinopathy. Invest Ophthalmol Vis Sci 1997; 38:36-47.
11. Ozaki H, Seo MS, Ozaki K, Yamada H, Yamada E, Okamoto N, Hofmann F, Wood JM, Campochiaro PA. Blockade of vascular endothelial cell growth factor receptor signaling is sufficient to completely prevent retinal neovascularization. Am J Pathol 2000; 156:697-707.
12. Aiello LP, Pierce EA, Foley ED, Takagi H, Chen H, Riddle L, Ferrara N, King GL, Smith LE. Suppression of retinal neovascularization in vivo by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF-receptor chimeric proteins. Proc Natl Acad Sci U S A 1995; 92:10457-61.
13. Sone H, Kawakami Y, Segawa T, Okuda Y, Sekine Y, Honmura S, Segawa T, Suzuki H, Yamashita K, Yamada N. Effects of intraocular or systemic administration of neutralizing antibody against vascular endothelial growth factor on the murine experimental model of retinopathy. Life Sci 1999; 65:2573-80.
14. Comer GM, Ciulla TA. Pharmacotherapy for diabetic retinopathy. Curr Opin Ophthalmol 2004; 15:508-18. Erratum in: Curr Opin Ophthalmol 2005; 16:195.
15. Awata T, Inoue K, Kurihara S, Ohkubo T, Watanabe M, Inukai K, Inoue I, Katayama S. A common polymorphism in the 5'-untranslated region of the VEGF gene is associated with diabetic retinopathy in type 2 diabetes. Diabetes 2002; 51:1635-9.
16. Wise GN. Retinal neovascularization. Trans Am Ophthalmol Soc 1956; 54:729-826.
17. Shima DT, Adamis AP, Ferrara N, Yeo KT, Yeo TK, Allende R, Folkman J, D'Amore PA. Hypoxic induction of endothelial cell growth factors in retinal cells: identification and characterization of vascular endothelial growth factor (VEGF) as the mitogen. Mol Med 1995; 1:182-93.
18. Pe'er J, Shweiki D, Itin A, Hemo I, Gnessin H, Keshet E. Hypoxia-induced expression of vascular endothelial growth factor by retinal cells is a common factor in neovascularizing ocular diseases. Lab Invest 1995; 72:638-45.
19. Gragoudas ES, Adamis AP, Cunningham ET Jr, Feinsod M, Guyer DR, VEGF Inhibition Study in Ocular Neovascularization Clinical Trial Group. Pegaptanib for neovascular age-related macular degeneration. N Engl J Med 2004; 351:2805-16.
20. Baffert F, Le T, Sennino B, Thurston G, Kuo CJ, Hu-Lowe D, McDonald DM. Cellular changes in normal blood capillaries undergoing regression after inhibition of VEGF signaling. Am J Physiol Heart Circ Physiol 2006; 290:H547-59.
21. Stitt AW, McGoldrick C, Rice-McCaldin A, McCance DR, Glenn JV, Hsu DK, Liu FT, Thorpe SR, Gardiner TA. Impaired retinal angiogenesis in diabetes: role of advanced glycation end products and galectin-3. Diabetes 2005; 54:785-94.
22. Cui TG, Foster RR, Saleem M, Mathieson PW, Gillatt DA, Bates DO, Harper SJ. Differentiated human podocytes endogenously express an inhibitory isoform of vascular endothelial growth factor (VEGF165b) mRNA and protein. Am J Physiol Renal Physiol 2004; 286:F767-73.