Introduction

Angiogenesis is the fundamental physiological process by which new blood vessels are generated from the pre-existing vasculature. Angiogenesis plays a critical role in the elaboration of the vasculature during development; however, this process is also triggered in hypoxic tissue in the adult organism [1]. Low oxygen concentrations lead to stabilization of the HIF-1 transcription factor, which then drives the transcription of a number of secreted pro-angiogenic factors—notably the growth factor VEGF-A [2]. While this response can be beneficial in conditions of tissue ischemia, it makes a negative contribution to a number of disease states, including proliferative diabetic retinopathy and wet age-related macular degeneration [3–5].

This pathological neovascularization has been most widely studied in cancer. Solid tumors become hypoxic at an early stage and secrete VEGF and other pro-angiogenic factors to stimulate tumor angiogenesis [6]. In addition to the effects of hypoxia, oncogenic changes to tumor cells can further contribute to tumor angiogenesis. Activating mutations of Ras, common in solid tumors, drive increased VEGF expression in cancer cells and repress the expression of anti-angiogenic factors [7–9]. Farnesyltransferase inhibitors (FTIs) are a novel class of anticancer therapeutics, designed to inhibit the farnesylation of Ras—a post-translational modification that targets Ras to its site of action at the plasma membrane. FTIs have been shown to have anti-angiogenic actions in tumor xenograft models [10–13]. Treatment of cancer cell lines with FTIs in vitro leads to reduced expression of both HIF-1 and VEGF [12, 14, 15], suggesting that FTIs are able to reverse the pro-angiogenic effects of oncogenic Ras in cancer cells.

Surprisingly, FTIs have recently been shown to inhibit VEGF-stimulated angiogenesis in a rat corneal angiogenesis model [10]. This suggests that there are general anti-angiogenic effects of FTIs that operate outside the tumor context, and which do not depend on targeting oncogenic Ras. How FTIs might be working in this context is unknown. Studies with FTIs suggest that they predominantly target transformed cells [16]. In keeping with this, research into the effects of FTIs on tumor angiogenesis has focused on responses in tumor cells, rather than the endothelial cells that receive tumor-derived pro-angiogenic signals. Here, we show that FTI treatment has profound and direct effects on normal endothelial cells, targeting multiple components of the angiogenic response. These findings significantly increase our understanding of the effects of FTIs on tumor angiogenesis and explain the general effects of FTIs on angiogenesis seen outside cancer. Importantly, they suggest that the therapeutic use of FTIs may extend beyond cancer to include treatment of pathological angiogenesis in conditions where Ras mutation is not a factor.

Materials and methods

Materials

The specific FTI L-744,832 was from Merck (Nottingham, UK) and was used at 10 μM unless stated otherwise. Sphingosine-1-phosphate was from Biolmol (Exeter, UK). Recombinant human VEGF-165 and FGF-2 were from R&D systems (Minneapolis, US), as were mouse (clone 390) and sheep antibodies to the platelet endothelial cell adhesion molecule 1 (PECAM-1). The 5-bromo-2-deoxyuridine (BrdU) mouse monoclonal antibody (BU-1) was from Upstate Biotechnology (Millipore, Dundee, UK). The FT-β subunit mouse monoclonal (B-7) was from Santa Cruz (Santa Cruz, US). Phosphospecific antibodies recognizing activated VEGFR2 (Tyr1175), p42/44 MAP kinase (Thr202/Tyr204), and Akt/PKB (Ser473) were from Cell Signaling Technologies, as was a rabbit polyclonal antibody recognizing total Akt/PKB.

Cell culture

Primary human umbilical vein endothelial cells (HUVEC) were prepared as described previously [17]. HUVEC were cultured in complete endothelial cell growth medium (ECGM, PromoCell) and supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin. Where indicated, cells were starved in endothelial cell basal medium supplemented with 0.1% (w/v) fatty acid free BSA (ECBM, PromoCell). HUVEC were used between passages three and six. Normal primary human dermal fibroblasts (NHDF, PromoCell) were cultured in DMEM with 10% (w/v) fetal bovine serum (FBS) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 300 μg/ml glutamine. Tissue culture plastic was pre-coated with 10 μg/ml fibronectin at room temperature for 30 min and acid-washed glass coverslips were pre-coated with 50 μg/ml fibronectin for 1 h at 37°C.

Immunofluorescence microscopy

For confocal immunofluorescence microscopy, cells were fixed for 15 min in 4% fresh paraformaldehyde in PBS and processed as described previously [18]. Confocal microscopy was performed using a Leica AOBS SP confocal laser-scanning microscope.

Proliferation assays and cell cycle analysis

Human umbilical vein endothelial cells were seeded on six well plates at 4 × 104 cells/ml and cultured in ECGM containing FTI or vehicle (DMSO). Medium and drug were refreshed every 24 h. Cells were harvested and counted using a hemocytometer at 24, 48, and 72 h to measure proliferation. For cell cycle analysis, cells were fixed in 70% ethanol at −20°C and stained with propidium iodide. DNA content was analyzed by flow cytometry.

Adhesion and migration assays

ECM cell adhesion array kits from Chemicon (Millipore, Dundee, UK) were used according to the manufacturer’s instructions. Cells were treated with FTI or vehicle in ECGM for 24 h before the experiment. Cell migration was measured using a modified Boyden chamber, according to the manufacturer’s instructions (Transwell, Corning Costar). Semi-permeable polycarbonate Transwell filters (6.5 mm diameter, 8 μm pore size) were coated with 10 μg/ml fibronectin for 16 h at 4°C. Filters were assembled with ECBM in the bottom chamber supplemented either nothing (control), fibroblast growth factor-2 (FGF-2; 25 ng/ml), VEGF-165 (40 ng/ml), or sphingosine 1-phosphate (S1P; 100 nmol/l). Cells were treated with FTI or vehicle for 16 h prior to the experiment and 3.0 × 104 cells were added to the upper chamber of each filter. After 4 h, the filters were fixed in 4% paraformaldehyde for 15 min and then permeabilized in 0.2% Triton X-100 for 5 min before staining with 5 μM DAPI for 5 min.

In vitro co-culture assay of angiogenesis

The effects of FTI on angiogenesis were examined using a well-characterized co-culture assay that leads to the formation of an anastomosing network of endothelial tubules over 10–14 days [19, 20]. HUVEC were mixed with NHDF at a ratio of 1:1 and with a total cell concentration of 3 × 104 cells/ml. The mixture was plated on fibronectin-coated coverslips and the cells were then cultured with vehicle or FTI in ECGM for up to 14 days to allow tubule formation. Medium and drug were refreshed every 2–3 days. In some experiments, cells were treated with 100 μM BrdU (Merck) for 2 h in ECGM prior to fixation and processing for immunofluorescence. To quantify vessel formation, cells were fixed in 70% ethanol at −20°C and labeled with anti-PECAM-1 antibody in 1% BSA for 1 h at 37°C. The labeled endothelial tubes were then stained using alkaline phosphatase-conjugated secondary antibody and BCIP/NBT solution (Sigma).

A modified version of the assay was used for experiments involving siRNA treatment of endothelial cells. NHDF were seeded alone onto glass coverslips at 3 × 104 cell/ml and allowed to grow to confluence over 4 days. On day 3, HUVEC were seeded at 4 × 104 cell/ml in 6-cm dishes. On day 4, these HUVEC were transfected with siRNA oligonucleotides using GeneFECTOR lipid according to the manufacturer’s protocol (Venn Nova, US). The control siRNA targeting lamin A/C (GGUGGUGACGAUCUGGGCUTT) has been characterized previously [21]. Two independent siRNAs targeting the FT-β subunit siRNA were designed and synthesized by Dharmacon (Thermo-Fisher, UK): FT1 (GCAAGUCGCGUGAUUUCUAUU) and FT2 (GCACUUCCAUUAUCUGAAAUU). After 3 h, the transfected endothelial cells were harvested and seeded onto the confluent monolayer of NHDF at a concentration of 3 × 104 cells/ml. At day 11, the cells were fixed and stained for PECAM as before.

Aortic ring assay

Thoracic aortas were removed from 8- to 12-week-old C57BL/6 mice after cervical dislocation and aortic ring assays were performed as previously described [22]. Angiogenesis was quantified by counting the number of microvessel sprouts that grew from each ring after 6 days.

Statistical analysis

Unless stated otherwise, all experimental data is the mean ± SEM of three entirely independent experiments. P values were determined by carrying out a paired student’s t-test (two-tailed).

Results

FTIs block endothelial cell proliferation

During angiogenesis, the normally quiescent endothelial cells of stable vessels must undergo rapid proliferation [1]. FTIs inhibit the proliferation of cancer cells; however, they generally have little effect on the proliferation of normal, untransformed cells [16]. Previous studies have provided conflicting data on the effects of FTIs on endothelial cell proliferation, with reports that proliferation of normal endothelial cells is either inhibited [12] or not inhibited [10] by FTI treatment. We examined the effects of FTI treatment on normal primary endothelial cells grown in culture. In agreement with Han et al. [12], we found that FTI treatment caused a marked inhibition of endothelial cell proliferation (Fig. 1a). We determined the basis of this effect by flow cytometry. FTI treatment led to a marked increase in the number of diploid endothelial cells with a consequent decrease in the haploid (G1/G0) population. We also saw a small but consistent peak representing tetraploid cells (Fig. 1b). FTIs can trigger apoptotic death in Ras-transformed cells [16, 23]; however, flow cytometry did not reveal a sub-G1 peak of apoptotic endothelial cells (Fig. 1b) and no cell death was observed in culture (data not shown). An increase in the diploid population is indicative of a G2/M cell cycle arrest. We examined this in detail by microscopy. FTI treatment did not increase the number of mitotic figures (Fig. 1c)—in fact, we were unable to find FTI-treated endothelial cells that had progressed beyond prophase. Instead, FTI-treated endothelial cells frequently exhibited multiple nuclei with chromosomal bridges between them (Fig. 1c). We conclude that FTIs block endothelial cell proliferation in early M-phase, with subsequent recondensation of chromatin to form multinucleate cells.

Fig. 1
figure 1

FTI inhibits endothelial cell proliferation. (a) Endothelial cells were cultured with FTI (solid circles) or with vehicle (DMSO; open circles). Proliferation was analyzed by counting cells at 24, 48, and 72 h after plating. (b) DNA content was analyzed using flow cytometry of cells treated for 72 h in the absence or presence of FTI. The panel shows a representative result from three independent experiments. (c) Endothelial cells were cultured for 72 h with or without FTI. The cells were then fixed and stained with DAPI (nuclei; blue) and PECAM-1 (plasma membrane; green). The far panel shows a magnified section of the FTI-treated cells (dashed box). In this image, the chromosomal bridges in the multinucleated cells can be clearly seen. Scale bar represents 20 μm

Full size image

FTI treatment blocks endothelial cell chemotaxis to VEGF

Endothelial cells participating in angiogenesis must first invade the surrounding tissue, and then undergo chemotactic migration toward the pro-angiogenic signal. This involves switching interactions from components of the basement lamina, like laminin, to interstitial matrix components, like fibronectin and collagen I [24]. The importance of these interactions to angiogenesis led us to examine whether FTI treatment affected the ability of endothelial cells to interact with extracellular matrix components. The adhesion of endothelial cells to a wide range of individual matrix proteins was measured, including components of the basement membrane and the interstitial matrix. FTI treatment had no effect on endothelial cell adhesion (Fig. 2a). We conclude that this aspect of the angiogenic process is not targeted by FTIs.

Fig. 2
figure 2

FTI does not affect cell adhesion or VEGF signaling, but inhibits endothelial cell migration toward VEGF. (a) Endothelial cells were treated with FTI (solid bars) or with vehicle (DMSO; open bars) for 24 h before being seeded onto the indicated extracellular matrix proteins. After 30 min, adherent cells were quantified by fluorimetry. (b) Endothelial cells were treated with FTI or with vehicle for 16 h before being seeded onto the fibronectin-coated upper chamber of a modified Boyden chamber. The lower chamber contained either normal cell medium or medium supplemented with 40 ng/ml VEGF, 100 nmol/l sphingosine-1-phosphate (S1P) or 25 ng/ml FGF-2, as indicated. After 4 h, cells that had migrated to the bottom chamber were fixed and counted. (c) Endothelial cells were treated with FTI or with vehicle for 24 h prior to stimulation with 40 ng/ml VEGF for the indicated times. Cells were analyzed by Western blotting with phosphospecific antibodies to measure the activation of VEGFR2, Akt/PKB, and p42/44 MAP kinase. Tubulin was measured as a loading control

Full size image

We then measured the effects of FTI treatment on the ability of endothelial cells to migrate toward three important chemotactic signals in angiogenesis—VEGF, sphingosine-1-phosphate, and FGF-2. FTI treatment had a small inhibitory effect on migration to spingosine-1-phosphate and FGF-2; however, this was not statistically significant (Fig. 2b). In contrast, FTI treatment completely inhibited chemotaxis toward VEGF (Fig. 2b). We examined whether FTI treatment affected VEGF signaling. Treatment with FTI had no effect on the activation of VEGF receptor-2 (VEGFR2/KDR; Fig. 2c), the major mediator of VEGF-induced chemotaxis in endothelial cells [25]. Further, FTI treatment had no effect on the downstream activation of either p42/p44 or PKB/Akt (Fig. 2c), suggesting that VEGF signaling was not compromised in these cells. We conclude that FTI treatment has a specific inhibitory effect on the chemotactic migration of endothelial cells toward VEGF; however, this is not through a gross inhibition of VEGFR2 signaling.

FTI treatment inhibits angiogenesis in an in vitro co-culture model

Having used isolated endothelial cells to show direct effects of FTI treatment on components of the angiogenic process, we wanted to examine the effects of FTIs in a more integrated system. We used a well-characterized in vitro angiogenesis model, where co-cultures of primary dermal fibroblasts and primary endothelial cells give rise to a branching network of well-defined tubules over a period of 10–14 days. Endothelial cells in this system first undergo a period of rapid proliferation, before migrating to form cord-like structures that then become stable tubules with a patent lumen [19, 20]. FTI treatment had a profound effect on tube formation in this assay. At day 7, when control assays had assembled the elongated sheets of endothelial cells, the corresponding FTI-treated cells were highly disorganized (Fig. 3). At day 14, when control assays had assembled a network of capillary-like structures, the FTI-treated cells had produced only a very few, highly truncated structures (Fig. 3).

Fig. 3
figure 3

FTI inhibits tube formation in a co-culture assay of angiogenesis. Assays were preformed in the absence or presence of FTI for the number of days indicated. Cells were stained with DAPI (blue) to reveal fibroblasts and endothelial cells, and for PECAM-1 (green) to reveal the endothelial cells. Scale bar represents 100 μm

Full size image

In our studies with isolated endothelial cells, we had seen effects of FTI on at least two components of the angiogenic process—proliferation and chemotaxis. Clearly, any inhibition of proliferation caused by FTI treatment in the early stages of the co-culture assay would carry through and potentially mask other activities. We wanted to see if we could divide the assay into its component stages and then extract detailed information about the actions of FTIs. We treated co-culture assays with FTI either for the first 4 days, or for days 4–7, and compared the effects on cell proliferation by using BrdU incorporation to label proliferating cells. As previously reported [26], endothelial cells were actively proliferating at day 4 of the assay, as were fibroblasts. FTI treatment caused a complete loss of proliferating endothelial cells at day 4 and significantly inhibited fibroblast proliferation (Supplementary Fig. 1). By day 7, endothelial cell proliferation had finished in control assays and hence no difference was observed with FTI treatment from days 4 to 7 (Supplementary Fig. 1). Having established the window of proliferation in the assay, we reasoned that treatment of co-cultures with FTI after the initial proliferative phase might reveal any additional actions of the drug. Cells treated between days 4 and 7 showed no obvious loss of cell number and the cord-like sheets were indistinguishable from control assays (Fig. 4a). Treatment of cells with FTI between days 7 and 14 caused significant defects—while there were no apparent differences in cell number, the FTI-treated co-cultures produced significantly fewer tubes. Quantification of these assays showed that FTI treatment led to a 67% reduction in tube length in the assay (Fig. 5). Close inspection of the tubes formed in the FTI-treated assays showed that they were almost completely devoid of the numerous long filopodial projections found on endothelial tubes in the control assays (Fig. 4b). Finally, treatment of co-cultures with FTI at day 14 had no significant effect on capillary-like structures—i.e. tubes that had formed were resistant to subsequent FTI treatment (Fig. 5).

Fig. 4
figure 4

FTI targets endothelial tube formation. (a) Assays were preformed in the absence or presence of FTI. In this assay, treatment with FTI was restricted to discrete periods during tube formation, to examine the effects of FTI on different aspects of the process. These treatments were carried out after the initial period of endothelial cell proliferation. Cells were stained with DAPI (blue) to reveal fibroblasts and endothelial cells, and for PECAM-1 (green) to reveal the endothelial cells. Scale bar represents 100 μm. (b) At high magnification, it can be seen that FTI-treatment between days 7 and 14 led to a loss of filopodial projections from tubes. Scale bar represents 20 μm

Full size image
Fig. 5
figure 5

Quantification of the effects of FTI treatment on tube formation. (a) FTI treatment between days 7 and 14, during the period of tube formation. (b) Treatment between days 14 and 16, after tubes had formed. Scale bar represents 400 μm. For each condition, the total length of the tubes formed was quantified, as well as the number of branch points per unit tube length. * P < 0.05

Full size image

The data from the co-culture assays make two important contributions. First, they allow a clear demonstration of the effects of FTI treatment on endothelial cell proliferation in the context of tube formation. Second, they demonstrate additional effects of FTI treatment on tube formation that are separate from the inhibition of endothelial cell proliferation.

FTI treatment directly targets endothelial cells in tube formation and acts through inhibition of farnesyltransferase

While FTI treatment had clear effects on endothelial cell proliferation and morphology in the co-culture assay, we also saw effects on fibroblast proliferation (Supplementary Fig. 1), which could potentially contribute to the defects in tube formation. To test the effects of FTIs on endothelial cell sprouting directly, we used the well-characterized aortic ring assay—an ex vivo assay of sprouting angiogenesis. In this assay, sections of aorta are embedded in collagen matrix in the presence of VEGF. Over a period of 6 days, the endothelial cells in the vessel are stimulated to form angiogenic sprouts that grow out of the vessel and into the collagen matrix [27]. FTI treatment had a profound effect on endothelial cell sprouting in this assay, and doses equivalent to those used in the co-culture assay caused a compete inhibition of sprout formation (Fig. 6).

Fig. 6
figure 6

FTI inhibits sprouting angiogenesis in the rat aortic ring assay. (a) Representative images of aortic rings cultured with increasing does of FTI. (b) Quantification of the effects of FTI-treatment on sprouting angiogenesis. * P < 0.05, ** P < 0.01, *** P < 0.001 (compared to control)

Full size image

To confirm the specificity of the effects of FTI treatment on endothelial cell function, we wanted to target farnesyltransferase activity in another way. We also wanted to do this selectively in endothelial cells in the co-culture assay without targeting this enzyme in the fibroblasts. To do this, we modified the co-culture assay to make it accessible to RNA interference. In the modified assay, we plated endothelial cells directly onto a confluent layer of primary fibroblasts. This allowed us to carry out siRNA silencing in the isolated endothelial cells before adding them to the assay. It also reduced the time the endothelial cells were in the assay, allowing us to maintain siRNA silencing throughout the tube formation. We used two independent siRNA oligonucleotides to the farnesyltransferase β-subunit (FTβ), each of which effectively silenced FTβ expression over the time course of the assay (Fig. 7b). Downregulation of farnesyltransferase in endothelial cells caused a marked inhibition of tube formation in the co-culture assay (Fig. 7). Interestingly, silencing of farnesyltransferase also caused a reduction in the number of branches formed by these tubes (Fig. 7)—a result that was not seen with FTI treatment of co-cultures (Fig. 5).

Fig. 7
figure 7

Specific targeting of endothelial cell farnesyltransferase inhibits tube formation. Endothelial cells were treated with siRNAs targeting the farnesytransferase β-subunit (FTβ) or with siRNA targeting the control gene, lamin. The cells were then harvested and seeded onto confluent lawns of NHDFs. At the end of the experiment, the cells were fixed and stained for PECAM-1 to reveal the tubes. (a) Representative images of tube formation. Scale bar represents 400 μm. (b) Western blotting of silencing of the FTβ-subunit in endothelial cells after 72 h using two independent siRNAs (FT1 and FT2). (c) Quantification of tube length and branch formation in cells treated with or without FTβ siRNA (FT1). Essentially identical results were seen with the second siRNA, FT2 (data not shown). * P < 0.05

Full size image

Taken together, we conclude that farnesyltransferase activity in endothelial cells is directly required for efficient tube formation and that inhibition of endothelial farnesyltransferase by FTI treatment makes a significant contribution to the anti-angiogenic effects of these drugs.

Discussion

Previous studies have shown that FTI treatment can reduce the secretion of pro-angiogenic factors by tumor cells [12–15]. Here, we show that FTI treatment also directly targets endothelial function in the angiogenic process, inhibiting both cell proliferation and chemotactic migration to VEGF. The inhibition of proliferation is associated with mitotic arrest and accumulation of diploid cells. The effects of FTIs on cancer cells extend beyond targeting Ras, and other farnesylated protein targets have been shown to contribute to their actions [28]. CENP-E and CENP-F are farnesylated proteins that are required for the capture and attachment of spindle microtubules to the kinetochore during mitosis [29–31]. Treatment of rapidly dividing cancer cell lines with FTIs leads to loss of CENP-E and CENP-F from the metaphase kinetochore, leading to delays in completion of mitosis and the appearance of “lagging chromosomes” that fail to align with the metaphase plate [30, 32]. The mitotic defect seen with FTI treatment of endothelial cells is more severe, with no cells apparently progressing to metaphase. We conclude that the anti-proliferative effect of FTI treatment on endothelial cells results from a severe defect in chromosomal alignment, presumably due to the targeting of CENP-E and CENP-F. The eventual recondensation of the duplicated chromosomes would explain the high percentage of diploid cells observed and the presence of frequent chromosomal bridges in these polynucleated cells. When compared to other non-transformed cells [16], endothelial cells appear to be highly sensitive to the anti-proliferative effects of FTI treatment.

The basis of the defect in migration toward VEGF is less obvious, as we found that FTI treatment has no effect on VEGFR2 activation and does not perturb downstream signaling pathways in VEGF-stimulated cells. The co-culture assay proved to be a powerful tool in analyzing the effects of drug treatment, allowing us to break down the process into individual stages and to examine the morphology of the endothelial cells at high resolution using confocal microscopy. Using this technique, we were able to resolve the effects of FTI treatment on the proliferation of endothelial cells from the later effects on tube formation. We were also able to observe a loss of filopodia on the FTI-treated tubes. Filopodial extensions allow cells to detect chemotactic gradients more efficiently [33]. The sensory filopodia on the tip cells of vascular sprouts cluster VEGFR2 [34], and these filopodia extend toward VEGF gradients in vivo [35]. The stimulation of endothelial cell migration by VEGF requires there to be a gradient of growth factor [34], and it would seem that endothelial cell filopodia are well-placed to sense this gradient. In keeping with this, genetic targeting of these filopodia leads to defects in directed sprout migration and causes abnormal vascular patterning [36, 37]. Loss of filopodia on FTI treatment would compromise the ability of endothelial cells to sense a VEGF gradient—explaining why these cells fail to migrate toward VEGF, even though VEGF-induced signaling is intact. Interestingly, FTI treatment did not affect chemotaxis toward sphingosine-1-phosphate or FGF-2, suggesting that filopodia are not required for sensing these chemotactic factors.

Our findings strongly suggest that farnesyltransferase activity is required for physiological angiogenesis. Studies with transgenic animals have shown that the activity of farnesyltransferase is required for embryogenesis, and mice lacking the catalytic FTβ subunit die at embryonic stage 11.5 or younger [38]. These embryos are highly disorganized, making it unclear if there are specific defects in angiogenesis. In healthy adult animals, angiogenesis is restricted to the female reproductive system [39]. Interestingly, farnesyltransferase activity is not required in adult mice. Genetic ablation of the FTβ subunit in 10 day-old mice inhibits the growth of Ras oncogene-dependent tumors but has no significant effects on health or longevity [38]. In keeping with this, FTI treatment has very low toxicity in adult mice and in patients [23, 40]. The ability of adult mice lacking farnesyltransferase activity to undergo efficient angiogenesis is untested; however, these animals show a significant delay in wound healing [38], which would be an expected consequence of a defect in angiogenesis [41].

In addition to its role in tumor progression, uncontrolled angiogenesis makes a critical contribution to a range of other pathologies. This includes diseases involving ocular neovascularization, such as proliferative diabetic retinopathy and wet age-related macular degeneration [3–5]. In proliferative diabetic retinopathy, new vessels grow across the retina in response to failure of existing capillaries and consequent tissue ischemia. This vessel growth disrupts vision. Wet age-related macular degeneration is a late-onset disease where sight loss is caused by neovascularization of the sub-retinal choroid, which can also lead to macular edema. Between them, these two conditions are the leading causes of blindness in the adult population [3–5]. The development of anti-angiogenic therapies for cancer has seeded the development of anti-angiogenic therapies for ocular neovascular diseases—mainly based on ligand traps for VEGF. The first of the new treatments to be approved have already made a significant clinical impact and additional therapies are currently undergoing clinical trial [42]. Here, we have shown that farnesyltransferase inhibitors can directly target endothelial cell functions in angiogenesis, explaining the reported anti-angiogenic actions of these compounds in the rat corneal angiogenesis model [10]. We propose that the anti-angiogenic effects of FTI treatment on tumor angiogenesis result from a combination of the ability of FTIs to suppress the secretion of pro-angiogenic factors by cancer cells [12, 14, 15], and the direct effects on endothelial cells themselves. We also propose that farnesyltransferase is a valid target for the development of new anti-angiogenic therapies, and suggest that current FTIs may be effective anti-angiogenic therapies outside of the cancer context.