sign in sign up
<<
Home , Growth factor, Placental growth factor, Sunitinib

(2011) 30, 1183–1193 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE 2 regulates endothelial sensitivity to

JC Welti1, M Gourlaouen1, T Powles2, SC Kudahetti2, P Wilson3, DM Berney2 and AR Reynolds1

1Tumour Group, The Breakthrough Research Centre, The Institute of Cancer Research, London, UK; 2Centre for Molecular , Queen Mary University of London, Barts & The London School Of Medicine & Dentistry, Institute of Cancer, John Vane Science Centre, Charterhouse Square, London, UK and 3Department of Medical Oncology, West Smithfield, London, UK

The vascular endothelial growth factor (VEGF) variety of human tumours (Ellis and Hicklin, 2008). inhibitor sunitinib has been approved for VEGF-A drives angiogenesis by signalling through two first-line treatment of patients with metastatic renal cognate receptors on local endothelial cells, VEGFR1 cancer and is currently being trialled in other cancers. and VEGFR2 (Ferrara et al., 2003; Olsson et al., 2006). However, the effectiveness of this anti-angiogenic agent Binding of VEGF-A to VEGFR2 activates key intra- is limited by the presence of innate and acquired drug cellular signalling molecules including ERK1/2 and resistance. By screening a panel of candidate growth PLCg. Pathways such as these coordinate the biological factors we identified fibroblast growth factor 2 (FGF2) processes necessary for VEGF to induce the formation as a potent regulator of endothelial cell sensitivity to of new vessels (Olsson et al., 2006). sunitinib. We show that FGF2 supports endothelial Numerous inhibitors of the VEGF signalling axis proliferation and de novo tubule formation in the presence have been developed, including the VEGF-neutralising of sunitinib and that FGF2 can suppress sunitinib-induced antibody, , and VEGF receptor tyrosine retraction of tubules. Importantly, these effects of FGF2 kinase inhibitors, such as sunitinib (Ferrara et al., 2004; were ablated by PD173074, a small molecule inhibitor of Ferrara and Kerbel, 2005; Ellis and Hicklin, 2008; FGF receptor signalling. We also show that FGF2 can Kerbel, 2008). These drugs are particularly effective in stimulate pro-angiogenic signalling pathways in endothe- VEGF-driven tumours such as clear-cell renal cancer. lial cells despite the presence of sunitinib. Finally, analysis Indeed sunitinib monotherapy doubles progression-free of clinical renal-cancer samples demonstrates that a large survival and extends overall survival in this disease and proportion of renal cancers strongly express FGF2. We is now standard first-line therapy (Motzer et al., 2007, suggest that therapeutic strategies designed to simulta- 2009). Despite these impressive results, variation in neously target both VEGF and FGF2 signalling may response between individuals occurs and, more impor- prove more efficacious than sunitinib in renal cancer tantly, resistance to therapy invariably develops (Motzer patients whose tumours express FGF2. et al., 2009; Rini and Atkins, 2009). Determinants of Oncogene (2011) 30, 1183–1193; doi:10.1038/onc.2010.503; sunitinib response and the mechanisms that mediate the published online 8 November 2010 onset of resistance to sunitinib are poorly understood. This presents a significant barrier to the development of Keywords: angiogenesis; growth factors; sunitinib; renal individualized therapy. cancer; resistance The tumour microenvironment contains numerous soluble factors, besides VEGF, that possess pro- angiogenic properties. These include growth factors that signal through tyrosine kinase receptors, such as Introduction members of the fibroblast growth factor (FGF), (PLGF) and -derived Tumour angiogenesis (the growth of new blood vessels growth factor (PDGF) families and factors that into tumours) is a key driver of tumour progression and signal through non-tyrosine kinase receptors, such as has received considerable attention as a therapeutic -8 (IL-8) and transforming growth-factor-b target (Hanahan and Folkman, 1996; Kerbel and (Relf et al., 1997; Ferrara, 2002; Presta et al., 2005). Folkman, 2002; Ferrara and Kerbel, 2005). One of the These factors may have an important role in the most highly studied pro-angiogenic growth factors is induction of tumour angiogenesis and may also be vascular endothelial growth factor (VEGF), particularly involved in altering the responsiveness of tumour blood the VEGF-A isoform, which is highly expressed in a vessels to inhibitors of the VEGF signalling axis (Bergers and Hanahan, 2008). In support of this latter concept, Correspondence: Dr AR Reynolds, Tumour Angiogenesis Group, FGF2 is upregulated in a mouse tumour model of acquired The Breakthrough Breast Cancer Research Centre, The Institute of resistance to DC101 (a VEGFR2 inhibitory antibody) Cancer Research, 237 Fulham Road, London SW3 6JB, UK. and inhibition of FGF2 in combination with DC101 led E-mail: [email protected] Received 21 June 2010; revised 5 August 2010; accepted 3 September to improved anti-tumour responses (Casanovas et al., 2010; published online 8 November 2010 2005). Moreover, inhibition of PLGF or PDGF has been FGF2 regulates sunitinib sensitivity JC Welti et al 1184 reported to enhance the anti-tumour efficacy of VEGF absence of sunitinib (Figure 1c). These data therefore pathway inhibitors in some murine tumour models identify FGF1, and more strikingly FGF2, as candidate (Bergers et al., 2003; Fischer et al., 2007; Crawford et al., factors that may be potent mediators of endothelial cell 2009). Another recent study demonstrated that IL-8 is resistance to sunitinib. upregulated in sunitinib-resistant tumour xenografts and that inhibition of IL-8 can re-sensitize these resistant tumours to sunitinib (Huang et al., 2010b). FGF2 suppresses the anti-angiogenic activity of sunitinib Sunitinib is an orally available tyrosine kinase inhibitor in a 3-dimensional assay of endothelial tube formation that potently inhibits several angiogenesis-related tyrosine As FGF2 was identified as the most potent factor able to kinase receptors that is, VEGF receptors (VEGFR1, drive resistance to sunitinib in an endothelial cell VEGFR2 and VEGFR3), PDGF receptors (PDGFRa proliferation assay (Figure 1c), we proceeded to validate and PDGFRb), and also the KIT receptor (Abrams this activity within an in vitro model of angiogenesis. et al., 2003; Mendel et al., 2003; O’Farrell et al., 2003). Human umbilical vein endothelial cells (HUVEC)- A recent study that examined the activity of sunitinib coated beads were embedded in a fibrinogen gel and against a panel of 317 kinases (representing 450% of then incubated with VEGF and different concentrations the known kinome) revealed that sunitinib has relatively of sunitinib. After 7 days, endothelial tubules had high affinity for at least 40 other kinases, including sprouted from VEGF alone-treated beads, but tubule myosin -chain kinase, MEK1 and MEK2 (Karaman formation was significantly inhibited by low concentra- et al., 2008). As activation of multiple kinase-signalling tions (10–100 nM) of sunitinib (Figures 2a and b) pathways is a requirement for growth factor-mediated confirming that sunitinib can inhibit VEGF-stimulated stimulation of angiogenesis, the relatively broad kinase tubule formation in this assay. Importantly, when the substrate specificity of sunitinib might be expected to assay was repeated in the presence of VEGF and FGF2, limit the potential for other growth factors to drive the number of tubules was enhanced 3–4 fold compared resistance to sunitinib through a mechanism that with beads cultured with VEGF alone (Figure 2c). involves direct stimulation of endothelial cells. These data are in agreement with previous reports Here we identify FGF2 to be a potent mediator of showing that FGF2 can enhance VEGF-mediated endothelial cell resistance to sunitinib. We further angiogenesis (Pepper et al., 1992; Giavazzi et al., demonstrate that FGF2 can suppress the anti-angio- 2003). Sunitinib was also able to suppress tubule genic activity of sunitinib by directly stimulating pro- formation in the presence of VEGF and FGF2 angiogenic signalling in endothelial cells. We also show (Figure 2c). However, tubule formation in the presence that FGF2 is expressed in a subset of human renal of VEGF, FGF2 and sunitinib was still significantly cancers. These data provide strong support for simulta- enhanced compared with tubule formation in the neous targeting of VEGF and FGF signalling as a presence of VEGF and sunitinib (Figure 2c). These therapeutic strategy in renal cancer. data confirm that FGF2 can act to suppress the anti- angiogenic effects of sunitinib. Having shown that FGF2 can suppress the effects of sunitinib on de novo tubule formation, we then Results addressed whether FGF2 could suppress the effects of sunitinib on pre-formed tubules. HUVEC-coated beads A screen of pro-angiogenic growth factors identifies FGF2 embedded in a fibrinogen gel were incubated from the as a mediator of endothelial cell resistance to sunitinib start of the assay with VEGF in the presence or absence Sunitinib is a potent inhibitor of VEGFR2 tyrosine kinase of FGF2. After 7 days, when tubules were clearly activity. In agreement with previous reports (Mendel visible, the medium was then additionally supplemented et al., 2003), nanomolar concentrations (10–100 nM)of with either vehicle or 10 nM sunitinib. The behaviour of sunitinib inhibited VEGF-induced phosphorylation of individual tubules was monitored by time-lapse micro- VEGFR2 in human endothelial cells (Figure 1a) and scopy for the following 48 h. In cultures incubated with VEGF-induced endothelial (Figure 1b). VEGF and vehicle, tubules grew in length through These concentrations are clinically relevant, as circulating extension at the tip of the tubule (Figure 2d and concentrations of sunitinib measured in patients are in the Supplementary Movie S1). In contrast, in cultures nanomolar range and do not typically exceed 200 nM incubated with VEGF and 10 nM sunitinib, tubules (Faivre et al., 2006; Britten et al., 2008). rapidly retracted (Figure 2e and Supplementary Movie To identify growth factors potentially involved in S2). These data are consistent with previously published resistance to sunitinib, we screened a panel of 19 pro- data showing that inhibitors of VEGF receptor tyrosine angiogenic growth factors for their ability to promote kinase activity can induce the retraction of existing endothelial proliferation in cell cultures incubated with vessels (Mancuso et al., 2006). However, in the VEGF and 100 nM sunitinib (Figure 1c). Although most combined presence of VEGF, FGF2 and 10 nM suniti- of the growth factors tested were able to induce a small nib, tubule retraction was abrogated (Figure 3f and degree of endothelial cell proliferation in the presence of Supplementary Movie S3). We quantified the ability of sunitinib, only FGF1 and FGF2 were able to restore FGF2 to suppress tubule retraction, by measuring the endothelial cell proliferation to levels that matched change in length of tubules after addition of 10, 25, 50 or (FGF1) or exceeded (FGF2) those observed in the 100 nM sunitinib. In the presence of VEGF alone,

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1185

Figure 1 A screen of pro-angiogenic growth factors identifies FGF2 as a potent mediator of endothelial cell resistance to sunitinib. (a) Nanomolar concentrations of sunitinib inhibit VEGF-stimulated phosphorylation of VEGFR2 in endothelial cells. HUVECs were stimulated with 10 ng/ml VEGF for 10 min in the presence of the indicated concentration of sunitinib. Lysates were analysed by western blotting for phosphorylated VEGFR2, total VEGFR2 and HSC-70 for loading control. (b) Nanomolar concentrations of sunitinib inhibit VEGF-induced proliferation of endothelial cells. Proliferation of HUVECs was assessed after 72 h incubation in the presence of 100 ng/ml VEGF and the indicated concentrations of sunitinib. Graph shows mean surviving fraction normalised to vehicle control±s.e.m. (c) A screen of pro-angiogenic growth factors demonstrates that FGF2 can suppress the anti-angiogenic activity of sunitinib. Proliferation of HUVECs was assessed after 72 h in the presence of no growth factor (no GF), 100 ng/ml VEGF and vehicle (VEGF þ vehicle), 100 ng/ml VEGF and 100 nM sunitinib (VEGF þ sun) or in the presence of 100 ng/ml VEGF, 100 nM sunitinib and 50 ng/ml of the indicated growth factor (for example, VEGF þ sun þ FGF2). Graph shows mean surviving fraction normalised to vehicle control±s.e.m. Ang1, 1; Ang2, angiopoietin 2; EGF, ; EPO, ; FGF1, fibroblast growth factor 1; FGF2, fibroblast growth factor 2; FGF5, fibroblast growth factor 5; HGF, ; IGF1, like growth factor 1; IGF2, insulin like growth factor 2; IL8, interleukin-8; PDGFAA, platelet derived growth factor AA; PDGFAB, platelet derived growth factor AB; Ple, ; PLGF1, placental growth factor 1; TGFb, transforming growth factor beta; TNFa, tumour necrosis factor alpha; TNFb, tumour necrosis factor beta; SDF1a, stromal derived factor 1 alpha. sunitinib induced tubule retraction in a dose-dependent molecule tyrosine kinase inhibitor PD173074 is a potent manner (Figure 2g). However, in the presence of VEGF inhibitor of FGF activity and FGF2, the ability of sunitinib to induce tubule (Mohammadi et al., 1998; Pardo et al., 2009). We tested retraction was significantly suppressed (Figure 2g). the relative ability of sunitinib and PD173074 to inhibit These data show that low concentrations of sunitinib VEGF- and FGF2-mediated endothelial cell prolifera- (10–100 nM) can induce the retraction of pre-formed tion. Inhibition of endothelial cell proliferation was tubules, but that this effect is suppressed by FGF2. assessed in the presence of 100 ng/ml VEGF or in the presence of 2.5 ng/ml FGF2, which is an FGF2 concentration that induces equivalent levels of endothe- PD173074 inhibits FGF2-mediated resistance to sunitinib lial proliferation to 100 ng/ml VEGF in our assay. We then tested whether the effects of FGF2 on Sunitinib concentrations in the range of 10–100 nM endothelial cell proliferation could be suppressed using inhibited VEGF-stimulated proliferation of endothelial an inhibitor of FGF receptor signalling. The small cells, but did not inhibit FGF2-stimulated proliferation

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1186

Figure 2 FGF2 suppresses the anti-angiogenic activity of sunitinib in an in vitro assay of endothelial tubule formation. (a, b) Sunitinib inhibits VEGF-stimulated endothelial tubule formation in vitro. Representative images of HUVEC coated beads (a) demonstrating tubule formation in cultures incubated for 7 days in the presence of VEGF plus vehicle (left) and inhibition of tubule formation in cultures incubated for 7 days with VEGF plus 100 nM sunitinib (right). (b) Graph shows dose-dependent inhibition of tube formation by sunitinib. Mean number of tubules per bead normalised to vehicle control±s.e.m. is shown (n ¼ 30 beads per data point). *Po0.05, **Po0.01 (Student’s t-test). (c) FGF2 enhances VEGF-mediated endothelial tubule formation in the presence of vehicle or sunitinib. Graph shows that tubule formation in cultures incubated for 7 days with VEGF and FGF2 is enhanced compared with cultures incubated for 7 days with VEGF alone, both in the presence of vehicle and in presence of the indicated concentrations of sunitinib. Mean number of tubules per bead normalised to vehicle control±s.e.m is shown (n ¼ 30 beads per data point). *Po0.05, **Po0.01 (Student’s t-test). (d–g) FGF2 suppresses sunitinib-induced retraction of pre-formed endothelial tubules in vitro. HUVEC- coated beads were incubated with VEGF or VEGF and FGF2 for 7 days. Cultures were then supplemented with vehicle or sunitinib at the indicated concentrations for a further 48 h to permit changes in tubule length to be assessed. Time-lapse microscopy demonstrates that VEGF induces tubule extension (d), the presence of sunitinib induces tubule retraction (e) and that FGF2 suppresses sunitinib-induced tubule retraction (f). Representative frames are shown at 0, 24 or 48 h. Arrows indicate the tips of tubules. (g) Quantification of change in tubule length after the indicated concentration of sunitinib was added for 48 h. Bars represent the percent increase or percent decrease in tubule length±s.e.m. (n ¼ 30 individual tubules per data point). *Po0.05, **Po0.01 (Student’s t-test).

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1187 of endothelial cells (Figure 3a). Conversely, PD173074 did not inhibit VEGF-stimulated proliferation of concentrations in the range of 10–100 nM inhibited endothelial cells (Figure 3b). These data confirm the FGF-stimulated proliferation of endothelial cells, but ability of sunitinib to selectively inhibit VEGF-mediated endothelial cell proliferation and the ability of PD173074 to selectively inhibit FGF2-mediated en- dothelial cell proliferation in this assay. We then tested whether PD173074 could restore sunitinib sensitivity to endothelial cells cultured in the presence of VEGF and FGF2. The combination of VEGF and FGF2 resulted in enhanced proliferation of endothelial cells compared with cells cultured with VEGF alone and also reduced the sensitivity of cells to sunitinib (Figure 3c). However, addition of 50 nM PD173074 to cells cultured with FGF2 and VEGF prevented the FGF2-mediated enhancement in endo- thelial cell proliferation and restored sensitivity to sunitinib (Figure 3c). The effects of PD173074 on tubule retraction were also examined. As observed previously, the presence of FGF2 inhibited the retraction of pre- formed tubules (Figure 3d). However, the addition of 50 nM PD173074 abrogated this effect of FGF2 and restored sunitinib sensitivity in this assay (Figure 3d). These data show that an inhibitor of FGF receptor signalling can prevent FGF2 from suppressing the sensitivity of endothelial cells to sunitinib.

FGF2 activates pro-angiogenic signalling pathways in endothelial cells in the presence of sunitinib Having confirmed that FGF2 can suppress the anti- angiogenic activity of sunitinib, we then addressed whether FGF2 can activate pro-angiogenic signalling pathways in endothelial cells in the presence of sunitinib. Stimulation of HUVECs with VEGF alone promoted strong phosphorylation of MEK1/2 and ERK1/2, but this was suppressed by sunitinib (Figure 4). In contrast, Figure 3 PD173074 inhibits FGF2-mediated resistance to sunitinib. in the presence of FGF2 alone or FGF2 and VEGF (a) Sunitinib inhibits VEGF-stimulated endothelial cell proliferation, combined, strong phosphorylation of MEK1/2 and but not FGF2-stimulated endothelial cell proliferation. VEGF- or ERK1/2 occurred in both the presence and absence of FGF2-stimulated proliferation of HUVECs was assessed after 72 h in the presence of the indicated concentrations of sunitinib. Graph sunitinib (Figure 4). Stimulation of HUVECs with shows mean surviving fraction normalised to vehicle control±s.e.m. VEGF alone also promoted phosphorylation of PLCg, (b) PD173074 inhibits FGF2-stimulated endothelial cell prolifera- which again was suppressed by sunitinib (Figure 4). tion, but not VEGF-stimulated endothelial cell proliferation. VEGF- Addition of FGF2 alone to HUVECs also promoted or FGF2-stimulated proliferation of HUVECs was assessed after phosphorylation of PLCg, although activation of this 72 h in the presence of the indicated concentrations of PD173074. Graph shows mean surviving fraction normalised to vehicle control± pathway appeared to be weaker than that observed with s.e.m. (c). PD173074 restores sunitinib sensitivity to endothelial cells VEGF. However, the weak activation of PLCg by cultured in the presence of FGF2. Inhibition of HUVEC proliferation FGF2 was not blocked by sunitinib. These data show by sunitinib was assessed after 72 h in the presence of (1) VEGF plus that although sunitinib can inhibit VEGF-mediated vehicle, (2) VEGF, FGF2 plus vehicle or (3) VEGF, FGF2 plus activation of ERK1/2 and PLCg, sunitinib does not 50 nM PD173074. Graph shows mean-surviving fraction normalised to vehicle control±s.e.m. In the combined presence of VEGF, FGF2 block FGF2-mediated activation of ERK1/2 and PLCg. and sunitinib endothelial cell proliferation was significantly enhanced compared with proliferation in the presence of VEGF and sunitinib alone (Po0.0001 ANOVA test), but this effect was ablated in the FGF2 expression in human renal cancer samples presence of PD173074. (d) PD173074 prevents effect of FGF2 on Sunitinib has been approved for the first-line treatment sunitinib-induced retraction of tubules. HUVEC-coated beads were of patients with metastatic renal cancer. Renal cancers incubated with VEGF or VEGF and FGF2 for 7 days. Cultures were then supplemented with vehicle, sunitinib (sun) and/or PD173074 are highly angiogenic and sunitinib is proposed to (PD) at the indicated concentrations for a further 48 h. Quantification suppress the growth of these tumours mainly through of change in tubule length after the indicated concentration of sunitinib inhibition of angiogenesis (Faivre et al., 2007; Huang was added for 48 h is shown. Bars represent the percent increase et al., 2010a). As our data identified FGF2 as a potent or percent decrease in tubule length that occurred±s.e.m. (n ¼ 30 individual tubules per data point). FGF2 suppressed sunitinib-induced suppressor of sunitinib’s anti-angiogenic activity, we tubule retraction (*Po0.05, **Po0.01, Student’s t-test) and this effect assessed the expression of FGF2 in human renal cancer was ablated by PD173074 (NS, no significant difference). samples. An FGF2-specific antibody was used to stain a

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1188 Table 1 Patient characteristics in the renal cancer TMA Number of patients 74 Mean age (age range) 61 (51–77)

Gender Male 57 Female 17

Histology Clear cell 74 Other 0

Number of metastatic sites 116 226 318 414

Risk factor (MSKCC) Good 27 Intermediate 24 Poor 23

Site of disease Lung 63 15 15 27

Figure 4 FGF2 activates ERK1/2 and PLCg signalling in et al., 2001; Fukata et al., 2005). Moreover, we clearly endothelial cells in the presence of sunitinib. HUVECs were left show that FGF2 expression in advanced renal cancer unstimulated or were stimulated with 25 ng/ml VEGF and/or can be associated with the tumour compartment or the 25 ng/ml FGF2 for 5 min in the presence of vehicle or 100 nM vascular compartment. sunitinib as indicated. Lysates were analysed by western blotting for phosphorylated VEGFR2, total VEGFR2, phosphorylated MEK, total MEK, phosphorylated ERK1/2, total ERK1/2, phosphorylated PLCg, total PLCg and HSC-70 for loading Discussion control. VEGF-stimulated activation of MEK, ERK1/2 and PLCg is inhibited by 100 nM sunitinib. However, sunitinib did not inhibit FGF2-stimulated activation of these pathways. Resistance to therapy is a major limiting factor in the successful treatment of cancers with anti-angiogenic agents that target the VEGF signalling axis (Bergers and Hanahan, 2008; Ebos et al., 2009). Stimulation of tissue microarray (TMA) composed of renal cancer tumour angiogenesis by other pro-angiogenic growth samples from a cohort of 74 patients with advanced factors in the tumour microenvironment may be one renal cancer who had progressed on conventional mechanism through which this resistance occurs. Pre- therapy, but who did not receive sunitinib vious work has shown that inhibition of FGF2 or therapy (more detailed information on these patients PLGF can overcome acquired resistance to VEGFR2 can be found in Table 1 and the Materials and methods inhibition in mouse tumour models (Casanovas et al., section). Interestingly, we found that FGF2 staining was 2005; Fischer et al., 2007). However, to the best of prominent in both tumour cells and in tumour blood our knowledge, the direct ability of growth factors vessels (Figures 5a–d and Table 2). Tumour-cell expres- to stimulate angiogenesis in human endothelial cells sion of FGF2 was localized predominantly to the nuclei when VEGFR2 is inhibited by sunitinib has not been of tumour cells (for an example see Figure 5a) and was clearly demonstrated. Here we screened a panel of 19 detectable in 36 of the 74 cases examined and absent pro-angiogenic factors and revealed that FGF2 can from the remaining 38 cases (for an example of negative potently suppress the anti-angiogenic activity of suniti- staining see Figure 5b). Staining of the tumour nib. We show that FGF2 is able to directly stimulate vasculature was localized to the cytoplasm of endothe- endothelial cell proliferation and endothelial tubule lial cells or associated with the vascular basement formation in the presence of sunitinib. We also show membrane (for an example see Figure 5c) and was that FGF2 can suppress sunitinib-induced retraction of detected in 41 of the 74 cases examined, but was absent pre-formed endothelial cell tubules. These data suggest from the remaining 33 cases (for an example of negative that the pro-angiogenic function of FGF2 may be staining see Figure 5d). These data confirm previous directly relevant for resistance to sunitinib. work showing that FGF2 is expressed in human renal It was surprising to find that the other pro-angiogenic cancers (Eguchi et al., 1992; Nanus et al., 1993; Slaton factors we tested showed significantly less activity in

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1189

Figure 5 Expression and localisation of FGF2 in human renal cancers. (a–d) FGF2 staining is evident in both renal cancer cells and the vasculature of human renal cancer. A tissue microarray comprising cores from 74 renal cancer patients was stained with an anti- FGF2 antibody. Shown are representative cores in which tumour nuclei stained positive (a) or negative (b) for FGF2 and cores in which tumour blood vessels stained positive (c) and negative (d) for FGF2. Arrowheads indicate blood vessels in (c, d). Note that in the example cores shown in (c, d) there are also FGF2-positive tumour nuclei apparent.

Table 2 Localisation of FGF2 staining scored in the renal cancer influence the response of endothelial cells to sunitinib TMA in vivo. Moreover, some growth factors may mediate Localisation of FGF2 staining No. of %of resistance to sunitinib by acting on other cell types, such patients patients as pericytes, tumour cells or infiltrating immune cells. Therefore, complementary work using appropriate Tumour nuclei positive 36 48.6 in vivo models of resistance to sunitinib may demon- Vessel positive 41 55.4 Tumour nuclei positive and vessel negative 14 18.9 strate a role for other growth factors in mediating Tumour nuclei negative and vessel positive 19 25.6 tumour resistance to sunitinib. Tumour nuclei positive and vessel positive 22 29.7 It is known that VEGF and FGF2 can activate Tumour nuclei negative and vessel negative 19 25.6 similar signalling pathways in endothelial cells in vitro, such as the Ras–Raf–MEK–ERK1/2 pathway and the Scoring of the renal cancer TMA is shown based on FGF2 localisation in all 74 patient samples. Row 1 shows the total number of samples PLCg–PKC pathway (Cross and Claesson-Welsh, 2001; that scored positive for tumour-nuclei staining. Row 2 shows the total Presta et al., 2005; Olsson et al., 2006). Moreover, number of samples that scored positive for vessel staining. In rows 3–6, essential roles for both of these pathways in the process the patients are subdivided into four groups based on positive and of angiogenesis have been demonstrated (Presta et al., negative staining in both compartments. 1991; Eliceiri et al., 1998; Meyer et al., 2003; Mavria et al., 2006). Here we show that although sunitinib can inhibit VEGFR2-mediated activation of ERK1/2 and comparison with FGF2. This may reflect a limitation of PLCg, sunitinib did not prevent FGF2-mediated activa- our in vitro screening approach and we cannot exclude tion of ERK1/2 and PLCg. These data suggest that the possibility that some of the angiogenic factors FGF2 can bypass inhibition of VEGFR2 signalling by screened may indeed have a role in resistance to activating at least two key pro-angiogenic signalling sunitinib in vivo. Of note, a recent study demonstrated pathways. It is likely that FGF2-mediated angiogenesis a role for IL-8 in mediating resistance to sunitinib in vivo also involves activation of other signaling pathways that (Huang et al., 2010b), whereas our screen did not we have not examined here. Although we do not suggest reveal a role for IL-8. It is likely that tumour-specific that FGF2-mediated activation of ERK1/2 and PLCg is microenvironmental effects, which are not recapitulated sufficient for FGF2 to activate angiogenesis in the in our assay, may permit certain growth factors to presence of sunitinib, these data demonstrate the

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1190 principle that pro-angiogenic signalling in response to suggest that therapeutic strategies designed to simulta- FGF2 can take place in endothelial cells in the presence neously target both VEGF and FGF2 signalling may of sunitinib. prove more efficacious than sunitinib in renal cancer We examined the expression of FGF2 in tumour patients whose tumours express FGF2. samples obtained from metastatic renal cancer patients. As well as confirming that FGF2 is commonly expressed in these tumours, we found that FGF2 can be Materials and methods differentially localized in patient samples. Expression of FGF2 in tumour nuclei was observed in 49% of the Reagents samples, whereas expression of FGF2 in tumour blood reagents were obtained from the following sources: vessels could be demonstrated in 55% of the samples. fetal calf (FCS) (Gibco, Invitrogen Ltd, Paisley, UK) Expression of FGF2 in the nuclei of tumour cells in and endothelial cell from bovine brain (Serotech, human cancer specimens has been reported in other Kidlington, Oxfordshire, UK). Antibodies were obtained from studies (Eguchi et al., 1992; Joy et al., 1997) and the following sources: total MEK (BD Pharmingen, San Jose, functional studies have shown that fibroblast growth CA, USA), pThr202/pTyr201-ERK1/2, pSer217/pSer221-MEK, pTyr783-PLCg,PLCg, pTyr1175-VEGFR2, VEGFR2 (Cell factors can have a signalling role in the nucleus (Bryant Signalling Technology, Danvers, MA, USA), FGF2 (Peprotech, and Stow, 2005). The origins of the vascular FGF2 London, UK), HSC-70 (Santa Cruz Biotechnology, Santa Cruz, staining we observed are currently unclear. As tumours CA, USA). All growth factors utilised were obtained from can secrete FGF2, this staining pattern may conceivably Peprotech, except for angiopoietin 2 and erythropoietin that be due to the accumulation of secreted FGF2 in and were obtained from R&D Systems (Abingdon, Oxfordshire, around tumour blood vessels. However, 26% of the UK). Sunitinib malate was obtained from LC laboratories cases we examined showed positive FGF2 staining in the (Woburn, MA, USA) and PD173074 was obtained from vasculature, whereas the tumour cells in the same Sigma (Poole, Dorset, UK). Unless otherwise stated, all other sample were negative for FGF2 staining (see Table 2). reagents were obtained from Sigma. Therefore, it is possible that the vascular FGF2 staining we observed may represent selective upregulation of Cell culture FGF2 expression in the tumour . In support HUVECs from pooled donors (TCS Cell Works, Buckingham, of this, expression of FGF2 by endothelial cells and Buckinghamshire, UK) were cultured in HUVEC-specific medium autocrine signalling of FGF2 in endothelial cells has (M199 medium supplemented with 20% FCS, 20 mg/ml endo- thelial cell mitogen, 10 mg/ml heparin and antibiotics). Cells been previously described (Schweigerer et al., 1987; were used for experiments at passage 4–8. Normal human Gualandris et al., 1996; Pintucci et al., 2002). It will now dermal fibroblast cells (TCS Cell works) were cultured in be important to examine whether the expression and Dulbecco’s Modified Eagle Medium supplemented with 10% localisation of FGF2 in renal cancers has any role in FCS and antibiotics. determining the response of renal cancer patients to sunitinib. If tumour or vascular expression of FGF2 in HUVEC proliferation assays renal cancer has a negative impact on response to HUVECs were plated on 96-well plates at a density of 1000 sunitinib, then this could represent a potential biomar- cells per well in HUVEC-specific medium. The next day the ker to identify renal cancer patients who are unlikely to medium was changed for M199 plus 10% FCS, supplemented respond to sunitinib. These patients might then be with 100 ng/ml VEGF and with sunitinib or PD173074 at the considered for treatment with one of the alternative indicated concentration or vehicle (0.1% DMSO; dimethyl targeted agents currently being tested in renal cancer, sulfoxide). For the purpose of the growth factor screen, cells such as novel agents that are designed to inhibit both were incubated with M199 medium plus 10% FCS, supplemented with 100 ng/ml VEGF, 50 ng/ml of an alternative growth factor VEGF receptor and FGF receptor signalling. and 100 nM sunitinib. After 72 h, cell viability was quantified using Several kinase inhibitors which target both VEGF the Cell Titre Glow cell-viability reagent (Promega, Southampton, and FGF receptor signalling are currently being trialled Hampshire, UK), according to the manufacturer’s instructions. in patients (Turner and Grose, 2010). Although our The plates were read in a luminescence plate reader (PerkinElmer, data suggest that this may be an attractive thera- Cambridge, Cambridgeshire, UK). peutic approach in renal cancer, it remains to be seen whether targeting both of these pathways in renal cancer Immunoblotting will be more efficacious than targeting the VEGF HUVECs were plated on 100 mm diameter plates (at a density pathway alone. The efficacy of this approach will likely of 5 Â 105 per plate) and cultured for 48 h. The cells were depend on several factors, including (1) the relative washed three times in serum-free M199 medium and then importance of VEGF and FGF2 signalling as onco- starved for 3 h in serum-free M199 medium. Treatment with genic drivers in a particular tumour, (2) the presence of the indicated concentration of sunitinib was applied 15 min alternative resistance mechanisms in a particular tumour before stimulation with the indicated concentrations of VEGF and (3) the level of toxicity experienced by patients and/or FGF2. After 5 or 10 min, plates were transferred to ice, washed twice with ice-cold phosphate-buffered saline and treated with these novel-kinase inhibitors. then lysates were collected in lysis buffer (1% IGEPAL In conclusion, our data suggest that the pro-angio- CA-630, 20 mM Tris pH 7.5, 150 mM NaCl, 10% glycerol, 1 mM genic function of FGF2 may be relevant for resistance to Na3VO4,10mM NaF, 1 mM 4-(2-aminoethyl) benzynesulphonyl sunitinib. This may be especially important in renal fluoride, 0.8 mM aprotinin, 0.05 mM bestatin, 0.015 mM E-64, cancer, in which FGF2 expression is prominent. We 0.02 mM leupeptin and 0.01 mM pepstatin). Cell lysates were mixed

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1191 with reducing Laemmli sample buffer, separated on 7% or with VEGF alone or VEGF and FGF2, as indicated, for 10% SDS–polyacrylamide gels and transferred onto Hybond– 7 days in an incubator with 371C/10% CO2. Plates were then ECL nitrocelluose membranes (GE Healthcare Life Sciences, transferred to the stage of the inverted Leica IX-70 microscope Little Chalfont, Buckinghamshire, UK) over night. Membranes and photomicrographs were captured of individual tubules were blocked at room temperature for 1 h in TBS-T (100 mM (time point ¼ 0 h). The medium was then supplemented with Tris–HCl, 0.2 M NaCl, 0.1% Tween 20 v/v) containing 5% milk sunitinib, PD173074 or vehicle (0.1% DMSO), as indicated, (Marvel, Premier International Food, Spalding, Lincolnshire, and plates returned to the incubator. After 48 h, photomicro- UK) followed by primary antibody incubation for 1 h at room graphs of the same tubules were recorded again (time temperature or overnight at 41C in TBS-T containing 5% bovine point ¼ 48 h). Sprout length was quantified from the photo- serum albumin. After washing, incubation with horse radish micrographs recorded at 0 and 48 h by using image analysis peroxidase-conjugated secondary antibodies was performed software (Adobe Photoshop, Uxbridge, UK). Percentage for 1 h at room temperature in TBS-T containing 5% milk. change in tubule length was calculated using the formula: Signal was visualized with the enhanced chemiluminescence 100 Â (a/b)À1, where a is the tubule length at 48 h and b the reagent (GE Healthcare Life Sciences). tubule length at 0 h.

Endothelial tube formation assays TMA Tube formation assays were performed using a modified Renal cancer samples were collected from patients with version of a previously published protocol (Nakatsu and advanced renal cancer who were enroled in a Phase III clinical Hughes, 2008). In brief, HUVECs were cultured for 24 h in trial of versus therapy (Ravaud et al., EGM2 complete medium (Lonza, Slough, Berkshire, UK) and 2008). Tissues were collected from patients before commencing then cells were trypsinised and resuspended in EGM2 at a treatment. The tissue was fixed in formalin, embedded in concentration of 2 Â 106 HUVECs per ml. Cytodex3 beads paraffin and used to generate a TMA consisting of three cores (GE healthcare Life Sciences) were coated with HUVECs by from each patient sample. For the purpose of this study, we incubating 2 Â 106 HUVECs with 3 Â 104 beads for 4 h with only analysed renal cancer samples of confirmed clear cell gentle agitation every 20 min. Coated beads were then diluted histology that were collected from 74 patients in the hormone to a volume of 5 ml in EGM2 and placed in a tissue-culture therapy arm of the trial (patients received daily megesterol incubator overnight. The next day, beads were washed in acetate or daily tamoxifen until disease progression or with- EGM2 and resuspended in 50 ml of sterile 2 mg/ml fibrinogen drawal from the trial). Following de-waxing and processing for in phosphate-buffered saline. A volume of 500 ml of bead- antigen retrieval, the TMA was stained with a rabbit anti- fibrinogen solution was then transferred to each well of a human FGF2 antibody (Peprotech) that was then detected 24-well plate (approximately 100 beads per well) containing using horse radish peroxidase-conjugated secondary antibo- 6.25 ml 50 U/ml thrombin. Fibrinogen gels were allowed to set dies and DAB staining. Scoring of the TMA was performed by for 15 min at 371C and then 2 Â 104 fibroblasts in EGM2 were two observers: Daniel Berney (a genitourinary pathologist) seeded on top of each well. After 3 h, medium was exchanged and Andrew Reynolds (the lead investigator). Scoring for EBM2 (Lonza) plus 2.5% FCS supplemented with growth was recorded separately for FGF2 expression in tumour nuclei factors (10 ng/ml VEGF and/or 50 ng/ml FGF2), sunitinib, and FGF2 expression in the vasculature. Individual patient PD173074 or vehicle (0.1% DMSO) as indicated. The plates cases were scored as positive if one or more of the three cores were re-fed every 2 days. from each patient stained positive for nuclear FGF2 or vascular FGF2. Analysis of tube formation assays The quantity of tubules that sprouted from HUVEC beads was Conflict of interest measured after 7 days of culture by fixing beads in 4% formalin and then counting the number of sprout tips per bead Thomas Powles is the recipient of an educational research under an inverted light microscope. A total of 15 beads were grant from Pfizer Global Pharmaceuticals. The other authors counted per well and all conditions were counted in duplicate declare no potential conflict of interest. wells. For time-lapse microscopy, tubule formation assays were established as described above and after 7 days the medium was supplemented with sunitinib or vehicle (0.1% DMSO) as indicated. The plates were then transferred to the Acknowledgements stage of an inverted Leica IX-70 microscope fitted with a heated chamber and a supply of 10% CO2. Images were We would like to thank Alan Ashworth, Clare Isacke and captured every 20 min for a total of 48 h. For the quantifica- Nicholas Turner for critical comments on the paper and tion of tubule extension/retraction, cultures were incubated Breakthrough Breast Cancer for research funding.

References

Abrams TJ, Lee LB, Murray LJ, Pryer NK, Cherrington JM. (2003). Bergers G, Hanahan D. (2008). Modes of resistance to anti-angiogenic SU11248 inhibits KIT and platelet-derived therapy. Nat Rev Cancer 8: 592–603. beta in preclinical models of human small cell lung cancer. Mol Britten CD, Kabbinavar F, Hecht JR, Bello CL, Li J, Baum C et al. Cancer Ther 2: 471–478. (2008). A phase I and pharmacokinetic study of sunitinib Bergers G, Song S, Meyer-Morse N, Bergsland E, Hanahan D. administered daily for 2 weeks, followed by a 1-week off period. (2003). Benefits of targeting both pericytes and endothelial cells in Cancer Chemother Pharmacol 61: 515–524. the tumor vasculature with kinase inhibitors. J Clin Invest 111: Bryant DM, Stow JL. (2005). Nuclear translocation of cell-surface 1287–1295. receptors: lessons from fibroblast growth factor. Traffic 6: 947–954.

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1192 Casanovas O, Hicklin DJ, Bergers G, Hanahan D. (2005). Drug Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge CE, resistance by evasion of antiangiogenic targeting of VEGF Campbell BT et al. (2008). A quantitative analysis of kinase signaling in late-stage pancreatic islet tumors. Cancer Cell 8: inhibitor selectivity. Nat Biotechnol 26: 127–132. 299–309. Kerbel R, Folkman J. (2002). Clinical translation of angiogenesis Crawford Y, Kasman I, Yu L, Zhong C, Wu X, Modrusan Z et al. inhibitors. Nat Rev Cancer 2: 727–739. (2009). PDGF-C mediates the angiogenic and tumorigenic proper- Kerbel RS. (2008). Tumor angiogenesis. N Engl J Med 358: 2039–2049. ties of fibroblasts associated with tumors refractory to anti-VEGF Mancuso MR, Davis R, Norberg SM, O0Brien S, Sennino B, treatment. Cancer Cell 15: 21–34. Nakahara T et al. (2006). Rapid vascular regrowth in tumors after Cross MJ, Claesson-Welsh L. (2001). FGF and VEGF function in reversal of VEGF inhibition. J Clin Invest 116: 2610–2621. angiogenesis: signalling pathways, biological responses and ther- Mavria G, Vercoulen Y, Yeo M, Paterson H, Karasarides M, Marais apeutic inhibition. Trends Pharmacol Sci 22: 201–207. R et al. (2006). ERK-MAPK signaling opposes Rho-kinase to Ebos JM, Lee CR, Kerbel RS. (2009). Tumor and host-mediated promote endothelial cell survival and sprouting during angiogenesis. pathways of resistance and disease progression in response to Cancer Cell 9: 33–44. antiangiogenic therapy. Clin Cancer Res 15: 5020–5025. Mendel DB, Laird AD, Xin X, Louie SG, Christensen JG, Li G et al. Eguchi J, Nomata K, Kanda S, Igawa T, Taide M, Koga S et al. (2003). In vivo antitumor activity of SU11248, a novel tyrosine kinase (1992). expression and immunohistochemical localization of inhibitor targeting vascular endothelial growth factor and platelet- basic fibroblast growth factor in . Biochem derived growth factor receptors: determination of a pharmacokinetic/ Biophys Res Commun 183: 937–944. pharmacodynamic relationship. Clin Cancer Res 9: 327–337. Eliceiri BP, Klemke R, Stromblad S, Cheresh DA. (1998). Meyer RD, Latz C, Rahimi N. (2003). Recruitment and activation of alphavbeta3 requirement for sustained mitogen-activated protein phospholipase Cgamma1 by vascular endothelial growth factor kinase activity during angiogenesis. J Cell Biol 140: 1255–1263. receptor-2 are required for tubulogenesis and differentiation of Ellis LM, Hicklin DJ. (2008). VEGF-: mechanisms of endothelial cells. J Biol Chem 278: 16347–16355. anti-tumour activity. Nat Rev Cancer 8: 579–591. Mohammadi M, Froum S, Hamby JM, Schroeder MC, Panek RL, Lu Faivre S, Delbaldo C, Vera K, Robert C, Lozahic S, Lassau N et al. GH et al. (1998). Crystal structure of an (2006). Safety, pharmacokinetic, and antitumor activity of bound to the FGF receptor tyrosine kinase domain. Embo J 17: SU11248, a novel oral multitarget tyrosine kinase inhibitor, in 5896–5904. patients with cancer. J Clin Oncol 24: 25–35. Motzer RJ, Hutson TE, Tomczak P, Michaelson MD, Bukowski RM, Faivre S, Demetri G, Sargent W, Raymond E. (2007). Molecular basis Rixe O et al. (2007). Sunitinib versus alfa in metastatic for sunitinib efficacy and future clinical development. Nat Rev Drug renal-cell carcinoma. N Engl J Med 356: 115–124. Discov 6: 734–745. Motzer RJ, Hutson TE, Tomczak P, Michaelson MD, Bukowski RM, Ferrara N. (2002). VEGF and the quest for tumour angiogenesis Oudard S et al. (2009). Overall survival and updated results for factors. Nat Rev Cancer 2: 795–803. sunitinib compared with in patients with metastatic Ferrara N, Gerber HP, LeCouter J. (2003). The biology of VEGF and renal cell carcinoma. J Clin Oncol 27: 3584–3590. its receptors. Nat Med 9: 669–676. Nakatsu MN, Hughes CC. (2008). An optimized three-dimensional Ferrara N, Hillan KJ, Gerber HP, Novotny W. (2004). Discovery and in vitro model for the analysis of angiogenesis. Methods Enzymol development of bevacizumab, an anti-VEGF antibody for treating 443: 65–82. cancer. Nat Rev Drug Discov 3: 391–400. Nanus DM, Schmitz-Drager BJ, Motzer RJ, Lee AC, Vlamis V, Ferrara N, Kerbel RS. (2005). Angiogenesis as a therapeutic target. Cordon-Cardo C et al. (1993). Expression of basic fibroblast growth Nature 438: 967–974. factor in primary human renal tumors: correlation with poor Fischer C, Jonckx B, Mazzone M, Zacchigna S, Loges S, Pattarini L survival. J Natl Cancer Inst 85: 1597–1599. et al. (2007). Anti-PlGF inhibits growth of VEGF(R)-inhibitor- O’Farrell AM, Abrams TJ, Yuen HA, Ngai TJ, Louie SG, Yee KW resistant tumors without affecting healthy vessels. Cell 131: et al. (2003). SU11248 is a novel FLT3 tyrosine kinase inhibitor with 463–475. potent activity in vitro and in vivo. Blood 101: 3597–3605. Fukata S, Inoue K, Kamada M, Kawada C, Furihata M, Ohtsuki Y Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L. (2006). VEGF et al. (2005). Levels of angiogenesis and expression of angiogenesis- receptor signalling - in control of vascular function. Nat Rev Mol related are prognostic for organ-specific metastasis of renal Cell Biol 7: 359–371. cell carcinoma. Cancer 103: 931–942. Pardo OE, Latigo J, Jeffery RE, Nye E, Poulsom R, Spencer-Dene B Giavazzi R, Sennino B, Coltrini D, Garofalo A, Dossi R, Ronca R et al. (2009). The fibroblast growth factor receptor inhibitor et al. (2003). Distinct role of fibroblast growth factor-2 and vascular PD173074 blocks small cell lung cancer growth in vitro and endothelial growth factor on tumor growth and angiogenesis. Am J in vivo. Cancer Res 69: 8645–8651. Pathol 162: 1913–1926. Pepper MS, Ferrara N, Orci L, Montesano R. (1992). Potent Gualandris A, Rusnati M, Belleri M, Nelli EE, Bastaki M, Molinari- synergism between vascular endothelial growth factor and basic Tosatti MP et al. (1996). Basic fibroblast growth factor over- fibroblast growth factor in the induction of angiogenesis in vitro. expression in endothelial cells: an autocrine mechanism for Biochem Biophys Res Commun 189: 824–831. angiogenesis and angioproliferative diseases. Differ 7: Pintucci G, Moscatelli D, Saponara F, Biernacki PR, Baumann FG, 147–160. Bizekis C et al. (2002). Lack of ERK activation and cell migration in Hanahan D, Folkman J. (1996). Patterns and emerging mechanisms FGF-2-deficient endothelial cells. FASEB J 16: 598–600. of the angiogenic switch during tumorigenesis. Cell 86: Presta M, Tiberio L, Rusnati M, Dell0Era P, Ragnotti G. (1991). Basic 353–364. fibroblast growth factor requires a long-lasting activation of protein Huang D, Ding Y, Li Y, Luo WM, Zhang ZF, Snider J et al. (2010a). kinase C to induce cell proliferation in transformed fetal bovine Sunitinib acts primarily on tumor endothelium rather than tumor aortic endothelial cells. Cell Regul 2: 719–726. cells to inhibit the growth of renal cell carcinoma. Cancer Res 70: Presta M, Dell0Era P, Mitola S, Moroni E, Ronca R, Rusnati M. 1053–1062. (2005). /fibroblast growth factor receptor Huang D, Ding Y, Zhou M, Rini BI, Petillo D, Qian CN et al. (2010b). system in angiogenesis. Cytokine Growth Factor Rev 16: 159–178. Interleukin-8 mediates resistance to antiangiogenic agent sunitinib Ravaud A, Hawkins R, Gardner JP, von der Maase H, Zantl N, in renal cell carcinoma. Cancer Res 70: 1063–1071. Harper P et al. (2008). Lapatinib versus hormone therapy in patients Joy A, Moffett J, Neary K, Mordechai E, Stachowiak EK, Coons S with advanced renal cell carcinoma: a randomized phase III clinical et al. (1997). Nuclear accumulation of FGF-2 is associated with trial. J Clin Oncol 26: 2285–2291. proliferation of human astrocytes and glioma cells. Oncogene 14: Relf M, LeJeune S, Scott PA, Fox S, Smith K, Leek R et al. (1997). 171–183. Expression of the angiogenic factors vascular endothelial cell

Oncogene FGF2 regulates sunitinib sensitivity JC Welti et al 1193 growth factor, acidic and basic fibroblast growth factor, tumor fibroblast growth factor, a mitogen that promotes their own growth. growth factor beta-1, platelet-derived endothelial cell growth factor, Nature 325: 257–259. growth factor, and pleiotrophin in human primary breast Slaton JW, Inoue K, Perrotte P, El-Naggar AK, Swanson DA, Fidler cancer and its relation to angiogenesis. Cancer Res 57: 963–969. IJ et al. (2001). Expression levels of genes that regulate metastasis Rini BI, Atkins MB. (2009). Resistance to targeted therapy in renal- and angiogenesis correlate with advanced pathological stage of renal cell carcinoma. Lancet Oncol 10: 992–1000. cell carcinoma. Am J Pathol 158: 735–743. Schweigerer L, Neufeld G, Friedman J, Abraham JA, Fiddes JC, Turner N, Grose R. (2010). Fibroblast growth factor signalling: from Gospodarowicz D. (1987). Capillary endothelial cells express basic development to cancer. Nat Rev Cancer 10: 116–129.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene