Pyrazolo–pyrimidine-derived c-Src inhibitor reduces angiogenesis and survival of squamous carcinoma cells by suppressing vascular endothelial growth factor production and signaling

Abstract

Src tyrosine kinase family cooperates with activated growth factor receptors to regulate growth, invasion and metastasis. The authors examined the influence of a novel c-Src inhibitor, 1l, derived from 4-amino-substituted-pyrazolo–pyrimidines, on tumor angiogenesis and on the angiogenic output of squamous carcinoma cells, A431 and SCC-4. The effect of 1l was assessed on growth and microvessel density in A431 tumors and its effect compared with the established c-Src inhibitor PP-1. The effects of c-Src inhibition were investigated on vascular endothelial growth factor (VEGF) expression and activity in tumor cells grown in vivo and in vitro, as well as on VEGF mediated signaling and on endothelial cell functions. Nanomolar concentrations of 1l decreased tumor volume promoted by A431 implanted in nude mice, without affecting in vitro cell tumor survival. This effect was related to 1l inhibition of VEGF production, and secondary to an effect on tumor microvessel density. The rabbit cornea assay confirmed that 1l markedly decreased neovessel growth induced by VEGF. In cultured endothelial cells, 1l inhibited the VEGF-induced phosphorylation on tyr416 of c-Src, resulting in a reduced cell proliferation and invasion. Consistently, 1l dowregulated endothelial nitric oxide synthase, MAPK-extracellular receptor kinase 1–2 (ERK1-2) activity and matrix metalloproteinases (MMP-2/MMP-9), while the tissue inhibitors of metalloproteinases (TIMP2/TIMP-1) were upregulated. These results demonstrate that nM concentrations of c-Src kinase inhibitors (1l and PP-1), by reducing the production of VEGF released by tumor cell and its endothelial cell responses, have a highly selective antiangiogenesis effect, which might be useful in combination therapies. © 2006 Wiley-Liss, Inc.

Angiogenesis, the formation of new capillaries, is associated with the progression of tumor growth and metastasis.1 The formation of new capillaries is a process that requires secretion of proteases, endothelial cell invasion, migration, proliferation and differentiation, which are cellular processes in part regulated by Src family of tyrosine kinases (SFKs).2 The SFKs consist of 8 members. Src, Yes and Fyn are ubiquitously expressed, while Lck, Hck, Fgr, Lyn and Blk have more tissue-restricted expression.3 SFKs are activated in response to stimulation of a variety of cell surface receptors such as tyrosine kinase receptors, integrin receptors, G-protein coupled receptors and by cellular stress.2 Recent studies have shown that c-Src, a nonreceptor tyrosine kinase, exhibits elevated protein levels and activity in numerous types of human cancers.2, 3, 4 Specifically, Src activity was found to be elevated in breast, pancreatic, ovarian, oesophageal, lung, gastric, colon and head and neck cancers.4, 5, 6, 7, 8 The frequency with which elevated expression and/or activity of Src occurs in epithelial cancers strongly suggests its implication in facilitating malignant progression. Src activity has been found to be a critical component of multiple signaling pathways that regulate proliferation, survival, metastasis and angiogenesis.2, 9 However, the mechanism by which Src activity contributes to cancer progression is still not well understood. Recently, numerous papers focus on Src effects on tumor metastasis and angiogenesis.9, 10, 11, 12 Because of its important role in oncogenic processes, it represents a therapeutic target ripe for exploitation.

Vascular endothelial growth factor (VEGF) has been identified as one of the most important factors mediating angiogenesis in physiological and pathological conditions. The key mechanism by which VEGF promotes angiogenesis and permeability is c-Src activation.11 Indeed, the tyrosine kinase c-Src coordinates both the VEGF-induced Ras-extracellular signal-related kinase (ERK) cascade, which mediates endothelial cell proliferation, survival, invasion and gene expression,13 and the VEGF-induced focal adhesion kinase-αvβ5 integrin cascade, which regulates vascular permeability.14 Activation of Src following VEGF interaction with its receptor VEGFR-2 has also been reported to mediate endothelial nitric oxide (eNOS) activity,15, 16 which is known to control both VEGF-mediated vascular permeability and the angiogenesis process.10, 17 The finding that expression of a dominant negative c-Src mutant inhibits VEGF induced angiogenesis in vivo suggests that c-Src might be a suitable pharmacological target for inhibition of angiogenesis.18, 19

Computer modeling of the pyrazolo–pyrimidine class, namely, 4-amino-substituted-1-(2-chloro-2-phenylethyl)-1H-pyrazolo[3,4-d]pyrimidines, predicted that the 1l compound would dock in the ATP pocket of the c-Src tyrosine kinase.20 Biochemical studies in tumor cells revealed that 1l potently inhibited the activity of c-Src and subsequently downregulated phosphorylation of many downstream signaling proteins at the cellular level such as mitogen activated protein kinases, MAPKs.20 Here we report that μM concentration of 1l or PP-1, a known c-Src inhibitor, significantly reduced squamous carcinoma cell A431 and SCC-4 survival and invasion. At low concentration, in the nanomolar range, c-Src inhibition markedly reduced VEGF-induced endothelial cell growth and invasion in vitro, and neovessel growth in vivo. This ultimately leads to reduction in tumor growth in immunodeficient mice. The molecular mechanisms involved the ability of c-Src to regulate both VEGF production in tumors and VEGF-mediated functions in endothelial cells.

Materials and methods

Reagents and chemicals

Bovine calf serum (BCS) was from Hyclone (Celbio, Milan, Italy), Dulbecco's modified Eagle's medium (DMEM), trypsin, gelatine and analytical grade chemicals were purchased from Sigma (Sigma-Aldrich, St. Louis, MO). VEGF-A and EGF were from Peprotech (Rocky Hill, NJ). PP-1 and L-NMMA were provided by Calbiochem (Inalco, Milan, Italy). The compound 1l (N-benzyl-1-(2-chloro-2-phenylethyl)-1H-pyrazolo[3,4-d]pyrimidin-4-amine, MW 363,85) has been synthesized as reported.18 The monoclonal anti-phosphor-c-Src (Tyr 416) antibody (Ab-1) and the polyclonal anti-v-Src were from Calbiochem, the monoclonal anti TIMP-1 and anti TIMP-2 were from Cell-Signaling Technologies (Beverly, MA), the monoclonal anti MMP-2 was from Calbiochem (La Lolla, CA) and the goat anti MMP-9 and the monoclonal anti-VEGF used for A431 were from Santa-Cruz (VEGF (C-1), Santa Cruz, CA), the monoclonal anti-VEGF-A (#7G7) used for SCC-4 was from RELIATech GmbH, the monoclonal actin antibody (I-19) was from Sigma, the monoclonal antibodies recognizing phosphorylated forms of p42/p44 (ERK1-2) MAPKs (Thr202/Tyr204; 9102) and total ERK1-2 were from Cell Signaling Technologies. PVDF membranes and enhanced chemiluminescence reagent was from Amersham Pharmacia Biotech (Milan, Italy). The anti-extradomain B of fibronectin (ED-B) antibody was kindly provided by Philogen S.p.A. Monteriggioni-Siena (Italy).

Cell culture

Postcapillary venular endothelial cells (CVEC) were obtained by using a bead-perfusion technique through the coronary sinus of bovine heart and cultured in DMEM supplemented with 10% BCS and antibiotics (100 U/ml penicillin, 100 μg/ml streptomycin) on gelatine coated dishes as previously described.17 The human epidermoid carcinoma A431 cells and the squamous carcinoma cell line SCC-4 were obtained from American Type Culture Collection. Squamous carcinoma cells, A431 were maintained in culture in DMEM supplemented with 4500 mg/l glucose and 10% FCS, SCC-4 were maintained in culture in Ham/F12 (1:1) supplemented with 0.5 μg/ml hydrocortisone and 10% FCS. Cells were split 1:5 twice a week.

MTT test

To evaluate cell viability, Vybrant MTT kit (Molecular Probe, Milan, Italy) was used. Cells (3 × 10³) were seeded in 96 multiwell plates with 10% serum for 18 hr, starved in 0.1% serum for 24 hr and then exposed to different concentration of the compound 1l or PP-1 for 24 hr. The last 4 hr, medium was removed and cells were incubated with fresh medium in the presence of 1.2 mM MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). Living cells reduce MTT to a strongly pigmented formazan product. After solubilization in DMSO (10 min, 37°C), the absorbance of formazan was measured with a microplate absorbance reader (Spectrafluor, Tecan Männedorf, Switzerland) at 540 nm. Data are reported as % of control.

Cell proliferation

1.5 × 10³ cells resuspended in 10% serum were seeded in each well of 96 multiplates. After adherence (5–6 hr), the supernatant was replaced with medium containing 0.1% serum to synchronize the cell cycle. After 24 hr cells were incubated with VEGF (20 ng/ml) in the presence or in the absence of different concentrations of 1l or PP-1 compound. After 48 hr, cells were fixed in 100% methanol and stained with Diff-Quik (Mertz-Dade, Behring, Milan, Italy). Data are reported as % of control, counted at 200× magnification.

Cell migration and invasion

Chemotaxis or chemoinvasion experiments were performed with the Boyden chamber technique (48-well microchemotaxis chamber) using polycarbonate filters (8 μm pore size, polyvinylpyrrolidone-free, Nucleopore).21 To assess the antimigratory and anti-invasive effects of 1l or PP-1, 1.2 × 10⁴ cells were pretreated for 1 hr with different concentration of the compounds and then placed in the upper compartment of the chamber. In the migration experiments the filters were coated with 100 μg/ml collagen and 10 μg/ml fibronectin, while in the invasion experiments the filters were coated with 1 mg/ml gelatine. VEGF (20 ng/ml) was used as the chemoactractive molecule in the lower compartment of the chamber. After incubation at 37°C for 6 hr for CVEC and 16 hr for tumor cell lines, the upper surface of the filter was scraped so as to remove nonmigrated cells. Filters were fixed and stained with Diff-Quick (Mertz-Dade). Data are reported as % of control (control: cells which had moved across the filter in 5 fields/well, counted at 400× magnification).

eNOS activity

To evaluate the eNOS activity 4 × 10⁵ CVEC were seeded in 60 mm culture dishes in 10% BCS. Cells were allowed to grow overnight and then were treated as described below. Cells were treated for 1 hr with 1l (50 nM) or PP-1 (500 nM) or L-NMMA (2 mM). Then 20 ng/ml VEGF was added for 1 hr. All determinations were performed in duplicate. eNOS activity was measured as described previously.22 Radioactivity values obtained from plates treated with L-NMMA were subtracted from each experimental point. eNOS activity is expressed as dpm/mg of protein.

Western blot

3 × 10⁵ cells were plated in 60 mm diameter dishes in 10% serum for 6 hr. After adhesion cells were serum starved overnight and then exposed to test substances following different protocols. To evaluate the effect of Src inhibitors on VEGF-induced Src and ERK1-2 phosphorylation, cells were pretreated with inhibitors for 1 hr and then stimulated for 10 min with VEGF (20 ng/ml, for endothelial cells) or EGF (20 ng/ml, for tumor cells). To measure the effect of Src inhibitors on TIMP-1 and -2 and MMP-2 and -9 expression, cells pretreated for 1 hr with the inhibitors were stimulated with VEGF (20 ng/ml) for 8 hr. To measure the effect of Src inhibitor on VEGF expression cells were treated with inhibitors for 8 hr. After incubation with test substances cells were lysed. Fifty microgram of proteins were mixed with 4× reducing SDS-PAGE sample buffer and denaturized at 100°C for 10 min. Electrophoresis was carried out in SDS/10% PAGE. Proteins were then blotted onto activated PVDF membranes and then the immunocomplex detected by enhanced chemiluminescence system. Results were normalized to those obtained by using an anti-total ERK1-2 and total Src (for phospho-ERK1-2 and phospho c-Src, respectively), or anti-actin (for phospho c-Src, VEGF, MMP-2 and -9 and TIMP-1 and -2).

To evaluate the VEGF expression in tumor xenograft, specimens were snap frozen in liquid nitrogen, and then homogenized with a dounce homogenizer using 1 ml of lysis buffer for sample (20 mM TrisCl, 40 mM Na pyrophosphate, 50 mM NaF, 5 mM MgCl₂, 100 mM Na vanadate, 10 mM EGTA, 1% Triton X-100, 0.5% NaDeoxycholate, 5 μg/ml leupeptin and 5 μg/ml aprotin). The samples were sonicated twice at 25% for 10 min and centrifuged at 10,000 g for 20 min. Eighty micrograms of protein were processed as previously reported for Western Blotting. Results were normalized with anti actin.

Angiogenesis in vivo: Rabbit cornea assay

Angiogenesis was studied in the cornea of albino rabbits and quantified by stereomicroscopic examination.17, 23 Experiments have been performed in accordance with the guidelines of the European Economic Community for animal care and welfare (EEC Law No. 86/609). Test substances were incorporated in slow-release pellets prepared from a casting solution of ethynil–vinyl copolymer (Elvax-40, DuPont-De Nemours, Milan, Italy). Pellets were then implanted into micropockets surgically produced in the lower half of the eye of anaesthetized New Zealand white rabbits (Charles River, Lecco, Italy). VEGF (200 ng) and 1l (5 μg) were delivered alone or in combination from one single pellet. Observation and quantification of the angiogenic response were performed by a slit-lamp stereomicroscope. The potency of angiogenic activity was evaluated on the basis of the number and growth rate of newly formed capillaries, and an angiogenic score was calculated by the formula [vessel density × distance from limbus] as described in Ref.20.

Tumor growth and vascularization in immunodeficient mice

Experiments have been performed in accordance with the guidelines of the European Economic Community for animal care and welfare (EEC Law No. 86/609) and National Ethical Committee. To assess the in vivo antiangiogenic/antitumor activity of 1l, female immunodeficient mice (5–8-week-old CD-1 nude mice, Charles River) were s.c. inoculated in the right flank with 10⁷ A431 cells in a volume of 50 μl.21 After 9 days, when tumors reached a volume of 200 mm³, animals were randomly assigned to 2 different experimental protocols (5 mice per group). At this time, peritumor treatment, i.e., the injection of the compound or vehicle close to the tumor mass, with 1l (1 μg/day/mice) or vehicle (5% DMSO in PBS) started. Daily treatment was performed for 10 consecutive days. Serial caliper measurements of perpendicular diameters were used to calculate tumor volume using the following formula: (shortest diameter × longest diameter × thickness of the tumor in mm). Data are reported as tumor volume in mm³. Animals were observed daily for signs of cytotoxicity and were sacrificed by CO₂ asphyxiation.

At day 10 animals were sacrificed and each tumor was split in 2 parts, one part was immediately frozen in liquid nitrogen for Western Blotting. The other part was embedded in Tissue-Tek O.C.T. (Sakura, San Marcos, CA), cooled in isopentane and frozen in liquid nitrogen for histology. Seven-micrometer-thick cryostat sections were stained with hematoxylin and eosin and adjacent sections were used for immunohistochemical staining with the anti-ED-B monoclonal antibody after fixation in absolute cold acetone. For each tumor ED-B positive stained vessel-like structures were randomly counted in 3 different sections. In each section, 7 counts were performed. Data (means ± SEM) are reported as total vessel-like structures counted/section.

Statistical analysis

Results are expressed as means ± SEM. Statistical analysis was performed using Student's t-test and analysis of variance. When a significant difference was detected, multiple comparison analysis was performed using the Student-Newman-Keuls test. A value of p < 0.05 was considered to denote statistical significance.

Results

c-Src inhibition influences A431 tumor vascularization

In a preliminary study, the 1l (object of this study), a new compound selected from a class of 4-amino-substituted-pyrimidines, was found to inhibit c-Src kinase activity in squamous carcinoma cell (A431) at micromolar concentrations.20 Here we demonstrate that 1l, at 200-fold lower concentration (50 nM), still exerts a significant inhibition (p < 0.01 vs. EGF) on c-Src phosphorylation, in 2 squamous carcinoma cell lines, A431 and SCC-4 (Fig. 1a). The extent of inhibition was similar to that exhibited by the known c-Src kinase inhibitor, PP-1 at 500 nM. However, low concentration (up to 500 nM) of c-Src inhibitors, either 1l or PP-1, had no effect on survival and invasion of both tumor cell lines (Figs. 1b and 2c), while at high concentrations (above 1000 nM) they drastically reduced survival and invasion of A431 and SCC-4 (A431 survival: EC₅₀ 4.69 μM for 1l and 6.5 μM for PP-1; SCC-4 survival: EC₅₀ 1.1 μM for 1l and 1.5 μM for PP-1; A431 invasion: EC₅₀ 2.54 μM for 1l and 4.77 μM for PP-1; SCC-4 invasion: EC₅₀ 0.94 μM for 1l and 0.85 μM for PP-1). However, the measurement of cell migration might reflect the decrease of cell survival induced by 1l.

Figure 1. Effect of c-Src inhibitors 1l and PP-1 on in vitro A431 and SCC-4 tumor cell survival, invasion and VEGF expression. a, c-Src phosphorylation in A431 and SCC-4 treated with EGF (20 ng/ml) in the presence of 1l (50 nM) or PP-1 (500 nM). Data are expressed as p-Src/Actin OD ratio. Figure insets show representative gel of western blot for anti-p-Src. Results were normalized with anti-Actin. A representative gel out of 3 are shown. ***p < 0.001 versus Control; ^##p < 0.01 versus EGF 20 ng/ml. b, and c, A431 and SCC-4 cells were treated with 1l or PP-1. Cell survival at 48 hr was assayed by MTT test (b). Cell invasion at 16 hr was assayed by Boyden chamber (c). The data represent inhibition (%) from two independent experiments in triplicate. Filled squares: 1l-treated A431; filled dots: PP-1 treated A431; empty squares: 1l-treated SCC-4; empty dots: PP-1 treated SCC-4. (A431 control: 1.71 ± 0.2 absorbance at 540 nm, for survival, and 49.5 ± 3.1 cell number counted/well, for invasion; SCC-4 control: 1.44 ± 0.03 absorbance at 540 nm, for survival, and 72 ± 7 cell number counted/well, for invasion). *p < 0.05, **p < 0.01, ***p < 0.001 versus the control condition (0.1% FCS). d, densitometric analysis of VEGF in A431 and SCC-4 treated for 8 hr with 1l (50 nM), PP-1 (500 nM). Data are the means of 3 blottings and are reported as VEGF-A/Actin OD ratio. Figure insets show a representative blot anti-VEGF-A. Results were normalized with anti-Actin. A representative gel out of 3 are shown. **p < 0.01, ***p < 0.001 versus Control.

Figure 2. Effect of 1l on in vivo tumor growth and vascularization. a, the antitumor activity of 1l was evaluated in nude mice inoculated with A431 cells and treated after the onset of tumor growth (day 9 from inoculation, 200 mm³ tumor volume). Peritumor treatment with 1l (1 μg/mice/day) or vehicle continued for 9 days. Data are reported as tumor volume in mm³ from treatment start (means ± SEM of 5 animals/group). *p < 0.05 vs. vehicle. b, the effect of 1 μg/day 1l (upper panels) on tumor angiogenesis at day 9 was compared to vehicle treated group (lower panels). Representative pictures of tumor sections stained with hematoxylin and eosin (left panels) and with the antibody specific for B-FN (right panels). A positive signal (brown) was visible in microvessels and in the matrix undergoing remodeling due to tumor cell activation. Original magnification 20×. c, representative gel of VEGF expression in tumor. Results were normalized with total actin. Bar graph represents the mean of 5 tumor samples expressed as VEGF-A/Actin ± SEM. ***p < 0.001 vs. control by analysis of variance.

Since SFKs are important signals for angiogenesis, we investigated whether their inhibitors would influence the production of VEGF, a growth factor that primes tumor cells growth.24 The results reported in Figure 1d show that A431 and SCC-4 were highly sensitive to c-Src inhibitors, as both 1l and PP-1 reduced endogenous VEGF production by more than 50%, at 50 and 500 nM, respectively (Fig. 1d). We also examined the efficacy of these inhibitors on in vivo tumors by implanting A431 cells in nude mice. In mice bearing A431 tumors, treatment with 1l (1 μg/day/mice) significantly reduced the growth of tumors compared with the control group. Tumors in 1l treated mice were significantly smaller than in control mice (p < 0.05 vs. vehicle group at day 6 and 7, approximately 50% lower than of control group) (Fig. 2a). However, we noted that in treated animals, stating from day 6, the rate of growth of tumors increased. One possible explanation is that A431 cells might acquire resistance to the compound, a phenomenon already reported for these cells.24 Given the effect of c-Src inhibitors on VEGF production we investigated the effect of 1l-mediated c-Src inhibition on tumor microvessel density. B-fibronectin (B-FN), the isoform containing extradomain B (ED-B), accumulates around neovascular structures in aggressive tumors and other tissues undergoing angiogenesis and remodeling.25 The quantification of newly formed vessels, performed after ED-B fibronectin immunostaining (Fig. 2b), highlights the antiangiogenic effect of 1l, since 24 ± 4 vessels were counted in the tumor mass of 1l group, while 64 ± 2.4 vessels were found in vehicle treated group (p < 0.0001, n = 5, and Fig. 2b). Importantly, at the molecular levels, tumors in 1l treated mice expressed lower levels of VEGF than in control mice (Fig. 2c). Quantification of blots revealed that VEGF in tumors of 1l treated mice was expressed 9-fold less than in control tumors. Together, the results demonstrate that c-Src inhibition reduced in vivo A431 tumor growth. This effect appeared to be related to its inhibition of VEGF production and secondary to an effect on tumor microvessel density, establishing a strict linkage between Src oncogene function and tumor angiogenesis.

c-Src inhibition reduces in vivo and in vitro VEGF-induced angiogenesis

The direct antiangiogenic activity of 1l was also evaluated on the neovascular growth induced by VEGF in the rabbit cornea assay. 1l was tested at 5 μg/pellet per se, to evaluate its potential inflammatory response, and in the presence of VEGF (200 ng/pellet), being the 2 compounds coreleased form the same pellet. 1l produced a significant inhibition (60% at 14th day) of VEGF induced vascularization, without eliciting inflammation (Figs. 3a–3c).

Figure 3. c-Src inhibitors reduce VEGF-mediated in vivo angiogenesis and in vitro endothelial cell growth and invasion. a, Angiogenesis was evaluated in the avascular rabbit cornea assay. The antiangiogenic activity of 1l (5 μg/pellet) was evaluated on the neovascular growth induced by VEGF (200 ng/pellet). Angiogenesis was followed by stereomicroscopic examination. Data are expressed as angiogenic score (means ± SEM) during time (days). Numbers are means from at least 4 implants for each experimental point. Representative pictures of corneal angiogenesis induced by VEGF alone (b), and in the presence of 1l (c). Photographs were taken at day 10 after surgical implantation, through a slit lamp stereomicroscope (×18). (d) Serum starved CVEC were exposed to 20 ng/ml VEGF in the presence of 1l at different concentrations or 500 nM PP-1 for 48 hr. Data are expressed as % of control (***p < 0.001 vs. Control: 146.75 ± 19 total counted cells/ well; #p < 0.05, ##p < 0.01, ###p < 0.001 vs. VEGF alone: 229 ± 20.4 total counted cells/well; means ± SEM of 3 experiments run in triplicate). (e) CVEC treated with 1l or PP-1 for 1 hr were seeded in the upper compartments of Boyden chambers, on porous filters coated with gelatine. The lower compartments contained basal media (0.1% BCS) or 20 ng/ml VEGF as chemoattractant. Data are reported % of control (**p < 0.01, ***p < 0.001 vs. Control: 24 ± 3.2 total counted cells/well; ##p < 0.01, ###p < 0.001 vs. VEGF alone: 45 ± 4 total counted cells/well, means ± SEM of 3 experiments run in triplicate).

We then investigated the effect of 1l on endothelial cell functions related to the angiogenesis process such as cell proliferation and invasion. First, we established the potential cytotoxic effect of 1l on cultured microvascular endothelial cells, CVEC, in comparison with the well-known c-Src inhibitor PP-1. In endothelial cells exposed to increasing concentrations of compounds (1 nM-1 μM), we observed a reduced cell viability in the range 500 nM-1 μM for compound 1l, and at concentration above 500 nM for PP-1 (Table I). The concentrations of compounds employed in the following experiments were devoid of nonspecific toxic effects.

Table I. Effect of 1l on Endothelial Cell Viability

Concentration, nM	1l	PP-1
Microvascular endothelial cells (CVEC) were treated with 1l or the selective c-Src inhibitor, PP-1. Cell viability at 24 hr was assayed by MTT test. The data are the mean ± SEM of 3 independent experiments. * p< 0.05, ** p< 0.01, *** p< 0.001 versus the control condition (0.1% BCS).
0	0.55 ± 0.04	0.57 ± 0.04
5	0.53 ± 0.02	0.59 ± 0.07
50	0.53 ± 0.05	0.51 ± 0.04
100	0.49 ± 0.08	0.51 ± 0.04
500	0.39 ± 0.004**	0.48 ± 0.05
1000	0.25 ± 0.07***	0.41 ± 0.04*

In quiescent CVEC (0.1% BCS), 1l (1 to 50 nM range) did not modify cell growth or invasion (Figs. 3d and 3e). When cells were stimulated to proliferate by VEGF (20 ng/ml), addition of 1l produced a marked inhibition of proliferation (Fig. 3d). The inhibitory effect elicited by 1l was concentration-dependent and significant in the range 10–50 nM (Fig. 3d). In cell invasion experiments, 1l at concentrations ranging from 10 to 50 nM, greatly affected the endothelial cell ability to invade in response to VEGF (Fig. 3e). Notably, PP-1 (500 nM) displayed inhibitory effects comparable to 50 nM 1l on VEGF-induced endothelial cell growth and invasion (Figs. 3d and 3e).

Inhibition of c-Src activity affects VEGF-promoted signals in microvascular endothelial cells

We then examined whether 1l, at non toxic doses, would influence endothelial c-Src kinase phosphorylation induced by VEGF. The c-Src signal, detected by immunoblot in the quiescent endothelium (lane 1, Fig. 4a), was enhanced by VEGF (20 ng/ml) application (lane 2, Fig. 4a), which doubled the specific phosphorylation on Tyrosine 416. The increase of c-Src kinase phosphorylation was inhibited by addition of either PP-1 (500 nM) or by 1l (maximal suppression in the10–50 nM range).

Figure 4. Effect of 1l and PP-1 on c-Src, ERK1-2 and eNOS activity. a, c-Src activity in CVEC treated with VEGF (20 ng/ml) with or without 1l. Data are expressed as p-c-Src/c-Src OD ratio. PP-1 (500 nM) was used as positive control. Figure inset shows representative gel of western blot for anti-pYSrc416. Results were normalized with anti-vSrc. A representative gel out of 3 is shown. ***p < 0.001 vs. Control; ^##p < 0.01 vs. VEGF 20 ng/ml. b, eNOS activity was measured in endothelial cell monolayers as conversion of [3H]L-arginine in [3H]L-citrulline. L-NMMA values were subtracted from each experimental point. Data are reported as dpm/mg protein. *p < 0.01 vs. Control response, ^#p < 0.05 and ^##p < 0.01 vs. VEGF alone. c, densitometric analysis of ERK1-2 in CVEC pre-treated for 1 hr with 1l (1, 10 and 50 nM), PP-1 (500 nM) or the selective MEKK inhibitor U0126 (10 μM) and then stimulated for 10 min with VEGF (20 ng/ml). Data are the means of 3 blottings and are reported as p-ERK1-2/ERK1-2 OD ratio. Figure inset shows a representative blot anti-pERK1-2. Results were normalized with anti-ERK1-2. A representative gel out of 3 is shown. ***p < 0.001 vs. Control; ^#p < 0.1, ^##p < 0.01, ^###p < 0.001 vs. VEGF 20 alone.

Next, we studied the influence of c-Src inhibition on eNOS and ERK, critical intracellular signals for the acquisition of the angiogenic phenotype induced by VEGF in endothelial cells, i.e., proliferation, invasiveness and ultimately neovessel formation. eNOS activity of CVEC, primed by VEGF (20 ng/ml), was severely reduced by both PP-1 (500 nM) and 1l (50 nM) (Fig. 4b). ERK1-2, which lies downstream of both eNOS and c-Src, was rapidly phosphorylated following VEGF application, as previously reported in CVEC.20 Both compounds reduced the VEGF-enhanced phosphorylation reaching significance in the 10–50 nM range (Fig. 4c). MEKK inhibitor U0126 (10 μM), the selective MAPK-ERK1-2 inhibitor, has been used as internal positive control.

MMPs and TIMPs production are regulated by Src

Because the acquisition of the angiogenic phenotype promoted by VEGF in endothelial cells is determined mainly by the balance between matrix metalloproteinases (MMPs) and their tissue inhibitor proteins (TIMPs), we decided to study the pattern of secretion of these proteins. In particular, we tested the effects of both PP-1 and 1l on the protein expression of MMP-2 and MMP-9, and on their inhibitors, TIMP-1 and TIMP-2. In microvascular endothelium, PP-1 (500 nM), as well as 1l (50 nM), reduced VEGF-induced MMP-2 and MMP-9 expression (Figs. 5b and 5e), while markedly upregulating TIMP-1 and TIMP-2 protein (see the OD ratio in Figs. 5c and 5f). Thus, both c-Src inhibitors regulate MMP/TIMP balance toward TIMP expression, therefore reducing endothelial proteolytic activity. Neither compound changed the pattern of secretion of the above proteins in quiescent cells (data not shown).

Figure 5. 1l and PP-1 inhibit VEGF-induced MMP-2 and MMP-9 expression and promotes TIMP-1 and TIMP-2 production. Western blottings for MMP-2, MMP-9, TIMP-1 and TIMP-2 expression in CVEC pretreated for 1 hr with 1l (50 nM) or PP-1 (500 nM) and then stimulated for 8 hr with VEGF (20 ng/ml). Results were normalized with actin. a and d, representative gels for MMP-2 /TIMP-2 or MMP-9 /TIMP-1 respectively are shown. In b, c, e and f, data are reported as means of three blotting for MMP-2, TIMP-2, MMP-9 and TIMP-1 OD, respectively. *p < 0.05, **p < 0.01, ***p < 0.001 versus Control; ^#p < 0.1, ^##p < 0.01 versus VEGF alone.

Discussion

In this work, by using a novel synthesized c-Src inhibitor, 1l, and the known c-Src inhibitor, PP1, we have investigated the molecular and cellular basis of c-Src kinase inhibitors as antiangiogenic agents in vitro, and in vivo in a preclinical model of human squamous cell carcinoma xenograft. We show that the growth of experimental tumors induced in immunosuppressed mice by implant of A431 is dependent on the new vessels formation promoted by the inherent overexpression of VEGF. 1l was found to inhibit the in vivo angiogenic potential of squamous cell carcinoma A431, by reducing VEGF released from the tumor mass (tumor as well as stroma cells), and to arrest tumor growth following its subcutaneous administration to mice. In fact, at the molecular level, 1l reduced endogenous VEGF expression in tumor cells, the in vivo angiogenesis induced by VEGF in the rabbit cornea and the in vitro VEGF stimulated proliferation, migration and invasion of endothelial cells. The signaling pathways associated with the inhibition of endothelial cellular effects, i.e., eNOS, and ERK1-2 activity, and the control of MMPs/TIMPs balance at transcriptional level were also affected by c-Src inhibition.

There are a number of reasons for investigating c-Src in the context of tumor growth and the dependence from angiogenesis. First, the c-Src protein has been reported to be overexpressed in a variety of human tumors and to play a key role in controlling their proliferation and invasiveness.4 Moreover, c-Src overexpression and activation have been correlated with a large number of growth-regulatory processes.26 Second, in epithelial tumors it cooperates with EGFR in growth signaling.24, 27 Particularly, the A431 cell line overexpress both the wild type EGFR and the aberrant extracellular forms of this receptor, and when implanted in mice, its growth is highly dependent on VEGF for angiogenesis.24 In line with this background, we found that in vivo c-Src inhibition by 1l decreased the tumor associated release of VEGF and reduced microvessel density, hence hampering tumor growth in mice transplanted with A431 cell. In fact, we report that, at the molecular level, 1l significantly reduced endogenous VEGF expression in squamous carcinoma cell lines, A431 and SCC-4 in vitro. Thus, inhibition of c-Src phosphorylation halts tumor cell growth by inhibiting primarily the angiogenesis process, demonstrating a linkage between c-Src oncogene function and tumor angiogenesis in squamous carcinoma cells. That c-Src inhibitors might directly reduce uncontrolled tumor cell growth can not be excluded. However, an interesting finding of this work was the blockade by low concentrations of 1l and PP-1 of VEGF mediated tumor angiogenesis and the in vivo and in vitro angiogenic process.

At variance from our conclusions, Criscuoli et al.10 have observed that ablation of c-Src (knock out mice for c-Src) affects only the development of metastasis rather than the primary tumor. Whether these differences are due to the approach for silencing c-Src (chemical inhibitors vs. gene ablation) or the type of epithelial tumor employed (squamous cell carcinoma vs. pulmonary) remains to be investigated. However that bulk of recent evidence on c-Src inhibitors indicates that these compounds predominantly suppress tumor angiogenesis to affect tumor growth.28, 29

The interaction of the c-Src kinase inhibitor 1l with vasculature is illustrated in the experiments on in vivo angiogenesis and on microvascular endothelial cells in culture under VEGF stimulation. 1l suppressed the formation of new capillaries in the rabbit cornea and significantly attenuated cell proliferation and invasiveness. Notably, the absence of effects of 1l at nM concentrations on quiescent endothelial cell indicates that the compound has no effects other than its c-Src inhibitory activity and specifically targets functions activated by VEGF. Although, VEGF dominates the development of tumor vasculature, given the upstream position of c-Src, we can not exclude that its inhibition may influence other signaling involved in vascular remodeling such as PDGF or FGF-2.30, 31

Signals associated with the VEGF pathway, such as eNOS, and MAPK-ERK1-2, were predictably down regulated by 1l, to an extent comparable with that observed for cell functions, suggesting that in microvascular endothelium c-Src activity mediates cell proliferation by controlling eNOS and ERK1-2 pathways (Fig. 6). VEGF stimulation of MAPK pathway is well established.32, 33 We and others have also reported that VEGF-induced activation of the MAPK cascade involves NOS dependent signaling events.22, 32, 33 The NOS/guanylate cyclase dependent activation of the MAPK cascade leads to VEGF-induced proliferation of endothelial cell.

Figure 6. Schematic representation. Potential inhibitory activity of 1l and PP-1 on c-Src-VEGF production pathway, controlling tumor mediated angiogenesis, and on VEGF-induced-c-Src-eNOS-MAPK signaling pathway, controlling endothelial cell functions related to invasiveness and angiogenesis.

Intriguing are the results showing the influence of 1l on the production of degradative enzymes stimulated by VEGF. VEGF, reported to disrupt endothelial barrier function,12 contributes to tumor cell extravasation and metastasis.11, 12 We found that VEGF disrupts the endothelial barrier, at least in part, by promoting MMP-2 and MMP-9 production. Pharmacological blockade of c-Src suppresses the VEGF-induced MMP production, while promoting the synthesis of TIMP-1 and TIMP-2. In particular, MMP-9 is a matrix degrading enzyme regulated in a c-Src-dependent manner at the transcriptional level through a GT box located downstream of the AP-1 site of its promoter. In this context, our results are fully consistent with this property of c-Src. TIMP-1 is also known to be regulated in a v-Src-dependent manner in transformed cells, and both TIMP-1 and TIMP-2 have been reported to be inhibited in PC3 human prostate cancer cells by μM concentrations of pyrrolopyrimidine c-Src inhibitors.34, 35 Conversely, in our model, c-Src inhibitors promote TIMP-1 and TIMP-2 expression, indicating an important role of c-Src kinase in modulating the matrix degradation ability of endothelial cells.

PP-1, used as a reference compound in this study, has been identified as a Src tyrosine kinase inhibitor and used extensively to investigate signaling pathways involving Src kinases.36 In our hands, the novel compound 1l inhibited c-Src phosphorylation displaying efficacy comparable with that of PP-1. Recently, however, reports have questioned the selectivity of this compound showing that PP-1 also directly inhibits other tyrosine kinases such as Kit, and Bcr-Abl in tumor cell lines and interferes with platelet derived growth factor receptor in smooth muscle cells.37

In conclusion, this study demonstrates that inhibition of c-Src kinase activity reduces VEGF induced-angiogenesis both in tumor and endothelial cells, indicating that c-Src inhibitors might be potential therapeutic agents for angiogenesis associated diseases. In this context, and particularly for epidermoid squamous carcinoma, for which conventional chemotherapy is poorly efficacious, the c-Src inhibitors might provide a novel strategy for combination therapy.

Acknowledgements

The authors thank Philogen S.p.A. Monteriggioni-Siena (Italy) for providing anti-ED-B antibody, and Dr. Morbidelli L. for performing rabbit cornea assay. SD was supported by funds from NuGO (FOOD-CT-2004-506360).