Antitumor and Antiangiogenic Activities of BMS-690514, an Inhibitor of Human EGF and VEGF Receptor Kinase Families Free

The HER family of receptor tyrosine kinases have been of great interest in cancer drug discovery because of their involvement in the biology of diverse cancer types. The epidermal growth factor (EGF) receptor is expressed in many epithelial cells and has been shown to be mutated in a subset of non–small cell lung cancer (1). Erlotinib and gefitinib are small molecule inhibitors of the EGF receptor (EGFR) kinase that have shown efficacy in non–small cell lung cancer patients with EGFR mutations (2). In addition, monoclonal antibodies to the EGFR (cetuximab and panitumumab) have been proven to provide effective treatments for patients with colorectal or esophageal cancer (3). A second HER kinase, HER2, has also been validated as a cancer drug target, and trastuzumab and lapatinib are approved agents for the treatment of breast cancer patients whose tumors have HER2 gene amplification or overexpression (4–6). A third member of the family, HER4, was recently shown to be mutated in 19% of melanomas, although the dependence of these tumors on HER4 remains to be confirmed (7). The remaining HER receptor, HER3, is capable of ligand binding but lacks enzymatic activity. HER3 has been shown to form heterodimers with other HER kinases as well as with non-HER receptor kinases and functions as an adaptor protein to facilitate the activation of downstream signaling (8, 9).

In addition to their relationship by virtue of sequence homology, the HER kinases are likely to be linked in their biological function. Many tumor cells coexpress 2 or more of the receptors, but the importance of receptor heterodimerization in tumor cell biology has yet to be fully appreciated. Yet receptor heterodimerization and receptor switching has already been proposed to account for intrinsic or acquired resistance to selective inhibition of EGFR or HER2 signaling (10, 11). The early development of kinase inhibitors that target the HER family led to the identification of erlotinib and gefitinib, both of which are highly selective for the EGFR and are 10- to 50-fold less potent in inhibiting HER2 (12, 13). Other HER kinase inhibitors, such as lapatinib, showed the feasibility of dual inhibition of EGFR and HER2 while retaining a high degree of overall kinase selectivity (14). However, these second-generation HER kinase inhibitors may still lack the necessary cell potency for both EGFR and HER2 to completely inhibit an EGFR/HER2 heterodimer. For example, the potency of lapatinib in EGFR-dependent cell lines is relatively modest, compared with its potency in inhibiting HER2-dependent cells (15, 16). Other pan-HER kinase inhibitors, that are irreversible binders, have recently been described but their application in the treatment of tumors that may depend on heterodimer signaling has not been explored (17–19). Thus, the utility of dual inhibitors of EGFR and HER2 in expanding the efficacy spectrum of HER inhibition remains to be shown.

The pyrrolotriazine core was previously shown to yield potent and selective inhibitors of a number of receptor tyrosine kinases including those of the HER family and the VEGF receptor 2 (VEGFR2; refs. 20–22). Continued optimization of a chemical series with that core led to the identification of BMS-690514, which, in addition to being a potent inhibitor of EGFR and HER2, also inhibits VEGFR2. VEGF signaling is an important aspect of the tumor angiogenesis process, and monoclonal antibodies and small molecule inhibitors of VEGF or its receptors have been established as a proven approach to inhibit tumor growth (23–25). In this report, we describe the characterization of the biological activities of BMS-690514 and of its potential as a novel anticancer agent that inhibits both HER receptor signaling in tumor cells and VEGFR signaling in the tumor vasculature.

Unless otherwise noted, all tumor cell lines were obtained from American Type Culture Collection and were maintained as recommended. SNU-216 and HUVEC (human umbilical vein endothelial cells) were obtained from the Korean Cell Line Bank and Clonetics, respectively. DiFi, PC9, KPL-4, and GEO were obtained from Drs. Z. Fan (MD Anderson Medical Center, Houston, TX), K. Nishio (National Cancer Center Hospital, Tokyo, Japan), J. Kurebayashi (Kawasaki Medical School, Kurashiki, Japan), and K. Mulder (Penn State College of Medicine, Hershey, PA; refs. 26–29). Sal2 cells were derived from a salivary gland tumor in a transgenic mouse that expressed a CD8HER2 fusion protein (30). Cell lines were maintained for no more than 40 passages or 3 months of continuous use. Antibodies specific for EGFR and HER2 were previously described (30). Anti-phosphotyrosine antibody (PY20) was obtained from BD Biosciences. Antibodies specific for HER3, HER4, mitogen-activated protein kinase (MAPK), phospho-MAPK, AKT, and phospho-AKT were obtained from Cell Signaling Technology. Mutations were introduced into EGFR and HER4 cDNA expression plasmids by using QuikChange Site-Directed Mutagenesis Kit (Stratagene). BMS-690514 (see Supplementary Fig. S1 for chemical structure) and BMS-582664 (brivanib alaninate) were synthesized by Bristol Myers Squibb Oncology Chemistry. Lapatinib and erlotinib were obtained from LC Laboratories. Recombinant enzymes and biochemical assays of protein kinase activity were as described previously (31). KinomeScan profiling of BMS-690514 was done by Ambit Biosciences (32). Cell proliferation was measured by the CellTiter 96 assay kit (Promega). BMS-690514 was added to the cells 24 hours after plating, and the cells were cultured for an additional 72 hours before the measurement was made.

The human EGFR and HER4 (JMa/Cyt1 isoform) cDNA sequences were subcloned into a mammalian expression plasmid with a cytomegalovirus promoter, and point mutations were introduced by the QuikChange Site Directed Mutagenesis Kit (Stratagene). LX1 cells, which do not express EGFR or HER2, were transfected with EGFR L858R mutant plasmid by Fugene 6 (Roche Applied Science). At 48 hours after transfection, cells were treated for 1 hour with BMS-690514, and cell lysates were prepared as described previously (31). Immunoprecipitation was conducted by an anti-EGFR antibody and the immunoprecipitates were analyzed in Western blots by using anti-phosphotyrosine and anti-EGFR antibodies. Antibody binding was detected by Li-COR Odyssey IR imaging system. HER4 expression plasmids were transfected into COS7 cells by Fugene 6, and the transfected cells were treated with inhibitors for 1 hour before cell lysates were prepared. HER4 phosphorylation was assessed by Western blot analyses of the cell lysates by using anti-phosphotyrosine and anti-HER4 antibodies. In studies that involved EGF stimulation, tumor cells that were approximately 80% confluent were cultured for 4 hours in medium that contained reduced (0.2% instead of 10%) FBS. BMS-690514 was added to the cultures and the cells were allowed to incubate for an additional hour. EGF (50 ng/mL) was added to stimulate receptor phosphorylation, and cell lysates were prepared 5 minutes later. Following Western blotting and detection, inhibition of receptor phosphorylation was quantitated by integration of pixel intensity, with normalization to total receptor proteins.

Animal studies were approved by the Bristol-Myers Squibb Animal Care and Use Committee and were in accordance with the American Association for Accreditation of Laboratory Animal Care. Tumors were implanted subcutaneously in female Balb/c athymic mice (Harlan Sprague-Dawley), and were allowed to reach approximately 100 mm³ before the initiation of compound treatment in groups of 8 to 10 animals each. Vehicles used were 40% propylene glycol, 50% water, and 10% Tween 80 for BMS-690514, sodium citrate buffer (pH 3.5) for BMS-582664, and 0.5% (w/v) hydroxypropyl methylcellulose, and 0.1% Tween 80 for erlotinib. Tumor measurement and data analyses were carried out as described (31).

As described previously, matrigel (BD Biosciences) was impregnated with recombinant VEGF and basic fibroblast growth factor (bFGF; PeproTech) to increase the extent of endothelial cell invasion and the sensitivity of the assay (33). Matrigel plugs were implanted subcutaneously in nude mice, and groups of 8 animals each were treated with BMS-690514 or BMS-582664 once daily for 10 days, before the plugs were excised and processed for immunohistochemistry as before (33).

Dynamic contrast-enhanced MRI (DCE-MRI) studies were conducted as described before (33). Athymic mice were subcutaneously implanted with L2987 human lung cancer cells and groups of 8 animals each were placed on study when tumor size reached 100 mm³. L2987 tumors were used because they had previously been shown to be highly vascularized, and their moderate sensitivity to BMS-690514 made it possible to study pharmacodynamic effects without the potential complication that may result from massive tumor cell death. MRI was done on a Bruker PharmaScan 4.7 Tesla with a 16 cm bore (Bruker Biospin). Gadopentetrate dimeglumine (GdDTPA Magnevist; Berlex Labs; 0.3 mmol/kg) was given as an intravenous injection (125 μL for 25 g mouse) over 4 seconds via syringe pump with an infusion rate of 2 mL/min. Images were analyzed as a pixel average of the regions of interest by Image Sequence Analysis (Bruker Biospin). Repetition times of 6,000, 1,200, 900, 600, 300, and 152.2 milliseconds were used to acquire images. The first 5 baseline points prior to contrast were averaged and used to normalize the tumor and muscle uptake values for each data set. Tumor values were normalized to their corresponding muscle values and results are reported as the area under the curve for the first 60 seconds (AUC₆₀). The vascular permeability surface area product (K^trans), a combination of blood flow, vascular density, and permeability, was calculated from the DCE-MRI data.

BMS-690514 was selected for further characterization because it is among the most potent inhibitors of EGFR, HER2, and the related HER4 in enzymatic assays, with IC₅₀ values of 5, 20, and 60 nmol/L, respectively (Supplementary Table S1). More extensive testing in a panel of kinase assays representing diverse families revealed that BMS-690514 also inhibits members of the VEGFR family with IC₅₀ values in the range of 25 to 50 nmol/L (Supplementary Table S1). Inhibition was also observed with FLT3, Lck, and CAMKII, but with reduced potency compared with inhibition of EGFR. A more comprehensive survey of the kinase selectivity of BMS-690514 was conducted in binding assays by using 315 kinases (KinomeScan). The results of that analysis confirmed that BMS-690514 is a highly selective kinase inhibitor (Supplementary Table S2). MAP3K3, Ret, and members of the Src non-receptor tyrosine kinase family were the other kinases with detectable binding to BMS-690514, but with reduced affinity relative to EGFR.

The potency and specificity of BMS-690514 in inhibiting the kinase targets was further assessed in human tumor cell lines that are known to be dependent on EGFR and HER2 signaling. EGFR mutations have been shown to occur in 5% to 20% of patients with non–small cell lung cancer, with exon 19 deletion and point mutation at codon 858 each accounting for approximately 40% of all mutations (34). Non–small cell lung tumor cells with exon 19 deletion (HCC4006, HCC827, and PC9) were highly sensitive to BMS-690514, which inhibits their proliferation with IC₅₀ values of 2 to 35 nmol/L (Table 1). However, as with other EGFR kinase inhibitors, there was reduced potency in a cell line (H1975) that has both an activating mutation and a T790M gatekeeper mutation (IC₅₀ = 1,000 nmol/L). Tumor cell lines with EGFR gene amplification (DiFi, NCI-H2073, A431) were also highly sensitive to inhibition by BMS-690514. With the exception of GEO colon and L2987 lung tumor cell lines, the majority of colon and lung tumor cell lines which express wild-type EGFR in diploid copy number are not sensitive to BMS-690514, as exemplified by HCT116 (IC₅₀ > 5 μmol/L; Table 1). These findings are consistent with the notion that tumor cell lines with EGFR gene amplification or activating mutation are highly dependent on receptor signaling for growth and survival. Tumor cell lines that are dependent on HER2 signaling were also found to be highly sensitive to BMS-690514. Breast and gastric tumor cell lines that have HER2 gene amplification (N87, SNU-216, AU565, BT474, KPL4, and HCC202) were inhibited with IC₅₀ values of 20 to 60 nmol/L (Table 1). A mouse salivary gland tumor cell line Sal2, expressing a chimeric receptor with the cytoplasmic sequence of HER2 constitutively dimerized by its fusion to the CD8 extracellular sequence, was also highly sensitive to inhibition by BMS-690514. Table 1 also shows examples of cell lines (LX1 and MRC5), that do not express EGFR or HER2, and were not sensitive to BMS-690514 (IC₅₀ > 10 μmol/L). BMS-690514 inhibited the VEGF-dependent proliferation of HUVEC, consistent with its potency in inhibiting the kinase activity of VEGFR2.

The target specificity of BMS-690514 was further confirmed by measuring inhibition of receptor phosphorylation and signaling in tumor cells. Autophosphorylation of EGFR with activating mutation, either exon 19 deletion or L858R missense mutation, was inhibited by BMS-690514 with nanomolar IC₅₀, consistent with its potency in inhibiting cell proliferation (Fig. 1A and B). In GEO colon tumor cells, EGF-dependent phosphorylation of wild-type EGF receptor was inhibited with an IC₅₀ value of approximately 200 nmol/L (Fig. 1C). Similarly, MAP kinase phosphorylation was also inhibited with comparable potency. In N87 gastric tumor cells, in which the HER2 gene is amplified and overexpressed, BMS-690514 inhibited HER2 phosphorylation with an IC₅₀ value of approximately 45 nmol/L, whereas both MAP kinase and AKT phosphorylation were also inhibited with comparable potency (Fig 1D). Thus, the potency with which BMS-690514 inhibited EGFR and HER2 phosphorylation is in good agreement with the potency with which cell proliferation was inhibited (Table 1). These data lend support to an on-target mechanism by which BMS-690514 inhibits the proliferation of a broad spectrum of tumor cells that depend on EGFR and HER2 signaling. In addition to inhibiting EGFR and HER2, BMS-690514 also inhibits the related HER4 kinase in a biochemical assay. As HER4 has recently been implicated in cancer cell signaling as well, the potency of HER4 inhibition by BMS-690514 was also characterized in a cell assay. As reported previously, HER4 is subject to proteolytic cleavage (39, 40). Both the full-length and cleaved receptors are extensively phosphorylated on tyrosines when overexpressed (Fig. 1E). Lapatinib, an inhibitor of HER receptor kinases, inhibited HER4 full-length receptor phosphorylation with an apparent IC₅₀ value of approximately 700 nmol/L (Fig. 1E). BMS-690514 was more potent in inhibiting HER4 phosphorylation, with an IC₅₀ value of approximately 60 nmol/L (Fig. 1E). The difference in potency between the 2 compounds is consistent with the difference in their intrinsic potency in HER4 enzymatic assays, as lapatinib was reported to inhibit HER4 kinase with an IC₅₀ value of 367 nmol/L, compared with 60 nmol/L for BMS-690514 (14). Phosphorylation of the cleavage product seems to be somewhat more sensitive to inhibition by both lapatinib and BMS-690514, which is not surprising, as the removal of the extracellular sequence is likely to impact the binding of the compounds to the kinase pocket. A number of mutations in HER4, recently identified in melanomas, were also characterized for their sensitivity to BMS-690514 (7). The mutated receptors analyzed were all comparable to wild-type HER4 in their sensitivity to BMS-690514 or lapatinib (Supplementary Table S3). As a control, HER2 was also overexpressed and was inhibited by BMS-690514 and lapatinib with comparable potency. These data show that BMS-690514 is a potent inhibitor of HER4 kinase activity in cells.

Whereas activating mutations in EGFR have been established as a predictive marker for sensitivity to EGFR kinase inhibitors, the molecular basis for sensitivity to small molecule inhibition among breast tumor cells has not been as clearly elucidated. When BMS-690514 was tested in a panel of breast tumor cell lines, there was a clear demarcation between cell lines that are sensitive (IC₅₀ values between 16 and 170 nmol/L) and those that are resistant (IC₅₀ > 3 μmol/L; Fig. 2). Of the 10 sensitive cell lines, 8 have been documented to have HER2 copy number gain and have high levels of the receptor protein (35–37). Only 1 cell line (CAMA-1) with measurable expression of HER2 was insensitive to BMS-690514. Expression of EGFR among cell lines in this panel is variable, and there does not seem to be any correlation between EGFR levels and sensitivity to BMS-690514. HER3, which has been implicated as an important signaling partner, is also expressed in most of these cell lines without any apparent impact on sensitivity or resistance to BMS-690514. The 2 sensitive cell lines that do not have HER2 gene amplification (MDA-MB-175-VIII and HCC1187) have high levels of receptor as well. MDA-MB-175-VIII was previously shown to be dependent on autocrine HER2/HER3 signaling and is sensitive to pertuzumab, which interferes with receptor heterodimerization (41). Thus, overexpression of HER2 seems to be sufficient to predispose breast tumor cell lines to inhibition by BMS-690514, again underscoring its intrinsic potency to that receptor target. The cell line HCC1187 expresses both EGFR and HER2, and was further characterized for the ability of the 2 receptors to form heterodimer. Stimulation of HCC1187 cells with EGF resulted in an increase in EGFR and HER2 phosphorylation that was reversed by treatment with BMS-690514 (Supplementary Fig. S2). Reciprocal immunoprecipitation and Western blot analyses revealed that EGF stimulated the formation of heterodimer, which was further enhanced by treatment with BMS-690514. The stabilization of EGFR/HER2 heterodimer was previously reported for other small molecule kinase inhibitors as well (42, 43).

The antitumor activity of BMS-690514 was evaluated in a number of tumor xenografts that are dependent on EGFR or HER2 signaling. When administered once daily by oral gavage over 14 to 21 days, BMS-690514 was well tolerated in nude mice, with a maximum tolerated dose of 90 mg/kg. The PC9 non–small cell lung tumor, with an EGFR exon 19 mutation, is highly sensitive to BMS-690514, with tumor regression observed at doses higher than 3 mg/kg, given once daily (Supplementary Fig. S3A; Table 2). At 30 mg/kg, BMS-690514 yielded similar antitumor activity as the maximum tolerated dose (100 mg/kg) of erlotinib. However, a more sustained suppression of tumor growth was achieved with BMS-690514; tumors treated with erlotinib (100 mg/kg) began to show regrowth 30 days after treatment cessation, whereas those given BMS-690514 remained unmeasurable for more than 100 days after the cessation of dosing. There was only minor weight loss associated with the high dose (60 mg/kg) group for BMS-690514, which resolved with continued treatment (Supplementary Fig. S4). Other EGFR-dependent tumors, including several with wild-type EGFR (GEO, A549, L2987, and H292) were also sensitive to BMS-690514, with activity ranging from significant growth delay to complete tumor stasis observed at doses of 15 to 90 mg/kg (Table 2). The exquisite sensitivity of PC9 tumors to BMS-690514 may in part reflect the distinct biology of tumor cells that have become dependent on the oncogene. PC9 and HCC827 lung tumor cells, with mutated EGFR, undergo PARP cleavage when treated with low concentrations of BMS-690514 (Supplementary Fig. S5). By contrast, L2987 cells with diploid wild-type EGFR required high concentrations of the compound for induction of PARP cleavage (Supplementary Fig. S5). When tested in the N87 gastric tumor model, which has HER2 gene amplification, BMS-690514 was also highly efficacious, with complete inhibition at doses of 30 mg/kg or higher (Supplementary Fig. S3B). Other tumor xenografts known to be dependent on HER2 signaling are also highly responsive to treatment with BMS-690514, with efficacy shown in some of these models at doses as low as 7.5 mg/kg, and tumor regression at higher doses (Table 2). In mice, BMS-690514 has excellent oral bioavailability and dose-dependent increase in plasma exposure over a broad range of doses (44). The plasma concentrations at low doses were sufficient to account for the antitumor activity observed in the highly sensitive xenograft models (Supplementary Fig. S6).

The potency of BMS-690514 in inhibiting the growth of tumor xenografts may be attributable to a combination of inhibition of tumor cell proliferation and inhibition of angiogenesis, because the compound is also a potent inhibitor of the VEGF receptor family. In the HCT116 colon xenograft, BMS-690514 yielded 50% tumor growth inhibition when administered at 60 mg/kg, even though the compound is inactive in inhibiting the proliferation of the cell line (Tables 1 and 2). It is therefore plausible that the modest antitumor activity in a model such as HCT116 results from the antiangiogenic property of the compound. The antiangiogenic activity of BMS-690514 was further evaluated in a matrigel plug assay. In this assay, endothelial cells are recruited to invade the matrix and proliferate in response to proangiogenic cytokines. As previously shown, brivanib alaninate (BMS-582664), a small molecule inhibitor of VEGFR2 and FGFR1, inhibited endothelial cell invasion by 60%, when used at the optimal dose of 100 mg/kg (ref. 33; Fig. 3). BMS-690514 treatment resulted a dose-dependent inhibition of endothelial cell recruitment, yielding 20% and 40% inhibition at 30 and 90 mg/kg, respectively, that is statistically significant (Fig. 3). In separate studies, selective EGFR inhibitors had no effect on VEGF/FGF-dependent endothelial cell recruitment in this assay (data not shown). BMS-690514 was also evaluated for its effect on tumor blood flow by using a pharmacodynamic imaging approach. Tumor-bearing mice were treated with BMS-690514 and tumor blood flow was measured by DCE-MRI with the aid of a contrast agent. In animals given 2 consecutive doses, BMS-690514 reduced the accumulation of contrast agent by 30% and 60%, when given at 30 and 90 mg/kg, respectively (Table 3). K^trans was reduced by 26% and 37%, respectively, at 30 and 90 mg/kg. Sampling at a later time point (after 3 doses) showed similar effects on blood flow but to a greater magnitude.

Optimization of the pyrrolotriazine series of HER kinase inhibitors led to the identification of BMS-690514, which is highly potent in inhibiting all 3 HER kinases (EGFR, HER2, and HER4). BMS-690514 is, in addition, a potent inhibitor of the VEGF receptor family. Outside of these 2 receptor kinase families, only a small number of additional protein kinases were found to interact with BMS-690514, and none of these other kinases is known to have a role in regulating tumor cell proliferation. In cell assays measuring receptor kinase inhibition, BMS-690514 showed potency that was comparable to its potency in cell proliferation assays, an observation that further confirms its on-target mechanism of action. The potency of BMS-690514 in inhibiting tumor cell proliferation reflects the role of EGFR and HER2 in epithelial cancer. EGFR and HER2 gene amplification in lung, gastric, and breast tumor cells predispose them to inhibition by BMS-690514. In addition, non–small cell lung tumors with activating mutations in EGFR are highly sensitive to inhibition by BMS-690514. These features, combined with its oral bioavailability in mice, resulted in an excellent antitumor activity profile of BMS-690514 in preclinical tumor models. The compound is efficacious at multiple doses below its maximum tolerated dose. Tumor regression was observed in xenografts that are highly dependent on EGFR or HER2 signaling, whereas tumor stasis was observed in several models that express wild-type EGFR. Among breast tumor cells, BMS-690514 is very potent in inhibiting the proliferation of cells with HER2 copy number gain. In addition, the compound is also effective in inhibiting those cell lines with diploid HER2 copy number but signal through heterodimer formation with EGFR or HER3. The latter observation highlights the difference in the therapeutic utility of BMS-690514 compared with antibodies against HER2. Trastuzumab is efficacious only in tumors with high HER2 expression, whereas pertuzumab inhibits the heterodimerization of HER2 with other HER receptors, and is relatively ineffective in tumor cells with HER2 gene amplification (45). The potential clinical utility of BMS-690514 in tumors with EGFR/HER2 heterodimer signaling warrants further investigation. In addition, the inhibition of HER2 kinase activity may have utility in the treatment of HER2-positive breast tumors that develop resistance to trastuzumab by the generation of the intracellular fragment p95 (46).

By virtue of sequence homology among the HER receptor family, BMS-690514 is also a potent inhibitor of HER4. The potency of BMS-690514 in inhibiting HER4 autophosphorylation is approximately 10 times greater than that for lapatinib, reflecting the difference in intrinsic potency of the 2 compounds as measured in biochemical assays. The role of HER4 in cancer has remained mostly unexplored, although recently it was reported that HER4 is mutated in approximately 19% of melanomas. The HER4 mutations that were analyzed here were previously reported to show enhanced focus-forming activity compared with the wild-type receptor when transfected in NIH3T3 cells (7). These specific mutations reside in the extracellular sequence and had no apparent impact on the sensitivity of the receptor to either lapatinib or BMS-690514. Further evaluation of the potential utility of BMS-690514 as a treatment for melanomas will have to await the availability of melanoma cell lines with HER4 mutations.

BMS-690514 inhibits the proliferation of endothelial cells in vitro and their activity in a matrigel plug assay, consistent with its potency in inhibiting VEGF receptor kinases. A functional demonstration of the antiangiogenic property of BMS-690514 was provided by the modulation of blood flow in tumor-bearing animals treated with the compound. In an independent study, it was recently observed that BMS-690514 decreased the microvascular density in a xenograft model (47). These observations together provide strong support for the antiangiogenic activity of BMS-690514. BMS-690514 is less potent than other compounds (e.g., brivanib) that were specifically optimized for VEGFR inhibition. As a result, BMS-690514 is not as broadly efficacious as other VEGFR kinase inhibitors, and its efficacy is restricted to tumors that depend on EGFR and/or HER2 signaling.

The data reported here have shown that BMS-690514 is a novel kinase inhibitor that combines potency against the HER kinases with inhibition of the VEGF receptor kinases. Its potency against all 3 HER kinases will make BMS-690514 an excellent tool for further broadening the therapeutic utility of targeting this receptor family, that has not been possible by using agents that are highly selective for single receptors. In addition, the ability to inhibit tumor angiogenesis may impart BMS-690514 enhanced efficacy. These properties make BMS-690514 a suitable candidate for exploring the clinical utility of combining tumor and host inhibition in a single treatment modality.

All authors except B. Krishnan are current employees of Bristol-Myers Squibb. B. Krishnan was an employee of Bristol-Myers Squibb when these studies were performed.

The authors thank Rolf Ryseck for DNA constructs and Kristen Kellar for assistance with the HUVEC proliferation assay.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.