SU5416, a selective inhibitor of the tyrosine kinase activity of the vascular endothelial growth factor (VEGF) receptor Flk-1/KDR, is currently in Phase III clinical trials for the treatment of advanced malignancies. In cellular assays, SU5416 inhibits the VEGF-dependent mitogenic/proliferative response of human umbilical vein endothelial cells (HUVECs). In tumor xenograft models, SU5416 inhibits the growth of tumors from a variety of origins by inhibiting tumor angiogenesis. In three different human tumor xenograft models, infrequent (once or twice a week) administration of SU5416 is efficacious despite the fact that it has a short plasma half-life (30 min), which suggests that SU5416 has long-lasting inhibitory activity in vivo. The goal of the present study was to determine the basis for the prolonged activity of SU5416. The results indicate that a short (3 h) exposure to 5 μm SU5416 (to mimic plasma levels of the compound as measured in patients who were receiving SU5416 therapy) produced long-lasting (at least 72 h) inhibition of the VEGF-dependent proliferation of HUVECs in culture, which indicate that SU5416 has long-lasting inhibitory activity in vitro as well as in vivo. SU5416 treatment of HUVECs did not affect surface expression of Flk-1/KDR or the affinity of the receptor for VEGF. Instead, the durability of the in vitro activity of SU5416 was shown to be attributable to its long-lasting ability to specifically inhibit VEGF-dependent phosphorylation of Flk-1/KDR and subsequent downstream signaling, although SU5416 is not an irreversible inhibitor of Flk-1/KDR tyrosine kinase activity. The long-lasting inhibition of cellular responses to VEGF was attributable to the accumulation of SU5416 in cells, as shown using radiolabeled compound, such that inhibitory cellular concentrations of SU5416 are maintained long after the removal of the compound from the medium. The long-lasting inhibitory activity of SU5416 in vitro is consistent with the finding that SU5416 has demonstrated evidence of biological activity in clinical studies when administered twice a week despite a short plasma half-life.
Angiogenesis, the sprouting of capillaries from preexisting blood vessels, is essential for the sustained growth of solid tumors (1 , 2) . Tumors that lack adequate vasculature become necrotic (3) and/or apoptotic (4 , 5) and do not grow beyond a limited size. However, tumors that undergo neovascularization acquire the ability to enter a phase of rapid growth and generally exhibit increased metastatic potential (6) . The significance of angiogenesis in the progression of human cancers has been highlighted by recent studies that relate angiogenic phenotype to patient survival (reviewed in Ref. 7 ). These studies found that the number of microvessels in primary tumors has prognostic significance in lung (8 , 9) , breast (10 , 11) , bladder (12) , and colon (13) carcinomas, and in tumors of the oral cavity (14) . Thus, inhibition of angiogenesis has been identified as an attractive approach for the treatment of human cancers (15 , 16) .
Although many pro- and antiangiogenic factors have been identified (reviewed in Refs. 17 and 18 ), several lines of evidence indicate that VEGF2 and its cognate receptor Flk-1/KDR play a major role in tumor angiogenesis (15) . In particular, genetic (19, 20, 21, 22, 23, 24) and biochemical (25, 26, 27) approaches have been used to demonstrate that disruption of VEGF signaling via Flk-1/KDR results in the selective inhibition of tumor growth in animal models. On the basis of these results, clinical studies have been initiated to determine the therapeutic potential of neutralizing antibodies directed against VEGF and small molecule inhibitors of Flk-1/KDR tyrosine kinase activity (reviewed in Ref. 7 ).
The first small molecule inhibitor of Flk-1/KDR tyrosine kinase activity to undergo large-scale clinical trials is SU5416 (28 , 29) . In biochemical studies, SU5416 is a potent, competitive (with respect to ATP) inhibitor of the tyrosine kinase activity of Flk-1/KDR with a Ki value of 0.16μ m (28 , 29) . Although SU5416 exhibited similar activity against platelet-derived growth factor receptor (Ki = 0.32μ m) in biochemical assays, SU5416 was a 20-fold less potent inhibitor of platelet-derived growth factor receptor phosphorylation than of Flk-1 phosphorylation in cells (28 , 29) . In contrast, SU5416 is a weak inhibitor of FGF receptor (Ki = 19.5μ m) and does not inhibit the epidermal growth factor receptor (28 , 29) . Moreover, the biochemical IC50s of SU5416 against the insulin-like growth factor I receptor, Met, Src, Abl, and Lck are at least 10μ m (data not shown). Thus, SU5416 is considered to be a selective inhibitor of Flk-1/KDR kinase activity.
In cell culture assays, SU5416 inhibits the KDR-mediated VEGF-dependent mitogenic response of HUVECs with an IC50 value of 40 nm but does not inhibit the mitogenic response of HUVECs stimulated with aFGF (28) . On the basis of its ability to potently inhibit VEGF-stimulated HUVEC proliferation in vitro, it was anticipated that SU5416 would have antiangiogenic activity in vivo, which has subsequently been directly demonstrated (28 , 30 , 31) . Consistent with its in vivo antiangiogenic activity, SU5416 inhibits the growth of a variety of tumor xenograft models, although it does not inhibit the proliferation of those same tumor cells in culture (28 , 29) .
In pharmacokinetic studies, SU5416 has a short (∼30 min) plasma half-life in both mice and rats (32) . Consequently, most studies of the in vivo activity of SU5416 have used daily administration of the compound. However, preliminary results using the A375 human melanoma tumor model indicated that SU5416 was also efficacious when given twice weekly (29) . The ability of infrequent administration of SU5416 to inhibit tumor growth in preclinical studies despite the short plasma half-life suggests that the inhibitory activity of SU5416 on Flk-1/KDR is long lived. Consistent with the results obtained in preclinical studies, a twice-weekly i.v. administration of SU5416 during clinical trials has demonstrated clinical activity as defined by increased time to disease progression in patients with advanced malignancies and AIDS-associated Kaposi’s sarcoma despite a short plasma half-life (33 , 34) . The goal of the present study was to determine the molecular basis for the long-lived inhibitory activity of SU5416 observed in preclinical models and in the clinic.
MATERIALS AND METHODS
SU5416 Chemical Synthesis.
SU5416 (3-[2,4-dimethylpyrrol-5-yl methylidenyl]-2-indolinone) was synthesized at the facilities of SUGEN, Inc., as described previously (35) .
Unless otherwise indicated, cell culture reagents were obtained from Life Technologies, Inc. (Gaithersburg, MD). A431 human epidermoid carcinoma cells and NCI-HT-29 human colon carcinoma cells were obtained from American Type Culture Collection (ATCC; Rockville, MD) and were propagated using standard tissue culture procedures in the medium suggested by ATCC. HUVECs were obtained from Clonetics (San Diego, CA) and were maintained in EGM (Clonetics) containing 2% FBS. NIH 3T3 cells, engineered to overexpress the murine VEGF receptor Flk-1 (3T3/Flk-1 cells), were obtained from Dr. Axel Ullrich (Max-Planck Institut fur Biochemie, Martinsried, Germany) and were maintained in DMEM supplemented with 10% FBS and 2 mm glutamine. Cells were propagated at 37°C in a humidified atmosphere of 5% CO2 using standard cell culture techniques.
s.c. Xenograft Model in Athymic Mice.
Female nu/nu mice (8–12 weeks old; 20 g) were obtained from Charles River (Wilmington, MA). Animals were maintained under clean-room conditions in sterile filter-top cages with Aspen Chip bedding housed on HEPA-filtered ventilated racks. Animals received sterile rodent chow and water ad libitum. Animal experiments were conducted as described previously (28) , in accordance to Institutional Animal Care and Use Committee guidelines in the SUGEN Animal Facility, which has been accredited by Association for Assessment and Accreditation of Laboratory Animal Care, International.
Tumor cells to be implanted into athymic mice were harvested at or near confluence by incubation with 0.05% trypsin-EDTA. Cells were pelleted by centrifugation at 450 × g for 5–10 min, and cell pellets were resuspended in sterile PBS to a concentration of 3–5 × 107 cells/ml. Cells (3–5 × 106/animal) were implanted s.c. into the hind flank region of mice on day 0 as described previously (20) . i.p. bolus injections (50 μl) of SU5416 in DMSO or vehicle control (DMSO alone) were started one day after the implantation of cells and continued with the regimens described until termination of the experiment, when tumors in the vehicle-treated control animals grew to an average size of approximately 1000 mm3 . Tumor growth was measured twice weekly using Vernier calipers for the duration of the treatment. Tumor volumes were calculated as the product of length × width × height. Ps were calculated using the two-tailed Student’s t test based on measurements taken on the last day of each experiment.
Endothelial Cell Proliferation Assays.
Endothelial cell proliferation assays were similar to the endothelial cell mitogenesis assays described previously (28) . Briefly, HUVECs were plated in 96-well flat-bottomed tissue culture plates (1.5 × 104 cells/100 μl/well) in starvation medium composed of EBM (Clonetics) containing 0.5% FBS. After an overnight incubation to quiesce the cells, the medium was replaced with starvation medium containing 3-fold serial dilutions of SU5416 (16.7–0.02 μm). Two or 3 h later, the medium containing SU5416 was removed and replaced with starvation medium containing VEGF (20 ng/ml; R&D Systems, Minneapolis, MN), aFGF (1 ng/ml; Boehringer Mannheim, Indianapolis, IN), or no ligand (media control) as indicated. At the indicated times after removal of SU5416-containing medium, the relative number of cells in triplicate wells was determined using the SRB (Sigma Chemical Co., St. Louis, MO) assay (36) .
Surface Expression of KDR.
Surface expression of KDR on HUVECs was assessed using standard flow cytometry procedures and commercially available antibodies. Briefly, serum-starved monolayers of HUVECs were incubated in the presence or absence of SU5416 at the indicated concentration (1 or 5μ m) and for the indicated time (3 or 24 h). After treatment, HUVECs were released from the tissue culture plates by incubation with 5 mm EGTA in Ca2+/Mg2+-free PBS. Released cells were harvested by centrifugation, washed in PBS containing 1% FBS, and then incubated for 1 h with a 1:100 dilution of a monoclonal antibody directed against an epitope on the extracellular domain of KDR (Sigma Chemical Co.). The cells were then washed with PBS containing 1% FBS to remove free primary antibody before being incubated for 45 min with a 1:100 dilution of PE-conjugated F(ab′)2 fragment of goat-antimouse IgG as secondary (2°) antibody (Caltag, Burlingame, CA). After a final wash to remove free secondary antibody, cells were analyzed for surface expression of KDR using a Becton Dickinson FACScan.
125I-VEGF Binding Assays.
VEGF binding to KDR and Flt-1 on HUVECs was assessed using 125I-VEGF. Briefly, confluent monolayers of HUVECs in 12-well tissue culture plates were serum starved overnight in starvation medium (EBM containing 0.5% FBS). The following day, the cells were incubated in the presence or absence of SU5416 at the indicated concentration (1 or 5 μm) and for the indicated time to determine the effect of a short (3 h) or long (24 h) exposure to compound. At the end of the treatment period, the cell monolayers were washed three times at room temperature with binding medium [DMEM containing 25 mm HEPES (pH 7.4) and 0.15% gelatin]. After the third wash, 0.5 ml of binding medium containing 3-fold serial dilutions of 125I-VEGF (100 μCi/μg; New England Nuclear, Boston, MA) ranging in concentration from 0.23 to 500 pm was added to duplicate wells to determine total 125I-VEGF binding. Separate 0.5-ml aliquots of the same dilutions of 125I-VEGF in binding medium containing 100 nm nonradioactive VEGF (R&D Systems) were added to another set of duplicate wells containing HUVECs to determine nonspecific 125I-VEGF binding. After a 2-h incubation at room temperature, the binding medium was removed, the cells were washed three times with ice-cold PBS containing 0.1% BSA (Sigma Chemical Co.), and the cells were lysed with the addition of 0.5 ml of 0.1 n NaOH. The amount of radioactivity in the binding buffer that was removed from the cells and that was in the cell lysate was counted in a Packard gamma counter to determine free and bound radioactivity, respectively. Specifically bound counts were determined by subtracting nonspecific from total counts at each free 125I-VEGF concentration. Scatchard analyses (37) of the data were used to determine dissociation constants (KD) for the binding of 125I-VEGF to Flt-1 and KDR.
Cellular Flk-1 and ERK 1/2 Phosphorylation Assays.
The effect of SU5416 on VEGF-dependent tyrosine phosphorylation of Flk-1 was determined as described previously (28) . Briefly, twenty confluent monolayers of 3T3/Flk-1 cells or HUVECs in 6-well or 10-cm tissue culture plates, respectively, were quiesced by serum starvation overnight in starvation medium (EBM containing 0.5% FBS). SU5416 was then added to 10 wells to a final concentration of 5μ m; the remaining 10 wells did not receive SU5416. After a short (2 or 3 h as indicated) exposure to SU5416, the medium was removed from five of the SU5416-treated wells, cell monolayer was washed, and fresh starvation medium was added; these samples were referred to as the “wash” SU5416 samples. The other five SU5416-treated samples did not have the compound washed out but, rather, were constantly exposed to SU5416 for the duration of the experiment. Immediately (0 h) or 6, 24, 32, or 48 h after the washout procedure, VEGF was added to a final concentration of 100 ng/ml to one well each of the untreated, the SU5416 wash, and the constant SU5416 wells; no VEGF was added to another untreated well that was subsequently referred to as the unstimulated sample. After a 10-min VEGF stimulation at 37°C, the medium was removed, the cell monolayer washed with ice cold PBS, and the cells scraped into HNTG buffer [20 mm HEPES (pH 7.5), 150 mm NaCl, 0.2% Triton X-100, and 10% glycerol) containing 1 mm NaVO4, 50 mm NaF, and protease inhibitors. Preparation of cell lysates, separation of cellular proteins (30 μg), and immunoblotting with antiphosphotyrosine antibody (SUGEN reagent) were performed as described previously (28) . Immunoreactive proteins were detected using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech, Piscataway, NJ). Membranes were then stripped with elution buffer (Pierce, Rockford, IL), and reprobed with an antibody directed against Flk-1/KDR (Santa Cruz Biotechnology, Santa Cruz, CA) to determine the amount of Flk-1/KDR in each lane.
A similar approach using antiphospho ERK 1/2 and anti-ERK 1/2 antibodies (New England Biolabs, Beverly, MA) was used to determine the effect of SU5416 on VEGF-dependent phosphorylation of ERK 1/2.
Flk-1 Receptor Kinase Assay.
The biochemical Flk-1 receptor tyrosine kinase assay was performed as described previously (28) . Briefly, detergent-solubilized proteins (50 μg) from 3T3/Flk-1 cells were added to each well of a polystyrene ELISA plate that had been precoated with a monoclonal antibody (SUGEN reagent) that recognizes Flk-1. After an overnight incubation at 4°C to allow the Flk-1 to be captured by the adsorbed antibody, the wells were washed three times with deionized water and once with TBST [50 mm Tris (pH 7.0), 150 mm NaCl, and 0.1% Triton X-100] to remove nonabsorbed proteins. Kinase reaction buffer [25 mm Tris, (pH 7.0), 100 mm NaCl, 10 mm MnCl2, 2% glycerol, 0.5 mm DTT, and 0.1% Triton X-100] was then added to each well. The kinase reaction resulting in the autophosphorylation of Flk-1 on tyrosine residues was initiated with the addition of ATP (10–30μ m final concentration) and was allowed to proceed at room temperature for 60 min. The reaction was then stopped with the addition of EDTA to a final concentration of 20 mm. The wells were washed as above and then probed for 45 min with a biotinylated monoclonal antibody directed against phosphotyrosine (UBI, Lake Placid, NY). The wells were then washed to remove unbound antiphosphotyrosine antibody prior to the addition of avidin-conjugated horseradish peroxidase H (Vector Laboratories, Burlingame, CA) for 30 min. After washing, substrate[ 3,3′,5,5′-tetramethyl benzidine dihydrochloride and H2O2 (Turbo-TMB; Pierce)] was added to the wells. After 15 min, reactions were stopped with the addition of H2SO4, and the relative amount of colored product in each well was determined based on A450 nm values measured using a 96-well microtiter plate reader. In experiments to determine the inhibitory effect of SU5416, 3-fold serial dilutions of SU5416 were added to the captured Flk-1 enzyme just prior to the addition of ATP to start the autophosphorylation reaction.
[14C]SU5416 Uptake and Washout Assays.
Confluent monolayers of HUVECs in 24-well tissue culture plates were serum starved overnight in starvation medium (EBM containing 0.5% FBS). To determine the rate of uptake of[ 14C]SU5416 by the HUVECs, on the following day, the medium was replaced with 1 ml of starvation medium containing 4 μm [14C]SU5416 (13.56 mCi/mmol; SynPep, Dublin, CA). At the times indicated in Fig. 6⇓ , the medium was harvested from duplicate wells, the cell monolayers were washed twice with PBS, and the cells were lysed with 0.4 ml/well of 0.1 n NaOH. The radioactivity in the starvation medium and cell lysates was determined by scintillation counting (Beckman counter) after the addition of 5 ml of Ready Safe Liquid Scintillation Cocktail (Beckman, Palo Alto, CA). To determine the time course of the washout of [14C]SU5416 from HUVECs, serum-starved cells in 24-well plates were initially incubated with starvation medium containing 4 μm [14C]SU5416 for 3 h. Cell monolayers were then washed twice with PBS, and the medium was replaced with fresh starvation medium without SU5416. At the indicated times after washout, medium was removed from duplicate wells, and the amount of cell-associated [14C]SU5416 was determined as described above. The concentration of cell-associated[ 14C]SU5416 was calculated using 8μ l/106 HUVECs as the cell volume, which was determined using the equation V= 4 /3Π r3 for a sphere (where V = volume and r = radius of cell) and an average cell radius of 12.5 μm for HUVECs released from tissue culture plates by trypsinization based on visual inspection of cells in a hemacytometer.
Long-lasting Effect of SU5416 on A431 Tumor Growth.
The majority of the animal experiments carried out with SU5416 have used daily administration of the compound because of the rapid clearance of the compound from plasma (28, 29, 30, 31) . However, twice-weekly administration of SU5416 at 50 mg/kg was as efficacious as daily administration of 25 mg/kg SU5416 in inhibiting the growth of A375 human melanoma xenograft tumors (29) . Because A375 xenografts are very sensitive to the antitumor activity of SU5416 (28) , the possibility existed that SU5416 was able to exert antitumor activity with infrequent administration because the A375 xenograft model was particularly sensitive to SU5416. To address this question, subsequent experiments were conducted using A431 human epidermoid carcinoma and NCI-HT-29 human colon carcinoma tumors, which are less sensitive than A375 tumors to the antitumor activity of SU5416 in athymic mice (29) .
As shown in Table 1⇓ , SU5416, administered i.p. daily at 25 mg/kg/day, inhibited A431 tumor growth by 51% (P = 0.0063), which is comparable with previous results with this tumor model (28) . Administration of SU5416 at 25 mg/kg twice weekly did not cause significant inhibition of tumor growth. In contrast, i.p. administration of SU5416 at 50 mg/kg once, two or three times weekly resulted in the statistically significant inhibition of A431 tumor growth comparable with that seen with daily administration of SU5416 at 25 mg/kg/day (Table 1)⇓ . A similar result was obtained with NCI-HT-29 tumors, in that once weekly administration of SU5416 at 50 mg/kg was as efficacious as daily administration of the compound at 25 mg/kg/day (Table 1)⇓ . Thus, the ability of SU5416 to inhibit the growth of tumors with infrequent dosing is evident in all of the three (A375, A431, and NCI-HT-29) tumor models tested, which suggests that the long-lived inhibitory activity of SU5416 would be evident with other tumor models susceptible to inhibition by daily administration of SU5416.
Long-lasting Effect of SU5416 on HUVEC Proliferation in Culture.
The durability of the inhibitory activity of SU5416 in an in vitro assay was directly determined using a HUVEC proliferation assay. In this assay, the ability of a 3-h exposure to SU5416 to provide prolonged inhibition of VEGF-induced proliferation of HUVECs was determined. The results of a representative assay are shown in Fig. 1⇓ .
At 0 h (immediately after the three-hour exposure to SU5416 and prior to addition of ligand), the SRB signal (indicative of relative cell number) is similar in all of the test wells regardless of SU5416 concentration and in all of the three treatment groups (Fig. 1⇓ A). With increasing time (48-h data are shown in Fig. 1⇓ B), fewer cells were detected in the media control samples, consistent with the inability of HUVECs to propagate or survive under serum-starvation conditions. The decrease in the number of cells with an increasing time of serum starvation was confirmed by a visual inspection of the cultures and did not vary with SU5416 concentration. In the aFGF-stimulated samples, there was an increase in the number of cells at the 48-h time point regardless of SU5416 treatment, which indicated that aFGF is still able to induce HUVEC proliferation in the presence of SU5416. This result is consistent with the demonstration that SU5416 does not block the mitogenic/proliferative response of HUVECs stimulated with aFGF (28) . In contrast, a dose-dependent effect of SU5416 was observed in samples stimulated with VEGF (Fig. 1⇓ C). In these samples, there was no decrease in the number of cells in wells initially exposed to low concentrations (<0.6μ m) of SU5416, which indicated that, in these wells, VEGF induced proliferation of the HUVECs despite the prolonged serum starvation. However, there was a marked decrease in the number of cells in wells exposed to the higher concentrations of SU5416, such that the number of cells in the VEGF-stimulated wells exposed to the two highest concentrations of SU5416 were the same as those in the media control wells. These data indicate that the two highest concentrations of SU5416 completely blocked the proliferative effect of the VEGF, and that the IC50 value for SU5416 in this assay is 1.5–2 μm. This IC50 value is higher than the 40-nm value reported for HUVECs exposed to SU5416 for the duration of the mitogenesis assay (48 h; Ref. 28 ) but is comparable with the IC50 value of ∼1μ m obtained in the mitogenesis assay when HUVECs are exposed to SU5416 during only the first 0.5–5 h of the assay.3
To confirm that an initial 3-h exposure to SU5416 was still having an inhibitory effect more than 48 h after removal of the compound, the cells were given a second stimulation with VEGF or aFGF 51 h after the washout, and the relative number of cells was determined based on the SRB signal 21 h later (72 h after the washout). As shown in Fig. 1⇓ C, at the 72 h time point there was a further decrease in the number of cells in unstimulated (media control) wells and in VEGF-stimulated wells that contained cells initially treated with the higher concentrations of SU5416. Again, SU5416 had no effect on the proliferative response of HUVECs stimulated with aFGF. Thus, even 72 h after a short exposure to SU5416, the effect of VEGF stimulation is specifically inhibited in HUVECs.
Effect of SU5416 on VEGF Binding via Flk-1/KDR.
To investigate the basis for the long-lived in vitro inhibitory activity of SU5416, the effect of SU5416 on the surface expression of KDR and its ability to bind VEGF were investigated. As shown in Fig. 2⇓ , similar levels of KDR are detected on the surface of untreated HUVECs or of HUVECs treated with 1 μm SU5416 for 3 h. Similar results were obtained with HUVECs incubated with a higher concentration of SU5416 (5 μm) and for a longer period of time (24 h; data not shown). Thus, SU5416 treatment of HUVECs does not decrease the surface expression of KDR.
Alternatively, SU5416 could inhibit VEGF-dependent HUVEC proliferation by blocking the binding of VEGF to its receptors through a secondary effect on receptor conformation. To address this question, the binding of 125I-VEGF to SU5416-treated and untreated HUVECs was investigated. As previously reported (38) , Scatchard analysis of specific 125I-VEGF binding data enabled the separate evaluation of the binding of VEGF to KDR and Flt-1 (a second VEGF receptor expressed on HUVECs). The dissociation constants (KD) determined for VEGF binding to Flt-1 and KDR were 17 pm and 390 pm, respectively, which are comparable with reported values (38) . As shown in Table 2⇓ , SU5416 treatment did not effect either the affinity (KD) or the number of functional KDR or Flt-1 binding sites on HUVECs.
Long-lasting Effect of SU5416 on Flk-1/KDR Phosphorylation and Subsequent Signaling.
Because the long-lasting inhibitory activity of SU5416 could not be attributed to alterations in VEGF binding to cellular receptors, the durability of the effect of SU5416 on VEGF-dependent Flk-1/KDR phosphorylation was investigated. As shown in Fig. 3⇓ , a 3-h pretreatment with SU5416, followed by a washout, was sufficient to inhibit VEGF-dependent phosphorylation of Flk-1 in 3T3/Flk-1 cells for 48 h, the longest time period tested. Importantly, the inhibition of Flk-1 phosphorylation observed in the washout sample at 48 h was comparable with that seen in the sample from cells exposed to SU5416 for the full 48 h (compare the third and fourth lanes of the 48-h sample).
Having demonstrated that SU5416 produced a long-lasting inhibition of VEGF-dependent Flk-1 phosphorylation, we next investigated the ability of SU5416 to confer long-term inhibition of VEGF-dependent signaling in HUVECs. Phosphorylation of ERK 1/2 (p42/p44MAP kinase) was selected as an indicator of VEGF-dependent Flk-1/KDR activation in these studies because it is known to be an early downstream target of Flk-1/KDR activation (39) , because a significant portion of this abundant protein is phosphorylated in response to VEGF stimulation of KDR in HUVECs, and because reliable reagents are available to detect phosphorylated ERK 1/2 separate from total ERK 1/2 in cell extracts. As shown in Fig. 4⇓ , even 32 h after the removal of SU5416 from the medium, VEGF-dependent phosphorylation of ERK 1/2 was inhibited to an extent similar to that observed in HUVECs exposed to SU5416 for the full 32-h period (compare the third and fourth lanes in the 32-h sample). Thus, the long-lived inhibition of the VEGF-dependent mitogenesis/proliferation of HUVECs by SU5416 in vitro is attributable to a long-lived inhibition of the VEGF-dependent signaling via Flk-1/KDR.
To determine whether the long-lived effect of SU5416 on ligand-dependent phosphorylation of ERK 1/2 was specific for VEGF stimulation, the ability of SU5416 to inhibit ERK 1/2 phosphorylation in response to aFGF was investigated. As shown in Fig. 5⇓ , SU5416 did not inhibit ERK 1/2 phosphorylation in aFGF-stimulated HUVECs (Fig. 5⇓ , right side of each panel) under conditions that VEGF-dependent ERK 1/2 phosphorylation was inhibited (Fig. 5⇓ , left side of each panel). This result, which is consistent with the inability of SU5416 to inhibit aFGF-dependent mitogenesis in HUVECs, demonstrates that the long-lasting inhibition by SU5416 retains its specificity for VEGF-dependent signals. It should be noted that time points beyond 48 h were not investigated because preliminary results indicated that the response of HUVECs to ligand stimulation in serum-starvation medium was more variable at later time points even in the absence of SU5416.
SU5416 Inhibition of Flk-1 Tyrosine Kinase Activity.
Previous biochemical studies have indicated that SU5416 is a potent, competitive (with respect to ATP) inhibitor of the tyrosine kinase activity of Flk-1 (Ki = 0.16μ m for SU5416, Km = 0.53 μm for ATP; Refs. 28 , 29 ). One possible explanation for the long-lasting inhibitory activity of SU5416 would be that it acts as an irreversible inhibitor of the Flk-1/KDR tyrosine kinase activity. To directly address this possibility, the durability of the inhibitory activity of SU5416 was determined in a biochemical kinase assay. In this assay, SU5416 inhibited Flk-1 kinase activity (IC50 = 0.75 μm when 30μ m ATP was used as substrate) only if SU5416 was present when the reaction was initiated with the addition of ATP. In contrast, little or no inhibitory activity was observed if SU5416 was removed by washing prior to the addition of ATP even after preincubations for as long as 30 min with high (20μ m) concentrations of compound. Thus, SU5416 is not an irreversible inhibitor of Flk-1 kinase activity.
Concentration of SU5416 in HUVECs.
The demonstration that SU5416 is not an irreversible inhibitor of Flk-1 tyrosine kinase activity but is able to inhibit VEGF-stimulated Flk-1 tyrosine phosphorylation for prolonged periods in cells suggested that SU5416 may be sequestered intracellularly. To directly address this possibility, the uptake and washout of SU5416 was determined in HUVECs incubated in starvation medium containing 4 μm[ 14C]SU5416.
As shown in Fig. 6⇓ A, SU5416 is rapidly taken up into HUVECs, reaching a plateau after 2 h. The cellular concentration of[ 14C]SU5416 after 3-h incubation in medium containing 4 μm[ 14C]SU5416 was determined to be approximately 450 μm, which indicated that the compound is concentrated in the cell. The majority of the cellular SU5416 present at the end of the uptake period also rapidly washes out of the cells. As shown in Fig. 6⇓ B, 30 min after replacing the SU5416-containing medium with compound-free medium, the cellular concentration of SU5416 dropped 10-fold compared with the amount present at the end of the 3-h uptake period. However, even after more than 90% of the compound washed out, the cellular SU5416 concentration was ∼25 μm. Importantly, there was no further decrease in the cellular concentration of SU5416, so that even 48 h after removal of compound, the cellular concentration of SU5416 was still ∼25 μm. This concentration exceeds the concentration of SU5416 required to inhibit Flk-1/KDR phosphorylation and VEGF-stimulated HUVEC proliferation. Also of note is the fact that the 14C recovered from cells at the end of the uptake and washout periods coeluted with SU5416 when subjected to high-performance liquid chromatography analysis,4 which indicated that the cell-associated radioactivity was SU5416. Thus, these data indicate that SU5416 is concentrated in cells, and that the cells maintain an inhibitory concentration of SU5416 for a prolonged period even when the compound is no longer present in the medium.
For comparison, similar studies were performed using[ 14C]SU6668, a compound structurally related to, but less hydrophobic than, SU5416 (40) . In these experiments, SU6668 was found: to be taken up at a much slower rate than SU5416; to achieve a lower final cellular concentration; and to rapidly and completely wash out of the cells when the compound was removed from the medium (data not shown). Consistent with the rapid and efficient washout of this compound, SU6668 did not have long-lasting activity in any of the functional assays described above.
The results of this study indicate that infrequent administration (once or twice weekly) of SU5416 effectively inhibits the s.c. growth of A431 human epidermoid carcinoma and NCI-HT-29 human colon carcinoma tumor xenografts in athymic mice. This confirms and extends previous results obtained with the A375 human melanoma xenograft model (29) , and is consistent with the observed clinical benefit derived from infrequent (twice weekly) dosing of SU5416 for the management of advanced malignancies (34) . Thus, there is both preclinical and clinical evidence to suggest that SU5416 has effects when administered infrequently. However, the explanation for the long-lived activity of SU5416, which has a short plasma half-life (32 , 34) , remained unclear.
Preclinical studies have demonstrated that SU5416 exerts its antitumor activity by inhibiting tumor angiogenesis (28 , 30 , 31) , and that its antiangiogenic activity is attributable to its ability to selectively inhibit VEGF signaling via Flk-1/KDR in vascular endothelial cells (28) . Therefore, the VEGF-dependent HUVEC proliferation assay was used to investigate whether the long-lived inhibitory activity of SU5416 could be detected in a cell-based assay. As detailed above, SU5416 has long-lasting and specific inhibitory activity in these cells. Thus, HUVECs, in addition to the 3T3/Flk-1 cells engineered to express high levels of Flk-1, were used to investigate the basis of the long-lived activity of SU5416.
Initial studies with SU5416 in the cell-based assays indicated that a 24 h exposure to SU5416 does not alter the surface expression of Flk-1/KDR or its ability to bind VEGF. In preliminary studies (data not shown), it was also determined that SU5416 blocked internalization of Flk-1/KDR in response to VEGF binding. These results confirm a recent report that ligand-stimulated phosphorylation of Flk-1/KDR is required for the receptor to be internalized (41) . These results indicate that SU5416 does not inhibit cellular responses to VEGF by down-regulating the number of functional VEGF receptors on the cell surface; but rather they confirm that SU5416 exerts its inhibitory activity by inhibiting the kinase activity of Flk-1/KDR without altering the ability of Flk-1/KDR to bind the ligand.
Subsequent studies with SU5416 in the cell-based assays indicated that a short exposure to SU5416 produced durable inhibition of VEGF-dependent phosphorylation of Flk-1. Consistent with its ability to block Flk-1/KDR phosphorylation, SU5416 was also able to inhibit VEGF-induced phosphorylation of ERK 1/2, a downstream target of Flk-1/KDR. Importantly, under conditions required to confer long-lasting inhibition of VEGF-dependent phosphorylation of ERK 1/2, SU5416 did not inhibit ERK 1/2 phosphorylation in response to stimulation with aFGF, which indicated that the inhibitory activity of SU5416 does not lose its selectivity.
Several other inhibitors of receptor tyrosine kinases have demonstrated long-lived inhibition of ligand-dependent receptor phosphorylation in biochemical assays and in cells (42, 43, 44, 45, 46, 47, 48) . Biochemical studies with these other inhibitors have indicated that they covalently associate with the receptor and act as irreversible inhibitors. However, previous studies with SU5416 have indicated that it is a competitive (not an irreversible) inhibitor of Flk-1/KDR tyrosine kinase activity (29) . The conclusion that SU5416 is not an irreversible inhibitor of Flk-1/KDR kinase activity is supported by results obtained during this study which indicate that SU5416 can be readily washed out of the Flk-1/KDR kinase domain.
The fact that SU5416 has long-lived biological activity although it is not an irreversible inhibitor of Flk-1 tyrosine kinase activity suggested that effective concentrations of SU5416 might be retained in the cell when it is no longer present outside the cell. Anecdotal evidence that SU5416 was retained in cells was initially provided by observations that cells incubated with SU5416 became orange (the color of SU5416), and that they remained orange even after extensive washing.3 This suggestion was confirmed in the present study using [14C]SU5416. These studies also demonstrated that cells concentrate SU5416. The cellular concentration of SU5416 in HUVECs incubated for 2 h in medium containing 4μ m compound was 450 μm. The ability of the cells to concentrate SU5416 likely explains why the IC50 of SU5416 in the HUVEC mitogenic assay is 0.04 μm (28) , which is below the Ki value of 0.16μ m for SU5416 in the biochemical Flk-1 kinase assay (29) .
A possible explanation for the ability of cells to concentrate SU5416 from extracellular medium may be the hydrophobic nature of the inhibitor, as indicated by its LogD value of >5. Thus, the inhibitor might be sequestered in the lipid membranes of the cell. Consistent with this hypothesis, preliminary cell fractionation experiments using HUVECs suggest that SU5416 is concentrated in the membrane fraction of the cells. From a reservoir in the cell membranes, the compound could partition into the cytosolic fraction of the cell so as to maintain inhibitory concentrations in the local environment of the membrane-associated Flk-1/KDR receptor kinase. It should be noted that the long-term inhibitory activity of SU5416 is specific to this compound; closely related, less hydrophobic compounds such as SU6668 (40) do not demonstrate long-lasting activity or concentration in similar studies.
An important result of these studies is that, although SU5416 has long-lasting inhibitory activity resulting from sequestration of the inhibitor in the cell, the specificity of the inhibitory activity is not affected. This is highlighted by the demonstration that the long-lasting effect remains selective for VEGF-dependent stimulation of Flk-1/KDR phosphorylation and intracellular signaling, with little or no effect on aFGF-stimulated effects consistent with the selectivity of the compound in biochemical kinase assays (28) . Thus, the fact that SU5416 has long-lasting inhibitory activity does not alter the selectivity of its inhibitory activity in vitro or in vivo.
In preclinical animal models, higher doses of SU5416 are required to achieve statistically significant efficacy with infrequent dosing (50 mg/kg once or twice weekly) than with daily dosing (25 mg/kg/day). Similarly, in HUVECs the IC50 of SU5416 in the mitogenic/proliferation assay is lower when the compound is present throughout the 48-h assay (IC50 = 40 nm) than when it is present only for a short period early in the assay (IC50 = 1 μm). Together, these data suggest that high systemic exposure to SU5416 for short periods of time is sufficient to confer a durable effect in vivo. Importantly, peak plasma levels of ∼17μ m have been measured at the end of the ∼1-h infusion period in patients receiving the recommended Phase II dose of SU5416 (145 mg/m2 ), and plasma levels above 5μ m have been detected for 2 h after the end of the infusion. Therefore, the data from this study provide a strong rationale for clinical regimens in which SU5416 is administered less frequently than once daily at 145 mg/m2 .
In summary, the data presented here indicate that short exposures to high concentrations of SU5416 (similar to those measured in the plasma of patients receiving the recommended Phase II dose of SU5416) were sufficient to inhibit the VEGF-dependent mitogenic response in HUVECs. These data also indicate that the in vitro inhibitory activity of SU5416 is durable, such that SU5416 caused a prolonged inhibition of Flk-1 phosphorylation (in 3T3/Flk-1 cells) and VEGF-dependent signaling (in HUVECs). A 24 h treatment of HUVECs with SU5416 did not, however, alter the number or affinity of VEGF receptors on the cell surface. Thus, we conclude that the long-lasting inhibition of VEGF-dependent activation of HUVECs by SU5416 is due to sustained inhibition of KDR phosphorylation secondary to the persistence of an inhibitory intracellular concentration of SU5416 for an extended period after compound has been removed from the medium.
We would like to thank Dr. Sara Courtneidge for helpful discussions during the course of these studies, and Drs. Courtneidge, Gerald McMahon, Beverly Smolich, and Peter Hirth for critical review of the manuscript. We would also like to acknowledge the contributions of Christina Mirabile and Sean Paxton in the preparation of the figures and manuscript.
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.
↵1 To whom requests for reprints should be addressed. at SUGEN, Inc., 230 East Grand Avenue, South San Francisco, CA 94080. Phone: (650) 553-8695; Fax: (650) 553-8308; E-mail:
↵2 The abbreviations used are: VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; aFGF, acidic FGF; Flk-1, fetal liver kinase-1; KDR, kinase insert domain-containing receptor; HUVEC, human umbilical vein endothelial cell; EGM, endothelial cell growth medium; EBM, endothelial cell basal medium; SRB, sulforhodamine B; ERK, extracellular signal-regulated kinase.
↵3 Unpublished observations.
↵4 C. Yang, personal communication.
- Received June 27, 2000.
- Revision received October 5, 2000.
- Accepted October 5, 2000.