Abstract
Phosphorylation of tyrosine residues on the epidermal growth factor (EGF) receptor (EGFr) is an important early event in signal transduction, leading to cell replication for major human carcinomas. CP-358,774 is a potent and selective inhibitor of the EGFr tyrosine kinase and produces selective inhibition of EGF-mediated tumor cell mitogenesis. To assess the pharmacodynamic aspects of EGFr inhibition, we devised an ex vivo enzyme-linked immunosorbent assay for quantification of EGFr-specific tyrosine phosphorylation in human tumor tissue specimens obtained from xenografts growing s.c. in athymic mice. When coupled with pharmacokinetic analyses, this measurement can be used to describe the extent and duration of kinase inhibition in vivo. CP-358,774 is an effective, orally active inhibitor of EGFr-specific tyrosine phosphorylation (ED50 = 10 mg/kg, single dose). It has a significant duration of action, producing, on average, a 70% reduction in EGFr-associated phosphotyrosine over a 24-h period after a single 100 mg/kg dose. Inhibition of EGFr phosphotyrosine in an ex vivo assay format effectively estimates the potency and degree of inhibition of EGFr-dependent human LICR-LON-HN5 head and neck carcinoma tumor growth. Substantial growth inhibition of human tumor xenografts was achieved with p.o. doses of the compound (ED50 = 10 mg/kg q.d. for 20 days). Combination chemotherapy with cisplatin produced a significant response above that of cisplatin alone with no detectable effects on body weight or lethal toxicity. Taken together, these observations suggest that CP-358,774 may be useful for the treatment of EGFr-driven human carcinomas.
For the majority of human carcinomas, growth factor receptors play an important role in tumorigenesis and progression to terminal disease states. The epidermal growth factor (EGF) receptor (EGFr) has been implicated in many human squamous cell carcinomas (Ozanne et al., 1986), such as non-small cell lung carcinoma and brain, bladder, breast, and ovarian carcinomas (Gullick, 1991). EGF at picomolar concentrations is mitogenic for cells overexpressing the receptor, and antibodies to EGFr abolish EGF-stimulated mitogenesis in LICR-LON-HN5 head and neck carcinoma (HN5; Modjtahedi et al., 1993b,c) and other tumor cells (Aboud-Pirak et al., 1988; Yoneda et al., 1991a). As an early event in the signal transduction process, the ligand transforming growth factor-α or EGF binds to EGFr on the surface of tumor cells and stimulates: 1) heterodimerization and homodimerization of EGFr molecules; 2) intermolecular cross-phosphorylation of intracytoplasmic tyrosine residues (EGFr autophosphorylation; Honegger et al., 1989); and 3) activation of the tyrosine kinase activity of EGFr. Apart from binding to the cognate ligand, all known functions of EGFr depend on tyrosine kinase activity. Point mutations in the kinase domain that abrogate ATP binding also abolish ligand-dependent kinase activity and abrogate EGF/transforming growth factor-α-induced mitogenesis (Moolenaar et al., 1988). An intact kinase domain is essential for activation of numerous downstream effectors, including phospholipase C-γ (Margolis et al., 1990; Nishibe et al., 1990; Wahl et al., 1990) phosphatidylinositol 3-kinase (Bjorge et al., 1990), and mitogen-activated protein kinase (Ahn et al., 1990), with the ultimate cellular response being DNA synthesis and cell division (Honegger et al., 1987). Transfection experiments have shown that EGFr overexpression alone may lead to constitutive activation of signal transduction, leading to uncontrolled mitosis (Di Fiore et al., 1987;Velu et al., 1987). The degree of EGFr overexpression has been shown to be related to tumorigenicity in some tumor systems (Santon et al., 1986; Velu, 1990). Recent studies of biopsy specimens suggest that overexpression of EGFr is associated with a poor prognosis in bladder (Neal et al., 1985) and breast (Sainsbury et al., 1985) carcinomas.
Despite homology with other tyrosine kinases, selective inhibitors have been identified (for a review, see Traxler, 1998). The EGFr tyrosine kinase therefore represents an attractive molecular target for pharmacological intervention. To monitor the effects of kinase inhibition, the degree of EGFr autophosphorylation was examined, because: 1) autophosphorylation of effector-specific tyrosine residues increases the velocity of the kinase reaction (Bertics and Gill, 1985); 2) autophosphorylation increases the affinity of the EGFr for its substrates, such as phospholipase C-γ (Magni et al., 1991), allowing these substrates to bind the activated receptor (docking site) and thereby become tyrosine phosphorylated; and 3) EGFr phosphotyrosine represents the last known biochemical event before committed steps toward cellular division are mediated by downstream effector mechanisms. For these reasons, we believe quantification of EGFr autophosphorylation is related to, and characterizes, inhibition of the kinase functionality.
CP-358,774 is a potent inhibitor of the EGFr tyrosine kinase with an IC50 value of 2 nM; CP-358,774 and its analogs have been shown to be direct-acting, reversible, ATP-competitive inhibitors of EGFr tyrosine phosphorylation (Moyer et al., 1997;Pustilnik et al., 1997). Specificity analysis has indicated >1000-fold selectivity against other tyrosine kinases, such as pp60v-src, pp145c-abl, the tyrosine kinase activities of the insulin and the insulin-like growth factor-1 receptors; selectivity has been shown against isolated kinases as well as in intact cells (Moyer et al., 1997). CP-358,774 inhibits autophosphorylation of the EGFr in a variety of EGFr-overexpressing tumor cells (IC50 = 20 nM) and produces inhibition of mitogenesis, inhibition of tumor cell division, and cell cycle arrest. In some cell types, such as DiFi, CP-358,774 induces concentration-dependent apoptosis in vitro.
Here, we report that CP-358,774 is an effective, orally active inhibitor of EGFr tyrosine autophosphorylation. CP-358,774 can effectively inhibit EGFr tyrosine phosphorylation in human tumors growing s.c. in athymic mice with an ED50 value of 10 mg/kg p.o. It has significant duration of action and produces substantial inhibition of human EGFr-overexpressing tumors growing s.c. in athymic mice. Moreover, the degree of inhibition of EGFr phosphotyrosine shows good agreement with the degree of tumor growth inhibition in treated animals. The results of these experiments were previously reported at the American Association for Cancer Research annual meeting (Pollack et al., 1997). The data suggest that CP-358,774 may be a useful new compound for therapy of human neoplastic diseases.
Materials and Methods
Mice.
Three- to 4-week-old female athymic mice (CD-1nu/nu) were used for human tumor xenografts. Mice were obtained from Charles River Laboratories (Wilmington, MA) and were housed in specific pathogen-free conditions, according to the guidelines of the American Association for Laboratory Animal Care; all studies were carried out with approved institutional experimental animal care and use protocols. During these studies, animals were provided pelleted food and water ad libitum and kept in a room conditioned at 70–75°C and 50 to 60% relative humidity with >15 fresh air changes per hour. Sentinel heterozygous littermates of the athymic animals were monitored routinely (3-week intervals) by serological assays and were found to be free of exposure to the following agents: murine hepatitis virus, ectromelia virus, and Sendai virus. For all studies, the mice were allowed to acclimate for 1 to 3 days after receipt of shipment; test animals were randomized before commencement of treatments and examined twice daily thereafter for compound-induced or tumor-related deaths. Moribund animals were sacrificed to reduce suffering.
Tumor Cell Lines.
The HN5 cells were obtained from Dr. M. J. O’Hare (Haddow Labs., Institute of Cancer Research, Sutton, Surrey, UK). All other cells were purchased from the American Type Culture Collection (Rockland, MD). All cell lines were free of reovirus type 3, pneumonia virus of mice, K-virus, Theiler’s virus, Sendai virus, ectromelia virus, and lactate dehydrogenase virus (Microbiological Associates, Bethesda, MD).
Cell Culture.
Cell lines were passaged by monolayer culture in 175-cm2 tissue culture flasks (Nunclon; Marsh Biomedical Products, Rochester, NY) in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated FBS (Hazelton Research Products, Inc., Lenexa, KS), 300 μg/ml glutamine, 100 U/ml penicillin G, 100 μg/ml streptomycin, and 10 μg/ml gentamycin at 37°C in a humidified 95% air/5% CO2 atmosphere. Routine periodic samples of cell culture broths tested negative forMycoplasma contamination (Microbiological Associates). For implantation in vivo, the tumor cells were harvested from exponentially growing cultures (60–80% confluence), detached by light trypsinization (0.25% trypsin and 0.02% EDTA, 1 min), washed in Hanks’ balanced salt solution (HBSS), resuspended in HBSS, mixed with the basement membrane preparation Matrigel (40234; Collaborative Biomedical Products, Bedford, MA), and held in an ice bath <1 h before injection.
Chemotherapeutants.
CP-358,774 [6,7-bis(2-methoxy-ethoxy)quinazoline-4-yl]-(3-ethynylphenyl)amine; MF = C22H23N3O4), a colorless, crystalline, anhydrous compound, was synthesized in our laboratories (Arnold and Schnur, 1998). In these studies, the hydrochloride salt (molecular weight = 429.9) was used in all cases, except for that represented in Fig. 7, which used the free base (molecular weight = 393.4), and the dosage levels shown represent the quantity of free base administered, excluding the contribution of the salt. The compound was formulated for i.p. or p.o. administration by dissolution of the dry powder in a small amount (10% of final volume) of dimethyl sulfoxide (DMSO), mixed by vortexing until dissolved; during vortexing, sufficient sterile, pyrogen-free physiological saline (0.15 N NaCl), containing 0.10% (w/v) Pluronic P105 (BASF Wyandotte, Parsippany, NJ), was added to produce a homogeneous fine suspension. The prepared dosage forms did not produce microbial colonies after incubation on brain-heart infusion agar and did not contain endotoxin detectable by the Limulusamoebocyte lysate assay (Associates of Cape Cod, Inc., Woods Hole, MA). Doxorubicin (Adriamycin; Rapid Dissolution Formula) was purchased from Adria Labs. (Columbus, OH). Cisplatin was obtained as a powder from Sigma Chemical Co. (St. Louis, MO). All dosage forms were freshly prepared for each day’s treatment. CP-358,774 and the reference agents were dosed according to the optimum formulation, route, and regimens, as empirically derived in previous studies; aggressive dosing parameters (single bolus treatments at maximum tolerated dosage levels) were used for maximum antitumor efficacy of the cytoreductive agents. Test animals were treated between 7 and 9 AM, immediately after a 12-h dark photoperiod (active phase), to control for variability introduced by circadian physiological cycles, according to the methods of Halberg et al. (1973).
EGFr Phosphotyrosine Determinations by Enzyme-Linked Immunosorbent Assay (ELISA).
To determine compound-induced inhibition of EGFr-associated tyrosine phosphorylation in human tumor explants from athymic mice, an ELISA specific for EGFr phosphotyrosine was developed. Tumor tissue was harvested at various times after dosing (usually 1 h) by careful dissection, immediately flash frozen in liquid nitrogen, and then homogenized in buffer formulated to prevent further tyrosine phosphorylation as well as enzymatic phosphatase activity. A double-antibody ELISA provided quantitative determinations of the degree of EGFr tyrosine phosphorylation after specific capture of EGFr protein.
Briefly, athymic mice with s.c. tumors (5–10 mm in diameter) were euthanized humanely, and tumors were excised with the use of small dissecting scissors and mosquito forceps, after which the tumor tissue was immediately flash frozen in liquid nitrogen and stored at −70°C before homogenization and immunoassay. Tumors were weighed, and for each 100 mg of tumor tissue, 1 ml of ice-cold, sterile lysis buffer was added. Lysis buffer contained (per liter) 50 ml of 1 M HEPES, pH 7.4, buffer, 37.5 ml of 4 M sodium chloride, 0.75 ml of 2 M magnesium chloride, 10 ml of 100 mM EDTA, 10 ml of glycerol, 10 ml of Triton X-100, 8 ml of 200 mM sodium orthovanadate, 4.2 g of sodium fluoride, 50 μg/ml phenylmethylsulfonyl fluoride, 25 mg of soybean trypsin inhibitor, 10 μg/ml leupeptin, and 10 μg/ml aprotinin. Tumors were homogenized with a Thomas Teflon pestle homogenizer attached to a power drill (or equivalent) and then clarified by centrifugation; the resulting supernatant liquid (800 μl) was transferred to microtiter plates in 200-μl aliquots and maintained at −70°C before assay.
Appropriate dilutions of tumor homogenates (1:20–1:40 dilutions) were made in blocking buffer containing (per liter) 50 g of bovine serum albumin, 10 g of ovalbumin, 0.90% NaCl, and 10 mM Tris · HCl buffer, pH 7.4. After dilution, 100-μl aliquots were transferred to microtiter wells containing adsorbed monoclonal antibody to EGFr protein (QIA08; Oncogene Science, Uniondale, NY). The plates were then incubated for 30 min at 30°C (or 3 h at room temperature) to allow efficient capture of the EGFr protein from the tumor homogenates. Microtiter wells were washed six times in a 1:10 dilution of Plate Wash Concentrate (PN 77 0550; DuPont NEN, Boston, MA). To detect phosphotyrosine residues, 100 μl of horseradish peroxidase-conjugated monoclonal antibody specific for phosphotyrosine (diluted 1:1000 in blocking buffer) was added to each well (PY54 conjugate, PT03; Oncogene Science), and plates were incubated for 1 h at 30°C. Microtiter wells were then washed six times in a 1:10 dilution of Plate Wash Concentrate, after which 100 μl/well of 3,3′,5,5′-tetramethylbenzidine substrate was added (50-76-04; Kirkegaard and Perry Laboratories, Gaithersburg, MD); color development was monitored over 30 min, after which all reactions were stopped with 100 μl/well of 0.09 M sulfuric acid. For quantification, absorbance was determined at 450 nm with a Bio-Rad (Hercules, CA) model 3550 microplate reader. EGFr phosphotyrosine content was calculated after normalization of each sample for total protein with a commercial kit (BCA Protein; Pierce, Rockford, IL).
The absorbance values for samples from each of the tumor-bearing animals (sample size, four mice/treatment group) were entered into a custom Microsoft Excel spreadsheet, where the endpoints (i.e., protein concentrations and phosphotyrosine levels) were calculated. In all cases, the EGFr-associated tyrosine phosphorylation was expressed as absorbance units/mg total protein. For statistical inferences, the relationships between groups (i.e., test versus control group) were identified using a computer program for the one-way ANOVA, where the α significance level was assigned at 0.05. P values were determined using Dunnett’s t statistic. A set of internal laboratory standards (i.e., aliquots from previously frozen tissue for both treated and control groups) was used to assess the quality and reproducibility of the immunoassay; in the course of 5 years’ routine testing, the results were highly reproducible (i.e., the coefficient of variation was <6.0%).
HPLC Determinations of CP-358,774 in Plasma and Tumor Tissue.
Determination of drug concentration was made by organic extraction (acetonitrile) of plasma and tumor samples, followed by HPLC. CP-358,774 in plasma and tumor tissues was extracted from 200-μl samples spiked with 100 μl of internal standard (CP-292,597; 0.8 ng/μl in acetonitrile) with 5 ml of methyl t-butyl ether using an Oberbach reciprocating shaker for 10 min. Before extraction, tumor tissue was homogenized in 4 parts deionized water to 1 part tumor specimen (v/m) using an Omni 2000 (Omni International, Gainesville, VA) tissue homogenizer. Samples were centrifuged at 3000 rpm for 5 min at 22°C using a Jouan centrifuge. The organic layer of each sample was transferred to a clean tube, and the methylt-butyl ether was evaporated to dryness in a Zymark Turbo-Vap at 60°C. All samples were reconstituted in 200 μl of mobile phase consisting of 70% water and 30% acetonitrile (v/v) brought to pH 2.4 with trifluoroacetic anhydride (Acros Organics). A 2-liter volume of mobile phase consisted of 1400 ml of Milli-Q deionized water, 600 ml of acetonitrile, and 550 μl of trifluoroacetic anhydride. The analytical column was a YMC Basic C-18 (4.6 mm × 150 mm, 3 μm). A pump (Thermo Separation Products Constametric 4100) was used to establish a 1.5 ml/min flow rate through the column. CP-358,774 was detected at 345 nm (AUFS 0.001) using an ultraviolet detector (Milton Roy Spectro Monitor 3100 variable wavelength detector). The retention time for CP-358,774 was 6.5. The lower limit of quantification of the assay was 10 ng/ml for plasma and 50 ng/g for tumor tissue.
Tumor Growth Inhibition Studies In Vivo.
The tumor growth inhibitory effects of CP-358,774 were measured in young athymic mice bearing established, palpable (2–4-mm diameter) human HN5 or A431 tumors. Tumors were induced in the left flank of 3- to 4-week old athymic mice by s.c. injection of 1 × 106 cultured, log phase HN5 or A431 cells in 0.20 ml of HBSS containing 50% Matrigel. Tumor size was measured in millimeters with Vernier calipers across two diameters three times/week, and the tumor volume (mm3) was calculated using the formula: tumor volume = (length × [width]2)/2, according to standard methods (Geran et al., 1972); results are expressed as tumor volume (TuV) in mm3. To calculate tumor growth inhibition, the following formula was used: inhibition (%) = (TuGcontrol − TuGtest)/TuGcontrol × 100%, where tumor growth (TuG) equals the final tumor size minus the pretreatment tumor size for individual treatment groups. This method of tumor implantation provided reproducible growth in athymic mice, enabling the determination of dose-response effects for a variety of chemotherapeutic agents. For each experiment, athymic mice were randomized on receipt of shipment and again after tumor implantation (i.e., before commencement of treatment). Data collected from the antitumor studies (e.g., tumor volume) were evaluated for statistical significance using one-way ANOVA (for significant antitumor activity,P < .05).
Results
Inhibition of EGFr Phosphotyrosine in HN5 Xenografts.
HN5 possesses many of the characteristics of EGFr-dependent squamous cell carcinomas both in vitro (Modjtahedi et al., 1993b,c) and in vivo (Modjtahedi et al., 1993a,b). In particular, monoclonal antibodies directed at the EGFr can completely block cellular proliferation in vitro, and for this reason, the tumor cell line was selected to evaluate a large series of EGFr tyrosine kinase inhibitors. When administered orally (by gavage) or parenterally (i.p.), CP-358,774 consistently produced significant, dose-related inhibition of HN5 EGFr tyrosine phosphorylation 1 h after dosing (Fig.1). Compared with vehicle-treated controls, a maximum of 80% reduction in phosphotyrosine was observed after dosing by p.o. or i.p. routes. In several preliminary experiments, the vehicle (10% DMSO, 0.85% NaCl, and 0.10% Pluronic P105) produced no inhibition of EGFr phosphotyrosine compared with water or saline treatments.
The data in Fig. 1 are a compilation of 28 (p.o.) and 22 (i.p.) independent experiments, attesting to the reproducible inhibition by this agent. The effective dose for 50% inhibition of the target receptor (ED50) was similar for p.o. and i.p. administration: 9.9 mg/kg p.o. and 9.2 mg/kg i.p. The minimal effective single dose eliciting statistically significant (P < .05) reduction in EGFr-associated phosphotyrosine was 5.5 mg/kg (39% reduction) and 2.8 mg/kg (47% reduction) for the p.o. and i.p. routes, respectively. These extrapolated ED50 values are within one order of magnitude of the IC50 value (20 nM) for the inhibition of EGFr phosphotyrosine by CP-358,774 in homogeneous populations of HN5 cells growing in vitro.
Relationship of Plasma Concentration to EGFr Phosphotyrosine Inhibition.
Figure 2 illustrates the relationship of plasma concentration of CP-358,774 to inhibition of EGFr-associated phosphotyrosine of HN5 xenografts. At 1 h post-treatment with a single dose, plasma concentrations of 2 to 10 μM CP-358,774 (0.79–3.9 μg/ml) were associated with a significant reduction in EGFr phosphotyrosine (∼40% reduction relative to vehicle-treated controls). At higher plasma concentrations (i.e., 10–100 μM, 3.9–39 μg/ml), the reduction in EGFr phosphotyrosine ranged from 65 to 75%. By interpolation, the effective plasma concentration for 50% inhibition of the target receptor was estimated at 8 μM (3.1 μg/ml) and ∼12 μM (4.7 μg/ml) for p.o. and i.p. dosing, respectively. In mouse plasma, 95% of CP-358,774 is bound to plasma proteins. Taking these data into account, at 1 h after the dose, 50% inhibition of EGFr-associated phosphotyrosine of HN5 tumors occurred at free plasma concentrations of 400 nM (160 ng/ml) for p.o. and 600 nM (240 ng/ml) for i.p. doses of CP-358,774.
Duration of Action of CP-358,774.
The duration of reduction in EGFr phosphotyrosine after a single 92 mg/kg dose of CP-358,774 was evaluated in the HN5 model (Fig. 3). After p.o. dosing, significant and substantial inhibition of EGFr phosphotyrosine (75–85%) was observed for 12 h; reduction was still measurable (25–40%), and statistically significant, after 24 h. To a similar degree of efficacy, parenterally (i.p.) dosed mice showed substantial inhibition of EGFr phosphotyrosine for 12 h; however, no reduction was observed at 24 h (data not shown). Calculation of the area under the curve for reduction in EGFr phosphotyrosine provides an estimation of the overall degree of inhibition over a 24-h period. Based on the assumption that complete inhibition (100%) of EGFr autophosphorylation over a 24-h period would produce inhibition of 2400%-h (100% “coverage”), p.o. dosing elicited inhibition of 1690%-h (70.4% coverage), whereas parenteral dosing (i.p.) showed an area under the curve of 1420%-h (59.0% coverage).
Within 1 h after p.o. dosing, peak plasma and tumor CP-358,774 concentrations were reached (124.6 μM and 54.8 μmol/kg, respectively). Plasma and tumor concentrations then declined rapidly until 6 h, after which concentrations remained low, although detectable, for several hours. The terminal elimination half-life in plasma after p.o. administration could not be determined because plasma concentrations from 6 to 24 h did not significantly decline; the mean tumor half-life after p.o. administration was estimated at 2.9 h. At 24 h after the dose, plasma and tumor concentrations were 38 μM and 4.0 μmol/kg, respectively. It is clear from Fig. 3 that although plasma and tumor CP-358,774 concentrations follow similar time courses, the EGFr-associated phosphotyrosine reduction does not decline with declining plasma and tumor levels and remains at a high level (80% inhibition) at 12 h after the dose. The reason for this is unclear but seems to be a consistent observation for this compound and related analogs.
To determine whether inhibition of EGFr tyrosine phosphorylation could lead to decreased expression or increased turnover of the surface-bound receptor, tumor homogenates were assayed for EGFr protein using a commercial kit (Oncogene Science). In several experiments, EGFr protein concentrations did not change within 24 h of a single dose of CP-358,774 or within 1 h in multiply dosed animals (n = 20 consecutive daily doses; data not shown). Moreover, it appeared from our data that in vivo receptor density remained relatively constant at 9.4 fmol/μg total protein among several experiments. Although we cannot conclude that transient changes in receptor density did not occur in these animals, it is apparent that a prolonged drug-induced reduction in EGFr-associated tyrosine phosphorylation could not be explained by concurrent reductions in receptor density. Similarly, tumor tissue concentrations of CP-358,774 correlated well with plasma concentrations. In Fig. 3, the use of athymic mice bearing bilateral tumors allowed the simultaneous measurements of EGFr phosphotyrosine, tumor tissue drug concentration, and plasma drug concentrations. On average, the peak tumor tissue concentration was 44 and 29% of the peak circulating plasma levels for p.o. and i.p. dosing, respectively, and CP-358,774 was not retained in tumors relative to plasma (Fig. 3). In tissue culture, the effects of CP-358,774 on EGFr phosphotyrosine in HN5 cells is freely reversible. Cellular EGFr phosphotyrosine levels return to levels found in untreated cells within minutes of removal of CP-358,774 from the culture medium. Thus there is no clear pharmacodynamic explanation for a persistent inhibitory effect of CP-358,774 in tumors growing in vivo.
Antitumor Effects of CP-358,774 (HN5).
The antitumor effects of CP-358,774 were determined in the EGFr-overexpressing human HN5 and human A431 epidermoid carcinomas. Both tumors have been shown to be inhibited by monospecific anti-EGFr antibodies in cell culture and in xenograft models (Fan et al., 1993; Modjtahedi et al., 1993a). Oral administration of CP-358,774 produced significant dose-related antitumor effects against established HN5 growing s.c. in athymic mice (Fig. 4). When test animals were dosed for 20 consecutive days beginning at 4 days after tumor implantation (tumor diameter, 2–4 mm), the minimal effective dose for significant antitumor effects was 5.7 mg/kg/day p.o., using one-way ANOVA (P < .05 with Dunnett’s test). Doses of 11 to 92 mg/kg/day p.o. produced substantial antitumor effects (i.e., >50% inhibition). During the course of dosing (days 4–23 after implantation), tumor-bearing mice treated with vehicle alone showed progressive enlargement of tumors; spontaneous regressions in vehicle-treated or untreated animals have not been observed in this model.
In the experiment described above, tumor sizes for CP-358,774-treated animals were significantly reduced, and this inhibitory effect was observed as long as the test animals were being treated. On the cessation of treatment, we have found that although tumors gradually enlarge, tumor growth rates do not generally equal those of the vehicle controls. Using the tumor size measurements taken only during the treatment period to calculate tumor growth inhibition as a percent of the controls, the aggregate therapeutic effects of the compound, relative to vehicle-treated controls, could be assessed (Fig.5). CP-358,774 produced dose-related inhibition of the s.c. growth of the human HN5 with an effective dosage for 50% inhibition (ED50) of 7 mg/kg/day p.o. Tumor stasis (100% tumor growth inhibition) was achieved with as little as 12.5 mg/kg/day; values of >100% inhibition indicate partial regression. The appearance of treated animals indicated that CP-358,774 was well tolerated, that no drug-induced fatalities were recorded, and that no losses in body weights were observed in several experiments. In subsequent studies, on this schedule and at dosages of 400 mg/kg/day, no deaths were observed, but deaths of athymic mice were observed at a dosage of 800 mg/kg/day after 8 days of dosing. Preliminary toxicological studies indicated that repetitive dosing in mice did not produce apparent toxicity or histological abnormalities in organs with high EGFr densities (i.e., liver).
We then compared the dose-response relationship for the pharmacodynamic effects (EGFr phosphotyrosine reduction in tumors 1 h after a single dose) and therapeutic effects (i.e., tumor growth inhibition during 20 days of consecutive daily dosing). In Fig.6, there was good correlation between these two endpoints (r2 = 0.92), which is strongly suggestive that the inhibition of HN5 tumor growth is related to the inhibitory effects of CP-358,774 on EGFr function.
To evaluate the therapeutic effectiveness of CP-358,774 on large, well established tumors, HN5 implants were allowed to grow to at least 1 cm in diameter (∼525 mm3) before the initiation of therapy. Treatment with CP-358,774 produced an immediate halt to tumor enlargement at all doses ≥11 mg/kg, and some tumor growth inhibition, although not as pronounced, was evident at a dose as low as 5.7 mg/kg (Fig. 7). Tumors in treated animals (11 mg/kg) appeared to be in stasis during the treatment or to slightly decrease in size during the dosing period. In addition, the static tumor profile appeared to extend beyond the treatment period, such that tumor sizes for the treated animals did not exceed pretreatment levels until at least day 60, which was 33 days after the cessation of treatment. Although tumors resumed growth at days 60 to 70, subsequent growth did not occur at the same rate as the controls. These data, taken together, suggest that treatment with CP-358,774 at effective dosage levels produces tumor stasis up to, and beyond, the dosing period and suggest that effective control of established tumors may be expected from chronic treatment regimens. In this and several other experiments, treatment with the maximally tolerated dose of doxorubicin had little therapeutic effect on large, well-established HN5 xenografts (Fig. 7), although tumor growth inhibition could readily be demonstrated with small tumor masses (tumor diameter, 2–4 mm; Fig. 8).
The p.o. administration of CP-358,774 also inhibited the growth of established human A431 carcinomas in athymic mice (Fig.8). The minimal effective dosage for significant antitumor effects was 6.25 mg/kg/day p.o. Dosages of 12.5 to 100 mg/kg/day p.o. produced dose-related inhibition with an observed ED50 value of 14 mg/kg/day p.o. Tumor growth was nearly completely inhibited with an apparent reduction in tumor size in animals dosed with 50 to 100 mg/kg/day (partial regression); however, after cessation of dosing, all tumors regrew progressively. Doxorubicin (Adriamycin), at its maximum tolerated dose for a single administration (15 mg/kg i.v.), was less effective than CP-358,774 at 50 mg/kg p.o. in this model.
Combinations of CP-358,774 and Cisplatin.
Current chemotherapy regimens use multiple agents for the treatment of carcinomas to obtain improved efficacy and avoid emergence of resistance. Combined therapy of experimental tumors with anti-EGFr antibodies and cytotoxic therapy can produce greater antitumor effects than that use of either modality alone (Baselga et al., 1993; Fan et al., 1993). We therefore examined the effects of CP-358,774 treatment in combination with a conventional cytotoxic drug (i.e., cisplatin) to characterize both the toxicity and efficacy of combination regimens.
In the HN5 model, the antitumor effects of cisplatin alone and in conjunction with CP-358,774 are shown in Fig.9. For this study, the cytoreductive agent cisplatin was dosed at a level (maximum tolerated dosage, 10 mg/kg i.v.) slightly below that which was lethally toxic to animals in preliminary studies, whereas CP-358,774 was dosed at the ED50 dose, which was considerably lower than the toxic dose via this route of administration. Regardless of the sequence of dosing of the two agents (e.g., A then B, A plus B, or B then A), there were no drug-induced deaths observed in this experiment. Furthermore, animals dosed with combinations of cisplatin and CP-358,774 did not show physical signs of toxicity on gross examination (i.e., ataxia, apparent weight loss, hypothermia, and so on).
As a single agent, cisplatin produced significant inhibition of tumor growth in all animals tested. A maximum of 40% inhibition occurred between 6 and 10 days after the dose, and these results were comparable to those obtained in previous studies with this agent; no regressions, however, occurred through the use of cisplatin. The inhibitory effects of cisplatin decrease as treatment is applied to larger tumors [i.e., less efficacy was observed as treatments were begun on days 4, 7, or 10 (top, middle, and bottom), respectively].
The growth inhibitory effects of CP-358,774 (9 mg/kg p.o. b.i.d. for 5 days) are also shown in Fig. 9. Antitumor effects of CP-358,774 were observed at ∼5 days after the commencement of dosing. After dosing, tumors grew progressively until the experiment was terminated; CP-358,774 does not lead to regressions under these dosing conditions in this tumor system.
The interactions of cisplatin and CP-358,774 for inhibition of HN5 tumor growth are shown in Fig. 9. To examine interactions between agents with dissimilar mechanisms of action, the sequence of compound administration was considered potentially important. Regardless of the dosing sequence, however, significant and substantial antitumor effects were observed (65–75% inhibition) whether the cytoreductive agent was dosed before CP-358,774 (top), concurrently with CP-358,774 (middle), or after CP-358,774 (bottom). In each case, the combination of cytoreductive agent and EGFr inhibitor showed greater antitumor effects over a more extended period of time than did the cytoreductive agent alone (additive effects). CP-358,774 may be administered concurrently with cisplatin with no apparent antagonism in drug action for the cytoreductive agent in this model. In addition, we observed no increased lethal cisplatin-induced toxicity or enhanced weight loss with these combinations of CP-358,774 and cisplatin.
Discussion
CP-358,774 is a potent, selective inhibitor of the tyrosine kinase of the human EGFr that blocks tumor cell division, produces cell cycle arrest, and initiates programmed cell death in EGFr-overexpressing human tumor cells (Moyer et al., 1997). To examine the therapeutic effects of this inhibitor, we selected a human tumor system (HN5) that is dependent on EGFr for mitogenic signals (Modjtahedi et al., 1993b,c). In parallel, we measured EGFr tyrosine phosphorylation to assess inhibition of the kinase activity in situ because the functions of this receptor in signal transduction require kinase activity. EGFr autophosphorylation represents the last known biochemical event before committed steps toward cell division are mediated by numerous downstream effectors.
CP-358,774 produces potent reduction in EGFr tyrosine phosphorylation within 1 h of dosing (ED50 = 9.9 mg/kg p.o.). The plasma CP-358,774 concentration associated with this effect was 8 μM (3.1 μg/ml); because 95% of CP-358,774 is bound to mouse plasma proteins, we calculated that 50% inhibition of EGFr phosphotyrosine occurred at free plasma concentrations of 400 nM (160 ng/ml) for p.o. doses of CP-358,774. Substantial reduction in EGFr phosphotyrosine was observed for at least 12 h after the dose (100 mg/kg p.o.), and significant reduction was noted at 24 h. The antitumor effects of CP-358,774 correlated well with inhibition of tumor-associated EGFr phosphotyrosine.
These results are comparable to those of Kunkel et al. (1996), who used Western blots to show 80 to 90% reduction in EGFr phosphotyrosine in A431 tumors as early as 15 min after i.p. administration of the EGFr inhibitor PD153035. Our findings differ in that we observed a longer duration of action for CP-358,774, whereas in the study of Kunkel et al. (1996), tyrosine phosphorylation returned to baseline after 3 h. At a dose of 40 mg/kg i.p., PD153035 produced a reduction in EGFr phosphotyrosine but no measurable changes in tumor growth when administered b.i.d from days 5 to 13 after s.c. implantation of A431. The importance of duration of receptor modulation to antitumor effects is underscored in data on a related compound. PD168393, an irreversible EGFr/erbB-2 inhibitor, suppressed EGFr phosphotyrosine for at least 24 h, whereas suppression by the reversible inhibitor PD174265 was short-lived (i.e., <8 h); only PD168393 showed antitumor activity (Fry et al., 1998). In our work, we were able to show substantial antitumor effects for the HN5 and A431 carcinomas at doses that can reduce EGFr autophosphorylation in extracted xenografts. Despite a number of EGFr-targeted therapeutants, little attention has been paid to assessment of receptor modulation in the growing tumor until very recently (Pollack et al., 1997; Vincent et al., 1998). Our data suggest that inhibition of EGFr, as evidenced by monitoring EGFr phosphotyrosine in situ, is directly related in this system to inhibition of established tumor growth (Fig. 6).
Although the benefit of an EGFr inhibitor has yet to be proved clinically, EGFr-specific therapeutants produce considerable inhibition of human xenografts in athymic mice. Monoclonal antibodies were shown by Masui et al. (1984) to prevent tumor growth in vivo. Inhibition by EGFr-specific monoclonal antibodies has been shown subsequently with a variety of therapeutic endpoints, including tumor growth inhibition, regression of tumor implants, reduced metastases, and increased lifespan (Aboud-Pirak et al., 1988; Mueller et al., 1991; Yoneda et al., 1991a; Fan et al., 1993a,b; Schnurch et al., 1994; Prewett et al., 1996). Striking antitumor effects were shown by Modjtahedi et al. (1993a,b; Dean et al., 1994) with rat monoclonal antibodies ICR62 and ICR16 and, very recently, by Yang et al. (1999) using humanized monoclonal antibody E7.6.3. These antibodies induced complete regressions of human xenografts if treatment was initiated on the day of tumor implantation and induced significant regressions in large (>1 cm) tumors.
Despite differences in methodologies and endpoints, our results are similar to those with monoclonal antibodies in the degree and timing of treatment effectiveness. The responses of EGFr-overexpressing HN5 and A431 carcinomas to treatment with CP-358,774 involved a marked cessation of tumor growth detected shortly after commencement of dosing. This was observed repeatedly in experiments involving 20 doses or as few as 5 daily doses (data not shown) and is a consistent finding with EGFr-targeted monoclonal antibodies (Masui et al., 1986; Yoneda et al., 1991a; Baselga et al., 1993; Modjtahedi et al., 1993a,b; Schnurch et al., 1994; Ciardiello et al., 1996; Yang et al., 1999). In addition, the responses are clearly potent (ED50 = 10 mg/kg p.o.), dose dependent, and substantial (i.e., >50% inhibition). CP-358,774 shares another pharmacological feature common to monoclonal antibodies, one that distinguishes these forms of therapies from conventional cytoreductive agents. When very large tumors were used (1-cm diameter), CP-358,774 produced stasis of tumor growth and a noticeable reduction in the size of treated tumors (i.e., partial regression). This is a confirmed observation for the monoclonal antibodies (Baselga et al., 1991; Yoneda et al., 1991a; Prewett et al., 1996); with doxorubicin, however, treatment is very effective when small tumors are treated (100–200 mm3; Fig.8) but is much less effective with large tumor masses (i.e., 500-1000 mm3; Fig. 7). Unlike monoclonal antibodies (Rodeck et al., 1987; Modjtahedi et al., 1993a,b; Dean et al., 1994), we have not seen complete regression in any of several xenograft therapy experiments to date; when treatment is stopped, tumors resume growth, although the growth rate appears somewhat less than that of the vehicle-treated controls.
Distinct differences exist between CP-358,774 and monoclonal antibodies as pharmacological agents, and these may emanate from differences in their respective mechanisms of action. Although CP-358,774 targets the intracellular domain of EGFr (Moyer et al., 1997), monoclonal antibodies act primarily as a blockade of the external EGF-binding domain. Immunoglobulins, however, may rely in part on recruitment of host effector mechanisms (immune cytolysis and antibody-dependent cellular cytotoxicity). Significant antitumor effects have been shown with Ig F(ab′)2 fragments by some investigators (Aboud-Pirak et al., 1988; Fan et al., 1993b), although not by others (Mueller et al., 1991), but there are studies that suggest monoclonal antibodies derive at least some therapeutic effects from immune recruitment (Herlyn and Koprowski, 1982; Bier et al., 1998). The cessation of tumor growth (i.e., stasis) is consistent with inhibition of EGFr signal transduction and with inhibition of subsequent cellular proliferation in vivo, but the contribution of immune mechanisms to antibody-induced regressions requires further clarification. In addition, the incubation of DiFi cells with CP-358,774 induces apoptotic cell death (Moyer et al., 1997), and the extent to which CP-358,774 induces apoptosis in vivo in the HN5 is currently under investigation.
The antitumor effects of CP-358,774 in conjunction with conventional cytoreductive agents (i.e., cisplatin) were similar to those observed for monoclonal antibody/cytoreductive combinations. Cisplatin and monoclonal antibodies have been shown to have additive and, in some cases, synergistic (supra-additive) antitumor effects (Aboud-Pirak et al., 1988; Fan et al., 1993a,b; Prewett et al., 1996). Additive effects of monoclonal antibodies have also been demonstrated for the tyrphostins (Yoneda et al., 1991b) and doxorubicin (Baselga et al., 1993; Baselga and Mendelsohn, 1994). Our observations with CP-358,774 and cisplatin indicated additive antitumor effects regardless of whether CP-358,774 was administered before, during, or after treatment with the cytoreductive; that is, we found significant differences in tumor growth inhibition for the combination compared with either agent alone. We did not observe supra-additive effects with this or other conventional chemotherapeutants, and the experiments were conducted under conditions that would have allowed synergistic responses to be detected. The major goal of the drug interaction studies, however, was to identify incompatibilities, of which there were none detected for either antitumor effects or toxicity endpoints.
Other small molecules have been reported to inhibit the EGFr. The tyrphostin RG-13022 inhibits EGFr autophosphorylation and EGFr-dependent tumor cell proliferation. At 400 μg/mouse/day, RG-13022 significantly inhibited MH-85 xenografts and produced increased lifespan (Yoneda et al., 1991b) but had little activity against HN5, albeit with treatment schedules and endpoints that differed somewhat from the original work (McLeod et al., 1996). Naamidine A, an alkaloid purified from a FijianLeucetta sp. sponge, showed potent antitumor effects against the A431 carcinoma with an ED50 value of ∼12.5 mg/kg/day when treatment was initiated 1 day after implantation (Copp et al., 1998). In addition, impressive tumor growth inhibition, stasis, and, in some models, regressions were achieved with the anilinoquinazoline ZD1839 against an array of human carcinoma xenografts; this compound has been advanced to clinical development (Woodburn et al., 1997). Although the treatment parameters and endpoints may differ for the EGFr inhibitors, these reports clearly indicate that the EGFr represents a potentially important molecular target for cancer therapy.
In conclusion, we found that CP-358,774 is a very potent and selective inhibitor of the EGFr tyrosine kinase and is capable of inhibiting EGFr tyrosine phosphorylation in human tumors in athymic mice after p.o. or i.p. administration. Most importantly, we observed that in tumor growth inhibition studies, the degree of inhibition of EGFr phosphotyrosine is clearly correlated with significant and substantial tumor growth inhibition. For these reasons, we believe that CP-358,774 may be useful for the therapy of EGFr-dependent human carcinomas.
Acknowledgments
We gratefully acknowledge the efforts of Drs. D. S. Salsburg and A. C. Swindell in biostatistical analyses and experimental design. In addition, we greatly appreciate the advice and support of Drs. Alan R. Proctor and Kelvin Cooper. Last, we would like to recognize the creativity and commitment of the late Rodney C. Schnur.
Footnotes
-
Send reprint requests to: Dr. Vincent A. Pollack, Department of Genomics, Targets and Cancer Research, Pfizer Central Research, Eastern Point Road, Groton, CT 06340. E-mail:vincent_a_pollack{at}groton.pfizer.com
-
↵1 Portions of this work were presented at the annual meeting of the American Association for Cancer Research, April 1997.
-
↵2 Present address: Department of Chemistry, BASF Bioresearch Corp., 100 Research Dr., Worcester, MA 01605-4314.
-
↵3 Address: OSI Pharmaceuticals, Inc., 106 Charles Lindbergh Blvd., Uniondale, NY 11553.
- Abbreviations:
- EGF
- epidermal growth factor
- EGFr
- epidermal growth factor receptor
- ELISA
- enzyme-linked immunosorbent assay
- HBSS
- Hanks’ balanced salt solution
- HN5
- LICR-LON-HN5 head and neck carcinoma
- Received May 18, 1999.
- Accepted August 5, 1999.
- The American Society for Pharmacology and Experimental Therapeutics