Effect of an Epidermal Growth Factor Receptor Inhibitor in Mouse Models of Lung Cancer

Gefitinib (Iressa, ZD1839) is a potent high-affinity competitive tyrosine kinase inhibitor aimed primarily at epidermal growth factor receptor (EGFR). Inhibitors in this class have recently been approved for clinical use in the treatment of advanced non–small cell lung cancer as monotherapy following failure of chemotherapy. We examined the efficacy of gefitinib on lung tumorigenesis in mouse models using both postinitiation and progression protocols. Gefitinib was given at a dose of 200 mg/kg body weight (i.g.) beginning either 2 or 12 weeks following carcinogen initiation. In the postinitiation protocol, gefitinib significantly inhibited both tumor multiplicity (∼70%) and tumor load (∼90%) in A/J or p53-mutant mice (P < 0.0001). Interestingly, gefitinib was also highly effective against lung carcinogenesis in the progression protocol when individual animals already have multiple preinvasive lesions in the lung. Gefitinib exhibited ∼60% inhibition of tumor multiplicity and ∼80% inhibition of tumor load when compared with control mice (both P < 0.0001). These data show that gefitinib is a potent chemopreventive agent in both wild-type and p53-mutant mice and that a delayed administration was still highly effective. Analyses of mutations in the EGFR and K-ras genes in lung tumors from either control or treatment groups showed no mutations in EGFR and consistent mutation in K-ras. Using an oligonucleotide array on control and gefitinib-treated lesions showed that gefitinib treatment failed to alter the activity or the expression level of EGFR. In contrast, gefitinib treatment significantly altered the expression of a series of genes involved in cell cycle, cell proliferation, cell transformation, angiogenesis, DNA synthesis, cell migration, immune responses, and apoptosis. Thus, gefitinib showed highly promising chemopreventive and chemotherapeutic activity in this mouse model of lung carcinogenesis. (Mol Cancer Res 2006;4(12):971–81)

Lung cancer is currently the leading cause of cancer mortality in the United States (1). Although surgical resection can be curative for many early-stage cancers, the overall 5-year survival rate for all stages of lung cancer is only 15% (2). Thus, prevention is an important part of lung cancer control especially in high-risk populations, such as current smokers and former smokers. Chemoprevention is defined as the use of specific natural or synthetic chemical agents to reverse, suppress, prevent, or delay the carcinogenic process either by blocking the development of early lesions or by inhibiting the progression to invasive cancer.

Epidermal growth factor receptor (EGFR) is a promising target for anticancer therapy because it is activated in a wide variety of cancers. Gefitinib (Iressa, ZD1839), a competitive EGFR tyrosine kinase inhibitor, was selected from a variety of substituted 4-(3-chloroanilino)-quinazolines as the most effective in inhibiting tyrosine kinases activity (3). It is an anilinoquinazoline compound with the chemical name 4-quinazolinamine, N-(3-chloro-4-flurophenyl)-7-methoxy-6-[3-(4-morpholinyl)propoxy]. Several in vitro and in vivo studies have shown that gefitinib can inhibit the growth of cancer cells and suppress the growth of xenografts (4, 5). Small-molecule EGFR inhibitors have recently been approved for clinical use in the treatment of advanced non–small cell lung cancer (NSCLC) as monotherapy following failure of chemotherapy (6). In two phase II trials, the response rates to gefitinib ranged from 11.8% to 18.4% (7, 8). Interestingly, the likelihood of a clinical response to gefitinib is greatly increased in lung cancer when activating mutations in the tyrosine kinase domain of the EGFR were found (9, 10). A smaller percentage of advanced lung cancers without EGFR mutations as well as other cancers lacking EGFR mutations, including head and neck, colon, and breast, seem to respond to gefitinib (11).

Mouse models allow vigorous preclinical characterization of efficacy and determination of mechanisms that can support phase II/III human chemopreventive clinical trials (12-14). Although gefitinib induces clinical responses in a subset of treated lung cancer patients, it has not been tested as a cancer prevention agent. Using a protocol to assess its role in prevention and a protocol to assess its role in progression, we treated a cohort of mice that had been exposed to benzo(a)pyrene (BaP), a carcinogen that induces lung tumors in mice. Treatment with gefitinib resulted in a substantial tumor prevention effect as well as a potent effect on blocking tumor progression. We also analyzed mutations in EGFR and K-ras in the resulting mouse lung tumors. Finally, we used Affymetrix microarrays (Affymetrix, Santa Clara, CA) to define genes whose expression was altered when comparing lesions from control and gefitinib-treated groups.

Before the chemoprevention experiment presented here, we did an 8-week toxicity assay with gefitinib (Fig. 1A). Of doses from 100 to 400 mg/kg body weight, we found that a dose of 200 mg/kg body weight did not cause mortality or significantly affect body weight. Thus, the daily dose used was 200 mg/kg for this study. The daily dose of gefitinib in humans is 250 to 500 mg.³

Based on the updated Guidance for Industry: Estimating the Maximum Safe Starting Dose in Initial Clinical Trials for Therapeutics in Adult Healthy Volunteers, U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research,⁴ the conversion factor is 12.3 for converting mouse dose in mg/kg to human equivalent dose in mg/kg. Assuming human body weight as 60 kg (as the guidance indicated), the mg/kg in human is 500 / 60 = 8.3 mg/kg. When we converted the dose we used in the study (200 mg/kg in mouse) by the factor of 12.3, the human equivalent dose is 200 / 12.3 = 16.3 mg/kg. It was just less than two times higher than the clinical dose in humans.

We began with the postinitiation protocol (Fig. 2A, a). Gefitinib treatments in either wild-type or p53^mut/wt mice were started 2 weeks after exposure to BaP and terminated at 20 weeks after BaP. It was continued for a total of 18 weeks and caused no significant body weight loss at the dose of 200 mg/kg. However, some treated mice showed light alopecia especially at the region near the eyelid (Fig. 1B). All the mice developed lung tumors following exposure to the carcinogen. The administration of gefitinib showed potent preventive effects on BaP-induced lung tumorigenesis. Gefitinib significantly reduced tumor multiplicity by 62% and 75% in wild-type and p53^mut/wt mice, respectively (Fig. 3A). On the total tumor load, which reflects both tumor multiplicity and tumor size, the treatment of gefitinib showed even more striking effects. Tumor load was decreased by 89% and 94% in wild-type and p53^mut/wt mice, respectively (Fig. 3B). Interestingly, the tumor size was in the range of 0.6 to 2.0 mm (diameter) in ∼80% of the animals in the control group. In contrast, the majority of the tumors (77%) found in the gefitinib-treated group were significantly smaller and in the range of 0.2 to 0.5 mm (diameter), although in the gefitinib-treated group all pulmonary lesions were hyperplasias or adenomas. These results indicate that gefitinib inhibits and/or reverses tumor progression. When we used the progression protocol (Fig. 2A, b), gefitinib treatments were started 12 weeks after exposure to BaP and terminated at 20 weeks after BaP. It was continued for a total of 8 weeks and no significant body weight loss was observed at the dose of 200 mg/kg. Gefitinib significantly reduced tumor multiplicity by 59.7% when compared with control group (Fig. 3C). On the total tumor load, the treatment of gefitinib showed even more striking effects. It decreased the tumor load by 81.4% when compared with control group (Fig. 3D). Gross and microscopic features of lung tumors from mice with or without gefitinib treatment are shown in Fig. 2B. All tumors seen in gefitinib-treated group were lung adenomas. Lung adenocarcinomas in control group are composed of cells with varying degrees of differentiation, with most cells appearing relatively undifferentiated relative to lung adenomas in treated mice (Fig. 2B, b). There is a complete loss of normal alveolar architecture, and the nuclear cytoplasmic ratio is increased in the carcinomas. Nuclear crowding and cytologic atypia are present, and there is heterogeneity of growth patterns. These data indicate that gefitinib is a potent chemopreventive and chemotherapeutic agent for lung tumorigenesis.

Mutations of EGFR have been reported to respond to gefitinib in humans. The main mutations were either in-frame deletions or amino acid substitutions clustered around the ATP-binding pocket of the tyrosine kinase domain focused on exons 18, 19, and 21 (single-nucleotide substitutions in exons 18 and 21 and in-frame deletions in exon 19; refs. 9, 10). In this study, exons 18, 19, 20, and 21 of EGFR from a series of both normal lungs and lung tumors with or without gefitinib treatment were sequenced by PCR direct sequencing. No mutations in these exons were found in DNAs from 5 control tumors and 5 gefitinib-treated lung tumors from A/J mice as well as from 20 adenomas and 5 adenocarcinomas from p53 transgenic mice. The results were confirmed in both sense and antisense direction analysis. Next, we determined the presence of K-ras mutations in lung tumors from mice with and without gefitinib. Mutation analysis of K-ras was done on a total of 11 lung tumors from wild-type mice that were treated with BaP and 9 tumors in mice treated with BaP plus gefitinib. The overwhelming majority of the tumors harbored mutations in the 12th codon of K-ras, and there was no significant difference in the incidence of mutation between tumors in mice treated with and without gefitinib [8 of 9 (88.9%) versus 10 of 11 (90.9%); Table 1].

To determine whether protein levels of EGFR are altered between normal lung and BaP-induced lung tumors, samples from each group were subjected to Western blot analysis. EGFR protein expression was not detected in three separate adenocarcinomas (diameter, >3 mm; ref. 15), although EGFR was expressed in two of the three normal lung samples (Fig. 4A). To confirm the result of higher protein level of EGFR in normal tissue than in tumor, we conducted the quantitative real-time PCR analyses for EGFR between normal tissues and BaP-induced lung tumors. The expression of EGFR in normal tissue was 2.64-fold higher than that in BaP-induced lung tumors (Fig. 4B). These results are consistent with the result from the Western blot analysis. We also determined the phosphorylation state of EGFR in the lung tumors and normal lung by Western blot using an antibody specific to phosphorylated Tyr¹⁰⁶⁸ of EGFR. Phosphorylation of EGFR in both tissues was not detectable; however, the phosphorylation of an equivalent amount of EGFR was significantly higher in 293T cells (Fig. 4A). This suggests that the phosphorylation levels of EGFR in tumors and normal lung are both very low. Finally, because gefitinib is tyrosine kinase inhibitor and may inhibit other growth factor receptors with tyrosine kinase domain by binding at their ATP site of the tyrosine kinase region, such as vascular endothelial growth factor receptor (VEGFR), fibroblast growth factor receptor (FGFR), platelet-derived growth factor receptor, and insulin-like growth factor receptor (IGFR), the mRNA expression levels of VEGFR, FGFR, platelet-derived growth factor receptor, and IGFR were analyzed using real-time PCR. Similarly, we did not observe increased expression of these growth factor receptors in mouse lung tumors. The mRNA expression in normal lung tissues is 2.83-, 4.00-, 3.72-, and 3.03-fold higher than those in lung tumor tissues in VEGFR, FGFR, FDGFR, and IGFR target genes, respectively (Fig. 4B).

To identify specific genes or pathways that may directly or indirectly contribute to the protective effect of gefitinib on tumorigenesis, three to four independent tumors and normal tissues from gefitinib treatment and control groups were used for this analysis. Oligonucleotide array using Affymetrix GeneChips revealed >3,500 genes that were differentially expressed when comparing control normal lung tissues and control lung tumors (data not shown). Among them, the expression of 208 genes, which were initially underexpressed or overexpressed in control lung tumors, was modulated toward to the normal lung tissue level in gefitinib-treated lung tumors (with the cutoff fold >1.5, P < 0.05; Fig. 5A). Among the 208 genes, 71 underexpressed genes were up-regulated and 137 overexpressed genes were down-regulated by gefitinib (Fig. 5A). Genes that were related to tumorigenesis were shown in Table 2. We selected six genes (Ar, Amid, Rspondin, Fas, G0s2, and Csf2rb1) for confirmation, and all of them were validated by real-time PCR analyses. Expression of Ar, Amid, Rspondin, Fas, and G0s2 was up-regulated toward the control lung normal tissue by gefitinib treatment, whereas expression of Csf2rb1 was down-regulated to normal tissue (Fig. 5B). GenMAPP search revealed that the levels of expression of several apoptotic genes were increased following gefitinib treatment, which may contribute to the efficacy of gefitinib in this mouse lung model (Fig. 5C). Furthermore, three additional pathways, including cell cycle control, G13 signaling, and transforming growth factor-β signaling, may also be involved in the efficacy of gefitinib against mouse lung tumorigenesis (Supplementary Figs. S1-S3).

We have also compared the gene expression changes in mouse models exposed to gefitinib with those that have been found to predict sensitivity to gefitinib in human NSCLC (16). As shown in Table 3, we found that gefitinib treatment changed the expression levels for 15 of these 131 genes in mice tumors (fold change, >1.5), which correspond to the expression profiles predicting sensitivity to gefitinib in human NSCLC. Among these 15 genes, 8 of them were modulated by gefitinib toward the normal level (fold change, >1.5). These genes may represent common targets of gefitinib in both human and mouse lung tumors.

Gefitinib is a small-molecule agent that is recognized as an inhibitor of the tyrosine kinase domain of the EGFR (3, 17, 18). Inhibition of the tyrosine phosphorylation by gefitinib results in decreased activation of various downstream elements (19). Gefitinib and other EGFR inhibitors can induce growth inhibition and growth delay in a wide range of tumor cell lines and human tumor xenografts. In some NSCLC patients, gefitinib can cause remarkably rapid and often profound responses. Small-molecule EGFR inhibitors have been recently approved for clinical use in the treatment of advanced NSCLC as monotherapy following failure of chemotherapy (6). Most of the initial clinical trials with the EGFR inhibitors have used individuals with metastatic disease who have failed standard therapies (11). Many of those clinical trials have shown limited efficacy as monotherapies (11). However, a more limited number of studies in earlier stages of cancer have shown more striking efficacy, implying that this class of compounds might be more useful in earlier clinical stages or in prevention (11). Interestingly, in advanced lung cancer, the efficacy of the EGFR inhibitors was most effective against tumors with an activating mutations in the tyrosine kinase domain of the EGFR (9, 10). However, some lung tumors without EGFR mutations and other cancers without EGFR mutations have proven sensitive to the EGFR inhibitors as well (11).

In view of the potential interest in the use of EGFR inhibitors in early-stage lung cancer and potentially in a prevention setting, we undertook the studies presented. We conducted a bioassay of gefitinib on BaP-induced lung tumorigenesis in A/J mice that mimic histopathology and stages of tumor progression with human lung adenocarcinomas (13, 14). Gefitinib was well tolerated throughout the duration of the experiment (Fig. 1). We used two different prevention protocols to examine the chemopreventive efficacy. The first protocol entailed treatment with gefitinib beginning 2 weeks after BaP for 18 weeks. In wild-type or p53^mut/wt mice, this resulted in profound decreases in tumor multiplicity (62% and 75%; Fig. 3A) and even more striking decreases in tumor volume (89 and 94%; Fig. 3B). Our second protocol done only in wild-type mice resulted in significantly decreased tumor multiplicity and volume (59.7% and 81.4%, respectively; Fig. 3C and D) despite the fact that treatment was started 12 weeks after the carcinogenic initiation with BaP. At 12 weeks after the BaP injection, multiple preinvasive lesions have developed in lung tissues. Based on our prior experience, lung adenomas tend to have tumors <1 mm in diameter, whereas lung adenocarcinomas will tend to have tumors >3 mm (15). In view of the striking decrease in average size of the tumors in the gefitinib group, this implies a striking change in the incidence of adenocarcinomas. These results have two important findings. Gefitinib was highly effective in lesions with or without a p53 mutation when given early. More importantly, late administration in wild-type mice was similarly effective as early treatment, implying that most of the effect occurs later during the progression stage, which is important because this is the most likely treatment protocol clinically. These results show that gefitinib can strongly inhibit and potentially reverse the process of lung tumorigenesis.

Activating mutations of EGFR tyrosine kinase domain have been reported to dramatically increase the response to gefitinib treatment in human lung cancer patients, although a significant number of tumors without such mutations were responsive as well (9, 10). The main mutations were either in-frame deletions or amino acid substitutions clustered around the ATP-binding pocket of the tyrosine kinase domain focused on exons 18, 19, and 21 (single-nucleotide substitutions in exons 18 and 21 and in-frame deletions in exon 19). Based on our prior experience, lung adenomas tend to have tumors <1 mm in diameter, whereas lung adenocarcinomas will most likely consists of tumors >3 mm (15). Accordingly, we sequenced the same regions flanking exons of 18, 19, 20, and 21 from 5 control normal lungs, 5 gefitinib-treated normal lungs, 5 control tumors, 5 treated tumors, 5 large BaP-induced tumors (>3 mm in diameter), and 20 small BaP-induced tumors (<1 mm in diameter). In both sense and antisense sequence analysis, we could not find mutations with these samples in exons 18, 19, 20, and 21. Gefitinib had activities in several cancer types, including recurrent glioblastoma, where there are no mutations in the EGFR tyrosine kinase region (20). There were no detectable levels of phosphorylated EGFR in mouse lung tumors examined, which is the primary postulated target of these EGFR inhibitors (Fig. 4A). To determine whether protein levels of EGFR differed between normal lung and BaP-induced lung tumors, we did Western blot analysis. We found that the expression levels in normal tissue were much higher than that in adenocarcinomas (Fig. 4A). The relative mRNA expression analysis of EGFR by real-time PCR also indicated that high expression of EGFR is probably not needed for the efficacy. The results were consistent with the finding that the efficacy of gefitinib in tumorigenesis was not dependent on the expression level or phosphorylation status of EGFR (4, 21). EGFR belongs to the ErbB family of tyrosine kinase protein. This family includes four members: EGFR, ErbB-2/Neu/HER2, ErbB-3/HER3, and ErbB-4/HER4. The members of the ErbB family have a similar structure with high degree of homology in the tyrosine kinase domain (18). Their homodimeric and heterodimeric dimerization is followed by activation of intrinsic protein tyrosine kinase activity, tyrosine phosphorylation, and activation of intracellular signal transduction pathway, such as phosphatidylinositol 3-kinase/AKT and ras/raf/MAPK/extracellular signal-regulated kinase kinase/MAPK pathway (22). It is possible that despite the absence of EGFR mutations in kinase domain and no detectable expression in tumors, gefitinib treatment dramatically reduced both the number and size of tumor nodules following exposure to a known carcinogen would suggest that in fact the mechanism of the chemoprevention of gefitinib may be related to other ErbB family members, such as ErbB-2, ErbB-3, and ErbB-4. It has been proved that gefitinib is able to affect signal transduction through different ErbB receptors (18).

To identify specific genes or pathways that may directly or indirectly contribute to the protective effect of gefitinib on gene expression in lung tumors from control and gefitinib-treated mice was investigated. Oligonucleotide array together with GenMAPP analysis revealed that gefitinib affects several cellular processes involved in the tumorigenesis, including apoptosis, cell cycle, cell proliferation, cell transformation, angiogenesis, DNA synthesis, cell migration, immune responses, etc. (Table 2). They are involved in the E2F pathway, tumor necrosis factor pathway, Fas pathway, nuclear factor-κB (NF-κB) pathway, caspase pathway, etc. Many of them are related to the process of apoptosis (Fig. 5C). Specifically, E2F1, Tnfrsf10b, Fas, Sesn1, Mib1, Nfkbi2, Ret, and Tnfrsf9 are found to be modulated toward levels observed in control normal lung. For example, E2F1 can induce apoptosis by a death receptor–dependent mechanism and by down-regulating Traf2 protein levels and inhibiting activation of antiapoptotic signals, such as NF-κB (23). E2F1 also can induce p53-dependent apoptosis by inducing ARF (CDKN2A), which neutralizes Mdm2 and stabilizes p53 that similarly may induce apoptosis (24). E2F1 can bind with the promoter region of human Apaf1 and induce apoptosis accompanied by casp9 (25). In this study, E2F1 was expressed at very low level in control tumors, whereas the gefitinib treatment up-regulated the expression by ∼7-fold (Table 2). Similarly, Fas belongs to tumor necrosis factor receptor superfamily. It activates caspase-3 (26) and caspase-8 for the cascade of caspases to lead to apoptosis (27). In the BaP-induced lung tumor, the expression level was 2.4-fold lower than normal tissue. The treatment with gefitinib reversed the expression toward to the normal level (Table 2). NF-κB transcription factor has been detected in numerous cell types that express cytokines, chemokines, growth factors, cell adhesion molecules, and some acute phase proteins in health and in various disease states. It can protect from or contribute to apoptosis. It was reported that complete and persistent inhibition of NF-κB has been linked directly to apoptosis (28). In contrast, in cells without NF-κB activity, p53-induced apoptosis was abrogated (29). IκB proteins (Nfkbia, Nfkbib, and Nfkbie) inactivate NF-κB by trapping it in the cytoplasm. They are necessary to generate the characteristic NF-κB activation profile (30). In this study, the expression level of Nfkbie was down-regulated by gefitinib to normal level (Table 2). Interestingly, gefitinib treatment changed the expression levels for 15 of those 131 genes (16) in mice tumors (fold change, >1.5; Table 3), which correspond to the expression profiles predicting sensitivity to gefitinib in human NSCLCs (16). Among them, eight were modulated by gefitinib toward to the normal level (fold change, >1.5). These genes may represent the common targets of gefitinib in both human and mouse lung tumors.

In summary, this study examined the effect of gefitinib on inhibiting pulmonary carcinogenesis in the A/J strain of mice, which is susceptible to lung tumor induction by chemical carcinogens, and in a p53-mutant strain of mouse. Gefitinib was able to inhibit tumor multiplicity and tumor size to a high extent whether given beginning 2 weeks or beginning 12 weeks after carcinogen treatment. Consistent mutation in K-ras was found in tumors that arose in both the placebo-treated and gefitinib-treated lungs. Gefitinib altered gene expression in a reproducible way for genes involved in cell growth and cell death. In many of the human studies in metastatic NSCLC, mutations in K-ras seemed to predict a lack of response to gefitinib (31). Thus, concurrent EGFR mutations and K-ras mutations were virtually never observed (31). However, the A/J lung tumor model has K-ras mutations in virtually 100% of the tumors examined either in wild-type or p53^mut/wt animals. Despite this lack of EGFR mutations and presence of K-ras mutations, we observed striking efficacy in both wild-type and p53-mutant animals even when gefitinib was used late during tumor progression. This is the first report of a chemopreventive effect of an EGFR inhibitor in mouse models of lung cancer. The results suggest that inhibiting EGFR may have potential as a chemopreventive approach.

BaP (99% pure) and tricaprylin were purchased from Sigma Chemical Co. (St. Louis, MO). Gefitinib was obtained from Chemoprevention Branch, National Cancer Institute (Bethesda, MD). A/J female mice at 6 weeks of age were obtained from The Jackson Laboratory (Bar Harbor, ME). The use of animals was approved by the Institutional Animal Care and Use Committee of the Washington University (St. Louis, MO).

A preliminary toxicity study was conducted with gefitinib (Fig. 1A), and a dose of 200 mg/kg body weight did not cause mortality or significantly affect body weight. Briefly, A/J female mice were divided into five dose groups: control, 100 mg/kg (32), 200 mg/kg, 300 mg/kg, and 400 mg/kg for gefitinib treatment by gavage, five times per week, in 0.2 mL of 1% Tween 80 solution. Five animals were used in each testing group. Animals were housed in plastic cages with hardwood bedding and dust covers in a high-efficiency particulate air–filtered, environmentally controlled room (24 ± 1°C, 12-h light/12-h dark cycle). Body weight was monitored weekly for the duration of the studies. All food and water were available ad libitum.

The A/J female mice were obtained from The Jackson Laboratory. A total of 95 mice was used in this study with 12 to 25 mice per group. Mice were given a single i.p. dose of BaP (100 mg/kg body weight, prepared just before injection) in 0.2 mL tricaprylin. Two or 12 weeks following the BaP injection, mice were randomized into vehicle control group and gefitinib treatment group. Treatments were initiated and continued for 18 or 8 weeks (Fig. 2A). In testing group, mice were given gefitinib at the dose of 200 mg/kg in 0.2 mL of 1% Tween 80 by gavage based on our preliminary toxicology results five times per week. The justification for the use of a 200 mg/kg dose of gefitinib five times per week is because it is the maximum tolerated dose that did not significantly affect body weight over an 8-week period. If this dosage is highly effective, future experiments will use the 200 mg/kg as a control to determine the effect of lowering doses to 100 or 50 mg/kg. In the control group, mice were give vehicle of 1% Tween 80 only. The body weight was measured every 2 weeks for the duration of the treatment. Mice were killed 20 weeks after exposure to carcinogen of BaP by CO₂ asphyxiation. Portions of tumor and normal tissue were frozen in liquid nitrogen and then transferred to −80°C until use. The remaining lung was fixed in Tellyesniczky's solution (15) overnight followed by 70% ethanol. The fixed lungs were evaluated by at least two investigators under a dissecting microscope to obtain the fixed surface tumor count, and individual tumor load was measured for volume calculation based on the following formula: mm³ = V = 4 / 3πr³. The total tumor load was obtained by adding the individual tumor load per lung, and the total tumor count was obtained by adding the individual tumor count per lung.

DNA was isolated from five samples in each of control normal tissue, control tumor, gefitinib-treated lung tissue, and treated tumor from the postinitiation protocol using the Trizol (Invitrogen, Carlsbad, CA). We also used DNA samples from BaP-induced lung tumors in p53 transgenic mice, including 20 adenomas (diameter, <1 mm) and 5 adenocarcinoma (diameter, >3 mm) from another bioassay (15). Sequences of PCR primers for EGFR were based on the data on the Web site⁵

with transcript ID ENSMUST00000020329 (Table 4). PCR mixtures with DNA template were subjected to 35 cycles of PCR amplification. Each cycle consisted of 1 min each at 95°C, 51°C, and 72°C. PCR products were resolved on 1.5% ethidium bromide–stained agarose gels and purified using QIAquick gel extraction kits (Qiagen Sciences, Germantown, MD). All the PCR fragments were automatically sequenced in both sense and antisense directions in the Protein and Nucleic Acid Chemistry Laboratories of Washington University by using Applied Biosystems Models 3100 and 3730 DNA sequencers (Perkin-Elmer, Foster City, CA). The mutations of 18, 19, 20, and 21 exons in EGFR were analyzed with the Sequencher software version 4.0.5 (Gene Codes Corp., Ann Arbor, MI).

Samples of extracted normal lung and lung tumors were Dounce homogenized in PBS in the presence of 1 mmol/L DTT, Complete protease inhibitor cocktail (Roche, Indianapolis, IN), and phosphatase inhibitor cocktail I and II (Sigma). Triton X-100 was added to a final concentration of 1%, and lysates remained on ice for 10 min. Lysates were centrifuged at 13,000 × g at 4°C. The supernatants were collected and added to lithium dodecyl sulfate sample buffer (Invitrogen) and 50 mmol/L DTT. Samples were electrophoresed on a 4% to 12% NuPAGE (Invitrogen) gel, transferred to polyvinylidene difluoride, and blotted with the indicated antibodies (EGFR, phosphorylated EGFR, and MAPK; Cell Signaling Technology, Danvers, MA). MAPK was used for internal control.

Relative gene expressions of EGFR (NM_207655), VEGFR (NM_008029), FGFR (NM_008010), platelet-derived growth factor receptor (NM_011058), and IGFR (NM_010513) were determined by quantitative real-time PCR analysis (33). The primers shown on Table 5 were designed using PrimerExpress version 2.0 (Applied Biosystems, Foster City, CA). Amplification of each target cDNA was done with SYBR Green Master Mix in Bio-Rad (Hercules, CA) Single-Color Real-time PCR Detection System according to the protocols provided. The control gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and target genes were amplified with equal efficiencies, and the −ΔCt (−ΔCt = Ct GAPDH gene − Ct target gene) values were used to quantitate the logarithmic transformation of the mRNA level of the target gene relative to that of GAPDH (33).

Total RNA from each sample was isolated by Trizol according to the manufacturer's protocols. Then, the samples of RNA were treated at GeneChip Core Facility in Washington University. Labeled cRNA was applied to the Affymetrix MOE430+2 GeneChips (Affymetrix), which contain 45,101 genes and expressed sequence tags on one array. Four samples in control normal lung and control tumor groups and three samples in gefitinib-treated tumor group were used for this array analysis. For each sample in control tumor and treated tumor, one to four tumors were combined (pooled) based on their size.

Four or three independent tumor samples were collected for each group. Array normalization and gene expression estimates were obtained using Affymetrix Microarray Suite 5.0 software. Differential expression was determined on the combined basis of statistical testing using t test and based on ratio with a cutoff of P < 0.05 and fold >1.5 being called positive for differential expression. For the selected genes, expression indexes were transformed across samples to an N(0, 1) distribution using a standard statistical Z-transform. These values were input to the GeneCluster program as reported (15), and genes were clustered using average linkage and correlation dissimilarity. Signal transduction pathways and other functional groupings of genes were evaluated for differential regulation using the visualization tool GenMAPP (University of California at San Francisco).⁶

It is a recently reported tool for visualizing expression data in the context of biological pathways (34). We imported the statistical results of our data set into the program and used GenMAPP to illustrate pathways containing differentially expressed genes (15). Differential gene expression was based on gefitinib treatment versus nontreatment expression change (fold change, >1.5).

Tumor multiplicity and tumor load were analyzed by two-sided Student's t test using Microsoft Excel 2002 Service Pack 3 to determine the differences in the number and size of lung tumors per mouse between each group. In all t tests, the level of statistical significance was set at P < 0.05.