Farnesyltransferase inhibitors are potent lung cancer ... - Nature

Oncogene (2003) 22, 6257–6265

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Farnesyltransferase inhibitors are potent lung cancer chemopreventive agents in A/J mice with a dominant-negative p53 and/or heterozygous deletion of Ink4a/Arf Zhongqiu Zhang1,6, Yian Wang1,6, Laura E Lantry2, Elizabeth Kastens1, Gongjie Liu2, Andrew D Hamilton3, Said M Sebti4, Ronald A Lubet5 and Ming You1,* 1 Department of Surgery and The Alvin J. Siteman Cancer Center, Campus Box 8109, Washington University School of Medicine, 660 South Euclid Avenue, St Louis, MO 63110, USA; 2Department of Pathology, Medical College of Ohio, OH 43614, USA; 3 Department of Chemistry, Yale University, New Haven, Connecticut, USA; 4Department of Oncology, Drug Discovery Program, H. Lee Moffitt Cancer Center and Research Institute, University of South Florida, Tampa, FL, USA; 5Chemoprevention Agent Development Research Group, National Cancer Institute, Bethesda, MD 20892, USA

Mutations in the Kras2 gene are seen in both human and mouse lung adenocarcinomas. The protein product (p21ras) encoded by the Kras2 gene must be post-translationally modified at a terminal CAAX motif in order to be biologically active. In this study, we systematically investigated the chemopreventive efficacy of two different farnesyltransferase inhibitors (FTIs): one is a peptidomimetic (FTI-276) and the other is an imidazole (L778–123). Both FTIs are designed to inhibit the posttranslational modification of p21ras proteins with a terminal CAAX motif. In a complete chemoprevention study, where the inhibitor was administered before carcinogen was given, and throughout the study, FTI276 treatment significantly reduced both the tumor multiplicity by 41.7% (Po0.005), and the total tumor volume by 79.4% (Po0.0001). In the late treatment study, where mice were treated with an inhibitor 12 to 20 weeks after carcinogen administration, FTI-276 treatment resulted in a 60% reduction in tumor multiplicity and 58% reduction in tumor volume. Next, we examined the chemopreventive efficacy of a new FTI, L-778,123, on lung tumor development in A/J mice and transgenic mice with a dominant-negative p53 mutation and/or heterozygous deletion of Ink4a/Arf. Treatment of mice with L-778,123 for a period of 10 weeks from 20 weeks to 30 weeks post carcinogen initiation resulted in an B50% decrease in tumor multiplicity in wild-type mice and mice with a dominant-negative p53 mutation and/or heterozygous deletion of the Ink4a/Arf tumor suppressor genes. Interestingly, tumor volume was decreased B50% in wildtype mice and in mice with an Ink4a/Arf heterozygous deletion, while tumor volume was decreased B75% in animals with a dominant-negative p53 and in mice with both a p53 mutation and heterozygous deletion of Ink4a/ Arf. This result suggests that FTI exhibited a significantly (Po0.05) more efficacious chemopreventive effect in animals with alterations of p53 and Ink4a/Arf as *Correspondence: M You; E-mail: [email protected] 6 These authors contributed equally to this work Received 3 February 2002; revised 26 March 2003; accepted 3 April 2003

contrasted with wild-type mice. Thus, FTIs are potent lung chemopreventive agents in both A/J mice and transgenic mice harboring a dominant-negative p53 and heterozygous deletion of Ink4a/Arf. In fact, L-778,123 is more effective in inhibiting primary lung progression in mice with a p53 mutation and/or an Ink4a/Arf deletion than in wild-type animals. Oncogene (2003) 22, 6257–6265. doi:10.1038/sj.onc.1206630 Keywords: farnesyltransferase inhibitor; Lung cancer; Chemoprevention; A/J mice; Transgenic mice; P53; Ink4a/Arf

Introduction Lung cancer is the leading cause of cancer death in men and women in the US (Minna et al., 1989). There is a strong correlation between lung cancer susceptibility and exposure to tobacco-related carcinogens (Sellers et al., 1992; Landis et al., 1998). Molecular changes in proto-oncogenes and tumor suppressor genes have been detected in all stages of lung carcinogenesis. Mutations in the ras oncogenes have been implicated in the pathogenesis of several human cancers including lung, colon, and pancreatic (Bos, 1989) cancers. In particular, nonsmall cell lung carcinoma (NSCLC), which accounts for 75% of lung cancers, has been shown to harbor Kras2 mutations in 30–50% of all cases (Rodenhuis and Slebos, 1992). Thus, blocking the effect of ras mutations has become one of the major strategies in cancer treatment and prevention. Similar genetic changes seen in human lung cancers have also been observed in mouse lung tumors. Kras2 is frequently activated in both mouse lung adenomas and adenocarcinomas (You et al., 1989). Alterations of other known tumor suppressors, such as p16, Rb, and p53, have been detected in lung tumors as well, although typically these have occurred later in tumor progression and often at lower frequencies (Herzog et al., 1997). We

Effect of FTI on lung tumorigenesis in transgenic mice Z Zhang et al

6258

have previously demonstrated an increased susceptibility to chemical induction of mouse lung, uterine, colon, and skin tumorigenesis in p53 transgenic mice (Lubet et al., 2000; Zhang et al., 2000, 2002a). The region containing the Ink4a/Arf locus on mouse chromosome 4 was frequently lost (50%) in lung carcinomas of various F1 hybrid mouse strains (Herzog et al., 1997). Furthermore, the Ink4a/Arf locus is homozygously codeleted in 12 of 16 (75%) mouse lung tumor cell lines (Herzog et al., 1996; McDoniels-Silvers AL et al., 2001). More recently, we provided functional evidence in vivo that heterozygous loss of Ink4a/Arf contributes to enhance lung tumor progression in mice (Zhang et al., 2002b). In general, genetic changes found in mouse lung tumors have striking similarities to those existing in humans. The Ras family of small GTP binding proteins is key to many signaling events in eukaryotes, including those for growth, differentiation, cytoskeletal organization, and apoptosis (Bos, 1989; Khosravi-Far and Der, 1994). All Ras proteins are synthesized as inactive precursors, which must be post-translationally modified to be biologically active (Casey et al., 1989; Hancock et al., 1989). The first step in the modification of Ras is the addition of a 15-carbon isoprenoid (farnesol) by farnesyltransferase (FTase) to the C-terminal cysteine of the CAAX motif (Hancock et al., 1989). Therefore, potential strategies have been sought that selectively interfere with farnesylation of constitutively activated oncogenic proteins like Ras (Gibbs et al., 1993; Tamanoi, 1993; Lerner et al., 1995a and b). There have been a series of reports on FTI as a potential cancer chemotherapeutic agent (Sebti and Hamilton, 2000). In cell culture models, FTIs reverse many of the biological properties of ras-transformed cells, particularly Ha-ras or N-ras, including inhibition of anchorage-independent growth, and morphological reversion (Prendergast and Oliff, 2000; Sebti and Hamilton, 2000; Reuter et al., 2000). These compounds also inhibited the growth of ras-transformed fibroblasts and human tumor cell lines transplanted into nude mice (Sebti and Hamilton, 2000; Reuter et al., 2000; Prendergast and Oliff, 2000). Most in vitro studies have predicted that FTI would be cytostatic rather than cytotoxic. Previous studies with FTI-276 employing the human lung adenocarcinoma cell lines CaLu-1 and A-549, which bear a mutant Kras2 oncogene, demonstrated a growth inhibition of CaLu1 cell xenografts (Sun et al., 1995, 1998). We recently reported the in vitro evaluation of FTI-276 in four lung epithelial cell line, two of which were immortalized, C10 and E10, the latter harboring a p53 mutation (Lantry et al., 2000a). The remaining cell lines, LM1 and LM2, were both tumorigenic and metastatic in mouse models, the latter of which harbors a 61st codon mutation in Kras2 (Lantry et al., 2000a). The most normal cell lines, the C10, were the least affected, with a 22% reduction in growth rate over control (Lantry et al., 2000a). The LM1 with wild-type ras had the greatest reduction in growth rate over control, at 64%, followed by the LM2 at 51%, and the E10 at 43% (Lantry et al., 2000b). In vivo, FTIs treatment of mice carrying activated forms of Oncogene

Ha-ras, N-ras, or K-ras showed either tumor regression or tumor stasis (Kohl et al., 1995; Mangues et al., 1998; Omer et al., 2000). For example, treatment of MMTVH-ras mice with the FTI L-744,832 caused rapid tumor regression, treatment of MMTV-N-ras mice with FTI L744,832 also caused tumor growth inhibition, while treatment of MMTV-Ki-rasB mice caused tumor stasis in mammary tumorigenesis (Kohl et al., 1995; Mangues et al., 1998; Omer et al., 2000). We had previously shown that FTI-276 exhibited a significant efficacy in a late treatment chemoprevention protocol in the A/J mouse lung tumor model (Lantry et al., 2000b). In the present study, we determined the chemopreventive effect of FTI-276 in a complete chemoprevention protocol in the A/J mice, and we also determined the efficacy of another FTI inhibitor (L778,123) in transgenic mice with various genotypes including wild-type mice, mice with a dominant-negative p53 mutation, mice with heterozygous deletion of the Ink4a/Arf locus and mice with both a p53 mutation and heterozygous deletion of the Ink4a/Arf locus. The significance of using transgenic mice that contain a dominant-negative p53 mutation and/or heterozygous deletion of the Ink4a/Arf is the fact that mouse lung tumor progression in these transgenic mouse models will involve three major genetic alterations (Kras2, p53, and Ink4a/Arf) commonly found in human lung NSCLC. Inactivation of p16 by hypermethylation and homozygous deletion has been detected in the majority of NSCLCs (Merlo et al., 1995). Other common changes include p53 mutations and Rb gene inactivation that have been seen frequently in invasive carcinomas (Chiba et al., 1990; Xu et al., 1991). In the present study, we crossed p53 transgenic mice with heterozygous Ink4a/Arf deficient mice to investigate the functional role of p53 and Ink4a/Arf genes in mouse lung tumorigenesis. Use of these mice with either a germline mutant p53 transgene (p53val135/wt), heterozygous Ink4a/Arf deficient, or both, on a A/J background, in conjunction with the lung-specific chemical carcinogens allows us to evaluate FTI as a chemopreventive agent against lung cancer in a potentially more relevant model than the standard A/J mouse model.

Results Effect of FTI-276 on 4-(methylnitrosamino)-1(3-pyridyl)-1-butanone (NNK)-induced lung tumorigenesis in A/J mice The effect of late treatment protocol for FTI-276 was previously examined in the A/J mouse lung tumor model. Mouse lung tumors were initiated with a single injection of NNK (100 mg/kg/body weight), and monitored for 18 weeks. The dependability of the pellet treatment was assessed in a pharmacokinetic study, in which mice treated with time-release pellets showed a cumulative average serum level of 1.68 mg/ml over a 30-day period (Lantry et al., 2000b). In this earlier study, FTI-276 was delivered daily from time-release


6259

a

Tumor Multiplicity

25

NNK 20 15

NNK+FTI-276

*

10 5

*

0

Complete treatment

Tumor load mm3

b

Late treatment

16

NNK

14

NNK+FTI-276

12 10 8 6 4

* *

2 0

Complete treatment

Late treatment

Figure 1 Chemopreventive and therapeutic effects of FTI-276 in the A/J mouse lung tumor model. For complete chemoprevention protocol, FTI pellets were given 7 days before NNK treatment until the end of bioassays. For late treatment protocol, mice were given NNK and monitored for 18 weeks. FTI-276 was given 18 weeks after NNK administration. The bioassays were terminated at the fourth week of treatment with FTI-276. (a) Chemopreventive effect of FTI-276 on lung tumor multiplicity. (b) Chemopreventive effect of FTI-276 on lung tumor load (error bars indicate s.d., *Po0.05)

pellets for 30 days to the animals after they had been initiated with NNK for 18 weeks. As summarized in Figure 1, and our published data (Lantry et al., 2000b), FTI-276 delivered by time-release pellet reduced the lung tumor incidence by 42% (Po0.01) (Lantry et al., 2000b), the multiplicity by 60% (Po0.01), and the tumor load by 58% (Po0.05) (Lantry et al., 2000b, Figure 1). In the present study, we examined the effect of complete chemoprevention treatment protocol for FTI276 and the results are summarized in Figure 1. FTI-276 pellets designed to deliver a daily dose equivalent to 40 mg/kg/body weight over 30 days were implanted subcutaneously into 5-week-old male A/J mice, once a month for four consecutive months. The mice were initiated with a tobacco specific carcinogen NNK during weeks 2 and 3. The multiplicity of tumors was significantly reduced in the treatment group with a reduction of 41.7% over control (Po.005) (Figure 1a). In control mice, the average tumor number per mouse was 17.7 7 6.0, with the majority of the tumors being in the greater than 1.0 mm size group. The mice receiving FTI-276 demonstrated an average of 10.375.4 tumors, with the majority of the tumors in the 0.5 mm size

group. Furthermore, the total tumor load (volume of tumor per mouse, mm3) in the treatment group was strongly reduced by 79.4% (Po0.0001) with mice receiving treatment having an average tumor load of 1.7971.06 mm3, compared to controls 8.6874.73 mm3 (Figure 1b). These results indicate that early treatment of A/J mice with FTI-276 is just as effective as late treatment, but did not prove to be more effective, suggesting that FTI-276 acts mainly by suppressing tumor growth during the stage of tumor progression. Characterization of a transgenic mouse lung tumor model with alterations in p53 and Ink4a/Arf p53 transgenic mice carrying a dominant-negative mutation (Ala 135 Val) were crossbred with heterozygous Ink4a/Arf-deficient mice. F1 mice with four genotypes, p53 þ / þ Ink4a/Arf þ / þ , p53 þ /Ink4a/Arf þ / þ , p53 þ / þ Ink4a/Arf þ /, and p53 þ /Ink4a/Arf þ /, were initiated with a single dose of benzo(a)pyrene [B[a]P] (100 mg/kg body weight) or vehicle control (tricapryline), and monitored for 30 weeks. In the tricaprylintreated group, lung tumor multiplicity was similar in all groups of mice with various genotypes (Figure 2). In B[a]P treated groups, B[a]P produced an average of 12– 13 tumors in wild-type (p53 þ / þ Ink4a/Arf þ / þ ) and heterozygous Ink4a/Arf-deficient (p53 þ / þ Ink4a/Arf þ /) mice, and 21–23 tumors in mice with a p53 dominant negative mutation and in compound mice with both the p53 mutation and the INK4A deficiency (Figure 2a). There is no significant difference in tumor multiplicity between wild-type and heterozygous Ink4a/Arf-deficient mice. However, there is a significant increase in tumor size in heterozygous Ink4a/Arf-deficient (p53 þ / þ Ink4a/ Arf þ /) mice compared with that from wild-type (p53 þ / þ Ink4a/Arf þ / þ ) mice after 30 weeks exposure to B[a]P (Po0.01) (Figure 2a). In addition, a synergistic effect of dominant-negative p53 combined with heterozygous loss of Ink4a/Arf was significantly more striking in the B[a]Ptreated groups causing an B30-fold increase in total tumor volume (Figure 2b). More notably, the majority of the large lung tumors (B80%) from the p53 þ /Ink4a/ Arf þ / mice were adenocarcinomas. In contrast, less than 10% of the lung tumors from the wild-type mice were adenocarcinomas, whereas B50% of the lung tumors from mice that were heterozygous Ink4a/Arfdeficient without the p53 mutation were adenocarcinomas. As shown in Figure 3 (b,d,f), mouse lung adenocarcinomas are composed of cells with varying degrees of differentiation with most cells relatively undifferentiated relative to lung adenomas (Figure 3a,c,e). There is a complete loss of normal alveolar architecture and the nuclear/cytoplasmic ratio is increased in the carcinomas (Figure 3b,d,f). Nuclear crowding and cytologic atypia are present and there is heterogeneity of growth patterns. Invasion into adjacent bronchioles or vessels is also observed (Figure 3b). These results suggest that dominant-negative p53 and Ink4a/Arf deficiency synergistically promote lung tumor progression. More importantly, this is the first transgenic mouse model that involves three major genetic Oncogene


6260

a Tumor Multiplicity

30 25

*

Vehicle

*

B[a]P

20

*

15 10 5 0

p53+/+p16+/+

p53+/+p16+/-

p53+/-p16+/+

p53+/-p16+/-

b Tumor load mm3

700 600 500

*

Vehicle B[a]P

400 300

*

Frequency and type of mutations in the Kras2 gene in tumors treated with and without FTI

200 100

*

0 p53+/+p16+/+

p53+/+p16+/-

p53+/-p16+/+

p53+/-p16+/-

Figure 2 Characterization of a transgenic mouse lung tumor model with alterations in p53 and Ink4a/Arf. p53 transgenic mice carrying a dominant-negative mutation (Ala 135 Val) were crossbred with heterozygous Ink4a/Arf-deficient mice. F1 mice with four genotypes, p53 þ / þ Ink4a/Arf þ / þ , p53 þ /Ink4a/Arf þ / þ , p53 þ / þ Ink4a/Arf þ /, and p53 þ /Ink4a/Arf þ / were initiated with a single dose of B[a]P (100 mg/kg body weight) or vehicle control (tricapryline), and monitored for 30 weeks. (a) Effect of genotypes on lung tumor multiplicity. (b) Effect of genotypes on lung tumor load. (p16 is used to represent Ink4a/Arf; error bars indicate s.d., *Po0.01). Approximately equal numbers of males and females were used, with no significant difference in tumor multiplicity between the sexes. In vehicle control groups, there are 20 p53 þ / þ Ink4a/Arf þ / þ , 18 p53 þ /Ink4a/Arf þ / þ , 21 p53 þ / þ Ink4a/Arf þ /, and 18 p53 þ /Ink4a/Arf þ / mice. In B[a]P treated groups, there are 48 p53 þ / þ Ink4a/Arf þ / þ , 43 p53 þ / þ Ink4a/Arf þ /, 35 p53 þ /Ink4a/ Arf þ / þ , and 19 p53 þ /Ink4a/Arf þ / mice

alterations (Kras2, p53, and Ink4a/Arf) commonly found in human lung NSCLC. Anti-cancer effect of FTI L-778,123 on B[a]P-induced lung tumorigenesis in p53 transgenic or/and heterozygous Ink4a/Arf-deficient mice In the second set of experiments, a new FTI, L-778,123, was examined for chemopreventive efficacy on lung tumor development in A/J mice and transgenic mice with a dominant-negative p53 mutation and/or heterozygous deletion of the Ink4a/Arf. Mice with four genotypes, p53 þ / þ Ink4a/Arf þ / þ , p53 þ /Ink4a/Arf þ / þ , p53 þ / þ Ink4a/Arf þ /, and p53 þ /Ink4a/Arf þ /, were initiated with a single dose of B[a]P (100 mg/kg body weight) or vehicle control (tricapryline), and monitored for 20 weeks. Starting at 20 weeks postinitiation with B(a)P, L-778,123 was administered daily, 5 days/week for 10 weeks. As shown in Figure 4a, L-778,123 delivered by subcutaneous injection reduced the lung Oncogene

tumor multiplicity by 43–47% (Po0.01) in all four genotypes of mice when compared with B[a]P control groups. The effects on tumor volume were more interesting and surprising (Figure 4b). Because wildtype mice (p53 þ / þ Ink4a/Arf þ / þ ) or mice with an Ink4a/ Arf deficiency (p53 þ / þ Ink4a/Arf þ /) genotypes had a decrease in total tumor volume of roughly 50% (Po0.01) (Figure 4b). In contrast, mice with a p53 mutation (p53 þ /Ink4a/Arf þ / þ or p53 þ /Ink4a/Arf þ /) had rough 75% reduction in tumor volume (Figure 4b). The significance of percent changes were calculated using a Z-statistics for proportions and the data showed that these reductions on tumor load in p53 þ /Ink4a/ Arf þ / þ and p53 þ /Ink4a/Arf þ / mice are significantly different from the tumor reductions in p53 wild-type mice (Po0.05, Po0.01, respectively). This result indicates that L-778,123 inhibits B[a]P-induced mouse lung tumor growth more effectively in mice with defects in p53 and Ink4a/Arf.

As shown in Table 1, mutation analysis showed that eight out of 11 lung tumors from wild-type mice treated with B[a]P, and 10 out of 11 lung tumors from the B[a]P/ L-778,123 group were positive for 12th codon Kras2 mutations. The frequency and type of mutations in the Kras2 gene was similar in the group treated with L-778, 123 to those in the B[a]P control group. The result is consistent with our previous observation that eight out of nine tumors from tumor-bearing animals treated with FTI-276 and all of the tumors (7/7) from the control group were positive for Kras2 mutations at the 12th codon of the Kras2 gene without a significant difference in the frequency and type of mutations in the Kras2 gene between the group treated with FTI-276 as those in the control group (Lantry et al., 2000b). Thus, there is no significant difference in the incidence and types of mutations between tumors from animals treated with an FTI (either FTI-276 or L-778,123) and tumors from control animals.

Discussion There are three striking aspects in the present study. First, the results from this study clearly demonstrate the efficacy of FTIs against lung cancer that contains a Kras2 mutation, a p53 mutation, and/or Ink4a/Arf deletion. This is important since there have been concerns that the post-translational modification of Kras2, unlike Ha-ras and N-ras, cannot be completely blocked by inhibition of farnesylation. Nonetheless, lung tumors with Kras2 mutations are susceptible to FTIs. This observation seems to support the evidence that the FTI’s primary targets may be RhoB (Du et al., 1999) and the proteins involved in the PI3K/AKT2 pathway (Jiang et al., 2000) rather than ras proteins themselves. Second, the efficacy of FTI –276 was similar when treatment was initiated either at the time of NNK


6261

a

b

c

d

e

f

Figure 3 B[a]P-induced lung tumors in A/J mice carrying a dominant negative p53 and heterozygous deletion of Ink4a/Arf. Photomicrographs of a lung adenoma at 4 (a), 20 (c), and 40 (e) magnifications from B[a]P-treated p53 þ / þ Ink4a/Arf þ / þ mice, and photomicrographs of a lung adenocarcinoma at 4 (b), 20 (d) and 40 (f) magnifications from B[a]P-treated p53 þ /Ink4a/ Arf þ / mice. Neoplastic cells in alveolar bronchiolar adenomas (e) are monomorphic, where neoplastic cells in alveolar bronchiolar carcinomas (f) are pleomorphic. Nuclei vary in size and contain multiple prominent nucleoli

administration or when treatment was initiated 18 weeks after NNK. These results imply that the primary inhibitory effects of FTIs are not during the phases of tumor initiation or early tumor progression, but rather on later phases of tumor progression. One should be aware that in the standard A/J model, virtually all of the tumors observed are adenomas. Third, in examining the therapeutic effects of L-778,123, we observed that the agent was more effective in inhibiting lung tumors with a dominant-negative p53 mutation or in compound mice with both a dominant-negative p53 mutation and heterozygous deletion of the Ink4a/Arf locus. This is quite important since mutations in p53 are commonly associated with resistance to therapy. Furthermore, p53 is perhaps the most commonly mutated gene in human tumors and is mutated early in tumor progression in certain cancers such as head and neck cancer and lung

cancer. Agents that are effective in the face of this mutation are particularly appealing. Previous reports on the efficacy of FTIs in cell culture vary widely between investigators. Even studies using the same Kras2 mutant cell line report an opposite effect when treated with FTIs. For example, Gu et al. (1999) found that in the HCT116 colon carcinoma cell line, an FTI (A-170634) inhibited Ras processing and blocked anchorage-dependent and independent growth. Whereas in another study with the same cell line, Nagasu et al. (1995), found that the FTI B965 did not block tumor growth, nor processing of Kras2. In our own in vitro studies, we found that FTI-276 exhibited growth inhibition in mouse lung epithelial cell lines: one contained an activating Kras2 mutation, while another had a Kras2 mutation and a p53 mutation. Employing anchorage-dependent growth as a primary end point, we Oncogene


6262

a Tumor Multiplicity

30 25

B[a]P B[a]P+L-778,123

20

*

*

15 10

*

*

5 0 p53+/+p16+/+

p53+/+p16+/-

p53+/-p16+/+

p53+/-p16+/-

b Tumor load mm3

700 600 500

B[a]P B[a]P+L-778,123

400 300

*

200 100 0

* p53+/+p16+/+

* p53+/+p16+/-

* p53+/-p16+/+

p53+/-p16+/-

Figure 4 Treatment with FTI (L-778,123) induced inhibition of lung tumor progression in A/J mice carrying a dominant-negative p53 or/and heterozygous deficient of Ink4a/Arf tumor suppressor genes. (a) Antitumor effect of L-778,123 on lung tumor multiplicity. (b) Antitumor effect of L-778,123 on lung tumor load. (p16 is used to represent Ink4a/Arf; error bars indicate s.d., *Po0.01). Approximately equal numbers of males and females were used, with no significant difference in tumor multiplicity between the sexes. In B[a]P-treated groups, there are 48 p53 þ / þ Ink4a/Arf þ / þ , 43 p53 þ / þ Ink4a/Arf þ /, 35 p53 þ /Ink4a/Arf þ / þ , and 19 p53 þ /Ink4a/ Arf þ / mice. In B[a]P-and L-778,123-treated groups, there are 21 p53 þ / þ Ink4a/Arf þ / þ , 19 p53 þ / þ Ink4a/Arf þ /, 21 p53 þ /Ink4a/ Arf þ / þ , and 10 p53 þ /Ink4a/Arf þ / mice

found that the dose required to reach the ED50 is greater for an FTI than a GGTI (Lantry et al., 2000a). These findings demonstrate the difficulties in interpreting results of in vitro studies where questions of pharmacokinetics, stromal interactions, and/or potentially unique combinations of genotypic changes routinely observed in vivo may not come into play. We have previously demonstrated that FTI-276 is chemopreventive in A/J mice using a late treatment protocol (Lantry et al., 2000b). The mechanism(s) includes a decrease in proliferation and an increase in apoptosis in the treated tumors, but with minimal effects on normal lung tissue (Lantry et al., 2000a). FTI-276 did

not strongly reduce the initiation of tumors, since the multiplicity of tumors was only slightly reduced. However, there was a significant difference in tumor size between the treated and untreated controls consistent with the decreased proliferation and increased apoptosis we previously observed. The present study carefully and systematically investigated the chemopreventive efficacy of two FTIs in the best available in vivo mouse lung tumor models. A complete chemoprevention study was performed with FTI-276 in A/J mice, and a significant inhibition of lung tumorigenesis was observed. This result is consistent with the concept that FTIs act primarily on tumor progression. In a transgenic mouse model with a dominant-negative p53 mutation and/or heterozygous deletion of the Ink4a/ Arf, we found that L-778,123 was not only effective in preventing lung tumorigenesis in mice with all genotypes but also exhibited efficacy (B75% reduction in lung tumor volume) in animals with either a dominantnegative p53 or both a p53 mutation and heterozygous deletion of Ink4a/Arf. There is increasing evidence that FTIs induce regression of tumors by multiple mechanisms, including the activation of apoptotic pathways or by cell cycle alterations (Barrington et al., 1998; Brassard et al., 2002; Crespo et al., 2002). Barrington et al. (1998) demonstrated that FTI could induce dramatic regression of mammary and salivary carcinomas in mouse mammary tumor virus (MMTV)-v-Ha-ras transgenic and p53-deficient mice. They also showed that the inhibitory effect was mediated by both the induction of apoptosis and an increase in G1 with a corresponding decrease in the S-phase fraction in a p53 independent manner (Barrington et al., 1998). A study of two human lung cancer cell lines showed that the ability of FTIs to inhibit bipolar spindle formation is not dependent on p53 mutation status since both wild-type and mutant p53 cells were equally affected (Crespo et al., 2002). Taken together, these results demonstrated that the ability of FTI to inhibit tumor growth and/or accumulate cells in G1 is independent of a p53 mutation or heterozygous loss of p16 or both. In fact, we have shown that FTIs exhibited an increased efficacy against lung progression in mice with a dominant-negative p53 or both a p53 mutation and heterozygous deletion of Ink4a/Arf. This observation is consistent with the finding that p53-deficient cells were sensitized to the effects of DNA-damaging agents as a result of the failure to induce expression of the cyclin-dependent kinase inhibitor p21 (Bunz et al., 1999). However, the exact mechanism for the increased efficacy of FTIs

Table 1 Analysis of Kras2 codon 12 mutations in mouse lung tumors from animals treated with FTI Genotype A/J A/J A/J A/J 1

(wild type) (wild type) (p53+/+Ink4a/Arf+/+) (p53+/+Ink4a/Arf+/+)

Data taken from Lantry et al. (2000b).

Oncogene

Treatment NNK NNK+FTI-276 B[a]P B[a]P+L-778,123

Frequency 1

7/7 (100%) 8/9 (89%)1 8/11 (72.7%) 10/11 (90.9%)

GGT-GAT

GGT-CGT

7 8 7 10

0 0 1 0


6263

against lung cancer progression in p53 and p53/Ink4a/ Arf is not clear at present. Finally, we described a new mouse model for human adenocarcinomas that includes alterations in commonly altered genes, for example, Kras2, p53, and Ink4a/Arf. Mutation or deletions of the Kras2, p53, and Ink4a/Arf tumor suppressor genes are the most common genetic defect detected in human lung cancer. p53 transgenic mice carrying a dominant-negative mutant p53val135/wt were crossed with heterozygous Ink4a/Arf-deficient mice, on an A/J background. When these mice were treated with lung-specific chemical carcinogens, lung tumors were generated that contain mutated Kras2, p53 or/and loss of Ink4a/Arf genes. In agreement with our previous studies (Zhang et al., 2000, 2002b), we found that mice with heterozygous deficiency of Ink4a/Arf exhibited a striking enhancement in lung tumor progression 30 weeks after carcinogen treatment when compared to their respective wild-type mice. A germline p53 mutation increased both tumor number and tumor volume 30 weeks postinitiation. More notably, our results showed that most of the tumors in compound p53 þ /Ink4a/Arf þ / mice were larger and histologically more advanced adenocarcinomas. Taken together, these results strongly suggest that germline p53 mutation and deficiency of Ink4a/Arf exhibit an enhanced susceptibility to lung tumorigenesis, and mutant p53 acts synergistically with the deletion of Ink4a/Arf to accelerate the development of undifferentiated malignant lung tumors. The models described here would appear particularly relevant to examining potential chemopreventive or chemotherapeutic agents employing an in situ model with a number of the known mutations relevant to lung adenocarcinomas in humans.

Materials and methods Chemicals NNK was purchased from Chemsyn Science Laboratories (Lenexa, KS, USA). B[a]P was purchased from Aldrich (Milwaukee, WI, USA). Tricaprylin was purchased from Sigma (St Louis, MO, USA). Custom-made time-release pellets of FTI-276 were prepared by Innovative Research of America (Tampa, FL, USA), FTI (L-778,123) was provided by Merck Research Laboratories (West Point, PA, USA). Animal models A/J mice were obtained from the Jackson Laboratories (Bar Harbor, ME, USA). The N10(A/J UL53–3) mice are 10 times backcross UL53– 3 mice (p53 transgenic mice) to the A/J strain. UL53–3 mice (on FVB background) carrying a 135val p53 mutation were obtained from National Institute of Environmental Health Sciences, Research Triangle Park, NC, USA. The animals were paired to set up breeding colonies for production of (A/J UL53–3) F1 mice and further to cross breed with A/J mice. After 10 generations of backcrossing to A/J mice, heterozygote N10(A/J UL53–3) p53 þ / mice were used as breeder to generate the N10(A/J UL53–3) N6(A/J Ink4a/ArfKO) F1 mice.

Ink4a/Arf nullizygous mice were obtained from Dana Faber Cancer Institute, (Boston, MA, USA) and have been rederived onto the A/J background through six generations of backcrossing. The A/J mouse is highly susceptible to lung tumor induction by mouse carcinogens. Since the tumor spectrum found in several knockout mice was similar including fibrosarcomas, lymphomas, and other types of sarcomas, the mouse background may partially contribute to the observed tumor spectrum. By using the A/J mouse as background, we are hoping that most of the resulting tumors developed in these mice will be lung tumors as seen in p53 transgenic mice (Zhang et al., 2000). After six generations of backcrossing to A/J mice, heterozygote N6(A/J Ink4a/Arf KO) mice were used as breeder to generate the N10(A/J UL53–3) N6(A/J Ink4a/Arf KO) F1 mice. Theoretically, N10(A/J UL53–3) mouse is 99.9% A/J background and N6(A/J Ink4a/Arf KO) mouse is 98.44% A/J background (Markel et al., 1997). Genotyping Tail clippings from each N10(A/J UL53–3) N6(A/J Ink4a/Arf KO) F1 mice were homogenized and incubated overnight at 371C in lysis solution (pronase 0.4 mg/ml, 10% sodium dodecylsulfate (w/v), 10 mm Tris, 400 mm NaCl, and 2 mm EDTA) followed by phenol–chloroform extraction and precipitation with ice-cold alcohol. To distinguish the p53 transgenic mice from nontransgenic littermates, DNA was extracted from the F1 mice tail by the proteinase K-sodium dodecylsulfate method, and was genotyped using polymerase chain reaction (PCR)–RFLP method as described previously (Zhang et al., 2000). Ink4a/Arf genotypes of the mice were tested by (PCR) method as described previously (Zhang et al., 2002b). Lung tumor bioassay in A/J mice For complete chemoprevention protocol, 5-week-old A/J mice were maintained in an environmentally controlled room (24711C, 12/12 h light/dark cycle), housed four per cage, in plastic cages with hardwood bedding and dust covers, and received AIN-76A Purified Diet, 100000 (Dyets Inc., Bethlechem, PA, USA) and water ad libitum. The mice were randomized into two groups of 15 animals each (Control and Treatment), prior to the beginning of the study. Two sets of pellets were designed to give 40 mg/kg/body weight/day, based on an average starting body weight of 22.5 g for the first month, and 27.5 g for months 2–4. The mice were anesthetized with one i.p. injection of a sterile solution of Ketamine (80 mg/ kg) and xylazine (16 mg/kg). The pellets were inserted under aseptic conditions between the ear and shoulder blade, alternating sides each month, and the incision was closed with a single sterile wound clip. The first dose of NNK was given on day 7, i.p. at 100 mg/kg/body weight. The second i.p. dose of NNK was delivered on Day 14. The health and weight of the animals were monitored on a biweekly basis for the duration of the experiment. Animals were terminated by CO2 asphyxiation on day 120, from the start of the chemoprevention regimen. For late treatment protocol, 4- week-old, male A/J mice were given one i.p. injection of NNK, at 100 mg/kg body weight in PBS, and monitored for 18 weeks. On the first day of FTI-276 treatment (18 weeks after NNK administration), the animals were randomized into treatment groups as follows: Group 1, 12 mice treated with FTI-276 time-release pellet; Group 2, 12 mice treated with pellet matrix control. The custom-made time-release pellets contained the equivalent concentration of FTI-276 to deliver 1.25 mg/day for 30 days, Oncogene


6264 based on 50 mg/kg /day to a 25 g mouse (Group 1), or matrix alone (Group 2). The pellets were inserted under sterile conditions, one per animal, s.c., at the right dorsal base of the neck. At the end of the fourth week of treatment with FTI276, the mice were euthanized by CO2 asphyxiation. To maintain the integrity of the DNA and RNA in fresh tumors, the lungs were harvested and evaluated by one investigator so that the tumors could be collected and snap frozen as quickly as possible. When available, 1–5 fresh tumors and a small portion of uninvolved lung were collected immediately upon termination and snap frozen. Blood samples were obtained from a representative number of mice for evaluation of circulating FTI-276 by HPLC. The lungs were fixed in Tellyesniczky’s (90% ethanol (70% v/v), 5% glacial acetic acid, 5% formalin (10% v/v buffered formalin)) overnight, and then kept in 70% ethanol for evaluation prior to paraffin embedding. The fixed lungs were evaluated by at least two investigators, under a dissecting microscope to obtain the fixed surface tumor count, and individual tumor size was measured for volume calculation based on the following formula: (mm3 ¼ V ¼ 4/3pr3). The total tumor count was obtained by adding the fixed and frozen tumor count. Unpaired Student’s t-tests were run to test for significance. Lung tumor bioassay on A/J mice harboring a dominant-negative p53 and/or heterozygous deletion of the Ink4a/Arf Six-week-old N10(A/J UL53–3) N6(A/J p16KO) F1 mice were randomized into control and treatment groups based on the genotype and gender. Control mice received a single dose of 0.1 ml tricaprylin as vehicle control. Treatment groups received a single i.p. dose of B[a]P (100 mg/kg body weight). At 20 weeks after B[a]P exposure, FTI treatment groups received 0.1 ml vehicle controls and groups 9–12 were given 0.1 ml FTI s.c., (70 mg/kg/day) once daily, 5 days/week for 10 weeks. FTI was dissolved in an aqueous solution containing NaCl to adjust the osmolarity to 270–300 mosmol and sodium citrate to give a pH of 5.4. The vehicle was water similarly adjusted with NaCl and sodium citrate to achieve the desired osmolarity and pH. All animals were killed 30 weeks after exposure to carcinogens (B[a]P) by CO2 asphyxiation. All animals were observed daily for clinical signs of ill health.

Moribund mice were terminated by CO2. All mice were killed or found dead were necropsied, and their tissues were fixed in 10% neutral-buffered formalin. Analysis of Kras2 mutations by PCR-direct sequencing DNA was isolated from lung tumors using the TRIzol reagent (Gibco BRL). The sequences of PCR primers for Kras2 exons 1 and 2 have been described previously (You et al., 1989). Briefly, a 100 ml reaction mixture containing 100 ng genomic DNA, 10 mm Tris-HCl, pH 8.5, 50 mm KCL, 2.5 mm MgCl2, 100 mm of each deoxyribonucleoside triphosphate (dATP, dCTP, dGTP, dTTP), 1.0 U of Taq DNA polymerase (Promega, Madison, WI, USA), and 40 pmol of each primer was overlaid with sterile mineral oil and subjected to 35 cycles of PCR amplification. Each cycle consisted of 1 min each at 941C, 551C, and 721C. The direct sequencing of PCR products was also carried out as described previously (You et al., 1989). Statistical analysis Students t-test was used to determine the differences in the number and size of lung tumors per mouse between treatment and control mice. Fisher’s exact test was used to determine the difference in the incidence. The significance of percent changes were calculated using Z-statistics for proportions pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi (Z ¼ ð^ p1 p^2 = p^ð1 p^Þð1=n1 þ 1=n2 ÞÞ), where p^1 is the proportional reduction due to FTI in the wild-type group, p^2 is the proportional reduction due to FTIs in the transgenic groups, p^ is the proportional reduction due to FTIs regardless of genotype, n1 is the number of wild-type animals, and n2 the number of transgenic animals. P-values corresponding to Zstatistics were looked up in a table. Acknowledgements We are grateful to Dr Roger W Wiseman for providing the original UL53-3 mice and Dr Ronald A DePinho for the original Ink4a/Arf/ mice and to Merck and Co., Inc. for supplying L-778,123. We acknowledge Dr William J Lemon and Sandya Liyanarachchi for statistical assistance. This work is supported by National Institutes of Health; Grants R01CA58554 and R01CA78797.

References Barrington RE, Subler MA, Rands E, Omer CA, Miller PJ, Hundley JE, Koester SK, Troyer DA, Bearss DJ, Conner MW, Gibbs JB, Hamilton K, Koblan KS, Mosser SD, O’Neill TJ, Schaber MD, Senderak ET, Windle JJ, Oliff A and Kohl NE. (1998). Mol Cell. Biol., 18, 85–92. Brassard DL, English JM, Malkowski M, Kirschmeier P, Nagabhushan TL and Bishop WR. (2002). Exp. Cell. Res., 15, 138–146. Bos JL. (1989). Cancer Res., 49, 4682–4689. Bunz F, Hwang PM, Torrance C, Waldman T, Zhang Y, Dillehay L, Williams J, Lengauer C, Kinzler KW and Vogelstein B. (1999). J. Clin. Invest., 104, 263–269. Casey PJ, Solski PA, Der CJ and Buss JE. (1989). Proc. Natl. Acad. Sci. USA, 86, 8323–8327. Chiba I, Takahashi T, Nau MM, D’Amico D, Curiel DT, Mitsudomi T, Buchhagen DL, Carbone D, Piantadosi S and Koga H. (1990). Oncogene, 5, 1603–1610. Crespo NC, Delarue F, Ohkanda J, Carrico D, Hamilton AD and Sebti SM. (2002). Cell Death Differ., 9, 702–709. Du W, Lebowitz PF and Prendergast GC. (1999). Mol. Cell. Biol., 19, 1831–1840. Oncogene

Gibbs JB, Pompliano DL, Mosser SD, Rands E, Lingham RB, Singh SB, Scolnick EM, Kohl NE and Oliff A. (1993). J. Biol. Chem., 268, 7617–7620. Gu WZ, Tahir SK, Wang YC, Zhang HC, Cherian SP, O’Connor S, Leal JA, Rosenberg SH and Ng SC. (1999). Eur. J. Cancer, 35, 1394–1401. Hancock JF, Magee AI, Childs JE and Marshall CJ. (1989). Cell, 57, 1167–1177. Herzog CR, Lubet RA and You M. (1997). J. Cell Biochem., 28/29S, 49–63. Herzog CR, Soloff EV, McDoniels AL, Tyson FL, Malkinson AM, Haugen-Strano A, Wiseman RW, Anderson MW and You M. (1996). Oncogene, 13, 1885–1891. Jiang K, Coppola D, Crespo NC, Nicosia SV, Hamilton AD, Sebti SM and Cheng JQ. (2000). Mol. Cell. Biol., 20, 139–148. Khosravi-Far R and Der CJ. (1994). Cancer Metastasis Rev., 13, 67–89. Kohl NE, Omer CA, Conner MW, Anthony NJ, Davide JP, deSolms SJ, Giuliani EA, Gomez RP, Graham SL and Hamilton K et al. (1995). Nature Med., 1, 792–797. Landis SH, Murray T, Bolden S and Wingo PA. (1998). CA Cancer J. Clin., 48, 6–29.


6265 Lantry LE, Zhang Z, Crist KA, Wang Y, Hara M, Zeeck A, Lubet RA and You M. (2000a). Exp. Lung Res., 26, 773–790. Lantry LE, Zhang Z, Yao R, Crist KA, Wang Y, Ohkanda J, Hamilton AD, Sebti SM, Lubet RA and You M. (2000b). Carcinogenesis, 21, 113–116. Lerner EC, Qian Y, Blaskovich MA, Fossum RD, Vogt A, Sun J, Cox AD, Der CJ, Hamilton AD and Sebti SM. (1995a). J. Biol. Chem., 270, 26802–26806. Lerner EC, Qian Y, Hamilton AD and Sebti SM. (1995b). J. Biol. Chem., 10, 26770–26773. Lubet RA, Zhang Z, Wiseman RW and You M. (2000). Exp. Lung Res., 26, 581–593. Mangues R, Corral T, Kohl NE, Symmans WF, Lu S, Malumbres M, Gibbs JB, Oliff A and Pellicer A. (1998). Cancer Res., 58, 1253–1259. Markel P, Shu P, Ebeling C, Carlson GA, Nagle DL, Smutko JS and Moore KJ. (1997). Nat. Genet., 17, 280–284. McDoniels-Silvers AL, Herzog CR, Tyson FL, Malkinson AM and You M. (2001). Exp. Lung Res., 27, 297–318. Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger PC, Baylin SB and Sidransky D. (1995). Nat. Med., 1, 686–692. Minna JD, Pass H, Glatstein E and Ihde DC. (1989). Cancer Principles and Practice of Oncology. J.B. Lippiencott: Pjiladelphia, pp. 591–687. Nagasu T, Yoshimatsu K, Rowell C, Lewis MD and Garcia AM. (1995). Cancer Res., 55, 5310–5314. Omer CA, Chen Z, Diehl RE, Conner MW, Chen HY, Trumbauer ME, Gopal-Truter S, Seeburger G, Bhimnathwala H, Abrams MT, Davide JP, Ellis MS, Gibbs JB, Greenberg I, Koblan KS, Kral AM, Liu D, Lobell RB,

Miller PJ, Mosser SD, O’Neill TJ, Rands E, Schaber MD, Senderak ET, Oliff A and Kohl NE. (2000). Cancer Res, 60, 2680–2688. Prendergast GC and Oliff A. (2000). Semin. Cancer Biol., 10, 443–452. Reuter CW, Morgan MA and Bergmann L. (2000). Blood, 96, 1655–1669. Rodenhuis S and Slebos RJC. (1992). Cancer Res., 52, 2665s–2669s. Sebti SM and Hamilton AD. (2000). Expert Opin Investig Drugs, 9, 2767–2782. Sellers TA, Potter JD, Bailey-Wilson JE, Rich SS, Rothschild H and Elston RC. (1992). Cancer Res., 52 (9 Suppl), 2694s–2697s. Sun J, Qian Y, Hamilton AD and Sebti SM. (1995). Cancer Res., 55, 4243–4243. Sun J, Qian Y, Hamilton AD and Sebti SM. (1998). Oncogene, 16, 1467–1473. Tamanoi F. (1993). Trends Biochem. Sci., 18, 349–353. Xu HJ, Hu SX, Cagle PT, Moore GE and Benedict WF. (1991). Cancer Res., 51, 2735–2739. You M, Candrian U, Maronpot RR, Stoner GD and Anderson MW. (1989). Proc. Natl. Acad. Sci., 86, 3070–3074. Zhang Z, Liu Q, Lantry LE, Wang Y, Kelloff GJ, Anderson MW, Wiseman RW, Lubet RA and You M. (2000). Cancer Res., 60, 901–907. Zhang Z, Li J, Lantry LE, Wang Y, Wiseman RW, Lubet RA and You M. (2002a). Cancer Res., 62, 3024–3029. Zhang Z, Wang Y, Herzog CR, Liu G, Lee HW, DePinho RA and You M. (2002b). Oncogene, 29, 5960–5966.

Oncogene