FOXO3 encodes a carcinogen-activated transcription factor frequently deleted in early-stage lung adenocarcinoma.

The FOXO family of transcription factors elicits cell cycle arrest, apoptosis, and resistance to various physiologic and pathologic stresses relevant to sporadic cancer, such as DNA damage and oxidative stress. Although implicated as tumor suppressors, FOXO genetic inactivation has not been observed in human cancer. In an investigation of the two major types of non-small cell lung cancer, here, we identify the FOXO3 gene as a novel target of deletion in human lung adenocarcinoma (LAC). Biallelic or homozygous deletion (HD) of FOXO3 was detected in 8 of 33 (24.2%) mostly early-stage LAC of smokers. Another 60.6% of these tumors had losses of FOXO3 not reaching the level of HD (hereafter referred to as sub-HD). In contrast, no HD of FOXO3 was observed in 19 lung squamous cell carcinoma. Consistent with the deletion of FOXO3 were corresponding decreases in its mRNA and protein levels in LAC. The potential role of FOXO3 loss in LAC was also investigated. The carcinogen (+)-anti-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) is strongly implicated as a cause of human lung cancer. Here, we show that FOXO3a is functionally activated and augments the level of caspase-dependent apoptosis in cells exposed to this DNA-damaging carcinogen. These results implicate FOXO3 as a suppressor of LAC carcinogenesis, a role frequently lost through gene deletion.

FOXO3a elicits a pro-apoptotic transcription program and cellular response to human lung carcinogen nicotine-derived nitrosaminoketone (NNK).

Long-term carcinogen exposure exerts continuous pressure on key mechanisms that repair or eliminate carcinogen-damaged cells giving rise to selective failures that contribute to lung cancer. FOXO3a is a transcription factor that elicits a protective response to diverse cellular stresses. Although implicated as a tumor suppressor, its role in sporadic cancer is uncertain. We recently observed that FOXO3a gene inactivation occurs frequently in carcinogen-induced lung adenocarcinoma (LAC). This suggests that FOXO3a may play a role in LAC suppression by eliciting a protective response to carcinogenic stress. Here we investigated this possibility by examining the role of FOXO3a in the cellular response to nicotine-derived nitrosaminoketone (NNK), a lung carcinogen implicated as a cause of human LAC. We show that restoration of FOXO3a in FOXO3a-deficient LAC cells increases sensitivity to apoptosis caused by a DNA-damaging intermediate of NNK. Prior to this cellular outcome, FOXO3a is functionally activated and mediates a large-scale transcription program in response to this damage involving a significant modulation of 440 genes. Genes most significantly represented in this program are those with roles in cell growth and proliferation>protein synthesis>gene expression>cell death>cell cycle. The results of this study show that FOXO3a directs an anti-carcinogenic transcription program that culminates in the elimination of carcinogen-damaged cells. This suggests that FOXO3a is a potential suppressor of carcinogenic damage in LAC.

FoxO3a gene is a target of deletion in mouse lung adenocarcinoma.

Lung adenocarcinomas (LAC) of smokers and never-smokers differ from one another in epidemiology, and clinical and molecular characteristics. The pathogenetic differences between these tumors are potential biomarkers and therapeutic targets. Mouse carcinogenesis models of human LAC are proven tools applicable for the identification of these molecular changes. Allelic loss frequency on human chromosome 6q is higher in LAC of smokers compared with never smokers. We analyzed the orthologous region on mouse chromosome 10 and found this region similarly was a more frequent site of allelic loss in carcinogen-induced LAC compared with non-induced or spontaneous LAC. We then conducted high resolution quantitative PCR-based deletion mapping of this region and identified the FoxO3a gene as the focus of bi-allelic or homozygous deletion (HD). HDs were detected in 5 out of 15 (33.3%) LAC cell lines and in 6 out of 75 (8%) carcinogen-induced primary LAC. FoxO3a was exclusively affected by HD in 7 of the samples examined, as loss of both alleles did extend to the nearest flanking genes of FoxO3a. Deletion of FoxO3a, either by HD or subclonal loss was detected in 38 out of 75 (50.7%) of carcinogen-induced LAC in contrast to only 1 out of 10 (10%) of LAC of untreated mice. Several of the samples also were subjected to direct sequence analysis; however, no intragenic mutations were detected. These results implicate FoxO3a as a selective target of deletion in mouse LAC. Significant association with carcinogenic induction suggests that deletion of FoxO3a contributes to the development of carcinogen-initiated tumors.

Adenylosuccinate synthetase 1 gene is a novel target of deletion in lung adenocarcinoma.

Tobacco smoke consists of numerous carcinogens whose effect on lung tumor development includes the induction of mutations in key genes as well as the induction of chromosome instability (CIN). Consequently, carcinogen-induced mouse lung adenocarcinomas (LAC) display many more recurrent site- and chromosome-specific changes in DNA copy number compared with noninduced LAC. Here we identified the Adenylosuccinate synthetase 1 (Adss1) gene located on distal chromosome 12q as a focus of bi-allelic or homozygous deletion (HD) in LAC. HDs of Adss1 were detected in 10 out of 84 carcinogen-induced mouse primary LAC and mouse LAC cell lines. In only four of these cases did the deletions affect either Siva1 or Inverted-formin 2 (Inf2), which immediately flank the Adss1 locus, indicating that Adss1 is a selective target of deletion in LAC. Losses of Adss1 not meeting the quantitative threshold of HD were detected in 36 out of 84 (42.9%) of the mouse tumors and cell lines. A similar frequency of ADSS1 deletion was observed in human LAC cell lines, suggesting relevance in human lung cancer. Adss1 losses were also found to be significantly associated with a more extensive CIN phenotype in the primary mouse tumors. These results implicate ADSS1 inactivation as a novel somatic alteration in lung carcinogenesis, and suggest that its selective deletion in LAC may be triggered by CIN.

Effects of the environmental mammary carcinogen 6-nitrochrysene on p53 and p21(Cip1) protein expression and cell cycle regulation in MCF-7 and MCF-10A cells.

The environmental pollutant 6-nitrochrysene (6-NC) is a potent mammary carcinogen in rats; it is more potent than numerous classical mammary carcinogens such as benzo[a]pyrene (BaP). The mechanisms that account for the remarkable carcinogenicity of 6-NC remain elusive. Similar to BaP, 6-NC is also known to induce DNA damage in rodents and in human breast tissues. As an initial investigation, we reasoned that DNA damage induced by 6-NC may alter the expression of p53 protein in a manner that differs from other DNA damaging carcinogens (e.g. BaP). Using human breast adenocarcinoma MCF-7 cells and immortalized human mammary epithelial MCF-10A cells, we determined the effects of 6-NC on the expression of p53 protein and its direct downstream target cyclin-dependent kinase inhibitor p21(Cip1) as well as on the cell cycle progression. Western blot analysis demonstrated that treatments of MCF-7 and MCF-10A cells with 6-NC for 12, 24 or 48h did not increase the level of total p53 protein; however, an increase of p21(Cip1) protein and a commitment increase of G(1) phase were observed in MCF-10A cells but not in MCF-7 cells. Further studies using 1,2-dihydroxy-1,2-dihydro-6-hydroxylaminochrysene (1,2-DHD-6-NHOH-C), the putative ultimate genotoxic metabolite of 6-NC, was conducted and showed a significant induction of p53 (p<0.05) in MCF-7 cells; however, this effect was not evident in MCF-10A cells, indicating the varied DNA damage responses between the two cell lines. By contrast to numerous DNA damaging agents such as BaP which is known to stimulate p53 expression, the lack of p53 response by 6-NC imply the lack of protective functions mediated by p53 (e.g. DNA repair machinery) after exposure to 6-NC and this may, in part, account for its remarkable carcinogenicity in the mammary tissue.

Array CGH analysis reveals chromosomal aberrations in mouse lung adenocarcinomas induced by the human lung carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone.

Exposure to genotoxic carcinogens in tobacco smoke is a major cause of lung cancer. However, the effect this has on DNA copy number and genomic stability during lung carcinogenesis is unclear. Here we used bacterial artificial chromosome array-based comparative genomic hybridization to examine the effect of NNK, a potent human lung carcinogen present in tobacco smoke, on the major genomic changes occurring during mouse lung adenocarcinogenesis. Observed were significantly more gross chromosomal changes in NNK-induced tumors compared with the spontaneous tumors. An average of 5.6 chromosomes were affected by large-scale changes in DNA copy number per NNK-induced tumor compared with only 2.0 in spontaneous lung tumors (p = 0.017). Further analysis showed that gains on chromosomes 6 and 8, and losses on chromosomes 11 and 14 were more common in NNK-induced tumors (p

Carcinogen-specific targeting of chromosome 12 for loss of heterozygosity in mouse lung adenocarcinomas: implications for chromosome instability and tumor progression.

Genotoxic carcinogens exert their tumorigenic effects in part by inducing genomic instability. We recently showed that loss of heterozygosity (LOH) on chromosome 12 associates significantly with the induction of chromosome instability (CIN) by the likely human lung carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and vinyl carbamate (VC) during mouse lung carcinogenesis. Here, we demonstrate the carcinogen specificity of this event and its effect on lung tumor evolution. LOH on chromosome 12 was observed in 45% of NNK-induced, 59% of VC-induced, 58% of aflatoxin B1 (AFB1)-induced, 14% of N-ethyl-N-nitrosourea (ENU)-induced and 12% of spontaneous lung adenocarcinomas. The frequency of LOH in each of the carcinogen-induced groups, except ENU, was significantly higher than in the spontaneous group (P<0.001). Deletion mapping revealed four potential candidate regions of 1-4 centiMorgans suspected to contain targeted tumor suppressor genes, with at least one expected to have a role in CIN. The relationship between LOH on chromosome 12 and additional chromosomal alterations occurring during lung tumor progression was also examined. LOH on chromosomes 1 and 14 were moderately frequent during malignant progression in tumors from all treatment groups, occurring in 21-35 and 18-33% of tumors. However, these alterations showed significant concurrence with LOH on chromosome 12 in VC-, NNK- and AFB1-induced tumors (P<0.05). The results suggest that a carcinogen-selective mechanism of lung cancer induction involves the frequent inactivation of genes on chromosome 12, including a stability gene that evidently promotes the evolutionary selection of additional chromosomal alterations during malignant progression.

Tea and cancer prevention: studies in animals and humans.

The role of tea in protection against cancer has been supported by ample evidence from studies in cell culture and animal models. However, epidemiological studies have generated inconsistent results, some of which associated tea with reduced risk of cancer, whereas others found that tea lacks protective activity against certain human cancers. These results raise questions about the actual role of tea in human cancer that needs to be addressed. This article is intended to provide a better perspective on this controversy by summarizing the laboratory studies in animals and humans with emphasis on animal tumor bioassays on skin, lung, mammary glands and colon, and the molecular and cellular mechanisms affected by tea. Finally, a recent small pilot intervention study with green tea in smokers is presented.

Inhibition of COX-2 in colon cancer cell lines by celecoxib increases the nuclear localization of active p53.

Inactivation of the p53 tumor suppressor gene usually involves somatic mutation or binding of viral oncoproteins to the p53 protein. However, several types of malignant and premalignant tissues harbor a genetically wild-type, but transcriptionally inactive, form of p53, often localized in the cytoplasm. Electrophilic prostaglandins (PGs) are known to sequester and inactivate p53 in the cytoplasm, an effect that is likely to occur when cyclooxygenase (COX)-2 levels become elevated during colon carcinogenesis. We determined the localization and expression of p53 in the presence of PGA(1) and celecoxib, a selective COX-2 inhibitor in human colon cell lines HCT-116 (wild-type p53) and HT-29 (mutant p53). In the absence of treatment, p53 protein accumulated preferentially in the nucleus in both cell lines. We observed that the total cellular levels of p53 protein increased with exposure time and concentration of PGA(1). By contrast, p21 protein levels remained unchanged as a function of time and concentration of PGA(1). In the presence of 20 micro M PGA(1), p53 accumulated preferentially in the cytosol. The nuclear:cytosol ratios of p53 were 31 and 2.1 in the controls and in the presence of PGA(1) in HCT-116 cells but were 22 and 4, respectively, in HT-29 cells. Treatment with 50 micro M celecoxib for 24 h did not significantly change p53 expression and localization. However, in the presence of 100 micro M celecoxib, p53 levels increased in the nucleus. The nuclear:cytosol ratios were then 31 (control) and 60 (100 micro M celecoxib) in HCT-116 cells and 22 (control) and 36 (100 micro M celecoxib) in HT-29 cells. These results indicate that electrophilic PGs cause wild-type p53 accumulation in the cytosol where it is inactive. Inhibition of COX-2 by celecoxib appears to alleviate this effect on p53 by reducing electrophilic PG synthesis. Thus, COX-2 inhibition of electrophilic PG formation appears to protect p53 tumor suppressor function.

Carcinogenic induction directs the selection of allelic losses in mouse lung tumorigenesis.

Recent evidence suggests that genome-wide allelic imbalances are inducible by carcinogens and may occur as cells adapt to carcinogenic exposure during tumorigenesis. We investigated the role of carcinogenic exposure on global and selected loss of heterozygosity (LOH) during mouse lung adenocarcinogenesis. Tumor induction by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) or vinyl carbamate (VC) resulted in a significant overall increase in the number of chromosomes affected by LOH per tumor when compared with spontaneous lung tumors. Allelic loss on chromosome 12 occurred at a frequency of 35% and 40% in NNK- and VC-induced tumors, respectively, compared with 8.3% in spontaneous tumors (P < 0.01). In contrast, spontaneous lung adenocarcinomas displayed LOH on chromosome 4 at a frequency of 77%, whereas a frequency of only 36% (P < 0.001) was observed in tumors induced by NNK. Sixty-four percent of the VC-induced tumors displayed LOH on chromosome 4. In addition, allelic losses on chromosomes 12 and 14 were significantly associated with an increase in chromosomal instability, suggesting that genes inactivated on these chromosomes may contribute to this effect. The results from this study demonstrate that genotoxic carcinogens increase chromosome instability, as evidenced by a significant increase in global LOH frequency, and significantly alter the selection of chromosomal alterations during lung tumor development.

A strong candidate gene for the Papg1 locus on mouse chromosome 4 affecting lung tumor progression.

Lung cancer is the leading cause of cancer death among both men and women, accounting for more than 28% of all cancer deaths. In fact, more people die of lung cancer than of colon, breast, and prostate cancers combined. Although lung cancer is largely induced by smoking, there is strong evidence for genetic susceptibility and gene-environment interactions in the development of lung cancer. Inbred mouse models offer an effective means of identifying candidate lung cancer susceptibility loci since genetic heterogeneity and enormous variation in exposure levels to environmental agents make it difficult to identify lung cancer susceptibility loci in humans. Papg-1 (pulmonary adenoma progression 1) was previously mapped to a region on mouse chromosome 4. This locus contains a candidate gene, Cdkn2a also referred to as Ink4a/Arf, which dually encodes two established tumor suppressors p16(INK4a) and ARF. Cdkn2a became a primary candidate for Papg-1 for two reasons: (1) two haplotypes of mouse Cdkn2a were found to segregate with differential genetic susceptibility to lung tumor progression in mice; and (2) in vitro studies showed that the p16(INK4a) allele from the BALB/cJ mouse had a significantly decreased ability to bind and inhibit CDK6 and to suppress cell growth when compared with the p16(INK4a) allele from the A/J mouse. Here, we report that mice with a heterozygous deficiency for the A/J Cdkn2a allele were significantly more susceptible to lung tumor progression than mice with a heterozygous deficiency for a BALB/cJ Cdkn2a allele, when compared to their respective wild type mice. These results offer strong evidence that naturally occurring variation of p16(INK4a) influences susceptibility to enhance lung tumor progression making it a strong candidate for the lung tumor progression locus, Papg-1.