Combined effects of asbestos and cigarette smoke on the development of lung adenocarcinoma: Different carcinogens may cause different genomic changes

The carcinogens in cigarette smoke are distinct from asbestos. However, an understanding of their differential effects on lung adenocarcinoma development remains elusive. We investigated loss of heterozygosity (LOH) and the p53 mutation in 132 lung adenocarcinomas, for which asbestos body burden (AB; in numbers per gram of dry lung) was measured using adjacent normal lung. All cases were classified into 9 groups based on a matrix of cumulative smoking (CS in pack-years; CS=0, 0<CS<25, ≥25 CS) and AB (AB=0, 0<AB<1,000, ≥1,000 AB). AB=0 indicates a lower level than the detection limit of ~100. LOH frequency increased only slightly with the elevation of CS in the AB=0 groups. In the AB>0 groups, LOH frequency increased as AB and/or CS was elevated and was significantly higher in the ≥1,000 AB, ≥25 CS group (p=0.032). p53 mutation frequency was the lowest in the AB=0, CS=0 group, increased as AB and/or CS rose, and was significantly higher in the ≥1,000 AB, ≥25 CS group (p=0.039). p53 mutations characteristic of smoking were frequently observed in the CS>0 groups contrary to non-specific mutations in the CS=0, AB>0 groups. Combined effects of asbestos and smoking were suggested by LOH and p53 analyses. Sole exposure to asbestos did not increase LOH frequency but increased non-specific p53 mutations. These findings indicate that the major carcinogenic mechanism of asbestos may be tumor promotion, acting in an additive or synergistic manner, contributing to the genotoxic effect of smoking. Since this study was based on a general cancer center’s experience, the limited sample size did not permit the consideration that the result was conclusive. Further investigation with a large sample size is needed to establish the mechanism of asbestos-induced lung carcinogenesis.


Introduction
Lung cancer is one of the leading causes of cancer-related death in both men and women worldwide, and adenocarcinoma is the most predominant histologic subtype in many parts of the world. Tobacco smoke is clearly the most important factor associated with the development of lung cancer, accounting for 80-90% of all cases. Asbestos is another significant inhaled carcinogen, contributing to the development of ~5-7% of all lung cancers (1). Many studies on asbestos-related lung carcinogenesis have analyzed the genotoxic effects of asbestos; asbestos fibers induce DNA damage, chromosome aberrations, mitotic disturbances and gene mutations (2). In addition, asbestos fibers can stimulate a range of other effects including cell proliferation, chronic inflammation, enhanced gene expression, such as c-fos and c-jun overexpression, and transformation (3,4). Despite these studies, the efficacy of asbestos-exposure as a complete lung carcinogen, independent of tobacco smoke, has not been demonstrated in humans, since lung cancers of asbestos-exposed individuals frequently occur in smokers and ex-smokers. The majority of asbestos-related lung cancers may result from the combined effects of asbestos and carcinogens in tobacco smoke, with the possibility of a synergistic relationship first proposed by Doll (5). Hence, the mechanism of asbestos-induced lung carcinogenesis still remains unclear.
Both loss of heterozygosity (LOH) and the p53 mutation are genetic alterations. LOH is frequently noted in cancer cells and is thought to occur through genetic instability at the chromosomal level. On the other hand, the p53 mutation is a genetic alteration at the nucleotide level. Mutation in the p53 tumor suppressor gene is the most frequently observed gene mutation in cancers. As described below, not only p53 mutations but also LOH spectra differ in different cancer types associated with different etiologies. Previously we compared the frequency of LOH on all autosomal chromosomes among non-small cell lung carcinomas (6,7) as well as p53 mutation patterns with adenocarcinoma cell morphology (8). The frequency of allelic loss on many chromosomal arms was commonly higher in squamous cell carcinomas than in adenocarcinomas. This result suggested that more cumulative genetic changes are associated with tumorigenesis in squamous cell carcinomas than contribute to adenocarcinomas, a pattern which may reflect a difference in the carcinogenic mechanisms responsible for the two histologies. In addition, we observed high frequencies of allelic losses on chromosomes 9p, 9q and 13q in squamous cell carcinomas, the majority of which were from smokers, and higher frequencies of allelic losses on these arms in adenocarcinomas from smokers than those from non-smokers. This loss of specific chromosomes associated with a particular histology is an example of LOH spectra reflecting etiology. The p53 mutational spectra differ among cancers of various organs, and its frequency and mutational spectra can be said to reflect carcinogenic patterns characteristic of exogenous or endogenous factors and thus may be helpful for identification of the responsible agents, including, among others, cigarette smoke, aflatoxin B1 and ultraviolet light. Hence, the analysis of p53 mutation can provide clues to the etiology of diverse tumors and to the function of specific regions of p53 (9,10).
The mutation pattern in smokers shows an excess of G:C to T:A transversions (34.2%), which are relatively uncommon in non-smokers or passive-smokers (16.6%) (11). These transversions often occur at codons 157, 158, 245, 248 and 273, experimentally identified as sites of adduct formation by benzo(a)pyrene, a single polycyclic aromatic hydrocarbon (PAH)-compound found in cigarette smoke. Other PAH-compounds also have a similar preference for adduct formation in these p53 codons (12,13).
In the present study, to elucidate the combined effects of asbestos-exposure and smoking on development of lung adenocarcinomas, we used 132 lung adenocarcinomas, for which we already obtained all detailed smoking histories, comprehensive LOH data for all autosomal chromosomes (7), and p53 mutation data.

Materials and methods
Patients and sample preparation. A total of 335 cases of lung adenocarcinoma were surgically removed at the Cancer Institute Hospital (CIH), Tokyo, Japan, between September 1989 and August 1996. Among the cases, fresh tumor tissues and corresponding normal lung and detailed smoking histories were successfully collected from 132 patients, which were used as materials in this study. Hence, they were collected semi-randomly without respect to asbestos-exposure status, and therefore provided a representative population for a cancer center in Japan. The clinicopathological data for these samples are summarized in Table I. We used a differentiation grading that was basically according to the former version of the Japanese Lung Cancer Society (14), as previously performed (15). Smoking history was surveyed intensively from patients and their families and presented as cumulative smoking (CS) in pack-years. The study protocol was approved by IRB of CIH and informed consent was obtained from all patients.
Measurement of asbestos-exposure. Asbestos-body burden (AB; in numbers per gram of dry lung tissue) was measured using paraffin blocks of corresponding normal lung tissues by a polarizing microscope (16). The detection limit, which means no AB was found on the measuring filter sample, was ~100 AB/g (dry lung) and expressed as 0 in this study.
A matrix of smoking-exposure and asbestos-exposure. To examine the dose-effect relationship of asbestos-exposure (presented as AB) and smoking-exposure (presented as CS in pack-years) on lung adenocarcinomas, we classified all cases into 9 groups based on a matrix of CS in pack-years: CS= 0 (n=54, 41%), 0<CS<25 (n=18, 14%), ≥25 CS (n=60, 45%), and AB: AB=0 (n=64, 48%), 0<AB<1,000 (n=28, 21%), ≥1,000 AB (n=40, 31%). Since the patients were selected consecutively from surgical tumor files in a general cancer center, only 4 cases (3.0%) exceeded 5,000 in AB. To investigate the mechanism of asbestos-induced lung carcinogenesis in a representative population for a cancer center, not a biased population heavily exposed to asbestos, we divided the cases between AB <1,000 and ≥1,000 AB.
LOH analysis. For LOH analysis, we performed Southern blotting. Experimental procedures and probes used were essentially the same as previously described (6,7). To facilitate the comparison, we used a fractional allelic loss (FAL) value, defined as: (number of chromosome arms with LOH)/(number of informative arms) for each case. Of 132 patients with adenocarcinomas, LOH data were available for 114 patients.
p53 mutation analysis. Analysis of p53 mutation was performed essentially as described elsewhere (8). Genomic DNA from fresh tumor samples was prepared and exons 4-8 and 10 of p53 were analyzed by polymerase chain reaction and DNA sequencing. Of the 132 patients with adenocarcinomas, p53 mutation data were available for 123 patients.
Statistical analysis. For statistical analysis, we used the t-test, Fisher's exact test, and Chi-square test, as appropriate. The twosided significant level was set at p<0.05. Data were analyzed with the statistical software Stata version 11 (StataCorp., College Station, TX, USA). Table II and Fig. 1A. LOH frequency increased only slightly correlating with the elevation of CS in the AB=0 groups, whereas, in the AB>0 groups, it increased as AB and/ or CS was elevated and was significantly higher in the ≥1,000 AB, ≥25 CS group than in the AB=CS=0 group (p=0.032). Details of cases with p53 mutations in lung adenocarcinomas are shown in Table III and summarized in Table IV. The p53 mutation rates of pathological stage I and II-IV lung adenocarcinomas were 32% (18 of 57) and 44% (29 of 66), respectively, not significantly different by Fisher's exact test (p= 0.19). p53 mutation frequency of lung adenocarcinomas classified by CS and AB are depicted in Fig. 1B. p53 mutation frequency was the lowest in the AB=CS=0 group (18%), increased as AB and/or CS rose, and was significantly higher in the ≥1,000 AB, ≥25 CS group (53%) than in the AB=CS=0 group (p= 0.039). Tobacco smoke, one of the most significant exogenous carcinogenic agents has been shown to frequently cause specific p53 mutations, especially G:C to T:A transversion (17) at specific codons described as 'hotspots', such as codon 157, 158, 245, 248 and 273 (13). p53 mutations characteristic of smoking, such as G:C to T:A transversion at the  Table II. FAL values (± SD) in lung adenocarcinomas, classified by AB and CS in pack-years.    tobacco-specific codons were frequently observed in the CS>0 groups, whereas non-specific mutations were often detected in the CS=0, AB>0 groups (Tables III and IV). In the ≥1,000 AB, CS=0 group, there was only one transversion and no tobaccospecific codons for the six p53 mutations. In contrast, in the AB=0, ≥25 CS group, there were five G:C to T:A transversions and five tobacco-specific codons among 13 p53 mutations. Fig. 2 shows p53 mutation spectra in lung adenocarcinomas, classified as smokers (A, n=33) or non-smokers (B, n=14) and asbestos-exposed (C, n=28) or not (D, n=19). Although p53 mutation spectra varied depending on the status of smoking history, they showed little difference between asbestosexposed or non-exposed. Whereas smokers had frequent G:C to T:A transversions, which are smoking-associated p53 mutations, non-smokers had frequent G:C to A:T transitions at CpG sites associated with spontaneous mutations, consistent with previous reports (9,17). With respect to tumor differentiation grade, a heavier smoking habit was associated with less-differentiated adenocarcinomas (Fig. 3A, p=0.0010, Chi-square test), in line with a previous study (18). On the other hand, there was no correlation between asbestos deposition and the differentiation grade (Fig. 3B, p=0.75).

Discussion
Both tobacco smoke and asbestos fibers are significant inhaled carcinogens which contribute significantly to lung adenocarcinoma development. We previously revealed that chromosome instability and LOH, rather than minisatellite and microsatellite instability, play major roles in the development of lung adenocarcinomas (19). The LOH and p53 spectra provide clues concerning the etiology and nature of carcinogenesis. To elucidate the carcinogenic mechanisms of two different inhaled carcinogens, asbestos and cigarette smoke, we investigated LOH on all autosomal chromosomes and measured asbestos burden (AB; asbestos body per gram of dry lung tissue) using corresponding normal lung tissue and investigated p53 mutation employing fresh tumor samples.
The p53 mutational spectra may be helpful for identification of the origins of the mutations that give rise to human   Percentages may not total 100, due to rounding. Del/Ins, deletion/insertion; AB, asbestos burden; CS, cumulative smoking.
cancers. For example, aflatoxin B1-associated hepatocellular carcinomas frequently have the specific p53 mutations: G:C to T:A transversions at the 3rd base of codon 249, AGG to AGT (Arg to Ser) (20). Another example of a clearly characteristic 'finger-print' mutation in p53 is the CC to TT double mutation in skin cancer (21). Exposure to UV light, a physical mutagen, produces distinctive pyrimidine dimers that, if unrepaired, can produce tandem mutations, most characteristically CC to TT transitions. Similar to these, the p53 mutational spectra can provide clues to the etiology of cancers. The possible role of asbestos-exposure in the genesis of p53 mutations in lung cancers is less well understood. Husgafvel-Pursiainen et al investigated p53 mutation of 105 lung cancers from smokers, comprising 53 squamous cell carcinomas, 39 adenocarcinomas and other 13 carcinomas, focusing on the presence or absence of asbestos-exposure (22). They found p53 mutations in 39% of asbestos-exposed patients with lung cancer while the percentage was 54% in patients not exposed to asbestos, indicating that the p53 mutations were less common among the cases with occupational asbestosexposure than in the non-exposed cases. These results have not been verified yet by another study, and need additional examinations of smoking status.
In adenocarcinoma without asbestos-exposure or smokingexposure, the p53 mutation rate was the lowest. It increased in correlation with the elevation of asbestos-exposure and/ or smoking-exposure. Adenocarcinomas associated with frequent smoking have characteristic p53 mutations, especially G:C to T:A transversions (17), at specific 'hotspot' codons (13). However, adenocarcinomas associated only with asbestos-exposure had non-specific p53 mutations, such as transitions which are thought to be caused by endogenous mechanisms associated with spontaneous events (9,17). Asbestos may work in a promoter-like manner. Production of reactive oxygen species and/or induction of tissue regeneration may be relevant.
Adenocarcinomas have different etiologies from squamous cell carcinomas, which can be reflected also in terms of LOH. As we revealed, LOH frequency was higher in squamous cell carcinomas than in adenocarcinomas (6,7). Poorly differentiated adenocarcinomas, which are often noted in smokers such as squamous cell carcinomas, have higher LOH frequency than differentiated adenocarcinomas, which have a relatively weaker association with smoking (23). Smoking induces complicated genetic changes in lung cancers.
One of the most intriguing recent discoveries in the field of lung cancer research is the identification of new driver mutations in lung adenocarcinomas, such as EGFR mutations (24,25) and ALK fusion (26). Both lung cancers with EGFR mutations or ALK translocations are characterized by negative or light smoking history. Lung cancers in non-smokers are considered to be less genetically complex than those in smokers and therefore they often have distinct characteristics developing on simple gene mutations for maintenance and survival. Consequently, patients with tumors harboring such simple oncogenic mutations represent good candidates who may stand to benefit from molecular-targeted drugs. To date, two-thirds of Japanese adenocarcinomas and a little more than half of Caucasian adenocarcinomas have mutually exclusive oncogenic mutations or other genetic alterations including EGFR, KRAS, MET, ALK and HER2 (27). Asbestos-associated alterations in chromosomal regions, such as 19p13 (28), 9q33.1 (29) and 2p16 (30) have been identified. Whereas the smoking status has a significant association with driver mutations in lung adenocarcinomas, the relationship with asbestos-exposure remains unclear.
In adenocarcinomas without asbestos-exposure, the LOH frequency increased only slightly, correlating with the elevation in smoking-exposure. On the other hand, in adenocarcinomas with asbestos-exposure, the LOH frequency increased as asbestos-exposure and/or smoking-exposure was elevated. This suggests that asbestos-exposure in concert with smoking-exposure increases LOH frequency.
In the present study, lung adenocarcinomas, for which asbestos-exposure and smoking-exposure data could be obtained, were examined for LOH and the p53 mutation. Combined effects of asbestos and cigarette smoke were suggested by these analyses. Asbestos-exposure alone did not increase the LOH frequency but increased non-specific p53 mutations. These findings suggest that the major carcinogenic mechanism of asbestos in lung adenocarcinomas may be as a promoter, contributing to the genotoxic effect of cigarette smoke. Since this study was based on a general cancer center's experience, the limited sample size does not permit consider- Figure 3. Cumulative smoking (CS) in pack-years (A) and asbestos burden (AB) (B) with reference to the histological differentiation grade. Although there was a significant relationship between CS and the differentiation grade (p= 0.0010, Chi-square test), there was no correlation between AB and the differentiation grade (p=0.75). Well-diff., well-differentiated; mod-diff., moderately differentiated; poorly-diff., poorly differentiated. ation that the result is conclusive. Further investigation with a large sample size is required to establish the mechanism of asbestos-induced lung carcinogenesis.