| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Environmental Epidemiology, Institute of Environmental Medicine, Karolinska Institute, S-171 77 Stockholm, [A. Y., G. P.]; Department of Biotechnology, KTH, Royal Institute of Technology, S-10044 Stockholm, [C. W., J. L.]; Department of Immunology, Microbiology, Pathology, and Infectious Diseases, Division of Pathology, Huddinge University Hospital, S-14186 Stockholm, [A. H.]; and Department of Cell Biology, Faculty of Health Sciences, S-58185 Linköping, [P. S.] Sweden
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
The p53 tumor suppressor gene is commonly mutated in human lung cancer (6)
. Mutations predominantly occur in exons 58, which are highly conserved through evolution and important for the function. The most common mutation in p53 in lung cancer is G:C
T:A transversions (7)
, which are associated with smoking (8
, 9)
. Two investigations have indicated unusual mutation patterns in the p53 gene in lung tumors among underground miners. In one study (10)
, 7 of 19 patients had mutations in exon 5 and 6, but none was a G
T transversion. In another study (11)
, 16 of 29 mutated cases had identical G
T transversions in codon 249 (AGG
ATG). Later studies in underground miners have not confirmed these results (12, 13, 14)
. Almost all of the lung tumors under investigation appeared in smokers, which made it difficult to assess in detail the role of radon exposure.
The aim of our study was to assess mutations in the p53 gene in lung tumors related to residential radon exposure and the joint effects of tobacco with radon. It was based on cases included in the nation-wide Swedish radon epidemiological study (5) and focused particularly on nonsmokers.
| Subjects and Methods |
|---|
|
|
|---|
50 Bq/m3) for each if the major histological types (squamous cell carcinoma, small cell carcinoma, and adenocarcinoma). Each of the three histological groups with radon exposure over 140 Bq/m3 contained less than 50 smoking cases, and consequently, all of the subjects in these categories were included (n = 111). Additionally, 50 squamous cell carcinoma (58% of the cases in this category) and 50 small cell carcinoma cases (76%) were selected randomly among smoking lung cancer cases with radon exposure up to 50 Bq/m3 as well as all of the adenocarcinoma cases (n = 39) in this exposure category. Patients treated with chemotherapy or radiation before the tissue sample was taken were excluded from the study (n = 27), and for 22 and 41 cases the tumor material was missing or insufficient, respectively, leaving 243 of the 333 originally selected cases for detailed analysis.
Exposure Information.
All of the study subjects or the next-of-kin were sent a questionnaire asking about smoking habits of the subjects, their spouses, and parents. The subjects lifetime occupational history and their residential addresses since 1947 were also investigated. Radon measurements were intended for all of the dwellings in which the subject had lived during a period of 2 years or more from 1947 up to 3 years before the end of the follow-up. In each dwelling, one detector was placed in a bedroom and another in the living room over a period of 3 months during the heating season. Radon was measured by solid-state
track detectors processed at the Swedish Radiation Protection Institute. Cumulative radon exposure since 1947 was estimated for each subject by adding the products of the measured radon level and the duration of residence in each dwelling. Time-weighted means radon concentrations were calculated by dividing the cumulative radon exposure by the total residential time in dwellings for which radon measurements were available.
Smoking habits were classified according to the time-weighted mean consumption of tobacco during the subjects lifetime. Conversions were made for pipe tobacco, cigarillos, and cigars. Subjects who stopped smoking 2 or more years before the end of the follow-up period were classified as ex-smokers. The occupation of each subject was grouped in one of four categories based on earlier evidence of occupational risks of lung cancer (16, 17, 18) . Subjects who lived for 10 years or more in any of the three largest cities in Sweden at some time between 1947 and the end of follow-up were classified as urban dwellers because of reported excess risks of lung cancer in these cities.
Molecular Analysis.
Microscopical slides and lung tumor samples were obtained from the pathology departments where the cancer cases were diagnosed. Paraffin blocks containing tumor tissue were selected from the original histological sections. Each block was then cut into (a) one section for demonstration of tumor tissue; (b) one section for IHC; (c) three 10-µm-thick sections for PCR analysis; and (d) one final section, again to confirm the presence of tumor. To avoid contamination between samples, the sectioning was performed with disposable knives and by cutting empty paraffin blocks between each new tumor.
The histopathological typing was reviewed and revised by one of us (A. H.). For immunohistochemical screening of abnormal p53 expression, sections of the tumors were mounted on glass slides, deparaffinized, placed in citrate buffer, and treated with microwaves for antigen retrieval. The immunohistochemistry staining was performed using a Tech Mate 500 immuno-stainer. DO-7 (Dako, code M 7001) was used as primary antibody (negative control Dako code H 0960) and developed using ChemMate Detection kit Peroxidase/DAB (Dako, code K 5001). The degree of staining of the nuclei of the cells was estimated on a scale from 0 to 3, and the distribution of the staining was classified as focal (more than 10 cells in a cluster and up to 20% of the tumor cell population) and general (more than 20% of the tumor tissue), respectively. The outcome was then classified as "reactive" and "nonreactive" regarding the binding of antibodies to the cell nuclei. No antibody binding (representing degree 0) and poor antibody binding (representing degree 1) in combination with focal distribution was considered as nonreactive.
Deparaffinized tumor tissues were digested with proteinase K, and genomic DNA was purified by phenol/chloroform extraction. Intronic primers for exons 58 of the p53 gene were used to amplify genomic tumor DNA. For SSCP3 analysis (19) , PCR products were labeled by including [32P]dATP in a secondary PCR-amplification for two cycles. Radiolabeled PCR products were diluted 20-fold with 0.1 mM EDTA/0.1% SDS, denatured, and loaded on 68% polyacrylamide gels containing 5% glycerol and 0.5 x mutation detection enhancement gels (FMC Bio-Products, Rockland, ME). DNA strands were separated for 1620 h at 68 W and autoradiographed. Mutations were detected as shifts in the mobility of the bands of separated single strands in the autoradiogram. PCR products showing altered mobility were eluted from the gels and reamplified for sequence determination. Sequencing was performed using Thermo Sequenase with 32P-radiolabeled dideoxynucleotides from Amersham Life Sciences. All of the analyses were performed in duplicates from at least two independent PCR amplifications.
Confirmative, direct solid-phase DNA sequencing of separately generated PCR products was performed (20) . Primers used for DNA amplification were situated in intronic sequences covering exons 58 of the p53 gene and labeled with biotin to facilitate direct solid-phase sequencing of PCR products by the use of paramagnetic beads. The biotinylated amplified fragments were immobilized onto streptavidin-coated solid support, and strand-specific alkali elution produced a clean template for sequencing. Solid-phase sequencing was performed by a robot with fluorescence-labeled primers, and an automated laser fluorescence instrument was used for sequence analysis. Alterations in the p53 gene were analyzed by comparing sequence data from the tumor tissues with wild-type sequence data.
Parallel analyses were performed using the SSCP combined with traditional DNA sequencing and direct solid-phase sequence analysis of the first 49 samples for quality assessment. For the rest of the tumors, only those samples showing confirmed altered mobility in polyacrylamide and/or mutation detection enhancement gels in the SSCP analysis were further analyzed by the solid-phase methodology.
Statistical Analysis.
The data were analyzed with the SAS software for PCs and STATA statistical software (Release 5.0; StataCorp, College Station, TX). Associations between exposure to radon and tobacco consumption and the presence of p53 mutations were described with maximum-likelihood estimates of relative risks and 95% confidence intervals based on logistic regression analyses. All of the ORs presented are adjusted for age (three categories: <60, 6069, and >69 years), gender, and smoking (three categories: nonsmoker, current smoker of <10 cigarettes/day, or at least 10 cigarettes per day) and/or residential radon exposure (three categories: <50 Bq/m3, 140400 Bq/m3, and >400 Bq/m3).
| Results |
|---|
|
|
|---|
|
Mutations of p53 in relation to histological type, radon exposure, and smoking are further elucidated in Table 2
. Squamous cell carcinomas showed the highest excessive risk for p53 mutations in relation to residential radon exposure (OR, 3.6; 95% CI, 1.012.8). On the other hand, the strongest association between p53 mutations and smoking was seen for small cell carcinomas, in which all of the 24 mutations were detected among smoking lung cancer cases.
|
|
Missense mutations accounted for 37 (60.7%) of 61 mutations, whereas silent base substitutions (3.3%), deletions (19.7%), insertions (1.6%), splice mutations (8.2%), and mutations leading to a stop codon (6.6%) accounted for the rest (Table 4)
. A total of 43 (70.5%) mutations were base substitutions, mostly G
T transversions and G
A transitions. All of the G
T transversions were detected in smokers and especially in the low radon exposure category. Three identical splice mutations were detected in intron 8, CGAGgt
CGAGtt, all in smoking cases highly exposed to residential radon. Four mutations were detected in codon 249, but none was of the same type as the hot spot mutation indicated by Taylor et al. (11)
.
|
A:T transitions and occurred at CpG sites.4
These base substitutions may result from spontaneous deamination of cytosine and 5-methylcytosine residues, which most often occur at CpG dinucleotides. Nine codons (i.e., 158, 175, 179, 193, 245, 249, 273, 278, and 287) contained more than one mutation.
In the IHC, 100 tumors were classified as reactive, of which only 33 showed altered mobility in the SSCP analysis (in two independent analyses; Table 5
). Twenty-five mutations detected in the SSCP analysis and direct sequencing were not observed in the immunohistochemistry staining. The concordance between the two methods was estimated to 62% and did not seem to be affected to any major extent by smoking or radon exposure. Assuming that the SSCP analysis in combination with DNA sequencing constituted the standard, the sensitivity and specificity of the immunostaining were estimated to 0.57 and 0.64%, respectively, and the positive and negative predicted values were 33 and 83%, respectively. Overall, the results regarding the association between radon exposure, smoking status, and occurrence of p53 mutations were consistent between the immunohistochemical staining and the SSCP analysis, as indicated above.
|
| Discussion |
|---|
|
|
|---|
In the present study, mutations in the p53 gene seemed to be associated with residential radon exposure. In most previous studies on lung cancer, radon exposure, and p53 mutations, cases with very high radon exposure in mines have been included, to identify possible radon-associated mutation hot spots (12, 13, 14 , 27) . In two studies, the p53 gene mutation prevalence in lung cancer cases among uranium miners was reported to be 37 and 56%, respectively (10 , 11) , and in a third study, it was estimated to be at least 40% (12) . Lung cancer cases exposed to approximately 80 Bq/m3 of domestic radon showed a p53 mutation prevalence of 28%, and those exposed to more than 300 Bq/m3 for more than 20 years had a prevalence of 41% (28) . Although the association between radon exposure and mutation prevalence in our study did not reach statistical significance, an association between p53 mutation prevalence and residential radon exposure at levels occurring in Sweden seems plausible.
We found p53 mutations to be associated with smoking status. A dose-response relation between tobacco consumption and p53 mutation prevalence was not seen, although the proportion of p53 mutations has been found to increase with the amount of tobacco smoking (23
, 29)
. In contrast, high tobacco consumption in combination with high radon exposure seemed to contribute to a decreased risk for p53 mutations compared with nonsmokers and smokers of less than 10 cigarettes per day with high radon exposure, indicating a possible antagonism of heavy tobacco consumption on the occurrence of radon-induced mutations. One explanation may be that heavy smokers have a thicker mucous membrane, preventing penetration of
-particles to target cells (30)
. It is also possible that lung cancer subjects smoking more than 10 cigarettes per day require a higher tobacco consumption to develop their lung cancer and, therefore, are less susceptible for p53 mutations, or that they to a larger extent develop their lung cancer via other mechanisms not involving the p53 gene. Interestingly, an association was suggested between passive smoking and p53 mutations among nonsmokers.
Compared with the p53 database, we detected a larger fraction of splice mutations (8 compared with 2% in the database) and deletions (20 and 9%, respectively) but fewer missense mutations (61 and 75%, respectively; Ref. 7
). In the p53 database, 23% of the mutations among smoking lung cancer cases are G
T transversions, which is similar to our study where 20% of the mutations among smokers were G
T transversions. For nonsmokers, 20% of the mutations reported in the database are G
T transversions compared with our study, in which no such transversions were detected among nine mutations in nonsmokers. Our study included a total of only 61 mutations, and the proportion of different mutations is uncertain. Furthermore, it is difficult to assess the quality of the information in the p53 database, and several sources of bias have been suggested, resulting from the selection of study subjects and analytical procedures (31)
.
Three identical CGAGgt
CGAGtt splice mutations in intron 8 were detected, all of which were among smokers with high radon exposure (two of the three mutations occurred among those exposed to over 400 Bq/m3). The three mutations are notable, inasmuch as only one mutation of this type has been detected in intron 8 among the more than 1000 reported mutations in lung cancer, suggesting a possible radon-specific mutation (7)
.
There are some methodological issues involved in a case-only study such as ours (32 , 33) . Bias resulting from the selection of lung cancer cases is unlikely because we used a sample from a consecutive series of incident cases in a defined population base. In certain exposure categories, all of which were nonsmoking lung cancer cases, and a stratified sample of the smoking lung cancer cases were selected from the nation-wide Swedish case-control study on residential radon and lung cancer. However, only cases with available and sufficiently large tumor samples were possible to analyze, which could lead to bias if treatment, availability (surgery and autopsy), or tumor size are related to the occurrence of p53 mutations. Confounding by other possible risk factors for p53 mutations, such as smoking status, age, and gender, was adjusted for in the analysis. There are substantial uncertainties in the estimation of radon exposure because of errors in radon measurements, duration of measurements, number of rooms measured, the measurement of radon instead of unattached and attached radon progeny, and measurements made recently to estimate exposure decades ago (34) . These uncertainties are likely to dilute a possible association between residential radon exposure and p53 mutations. Next-of-kin constituted the source of questionnaire information for most of the cases, which may raise concerns regarding data quality. However, several Swedish studies (35, 36, 37) show that information of high quality about a deceased subjects residential history and smoking habits can be obtained from a next-of-kin with the methods used.
There were some discrepancies in results on p53 mutations in individual tumors between the IHC staining and the SSCP analysis followed by DNA sequencing, although the overall picture in relation to residential radon and smoking was consistent. The concordance between p53 protein nuclear accumulation and p53 gene mutation data was estimated to be 63%, which is in the range (5379%) of what previous studies have reported (10 , 38, 39, 40, 41, 42, 43) . It is probable that the mutations detected by repeated SSCP and sequence analysis are true mutations. Although tissue samples were selected to contain tumor tissue to a considerable extent, relative paucity of tumor cells may contribute to explaining the negative SSCP analysis despite positive results from the IHC. Another possibility may be accumulation of wild-type p53 in tumor cells, at levels sufficient to be detected in the IHC. On the other hand, splice mutations or deletions, which have been reported to cause accumulation of the p53 protein to a lesser extent than missense mutations, may partly explain the negative results in the IHC but positive results in the SSCP and sequence analyses.
| Acknowledgments |
|---|
| Footnotes |
|---|
1 Supported by a grant from the Swedish Cancer Society. ![]()
2 To whom requests for reprints should be addressed, at Division of Environmental Epidemiology, Institute of Environmental Medicine, Karolinska Institute, Box 210, S-171 77 Stockholm, Sweden. ![]()
3 The abbreviations used are: SSCP, single-stranded conformation polymorphism; OR, odds ratio; CI, confidence interval; IHC, immunohistochemistry analysis. ![]()
4 CpG site is a shorthand notation for a cytosine followed by a guanine. Cytosine bases within CpG sites are frequently methylated and susceptible to spontaneous deamination, which leads to a base substitution of cytosine to thymidine. ![]()
Received 10/13/98; revised 2/16/99; accepted 3/ 8/99.
| References |
|---|
|
|
|---|
-particles. Carcinogenesis (Lond.), 18: 511-516, 1997.
-thiotriphosphate nucleotides. Clin. Chem., 41: 1461-1466, 1995.This article has been cited by other articles:
![]() |
A. Ruano-Ravina, R. Perez-Becerra, M. Fraga, K. T. Kelsey, and J. M. Barros-Dios Analysis of the relationship between p53 immunohistochemical expression and risk factors for lung cancer, with special emphasis on residential radon exposure Ann. Onc., January 1, 2008; 19(1): 109 - 114. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. M. Barros-Dios, M. A. Barreiro, A. Ruano-Ravina, and A. Figueiras Exposure to Residential Radon and Lung Cancer in Spain: A Population-based Case-Control Study Am. J. Epidemiol., September 15, 2002; 156(6): 548 - 555. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. H. Vahakangas, W. P. Bennett, K. Castren, J. A. Welsh, M. A. Khan, B. Blomeke, M. C. R. Alavanja, and C. C. Harris p53 and K-ras Mutations in Lung Cancers from Former and Never-Smoking Women Cancer Res., June 1, 2001; 61(11): 4350 - 4356. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Hainaut and G. P. Pfeifer Patterns of p53 G{->}T transversions in lung cancers reflect the primary mutagenic signature of DNA-damage by tobacco smoke Carcinogenesis, March 1, 2001; 22(3): 367 - 374. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Husgafvel-Pursiainen, P. Boffetta, A. Kannio, F. Nyberg, G. Pershagen, A. Mukeria, V. Constantinescu, C. Fortes, and S. Benhamou p53 Mutations and Exposure to Environmental Tobacco Smoke in a Multicenter Study on Lung Cancer Cancer Res., June 1, 2000; 60(11): 2906 - 2911. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Williams, F. Ponten, C. Moberg, P. Soderkvist, M. Uhlen, J. Ponten, G. Sitbon, and J. Lundeberg A High Frequency of Sequence Alterations Is Due to Formalin Fixation of Archival Specimens Am. J. Pathol., November 1, 1999; 155(5): 1467 - 1471. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Cell Growth & Differentiation |