|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Cyclin D1 Overexpression in Bronchial Epithelia of Patients With Lung Cancer Is Associated With Smoking and Predicts Survival
From the Department of Clinical Research and Institute of Pathology, Institute of Thoracic Surgery, Institute of Medical Oncology, Inselspital Bern, University of Bern, Switzerland; and Roy Castle Centre, Liverpool, UK. Address reprint requests to D.C. Betticher, MD, Institute of Medical Oncology, University Hospital of Bern, 3010 Bern, Switzerland; email: daniel.betticher{at}insel.ch.
Purpose: Cyclin D1 is overexpressed in almost 60% of resectable non–small-cell lung cancer (NSCLC). In the absence of cyclin D1 gene amplification, overexpression is characterized by allelic imbalanced transcript levels. Methods: The aims were to study cyclin D1 expression by immunohistochemistry and allelic balance of transcripts in tumor-free bronchial epithelia from patients with resectable NSCLC by using monoclonal antibodies (48 patients and 288 sites), microdissection/reverse transcriptase polymerase chain reaction/restriction fragment length polymorphism analyses (24 patients and 144 sites). Derived data were related to patient characteristics—in particular, smoking habits. Results: In 167 (58%) of 288 sites, cyclin D1 was overexpressed, with cytoplasmic and nuclear sublocalization in 53% and 7% of all sites, respectively. Nuclear overexpression was more frequent in premalignant versus normal or hyperplastic epithelia (55% v 3%; P < .0001). Allele-specific expression imbalances were found in 69 (48%) of 144 sites; in particular, those in which cyclin D1 was overexpressed (P = .004). In 14 (58%) of 24 patients, balanced or imbalanced transcript ratios and degree of expression were consistent at all sites for the same patient, whereas in another 10 patients, transcript balances and cyclin D1 expression patterns varied across the sites. Nuclear cyclin D1 expression in at least one site (14 of 48 patients) was linked to heavy smoking (> 40 pack-years; P = .02) and shorter overall survival (P = .01). Conclusion: Allele-specific, probably damage-driven, deregulation of the cyclin D1 gene may precede and perhaps facilitate the spread of preneoplastic clones across the bronchial epithelial surface in a significant number of patients. Cyclin D1 expression at multiple bronchial sites may identify a subgroup of heavy-smoking patients with poor outcome.
IT IS today commonly accepted that more than 80% of lung cancers can be attributed to tobacco exposure.1 The tracheobronchial tree of the persistent smoker is repeatedly exposed to the particular carcinogenic agents present in tobacco smoke. These agents, at least in part, drive the accumulation of successive mutations that force progression from normal tissue to invasive carcinoma. Much evidence has been provided that human cancer is the end result of the stepwise accumulation of multiple genetic alterations. The progression from a normal to a malignant tissue is a process of increasing structural disorder, and this disorder is driven by the sequential genetic damage acquired by an increasingly abnormal clone of cells. Within a tissue that is heavily exposed to carcinogenic agents, it might be expected that this process would be relatively common and that we might observe various developing preneoplastic lesions within such a tissue. Indeed, this would seem to be the case. Fifty years ago, Slaughter et al2 proposed that such observations represented evidence for a process termed field cancerization: This process describes the diffuse mucosal changes of the tracheobronchial tree that occur over years of exposure to a carcinogen. This concept may explain the high incidence of secondary lung cancer after successful resection of a tumor. In fact, the lifetime incidence of a second primary tumor, defined as a second lung cancer of different histology, is more than 10%, with an annual incidence of 1% to 5%, depending on the subgroup. This is a significantly higher rate than even in the heaviest-smoking population.3 We have therefore considered two concepts: first, that multiple genetic lesions, mutations that have occurred sequentially over time within a cell, cause malignancy, and second, that histologically identifiable, preneoplastic lesions may occur at different sites across a particular organ or tissue, with a frequency at least partially related to the exposure of that tissue to carcinogenic agents. Such concepts enable us to build models of neoplastic transformation and to test those models through the observation of preneoplastic lesions in patient-derived material. In the first instance, we can assume that the transformation process is driven through selective advantage. Simply, under given conditions, if a mutation confers onto a cell an increased proliferation rate over its neighbors—for example, through the alteration of cell-cycle regulators over its neighbors or a reduced tendency to die—then that cell will form an expanding clonal population of mutant cells. New mutations conferring further growth advantage within members of this colony will result in further subclonal expansion and perhaps even a wider geographical spread. As this process continues, driven by repeated carcinogen exposure, we can envisage the generation of a series of clones, with each new subclone being more aggressive than its predecessors. In the case of the lung cancer patient, at least one of these daughter clones will have developed an invasive phenotype and moved to full malignancy. There are two nonexclusive ways in which a field cancerization may be generated. Preneoplastic lesions across the tracheobronchial tree of a patient may have a common origin, being all related to one geographically disseminated clone, or lesions may arise independently; they may be clones that do not share a mutational heritage. Perhaps more likely is a situation in which both mutationally related and mutationally unrelated clonal lesions develop within the epithelia of particular patients. There are a number of ways to examine this question of origin, but we must always bear in mind that the failure to identify a common mutation among geographically or temporally disseminated lesions does not imply independence of origin, because we may simply not have detected the critical early alterations. Conversely, the observation of common changes—for example, shared breakpoints, losses of heterozygosity, or point mutations—might imply such clonal relatedness. We must be aware, although it is perhaps unlikely, that such changes could alternatively reflect a genetically acquired tendency toward the mutational events. A number of investigations have demonstrated that nonmalignant lung tissue may harbor genetically damaged clones of cells. In recent studies, loss of one allele of chromosomes 3p, 9p, and 17p was observed in normal bronchial mucosa at different locations of current and former smokers.4–8 Similarly, the mutation TP53 has been detected in the histologically normal lung tissue of lung cancer patients.9 Conversely, loss-of-heterozygosity analysis of multiple sites from a number of patients failed to show a common clonal change in such lesions, an observation that would be consistent with independence of origin.7 Our analysis of cell-cycle gene alterations in the primary tumors of non–small-cell lung cancer (NSCLC) patients, in particular those focused on the G1 restriction point control elements, suggested that the deregulation of cyclin D1 expression was an important characteristic of this disease.10,11 We and others extended the study of cyclin D1 expression and were able to report increased levels of the protein in the tumor-free bronchial epithelia of patients.12,14 In our initial study, we were able to show, through the analysis of a cDNA-located restriction fragment length polymorphism (RFLP), that overexpression of cyclin D1 in primary lesions was associated with imbalance in the tumor tissue in the level of transcripts from each parental CCND1 (cyclin D1) allele. Given that in a normal or quasi-normal situation, genes are likely to be upregulated biallelically, such disease-specific imbalances in expression in tumor tissue indicate the presence of cis-located genetic alteration: DNA-damage events that result in the upregulation or alternative processing of one allele. In this study, we have extended our investigations of the involvement of cyclin D1 in early disease by examining expression at the DNA, RNA, and protein level in microdissected bronchial epithelial cells at six different sites distant from the primary NSCLC. The analysis of CCND1 expression patterns has allowed us to augment our understanding of lung cancer pathogenesis and to identify two distinct groups of patients with apparently distinct outcome.
Patient Characteristics and Tumor Specimens Tumor samples and bronchial epithelia were collected prospectively from consecutive patients (37 men and 11 women; median age, 65.3 years; range, 30.6 to 80.7 years) who underwent resection of staged resectable NSCLC at the University Hospital of Berne, Switzerland (segmentectomy [n = 2], lobectomy [n = 32], or pneumonectomy [n = 14]). Seventeen patients were in stage I, 12 in stage II, and 19 in stage III (International Union Against Cancer).15 All tumors were classified, according to the standard criteria of the World Health Organization,16 by the local pathologists and then reviewed by our pathologist (M.G.); there were 20 squamous carcinomas, 24 adenocarcinomas, three large-cell carcinomas, and one poorly differentiated NSCLC. Tumor size was measured on the fresh specimen by the pathologists. Twenty-six of 48 patients are alive with a median observation time of 16 months (range, 0.5 to 33 months). The specimens (primary tumor and six sites distant to the tumor) were snap-frozen in liquid nitrogen, embedded in TissueTek (Sakura Finetek, Zoeterwoude, the Netherlands) and kept at -80°C. Simultaneously, a piece of each specimen was processed in formalin and embedded in paraffin for immunohistochemistry.
Microdissection
DNA Extraction and Polymerase Chain Reaction Amplification
RNA Extraction and One-Step Reverse Transcriptase PCR
RFLP Analysis
Immunohistochemistry Control sections were treated accordingly, and the primary antibody mix was replaced with buffer. Sections were then incubated in a 1/300 dilution (same diluent as above) of biotinylated goat antimouse immunoglobulins (Dako) and thereafter with avidin-biotin complex/horseradish peroxidase (1/200 in TBS; Dako). Finally, sections were developed in 3,3-diaminobenzidine (Sigma), counterstained with hemalaun, and mounted. Cyclin D1 staining of cells such as lymphocytes, plasma, and endothelial cells, which were mainly negative, was used as an internal control. Throughout the study, sections from an NSCLC known to stain positive for cyclin D1 were analyzed in parallel to serve as positive controls. The sections were reviewed by our pathologist (M.G.) in a blinded fashion, considering histology, grading of differentiation, and cyclin D1 staining. Epithelial changes were graded according to the criteria reviewed by Travis et al.16 Squamous metaplasia is the first clear-cut morphologic change in the respiratory epithelium induced by smoking. Because our interest was focused on early lesions, we grouped them in a common group of preinvasive lesions. In detail, regarding cyclin D1 staining, the following criteria were used for cytoplasmic staining, nuclear staining, or both: for the intensity, -, +, ++, and +++; and for the frequency, 0% to 5%, 5% to 50%, and 50% to 100%. For data grouping, a bronchial site was considered to be positive for nuclear staining with + to +++ intensity and more than 5% positive cells and to be positive for cytoplasmic staining with a ++ to +++ intensity and more than 50% positive cells.
Statistical Analyses
Patient Characteristics and Bronchial Site Description Six bronchial epithelial specimens distant to the primary tumor (in different lobes, segments, or both), preserved at the time of tumor resection (snap-frozen and formalin-fixed), were obtained from 48 patients with a resectable NSCLC; thus, 288 bronchial sites were studied. Patient and tumor characteristics are listed in Table 1  . Normal and hyperplastic epithelia were most common (n = 35 and n = 232), whereas preinvasive lesions occurred in 20 of 288 sites, respectively, in 19 of 48 patients. Patients in whom we observed preinvasive lesions were more likely to have a higher stage of disease (Table 1) and a worse survival (Fig 1).
Cyclin D1 Expression in Bronchial Epithelia Cyclin D1 overexpression was found in 167 of 288 and 22 of 48 bronchial sites and tumors, respectively (Fig 2 ). Cyclin D1 was predominantly localized to the cytoplasm (in 53% of all sites studied), whereas nuclear cyclin D1 was less frequent (7% of all sites). The results are listed in Table 2. The frequency of nuclear cyclin D1 overexpression increased significantly with the grade of the preinvasive lesion from normal/hyperplastic (3%) to metaplastic/dysplastic epithelia (55%; P < .0001) and was not different at a distance from the primary tumor. In four patients only, we detected no cyclin D1 overexpression at all bronchial sites analyzed; no altered cyclin D1 expression was found in the corresponding tumor. Nuclear cyclin D1 expression in at least one bronchial site (14 of 48 patients) was linked to heavy smoking (> 40 pack-years; P = .02). Cytoplasmic cyclin D1 staining was correlated with a better remission duration and overall survival.
Alterations of the CCND1 Gene in Bronchial Epithelia A polymorphism of cyclin D1 located in the 3' UTR allows a PCR-based (or RT-PCR-based) RFLP determination of allelic balance. The 48 patients were first genotyped for the CCND1 gene polymorphism18; 24 (50%) of 48 were informative. As we previously demonstrated by using Southern blotting and PCR multiplex in NSCLC primary tumors, an allelic imbalance at the CCND1 locus correlates, in most cases, with amplification of the gene.11 Similarly, in this study, we found in five (21%) of 24 tumors a CCND1 allelic imbalance compatible with gene amplification (Fig 3 ). In four of these cases, cyclin D1 was overexpressed. To investigate whether this alteration also occurred in other bronchial sites, we performed RFLP DNA analyses in microdissected epithelia from a number of informative patients (n = 7; 42 bronchial sites). In two (5%) of 42 sites analyzed, we found allelic imbalances consistent with gene amplifications. In both patients, the sites with the imbalance were distant to the primary tumor, allowing exclusion of contamination of the analyzed cells by primary tumor cells. In both sections, immunostaining of epithelial cells revealed overexpression of cyclin D1. On the basis of these results, we conclude that CCND1 gene amplification may occur in premalignant epithelia, but it may represent a rare event in the pathogenesis of lung cancer and is therefore unlikely to explain the large number of sites exhibiting cyclin D1 overexpression.
Allele-Specific Cyclin D1 Expression As a next step, we looked at expression at the transcript level. To generate pure populations of epithelial and tumor cells, all bronchial sites and tumors of the 24 informative patients were microdissected. The allelic balance of CCND1 cDNA was determined by RFLP analysis of RT-PCR products. We found imbalanced expression in 12 of 24 tumors and 69 of 144 sites, respectively (Fig 4 ). Imbalanced expression was associated with cyclin D1 overexpression (P = .006). Taken together, three groups of patients could be distinguished: (1) in a first small group (n = 2), balanced allelic expression was found at all sites, and the tumor and no cyclin D1 overexpression was found at any site, including the tumor; (2) in a second group (n = 10; 42%), we detected a similar allelically imbalanced RNA expression and cyclin D1 overexpression at all sites and the tumor; and (3) the third group of patients (n = 12; 50%) was characterized by a variable status of balanced or imbalanced allelic expression and cyclin D1 expression or overexpression in different samples. Sites exhibiting an identical cyclin D1 overexpression in terms of intensities, frequencies, and subcellular localizations (see description in Methods) were considered to belong to the group with similar alterations. In all patients with similar allelic imbalances at all sites (n = 12), the same imbalance was found in the tumor (with the exception of three patients).
Association of Cyclin D1 Expression and Clinical Outcome Patients with similar cyclin D1 overexpression at all sites (n = 14) were compared with patients with variable cyclin D1 expression at different bronchial sites (n = 34). There was no association found for sex, age, histology (including differentiation), stage of disease, tumor size, or altered cyclin D1 expression. Cigarette smoking correlated with altered cyclin D1 expression: heavy smokers (> 40 pack-years) were more likely to belong to patients with similar cyclin D1 altered expression at all sites (P = .03; Table 3 ).
Patients with no nuclear cyclin D1 overexpression at any site (n = 34) were compared with patients who exhibited nuclear cyclin D1 staining at at least one site (n = 14). This last group of patients was characterized by a shorter overall survival (Fig 5 ). Patient characteristics for both groups are listed in Table 1 . The probability of shorter survival time was maintained after adjustment for age, stage, number of pack-years, and sex (P = .006; Table 3). The estimated relative risk of shorter survival was 0.09 in the group with nuclear cyclin D1 staining (95% confidence interval, 0.016 to 0.5). On the other hand, cytoplasmic cyclin D1 appears to be associated with a prolongation of remission duration in patients with overexpressed cyclin D1 in at least one epithelial site (Fig 6).
Our results from the analysis of geographically disseminated bronchial sites indicate that the alteration of cyclin D1 expression may be an early step in the neoplastic transformation of bronchial epithelia. Further, they support the model that such deregulation may, in some patients, predate geographical dissemination and expansion of histologically normal but genetically abnormal epithelial clones. The results are in agreement with our previous work on the resection margins of NSCLC patients12 and with those of Brambilla et al and Lonardo et al.13,14 This study, together with the previous work, confirms by immunostaining that cyclin D1 is overexpressed at high frequency in tumor-free epithelia. A deeper analysis was therefore necessary to study the nature and significance of this overexpression. Our analysis indicates that CCND1 plays a key role in the pathogenesis of NSCLC. Underpinning the overexpression of this gene in tumors, at high frequency, we see a bias in transcription toward one or another parental allele.11 We have extended this observation from primary tumors to geographically distant, histologically tumor-free epithelial sites in the same patients, and we have shown that the same abnormal bias in expression may be seen at multiple distant epithelial sites. At the DNA level, we have also demonstrated that allelic imbalance of CCND1, consistent with gene amplification events, may be detected in preneoplastic tissue. The association of cyclin D1 overexpression and the allelically imbalanced mRNA expression observed indicates that CCND1 induction is caused by specific genetic events that affect just one parental allele. Thus, the increased expression seen in the epithelial cells is not simply a result of nonspecific upregulation of a gene involved in rapid cell growth triggered, for instance, by cigarette smoking, but is likely to be the result of genetic alterations that directly affect one homolog of CCND1. Such alterations might include mutations that affect the splicing of the gene, the stability of the mRNA, or, perhaps, translocations or epigenetic modifications that result in the relative overexpression of one or other parental allele. In addition to the observation that we may see multiple independent sites with a consistent transcript imbalance, we have also seen different epithelial regions from the same patient with imbalances in one or another parental allele, depending on the site. This observation again underlines the importance of deregulation of CCND1 at the earliest stages of neoplastic transformation; it strengthens the argument that DNA damage or allele-specific modification generates the transcript imbalances that we see. We previously demonstrated,11 in a subset of tumors with CCND1 amplification and cyclin D1 overexpression, that the transcript imbalance seen in cDNA seemed to favor the nonamplified allele. Because the polymorphic restriction site is located distant to the coding sequence, within the long 3' UTR, the most likely explanation for this finding is that the mRNA encoded by the amplified allele is preferentially spliced, precluding amplification of the RFLP site. In our subsequent work,10 we described a further transcript (transcript b) that does not include the conventional 3' UTR, but terminates within intron 4. The encoded peptide from such a transcript would lack the exon 5–specified PEST region. Further studies in bronchial epithelia would be of interest to examine the balance of these transcripts in the target tissue. Another interesting finding was the subcellular localization of cyclin D1, which was, in some cases, exclusively in the cytoplasm of the bronchial cells. Cytoplasmic cyclin D1 has been reported in a series of solid tumors, such as lung, breast, and colorectal cancer,11,20–23 and is, at least in some tumors, believed to be associated with cancer development.22 In our study, nuclear cyclin D1 sublocalization was associated with preinvasive lesions. The transfer of cyclin D1 to the nucleus has been shown to be induced by p21 and p27 proteins—both cyclin-dependent kinase inhibitors.20,24,25 In our previous work on p21 and cyclin D1 in NSCLC,26 we described an association of p21 overexpression and cyclin D1 nuclear accumulation. In in vitro studies, we demonstrated that when NIH3T3 cells were transfected with a mutant form of cytoplasmic cyclin D1, enforced expression of p21 drove the mutant cyclin D1/cyclin-dependent kinase 4 complex into the nucleus.24,25,27 A similar regulatory mechanism has been described for p27.24 Recently, all trans-retinoic acid has been reported to increase the protein level of p27.28 Taking these data together, we may hypothesize that all trans-retinoic acid would be able to force the transfer of cyclin D1 to the nucleus through the expression of cyclin-dependent kinase inhibitors. It is tempting to speculate that this might provide at least a partial hypothetical explanation of why vitamin A and its derivatives apparently promote the development of lung cancer, as seen in large randomized prevention trials in smokers.29,30 Our finding of allelic expression imbalances being strongly associated with cyclin D1 overexpression may indicate alternate splicing of cyclin D1 transcripts. Langenfeld et al31 showed, based on experiments performed on cell lines of the gastrointestinal tract, that retinoic acid–induced cyclin D1 proteolysis reduced growth at the G1 phase. These results are consistent with the well-known activity of retinoids to prevent second primary tumors in the aerodigestive tract (treatment of premalignant lesions such as oral leukoplakia). Because retinoic acid derivative intake increases the risk of lung cancer, another mechanism has to be postulated to explain their effect in this disease. In light of our findings, we have therefore discussed the possibility that effects on cyclin D1 subcellular localization might play a part in this process. In our previous report,10 we described two distinct transcripts leading to similar proteins, both of which were detected with the antibodies used. The protein originating from transcript b lacks the PEST region; thus, a change in the half-life of this protein is postulated. Further experiments in the bronchial epithelia demonstrated the existence of both transcripts and showed that the transcript a/b ratio differs from site to site. Further analyses of this finding are under way. Metaplasia, dysplasia, and carcinoma-in-situ dispersed throughout the tracheobronchial tree are frequently observed alterations in smokers and are believed to represent initiated and premalignant stages of malignant cells. It has been reported that high-grade dysplasias and carcinoma-in-situ demonstrate a significant increase of total number of allelic losses as compared with hyperplastic and normal epithelium. In fact, specific parental alleles were lost in chromosomal deletions present in preneoplastic lesions at different sites and their accompanying cancers.6 Others, however, have not noted such similarities in allele-specific mutations in preinvasive lesions.7 These apparently contradictory findings might be explained at least in part by our results, because we see an apparently common relationship between lesions in some patients but do not observe such relationships in others. Indeed, both mechanisms probably play a major role in the disease, varying lesion by lesion and patient by patient.
In our study, metaplastic and dysplastic epithelia were more likely to show cyclin D1 overexpression localized to the nucleus, a finding that was associated with smoking habit and poorer outcome, in particular, with shorter overall survival (5 v 18 months; P = .01), a large difference that remained in multivariate analysis (Fig 5 In conclusion, our study of the bronchial epithelia of patients with resectable NSCLC is the first to show that, first, cyclin D1 is overexpressed in most cases, and the protein is mainly localized to the cytoplasm; second, nuclear localization of cyclin D1 is associated with premalignant transformation; third, the distribution of similar allelically imbalanced expression, together with cyclin D1 overexpression, supports the concept of the clonal dispersion of a malignant or premalignant cell across the lung; and, finally, bronchial epithelia with nuclear cyclin D1 staining occurs especially in heavy smokers and predicts poor outcome. Further independent investigations will, therefore, be valuable in confirming these associations. The work further strengthens the argument that cyclin D1 might be an effective lung cancer therapeutic target. It also raises the possibility of using cyclin D1 as a molecular marker to identify high-risk individuals and may allow us to develop effective surveillance and early intervention strategies.
We are grateful to B. Hügli and B. Steiner for their technical help.
This work was funded by the Swiss Cancer League (KFS-703-8-1998). J.H. is funded by the Roy-Castle-Foundation, UK.
1. Shopland DR, Eyre HJ, Pechacek TF: Smoking-attributable cancer mortality in 1991: Is lung cancer now the leading cause of death among smokers in the United States? J Natl Cancer Inst 83:1142–1148, 1991 2. Slaughter DP, Southwick HW, Smejkal W: "Field cancerization" in oral stratified squamous epithelium: Clinical implications of multicentric origin. Cancer 6:963–968, 1953[CrossRef][Medline] 3. Grover FL, Piantadosi S: Recurrence and survival following resection of bronchioloalveolar carcinoma of the lung: The Lung Cancer Study Group experience. Ann Surg 209:779–790, 1989[Medline]
4. Wistuba II, Lam S, Behrens C, et al: Molecular damage in the bronchial epithelium of current and former smokers. J Natl Cancer Inst 89:1366–1373, 1997
5. Mao L, Lee JS, Kurie JM, et al: Clonal genetic alterations in the lungs of current and former smokers. J Natl Cancer Inst 89:857–862, 1997
6. Kishimoto Y, Sugio K, Hung JY, et al: Allele-specific loss in chromosome 9p loci in preneoplastic lesions accompanying non-small-cell lung cancers. J Natl Cancer Inst 87:1224–1229, 1995 7. Gazdar AF, Lam S, Wistuba II: Molecular and cytologic techniques of early detection, in: Pass HI, Mitchell JB, Johnson DH, et al (eds): Lung Cancer: Principles and Practice (ed 2). Philadelphia, PA, Lippincott, Williams & Wilkins, 2000, pp 407–424 8. Caballero OL, Cohen D, Liu Q, et al: Loss of chromosome arms 3p and 9p and inactivation of P16 (INK4a) in normal epithelium of patients with primary lung cancer. Genes Chromosomes Cancer 32:119–125, 2001[CrossRef][Medline] 9. Franklin WA, Gazdar AF, Haney J, et al: Widely dispersed p53 mutation in respiratory epithelium: A novel mechanism for field carcinogenesis. J Clin Invest 100:2133–2137, 1997[Medline] 10. Betticher DC, Thatcher N, Altermatt HJ, et al: Alternate splicing produces a novel cyclin D1 transcript. Oncogene 11:1005–1011, 1995[Medline] 11. Betticher DC, Heighway J, Hasleton PS, et al: Prognostic significance of CCND1 (cyclin D1) overexpression in primary resected non-small-cell lung cancer. Br J Cancer 73:294–300, 1996[Medline] 12. Betticher DC, Heighway J, Thatcher N, et al: Abnormal expression of CCND1 and RB1 in resection margin epithelia of lung cancer patients. Br J Cancer 75:1761–1768, 1997[Medline]
13. Brambilla E, Gazzeri S, Moro D, et al: Alterations of Rb pathway (Rb-p16INK4-cyclin D1) in preinvasive bronchial lesions. Clin Cancer Res 5:243–250, 1999
14. Lonardo F, Rusch V, Langenfeld J, et al: Overexpression of cyclins D1 and E is frequent in bronchial preneoplasia and precedes squamous cell carcinoma development. Cancer Res 59:2470–2476, 1999 15. Mountain CF: The new International Staging System for Lung Cancer. Surg Clin North Am 67:925–935, 1987[Medline] 16. Travis W, Colby T, Corrin B, et al: Histological Typing of Lung and Pleural Tumours (ed 3). Berlin, Germany, Springer, 1999 17. Fink L, Seeger W, Ermert L, et al: Real-time quantitative RT-PCR after laser-assisted cell picking. Nat Med 4:1329–1333, 1998[CrossRef][Medline]
18. Heighway J: HaeIII polymorphism within 3' untranslated region of PRAD1. Nucleic Acids Res 19:5451, 1991 19. Cox D: Regression models and life tables (with discussion). J R Stat Soc B 34:187–220, 1972 20. McKay JA, Douglas JJ, Ross VG, et al: Cyclin D1 protein expression and gene polymorphism in colorectal cancer: Aberdeen Colorectal Initiative. Int J Cancer 88:77–81, 2000[CrossRef][Medline] 21. Marsh KL, Varley JM: Frequent alterations of cell cycle regulators in early-stage breast lesions as detected by immunohistochemistry. Br J Cancer 77:1460–1468, 1998[Medline] 22. Arber N, Hibshoosh H, Moss SF, et al: Increased expression of cyclin D1 is an early event in multistage colorectal carcinogenesis. Gastroenterology 110:669–674, 1996[CrossRef][Medline] 23. Palmqvist R, Stenling R, Oberg A, et al: Expression of cyclin D1 and retinoblastoma protein in colorectal cancer. Eur J Cancer 34:1575–1581, 1998[CrossRef][Medline] 24. Diehl JA, Sherr CJ: A dominant-negative cyclin D1 mutant prevents nuclear import of cyclin-dependent kinase 4 (CDK4) and its phosphorylation by CDK-activating kinase. Mol Cell Biol 17:7362–7374, 1997[Abstract]
25. LaBaer J, Garrett MD, Stevenson LF, et al: New functional activities for the p21 family of CDK inhibitors. Genes Dev 11:847–862, 1997 26. Vonlanthen S, Heighway J, Kappeler A, et al: p21 is associated with cyclin D1, p16INK4a and pRb expression in resectable non-small cell lung cancer. Int J Oncol 16:951–957, 2000[Medline] 27. Cheng M, Olivier P, Diehl JA, et al: The p21(Cip1) and p27(Kip1) CDK ‘inhibitors’ are essential activators of cyclin D-dependent kinases in murine fibroblasts. EMBO J 18:1571–1583, 1999[CrossRef][Medline]
28. Sueoka N, Lee HY, Walsh GL, et al: Posttranslational mechanisms contribute to the suppression of specific cyclin:CDK complexes by all-trans retinoic acid in human bronchial epithelial cells. Cancer Res 59:3838–3844, 1999
29. Omenn GS, Goodman GE, Thornquist MD, et al: Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334:1150–1155, 1996
30. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers: The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. N Engl J Med 330:1029–1035, 1994
31. Langenfeld J, Kiyokawa H, Sekula D, et al: Posttranslational regulation of cyclin D1 by retinoic acid: A chemoprevention mechanism. Proc Natl Acad Sci U S A 94:12070–12074, 1997 Submitted March 20, 2002; accepted February 6, 2003.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 tests for categorical features and Mann-Whitney U tests for continuous ones. Overall survival was analyzed with Kaplan-Meier survivor function estimates, and simple comparisons between the two groups were made with the log-rank test. Cox regression
