Published ahead of print on June 9, 2005, doi:10.1164/rccm.200503-439OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200503-439OC 3p Microsatellite Alterations in Exhaled Breath Condensate from Patients with Non–Small Cell Lung CancerInstitute of Respiratory Disease, University of Foggia, Foggia; Istituto di Scienze delle Produzioni Alimentari, Centro Nazionale delle Ricerche; Bari Institute of Respiratory Disease, University of Bari; Clinical Experimental Oncology Lab, National Cancer Institute; and Department of Thoracic Surgery, San Paolo Hospital, Bari, Italy Correspondence and requests for reprints should be addressed to Giovanna Elisiana Carpagnano, M.D., Via De Nicolo 5, 70121 Bari, Italy. E-mail: ge.carpagnano{at}unifg.it
The still-high mortality for lung cancer urgently requires the availability of new, noninvasive diagnostic tools for use in early diagnosis and screening programs. Recently, exhaled breath condensate (EBC) has been proposed as a useful tool to obtain biological information on lung cancer disease. This study provides, for the first time, evidence that DNA alterations already described in lung cancer are detectable in EBC from patients with non–small cell lung cancer (NSCLC) and in healthy subjects. Thirty patients with histologic evidence of NSCLC and 20 healthy subjects were enrolled in the present study. All subjects had allelotyping analysis of DNA from EBC (EBC-DNA) and from whole blood (WB-DNA) of a selected panel of five microsatellites (D3S2338, D3S1266, D3S1300, D3S1304, D3S1289) located in chromosomal region 3p. Results from healthy subjects and subjects with cancer, and from EBC and WB, were compared. In addition, the relationships with smoking habit and clinicopathologic tumor features were considered. Microsatellite alterations (MAs) were found in 53% of EBC-DNA and in 10% of WB-DNA loci investigated in patients with NSCLC (p < 10–6); conversely, MAs were present only in 13% of EBC-DNA and in 2% of WB-DNA informative loci in healthy subjects. In patients with NSCLC, a direct association between number of MAs detected in EBC-DNA and tobacco consumption was observed. We conclude that EBC-DNA is highly sensitive in detecting MA information unique to patients with lung cancer. Furthermore, MA information seems to be directly related with tobacco consumption, and is potentially applicable to screening and early diagnostic programs for patients with NSCLC.
Key Words: breath condensate • DNA • microsatellite • non–small cell lung cancer Lung cancer remains the most frequent type of cancer and cause of cancer death in men, and its incidence recently has increased significantly in women as well (1). Although a large number of potential causes for lung cancerogenesis have been hypothesized, an important role has been universally reported for tobacco use, which accounts for more than 80% of lung cancer development (2). Despite the efforts to find new, more sensitive diagnostic tools able to anticipate the diagnosis of this cancer, a significant percentage of patients will never undergo surgery because of the wide extent of disease at diagnosis. To try to anticipate the diagnosis, screening programs in asymptomatic, high-risk population groups have been considered; toward this goal, cytology of the sputum, circulating tumor biomarkers, chest radiograph, nuclear magnetic resonance (MNR), and similar techniques have been attempted, sometimes with interesting but still inconclusive results (3). A completely different approach involves locating biomolecular markers of lung cancerogenesis or tumor progression, potentially permitting the individualization of early cell damage and noninvasive staging of the disease. In fact, currently it is generally accepted that lung cancer results from the occurrence of a number of genetic alterations in oncogenes and tumor suppressor genes (4–6) that are potential markers either for screening procedures or for earlier detection in patients with non–small cell lung cancer (NSCLC) (5, 7, 8). Several DNA alterations that take place during the development of cancer (gene mutation, microsatellite instability [MI], promoter methylation, and overexpression) have been identified in different biological samples of patients with lung cancer (9). However, tissues used for molecular studies are difficult to harvest because they require highly invasive techniques, which are poorly suited for wider screening programs. Recently, Gessner and colleagues (8) have demonstrated the possibility of detecting human DNA in exhaled breath condensate (EBC). EBC is a fluid that comes from the airways that can be collected by means of an easy, completely noninvasive, repeatable procedure, and which is well accepted by patients (10–15) For this reason, the analysis of genetic characteristics in EBC could represent a means of noninvasive identification of early markers of NCSLC. Recent insights into the molecular basis of cancer have recognized the occurrence of some triggering molecular events likely to result in the development of lung cancer, including MI and loss of heterozygosity (LOH) (16, 17). Microsatellites are repetitive nucleotide sequences of varying lengths, which are scattered throughout the genome, between and within genes. They have been used as markers for genetic mapping because they are highly polymorphic and stably inherited (18–20). Previous studies have shown the presence of microsatellite alterations (MAs) in NSCLC with a variable frequency depending on the number and loci studied and on the clinicopathologic characteristics of the patients (21, 22). Therefore, MA also has been recently proposed as an early marker in lung carcinogenesis (23, 24). The aim of the present study was to check for the feasibility of MA analysis in the EBC-DNA from patients with NSCLC. Furthermore, differences between results from DNA of healthy subjects and of subjects with cancer, and of EBC-DNA and of whole blood DNA (WB-DNA), were compared. Finally, the relationships between smoking habit and clinicopathologic tumor features were considered.
Characteristics of Patients The study population consisted of 30 patients (19 men; mean age ± SD, 63 ± 8 years) who had a histologic diagnosis of NSCLC at the Department of Thoracic Surgery, San Paolo Hospital, Bari, and at the Department of Respiratory Disease, Foggia University. Twenty healthy control subjects, with chest scans negative for cancer, were also enrolled (13 men; mean age ± SD, 61 ± 7 years). Written, informed consent was obtained from all subjects on approval of the study by the ethic committees of the two institutions. All the patients were enrolled in the study immediately after cytohistologic diagnosis before any of them had received any form of anticancer therapy, including primary surgery. The patients underwent standard diagnostic and staging procedures consisting of the following: physical examination; serum chemistry analysis; brain, chest, and abdomen computed tomography scans; radionuclide bone scan; and bronchoscopy. The diagnosis of NSCLC was made either by bronchoscopic biopsy or by transthoracic needle aspiration. Assessments of tumor (T) and node (N) status were based on the International Union against Cancer TNM staging system (25). Overall, patients with NSCLC were classified as stage I in eight cases, stage II in six cases, stage III in seven cases, and stage IV in nine cases. At the time of enrollment into the study, 27 patients were current smokers (with an average tobacco consumption estimated at 43 pack-years; range, 20–100), whereas 3 were ex-smokers (tobacco consumption stopped from 7.6 ± 4.2 years), with a previous average tobacco consumption estimated at 47 pack-years (range, 20–106). Patients who were current smokers were divided into three groups on the basis of tobacco consumption expressed in pack-years (Group 1, < 20 pack-years; Group 2, 20–50 pack-years; Group 3, > 50 pack-years). Ex-smokers were aggregated to Group 1 because of the length of time that had elapsed from smoking cessation. Ten healthy control subjects were also smokers, with a mean tobacco consumption estimated at 44 pack-years. Patients and healthy subjects with concomitant diseases (e.g., infectious, autoimmune disease) were excluded from the study. All the patients and the healthy subjects enrolled underwent EBC and WB collection.
EBC and WB Collection In 10 cases with NSCLC, just after collection, an aliquot of EBC was centrifuged and analyzed for cell viability by trypan blue dye assay performed in a Burker chamber. A mean number of 83 x 106 cells/ml with a mean of 80% of viable cells were found. A further cytologic examination of EBC showed a lymphomononuclear origin of cells; however, most part of the cells resulted disrupted and from the debris size they appear to be epithelial cells. At the same time as the EBC collection, a paired peripheral WB sample (3 ml) was collected from 20 healthy subjects and 28 patients; samples were put into ethylenediaminetetraacetic acid tubes and immediately stored at –80°C.
Microsatellite Analysis All the samples of the EBC DNA were positive for ß-actin gene fragments. EBC- and WB-DNA was amplified by fluorescent polymerase chain reaction (PCR). The analysis of MAs was performed using the following five polymorphic microsatellite markers from chromosome 3p, which account for "hot spots" of deletions in lung cancer believed to be involved in the carcinogenesis of lung cancer: 3p24.2 (D3S2338), 3p23 (D3S1266), 3p14.2 (D3S1300, fragile histidine triad [FHIT] locus), 3p25-26 (D3S1304), and 3p21 (D3S1289) (26, 27). Nucleotide sequences of primers for microsatellite analysis are available through the Genome database (http://www.ncbi.nlm.nih.gov/genemap-99). One of each paired primer was fluorescent-labeled with FAM and EXE (ABI Prism Linkage Mapping Set, Applied Biosystems, Foster City, CA). A total of 50 ng of DNA from EBC and WB was used for each PCR amplification. PCR amplification was performed on 10 ng of EBC- and WB-DNA in duplicate, in a 10-µl final volume, and performed on a GeneAmp 9700 thermal cycler (Applied Biosystems) by combining the template with 0.5 U of AmpliTaq GOLD (Applied Biosystems) in PCR buffer, 0.2 mM of each primer, 125 µM of deoxyribonucleosides triphosphate (dNTPs). The PCR protocol consisted of 35 cycles of 10 minutes at 94°C, 50 minutes at 94°C, 1 minute at 52°C, 2 minutes at 72°C, and 30 minutes at 60°C. A negative control (buffer and enzyme without DNA template) was included in every PCR series. PCR products for each clinical specimen were analyzed by laser fluorescence using an ABI Prism DNA sequencer 310 equipped with GeneScan 2.1 software (Applied Biosystems). This technique allowed for sensitive and quantitative allele ratio estimation by measuring the peak height of both alleles, as previously described (28). Our assay is based on the detection of an alteration in the allele ratio in the EBC-DNA of healthy subjects and of patients with NSCLC when compared with the allele ratio in the paired blood cell DNA of the same healthy subject and patient. LOH and the presence of allele shifts indicating genomic instability were recorded in the various samples of breath condensate and compared with the profile obtained in the DNA from blood cells. LOH was scored when a reduction of at least 30% in allele intensity in the experimental sample was seen. MI was defined as the appearance of a clear novel band that was absent in the lane from the healthy control blood DNA. In our systematic study, each result of amplification was confirmed by at least two independent analyses.
Statistical Analysis
The feasibility of MA analysis in EBC-DNA has been preliminarily evaluated. To this end, EBC-DNA and paired WB-DNA from 20 healthy subjects have been tested and compared for MI and LOH analysis. Good-quality DNA in terms of integrity and amount (mean quantity, 20 ng/µl) was obtained in all EBC samples. MAs were found only in 7 of 20 EBC-DNA and in 2 of 20 WB-DNA samples from healthy subjects. In the EBC-DNA from one subject only, a maximum of three MAs has been shown.
Microsatellite Analysis in EBC-DNA from Patients with NSCLC Considering each microsatellite as informative when heterozygosity was evident (see METHODS), only 19 patients had informative results in all five considered loci; regarding the total number of analyzed loci (five for each of 30 patients), 90% (135/150) of the analyses resulted in informative microsatellites. Four patients did not show alterations in any of the considered markers, five had only one MA, nine presented four markers altered at the same time, whereas none presented simultaneous alteration in all five considered loci. Table 1 shows the results of the microsatellite analysis in EBC-DNA for each locus and patient.
LOH was found in 24% (33/135) and MI in 29% (39/135) of the informative loci studied. The most frequently altered microsatellites in EBC-DNA were D3S1300 (in 61% of the informative DNA samples for that locus) and D3S2338 (in 59% of the informative DNA samples for that locus).
Microsatellite Analysis in WB-DNA from Patients with NSCLC Table 2 shows the results of the microsatellite analysis in the WB-DNA for each locus and patient. LOH was found in 3% (4/127) and MI in 5.5% (7/127) of the informative loci studied; the most frequently altered microsatellite was D3S1304 (20% of the informative DNA samples for that locus).
Comparison of Microsatellite Analysis in EBC-DNA and WB-DNA from Patients with NSCLC Figure 1 shows some typical examples of MI and LOH in EBC-DNA and its paired WB-DNA from one patient.
Data on relationships between EBC-DNA and WB-DNA are summarized in Table 3. EBC-DNA and WB-DNA provided a similar spectrum of informative loci in all the patients (100% of agreement in terms of capability to individualize informative loci); however, a significantly higher number of MAs were present in EBC-DNA with respect to WB-DNA (53 versus 10% of MAs, respectively; p < 10–6). All MAs found in WB-DNA were also evident in EBC-DNA; conversely, 59 of 127 MAs shown in EBC-DNA were not found in WB-DNA.
MAs and Clinicopathologic Features We finally analyzed MAs in relation to clinicopathologic characteristics of the patients. The percentage of patients with at least one MA resulted in similar EBC-DNA from patients with adenocarcinoma with respect to those with squamous carcinoma (86 vs. 89%, respectively). In particular, MI was present in a similar percentage of the 2 histotypes (30 in adenocarcinomas vs. 27% in squamous carcinomas). However, LOH resulted more frequently in squamous than in adenocarcinoma cases (36 vs. 19%, respectively; p = 0.05). Finally, no relationship between number and type of MAs in EBC-DNA and tumor stage was evident.
Regarding WB-DNA, we found a significant relationship only between MA and tumor stage; in fact, MAs were present in 25% of patients with stage I, 17% of patients with stage II, 100% of patients with stage III, and 89% of patients with stage IV disease ( Finally, relationships with smoking habit were considered. The number of MAs present in EBC-DNA and an increase in tobacco consumption were directly related in healthy and patient subgroups (Figure 2). In particular, the frequency of MAs increased from patients in Group I (< 20 packs/year) compared with those in Group III (> 50 packs/year), with a mean number of MAs of 1.4 ± 1.3 versus 3.1 ± 0.9, respectively (p = 0.02). The D3S1300 locus was the most frequently altered in heavy smokers with NSCLC, and this probability decreased from Group III to Group I (from 86% in Group III to 77% in Group II to 0% in Group I).
In addition, the number of MAs in EBC-DNA was significantly higher in Group 1 patients than in healthy subjects consuming the same amount of tobacco per year (mean number of MAs, 1.4 ± 1.3 vs. 0.3 ± 0.6, respectively; p = 0.03).
In the present study, the genetic alterations of microsatellites on chromosome locus 3p were explored in EBC-DNA samples from 30 patients with NSCLC and 20 healthy control subjects. The results of this study showed that 89% of patients with NSCLC exhibited genetic alterations, either MI or LOH, in their EBC-DNA, whereas only 35% of the healthy subjects exhibited these alterations. This evidence suggests interesting clinical perspectives for the analysis of specific lung cancer genetic markers in an easily accessible and organ-specific biological specimen, such as EBC. Some genetic events seem to trigger lung cancerogenesis and precede the morphologic transformation of cells (24). Therefore, identification of genetic alterations in apparently normal cells is the goal for early diagnosis and optimal treatment of this tumor for which only a very weak arsenal seems to be available today (7). Although several genetic alterations involved in the oncogenesis of lung cancer have been candidate markers for early diagnosis, they are mainly harvested in the cells of tumor tissue, which is usually accessible only at the time of surgical resection (i.e., when the tumor is already in an advanced phase) (20, 29). The cancerization theory assumes that genetic alterations are also present in nonmalignant lung tissue adjacent to the tumor and on the entire field of the bronchial tree exposed to the carcinogenic damage (20). For this reason, molecular alterations typical of lung cancer also have been recently explored in cells coming from the airways and collected through bronchoalveolar lavage and induced sputum (9, 23, 24, 29). Recently, Gessner and colleagues (8) demonstrated the possibility of identifying genetic alterations, including the mutations of p53 exons 5-8, in breath condensate. The possibility of collecting this sample in an easy, completely noninvasive, and cheap manner, that is also well accepted by individuals, makes breath condensate a suitable tool for broader routine genetic screenings of the population at risk and for earlier identification of lung tumor. Among the various molecular markers, a growing interest has been recently generated by the analysis of MAs—namely, LOH and MI (23, 30, 31)—involved in the early events and important steps of lung carcinogenesis (23). The MAs more closely related to NSCLC have been identified on the short arm of chromosome 3 (3p) where interesting tumor suppressor genes are located, including transforming growth factor ß type II receptor and FHIT (30, 31). The allelic losses on this locus that usually result in the inactivation of these genes have been demonstrated to be involved in tumor initiation and progression of several cancers, including those of the lung (7). The exact mechanism causing MAs in lung cancer remains still unknown, even if it seems to differ from the process causing mismatch repair defects (32). MAs have been largely studied in the DNA of the serum, induced sputum, bronchoalveolar lavage, and tumor tissue of patients with NSCLC (9, 23, 24, 28, 29, 33). The frequency of the reported MAs in lung cancer varies from 2 to 55% as a function of the specific microsatellite loci examined and clinicopathologic characteristics of the series analyzed (21, 22). This study has for the first time investigated the possibility of detecting the MAs present in the EBC-DNA and comparing the obtained results with those in paired WB-DNA. The limited presence of genetic alterations in EBC-DNA samples of healthy subjects further supports the hypothesis that this new approach could be highly specific for detection of molecular alterations related to cancerogenesis. Like other studies, which reported a higher frequency of LOH at locus 3p in cancer cells, a greater percentage of MI was observed in the EBC-DNA of patients with NSCLC enrolled in this study in comparison with healthy individuals (24, 28, 34, 35). As previously reported, a larger number of MAs have been observed in adenocarcinoma than in squamous carcinoma, probably due to the association of adenocarcinoma with large airways, in more direct contact with EBC (24). Conversely, in squamous carcinoma, a prevalence of LOH was found. These findings, which conflict with those of Park and coworkers (20), who found no correlation between presence of MAs and histologic subtype of NSCLC, seem to support the hypothesis of Zhou and colleagues (23), suggesting that the mechanisms of tumorigenesis could be different in various histologic lung subtypes. It has long been known that the concentration of free-circulating DNA in plasma is greater in patients with tumor (28, 33). This free-circulating WB-DNA also comes from tumor cells, although the way it is released into the bloodstream remains to be definitively clarified. Several studies have demonstrated the presence of genetic alterations in the WB-DNA of patients with cancer, thus supporting the possibility of recognizing it through the use of molecular tests (33, 38, 39). However, in the present study, WB-DNA was investigated and showed only 10% of MAs, most of them located in the same locus as the paired EBC-DNA. As expected, WB-DNA contained fewer MA than EBC-DNA. No significant difference in terms of percentage of MAs at the different tumor stages was observed in DNA of the breath condensate. Conversely, the percentages of MAs in WB-DNA increased as a function of tumor stage, perhaps due to the amount of circulating DNA present in blood at higher disease stages. This suggested an important role for the microsatellite analysis of the plasma DNA, not only in follow-up but also for early diagnosis of NSCLC, as suggested by Sozzi and colleagues (33). This seems to suggest a possible role for EBC microsatellite analysis as an early marker of carcinogenesis or of susceptibility. Recently MA has been assumed to be the expression of carcinogen exposure, including cigarette smoke (20, 24, 27). A high incidence of MAs has, in fact, been reported in both former and current smokers (23, 24, 36, 37). This study confirms the presence of a parallel increase of MA number and tobacco consumption already reported by other investigators (41–43); we further show that these alterations can be better detected in EBC-DNA. These results are further corroborated by the evidence of the prevalent alteration of one loci, DS1300, which could have specific pathogenetic relevance. This marker is in the fragile site FRA3B of FHIT gene (in intron 5), thus supporting the strong association between this gene inactivation and carcinogenesis. Studies are ongoing to elucidate this aspect and to verify the possibility of using the alterations of the D3S1300 locus as a marker of exposure to tobacco carcinogens. This information seems to us one of the most important results of our study, directly suggesting that our assay performed on an airway product (i.e., EBC) could be a better and direct expression of lung cell exposure to carcinogens. Studies are ongoing on larger series of healthy individuals to validate the idea that EBC could be used to quantify the DNA alterations due to the cumulative exposure to smoking carcinogens. In conclusion, our results provide evidence that it is possible to investigate somatic MAs in the breath condensate DNA of subjects with NSCLC. The noninvasiveness of breath condensate collection and the important role of MA in lung carcinogenesis make these results potentially relevant not only for the follow-up of patients with NSCLC but also for the screening of high-risk populations. Further cytogenetic investigations that examine a larger number of microsatellite markers in patients and healthy subjects are needed to verify the potential of these findings in terms of clinical application.
Supported in part by special project "Programma Italia-USA–Farmacogenomica Oncologica" 2004. Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form March 21, 2005; accepted in final form June 6, 2005
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 tests were used to compare qualitative data. A Wilcoxon test was used for comparison between categories. Data were expressed as means ± SD. Significance was defined as a p value of less than 0.05. 
