Clones and Subclones in the Lung Cancer Field

Based on a variety of clinical (1,2), histologic (5-7), and molecular (8-10) data, lung cancer, like many other epithelial cancers, is suggested to develop as a multistep tumorigenesis process in a field of cancerization (11,12). Carcinogen exposure (e.g., tobacco) is believed to induce genetic damage in target cells that accumulates over years of exposure. Moreover, chronic tissue injury leads to chronic wound healing, whereby cells carrying genetic alterations that are important for survival and preferential regrowth expand at the expense of their neighbors. These surviving colonies continue to accumulate new tissue-relevant genetic alterations and eventually evolve into an invasive lesion. Since the whole aerodigestive tract is similarly exposed to carcinogen, one might predict that such clonal outgrowths develop throughout the exposed tissue. Until recently, characterization of the fine structure of such a multifocal and multistep tumorigenesis process in lung tissue was technically hampered by the need for tissue dissociation either for cell culture or for bulk molecular genetic or cytogenetic analyses. The advent of molecular techniques that could be used on smaller amounts of tissue, including polymerase chain reaction on microdissected epithelium (15-17), increased the spatial resolution of the analyses. Moreover, the identification of molecular changes that occurred at relatively high frequencies in lung tumors provided informative markers that could be used on small tissue samples to query the nature of the lung cancerization field (18,19). As a result, several groups have now demonstrated the presence of clonal outgrowths in the epithelial field of lung tumors, with higher frequencies of genetic changes detectable in dysplastic lesions than in normal-appearing and metaplastic lesions (20-24). These clonal outgrowths frequently exhibit different molecular changes from the primary tumor, suggesting that some foci are evolving through a distinct multistep tumorigenesis pathway in the lung cancerization field. Moreover, this group of investigators as well as our own have shown that such clonal patches can be observed by use of similar techniques in lung epithelium of individuals at increased risk for lung cancer who have had a long period of tobacco exposure (25,26). In this issue of the Journal, Park et al. (27) have taken clonal analysis of the normal and histologically altered lung epithelium in the field of lung cancers to an even finer level. This group of investigators has carried out serial microdissection (laser capture and manual micromanipulator-assisted scraping) of sequential patches of 200-1000 bronchial epithelial cells from paraffin-embedded, resected lung tumor tissue or scrapings from the surface of resected lung tumor tissue by use of a recently described epithelial aggregate separation and isolation procedure (28). The isolated DNA from each of these small cell patches was then subjected to polymerase chain reaction analysis using primers for 12 microsatellite markers at seven chromosomal regions frequently altered in lung cancer. Of interest, molecular evidence for clonal outgrowth was observed in at least one focus of nonmalignant epithelium from 13 of 19 lung resections examined, and one third of the foci examined showed at least one molecular change (e.g., loss of heterozygosity or microsatellite alterations). Although the finding of clonal outgrowths in lung epithelium in the field of lung cancers is not a new finding, two observations in this study are particularly fascinating. First, adjacent patches of 200-1000 lung epithelial cells can show distinct molecular changes, suggesting that they may represent different evolving clones (or subclones). The authors estimate, for example, that most of the clonal patches that they examined contained approximately 90 000 cells. Second, the frequency of such subclonal changes is even higher than that previously observed when larger regions of epithelium were examined by similar techniques. The authors suggest that their use of fluorescence endobronchoscopy (29) might have partially accounted for the unusually high frequency of detected clonal outgrowths; however, it would not account for the frequent observation of small adjacent but molecularly distinct foci. These findings complement our own studies using chromosome in situ hybridization techniques, where even smaller patches of monosomic and polysomic outgrowths can be detected in the lung tissue in individuals at risk for lung cancer (24,30,31). Together, these findings raise the possibility that, during years of chronic exposure, the tobacco-exposed lung may evolve into a mosaic of clonal (and subclonal) outgrowths. Repeated carcinogenic insult might be expected to initiate new genetic changes that allow selective outgrowth of surviving cells during the wound-healing process. Moreover, some have recently postulated that altered epithelial cell-extracellular matrix interactions associated with tissue damage could further enhance the genetic instability process (32), perhaps helping to explain the continued risk of lung cancer for many years beyond smoking cessation (33). Like many exciting observations in research, when one reexamines the older literature, one wonders why this observation should be considered a surprise. The pioneering work of Mintz and Illmensee (34) using chimeric mice demonstrated that most tissues originate from a limited number of stem cells that are responsible for building distinct tissue regions (35). With the use of molecular tools to distinguish allele-specific X-chromosome inactivation, for example, it has been estimated that the human urothelium is derived from just 200-300 founder cells (36). Similarly, Smith and co-workers (37,38) provided evidence that each breast duct represents a unique clonal outgrowth and that, during the breast tumorigenesis process, subclones may arise within already clonal fields. Evidence for multiclonal and subclonal outgrowths in the fields of human tumors has also been well documented in tissues particularly amenable for spatial analyses [ e.g., the esophagus (39); bladder (40); and prostate, cervix, and skin (41)]. Similarly, Brash (42), using p53 immunostaining as a marker of clonality (in this case, altered p53 [ also known as TP53]-staining patterns were tightly associated with p53 gene alterations), observed small clonal outgrowths in the skin of mice after chronic UV light exposure. In the latter case, Brash suggested that the p53 alteration itself provided a survival and growth advantage for those damaged cells during the chronic exposure and wound-healing process. Of interest, after cessation of UV light exposure, some of these clones regressed and some remained and continued to evolve into invasive carcinoma. The findings described by Park et al. have important implications for early detection, risk assessment, and chemoprevention studies. One of the dilemmas facing chemoprevention studies for lung cancer is identifying those individuals at highest cancer risk who might best benefit from chemopreventive intervention. For example, individuals with a 20-pack-year smoking history (pack-years = number of packs smoked per day × number of years of smoking) are known to be at increased risk for lung cancer. In fact, nearly two thirds of such individuals who are current smokers will show metaplasia or dysplasia at one of six random biopsy sites during an endobronchial examination (43). However, since their lifetime risk of developing lung cancer is on the order of 10%-15% (44), lung chemoprevention trials with cancer as an end point requires large numbers of subjects and long follow-up times, sometimes with untoward results (45). As a result, there is a need for biomarkers that can provide early cancer detection, identify individuals at increased cancer risk, and provide surrogate markers of response to intervention (46). The ability to sensitively detect clonal outgrowths in the lung tissue at risk might be useful in these settings. Lung epithelium in the field of a lung tumor might represent a field at 100% risk for developing cancer. Thus, markers such as the presence of clonal outgrowths in the adjacent epithelium might reasonably be considered as candidates for risk assessment. After all, individuals with a primary lung cancer are known to be at high risk for second primary tumor development (47). Moreover, the presence of histologically abnormal epithelium in the lung is thought to portend higher lung cancer risk, and the findings of Gazdar's laboratory and others have shown that increased histologic change is associated with increased frequency and degree of molecular change. A similar argument might be made for using lung clonal assessments for predicting risk in individuals with a history of carcinogenic exposure. In the past, the general thinking by investigators in the field has been that molecular evaluation of lung biopsy specimens might provide information regarding that lesion's likelihood of evolving into invasive carcinoma. That is, if tumor development requires a certain number of relevant genetic alterations, then the determination of the accumulated number of genetic alterations in a particular biopsy specimen might provide an indication of how far that lung region has evolved in the tumorigenesis process, and this might provide a lung cancer risk estimate for an individual. Similarly, if one can revisit that lung site during a chemopreventive intervention, then one might be able to directly monitor efficacy by determining whether tumor evolution is slowed or even reversed at that site. One problem with such an approach is that it is not known ahead of time where a lung tumor is most likely to occur; thus, the biopsy process must be considered somewhat random. Some groups have moved toward fluorescence endobronchoscopy to guide them toward more advanced lesions; however, this remains an area of continued controversy. Moreover, many lung adenocarcinomas arise in the lung periphery and are difficult to assess with endobronchoscopy. Even at that, the results of this study present an additional wrinkle in this approach. If the detected clones are as tiny as those reported by Park et al., the first biopsy itself might remove the indicator clone and provide a “local cure.” An alternative approach to consider in this setting is to view risk assessment and response to intervention in a more global manner (30). That is, if the risk of lung cancer development is related to the time and degree of exposure and one's inherent susceptibility to carcinogenic insult, then a downstream measure of total accumulated damage in the target tissue might provide useful lung cancer risk information for an individual. In the setting of Park et al., it might be possible that the measurement of the degree of subclonal outgrowth in the lung tissue might provide such a marker. If the tumorigenesis process is more global in the lung field, then random biopsies and subclonal outgrowth assessments might provide surrogate risk information for the whole tissue. Preliminary observations from our own laboratory support the notion that the rate of genetic change at different assessable sites within the lung is somewhat uniform in individuals at risk for lung cancer. Similarly, during a chemopreventive intervention, one might use subclonal outgrowth analyses (in combination with other phenotypic biomarkers that address the mechanism of chemopreventive response in the tissue) as an indicator of chemopreventive impact on the tumorigenesis process in the field at risk rather than on one particular site. The success of such an approach assumes many factors that have yet to be addressed. For example, if the various forms of lung cancer have different patterns and frequencies of clonal outgrowth, then one measurement might have different implications for the risk of different types of lung cancer. Nevertheless, the results of the report by Park et al. provide important insights into the nature of the lung cancerization field and provide important tools with which to carry out lung cancer risk assessment and chemoprevention studies. Supported in part by Public Health Service grant CA48369 and CA70907 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services.