Epigenetics in Lung Cancer*
Genetic alterations are a hallmark of human cancer. Changes in DNA methylation, an epigenetic modification that is present in mammalian cells, are also characteristic of human cancer. The CpG dinucleotide, which is usually underrepresented in the genome, is clustered in the promoter regions of some genes. These promoter regions have been termed CpG islands. CpG islands are protected from methylation in normal cells, with the exception of genes on the inactive X chromosome and imprinted genes.1 This protection is critical, since the methylation of promoter region CpG islands is associated with a loss of expression of these genes. The following three different alterations in DNA methylation are common in human cancer: (1) global hypomethylation, often seen within the body of genes; (2) dysregulation of DNA methyltransferase I, the enzyme involved in maintaining methylation patterns, and potentially other methyltransferases; and (3) regional hypermethylation in normally ummethylated CpG islands.
At the outset, it is important that the term epigenetics be defined. Although this term had been coined before, Holliday2 provided definition for this concept in reviews 15 years or so ago. As used today, this term refers to a change in gene expression that is heritable (ie, that is can be passed on through cell division) but that does not involve a change in DNA sequence. This contrasts with a true genetic alteration. In this context, it is helpful to look at the entire repertoire of changes that accomplish the gains or losses of gene function that result in the transformed phenotype. These changes are outlined in Figure 1 and span from those that are completely permanent to those that are very dynamic. Epigenetic changes are semi-permanent.
Work performed over the past several years has demonstrated that promoter hypermethylation commonly silences tumor suppressor genes in human cancer, serving as an alternative mechanism for the loss of tumor suppressor gene function. While this growth in interest does not displace or diminish the importance of the genetic changes in cancer, this overview will only discuss the role of epigenetic changes in cancer, specifically promoter region DNA methylation. However, it is worth noting some potential reasons for this increased interest in DNA methylation changes in cancer, and in lung cancer in particular. These reasons include the awareness of a bone fide mechanism of gene inactivation, and the fact that the detection of this change is now technically easier to characterize than mutations.3 These are, however, not biological reasons for the increasing number of publications on DNA methylation in cancer. In addition, epigenetic changes are a very powerful modulator of cellular phenotype, and are in fact a normal process gone bad and not an abnormal process, such as gene deletion, rearrangement (except for Ig gene rearrangement, which is not normal), or point mutation. As an example of this normal modulator of the cell phenotype, the wide variety of histologic differentiation observed during embryogenesis is mediated through epigenetic mechanisms, not genetic mechanisms.
The relative cancer specificity of changes in CpG island methylation makes methyltransferase I an attractive target for cancer therapeutics. In addition, an examination of the specific changes in functionally important genes can provide a molecular understanding of the biology of an individual tumor. For example, the most studied of all tumor suppressor genes for promoter hypermethylation is the p16 gene, currently designated CDKN2A, which is a cyclin-dependent kinase inhibitor that functions in the regulation of the phosphorylation status of the Rb protein. Frequent homozygous deletion and some point mutations have suggested an important role for this gene in neoplasia. However, numerous follow-up studies have reported very low rates of inactivation of this gene in tumors of many histologic types. In fact, hypermethylation associated with th loss of expression of the CDKN2A gene has been found to be one of the most frequent alterations in neoplasia. Initial reports456 have described methylation of the p16 gene in cancers of the lung, head, and neck, in gliomas, and in colorectal and breast carcinomas. Many additional studies78 have documented the importance of this change in the pathogenesis of non-small cell lung cancer, for which this gene may be the most extensively studied, and have suggested that this is an early event leading to the development of lung cancer.
Many other genes have been described to be methylated in non-small cell lung cancer.9 The most widely studied and prominent genes in lung cancer include RARbeta,10 RASSF1A,1112 methyl-guanine-methyltransferase,13 and DAP-kinase.9141516 The function of each gene is different, with roles in receptor-mediated signaling, ras signaling, DNA repair, and apoptosis. The growing number of genes (incompletely described here) has necessitated looking at multigene profiles of cancers, and lung cancer is no exception. Indeed, by looking at multiple genes in non-small cell lung cancer, it can be seen that the majority of cancers have one, and in most cases more than one, aberrantly methylated CpG island that silences critical genes,9 and we have recently found similar changes in 42 adenocarcinomas. In both cases, the methylation of each gene appeared to segregate independently of changes at other loci and suggested that increasing the number of genes examined will further decrease the number of tumors without an identified aberrantly methylated CpG island. It is likely that all tumors harbor both genetic and epigenetic changes that compliment to form the transformed phenotype.
The profile of DNA methylation changes differs according to the cell of origin. While this can be seen by looking at profiles of lung, colon, or breast cancer examined in separate studies,917 it can be best appreciated by examining the methylation changes in multiple loci in different cancer types or cancer cell lines. In a large panel of cancer cell lines,18 this approach of looking at the methylation changes in 15 genes was able to cluster cell lines according to tumor of origin. This probably should not come as a surprise. However, it is not known whether these differences resulted from baseline tissue-specific changes in gene expression or chromatin structure that make some genes more susceptible to methylation, or whether the selective importance of these genes is different in different tumor types.
However, the number of changes observed in these tumors offers another advantage in studying the biology of neoplasia. Carcinogenesis is a multistep process that has been described in terms of a tumor progression model, most clearly in the genetic changes in colon cancer. We have recently begun to examine the changes in promoter region methylation at mulitiple loci of invasive tumors, normal tissues, and preinvasive lesions from these same tissues. The timing of these events probably varies much as has been proposed for the genetic models of tumor progression. In addition, the number of changes present in a lesion increases as the histologic changes worsen. We suggest that these observations represent the epigenetic progression that parallels the histologic progression observed in multiple organ sites. This is shown in Figure 2 .
Because of the frequency and timing of these epigenetic changes, promoter hypermethylation events provide one of the most promising markers for molecular detection. DNA-based markers have advantages because of the inherent stability of DNA compared with RNA and some proteins. The constant position of the abnormal CpG methylation in genes of interest allows a simpler detection strategy than is possible for many of the common mutations in cancer. Such mutations, even for the same tumor type, may differ widely in position within the gene from patient to patient. In contrast, for any given gene, a single assay for the detection of promoter methylation abnormalities can work for all patients. Methylation has been detected14 in the serum of patients with lung cancer, demonstrating the use of molecular detection. Some studies1920 have suggested that gene-specific methylation changes in sputum could be good intermediate markers for the early detection of lung cancer and for defining the efficacy of chemopreventive interventions. These approaches have utilized this multigene approach, which continues to examine increasing numbers of genes as our knowledge of the epigenetic changes in lung cancer grows.21 Further work will define the importance and timing of these changes in the neoplastic transformation occurring in the lung tissue, and the use of such changes in the early detection of lung cancer and in monitoring changes associated with therapy.