Human Breast Adenocarcinoma: Dna Content, Chromosomes, Gene Expression and Prognosis

s 173 CHROMOSOMES, PLOIDY AND GENETIC IMBALANCES OF LUNG CANCER Iver Petersen Institute of Pathology, Charité Medical School, Humboldt University, Berlin, Germany Lung cancer is a heterogeneous and highly aggressive disease which is reflected by a wealth of genetic alterations on the DNA and RNA level. In an attempt to understand this apparent chaos we used screening methods like Comparative Genomic Hybridization (CGH), Suppression Subtractive Hybridization (SSH) and cDNA microarrays. In addition, cytogenetic information on lung tumors were retrieved from the Mitelman Database of Chromosome Aberration and analyzed for chromosome numbers and alterations. From the Mitelman database, in total 660 lung tumors were identified which 446 were histologically typed. The analysis was then mainly restricted to 160 adenocarcinomas (ADC), 145 squamous cell carcinomas (SCC), 14 ADC-SCC, 38 large cell lung carcinomas (LCLC) and 49 small cell lung carcinomas (SCLC). All showed at least subtle chromosomal abnormalities indicative of aneuploidy. About 30% and 22% of the near diploid ADC and SCC, respectively, carried only single chromosome change, in particular loss of chromosome Y and gain of chromosome 7, in contrast to only 8% of LCLC being generally highly aneuploid and carrying the highest chromosome numbers of all lung cancer subtypes. Except for 1 case (2%), all SCLC were highly aneuploid although 27% carried a near diploid chromosome number. DNA measurement of primary SCLC may indicate an even higher percentage of pseudodiploid cases in up to 90% of cases. Except for the near diploid cases, SCC were almost invariably hyperdiploid. In contrast, hypodiploid tumors were present in the ADC and SCLC subgroups, both being associated with a high degree of aneuploidy. Beside the near diploid cases, the histogram of the lung carcinomas according to their chromosome numbers showed a second peak in the near triploid range. CGH revealed typical patterns of chromosomal imbalances in each lung cancer subtype and also specific alterations that were significantly associated with tumor progression and differentiation [1–6]. Amazingly, there were even chromosomal imbalances detectable that correlated with organ specific metastasis to the brain [5]. The highest prevalence of alterations were observed in SCLC. The data confirms that aneuploidy is a key factor in lung carcinogenesis being early detectable and also associated with tumor progression. Different chromosome numbers and imbalances are associated with lung cancer subtypes, their variation by chromosomal instability may cause transition in tumor morphology and differentiation. The expression analysis is able to translate the genetic imbalances into disregulations of specific genes thus carrying the potential to identify candidates for diagnostic and therapeutic purposes [7–10]. Our cDNA microarray study [8] using a 24,000-element chip representing more than 17,000 unique genes on 67 lung cancer specimens including five SCLC from 56 patients showed that the major subtypes, i.e. squamous, adeno-, large cell and small cell carcinomas clustered into individual subgroups apart from normal lung. For the clustering a subset of 918 cDNA clones was chosen that discriminated best between the tumors form different patients (compared to tumor pairs from one individual). Doing so, the above mentioned tumor subgroups were associated with the upor downregulation of a cluster of genes being most characteristic each tumor type. Squamous cell carcinoma (SCC) of the lung showed characteristics of a “true” squamous epithelium with expression of genes like p63, and cytokeratins 5, 13, and 17. Large cell carcinomas showed expression of genes involved in tissue remodeling. Adenocarcinomas, the largest subcollective, separated into three subgroups that were significantly different in survival. In group 3 adenocarcinomas with bad survival, genes involved in lung differentiation like TTF1 were downregulated. Together with large cell carcinomas the gene expression pattern suggested an epithelial– mesenchymal transition. In summary, CGH and expression profiling are powerful tools for lung cancer characterisation and the identification of new diagnostic and therapeutic candidate genes. Microarrays may be used to supplement the conventional classification, however, the analysis of specific genes by RT-PCR or immunohistochemistry may prove to be an easier and cheaper alternative in this respect. Finally, our studies inspired two models for lung cancer progression, one being associated with small cell (neuroendocrine) dedifferentiation [11], the other with large cell (mesenchymal) dedifferentiation.