Approaches for the localization and identification of human cancer genes

Abstract

Chromosomal aberrations have been recognized as important observations that underpin the concept of a mutator phenotype in cancer. In this thesis, two basic strategies were applied to map the location of cancer-related genes in several types of cancers. The first strategy was to investigate the global chromosomal alterations in cancer cells using comparative genomic hybridization (CGH) and single nucleotide polymorphism arrays. The second strategy was to search for specific chromosomal rearrangements in the tumor cells, as well as constitutional cells in familial cancer using karyotyping analysis. To further approach the molecular mechanisms of the disease, mutation screening and expression profilings were performed.

Using the first strategy, we found significant concordance of deletions in chromosomes 3 and I I in 34 tumors (94%) and 31 tumors (86%), respectively, in von Hippel-Lindau (VHL) associated pheochromocytomas, suggesting that they are involved in two different but necessary and complementary genetic pathways. The loss of chromosome I I appeared to be specific for VHL-related pheochromocytoma as it was not present in any of the 10 VHL-related CNS hemangioblastomas studied and was significantly less common when compared with sporadic and MEN2-related pheochromocytomas.

In the series of 23 cases of SCLC analyzed, CGH alterations were frequently detected ranging from 9 to 22 changes in the individual tumors, suggesting the malignant phenotype of this tumor type. The identification of subchromosomal losses and gains, as well as the high-level amplifications of 1p32-33 and 2p22-24, provides starting points for the exact characterization of molecular events involved in SCLC tumor progression. Here, we also evaluated the SNP array hybridization for the detection of genome-wide LOH in 17 matched pairs of SCLC and normal control DNA samples. We showed that SNP hybridization assay has a higher resolution than CGH for detection of losses, but that gains seen by CGH are not readily detectable by the SNP array approach.

Using extensive genetic approaches, we characterized a novel translocation t(1;3)(q32;q13.3) associated with familial renal cell carcinoma. By FISH mapping, the breakpoints were refined to a 5 cM region in 3q13.3 and a 3.6 cM region in 1q32. We also proposed a three-step genetic model for the tumorigenesis in this family.

Using karyotyping approaches, we identified a novel balanced translocation t(3;7)(p25;q34) in a follicular thyroid carcinoma (FTC), which was then led to the identification of a fusion oncogene. This fusion involves a novel gene, which is named FTCF, located at 7q34 that is fused to the 5 region of the PPAR gamma 1 gene which resides at 3p25. FTCF was shown to be ubiquitously expressed in normal tissues and the corresponding protein was localized to the nucleus.

In addition, we determined the expression profiles of FTCs using expression array. A total of 76 genes were consistently differential expressed in FTCs as compared to normal thyroid, and were mainly related to functions of cell adhesion, immune response, cell cycle, signal transduction, and fatty acid metabolism. We also compared the expression patterns of FTCs with and without PAX8-PPAR gamma 1 fusion oncogene, and the clustering analysis separated the two tumor subsets well. The FTCs with PAX8-PPAR gamma 1 fusion displayed a subset of deregulated genes related to cell cycle, translational control and FYN-mediated signaling cascade(s). Taken together, the results suggest that the genetic subsets of FTCs have distinct expression profiles, thus implying the involvement of different genetic pathways in follicular thyroid tumorigenesis.