Delection mapping of human 3p in major epithelial types of cancer and fine localization of candidate tumor suppressor genes

Abstract

Allele loss and deletion mapping using microsatellite markers and the detection of homozygous deletions represented until now the most powerful method to localize potential TSGs. Loss of heterozygosity (LOH) involving several chromosome 3p regions accompanied by chromosome 3p deletions are detected in almost 100% of renal cell carcinoma (RCC), small (SCLCs) and more than 90% of non-small (NSCLC) cell lung cancers. These 3p genetic alterations led to the conclusion that short arm of human chromosome 3 contains several tumor suppressor gene(s) (TSGs). A number of studies were done to perform fine mapping and localize TSG more precisely but they had low efficiency. Reports differ, for example, on the extent of 3p losses in different tumors, with some papers reporting large terminal deletions, and others claiming interstitial deletions. Reports for the frequency of LOH also differ for the same marker in the same type of tumor.

One of the reasons for these discrepancies is that admixtures of stroma, blood vessels, lymphocytes and other normal cells in the tumor samples were unavoidable sources of error in LOH studies of solid tumors. To test the potential impact of this problem, we developed novel approach that we call AlleleTitration-Assay (ATA). In ATA experiments, we prepared artificial mixtures of mouse-human microcell hybrid cell lines that carried different alleles of the same chromosome 3 marker. We have demonstrated that normal tissue admixtures will be less of a problem when LOH affects an H allele than with an L allele. The results suggested that about 50% of the L-allele deletions in tumor samples might go undetected. We suggested new L-allele rules for the evaluation of LOH experiments to avoid this bias. Comparative genome hybridization (CGH) method allows analyzing the whole chromosome but CGH is still not sensitive enough to detect deletion region smaller than 1-2 Mb and cannot detect allele losses. Using both CGH and LOH will lead to exploitation of their advantages and will limit their disadvantages. ATA rules combined with CGH and LOH analysis of RCC cell lines and biopsies confirmed the presence of interstitial deletions and opened the way for the mapping of candidate TSGs. We concluded that there are two main frequently affected regions in 3p that can harbor candidate tumor genes: 3p21.3 telomeric or 3p21.3T, including AP20 region and 3p21.3 centromeric or 3p21.3C, including LUCA region.

Since the resolution of CGH method is rather low, we decided to use a much more sensitive, rapid and quantitative method, termed quantitative real-time PCR to evaluate 3p genetic changes in carcinogenesis. This technology does not require polymorphic markers, and any marker is informative for any cancer case. Two nonpolymorphic STS markers (NLJ-003 and NL3-001) located in the AP20 and LUCA regions, respectively, were used for quantitative real-time PCR as TaqMan probes. LOH analysis was verified using L-allele rules, real-time quantitative PCR and Southern hybridization. Significant (85%) correlation was found between DNA copy numbers for NLJ-003 and NL3-001 markers and LOH data for adjacent polymorphic loci. The real-time PCR data were consistent with the Southern data too.

The results of the study allows to make at least two conclusions. First amplification of 3p is very common (15%-42.5%) in studied cancers and probably in other epithelial malignancies. Therefore, microsatellite deletion data should be evaluated carefully as allelic imbalances mean not only deletions but also amplifications. Second, the data showed that aberrations of either NLJ-003 or NL3-001 were detected in more than 90% of all studied cases. Homozygous deletions were detected in 10%-18% of all cases in NLJ3-001 or NL3-001 loci. The exceptionally high level of chromosome aberrations in NLJ003 and NL3-001 loci suggests that multiple TSG(s) involved in different malignancies are located very near to these markers.

Careful analysis of 15 homozygous deletions in NL3-001 allowed to establish that the smallest region homozygously deleted in 3p21.3C is located between D3S1568 (CACNA2D2 gene) and D3S4604 (SEMA3F gene) and contains 17 genes previously defined as lung cancer candidate TSG(s). Mapping of 19 homozygous deletions in AP20 region resulted in localization of the smallest region homozygously deleted in 3p21.3T. It was flanked by D3S1298 and D3S3623. Only 4 genes are located there, namely APRG1, ITGA9, HYA22 and VILL, which need to be analysed.