We have performed CDKN2A mutation analysis on subjects from melanoma families from the NCI Family Cancer Registry studied by Peggy Tucker and Alisa Goldstein of the Genetic Epidemiology Branch (GEB) for many years. We recently identified a novel mutation within the amino acid coding portion of CDKN2A (D153cspm) that alters splicing in the same manner as an earlier mutation we identified (IVS2+1g>t). Within the NCI families, two prominent pedigrees, with many cases of melanoma and some of the highest lod scores for linkage to the p16 locus on chromosome 9, do not have mutations within the mRNA encoding portion of CDKN2A, nor in the more than two kilobases of upstream potential regulatory region sequences based on analysis in multiple laboratories, including ours. While it is possible that these two linked but mutation negative families are due to another gene in this region, a more likely explanation is that they contain a mutation that is not detectable with the PCR-based screening techniques we are using. We have used a number of different primer pairs and sequenced both genomic DNA and cDNA, so it is unlikely that there are undetected mutations in the mRNA encoding portions of the gene. These linked but mutation negative families are extremely unlikely to have larger genomic deletions in the region because they carry the CDKN2A amino acid coding polymorphism A148T that segregates with disease in the families, and the affected individuals are heterozygous for microsatellite markers flanking the locus. They may have an intronic mutation that affects splicing but is not included when sequencing the exons, similar to the deep intron splicing mutation recently identified in a number of melanoma families, particularly from Great Britain. We have analyzed our samples directly for this intronic mutation, and have performed RT-PCR amplifying the entire coding portion of the gene, and do not observe any variant products that would represent splicing differences. There are no cells available from members of these families that express p16 (measurable by Western blot), such as fibroblasts or melanoma cell lines, and Westerns performed on lymphocyte lysates have shown no detectable p16 protein. We are continuing to study these two families and to collaborate with GEB investigators conducting melanoma case-control studies. Capitalizing on the presence of the A148T missense mutation and a 3'UTR SNP within CDKN2A, we have sequenced RT-PCR products from subjects from these two mutation negative families to evaluate whether both alleles are equally expressed. We have found that only one allele is being expressed in many subjects. We are still blinded as to which subjects are affected, but according to Dr Goldstein, the lack of expression of one allele largely segregates with melanoma in these families. It is possible this represents """"""""normal"""""""" background levels of differential allele expression, but more likely is the result of a regulatory or splicing mutation that prevents expression from one of the two chromosomes and that indeed these two families have some defect in CDKN2A. We have recently begun dHPLC-based screening of both introns of this gene to search for mutations that might affect splicing or regulation of this gene. We have also developed Taqman real-time assays to measure allele-specific expression to test whether differential allele expression of CDKN2A is a more general phenomenon, not restricted to these two families.
| Goldstein, A M; Struewing, J P; Fraser, M C et al. (2004) Prospective risk of cancer in CDKN2A germline mutation carriers. J Med Genet 41:421-4 |
| Baccarelli, A; Calista, D; Minghetti, P et al. (2004) XPD gene polymorphism and host characteristics in the association with cutaneous malignant melanoma risk. Br J Cancer 90:497-502 |
| Rutter, Joni L; Bromley, Christina M; Goldstein, Alisa M et al. (2004) Heterogeneity of risk for melanoma and pancreatic and digestive malignancies: a melanoma case-control study. Cancer 101:2809-16 |
