Most Genetic Epidemiology Branch investigations evaluate the contributions of host susceptibility and environmental exposure in the development of cancer. In family studies, the host susceptibility measure is frequently an alteration in specific gene(s). These studies tend to be very long term with varying activity. Although two genes associated with melanoma susceptibility have been identified (CDKN2A and CDK4), alterations in these genes are found in only a small percentage of melanoma-prone families. The search for other genes continues; in collaboration with an international consortium (GenoMEL), a search for a new melanoma susceptibility gene on 1p22 continues. In collaboration with GenoMEL, we evaluated four features previously reported in individual groups to be associated with likelihood of the family having a CDKN2A mutation in 385 families from 17 groups in 3 continents. One hundred fifty of the families had CDKN2A mutations; 235 did not. The features evaluated were number of individuals with melanoma in the family; presence of individuals with multiple primary melanomas within the family; early age of onset of melanoma; and presence of pancreatic cancer in the family. Overall, each individual feature was highly significantly associated with CDKN2A status of the families.There were, however, clear differences in geographic areas when the features were simultaneously evaluated. In Australia, there was no significant association with pancreatic cancer. In addition, to have an increased risk of CDKN2A mutation, the age of onset of melanoma in the family was less than 40, at least two individuals had to have multiple primary melanomas, and more that 6 individuals in the family were affected. In North America, Age of onset less than 40 and one or more individuals with multiple primary melanoma were significantly associated with CDKN2A status. In Europe, pancreatic cancer, one or more individuals with multiple primary melanomas, age under 50 and 4 or more family members were associated with CDKN2A status. GenoMEL is collecting data to evaluate these differences. Among 466 families, of whom 178 had mutations in CDKN2A affecting p16, 7 had mutations affecting ARF, and 5 had mutations in CDK4, the relationship between other cancers and mutations were evaluated. Across the geographic locales, different founder mutations were frequent. The mutations affecting CDK4 and ARF were similar, between 2-3%. The relationship between pancreatic cancer and CDKN2A mutation was strong, but varied by specific mutation. There was no apparent association with either neural tumors or uveal melanoma and CDKN2A mutations. A marginally significant association of ARF and neural tumors, based on small numbers, was seen. In the American families, vairants in MC1R were related to melanoma risk in families with CDKN2A mutations. Multiple variant alleles of MC1R were associated with multiple primary melanomas. We continue to accrue and evaluate new families. Genetic analyses of Italian melanoma families from the Emilio Romagna area have shown that mutations in CDKN2A are quite rare, but variants of MC1R are important risk factors in individuals without the fair hair/skin phenotype. A whole genome scan has been completed; linkage analyses are almost completed. We have continued to evaluate families of individuals with heritable retinoblastoma and melanoma. The study of familial chordoma, a rare, low-grade, malignant bone tumor derived from remnants of the notochord, was expanded to include additional families. Additional efforts to identify more families have yielded another family. Sequencing of candidate genes in the identified linkage interval continues. Studying families with lymphoproliferative cancers has been a long-standing interest. We have conducted a linkage analysis on suitable CLL kindreds which identified several areas of interest for further study.
Showing the most recent 10 out of 29 publications
