Functional analysis of CDKN2A/p16INK4a 5'-UTR variants predisposing to melanoma

Germline CDKN2A mutations are observed in 20-50% of melanoma-prone families. We identified melanoma patients that were heterozygous for non-coding germline variants in the 5'-UTR of CDKN2A (c.-21C > T; c.-25C > T&c.-180G > A; c.-56G > T; c.-67G > C) and examined their impact on the p16(INK4a) 5'-UTR activity using two luciferase-based reporter vectors that differ in basal transcription level and that were transfected into the melanoma-derived WM266-4 and in the breast cancer-derived MCF7 cells. The wild-type 5'-UTR sequence, containing a reported SNP (c.-33G > C) and a known melanoma-predisposing mutation (c.-34G > T), was included as controls. Results revealed that the variants at -21 and -34 severely reduced the reporter activity. The variants at -56 and at -25&-180 exhibited a milder impact, while results with c.-67G > C were dependent on the plasmid type. Quantification of the luciferase mRNA indicated that the effects of the variants were mainly post-transcriptional. Using a bicistronic dual-luciferase reporter plasmid, we confirmed that c.-21C > T and c.-34G > T had a severe negative impact in both cell lines. We also applied a polysomal profiling technique to samples heterozygous for the 5'-UTR variants, including patient-derived lymphoblasts. Analysis of allelic imbalance indicated that in addition to the c.-21C > T variant, the c.-56T > G and c.-67G > C variants also reduced mRNA translation efficiency. Overall, our results suggest that the c.-21C > T sequence variant is a melanoma-predisposing mutation. The c.-25C > T&c.-180G > A and particularly the c.-56G > T variants showed a range of intermediate functional defects in the different assays, and were not observed in the control population. We propose that these variants should be considered as potential mutations.

Germline CDKN2A mutations are observed in 20-50% of melanoma-prone families. We identified melanoma patients that were heterozygous for non-coding germline variants in the 5′-UTR of CDKN2A (c.-21C > T; c.-25C > T&c.-180G > A; c.-56G > T; c.-67G > C) and examined their impact on the p16INK4a 5′-UTR activity using two luciferase-based reporter vectors that differ in basal transcription level and that were transfected into the melanoma-derived WM266-4 and in the breast cancer-derived MCF7 cells. The wild-type 5′-UTR sequence, containing a reported SNP (c.-33G > C) and a known melanoma-predisposing mutation (c.-34G > T), was included as controls. Results revealed that the variants at -21 and -34 severely reduced the reporter activity. The variants at -56 and at -25&-180 exhibited a milder impact, while results with c.-67G >C were dependent on the plasmid type. Quantification of the luciferase mRNA indicated that the effects of the variants were mainly post-transcriptional. Using a bicistronic dual-luciferase reporter plasmid, we confirmed that c.-21C > T and c.-34G > T had a severe negative impact in both cell lines. We also applied a polysomal profiling technique to samples heterozygous for the 5′-UTR variants, including patient-derived lymphoblasts. Analysis of allelic imbalance indicated that in addition to the c.-21C > T variant, the c.-56T > G and c.-67G > C variants also reduced mRNA translation efficiency. Overall, our results suggest that the c.-21C > T sequence variant is a melanoma-predisposing mutation. The c.-25C > T&c.-180G > A and particularly the c.-56G > T variants showed a range of intermediate functional defects in the different assays, and were not observed in the control population. We propose that these variants should be considered as potential mutations. © The Author 2010. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org.