A common founder for the V126D CDKN2A mutation in seven North American melanoma-prone families

One of the most common melanoma-related CDKN2A mutations reported in North America is the V126D mutation. We examined nine markers surrounding CDKN2A in three American and four Canadian families carrying the V126D mutation. All seven families had a haplotype consistent with a common ancestor/founder for this mutation. In addition, the mutation appears to have originated 34–52 generations ago (1-LOD-unit support interval 13–98 generations). © 2001 Cancer Research Campaign http:///www.bjcancer.com

inhibiting cyclin D1-CDK4 in a reconstituted pRb-kinase assay, and for increasing the proportion of G1-phase cells following transfection (Parry and Peters, 1996). Previous examination of three American melanoma-prone families with this mutation suggested the possibility of a common haplotype . We have now examined additional markers in these three families, including marker D9S974, a marker extremely close to the CDKN2A gene (Randerson-Moor et al, 2001). In addition, we have examined nine markers surrounding the CDKN2A gene in four Canadian families carrying the V126D mutation. The results show that all seven families have a haplotype consistent with a common founder for this mutation. In addition, although based on only seven families, the mutation appears to have originated approximately 34-52 generations ago (1-LOD-unit support interval 13-98 generations).

Families
Details about the seven North American families have been presented previously (Hussussian et al, 1994;Goldstein et al, 2000;Liu et al, 1999). For all participants, written informed consent was obtained prior to participation under Institution Review Board approved protocols. Briefly, the four Canadian families had an average of three melanoma patients per family (Table 1). In addition, patients from two of the families had multiple primary melanomas. The three American families all had patients with multiple primary melanomas; two of the families had at least one patient with pancreatic cancer. Although six of the seven families had German/English ancestries, little is known about the ancestral pathway for melanoma in these families prior to their arrival in North America.

Genotyping
Nine markers were genotyped for the analysis to determine which alleles from loci flanking CDKN2A were transmitted with CMM in each of the families: IFNA, D9S736, D9S1749, D9S974, D9S942, D9S1748, D9S1604, D9S171, and D9S126. Markers D9S1749 (approximately 0.0105 M or 1.05 cM distal) and D9S974 (approximately 0.012 cM proximal) were the closest flanking markers to exon 2 of the CDKN2A gene (Randerson-Moor et al, 2001). Allele sizes for all markers are comparable with those from haplotype studies of Ciotti et al (2000) and Pollock et al (except for D9S736 and D9S126) (1998).

Dating the mutation
To estimate when the V126D mutation originated, we used a maximum likelihood (MLE) method developed by D. Goldgar (Neuhausen et al, 1996(Neuhausen et al, , 1998 and previously applied to G101W (Ciotti et al, 2000). Briefly, the joint likelihood of the V126D haplotypes was written as a function of the recombination fraction between the disease and each marker, the number of generations (G) since the mutation arose, and the mutation rate (0.0006 for all markers except D9S1749 [0.01] and D9S942 [0.002]) and allele frequencies at each marker locus (Ciotti et al, 2000) ( Table 2). The MLE method was used to find the value of G that best fitted the pattern of haplotype sharing at the nine marker loci. When haplotypes could not be determined with certainty, all possible haplotypes consistent with the observed multilocus genotypes were considered in the analysis. Approximate support intervals were calculated by finding the value of G on either side of the most likely value that had a ≥ 10-fold decrease in likelihood.

RESULTS
Haplotype analysis using 9 polymorphic markers spanning the CDKN2A locus was performed on index cases and additional family members (when available) to determine whether carriers from the seven families harbored the same mutation identically by descent. Table 3 shows the disease haplotypes or genotypes for the seven families. Both alleles are indicated for markers for which segregating alleles could not be unambiguously determined. All seven families showed a haplotype or genotype consistent with a single genetic origin for the V126D mutation. The D9S1749-D9S1604 haplotype 16/17-6-11-9-2 appears to be common across all families, after allowing for recombination over time. One American family (K) had the 5 allele at D9S974, rather than the 6 allele seen in all other families. D9S1749, previously shown to vary in allele size because of replication slippage resulting in the loss or gain of one or more repeat units during meiosis (Pollock et al, 1998), showed either the 16 or 17 allele co-segregating in all but one family. Family 103 showed allele 14 co-segregating with CMM. Allowing for replication slippage in D9S1749, families J and L shared a common haplotype from IFNA to D9S126. It was not possible to further assess the extended sharing of the disease related haplotype in other families because the co-segregating alleles could not be unequivocally determined.
Results from the MLE method suggested that the V126D mutation originated approximately 34-52 generations ago (1-LOD-unit support interval 13-98 generations) or approximately 680-1040 years ago (1-LOD-unit support interval 260-1960 years) using a 20-year generation interval or 1020 to 1560 years ago (1-LOD-unit support interval 390-2940 years) using a 30-year generation interval. The maximum likelihood estimates for alleles 16 and 17 of D9S1749 were equivalent; thus a range in the estimate of the mutation origin is presented (e.g. 34-52 generations).

DISCUSSION
The V126D mutation appears to have originated from a common founder or ancestor, as is the case with most recurrent CDKN2A mutations studied to date. Only 23ins24, a 24 base-pair duplication, has been shown to have multiple origins (Pollock et al, 1998), probably due to the inherent instability of the wild type CDKN2A 5' tandem repeat region. Although there are many recurrent CDKN2A mutations, only two -G101W and 113insArg -have been evaluated to determine their ages of origination. Using the same MLE methods as was used in the current study, Ciotti et al (2000) and Hashemi et al (2001) concluded that the G101W and 113insArg mutations both originated approximately 100 generations ago. Using the same maximum likelihood estimate method, we estimated that the V126D mutation originated approximately 34-52 generations ago (1-LOD-unit support interval 13-98 generations). Given the relatively small number of families in the present study and the sensitivity of the MLE method to the marker mutation rates, we also employed an approach proposed by Neuhausen et al (1996) to evaluate the variability in the estimated age of the mutation. The age of origin for the mutation was reestimated assuming marker mutation rates that were an order of magnitude (i.e. 10x) lower and higher (Neuhausen et al, 1996) than the values used in the original analysis. The estimated age of the V126D mutation was reduced to 9 generations when a 10-fold increase in marker mutation rates was assumed and 71 generations   when a 10-fold decrease in rates was applied (combined 1-LODunit support interval 3-145 generations). These additional findings in conjunction with the original results suggest the possibility of a more recent origin for V126D relative to that seen for the G101W and 113insArg mutations. We cannot, however, preclude a more remote origin based on the greater imprecision in the estimated age resulting from the smaller numbers of families available for the current study.
American family K had the 5 allele at D9S974, rather than the 6 allele seen in all the other families. Since D9S974 is the marker closest to the V126D mutation (0.012 cM), a recombination event so close to the mutation would likely indicate a remote origin for the mutation. Replication slippage, conversely, does not necessarily imply an ancient origin for the mutation. Unfortunately it is not possible to determine whether the 5 allele, only 2 base pairs smaller than the consensus 6 allele, resulted from recombination or Common founder for V126D mutation 529 British Journal of Cancer (2001) 85(4), 527-530 © 2001 Cancer Research Campaign  *Both alleles are indicated for markers for which segregating alleles could not be unequivocally determined. **Alleles that are part of the common disease-related haplotype are shown in boldface. ***Line between D9S1749 and D9S974 represents location of V126D mutation.
replication slippage. Thus, this allele change provides little additional evidence for helping determine the origin of the V126D mutation.
Most recurrent CDKN2A mutations observed in North America can be traced back to a European country or region of origin. For example, the recurrent M53I mutation, which has been found at high frequency in North America, Great Britain, and Australia, appears to have originated in Great Britain (Pollock et al, 1998;Liu et al, 1999). Similarly, the G101W mutation, which is very common in the United States, appears to have originated in southwestern Europe; it is the most common mutation detected in France and Italy (Ciotti et al, 2000;Soufir et al, 1998;Ghiorzo et al, 1999;Ruiz et al, 1999). In contrast, the V126D mutation does not appear at high frequency in any other countries besides the United States and Canada. The mutation has been observed in Australia, France and Italy but only rarely in each of these countries. This phenomenon may reflect selective mutation testing in the various countries or it may be related to the origination of this particular mutation. For example, the major CDKN2A mutation testing from Great Britain to date has occurred in Northcentral England and Scotland where the V126D mutation has not been observed (MacKie et al, 1998;Newton Bishop et al, 1999). Also, very little data from Germany on CDKN2A mutation testing has been published. Testing in other areas of Great Britain, Germany or other parts of continental Europe, however, might reveal the mutation. Although six of the seven families immigrated to North America from Germany and England, the ancestral pathway for melanoma in these families cannot be determined. Additional families from North America as well as from other geographic areas may help determine the geographic origin for this recurrent yet puzzling CDKN2A mutation.