Founder Effect of the RET C611Y Mutation in Multiple Endocrine Neoplasia 2A in Denmark: A Nationwide Study

Patients

This study included 21 apparently unrelated MEN2A families identified from a nationwide Danish RET cohort from 1994 to 2014 (Table 1) (4).

The Danish RET cohort initially contained 36 MEN2 families. One family has been excluded, as the I852M sequence change is no longer considered pathogenic (5,6). For the purpose of this study, all MEN2A families carrying unique mutations within the cohort (n = 6) and all MEN2B families (n = 6), three of which have been described elsewhere (7 –10), were excluded. This resulted in 23 MEN2A families. In two C634R families (family nos. 23 and 25), DNA was not available. Thus, a total of 21 families were eligible for inclusion. Twelve, two, two, three, and two families carried the C611Y, C618F, C618Y, C620R, and C634R mutations, respectively (Table 1). Several of these have been reported previously (11 –15). Based on the availability of DNA, one individual from each family was chosen for analysis.

The investigation was approved by the Capital Region Committee on Health Research Ethics (H-4-2010-050) and the Danish Data Protection Agency (2015-41-4464).

RET testing

The primary testing method was Sanger sequencing, which included exons 8–11 and 13–16. This has recently been described in detail (4). In the literature, C611Y mutations due to both the TGC-TAT and the TGC-TAC codon change have been described (5). Herein, only the latter is reported and discussed.

Single nucleotide polymorphism chip data

CytoScan assay (Affymetrix, Santa Clara, CA) was performed on DNA from whole blood according to the manufacturer's instructions. Probe-level analysis was performed on the resulting raw data files, using Chromosome Analysis Suite software by Affymetrix (ChAS 3.2), and the genotype calls were exported. To obtain estimates of population allele frequencies, the single nucleotide polymorphism (SNP) chip data were merged with publicly available genotypes from the Central European population from the HapMap 3 data (16). SNPs with a minor allele frequency <5% and with more than 2% missing data were removed. After merging, 385,743 SNPs on the autosomes remained. High-linkage disequilibrium (r ² > 0.8) was removed using PLINK (17).

Identity by descent analysis

To determine if individuals were related between families, the amount of alleles sharing identity by descent (IBD) was estimated globally using the SNP chip data. The method of moment estimator from PLINK was used (17). To determine if the RET mutations had a recent common ancestor, the IBD sharing across the genome was estimated using Relate, which gives a probability of IBD sharing at each position in the genome (18). The probability of sharing at least one allele IBD at the RET locus was reported. If the RET mutations had a recent common ancestor, the region around the mutations was expected to be IBD between families, even if the families were distantly related.

Limitations

This study shares limitations that are inherent to molecular studies of rare diseases. To increase the sample size, all relevant MEN2A families were included on a nationwide level. This included 12 families with 108 RET-proven C611Y carriers, being the largest C611Y cohort reported to date.

Two relevant C634R families were excluded from analysis, as DNA (RET tests performed in 1995) was unavailable. This accounted for 50% (2/4) of the C634R families, but only 9% (21/23) of the total number of families relevant for the study. Importantly, DNA was available from all the C611Y families.

The local IBD sharing analysis can detect haplotypes with a common ancestor within the last 10–20 generations (18). Thus, relatedness among family nos. 14, 15, 18, 19, 20, 22, and 24 cannot be ruled out completely. However, this does not compromise the finding of a common haplotype between all C611Y families and another between both C618Y families.

Founder effect

In the present study, all 12 families with the C611Y mutation shared a common haplotype, suggesting that the Danish families originate from a recent common ancestor. A recent Norwegian study reported two C611Y families, where at least one family has known Danish ancestors (EM Opsahl 2017, pers. commun.) (1). C611Y carriers have also been reported in Germany, the Netherlands, Portugal, Czech Republic, and Hungary (19 –23). Although unclear, one might speculate that some of these carriers, especially those from countries in proximity to Denmark, might be related to the Danish families. Relatedness to the C611Y carriers reported in China, Japan, Iran, and the United States seems less likely (24 –27).

A corresponding founder effect has been investigated in other large haplotype studies (28 –32). A recent study of four Brazilian and eight Greek families suggested common ancestors for the G533C mutation (28). Another Brazilian study of eight M918V families suggested a common founder mutation with origin in Portugal (29). Conversely, a Spanish study of 14 families with the C634Y mutation and an international collaborative study of 32 families with the C634R mutation failed to identify common haplotypes (30,31). Additionally, a large Italian multicenter study reported 49 V804M families but excluded a founder effect for this mutation (32).

De novo mutations

The likelihood of de novo mutations in the C611Y cohort, comprising 108 RET proven mutation carriers, seems infinitesimal, as all the families share the same haplotype. Several studies have reported C611Y carriers (1,11,14,19 –27,33 –43). Only one of these studies has performed haplotype analysis of families with the C611Y mutation. The study found a common haplotype in three of four apparently unrelated C611Y families in Portugal (21). Along with the present results, this indicates that the C611Y mutation rarely arises de novo.

As for the other included families carrying the C618F, C618Y, C620R, and C634R mutations, sample sizes are too small for meaningful comments on the de novo occurrence.