Archeogenetics of F11 p.Cys38Arg: a 5400-year-old mutation identified in different southwestern European countries

TO THE EDITOR:

Factor XI (FXI) is a key element for the amplification of thrombin generation, which participates in the contact pathway. This explains why FXI deficiency has minor physiological relevance but protects against thrombotic events, supporting FXI as an excellent target for antithrombotic treatments.^1,2 

FXI deficiency (MIM #612416) has been considered a rare disorder that might reach relatively high prevalence in certain populations. The best example is within the Ashkenazi Jews, demonstrating an incidence of 8% of heterozygous individuals for 2 mutations: p.Glu117Ter and p.Phe283Leu, which account for 98% of alleles.³  So far, 220 different causative mutations have been identified throughout the FXI gene (F11) (http://www.factorix.org/). Recently, the exome-based data obtained from >60 000 individuals (Exome Aggregation Consortium) have revealed new information regarding genetic variation in F11, showing (a) profound differences in heterozygous frequency among populations, and (b) evidence of recurrent and ethnic-specific mutations: (1) p.Phe223Leu in Africans (23.5% of all mutated alleles); (2) p.Gln263Ter and p.Leu424CysfsTer in East Asians (28.2% and 20.5%, respectively); and (3) p.Ala412Thr in Latinos (25%).⁴ 

The screenings of FXI deficiency for 20 years in Yecla, a small town in the southeast of Spain, have revealed a recurrent point mutation, c.166T>C p.Cys38Arg, identified in 24 out of 46 unrelated cases (52%), and reached a frequency of 2% in the general population.⁵  This mutation was initially described in the French Basque country as the most prevalent genetic defect causing FXI deficiency (8 out of 12 cases: 66.7%) and also with high frequency in the general population (1%).⁶  The same mutation was identified in French Brittany in 3 out of 12 cases (25%).⁷  One Portuguese patient also had this mutation.⁸  Here, we detected this mutation in 14 out of 55 (25.5%) unrelated cases with FXI deficiency from Barcelona (in the northeast of Spain but 487 km away from Yecla). However, it has never been described in patients with FXI deficiency from non-European populations^9-11  and was also absent from large UK or Italy cohorts of patients (supplemental Figure 1, available on the Blood Web site).^12,13 

Because p.Cys38Arg mutation was thought to be related to a founder effect in the French Basque country,⁶  the present study was performed to confirm a potential common founder effect and to estimate its occurrence time.

This study was conducted in 64 subjects with the p.Cys38Arg mutation (30 from Yecla; 27 from Barcelona; 3 from the French Brittany; and 4 from the French Basque Country) well characterized by molecular, functional, and biochemical methods previously described.⁵  In addition, 20 unrelated noncarrier family trios from Yecla were enrolled as controls. All subjects gave their informed consent to enter the study, which was performed according to the Declaration of Helsinki (Edinburgh 2000) (see supplemental Materials for additional information).

The whole F11 (23 Kpb) was sequenced by next-generation sequencing (supplemental Materials).⁵  All carriers shared a unique intragenic haplotype. The selection of single-nucleotide polymorphisms with minor allele frequency (MAF) < 0.6 revealed the common F11 haplotype containing 13 polymorphisms, rs566328623 (MAF 0.007) being the most informative variant linked to the p.Cys38Arg mutation (Table 1).

Interestingly, rs4253413 c.486-361C>T (MAF 0.217) was exclusively detected in all carriers from Yecla, but was absent from all other subjects.

Four microsatellite markers covering 1.7 Mbp upstream and 1.8 Mbp downstream of F11 (supplemental Figure 2; supplemental Materials) were also evaluated.

Extragenic haplotypes were established from homozygous cases and by allele segregation among family. In cases without information, we chose the allele associated with the lowest number of recombination events.

Microsatellite analysis revealed at least 23 potential different extragenic haplotypes among carriers (Table 2), none present in any of the 60 healthy subjects from the family trios coming from Yecla (Fisher’s exact test, P < .0001).

The extragenic diversity was lower among carriers from small or isolated populations as Yecla and Brittany than in carriers from the big city of Barcelona (Table 2), as also the linkage of disequilibrium analysis showed (supplemental Table 1; supplemental Materials).

Although the extragenic haplotype was only partially determined in carriers from the Basque Country, microsatellite data showed a remarkable genetic diversity, particularly in the markers closer to F11 (supplemental Table 2).

Estimation of the p.Cys38Arg mutation age was carried out using the Bayesian method proposed by Rannala and Reeve¹⁴  and implemented by Disease Mapping Linkage disEquilibrium software (DMLE+2.3).¹⁵  The applied map distance, obtained from deCODE map,¹⁶  was 3.1 cM/Mb. The growth population rate (r) was obtained from the census results in Yecla during the last 551 years (in 1450, the population was 1400; in 2001, the population was 30 824), considering 25 years for a generation and according to the following formula: r = (LnP₁ − LnP₀)/g (P = population, g = generations). Thus, Yecla’s population growth rate was estimated to be 0.14. For Europe, we used the growth population rate of the last 2000 to 3000 years (r = 0.034) described elsewhere.¹⁷ 

The proportion of disease-bearing chromosomes, f = 0.1% in Europe and f = 4% in Yecla, was calculated by (1) estimating the frequency of the p.Cys38Arg mutation (in Europe 1/35 000 [http://exac.broadinstitute.org/variant/4-187192873-T-C], and in Yecla 2/100⁵ ); (2) considering population sizes (Europe: 700 × 10⁶ and Yecla: 35 000); and (3) the number of chromosomes analyzed (46 for the European cohort and 27 for Yecla).

The density peaked at 216 generations (95% confidence interval: 149-317) when estimating a proportion of 0.2% for the mutated allele according to data from the Exome Aggregation Consortium. The estimated age of this mutation was ∼5400 years. The estimated time of mutation occurrence in Yecla was ∼30 generations (95% confidence interval: 19-50), which represents 625 years (supplemental Figure 3).

FXI deficiency denomination as hemophilia C has certainly contributed to broaden the conclusions on the rarity of this disorder. Textbooks on general medicine, clinical hematology, and hemostasis still report a low prevalence for FXI deficiency (1/10 000-50 000) in the general population. However, increasing evidence from epidemiological studies,^5,11,13  the limitation of current diagnostic methods to identify FXI deficiency,¹⁸  the clinical mildness of bleeding in even cases with severe FXI deficiency,¹⁹  and our current study strongly suggest that the incidence of this disorder might be underestimated.

Our study shows new hot spots of FXI deficiency in Europe raised by a new recurrent and significantly older mutation than that found in the Ashkenazi Jews. The recurrent Ashkenazi’s p.Glu117Ter and the p.Phe283Leu mutations, deriving from relatively few founders that have maintained rather high endogamous rates over time,^3,20  were dated back from 120 to 185 and 31 to 100 generations ago, respectively.^20,21  We estimated that the p.Cys38Arg variant occurred ∼149 to 317 generations ago. The remarkable extragenic genetic diversity observed in the French Basque Country suggests that this variant might have appeared in this population. Then, it was spread out to the north (reaching the French Brittany) and to the south (arriving to Spain and Portugal). Our study reveals that this variant arrived to Yecla in the thirteenth century, probably during the Christian Reconquest of Southern Spain, with a new intragenic marker. It is also important to consider possible positive selection mechanisms to explain the high prevalence of this mutation in certain populations (Yecla or the French Basque Country), as it has been described for other hemostasis genetic disorders.²²  Resistance to infection, hemostasis, or iron conservation has been interpreted as an adaptive profile that might positively select certain gene variations.²³  Within this framework, we think that this potential positive selection might not be explained by the antithrombotic protective effect of FXI deficiency.²  A real positive selection could be based on the evidence supporting that FXI deficiency may confer a survival advantage in sepsis by altering the cytokine response to infection and blunting activation of the contact system.²⁴  Alternatively, drift hypothesis cannot be ruled out, because these regions were geographically isolated, with low inward migration and a high degree of consanguinity, as it has been described for other founder mutations involved in different diseases.²⁵