Geographical genetic variability: a factor to consider when assessing clinical implications of PRDM9

Due to its role during meiosis, variations in the PRDM9 nucleotide sequence have been associated with different pathologies of meiotic origin: infertility (Irie et al. 2009), de novo genomic disorders (Berg et al. 2010; Borel et al. 2012), and childhood leukemogenesis (Hussin et al. 2013) (OMIM accession number: 609760). So far, alleles of the PRDM9 zinc finger array have been determined in studies of meiotic recombination, primarily in Icelanders (Kong et al. 2010) and in individuals of central European or African descent (Baudat et al. 2010; Berg et al. 2010, 2011; Kong et al. 2010; Borel et al. 2012; Hussin et al. 2013). To determine the variability in the zinc finger array of PRDM9 in a Spanish control population (n = 103) and in parents of patients with a de novo 22q11.2 deletion (n = 16), this genetic region was investigated through Sanger sequencing analysis. Frequencies of alleles found in our control and case populations are shown in Table 1. Four new zinc finger coding repeat types and eight new alleles (Table 1) have been identified. Three of the new zinc finger coding repeat types (spm1 referred as s mutant, spm3 as j mutant, and spm4 as d mutant in (Jeffreys et al. 2012)) and two of the new alleles (L39, L44) had been previously described as de novo somatic/germ-line rearranged PRDM9 mutants that occur in blood/sperm of individuals carrying known alleles in a publication on PRDM9 instability (Jeffreys et al. 2012). We show that these “blood/ sperm mutants” can also be found as a polymorphic variant in our population. Considering this fact and the number (in the order of hundreds) of PRDM9 “blood/sperm mutants” found by Jeffreys et al. 2012, one can only expect that this gene’s population variability is higher than the variability reported up to now. When analyzing the possible clinical implications of PRDM9 as a risk factor for genomic instability, Berg et al. (2010) obtained results that suggested that rare allelic forms of this gene confer a protective effect against rearrangements in three highly unstable minisatellites and de novo CMT1A and HNPP rearrangements. Later, Borel et al. (2012)) showed that the majority of the de novo microdeletions they studied were not dependent on non–A alleles. In accordance to all these previous results, we also observe a higher frequency of the A allele in our case population (87.5%) than in our control population (72.33%) (Table 1). The sample size of our case group, transmitting parents of a de novo 22q11.2 deletion, however, is too small to achieve statistical significance. Also regarding PRDM9 clinical relevance, it has been suggested that differences in PRDM9 allelic frequencies between populations correlates with differences in the susceptibility to de novo rearrangements (Berg et al. 2010). If true, this hypothesis could be extended to other diseases in which an implication of PRDM9 is assumed. For example, a recent study (Hussin et al. 2013) associates rare allelic forms of PRDM9 with childhood leukemogenesis. The authors observed a statistically significant higher frequency of both non–A alleles and k-finger alleles (alleles containing a k zinc finger coding repeat type) in parents of children with acute lymphoblastic leukemia (ALL) (32.7% and 13.5%, respectively) compared to controls (15.1% and 3.3%, respectively). In our control population, non–A alleles and k-finger alleles (C, D, L19, and L20) were observed at an allelic frequency of 27.67% and 9.2%, respectively. It would be interesting to replicate this study in a population with a higher incidence of k-finger alleles, such as the one described in this article, in order to confirm the implication of these alleles in the generation of ALL. Given the differences observed among only three different populations (Table 1) and considering the many “blood/sperm mutants” found by Jeffreys et al. (2012), we believe that future studies aiming to increase understanding of the role of PRDM9 alleles in determining human susceptibility to meiotic disorders or other diseases would benefit from having a broader picture of PRDM9 variability in human populations.

Due to its role during meiosis, variations in the PRDM9 nucleotide sequence have been associated with different pathologies of meiotic origin: infertility (Irie et al. 2009), de novo genomic disorders (Berg et al. 2010;Borel et al. 2012), and childhood leukemogenesis (Hussin et al. 2013) (OMIM accession number: 609760). So far, alleles of the PRDM9 zinc finger array have been determined in studies of meiotic recombination, primarily in Icelanders (Kong et al. 2010) and in individuals of central European or African descent (Baudat et al. 2010;Berg et al. 2010Berg et al. , 2011Kong et al. 2010;Borel et al. 2012;Hussin et al. 2013). To determine the variability in the zinc finger array of PRDM9 in a Spanish control population (n = 103) and in parents of patients with a de novo 22q11.2 deletion (n = 16), this genetic region was investigated through Sanger sequencing analysis. Frequencies of alleles found in our control and case populations are shown in Table 1. Four new zinc finger coding repeat types and eight new alleles (Table 1) have been identified. Three of the new zinc finger coding repeat types (spm1 referred as s mutant, spm3 as j mutant, and spm4 as d mutant in (Jeffreys et al. 2012)) and two of the new alleles (L39, L44) had been previously described as de novo somatic/germ-line rearranged PRDM9 mutants that occur in blood/sperm of individuals carrying known alleles in a publication on PRDM9 instability (Jeffreys et al. 2012). We show that these "blood/ sperm mutants" can also be found as a polymorphic variant in our population. Considering this fact and the number (in the order of hundreds) of PRDM9 "blood/sperm mutants" found by Jeffreys et al. 2012, one can only expect that this gene's population variability is higher than the variability reported up to now.
When analyzing the possible clinical implications of PRDM9 as a risk factor for genomic instability, Berg et al. (2010) obtained results that suggested that rare allelic forms of this gene confer a protective effect against rearrangements in three highly unstable minisatellites and de novo CMT1A and HNPP rearrangements. Later, Borel et al. (2012)) showed that the majority of the de novo microdeletions they studied were not dependent on non-A alleles. In accordance to all these previous results, we also observe a higher frequency of the A allele in our case population (87.5%) than in our control population (72.33%) ( Table 1). The sample size of our case group, transmitting parents of a de novo 22q11.2 deletion, however, is too small to achieve statistical significance.
Also regarding PRDM9 clinical relevance, it has been suggested that differences in PRDM9 allelic frequencies between populations correlates with differences in the susceptibility to de novo rearrangements (Berg et al. 2010). If true, this hypothesis could be extended to other diseases in which an implication of PRDM9 is assumed. For example, a recent study (Hussin et al. 2013) associates rare allelic forms of PRDM9 with childhood leukemogenesis. The authors observed a statistically significant higher frequency of both non-A alleles and k-finger alleles (alleles containing a k zinc finger coding repeat type) in parents of children with acute lymphoblastic leukemia (ALL) (32.7% and 13.5%, respectively) compared to controls (15.1% and 3.3%, respectively). In our control population, non-A alleles and k-finger alleles (C, D, L19, and L20) were observed at an allelic frequency of 27.67% and 9.2%, respectively. It would be interesting to replicate this study in a population with a higher incidence of k-finger alleles, such as the one described in this article, in order to confirm the implication of these alleles in the generation of ALL.
Given the differences observed among only three different populations (Table 1) and considering the many "blood/sperm mutants" found by Jeffreys et al. (2012), we believe that future studies aiming to increase understanding of the role of PRDM9 alleles in determining human susceptibility to meiotic disorders or other diseases would benefit from having a broader picture of PRDM9 variability in human populations. sored by predoctoral fellowship from the Conselleria d'Educaci o, Cultura i Universitats (Govern Illes Balears), and Fons Social Europeu. We thank C. Juan and J. Pons for their helpful comments on technical aspects of this work.

Conflict of Interest
None declared.