Mutations in the UBA domain of SQSTM1 are a common cause of Paget's disease of bone. Here we show that the most common disease-causing mutation (P392L) is carried on a shared haplotype, consistent with a founder effect and a common ancestral origin.
Introduction: Paget's disease of bone (PDB) is a common condition with a strong genetic component. Mutations affecting the ubiquitin-associated (UBA) domain of sequestosome 1 (SQSTM1) have recently been shown to be an important cause of PDB. The most common mutation results in a proline to leucine amino acid change at codon 392 (P392L), and evidence has been presented to suggest that there may be a recurrent mutation rather than a founder mutation on an ancestral chromosome. Because marked geographical differences exist in the prevalence of PDB, we have investigated the frequency of SQSTM1 mutations in different populations and looked for a founder effect on chromosomes bearing SQSTM1 UBA domain mutations.
Materials and Methods: We conducted mutation screening of SQSTM1 and performed haplotype analysis using the PHASE software program in 83 kindreds with familial PDB, recruited mainly through clinic referrals in the United Kingdom, Australia, and New Zealand. Similar studies were conducted in 311 individuals with PDB who did not have a family history and 375 age- and sex-matched controls from the United Kingdom.
Results: The proportion of patients with familial PDB who had SQSTM1 UBA domain mutations varied somewhat between referral centers from 7.1% (Sydney, Australia) to 50% (Perth, Australia), but the difference between centers was not statistically significant. Haplotype analysis in 311 British patients with PDB who did not have a family history and 375 age- and sex-matched British controls showed that two common haplotypes accounted for about 90% of alleles at the SQSTM1 locus, as defined by common single nucleotide polymorphisms (SNPs) in exon 6 (C916T, G976A) and the 3′UTR (C2503T, T2687G). These were H1 (916T-976A-2503C-2687T) and H2 (916C-976G-2503T-2687G). There was no significant difference in haplotype distribution in PDB cases and controls, but the P392L mutation was found on the H2 haplotype in 25/27 cases (93%), which is significantly more often than expected given the allele frequencies in the normal population (odds ratio, 13.2; 95% CI, 3.1-56.4; p < 0.0001). Similar findings were observed in familial PDB, where 12/13 (92%) of P392L mutations were carried on H2 (odds ratio 17.2; 95% CI, 2.2-138; p = 0.001).
Conclusions: These results provide strong evidence for a founder effect of the SQSTM1P392L mutation in PDB patients of British descent, irrespective of family history. Our results imply that these individuals share a common ancestor and that the true rate of de novo mutations may be lower than previously suspected.
PAGET'S DISEASE OF bone (PDB; MIM 167250, 602080) is a common disorder, affecting between 1% and 3% of individuals >55 years of age in white populations.(1, 2) The cause of PDB is not completely understood, but genetic factors play an important role, reflected by the fact that 15–40% of affected individuals have at least one affected first-degree relative.(3–5) Numerous extended families have been described where PDB is inherited in an autosomal dominant manner, and several susceptibility loci have been identified by genome-wide searches on chromosomes 5, 2, 9, 10, and 18.(6–11) Positional cloning studies have shown that mutations in the sequestosome 1 gene (SQSTM1; MIM 601530) are the cause of 5q35-linked PDB.(12, 13) A proline-leucine mutation affecting codon 392 (P392L) in the ubiquitin-associated (UBA) domain of SQSTM1 was initially identified as a cause of familial and sporadic PDB in the French-Canadian population,(12) and subsequently, other mutations affecting the UBA domain of SQSTM1 were described in familial and sporadic PDB.(13) To date, 11 different mutations of SQSTM1 have been identified in PDB, and all of these cluster in the UBA domain.(14–19) While these studies show that SQSTM1 mutations cause PDB in many different populations, it remains unclear whether these represent recurrent mutations, as originally suggested by Laurin et al.,(12) or whether there is evidence of a founder mutation carried on an ancestral chromosome. We have investigated this possibility by looking for evidence of a founder effect in PDB patients from Britain and from Australia and New Zealand whose white populations are mainly of British descent.
MATERIALS AND METHODS
Patients with familial and sporadic PDB and normal controls were recruited as previously described.(13) Patients were classified as having “sporadic” PDB if they did not have a positive family history of the disease and familial PDB if they knew of at least one other blood relative who had been diagnosed with PDB. In all cases, PDB was diagnosed on the basis of standard clinical criteria.(20) Subjects with familial PDB included in this study comprised 69 families described by Hocking et al.(15) and an additional 14 families recruited from a clinic population in Sydney, Australia. Subjects with sporadic PDB included in the study were recruited from routine clinic referrals in Aberdeen and Liverpool. All of the PDB patients and controls were white. Pedigree diagrams for all of the families are available on the Bone Research Group website at the University of Aberdeen. (http://www.abdn.ac.uk/medicine_therapeutics/bone/paget%20pedigrees.shtml).
Mutation screening and single nucleotide polymorphisms genotyping
DNA was extracted from blood as previously described.(13) Mutations and polymorphisms in SQSTM1 were identified by conducting mutation screening of the proximal promoter and upstream enhancer of SQSTM1(21) (from −1855 bp to the transcription start site), the entire coding region including the intron-exon boundaries, and the 3′ untranslated region (3′UTR) by automated DNA sequencing of PCR products generated from genomic DNA. The DNA sequencing was carried out on a MegaBace 1000 DNA sequencer (Amersham), using DYEnamic ET terminator chemistry according to the manufacturer's instructions. The primer sequences and annealing temperatures used for PCR amplification of the promoter and 3′UTR are shown in Table 1, and those used to amplify the coding regions were given previously.(13) Genotyping of the single nucleotide polymorphisms (SNPs) discovered during mutation screening was carried out by automated DNA sequencing as described above, using the forward or reverse PCR primers as the sequencing primer.
Table Table 1. Primers for Mutation Screening of SQSTM1 Promoter and 3′ UTR
Haplotype analysis was carried out using the PHASE haplotype analysis program (version 2.0.2).(22) Comparison of haplotype and genotype frequencies between groups was by χ2 test using Minitab release 12.23.
We identified several mutations affecting the UBA domain of SQSTM1 as previously described,(13, 15) as well as two known polymorphisms in exon 6(12) at position 916 (C916T/Asp292Asp; refSNP ID rs4935) and 976 (G976A/Arg312Arg; rs4797) and a further two known polymorphisms in the 3′UTR at positions 2503 (C2503T; rs10277) and 2687 (T2687G; rs1065154). No polymorphisms or mutations were detected in the promoter region. Genotype frequencies for the exon 6 and the 3′ UTR polymorphisms in sporadic PDB cases without SQSTM1 mutations and controls are shown in Table 2. All SNPs were in Hardy-Weinberg equilibrium. There was no significant difference in distribution of the genotypes between cases and controls for any of the polymorphisms studied.
Table Table 2. Geotype Frequency of Common SNP at the SQSTM1 Locus in PDB Cases Without SQSTM1 Mutations and Controls
Haplotype analysis showed that all four SNPs were in strong linkage disequilibrium and that two haplotypes defined by these polymorphisms accounted for ∼90% of alleles at the SQSTM1 locus (Fig. 1). These were 916T-976A-2503C-2687T (H1) and 916C-976G-2503T-2687G (H2). The remaining alleles were accounted for by nine rare haplotypes with individual frequencies of between 0.3% and 2.1%.
Details of the origin of patients with familial PDB and the proportion of families in each center that were found to have UBA domain mutations of SQSTM1 are summarized in Table 3. There was some variation in the proportion of patients with SQSTM1 mutations in different centers mainly because of a generally lower frequency of SQSTM1 mutations in families from Sydney and Geelong, although the difference between centers was not statistically significant.
Table Table 3. Origin of PDB Families and Prevalence of SQSTM1 Mutations in Different Centers
Table 4 shows the frequency of the common haplotypes in 311 sporadic PDB cases from the United Kingdom and 375 controls, who were age- and sex-matched for the cases, and 83 kindreds with familial PDB. There was a slight excess of the H2 haplotype in sporadic PDB cases compared with controls, but this was not significant (χ2 = 3.4; p = 0.067), and when PDB patients with known UBA domain mutations of SQSTM1 were excluded from the analysis, the frequency of H1 and H2 was almost identical in cases and controls. The frequency of H1 and H2 haplotypes was similar in PDB families as in the controls and sporadic PDB cases.
Table Table 4. Haplotype Frequencies in Familial and Sporadic PDB Cases and Normal Controls
Analysis of the relationship between UBA domain mutations and haplotype background in sporadic and familial PDB is shown in Table 5. There was a highly significant difference in the frequency of H1 and H2 haplotypes in subjects who carried the P392L mutation. We found that 25 of the 27 P392L mutations (92.6%) in sporadic PDB cases were carried on the H2 background, which is 13.2 times more often than expected on the basis of population frequency of H1 and H2 alleles (95% CI, 3.5-63.8; p < 0.0001). Similar findings were observed in familial PDB where we found that 12 of the 13 families with P392L mutation carried the mutation on a H2 background, which is 17.2 times more often than expected (95% CI, 2.2-138; p = 0.001). There was no clear trend for clustering of other UBA domain mutations on one or other haplotype, although the numbers were small, and the study did not have sufficient power to detect a founder effect for the other mutations, even if such an effect was present.
Table Table 5. Haplotype Background in Familial and Sporadic PDB Associated with SQSTM1 Mutations
This study shows that the P392L mutation in SQSTM1 is carried on a common haplotype background in the vast majority of patients with PDB, irrespective of family history, indicating that the mutation is likely to be identical by descent in these individuals. Over 90% of P392L mutations in our population were found on the H2 haplotype background, which differs from the findings reported by Laurin et al.,(12) who found that the P392L mutation occurred in eight families on what they termed the PDB3H1 haplotype (corresponding to our H1 haplotype) compared with three families where the mutation occurred on the PDB3H2 haplotype (corresponding to our H2 haplotype).
An especially interesting finding to emerge from our study was that the vast majority of patients without a family history of PDB who had the P392L mutation carried it on the same haplotype background (H2) as the familial cases. While these patients were not aware of a family history of PDB, we did not undertake clinical or biochemical screening of their parents or other relatives, so it is entirely possible that these individuals may, in fact, have had occult familial PDB. This view is supported by the findings of population-based studies performed in The Netherlands that have indicated that the vast majority of subjects with mild PDB in the community are undiagnosed(23) and that many relatives of patients with familial PDB are also undiagnosed, unless specifically investigated for the presence of PDB.(17) Our data are consistent with the findings of Laurin et al., who also reported P392L mutations in 18/112 (16%) French-Canadian patients with apparently “sporadic” PDB, all of whom carried the mutation on one of the two common haplotype backgrounds referred to previously.
Our observations could partly account for the relatively high incidence of PDB in British migrants to countries such as Australia and New Zealand(24) and are consistent with the view that some founders of these populations were carriers of PDB-causing mutations. It is possible that the apparent reductions in prevalence of PDB that have been observed in some of these countries recently(25) are partly caused by dilution of the gene pool with migrants from other European countries or geographical regions such as Asia, where PDB is rare.(24) Although we did not have detailed information on the ancestry of the PDB patients who took part in this study, all were of white decent, and it is known that many founders of the populations in Australia and New Zealand were from the British Isles. This is consistent with a founder effect of the P392L mutation in PDB patients of British ancestry. Further studies will need to be performed in PDB patients from other European countries to determine whether our findings are specific to subjects of British descent or whether they also apply to PDB subjects from elsewhere in Europe.
In summary, haplotype analysis of the SQSTM1 gene in patients with familial and sporadic PDB has shown that the P392L mutation is transmitted on a common haplotype in the vast majority of cases. This is consistent with a strong founder effect, indicating that many of our patients may be linked by a common ancestor. These findings do not exclude the possibility that the UBA domain of SQSTM1 is a mutational hot spot as suggested by Laurin et al.,(12) but indicates that the new mutation rate may be lower than was previously thought.
This study was supported in part by grants from the Arthritis Research Campaign (R0589, R0616, R0544); the National Association for Relief of Paget's Disease (UK); the Medical Research Council; and the Paget's Disease Charitable Trust (Auckland, New Zealand). We acknowledge the contribution of Grace Taylor, Angela Duthie, and Sheena Main to the DNA extraction and genotyping.