Background There is a shortage of OA susceptibility loci in spite of large collaborative GWAS. Therefore, it is possible that other polymorphisms could contribute to OA genetics. Microsatellites are good candidates because due to their multiple alleles they lack sufficient LD with SNPs to be detected in GWAS. In addition, a microsatellite in BMP5 is specially promising because it has been found associated with hip OA in British women during investigation of a previously identified hip OA linked locus in 6p12.3-q131. However, no replication studies have been done. In addition, this microsatellite has shown a modifier effect on expression of the BMP5 gene, which is involved in joint and chondorcyte functions.
Objectives We aimed to explore association of the BMP5 microsatellite with knee OA
Methods We have studied 1003 patients (716 women and 287 men) that have undergone total joint replacement because of primary knee OA and 1543 controls (758 women and 785 men) of similar age. Patients with OA and controls were recruited in Oxford (UK; 360 patients, 698 controls), Thessaly (GR; 369 patients, 383 controls) and Santiago de Compostela (ES; 274 patients, 462 controls). Genotypes of the D6S1276 microsatellite were determined by capillary electrophoresis after amplification including a FAM labeled primer. Comparison of frequencies was done with the omnibus test implemented in CLUMP and with the 2xn Mantel-Haenszel approach implemented in R to account for sample collection stratification.
Results Valid genotypes were obtained in 93.5% of the samples and this percentage was similar in patients and controls. Allele frequency distribution was similar to the already described for Europeans. There were 9 different alleles corresponding to 5 to 13 TCTA repeats. The most common allele contained 10 repeats, followed for the 9 and 8 repeat alleles.
Comparison of allele frequency distributions by sample collections showed a significant difference between British knee OA patients and controls (P=0.002), but not among the Greek or Spanish. Combined analysis of the three collections, showed a significant difference between patients and controls (P=0.006). When the results were stratified by gender, there were significant differences in the Greek (P=0.03) and British (P=0.04) women; and almost significant in British men (P=0.05). In the combined analysis only women showed a different distribution (P=0.001). Specific allele analysis showed that there was an excess of alleles with 9 repeats in control women (P=0.023) or men (P=0.02; and P=0.002 all together) and an excess of alleles with 8 repeats in knee OA women patients (P=0.004), but not men.
Conclusions We have found significant association between the intron 1 BMP5 microsatellite and knee OA. This result together with the reported association of the same microsatellite with hip OA1, indicate that this is a probable OA susceptibility locus that has escaped detection in GWAS due to its many alleles. Other microsatellite, in the ASPN gene, has already shown strong evidence of association with OA2. Further studies will determine the degree to which microsatellites contribute to OA susceptibility.
Wilkins JM, et al. BMC Med Genet. 2009;10:141
Nakamura T, et al. Hum Mol Genet. 2007;16:1676-81
Disclosure of Interest None Declared