Genetic analysis of sight-threatening diabetic retinopathy in southern Chinese patients with type 2 diabetes

Abstract

Background and objectives: Recent genetic studies, including candidate gene association studies and genome-wide association studies (GWAS), have identified a list of potential susceptibility loci for diabetic retinopathy (DR). Whether the identified susceptibility variants would also show similar effects in the Southern Chinese population remains to be investigated. The objective of this thesis was to evaluate these reported DR-associated genetic variants in the Southern Chinese population and investigate the effect of glycaemic control on the genetic susceptibility to DR. Major findings: Significant associations of 5 genetic variants identified from previous candidate gene association studies for DR were successfully validated in a cross-sectional case-control study of Southern Chinese patients with T2DM, which involved 673 sight-threatening DR (STDR) cases and 1596 non-STDR controls. Two variants showing significant associations with STDR were found in the well-studied VEGFA locus. The other 3 STDR-associated variants were detected in the NOS3, CHN2 and IRS2 loci. Although the association of CHN2 rs39059 with STDR was attenuated after adjustment for glycated haemoglobin (HbA1c), meta-analysis of CHN2 rs39059 with previous studies confirmed the association and demonstrated an even stronger association with DR. Sixty-nine novel variants showing top associations with DR in previous GWAS were examined in a cross-sectional case-control study, which involved 1051 STDR cases and 2042 non-STDR controls. The associations of COL5A1 rs59126004, IGSF21-KLHDC7A rs3007729, CREB5 rs11765845 and LOC728275-LOC72831 rs227455 with STDR were successfully confirmed in the current study, with COL5A1 rs59126004 being the most significant finding. The associations of COL5A1 rs59126004 and IGSF21-KLHDC7A were further confirmed in the subgroup analysis examining the associations with proliferative DR (PDR). Same 69 variants were also examined for their interactions with glycaemic control on the risk of STDR. The effect of glycaemic control on genetic susceptibility to DR was examined with both dichotomised HbA1c (HbA1c <7% versus ≥7%) and continuous HbA1c. COL5A1 rs59126004, CAMK4 rs2300782 and KIAA1804-KCNK1 rs6662352 all showed significant interactions with dichotomised HbA1c on the risk of STDR, with COL5A1 rs59126004 showing the most significant result. COL5A1 rs59126004 also showed significant interaction with dichotomised HbA1c on the risk of PDR. Stratified analysis showed that COL5A1 rs59126004 was only associated with STDR or PDR in patients with adequate glycaemic control (HbA1c level <7%), with 42% and 63% reduced risk of STDR and PDR, respectively, for carrying each additional C allele of rs59126004. Both associations in the stratified analysis were able to survive the stringent Bonferroni correction. Significant interactions with continuous HbA1c on the risk of STDR were detected at COL5A1 rs5912600, MALRD1 rs17670074, CREB5 rs11765845 and KIAA1804-KCNK1 rs6662352. Only COL5A1 rs59126004 and MALRD1 rs17670074 showed significant interactions with continuous HbA1c on the risk of PDR. Conclusions: The findings of this thesis have validated several susceptibility variants identified from previous genetic studies for DR, and provided novel evidence on the role of glycaemic control in modulating the genetic susceptibility to DR. These genetic risk factors may be employed for risk profiling in patients with T2DM and facilitate the development of personalised DR risk management.