Abstract
BACKGROUND/OBJECTIVES: Observational research indicated a potential relationship between inflammatory cytokines and early age-related macular degeneration (AMD), but causal associations remain unclear. SUBJECTS/METHODS: Using summary data from genome-wide association studies (GWAS) of 41 different inflammatory cytokines and AMD (including early and late stages), we employed a two-sample Mendelian randomisation (MR) design. Analytical methods including inverse variance weighted (IVW), weighted median, weighted mode and MR-Egger were conducted to detect the causal associations. 'mRnd' online analysis tool was used to evaluate the statistical power of MR estimates. MR-Egger, MR pleiotropy residual sum and outlier (MR-PRESO), Cochran's Q and leave-one-out method were applied for sensitivity analysis. RESULTS: Our study selected 452 single nucleotide polymorphisms (SNPs) as instrumental variables (IVs) for following analysis. The IVW results showed that inflammatory cytokines, such as IL-10, GCSF, SCF, IL-1ra, SCGFb and MIG were causally associated with AMD, and MIG, SCGFb and Dry AMD were causally related, while IL-1ra, GCSF and Bngf were causally related to Wet AMD. Interestingly, however, our results suggest no strong causal effect between other inflammatory cytokines and AMD. The analysis results did not reveal horizontal pleiotropy or heterogeneity, and the results are considered robust. The approved drugs or kinase inhibitors predicted by DsigDB to be druggable were obtained mainly from target genes such as ALK, F12 and HTR7. CONCLUSIONS: Our MR study suggested that specific inflammatory cytokines may have an impact on overall AMD, but no causal relationship was observed with early AMD. The relationship between them needs further study.