Alzheimer's disease as a causal risk factor for diabetic retinopathy: a Mendelian randomization study

Abstract

Objectives: This study aims to investigate the causal relationship between Alzheimer's Disease (AD) and Diabetic Retinopathy (DR).

Methods: Employing Mendelian Randomization (MR), Generalized Summary-data-based Mendelian Randomization (GSMR), and the MR-Steiger test, this study scrutinizes the genetic underpinnings of the hypothesized causal association between AD and DR, as well as its Proliferative DR (PDR) and Non-Proliferative DR (NPDR) subtypes. Comprehensive data from Genome-Wide Association Studies (GWAS) were analyzed, specifically AD data from the Psychiatric Genomics Consortium (71,880 cases/383,378 controls), and DR, PDR, and NPDR data from both the FinnGen consortium (FinnGen release R8, DR: 5,988 cases/314,042 controls; PDR: 8,383 cases/329,756 controls; NPDR: 3,446 cases/314,042 controls) and the IEU OpenGWAS (DR: 14,584 cases/176,010 controls; PDR: 8,681 cases/204,208 controls; NPDR: 2,026 cases/204,208 controls). The study also incorporated Functional Mapping and Annotation (FUMA) for an in-depth analysis of the GWAS results.

Results: The MR analyses revealed that genetic susceptibility to AD significantly increases the risk of DR, as evidenced by GWAS data from the FinnGen consortium (OR: 2.5090; 95% confidence interval (CI):1.2102-5.2018, false discovery rate P-value (P_FDR)=0.0201; GSMR: b_xy=0.8936, b_{xy_se}=0.3759, P=0.0174), NPDR (OR: 2.7455; 95% CI: 1.3178-5.7197, P_FDR=0.0166; GSMR: b_xy=0.9682, b_{xy_se}=0.3802, P=0.0126), and PDR (OR: 2.3098; 95% CI: 1.2411-4.2986, P_FDR=0.0164; GSMR: b_xy=0.7962, b_{xy_se}=0.3205, P=0.0129) using DR GWAS from FinnGen consortium. These results were corroborated by DR GWAS datasets from IEU OpenGWAS. The MR-Steiger test confirmed a significant association of all identified instrumental variables (IVs) with AD. While a potential causal effect of DR and its subtypes on AD was identified, the robustness of these results was constrained by a low power value. FUMA analysis identified OARD1, NFYA, TREM1 as shared risk genes between DR and AD, suggesting a potential genetic overlap between these complex diseases.

Discussion: This study underscores the contribution of AD to an increased risk of DR, as well as NPDR and PDR subtypes, underscoring the necessity of a holistic approach in the management of patients affected by these conditions.

Keywords: Alzheimer’s disease; FUMA; Mendelian randomization; causal association; diabetic retinopathy.

Figure 1

An overview of the study design.

Figure 2

Mendelian randomization results of causal effects between AD and DR. (A) MR analysis with AD as exposure. (B) MR analysis with DR, NPDR and PDR as exposure.

Figure 3

Manhattan plot of the gene-based test as computed by MAGMA based on input GWAS summary statistics of AD (A), DR (B), NPDR (C) and PDR (D). Genome wide significance (red dashed line in the plot) was defined at P = 0.05/N.

Figure 4

Genetic risk loci identified by FUMA analysis. (A) Genetic risk loci for AD. (B) Genetic risk loci for DR. (C) Genetic risk loci for NPDR. (D) Genetic risk loci for PDR. Genomic risk loci was displayed in the format of “chromosome: start position–end position”. Each genomic locus was represented by a series of histograms, arranged from left to right to display the size of the locus, the number of candidate SNPs, the number of mapped genes, and the number of known genes located within it.

Figure 5

The analysis of tissue enrichment by FUMA using GTEx v8 (n=54 tissues). (A) Tissue enrichment for AD. (B) Tissue enrichment for DR. (C) Tissue enrichment for NPDR. (D) Tissue enrichment for PDR.