In a recent study published in the journal Neurology, researchers assessed the genetic variation among targets of Alzheimer’s disease risk and anti-diabetic drugs.
Alzheimer’s disease (AD) and type 2 diabetes mellitus (T2DM) are highly prevalent diseases experienced by the elderly population. A study has shown that T2DM patients are at a 53% higher risk of reporting AD. Furthermore, AD has been deemed type 3 diabetes due to the impaired glucose control and insulin resistance in the brain. Several clinical trials have assessed the impact of anti-diabetic drugs in AD patients.
About the study
In the present study, researchers performed Mendelian randomization (MR) to assess the impact of genetic variation observed in anti-diabetic drug targets on the risk of AD.
The team used a two-sample MR design to extract data related to exposure and outcome from two independent non-overlapping patient groups. Genetic variants present within the genes that were responsible for encoding the protein targets of anti-diabetic drugs were assessed. These variants were detected in the genome-wide association studies (GWAS) dataset for blood glucose levels and were used as a proxy for the use of anti-diabetic drugs.
Reduced blood glucose levels are an established response toward anti-diabetic treatment; hence, it was used as an important biomarker in the study. The causal estimation was verified using three MR model assumptions: (1) relevance was assessed based on the significant correlation between instrumental variables (IVs) and target proteins, (2) exchangeability was assessed as per independence of IVs on confounding factors, and (3) exclusion restriction was examined as the absence of any direct impact of IVs on AD risk, except that via drug targets.
Furthermore, the team extracted IV-exposure association from a GWAS of blood glucose examined on participants from UK Biobank who had European ancestry. In the association setting, the team employed a mixed linear method based on a model to adjust for population stratification according to principle components and relatedness as per a genetic relationship matrix.
The AD summary statistics were also extracted from a GWAS, wherein the phase 1 involved the meta-analysis of clinically diagnosed late-onset AD and control cases. In phase 3, AD-by-proxy and control cases were also meta-analyzed. The dataset that included clinically diagnosed AD cases were used for primary analysis, while those comprising AD / AD-by-proxy cases were used in a sensitivity analysis.
The team identified a total of seven classes of anti-diabetic drugs, which included glucagon-like peptide-1 (GLP-1), thiazolidinediones (TZDs), sodium-glucose cotransporter (SGLT2) inhibitor, dipeptidyl peptidase 4 (DPP-4) inhibitor, metformin, insulin/insulin analogs, and sulfonylureas. Furthermore, the cis-variants present within each encoding gene were identified and the variants related to blood glucose were retained.
Furthermore, a positive control analysis was conducted with insulin resistance, insulin secretion, T2DM, and obesity-related traits as outcomes. IVs for insulin or insulin analogs were selected within the insulin receptor (INSR); hence it was expected to affect the function of the insulin receptor while reducing insulin resistance.
The study results showed a significant association between reduced risk of T2DM and genetic variation among the drug targets of insulin analogs, GLP-1, sulfonylureas, and TZDs. Furthermore, genetic variation among the drug targets of GLP-1 analogs and sulfonylureas was correlated to enhanced insulin secretion. In contrast, TZD and insulin or insulin analogs were associated with reduced insulin resistance.
Concerning obesity-related traits, the estimates of sulfonylureas and TZD indicated an increase in body mass index (BMI), hip circumference (HIP), and waist circumference (WC). However, estimates for GLP-1 analogs and insulin or insulin analogs varied across the three traits related to obesity.
The team also noted that genetic variation within sulfonylurea targets correlated with decreased AD risk. Moreover, there was a suggestive correlation between lower AD risk and genetic variation among GLP-1 analog targets across the different MR methods. The team also observed no heterogeneity within substantial pleiotropy or IVs for all the estimates.
Colocalization analysis was conducted for GLP-1 analogs and sulfonylureas within the genes that encoded the drug targets. The team did not find any significant evidence that indicated colocalization between AD and blood glucose within the genetic regions related to sulfonylureas. In contrast, colocalization was noted in the regional association plots. In the case of GLP=1 analogs, there was no evidence to suggest either a distinct colocalization trend or a common variant within the GLP-1 receptor.
Overall, the study findings showed a significant association between genetic variation among sulfonylurea targets and reduced risk of AD. The researchers believe further research could analyze the underlying association mechanism between AD and sulfonylureas.