Impact of ACLY on type 2 diabetes in UK Biobank: a Mendelian randomisation study

Abstract

Background: Cholesterol lowering therapies have greatly reduced the burden of cardiovascular disease. Bempedoic acid inhibits ATP citrate lyase (ACLY), a key enzyme in the cholesterol biosynthesis pathway, and has shown promising results in early-phase randomised controlled trials. Targeting this mechanism may therefore provide a further therapeutic option for lowering LDL cholesterol (LDL-C) and cardiovascular risk. In contrast to statin therapy, randomised comparisons suggest that the incidence of newonset diabetes is lower with bempedoic acid than with placebo. Mendelian randomisation offers the potential to use genetic variants in the ACLY gene to mimic the on-target effects of bempedoic acid. We used this approach to explore the association between genetic inhibition of ACLY and risk of type 2 diabetes (T2DM). Purpose: To assess the impact of lifelong genetic inhibition of ACLY associated with lower LDL-C levels on T2DM and glycaemic control (alongside effects on coronary heart disease [CHD]), investigate whether the association varies by background T2DM risk, and explore potential mediating factors. Methods: We undertook a 2-sample Mendelian randomisation study of ∼0.3M unrelated White British individuals in UK Biobank, including 26,495 cases of T2DM and 37,272 cases of CHD. Causal effects on these outcomes were estimated using the inverse variance weighted approach, based on 9 independent LDL-C associated ACLY (±500 kb) genetic variants weighted by their LDL-C effects from previously published studies. Results for T2DM and CHD were then meta-analysed with comparable analysis of data from DIAGRAM (n=26,676 T2DM cases) and CARDIo-GRAMplusC4D (n=60,801 CHD cases) respectively. Mediation analyses were conducted to explore the relevance of adiposity traits, such as body mass index, on the observed associations. Subgroup analyses considering the effects in individuals at varying risk of T2DM (based on polygenic risk), as well as the effects of ACLY on glycaemic control in those with established diabetes were undertaken. Results: Genetic variation in ACLY was associated with a 16% lower risk of T2DM (odds ratio [OR]: 0.84; 95% CI; 0.76-0.92) per 10mg/dL lower ACLYpredicted LDL-C level. The magnitude and direction of this effect was comparable to that observed with CHD (OR: 0.84; 95% CI; 0.77-0.91). There were no differences in the estimated effects of ACLY on T2DM across tertiles of T2DM polygenic risk (p-trend=0.63). Furthermore, the association between ACLY and T2DM was not materially mediated through effects on body mass index. Analyses exploring additional potential mediating factors, the relevance of inflammation, and the effects of ACLY on glycaemic control in individuals with diabetes are also reported. Conclusion: This Mendelian randomisation study shows that genetic variants that mimic inhibition of ACLY and are associated with lower LDL-C levels lower the risks of developing T2DM as well as of CHD events.