Skip to content

Drug-Target Mendelian Randomization of Lipid-Modifying Therapies

Summary

Richardson et al. (2022) used cis-Mendelian randomization to compare the genetically predicted effects of 8 lipid-modifying drug targets — spanning LDL-cholesterol, HDL-cholesterol, and triglyceride pathways — on coronary artery disease (CAD) risk and on 249 NMR-derived circulating metabolic traits in up to 115,082 UK Biobank participants. Despite broadly consistent effects on CAD risk, drug targets acting on different lipoprotein classes produced very different downstream metabolomic signatures. This divergence has direct implications for selecting pharmacodynamic biomarkers in early drug development.[1]

Design

Genetic risk scores were constructed for 8 targets grouped by their primary lipoprotein effect: LDL cholesterol — HMGCR, PCSK9, NPC1L1; HDL cholesterol — CETP; triglycerides — APOC3, ANGPTL3, ANGPTL4, LPL. Each score's effect was estimated on CAD risk and, separately, on 249 metabolic traits (lipoprotein subfractions, fatty acids, amino acids, glycolysis-related metabolites, inflammatory markers) measured by the Nightingale NMR platform.[1]

Findings

  • All 8 scores showed evidence of an effect on CAD risk except ANGPTL3.[1]
  • Genetically predicted effects on the 249-trait metabolic signature were highly concordant within a lipoprotein class: the linear fit between HMGCR- and PCSK9-proxied effects across all 249 traits was r²=0.91.[1]
  • Effects were much less concordant across lipoprotein classes: e.g. HMGCR vs. each of the four triglyceride targets all had r²<0.02.[1]
  • Glycoprotein acetyls (GlycA), an inflammatory biomarker, illustrated this divergence directly: genetically predicted LDL-lowering (HMGCR/PCSK9/NPC1L1) had only a weak effect on GlycA, whereas genetically predicted triglyceride-lowering had a strong GlycA-lowering effect.[1]

Interpretation (as stated by the authors): target-class-consistent CAD risk reduction does not imply target-class-consistent metabolic/biomarker effects — a distinction relevant to selecting dose-ranging biomarkers and anticipating off-target pharmacodynamic effects in lipid-drug development.[1]

See Also

Citations

[1] Richardson, Leyden, Wang, Bell, Elsworth, Davey Smith, Holmes (2022), "Characterising metabolomic signatures of lipid-modifying therapies through drug target mendelian randomisation", PLOS Biology, 20(2), e3001547. Supports: design and all findings above. Location: Abstract; Results (CAD risk estimates; cross-target r² comparisons; GlycA case study).