Genetically proxied inhibition of angiotensinogen synthesis is associated with lower cardiovascular risk
Panagiotis ZangasiD, Murad OmaroviD, Lanyue ZhangiD, Marios K. GeorgakisiD
Abstract
Background and AimsAngiotensinogen synthesis inhibitors have shown promising blood pressure-lowering effects in early-stage trials, but their impact on cardiovascular outcomes remains unknown. We investigated associations between genetic variants mimicking angiotensinogen synthesis inhibition and cardiovascular phenotypes. MethodsWe developed a genetic proxy for hepatic angiotensinogen synthesis downregulation comprising AGT variants that lower liver AGT expression (N=1,183) and circulating angiotensinogen levels (N=47,745), selected to mimic the effects of RNA-based angiotensinogen-targeting therapies. Using drug-target Mendelian randomization, we assessed effects on coronary artery disease (210,842 cases, 1,167,328 controls), stroke (110,182 cases, 1,503,898 controls) and heart failure (207,306 cases, 2,151,210 controls), along with vascular endophenotypes and safety outcomes. ResultsMirroring pharmacological angiotensinogen synthesis inhibitors in trials, the AGT genetic instrument was associated with lower systolic (SBP, -0.60 [-0.71;-0.48] mmHg) and diastolic blood pressure (DBP, -0.40 [-0.46;-0.33] mmHg), and higher renin and potassium levels. Genetically proxied angiotensinogen synthesis inhibition was associated with lower odds of coronary artery disease (OR per mmHg SBP reduction: 0.954 [0.937-0.972]), stroke (OR: 0.949 [0.928-0.970]) and heart failure (OR: 0.972 [0.957-0.987]) with effect sizes proportional to the SBP-lowering effects of genetic proxies for other renin-angiotensin-aldosterone system drug classes. We found additional associations with lower burden of atherosclerosis, cerebral small vessel disease, and adverse cardiac remodeling on imaging endophenotypes. Aside from hyperkalemia, we detected no links to major safety concerns, including impaired kidney function. ConclusionsGenetic downregulation of angiotensinogen synthesis is associated with lower cardiovascular disease burden without concerning safety signals, supporting the potential of angiotensinogen inhibitors to reduce cardiovascular risk. Structured graphical abstractO_ST_ABSKey QuestionC_ST_ABSIs there human genetic evidence suggesting that inhibition of hepatic angiotensinogen synthesis can reduce long-term cardiovascular risk? Key FindingGenetically proxied angiotensinogen synthesis inhibition is associated with lower risk of coronary artery disease, stroke and heart failure, as well as favorable effects on cardiac and cerebrovascular pathologies, without raising major safety concerns. Effect estimates were comparable in magnitude to those observed for genetic proxies of approved RAAS-blocking therapies. Take-home MessageHuman genetic evidence supports the hypothesis that angiotensinogen synthesis inhibition may reduce both cardiovascular event risk and chronic subclinical vascular disease burden, providing a strong rationale for prioritizing angiotensinogen inhibitors in cardio- and cerebrovascular outcome trials. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=119 SRC="FIGDIR/small/26347887v1_ufig1.gif" ALT="Figure 1"> View larger version (49K): org.highwire.dtl.DTLVardef@264dd8org.highwire.dtl.DTLVardef@c0c6f6org.highwire.dtl.DTLVardef@46752dorg.highwire.dtl.DTLVardef@1b71c12_HPS_FORMAT_FIGEXP M_FIG Genetically proxied angiotensinogen synthesis inhibition and cardiovascular risk reduction. Graphical overview of the study design. Created in https://www.biorender.com/ AGT: angiotensinogen, pQTL: protein quantitative trait loci, eQTL: expressive quantitative trait loci, LD: linkage disequilibrium, MR: Mendelian Randomization, SBP: systolic blood pressure, DBP: diastolic blood pressure, UKBB: UK Biobank, CAD: coronary artery disease, RAAS: renin-angiotensin-aldosterone system, CVD: cardiovascular disease, ICH: intracerebral hemorrhage, SAH: subarachnoid hemorrhage, cIMT: carotid intima-media thickness, cSVD: cerebral small vessel disease, HF: heart failure, HFrEF: heart failure with reduced ejection fraction, HFpEF: heart failure with preserved ejection fraction, MRI: magnetic resonance imaging, eGFR: estimated glomerular filtration rate, UACR: urinary albumin-to-creatinine ratio, MVP: Million Veteran Program C_FIG
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 0
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (9 Mar 2026) · CC BY 4.0
AGT variants mimicking hepatic angiotensinogen synthesis inhibition are linked, via drug-target Mendelian randomization, to lower coronary artery disease, stroke and heart failure risk and higher potassium.
RNA-based angiotensinogen synthesis inhibitors (e.g., zilebesiran) reduce blood pressure in phase II trials, yet their effect on cardiovascular events is unknown and the first outcome trial, ZENITH, will not read out before 2030. Drug-target genetic proxies provide early human evidence on efficacy, safety and relative benefit versus approved RAAS blockers, which can guide development decisions ahead of trial results.
Key results
ShownGenetically mimicked AGT synthesis inhibition, scaled per 1 mmHg SBP lowering, was associated with lower odds of all three primary outcomes. The result held with weighted median and MR-Egger methods, with outcome GWAS that excluded UK Biobank (UKBB), and with the complementary instruments.
IVW OR per 1 mmHg SBP reduction: CAD 0.954 (95% CI 0.937; 0.972, p=3.8*10^-7), stroke 0.949 (0.928; 0.970, p=2.1*10^-6), HF 0.972 (0.957; 0.987, p=2.4*10^-4). Weighted median ORs 0.960, 0.950 and 0.974. No Egger intercept signal. HF showed between-variant heterogeneity (p<0.05). · Angiotensinogen genetic perturbation is…
ShownThe cis-AGT instrument behaves like the drug. It lowers SBP and DBP, raises renin and potassium, and its liver expression signal colocalizes with plasma AGT levels. Its scaled blood pressure effect resembles zilebesiran in KARDIA-1, and a negative control showed no association.
46-variant Olink instrument (all F>10) explained 15.2% of plasma AGT variance; colocalization PPH4 99%. Per 1-SD AGT reduction: SBP −0.597 mmHg (p = 3*10^-23), DBP −0.396 mmHg, renin beta 0.089, potassium beta 0.026. Scaled UKBB score: 2.6/2.3 mmHg SBP/DBP vs 3.3/2 mmHg for zilebesiran. Left-handedness negative control null. · Genetic instruments; The genetic instrument proxies pharmacological…
SuggestedPer unit of SBP lowering, the AGT effects on CAD, stroke and HF did not differ significantly from genetic proxies for ACE inhibitors, angiotensin receptor blockers and renin inhibitors. AGT was the only target associated with all three outcomes.
Comparator instruments had 3 variants in ACE, 1 in AGTR1 and 2 in REN, selected at p<10^-5. NR3C2 had no qualifying variants and was excluded. Cochran Q-derived p >0.1 for each primary outcome. Only the AGT SBP-lowering instrument reached p<0.05 for CAD, stroke and HF. · Comparison with approved RAAS blockers; Angiotensinogen genetic perturbation is…
ShownThe instrument was also linked to lower risk of MI, angina, ischemic and small-vessel stroke, and SAH. It was linked to less coronary calcification, fewer cerebral small vessel disease MRI markers, less carotid plaque and lower left ventricular mass, but not to HF subtypes.
Per 1 mmHg SBP reduction: MI OR 0.923 (p=1.2*10^-7), CAC beta −0.171, ischemic stroke OR 0.946, small-vessel stroke OR 0.941 (p=0.047), SAH OR 0.883, WMH beta −0.042, mean diffusivity −0.045, LVM standardized beta −0.068 (p=1.8*10^-9), LVmeanWT −0.042. Null for HFrEF, HFpEF and NT-proBNP. · Angiotensinogen genetic perturbation is…
SuggestedApart from higher potassium, the primary instrument showed no adverse kidney signal and no consistent liver signal. A 326-outcome PheWAS found mostly protective associations, 13 of them circulatory endpoints, plus a few risk signals with no clear pattern.
Per 1-SD plasma AGT reduction: CKD OR 0.966 (0.936; 0.998, p=0.035). AST beta 0.012 (p=0.003) with no ALT, ALP, gamma-GT or bilirubin change. The SomaScan instrument was linked to higher gamma-GT and lower eGFR and UACR. PheWAS: 40 outcomes at FDR p<0.05 (33 lower odds, 7 higher, e.g., atrioventricular block, bladder cancer). · Safety-related outcomes; Phenome-wide analyses highlight repurposing…
Methods, data and reproducibility
Drug-target two-sample Mendelian randomization with cis-AGT instruments (Olink plasma pQTLs primary; SomaScan pQTLs and liver eQTLs for sensitivity), scaled per mmHg SBP lowering for cardiovascular and imaging outcomes and per SD of plasma AGT for safety outcomes. Also colocalization, UK Biobank genetic scores compared with KARDIA-1, ACE/AGTR1/REN comparator proxies and a 326-outcome PheWAS.
- Olink-measured plasma angiotensinogen GWAS in UK Biobank (primary instrument source) · human (UK Biobank, European ancestry) · n = 47,745
- SomaScan-measured plasma angiotensinogen GWAS from deCODE genetics (sensitivity instrument) · human (Icelandic) · n = 35,559
- Liver eQTL meta-analysis of 4 studies for hepatic AGT expression (sensitivity instrument) · human (European ancestry) · n = 1,183
- CAD GWAS meta-analysis combining CARDIoGRAMplusC4D, UK Biobank and Biobank Japan · human · n = 210,842 cases, 1,167,328 controls
- GIGASTROKE stroke GWAS, including ischemic stroke subtypes · human · n = 110,182 cases, 1,503,898 controls
- Heart failure GWAS meta-analysis across eight cohorts/consortia · human · n = 207,306 cases, 2,151,210 controls
- UK Biobank individual-level genotypes, blood pressure and plasma AGT for genetic score analyses · human · n = 384,908 (BP, untreated); 44,428 (plasma AGT)
- FinnGen, MVP and UK Biobank PheWAS meta-analysis of 326 binary EHR-derived outcomes · human · n = 730,332 to 1,546,730 per outcome
- Methods & tools
- TwoSampleMR R package (0.6.19), coloc R package (Bayesian colocalization), plink 2.00a (genetic scores), Inverse variance weighted (IVW) MR, MR Egger regression, Weighted median MR, Wald ratio, Cochran's Q test, Benjamini-Hochberg FDR correction, Olink proteomics platform, SomaScan aptamer assay, Clumping and thresholding genetic scores
- Reagents & environment
- TwoSampleMR R package version 0.6.19; plink 2.00a for genetic scores; coloc R package (version not given); genomic coordinates on GRCh37-hg19.
- Reproducibility gaps
- GWAS accessions appear only in Supplementary Table 2. Vascular dementia summary statistics are available on request from MEGAVCID. UK Biobank data require an approved application. No analysis code repository is reported.
Limitations (6)
- Although the liver eQTL resource is the largest available (N=1,183 from four studies), it lacks the power to detect additional hepatic expression-modifying AGT variants; the eQTL instrument had only 2 variants. · stated by the authors
- MR reflects lifelong genetic lowering of AGT, which may differ from short-term drug exposure. It also cannot capture effects of the liver-targeted delivery chemistry, such as GalNAc conjugation. · stated by the authors
- Effects on neighboring genes, horizontal pleiotropy or linkage disequilibrium cannot be fully ruled out. Most variants had only modest blood pressure effects, which risks weak instrument bias, though that bias would push estimates toward the null. · stated by the authors
- Most contributing GWAS were of European ancestry, which limits generalization to other ancestries. Kidney results differed between the Olink and SomaScan instruments and need confirmation in trials. · stated by the authors
- The comparison with approved RAAS drugs relies on comparator proxies with very few variants at a relaxed p<10^-5 threshold (1 variant for AGTR1, none for NR3C2). The lack of heterogeneity therefore gives weak support for non-inferiority. · noted in this summary
- Secondary and safety outcomes used an uncorrected p<0.05 threshold across many tests, and several associations lie close to it (small-vessel stroke p=0.047, carotid plaque p=0.049, SAH p=0.027). · noted in this summary
Who should read:Read this if you evaluate RNA-based angiotensinogen inhibitors such as zilebesiran, design drug-target Mendelian randomization studies, or weigh RAAS blockade for cerebral small vessel disease and cardiac remodeling.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Analysis of the Combined Effect of rs699 and rs5051 on Angiotensinogen Expression and Hypertension
Nicholas R. Powell, Tyler Shugg, … Todd C. Skaar·Chronic Diseases and Translational MedicinePeer-reviewed·Dec 26, 2023
- Genetic reprogramming by brief inhibition of the renin-angiotensin system in spontaneously hypertensive rats leads to persistently reduced kidney renin and low blood pressure.
Sean Byars, Priscilla Ribeiro Prestes, … Stephen B. Harrap·bioRxiv·Jul 30, 2023
- Gene polymorphism associated with Angiotensinogen(M235T), Endothelial lipase (584C/T) and susceptibility to coronary artery disease: A meta-analysis
Hongyan Zhao, Ranzun Zhao, … Jidong Rong·Bioscience ReportsPeer-reviewed·Jul 1, 2020
- Norvaline reduces blood pressure and induces diuresis in rats with inherited stress-induced arterial hypertension.
M. A. Gilinsky, Y. K. Polityko, … S. E. Naumenko·bioRxiv·Jun 21, 2019
- ACE inhibition and cardiometabolic risk factors, lung ACE2 and TMPRSS2 gene expression, and plasma ACE2 levels: a Mendelian randomization study
Dipender Gill, Marios Arvanitis, … Stephen Burgess·Royal Society Open SciencePeer-reviewed·Nov 1, 2020
- Early inflammatory response mediated by Angiotensin II in cardiac arteries of normotensive mice.
T. C. d. S. Oliveira, K. A. d. S. Viegas, … S. Lacchini·bioRxiv·Nov 6, 2023