In vivo PCSK9 base editing (VERVE-102) as one-time treatment for familial hypercholesterolemia: lessons from HEART-2

In vivo adenine base editing to disrupt PCSK9 in hepatocytes represents a paradigm-shifting approach to cardiovascular disease — a single infusion to permanently lower LDL-cholesterol by mimicking the cardioprotective PCSK9 loss-of-function phenotype observed in naturally occurring human carriers. Verve Therapeutics has developed VERVE-102, an LNP-delivered adenine base editor targeting PCSK9 in liver, now in Phase 1b clinical trials (HEART-2). The c.399+1G>A splice donor variant in ClinVar illustrates the principle: loss of PCSK9 function (whether from natural variants or therapeutic editing) increases hepatocyte LDL receptor surface expression and reduces circulating LDL-C. Human genetic epidemiology provides extraordinary validation: heterozygous PCSK9 loss-of-function carriers have ~28% lower LDL-C and ~88% reduced coronary heart disease risk (Cohen et al., 2006, PMID: 16554528).

The HEART-1 Phase 1 trial of VERVE-101 (first-generation, using a different editing approach) demonstrated dose-dependent PCSK9 protein reduction (up to 84%) and LDL-C reduction (up to 55%) in patients with heterozygous familial hypercholesterolemia. VERVE-102 uses an improved GalNAc-targeted LNP formulation with an adenine base editor to install a splice-site disruption in PCSK9. Preclinical studies in non-human primates showed durable (>2 year) PCSK9 knockdown after a single IV dose. The genetic validation is uniquely strong: PCSK9 is one of the best-validated drug targets in human genetics, with decades of Mendelian randomization, clinical trial (evolocumab, alirocumab), and human genetics data supporting LDL-C lowering as causal for coronary disease reduction. Natural homozygous PCSK9 loss-of-function individuals (e.g., with Q152H) are healthy with very low LDL-C, supporting the safety of near-complete knockdown.

Off-target editing remains the primary safety concern for a therapy intended for potentially millions of patients (FH prevalence ~1 in 250). While adenine base editors have lower off-target DNA editing than Cas9 nucleases, RNA off-target deamination and rare off-target genomic edits must be comprehensively characterized. Liver inflammation (transaminase elevation) was observed in HEART-1, likely related to the LNP delivery vehicle and/or immune recognition of the editing machinery. The irreversibility of gene editing means that any unexpected long-term consequences cannot be reversed — unlike statins or PCSK9 monoclonal antibodies, which can be discontinued. For the specific c.399+1G>A variant, patients who already carry a pathogenic PCSK9 loss-of-function allele may have different risk-benefit considerations if they are already partially PCSK9-deficient. Regulatory and ethical frameworks for deploying gene editing in common diseases (beyond rare monogenic disorders) are still evolving.