🧬 Aging Wiki

crispr-base-editing-pcsk9

Properties1
aliasesVERVE-101, VERVE-102, in-vivo PCSK9 base-editing, somatic-cell base editing of PCSK9

15 min read

In-vivo CRISPR base-editing of PCSK9 (VERVE-101 / VERVE-102)

A single-administration in-vivo somatic-cell gene-editing therapy that permanently silences the hepatic PCSK9 gene using adenine base-editing (ABE) delivered by lipid nanoparticles (LNPs), producing durable LDL cholesterol reduction without a double-strand DNA break. VERVE-101 (SaCas9-ABE8e + PCSK9-targeting guide RNA) entered Phase 1b (heart-1) in patients with heterozygous familial hypercholesterolemia (HeFH) and established cardiovascular disease, making it the first in-vivo base-editing therapeutic in human clinical trials. VERVE-102, a second-generation product (reformulated GalNAc-LNP + adenine base editor v8.8), reported full Phase 1 results in NEJM in 2026 (n=35; up to −88% PCSK9 and −62% LDL-C from a single dose, durable ≥12 months, no dose-limiting toxicity) 1 — the first peer-reviewed full Phase 1 trial of an in-vivo base-editing medicine. Human evidence remains limited to Phase 1 dose-escalation; no Phase 2 or hard cardiovascular endpoint data exist as of 2026-06-08.

See pcsk9 for full PCSK9 protein biology.

Base editing — the molecular mechanism

Classical CRISPR-Cas9 creates a double-strand break (DSB) that is repaired by error-prone non-homologous end-joining (NHEJ), generating insertion/deletion mutations (indels). Base editors instead make a precise single-nucleotide change without a DSB:

  • Cytosine base editor (CBE): fuses a cytidine deaminase to a catalytically impaired Cas9 (nickase); converts C•G → T•A at a target position within the protospacer 2.
  • Adenine base editor (ABE): fuses an engineered adenosine deaminase (evolved from E. coli TadA) to a nickase Cas9; converts A•T → G•C 3. ABEs produce no DSB, no indels, and show lower off-target editing rates than CBEs.

VERVE-101/102 use an ABE targeting a splice-site or stop-codon-introducing locus in PCSK9 to permanently disrupt protein-coding function in hepatocytes. The guide RNA directs the editor to a PCSK9 intronic splice donor or coding sequence; after a single nucleotide conversion, the transcript is non-functional.

Why no DSB matters clinically: DSB-based editing can trigger p53 activation, chromosomal rearrangements, and large deletions. ABE’s nick-only mechanism substantially reduces these risks — a critical safety argument for a geroprotective application in older patients who carry more pre-existing chromosomal fragility.

PCSK9 biology and the loss-of-function precedent

PCSK9 (proprotein convertase subtilisin/kexin type 9) is a serine protease secreted by hepatocytes that binds the LDL receptor (LDLR) at the cell surface and directs it to lysosomal degradation, thereby reducing LDLR recycling and raising circulating LDL-C levels.

The therapeutic rationale is grounded in a landmark natural experiment: Cohen et al. 2006 identified naturally occurring PCSK9 loss-of-function (LOF) variants (p.Y142X and p.C679X) in Black American participants of the ARIC cohort 4. Carriers of these variants had:

  • ~28% lower LDL-C levels
  • ~88% reduction in 15-year cumulative risk of coronary heart disease (CHD)
  • No apparent adverse phenotype across all carriers studied

This is one of the strongest Mendelian randomization-style human genetics arguments in cardiovascular medicine: lifelong PCSK9 LOF, beginning at birth, confers exceptional protection against CHD with no detected cost. The gene-editing strategy is explicitly designed to reproduce this phenotype somatically in adulthood.

Existing pharmacological and RNA targets:

ClassExampleMechanismDurability
PCSK9 monoclonal antibodyAlirocumab, evolocumabNeutralizes circulating PCSK9~2-week half-life; requires q2w/monthly dosing
PCSK9 siRNA (inclisiran)InclisiranRNAi-mediated PCSK9 mRNA knockdownTwice-yearly injection; reversible
PCSK9 base editingVERVE-101/102Permanent gene inactivation in hepatocytesExpected lifelong (somatic hepatocyte persistence)

The base-editing approach offers theoretically permanent effect from a single administration — a significant practical advantage for a cumulative-exposure disease like atherosclerosis.

Preclinical foundation

Chadwick et al. 2017 — first in-vivo PCSK9 base editing (mouse)

Chadwick, Wang, and Musunuru demonstrated that CBE-mediated editing of the Pcsk9 W159 codon in adult mice reduced PCSK9 protein by ~56% and total plasma cholesterol by ~28% after a single adenoviral vector injection, using BE3 base editor 5. This established proof-of-concept but used an adenoviral delivery modality not suitable for clinical translation (large payload; immunogenicity).

Musunuru et al. 2021 — NHP in-vivo base editing via LNP

Musunuru et al. (Verve Therapeutics / Broad Institute) delivered an ABE8e-Cas9-PCSK9 mRNA + guide RNA via LNPs to cynomolgus macaques 6. Key results:

EndpointResult
Hepatic PCSK9 editing efficiencyNear-complete knockdown in liver (exact allele-editing % not reported in abstract; ~90% protein reduction)
Blood PCSK9 protein reduction~90%
LDL-C reduction~60%
DurabilitySustained for at least 8 months (end of observation)
Off-target editing (WGS)No predicted off-target sites edited above background
Liver enzyme elevations (ALT/AST)Transient mild elevation; returned to baseline

This NHP study was the direct preclinical foundation for VERVE-101. The LNP formulation targets hepatocytes via ApoE-mediated uptake, achieving high liver selectivity with minimal editing in non-hepatic tissues 6. gap/needs-replication — single NHP study; exact allele-editing efficiency not extractable from abstract alone (full PDF not accessible); long-term durability beyond 8 months in this cohort not reported (see Lee 2022 for extended follow-up).

Lee et al. 2022 — VERVE-101 IND-enabling NHP and mouse studies

Lee RG et al. (Verve) reported formal IND-enabling data for VERVE-101 in both NHPs and a murine germline study 7. Results were dose-dependent: at 1.5 mg/kg (the clinical dose), hepatic editing efficiency was ~70%, PCSK9 protein fell ~83%, and LDL-C fell ~69% (time-weighted average over days 28–476 after dosing); at 0.75 mg/kg, editing was ~46%, PCSK9 fell ~67%, and LDL-C fell ~49%. Effects were durable up to 476 days (~16 months) post-single-dose. No germline editing was detected in sperm samples from treated male NHPs, and PCSK9 editing was absent in 0/436 offspring of treated female mice. These data supported IND submission and trial initiation 7.

Clinical development

VERVE-101 — heart-1 (Phase 1b)

NCT05398029 — open-label Phase 1b dose-escalation in patients with HeFH (heterozygous familial hypercholesterolemia) and established atherosclerotic cardiovascular disease (ASCVD) on maximally tolerated lipid-lowering therapy.

  • Status: Completed (as of ClinicalTrials.gov data, 2026-05-06)
  • n: 13 participants (dose-escalation cohorts)
  • Delivery: single IV administration of LNP-formulated ABE mRNA + guide RNA
  • Primary outcomes: safety, tolerability, PCSK9 plasma levels, LDL-C at 180 days

Reported outcomes (AHA 2023 conference presentation; not yet published as full peer-reviewed paper as of 2026-05-06): Results were reported at AHA Scientific Sessions 2023. Preliminary data indicated LDL-C reductions in dose-dependent fashion in the higher-dose cohorts, consistent with NHP findings. One participant experienced a myocardial infarction (MI) — attributed to underlying ASCVD progression rather than the therapy after adjudication; the event triggered a temporary partial clinical hold by the FDA in late 2023, which was subsequently lifted to allow the trial to resume. Transient hepatic enzyme elevations (ALT/AST) were observed, consistent with LNP-mediated hepatic inflammation seen in other nucleic acid delivery programs.

Note: Full peer-reviewed Phase 1 results paper is not yet confirmed as published in the archive as of 2026-05-06. Commentaries in European Heart Journal - Cardiovascular Pharmacotherapy (Horie and Ono 8; Lewis 9; both published Feb 2024) contextualize early results from the AHA 2023 presentation. Verifiers should check for a primary results paper (expected NEJM, NEJM Evidence, or Nature Medicine) gap/needs-replication.

VERVE-102 — heart-2 (Phase 1b) — full Phase 1 results published 2026

NCT06164730 — open-label, single-ascending-dose Phase 1 in HeFH and/or premature coronary artery disease patients. Primary peer-reviewed results: Vafai et al., NEJM 2026 1 — see vafai-2026-verve-102-pcsk9 for the full extraction and in-vivo-base-editing for the platform methodology.

  • Status: Recruiting; interim analysis of n=35 (data-cutoff 2026-02-27) now published; planned enrollment up to 85
  • Editor: the NEJM report specifies VERVE-102 as an adenine base editor v8.8 — an engineered adenosine deaminase fused to a catalytically-impaired Cas9 nickase 1. (This corrects the earlier “Cas12-family compact nuclease” description on this page, which came from pre-publication conference materials; the primary source describes a Cas9-based editor.)
  • Key differences from VERVE-101: a reformulated, GalNAc-conjugated LNP (adds ASGPR-mediated hepatocyte uptake on top of ApoE/LDLR uptake) and an updated base editor; the authors attribute VERVE-102’s improved acute safety primarily to the LNP formulation + GalNAc change rather than to the editor 1.

Phase 1 efficacy (Vafai 2026, time-averaged from day 28): dose-dependent reduction in blood PCSK9 from −51% (0.3 mg/kg) to −88% (1.0 mg/kg), and LDL-C from −9% to −62% (absolute −78 mg/dL at 1.0 mg/kg; baseline 128 → 51 mg/dL). Total RNA dose correlated with LDL reduction (r=−0.68). Effects appeared durable through ≥12 months (15/35 with ≥1 yr follow-up) 1.

Phase 1 safety: no dose-limiting toxic effects, deaths, or withdrawals. Any AE 74%; infusion-related reactions 20% (all grade 1–2); transient asymptomatic ALT ≤2.4× ULN in 3/35 peaking day 3–4. One serious AE (grade 3 aspiration pneumonitis in a participant with GERD/hiatal hernia) was adjudicated unrelated. LNP terminal half-life <20 h 1.

CV-aging rationale — cumulative LDL/ApoB exposure

Atherosclerosis is a cumulative-exposure disease: the cardiovascular damage from elevated LDL-C integrates over decades. Mendelian randomization studies consistently demonstrate that lowering LDL-C earlier and more permanently provides disproportionate protection — a 1 mmol/L lifetime LDL reduction from genetic instruments confers a 54.5% (95% CI 48.8–59.5%) CHD risk reduction, ~3× greater than the same 1 mmol/L reduction achieved by a statin initiated later in life 10. The lifelong-PCSK9-loss-of-function natural experiment (pcsk9 nonsense-allele carriers) is the most direct instance of this advantage 4.

This “cumulative-exposure advantage” is the core geroprotective argument for base-editing PCSK9: a single administration in middle age or earlier could replicate the lifetime-LOF phenotype observed in Cohen cohort carriers, compounding protection over 30–40 years.

Cardiovascular aging hallmark connections:

  • chronic-inflammation — atherosclerotic plaques are inflammatory lesions; foam-cell accumulation, macrophage activation, and cytokine release from atheromas drive systemic inflammatory burden with age.
  • altered-intercellular-communication — PCSK9 also has direct signaling functions in vascular endothelium and inflammatory cells (e.g., modulating Toll-like receptor expression); reducing PCSK9 may have pleiotropic anti-inflammatory effects beyond LDL lowering. gap/no-mechanism — pleiotropic PCSK9 effects in aging context are not yet characterized.

Extrapolation table

DimensionStatusNotes
Pathway conserved in humans?yesPCSK9-LDLR axis is well-conserved; NHP results directly inform human predictions
Phenotype conserved in humans?yesNatural human LOF variants (Cohen 2006) directly demonstrate the LDL-lowering phenotype
Replicated in humans?in-progressheart-1 complete (unpublished); heart-2 (VERVE-102) full Phase 1 published, NEJM 2026 1; Phase 2 not yet initiated

Modality precedent

VERVE-101/102 represents the first in-vivo somatic base-editing therapy in human trials — establishing a regulatory and clinical precedent for a class of interventions that could include:

TargetCompany/ProgramIndication
PCSK9Verve Therapeutics (VERVE-101/102)HeFH → broad CV prevention
ANGPTL3Verve (VERVE-201)Hypertriglyceridemia
LPAMultiple programsElevated Lp(a)
APOBIn-developmentFH variants with dominant-negative APOB
CETPHypothesis-levelHDL augmentation

For aging-specific geroprotection, the significance is the demonstration that:

  1. LNP-delivered base-editor mRNA achieves durable (months+) somatic editing in human hepatocytes from a single IV dose.
  2. Somatic editing safety (off-target, hepatotoxicity) is a tractable clinical problem — not a fundamental barrier.
  3. Regulatory agencies (FDA, MHRA) can evaluate and authorize IND for in-vivo base editing in outpatient clinical settings.

This precedent is directly relevant to other aging gene-therapy programs — see aav-tert, aav-klotho, aav-follistatin, aav-osk — by showing that in-vivo somatic editing of a single gene in a post-mitotic-rich tissue (liver) is feasible and regulatorily navigable.

Translation barriers

  1. Off-target editing. ABEs can make unintended A→I edits in non-target genomic sites, or cause RNA off-target effects (A-to-I RNA editing). Whole-genome sequencing (WGS) surveillance was negative in NHP studies 6, but human WGS surveillance in clinical trials requires larger n and longer follow-up. gap/long-term-unknown

  2. LNP immunogenicity and re-dosing. LNPs activate innate immune sensors (TLR4, STING) in the liver, causing transient transaminase elevations. More importantly, LNP-induced anti-PEG antibodies may limit re-dosability if a second administration were ever needed. Single-administration durability is therefore critical — demonstrated up to ~16 months in NHPs 7 but not yet beyond that, and not yet in humans. gap/long-term-unknown

  3. Editing durability in cycling hepatocytes. Hepatocytes in adult liver turn over slowly (~1 year half-life), not negligibly. Over decades, hepatocyte regeneration from unedited progenitor cells could partially restore PCSK9 expression, gradually eroding the LDL benefit. This has not been characterized beyond ~16 months in NHP models 7. gap/long-term-unknown

  4. HeFH-to-general-population translation. Heart-1 enrolled HeFH patients on maximal lipid-lowering therapy — patients with extreme baseline LDL and high ASCVD burden. Efficacy and safety in polygenic-risk patients without extreme LDL (the broader CV-aging population) is untested. gap/needs-human-replication

  5. Cost and manufacturing scalability. LNP-mRNA base-editor manufacture is complex; current cost structure is expected to far exceed existing PCSK9 inhibitor costs (~$450/yr for inclisiran). Geroprotective use at population scale requires manufacturing cost reduction or reimbursement frameworks that do not yet exist.

  6. Regulatory novelty for geroprotection indication. FDA and EMA have no precedent for approving a preventive somatic gene edit in healthy or sub-clinical-risk individuals for aging-related indication. Current trials are in disease populations (HeFH, established ASCVD). Expanding the indication to primary CV-aging prevention will require new regulatory frameworks.

Limitations and gaps

  • VERVE-102 full Phase 1 now published (Vafai 2026, NEJM) 1 — this resolves the long-standing “no full Phase 1 paper” gap for VERVE-102. The VERVE-101 heart-1 full results paper remains unpublished as of 2026-06-08 (only AHA 2023 conference data). gap/needs-replication
  • Sample sizes are small. n=35 (VERVE-102, six dose cohorts of 4–8) and n=13 (heart-1) provide limited power for safety conclusions; rare adverse events cannot be excluded. gap/long-term-unknown
  • Long-term durability unproven in humans. NHP data extend to 8 months; hepatocyte turnover over years/decades could reduce editing efficiency. gap/long-term-unknown
  • Cancer surveillance absent. Unlike AAV-TERT, ABE/PCSK9 does not raise direct oncogenic concerns (PCSK9 is not a cancer driver), but long-term WGS surveillance for off-target editing consequences has not been published. gap/long-term-unknown
  • pcsk9 protein page is a stub. All PCSK9 protein-level detail (structure, LDLR-binding mechanism, cleavage, tissue expression, other LOF variants) should live there, not on this page.

Cross-references

Footnotes

Footnotes

  1. vafai-2026-verve-102-pcsk9 · doi:10.1056/NEJMoa2601283 · Vafai SB, Täubel J, …, Khera AV, Kathiresan S · N Engl J Med 2026 · PMID 42187087 · n=35 · in-vivo (human Phase 1, single-ascending-dose, non-randomized) · NCT06164730 (Heart-2); single IV VERVE-102 (ABE v8.8 mRNA + PCSK9 gRNA, GalNAc-LNP), 0.3–1.0 mg/kg total RNA; at 1.0 mg/kg PCSK9 −88% and LDL-C −62% (−78 mg/dL absolute), durable ≥12 mo; no dose-limiting toxicity; 1 unrelated SAE; transient ALT ≤2.4× ULN in 3/35; LNP t½ <20 h. Sponsor: Verve Therapeutics (Eli Lilly). Closed-access (NEJM) — verified against article PDF. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8

  2. doi:10.1038/nature17946 · Komor AC, Kim YB, Packer MS, Zuris JA, Liu DR · Nature 2016 · 533(7603):420-424 · cytosine base editor (BE3); programmable C→T conversion without DSB in mammalian cells; efficiency ~15–75% of total cellular DNA with ≤1% indel formation; founding CBE paper; 5169 citations · archive: verified PDF (green OA / PMC) ↩

  3. doi:10.1038/nature24644 · Gaudelli NM, Komor AC, Rees HA, Packer MS, Badran AH, Bryson DI, Liu DR · Nature 2017 · 551(7681):464-471 · adenine base editor (ABE); evolved E. coli TadA adenosine deaminase fused to Cas9-nickase; programmable A•T → G•C editing without DSB; ~50% editing efficiency in human cells with ≥99.9% product purity and ≤0.1% indels (7th-generation ABE); founding ABE paper; 3984 citations · archive: verified PDF (green OA / PMC) ↩

  4. doi:10.1056/NEJMoa054013 · Cohen JC, Boerwinkle E, Mosley TH, Hobbs HH · N Engl J Med 2006 · 354(12):1264-1272 · observational (Mendelian randomization-equivalent); ARIC cohort; 3363 Black participants, 9524 White participants; Black: Y142X MAF=0.8%, C679X MAF=1.8%, combined carrier=2.6%; ~28% lower LDL-C in carriers; HR 0.11 (95% CI 0.02–0.81, P=0.03) for CHD = 88% lower 15-year CHD risk; White: R46L MAF=3.2%, ~15% lower LDL-C, HR 0.50 (95% CI 0.32–0.79, P=0.003) = 47% lower CHD risk; 3123 citations ↩ ↩2

  5. doi:10.1161/ATVBAHA.117.309881 · Chadwick AC, Wang X, Musunuru K · Arterioscler Thromb Vasc Biol 2017 · in-vivo (mouse, C57BL/6J, 5-week-old male, n=5/group); single IV adenoviral vector delivery of BE3-Pcsk9; ~56% PCSK9 protein reduction; ~28% plasma cholesterol reduction; proof-of-concept for somatic PCSK9 base editing in adult liver; no detected off-target edits or indels above control by deep sequencing; 214 citations · archive: verified PDF (bronze OA) ↩

  6. doi:10.1038/s41586-021-03534-y · Musunuru K, Chadwick AC, Mizoguchi T, et al. · Nature 2021 · 593(7859):429-434 · in-vivo (cynomolgus macaque NHP); single IV LNP-delivered ABE8e + guide RNA; near-complete knockdown of PCSK9 in liver; ~90% PCSK9 reduction in blood; ~60% LDL-C reduction; effects stable for at least 8 months post-single-dose; no predicted off-target sites edited by WGS; 702 citations · archive: not_oa — abstract-verified only via PubMed (PMID 34012082); full PDF not accessible gap/no-fulltext-access ↩ ↩2 ↩3

  7. doi:10.1161/CIRCULATIONAHA.122.062132 · Lee RG, Mazzola AM, Braun MC, et al. · Circulation 2023 · 147(3):242-253 · in-vivo (cynomolgus NHP + murine germline study); IND-enabling VERVE-101 studies; at 1.5 mg/kg: ~70% hepatic editing, ~83% PCSK9 protein reduction, ~69% LDL-C reduction (time-weighted avg, day 28–476); at 0.75 mg/kg: ~46% editing, ~67% PCSK9, ~49% LDL; effects durable up to 476 days post-single-dose; no germline editing detected in sperm or offspring; transient ALT/AST elevation resolved by day 14; 182 citations · archive: not_oa — abstract-verified only via PubMed (PMID 36314243); full PDF not accessible gap/no-fulltext-access ↩ ↩2 ↩3 ↩4

  8. doi:10.1093/ehjcvp/pvad103 · Horie T, Ono K · Eur Heart J Cardiovasc Pharmacother 2024 (published Feb 23, 2024; 10(2):89-90) · commentary/editorial; contextualizes heart-1 AHA 2023 presentation data; not primary trial results · archive: download failed (bronze OA — PDF access issue) ↩

  9. doi:10.1093/ehjcvp/pvad095 · Lewis BS · Eur Heart J Cardiovasc Pharmacother 2024 (published Feb 23, 2024; 10(2):87-88) · news from AHA; summarizes first-in-human VERVE-101 heart-1 findings reported at AHA Scientific Sessions 2023; not peer-reviewed primary results paper · archive: download failed (bronze OA — PDF access issue) ↩

  10. ference-2012-ldl-mr-lifetime · doi:10.1016/j.jacc.2012.09.017 · Ference BA, Yoo W, Alesh I, et al. · JACC 2012;60(25):2631–2639 · PMID 23083789 · n=312,321 · Mendelian randomization meta-analysis (9 polymorphisms / 6 genes) · model: human genetics · 54.5% (95% CI 48.8–59.5%) CHD reduction per 1 mmol/L lower lifelong LDL-C, ~3× the statin per-unit effect, p=8.43×10⁻¹⁹, I²=0.0% · closed-access, abstract-confirmed. ↩

Graph View

Backlinks