2026-06-08

[2026-06-08] log

[2026-06-08] ingest | VERVE-102 in-vivo PCSK9 base editing — Phase 1 (Vafai 2026, NEJM) + method ramifications (user-prompted)

User supplied a new study (NEJMoa2601283, PMID 42187087, DOI 10.1056/NEJMoa2601283) with a local article PDF, explicitly asking to weight the method’s ramifications equally with the work done. This is the first peer-reviewed full Phase 1 report of an in-vivo base-editing medicine. Extraction was done by the main agent directly from the article PDF (read end-to-end, 11 pp). The study page and method page are seeded verified: false + ⚠️ banner and queued for an independent verifier pass against the same PDF.

Why this slots cleanly

The wiki already had a mature intervention page, crispr-base-editing-pcsk9 (verified 2026-05-06), whose verified-scope explicitly flagged “no peer-reviewed full Phase 1 paper has appeared.” This ingest fills that exact gap for VERVE-102 (VERVE-101/heart-1 full paper remains unpublished). All target biology (pcsk9, ldlr, lipoprotein-metabolism, atherosclerosis, familial-hypercholesterolemia) already existed and is verified — this was a tighten-the-chain + new-method-page exercise, not a broad seed.

Added

• added: studies/vafai-2026-verve-102-pcsk9.md (type: study, in-vivo/human Phase 1, n=35, verified:false). Full Design/Endpoints/Results/extrapolation/caveats extraction. Headline: single IV VERVE-102 (ABE v8.8 mRNA + PCSK9 gRNA, GalNAc-LNP), 0.3–1.0 mg/kg total RNA; dose-dependent PCSK9 −51%→−88% and LDL-C −9%→−62% (−78 mg/dL absolute at 1.0 mg/kg); durable ≥12 mo (15/35 ≥1 yr); no dose-limiting toxicity; 1 unrelated SAE (aspiration pneumonitis); transient ALT ≤2.4× ULN in 3/35; LNP t½ <20 h; total-RNA-dose↔LDL r=−0.68.
• added: methods/in-vivo-base-editing.md (type: method, category gene-editing, verified:false) — the “method ramifications” deliverable. Covers ABE/CBE chemistry (DSB-free nickase + deaminase), LNP-mRNA-GalNAc delivery + hepatotropism, the off-target landscape (Cas-dependent/independent DNA, transcriptome-wide RNA A-to-I, bystander, germline), surveillance tiers (GUIDE-seq/CIRCLE-seq + error-corrected sequencing → cross-links duplex-sequencing), durability-through-turnover, and evidence-weight implications (proxy vs allele-level editing; permanence cuts both ways). Target-validation cross-link to mendelian-randomization.

Updated (propagation)

• updated: interventions/gene-therapy/crispr-base-editing-pcsk9.md — rewrote the VERVE-102 section with published Phase 1 efficacy/safety; updated intro, extrapolation table, and Limitations to mark the VERVE-102 publication gap resolved (heart-1 still open); added [^vafai2026] footnote + vafai-2026-verve-102-pcsk9/in-vivo-base-editing cross-refs; bumped literature-checked-through 2026-05-08 → 2026-06-08 with a dated verified-scope addendum. Correction: the page previously described VERVE-102 as switching to a “Cas12-family compact nuclease” (from pre-publication conference materials); the NEJM primary source states an ABE v8.8 with a Cas9 nickase — corrected on the page and noted as a supersession.
• updated: molecules/proteins/pcsk9.md — “In vivo CRISPR base editing” section now carries the published Phase 1 headline + [^vafai2026] footnote + in-vivo-base-editing link (page stays verified:true; new claim is footnoted and mirrors the queued study page).

Conceptual frames touched (paper-impact enumeration)

• Intervention modality: durable single-dose (“one-and-done”) somatic editing vs lifelong drug/mAb/siRNA — see interventions-by-modality.
• Comparative/aging framing: gene editing as the durable end of the intervention spectrum / “Tier 3” longevity strategy — longevity-escape-velocity.
• Method evidence-weighting: new reusable guidance for how future study pages should weight in-vivo editing claims (editing-measurement tier, off-target depth, durability vs turnover).

Gaps surfaced / follow-ups

• Independent verifier pass queued for both new pages against the local article PDF (study page is the primary claim home; method page’s vafai-derived numerics are a subset).
• #gap/needs-replication — a second, independent in-vivo base-editing PCSK9 Phase 1 trial is cited within Vafai 2026 (Nat Med 2026); ingest when resolved to a primary source. DOIs for secondary in-vivo-editing programs (nexiguran ziclumeran TTR; the Nat Med trial) not yet captured — add on the verifier pass.
• #gap/long-term-unknown — human durability beyond ~3 yr and genome-wide off-target consequences over an oncogenic latency are unestablished for any in-vivo base-editing program.
• #stub — GUIDE-seq / CIRCLE-seq off-target-nomination assays referenced but unbuilt.
• study-design enum note: a non-randomized single-ascending-dose Phase 1 has no exact enum value (in-vivo | in-vitro | observational | rct | meta-analysis | review). Used in-vivo per existing convention (single-arm human Phase 1 → in-vivo, body clarifies non-randomized). Possible future schema value for non-randomized clinical trials.