AAV Gene Delivery

Definition

Adeno-associated virus (AAV) vectors are the dominant platform for cardiac gene therapy. They are non-integrating (largely episomal), low-immunogenicity, replication-deficient viral vectors capable of transducing post-mitotic cardiomyocytes. AAV9 is the most widely used serotype for cardiac gene delivery due to its cardiotropism. As of January 2024, AAV9 was used in 92% of published cardiac arrhythmia gene therapy studies; 88% of those studies relied on small-animal (mouse/rat) models — the dominant translational gap.

Key Concepts

Vector Biology

• Packaging capacity — 4.7 kb ceiling: Standard AAV vectors accommodate ~4.7 kilobases of genetic payload. This precludes direct gene replacement for large disease-causing genes: RYR2 (~15,000 nt) and SCN5A (~6,048 bp) both exceed this limit, making packaging size the principal determinant of therapeutic strategy selection. (sources/gene-therapy-arrhythmia-2025)
• Workarounds for large genes:
  • Dual-vector (trans-splicing) and split-vector systems: Divide large constructs across two AAV vectors that reconstitute in the cell. Performance is inconsistent relative to single-vector delivery. (sources/gene-editing-cv-tcm-2025)
  • Mini-gene constructs: Truncated functional domains; unpredictable expression dynamics.
  • Pathway-targeted therapy: Bypasses size constraints by targeting shared molecular nodes — e.g., MOG1 (Nav1.5 trafficking chaperone) for BrS instead of full-length SCN5A; CASQ2 overexpression for CPVT1 instead of RYR2. (sources/gene-therapy-arrhythmia-2025)
• AAV9 — de facto standard serotype: Used in 92% of cardiac arrhythmia gene therapy studies. Demonstrated efficacy for CASQ2 replacement (CPVT), SupRep for LQTS, PKP2 replacement (ARVC), CaMKII peptide inhibitor delivery, SCN5A base editing (LQT3), and CRISPR-SaCas9 for CPVT1. (sources/gene-therapy-arrhythmia-2025, sources/aav-gene-therapy-arrhythmia-hr-2024)
• AAV2i8 — chimeric next-generation vector: Combines AAV2 and AAV8 capsid elements; broad striated muscle tropism; reduced liver transduction compared to AAV8 → improved cardiac specificity for systemic delivery. Not yet used in arrhythmia studies. (sources/aav-gene-therapy-arrhythmia-hr-2024)
• AAVMYO — striated muscle-specific vector: Improved tropism for cardiac and skeletal muscle; reduces mosaicism by improving homogeneity of cardiac transduction compared to AAV9; used in combination with cardiac-specific peptides. (sources/gene-editing-cv-tcm-2025)
• Non-viral alternatives:
  • Lipid nanoparticles (LNPs) encapsulating mRNA: Transient, non-integrating expression; low immunogenicity; repeated dosing feasible. Safety demonstrated in EPICCURE trial (VEGF-A mRNA).
  • Naked mRNA: TBX18 mRNA (unformulated) established biological pacing in rats and pigs — proof-of-concept for mRNA-based cardiac reprogramming. (sources/gene-therapy-arrhythmia-2025)

Routes of Delivery

• Intravenous: Least invasive; suitable for systemic disease. Significant off-target liver accumulation; lower cardiac transduction efficiency — requiring higher doses with greater immune and hepatic exposure. (sources/gene-therapy-arrhythmia-2025)
• Intracoronary: Catheter-delivered directly to coronary circulation; better cardiac uptake and reduced hepatic distribution than IV. Less effective in the presence of coronary artery disease (reduced perfusion limits vector distribution). Requires cath lab access. (sources/gene-therapy-arrhythmia-2025)
• Intramyocardial: Targeted regional delivery; used intraoperatively. Most precise spatial control but invasive; limited spread from injection site; risk of localised inflammation and arrhythmias at injection site. (sources/gene-therapy-arrhythmia-2025)
• No head-to-head RCT defines the optimal route for any cardiac gene therapy indication — route selection is currently extrapolated from animal studies and manufacturing constraints. (sources/gene-therapy-arrhythmia-2025)

Delivery Challenges & Limitations

• Immune responses — single-dose constraint: Pre-existing or treatment-induced anti-AAV neutralising antibodies limit efficacy and prevent re-dosing with the same serotype. Mitigations under development: transient immunosuppression, capsid re-engineering, extracellular vesicle shielding — none validated in humans. For chronic cardiac diseases requiring lifelong therapy (HCM, LMNA cardiomyopathy), the single-dose constraint is a major unresolved barrier. (sources/gene-therapy-arrhythmia-2025)
• Mosaicism — >70% correction threshold for structural disease: In vivo cardiac gene editing produces heterogeneous transduction — cells in lower-penetration regions remain unedited. In structural cardiac conditions (DMD, ARVC), >70% phenotypic correction is required to restore function. Below this threshold:
  • Heterogeneous electrophysiological properties between edited/unedited cells → arrhythmia risk
  • Maladaptation in partially corrected tissue → compensatory fibrosis
  • Partially edited cells may express neoantigens → enhanced immunogenicity
  • In DMD and ARVC specifically, sub-threshold mosaicism has caused myofiber necrosis and fibrosis (sources/gene-editing-cv-tcm-2025)
• Episomal dilution in paediatric patients: AAV remains primarily episomal. In growing paediatric patients, cardiomyocyte division dilutes episomal transgene copies over time — durability of childhood-administered therapy is uncertain, and re-treatment is blocked by pre-existing immunity. (sources/gene-therapy-arrhythmia-2025)
• Integration risk: Rare genomic integration events have been reported despite AAV's predominantly episomal persistence — theoretical insertional mutagenesis risk requiring long-term follow-up. (sources/gene-therapy-arrhythmia-2025)
• Human translation gap: AAV9 cardiac tropism is established in mice but human myocardium transduction efficiency is not explicitly confirmed. Human anatomical scale, receptor expression variation, active disease states, and antibody prevalence are all proposed to reduce efficacy. Sex, age, delivery route, and dose independently influence AAV performance — all require characterisation before human application. (sources/gene-editing-cv-tcm-2025)

Preclinical Efficacy Data

• Meta-analysis (26 studies, search to Jan 2024): AAV gene therapy reduced AF inducibility by 81% (OR 0.19, 95% CI 0.08–0.45; I²=0%; p<0.01) and combined VA inducibility by 94% (OR 0.06, 95% CI 0.03–0.11; I²=27.3%; p<0.01). Inherited VA: 96% reduction; acquired VA: 89%. Low heterogeneity in the AF analysis (I²=0%) suggests consistent effect across diverse molecular targets. (sources/aav-gene-therapy-arrhythmia-hr-2024, rating: medium)
• Publication bias: No study in the 2024 systematic review reported a failure of AAV-mediated gene therapy. Funnel plots trend symmetric but n=26 is insufficient to exclude bias definitively. The consistently positive ORs almost certainly overestimate true effect size. (sources/aav-gene-therapy-arrhythmia-hr-2024)
• LQT3 base editing — 20% correction sufficient: AAV9 dual-vector adenine base editing of SCN5A p.T1307M at ~20% editing efficiency prevented fatal arrhythmias in mice — substantially below the >70% structural correction threshold. Mechanism: electrotonic coupling spreads electrophysiological benefit from edited cells to adjacent unedited cells, a population-level effect not available for structural protein repair. (sources/aav-gene-therapy-arrhythmia-hr-2024)

Gene Targets by Disease

• AF targets (7, all preclinical): SERCA2a (Ca²⁺ cycling, intrapericardial delivery, rabbit), TASK-1 (K⁺ channel silencing, pig — strongest translational model), NLRP3 (inflammasome knockdown, cardiomyocyte-specific), miR-27b (Smad-2/3 anti-fibrotic pathway), IGF1 (atrial fibrosis inhibition), Myl4 (familial AF gene replacement), SIRT3 (alcohol-AF pathway via SIRT3-AMPK and mitochondrial dynamics). All address AF prevention — treatment of established persistent/permanent AF has not been tested. (sources/aav-gene-therapy-arrhythmia-hr-2024)
• Acquired VA targets (4, all preclinical): TBX5 (transcription factor restoration post-DCM/ischaemia); dystrophin activation via CRISPR/dCas9 (restored Nav1.5 membrane localisation → normalised conduction); adiponectin overexpression in left stellate ganglion (inhibited post-MI neural activity — potential non-surgical sympathectomy alternative); SERCA2a (post-MI electrophysiological improvement, AAV1, pig). (sources/aav-gene-therapy-arrhythmia-hr-2024)
• Inherited VA targets (15 studies, CPVT/ACM/PRKAG2/BrS/LQT3):
  • CPVT: CASQ2 replacement (single injection → curative ≥1 year); triadin replacement; allele-specific RYR2 silencing; CRISPR-SaCas9 RYR2-R4496C; engineered CaM protein; CaMKII inhibitory peptide (note: 10% QTc prolongation under β-agonist stimulation — proarrhythmic signal).
  • ACM: PKP2 replacement (only arrhythmia target in clinical trial); PLN-R14del CRISPR; GJA1-20k (Cx43 trafficking); BAG5.
  • PRKAG2 syndrome: CRISPR correction at postnatal day 4 and 42 → 2× reduction in ventricular preexcitation.
  • BrS: MOG1 (Nav1.5 chaperone) — pathway workaround for SCN5A size constraint.
  • LQT3: SCN5A p.T1307M adenine base editing — 20% efficiency sufficient (see above). (sources/aav-gene-therapy-arrhythmia-hr-2024)

Clinical Translation

• Only PKP2 has reached clinical trial for arrhythmia (as of Jan 2024): Among 22 identified molecular targets, PKP2 gene replacement (ACM — LEXEO/Rocket trial) is the sole arrhythmia target in a clinical trial. SERCA2a entered clinical trial (AskBio Phase 1) for heart failure — not arrhythmia. All remaining 20 targets are preclinical. (sources/aav-gene-therapy-arrhythmia-hr-2024)
• Future directions: Next-generation cardiotropic capsids with enhanced cardiac specificity; antibody-oligonucleotide conjugates; image-guided and magnetically guided delivery systems; extracellular vesicle-shielded vectors to evade pre-existing immunity. (sources/gene-therapy-arrhythmia-2025)

Danon Disease — Longest Published AAV Cardiac Follow-Up (RP-A501)

• RP-A501 (rAAV9-LAMP2B) in Danon disease (n=7 males; 24–54 months; NEJM 2025): Single IV infusion at low-dose (6.7×10¹³ gc/kg) or high-dose (1.1×10¹⁴ gc/kg); 3-drug immunomodulatory regimen required (prednisone + calcineurin inhibitor + rituximab); 7/7 alive at 4.5 years; LAMP2B protein confirmed in endomyocardial biopsy (6/6 at 12 months; 5/6 at 24–36 months). (sources/aav9-danon-nejm-2025, rating: high)
• Efficacy in 6 evaluable patients (LVEF ≥40% at baseline): LVM index median −23%; troponin I median −84%; natriuretic peptides median −57%; NYHA and KCCQ-12 improved ≥5 points in all 6; adult patients now aged 21–24 years — beyond the natural Danon disease transplant/death age of 19–21 years. (sources/aav9-danon-nejm-2025, rating: high)
• Durability of expression confirmed: Vector DNA and RNA largely sustained at 4.5 years — consistent with non-dividing adult cardiomyocytes preventing episomal dilution. This directly addresses the paediatric dilution concern: in children and adults with non-dividing cardiomyocytes, episomal AAV is stable; the dilution problem is specific to rapidly dividing infant cardiomyocytes (as in infantile-onset Pompe disease). (sources/aav9-danon-nejm-2025, rating: high)
• Grade 4 complement-mediated TMA — the dominant safety signal: Occurred in 1 adult patient receiving high-dose RP-A501 with LVEF 32% at baseline; resulted in thrombocytopenia, AKI requiring RRT, and ultimately cardiac transplantation at 5 months. Complement activation (sC5b-9) was more robust in adult than pediatric patients; mechanism is anti-capsid antibody-driven complement activation. LVEF <40% now an exclusion criterion; low-dose standardised for Phase 2. (sources/aav9-danon-nejm-2025, rating: high)
• Immunomodulatory regimen required — and causes significant AEs: Unlike the Pompe GC301 trial (prophylactic prednisolone only), RP-A501 required a 3-drug immunosuppressive regimen. This caused grade 3 glucocorticoid-induced skeletal myopathy in 3 patients (exacerbation of underlying Danon disease) and salmonella sepsis in 1. Immunosuppression is a critical component but carries its own morbidity — a major practical challenge in a population with pre-existing neuromuscular disease. (sources/aav9-danon-nejm-2025, rating: high)
• Anti-LAMP2B antibodies detected (no clinical sequelae): Binding antibodies to LAMP2B developed in several patients; no LAMP2B-specific T-cell response. This contrasts with GC301 (Pompe) where no anti-GAA antibodies emerged. The mechanism of differential transgene immunogenicity across disease targets is unresolved. (sources/aav9-danon-nejm-2025, rating: high)

AAV-Based GET Clinical Fatalities in DMD — 2026 ACC Summary

• DMD — innate immune response fatality (clinical trial): A 27-year-old with advanced DMD died 8 days after receiving AAV-based CRISPR epigenome editing therapy; cause was an innate immune response to the AAV vector that aggravated his baseline poor state of health from advanced disease. (sources/gene-editing-acc-2026 — very high)
• First two boys treated with AAV-CRISPR-Cas12: In a separate DMD clinical trial, the first two boys showed minimal evidence of long-term dystrophin rescue — despite initial expression. (sources/gene-editing-acc-2026 — very high)
• ACC 2026 conclusion: These events highlight the challenges of clinical translation of AAV-based GETs for myocardial and vascular diseases and suggest that ongoing efforts to develop nonviral methods for transient delivery into these tissues will be more fruitful. (sources/gene-editing-acc-2026 — very high)
• Strategic implication (ACC 2026): Liver-targeted diseases use LNP (re-dosable; lower immunogenicity; no viral genome; GalNAc ensures hepatic specificity); cardiac-targeted diseases require AAV but must contend with all AAV limitations. LNP modifications for non-liver targets are an active research frontier. (sources/gene-editing-acc-2026 — very high)

Pompe Disease — First Human IV AAV9 Clinical Data (GC301)

• GC301 (rAAV9-coGAA) in infantile-onset Pompe disease (n=4 infants; NEJM 2025): Single IV infusion at 1.2 × 10¹⁴ vg/kg + prophylactic prednisolone; 3/4 patients survived 52 weeks with improved cardiac (LVM index, LVEF) and psychomotor outcomes; 2 patients achieved normal HINE scores by 12 months of age; Patients 3 and 4 walked without assistance by 22–25 months. (sources/aav9-pompe-nejm-2025, rating: high)
• Key immune finding — no anti-GAA antibodies detected: Unlike enzyme-replacement therapy (which uniformly induces anti-rhGAA IgG in both CRIM-positive and CRIM-negative patients), no anti-GAA binding antibodies and no GAA-specific cellular responses were detected in any patient throughout the 52-week period. This negative immune response to the transgenic product is reported with other AAV-based therapies; mechanism likely involves hepatic expression promoting regulatory tolerance but remains incompletely understood. (sources/aav9-pompe-nejm-2025, rating: high)
• AAV9 crosses the blood–brain barrier: Preclinical and now clinical evidence (normalized HINE scores) that systemic IV AAV9 achieves CNS transduction — unlike rhGAA, which cannot cross the BBB and leaves CNS glycogen accumulation untreated (causing white matter abnormalities from ~2 years and progressive cognitive decline in ERT-treated survivors). (sources/aav9-pompe-nejm-2025, rating: high)
• Respiratory infection impairs gene therapy efficacy: GAA activity declined to near-baseline during intercurrent pneumonia in Patient 1; LVM index increased transiently during infections. Pneumonia occurred in all 4 patients. AAV9 does not transduce alveolar cells — respiratory tract vulnerability is not corrected by the vector. (sources/aav9-pompe-nejm-2025, rating: high)
• Paediatric re-dosing problem applies: Anti-AAV9 IgG developed and persisted in all 4 patients post-infusion. Episomal transgene dilution during infant growth combined with pre-existing immunity blocking re-dosing creates the same therapeutic catch-22 described for paediatric channelopathies — critical for IOPD where treatment begins in the first 6 months of life. (sources/aav9-pompe-nejm-2025, rating: high)

Contradictions / Open Questions

• IV vs. intracoronary route — no head-to-head data: Intravenous delivery offers the lowest procedural risk but produces significant off-target liver transduction and lower cardiac uptake. Intracoronary delivery improves cardiac specificity but requires cath lab access and is less effective with coronary disease. No head-to-head RCT defines the optimal route for any cardiac gene therapy indication; route choice is extrapolated from animal studies and manufacturing constraints. (sources/gene-therapy-arrhythmia-2025)
• Re-dosing prevented by immune responses: Anti-AAV neutralising antibodies preclude re-dosing with the same serotype. For chronic cardiac diseases (HCM, LMNA cardiomyopathy) requiring durability beyond 5–10 years, the single-dose constraint is a major unresolved barrier. Capsid re-engineering and extracellular vesicle shielding are proposed but unvalidated in humans. (sources/gene-therapy-arrhythmia-2025)
• Episomal dilution vs. re-dosing impossibility in paediatrics: Childhood-administered AAV therapy will be progressively diluted as cardiomyocytes expand, yet pre-existing immunity prevents re-treatment. This creates a therapeutic catch-22 specific to paediatric channelopathies. (sources/gene-therapy-arrhythmia-2025)
• Publication bias inflates preclinical effect sizes: No study in the 2024 systematic review reported failure. The OR 0.06 for combined VA inducibility almost certainly overestimates clinical translation readiness. This makes it difficult to identify which of the 22 targets are genuinely promising versus artefactually positive. (sources/aav-gene-therapy-arrhythmia-hr-2024, rating: medium)
• Editing efficiency threshold discrepancy — 20% vs. >70%: Base editing of SCN5A p.T1307M at 20% efficiency prevented fatal LQT3 arrhythmias (Qi 2024), conflicting with the >70% structural correction threshold in DMD and ARVC. The most likely explanation is disease-specific: electrophysiological rescue is a population-level emergent property (electrotonic coupling spreads benefit to neighbouring unedited cells), whereas structural repair requires per-cell correction. This distinction has direct implications for which diseases are tractable with lower editing efficiency. (sources/aav-gene-therapy-arrhythmia-hr-2024, sources/gene-editing-cv-tcm-2025)
• Anti-transgene immune tolerance in AAV-Pompe disease vs. anti-rhGAA in ERT: GC301 (AAV9-GAA) induced no anti-GAA antibodies in any patient, while ERT universally induces anti-rhGAA IgG regardless of CRIM status. This discrepancy may reflect hepatic cross-presentation inducing regulatory T-cell tolerance for the AAV-expressed transgene. (sources/aav9-pompe-nejm-2025, rating: high)
• Differential transgene immunogenicity — Danon vs. Pompe: RP-A501 (AAV9-LAMP2B) induced anti-LAMP2B binding antibodies in several patients, while GC301 (AAV9-GAA) induced no anti-GAA antibodies in any patient, despite both delivering therapeutic proteins via systemic IV AAV9. The immunological divergence is unexplained — potentially related to differences in hepatic expression levels, protein structure, or endogenous protein absence (Danon males have zero LAMP2 vs. Pompe patients who may have residual GAA depending on CRIM status). This contrast has implications for how immunomodulatory regimens are designed for future AAV cardiac gene therapies. (sources/aav9-danon-nejm-2025, sources/aav9-pompe-nejm-2025, rating: high)
• Complement-mediated TMA — adults more than children, dose-dependent: In the Danon RP-A501 trial, adult/adolescent patients showed more robust complement activation than pediatric patients; TMA occurred only in the high-dose adult with low LVEF. The pediatric patients had no serious treatment-related AEs. Whether this reflects immune maturity, dose, LVEF, or residual anti-AAV9 antibody load requires prospective characterisation. (sources/aav9-danon-nejm-2025, rating: high)

Connections

• Related to concepts/SupRep-Therapy
• Related to concepts/Gene-Silencing-Therapy
• Related to concepts/CRISPR-Cas9-in-Channelopathies
• Related to concepts/Gene-Editing-Risk-Benefit-Framework
• Related to concepts/Epigenetics-Cardiac-Arrhythmia
• Related to concepts/Pompe-Disease
• Related to concepts/Danon-Disease
• Related to entities/CPVT
• Related to entities/Long-QT-Syndrome
• Related to entities/Brugada-Syndrome
• Related to entities/ARVC
• Related to entities/RYR2
• Related to entities/SCN5A
• Related to entities/Genecradle-Therapeutics

Sources

• sources/gene-editing-acc-2026
• sources/gene-therapy-arrhythmia-2025
• sources/gene-editing-cv-tcm-2025
• sources/aav-gene-therapy-arrhythmia-hr-2024
• sources/aav9-pompe-nejm-2025
• sources/aav9-danon-nejm-2025