top of page

Cardiac Gene Therapy Delivery: How AAV Targeting Works

19 hours ago
11 min read

When I read a cardiac gene therapy delivery study, I want to know where in the heart the increased expression occurs.


A signal concentrated around a few injection sites has a different therapeutic meaning from expression across a broad area of diseased myocardium. Ischemia and fibrosis add further complications by changing perfusion and tissue structure. Choosing the therapeutic gene leaves a substantial delivery problem: which cells will encounter the vector, and will the effect reach enough of the diseased tissue?


Cardiac gene therapy delivery concept with a stylized heart, delivery particles and layered barriers representing access, retention and coverage
Abstract concept of cardiac delivery barriers; not measured distribution or efficacy. Concept illustration | Not a paper figure. LuTra Studio

A review published in npj Biomedical Innovations on November 2, 2026 brought these delivery questions together. It prompted me to look at administration alongside vector design, and at why hydrogels, microneedles and patches have become part of this research. Review


What goes inside an AAV vector?


AAV stands for adeno-associated virus, a different virus from adenovirus. In a recombinant AAV vector, a protein capsid surrounds an engineered DNA genome. The capsid participates in delivery; the DNA carries the intended expression instructions. Technical guide


A typical AAV capsid contains 60 subunits comprising the related proteins VP1, VP2 and VP3, arranged in an icosahedral shell. Individual particles need not have identical VP proportions. Beyond containing the genome, capsid features participate in intracellular trafficking and nuclear localization, so an engineered capsid can affect events after uptake as well as initial cellular interactions. Primary structural study


The GOI and expression cassette


The gene of interest (GOI) is the introduced gene we want the cell to express. In AAV9-VEGF, AAV9 identifies the capsid, while the DNA inside encodes vascular endothelial growth factor (VEGF). The receiving cell makes VEGF after reading that DNA.


The vector genome is bounded by ITRs; the expression cassette lies within them and contains the promoter, GOI and sequences needed for RNA processing. ITRs are cis-acting replication and packaging elements, distinct from transcriptional regulatory sequences. A simplified protein-expression vector architecture is:


5′ ITR → promoter/regulatory region → GOI coding sequence → poly(A) signal → 3′ ITR


Element

Role

ITRs

Flanking sequences involved in genome replication and packaging during production

Promoter

Regulates transcription in the receiving cell

GOI

Specifies the intended protein; only part of the vector DNA

Poly(A) signal

Supports RNA 3′-end processing; distinct from a preformed mRNA poly(A) tail


Other designs add enhancers, introns, UTRs or additional regulatory elements. In a conventional rAAV design, the expression cassette replaces the viral rep and cap regions between the ITRs; production functions are supplied separately. Capsid protein on the particle therefore does not imply that the intended DNA cargo contains the cap gene. The approximately 4.7-kb conventional single-stranded vector genome budget must accommodate the regulatory elements and ITRs as well as the GOI. Vector design reference


AAV protein capsid containing single-stranded DNA beside a simplified vector genome ordered as 5′ ITR, promoter, GOI, poly(A) and 3′ ITR
Conceptual AAV capsid and vector DNA: ITRs flank the vector genome, with the expression cassette between them. Single-stranded example; dimensions, protein arrangement and sequence lengths are not to scale. LuTra Studio

Packaging happens before administration


The production system supplies the vector genome, capsid proteins and supporting functions needed to package DNA. By the time the vector is administered, capsid and cargo have already been paired. Each can be optimized separately, but their performance has to be established together.


AAV can also carry DNA encoding functional RNAs or gene-editing components. The examples here focus on protein expression; using AAV does not, by itself, mean that genome editing is taking place.


The distance between delivered DNA and active protein


The sequence is roughly cellular contact and uptake → intracellular trafficking and endosomal escape → nuclear delivery and uncoating → transcription-competent DNA → RNA → protein. Actual routes and efficiencies depend on the capsid and cellular context.


Six-stage protein-coding ssAAV pathway: uptake, endosomal escape, nuclear delivery and uncoating, double-stranded DNA formation, transcription and RNA processing, then cytoplasmic translation
Simplified protein-coding ssAAV pathway. Complementary-strand synthesis illustrates one route to a double-stranded template; complementary-genome annealing is another. Transcription is nuclear and translation cytoplasmic; no fixed efficiency or chromosomal integration is implied. LuTra Studio

Following cell-surface interactions and endocytic uptake, vector must navigate intracellular compartments, escape the endosomal pathway and deliver its genome to the nucleus. AAV trafficking experiments established that post-entry processing can limit transduction. Primary study


Uncoating makes the genome available for subsequent processing. A 2026 AAV2 imaging study in HeLa cells found different trafficking profiles among tested particles and variants. These cell experiments help explain trafficking mechanisms; efficiency and timing in human cardiomyocytes still need to be established. Primary study


Conventional single-stranded AAV (ssAAV) must form a double-stranded transcription template. This can involve host synthesis of a complementary strand or annealing of complementary plus- and minus-strand vector genomes delivered by separate particles. Their relative contributions depend on the tissue and experimental context; mouse liver experiments demonstrated an important contribution from complementary genomes. Delivered DNA therefore cannot be converted directly into an expected expression level. Primary template-formation study


Self-complementary AAV (scAAV) packages DNA that can pair with itself, reducing dependence on complementary-strand synthesis while approximately halving unique-sequence capacity. Uptake, trafficking and nuclear delivery remain necessary. Foundational scAAV study


The cell then transcribes and processes RNA, exports mature mRNA and translates it on ribosomes in the cytoplasm. For AAV9-VEGF, the next measurements are active VEGF and its effects in tissue. Detecting vector DNA alone leaves those questions open.


Where the DNA persists


rAAV DNA can persist outside chromosomes as episomal DNA, but integration can also occur. Primate studies have assessed both forms, and later work suggests integrated genomes may contribute to long-term expression in particular liver settings. Neither complete exclusion of integration nor permanent expression should be assumed. Persistence and integration study, Long-term primate liver study


These liver findings require separate validation in cardiac tissue. Long-term expression, permanent genetic correction and durable clinical benefit each call for their own evidence.


What does targeting actually change?


With an AAV targeting claim, I first look at what was measured: vector distribution or gene expression. Capsids influence cellular interactions and transduction, while promoters influence transcription after entry. An early mouse study combined altered capsids with a cardiac-associated promoter to work on both. Primary study


Researchers compare natural capsids or screen engineered variant libraries for performance in target cells and tissues. Tissue tropism describes a preference under the tested conditions, rather than absolute organ specificity. AAV9, for example, transduces adult mouse heart but also liver. Cardiac capsid screening, AAV9 expression study


Coronary delivery, local injection and epicardial materials change where and how long tissues encounter vector. Anatomical proximity does not establish cardiomyocyte-specific recognition. The hydrogels and patches discussed later primarily address this exposure layer.


A cardiac-associated promoter acts after delivery, helping control which cells transcribe the cargo. Particles may still encounter other organs, so expression in those tissues needs to be measured. Primary study


Using miRNA to reduce off-target expression


Adding recognition sites for an endogenous microRNA can reduce cargo expression in cells containing that miRNA. An AAV9 study used miR-122 recognition sites to suppress liver expression while preserving cardiac expression in adult mice. This is expression detargeting; it does not establish that vector particles avoided the liver or that hepatic risk was eliminated. Primary study


The miRNA itself can vary with biological state. A 2020 study found variable cardiac miR-122 levels across mouse strains and human cardiomyopathy specimens, with suppression of intended cardiac cargo expression in some settings. That makes the miRNA state of the intended cells part of the design problem. Primary study


Design layer

Main question

Inference to avoid

Capsid selection and engineering

Which cells support useful transduction?

Cardiac preference means heart-only delivery

Administration and materials

Where and for how long does exposure occur?

Local exposure establishes cell specificity

Promoter

Where is transcription supported?

Lower expression means less vector exposure

miRNA regulation

Where is cargo expression suppressed?

Lower signal establishes zero exposure or risk


Vector DNA, cargo RNA, protein and function should be measured separately, with cell identity resolved. Vector DNA detected by quantitative PCR (qPCR) in a whole-heart homogenate does not establish productive cardiomyocyte transduction or the fraction of cells transduced. A heart-to-liver ratio also needs organ-level absolute measurements to distinguish increased target signal from reduced off-target signal.


For a related look at cellular uptake and productive RNA delivery, see LNP Protein Corona: How Blood Proteins Reshape RNA Delivery.


Model choice can also change the ranking of candidate capsids. In a 2023 study, capsids selected in human cardiomyocytes showed promise in human models but performed relatively poorly in mouse heart. Cell entry also did not consistently predict expression. To judge how far such a result might translate, we need to know the species, cellular maturity, route and assay endpoint. Primary study


Cardiac gene therapy delivery starts with the treatment area


Access, retention and coverage describe different parts of delivery: reaching the intended cells, maintaining useful exposure, and covering the tissue relevant to the disease. A whole-heart signal can hide important differences among them.


Vector retention should also be measured separately from transgene persistence. A vector retained in a material may remain available for release, or may be trapped. Expression after successful transduction describes a different process. Those distinctions should guide the assays chosen to evaluate a platform.


The required distribution should follow the disease and mechanism. A localized infarct and a diffuse cardiomyopathy create different coverage questions. Development teams should specify the target cells and regions before deciding whether a concentrated signal or a broader distribution constitutes success. This helps explain why intravenous, coronary, intramyocardial and epicardial approaches have different uses and limitations. Vector in the circulation still has tissue and cellular barriers to cross; vector released at the surface needs to penetrate far enough into the myocardium. Route comparison


Route

Priority development question

Intravenous

How much useful cardiac delivery is achieved at the selected dose, alongside exposure elsewhere?

Coronary circulation

Does regional perfusion produce the needed coverage at an acceptable procedural risk?

Intramyocardial injection

How far does functional delivery extend beyond each injection site?

Epicardial patch

Do surface contact, release direction and myocardial penetration match the lesion?

Intrapericardial administration

Do cavity distribution, clearance and transport across the epicardium support the intended coverage?


Epicardial patches and intrapericardial administration require separate distribution and procedural assessments. A dose increase should be judged by the additional functional delivery it produces alongside off-target exposure.


Immune barriers influence effective delivery


Pre-existing neutralizing antibodies can interfere with transduction, while capsid-associated cellular immunity can affect transduced cells and expression persistence. Healthy-population serum studies found differences in antibodies and neutralizing activity across AAV types; an early liver-directed hemophilia trial also suggested that capsid-related T-cell responses could limit expression. These are serum and hepatic gene-therapy findings, rather than cardiac efficacy estimates, but they establish why capsid, route, dose and immune assessment must be considered together. Neutralizing-factor study, Human immune-response study


What hydrogels and patches offer local delivery


Hydrogels have material effects of their own


A 2023 Biomaterials study combined AAV9-VEGF, polyaniline nanorods and alginate in an injectable hydrogel, evaluated in rats with myocardial infarction. The investigators reported angiogenesis and cardiac functional findings in a system that also had conductive and antioxidant properties. Primary study


For studies like this, I pay close attention to the blank-material, free-vector and combination controls. The material may alter release and may also protect tissue independently. Release measurements, vector activity after release and cardiac function need to be read together before deciding which component to improve.


Microneedles redistribute the dosing sites


A 2020 Science Advances study used phase-transition microneedles to deliver AAV in rats. The investigators reported more even local expression than direct injection, and angiogenic and functional findings with VEGF as the cargo. Primary study


Microneedles spread dosing across an area of the myocardial surface. Moving from rats to humans brings questions about wall thickness, lesion size, contact during cardiac motion and the response to placement. It also requires a procedure to access the heart surface. The reported effects remain preclinical.


The cargo changes what a scaffold study can tell us


In 2025, Matsumura and colleagues incorporated VEGF-encoding AAV into an electrospun biodegradable elastomeric scaffold. Patches were placed three days after myocardial infarction in rats, with eight-week follow-up. The VEGF-AAV patch group showed higher ejection fraction and smaller end-diastolic area than control-patch or sham groups. This is an elastomeric scaffold platform, with a different structure and administration method from injectable hydrogels. Primary study


An earlier 2017 study used an elastomeric AAV patch encoding green fluorescent protein (GFP) to investigate sustained local delivery, while also assessing cardiac function and remodeling. GFP is a reporter; functional changes in that experiment should not be interpreted as therapeutic activity of GFP. Primary study


GFP helps us observe delivery; VEGF introduces a therapeutic mechanism to evaluate. Material controls are also needed to distinguish the mechanical support provided by the scaffold.


CUPID 2 brings the discussion back to patients


SERCA2a is a cardiac isoform of sarcoplasmic/endoplasmic reticulum calcium ATPase involved in cellular calcium reuptake; the gene-transfer strategy aimed to improve calcium handling. CUPID 2 randomized 250 patients with heart failure and reduced ejection fraction to intracoronary AAV1/SERCA2a or placebo in a double-blind phase 2b trial; 243 entered the modified intention-to-treat analysis. The primary endpoint concerned recurrent heart-failure events: hospitalization or ambulatory treatment for worsening heart failure. The tested regimen did not improve this endpoint (hazard ratio 0.93; 95% CI 0.53–1.65; p = 0.81). Trial publication


CUPID 2 shows how much remains to be established between a therapeutic mechanism and patient benefit. Its result cannot assign the failure solely to delivery. A subsequent program still needs to connect the selected target and functional vector delivery to the clinical outcome being tested.


A product has to survive preparation and administration


For product development, I would next ask how much usable vector remains after handling. The amount loaded, released or left in a device, together with activity after release, determines what cells actually encounter. Vector genomes (vg), total capsids and functional transducing dose describe different aspects of that dose.


FDA guidance on gene-therapy IND CMC information calls for support for product identity, quality, purity and strength, including potency. Specific assays for a combination delivery platform still require product- and risk-based development; the proposed measurements above are not a universal regulatory release panel. FDA CMC guidance


Storage, transport, preparation and device compatibility affect whether that dose can be reproduced. For patches and other locally administered platforms, operator variability and procedural burden are part of development as well.


Hydrogels and patches interest me because they offer ways to adjust local exposure. Before advancing a platform, though, I want to see how much effect reaches the intended cells, how much of the lesion it covers, and whether it survives the practical steps of administration. An increase in cardiac expression becomes useful for development when those measurements accompany it.


References


  1. Kabboul G, et al. Challenges and innovations in cardiac AAV gene delivery as a platform for cardiac regeneration. npj Biomedical Innovations. 2026;3:64. Source

  2. Shi H, et al. Microneedle-mediated gene delivery for the treatment of ischemic myocardial disease. Science Advances. 2020;6:eaaz3621. Source

  3. Wu C, et al. Injectable polyaniline nanorods/alginate hydrogel with AAV9-mediated VEGF overexpression for myocardial infarction treatment. Biomaterials. 2023;296:122088. Source

  4. Matsumura Y, et al. Incorporation of Adeno-Associated Virus Encoding Vascular Endothelial Growth Factor into a Biodegradable Elastomeric Scaffold for Improved Function in the Ischemic Rat Heart. ACS Biomaterials Science & Engineering. 2025;11:2226–2235. Source

  5. Gu X, et al. Sustained viral gene delivery from a micro-fibrous, elastomeric cardiac patch to the ischemic rat heart. Biomaterials. 2017;133:132–143. Source

  6. Greenberg B, et al. Calcium upregulation by percutaneous administration of gene therapy in patients with cardiac disease (CUPID 2): a randomised, multinational, double-blind, placebo-controlled, phase 2b trial. The Lancet. 2016;387:1178–1186. Source00082-9)

  7. Addgene. Adeno-Associated Viral Vector Guide. Source

  8. Müller OJ, et al. Improved cardiac gene transfer by transcriptional and transductional targeting of adeno-associated viral vectors. Cardiovascular Research. 2006. Source

  9. Geisler A, et al. microRNA122-regulated transgene expression increases specificity of cardiac gene transfer upon intravenous delivery of AAV9 vectors. Gene Therapy. 2011;18:199–209. Source

  10. Kraszewska I, et al. Variability in Cardiac miRNA-122 Level Determines Therapeutic Potential of miRNA-Regulated AAV Vectors. Molecular Therapy Methods & Clinical Development. 2020. Source

  11. Kok CY, et al. Development of new adeno-associated virus capsid variants for targeted gene delivery to human cardiomyocytes. Molecular Therapy Methods & Clinical Development. 2023;30:459–473. Source

  12. Wörner TP, et al. Adeno-associated virus capsid assembly is divergent and stochastic. Nature Communications. 2021;12:1642. Source

  13. Douar AM, et al. Intracellular Trafficking of Adeno-Associated Virus Vectors: Routing to the Late Endosomal Compartment and Proteasome Degradation. Journal of Virology. 2001;75:1824–1833. Source

  14. Hutchinson TE, et al. Intracellular trafficking and nuclear compartmentalization of AAV gene delivery. Molecular Therapy Advances. 2026;34:201770. Source

  15. McCarty DM, Monahan PE, Samulski RJ. Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis. Gene Therapy. 2001;8:1248–1254. Source

  16. Nowrouzi A, et al. Integration Frequency and Intermolecular Recombination of rAAV Vectors in Non-human Primate Skeletal Muscle and Liver. Molecular Therapy. 2012;20:1177–1186. Source

  17. Greig JA, et al. Integrated vector genomes may contribute to long-term expression in primate liver after AAV administration. Nature Biotechnology. 2024;42:1232–1242. Source

  18. Nakai H, Storm TA, Kay MA. Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo. Journal of Virology. 2000;74:9451–9463. Source

  19. Boutin S, et al. Prevalence of Serum IgG and Neutralizing Factors Against Adeno-Associated Virus (AAV) Types 1, 2, 5, 6, 8, and 9 in the Healthy Population: Implications for Gene Therapy Using AAV Vectors. Human Gene Therapy. 2010;21:704–712. Source

  20. Manno CS, et al. Successful transduction of liver in hemophilia by AAV-Factor IX and limitations imposed by the host immune response. Nature Medicine. 2006;12:342–347. Source

  21. U.S. FDA. Chemistry, Manufacturing, and Control (CMC) Information for Human Gene Therapy Investigational New Drug Applications (INDs). Final guidance, January 2020. Source

Comments


bottom of page