Dyno Therapeutics, Inc. has announced new preclinical findings from three artificial intelligence-engineered adeno-associated virus capsid programmes, demonstrating improved gene delivery to brain and muscle tissue alongside advances in avoiding pre-existing antibodies. Disclosed on October 27, 2026, the research is being presented at the 33rd Annual Congress of the European Society of Gene and Cell Therapy in Hamburg, Germany, with results including 152-fold higher brain transduction than AAV9 in non-human primates and immune-evasive capsids that tested seronegative against up to 85% of human serum samples.
The findings address several important technical challenges in gene therapy, where delivery efficiency, unwanted tissue exposure and immune responses can limit the usefulness of investigational treatments. Dyno Therapeutics is developing engineered capsids, the protein shells that transport genetic material into cells, to improve the ability of gene therapies to reach selected organs after systemic administration. Its latest research focuses on muscle-specific delivery, cross-species brain and muscle targeting, and reducing recognition by pre-existing antibodies.
However, the reported results remain preclinical. The studies establish differences in vector distribution, cellular transduction and laboratory measures of antibody recognition, but they do not demonstrate that the engineered capsids improve human treatment outcomes or reduce clinical adverse events. Their value will depend on whether these advantages can be reproduced across relevant therapeutic programmes and translated into acceptable safety and efficacy profiles.
Why are Dyno Therapeutics’ three ESGCT 2026 presentations important for the development of systemic gene therapies?
Adeno-associated viruses have become widely used delivery vehicles in gene therapy because they can transport therapeutic genetic material into different cell types. However, naturally occurring and previously engineered capsids have limitations that can restrict treatment efficiency. These include incomplete penetration of target tissues, accumulation in unwanted organs and neutralisation by antibodies already present in the patient.
Dyno Therapeutics is attempting to address these problems through AI-guided capsid engineering. Its approach combines computational design with high-throughput experimental testing, allowing researchers to evaluate large numbers of potential vector sequences against specific biological objectives. The latest presentations examine whether this process can produce improvements across several different aspects of gene delivery.
The research is being presented through three scientific posters scheduled for October 28 and 29, 2026. The projects investigate dual-mechanism muscle-targeting capsids, a vector that engages alkaline phosphatase to reach the central nervous system and muscle, and an iterative strategy for engineering capsids that evade pre-existing antibodies. Although each programme addresses a different challenge, they share the objective of improving the efficiency and selectivity of genetic material delivery.
Adrian Veres, chief scientific officer and co-founder of Dyno Therapeutics, has identified gene delivery as a central obstacle to realising the broader clinical potential of genetic medicines. He has positioned the latest findings as progress towards more effective delivery systems for muscle and neurological diseases. Nevertheless, the studies remain at the technology-development stage, and their clinical implications require further investigation.
How did Dyno Therapeutics achieve approximately 40-fold higher skeletal muscle transduction in non-human primates?
The first presentation, poster P0116, focuses on a new generation of capsids designed to engage more than one mechanism involved in cellular entry. The programme combines integrin binding with a proprietary muscle receptor mechanism to improve delivery to skeletal and cardiac muscle. The approach is intended to increase transduction of target tissues while limiting distribution to organs that are not the intended treatment site.
According to Dyno Therapeutics, lead capsids were initially identified through pooled testing in non-human primates. Researchers subsequently evaluated selected candidates individually in both mice and non-human primates, providing additional evidence that the observed effects were not confined to the original screening experiment. This sequential testing is relevant because pooled experiments can identify promising sequences without fully characterising their individual performance.
At a dose of 5 × 10¹² vector genomes per kilogram, one lead capsid demonstrated approximately 20-fold higher skeletal muscle biodistribution than MyoAAV-4E, an engineered vector described in a 2021 scientific publication. The same comparison showed approximately 40-fold higher skeletal muscle transduction on average in non-human primates. Dyno Therapeutics also reported improved cardiac muscle tropism, suggesting potential relevance to diseases involving both skeletal and cardiac tissue.
Biodistribution and transduction are related but distinct measurements. Biodistribution describes where vector material is detected, while transduction reflects successful delivery and expression of genetic material within cells. A vector that reaches a tissue more efficiently does not necessarily generate a proportionate increase in functional therapeutic protein.
The 40-fold transduction result is therefore an important experimental finding, but its interpretation depends on the specified comparator, tissue measurements and dosing conditions. It should not be taken to mean that a future gene therapy will be 40 times more effective in patients. Therapeutic outcomes will also depend on the genetic payload, required expression levels and the biological characteristics of the disease being treated.
Could dual-mechanism AAV capsids reduce the high-dose challenges associated with muscular gene therapy?
Systemic delivery to skeletal muscle presents a demanding engineering problem because the tissue is distributed throughout the body. An effective treatment may need to deliver genetic material to a substantial proportion of muscle fibres, including tissues that are difficult to reach efficiently. Increasing vector dose can improve exposure but may also increase manufacturing requirements and the potential for unwanted biological effects.
Several clinical gene therapy programmes targeting muscular disorders have required substantial systemic vector doses. Higher exposure can increase the importance of immune responses, liver monitoring and other safety considerations. Improving delivery efficiency could therefore be valuable if it allows therapeutic expression at lower administered doses.
Dyno Therapeutics has designed its dual-mechanism capsids to improve the proportion of vector activity occurring in intended muscle tissues. The use of multiple receptor interactions may provide an advantage over delivery systems relying primarily on a single entry mechanism. However, the relationship between receptor engagement and clinical performance will vary according to tissue characteristics and disease biology.
Earlier research presented by Dyno Therapeutics at the World Muscle Society Congress in October 2026 also examined related multi-mechanism capsids. The company reported stronger muscle delivery and improved liver detargeting in mouse and non-human primate studies, supporting the broader rationale for its design strategy. These findings are consistent with the direction of the ESGCT programme but remain experimental evidence rather than clinical safety validation.
A reduction in the amount of vector required for effective treatment could have commercial as well as clinical implications. Gene therapy manufacturing is technically demanding, and lower doses may reduce the amount of vector material needed for each treatment. Nevertheless, any manufacturing advantage would depend on whether the capsids can be produced efficiently and maintain their performance when combined with therapeutic payloads.
The most important future comparison will involve therapeutically relevant doses and functional outcomes. Showing increased transduction at a low experimental dose is useful, but clinical development must establish how much expression is required to produce benefit and whether the associated exposure is acceptable.
What does the 152-fold increase in brain transduction reveal about Dyno Therapeutics’ ALPL-targeting AAV platform?
The second presentation, poster P0515, examines an AI-designed capsid targeting alkaline phosphatase, known as ALPL, as a mechanism for reaching the central nervous system. Dyno Therapeutics reported substantial improvements in brain delivery following systemic administration in non-human primates. The approach is intended to address the difficulty of transporting genetic material across the blood-brain barrier.
The blood-brain barrier protects neural tissue by restricting the passage of many substances from the bloodstream. This protective function also complicates treatment of inherited neurological disorders because systemic medicines and delivery vehicles may not reach the brain efficiently. A capsid capable of crossing this barrier could expand the range of diseases that can be investigated through intravenous gene delivery.
Dyno Therapeutics reported that its ALPL-targeting vector achieved 152-fold higher brain transduction than AAV9 in non-human primates. The company also observed up to 50% neuronal transduction in the assessed tissue regions, indicating that the engineered capsid could reach a substantial proportion of certain neuronal populations under the experimental conditions. These findings suggest an improvement in central nervous system delivery relative to the specified conventional comparator.
However, the reported percentage should not be interpreted as uniform delivery to half of all neurons throughout the brain. Transduction can vary between anatomical regions, neuronal populations and experimental conditions. A clinically useful neurological therapy may require delivery to specific cell types or brain regions rather than broad distribution alone.
The 152-fold comparison also reflects a particular experimental framework. It does not establish that all therapeutic payloads will exhibit the same increase in expression or that the improved distribution will generate proportional clinical benefit. Further studies will need to examine the interaction between the capsid, its genetic cargo and the biological characteristics of intended treatment targets.
Why is Dyno Therapeutics combining central nervous system delivery with muscle-targeting capabilities?
An additional feature of the ALPL-targeting capsid is its reported ability to reach skeletal muscle as well as the central nervous system. Dyno Therapeutics disclosed that the vector transduced approximately 55% of quadriceps muscle fibres in non-human primates. The external comparator used for that muscle assessment reached approximately 5% of the measured fibres.
The result raises the possibility of developing delivery systems capable of reaching more than one therapeutically relevant tissue. Some inherited disorders involve both neurological and muscular manifestations, creating a rationale for investigating vectors with combined tissue targeting. However, the usefulness of such an approach would depend on the particular disease and the expression requirements of its therapeutic payload.
Dyno Therapeutics also reported sevenfold liver detargeting compared with AAV9 in the ALPL-targeting study. The finding is important because unwanted vector accumulation in the liver can contribute to off-target exposure and complicate systemic gene therapy development. Reducing this distribution could potentially improve the relationship between exposure in intended and unintended tissues.
Nevertheless, lower liver biodistribution is not equivalent to demonstrated freedom from liver toxicity. Systemic adeno-associated virus therapies can produce complex immune and inflammatory responses, and the risk profile depends on several factors beyond the proportion of vector reaching a particular organ. The experimental finding therefore supports further safety investigation without establishing a clinical safety advantage.
The ability to reproduce delivery performance across animal species is another important consideration. Dyno Therapeutics has designed the capsid to engage a biological mechanism relevant across species, potentially reducing some limitations associated with vectors that perform well in mice but poorly in primates. Even so, receptor expression and other aspects of vector biology can differ between non-human primates and people, making direct human translation uncertain.
How could antibody-evasive AAV capsids expand the population eligible for future gene therapies?
The third ESGCT presentation, poster P0148, addresses pre-existing immunity against adeno-associated virus vectors. Some individuals have antibodies generated through previous natural exposure to related viruses, and these antibodies may interfere with the delivery of gene therapy vectors. As a result, antibody screening can affect eligibility for particular treatments.
Dyno Therapeutics used a three-round AI-guided engineering strategy that included mapping antibody-recognition regions across the capsid surface. The objective was to modify vector sequences to reduce antibody recognition while preserving the properties required for efficient delivery to target tissues. This creates a design challenge because structural changes that reduce immune recognition can also interfere with receptor binding or other aspects of vector function.
According to the company, newly engineered capsids tested seronegative against up to 85% of the human serum samples evaluated, compared with 49% for AAV9 under the reported testing conditions. Dyno Therapeutics also reported preservation of central nervous system and muscle receptor engagement. These findings suggest that some engineered variants may be less susceptible to recognition by pre-existing antibodies.
However, seronegativity in a laboratory assay does not establish that the same proportion of patients would qualify for treatment. The result depends on the serum panel, assay methodology, antibody threshold and specific capsid being evaluated. The announcement does not provide enough detail to determine how consistently the finding would apply across different human populations.
The distinction between antibody binding and functional neutralisation is also relevant. Antibodies detected in laboratory assays may differ in their capacity to prevent vector transduction, while immune responses involve mechanisms beyond circulating antibodies alone. Clinical eligibility would depend on the validated testing approach and the requirements of the particular investigational or approved therapy.
Even if antibody-evasive capsids improve access to initial treatment, repeat dosing would remain a separate challenge. Administration of a viral vector can itself generate an immune response, potentially affecting subsequent exposure. The reported preclinical results do not establish that repeat administration is feasible or safe.
What safety challenges must Dyno Therapeutics address before improved AAV delivery can translate into human treatments?
The development of more efficient AAV capsids has potential implications for safety, but improved distribution does not automatically produce a safer therapeutic product. The United States Food and Drug Administration has identified immune responses and neutralising antibodies as important concerns in AAV-based gene therapy. Regulators also examine liver-related complications and other systemic effects that can emerge after vector administration.
AAV-associated toxicity is influenced by several factors, including vector dose, capsid characteristics, genetic payload, patient health and immune status. Consequently, differences in laboratory biodistribution provide only part of the information needed to characterise clinical risk. Developers must also investigate how vector exposure interacts with the intended therapeutic protein and the disease being treated.
For Dyno Therapeutics, the preclinical challenge is to demonstrate that improvements in tissue targeting are consistent across relevant experimental systems. The next stage would involve appropriately designed toxicology studies and other assessments supporting the proposed clinical use of a selected vector. Safety findings from one capsid or genetic payload cannot automatically be transferred to another combination.
Manufacturing quality will also become increasingly important as promising vectors advance towards clinical applications. AAV production requires controls over vector identity, purity, potency and consistency. Changes in capsid sequence may affect production yield, stability or other characteristics that influence the feasibility of commercial manufacturing.
These considerations are particularly relevant to a platform company developing multiple capsids for use by different pharmaceutical partners. Each therapeutic programme may have distinct requirements concerning tissue targeting, dosing and manufacturing. Strong performance in a general screening platform therefore needs to be translated into product-specific evidence.
How do the new ESGCT findings strengthen Dyno Therapeutics’ partnerships with Astellas Pharma and Roche?
Dyno Therapeutics operates a platform-focused business model that includes collaborations with pharmaceutical developers seeking improved delivery technologies. Its established partners include Astellas Pharma Inc. and Roche, while the company has also worked with NVIDIA on artificial intelligence-related technology. These relationships demonstrate commercial interest in applying engineered capsids to therapeutic development programmes.
In April 2026, Astellas Pharma exercised an option to license a Dyno Therapeutics capsid for skeletal muscle-targeted gene delivery. The arrangement included a $15 million licensing fee and potential future development, regulatory and commercial milestone payments, together with royalties on qualifying products. The exercise of that option represented a commercial licensing milestone distinct from the new preclinical findings.
Dyno Therapeutics has also expanded its relationship with Roche in central nervous system gene delivery. In October 2024, the companies agreed to an expanded collaboration involving a $50 million upfront payment. Roche subsequently exercised a licence in early 2025 to develop a neurological gene therapy using a Dyno Therapeutics capsid under an earlier partnership.
These agreements provide examples of how the company can potentially monetise its technology without independently developing every therapeutic programme through regulatory approval. Pharmaceutical partners can license particular delivery systems and take responsibility for further product development. Dyno Therapeutics may receive licensing consideration, milestones or royalties depending on the relevant agreement.
However, the October 27 ESGCT announcement did not disclose a new licensing transaction or additional commercial payment. The reported data may support future partnering discussions, but they should not be treated as evidence of new revenue. Its immediate significance lies in expanding the experimental evidence available to potential technology users.
The company also makes selected capsids available through direct licensing and its Dyno Frontiers Network. This creates an additional channel for engaging genetic medicine developers, including smaller organisations that may lack extensive delivery engineering capabilities. Future commercial success will depend on whether the technology can support viable therapeutic programmes and generate repeatable licensing demand.
Could AI-guided capsid engineering reduce the time and cost of developing genetic medicines?
Traditional capsid engineering can involve extensive experimental screening to identify variants with desired biological characteristics. Different vector properties may need to be optimised separately, including tissue selectivity, transduction efficiency, immune recognition and manufacturing suitability. Improving one property can sometimes create trade-offs in another, making the design process complex.
Dyno Therapeutics is using artificial intelligence to connect experimental measurements with computational predictions of sequence performance. The approach allows researchers to design new variants based on information generated from previous rounds of testing. The company’s broader platform incorporates high-throughput in vivo experimentation to refine these predictions.
The three ESGCT programmes demonstrate applications of this strategy to different biological challenges. Dual-mechanism muscle vectors address receptor engagement, ALPL-targeting capsids investigate tissue distribution, and antibody-evasive variants focus on immune recognition. The ability to optimise multiple properties could be commercially relevant if it reduces the number of experimental cycles required to identify suitable candidates.
However, the announcement does not provide a quantified comparison of research costs or development timelines against conventional vector engineering approaches. It also does not establish that AI-designed capsids have higher clinical success rates than alternatives. Such claims would require comparative evidence extending beyond the reported preclinical findings.
The ultimate economic benefit will depend on the full development pathway, including manufacturing, toxicology, clinical testing and regulatory review. Improving the efficiency of vector discovery may reduce one source of research expenditure without eliminating other substantial costs. The value of AI-guided engineering will therefore need to be assessed through successful translation into development programmes rather than computational performance alone.
What will determine whether Dyno Therapeutics’ ESGCT 2026 capsid results lead to future clinical applications?
The first requirement will be further validation of the reported delivery and immune-evasion characteristics. The company has demonstrated promising preclinical performance, including substantial improvements over specified reference capsids, but results will need to remain consistent across additional studies. Experimental reproducibility will be particularly important when different vector batches, genetic payloads and disease models are introduced.
The next development question concerns therapeutic relevance. A capsid that achieves high neuronal or muscle transduction must still deliver a functional genetic payload at an exposure level appropriate for the intended disease. The relationship between transduction measurements and actual therapeutic benefit will vary according to the biological target and required expression threshold.
Human translation remains the largest unresolved scientific step. Non-human primate studies provide valuable information about vector delivery, but they cannot establish human tissue distribution, immunogenicity or clinical effectiveness with certainty. Regulatory development will require product-specific evidence addressing these limitations.
Commercial progress will also depend on whether pharmaceutical and biotechnology partners incorporate the new capsids into active treatment programmes. Licensing agreements, candidate selection and subsequent clinical development would provide stronger evidence of practical platform value than conference presentations alone. Manufacturing feasibility and regulatory readiness would influence which vectors are selected.
For Dyno Therapeutics, the ESGCT 2026 data expand the experimental foundation of its AI-guided gene delivery platform across several clinically relevant challenges. The reported increases in brain and muscle transduction, together with reduced antibody recognition, provide a basis for further technology development. The decisive test will be whether these engineered capsids can support human gene therapies with sufficient efficacy, acceptable safety and a commercially viable manufacturing profile.
