Background

In the field of gene therapy, adeno-associated virus (AAV) is recognized as a widely used vector due to its safety and excellent gene delivery efficiency. Various treatments, including Zolgensma for spinal muscular atrophy, have established themselves in the market based on AAV vectors. However, the inherent broad tissue tropism of the virus has repeatedly acted as an obstacle in applying tumor-targeted therapies.

This is because most AAV particles administered via intravenous injection accumulate in liver tissue. This liver uptake not only increases the risk of toxicity to normal hepatocytes but also drastically reduces the effective dosage reaching primary or metastatic tumor sites. Researchers attempted to directly insert target ligands by genetically recombining capsid protein sequences, but this resulted in adverse effects such as reduced virus assembly yield and particle structural instability. Consequently, there was an urgent need to develop a gene delivery vehicle that can precisely target only tumors without damaging normal tissue, especially for patients with neuroblastoma (NB), a pediatric solid tumor with high recurrence and frequent metastasis.

Key Findings

Researchers found a breakthrough with 'AAV-STITCH', a platform that integrates the SpyTag system—a protein covalent bonding technology—onto the viral surface. The concept involves post-modifying the capsid proteins that form the viral shell to precisely attach desired target molecules.

To eliminate liver-binding characteristics, researchers constructed a variant called 'AAV9-W503A' by replacing the 503rd tryptophan residue of the AAV9 capsid with alanine. This variant, with its galactose receptor-binding site disrupted, completely loses the ability to infect normal hepatocytes. The researchers inserted a 13-amino acid SpyTag peptide into the virus surface-exposed loop and chemically conjugated a single-chain antibody fragment (scFv) that recognizes GD2 (disialoganglioside), an antigen on the surface of neuroblastoma, in a concentration-dependent manner. The assembled vector is named AAV-STITCHαGD2.

In pseudometastatic xenograft mouse model experiments, AAV-STITCHαGD2 successfully achieved selective infection of GD2-positive NB tissue without affecting normal liver tissue. This result completely eliminated the liver uptake side effects observed with existing wild-type vectors.

The therapeutic efficacy was also potent. When researchers loaded a suicide gene that induces cell death into AAV-STITCHαGD2 and administered it, tumor growth was inhibited and survival periods were significantly extended in subcutaneous tumor and simulated metastasis mouse models. In conditions combined with standard chemotherapy for recurrent neuroblastoma (NB), complete remission responses, where tumors disappeared without a trace, were even observed.

Significance and Outlook

This research has opened a path to freely control target tissue tropism post-assembly without complex genetic manipulation of the viral vector capsid. It is evaluated as expanding the scope of AAV-based gene therapy into the field of solid tumor treatment, which was previously difficult to administer systemically via intravenous injection due to liver toxicity.

The platform's flexibility is also notable. By simply replacing the type of scFv antibody to be conjugated while keeping the viral body intact, the target can be easily switched to other intractable cancers that express GD2, such as glioblastoma or melanoma, in addition to NB. Since the virus assembly process itself is not disrupted during production, a process design that covalently bonds target proteins while maintaining high viral titers is also advantageous for mass production.

However, hurdles to commercialization remain. The risk that SpyTag, a bacteria-derived peptide at the conjugation site, will trigger unwanted immune responses in the human immune system must be carefully evaluated. Follow-up evaluations to confirm long-term safety and in vivo circulation half-life in large animal models, which are more similar to the human biological environment than rodent models, are identified as prerequisites for entering clinical trials.