How Top Biotechs Use Deep Learning to Drive Development of Cell and Gene Therapy

Gene therapy is a rapidly growing field that has the potential to revolutionize the way many diseases are treated. One of the most promising gene therapy platforms is based on adeno-associated virus (AAV) vectors. However, despite their many advantages, AAV-based gene therapies also have several safety concerns, including the risk of empty capsids and resultant effects.

Empty capsids are AAV particles that do not contain genetic material and have the potential to lead to immunological responses without providing any therapeutic benefit¹.   Moreover, they may compete with filled capsids for the same cellular receptors, reducing the efficacy of the therapy by limiting the number of receptors available for the functional capsids to bind.   

The Costly Consequences of Empty Capsids in AAV Gene Therapy Development

The importance of screening AAV-based gene therapies for empty capsids can not be overstated.  In fact, in 2021, the FDA’s Cellular, Tissue, and Gene Therapies Advisory Committee (CTGTAC) shared a report and draft guidance on safety issues for AAV-based gene therapies as this is a notable concern that needs to be addressed³.   

Specifically the document reads; “The empty capsids increase overall antigenic load and potentially exacerbate capsid-triggered innate and adaptive immune responses. The empty capsids can contribute to the peptides presented by major histocompatibility complex molecules, with consequent recognition and clearance of transduced cells by capsid-specific cytotoxic T cells”.

Here is a quick breakdown of the costly consequences of empty capsids in AAV gene therapy development.

Reduced efficacy

Incomplete capsids squander available receptors, leading to a decrease in therapy efficacy and thus necessitating higher doses of the product. This exacts an increased financial burden on production costs - potentially undermining its commercial viability.

Increased costs of manufacturing

Mitigating the presence of empty capsids through purification adds to production costs, as this requires a higher dosage for therapy. The increased need for larger doses of medicine due to the abundance of empty shells can be also be incredibly costly.

Decreased commercial potential

Unfilled capsids have the potential to severely reduce a product's overall therapeutic benefit. An unsatisfactory therapeutic effect will lead patients away from that therapy and towards other options.  

Increased FDA regulatory approval hurdles

Empty capsids can elicit an immune response, which could counter the therapeutic intended application. This will force the FDA to require additional information concerning its safety and efficacy before granting approval of the gene therapy.

The Impact of Empty Capsids and Vector Dose on AAV Safety

According to the FDA's 2021 report: “In addition to stimulation of innate and adaptive immune responses, AAV empty capsids may compete with full capsids for receptor binding on target cells, which could necessitate an increase in the required vector dose”

Increasing vector dose will lead to unintended off-target effects, toxicity and immunogenicity, while also increasing total costs of production.   But by reducing the amount of empty capsids, the doses required for achieving therapeutic efficiency will be lowered, resulting in improved safety and decreased costs of gene therapy production. Therefore, in order to keep dose requirement low, it's essential to develop the most pure and potent AAV gene therapy product.

To date, the FDA has not formalized guidance to specify an upper limit for residual empty capsids in AAV-based gene therapies anddose regulations. However, the gene therapy field has already recommended that industry establish a total capsid titre as a part of FDA applications for safety at high vector doses⁴.‍

Tackling the AAV Safety and Empty Capsid Problem with Artificial Intelligence

Empty capsids are largely caused by truncation of the AAV construct upon delivery. Recent research has shown that these truncations can occur due to the secondary structure of the genome, specifically that DNA inverted repeats are hotspots for genome instability because they can form stable hairpin or cruciform structures that interfere with DNA replication.⁵ 

Advancements in artificial intelligence and machine learning have provided us with an opportunity to decrease the presence of empty capsids and generate the most pure and potent AAV-based gene therapy enabling safer gene therapy options. 

AI can play a valuable role in modeling and predicting truncation and designing constructs with a reduced propensity for truncation, improving the manufacturability of AAV-based gene therapies. More specifically, deep learning algorithms are an excellent option to model truncation propensity due to their flexible architecture, a variable receptive field for inputs, and ability to self-learn important drivers of signal from the information. Once trained, they provide an in silico mechanism for prediction and subsequent optimization of the construct design. 

‍Our artificial intelligence work at Form has shown that combining biological knowledge, machine learning, and codon optimization can reduce truncation by as much as 70% in the coding region for some transgenes⁶. This improvement and the continued advancement of predictive DL algorithms should help clear a path for better manufacturability of AAV-based gene therapies and create more viable therapeutics in the years to come.

FAQs

1. What are empty capsids in AAV-based gene therapy? Empty capsids are AAV particles that do not contain genetic material and have the potential to lead to immunological responses without providing any therapeutic benefit¹.

2. What are the consequences of having empty capsids in AAV-based gene therapy? The consequences of having empty capsids in AAV gene therapy development is that it reduces efficacy, increases costs of manufacturing, decreases commercial potential, and increases FDA regulatory approval hurdles. 

3. How is AI used in AAV construct design? AI can predict truncation and design constructs with a reduced propensity for truncation, improving the manufacturability of AAV-based gene therapies.

AI Disclosure: This content was generated with assistance from AI tools for copywriting.