Characterizing Full, Empty, and Truncated AAV Capsids

In drug development, we’ve talked for over a decade about the “translation gap.”

Awareness about the translation gap – the fact that many promising pre-clinical candidates fail to deliver therapeutic success in the clinic – is at an all-time high, and many industry-leading companies are hard at work on the problem, developing more clinically-relevant model systems, implementing AI solutions, and more.

Yet, while one gap is beginning to close, another has opened. 

This new gap, the “accessibility gap,” has materialized alongside biopharma’s most promising, new therapeutic class, gene therapies. Accessibility goes far beyond the (relatively) simple questions about pre-clinical/clinical efficacy and safety, expanding into unique and challenging territory, such as manufacturing woes, regulatory hurdles, and IP/licensing issues, all of which ladder up to the largest contributor to inaccessibility for patients: drug pricing. Look no further than Hemgenix, labeled the most expensive drug in the world ($3.5 million per treatment) after its FDA approval in November of 2022.¹

What Do We Do With Gaps? Build Bridges.

There has been much justification for these exorbitant price tags. 

Yes, gene therapies are complex to develop and manufacture. Sure, they require new (expensive) analytical techniques for R&D, QC, and more.

But simply accepting the status quo is antithetical to the very nature of our industry. 

Delivering accessible gene therapies, such as CRISPR-based therapeutics, to patients who need them requires not only imagination but a re-imagination of how gene therapy development is currently done. 

A task force led by Jennifer Doudna is doing just that by highlighting the reasons that “...some companies have discontinued work on genetic therapies because of ‘challenging’ manufacturing and delivery problems or ‘confounding’ fundraising efforts.”² In their recent report, the task force proposes making IP from academic institutions more affordable and a mixed organizational model consisting of non-profit organizations, academic institutions, and public companies to advance CGTs into the clinic.³‍

The last component of the bridge over the accessibility gap focuses on reducing the manufacturing costs of gene therapies. The key to this, the task force proposes, is the development of platform technologies and standardized SOPs that regulatory bodies can “know and trust,” rather than “reinventing the wheel” with novel solutions for the same gene therapy manufacturing woes. High manufacturing costs have also been in the sights of Peter Marks, director of the Center for Biologics Evaluation and Research at the FDA, who announced the launch of an Operation Warp Speed-style initiative focused on rare diseases.⁴

Fixing the Cost of Biomanufacturing: A Focus on AAV-Based Gene Therapies

In addition to the biomanufacturing issues mentioned above, a major contributor to the cost of manufacturing gene therapies is the reliance on complex biological systems – viral vectors – for production. Viral vectors are used widely for the delivery of therapeutic genes in humans. 

Of all viral vector types, the most common in use for gene therapies are adeno-associated virus (AAV) (42%),  lentivirus (30%), and adenovirus (13%). Other viral vector types in use are poxi, herpes simplex, and retroviruses. 

To better understand the rise of AAV-based gene therapies, let's look at some basics of the virus and the advantages of using it for gene therapies.

AAV Fundamentals and Benefits

AAVs are the leading gene delivery system, with a strong track record of safety and efficacy, validated by several recent approvals of AAV-based therapies.

An AAV is made from just two components: a 20-sided icosahedron, called an AAV capsid, and up to 4,700 bases of single-stranded DNA.⁵ AAVs can package custom DNA inside of capsids, inject them into the body, and treat genetic disorders by supplying cells with “repaired” genes.

AAVs are also naturally diverse. At least 13 serotypes have been discovered so far. Each one has a unique capsid, dotted with sugars and proteins, that together determine which tissues the AAV targets. AAV9 delivers its genetic payload to the liver, muscles, and lungs, whereas AAV2 targets skeletal muscle, retina, liver, and neurons. Their ability to deliver transgenes into differentiated cell populations is another major advantage. 

Where AAV Manufacturing Cost Goes Up: Packaging Problems

It’s not all “sunshine and lollipops” when it comes to using AAVs for gene therapy. 

The pre-clinical and clinical advantages above are balanced by the unforeseen complications arising as gene therapy developers navigate small-batch manufacturing and scale-up.⁶ One major contributor is failure in the replication and packaging of recombinant AAV (rAAV) genomes into assembled capsids, which causes decreased manufacturing efficiency and yield.

These inefficiencies account for millions of dollars in capital and months/years of manufacturing trial and error per pipeline candidate. In addition, they can cause safety issues, as failed packaging can lead to non-therapeutic impurities, which can induce potent immune responses in clinical trials. These and others are the exact problems contributing to the ongoing accessibility gap. Solving replication and packaging issues in construct design would address a multi-billion dollar problem for the gene therapy industry.

Let’s dive further into the nature of these impurities, the consequences, and the current analytical techniques used to detect them.

Distinguishing full, empty, and truncated AAV capsids

There are a number of impurities that can arise during AAV-based therapy manufacturing. 

Empty AAV capsids are AAVs that contain no genomic DNA. They are a common impurity found even in clinical-grade AAV preparations.⁷ Initially, AAV gene therapy drug products were classified as empty AAV or full AAV capsids, the assumption being that the full capsids were filled with the correct payload. 

We’ve learned through high-fidelity long-read sequencing and other molecular methods that the full capsids can be filled with various components, including truncated genomes. Like empty capsids, these are non-therapeutic impurities. Recent briefing documents published by the FDA on Sarepta Therapeutics’ Duchenne muscular dystrophy drug highlighted that. 

These full but non-therapeutic capsids are problematic for two main reasons: They obscure dose calculations and contribute to immunogenic load, which can lead to safety concerns. To prevent such problems, the number of capsids containing full genomes must be accurately quantified to estimate how much effective payload is delivered. Without that breakdown, the dose may be under the therapeutic threshold but very close to the safety threshold with the number of viral particles being delivered, especially for systemic therapies.

During manufacturing, there are downstream techniques – size-exclusion and anion-exchange chromatography – that enable the removal of some empty capsid impurities but may not remove partially-filled capsids.⁷ Prior to manufacturing, gene therapy developers can examine the presence of empty or partially filled capsids using analytical ultracentrifugation, scanning electron microscopy, and other methods, but they are not high-throughput. 

Full/empty capsid ratios and partially filled capsid analysis are well-accepted as necessary critical quality attributes for gene therapy candidates, yet the assays are not broadly applicable. For instance, full/empty capsid ratios can be analyzed by comparing viral genomes quantified by qPCR or ddPCR to viral particles quantified by ELISA. However, ELISA antibodies are not available for all AAV serotypes, limiting the widespread standardization necessary.

In part, bringing down AAV manufacturing costs relies on reducing the production of these impurities and an industry-wide standardization of analytical methods to detect them and/or eliminate them from AAV preparations. The production method used for Hemgenix, for instance, relied on a baculoviral expression system that produces full-to-empty capsid ratios between 50% and 80%.^8,9 Improving this ratio or reducing the need for downstream purification could significantly reduce the cost of gene therapies.

AAV Payload optimization to improve potency, increase specificity, and reduce immunogenicity

Efforts are being made to engineer better AAVs to prevent these manufacturing issues and enable a workaround for current analytical issues. If the payload that is translated from R&D can be analyzed and enhanced, the proportion of AAV capsids filled with full genomes can be increased – allowing for more therapeutic payload to be delivered with less non-therapeutic payload. This would increase safety, efficacy, and accessibility.

One approach to AAV improvement has focused on capsid engineering for increased manufacturability and reduced immunogenicity. Current methods used for novel capsid discovery focus on rational design, the discovery of AAVs in natural populations, and directed evolution, though these are primarily used to discover or improve receptor or tissue specificity.¹⁰ In silico design, using AI and ML-powered computation methods, have enabled capsid design directed, not only at specificity for a target receptor, but for improved manufacturability. A recently published pre-print by Eid et al. demonstrates the multi-trait optimization of an AAV capsid using a generalizable ML program called Fit4Function.¹¹ The group identified AAV capsid candidates that were produced with high yield, transduction efficiency, and biodistribution.

Another approach to AAV improvement has focused on the engineering of rAAV genomes. For instance, unmethylated CpG dinucleotide-based motifs are immunogenic, and reducing or fully methylating these sequences is a key attribute for reducing the activation of the innate immune system. In addition, secondary and tertiary conformations in rAAV genomes have been found to cause replication errors during capsid filling, leading to the packaging of truncated genomes into AAV capsids. As discussed above, these partially-filled AAV capsids are non-therapeutic, can lead to higher dosing of AAV-based therapies, and ultimately cause safety issues due to immunogenicity. To deal with these issues, we at Form Bio have developed an ML algorithm to predict truncation propensity in AAV constructs, enabling gene therapy developers to optimize constructs for improved manufacturability and reduced immunogenicity.

How Form Bio Can Steer You Down the Right Road to IND Approval

AAV development has many key steps – from finding early efficacy and establishing safety to validating manufacturability – all of which can be enhanced and accelerated with AI and modern computational tools. 

We can help streamline your gene therapy development and manufacturing process, from research to commercialization, helping prepare for successful Investigational New Drug (IND) submission.  Our AI-powered platform, FormSight^AI can help you characterize your AAV product and simulate bioreactor runs without investing large amounts of time or capital, respectively. These solutions are ideal for early gene therapy discovery and development when construct design is your primary focus. 

In addition, our robust platform facilitates data management helping you communicate and share your scientific data more effectively with regulatory agencies getting your life saving and less costly gene therapy to patients fast.