How Are AAV Vectors Transforming Gene Therapy?

Estimates suggest that genetic factors contribute to the development of common diseases by 40%. For rare diseases, this figure doubles to 80%, most of which have no FDA-approved treatment.^1,2 ‍

Recently, this large unmet need has begun to be addressed with gene therapies, which have given drug developers the tools to address disease-causing mutations with precision, accuracy, and a single treatment versus treating disease symptoms. These tools are multi-faceted solutions with various modalities for genome editing and therapeutic protein replacement.  

Some examples include, two FDA approved cell-based gene therapies – Casgevy and Lyfegenia – which use genome editing for the treatment of sickle cell disease.³  Meanwhile, advancements in gene replacement approaches are evident with FDA approved gene therapy, Zolgensma, for the treatment of spinal muscular atrophy (SMA).⁴‍

While the power to cure all genetic diseases, including rare ones, is an alluring lens to view the future, there are modern-day barriers. Much like other therapeutic modalities, the translation gap remains, making it difficult for developers to go from proof-of-concept disease models in the biological lab to quickly demonstrate safety and efficacy in clinical trials with experimental gene therapies. 

Here, we discuss the different gene therapy delivery modalities, the current challenges with their deployment, and how to overcome them. Our primary emphasis is on adeno-associated virus (AAV)-based therapies, recognized for wide-spread application and significant potential in this delivery method.

AAV Functionality: From Protein Shell to Therapeutic Vehicle

AAV is a protective protein shell enveloping a small, single-stranded DNA genome of around 4.8 kilobases (kb) and relies on co-infection with other viruses, primarily adenoviruses, to undergo replication.⁵ The single-stranded genome of AAV comprises three key genes: Rep (Replication genes for Rep78, Rep68, Rep52, and Rep40 ), Cap (Capsid gene for VP1, VP2, and VP3), and aap (Assembly), flanked by inverted terminal repeats (ITRs).⁶ Using three promoters, alternative translation start sites, and differential splicing, these genes generate at least nine gene products mediating viral genome replication and packaging and target cell binding and internalization.⁷ 

Mechanism of Action in AAV Gene Therapy

A recombinant AAV (rAAV) is generated by replacing viral genes with therapeutic transgenes still anchored by the viral subtype-specific ITRs. This leads to the production of AAV capsids capable of entering cells and delivering transgenic AAV genomes to the nucleus. Once there, the recombinant genome, flanked by ITRs, circularizes to form an episome from which transgenes can be expressed.⁸ The persistence of these episomes and transgene expression is inversely proportional to the replication rate of the transduced cells: The faster the cells divide, the more the episomes are diluted, and the less transgene expression is achieved. 

Defining AAV Capsid vs. AAV Construct vs AAV Vector

For rAAV-based therapies, the terms “capsid,” “construct,” and “vector” are relevant, widely used terms in gene therapy design and development.

The capsid is the protein coat surrounding the AAV viral genome, which, in the case of rAAV, engineered to contain a therapeutic transgene. AAV capsids mediate entry to the target cell. 

An rAAV construct refers to designing a DNA sequence that will be packaged into a rAAV capsid structure. The rAAV construct typically includes therapeutic transgenes and regulatory elements (promoters, enhancers, etc.) that control the expression of the transgene. 

The viral vector refers to the entire delivery system: An rAAV construct inside the AAV capsid, serving as a vehicle for therapeutic gene delivery into target cells. 

Approved Gene Therapies Leveraging AAV as a Delivery System

As of January 31, 2024, There are 34 distinct cell and gene therapies FDA-approved. Over 75% of these (26 out of 34) are allogenic or autologous cell therapies, and the remaining 8 are viral gene therapies – including adenovirus, AAV, and herpes simplex virus-1 (HSV-1) – engineered to deliver a therapeutic genetic payload to specific cells⁹.  

The 5 approved AAV gene therapies, year of approval, and serotype include the following:

1. Spark Therapeutics;  Approved in 2017, Luxturna is indicated for the treatment of RPE65 mutation-associated retinal dystrophy. Serotype AAV2.
2. Novartis; Approved in 2019, Zolgensma is indicated for the treatment of survival motor neuron 1 (SMN1)-associated spinal muscular atrophy (SMA).  Serotype AAV9.
3. CSL Behring LLC; Approved in 2022, Hemgenix is indicated for the treatment of adults with Hemophilia B.  Serotype AAV5.
4. Sarepta Therapeutics, Approved in 2023, Elevidys is indicated for the treatment of Duchenne muscular dystrophy (DMD) with a confirmed mutation in the DMD gene. Serotype AAVrh74.
5. BioMarin Pharmaceutical Inc.; Approved in 2023, Roctavian is indicated for the treatment of adults with severe hemophilia A.  Serotype AAV5.

However, this success has been accompanied by several major challenges that are currently standing in the way of realizing the full potential of AAV-based gene therapies. 

Natural AAV infection is not associated with any specific pathophysiology in humans, yet the increased number of AAV-based gene therapies being tested has brought several safety concerns to the surface. Several highly publicized deaths in clinical trials with AAV-based therapies sparked concern over the high dosing used in patients and some of the contaminants present in AAV preparations.¹⁰ Thus, balancing efficacy and safety is particularly challenging for viral delivery-based gene therapies.

Challenges in Vector Design and Packaging of AAV Gene Therapy

To ensure the effectiveness and safety of AAV gene therapy, preclinical developers need to overcome substantial operational risks associated with the vector design and manufacturing process. Failure to address these risks adequately can result in a significant amount of time and money spent on therapeutic paths that face late failure rates. 

AAV Packaging and Manufacturing Challenges

Generally, the packaging size of an rAAV vector is limited to 5 kb (including the ITR sequences) and smaller. If vectors are larger than that, viral production yields are greatly reduced, and those vectors that are packaged are truncated. Truncated rAAV constructs are non-therapeutic and dilute the purity of rAAV preparations.

Construct truncations and other impurities, such as chimeric vectors or host or helper plasmid DNA, can also arise due to unforeseen construct design choices. The generation of these non-therapeutic impurities during production can sometimes be caused by DNA secondary structure and the stalling of the replication machinery.¹¹ With these contaminants comes the need for downstream enrichment, which can be technically challenging and expensive and inefficient, which ultimately raises the overall cost of gene therapies that are approved for commercialization.

Tissue Tropism Challenges

When systemically administered to humans, most natural AAVs accumulate in the liver and can efficiently transduce hepatocytes. However, different AAV serotypes enable delivery to various other tissues helping drive specificity.

Recently, there’s been significant focus on delivery to the CNS which may require crossing the blood-brain barrier or delivery directly to the brain or spinal cord. Achieving efficient and uniform targeting has been a continued challenge for AAV developers, as well as deselecting the liver and enriching for the target tissue such as the kidney, heart, or muscle.¹²

Improving Effectiveness and Accelerating Speed of AAV Gene Therapy Development

Overcoming the efficacy and safety challenges requires a multi-faceted approach focused on optimizing AAV constructs and capsids for the manufacturing scale-up beyond the small-scale production required in the preclinical stages.

Improving AAV Construct Yield, Safety and Expression

Recombinant DNA is nothing new, and regulatory elements and transgenes involved in construct design are often equated to interchangeable modules, where they can be put in different contexts and behave similarly.

With AAV-based gene therapies, there is significant variation in how different constructs can behave at every step of development and manufacturing. 

For instance, expression of mRNA and protein can be challenging to predict and is highly dependent on the transduced cell types, secondary structure, metabolic state of the cell and more. There’s also an enhanced risk of immunotoxicity associated with CpG islands, which may not affect the overall expression of the transgene but can cause significant safety issues.

‍Partially filled AAV capsids have also been found to be a common impurity in AAV production runs.¹³ These can arise due to secondary structures and certain sequence motifs present in rAAV constructs. These lead to AAV capsids being filled with partial constructs and these can ultimately lead to the production of an immunogenic truncated transgene or the need to administer high doses to reach desired levels of efficacy. 

Developing ways to perform multi-factorial optimization for all of these potential complications early in development can save significant time and effort for gene therapy developers. In silico solutions are currently in development to predict gene expression, compare multiple constructs across different parameters (e.g., yield, safety, and expression), packaging, truncation, and more.

Improving AAV Vector Design and Product Characterization

Optimizing construct design early in development to improve the quality and yield of manufactured AAV products is one way to design for manufacturing right from the start. In addition, using techniques that efficiently and comprehensively characterize AAV manufacturing runs can provide the data to further optimize manufacturing processes. One example is long-read sequencing, which offers a more complete understanding of construct variants (e.g. mutations, truncated sequences, etc.) within manufactured rAAV.

‍In silico technology also provides algorithms that enable AAV yield and quality predictions for bioreactor runs and provide improved AAV product characterization to ensure a thorough understanding of product composition. 

Improving AAV Capsid Tropism and Capacity

Significant effort has also been made to improve the tissue tropism or cell specificity of AAV serotypes. Rational design, directed evolution, and in silico approaches have been used to engineer capsids with improved transduction of muscle cells, lung alveoli, and retinal endothelial.^14,15‍

As AAVs have small packaging capacities, there have also been construct design efforts that minimize the size of the transgene and regulatory elements required for efficacy. In principle, AAV capsid engineering could also help increase the capacity of currently available AAVs.

Form Bio’s in silico Partner Program to Advance AAV Gene Therapy Development

Form Bio now offers powerful in silico capabilities from disease models to IND for gene therapy development. This range of capabilities allows us to be a co-development partner with gene therapy companies in a broad, program-like manner from the early stages of drug design through preclinical development and regulatory approval to begin clinical development. 

We are widening our moat by securing and sequencing what we believe is the most comprehensive aggregation of high-fidelity NGS data on gene therapy constructs, with a variety of genes of interest, promoters, etc. and incorporating this data into our training and inference models to understand the following:

• manufacturing output
• construct  design flaws 
• secondary and tertiary structures that impede translation
• immunogenicity

‍AI Disclosure: Feature image was generated by an AI image tool MidJourney.