Axial - Regenxbio
Surveying great inventors and businesses
As always, if you or someone you know has a great idea or company in life sciences, Axial would be excited to get to know you and possibly invest in your vision and company — firstname.lastname@example.org
Regenxbio is a fascinating business that is powering gene therapies that are curing disease. The company is an interesting case study on how to set a standard in life sciences, in the case of Regenxbio making adeno-associated virus (AAV) the standard vector for most gene therapies. The founders used this to build a unique business. Founded in 2009 by James Brown (COO) and Kenneth Mills (CEO), Regenxbio formed from a collaboration between the biotechnology consultancy firm FoxKiser, Penn, and the Wilson Lab. James and Kenneth were both trained as chemists and worked at a diagnostics company called IGEN in the late 1990s and early 2000s then both at Meso Scale Diagnostics, a joint venture between IGEN and Meso Scale Technologies. Before forming Regenxbio, they were both working at FoxKiser.
I’ve tracked Regenxbio since 2013, but a recent conversation with my buddy Daanish stimulated me to dig in deeper and understand how the company set the standard for AAV gene therapies and what set of features can be extracted to understand how other companies can do something similar.
While at FoxKiser, Kenneth got involved with work around gene therapies. An outsider’s perspective as a chemist and industry experience in diagnostics was instrumental in the founders’ ability to rigorously analyze the gene therapy field and not be stained/discouraged by past failures. James Wilson (inventor of Regenxbio’s core platform) pioneered gene therapies and was building momentum for the class of medicine throughout the 1990s. The death of Jesse Gelsinger in 1999, a patient with ornithine transcarbamylase (OTC) deficiency treated with an adenovirus therapy, put a halt to progress in the field and set it back by at least 10 years. With a fresh pair of eyes, the founders saw AAVs, whose utility was discovered in the mid-2000s, as a safer vector that were seeing major advancements in their ability to express payloads at a much higher level than previous versions and had the increasing potential to have a therapeutic effect across more diseases.
Through their work at FoxKiser, the founders began collaborating with Wilson. After forming Regenxbio, the founders licensed out a suite of AAV patents from Penn and GlaxoSmithKline (GSK), who had some of the rights previously and sponsored nonclinical AAV research at the Wilson Lab.
Externally, key events occurred that drove the company’s growth. The ‘Why now?’ was really driven by the increasing capabilities of AAVs and the founders' courage to form a company in the field before it was obvious. Success of AAV2 carrying REP65 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4169122/, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2829748/, https://www.ncbi.nlm.nih.gov/pubmed/19953081) followed by the success of AAV8 (Wilson was involved) in Hemophilia B (https://www.nejm.org/doi/full/10.1056/NEJMoa1108046, https://www.nejm.org/doi/full/10.1056/NEJMoa1407309) were the major events that set up Regenxbio to license out their technology to a wide set of companies.
Regenxbio’s initial approach was licensing, then they made the transition to internal development in 2014/2015. This ultimately has led to a bimodal business model where the company licenses AAVs outside their core areas of focus and has an internal pipeline focused on metabolic disease, CNS, and ophthalmology. Simply, the company focuses on three areas to grow - making sure they can make proteins safely and durably and make manufacturing easy:
The key early driver for their success to license out to so many companies was focusing on safety. Given the past history of gene therapies, Regenxbio’s value proposition was safe therapies that can be clinically useful. With a few gene therapies now approved, the next generation of companies can now focus on new features to develop new vectors with better delivery and efficacy profiles. From this, there are a set of features that have enabled Regenxbio to create the standard in gene therapies:
Owned a large piece of land in AAV IP especially when others were skeptical
First-mover in the clinic
Showed utility in a field where failure was expected (i.e. low bar to overcome)
Focus on safety, not efficacy
Initial plan to license out to other companies
Buy-in from the leading lab and clinicians in the field
Standards are a really great way to build layers of abstraction that power new companies and build powerful moats (i.e. commoditize your complement). This is really true in software. There is a massive opportunity in life sciences especially in drug development to standardize various processes - this is where software is a major enabler of new methods. Standards enable faster versioning of medicines and other products. Standards speed up development Standards derisk businesses attracting more and new types of capital and talent. This case study serves to figure out what set of features allow a company to set a standard in life sciences.
Around its founding, Regenxbio took a one-step approach to translate AAVs into medicine. Rather than overengineer a new modality, taking the obvious approach and getting to the clinic as quickly as possible put the company into the position it is today. By partnering with the Wilson Lab, who at the time had the perception of leading a dead field - gene therapies, Regenxbio got access to a deep portfolio of exclusive AAV patents. This probably was the most important decision the founders made for the company. They essentially bought IP low when everyone else was down on gene therapies then developed it and built an incredible business.
So far, Regenxbio’s biggest success has been powering AveXis’ (now Novartis) gene therapy, Zolgensma to cure spinal muscular atrophy (SMA). AveXis was one of the earlier companies Regenxbio engaged with. The logic to go after SMA first rather than say Parkinson’s or something was that SMA’s mechanism was well understood and could be solved well through protein expression.
The initial model the founders of Regenxbio pursued was licensing. After executing a few deals and waiting a few years for others to take clinical risk, the company added an internal pipeline. The key risk they took was focusing on AAVs and gene therapies so early on. That first mover advantage gave the company a buffer to make some mistakes along the way and figure out who to work with and at what terms. Being early on with a new modality helped Regenxbio set licensing terms at really attractive rates.
AAVs are effective therapies but are can’t miss like the Skywalker/Death Star shot. The stakes actually might be higher - the life of a patient is more important than the fate of the republic. After an initial dose, the immune system forms a memory against the AAV and inhibits re-dosing. As a result, a lot of work upfront has to be done to map preclinical work onto clinical success. For gene therapies versus antibodies or small molecules, that means a lot more mice studies and especially non-human primate work. It’s harder to iterate across products for gene therapies in humans. It’s a modality where groups have to measure many times before making the cut. The key experiment for companies like Regenxbio is to find the minimally effective dose in models before going into humans. After that, they can go the distance to a pivotal trial and hopefully approval.
The most important part of Regenxbio’s technology is to understand AAVs. AAV gene therapies can infect a wide set of cell types without a large immune response (versus adenoviruses) and without the risk of random insertions within the genome (versus lentivirus). In short, Regenxbio took on developmental risk of AAVs but had minimal-to-no risk around discovery. The company didn’t need to rely on huge breakthroughs to make their business model work.
The components of AAVs are:
Two inverted terminal repeats (ITR) that enable the creation of a double-stranded DNA. AAVs have a single-stranded genome. Integration requires dsDNA.
Between the two ITRs are the Rep and Cap genes, which encode the replication and capsid (viral packaging) proteins, respectively, which are important for viral delivery and effectiveness
A separate plasmid, called the transfer plasmid, is used to deliver the therapeutic transgene, flanked by two ITRs (i.e. in cis)
A third plasmid, helper plasmid, is used to express essential genes for the AAV. This is incredibly important because natural AAVs cannot replicate and enter the lytic cycle alone.
Various promoters are designed to ensure expression of these components is high
All of these components are often put into HEK293 cells to manufacture recombinant viral particles
A major rate limiting step for AAVs is packaging capacity, about 4.7 kb. Some companies create payloads that are small enough to fit into an AAV. A method used to increase the capacity of AAVs is to split the transgene into two transfer plasmids and have them spliced together during manufacturing.
The second part of AAV gene therapies is manufacturing. HEK293 cells are used because they grow easily (although improvements can still be made here) and provide adenoviral E1 genes to help generate AAV particles. The three components described above are transfected into the cells. About 2-3 days later, the cells are harvested and cycled through freezing and thawing to induce viral particle release, which are collected and purified. The two key metrics are viral titer (how many particles are there be mL) and purity (minimize the concentration of the lysate, which could be toxic).
The third part is different variants of AAVs called serotypes, which is determined by the capsid sequence. AAVs come in many stripes and colors having different infection efficiencies for certain cell types (called tropism) across serotypes. There are 11 AAV serotypes with 1-9 the most widely used for therapeutic gene therapies.
What’s really exciting here is the ability to engineer new serotype variants to alter delivery. Companies like 4D, AskBio, Dyno, and Spark emerged to screen (rationally, directed evolution, or both) different capsids and create new AAV variants. Regenxbio does some of this, but their real advantage was showing clinical success that increased demand for their AAVs. Regenxbio probably just buys one of these companies sooner or later. Capsid sequences can be mixed by AAV serotypes to create new vectors - this pseudotyping can create a AAV1/8 vector from AAV1 and AAV8 that has different properties than the individual vectors. Synthetic capsid sequences can also be used.
Finally, a major advantage of AAVs is their ability not to integrate into the genome and potentially disrupt an essential gene. As long as the Rep gene is used on a different plasmid (i.e. in trans), the AAVs manufactured will generate dsDNA that will be extra-chromosomal, called an episome, that is not integrated within the genome but is expressed reliably by the cell. These episomes are pretty complex and large structures and can express proteins for long periods of time (the AveXis and Spark clinical data will show for how long in humans) in non-dividing cells.
Regenxbio’s platform and products rely on these 4 parts. This case study will focus on the company’s platform and how it built a business model around it instead of the core pipeline. It’s headline wet AMD program is exciting but is facing stiff competition from Regeneron with Eylea. Really, Regenxbio is one of the few biotechnology companies where after ~10 years its platform is worth more than its pipeline:
Regenxbio calls its platform NAV. Around its founding, the company took a one-step approach to translate AAVs into medicine. Rather than overengineer a new modality, taking the obvious approach and getting to the clinic as quickly as possible put the company into the position it is today. By partnering with the Wilson Lab, who at the time had the perception of leading a dead field - gene therapies, Regenxbio got access to a deep portfolio of exclusive AAV patents. This probably was the most important decision the founders made for the company. They essentially bought IP low when everyone else was down on gene therapies then developed it and built an incredible business. At the start of the company, the NAV platform searches for capsid sequences across a wide set of tissues from humans and non-human primates. The team would find an AAV serotype that had a high prevalence in a certain tissue then do engineering on the various components of the AAV and develop manufacturing methods to scale up production. A lot of exciting and hard work. The platform, from search to manufacturing, optimizes for a set of features of an AAV:
Tissue specificity and distribution
High and durable gene expression
Low immune response
High levels of integration
These five parts are the major driver for any gene therapy company. In the late 2000s, Regenxbio was the best in the AAV field to generate vectors with useful features. A few years later around 2012/2013, Regenxbio’s vectors accounted for over 75% of all gene therapy INDs.
Today, Regenxbio’s relationship with the Wilson Lab still generates new AAVs and their partnership with WuXi AppTec is working on scaling up manufacturing - the process is capital intensive and requires a lot of steel so having partners allows Regenxbio to focus on what it is really good at: generating useful AAVs and licensing them out:
Gene therapies have created multi-billion companies and the field has barely gotten started. When Regenxbio was founded, the underlying market structure of gene therapies was non-existent - fragmented on both sides. The founders had to create a new market and have a vision on what is important and what is not. A focus on safety gave the company a high value proposition at the time and enabled their vectors to get into the clinic sooner rather than later.
The market for Regenxbio was any disease that is genetic. That’s a pretty large market by any standards. The gene therapy market has $100Ms in revenue right now and will likely have billions and maybe tens of billions of revenue pretty soon (3-6 years).
So far, Regenxbio’s biggest success has been powering AveXis’ (now Novartis) gene therapy, Zolgensma to cure spinal muscular atrophy (SMA). AveXis was one of the earlier companies Regenxbio engaged with. The logic to go after SMA first rather than say Parkinson’s or something was that SMA’s mechanism was well understood and could be solved well through protein expression. SMA is a neuromuscular disorder that is caused by a mutation in the SMN1 gene (important for the survival of motor neurons). This leads to lower levels of SMN protein. An AAV that delivers an SMN1 transgene with high levels of expression could cure the disease. The team at AveXis did incredible work to bring the therapy through the clinic and approval in 2019. With about 20K people in the US with the disorder, they now have hope with a curative therapy. Pricing the medicine at a little over $2M per patient, the market size for gene therapies in SMA alone is well over $40B. Being one of the first gene therapies approved, AveXis and Novartis had to do a lot of the heavy lifting to figure out new pricing models for curative therapies. The field will benefit from this work as more therapies come through the clinic that can actually cure disease. Working with payors, Novartis essentially came to an agreement where if it doesn’t work, you get your money back.
The initial model the founders of Regenxbio pursued was licensing. After executing a few deals and waiting a few years for others to take clinical risk, the company added an internal pipeline. The key risk they took was focusing on AAVs and gene therapies so early on. That first mover advantage gave the company a buffer to make some mistakes along the way and figure out who to work with and at what terms. Being early on with a new modality helped Regenxbio set licensing terms at really attractive rates. Having a monopoly over AAVs at least for a few years helped the company get great royalty rates. Rarely are there monopolies in life sciences, but the market that the founders pursued was amenable for Regenxbio to set the standard for gene therapies.
Starting off, the company owned a large amount of IP covering several naturally occurring AAVs. Around 2009, the two AAVs with the most attention were AAV1 and AAV2. Regenxbio pushed forward with AAV7/8/9 that had broader distribution profiles. That’s another key risk the founders took - focusing on different AAV serotypes. Zolgensma is an AAV9 therapy. Many of Regenxbio’s initial deals were focused on AAV7/8/9.
Beyond taking a stance on the technology, the second hard part of executing a licensing business is figuring out who to work with. Licensing out technology is exciting because there is an upfront payment, maybe some fees along the way, and an ambiguous royalty rate. The terminal value of any licensing deal is that royalty rate. Simply, a company like Regenxbio had to take risks with what companies to work with. Those companies are the ones that will do the hard work and make sure a therapy gets approval. Licensing out AAV9 for SMA to another company and not AveXis probably wouldn’t have worked out for Regenxbio.
The best licensing deal Regenxbio has been able to complete so far has been with AveXis. That deal alone makes Regenxbio very undervalued. In 2014, Regenxbio exclusively licensed an AAV9 vector for SMA to AveXis. These terms can be found in Regenxbio’s S-1 and other SEC filings. The best companies I’ve partnered with compiled all of the public data and terms for their modality:
An initial fee of $2M
An annual maintenance fee
Up to $12.25M in milestone fees
Mid-single to low-double digit royalties on net sales of products AveXis develops with the technology
Regenxbio executed deals like this across well more than 20 programs across AAV vectors 7-10 (images below). The company has no approved products themselves but will have $100Ms in sales. Over the next decade, Regenxbio will face more competition. They set the standard for gene therapies and placed their bets in various partners, but companies like Bluebird, BioMarin, and uniQure are knipping at Regenxbio’s heels. Nothing comes easy. Beyond that, Regenxbio’s AAVs face competition from other modalities. This is a key point for the future of medicine - clinicians will have a diverse arsenal of medicines from ASOs to gene therapies to mAbs to treat disease. Competition is getting fiercer in some spots. For Regenxbio, their internal wet AMD program has a high hurdle to jump over that’s called Eylea. Ultimately, Regenxbio set a standard for gene therapies and used their timing and skill to build a unique business.
Source: Zack Fink