Solid BioSciences, Sarepta, and Regenxbio ($sldb, $srpt, $rgnx)
Gene therapies for the largest gene in the body, protein-protein interaction choices, and the importance of capsid design.
I wrote a few months ago on Sarepta, a company that’s known for its Duchenne Muscular Dystrophy (DMD) treatments. DMD is a devastating disease caused by an inability to produce any dystrophin protein, leaving patients wheelchair bound by age 12 and with an average life expectancy of under 30. Sarepta’s treatments can be divided into two buckets: exon-skipping therapies, which skip over a patient’s mutation during transcription so that a shortened but functional dystrophin protein can be produced, and its gene therapy, ELEVIDYS, which delivers a truncated version of the dystrophin gene to the patient’s skeletal and cardiac muscle cells.
Elevidys is currently the only FDA-approved gene therapy for treatment of DMD, an approval that wasn’t without controversy. The therapy failed to cause a statistically significant change in the primary endpoint measured (NSAA, a test to evaluate motor abilities), but did lead to improvements in other metrics such as time to rise from floor and time to climb four steps. While Elevidys presents a step forward, there’s still much be desired when it comes to treatment.
Fortunately, Sarepta’s not the only company working on therapeutics for the disease, and several others (Solid Biosciences, Regenxbio, and Genethon) are working on gene therapy treatments that are structured differently from Sarepta’s.
Solid Biosciences has a good overview of how gene therapies work in general:
“Gene transfer, a type of gene therapy, is designed to address diseases caused by mutated genes through the delivery of functional versions of those genes, called transgenes. The transgenes are then utilized by the body to produce proteins that are absent or not functional prior to treatment, potentially offering long-lasting clinical benefit. In addition to a transgene, our gene transfer candidates include a viral capsid or vector (a protein shell utilized as a vehicle to deliver a transgene to cells in the body) and a promoter (a specialized DNA sequence that directs cells to produce the protein in specific tissues).”1
The theory behind a gene therapy for Duchenne is straightforward: if you can deliver a functional version of the dystrophin gene then symptoms should markedly improve. The practice is complicated, however, by the large size of the dystrophin gene (it’s the largest in the human body), which means companies and researchers are forced to make a judgement call on which portions to include.
This judgement call is somewhat simplified by the existence of Becker’s Muscular Dystrophy (BMD), a milder form of muscular dystrophy where patients have a much longer life expectancy and progressive muscle weakness is not as severe. BMD patients do produce some dystrophin, but in a shortened form. These patients enable researchers to back into a shortened gene construct that could prove effective for those with Duchenne; getting DMD patients to a place where their symptoms present as a mild form of BMD would be an enormous step forward. Unfortunately, the size of the dystrophin gene in these BMD patients is still too large to be inserted into an AAV capsid (which is why the transgene designed by Solid Biosciences differs from the transgene designed by Sarepta differs from the transgene designed by Regenxbio), but nonetheless serves as a useful roadmap.
The viral vector and promoter choice are other critical aspects of designing a gene therapy. If you get the promoter choice wrong the gene will be translated in the wrong tissue of the body; if you get the vector choice wrong you can end up delivering the gene to the wrong cells and causing a severe immune response in patients. The below slide from a Regenxbio presentation gives a good overview of how Sarepta and its competitors differ in AAV vector, promoter, and transgene construction:2
When thinking about choices on transgene construction, it’s useful to keep in mind that proteins don’t act in a silo but instead interact with each other. Dystrophin plus the proteins that help it do its job form a larger dystrophin-associated protein complex (DAPC). Getting dystrophin expression in patients with DMD is hopefully fruitful, but you also want the proteins that typically interact with the dystrophin protein to be able to do so! This is done by including certain protein-binding sites that are part of the dystrophin gene. Decisions about transgene design, then, really come down to decisions about which proteins you think it’s most vital for dystrophin to be able to interact with. [Those less scientifically inclined should feel free to skip the next two paragraphs. The bottom line is companies make different choices about which protein-protein interactions to prioritize, which in turns impacts the therapy’s efficacy]
Genethon’s transgene construction doesn’t differ from Sarepta’s, which is explained by the seven year research agreement and five year license & collaboration agreement between the two organizations for Genethon’s DMD program. Regenxbio’s management emphasizes their decision to include the CT-domain as part of its transgene. The CT-domain comes at the tail-end of the dystrophin protein, isn’t included in any competitors’ transgenes, and plays an important role in interactions with the α-dystrobrevin protein and five syntrophin proteins (all parts of the DAPC). The hope behind including the CT-domain is three-fold: in mice it increased the half-life of the therapy (the longer the capsid-transgene combo remains in the muscles the more functional protein that can be expressed!), enabled better muscle recovery, and improved restoration of the DAPC (which makes sense; it’s hard to have the entire complex if the associated proteins aren’t able to bind anywhere).
Solid Biosciences’ transgene doesn’t contain the CT-domain, but instead includes the nNOS binding domain. nNOS (neuronal nitric-oxide synthase) is an enzyme that’s believed to ensure enough blood gets to the skeletal muscles during exercise. This is something that doesn’t happen for those with DMD and is a direct contributor to muscular dystrophy rather than only a coinciding event.3 This enzyme can’t be recruited without a-syntrophin, a protein that’s part of the DAPC and can’t bind to dystrophin without spectrin-like repeats 16 and 17. Regenxbio’s construct allows for syntrophin interaction at the CT-domain, but its lack of R16/17 (which can be seen in the above slide) means its construct can’t recruit nNOS.
This can all sound quite abstract, but the reason Solid’s choice is so interesting is because the FDA specifically flagged Sarepta’s construct as not including R16/17:
“The extreme truncations in Sarepta’s microdystrophin [microdystrophin is defined as any shortened version of the dystrophin protein] protein result in absence of important functional domains. For example, Sarepta’s micro-dystrophin does not bind either neuronal nitric oxide synthase or α-syntrophin, two proteins known to play a synergistic role to protect muscle cells. Recruitment of neuronal nitric oxide synthase by wild-type dystrophin at the sarcolemma through spectrin-like repeats 16 and 17 (R16/17) helps control local blood flow by antagonizing sympathetic vasoconstriction. It is therefore unclear to what extent Sarepta’s micro-dystrophin can function similarly to wild-type dystrophin or to shortened forms of dystrophin in patients with BMD.”4
At least currently, Solid including R16/17 seems like a good bet. Its day 90 data for the first three candidates (note the small sample size!) in the Phase 1/2 trial was very positive, with microdystrophin expression levels higher than what Sarepta and Regenxbio reported at the same point in time.
There’s also a potential scenario where Solid’s therapy doesn’t end up succeeding but there’s tremendous value created through its proprietary AAV-capsid. SGT-003, Solid’s current gene therapy candidate, is the company’s second attempt at developing a gene therapy to treat DMD. Its first attempt, SGT-001, had a similar focus on nNOS and a-syntrophin but utilized a different capsid, AAV9. SGT-001 was plagued by safety issues, which led to frequent pauses in the clinical trial and eventually a more permanent program pause in 2022.5 SGT-003 uses a capsid designed in-house, AAV-SLB101, that has improved skeletal muscle tropism (ability to infect the skeletal muscle) when compared to AAV9. Improved tropism matters because an AAV capsid is a virus and so elicits an immune response. The better the tropism, the lower the dose can be, and so the less you have to worry about an immune response affecting the therapy’s efficacy/safety. Solid has licensed out this proprietary capsid to 15 academic labs and corporations, and so stands to benefit enormously if these labs/corporations bring a therapy using this capsid to market. The company’s additionally putting effort into developing capsids to tackle cardiac and neuromuscular diseases, with a similar intent to license them out if successful.6
Sarepta’s CEO has been overly dismissive of any competitors’ efforts on earnings calls, but in doing so he makes a few good points on the company’s short-term advantages.7 As exciting as Solid’s results are, so far only 6 patients have been dosed with the therapy, which means there’s very limited data on both safety and efficacy. That puts those with DMD in a challenging situation: choose a therapeutic with a better known safety-profile that has limited efficacy or choose one that doesn’t yet have a ton of safety data but could end up being a better treatment.8 Moreover, Sarepta has a real distribution advantage given its years of serving DMD patients through exon-skipping therapies. Its existing relationships with infusion centers, which are used for both exon-skipping and gene therapy treatments, means the company is just better positioned than competitors when it comes to rolling its gene therapy out at scale. Furthermore, even if Regenxbio and Solid both have clinical trials that go well, the longer they take to get to market the more existing DMD patients Sarepta has the opportunity to dose. Gene therapies are intended to be one time, in large part because a patient’s immune system will recognize the AAV virus the second time and do a much more effective job at stopping it from reaching the intended cells. Solid management argues that its capsid is differentiated enough from Sarepta’s AAVrh74 capsid that redosing opportunities may exist for the patients who have already received Elevidys, but then one runs into the question of whether insurance will cover a second treatment that costs in the low millions, particularly if it displays a questionable level of efficacy.9
Disclaimer: The information in this post is not intended to be and does not constitute investment or financial advice. You should not make any decision based on the information presented without conducting independent due diligence
Solid Biosciences Annual Report, Pg 1.
You can find the entire presentation here. (“RGX-202: AFFINITY DUCHENNE® PIVOTAL TRIAL AND INTERIM FUNCTIONAL DATA”
Some of Solid’s safety issues highlighted the importance of high full to empty capsid ratios. Manufacturing gene transfer therapies is far from easy, and the AAV capsids produced don’t always end up containing the intended transgene (hence they are called ‘empty’). This really matters with AAV-based gene therapies because an empty capsid stresses the immune system with no clinical benefit in return. The worse the full to empty ratio, the greater dose you have to give patients, and the more the immune system is unnecessarily stressed.
If Solid becomes primarily a capsid-design company it’d be a fascinating instance of a biotech going from a therapeutic company to a pure platform play, something that typically happens in reverse! Solid’s efforts also highlight the importance and value of what Dyno Therapeutics is working on.
When asked about Elevidys competitors on Sarepta’s most recent earnings call the CEO said: “As far as modeling….it's way too early to be modeling competition from people that have 1, 2 and 3 patients in these early experimental therapies. Probably haven't even figured out manufacturing at that point. Certainly, we wouldn't be modeling anything this decade in any event. It's something we can look to if people will actually start progressing.”
Better-known doesn’t mean entirely safe. Sarepta this week announced that a patient had unfortunately died from acute liver failure after his treatment with Elevidys. A significant challenge with AAV-vectors is the stress they can place on the liver, which highlights the vital importance of capsid design.
Solid Biosciences Presentation, pg 15.