By Anna Rose Welch, Editorial & Community Director, Advancing RNA
If Shakespeare's famous Sonnet 18 had been written in the form of a dialogue on comparability with the FDA, I like to imagine it would have gone something like this:
Industry: “Shall I compare my gene therapy to a summer's day?”
FDA: “No. It should be as highly similar to itself as possible.”
Unfortunately, this was not quite how comparability was broached during Sarepta’s FDA Advisory Committee meeting for its AAV gene therapy in Duchenne Muscular Dystrophy (DMD). That said, as I prefaced in my previous article of takeaways from this event, there was plenty of Shakespearean drama-worthy dialogue between the advisory committee, FDA, and Sarepta, all of which culminated in an uncomfortably close vote in favor of the therapy’s accelerated approval. Following two more months of FDA deliberation, the product ultimately won an accelerated approval to the joy of the DMD community.
Though there’s a lot that can be said about the adcom, I found it, first-and-foremost, an incredible case study on how our biggest CMC discussions are translating into real-life regulatory dialogue and decision-making. In fact, I’d go so far as to declare this adcom one of the most momentous occasions we’ve witnessed so far in the CGT industry — especially for those of us in the manufacturing world. To my knowledge, this was our first publicized example of a clinical (i.e., not analytical) comparability study.
There are a few must-know details about the why & how of Sarepta’s comparability exercise, which I’ll outline for you here. But it wasn’t the facts and figures underpinning the comparability assessment itself that I found the most noteworthy; rather, this adcom served as an important reminder about the overall role comparability can and likely will play in the regulation of our advanced therapies moving forward. I’ve said it before and will say it again: our therapies and their regulatory risk vs. benefit equations are all unique, which makes it difficult to broadly extrapolate meaning from a single adcom/product. But if we look at this “flagship” example alone, I’d argue that the regulatory approach to comparability could look and feel a bit different than what we might have imagined.
A Tale of Two Processes: The How & Why of Sarepta’s Comparability Exercise
Sarepta’s clinical gene therapy product was made using two GMP-compliant manufacturing processes, “affectionately” named Process A and Process B.
The product created using Process A was made by Nationwide Children’s Hospital for Sarepta’s early-phase clinical studies (a.k.a. Study 101 and Study 102).
Prior to the company’s pivotal trial, Sarepta changed its manufacturing partner and process — which, as the FDA specified in its comparability guidance, is “generally considered a major change.”
Process B — which is Sarepta’s commercial manufacturing process — was run by the CDMO Catalent. The product currently being investigated in part 1 of Sarepta’s Phase 3 study (a.k.a. Study 301) was made using Process B.
Of course, this tale of two processes doesn’t end here. The change from Process A to Process B ultimately resulted in a product with a different purity profile. In particular, the commercial product made by Process B boasts a much larger percentage of empty capsids. (What that percentage is, however, has not been publicly shared.)
Even those of us who are only remotely connected to the AAV gene therapy world likely know that the full/partial/empty capsid ratio has become a hotter and hotter topic of discussion and debate. With any AAV gene therapy, the goal is to make sure you are dosing patients with capsids that are full of all “the right stuff” — a.k.a. the entire and proper transgene (i.e., cargo) and regulatory elements — as opposed to empty or partially filled capsids that contain transgene fragments or other impurities (e.g., host cell or helper DNA). The clinical impact and safety of partial and empty capsids remain areas of ongoing research. However, at this current moment in time, the higher the percentage of empty capsids, the greater the concerns about a therapy’s clinical efficacy and safety/immunogenicity.
We can see these concerns clearly reflected in Sarepta’s adcom, during which the FDA reviewers posed questions about whether the larger percentage of empty capsids from Process B would interfere with transduction efficiency, and in turn, the therapy’s efficacy. They also raised questions about whether the greater number of capsids needed to ensure patients received the necessary dose of the gene therapy would cause serious immune responses/adverse events. (Please note, there were no deaths in this trial and only a small number of serious adverse events. Even the patient advocate on the committee emphasized that he was “exceedingly pleased” with this therapy’s safety profile compared to other AAV therapies and with Sarepta’s efforts to understand and mitigate the risk of future AEs.)
Now, all this background information raises two key questions, the first being: Why did this change transpire in the first place — especially if NCH offered a product with a better packaging profile under GMP compliance?
Based on the FDA briefing documents alone, we can speculate that scalability — both in terms of meeting current U.S. and perhaps even global demand in the long-term — was a key reason for the change in process. As was noted in the briefing documents, moving to Process B involved a change to a “scaled-up purification method that incorporates chromatography-based methods for empty/full separation.” This is compared to process A, which was “an AAV-based purification process that allows near-complete removal of empty AAV capsids.”
Though the FDA was predominantly focused on the downstream process, I think it’s also important to step back and consider the impact that the upstream process also has on the quality and scalability of a therapy overall. A quick look at NCH’s vector production process (Process A) reveals that they offer an adherent process. It’s pure speculation on my part that the transition to Catalent also included a change to suspension culture. But it’s a safe bet, seeing as more companies today are striving to achieve viral vector production in suspension culture at 1,000- or even 2,000-liter scale, for the sake of scalability. Generally speaking, suspension culture may still leave quite a bit to be desired from a yield perspective today; but I’d imagine achieving the necessary yields to dose the commercial market using an adherent process would call for a Manhattan-sized collection of cell culture vessels.
McKinsey said it best here in its 2022 viral vector report: “The choice among current upstream production systems remains a trade-off between flexibility, scalability, and quality.” Based on the little we can piece together here, it seems that Sarepta’s experience beautifully demonstrates the current tradeoff we may face between achieving the necessary scale and quality.
No doubt, many of us are also wondering how we ended up with a clinical and not an analytical comparability study. As we already know, the percentage of full viral particles in Process B product was “significantly different" compared to that of Process A product. But a closer look at the FDA materials reveals an intriguing analytical detail. Though the company used the same commercial test method in their comparability study to test materials from each process, samples from Process A and B were tested “at different times and at different contract testing labs, without method transfer; therefore, no direct side-by-side analysis was conducted.”
If we refer to page 17 of the FDA’s comparability guidance, it specifies that we test pre- and post-change samples “side-by-side,” or “head-to-head,” in the same experiment. (Got retains?) Method transfer is also essential when relying on multiple facilities.
Ultimately, the documented lack of analytical “seaworthiness” translated into an open label clinical bridging study, referred to throughout the adcom as Study 103. This bridging study compared the safety and expression of Sarepta’s micro-dystrophin in 40 patients receiving Process B product with that of the Process A product in Study 102. (FYI, Study 101 and 102 had a combined total of 45 patients.)
I’d encourage you to check out the FDA’s briefing materials for greater clinical detail on AEs seen throughout all arms of the trial; there were a small handful of cardiac-related AEs seen in patients undergoing treatment with Process B product. However, overall, our first foray into a clinical comparability study demonstrated that Process B product, despite having a higher number of empties/partials, was clinically comparable to Process A product in terms of protein expression and toxicity.
What Can This Adcom Teach Us?
As nerve wracking as it may be to witness, Sarepta’s adcom reaffirmed a difficult but important truth we all will have to navigate: Our industry will have to get much more comfortable with the concept of manufacturing changes despite the inherent risks — especially if we want to achieve our long-term goals of improving scale, titers, quality, and immunogenicity/safety profiles for many of our therapies. As I wrote in a previous article on the importance of change, many of the technologies and materials we may need to improve our therapies won’t always be there for us in the earliest stages of clinical development when comparability is a “lighter” lift. Sarepta’s comparability exercise ultimately revealed the extent to which our industry’s overall technical capabilities and analytical approaches are not yet as “grown up” as we may need/want them to be.
As part 1 of this article series demonstrated, this was far from a clear-cut adcom meeting all around — from a nonclinical, clinical, and a CMC perspective. For obvious reasons, manufacturing was much more closely scrutinized in Sarepta’s meeting than in previous FDA adcom meetings. In fact, there were at least two adcom committee members who voted against accelerated approval in part thanks to the fact that Sarepta’s ongoing pivotal trial (Study 301) is the first and only trial investigating the Process B product in a blinded setting.
But all that said, I walked away from this adcom and its overall discussions on manufacturing comparability feeling surprisingly heartened. In the grand scheme of the entire adcom, manufacturing and comparability received much less discussion and intense scrutiny throughout than I would’ve anticipated, especially given the significant role comparability currently plays in CMC-oriented discussions.
Do not get me wrong: Sarepta’s change raised the FDA’s eyebrows — and for important reasons. We must do our due diligence to demonstrate that we are providing a safe, efficacious, and high-quality product, no matter what it takes. For some of us, particularly in these earlier days, doing our due diligence may mean performing a clinical comparability study — spared no expense. But this adcom served as an important reminder to all of us who live and breathe CMC every day that manufacturing changes and comparability exercises are just one of the many critical puzzle pieces comprising an advanced therapy.
We hear it said regularly that a CGT therapy’s approval is dependent on two important factors — its unique risk vs. benefit ratio and the totality of the evidence. Throughout the adcom, Buddy Cassidy, the patient advocate on the advisory committee, provided some of the most thoughtful, persuasive, and well-educated arguments a patient advocate has ever offered in favor of a therapy’s benefits. Frankly, I got chills every time he opened his mouth — and I don’t think I was the only one.
Likewise, despite the FDA’s seeming resistance to offering an accelerated approval for this therapy, the agency and Peter Marks demonstrated that they do live by their word when it comes to taking the totality of the evidence into account. At one point during the discussion, an adcom member asked if the FDA would ever compel a company to go with one manufacturing process over another — particularly in instances like Sarepta’s where comparability between two processes raises concerns.
Though the answer is a straightforward no seeing as the FDA’s recommendations are not legally binding, I appreciated the thoughtfulness and detail the FDA representative offered in his response — along with his friendly reminder that we alone are the experts on our therapies:
“The FDA does not have a specific recommendation as to how a sponsor chooses to manufacture their product. Part of the reason as to why it would be difficult to mandate a manufacturing strategy is that every indication is different; every route of administration is different. Currently, we do not mandate any applicants to use, for example, any specific vector serotypes or methods, nor do we mandate the extent of the empty or full capsids. We look at the data in its totality, assess the therapy’s safety and efficacy, and consider its overall risk vs. benefit.”
Every adcom meeting provides a valuable opportunity for all of us and the FDA to step outside the silos of a functional area and see the whole picture of a therapy. Making the case for a therapy to the FDA is never light lifting, and it certainly was not for our valiant brethren at Sarepta — especially considering the FDA almost didn’t convene an adcom meeting to discuss Sarepta’s product in the first place. But I’m incredibly grateful this one transpired. For all of us striving to find our way out of the less well-lit corners of the current CGT regulatory CMC “black box,” there really could not have been a better case study than this.
Miss Part 1? Check it out here!
Like what you’re reading? Sign up to continue following ARW’s writings on the CGT + mRNA manufacturing paradigm.