2020: the year the FDA ate our field for lunch (and why)

by Katy Spink, DHC’s COO and Managing Partner
Addendum below* added on 1/5/2021

The end of 2020 didn’t arrive a moment too soon for most people I know, and the field of cell and gene therapy at large must have felt the same. 2020 seemed at times to be no more than a litany of one disappointment/rejection/delay after another for our young field. When taken in the aggregate, though, we can see a pattern, and once identified, I believe it’s a pattern that is relatively easy to correct. Let’s take a look at what this year’s C&GT issues have in common.

I’ll start you off with a hint: there has been chatter over the last few years concerning CMC as the Achilles heel of cell and gene therapy. Here’s what former FDA commissioner Scott Gottleib said back in May of 2018:
“In contrast to traditional drug review, where 80 percent of the review is focused on the clinical portion of that process, and maybe 20 percent is focused on the product issues, I’d say that this general principal is almost completely inverted when it comes to cell and gene therapy . . . The more challenging questions relate to product manufacturing and quality . . .”

That starts to sound like an accurate prediction when we look at the percentage of highly-publicized late stage review issues during 2020 that point back to CMC-related concerns. Even two delays that at first blush appeared to be based on clinical data are, when we look deeper, largely rooted in analytical comparability issues.

Let’s begin by considering the four that were clearly and publicly acknowledged (in the press coverage) to be CMC-based issues:

• First, BMS’ ide-cel (CAR-T gene modified cell therapy for myeloma): The FDA issued a Refusal to File in May because the CMC section “lacked the necessary detail to process the application.” Although BMS provided very little additional information about the nature of the missing data, according to an 8K filed by BMS’ partner Bluebird at the time, it seems to have related to process controls and validation for both lentiviral and drug product manufacturing. The amount of detail required by the FDA in this section appeared to take the sponsor by surprise, given that “the company was not aware that the FDA had questions about the level of detail in the CMC module until we received the Refusal to File letter.”
• BMS’ frustrating year continued in the form of the one-two punch dealt to liso-cel (the CAR-T gene modified cell therapy for lymphoma): In May, an announcement delayed the PDUFA date by three months (from August to November) as a result of a ‘major amendment’ of the BLA post-filing and acceptance. It is unclear if additional information in the amendment was CMC-related, though proximity and similarity (‘additional data’) to the ide-cel announcement a week later does cause me to wonder if the two issues could be related. By the end of the year we saw a 483 issued for deficiencies at BMS’ Bothell, WA plant (483 was issued in October but not made public until December) and delays to an inspection of their Texas CMO lead to further delays in FDA’s decision-making beyond the November PDUFA date. (This series of events has been discussed in the press most often through the lens of Celgene shareholder concerns regarding the deadline for $9/share Contingent Value Rights (CVR) tickets but I prefer to remain focused on the “why” behind the issues.)
• Moving on to Sarepta’s SRP-9001 (a AAV gene therapy for Duchenne Muscular Dystrophy): The FDA requested an additional potency assay during the Type C meeting in September. At the time, it was unclear whether this issue would delay the initiation of a planned large, multicenter Phase 3 trial. Subsequently, though the sponsor claimed to have resolved the issue with the FDA, they simultaneously changed plans to run a much smaller open label study “intended to accelerate Sarepta’s validation of commercial process material.” The change was cited by the sponsor as being due to “the risks that the ongoing CoVID-19 pandemic posed to the execution of a large, multi-center trial,” which is certainly a valid concern, but it does raise the question of the degree to which this change in strategy may have been influenced by the need for additional validation of the revised potency assay strategy before embarking on a large pivotal trial.
• Last but not least, Iovance’s Lifileucel (TIL cell therapy for melanoma): In October, following a Type B meeting with FDA, a delay in the BLA filing was announced, pushing it from 2020 to 2021 based on “the Company and the FDA hav[ing] not been able to agree on the required potency assays to fully define its TIL therapy, which is required as part of a BLA submission.” [Note the common theme on these last two of product potency.  Spoiler alert: this is won’t be the last time that issue comes up in this article.]

That sums up the clear CMC-related issues. Now, onto two situations that were viewed primarily in the press and by the public as clinical rejections, but which both, upon closer inspection, have a significant CMC component to them as well.

Mesoblast’s Ryoncil (MSC cell therapy for GVHD) was rejected by the FDA in October over apparent clinical issues, with the FDA requesting an additional randomized controlled trial. A careful reading of the FDA commentary, though, suggests that the clinical data might have been considered adequate despite its weaknesses…had the sponsor provided better characterization of product CQAs and their link to clinical effectiveness, especially as it relates to comparability between product lots derived from different donors. Among the evidence for this interpretation are:

1) The FDA dedicating the entire morning session of the Advisory Committee meeting to a discussion of “Product Characterization.”

2) The relatively scathing comments in the FDA’s briefing book for the AdComm, e.g.:

“FDA’s position is that the product attributes the Applicant has identified as related to potency and activity, however, do not have a demonstrated relationship to the clinical performance of specific DP lots” and

“the lack of CQAs related to clinical effectiveness limits the ability of the Applicant’s analytical methods to demonstrate that DP made after these changes maintains the same potency and quality as DP made before these changes”

3) The description in the briefing book of two potential paths to approval (I am paraphrasing, here) being either the use of a robust, matrixed approach to potency with clear linkages to hypothesized MOAs or a clear demonstration of efficacy and a well-controlled manufacturing process, even if MOAs are not clearly understood.

Taken together, the implication seems to be that Mesoblast may have been able to overcome the weaknesses of its clinical data had it had a more robust approach to product characterization, especially as it relates to potency.

The final example of CMC as C&GT’s Achilles heel in 2020 is BioMarin’s Roctavian (AAV gene therapy for hemophilia A). Roctavian was rejected by FDA in August, citing a need for longer-term data from the Phase 3 study to support approval. However, the company and the market appeared to be expecting that the long-term data from the Phase 1/2 study would be sufficient to support approval, as we saw in the CEO’s comments implying that the FDA “moved the goalposts” on this: “We had an agreement with the FDA over a year ago…regarding the hurdle we’d have to pass for accelerated approval,” and “Not only did they move the goalpost for accelerated approval, but they’re now apparently moving the goalpost for full approval.”

One possible (perhaps even likely) conclusion is that the FDA’s apparent change of heart regarding willingness to rely on the Roctavian Phase 1/2 data to support long term efficacy was the result of seeing something in the data that raised concerns about potential lack of comparability between the Phase 1/2 material (manufactured at a CMO) and the Phase 3 material (manufactured in-house; apparently with process changes to starting materials, cell culture and purification methods). We don’t have a clear window into whether that comparability hurdle was addressed, but it would certainly fit the expectations of Occam’s Razor.

I also find it intriguing that at least four of the six examples above (the two BMS products being the possible exceptions) clearly relate to product characterization (analytical) considerations, with particular emphasis on matters of product potency. Potentially, ide-cel may have also been at least partially analytical in nature, considering the ‘process control and validation’ issues that were cited. (Full disclosure: I am not without an interest in this possibility, given that Dark Horse Consulting is ideally suited for providing the analytical cross-checks that savvy sponsors are currently considering. We even released a recent case study along these lines, which can be found here.)

I know I’ll be watching the industry very closely for further signs of this trend, or work that has been carefully undertaken to avoid this trend. Even though it was a frustrating year in so many respects, I take comfort in the degree to which we’re seeing patterns, and manageable ones at that. We know how to solve these problems and, if we’re brutally honest with ourselves, we can acknowledge that at least some of these cropped up when external expectations collided with a lowering—intentional or not—of internal standards. More robust analytical reviews, earlier in the process, will go a long way towards identifying CMC vulnerabilities…and will, to my mind, offer a path to resolving them. A robust analytical strategy would be a…ahem…“potent” addition to any sponsor’s product development plan to avoid falling victim to similar issues themselves.

* ADDENDUM, 1/5/2021
Everyone who is following C&GT news (likely all of you reading this article!), will have seen the BMS news this week.

We now know that a plant inspection of the Lonza Texas facility was, in fact, undertaken in early December, but due to shortfalls identified in that inspection, FDA did not approve. Chalk another CMC disappointment up to 2020, I suppose.

This news primarily received attention in the media as the final nail in the coffin of the Celgene CVRs (to the tune of $6.4B for investors), but more importantly, it represents yet another disappointing delay in bringing a valuable and potentially life-saving treatment to patients. I can only hope that the resulting media attention on the relatively ‘unsexy’ topic of C&GT CMC will cast a spotlight for both sponsors and investors on the need to do better in the future.

‍