Blog | June 30, 2023

"Our CMC, Ourselves:" 3 Iterative [Self] Improvements In Cell & Gene Manufacturing

Source: Cell & Gene Collaborative
Anna Rose Welch Headshot

By Anna Rose Welch, Editorial & Community Director, Advancing RNA


A few months ago, I spent a week in London. Each time I ended up on “The Tube,” I found myself ceaselessly charmed by the poeticism of the phrase “mind the gap,” which was repeated every time the train opened its doors. Hence why I couldn’t help but smile when I was met with this phrase once more during a presentation, “‘Mind the Gap:’ Regulatory Considerations for Clinical & CMC Product Development” at the recent PDA ATMP Conference.

In the CGT CMC space, we know the importance of keeping our friends close and our clinical teams closer. Our therapies’ reputations in the eyes of the regulatory agencies, patients, and treatment centers depend upon our operational seamlessness. But there are many well-known reasons as to why CMC often ends up falling behind clinical development. Overall, the PDA ATMP conference agenda carried us nicely through the challenges that broaden the gap between CMC and the clinic, including analytical development, vendor and supply-chain management, and/or raw material control.

However, the discussions throughout the two days also emphasized the ways in which we can narrow the gap between who we are as an industry/what we know today and who we hope to be/what we hope to know in the future. We’ve heard it said many times that our CMC efforts are — or at least should be — an iterative effort. But I loved the way the FDA’s Tiffany Lucas instead chose to define the iterative work we do on manufacturing as akin to the work we do on ourselves.

“We are always going back and making ourselves better,” she explained. “If you’re one of those people who naturally asks, ‘How can I make myself better?’ and regularly evaluates what you can and cannot control, then CMC is the perfect place for you to be. These are the questions we at the FDA want you to ask and understand about your manufacturing process.”

Here, I’ll outline the three areas in which I felt the discussions at the PDA ATMP conference revealed how we can better “close the gap” between where we currently are and where we know our manufacturing processes (and our brains) can be in the future.

  1. Closing The Gap Between Our Minds & Molecules’ Structure/Function

When we’re an expert at something, we often use the phrase, “I know this subject/topic like the back of my own hand.” In the ATMP space today, we certainly hope we’ll be able to say the same thing about the structure/function of our products someday. Though it’s unrealistic to expect we will/can know our predominantly non-commercialized products like the backs of our hands today, the FDA still expects us to exhaust every analytical method at our disposal to get to know our products ASAP. And yes, this means we will — and must — encounter (and share!) the information we have no idea what to do with yet. In other words, navigating our current process- product- & function labyrinth and emerging smarter on the other side involves being analytically fearless. But it takes a lot more than just being gung-ho and willing to embrace the analytical challenges and uncertainties. Part of this fearlessness means planning for the failures and/or inconsistencies that can and likely will arise from insufficient investment in analytical development; undesirable CDMO change management scenarios; jumbled analytical transfers; and/or retesting limitations. (#GotRetains?)

What I most appreciated about the PDA ATMP conference was the realism with which the speakers addressed the analytical hurdles we’ll no doubt encounter. As the FDA’s Mikhail Ovanesov explained, many small companies today don’t have the financial luxury to invest enough in early analytical development, standards implementation, and method validation. But as we often say across the entire CMC space, we can spend a lot now or spend way more later after something goes wrong. As Ovanesov advised, relying on poorly selected and/or uncontrolled analytical methods in early clinical development could limit our ability to rely on early-phase clinical data in our BLA, unless we want to retest these earlier lots using our BLA-described commercial methods. Similarly, should the FDA have lingering questions about the suitability of our methods upon approval of a BLA, we may find ourselves facing some difficult post-approval compromises. For example, we may need to achieve and demonstrate narrow release or in-process control limits that are not reflective of our existing manufacturing capabilities.

In addition to the costly and/or strategy-busting issues that can come from insufficient attention to analytical development, the conference also homed in on the analytical challenges of navigating outsourcing relationships. We often fixate on our partnership entrance strategies — and having a waterproof technical and quality agreement in place is, of course, essential. As Stephan Krause of BMS explained, we need to thoroughly research and analyze our potential partners’ analytical platforms before signing on the dotted line. After all, many of these platforms were most likely not built based on cell and gene therapy products. Given the novel intricacies of CGT products, it’s also exceptionally important we dictate within our technical and quality agreements the transparency and input we as sponsors expect throughout development — particularly when it comes to change management and emergency reviews and resolution processes.

However, though we may be true romantics at heart and want to believe our first CDMO will be a long-term relationship, we should never forget to have an exit strategy in place. In fact, as Krause advised, we should always enter a new CDMO relationship with a healthy dose of realism. Assuming our next CDMO will be better than the previous may not be the most helpful mindset. Rather, we should be thinking carefully and realistically about how such a change may impact our analytical development program. This means considering the similarity or sameness of testing methods; understanding the need for any retesting/method comparison studies; and planning in advance for any clinically meaningful changes in quality via a new platform that may require a specification revision.     

  1. Closing The Gap Between Our Suppliers & The Supply Chain of our Dreams

I’ll admit it was a little weird to type the phrase “the supply chain of our dreams.” Despite its critical importance and its impact on product quality and development cost, it never ceases to surprise me how often this topic falls by the wayside in industry discussions. That’s why I found it refreshing that supply chain management, optimization, and raw material control/risk mitigation got some much-deserved time in the spotlight throughout the conference.

The problems with supply chain management are well broadcast. Over the past few years — and to this day — our industry continues to grapple with supply shortages and a lack of high-quality GMP or dual-source GMP supply options for critical materials. But if there was one point emphasized throughout the many discussions, it’s that getting our supply chains up to speed doesn’t have to be — and shouldn’t be — a waiting game. We need to take a more proactive approach in our evaluation of risk for each raw material as we progress through development. As Ryan Murray of ValSource explained, it’s not unusual in a sole-source environment for us to grin-and-bear-it and assume we’ll just “figure out” how to alter our tricky supply reality later. While taking such a risk in early development may be the best — or only — option available at that moment, it’s important to remember that our understanding of our process and how each material impacts said-process becomes more thorough over time. In turn, our risk assessment of each material and our relationships with our suppliers should evolve to reflect that additional data/experience.

As several speakers went on to explain, we as sponsors shouldn’t hesitate to encourage our supplier partners to “professionalize” their own processes by investing in GMP production and breaking away from the high-volume mAB-style approach to material supply. Just as we tend to gravitate towards the mAB model of “always be scaling-up,” raw material supply for ATMPs tends to be placed into the same cookie-cutter style formula. As WindMIL Therapeutics Kimberly Noonan argued, we must keep working with our suppliers to reduce the scale at which they produce and ship raw materials. Not only will “right-sizing materials” for smaller batch ATMPs enable us to automate our processes more seamlessly, but it can also cut down on material waste — a factor that ties nicely into… 

  1. Closing the Gap Between Our COGS & Our Patients

This is an evergreen topic of discussion in this space. But I daresay it’s even more topical today given the $3.2 million price tag for Sarepta’s newly approved DMD gene therapy. As you may have already guessed, our raw material costs account for a large portion of our COGS. In fact, for a cell or gene therapy requiring plasmids, materials can account for upwards of 60-70 percent of a product’s COGS.

Though we often talk about the cost-savings of closing and automating our processes and reducing our labor needs, Genentech’s Andy Case argued that raw material optimization can be just as, if not more, impactful in reducing your final COGS. In many cases, only 10-15 materials are the greatest cost generators on our bill of materials. As such, our supplier partnerships should not only be centered around improving the availability/scale/format, and quality of these critical materials, but also around developing a roadmap for bringing costs down through development. As Case pointed out, there’s typically no going back cost-wise should you wait to negotiate those costs after success in Phase 2 trials.

Overall, it comes down to thinking more wholistically about the role of your supply chain throughout your development. For those of us in process development, we’re focused predominantly on how to make sure a process works, as opposed to how our materials could function more cost-effectively within that process. Moving forward, we can get smarter earlier about identifying better, more cost-effective, and/or more reliably attainable materials that will result in a comparable product.

Though we typically emphasize the reduction in labor costs we’ll realize with automation, there’s also a lesser-proclaimed material optimization benefit to process automation. At face value, there will inevitably be high costs associated with pursuing automation early, particularly if we need to engineer custom materials and/or rely on a single use assembly that also requires us to adhere to restrictive (and expensive) supply agreements. But as Case added, automation enables us to better monitor raw material quantities and waste less, seeing as our materials can be used across multiple batches. Couple greater material usage efficiency with a reduction in labor costs and fewer manufacturing failures, the long-term financial benefits of earlier implementation can outweigh the initial costs.  

It may seem excessively humbling to approach our plans/strategies with just as much skepticism as hope. But throughout the conference, the phrase “be honest with yourself” was regularly repeated, suggesting that a large part of closing the gap between where we are today and where we want to be demands we carefully identify and take ownership over what can go wrong from the get-go. As the FDA’s Lucas reiterated, “There’s a lot of self-reflection that goes into this risk-management approach. ‘Where could potentially go wrong? How are we controlling and/or investigating these problems? Where am I relying on someone else to manage these problems?’”

In fact, such self-reflection may demand an additional shift in mindset, particularly when we’re approaching the FDA to discuss our approaches/strategies. As one speaker emphasized, it’s just as important — if not more important — to ask our reviewers what could potentially go wrong with our strategies, in addition to whether they agree with them or not. Admittedly, we may not get the answer we want. We may not get an answer at all. But if I learned one thing from the PDA ATMP conference, it’s that we still have plenty of room to grow.