Cell Therapy New Product Introduction: On Demand

Matthew M. Hewitt (00:08):
Just brief summary in case anybody's not aware and we will make these slides available after the talk. Feel free to take pictures or you can wait for the deck. Charles River currently has, this is a little bit out of date, we're over 120 sites in 23 countries. Last check in '21, we were a part of 86% of all therapeutics approved by FDA. Within oncology, that's 93%. We have various different business units that work within the company to provide an end-to-end solution for COO and CDMO, as well as the testing required to release them.

(00:42)
The vision is obviously centered around patients and whatever services we can bring to bear across all modalities. Of course, we're here today to talk primarily around advanced modalities in cell and gene therapy, but ultimately this is all about the patients for us. Now, there will be a test at the end of the talk, so please read all of this. Don't listen to me, just read. But what this is telling you, is a quick overview of our advanced therapeutics portfolio and how it sits today.

(01:11)
We also have cell solutions business that provides us with cells. We always say it's hard to do cell therapy without cells. These are research leukopaks, GMP leukopaks, mobilized, double mobilized, and so on. Isolated material, disease state material. We have gene therapy manufacturing, which includes plasmids from RUO all the way through full GMP, including off the shelves for [inaudible 00:01:37] and soon to be AV. We also can do obviously the transfer plasmas for the GOIs, viral vector manufacturing from RUO to full GMP. And then in cell therapy, I would say we have the typical experience with many CDMOs that have been doing this for 20 years across various different cell types in immuno-oncology, regenerative medicine, and other applications.

(01:58)
Our cell therapy facility in Memphis is commercially licensed, which I'll get to on the next slide. This is a quick map of not only our CDMO locations, but also our testing locations. Within CDMO we have about 800 employees, another 750 or so in testing, and then we're across United States, Europe, and we have some operations in Asia as well.

(02:22)
Last year our Memphis facility for cell therapy was inspected and granted a commercial license from EMA for both allogeneic and autologous therapy distribution in Europe. We just underwent some more inspections for a subsequent program, so we had EMA and FDA come, and we hope to be able to speak about that and the results of that, and those inspections soon.

(02:45)
I think hopefully everybody has been following the field this year. Up until the end of '22, we had around I think 12 or so therapies in total across cell and gene. That was since the first approvals back in 2017. This year, if we read the t-leaves for the ones that aren't yet approved, we're looking at probably about eight approvals this year in the US. To go from 12 to 20 in a single year, we're starting to approach this level that Scott Gottlieb gave this now I think famous infamous speech back in 2017, that by 2025 we should start to see 10 to 20 approvals per year. It seems like we're very well on track to potentially hit that. With almost 1700 clinical trials going on worldwide, a lot of those happening in North America, even if we fully risk adjust all those clinical trials, there's going to be a good number of products coming to the market soon.

(03:45)
I'll talk briefly about some of these articles that have come out recently. There's been several talking about the difference between CDMO capability versus capacity, and how we deal with that, and I think a system to deal with that in general. But first, one slide we like to show at some of these talks, is the different modalities and how they've evolved over the decades. We began, let's say long ago or not so long ago, primarily in small molecules and we were using chemistry to affect biology. Then we moved on to large molecule and biologics, and we were using elements of biology to modify biology. And now, we're in the advanced therapeutic space, we're using biology to modulate biology. We're getting progressively more complex as we can deal with progressively more complex diseases. That's what we see and it's also what leads into the capability versus capacity.

(04:36)
Now when you look at a CDMO, I think it's important to understand where you want your capabilities, versus where you want your capacity. Now, I want my capabilities to definitely be in development. I want my PD teams, my AV teams to have the necessary skills and experience to develop the proper process that's fit for purpose for your program. On the other side of that, the analytics, we want the capabilities to develop the right analytics and I would say in the grand scheme of things, a lot of times we end up with more challenges and analytics than process when it comes to trying to commercialize an advanced therapeutic.

(05:14)
But when you get into GMP, I think one of the last things I want is a bunch of capabilities in GMP. I want execution in GMP. There should be really no thinking, is what I always say. We should be brains and developments, hands of GMP. The brains think, they develop, the hands do and execute. If you start thinking in GMP, and I think that was Ali's question on the last one, you may run into some trouble because people potentially could always find a better way to do it. But when you're in a GMP environment and your entire data or existence in this case is governed by a batch production record, going against that can have consequences and there are proper ways to certainly go through change controls to do that. But so it's important to really talk to your CDMOs and ensure they have the proper capabilities in the right place and they have the proper capacity to execute your programs in another, whether that's early phase, late phase or commercial.

(06:16)
What's needed when we're starting to look at how to transfer programs in and what are the pain points? A lot of times pain points, the first thing that comes up is timelines. Everybody's timeline is yesterday. These programs move very quickly. When you start to get good data, there's always a portion where you really want to accelerate to take advantage of any competitive advantage you might have to get the market sooner than your competition. An example of this that's in the news, is Bluebird and Vertex are vying to be the first approved product for sickle cell. Their PDUFA dates are 12 days apart. I think the one that has first, which is Vertex on December 8th, would probably like to get approval on December 8th and then probably like to have Bluebird get approval on December 20th, just to have the bragging rights that they were the first. There is always that context of timelines.

(07:10)
GMP documentation, this is something that we spend a lot of time on when we do MPI and we will dive into this a little bit. A lot of times your problems can start and end with how good the documentation you have is in GMP. The one thing that I think we've lived by for far too long in this space, is the processes the product. That's a problem. If you have specific tribal knowledge that you need specific operators for, you will not only have trouble transferring that process to any CDMO or any facility, you will have trouble scaling that process. I'm sure everybody was at the morning chat where we're talking about the fact that only about 30% of patients at the moment who are eligible for cell therapy actually receive one. And so if we're going to actually scale, we're going to have to get away from nuance, and special travel knowledge, and we're going to have to be able to write it down in a batch record, and make it approachable for non-advanced degree people to follow.

(08:07)
When we look at this, we start off with the process development, which is obviously a key bit, and then move into analytical. Ideally, you're going to have a CDMO or a manufacturer with in-house QC. This is obviously very, very important for indications that are in oncology because just because we're manufacturing a cell therapy, does not mean that the disease stops progressing.

(08:33)
There is a very good poster that illustrates this from 2021, from Sheba Medical Center, from the European Hematology Association, where they followed one of the commercial CAR-Ts that is being manufactured here in the US, but then the patient is actually in Israel. They have to ship the cells from Israel to the US. They have to manufacture, release in the US, and ship back to Israel. And it turns out that the turnaround times for this for so long that a significant portion of the patient's disease were progressing to the point where they had to receive local therapy. Now you can imagine if they did not have any in-house QC and we had to ship it, you then run the risk of sample loss. You run additional timelines for shipping, ascension into the system, and then ultimately handling to get it done.

(09:21)
In-house QC is pretty important. Certainly, appropriate manufacturing platforms, this is something that's becoming more and more important. I think relating back to timelines, everybody wants to get to clinic quickly and we certainly see a lot of processes still going into the clinic in more of an open manual or semi-automated, parts closed, parts open processes. Now increasingly when you get the later phase and ultimately into the commercialization of therapies, regulators are asking us to close processes. It's a far better risk profile. It potentially makes them a little bit more robust. I've done some work on this myself in clinical comparability and it does against some of the open systems.

(10:04)
It's something to consider. I think there's no disagreement that we need more closed processing, it's a matter of when you implement it. That's got to be a discussion that you should have when you are a therapeutic developer and thinking about this, and also discuss it with the CDMO around what's fit for purpose if you're using one.

(10:22)
Strong quality is a key here in GMP. We want to make sure that we have the proper controls that are phase appropriate. There are going to be changes that need to be made, especially in phase one programs, as you're just getting your feet under you, versus say some of the more locked process analytical suites that you would have in late phase and commercial.

(10:43)
The new product introduction framework is critical for this because again, a lot of your problems can stem from improperly transferring things into GMP without proper documentation and activities. Because when you do that, there can be follow on effects with mistakes made by the manufacturing teams, which starts kicking out deviations to your quality team, and then starts to strain them. And then of course you need to have a framework and understanding of what you need to go to commercial and that sometimes comes with experience. That is one thing that probably only comes with experience.

(11:17)
This is a quick paper just showing that there are differences in biologics where we've had in monoclonals where we've had a lot more experience over the years just in a sheer amount of time versus CGT. The two things we see is there's differences in CMC issues. I can tell you in talking with the authors, about 80% of these are clustered in analytical versus process. That's not super surprising, so that's why we're starting to put a lot more focus on analytical.

(11:46)
On safety, this is likely pointing to the fact that the in vivo models we're using for the translational work are not perfect in translating in the humans. A lot of times we're using immune efficient mice. At Charles River we also have the CRO side, so we're continuing to work on both. Building better models on the in vivo side, as well as seeing what we can leverage on the in vitro side to not only further reduce animal use, but also hopefully improve the translatability of the clinic.

(12:18)
When you have the limited capabilities, this is primarily in your NPI. If you decide that you want to short circuit some of this, then this can be a problem because you end up... I'd say most of the time you're going to end up with some poor timelines, miss dates because you'll end up having to go back and redo things. You're probably going to have some challenges around increased costs as you continue to have to redo a lot of these activities. And then ultimately, it'll delay your time to clinic and ultimately have an impact on activating sites and ultimately dosing patients.

(12:53)
In the past, there hasn't been NPIs and then when we started putting in NPIs, it's primarily around documentation, assay qualification, and then transfer activities into the suite. I think that we're generally around 10 and a half months for this. We're looking to get this down a little bit shorter. This is primarily moving away from somewhat of the one-off cell therapies we see. Now that we've gained a lot of the travel knowledge, we can more or less bend out where a lot of these therapies typically go in terms of manufacturing, so CAR-T's, G-REX, a lot of times Prodigy now. And if you're in the Prodigy, we have a lot of experience with that. We've done that. We have the EPO set up, we have the full finish set up on it. That one we can just put it right in. It's a lot easier than doing some manual process development. And in that case, we can get this down even further, hopefully by a couple months.

(13:45)
This is yet another table which we'll have a test on, but the point here is these are all the activities that are happening to get you from onboarding and into GMP. A lot of this is around evaluating process upfront and not doing things in series, but doing things as much as possible in parallel.

(14:05)
What we have as an example, is we have two MSAT teams, one with our development team, one with our GMP team, and the GMP side is also split into tech ops. And so we do there is, we do a lot of side-by-side training on the process for MSAT and PD. We then start to bring in a couple of the operators for the manufacturing side to have them give some feedback on the process, as well as give feedback on the batch record generation, things like additional pictures, additional diagrams as needed in the BPRs. And then ultimately, this is going to smooth your transition into GMP because ultimately we have to go on the training runs and engineering runs in order to get into clinical.

(14:47)
Ultimately, the MPI is about execution, but if you don't have the proper capture, in terms of making sure we properly understand what we're taking in and then you don't have the proper planning, which is on the last slide, on activities, then it's very hard to execute in an organized fashion. The planning is as important as the execution. In this case, we put an emphasis on all three of these and ultimately to get everything into GMP much more quickly.

(15:20)
Again, this is just a quick run through of how we do the NPI. I'm just going to highlight a couple things. One is that the in-house analytics, in our case, what we tend to do, is our analytical team has two sides. We have an analytical development team that does all the development and then we have a side of an analytical team that sits in a GMP analytical lab and they'll also qualify the assay if it's part of a phase one, they'll validate it if it's part of a late phase, and then also commercial. This speeds up the time to GMP execution.

(15:55)
Then training and engineering, we have MSAT and manufacturing the same room, and then ultimately in the same suite as we're executing those runs. This is important that we have them collaborating with each other and doing that knowledge transfer upfront, rather than piecemeal at some point.

(16:15)
When you're doing the tech transfer, obviously there's a lot of SOPs that come out of this, as well as other documentation. I think we've already talked about facility fit a good bit, standard test methods. Testing plans is something that comes up quite frequently that folks don't think about. We want to make sure that the testing plan of the suite and the facility are robust, but at the same time we don't want to make it so complex it becomes essentially a Rube Goldberg machine and we just have our QCM teams in there doing EM monitoring all the time. This is a balance. And of course bombs, PFDs, the master batch records, et cetera, and the risk assessments are key for this.

(16:59)
Batch records right now, obviously we like many others are transitioning away from paper. It's just something we have to get away from if we're going to do proper review, in terms of lot disposition and release. A lot of our clients, if you think about some of these backed records can run several hundred pages and in some cases upwards towards a thousand, so whatever the number of pages is, multiply it by two because our QA has to review it and then their QA has to review it. Moving to EBRs is a huge thing for us, especially in the autologous space because if we're going to move to potentially tens of thousands of doses per year, there's no way to review that by paper. That's a huge priority for us and we're working with a vendor right now to move to that.

(17:48)
I think the KPIs that we're doing, and there's a big discussion right now similar to biologics that advanced therapeutics manufacturing needs to move away somewhat from fee for service and into somewhat of a performance based pricing manufacturing metric. And so that's what we're doing as part of the late phase in commercial, is we're setting up KPIs. We can set expectations for what's expected of us.

(18:13)
Now part of that in the beginning is going to be around what's the expectation to actually get your process into GMP? That's a critical one. Everybody's again, worried about on time delivery. Timelines are king always. What's the timeline adherence? Again, we keep hearing about timelines. We want to make sure we deliver on time. We want to make sure that the overall project stays on time.

(18:35)
Process flexibility, I think that again, we all go in with the best laid plans, as we've been talking about. Things change, we have to be a little bit flexible. We have to be nimble so that we can adjust as needed. And then ultimately, there's going to be tech transfers, at least from the client to the facility, and sometimes there may be more. We need to be able to move these processes around efficiently. And again, that comes down to the systems that we use and the processes we put in place.

(19:06)
Again, this is more around we want to go through a defined process to get you in the GMP. GMP readiness is the key. We want to make sure you have good process or robustness. If you can't consistently manufacture the product, if the GMP record isn't sufficient to the point where we could hopefully pull somebody in and have them read through it and understand what you're trying to do, and I'll touch on that in a minute, then we have to go back and do some additional modifications to de-risk it.

(19:38)
There are, again, coming back to the process of the product, we want to get away from nuance. There is some stuff that when developers are working on their process for several years, it's their baby. There becomes certain things that they say, "Well this is common knowledge." Well, it's common knowledge to them because they've been doing it for three to four years. But to be put in the hands of somebody new and then all of a sudden we don't have common knowledge, so that's where this nuance comes in.

(20:06)
I don't know if anybody has seen this video. It's about seven years old. It's becoming part of our training program, how to make a peanut butter and jelly sandwich. If you haven't seen it, look it up. It's a really good explanation of GMP and how we try to structure our batch records. In this video, if you haven't seen and you can't tell, the dad asks his daughter and his son to write him instructions on how to make a peanut butter and jelly sandwich, but he does literally everything they say. When they say put the jelly on top of the bread, that's what he does. He puts the jar of jelly on top of the bread. And of course they proceed to get very frustrated with him and they go through several iterations of this over about six minutes. And ultimately, he does make a peanut butter jelly sandwich. Not to spoil the end for you, but I wouldn't say it was perfect.

(20:56)
That's the point though, is that if we go by processes the product, we end up with a sandwich that looks like that. We have to take the nuance out of this process and put it into something that everybody can follow and it's a well-informed set of directions.

(21:14)
There's another big initiative, and I think it was touched on the last talk, a little bit around workforce development. I sit on several of the committees myself. We've been working at multiple levels on this and I'll touch on that more in a minute. Traditionally, GMP has been seen as needing a four year degree, generally a bachelor's. Now, we hire a lot of Bachelor's folks and they're great, but most of them don't have their life goal is getting into a bunny suit several times a week for four to six hours a day. That's just not in their long-term career plans. We believe, and we've already started to do this to some degree, that we need to increase our training and then subsequently lower the educational requirements for a lot of our manufacturing associates so we can access a larger pool of talent.

(22:04)
The example I always give for this and when people say, "Well, I didn't learn about soul culture until I was a sophomore or junior in college." And I say, "Okay, well if you have a job in the US, generally you have to go get a drug test." Usually, that's done by someone like LabCorp. I'm going to ask you to think about the people's educational level running those tests. They don't have very many four year degrees in there, yet we rely on them to properly process our drug tests and come back negative so we can get gainful employment. I don't see this as any different. I think this is a trade that we can teach people and I think that we get people interested in early in life.

(22:43)
The other thing we hear is, "Well you need specialized equipment." If you go to any technical center in the US in any county, they have HVAC, they have plumbing, they have mechanics. Last I checked, they need specialized equipment for that as well. This seems like something that we could do, even at the high school level. And to that effect, we've actually, at Charles River, near our Memphis, Tennessee facility, invested into a technical center in Memphis for high schoolers to begin to learn about processing and biological materials.

(23:11)
I think when we boil it down, what do I need for somebody to walk into the facility? I need somebody who knows how to gown in and out of the clean rooms. For anybody that's done it, it's not just like put a hair cap on, and lab coat, and in you go. It's a pretty complex process. You have to really practice it and be proficient at it. I need somebody that knows how to do aseptic technique and I need somebody who knows how to do basic soul culture. Two and three go together, but those are the basic skills we need for it to have people to be productive in GMP. Everything else is going to be very site specific. What's in your QMS, what's in your training records, your on-the-job training? But those three things are very portable for those folks and would bring a lot of value.

(24:02)
The way we look at this is... I have a boss and her MO, she gives me a book every year to read. It's never Twilight or something maybe fun. It's always she wants to teach me something. One year I got the Good Job Strategy and it's actually a good book if you get a chance to read it and have the inclination to, but it talks a lot about different ways to invest in employees.

(24:25)
And so obviously, we want to invest in employees, and empower them, and cross-train them, and everything else, like I think was talking in the last talk. But at the same time, we have to realize that many times in GMP manufacturing we do have a decent bit of turnover at the lower levels. The way that this book talks about mitigating against that, is having a really, really good robust training program. Because if you can have a really good robust training program and then the correct documentation that makes these processes and batch records more approachable, then if you do have folks leave, you can slot others in very quickly because you have the proper training. Essentially, you're turning people to somewhat into interchangeable pieces. We don't like to use that term, but ultimately, again, we're talking about stuff where we need to be able to properly train folks.

(25:17)
All right, so the last thing I'll touch on real quick, is we just had some new guidance come out and we're on a couple of the committees that are writing some responses to it. I think that we're making some good progress on this front, in terms of having FDA and EMA give us some pretty good direction around where the field's going to go. Some of it I think we're going to agree with, some of it we're not going to agree with. But ultimately, I think that it opens up a larger conversation, which is that right now the mechanisms for interaction are solely limited, direct interactions are solely limited to IND holders. But there are a lot of questions outside of single INDs that are quite interesting to answer, and I think some of them are actually being asked for some of the solutions providers in the exhibition hall here.

(26:07)
A lot of us want to do parallel manufacturing of therapies, not only of different patients, but of different therapies because ultimately if we're going to scale efficiently, and you're in a closed system, we have to figure out a way to more efficiently use our manufacturing capacity. And right now, as a CDMO in an academic institution, those two entities, we're pretty interested in doing that, but there's not a good mechanism for us to interact with the agency on this. I think that's some of the discussions we're having with them now is can we open up an avenue for this type of discussion? Because there is a way to do parallel processing across this, where we are able to analytically differentiate therapies just like we do right now for fit for purpose and fit for facility.

(26:54)
I think this is something that we'll continue to work on in the future, but I do want to highlight that we are thinking about when do we get out of the hundreds of batches per year, the thousands of batches per year, and into the tens and hundreds of thousands of batches per year? We're just not going to be able to do it by continuing to build massive amounts of spaces. We're going to actually have to get more efficient with our manufacturing capacity.

(27:16)
We've got a couple minutes left, so I'll close up by saying that Charles River, we have a lot of experience in this space, obviously, end-to-end, commercially licensed in the cell therapy space, soon to be in viral vector. We like to believe we're a good partner and if anybody needs anything, please let us know. And from there, thank you and I'll take any questions.

Speaker 2 (27:41):
We have a couple of minutes for maybe a couple of quick questions.

Speaker 3 (27:52):
You're talking about the risking and what I have learned in the last couple years, is that the biggest risk is the fact that this is a personal medicine and every person is different.

Matthew M. Hewitt (28:04):
Right.

Speaker 3 (28:04):
We should not say patient to patient variability. The nature of the patient is viable. What efforts is doing Charles River Lab with all those resources trying to relate the history, the path, the treatment, and the condition of that patient to the success rate of the treatment so you can do something to be able to respond to the reality of the patient material that is viable?

Matthew M. Hewitt (28:37):
In a prior life, I was at Lonza and the shiny little white pill thing out there, the Cocoon, I was part of the team that launched it. I ran the first clinical comparability on the Cocoon, so we were actually moving away. It was a CD19 CAR-T process that was used in Israel, and they had some really nice data, and they were always using G-REX. That was their consistent method, open manual.

(29:04)
We did the clinical comparability with the Cocoon and what we found when I did the clinical comparability was, is that when we did the mean or the average number of cells that were produced, your standard deviation or even the SEM was actually larger than the mean. And so what that typically tells you, is that you don't really have a truly stable process because you're getting a mean, but it's the result of having almost no cell production or just 10, 12 billion.

(29:32)
When we moved then into the closed automated system in the Cocoon, we actually saw that we produced a little bit lower number of cells. We still made dose several times over, but ultimately the standard deviation now went way down. What we start to see there is that whether we like it or not, when we do things in the open manual method, even if we do really good training, there are subtle differences from human to human. That's why you see a lot of effort, I wouldn't say to eliminate the human because I think that can lead to some sticky situations around regulatory, but I think certainly minimize the touch points that are required. That way we don't have this nuance.

(30:14)
That's the last bit of nuance, is to really move the closed systems. Again, I don't think anybody is arguing that that's not the case. It's just a matter of when you are going to implement that. There's some other analytics that we're doing around starting material to try to define that ultimate question, which is what is a good donor and what is not a good donor? That's the other side of it.

Speaker 2 (30:37):
Thank you very much.