Bodil Willumsen: Head of EU supplier relationship team at Novartis, responsible for lentiviral vector development for Kymriah from first Novartis patient through market approval to post launch change. Later years venturing into AAV gene therapy, w Luxturna and lately w consulting jobs related to AAV.
CDMO selection
1) What is important in CDMO selection and what has changed since you started the CMO selection process for Kymriah in 2013?
2) So, what would be your recommendation for companies looking for a CDMO today?
Tech transfer to CDMO
3) Once you have selected a CDMO, do you have any recommendations for how to go about the tech transfer, particularly if you are a startup with few people?
4) Lastly, based on your experience, what are some of the keys to success and smooth development?
Transcript
Fabio:
Hi everyone, I’m here with Bodil Willumsen. She used to be the head of EU supplier relationship and Novartis, so she was responsible for lentiviral development for Kymriah from the very first patient treated all the way to market approval and post-approval. And then lately she was also involved with AVV gene therapy with Luxturna and also with other consulting projects. So I invited Bodil because she’s an expert in CDMO selection and that’s a very important issue that most, if not all, cell and gene therapy startups have. So Bodil, thank you for being here.
Bodil Willumsen:
Thank you, it’s a pleasure. Glad that we could find a time that worked out.
Fabio:
Thank you. So, as I said, CDMO selection is very important. What is important when it comes to selecting the right CDMO? And has anything changed from when you started with Kymriah in 2013? So it’s basically 10 years now.
Bodil Willumsen:
It’s a long time ago. So yes, in 2013, the cell engine was obviously in its infancy. And there were very few CMOs that were offering services in that field. For instance, for plasmids there were a few, but they were not manufacturing to GMP at that time. So we had to work with generally our manufacturers to help them enhance their quality standards. For LentiViral Vector that I was also responsible for, there were two companies working with Lenti basically, and there was a few academic players. So… What mattered to us when we selected our CMO back then for the Lentiviral Vector was first of all, their technical capabilities and their IP position. But I also would mention that what was very important was maybe the more soft skills, their openness to the collaboration, their willingness to invest upfront to show that their technology could be used for our purpose. And also their flexibility in offering a reasonable contract that was a win-win for both parties. That was very important. I mean, this was a relatively small company at the time and we made a leap of faith to work with them and it just worked out very well. So for me in my work history, this is a prime example of how important it is to select the right CMO. So nowadays the field has matured and there’s a lot more CMOs in this space. And there’s been a lot of talk about the limited manufacturing for instance and also in the cell field. And so a lot of new players have come up. Some of them have a track record, maybe a good one, maybe a bad one. Some of them may not have a track record at all. There is a lot to consider. There could also be some of these CMOs that are established in the antibody field. Then they’re opportunistically trying to join this new field of cell and gene. You need to think carefully before you select.
Fabio:
Excellent. So, what would be your recommendation to selecting one if you are in a start-up, in a small start-up, at the beginning of the journey? What would be your recommendation today, given that there are way more players and probably overload of information and you don’t know where to start?
Bodil Willumsen:
Yeah, I think one of the most important thing is whether you have your own process that you want to move forward with or whether you want to rely on a platform process from a CDMO because if you need a platform, of course, that limits the number of CMOs that you can choose. Often it will probably be a hybrid scenario where you have some process but it needs to be adjusted maybe to the scale and to the industrial conditions and to the quality needed for clinical. So that I think is a very important one and then there could be a long list of things. I think that of course the expertise and the track record as we just talked about. As there are so many new players in for instance the AAV field, do you want to be the first one in that field for that CMO? Maybe not. Or maybe you’re willing to give them a chance if they’re otherwise the right fit, if they have a good track record in adjacent fields, maybe for other vector types and so on. That would be one thing. Then of course, lots of other technical requirements such as bioreactor type and size, analytical capabilities and so on. For a viral vector there might be time savings if you select a CDMO that can also offer other services that could be the plasmids manufacturing or it could be the fill and finish. That might offer some time savings and you don’t have to work with so many different CMOs. You also have to consider their manufacturing capacity. How many lines do they have? How busy are they? Do you need to commit to your manufacturing slots way in advance? Sometimes you do, at least in the past. That has been my experience, that you need to lock in your manufacturing, let’s say six or 12 months in advance. And so you wanna talk to the CMO and understand what are the conditions and what is the flexibility if you need to shift your manufacturing slot. Of course there’s price. Price is such an easy thing to look at, but it should never be the overriding factor. Because later on if you want to change your manufacturing process or change to a new CMO, the cost of doing that is so much higher than maybe paying slightly more for your first batch. Timelines of course, how, as I said, when do you need to book your slot, how much, how far in advance. The quality track record, the size of the CMO. If you’re a very small startup, you might not want to work with a big rigid CMO, because it might not be a good culture fit. Or… you might choose to do exactly that because they know what they’re doing and they have this good well-tested platform. Finally, I don’t want to underestimate the collaboration style. How willing, how flexible and how communicative are they? These kind of soft things that you may want to go visit them before you can really truly assess. And then, of course, there can be the geographical location. But again, I wouldn’t say that this is a super important one. So there is a lot to consider probably a lot more than what I just listed and you may want to actually liaise with a consultant who knows the field to make sure that you do the right selection upfront the best that you can.
Fabio:
Excellent. Well, thank you Bodil. So now I’m thinking still, assuming you are in a startup, imagine in a startup with a few people at the very beginning, and you might have a process, as you described sometimes it’s a bit of a hybrid, but once you selected the CDMO, maybe… consultant, what would be your recommendation in terms of tech transfer? Because then you are still a small startup and you still need to transfer your process to the CDMO that you selected. How do you go about it?
Bodil Willumsen:
Yeah, thanks. So you may want to consider, you probably often want to consider to do a quality audit upfront. The case where you might… leave that out is if you are sure that they have a very good quality track record and they’ve been inspected by various authorities, FDA, European authorities, whatever, then maybe you can leave it out but you definitely want to make sure that they have a good quality system in place. Then you need to set up the team and for me essential is that you have a project manager on either side that function as the single point of contact companies and they will speak often, they will set the timelines and the goals, the well-defined goals and they will set the team size of the team can vary a lot from one person to 10. I don’t know if you have experts that you want to include in manufacturing, maybe quality, maybe regulatory, then the teams need to be defined and the experts in both companies need to speak directly to each other and I would recommend that in email conversation that they always include the project manager on there. It’s just to make sure that there is one person who has the complete overview on either side. Then I would always recommend beginning with the end in mind. And what I mean is, if your company has a research process, a process you made in your own lab, be open about it and allow time for the CDMO to test it at small scale before proceeding to full scale and to make adaptations. You may think that it works well but it’s something else to do it in routine settings and to high quality standards. So do allow time for the CDMO to do what they’re good at and to adapt the process before you make your GMP patch for phase one. And I know this can be difficult. Believe me, I know because there’s always time pressure in traditionally CMC. There is always this pressure from the other parts of the company to do it as quickly as possible. But my recommendation is to leave time and to make sure that you have a good process in place as much as you can. Also You may want to consider the manufacturing line that you use if you have the luxury of choosing, because change of manufacturing line later on will require comparability studies. So it will be costly and take time. So if you’re able, even if you use a different manufacturing line afterwards that is right next to the one that you used and uses the same single use equipment, you still need to do comparability. That’s my experience. If you can start in the right line that you can also use later, maybe even into commercial, but at least later on, that will be an advantage.
Fabio:
Avoiding having to change the process. Maybe some changes would be inevitable but you want to reduce it as much as possible. I guess there is even more pressure when it comes to CMC in cell and gene therapy compared to other traditional biologics because it’s so much easier and so much more kind of established with other products that people underestimate that when it comes to cell therapy in particular it’s very challenging.
Bodil Willumsen:
Exactly.
Fabio:
Sometimes it just cannot be done fast enough, as you would normally do with another antibodies or another kind of traditional biologics. So I understand. And with that in mind, based on your experience, what can make the process and this tech transfer smoother and more successful?
Bodil Willumsen:
Yeah, thanks. Some of the examples that I’ve been following over the years. So several of the CAR-T players that made it to the market have had trouble delivering enough treatments to the market after they had approval. So many have of course struggled to build out their cell manufacturing capacity either locally or around the globe. And this includes Novartis, of course. But there are also several examples where the lentiviral vector was the limiting factor in not being able to meet the demand. And so while there’s been a lot of debate, as we discussed, and there’s been a lot of talk about limited manufacturing capacity for vectors, I don’t think this is the issue in these specific cases. I think the issue here was that once you launch with a specific, maybe low-yielding vector process, it takes years to be able to change. First you need to develop a better process. You need to go through all the same motions that you did already. You need to validate. You need to do PPQ runs, do analytical bridging, do comparability studies, including comparative stability studies. There’s a big package that you need to put in place. It takes time. And then you can start your regulatory process. And in let’s say Europe or US, you may be able to get some changes approved in half a year. There can be other countries where it takes a year. So this whole thing about changing your process after you launch is a big demanding process that takes years. It’s very lengthy. So you’re really better off launching with a more mature process if at all you can, as we discussed. So in essence… What is important is to select the right or the best CDMO that you can with the best technical capabilities and with the most collaborative people as well. Then… Again, you need to allow the CDMO some time to adapt the process to make it more robust, to make it more industrial. Keep in mind that what may have worked well in your lab may not work in an industrial setting and may not produce good GMP material for clinical studies. This is definitively also the case for cell processes such as CAR-T, that it may work well with a few highly skilled technicians in your lab, but if you go to a bigger facility with more different people, it may no longer work because it’s a tricky process. If you find, as we said, that you need to change your process, and you probably always will, but do it as soon as feasible. Do big changes if you can before phase two. The later you make the changes, the higher the bar will be for implementing it. And overall, as you also alluded to, CMC has a more dominant role in cell and gene therapy than in classical biologics, because the biology is just so much more complicated. Of course, it’s getting more and more of a commodity, but it’s not there yet. Right. It’s not at the stage where antibodies are. So do make a point of negotiating of negotiating with the project team and with your clinical and translational medicine colleagues or with your management time for the CMC development upfront in the early phases of your drug.
Fabio:
Excellent. Well, thank you Bodil, thank you so much for this very interesting interview and for sharing your experience with us. I’d like to thank you and I look forward to our workshop where we can learn and deep dive into how to select a CDMO and how to go about the tech transfer process.
Bodil Willumsen:
Indeed. Bye. Thanks.
Do you want to be the first one to know about new interviews, training courses and other exclusive opportunities? Register here