Q&A highlights: Decentralized manufacturing: enabling scalable and cost-effective point of care for cell and gene therapies
The transformative potential to change the way human disease is treated using cell and gene therapy has motivated rapid technological advancements towards streamlining manufacturing processes. In this presentation, Qasim Rafiq, PhD (Associate Professor, University College London, UK) and Xavier de Mollerat du Jeu, PhD (Director, Product Management, Cell and Gene Therapy, Thermo Fisher Scientific, MA, USA) discussed how standardization via automation and introduction of quality control via closed system manufacturing will contribute to overcoming barriers and facilitate the reach of this treatment to broader patient populations.
Watch the webinar on-demand now or read on for highlights of the Q&A session following Qasim and Xavier’s presentation.
What type of in-line monitoring are you thinking about? Are you doing functional potency looking at growth? What kind of tools are you using right now?
Qasim Rafiq (QR): At the moment we are using the traditional sensors that you would find within a stirred-tank bioreactor, such as pH, dissolved oxygen, dissolved carbon dioxide and temperature. What we are starting to see, and this will be certainly important for decentralized manufacturing moving forward, is the emergence of new technologies and innovations in the process analytical technology space, for example capacitance probes. We are working very closely in collaboration with one of the leading sensor companies, Adwa Instruments (Szeged, Hungary), to look at capacitance sensors. In addition, what I would love to see is the emergence of online glucose monitoring and lactate monitoring, and more broadly metabolite monitoring.
Beyond that, one of the things we are looking at is the use of Raman spectroscopy for ideally non-invasive online monitoring of key features, whether that’s metabolites, cell density and so on. I think if we could get to the point where we can measure functionality, or even a phenotype, that would be ideal.
Should we be actively developing technology to support decentralized manufacturing? More specifically, would this technology be different between centralized and decentralized?
QR: Very good question. From my perspective, I think we absolutely should, and I don’t think we need to classify a centralized or decentralized technology. Although I think there will be a preference depending on the manufacturing paradigm for research in technology, clearly we are not going to need a 20,000-liter stirred-tank bioreactor for a decentralized patient-specific therapy. But what we are starting to see now, and it is great to see, is organizations like Thermo Fisher Scientific and others starting to innovate specifically for the cell and gene therapy space. For me, that is probably the most exciting thing and I think it will drive the industry forward. Up until the last couple of years, there hasn’t really been an incentive for large technology companies or SMEs to innovate for cell and gene therapies because there just really wasn’t a market need. But now we are starting to see several approvals within this space; we are starting to see the technology gaps. As I identified in my talk, we have large biopharma technologies and we have very small-scale, non-automated, poorly controlled technologies such as T-flasks and culture bags. However, we are now starting to see technologies being specifically developed for cell and gene therapies. I think primarily we had autologous in mind, so you are seeing a technology that Thermo are producing, and many others, that will support cell and gene therapy workflows, and that for me is where I believe we will move forward. That will certainly develop as we start to consider a decentralized manufacturing approach.
CAR-T has been promoted for hospital exemption studies in the EU. Is such a development beneficial to progress CAR-T development, or is it risky for patients, or both?
QR: From my perspective I think that was always inevitable. One of the issues in the EU is the use of hospital exemptions and we have seen with certain therapies it has proven commercially disastrous where hospital exemptions are employed.
However, I do think that hospital exemption in the grand scheme of things will benefit CAR-T development moving forward, and that will generate significant patient data. This harks back to what I mentioned in my presentation where even in the USA we are seeing a big push from academic medical centers to be able to be the ones who are manufacturing these therapies on site in a decentralized fashion to deliver point-of-care production.
Ultimately, there is a concern about the risks and the safety aspects of such an approach. For example, how cost effective will that be and whether it would be worthwhile spending that money elsewhere. But ultimately, from a clinical CAR-T perspective, I believe it will be beneficial to generate that data. However, one of the issues with hospital exemptions is that we must consider the impact it has on commercial outcomes and outputs.
I am always worried that hospital exemptions may limit commercial innovation and drive the sector towards a point where it is not sustainable because you can just deliver many of these therapies via hospital exemption.
For decentralized production models, what functions, if any, are you looking at centralizing (e.g., a quality testing warehousing, QC, employee training, etc.)?
QR: That’s a great question and I think it’s maybe some or all of the above. There are a number of models being proposed and we have looked at a number of different models where everything is decentralized due to a more hub-and-spoke type model. Quality is certainly one of those where it may make sense to centralize testing and so on, especially when considering QC testing.
Employee training is an important aspect and one that, again, having a centralized mechanism for preparation and delivery, could be delivered centrally to ensure a certain level of quality and standards.
Warehousing is also an interesting one. If we start to look at adopting a lean approach, then ideally you want to minimize the amount of warehousing and therefore that is more amenable for a decentralized approach because you want to have minimal stock in place. However, lean is always a challenge for healthcare industries and therapy production perhaps because of how unstable some of these products are, in addition to the time critical nature of certain reagents and products as well. From our perspective, we have a range of models and in the UK, we are starting to see the emergence of advanced therapy treatment centers, which will provide an interesting model for how decentralized production, certainly within Europe, proceeds moving forward. We know the MHRA, for example, are very keen to and they have started engaging with stakeholders to establish workshops around the concept of decentralized point-of-care manufacture.
From the patient’s perspective, does it matter whether the drug is manufactured in a decentralized or centralized approach?
QR: From the patient’s perspective, I think the answer originally might be no. The patient receives a therapy, whether it’s manufactured in a decentralized or centralized production platform, system or approach, is perhaps not important. However, when you consider the opportunity to significantly reduce vein-to-vein production time, then I think it does. Particularly when you look at patients who are suffering from ALL or AML, these are patients in the end stages of the treatment options available to them. Their disease progresses very quickly, and I think being able to reduce the vein-to-vein time by a few days could be significant. Thus, from that perspective it does because it allows the patient to receive the therapy faster and therefore reduces issues around logistics and supply chain, and increases confidence that the therapy can be delivered as and when it’s needed.
In some ways it does have an impact on the patient, even if the patient themselves do not necessarily see the immediate benefit. In terms of being able to access the therapy sooner it could be a significant advantage.
You indicated that process control could be responsible for some autologous failure. Do you mean the failure wasn’t the quality of the patient cells at all, but poor process control of acceptable set quality, or that dynamic process control could compensate for the compromised performance of the patient cells?
QR: I think it certainly is the latter, so that dynamic process control can compensate from the quality issues that we see. But also, it’s a case where invariably and inevitably, we know that there will be differences between patient material – that goes without saying. However, I do think much of the variation we are seeing in these processes now, particularly where there is no process monitoring, or no process control/limited process control, much of that variation is actually because of our omissions, lack of process control or the use of undefined reagents. Therefore, my hypothesis or my belief is that once we inevitably see variation between patients, we can significantly reduce that level of variation down to a manageable extent by implementing effective process control. Where there would naturally be a variation, we can reduce that using process control.