The biopharmaceutical industry is increasingly interested in a range of emerging therapies. ‚ÄúWe‚Äôre really starting to get beyond the monoclonal antibody,‚ÄĚ said Patricia Seymour (senior consultant with BioProcess Technology Consultants) in her introduction to a lunchtime BPI Theater roundtable at the 2016 Biotechnology Industry Organization annual convention in San Francisco, CA, this past June. The discussion brought together three industry insiders for strategic outsourcing to talk about emerging biotherapies and their manufacturing challenges:
- Mark Angelino (senior vice president of pharmaceutical sciences at bluebird bio)
- Chris Chen (chief executive officer of WuXi Biologics)
- Andreas Weiler (global business unit head of emerging technologies at Lonza).
Seymour first pointed out the differences between allogeneic and autologous cell therapies: The former is patient-specific; the latter involves ‚Äúsomewhat immortalized cell lines‚ÄĚ from donors. They have different types of manufacturing concerns. Similarly, gene therapy platforms span a range of viral vectors and other delivery modes, some products overlapping with cell therapies. ‚ÄúGene editing is an exciting technology we are just starting to deal with from a manufacturing perspective,‚ÄĚ Seymour said. Finally, she introduced antibody‚Äďdrug conjugates (ADCs), which involve challenges related to safety and heterogeneity, as well as supply chain management.
Cell therapy manufacturing typically begins with harvesting cells from a patient or donor. That can create a need for ‚Äúminiature manufacturing environments‚ÄĚ to be built up around processing cells before they can be given to patients. Seymour said companies need to be aware of supply chain, manufacturing control, and quality issues when it comes to such processes. ‚ÄúTo date,‚ÄĚ she said, ‚Äúthese technologies have to rely on open manual processing at relatively small scales, and they use research-grade materials. But we need to move toward more robust, reproducible processes at larger scales.‚ÄĚ Compliance with good manufacturing practices (GMPs) will be important for processing clinical and commercial cell therapies.
Next, she focused on two types of gene therapy viral vectors: adenoassociated viruses (AAVs) and lentiviral vectors (LVVs) for delivering genes. She referred attendees to an¬†article in BPI‚Äôs April 2016 cell therapy supplement as ‚Äúa really wonderful overview of the gene therapy space‚ÄĚ (1). Seymour described this development pipeline as ‚Äúa very active field looking at multiple types of indications.‚ÄĚ Those include cancers and inherited disorders as well as others ‚ÄĒ with several products in development for each. Manufacturing challenges come with producing, amplifying, expanding, purifying, and polishing viral vectors as drug substances. But gene therapy drug-product manufacturing has its own concerns, she said, ‚Äúbecause many fill‚Äďfinish houses aren‚Äôt willing to accept this type of technology based on cleanability issues.‚ÄĚ Thus, sponsors of such products may have trouble finding contract manufacturing organizations (CMOs) that will take on their projects.
In reference to ADCs, Seymour pointed out that their history is longer than some might think. ‚ÄúI started my career [working on them] at Immunogen back in the 1980s. I remember when the maytansinoid¬†technology first came out, and now it‚Äôs a commercial product. It was exciting then, it‚Äôs still exciting. But when is the next product going to be approved, what platform will it use?‚ÄĚ
ADC products are complicated, with complex supply chains that make them expensive to manufacture because of all of their different parts and related chemistry, manufacturing, and controls (CMC) issues. ‚ÄúOne ADC essentially is the equivalent of four or five CMC programs,‚ÄĚ Seymour said. ‚ÄúSo if you know how much it costs to develop an antibody and a CMC platform, then you add the toxin and other components, and you can see how the cost of goods (CoG) can increase substantially.‚ÄĚ That often becomes an issue in portfolio management, she explained, when companies want to balance the portfolio with a product mix of affordable CoG.
Mark Angelino: Bluebird bio was funded and rebranded in 2010. specializing in gene therapies. It uses an ex vivo LVV to transduce a patient‚Äôs own cells. After processing, the modified cells are returned to patients as a curative therapy. The company currently has four products in clinical trials for childhood adrenoleukodystrophy, beta thalassemia, sickle-cell disease, and oncology. The company also added a gene-editing platform two years ago. ‚ÄúAs if the complexity of the manufacturing platform wasn‚Äôt enough,‚ÄĚ Angelino said, ‚Äúwe bolted on a gene-editing platform a couple years ago from an acquisition. So we are also having gene editing play at the preclinical stage.‚ÄĚ
Andreas Weiler: Founded 120 years ago in the Swiss Alps, Lonza moved into biotechnology 30 years ago and ADCs in particular 10 years ago, when it also built a bioconjugate facility. Its ADC program leverages the company‚Äôs biologic experience and small-molecule experience in combination. To move into gene therapies, the company acquired Vivante GMP Solutions in Texas six years ago. So it now has capabilities across the breadth of emerging technologies. Lonza has manufactured¬†multiple batches and commercialized a few products.
Chris Chen: WuXi Biologics has a wealth and depth of expertise in multiple platforms for manufacturing ADCs, antibodies, and gene therapies.
Forecasting: To begin by addressing patient demands and needs for new therapies, Seymour asked Angelino to describe his company‚Äôs perspective. ‚ÄúIt varies depending on the indication,‚ÄĚ was his answer. ‚ÄúWhen we started six years ago, early indications were for ultrarare diseases.‚ÄĚ To that point, he mentioned stem-cell programs and bluebird bio‚Äôs lead central nervous system (CNS) treatment for cerebral adrenoleukodystrophy (C-ALD), which has a worldwide incidence of 1 in 20,000 newborn males each year. The company‚Äôs lentiviral C-ALD product is currently in an upcoming phase 2‚Äď3 trial.
Another bluebird bio target indication is beta-thalassemia (for which a different lentiviral product is in phase 2‚Äď3 trials as well), with what Angelino described as a larger patient population. The same vector is also in phase 1‚Äď2 trials for sickle cell disease, which Angelino described as affecting an even larger patient population. ‚ÄúThen in oncology,‚ÄĚ he said, ‚Äúyou may have thousands if not tens of thousands of patients. So our approach has been a stepwise gradation.‚ÄĚ Bluebird seeks to prove its technology in smaller indications before expanding to larger markets.
Seymour asked the others what they see among their CMO customers and how their companies are addressing those needs. Weiler spoke about oncology. ‚ÄúWhen you treat a patient with chemotherapy, you fight not only the disease, but you also fight his whole body.‚ÄĚ He contrasted that with ADCs, which offer patients a better quality of life with fewer side effects. ‚ÄúAnother nice approach enables a patient‚Äôs own immune system to recognize the cancer and destroy the cancer, so you have even fewer side effects because the patient‚Äôs own body fights the cancer. In my opinion, it‚Äôs a total paradigm shift.‚ÄĚ¬†He too mentioned smaller patient populations for rare and ultrarare diseases. ‚ÄúThe beauty of some of these treatments is that they might be curative.‚ÄĚ
Chen said that patient populations might be relatively small for cell and gene therapies, but for ADCs the market could be as large as or even larger than that for monoclonal antibodies (MAbs). ‚ÄúWe‚Äôve actually seen,‚ÄĚ he said, referring to clinical trials, ‚Äúthat sometimes to get the maximum effect, people even use a higher dose than for traditional antibodies.‚ÄĚ Successful ADCs could be manufactured at 500 kg to a metric ton each year, ‚Äúand that would be a huge burden on the supply chain. Right now, the global capacity for ADCs is not there. So it‚Äôs going to be interesting over the next five to 10 years to see whether ADCs become blockbuster products. With multiples of those, there would be a huge shortage of ADC commercial manufacturing capacity, both for drug substances and drug products.‚ÄĚ He believes the forecast is similar for bispecific and bimodal antibodies.
‚ÄúThese technologies take us beyond the sort of pellet-gun approach of chemotherapy, where you just kill as many rapidly dividing cells as possible,‚ÄĚ Seymour affirmed, ‚ÄúNow it is very focused, and patients get the best benefit with the fewest side effects. But it‚Äôs still a work in progress. For ADCs, the drug/antibody ratio and free-toxin issues can become problematic. But how focused can gene therapies be? It helps to clarify the limitations of existing medicines, and that can drive development of these new therapies.
‚ÄúMark,‚ÄĚ she said, ‚Äúyou mentioned that a lot of bluebird bio products are truly intended to be curative. But what does that mean from a manufacturing and product-reimbursement perspective? If you cure your patient population, then you don‚Äôt have a market anymore. Might you need to reapply some such therapies if cells stop expressing what you want them to express?‚ÄĚ
Angelino acknowledged that such effects have been seen with some early¬†product candidates in the field. But he countered, ‚ÄúThe hope is getting to where we may not need reapplication. When it comes to forecasting capacity, it depends. In some case you get a peak and then a plateau to what is essentially the treatable population. In other cases, we see a more traditional pharmaceutical industry approach to forecasting, so the peak is much higher, and then it trails down. A lot of that has to deal with how you set up capacity and multiproduct facilities. How do the peaks and troughs overlay, and will they allow you to use your capacity maximally? That‚Äôs a lot of what we deal with daily, weekly, and monthly, as we determine what we need from a portfolio to make it sustainable.‚ÄĚ
Seymour wondered whether CMOs must examine their clients‚Äô approaches to forecasting and build infrastructure to support them given the uncertainty of demand, potential peaks and fluctuations, and stabilization over time. Weiler pointed to the pipeline of emerging therapies, with about 1,100 products currently in preclinical and clinical testing. Then he asked who knew how many such products had been commercialized so far. ‚ÄúIt‚Äôs actually less than a dozen. Most of these drugs are innovated by smaller players, so I think this has huge implications for the market. A CMO has to consider what is in the preclinical pipeline right now to prepare itself and build the right capacity to serve those clients. If you follow this courageous model ‚ÄĒ invest now into cell and gene therapies and bioconjugates ‚ÄĒ then you can lose or win big.‚ÄĚ
In 2006, when Lonza invested US$50 million into ADCs, Weiler pointed out, few other CMOs were building such capacity. ‚ÄúAt the end of the day, it played out,‚ÄĚ he said, ‚ÄúBut it could go the other way. So a CMO has to take a certain risk and look deeply into process innovation. That‚Äôs your strength.‚ÄĚ He also pointed to a potential paradigm shift related to intellectual property (IP). ‚ÄúWhen you look into these new products, the IP is not only in the product, but also the process. This is very different from¬†most drugs on the market. So CMOs can gain a big advantage through process innovation, which is critical to the success of innovators in the market.‚ÄĚ
Chen agreed. ‚ÄúAs a CMO, we manage capacity. But there‚Äôs much less risk than a product company faces because a given product could fail. But we manage an entire portfolio. As Andreas said, there are a lot of programs out there.‚ÄĚ He described WuXi‚Äôs perspective as essentially, If you build it they will come. ‚ÄúAs long as we are bullish on the industry, then the payback will be here. Eventually, you have a leap of faith to go through.‚ÄĚ He called the ADC supply chain ‚Äúa nightmare,‚ÄĚ with four or five CMC programs per product in development. So his company ‚Äúput everything within 100 miles,‚ÄĚ from synthesis of high-potency toxins to bioprocessing of MAbs to conjugation to fill and finish. ‚ÄúSo WuXi can manage the entire ADC supply chain in a very short distance.‚ÄĚ
Process Control and Quality by Design (QbD): Seymour returned to the concept of a process defining a product. ‚ÄúAre we actually going backward? We‚Äôve made so many strides in well-characterized biologics for proteins. But for emerging therapies, are we actually going back to the old model?‚ÄĚ
For autologous cell therapy products (patient-specific treatments), Weiler described two business models: either centralized or decentralized. ‚ÄúYou can have a centralized business model in which you take the blood of a patient, genetically modify it, expand the cells, then transport it back to the patient for reinjection.‚ÄĚ The advantage of that model, he said, is greater control for regulatory compliance. But in a decentralized model with a fully automated reactor that could perform the full bioprocess, ‚Äúyou would have hospital centers that treat patients right there: take the blood, change it, and reinject through fully automated processing. We are not there right now, but we are moving fast in that direction to offer both centralized and decentralized models.‚ÄĚ
Seymour agreed. ‚ÄúThat is an area manufacturing professionals need to focus on. We talk about big facility expansions, hundreds of thousands of square feet, and tens of thousands of liters for proteins. But now we‚Äôre going to need a hood, a few hands, and benches to create these little local manufacturing environments. It‚Äôs truly a personalized medicine approach, and all the regulatory controls around that are still being sorted out around the world.‚ÄĚ The concept brings up questions that regulators and companies are still struggling to answer ‚ÄĒ over batch definition and release, for example. Seymour asked Angelino to address those challenges from his innovator company‚Äôs perspective.
The biotechnology field has spent a long time trying to get away from the process‚Äďproduct problem, he said. In these early days of emerging therapy development, however, ‚Äúthere‚Äôs a conservatism that says, ‚ÄėWell, if we don‚Äôt change the process, then we won‚Äôt affect anything.‚Äô But I think that‚Äôs dangerous.‚ÄĚ Regulatory agencies want to see processes evolve, especially as technology advances. So companies need to work with regulators to define and show what actually changes. ‚ÄúWe‚Äôre getting more latitude as the science catches up. Challenges will depend on indications. We have some ultrarare indications; we‚Äôve got some broader indications.‚ÄĚ The former may be satisfied with centralization; the latter could lend itself better to a decentralized approach.
‚ÄúThat makes me think of process control,‚ÄĚ Seymour said. Perhaps automated technologies and manufacturing platforms (the decentralized option) lend themselves to starting QbD from the very beginning rather than more organically developing it to the proof-of-concept stage, then tightening the design space in a later phases. ‚ÄúIf you can define all of that early,‚ÄĚ she suggested, ‚Äúand then go out into all of these miniature manufacturing environments, does that provide more flexibility?‚ÄĚ
‚ÄúWe‚Äôve all grown up with QbD now,‚ÄĚ Angelino said. ‚ÄúI come from a¬†more traditional setting, where there‚Äôs an expectation to use it.‚ÄĚ He believes there is much value in trying to learn as much as possible about an emerging therapy early on. But he cautioned, ‚ÄúThe science hasn‚Äôt quite caught up.‚ÄĚ At this point, he says, people in the field are learning as much as they can; it may be too early yet to define attributes as precisely as we‚Äôd like. ‚ÄúIt‚Äôs a bit of both worlds right now.‚ÄĚ
‚ÄúThat‚Äôs a good point,‚ÄĚ said Weiler. He highlighted the ‚Äútwo-dimensional field‚ÄĚ of cell-therapy bioprocessing ‚ÄĒ with trays and flasks and other means of culturing and expanding attachment-dependent cells. ‚ÄúYou need huge facilities and a lot of manual processing. To move into a 3D process (bioreactors) takes a lot of QbD. But at the end of the day, you need to show comparability: that the cells you have grown on a 2D platform are the same as those you are growing in a 3D bioreactor.‚ÄĚ His company is currently working to transfer many 2D processes into 3D platforms using microcarriers, and QbD is helping it do so. ‚ÄúThis is a challenge right now, but we‚Äôre making big progress.‚ÄĚ
‚ÄúFor smaller indications ‚ÄĒ ultrarare diseases ‚ÄĒ you probably don‚Äôt have the resources to apply QbD from the beginning,‚ÄĚ Chen remarked. ‚ÄúFor much larger indications, QbD probably makes more sense.‚ÄĚ
Innovation: Seymour sees a strong potential here for an important service offering. ‚ÄúCan CMOs offer the QbD process, which can be labor intensive and costly? What are some other offerings that CMOs are building into their business models to meet some of these demands?‚ÄĚ For example, she wondered whether Lonza‚Äôs facility in Walkersville, MD, could eventually use thousand-liter bioreactors for cell therapy expansion. ‚ÄúHow do you look at your manufacturing networks to do those types of things?‚ÄĚ
Such scales will be unnecessary, Weiler countered. ‚ÄúIn a 200-L bioreactor right now, we can manufacture as many stem cells as in a football-field‚Äďsized facility using 2D processing. That‚Äôs the future for many allogeneic processes.‚ÄĚ But autologous processes will need innovative¬†solutions for manufacturing. He asked how batch consistency can be ensured when each patient constitutes a single batch. Here is where the challenge of managing a supply chain comes to the fore. ‚ÄúIt is a strength of a good CMO to look early enough into process innovation. I don‚Äôt think we should compete with the innovators, but we should enable them to bring drugs to market as fast as possible.‚ÄĚ
Chen elaborated: ‚ÄúCMOs are beginning to offer solutions. It‚Äôs not just helping sponsors with one thing. For some smaller companies, we provide guidance from the beginning of a program.‚ÄĚ That means design and implementation of a manufacturing process, including scale-up or -out. ‚ÄúEssentially we guide them through an entire product lifecycle approach. For ADCs, we even offer a discovery service, so we can help a client assemble a different type of warhead, a different type of conjugative chemistry, or different types of antibodies ‚ÄĒ essentially make 40‚Äď50 different ADCs the client can use in pharmacology studies and find the best one to develop. So we go from making one milligram to one gram to 100 grams and eventually one kilogram.‚ÄĚ
Weiler expanded on the difficulty of ADC development. Antibodies are relatively huge molecules; the linkers and toxins are tiny. The chemistry of combining them is not too difficult. But determining where the small molecules connect to the large ones, ‚Äúthat‚Äôs the real crux. You have a heterogeneous component, so we‚Äôre looking into site-specific conjugations.‚ÄĚ That would produce¬†more homogeneous compounds, with only one or two toxins sticking on to the same position on each MAb structure. Weiler described such an approach as the next generation of ADCs. Current marketed products are first-generation heterogeneous mixtures.
‚ÄúHowever, until now nobody could show that site-specific conjugates have more efficacy in clinical trials. That‚Äôs what everyone is nervous about now. Let‚Äôs see how this new generation of ADCs will do. Some people say ADCs are dead ‚ÄĒ but clearly the pipeline is growing dramatically with these new technologies. And we also see new peptide conjugates and combination therapies, which are exciting.‚ÄĚ
Innovators often have an antibody and an indication in mind, but need to partner or license the technology of a toxin and payload. CMOs can provide process assistance and scale-up, but some sponsors may solve their own process problems.
‚ÄúMark [Angelino] and I worked together at Millennium Pharmaceuticals,‚ÄĚ Seymour said. ‚ÄúGE Healthcare‚Äôs Xcellerex platform originated out of that organization, and now it‚Äôs a well-known and widely implemented single-use bioreactor platform. There are other examples: GE‚Äôs Wave bioreactor came out of Schering-Plough, Thermo Scientific‚Äôs HyClone SUB system came out of Baxter Healthcare. These companies had unmet needs, were able to address them in-house, then realized that they didn‚Äôt want to be technology companies in that sense (equipment manufacturers) and spun them out. But where is that going with emerging therapies? Are CMOs willing to make such investments, or will those have to come from the product-sponsor side?‚ÄĚ
Angelino believes it will be both. ‚ÄúAs much as we aspire to be a drug-product company, we have invested a lot in technology. We know what we need ‚ÄĒ to improve product yields, improve analytics, and so on ‚ÄĒ and we‚Äôre trying to partner with academics and companies with technologies. We‚Äôre looking at this as they did biologics and MAbs in the 1990s.¬†Remember when it was a big deal to get a gram per liter expression? As for analytical technologies, our ability to characterize biologics today is so profoundly different than it was. So our view now is that the technology will catch up, and you‚Äôre either part of it or you‚Äôre going to be left behind. As soon as it becomes available, regulatory expectations will go there. So we‚Äôre doing a bit of both right now.‚ÄĚ
‚ÄúInnovation comes from innovators and their ideas,‚ÄĚ Weiler pointed out. ‚ÄúWe have to adopt their processes for their needs.‚ÄĚ Again he brought up the product‚Äďprocess link: ‚ÄúThe media, the growth factors, the bioreactor or 2D reactor ‚ÄĒ a lot of things have to come together. So every process is individualized, and we optimize them together with our clients. The synergistic effects are important: A CMO can leverage the experience it has gained over the years by working with other clients. A CMO works with many different technologies from many different innovators and partners. So it can use the best techniques to make its clients successful, which I think is even more important for these technologies in which the process plays such an important role. Right now they are where biologics were 15‚Äď20 years ago. We will see huge improvements in productivity over time.‚ÄĚ
Fast-Track Concerns: Many emerging therapies are achieving fasttrack regulatory status because they are intended to address highly unmet medical needs, they are innovative approaches, and they are proving be efficacious in small trials. Seymour asked Angelino, ‚ÄúWhat keeps you up at night when you get fast-track status but your CMC is two years behind?‚ÄĚ
He pointed to the European Medicines Agency‚Äôs (EMA‚Äôs) adaptive pathway. ‚ÄúThe acceleration is often in the context of clinical testing.‚ÄĚ On the CMC side, meanwhile, regulators are still holding the line. ‚ÄúSo with a gene therapy, we‚Äôve been very upfront about that: ‚ÄėIf you want to use adaptive trials and get accelerated, then you‚Äôre going to have to get creative on some of the quality and CMC aspects as well.‚ÄĚ That doesn‚Äôt mean cutting corners, he¬†insisted. Instead, the company (with regulator help) must distinguish between what information is necessary and what is simply nice to have. Because of the potential benefit for patients, regulators are willing to work with companies on the fast track.
‚ÄúIt still does give us nightmares from a CMC perspective when everything is accelerated,‚ÄĚ Angelino said, ‚ÄúThis gets to the earlier question of product and process. The agencies have to manage the fact that we‚Äôre going to change these products throughout their lifecycle. Even though they‚Äôre complex ‚ÄĒ maybe even more complex than what we‚Äôve seen in the past ‚ÄĒ regulators see the advantages. We‚Äôre learning in the clinic that after a small study of 15 patients, we can adapt and possibly make an even better product. The agencies see that, so this collaboration can and does happen.‚ÄĚ
‚ÄúIt‚Äôs a dilemma,‚ÄĚ Weiler agreed. ‚ÄúWe have a lot of processes in late clinical phases based on older manufacturing technologies. So probably the product launch will happen with those technologies while we‚Äôre talking to the agencies about switching over to a modern technology ‚ÄĒ for example, moving from 2D to 3D processes. Many validated processes use 2D culture, so there will be scale-up issues in the future if patient populations increase. The authorities have to be open to changing technology. We are in a lot of discussions and see a lot of openness, especially in the Japanese market.‚ÄĚ
‚ÄúIn the end,‚ÄĚ Chen said, ‚Äúit‚Äôs still the sponsor, the innovator, who needs¬†to decide whether to take a risk now or take it later on. This is a balance of that.‚ÄĚ
Seymour suggested that investing in QbD early on could help accelerate a CMC process. ‚ÄúIf you are given this opportunity to file sooner, you risk not having a complete package that you otherwise would have collected over the full development cycle time. My understanding is that the agencies are very receptive to this because they want to get these products to market but safely.‚ÄĚ
Again she brought up forecasting. Most emerging technologies are still early in their clinical phases. ‚ÄúThere‚Äôs usually a lot of excitement with some early phase data, then some not-so-exciting news as they go into larger trials and maybe have some clinical failures.‚ÄĚ This will affect CMO business. Companies planning for commercial launch are ramping up resources, but then a product might fail. ‚ÄúThe range of probability is enormous,‚ÄĚ Seymour said. ‚ÄúThinking about that early on, how do you build your network? Where do you build your capacity?‚ÄĚ
Speaking from the innovator perspective, Angelino said his company made the choice ‚Äúto be virtual at this point, and that was because our lead indications were our smallest indications. We do see the value of building capacity at some point, but that capacity can be driven by something of volume or at least a coalition of a few assets that provide the necessary volume.‚ÄĚ He pointed out the difference between growth ‚Äúto the right, which is toward commercial launch‚ÄĚ and growth ‚Äúto the left, from a platform perspective.‚ÄĚ A sustainable business model requires more than just one product. It requires enough volume to sustain a certain capacity. ‚ÄúFor us, it‚Äôs been trying to grow both to achieve that.‚ÄĚ
‚ÄúThat‚Äôs one beauty of cell therapies,‚ÄĚ Weiler said, ‚Äúcompared with MAb manufacturing. You don‚Äôt need big facilities; you have more or less a cleanroom that you can build up and use for many different manufacturing processes (thanks to single-use equipment). So it is¬†important for a CMO to have those available.‚ÄĚ The associated capital investment risk is much less than it would be for a traditional mammalian-cell‚Äďculture facility for MAb manufacturing, which could cost $300‚Äď400 million. ‚ÄúYou have to take a certain risk as a CMO, and I think WuXi and Lonza are doing this. But the risk of clinical failure with these products is not as high as it would be for mammalian-cell biomanufacturing.‚ÄĚ
Chen agreed. ‚ÄúA CMO can manage multiple programs, so the risk is lower because it averages out. Right now, based on the current industry outlook, many programs are coming into the clinic. Even if one struggles, other programs are ready to come in quickly. So there is no problem over the short term, but managing capacity is always a concern over the long term, both for CMOs and innovator companies.‚ÄĚ
Barriers to Success: Seymour pointed out that manufacturing capacity can be strained if more products succeed than expected. She asked what obstacles stand in the way of emerging therapies advancing through clinical trials. ‚ÄúWhat about access to talent and capital? These are new technologies. How do you make sure you have the best trained people to work on these products?‚ÄĚ Sponsors and CMOs are potentially competing with each other for talent. ‚ÄúIf you look back five or six years with bluebird bio,‚ÄĚ Angelino said, ‚Äúwe chose to search within academia. These were technologies coming from academia, and at that point for us it was not about getting to GMP as much as securing the technology itself.‚ÄĚ He said that as a sponsor company grows, it can add pieces ‚Äútoward the end game, which is commercial manufacturing.‚ÄĚ In recent years, AAV and LVV companies have taken off, and around Boston (for example), they‚Äôre all fighting for the same people. Companies need to focus on retaining people, he said, because it‚Äôs difficult to sustain growth as technology advances rapidly.
Weiler agreed that there is ‚Äúa war for talent in this business right now.‚ÄĚ¬†It‚Äôs a job-hunter‚Äôs market, where people with the necessary knowledge and skills have many opportunities. ‚ÄúBut it‚Äôs also important to train people properly.‚ÄĚ He offered the example of chimeric antigen receptor (CAR) T-cell manufacturing: ‚ÄúYou take blood from a patient and genetically modify that blood. At the same time, the patient is treated with hemotherapy while your people are processing and expanding their modified cells. Then you have to reinject them into the patient.‚ÄĚ All of that places great stress on workers, who have one chance to get everything right. ‚ÄúYou need to train these people extremely well because it‚Äôs one shot, and you have to deliver the right quality in the right time. That‚Äôs extremely important, so I believe it‚Äôs all about training, training, training.‚ÄĚ
As for other bottlenecks, Weiler pointed to scale-up. ‚ÄúBut there is no problem I can see that cannot be solved in the next couple of years. Many of these treatments address unmet medical needs, so there will be solutions.‚ÄĚ
Chen sees manufacturing capacity as a potential bottleneck. ‚ÄúI think there is still very limited capacity on the cell therapy and gene therapy side. So either the established players must invest more, or newer players must come into this industry. It‚Äôs very important to support future growth.‚ÄĚ
Looking to the Future: Seymour opened up the session to the audience. The first question was about the overall industry forecast: ‚ÄúIf it‚Äôs taken many years to come up with a homogeneous ADC product ‚ÄĒ and regarding cell therapy products, cells are going to be more heterogeneous as we modify them (even selecting for the right population to begin with) ‚ÄĒ do you think it will take half a century to get there?‚ÄĚ
Each new therapeutic mode can build upon the knowledge of those that came before. And Angelino said he thinks it will be different this time around. ‚ÄúThese are profound step-ups of risk and reward, so they won‚Äôt be perfect products the first time out, but they will be products. These drugs are¬†making a difference.‚ÄĚ Over the past three years, he said, more companies have been entering this space. ‚ÄúThe second wave is already seeing big changes in the risk profile, with a lot more understood as time goes by. A couple of big difficulties are starting to get addressed, and now you‚Äôre seeing profound effects for patients.‚ÄĚ
Weiler pointed to the possibility of curative approaches. ‚ÄúThat excites me every day when I come to work. There is a new kind of product coming to market using your own immune system to fight disease. That has never happened before. Patients treated with ADCs can have a normal life, which is remarkable. So I hope it will not take as long, but especially with ADCs we need a few next shots. We will see those, those will come, and we hope that will be soon.‚ÄĚ
Product and Process: The next question came from a woman working at a start-up company in Taiwan. ‚ÄúWe are working on a new ADC platform that may solve some homogeneity problems,‚ÄĚ she said. ‚ÄúHow soon should we bring in processing experts to help us get it into clinical trials quickly?‚ÄĚ
Chen told her, ‚ÄúYou need to look at the processing as early as possible. ADC processing is key.‚ÄĚ And Seymour confirmed that: ‚ÄúAssembling the right team early is always important. Understanding your process as well as you can early on helps minimize risk down the road. So when you do make process changes, you understand what you are changing and do it intentionally. You have control over that, so you don‚Äôt have to redo or invalidate some clinical results (even nonclinical results), which can be very timely and costly.‚ÄĚ
Failure Rates: The final question came from a PhD student from the Rheinisch-Westf√§lische Technische Hochschule University in Aachen, Germany. He wondered how the failure rate for biologics in clinical trials compares with that of synthetic, small-molecule chemical drugs.
‚ÄúWith small molecules,‚ÄĚ said Angelino, ‚Äúsome 10% will make it through to commercialization, primarily because you have preclinical and safety failures.‚ÄĚ The numbers are¬†better for biologics, which tend to fail later for efficacy reasons ‚ÄĒ making those programs more costly in the long run. ‚ÄúFor rare diseases, the numbers are higher than in chronic indications ‚ÄĒ just because of some latitude around risk and reward.‚ÄĚ For emerging therapies, the data aren‚Äôt yet clear. ‚ÄúFor a monogenetic disease, if you can fix that one specific mutation, then you should have an effect. But will that translate to some larger failure rate down the line? That‚Äôs yet to be seen.‚ÄĚ
The industry would love to have some reliable data around this question, Weiler said. ‚ÄúIn 1990, about one out of four products made it from phase 1 on to market approval; right now, it is one out of eight (a total average for biologics and small molecules). But when I look into cell therapy, for example, there is no general failure rate from phase 1 into phase 2, when you move directly into efficacy. These things are your own stem cells, often, so they are safe.‚ÄĚ
‚ÄúProteins don‚Äôt fail as early as small molecules,‚ÄĚ Seymour added, ‚Äúbecause¬†they are safer, in general, as are certainly some cell therapies. That means you spend more money longer into the development cycle. From a small-molecule perspective, you want to kill programs quickly and early because you don‚Äôt want to invest a lot of money in a product that fails. But you end up taking biologics very far into development, into large pivotal trials, spending a lot of money, and then you can get the disappointing news. So it‚Äôs tough. How much are you willing to risk to prove the reward?‚ÄĚ
Weiler concluded, ‚ÄúUntil 2010, there was a huge decline in new drug and biologics license application (BLA) approvals. Since 2010, the focus has shifted a bit more toward more targeted, personalized medicines, and we see a huge uptake in new drug approvals and BLA approvals. Last year was a record year, and this year is expected to be another record year. But that also has to do with the changing focus of pharmaceutical companies: It‚Äôs not¬†only the focus on blockbusters anymore; it‚Äôs more the focus on targeted, personalized medicines.‚ÄĚ
1 Pettitt D, et al. Emerging Platform Bioprocesses for Viral Vectors and Gene Therapies. BioProcess Int. 14(4) 2016: S8‚ÄďS17.
Cheryl Scott is cofounder and senior technical editor of BioProcess International, PO Box 70, Dexter, OR 97431; 646-957-8879; firstname.lastname@example.org; www.bioprocessintl.com. Patricia Seymour is a senior consultant with Bio Process Technology Consultants (BPTC), 12 Gill Street, Suite 5450, Woburn, MA 01801-1728; 1-781-281-2707; email@example.com; www.bptc.com.