Where Are Today’s Biomanufacturing Bottlenecks and How Is Industry Looking to Overcome Them?

View PDF

Moderator Dan Stanton, with Ramon LeDoux (YMC Process Technologies), Courtney Morgret (AbbVie), and Kiran Chin (MKA Insights)

Moderator Dan Stanton began the discussion by talking about bottlenecks within manufacturing. “The term bottleneck brings up imagery that describes the sort of problems we have when it comes to the manufacturing flow,” he stated. He posed the first question to Kiran Chin: “Your presentation highlighted companies that ask the right questions, and you spoke of the importance of doing so when it comes to technologies and strategies. In your experience, do bottlenecks feature prominently in those important questions?”


Chin didn’t think so. “They should, but they don’t. Everybody gets so focused on what’s ahead of them that they fail to realize what’s behind them,” she responded. Questions that need to be asked include, “What are our bottlenecks?” and “Where are we struggling to meet the customer demands or the customer challenges?”

Stanton asked Ramon LeDoux and Courtney Morgret about their experiences in upper-level management realizing the importance of bottlenecks ahead of time. Morgret thinks it all comes down to the bottom line. “The intricacies of what it takes to get a product to a patient are known,” she said, “but there are so many different areas that need to come together to make that happen. It’s known and understood, but vast.”

LeDoux believes that from a process development and manufacturing perspective, it’s clearly understood that bottlenecks are a source of frustration. “It’s a real challenge,” he said, “because nobody wants to implement a new process that can’t be scaled up.” He added that there needs to be enough material to work with, but that “with such a focus on speed to market, the implementation of newer technologies often takes a back seat to the speed-to-market issue.”

LeDoux added that companies address such issues by scaling based on current technologies in their platform. In an existing facility, for example, the bottleneck might be in the amount of buffer the company can prepare or use at a given time and the size of chromatography columns needed to support equipment that it already has. Such a company would prevent potential bottlenecks by refraining from changing the technology and simply scaling larger.

Chin added, “I think that the phrase unmet need is what we use to describe bottlenecks. How do we address this challenge, and how do we continue to grow? When we think about the unmet need, I think that is on the radar of all the executives that I speak to. If we don’t meet that customer need, we’re going to fail to thrive.”

Overcoming Bottlenecks in Biomanufacturing
Stanton shifted the question to biomanufacturing bottlenecks. “Where are they occurring?”
Chin pointed to a lack of available talent. “The biggest challenge right now seems to be in getting trained talent through the doors to do the job we need done. What we’re seeing is people jumping from organization to organization just because the retention programs are not so great.”

Morgret added that the bottlenecks in antibody–drug conjugate (ADC) development lie in the significant lead times required to get such products to patients. Supplier schedules require a certain agility to respond to changing clinical needs and shifting timelines. She believes that having the flexibility to manage all the different manufacturing modes can be a challenging proposition.

Stanton highlighted the complexity of manufacturing ADCs. “When you’ve got an antibody and a small molecule, and you’ve got the bioconjugation left to do, does that theoretically give you three times the opportunity to have bottlenecks and three times the problems? Is that how it equates?”

Morgret agreed. “In some ways the individual entities can create just as much of a bottlenck as a standalone MAb can.”

Stanton asked how a contract development and manufacturing organization (CDMO) manages those three different loads. Morgret thinks that success lies is having a good understanding of the clinical demand and program timelines as well as dosing strategies and clinical study design. That can allow contractors to “switch gears and supply other programs as needed,” she said.

LeDoux believes there will always be a bottleneck in biopharmaceutical manufacturing. He pointed out that downstream processing currently has the most bottlenecks. “One of those very clear areas is in protein A capture of the antibodies. It’s a very templated process that lends itself to economic stress and economic drivers.” Additionally, it is a very expensive unit operation because of the cost of the resin, so there is economic pressure to drive that cost down and reduce the scale of capture. Once that process is more optimized, the bottleneck simply moves.

Chin added that roadblocks don’t exist for very long because people find workarounds. “I feel that the bottlenecks keep shifting and are temporal. A bottleneck exists for a short period until someone finds a way to innovate and work past that hurdle. That’s what I love about this industry.”

Heading Toward Continuous Manufacturing
Stanton asked Morgret and LeDoux how they felt about innovators in the bioprocess industry. “How well are they addressing such issues?”

LeDoux said that the industry is home to “a huge amount of intellectual horsepower. Conversely though, there’s also a reluctance to change because of the regulatory environment. The FDA has been touting continuous manufacturing for quite a bit of time. Yet we’ve seen upstream again lead downstream, but without discussion around a truly integrated continuous approach. I always go back to antibody manufacturing because it’s such an easy process. Technically, it’s probably one of the easiest processes we work with. It’s the most templatable, so in many respects it lends itself to leading the curve toward establishing a continuous manufacturing model.”

Yet it’s very difficult to get to that level because of the need for speed to market and risk mitigation. Nobody wants to be the first to adopt leading-edge technology. It’s always difficult to be an early adopter. One interesting thing, LeDoux noted, is that continuous processing also is a niche-oriented opportunity. If you have a blockbuster that you expect to be worth several billion dollars a year, then you will be reluctant to implement a new process that could slow down regulatory approval and speed to market. Conversely, if you have a small-market molecule in development — a very specific antibody that is almost an orphan drug status, for example — then the value of continuous manufacturing perhaps wouldn’t be apparent.

LeDoux added, “You have a window of opportunity to establish a new process and huge barriers to overcome because nobody’s done it yet. There’s a lot of discussion, there’s a lot of work being done in process development and continuous manufacturing, but it’s very difficult to get to the next step with it.”

Chin agreed. “In my experience, it’s about having the proof of concept completed as a standalone activity so that when that program presents itself, we can apply it. It’s very challenging to impose a risk to either supply or timeline. If we can do a proof of concept ahead of time for a new technology, then that gives us a fighting chance to implement it.”
Stanton found it interesting how continuous processing was brought up as a potential way to debottleneck the upstream to downstream transition. He asked whether companies mostly are developing scale-up processes or instead are focusing on optimizing upstream steps.

LeDoux wasn’t certain but suggested that the answer simply might rest on the need to manage risk. “It’s not such a huge paradigm shift to go from fed-batch or batch to profusion. The upstream side always seems to be ahead of the downstream side.”
Chin believes that greater pace of innovation happens upstream because capital equipment costs to purchase large bioreactors are far greater than just modifying a process. She pointed to how yields are improved upstream.

“I think the reason downstream isn’t innovated as frequently or at the pace of upstream is that, if it isn’t broken, don’t fix it. And there really is only one solution: It’s protein A for the most part. There’s nothing else that’s been challenging that until today,” she said.

Adopting Single-Use Technologies
Stanton then asked whether adoption of single-use technologies in upstream production might help debottleneck processes downstream.

Chin is a proponent of a single-use solutions that enable companies to work with smaller volumes of materials. Morgret added that single-use technologies help with cleaning and remove the need to develop analytical methods for containment of highly potent components. “Single use definitely can help debottleneck some of those processes.”

LeDoux argued that it doesn’t address the entire problem, however. As expression titers rise, he said, single use is only part of the solution if a developer doesn’t address downstream process efficiencies.

Stanton asked whether using disposable bioreactors upstream can cause different problems farther down the line than if stainless steel had been used. LeDoux didn’t think so, saying that if utilities are available to support stainless steel use in downstream manufacturing, then such utilities can be coupled to single-use bioreactors as well. “I don’t know that there would be any reason why one would necessarily influence the other directly,” he said.

Morgret added that when single-use technologies are used in conjugation unit operations, solvents often are present. In such cases, single use needs to be implemented early on so that a company can demonstrate acceptable levels of leachables and extractables and process stability.

According to Chin, automation shows great potential. “I think single use is a solution, but it is not the perfect solution. We need to start pulling software into what we do as an industry because the software probably will fix a lot of what’s broken, and it’ll create the level of transparency that is missing in our processes today,” she said.
Looking at Automation

Stanton noted that he has heard automation discussed more and more these days, specifically focused in the cell and gene therapy space. “They’ve got separate issues from the manufacture of MAbs, but they seem to have gone straight through software and automation to try to address them. Do you find that is the case? And why is that part of industry looking at software and automation when MAb manufacture isn’t so much?”

“Because cell and gene therapies incorporate the patient as part of the supply chains,” said Chin, “there is greater risk in not having a level of transparency. The risk to a single patient is so much greater than if you’re building out batches and testing and validating [them] because you have a direct line to a person that you’re hoping to help. This could go in so many different directions because it is a relatively uncharted territory, so companies are realizing very quickly that software and automation can provide a level of transparency and risk mitigation that doesn’t exist today. I think they are faster to adopt those solutions because they have a greater need, a bigger monkey on their back than I would venture to guess is the case for traditional biopharmaceuticals. It’s that one patient–one solution that’s a big responsibility for anyone.”

LeDoux thinks that the bulk of antibody manufacturers are changing their strategy, too. “Typically, if you think back to a few years ago, everything was based on islands of automation. You had unit operations. Each of those operations had their own control platform, and they didn’t necessarily speak to a distributed control system (DCS). I think over 50% of manufacturing now incorporates a DCS type of solution. All the batch records are mustered in a single location. They are electronically based.”

Stanton asked whether that would include incorporating real-time monitoring. LeDoux thinks that it will, adding that “I don’t know how centralized those data are in terms of all of the assays done. I’m thinking of pure manufacturing steps, though, with a tendency toward going to a DCS approach, a plant-wide system rather than islands of automation.”

Morgret commented that having real-time monitoring and all parameters available in an electronic system is useful in early development as a company is scaling up and moving into a new facility. The company may have only limited process understanding for an accelerated program, for example, so additional monitoring early on could help ensure that product made at scale is the same as what was made in the laboratory. Regulatory agencies also expect that such monitoring continues throughout the course of a product life cycle, making data available in a central location for tracking and trending.

Real-Time Analytics: Stanton asked how the lack of real-time analytics may contribute to existing bottlenecks. “If the industry can incorporate real-time analytics and overcome some issues with assays and such, then we’re into clear sailing. Or am I too optimistic?”
Morgret responded that ideally, to run a process, technicians could just set a reaction time and temperature and let it run. But the reality is that in the early stages, a company just doesn’t have the details figured out yet. So real-time analytics are important in the early stages of product and process development. “When transferring a process to a new facility, you have to get those assays validated in the quality control (QC) laboratory and incorporate the time it takes to run them. The ability to do that as fast as possible is imperative.”

Future Expectations
Stanton concluded the discussion by asking whether anything in the manufacture of antibodies (or perhaps in emerging therapies) might be “a ticking time-bomb” problem waiting to explode. He asked whether the panelists worry about another type of bottleneck in the future.

Uncertain as to whether it’s an actual bottleneck, Chin noted that the automation component should solve some challenges specific to transparency. “I don’t see it as a bottleneck. I just wish industry would adopt it faster.”

LeDoux added that “the whole economic model around antibody manufacturing in development has the potential to change, and that’s obviously a huge driver for the industry over time.

Leave a Reply