The current discussion of downstream bottlenecks arises from the ease with which up-stream activities generate protein and the desire for downstream operations to match upstream productivity.
“In theory there really is no mass limit upstream,” says Günter Jagschies, Ph.D., senior director at GE Healthcare Life Sciences. “But when you reach purification you are limited to how much material downstream equipment can handle.”
Many legacy facilities, even newer plants, suffer from downstream bottlenecks or, perhaps more aptly, “capacity mismatches” between upstream and downstream operations. “Legacy plant design issues are what led to the discussion of downstream bottlenecks in the first place,” Dr. Jagschies notes.
The very concept of bottlenecks is relative. Overall process efficiencies are as much as 20-fold higher, in fully optimized facilities, than a decade ago. What we now refer to as bottlenecks might have seemed lightning-fast to bioprocessors in 2003.
Productivity, whether calculated by reactor volume or kilos of finished product per unit time, have also skyrocketed. Upstream and downstream bioprocessing may actually be reaching a healthy equilibrium of sorts. Dr. Jagschies observes that ten years ago “the economics on the upstream side were lousy. Now you have better economic balance between upstream and downstream.”
Cell culture-related activities take far longer than all the remaining unit operations combined. Yet, bioprocessors recognize the unique issues related to downstream bottlenecks, which are akin to what chemists call a “rate-limiting step” in a reaction mechanism.
Bottlenecks are operations that are time-consuming relative to steps that precede or follow, often resulting in equipment or facility under-utilization, product instability, reduced efficiency, and higher process-related costs.
“When we speak with downstream processing groups, they still mention the classic bottlenecks of chromatography, buffer capacity, and cleaning,” says Kimo Sanderson, vp for client services at Asahi Kasei Bioprocess. “So yes, downstream bottlenecks are real, but we’re seeing some relief in terms of their severity.”
Single-use equipment, in-line buffer dilution, high-capacity chromatography resins, and a deeper appreciation for modeling and scheduling software, have helped to mitigate some of the more serious bottlenecks, Sanderson adds. “And we’re beginning to see the idea of debottlenecking taken to the next level, in the form of continuous downstream processing,” notes Sanderson.
A downstream bottleneck is any step that requires more time or resources than it “should.” Perception plays into this view, however. Rising upstream productivity can make certain downstream operations appear inefficient when in fact they are just as efficient, or in some cases more efficient, than they were a few years back.
“It’s not always fair to call these bottlenecks, but when in the heat of the moment you can see, smell, taste, and hear the inefficiency of a particular step, you want to improve it,” Sanderson remarks.
Despite advances in downstream equipment, the overall perception is that purification will never quite catch up with protein expression due to cost, facility, and time constraints. Separation products are expensive and difficult to scale, and processing floor space cannot be created from thin air.
“And the longer a target protein remains in the presence of contaminating proteins, the greater the risk to its stability, product yield, and purity,” says Yamuna Dasarathy, Ph.D., director of chromatography marketing at Pall Life Sciences. “Any technology that accelerates the target-isolation process will mitigate the risk and de-bottleneck the entire process flow.”
Dr. Dasarathy identifies chromatography as one of the chief downstream bottlenecks. Column packing, testing, qualifying, and validation of the packing followed by processing of the feed stream on the column all add to the cycle time and slow down the purification process flow.
To avoid or mitigate bottlenecks, Dr. Dasarathy suggests using flexible facilities, single-use technologies, pre-packed chromatography columns, and membrane chromatography technology. Reducing the number of purification steps and/or unit operations (process intensification) is another strategy, but difficult to achieve.
Christine Gebski, head of the Poros resin business unit at Life Technologies, identifies several significant areas where bottlenecks tend to crop up.
In clarification, which handles the largest downstream volumes, processors typically employ multiple unit operations. “There’s no one single solution,” Gebski explains. “It usually takes a combination of unit operations to move from cell culture into a clarified harvest material suitable for further downstream processing.”
Other bottlenecks arise in “supporting” unit operations between chromatography steps, where tangential flow filtration may be required to change pH or buffer systems. Such steps are typically undertaken due to limitations in chromatography resin performance, which are often related to low operating flow rates.
Still other bottlenecks result from unit operations implemented in early development, but which have not kept pace with improvements in surrounding steps. “They become the status quo,” Gebski adds.
Even when the rising tide of process efficiency raises the productivity of all unit operations, quality issues arise, related to impurity levels that require lengthier purification or additional steps. Higher titers will generally come at the cost of higher impurity levels, but in some situations the contaminants may exist at levels that are disproportionate to improvements in protein titer.
Fabien Rousset, deputy director and head of biopharma technologies at Novasep, prefers to refer to delays and holdups during production as “efficiency bottlenecks” rather than process or unit operation bottlenecks. “Nowadays, the objective is to reduce holding times and improve productivity,” he says.
Traditional biopharmaceutical processes consist of discrete, discontinuous unit operations, which tends to slow down production and adversely affect product shelf-life and stability. Achieving higher productivity is possible, although optimization of individual scale is of limited value. “The use of continuous technologies in USP and DSP is certainly the best way to overcome these limitations,” Rousset says.
With the increasing adoption of perfusion cell culture upstream, he believes that continuous downstream processing will be the next step, resulting in fully continuous production. This significant mindset change is what drove Novasep to develop its BioSC® continuous chromatography system, which reduces process time while optimizing buffer and resin consumption, explains Rousset.
Another bottleneck Rousset identifies is “time-to-API,” which is related to the duration of process development—the “time bottleneck” in his parlance. The development of dedicated manufacturing platforms, as is now common in monoclonal antibody production, is a good way to address time-to-API.
“The main challenge, then will be to reconcile these efficiency and time bottlenecks,” Rousset adds. “The one-size-fits-all platform processes must address not only development time constraints, but also large-scale manufacturing efficiencies. This should in turn drive the industry toward adopting continuous technologies in standard bioprocesses.”
Disposable processing has, in some cases, drastically reduced development timelines and validation costs during process development and clinical phases. However, their application at commercial scale continues to generate issues, according to Rousset. “Disposables may not be cost-effective at large production scale, and are often replaced by reusable equipment.”
In those situations, the availability of materially equivalent products must be addressed from the earliest development phases to ease scale-up efforts afterwards.
It is now more or less a given that downstream bioprocess bottlenecks arise from rising upstream titers. Where large cell culture volumes once required larger column diameters for capture, the challenge now is dealing with more concentrated process streams.
Francis Bach, director for purification technologies at Sartorius Stedim Biotech, believes that most downstream process bottlenecks are company- or process-specific. “They are almost self-induced in many cases,” he observes. “The term ‘bottleneck’ is a generic label to get people to stop and think about what they’re doing and put in better processes.”
Many bottlenecks arise while processes are still nascent, and persist throughout the molecule’s development and manufacturing life. Development-stage companies are constantly rushing to produce material, whether a few milligrams for toxicology studies or a few grams for an early-stage human trial, Bach says.
Processes developed for preclinical materials are applied to clinical-grade proteins, sometimes with minimal tweaking. And if the molecule advances to human studies, many aspects of the “legacy” process remain.
“So if you look at it, clinical trials are actually the bottleneck,” says Bach, referring to what is by far the “slow step” in drug development, as well as to its after-effects in the case that the drug is approved. “This is no doubt a radically different perspective than what you’ve heard,” he continues.
Post-approval bottlenecks depend on the molecule. Many manufacturing hold-ups have been eliminated through platforming the manufacture of molecule classes, e.g., monoclonal antibodies.
“Industry has done a nice job selecting platforms for mAbs, in selecting processes, vendors, and suppliers to help minimize bottlenecks,” believes Bach. For marketed drugs manufactured at large scale, the most significant bottlenecks, he maintains, are reducing buffer consumption and volumes through the use of smaller columns, more efficient filtration, and membrane chromatography.
Manufacturers of non-mAb proteins still face challenges, Bach says, because the molecular diversity of their products makes platforming difficult.
One way to avoid downstream bottlenecks, according to Bach, is for bioprocessors to work with vendors on new purification products and implementations thereof. Collaboration also helps vendors to improve their offerings.
Many drug sponsors are averse to this sort of relationship, fearing that intellectual property will be compromised. Yet Bach says “there are no secrets” in the operation of most purification steps. “From the vendor’s perspective, it’s just another combination of buffers and a molecule.”