Continuous Manufacturing: A Changing Processing Paradigm

Although continuous manufacturing is well established for bulk chemicals, complex automation and validation challenges limit continuous applications for biologics manufacturing. While only a few approved biologic products in the market-less than 10%, according to estimates by Eric Langer, managing partner, BioPlan Associates-are manufactured through perfusion or continuous downstream processing, the field is evolving (1). Most of these processes are done piecewise, however, and are not truly continuous. 

Langer points out that leaders such as Genzyme, Bayer, Janssen, Merck-Serono, Novartis, and Lonza for Eli Lilly have manufactured approximately 19 marketed recombinant protein products/monoclonal antibody (mAb) products using perfusion or elements of continuous processing, and these products are predominantly blockbusters with annual revenues totaling approximately $20 billion (1). Actual use of perfusion for commercial manufacture may even be more widespread, but bioprocessing manufacturers are becoming increasingly reticent to publicly share details about their processes, he says.

Presenters of a BioPharm International webinar, “The Future of Continuous Downstream Bioprocessing,” cite continuous biomanufacturing initiatives at an Amgen Singapore plant as an example of how manufacturers are embracing the changing processing paradigm (2). GlaxoSmithKline, Johnson & Johnson, Genzyme, Bristol-Myers Squibb, AstraZeneca, Samsung BioLogics, and Novartis AG are among the pharmaceutical companies and contract organizations building or expanding biotechnology facilities to make drugs in innovative ways. BioMarin and Vertex also plan to incorporate elements of continuous processing in some of their operations.

While adoption of continuous processes may seem slow, single-use technologies themselves are still fairly new. Single-use process technologies took “20 years to be fully adopted and accepted in clinical manufacturing, and they are only just making their way through commercial manufacture,” says Andrew Sinclair, president and founder of Biopharm Services, a technology company that serves the biopharmaceutical manufacturing business. “About five years ago, as the industry matured, the manufacturers were focusing on flexibility, cost, and capital efficiency in manufacturing,” he says. “It was at that point that continuous technology was considered for downstream processing, so we are really at the early stage of a normal technology adoption cycle.”

Using continuous processing for biologics presents a way to address productivity improvement challenges, and can also offer the opportunity to implement standardization for biologics (3). Employing straight-through processing in mAb purification could significantly improve throughput with a much smaller manufacturing footprint, presenters of the aforementioned webinar conclude. Case studies with mAbs have shown similar yield and purity compared with batch runs and could be associated with an increase in productivity, they note. Other advantages of continuous manufacturing include consistent product quality; smaller equipment; streamlined processes; low process cycle times; reduced operating costs; increased flexibility; elimination of hold tanks and unit operations that do not provide value; high equipment utilization rates; high volumetric productivity; more automation coupled with less human interaction; an increased use of single-use equipment; and reduced inventory and storage needs (1, 3).

FDA support for quality improvements
FDA has been a strong supporter of continuous processing as early as 2004 when it released Pharmaceutical cGMPs for the 21st Century–A Risk-Based Approach. In addition to reduced inventory, lower capital costs, a smaller ecological footprint, and more flexible operation, FDA is an advocate of the fact that continuous manufacturing reduces manual handling of products and allows for better process control. An agency representative said at an International Forum and Exhibition on Process Analytical Chemistry Annual Meeting in 2012 that the on-line monitoring characteristic of continuous processing can facilitate real-time testing approaches and can support FDA’s quality-by-design (QbD) initiatives (4). Biotechnology products, when run continuously, can be sampled regularly, and any process contamination can be resolved more easily without having to discard entire batches.

To take regular samples, a robust in-line monitoring system linked to feedback and feedforward controllers is necessary, points out Sinclair. Traditional sampling and offline testing typically done for a batch process in a continuous system is challenging, says Robert Snow, principal engineer at Genzyme, “as it only represents a point in time, not the complete ‘batch,’ given the continuous nature.” For this reason, Snow attests, for continuous manufacturing, online analytical testing methods are preferable. “Critical process parameters are also needed to control the process and maintain the steady-state condition,” comments Snow.

Although FDA meeting notes have claimed that “There are no regulatory hurdles for implementing continuous manufacturing (4),” there are some regulatory issues that persist with the process for biologics, such as performing quality assurance/quality control with products manufactured through a continuous system as well as defining lots and batches (1). The main concern lies in the fact that a batch can be defined based on quantity manufactured or duration of the process (5). Sinclair says the issue of batch identity is up to the user to define prior to manufacture and that “the real regulatory issue is one of traceability when there is a problem that has consequences of a recall for a manufacturer.”

While many of the regulatory expectations related to quality for continuous manufacturing are the same as those for batch processing, Allison et al. points out that in continuous processing, sampling considerations will differ from batch; deviations might need to be handled differently; variability should be controlled; manufacturing changes managed; and importantly, “the rationale for testing of a continuous batch must be reconciled against the traditional paradigm” (5). Konstantin B. Konstantinov writes that FDA is in the position to offer incentives to vendors “to develop required continuous unit operations (particularly downstream) and PAT [process analytical technology] instrumentation (3).” These technologies could offer enhanced information about product quality attributes, such as activity, aggregation, glycosylation, and impurity levels, he adds.

Making the switch: Perceived barriers to implementation
There are often too many steps in a batch model, and there are limitations in flexibility, as process controls are not dynamic (2). Batch processing is becoming a process bottleneck, and the industry is looking for more efficient ways to manufacture biologics. Continuous manufacturing options offer these benefits, according to industry experts, and a continuous unit operation can process a continuous flow input for a prolonged period of time (3). “There is also an added benefit related to actual run time,” explains Snow. “In a batch system there are many non-productive days associated with startup and turnaround between batches.”

Even though continuous manufacturing/perfusion has been around since the 1980s–first used for fused-cell hybridoma culture–it is likely to enjoy a comeback now that so many biologic therapies are in the pipeline, writes Langer. Implementing a switch from batch to continuous processing, however, could be perceived as a challenge. If overall bioprocess yields continue to double every five years, switching from batch bioprocessing to continuous methods may prove even more difficult (1). Making the switch, says Snow, will require an understanding of the continuous process “well enough to achieve steady-state conditions.”

James Evans, former associate director of the Novartis-MIT Center for Continuous Manufacturing and current director of API technical services at Hospira, said at an American Association of Pharmaceutical Scientists meeting in 2010 that implementing integrated continuous biomanufacturing would “require the implementation of quality-by-design principles, new product development processes, new facility layouts,” and improvements in the technical skills of the engineers and other professionals who run these processes (6). “Upfront investment [for continuous manufacturing] is more significant than batch processing,” Evans told BioPharm International. To properly deliver continuous manufacturing, a company must invest upfront in process understanding, he adds, and “with failure rate of products in the clinic, this could become a significant hurdle.”

George Barringer, PhD, business development executive for Stratophase asserts that “Resources historically allocated to the many manual operations in batch and fed-batch systems will need to be reallocated to positions supervising a highly automated hands-free process. Equipment and instrumentation will have to be modified and developed to handle a continuous flow of material from raw input to finished product,” he says, and “accurate, on-line, real-time monitoring instrumentation will be required for continuous manufacturing to succeed.”

Resistance to adoption of continuous manufacturing upstream has “historically been based on the unreliability and complexity of equipment such as cell retention devices, bioreactors, and associated control equipment,” say Steve Tingley, vice-president, BioProcess Sales and Marketing, Repligen, and John Bonham-Carter, director, ATF Systems, but they add that many of these concerns are no longer relevant because of the invention of more reliable bioreactor controllers and devices. “Adoption of continuous downstream is limited by technological immaturity,” say Tingley and Bonham-Carter. They also note that continuous capture typically requires sterile chromatography columns. “For downstream polishing steps, the infancy of multi-column chromatography systems is a limitation,” they add.

A change from batch to continuous manufacturing may necessitate new equipment, process control parameters, and control strategies to establish product equivalency, notes Allison et al. (5). Plus, if a manufacturing change occurs post-drug-approval, changes should be summarized and justified with bioequivalence studies, the authors write.

Costs and benefits of continuous manufacturing
The complexity of continuous manufacturing could make the switch unattractive to prospective users, and there are also concerns surrounding process development control, contamination risks, and scale-up potential of traditionally batch-based systems (1). In addition, there is some trepidation surrounding the integration of upstream and downstream processes and the interaction between unit operations, as it is well documented that changes upstream affect other steps down the line (7).

There are also some worries about the overall novelty of the technology in the continuous space, as not all equipment needed to perform continuous processes may yet be available for GMP manufacturing. Furthermore, both stainless-steel material and single-use equipment were not necessarily meant for constant use, and this type of wear and tear may accelerate product breakdown (1).

Upstream processing
Perfusion is increasingly being adopted by a number of manufacturers, wherein material is simultaneously charged and discharged from a processing system. Perfusion implementation has been dramatic over the past few years, note Tingley and Bonham-Carter. “Currently, only a minority of bioreactors are specified with perfusion capability; however, there is a clear awareness that new facilities must be flexible and adaptable to perfusion technology,” assert Tingley and Bonham-Carter. “In clinical facilities today, we estimate that approximately 20% of biologics utilize perfusion in their upstream operations.”

Generally, product variability is reduced with perfusion. Rick Johnston, principal of Bio-G, writes, however, that there is greater variability for drugs in short runs such as what would be typically seen for orphan drugs. He says that until perfusion becomes standard practice in the industry, batch-fed operations may continue to reign for companies developing a large portfolio of short-run manufacturing products, and “scaling process development is easier using fed-batch technologies” (1). Langer also writes that, based on data from his yearly report, “there has been no distinct trend for bioprocessing professionals to select perfusion over fed batch” (1). Even so, because perfusion involves continuously introducing fresh media, there is less accumulation of toxic waste products, degraded DNA, and other debris than is associated with products that remain in a bioreactor for extended periods of time during batch processing. In addition, writes Langer, “lag phases are eliminated because cell culture is always at or near peak efficiency” (1). Elimination of process collection and hold steps is a major benefit of continuous processing, concurs Snow, as product degradation and aggregation typically happen in these phases. Langer notes that perfusion “can enable manufacture of inherently less-stable and labile proteins” and is less stressful in general to proteins, resulting in more consistent, better-quality products than that which would be produced with batch processing (1). BioPlan predicts that in five years, 50% or more of new bioprocessing lines and facilities will incorporate some elements of continuous processing, and it will likely be in the area of perfusion (1).

While perfusion is widely used in bioprocessing, some upstream technologies still need to be developed and optimized, writes Konstantinov. These include automatic cell density control, foam control, oxygenation, and ventilation-and cell lines must be stable and robust enough to maintain high productivity over time (3).

Downstream processing
Many industry representatives recognize downstream manufacturing as the most difficult component of continuous processing, and experience with it is limited (3). Sinclair says that although perfusion bioreactors have been running upstream for decades, and the technology, supply chain, and operating experience already exists, for downstream manufacturing, “the supplier side of the technology is immature.” Despite these technology gaps, downstream processing in a continuous model is much more expedient, Sinclair points out; processing takes hours instead of days, potentially leading to improved product quality.

As the mass of protein to be purified increases, required buffer and elution volumes in batch-capture chromatography increase (8). A continuous capture option could be a good alternative to reduce column size and lower media utilization. To reduce consumption of chromatography media such as buffers, resins, and solvents, many companies are looking into multi-column chromatography, specifically, multi-column countercurrent solvent gradient purification (MCSGP). In batch chromatography, a resin is often not loaded at maximum capacity, and capture columns generally have to be cycled multiple times. A column in batch is typically loaded only approximately 80%, whereas multi-column chromatography can allow complete saturation and can increase working chromatography capacity (1). Therefore, multi-column systems are ideal for those looking to maximize productivity and improve resin capacity utilization. “The common method for continuous chromatography is to sequence multiple columns in series–by doing so, the columns can be loaded to 100% capacity and any product flow-through will be captured by the next column in series,” comments Snow. Sinclair notes the use of continuous columns results in “a 20-30% increase in resin utilization when compared to a batch process.”

Low-flow rates characteristic of continuous processing allow for small column size, notes Snow. The column size can be reduced 10 to 20 times smaller than normal, estimates Sinclair. As a result, he says, “The ability to match the scale of the plant to any of the commercial throughputs means that these facilities can readily meet demand.”

A multi-column chromatography system starts with a complex mixture, which is loaded into the system and goes through a series of washing, elution, regeneration, and equilibration steps. These procedures repeat in subsequent columns, helping to balance downstream productivity with upstream titers. Several equipment manufacturers, such as Novasep, Tarpon Biosystems, GE Healthcare, Semba, and ChromaCon, offer continuous chromatography systems.

Konstantinov writes that in addition to multi-column technology, downstream operations could benefit from novel continuous viral inactivation and ultrafiltration/diafiltration unit operations. To standardize downstream operations, he says one option would be “the rational design of highly specific ligands for the capture step analogous to Protein A in mAb purification” (3).

Costs of adoption
While it may be a significant investment to totally replace legacy batch systems, over time, continuous biomanufacturing and perfusion are thought to offer cost savings within the range of 30–50% (8, 1). A traditional manufacturing facility is estimated to cost $150 million to construct, while a continuous plant would cost significantly less, analyst Marcus Ehrhardt from PwC told The Wall Street Journal (9). Ehrhardt told BioPharm International, however, that the price of construction ranges depending on capacity, technology, location, etc., and that the estimated price applies to small-molecule drug manufacturing facilities only. Biologics plants are much more expensive, he asserted. The initial cost of construction for a continuous plant could exceed $30 million; this was the amount for the recent Vertex continuous manufacturing plant. GlaxoSmithKline’s new hybrid continuous-batch synthesis manufacturing plant in Jurong, Singapore-announced in October 2014 and predicted to open in 2016-cost 19 million euros (approximately $25 million) to build. Bonham-Carter attests that, for existing facilities looking to retrofit, the cost of new continuous/perfusion equipment is not that high, but “the real cost is adapting the facility and the validation workload.” Understanding this cost and ensuring the gain from the new technology is sufficient prior to implementation is necessary, he says. As a result, for facilities looking at small or hybrid steps toward continuous operation, “Output can be boosted 10–30% in a traditional fed-batch facility for equipment costs of single-digit millions.”

The costs of operation for a continuous manufacturing plant vary based on whether site managers use single-use equipment or stainless-steel operations. Perfusion typically costs $44.1 million when using stainless steel per 500 kg/year, but using single-use technologies for perfusion reduces operating costs approximately $11 million, according to data from Refine Technology (1). “[Continuous] facilities are much smaller for a given throughput, which results in smaller capital and reduced operations,” notes Sinclair. “As a consequence, they should be quicker to build, especially when combined with single-use technologies.” Bonham-Carter notes that using smaller reactors, coupled with less downtime through continuous use at high cell density, “means capital demands may be at least 10-fold lower than historically required for a similar annual output.”

Recent modernization efforts
Some companies need only upgrade existing facilities to incorporate continuous methods, rather than build from the ground up. Hemispherx Biopharma, which produces a natural alpha interferon for the treatment of refractory or recurring external genital warts, halted the manufacture of its drug in 2008 because of high labor costs and low capacity, which the company said in a release significantly limited the commercialization potential of the treatment (10). After $8 million in facility upgrades, the company integrated continuous manufacturing throughout the process. It scaled up from hundreds of small flasks to a 600-L bioreactor, effectively eliminating approximately 80% of the workforce previously needed for this part of the process. While the process improvement may also improve cost efficiency, the company said that FDA will still need to reaffirm the amended biologic license application for the facility before the sales of Hemispherx’s drug can commence.

GlaxoSmithKline has employed a number of reactor technologies that are helping it make the move to continuous manufacturing, says Mark Buswell, vice-president of advanced manufacturing technology at GlaxoSmithKline, including Ehrfeld reactor systems and Corning reactors. The company’s focus is currently on continuous manufacturing for small molecules, such as its respiratory APIs. In terms of drug product, the company has made advances in continuous blending and wet granulation to overcome particle flow and segregation issues. “For drug product manufacturing, a lot of the unit operations, such as tablet compression, are actually continuous operations,” Buswell points out. “The industry has historically chosen to operate them in batch mode, however, mostly to maintain genealogy of material-and the industry needs to overcome this mindset.” Buswell says the company is interested in continuous fermentation for large molecules, but that this “is a ways off.”

Amgen’s $200 million facility in Singapore is a recent example of a facility built to incorporate continuous processing. The company’s continuous “Next-Generation Biomanufacturing Facility” uses single-use bioreactors, disposable plastic containers, real-time quality analysis, and also boasts modular design. It is expected to use fewer resources and increase bulk production capabilities, which is estimated to result in “a cost reduction of 60% or more per gram of protein,” according to an Amgen statement (11). The facility is slated to start the production of biologic therapies in 2017.

In a 10-K filing released in 2014, Vertex said it plans to use continuous manufacturing process in its new Boston, MA facility to manufacture co-formulated lumacaftor and ivacaftor tablets for the treatment of cystic fibrosis (12). While the company is among the few that are spearheading process improvements via the continuous pathway, it acknowledges there may be bumps in the road to production, and even says that any failure to establish and validate its manufacturing process could negatively affect its ability to obtain approval for or launch of lumacaftor in combination with ivacaftor. “While we believe continuous process manufacturing has a number of benefits, we have not previously designed a continuous manufacturing process to produce commercial quantities of a pharmaceutical product, and we believe that we would be the first company to seek approval for an NDA [new drug application] or an MAA [marketing authorization application] using this method of manufacturing. As a result, we also have designed and tested a non-continuous process for manufacturing co-formulated lumacaftor and ivacaftor tablets that we would seek to utilize if we experience delays associated with the continuous manufacturing process.”

End-to-end
In 2014, Genzyme filed a patent, called “Integrated Continuous Manufacturing of Therapeutic Protein Drug Substances,” for a platform that will integrate upstream and downstream processes (13). The platform flows a liquid culture medium from a perfusion bioreactor into a preliminary multi-column chromatography system (MCCS1), then the eluate of the MCCS1 containing the recombinant therapeutic protein is continuously fed into a second multi-column chromatography system (MCCS2), where it is purified and polished. In an abstract presented at the Biomanufacturing and Process Development’s Continuous Bioprocess Symposium in 2014, Genzyme said the platform performed protein drug capture, viral inactivation, and in-line buffer dilution, along with intermediate and final polishing purification steps to generate the desired drug substance. The presenter concluded that the findings demonstrated “the potential of integrated fully continuous bioprocessing as a universal platform for the manufacture of various kinds of therapeutic proteins” (14).

Novartis gave MIT $65 million in 2007 for a decade-long research program to investigate an end-to-end approach to continuous processing (15), and the Novartis-MIT Center for Continuous Manufacturing successfully churned out the first end-to-end, completely continuous synthesis and formulation of a treatment for high blood pressure in 2013, aliskiren hemifumarate (16). While this drug is not a biologic, Bernhardt L. Trout, director of the Novartis-MIT Center for Continuous Manufacturing, says that end-to-end processing can absolutely be applied to biologics. Trout’s focus is not strictly on a continuous flow; rather, his team focuses on “a systems approach, end-to-end integration, and model-based control.” He says that all biologics could potentially be candidates for continuous processing. “Our concept of continuous is based on where the whole industry should be in the future,” Trout told BioPharm International. MIT is also currently working on new technologies for the continuous manufacture of biologics in the chemical engineering department; this work is separate from the work being done in the Novartis-MIT Center. “There is no current technical limitation to doing an end-to-end process for a biologic therapy-it’s just a matter of someone doing it,” he adds.

Typically, the only continuous unit operation in a hybrid system is a bioreactor that has been fashioned with a cell retention device (3). While many commercial drugs are made through perfusion, the new area of interest is running the downstream portion of the process continuously. Even though “this concept is in its infancy,” according to Sinclair, “Many major biopharma companies are evaluating downstream continuous processing for new clinical projects and many more (such as Pfizer, Genzyme, Bayer, etc.) have published papers about its potential.” Full process implementation of continuous processing is still in its early days, say Tingley and Bonham-Carter.

There may even be some products that are much more suited to closed processing or hybrid systems than fed-batch systems, stresses Sinclair. “For example, any proteins that are not particularly stable and those that auto-catalyze (such as blood-clotting proteins) cannot be easily made in a batch mode and benefit greatly from running a purification train continuously.” Stable proteins could also benefit from being run continuously, as glycosylation changes in stable proteins are possible in batch systems (3). Despite the fact that all biologics could be manufactured via continuous processes, not all necessarily would be suitable candidates for this method. Elements such as product and process compatibility with available technologies, as well as product and market demand, could factor into the decision to incorporate continuous techniques in biological therapy manufacturing, says Snow.

Companies that are incorporating end-to-end processing in their R&D operations have the opportunity to be “champions in advancing the best continuous processing technologies to clinical and commercial production over the next decade,” according to Tingley and Bonham-Carter. “Continuous bioprocessing will not just prove to be feasible, but will also become a strategic core competence to therapeutic drug manufacturers. Process speed, tighter quality control, and lower costs achieved by combining highly skilled staff with automated processes provide a competitive advantage for the next generation of biological medicines,” they conclude.

References
1. BioPlan Associates, “Continuous Bioprocessing and Perfusion: Increased Adoption Expected,” in 11th Annual Report and Survey of Biopharmaceutical Manufacturing, E. Langer, Ed. (April 2014), pp. 103-116.
2. S. Ghose, R. Nordberg, and A. Forss, “The Future of Continuous Downstream Processing,” on-demand webcast on BioPharm International, https://event.on24.com/eventRegistration/EventLobbyServlet?target=registration.jsp&eventid=890993&sessionid=1&key=63C475F147ABC8B55EFC1F0A20D92723&sourcepage=register, accessed Feb. 20, 2015.
3. K.B. Konstantinov and C.L. Cooney, “White Paper on Continuous Bioprocessing,” presentation at the International Symposium on Continuous Manufacturing of Pharmaceuticals: Implementation, Technology & Regulatory (Cambridge, MA, 2014).
4. S. Chatterjee, “FDA Perspective on Continuous Manufacturing,” presentation at the IFPAC Annual Meeting (Baltimore, MD, 2012).
5. G. Allison et al., “Regulatory and Quality Considerations for Continuous Manufacturing,” presentation at the International Symposium on Continuous Manufacturing of Pharmaceuticals: Implementation, Technology & Regulatory (Cambridge, MA, 2014).
6. P. Thomas, “With Continuous Operations, Can Drug Manufacturing Become a Rock Star?,” http://www.pharmaqbd.com/novartis-mit_evans_continuous_manufacturing/, accessed on February 23, 2015.
7. C. Challener, "Bioprocessing & Sterile Manufacturing," supplement to Pharm. Technol. 38 (2014). 
8. M. Bisschops et al., "Single-Use, Continuous-Countercurrent, Multicolumn Chromatography," supplement to BioProcess Internat. (2009), http://www.bioprocessintl.com/downstream-processing/chromatography/single-use-continuous-countercurrent-multicolumn-chromatography-184186/, accessed Feb. 15, 2015.
9. J.D. Rockoff, “Drug Making Breaks Away From Its Old Ways,” The Wall Street Journal, Feb. 8, 2015, http://www.wsj.com/articles/drug-making-breaks-away-from-its-old-ways-1423444049, accessed Feb. 22, 2015.
10. Hemispherx Biopharma, “Hemispherx Biopharma Announces the Completion of the Newly Upgraded Alferon Facility: Commercial Production of Alferon is in the Final Stage Integrating Continuous Flow Manufacturing Technology,” Press Release, March 2, 2015.
11. Amgen, “Amgen Outlines Strategy, Growth Objectives And Capital Allocation Plans,” Press Release, Oct. 28, 2014.
12. Vertex, Form 10-K Annual Report, Filed Feb 11, 2014, http://investors.vrtx.com/secfiling.cfm?filingID=875320-14-15& CIK=875320, accessed March 1, 2015.
13. Genzyme, Integrated Continuous Manufacturing of Therapeutic Protein Drug Substances, US patent 20140255994 A1, March 3, 2014.
14. R. Patil, “Integrated Fully Continuous Production of Recombinant Therapeutic Proteins,” presentation at the Biomanufacturing and Process Development’s Continuous Bioprocess Symposium (Durham, NC, 2014).
15. T. Wallack, “Novartis to give MIT $65m to find new way to produce drugs,” The Boston Globe, Sept. 28, 2007, http://www.boston.com/business/articles/2007/09/28/novartis_to_give_mit_65m_to_find_new_way_to_produce_drugs/?page=full, accessed March 1, 2015.
16. M. Peplow, “Organic synthesis: The robo-chemist,” Nature 512, pp. 20-22 (August 07, 2014), doi:10.1038/512020a.

Article DetailsBioPharm International
Vol. 28, No. 4
Pages: 20–27, 41
Citation: When referring to this article, please cite it as R. Hernandez, “Continuous Manufacturing: A Changing Processing Paradigm,” BioPharm International28 (4) 2015.