Cynthia A. Challener is a contributing editor to Pharmaceutical Technology.

Cynthia A. Challener

Mammalian cell lines, particularly Chinese hamster ovary (CHO) cell lines, have become the standard expression systems for the production of biologic drug substances from recombinant proteins to more complex monoclonal antibodies (mAbs). CHO cells have played a significant role in the manufacture of revolutionary drugs for the treatment of many diseases, and their use is still the focus of major investment among biopharma companies worldwide, according to Barry Holtz, president of iBio. Although the standard, they do possess limitations that need to be addressed as the biopharmaceutical industry evolves to meet government, payer, and patient expectations for cost-effective, safe, and efficacious medicines. In addition, conventional mammalian cell lines may be inappropriate for the production of next-generation medicines such as bi/multi-specific antibodies and gene and cell therapies.

	High cost and long development timelines are two major drawbacks of conventional expression systems, according to Mike Laird, senior director and principal scientist for process development at Genentech. These systems also have the potential for low expression levels for some novel protein structures.

	The biggest issue, asserts Holtz, is the time involved in the development of mammalian expression systems. “Traditional mammalian cell-based systems require the development of stable cell lines that perform well, which requires the completion of multiple evaluations in small reactors over several months. While the product is well-characterized at that point, scale up to larger reactors is often needed, and the environmental changes in bigger vessels, whether single use or stainless steel, can impact the post-translational modification of the protein, which can cause problems and delays in the business timeline. At each step, the protein must be extensively characterized to assure efficacy and potency. All of this effort adds significant expense as well.”

	Therapeutic protein production using CHO is expensive, agrees Mark Emalfarb, founder, CEO, president, and director of Dyadic International. “Mammalian expression systems require costly upfront investments in manufacturing facilities and high material and production costs. In addition, CHO expression entails a relatively low mAb yield (low-single-digit g/L/d) and a long cycle time. Furthermore, CHO cell lines typically require two viral purification steps, which are not necessary for some alternative systems, such as the 

 (C1) fungal system we are developing. C1 has no viruses and thus the need for those purification steps is eliminated,” he observes.

	Another important issue relates to the fact that the current generation of mammalian expression systems has not seen the complex, non-natural protein formats currently found in discovery and thus their synthesis, folding, and secretion machinery has not evolved to handle such proteins, according to Andy Racher, associate director of future technologies at Lonza Pharma & Biotech. “These systems have limited ability to produce these new proteins with clinically relevant attributes and in clinically relevant amounts,” he notes. In addition, many new proteins contain three or four rather than one or two different polypeptides, and current expression vector formats are challenged by these new protein heteromers.

	“It is time to realize the limitations of CHO and look beyond it to explore newer and potentially more efficient drug development and production methods,” asserts Emalfarb.

	As the need to make biologic drugs more accessible and affordable to patients increases, the industry is ramping up its investigation of other manufacturing methods. The drawbacks of mammalian expression systems are also driving the exploration of new and alternative technologies to move many next-generation biologic drug candidates through later development stages.

	A number of “new” technologies will become more routine as the demands for new “designed” proteins strain the capabilities of conventional CHO cell systems, according to Holtz. One important approach is the engineering of new mammalian cell lines using new genetic editing technologies and the innovative design of gene constructs to optimize yields. He also notes that techniques to evaluate libraries of cell lines will help optimize expression and yield.

	Lonza, for instance, has developed a suite of multigene vectors where three, four, and possibly more different genes can be easily inserted into a single expression vector. “By putting all the genes into a single vector, all genes are ensured of being inserted into a transcriptionally active locus in the genome and being transcribed at high levels,” Racher says.

	There are also efforts to develop entirely different expression systems based on plants, baculovirus, bacteria (such as 

 in the Phoenix system), and yeast, many of which have already been demonstrated to get proteins to the clinic and licensure, according to Holtz.

	“In the not-too-distant future, it is likely that drug companies will evaluate two or more expression systems simultaneously as a routine best-practices approach in early stage development,” he comments.

	iBio’s plant-based system offers rapid evaluation of protein expression at a very low cost, according to Holtz. Because vectors are used to transfect the plant leaf cells, multiple constructs can be evaluated in parallel. Once infected, the plants produce the required proteins in less than seven days. At that point they are harvested, homogenized, and a clarified protein extract is ready for traditional protein separation and purification. In addition, scale-up is seamless and reproducible; each 10-g plant is an individual bioreactor, so it is only a matter of growing more plants and there are no issues around changes in protein structure or posttranslational modification, according to Holtz. He also notes that plant bioreactors are grown with no human or animal-derived materials and are not handled at any time by humans, which eliminates the chance that mammalian adventitious viruses will be present.

	Production of plant-made biologics been scaled-up by several companies in new facilities that can produce hundreds of kilos of mAbs and other therapeutic proteins per year. Successful antibodies (cancer vaccines and others) and other therapeutic proteins such as enzyme replacement therapies have been successfully taken to advanced clinical trials and some to licensure. In all cases, there have been no reports of adverse events associated with production of therapies in plants,” Holtz states.

	iBio grows 2.2 million plants continuously at its Texas facility and has worked with a variety of clients to produce mAbs, fusion proteins, antibody-drug conjugates, and vaccines, including virus-like particles (VLPs). “We have invested in increased product and process facilities and a cGMP-compliant pilot plant that-coupled with our large-scale manufacturing facility-assures clients that they can develop their protein through clinical trials and then be supported for the commercial launch of their products,” says Holtz. iBio will also transfer the technology to clients if they want to build and operate their own facilities.

	Emalfarb believes that the C1 fungal expression system may one day be a safe and efficient approach to speeding up the development, lowering the production costs, and improving the performance of vaccines and biologic drugs at flexible commercial scales.

	Dyadic’s Cl gene expression platform is based on technology originally developed for industrial biotech applications, such as biofuel and enzyme production, and sold to DuPont for $75 million in December 2015. The genetically modified strain of 

 is designed to produce enzymes and other proteins at a rapid rate. The company retained the rights to apply C1 to human and animal biopharma applications and has been investigating its use for the production of mAbs with humanized glycostructures; non-glycosylated mAbs, antibody fragments, FC fusion proteins, next-generation biologics, and other therapeutic proteins for which glycosylation structures are undesirable; and antigens, vaccines, and VLPs.

	“We are applying the power of an industrially proven gene-expression system that has been used by the likes of Abengoa, BASF, Dyadic, DuPont, and Shell Oil, among others, to produce industrial enzymes and proteins at greater than 100 g/L of total protein at up to 80% purity (80 g/L) at commercial scales greater by 25 times (500,000-L scale) or more than some of the largest CHO bioreactors (12,000-L scale) in one-half to one-third the time,” Emalfarb explains, noting that there is still room for yield improvement with C1. He adds that Dyadic has to date achieved a productivity for mAbs of 9 g/L in 90 hours or a 2.4 g/L/d production rate, which can be compared to 4 g/L in 336 hours or a 0.30 g/L/d for typical CHO processes, an eight-fold improvement.

	Emalfarb notes that the media cost for C1 is a fraction of that for CHO, there is no need for viral inactivation, and C1-expressed proteins are secreted from the cells in a purer form than those produced by CHO cells so are likely to be quicker and easier to purify.

	Dyadic is currently working with pharmaceutical companies that are researching its C1 platform to speed up the development and lower the cost of biologics, enable the development and commercialization of genes that are difficult to express at reasonable yields in CHO and other cell lines, and apply C1 for the production of larger quantities of proteins earlier in discovery and development. The company and its partners are also investigating the possibility of getting difficult-to-express genes that have potential as new and novel cures-but have been shelved due to lack of expression into the clinic-in a commercializable and affordable way, according to Emalfarb.

	In late 2017, Lonza introduced a new yeast-based expression system for the production of next-generation biologics. Its XS Pichia 2.0 Expression and Manufacturing Platform, based on 

, was designed to combine the best features of bacterial and mammalian systems in one system: fast and easy strain development and robust and rapid fermentation combined with a highly pure secreted product for a simple downstream processing, according to Christoph Kiziak, research and technology lead for microbial technology at Lonza Pharma & Biotech.

	“The driving force was to rethink the whole production strategy for producing proteins to circumvent the main bottlenecks of bacterial systems (e.g., intracellular production, endotoxin), CHO systems (e.g., time-consuming, viral clearance), and yeast systems (e.g., use of methanol, hyperglycosylation), while maintaining the use of 

 due to the general advantages and regulatory acceptance of this yeast cell,” Kiziak says.

	The “auto-inducible” setup of the new system makes it convenient for high-throughput clone screening, which results in highly pure material for preliminary quality analysis of the product at an early time point.

	In addition, fermentation development follows a product-specific model-based approach, which allows yeast fermentations to be performed in two to three days with a high volumetric productivity, according to Kiziak. It can also be expanded by an 

 model for process productivities over a wide production window. “This predictive model allows us to take into account production plant and process-specific limitations at any stage of development and provides high flexibility and quality for later production,” he explains.

	The methanol-free process also avoids the negative impact on cell viability and product quality of the commonly used AOX1 system, according to Kiziak. There is no need for explosion-proof facilities and the lower oxygen transfer rates compared to the AOX1 system provide additional flexibility regarding production plant requirements. There is also no need for endotoxin or viral clearance testing.

	Furthermore, the product is secreted into the culture supernatant, where the minimal medium together with the low host-cell protein background provides an ideal starting point for an efficient downstream process.

	To date, Lonza has focused on producing multi-specific novel antibody mimetics from various sources using the XS Pichia 2.0 and has achieved productivities of more than 2 g/L per day. The company is working to make the system even more customer friendly and easy to apply and to refine the model-based approach in order to improve the accuracy of the prediction of fermentation processes. Additional promoters with different strengths and induction profiles are also being developed to allow the tuned expression of helper factors, auxiliary proteins, heteromeric products, enzyme cascades, etc., which will expand the applicability of the XS Pichia 2.0 in the future, according to Kiziak.

	Regardless of the technology, there is a general acceptance that existing mammalian expression technology can no longer meet the needs of the biopharmaceutical industry. “One way to make healthcare more accessible and affordable to patients could be changing the cell lines we use for manufacturing,” Emalfarb observes. “Our goal is to bring affordable medicines to more patients, in addition to improving processes to develop new treatments,” he adds.

	The industry isn’t there just yet, however, according to Laird. “At this time, we are not aware of novel expression technologies appropriate for commercialization of next-gen biologics with significantly reduced costs, timelines, or complexity that can also ensure consistent post-translational modifications such as glycosylation. Although some new complex molecules could be harder to express using current or conventional mammalian expression systems, we think these systems are and will be the best approaches to express proteins for therapeutic purposes for the foreseeable future, especially given the vast knowledge from current advances in genome sequencing and CRISPR [clustered regularly interspaced short palindromic repeats] gene-editing technology that can be used to modify these conventional mammalian expression systems,” he explains.

	“With that said, we are very open to evaluating novel expression systems and will feverishly pursue new technologies as they become available,” Laird asserts. “We all have the same goal of delivering medicines to patients as quickly and efficiently as possible,” he concludes.

	Genentech is currently focused on the development of targeted integration mammalian cell lines to enable faster, more consistent medicine development with higher productivities. To date, engineered host-cell lines to optimize performance, increase productivity, ensure product quality, increase the ability to produce complex formats, and decrease timelines have been achieved, according to Laird.


	Vol. 42, No. 7
	July 2018
	Page: 28–31

	When referring to this article, please cite it as C. Challener, “The Search for Next-Gen Expression Systems" 

Articles

Biopharma seeks alternatives that meet the needs for next-gen biologic drug production.

The Search for Next-Gen Expression Systems

Demand for controlled-release formulations is expanding as patients’ expectations for greater convenience are increasing. Extended-release products have reduced dosing frequencies, which are particularly beneficial for pediatric and geriatric patients, but are appreciated by most patients and in general lead to improved medication adherence (1). Reformulation of branded drugs in controlled-release formats is one of the most common methods for extending the product lifecycle. Generic drugs are also often formulated with controlled-release properties (2).

	Controlled-release behavior can be achieved using polymer coatings on solid dosage forms or by the incorporation of various types of polymer matrix systems, enzyme-activated systems, or systems that respond to changes in physical conditions within the formulation. Mechanisms include dissolution, diffusion, osmotic pressure, maintaining a hydrologic or hydrodynamic balance, and ion exchange. Newer approaches are currently being developed based on nanotechnology and novel matrix technologies and include gastro-retentive drug delivery for controlled release, buoyant systems, and mucoadhesive systems for regulated release of antiviral medications (1,2).

	Excipients are key ingredients in controlled-release formulations. They act as matrix formers in matrix delivery systems and as polymeric membranes for drug powders and multi-particulate systems in oral solid dosage (OSD) forms, according to Bujar Muça, technical product manager, IMCD. Plasticizers and pore formers impact drug release in coated products, and, often, two-polymer chemistries are used to provide tailored drug release. In oral liquid products, excipients can also provide taste masking and protection for the gastric mucosa, he adds. In specific cases, excipients can provide abuse deterrence and aid in the avoidance of alcohol-induced dose dumping, according to Evonne Brennan, technical marketing manager at IMCD.

	“Ultimately,” says Shawn Branning, strategic marketing leader for controlled release at DuPont Nutrition & Health, “excipients that provide controlled-release characteristics to oral solid dosage forms allow not only control of the rate of the API release, but also targeting of where that release occurs.” He adds that when considering OSD forms, controlled-release formulations can refer to various types of sustained-release or targeted-(delayed) release profiles.

	Several factors must be considered when selecting specific excipients for controlled-release formulations, according to Chhanda Kapadia, technical manager, IMCD. Dosage form type and design need to be considered along with the solubility of the API to ensure that the desired dissolution profile can be achieved with a given excipient, and/or that a good 

 correlation is obtainable. Issues that relate to patient compliance and safety should also be considered, according to True Rogers, technologies leader at Dow Pharma Solutions.

	Ethylcellulose remains the polymer of choice in multi-particulate drug delivery systems with or without an additional coating platform, the use of which is dependent on the characteristics and dosage form type of the drug, according to Brennan. Methacrylate polymers are broadly used in delayed-release dosage forms and also provide gastro-resistant functionality. In monolithic delivery systems, hydroxypropyl methylcellulose (HPMC) continues to be the excipient of choice, although the use of two-polymer matrix systems is growing. “This approach can provide unique synergies and enable fine-tuning of the drug release pattern via modification of polymer entanglement and matrix texture,” Muça explains.

	The key for pharmaceutical formula tors, concludes Rogers, is to have proven and flexible polymers such as cellulosics along with knowledgeable partners that can help provide solutions to formulation challenges.

	Many APIs with unique formulating challenges

	Formulators can, in fact, face many challenges-from achieving desirable dissolution and release profiles for poorly soluble drugs to risk management of dose dumping to the selection of biocompatible materials with the desired controlled-release properties (1). Controlled-release products also require high API dosages, which can in some cases be difficult to formulate into the desired dosage forms with acceptable properties (2).

	Newer APIs present some unique formulating difficulties, according to Rogers, such as small therapeutic windows or sensitivity to certain pH environments. On the other hand, formulating controlled-release alternatives to already-marketed products can be challenging if it is necessary to select different polymer chemistries and excipients to avoid patent infringement issues, according to Lies d’Olieslager, technical product manager, IMCD. Consistently matching the drug release profile, the scale-up of controlled-release coating processes (which requires extensive process knowledge), the curing of aqueous coating systems-particularly under dynamic conditions-and resistance to ethanol-induced dose dumping are additional challenges, she adds.

	“Controlled-release excipients are excellent solutions to these particular challenges,” Rogers asserts. For instance, he notes that some excipients can now offer dual functionality, both controlling the release of the API and providing increased solubility to poorly soluble APIs. Understanding polymer chemistry and exploring synergies are also important to overcoming these difficulties, according to Brennan. “Formulator experience is crucial; with movement of personnel, the loss of expertise can impact development times in an already costly process,” she says.

	Given that excipients are key ingredients in controlled-release formulations, it may seem surprising that excipient choices are somewhat limited, particularly for certain patient populations, notably children. Suppliers are developing “novel” excipients that may provide solutions and support to move existing formulations into new areas, according to Mary Tanenbaum, technical manager, IMCD, but excipients do not have their own approval pathway. “Pharmaceutical companies are hesitant to take on the responsibility of toxicology and clinical trials with an untested material and run the risk of jeopardizing approval of their new drug applications,” Tanenbaum observes.

	Most new excipients, therefore, are modifications or combinations of previously approved excipients. For instance, Muça points to granular Carbopol developed by Lubrizol to address flowability issues in direct-compression processing of matrix delivery systems. The Dow Chemical Company has also developed and commercialized Methocel DC2, a more flowable morphology of HPMC for streamlined direct-compression manufacture of modified-release matrix tablets, according to Rogers. In addition, custom grades of Benecel HPMC from Ashland fill a gap in HPMC polymer choice for matrix systems, providing formulators with more predictable release profiles and reducing batch-to-batch variability previously observed due to HPMC blending, according to Brennan.

	Meanwhile, FMC Health & Nutrition, now part of DuPont Nutrition & Health, has promoted the use of surfactants for in-situ curing while coating with Aquacoat ECD, a 30% by weight aqueous dispersion of ethylcellulose polymer. “This pseudolatex dispersion has advantages over solvent systems because it has a high-solids content with low viscosity, and thus a shorter processing time,” notes Engin Sari, technical product manager, IMCD. DuPont has also developed a coating with a robust sustained-release profile in the presence of alcohol that helps formulators create products with resistance to alcohol-induced dose dumping, he adds. Aquacoat ARC is a blend of Aquacoat ECD and guar gum in an aqueous system for multi-particulate drug delivery systems.

	Collaboration is essential to finding solutions

	IMCD represents a wide portfolio of products and continually pursues a training program covering excipient functionality to ensure ongoing knowledge transfer, according to Brennan. “Engaging with formulators to review any challenges they face is key to successful development of solutions. Continued education is also fundamental; excipient manufacturers regularly provide information on product functionality, product consistency, and product variability, and we need to stay up-to-date,” she comments.

	As an industry group, the Controlled Release Society provides a forum for networking and the development of connections within other formulation groups. It does not provide an avenue for the discussion of intimate formulation details, however, which remains challenging, according to Brennan. “Using confidentiality agreements is one approach, but these take time. Good relationships with development teams, therefore, continue to be of primary importance,” she says.

	Excipient manufacturers need to work side-by-side with drug product manufacturers as early as possible in the drug product development process in order to not only identify the issues, but to mitigate the impact of the issues proactively, agrees Rogers. He adds that excipient manufacturers must understand every detail of their processes and polymers in order to achieve the best solutions for their customers.

	“Dow Pharma Solutions works closely with our customers to understand their specific needs, and, over time, we have developed new solutions to meet controlled release formulation challenges as they have arisen,” Branning remarks. “Partnering with our customers not only helps us in suggesting the right solutions for their formulations, but also helps drive our development of new excipients to meet the ever-changing needs of the market. We have found this type of partnership to be the best strategy for addressing the challenges faced by the pharmaceutical market today,” he concludes.

	The Dow Chemical Company has also been busy combining different businesses to expand its excipient offerings to the broader market. The excipient portfolios of The Dow Chemical Company, DuPont, and FMC’s Health & Nutrition division will be combined when the three businesses are merged to form DuPont Nutrition and Health. “The new company will have the expert teams from all three businesses working together and with customers and channel partners to understand their needs and unresolved challenges as well as provide unique combination solutions,” Branning says.

	1. Future Market Insights, “Controlled-Release Drug Delivery Technology Market: Global Industry Analysis and Opportunity Assessment 2017–2027,” Report Description (to be published July 2018).
	2. Future Market Insights, “Oral Controlled Release Drug Delivery Technology Market: Diffusion Controlled Release System Anticipated to be the Most Attractive Segment During the Forecast Period: Global Industry Analysis 2012–2016 and Opportunity Assessment 2017–2027,” Press Release, July 20, 2017.


	Vol. 42, No. 7
	July 2018
	Page: 24–26

	When referring to this article, please cite it as C.Challener, “Selecting Excipients for Controlled Release" 

With the right excipients, formulators can control when, where, and how an API is released.

Selecting Excipients for Controlled Release

Flow chemistry for the production of small-molecule APIs is recognized to offer significant benefits, particularly for processes that involve hazardous reagents or that are highly exothermic. The ability to provide greater process control combined with the use of minimal quantities of reagents at any one time increases the safety of such processes, in many cases enabling the production at commercial scale of valuable compounds that cannot be manufactured practically in batch mode. Continuous processing is new to API manufacturing, however, and additional tools and technologies must still be developed to make this promising approach a practical, widespread reality.

	A few solutions that have been recently introduced or are under development to facilitate the adoption of flow chemistry for API manufacturing are described as follows.

	Continuous processes are required not only for the reactions used in the synthesis of small-molecule APIs, but also for the unit operations connecting reactions. The development of continuous work-up solutions is one area that has been lagging.

	Zaiput Flow Technologies was founded at the end of 2012 to commercialize liquid-liquid separation/extraction technology suitable for continuous processing that was developed by Andrea Adamo while working at the Massachusetts Institute of Technology. The company’s core capability, according to Adamo, is engineering innovation in the context of fluid-related operations.

	Zaiput’s membrane-based separation technology exploits the concepts of hydrophobicity and hydrophilicity and unlike traditional devices does not require gravity to achieve separation. Differences in the wetting properties of two liquids (or a gas and a liquid) on the porous membrane allow for separation of immiscible solutions, even for solutions with compounds that have the same density. One material wets the membrane, filling all of the pores (the wetting phase). Application of a controlled-pressure differential “pushes” the wetting phase through the pores while leaving the not-wetting phase behind.

	“This technology is the only solution available for continuous manufacturing because these devices can be operated in a pressurized line to enable multistep synthesis without interruptions,” Adamo, who is the CEO of the company, asserts. He adds that the systems have excellent scalability and thus provide a single solution from the bench to the plant, reducing the time and money needed for optimization and scale up.

	Zaiput has sold its technology to pharmaceutical companies around the world, and even sent devices into space on the International Space Station. The technology was included in the CRS-13 Mission in December 2017 to confirm its ability to perform in zero-gravity conditions. The devices were found to successfully separate immiscible liquids in space while under remote operation and observation.

	The company has an ongoing collaboration with Snapdragon Chemistry, a contract-development company focused on facilitating the adoption of flow chemistry and is starting one with the “Medicines 4 All” program operated by Virginia Commonwealth University (VCU) with support from the Gates Foundation. “Zaiput has innovation at its heart, and we are always looking for new additions and new ideas,” Adamo asserts.

	For continuous manufacturing, where many of pieces of equipment have to be operated, monitored, and adjusted simultaneously, automation is essential. Automation has long been available on plant-scale but is now coming into the pilot plant and laboratory, according to Matthew M. Bio, president and CEO of Snapdragon Chemistry. “With laboratory automation enabled, flow chemical process development is accelerated through automated experimentation. At the pilot scale, automation is being applied to ensuring safety and consistent product quality,” he explains.

	In-line process analytical technologies (PATs) such as infrared (IR) spectroscopy, Raman spectroscopy, and bench-top nuclear magnetic resonance are also being developed specifically for flow chemistry applications, according to Bio. “In the lab, in-line PATs are providing rich data sets that can be mined for process understanding, while in the pilot plant, they provide real-time process monitoring and enable feedback control of continuous processes.”

	Automated flow systems, coupled with embedded in-line PAT, are being used at Snapdragon to automate the discovery and development of continuous manufacturing processes. “We are able to rapidly define process design spaces through automated process characterization,” says Bio. The company is also developing autonomous process development tools where optimization algorithms are used to discover optimal reaction conditions. The data derived from these flow systems can be used for machine learning applications.

	Flow chemistry is often performed in microreactors, small reactors with small volumes and high throughputs. Anton Paar, in conjunction with pharmaceutical companies such as Janssen, Patheon, Astra Zeneca, and Lonza, has used additive manufacturing to produce a microreactor as part of the international research project “CC Flow.” Initial research and simulations were performed by Professor Oliver Kappe and his team at the University of Graz.

	The reactor was designed for the difluoromethylation of a lithiated nitrile with fluoroform (CHF3), a known greenhouse gas. “Three prototypes were developed before the final design and the perfect internal and external dimensions were determined,” explains Stefan Pfanner, specialist for additive manufacturing and direct metal laser sintering at Anton Paar.

	The reactor was produced via direct metal laser sintering using 316L stainless-steel powder with dimensions of 164 mm by 93 mm by 3 cm (thickness). Stainless-steel was chosen based on the chemical, mechanical, and thermal stability and thermal conductivity required for the reaction, according to Pfanner. A serpentine cooling core is surrounded by reaction channels that have an inner diameter of 0.8 mm and a total length of four meters. There are four inlets, two defined reaction zones, and one outlet.

	Anton Paar plans to develop various tailor-made reactor types for different industrial requirements, according to Gunter Kole, business area manager for analytical and synthetic chemistry at the company. “One of the advantages of 3-D printing is that any type of design can be produced, which has not been possible in the past with conventional manufacturing technologies,” he notes.

	In November 2017, Corning and the University of Liège in Belgium announced the establishment of the first lab in Europe qualified for use of the Corning Advanced Flow reactor (AFR) (1).

	The lab is located in the university’s Center for Integrated Technology and Organic Synthesis (CiTOS), which provides customers with continuous-flow demonstrations, experimental trials, feasibility testing, and chemical reaction process development services. Other labs have previously been qualified in the United States and China.

	Earlier in 2017, Angelini Pharma installed Corning’s AFR technology at its Aprilia facility for use in API manufacturing (2). The company started with a pilot line (G1 glass and G1 silicon carbide) in 2015 and decided to switch to the G4 production reactor shortly afterwards. The production reactor meets pharmaceutical quality and FDA requirements. The company uses flow chemistry to perform reactions it could not do previously on a commercial scale in batch mode (2).

	The AFR technology is specially designed to enable the conversion of batch chemical processes to continuous processes. The reactors consist of engineered fluidic modules that integrate heat-transfer and mass-transfer into a single piece of equipment. As a result, they offer at least 100 times enhancement in mixing, 1000 times improvement in heat transfer performance, and efficient scale-up from the lab to full-scale production, according to the company (1). Processes, therefore, can benefit from increased efficiency, scalability, yields, and quality with reduced performance variability and cost.

	Another change in technology that is impacting continuous manufacturing today, according to Bio, is the approach to scale up and the design of production facilities for continuous manufacturing. He points to Eli Lilly’s Kinsale plant in Ireland as an example of a new approach to pharmaceutical drug-substance manufacturing in which the use of small-footprint modular components in fume hoods results in a flexible facility where process trains can be easily reconfigured for multi-product applications. “The reduced size and cost of continuous manufacturing equipment together with the ability to adjust output by numbering out and numbering up will enable responsive supply chains for pharmaceutical API,” Bio states.

	Snapdragon has developed a scalable-by-design approach to flow chemical process development. “We have invested in the development and characterization of laboratory reactors that mirror the performance of production-scale reactors. Our custom software platform FlowChem WebApp continuously logs data from these reactors and interacts with a suite of [Internet-of-Things] IoT-enabled flow components for automated experimentation and even autonomous, algorithm-driven, reaction optimization. These tools allow for rapid translation of laboratory results to production scale,” says Bio.

	Snapdragon has also invested in the development and scaling of photochemical processes, recently completing the design and construction of a production-scale 5 kW flow photoreactor that is built to work with a range of different wavelengths and is able to operate under pressure for gas/liquid photoreactions.

	Pumping technologies for flow chemical applications, particularly at lab and pilot scale, are still inadequate, according to Bio. “The syringe pumps typically used in laboratory flow demonstrations don’t accurately mirror the performance of plant-scale pumping technologies, but to develop plant-ready processes, pump performance characteristics must be taken into account,” he explains. Some manufacturers are beginning to look at this issue, though. The Tacmina Smoothflow technology is particularly useful for developing flow processes with organometallic reagents, according to Bio. Snapdragon has been working with Tacmina to evaluate this pump in flow chemistry applications. The company is also evaluating microannular gear pumps as another scalable pump technology for flow chemistry.

	1. Corning, “Corning and University of Liège Announce First Advanced-Flow Reactor Qualified Lab in Europe,” Press Release, Nov. 3, 2017.
	2. Corning, “First Industrial Use of AFR Technology in Italy,” Press Release, March 2017.


	Vol. 42, No. 6
	June 2018
	Pages: 22–24

	When referring to this article, please cite it as C. Challener, “New Tools Outline Advances for Continuous API Processing" 

From separation systems to reactor technology, new tools are increasing the feasibility of continuous API production.

New Tools Outline Advances for Continuous API Processing

The purity (aggregates, host-cell proteins, host-cell DNA, etc.) and structure (sequence of amino acids, post-translational modifications, folding, etc.) of proteins directly impact the safety and efficacy of any formulated drug product based on them. Because proteins are generally produced as a mixture of similar but different biomolecules using living cells and subjected to extensive purification steps, variations in their structure and purity can occur from batch to batch.

	The International Council for Harmonization (ICH) guidance Q6B, 

Specifications: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products

 (1) established guidelines for setting specifications for biologic drug substances, including determination of their composition, physical properties, and primary and higher-order structures, using appropriate methodologies. Comprehensive characterization of proteins is often challenging, however, given their complexity. In many cases, singular analytical methods cannot provide the complete picture for specific attributes. Orthogonal methods-different methods intended to measure similar attributes-are often necessary to provide independent confirmation of protein properties. Protein particle characterization is no exception.

	The key when selecting orthogonal analytical methods is to choose technologies that overlap in complementary ways to provide a complete characterization of the protein in question, according to Kent Peterson, president and CEO of Fluid Imaging Technologies. The first step is to identify the information that is required for protein characterization and which techniques can provide that information-and the information that each cannot provide. “It is important to keep in mind that different techniques may provide the same information, but produce different results due to their varying technologies,” Peterson says.

	Proteins, as well as all types of monoclonal antibodies and gene and cellular therapeutics, differ in terms of their structures, biological activities, and purity profiles. Consequently, while different sets of orthogonal technique strategies may be used for different biologics, true “toolbox” approaches or platforms may not be completely successful, according to Mike Merges, director of strategic growth for Catalent Biologics.

	As a result, Catalent, when developing strategies for assay or method development and optimization, considers a combination of one factor at a time (OFAT) to define individual factors, at least initially, and design of experiment (DoE) to look at multiple and interacting factors. “The use of fractional factorial DoE as soon as is practical allows for a more rapid and robust development of a set of orthogonal methods,” Merges adds.

	When it comes to particle analysis, Peterson notes that different techniques rely on different detection principles ranging from fairly basic (visual/microscopic inspection) to high-tech methods (resonant mass measurement [RMM], analytical ultracentrifugation [AUC]) and cover a wide range of particle sizes. “No single technique provides all of the answers. Two different techniques provide more data than just one,” he observes.

	Effective sets of orthogonal methods “follow-the-molecule,” according to Merges. Timeliness, cost-effectiveness, and accuracy are other important factors, according to Peterson.

	In addition, Merges notes that the ability to validate the techniques and transfer them from the development phase/lab to the quality control phase/lab or even to an outside contract research organization are essential.

	“Validation is accomplished in a phase-appropriate manner,” Merges continues. The overall guidance for all phase-appropriate levels is ICH Q2 (R1) (2),

	although different technical platforms (e.g., enzyme-linked immunosorbent assay [ELISA] or bioassay potency tests) may have specific levels of adherence to the ICH guidelines and/or other pieces of guidance, such as 

	“The establishment of appropriate validation acceptance criteria should be based upon data-driven decisions. Those data are best generated via a pre-validation exercise conducted prior to the drafting of each phase-appropriate validation protocol,” he explains.

	Orthogonal techniques for particle characterization

	When it comes to particle characterization, Petersen points out that collection of imaging data is one of the accepted techniques for acquiring size, shape, and other morphological information used to differentiate protein aggregates from other particles in the formulation. “The nature of the particles in protein formulations and the small sample volumes pose a challenge in that throughput, accuracy, and specificity are met by very few technologies, however,” he comments.

	Different techniques have different strengths and weaknesses for addressing these challenges. “For example,” Petersen observes, “particle count can be obtained by various particle analyzers, including single-function particle counters, light obscuration (LO) instruments, and flow imaging microscopes (FIMs). However, only some FIMs offer light and dark pixel thresholding that enables the detection of translucent proteins, while these proteins go undetected by other technologies.” 

	More specifically, LO provides particle count and concentration for large sample volumes. It provides the ability to analyze large volumes relatively quickly, has been around for a long time, and is seen as the first step in particle analysis, according to Peterson.

	It does not capture morphological information about the particles contained in the sample, however, nor does it accurately detect or record translucent to semi-translucent particles, according to Peterson. “In LO, the signal from each particle is converted to a sphere of correlative size, and what is actually measured is the equivalent spherical diameter (ESD) of the generated sphere. We know, however, that not all protein particles are spherical in shape, so this method does not produce the most accurate size distribution,” he explains. LO is an inexpensive and quick method that has been in use for decades despite these limitations due mostly to the fact that it remains a USP compendial method.

	FIM, on the other hand, enables the detection of particle sizes ranging from 300 nm-5 mm and reveals morphological characteristics of the particles imaged. In addition, FIM can detect translucent, semi-translucent, and opaque particles and allows for the differentiation of protein aggregates and non-proteinaceous particles.

	Currently, light obscuration and membrane microscopy are required by USP regulations, and flow imaging microscopy is being rapidly embraced globally, according to Peterson. The combination of technological advancement and the industry’s acceptance of the need for a better method to fill the gap where methods like LO left a lot of unanswered questions-specifically for smaller micro- and nanoparticle analysis-have paved the way for newer methods to become widely accepted, he explains. “Better data from imaging flow microscopy provides morphological characteristics of the particles analyzed and answers these questions while providing a more detailed analysis,” he says.

	Nanoparticle analysis lacks many orthogonal components. Historically, imaging in flow could not address this size range, according to Peterson, so information on size and shape and transparency through the imaging of nanoparticles had been missing. Newer technologies such as nano-flow imaging from his company are addressing this need. Automated membrane microscopy from HALO Labs and hyperspectral imaging from Spheryx are also offering promising solutions. “Until recently, there has been no imaging capability for flow imaging microscopy of particles in the nanometer size range. The FlowCam Nano imaging instrument from Fluid Imaging Technologies is an oil immersion flow microscopy technique that provides optical resolution down to the hundreds of nanometers,” Peterson says.

	The breakthrough technologies described herein are examples of solutions being developed by instrument manufacturers in response to the needs of formulation development scientists for orthogonal analytical techniques for protein characterization. Bringing them to the market is increasingly difficult, however, according to Peterson. “There seems to be a growing lag time between commercial implementation and regulatory acceptance and requirements. These delays may be jeopardizing the timely, widespread use of technologies that can significantly contribute to the safely and efficacy of biotherapeutic drugs,” he asserts.

Test Procedures and Acceptance Criteria for Biotechnological/Biological Products

, Step 4 version (ICH, 1999).
	2. ICH, Q2(R1) 

Validation of Analytical Procedures: Text and Methodology

, Step 4 version (ICH, 1994) (Complementary Guideline on Methodology dated 6 November 1996 incorporated in November 2005).
	3. 

 General Chapter <1033>, “Biological Assay Validation” (US Pharmacopeial Convention, Rockville, MD, 2010).
	4. 

 General Chapter <1034>, “Analysis of Biological Assays” (US Pharmacopeial Convention, Rockville, MD, 2012).

Vol. 42, No. 4
	April 2018
	Pages: 56–58

	When referring to this article, please cite it as C. Challener, "Multiple Views Deliver Complete Characterization of Protein Particles" 

Complex protein formulations require multiple analytical techniques to achieve full characterization.

Complete Protein Particle Characterization from Multiple Views

Biologic drug substances are large, complex molecules prepared from living systems via complex cell-culture or fermentation processes. Given this complexity, it can be difficult to achieve consistent and predictable performance, with numerous factors from raw material variability to operator training to equipment selection influencing process outcomes. Biopharmaceutical manufacturers continuously seek new manufacturing technologies and approaches to improve process robustness and enhance product consistency and quality. Recent pursuits include continuous processing, the implementation of process analytical technology (PAT) and automation systems, the adoption of single-use equipment, development of enhanced modeling capabilities, and leveraging advances in data management and analysis. In the end, however, maximizing the benefits of these new approaches and technologies requires extensive collaboration within multifunctional teams.

	Uniform production of biologic drug substances from batch to batch requires consideration of three key attributes, according to Elise Mous, director of sales and marketing and business development for Capua BioServices: the people involved, including their training, know-how, experience, and flexibility; the process itself, because each has its own unique challenges, characteristics, and requirements; and the desired product, for which the final specifications determine the end-game.

	 “The first step is to find the right approach needed to orchestrate all of the necessary requirements in the right manner to ensure a successful, scalable, and robust recipe,” she says. Those requirements include expression strain requirements/characteristics; raw material requirements; plant/equipment design needs; automation needs; training requirements; utilities requirements; key process parameters; regulatory constraints/requirements; and environmental, health, and safety constraints/requirements. For contract development and manufacturing organizations (CDMOs), much of this information can be learned either from past experience at the plant or from the client.

	Attention to details can also make a significant difference in the robustness of a process, according to Mous. “It is important to analyze the potential risks and uncertainties upfront and define a custom onboarding trajectory for transfer, development, scale-up, and manufacturing campaigns, which includes definition and agreement of the joint objectives and journey required to achieve them,” she observes. Development of scalable, robust technologies as early as possible and establishment of a sampling plan and analytics are also important. Finally, attention should be paid to continuous improvement so that learning from the process and the data is ongoing. “Of course,” she adds, “safety and quality requirements can never be compromised.”

	Because CDMOs work on both long-term commercial and short-term clinical and commercial projects, the amount of available process data can vary significantly. Fujifilm Diosynth Biotechnologies tailors its approach to uniformity- and consistency-control based on the amount of data that is gathered, according to Abel Hastings, director of process sciences at the company’s North Carolina facility.

	 “For processes with long histories and substantial data sets, we can leverage traditional statistical control. For processes with smaller datasets, we aim to improve early responsiveness to process control and consistency monitoring by applying a logic-based, piece-wise approach,” he explains. This methodology requires three main steps: process parameter to attribute linkage, development of an ideal distribution of expected data, and comparison of data against expectations.

	Mapping out the expected process parameters to attribute linkage can form the basis for determining which data are most valuable to not only confirm that the process is running reliably, but also ultimately lead to tighter process control, even in circumstances where data are limited, according to Hastings. “This approach can also help teams focus on ways to capture the most informative information,” he adds.

	Once parameter-attribute linkages are established, Fujifilm’s teams discuss and document the expected data results distribution. For example, they consider whether a normally distributed histogram or a sine-wave distribution should be expected, and whether the data might be skewed. “Developing expectations based on the factors and control elements can help teams establish a basic understanding of what ‘normal’ should look like,” Hastings explains. Once actual data are generated, they are compared to the expectations. “Even in the absence of statistical significance, this approach can allow processes to move toward increased process control and uniformity early in their lifecycles,” he observes.

	Effective application of that actual data is essential to enabling meaningful interpretation, Hastings adds. “Most major equipment manufacturers have developed strong and deep data systems that gather information to help ensure reliable operation. In some instances, however, more data does not lead to more knowledge,” he asserts.

	Because CDMOs work on both long-term commercial and short-term clinical and commercial projects, the amount of available process data can vary significantly.

	To ensure the right interpretation can be made, Fujifilm recommends assessing each data acquisition against the corresponding parameter-attribute pair. “Teams should aim to identify any points of ambiguity. Common points of concern can result from misalignment between parameters and attributes, limits of resolution in data, and data acquisition frequency,” Hastings explains.

	An example of a parameter/attribute misalignment would be the difference between a temperature requirement for a shake flask operation and data recording of the temperature of an incubator. “While the temperatures may be the same, they are not the same element,” he says. Issues with data resolution can arise with many digital read-outs, which may clip or round data, potentially resulting in over-interpretation or under-appreciation of the degree to which the data may be skewed.

	Finally, Hastings notes that it is important for users to recognize that automated data systems recorded at intervals may not give an accurate reflection of processes that change at higher frequency. “Teams should consider the rate of change of a process step before selecting a sampling frequency,” he says. To avoid these problems and similar risks, Fujifilm strongly recommends a cross-functional review of processes--from scientist to automation engineer--in order to ensure all data risks are considered.

	Automation helps to ensure consistency by minimizing human errors and enabling repeatability and reliability of operations from one batch to another.

	In addition to types and amount of data and the way they are collected, the selection of biomanufacturing equipment can have a substantial impact on process robustness. Which aspects are most important can vary from one process to another, according to Gerald Hofmann, head of technology development at Capua BioServices. In addition, design attributes to be considered should not be limited to the performance during the actual cell-culture or fermentation process. “The hygienic design of the equipment, its maintenance needs, and what is required for an effective cleaning regime must also be well understood,” he says.

	Perhaps most importantly, selecting the equipment design that will provide the most robust process cannot be achieved without proper understanding of the full process requirements, according to Hofmann. “Making sure that these process requirements, and the needs of the user, are also well understood by equipment manufacturers can also have a measurable impact,” he comments. New equipment should therefore be selected to fit for a defined purpose in accordance with user requirements and critical process parameters.

	When working with existing equipment, it is important to understand the opportunities and limits presented by the design and the potential for any modifications that can make the equipment more fit-for-purpose. Appropriate installation qualification, operational qualification, and process qualification protocols should also be prepared, reviewed, approved, and followed; and a proper maintenance plan should be developed and implemented, according to Hofmann.

	Automation helps to ensure consistency by minimizing human errors and enabling repeatability and reliability of operations from one batch to another. Increasing adoption of automation and PAT in biopharmaceutical manufacturing is intended to reduce drug variability, increase yield, drive down costs, and maximize safety.

	As CDMOs, both Fujifilm and Capua BioServices have a unique perspective because they work with a spectrum of processes, including products undergoing their first at-scale batches, decade-old processes being relocated to another facility, and commercial processes with a long history of continuous improvement. “The strategy necessary to ultimately implement automation ideals can often be challenging for processes striving to simply produce clinical batches,” says Hastings. “It is beneficial to find the right balance between manual versus automated operations; continuously evaluating investment versus added value,” adds Mous.

	“When installing new processes into our facility, we aim to consider the full lifecycle of the process while also meeting the short-term goals of the project. To do so and overcome some of the challenges with implementation of automation, we design our process development work to simulate what can be both implemented as a manual operation and be transitioned into an automated step,” says Hastings.

									In response to the growing complexity and globalization of biopharmaceutical manufacturing, in 2017 FDA’s Office of Pharmaceutical Quality (OPQ) established the Manufacturing Science and Innovation Center of Excellence (CoE) “to promote internal and external scientific collaboration in manufacturing science, facilitate research communication and management, and advance OPQ’s research culture and capabilities in manufacturing science” (1). The CoE is intended to build on FDA’s history of research on bioprocessing issues, particularly viral clearance, quality by design, bioreactor control, and unit operation linkage.

									The CoE’s mission is to “promote biopharmaceutical science and innovation by addressing cross-cutting science and regulatory issues in bioprocessing through research that supports guidance and policy development,” which it will accomplish by identifying technology gaps in bioprocessing methods and directing research to address them, with a focus on improving robustness and eliminating potential failure modes association with common unit operations. Some currently perceived gaps include real-time monitoring of protein critical quality attributes, adventitious agent control, viral clearance robustness, and the linkage between upstream and downstream processes.

									FDA also believes that the CoE can help address industry competitiveness. In addition, because advances in manufacturing technologies can simultaneously improve efficiency/productivity and create regulatory challenges, the CoE will enable FDA scientists/reviewers to better understand new technologies as they are being developed. As a result, the agency will be better positioned to evaluate their potential impacts on quality and safety.

									1. K. Brorson and S. L. Lee, “

OPQ Establishes Manufacturing Science CoE

	Underlying all activities related to the development of robust processes for the production of biologic drug substances is the need for collaboration among scientists and engineers involved in all aspects of a project. “Selection, training, and coaching of the project team, from the operators to the process engineers to the plant engineers to the analytical scientists, is essential to achieving effective processes. Success is not possible without such a multidisciplinary approach,” asserts Mous.

	In many cases, finding the right solutions requires a bit of creative thought and collaborative work between the process development teams and manufacturing and/or quality-control operators, agrees Hastings. “We strongly recommend that cross-functional teams spend time in one-another’s area in order to build a solid understanding of flexibility needs, constraints, and long-term opportunities,” he comments.

	Through this type of collaboration, it is possible for teams to find the right balance between speed, risk and reward, and cost, according to Mous. “In some cases, a project can be started quickly with a higher risk, while in others, investment in proper preparation upfront is the more appropriate approach,” she explains. Cross-functional teams are also better able to determine whether making facility or process modifications is more practical. “These teams know how to look for the best fit in terms of not only the best fit with the process, but the best fit with the budget,” Mous concludes.

	When referring to this article, please cite it as C. Challener, “Consistent API Quality Calls for Collaboration" 

Bioprocess understanding, the right equipment, and automation help, but multifunctional teamwork is the key to API production success.

Consistent API Quality Calls for Collaboration

Access to multiple analytical techniques is essential for fully characterizing complex protein formulations.