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

Cynthia A. Challener

Biopharmaceuticals, while highly effective medicines, are susceptible to aggregation upon manufacturing, distribution, and storage, which can result in immunogenicity and/or reduced efficacy. Stabilization of proteins and other biologic drugs is, therefore, crucial for ensuring both safety and efficacy.

	The molecular structures of therapeutic proteins and the critical sites within these proteins that are prone to oxidation, deamidation, hydrolysis, etc., and can result in aggregation and fragmentation, are well known. The development process, therefore, begins with efforts to stabilize therapeutic proteins using protein engineering. However, the improvement in stability is generally limited because of the unstable nature of proteins; the stabilization energy for the native state is typically between 5 and 20 kcal/mol, which is equivalent to that of a few hydrogen bonds. Because the folded state is only marginally more stable than the unfolded state, any changes in the protein environment may trigger protein degradation, aggregation, and or inactivation.

	The use of excipients is, therefore, common during the entire manufacturing process and in the final formulation to achieve protein stabilization through retardation of chemical degradation processes and prevention of aggregation, according to Kristina Kemter, project leader in the Research and Development department of Leukocare. Selection of appropriate stabilization agents, however, is challenging, and made more so by the limited choice of substances that have been approved for this use.


	Protein stability is a result of achieving a balance between destabilizing and stabilizing forces. The destabilizing forces are mainly due to the large increase in entropy of unfolding; the stabilizing forces are provided by intra-protein and protein-solvent interactions. Disruption of these interactions will shift the balance and destabilize a protein, and many stress factors are known to disrupt this delicate balance and affect protein stability.

	These stress parameters include temperature, pH, ionic strength, metal ions, surface adsorption, shearing, shaking, additives, solvents, protein concentration, purity, morphism, pressure, and freeze/thaw-drying, according to Kemter. As a result, a variety of physical and chemical processes can lead to aggregation of proteins. She also notes that there are significant differences between freeze-dried and liquid protein products in terms of the stress exposures upon manufacturing (freeze drying) and during storage of the final products.

	Liquid formulations in general have a higher risk for physical and chemical instability during storage than dried formulations due to their high mobility and the increased likelihood of chemical reactions and physical instability. For liquid-protein preparations, stability can be enhanced by selection of pH-buffering salts, and often amino acids can also be used. It is often interactions at the liquid/air interface or liquid/solid interface (with the packaging) that lead to aggregation following adsorption and unfolding of the protein, according to Jens T. Bukrinski, a formulation scientist with Novozymes. Shear stresses during spray drying can also lead to exposure of hydrophobic protein regions upon spray drying that allow initiation of aggregation, according to Kemter.

	Stability, particularly thermal stability, can be increased greatly when the protein is dried. For example, lyophilized formulations are generally more shelf-life stable than liquid formulations; however, protein molecules can potentially form large amounts of aggregates during the lyophilization process, according to Jay Kang, director of analytical and formulation development at Patheon. In the freezing phase, the formation of numerous ice crystals creates a large water/solid surface, a highly concentrated solution, and often an altered pH environment. All these conditions can cause protein aggregation. During drying, removal of water puts significant stresses on protein molecules and can induce aggregate formation.


	Temperature and pH have the greatest influence on both physical and chemical protein stability. High temperatures can lead to thermal denaturation and subsequent aggregation as well as accelerate chemical degradation pathways (i.e., side chain oxidation, hydrolysis, and deamidation) that may result in aggregation. In addition, most proteins are generally stable only in a narrow pH range, and physical and chemical pathways to aggregation are possible outside of that window, according to Kemter. Furthermore, the effect of pH on the chemical stability of a protein can be altered in the presence of excipients.

	For instance, the effect of salts strongly depends on the pH of the solution, which dictates the charged state of ionizable groups in the protein. “Salts may stabilize, destabilize, or have no effect on proteins depending on the type and concentration of salt, the nature of ionic interactions, and the presence and amount of charged residues in proteins on fully exposed surfaces and/or in fully or partially buried interior sections,” Kemter says.

	Depending on the type and concentration, metal ions may also stabilize or destabilize a protein because their interactions with proteins are highly protein-dependent. Metal ions may also significantly affect protein stability without affecting much of its secondary structure. “Trace amounts of metal ions in protein formulations may catalyze the oxidation of proteins, namely via the Fenton pathway, targeting in particular the residues methionine, cysteine, histidine, tryptophan, tyrosine, proline, arginine, lysine, or threonine. The catalysis depends on the concentration of the metal ions, and the metal-catalyzed reaction can be facilitated in the presence of reducing agents such as ascorbate or sulfhydryl (RSH) compounds,” Kemter explains. Oxidation of proteins by metal ions, oxygen, or reducing agents can lead to aggregation.

	Similarly, chelating agents such as ethylenediaminetetraacetic acid (EDTA) and citric acid can bind and destabilize proteins or bind harmful metal ions and prevent oxidation. The net effect depends on the metal ions, oxidation mechanism, and the type of chelating agent, according to Kemter.

	Protein concentration can also influence aggregation and in some cases chemical degradation, depending on the mechanism. “More biopharmaceuticals are being formulated at very high concentrations today. Higher concentrations create more chances for unwanted self-association, and consequently high viscosities, which can make both processing and injection more difficult,” Bukrinski observes. High pressure can also cause protein unfolding during freeze-drying or spray-freeze drying.

	Chemical degradation pathways include not only oxidation, hydrolysis, and deamidation, but also isomerization, succinimidation, disulfide bond formation and breakage, non-disulfide crosslinking, and deglycosylation, and often occur simultaneously, according to Kemter. She notes that deamidation of asparagine and glutamine residues is most common, and the rate, mechanism, and location of deamidation are all pH-dependent. Oxidation is another important chemical degradation mechanism, particularly the oxidation of the thio groups in methionine and cysteine residues. The formation of disulfide bond linkages or thio-disulfide exchanges can result in protein aggregation or polymerization.

	“All of these influences can occur--either simultaneously or separately--when different types of stresses are applied to a protein, such as during isolation and purification of a protein; drying of a protein by lyophilization, spray-drying, spray-freeze drying or foam drying; storage of a protein in solution; after drying; as well as reconstitution after drying,” states Kemter.


	Stabilizing excipients are added to formulations to slow down or prevent protein aggregation through different mechanisms, including strengthening of protein-stabilizing forces, destabilization of the denatured state, and direct binding to the protein, which are applied during isolation and purification, drying (lyophilization, spray-drying, spray-freeze drying, foam-drying), storage in solution or after drying, and reconstitution after drying, according to Kemter.

	Six categories of excipients are commonly used to stabilize proteins against aggregation, according to Kang: buffers, salts, amino acids, polyols/disaccharides/polysaccharides, surfactants, and antioxidants. These excipients prevent aggregation through several mechanisms. “First, pH is critical to protein stability and must be controlled to an optimal value through the use of appropriate buffers. Salts and amino acids increase the ionic strength of solutions while minimizing electrostatic interactions between protein molecules.”

	Polyols/disaccharides/polysaccharides, on the other hand, stabilize protein molecules through preferential hydration or preferential exclusion, in which the excipient is excluded from the protein surface, allowing the rearrangement and stabilization of a stabilizing hydrate shell around the protein surface and preventing protein molecules from interacting. Surfactants, meanwhile, protect proteins against interaction with hydrophobic surfaces such as air/water interfaces and container surfaces, and antioxidants protect proteins against oxidation.

	Specific examples include trehalose and sucrose (disaccharides); mannitol and sorbitol (sugar alcohols); histidine, glycine, and arginine (amino acids); polysorbate 20, polysorbate 80, and proteins like human serum albumin (surfactants); sodium chloride (salt); and dextran and polyethylene glycol (polymers). Each type has a completely different stabilizing mechanism.

Alkylsaccharides as alternatives to polysorbates


	The quality of the excipients can have a dramatic impact on protein stability. For example, common surfactants such as polysorbates are well known to contain trace amounts of peroxides that can become significant during storage and consequently compromise protein stability through oxidation of susceptible amino acids, particularly exposed methionine, according to Kang. “A good practice is to use compendial excipients in the development study and control the quality/storage of the excipients from the very beginning, so no surprises are encountered in GMP manufacturing and subsequent long-term storage,” he says.

	The issues with polysorbates have been a focus of Aegis Therapeutics. “The oxidative damage to biotherapeutic proteins can directly reduce their efficacy, but perhaps more importantly introduce unwanted immunogenicity, which can give rise to neutralization of activity and in more serious cases, neutralization of any residual biological activity associated with the patient’s own corresponding proteins,” observes Edward T. Maggio, president and chief executive officer of the company.

	The polysorbates do an excellent job of preventing aggregation and permitting the creation of high-concentration formulations of biotherapeutics, but even though they are ideally highly purified when first incorporated, Maggio notes that formation of chemically reactive species occurs immediately upon any contact with oxygen during manufacturing or storage of the final product.

	Aegis has shown that alkylsaccharides are as effective, and in some cases, more effective than the polysorbates for the prevention of aggregation, and they are completely free of any oxidative damage problems. Comprising a sugar attached to a long chain fatty acid, alkylsaccharides break up rapidly in the body into a sugar, typically glucose, and a fatty acid. They are essentially nontoxic food components, and are considered generally recognized as safe (GRAS) for food purposes. Both Big Pharma and smaller biotech companies (e.g., Roche and Biodel) are investigating the use of alkylsaccharides as replacements for polysorbates, according to Maggio. He also notes that there is extensive safety and toxicity study data in the drug master file covering the alkylsaccharides to which licensees of the technology receive a right of reference for their regulatory filings.


	To assist formulators of biologic drugs with the stabilization and protection of therapeutic proteins in dry and liquid formulations, Leukocare developed its Stabilizing and Protecting Solutions (SPS) platform of excipient formulations. The proprietary SPS platform has been proven to stabilize different biological macromolecules during freeze drying and spray drying, during storage at increased temperature in dry or liquid formulations, and even during sterilization processes, according to Kempter. She adds that the modular concept of the SPS facilitates the possibility to adjust excipient compositions to the specific requirements of the biomolecule to protect from various types of stress exposures.

	SPS formulations solely comprise excipients (mainly based on amino acids) listed in pharmacopeia that are also in general listed by FDA as inactive ingredients. Interactions between multiple functional groups of the excipients in solution and particularly during drying result in a strong amorphous character of the dried product, according to Kemter. The osmolytic excipients of the compositions may also support stabilization of the hydration shell around proteins in liquid formulations according to the preferential exclusion theory. In dried formulations, and according to the theory of water replacement, the excipients may substitute the stabilizing hydrogen bonds between proteins and water through the formation of stabilizing electrostatic interactions or hydrogen bonds with the proteins upon drying.


	rHSA is the most common protein found in human plasma and is known for ligand binding and colloidal stabilization, properties that are also ideal for the stabilization of proteins and other biologics. In addition, because rHSA is a natural protein in the body, it is safer than other alternative surfactant excipients, such as synthetic polymers, because it has a low risk for immunogenicity and there is a natural pathway for its elimination from the body.

	In addition to preventing protein aggregation by blocking undesired adsorption onto vial surfaces or air/liquid interfaces, rHSA has also been shown to prevent self-association of protein drugs through preferential hydration or preferential exclusion effects and to prevent the formation of micrometer-sized particles to increase the solubility of poorly soluble peptide-based drugs. Furthermore, rHSA exhibits antioxidant properties used for the stabilization of radiolabeled drugs. “As a result of this multifunctionality, in some cases, it is possible to reduce the number of excipients required to stabilize a biologic drug,” Bukrinski says. He also notes that through covalent linking of a biologic API to rHSA or through encapsulation in albumin particles, it is possible to extend the circulatory half-life and control the release of drugs. Novozymes offers recombinant HSA products specifically produced for the biopharmaceutical industry.

Choosing the right stabilizing excipients


	Because every protein is different and reacts differently with different types of excipients, the selection of the proper excipients for a given formulation is a complex and challenging task. Some excipients may be effective in one formulation and cause protein degradation and aggregation in another. “For instance, commonly used polysorbate detergents often disrupt the protein complexes in virus-like particles, and therefore an alternative such as recombinant human serum album or a carbohydrate-based polymer is needed,” Bukrinski comments.

	The choice of the right excipients for stabilization of a desired protein must take into account the very complex physical and chemical degradation behaviors and the potential multifunctional influences that each excipient can have on protein stability, according to Kemter. “The formulation scientist must have a thorough knowledge of several factors: how to optimize the physical and chemical stability of the active ingredient; how, rationally, to include specific excipients in the formulation; how to obtain the optimum conditions for stability; how to prevent stability issues during up-scaling; and, finally, how to design a formulation that is suitable for the intended route of administration, that is, one that allows the absorption barrier to be overcome.”

	Unfortunately, the structural differences among different proteins are so significant that generalization of universal stabilization strategies has not been successful to date. In addition, while it is desirable to keep the number of functional excipients as low as possible, the vast amount of potential destabilizing effects often necessitates the addition of several excipients.

	Many scientists select excipients empirically, which takes a long time and puts a strain on resources, according to Kang. “It is more effective to perform a preformulation study to identify degradation pathways and stability-indicating analytical methods before testing major categories of excipients for their ability to protect protein molecules against specific degradations. Further study using a design-of-experiment (DoE) approach enables determination of the optimal levels of the most effective excipients and their interactions for the final formulation,” he asserts. In fact, Patheon has been able to shorten the development timeline by designing formulations based on scientific rationale and the screening of large quantities of excipients using DoE.

	Even so, Kemter believes that there remains a significant need for improved methods for the folding of proteins and the prevention of unfolding during production and storage and that can be generally applied to the majority of biopharmaceuticals. “Such technology would eliminate the need for so many individual investigations,” she notes.


	Typical studies for determination of the stability of biopharmaceutical formulations include determination of the level of aggregates and insoluble particles of various sizes; detection of increased immunogenicity; determination of the level of amino acid oxidation, in particular, methionine and tryptophan; and evaluation of the fragmentation of polypeptide chains, all for samples subjected to both stressed and unstressed conditions, according to Maggio.

	Most of the tests are accelerated aging studies, such as heating at elevated temperatures for 1 to 4 weeks, shaking at high temperature with an air/liquid interface for several days, and freezing and thawing for multiple cycles. Temperature ramping experiments are also common for determination of the melting and aggregation temperatures under different conditions, but Bukrinski has on several occasions observed that results from such studies do not always correlate well with shelf-life stability.

	A whole host of advanced techniques are employed to analyze the results obtained in the above tests. Some examples include polyacrylamide-gel electrophoresis (PAGE), particle-size analysis using micro-flow imaging (MFI), dynamic and static light scattering, size-exclusion chromatography (SEC), ion-exchange chromatography (IEX), and reversed-phase (RP) high-pressure liquid chromatography (HPLC), liquid chromatography/mass spectrometry (LC/MS), peptide and glycan mapping, circular dichroism (CD), Fourier transform infrared spectroscopy, and other advanced MS methods.

	A newer method that, according to Bukrinski, is increasingly used as an orthogonal method for the validation of SEC results is field flow fractionation (FFF). Instead of passing through a packed column, the fluid passes through a narrow channel and a force applied perpendicular to the channel causes separation of the particles based on their different mobilities.

	Not all methods will work equally well for each biologic API. Therefore, the first step in any study is to determine which methods provide accurate results for the protein in question. “It is very important to first determine the appropriate methods for stability evaluation, followed by optimization of the pH and ionic strength for the formulation. It is only at that point that screening for appropriate excipients can be effective,” Bukrinski says.

	Because so many stability tests must be performed for the development of an effective biopharmaceutical formulation, he also notes that the ability to perform rapid, high-throughput screening is imperative. “We perform as many analytical methods as we can using robotics systems that can evaluate small quantities of samples in microplates. Often times, however, tight timelines will dictate a less optimal formulation development process,” notes Bukrinski.

	“Two of the most efficient screening tools include biophysical characterization and high-throughput degradation analysis under stressed conditions,” adds Kang. “Biophysical tools, such as differential scanning calorimetry (DSC) and CD, directly measure the effect of excipients on the higher order structure of the proteins, which is often a good indicator of protein stability. High-throughput methods such as plate-based dynamic light scattering (DLS) can screen a large amount of excipients in a short period time.”

	Compared to small molecules, however, protein degradation involves multiple pathways, and a simple Arrhenius equation is generally not applicable, according to Kang. It is, therefore, difficult to predict long-term real-time stability through stressed and accelerated stability studies. “New mathematical models are currently being developed to establish a correlation between protein stability at different storage temperatures, and if successful, will greatly enhance formulation development and shelf-life prediction for protein drugs,” he states.

Automated stability determination using chemical denaturation


	Because temperature has such a significant influence on protein aggregation, it is important to obtain accurate thermodynamic data on proteins. However, temperature denaturation is often irreversible and accompanied by aggregation and precipitation and, therefore, not suitable for thermodynamic analysis.

	On the other hand, isothermal chemical denaturation is generally a reversible process for most biologic drugs and has been proven to provide reliable thermodynamic data of use for the evaluation of protein stability. AVIA Biosystems has developed an automated protein denaturation system for protein stability analysis that uses intrinsic (or extrinsic) fluorescence to monitor the protein conformational changes associated with protein unfolding (denaturation). Using the instrument, the identification of solvent conditions that maximize the structural stability of biologics can be completely automated, according to the company. Proteins and other compounds can also be ranked by their stability and propensity to aggregate.

Few approved excipients creates challenges


	Although the large number of stability tests and analyses required for the development of stable protein formulation present difficulties for biopharmaceutical companies, one of the biggest challenges for formulators is the limited number of known and approved excipients for protein stabilization, according to Kemter. In addition, many biopharmaceutical manufacturers will only use excipients that are approved internally, and generally only consider other excipients that are approved but have not been previously used by the company if no solution is possible with the existing choices, according to Bukrinski. “New excipients--whether just new to the company or compounds that have not been used as excipients before--will only be used if they have significantly high value and can compensate for the extended approval process. There must be a justification for the additional cost and risk,” he says.

	Kang agrees that the biggest challenge in developing new stabilizing excipients for biologics is the additional safety data required to introduce a novel excipient to a pharmaceutical product, particularly a parenteral formulation, which is the form in which most protein drugs are formulated. “The resources and time associated with this requirement makes formulation scientists hesitant to try new excipients,” he observes.

Pharmaceutical Technology's APIs, Excipients, and Manufacturing Supplement

:
	When referring to this article, please cite it as C. Challener, “Excipient Selection for Protein Stabilization," 

 APIs, Excipients, and Manufacturing Supplement 39 (18) 2015.

Articles

The complex task of stabilizing proteins is made more challenging due to the limited number of approved excipients.

Excipient Selection for Protein Stabilization

Despite a few significant investments in continuous manufacturing facilities by pharmaceutical companies, including Vertex Pharmaceuticals, Johnson & Johnson, GlaxoSmithKline, and Novartis (1), the adoption of flow chemistry for commercial production of APIs generally remains in the early stages. FDA has encouraged the adoption of continuous manufacturing since 2004, but specific guidelines are lacking from the agency and other regulators around the globe. Both former FDA Commissioner Margaret Hamburg (1) and Center for Drug Evaluation and Research Director Janet Woodcock (2) have been more vocal about the issue, particularly in relation to the proposed 21st Century Cures Act. This legislation requires FDA to support the development and implementation of continuous manufacturing for drugs and biologicals as one of several approaches to speeding up drug development and commercialization (3). In addition to a lack of comprehensive regulatory guidance, however, a dearth of industry personnel with expertise in flow chemistry is a hindrance to rapid adoption of continuous technologies.

Early adopters and fast followers drive change


	“While the industry as a whole is in the early stages of adopting flow chemistry for small-molecule API manufacturing, the early adopters and fast followers not only recognize that flow chemistry is the future of manufacturing, but also believe that they can implement it effectively,” asserts Tim Jamison, professor of chemistry and incoming department head at the Massachusetts Institute of Technology (MIT) and CEO of Snapdragon Chemistry, a new company dedicated to catalyzing the adoption of continuous flow synthesis. “Ultimately, when these companies realize the many benefits of flow chemistry, including reduced operating costs, footprint, and capital expenditures combined with improved process efficiencies, control, and product quality, investors likely will modify their expectations and demand this increased value from the industry as a whole. The rest of the industry will then have to scramble to catch up, and the early adopters and fast followers should reap the rewards of their forward-thinking actions. At that point, the entire industry--out of necessity to remain competitive--likely would shift its view on flow technology,” he explains.

	Most large pharmaceutical companies, according to Peter Poechlauer, innovation manager with Patheon, have at least installed advocate groups with a mission to showcase successful applications of flow processes. Dominique Roberge, head of chemical technologies with Lonza Pharma & Biotech agrees that flow chemistry has become an accepted technology for small-molecule manufacturing, largely for the development of new chemical reactions that have not been feasible in batch operations, to reduce the cost of goods, and to decrease capital expenditures (CAPEX) via process intensification. “These projects are significantly more focused and give a better understanding of what is achievable via flow chemistry. As a result, we have moved past focusing on feasibility studies only and are now evaluating projects that are more mature for tech transfer and scale-up,” he notes.

	“The decision to develop a certain step as a flow process is, however, opportunistic and may be motivated by the speed provided in early development, the need for smooth scale-up of hazardous reactions, and/or the savings in investment for a new drug whose future is still uncertain,” Poechlauer observes. In addition, he notes that these criteria apply to single steps of multi-step pharmaceutical syntheses, and therefore hybrid approaches that combine continuous and batch operations are most common; few companies have developed end-to-end continuous syntheses of pharmaceuticals that combine drug-substance manufacturing and drug-product formulation.

	“Continuous flow manufacturing occupies a similar position to where a technology like spray drying was 10-15 years ago. It shows great promise but the ‘mainstream’ commercial viability within the pharmaceutical industry has yet to mature,” says Patrick Kaiser, a principal scientist in the process development business of SAFC. He believes that, like spray drying, a clear commercial pathway and, more importantly, a clear regulatory pathway will drive more entrants as developers embrace the technology’s promise. Even so, Kaiser expects there will be limitations to its full embrace, because certain systems lend themselves to be more relevant to continuous processing, while others make less sense to perform via continuous means, and this dissonance complicates the pathway forward. “The reality is that CMOs must embrace such disruptive technologies moving forward to ensure long-term competitiveness in the marketplace,” he concludes.


	A positive indication for the future of flow chemistry is its increasing use for different types of reactions and downstream separation/purification operations. With respect to chemical reactions, using flow processes allows better control of yield and selectivity, which have a direct influence on purification and separation steps, according to Roberge. “The best approach is to develop new processes that can and will lead to a significant improvement in the synthetic route, but will typically not work in a traditional batch process,” he comments. His examples include various oxidation reactions (with molecular oxygen or hydrogen peroxide), azide chemistry, and high-temperature/pressure reactions developed via microwave chemistry.

	Jamison adds that the use of flow chemistry for photochemical and electrochemical reactions is also exciting, because these reaction classes are typically difficult to carry out and control and also challenging to translate from laboratory scale to pilot or manufacturing scale, but afford significant opportunities to access completely novel chemical scaffolds and greatly streamline current synthetic routes. “By using flow chemistry to gain better control, predictability, and scalability for these reaction classes, chemists can increase their utilization, which will ultimately result in significant advances and broader adoption in the pharma industry,” Jamison states.

	Integration of chemical synthesis reactions in flow with an increasing diversity of work-up operations in flow, such as continuous extraction, membrane processes, and the crystallization and separation of solids, is also an important development, according to Poechlauer. He further notes that the application of parallel, analytical instruments with sufficiently short response times have been developed, allowing efficient control of these processes.

	The development of a complete toolbox of flow reactors that can be used for all types of reaction rates and phases (e.g., liquid-liquid, solid-liquid, etc.) is also necessary for the broad application of flow chemistry to be achieved, according to Roberge. To address some of this need, Lonza has developed the pulsating coil reactor for liquid-liquid phase reactions and an efficient de-plugging system based on ultrasound technology. “Ultimately, true innovation in this field will come from a few key players in the market who have the variety of experience and infrastructure to optimize multiple reaction platforms,” asserts Roberge.


	In fact, the inadequate supply of scientific talent and expertise necessary to implement continuous flow technology at scale is perhaps the largest factor hindering more rapid adoption of flow chemistry for small-molecule API synthesis, according to Jamison. “It’s not as simple as asking current chemists to start working with continuous flow technology. That is like asking a saxophonist to play oboe. While both instruments are woodwinds, a saxophone uses one reed, while the oboe uses two. Thus, saxophonists can certainly become oboists, but it is not automatic; there will be a learning curve in most, if not all cases. Currently, flow chemistry is generally not in most university chemistry curricula. Thus, there will continue to be a lack of expertise in this area until this situation is changed,” he explains. “In the long run, industry demand will accelerate such changes; as the industry requires more expertise in this area, education/training standards will shift to meet this demand. This shift will not occur immediately, however, and there will likely be a short- to mid-term lack of human resource supply,” Jamison says.

	Rhony Aufdenblatten, manager of small-molecule business development with Lonza Pharma & Biotech, agrees that flow chemistry remains a specialized technology, because it requires specific technical know-how that can only be developed over years of manufacturing different chemical products. “The key challenge for any small-molecule development program is management of scale-up of the lab process for industrialization. Moving this type of scale-up into a new platform like flow chemistry can only be handled by the few players who have experience working with a variety of chemistries and processes,” he observes.

	Poechlauer is not convinced that continuous processing is “experts only” territory any longer, but he also believes that this perception certainly affects decisions regarding adoption of the technology. As a result, he does believe that CMOs with a proven track record in continuous processing may be favored as demand for this capability increases.


	Two other factors that are influencing the rate of adoption of flow chemistry for API synthesis include a lack of clear regulatory guidelines and the existing batch-based manufacturing infrastructure. Although FDA representatives recently made a number of public comments in support of continuous manufacturing in the pharmaceutical industry, from a regulatory standpoint, there remains a need to develop clear, harmonized guidelines accepted across the various regulatory authorities that will facilitate the development of continuous manufacturing routes in a manner that guarantees a consistent way to monitor/regulate their output, according to Jamison. In addition, while there are a growing number of flow processes being filed with auditors despite this lack of clarity, Poechlauer notes that there is little experience with respect to the auditing of continuous process steps.

	The International Conference on Harmonization (ICH) may be an effective body for achieving international guidelines. The organization may in fact be making continuous flow manufacturing a focus issue over the next year to 18 months.

	The impact of existing batch plants is less straightforward. For companies with unused batch capacity, Aufdenblatten notes that additional motivation for further investment into large-scale flow chemistry will be needed to overcome the additional CAPEX for flow infrastructure. “Typically, a gain in yield of 2-3% will not be sufficient; reduced cost of goods, safer processes, and breakthroughs in process platforms must also be considered,” he explains.

	Jamison points out, though, that existing infrastructure is in various stages of maturity, and new capacity (whether batch or continuous) could be established naturally, with continuous plants being built in place of new batch plants, as appropriate. “In addition,” he says, “continuous manufacturing plants could have a 10- to 100-fold smaller footprint than a batch plant of comparable output. Therefore, the CAPEX investment is smaller, which might sway the economic analysis to favor the business case of mothballing a significant number of existing batch-mode plants.”

	The mentality of the process development function must also be considered, according to Kaiser. The use of flow chemistry earlier in development requires both the innovator and CMO to be open to using continuous systems as an option in their process development efforts. Doing so potentially requires developing an initial batch process backed up by a second-generation flow process, perhaps simultaneously if the drug is on an accelerated approval pathway, which is not a small shift in today’s “fast-fail” drug development business. “Development companies are generally averse to investing too much in the manufacturing process until there is good clinical data to show manufacturing on a larger scale is necessary. Unfortunately, waiting too long to develop a continuous process also complicates a company’s regulatory strategy and potentially a challenge with different impurity profiles from different manufacturing processes,” he comments. The best strategy to combat this challenge, according to Kaiser, is to have experienced flow chemistry experts recognize where continuous systems provide the greatest opportunity for results early on in the process development effort.

End-to-end implementation of flow chemistry


	Snapdragon Chemistry-a new company launched in early May 2015 by CEO Tim Jamison, a professor of chemistry and incoming department head at the Massachusetts Institute of Technology (MIT), with Boston University assistant professor of chemistry Aaron Beeler-intends to provide drug companies with assistance in developing flow chemistry strategies and systems. 

	“Snapdragon Chemistry has the capability to help a drug-maker from end-to-end in implementing continuous flow manufacturing,” Jamison says. Backed by a scientific advisory board that includes Steve Buchwald and Klavs Jensen, well-respected MIT chemistry and chemical engineering professors, respectively, the company is offering solutions services that include evaluation of entire portfolios of molecules in development to identify the greatest opportunities for implementation of continuous flow chemistry. The company also offers the translation of desired organic syntheses end-to-end into a continuous-flow format, including the selection and/or design and development of laboratory- to commercial-scale continuous reactors.

	“Our goal is to help adapt flow technology solutions to the context of a specific company. We have the deep technical expertise across chemical synthesis and engineering to help companies make the appropriate decisions on where to invest in flow chemistry and increase their probability of success in implementing the technology,” he adds. Snapdragon Chemistry has already formed partnerships with companies, such as drug-discovery service provider Paraza Pharma and equipment manufacturer Zaiput Flow Technologies, to bring the most innovative and effective continuous flow solutions to clients, according to Jamison.


	SAFC has the capability to manufacture products using continuous systems on a 1-50 kg/day basis and currently is producing monomers, reagents, and building blocks for pharmaceutical and high-technology applications. The company is also investigating the addition of cGMP capacity that will include continuous manufacturing technologies.

	Lonza has produced many intermediates on kilogram and hundreds of kilogram scales for APIs in clinical trials. In 2013, for instance, the company completed a technology transfer to a large-scale manufacturing process under high-temperature and pressure conditions that were not possible in a batch process, according to Rhony Aufdenblatten, manager of small-molecule business development with Lonza Pharma & Biotech. Lonza has also used flow chemistry for the multi-ton production of a niacin product under GMP conditions.


	1. J.D. Rockoff, “Drug Making Breaks Away From Its Old Ways,” 

, February 8, 2015, www.wsj.com/articles/drug-making-breaks-away-from-its-old-ways-1423444049, accessed June 2, 2015.
	2. J. Wechsler, “

Congress Encourages Modern Drug Manufacturing

 website, May 1, 2015, , accessed June 2, 2015.
	3. Energy & Commerce Committee, United States House of Representatives, “

Full Committee Vote on the 21st Century Cures Act

,” May 19, 2015, , accessed June 2, 2015.


	When referring to this article, please cite it as C. Challener. “Lack of Expertise Hinders Adoption of Continuous API Synthesis,” 

Early adopters look to speed up the use of continuous API manufacturing, but lack of expertise hinders adoption of continuous API synthesis.

Accelerating Continuous API Synthesis

Perfusion technology was first introduced in the late 1980s to boost the low product concentrations obtained from the early cell lines used for biopharmaceutical manufacturing. Dramatic increases in titers for batch and fed-batch cell-culture processes over the next two decades largely eliminated the need for perfusion and interest waned. A key driver today is the production of unstable proteins that require low residence times in the bioreactor. Recently, however, rising cost pressures have been driving the need to achieve greater efficiency and productivity and reduce costs. Continuous bioprocessing is seen as a possible solution and, consequently, perfusion is once again attracting significant attention.

	In perfusion mode, high cell numbers are sustained for much longer periods by constantly feeding fresh media and removing spent media while the cells remain in culture. With this approach, optimum conditions for growth and production are maintained by supplying the appropriate nutrients and removing toxic waste products. Because the product is also regularly removed and separated from the waste products that can cause degradation, perfusion is highly beneficial for biologic APIs that are unstable under production conditions.

	Different methods are used to remove the spent cells, including centrifugation, in which the cells

	are separated and then returned to the reactor; alternating tangential-flow (ATF) filtration, in which the cells remain in the reactor and the spent media and product are collected in the filtrate; and the use of adherent cells that bind to capillary fibers or membranes in the bioreactor, which allows easy separation from the spent media.

	Introduction of the ATF technology (by Refine Technology; now owned by Repligen) was quite significant for perfusion processes, according to Parrish Galliher, CTO of Xcellerex, which is part of GE Healthcare’s Life Sciences business. “ATF and other filtration perfusion processes allow for complete clarification of the product, whereas with centrifugation, approximately 90% of the cells are recycled back to the bioreactor, and an additional filtration step is required to remove the remaining 10% of the cells from the product.” He also notes that advances in single-use technologies for perfusion have greatly simplified the process because sterilization and cleaning systems are no longer required, which in turn has made it much easier to set up and evaluate perfusion as a viable option for cell culture.


	As indicated above, perfusion cell culture has been used in the industry to produce biologics clinically and commercially for more than 25 years. Currently, a handful of companies manufacture biopharmaceutical drug substances using perfusion technology, largely because these products are not stable in the bioreactor and must be removed quickly to minimize degradation/modification, according to Galliher.

	Many more companies are looking at perfusion at the process development and clinical scales, however. This interest goes beyond the desire to improve the product quality of sensitive biologics; perfusion has the potential to reduce the scale at which reactions can be performed and increase productivity through better equipment utilization. “With the recent growing interest in continuous bioprocessing, there is renewed interest from manufacturers, equipment vendors, and academics in evaluating and pursuing perfusion technology,” says Chris Hwang, senior director of late-stage process development at Genzyme.


	Properly designed perfusion processes can significantly increase volumetric productivities (grams/L of bioreactor working volume per day) such that the bioreactor scale can be significantly reduced, which facilitates adoption of disposable technologies. The result is an increase in operational and capacity flexibility, reduced capital and operational costs, and increased speed (facility readiness and manufacturing operation), according to Hwang.

	In addition, due to the measurable decrease in production scale, it is possible to develop and scale up processes to production scale in development labs using the same control/automation systems and eliminate process scale-up in manufacturing, which significantly decreases technology-transfer risks. Site-to-site transfer risks can also be reduced because perfusion processes can be used to produce any protein (stable or non-stable), which facilitates standardization.

	Product quality is also important. In addition to minimizing the risks associated with product degradation/modification due to the reduced product residence time in the bioreactor compared with what occurs in a fed-batch system, perfusion processes can achieve steady-state product quality and simplify downstream operations, which all result in decreased heterogeneity and greater product consistency.

	The jury is still out, however, on whether perfusion and continuous biomanufacturing can truly increase development speeds and decrease costs for products that do not have stability/degradation issues, according to Galliher. “Perfusion is very easy to achieve on the lab scale, but there are many reports of additional challenges to its use on the commercial scale.

	Opinions vary widely among those who have been involved in implementing larger commercial perfusion processes, with some eager to use the technology again and others who are not,” Galliher says. The potential advantages are sufficiently significant, though, and many people are evaluating perfusion to determine its applicability for their biopharmaceutical processes.

Integration with downstream continuous processes


	Integration of a perfusion bioreactor process with continuous downstream processes (integrated continuous biomanufacturing, or ICB) increases the advantages of perfusion, according to Hwang. “If all of the unit operations between the bioreactor and downstream purification are performed continuously, one can minimize hold times and eliminate many unit operations, and all of the downstream equipment can also be miniaturized and single-use technologies can be utilized. The application of appropriate process analytical technology (PAT) enables automation of the entire production process without operator intervention, except to ensure uninterrupted medium and buffer supply,” he explains.

	Genzyme/Sanofi has successfully demonstrated at production scale the integration of a perfusion bioreactor with continuous capture for model enzymes and monoclonal antibodies with the products Cerezyme (imiglucerase for injection), Fabrazyme (agalsidase beta), and Myozyme (alglucosidase alfa). In addition, proof-of-concept of a fully continuous process (end-to-end) has also been demonstrated at the lab scale, according to Hwang.


	Despite changing perceptions of perfusion and continuous processing, significant challenges do remain. Some are very practical issues. A lack of adequate sensor technology for the on-line direct evaluation of product quality parameters-protein folding, aggregation, glycosylation, oxidation, contamination, etc.-is a risk, according to Galliher. “Product quality at present can only be monitored indirectly on-line through determination of solution temperature, pH, conductivity, etc., and the product quality cannot be determined until it is sampled and analyzed.

	As a result, more business risk is accumulating the longer a process is run, and that risk is compounded even further if perfusion is directly integrated with multiple continuous downstream processes that occur without product sampling and analysis,” he explains. Robust in-line product quality sensors will reduce this operational risk.

	Process quality by design (QbD) and design-space studies can also be more complex, because the ability to maintain product quality over an extended period of time must be verified for a full production run, which can go on for weeks or months. These longer run times can extend development times, according to Parrish, because QbD screening and optimization experiments take longer. It is possible to reduce the number of process parameters that are investigated to keep the development time similar to that of a batch/batch-fed process, but that may lead to the development of a less robust process. The positive view that FDA has of continuous manufacturing presents an opportunity to evolve the guidelines for QbD studies when applied to continuous processing.

	Unexpected costs can also reduce the positive impacts of perfusion. As an example, Galliher notes the benefit that a perfusion bioreactor can be scaled down by as much as 10-fold compared to a conventional bioreactor, enabling adoption of single-use technology with a smaller footprint and lower costs. Because the process is continuous, additional sterile-feed and sterile-harvest collection vessels are required to keep the process operating. “These additional vessels take up room and carry additional costs that must be considered when determining the overall benefits of continuous biomanufacturing,” Galliher observes. Concentrates of cell-growth media and buffers coupled with in-line dilution can reduce these additional costs. Hwang also notes that in an ICB process, many of these tanks would be eliminated.

	Other challenges have more to do with perceptions and company culture. “For companies that have historically relied on batch upstream and downstream processes for protein production, moving to perfusion/continuous processes can be difficult due to management concern about implementing what are considered to be new technologies for the company, particularly with respect to the potential for increased risk and development times,” Hwang explains. There is also often a desire to fully utilize existing production capacity, and it can be very challenging to overcome the deeply entrenched batch technology mindset throughout an organization. For existing products, there are additional concerns about implementing continuous processes as part of lifecycle-management strategies with respect to product comparability.

	To overcome these challenges, Hwang recommends that, to gain experience, companies invest resources and capital for the testing of continuous platforms that is de-coupled from product development. That way, risks can be minimized because a new platform won’t be considered ready for manufacturing until the expected process benefits can indeed be achieved.


	Given the heightened interest in perfusion and continuous bioprocessing, it is not surprising that equipment and device manufacturers are working to develop improved solutions for perfusion processes. “Key goals are to make perfusion more robust, scalable, and easier to operate,” says Galliher. GE is looking at all aspects of the system, such as in-line dilution systems and improvements of the filtration process (prevention of clogging) to increase flow rates for larger-scale, single-use filtration systems.

	Both Galliher and Hwang agree that the development of novel, robust PAT sensors for in-line monitoring of process performance and product attributes in real time will be crucial for the more widespread adoption of perfusion and other continuous bioprocesses on a commercial scale. For full integration into disposable platforms for protein production, pre-sterilized production-scale perfusion systems must also be developed, according to Hwang.

	“FDA has been encouraging the industry to transition to continuous manufacturing (CM) since 2011 because the agency sees CM as a key enabler for modernizing pharmaceutical manufacturing. I anticipate that with the continued improvement of bioreactor cell densities and further increases in volumetric productivities, there will be growing interest in the use of perfusion technology for protein production,” says Hwang. “Coupled with the potential significant strategic advantages of ICB, I suspect that industry adoption will continue to increase and further fuel advancement of the intensification of perfusion and continuous bioprocesses,” he concludes.

:
	When referring to this article, please cite it as C. Challener. “The Potential of Perfusion,” 

Potential for improved product quality and cost/time savings is reviving interest in perfusion technology.

The Potential of Perfusion

Pursuing Perfusion

FDA, Congress, and early adopters look to speed up the use of continuous API manufacturing.

Lack of Expertise Hinders Adoption of Continuous API Synthesis

Many transformations that afford important categories of organic building blocks useful for the synthesis of pharmaceutical intermediates and APIs cannot be performed due to hazardous conditions. Reactions that require the use of unstable reagents lead to the formation of highly reactive intermediates, and those that are highly exothermic typically are not suitable for implementation on a large scale because of the hazards they pose.

	Research is ongoing in both the industrial and academic setting to develop alternative approaches that will enable these transformations to be performed safely on a commercial scale. Such approaches include the development of new forms of hazardous reagents that are stable and can be readily manipulated in the plant setting; identification of new reaction conditions, such as the use of flow chemistry, to reduce the quantities of hazardous reagents and/or provide highly controlled conditions for energetic reactions; and the identification of entirely new routes that do not have any of these issues.

	Examples of recent approaches to overcome hazardous transformations are described in the following sections.
	
	


	In the Vilsmeier (or Vilsmeier-Hack) reaction, aryl aldehydes and ketones are obtained from substituted amides via their reaction with phosphorous oxychloride and electron-rich aromatic compounds. The substituted chloroiminium ion formed from the reaction of the substituted amide with phosphorous oxychloride is referred to as the Vilsmeier reagent (VR). While this reaction can provide a wide range of interesting carbonyl compounds, its use on a large scale can be difficult, because VRs are often irritants and also hazardous reagents due to their high thermal energies of decomposition.

	Researchers at Novartis Pharma employed flow chemistry to address this issue (1). They achieved the in-line formation and immediate consumption of VRs using a combination of batch and flow technologies and applied the new process to the synthesis of the anti-diabetic drug vildagliptin.

	Researchers at AMRI also performed a potentially hazardous cyclopropane ring-opening reaction in a continuous-flow reactor to increase the safety of the process (2). Subsequent copper-mediated Diels-Alder and palladium-catalyzed Negishi coupling reactions were also used in the optimized synthesis of taxadienone on a decagram scale.

	At Bristol-Myers Squibb, a hazardous analysis of a reaction involving the conversion of a tertiary alcohol to a hydroxypyrrolotriazine intermediate identified the potential for thermal runaway (3). To mitigate this potential, a continuous oxidative rearrangement process was developed that involves the mixing of three different, stable feed streams in a specific order using in-line static mixers. A “plug flow” design achieved using heat exchangers ensures sufficient residence times for completion of the reaction. The continuous process has been employed on the pilot plant and commercial manufacturing scales for the production of the investigational liver cancer drug brivanib alaninate.


	Advances in process analytical technology (PAT) not only help speed up process development and optimization, they can also facilitate the control, and thus increase the safety, of synthetic organic reactions. Researchers at Merck Sharp & Dohme recently reported the successful application of PAT for determining the appropriate timing of reagent addition and reaction quenching (4). The technologies discussed included process video microscopy (PVM), a new focused-beam reflectance measurement (FBRM) method, miniature process infrared (IR) spectroscopy, and a flow IR sensor. PVM was used as a nondestructive method for analyzing particle morphologies in real time, while the new FBRM technique has improved reliability due to its greater resistance to probe fouling. The miniaturized IR system was found to be easier to use at scale-up and more robust, and the miniature IR flow sensor was a more cost-effective tool with a faster response, and thus useful for continuous flow processes.

	The IQ Consortium, an organization of pharmaceutical and biotechnology companies with the mission of advancing science-based and scientifically-driven standards and regulations for pharmaceutical and biotechnology products worldwide, also recently reported on the current state of PAT use in the development of APIs (5). In the paper, the authors point out that PAT can minimize the hazards posed to operators when sampling hazardous materials and also provide more reliable data on processes that involve the use of hazardous materials. Through greater process understanding, increased process control-often through the use of simplified monitoring and control systems-is possible at the commercial production scale.


	One of the most effective methods for eliminating the concerns associated with hazardous reagents is to develop stable derivatives or totally new alternatives that accomplish the desired transformation without the safety concerns of the original compounds. Several examples have appeared in literature recently.

	Michael Willis at the University of Oxford developed a method for the synthesis of sulfonamides via 

 formation of intermediate sulfinates from magnesium-, lithium-, and zinc-based reagents and DABSO, a surrogate for sulfur dioxide (6). The sulfinates were reacted with N-chloroamines, which were also generated in situ by reacting amines with aqueous bleach. While sulfonamide groups are present in biologically relevant compounds, they are often accessed through the corresponding sulfonyl chlorides, which can be unstable. The new method provides a route that avoids sulfanyl chlorides. Primary and secondary amines, anilines, and amino acids were all suitable for the reaction, which proceeded most efficiently with Grignard reagents.

	Sodium metal is often used effectively for Birch reductions and the reduction of esters (Bouveault-Blanc reduction), but does have associated safety concerns at larger scale given the pyrophoric nature of sodium metal and the volatility of ethanol. Researchers at the University of Manchester and Pentagon Fine Chemicals reported that a dispersion of sodium metal particles with diameters ranging from 5-15 μm in a mineral oil is a nonpyrophoric, free-flowing powder alternative that can be handled in air and provides results similar to sodium metal (7). In particular, the Bouveault-Blanc reduction was found to proceed in high yields at 0 °C in hexane for a wide range of aliphatic ester substrates with 2.5 equivalents of isopropanol rather than ethanol and just 4.5 equivalents of the sodium dispersion. Notably, alkenes, aryl amines and fluorides, acidic protons, and sterically hindered carbon centers were tolerated

	Direct aryl C-H chlorination is another desirable reaction for the preparation of aromatic chlorides, but the most reactive methods typically require the use of hazardous reagents such as chlorine (Cl

). Milder reagents are available, including N-chlorosuccinimide (NCS) and 1,3-dichloro-5,5-dimethylhydantoin (DCDMH), but their applicability is limited due to their reduced reactivity.

	Other reagents that have been employed also suffer from unattractive features such as toxicity, hydroscopicity, light/heat sensitivity, and explosive properties among others. Phil S. Baran and colleagues at The Scripps Research Institute and Bristol-Myers Squibb have developed the guanidine-based chlorinating reagent 2-chloro-1,3-bis(methoxycarbonyl)guanidine (CBMG), which they refer to as Palau’chlor, as an alternative for the mild and safe direct chlorination of nitrogen-containing heterocycles and certain arenes, conjugated π-systems, sulfonamides, and silyl enol ethers (8). The new reagent is an air-stable solid that is thermally stable below 100 °C, yet achieves transformations that have only been possible in the past with highly reactive, hazardous chlorinating reagents. Importantly, the cost of the reagent and its functional group tolerability are similar to that of NCS.

, 19 (4), 551-554 (2015).
	2. S. G. Krasutsky, et al., 

. 19 (1), 284-289 2015.
	3. Thomas L. LaPorte, et al., 

. 18 (11), 1492-1502 (2014).
	4. George Zhou, et al., 

. 19 (1), 227-235 (2015).
	5. John D. Orr ,et al., 

. 19 (1), 63-83 (2015).
	6. M.C. Willis, et al., Angew. 

. 54 (4), 1168-1171 (2015)
	7. D.J. Proctor, et al., 

. 79 (14), 6743-6747 (2014).
	8. Phil S. Baran, et al., 

Safer reagents and reaction conditions are making many hazardous transformations possible.