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

Cynthia A. Challener

Innovative equipment, analytical techniques, software, and modeling systems are improving the tableting process.

Articles

Advances in Tableting

					Thinkstock Images/Stockbyte/GETTY IMAGES

	The market for formulated drugs based on highly potent active pharmaceutical ingredients (HPAPIs) is growing at a rapid pace, largely due to the development of highly targeted therapies based on antibody-drug conjugates, which can include cytotoxic small-molecule components. The manufacture of this expanding field of HPAPIs is challenging and requires specific know-how, facilities, equipment, and procedures designed to mitigate the risk associated with producing and handling potent compounds. Standards and technologies are continually changing, and HPAPI manufacturers must remain vigilant and prepared to adopt and implement the latest designs, equipment, training, and procedures to reduce the risks posed by HPAPIs.


	Although many challenges exist for high-containment API manufacturing, the variability and uncertainty associated with each compound present the greatest risks, according to Waldo Mossi, general manager of Helsinn Advanced Synthesis. “The importance of occupational exposure limits (OELs) is widely neglected in discovery research and early development,” he states. He explains that many companies use a one-size-fits-all approach to handling and managing the containment of bulk drug substances. Each individual process, however, offers different challenges, and no two new chemical entities (NCEs) are alike. The situation is aggravated by the lack of universally accepted definitions for various compound types, such as highly active, highly potent, and cytotoxic agents, which can lead to confusion between sponsor companies and custom-manufacturing organizations (CMOs), according to Mossi.

	To manage the variation and address the uncertainty associated with new substances, Helsinn continues to strive for design of toxicology testing and safety evaluation from the early stages of process development. The company also uses a comprehensive, science-based OEL evaluation approach from the start of an HPAPI project, and works with experienced industrial hygienists to assign initially conservative OELs to each potent compound that will enter its facility. “Since a certain level of risk will always exist when working with HPAPIs, it is important to foster a strong company culture of excellence in protecting employees, products, and the environment. Our comprehensive approach to process and compound evaluation helps to clearly define the needs and objectives in handling each process step,” Mossi observes. 

	Fortunately, as an HPAPI project proceeds through the development lifecycle and into clinical trials, the understanding of the risks associated with the potent compound increases and risk mitigation generally becomes less difficult, according to Patrick Klipstine, director of SAFC’s Madison, WI site. “During the development process, SAFC pursues ongoing internal evaluations and works with third parties to bolster this process. As the definition of potency becomes better defined during the development cycle, our process engineers and environmental, health, and safety (EHS) representatives can make appropriate modifications to the manufacturing engineering controls,” he notes. 

	Manufacturing and process continuity are also crucial during scale up to ensure that risks are minimized, according to Mossi. “Laboratories and small-scale GMP equipment should be designed so that they are aligned with the large-scale equipment used for commercial production in order to ease the transition and reduce uncertainty and risk during scale up,” he says. At Helsinn, the risk associated with scale up is reduced through facility design and investigated during early design of experiment (DOE) analyses.


	In any chemical manufacturing plant, the protection of operators is a top priority. In facilities producing HPAPIs, providing operator protection is absolutely critical and the top priority, according to Klipstine, which means that appropriate engineering controls are in place and personal protective equipment is available. In addition, every unit operation must be considered with regard to both the chemistry and potential occupational exposure. “Chemical processing steps are evaluated on their merits with regard to sound chemical process hazard criteria, and then each process is developed to allow for the safest execution of the process within the identified equipment train, taking into consideration compatibility with materials of construction, thermal output, gas evolution, waste stream management, etc.” Klipstine notes. With these considerations taken into account, SAFC then applies engineering controls for containment to mitigate the risk for occupational exposure. Specifically, the company has adopted a risk-minimization strategy that is systematically constant, but allows for different outcomes depending on the chemistry and potency of each API.

	“Each opportunity that turns into a project goes through a defined risk assessment process, with experts in our process development, EHS, and process engineering groups closely collaborating to provide a robust process from both a chemical engineering and process engineering standpoint. One of SAFC’s first principles is that all processing of powders and liquids are conducted in closed systems that have been verified to be effective for the prevention of occupational exposure. Second tier to these systems are robust training programs that have been designed for specific unit operations,” he adds.

	Helsinn also emphasizes the use of fully closed systems and isolation to avoid or mitigate areas of greatest risk. The company fosters the approach of contained chemistry, which means that equipment for each individual process (e.g., balances, rotary dryers, pressure filter dryers, and slurry vessels) is installed inside an isolator that has been qualified for occupational exposure levels down to nanogram levels according to the International Society for Pharmaceutical Engineering’s (ISPE) Standardized Measurement of Equipment Particulate Airborne Concentration (SMEPAC) methodology. The use of such an approach, according to Mossi, was made possible by a clean atmosphere design supported by accurate general ventilation using double-pass, high-efficiency particulate arrestance (HEPA) filtration. 

	It should also be noted, according to Mossi, that in addition to variable chemistry, operational risk depends greatly on several factors, including the company culture, personnel training, proper operational execution, and the design and engineering of the facility. In general, the greatest challenges are typically associated with operations such as sampling, loading/unloading of the reactor, and transfer of the material. “Powder handling presents the highest probability for potential worker exposure, and it is important to carefully study optimal methods for minimizing and, wherever possible, removing powder handling operations from an HPAPI process,” he states. At Helsinn, if powder handling is necessary, effective and consistent safeguards are factored in with redundancies to mitigate the risk even further. 

	SAFC handles both liquid and powder HPAPIs under a defined set of unit operations to minimize the potential for occupational exposure. “By using one common system defined appropriately for scale to ensure containment, our chemists can be assured that the processes they are executing are appropriately identified for the defined potency,” Klipstine explains. To reduce risk, all large-scale isolation of powders is conducted in jacketed filter dryers where solids can be filtered and dried without the necessity for discharge from the drying unit operation. Once dry, the HPAPI is discharged using glove-box containment techniques directly into predefined drug-substance packaging using ILC Dover continuous liner technologies, according to Klipstine. As a best practice, Helsinn investigates each manufacturing process step for hazard and safety together with the aid of an outside industry expert as part of its DOE analysis. 


	Of significance to HPAPI producers is having a thorough understanding of the cleaning procedures required to meet allowable carryover limits for multipurpose equipment, according to Klipstine. “Controlling cross-over contamination mitigates any potential risk to patients,” he asserts. Mossi agrees that safety cleaning verification at each stage of a manufacturing process and GMP cleaning validation is crucial. SAFC’s philosophy is to apply a continuous improvement mentality so that its systems will exceed current industry standards. Even so, one challenge the company faces regularly as a CMO with a multipurpose facility relates to servicing customers ranging from virtual biotechnology firms to large pharmaceutical manufacturers that have a wide range of expectations regarding handling and cleaning verification.


	Another challenge for CMOs that offer HPAPI manufacturing services is the continual evolution of industry standards and technologies. “Companies that want to participate in this market must adopt these newer technologies,” Klipstine asserts. SAFC, for example, had to transition to more robust analytical technologies with improved sensitivity and detection levels that allow for the determination of potential API carryover at part-per-billion levels.

	On the other hand, as the industry continues to mature, consultants and innovative equipment manufacturers can help design state-of-the-art engineering controls to better suit specific facility containment requirements, according to Mossi. He notes that Helsinn’s recent facility expansion was custom-designed to support workflow, ergonomics, and safety, while containing several unit operations within only a few isolators.


	Cynthia A. Challener, PhD, is a contributing editor to 

: When referring to this article, please cite as C. Challener, “Minimizing Risk during HPAPI Manufacture,” 

Protecting workers, patients, and the environment requires advanced technologies.

Minimizing Risk during HPAPI Manufacture

	The vast majority of new small-molecule drug candidates are chiral, and methods must be identified for obtaining the desired enantiomer in high purity. A common approach is to separate the enantiomers of a racemate. Certain chemo- and biocatalysts mediate transformations that provide chiral chemicals with high enantioselectivities; however, in most, both the desired and undesired enantiomers are obtained. Diastereomeric resolutions via salt formation, kinetic resolution, and chiral chromatography are techniques commonly used at commercial scale. All of these methods have drawbacks, and pharmaceutical companies are always seeking alternatives that carry lower costs and can be completed in shorter times


	Many compounds with bioactivity exist in two forms; enantiomers only differ by the way in which they rotate light. These enantiomers have identical chemical and physical properties, but interact differently with other chiral molecules, including many important biochemicals. Therefore, when a new candidate API is identified, a synthetic route that produces only the correct enantiomer must be developed.

	Finding efficient methods for the production of chirally pure APIs is a continually important topic in the pharmaceutical industry, according to Robert Hof, chief operating officer of contract research organization Syncom, based in Groningen, The Netherlands. Chiral racemates remain one of the major sources for the single enantiomers required for chiral APIs and drugs. “Resolution is absolutely necessary for separating enantiomers,” asserts R.M. Kellogg, a cofounder of Syncom. Fortunately, he notes that striking advances have been made in diastereomeric resolutions, kinetic resolutions (stoichiometric, catalytic, and parallel), and chromatographic techniques. “The availability of many different methodologies allows the selection of the best methodology for a particular compound,” Kellogg says.

	Chiral resolution by either diastereomeric salt formation or biocatalysis remains one of the most applied technologies due to its relative simplicity and scalability, according to Hof. “In particular, when racemization loops or-even better-dynamic kinetic approaches can be implemented for compounds with racemization-prone functionalities, 100% yield with 100% enantiomeric excess (pure enantiomer) can be achieved,” Hof adds.

	Resolution by diastereomeric salt formation is often applied to amines and acids because they have reactive groups that can form salts with chiral resolving agents. Alcohols are often resolved using biocatalysts. For apolar compounds including alcohols, many amines, thiols, and sulfides, biocatalysis, sometimes combined with organometallic catalysts, can be the method of choice, according to Kellogg.


	When a racemate is separated via diastereomeric salt formation, the two enantiomers react with a single enantiomer of a chiral compound that will form a salt with both enantiomers. Because each component of the salt is chiral, the product has two chiral centers and is considered to be a diastereomer. Unlike enantiomers, diastereomers have different physical properties and often can be readily separated from one another via crystallization.

	While there are well-established, inexpensive basic (e.g., cinchonidine and phenethylamine) and acidic (e.g., tartaric and camphor-1-sulfonic acids) resolving agents that are available on a large scale, efforts are continually being invested in the development of novel resolving agents that can be produced cost-effectively in quantities required for the production of commercial APIs. Syncom, for example, has developed strongly acidic chiral phosphoric and sulfonic acids. Others have also developed more economical, scalable routes for the production of older resolving agents that previously were not cost-effective for commercial API synthesis.

	One challenge with the diastereomeric salt approach, however, is that the resolving agent must be relatively cheap or readily recyclable, according to Hof. “Large-scale manufacturing is all about process simplicity and a low cost of goods. For chiral resolution, therefore, the key is identifying a relatively cheap or easily recyclable resolving agent and, ideally, racemization conditions for the unwanted isomer,” he says.


	An alternative approach is to find a way to separate enantiomers without the need to form diastereomeric salts. In fact, there are many types of chiral compounds that cannot form salts or for which it is difficult to do so due to steric hindrance, reduced reactivity, or other reasons.

	For these compounds, Syncom has focused on attrition-enhanced deracemization, or Viedma ripening, in which chiral compounds that are conglomerates (i.e., the enantiomers crystallize separately) are resolved efficiently without the need to use a chiral resolving agent. The separation is achieved by grinding the solid racemate of a conglomerate under racemizing conditions. Separation of the enantiomers occurs because in a conglomerate, each enantiomer has a greater affinity for the same enantiomer and ultimately the two enantiomers will crystallize separately to form a mechanical mixture. This process is accelerated by creating very small crystals for more rapid dissolution and when performed under racemizing conditions provides a pure, single enantiomer in high yield.

	Because few organic compounds are conglomerates, it is necessary to form derivatives of most substances to obtain conglomerates suitable for attrition-enhanced deracemization. A demonstration of this technique is an alternative synthesis of clopidogrel developed in the Syncom laboratories (1). “This need has stimulated search techniques for conglomerates (second harmonic generation techniques) and also more profound studies of crystallization behavior in order to predict the formation of conglomerates,” notes Kellogg.

	In addition, Syncom is working with various partners to develop standard pharmaceutical multipurpose manufacturing equipment for performing commercial-scale, attrition-enhanced deracemization. The company is also working to develop attrition-induced resolutions of non-racemisable conglomerates, and recently reported the resolution of a salt of omeprazole, a drug for treatment of the symptoms associated with gastroesophageal reflux disease (GERD) and other conditions caused by excess stomach acid (2). 

Enantiospecific cocrystallization in solution


	Researchers at Université Catholique de Louvain (UCL) in Belgium have developed another new resolution technology that does not require the use of chiral resolving agents. This approach is based on the fact that one enantiomer of a chiral compound will form a cocrystal with only one of the two enantiomers of a chiral coformer, allowing the resolution of a racemic mixture via enantiospecific cocrystallization in solution (3). 

	The technique was demonstrated using an enantiomeric mixture of 2-(2-oxopyrrolidin-1-yl)butanamide. The S-isomer is an API (levetiracetam) for the treatment of epilepsy that must be separated by chiral chromatography because no effective chiral resolving agent has been found. A structure-based cocrystal screen was performed to identify coformer candidates, and then ternary-phase diagrams were constructed for different temperatures to identify the ideal coformer and conditions for enantioenrichment. Notably, after just one cocrystallization step, 70% of the S-enantiomer was separated. 

	This approach to chiral resolution is attractive because it does not require the use of resolving agents and can be applied to compounds that do not readily or cannot form diastereomeric salts, many of which would require separation via chiral chromatography. The coformer technique is more cost-effective than chiral chromatography and can be performed using typical crystallization equipment (3). Syncom also participates in a project funded by the Dutch National Science Foundation in which co-crystallization together with attrition-induced deracemization is used to resolve commercial products, such as Naproxen.


	Even with these attractive alternative technologies, the search for improved chiral resolution methods is far from over. “Nature sometimes separates enantiomers for us. Roughly 10----15% of chiral organic compounds are conglomerates for which the enantiomers crystallize separately from one another. Could we imitate nature and learn how to crystallize compounds in a single step as single enantiomers and at the same time separate them? This scientific challenge remains as yet unanswered and is waiting to be met,” Kellogg asserts.

 13 (6), 1195-1198 (2009).
	2. R.M. Kellogg, et al., 

 17 (6), 946-950 (2013).
	3. G. Springuel and T. Leyssens, 

When referring to this article, please cite it as:
	C. Challener, " Chiral Resolution with and without Resolving Agents," 

Resolution technologies remain crucial for commercial-scale chiral API production.

Chiral Resolution with and without Resolving Agents

	During biosimilar development, it is necessary to demonstrate that the physicochemical properties, efficacy, and safety (e.g., immunogenicity) of the biosimilar are similar to those of the targeted branded biologic. Therefore, in addition to structural and biochemical analyses of biosimilars for confirmation of the physical and chemical characteristics of the biosimilar, it is necessary to demonstrate that the pharmacodynamics (PD), pharmacokinetics (PK), and immunogenicity of the biosimilar are also comparable to those of the branded drug. Consequently, functional assays, such as ligand-binding assays (LBAs) and mechanism of action (MOA) cell-based assays (CBAs), must be developed to effectively assess biosimilars. In many cases, LBAs are crucial for determining PK and PD behavior, and they are also often used as orthogonal assays to CBAs.


	Most biotherapeutics rely on specific binding activities to achieve a clinical effect. Therefore, matching the binding activities of a biosimilar candidate to those of the reference product is part of the process of developing a biosimilar and something that is expected by both the European Medicines Agency (EMA) and FDA, according to Gino Grampp, R&D policy director for Amgen. Physicochemical assays for testing of originator and biosimilar molecules, however, can only assess the actual structure of the molecules, and do not provide information on the functional conformation of the molecules and how they interact with the intended target. “Only a functional assay, such as an LBA or CBA, can fully address this issue,” says Michael Sadick, senior manager for large molecule analytical chemistry with Catalent Pharma Solutions. 

	For monoclonal antibodies (mAbs), which make up a large portion of the biosimilars currently under development, it is important to remember that the different ends of the molecule are involved in binding with different activities (the Fab and Fc regions, which bind to the antibody target/ligand and Fc receptors [CD64, CD32, CD16, and FcRN], respectively). While 

 biological activity assessments provide vital holistic information on biosimilarity, a combination of ligand- and Fc-binding assessments may provide more sensitive and mechanistic values for biosimilarity, according to Sadick.

	FDA, in its May 2014 draft guidance on how to show biosimilarity using clinical pharmacological data, indicated that LBAs, concentration and activity assays, and PD assays were important for biosimilar product development. In particular, according to Grampp, FDA suggested in the 2014 guidance that analytical assays should be able to specifically detect active drug, for example, via binding assays that use reagents specific for the binding sites for the drug (e.g., the complementarity determining regions [CDRs] of an antibody). “In this case, the binding assay is used not just to detect the molecule concentration, but to test the in-vivo properties of the molecule, and therefore, a well-designed and sensitive assay can be of great benefit to a biosimilar developer,” Grampp observes. 

	In October 2014, the American Association of Pharmaceutical Scientists (AAPS) Ligand-Binding Assay Bioanalytical Focus Group Biosimilars Action Program Committee published a whitepaper with recommendation for the development and validation of ligand-binding assays for PK assessments. Sadick notes that the AAPS whitepaper has similar conclusions regarding LBAs for bioanalytical testing as appear in the FDA 2014 guidance. “Biosimilarity testing for bioanalytical samples [PK/PD studies] is pretty much fully dependent on ligand-binding analyses. Binding assays are useful for not only assessment of the similarity of the final structures of the originator and biosimilar molecules, but also for assessment of the similarity of the antigenicity of the originator and biosimilar molecules. Binding assays work very well for the detection of antidrug antibodies, and for testing which of those antibodies may have drug-activity-neutralizing potential,” says Sadick.


	Binding activities can be determined by either cell-based assays or ligand-binding assays. Ligand-binding assays are of great importance for the analysis and comparison of the originator and biosimilar therapeutic for both testing of the drug before administration to preclinical or clinical recipients, as well as for testing of an active drug in preclinical or clinical recipient sera, according to Sadick. The latter would include both assessment of the function of the drug in the serum, as well as the activity of potentially induced host neutralizing anti-drug antibodies (nAbs). “Very importantly, for bioanalytical studies there are very few cell-based bioassays capable of working with sera derived from preclinical or clinical recipients. Thus, for bioanalytical assessments of biosimilarity, ligand-binding assays are in many cases the only functional assays available,” Sadick states. 

	In addition, ligand-binding assays enable the characterization of the potency of different biosimilars and provide crucial information on the affinity (dissociation rate constant, KD) and selectivity of biosimilars to specific targets, according to Roger Bosse, senior product leader for research reagents with PerkinElmer. He adds that regulatory authorities recommend simple and reliable methods for the characterization of biosimilars, and LBAs are one type of alternative.

	For drug substance/drug product testing, on the other hand, Sadick notes that cell-based functionality is suggested whenever possible. “For ligand-type therapeutic molecules (e.g., erythropoietin), ligand-binding assessments should probably not be used as a first choice for biosimilarity testing, because cell-based assessments are more biomimetic to the ligand mechanism of action as compared to ligand-binding assays and tend to be more sensitive to small differences or changes,” he says. If a cell-based bioassay cannot be established, however, Sadick indicates that ligand-binding assays would be the first choice for biosimilarity assessment. 

	In addition, for mAbs, ligand-binding assays do provide a valuable assessment of biosimilarity and should be included, orthogonally, with cell-based functionality assessments.


	It is, in fact, important to remember that the totality of evidence-a stepwise approach to evidence development that includes both structural and functional analyses-will ultimately be required to address residual uncertainties and establish biosimilarity, according to Margaret Karow, executive director of biosimilars process development at Amgen. “While target binding assays are a critical component of the analytical comparison of biosimilars to their corresponding reference products, the assays do not replace other analytical assessments required, nor do they provide information about the activities of the molecules, such as cellular signaling or effector functions (antibody-dependent cell-mediated cytotoxicity and complement-dependent cytotoxicity). Furthermore, the assays cannot predict the extent of clinical benefit or potential differences in safety, such as those identified risks for the reference product or unidentified risks for the biosimilar, and lastly, they are not predictive of a molecule’s potential immunogenicity,” she states.


	There are essentially two general methods for assessing ligand binding: ELISA (enzyme-linked immunosorbent assay)-type assays and kinetic binding assays. Classical ELISA assays were based on horseradish peroxidase or alkaline phosphatase and relied on color detection using a microplate colorimeter, which often suffered from poor sensitivity, particularly for samples containing serum, according to Sadick. Today, he notes that the final reagent/moiety of an ELISA test can be analyzed using several other techniques, including time-resolved fluorescence, electrochemiluminescence, and proximity-bead-based assays. “Each of these assay approaches provides differing strategies for increased sensitivity and decreased background or serum interference,” Sadick says. Matthew Reuter, biotherapeutics business development with PerkinElmer, adds, “These novel nonradiometric homogenous no-wash platforms aid in assay transferability, reproducibility, scalability, and overall cost reduction. They are also very robust and meet development criteria, which is equally important because typically LBAs are developed by upstream analytical method development groups and subsequently transferred to downstream QC-GMP lot release groups.”

	Ligand-binding kinetic assays for biosimilarity testing typically rely on surface plasmon resonance (SPR), according to Sadick. SPR measures binding in terms of the “change in mass” that occurs as the ligand and binding partner attach to one another and then separate. Advances in SPR analysis have resulted in increased robustness of the assays and reduced operator influence, as well as increased throughput for the determination of binding affinities. Biolayer interferometry (BLI) is a newer method that, like SPR, measures binding in terms of the “change in mass,” but uses eight biosensor tips rather than microfluidics. BLI also allows for hybrid binding for quantitative ELISA-like binding determinations, which may be applicable for PK/PD bioanalytical analyses, according to Sadick.


	One of the biggest challenges with LBAs is determining when they are appropriate. “LBAs are certainly easier for starting a development campaign, but recent regulatory guidelines indicate that a cell-based MOA bioassay should accompany the development of a drug candidate as early as possible,” says Reuter. If an LBA is used for determining the MOA or bioactivity of a molecule, it must eventually be translated to a cell-based format, and therefore it is necessary to demonstrate perfect correlation between the LBA and the cellular assay. “Adapting traditional complex cellular functional assays into simpler binding assays that rely on purified cellular components (e.g., cell surface receptors) is another important challenge,” Reuter adds. PerkinElmer is developing both binding and functional assays for characterization and cell-based determination of the MOAs of biological drugs (including biosimilars) that bind to specific cell-surface receptors, antigens, or soluble ligands using various technologies. 

	Target solubility is another issue, according to Karow. “High-throughput, kinetic assessment of target binding on cells for those targets that cannot be produced as soluble proteins remains a challenge for the industry,” she observes. Several companies are, however, developing technologies designed to overcome this problem, including high-throughput, high-performance, real-time, label-free analysis systems.

	For Catalent, the need to derivatize therapeutic mAbs, or the “test” moieties, to immobilize them has been a concern. “It is more optimal to immobilize the receptors so that the mAb samples could be analyzed without any derivatization,” says Sadick. Therefore, the company immobilized all six recombinant Fc receptors on nickel-coated biosensor tips for use in BLI analyses. “With this approach to binding kinetic analysis, the KD values for mAbs, both originator molecules and biosimilars, can be accurately and reproducibly determined,” he asserts.

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

Ligand-binding assays are fundamental to characterizing biosimilars.

Ligand-Binding Assays and the Determination of Biosimilarity

	Approaches to drug development are changing, as pharmaceutical companies use advanced technologies to achieve targeted goals. Today, rational drug development strategies based on modeling, simulation, and biomarkers complement such traditional drug discovery methods as structure-activity relationship (SAR) studies. Biomarkers are also being used to evaluate drug performance during preclinical testing and clinical trials. Assay development and validation, however, can be more challenging for biomarkers than it is for APIs. This article will examine some of these issues.
	
	Many of the challenges of working with biomarkers stem from the fact that, unlike APIs in neat solutions, they are endogenous compounds in biological fluids. For example, isobaric analytes such as bile acids in matrices can lead to higher background noise and interference in biomarker assays. In addition, biomarkers are typically present in trace amounts, unlike APIs, according to Stephanie Pasas-Farmer, senior director of bioanalytical labs with Frontage Laboratories.

	Furthermore, analytical method development for APIs tends to be more straightforward because reference materials with known concentrations are available and can be used to spike into the matrix. With biomarker assays, internal standards are not often available, and, generally, an endogenous level of the biomarker is already in the matrix, which makes quantitation a challenge, according to April Brys, director of biomarker services with Battelle.

	Regulatory requirements for analytical methods also differ for APIs and biomarkers. API analytical method development and validation must meet US Pharmacopeial Convention requirements or guidelines set by the International Conference on Harmonization (ICH).

	Biomarker assay development and validation, meanwhile, depends on the assay’s intended use. If it is being used as a diagnostic tool, it must meet requirements of the Clinical Laboratory Improvement Amendments (CLIA).  If it is being used to establish a pharmacokinetic/pharmacodynamics (PK/PD) endpoint, it must meet good laboratory practice (GLP) standards.

	Many of the same analytical techniques, however, are used for API methods and biomarker assays. “Small-molecule biomarker assays are typically developed using a liquid chromatography/mass spectrometry (LC/MS) platform, which provides a great balance of sensitivity, selectivity, robustness, precision, accuracy, and ease of operation,” states Patrick Bennett, executive director of biomarker operations at PPD. The LC/MS platform is also beneficial for biomarker analysis because it can be used to perform multi-component analyses simply, including metabolomics-based analyses of hundreds of molecules per sample.
	
	


	For biomarker assays, regulatory agencies prefer that calibration standards and quality control samples be prepared using the same biological matrix (whether blood, urine, plasma, or other biological fluid) found in the samples to be analyzed. However, they should be free of biomarker, which can influence the accuracy of readings. Unfortunately, it is difficult to obtain analyte-free matrices.
	
	One option, according to Pasas-Farmer, is to use a surrogate matrix that is biomarker free. A second option is to remove the endogenous analyte/biomarker from the matrix through chemical treatment, charcoal stripping, selective extraction, or some other method. In both of these cases, however, the chosen matrix will be different from the study sample matrix. “With surrogate, stripped, and depleted matrices, there may be issues with solubility, stability, non-specific binding, or ionization matrix effects due to the differences in the modified matrix and the authentic matrix,” says Bennett. Consequently, it will be necessary to design validation experiments to confirm that the ability to quantify analytes in a modified matrix is similar to that for analytes in the authentic matrix. If an authentic matrix is combined with standard addition, Bennett adds, experiments and processes will be necessary to determine the baseline and final concentrations and generate bridging data from newer lots.

	These issues may be avoided by using a standard addition approach with the true biological matrix once the basal levels of the endogenous biomarker have been calculated for the specific lot of control matrix. Stable, labeled reference material can be used as a surrogate standard, and labeled internal standards can be effective. These approaches, however, can also be challenging because, often, reference standards are not available, or are difficult to synthesize, for biomarker assays. Even if they are available, whether they are labeled or unlabeled, they can be expensive.

	In some cases, depending on the detection method and the type of matrix, samples may have to be purified before a biomarker assay can be completed. This, in turn, can create various problems, including recovery and matrix interference issues, notes Brys.
	
	


	Before method validation, it is crucial to identify the intended use of the biomarker assay. “The investigator must understand the ultimate use of the data generated,” says Pasas-Farmer. If biomarker data are to be used for early screening and evaluation of the mechanism of action (MOA) of a drug candidate, then a qualified assay that is semi-quantitative or relatively quantitative can be used, she says, but if the biomarker data are used to support claims for the safety/efficacy of dosing in a pivotal preclinical or clinical study, and if they will be used in a regulatory filing in support of a new drug application (NDA), then a fully validated assay should be implemented.
	
	Finally, if the assay will be used as a diagnostic tool to determine the disease state of an individual, then CLIA validation requirements must be met, she explains.

	Discussions are still ongoing, however, about whether biomarker assays should be fully validated, or if qualification of an assay can still be used if the new chemical entity is in the early stages of development. Bennett notes, for example, that, even if no compliance is deemed necessary, extensive method performance qualification may still be warranted to ensure that the biomarker assay can provide adequate confidence in the data required.

	Validation of an API analytical method is typically focused on the validation of specific assay performance parameters such as the precision, robustness, linearity, and lower limit of quantification (LLOQ)/lower limit of detection (LLOD). In contrast, validation of biomarker assays involves determination of “fitness for purpose” and must demonstrate that the assay meets the predetermined purpose, according to Brys, who defines the stages of biomarker validation as follows:

		Defining the purpose and selection of the assay

		Identifying the critical reagents and standards

		Writing the procedure and validation plan

		Performing the actual study validation, which typically includes assessment of fitness for purpose, robustness of the assay in the clinical setting, standardization of patient sampling, and the collection, storage, and stability of the samples.

		Evaluating the assay during routine use

		Identification of batch-to-batch issues.

	If stable labeled material is used as a surrogate reference standard, Bennett says, additional experiments must be performed immediately before each analytical run, to demonstrate similar responses for equimolar amounts of labeled and unlabeled analyte. Because biomarkers are endogenous and the goal is to measure the change in the biomarker concentration in response to some other change (e.g., use in diseased vs. healthy patients or in comparing different therapies), he says, experiments must also be run to determine the ability to detect a statistically meaningful change.
	
	


	While it might seem obvious, Bennett stresses that, even before assay development starts, it is crucial to ensure that the correct biomarker(s) is being evaluated for the right endpoint(s). Brys adds that selected biomarkers should have a strong biological rationale and correlation with efficacy in preclinical studies and, thus, focusing on a few biomarkers is more effective. “A basic starting point for the development of any biomarker assay is an understanding of relevant precedents for the assay,” agrees Donna Eash, director of client services with Frontage.
	
	As important is ensuring that the extent of qualification or validation is aligned with the objective of the biomarker data that is obtained, according to Bennett. “If the data are used for dose selection, formulation comparisons, or research, then compliance may not be as important as ensuring that the method meets an appropriate acceptance criteria. However, if the data are to be used for patient enrollment, safety, or efficacy, meeting required regulatory compliance standards also should be included in the strategy,” he observes.
	Developers must also outline an effective development plan, considering that endogenous biomarkers are often identified in discovery stages. “A good strategy must be in place for initial discovery into preclinical, followed by preclinical into clinical, as each stage will have its own level of rigor, and follow the fit-for-purpose assay qualification and validation steps,” Eash notes.

	Given the low levels of biomarkers present in most samples, Brys adds, it is important to identify an analytical method with sufficient sensitivity. Access to actual study or patient samples can be important during the pre-qualification/validation stage to ensure that the right concentration range(s) are being evaluated and to determine differences in the concentrations of the biomarker(s) in treated and untreated samples, according to Bennett. He also reiterates that validation experiments should be performed to quantify any matrix effects and non-specific binding that may occur when using modified or surrogate matrices.

	The following are also important to the successful development and validation of biomarker assays, Brys says:

		Evaluating biomarker stability, particularly during storage

		Proper determination and evaluation of sample collection and storage methods

		For biomarker assays that are intended for use in diseased patients, the validation process must also include screening patients to evaluate how the biomarker is affected by the disease state and how the disease state will affect the analysis, according to Pasas-Farmer.


	New advances in analytical technology, such as multiplexing systems that allow for 10–100 biomarkers to be analyzed in a single sample, offer a number of benefits, including lower costs and fewer invasive procedures, says Pasas-Farmer.
	
	High resolution/accurate mass (HR/AM) instruments are now commercially available with triple quadrupole-like sensitivity and robustness. Bennett says that they offer significantly higher resolving power with a much greater signal-to-noise ratio, even when using single-ion monitoring (SIM). SIM is advantageous because it does not require the fragmentation and subsequent loss of signal that occurs in LC-MS/MS experiments. This means that, for many analytes, particularly those with poor or excessive fragmentation or high background noise levels, even when MS/MS is used, it allows for greater selectivity and sensitivity, he says.  Examples would include bile, which has poor fragmentation, and steroids, with excessive fragmentation.

	Easier and more robust nanospray technologies have also allowed researchers to achieve significantly better LLOQs than was possible with high-flow liquid chromatography and traditional electrospray ionization techniques, Bennett says. Pairing this technology with either HR/AM or triple quadrupole instruments can provide sensitivity benefits that are further magnified when multidimensional chromatography and an immunoaffinity cleanup step are used, he adds.


	Regulatory expectations for biomarker methods are changing, and Bennett expects them to continue to evolve over the next three to five years. FDA’s recent guidance suggests that the agency will take a more rigorous and less semi-quantitative approach to overseeing assay qualification and validation. In addition, Incurred Sample Reanalysis, which is used to measure the quality of data and robustness of an assay and has been required for small- and large-molecule drug substances, is now being requested for biomarker development, Pasas Farmer points out.


	Cynthia A. Challener, PhD, is a contributing editor at 

 When referring to this article, please cite it as C. Challener, “Developing and Validating Assays for Small-Molecule Biomarkers,” 

Working with biological matrices and understanding the intended use are crucial.

Developing and Validating Assays for Small-Molecule Biomarkers

	Single-use systems (SUS) have become established in biopharmaceutical drug development. As confidence in the technology has increased, the adoption of disposable equipment for commercial manufacturing has begun to take place. With this shift has come a dependence on suppliers that is related to production capabilities, a situation that did not previously exist. This reliance on suppliers translates to increased risk, and consequently biopharmaceutical manufacturers are focused on managing this risk and ensuring security of supply. Dual sourcing of single-use systems and components was initially proposed as the key solution. After further consideration of the issues surrounding the supply of single-use systems, however, many in the industry believe that dual sourcing presents many challenges that can often be avoided by taking alternative approaches.
	
	


	Today, single-use systems are used throughout the entire process chain from upstream to downstream and from bench scale to production scale. In the past, when drugs were manufactured in stainless steel, the manufacturer was in control of the infrastructure once its facility was in place, sourcing only critical disposable components such as filters and chromatography sorbents. Although disposable systems do provide numerous advantages, such as reduced contamination risks, shorter set-up times, and lower capital costs, biopharmaceutical manufacturers do not have that same level of control over the disposable systems as they did with traditional stainless-steel equipment because the systems and components must be purchased from a supplier, according to John Briggs, director of quality, regulatory, and compliance at ASI-Life Sciences. As a result, adds Mario Philips, vice-president of single-use technologies at Pall Life Sciences, there is a greater dependence on the single-use supplier, which increases the risk for the manufacturer and thus requires greater vendor transparency and supply-chain security.
	
	“Unlike with traditional equipment, where almost everything is physically located at the facility, single-use systems are delivered to the site from a supplier. If the user operates on a just-in-time basis with no inventory, any delay in delivery could put the manufacturing schedule at risk. Consequently, an SUS is a critical part of drug substance manufacturing, and there is a greater need to understand and secure the sourcing of these systems,” says Roman Rodriguez, global market development manager for single-use technologies with EMD Millipore. As such, it is logical to consider qualifying a back-up supplier as part of the SUS risk-management system, and dual sourcing is a response to a need for anticipating unforeseen issues and scenarios.
	
	The implementation of single-use systems in commercial manufacturing has had the additional consequence of shifting the purchasing of these systems from those in development to sourcing professionals. As a result, according to Tony White, a founding director of the BioPhorum Operations Group (BPOG), some biopharmaceutical manufacturers are now starting to consider single-use technologies as a strategic commodity and managing their purchases across their manufacturing network, which often leads to the identification of risks that were not previously perceived. “In addition,” says White, “they are realizing that they have been single-sourcing key components by default--a position they do not find very comfortable.”
	
	Past experience with SUS sourcing problems is also a driver for the increasing focus on multisourcing, according to Terry Hudson, director of process development engineering with Genentech. “At any conference with single-use as a topic, you can find a case study or conversation about a single-use component being changed with either no prior notification or inadequate notification to the end user, and in several cases that has been tied to a negative impact on production,” he explains. In addition, Hudson notes that companies are also learning that occasionally a given process may be more sensitive to a leachable or film type than other processes, and it is beneficial to have an alternative film that can be used with that process.  
	
	There are regulatory drivers as well. Regulatory authorities have had a clear and simple message with respect to single-use products, according to Ross Acucena, product strategy manager-ReadyToProcess with GE Healthcare Life Sciences, which is that their use is analogous to outsourcing crucial elements of GMP operations, such as presterilization. “The role of the supplier is therefore now more critical, and regulators expect drug manufacturers to mitigate any risk with close partnerships and thorough knowledge of their suppliers’ quality systems and supply chains,” he says. Furthermore, the increased focus on suppliers has highlighted the fact that SUS supply chains are both broad and deep because of the integration of many components into the end products. “Single-use customers recognize that it is difficult to fully control the supply chain, which presents a potential risk to their operations and subsequent supply of needed drugs to patients,” Acucena asserts. The industry response has been to increase its focus on mitigating such risk through multisourcing.  
	
	


	The ultimate goal of multisourcing is to have the ability to seamlessly transfer from one single-use product source to another without risk to patient safety, according to Acucena. “There is a strong need to characterize and understand this issue for all parties in the complex supply chains that exist within the industry. Upon investigation, it may materialize that some parties have a robust supply chain where risk is small. The converse is that unacceptably high levels of risk are detected, requiring either dual-source or multisource strategies, or even exiting a specific relationship,” he observes.
	
	In an ideal state, according to Hudson, the multisourced component would be supplied by two different companies with different supply chains to minimize the potential for a single change impacting both components. At a minimum, multisourcing needs to provide assurance that a change or disruption to one source will not also occur at the second source. Thus, risk mitigation can be achieved by using one vendor that has two plants or by using two separate vendors. “The latter option also provides some commercial leverage and tends to be the preferred approach for end users,” Philips says, “but a lack of industry standards for single-use systems makes it more difficult, if not impossible.” Not surprisingly, suppliers prefer the dual site or manufacturing redundancy concept, according to Spencer Parkinson, senior product manager for bioproduction with Thermo Fisher Scientific. “If it is truly assurance of supply that is of interest, then more than one production site addresses this need, while allowing a supplier to provide ‘best-in-class’ products,” he notes.
	
	For Rodriguez, multisourcing is the basic view, and the real issue relates to supply chain understanding, assurance, and control, as well as taking action to reduce inherent risks. “Doing so is not as easy as it seems due to supply chain complexity down to the resin suppliers and the varying approaches that can be taken. It really starts with a close interaction with SUS integrators on a whole supply chain view. An alternative supplier’s products must be assessed with the same science-based qualification process, and if different suppliers adopt slightly different approaches, the end-user must determine what is acceptable,” he says. As an example, Rodriguez points to bioburden levels, which can be tested on every single lot or using a statistical approach. Both methods may be acceptable depending on the end-user’s specifications.
	
	


	There are numerous challenges when considering true dual sourcing, and many relate to the complexity of the single-use supply chain. The SUS business is highly customized, and there are thousands of single-use components, many of which only work with other components made by the same supplier. Of course, not all those components need the same scrutiny depending on their criticality and demand, according to Rodriguez, but due to the huge number of components and suppliers, it is a complex task to track and understand the supply chain to secure and control it. As an integrator, EMD Millipore has prioritized and ranked SUSs based on criteria (i.e., criticality, volume used, lead time, reliability of the suppliers) and customer interaction.
	
	Qualification and validation are, in fact, the critical criteria that must be met to achieve dual sourcing, according to Acucena. “To achieve dual sourcing, one must have documented evidence that a change in input (product contact materials) will not affect the output (product quality, efficacy, safety), and it is ultimately through qualification and validation that such assurance is achieved,” he notes.
	
	ASI-Life Sciences is taking a close look at how to demonstrate functional equivalency of materials obtained from different suppliers. “At this point there is no standard or definition for functional equivalency, but it is necessary to define what constitutes functionally equivalent materials in order to ensure change control,” Briggs observes.

	“The challenge lies in the fact that proprietary plastics/elastomers are used by different suppliers, and thus composition alone cannot be used to determine equivalency, and other characteristics, such as the physical and chemical properties of the single-use product and its performance under process conditions must be evaluated. In addition, the ability to integrate the product into an end-user process and the quality system of the supplier must also be considered,” he comments.
	
	Standardization is frequently raised as another issue for dual sourcing of single-use systems. Philips notes that with limited standards in place, variation in SUS design continues to be significant, and it is much harder to have contingency in supply. Hudson agrees that the main challenge to dual sourcing is the current business model used by most single-use companies, which involves the development of hardware for a SUS that will only work with their single-use components. “This approach for end users is a significant impediment to deploying such systems into a GMP environment. Until companies are willing to standardize enough on their design that they fit more than one single-use assembly, including an assembly from another company, it will be difficult to achieve true dual sourcing,” says Hudson.  

	Parkinson, on the other hand, argues that standardization would lead to a halt in innovation. “Single-use products are not commodities and will not be in the foreseeable future; innovation is critical to improving the technology. While it also leads to the development of different designs and materials by each vendor, improvements would not occur otherwise” he states.
	
	The continued advancement of single-use technology adds to the complexity, however, with new products constantly introduced that require change management. “We are aiming at full process integration by adding supply chain considerations from the lead times associated with single-use items, which requires discussions with the end-user because it is an integral part of technology transfers into manufacturing operations,” Rodriguez says. “Obviously the primary action is to communicate and build a basis of understanding about the supply chain itself and the potential associated risks, which naturally leads to actions to minimize those risks at every level of the supply chain, starting with the single-use integrator. Focusing on the product attributes, quality, and specifications may not be enough, however, as there are other steps involved such as sterilization and transportation (sometimes international) that may impact the product and lead time. Ultimately, therefore, the first challenge is to get a clear understanding of the supply chain, to communicate about it, and then establish an action plan on a case-by-case basis,” he continues.
	
	Thoroughly understanding the supply chain is, however, an intense endeavor that requires significant resources, according to Acucena. He in particular points to the case where two SUS suppliers purchase equipment from the same precursor company, and thus qualifying the two SUS suppliers does not achieve dual sourcing. He also notes that there are significant costs associated with qualifying dual-source suppliers, and the value of such an investment is an important consideration. “In essence,” says Acucena, “dual-sourcing or multisourcing can be likened to an ‘insurance policy’ that is relatively expensive to buy.”
	
	


	While the initial reaction of many biopharmaceutical manufacturers to concerns over single-use supply chain security was to look at dual-sourcing solutions, the focus has since shifted to the broader issue of security of supply and how best to manage it, according to Acucena. “The need is security of supply,” agrees Kevin Ott, executive director of the Bio-Process Systems Alliance (BPSA). He adds that dual sourcing is one possible solution, but there are others with fewer validation challenges, including redundant supply sources from a single vendor and inventory plans, for example. “The question,” he continues, “also becomes complicated by how far back in the supply chain one needs to go. That depends on volume versus demand, structure of the supply chain, etc.”
	
	For example, Ott notes that suppliers of biocontainers that can assemble containers in two locations and carry sufficient inventory of finished units and film to cover film supply disruptions do not necessarily need to be dual sourced to ensure continuity of supply, just perhaps be “dual located.”
	
	“The justification for dual sourcing should depend on the risk and the level of simplicity, and single sourcing is in many cases appropriate as long as the relationship between the supplier and the end user is a partnership,” asserts White. He notes that many automotive manufacturers only single source strategically important materials because the investment in establishing a reliable supplier is significant and not worth doubling.
	
	One alternative to dual sourcing is, according to Philips, to limit the number of contact materials in single-use systems and aim to create interchangeability for any layer (e.g., films, filters, valves, and tubes) that comes in contact with a drug product. Pall Life Sciences has taken a step forward in this direction by vertically integrating its own production processes one step back in the supply chain. The company uses its own in-house extruded film to make the supply chain more robust and is working to mitigate any potential changes from resin suppliers. “While we can’t prevent change, we are prepared to manage that change, and if we fully understand our supply chain, we have the power to support customers through any change that may occur through our supply strategy, which adds up to an increased ability to offer a greater level of supply-chain security and continuity to users,” remarks Philips.
	
	Business continuity planning and inventory management are additional mechanisms for reducing risk, according to Acucena. He notes that suppliers are now engaged in an increased focus on these aspects, and thus the crucial first step of increasing industry awareness is already taking place. He also notes that the experience with reducing the risk associated with the single-sourcing of critical chromatography media, such as protein A, can be translated to single-use systems. “GE Healthcare sought to satisfy the industry’s need to mitigate this risk through a strong focus on business continuity planning (in accordance with ISO 22301 and achieving conformance to this standard), including qualifying dual sources for critical raw materials and a significant investment in safety stock. These efforts have delivered supply-chain security improvements that should help to eliminate end-user dual sourcing. Similar approaches should be applicable to single-use technology,” he explains.
	
	Both White and Ott also note that focusing on dual sourcing at this stage in the industry’s development may not be appropriate. “Many biopharmaceutical manufacturers are not yet in a good position to dual source, as they are just starting to understand the extent of their complete single-use portfolios across their networks,” White says. Ott adds that it may be premature or counterproductive to imagine risks. “It is better to actually study, evaluate, and understand your own supply chain and the real risks. Once that is done, the need to establish truly independent dual sources and the difficulty that such an approach represents with respect to validation and regulatory filing begins to look like much less of an issue than might be imagined,” he comments.
	
	


	Supply-chain security is a top challenge for both BPSA and BPOG moving forward into 2015. The Parenteral Drug Association (PDA) is also working on concepts to facilitate multisourcing. While these initiatives are important, Philips would prefer to see vendors, users, and regulators working together to bring one set of standards forward, rather than multiple associations working on multiple standards. “The biopharmaceutical industry is still a relatively immature industry, and substantial changes can be expected over the coming decade. Standardization is going to be a key element of true multisourcing, and having one body regulating the industry will make supply chains more transparent and keep user expectations realistic,” he says.
	
	Many of the industry-wide initiatives already underway to provide standardization and guidance on change notification, leachable/extractable testing, and other end-user needs for increased adoption of single-use systems, and particularly for components and systems that already allow some interchangeability such as buffer bags and tubing connectors, will help speed up multisourcing, according to Hudson. On the other hand, where the supplier’s business model is to design a system that only works with its components, education is the key. “Genentech recently presented a proof-of-concept design for a single-use bioreactor that could work with top- and bottom-mounted bags from two different suppliers. The next step is convincing management at suppliers that designs such as this one will lead to higher future sales because they help open the doors to more GMP applications and greater adoption of single-use technology in general,” he states. Genentech is also partnering with several single-use supplier companies to gain greater visibility into their supply chains, to provide recommendations on what it considers “multisourced”, and to communicate what applications of single-use simply will not occur until they can be multisourced.
	
	EMD Millipore, meanwhile, in addition to implementing a business continuity plan and supply risk management and emergency management programs focusing on products, suppliers, and its sites, respectively, has teamed with a third party (Resilinc Corporation) to map its supply chain for key products and suppliers to identify and mitigate risks with suppliers, according to Rodriguez.
	
	In its quest to identify a practical approach for determining the functional equivalency of unassembled components, ASI-Life Sciences submitted platinum-cured silicone tubing with similar specifications, mechanical properties, and performance profiles under typical use conditions obtained from two suppliers to Chemic Laboratories for comparative analysis of their extractable/leachable profiles. “Tubing was selected for this exercise because it is a common component of single-use systems and can be evaluated in a fairly straightforward manner,” Briggs observes. Importantly, no significant differences in the extraction profiles were observed for the two tubing samples. “These results indicated that the materials were comparable within the variability of the analytical techniques used,” Briggs says. He also recognizes that tubing is a universal single-use component, and establishing the functional equivalency of other, more proprietary single-use systems and components will be more challenging. “Our goal is to initiate a dialogue around the practical considerations for achieving dual sourcing of single-use systems.” 


	Vol. 38, Issue 12
	Pages: 24–26, 57
	Citation: When referring to this article, please cite it as C. Challener, "Securing the Single-Use Supply Chain," 