	As part of its ongoing support for the development and approval of competitive biotech therapies, FDA has updated and clarified standards and procedures for ensuring the quality and similarity of biosimilars. A new draft guidance recommends a process for developing comparative analytical assessment plans to demonstrate biosimilarity, with added details on procedures for conducting such assessments and for documenting quality in a range of production lots to make biosimilar development and production more predictable and efficient (1).

	The new advisory replaces an earlier guidance on quality considerations for biosimilars issued in 2015 and updates a 2017 draft guidance on statistical approaches, both withdrawn later by FDA due to industry concerns and objections. This revised document presents strategies for conducting comparative analytical studies to assess biosimilarity to a reference product and for presenting scientific and technical information in the chemistry, manufacturing, and controls (CMC) section of biosimilar applications. There’s more detail on how comparative analytical assessments should cover key processes, including a product’s expression system, manufacturing process, physicochemical properties, functional activities, target binding, impurities, reference product and reference standards, finished drug product, and stability.

	To address concerns about lot-to-lot variability, FDA advises biosimilar makers to include data from at least 10 reference product lots acquired over several years to fully assess reference product drift. Similarly, sponsors should assess 6–10 lots of the proposed biosimilar, including both investigational and commercial-scale lots, validation lots, and lots manufactured at difference scales. Additional sections clarify the submission of reference standards and data on functional activity to demonstrate any structural differences with a reference product. And FDA provides the usual caveat that it will consider alternative approaches from sponsors in analyzing and presenting requested data. 

	FDA issued the revised quality assurance advisory soon after publication of a final guidance on developing interchangeable biosimilars, a much-anticipated and hotly debated option for manufacturers (2). That document updates a draft guidance from January 2017 that attracted voluminous comment from multiple stakeholders. The new version instructs manufacturers on conducting switching studies, which FDA says it usually will require to approve an interchangeable label. While some firms consider extensive switching studies unnecessary, FDA justifies this approach as important for demonstrating fully that safety and efficacy are not affected by changing treatment from brand to interchangeable, a finding that is considered important to gain the trust of physicians and patients in pharmacy-level switching.

	Important for biosimilar makers, the new guidance supports the use of non-US-licensed comparator products to conduct tests and to produce the data needed to demonstrate interchangeability. Even though FDA wants data from bridging studies to foreign-made reference products, manufacturers still may benefit from leeway to obtain samples of biologics that often are less costly outside the United States and can help reduce the cost and time for developing interchangeable therapies. Sponsors developing interchangeable products may be eligible to gain a year of market exclusivity for being the first to develop such a copycat therapy.

	The new guidance also appears less didactic by dropping the frequent use of terms such as “residual uncertainty” and “fingerprint-like” similarity between reference and interchangeable products, which were common in the earlier advisories. FDA here aims to set general standards and requirements for these therapies but notes, however, that the specific data and testing needed to document interchangeability may vary with the structural and functional complexity of the product and with clinical experience with the reference product. Sponsors may seek to justify an exemption from switching studies, and FDA says it will take a flexible approach. 

	In moving forward on standards for interchangeable products, FDA sets the stage for the development and approval of competitive insulin products, which are scheduled to transition to biologic status in March 2020. There is great anticipation that interchangeable insulin therapies will provide important alternatives to current costly diabetes treatments as multiple sponsors move into this field. These issues were discussed at an FDA public meeting in May on the future of biosimilar insulins, just days after the appearance of the interchangeable biologics guidance. As a relatively simple biotech product that has been available as a drug for 100 years (but ineligible for generic competition), manufacturers expect to be able to utilize a rather streamlined and straight-forward process for documenting similarity and interchangeability for these widely used therapies. Firms planning to produce competitive insulin therapies propose that FDA require data from only one, small immunogenicity study, as opposed to multiple switching studies, and that analytic characterization, bridging studies, and reliance on pharmacokinetic and pharmacodynamic data should provide ample support for interchangeable insulins.

	FDA says it plans additional guidance on data needed on product container closure systems and delivery device constituent parts to support the presentation of a proposed interchangeable product. Of interest is how a biosimilar or interchangeable determination may be affected by delivery of insulin through a pump or over-the-counter device.

	So far FDA has not approved any biosimilars as interchangeable, but this is expected to change with this added clarification of regulatory process and requirements. The rising price of insulin therapies has become one of the most contentious issues in the escalating drug pricing debate, with Congressional hearings generating multiple legislative proposals on this issue, many supporting the development of biosimilar and interchangeable products. 

Development of Therapeutic Protein Biosimilars: Comparative Analytical Assessment and Other Quality-Related Considerations, Guidance for Industry

Considerations in Demonstrating Interchangeability With a Reference Product, Guidance for Industry 

	When referring to this article, please cite it as J. Wechsler, "FDA Revamps Biosimilar Quality Requirements," 

Articles

Regulatory Beat

New guidance documents clarify production standards and processes for developing interchangeable biologic drugs. 

FDA Revamps Biosimilar Quality Requirements

Q. We are preparing for our first inspection by an overseas regulatory agency. This will be a challenge from several perspectives: language, culture, and presentation style. We believe we have covered the first two issues through highly experienced translators and cultural training sessions with coaches. Our concern is about how to present and respond in an appropriate manner. Can you give any advice?

	A. Hosting a foreign agency for the first time is always challenging, but as with all inspections, nothing beats preparation, practice, practice, and more practice. Even seasoned personnel can become flustered, nervous, or even unsure how to respond to unfamiliar questions or requests, especially when asked in a foreign tongue. One tool proven to be particularly useful in such situations is a playbook. Playbooks may be known by other terms, but essentially these are aide-memoires that function as references for company representatives who need to present or answer questions from inspectors or auditors. The playbook may include prepared answers to anticipated questions the auditor may ask and/or detailed information about process steps.

	How detailed should these playbooks be and in what format? Here is what has stood the test of time:

		Always start with a succinct, logical, and convincing few lines of text and/or graphics.

		If necessary, have a second playbook with more detailed, supporting information.

		Have a third party (i.e., someone unfamiliar) review your playbook and/or provide feedback.

	The first point is the most important. As we all are experts in what we do, we can talk in great detail about it. But inspectors do not have unlimited time. They want to receive answers that provide them with clear and easy-to-comprehend information. This is best illustrated with an example. The inspector may ask: ‘What is your bioburden control strategy for this product?’ Without a playbook, the answer may go like this: ‘In step 5 after vessel 234 on the second floor, remember you saw this reactor on the tour of the facility. We have installed a 0.2-micron filter …’

	With a playbook, the answer would be more similar to this: ‘We have a bioburden control strategy that takes into account environment, process equipment, material handling, product properties, and regulatory expectations. For the five steps produced in our facility, we have prepared a strategy paper, which has subsequently been verified during the validation studies and ongoing monitoring.’

	Of course, should the inspector then want to know more details, this is when all the evidence can be presented. Again, it may be useful to have a playbook ready for this, especially if many activities should be organized in a logical sequence and timeline.

	Playbooks have several benefits, as they help those responding in an inspection to give clear answers. They also help prepare the site personnel to find answers to difficult questions (e.g., why certain deviation investigations are overdue, or why a particular change hasn’t been documented as per requirements). They also help summarize convoluted event histories or complex situations. Often, one has to overcome obstacles to reach a successful outcome, and playbooks help to focus on the positives, rather than to overemphasize the difficulties experienced.

	Then why should one have these playbooks reviewed by someone from the outside? Well, the inspector will be exactly such a person (i.e., an outsider who isn’t necessarily familiar with the terminology, the facility, or the processes used at the site). Therefore, by having an outside third party review the playbook, you essentially perform a dry run of the inspection.

	In summary, these playbooks are a useful tool in any inspection or audit, whether foreign or local, but writing these in a clear and concise manner is anything but child’s play.

	When referring to this article, please cite it as S. Schmitt, "Playbooks Are Not Just Child’s Play," 

Ask the Expert

Cultural and language discrepancies during an audit can be resolved using what many call a “playbook,” says Siegfried Schmitt, PhD, vice-president, technical, Parexel Consulting.


Playbooks Are Not Just Child’s Play

The concept of quality by design (QbD) has been part of the biopharmaceutical industry for more than a decade. While similar approaches to QbD are used in the development and manufacturing of large-molecule and small-molecule drugs, the complex nature of large-molecule drugs creates more factors that may affect process performances and the critical quality attributes (CQAs) of the drug.

	Because biomanufacturing involves living cells, there are more reactions happening in a bioreactor than usually happen in chemical reactions, according to Bill Whitford, strategic solutions leader, bioprocess, GE Healthcare. “While we often understand nearly everything about a small-molecule reactor development pathway, we have only a partial understanding of all the interacting salvage, degradation, and biosynthetic pathways of mammalian cells. This includes even many relevant metabolic pathways, inductions, flux, rates, and turnover.  Finally, we often develop and control bioprocesses using very surrogate key quality indicator (KQI) and critical process parameters (CPPs). Rather than measuring precisely what we know (or suppose) to be relevant, we monitor such general parameters as pH, oxygen, trypan-blue exclusion, and such token or generally representative metabolites as glucose or lactate,” says Whitford.

	According to the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ Consortium), a collective of more than 35 biopharmaceutical companies, there are also differences between the use of QbD for large molecules versus small molecules when considering antibody-drug conjugates (ADCs). ADCs are complex molecules composed of an antibody linked to a cytotoxic small-molecule drug. As noted by the IQ Consortium, companies that develop ADCs state that CPPs and CQAs may also differ between the drug-linker (DL) component of the ADC and the related small-molecule drug because the DL represents a small portion of the total mass of the ADC. Unlike small-molecule drugs, the dosing frequency and scheme of the ADCs also may dilute the impurities in DL intermediates. 

	Another difference between small-molecule and large-molecule drugs, according to the IQ Consortium, is that in early development, a platform process approach may be applied, which may accelerate timelines and reduce development burden.

	QbD is used in cell-culture expansion, cell harvesting, and cell-culture media development. However, according to the IQ Consortium, many companies find that clone selection and the development of the production-stage bioreactor process benefit most from the application of QbD due to the following: “In clone selection, the final producer cell line is defined, which is then used for the commercial process and throughout the lifecycle of a product, while the N-stage production bioreactor is considered to be the most critical step with respect to the CQAs and essentially determines the final product quality from an upstream perspective. In principle, all upstream process steps can benefit from a QbD approach and examples where companies apply those also depend on the assessment of (business) risks,” a representative of the IQ Consortium says. Predefining development goals for titer ranges and product quality profiles, as well as assessing risk, also benefit from QbD.

	The development of robust processes that include good cell culture performance can be done through design of experiments (DoE). In early stage development, QbD can be used to identify a quality target product profile (QTPP), according to the IQ Consortium. Late-stage applications include process characterization studies to decide on one-factor-at-a-time (OFAT) experiments versus DoE-based experiments and the defining and classifying of parameter ranges. “Currently, a claim of a design space is not pursued by most companies. However, achieving in-depth understanding is a goal (e.g., by advanced data analysis approaches). Also, detailed risk assessments can play an important role in late-stage activities. Commonly, the knowledge gained in process characterization studies (PCS) will be included in the control strategy,” the IQ Consortium states.

	Performing small-scale DoE by applying verified scale factors and using a risk-based approach are key to implementing QbD in upstream processing, says Whitford. “The final product will be a verified design space, clearly communicated-being mindful of using accepted terms and definitions-to regulators in the submission.  Currently, upstream process development focuses on digitally transformed and intensified biomanufacturing,” Whitford states.

	Bioreactor process development has been positively impacted by the use of QbD and empowered by process analytical technology (PAT), according to Whitford, by enabling continuous monitoring of process parameters and product characteristics throughout development and manufacturing. “In the pursuit of process intensification, the influence of altering process parameters on product characteristics can be rapidly assessed.  This contributes to increasing an understanding of what types and ranges of bioreactor parameters provide acceptable product quality. This mapping of the relationship between bioreactor process characteristics and product CQAs will define a bioreactor design space, an important element in both ensuring a robust process as well as current FDA expectations in filings,” says Whitford.

	Using QbD to develop cell culture processes

	Because cell culture has a direct impact on production efficiency and product quality, using DoE can help determine the relationship between cell-culture process parameters and product quality, which then helps determine the design of the operation space for a process. DoE can also help determine the needed sensor technologies to establish robust closed-loop model predictive controls, according to Whitford. “Unlike many other production processes, biomanufacturing has classically been hindered by a lack of real-time product attribute measurement. Even in nominally identical bioprocesses, we can see diversity in product quality due to variability in (e.g., such raw materials as cell culture media). Using a QbD/PAT strategy directs the Â­implementation of advanced near-real time monitoring through the application of newer, automated sampling techniques and multivariate data analysis. This supports advanced

 modeling that can not only help control the process, but predict product quantity and quality outcomes,” says Whitford. 

	Implementing a feedback quality control strategy, as opposed to a recipe-based control strategy, when using QbD in developing cell culture processes may be the best approach, says Brandon Downey, principal engineer, R&D at Lonza, because of the challenges cell-culture processes present. “Many [cell culture] processes have successfully used a recipe-based control strategy, determined using a design space approach, to manufacture quality product. Nevertheless, cell culture processes pose some unique challenges when attempting to employ a recipe-based control strategy,” says Downey. And, the large number of raw materials used in cell culture can make understanding the variation of each component difficult. “Although fundamental understanding of how variations in common raw material attributes impact a cell culture process is increasing, this knowledge is still far from complete enough to encompass the variation that has been observed in the cell culture media,” he states.

	The behavior of the cull culture process may also be impacted by variations in unmeasured impurities in raw materials, says Downey, making it difficult to determine the total impact of raw material variation on the cell culture process. “Finally, the cells in the production cell culture step uniquely determine many of the structural attributes of biologic molecules which could (and in some cases have been clinically shown to) influence the function of the molecule.”

	When it comes to cell line selection, QbD can be used in the selection of a final clone during early-stage development by using prior knowledge to predefine clone selection criteria including cell-line stability, culture performance, and product quality attributes. “As the selection process progresses, more clone-specific knowledge is gained, and the process continues until the clone that best fulfills the selection criteria is identified,” according to the IQ Consortium.

	Whitford stresses that even though many bioprocessing platforms have been around for decades, individual process optimization is beneficial to individual cell strains, recombination techniques, cloning approaches, and culture systems. Continuous monitoring is being enabled by PAT-derived systems, says Whitford. In addition, highly accelerated bioprocess development is supported by new fully-automated, single-use multi-parallel bioreactors. “Employing some of the process monitoring and design approaches … in the context of high throughput and automated systems for process development is now very popular.  For example, some vendors offer as many as 24 fully featured, single-use, 250 mL mini-bioreactors supporting process development and optimization as well as selection or development in scale-down studies,” says Whitford. 

	Medium selection, however, varies based on practices and needs, according to the IQ Consortium. “For example, companies that employ a platform process approach, including a platform cell culture media system, typically do not require media selection for molecules that fit the platform process. However, for atypical modalities that do not fit the predefined selection criteria, media screening can be done to select the condition which yields attributes that best fit the criteria,” says the IQ Consortium.

	Using a QbD approach to define objectives for establishing a new Â­manufacturing facility may be beneficial. Risk assessments can be used to evaluate different equipment against the predefined objectives. The knowledge gained through this process can show how different equipment may impact the objectives, according to the IQ Consortium.

	QbD can also be applied to the evaluation of equipment performance. The suitability of equipment can be established through experimental designs, says Whitford. “This is generally accomplished through either limited experimentation at scale, or through studies in scale-down models, employing prior understanding of the scale-factors involved in extrapolating results to full-scale manufacturing.  Premier single-use system vendors have for years been supplying much information to aid in such studies. One example here is CFD data on single-use mixers and various scales, fill volumes, impeller speeds, and viscosity of diluents.  We have seen such techniques applied to the selection of single-use bioreactors and mixers, perfusion apparatus and even dynamically feed-back controlled inline buffer conditioning systems.”

	When it comes to late-stage development, the IQ Consortium agrees that small-scale models (SSMs) are essential to QbD when scale-down models must be qualified to show they represent large-scale operations. “SSMs are used in PCS to understand the process and determine parameter ranges and classifications. Given that the PCS experiments are often large and complex, it is important to balance between representativeness and throughput of the SSMs,” according to the IQ Consortium.

	Designing quality into the biopharmaceutical process early on can be difficult because the nature of the protein may not be fully understood, says Whitford. There is also a range of variants, such as glycosylation in monoclonal antibodies (mAbs), that occurs when biologics are generated, which can affect both biological activity and/or the secondary or tertiary structure of the molecule. “This can influence not only the entity’s activity but its stability in storage,” says Whitford.  “Secondly, we don’t yet fully understand the conditions of cell culture that will control those molecular properties. Often, they are discovered more empirically than by pathway and structure understanding. Finally, we are only beginning to implement comprehensive monitoring of the bioreactor process, metabolites, and products that can provide robust specific, relevant, and predictive process parameter values and product quality assessment,” Whitford adds.

	A robust cell culture platform process can ensure that CQA profiles are consistent from the early stages of development through commercialization, according to the IQ Consortium. It may be difficult, however, for companies without a platform process in place to incorporate quality in early-stage development without performing additional experimentation to determine the robustness of their process. This experimentation can be done using a QbD approach and DoE methodology, says the IQ Consortium, when “molecules cannot conform to the platform.”

	This work, however, could have an impact on investigational new drug application submission timelines. “There is a limitation to how much quality can be designed into the process during early-stage development, and in some cases, target quality characteristics may not be clear. Some companies question what effort associated with this should be taken early in development. For example, early-stage CQAs are more focused on the safety aspect of the drugs and as clinical data and process experience are gathered over the course of the project, this informs the efficacy aspect thus modifying the CQAs in late-stage development,” according to the IQ Consortium. They also warn that using in-licensing molecules may present unanticipated challenges. “This might, depending on the development stage of the in-licensed molecule, shift focus from ‘design’ to comparability.”

	The implementation of QbD into the bio/pharmaceutical industry aligns with regulatory and industry efforts to ensure the efficacy, safety, and quality of medications. The complex nature of biologics adds additional complexity to developing and manufacturing these innovative therapies. Using a QbD approach from early-stage development through commercialization can ensure that upstream processes are efficient and reliable.

	“As biologics increasingly take the form of novel modalities (e.g., not mAbs), it will become increasingly important to be able to control the quality of biologics so that the structure-functional relationships of these new classes of compounds can be better understood in the clinic. Furthermore, as medicines become more personalized, the ability to tailor the structural attributes of biologics molecules that are dictated by the process will become increasingly important. QbD will likely continue to play a foundational role in creating the processes which can robustly deliver these compounds,” concludes Downey.


	Vol. 32, No. 7
	July 2019
	Pages: 17-22

	When referring to this article, please cite it as S. Haigney, "Applying QbD to Upstream Processing," 

Features

Using a QbD approach from early-stage development through commercialization can ensure that upstream processes are efficient and reliable.

Applying QbD to Upstream Processing

	A bipartisan effort by the Senate Health, Education, Labor and Pensions (HELP) committee designed to lower healthcare and medication costs has led to a contentious debate between FDA and standards-setting organization, the US Pharmacopeial Convention (USP). The committee voted 20–3 on June 26, 2019 to advance legislation with proposals that include changes to patent policy, ending “surprises” in medical bills, increasing transparency in healthcare, and reducing roadblocks that delay generic drugs and biosimilars getting to market (1).

	It is provisions in Section 207-designated as promoting biological drug innovation-that are proving controversial, pitting FDA and USP in an exchange of policy statements. According to HELP committee documents, Section 207 is intended to prevent delays in the licensure of biosimilar and interchangeable products by excluding biological products subject to regulation under the Public Health Service Act from requirements to follow 

compendial standards. A similar provision to make biologics exempt from meeting USP standards was removed from the final 21st Century Cures Act of 2016 (2).

	“A biological product is so inherently complex and variable that the established structure of the 

 monograph standards process does not serve it well, and in fact, can impede technological progress or innovation,” noted Steven Kiozlowski, director of the Office of Biotechnology Products in FDA’s Office of Pharmaceutical Quality in an interview published by FDA (3).

	Due to the inherent differences in biological products, the “sameness” standard used in 

monographs for chemical-based drugs cannot apply to biologic drugs, FDA argues.

	USP, backed by patient groups and pharmacists, argues that the elimination of required standards would harm patients. In a letter to the committee, USP noted the role standards play in providing quality benchmarks; supply chain security; reliability and predictability for drug product development, manufacturing, and distribution; and promoting transparency and accountability that leads to patient trust (4).

	“Essential to the framework that safeguards the quality and safety of medicines in the United States is the principle that public quality standards, required under the law, establish and articulate quality expectations for medicines, including biologics,” USP noted in the letter. 

	FDA has no legal authority to control the price of drugs. However, the agency notes it is “committed to facilitating increased competition in the market for prescription drugs through the approval of lower-cost, generic medicines” and “makes sure that safe and effective drugs are available to improve the health of consumers” (5). That’s a lot of priorities to balance in a high-stakes, highly partisan environment.

, US Senate, 116th Congress, 1st Session (Washington, DC).
	2. J. Wechsler, 

“FDA Challenges USP Standards for Biologics,”

CDER Conversation: Ensuring That Standardization Does Not Impede Biological Product Innovation

Letter to Senate Committee on Health, Education, Labor and Pensions (HELP)

	When referring to this article, please cite it as R. Peters, “A Biologics Partisan Divide," 

From the Editor

FDA and USP take sides in debate on biologic drug standards.

A Biologics Partisan Divide

Continuous bioprocessing provides numerous benefits, from high product quality and consistency to smaller physical footprints, increased flexibility, and potentially lower capital and operating expenses. It also provides the greatest benefits if real-time monitoring of process parameters and product quality, including detection and quantitation of residual impurities, is performed. While advances in analytical methods and automation are occurring, practical implementation of real-time monitoring of product- and process-related impurities has yet to be achieved.

	“In an ideal world,” asserts Byron Kneller, senior director for analytical and formulation development with AGC Biologics, “we would have simple, rapid, physicochemical tests for residual impurities.” 

	Unfortunately, there are different types of residual impurities. They can be classified as product- and process-related, and as chemical impurities (typically small-molecule process additives such as isopropyl β-D-1-thiogalactopyranoside, antifoams, antibiotics, etc.) or biological impurities from the host cell or processing steps (host-cell proteins [HCP], host-cell DNA, residual Protein-A, etc.).

	“Technologies that utilize mass spectrometry (MS) capabilities are, at this point, the seemingly best choice for residual impurity detection, characterization, and/or quantitation when coupled with liquid chromatography (LC) systems,” observes Amit Katiyar, director of analytical and formulation sciences for Thermo Fisher Scientific. Multi-attribute methods (MAM) capable of detecting, quantifying, and characterizing multiple product quality attributes are currently being developed. 

	A triple-quadrupole system would be best suited for quantitative analysis of small molecules/peptides, with quadrupole time-of-flight systems more appropriate for the characterization and semi-quantitative analysis of larger molecules, according to Charles Heise, senior staff scientist for bioprocess strategy and development at Fujifilm Diosynth Biotechnologies.

	An amalgamation of both technologies for a single quantitative measurement covering a wide mass range would be an ultimate solution, he adds. A series of standards could be used to generate standard curves for target impurities and the software needed for identification and quantification of peaks. Aspects of high-throughput proteomics analysis may also be applicable to the continuous manufacturing environment.

	Other analytical techniques Heise highlights include in-line spectroscopic measurements such as Raman, Fourier-transform infrared (FTIR), ultraviolet/visible (UV/Vis), and refractive index methods that can give quantitative as well as qualitative, real-time analysis. “The particular technology used will depend on the impurity, its expected concentration range, and whether monitoring is performed during steady-state conditions,” he observes. New methods or advances in current technologies may also become available that allow in-line measurements using nuclear magnetic resonance (NMR) imaging, sensor arrays, or LC systems.

	The final option, Heise says, is to do no process monitoring and rely solely on the quality-by-design data generated during process development to define the operating ranges and edges of failure.

	“In practice, the choice of method may be constrained by the availability of the instrumentation, turn-around time, analyst training, and potential interference from the complex matrices that may be involved in the downstream purification,” comments Ian Parsons, director of analytical development for biologics at Charles River Laboratories. His company has focused primarily on the use of high-performance LC (HPLC) and the enzyme-linked immunosorbent assay (ELISA), with MS and gas chromatography (GC) also used frequently. 

	“The pharmaceutical industry is still heavily utilizing traditional separation techniques such as size-exclusion, ion-exchange, and reversed-phase chromatography in addition to immunoassays such as ELISA and electrophoresis as the primary means for quantifying product- and process-related impurities,” notes Katiyar. 

	Often tests for process-related impurities require specialized reagents such as anti-HCP antibodies or natural products like limulus amebocyte lysate and can be more complex, adds Kneller. In addition, many impurities are heterogeneous (e.g., HCP) and thus not amenable to detection and quantitation using simple tests.

	In practice for continuous processes, in-line confounded (e.g., spectroscopy) or secondary parameter (e.g., off-gas, pH, basic physical property) measurements are most user friendly, but accuracy of residual impurity detection is sacrificed for real-time control, as a result of the heterogeneous nature of the signal, or to allow for detection of a secondary response, according to Heise. 

	He also notes that off-line LC–MS, where mass windows can be limited by buffer solution and scanning rates, or LC–evaporative light scattering detection (ELSD), is typically used for small molecules, but these are not sensitive enough and can take too long to run for large-molecule manufacture.

	These approaches could potentially be substituted for sensor technologies with equivalent modalities, according to Heise, such as the Octet system from Fortebio, which relies on biolayer interferometry. Similarly, he says that ELISAs for sensitive quantitation of biological impurities that have slow response times (material quarantining needed) could also be replaced with equivalent ‘dip-and-read’ methods.

	The new technologies (e.g., MAM, MS for HCP analysis, automation for DNA, HCP, and Protein A residuals) are just starting to be introduced in the process development space, according to Katiyar. “The transition from exploratory to fully validated methods will be slower for such methods in a cGMP environment where method performance, equipment validation, and data-integrity measures need to be at the highest level. There will also be a significant cost aspect of not only reconfiguring current laboratories to accommodate the equipment, but also for training and/or hiring new personnel with expertise in these new technologies,” he says. 

	At the process level, identifying the residual impurities present and the appropriate analytical assays, with the required detection limits and sensitivity ranges, to monitor them across the process is critical, stresses Dan Pettit, senior staff scientist for analytical development at Fujifilm Diosynth Biotechnologies. At the operations level, he notes that identifying suitably equipped labs to develop and validate methods for GMP analysis is vital because of the heterogeneity of process impurities and the consequent analytical issues they pose. 

	Heterogeneous process-related impurities, specifically HCPs, are often quantified in early-phase projects using generic kits that may result in key impurities going undetected during purification process development. “These impurities have the potential to cause problems at production scale,” Kneller says. Current ELISA techniques lack the ability to identify specific immunogenic proteins, thereby making the process development aspect less strategic in design, agrees Michael Farris, scientific manager of analytical and formulation sciences with Thermo Fisher Scientific. “Building toxicology and immunogenicity databases for various HCPs and other process impurities would help fine-tune process development and, in itself, provide a deeper insight into process performance and robustness,” he asserts.

	Identifying/developing suitably sensitive methods is also challenging, according to Pettit. “Typically, ppm or ppb levels of analyte in a complex mixture must be detected, and therefore, isolation/derivatization of the analyte may be required. These activities tend to be prohibitively expensive to outsource,” he observes. 

	In addition, isolation of impurities prior to quantitation may introduce artifacts or cause the generation of various altered states during the isolation process. For example, physical manipulation of samples prior to the detection of multiple analytes may alter the HCP antigen profile and relative abundance of the analytes themselves, according to Farris. Additional studies to understand the stability of isolated product(s) may therefore also be needed. 

	Other process impurities such as Long R3 IGF-1, which is a component of some cell-culture processes, have a tendency to adhere to certain plastics, thereby making accurate quantitation more difficult, Farris adds. Sampling instructions must be clearly defined so as to not artificially deplete the analyte during sampling and/or storage prior to analysis. 

	Parsons agrees that the main challenges are focused around developing methods of sufficient sensitivity that also minimize product and matrix-related interference, and thereby accomplish sufficient recovery of the analytes. In numerous instances, Farris points out that the quantitation and/or characterization of impurities is hindered by the API and/or the matrix composition. 

	“Immunoassays, such as ELISA, used for quantitation of host-cell proteins and leachables like Protein A are sensitive to extreme pH, high salt concentrations, and certain detergents. The typical means for overcoming the signal suppression or interference associated with their presence is to dilute the sample, which acts to decrease the sensitivity of the method to maintain acceptable precision and accuracy,” Farris explains. As a consequence, sensitivity constraints imposed by matrix/API interference must be a point of consideration when demonstrating that a method is fit for use for process characterization and/or process validation activities.

	For products with small total batch volumes, the quantities of material required for analytical method development and sample analysis can also present a challenge, Parsons notes. “If method sensitivity is also an issue, an approach can be taken to spike an aliquot of upstream material with a higher known concentration of analyte and then demonstrate fold-removal of the analyte across the downstream purification step(s),” he comments.

	For continuous processes, the turnaround time of typical assays is a key issue, according to Kneller. “The challenge is to increase our testing throughput, decrease our reliance on heterogeneous reagents, and to find ways to either simplify the method types used or engineer robust, easy-to-use systems for more complicated analyses (e.g., mass spectrometry),” he says. 

	Real-time monitoring of representative material for modeling the residence time distribution in continuous processes is also required for process control, according to Pettit. “The range of appropriate analytical sensors for continuous monitoring of product specific critical quality attributes and impurities is limited. Additionally, sensor drift over the operating time, any on-going sensor calibration needs, and system suitability testing for quality critical attribute monitoring during continuous operation must be resolved,” he adds. “Here again, sensor technology could be the way to go (e.g., online Octet dip-and-read-type systems); however, the technology is not sufficiently developed yet. The final technology offerings need to be diverse, robust, accurate, reliable, and rapid,” he concludes.

	For off-line analysis of residuals from a continuous process, Parsons notes that methods for aliquot/sample withdrawal are needed. In addition, the sampling frequency should be sufficient to provide a statistical sampling of the material flowing through the process in order to account for any transient variation that may occur while avoiding significant depletion of the process flow.

	The analytical testing strategy associated with continuous processes can, however, be more demanding in terms of the number of sampling time points if in-line or on-line monitoring is not possible, according to Farris. “Process analytical technologies have generally been geared more toward product quality, cell-culture viability, and growth than active monitoring of process impurities. If a deeper understanding of any particular impurity is required, more complicated assays will likely be required,” he observes.

	Residuals analysis when moving from batch to continuous

	When moving from batch to continuous processing, it is likely that the same analytical methods for residual impurities can be used, but potentially in higher throughput versions. Offline analysis using LC, GC, ELISA, and MS should be possible as long as an appropriate sampling approach can be implemented, Parsons observes. 

	The throughput capability and time-to-result for the analytical method employed are also important aspects if at some point the impurity is deemed process critical, according to Darshini Shah, senior scientist and group lead for downstream process development at Thermo Fisher Scientific. Automation of methods (e.g., liquid-handling or use of systems such as the Octet for HCP analysis) can provide higher throughput both in terms of overall speed and analyst effort, according to Kneller.

	The analytical target profile must also be evaluated to ensure methods are capable of monitoring fluctuating analyte levels throughout continuous processing. “Retention rates or biomass removal may result in periods of correspondingly lower impurity levels, and the analytical method must be sensitive enough and support a large enough dynamic range to be able to actively monitor the process range without the need to constantly repeat analyses to better target the operating range of the assay,” Shah explains. 

	In addition, the typical approach in batch processing involves analysis of a small discrete sample number, with reference standards bracketing the samples and system suitability tests performed prior to analysis, according to Heise. Rotating through multiple identical analyzers would allow continuous monitoring, but the solution would be costly. “Validation of both system suitability tests and the analytical methods for continuous control/monitoring will need additional work to demonstrate that results do not require bracketing with reference standards and that the detection methods do not drift or get poisoned over time when operating for 90+ days,” he says. Failure mode mitigation strategies will also need to be developed, for example with respect to system suitability failure.

	In addition, Parsons points out that for in-line and on-line analysis, the inherent requirements of sensitivity and specificity for analysis of process residuals suggests that many of the current and potential online methods would have limitations and may perhaps be able to best provide supportive general information, rather than specific quantitation of residuals. “Offline analysis using traditional validateable techniques might still be required for specific and sensitive quantitation of the process residuals. Nevertheless, if a continuous process can be shown (perhaps by an online monitoring method) to be robust, then reduced sampling of the continuous process for quantitation of residuals may be justifiable,” he notes.

	In fact, because continuous processes are expected to operate under steady-state conditions, control through predictive modeling or by exception should be possible, Heise adds. “Multi-variant analysis during development may identify simple in-line monitoring techniques that can control the process to ensure residual impurities do not pass through the whole process. However, the dynamics of the process will be different at the outset until the steady state is reached. A testing strategy defining the frequency of measurement pre- and post-steady state will be required to identify when process equilibrium has been reached and when it begins to fail,” he observes.

	“Ideally,” Heise continues, “we don’t really want to be testing anything real-time and on-line. Instead, we want to have confidence that the process will deliver consistently over a defined time scale.” 

	Potential for on-line/in-line analysis of residual impurities

	Significant development is still needed before on-line/in-line analysis of residual impurities will be widely implemented. Shah is not aware of any on-line/in-line tools capable of monitoring residual impurities for either quantitation or characterization. “Offline analysis is the predominant means of analysis and the only on-line/in-line tools are related to product quality aspects,” he says. 

	Current solutions exist around automated sampling for at-line analytics, where process changes occur over a longer time span than the analysis time, such as in the case of mammalian bioreactor metabolite analysis, according to Pettit. 

	On-line process analytical techniques such as near-infrared spectroscopy (NIR) or mid-infrared spectroscopy (mid-IR) and UV may provide some general information on clearance of residuals but are unlikely to be specific or sensitive enough to enable detailed quantitation of analytes, according to Parsons. 

	Parsons does note that low-field NMR is a potentially promising technique for on-line analysis of process residuals. “The system is simple to operate, potentially applicable to both batch and continuous processes, and could be performed by stop-flow analysis with direct connection of the flow path to the downstream purification process,” he says. Attractive features of NMR in this regard are its inherently quantitative response factor and its relatively high resolving power, and hence higher specificity than a technique such as UV, while drawbacks include barriers to implementation of a conceptually complex analytical technique and limitations on sensitivity. 


	Vol. 32, No. 7
	July 2019
	Pages: 28–31

	When referring to this article, please cite it as C. Challener, “Analysis of Residual Impurities in Continuous Manufacturing," 

Real-time monitoring of product- and process-related impurities remains a challenge.

Analysis of Residual Impurities in Continuous Manufacturing

	Gene therapies are becoming a fast-growing area in the biopharmaceutical industry, creating a crowded pipeline of products in need of manufacturing. FDA expects the agency will receive more than 200 investigational new drug applications for cell and gene therapies by 2020, with the agency predicting that it will be approving 10–20 cell- and gene-therapy products each year by 2025 (1).

	Because gene therapies use the patient’s own body to manufacture the active ingredient, the treatment is specific to that patient. This tailoring can often be a more effective and longer-lasting treatment, according to Chris Murphy, general manager for viral vector services at Thermo Fisher Scientific. “Because of these developments, novel, life-changing therapies are being realized that may transform the lives of patients in need.”  

	This growth is causing contract development and manufacturing organizations (CDMOs) to take a look at their portfolios as their clients require more and more biologics-related services. Catalent has been focused on the clinical pipeline of gene therapies and anticipating the need for commercial manufacturing. “There is an abundance of gene therapy product candidates in the pipeline, and on the heels of market approvals of gene therapy products in 2017 and the influx of venture capital funding for gene therapy companies, FDA approved a second gene therapy this year. We see this as a watershed moment, where gene therapies are gaining regulatory acceptance and setting a precedent for other programs in the pipeline,” says Bernie Clark, vice-president, marketing and strategy, Catalent Biologics & Specialty Drug Delivery.

	According to Clark, outsourcing the manufacture of gene therapies requires relatively low volumes to fulfill demand, making it an attractive option. To better service its clients’ outsourcing needs in this area, Catalent acquired Paragon Bioservices, Inc., a viral vector development and manufacturing company for gene therapies with expertise in adeno-associated virus vectors, in May 2019 (2).

	Clark believes the acquisition will help Catalent accelerate its growth. “Paragon offers its partners GMP-compliant manufacturing capacity, scale-up expertise, and customized downstream processing without the added time, costs, and risks of building a new viral facility,” says Clark. “The acquisition also brings complementary capabilities that will fundamentally enhance our biologics business and our end-to-end integrated biopharmaceutical solutions for customers. We have also gained the experienced leadership and dedicated team at Paragon, who will continue to run the business going forward, bringing their expertise and capabilities to the Catalent family.”

	Thermo Fisher Scientific Inc. took a major step into the viral-vector manufacturing arena with the announced purchase of Brammer Bio, a leading viral vector CDMO, with 600 employees and locations in Massachusetts and Florida. The purchase agreement, announced in a March 24, 2019 press statement, is for approximately $1.7 billion in cash. Brammer Bio joins Patheon, acquired in 2017, and Fisher Clinical Services, which have been combined to form the Thermo Fisher’s Pharma Services business (3).

	The acquisition enhances the company’s pharmaceutical and biotechnology customer value proposition by combining cell-culture media, single-use bioprocessing capabilities, analytical instruments, and related consumables. “Brammer Bio is an exciting addition to our Pharma Services business. By sharing our combined capabilities, expanding commercial scale and broadening deep customer relationships, we will strengthen our position as a trusted partner to our pharma and biotech customers. We will accelerate advancements in gene and cell therapy, meet the increasing demand of the customers we serve and bring life-changing medicines to patients in need,” says Murphy.

	Viral vectors are crucial to gene therapy because they deliver the therapy into the patient. “Viruses are evolutionarily designed to enter mammalian cells and reproduce themselves. Gene therapy harnesses this feature to transport the genetic material-or ‘active ingredient’-to target tissue or cells. Viral vectors are engineered to be safe for humans and can target specific cells or tissues in our bodies to maximize the effect of the treatment,” says Murphy. Thermo Fisher Scientific’s acquisition of Brammer Bio’s viral vector process development and manufacturing leverages capabilities in the company’s biologics business.

	Clark also emphasizes the importance of viral vectors. “During gene therapy biomanufacturing, cells package the therapeutic genetic material into viral vesicles and secrete the vesicles into the media to be purified, then undergo formulation development to be ultimately filled into a vial or syringe,” says Clark. According to Clark, however, few innovators have the capacity, expertise, or resources to manufacture viral vectors. Catalent’s acquisition of Paragon helps the company manufacture viral vectors for their clients. “There are several types of viral vectors that can be utilized for gene therapy. Adeno-associated virus (AAV) is the vector most commonly used today in clinical trials and approved gene therapies, due to its safety and tissue-specific targeting abilities, and Paragon is one of the leading development and manufacturing providers for AAV vectors,” says Clark.

	In May 2019, Thermo Fisher Scientific announced that it is investing more than $50 million into its global bioproduction capabilities. The expansion will provide additional capacity for manufacturing single-use bioprocess container (BPC) systems. In Cramlington, United Kingdom, the company will expand assembly capacity and add BPC systems manufacturing. The proximity of these capabilities to customers in Europe will shorten lead times and improve overall global efficiency, according to the company (4).

	In the United States, the company will expand cleanroom space for BPC chamber and related assembly production processes at its site in Logan, UT, and further expand capacity at its site in Millersburg, PA. Construction is expected to be completed by the end of 2020.

	“The demand for our bioproduction products and services continues to outpace the market,” said Cory Stevenson, president of Thermo Fisher Scientific’s bioproduction business, in a company press release. “These investments will expand capabilities across our existing bioproduction network while we look to extend our footprint into new regions to meet increasing customer demand for our industry-leading single-use technologies.”

	To further develop its biologics offerings, Catalent is launching OneBio Suite, a new offering for the integrated development, manufacturing, and clinical supply of biologic drugs. The Suite is designed to address challenges that arise when accelerating programs to clinic or market, while also Â­reducing Â­complexity and risks from projects. The OneBio Suite is built on Catalent’s track record in biologic drugs development, which includes more than 115 global clinical trials and 11 commercially marketed monoclonal antibodies using the company’s GPEx cell line development technology, and 20 approved products through fill/finish and commercial supply to global markets, the company reports. According to Clark, the company has multiple initiatives in the works to shorten timelines and increase efficiency (5).

	“Time is often lost for sponsors on the path to clinic from contract negotiation, site inspections, handoffs, and poor communication between multiple vendors,” commented Clark in a press statement announcing the service. “Through our new OneBio Suite, Catalent is uniquely positioned to provide an integrated offering that can accelerate biologic development potentially shaving weeks to months off standard timelines and allowing our customers to get to clinic and market faster.”

	1.  FDA, “Statement from FDA Commissioner Scott Gottlieb, M.D. and Peter Marks, M.D., Ph.D., Director of the Center for Biologics Evaluation and Research on New Policies to Advance Development of Safe and Effective Cell and Gene Therapies,” 

	2. Catalent, “Catalent To Acquire Gene Therapy Leader Paragon Bioservices, Inc. for $1.2 Billion,” Press Release, April 15, 2019.

	3. Thermo Fisher Scientific, “Thermo Fisher Scientific to Acquire Brammer Bio, a Leader in Viral Vector Manufacturing,” Press Release, March 24, 2019.

	4. Thermo Fisher Scientific, “Thermo Fisher Scientific to Invest $50 Million to Expand Bioproduction Capabilities,” Press Release, May 20, 2019.

	5.  Catalent, “Catalent to Launch OneBio Suite For Integrated Biologics Development, Manufacturing and Supply at BIO International 2019, Press Release, May 23, 2019.

	When referring to this article, please cite it as S. Haigney, "Gene Therapies Propel Outsourcing Investment," 

Perspectives on Outsourcing

The increasing growth in the cell- and gene-therapy markets is inspiring CDMOs to expand their services in this emerging biologic drug arena. 

Gene Therapies Propel Outsourcing Investment

Over the past few years, after some notable failures at a number of companies (1), global regulators have been paying much closer attention to how pharmaÂ­ceutical manufacturers safeguard the integrity of their data, to prevent accidental manipulation as well as fraud. Data integrity is the foundation of the current good manufacturÂ­ing practices (cGMPs). According to FDA, between January 2015 and May 15, 2016, 21 out of the 28 warning letters issued to pharmaceutical manufacturers centered around problems with data integrity (2).

	As Orlando López, Kansas-based senior site automation specialist for a Big Pharma company, has noted (3), conÂ­nections between different data points (i.e., initial data capÂ­ture, metadata, and records) cannot be weakened or broken because they provide the only objective evidence that operaÂ­tions meet regulatory requirements and are being managed responsibly. These intact connections are also needed for process validation and continuous process verification.

	FDA’s 2016 draft guidance clarified a number of best practices, such as the need for paper and electronic recordkeeping to use the same basic practices, and for an administrative role for managing data, separate from funcÂ­tions that are generating and recording that data. FDA also introduced the ALCOA acronym to emphasize the fact that data must be “attributable, legible, contemporaneous, original (or a true copy), and accurate.”

	In December 2018, FDA updated this guidance (4). “The new guidance serves to clarify data integrity requirements that have consistently challenged organizations. It firms up expectations around data/metadata and the data lifecycle within a risk-based structure of systems and design controls,” says Kir Henrici, CEO of the Henrici Group, a consultant for the Parenteral Drug Association (PDA), which pubÂ­lished best practices on laboratory data integrity in 2018 and plans to publish manufacturing data integrity guidance by the end of 2019.  FDA has emphasized the need for careful risk manageÂ­ment, the importance of audit trails and access controls, and the need to save all records that could be relevant to cGMP compliance.

	FDA’s latest data integrity guidance also underscores the importance of senior management support and involvement in data integrity efforts. “Leadership and culture are key, and, if underdevelÂ­oped, are the starting points for comÂ­pliance risks and gaps,” notes Henrici.

	Training and ongoing communicaÂ­tion with senior management are crucial to establishing and maintaining this support, says López. Expressing potenÂ­tial losses or liabilities in dollars and cents terms can be especially effective, he says, noting a 2017 hacking inciÂ­dent at Merck & Co., which cost $105 million to fix (5) and might have been prevented if investment had been made in data security. He also recommends keeping senior management aware of regulatory citations in a way that allows them to connect data issues to lost batches, product recalls, incorrect labels, or formulation changes. “Senior managÂ­ers needs to hear about money. It gets their attention immediately,” he says.

	It is also important to take a systematic, risk-based approach to data integrity proÂ­grams, says Henrici. Some organizations may roll out data integrity programs in fragments, or as a reaction to a specific event, but then allow efforts to trail off, she says. “Pharmaceutical companies need to structure a systematic, risk-based approach that integrates the quality manÂ­agement system (QMS), which assures the efficacy and sustainability of the entire data integrity initiative.”

	Process mapping is especially important in ensuring that the right documentation is easily accesÂ­sible to staff, but also to regulators. “Companies should track their ‘qualÂ­ity decisions’ to identify where non-cGMP records (e.g., emails) will be required to demonstrate cGMP comÂ­pliance,” says Henrici.

	“You need to understand each speÂ­cific process. Then, if you do a good risk assessment, to record which systems and which data you will focus on, you will be in a much better position to demonstrate full cGMP controls to the authorities,” says López. “If you have done both the risk assessment and the process mapping correctly, there should be no doubt that you are implementing controls to the correct records,” he says. Decisionmaking processes must be documented, says Henrici-not only reported data but any data that support quality decisions-and they need to be justified, retrievable, and reviewable, she says. Maintaining a clear audit trail and using the best approaches to computer system validation will be crucial and will require new training and staffing practices, says Henrici. “Next-generation quality assurance will consist of multiple disciplines or working teams to keep pace with data integrity requireÂ­ments in the era of smart manufacturing, big data management, and artificial intelÂ­ligence,” she says.

	In a sense, however, data integrity is only the first step in a continuum. “Pharmaceutical industry guidelines often refer to data reliability, which is a step beyond data integrity,” says López. “But, ultimately, the goal for our industry is data quality, which is a step beyond reliability. We canÂ­not reach that point with the guideÂ­lines that we have now,” he says. “And cGMP requirements alone will not suffice.” (6) Another challenge is the fact that existing pharmaceutical industry data guidelines don’t go deep enough, to the engineers and IT proÂ­fessionals who are implementing the systems, says López. “ALCOA may give them the ‘why,’ but it doesn’t give them the ‘how,’” he says.

	Annex 11, first published in 1986, addresses data integrity in a more explicit way than the US 21 Code of Federal Regulations (CFR) Part 11, he notes. López recalls one Big Pharma company that correlated Annex 11 with Part 11 computer system requireÂ­ments, which made things easier for engineers and IT professionals impleÂ­menting systems.

	Currently, there are a number of different data integrity guidelines. In pharma, these include regional ones as well as those developed by proÂ­fessional societies such as PDA and the European Compliance Academy (ECA). In addition, the software industry has codes of its own, as does the manufacturing industry in genÂ­eral. The focus of each of the relevant pharma guidelines is different, López says, and the definition of data integÂ­rity in software engineering standards (i.e., as mainly about data security and not changing the records), differs somewhat from its definition in the pharmaceutical industry. The result can be confusing to IT professionals, at least initially, and take time to sort out.

	“I would love to see integraÂ­tion between pharma standards and the software engineering standards, because currently the industry is not speaking the same language as softÂ­ware development academics. For example, for the pharma industry, valiÂ­dation is a process associated with the system lifecycle (SLC), which conÂ­tinues from the start of a project until decommissioning. For software develÂ­opment, validation is just the testing phase of the SLC. So when you hire a new graduate, he or she won’t underÂ­stand the difference,” says López.

	Without understanding the pharma validation concept and the relationÂ­ship between software engineering and software quality, people in pharma’s IT and engineering departments will have different understandings and the resulting programs will be incomplete, he says. López recalls seeing such a situation when asked to consult for one mid-sized pharmaceutical comÂ­pany several years ago. Data security had been made part of the system’s user requirements, but, as one looked at the different processes throughout the product lifecycle, there was one point where the computer system was missing provisions for security, so that anyone could make changes to the data. “There was testing, but no transparency, and in the design document, there had been no steps taken to ensure security throughout the lifecycle,” he says.

	Common standards and definiÂ­tions will be the key to moving the industry beyond data integrity to data quality, López says, explaining that the origin of the International Society for Pharmaceutical Engineers’ (ISPE’s) GAMP 5 guidance (7) is the International Standards Organization’s (ISO’s) 9000-3, which is also the founÂ­dation of the pharmaceutical industry’s guidelines. “Why are we speaking difÂ­ferent languages?” he asks. “We need to synchronize all the different stanÂ­dards (e.g., ISO, ISPE, The Institute of Electrical and Electronics Engineers [IEEE]), and pharma’s so that we are speaking one common language and so that we will understand each other,” says López. Without that common understanding, gaps will persist, some worse than others, he says.

	The pharmaceutical industry’s approach to data integrity and quality will need to evolve as the industry moves to a Pharma 4.0 model and adopts data analytics and artificial intelligence. At INTERPHEX 2019 in April, Henrici spoke of the need for a “Data Integrity 4.0” strategy, a framework for matching IT and data integrity rules. She also noted that FDA approved 12 artificial intelligence algorithms in 2018.

	If the industry is to reap the benÂ­efits of using algorithms and artificial intelligence, she said, companies will need to create multifunctional data governance teams to bring different perspectives to this effort, and facilitate communication between industry and regulators, and data will need to be safe, secure, and relevant (8).

	In the end, the required knowlÂ­edge is already there for implementing new IT approaches, says López, who developed an analytics and business intelligence platform at a Big Pharma company several years ago, before the term “Big Data” had even become a buzz phrase. He advises breaking the problem down to its simplest level: the data wave and input/output (I/O). “With a Big Data project, instead of one, you might have 20 data waves and I/Os. The pharmaceutical industry has been working on all the technoloÂ­gies that industries are exploring (e.g., wireless, industrial internet of things, and Big Data). In the end, each type involves software and hardware, input and output, and there are certain tools best suited for each situation,” he says. “On the fundamental level, we need to understand the relationship between the system, process mapping, data wave sets, and intended use, and simply apply what we have learned during the past 30 years to the new technologies,” adds López (9).

	The industry has come a long way in improving data integrity since the Barr decision in 2005 (1), a landÂ­mark case that resulted in mandates for recordkeeping and the investigaÂ­tion of out-of-specification conditions. New guidance and better integration of existing best practices-not only those designed for pharma’s end users, but for implementation specialists in IT and engineering-promise to push pharmaceutical manufacturers beyond data integrity in the future.

Data Integrity: Detecting and Mitigating Risk

,” presented at the Life Sciences Trainers and Educators Network annual conference (2017).
	2. S. Barkow, “

Current Expectations and Guidance Including Data Integrity and Compliance With cGMPs

,” presented at the International Society for Pharmaceutical Engineers Data Integrity Workshop (Bethesda, MD, June 5, 2016).
	3.  O. López, “

Defining and Managing Raw Manufacturing Data

Data Integrity and Compliance with cGMP, Q&A–Guidance for Industry

 (Silver Spring, MD, December 2018).
	5. C. Souza, 

 38 (12) 28–29, 38 (2018).
	6.  O. Lopez, “Comparison of Health Authorities’ Data Integrity Expectations,” presented at the 4th Annual Data Integrity Validation Conference (Boston, MA, August 2018).
	7.  C. Plagiannos, “

What is GAMP 5 and How Do I Use It Effectively

, montrium.com, Nov. 30, 2015.
	8. A. Shanley, “

, April 15, 2019.
	9. O. López, “Electronic Records Integrity in Data Warehouse and Business Intelligence” in 

Data Integrity in Pharmaceutical and Medical Devices Regulation Operations

 (CRC Press, Boca Raton, FL, 1st ed., 2017), 341–351.

Vol. 32, No. 7
	July 2019
	Pages: 41–45

	When referring to this article, please cite it as A. Shanley, “Removing Gaps in Data Integrity," 

FDA guidance is expected to improve industry practices, but work is also needed to bridge disparate industry and software engineering standards. 

Removing Gaps in Data Integrity

The biopharmaceutical market is challenged by growth, restructuring, and time-to-market pressures. For conÂ­tract development and manufacturing organizations (CDMOs), development timelines for standard vanilla antibodies have been cut in half over the past few years, says Uwe Gottschalk, chief scientific officer at Lonza. “That means going from DNA to clinical trials within one year,” he says, “and development times for all the individual steps entailed, from cell-line construction to a new manuÂ­facturing process, have all been shrinking dramatically.”

	For downstream bioprocessing, the greatest test has been addressing ongoing productivity gains upstream. “There is a gap, and it is widening, between what is coming out of the fermenters and the capabilities of downstream processing systems,” says Gottschalk. While upstream processing is volume-driven, unit operations downstream are mass driven, he says. One can make a kilogram of antibodies in a 1000-L bioreactor, or increase the output to 10 kilograms by increasing expression levels and overall yield, says Gottschalk. But to handle the larger amount downstream, at any capacity, will require chromatography columns that are 10 times larger, or that are run in 10 cycles, he says.

	Market pressures have driven improved process technologies, which some observers refer to as “next-gen” bioprocessing, a catÂ­egory that includes process intensification and the connection of different unit operations, as well as continuous processing, both up and downstream. “These improvements allow manuÂ­facturing to run more efficiently and produce higher yields in less time and space, reducing capital investment,” says Andrew Bulpin, head of process solutions at MilliporeSigma. “By 2020, we expect that approximately 20% of today’s molecular pipeline will be manufactured using elements of next-generation bioproÂ­cessing or continuous manufacturing,” he says.

	Downstream, these improvements include combined unit operations, alternatives to chromatography, new process modelÂ­ing techniques that allow for the visualization of “the golden batch,” and analytical approaches including greater use of process analytical technologies (PAT). While the gap with upstream may still be there, it is being addressed. This article highlights trends.

	Behind many of the new advances is process intensification, which improves facility productivity by focusing on terms of kilograms of product manufactured per year, per square meter of facility footprint, says Peter Levison, executive director of business development at Pall Corp. “Chromatography is a potential bottleÂ­neck downstream because sorbents have a finite binding capacity, but process intenÂ­sification allows feed to be pre-concenÂ­trated prior to adsorption,” he says. “In addition, one can move from batch to continuous chromatography to maximize the adsorptive capacity of the sorbent,” he says, noting the benefits of single-pass tangential flow filtration (TFF). As higher concentration formulations become the rule, filtration systems will have to handle concentrated and more viscous solutions without damaging product, he says, and virus and sterile filters are being developed to address these challenges.

	Bulpin also notes the importance of process intensification “Our work with customers has shown that facility throughput can increase by up to 75%, simply by converting one- or two-unit operations,” he says.

	Alex Chatel, product manager at Univercells, singles out TFF systems as a major process-intensification-based advancement. TFF enables a large increase in concentrations in a single pass, so that feed streams needn’t be recycled many times over through the membrane, he says. Bulpin recalls one case in which a manuÂ­facturer installed a single-pass TFF device, which now provides inline feed stream concentration and/or dilution, elimiÂ­nating the need for holding tanks and intermediate process steps. “With slight modifications to a couple of unit operaÂ­tions, companies can incorporate a more continuous approach while reducing their manufacturing footprint and scale,” he says.

	Using process intensification, unit operations are being integrated so that, for example, cell removal and early clarificaÂ­tion may be captured in one step, instead of the three or four that were previously required (i.e., typically centrifugation folÂ­lowed by depth filtration, sterile filtration and capturing column), says Gottschalk. As a result, some downstream processes have been reduced from 10 to three or four steps, and, in the future, he believes, two or three steps will be possible.

	The move to continuous chromatogÂ­raphy (i.e., having multiple columns running side by side) has also been a sigÂ­nificant advance, says Chatel. Univercells, for example, is developing continuous-process-based platforms and is working on virus manufacturing platforms in projects sponsored by the Gates Foundation. “All are batch processes because they depend on binary loops and they need regeneration, but they are run so that when one is purifying product, the other column is being regenerated, resulting in a continuous stream of prodÂ­uct output. Advances in engineering and mathematics enabled this to happen, and have drastically reduced the amount of resin required,” he says.

	The result of all these improvements has given manufacturers access to a “scaleÂ­able toolbox” that can enable batch, conÂ­tinuous, or hybrid process development and scale up using stainless steel or single-use systems, says Levison.

	Examples he cites include, not only processes that enable continuous chroÂ­matography, but depth filters designed to clarify higher cell density fed batch culÂ­ture; continuous diafiltration and continÂ­uous concentration; and the introduction of large-scale prepacked chromatography columns and new single-use chromatogÂ­raphy skids to complement them. “With associated products such as sterile conÂ­nectors, biocontainers, and single-use filÂ­ters, a closed, fully connected downstream processing train can now be assembled, allowing for rapid scale-up and process implementation,” says Levison.

	There has been a shift to single use downstream, Gottschalk says. In some cases, single-use systems can run at close to 100% variable cost eliminatÂ­ing the need to depreciate costs for the facility and stainless-steel equipment, he says. Gottschalk sees increased use of convective, instead of throughput-limited chromatography media, as a key trend. “Filters are developed with the ligands that would normally go on chromatography resins, and used in a membrane, monolith, or depth filter,” he explains.

	Although some research groups and companies, most notably Merck & Co., have been exploring end-to-end continuÂ­ous biomanufacturing, more manufacÂ­turers are using continuous processing strategically, both up and downstream. Univercells has developed a continuÂ­ous production system in which cells are grown in batch mode but different process steps are run continuously. “We need to get away from the old approach, in which you send all the contaminants downstream,” says Chatel. As Gottschalk says, continuous processing was always used upstream for sensitive molecules that could not endure conditions in the fermenter (e.g., enzymes such as Factor A for hemophilia treatments). Perfusion fermentation was essential in order to keep the molecule intact, but the approach has since been used with other more stable molecules such as antibodies, and is now moving into the broad bioÂ­pharma world, he says.

	Continuous processing is already a reality for single-use systems, Levinson says. “We can currently clarify cell culÂ­ture media continuously using depth filÂ­tration; purify product using continuous chromatography; carry out continuous virus inactivation; and then complete the ultrafiltration and diafiltration stages using continuous single-pass filtration technologies,” he says.  Trends for the future relate more to the modularity of the continuous downÂ­stream platform, with connectivity between each unit operation and inteÂ­grated PAT and automation and control platforms, Levison says. For a truly conÂ­tinuous upstream process, a perfusion bioreactor could be used upstream, but it would require continuous cell culture fluid harvesting. Current cell-retention technologies based on hollow fiber filtraÂ­tion are prone to fouling, he says, so the composition of the harvest is variable and changes over time. One alternative could be to use acoustic wave separaÂ­tion (AWS), and platforms using AWS should be commercially available in the near future, he says.

	The biggest change in bioprocess autoÂ­mation has been the introduction of the model-driven approach, says Bulpin, who explains that this approach allows users to define process, interactions, and contexts in form of parameterized data. This helps to reduce complexity, increase flexibility and reusability, and enable dynamic re-configuration of systems and processes. Model-driven design can lead to increased quality by increasing visÂ­ibility, providing higher abstraction layers with less custom code, and empowerÂ­ing the domain experts by focusing on the process rather than underlying techÂ­nology, Bulpin says. In the long term, he expects it will lower the sensitivity to change and help to bridge the gap between business and technical domains.

	At the same time, analytical methÂ­ods are improving, and use of PAT is increasing. Chatel has seen advances in sensors for pH, dissolved oxygen, and traditional measurements, that have allowed companies to adopt monitorÂ­ing and control practices for bioreactors and other steps such as chromatography. “Future sensors will be able to detect metabolic concentrations in cell cultures (e.g., measure titers at line and provide feedback control over feeding strategies) to reduce burden of contaminants and ultimately lead to more lean processes and reduce processing time,” he predicts.

	Noting the increased use of Raman spectroscopy-based methods, Gottschalk says, “We can now monitor the most relevant critical parameters online or at-line so that, after we have manufacÂ­tured the batch, we know whether it is acceptable and can be approved or not. In biopharma, we have one or two critiÂ­cal assays that need to be run at the end of production, such as the adventitious virus test, which cannot currently be brought online, but most other analytiÂ­cal tests can be run online,” he says.

	Bupin sees artificial intelligence (AI), primarily machine learning, as drivÂ­ing future improvements in PAT. “This approach should significantly improve quality and process robustness; combined with near-real-time data acquisition, it will enable closed loop for the process execution,” he says, improving commuÂ­nications and interaction between enterÂ­prise systems and manufacturing field I/O (inputs/outputs), including sensors, actuators, analyzers and drives.

	As more bioprocessing steps are automated, the approach is also proving important for sampling, says Gottschalk. Lonza is working on data mining and new approaches using machine learning and other artificial intelligence concepts. “The goal is to use advanced concepts and multivariate data analysis to ensure that we run only the ‘golden batches,’” he says.

Although downstream efficiency still lags behind upstream, engineering-driven innovation is breaking through boundaries.

BioPharm International-07-01-2019

FDA Revamps Biosimilar Quality Requirements

Playbooks Are Not Just Child’s Play

Applying QbD to Upstream Processing

A Biologics Partisan Divide

Analysis of Residual Impurities in Continuous Manufacturing

Gene Therapies Propel Outsourcing Investment

Removing Gaps in Data Integrity

Catching Up Downstream

Design Considerations for a Commercial Cell and Gene Therapy Facility

Catalytic Site Analysis and Characterization of a Solvent-Tolerant Aldo-Keto Reductase

Up-to-Date Systems Streamline Operations

BioPharm International, July 2019 Issue (PDF)