Approaches and Lessons Learned for Analytical Method Transfers to Manufacturing Sites at Various Stages of Development

Overview

Analytical method transfers to external testing sites are required by health authorities to ensure both quality of clinical materials provided to patients, and effective testing during registrational stability studies. With the different approaches available to satisfy these regulatory requirements, factors such as stage of process development, complexity of the control strategy, prior experience, and infrastructure/analytical skillsets available should be considered. This article will review different approaches currently in use for analytical method development and method transfer from innovator drug development sites to external contract manufacturing organizations (CMOs). Experiences from recent analytical transfers to support a Drug Substance (DS) manufacturing process will be provided, however the strategies utilized are also applicable to transfer of drug product control strategies. Approaches and lessons-learned for consideration, such as timelines, use of non-QC friendly analytical methods, contingency planning, documentation and release vs. For-Information Only (FIO) testing to facilitate transfers will be shared.

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers
from American Pharmaceutical Review – all delivered right to your inbox! Sign up now!

Introduction

Recently, industry has witnessed an increased use of CMOs by innovator pharmaceutical companies to manufacture clinical supplies of drug substance and their intermediates.1,2 As a result, there has been a challenge, to transfer Fit-for-Purpose (FFP) analytical methods which have limited robustness studies and batch testing experience. As modelling and knowledge of how drug targets attach to disease receptor sites has advanced, there has been a corresponding increase in the complexity of the small molecule targets.3 This, combined with the need for low-level monitoring (e.g. parts per million) of potentially genotoxic impurities (GTIs),4 has resulted in analytical control strategies becoming increasingly more challenging utilizing analytical detection techniques outside of traditional QC-friendly LC/UV and GC/FID detection. Manufacturing processes evolve during clinical development to provide increased scale of production, robustness, and enhanced cost-effectiveness for delivering commercial material. The analytical methods also evolve and the different versions are transferred to the CMOs multiple times (including other CMOs) to support the different campaigns during process development. Clearly, a challenge exists, to ensure the quality of clinical campaigns/materials in a FFP phase-appropriate manner; while embracing the more time consuming, business complexity and scrutiny associated with developing and transferring methods to support process validation and commercial manufacture as outlined in ICH Q71.

Analytical control/testing strategies typically evolve throughout clinical development as the critical process parameters (CPPs) and critical quality attributes (CQAs) of the process and DS become better understood. The lack of experience with initial processes developed to manufacture clinical materials necessitates the reliance upon an appropriate set of DS specifications to ensure quality. As additional process knowledge and understanding of the CPPs and CQAs become available, the business risk of generating DS batches that fail specifications is greatly reduced. A robust analytical data set is needed to generate process knowledge, thus a greater number of analytical methods are utilized in development than are typically used in the final commercial process. The data from these additional methods help refine the commercial control strategy to ensure commercial viability. Much of this supplemental data can be acquired using FIO or In Process Control (IPC) testing methods to generate useful process knowledge. FIO tests do not require validation or formal Technology Transfer (TT); however often they use the IPC or release method conditions. Generating meaningful process knowledge allows the process team to implement effective and lean analytical strategies for the validation campaign of the commercial process.

Evolving control strategies represent a challenge to the process team as they represent the culmination of considerable knowledge collected across multiple campaigns run potentially at multiple CMOs over the clinical program. At BMS the preferred strategy is to utilize CMOs to provide all analytical campaign support for the batch release, as well as IPC/FIO testing. This strategy allows limited, internal analytical scientists to focus on collaborating with their chemistry and engineering colleagues on in-house process development studies. Though this approach does create more effective and productive internal collaborations, it does generate several challenges including assurance that the CMOs provide accurate, reliable analytical data from the FIO, IPC and intermediate release methods and containing the analytical costs associated with the campaign within a reasonable budget.

There is a need to work effectively with the CMOs to ensure that there is confidence in the data that they generate. Therefore, limited qualification of these methods or default acceptance criteria are typically used to manage the scope of the analytical package. During clinical development, regulatory (e.g. “release”) specifications are needed only for the final DS (where methods require more qualification rigor) with intermediate specifications typically not required until the validation (commercial) stage of development. A ‘use test’ of the materials (intermediates, reagents, etc.) is typically employed as a further control of quality and source of knowledge gained from each step of the clinical campaign. For the validation campaign (commercialization), methods for intermediates, IPCs and starting materials as well as the API all require registrational validation.1,5

Method Transfer Approaches

Analytical method transfers6 should not be a check-box exercise for the Receiving Laboratory (RL) aimed only at fulfilling any regulatory requirements and/or expectations. The goal of the method transfer activity is to provide the RL with sufficient method experience such that results the laboratory produces are accurate and scientifically sound. The ultimate goal of a method transfer is to demonstrate that the method produces consistent results irrespective of which laboratory was used to produce the results. Different approaches7 to method transfers can been used, such as qualification, validation, co- validation, verification, and transfer. Table 1 provides a list of the various approaches as well as recommendations and comments for each approach. In considering which to use for a transfer, one should assess several practical aspects of the transfer process such as representative sample material availability, equipment, time and resources required. Technical aspects, including complexity, of the methods should be assessed by the Expert Lab (EL) to ensure the RL gains key information to replicate it and gain experience prior to formal transfer.

There are many factors to consider when assessing the risks associated with a method transfer.8 With any method, factors related to instrumentation, analyst experience, and timelines can all factor into the overall risk of a successful transfer. Many of these factors are summarized in the Ishikawa diagram in Figure 1. To assess and plan for the degree of training required consideration of several factor areas should be made, such as RL familiarity with the method/technique, method complexity, and accuracy level required for the RL. A simple training assessment tool is illustrated in Figure 2.

Strong collaboration between the entire development team of method developers, process chemists, engineers, outsourcing teams, quality organizations and CMOs is needed to perform successful TTs and effectively resolve the technical challenges that arise during the transfer. This collaboration should begin several months in advance of the campaign by capturing the critical information for the method performance in the initial Request for Proposal (RfP) sent to the CMO. Clearly outlining the method required for the campaign, the specific instrumentation needed and transfer requirements are crucial to establishing aligned expectations with the RL. However, during process development it is sometimes difficult to provide detailed information for each method as the methods quality attributes may still be under development and evolving at the time of the RfP. Outlining expectations for the qualification/validation of methods, the approach to method transfer that will be utilized, and the expected timelines can be agreed upon with the CMO early in the process while the methods are still being developed. A summary of the roles and responsibilities for both the EL and the RL is shown in Figure 3. Analytical transfers to CMOs with prompt, effective planning and clear communications can help transfer activities stay off the project critical timeline path.

There are different drivers and timelines for clinical campaigns compared to validation campaigns as outlined in ICH Q7.1 These considerations lead to different activities and timelines as shown in Figure 4. Although analytical methods during the early stage of development may be considered FFP, and may not be part of a long term strategy, these methods have both benefits and drawbacks:

1. The methods enable effective use of an array of analytical instrumentation for quick and insightful information,
2. Acceptable error from un-optimized methods can still generate valuable information,
3. The CMO may not have the skillset or equipment to execute specific methods,
4. A FFP method may not be robust enough to be used in a commercial setting.

Balancing these benefits and drawbacks may require more complex analytical testing coordination for clinical materials such as:

1. Testing at alternative laboratories (3rd party),
2. Accepting less precise and/or accurate data,
3. Agreement on specific training or instrumentation purchase at the CMO as either a short term or long term investment
4. Having an appropriate backup strategy.

Addressing Challenges at the CMOs

Even with careful planning, problems are likely to be encountered when transferring and executing pre-commercial methods to CMOs. The analytical testing lab is a cost center for the CMO and as such its resources are typically lean. Therefore, it is not uncommon to find that the resources available to resolve issues that may arise are often stretched quite thin, which presents a significant risk to the project timeline and successful technology transfer. Effective collaboration with the CMOs is key to solving problems effectively and ensuring timelines are met and the process knowledge generated is meaningful. Troubleshooting a method issue can typically be done remotely; however, when tight timelines are needed or when specific method transfers are considered high risk (as outlined in Figure 3) it is recommended that the EL travel to the CMO to ensure issues are resolved timely and effectively. Investigations are needed when the acceptance criteria for a method transfer or validation cannot be met in a CMO laboratory. A typical workflow is shown in Figure 5. If the method in question is needed for a critical IPC or batch release, EL will need to work closely with the CMO’s Quality Assurance (QA) team to ensure that the laboratory is qualified. The investigation plan, any revisions to the transfer protocol/method, and the final investigation report will need to be approved by the CMO’s QA team. In the remainder of this section several challenges recently encountered in transferring methods to CMOs will be highlighted with the lessons learned. Additionally, a set of overall findings that can be applied in a preventive way are shared in Figure 6.

Example 1: Small differences in analytical practices at a CMO’s site can cause significant transfer difficulties. For example, a low recovery issue was observed when a GTI trace analysis method for 2-hydroxypyridine-N-oxide (HOPO) was being transferred to a CMO. Investigations into the instrument performance/parameters, method conditions, materials, and sample preparations failed to provide a root cause and could not improve the initial low recovery values obtained at the CMO. Timeline concerns necessitated that the subject matter expert (SME) travelled to the CMO to oversee the investigation. During this on-site interaction with the CMO, the EL analyst observed that the CMO staff utilized an aluminum weighing boat to weigh the sample. The weighing boat and sample were then transferred to the volumetric flask and sonicated before final dilution of the sample. As HOPO can chelate to aluminum, the use of the aluminum weighing boats was the source of the low recovery observed (see Figure 7a). This example highlighted that it is beneficial to be familiar with practices for common lab unit operations.

Example 2: For many development LC methods selection of a robust HPLC column is critical to pass NDA validation. For this example, a new C18 (end-capped) HPLC column and two used columns were evaluated in the EL for the HPLC release method of an early intermediate. The system suitability results for these three columns was a resolution > 3.0 between the main peak and a key impurity. The system suitability resolution acceptance criteria chosen was ≥ 2.0. During the RL validation of the method using three new columns from different column silica lots, the resolution was found to be 1.3 - 2.4. A systematic investigation by the CMO, confirmed that the low resolution values observed were caused by the HPLC column, rather than the HPLC instrument, solvents, or procedures. An in-house investigation confirmed the resolution with newly purchased lots of HPLC columns and hypothesized that exposed silanol groups on the surface of the underlying silica particles significantly improved the resolution and the silica manufacturing process may have changed (with more efficient end capping). Subsequently a simple procedure to remove some of the end capping using elevated levels of TFA in mobile phase was developed to precondition new HPLC columns to give a resolution of > > 2.0. A chromatography overlay of the different columns is shown in Figure 7b. This example highlights the impact that subtle changes in the stationary phase manufacture can have. Therefore robustness should be evaluated with old and new column batches for future commercial testing methods.

Example 3: The close collaboration between the EL and the CMO is critical particularly when the CMO encounters a method transfer/ validation issue. An unexpected challenge was encountered after transferring an IPC HPLC method for an intermediate step. This method had previously been successfully qualified at the RL (CMO’s analytical R&D laboratory) which was then followed by an internal transfer to the CMO’s QC laboratory. Surprisingly, the QC laboratory observed a co-elution of starting material peak and a new unknown peak, which was not observed in either the EL or the RL. The investigation determined that the unknown peak was introduced from a different lot of reagent than the one used in the other laboratories. Additionally the intermediate HPLC release method could separate the starting material peak and the unknown peak. In order to keep the project timeline the release method was also validated for use as the IPC. The validation was completed at the EL and transferred to the CMO through a transfer waiver with familiarity testing. This example highlighted the collaboration value when solving ‘last minute’ unexpected analytical problems.

Example 4: Adding detailed descriptions and controls into the method procedure can help the performance of less robust methods. An HPLC release method of an unstable di-bromo starting material (SM) was developed and successfully validated at the EL. The subsequent transfer led to challenges with reproducing the method at the RL where the chromatogram contained two new unknown impurity peaks. The investigation determined the RL had generated two SM degradants (see Figure 7c) from using a different quality/grade of tetrahydrofurn diluent and bottled tri-fluoroacetic acid (TFA) vs ampoule. Specific instructions on the grade/age of materials along with the mobile phase preparation, diluent preparation and storage conditions were added to control the degradation risk.

Conclusions

The transfer of analytical methods to the various CMOs during development when the process, methods and the analytical control strategy have not been finalized or optimized has been shown to be challenging. However, with appropriate assessment, planning, clear roles and responsibilities, proactive communication and prioritization successful transfer can be achieved to enable the clinical campaign needs. These transfers leverage learnings and process knowledge to help refine the commercial analytical control strategy and optimize method critical quality attributes. Although encompassing more partners and requirements at the validation stage, successful transfers can be accomplished using similar up front preparations, organized regular communications and use of appropriate risk assessment tools. Regularly sharing learning experiences and use of consistent work practices can simplify and enable timely transfers that reduce the risk of issues arising.

Author Biographies

Peter Tattersall, PhD, is a Senior Principal Scientist in Chemical and Synthetics Development department within Product Development at Bristol-Myers Squibb Company in New Jersey. He received his BSc. And

Ph.D. from the University of Manchester, UK. He previously worked in Analytical Development at AstraZeneca, Wilmington. He joined Bristol- Myers Squibb in 2003 where he is an analytical team leader within Chemical and Synthetic Development. He supervises a small group of analytical chemists working on method development, qualification and transfer in process support of drug substance synthesis.

Xuejun Xu, PhD, is a Senior Research Investigator in Chemical and Synthetics Development department within Product Development at Bristol-Myers Squibb Company in New Jersey. He received his BSc from Wuhan University, China, and Ph.D. from Dalhousie University, Canada. He joined Bristol-Myers Squibb in 2006 and his work focuses on method development, validation, and transfer in process support of drug substance synthesis.

Brent Kleintop, PhD is a Group Director in the Chemical and Synthetics Development department within Product Development at Bristol-Myers Squibb Company in New Jersey. He received his Ph.D. from the University of Florida. He has been with BMS since 1997 and has also worked for Wyeth Research (now Pfizer) and Finnigan MAT (now ThermoFinnigan). Brent’s interests and experience focus on leveraging innovative analytical approaches to solve challenging problems encountered in developing clinical and commercial drug substance and drug product processes, and transfer of methods to commercial manufacturers.

Acknowledgements

The authors wish to acknowledge colleagues who collaborated on the method transfers described in this manuscript. Adrian Doggett, Morgan O’Sullivan, Amy Bu, Kieran O’Conner, Rachel Wall, Diarmuid Scanlon, Li Li, Neil Tang, John Hayes, Michelle Kubin, and Qinggang Wang all worked on the transfers described in this manuscript. The authors also recognize their chemistry and chemical engineering colleagues such as Al Delmonte, Ken Fraunhoffer, and Ben Cohen who provided important expertise in their areas of specialty in solving some of the problems encountered during these transfers. Finally, thanks to Jonathan Shackman, Tom Raglione, and Scott Miller for their input in the manuscript review.

References

1. ICH, Q7(R1), Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients (Sep 2016); FDA Guidance for Industry, Analytical Procedures and MethodsValidation for Drugs and Biologics, July 2015; EudraLex, Volume 4, EU Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use, Part 1, Chapter 6: Quality Control, 2017
2. Future Pharma Partner Models - Outsourcing Trends in API Development & Manufacturing, S. Kuehn, C. A. Challener, E. Branch, C. Cao and J. Sticker American Pharmaceutical Review. 2016, 19 (5); “Successful API Development: Lessons from a CMO”, E. Price, Pharmaceutical Outsourcing, 2017 18, (1); “Specialty Markets and Services Drive API Growth.” A. Shanley, Pharmaceutical Technology. (2016) 40, (3) 26-30
3. “Current Complexity: A Tool for Assessing the Complexity of Organic Molecules”; J. Li and M. D. Eastgate; Organic and Biomolecular Chemistry., 2015, 13, 7164.; “Designing for Efficiency: A Discussion on Selecting Synthetic Approaches for the Synthesis of Drug Candidates”; M. D. Eastgate, M. A. Schmidt and K. R. Fandrick; Nature Reviews Chemistry, 2017, 1, 16.
4. ICH M7(R1) Assessment and Control of DNA Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk (March 2018)
5. ICH Q2BValidation of Analytical Procedures: Methodology (Nov 1996)
6. Chapter 8 Analytical Method Transfer, M. Swartz, I. Krull, Handbook of validation, CRC press, 2012; Chapter 13 Analytical Method Transfer,” S. Scypinski and J. Young, Handbook of Modern Pharmaceutical Analysis, S. Ahuja and S. Scypinski, Academic Press, 2011; B. Davani, Chapter Pharmaceutical Analysis for Small Molecules, John Wiley and Sons, Inc., 2017;“The transfer of analytical procedures”J. Ermer, M. Limberger, K. Lisb and H. Wätzigc. Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 262–276
7. CH Q2B Validation of Analytical Procedures: Methodology (Nov 1996); USP 41, General Information Chapter: Transfer of Analytical Procedures, Rockville, General Information Chapter: Verification of Compendial Procedures, Rockville; “Covalidation Strategies to Accelerate Analytical Method Transfer for Breakthrough Therapies”Pharm. Tech. 41 (4) 38-48.
8. ICH Q9 Quality Risk Management (June 2006); “Analytical method transfer using equivalence tests with reasonable acceptance criteria and appropriate effort: Extension of the ISPE concept” L. Kaminski, U. Schepers and H. Wätzig., Journal of Pharmaceutical and Biomedical Analysis 53 (2010) 1124–1129; “Applying Quality by Design Principles to Analytical Methods to Gain Enhanced Method Understanding”, K. Barnett, K. Doyle, K. Wang, J. Morgado and J. Harwood; American Pharmaceutical Review. 2015, 18 (3).