Drug discovery, preclinical, and development phases can take years, cross multiple organizations, facilities, researchers, and technologies. Maintaining the integrity of bioanalytical methods is paramount. Roger Hayes, PhD, vice-president, general manager of Laboratory Sciences at MPI Research, shares best practices in an interview with Pharmaceutical Technology.
Best practices: Preclinical to Phase I
PharmTech: What would you identify as some key best practices to ensure successful bioanalytical methods transfer from preclinical to Phase I development?
Hayes: The preclinical space incorporates both discovery and safety assessment phases, each with different requirements of bioanalytical method quality and sensitivity. A common term is fit-for-purpose, which refers to the quality of the method required to answer questions or regulatory requirements at a particular stage of drug development. Safety assessment studies are normally performed in accordance with GLP regulations and, as such, require fully validated bioanalytical methods. In contrast, methods that support candidate selection or lead optimization are not validated, but the methods do require sufficient quality to provide confidence in the results.
Because preclinical safety assessment involves higher drug exposures than those anticipated in the clinic, the bioanalytical method has a higher limit of quantitation that offsets the limited volumes typically available for sample extraction and subsequent concentration. Moreover, the number of samples collected for preclinical analysis is much lower than the high-throughput environment of clinical drug development. Therefore, assay robustness is not of high concern up until the transition to clinical bioanalysis.
Assay robustness is best achieved by adequate sample cleanup and chromatographic selectivity that separates matrix interferences and biotransformed products from the analyte of interest. Chromatography method development following preclinical bioanalysis is usually sufficient for first-in-human clinical trials. Some consideration is typically given to shorter analytical run times to accommodate higher sample numbers and a desire for more rapid data turnaround. If the first clinical studies use normal, healthy volunteers, the sample composition is not likely to be much different from preclinical studies, but patient samples will likely be more heterogeneous. Sample cleanup will, therefore, require more sophistication; a simple protein precipitation approach used in preclinical bioanalysis may give way to solid phase extraction or phospholipid removal methodologies. To achieve lower limits of quantitation, higher sample volumes are typically extracted and concentrated. Fortunately, sample volume is not normally a limitation in clinical bioanalysis, with the exception of pediatric or certain oncology trials.
Even with adequate sample cleanup and chromatographic selectivity, it is important to sufficiently compensate for analyte recovery or ionization differences by an appropriate internal standard, most often a stable isotope-labeled analog of the drug. Indeed, a stable labeled internal standard is essential for clinical analysis because of the heterogeneity of humans compared to preclinical species.
PharmTech: Can you outline specialized considerations for bioanalytical methods transfer when the drug substance is a macromolecule, such as pharmacokinetic and immunogenicity measurements, or the use of specific reagents (i.e., antibody pairs) in the method development?
Hayes: Two key elements in successfully transferring a bioanalytical method for a protein therapeutic are the identification and sourcing of critical reagents. The identity of the test article will have a direct impact on the complexity of the ligand-binding assay (LBA), whether the assay is to support pharmacokinetic (PK) measurements or determine anti-test article antibody (ADA) response. Moreover, critical reagents used for capture and detection in PK assays, or as surrogate positive controls for ADA assays, need to be reliably sourced. The success of the LBA method is highly contingent on the quality and integrity of the critical reagents. When critical reagents are not commercially available, a sponsor will be expected to initiate the generation of critical reagents either internally or externally (outsourced).
Over the lifetime of the assay, there needs to be monitoring and quality control of the reagents. Reagent lot-to-lot differences may lead to significant variability in assay performance. When such variability is observed, the assay may require modification to return assay performance to established acceptable limits. The lot-to-lot variability may be substantial when the reagent is a polyclonal antibody that was generated from different immunizations or different animals. Alternatively, the use of monoclonal antibodies, where all lots are obtained from the same hybridoma cell line, is one approach that sponsors and CROs may choose to mitigate such differences.
PharmTech: What are some problems in data/information sharing that might be encountered in bioanalytical methods transfer and how may they be resolved? For example, how are laboratory variations (e.g., methods of pipetting, storage of reagents, analyte stock solution, specialized equipment, reagent purity and preparation) addressed that might influence method variability?
Hayes: There are many different reasons for a sponsor to outsource bioanalysis. But once an outsourcing relationship is established, the level of communication must align with the sponsor. At its foundation, the relationship requires the exchange of information in an accurate and timely manner. Over the course of the relationship, various activities occur, including assay development, validation, and sample analysis. The sponsor will expect that these activities are performed by well-trained bioanalytical scientists. The CRO will provide individual training, and because efficiency is a key requirement of CRO operations, harmonizing techniques and streamlining bioanalytical processes across the laboratory will likely be implemented. To reduce and even eliminate technical differences in pipetting, for example, using dependable automated liquid handling systems is strongly recommended. With automation, analyst fatigue resulting in human error is dramatically reduced and productivity is increased.
The assay format and platform chosen for a given method are influenced by many factors. For instance, preclinical assay formats are largely dependent on availability of critical reagents, while dose and sensitivity may dictate the platform used. Clinical programs, on the other hand, require specific human protein recognizing reagents capable of selectively detecting human protein therapeutics. Because of the need to ensure method transferability across different facilities, it is unlikely that novel assay platforms will be selected for the initial method.
As discussed above, the success of preclinical and clinical bioanalysis is determined by the quality of an assay’s critical reagents. During preclinical development, the life span of an assay is often short term and predictable; therefore, reagent availability is the only concern best mitigated with effective relationships with reagent vendors. Over prolonged clinical studies, implementation of more stringent physicochemical and biophysical characterization methods for critical reagents is essential to ensure the integrity of the bioanalytical method and data quality.
Achieving acceptable standards
PharmTech: How is the acceptability of a method transfer best evaluated? How are factors such as site transfer considered?
Hayes: The pharmaceutical industry is highly regulated and regulatory authorities normally offer guidance documents that describe best practices for the validation and subsequent application of bioanalytical methods. The fundamental parameters of a method that require being in control include accuracy, precision, selectivity, sensitivity, reproducibility, and stability. The quality of an analytical method is normally assessed by the performance of standards and quality control samples (QCs). Current guidance recommends an acceptance criterion of 15% for accuracy, precision of all standards, and QCs, with the exception of the lower limit of quantitation (LLOQ), where the acceptance criterion is increased to a 20% deviation. However, standards and QCs may not adequately reflect the composition of study samples from dosed subjects (i.e., incurred samples). The limitation of using spiked control samples to assess ‘real’ samples is known, and there is general acceptance that confirmatory reanalysis of incurred samples can be used to assess the reproducibility of a method.
Method transfers are very common and, in fact, the majority of regulated bioanalysis is now performed at CROs. A method may be validated either in-house or at another CRO, and because of resource or quality constraints the sponsor might request that a new CRO use the previously validated method with minimal additional work. It is important to note that a regulator expects to be able to verify all aspects of the method validation during an inspection at the site performing regulated bioanalysis. If all raw data generated during method validation are available on-site for review, and the regulator can verify that the original method is being executed in the same validation manner, this would eliminate the need for repeating the method validation at a new CRO. Unfortunately, it is rare that all raw data are transferred to the receiving CRO, when what typically is provided is only the validation report. Moreover, it is difficult to assure a regulator that the method is being executed in exactly the same manner as might be documented in a validation report. The path of least resistance is to revalidate the method in its entirety to assure that all source records are available in the event of a regulatory audit.
A regulator will also typically require some assessment of comparability between methods, particularly if two or more methods are used within the same study. Comparability can be performed by using samples of known concentration (i.e., QCs and standards), and/or existing study samples. The acceptance criterion for comparability is commonly 20% relative error for small molecules and 30% for large molecules (e.g., protein therapeutics).