2021 White Paper on Recent Issues in Bioanalysis: Mass Spec of Proteins, Extracellular Vesicles, CRISPR, Chiral Assays, Oligos; Nanomedicines Bioanalysis; ICH M10 Section 7.1; Non-Liquid & Rare Matrices; Regulatory Inputs (Part 1A – Recommendations on Endogenous Compounds, Small Molecules, Complex Methods, Regulated Mass Spec of Large Molecules, Small Molecule, PoC & Part 1B - Regulatory Agencies' Inputs on Bioanalysis, Biomarkers, Immunogenicity, Gene & Cell Therapy and Vaccine)

Index Part 1A

Introduction

SECTION 1 – Regulated Mass Spectrometry of Large Molecules

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

SECTION 2 – Endogenous Compound & Complex Methods

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

SECTION 3 – Regulated Bioanalysis for Small Molecules & Point of Care

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

SECTION 4 – Mass Spectrometry of Proteins

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

Index Part 1B

SECTION 5 – Input from Regulatory Agencies on Regulated Bioanalysis & BMV

SECTION 6 – Input from Regulatory Agencies on Immunogenicity, Biomarkers, Gene & Cell Therapy and Vaccines

References

Introduction

The 15th edition of the Workshop on Recent Issues in Bioanalysis (15th WRIB) was held between 27 September and 1 October 2021. Even with a last-minute move from in-person to virtual, an overwhelmingly high number of nearly 900 professionals representing pharma and biotech companies, contract research organizations (CROs), and multiple regulatory agencies still eagerly convened to actively discuss the most current topics of interest in bioanalysis. The 15th WRIB included three Main Workshops and seven Specialized Workshops that together spanned 1 week in order to allow exhaustive and thorough coverage of all major issues in bioanalysis, biomarkers, immunogenicity, gene therapy, cell therapy and vaccines.

Moreover, in-depth workshops on biomarker assay development and validation (BAV) (focused on clarifying the confusion created by the increased use of the term “Context of Use [COU]”); mass spectrometry of proteins (therapeutic, biomarker and transgene); state-of-the-art cytometry innovation and validation; and, critical reagent and positive control generation were the special features of the 15th edition.

As in previous years, this year's WRIB continued to gather a wide diversity of international, industry opinion leaders and regulatory authority experts working on both small and large molecules to facilitate sharing and discussions focused on improving quality, increasing regulatory compliance, and achieving scientific excellence on bioanalytical issues.

The active contributing chairs included: Dr. Eugene Ciccimaro (BMS), Dr. Anna Edmison (Health Canada), Dr. Fabio Garofolo (BRI), Dr. Swati Gupta (AbbVie), Dr. Shannon Harris (HilleVax), Dr. Carrie Hendricks (Sanofi), Ms. Sarah Hersey (BMS), Dr. Steve Keller (AbbVie), Dr. Lina Loo (Pfizer), Dr. Mark Ma (Alexion), Dr. Joel Mathews (Ionis), Dr. Meena (Stoke), Dr. Manoj Rajadhyaksha (Alexion), Dr. Ragu Ramanathan (Vertex), Dr. Susan Spitz (Incyte), Dr. Dian Su (Mersana), Dr. Matthew Szapacs (Abbvie), Dr. Albert Torri (Regeneron), Dr. Jian Wang (Crinetics), Dr. Jan Welink (EU EMA), and Dr. Yuling Wu (AstraZeneca).

The participation of major and influential regulatory agencies continued to grow at the 15th WRIB during its traditional Interactive Regulators' sessions including presentations and panel discussions on:

• Regulated Bioanalysis and BMV Guidance/Guidelines: Dr. Seongeun Julia Cho (US FDA), Dr. Arindam Dasgupta (US FDA), Dr. Anna Edmison (Health Canada), Dr. Elham Kossary (WHO), Mr. Gustavo Mendes Lima Santos (Brazil ANVISA), Dr. Sam Haidar (US FDA), Dr. Sandra Prior (UK MHRA), Dr. Mohsen Rajabi Abhari (US FDA), Dr. Diaa Shakleya (US FDA), Dr. Catherine Soo (Health Canada), Dr. Nilufer Tampal (US FDA), Mr. Stephen Vinter (UK MHRA), Dr. Yow-Ming Wang (US FDA), Drs. Jan Welink (EU EMA), Dr. Jinhui Zhang (US FDA)

• Biotherapeutic Immunogenicity, Gene Therapy, Cell Therapy and Vaccines: Dr. Nirjal Bhattarai (US FDA), Dr. Elana Cherry (Health Canada), Dr. Isabelle Cludts (UK MHRA), Dr. Heba Degheidy (US FDA), Dr. Soma Ghosh (US FDA), Dr. Akiko Ishii-Watabe (Japan MHLW), Dr. Susan Kirshner (US FDA), Dr. Kimberly Maxfield (US FDA), Dr. Joao Pedras-Vasconcelos (US FDA), Dr. Mohsen Rajabi Abhari (US FDA), Dr. Vijaya Simhadri (US FDA), Dr. Dean Smith (Health Canada), Dr. Therese Solstad (EU EMA/Norway NoMA), Dr. Daniela Verthelyi (US FDA), Dr. Meenu Wadhwa (UK MHRA), Ms. Leslie Wagner (US FDA), Dr. Günter Waxenecker (Austria AGES), Dr. Haoheng Yan (US FDA), Dr. Lucia Zhang (Health Canada)

• Biomarkers/CDx and BAV Guidance/Guidelines: Mr. Abbas Bandukwala (US FDA), Dr. Soma Ghosh (US FDA), Dr. Shirley Hopper (UK MHRA), Dr. Kevin Maher (US FDA), Dr. Yoshiro Saito (Japan MHLW), Dr. Yow-Ming Wang (US FDA)

All the traditional “working dinners” attended by both industry key opinion leaders (KOL) and regulatory representatives were held in a virtual format this year, and the extensive and fruitful discussions from these “working dinners” together with the lectures and open panel discussions amongst the presenters, regulators and attendees culminated in consensus and recommendations on items presented in this White Paper.

A total of 66 recent issues (‘hot’ topics) were addressed and distilled into a series of relevant recommendations. Presented in the current White Paper is the background on each issue, exchanges, discussions, consensus and resulting recommendations.

Due to its length, the 2021 edition of this comprehensive White Paper has been divided into three parts for editorial reasons. This publication covers Part 1 recommendations.

Part 1 – Volume 14 Issue 9 (May)

Regulated Mass Spectrometry of Large Molecules (six topics)

Endogenous Compounds & Complex Methods (five topics)

Regulated Bioanalysis for Small Molecule & Point of Care (six topics)

Mass Spectrometry of Proteins (six topics)

Input from Regulatory Agencies on Regulated Bioanalysis & BMV

Input from Regulatory Agencies on Immunogenicity, Biomarkers, Gene Therapy, Cell Therapy and Vaccines

Part 2 – Volume 14 Issue 10 (May)

Biomarkers & CDx Development & Validation (nine topics)

Cytometry Validation & Innovation (eight topics)

LBA Regulated Bioanalysis, Critical Reagents & Positive Controls (nine topics)

Part 3 – Volume 14 Issue 11 (June)

Gene Therapy, Cell Therapy & Vaccines (nine topics)

Immunogenicity of Biotherapeutics (eight topics)

SECTION 1 – Regulated Mass Spectrometry of Large Molecules

Surinder Kaur¹, Stephen C Alley², Matt Szapacs³, Amanda Wilson¹⁸, Eugene Ciccimaro⁴, Dian Su⁵, Neil Henderson⁷, Linzhi Chen⁶, Fabio Garofolo³³, Shawna Hengel², Wenying Jian⁸, John F Kellie⁹, Anita Lee¹⁰, John Mehl⁹, Joe Palandra¹¹, Haibo Qiu¹², Natasha Savoie³⁹, Diaa Shakleya¹³, Ludovicus Staelens¹⁴, Hiroshi Sugimoto¹⁵, Giane Sumner¹², Jan Welink¹⁶, Robert Wheller¹⁷, Y-J Xue⁴, Jianing Zeng⁴, Jinhui Zhang¹³ & Huiyu Zhou¹⁹

Authors are presented in alphabetical order of their last name, with the exception of the first 7 authors who were session chairs, working dinner facilitators and/or major contributors.

The affiliations can be found at the beginning of the article.

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 14th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 15th WRIB. The background on each issue, discussions, consensus and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

Regulated Bioanalysis/ICH M10 of Large Molecules by Mass Spectrometry

Multiple factors can potentially affect overall variability in biotherapeutic liquid chromatography mass spectrometry (LCMS) pharmacokinetic (PK) assays. Do we need specific acceptance criteria in the guidance? What does recovery mean in the context of hybrid PK assays and what is being done to evaluate this? Should the consistency of the extraction step only be evaluated if it is used in the assay (consistent with liquid chromatography [LC] versus ligand binding assay [LBA] in bioanalytical method validation [BMV])? For surrogate PK or biomarker assays, what other validation parameters not covered by BMV are suitable for tissue protein assays?

Internal Standard Selection for Hybrid Assays in Regulated Bioanalysis/ICH M10

Hybrid PK assays have now been used in approval packages in the US (Biologics License Applications [BLA] and Emergency Use Authorizations [EUA]) and Europe (Marketing Authorization Application [MAA]). Are there remaining reservations for industry or regulators regarding the suitability of hybrid assays in marketing authorizations? What are the choices for internal standards (IS) for hybrid assays (stable isotope label (SIL) peptide, winged peptide, and intact protein)? Are there situations where intact proteins have assay robustness advantages?

Regulatory Feedback on Nanomedicine Bioanalysis by Mass Spectrometry

With many drug programs there can be a large potential list of analytes to measure. What does a risk-based approach for the analytes initially measured look like in order to move the program forward?

Bioanalytical Mass Spectrometry Strategies for CRISPR Quantification

For the biodistribution of gene therapies (GTx), other than those mentioned in the draft ICH S12 guidance [29] (i.e., genetic material, DNA/RNA of the GTx product and, if appropriate, expression product) what should we be measuring in biodistribution studies? Is the measurement of the CAS9 protein and guides for clustered regularly interspaced short palindromic repeat (CRISPR) therapeutics in a biodistribution study needed? Is it necessary to understand the delivery vehicle in the biodistribution study? Is anti-drug antibody (ADA) required for the different components that will be measured in the biodistribution study?

Extracellular Vesicle Bioanalysis by Mass Spectrometry

Will interference from matrix proteins impact exosome analysis (i.e., measurement of the proteins and metabolites within exosomes)? Will recently developed exosome separation size-exclusion chromatography (SEC) kits provide enough recovery with good consistency for quantitative protein analysis? How does the exosome's heterogeneity impact separation and analytical method selection?

Immunocapture Platform Considerations for Intact Mass LCMS

Are quadratic fit calibration curves appropriate for clinical assays? In LBA, background is common and background subtraction is routine. Should the background be subtracted for LCMS? Is there a need, or a motivation, for using intact mass approaches for quantification of biotherapeutic co-mixtures?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

Regulated Bioanalysis/ICH M10 of Large Molecules by Mass Spectrometry

Hybrid assays (immunoaffinity [IA]-LCMS) have gained increasing acceptance in biopharmaceutical drug development as an alternative, or in some cases, a preferred bioanalytical approach to conventional LBA for the quantification of protein therapeutics [6,8,11,14,17,20,23,26]. This has been driven in part by the increasingly complex large molecule modalities and combination therapies in development in the industry. There was a unanimous industry consensus that hybrid methods are no longer seen as novel technologies and there have not been many questions from the health authorities for using these techniques, if appropriate. Current health authority assay validation guidelines do not directly specify requirements for hybrid assays [30–32]. However, considerations for large molecule hybrid assay development and validation criteria for PK assays have been recommended in recent industry White Papers, based on a combination of existing criteria for ligand binding and chromatographic methods [23,26]. The minimum recommended requirements included specificity of the surrogate peptide(s), digestion efficiency, LBA accuracy and precision, and 30% difference between replicate analyses incurred sample reproducibility (ISR) criteria (which should be targeted until more experience with hybrid assays has accrued). Dilutional linearity and capture reagent lot-to-lot variation should also be considered [23,26]. Furthermore, to understand the capture efficiency, matrix should be spiked with the analyte both before and after the immunocapture. For the recovery assessment, which is tied to assay reproducibility, the goals are to gain consistency of the method throughout the assay range and potentially provide a lower limit of quantitation (LLOQ), therefore optimization of the assay should be considered beyond a numerical result.

With the wider adoption of hybrid assays in nonclinical and clinical studies, it is important to ensure appropriate acceptance criteria for precision and accuracy. Due to the greater analytical complexity involved in hybrid assays compared with conventional LCMS methods, White Papers have recommended the LBA acceptance criteria of “20/25” for protein quantification instead of the “15/20” criteria for typical small molecule LCMS methods [23,26,33,34]. Additionally, consensus between the industry and the regulators has been to continue refining assay validation criteria with increasing industry experience when implementing these methods in regulated studies.

To further support these previous recommendations, data highlighting assay performance from preclinical studies and clinical studies with more than 25,000 analyzed PK samples for a recently approved therapeutic were discussed. The preclinical assay was a generic hybrid “total” assay format [35] used in all regulated studies for antibody-based biotherapeutics since 2014. Representative quality control (QC) sample performance data presented for two generic assay formats across 16 studies, over a period of three years, showed ±20% acceptance criteria were optimal. It is noteworthy that tighter ±15% performance was observed during the original assay validations [35]. Thus, if tighter ±15% acceptance criteria are implemented based on well-controlled validation experiments, performance may not translate during study sample analysis over longer periods.

Assay performance data from a specific capture format, covering five clinical studies over a three-year period, was also discussed. Both “total” and “free” format analysis [36] has been in use since 2016. Hybrid PK assay performance results showed ±20% acceptance criteria were optimal and representative of the majority of similar assays in use. Again, tighter ±15% performance was observed during assay validation; a well-controlled experiment over a much shorter time frame. Interestingly, a bispecific “XY” hybrid PK assay used in one of the clinical studies showed ±15% acceptance criteria were optimal, but this is not typically observed in similar assay types.

Therefore, based on these new data, it was confirmed that for hybrid assays, the sources of variability are more complex than typical small molecule LCMS assays, resulting in the optimal criteria of ±20% CV/bias (25% at assay LLOQ). Data gathered across the industry to date indicate that the acceptance criterion of ±15% bias is not generally applicable due to the more complex biological nature of these methods (either immunocapture and/or digestion). Furthermore, there is no push from regulators for acceptance criteria of ±15% bias. Indeed, the regulators only request that acceptance criteria are justified with data generated during assay validation and study sample analysis.

Logistically, hybrid assay development can have challenges. Standard operating procedures (SOPs) should be in place outlining hybrid assay validation procedures and criteria; SOPs pertaining to small molecule LCMS methods may have criteria that do not match those derived during hybrid method validation, resulting in inspection findings.

Internal Standard Selection for Hybrid Assays in Regulated Bioanalysis/ICH M10

For hybrid assays, which combine aspects of ligand binding and mass spectrometry (MS) assays, the choice of IS comes at an interesting intersection: internal standards are recommended for small molecule MS assays [30–32] while unusual for large molecule LBAs and therefore offer very little precedence. Prior recommendations for IS selection have been made in 2016 [14] and 2019 [23]. Current discussions are built on those recommendations and illustrate the evolving trends in IS selection, assay design, and impact on the ability to meet LBA acceptance criteria.

Reagent availability is the first consideration for hybrid assay development. Historically, “winged peptides” were seen as a potential option as an IS. More recently, their use has gone out of favor as no advantages have generally emerged using this strategy although two posters were presented at the 15th WRIB [37,38]. Currently, most assays reported using SIL-peptides and although labelled full length proteins are also used, they are less widely available. In preclinical (discovery) studies, labs are also utilizing generic IS, often another similar protein. Commercially available SILuMAb [39] has been a convenient reagent, providing a readily accessible supply of a SIL generic antibody that can be used for the quantitation of antibody therapeutics. Non-antibody large molecules do not have such a convenient IS source but SIL-peptides and SIL-therapeutics can be used as alternatives (also appropriate for use with antibody therapeutics). SIL-peptides are readily available from custom synthesis vendors and could include quantitation peptides or winged peptides with protease cleavage sites that allow the release of the quantitation peptide. SIL-therapeutics are much more difficult reagents to obtain and may require in-house production due to the lack of vendors that can consistently produce these reagents.

Assay design and optimization can help overcome inconsistencies that may arise from the choice and availability of IS. SIL-therapeutics allow better control of sample preparation including affinity capture, proteolysis, and chromatography, while general SIL-intact antibody (SILuMAb) and winged and quantitation peptides offer progressively less control over these steps. In several case studies, examples of each choice of IS were provided, demonstrating assay performance within LBA acceptance criteria after a variety of assay design and optimization efforts.

A case study was discussed where a generic hybrid assay for use in mouse serum was developed and validated in six labs. Two IS choices were compared: SIL-tryptic peptide and SILuMAb [40]. Assay development included the assessment of protease digestion efficiency. Identical peptide sequences were used for the analyte, SIL-peptide and SILuMAb measurement. Both SIL-peptide and SILuMAb IS choices led to accurate and precise assays, and the authors noted that 4 of 6 labs met chromatography acceptance criteria.

The second case study described a hybrid assay for two antibodies in clinical samples, complementarity-determining region (CDR)-tryptic peptides for trastuzumab and pertuzumab, using SILuMAb and affinity LCMS [41]. Digestion efficiency and recovery were both evaluated during assay development. The authors noted that both antibodies met LBA acceptance criteria, but the trastuzumab peptide performance was worse potentially due to wider divergence of IS retention time.

A three-antibody assay (bevacizumab, nivolumab, and pembrolizumab) in clinical samples using a hybrid assay was the third case study discussed [42]. The IS utilized the addition of a fourth antibody (tocilizumab) during the pre-affinity capture step and a SIL-peptide post column. All antibody and SIL-peptides had different sequences. The use of 2 IS resulted in better performance than without an IS at all, especially for precision, due to inconsistent recovery (corrected by the antibody IS) and matrix effects across the chromatogram (corrected by the post-column SIL-peptide).

Finally, a ∼6000 Da therapeutic protein hybrid assay was evaluated for an insulin biosimilar and two metabolites using two chemically synthesized SILs for the parent and metabolites. The protein did not require proteolytic digestion [43] which is an option when a high sensitivity assay is not required. Extensive optimization of multiple steps was performed in assay development to devise an assay meeting LBA acceptance criteria for validation and sample analysis. Assay performance was well within LBA acceptance criteria and thousands of clinical samples were analyzed with an ISR passing rate of 94.3%.

The use of an IS is considered a requirement for PK assays where the reference standard is the same as the analyte. For biomarkers, the use a recombinant protein as a standard is allowed if available to measure an endogenous analyte (potentially with post-translational modification (PTM), see Section 4). A thorough understanding and optimization of the digestion step are needed for all assays. Finally, it was recommended to include the rationale for the selection of the assay strategy in regulatory submissions (e.g., why a platform was selected, what analytes are being measured).

Regulatory Feedback on Nanomedicine Bioanalysis by Mass Spectrometry

In recent years, there have been a number of breakthroughs producing sophisticated drug delivery initiatives. Developing strategies for targeted drug delivery is now an integral part of drug development for many programs. Nanoparticles may be a promising solution for drugs that have not progressed in development due to unfavorable drug metabolism and pharmacokinetics (DMPK) properties or may improve the Therapeutic Index (TI) and duration of action. These particles come in a range of shapes and sizes (e.g., liposomes ∼100 nm, dendrimer ∼1 nm). Currently, there is a draft regulatory guidance available for using these types of nanomaterials [44], as well as previously published White Paper recommendations [10,13,23]. Consensus was reached that for analytes such as transgene products and modified messenger RNA (mRNA), methods should meet regulatory requirements and FFP approaches seem reasonable for the measurement of the nanoparticle protein corona.

Each different delivery vehicle presents a unique set of challenges and often there is a requirement not only to understand the drug substance exposure but to also understand the relationship between the drug and its carrier. A nanomedicine is formulated by either encapsulating the drug inside the nanoparticle or linking the drug to a scaffold. The nanoparticle:drug moiety is administered to the patient, leading to the encapsulated or linked drug and released drug moiety in circulation. It was agreed that bioanalytical methods must then be developed to measure encapsulated or linked drug and free drug. Nanomedicines are designed to “leak” the drug; therefore, sample handling should be designed to minimize diffusion to prevent unwanted release when measuring free drug. Another challenge is that, often, relative total and released drug ranges can be very different resulting in a small amount of unwanted drug released from the nanomedicine having a significant impact on the ability to accurately measure free drug.

New data for a dendrimer macromolecule containing lysine branched nanoparticles highlighted the increased experience with these delivery molecules. It was used to discuss the ongoing bioanalytical challenges and strategies when measuring drug exposure and examining what the body does to the dosed delivery molecules. Biodegradable chemistries allow for the release of the active drug in aqueous conditions. The therapeutic could contain enough drug in a single administration to last for 1 month, with release that provides the right efficacy and safety profile. For a dendrimer, two assays are developed: one for total drug and one for the released drug. The total assay calibration standards can be prepared with the drug or the dendrimer, if needed to account for the incomplete release of the drug. Samples can be treated with hydrazine or sodium hydroxide prior to a LLE or protein precipitation extraction and injection on an LCMS. For the measurement of the released drug, calibration samples are prepared using the released drug and LLE release should be assessed in vitro to identify optimal sample handling conditions that minimize unwanted release. In the highlighted case study, the final processing instructions required that the pH was adjusted with citric acid and that samples were kept on ice and processed within 1 hour of thawing. Freeze-thaw experiments also determined that only a single freeze-thaw cycle was permitted. It is recommended that the freeze-thaw assessment is performed by spiking QC samples with dendrimer containing a known level of released drug (∼0.1%). These QCs should also be used to monitor for control of unwanted release during sample extraction. Methods for the measurement of total and released drug were successfully validated for use in a preclinical study. The lessons learned from the preclinical validation were then applied to the method used for clinical studies, where multiple aliquots of samples were prepared to mitigate freeze-thaw instability, and samples were kept on ice and processed within the required time frame.

mRNA is inherently unstable in the presence of nucleases (e.g., high in blood), so protection strategies may need to be employed for successful drug dosing. One of these strategies could be the use of lipid nanoparticles (LNP) [45]. LNPs are known to elicit ADA/inflammatory responses that can potentially affect PK/pharmacodynamics (PD) of a dosed modified mRNA and/or the in vivo expressed biologic (IVEB) translated from the mRNA. Some studies have shown that through the functionalised incorporation of anti-inflammatory agents into the LNPs, the inflammatory impact upon PK/PD can be negated. The bioanalysis of an LNP typically uses a surrogate such as the cationic lipid component; however, an incorporated anti-inflammatory agent (or functionalised moiety) could be used as an alternative surrogate. When assessing the cationic lipid, it is important to overcome analyte instability; samples should be collected and processed using acidification, and all work should be performed on wet ice [13,23]. Furthermore, for tissue samples, frozen tissue bead beating homogenisation should be used to avoid excessive heat generation. Table 1 provides an overview of analytes from dosed LNP-modified mRNA and recommended bioanalytical approaches. Table 2 summarizes the bioanalytical validation requirements at different stages of drug development.

Upon entering the body, nanomaterials circulating in the blood can interact with multiple plasma proteins to form a biomolecular corona, otherwise known as a protein corona. Draft regulatory guidance requires the determination of “the major binding proteins involved in the formation of the corona over time” for the impact on cellular uptake and safety [44]. To perform cross-comparative protein corona evaluations, the industry may need to apply a consistent framework of experimental design [46]. In the case study discussed, two nearly identical LPNs were evaluated; their only difference was in the length of the cationic lipid. The different LNPs were incubated in human plasma for different lengths of time (e.g., 10 min and 60 min), prior to sample processing, first using SEC, then with sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) before being injected on an LCMS. Results demonstrated that the protein coronas that were formed differed substantially in the types of plasma proteins interacting with the LNP surface. Moreover, the composition of the protein corona on either LNP appeared to change over the time period investigated, emphasizing the need for harmonized framework for performing these types of evaluations across the industry (e.g., plasma type used, ex vivo incubation conditions and duration of experiments). Protein corona development has been shown to rearrange the structure and composition of LNPs (e.g., Apolipoprotein E), therefore understanding the protein “coronasome” of nanoparticles could lead to safer and potentially more effective therapeutic outcomes [47].

It was agreed that currently, analyte measured, assay choice and qualification/validation status are driven by intended purpose of the data generated; indeed, appropriate steps may be needed to collect and store samples for retrospective bioanalysis should regulators push back at a later date, rather than taking the “measure it all” approach (e.g., for a therapeutic dosed in a LNP, track the LNP, protein produced, ADA to the produced protein, etc.) which may be prohibitive from a resource perspective. However, continued discussions and experience are still needed to build an informed body of evidence.

Bioanalytical Mass Spectrometry Strategies for CRISPR Quantification

To fully understand the editing efficacy in therapeutic genome editing projects, one must assess the PK of several components in the editing machinery and their temporal relationship, which requires a complex bioanalytical strategy using multiple technologies. Genome editing using CRISPR has great potential to treat many diseases by using the ribonucleic protein (RNP) editing machinery comprising of CRISPR associated protein (Cas protein) and a small guide RNA, to revert the mutation. The Cas protein is an enzyme that acts as molecular scissors and when combined with the guide RNA, it is directed to the site of edit.

Like other therapeutic proteins, Cas protein can be analyzed either by using LBA, LCMS or a hybrid approach. Isolated from different bacterial sources, Streptococcus pyogenes (Sp) and Staphylococcus aureus (Sa) Cas9 proteins are two of the major types currently being explored as CRISPR therapeutics. The sequence homology between Cas proteins is less than 20%. The measurement of intact Cas protein can be challenging and can be limited by the availability of suitable and specific LBA reagents. LCMS offers an alternative approach for measuring Cas9 and can allow for the understanding of intact Cas9 protein kinetics. Previous recommendations for CRISPR/Cas9 genome editing bioanalysis [25] were updated with new data from industry experience in the discovery setting, where a flexible measurement approach to quantify Cas9 is required. In the case study discussed, limited commercial kits and reagents were available and bespoke reagent generation was not available in time. A LBA assay was developed using two commercially available N-terminal antibodies for both capture and detection, since the commercially available C-terminal antibodies were not optimal, limiting assay recovery and sensitivity. Selectivity using the N-terminal approach was not optimal (both intact and truncated protein was detectable) and, if bound, there was the potential for guide RNA interference. Therefore, a hybrid IA-LCMS approach was considered using an N-terminal capture antibody with C-terminal peptide monitoring after protease enzyme digestion and an action plan was developed to: 1) identify a surrogate C-terminal peptide; 2) establish a fit-for-purpose (FFP) workflow; 3) optimise the immunocapture LCMS method; and, 4) measure protein concentration in samples derived from both in vitro and in vivo studies.

A Basic Local Alignment Search Tool (BLAST) search can be used to identify a peptide unique to SaCas9, located towards the C-terminus and not found in the proteome of relevant model organisms (preclinical species and human) or BSA (used in the IA protocol). Protease choice should consider the avoidance of methionine and cysteine due to the variation of oxidation state or alkylation. The peptide of choice should have good properties to provide high peak area and a good peak shape.

Following peptide selection, the hybrid assay workflow can be optimized. For the IA enrichment step, the diluent, dilution factor, additives, antibody selected and matrix effects should be considered. The subsequent digestion can be performed using an elution or on bead digestion, but the choice and amount of protease is important, as well as the introduction (or not) of alkylation. Finally, based on the surrogate peptide choice, the LCMS conditions should be optimized.

It was agreed that CRISPR technology offers an opportunity to treat diseases in vivo, especially those that are dependent on single point mutations. Complex bioanalytical strategies are required to fully characterize CRISPR therapeutics whereby there are several potential components to measure (Cas9 protein, guide RNA, delivery vector, genome edit and the downstream transgene product). FFP discovery approaches are needed for multiple component understanding, and currently the majority of in vivo-delivered CRISPR therapy studies are in the realm of early-stage discovery work with the goal of building out an optimal bioanalytical strategy.

Extracellular Vesicle Bioanalysis by Mass Spectrometry

Initial recommendations on the potential clinical utility of extracellular vesicles (EV) were issued in the 2020 White Paper in Bioanalysis [26,27]. The discussion primarily focused on preanalytical variables, EV isolation methods, application of hybrid and flow cytometry assays for EV quantification, and their impact on EV biomarker validation. The potential to use EV liquid biopsies as a surrogate biomarker of tissue disease and response to treatment drew significant attention and enthusiasm due to the challenges of tissue analysis itself. However, small particle sizes, interference from other matrix proteins and reproducible separation are a few key technical challenges for consistent and robust EV bioanalytical analysis. In addition, EVs separated from plasma, serum and other biofluids cannot be differentiated by their tissue or cell origin [48]. There was consensus that this added challenge may further complicate data analysis.

Current discussions compared the utility of multiple EV separation methods for quantitative MS analysis based on literature and internal experiences. In addition, the possibility of using targeted proteomics was discussed to quantify big panels of proteins related to disease and treatment responses for exosome samples [49]. It was agreed that, although hybrid assays are highly specific and sensitive, one drawback is that they can only be used to quantify proteins when critical reagents are available. Targeted proteomics expands the possibility to quantify proteins without the need for capture reagents. Successful application of targeted proteomics can accelerate biomarker selection to enable more rapid validation of hypotheses.

It was confirmed that common methods for isolating exosomes include ultracentrifugation, ultrafiltration, density gradient centrifugation, affinity isolation, precipitation and SEC. SEC for exosome isolation has become more prevalent for MS analysis [50]. In theory, SEC methods preserve vesicle structure, integrity, and biological activity while potentially providing high purity, high yield, at a relatively inexpensive cost. This approach allows for fast turnaround times, taking 10 to 20 minutes per sample. SEC efficiently removes over 95% of blood proteins, which makes it clinically applicable when compared with other techniques. The removal of most of the blood proteins (e.g., albumin) is also critical for LCMS-based proteomic studies. Finally, there is no polymer interference for MS analysis. More and more commercially available separation kits and tools can help the throughput and standardization of the EV separation process for possible use in clinical trials and beyond. The process cannot differentiate exosome subpopulations through size therefore the current recommendation is to combine all relevant fractions. Sucrose density gradient centrifugation allows separation of the subpopulations through bottom load only, each having a different size distribution and distinct molecular and biological properties [51].

However, consistent preparation of exosomes is still challenging for protein quantification. Although a significant amount of plasma proteins can be separated from the exosomes, matrix proteins are typically still quite abundant in the enriched exosome samples and data interpretation should take matrix proteins into consideration. Furthermore, protein recovery in the exosome samples can be relatively low compared to that in plasma and formalin-fixed, paraffin-embedded (FFPE) tissue samples.

It was agreed that the science in this field continues to evolve in the direction recommended by the 2020 White Paper in bioanalysis [26,27]. More organizations are exploring this methodology with positive results, but it is still an evolving field with a growing body of data.

Immunocapture Platform Considerations for Intact Mass LCMS

Intact mass approaches for the bioanalysis of biotherapeutics enables quantitative measurement while simultaneously providing important information about drug biotransformation following dosing. Information about the ex vivo and in vivo fate of biotherapeutics can help with candidate selection and risk mitigation. Immunocapture is an important step in the sample preparation process for intact mass analysis of therapeutics from a number of biological matrices. Various immunocapture platforms can be applied. In contrast to multiple-reaction monitoring (MRM), intact mass analysis does not rely on tandem MS to reduce spectral background and therefore sample preparation using immunocapture combined with LC separation is the primary mode for overcoming background interference [20,23,26].

The discussion focused on issuing recommendations for determining optimal platform selection for quantitative performance. Plate-based, bead-based, tip-based and cartridge-based platforms were evaluated for the immunocapture of biotherapeutics from serum. As part of this evaluation, the impact of bead quantity on assay performance and figures of merit were investigated. The evaluated platforms can each provide advantages for intact mass analysis, which may be beneficial depending on sensitivity, dynamic range, sample volume, loading capacity and platform flexibility. However, a trade-off between dynamic range and sensitivity was observed; the higher sensitivity platforms produced lower background signal and superior LLOQ but resulted in more limited dynamic range and upper limit of quantitation (ULOQ). This trade-off can limit the use of a single set of immunocapture conditions to simultaneously achieve low LLOQ and wide dynamic range. Therefore, laboratories may need to consider investing in multiple platforms.

When evaluating the standard fit, the simplest model should be used. It was agreed that given the immunocapture step, a quadratic fit may be appropriate in some cases. There was consensus that the data should guide the selection of the curve fit. A final recommendation was that considerations on background or ghost peak issues (similar to having an interference peak at the LLOQ) may warrant a change of the capture antibody or further optimization of the chromatography, if possible. Table 3 summarizes overall results for key parameters.

Intact mass approaches can also be applied to the quantification of biotherapeutic mixtures in serum. Biotherapeutic mixture analysis is an important application because of the expanded use of co-dosed therapies applied for disease treatment. High-capacity capture approaches are best suited for co-mixture bioanalysis, as low capacity approaches would limit the dynamic range for capturing multiple antibodies and gaining meaningful measurement of their modifications. Intact mass immunocapture can be used for combination therapies if appropriate, although patient specific differences in biotransformation may be seen.

RECOMMENDATIONS

Below is a summary of the recommendations made during the 15th WRIB:

Regulated Bioanalysis/ICH M10 of Large Molecules by Mass Spectrometry

• Hybrid assays are no longer seen as novel technologies. They have been used in Regulated studies and have been accepted by global health authorities, if appropriately validated and scientifically justified.

• For hybrid PK methods the sources of variability are more complex than for small molecule LCMS, resulting in the acceptance criteria of ±20% CV/bias (25% at assay LLOQ) as per 2019 and 2020 White Paper in Bioanalysis recommendations [23,26]:

  	○	Data gathered across industry to date indicate that the acceptance criteria of ±15% bias is not generally applicable due to the biological nature of these methods (either immunocapture and/or digestion).
  	○	There is no push from regulators for acceptance criteria of ±15% bias. The regulators only request that acceptance criteria are justified with data generated during the assay validation and study sample analysis.

• SOPs should be in place outlining hybrid assay validation procedures and criteria; SOPs pertaining to small molecule LCMS methods may have criteria that do not match those derived during the hybrid method validation, and result in inspection findings.

• It is important to demonstrate consistent immunocapture recovery throughout the assay range and not to establish a numerical recovery parameter.

Internal Standard Selection for Hybrid Assays in Regulated Bioanalysis/ICH M10

• Historically, “winged peptides” were seen as a potential option as an IS. More recently their use has gone out of favor as no advantage generally emerged with this strategy.

• Currently, most assays use SIL-peptides or, if available, labelled full length proteins. In preclinical (discovery) studies labs are also utilizing generic internal standards.

  	○	A reliable and scientifically justified IS is considered a “must have” in PK assays used in regulated bioanalysis.
  	○	A thorough understanding/optimization of the digestion step is needed for all hybrid assays used for regulated studies.

• The rationale for selection of the assay strategy should be included in regulatory submissions (e.g., why a platform was selected, what analytes are being measured; see Section 5 for regulatory input).

Regulatory Feedback on Nanomedicine Bioanalysis by Mass Spectrometry

• For analytes such as transgene products and modified mRNA, methods should meet regulatory requirements and FFP approaches seem reasonable for measurement of the nanoparticle protein corona.

• Nanomedicine bioanalytical methods require consideration of handling procedures for the accurate and precise measurement of encapsulated or linked drug and released drug:

  	○	To overcome analyte instability, pH and/ or temperature can be used.
  	○	For tissue samples, frozen tissue bead beating homogenization should be used to avoid excessive heat generation.

• Currently, analyte measured, assay choice and qualification/validation status are driven by intended purpose of the data generated; indeed, appropriate steps to collect and store samples may be needed for retrospective bioanalysis should need arise, rather than taking the “measure it all” approach (e.g., for a therapeutic dosed in an LNP, track the LNP, protein produced, ADA to the produced protein, etc.) which may be prohibitive from a resource perspective.

• This is still an innovative field and continued discussions and experience are needed to build a body of evidence to learn from and to update the recommendations as science and regulations evolve.

Bioanalytical Mass Spectrometry Strategies for CRISPR Quantification

• In the absence of suitable commercially available LBA reagents, current data shows that hybrid IA-LCMS assays can be used to measure Cas9.

• The selection of a unique C-terminal surrogate peptide in combination with immunocapture with an N-terminal antibody is crucial for developing a successful hybrid assay for monitoring intact Cas9.

• Currently, the majority of invivo-delivered CRISPR therapies are in the realm of early stage discovery work with the goal of building an optimal bioanalysis strategy.

Extracellular Vesicle Bioanalysis by Mass Spectrometry

• Targeted proteomics expands the possibility to quantify proteins without the need for capture reagents and may be successfully applied to EV analysis.

• Recent data show that SEC kits are recommended for EV separation. However, consistency of the performance for quantitative analysis is unknown given that exosomes may have different tissue origins and are heterogeneous.

• Evaluation of the impact of matrix (selectivity assessment) can be performed as follows:

  	○	Isolate exosomes from different patients.
  	○	Use different bodily fluids and get a mixture of exosomes using ultracentrifugation and SEC.

• Since consistent preparation of exosomes is still challenging for protein quantification, data interpretation should take matrix proteins into consideration as a way for data normalization.

• Protein recovery in the exosome samples can be relatively low compared to that in plasma and FFPE tissue samples.

Immunocapture Platform Considerations for Intact Mass LCMS

• When choosing the immunocapture platform, a trade-off between dynamic range and sensitivity should be considered:

  	○	Higher sensitivity platforms produce lower background signal and superior LLOQ but result in more limited dynamic range and ULOQ.
  	○	Laboratories may need to consider investing in multiple immunocapture platforms.

• For hybrid intact mass assays, when evaluating the standard curve fit, the simplest model should be used. However, given the immunocapture step, quadratic fit may be appropriate in some cases. The data should guide the selection of the curve fit.

• Background or ghost peak issues (similar to having an interference peak at the LLOQ) may warrant consideration of changing the capture antibody.

• Intact mass immunocapture can be used for combination therapies if appropriate, although patient specific differences in biotransformation may be seen.

SECTION 2 – Endogenous Compound & Complex Methods

Jian Wang²¹, Scott Summerfield²², Olga Kavetska³⁰, Lieve Dillen²⁴, Ragu Ramanathan²⁰, Mike Baratta²³, Arindam Dasgupta¹³, Anna Edmison²⁵, Luca Ferrari²⁶, Sally Fischer¹, Daniela Fraier²⁶, Fabio Garofolo³³, Sam Haidar¹³, Kathrin Heermeier²⁷, Christopher James²⁸, Allena Ji²⁹, Lina Luo³⁰, Gustavo Mendes Lima Santos³¹, Noah Post³², Anton I Rosenbaum³⁴, Natasha Savoie³⁹, Sune Sporring³⁴, Sekhar Surapaneni⁴, Stephen Vinter³⁵, Katty Wan³⁰, Jan Welink¹⁶ & Eric Woolf³⁶

Authors are presented in alphabetical order of their last name, with the exception of the first 5 authors who were session chairs, working dinner facilitators and/or major contributors.

The affiliations can be found at the beginning of the article

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 14th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 15th WRIB. The background on each issue, discussions, consensus and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

Chiral Methods for Method Development & BMV

What approach should be applied to the development of chiral drugs and the implementation of chiral bioanalytical methods? How are early data generated compared to data generated from toxicology studies or early clinical studies where fully validated methods are used? Is the application of derivatization for chiral analysis a vanishing approach with improved MS instrumentation and sensitivity? Are chiral methods more prone to interference from concomitant medications compared to achiral methods? Are chiral methods needed to determine bioequivalence of chiral drugs? When and why can methods be transitioned from chiral to achiral analysis? Is the sum of R and S enantiomers equivalent to the chiral drug obtained by achiral methods? If one enantiomer is too insignificant (e.g., <10%) and consistent, can a switch to an achiral method be justified?

ICH M10 Section 7.1 for Endogenous Compound Quantification

In chromatographic endogenous compound methods, is it recommended to apply criteria for the relationship between endogenous levels in the matrix used for standard addition/background subtraction and achievable LLOQ, such as the ratio of endogenous analyte concentration to LLOQ concentration should be <4? Since the 2020 White Paper [26] contained no recommendation for a slope criterion for parallelism studies, although it was recommended to ensure surrogate matrix and authentic matrix behaved “similarly”, what criteria should be used to show similarity?

Tissue Analysis, Rare Matrices & Atypical Sample Collection in Regulated Bioanalysis

What are the recommendations for validation of scarce matrices when the analytical methods support exploratory endpoints versus primary or secondary endpoints? What is the best practice for tissue homogenate QC preparation supporting studies where multiple tissues may be tested, such as tumor biopsies in oncology (e.g., tissue pools, separate QCs, etc.)? What are the main challenges to bioanalytical assays for very low samples and what technologies will help?

Quantitation of Intracellular Disposition of Oligonucleotides & Sensitivity/Specificity Challenges

How is nonspecifically bound small interfering RNA (siRNA) controlled or monitored in assays for RNA-induced silencing complex (RISC)-loaded siRNA? Is normalization needed when assessing assay robustness (internal reference - miRNAs)? Given that RISC-loaded siRNA is considered an important parameter to understand PD (both potency and duration of action), how is bioanalytical quality safeguarded?

What are the recommendations for the preferred antisense oligonucleotides (ASO)/siRNA assay to support regulated work? Is it only sensitivity driven? Quantification of full-length oligonucleotides and their metabolites can be performed using different bioanalytical technologies although the selectivity of these technologies is different. In case challenging sensitivity requirements emerge during drug development, how can technologies be switched to avoid the risk of generating poorly comparable data? Is the technology suitable for application in regulatory settings? How widely is it applied in the industry?

Recent Developments of Endogenous Compounds & Fit-for-Purpose Validation

Can a “biomarker cocktail” be used to assess multiple biomarker transporters by high resolution MS (HRMS) simultaneously in the same clinical study to evaluate drug–drug interaction (DDI) risk? The complexity of renal transporter biomarkers includes both urine and plasma assay development, but parallelism establishment over various dynamic ranges for multiple analytes in both urine and plasma is challenging. Is there a need to demonstrate parallelism mathematically by using slope comparison? Are there any opportunities for HRMS implementation to small molecule biomarker assays? Can the quantification of clinically relevant lipid and metabolic biomarkers using highly multiplex, metabolomic-style methods be validated? What is an appropriate use of surrogate analytes, longitudinal QC in clinical PD bioanalysis? Can we bridge biomarker data between studies enabling meta-analyses, analytical methods, collection methods (e.g., traditional versus home sampling)? What are the key considerations?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

Chiral Methods for Method Development & BMV

Individual isomers of chiral drugs may display different PD/toxicodynamic (TD) properties or absorption, distribution, metabolism, and excretion (ADME) profiles. Therefore, it is important to understand the PD, toxicity, and ADME properties of individual isomers as well as the potential for undergoing isomerization before selecting an individual isomer or racemate for drug development. Chiral centers are potentially responsible for the formation of chiral metabolites and the activity/potency, safety, and clearance needs to be evaluated early. Often, the decision to follow metabolites is made based on in vitro metabolism data, scouting in the clinical multiple ascending dose (MAD) study or following human ADME (prior to Phase III).

The recent regulatory guidance/guidelines on developing chiral drugs provide general direction for manufacturers to identify and characterize each individual isomer and provide a justification for developing racemate drugs [52–54]. It was agreed that there is a need for updated guidance to clearly dictate when chiral assays are needed. Currently, the EU guidance [53] is stricter on chiral requirements than the US one [52]. To facilitate international submissions, harmonization is desirable. Clarification for the need for chiral assays in bioequivalence (BE) studies could be addressed in the final draft of the ICH M13 guidance, which is not yet at step 2 [55]. Currently, the FDA addresses the need for chiral assays in product-specific or PK endpoint guidance documents. Regulators mentioned that they see few chiral assays for BE studies unless different activities (PD/safety) or PK profiles are expected between enantiomers.

Discussions expanded on the recommendations first issued in the 2017 White Paper in Bioanalysis [16] and focused on considerations for developing individual isomers and various challenges in developing and validating bioanalytical methods for chiral drugs, along with case studies for successfully implementing the strategy. In order to understand the PK properties of individual isomers, it is important to develop selective and sensitive methods to quantitate individual isomers and enable characterization of their PK and disposition properties in nonclinical and clinical studies. It was agreed that even with an enantiomer pure compound, chiral conversion should be addressed with a need for chiral separation in relation to differences in potency and/or clearance between the enantiomers. Typically, achiral assays are used as the primary validated assay and chiral assays are qualified for further understanding. This decision is multifunctional, driven by the internal decisions based on the understanding of the compound's PK, disposition and elimination and often followed by scientific consultations or during planned interactions with health authorities.

It was confirmed that chiral bioanalytical methods used to quantitate individual isomers are much more difficult to develop than achiral methods [16]. The coupling of chiral chromatographic methods with MS is often challenging and requires considerable method development efforts and time to validate sensitive, selective, and reproducible assays. Special requirements for chiral separation include the high resolving power for baseline resolution of chiral drugs and their metabolites in order to mitigate possible ion suppression with co-eluting enantiomers; compatibility with aqueous mobile phases or buffer systems that are amendable for MS detection; cleaner samples for maintaining the quality of separations and column life; high sensitivity (low ng/ml or even pg/ml) for highly potent chiral drugs, and, a reasonable separation time (within 10 min). In addition, chiral inversion (e.g., pH-dependent inter-conversion) may require stability assessments and stabilization following sample collection [56,57]. Overall, chiral assays are substantially more resource and time consuming which impacts clinical trial execution timelines and drug development time.

Given the challenges with developing chiral methods, a tiered approach to their development and degree of validation was recommended. Using a tiered approach, a qualified chiral assay could be used for discovery and exploratory study support, which could later be validated for the GLP toxicology and clinical trials if required. Once the in vivo inversion is fully characterized, the method could move to a validated achiral assay for long-term use. However, chiral assays may still be needed for metabolites that significantly contribute to the efficacy and safety profile.

Diastereomers do not require chiral chromatographic conditions for their separation. The majority of these compounds have multiple fixed chiral centers, thus, inversion at one center results in the formation of diastereoisomers. Similarly, if an additional chiral center is formed via metabolism in a molecule with one chiral center, stereoisomers of the metabolite will be diastereomeric. Conventional chromatographic conditions used for metabolite profiling should consider this point and profiling runs should be long enough to take into account the need to look for diastereomeric metabolites.

Many techniques are available for chiral separation. Gas-phase chromatography is suitable for volatile compounds but has largely been replaced by high performance liquid chromatography (HPLC), supercritical fluid chromatography (SFC) and capillary electrophoresis (CE). There are, however, few laboratories with non-LC development experience. LCMS is the most widely used, meeting almost all the demands of chiral bioanalysis due to its high sensitivity, specificity, automation and good reproducibility. LCMS is well suited for aqueous biological samples with no required derivatization. Most LC is transitioning to ultra-performance liquid chromatography (UPLC) with higher flow rates.

Furthermore, there are various separation modes with extensive column selections. These include protein-based columns used for broad selectivity and suitable for reverse phase mobile phases; polysaccharide-based columns (e.g., Chiralcel and Chiralpak series) suitable for both normal and reverse phase chromatography; macrocyclic glyco-peptide-based columns (e.g., Chirobiotic V and T series) suitable for normal and reverse phase chromatography and polar organic modes with promising chiral selectors having long-term stability; and, cyclodextrin-based columns (i.e., native, derivatized and aromatically derivatized columns) also suitable for normal and reverse phase chromatography and polar organic modes. LC chiral column screening and selection is the most challenging component of chiral LCMS method development. It is a time consuming and expensive “trial and error” screening process because it is hard to predict the retention behavior of compounds on chiral stationary phases. Chiral columns are generally less durable and stable than regular columns; any accumulation of endogenous components can cause the merging of two chiral peaks. Extensive sample clean-up using SLE, LLE or SPE is recommended; protein precipitation is not recommended. The use of a pre-column trap (short reverse phase column) in a linear or 2D configuration to protect the chiral column can extend the life of the column as well as improve assay robustness.

Mobile phase selection is another important part of chiral assay development. Reversed-phase liquid chromatography (RPLC) is suitable for the majority of chiral separations because it is easy to execute and amendable to MS detection. Polar organic LC (e.g., methanol, ethanol, hexane) is ideally suited for large chiral molecules. Normal-phase LC offers superior separation than RPLC but it is less frequently used due to it being less amendable to MS; post-column solvent addition for ionization is needed. Mobile phase ionic strength, pH values, and the use of modifiers can impact chiral separation and need to be optimized.

ICH M10 Section 7.1 for Endogenous Compound Quantification

There have been multiple challenges with overcoming high endogenous analyte concentrations in authentic matrix standard curves when quantifying biomarkers in biological samples. It is recommended to have the same matrix for the standard curve as the study samples [30–32].

In general, it was confirmed that authentic matrix lots with lower endogenous levels should be selected [19]. Variability of endogenous levels can be a challenge when selecting matrix lots. Additional recommendations have also been published for the analysis of endogenous compounds as planned by the ICH M10 BMV draft guideline [26,32]. Section 7.1 of the ICH M10 BMV draft guideline proposes four approaches for assaying endogenous analytes: 1) standard addition; 2) background subtraction; 3) surrogate analyte; and, 4) surrogate matrix [32]. The standard addition approach involves constructing a calibration curve for each sample by adding increasing concentrations of standard solutions in to aliquots of the study sample. The sample concentration is determined from the intercept of the X-axis. A surrogate analyte (e.g., SIL-analyte) may be selected to spike the standard curve if it is demonstrated that the surrogate analyte behaves similarly to the analyte of interest [32]. In the background subtraction approach, a series of standards are spiked into authentic matrix containing an endogenous analyte level. The endogenous level of the matrix is assessed by extracting at least 3 replicates (with IS) and the result subtracted from the standard curve. All study sample concentrations can then be measured using the regression equation. To facilitate this procedure, a laboratory information management system (LIMS) can be leveraged in two ways to perform the calculations. The first approach is appropriate if the IS peak response is consistent throughout the run. After subtracting the average endogenous peak area of blank (with IS) from the spiked standards and spiked QCs, the original peak areas are replaced with the modified peak areas and then the data is imported into the LIMS for data regression. No modifications to study samples are required as they are not spiked samples. In the second approach, suitable for various conditions even when IS area is inconsistent but tracks, the corrected area is calculated from the subtracted peak area ratio (spiked standards or QCs) multiplied by its corresponding IS peak area. This is not required for study samples.

When trying to establish a new method using authentic matrix, it is hard to know what assay LLOQ will work for the available matrix lots. It was suggested to screen at least 15–20 lots of normal matrix samples to estimate the range using a surrogate matrix standard curve (regardless of parallelism). The ratio of the endogenous level to the desired LLOQ should be very low (e.g., less than 4); the lower the ratio, the better to achieve the desired LLOQ. Furthermore, when the ratio of endogenous/LLOQ exceeds 4, it can be hard to distinguish between the spiked LLOQ and the blank endogenous level as the difference is within the allowed variability of the signal. If a low ratio of endogenous level/LLOQ cannot be attained, then the LLOQ can be adjusted to a higher concentration that would be appropriate for the study.

If the use of authentic matrix is not an option, either because the endogenous concentration does not meet the sensitivity needs of the assay or because of the limited quantities of the matrix, the fourth approach recommended in Section 7.1. of the ICH M10 draft guideline [32] should be considered: a surrogate matrix approach. In this approach, both parallelism and matrix effect should be assessed. Often if this approach is used, the standards are prepared in the surrogate matrix, and QCs are prepared using the standard addition or background subtraction methods.

Relative matrix effect determines if spiked analyte can be added to the endogenous analyte and accurately recovered from different lots of authentic matrix using a surrogate matrix standard curve for regression.

It is suggested that the parallelism between the two matrices should be evaluated to be similar by screening at least 15 normal biological matrix lots for estimation of the concentration range of the endogenous analyte using the standard curve in a surrogate matrix or the endogenous concentration as measured by extrapolation of the standard addition calibration to the authentic matrix prior to assessment of parallelism [13]. A slope criterion has previously been suggested to assess parallelism for LCMS methods [26] when screening surrogate matrices. It is based on a mathematic derivation, linear response between instrument response and analyte concentration in LCMS assays and the ±15% precision and accuracy criteria (%difference between surrogate matrix and authentic matrix slopes should be within -13% to 17.6%). Two case studies were discussed where this approach was successfully used in both CSF and plasma, and the subsequent validation met all appropriate criteria. However, no consensus on the use of the slope criterion was achieved, with an agreement to continue the evaluation of this criterion. The extent of the validation work should be consistent with the purpose of quantification.

If neither authentic matrix nor surrogate matrix is suitable for use, dilution of the authentic matrix with a suitable surrogate matrix might be an alternative solution for the standard curve if supported by a strong scientific justification. The assessment of parallelism with the dilution QC is especially important when the reduction of endogenous compound level is expected in the disease condition.

Tissue Analysis, Rare Matrices & Atypical Sample Collection in Regulated Bioanalysis

A little over two decades ago, tandem mass spectrometry disrupted the field of quantitative bioanalysis, rapidly replacing LC-ultraviolet as the dominant platform for PK assays. The key drivers for this change were the combination of high selectivity and sensitivity, and now MS sensitivity is often sufficient to facilitate drug extraction from small sample volumes (low microliter range). Despite significant advancements in detection sensitivity, there are still bioanalytical challenges to resolve. Many small volume samples are derived from patient tissue biopsies (e.g., tumors), liver tissue (e.g., fine-needle aspiration), blood/plasma microsampling or non-blood derived biofluids (e.g., tears). Often the analyte quantification is highly informative for drug discovery and development decision-making and often allows for a better understanding of the drug disposition in the targeted organ or tissue and tissue/blood ratio. It is challenging, however, to ensure consistency of such sampling and homogeneity of the biospecimen. Assessing the key characteristics of assay performance is therefore very important while keeping in mind that the control matrices are often difficult to source in significant quantities for method development, validation and study support (e.g., spiked QCs). The detection response of some analytes may also be low which further increases the analytical challenge. Hence analyte quantification from rare and small volume matrices greatly tests the ingenuity and resourcefulness of bioanalytical scientists.

New case studies were discussed in reference to previously issued recommendations on tissue analysis and rare matrices [13,16]. In the first case study, an LCMS method validated in peripheral blood mononuclear cells (PBMCs) was adapted in order to measure the analyte at the site of action to confirm tissue penetration and modulation of pathways. The new method needed to be suitable for multiple tissue types, using an approximate biopsy amount of 10 mg. Since the control matrices were scarce, the original assay needed to be adapted for potentially smaller sample sizes. Two options for tissue homogenization and protein extraction were evaluated. In the first, tissue samples were treated with beads and lysis buffer using an Omni BeadRuptor, which is suitable for the sample size but typically generates high heat, which is hard to dissipate. Three tissue types were assayed in duplicate using LCMS, and the results did not demonstrate adequate precision. Therefore, the GentleMACS technology was attempted. The conditions created with this instrument are less harsh, but it is typically used for larger tissue samples (>300 mg). However, even with the smaller sample amounts, duplicate analyses for all three tissues were precise. Therefore, a surrogate matrix of 1 mg/ml bovine serum albumin (BSA) in lysate buffer was used for the standards and a QC homogenate containing the four most likely tissue types was used as the QC matrix. Once sample extraction was optimized, the assay was successfully used to measure target engagement in a clinical study. It was recommended that each tissue type be assessed; a QC mix like the one in this example may be suitable if the clinical team is made aware of what the validation process can or cannot test so that results are properly interpreted.

In the second case study, atypical sample collection methods were under discussion with the study team, and the impact of the method on the assay was assessed. For example, collection of tears or bronchoalveolar lavage effluent could theoretically provide a quantitative measurement if normalisation (by gravimetric analysis or a parallel measurement) was feasible. The biofluid was to be collected onto an absorbent substrate, retaining a small volume of aqueous sample. It was determined that there was a dependency on both the collected volume and the measured concentration, therefore only a qualitative measurement was achievable.

Rare or tissue matrices are most likely to be associated with exploratory endpoints, so FFP approaches are suitable and highly recommended (see also Section 4). For primary and secondary endpoints, the level of validation should be agreed upon with end users to ensure awareness of any elements that differ from full validation, so results are not over-interpreted. Regulators recommended that when working with rare matrices, consultations with health authorities should be done early on to clarify the kind of data that will be submitted based on assay feasibility. Mixed matrices can be used (e.g., standard in plasma or abundant tissue and QCs in specific tissue homogenate) or an approach where homogenates are diluted in plasma may be appropriate when multiple matrices are needed. For rare matrices, there are currently challenges obtaining cerebrospinal fluid (CSF); two mitigation plans were proposed: (1) using a consistent ratio of CSF:plasma in all samples/QCs, and (2) increasing assay sensitivity to dilute out nearly all CSF with plasma. Adsorption challenges with CSF should be considered. However, the use of plasma is a case-by-case decision, depending on the biomarker levels in plasma and CSF.

There are many emerging technologies that allow for the reduction of matrix usage. Acoustic droplet ejection (ADE) can be used to miniaturize an assay by handling low sample volumes (typically nanoliters or picoliters) without any physical contact [58]. It creates 200–500 consistently-sized droplets/second from any well to any other well on a plate. There is no impact on sample integrity because of the low energy needed for transfer. The technology can also adjust for many volumes and viscosities, therefore can be used for many matrix types from plasma to tissue homogenates. The advantages of ADE include improved accuracy and precision, no carryover, and reduced sample volumes. A case was discussed where this technology was coupled to a process beginning with manual pipetting, followed by an automated decapper and Bravo instrument to reformat the samples into a 384-well plate. The ADE then added the assay buffer to the polymerase chain reaction (PCR) plate, as well as added the standards and QC samples and unknowns. The volume on the PCR plate was normalized to 1 μl by the ADE before analysis by the Gyrolab. Another automatic dispensing technology suggested was the Tecan D300e, which dispensed standards directly into matrix for the assay without needed to perform serial dilutions. Miniaturizing bioanalytical methods results in potential saving of rare matrices. Challenges include (1) ensuring homogeneous mixing of small sample volumes (acoustic mixing may be suitable); (2) spiking small volumes of standard to avoid high % organic; (3) liquid handling, transfer of small volumes; (4) sensitivity of the assay.

Quantitation of Intracellular Disposition of Oligonucleotides & Sensitivity/Specificity Challenges

There are several oligonucleotide modalities being investigated as therapeutics, such as ASOs, siRNAs, micro RNAs, short hairpin RNAs, etc. [19,59,60]. Therapeutic oligonucleotides that silence their target mRNA transcript require cellular uptake, making this a key point for assessing activity. Following cellular uptake, these oligonucleotides are trapped in the endosomal compartment; subsequent release into the cytosol and/or nucleus is mandatory to access the silencing machinery. Two major silencing mechanisms of action were described based on: 1) RNaseH1-mediated degradation for ASOs, or 2) AGO2-mediated degradation for siRNA (the AGO2 protein is a nuclease enzyme that is part of the RISC) [61,62].

Selective delivery of oligonucleotides into the tissue of interest can be enhanced by conjugation strategies, especially for siRNAs which show little cell penetration. The N-acetylgalactosamine (GalNAc) moiety has been introduced as a liver targeting approach since it binds with high affinity to the asialoglycoprotein receptor (ASGPr). This receptor is almost exclusively expressed on hepatocytes. GalNAc conjugated oligonucleotides have been shown to increase liver exposure and to knock down the target mRNAs at reduced dose in the clinic [63–66].

Investigations on two major aspects of intracellular oligonucleotide delivery were discussed where the quantitative analysis was supported by both LC-fluorescence and LCMS assays. First, the intracellular liver uptake and the function of the ASGPr were examined. Both in vitro (in combination with transporter inhibitors) and in vivo (wild type and ASGPr-knock out animals) experiments provided additional insights on the impact of this receptor on intracellular drug disposition and more in general on the pharmacokinetics of these drugs. Assays to evaluate the abundance of ASPGr in cells and/or animal models will add additional insight in the role of the receptor and the relation to the intracellular concentrations.

Secondly, to further elucidate the details of the intracellular disposition, it is key to understand endosomal release and cytosolic delivery. In vitro experiments to enhance endosomal release to allow better exploration of downstream RISC binding were designed. RISC binding of siRNAs was investigated following immunoprecipitation of the RISC complex by AGO2 antibodies. Mitigation of nonspecific binding, stability and (un)availability of critical reagents are important to evaluate [67–70]. The nonspecific binding can be to the beads (spike siRNA to beads in absence of antibody) and the antibody. Therefore, it was recommended to include a non-specific antibody or control IgG.

The RISC binding of the antisense strands of siRNAs is critical for target mRNA degradation and therefore considered to provide a better PK/PD relationship compared to total tissue concentrations. Typically, RISC loaded siRNA are in the low ng/g levels in the presence of high μg/g total tissue levels (<0.1% loaded). In addition, phosphorylation of the 5′ nt of the antisense strand is required to bind to AGO2, the catalytic unit in RISC, responsible for endonuclease activity. Analysis of the 5′P of the antisense strand was proposed as a surrogate marker for cytosolic released siRNA and therefore was included in some investigations. Furthermore, endogenous miRNAs have been proposed (miR122 in liver) for normalization. However, understanding RISC saturation and potential displacement of endogenous miRNAs is still part of ongoing investigations and might limit the use of endogenous miRNAs for normalization.

The sensitivity of LCMS methods for quantification of ASOs has significantly improved recently and reaching sub-ng/ml LLOQ is now feasible. This is mainly due to improved MS equipment, consolidated extraction techniques (using LLE and/or ion exchange SPE) and the adoption of state-of-the-art chromatographic technologies (ion pair-reversed phase chromatography incorporating hexafluoroisopropanol (HFIP) or other mobile phase modifiers). LCMS sensitivity is typically sufficient to support preclinical studies, particularly for analysis of tissues where the ASOs exhibit their efficacy and are present at high concentrations. However, sensitivity might not be sufficient in the clinical phase where the administered doses and the observed exposures are expected to be lower. Sensitivity can also be challenging in certain matrices, e.g., in plasma or serum when the administration of ASOs is not systemic.

Bioanalytical assay selection should be driven by the stage of development, assay development time and the need of critical reagents. It was noted that there is currently limited experience with quantitative PCR (qPCR) for regulated work. In all cases defined above, switching technology (e.g., LCMS to hybridization enzyme-linked immunosorbent assay ELISA [hELISA]) in order to meet a sensitivity target should be carefully evaluated because the selectivity of these assays can be different and a direct comparison of the obtained exposures may not be possible. Long analytical run times are often observed resulting in low throughput and high cost per sample. Selectivity is one of the most difficult aspect for oligonucleotide assay development, even with the advantages of LCMS, particularly in cases of co-eluting minor metabolites without truncation of nucleotides (desulfurization, defluorination, deamination). Skilled analysts are needed to mitigate these issues. If switching technologies is unavoidable, bridging experiments should be performed. It was recommended to analyze the same set of samples with both technologies to understand the relationship between them. The same results may not be provided, but there should be understanding of how the differences can impact data interpretation. Another consideration when designing bridging experiments is that if tissue samples are homogenized for one type of technology, they might not be usable for the other type.

According to the MIST guidance [71] and previous White Paper in Bioanalysis recommendations [19,23,26], metabolites should be quantified in case these significantly contribute to the pharmacological and toxicological effect. Again, the selection of the technology to use for their quantification should be very carefully evaluated because of potential selectivity issues. Additional complexity is due to the fact that the metabolic profiles of ASOs in plasma and in tissues are typically very different, resulting in different sensitivity requirements. For targets such as liver, it is not important to use the same tissue homogenate for PK/PD measurements, but in central nervous system tissue there can be large changes in PK/PD, even when sampling is done at very minimal distances for the PK and PD samples.

Considering all the above concerns and challenges, a proposed strategy is to generate preliminary metabolite identification information in serum and tissues before the quantitative technology is selected and bioanalytical methods are fully validated to support regulatory toxicological studies. The pharmacological activities of predictable shortmers should also be evaluated in order to determine what analytes require quantification. The selectivity of LCMS and hELISA assays versus shortmers should be assessed when selecting the assay platform.

Recent Developments of Endogenous Compounds & Fit-for-Purpose Validation

Quantification of endogenous biomarkers in clinical studies requires careful evaluation of accuracy, precision, selectivity, specificity as well as reproducibility [13,16,19,25]. There are many challenges with the development of multiplexed assays for small molecule endogenous analytes, such as with chromatographic separation due to different chemical properties and numerous isobaric interferences. When deciding on the appropriate cocktail of analytes, it was suggested to group them by similar chemical properties in order to facilitate combined extraction and chromatographic separation. The discussion of the selectivity/specificity of endogenous compounds is not new but still very relevant for this class of molecules because development is challenging and laborious due to numerous isobaric analogues. The establishment of different dynamic ranges based on distinct endogenous levels for each analyte is also a challenge and can induce complex standard and QC preparation procedures which can be prone to errors but ranges should be selected based on the anticipated study population.

The development of plasma lipid multiplex assays adds additional challenges, particularly due to more difficult extraction procedures [13,16,19]. These can be mitigated by using an IS and pooling lots of authentic matrix as a QC sample for every batch to monitor the assay performance. Inherent structural complexities can affect the level at which the quantification assay can differentiate between different lipids. Highly multiplex methods for structurally similar analytes present several bioanalytical challenges due to lack of readily available authentic standards or matching SIL-IS for all analytes of interest. Furthermore, there can be a large variation in concentration between lipid species, adding complexity to the development of the method. Solutions proposed include use of a surrogate analyte (or analytes) to preliminarily define the linear range, use of surrogate matrix to define dilutional linearity for each endogenous analyte and use of a fold-change between samples as the relative quantification method instead of relying on a standard curve within intended context of use.

The characterization of the performance of a multiplex bioanalytical method for quantification of greater than 30 phosphatidylinositols (PI), a class of low-abundance plasma phospholipids, was discussed [72]. PI can considerably impact the charge-dependent interactions of high-density lipoprotein with lipases and other proteins since they are negatively charged phospholipids. Consequently, a thorough description of the quantitative properties of a bioanalytical method for the measurement of human plasma PI is important [72]. The odd-chain PI species that are not normally present in human plasma were used as surrogate analytes due to high levels of endogenous PI levels to assess assay performance and establish the dynamic linear range for PI lipids [72]. A high throughput method was developed, qualified, transferred to an automation platform and applied to sample testing in two clinical trials in healthy and stable coronary artery disease (CAD) subjects [73,74]. The method demonstrated acceptable precision and accuracy (±30%) over a linear range of 1–1000 nM for the surrogate analyte and dilutional linearity (8-fold) for endogenous PI. To correct for batch effects, the Systematic Error Removal using Random Forest (SERRF) normalization algorithm was employed [75]. Moreover, SERRF was employed to bridge the raw values between the two clinical studies, enabling quantitative comparison of absolute values. The comparison of the two studies revealed that healthy subject levels of PI are consistently higher across PI species compared to CAD subjects [73–75]. The discussion based on this case study showed the importance of the approaches used for the development, characterization and FFP validation of bioanalytical methods for the quantification of clinically-relevant lipid and metabolic biomarkers. In fact, these highly multiplex (>30 analytes), metabolomic-style methods were not suitable for a full validation as prescribed by regulatory guidance [30,31]. Furthermore, this work demonstrated how advancements in several areas of technology such as LC, MS, automation and data processing could enable discovery of clinically relevant biomarkers and FFP validation for such multiplex methods.

There is increased interest in using endogenous probes, in conjunction with DDI decision trees from the regulatory agencies, to discharge DDI risk quickly in early clinical studies. Recommended organic cation transporters (OCT1 & OCT2) and multidrug and toxin extrusion protein (MATEs) DDI risk thresholds are known to be conservative, and high false positive rates are reported [76–78]. Previous work demonstrated that N¹-methylnicotinamide (1-NMN) could be a useful endogenous biomarker of OCT2 or MATEs mediated DDIs [79,80], however, current investigations also revealed 1-NMN as dual liver OCT1 and renal OCT2 substrate [81]. Therefore, a more specific urinary biomarker, N¹-methyladenosine (m1A) was proposed as an endogenous biomarker for renal OCT2 or MATEs inhibition for early discharge of renal transporter DDI risks. A multiplexed hydrophilic interaction liquid chromatography (HILIC)-MS/HRMS method for simultaneous measurement of m1A, 1-NMN, isobutyryl-L-carnitine (IBC), and carnitine in human plasma and urine samples was developed to evaluate multiple transporter biomarkers in the same study.

The development, FFP validation and application to a clinical trial study of a highly sensitive HILIC-MS/HRMS assay were discussed. The surrogate matrix approach was used to avoid the interference from endogenous analytes and ensure good accuracy and precision for the assay. HRMS provided specificity and confirmation of the four analytes and the isobaric endogenous compounds. Plasma standard and QC samples were prepared in 3-times charcoal stripped human plasma; urine calibration and QC samples were prepared in water as the surrogate matrix. The assay performed accurately with all accuracies and precisions within ±20% of their nominal values. The assay was relatively simple and reliable with good selectivity and sensitivity. The case study discussed demonstrated the assay application in a DDI clinical trial with a company drug candidate that indicated strong OCT2 and MATEs inhibition in vitro, which exceeded FDA risk threshold criteria at the projected human dose. The 4-plex assay enabled the performance comparison including biomarker specificity and sensitivity, of m1A and 1-NMN as OCT2 and MATEs biomarkers, IBC and carnitine as OCT1 biomarkers, as a “biomarker cocktail”, which increased the confidence in the DDI risk assessment and offered options for earlier characterization and clinical safety projections for OCT2, MATEs, and OCT1-mediated DDIs.

RECOMMENDATIONS

Below is a summary of the recommendations made during the 15th WRIB:

Chiral Methods for Method Development & BMV

• Regulatory guidance for chiral methods is broad and there is a need for specific guidance to clearly dictate when chiral assays are needed.

  	○	EU guidelines are stricter on chiral requirements than US ones and there is a need to harmonize to facilitate international submissions.
  	○	Clarification is required for chiral assays in BE studies in the final ICH M13 guideline.
  	○	Currently, more information can be found in the FDA Product Specific Guidance or PK Endpoint Guidance.
  	○	Regulators see few chiral assays for BE studies unless different activities or PK profiles are expected between enantiomers.

• In order to understand the PK properties of individual isomers, it is important to develop selective and sensitive methods to quantitate individual isomers and enable characterization of PK and disposition properties in nonclinical and clinical studies.

  	○	Chiral centers potentially responsible for the formation of chiral metabolites need to be evaluated early. Evaluate activity/potency, safety, clearance.
  	○	Even with an enantiomer pure compound, chiral conversion should be addressed with a need for chiral separation in relation to differences in potency, safety and disposition between the enantiomers.
  	○	Often the decision to follow metabolites is made based on early in vitro drug metabolism assessments, scouting in the clinical MAD study or following hADME (prior to Phase III).

• When developing chiral assays the following challenges should be considered:

  	○	Coupling of chiral chromatographic methods with MS is often challenging and requires considerable method development, resources and time.
  	○	Column technology, durability and reproducibility.
  	○	Use of non-LC based methods (e.g., GCMS or SFC-MS) for chiral quantitation is further complicated by the fact that few labs have experience developing these types of assays.

• Diastereomers do not require chiral chromatographic conditions for their separation.

  ○

  Conventional chromatographic conditions used for metabolite profiling should consider this point and profiling runs should be long enough to take into account the need to look for diastereomeric metabolites.

• Achiral assays are used as primary validated assays and chiral assays are qualified for mechanistic understanding, unless differences in potency or safety of disposition would require individual stereoisomer quantification. Scientific consultations or planned interactions with health authorities is recommended.

ICH M10 Section 7.1 for Endogenous Compound Quantification

• ICH M10 BMV draft guideline Section 7.1 provides an exhaustive guidance for the analysis of endogenous compounds and its recommendations have been welcomed by the global bioanalytical community due to the lack of regulatory guidance in this field for both PK and biomarker assays.

  ○

  Section 7.1 has already been widely applied in the industry even if this is still a draft guideline.

• It is recommended to have the same matrix for the standard curve as the study samples.

  	○	High endogenous levels can cause limitations to the LLOQ level of assay.
  	○	The ratio of the endogenous level to the desired LLOQ should be very low (suggested less than 4); the lower the ratio, the better to achieve the desired LLOQ.

• Endogenous levels should be screened to determine which lot is appropriate for preparing the standard curve to achieve the desired LLOQ.

  ○

  It was suggested to screen at least 15–20 lots of normal matrix samples to estimate the range using a surrogate matrix standard curve.

• It is challenging to distinguish between the spiked LLOQ and the blank endogenous level if the difference is within the allowed variability of the signal. If a low ratio cannot be attained, then the LLOQ can be adjusted to a higher concentration that would be appropriate for the study.

• The surrogate matrix approach is often used for the standards, and QCs are prepared using the standard addition or background subtraction methods.

• Most use one of two approaches for parallelism assessments; both values within 20% (assay variability).

  	○	Endogenous concentration as measured with a surrogate matrix calibration curve.
  	○	Endogenous concentration as measured by extrapolation of the standard addition calibration to the authentic matrix.

Tissue Analysis, Rare Matrices & Atypical Sample Collection in Regulated Bioanalysis

• These matrices are most likely to be associated with exploratory endpoints so FFP approaches are suitable.

• For primary and secondary endpoints, the level of validation should be agreed upon with end users to ensure awareness of any elements that differ from full validation, so that results are not over interpreted.

• When working with rare matrices, it is important that the regulators know what kind of data are presented, and what the data can inform. Consult early with regulators for their input.

• Ideally assess each tissue type separately. A homogenate of mixed matrices is not recommended but may be suitable only if agreed upon in advance with the clinical team and they are made aware of what the validation process can or cannot test so that results are not over interpreted.

• Mixed matrices can be used (e.g., standard in plasma or abundant tissue and QCs in specific tissue homogenate) or an approach where homogenates are diluted in plasma may be appropriate when multiple matrices are needed.

• Getting blank CSF is a widespread challenge. Hence two approaches are recommended: (1) using a consistent ratio of CSF:plasma in all samples/QCs, etc.; and (2) increasing assay sensitivity to dilute out nearly all CSF with plasma. Moreover, as reminder, it is always important to consider that adsorption can be a challenge with CSF.

• There was consensus that miniaturizing bioanalytical methods results in potential saving of rare matrices. Challenges include (1) ensuring homogeneous mixing of small sample volumes (acoustic mixing may be suitable); (2) spiking small volumes of standard to avoid high % organic; (3) liquid handling, transfer of small volumes; (4) sensitivity of the assay.

Quantitation of Intracellular Disposition of Oligonucleotides & S&S Challenges

• RISC loaded siRNA: low ng/g levels in the presence of high μg/g total tissue levels (<0.1% loaded).

• There can be nonspecific binding to beads (spike siRNA to beads in absence of antibody) and antibody therefore, it was recommended to include a non-specific antibody or control IgG.

• Endogenous miRNAs have been proposed (miR122 in liver) for normalization.

• Bioanalytical assay selection is driven by the stage of development, assay development time, and need of critical reagents; there is currently limited experience with qPCR for regulated work.

• Selectivity is one of the most difficult aspect for oligonucleotide assay development, even with the advantages of LCMS, particularly in cases of co-eluting minor metabolites without truncation of nucleotides (desulfurization, defluorination, deamination). Skilled analysts are needed to mitigate these issues.

• If switching technologies is unavoidable, bridging experiments should be performed. It was recommended to analyze the same set of samples with both technologies to understand the relationship between them. The same results may not be obtained, but there should be understanding of how the differences can impact data interpretation.

• If tissue samples are homogenized for one type of technology, they might not be suitable for the other type. This should be considered when designing bridging experiments.

• For targets such as liver, it is not important to use the same tissue homogenate for PK/PD measurements, but in CNS there can be large changes in PK/PD, even when sampling is done at very minimal distances for the PK and PD samples.

Recent Developments of Endogenous Compounds & Fit-for-Purpose Validation

• It is recommended to carefully consider the main bioanalytical challenges when developing multiplexed assays for several endogenous small molecules:

  	○	Chromatographic separation due to different chemical properties and numerous isobaric interferences;
  	○	Different dynamic ranges based on different endogenous levels for each analyte.

• When preparing the standards and QCs for several endogenous compounds (cocktail), similar chemical properties should be considered to facilitate combined extraction and chromatographic separation. However, it is recognized that in case of different chemical properties, it might be challenging.

• Discussion on the selectivity/specificity of bile acid biomarkers is not new but still very relevant for this class of molecules and it is always challenging and laborious due to numerous isobaric analogues.

• Dynamic range can be managed but it is important to consider that it can generate complex calibration and QC preparations, more prone to errors. Normal reference ranges for analytes being studied should be determined in the specific population of interest to differentiate true biomarker changes compared to inter-subject variability. Inter-occasion variability dependent on the time of sampling may also be a concern. For bile acid biomarkers, taking a baseline sample at multiple time points may be needed.

• With high endogenous level analytes in urine, water may be the better surrogate matrix compared with others (e.g., synthetic human urine) by diluting urine samples upfront.

• Development, characterization and qualification of methods for quantification of clinically-relevant lipid and metabolic biomarkers using highly multiplex, metabolomic-style methods is typically not amenable to full validation.

• Key considerations for highly multiplex methods include:

  	○	Use of longitudinal QCs is essential, especially if studies are to be run at different sites and to enable comparison between studies, bridging between studies using bioinformatics approaches (e.g., SERRF and others).
  	○	QCs can consist of pooled study samples, pooled commercial samples, or standardized matrices such as SRM1950 plasma; spiking in standards may be appropriate for analytes depending on their levels relative to the study samples.
  	○	Establish the quantitative range using both surrogate and endogenous analytes (dilutional linearity).
  	○	Examine the appropriate level for QCs relative to the study samples and which level is best for data correction. This can be achieved by diluting the QCs and/or by spiking-in analytes of interest.
  	○	Assess impact of high concentrations of endogenous analytes on quantification.
  	○	Determination of linear quantitative range is essential when establishing methods for analytes of the same class that are present in vastly different concentrations (e.g., lipid species).
  	○	Fold-change accuracy should be evaluated with QCs at different levels to determine that linearity is maintained throughout the batch and between batches.

SECTION 3 – Regulated Bioanalysis for Small Molecules & Point of Care

Jian Wang²¹, Anna Edmison²⁵, Stephen Vinter³⁵, Ragu Ramanathan²⁰, Arindam Dasgupta¹³, Lieve Dillen²⁴, Luca Ferrari²⁶, Sally Fischer¹, Daniela Fraier²⁶, Fabio Garofolo³³, Sam Haidar¹³, Kathrin Heermeier²⁷, Christopher James²⁸, Allena Ji²⁹, Olga Kavetska³⁰, Lina Luo³⁰, Gustavo Mendes Lima Santos³¹, Noah Post³², Anton I Rosenbaum³⁴, Natasha Savoie³⁹, Scott Summerfield²², Sune Sporring³⁴, Sekhar Surapaneni⁴, Katty Wan³⁰, Jan Welink¹⁶ & Eric Woolf³⁶

Authors are presented in alphabetical order of their last name, with the exception of the first 4 authors who were session chairs, working dinner facilitators or major contributors.

The affiliations can be found at the beginning of the article.

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 14th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 15th WRIB. The background on each issue, discussions, consensus and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

Dealing with GLP, GCP & GCLP Frameworks in Regulated Bioanalysis

How is quality oversight ensured for Good Clinical Laboratory Practices (GCLP) bioanalysis? Is an organizational framework needed for GCLP similar to Good Laboratory Practices (GLP) facilities, i.e., a harmonized/updated international guidance document? What level of validation is required for bioanalysis (PK and ADA) in non-pivotal studies? Can we clarify the distinction of pivotal versus non-pivotal studies in guidance documents?

Importance of Incurred Sample Stability in Regulated Bioanalysis

How is time zero (t0) determined for incurred sample stability (ISS) evaluations? For analytes with stability liability or unstable analytes that used stabilizers, how can we monitor ISS? If ISR is done quickly (following the initial analysis), is it necessary to run additional ISR from the same set of samples at a later time as a measure for ISS? Is ISS necessary when reference standards are unavailable?

Challenges When Changing Platforms (LBA to LCMS) in Regulated Bioanalysis

What can be done when LBA/immunoassay cross validation with LCMS does not pass due to an unknown metabolite (e.g., a metabolite co-measured in the LBA assay)? What can be done if the metabolite has identical mass with the parent and co-elutes in a short LC gradient (e.g., 5 min), but can be separated in a 30–40 min LC gradient? What if the metabolite is as active as parent? What if the metabolite is inactive? What if the bioanalytical assay was an LBA assay, where it would not be possible to distinguish the contributions?

Patient-Centric Approaches & Point of Care in Regulated Bioanalysis

The ICH S3A Q&A 2017 [82] recommends an assessment of the comparability of the exposure measurement in a given matrix between microsampling (capillary) and conventional sampling (venous) when moving from one technique to the other during the early stage of toxicology studies. In the case where the comparability has been performed in the toxicological species, is it common practice to repeat it in the clinical setting? When the comparability is performed in the clinical setting, what approach is used (e.g., ad hoc study or bridging venous vs capillary in the same study)? How popular is the use of the Tasso and Mitra devices for clinical studies? For which type of studies are they used (e.g., pediatric, Duchenne muscular dystrophy patients, critically ill patients, oncology)? Are Tasso and Mitra devices used for a subset of studies or entire programs (e.g., only Phase II–III or for disease populations)? How is comparability with traditional sampling approached, considering that there might also be a change in matrix (e.g., blood instead of plasma, dry versus liquid matrix)? Are additional validation tests performed (e.g., plasma protein binding ratio, blood/plasma (or serum) partitioning) and hematological parameters evaluated so that the systemic exposure can be evaluated appropriately from each measurement using different matrices? What are the challenges during filing for the above approaches (e.g., were there early discussions with regulators before filing, how high is the filing failure rate)? When home nursing is chosen to conduct a clinical trial (as opposed to clinical visits), what are the regulatory expectations for proving the chain of custody, e.g., recording information on administered dose, sample collection, and shipment date and time in a form or as a printout or any other way? How is this controlled? Once we show equivalent data with point of care (POC) versus central lab tests, what are the barriers to implementation of these technologies as a tool for use in clinical trials (not diagnostics)? Can validation of microsampling during assay development be sufficient for implementation in clinical trials? What do others see as barriers to the implementation of these technologies? How else can we improve our clinical trials and patient convenience especially during the current pandemic?

Regulatory Standards to Perform Bioanalysis in China

What approaches are being used to cross-validate PK and ADA assays when different bioanalytical laboratories are used to support China and ex-China cohorts in the same study? What rationales have been successfully used to support Human Genetic Resources Administration of China (HGRAC) sample export applications? What is the best approach to facilitate the China Inspection and Quarantine (CIQ) investigator sample export inspection? What is the best approach to support China inspection requests to reproduce PK calculations in real-time? For a global trial, where analysis of samples collected in China is, due to China export restrictions, conducted by a laboratory in China and where all samples collected outside of China are analyzed by a different laboratory, outside of China, is it necessary to present and analyze data by laboratory, or is aggregation acceptable if cross-validation criteria have been met? Is the recommendation different for PK data versus immunogenicity data?

Bioanalytical Challenges for Oncology Drug Development

Is the complexity of oncology drug regimens sufficiently considered in the development of BMV guidance(s)? Does bioanalytical wet lab testing of every possible analyte combination still make sense in selectivity testing? How can bioanalytical co-med stability be realistically tested, given the extensive number of combinations in modern oncology trials and the complexity of drug regimens in cancer treatment?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

Dealing with GLP, GCP & GCLP Frameworks in Regulated Bioanalysis

There are a variety of Good Practices (GxP) that are required during drug development. The scope of the GLP regulations is non-clinical safety studies [83,84], which are executed in the preclinical space while clinical studies must be performed using Good Clinical Practices (GCP) [85]. Both GCP and GLP provide frameworks for the execution of clinical or preclinical studies in order to ensure the quality, integrity and reliability of study data. The GCPs discuss the 13 main principles which include the informed consent of trial subjects as well as their rights, safety, and well-being, confidentiality, staff experience and training, good documentation, and a principle that requires systems and procedures in place to assure the quality of the trial. However, there are no specific details as to these systems and procedures. The GLPs go into more specific details regarding the systems and procedures needed around staff experience and training, equipment, good documentation, facilities and other test system related elements.

Bioanalytical laboratories provide analyses for both preclinical and clinical studies. While the required framework for bioanalysis is clear in the preclinical space due to detailed information provided by the GLP regulations, there is more room for interpretation in the clinical space. As a result, EMA [86] released a guidance document clarifying that the analysis or evaluation of clinical trial samples should be conducted in accordance with the principles of GCP. The ICH M10 BMV draft guideline [32] also states that for studies that are subject to GLP or GCP, the bioanalysis of study samples should also conform to their requirements.

Although WHO published a guidance referencing GCLP [87], this was published on behalf of the Special Programme for Research and Training in Tropical Diseases and should not be applied in other settings. GCLP is not recognized by regulatory agencies as a GxP in either North America or Europe [14,27].

On-going discussions focused on the perspective of the GxP framework needed for pivotal and non-pivotal studies and linking the levels of validation needed for methods used to support studies conducted within each framework. There was unanimous consensus that no further regulations are needed beyond GCP and GLP for studies and assays. The EMA Reflection Paper [86] gives guidance and flexibility on approaches for use with a variety of assays (e.g., method validation – “analysis should be performed using appropriately validated methods with defined acceptance criteria where appropriate.”) that covers all endpoints. If analysis is performed as part of a clinical trial, then GCP compliance is required and consideration must also be given to any relevant guidance documents that are followed.

The outsourcing of experimental work has become more prevalent, particularly for GLP studies. As a result, fewer GLP studies and fewer bioanalytical phases of GLP studies are being conducted in-house at the sponsor, raising the question of the value of maintaining OECD GLP certification for those who fall under these regulations. One proposed option is keeping GLP systems and using the GLP-style quality oversight for GCP studies as well. Among the small differences are the level of internal auditing and the extra demands of GCP, such as compliance with the informed consent. Another option is giving up the GLP certification and developing specific systems for GCP quality oversight. While the first option is simpler in some ways, the second option is open to develop specific systems required in the analysis of clinical samples (e.g., for biomarkers) such as flow cytometry, proteomic, metabolomic and genomics techniques.

Key recommendations were provided on the topic of the validity and appropriateness of the language “GxP method” and/or “GLP Validation”, and how these are considered in terms of “pivotal” and “non-pivotal” studies. Indeed, it was noted that since the initial recommendations in 2017, there are still issues with the inappropriate use of the terms “GxP method” and/or “GLP Validation” within the global bioanalytical community [16] whereas, the term “Scientific Validation” seems to have been completely eradicated and replaced with the more appropriate term “FFP Validation” [16]. Unless stipulated in national regulations, there is no requirement to perform method validation in compliance with GLP. Data should be accurately recorded and stored in a manner that protects its integrity, with suitable quality systems in place for their management [88]. Methods developed under a GxP framework can be validated, either fully or FFP. “GxP compliant” is not synonymous with “validated”.

Guidance from health authorities is important and covers many aspects. For clinical validation and study conduct, bioanalytical laboratories typically work very close to what GLPs require without actually claiming GLP. The ICH M10 BMV draft guideline [32] focuses on “studies submitted” rather than the use of “pivotal studies”, to better represent those studies where a full validation is required. Studies not meant for regulatory submission are at the discretion of the sponsor, therefore there was concern from industry that this may be burdensome given that most studies are included in submissions.

The level of validation required for non-pivotal studies was discussed. Guidance documents from health authorities are most important for determining compliance. For example, guidance documents distinguish between “non-pivotal” or “exploratory” versus “pivotal” studies, followed by definitions that are not identical between guidelines [30,31,90]. That brings the question of which studies are considered pivotal and what level of scrutiny should be applied in non-pivotal studies. It was concluded that a distinction should be made on the assay rather than the study level, depending on the intended use of the data produced. Non-pivotal assays need to be suitably validated so they are fit for purpose as they support decision making which can impact subjects and patients, so if a fit for purpose approach is used, the level of validation must be monitored as assays develop to ensure data is reliable to support decision making. It was agreed that the intended use of study data should feed the definition of “pivotal”.

Importance of Incurred Sample Stability in Regulated Bioanalysis

ISS is currently not a regulatory requirement. However, investigating ISS may be justified in some cases because the stability of an analyte in the matrix may not be completely predicted with the spiked QC samples. The discussion built upon prior industry and regulator recommendations on ISS [5,6,10,15] and focused on strategies that can be implemented during method development to evaluate potential ISS risks. Routine chemical structural alert analysis of analyte, examination of in vitro/in vivo metabolic pathways and testing known labile metabolites are some commonly adopted and effective measures used before incurred samples are available for evaluation. Some commonly occurring stability alerts include prodrugs, unstable N-oxides or glucuronides, and lactone-ring structures. When instability is confirmed, typically with an unknown metabolite reverting back into the analyte, stabilization strategies should be evaluated during method development and effectiveness of stabilization should be monitored during sample analysis by conducting ISR. It is important to determine at which stage of sample handling to add the stabilizer; if it is added to the plasma or serum, ISR may not detect any stability issues in case these originated in whole blood.

Despite the best efforts to stabilize the analytes of interest during method development, ISS issues may still occasionally be detected during sample analysis, discovered largely through inconsistent ISR results or when assaying analytical repeats. Evaluation of ISS is warranted when the investigation indicates the stability of incurred samples may be contributing to the observed inconsistent results. Using incurred clinical samples to conduct stability assessments should be done in compliance of GCP with an investigational plan detailing experimental procedures. When observed issues are seen in a subset of subject samples, consideration should be given to using non-pooled subject samples to avoid dilution effects on the problematic samples. It may be necessary to perform stability tests other than long-term stability (LTS) based on specific observations in the study (e.g., freeze/thaw, bench top and autosampler stability) [91]. It was recognized that it is not always possible to conduct LTS evaluations using incurred samples because the lack of t0 (baseline) results at the time when the issue was identified.

Based on the investigational outcome, it is essential to perform an impact assessment and/or justify any actions taken. If incurred sample instability is confirmed, the stability obtained from incurred samples should set stability limits for sample analysis. Addition of or improvements with the use of stabilizers at the clinical site may also be necessary, as the addition of stabilizer to sample tubes can prove challenging in the clinical setting. To facilitate the procedure, it was suggested that the use of a dual pipet that delivers plasma and buffer at the same time can be an option, or stabilizers can be added earlier during tube preparation.

Challenges When Changing Platforms (LBA to LCMS) in Regulated Bioanalysis

Over the course of a drug development program, it may be necessary to change the assay platform as in the case study discussed, where bioanalysis in the GLP toxicity studies as well as in the clinical single-dose and multiple-dose studies was performed with immunoassays (alphaLISA). The project was further developed into a combination project, and co-dosed with a compound that cross reacted with the antibodies used in the original immunoassay. Therefore, the assay format was changed to LCMS for the non-clinical and clinical studies. A hybrid assay was developed and validated in rats and humans. Both assays performed well in the validation studies, and all parameters were fulfilled for spiked QC samples.

It was agreed that, to be able to compare prior data with data from the new platforms, cross-validation of the methods is required. However, there can be challenges when interpreting the cross-validation results, especially when, as expected, LBA and LCMS methods do not give identical results. In the case study discussed, the two different assay platforms were compared for incurred samples obtained in a rat toxicology study and from a clinical Phase I trial. It was concluded that the case study discussed is in agreement with the recommendations in White Paper in Bioanalysis [8,11].

The cross-validation of the immunoassay and hybrid assay from the example showed that the results were comparable for spiked QCs, but when comparing non-clinical and clinical study samples, the hybrid assay gave lower results. There was a time dependent difference in the results, indicating a potential metabolite formation that is expected to be structurally very similar to parent and consequently co-measured in immunoassay. This conclusion illustrates the importance of knowing what the LBA assay measures and how the assays relate to each other.

When using two different platforms in the same program, it was recommended to avoid switching methods within the same study. Between studies, cross-validation is a “nice to have” since LBA and LCMS do not measure exactly the same analyte, and the cross-validation will provide information regarding the differences in what is being measured. If data are obtained from different methods across studies that are going to be combined or compared, the focus should be on knowing what the assays measure and how the assays relate to each other (determining bias), not with getting the same number as a result. Due to the variety of molecules and platforms, the amount of cross-validation needed is case-by-case and cannot be standardized.

Patient-Centric Approaches & Point of Care in Regulated Bioanalysis

Patients are willing to take part in clinical trials if they believe that the trials are beneficial to them or to others. However, they are often heavily burdened by the number of journeys to the clinical site (up to several weeks after the treatment ends), the trial duration, and the need for invasive procedures such as blood and other matrix collections. The burden on participants in a clinical study may lead to slow recruitment, inequitable access to clinical trials and trial diversity. Once participants are recruited onto a study this burden can result in patient dropouts, even when they are financially compensated for the expenses incurred. It is not uncommon for a clinical trial to be significantly delayed due to slow recruitment or to ultimately fail, as the ability to enroll and retain the required number of patients to allow meaningful data to be obtained for the drug under investigation is compromised. This is especially notable in pediatric development. Patient centricity plays a pivotal role in increasing the success of clinical trials, as it aims to establish processes and find solutions best suited to patient needs. This reduces the barriers to their participation, which in turn allows more patients to enroll in the clinical trial. Home sampling or microsampling technologies are also available that can improve patient convenience and participation, leading to faster clinical trial enrollment and more diverse clinical trials. Furthermore, these alternative ways of collecting samples can enable development of medicines for neonates and young children (<2 years of age), help with collecting samples from vulnerable populations, and increase diversity in clinical trials by making trials more accessible through reduction in clinic visits. Another area that has not yet been fully explored is leveraging these technologies to increase our understanding of disease biology by providing longitudinal sampling and biomarker data during episodic diseases where patient sampling may not be possible without the availability of home sampling.

Bioanalysts can play a pivotal role in support of clinical programs to improve the patient experience and/or facilitate operational difficulties in logistics. In a case study discussed, COVID-19 restrictions severely limited the possibility of onsite visits by patients. In order to mitigate the use of site nursing staff conducting home visits to collect samples, additional investigation into the stability in matrix was performed to de-risk the project; stability was proven for room temperature and normal light conditions with sufficient duration to collect good quality samples in a non-controlled environment. Working in close collaboration with sample operations, alternative ways of collecting samples were used and meaningful requisition forms were defined for regulatory compliance documentation, so that the clinical trials could take place in the midst of the COVID-19 pandemic.

Microsampling is a minimally invasive technique that is becoming more prevalent to withdraw blood from patients in clinical trials. A dry blood sample can be collected using the dried blood spot or the volumetric adsorptive micro sampling technologies, while a liquid microsample can be obtained via capillary sampling or by the use of the touch activated phlebotomy device. Microsampling is a more convenient sampling technique compared with traditional venipuncture as it minimizes patients discomfort while allowing for the collection of a good quality sample [16,23,26]. Study data were discussed from a pediatric study looking at a rare genetic neurodevelopmental disorder, designed to determine if exposure was comparable in venous and capillary samples as required by regulatory guidance [82]. In the context of this study, the use of capillary sampling had the advantage of alleviating the burden of blood collection while allowing for home visits to substitute for site visits on non-dosing days, denser PK sampling over a longer period of time, and the use of microsampling (capillary sampling) in future pediatric studies. Comparable results were obtained from capillary and venous samples using a hELISA assay developed and fully validated according to regulatory guidance [30,31].

Traditionally, venous blood draws are standard in clinical trials, where blood is collected via venipuncture for all timepoints and transferred to a central laboratory. This process is inconvenient, uncomfortable and potentially painful, requiring specialized staff in a controlled setting to collect large sample volumes. Several microsampling techniques were evaluated in an antibody therapeutic study in order to determine the utility of these techniques in clinical and non-clinical studies [92]. Some of the implementation challenges that were considered when comparing these technologies to blood samples were the impact of different collection sites (finger pricks vs arm capillary collection); bioanalytical requirements (sensitivity, stability, extractability, sample processing and shipping conditions); logistics (shipping requirements, time and date collection, patient compliance, sample quality, clinical site and patient training); and business or regulatory related (increased costs for conduct, complex protocol, CRO readiness, international regulatory approval). It was recommended to determine if the capillary/venous ratio is the same in preclinical versus clinical samples; if not, then additional validation should be carried out to ensure that data can be properly interpreted across different stages of the program; if yes, microsampling should be able to be used in clinical trials. It was noted that although there is some experience with using Tasso and Mitra for regulated clinical bioanalysis. The predominant challenges that still need to be addressed are patient compliance, time stamping of samples, CRO capabilities/implementation as well as sample stability during shipment. One patient compliance concern is with patients providing a blood sample from another person. It was suggested that to circumvent this issue, a DNA test on a witnessed blood sample can be performed, with repeated testing on further unwitnessed samples. Logistical challenges can be overcome with a study design that takes into account chain of custody issues so all risks are addressed by the clinical trial protocol to ensure GCP requirements are met and data reliability is maintained.

The increasing focus on personalized healthcare and patient convenience has placed greater emphasis on POC technologies with fast turnaround times as well as technologies that enable home sampling, reducing the burden of visits to the clinic. Although POC technologies have been around for years, their use has been limited to diagnostics and has not been used as a tool for drug development. POC technologies have made great progress in the last decade and can provide near real time results with reliable precision and accuracy comparable with central laboratories. Some applications of POC technologies are for enrollment and stratification for acute diseases (e.g., acute respiratory distress syndrome, sepsis); mitigation and management of adverse events (e.g., Cytokine Release Syndrome) [93]; therapeutic drug monitoring and dose adjustment; and immediate availability of data, resulting in 3–8 weeks earlier data analysis. These technologies can provide immediate actionable results that have the potential to positively impact patient care and clinical trial timelines. Studies should be conducted to show comparability of POC tests to an established method. Once equivalent data with POC versus central lab tests are demonstrated, implementation should be possible.

The challenge is the uptake and implementation of POC technologies into our current healthcare and established clinical trial models. Some of the challenges were evaluated and addressed in two case studies including sensitivity, comparability of data to clinical labs, cartridge stability/reproducibility and the use of blood versus serum/plasma. Logistically, accessibility and the ease of use by a bedside nurse, device training and kit supply, material flow, storage, and logistics for unused and used cartridges, capacity/capability at sites to process samples (e.g., centrifuge if needed, etc.), data transfer and verification; and data blinding are all considerations when setting up a study.

Regulators recommended that the sponsors should contact them early in the program to discuss implementing microsampling or POC technologies to ensure a reliable protocol is implemented with mitigated risks to patients and results in submitted studies.

Regulatory Standards to Perform Bioanalysis in China

China currently commands an estimated 30% of the world market for pharmaceuticals; pharmaceutical sales in China estimated to reach $161.8 billion by 2023. As such, China represents a significant market opportunity for Pharma [94]. Given the huge market, China has, for many companies, moved from a secondary market to a primary market. Thus, inclusion of China among the first wave of filings is an approach many companies have taken. To include China in early submissions, it has become necessary to conduct clinical studies in China. These include both standalone Phase 1 studies as well as the inclusion of cohorts of patients from China in global Phase II/III studies. Often these studies collect samples that require bioanalysis, either for assessment of pharmacokinetics, immunogenicity, or biomarkers. In contrast to the United States, where clinical studies fall under the auspices of the FDA, in China, multiple government agencies are involved in the aspects of studies related to bioanalysis. These agencies are the National Medical Products Administration (NMPA), the HGRAC, and the CIQ.

There are several study approval steps. The first step is obtaining Clinical Trial Authorization by the NMPA. The NMPA is responsible for the regulation of drugs and medical devices and provides clinical study approval. Once completed, HGRAC approval(s) are required. The HGRAC is under the Ministry of Science and Technology and is responsible for the protection of human genetic resources through the regulation of the collection and use of any source of genetic material such as cells, blood specimen, etc., as well as information or data relating to such material. All clinical trials that involve collection and usage of human biospecimens, including uses unrelated to genetics, must be approved by HGRAC. HGRAC must approve laboratories planned to be used to analyze samples from each study, regardless of whether the laboratories are in or outside of China. If ex-China, it must approve the export of samples. The regulations have been in effect since 1998, but have been recently updated and strengthened in mid-2019 [95]. In the initial application, the laboratories to be used for non-exploratory analyses must be identified with the rationale if sample exportation is requested. A supplementary application is required to identify laboratories to be used for exploratory analyses, including associated intellectual property and data sharing considerations. Finally, the HGRAC shipment export application must be completed. If exportation is approved in the initial application, a separate application must be submitted when each shipment is ready for export. This is combined with a CIQ application. New regulations require the Lead Investigator to apply for export approval and inspection [96], submitted to the local CIQ office prior to each shipment but after receipt of the HGRAC shipment export approval; formerly a central lab could file this application. This final application triggers a customs inspection at the Lead Investigator site.

Depending on the study objectives (primary, secondary, and tertiary objectives/endpoints), regulations in China require that data from studies be shared with clinical investigators. This requirement is part of the initial HGRAC application process. In the case of exploratory objectives, the data as well as the intellectual property associated with the objectives must be shared with the investigator [95]. This is included in the HGRAC supplementary application. When outlining the objectives, samples for future biomedical research should not be included. Sample disposal timelines should be defined (1–2 years after analysis is complete). Finally, including PK or immunogenicity assessments as an exploratory endpoint adds a layer of complexity to the study approval process.

In addition to the complexities of the Chinese regulatory system, due to restrictions placed on the export of biological specimens, samples collected in China often need to be analyzed in a bioanalytical laboratory in China. Prior to 2019, in many cases, justification for sample exportation was accepted; new HGRAC regulations implemented in 2019 have led to an increased rate of rejections for export requests. As a result, PK and ADA samples collected in China need to be analyzed in China.

Prior to 2015, a large percentage of bioanalytical work in China was conducted by academic institutions; bioanalytical support was bundled by investigators with clinical study conduct. In 2015, China instituted a self-assessment program that mandated that sponsors assess submitted studies for compliance to regulatory expectations. This was implemented prior to an enhanced inspection program and resulted in serious issues identified during regulatory inspection disqualifying sponsors from future submissions for multiple years. This led to a shift in the bioanalytical landscape toward “commercial” bioanalytical labs, some as part of multinational organizations. PK and ADA methods used to support China studies should be validated in accordance with local BMV guidances [97]. China-specific guidance is largely in line with EMA and FDA guidances [30,31] and ICH M10 BMV draft guideline [32]. Once implemented, it will result in completely consistent guidance for PK type assays.

Audits of China studies by NMPA are common and should be anticipated. These audits are typically end to end, beginning with the clinical site and sample collection, to the bioanalytical laboratory – usually with a focus on chain of custody of samples – through to the PK calculations. Unique to China audits is the need to reproduce PK calculations in the presence of the auditor. Report modeling efforts should be treated separately from the clinical study report. The audit experience should be considered when selecting a China bioanalytical laboratory.

For global studies with China cohorts, the inability to export samples requires that multiple laboratories (China and ex-China) could be used to support these studies. As such, current regulatory guidance requires that assays at these laboratories be subject to cross-validation if the data from the multiple laboratories are to be aggregated. This results in two possible scenarios when considering the need for cross-validation. The first scenario involves subjects in the study are only from China. In this case, if the assay used outside of China can be established within China, a cross-validation may not be required. It is recommended that the China lab conducts and documents a full assay validation. The need for cross-validation is issue driven and should be assessed by sponsor. If the assay used in China is different from that used to support ex-China studies, cross-validation is required. The second scenario involves a global study with separate labs supporting China and ex-China work; cross-validation is required. However, if cross-validation fails or is not possible, the ability to aggregate China data with ex-China Data from global studies may be limited.

The cross-validation of PK assays may be conducted with spiked and/or post-dose samples, however importing post-dose samples into China may be challenging due to HIV/HCV/HBV and currently COVID testing requirements. A data analysis plan and acceptance criteria should be developed a priori. Data analysis should include the assessment of bias. Applying ISR criteria alone does not satisfy this requirement. It was noted that it is important to distinguish within-lab vs between-lab variability (e.g., if lab 1 is having a high day and lab 2 is having a low day, cross-validation may fail, but if done across multiple days, cross-validation may pass).

The cross-validation of immunogenicity assays is not explicitly covered in regulatory guidance [98]. Immunogenicity assays typically use a “surrogate” as positive control and testing samples spiked with surrogate may or may not yield results representative of patients' samples. Ideally, previously tested patient samples should be used to compare immunogenicity results between labs but the importation of patient samples into China may be logistically challenging especially in the COVID era. Immunogenicity assays utilize a cut point to differentiate between positive and negative samples. Cut points are typically determined on a lab by lab basis and are set based on the control matrix available at the lab. Cut points for the same assay may differ significantly between labs, hence impacting the immunogenicity rate generated at the different labs. Therefore, demonstration of consistent immunogenicity results between labs may be challenging.

Bioanalytical Challenges for Oncology Drug Development

Increasing understanding of cancer biology and genetics continue to drive innovations in oncology drug treatment. Advances in immuno-oncology over the last decade have joined targeted therapies and classical chemotherapy as a third paradigm in cancer drug treatment [99]. Oncology represents an increasingly large proportion of research and development budgets (∼35% in 2019) and includes many novel therapeutic approaches and drug entities [100]. Drug modalities in development for oncology applications now include small molecules [101], monoclonal antibodies (mAbs) [102], antibody–drug conjugates (ADCs), bispecific T cell engagers, chimeric antigen receptor (CAR)-T cell therapies, siRNA, proteolysis-targeting chimeras and viral vectors. Successful cancer treatments almost always involve combinations of different agents, and even if new drugs are initially registered as single agents, strong efforts are made to identify combinations that will bring additional therapeutic benefits [103,104]. While there is hope for combinations of relatively non-toxic targeted or immune-oncology agents, many new agents are being studied with established combination therapies including cytotoxic agents, and even some newer agents have narrow therapeutic margins requiring close monitoring of drug exposure.

Many oncology treatment regimens require specific supportive or prophylactic care to avoid side-effects such as infusion reactions, neutropenia, viral or bacterial infections, and to control pain, nausea and vomiting and anxiety. In addition, many cancer patients present with pre-existing conditions unrelated to their cancer, but which also require continuing drug management. Considering all of these factors, many cancer patients receive multiple drugs in addition to their combination oncology treatments resulting in a complex polypharmacy.

There are specific bioanalytical challenges encountered during the clinical development of many of these new oncology agents. These include the development of bioanalytical methods for novel drug entities, which may require analysis of multiple drug-related species, use of methods such as PCR, hELISA, HRMS, or flow cytometry, or the development of extremely high sensitivity methods. Sometimes PK/PD relationships for novel drug entities are not well-understood and may require analysis of unusual or difficult to obtain matrices. Clinical exploration of multiple combinations with other agents can present difficulties in accessing proprietary bioanalytical methods for competitor's molecules and obtaining reagents and reference materials, especially for recently registered and novel agents.

Current wording in regulatory guidances on selectivity and co-administered compound stability testing for “specific drug regimens” [30–32,105] can be difficult to follow due to extreme polypharmacy and complexity in some clinical programs. Selectivity testing experiments are relatively straightforward to set-up and so can be performed for multiple combinations; however, it is unclear if all of these bioanalytical experiments make sense, particularly with novel agents and highly selective modern analytical techniques. Furthermore, at the time of this publication, only one instance has been reported in the literature where selectivity was compromised in an LBA assay [100]. Recommendations on co-administered compound stability were issued in the 2010 White Paper in Bioanalysis [3] when regulatory agencies began questioning the stability of samples containing co-administered compounds. Since then, the topic has been revisited numerous times to determine if any additional evidence had been discovered to warrant stability concerns [4,7].

Most recently, the scientific rationale for the requirement to assess co-med stability was again discussed during the public consultation of the ICH M10 BMV draft guideline [24]. The current ICH M10 BMV draft guideline includes the recommendation for co-administered compound stability assessment for studies with fixed-dose combinations or when multiple drugs are administered under a specific dosing regimen [32], although no convincing examples have been reported where analyte stability was affected by co-administered compounds in multiple publications to date [3,4,7,24,106]. The ICH M10 BMV draft guideline does not require stability experiments for all co-medications regularly administered to patients, however, the ICH M10 draft guideline and other current regulatory guidelines still include the conduct of stability experiments for co-medications administered to patients as part of “specific drug regimens”. This later requirement is particularly problematic for oncology studies given the complex polypharmacy seen in many oncology treatments, the different treatment regimens used for different indications, and as routinely administered supportive care drugs may vary for individual patients and accordingly to institutional guidelines. Considering the lack of examples showing impact to stability, and as a significant percentage of oncology drug combinations do not advance beyond Phase I studies, a reasonable strategy is to delay consideration of co-administered compound stability testing until combinations enter registrational studies. Regulators also confirmed that lack of co-administered compound stability could be accepted but it would be important to provide a rationale if not performed.

Finally, the pace and design of oncology clinical studies has changed greatly with studies now often targeting accelerated registration strategies and single protocols having multiple branches and sub-protocols to explore different indications, drug combinations and clinical pharmacology endpoints. These approaches can also complicate bioanalytical support strategies, for example due to accelerated study plans limiting time for implementation of method improvements, or the need for multiple bioanalytical methods for co-administered oncology agents. In addition, for small molecule drugs, implementation of bioanalytical methods for metabolites is complicated by accelerated timelines as human ADME or metabolite profiling data may not be available early enough to guide development of bioanalytical methods for circulating human metabolites.

RECOMMENDATIONS

Below is a summary of the recommendations made during the 15th WRIB.

Dealing with GLP, GCP & GCLP Frameworks in Regulated Bioanalysis

• No further regulations are needed beyond GCP and GLP for studies and assays. The EMA Reflection Paper [86] gives guidance and flexibility on approaches for use with a variety of assays (e.g., method validation – “analysis should be performed using appropriately validated methods with defined acceptance criteria where appropriate.”) that covers all endpoints.

  ○

  If analysis is performed as part of a clinical trial, then GCP compliance is required and consideration must also be given to any relevant guidance documents that are followed.

• Appropriateness of the language for BMV:

  	○	Unless stipulated in national regulations, there is no requirement to perform method validation in compliance with GLP. Data should be accurately recorded and stored in a manner that protects its integrity, with suitable quality systems in place for their management [88].
  	○	Methods developed under a GxP framework can be validated, either fully or FFP.
  	○	“GxP compliant” is not synonymous with “validated”.
  	○	It was confirmed that the inappropriate term “Scientific Validation” [16] seems to have been completely eradicated and has been successfully replaced with correct term “FFP Validation”.

• Different guidance/guidelines have different definitions of pivotal studies. The intended use of study data should feed the definition of pivotal.

• ICH M10 BMV draft guideline focuses on “studies submitted” so those not submitted are at the discretion of the sponsor, but point brought forth that most studies are included in submissions, so this can get burdensome.

• Non-pivotal assays need to be suitably validated so they are fit for purpose as they support decision making which can impact subjects and patients, so if a fit for purpose approach is used, the level of validation must be monitored as assays develop to ensure data is reliable to support decision making.

Importance of Incurred Sample Stability in Regulated Bioanalysis

• It was confirmed that ISS is not a regulatory requirement. However, it could be very useful during an investigation mainly if ISR fails and stability issues are suspected.

  ○

  During ISS investigations, it is very important to determine at which stage to add a stabilizer, indeed if added to plasma or serum, ISR may not detect any stability issues in case these originated in whole blood.

• The main challenge is still the presence of an unknown metabolite reverting back to analyte and this possibility should be considered during the ISS investigations.

• Recommendations for the use of stabilizers are:

  	○	Using a dual pipet that delivers plasma and buffer at the same time.
  	○	Add stabilizers early during tube preparation.

Challenges When Changing Platforms (LBA to LCMS) in Regulated Bioanalysis

• It is important to avoid switching methods within a study.

• Between studies, cross-validation is a “nice to have” since LBA and LCMS often do not measure exactly the same analyte and cross-validations will provide information regarding the differences in what is being measured.

• If data are obtained from different methods across studies that are going to be combined or compared, the focus should be on knowing what the assays measure and how the assays relate to each other (determining bias), rather than getting the same number from both methods.

• The amount of cross-validation needed is really case-by-case, and cannot be standardized based on the above recommendations.

Patient-Centric Approaches & Point of Care in Regulated Bioanalysis

• Regulators encouraged sponsors to contact them early in the program to discuss implementing microsampling or POC technologies in order to ensure a reliable protocol is implemented with mitigated risks to patients and results in submitted studies.

Microsampling

• Use of at-home or microsampling technologies is recommended since they are available and they can improve patient convenience and participation, enable development of medicines for neonates and young children (<2 years of age), help with collecting samples from vulnerable populations, and increase diversity in clinical trials by making trials more accessible through reduction in clinic visits.

• However before using microsampling it is important to determine if information can be extrapolated (e.g., using capillary and venous samples if the capillary/venous ratio is the same in preclinical versus clinical samples). If not, then additional validation should be carried out to understand the reason. If yes, microsampling could be used in clinical trials.

Point of Care Technologies

• Studies should be conducted to show comparability of POC tests to an established method. Once equivalent data with POC versus central lab tests are demonstrated, implementation should be possible.

• The predominant challenges with POC technologies that still need to be addressed are patient compliance (including concerns of a subject providing blood from a different person), time stamping of samples, CRO capabilities/implementation as well as sample stability during shipment.

• Study design should include all variables to meet GCP.

• Logistical challenges can be overcome with a study design that takes into account chain of custody issues in order that all risks are addressed by the clinical trial protocol to ensure GCP requirements are met and data reliability is maintained.

Regulatory Standards to Perform Bioanalysis in China

• Cross-validation is needed for analysis performed in China and ex-China.

  ○

  Consider that matrix source differences can cause challenges during these cross-validations.

• Importing/exporting samples into/out of China is still challenging and with no short-term solutions. Moreover, there are different agencies involved in China as well, which also complicates matters.

Bioanalytical Challenges for Oncology Drug Development

• For selectivity testing it is possible to test many combinations, but it is unclear if “wet lab” experiments make sense in many instances, particularly with novel agents and highly selective analytical techniques, and as interference is rarely seen with modern analytical methods.

• Due to the lack of examples showing impact to stability, assuming that the stability of each individual analyte is demonstrated, industry suggests that the scientific risk of not performing co-administered compound stability testing in Phase 1 is minimal. A reasonable strategy is to delay consideration of co-administered compound stability testing until combinations enter registrational studies.

  ○

  Regulators confirmed that the lack of co-administered compound stability could be supported with a rationale of why it was not performed.

SECTION 4 – Mass Spectrometry of Proteins

Matt Szapacs³, John Mehl⁹, Eugene Ciccimaro⁴, Dian Su⁵, John F Kellie⁹, Stephen C Alley², Mike Baratta²³, Linzhi Chen⁶, Fabio Garofolo³³, Neil Henderson⁷, Shawna Hengel², Wenying Jian⁸, Surinder Kaur¹, Anita Lee¹⁰, Joe Palandra¹¹, Haibo Qiu¹², Natasha Savoie³⁹, Diaa Shakleya¹³, Ludovicus Staelens¹⁴, Hiroshi Sugimoto¹⁵, Giane Sumner¹², Jan Welink¹⁶, Robert Wheller¹⁷, Amanda Wilson¹⁸, Y-J Xue⁴, Jianing Zeng⁴, Jinhui Zhang¹³ & Huiyu Zhou¹⁹

Authors are presented in alphabetical order of their last name, with the exception of the first 5 authors who were session chairs, working dinner facilitators and/or major contributors.

The affiliations can be found at the beginning of the article.

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 14th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 15th WRIB. The background on each issue, discussions, consensus and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

Hybrid Assays to Quantify Therapeutic Proteins

How to overcome ADA interference in the development of a sensitive, accurate and reproducible hybrid assay for therapeutic proteins? How should run acceptance criteria be set up based on validation precision and accuracy criteria? In general, increases in sensitivity can be achieved applying nano- or micro-scale chromatography and/or HRMS for therapeutic protein quantification when protein-level clean-up is negated or crude. What is holding the industry back from using these tools more widely? What are the current advancements in trypsin digestion?

PTM/Glycosylation Analysis for Biomarkers & Biotherapeutics

For biomarkers and biotherapeutics, what PTM analyses are taking place? What MS systems are being used (LCMS versus HRMS)? What is the analytical rigor needed? Has any PK/PD modeling been established or implemented using PTM bioanalytical data?

Hybrid Assays to Quantify Protein Biomarkers

How can the workflow to develop a tissue assay be applied to help address different bioanalytical challenges? Are there any new recommendations for protein biomarker validation? How can matrix selection improve the sensitivity for low abundance protein biomarkers? How should parallelism be assessed for hybrid biomarker assays?

Hybrid Assays for Target Engagement Assessment

Is there a preference from clinical and modeling teams to have direct target engagement (TE) assays over downstream TE assays when possible, or vice versa? For non-covalent drug:target interactions, what type of experiments should be done in method development to rule out artifacts from sample preparation? Often drug levels are overwhelmingly higher than its target and in such cases drug-bound and/or total target assays may be more suitable. Are there any benefits to measuring free target given the challenges and shortfalls with free assays?

Quantification of ADA by Hybrid Assays

What is the typical extraction efficiency and how does it compare with LBA? Can we improve the efficiency of immunocapture, and therefore maximize the recovery of ADA and sensitivity of detection? ADA positive controls (PC) are considered critical for LCMS ADA assays and isotyping. What is the best strategy for PC selection and application? How can the drug interference be reduced and the drug tolerance of the assay be improved?

Hybrid Assays to Quantify Transgene Proteins

What is the advantage of using hybrid methods? How should the transgene protein be quantified?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

Hybrid Assays to Quantify Therapeutic Proteins

Several recommendations have been published on the use of hybrid assays to overcome the ADA issue as well as the advantages and challenges of using hybrid assays as a complementary technique for biotherapeutic PK assessments [20,23,26]. ADA interference can present a challenge to the development of a sensitive, accurate and reproducible LBA for the accurate quantification of protein therapeutics. ADAs may compete with the capture/detection antibody for target binding and can substantially impact the quantification of proteins that induce strong immune responses [107–109].

A case study was discussed on the evaluation and implementation of a hybrid assay to support clinical PK studies of a therapeutic protein. By combining acid dissociation, immunoaffinity capture and LCMS detection, a hybrid assay was developed which tolerated 200 μg/ml of rabbit antidrug IgG. An ELISA assay only tolerated 2.5 μg/ml of the same monoclonal antibody. The acid dissociation step proved to be critical to reducing ADA interference. A head-to-head comparison of ADA-tolerance was then performed in a spike recovery experiment using 58 clinical samples that were selected based on ADA isotypes and titers. The hybrid assay resulted in 94.8% accuracy of samples within ±20.0% bias. In the subgroup of 28 samples that had high titers for at least one ADA isotype, 89.3% of the hybrid assay results were within ±20.0% of the theoretical value while only 17.9% and 14.3% of ELISA and electrochemiluminescence assays (ECLA) results were within ±20.0% of the theoretical value, respectively. The hybrid assay demonstrated superior accuracy over ELISA and ECLA, suggesting better ADA-tolerance. The effects of ADA on the LBA assay were diverse and showed no correlation with the measured ADA titers, probably due to the complex and dynamic nature of the ADA response over the course of the study and potential drug interference on the ADA measurement. When implemented for clinical PK sample analysis, the hybrid assay showed an ISR pass rate of 83.7%, with lower ISR reproducibility generally associated with higher ADA titers. It was agreed that this case study further supports the advantage of hybrid assays in overcoming ADA interference while underscoring the remaining challenge in achieving reliable quantification of therapeutic proteins in the presence of high ADA titers [20,23,26].

Methods using nano- or micro-scale chromatography (also known as nano and microflow) in hybrid assay and reagent-free MS methods for biotherapeutics are less common, due to the more nuanced nature of the set-up and technical requirements. There are cases, however, where LBA assays prove challenging. For example, issues arise with specificity when high sensitivity is needed for LBA assays. In this event, hybrid assay and reagent-free MS methods have been shown to be successful for antibody quantification to support PK analysis in clinical studies, even when dosed in the low μg/kg range with the added benefit of reducing or removing issues with selectivity and specificity.

Sensitivity gains are often observed when chromatographic flow rates are lowered, and ion transfer efficiency increases. As discussed in Section 1, the need for an appropriate IS with as little variability as possible is crucial for successful analysis. This is further exemplified in cases where nano- or micro-scale chromatography is used, adding complexity with potential incorporation of loading/trap columns, and the delicate nature of the equipment and sensitivity to source build-up, ion suppression, etc. Capture and digestion steps in large molecule hybrid assays can attribute to assay variability. To attain adequate sensitivity requirements for clinical sample analysis, hybrid assays typically include a capture step that is specific to the therapeutic (e.g., anti-idiotype, antigen, etc.). Analyte-based IS (i.e., peptides) do not account for these two crucial steps, which can ultimately impact robustness. Therefore, to account for specific capture and digestion, a heavy labeled intact antibody is preferred for use as an IS.

Building on previous recommendations on nanoflow/microflow LCMS and sensitivity [8,23], a case study was evaluated and discussed that focused on comparison between LBA and a hybrid nano-scale LCMS assay. A therapeutic mAb used in a clinical study was initially analyzed using an ELISA method with an LLOQ of 15 ng/ml and an anti-idiotype capture and detection reagent. At the anticipated dose of the mAb, the assay sensitivity was deemed insufficient (i.e., majority of samples were below LLOQ based on the ELISA assay). At the same time, selectivity and sensitivity issues prevented lowering the LLOQ. A consistent signal was observed in some patient predose samples through multiple time points/cycles. The observed signal appeared additive to therapeutic levels and indicated strong interference from this patient population in the assay. Furthermore, a desired LLOQ of <1.0 ng/ml was difficult to achieve using a traditional hybrid assay. Therefore, a hybrid nanoflow LCMS assay was developed and a LLOQ of 0.5 ng/ml was achieved. The use of HRMS increased the observed signal-to-noise ratio (SNR); a loading column reduced the run time to 15 minutes and diverted salts to waste, thereby protecting the analytical column and tips. Multiple aspects of the method were considered to reduce interference and increase sensitivity. A large sample volume (>100 μl) allowed for greater sensitivity while remaining compatible with the affinity capture. A therapeutic-specific capture reagent was used to decrease the background noise. Mass detection of the CDR peptide improved selectivity and specificity. When coupling it to the nanoflow, it increased the ion transmission and further reduced the background noise. Assay performance during the validation was robust and reproducible, with accuracy and precision within 10%. To interrogate each specific problem observed in the original data set and confirm true (>LLOQ) ELISA results, 3 categories of clinical samples were pooled and tested with the new hybrid nano-scale LCMS assay: “true” signals in the ELISA method, predose samples with identified interference and BLQ samples. In general, the hybrid values were as expected and without the interference observed in ELISA. Samples with consistent interference in pre-dose timepoints were BLQ and subsequent time points were in line with exposure compared to other subjects. The LLOQ of 0.5 ng/ml improved the ability to characterize the full PK profile. It was agreed, based on current industry experience and the case study discussed that nanoflow hybrid LCMS assays can be successfully validated and used in routine, high throughput sample testing to support clinical studies. It is recommended that IS choice, column/emitter lifespan/consistency, and resolution of the mass analyzer should all be considered during method development to achieve increased robustness. This confirms previous recommendations that ultra-sensitive low flow hybrid assay have proven to successfully eliminate interference observed in specific ELISA assays [8,23]. However, nano/micro-flow LCMS is not widely available, and this has to be taken into consideration if the assay is eventually to be transferred to a CRO.

One of the major bottlenecks in bottom-up workflows is the time required to digest the protein. However with the advent of a new generation of heat stable trypsins, digestion in as little as an hour within biological matrix is offered. The evaluation of three next generation trypsins was discussed, focusing on the magnitude/robustness of peptide liberation versus a standard overnight digestion. Promega Rapid Digestion, Promega Trypsin Gold and Thermo Smart Digest Soluble were tested with a sample of SILuLite SigmaMAb reconstituted in rat plasma. Various digestion times and trypsin ratios were assessed. Promega Rapid Digestion demonstrated superior digestion efficiency after 1 hour, enabling single day protein digest workflows, even when combined with immunoaffinity purification (IP) and SPE. The data confirmed that maximal efficiency was achieved even at high protein:enzyme ratios. Finally, it was determined that reduction/alkylation could be negated for non-cysteine containing peptides, allowing for a boost in sensitivity and further time savings during the workflow. However, based on previous discussions outlined in Section 1, digestion efficiency should be optimized on a case-by-case basis.

HRMS has provided an additional tool for potentially increasing sensitivity of assays through improved selectivity, however, it does come with associated challenges [20,23,26]. The instrumentation is not always available in bioanalytical laboratories, and traditionally, experience and training of scientists has been geared towards the triple quadrupole platform. Furthermore, the software used to run an HRMS must meet software validation standards and be 21 CFR Part 11 compliant. An example was provided to demonstrate that HRMS is a viable alternative to traditional LCMS in quantitative protein digest assays.

When a protein LCMS assay is to be utilized within a regulated study, the robustness of the workflow becomes an ever more important consideration. Bottom-up workflows by nature are complex, consisting of multiple steps spanning protein purification, peptide liberation and peptide purification, all of which can add a source of variability. It was agreed that control of the assay to mitigate the impact of this variation should be achieved through careful optimization of conditions, including implementing a strategy for identification of poorly controlled steps and employing an iterative approach to the methodology over the assay life cycle.

PTM/Glycosylation Analysis for Biomarkers & Biotherapeutics

Reagent-free LCMS methods are ideally suited to study PTM dynamics on endogenous proteins as a read out of TE from in vivo studies [20,23,26,110]. Assays for measuring biomarker PTM levels from serum or tissues may be presented for potential application to other PTM biomarkers that demonstrate mechanism of action or drug efficacy. Similar workflows for PTM quantitation for biotherapeutics may be key for understanding protein biotransformation or quality attributes at the amino acid level. LCMS protein approaches have been long utilized for peptide-based characterization of biotherapeutics in chemistry, manufacturing and controls (CMC). However, recent advancements in hybrid assays have enabled similar measurements for mAbs and other biopharmaceuticals from in-life studies [111]. In addition to peptide-level analyses, top-down characterization approaches can be utilized, both for biotherapeutics and for identifying co-eluted or background proteins [112].

Glycosylation is a PTM that can have significant influence on the solubility, stability, and tertiary structure of a therapeutic protein. Glycosylation can impact not only the therapeutic protein's biological activities, but also PK/PD [113], efficacy, safety, and immunogenicity. Several monosaccharides can be incorporated into therapeutic proteins (via asparagine, serine, or threonine residue); glycosylation can lead to heterogeneity and complexity: variable composition, linkage, branching points, and configuration of monosaccharides at each glycosylation site. It was agreed that characterization of individual glycoforms' PK properties has become a major focus for successful therapeutic protein development [27,114,115]. However, it was also stated that, with the heterogenous and complex nature of each therapeutic protein, it is very challenging to characterize the PK properties of individual glycoforms due to differential clearance rate in the systemic circulation.

With LBA more often used for the quantitation of total therapeutic protein concentrations rather than that of individual glycoforms, MS-based bioanalytical approaches have been widely utilized for individual glycoform profiling [116]. To date, there are three MS-based glycoform profiling approaches: 1) intact glycoprotein analysis for the overall glycosylation profiling; 2) glycopeptide analysis for identification of individual glycoforms at each glycosylation site with more detailed glycan structural information; 3) cleaved glycan analysis for individual glycan structures and overall glycan profiling. Although glycopeptide analysis is commonly used, the two other approaches can also provide complementary information for full characterization of glycosylation. Two case studies were discussed to evaluate the utility of MS-based glycoform profiling and provide recommendations.

The first case study exemplified glycopeptide PK profiling in rats for four glycoengineered mAbs: the reference mAb with common Chinese hamster ovary cell glycosylation, Man5 (mainly high-mannose glycans with five mannoses), ST3 mAb (galactosylated and highly sialylated) and monoantennary glycan [117]. Results suggested that among the four glycosylated mAbs dosed, M5 mAb exhibited a faster clearance rate. Multiple glycoform kinetics resolved within one mAb preparation with the differential clearance: Man5G0, Man5 > G0F ∼ G0F-N.

The second case involved free N-glycan PK profiling in rabbits [118]. Most glycan PK profiles assayed with the hybrid assay were comparable to the overall profile based on the LBA data. However, 2 glycans were statistically different. It was concluded that hybrid assays can measure the PK profile of individual N-glyco-variants with high sensitivity and reliable results, allowing insights into profiles that are not available with traditional LBA platforms.

Hybrid Assays to Quantify Protein Biomarkers

FFP BAV has been discussed frequently and recommendations, available in White Papers, include performing precision (inter-assay, intra-assay), dilution linearity, spike recovery freeze-thaw stability, processed sample stability, carryover, interference and biological variability [22,25,26]. It was confirmed that these recommendations are still upheld and currently used by the global bioanalytical community.

Although the application of LCMS to biomarkers was originally hampered by lack of sensitivity and, in some cases, selectivity, several techniques have been developed to overcome these challenges. The application of IP and low flow LC-HRMS has led to increased sensitivity and selectivity, allowing the use of LCMS approaches for low abundance biomarker quantification. The choice of approach should be driven by the amount of sensitivity required, i.e., peptide IP (when coupled with protein IP) is more sensitive than protein IP, which is more sensitive than non-IP peptide purification. However, for all approaches, adequate choice of the signature peptide is paramount. The type of data needed may limit and dictate the choice of technology. For example, when a pharmacologically active concentration is needed, a protein IP step is usually required.

For low-abundant biomarkers, the selection of the matrix is important. A case study was discussed to demonstrate that some biomarkers have much higher concentrations observed in e.g., plasma over serum. The biomarker existed in different conformations, influenced mainly by calcium concentrations. Therefore, ethylenediaminetetraacetic acid (EDTA) was selected as anticoagulant, complexing with the calcium and the conformation was stabilized. Furthermore, the biomarker existed both intracellularly and extracellularly, with intracellular amounts far exceeding extracellular quantities; extra attention should be paid to sample hemolysis and the magnitude of the effect should be evaluated. In the case discussed, plasma was selected as the matrix of choice, however, it was not as clean as serum due to micro-aggregates, which can cause loss of magnetic beads during incubation and cause variability. Filtration of the plasma samples before the immunocapture can solve this issue given that the biomarker does not bind to these micro-aggregates. The same biomarker assay was developed for urine. With this matrix, there were indications of aggregation and absorption to plastic. Low-bind plastics and surfactants can help mitigate these issues. Urine can also have high salt content, impacting the conformational changes as previously described. Leveraging the information from the plasma assay, EDTA was introduced, which stabilized the conformation and solved the absorption issue. However, this resulted in a much more complex sampling process that requires exact sample volume to be collected for addition of the anti-sticking solution. For the calibration curve, surrogate matrix was used. An animal species was selected where the signature peptide was not present. A matrix comparability test between plasma and urine was conducted to assess the potential effect for using the same surrogate matrix in the urine assay. No adverse effect was observed and thus EDTA plasma was used as the matrix for calibrators in both assays. Meanwhile, endogenous QC samples were used as required with the recombinant protein.

Current industry practice and recommendations on tissue protein biomarker quantitation were discussed and reviewed. This topic was first discussed in 2017 [17] where recommendations were based on early experience with hybrid assays. Discussions continued in 2018 [20] and 2019 [23]. It was agreed that the previous recommendations are still valid and currently used by the industry. Indeed, a typical assay development workflow for tissue assays begins with digestion of the recombinant protein and selection of the surrogate peptide, followed by LCMS method set up. Antibody screening is then cycled with tissue homogenization and extraction optimization until a method is developed with adequate sensitivity and recovery. The next steps are selection of the surrogate matrix and the establishment of the range, followed by FFP validation with parallelism, recovery and reproducibility experiments. This workflow can be adapted to accommodate a variety of assay challenges [119–121]. Non-liquid matrices, specifically tissue methods, are semi-quantitative and currently there is no strategy for full validation due to issues of authentic standard incorporation and extraction during sample preparation. How tissues are prepared and perfused prior to harvesting and blood contamination all contribute to variability. Additionally, stability in tissue is difficult to confirm due to biological heterogeneity and challenges for traceability from lack of representative authentic standards. Therefore, methods for tissue measurement of biomarkers should be fit for purpose, which is appropriate since the data are typically used for internal decision making.

Parallelism experiments for biomarker hybrid assay were recommended [17,20,26] in disease states where possible. However, it was specified that it is not necessary to assess parallelism for each step of sample preparation (e.g., digestion). Similar challenges exist for the evaluation regarding the lack of endogenous material as a reference standard. Sources of recombinant reference could include peptide fragments, cell line-derived material, or material harvested from other species. Preparing the calibrants using a dilution series of in-study samples could be considered. When determining the minimum required dilution (MRD), large sample dilutions may improve capture efficiency and minimize matrix effects but will impact sensitivity. For immunocapture antibody selection, it was recommended to screen using blank matrix and matrix spiked with protein. Finally, it was confirmed that precision and relative accuracy should be measured using endogenous material as QC (eQC) [18,21,24,27]. The FFP validation should be considered when setting precision and relative accuracy criteria. This is based on understanding the protein biology and how it is expected to change due to biological variability and administration of the treatment. Run acceptance can be based on statistical criteria or set at broader criteria (+5%) than PK assays. However, setting arbitrary criteria is not recommended. Variability in these assays is generally greater and they may not be able to differentiate small fluctuations in biomarker levels.

Hybrid Assays for Target Engagement Assessment

As a direct indication of the interaction of drug with their target biomolecules, TE plays a critical role in drug development including target validation, efficacy evaluation, PK/PD correlation, biomarker selection, dose prediction, toxicology assessment, etc. [122], and is typically requested for early clinical phases. While many analytical platforms can be used, hybrid assays have emerged as a versatile, powerful tool for TE measurements [17,26,123–125]. Three assay formats are possible: free target, drug-bound target, and total (free + bound). The free assay can be challenging to develop as the capture antibody can bind an epitope competitive to the drug or the immunocapture reagent might cause dissociation of the drug-target complex. In addition, a small degree of dissociation from the highly abundant drug-target complex can result in significant overestimation of free target [17,26,126].

There are several considerations to take into account during method development. Surrogate matrix is commonly used for calibration standard preparation for biomarker assays when the authentic matrix contains measurable amounts of endogenous biomarker (target). When surrogate matrix is used, the similarity with matrix effect and extraction recovery in both the surrogate and original matrices should be demonstrated with a parallelism experiment [126]. Calibrators are typically prepared using recombinant protein. It is important to understand the biology of the target (binding partners) and interferences if there is a ligand binding component in the method. The recombinant reference protein may not resemble the endogenous form due to differences in structure and native state (different proteoform, activation state, PTMs, binding partners, folding, etc.), altering the binding characteristics. In addition to a mismatch of immunocapture efficiency between the endogenous and reference protein, differences in digestion or MS detection may also contribute to assay inaccuracy [127]. Consensus was reached during the discussions that when selecting assay controls, in-vitro conditions may not be representative of in-vivo conditions. Additionally, to rule out artifacts from sample preparation the use of orthogonal methods should be considered. When selecting the assay type needed, the binding kinetics of the therapeutic in these measurements should be considered since it may be possible to get free target measures from drugs with a slow K_off.

Two case studies were discussed for enabling progression from toxicology TE to clinical PD assessment using LCMS. In both cases, a total target assay and a combination assay for simultaneously measuring drug-bound target and total target were developed based on the hybrid platform and validated as FFP according to prior recommendations [17,26]. Key assay development included immunocapture antibody and signature peptide selection with the aim that the same assay could be adapted to all species (rat, monkey and human) and all matrices (plasma, vitreous humor and aqueous humor). The selected surrogate peptides were conserved among different species, which made the nonclinical assays easily adaptable to the clinical assays. For the total target measurement, an anti-target antibody was used as the immuno-capture reagent for all species. For the bound target measurement, an anti-human Fc antibody was used as the immunocapture reagent for animal species while an anti-idiotype mAb was selected for human samples. Among various assay performance qualifications, drug interferences and parallelism were assessed thoroughly, and preset criteria were achieved. From preclinical toxicology studies it was found that target level increased by 2.9–12 fold in the presence of drug, clearly demonstrating target engagement in vivo. Results from the drug-target engagement in serum and ocular samples were used for PK/PD correlation, human dose projection, and proof of drug-target interactions in early clinical development.

Quantification of ADA by Hybrid Assays

Evaluating immunogenicity of therapeutic protein products is critical for drug development since ADAs may pose problems for patient safety or change PK profiles and reduce efficacy [128]. The continuous improvement and recent applications of hybrid assays have demonstrated its potential as a good orthogonal and/or complementary approach to LBA for immunogenicity analysis [129–134]. An additional benefit of these assays is ADA isotyping, recommended in guidance by regulators [90,135]. Although it is generally considered that the predictability of animal studies for evaluation of immunogenicity in humans is low, immunogenicity assessments in animals are often conducted to assist in the interpretation of animal study results (e.g., toxicology studies) and in the design of subsequent nonclinical and clinical studies [6,11,14,17,20].

To demonstrate the advancements in this field, a case study was discussed where a hybrid assay was established to isotype and semi-quantify monkey ADAs to fully human mAb drug candidates. ADAs were isolated from serum samples using an immunocapture step with the Fab of the full-length mAb crosslinked to magnetic beads to minimize matrix interference. During development, the use of crosslinking to beads was compared to the use of the biotinylated drug and crosslinking to the beads doubled the recovery rate of ADA capture. Non-specific binding and background noises are big challenges to most of the ADA assays. Although some of the background interference can be mitigated using Fab only as capturing reagents, it only works for IgG4 with no improvement for other Ig isotypes, e.g., IgM [136–138]. A positive mAb control against the human immunoglobulin kappa light chain was used as a calibration standard for ADA semi-quantification, however, a more quantitative strategy with better PCs (e.g., heavy isotope labeled antibody) and isotope labeled peptide standards could provide more accurate and reliable quantitation for each ADA isotype. The results indicated that: 1) IgG1 was the most abundant isotype in ADA positive samples; 2) IgG2 and IgG4 were found at lower levels; 3) IgG3 and IgA were found at very minor levels. Furthermore, the levels of total ADA determined by the LCMS methods were compared with the results obtained using a traditional LBA. The study demonstrated that IA-LCMS is a promising technology for detecting and isotyping ADAs and provide additional immunogenicity information for both nonclinical and clinical immunogenicity assessments. Drug tolerance was not discussed in detail but should be considered for development of hybrid assays for ADA isotyping.

It was agreed that several lessons were learned from this case study. During method development and validation, it is important to consider that these are semi quantitative/relative quantitative methods using a surrogate control. Extraction efficiency can impact sensitivity and varies by isotype; the methods used to assess extraction efficiency using surrogate controls may not reflect the real samples. The selection of PCs is important. Using a positive control that has a high affinity will result in high extraction efficiency. When developing the assay, it is important to be mindful that immune response is not always to the CDR but could be to other parts of the molecule. Finally, drug tolerance may be improved if the capture reagent is covalently bound to the beads; more stringent diluent conditions can be used.

Hybrid Assays to Quantify Transgene Proteins

The increase in gene therapy therapeutics has resulted in new challenges and opportunities in both the modelling and translational as well as the bioanalytical space. One such bioanalytical opportunity is the need for measurement of transgene expressed protein in both soluble matrices and, more importantly, target tissue [17,20]. These measurements can be critical as they can inform on protein levels in healthy volunteers, which can help set efficacious levels and dose projections for the gene therapy. In addition, sufficient therapeutic transgene protein levels must be achieved in the systemic circulation or the targeted tissue at the efficacious doses for any successful gene therapy programs. Therefore, the quantification of transgene expressed proteins becomes of particular interest in nonclinical studies to select the optimal dosage and build the PK/PD relationship for clinical development.

In vivo and ex vivo gene therapy modalities, lentiviral or adeno-associated virus vectors are engineered to deliver the transgenes to the target site of action. To assess the biodistribution and persistence of the gene therapy molecule, the viral genome biodistribution and transduction efficiency are usually evaluated by the viral genome DNA and transgene mRNA level by qPCR and RT-qPCR methods, respectively.

Hybrid assays have demonstrated high selectivity and broad dynamic range compared with conventional LBA for protein bioanalysis in gene therapy. For example, hybrid assays can be a powerful tool to differentiate the wild type and mutated transgene protein (e.g., amino acid correction) [139,140] or detect the post-translational modification (e.g., glycosylation pattern) [141,142]. In addition, hybrid assays are less stringent on the critical capture/detection reagent for the transgene protein. This advantage also brings up the assay tolerability in the presence of ADA (anti-transgene protein) in the nonclinical study [17,20].

Although the conventional LBA-based assay is the current gold standard for transgene protein detection, hybrid assays can be used as a complementary assay to characterize the transgene protein biodistribution of the gene therapy molecule [17,20]. Obtaining accurate and robust measurements in tissues however can be quite challenging and becomes even more so when considering clinical application. Considerations need to be given to many key items including the bioanalytical approach, reference standard, acceptance criteria, validation assessments for tissue samples, data normalization, just to name a few.

Tissue collection for biodistribution studies need to be well coordinated for each assay. The tissue collection list needs to be fit-for-purpose depending on the target site of action in the project. Tissue samples need to be collected, flash frozen and stored in a freezer until analysis. For mRNA assays, tissue samples need to be collected in RNA-protect tissue tubes or RNAlater solution.

Online peptide immunoaffinity has been previously shown to enhance sensitivity in protein IA due to sample cleanup but specifically offers unique value for workflows where bypassing protein-based capture is advantageous. This can be due to concerns with multiple isoforms or differences in transgene protein, or when the use of harsh lysis conditions/detergents for tissues is not well suited for efficient or robust protein-based affinity. Pellet digestion, through crashing of proteins by organic solvents such as acetonitrile, helps to remove the harsh lysis buffer components.

For transgene proteins, there is an inherent desire to treat the assay as if it were for PK in a clinical setting. However, the unique considerations for the assay and calibrators align with the biomarker type assay (e.g., stability assessments: freeze-thaw, benchtop, re-injection, processed sample). Additionally, a measure of peptide IA efficiency should be incorporated as part of the daily system suitability.

Regarding the reference standard, a certificate of analysis should contain the concentration and how it was determined, the lot number and date, the storage buffer, the shipment and storage conditions and the purity, typically by SDS-PAGE. Single use aliquots should be considered to limit the impact of freeze-thaw issues for freshly prepared standards and/or QCs. Endogenous controls should be implemented; standard and QC preparation should be tested. A reference for normal expression levels should be implemented, along with retest instructions for the reference material.

When developing a reference standard stability plan, it was noted that commercial vendors often use generic 12-month language at -20°C to -70°C, therefore stability durations and testing frequency should be considered beyond those initial time estimates. Testing parameters should reaffirm the conditions used in the original certificate of analysis and should provide a concentration assignment (e.g., bicinchoninic acid analysis), size confirmation (e.g., Western blot), and comparison to an external reference (e.g., endogenous control).

RECOMMENDATIONS

Below is a summary of the recommendations made during the 15th WRIB:

Hybrid Assays to Quantify Therapeutic Proteins

• Hybrid assays have a major advantage in overcoming ADA interference and can be successfully used for the analysis of clinical samples.

  ○

  Combining acid dissociation, IA capture and LCMS detection, hybrid assays may tolerate high levels of ADAs.

• Nanoflow hybrid LCMS assays can be successfully validated and used in routine, high throughput sample testing to support clinical studies.

  ○

  For increased robustness, IS choice, column/emitter lifespan/consistency, and resolution of the mass analyzer should all be considered during method development.

• When protein LCMS assays are utilized within a regulated study, the robustness of the workflow becomes a very important consideration.

  ○

  Controls should be built into the assay to mitigate the impact of variations by optimization of conditions, including implementing a strategy for identification of poorly controlled steps and employing an iterative approach to the methodology over the assay life cycle.

PTM/Glycosylation Analysis for Biomarkers & Biotherapeutics

• Reagent-free LCMS methods can be successfully used to study PTM dynamics on endogenous proteins as a read out of target engagement from in vivo studies.

• The heterogenous and complex nature of each therapeutic glycoprotein cause challenges to characterizing the PK properties of individual glycoforms due to differential deglycosylation in the systemic circulation.

• Hybrid assays can measure the PK profile of individual N-glyco-variants with high sensitivity and reliable results and provide insights into profiles not available with traditional LBA platforms.

Hybrid Assays to Quantify Protein Biomarkers

• Prior White Paper in Bioanalysis recommendations for protein BAV are still upheld.

• The choice of IP approach for low abundance biomarker quantification should be driven by the amount of sensitivity required, i.e., peptide IP is more sensitive than protein IP, which is more sensitive than non-IP peptide purification.

  	○	For all IP approaches, adequate choice of the signature peptide is paramount.
  	○	If a pharmacologically active concentration is needed, then a protein IP step is required.
  	○	For low-abundant biomarkers, the matrix selection is important. Some biomarkers have much higher concentrations observed in e.g., plasma over serum.
  	○	The magnitude of the effect of hemolysis should be evaluated.

• The typical assay development workflow for biomarker tissue assays as described in previous White Paper recommendations is still valid and can be adapted to accommodate a variety of assay challenges. Steps include:

  	○	the digestion of the recombinant protein and selection of the surrogate peptide.
  	○	LCMS method set up, optimization of performance and response of selected peptides.
  	○	immunocapture antibody screening cycled with tissue homogenization and extraction optimization until a method is developed with adequate sensitivity and recovery.
  	○	surrogate matrix selection and the establishment of the range should follow analytical considerations of assay performance and be FFP.
  	○	quantitation of protein should follow qualification by parallelism, recovery and reproducibility experiments.

• Evaluate parallelism in disease state where possible and follow previous White Paper in Bioanalysis recommendations for parallelism in LCMS:

  	○	Analogous to LBA since similar challenges re: lack of endogenous material for reference standard.
  	○	Only aspects of an assay prone to matrix effects require assessment of parallelism.
  	○	Consider using pooled in-study control samples as the dilution series to assess study sample and reference measurement concordance.

• Large sample dilutions may improve capture efficiency and minimize matrix effects but will impact sensitivity.

• For immunocapture antibody selection, screening is performed with blank matrix and matrix spiked with protein.

• Precision and relative accuracy should be measured using endogenous material QCs.

Hybrid Assays for Target Engagement Assessment

• Direct TE is usually requested for early clinical phases and it is important to understand the biology of the target (binding partners) and interferences if there is a ligand binding component in the method.

• Binding kinetics of the therapeutic will drive assay strategy in these measurements.

• If the drug has a slow K_off, it is possible to get accurate free target measurements when the drug:target equilibrium is maintained within the timeframe and handling procedure of analysis.

• When selecting assay controls, consider that in vitro conditions are often not reflected by in vivo conditions.

• To rule out artifacts from sample preparation and confirm measurements, consider the use of orthogonal methods such as flow cytometry, if possible.

Quantification of ADA by Hybrid Assays

• It was confirmed that hybrid assays can be used successfully as an alternative/orthogonal approach to LBA for immunogenicity analysis.

• An additional benefit of these assays is for ADA isotyping, recommended in guidance by regulators.

• During method development and validation, it is important to consider that these are semi quantitative/relative quantitation methods using a surrogate control.

• Extraction efficiency would impact sensitivity and varies by isotype. The methods used to assess extraction efficiency using surrogate controls may not reflect that in a real sample.

• Positive control selection is extremely important also for hybrid assays. Using a positive control that has a high affinity may result in high extraction efficiency.

• Immune response is not always to the CDRs and could be to other parts of the molecule.

• Drug tolerance may be improved if the capture reagent is covalently bound to the beads; more stringent diluent conditions can be used.

Hybrid Assays to Quantify Transgene Proteins

• A hybrid assay was successfully used for the measurement of transgene expressed protein in both soluble matrices and target tissue as an alternative/orthogonal approach to LBA.

  	○	Generally, tissue protein extraction buffer containing protease inhibitors is incompatible with the tryptic digestion without the immunocapture purification.
  	○	Pellet digestion is recommended for sample cleanup as long as the transgene protein expression level is sufficient for the quantification.

• If a transgene protein has no endogenous counterpart or a specific capture antibody against the engineered transgene protein is available, it is possible to use specific peptides unique to the transgene protein. Otherwise, measurement of total transgene protein is based on the assumption that any post-treatment increases are due to the transgene protein.

  ○

  To discriminate between endogenous versus engineered transgene protein, the use of mRNA assay can be considered as a backup.

• For validation of transgene proteins methods, FFP validation should be considered.

  	○	BMV guidelines for PK assays can be a starting point but unique considerations for the assay and calibrators align more with biomarker type assays and BAV.
  	○	Additionally, a measure of peptide IA efficiency should be incorporated as part of the daily system suitability.

SECTION 5 – Input from Regulatory Agencies on Regulated Bioanalysis & BMV

Seongeun (Julia) Cho¹³, Arindam Dasgupta¹³, Anna Edmison²⁵, Sam Haidar¹³, Elham Kossary³⁸, Gustavo Mendes Lima Santos³¹, Sandra Prior³⁶, Mohsen Rajabi Abhari¹³, Diaa Shakleya¹³, Catherine Soo²⁵, Stephen Vinter³⁶, Yow-Ming Wang¹³, Jan Welink¹⁶ & Jinhui Zhang¹³

Authors are presented in alphabetical order of their last name.

The affiliations can be found at the beginning of the article.

The highlight of each WRIB conference is the annual input provided by international regulators on regulated bioanalysis and BMV. This year, regulators from US FDA, EU EMA, UK MHRA, Health Canada and Brazil ANVISA and WHO presented topics of interest to the Global Bioanalytical Community.

US FDA

OSIS Response to the Pandemic & Observations from Remote Record Reviews

Main functions of the Office of Study Integrity and Surveillance (OSIS) include: (1) to select sites for inspection based on various risk and surveillance criteria; (2) inspect sites to ensure quality and integrity of studies; (3) report inspections by writing and reviewing establishment inspection reports; and, (4) support CDER's approval decisions with data reliability recommendations and evaluations of subject safety. In the face of the pandemic, in April 2021, the US FDA issued a guidance document on Remote Interactive Evaluations [143]. This document describes the various interactive and virtual tools and approaches available for FDA's remote interactive evaluation. In addition, in May 2021, a Resiliency Roadmap for FDA Inspectional Oversight was issued explaining the agency's inspectional activities during the COVID-19 pandemic and its plan toward a more consistent state of operations [144]. Under the COVID-19 public health emergency, FDA has limited inspections to prioritized domestic facility inspections and those that are deemed mission-critical. FDA has been applying risk management approaches to determine when to request a facility to participate in a remote interactive evaluation. The guidance also notes that the FDA does not accept regulated entities' requests to conduct a remote interactive evaluation. In alignment with the agency's remote interactive evaluation approach, OSIS in CDER implemented remote record review as an alternative tool to on-site inspections. While it is not meant to replace or represent FDA onsite inspections, the remote record review allowed OSIS to continue to support the Office's mission and CDER's application review. It was initiated in June 2020 and includes reviewing bioavailability, bioequivalence, and GLP studies supporting NDAs, ANDAs, BLAs, and INDs. In fiscal year 2020, OSIS conducted 167 site evaluations, the majority of which were through remote evaluations.

Remote record review involves voluntary participation of a site and includes review of source data, site records, facilities, electronic systems, processes, etc., by video and teleconferences with site staff. Prior to the remote record review, the site is provided with the list of documents for OSIS’ review, which are then uploaded to a secure FDA cloud repository. Questions and concerns are discussed during virtual interactions and written observations are provided at closeout as applicable. Following evaluation of observations and the site's response, OSIS provides its conclusions and recommendations to the CDER review division regarding data reliability and human subject protection.

From the remote record reviews conducted since 2020, 31% had observations issued. Some examples of observations included:

• samples reanalyzed without documented causes

• no investigations performed despite SOP requirement

• no records of timing for sample retrieval and processing

• anchor points excluded in calibration curve fitting inconsistently without established criteria

• subject case report forms not accurately reflecting source information

Moreover, several specific observations made during remote record reviews were shared in more detail to help industry improve quality and compliance with regulations. In the first observation, the laboratory did not report all validation data for estimation of assay accuracy and precision (A&P). A&P assessments conducted during method validation excluded two A&P runs from the global A&P calculations without evidence or documentation of an assignable cause. The lab's SOP on method validation allowed for exclusion of QC data in A&P runs from global A&P calculations without evidence or documentation of an assignable cause. OSIS' scientific evaluation was that A&P runs should not have QC acceptance criteria because they are predictive of how well the method being validated will perform during sample analysis and the QC samples serve to mimic study samples. Not including “bad” QC data and including only “good” QC data during A&P assessment gives a false estimate of the accuracy and precision of the method, presenting it as better than it actually is. All data generated for estimation of accuracy and precision should be included in the A&P assessment except those where documented errors are due to an “assignable cause.” However, firms have an option of including A&P assessment without the “bad” QC data in addition to presenting the totality of the results. In the observation, the firm was asked to recalculate the accuracy and precision after including all QC values from all A&P runs. Inclusion of unreported results led to failure of the experiment to meet acceptance criteria.

The second observation was issued because investigational medicinal product (IMP) reserve samples were not retained from all shipments. Specifically, 20 kits were received from the IMP supplier as part of a third shipment for the study. However, all 20 IP kits were dispensed to study subjects for dosing and none was retained as reserve sample, as required by regulations [145]. For all BE studies not conducted under an investigational new drug (IND), the testing site is responsible for retaining reserve samples. If the study is conducted under an IND, the sponsor is responsible for retaining reserve samples [146]. Representative reserve samples have to be retained from all shipments of drug products received by the testing site. Additional guidance is available provides clarification on the quantity of reserve samples that have to retained but does not apply to the other requirements like reserve sample requirements under 21 CFR 211.170 current good manufacturing practices. “Requests for a reduction in the quantity of reserve samples for studies involving multiple shipments and study sites (e.g., comparative clinical endpoint studies), including any unusual circumstances that may prevent a particular study from retaining samples from each shipment, should be addressed on a case-by-case basis” [147]. Because the firm did not retain reserve samples for reference and test products from the third shipment, authenticity of the drug products used in that part of the study could not be verified. The firm agreed with the finding and indicated that the sponsor instructed them not to retain reserve samples from the third shipment. As a corrective action to prevent recurrence in future studies, the site planned to retain reserve samples from all shipments as required irrespective of the sponsor's instructions.

Other observations were issued since on the selectivity of the biologic PK assays in the study matrix was not fully demonstrated. In one case, the firm did not evaluate the PK assay specificity for interference from ADAs. In another, the firm did not evaluate the effect of hemolyzed matrix for a US approved drug in the PK assay. Finally, a firm did not evaluate the LTS for a US approved drug. Additional findings were related to not reporting method validation data, not having established criteria to exclude data, and missing original or raw data.

BMV of Biotherapeutics using Mass Spectrometry

In the last 10 years, there have been plenty of literature and recommendations published on the use of MS for quantification of large molecules as alternative/orthogonal approach to LBA [3–6,8,11,14,18,20,24,26,27,33,34,148]. When using MS-based assays for biotherapeutic quantification in regulatory submissions, it is beneficial to provide scientific justification on the reasons for: deciding to use LCMS rather than LBA (e.g., lack of reliable critical reagents for LBA); top-down versus bottom up; type of IA capture procedure, and IS. Another key consideration is the selection of the signature peptide(s). The documented justification of the choice should include how many peptides and transitions/ions are chosen, the chemical and metabolic stability of the signature peptides, the MS acquisition mode, and the possibility of using HRMS for retrospective analysis [148]. There has been recent consensus on the use and management of critical reagents also for hybrid assays similar to the LBA [149,150]. The effect of ADA and biotransformation have been also recently been discussed and should be considered in the scientific justification [36,111]. Finally, recommendations have been given for comparison and cross-validation for measuring the same analyte, acceptance criteria, and identifying possible patterns in the data [8].

Bioanalytical Regulatory Requirements for Diverse Complex Matrices

Pharmaceutical quality ensures a product of any kind consistently meets the expectations of the user. Patients expect safe and effective medicine with every dose they take. Pharmaceutical quality assures every dose is safe and effective, free of contamination and defects. “Validated bioanalytical methods for the quantitative evaluation of analytes and biomarkers in complex matrices and biological tissue are critical to successfully conduct nonclinical, biopharmaceutics, and clinical pharmacology studies. These validated methods provide critical data to support the safety and effectiveness of drugs and biologic products. Validating the analytical method ensures that the data are reliable by addressing certain key questions, including: does the method measure the intended analyte, what is the variability associated with these measurements, what is the range in measurements that provide reliable data, how do sample collection, handling, and storage affect the reliability of the data from the bioanalytical method” [30]. The Office of Testing and Research (OTR) is responsible for guidance support, medical counter measures, product performance, public health, and surveillance. The OTR recently published a bioanalytical preclinical study of dexamethasone and metabolite quantification in rabbit plasma, aqueous and vitreous humor and retina tissue using UPLC-MS [151]. The regulatory objective was product specific guidance for intravitreal implants with an outcome of developing a non-invasive surrogate biomarker of intra-ocular exposure for assessing BE of a generic implant product. The research outcome was to develop an alternative bioequivalence approach based on plasma PK in combination with in vitro formulation characterization and drug release testing to assess bioequivalence between Ozurdex and its generic products. The study determined that the release of dexamethasone from the intravitreal implant at a plasma concentration of 0.24 ng/ml, aqueous humor of 177 ng/ml, vitreous humor of 318 ng/ml, and retinal implant harvested tissue of 2440 ng/ml. Another bioanalytical study for regulatory guidance support was presented for the in vitro BE analysis through regression-based and ratio-based quantification approaches. The regulatory objective was to provide guidance for in vitro BE analysis and to provide a science-based standard to guide application sponsors in implementing a valid analytical calibration model to best ensure acceptable accuracy and precision. The study found there was no statistically significant difference between ratio-based and regression-based analysis at low, medium, and high calibrator concentrations for acyclovir, vancomycin, and triamcinolone acetonide. An analytical method developed for alcohol-based hand sanitizer products was shared. The study endpoint was drug product quality to support the US FDA's public health mission to ensure the quality and safety of alcohol-based hand sanitizer drug products (liquids, gels, wipes, foams and sprays). The research outcome was to develop an advanced analytical testing method for ethanol and isopropanol as active ingredients and to monitor level 1 and 2 impurities. The method was successfully validated for the quantitation of ethanol, isopropanol, and level 1 impurities (acetaldehyde, methanol, benzene and acetal) over three days by two different analysts following ICH Q2(R1) guidelines [152]. A final study was presented regarding a recent medical counter-measure study from the OTR for the preclinical determination of galantamine for use in the event of a nerve agent attack [153]. Both galantamine oral dissolving film and transdermal drug delivery systems (TDDS) were bioavailable. The galantamine oral dissolving films provided a fast drug release and the TDDS formulations provided prolonged drug release. In summary, the current bioanalytical method validation guidance supported by bioanalytical studies and emerging bioanalytical tools can advance the regulatory science of bioanalytical method validation. FFP analytical method development of complex biological and pharmaceutical matrices by advanced technologies will support the regulators with reliable scientific data for regulatory actions. Implementation of bioanalytical guidances [30–32] are the bridge to quality drug products.

EU EMA

ICH M10 BMV Guideline

Updates were provided on the status of the ICH M10 BMV guideline [32]. The ICH M10 BMV guideline was endorsed by the ICH management committee in October 2016. The guidance is currently in stage 3 (regulatory consultation and discussion) following public consultation. To deal with the high number of comments received, topics were first discussed by the M10 regulators to find consensus. Thereafter, their position was discussed with M10 industry members with a goal of obtaining consensus between regulators and industry members. A total of 2477 comments were received and discussions focus only on topics with major comments or new information. When no consensus between industry and regulators is possible, the regulators' position will prevail. Due to ICH policy, issues under discussion could not be presented at the 15th WRIB. Travel limitations brought on by the COVID-19 pandemic caused some delays with the ICH Working Group discussions, therefore the numbers of teleconference meetings was increased to further limit delays with issuance of the document. The anticipated completion date for the final step 4 sign off and adoption of the guideline is May 2022.

Regulatory Findings & Observations: EU Applications

Recent regulatory findings and observations from EU applications were discussed. Topics included the use of medium QCs, QC storage conditions, and ISR. An issue regarding medium QC levels was identified in the 2019 White Paper in Bioanalysis and recommendations were issued [27]. During validation and analysis of study samples, some applicants started to use QC sample concentration levels at LLOQ, 3 × LLOQ (low QC), ∼5–10% of the calibration curve range (medium QC), and ∼75% of the calibration curve range (high QC). Therefore, concerns were raised by regulators that a medium QC around 30–50% of the calibration curve range is lacking to monitor the precision and accuracy of the 10% to 75% range of the curve. To address this issue, regulators are now seeing data for 2 medium QC levels, one for the geometric mean (at ∼5–10% of the calibration curve range) and one for the arithmetic mean (at 30–50% of the calibration curve range).

Another previous topic identified was the storage of QCs for stability evaluations in a single tube or in several tubes. Taking multiple samples from a single tube can be considered a replicate evaluation instead of an independent evaluation. Since this observation, no major objections or concerns have been raised to date potentially due to storage tubes. This may be due to QCs being stored as expected or to no in-depth assessment of the SOPs since the issue is no longer considered a major issue.

ISR guidance was first provided by the EMA in 2011 [31] and has continued to be addressed as part of the clinical pharmacology and pharmacokinetics: questions and answers database [154]. Due to decentralized repeat use and mutual recognition procedures, it is still possible to receive studies carried out before the bioanalytical guideline came into force. Therefore, the absence of ISR should be handled by providing justification using possible metabolite back conversion, the obtained PK data in the study, the 90% confidence interval observed in the BE study, data from repeat analysis, and any other ISR data for the same drug compound obtained from the same laboratory. However, although the number of failed ISR evaluations is low, they cannot be waived by using this type of justification. An update was given on a case of failed ISR during the 2018 Regulatory Inputs [20]. The validated bioanalytical method, although within criteria, accuracy and precision, showed high variability; ISR data was well below 67%. The advice from the regulators was to revalidate the method using another CRO and to reanalyze study samples. The bioanalytical method was successfully revalidated as suggested with accuracy and precision well within criteria. ISR data for the reanalysis was well above 67%. The main hurdle for this achievement was obtaining sufficient stability data (greater than 3 years) but the final outcome was accepted.

The conclusion by regulators was that the number of major objections related to analytical issues is low; concerns raised are, in general, minor issues or clarifications. No applications were rejected due to inaccurate or imprecise bioanalytical methods.

UK MHRA

Requirements for Bioanalytical Electronic Data

Raw data refers to all original test facility records and documentation, or verified copies thereof, which are the result of the original observations and activities in a study [84]. Furthermore, the term source data refers to all information in original records and certified copies of original records of clinical findings, observations, or other activities in a clinical trial are necessary for the reconstruction and evaluation of the trial [85]. Derived data obtained by data processing should be collected and maintained in a secure manner, such that they are attributable, legible, contemporaneously recorded, and accurate whether raw data or a verified copy [155].

Data should be accessible for users within appropriately assigned functions, allow access for audits and inspections, and permit full reconstruction of these activities. Data integrity must be ensured with suitable security and audit trails in place. Data are required to be retainable, archived and backed-up. Suitable access to systems is also applicable to Quality Control and Quality Assurance groups within laboratories as well as regulators. For inspections, the accessibility requirement is for direct access of a copy and/or live system with suitable access rights to allow interaction with the data and reconstruction of data generation activities. Inspections require the review of audit trails and available reports as well as a verification of the archive process and how the data are protected at all times. How data can be accessed is discussed during inspection planning and should not be limited to just raw/source data; supporting system access is required as well. The use of and access to electronic data has allowed for remote inspections to be conducted using a combination of outputs (e.g., reports, submission of data and access to documents/systems) and shared desktops. Electronic data can be submitted to MHRA inspectors for upload into their own analytical software for inspection. Example inspection findings were shared. In one case, the laboratory did not provide sufficient access to electronic source data to allow verification of procedures by quality assurance (or quality control) personnel. Electronic data for audit was in effect a pdf document. Quality department staff did not have sufficient access or alternative methods to allow them to check SOP compliance within electronic systems to reconstruct activities. At another firm, audit trails were in use, but their content was not part of the internal review process. Furthermore, data in the form of text files could be saved outside of the secured network and then inserted into the secured folder without detection. An archiving finding detected that a laboratory was requested to perform further analysis by the sponsor. However, only paper print outs had been retained as electronic data had been deleted. Therefore, the data analysis could not be conducted. In summary electronic data systems should be designed to meet regulatory requirements and guidance. Consideration should be given to how supporting processes, access and use of the systems can any potentially impact on their effectiveness. MHRA inspectors will work closely with laboratories during planning to ensure suitable access is available for inspection.

Bioanalytical Observations & Findings

MHRA bioanalytical inspections over the last 18 months, globally and within the UK, have been a mixture of on-site inspections, fully remote inspections and a hybrid using both approaches. Although a remote inspection cannot fully replace an on-site inspection, the use of technology has permitted an effective inspection approach. Bioanalytical method validation guidance states that at least two QC sample levels should fall within the range of concentrations measured in study samples [31], however data has been presented where this was not respected. One finding was issued due to changes in criteria where a laboratory updated internal criteria for a positive/negative result without notifying the sponsor. The new criteria did not match those used at other laboratories analyzing the same sample types, creating data interpretation issues when the data was collated. A finding occurred when systems in place at the laboratory did not contain sufficient detail on how to manage different types of matrix samples leading to inappropriate storage at -20°C, which was not identified by the laboratory until after investigation into multiple analytical failures. A finding regarding sample stability was issued because the total time that a processed sample is stored must be concurrent (i.e., in this case autosampler and other storage times were added together).

Other common findings included:

• working practices that were not reflected by the procedures they claim to be following

• failure to follow procedures

• failure of quality activities to identify the non-conformances

• document design did not capture key information to allow reconstruction of study activities (e.g., location and times of sample storage)

• key areas of risks to data integrity were not identified or addressed

In summary, inspections continue to identify activities that fail to follow recognized guidance. Ineffective change control has resulted in issues for laboratories, especially where new activities are introduced. Communication (or lack of) has resulted in problems with results reporting, or regulatory findings due to lack of evidence/information being shared or recorded.

Biotherapeutic International Standards

Recent advances in Biotherapeutic International Standards were presented by the National Institute for Biological Standards and Control (NIBSC), a center of the MHRA. The NIBSC is responsible for the standardization and control of biological medicines playing a national and international role in assuring the quality of biologicals. International Standards provide a mechanism to assign biological potency to many biological products and to maintain a global system of secondary standards.

Biotherapeutic mAbs are a fast-growing group of medicines, and the first biosimilar mAb was approved in 2013. Advances in analytical technology have improved the knowledge of product heterogeneities and supported the development of a robust regulatory framework for these products with more than 30 biosimilar mAbs currently approved. Bioactivity is a critical quality attribute and bioassays are used during the biosimilar comparability studies (to the innovator clinical product, known as the reference medicinal product) and to assess product consistency during the life cycle of the mAb (using the manufacturer's proprietary reference standards). Manufacturing process changes post-approval are common and have the potential to impact product characteristics including bioactivity [156–158]. Multiple bioactivities typically contribute to the mAb mechanism of action and its clinical effects, and these may drift overtime as a result of manufacturing process changes, potentially leading to divergence amongst products. The regulatory framework cannot identify potential cumulative bioactivity drifts across products, jurisdictions, and over time, and ensuring consistency in bioactivity has emerged as a new regulatory challenge. WHO International Standards are developed as new tools to ensure consistency in the potency of mAbs. They are established by the Expert Committee in Biological Standardization (ECBS) and are publicly available lyophilized preparations optimized for long-term stability. They define bioactivity in International Units and support the performance and calibration of bioassays and in-house reference standards, facilitating global harmonization and traceability. Data from the international collaborative studies for the 1st WHO International Standard for rituximab and trastuzumab illustrate the benefits of using these reference preparations across multiple bioactivities, such as target binding or associated downstream cellular effects and Fc-effector functions like antibody dependent cell-mediated cytotoxicity (ADCC) or complement dependent cytotoxicity (CDC). However, it is emphasized that WHO International Standards for mAbs do not play a regulatory role in defining biosimilarity, do not define specific activity (units of activity per mass), and have no role in clinical labelling or product dosing [159], as these are the specific roles of the reference medicinal product under the scope of the product's license specifications. The use of International Standards is promoted by regulatory guidelines that advise the use of biological assays expressed in units of activity calibrated against an International Standard when available and appropriate for the assay utilized [115]. WHO International Standards have already been established for rituximab, infliximab, adalimumab, bevacizumab and trastuzumab and are widely used by manufacturers, governmental institutions, Pharmacopoeias, biotech companies and CROs. The current pipeline for development includes, amongst others, cetuximab and ustekinumab.

In summary, WHO International Standards for mAbs play a unique role in supporting bioassays and manufacturer reference standards allowing global harmonization and traceability both horizontally, across products and batches, and longitudinally, over time [160].

Health Canada

Observations During the Review of Pharmaceutical Drugs

Therapeutic Products Directorate (TPD) is Canada's regulator of prescription pharmaceutical drugs for human use. The Division of Biopharmaceutics Evaluation (DBE) is responsible for the review of BE studies and pivotal comparative bioavailability (BA) studies to support, for example, the bridging of formulations, food effects, fixed dose combination products, line extensions, formulation changes, alternate modes of administration (e.g., nasogastric tube, sprinkle study), and manufacturing site changes for a modified-release product. The presentation focused on situations in which the validated concentration range may not be representative of study sample concentrations and strategies to address samples that are above ULOQ. There are two scenarios typically used to deal with this situation. In the first, the calibration range is extended as per BMV guidelines [30–32], with new QC levels that appropriately reflect the study samples, and a new high QC concentration that should be validated (including matrix-based stability). In the second scenario, the calibration range is not extended but a validated dilution method is used to dilute samples that are above the ULOQ. Dilution QCs (DIQC), spiked above the highest concentration and diluted with the same dilution factor as study samples, are analyzed in analytical runs containing diluted study samples. A new high QC (typically sponsors use the DIQC) is used for matrix-based stability. The majority of submissions reviewed use an adaptation of the latter scenario. When reviewing this data, regulators will consider various factors to determine whether the approach used is acceptable (e.g., the number of affected samples that are C_max values, if stability data supports the study samples, the purpose of the study (supportive only or pivotal), and if QCs are still relevant to study samples).

Three case studies were presented to illustrate different approaches used and how the approaches were assessed by regulators. The first used a validated calibration curve range of 10 – 10000 ng/ml and QCs at 30, 500, 2000, 8000 ng/ml; the DIQC was 50,000 ng/ml. Study sample concentrations ranged from 3230–14,200 ng/ml resulting in only one QC level within the range of samples. Prior to submitting to the regulator, new QCs at 5000 ng/ml and 15000 ng/ml (DIQC) were prepared since all samples that were >ULOQ were <15,000 ng/ml and after dilution would fall between 5000–8000 ng/ml; the study samples were reanalyzed. The approach was acceptable and the regulator concluded that the existing stability was adequate based on the number of samples/subjects affected and the type of study (supportive only).

The second case study is another example of an acceptable approach used to address the situation of study samples being above the ULOQ. The validated calibration curve range could not be extended due to carryover concerns and 44% of study samples were above ULOQ. A DIQC covering maximum study sample concentrations was prepared and samples were pre-diluted and analyzed successfully. Matrix based stability (long-term, freeze-thaw, bench-top) was conducted using the DIQC.

The final case study involved a successful revalidation to change the calibration curve range from 0.05–10 ng/ml (QCs 0.14, 3 and 8 ng/ml) to 0.05–250 ng/ml (QCs 1.5, 87 and 200 ng/ml) prior to sample analysis. The study sample concentrations ranged from 41–900 ng/ml. A DIQC of 977 ng/ml was employed successfully for the above ULOQ study samples; however the stability assessments were not repeated with the new QCs. The original high QC of 8 ng/ml that was used for stability was not representative of the study sample concentrations and was not accepted.

Challenges Experienced During Review of Biologic Drugs

The Clinical Evaluation Divisions (CED) of the Biologics and Radiopharmaceutical Drugs Directorate (BRDD) at Health Canada presented challenges encountered during the review of biologic drugs. Key responsibilities for the CED are to review pivotal and non-pivotal nonclinical and clinical (safety and efficacy) studies, review pivotal comparative PK or PK/PD studies (e.g., biosimilars, formulation changes, bridging studies, line extensions, change in routes of administration), review pharmacometrics/population PK analysis results (e.g., to describe PK characteristics of a biologic drug, to support changes in dose/dosing regimen, to extrapolate dose/dosing regimens to pediatric indications, to justify/predict the dose in special populations), and review real world data/evidence and literature data (submissions relying on third-party data). Data integrity refers to data that are complete, consistent, accurate and reliable. It includes Good Documentation Practices and Good Data Management Practices and should be assured throughout the life cycle of a drug product. Data generated from clinical studies should be robust and accurate in order to support the efficacy, safety and quality of a drug product throughout its life cycle. If the integrity and trustworthiness of the data is in doubt, the generated clinical data is of limited value. Ensuring data integrity increases the regulator's confidence in the authenticity of the data, which in turn expedites the review process. Data integrity is a review challenge. Using case studies, examples were presented to demonstrate what was considered poor analytical practice.

In the first case study, hemolysis was observed in 5% of plasma samples collected from a PD study. During validation, however, the sponsor failed to demonstrate a lack of hemolytic effect as HQC samples failed to meet acceptance criteria in 0.5%, 1% and 2% hemolysed plasma samples. No further investigation was performed (e.g., concentration dependence). When requested to address the concerns of hemolysis, the sponsor responded that the method was appropriate for use because selectivity criteria were met at the LLOQ level. Furthermore, the sponsor stated that samples showing hemolysis had been repeated using an unvalidated assay and yielded similar results; hence, the repeated results were used to replace non-reportable values. The sponsor deemed that there was no impact on the study outcome because both sets of results demonstrated biosimilarity. However, Health Canada could not confirm this outcome due to the unvalidated method used.

In a second case, inadequate QC levels were used for sample analysis. In this example, due to sensitivity issues, the LLOQ was raised to the low QC concentration; however, no re-adjustment of the low QC concentration was carried out and a narrowed acceptance criteria was applied. As sample analysis was carried out using only medium and high QC samples, for some runs, only one QC level fell within the study sample concentrations. Health Canada questioned the adequacy of the QCs to accept or reject the runs.

Finally, inadequate descriptions of analytical repeats are still a major challenge during study data review. Often, only a list of repeat codes is submitted with unclear justification (e.g., client-requested reassay). The reasons for the repeats, a description of analytical issues and corrective measures taken, if any, are often not provided. In addition, there are differences seen in how samples are repeated and repeat values are reported and, the sponsor does not provide an explanation for why repeats are handled differently. This results in ambiguous reporting practices and raises concerns of bias. The incomplete documentation raises questions on whether the omissions were inadvertent or intentional and what other information could be missing. Health Canada's recommendation is to ensure SOPs are followed. Any SOP deviations should be reported along with a discussion of the consequences of such deviation(s). If corrective actions were taken, details should be provided. If BRDD is seeking clarification, consider whether one-line responses resolve data integrity concerns. Comprehensive reporting of data, particularly reassays was recommended since decisions on safety and efficacy cannot be made based on a list. Oversight of the entire development program of the biologic drug should be ensured. Lack of oversight causes confusion and unnecessarily increases the burden of review.

Brazil ANVISA

Updated Regulation & ICH Initiatives

The Anvisa regulation for the conduct of relative bioavailability/bioequivalence studies (BA/BE) was under public consultation [161]; an overview of the changes was presented. For PK study design, multiple dose studies can only be adopted when justified. For medicines containing vitamins, submissions associated with active pharmaceuticals that claim therapeutic properties and that are not food supplementation must present BA/BE studies. In addition to the primary parameters for determining bioequivalence of transdermal medicines (AUC_0–t and C_max), partial AUC should be considered by calculating the area from zero to half the dosing interval (initial partial AUC) and then to the last collection time (final partial AUC). Irritation and sensitization of the skin should also be evaluated. The use of achiral bioanalytical methods is acceptable for most BA/BE studies, however, the individual quantification of enantiomers must be performed when all the following requirements are met or unknown: 1) the enantiomers exhibit different pharmacokinetic and pharmacodynamics characteristics; and, 2) the AUC ratio of the enantiomers is modified due to differences in absorption speed. In regards to high variability drugs, the extent of the expansion must be defined based on the intra-individual variability of the comparator drug, observed in the BE study, according to the equation [LI,LS] = exp [±k·sWR]. The maximum extension accepted is 69.84–143.19%. On the topic of narrow therapeutic index drugs, the acceptance range for AUC should be set to 90.00–111.11%; the same may happen with C^max where it is of particular importance to safety, efficacy or monitoring of the drug level. Finally, for PD studies for topical corticosteroids, vasoconstriction assays, also known as the skin whitening assay, must be performed. Anvisa initiatives with ICH were presented. Anvisa is actively participating in almost 30 Expert Working Groups, 4 of which are related to bioanalysis. Anvisa is evaluating the impact of implementing ICH M9 [162], ICH M10 [32], ICH M12 [163] and ICH M13 [55].

WHO

Inspections of Bioequivalence Studies

During a BE study inspection, both clinical sites and bioanalytical sites can be inspected. BE study inspections are performed to verify that the CRO has adequate quality control and quality assurance systems in place which ensure the conduct of BE studies in compliance with applicable requirements. Inspections ensure that human subjects were protected from undue hazard or risk throughout clinical trials and recognized ethical standards were applied. Confirmation that study data and the respective results are scientifically valid and accurate is also an objective. When a medicinal product application is received and successfully screened, the CRO is selected based on a risk-based approach and listed on the WHO inspection plan. No inspection occurs without a submission dossier. A notification is sent in advance to confirm the availability of the site staff and the inspection plan is agreed cooperatively with the CRO. Study and site quality management documentation is requested for review prior to the inspection. Examples of requested documents include the CRO master file, list of computerized systems and SOPs, study reports, including study data such as chromatography data. The onsite inspection, typically lasting 3–5 days, consists of an opening meeting, tour of facility, staff interviews, review of documentation, quality management documentation, study documentation, source data including meta data on computerized systems, daily wrap-up sessions and a closing meeting. The inspection report is sent to the site 30 days from the last day of inspection and contains site and study product information, deficiencies and the respective classification (based on a specified definition), WHO references, and a cover letter with instructions on how to respond. A comprehensive corrective action/preventative action (CAPA) response is expected within 30 days of the receipt of the report. The CAPA plan is assessed by the inspection team and a closing letter is issued to the site. If the responses to the findings are acceptable, the site will be considered compliant, and a WHO Public Inspection Report (WHOPIR) will be published. If the response is deemed inadequate, the site will be considered as non-compliant and a WHOPIR will not be published. In the latter case, a follow-up inspection may be scheduled or the WHO product application may be rejected. In case of critical deficiencies (e.g., data integrity related findings), a Notice of Concern can be published. Inspections should be closed within 6 months from the date he inspection report was submitted. A WHO Public Inspection Report (WHOPIR) is a summary of the inspection report of a CRO, and is valid for 3 years. The WHOPIR is not a complete report but represents a summary of the highlights in the report. All confidential and proprietary information is removed. A list of applicable references is available within the inspection report.

SECTION 6 – Input from Regulatory Agencies on Immunogenicity, Biomarkers, Gene & Cell Therapy and Vaccines

Abbas Bandukwala¹³, Elana Cherry²⁵, Isabelle Cludts³⁶, Soma Ghosh¹³, Shirley Hopper³⁶, Akiko Ishii-Watabe³⁸, Susan Kirshner¹³, Kevin Maher¹³, Kimberly Maxfield¹³, Joao Pedras-Vasconcelos¹³, Mohsen Rajabi Abhari¹³, Yoshiro Saito⁴⁰, Dean Smith²⁵, Therese Solstad⁴¹, Daniela Verthelyi¹³, Meenu Wadhwa³⁶, Leslie Wagner¹³, Yow-Ming Wang¹³, Günter Waxenecker⁴², Haoheng Yan¹³ & Lucia Zhang²⁵

Authors are presented in alphabetical order of their last name.

The affiliations can be found at the beginning of the article.

The highlight of each WRIB conference is the annual input provided by regulators on immunogenicity, biomarkers, gene and cell therapy and vaccines. This year, regulators from US FDA, EU EMA, UK MHRA, Norway NoMA, Austria AGES, and Health Canada presented topics of interest to the Global Bioanalytical Community.

Immunogenicity

US FDA

New Tools to Inform Immunogenicity Risk Assessments: Possibilities & Challenges

All the information needed to assess the immunogenicity risk of a product should be summarized in the Integrated Summary of Immunogenicity (ISI), submitted in eCTD Section 5.3.5.3. This evolving document should nucleate all the information needed to perform an assessment of immunogenicity risk that is stage- or program-appropriate including the initial immunogenicity risk assessment, proposed tiered strategy and bioanalytical assays, clinical study design and sampling strategy, clinical immunogenicity data analysis, conclusions and risk mitigation, and post-marketing/life-cycle management plans. Early on, the immunogenicity risk assessment informs clinical trial design regarding study population, size, length, sampling times. In addition, it should include documentation for ADA assay development and testing, neutralizing antibody (NAb) assay development and testing, epitope specificity, and when appropriate cross reactive-ADA assay development and testing, and considerations regarding tolerization induction.

There are a host of in silico and in vitro tools that are used to inform the immunogenicity risk assessment prior to the clinical use for target selection and deimmunization of therapeutic candidates. These tools have the potential to improve the immunogenicity risk assessment but are currently in different stages of development. Those that are considered well-developed include methods that examine innate immune activation (in vitro), methods that assess binding to major histocompatibility complex (MHC) (insilico, in vitro), and methods that assess T cell activation (invitro, in vivo). Those lagging behind include methods to assess the risk of B cell epitopes and modeling tools to integrate product molecular and quality attributes, patient background, and clinical treatment. The FDA recently finalized a guidance on synthetic generic peptides [164] which discusses pre-clinical tools to assess the immunogenicity risk for generic peptides. The document outlines that risk is a measure of probability (due to population, treatment, product) and consequences (i.e., safety, efficacy). It was noted that if the active pharmaceutical ingredient (API) remains the same in the generic peptide, then the residual uncertainty is primarily due to the API sequence-related impurities and/or manufacturing process-related Impurities. Detection of unknown innate immune response modulating impurities (IIRMI) requires highly sensitive orthogonal assays with demonstrated capability of detecting clinically relevant levels of IIRMI. Whole blood, PBMC, purified immune cells (dendritic cells, monocytes, B cells), cell lines (monocytic/transformed or engineered to express specific pattern recognition receptors) can be used. PBMC, whole blood, and dendritic cells have high donor to donor variability, cellular viability, and unstable receptor expression therefore no less than 20 donors are recommended. Cell lines expressing defined sets of innate immune receptors are more reproducible but have a limited receptor array and poor response to aggregates or particulates. Formulation can also impact assay sensitivity for IIRMI detection. Readouts for these assays typically include proliferation, cell marker expression, cytokine secretion, and gene expression arrays. Regarding the assays to evaluate impurities with variations in the peptide sequence, in silico epitope prediction has several advantages, such as high throughput, allows for assessment of theoretical modifications, and multiple MHC coverage. There are, however, several limitations, particularly for evaluating unusual or non-natural amino acids or post-translational modifications. MHC are highly polymorphic and frequency is regional. Data supporting MHC binding is increasingly available, however training datasets on certain MHC alleles can be limited. In addition, the backend of the computational algorithm and validation is different for each provider and proprietary, which can become a challenge for regulators. In vitro assays examining MHC binding and T cell responses to the peptide impurities are also frequently used to assess the risk of these impurities. A number of challenges remain with these assays as they can have multiple formats: human blood-derived cell-based assays (DCs as APCs; CD4 T cells as effector cells), and protein-specific T cell amplification. Detailed and clear experimental design and culture conditions and justification of the method used, samples (size, MHC, variability, target population), culture conditions, concentration of the product used, readout selection (proliferation, cytokines, cell markers), and suitability controls that confirm sensitivity for naïve T cell responses is helpful for assessors when considering these assays. Regardless of the method used, discuss the assessment strategy early with regulators. Assessing the risk of product- and process-related impurities is not sufficient to determine the immunogenicity risk but can support an assessment of “relative” immunogenicity risk as compared to the product that was used in clinical trials. Orthogonal assays for product- and process-related impurities can contribute to the totality of evidence used to assess the potential immunogenicity risk of new impurities following a manufacturing change. As these tools become more robust and the relationship with clinical immunogenicity outcome strengthens, it might be possible to use them to aid in the design of smaller, more targeted clinical trials that assess product immunogenicity.

Immunogenicity Risk Assessment & COVID19 – Current Perspectives from CDER Office of Biotechnology Products

Sponsors should perform an Immunogenicity risk assessment on program and product risk factors as per 2014 FDA immunogenicity risk assessment guidance [165]. Principles to consider include analyzing product/CMC-related factors which could affect the immunogenic potential of the product, patient demographic factors that could enhance the likelihood of the patient population to produce an immune response to the product, and trial design-related factors that may increase likelihood of the study conditions to facilitate an immunogenic response. The immunogenicity risk assessment considerations for biologics for the treatment of Covid-19 should follow similar principles. Product/CMC-related factors should consider whether the amino acid sequence is derived from human or a non-human source. The degree of homology/ and cross-reactivity to an endogenous protein, is important for sponsors to consider given that non-human amino acid sequences are more immunogenic and thus more likely to lead to ADA that can impact product efficacy, while human sequences are less immunogenic but have a higher risk of leading to a deficiency syndrome in patients if the endogenous protein has a non-redundant biological function, impacting not only efficacy but also safety of the product.

Another factor to consider is whether the proposed product is a “first-in-class” product with little immunogenicity data available, compared to product with immunogenicity data available for other clinical indications. For both “first-in-class” and “prior-in-class experience” products, quality control strategy and product quality profiles are important to ensure the product has as low an immunogenic potential as possible. For an approved drug with a new administration route, Sponsors should consider the potential impact of the novel route on the product quality profile and immunogenic potential of the product.

Regarding patient demographics, sponsors should consider the age of the patient populations and known co-morbidities. Pediatric patients and adult patients often differ in the type and breadth of immunogenic responses to therapeutic drugs as well as degree of disease severity during SARS-CoV2 infection. Responses can also differ in younger adult individuals compared to older adults. The presence of pre-Covid-19 auto-immune or inflammatory disease, post-acute sequelae of SARS-COV2 infection [PASC]) is an additional important consideration, along with the potential clinical consequences of product immunogenicity in the specific patient population, and their impact on safety and efficacy.

Trial design-related factors can also impact the immunogenicity risk, including the stage of Covid-19 disease the product is being administered at, such as in the asymptomatic/pre-symptomatic stage, early acute inflammatory, late acute inflammatory stage. If the product is proposed to treat PASC or SARS-CoV2 reinfection, this may modify the overall immunogenicity risk of the product.

Finally, there are additional considerations for the immunogenicity risk assessment for clinical studies impacted by intercurrent SARS-CoV2 infections. The spectrum of clinical syndromes associated with Covid-19 disease could impact immunogenicity of drug products under study for unrelated clinical indications. Given our current lack of understanding on the potential immunomodulatory impact of Covid-19 and PASC, ongoing clinical trials whose study population has been infected by SARS-CoV2 may consider collecting information on whether the patient had Covid-19 and/or PASC. They should also analyze whether infection status impacted product immunogenicity rates (pre- versus post-Covid-19).

ADA Assay Comparability

Immunogenicity at the FDA is reviewed by CDER for monoclonal antibodies, growth factors, fusion proteins, cytokines, enzymes, therapeutic toxins and by CBER for allergenics, blood and blood components including clotting factors, cellular and gene therapies, and vaccines. The 2019 immunogenicity guidance discusses limitations of comparing ADA incidence across products [90]. In general, comparison is not recommended. However, in cases where it is necessary, “a head to-head clinical study from which samples obtained are tested using an assay demonstrated to have equivalent sensitivity and specificity for antibodies against both therapeutic protein products.” Assays used should be comparable across laboratories. When different sets of assays are used within same immunogenicity program, immunogenicity data supporting labelling should be based on the analysis of pivotal study samples tested using the latest iteration of fully validated immunogenicity assays. This could include an integrated immunogenicity dataset from multiple clinical studies, using samples initially analyzed with an earlier set of assays. In this case, it is recommended that samples be reanalyzed using the latest validated assays by performing a bridging study that includes prior ADA positives, negatives and borderline samples. When this occurs, a justification for the choice of dataset should be provided and data concordance for prior positives (strong, and borderline) versus negatives examined. Immunogenicity assay comparability using platform agnostic methods to compare immunogenicity rates across different drug programs and different sponsors is currently not recommended.

FDA Regulatory Perspective on Neutralizing Antibody Assays

A NAb is a subset of ADA that can interfere with the clinical activity of a therapeutic protein product by preventing product binding to the target, interfering with the pharmacological activity (such as receptor–ligand interaction), or preventing product cellular uptake. The NAb directly inhibits drug function (loss of efficacy) and potentially inhibits biological function of the endogenous counterpart (safety). NAbs may be more effective in directly impacting efficacy (e.g., IFN-β and IL-2) compared to non-neutralizing binding ADA. When developing a NAb assay strategy, a risk-based approach should be used. NAb assays are critical when neutralizing immunogenicity poses a high-risk to patient safety. In these cases, a qualified/validated assay should be implemented early (Phase I). For medium to low risk, a validated assay is needed prior to testing clinical Phase III study samples. Two types of assays are generally used for NAb assessments: cell-based assays (CBA) and non-cell-based competitive LBAs (CLBA). FDA recommended that neutralization assays use a CBA format; an alternative strategy might be acceptable but regulators should be consulted. Factors that affect the selection of an appropriate assay format are the mechanism of action (MOA) of the therapeutic protein product, as well as selectivity, sensitivity, precision, and robustness of the assay. Further details are described by Wu B et al. [166]. In selected cases, where there is a highly sensitive PD marker or an appropriately designed PK assay (or both) that generate data that inform clinical activity, it may be possible to use these in lieu of a NAb assay. This determination should be made in consultation with regulators. Examples for assay format selection were presented. One example was presented for the mAb trastuzumab; the MOA is the inhibition of cell proliferation and ADCC [167]. FDA has approved 5 trastuzumab biosimilar products for which the NAb assays were a ADCC CBA or a CLBA. Another example, the enzyme replacement therapy for Sebelipase alfa (recombinant LAL) for treating Lysosomal Acid Lipase (LAL) Deficiency, has an MOA of cellular uptake and enzymatic activity in the lysosome. The NAb assays include a cellular uptake assay and an in vitro enzyme assay. For the novel modality of caplacizumab-yhdp, which is an anti-vWF nanobody (modified) for treating aTTP, the MOA is the inhibition of the interaction between vWF and platelets. The NAb assays used based on this MOA are a functional NAb assay (CLBA) with limited drug tolerance and an epitope characterization assay solely detecting ADA binding to the CDRs of the nanobody. Several NAb assay validation issues were discussed. 1) For assay cut point, NAb assay cut point is recommended to set at 1% false-positive rate confirmed using ADA positive samples. 2) If NAb assay sensitivity is poor, some approaches to resolve the issue may be analyzing the activity curve to select the optimal therapeutic product concentration in the assay; changing from a bioassay to CLBA, if applicable; selecting a PC with a higher neutralizing activity to provide better estimate of the assay sensitivity; and integrate the clinical data interpretation for assay suitability justification. 3) If precision of the cell-based assay is poor, consider further optimization, analyst training, and use of more replicates. 4) Matrix components may enhance or inhibit the activity of a therapeutic protein product in bioassays. Test results from baseline pre-exposure samples may be informative. 5) In regard to specificity, a confirmatory NAb assay is not common, but it should be considered in certain cases. 6) For drug tolerance, consider introducing procedures to remove the drug in patient samples, changing from a bioassay to CLBA, if applicable, and/or integrate the clinical data interpretation for assay suitability justification.

FDA CDER Immunogenicity Review Committee (IRC)

The IRC is a cross-office committee that serves in a broad oversight capacity to the OND review divisions, OCP, OBP, and other offices and divisions in CDER. They consider development programs to evaluate BLAs, NDAs and ANDAs with product-specific immunogenicity concerns. The IRC mission is to serve as a platform to ensure the application of consistent principles and policies for immunogenicity risk management and to provide a forum for defining and leading multidisciplinary and integrated approaches to immunogenicity risk assessments, clinical trial expectations, and assay development. The IRC framework at the product class level is to provide recommendations on the development of immunogenicity risk assessments. The policy and research level role is to identify broader trends, inconsistencies and gaps in immunogenicity evaluations. Their application/regulatory level role is to provide evidence-based and consistent recommendations. They also communicate regulatory recommendations, communicate frameworks, and foster filling knowledge gaps. IRC meetings occur monthly and address questions related to immunogenicity that are interdisciplinary in nature. Meetings may discuss application-specific review issues, policy-related issues, emerging immunogenicity-related issues, preliminary comments to meeting requests, immunogenicity related-research projects or proposals, presentations, letters, or other documents. Some past and ongoing IRC topics include the role of genotype/phenotype (residual enzyme activity) and cross-reactive immunological material-status for enzyme replacement therapies for inborn errors of metabolism; the immunogenicity risk of enzyme replacement therapies in single dose first-in-human studies using healthy volunteers; and, the design of PMR/PMC to detect long-term toxicities and clinical management of pegylated products. Other topics include determining the immunogenicity risk assessments for proposed biosimilars to insulin products, immunogenicity information and product labeling, and use of biological assays to evaluate product-related impurities that may alter the immunogenic potential of generic products.

Recent FDA Observations of Immunogenicity Assays

Recent observations from remote record review of immunogenicity assays were presented. The critical validation parameters that were evaluated are cut point (screening, confirmatory, titer, neutralizing antibody), assay positive controls (low positive controls and high positive controls), precision (screening, confirmatory, neutralizing antibody), assay sensitivity, selectivity and specificity, hemolysis and lipemia effect, drug tolerance, and stability (freeze/thaw, short-term). Common observations reported include improperly established cut points and relevant low positive controls that were not used in confirmatory assays. The inadequate precision of confirmatory assays and inadequate drug tolerance were also observed. Some examples of failed runs and failed QCs or PCs not being included in global precision and accuracy statistics during method validation were seen, as well as not having established acceptance criteria to exclude data during sample analysis. In one case, the firm did not properly determine the NAb cut point, therefore, the firm was not able to detect true NAb-positive post-dose samples. In another case, the firm did not establish an adequate confirmatory cut point (CCP). The CCP was established using a 99.9% confidence interval. Samples that were reported as conclusive were analyzed in tier 3 and several samples were confirmed positive using a new CCP calculated using a 99% confidence interval. An observation regarding assay precision was shown where the inter-assay precession of the confirmatory assay was not determined. To confirm the positives, the treated to untreated ratio should be <0.2606 which resulted in the high PC CV% >30%. However, the low PC was not confirmed positive. Cases of inadequate drug tolerance were also seen where the presence of drug at the observed concentrations was shown to inhibit detection of ADAs in the screening assay.

Austria AGES/EU EMA

Experience with the Integrated Summary of Immunogenicity in the EU

Immunogenicity data are generated throughout the whole course of drug development by researchers, nonclinical experts, assay developers, biotechnologists, statisticians, and clinicians. Immunogenicity data are also dispersed to numerous locations. Therefore, “it is recommended that this summary is placed in Chapter 2.7.2.4 Special Studies or, if more detailed, in Chapter 5.3.5.3 of the CTD. Both the planning and the evaluation of immunogenicity studies of a biological product are multidisciplinary exercises. Therefore, it is recommended that the applicant includes an integrated summary of immunogenicity in the application, including a risk assessment to support the selected immunogenicity program. The summary should be concise and contain links to the appropriate chapters of the application” [168]. To evaluate unwanted immunogenicity, in silico assays can be used for early de-risking (screening) and may be demanded to help understanding clinical signals. There are several platforms (public, academic), partly with mathematical models, however non-reproducible results have been seen on public platforms. Multiple molecules have already been engineered aiming reduced immunogenicity, but none has yet been challenged in a clinical setting. Usually animal models lack predictivity for immunogenicity in humans but for upcoming modalities (e.g., cellular, gene therapeutics) there is an urgent need for new analytical methods. Generally, in vivo models should be exposed to preparations that are representative of the clinical trial material as the use of artificially degenerated material likely generates misleading data. To present immunogenicity data, adapt the format to the needs. The integrated summary of immunogenicity is one option. There is no strict requirement for an ISI, but it may help to conclusively present multidisciplinary data and make data comprehensible.

UK MHRA

Development of Reference Material as Positive Controls for ADA Assays

Testing for immunogenicity is a regulatory expectation for product approval [90,135] and requires validated assays for ADA measurement and evaluation of ADA impact on clinical outcomes. The role of the PC in assay development and validation is important for sensitivity, selectivity, specificity, and drug tolerance evaluation. It is also an assay performance indicator. PCs are monoclonal or polyclonal, can be sourced commercially or generated in-house, and be of animal or human origin. Currently, assays are semi-quantitative as there is no reference material for antibody assays, with the exception of the WHO EPO antibody panel. For some mAb therapeutics, e.g., TNF-α antagonists, clinical decisions could be facilitated by Therapeutic Drug Monitoring (TDM), which includes measurement of drug levels and detection of ADA (pos/neg, concentration or titer). In clinical settings there are options for the use of commercial kits (mostly ELISA; some lateral flow) or in-house assays. Different commercial kits use different positive controls (e.g., mouse mAb, human mAb), different units (AU, AU/ml, ng/ml) and different cut-off criteria; therefore assays are not interchangeable, and the interpretation of results can be problematic. There is clearly a need for reference materials to standardize ADA testing. The NIBSC ADA standardization project was endorsed by WHO ECBS in 2016 as the proposed first WHO International Standards (or reference panels) for antibodies for use in immunogenicity assessments of biotherapeutic products [169]. The aim is to provide reference antibody or antibody panels as positive controls to standardize ADA testing across different assay platforms and laboratories for several mAb therapeutics. These standards could potentially be used by product manufacturers, commercial kit suppliers, clinical laboratories, researchers, regulatory agencies, and/or control laboratories.

Biological reference materials are only useful if they retain activity over time, and NIBSC has world leading research and development capability in freeze-drying methods for biologicals. Multi-center collaborative studies are essential for demonstrating that the proposed standard is fit for purpose. Aims of collaborative studies may include confirmation/measurement of activity, demonstration that the candidate is suitable for calibration of other standards and assessment of relevant test samples, comparison of two or more candidate standards, value assignment, assessment of stability, and assessment of different assay methods.

Development of an International Standard for infliximab ADA is currently underway. The ADAs for the lyophilized material were provided by the ABIRISK consortium and consist of purified human mAbs, isolated and cloned from patients' PBMCs. The study steps for the assessment of ADA material include development of binding and neutralization assays; comparison of mAbs sourced from collaborators and a commercial supplier; trial fills; selection of formulation; 3 months stability testing; and definitive fills. Five formulations were tested (sodium citrate-based & glutamic acid-based) and materials assessed for binding and neutralizing activities (compared with bulk material), protein determination, integrity by DW-SE-HPLC, and short-term stability. In the definitive fill, the final antibody content is 50 μg per ampoule. The collaborative study for infliximab ADA is ongoing, with 21 labs in 11 countries recruited, and is expected to bring in data from binding assays using various platforms and from neutralization assays (cell-based/non-cell-based). Material to be tested has been sent to participants and included lyophilized preparations, liquid mAb preparations and serum samples from healthy controls and infliximab-treated patients (ADA pos or neg, different titers, no or negligible amount of circulating drug). Participants were asked to perform 3 independent assays and report quantitative results (titers or concentration) for the samples, relative to in-house or kit standards and relative to the lyophilized material provided. All data will be statistically analyzed at NIBSC. ADA standards could allow for assay performance monitoring, facilitate the immunogenicity assessment of originators and biosimilars, facilitate monitoring of immunogenicity as part of pharmacovigilance, and improve harmonization of ADA tests (including commercial kits). They could enable evaluation of the clinical and cost effectiveness of TDM-based treatment strategies and facilitate clinical decisions when linked with measurement of drug trough levels. This may lead to better patient outcomes by applying appropriate clinical criteria (e.g., discontinuation of treatment, dosing escalation or de-escalation, switch to another therapeutic) and allowing individualized treatments.

Biomarkers

US FDA

FDA's Perspective on Liquid Biopsy In Vitro Diagnostic Tests: Challenges & Opportunities

The FDA review of in vitro diagnostic devices for liquid biopsy tests evaluates intended use, validation of pre-analytical conditions, analytical validation, and clinical validation. The intended use evaluates what the device measures, specimen type(s), intended population, clinical indications, distribution of test system, and instrument system. Key pre-analytical considerations should evaluate factors influencing quality and quantity of cell free and circulating tumor DNA for reproducible and accurate results such as performance of specimen collection device (blood collection tubes), specimen processing and preparation, and collected information on patient characteristics that could impact tumor shedding, etc. Types of studies needed for analytical validation depend on the technology and the intended use claims including but not limited to sensitivity (limit of blank, limit of detection), specificity (exclusivity/cross reactivity/interference), precision (repeatability/reproducibility), accuracy (reference method comparison), robustness, stability, and Contrived Sample Functional Characterization Study if non-clinical or contrived samples are used for some analytical validation studies to supplement the intended use clinical specimens due to the low prevalence of the biomarker/analyte, etc. For devices with quantitative claims, additional experiments such as LLOQ and linearity should be performed.

For clinical validation of a companion diagnostic (CDx) assay, the recommended approach is for the drug and CDx assay to be validated in the same pivotal clinical trial. In many instances the clinical trial assay (CTA) used for patient enrollment is not the final to be marketed assay (final CDx). In these circumstances, a bridging study is needed to demonstrate that the efficacy estimated from patients enrolled using the CTA is the same as the efficacy based on testing patients with the final CDx assay. Therefore, all the enrolled patient samples (i.e., all biomarker-positives) and a random subset of biomarker-negatives should be retested with the final CDx assay. The clinical validation of the CDx should be based on the primary efficacy population of the drug. To demonstrate robust clinical performance of the CDx assay, a high proportion of evaluable samples from the drug efficacy population should be re-tested in the CDx bridging study. When using multiple CTAs for enrollment (e.g., for rare biomarkers), each of the CTAs should be fully specified, and their cutoffs should be established and locked down prior to using these tests for patient enrollment in the registrational study. It is strongly advised that the sample sets used for training and validation are independent of each other. Test sites should meet minimum analytical validation requirements (i.e., minimum acceptable performance characteristics for assay sensitivity, specificity, accuracy, precision, minimum assay input studies, etc.) such that patients are enrolled using local tests that identify the appropriate patient population. Information on CTAs including test methodology/name, test cut-off, specimen type used, and brief analytical validation should be available to evaluate if test results are comparable and identify the same group of patients that are receiving the drug. Biomarker calling rules or variant classification rules, if applicable, should be pre-specified such that the same biomarker classification rules are used to determine biomarker positivity across all enrollment sites. All clinical studies and sites should follow the same protocol for specimen collection, banking, and processing. Both tissue and blood samples should be banked from all enrolled patients to evaluate the concordance between tissue and liquid biopsy test results and to evaluate the potential differences in drug efficacy between subgroups of patients (i.e., tissue+/plasma+, tissue+/plasma-, and tissue-/plasma+) as defined by each test. This information is needed for drug and device labeling.

Approved liquid biopsy PCR tests with CDx claim(s) include the Cobas EGFR Mutation Test v2 (refer to Summary of Safety and Effectiveness Data for P150044 and P150047) and the therascreen PIK3CA RGQ PCR kit (refer to Summary of Safety and Effectiveness Data for P190004). Approved liquid biopsy next generation sequencing tests with CDx and tumor profiling claims are the Guardant360 CDx Test (refer to summary of Safety and Effectiveness Data for P200010) and the Foundation One Liquid CDx test (refer to Summary of Safety and Effectiveness Data for P190032, P200006, and P200016).

On the horizon are liquid biopsy devices for detection of molecular residual disease. Here, considerations are clinical uses and clinical study designs with claims such as prognosis, escalation/de-escalation of treatment, etc., and validation studies to evaluate the pre-analytical and analytical performance. For liquid biopsy cancer screening and early detection tests specifically, the considerations are the elements of clinical study design (sample size, study endpoints, etc.). The evaluation of benefit versus risk (B/R) is critical and should be based on per cancer versus multiple cancers pooled together and on head-to-head comparison with cancers having standard of care testing versus those without existing standard of care tests. The evaluation of the risks, including anxiety as a result of the ‘diagnostic odyssey’, and potentially harmful side-effects of follow-up procedures to confirm the cancer, is also critical. Additionally, these tests also need studies to evaluate the pre-analytical and analytical performance.

Validation of Flow Cytometric In Vitro Diagnostic Devices

There are several device characteristics, unique to flow cytometry, that should be considered during the design of a validation protocol. This process begins with the definition of the Intended Use (IU) that defines the analytes and measurands (CD3, CD19, ± versus %, cells/μl, or molecules/cell), specimen type, and the instrument. For in vitro diagnostic devices (IVDs), the Indication for Use is typically combined with the IU and defines the intended patient population (condition, gender, age), point of care, and the clinical use (diagnostic, prognostic, predictive, monitoring, screening or confirmatory). Key components of IVD validation include analytical specificity, measuring range, analytical sensitivity [170], linearity [171], precision [172], specimen stability [173], and accuracy which can be applied to flow cytometry. Analytical specificity for cytometry is dependent on detection reagents (antibody clonality, class, fluorochrome conjugate, cluster of differentiation [CD] designation). Non-CD markers can be characterized for specificity by Western blots, competitive inhibition, lineage specific staining, etc. Antibody reagents are typically optimized with the creation of antibody binding curves. Special consideration is needed for multicolor staining (stearic interference, FRET, fluorescence spillover) in order to ensure the specificity of the device. Analytical specificity also has an instrument component which requires consideration of cytometry configuration and standardization of photomultiplier tube voltages and compensation. Other components to be included in the validation include the protocol (covering data acquisition and analysis) and matrix comparison and interference studies. Specimen stability should be established early to avoid the impact of specimen age associated artifacts upon the validation results. Precision studies should address repeatability (1 site) and reproducibility (3 sites, at least 1 in the US). Specimens to use should be drawn from the IU population and include analyte concentrations across AMR and around clinical decision points.

For flow cytometry, the concept and validation of “analytical sensitivity” may require consideration of several distinct, but related concepts, as applicable: 1) rare event counting with cell enumeration (for semi-quantitative devices, e.g., MRD with definition of limit of blank/ limit of detection/LLOQ following CLSI EP17-A2 [170]; 2) separation of dim events from background, which may include validation of the limits of fluorescence detection, and incorporation of fluorescence intensity standards. Validation of linearity for flow cytometry also requires consideration of several concepts, as one or more may be relevant, depending on the IU. The approach to validation of flow cytometric linearity will differ, depending upon whether one is assessing the linearity of cell enumeration (cells/uL following CLSI EP06-A2) [171] or photomultiplier tube linearity. When possible, analytical performance may be tested using method comparison versus predicate (e.g., comparison to a previously cleared IVD having the same IU and technological characteristics). Clinical accuracy may be assessed relative to diagnosis or clinical outcome. This is straight-forward when the phenotype is pathognomonic, the phenotype is invariable, or there is a single definition of the clinical condition.

21st Century Cures: Biomarker Qualification & Analytical Guidance

21st Century Cures and PDUFA VI increasingly place FDA as an active participant in drug development, broadening our traditional regulatory role. It formalizes a three-step submission process with a Letter of Intent (LOI), Qualification Plan (QP), and Full Qualification Package (FQP). The FDA submission decision (Accept or Not Accept) and a transparent process mean that all stakeholders are aware of the tools in development, stage, and FDA determinations and recommendations. For acceptance of a biomarker with 21st Century Cures, the acceptance decision for each submission (LOI, QP, FQP) is based upon the scientific merit for the proposed biomarker addressing an impactful drug development need, having enough information to suggest a likelihood of success, and the feasibility of the proposed analytical biomarker measurement approach [174]. Modification of the qualification determination is based upon new information that alters the conclusions that supported the original determination. Analytical guidance is given for biomarker qualification, context of use, and technologies. It does not replace requirements for FDA submissions or other guidance documents. Guidance is also given for pre-analytical variables (sample collection, handling, processing, storage, reference standards, performance characteristics, measurement method description, statistical considerations, and acceptable performance [174]. Key analytical characteristics are accuracy/relative accuracy, measurement range, precision, repeatability, reproducibility, analytical specificity, and limits of detection/limits of quantitation. Accuracy should be evaluated with the biomarker of interest, phantoms, other reference materials, retrospective data, and animal testing. Repeatability and reproducibility should be evaluated at different sites, with different users, and different measurement device manufacturers. Other characteristics may be added based on COU, measurement method technology, and multiple methods.

UK MHRA

Using Biomarker-Based Assays in Drug Trials – A Regulator's Perspective

The new EU In Vitro Diagnostic Medical Device Regulations (IVDR) will fully apply in EU Member States and Northern Ireland from 26 May 2022 [175]. For Great Britain, Directive 98/79/EC on in vitro diagnostic medical devices (EU IVDD) will continue to be given effect in UK law through the UK Medical Devices Regulations 2002 [176]. CE marking will continue to be recognized in Great Britain until 30 June 2023 whether in conformance with EU IVDD or EU IVDR. The UK Conformity Assessed (UKCA) mark will be required in Great Britain from 1 July 2023 in order to place an IVD on the Great Britain market. Until then, manufacturers can use the UKCA mark on a voluntary basis. Requirements are based on those of the EU IVDD. Importantly, the UKCA mark is not recognized in the EU, European Economic Area or Northern Ireland markets. CE marking is required in Northern Ireland. Regarding clinical study terminology, ‘clinical trial’ typically refers to a clinical trial of an investigation medicinal product (CTIMP). For medical devices, the term ‘clinical investigation’ is used instead. Trials which determine the clinical performance of an assay are ‘IVD performance evaluation studies’. More than one of these categories could apply to a single clinical study. Biomarker assays are commonly used in CTIMPs for a variety of purposes including characterization of the study population (results may or may not be used for enrolment decisions), measurement of endpoints (primary or key secondary endpoints such as agreed surrogate endpoints, exploratory endpoints), and management of trial subjects (e.g., safety biomarker as part of criteria for dose reduction or stopping). All biomarker assays used in CTIMPs as IVDs to enroll subjects or allocate study treatment must have the relevant mark of conformity unless an exemption applies. Where clinical performance of the IVD is yet to be demonstrated, the mark of conformity need only be for the analytical performance of the IVD (i.e., biomarker detection). This will include reagents, equipment, calibrators, controls and software and likely to be via self-certification under the current UK regulations. The MHRA is considering whether there should be additional regulatory scrutiny of the analytical performance for some biomarker assays used in clinical trials. CTIMPs which determine the clinical performance of the biomarker assay need to be additionally registered as IVD performance evaluation studies, which results in regulatory scrutiny of analytical performance. However, not all biomarker assays used in CTIMPs are a ‘device for performance evaluation’. For biomarker assays used in CTIMPs to enroll subjects or allocate study treatment, the potential risk to study subjects of an incorrect biomarker result should be considered.

Japan MHLW

Points to Consider Document on Biomarker Assay Validation (BAV) in Japan

The Japan points to consider document on biomarker assay validation (BAV) was finalized in March 2021 by the government-private research group [177] with a scope of harmonizing BAV in Japan close the Critical Path Institute document [178]. The molecules in scope include endogenous metabolites, peptides, and proteins and the methods in scope were MS and LBA, but excluding immunohistochemistry, flow cytometry, genomics, and MS imaging. The document applies to biomarkers as drug developmental tools (excluding CDx, clinical chemistry) and biomarkers used for regulatory decision making (excluding ones for exploration and internal decision making). Assays for biomarkers to be used for drug efficacy and/or adverse reactions as a part of 1) endpoints for clinical evaluation, and 2) supporting the post-marketing evaluation should be fully assessed and validated analytically. The principles for BAV assessments (parameters and acceptance criteria) are based on their intended use or characteristics (difficult to uniformly establish acceptance criteria, in contrast to BMV). The document describes “Points to consider” and “Recommendations” with no general acceptance criteria (case by case basis). Necessary validation parameters should be determined by the sponsor and performed based on the fit-for-purpose principle. Pre-determined acceptance criteria should be described in the validation protocol according to the biomarker's COU. For accuracy and precision, it was recommended to perform the evaluation using QC samples prepared directly in authentic matrix, diluting an authentic matrix with a surrogate matrix, spiking an authentic matrix with known concentrations of a reference standard, or spiking a surrogate matrix with known concentrations of a reference standard, depending on the target biomarker concentrations. At least three replicates each of at least four concentration levels in different batches should be analyzed. Even when QC samples are prepared with a surrogate matrix, the use of an authentic matrix from at least one lot is advisable. For low-molecular-weight analytes or when an appropriate reference standard is not available (e.g., same structure), relative accuracy can be allowed. When matrices containing endogenous substances are used to prepare QC samples, concentrations of the endogenous substances in the matrices should be analyzed to select a matrix that does not affect the assay or to calculate accuracy using a pre-defined formula. Recommendations for parallelism were described. Parallelism is unnecessary in the case of a chromatographic method of an endogenous molecule using SIL-standard and if authentic biological matrix is used. Otherwise, it is recommended to consider the evaluation of parallelism with multiple lots of matrices. If high concentration study samples are not available during an LCMS validation, it is recommended to spike the same endogenous molecule or its stable isotope; for LBA, spiking recombinant proteins is allowable. If low concentration study samples are not available during validation, the use of high concentration samples diluted with surrogate matrix may be used, and the evaluation repeated when low concentration samples are available. Regarding concentration levels, the actual concentration range should be considered; it is recommended to use ≥3 levels with >1 replicates. A glossary of 55 words was included, referring to domestic BMV guidelines [30–32]. Definitions included fit for purpose, parallelism, and surrogate reference standard. An addendum was provided as an example for the evaluation of analytical methods for purposes other than inclusion in an application dossier. However, sponsors should decide on the required evaluation parameters and their acceptance criteria, based on the COU of each biomarker. Future plans will focus on BAV for qPCR and flow cytometry.

Gene Therapy, Cell Therapy & Vaccines

US FDA

Immunogenicity of CAR-T Cells

A wide variety of gene therapy products are regulated by the Office of Tissues and Advanced Therapies (OTAT) in CBER such as plasmids, bacterial vectors, viral vectors, genome editing cells, genetically modified cells, etc., for a wide variety of indications. For CAR-T cells, there are concerns with inflammatory toxicities, specificity (on-target off-tumor and off-target), resistance/relapses, and immunogenicity of transgenes and genome-editing components. A CAR-T cell therapy case study was discussed where humoral and cellular immune responses in Anti-CAIX CAR-T cells were detected during treatment of renal cell carcinoma [179]. Eleven patients were treated with scFv derived from a murine antibody. Anti-CAR antibodies were detected in almost all patients with increasing titers remaining detectable throughout the evaluation period. Anti-CAR-T cell response was detected in 7/9 patients. Interestingly, 2 patients showed T cell response against retroviral vector backbone. CAR-T cells did not persist for an extended period of time (less than 2 months). A second case study of immune response against Cas9 during genome-edited T-cell receptor (TCR)-T cell therapy was described for Autologous anti-NY-ESO-1 TCR-T cells [180]. CRISPR/Cas9 is used to remove endogenous TCR and PD-1. Residual Cas9 levels were below the limit of detection of the assay at the end of production. Anti-Cas9 antibodies and T cells were detected in all 3 patients at baseline. Levels of Cas9 antibodies did not increase after the gene therapy, and modified T cells persisted for up to 9 months. This suggests that immunogenicity concerns of Cas9 can be mitigated by optimization of the manufacturing process. For FDA approved CAR-T cell products Kymriah, Yescarta, Tecartus, Breyanzi, and Abecma, immunogenicity (humoral response against CAR) was seen in the clinical studies but its impact on safety and efficacy was not found or was not conclusive. Due to the relatively small number of samples that were evaluated and since T cell response was not assessed, these findings are not conclusive. Data so far obtained from approved therapies is promising, but CAR-T cells are getting increasingly more complex in design and manufacturing. Furthermore, the approved products are autologous and the immunogenicity risks remain unclear with allogeneic products. It is important to note that the CAR-T cell field is still very young and all the risks associated with this product is not fully understood. CAR-T cell immunogenicity and its impact on safety and efficacy is not completely known. Each product type is unique; thus risks may vary so product-specific immunogenicity risk assessments should be conducted. Increased characterization of product and manufacturing processes may also help identify product attributes to predict immunogenicity and reduce product- and process-related risks.

Understanding & Navigating the Immune Responses to CRISPR-Associated Nuclease Cas9

When ADAs develop due to an immune response, these may or may not affect the activity of the drug. Antibodies that affect drug activity by binding to active protein domains are called NAbs, however non-neutralizing antibodies are not necessarily benign as they can affect the PK/PD profile and cause loss of tissue targeting. ADAs can also cross-react with endogenous proteins or elicit anaphylactic reactions. An important research focus of CRISPR Cas technology, and of regulatory safety concern, has been the unpredictable off-target effects that could lead to lethal consequences. As in vivo clinical applications expand, immunogenicity is likely to be a key regulatory concern. Cas proteins are of bacterial origin and thus, in the high immunogenicity risk category per FDA Guidance [165]. Speculation regarding immune responses to Cas proteins has existed since the invention of the technology. In 2018, three publications reported pre-existing T and B cell responses to Cas9 [181–183]. Despite the use of different methods, cut-points, samples, etc., all three studies found that at least a fraction of the population exhibits pre-existing immunity to Cas9. Cas immunogenicity is not the same as with therapeutic proteins. Cas9 may be introduced in vivo as DNA, mRNA or RNP. Delivery by Cas9 DNA/mRNA would engage MHC class I molecules and CD8⁺ T cells. RNPs would engage the MHC class II and CD4⁺ T cells. Immunogenicity data should be assessed considering the role of MHC restriction in T cell activation. The appropriate T cell subpopulation (CD4⁺/CD8⁺) and MHC Class (I/II) should be used for assessment. The potential promiscuous epitopes on Cas9 can activate IFN-γ, which activates innate and adaptive immune responses; this can trigger class-switching of B cell receptors/antibodies from IgM to IgG2, TNF-α which is associated with maturation of dendritic cells permitting antigen presentation, and IL-2 which induces clonal expansion of effector T cells primed with the antigen. Reproducible, validated assays with defined cut points to study immune responses to Cas9 are currently lacking. For instance, if studies do not have consistent and validated methods and statistical tools for determining a “cut-point” for identifying a subject as positive, there will be wide variations in the results. Model systems with meaningful clinical association representative of the population of interest are needed. For instance, cells for ex vivo assays should be sourced from a cohort of donors that is of the population that will receive the treatment representative (e.g., vis-à-vis HLA distribution) [184]. Well-characterized GMP quality reagents with controlled contaminants are also needed. For instance, academic researchers procure laboratory grade Cas9 from numerous suppliers. These have varying amounts of endotoxin which can affect responses in assays used to measure T cell responses [185]. There are published data relating to likely clinical consequences of Cas9 immunity. In vivo consequences of pre-existing immunity to SaCas9 were evaluated in the context of liver genome editing with AAV packaging CRISPR-Cas9 in a mouse model [186]. Efficient genome editing occurred in mouse liver with pre-existing SaCas9 immunity. However, genome editing was accompanied by an increase in CD8⁺ T cells in the liver and a cytotoxic T cell response. The results were hepatocyte apoptosis, loss of recombinant adeno-associated virus genomes, and complete elimination of genome-edited cells. Given the large size of Cas9 and poor homology to human proteins modulating Cas9, immunogenicity assessment appears to be a daunting task. Mixed antigen peptide pools used in conjunction with T cell proliferation assays suggest that the problem, while challenging, may be manageable. It is however important that the assays and analytical tools are validated and fit-for-purpose (i.e., suitable for the question that needs to be answered) as they will be central to clinical decision making.

Advancement in Bioanalytical Techniques to Improve Cell Therapy Product Quantification: Regulatory Considerations

One of the technologies that has potential to improve cell therapy product quantification is flow cytometry. It has the potential to be used in product quality, viability, identity, stability, cell count, purity, potency, and safety. It has led to observations that some populations have brighter or dimmer signals compared to those expressed in normal cells. Flow cytometry has allowed recognition of the correlation between the intensity of the fluorescent signal and the amount of antibody bound per cell (i.e., the number of antigen sites expressed) leading to a full understanding of the translation of “dim” and “bright” into real mass units of fluorescence intensity. There is an increased interest in quantifying the expression and activities of a variety of proteins and enzymes for diagnostic, prognostic, and therapeutic purposes. Quantitative flow cytometry (QFCM) is a method to assign numeric values to the fluorescent signals generated from stained cells. The advancement in bioanalytical techniques within the past decade have resulted in the development of flow cytometric standardization methods and materials that allow quantifying fluorescence intensity with improved levels of control and inter-laboratory precision. Fluorescent signals can be converted to “antibody bound per cell” values using biological reference standards and beads with precise and well-defined numbers of antibody binding sites or quantitative fluorescent beads labeled with a known number of fluorescence molecules. In addition, mean fluorescence intensity (MFI) can be converted to known fluorescent units (e.g., molecules of equivalent soluble fluorochrome) which allow comparison between analysis on different days and/or using different flow cytometers. There are sources of uncertainty, however, including pre-acquisition steps (sample handling, antibody variability, and staining procedure). Acquisition uncertainty can arise from instrument calibration, maintenance, number of cells acquired, controls, standardization procedures, and operators. Finally, gating strategy, data analysis, and method used for quantitative measurements (reference standards versus beads) introduce post acquisition uncertainty. Challenges in QFCM include antigen source (whole blood, whole tissue, cell extract, sorting, etc.), cell culture (long term versus short term), single day versus multiple time points to follow up, cell stimulation or treatment, antibody lot-to-lot variability, and cross experimental standardization. In summary, the field of clinical QFCM based on MFI evaluation requires strict standardization procedures as compared to those needed for the qualitative flow cytometry analysis. Instrument calibration is crucial for quantification of fluorescence measurements by flow cytometry and should be performed routinely. Adherence to quality control measures will minimize subsequent assay uncertainty. A full understanding of the sources of uncertainty, their relative contributions, and areas of improvement will lead to development of robust flow cytometry based quantitative methods.

The Vaccine Assay Validation Challenge: A Reviewers Perspective

The bioanalytical assay has a central role in vaccine development from basic research through post-approval changes. Assay evolution occurs in conjunction with product development following a life cycle of development, qualification, validation, and revalidation. For bioanalytical assay validation, the intended use needs to be considered in the context of a specific study before the assay can be considered validated. The intended use needs to be considered for critical parameters, study design and evaluation instead of indiscriminately following regulatory guidance. Guidance documents are just that, meant as a guide and not dogma, but each assay should be evaluated for its particular use. A common observation is regarding insufficient time spent during development or initially applying incorrect criteria. Often, a retrospective investigation into the cause of a failed evaluation along with additional experience with the assay concludes that the criterion was not appropriately established. Therefore, when establishing validation criteria, it is important to fully understand the limits of the method and use appropriate statistical analyses; the early involvement of a statistician is highly recommended. Critical reagents are also a challenge. Well characterized reagents and controls are needed for determination of critical features, fitness for use, and replacement. Three levels of criteria are needed for validation. System suitability criteria should set the acceptance or rejection of data generated with a given assay. If the assay fails, all data from that assay should be rejected. Sample acceptance criteria are used to ensure samples behave as expected in the assay. Sample failures would result in only the sample being rejected. Validation criteria should be used to decide if an assay is suitable for its intended purpose. A prospective study of assay performance is only a snapshot and should only use data from assays that meet the system suitability criteria based on the intended use of the assay. Common deficiencies for system suitability include too few criteria for the complexity of the assay and ranges that are not statistically justified. Validation should be considered a demonstration of the performance established during qualification. Criteria need to be established prior to validation. Specifications are used to decide whether the test material is acceptable for use and is based on the clinical lot performance and results from routine manufacturing. Specifications should be established with results from a validated method and must be within the working range of the assay. In conclusion, the intended use must be clearly defined. The validation needs to include only those parameters relevant to the purpose and the method. Validation is not a destination; it is one point in the assay life cycle.

Health Canada

Where Clinical & Quality Analysis Meet – Dose Ranging Immunogenicity Data Supports Vaccine Development & Harmonized Specifications

Broadly characterized immunogenicity in early dose ranging trials, linked to stability indicating critical quality attributes (CQA) required for clinical performance (e.g., potency), supports clinical and product development and expedited Covid-19 authorizations. Well-characterized immunogenicity in dose ranging studies also supports correlates of protection (CoP) analyses and expedites future development. There is a regulatory tendency to tighten specifications that lack a robust clinical basis (e.g., only Phase III lot data and manufacturing capability). The absence of clinically-linked/patient-centric specifications may be a disincentive for assay and process improvement, since that can also result in agency requests to tighten specifications. Robust immunogenicity characterization in early phase dose ranging studies and clinical-based/patient-centric specifications offer many advantages and enable harmonized specifications. An example is Shingrix Varicella-zoster virus subunit (VZV gE), a shingles vaccine with AS01B adjuvant [187]. Cell-mediated immunity (CMI) responses were overlapping over the dose range. Observed antibody titer differences were not considered to be clinically significant (robust antibody response at lowest dose). CMI and antibody characterization resulted in a much wider, harmonized potency specification (FDA, EMA and Health Canada) than was supported by Phase III lots. Clinical assays in dose ranging Covid-19 vaccine trials supported rapid development and manufacturing scale up. They characterized the binding antibody and live virus and pseudo-virus NAb. The characterization of CMI included Ag-specific CD4⁺ and CD8⁺ T cells, Th1 and Th2 cytokine profiles, with initial concerns for the potential of vaccine induced enhanced disease, but also established the manufacturing platform immuno-profile. PCR assays were used for SARS-CoV-2 disease endpoint conformation. The use of animal models supported Phase I (i.e., immunogenicity inbred mice) and later trials with challenge studies in ACE2 tg-mice, ferrets, Syrian Golden hamsters and non-human primates. Dose ranging clinical studies with stability indicated CQA for mRNA vaccines permitted authorization of robust specifications that enabled rapid authorization of scale up and marketing. Suboptimal dose ranging regimes in animal models was essential to enabled CoP studies. Dose ranging immunogenicity in Phase I and II with Phase III efficacy data enabled CoP analyses by manufacturers, COVAX, academics [188], Warp Speed, and BARDA. Initial preclinical and clinical studies support NAb as a potential CoP, but not as the only critical immune effector mechanism. Early phase, well characterized immunogenicity dose ranging studies support robust and defendable harmonized product specifications that are less prone to agency pressures to tighten over the product life cycle. A key lesson from Merck's experience with the 2014 Ebola outbreak and the general experience from the COVID-19 pandemic is that, if manufacturers authorize and encourage submission data sharing between key regulators, as well as regulatory coordination of questions and responses, this would further harmonize regulatory decisions. For new COVID-19 vaccine considerations, Health Canada considers all well-designed clinical trials (including placebo controlled disease endpoint designs), but will not require COVID-19 placebo trials in the current context. For new vaccines without disease endpoint data, alternative designs including immunobridging against an authorized active comparator will be considered. Health Canada and MHRA are aligned with the non-inferiority immunogenicity/superiority considerations outlined in the International Coalition of Medicines Regulatory Authorities (ICMRA) summary of their June 24, 2021 meeting [189]. Health Canada, MHRA and the ACCESS Consortium Partners are also aligned regarding cross-platform immunobridging [89]. The use of WHO standards in neutralization studies is strongly recommended. In addition, applicants proposing cross-platform immunobridging for new vaccines are expected to provide relevant animal challenge studies to support proof of concept against variants of concern (VOCs) being strongly recommended; characterization of comparative immunogenicity profiles, including cell-mediated immunity; characterization of comparative in vitro neutralization against VOCs; safety data in at least 3000 adults of all ages with a median follow-up of at least two months post-final dose; and, post-authorization effectiveness studies.

Norway NoMA / EU EMA

Navigating the European Regulatory Landscape for Gene Edited Cell Therapy Products & Vaccines

The EU EMA is the regulatory body responsible for the scientific evaluation and supervision of medicines developed by pharmaceutical companies for use in the EU (human and veterinary). The EMA has several expedited pathways. Accelerated assessment (AA) reduces assessment time for major public health interests with unmet medical needs. A request for accelerated assessment should be made at least 2–3 months before MAA submission, when high level results from pivotal studies are available. Justification should be provided by the applicant to substantiate the claim [190]. Major public health interest is typically to be shown by demonstrating existence of unmet medical needs, how the product could address the unmet medical needs, and the strength of evidence expected at the time of the MAA. The PRIority MEdicines (PRIME) pathway is for better use of existing regulatory and procedural tools and reinforcing the concept of accelerated assessment by early identification of products fulfilling the criteria for AA [191]. Entry to the scheme is at two different stages in development: at the earlier stage of proof of principle (prior to Phase II/exploratory studies) focusing on SMEs or at proof of concept (prior to Phase III/confirmatory studies). It must be based on adequate data to justify a potential major public health interest. There is enhanced support, e.g., through iterative scientific advice and timely appointment of Rapporteurs. Confirmation of eligibility to AA prior to submission can be provided and applicants not eligible to PRIME can still request accelerated assessment. Switching from AA to standard can occur [192]. The main reasons for a switch are major objections not resolvable under accelerated assessment or applicants facing challenges to complete quality, bioanalytical assays, manufacturing development and data requirements during development of medicines for early access. “The EMA is committed to enabling early patient access to new medicines, particularly those that target an unmet medical need or are of major public health interest” [193]. Different procedures are available to establish an early dialogue with regulators and support prospective planning. “EMA provides medicine developers advice on the most appropriate way to generate robust evidence on a medicine's benefits and risks” [194]. There is also a parallel scientific advice or consultative advice with EMA and US. “Pre-submission meetings between applicants and the EMA and (Co-)Rapporteurs occurs approximately 7 months prior to the anticipated date of submission of the application” [195]. There is an opportunity for applicants to obtain procedural, regulatory and legal advice from the EMA and the Rapporteurs, and discuss issues specific to their upcoming application. An EMA product team is available to address any questions regarding the MAA. Any existing prior information should be included in the CTD in the section where the product specific information otherwise would be, including arguments on how the information is relevant. Prior knowledge is a term used in ICH and EMA guidelines and includes “company knowledge from development and manufacturing experience that might be a good basis for shifting the time-point for completion of certain quality studies” [196]. It may make some development studies redundant and can stem from different platforms, e.g., monoclonal antibodies, viral vectors, vaccines or oligonucleotides. Process validation is a life cycle activity. For products in an early access program, consideration should be given towards providing sufficient information to support approval and if it is a departure from traditional requirements. Several tools which can facilitate flexibility are process-validation protocols, concurrent validation, deferral of the submission of certain process validation data, “decoupling active substance and finished product process validation, and continuous process verification [197,198]”. A risk-based approach will allow for a reduced comparability package focusing only on the relevant information. Based on this, a justified set of release, (accelerated) stability and/or characterization data can be used to demonstrate comparability. It is essential that an appropriate pre-specified plan with a justification for the statistical approach is chosen [199]. Comparability acceptance criteria proposed for the relevant quality attribute selected according to a risk-based approach should be provided in the regulatory submission. Inclusion of side-by-side analysis of individual values with accompanying descriptive statistics to summarize data (e.g., min-max and 3*sigma ranges) is recommended [200]. Suitable graphical representations (e.g., individual values scattergrams) could be provided. In cases where very few batches are available (sometimes in combination with large variability a statistical tool may not be useful) comparison with historic ranges may be the best approach [199,200].

Japan MHLW

Evaluation of Anti-SARS-CoV-2 Antibody Tests in the COVID-19 Project of Japan

A variety of anti-SARS-CoV-2 antibodies (against spike protein, against nucleocapsid protein, and other proteins) are induced during infection with SARS-CoV-2. The antibodies in biological matrices are a mixture of these antibodies with different antigens, epitopes and binding affinities. Antibody tests use S protein and/or N protein as the antigen. In many cases, specific domains of the proteins or engineered proteins to improve specificity among coronaviruses are used. In Japan, anti-SARS-CoV-2 antibody tests are not categorized as in vitro diagnostic tests that need regulatory approval even for emergency use. Therefore, there are concerns about their analytical performance and the consistency of their quality. A study was conducted to evaluate the analytical performance of anti-SARS-CoV-2 antibody tests (i.e., cut-off titer to determine antibody positive and its precision) by using original reference standard (NIHS-RS) prepared from COVID-19 patients. The NIHS-RS was prepared from 34 samples of COVID-19 patient sera. The antibody titers of NIHS-RS and the WHO international standard were compared by commercial ELISA kits and confirmed that the NIHS-RS had sufficient titer for the intended use. The collaborative study using NIHS-RS revealed that antibody assay kits distributed in Japan can generally detect antibodies against SARS-CoV-2. The maximum dilution factor of the reference standard that gives a positive result differed for each kit, suggesting that the positive judgment criteria, in addition to limit of detection, differ depending on the kit. Concerns were not identified in precision of the results obtained from all kits. Factors that affected the analytical performance of each kit may include the differences in response in each kit (due to the differences in antigens and detection reagents used), and the difference in cut-off value (due to the clinical specimens used for criteria setting). Exceptionally, some kits for IgM did not obtain a positive result even using the highest concentration of reference standard. Although the analytical sensitivity was considered insufficient, the companies of these IgM kits provided comments that each product has sufficient analytical sensitivity in the in-house evaluation. Standardization of anti-SARS-CoV-2 assay by using an international reference standard and appropriate quality control of each kit are critical for future use.