2024 Hybrid European Bioassay Conference

Title: The Road to In Vitro Potency Assay

Abstract: Coming Soon

Title: Introduction of an Automated Assay Trending Platform for Assay Monitoring

Abstract: Assay trending is key to monitoring the performance of an assay, and a complete data set is valuable in identifying ongoing trends over the assay lifecycle and possible root causes of assay failures, out of specification or out of trend results. It is also required by regulators and is recommended in USP <1220>. Immunocore’s current approach to trending assay data requires transcription from multiple data sources, which can be laborious and potentially introduce error. We aim to introduce an automated process to trend data, reducing errors and time spent on data entry. Our platform would encompass a capability to efficiently monitor data and integrate analysis platforms for exploration of impact of various assay parameters. We consider challenges such as compatibility with software and validation of entries, opportunities to incorporate outsourced assay data, and where this platform may have improved previous troubleshooting efforts.
Over routine use of Immunocore’s platform Activity ELISA methods, repeated failures of the goodness of fit criteria, sum of squares non-linearity, were observed. This was deemed to be due to a recurring hook effect, which was amplified through using different batches of TMB with different rates of colour development between batches. To address this, the goodness of fit criteria was changed to R2, and newer iterations of the assay include a pre-read prior to adding the stop solution to control development. We consider if an automated data trending platform could have identified the root cause of the failures and led to earlier control and identification of a more appropriate goodness of fit metric.

Title: Case Study: Optimization of an Early Phase Enzyme-Based Potency Assay to Detect NN1 Activity

Abstract: Potency assays are crucial for developing biopharmaceutical products as they enable quantification of a drug’s specific activity. These assays closely mirror the in vivo biological mechanism-of-action, thereby establishing a correlation between the drug’s measured activity and clinical efficacy. Cell-based systems are commonly used to develop potency assays, but they may not be ideal for all drug targets. In such cases, a biochemical approach can be employed. Despite the absence of cells, biochemical assays are equally challenging, because slight changes in reaction conditions or concentrations can significantly affect activity and variability.

In this project, the drug target (NN1) was an enzyme secreted into the extra-cellular space, thus justifying the selection of a biochemical assay. The initial method exhibited significant variation and a low signal-to-background ratio of approximately 1.4. Consequently, further optimization was necessary, this case study focuses on the following parameters.

Firstly, the kinetic profile of the enzymatic reaction was analyzed to identify the timeframe where the reaction is linear. Secondly, the concentration of cofactors required in the assay was determined to increase the assay window. Thirdly, plate-homogeneity and sample positioning were optimized to enhance accuracy. Lastly, laboratory automation was employed to minimize assay variability.

After optimization, variation was reduced and the signal to background ration was increased to ca. 7.0, and the overall accuracy of the method was found to be 108%.

Title: Functional Design of Experiment (DoE) for Potency Assay Optimization and In-Silico Simulation

Abstract: For biotherapeutic analytics, robust and reliable potency assays are required. Design of experiment (DoE) approaches can be used to investigate the impact of multiple assay parameters. Currently, specific assay features (e.g., half effective concentration) are modelled independently from each other. A joint interpretation of several assay features is thus difficult. In our innovative functional DoE approach, we use the functional relationship of the assay features to describe the sigmoidal dose-response curve. With the composed functional form, the direct impact of assay parameters on the dose-response curve shape can be modelled. Moreover, a multivariate desirability can be defined and used for assay optimization and assessment of robustness. We believe that functional modelling contributes to understanding the joint influence of assay parameters and helps to design robust biotherapeutic analytics.

Title: Reproducibility is Quintessential – Using R and Quarto for Bioassay Development

Abstract: An analysis of studies of reproducibility within life sciences estimates that the “prevalence of irreproducible preclinical research exceeds 50%”1. Reproducibility is a fundamental principle of the scientific method, how do failures like this exist, especially in our field?

One solution is to blend our data, statistical analysis, and interpretation into a systematic manner. For over 25 years, the analysis and analysis-ready data sets has been practices in the clinical field, thanks to the introduction of the Clinical Data Interchange Standard Consortium (CDISC). The tools R and Quarto can help make this possible for bioassay development.
This presentation will outline the reproducibility and automation benefits of the open-source scientific and technical publishing system Quarto using a bioassay development case-study.

Reference:
1Freedman LP, Cockburn IM, Simcoe TS (2015) The Economics of Reproducibility in Preclinical Researc h. PLoS Biol 13(6): e1002165. https://doi.org/10.1371/journal.pbio.1002165

Title: CombiStats Software – New Web Application

Abstract: CombiStats is a software programme developed by the European Directorate for the Quality of Medicines & HealthCare (EDQM) for the statistical evaluation of biological dilution assays in accordance with Chapter 5.3 of the European Pharmacopoeia.

Initially designed for Official Medicines Control Laboratories (OMCLs), CombiStats is now also available to any interested organization or person. In 2023, 960 licenses were issued, and the software used in 34 countries in Europe and 34 others around the world. CombiStats has thus evolved into an internationally recognized reference tool in its domain, which facilitates the mutual recognition of data and results by all interested parties.

In May 2024, the EDQM will launch the first web-application of CombiStats. The following statistical models will include:

• Linear regression models for quantitative results: parallel line and slope ratio analysis models,
• Sigmoid models for quantitative results: 3-exponenƟal regression model, 4- and 5-parameter logistic regression models,
• 4-parameter logistic regression model for quantal (pass/fail) results, including several data transformations (e.g. probit, logit), effective doses,
• Analysis of single-dose assays using the Wilcoxon-Mann-Whitney test.

Users will also benefit from a secured access to the application (user ID and password), an audit trail and creation of statistical reports in PDF.

The purpose of the poster will be to communicate on the new CombiStats version and to exemplify some of the features above mentioned.

Title: Biological Assays Linearity: Making the Link with Assay Intended Use to Derive Fit-for-Purpose Acceptance Criteria

Abstract: Whenever a quantitative biological assay is developed, it must have demonstrated adequate performances before being used to support clinical trials or quality control programs. Among assay properties, linearity is defined as its ability to provide results that are directly proportional to the amount of analyte in the sample (or to its known relative potency). In absence of reference material, the dilutional linearity can also be used to confirm the relative accuracy of the assay, by comparing measurements obtained on independent dilutions of the same starting material.

To demonstrate the dilutional linearity of an assay, the dose-proportionality approach is used. High positive samples are diluted independently in a negative matrix (e.g., pool of negative samples) at several fractional dilution factors. Assuming lognormality of measurements, a linear regression is used to adjust the test results on the dilution factor (log_10⁡〖(result)=〗 β_0+β_1 log_10⁡〖(dilution〗)). In this view, assay dilutional linearity is evaluated as the closeness of the observed slope to the perfect slope of -1. Setting an acceptable deviation range around this perfect slope is all about finding the right balance between unavoidable assay limitations and making sure the assay is fit-for-purpose.

Data from two different biological assays are used to illustrate dilutional linearity evaluation design and results interpretation. Building-up on acquired knowledge on assay behavior, we propose a simple statistical method to derive validation acceptance criteria considering assay use in group comparison.

Contributing Authors:
Lepers C., Bellanger A., Le Bouter M., Petrof O., Taverne C., Verniquet M.
Vx Assay Statistics, GSK, Rixensart, Belgium

Title: Implementation of suitable SSTs and Outlier Detection Rules: Exemplary Solutions for Different Biossay Formats

Abstract: The setting of suitable assay and sample acceptance criteria is crucial to control the performance of any bioassay system to distinguish between valid and invalid analyses. Although it is not optimal to define acceptance criteria solely based on assay performance of the respective system, the capability of the assay should be considered, to control the failure rate and avoid unnecessary repetitions. Furthermore, the precision and accuracy of an assay can be improved by a good replication strategy and performance of statistical outlier testing including evaluation of block effects.

Due to the complexity of biological assays, many parameters may be relevant, but care should be taken regarding setting too many or too tight acceptance criteria. Exemplarily, the implementation of SSTs will be presented for single bioassay formats (reflecting different modes of action), including the evaluation of suitability of statistical measures and tests.

Title: Total Analytical Error: The Not Any-More Missing Link Between Validation Guidelines Such as ICHQ2(R2), ICHM10, USP 1033, USP 1210, and Many Others

Abstract: Evolution in the design, modeling and statistical requirements involved in the assessment of the fitness of purpose of (bio)-analytical methods and bioassays are booming in our regulatory environment: the current second revision of ICH Q2 (R2), ICH M10, USP <1210>, USP <1033> and corresponding EMA and FDA guidelines. Although all these guidance documents may have different origins, different names, and different area of applications they all focus on the same aim: ensuring that the developed measurement system, irrespective of the technology used, provides reliable reportable results for the next step of the assay life cycle.

A common recent vision to assess this reliability is emerging in these documents, sometimes called: Total Error, Total Analytical Error, Combined Approach for Accuracy and Precision, and so on.
This presentation aims at clarifying, using examples of application:

• What is Total Analytical Error?
• What experimental designs are appropriate to assess the Total Analytical Error for validation studies?
• How to estimate or compute Total Analytical Error for different format of the reportable result?
• How to use Total Analytical Error to decide about the validity of the (bio) analytical method or bioassay?

Finally, the relationship between the validation step and the routine release of results will be highlighted by introducing another concept which is Measurement Uncertainty that answers the question: Where the true value of my reportable result has high probability to be, conditional to the observed reportable result?
Using Total Analytical Error and Measurement Uncertainty will make coherent each step of the life cycle of analytical methods, starting from the definition of the Analytical Target Profile (ATP), through the validation step and up to the everyday use of the (bio) analytical methods or bioassays. In the end; laboratories will have the knowledge of the reliability of the reportable results they will use to make crucial decisions.

Title: From Manual to Automated: Case Studies for Increasing Bioassay Precision and Throughput by Adding Modular Components in Bioassay Workflow

Abstract: For Quality Control (QC) of biopharmaceuticals, release of batches and stability assessments require in vitro cell-based (bio-)potency assays (CBA). These assays inherit an intrinsic variability, that primarily arises from the use of living cells and is further influenced by the number of pipetting steps, the complexity of the assay, and the experience of the analyst.

Despite rapid advances in full assay automation in recent years, implementation efforts remain challenging, particularly for a QC release testing in a GMP environment. Therefore, we focused on semi-automation of potency assays in a modular manner. Since bioassays can be broken down into smaller steps, this approach allowed us to automate only specific parts of the assay.

Here, we present several case studies of implementations of modular workflow automation to improve the reliability, precision, and throughput of potency assays.

Title: The Importance of Measuring Protein Interactions under Physiological Conditions by the Example of von Willebrand Factor

Abstract: Von Willebrand factor (VWF) is a key protein in the hemostatic process by forming a non-covalent complex with coagulation factor VIII (FVIII) to protect it from early degradation and mediating platelet adhesion to collagen at a site of injury to achieve wound closure. Both functions are regulated by the conformation state of VWF which is driven by the blood flow condition. Albeit each monomer of VWF contains a FVIII-binding site, according to weight ratio, only one FVIII molecule is bound per 50 VWF monomers suggesting that VWF is not saturated with FVIII in the circulation. Most available assay systems determining the VWF-FVIII interaction are solid-phase-based, however upon immobilization to a surface, FVIII dissociates from VWF thus reducing the maximum amount of FVIII capable to bind to VWF. Therefore, selecting an appropriate in vitro assay system is critical to determine its FVIII binding capacity as close as possible to the physiological situation. VWF´s function to mediate platelet adhesion is regulated by a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13 (ADAMTS13) by controlling VWF multimer size. Demonstrating the ADAMTS13-mediated cleavage of VWF under physiological shear has proven difficult. Traditional analytical methods utilize denaturing environments, thereby poorly replicate the natural vascular environment. We aimed to develop a sensitive methodology to visualize and quantitate VWF proteolysis by a recombinant ADAMTS13 (rADAMTS13) under arterial shear flow in human blood. Methods: To determine VWF-FVIII under physiological conditions, we developed a new assay method based on surface plasmon resonance (SPR) technology which measures the binding of FVIII to VWF in solution without involving any surface binding step. The interaction with FVIII was determined for various sources of VWF including a plasma derived VWF (pdVWF), a recombinant VWF (rVWF) and a partially processed pro-rVWF. Moreover, the FVIII binding capacity was also tested for VWF present in different human plasma matrices. A mixture of VWF to be analyzed and rFVIII is injected into the flow cell of a SPR sensor chip with immobilized pdVWF. In this indirect assay system, the rFVIII that forms a complex with rVWF in solution is not available for binding to the immobilized pdVWF. VWF-mediated platelet adhesion was measured by a microfluidics-based approach utilizing the BioFlux 1000Z system (Fluxion Biosciences, Oakland, California, USA), where flow channels were coated with 143 μg/mL collagen type I (Chrono-Par Collagen, Chrono-Log Corporation). rADAMTS13 was added to healthy donor blood at different concentrations and the time course of VWF-mediated platelet adhesion to immobilized collagen was determined by microscopy using fluorescent labeled platelets (4µL/mL Calcein-AM (4µM) dissolved in DMSO). Results: Under the applied conditions, rVWF and pdVWF were endowed with a 30-times higher FVIII binding capacity than could be anticipated from the physiological ratio which is near to the theoretical maximum FVIII binding capacity of VWF. Similar results were obtained with VWF from various plasma species. By contrast, the pro-rVWF preparation clearly bound less FVIII. Optimized analytical techniques enabled visualization and quantification of VWF proteolysis based on platelet binding under shear flow. Addition of 1.875-7.5 IU/mL rADAMTS13 to blood reduced the VWF-mediated platelet adhesion to collagen in a concentration dependent manner. Repeated testing validated the sensitivity. Statistical analysis quantified inter-sample variability. Conclusion: We successfully developed two analytical methods to elucidate important functionalities of VWF under close to physiological conditions. First, our newly developed analytical method based on surface plasmon resonance technology allows measuring the interaction of VWF with FVIII on a quantitative basis under conditions similar to the natural milieu. The assay is therefore also expected to more accurately define impaired FVIII binding in mutant VWF variants of patients with von Willebrand disease. Secondly, we successfully established a powerful methodology harnessing microfluidics to gain fundamental insights into rADAMTS13 function under physiologically relevant shear flow conditions. Further enhancement of the techniques, increased biological sampling, and exploration of collagen types could build on these findings. By this example, we show the importance of developing and selecting appropriate analytical methods to characterize the functionality of proteins under near to physiological conditions which provides important information to guide a targeted development of new drug candidates. Conflict of Interest: Author and co-authors are full-time employees and shareholders of Baxalta Innovations GmbH, Vienna, Austria, a member of the Takeda group of companies. Funding statement: This study was funded and sponsored by Takeda Development Center Americas, Inc. Contributing Authors: Gerald Schrenk, Michaela Schaedler, Sylvia Peyrer-Heimstaett, Manfred Billwein Baxalta Innovations GmbH, a member of the Takeda group of companies

Title: Using Product Specification Limits to Define the Right Number of Cell Plates Used per Sample

Abstract: The final specification for the bioassay of a product is, in most cases, unknown when the bioassay is to be developed. As such, we typically have wider specification limits during development than for a marketed product. However, based on experience from other biotech products we may have an idea of what the final specification may end up being. A method should ideally be developed to support the final specification without the need for multiple testing and averaging of results. However, limited time resources available during development may result in a non-ideal method. This makes it even more important to have a strategy for the final development of a method prior to validation and filing. At this point it is important to: 1) Know the specification that you are optimizing towards. If you do not know be conservative and optimize towards meeting a strict specification. 2) Have a strategy for defining the number of plates used per sample (why use 3 plates if your data supports using 2?) and how these plates should be analysed: Independently in separate assays or within the same assay. And 3) Know the position effects of your plate. This talk will provide a walkthrough of the above concepts in theory and in practice using anonymized data from an in house bioassay.

Title: Why Agonistic Antibodies Remain Challenging to Identify & Characterize? Development & Qualification of a Tailored Bioassay Approach for Assessing Agonistic Activity of Immune-checkpoint Antibodies

Abstract: Checkpoint blockade antibodies are established cancer therapeutics. A contrasting class of therapeutics, agonistic checkpoint antibodies, has emerged particularly for attenuating inflammation in autoimmune diseases. Despite showing clinical promise, their development remains challenging due to limitations in conventional checkpoint assay designs that often fail to replicate physiologically relevant conditions in vitro and many times are confined to assessing only inhibitory responses.

Here, we introduce a novel checkpoint assay design in a co-culture format involving Fcγ receptor-expressing cells, directly measuring the stimulatory activity of agonistic antibodies. Through testing various classes of Fcγ receptors and diverse cell backgrounds, we have established a specific and robust assay design for several key checkpoint receptors such as CD40, CD137, PD-1, and BTLA. Notably, we have identified previously unrecognized agonistic activity in a clinically approved blockade antibody, underscoring the importance of exploring similar activity in other therapeutics of this nature. The novel thaw-and-use optimized assay format exhibits high reproducibility with ease of execution and has been qualified for potency and stability studies within a quality-controlled environment.

Title: Apples and Oranges: Case Studies on Similarity, Comparability and Equivalence Regarding Potency Determination

Abstract: The question of “how same or different are A and B” is a central, recurring topic that presents itself in all shapes and forms during analytical development, including for potency-assays.

E.g. differences in method-execution, comparability of cell lines, cellular vs binding assays, old vs new reference standard, … etc.

How to evaluate the “A vs. B” is case-dependent, and so are the consequences when concluding that A and B are same, “similar enough”, or different.

This presentation shows examples from our everyday life in potency assay development. The main focus is on Reference Standards for the same product with different potencies, and the approach we chose as a consequence of non-equivalence.

Title: Navigating Phase-Appropriate Potency Testing for Cell and Gene Therapy Products

Abstract: Phase-appropriate potency testing is critical for the development and advancement of cell and gene therapy (CGT) products through clinical trials. This talk will provide insights into the importance of tailoring potency assays to specific phases of product development, ensuring accurate and meaningful assessment of product efficacy and alignment with the current regulatory guidelines from FDA and EMA.

Covering a range of CGT modalities, key considerations will be presented for designing and implementing phase-appropriate potency assays. By aligning potency testing with the stage of clinical development, sponsors can optimize resources, expedite regulatory approval, and mitigate risks associated with product variability. Furthermore, the talk will address challenges and strategies related to potency assay validation, standardization, and technology transfers. Attendees will gain valuable insights into navigating the complexities of phase-appropriate potency testing, empowering them to drive innovation and ensure the safety and efficacy of CGT products throughout their development lifecycle.

Title: Innovative Bioassays for the Assessment of Therapeutic Anti-SARS-CoV-2 Neutralizing Antibody

Abstract: Introduction: Recombinant anti-SARS-CoV-2 neutralizing antibodies (NAbs) against the receptor-binding domain (RBD) have played a therapeutic role during the COVID-19 pandemic with the primary intent to block viral entry to ACE2-expressing host target cells. However, these very ‘neutralizing’ antibodies may also opsonize the endocytosis of viruses into FcγR-bearing immune cells greatly influencing subsequent enhancement of infection. We reported development of innovative bioassay system to allow comprehensive analysis of factors relevant to efficacy vs safety attributes of these therapeutic recombinant NAbs.

Methods: Different neutralization vs infection cell models were developed based on the strategies that allow step-by-step analysis of the binding of SARS-CoV-2, pseudotyped viruses, S1 subunit, or RBD protein to various target or effector cells, its signaling transduction and inflammatory cytokine production, its regulation by NAbs, and its control mechanism. Various ACE2- and FcγR-bearing primary or engineered cells, pseudo-viruses (wild type and 8 S variants) and innovative cell binding inhibition assays, synergistic antibody-dependent enhancement (ADE) assays were used to evaluate the NAb products.

Results: For the 15+ NAb products, their activities of RBD binding, cell-based ACE2 binding inhibition and pseudovirus neutralization closely paralleled with those seen in live virus plaque reduction neutralization test (PRNT). And the former assays were validated and established as part of a product release control system. Various cell-based FcγR binding, C1q binding, ADCC, and ADE assays were established for product characterization. NAb products demonstrated significantly different ADE activities which were assayed in the context of uptake of NAb opsonized pseudotyped viruses by the FcγR-expressing cells. Interestingly, it was first reported that ADE activities were boosted dramatically with the specific cocktail NAb products.

Conclusions: We describe a comprehensive system of innovative bioassays evaluating quality attributes associated with efficacy vs safety of SARS-CoV-2 NAbs. While the RBD-NAbs and cocktail all exhibit potent neutralizing activities against new SARS-CoV-2 variants, a major safety concern is the synergistic ADE effect, which should be evaluated individually.

Key words: Therapeutic antibody, neutralization, antibody-dependent enhancement

Title: Limiting Potency Bias from Allowed Non-Similarity While Protecting the Similarity Pass Rate

Abstract: Protecting bioassay-based estimates of relative potency against bias (due to allowed non-similarity) and against unacceptably high similarity failure rates, while allowing for changes in assay capability (precision) appears to be impractical. While power calculations for both detecting non-similarity (via difference tests) and for passing similar samples (via equivalence tests) are particularly helpful, these calculations are not simple. This presentation will illustrate some tools for these calculations and how they support modifying similarity criteria based on data from bioassay.

Title: Using a Design of Experiments (DOE) to Optimize the Operating Conditions of a Serum Bactericidal Assay

Abstract: DOE enables the simultaneous assessment of multiple input factors and, importantly, identifies their interactions which is hardly explored via a traditional “one-factor-at-a-time” approach. DOE is therefore an efficient approach which allows scientists to extract more information from their data at lower cost. Several elements should be considered when selecting an experimental design: purpose of experiment, format of assay, sources of variability, practical constraints, available resources… This work describes how statisticians used an optimal block design to set up a doable experiment with reasonable budget that helped scientists select the most appropriate conditions when operating their serum bactericidal assay. Multiple experimental parameters (e.g., incubation time, reagent concentration…) have been included in the DOE and their impact on several critical responses (e.g., killing activity, sample titer…) have been analyzed to support the decision making.

Contributing Authors:
Tran-Ly-Binh A.(1), Bellanger A.(1), Dozot M.(2), Dheur M-S.(2), White J.(3)
(1) Vx Assay Statistics, GSK, Rixensart, Belgium
(2) Vx-Clinical Laboratory & Assay Portfolio, GSK, Rixensart, Belgium,
(3) Clinical Statistics, GSK, Wavre, Belgium

Title: The Critical Step of Linking Mechanism of Action to the Potency Assay of a Cell Therapy

Abstract: Coming Soon

Title: Dual-Targeting Antibody-Based Drugs: Development of Robust, Orthogonal Techniques for Binding Assessment with AQbD Approach

Abstract: Bispecific antibodies (bsAbs) are antibodies with two binding sites directed at two different antigens or two different epitopes on the same antigen. BsAbs display various mechanisms of action, which directly translates into the fact that therapeutic effects of bsAbs are superior to those of monoclonal antibodies. For this reason, recently more attention has been paid to dual-targeting antibody-based drugs. Simultaneously, in recent years, several regulatory updates have been introduced with regard to analytical methods lifecycle. More focus has been put into the early stages of method development: Analytical Quality by Design (AQbD) approach has been proposed.

Here, we present case study, describing development of two orthogonal methods for bsAbs binding evaluation, utilizing ELISA and SPR techniques. The goal of this case study was binding assessment for different dual-specific compounds with anti-CD3 and anti-CD20 activity. Following the newest updates, critical quality attributes (CQAs) of considered bispecific antibody-based drugs were investigated and analytical target profile (ATP) was created. Subsequently, preliminary risk analysis was performed. As a result, potential areas of an assay that require special care were appointed, so as to minimize probability of performance change in the future. Throughout the optimization stage, Design of Experiments (DoE) approach was applied, which investigated the most important interactions between considered factors and reduced the number of analyzes required to be executed. In the frames of development, statistical approaches, such as replication strategy and outlier detection, were implemented to improve accuracy and precision of assays. The final stage was defining a preliminary analytical control strategy (ACS), which ensures desirable methods’ performance and adequate quality of obtained results.

Bio coming soon.