The best bioanalytic tools support bioprocess development by enabling scientists to measure and monitor samples throughout an entire bioprocess quickly and efficiently. Sofia Carvalho is a senior research associate at Portugal’s Instituto de Biologia Experimental e Tecnológica (iBET, https://ibet.pt), a nonprofit biotechnology research organization that works to advance biologicals and biotherapeutics research and development by connecting academia and industry. In an April 2023 webinar, she discussed how advances in bioanalytical tools for viral and viral-like sample characterization can improve the future of vaccine development.
Animal cells often are used to produce viruses for vaccine applications. To help ensure high quality for vaccine products, scientists must leverage analytical instruments and methods that can characterize product quality attributes. iBET’s team uses several tools, including measurement of viral titers, determination of viral structure, and solutions for characterization of omics profiles.
Traditional bioanalytical tools for rotavirus and other vaccines have several drawbacks. Virus-titer measurements based on the cell-culture infectious dose 50% (CCID50) assays are time consuming, labor intensive, and subject to high variability. Other methods such as polymerase chain reaction (PCR) or fluorescence based, are faster and more sensitive but they still are cell-based assays with high variability and low throughput.
Biolayer interferometry (BLI) analysis improves upon the shortcomings of other methods. BLI is a label-free method that uses fiberoptic-based biosensors to measure biomolecular interactions in real time. Each biosensor is coated with an agent that immobilizes antibodies or other receptors of interest. Analyte binding with those biomolecules results in a “biolayer thickness shift” that a BLI instrument registers as a change in optical wavelength. The instrument monitors changes in biolayer density over time to measure binding responses.
BLI analysis can be performed using an Octet platform, which enables users to process up to 96 samples in parallel and perform protein quantitation in under five minutes. The Octet platform allows for real-time analysis because it monitors binding interactions continuously. It can measure analytes even in crude samples without sample labeling.
Characterizing Rotavirus Vaccines
Carvalho presented two case studies developed at iBET, the first of which was performed in collaboration with GSK. The team used BLI technology to quantify live, attenuated rotavirus particles in Rotarix vaccine products from different stages of manufacturing. Scientists used two different strategies, each exploring proteins present in rotavirus. The first strategy applied antibodies for rotavirus antigen (VP7) to the biosensor surface. The second strategy targeted the VP4 rotavirus spike protein, using glycoreceptors.
In a typical workflow, users perform a equilibration step with the biosensor and a buffer to establish baseline measurements. Antibodies or other receptors are then immobilized onto the biosensor tip. A second baseline guarantees that unattached molecules are removed. Then the sensors are dipped into samples containing the analyte (in this case, rotavirus particles).
When selecting the best biosensor and ligand combination, several parameters should be evaluated, including an analyte’s binding magnitude and association profile. Different profiles can be observed depending on the chosen strategy. For instance, when applying a glycoreceptor to the sensor surface, Carvalho’s team observed a nonstandard three-step binding event, suggesting conformational changes on the receptor during association. But when scientists applied an anti-VP7 antibody to the sensor surface, no such events were observed. The initial binding magnitude of the VP7 protein to the antibody proved to be higher than that obtained with the glycoreceptors. Thus, the team developed assays using the anti-VP7 antibody strategy.
To optimize a binding response, scientists must define the optimal antibody loading concentration by screening several dilutions. Carvalho’s team observed biosensor saturation for antibody concentrations >15 μg/mL. Below that threshold, the binding response depended on the antibody concentration, with 2 μg/mL presenting the optimal binding value.
Carvalho emphasized the importance of preventing nonspecific binding. One way to test for that concern is to block the ligands loaded on a biosensor and then to measure virus association. Those values then can be compared with results from viral binding with “naked” (unloaded) sensors. Blocking proved to be an effective strategy for Carvalho’s team. During testing, rotavirus particles showed negligible nonspecific binding to naked biosensors.
To ensure that the method was specific for intact rotavirus particles, the team also performed a binding study comparing values from whole virions and free VP7 proteins, which can be present in rotavirus-vaccine in-process samples. Free proteins showed a low initial binding rate. By contrast, whole rotavirus particles presented a higher initial binding rate and response, indicating that the BLI method would be effective for measuring rotavirus binding behavior.
Carvalho also demonstrated how the iBET team used the Octet system to quantify rotavirus particles within samples from different steps in the vaccine purification process. The team applied BLI and CCID50 to several in-process samples. A Bland-Altman plot of the resulting data enabled measurement comparison. Similar viral-titer values indicated that the BLI method accurately quantified the numbers of whole rotavirus particles in the process samples.
Developing a New Tool For the second case study, Carvalho cited iBET’s participation in EDUFLUVAC, a project funded by the European Commission. The major goal of the project was to develop a paninfluenza vaccine based on virus-like particles (VLPs). The concept of a “universal” flu vaccine relies on the display of multiple hemagglutinin/neuraminidase epitopes that are from different influenza strains and/or subtypes to maximize virus diversity on the surface of the scaffold, generating broadly reactive antibodies.
Fast quantification of hemagglutinin protein is difficult, because analytical methods such as single-radial immunodiffusion (SRID) and hemagglutination assays (HAs) are impractical for in-line process monitoring. For instance, SRID provides low sensitivity and is a time-consuming process. HAs are relatively straightforward, but they require fresh red blood cells (RBCs), which are unstable. Moreover, such assays are less accurate for quantification of VLPs than for quantitation of true virus particles. To develop a fast, functional assay that could work for VLP-based products, the iBET team explored the BLI method as an alternative to HAs.
Hemagglutinin proteins play an important role in virus binding to RBCs. Receptor subtypes differ among hosts, affecting infection and replication of influenza viruses among species. The scientists used BLI to speed up and improve VLP quantification. They created an assay like the one used in the rotavirus case study. But instead of antirotavirus antibodies, they used sialic-acid receptors conjugated to streptavidin-coated biosensors.
Different downstream processing steps yield different purity levels among samples, which affects response behavior. In this case, samples were grouped by downstream processing step. Most sample quantifications ranged within the HA error range. Carvalho noted that early-stage samples contained culture-media compounds that sometimes interfered with concentration analysis. But as downstream processing progressed, sample purity increased and concentration values from the HA and BLI method became more comparable.
Because iBET developed VLPs with multiple influenza antigens, the organization also sought to evaluate specificity among multiple hemagglutinin subtypes and/or groups. Carvalho’s team analyzed several strains from the same subtype against the same calibration curve. The team chose one trivalent, one pentavalent, and three monovalent VLP samples to quantify. Their results showed that group quantification is possible within a single calibration curve.
Lessons Learned The BLI method was developed to provide a fast and robust assay for rotavirus and influenza quantification using stable and homogeneous receptors that are compatible with bioprocess development time constraints. BLI analysis using the Octet platform enabled the team to analyze rotavirus samples relatively rapidly, decreasing wait times from days to minutes. Influenza analysis benefited also, dropping response times from an hour to 15 minutes. The Octet platform also increased throughput, enabling scientists to analyze many more samples with fewer costs and resources. It presents a broad linear dynamic range and is suitable for quantification of in-process samples. The platform enables analysis of the entire bioprocess, from harvest to final product.
Questions and Answers
How should BLI users decide whether to develop an antibody- or receptor-binding method? The best choice depends on the viral particle you are using and on the type of method you want to develop. If you want to measure a specific binding event, it is easier to use an antibody than a receptor. However, sometimes antibodies are unavailable. For example, if you want to develop an analytical method that is suitable for several strains of influenza, using antibodies will be more difficult because you will need several types. For rotavirus, we worked with one serotype using an available antibody.
How would you approach analysis of VLPs on the Octet platform? The best setting to use depends on the VLPs you are using. Some of the multivalent influenza VLPs we studied were unstable, so we used a different setting for them than we did for other viruses. For example, we used room temperature with maximum agitation because our in-process samples tended to settle to the bottom of 96-well plates.
Are the Octet methods stability indicating? The Octet platform can be used to develop stability-indicating methods. For instance, we can use the platform to inactivate the antigen of an influenza virus and then show that the denaturated form does not bind to the immobilized ligands.
Find the full webinar online at www.bioprocessintl.com/category/webinars.