There is an emergent need for alternative technologies to support the characterization of molecules that are increasingly complex. We present several enabling mass spectrometry-based technologies that will dramatically increase the throughput in biologics R&D. We present the utility of microdroplets to perform real-time digestions, high-resolution ion mobility separations to unravel structural complexities, and multiplexing with isobaric labeling as a new paradigm for quantitation with phenomenal precision and unprecedented speeds.

Bioprocess analytics have been challenged by the rapid growth of biomolecules. The application of assay automation tools and digital infrastructure allows high-throughput testing and rapid availability of analytical data for process stakeholders. Our automation capabilities include the utility of robotic liquid handlers for sample extraction, preparation and/or pre-treatment for labor intensive workflows such as glycan analysis, capillary gel electrophoresis and protein purification. After analysis, data piping from software to electronic notebooks; and real time visualization of results, trends, analytical methods and molecular information. Here, we discuss the overview our assay automation and digital capabilities with some key case studies.

How can High-Resolution Ion Mobility Mass Spectrometry (HRIM-MS) address the challenges of large molecule characterization? This seminar will discuss the benefits of HRIM-MS in enhancing structural characterization of large molecule therapeutics, while increasing throughput and reducing reliance on LC-based methods. Specific focus will be on how HRIM-MS provides more complete information on coeluting structural isomers, as well as and on emerging workflows.

Faster development of analytics is rendered possible by advanced machine learning methods. The acceleration becomes possible by reducing the number of experiments required to develop the analytical methods. The number of experiments can be reduced because, e.g., the correlations in the data renders parts of the observations redundant or historic data can be exploited to develop the current analytical test. We will show results for advanced spectral data analysis.

Raman and NIR spectroscopy are emerging methods for the characterization of biopharmaceutical therapeutics. Despite recent developments in analytical and spectroscopic methods, rapid establishment of accurate models for spectra processing remains a challenge due to statistical and biological uncertainties, time-consuming analysis. In that regard, we employed genetic algorithms to optimize Raman and NIR processing pipelines. This innovative approach demonstrates improvement in time, cost, versatility, and correlations across various categories of analytes.

Automating sample preparation on robotic liquid handlers provides high-throughput capability and consistency during drug research and development. While some methods require minimal adjustments post-implementation, complex workflows often require continuous optimization and advancement. Presented here is the evolution of an automated sample preparation method for peptide mapping. What started as a semi-automated workflow has transformed into a fully automated process with optional front-end features and integrated analytical instrumentation.

The raw materials used in the manufacture of biopharmaceuticals are an important consideration for ensuring product quality. This presentation will describe key elements of our raw material control strategy including a framework for identifying important material attributes and a data analytics initiative for trending material performance. A case study will be presented to highlight how these tools are used to understand the relationship between material attributes and product quality attributes.

Potency assay for Influenza vaccines requires production of WHO reference materials (standard and strain-specific antibodies) each time when virus strains for vaccine composition is changed during the annual WHO strain selection process. Production of reagents is considered as a major factor delaying the vaccine supplies for the seasonal and potential pandemic influenza outbreaks.  It necessitates tremendous efforts to manage critical reagents for the timely supply of vaccines each season.

High-resolution NMR is a key technology that provides critical information about protein structure and dynamics in physiologically relevant conditions.  Recent advances in acquisition and analysis enable us to see intact antibodies at natural abundance.¹  Sensitive to changes in higher order structure, NMR is suited for similarity assessment of biologics and biosimilars.²  With intrinsically high information content NMR has the potential to reduce the number of techniques needed to characterize therapeutic drugs.

Previously in early biopharmaceutical drug discovery processes, only a finite selection of leads could be characterized due to limited material availability. In order to increase our screening capacity, automated sample preparation, assay miniaturization & improved data management strategies have been implemented. This has allowed us to create a data-rich environment enabling selection of high-quality leads. A high-level overview of this high-throughput workflow will be presented.

Biologics discovery projects require sensitive and high-throughput structural assays to select proteins with favorable developability profiles early in the discovery process. With pipelines frequently focusing on multiple distinct therapeutic modalities, these techniques must also be generally applicable. I will describe our strategy for implementing high-throughput analytics in our workflows with focus on sample consumption, throughput, and maximization of analytical diversity.

Therapeutic antibody development requires co-optimization of biophysical properties that exhibit strict tradeoffs. This process remains challenging due to difficulties interpreting deep sequencing data. This presentation discusses advances in computational analysis of deep sequencing data, comparing several sequence-based feature extraction methods for analysis via machine learning models, ultimately achieving accurate predictions of analog binding strengths from binary sequencing labels and enabling the tunable co-optimization of antibody properties from high-throughput experimentation.

The use of the ZipChip device coupled to high resolution Orbitrap mass spectrometry is described for the characterization of charge variants of monoclonal antibodies under native conditions. The ZipChip provides excellent separation performance and facilitates impressive sensitivity based on operation under nanoelectrospray conditions, allowing for deep characterization from nanogram quantities of material. 

Time-domain NMR (TD-NMR or wNMR) is a non-invasive and non-destructive method to analyze therapeutic biologics. Measurements can be done with small bench-top devices. On top of this: results are obtained within seconds. Thus, this technique is a perfect match for in-line analytics during manufacturing, especially continuous processing, or for real-time drug release. See here examples from our therapeutic protein and gen vector pipeline.

The early stage of development for a new biotech product is often hindered by lack of availability of materials and resources. For the most promising candidates, the subsequent race through clinical phases may not provide the opportunity to correct errors that were made in early development, leaving us to live with inadequate shelf life or storage configurations, or rather costly surprises. This presentation is intended to enable the sponsor to utilize the data they have, in conjunction with risk assessments, to mitigate issues before they become major hurdles. The goal is to look beyond the numbers to get an understanding of what is really happening with the product.

During this presentation, I will show how two approaches, the ReFOLD assay and modulated scanning fluorimetry (MSF), can be used to assess protein refoldability to select therapeutic antibody candidates that are resistant to aggregation during storage. Further, I will show how other biophysical methods and molecular dynamics (MD) simulations can support and explain the observations made with the ReFOLD assay and MSF.

The iLine F is an on-line platform that provides continuous monitoring and near real-time measurements of VCC and viability. By employing a label-free approach, this platform utilizes the optical phenotype of expanding T cells and machine learning to classify cell viability and measure VCC. With non-destructive sampling, the iLine F eliminates the requirement for invasive bioreactor sampling steps, thereby reducing the risk of product contamination through closed processing. Finally, the iLine F can also interface with process control systems to enable automated bioreactor operation, whereby enabling operators to inoculate the cell culture systems and walk away until harvest is required.

As the use of model-based applications supporting In-line Monitoring (ILM) and Real-Time Release Testing (RTRT) strategies increase, the need for efficient and robust model lifecycle management practices increases. Data integration and automated workflows provide a means for reducing long-term cost associated with multivariate models and enabling a more sustainable resource model. This talk will review effort at Biogen to design PAT systems and supporting workflows to enable automated modeling approaches.

Recombinant adeno-associated viruses (rAAVs) deliver therapeutic DNA for gene therapy. However, rAAV manufacturing is imperfect, producing a small portion of recombinant viruses with the therapeutic gene. Here, we developed nanofluidic resonators for characterizing rAAV by measuring their mass. We juxtapose our approach with techniques ranging from ddPCR and analytical ultracentrifugation to static light scattering. With our rAAV characterization approach, we aim to enable real-time quality control in continuous rAAV manufacturing.