We applied structure-based constrained optimizations to affinity-mature VHH-72 for the SARS-CoV-2 spike protein while retaining original affinity for SARS-CoV-1. ADAPT-designed mutants improved spike binding and virus neutralization for circulating SARS-CoV-2 variants relative to the parental VHH72-Fc. Remarkably, designed mutants restored binding to the Omicron variant, which was lost in the parental antibody due to spike-RBD mutations. Dual-affinity optimization against sarbecoviruses from different clades can broaden specificity within a given clade.

In order to meet patient-centralized needs, the diversity of protein therapeutics has increased considerably. New formats drive the evolution of the clinical and regulatory landscapes. In this presentation we explore early-phase development strategies, covering both fundamentals (platforms) and specialized considerations (novel formats). Topics include: Established/successful strategies; Drivers and typical challenges for new strategies; Desirable expertise/capabilities; Transforming risk perceptions into innovation opportunities.

Developability assessment of biotherapeutics is a powerful risk mitigation tool. We consider assessment of candidate Manufacturability as a part of our Developability process. Manufacturability assessments of potential lead molecules starts with early productions and observations in analytical biophysical assays adding a filter to rate the ideal candidates. Here we give insights into our Manufacturability process and comment on applied methods generating potential lead molecules for clinical development.

We describe the application of a high-throughput B cell culture screening approach which combines automation and the isolation of single, antigen-specific IgG-secreting cells. Efficient mining of the antibody immune repertoire using this platform enables the identification of rare antibodies with desirable characteristics. We show how the technology was used in the discovery of a highly potent antibody to human interleukin (IL)-22, a cytokine with multiple functions in inflammatory and tissue responses.

Machine learning has emerged as a powerful tool for parallelized protein mutation effect prediction. Seismic Therapeutic is shifting how drugs for immunology are discovered and developed by integrating machine learning techniques with structural biology, protein engineering, and translational immunology. We’ll discuss how machine learning can be used to optimize for several critical drug properties simultaneously, thereby reducing the number of iterations needed to craft a molecule suitable for the clinic.

Next-generation multi-specific antibody therapeutics combine functional activities of several mono-specific antibodies and have emerged as successful treatment options for complex, multifactorial illnesses including cancer and inflammatory diseases. Here we report on combining experimental observations with novel in silico approaches to predict correct pairing of light chains with their heavy chain partners to optimize bispecific antibodies in the Y-shaped format.

This presentation covers an overview of GSK Discovery's Biopharm developability capability, which builds throughout the discovery process. This includes the ability to generate large panels of leads in a high-throughput, automated manner, enabling parallel biological and developability screening. The resultant multi-faceted lead selection process also highlights risk areas for downstream CMC and in vivo groups, such that de-risking packages can be moved off critical path, accelerating the drug discovery process. 

We have implemented a pipeline for ML model generation and property prediction for antibodies/VHHs to evaluate sequences and 3D models with regard to their diversity and developability properties, such as liabilities, PTMs, immunogenicity risks, PK properties, and compatibilities with formulations. This pipeline does not only allow to select sequences from high-throughput screening approaches but is also utilized for optimization towards the desired overall developability profile.

1+1 does not equal 2. In this presentation, we will summarize the progress made to assess the developability of our antibody-based future drugs. Biophysical properties are predicted by analyzing sequences, computing parameters on structural models, and performing molecular dynamics. However, while progressing in analyzing single binders, the behavior of bi- or tri-specific formats in different flavors remains challenging. Shape, size, and linkers have their importance.

ADCs are complex engineered therapeutics that combine the precision of the antibody with the activity of the payload, thus lowering the risks of exposure of normal tissues to the potent agents and broadening the potential therapeutic windows compared to traditional chemotherapies. The pace of ADC development is accelerating with the number of investigational agents in human trials growing rapidly. Here we will discuss the developability of ADCs and our approach to speeding up the development of this therapeutic modality which is key to first-in-human and delivering new and better drugs to patients faster.

ON203 is an antibody targeting the oxidized form of macrophage migration inhibitory factor (oxMIF). ON203 is the result of an extensive screening process of mutations in the VH and VL domains of imalumab to improve physicochemical properties such as surface-hydrophobicity and aggregation propensity, while maintaining affinity & specificity. The in vitro efficacy and safety were shown by enhanced ADCC and reduced non-specific cytokine release. This translated into improved tumor suppression in vivo.