Protein aggregation impacts industrial protein production and formulation. Aggrescan3D (A3D) was developed to aid in understanding and engineering aggregation in globular proteins, becoming one of the most employed structure-based predictors of aggregation due to its accuracy, ease of use, and comprehensive modules to assist in aggregation study and protein redesign. Here we present Aggrescan4D (A4D), which largely extends A3D’s functionality by incorporating pH-dependent aggregation prediction, and an evolutionary-informed automatic mutation protocol to engineer protein solubility.

A key aspect of pharmaceutical development is to guarantee the stability of products during long-term storage and shipment. Advance kinetic modeling (AKM) is relying on short-term accelerated stability studies to generate Arrhenius-based kinetic models that can be used for stability forecasts. The AKM methodology was evaluated on key stability indicating attributes of different types of biotherapeutics and was demonstrated to be a universal and reliable tool for stability predictions.

The interfacial aggregation of biologics poses significant challenges in drug development and delivery, enhanced by the need for high-concentration formulations. We present a microfluidic platform capable of capturing in real time protein particle formation upon application of an oil-water interface. We apply the device to analyse high-concentration protein solutions and design stabilisation strategies against interfacial aggregation beyond traditional surfactants.

Patient convenience and decentralized care get increasing attention in the pharmaceutical industry. Hence, high concentration liquid formulations, favorable solution behavior and protein stability are essential to enable the future of drug development. Predicting the behavior of proteins at high concentrations and the associated risks like high viscosity or aggregation remains challenging - especially with growing complexity in molecular designs required to address medical needs. We will present an early screening and selection process based on high throughput, low mass requiring assays that allows us to assess critical solution parameters and predict developability risks early in drug discovery and over a range of different molecule formats.

Therapeutic protein formulations (TPFs) are promising for treating complex diseases, but face stability challenges due to external physical factors (e.g., heat or mechanical stress). Current batchwise stability studies are inadequate, lacking control and failing to mimic real-life destabilizing conditions. An automated microfluidic platform has been developed to study TPF stability by applying various physical stresses, revealing colloidal and structural changes not detected by conventional methods. This innovative approach generates advanced stability data, enhancing understanding of formulation instability and guiding improved mitigation strategies to enhance clinical outcomes and patient satisfaction.

The application of surfactants, mainly polysorbates, is a common practice to prevent surface- or agitation-induced protein aggregation in liquid formulation. However, polysorbates, despite their common application, bring along disadvantages including chemical and enzymatic instability. This presentation will provide an overview of the protein stabilising capability of surfactants against agitation- and interface-induced stresses and corresponding assays for its evaluation. Furthermore, a focus is set to alternative surfactants suitable to replace polysorbates.

The presence of endotoxin must be reliably detected in pharmaceutical formulations to ensure patient safety and reduce the risk of drug-induced septic shock. However, recent phenomena have demonstrated that particular formulation excipients, over time, can mask endotoxin from gold-standard detection assays (i.e., LAL assays). The current work presents recent data that aims to identify the underlying mechanism of masking.

• AI/ML applications for predicting formulation stability and developability of biological drug candidates.
• AI-Driven formulation design for enhanced developability profiles.
• Integrating AI/ML into early formulation development for improved developability assessment.

Advances in molecular format complexity and the need for higher protein concentrations in biotherapeutics present significant formulation challenges. Our in silico pipeline streamlines the development of stable liquid formulations, saving time and cost. By employing physics-based simulations, we predict protein behaviour in diverse conditions, facilitating the pre-selection of optimal excipients and conditions for specific active pharmaceutical ingredients, thereby enhancing the success of formulation development.

Computational approaches have gained popularity over the last decade for the screening of biologics molecules. We will present how in silico models trained on legacy data using sequence, structure, and/or deep learning derived features perform at predicting various physicochemical properties of candidates and how these help in improving the overall quality of the biotherapeutic pipeline.