Application of machine learning to predict monomer retention of therapeutic proteins after long term storage

Predicting the Long-Term Stability of Biologics with Short-Term Data

Understanding the long-term stability of biologics is crucial to ensure safe, effective, and cost-efficient life-saving therapeutics. Current industry and regulatory practices require arduous real-time data collection over three years; thus, reducing this bottleneck while still ensuring product quality would enhance the speed of medicine to patients. We developed a parallel-pathway kinetic model, combined with Monte Carlo simulations for prediction intervals, to predict the long-term (2+ years) stability of biotherapeutic critical quality attributes (aggregates, fragments, charge variants, purity, and potency) with short-term (3-6 months) data from intended, accelerated, and stressed temperatures. We rigorously validated the model with 18 biotherapeutic drug products, composed of IgG1 and IgG4 monoclonal antibodies, antibody-drug conjugates, dual protein coformulations, and a fusion protein, including high concentration (≥100 mg/mL) formulations, in liquid and lyophilized presentations. For each drug product, we accurately predicted the long-term trends of multiple quality attributes using just 6 months of data. Further, we demonstrated superior stability prediction via our methods compared with industry-standard linear regression methods. The robust and repeatable results of this work across an unprecedented suite of 18 biotherapeutic compounds suggest that kinetic models with Monte Carlo simulation can predict the long-term stability of biologics with short-term data.

DOTAD: A Database of Therapeutic Antibody Developability

The development of therapeutic antibodies is an important aspect of new drug discovery pipelines. The assessment of an antibody's developability-its suitability for large-scale production and therapeutic use-is a particularly important step in this process. Given that experimental assays to assess antibody developability in large scale are expensive and time-consuming, computational methods have been a more efficient alternative. However, the antibody research community faces significant challenges due to the scarcity of readily accessible data on antibody developability, which is essential for training and validating computational models. To address this gap, DOTAD (Database Of Therapeutic Antibody Developability) has been built as the first database dedicated exclusively to the curation of therapeutic antibody developability information. DOTAD aggregates all available therapeutic antibody sequence data along with various developability metrics from the scientific literature, offering researchers a robust platform for data storage, retrieval, exploration, and downloading. In addition to serving as a comprehensive repository, DOTAD enhances its utility by integrating a web-based interface that features state-of-the-art tools for the assessment of antibody developability. This ensures that users not only have access to critical data but also have the convenience of analyzing and interpreting this information. The DOTAD database represents a valuable resource for the scientific community, facilitating the advancement of therapeutic antibody research. It is freely accessible at http://i.uestc.edu.cn/DOTAD/ , providing an open data platform that supports the continuous growth and evolution of computational methods in the field of antibody development.

Designing Robust Monoclonal Antibody Drug Products: Pitfalls of Simplistic Approaches for Stability Prediction

Thermal stability attributes including unfolding onset (Tonset) and mid-point (Tm) are often utilized for efficient development of monoclonal antibody (mAb) products during lead selection and formulation screening workflows. An assumption of direct correlation between thermal and kinetic physical stability underpins this basic approach. While literature reports have substantiated this general approach under specific conditions, clear exceptions have been highlighted alongside. Herein, a set of mAbs formulated under diverse solution conditions to generate a broad array of thermal and kinetic stability profiles were systematically analyzed. Sequence modifications in the Fc region were purposefully engineered to generate a set of low-melting mAbs. A diverse set of excipients were subsequently utilized and shown to modulate the Tm over a wide range. While a general correlation between high Tm and low aggregation rate was observed under accelerated conditions, the predictive utility of Tm under relevant product storage conditions was inadequate at best. Critically, Tm data did not correlate with long-term aggregation rates under refrigerated or room temperature conditions. Even under accelerated conditions, Tm appeared to be a poor predictor of aggregation once it exceeded the solution storage temperature (40°C) by ∼15°C, similar to conditions routinely encountered in the development of canonical mAbs (Tm > 60°C). Pitfalls of simplistic correlative approaches are discussed in the context of practical biologics product development.

Characterization of antibody surface properties and selection of a diverse subset for development of a generic size-exclusion chromatography method

Aggregates are an important quality attribute for biotherapeutics because they can affect the safety and efficacy of the drug and they are therefore routinely monitored, mainly with size-exclusion chromatography (SEC). However, there is often a need to tailor a SEC method to a specific molecule. This study describes the development of a generic SEC method tested on 138 antibodies with the variable domains taken from clinical stage antibodies. We report on the discovery of a subset of 12 antibodies that represents the full range of physiochemical properties found in the 138 antibodies. This subset is shown to be an efficient and reliable test set when developing chromatographic methods for antibodies. An understanding of the nature of the analyte-stationary phase interactions was gained when using this set with its wide range of physiochemical properties. Highly hydrophobic antibodies interact strongly with some modern silica hybrid materials causing the elution time to increase significantly, while a hydrophilically modified hybrid surface showed highly reduced interactions for the hydrophobic antibodies. Highly hydrophilic antibodies, on the other hand, exhibited asymmetric peaks to a certain extent on all stationary phases, while the elution time was not affected. The developed SEC method was shown to have satisfactory performance in terms of linearity, repeatability, range, and accuracy and exhibit very narrow distributions of elution time and peak symmetry when testing the 138 antibodies indicating its generic performance.

Evaluating the chaos game representation of proteins for applications in machine learning models: prediction of antibody affinity and specificity as a case study

CONTEXT

Machine learning techniques are becoming increasingly important in the selection and optimization of therapeutic molecules, as well as for the selection of formulation components and the prediction of long-term stability. Compared to first-principle models, machine learning techniques are easier to implement, and can identify correlations that would be hard to describe at a mechanistic level, but strongly rely on high-quality input training data. Here, we evaluate the potential of the "chaos game" representation to provide input data for machine learning models. The chaos game is an algorithm originally developed for the production of fractal structures, and later on applied also to the representation of biological sequences, such as genes and proteins. Our results show that the combination of the chaos game representation with convolutional neural networks results in comparable accuracy to other machine learning approaches, thus indicating that chaos game representations could be a valid alternative to existing featurization strategies for machine learning models of biological sequences.

METHODS

We implement the chaos game in Python 3.8.10, and use it to produce fractal as well as novel expanding representations of protein sequences. We then feed the resulting images to a convolutional neural network, built in Python 3.8.10, using TensorFlow 2.9.1, Keras 2.9.0, and the scikit-learn 1.1.1 packages. We select as case study a recently published dataset for the antibody emibetuzumab, with the objective of co-optimizing antibodies variants with both high affinity and low non-specific binding.

Rapid Estimation of Size-Based Heterogeneity in Monoclonal Antibodies by Machine Learning-Enhanced Dynamic Light Scattering

Aggregation of monoclonal antibody therapeutics is a serious concern that is believed to impact product safety and efficacy. There is a need for analytical approaches that enable rapid estimation of mAb aggregates. Dynamic light scattering (DLS) is a well-established technique for estimating the average size of protein aggregates or for evaluating sample stability. It is usually used to measure the size and size distribution over a wide range of nano- to micro-sized particles using time-dependent fluctuations in the intensity of scattered light arising from the Brownian motion of particles. In this study, we present a novel DLS-based approach that allows us to quantify the relative percentage of multimers (monomer, dimer, trimer, and tetramer) in a monoclonal antibody (mAb) therapeutic product. The proposed approach uses a machine learning (ML) algorithm and regression to model the system and predict the amount of relevant species such as monomer, dimer, trimer, and tetramer of a mAb in the size range of 10-100 nm. The proposed DLS-ML technique compares favorably to all potential alternatives with respect to the key method attributes, including per sample cost of analysis, per sample time of data acquisition along with ML-based aggregate prediction (<2 min), sample requirements (<3 μg), and user-friendliness of analysis. The proposed rapid method can serve as an orthogonal tool to size exclusion chromatography, which is the current industry workhorse for aggregate assessment.

How can machine learning and multiscale modeling benefit ocular drug development?

The eyes possess sophisticated physiological structures, diverse disease targets, limited drug delivery space, distinctive barriers, and complicated biomechanical processes, requiring a more in-depth understanding of the interactions between drug delivery systems and biological systems for ocular formulation development. However, the tiny size of the eyes makes sampling difficult and invasive studies costly and ethically constrained. Developing ocular formulations following conventional trial-and-error formulation and manufacturing process screening procedures is inefficient. Along with the popularity of computational pharmaceutics, non-invasive in silico modeling & simulation offer new opportunities for the paradigm shift of ocular formulation development. The current work first systematically reviews the theoretical underpinnings, advanced applications, and unique advantages of data-driven machine learning and multiscale simulation approaches represented by molecular simulation, mathematical modeling, and pharmacokinetic (PK)/pharmacodynamic (PD) modeling for ocular drug development. Following this, a new computer-driven framework for rational pharmaceutical formulation design is proposed, inspired by the potential of in silico explorations in understanding drug delivery details and facilitating drug formulation design. Lastly, to promote the paradigm shift, integrated in silico methodologies were highlighted, and discussions on data challenges, model practicality, personalized modeling, regulatory science, interdisciplinary collaboration, and talent training were conducted in detail with a view to achieving more efficient objective-oriented pharmaceutical formulation design.

Machine Learning Analysis Provides Insight into Mechanisms of Protein Particle Formation Inside Containers During Mechanical Agitation

Container choice can influence particle generation within protein formulations. Incompatibility between proteins and containers can manifest as increased particle concentrations, shifts in particle size distributions and changes in particle morphology distributions. In this study, flow imaging microscopy (FIM) combined with machine learning-based goodness-of-fit hypothesis testing algorithms were used in accelerated stability studies to investigate the impact of containers on particle formation. Containers in four major container categories subdivided into eleven container types were filled with monoclonal antibody formulations and agitated with and without headspace, producing subvisible particles. Digital images of the particles were recorded using flow imaging microscopy and analyzed with machine learning algorithms. Particle morphology distributions depended on container category and type, revealing differences that would not have been obvious by analysis of particle concentrations or container surface characteristics alone. Additionally, the algorithm was used to compare morphologies of particles generated in containers against those generated using isolated stresses at air-liquid and container-air-liquid interfaces. These comparisons showed that the morphology distributions of particles formed during agitation most closely resemble distributions that result from exposure of proteins to moving triple interface lines at points where container-air-liquid interfaces intersect. The approach described here can be used to identify dominant causes of particle generation due to protein-container interactions.

Exploring the potential of machine learning for more efficient development and production of biopharmaceuticals

Principles of Industry 4.0 direct us to predict how pharmaceutical operations and regulations may exist with automation, digitization, artificial intelligence (AI), and real time data acquisition. Machine learning (ML), a sub-discipline of AI, involves the use of statistical tools to extract the desired information either through understanding the underlying patterns in the information or by development of mathematical relationships among the critical process parameters (CPPs) and critical quality attributes (CQAs) of biopharmaceuticals. ML is still in its infancy for directly supporting the quality-by-design based development and manufacturing of biopharmaceuticals. However, adoption of ML-based models in place of conventional multi-variate-data-analysis (MVDA) is increasing with the accumulation of large-scale data. This has been majorly contributed by the real-time monitoring of process variables and quality attributes of products through the implementation of process analytical technology in biopharmaceutical manufacturing. All aspects of healthcare, from drug design to product distribution, are complex and multidimensional. Thus, ML-based approaches are being applied to achieve sophistication, accuracy, flexibility and agility in all these areas. This review discusses the potential of ML for addressing the complex issues in diverse areas of biopharmaceutical development, such as biopharmaceuticals design and assessment of early stage development, upstream and downstream process development, analysis, characterization and prediction of post translational modifications (PTMs), formulation and stability studies. Moreover, the challenges in acquisition, cleaning and structuring the bioprocess data, which is one of the major hurdles in implementation of ML in biopharma industry, have also been discussed. Regulatory perspectives on implementation of AI/ML in the biopharma sector have also been briefly discussed. This article is a bird's eye view on the recent developments and applications of ML in overcoming the challenges for adopting "Industry - 4.0" in the biopharma industry.

Machine learning prediction of antibody aggregation and viscosity for high concentration formulation development of protein therapeutics

Machine learning has been recently used to predict therapeutic antibody aggregation rates and viscosity at high concentrations (150 mg/ml). These works focused on commercially available antibodies, which may have been optimized for stability. In this study, we measured accelerated aggregation rates at 45°C and viscosity at 150 mg/ml for 20 preclinical and clinical-stage antibodies. Features obtained from molecular dynamics simulations of the full-length antibody and sequences were used for machine learning model construction. We found a k-nearest neighbors regression model with two features, spatial positive charge map on the CDRH2 and solvent-accessible surface area of hydrophobic residues on the variable fragment, gives the best performance for predicting antibody aggregation rates (r = 0.89). For the viscosity classification model, the model with the highest accuracy is a logistic regression model with two features, spatial negative charge map on the heavy chain variable region and spatial negative charge map on the light chain variable region. The accuracy and the area under precision recall curve of the classification model from validation tests are 0.86 and 0.70, respectively. In addition, we combined data from another 27 commercial mAbs to develop a viscosity predictive model. The best model is a logistic regression model with two features, number of hydrophobic residues on the light chain variable region and net charges on the light chain variable region. The accuracy and the area under precision recall curve of the classification model are 0.85 and 0.6, respectively. The aggregation rates and viscosity models can be used to predict antibody stability to facilitate pharmaceutical development.

State-of-the-Art Review of Artificial Neural Networks to Predict, Characterize and Optimize Pharmaceutical Formulation

During the development of a pharmaceutical formulation, a powerful tool is needed to extract the key points from the complicated process parameters and material attributes. Artificial neural networks (ANNs), a promising and more flexible modeling technique, can address real intricate questions in a high parallelism and distributed pattern in the manner of biological neural networks. The data mined and analyzing based on ANNs have the ability to replace hundreds of trial and error experiments. ANNs have been used for data analysis by pharmaceutics researchers since the 1990s and it has now become a research method in pharmaceutical science. This review focuses on the latest application progress of ANNs in the prediction, characterization and optimization of pharmaceutical formulation to provide a reference for the further interdisciplinary study of pharmaceutics and ANNs.

Long-term stability predictions of therapeutic monoclonal antibodies in solution using Arrhenius-based kinetics

Long-term stability of monoclonal antibodies to be used as biologics is a key aspect in their development. Therefore, its possible early prediction from accelerated stability studies is of major interest, despite currently being regarded as not sufficiently robust. In this work, using a combination of accelerated stability studies (up to 6 months) and first order degradation kinetic model, we are able to predict the long-term stability (up to 3 years) of multiple monoclonal antibody formulations. More specifically, we can robustly predict the long-term stability behaviour of a protein at the intended storage condition (5 °C), based on up to six months of data obtained for multiple quality attributes from different temperatures, usually from intended (5 °C), accelerated (25 °C) and stress conditions (40 °C). We have performed stability studies and evaluated the stability data of several mAbs including IgG1, IgG2, and fusion proteins, and validated our model by overlaying the 95% prediction interval and experimental stability data from up to 36 months. We demonstrated improved robustness, speed and accuracy of kinetic long-term stability prediction as compared to classical linear extrapolation used today, which justifies long-term stability prediction and shelf-life extrapolation for some biologics such as monoclonal antibodies. This work aims to contribute towards further development and refinement of the regulatory landscape that could steer toward allowing extrapolation for biologics during the developmental phase, clinical phase, and also in marketing authorisation applications, as already established today for small molecules.

Towards an improved prediction of concentrated antibody solution viscosity using the Huggins coefficient

The viscosity of a monoclonal antibody solution must be monitored and controlled as it can adversely affect product processing, packaging and administration. Engineering low viscosity mAb formulations is challenging as prohibitive amounts of material are required for concentrated solution analysis, and it is difficult to predict viscosity from parameters obtained through low-volume, high-throughput measurements such as the interaction parameter, k_D, and the second osmotic virial coefficient, B₂₂. As a measure encompassing the effect of intermolecular interactions on dilute solution viscosity, the Huggins coefficient, k_h, is a promising candidate as a parameter measureable at low concentrations, but indicative of concentrated solution viscosity. In this study, a differential viscometry technique is developed to measure the intrinsic viscosity, [η], and the Huggins coefficient, k_h, of protein solutions. To understand the effect of colloidal protein-protein interactions on the viscosity of concentrated protein formulations, the viscometric parameters are compared to k_D and B₂₂ of two mAbs, tuning the contributions of repulsive and attractive forces to the net protein-protein interaction by adjusting solution pH and ionic strength. We find a strong correlation between the concentrated protein solution viscosity and the k_h but this was not observed for the k_D or the b₂₂, which have been previously used as indicators of high concentration viscosity. Trends observed in [η] and k_h values as a function of pH and ionic strength are rationalised in terms of protein-protein interactions.

An accelerated surface-mediated stress assay of antibody instability for developability studies

High physical stability is required for the development of monoclonal antibodies (mAbs) into successful therapeutic products. Developability assays are used to predict physical stability issues such as high viscosity and poor conformational stability, but protein aggregation remains a challenging property to predict. Among different types of stresses, air–water and solid–liquid interfaces are well known to potentially trigger protein instability and induce aggregation. Yet, in contrast to the increasing number of developability assays to evaluate bulk properties, there is still a lack of experimental methods to evaluate antibody stability against interfaces. Here, we investigate the potential of a hydrophobic nanoparticle surface-mediated stress assay to assess the stability of mAbs during the early stages of development. We evaluate this surface-mediated accelerated stability assay on a rationally designed library of 14 variants of a humanized IgG4, featuring a broad span of solubility values and other developability properties. The assay could identify variants characterized by high instability against agitation in the presence of air–water interfaces. Remarkably, for the set of investigated molecules, we observe strong correlations between the extent of aggregation induced by the surface-mediated stress assay and other developability properties of the molecules, such as aggregation upon storage at 45°C, self-association (evaluated by affinity-capture self-interaction nanoparticle spectroscopy) and nonspecific interactions (estimated by cross-interaction chromatography, stand-up monolayer chromatography (SMAC), SMAC*). This highly controlled surface-mediated stress assay has the potential to complement and increase the ability of the current set of screening techniques to assess protein aggregation and developability potential of mAbs during the early stages of drug development.

Abbreviations:AC-SINS: Affinity-Capture Self-Interaction Nanoparticle Spectroscopy; AMS: Ammonium sulfate precipitation; ANS: 1-anilinonaphtalene-8-sulfonate; CIC: Cross-interaction chromatography; DLS: Dynamic light scattering; HIC: Hydrophobic interaction chromatography; HNSSA: Hydrophobic nanoparticles surface-stress assay; mAb: Monoclonal antibody; NP: Nanoparticle; SEC: Size exclusion chromatography; SMAC: Stand-up monolayer chromatography; WT: Wild type

Machine learning directed drug formulation development

Machine learning (ML) has enabled ground-breaking advances in the healthcare and pharmaceutical sectors, from improvements in cancer diagnosis, to the identification of novel drugs and drug targets as well as protein structure prediction. Drug formulation is an essential stage in the discovery and development of new medicines. Through the design of drug formulations, pharmaceutical scientists can engineer important properties of new medicines, such as improved bioavailability and targeted delivery. The traditional approach to drug formulation development relies on iterative trial-and-error, requiring a large number of resource-intensive and time-consuming in vitro and in vivo experiments. This review introduces the basic concepts of ML-directed workflows and discusses how these tools can be used to aid in the development of various types of drug formulations. ML-directed drug formulation development offers unparalleled opportunities to fast-track development efforts, uncover new materials, innovative formulations, and generate new knowledge in drug formulation science. The review also highlights the latest artificial intelligence (AI) technologies, such as generative models, Bayesian deep learning, reinforcement learning, and self-driving laboratories, which have been gaining momentum in drug discovery and chemistry and have potential in drug formulation development.

Predicting drug-microbiome interactions with machine learning

Pivotal work in recent years has cast light on the importance of the human microbiome in maintenance of health and physiological response to drugs. It is now clear that gastrointestinal microbiota have the metabolic power to promote, inactivate, or even toxify the efficacy of a drug to a level of clinically relevant significance. At the same time, it appears that drug intake has the propensity to alter gut microbiome composition, potentially affecting health and response to other drugs. Since the precise composition of an individual's microbiome is unique, one's drug-microbiome relationship is similarly unique. Thus, in the age of evermore personalised medicine, the ability to predict individuals' drug-microbiome interactions is highly sought. Machine learning (ML) offers a powerful toolkit capable of characterising and predicting drug-microbiota interactions at the individual patient level. ML techniques have the potential to learn the mechanisms operating drug-microbiome activities and measure patients' risk of such occurrences. This review will outline current knowledge at the drug-microbiota interface, and present ML as a technique for examining and forecasting personalised drug-microbiome interactions. When harnessed effectively, ML could alter how the pharmaceutical industry and healthcare professionals consider the drug-microbiome axis in patient care.

Artificial Intelligence in Pharmaceutical Field - A Critical Review

Artificial intelligence is an emerging sector in almost all fields. It is not confined only to a particular category and can be used in various fields like research, technology, and health. AI mainly concentrates on how computers analyze data and mimic the human thought process. As drug development involves high R & D costs and uncertainty in time consumption, artificial intelligence can serve as one of the promising solutions to overcome all these demerits. Due to the availability of enormous data, there are chances of missing out on some crucial details. For solving these issues, algorithms like machine learning, deep learning, and other expert systems are being used. On successful implementation of AI in the pharmaceutical field, the delays in drug development, and failure at the clinical and marketing level can be reduced. This review comprises information regarding the development of AI, its subfields, its overall implementation, and its application in the pharmaceutical sector and provides insights on challenges and limitations concerning AI.

Machine learning reveals hidden stability code in protein native fluorescence

Conformational stability of a protein is usually obtained by spectroscopically measuring the unfolding melting temperature. However, optical spectra under native conditions are considered to contain too little resolution to probe protein stability. Here, we have built and trained a neural network model to take the temperature-dependence of intrinsic fluorescence emission under native-only conditions as inputs, and then predict the spectra at the unfolding transition and denatured state. Application to a therapeutic antibody fragment demonstrates that thermal transitions obtained from the predicted spectra correlate highly with those measured experimentally. Crucially, this work reveals that the temperature-dependence of native fluorescence spectra contains a high-degree of previously hidden information relating native ensemble features to stability. This could lead to rapid screening of therapeutic protein variants and formulations based on spectroscopic measurements under non-denaturing temperatures only.