In silico selection of therapeutic antibodies for development: viscosity, clearance, and chemical stability

A framework for the biophysical screening of antibody mutations targeting solvent-accessible hydrophobic and electrostatic patches for enhanced viscosity profiles

The formulation of high-concentration monoclonal antibody (mAb) solutions in low dose volumes for autoinjector devices poses challenges in manufacturability and patient administration due to elevated solution viscosity. Often many therapeutically potent mAbs are discovered, but their commercial development is stalled by unfavourable developability challenges. In this work, we present a systematic experimental framework for the computational screening of molecular descriptors to guide the design of 24 mutants with modified viscosity profiles accompanied by experimental evaluation. Our experimental observations using a model anti-IL8 mAb and eight engineered mutant variants reveal that viscosity reduction is influenced by the location of hydrophobic interactions, while targeting positively charged patches significantly increases viscosity in comparison to wild-type anti-IL-8 mAb. We conclude that most predicted in silico physicochemical properties exhibit poor correlation with measured experimental parameters for antibodies with suboptimal developability characteristics, emphasizing the need for comprehensive case-by-case evaluation of mAbs. This framework combining molecular design and triage via computational predictions with experimental evaluation aids the agile and rational design of mAbs with tailored solution viscosities, ensuring improved manufacturability and patient convenience in self-administration scenarios.

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.

Cellular Neonatal Fc Receptor Recycling Efficiencies can Differentiate Target-Independent Clearance Mechanisms of Monoclonal Antibodies

In vivo clearance mechanisms of therapeutic monoclonal antibodies (mAbs) encompass both target-mediated and target-independent processes. Two distinct determinants of overall mAb clearance largely separate of target-mediated influences are non-specific cellular endocytosis and subsequent pH-dependent mAb recycling mediated by the neonatal Fc receptor (FcRn), where inter-mAb variability in the efficiency of both processes is observed. Here, we implemented a functional cell-based FcRn recycling assay via Madin-Darby canine kidney type II cells stably co-transfected with human FcRn and its light chain β2-microglobulin. Next, a series of pH-dependent internalization studies using a model antibody demonstrated proper function of the human FcRn complex. We then applied our cellular assays to assess the contribution of both FcRn and non-specific interactions in the cellular turnover for a panel of 8 clinically relevant mAbs exhibiting variable human pharmacokinetic behavior. Our results demonstrate that the interplay of non-specific endocytosis rates, pH-dependent non-specific interactions, and engagement with FcRn all contribute to the overall recycling efficiency of therapeutic monoclonal antibodies. The predictive capacity of our assay approach was highlighted by successful identification of all mAbs within our panel possessing clearance in humans greater than 5 mL/day/kg. These results demonstrate that a combination of cell-based in vitro assays can properly resolve individual mechanisms underlying the overall in vivo recycling efficiency and non-target mediated clearance of therapeutic mAbs.

DSC Derived (Ea & ΔG) Energetics and Aggregation Predictions for mAbs

The Arrhenius energy of activation of unfolding Ea unfolding and Gibbs free energy of unfolding ΔG unfolding have been calculated utilizing DSC differential scanning calorimetry for 4 mAbs (1 biosimilar) in 3 formulations. DSC derived ΔTm melting temperature changes for each mAb domain (CH2, Fab, CH3) at calorimetric scan rates at 60 °C, 90 °C, 150 °C and 200 °C / hr. were utilized to calculate the kinetic Eaunfolding. The DSC derived Ea trend with observed aggregate formation and can be used to predict%HMW formation post 9-month storage at 5 °C and 40 °C for all formulations analyzed. Additionally, thermodynamic ΔG unfolding energies were also derived (Tm, ΔCp and ΔH measurements) for each mAb at every scan rate to observe scan rate dependence of ΔG and for extrapolation to 0 °C/hr. (to report ΔG at true equilibrium conditions). Both derived thermodynamic ΔG and kinetic Ea energies were combined to build full energetic landscapes for mAb unfolding and aggregation. Statistical multivariate analysis of kinetic (Ea CH2, Ea Fab, Ea CH3) energies, thermodynamic (ΔG5 °C and ΔG40 °C) energies and in-silico modeled surface properties was also performed. Analysis revealed key significant parameters contributing to aggregation. These parameters were utilized to build predictive aggregation models for 25 mg/mL mAb formulations stored 9-months at 5 °C and 40 °C.

Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies

Background

Early assessment of antibody off-target binding is essential for mitigating developability risks such as fast clearance, reduced efficacy, toxicity, and immunogenicity. The baculovirus particle (BVP) binding assay has been widely utilized to evaluate polyreactivity of antibodies. As a complementary approach, computational prediction of polyreactivity is desirable for counter-screening antibodies from in silico discovery campaigns. However, there is a lack of such models.

Methods

Herein, we present the development of an ensemble of three deep learning models based on two pan-protein foundational protein language models (ESM2 and ProtT5) and an antibody-specific protein language model (PLM) (Antiberty). These models were trained in a transfer learning network to predict the outcomes in the BVP assay and the bovine serum albumin binding assay, which was developed as a complement to the BVP assay. The training was conducted on a large dataset of antibody sequences augmented with experimental conditions, which were collected through a highly efficient application system.

Results

The resulting models demonstrated robust performance on canonical mAbs (monospecific with heavy and light chain), bispecific Abs, and single-domain Fc (VHH-Fc). PLMs outperformed a model built using molecular descriptors calculated from AlphaFold 2 predicted structures. Embeddings from the antibody-specific and foundational PLMs resulted in similar performance.

Conclusion

To our knowledge, this represents the first application of PLMs to predict assay data on bispecifics and VHH-Fcs.

Stability of Protein Pharmaceuticals: Recent Advances

There have been significant advances in the formulation and stabilization of proteins in the liquid state over the past years since our previous review. Our mechanistic understanding of protein-excipient interactions has increased, allowing one to develop formulations in a more rational fashion. The field has moved towards more complex and challenging formulations, such as high concentration formulations to allow for subcutaneous administration and co-formulation. While much of the published work has focused on mAbs, the principles appear to apply to any therapeutic protein, although mAbs clearly have some distinctive features. In this review, we first discuss chemical degradation reactions. This is followed by a section on physical instability issues. Then, more specific topics are addressed: instability induced by interactions with interfaces, predictive methods for physical stability and interplay between chemical and physical instability. The final parts are devoted to discussions how all the above impacts (co-)formulation strategies, in particular for high protein concentration solutions.'

Sequence-Based Viscosity Prediction for Rapid Antibody Engineering

Through machine learning, identifying correlations between amino acid sequences of antibodies and their observed characteristics, we developed an internal viscosity prediction model to empower the rapid engineering of therapeutic antibody candidates. For a highly viscous anti-IL-13 monoclonal antibody, we used a structure-based rational design strategy to generate a list of variants that were hypothesized to mitigate viscosity. Our viscosity prediction tool was then used as a screen to cull virtually engineered variants with a probability of high viscosity while advancing those with a probability of low viscosity to production and testing. By combining the rational design engineering strategy with the in silico viscosity prediction screening step, we were able to efficiently improve the highly viscous anti-IL-13 candidate, successfully decreasing the viscosity at 150 mg/mL from 34 cP to 13 cP in a panel of 16 variants.

A systemic approach to identifying sequence frameworks that decrease mAb production in a transient Chinese hamster ovary cell expression system

Monoclonal antibodies (mAbs) are often engineered at the sequence level for improved clinical performance yet are rarely evaluated prior to candidate selection for their "developability" characteristics, namely expression, which can necessitate additional resource investments to improve the manufacturing processes for problematic mAbs. A strong relationship between primary sequence and expression has emerged, with slight differences in amino acid sequence resulting in titers differing by up to an order of magnitude. Previous work on these "difficult-to-express" (DTE) mAbs has shown that these phenotypes are driven by post-translational bottlenecks in antibody folding, assembly, and secretion processes. However, it has been difficult to translate these findings across cell lines and products. This work presents a systematic approach to study the impact of sequence variation on mAb expression at a larger scale and under more industrially relevant conditions. The analysis found 91 mutations that decreased transient expression of an IgG1κ in Chinese hamster ovary (CHO) cells and revealed that mutations at inaccessible residues, especially those leading to decreases in residue hydrophobicity, are not favorable for high expression. This workflow can be used to better understand sequence determinants of mAb expression to improve candidate selection procedures and reduce process development timelines.

Effects of nitrogen fertilizer on the physicochemical, structural, functional, thermal, and rheological properties of mung bean (Vigna radiata) protein

Nitrogen fertilizer can affect the seed quality of mung bean. However, the effects of nitrogen fertilizer on the properties of mung bean protein (MBP) remain unclear. We investigated the effects of four nitrogen fertilization levels on the physicochemical, structural, functional, thermal, and rheological properties of MBP. The results showed that the amino acid and protein contents of mung bean flour were maximized under 90 kg ha-1 of applied nitrogen treatment. Nitrogen fertilization can alter the secondary and tertiary structure of MBP. The main manifestations are an increase in the proportion of β-sheet, the exposure of more chromophores and hydrophobic groups, and the formation of loose porous aggregates. These changes improved the solubility, oil absorption capacity, emulsion activity, and foaming stability of MBP. Meanwhile, Thermodynamic and rheological analyses showed that the thermal stability, apparent viscosity, and gel elasticity of MBP were all increased under nitrogen fertilizer treatment. Correlation analysis showed that protein properties are closely related to changes in structure. In conclusion, nitrogen fertilization can improve the protein properties of MBP by modulating the structure of protein molecules. This study provides a theoretical basis for the optimization of mung bean cultivation and the further development of high-quality mung bean protein foods.

ProtT5 and random forests-based viscosity prediction method for therapeutic mAbs

Viscosity is a key characteristic of therapeutic antibodies for subcutaneous administration which requires low volume and high concentration formulations. It would be highly beneficial to accurately predict the viscosity of newly developed therapeutic antibodies in the early stages of development. In this work, a ProtT5-XL-UniRef50 (ProtT5) and Random Forests (RF)-based prediction method was proposed for accurately predicting the viscosity of monoclonal antibodies, with only corresponding sequences needed. Starting from the given heavy and light chain V-region sequences, corresponding features were first extracted from the ProtT5 pretrained model. Kernel principal analysis (Kernel-PCA) was then used for reducing the extracted 2048-D (1024-D for each sequence) feature vector to a reasonable level for efficient training of the RF-regressor. Then, the RF model was constructed on 40 commercially available therapeutic antibodies and tested with 3-folds cross-validation. Test results show that the model could reproduce the viscosity value at a high level (Pearson correlation coefficient (PCC) = 0.928). Performance on classifying high (>30 cP) and low (<30 cP) viscosity is much more satisfactory, the Accuracy (ACC) and the area under precision-recall curve (AUC) of the classification model from validation tests are 0.975 and 1.000, respectively. Compared to 5 existing state-of-the-art viscosity prediction methods, the proposed method performs best which would facilitate high concentration antibody viscosity high-throughput screening.

A minimal physiologically based pharmacokinetic model to study the combined effect of antibody size, charge, and binding affinity to FcRn/antigen on antibody pharmacokinetics

Protein therapeutics have revolutionized the treatment of a wide range of diseases. While they have distinct physicochemical characteristics that influence their absorption, distribution, metabolism, and excretion (ADME) properties, the relationship between the physicochemical properties and PK is still largely unknown. In this work we present a minimal physiologically-based pharmacokinetic (mPBPK) model that incorporates a multivariate quantitative relation between a therapeutic's physicochemical parameters and its corresponding ADME properties. The model's compound-specific input includes molecular weight, molecular size (Stoke's radius), molecular charge, binding affinity to FcRn, and specific antigen affinity. Through derived and fitted empirical relationships, the model demonstrates the effect of these compound-specific properties on antibody disposition in both plasma and peripheral tissues using observed PK data in mice and humans. The mPBPK model applies the two-pore hypothesis to predict size-based clearance and exposure of full-length antibodies (150 kDa) and antibody fragments (50-100 kDa) within a onefold error. We quantitatively relate antibody charge and PK parameters like uptake rate, non-specific binding affinity, and volume of distribution to capture the relatively faster clearance of positively charged mAb as compared to negatively charged mAb. The model predicts the terminal plasma clearance of slightly positively and negatively charged antibody in humans within a onefold error. The mPBPK model presented in this work can be used to predict the target-mediated disposition of a drug when compound-specific and target-specific properties are known. To our knowledge, a combined effect of antibody weight, size, charge, FcRn, and antigen has not been incorporated and studied in a single mPBPK model previously. By conclusively incorporating and relating a multitude of protein's physicochemical properties to observed PK, our mPBPK model aims to contribute as a platform approach in the early stages of drug development where many of these properties can be optimized to improve a molecule's PK and ultimately its efficacy.

Antibody association in solution: cluster distributions and mechanisms

Understanding factors that affect the clustering and association of antibodies molecules in solution is critical to their development as therapeutics. For 19 different monoclonal antibody (mAb) solutions, we measured the viscosities, the second virial coefficients, the Kirkwood-Buff integrals, and the cluster distributions of the antibody molecules as functions of protein concentration. Solutions were modeled using the statistical-physics Wertheim liquid-solution theory, representing antibodies as Y-shaped molecular structures of seven beads each. We found that high-viscosity solutions result from more antibody molecules per cluster. Multi-body properties such as viscosity are well predicted experimentally by the 2-body Kirkwood-Buff quantity, G22, but not by the second virial coefficient, B22, and well-predicted theoretically from the Wertheim protein-protein sticking energy. Weakly interacting antibodies are rate-limited by nucleation; strongly interacting ones by propagation. This approach gives a way to relate micro to macro properties of solutions of associating proteins.

Molecular surface descriptors to predict antibody developability: sensitivity to parameters, structure models, and conformational sampling

In silico assessment of antibody developability during early lead candidate selection and optimization is of paramount importance, offering a rapid and material-free screening approach. However, the predictive power and reproducibility of such methods depend heavily on the selection of molecular descriptors, model parameters, accuracy of predicted structure models, and conformational sampling techniques. Here, we present a set of molecular surface descriptors specifically designed for predicting antibody developability. We assess the performance of these descriptors by benchmarking their correlations with an extensive array of experimentally determined biophysical properties, including viscosity, aggregation, hydrophobic interaction chromatography, human pharmacokinetic clearance, heparin retention time, and polyspecificity. Further, we investigate the sensitivity of these surface descriptors to methodological nuances, such as the choice of interior dielectric constant, hydrophobicity scales, structure prediction methods, and the impact of conformational sampling. Notably, we observe systematic shifts in the distribution of surface descriptors depending on the structure prediction method used, driving weak correlations of surface descriptors across structure models. Averaging the descriptor values over conformational distributions from molecular dynamics mitigates the systematic shifts and improves the consistency across different structure prediction methods, albeit with inconsistent improvements in correlations with biophysical data. Based on our benchmarking analysis, we propose six in silico developability risk flags and assess their effectiveness in predicting potential developability issues for a set of case study molecules.

Exploring molecular determinants and pharmacokinetic properties of IgG1-scFv bispecific antibodies

Bispecific antibodies (BsAbs) capable of recognizing two distinct epitopes or antigens offer promising therapeutic options for various diseases by targeting multiple pathways. The favorable pharmacokinetic (PK) properties of monoclonal antibodies (mAbs) are crucial, as they directly influence patient safety and therapeutic efficacy. For numerous mAb therapeutics, optimization of neonatal Fc receptor (FcRn) interactions and elimination of unfavorable molecular properties have led to improved PK properties. However, many BsAbs exhibit unfavorable PK, which has precluded their development as drugs. In this report, we present studies on the molecular determinants underlying the distinct PK profiles of three IgG1-scFv BsAbs. Our study indicated that high levels of nonspecific interactions, elevated isoelectric point (pI), and increased number of positively charged patches contributed to the fast clearance of IgG1-scFv. FcRn chromatography results revealed specific scFv-FcRn interactions that are unique to the IgG1-scFv, which was further supported by molecular dynamics (MD) simulation. These interactions likely stabilize the BsAb FcRn interaction at physiological pH, which in turn could disrupt FcRn-mediated BsAb recycling. In addition to the empirical observations, we also evaluated the impact of in silico properties, including pI differential between the Fab and scFv and the ratio of dipole moment to hydrophobic moment (RM) and their correlation with the observed clearance. These findings highlight that the PK properties of BsAbs may be governed by novel determinants, owing to their increased structural complexity compared to immunoglobulin G (IgG) 1 antibodies.

Predicting deamidation and isomerization sites in therapeutic antibodies using structure-based in silico approaches

Asparagine (Asn) deamidation and aspartic acid (Asp) isomerization are common degradation pathways that affect the stability of therapeutic antibodies. These modifications can pose a significant challenge in the development of biopharmaceuticals. As such, the early engineering and selection of chemically stable monoclonal antibodies (mAbs) can substantially mitigate the risk of subsequent failure. In this study, we introduce a novel in silico approach for predicting deamidation and isomerization sites in therapeutic antibodies by analyzing the structural environment surrounding asparagine and aspartate residues. The resulting quantitative structure-activity relationship (QSAR) model was trained using previously published forced degradation data from 57 clinical-stage mAbs. The predictive accuracy of the model was evaluated for four different states of the protein structure: (1) static homology models, (2) enhancing low-frequency vibrational modes during short molecular dynamics (MD) runs, (3) a combination of (2) with a protonation state reassignment, and (4) conventional full-atomistic MD simulations. The most effective QSAR model considered the accessible surface area (ASA) of the residue, the pKa value of the backbone amide, and the root mean square deviations of both the alpha carbon and the side chain. The accuracy was further enhanced by incorporating the QSAR model into a decision tree, which also includes empirical information about the sequential successor and the position in the protein. The resulting model has been implemented as a plugin named "Forecasting Reactivity of Isomerization and Deamidation in Antibodies" in MOE software, completed with a user-friendly graphical interface to facilitate its use.

Modulation of the high concentration viscosity of IgG1 antibodies using clinically validated Fc mutations

The self-association of therapeutic antibodies can result in elevated viscosity and create problems in manufacturing and formulation, as well as limit delivery by subcutaneous injection. The high concentration viscosity of some antibodies has been reduced by variable domain mutations or by the addition of formulation excipients. In contrast, the impact of Fc mutations on antibody viscosity has been minimally explored. Here, we studied the effect of a panel of common and clinically validated Fc mutations on the viscosity of two closely related humanized IgG1, κ antibodies, omalizumab (anti-IgE) and trastuzumab (anti-HER2). Data presented here suggest that both Fab-Fab and Fab-Fc interactions contribute to the high viscosity of omalizumab, in a four-contact model of self-association. Most strikingly, the high viscosity of omalizumab (176 cP) was reduced 10.7- and 2.2-fold by Fc modifications for half-life extension (M252Y:S254T:T256E) and aglycosylation (N297G), respectively. Related single mutations (S254T and T256E) each reduced the viscosity of omalizumab by ~6-fold. An alternative half-life extension Fc mutant (M428L:N434S) had the opposite effect in increasing the viscosity of omalizumab by 1.5-fold. The low viscosity of trastuzumab (8.6 cP) was unchanged or increased by ≤ 2-fold by the different Fc variants. Molecular dynamics simulations provided mechanistic insight into the impact of Fc mutations in modulating electrostatic and hydrophobic surface properties as well as conformational stability of the Fc. This study demonstrates that high viscosity of some IgG1 antibodies can be mitigated by Fc mutations, and thereby offers an additional tool to help design future antibody therapeutics potentially suitable for subcutaneous delivery.

Developability considerations for bispecific and multispecific antibodies

Bispecific antibodies (bsAb) and multispecific antibodies (msAb) encompass a diverse variety of formats that can concurrently bind multiple epitopes, unlocking mechanisms to address previously difficult-to-treat or incurable diseases. Early assessment of candidate developability enables demotion of antibodies with low potential and promotion of the most promising candidates for further development. Protein-based therapies have a stringent set of developability requirements in order to be competitive (e.g. high-concentration formulation, and long half-life) and their assessment requires a robust toolkit of methods, few of which are validated for interrogating bsAbs/msAbs. Important considerations when assessing the developability of bsAbs/msAbs include their molecular format, likelihood for immunogenicity, specificity, stability, and potential for high-volume production. Here, we summarize the critical aspects of developability assessment, and provide guidance on how to develop a comprehensive plan tailored to a given bsAb/msAb.

Reduction of monoclonal antibody viscosity using interpretable machine learning

Early identification of antibody candidates with drug-like properties is essential for simplifying the development of safe and effective antibody therapeutics. For subcutaneous administration, it is important to identify candidates with low self-association to enable their formulation at high concentration while maintaining low viscosity, opalescence, and aggregation. Here, we report an interpretable machine learning model for predicting antibody (IgG1) variants with low viscosity using only the sequences of their variable (Fv) regions. Our model was trained on antibody viscosity data (>100 mg/mL mAb concentration) obtained at a common formulation pH (pH 5.2), and it identifies three key Fv features of antibodies linked to viscosity, namely their isoelectric points, hydrophobic patch sizes, and numbers of negatively charged patches. Of the three features, most predicted antibodies at risk for high viscosity, including antibodies with diverse antibody germlines in our study (79 mAbs) as well as clinical-stage IgG1s (94 mAbs), are those with low Fv isoelectric points (Fv pIs < 6.3). Our model identifies viscous antibodies with relatively high accuracy not only in our training and test sets, but also for previously reported data. Importantly, we show that the interpretable nature of the model enables the design of mutations that significantly reduce antibody viscosity, which we confirmed experimentally. We expect that this approach can be readily integrated into the drug development process to reduce the need for experimental viscosity screening and improve the identification of antibody candidates with drug-like properties.

Assessment and incorporation of in vitro correlates to pharmacokinetic outcomes in antibody developability workflows

In vitro assessments for the prediction of pharmacokinetic (PK) behavior of biotherapeutics can help identify corresponding liabilities significantly earlier in the discovery timeline. This can minimize the need for extensive early in vivo PK characterization, thereby reducing animal usage and optimizing resources. In this study, we recommend bolstering classical developability workflows with in vitro measures correlated with PK. In agreement with current literature, in vitro measures assessing nonspecific interactions, self-interaction, and FcRn interaction are demonstrated to have the highest correlations to clearance in hFcRn Tg32 mice. Crucially, the dataset used in this study has broad sequence diversity and a range of physicochemical properties, adding robustness to our recommendations. Finally, we demonstrate a computational approach that combines multiple in vitro measurements with a multivariate regression model to improve the correlation to PK compared to any individual assessment. Our work demonstrates that a judicious choice of high throughput in vitro measurements and computational predictions enables the prioritization of candidate molecules with desired PK properties.

Optimization of therapeutic antibodies for reduced self-association and non-specific binding via interpretable machine learning

Antibody development, delivery, and efficacy are influenced by antibody-antigen affinity interactions, off-target interactions that reduce antibody bioavailability and pharmacokinetics, and repulsive self-interactions that increase the stability of concentrated antibody formulations and reduce their corresponding viscosity. Yet identifying antibody variants with optimal combinations of these three types of interactions is challenging. Here we show that interpretable machine-learning classifiers, leveraging antibody structural features descriptive of their variable regions and trained on experimental data for a panel of 80 clinical-stage monoclonal antibodies, can identify antibodies with optimal combinations of low off-target binding in a common physiological-solution condition and low self-association in a common antibody-formulation condition. For three clinical-stage antibodies with suboptimal combinations of off-target binding and self-association, the classifiers predicted variable-region mutations that optimized non-affinity interactions while maintaining high-affinity antibody-antigen interactions. Interpretable machine-learning models may facilitate the optimization of antibody candidates for therapeutic applications.

The origins of nonideality exhibited by monoclonal antibodies and Fab fragments in human serum

The development of therapeutic antibodies remains challenging, time-consuming, and expensive. A key contributing factor is a lack of understanding of how proteins are affected by complex biological environments such as serum and plasma. Nonideality due to attractive or repulsive interactions with cosolutes can alter the stability, aggregation propensity, and binding interactions of proteins in solution. Fluorescence correlation spectroscopy (FCS) can be used to measure apparent second virial coefficient (B2,app ) values for therapeutic and model monoclonal antibodies (mAbs) that capture the nature and strength of interactions with cosolutes directly in undiluted serum and similar complex biological media. Here, we use FCS-derived B2,app measurements to identify the components of human serum responsible for nonideal interactions with mAbs and Fab fragments. Most mAbs exhibit neutral or slightly attractive interactions with intact serum. Generally, mAbs display repulsive interactions with albumin and mildly attractive interactions with IgGs in the context of whole serum. Crucially, however, these attractive interactions are much stronger with pooled IgGs isolated from other serum components, indicating that the effects of serum nonideality can only be understood by studying the intact medium (rather than isolated components). Moreover, Fab fragments universally exhibited more attractive interactions than their parental mAbs, potentially rendering them more susceptible to nonideality-driven perturbations. FCS-based B2,app measurements have the potential to advance our understanding of how physiological environments impact protein-based therapeutics in general. Furthermore, incorporating such assays into preclinical biologics development may help de-risk molecules and make for a faster and more efficient development process.

Prediction of Antibody Viscosity from Dilute Solution Measurements

The high antibody doses required to achieve a therapeutic effect often necessitate high-concentration products that can lead to challenging viscosity issues in production and delivery. Predicting antibody viscosity in early development can play a pivotal role in reducing late-stage development costs. In recent years, numerous efforts have been made to predict antibody viscosity through dilute solution measurements. A key finding is that the entanglement of long, flexible complexes contributes to the sharp rise in antibody viscosity at the required dosing. This entanglement model establishes a connection between the two-body binding affinity and the many-body viscosity. Exploiting this insight, this study connects dilute solution measurements of self-association to high-concentration viscosity profiles to quantify the relationship between these regimes. The resulting model has exhibited success in predicting viscosity at high concentrations (around 150 mg/mL) from dilute solution measurements, with only a few outliers remaining. Our physics-based approach provides an understanding of fundamental physics, interpretable connections to experimental data, the potential to extrapolate beyond training conditions, and the capacity to effectively explain the physical mechanics behind these outliers. Conducting hypothesis-driven experiments that specifically target the viscosity and relaxation mechanisms of outlier molecules may allow us to unravel the intricacies of their behavior and, in turn, enhance the performance of our model.

Characterizing Protein-Protein Interactions and Viscosity of a Monoclonal Antibody from Low to High Concentration Using Small-Angle X-ray Scattering and Molecular Dynamics Simulations

Understanding protein-protein interactions and formation of reversible oligomers (clusters) in concentrated monoclonal antibody (mAb) solutions is necessary for designing stable, low viscosity (η) concentrated formulations for processing and subcutaneous injection. Here we characterize the strength (K) of short-range anisotropic attractions (SRA) for 75-200 mg/mL mAb2 solutions at different pH and cosolute conditions by analyzing structure factors (Seff(q)) from small-angle X-ray scattering (SAXS) using coarse-grained molecular dynamics simulations. Best fit simulations additionally provide cluster size distributions, fractal dimensions, cluster occluded volume, and mAb coordination numbers. These equilibrium properties are utilized in a model to account for increases in viscosity caused by occluded volume in the clusters (packing effects) and dissipation of stress across lubricated fractal clusters. Seff(q) is highly sensitive to K at 75 mg/mL where mAbs can mutually align to form SRA contacts but becomes less sensitive at 200 mg/mL as steric repulsion due to packing becomes dominant. In contrast, η at 200 mg/mL is highly sensitive to SRA and the average cluster size from SAXS/simulation, which is observed to track the cluster relaxation time from shear thinning. By analyzing the distribution of sub-bead hot spots on the 3D mAb surface, we identify a strongly attractive hydrophobic patch in the complementarity determining region (CDR) at pH 4.5 that contributes to the high K and consequently large cluster sizes and high η. Adding NaCl screens electrostatic interactions and increases the impact of hydrophobic attraction on cluster size and raises η, whereas nonspecific binding of Arg attenuates all SRA, reducing η. The hydrophobic patch is absent at higher pH values, leading to smaller K, smaller clusters, and lower η. This work constitutes a first attempt to use SAXS and CG modeling to link both structural and rheological properties of concentrated mAb solutions to the energetics of specific hydrophobic patches on mAb surfaces. As such, our work opens an avenue for future research, including the possibility of designing coarse-grained models with physically meaningful interacting hot spots.

Utilizing PK and PD Biomarkers to Guide the First-in-Human Starting Dose Selection of MTBT1466A: A Novel Humanized Monoclonal Anti-TGFβ3 Antibody for the Treatment of Fibrotic Diseases

MTBT1466A is a high-affinity TGFβ3-specific humanized IgG1 monoclonal antibody with reduced Fc effector function, currently under investigation in clinical trials as a potential anti-fibrotic therapy. Here, we characterized the pharmacokinetics (PK) and pharmacodynamics (PD) of MTBT1466A in mice and monkeys and predicted the PK/PD of MTBT1466A in humans to guide the selection of the first-in-human (FIH) starting dose. MTBT1466A demonstrated a typical IgG1-like biphasic PK profile in monkeys, and the predicted human clearance of 2.69 mL/day/kg and t1/2 of 20.4 days are consistent with those expected for a human IgG1 antibody. In a mouse model of bleomycin-induced lung fibrosis, changes in expression of TGFβ3-related genes, serpine1, fibronectin-1, and collagen 1A1 were used as PD biomarkers to determine the minimum pharmacologically active dose of 1 mg/kg. Unlike in the fibrosis mouse model, evidence of target engagement in healthy monkeys was only observed at higher doses. Using a PKPD-guided approach, the recommended FIH dose of 50 mg, IV, provided exposures that were shown to be safe and well tolerated in healthy volunteers. MTBT1466A PK in healthy volunteers was predicted reasonably well using a PK model with allometric scaling of PK parameters from monkey data. Taken together, this work provides insights into the PK/PD behavior of MTBT1466A in preclinical species, and supports the translatability of the preclinical data into the clinic.

AB-Gen: Antibody Library Design with Generative Pre-trained Transformer and Deep Reinforcement Learning

Antibody leads must fulfill multiple desirable properties to be clinical candidates. Primarily due to the low throughput in the experimental procedure, the need for such multi-property optimization causes the bottleneck in preclinical antibody discovery and development, because addressing one issue usually causes another. We developed a reinforcement learning (RL) method, named AB-Gen, for antibody library design using a generative pre-trained transformer (GPT) as the policy network of the RL agent. We showed that this model can learn the antibody space of heavy chain complementarity determining region 3 (CDRH3) and generate sequences with similar property distributions. Besides, when using human epidermal growth factor receptor-2 (HER2) as the target, the agent model of AB-Gen was able to generate novel CDRH3 sequences that fulfill multi-property constraints. Totally, 509 generated sequences were able to pass all property filters, and three highly conserved residues were identified. The importance of these residues was further demonstrated by molecular dynamics simulations, consolidating that the agent model was capable of grasping important information in this complex optimization task. Overall, the AB-Gen method is able to design novel antibody sequences with an improved success rate than the traditional propose-then-filter approach. It has the potential to be used in practical antibody design, thus empowering the antibody discovery and development process. The source code of AB-Gen is freely available at Zenodo (https://doi.org/10.5281/zenodo.7657016) and BioCode (https://ngdc.cncb.ac.cn/biocode/tools/BT007341).

Biophysical Determinants for the Viscosity of Concentrated Monoclonal Antibody Solutions

Monoclonal antibodies (mAbs) are particularly relevant for therapeutics due to their high specificity and versatility, and mAb-based drugs are hence used to treat numerous diseases. The increased patient compliance of self-administration motivates the formulation of products for subcutaneous (SC) administration. The associated challenge is to formulate highly concentrated antibody solutions to achieve a significant therapeutic effect, while limiting their viscosity and preserving their physicochemical stability. Protein-protein interactions (PPIs) are in fact the root cause of several potential problems concerning the stability, manufacturability, and delivery of a drug product. The understanding of macroscopic viscosity requires an in-depth knowledge on protein diffusion, PPIs, and self-association/aggregation. Here, we study the self-diffusion of different mAbs of the IgG1 subtype in aqueous solution as a function of the concentration and temperature by quasi-elastic neutron scattering (QENS). QENS allows us to probe the short-time self-diffusion of the molecules and therefore to determine the hydrodynamic mAb cluster size and to gain information on the internal mAb dynamics. Small-angle neutron scattering (SANS) is jointly employed to probe structural details and to understand the nature and intensity of PPIs. Complementary information is provided by molecular dynamics (MD) simulations and viscometry, thus obtaining a comprehensive picture of mAb diffusion.

Simplifying complex antibody engineering using machine learning

Machine learning is transforming antibody engineering by enabling the generation of drug-like monoclonal antibodies with unprecedented efficiency. Unsupervised algorithms trained on massive and diverse protein sequence datasets facilitate the prediction of panels of antibody variants with native-like intrinsic properties (e.g., high stability), greatly reducing the amount of subsequent experimentation needed to identify specific candidates that also possess desired extrinsic properties (e.g., high affinity). Additionally, supervised algorithms, which are trained on deep sequencing datasets obtained after enrichment of in vitro antibody libraries for one or more specific extrinsic properties, enable the prediction of antibody variants with desired combinations of extrinsic properties without the need for additional screening. Here we review recent advances using both machine learning approaches and how they are impacting the field of antibody engineering as well as key outstanding challenges and opportunities for these paradigm-changing methods.

A Proposed Global Medicines Agency (GMA) to Make Biological Drugs Accessible: Starting with the League of Arab States

Medical anthropology teaches us of historical disparity in the accessibility of medicines in the developing world due to their lack of availability and affordability, more particularly of biological drugs, including therapeutic proteins, gene therapy, CRISPR-Cas9, mRNA therapeutics, CART therapy, and many more. This challenge can be resolved by establishing an independent regulatory agency, proposed as the Global Medicines Agency (GMA), with a charter to allow originators from the Stringent Regulatory Agency (SRA) countries to receive immediate registrations applicable to all member states, expanding the market potential as an incentive. For non-SRA countries, it will be limited to biological drugs that are allowed their copies to be made, only biosimilars. A transparent approval process will involve using a rapporteur, a third-party product-related current Good Manufacturing Practice (cGMP), and assurance of the integrity of samples tested for analytical similarity and clinical pharmacology testing. GMA membership will be open to all countries. Still, it is suggested that the League of Arab States, representing 22 states with a population of 400 million, takes the lead due to their cultural and language homogeneity, which is likely to provide a concurrence among the member states. However, some states, like the Gulf Cooperative Council, are already accustomed to this approach, albeit with a different perspective. The target drugs are biotechnology and gene therapy pharmaceuticals, and their scope can be expanded to any drug.

The Evolution of Commercial Antibody Formulations

It has been nearly four decades since the first therapeutic monoclonal antibodies were approved and made available for widespread human use. Herein, US and EU approved antibody formulations are reviewed, and their nature and compositions are evaluated over time. From 1986 through Jan 2023, significant formulation trends have occurred and to represent this, 165 commercial antibody therapeutic formulations were binned into 5 different periods of time. Overall, we have observed the following: 1) The average formulation pH has decreased in recent years by over 0.5 units along with a decrease in variability that is largely driven by non-high concentration liquid in vial presentations for IV administration, 2) The use of certain excipients and buffers such as histidine, sucrose, metal chelators, arginine and methionine has become significantly more common, whereas formulations that contain phosphate, salt, no sugar or no surfactant have fallen out of favor, 3) Overall formulation space has increasingly become more homogenous and has converged in terms of formulation pH and excipient preferences regardless of formulation concentration, drug product presentation, and route of administration, 4) The average calculated isoelectric point (pI) has decreased 0.26 units, and 5) Overall, the average formulation pH and calculated pI for all commercial antibodies surveyed was 6.0 and 8.4, respectively. These trends and formulation convergence may be driven by multiple factors such as advancements in high-throughput computational and analytical technologies, the increased emphasis and understanding of certain developability attributes and formulation principles during lead selection and formulation development, and the adoption of low-risk development platform approaches.

Differences between Omicron SARS-CoV-2 RBD and other variants in their ability to interact with cell receptors and monoclonal antibodies

SARS-CoV-2 remains a health threat with the continuous emergence of new variants. This work aims to expand the knowledge about the SARS-CoV-2 receptor-binding domain (RBD) interactions with cell receptors and monoclonal antibodies (mAbs). By using constant-pH Monte Carlo simulations, the free energy of interactions between the RBD from different variants and several partners (Angiotensin-Converting Enzyme-2 (ACE2) polymorphisms and various mAbs) were predicted. Computed RBD-ACE2-binding affinities were higher for two ACE2 polymorphisms (rs142984500 and rs4646116) typically found in Europeans which indicates a genetic susceptibility. This is amplified for Omicron (BA.1) and its sublineages BA.2 and BA.3. The antibody landscape was computationally investigated with the largest set of mAbs so far in the literature. From the 32 studied binders, groups of mAbs were identified from weak to strong binding affinities (e.g. S2K146). These mAbs with strong binding capacity and especially their combination are amenable to experimentation and clinical trials because of their high predicted binding affinities and possible neutralization potential for current known virus mutations and a universal coronavirus.Communicated by Ramaswamy H. Sarma.

Subclass Effects on Self-Association and Viscosity of Monoclonal Antibodies at High Concentrations

The effects of a subclass of monoclonal antibodies (mAbs) on protein-protein interactions, formation of reversible oligomers (clusters), and viscosity (η) are not well understood at high concentrations. Herein, we quantify a short-range anisotropic attraction between the complementarity-determining region (CDR) and CH3 domains (K_CDR-CH3) for vedolizumab IgG1, IgG2, or IgG4 subclasses by fitting small-angle X-ray scattering (SAXS) structure factor S_eff(q) data with an extensive library of 12-bead coarse-grained (CG) molecular dynamics simulations. The K_CDR-CH3 bead attraction strength was isolated from the strength of long-range electrostatic repulsion for the full mAb, which was determined from the theoretical net charge and a scaling parameter ψ to account for solvent accessibility and ion pairing. At low ionic strength (IS), the strongest short-range attraction (K_CDR-CH3) and consequently the largest clusters and highest η were observed with IgG1, the subclass with the most positively charged CH3 domain. Furthermore, the trend in K_CDR-CH3 with the subclass followed the electrostatic interaction energy between the CDR and CH3 regions calculated with the BioLuminate software using the 3D mAb structure and molecular interaction potentials. Whereas the equilibrium cluster size distributions and fractal dimensions were determined from fits of SAXS with the MD simulations, the degree of cluster rigidity under flow was estimated from the experimental η with a phenomenological model. For the systems with the largest clusters, especially IgG1, the inefficient packing of mAbs in the clusters played the largest role in increasing η, whereas for other systems, the relative contribution from stress produced by the clusters was more significant. The ability to relate η to short-range attraction from SAXS measurements at high concentrations and to theoretical characterization of electrostatic patches on the 3D surface is not only of fundamental interest but also of practical value for mAb discovery, processing, formulation, and subcutaneous delivery.

Physiologically Based Pharmacokinetic Modeling to Characterize the Effect of Molecular Charge on Whole-Body Disposition of Monoclonal Antibodies

Motivated by a series of work demonstrating the effect of molecular charge on antibody pharmacokinetics (PK), physiological-based pharmacokinetic (PBPK) models are emerging that relate in silico calculated charge or in vitro measures of polyspecificity to antibody PK parameters. However, only plasma data has been used for model development in these studies, leading to unvalidated assumptions. Here, we present an extended platform PBPK model for antibodies that incorporate charge-dependent endothelial cell pinocytosis rate and nonspecific off-target binding in the interstitial space and on circulating blood cells, to simultaneously characterize whole-body disposition of three antibody charge variants. Predictive potential of various charge metrics was also explored, and the difference between positive charge patches and negative charge patches (i.e., PPC-PNC) was used as the charge parameter to establish quantitative relationships with nonspecific binding affinities and endothelial cell uptake rate. Whole-body disposition of these charge variants was captured well by the model, with less than 2-fold predictive error in area under the curve of most plasma and tissue PK data. The model also predicted that with greater positive charge, nonspecific binding was more substantial, and pinocytosis rate increased especially in brain, heart, kidney, liver, lung, and spleen, but remained unchanged in adipose, bone, muscle, and skin. The presented PBPK model contributes to our understanding of the mechanisms governing the disposition of charged antibodies and can be used as a platform to guide charge engineering based on desired plasma and tissue exposures.

Surface patches induce nonspecific binding and phase separation of antibodies

Nonspecific interactions are a key challenge in the successful development of therapeutic antibodies. The tendency for nonspecific binding of antibodies is often difficult to reduce by rational design, and instead, it is necessary to rely on comprehensive screening campaigns. To address this issue, we performed a systematic analysis of the impact of surface patch properties on antibody nonspecificity using a designer antibody library as a model system and single-stranded DNA as a nonspecificity ligand. Using an in-solution microfluidic approach, we find that the antibodies tested bind to single-stranded DNA with affinities as high as KD = 1 µM. We show that DNA binding is driven primarily by a hydrophobic patch in the complementarity-determining regions. By quantifying the surface patches across the library, the nonspecific binding affinity is shown to correlate with a trade-off between the hydrophobic and total charged patch areas. Moreover, we show that a change in formulation conditions at low ionic strengths leads to DNA-induced antibody phase separation as a manifestation of nonspecific binding at low micromolar antibody concentrations. We highlight that phase separation is driven by a cooperative electrostatic network assembly mechanism of antibodies with DNA, which correlates with a balance between positive and negative charged patches. Importantly, our study demonstrates that both nonspecific binding and phase separation are controlled by the size of the surface patches. Taken together, these findings highlight the importance of surface patches and their role in conferring antibody nonspecificity and its macroscopic manifestation in phase separation.

Low-data interpretable deep learning prediction of antibody viscosity using a biophysically meaningful representation

Deep learning, aided by the availability of big data sets, has led to substantial advances across many disciplines. However, many scientific problems of practical interest lack sufficiently large datasets amenable to deep learning. Prediction of antibody viscosity is one such problem where deep learning methods have not yet been explored due to the relative scarcity of relevant training data. In this work, we overcome this limitation using a biophysically meaningful representation that enables us to develop generalizable models even under limited training data. We present, PfAbNet-viscosity, a 3D convolutional neural network architecture, to predict high-concentration viscosity of therapeutic antibodies. We show that with the electrostatic potential surface of the antibody variable region as the only input to the network, the models trained on as few as couple dozen datapoints can generalize with high accuracy. Our feature attribution analysis shows that PfAbNet-viscosity has learned key biophysical drivers of viscosity. The applicability of our approach to other biological systems is discussed.

Characterizing Experimental Monoclonal Antibody Interactions and Clustering Using a Coarse-Grained Simulation Library and a Viscosity Model

Attractive protein-protein interactions in concentrated monoclonal antibody (mAb) solutions may lead to the formation of clusters that increase viscosity. Here, we propose an analytical model that relates mAb solution viscosity to clustering by accounting for the contributions of suboptimal mAb packing within a cluster and cluster fractal dimension. The influence of short-range, anisotropic attractions and long-range Coulombic repulsion on cluster properties is investigated by analyzing the cluster-size distributions, cluster fractal dimensions, radial distribution functions, and static structure factors from a library of coarse-grained molecular dynamics simulations. The library spans a vast range of mAb charges and attractive interactions in solutions of varying ionic strength. We present a framework for combining the viscosity model and simulation library to successfully characterize the attraction, repulsion, and clustering of an experimental mAb in three different pH and cosolute conditions by fitting the measured viscosity or structure factor from small-angle X-ray scattering. At low ionic strength, the cluster-size distribution is impacted by strong charges, and both the viscosity and net charge or structure factor and net charge must be considered to deconvolute the effects of short-range attraction and long-range repulsion.

Embedding Dynamics in Intrinsic Physicochemical Profiles of Market-Stage Antibody-Based Biotherapeutics

Adequate stability, manufacturability, and safety are crucial to bringing an antibody-based biotherapeutic to the market. Following the concept of holistic in silico developability, we introduce a physicochemical description of 91 market-stage antibody-based biotherapeutics based on orthogonal molecular properties of variable regions (Fvs) embedded in different simulation environments, mimicking conditions experienced by antibodies during manufacturing, formulation, and in vivo. In this work, the evaluation of molecular properties includes conformational flexibility of the Fvs using molecular dynamics (MD) simulations. The comparison between static homology models and simulations shows that MD significantly affects certain molecular descriptors like surface molecular patches. Moreover, the structural stability of a subset of Fv regions is linked to changes in their specific molecular interactions with ions in different experimental conditions. This is supported by the observation of differences in protein melting temperatures upon addition of NaCl. A DEvelopability Navigator In Silico (DENIS) is proposed to compare mAb candidates for their similarity with market-stage biotherapeutics in terms of physicochemical properties and conformational stability. Expanding on our previous developability guidelines (Ahmed et al. Proc. Natl. Acad. Sci. 2021, 118 (37), e2020577118), the hydrodynamic radius and the protein strand ratio are introduced as two additional descriptors that enable a more comprehensive in silico characterization of biotherapeutic drug candidates. Test cases show how this approach can facilitate identification and optimization of intrinsically developable lead candidates. DENIS represents an advanced computational tool to progress biotherapeutic drug candidates from discovery into early development by predicting drug properties in different aqueous environments.

An Intercompany Perspective on Practical Experiences of Predicting, Optimizing and Analyzing High Concentration Biologic Therapeutic Formulations

Developing high-dose biologic drugs for subcutaneous injection often requires high-concentration formulations and optimizing viscosity, solubility, and stability while overcoming analytical, manufacturing, and administration challenges. To understand industry approaches for developing high-concentration formulations, the Formulation Workstream of the BioPhorum Development Group, an industry-wide consortium, conducted an inter-company collaborative exercise which included several surveys. This collaboration provided an industry perspective, experience, and insight into the practicalities for developing high-concentration biologics. To understand solubility and viscosity, companies desire predictive tools, but experience indicates that these are not reliable and experimental strategies are best. Similarly, most companies prefer accelerated and stress stability studies to in-silico or biophysical-based prediction methods to assess aggregation. In addition, optimization of primary container-closure and devices are pursued to mitigate challenges associated with high viscosity of the formulation. Formulation strategies including excipient selection and application of studies at low concentration to high-concentration formulations are reported. Finally, analytical approaches to high concentration formulations are presented. The survey suggests that although prediction of viscosity, solubility, and long-term stability is desirable, the outcome can be inconsistent and molecule dependent. Significant experimental studies are required to confirm robust product definition as modeling at low protein concentrations will not necessarily extrapolate to high concentration formulations.

Protein Stability After Administration: A Physiologic Consideration

Regulatory authorities and the scientific community have identified the need to monitor the in vivo stability of therapeutic proteins (TPs). Due to the unique physiologic conditions in patients, the stability of TPs after administration can deviate largely from their stability under drug product (DP) conditions. TPs can degrade at substantial rates once immersed in the in vivo milieu. Changes in protein stability upon administration to patients are critical as they can have implications on patient safety and clinical effectiveness of DPs. Physiologic conditions are challenging to simulate and require dedicated in vitro models for specific routes of administration. Advancements of in vitro models enable to simulate the exposure to physiologic conditions prior to resource demanding pre-clinical and clinical studies. This enables to evaluate the in vivo stability and thus may allow to improve the safety/efficacy profile of DPs. While in vitro-in vivo correlations are challenging, benchmarking DP candidates enables to identify liabilities and optimize molecules. The in vivo stability should be an integral part of holistic stability assessments during early development. Such assessments can accelerate development timelines and lead to more stable DPs for patients.

Identifying developability risks for clinical progression of antibodies using high-throughput in vitro and in silico approaches

With the growing significance of antibodies as a therapeutic class, identifying developability risks early during development is of paramount importance. Several high-throughput in vitro assays and in silico approaches have been proposed to de-risk antibodies during early stages of the discovery process. In this review, we have compiled and collectively analyzed published experimental assessments and computational metrics for clinical antibodies. We show that flags assigned based on in vitro measurements of polyspecificity and hydrophobicity are more predictive of clinical progression than their in silico counterparts. Additionally, we assessed the performance of published models for developability predictions on molecules not used during model training. We find that generalization to data outside of those used for training remains a challenge for models. Finally, we highlight the challenges of reproducibility in computed metrics arising from differences in homology modeling, in vitro assessments relying on complex reagents, as well as curation of experimental data often used to assess the utility of high-throughput approaches. We end with a recommendation to enable assay reproducibility by inclusion of controls with disclosed sequences, as well as sharing of structural models to enable the critical assessment and improvement of in silico predictions.

Predictive modeling of concentration-dependent viscosity behavior of monoclonal antibody solutions using artificial neural networks

Solutions of monoclonal antibodies (mAbs) can show increased viscosity at high concentration, which can be a disadvantage during protein purification, filling, and administration. The viscosity is determined by protein-protein-interactions, which are influenced by the antibody's sequence as well as solution conditions, like pH, buffer type, or the presence of salts and other excipients. To predict viscosity, experimental parameters, like the diffusion interaction parameter (kD), or computational tools harnessing information derived from primary sequence, are often used, but a reliable predictive tool is still missing. We present a modeling approach employing artificial neural networks (ANNs) using experimental factors combined with simulation-derived parameters plus viscosity data from 27 highly concentrated (180 mg/mL) mAbs. These ANNs can be used to predict if mAbs exhibit problematic viscosity at distinct concentrations or to model viscosity-concentration-curves.

In Silico Prediction Method for Protein Asparagine Deamidation

In silico prediction methods were developed to predict protein asparagine (Asn) deamidation. The method is based on understanding deamidation mechanism on structural level with machine learning. Our structure-based method is more accurate than the sequence-based method which is still widely used in protein engineering process. In addition, molecular dynamics simulation was applied to study the time occupancy of nucleophilic attack distance, which is hypothesized as the most important step toward the rate-limiting succinimide intermediate formation. A more accurate prediction method for distinguishing potentially liable amino acid residues would allow their elimination or reduction as early as possible in the drug discovery process. It is possible that such quantitative protein structure-property relationship tools can also be applied to other protein hotspot predictions.

Assessing developability early in the discovery process for novel biologics

Beyond potency, a good developability profile is a key attribute of a biological drug. Selecting and screening for such attributes early in the drug development process can save resources and avoid costly late-stage failures. Here, we review some of the most important developability properties that can be assessed early on for biologics. These include the influence of the source of the biologic, its biophysical and pharmacokinetic properties, and how well it can be expressed recombinantly. We furthermore present in silico, in vitro, and in vivo methods and techniques that can be exploited at different stages of the discovery process to identify molecules with liabilities and thereby facilitate the selection of the most optimal drug leads. Finally, we reflect on the most relevant developability parameters for injectable versus orally delivered biologics and provide an outlook toward what general trends are expected to rise in the development of biologics.

Structure-Based Optimization of Antibody-Based Biotherapeutics for Improved Developability: A Practical Guide for Molecular Modelers

A great effort to avoid known developability risks is now more often being made earlier during the lead candidate discovery and optimization phase of biotherapeutic drug development. Predictive computational strategies, used in the early stages of antibody discovery and development, to mitigate the risk of late-stage failure of antibody candidates, are highly valuable. Various structure-based methods exist for accurately predicting properties critical to developability, and, in this chapter, we discuss the history of their development and demonstrate how they can be used to filter large sets of candidates arising from target affinity screening and to optimize lead candidates for developability. Methods for modeling antibody structures from sequence and detecting post-translational modifications and chemical degradation liabilities are also discussed.

Development of in silico models to predict viscosity and mouse clearance using a comprehensive analytical data set collected on 83 scaffold-consistent monoclonal antibodies

Biologic drug discovery pipelines are designed to deliver protein therapeutics that have exquisite functional potency and selectivity while also manifesting biophysical characteristics suitable for manufacturing, storage, and convenient administration to patients. The ability to use computational methods to predict biophysical properties from protein sequence, potentially in combination with high throughput assays, could decrease timelines and increase the success rates for therapeutic developability engineering by eliminating lengthy and expensive cycles of recombinant protein production and testing. To support development of high-quality predictive models for antibody developability, we designed a sequence-diverse panel of 83 effector functionless IgG1 antibodies displaying a range of biophysical properties, produced and formulated each protein under standard platform conditions, and collected a comprehensive package of analytical data, including in vitro assays and in vivo mouse pharmacokinetics. We used this robust training data set to build machine learning classifier models that can predict complex protein behavior from these data and features derived from predicted and/or experimental structures. Our models predict with 87% accuracy whether viscosity at 150 mg/mL is above or below a threshold of 15 centipoise (cP) and with 75% accuracy whether the area under the plasma drug concentration-time curve (AUC0-672 h) in normal mouse is above or below a threshold of 3.9 × 106 h x ng/mL.

Utility of physiologically based pharmacokinetic modeling to predict inter-antibody variability in monoclonal antibody pharmacokinetics in mice

In this investigation, we tested the hypothesis that a physiologically based pharmacokinetic (PBPK) model incorporating measured in vitro metrics of off-target binding can largely explain the inter-antibody variability in monoclonal antibody (mAb) pharmacokinetics (PK). A diverse panel of 83 mAbs was evaluated for PK in wild-type mice and subjected to 10 in vitro assays to measure major physiochemical attributes. After excluding for target-mediated elimination and immunogenicity, 56 of the remaining mAbs with an eight-fold variability in the area under the curve (A U C 0 - 672 h : 1.74 × 106 -1.38 × 107 ng∙h/mL) and 10-fold difference in clearance (2.55-26.4 mL/day/kg) formed the training set for this investigation. Using a PBPK framework, mAb-dependent coefficients F1 and F2 modulating pinocytosis rate and convective transport, respectively, were estimated for each mAb with mostly good precision (coefficient of variation (CV%) <30%). F1 was estimated to be the mean and standard deviation of 0.961 ± 0.593, and F2 was estimated to be 2.13 ± 2.62. Using principal component analysis to correlate the regressed values of F1/F2 versus the multidimensional dataset composed of our panel of in vitro assays, we found that heparin chromatography retention time emerged as the predictive covariate to the mAb-specific F1, whereas F2 variability cannot be well explained by these assays. A sigmoidal relationship between F1 and the identified covariate was incorporated within the PBPK framework. A sensitivity analysis suggested plasma concentrations to be most sensitive to F1 when F1 > 1. The predictive utility of the developed PBPK model was evaluated against a separate panel of 14 mAbs biased toward high clearance, among which area under the curve of PK data of 12 mAbs was predicted within 2.5-fold error, and the positive and negative predictive values for clearance prediction were 85% and 100%, respectively. MAb heparin chromatography assay output allowed a priori identification of mAb candidates with unfavorable PK.

Blueprint for antibody biologics developability

Large-molecule antibody biologics have revolutionized medicine owing to their superior target specificity, pharmacokinetic and pharmacodynamic properties, safety and toxicity profiles, and amenability to versatile engineering. In this review, we focus on preclinical antibody developability, including its definition, scope, and key activities from hit to lead optimization and selection. This includes generation, computational and in silico approaches, molecular engineering, production, analytical and biophysical characterization, stability and forced degradation studies, and process and formulation assessments. More recently, it is apparent these activities not only affect lead selection and manufacturability, but ultimately correlate with clinical progression and success. Emerging developability workflows and strategies are explored as part of a blueprint for developability success that includes an overview of the four major molecular properties that affect all developability outcomes: 1) conformational, 2) chemical, 3) colloidal, and 4) other interactions. We also examine risk assessment and mitigation strategies that increase the likelihood of success for moving the right candidate into the clinic.

High concentration formulation developability approaches and considerations

The growing need for biologics to be administered subcutaneously and ocularly, coupled with certain indications requiring high doses, has resulted in an increase in drug substance (DS) and drug product (DP) protein concentrations. With this increase, more emphasis must be placed on identifying critical physico-chemical liabilities during drug development, including protein aggregation, precipitation, opalescence, particle formation, and high viscosity. Depending on the molecule, liabilities, and administration route, different formulation strategies can be used to overcome these challenges. However, due to the high material requirements, identifying optimal conditions can be slow, costly, and often prevent therapeutics from moving rapidly into the clinic/market. In order to accelerate and derisk development, new experimental and in-silico methods have emerged that can predict high concentration liabilities. Here, we review the challenges in developing high concentration formulations, the advances that have been made in establishing low mass and high-throughput predictive analytics, and advances in in-silico tools and algorithms aimed at identifying risks and understanding high concentration protein behavior.

Monoclonal Antibody Pharmacokinetics in Cynomolgus Monkeys Following Subcutaneous Administration: Physiologically Based Model Predictions from Physiochemical Properties

An integrated physiologically based modeling framework is presented for predicting pharmacokinetics and bioavailability of subcutaneously administered monoclonal antibodies in cynomolgus monkeys, based on in silico structure-derived metrics characterizing antibody size, overall charge, local charge, and hydrophobicity. The model accounts for antibody-specific differences in pinocytosis, transcapillary transport, local lymphatic uptake, and pre-systemic degradation at the subcutaneous injection site and reliably predicts the pharmacokinetics of five different wild-type mAbs and their Fc variants following intravenous and subcutaneous administration. Significant associations were found between subcutaneous injection site degradation rate and the antibody's local positive charge of its complementarity-determining region (R = 0.56, p = 0.0012), antibody pinocytosis rate and its overall positive charge (R = 0.59, p = 0.00063), and antibody paracellular transport and its overall charge together with hydrophobicity (R = 0.63, p = 0.00096). Based on these results, population simulations were performed to predict the relationship between bioavailability and antibody local positive charge. In addition, model simulations were conducted to calculate the relative contribution of absorption pathways (lymphatic and blood), pre-systemic degradation pathways (interstitial and lysosomal), and the influence of injection site lymph flow on antibody bioavailability and pharmacokinetics. The proposed physiologically based modeling framework integrates fundamental mechanisms governing antibody subcutaneous absorption and disposition, with structured-based physiochemical properties, to predict antibody bioavailability and pharmacokinetics in vivo.

Non-specificity as the sticky problem in therapeutic antibody development

Antibodies are highly potent therapeutic scaffolds with more than a hundred different products approved on the market. Successful development of antibody-based drugs requires a trade-off between high target specificity and target binding affinity. In order to better understand this problem, we here review non-specific interactions and explore their fundamental physicochemical origins. We discuss the role of surface patches - clusters of surface-exposed amino acid residues with similar physicochemical properties - as inducers of non-specific interactions. These patches collectively drive interactions including dipole-dipole, π-stacking and hydrophobic interactions to complementary moieties. We elucidate links between these supramolecular assembly processes and macroscopic development issues, such as decreased physical stability and poor in vivo half-life. Finally, we highlight challenges and opportunities for optimizing protein binding specificity and minimizing non-specificity for future generations of therapeutics.