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.

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.

Improved Diffusion Interaction Parameter Measurement to Predict the Viscosity of Concentrated mAb Solutions.

During the developability assessment of therapeutic monoclonal antibody (mAb) candidates, utilization of robust high-throughput predictive assays enables rapid selection of top candidates with low risks for late-stage development. Predicting the viscosities of highly concentrated mAbs using limited materials is an important aspect of developability assessment because high viscosity can complicate manufacturability, stability, and administration. Here, we report a high-throughput assay measuring protein-protein interactions to predict mAb viscosity. The diffusion interaction parameter (kD) measures colloidal self-association in dilute solutions and has been reported to be predictive of the mAb viscosity at high concentrations. However, kD of Amgen early stage IgG1 mAb candidates measured in 10 mM acetate at pH 5.2 containing sucrose and polysorbate (denoted A52SuT) shows only weak correlation to their viscosities at 140 mg/mL in A52SuT. We hypothesize that kD measured in A52SuT reflects primarily long-range electrostatic repulsions because most of these mAb candidates carry strong net positive charges in this low ionic strength formulation with pH (5.2) well below pI values of mAb candidates. However, the viscosities of high concentration mAbs depend heavily on short-range molecular interactions. We propose an improved kD method in which salt is added to suppress charge repulsions and to allow for detection of key short-range interactions in dilute solutions. Salt types and salt concentrations were screened, and an optimal salt condition was identified. This optimized method was further validated using two test mAb sets. Overall, the method improves the Pearson R2 between kD and viscosity (6-230 cP) from 0.24 to 0.80 for a data set consisting of 37 mAbs.

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.

Microfluidic viscometer by acoustic streaming transducers.

Measurement of fluid viscosity represents a huge need for many biomedical and materials processing applications. Sample fluids containing DNA, antibodies, protein-based drugs, and even cells have become important therapeutic options. The physical properties, including viscosity, of these biologics are critical factors in the optimization of the biomanufacturing processes and delivery of therapeutics to patients. Here we demonstrate an acoustic microstreaming platform termed as microfluidic viscometer by acoustic streaming transducers (µVAST) that induces fluid transport from second-order microstreaming to measure viscosity. Validation of our platform is achieved with different glycerol content mixtures to reflect different viscosities and shows that viscosity can be estimated based on the maximum speed of the second-order acoustic microstreaming. The µVAST platform requires only a small volume of fluid sample (∼1.2 µL), which is 16-30 times smaller than that of commercial viscometers. In addition, µVAST can be scaled up for ultra-high throughput measurements of viscosity. Here we demonstrate 16 samples within 3 seconds, which is an attractive feature for automating the process flows in drug development and materials manufacturing and production.

High-Throughput Mass Spectrometry for Biopharma: A Universal Modality and Target Independent Analytical Method for Accurate Biomolecule Characterization.

Reversed-phase liquid chromatographic mass spectrometry (rpLC-MS) is a universal, platformed, and essential analytical technique within pharmaceutical and biopharmaceutical research. Typical rpLC method gradient times can range from 5 to 20 min. As monoclonal antibody (mAb) therapies continue to evolve and bispecific antibodies (BsAbs) become more established, research stage engineering panels will clearly evolve in size. Therefore, high-throughput (HT) MS and automated deconvolution methods are key for success. Additionally, newer therapeutics such as bispecific T-cell engagers and nucleic acid-based modalities will also require MS characterization. Herein, we present a modality and target agnostic HT solid-phase extraction (SPE) MS method that affords the analysis of a 96-well plate in 41.4 min, compared to the traditional rpLC-MS method that would typically take 14.4 h. The described method can accurately determine the molecular weights for monodispersed and highly polydispersed biotherapeutic species and membrane proteins; determine levels of glycosylation, glycation, and formylation; detect levels of chain mispairing; and determine accurate drug-to-antibody ratio values.

Synthesis of Zwitterionic and Trehalose Polymers with Variable Degradation Rates and Stabilization of Insulin.

Polymers that stabilize biomolecules are important as excipients in protein formulation. Herein, we describe a class of degradable polymers that have tunable degradation rates depending on the polymer backbone and can stabilize proteins to aggregation. Specifically, zwitterion- and trehalose-substituted polycaprolactone, polyvalerolactone, polycarbonate, and polylactide were prepared and characterized with regards to their hydrolytic degradation and ability to stabilize insulin to mechanical agitation during heat. Ring-opening polymerization (ROP) of allyl-substituted monomers was performed by using organocatalysis, resulting in well-defined alkene-substituted polymers with good control over molecular weight and dispersity. The polymers were then modified by using photocatalyzed thiol-ene reactions to install protein-stabilizing carboxybetaine and trehalose side chains. The resulting polymers were water-soluble and exhibited a wide range of half-lives, from 12 h to more than 3 months. The polymers maintained the ability to stabilize the therapeutic protein insulin from activity loss due to aggregation, demonstrating their potential as degradable excipients for protein formulation.

PEG Analogs Synthesized by Ring-Opening Metathesis Polymerization for Reversible Bioconjugation.

Poly(ethylene glycols) (PEGs) with protein-reactive end-groups are widely utilized in bioconjugation reactions. Herein, we describe the use of ring-opening metathesis polymerization (ROMP) to synthesize unsaturated protein-reactive PEG analogs. These ROMP PEGs (rPEGs) contained terminal aldehyde functionality and ranged in molecular weight from 6 to 20 kDa. The polymers were readily conjugated to free amines on the protein hen egg-white lysozyme (Lyz). Biocompatibility of the unsaturated PEGs was assessed in vitro, revealing the polymers to be nontoxic up to concentrations of at least 1 mg/mL in human dermal fibroblasts (HDFs). The resulting unsaturated rPEG-lysozyme conjugates underwent metathesis-based depolymerization, resulting in decreased molecular weight of the conjugate.

Synthesis and Biological Evaluation of a Degradable Trehalose Glycopolymer Prepared by RAFT Polymerization.

There is a significant need for new biodegradable protein stabilizing polymers. Herein, the synthesis of a polymer with trehalose side chains and hydrolytically degradable backbone esters and its evaluation for protein stabilization and cytotoxicity are described. Specifically, an alkene-containing parent polymer is synthesized by reversible addition-fragmentation chain transfer polymerization, and thiolated trehalose is installed using a radical-initiated thiol-ene reaction. The stabilizing properties of the polymer are investigated by thermally stressing granulocyte colony-stimulating factor (G-CSF), which is expressed and purified using a custom-designed G-CSF fusion protein with a polyhistidine-tagged maltose binding protein. The degradable polymer is shown to stabilize G-CSF to 66% after heating at 40 °C. Poly(5,6-benzo-2-methylene-1,3-dioxepane (BMDO)-co-butyl methacrylate-trehalose) is degraded and its cellular compatibility is investigated. While the polymer is noncytotoxic, cytotoxic effects are observed from the degraded products in fibroblasts and murine myeloblasts. These data provide important information for future use of BMDO-containing trehalose glycopolymers for biomedical applications.

Substituted Polyesters by Thiol-Ene Modification: Rapid Diversification for Therapeutic Protein Stabilization.

Many proteins, especially those used as therapeutics, are unstable to storage and shipping temperatures, leading to increased costs in research and industry. Therefore, the design and synthesis of novel stabilizers is an important area of investigation. Herein we report new degradable polymers that stabilize proteins to environmental stressors such as refrigeration and elevated temperature. Specifically, polycaprolactones with different pendant groups were synthesized and surveyed for their ability to stabilize an important therapeutic protein to storage and shipping conditions. Ring-opening polymerization (ROP) of an allyl-substituted caprolactone monomer was carried out using the organocatalyst 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) to yield a well-defined, alkene-substituted degradable polymer, which was used as a common backbone to control for the degree of polymerization. Relevant side chains such as trehalose, lactose, glucose, carboxybetaine, and oligo(ethylene glycol) were installed via postpolymerization thiol-ene reactions. These degradable polymers were then employed as excipients for the stabilization of the therapeutic protein granulocyte colony-stimulating factor (G-CSF) against storage at 4 °C and shipping temperatures of 60 °C. The best stabilization was observed using the trehalose- and zwitterion- substituted polyesters. Both the trehalose- and carboxybetaine-substituted pCL were further investigated with regard to molecular weight dependence, and it was found that the molecular weight was minimally important for stabilization to refrigeration, but critical for G-CSF stabilization at elevated temperatures. Both high performing zwitterionic and trehalose polyesters were also degraded, and the polymers and degradation products were shown to be noncytotoxic. This work provides potential biocompatible polymers for stabilization of the important therapeutic G-CSF, as well as a general platform for the future discovery of new polymeric protein stabilizers.

Controlled Radical Polymerization as an Enabling Approach for the Next Generation of Protein-Polymer Conjugates.

Protein-polymer conjugates are unique constructs that combine the chemical properties of a synthetic polymer chain with the biological properties of a biomacromolecule. This often leads to improved stabilities, solubilities, and in vivo half-lives of the resulting conjugates, and expands the range of applications for the proteins. However, early chemical methods for protein-polymer conjugation often required multiple polymer modifications, which were tedious and low yielding. To solve these issues, work in our laboratory has focused on the development of controlled radical polymerization (CRP) techniques to improve synthesis of protein-polymer conjugates. Initial efforts focused on the one-step syntheses of protein-reactive polymers through the use of functionalized initiators and chain transfer agents. A variety of functional groups such as maleimide and pyridyl disulfide could be installed with high end-group retention, which could then react with protein functional groups through mild and biocompatible chemistries. While this grafting to method represented a significant advance in conjugation technique, purification and steric hindrance between large biomacromolecules and polymer chains often led to low conjugation yields. Therefore, a grafting from approach was developed, wherein a polymer chain is grown from an initiating site on a functionalized protein. These conjugates have demonstrated improved homogeneity, characterization, and easier purification, while maintaining protein activity. Much of this early work utilizing CRP techniques focused on polymers made up of biocompatible but nonfunctional monomer units, often containing oligoethylene glycol meth(acrylate) or N-isopropylacrylamide. These branched polymers have significant advantages compared to the historically used linear poly(ethylene glycols) including decreased viscosities and thermally responsive behavior, respectively. Recently, we were motivated to use CRP techniques to develop polymers with rationally designed and functional biological properties for conjugate preparation. Specifically, two families of saccharide-inspired polymers were developed for stabilization and activation of therapeutic biomolecules. A series of polymers with trehalose side-chains and vinyl backbones were prepared and used to stabilize proteins against heat and lyophilization stress as both conjugates and additives. These materials, which combine properties of osmolytes with nonionic surfactants, have significant potential for in vivo therapeutic use. Additionally, polymers that mimic the structure of the naturally occurring polysaccharide heparin were prepared. These polymers contained negatively charged sulfonate groups and imparted stabilization to a heparin-binding growth factor after conjugation. A screen of other sulfonated polymers led to the development of a polymer with improved heparin mimesis, enhancing both stability and activity of the protein to which it was attached. Chemical improvements over the past decade have enabled the preparation of a diverse set of protein-polymer conjugates by controlled polymerization techniques. Now, the field should thoroughly explore and expand both the range of polymer structures and also the applications available to protein-polymer conjugates. As we move beyond medicine toward broader applications, increased collaboration and interdisciplinary work will result in the further development of this exciting field.

Therapeutic protein-polymer conjugates: advancing beyond PEGylation.

Protein-polymer conjugates are widely used as therapeutics. All Food and Drug Administration (FDA)-approved protein conjugates are covalently linked to poly(ethylene glycol) (PEG). These PEGylated drugs have longer half-lives in the bloodstream, leading to less frequent dosing, which is a significant advantage for patients. However, there are some potential drawbacks to PEG that are driving the development of alternatives. Polymers that display enhanced pharmacokinetic properties along with additional advantages such as improved stability or degradability will be important to advance the field of protein therapeutics. This perspective presents a summary of protein-PEG conjugates for therapeutic use and alternative technologies in various stages of development as well as suggestions for future directions. Established methods of producing protein-PEG conjugates and new approaches utilizing controlled radical polymerization are also covered.