Computational Screening for mAb Colloidal Stability with Coarse-Grained, Molecular-Scale Simulations.

Monoclonal antibodies (mAbs) are an important modality of protein therapeutics with broad applications for numerous diseases. However, colloidal instabilities occurring at high protein concentrations can limit the ability to develop stable, high-concentration liquid dosage forms that are required for patient-centric, device-mediated products. Therefore, it is advantageous to identify colloidally stable mAbs early in the discovery process to ensure that they are selected for development. Experimental screening for colloidal stability can be time- and resource-consuming and is most feasible at the later stages of drug development due to material requirements. Alternatively, computational approaches have emerging potential to provide efficient screening and focus developmental efforts on mAbs with the greatest developability potential, while providing mechanistic relationships for colloidal instability. In this work, coarse-grained, molecular-scale models were fine-tuned to screen for colloidal stability at amino-acid resolution. This model parameterization provides a framework to screen for mAb self-interactions and extrapolate to bulk solution behavior. This approach was applied to a wide array of mAbs under multiple buffer conditions, demonstrating the utility of the presented computational approach to augment early candidate screening and later formulation strategies for protein therapeutics.

Native and Non-Native aggregation pathways of antibodies anticipated by cold-accelerated studies.

Assessment of cold stability is essential for manufacture and commercialization of biotherapeutics. Storage stability is often estimated by measuring accelerated rates at elevated temperature and using mathematical models (as the Arrhenius equation). Although, this strategy often leads to an underestimation of protein aggregation during storage. In this work, we measured the aggregation rates of two antibodies in a broad temperature range (from 60 °C to -25 °C), using an isochoric cooling method to prevent freezing of the formulations below 0 °C. Both antibodies evidenced increasing aggregation rates when approaching extreme temperatures, because of hot and cold denaturation. This behavior was modelled using Arrhenius and Gibbs-Helmholtz equations, which enabled to deconvolute the contribution of unfolding from the protein association kinetics. This approach made possible to model the aggregation rates at refrigeration temperature (5 °C) in a relatively short timeframe (1-2 weeks) and using standard characterization techniques (SEC-HPLC and DLS).

AF4 and PEG Precipitation as Predictive Assays for Antibody Self-Association.

Monoclonal antibodies (mAbs) are often formulated as high-protein-concentration solutions, which in some cases can exhibit physical stability issues such as high viscosity and opalescence. To ensure that mAb-based drugs can meet their manufacturing, stability, and delivery requirements, it is advantageous to screen for and select mAbs during discovery that are not prone to such behaviors. It has been recently shown that both these macroscopic properties can be predicted to a certain extent from the diffusion interaction parameter (kD), which is a measure of self-association under dilute conditions.1 However, kD can be challenging to measure at the early stage of discovery, where a relatively large amount of a high-purity material, which is required by traditional methods, is often not available. In this study, we demonstrate asymmetric field-flow fractionation (AF4) as a tool to measure self-association and therefore identify antibodies with problematic issues at high concentrations. The principle lies on the ability to concentrate the sample close to the membrane during the injection mode, which can reach formulation-relevant concentrations (>100 mg/mL).2 By analyzing a well-characterized library of commercial antibodies, we show that the measured retention time of the antibodies allows us to pinpoint molecules that exhibit issues at high concentrations. Remarkably, our AF4 assay requires very little (30 µg) sample under dilute conditions and does not need extensive sample purification. Furthermore, we show that a polyethylene glycol (PEG) precipitation assay provides results consistent with AF4 and moreover can further differentiate molecules with issues of opalescence or high viscosity. Overall, our results delineate a two-step strategy for the identification of problematic variants at high concentrations, with AF4 for early developability screening, followed by a PEG assay to validate the problematic molecules and further discriminate between opalescence or high-viscosity issues. This two-step antibody selection strategy enables us to select antibodies early in the discovery process, which are compatible with high-concentration formulations.

Ultradilute Measurements of Self-Association for the Identification of Antibodies with Favorable High-Concentration Solution Properties.

There is significant interest in formulating antibody therapeutics as concentrated liquid solutions, but early identification of developable antibodies with optimal manufacturability, stability, and delivery attributes remains challenging. Traditional methods of identifying developable mAbs with low self-association in common antibody formulations require relatively concentrated protein solutions (>1 mg/mL), and this single challenge has frustrated early-stage and large-scale identification of antibody candidates with drug-like colloidal properties. Here, we describe charge-stabilized self-interaction nanoparticle spectroscopy (CS-SINS), an affinity-capture nanoparticle assay that measures colloidal self-interactions at ultradilute antibody concentrations (0.01 mg/mL), and is predictive of antibody developability issues of high viscosity and opalescence that manifest at four orders of magnitude higher concentrations (>100 mg/mL). CS-SINS enables large-scale, high-throughput selection of developable antibodies during early discovery.

Characterization of Opalescence in low Volume Monoclonal Antibody Solutions Enabled by Microscale Nephelometry.

Monoclonal antibody (mAb)-based drugs are often prone to unfavorable solution behaviors including high viscosity, opalescence, phase separation, and aggregation at the high concentrations needed to enable patient-centric subcutaneous dosage forms. Given that these can have a detrimental impact on manufacturability, stability, and delivery, approaches to identifying, monitoring, and controlling these behaviors during drug development are critical. Opalescence presents a significant challenge due to its relationship to liquid-liquid phase separation. Quantitative characterization of opalescence via turbidimetry is often restrictive due to large volume requirements (>2 mL) and alternative microscale approaches based on light transmittance (Eckhardt et al., J Pharm Sci Technol. 1994, 48: 64-70) may pose challenging with respect to accuracy. To address the need for accurate and quantitative microscale opalescence measurements, we have evaluated the use of a 'de-tuned' static light scattering detector which requires <10 µL sample per measurement. We show that tuning of the laser power to a range far below that of traditional light scattering measurements results in a stable detector response that can be accurately calibrated to the nephelometric turbidity unit (NTU) scale using appropriate standards. The calibrated detector signal yields NTU values for mAbs and other protein solutions that are comparable to a commercial turbidimeter. We used this microscale approach to characterize the opalescence of 48 commercial mAb drug products and found that the majority have opalescence below 15 NTU. However, in products with mAb concentrations greater than 75 mg/mL, a broad range of opalescence was observed, in a few cases greater than 20 NTU. These measurements as well as nephelometric characterization of several IgG1 and IgG4 mAbs across a broad pH range highlight subclass-specific tendencies toward opalescence in high concentration solutions.

Machine Learning Feature Selection for Predicting High Concentration Therapeutic Antibody Aggregation.

Protein aggregation can hinder the development, safety and efficacy of therapeutic antibody-based drugs. Developing a predictive model that evaluates aggregation behaviors during early stage development is therefore desirable. Machine learning is a widely used tool to train models that predict data with different attributes. However, most machine learning techniques require more data than is typically available in antibody development. In this work, we describe a rational feature selection framework to develop accurate models with a small number of features. We applied this framework to predict aggregation behaviors of 21 approved monospecific monoclonal antibodies at high concentration (150 mg/mL), yielding a correlation coefficient of 0.71 on validation tests with only two features using a linear model. The nearest neighbors and support vector regression models further improved the performance, which have correlation coefficients of 0.86 and 0.80, respectively. This framework can be extended to train other models that predict different physical properties.

A single molecular descriptor to predict solution behavior of therapeutic antibodies.

Despite the therapeutic success of monoclonal antibodies (mAbs), early identification of developable mAb drug candidates with optimal manufacturability, stability, and delivery attributes remains elusive. Poor solution behavior, which manifests as high solution viscosity or opalescence, profoundly affects the developability of mAb drugs. Using a diverse dataset of 59 mAbs, including 43 approved products, and an array of molecular descriptors spanning colloidal, conformational, charge-based, hydrodynamic, and hydrophobic properties, we show that poor solution behavior is prevalent (>30%) in mAbs and is singularly predicted (>90%) by the diffusion interaction parameter (k D), a dilute-solution measure of colloidal self-interaction. No other descriptor, individually or in combination, was found to be as effective as k D. We also show that well-behaved mAbs, a substantial subset of which bear high positive charge and pI, present no disadvantages with respect to pharmacokinetics in humans. Here, we provide a systematic framework with quantitative thresholds for selecting well-behaved therapeutic mAbs during drug discovery.

Impact of cysteine variants on the structure, activity, and stability of recombinant human α-galactosidase A.

Recombinant human α-galactosidase A (rhαGal) is a homodimeric glycoprotein deficient in Fabry disease, a lysosomal storage disorder. In this study, each cysteine residue in rhαGal was replaced with serine to understand the role each cysteine plays in the enzyme structure, function, and stability. Conditioned media from transfected HEK293 cells were assayed for rhαGal expression and enzymatic activity. Activity was only detected in the wild type control and in mutants substituting the free cysteine residues (C90S, C174S, and the C90S/C174S). Cysteine-to-serine substitutions at the other sites lead to the loss of expression and/or activity, consistent with their involvement in the disulfide bonds found in the crystal structure. Purification and further characterization confirmed that the C90S, C174S, and the C90S/C174S mutants are enzymatically active, structurally intact and thermodynamically stable as measured by circular dichroism and thermal denaturation. The purified inactive C142S mutant appeared to have lost part of its alpha-helix secondary structure and had a lower apparent melting temperature. Saturation mutagenesis study on Cys90 and Cys174 resulted in partial loss of activity for Cys174 mutants but multiple mutants at Cys90 with up to 87% higher enzymatic activity (C90T) compared to wild type, suggesting that the two free cysteines play differential roles and that the activity of the enzyme can be modulated by side chain interactions of the free Cys residues. These results enhanced our understanding of rhαGal structure and function, particularly the critical roles that cysteines play in structure, stability, and enzymatic activity.

Improved sensitivity of an acid sphingomyelinase activity assay using a C6:0 sphingomyelin substrate.

Short-chain C6-sphingomyelin is an artificial substrate that was used in an acid sphingomyelinase activity assay for a pilot screening study of patients with Niemann-Pick disease types A and B. Using previously published multiplex and single assay conditions, normal acid sphingomyelinase activity levels (i.e. false negative results) were observed in two sisters with Niemann-Pick B who were compound heterozygotes for two missense mutations, p.C92W and p.P184L, in the SMPD1 gene. Increasing the sodium taurocholate detergent concentration in the assay buffer lowered the activity levels of these two patients into the range observed with other patients with clear separation from normal controls.

Carbohydrate-mediated polyethylene glycol conjugation of TSH improves its pharmacological properties.

Thyrogen (thyrotropin alfa for injection), recombinant human TSH (rhTSH), has been successfully used to enhance diagnostic radioiodine scanning and thyroglobulin testing in the follow-up of patients with thyroid cancer and as an adjunctive treatment for radioiodine thyroid remnant ablation. However, the short half-life of rhTSH in the circulation requires a multidose regimen. We developed novel sialic acid-mediated and galactose-mediated conjugation chemistries for targeting polyethylene glycol (PEG) to the three N-linked glycosylation sites on the protein, to prolong plasma half-life by eliminating kidney filtration and potential carbohydrate-mediated clearance. Conjugates of different PEG sizes and copy numbers were screened for reaction yield, TSH receptor binding, and murine phamacokinetics/pharmacodynamics studies. The best performing of these products, a 40-kDa mono-PEGylated sialic acid-mediated conjugate, exhibited a 3.5-fold longer duration of action than rhTSH in rats, as a 5-fold lower affinity was more than compensated by a 23-fold extension of circulation half-life. Biochemical characterization confirmed conjugation through the sialic acids. Correlation of PEG distribution on the three N-linked glycosylation sites and the PEG effect on receptor binding supported the previously reported structure-function relationship of rhTSH glycosylation. This long-acting rhTSH has the potential to significantly improve patient convenience and provider flexibility while reducing potential side effects associated with a sudden elevation of serum TSH.

Site-specific PEGylation of human thyroid stimulating hormone to prolong duration of action.

Recombinant human thyroid stimulating hormone (rhTSH or Thyrogen) has been approved for thyroid cancer diagnostics and treatment under a multidose regimen due to its short circulating half-life. To reduce dosing frequency, PEGylation strategies were explored to increase the duration of action of rhTSH. Lysine and N-terminal PEGylation resulted in heterogeneous product profiles with 40% or lower reaction yields of monoPEGylated products. Eleven cysteine mutants were designed based on a structure model of the TSH-TSH receptor (TSHR) complex to create unique conjugation sites on both α and ß subunits for site-specific conjugation. Sequential screening of mutant expression level, oligomerization tendency, and conjugation efficiency resulted in the identification of the αG22C rhTSH mutant for stable expression and scale-up PEGylation. The introduced cysteine in the αG22C rhTSH mutant was partially blocked when isolated from conditioned media and could only be effectively PEGylated after mild reduction with cysteine. This produced a higher reaction yield, ~85%, for the monoPEGylated product. Although the mutation had no effect on receptor binding, PEGylation of αG22C rhTSH led to a PEG size-dependent decrease in receptor binding. Nevertheless, the 40 kDa PEG αG22C rhTSH showed a prolonged duration of action compared to rhTSH in a rat pharmacodynamics model. Reverse-phase HPLC and N-terminal sequencing experiments confirmed site-specific modification at the engineered Cys 22 position on the α-subunit. This work is another demonstration of successful PEGylation of a cysteine-knot protein by an engineered cysteine mutation.

Detection of high-molecular-weight amyloid serum protein complexes using biological on-line tracer sedimentation.

The systemic amyloidoses are a rare but deadly class of protein folding disorders with significant unmet diagnostic and therapeutic needs. The current model for symptomatic amyloid progression includes a causative role for soluble toxic aggregates as well as for the fibrillar tissue deposits. Although much research is focused on elucidating the potential mechanism of aggregate toxicity, evidence to support their existence in vivo has been limited. We report the use of a technique we have termed biological on-line tracer sedimentation (BOLTS) to detect abnormal high-molecular-weight complexes (HMWCs) in serum samples from individuals with systemic amyloidosis due to aggregation and deposition of wild-type transthyretin (senile systemic amyloidosis, SSA) or monoclonal immunoglobulin light chain (AL amyloidosis). In this proof-of-concept study, HMWCs were observed in 31 of 77 amyloid samples (40.3%). HMWCs were not detected in any of the 17 nonamyloid control samples subjected to BOLTS analyses. These findings support the existence of potentially toxic amyloid aggregates and suggest that BOLTS may be a useful analytic and diagnostic platform in the study of the amyloidoses or other diseases where abnormal molecular complexes are formed in serum.

Fluorescence-detected sedimentation in dilute and highly concentrated solutions.

Analytical ultracentrifugation (AUC) is a powerful, first-principles method for characterizing macromolecules in solution. The recent development of fluorescence-detected sedimentation for the AUC (AU-FDS) has extended the sensitivity and selectivity of the instrument which, in turn, has enabled the study of both higher affinity interactions and the sedimentation of one component in complex, concentrated solutions. While still in its infancy, AU-FDS is becoming more widespread as shown by the increasing number of literature reports citing its use. While AU-FDS enables the analysis of systems not amenable to absorbance or interferometric detection, its use is not without limitations. In most cases, preparing samples for AU-FDS analyses requires chemical conjugation with fluorescent dyes, a step that may influence the size or shape of a molecule sufficiently to alter its transport during sedimentation. Careful preparation and characterization of the amount of free dye and the degree and site specificity of labeling is required for robust interpretation of AU-FDS data. In some cases, studies of the effect of labeling on the structure, activity, or association properties of the macromolecule may be warranted. However, these complications are of minor consequence compared to the unique information that can be obtained by AU-FDS. In particular, its ability to provide direct, physical characterization of the thermodynamic behavior of molecules in complex and concentrated solutions makes AU-FDS a powerful technology for understanding the physical underpinnings of living systems.

The modulation of transthyretin tetramer stability by cysteine 10 adducts and the drug diflunisal. Direct analysis by fluorescence-detected analytical ultracentrifugation.

Transthyretin (TTR) is normally a stable plasma protein. However, in cases of familial TTR-related amyloidosis and senile systemic amyloidosis (SSA), TTR is deposited as amyloid fibrils, leading to organ dysfunction and possibly death. The mechanism by which TTR undergoes the transition from stable, soluble precursor to insoluble amyloid fibril and the factors that promote this process are largely undetermined. Most models involve the dissociation of the native TTR tetramer as the initial step. It is largely accepted that the TTR gene mutations associated with TTR-related amyloidosis lead to the expression of variant proteins that are intrinsically unstable and prone to aggregation. It has been suggested that amyloidogenicity may be conferred to wild-type TTR (the form deposited in SSA) by chemical modification of the lone cysteine residue (Cys(10)) through mixed disulfide bonds. S-Sulfonation and S-cysteinylation are prevalent TTR modifications physiologically, and studies have suggested their ability to modulate the structure of TTR under denaturing conditions. In the present study, we have used fluorescence-detected sedimentation velocity to determine the effect of S-sulfonate and S-cysteine on the quaternary structural stability of fluorophore-conjugated recombinant TTR under nondenaturing conditions. We determined that S-sulfonation stabilized TTR tetramer stability by a factor of 7, whereas S-cysteinylation enhanced dissociation by 2-fold with respect to the unmodified form. In addition, we report the direct observation of tetramer stabilization by the potential therapeutic compound diflunisal. Finally, as proof of concept, we report the sedimentation of TTR in serum and the qualitative assessment of the resulting data.

Expression, purification, and in vitro cysteine-10 modification of native sequence recombinant human transthyretin.

Transthyretin (TTR) is a serum protein that is also a prominent component of deposits in two different types of systemic amyloid disease, senile systemic and familial TTR amyloidoses. Studies of recombinant TTR (rTTR) have provided many insights into the relationship between protein structure and amyloidogenicity. Yet, there is no existing recombinant system that results in high yield production of a protein that is identical in primary structure to human TTR. To date, most published studies have generated rTTR using the human gene sequence, which is poorly expressed in Escherichia coli. In addition, the gene sequence has been flanked by a 3' AUG start codon to initiate translation, resulting in the expression of a protein containing an N-terminal methionine residue not present in the human protein. We present an improved technique which can be used to generate large quantities of human native sequence TTR. Our recombinant system utilizes a gene containing codons altered for efficient expression in E. coli and an N-terminal polyhistidine tag for simplified purification. Optimization of this system was accomplished by generating a modified polyhistidine tag that was efficiently removed by dipeptidyl aminopeptidase I (DAPase). This is the first report detailing an effective and useful method for producing rTTR containing an amino acid sequence identical to human TTR. Furthermore, we describe the thiol modification of the recombinant protein to achieve exact replication of the several prominent post-translationally modified forms of TTR that have been identified in human serum.

Detailed structural analysis of amyloidogenic wild-type transthyretin using a novel purification strategy and mass spectrometry.

Wild-type transthyretin (TTR), normally a soluble plasma-circulating protein, can be amyloidogenic, i.e., form tissue-deposited fibrillar material in the extracellular matrix of various organs throughout the body. Senile systemic amyloidosis (SSA) is one such pathology and features TTR-containing amyloid deposits that are found primarily in the heart. The cause for this transition from soluble to insoluble protein in SSA is yet to be determined as specific structural features that might favor TTR fibrillogenesis have not yet been identified. The precise characterization of ex vivo fibril deposits might provide insight, but structural analyses of TTR from amyloid deposits have been hindered thus far by the lack of purification strategies that overcome the insolubility of the tissue-derived protein without degrading it. Consequently, the true biochemical nature of deposited TTR remains in question. In this study, we provide detailed analyses of both the soluble (serum) and deposited (tissue) forms of TTR from cases of SSA. In the serum, a distribution of mixed disulfides, specifically S-sulfonated and S-cysteinylated forms of TTR, as well as the unmodified protein were identified. The relative levels of the three TTR species in the SSA group were comparable to amounts present in sera from age-matched control groups. For characterization of the amyloid deposited TTR, we investigated cardiac tissue samples obtained from three separate cases of SSA. We report a novel chromatographic purification strategy performed under nonreducing conditions (to maintain cysteine disulfide status) and the use of this procedure in conjunction with detailed mass spectrometric analysis of TTR from the amyloid deposits. A series of C-terminal TTR fragments with N-termini ranging from amino acids 46 to 55 were identified. We also determined that the deposits in all samples contained Cys10 disulfide-linkedhomodimers composed of full-length TTR monomers. This last finding suggests an important role for Cys10 conjugation in the transition from soluble TTR to the pathological amyloid fibril.

Valence and anion binding of bovine ribonuclease A between pH 6 and 8.

Several studies have shown that divalent anion binding to ribonuclease A (RNase A) contributes to RNase A folding and stability. However, there are conflicting reports about whether chloride binds to or stabilizes RNase A. Two broad-zone experimental approaches, membrane-confined electrophoresis and analytical ultracentrifugation, were used to examine the electrostatic and electrohydrodynamic characteristics of aqueous solutions of bovine RNase A in the presence of 100 mM KCl and 10 mM Bis-Tris propane over a pH range of 6.00-8.00. The results of data analysis using a Debye-Huckel-Henry model, compared with expectations based on pK(A) values, are consistent with the binding of two chlorides by RNase A. The decreased protein valence resulting from anion binding contributes 2-3 kJ/mol to protein stabilization. This work demonstrates the utility of first-principle valence determinations to detect protein solution properties that might otherwise remain undetected.