Predicting Formulation Conditions During Ultrafiltration and Dilution to Drug Substance Using a Donnan Model with Homology-Model Based Protein Charge.

In the manufacturing of therapeutic monoclonal antibodies (mAbs), the final steps of the purification process are typically ultrafiltration/diafiltration (UF/DF), dilution, and conditioning. These steps are developed such that the final drug substance (DS) is formulated to the desired mAb, buffer, and excipient concentrations. To develop these processes, process and formulation development scientists often perform experiments to account for the Gibbs-Donnan and volume-exclusion effects during UF/DF, which affect the output pH and buffer concentration of the UF/DF process. This work describes the development of an in silico model for predicting the DS pH and buffer concentration after accounting for the Gibbs-Donnan and volume-exclusion effects during the UF/DF operation and the subsequent dilution and conditioning steps. The model was validated using statistical analysis to compare model predictions against experimental results for nine molecules of varying protein concentrations and formulations. In addition, our results showed that the structure-based in silico approach used to calculate the protein charge was more accurate than a sequence-based approach. Finally, we used the model to gain fundamental insights about the Gibbs-Donnan effect by highlighting the role of the protein charge concentration (the protein concentration multiplied with protein charge at the formulation pH) on the Gibbs-Donnan effect. Overall, this work demonstrates that the Gibbs-Donnan and volume-exclusions effects can be predicted using an in silico model, potentially alleviating the need for experiments.

Rational design of peptide affinity ligands for the purification of therapeutic enzymes.

Non-mAb biologics represent a growing class of therapeutics under clinical development. Although affinity chromatography is a potentially attractive approach for purification, the development of platform technologies, such as Protein A for mAbs, has been challenging due to the inherent chemical and structural diversity of these molecules. Here, we present our studies on the rapid development of peptide affinity ligands for the purification of biologics using a prototypical enzyme therapeutic in clinical use. Employing a suite of de novo rational and combinatorial design strategies we designed and screened a library of peptides on microarray platforms for their ability to bind to the target with high affinity and selectivity in cell culture fluid. Lead peptides were evaluated on resin in batch conditions and compared with a commercially available resin to evaluate their efficacy. Two lead candidates identified from microarray studies provided high binding capacity to the target while demonstrating high selectivity against culture contaminants and product variants compared to a commercial resin system. These findings provide a proof-of-concept for developing affinity peptide-based bioseparations processes for a target biologic. Peptide affinity ligand design and screening approaches presented in this work can also be easily translated to other biologics of interest. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 34:987-998, 2018.

Competitive adsorption of labeled and native protein in confocal laser scanning microscopy.

Confocal laser scanning microscopy has been previously applied to the study of protein uptake in porous chromatography resins. This method requires labeling the protein with a fluorescent probe. The labeled protein is then diluted with a large quantity of native protein so that the fluorescence intensity is a linear function of the labeled protein concentration. Ideally, the attachment of a fluorescent probe should not affect the affinity of the protein for the stationary phase; however, recent experimental work has shown that this assumption is difficult to satisfy. In the present study, we present a mathematical model of protein diffusion and adsorption in a single adsorbent particle. The differences in adsorption behavior of labeled and native protein are accounted for by treating the system as a two-component system (labeled and native protein) described by the steric mass action isotherm (SMA). SMA parameters are regressed from experimental linear gradient elution data for lysozyme and lysozyme-dye conjugates (for the fluorescent dyes Cy3, Cy5, Bodipy FL, and Atto635). When the regressed parameters are employed in the model, an overshoot in the labeled lysozyme concentration is predicted for Cy5- and Bodipy-labeled lysozyme, but not for Atto635-labeled lysozyme. The model predictions agree qualitatively well with recent work showing the dependence of the concentration overshoot on the identity of the attached dye and provide further evidence that the overshoot is likely caused by the change of binding characteristics due to the fluorescent label.

Investigation of protein retention and selectivity in HIC systems using quantitative structure retention relationship models.

In the present work, the effect of stationary phase resin chemistry and protein physicochemical properties on protein binding affinity in hydrophobic interaction chromatography (HIC) was investigated using linear gradient chromatography and quantitative structure-retention relationship (QSRR) modeling. Linear gradient experiments were carried out for a set of model proteins on four different HIC resins having different backbone and ligand chemistry. The retention data exhibited significant differences in protein binding affinity, not only across the phenyl and butyl ligand chemistries, but also for the different backbone chemistries found in the Sepharose (cross-linked agarose) and the Toyopearl 650 M (polymethacrylate) series of resins. QSRR models based on a Support Vector Machine (SVM) approach were developed for the linear retention data using molecular descriptors based on protein crystal structure and primary sequence information as well as a set of new hydrophobicity descriptors based on the solvent accessible protein surface area. The results indicate that the QSRR models were successfully able to capture and selectivity predict the changes observed in these systems. Furthermore, the new descriptors resulted in physically interpretable models of protein retention and provided insights into the factors influencing protein affinity in these different HIC systems. The approach put forth in this study provides a framework for developing predictive tools and for gaining insight into protein selectivity in hydrophobic interaction chromatography.

Multidimensional high-throughput screening of displacers.

A multidimensional, batch high-throughput screening (MD-HTS) protocol was developed to investigate the effects of various parameters on the selectivity of ion-exchange protein displacement systems. A variety of molecules were screened, and the results were employed to provided insights into the influence of displacer chemistry and concentration, resin chemistry, and mobile-phase salt counterion on the efficacy and selectivity of these nonlinear chromatographic systems. These results open up the possibility of tailoring the selectivity of displacement separations by choosing appropriate combinations of operating conditions using the MD-HTS technique. The screens were also employed for the identification of displacers and conditions for the separation of a challenging protein mixture by selective displacement chromatography. Column displacements were carried out with potential lead compounds identified from the MD-HTS screens, and the results confirmed that selective displacement could indeed be achieved for this model mixture. Furthermore, the results indicated that this approach is particularly useful when the order of elution is not changed, but the inherent selectivity is increased in the presence of the displacer. The results presented in this paper demonstrate the utility of the MD-HTS technique for rapid method development in protein ion-exchange displacement chromatography.

Investigation of DNA-binding properties of an aminoglycoside-polyamine library using quantitative structure-activity relationship (QSAR) models.

We have recently developed a novel multivalent cationic library based on the derivatization of aminoglycosides by linear polyamines. In the current study, we describe the DNA-binding activity of this library. Screening results indicated that several candidates from the library showed high DNA-binding activities with some approaching those of cationic polymers. Quantitative Structure-Activity Relationship (QSAR) models of the screening data were employed to investigate the physicochemical effects governing polyamine-DNA binding. The utility of these models for the a priori prediction of polyamine-DNA-binding affinity was also demonstrated. Molecular descriptors selected in the QSAR modeling indicated that molecular size, basicity, methylene group spacing between amine centers, and hydrogen-bond donor groups of the polyamine ligands were important contributors to their DNA-binding efficacy. The research described in this paper has led to the development of new multivalent ligands with high DNA-binding activity and improved our understanding of structure-activity relationships involved in polyamine-DNA binding. These results have implications for the discovery of novel polyamine ligands for nonviral gene delivery, plasmid DNA purification, and anticancer therapeutics.

A priori prediction of adsorption isotherm parameters and chromatographic behavior in ion-exchange systems.

The a priori prediction of protein adsorption behavior has been a long-standing goal in several fields. In the present work, property-modeling techniques have been used for the prediction of protein adsorption thermodynamics in ion-exchange systems directly from crystal structure. Quantitative structure-property relationship models of protein isotherm parameters and Gibbs free energy changes in ion-exchange systems were generated by using a support vector machine regression technique. The predictive ability of the models was demonstrated for two test-set proteins not included in the model training set. Molecular descriptors selected during model generation were examined to gain insights into the important physicochemical factors influencing stoichiometry, equilibrium, steric effects, and binding affinity in protein ion-exchange systems. The a priori prediction of protein isotherm parameters can have direct implications for various ion-exchange processes. As proof of concept, a multiscale modeling approach was used for predicting the chromatographic separation of a test set of proteins using the isotherm parameters obtained from the quantitative structure-property relationship models. The simulated column separation showed good agreement with the experimental data. The ability to predict chromatographic behavior of proteins directly from their crystal structures may have significant implications for a range of biotechnology processes.

Parallel screening of selective and high-affinity displacers for proteins in ion-exchange systems.

This paper employs a parallel batch screening technique for the identification of both selective and high-affinity displacers for a model binary mixture of proteins in a cation-exchange system. A variety of molecules were screened as possible displacers for the proteins ribonuclease A (RNAseA) and alpha-chymotrypsinogen A (alpha-chyA) on high performance Sepharose SP. The batch screening data for each protein was used to select leads for selective and high-affinity displacers and column experiments were carried out to evaluate the performance of the selected leads. The data from the batch displacements was also employed to generate quantitative structure-efficacy relationship (QSER) models based on a support vector machine regression approach. The resulting models had high correlation coefficients and were able to predict the behaviour of molecules not included in the training set. The descriptors selected in the QSER models for both proteins were examined to provide insights into factors influencing displacer selectivity in ion-exchange systems. The results presented in this paper demonstrate that this parallel batch screening-QSER approach can be employed for the identification of selective and high-affinity displacers for protein mixtures.

Identification of chemically selective displacers using parallel batch screening experiments and quantitative structure efficacy relationship models.

Parallel batch screening experiments were carried out to examine how displacer chemistry and salt counterions affect the selectivity of batch protein displacements in anion exchange chromatographic systems. The results indicate that both salt type and displacer chemistry can have a significant impact on the amount of protein displaced. Importantly, the results indicate that, by changing the displacer, salt counterion, or both, one can induce significant selectivity changes in the relative displacement of two model proteins. This indicates that highly selective separations can be developed in ion exchange systems by the appropriate selection of displacer chemistry and salt counterion. The experimental batch screening data were also used in conjunction with various molecular descriptors to generate quantitative structure efficacy relationship (QSER) models based on a support vector machine feature selection and regression tool. The models resulted in good correlations and successful predictions for an external test set of displacers. A star plot approach was shown to be a powerful tool to aid in the interpretation of the QSER models. These results indicate that this modeling approach can be employed for the a priori prediction of displacer efficacy as well as for providing insight into displacer design and the selection of proper mobile-phase conditions for highly selective separations.