Highland games: A benchmarking exercise in predicting biophysical and drug properties of monoclonal antibodies from amino acid sequences.

Biopharmaceutical product and process development do not yet take advantage of predictive computational modeling to nearly the degree seen in industries based on smaller molecules. To assess and advance progress in this area, spirited coopetition (mutually beneficial collaboration between competitors) was successfully used to motivate industrial scientists to develop, share, and compare data and methods which would normally have remained confidential. The first "Highland Games" competition was held in conjunction with the October 2018 Recovery of Biological Products Conference in Ashville, NC, with the goal of benchmarking and assessment of the ability to predict development-related properties of six antibodies from their amino acid sequences alone. Predictions included purification-influencing properties such as isoelectric point and protein A elution pH, and biophysical properties such as stability and viscosity at very high concentrations. Essential contributions were made by a large variety of individuals, including companies which consented to provide antibody amino acid sequences and test materials, volunteers who undertook the preparation and experimental characterization of these materials, and prediction teams who attempted to predict antibody properties from sequence alone. Best practices were identified and shared, and areas in which the community excels at making predictions were identified, as well as areas presenting opportunities for considerable improvement. Predictions of isoelectric point and protein A elution pH were especially good with all-prediction average errors of 0.2 and 1.6 pH unit, respectively, while predictions of some other properties were notably less good. This manuscript presents the events, methods, and results of the competition, and can serve as a tutorial and as a reference for in-house benchmarking by others. Organizations vary in their policies concerning disclosure of methods, but most managements were very cooperative with the Highland Games exercise, and considerable insight into common and best practices is available from the contributed methods. The accumulated data set will serve as a benchmarking tool for further development of in silico prediction tools.

Virus removal robustness of ion exchange chromatography.

Virus removal by ion exchange chromatography enhances the safety profile of therapeutic protein products. The robustness of virus removal depends on electrostatic binding between virus and oppositely charged chromatography media. However, model retrovirus Xenotropic Murine Leukemia Virus (XMuLV) binding remains robust even when virus and media are both positively charged. We investigated this counter-intuitive phenomenon using side-by-side comparison of virus-media binding behavior of XMuLV versus parvovirus, two viruses very different in size and structure but comparable in isoelectric point. When both viruses were negatively charged, XMuLV bound to positive anion exchange media with higher strength than parvovirus. When both viruses were positively charged, XMuLV remained tightly bound to positive media but parvovirus was dissociated. Likewise, XMuLV binding to media was much stronger than parvovirus under cation exchange conditions. These findings suggest that XMuLV binding could be enhanced by localized charge distribution not possessed by parvovirus, which is an important consideration for designing chromatography processes with robust virus removal capacity.

Double-peak elution profile of a monoclonal antibody in cation exchange chromatography is caused by histidine-protonation-based charge variants.

We have systemically investigated unusual elution behaviors of an IgG4 (mAb A) in cation exchange chromatography (CEX). This mAb A exhibited two elution peaks under certain conditions when being purified by several strong CEX columns. When either of the two peaks was isolated and re-injected on the same column, the similar pattern was observed again during elution. The protein distribution between the two peaks could be altered by NaCl concentration in the feed, or NaCl concentration in wash buffer, or elution pH, suggesting two pH-associated strong-and-weak binding configurations. The protein distributions under different pH values showed good correlation with protonated/un-protonated fractions of a histidine residue. These results suggest that the double-peak elution profile associates with histidine-protonation-based charge variants. By conducting pepsin digestion, amino-acid specific chemical modifications, peptide mapping, and measuring the effects of elution residence time, a histidine in the variable fragment (Fab) was identified to be the root cause. Besides double-peak pattern, mAb A can also exhibit peak-shouldering or single elution peak on different CEX resins, reflecting different resins' resolving capability on protonated/un-protonated forms. This work characterizes a novel cause for unusual elution behaviors in CEX and also provides alternative avenues of purification development for mAbs with similar behaviors.

Trace levels of the CHO host cell protease cathepsin D caused particle formation in a monoclonal antibody product.

Chinese hamster ovary (CHO) cells are often used to produce therapeutic monoclonal antibodies (mAbs). CHO cells express many host cell proteins (HCPs) required for their growth. Interactions of HCPs with mAbs can sometimes result in co-purification of trace levels of 'hitchhiker' HCPs during the manufacturing process. Purified mAb-1 product produced in early stages of process optimization had high HCP levels. In addition, these lots formed delayed-onset particles containing mAb-1 and its heavy chain C-terminal fragments. Studies were performed to determine the cause of the observed particle formation and to optimize the purification for improved HCP clearance. Protease activity and inhibitor stability studies confirmed that an aspartyl protease was responsible for fragmentation of mAb-1 resulting in particle formation. An affinity resin was used to selectively capture aspartyl proteases from the mAb-1 product. Mass spectrometry identified the captured aspartyl protease as CHO cathepsin D. A wash step at high pH with salt and caprylate was implemented during the protein A affinity step to disrupt the HCP-mAb interactions and improve HCP clearance. The product at the end of purification using the optimized process had very low HCP levels, did not contain detectable protease activity, and did not form particles. Spiking of CHO cathepsin D back into mAb-1 product from the optimized process confirmed that it was the cause of the particle formation. This work demonstrated that process optimization focused on removal of HCPs was successful in eliminating particle formation in the final mAb-1 product.