Chinese hamster ovary cell-specific biopharmaceutical glycoform predictions through discretized reaction network modeling
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Chinese hamster ovary (CHO) cells produce more biopharmaceuticals than any other cell line due to the CHO cells’ many advantageous characteristics, including their ability to glycosylate biopharmaceuticals with human-compatible glycans. The biopharmaceutical glycoform affects the product efficacy, half-life, and immunogenicity; therefore, biopharmaceuticals must have a consistent glycoform to ensure therapeutic efficacy and patient safety. One challenge associated with ensuring consistent glycosylation is that the CHO-specific glycosylation reaction network is a non-template driven cellular process with many variables, making predictive glycoform modeling difficult. This research addresses this challenge through the design of a computational glycosylation tool using a novel modeling technique that predicts biopharmaceutical glycoforms and thereby generates experimentally-relevant CHO cell line-specific information to improve biopharmaceutical manufacturing. ☐ Achieving this goal requires a fundamental understanding of CHO cellular biology, processes, and reaction networks. The recently sequenced and annotated CHO genome facilitates a detailed, mechanistic understanding of CHO cell-specific biology. Recent and emerging genome sequencing technologies were characterized, the differences between the technologies were highlighted, and the reported CHO biopharmaceutical applications of these sequencing technologies were examined with a focus on the sequencing and annotation of the CHO-K1 cell and Chinese hamster (CH) genomes. We improved CHOgenome.org, the centralized CHO community’s public database repository through the addition of the CHO and CH genomes to the website databases and through the creation of additional genomic and proteomic bioinformatics tools. The reported CHO research community’s use of these bioinformatics tools was also described. ☐ Controllability of the biopharmaceutical product quality is essential for an approved biotherapeutic and predicting the results of cell-engineered modifications on the glycoform could aid future product quality control methods. This work details the use of a novel Discretized Reaction Network Modeling using Fuz zy Parameters (DReaM-zyP) modeling technique that was then used to create Glyco-Mapper, an innovative systems biology glycosylation prediction tool. The Glyco-Mapper input variables consist of glycosylation gene parameters and the media’s nutrient composition, enabling Glyco-Mapper to replicate cell line-specific reference glycoforms and predict the glycoform changes resulting from various cell engineering modifications. The modifications Glyco-Mapper has successfully predicted include the altered expression of glycosylation, nucleotide sugar transport, and metabolism genes, as well as modified nutrient feeding strategies. Glyco-Mapper’s ability to replicate cell line-specific reference glycoforms and accurately predict the reference-specific engineered glycoforms provides a streamlined tool to design cell lines to control specific product quality attributes. ☐ In this work, CHO-produced biopharmaceutical glycoforms from literature were used to validate the Glyco-Mapper’s predictive glycoform output. Glyco-Mapper predicted the engineered glycoforms with an accuracy, sensitivity, specificity, and predictive accuracy of the glycans changing experimental measurements as a result of the engineering strategy of 96%, 85%, 97%, and 85%, respectively. A non-mAb model biopharmaceutical reference glycoform was replicated using Glyco-Mapper, a novel gene knockdown (GnT-II) was predicted, and the predicted glycoform was experimentally confirmed with an accuracy and specificity of 95% and 98%, respectively. Additional glycoform predictions are presented and continued investigation of these predictions is recommended to further validate and improve the Glyco-Mapper tool. Glyco-Mapper is a novel CHO-specific glycosylation tool that predicts biopharmaceutically-relevant glycoforms and generates CHO cell line-specific information that can be used to improve biopharmaceutical manufacturing through enhanced product quality control.