Model-based analysis of N-glycosylation in Chinese hamster ovary cells.
Model-based analysis of N-glycosylation in Chinese hamster ovary cells.
复制标题
DOI：
10.1371/journal.pone.0175376
复制
发表时间：
2017
期刊：
影响因子：
3.7
通讯作者：
Betenbaugh MJ
中科院分区：
文献类型：
--
作者：
Krambeck FJ;Bennun SV;Andersen MR;Betenbaugh MJ
中国仓鼠卵巢（CHO）细胞是生产用于生物治疗药物的糖基化重组蛋白的金标准。其糖基化模式与人类版本的相似性使该细胞系的产物具有良好的药代动力学特性和较低的引起免疫原性反应的可能性。由于多糖结构是细胞内酶协同作用的产物，因此很难预先预测遗传操作的影响如何改变细胞的多糖结构和治疗特性。因此，能够预测糖基化的定量模型已经成为处理糖基化过程复杂性的有前途的工具。例如，本研究中使用的相同模型的早期版本被其他人成功地用于预测酶活性的变化，这些变化可能会产生期望的聚糖结构变化。在这项研究中，我们利用该模型的更新版本，通过对先前发表的质谱数据的解释，对10个中国仓鼠卵巢（CHO）细胞系（包括一个野生型亲本和9个CHO突变体）的n -糖基化进行了全面分析。更新的n -糖基化数学模型包含多达50,605个聚糖结构。调整该模型中的酶活性以匹配n -聚糖质谱，产生糖基化过程的详细预测，酶活性谱和每个细胞系的完整糖基化谱。这些特征与之前报道的生物化学和遗传数据一致。基于模型的结果还预测了以前未发表的细胞系的糖基化特征，表明糖基化酶活性的变化比直接由基因突变引起的变化更复杂。该模型预测，CHO细胞系具有调节机制，允许它们调节糖基化酶活性，以减轻已知存在于这些突变细胞系中糖基化功能的主要丧失或获得的副作用。CHO细胞糖基化的定量模型具有预测糖工程操作如何影响糖型分布以提高糖蛋白产物治疗性能的潜力。
The Chinese hamster ovary (CHO) cell is the gold standard for manufacturing of glycosylated recombinant proteins for production of biotherapeutics. The similarity of its glycosylation patterns to the human versions enable the products of this cell line favorable pharmacokinetic properties and lower likelihood of causing immunogenic responses. Because glycan structures are the product of the concerted action of intracellular enzymes, it is difficult to predict a priori how the effects of genetic manipulations alter glycan structures of cells and therapeutic properties. For that reason, quantitative models able to predict glycosylation have emerged as promising tools to deal with the complexity of glycosylation processing. For example, an earlier version of the same model used in this study was used by others to successfully predict changes in enzyme activities that could produce a desired change in glycan structure. In this study we utilize an updated version of this model to provide a comprehensive analysis of N-glycosylation in ten Chinese hamster ovary (CHO) cell lines that include a wild type parent and nine mutants of CHO, through interpretation of previously published mass spectrometry data. The updated N-glycosylation mathematical model contains up to 50,605 glycan structures. Adjusting the enzyme activities in this model to match N-glycan mass spectra produces detailed predictions of the glycosylation process, enzyme activity profiles and complete glycosylation profiles of each of the cell lines. These profiles are consistent with biochemical and genetic data reported previously. The model-based results also predict glycosylation features of the cell lines not previously published, indicating more complex changes in glycosylation enzyme activities than just those resulting directly from gene mutations. The model predicts that the CHO cell lines possess regulatory mechanisms that allow them to adjust glycosylation enzyme activities to mitigate side effects of the primary loss or gain of glycosylation function known to exist in these mutant cell lines. Quantitative models of CHO cell glycosylation have the potential for predicting how glycoengineering manipulations might affect glycoform distributions to improve the therapeutic performance of glycoprotein products.