基于定量构效分析的生物分离研究新方法及抗体药物分离

Based on the complexity of bio-macromolecules and the limited understanding on separation rules, with the chromatographic separation as the model, new analysis method of quantitative structure-activity relationship (QSAR) would be introduced to study the relations between protein structure, resin properties, separation conditions and the separation performance. The understanding on the essence of protein chromatographic separation would be enhances, and the predictability of process development would be improved. Mixed-mode chromatography will be focused as new technology and ion exchange chromatography and hydrophobic interaction chromatography as the control. Firstly, a library of model proteins will be set. A series of adsorption and separation experiments will be carried out, and the systematic data would be obtained accurately to describe the separation performance. Secondly, with the aid of molecular simulation, new multi-dimensional descriptors would be constructed to characterize the interactions between proteins and resins in a better way. After the screening with statistical analysis, a QSAR correlation between key descriptors and the separation parameters would be obtained. Finally, the QSAR analysis will be used to develop new processes for antibody purification. The key factors affecting the specific separation performance would be identified, which will be helpful to design new ligands and optimize the separation conditions. New process will be developed to separation efficiently the structural similar components of antibodies. In this project new methods based on QSAR analysis would be established to predict rationally the separation performance from the properties of proteins and resins, which would certainly improve the ligand design and the optimization of downstream process. The methods developed are universal and can be extended to other separation processes of pharmaceutical proteins.

针对生物大分子的复杂性和分离规律认识的有限性，以层析分离为代表，引入定量构效关系（QSAR）分析，探讨蛋白结构、介质性质和分离条件三者与分离性能的相关性，强化分离本质的认识，提高分离过程的可预见性。以新型混合模式层析为重点，离子交换层析和疏水相互作用层析为对照，一方面，构建模型蛋白库，开展系列吸附和分离实验，获得系统的分离性能数据；另一方面，借助分子模拟，构建新型多维描述符，合理表征蛋白—介质相互作用，优化统计分析，实现关键描述符—宏观分离性能的定量关联。在此基础上，针对抗体药物分离，基于QSAR分析，预测特定分离性能的核心影响因素，设计新型分离介质，优化分离过程，实现结构类似组分的高效分离。本项目将建立基于QSAR分析的生物分离研究方法，实现性质到性能的合理预测，为新型介质研发和分离过程优化提供指导，促进理性的定向设计。所建立的方法具有普适性，可推广到其它药用蛋白的分离纯化过程。

针对抗体等蛋白药物的高效分离，集中于定量构效关系（QSAR）分析和分离新方法，一方面从蛋白吸附性能表征入手，筛选模型蛋白库，测定介质吸附性能，鉴定特定蛋白对象；另一方面从QSAR分析入手，考察关联算法，构建分析平台，探讨蛋白-配基相互作用；在此基础上，设计新型分离配基，探讨分离新方法，优化分离过程，提高分离效率。主要成果包括三个方面：第一，蛋白吸附特性和单抗电荷异质体分离鉴定：选取了20种分子结构、等电点和大小各异的模型蛋白，测定了多种混合模式介质的吸附性能；针对抗体不同种类、亚型和片段，系统考察了不同pH和盐浓度的影响；针对抗体电荷异质体，实现了高效分离和结构鉴定，为QSAR分析提供了数据基础。第二，QSAR分析和配基-蛋白相互作用：分别针对20种模型蛋白、抗体及片段、抗体电荷异质体，筛选了蛋白和配基的特征分子描述符，形成了2D和3D等多种维度的分子描述；考察了支持向量机、随机森林等多种关联算法，构建了QSAR分析平台，并引入机器学习新方法，实现了蛋白/配基分子特性和分离行为的有效关联。第三，新型介质研发和蛋白分离新方法：针对抗体和重组人血白蛋白，基于蛋白的分子特征，筛选和优化设计了新型混合模式配基，提高了分离效率，单步分离纯度达到90%以上；针对抗体电荷异质体分离，探讨了连续流层析分离新方法，实现了微量组分的高效富集；建立了层析模型表征，辅助分离过程设计，并形成了集成软件包，显著提高了过程开发效率。本项目理论联系实际，形成“基于定量构效分析”的生物分离研究新方法和定向设计的新思路，对多功能配基设计、新型介质研发、抗体等重组蛋白药物分离纯化等具有很好的指导和推广作用。