Bermudez-Lekerika Paola, Gualdi Francesco, Le Maitre Christine L, Piñero Janet, Oliva Baldomero, Gantenbein Benjamin
Tissue Engineering for Orthopaedics & Mechanobiology, Bone & Joint Program, Department for BioMedical Research (DBMR), Faculty of Medicine, University of Bern, Murtenstrasse 35, Bern CH-3008, Switzerland.
Graduate School for Cellular and Biomedical Sciences (GCB), University of Bern, Mittelstrasse 43, Bern CH-3012, Switzerland.
Comput Struct Biotechnol J. 2025 Apr 14;27:1600-1613. doi: 10.1016/j.csbj.2025.04.015. eCollection 2025.
Protein-protein interaction (PPI) networks provide a static map of functional protein interactions, which when combined with algorithms, can prioritize key protein candidates which experimental studies cannot capture. This study, aimed to construct knowledge-based nucleus pulposus (NP)-specific PPI networks which could be deployed to investigate complex protein interactions in human NP cells and tissues following IL-4 and IL-10 stimulation. NP-specific PPI networks were developed based on mass spectrometry (MS) and secretome datasets from human NP cells. These networks were validated using in vitro and ex vivo experimental data sets. Genes Underlying Inheritance Linked Disorders (GUILD) genome-wide network-based prioritization framework was employed for protein candidate prediction under no treatment baseline and IL-4, IL-10 and IL-1β single or combined stimulating scenarios. These secretome-based in vitro PPI networks were able to reproduce the no-treatment candidate prioritization baseline. Whereby within NP cells from discs isolated due to traumatic injury biglycan was identified whilst in degenerate samples decorin was highlighted. Furthermore, experimentally observed IL-4 pleiotropic behaviour was predicted by IL-1 receptor-like 1 prioritization. PPI network-based IL-4 and IL-10 conditions offered novel insights of potential candidates, including collagen IV and fibroblast growth factor intracellular binding protein (FIBP) as key candidates within IL-4 activation pathways, whereas urocortin 3 and neural growth factor were identified following IL-10 stimulation. Additionally, MS based PPI network propagation offered a more extensive, module-based structure networks with lower edge degree and biological variability. Overall, multiple proteomic experimental approaches are required to successfully validate in-silico prediction models to understand the complex interactions between the plethora of proteins involved in IVD degeneration.
蛋白质-蛋白质相互作用（PPI）网络提供了功能性蛋白质相互作用的静态图谱，与算法相结合时，可对实验研究无法捕捉的关键蛋白质候选物进行优先级排序。本研究旨在构建基于知识的髓核（NP）特异性PPI网络，用于研究白细胞介素-4（IL-4）和白细胞介素-10（IL-10）刺激后人NP细胞和组织中的复杂蛋白质相互作用。基于人NP细胞的质谱（MS）和分泌蛋白质组数据集开发了NP特异性PPI网络。这些网络使用体外和离体实验数据集进行了验证。在无治疗基线以及IL-4、IL-10和IL-1β单一或联合刺激情况下，采用基于全基因组网络的遗传性连锁疾病相关基因（GUILD）优先级框架进行蛋白质候选物预测。这些基于分泌蛋白质组的体外PPI网络能够重现无治疗候选物优先级基线。由此，在因创伤性损伤而分离的椎间盘NP细胞中鉴定出了双糖链蛋白聚糖，而在退变样本中则突出显示了核心蛋白聚糖。此外，通过IL-1受体样1优先级预测了实验观察到的IL-4多效性行为。基于PPI网络的IL-4和IL-10条件提供了潜在候选物的新见解，包括IV型胶原和成纤维细胞生长因子细胞内结合蛋白（FIBP）作为IL-4激活途径中的关键候选物，而在IL-10刺激后鉴定出了尿皮质素3和神经生长因子。此外，基于MS的PPI网络传播提供了更广泛的、基于模块的结构网络，具有更低的边度数和生物学变异性。总体而言，需要多种蛋白质组学实验方法来成功验证计算机预测模型，以了解参与椎间盘退变的大量蛋白质之间的复杂相互作用。
