Sharma Divya, Arumugam Sivakumar
School of Bioscience and Technology, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
School of Bioscience and Technology, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
Comput Biol Med. 2025 Oct;197(Pt B):111068. doi: 10.1016/j.compbiomed.2025.111068. Epub 2025 Sep 12.
The dysregulation of the Wnt/β-catenin signaling pathway serves as a central driver of Colorectal cancer (CRC) initiation and progression. Tankyrase (TNKS), a PARP family enzyme, facilitates Wnt-driven tumorigenesis by PARylating and destabilizing Axin, thereby promoting β-catenin accumulation. In APC-mutated CRC, TNKS inhibition restores β-catenin degradation, highlighting its potential as a therapeutic target. To address this gap, an integrative QSAR model was constructed to identify novel TNKS inhibitors with favorable pharmacokinetic and therapeutic efficacy. A dataset of 1100 TNKS inhibitors was retrieved from the ChEMBL database and subjected to ligand-based QSAR modeling to predict potent chemical scaffolds based on 2D and 3D structural and physicochemical molecular descriptors. To enhance model reliability, Machine learning (ML) approaches such as feature selection and random forest (RF) classification models were applied. The models were trained, optimized, and rigorously validated using internal (cross-validation) and external test sets, achieving a high predictive performance (ROC-AUC) of 0.98. Virtual screening of prioritized candidates was complemented by molecular docking, dynamic simulation, and principal component analysis to evaluate binding affinity, complex stability, and conformational landscapes. This strategy led to the identification of a potential TNKS inhibitor, Q1 (Olaparib), as a possible repurposed drug against TNKS. Network pharmacology further contextualized TNKS within CRC biology, mapping disease-gene interactions and functional enrichment to uncover TNKS-associated roles in oncogenic pathways. Collectively, these findings underscore the effectiveness of combining machine learning and system biology to accelerate rational drug discovery. Olaparib emerges as a promising candidate for TNKS-targeted therapy, providing a strong computational foundation for experimental validation and future preclinical drug development.
Wnt/β-连环蛋白信号通路的失调是结直肠癌（CRC）起始和进展的核心驱动因素。端锚聚合酶（TNKS）是一种聚（ADP-核糖）聚合酶（PARP）家族酶，通过对轴蛋白进行聚（ADP-核糖）化修饰并使其不稳定，促进Wnt驱动的肿瘤发生，从而促进β-连环蛋白的积累。在APC突变的CRC中，TNKS抑制可恢复β-连环蛋白的降解，突出了其作为治疗靶点的潜力。为了填补这一空白，构建了一个综合定量构效关系（QSAR）模型，以识别具有良好药代动力学和治疗效果的新型TNKS抑制剂。从ChEMBL数据库中检索了1100种TNKS抑制剂的数据集，并基于二维和三维结构以及物理化学分子描述符进行基于配体的QSAR建模，以预测有效的化学支架。为了提高模型的可靠性，应用了机器学习（ML）方法，如特征选择和随机森林（RF）分类模型。使用内部（交叉验证）和外部测试集对模型进行训练、优化和严格验证，实现了0.98的高预测性能（ROC-AUC）。通过分子对接、动态模拟和主成分分析对优先候选物进行虚拟筛选，以评估结合亲和力、复合物稳定性和构象格局。该策略导致鉴定出一种潜在的TNKS抑制剂Q1（奥拉帕利），作为一种可能重新用于TNKS的药物。网络药理学进一步将TNKS置于CRC生物学背景下，绘制疾病-基因相互作用和功能富集图谱，以揭示TNKS在致癌途径中的相关作用。总的来说，这些发现强调了将机器学习和系统生物学相结合以加速合理药物发现的有效性。奥拉帕利成为TNKS靶向治疗的有希望的候选药物，为实验验证和未来临床前药物开发提供了坚实的计算基础。
