Scholarlink

<p>Geranylgeranyl pyrophosphate synthase (GGPPS) is an attractive target for the treatment of multiple myeloma. However, no GGPPS-targeting agents have advanced to clinical trials, primarily because most known inhibitors contain a diphosphate moiety that affects the pharmacokinetic profiles. Thus, the discovery of novel, diphosphate-free GGPPS inhibitors is necessary. In the present work, we employed a hybrid virtual screening strategy integrating ligand-based deep neural network (DNN) models and structure-based molecular docking to identify new hit compounds. The DNN model was pre-trained on a large CX LogP dataset and subsequently fine-tuned using a curated dataset of 210 GGPPS inhibitors, yielding better performance than a model trained directly on the GGPPS dataset. For the structure-based method, molecular dynamics simulations were performed to refine the original receptor structure, improving the ability of docking scores to distinguish between highly active and weakly active inhibitors. A virtual screen of nearly 1.85&#xa0;million compounds was conducted, and 29 compounds were selected and tested using <i>in vitro</i> RPMI-8226 cells evaluation assays. Two compounds, which lack diphosphate groups, exhibited &gt; 50% inhibition at a concentration of 10 µM. These two diphosphate-free hits can serve as potential inhibitors for the development of novel multiple myeloma therapeutics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Discovery of geranylgeranyl pyrophosphate synthase (GGPPS) inhibitors against multiple myeloma by virtual screening

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • Zijian Qin,
  • Haiyue Wu,
  • Jing Jin,
  • Zhichao Zheng,
  • Changjiang Huang

摘要

Geranylgeranyl pyrophosphate synthase (GGPPS) is an attractive target for the treatment of multiple myeloma. However, no GGPPS-targeting agents have advanced to clinical trials, primarily because most known inhibitors contain a diphosphate moiety that affects the pharmacokinetic profiles. Thus, the discovery of novel, diphosphate-free GGPPS inhibitors is necessary. In the present work, we employed a hybrid virtual screening strategy integrating ligand-based deep neural network (DNN) models and structure-based molecular docking to identify new hit compounds. The DNN model was pre-trained on a large CX LogP dataset and subsequently fine-tuned using a curated dataset of 210 GGPPS inhibitors, yielding better performance than a model trained directly on the GGPPS dataset. For the structure-based method, molecular dynamics simulations were performed to refine the original receptor structure, improving the ability of docking scores to distinguish between highly active and weakly active inhibitors. A virtual screen of nearly 1.85 million compounds was conducted, and 29 compounds were selected and tested using in vitro RPMI-8226 cells evaluation assays. Two compounds, which lack diphosphate groups, exhibited > 50% inhibition at a concentration of 10 µM. These two diphosphate-free hits can serve as potential inhibitors for the development of novel multiple myeloma therapeutics.