AUTHOR=Wen Xianhui , Cui Miaomiao , Zhang Junhua , Huang Hai TITLE=Liquid-liquid phase separation in gastric cancer: identifying novel biomarkers and therapeutic targets through gene signature analysis JOURNAL=Frontiers in Immunology VOLUME=Volume 16 - 2025 YEAR=2025 URL=https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2025.1620390 DOI=10.3389/fimmu.2025.1620390 ISSN=1664-3224 ABSTRACT=Background and objectiveLiquid-liquid phase separation (LLPS) plays an important role in the development of many tumors, including gastric cancer, but its prognostic value is unclear. The aim of this study was to explore the prognostic significance of LLPS-related genes in gastric cancer to provide a basis for improving the accuracy of prognostic prediction and finding potential therapeutic targets in gastric cancer.MethodsClinical and transcriptomic data of gastric cancer were downloaded from TCGA and GEO databases, and LLPS-related genes were extracted from PhaSepDB. Unsupervised clustering was used to identify molecular subtypes based on LLPS gene expression. LLPS gene features were constructed and validated by LASSO Cox regression, and their staging prediction value was also evaluated by machine learning methods. Key genes were validated by qRT-PCR, Western blot, immunofluorescence, and functional experiments (shRNA knockdown, CCK-8, clone formation, and scratch assay).ResultsTwenty LLPS-associated genes showed significant mRNA expression, copy number variation, somatic mutation, and interaction network alterations in gastric cancer tissues. Two LLPS molecular isoforms with different survival outcomes and immune microenvironment characteristics were identified. A four-gene LLPS prognostic signature consisting of DACT1, EZH2, PAK2, and PSPC1 was constructed, and the high-risk group had a poorer prognosis and was prone to drug resistance. Machine learning analysis further confirmed the predictive value of this gene signature. Functional experiments showed that knockdown of PSPC1 significantly inhibited the proliferation (inhibition rate >50%, P <0.001) and migration ability (P<0.0001) of gastric cancer cells. Immunofluorescence confirmed the localization and aggregation characteristics of DACT1 and PSPC1.ConclusionThis study revealed the important role of LLPS in gastric cancer, and the constructed four-gene LLPS signature is expected to be a novel biomarker for prognostic assessment and treatment of gastric cancer. PSPC1 plays a key role in gastric cancer progression, and has the value of a potential therapeutic target.