Related Experiment Video
Updated: May 6, 2026
Intramucosal Inoculation of Squamous Cell Carcinoma Cells in Mice for Tumor Immune Profiling and Treatment Response Assessment
Published on: April 22, 2019
Construction of a Vasculogenic Mimicry-related RiskScore Model to Assess Patient Prognosis and Immunotherapy Response
Yujia Zhai1, Xinyu Gu2, Di Huang3
1Department of Oral Medicine, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, 450052, China.
Introduction:
Vasculogenic mimicry (VM) plays vital roles in tumor development that are closely relevant to patient adverse outcomes and chemoresistance. This study aimed to identify a novel VM-associated signature to forecast the prognosis and immunotherapy response of head and neck squamous cell carcinoma (HNSC) patients.
Materials And Methods:
HNSC samples were derived from TCGA and GEO databases. VM-related genes (VMGs) were acquired from previous literature. VMGs' score was estimated using the "GSVA" package. The critical gene module was recognized by the "WGCNA" package. Differentially expressed genes (DEGs) were determined utilizing the "limma" package, and functional enrichment analysis was conducted by the "clusterProfiler" package. Hereafter, employing univariate Cox, lasso Cox, and multivariate stepwise regression analysis, the independent prognostic VMGs were identified to develop the RiskScore model and verify their predictive performance. Moreover, the immune infiltration, immunotherapy response, and drug sensitivity were analyzed. Finally, the expression of the selected key genes was evaluated in vitro using qRT-PCR in HNSC lines.
Results:
The tumor group showed a higher VMGs score than the normal group. 590 key module genes were recognized by WGCNA, and then intersected with 6160 DEGs to obtain 293 candidate genes that were mainly involved in the PI3K-Akt and extracellular matrix (ECM)-relevant pathways. Thereafter, 9 independent prognostic VMGs (CHSY1, TNFAIP6, PRELP, HTRA1, RNF144A, COL8A2, DCHS1, FMOD, NOSTRIN) in HNSC were identified and selected to establish a RiskScore model, with good robustness in predicting patient outcomes. Compared with the low-risk group, the high-risk group showed an adverse prognosis, lower immune infiltration, and worse immunotherapy response. Besides, RiskScore was negatively correlated with several chemotherapeutic drugs such as FTI-277, Obatoclax Mesylate, Embelin, etc. Experimental validation using qRT‑PCR confirmed that most of the signature genes (including CHSY1, TNFAIP6, HTRA1, COL8A2, FMOD) were significantly upregulated in an oral squamous cell carcinoma line compared to normal keratinocytes.
Discussion:
Our present study established a novel 9-VMGs RiskScore to predict the prognosis, immune infiltration features, immunotherapy response, and drug sensitivity for HNSC patients. This study provided a theoretical foundation for further exploration of HNSC pathogenesis, contributing to personalized treatment and drug selection for HNSC.
Conclusion:
The VM-related RiskScore we developed serves as a reliable prognostic and predictive tool, providing valuable guidance for risk stratification and immunotherapy decision-making in HNSC patients.
