mRNAsi-related genes can effectively distinguish hepatocellular carcinoma into new molecular subtypes

Canbiao Wang; Shijie Qin; Wanwan Pan; Xuejia Shi; Hanyu Gao; Ping Jin; Xinyi Xia; Fei Ma

doi:10.1016/j.csbj.2022.06.011

mRNAsi-related genes can effectively distinguish hepatocellular carcinoma into new molecular subtypes

Comput Struct Biotechnol J. 2022 Jun 8:20:2928-2941. doi: 10.1016/j.csbj.2022.06.011. eCollection 2022.

Authors

Canbiao Wang¹, Shijie Qin^{1

2}, Wanwan Pan¹, Xuejia Shi¹, Hanyu Gao¹, Ping Jin¹, Xinyi Xia², Fei Ma¹

Affiliations

¹ Laboratory for Comparative Genomics and Bioinformatics & Jiangsu Key Laboratory for Biodiversity and Biotechnology, College of Life Science, Nanjing Normal University, Nanjing 210046, China.
² Institute of Laboratory Medicine, Jinling Hospital, Nanjing University School of Medicine, the First School of Clinical Medicine, Southern Medical University, Nanjing, Jiangsu 210002, China.

Abstract

Background: Recent studies have shown that the mRNA expression-based stemness index (mRNAsi) can accurately quantify the similarity of cancer cells to stem cells, and mRNAsi-related genes are used as biomarkers for cancer. However, mRNAsi-driven tumor heterogeneity is rarely investigated, especially whether mRNAsi can distinguish hepatocellular carcinoma (HCC) into different molecular subtypes is still largely unknown.

Methods: Using OCLR machine learning algorithm, weighted gene co-expression network analysis, consistent unsupervised clustering, survival analysis and multivariate cox regression etc. to identify biomarkers and molecular subtypes related to tumor stemness in HCC.

Results: We firstly demonstrate that the high mRNAsi is significantly associated with the poor survival and high disease grades in HCC. Secondly, we identify 212 mRNAsi-related genes that can divide HCC into three molecular subtypes: low cancer stemness cell phenotype (CSCP-L), moderate cancer stemness cell phenotype (CSCP-M) and high cancer stemness cell phenotype (CSCP-H), especially over-activated ribosomes, spliceosomes and nucleotide metabolism lead to the worst prognosis for the CSCP-H subtype patients, while activated amino acids, fatty acids and complement systems result in the best prognosis for the CSCP-L subtype. Thirdly, we find that three CSCP subtypes have different mutation characteristics, immune microenvironment and immune checkpoint expression, which may cause the differential prognosis for three subtypes. Finally, we identify 10 robust mRNAsi-related biomarkers that can effectively predict the survival of HCC patients.

Conclusions: These novel cancer stemness-related CSCP subtypes and biomarkers in this study will be of great clinical significance for the diagnosis, prognosis and targeted therapy of HCC patients.

Keywords: Cancer stem cell; Hepatocellular Carcinoma; Molecular subtype; Prognosis; mRNAsi.