Discovering Rare Genes Contributing to Cancer Stemness and Invasive Potential by GBM Single-Cell Transcriptional Analysis.

Lin Pang,Bo Pang,Min Yan,Yun Xiao,Liwen Xu,Huating Yuan,Yanyan Ping,Gaoming Liao,Xia Li,Feng Li,Jing Hu

doi:10.3390/cancers11122025

Abstract

Single-cell RNA sequencing presents the sophisticated delineation of cell transcriptomes in many cancer types and highlights the tumor heterogeneity at higher resolution, which provides a new chance to explore the molecular mechanism in a minority of cells. In this study, we utilized publicly available single-cell RNA-seq data to discover and comprehensively dissect rare genes existing in few glioblastoma (GBM) cells. Moreover, we designed a framework to systematically identify 51 rare protein-coding genes (PCGs) and 47 rare long non-coding RNAs (lncRNAs) in GBM. Patients with high expression levels of rare genes like CYB5R2 and TPPP3 had worse overall survival and disease-free survival, implying their potential implication in GBM progression and prognosis. We found that these rare genes tended to be specifically expressed in GBM cancer stem cells, which emphasized their ability to characterize stem-like cancer cells and implied their contribution to GBM growth. Furthermore, rare genes were enriched in a 17-cell subset, which was located in an individual branch of the pseudotime trajectory of cancer progression and exhibited high cell cycle activity and invasive potential. Our study captures the rare genes highly expressed in few cells, deepens our understanding of special states during GBM tumorigenesis and progression such as cancer stemness and invasion, and proposes potential targets for cancer therapy.

Highlights

Glioblastoma (GBM) is the most common primary brain cancer in adults and the leading cause of brain cancer-related deaths, with median overall survival of only 12~18 months [1]
Liu and colleagues deployed single-cell RNA-seq of the human brain to provide a comprehensive annotation and quantification of long non-coding RNAs at a greater resolution [18]. In contrast to their low expression levels detected in bulk samples, lncRNAs were expressed at levels comparable with those of mRNAs in individual cells and showed the cell-type and single-cell specificity, which was confirmed by in situ hybridization
We investigated the distribution of average expression levels of detected protein-coding genes (PCGs) and lncRNAs across cells

Summary

Introduction

Glioblastoma (GBM) is the most common primary brain cancer in adults and the leading cause of brain cancer-related deaths, with median overall survival of only 12~18 months [1]. Liu and colleagues deployed single-cell RNA-seq of the human brain to provide a comprehensive annotation and quantification of long non-coding RNAs (lncRNAs) at a greater resolution [18] In contrast to their low expression levels detected in bulk samples, lncRNAs were expressed at levels comparable with those of mRNAs in individual cells and showed the cell-type and single-cell specificity, which was confirmed by in situ hybridization. Torre et al performed single-cell analysis on a melanoma cell line to explore rare cell gene expression patterns [19] They observed that some resistance markers such as EGFR, AXL, WNT5A and NGFR, which showed overall low average expression across all cells, had high expression levels in a subset of cells. These results indicate that genes with cell subset-specific expression patterns would have crucial effects on pivotal biological activities

Methods

Results

Conclusion