The Most Comprehensive Study at Single-Cell Resolution: A Giant Step toward Understanding Gastric Cancer

Fei-Yu Diao

DOI: https://doi.org/10.1055/s-0042-1758763

2022-12-10

Global Medical Genetics

Abstract:In 2020, gastric cancer is the fifth common cancer and the fifth leading cause of cancer death in the world.[1] It has the highest incidence and mortality rate in Asian countries, such as China, Japan, and South Korea.[2] [3] Heterogeneity at the histologic, transcriptomic, genomic, and epigenomic levels exists between gastric cancer patients (interpatient heterogeneity) and within individual tumor mass (intertumoral heterogeneity). It leads to different cancer biological behaviors and treatment response.[4] Therefore, biomarkers developed based on the heterogeneity of gastric cancer play an important role in guiding clinical treatment and improving patient prognosis.[5] [6] Although some current cancer genome projects, The Cancer Genome Atlas (TCGA) and Asian Cancer Research Group, have made great progress in facilitating the molecular typing of gastric cancer, their role in improving the prognosis of gastric cancer patients is limited. Therefore, to conduct high-resolution studies at molecular level in a wide range of patients to guide the clinical treatment of gastric cancer is necessary. Previous "bulk-transcriptome" studies have found that each gastric cancer case has a unique expression profile contributed by cancer cells and resident cell types of tumor microenvironment (such as cancer-associated fibroblasts, immune cells, and endothelial cells, etc.),[7] but the underlying molecular mechanisms of how tumor microenvironment resident cells drive tumor phenotype evolution and clinical progression remain unknown. With the advances in bioinformatics, bulk sequencing data has been successfully decomposed into lineage-specific constituent programs, but this approach fails to discern rare cell populations, fine-scale tissue lineages, cell–cell interactions, and relationships between lineages.[8] Single-cell RNA sequencing (scRNA-seq) is the primary tool for addressing these issues. It can detect gene expression in thousands of cells simultaneously, enabling comprehensive analysis of different cell types in tumor mass under different conditions. Indeed, scRNA-seq on gastric cancer tissues from various sources has provided unique insights of cancer biology. However, these current scRNA-seq studies are limited by the number of samples and cells, as well as the dissociation requirements for tissues, which had led to the loss of many key information, especially spatial information. Thus, digital spatial analysis, in situ sequencing, and multiplexed error-robust fluorescence in situ hybridization platforms have been developed to maximize the preservation of spatial information, and thereby allowing the in-depth analysis of tumor–tumor microenvironment interactions. In a study recently published in Cancer Discovery , titled "Single-Cell Atlas of Lineage States, Tumor Microenvironment, and Subtype-Specific Expression Programs in Gastric Cancer," Kumar et al[9] delineated a comprehensive single-cell atlas of gastric cancer specimens across clinical stages and histologic subtypes by scRNA-seq and complemented this atlas by spatial transcriptomics, orthogonal validation in independent bulk RNA-seq cohorts, and functional demonstration using patient-derived organoids and in vivo models. This scRNA-seq study discovered several new rare cell populations undergoing state transition in gastric cancer, associations between plasma cell and cancer-associated fibroblast sublineages, and gastric cancer clinical stages or histological subtypes, and gastric cancer-associated cell type-specific expression programs. One of the strengths of this study is the large number of cells analyzed by scRNA-seq (more than 200,000 cells), which is much higher than the sum of all previous gastric cancer scRNA-seq studies. Additionally, the clinical specimens collected in this study covered multiple clinical stages and subtypes of gastric cancer and also included a comparative analysis of organoids. Most importantly, the main findings of this study are orthogonally verified by new spatial transcriptomics technology (digital spatial profiling [DSP]). Previous scRNA-seq studies compared tumors and normal tissues in a cell-lineage-specific manner, such that their tumor profiles were composite tumor profiles assembled from different error signatures expressed by different lineages. By using DSP technology, this study found that chief cells and intestinal-type cells contributed the largest cancer-associated gene expression differences in cancer epithelial component. The most noteworthy is that the intestinal-type epithelial cells (EpiInt1) highly expressed oncogenes of gastric cancer, suggested that EpiInt1 population was a key epithelial cell population in the transition of metaplastic gastric epithelial cells into cancer cells. However, the platform used in this study is unable to detect single-cell alterations at both the DNA and RNA levels on a genome-wide sca -Abstract Truncated-