美国斯坦福大学Aaron M. Newman、Andrew J. Gentles等研究人员合作绘制出人类实体肿瘤中临床上不同的细胞状态和生态系统图谱。这一研究成果于2021年9月30日在线发表在国际学术期刊《细胞》上。

研究人员报道了EcoTyper，这是一个机器学习框架，用于从大量的、单细胞的和空间分辨率的基因表达数据中大规模地识别和验证细胞状态和多细胞群体。当应用于16种人类癌症的12个主要细胞系时，EcoTyper确定了69种转录定义的细胞状态。大多数状态是肿瘤组织所特有的，在不同的肿瘤类型中普遍存在，并具有明显的预后作用。

通过分析细胞状态的共同发生模式，研究人员发现了10个临床上不同的多细胞群落，具有出乎意料的强保护性，包括三个与不利生存有关的骨髓和基质元素，一个在正常组织中富集，以及两个与早期癌症发展有关。这项研究阐明了人类癌细胞组织的基本单位，并为任何组织细胞生态系统的大规模分析提供了一个框架。

据悉，确定细胞如何随其局部信号环境的变化而变化，并组织成不同的细胞群落，对于理解发育、衰老和癌症等不同过程至关重要。

附：英文原文

Title: Atlas of clinically distinct cell states and ecosystems across human solid tumors

Author: Bogdan A. Luca, Chloé B. Steen, Magdalena Matusiak, Armon Azizi, Sushama Varma, Chunfang Zhu, Joanna Przybyl, Almudena Espín-Pérez, Maximilian Diehn, Ash A. Alizadeh, Matt van de Rijn, Andrew J. Gentles, Aaron M. Newman

Issue&Volume: 2021-09-30

Abstract: Determining how cells vary with their local signaling environment and organize intodistinct cellular communities is critical for understanding processes as diverse asdevelopment, aging, and cancer. Here we introduce EcoTyper, a machine learning frameworkfor large-scale identification and validation of cell states and multicellular communitiesfrom bulk, single-cell, and spatially resolved gene expression data. When appliedto 12 major cell lineages across 16 types of human carcinoma, EcoTyper identified69 transcriptionally defined cell states. Most states were specific to neoplastictissue, ubiquitous across tumor types, and significantly prognostic. By analyzingcell-state co-occurrence patterns, we discovered ten clinically distinct multicellularcommunities with unexpectedly strong conservation, including three with myeloid andstromal elements linked to adverse survival, one enriched in normal tissue, and twoassociated with early cancer development. This study elucidates fundamental unitsof cellular organization in human carcinoma and provides a framework for large-scaleprofiling of cellular ecosystems in any tissue.

DOI: 10.1016/j.cell.2021.09.014

Source: https://www.cell.com/cell/fulltext/S0092-8674(21)01061-8