Hayford Corey E, Tyson Darren R, Robbins C Jack, Frick Peter L, Quaranta Vito, Harris Leonard A
Chemical and Physical Biology Graduate Program, Vanderbilt University School of Medicine, Nashville, Tennessee, United States of America.
Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tennessee, United States of America.
PLoS Biol. 2021 Jun 1;19(6):e3000797. doi: 10.1371/journal.pbio.3000797. eCollection 2021 Jun.
Tumor heterogeneity is a primary cause of treatment failure and acquired resistance in cancer patients. Even in cancers driven by a single mutated oncogene, variability in response to targeted therapies is well known. The existence of additional genomic alterations among tumor cells can only partially explain this variability. As such, nongenetic factors are increasingly seen as critical contributors to tumor relapse and acquired resistance in cancer. Here, we show that both genetic and nongenetic factors contribute to targeted drug response variability in an experimental model of tumor heterogeneity. We observe significant variability to epidermal growth factor receptor (EGFR) inhibition among and within multiple versions and clonal sublines of PC9, a commonly used EGFR mutant nonsmall cell lung cancer (NSCLC) cell line. We resolve genetic, epigenetic, and stochastic components of this variability using a theoretical framework in which distinct genetic states give rise to multiple epigenetic "basins of attraction," across which cells can transition driven by stochastic noise. Using mutational impact analysis, single-cell differential gene expression, and correlations among Gene Ontology (GO) terms to connect genomics to transcriptomics, we establish a baseline for genetic differences driving drug response variability among PC9 cell line versions. Applying the same approach to clonal sublines, we conclude that drug response variability in all but one of the sublines is due to epigenetic differences; in the other, it is due to genetic alterations. Finally, using a clonal drug response assay together with stochastic simulations, we attribute subclonal drug response variability within sublines to stochastic cell fate decisions and confirm that one subline likely contains genetic resistance mutations that emerged in the absence of drug treatment.
肿瘤异质性是癌症患者治疗失败和获得性耐药的主要原因。即使在由单一突变致癌基因驱动的癌症中，对靶向治疗的反应差异也众所周知。肿瘤细胞中额外基因组改变的存在只能部分解释这种差异。因此，非遗传因素越来越被视为癌症肿瘤复发和获得性耐药的关键因素。在此，我们表明遗传和非遗传因素在肿瘤异质性实验模型中均对靶向药物反应差异有影响。我们观察到，在常用的表皮生长因子受体（EGFR）突变非小细胞肺癌（NSCLC）细胞系PC9的多个版本和克隆亚系之间以及内部，对EGFR抑制存在显著差异。我们使用一个理论框架解析这种差异的遗传、表观遗传和随机成分，在该框架中，不同的遗传状态会产生多个表观遗传“吸引盆”，细胞可在随机噪声驱动下在这些“吸引盆 ”之间转变。通过突变影响分析、单细胞差异基因表达以及基因本体（GO）术语之间的相关性将基因组学与转录组学联系起来，我们建立了驱动PC9细胞系不同版本间药物反应差异的遗传差异基线。将相同方法应用于克隆亚系，我们得出结论，除一个亚系外，所有亚系的药物反应差异均归因于表观遗传差异；另一个亚系则归因于基因改变。最后，通过克隆药物反应测定以及随机模拟，我们将亚系内亚克隆药物反应差异归因于随机细胞命运决定，并确认一个亚系可能包含在无药物治疗情况下出现的遗传耐药突变。
