Cancerous tumors are composed of cells that may look the same but are in fact incredibly diverse. Mutations and changes in gene activity give rise to competing sub-populations of tumor cells with different behaviors, such as ability to resist anticancer treatments.
Researchers now have developed a way to zero in on the subpopulations of cells within a tumor that are driving important disease behaviors.
“If we can precisely identify which cell subpopulations are responsible for drug response, tumor progression and metastasis, then we can look deeper to identify the mechanisms and develop better targeted therapies,” said Zheng Xia, Ph.D., senior author of a paper in Nature Biotechnology describing the new method. Xia is an assistant professor of biomedical engineering in the OHSU School of Medicine.
Image: scanning electron microscope view of lung cancer cells (Anne Weston/Francis Crick Institute CC BY-NC 4.0)The method expands the power of single-cell analysis, which can identify cell types and cell states of different subpopulations by measuring and sequencing RNA, the genetic messenger molecule transcribed from active genes.
Single-cell analysis has been limited by small sample sizes, resulting in a lack of statistical power to identify the tumor cell subpopulations driving patient-relevant outcomes such as drug resistance and metastasis, or spreading to other body parts. Meanwhile, such valuable clinical information is widely available – at the whole-tumor level – in massive public databanks such as The Cancer Genome Atlas. But these databanks have amassed gene sequencing and other information from bulk methods, which average the properties of the cells.
Xia and his team devised a computational method to overcome the information gap. “We are using bulk data as a bridge to link clinical information to single-cell findings and guide the identification of critical single-cell subpopulations,” Xia said. They’ve dubbed the method “Scissor”, and say it could enable broad application of widely available clinical information in single-cell data analysis to unlock the most disease-relevant cell subpopulations for cell-targeted therapies.
In a demonstration of the power of Scissor, the researchers identified an aggressive cancer cell subpopulation in lung adenocarcinoma tumors that was associated with worse survival outcomes. They had single-cell data from just two cancer patients, but clinical outcomes data from 471 patient samples in The Cancer Genome Atlas.
In the Scissor-selected cells associated with worse survival, the researchers found a distinct pattern of gene activity: 23 genes were up-regulated and 205 were down-regulated. The significance of the up-regulated gene signature enriched in hypoxia pathways held up in an analysis of six independent lung cancer datasets. In 5 of the 6, patients with higher signature scores had significantly worse survival than the patients with lower signature scores. The findings could point the way to potential drug targets, Xia said.
Scissor analysis of melanoma data revealed a subpopulation of immune T cells associated with favorable response to the immunotherapy drugs called checkpoint inhibitors.
The researchers identified 17 genes that were up-regulated and 120 that were down-regulated in the Scissor-selected T cells. The gene activity signature held up as predictive of response when the researchers looked at data from an independent set of patients.
Beyond cancer, the researchers showed how Scissor could effectively interpret data from Alzheimer’s disease and from a type of muscular dystrophy.
One of the strengths of Scissor is the way it automatically selects cell subpopulations from single-cell data that are driving the tumor behavior or clinical outcome of interest. “It is hypothesis free,” Xia said. “You don’t need to hypothesize which cell type is responsible.”
∞
Further reading:
Identifying phenotype-associated subpopulations by integrating bulk and single-cell sequencing data by Duanchen Sun, Xiangnan Guan, Amy E. Moran, Ling-Yun Wu, David Z. Qian, Pepper Schedin, Mu-Shui Dai, Alexey V. Danilov, Joshi J. Alumkal, Andrew C. Adey, Paul T. Spellman, and Zheng Xia. Nature Biotechnology (Nov. 11, 2021)
