Quantifying Breast Cancer Heterogeneity Through Single-Cell Transcriptomics

Background and Objectives

Tumors are complex systems characterized by variations across genetic, transcriptomic, phenotypic, and microenvironmental levels. This study introduced a novel framework for quantifying cancer cell heterogeneity using single-cell RNA sequencing data. The framework comprised several scores aimed at uncovering the complexities of key cancer traits, such as metastasis, tumor progression, and recurrence.

Methods

This study leveraged publicly available single-cell transcriptomic data from three human breast cancer subtypes: estrogen receptor-positive, human epidermal growth factor receptor 2-positive, and triple-negative. We employed a quantitative approach, analyzing copy number alterations (CNAs), entropy, transcriptomic heterogeneity, and diverse protein-protein interaction networks (PPINs) to explore critical concepts in cancer biology.

Results

We found that entropy and PPIN activity related to the cell cycle could distinguish cell clusters with elevated mitotic activity, particularly in aggressive breast cancer subtypes. Additionally, CNA distributions varied across cancer subtypes. We also identified positive correlations between the CNA score, entropy, and the activities of PPINs associated with the cell cycle, as well as those linked to basal and mesenchymal cell lines.

Conclusions

This study addresses a gap in the current understanding of breast cancer heterogeneity by presenting a novel quantitative approach that offers deeper insight into tumor biology, overcoming some limitations of traditional marker-based methods. Using single-cell RNA sequencing data, this work introduces a novel scoring framework that quantifies key cancer traits, such as CNAs, transcriptomic heterogeneity, entropy and activities of PPIN associated with biological processes relevant to cancer biology. The proposed methodology allows exploring these features at the individual cell level, revealing intra- and inter- tumor heterogeneity that may be relevant for tumor evolution and treatment response. We applied this methodology to human scRNA-seq datasets from ER+, HER2+, TN breast cancer subtypes. Our analysis revealed significant differences in several scores across the subtypes. Overall, this approach enables a better understanding of breast cancer heterogeneity, with the potential to identify novel therapeutic targets and strategies.