Decoupling epithelial-mesenchymal transitions from stromal profiles by integrative expression analysis.

Area of Science:

• Oncology
• Computational Biology
• Genomics

Background:

• Epithelial-to-mesenchymal transition (EMT) is a known driver of cancer metastasis.
• Distinguishing EMT in cancer cells from normal stromal cells in the tumor microenvironment is challenging.
• Existing methods struggle to isolate EMT-specific gene expression profiles.

Purpose of the Study:

• To develop a computational framework to differentiate mesenchymal signatures between cancer and stromal cells.
• To analyze Epithelial-to-mesenchymal transition (EMT) patterns across a large cohort of tumors.
• To investigate the association between EMT patterns and clinical features, including metastasis.

Main Methods:

• Utilized published single-cell RNA sequencing data for direct comparison of mesenchymal signatures.
• Developed a deconvolution framework using bulk RNA sequencing datasets to separate cancer and stromal cell profiles.
• Applied the framework to analyze EMT characteristics across hundreds of tumor samples.

Main Results:

• Identified three distinct patterns of Epithelial-to-mesenchymal transition (EMT).
• Observed specific EMT patterns associated with squamous, gynaecological, and gastrointestinal cancers.
• Found that in most cancer types, EMT patterns were not significantly correlated with increased metastasis risk.

Conclusions:

• The developed computational framework effectively deconvolves mesenchymal signals from cancer and stromal cells.
• Epithelial-to-mesenchymal transition (EMT) plays a complex role in cancer progression, not always directly correlating with metastasis.
• Other stages of the metastatic cascade may represent a more significant bottleneck for cancer spread than EMT alone.