Abstract:<p class="a-plus-plus">Colorectal cancer (CRC) progression is associated with cancer cell dedifferentiation and sternness acquisition. Several methods have been developed to identify sternness signatures in CRCs. However, studies that directly measured the degree of dedifferentiation in CRC tissues are limited. It is unclear how the differentiation states change during CRC progression. To address this, we develop a method to analyze the tissue differentiation spectrum in colorectal cancer using normal gastrointestinal single-cell transcriptome data. Applying this method on 281 tumor samples from The Cancer Genome Atlas Colon Adenocarcinoma dataset, we identified three major CRC subtypes with distinct tissue differentiation pattern. We observed that differentiation states are closely correlated with anti-tumor immune response and patient outcomes in CRC. Highly dedifferentiated CRC samples escaped the immune surveillance and exhibited poor outcomes; mildly dedifferentiated CRC samples showed resistance to anti-tumor immune responses and had a worse survival rate; well-differentiated CRC samples showed sustained anti-tumor immune responses and had a good prognosis. Overall, the spectrum of tissue differentiation observed in CRCs can be used for future clinical risk stratification and subtype-based therapy selection.</p>

Single-cell genomic profile-based analysis of tissue differentiation in colorectal cancer