A Gene Signature Derived from the Loss of CDKN1A (p21) Is Associated with CMS4 Colorectal Cancer

Santiago Bueno-Fortes; Julienne K Muenzner; Alberto Berral-Gonzalez; Chuanpit Hampel; Pablo Lindner; Alexandra Berninger; Kerstin Huebner; Philipp Kunze; Tobias Bäuerle; Katharina Erlenbach-Wuensch; José Manuel Sánchez-Santos; Arndt Hartmann; Javier De Las Rivas; Regine Schneider-Stock

doi:10.3390/cancers14010136

A Gene Signature Derived from the Loss of CDKN1A (p21) Is Associated with CMS4 Colorectal Cancer

Cancers (Basel). 2021 Dec 28;14(1):136. doi: 10.3390/cancers14010136.

Authors

Santiago Bueno-Fortes¹, Julienne K Muenzner^{2

3}, Alberto Berral-Gonzalez¹, Chuanpit Hampel^{2

3}, Pablo Lindner^{2

3}, Alexandra Berninger^{2

3}, Kerstin Huebner^{2

3}, Philipp Kunze^{2

3}, Tobias Bäuerle⁴, Katharina Erlenbach-Wuensch³, José Manuel Sánchez-Santos^{1

5}, Arndt Hartmann³, Javier De Las Rivas¹, Regine Schneider-Stock^{2

3}

Affiliations

¹ Bioinformatics and Functional Genomics Group, Cancer Research Center (CiC-IMBCC, CSIC/USAL/IBSAL), Consejo Superior de Investigaciones Científicas (CSIC) and University of Salamanca (USAL), 37007 Salamanca, Spain.
² Experimental Tumor Pathology, University Hospital of the Friedrich-Alexander University Erlangen-Nürnberg, 91054 Erlangen, Germany.
³ Institute of Pathology, University Hospital of the Friedrich-Alexander University Erlangen-Nürnberg, 91054 Erlangen, Germany.
⁴ Preclinical Imaging Platform Erlangen (PIPE), Institute of Radiology, University Hospital Erlangen-Nuremberg, 91054 Erlangen, Germany.
⁵ Department of Statistics, University of Salamanca (USAL), 37008 Salamanca, Spain.

Abstract

The epithelial-mesenchymal transition (EMT) is associated with tumor aggressiveness and increased invasion, migration, metastasis, angiogenesis, and drug resistance. Although the HCT116 p21-/- cell line is well known for its EMT-associated phenotype, with high Vimentin and low E-cadherin protein levels, the gene signature of this rather intermediate EMT-like cell line has not been determined so far. In this work, we present a robust molecular and bioinformatics analysis, to reveal the associated gene expression profile and its correlation with different types of colorectal cancer tumors. We compared the quantitative signature obtained with the NanoString platform with the expression profiles of colorectal cancer (CRC) Consensus Molecular Subtypes (CMS) as identified, and validated the results in a large independent cohort of human tumor samples. The expression signature derived from the p21-/- cells showed consistent and reliable numbers of upregulated and downregulated genes, as evaluated with two machine learning methods against the four CRC subtypes (i.e., CMS1, 2, 3, and 4). High concordance was found between the upregulated gene signature of HCT116 p21-/- cells and the signature of the CMS4 mesenchymal subtype. At the same time, the upregulated gene signature of the native HCT116 cells was similar to that of CMS1. Using a multivariate Cox regression model to analyze the survival data in the CRC tumor cohort, we selected genes that have a predictive risk power (with a significant gene risk incidence score). A set of genes of the mesenchymal signature was proven to be significantly associated with poor survival, specifically in the CMS4 CRC human cohort. We suggest that the gene signature of HCT116 p21-/- cells could be a suitable metric for mechanistic studies regarding the CMS4 signature and its functional consequences in CRC. Moreover, this model could help to discover the molecular mechanisms of intermediate EMT, which is known to be associated with extraordinarily high stemness and drug resistance.

Keywords: CAM model; CDKN1A; HCT116 cells; SNAI2; colorectal cancer; consensus molecular subtypes (CMS); epithelial–mesenchymal transition (EMT); intermediate EMT.