Ion Channel-Extracellular Matrix Interplay in Colorectal Cancer: A Network-Based Approach to Tumor Microenvironment Remodeling

Abstract

The progression of colorectal cancer (CRC) is driven by dynamic interactions between tumor cells and their microenvironment, particularly the extracellular matrix (ECM). Ion channels, critical regulators of cellular signaling, have emerged as mediators of ECM remodeling and tumor aggressiveness. In this study, we integrate transcriptomic data from 185 CRC tumors and 157 adjacent normal tissues with network modeling to dissect the interplay between ion channels and the ECM. We identified 4036 differentially expressed genes (DEGs), including 188 ion channel-associated DEGs (IC-DEGs) enriched in ECM-related pathways, such as collagen assembly, matrix metalloproteinase regulation, and mechanotransduction. Structural equation modeling revealed an active CRC-ion channel module (CRC-IC) comprising 482 nodes and 422 edges, highlighting dysregulated interactions between ECM components (e.g., COL1A1, COL5A2, VCAN, LAMA4, LA-MA5, LAMC1), ion channels (e.g., TRPM5 and SLC16A1), and cytoskeletal regulators. Key nodes, including CHST11 and VCAN, were associated with ECM sulfation, tumor invasiveness, and immune evasion. Notably, survival was associated with MAPK1, SLC16A1, and ABCB4 in relation to patient prognosis. Our findings underscore the pivotal role of ion channels as co-factors in ECM dynamics in CRC, offering mechanistic insights into tumor-stroma crosstalk and identifying potential therapeutic targets to disrupt microenvironment-driven progression.

Keywords: causal network inference; colorectal cancer; extracellular matrix; ion channels; transcriptomics.

Figure 1

Volcano plot. A total of 4036 DEGs were identified, including 2050 downregulated (blue) and 1986 upregulated (red) genes, with an adjusted p-value < 0.01. Vertical dashed lines mark the log₂ Fold Change thresholds of ±1 (up- and downregulated, respectively), and the horizontal dashed line indicates the significance threshold of −log₁₀(adjusted p-value) = 2 (p < 0.01).

Figure 2

Identification of IC-DEGs in CRC. A total of 188 IC-DEGs were identified by intersecting the 4036 CRC DEGs with a panel of 720 ion-channel genes using Venny v2.1, https://bioinfogp.cnb.csic.es/tools/venny/index.html (accessed on 5 March 2025).

Figure 3

Reactome pathways identified by the 188 IC-DEGs. The dots represent the pathways, with dot size indicating the total number of genes and color denoting the number of DEGs in the pathway.

Figure 4

CRC-IC module and community. (A) Full CRC-IC module. The overall module obtained from the structural equation modeling (SEM) fitting is illustrated. (B) Reduced graph. Red nodes represent overexpressed genes, and blue nodes indicate downregulated genes. Red edges denote significant direct effects with a positive association (p-value < 0.05 and b effect > 0), and blue edges indicate significant direct effects with a negative association (p-value < 0.05 and b effect < 0).

Figure 5

Functional enrichment analysis of communities linked to the CRC-IC reduced graph. The figure illustrates the top enriched Gene Ontology (GO) terms (Molecular Function: MF, Biological Process: BP, and Cellular Component: CC) and pathways for communities 35, 36, 19, 29, 30, and 7. These communities were selected based on their overlap exceeding 40% of their genes with the CRC-IC module’s reduced graph, emphasizing their critical role in the core module’s biological processes.

Figure 6

The most relevant genes associated with overall survival (OS) in colon carcinoma. Low and high expression levels are derived (dichotomized) by a ROC analysis.

Figure 7

The study workflow. The dataset selection step is highlighted in yellow, RNA-seq bioinformatics analysis in light blue, and causal network analysis step is highlighted in green.

Figure 8

Conceptual workflow. The box (4) shows the network (i.e., the active CRC-IC module) achieved by the graph-based algorithms, from the DEG Reactome enrichment, before the SEM fitting. The boxes (5) and (6) show an active CRC-IC module, red dots represent overexpressed genes, while blue dots indicate down-expressed genes. Red arrows denote significant direct effects with a positive association, and blue arrows indicate significant direct effects with a negative association.