Computational network analysis of two popular skin cancers provides insights into the molecular mechanisms and reveals common therapeutic targets

Abstract

Basal Cell Carcinoma (BCC) and Actinic Keratosis (AK) are prevalent skin conditions with significant health complications. The molecular mechanisms underlying these conditions and their potential shared pathways remain ambiguous despite their prevalence. Therefore, this study aims to elucidate the common molecular pathways and potential therapeutic targets for BCC and AK through comprehensive computational network analysis. Linkage analysis was performed to identify common liable genes between BCC and AK. Protein-protein interactions (PPIs), Topological properties, GO enrichment, pathway enrichment, and gene regulatory network analyses were also performed to reveal potential molecular mechanisms and pathways. Furthermore, we evaluated protein-drug interactions (PDIs) to identify potential therapeutic targets. Our analysis revealed 22 common genes between BCC and AK: TP53, EGFR, CDKN2A, MMP9, PTGS2, VDR, BCL2, MMP2, EZH2, TP63, FOXP3, MSH2, MMP14, FLG, MC1R, CDKN2B, TIMP3, TYR, SOX10, IRF4, KRT17, and NID1. PPI network analysis highlighted TP53 and EGFR as central hubs, validated using RNA-seq data. Co-expression and physical interaction analysis revealed a strong interplay between the common genes at the transcriptional and functional levels. GO analysis identified skin cancer-relevant terms: "skin development", "immune system development", and "response to radiation" as significantly enriched biological processes, while pathway enrichment analysis highlighted several cancer-related pathways enrichment. Gene regulatory network analysis revealed complex interactions between genes, miRNAs, and transcription factors, with TP53, BCL2, and EGFR playing central roles. PDI network analysis identified ibuprofen as a potential therapeutic agent targeting PTGS2 and BCL2, while other proteins VDR, MMP2, MMP9, and TYR showed interactions with multiple drugs. This computational analysis provides valuable insights into the shared molecular mechanisms of BCC and AK, revealing common pathways and potential therapeutic targets for developing novel treatment strategies and repurposing existing drugs for these prevalent skin cancers. Therefore, these findings may guide future research in understanding and developing targeted therapies for both conditions.

Keywords: Actinic keratosis; Basal cell carcinoma; Bioinformatics; Drug repurposing; Protein-drug interaction; Protein-protein interaction; Skin cancers.

Fig. 1

Graphical representation of current study workflow where Protein-protein interactions (PPI), Gene Regulatory Network (GRN), Protein-drug interactions (PDI), Co-expression, and Physical Interactions (C&I) have been thoroughly analyzed.

Fig. 2

Identification and visualization of overlapping genes. (A) Identification of overlapping genes using Fisher's exact test. (B) Visualization of overlapping genes using Venn diagram which shows 22 common genes between Basal cell carcinoma and Actinic keratosis genes dataset.

Fig. 3

Protein-Protein interaction network (A) InnateDB database and (B) STRING database using NetworkAnalyst 3.0 among common 22 genes to represent condensed interconnection where TP53, EGFR, BCL2, EZH2, MSH2, CDKN2A were the most common and significantly connected in both of the two databases.

Fig. 4

Topological Properties of the top 22 responsible genes based on the PPI network. (A) represent the Closeness Centrality. (B) represent the avg. Cluster coefficient. (C) represent the Betweenness Centrality. (D) represent the Topological coefficient. The red dots represent the 22 genes together with highly connected neighbors (black) that have similar topological properties.

Fig. 5

Co-expression & Physical Interaction between 22 genes using GeneMANIA. (A) Whole network, (B) Co-expression, (C) Physical Interaction, (D) Pathway. Smaller size node represents a lower degree of association and larger node size represents a higher degree of association among genes.

Fig. 6

Visualization of GO-terms enrichment analysis and their regulatory functions. 6 Major GO terms were significant with FDR ≤0.05.

Fig. 7

Pathway analysis revealed (A) KEGG and (B) Reactome Pathways enrichment.

Fig. 8

Gene regulatory network analysis revealed (A) Gene-miRNA interaction, (B) TF-gene Interaction, and (C) TF-miRNA Coregulatory Network.

Fig. 9

Graphical representation of PDI network using Networkanalyst where PTGS and BCL2 interacted with common drug ibuprofen. Multiple other drugs interacted with proteins VDR, MMP2, MMP9, TYR, and MC1R.

Fig. 10

Validation of hub genes from GEO datasets. (A) and (B) Significantly upregulated and downregulated genes of BCC and AK, respectively. (C) Seven validated hub genes of topological centralities using dysregulated genes of BCC, and AK from RNA-seq data.