PhD Scientific Days 2022

Pathology and Oncology III. (Poster discussion will take place in the Aula during the Coffee Break)

Correlation analysis of normal pancreas and malignant pancreatic ductal adenocarcinoma on the transcriptome level

Text of the abstract

Introduction: Gene co-expression correlations frequently signal shared biological functions with coordinated regulation. We hypothesized that maintained correlations might be essential for cellular survival, representing potential vulnerabilities of cancer cells.

Aims: We aimed to reveal correlations preserved in pancreatic ductal adenocarcinomas (PDAC) across normal and tumor tissues.

Methods: We searched the NCBI GEO for raw microarray data and the TCGA project for RNA-seq data. The microarray dataset consisted of 248 tumor and 108 normal samples allowing the analysis of 12,210 genes. The RNA-seq dataset incorporated 177 tumor, four normal samples from TCGA, and 248 normal samples from GTEx, enabling the analysis of 21,479 genes. Genes with an altered expression were identified with a Mann-Whitney U test at p<0.01, and a Pearson correlation was performed to identify preserved correlations.

Results: Altogether 371 significant correlations involving 262 genes were preserved across normal samples and tumors in both RNA-seq and gene chip platforms. The identified close-knit gene network is mainly responsible for extracellular matrix organization. Seven genes (SPARC, COL6A3, MMP2, HTRA1, FN1, PALLD, and COL3A1) were heavily overrepresented in maintained correlations, some of them participating in as many as 58 interactions. High expression of 28 genes was linked to poor disease outcome at FDR ≤ 10%, out of which FN1, an extracellular matrix component, was both overrepresented in maintained correlations and associated with worse overall survival (p = 0.00097, FDR ≤ 5%). The growing expression of two genes, MYL12A and MYL12B, across normal tissues, primary, and metastatic tumors may drive the acquisition of motility by cancer cells.

Conclusion: Our results propose novel prognostic biomarkers of PDAC and pinpoint fundamental cellular interactions as potential targets for combination therapies. The presence of significant correlations across different data platforms substantiates the validity of our findings.