Jelena Karanović*, Tamara Babić, Sandra Dragičević and Aleksandra Nikolić
Institute of Molecular Genetics and Genetic Engineering, University of Belgrade, Serbia
jelena.karanovic [at] imgge.bg.ac.rs
Abstract
Transforming growth factor beta (TGFB) signaling, a central pathway regulating cell growth, differentiation, and immune responses, is enrolled in gastrointestinal (GIT) tumors. Non-canonical transcripts significantly contribute to the complexity of TGFB regulation, adding layers of transcriptional and post-transcriptional control. This study aimed to identify common deregulated transcripts of the TGFB pathway across seven GIT tumors and to explore similarities and differences in their regulatory landscapes.
Expression data for coding and non-coding transcripts of six canonical TGFB pathway genes (TGFBR1, TGFBR2, SMAD2, SMAD3, SMAD4, and SMAD7) were retrieved from GTEx (normal tissue) and TCGA (primary tumor and solid normal tissue) databases via the XENA browser. For each tumor type (head and neck, esophagus, stomach, liver, pancreas, colon, and rectum) differential expression analysis was performed using R (v4.5.2). Transcripts deregulated in all seven tumors were subjected to principal component analysis (PCA) to identify transcriptional patterns, and to in silico annotation of coding potential (CPC2) and subcellular localization (LNClocator, iLoc-LncRNA). Enrichment analysis was performed using the gprofiler2 package.
Out of 59 significant transcripts, four non-canonical (SMAD2: ENST00000586040.5 and ENST00000356825.8; SMAD3: ENST00000439724.7; SMAD4: ENST00000591126.5) were deregulated across all tumors. They were predicted to be protein-coding, with predominant cytoplasmic localization, while the SMAD3 transcript could also localize to nuclear compartments. PCA revealed that SMAD2 transcripts differentiated samples along PC1 (47.7% variance) synergistically, whereas SMAD3 and SMAD4 transcripts separated samples along PC2 (31.4% variance) antagonistically, distinguishing normal from patients-derived samples (tumor and solid normal). All GIT tumors displayed broadly similar expression profiles relative to normal tissues, with subtle inter‑tumor differences, most pronounced in colon and rectum compared to other tumors. Enrichment analysis highlighted aberrant TGFB signaling, coupled with dysregulated cell cycle control and genomic instability, as a convergent oncogenic mechanism across GIT tumors, integrating developmental pathways, immune modulation, and malignant transformation.
In conclusion, four deregulated non‑canonical TGFB transcripts represent a common feature of GIT tumors, with subtle inter‑tumor differences most evident in colon and rectum relative to other tumors. These findings highlight their potential as biomarkers for tumor classification and therapeutic targeting, warranting validation in future investigations across diverse patient cohorts.
Keywords: gastrointestinal tumors, TGFB, transcripts
Acknowledgement: This work was funded by the Ministry of Science, Technological Development and Innovation of the Republic of Serbia (Contract No. 451-03-33/2026-03/200042).