ABSTRACT
Melanoma incidence is rising, and accurate risk stratification remains challenging because of the molecular heterogeneity underlying disease progression. Specifically, the transition from benign nevi to malignant melanoma and the acquisition of the aggressive ulcerated phenotype represent critical barriers in clinical management that require novel biomarkers. In addition, there is a critical need for novel biomarkers to identify patients with primary melanoma who, despite complete surgical resection, remain at high risk of recurrence and would benefit from adjuvant therapy.
We applied an integrative systems biology approach to decipher the miRNA-dependent regulatory architecture of melanoma. Following PRISMA guidelines, we conducted a robust meta-analysis of six independent transcriptomic studies, overcoming inter-study heterogeneity. This was coupled with network inference algorithms to construct validated miRNA-mRNA interactomes and identify dynamic functional modules.
In the diagnostic scenario, we identified a consensus signature of 24 miRNAs. Network topology analysis revealed hsa-miR-142-5p as a master regulator that orchestrates the dismantling of the Oncogene-Induced Senescence (OIS) barriers by targeting CDK6, SIRT1, and TGFBR2. In the prognostic scenario (ulceration), we identified a specific “stress-adaptive” signature of 23 miRNAs. Notably, the upregulation of hsa-miR-223-3p emerged as a key driver of invasiveness by suppressing the motility-limiting tumor suppressor RHOB, while the concurrent loss of hsa-miR-200c and hsa-miR-489-3p may unleash Epithelial-Mesenchymal Transition (EMT) and EGFR-driven survival pathways.
Our study supports a double-switch mechanism where specific miRNA alterations first drive senescence escape and subsequently promote survival in the hypoxic ulcerated niche. These signatures offer robust biomarkers for diagnosis and highlight the miR-142-5p/CDK6 and miR-223-3p/RHOB axes as potential therapeutic targets for precise intervention.
Full Text
The Full Text of this preprint is available as a PDF (3.9 MB). The Web version will be available soon.