An EV-Guided Multi-Compartment Proof-of-Concept Framework for Biomarker Prioritization in Cholangiocarcinoma.
Cholangiocarcinoma (CCA) is a highly heterogeneous malignancy in which numerous biomarker candidates have been reported, yet few progress to clinical use. Beyond biological complexity, this low translational yield reflects the lack of systematic criteria for prioritizing biomarkers during the discovery stage. In particular, tumor-derived signals identified in tissue often fail to persist in clinically accessible biofluids, as cross-compartment signal behavior is rarely evaluated explicitly.
We developed an extracellular vesicle (EV)-guided, multi-compartment proof-of-concept framework to assess biomarker robustness and translatability early in discovery. EV proteomes from three biologically distinct CCA cell lines and a normal cholangiocyte were analyzed using multivariate and machine-learning-assisted approaches to identify conserved EV-associated features. These were integrated with public transcriptomic, epigenetic, copy-number, promoter usage, and miRNA regulatory data. Tissue relevance was assessed using TCGA/GTEx RNA-seq datasets, and exploratory signal behavior was examined in pooled serum- and urine-derived EVs from CCA patients and controls.
EV proteomics revealed marked molecular heterogeneity across CCA models but identified a small subset of conserved EV-associated proteins. SERPINF2 was used as a representative example, showing consistently reduced EV-associated abundance across all CCA models with coordinated regulation across multiple molecular layers. SERPINF2 expression was independent of patient sex and tumor stage and clearly distinguished tumor from normal bile duct tissue. Exploratory biofluid analyses demonstrated compartment-dependent signal behavior, with SERPINF2 depletion detectable in urine-derived EVs but not in serum-derived EVs.
Rather than validating a single biomarker, this study presents an EV-guided, multi-compartment framework for prioritizing biomarker candidates at the discovery stage. By explicitly accounting for tumor heterogeneity and compartment-specific signal preservation, this proof-of-concept approach provides a practical decision-support strategy for identifying biomarkers with greater translational potential in heterogeneous cancers such as CCA.
We developed an extracellular vesicle (EV)-guided, multi-compartment proof-of-concept framework to assess biomarker robustness and translatability early in discovery. EV proteomes from three biologically distinct CCA cell lines and a normal cholangiocyte were analyzed using multivariate and machine-learning-assisted approaches to identify conserved EV-associated features. These were integrated with public transcriptomic, epigenetic, copy-number, promoter usage, and miRNA regulatory data. Tissue relevance was assessed using TCGA/GTEx RNA-seq datasets, and exploratory signal behavior was examined in pooled serum- and urine-derived EVs from CCA patients and controls.
EV proteomics revealed marked molecular heterogeneity across CCA models but identified a small subset of conserved EV-associated proteins. SERPINF2 was used as a representative example, showing consistently reduced EV-associated abundance across all CCA models with coordinated regulation across multiple molecular layers. SERPINF2 expression was independent of patient sex and tumor stage and clearly distinguished tumor from normal bile duct tissue. Exploratory biofluid analyses demonstrated compartment-dependent signal behavior, with SERPINF2 depletion detectable in urine-derived EVs but not in serum-derived EVs.
Rather than validating a single biomarker, this study presents an EV-guided, multi-compartment framework for prioritizing biomarker candidates at the discovery stage. By explicitly accounting for tumor heterogeneity and compartment-specific signal preservation, this proof-of-concept approach provides a practical decision-support strategy for identifying biomarkers with greater translational potential in heterogeneous cancers such as CCA.
Authors
Kotawong Kotawong, Roytrakul Roytrakul, Phaonakrop Phaonakrop, Na-Bangchang Na-Bangchang, Chaijaroenkul Chaijaroenkul
View on Pubmed