Differential prognostic roles and clinical implications of mitochondrial and genomic tRNA-derived fragments in colorectal liver metastases.

BACKGROUND

Colorectal liver metastases (CRLM) are the leading cause of colorectal cancer (CRC)-related mortality. Transfer RNA-derived fragments (tRFs), a novel class of small non-coding RNAs (sncRNA), regulate gene expression, stress response, and immune functions in cancer. While increasingly implicated in CRC progression, their prognostic significance in CRLM remains unknown. This study investigates the abundance and prognostic value of genomic (ge) and mitochondrial (mt) tRFs in CRLM.

METHODS

Tumor samples from CRLM patients who underwent curative liver resection between January 2012 and December 2015 were retrospectively analyzed. Small RNA sequencing (sRNA-seq) quantified ge- and mt-tRF expression in tumor tissue. Event-free survival (EFS) was the primary outcome. Associations between tRF expression and EFS were evaluated using Cox regression, spline modeling, and network analysis.

RESULTS

Among 588 screened samples, 40 met eligibility criteria (18 females [45%], median age 64 [42-79]). A total of 432 tRFs were identified, with ge-tRFs (67%) more abundant than mt-tRFs (33%). Spline regressions classified tRFs into ten prognostic groups. High ge-tRF abundance was predominantly associated with unfavorable EFS (FDR < 0.2; 94%), while mt-tRFs were significantly (p < 0.001; χ test) more often linked to favorable EFS (FDR < 0.2; 26%). Network analysis of tRF abundance correlations revealed higher intra-mitochondrial network density compared to the intra-genomic tRF network. No significant structural differences were observed between prognostically significant vs. non-significant or favorable vs. unfavorable tRFs. Key tRF candidates, including tRHalve3-His-CAU and tRNAleader-Gln-UUG (mt-tRFs), as well as tRFmisc-Tyr-GTA (ge-tRF), remained independent prognostic markers after adjusting for clinical covariates.

CONCLUSION

This study provides the first comprehensive characterization of tRF expression in CRLM, identifying distinct prognostic roles for ge- and mt-tRFs. While ge-tRFs correlated with poor prognosis, several mt-tRFs were linked to favorable outcomes, highlighting their potential as novel prognostic biomarkers and therapeutic targets.