BACKGROUND

Genomic instability (GI) is among the top ten characteristics of malignancy. Long non-coding RNAs (lncRNAs) are promising cancer biomarkers that are reportedly involved in GI. So far, the clinical value of GI-related lncRNAs (GIlncs) in papillary thyroid cancer (PTC) has not been clarified.

METHODS

Integrative analysis of lncRNA expression and somatic mutation profiles was performed to identify GIlncs. Analysis of differentially expressed lncRNAs in the group with high- and low- cumulative number of somatic mutations revealed significant GIlncs in PTC. Univariate and multivariate Cox proportional hazard regression analyses were performed to identify hub-GIlncs.

RESULTS

A computational model based on four lncRNAs (, , , and ) was identified as a quantitative index using an discovery cohort. GILS score was significantly associated with poor prognosis, as validated in the TCGA dataset and further tested in our local RNA-Seq cohort. Moreover, a combination of clinical characteristics and the composite GILS-clinical prognostic nomogram demonstrates satisfactory discrimination and calibration. Furthermore, the GILS score and , , , and were also associated with driver mutations and multiple clinical-pathological variables, respectively. Moreover, RNA-Seq confirmed the expression patterns of , , , and in PTC and normal thyroid tissues. Biological experiments demonstrated that downregulated or overexpressed affect PTC cell proliferation, migration, and invasion . Activation of the stromal and immune cell infiltration was also observed in the high group in the PTC microenvironment.

CONCLUSION

In summary, we identified a signature for clinical outcome prediction in PTC comprising four lncRNAs associated with GI. A better understanding of the GI providing an alternative evaluation of the progression risk of PTC. Our study also demonstrated as a novel oncogenic lncRNA and verified its phenotype in PTC.