A Two-Circular RNA Signature of Donor Circfoxn2 and Circnectin3 Predicts Early Allograft Dysfunction after Liver Transplantation

Kun Wang,Xuyong Wei,Qiang Wei,Di Lu,Wangyao Li,Binhua Pan,Junli Chen,Haiyang Xie,Shusen Zheng,Xiao Xu

DOI: https://doi.org/10.21037/atm.2019.12.132

IF: 3.616

2020-01-01

Annals of Translational Medicine

Abstract:BACKGROUND:Early allograft dysfunction (EAD) following liver transplantation is associated with poor recipient and graft survival. In recent years, circular RNAs (circRNAs) have emerged as important components of endogenous RNAs. This study aims to explore the expression profile and predictive value of graft circular RNAs for EAD after liver transplantation.METHODS:RNA sequencing was conducted to identify the circRNA profile in donor liver tissues. Furthermore, quantitative real-time polymerase chain reaction (qRT-PCR) was used to identify candidate circRNAs. A novel model combining circular RNA signature was established to predict EAD based on the multivariate analysis.RESULTS:A total of 442 circRNAs were differentially expressed between the EAD and non-EAD groups, of which, 223 were significantly upregulated and 219 were downregulated in the EAD group (Fold change >2, P<0.05). qRT-PCR validation indicated that circFOXN2 and circNECTIN3 levels in the EAD group were significantly lower than those in the non-EAD group (P=0.038, 0.024, respectively; n=115). Among the 115 recipients, 32 recipients with high circFOXN2 expression were classified as circular RNA signature A and the rest recipients with low circFOXN2 expression were categorized into circular RNA signature B (n=33, high circNECTIN3 expression) and C (n=50, low circNECTIN3 expression). The incidence rates of EAD in signature A, B and C were significantly different (3.1%, 21.2% and 42.0%, respectively; P=0.000). According to the multivariate analysis, a novel predictive model for EAD was developed based on CIT (P=0.000) and circular RNA signature (P=0.013). The novel model displayed a high predictive value for EAD with areas under the curve (AUC) of 0.870 (95% CI: 0.797-0.942).CONCLUSIONS:Donor circFOXN2 and circNEXTIN3 were associated with the incidence of EAD. The novel model combing the two-circular RNA signature had a high predictive value for EAD.