Development of alternative splicing signature in lung squamous cell carcinoma

Abstract

Increasing evidence demonstrated that alternative splicing (AS) played a vital role in tumorigenesis and clinical outcome of patient. However, systematically analysis of AS in lung squamous cell carcinoma (LUSC) is lacking and greatly necessary. Thus, this study was to systematically estimate the function of AS events served as prognostic indicators in LUSC. Among 31,345 mRNA AS events in 9,633 genes, we detected 1,996 AS in 1,409 genes which have significantly connection with overall survival (OS) of LUSC patients. Then, prognostic model based on seven types of AS events were established and we further constructed a combined prognostic model. The Kaplan-Meier curve results suggested that seven types of AS signatures and the combined prognostic model could exhibit robust performance in predicting prognosis. Patients in the high risk group had significantly shorter OS than those in the low risk group. The ROC showed all prognostic models had high accuracy and powerful predictive performance with different AUC ranging from 0.837 to 0.978. Moreover, the combined prognostic model had highest performance in risk stratification and predictive accuracy than single prognostic models. Finally, a significant correlation network between survival-related AS genes and prognostic splicing factors (SFs) was established. In conclusion, our study provided several potential prognostic AS models and constructed splicing network between AS and SFs in LUSC, which could be used as potential indicators and treatment targets for LUSC patients.

Switched alternative splicing events as attractive features in lung squamous cell carcinoma

“…It is well believed that subtype-specific and progression-specific splicing aberrations would usher us into an unprecedented era of cancer studies [16]. Recently, a few studies in LUSC have done some integrative analysis about the prognostic value of AS [40,41]. However, the comprehensive evaluation of those AS events involved in splicing pattern shifts or the LNM status of LUSC is still lacking.…”

“…Zhao et al and Li et al systematically identified the prognostic AS signature in NSCLC, the former group conducted prognostic AS models from the perspective of different sexes [48], and the latter group firstly created prognostic predictors with one type or all types of AS events in LUSC [70]. Yan et al and Liu et al also built prognostic models for seven types of AS events one by one recently [40,41]. In this aspect, our study used combined algorithms and constructed a final prognostic risk score model with a few AS events, which focused on detailed indicators rather than an overview of AS-prognosis significance.…”