Introduction
Lung cancer is one of the common human malignancies and is the leading cause of cancer-associated deaths worldwide, resulting in approximately 1.7 million deaths (18.4%) worldwide in 2018 [
1] and about 140,000 deaths per year in the USA in 2019 and 2020 [
2,
3]. The non-small cell lung cancer (NSCLC) mainly included lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), and large cell lung carcinoma (LCLC) [
4]. LUSC accounts for approximately 30% of all lung cancer cases [
4]. Unfortunately, Patients tend to be older, typically diagnosed at advanced stage, and lack of effective molecular-targeted drugs [
5]. Thus, they have a poor prognosis with five-year survival rate of < 15% [
6]. Currently, the pathogenesis and progressive mechanisms of LUSC remain unclear. Though great progress has been made in diagnosis and targeted treatment of lung cancer, its clinical outcome is still unsatisfactory [
7‐
9]. Therefore, it is necessary to further systematically understand the biological mechanisms in LUSC, which could facilitate the development of target therapy and prognostic biomarkers of LUSC patients.
Alternative splicing (AS) is the process of removing introns from most human multi-exon genes and alternatively including or excluding specific exons. AS of pre-mRNA is one of the most extensive and sophisticated mechanisms to interpret proteome diversity and produce mature mRNAs and protein variants structurally and functionally [
10]. In addition to protein diversity, the translation of mRNA isomer was also downregulated by AS events through resulting in degradation of premature stop codon [
11]. Hence, AS was an indispensable procedure and alterations in splicing patterns had a close relationship with protein functions. In the past few years, extensive genomic and functional investigations had found that initiation of specific isoforms and splicing defects were driving factors for cancer [
12,
13]. For the past few years, accumulating evidence has illustrated that unusual AS events could exert a straight role in the process of biogenesis and deterioration of cancers by involving in cell proliferation, migration, immune escape and other process [
14].
Recently, many studies have described the perturbation of AS events in various cancer, including lung cancer. For example,
Kong-Beltran et al. observed met exon 14 in lung cancer which could contribute to protein regions' deletion that restricts its kinase catalytic activity [
15]. More recently, analyses of AS have also shown prognostic value for a variety of cancer types, including non-small cell lung cancer [
7], ovarian cancer [
16], breast cancer [
17] and so on. Currently, several studies have mainly concentrated on identifying different AS events between cancer tissues and normal controls or prognostic AS events [
18,
19]. Nevertheless, few articles systematically reported the LUSC-specific AS events associated with survival of patients.
Moreover, several studies have shown that alteration in splicing factors might facilitate to activate oncogenes and tumor pathways or alternatively destruct the effect of tumor suppressors [
20,
21]. Hence, it is imperative to draw a comprehensive regulatory network of SFs [
22,
23]. Considering the close connection between AS and SFs and the fact that they are only superficially understood, it is imperative to investigate their prognostic property, as well as the regulatory mechanism in LUSC.
The Cancer Genome Atlas (TCGA) project provides abundant resources, such as Exon, splice, and transcript isoform expression levels, to investigate AS patterns of cancers [
24]. We systematically profiled the genome-wide LUSC-specific alternative splicing events from TCGA. The purpose of this study is to explore the roles of alternative splicing events that could be considered as prognostic biomarkers in LUSC. Findings in this study would contribute to exploit novel and appropriate therapeutic treatments for LUSC.
Discussion
Alternative splicing (AS) is an important process by which a single pre-mRNA precursor generates a large number of mature mRNAs and protein isoforms structurally and functionally [
27,
28]. Accumulated evidence revealed that the plasticity of AS participated in cell metastasis, apoptosis, invasion, proliferation, immune evasion, and drug reliance of tumors by promoting cancer cells to produce isoform conversion [
29‐
31]. So far, majority of studies focused on the exploration of AS events as biomarker for cancers have demonstrated that several AS events and spliced variants could be indicators to diagnose and predict cancers. For example, Kelley et al. revealed that the aberrant splicing expression of GSN gene had an obviously higher expression in tumor tissues than in adjacent tissues and regulated the HNSCC’s cell proliferation process [
32]. Recently, a report suggested that KRAS-4A and KRAS-4B (KRAS isoforms) had significantly related to poor survival of CRC patients, especially microsatellite stable primary CRC [
33]. Many studies demonstrated that CD44 could present multiple isoforms through variable mRNA splicing and CD44 isoforms play critical roles in tumor initiation and be considered as potential treatment target in CRC compared with CD44 [
34,
35].
Currently, a lot of studies have also reported abnormal AS events that exerted crucial functions in lung cancer by promoting the initiation and progression of lung cancer. For example, a microarray study found differential splicing of MACF1, VEGFA, NUMB, and APP between NSCLC and control tissues in 29 patients [
8]. In addition, AS variants were also related to drug sensitivity of lung cancer. It is previously reported that patients with lung adenocarcinoma with carcinogenic mutations at MET exon 14 RNA splice acceptor and donor sites could benefit from treatment with MET inhibitors crizotinib and cabozantinib, identifying a novel therapeutic target for lung adenocarcinoma [
9]. Although some researchers have identified several prognostic alternative splicing events in LUAD and LUSC [
7], with the development of high-throughput sequencing technique, novel prognosis-related AS events and potentially therapeutic targets needed to be explored further.
Here, we identified AS signatures and established regulatory network between AS events and SFs in LUSC through the analysis of TCGA program to gain systematic and comprehensive perception into RNA splicing patterns. In this study, a total of 1996 AS events were obviously related to LUSC patients' survival. Among the seven types of AS models, AD events showed the highest predictive power in survival prediction of LUSC patients than other six types of AS models. Moreover, we constructed a combined prognostic model composed of different splicing patterns of 14 genes including PLEKHG5, FANCA, ZMYND8 and so on. The combined prognostic model had higher predictive performance than any single type in seven AS models. In recent studies, some of these genes have been reported to exert carcinogenic or suppressive roles in cancers. For example, PLEKHG5 is a novel prognostic biomarker in glioma patients and could promote glioma migration and invasion [
36]. FANCA was a prognostic factor in LUAD [
37].
What's more, we identified some key SFs which might exert essential roles in the development and progression of cancers through modulating their corresponding AS events. With convincing results in this study that there was a positively correlation between most of poor prognosis-related AS events and SFs, while there was a negatively correlation between most of favorable prognosis-related AS events and SFs in LUSC, tt was well known that splicing factors could precisely regulate the splicing process by combining with specific genes’ the splice-regulatory sequence elements. This study provided an efficient approach to elucidate the underlying mechanism of AS events involved in survival of LUSC patients.
However, there are also some limitations in this study. Firstly, we could not verify the prognostic value of our combined prognostic model in other databases due to lack of available data. Secondly, the specific and valuable mechanisms of AS events and key SFs network need to be further explored. Finally, it is also need to further explore the interaction and regulation of AS events with other levels such as somatic mutation, copy number variation and DNA methylation.
In summary, we analyzed prognosis-related AS events and established AS signatures to predict survival of LUSC patients. Besides, an interesting splicing correlation network offered novel perceptions into how abnormal AS events were potentially modulated via pivotal SFs. These prognosis-related AS events and SFs provided us many valuable therapeutic targets for future validations and ultimately clarified the underlying mechanisms of AS in LUSC tumorigenesis.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.