Introduction
Hepatocellular carcinoma (HCC) is known to be the fifth most common cancer and is the third cause of cancer-related mortalities worldwide. Thus, it is considered a huge threat to people’s health [
1]. Surgical resection and liver transplantation are the most suitable methods for treatment of early-stage HCC. However, for non-resectable patients, radiofrequency ablation (RFA) and stereotactic body radiation therapy (SBRT) are deemed as first-line local treatment options [
2,
3]. Nonetheless, many patients with HCC treated with resection or local ablation show early relapse after treatment [
4]. In the past decade, immunotherapy has become popular for successfully treating various cancers [
5‐
8]. Interestingly, cumulative evidence revealed that immune-mediated mechanisms are deeply involved in the progression of HCC [
9]. This suggests the great potential of immune-based therapies for treating patients with HCC. However, few studies report only satisfactory immunotherapy efficacy for HCC. The failure of immunotherapy could be attributed to the complex composition of the tumor microenvironment (TME). Thus, it is important to understand the tumor immune microenvironment (TIME) for identifying pivotal therapeutic targets and prognostic biomarkers for HCC.
DDX1 is a member of the DEAD-box RNA helicase family and participates in various biological processes, such as mRNA translation, tRNA splicing, rRNA processing, microRNA maturation, and repair of DNA double-strand breaks (DSBs) [
10‐
14]. In addition to its role in transcriptional regulation, a recent study reported that DDX1 interacts with eIF3a and eIF4b to decrease insulin translation [
15]. Furthermore, some studies found that DDX1 can inhibit viral replication [
16,
17]. Moreover, researchers have found that DDX1 promotes tumorigenesis in various carcinomas, such as retinoblastoma, neuroblastoma, testicular carcinoma, colorectal cancer, and breast cancer [
18‐
21]. Nevertheless, the underlying mechanism of DDX1 in the progression of HCC remains obscure. In this research, we tried to revealed the potential role of DDX1 in the TIME of HCC.
Materials and methods
Data collection and analysis
We downloaded the raw RNA sequence data and associated clinical data of patients with HCC from the The Cancer Genome Atlas(TCGA) database. We compared the differential expression of DDX1 between HCC and normal tissues by paired t-test and Mann–Whitney U test. Subsequently, an independent sample t-test was used to analyze the correlation of DDX1 expression with clinicopathologic characteristics, including gender, T stage, pathologic stage, histologic grade, age, alpha fetoprotein (AFP) level and Child-Pugh grade. Analysis and plots were carried out using R packages “rstatix,” “car,” “tidyverse,” “ggplot2,” and “reshape2.” A P value of < 0.05 was considered statistically significant.
Survival analysis and external validation
Based on the average DDX1 expression, the Kaplan Meier (KM) survival curve was used to analyze the prognosis of patients with HCC with different levels of DDX1 mRNA expression. The analyses were conducted using the “survival” and “survminer” packages. Furthermore, we downloaded the RNA-sequencing expression profiles and corresponding clinical information of the liver cancer dataset (RIKEN, JP) from the International Cancer Genome Consortium (ICGC) database. According to the level of DDX1 expression in patient samples, two groups were created. The log-rank test was used to evaluate the difference in overall survival (OS) between the groups. The timeROC analysis was used to compare the predictive accuracy. Log-rank test and Cox regression methods were used to calculate the KM curve P values and hazard ratio (HR) with 95% confidence (CI) intervals.
Genetic mutation and methylation analysis
The cBioPortal (
www.cbioportal.org) is an online exploratory tool for visualizing and analyzing cancer genomic data in multidimensional ways [
22]. In this study, we used three different datasets (INSERM, Nat Genet 2015; AMC, Hepatology 2014; TCGA, Firehose Legacy) in cBioPortal to characterize the genomic profiles of DDX1. We categorized the patients into two groups (no mutation and mutation of DDX1) and used the Kaplan–Meier survival curves for evaluating the effect of mutations on survival.
Metasurv (
https://biit.cs.ut.ee/methsurv/) provides survival analysis based on CpG methylation patterns [
23]. To gain a comprehensive understanding of methylation and prognosis, we accessed the survival rate of the CpG methylation sites in DDX1. Survival was expressed using a Kaplan–Meier (KM) curve. In addition, the UALCAN (ualcan.path.uab.edu/index.html) database was employed to access the correlation of promoter methylation level of DDX1 in HCC with clinicopathological characteristics. Statistical significance was defined as a difference of < 0.05.
Correlation between DDX1 expression and immune infiltration
With the use of QUANTISEQ algorithm, we determined the association between DDX1 expression and tumor-infiltrating immune cells (TIICs), which including CD4
+ T cells, CD8
+ T cells, macrophages and so on. The R package “immunedeconv” was used to reliably estimate immune infiltration [
24].
TIMER (
https://cistrome.shinyapps.io/timer/) is a public website that analyzes the abundance of TIICs and gene expression across various cancers [
25]. According to TIMER dataset, the expression level of DDX1 was associated with the degree of TIICs, including as CD8
+ T cells, CD4
+ T cells,macrophages and so on. The above results were executed by the Gene module and visualized by scatter plots. We measured the correlation between TIICs and DDX1 expression using Spearman’s correlation.
Biological function enrichment analysis
The differentially expressed mRNAs were screened using the Limma package based on the RNA sequence and associated clinical information downloaded from the TCGA database. To correct the false-positive results, the adjusted P-value was analyzed. We set the thresholds for differentially expressed genes (DEGs) as adjusted
P-value of < 0.05 and |log2 fold change| ≥ 1.5. Kyoto Encyclopedia of Genes and Genomes (KEGG;
https://www.kegg.jp/) is a sophisticated database resource for the systematic analysis of gene functions, which links genomic information with higher order functional information [
26‐
28]. Subsequently, the ClusterProfiler R package was utilized to analyze the pathways enriched in DEGs based on Gene Ontology (GO) and KEGG databases.
Protein interaction screening
STRING (version 11.5) is an online tool that evaluates the interactions of genes [
29]. In this study, the STRING dataset was used to search co-expressed genes and construct Protein-Protein Interaction Networks(PPI) networks with an interaction score ≥ 0.15. We identified the top 50 DDX1-interacted proteins. Moreover, using the Similar Genes Detection module of GEPIA (
http://gepia.cancer-pku.cn/index.html), we identified the top 100 DDX1-related genes in liver cancer samples from TCGA. We then utilized Online tool (
http://bioinformatics.psb.ugent.be/webtools/Venn/) to carry out an intersection analysis between the top 50 DDX1-binding proteins and the top 100 genes related to DDX1 expression. Finally, we investigated the association of DDX1 with the selected genes via the Correlation Analysis module of GEPIA. Results were presented using scatter plots and Pearson correlation coefficient.
GSEA analysis
RNA-seq data collected from the TCGA database were analyzed using Gene Set Enrichment Analysis (GSEA) to preliminarily classify the genes based on their correlation with DDX1 expression. By GSEA, we carried out a KEGG enrichment analysis to explore the underlying biological roles of DDX1. A false discovery rate (FDR) < 0.05 and a nominal P-value < 0.05 were considered statistically significant.
Discussion
The DEAD-box RNA helicase family is known to play critical roles in various RNA metabolic processes. Recent studies indicate that DDX members are dysregulated in multiple cancers and function as key players in tumor progression [
30]. Hu et al. identified that DDX51 regulates cellular proliferation in esophageal squamous cell carcinoma through the PI3K/AKT/mTOR pathway [
31]. Jiang et al. proved that knock-down of DDX46 caused a significant reduction in cell invasion and migration in osteosarcoma [
32]. DDX1 activates the transcription of 12p stem cell genes in testicular tumorigenesis [
19]. Tanaka et al. proved the role of DDX1 in promoting colorectal tumorigenesis in vitro and in vivo [
20]. Researchers have proved that high DDX1 expression is associated with improvement in local control, distant metastatic-free survival, and OS when compared with low DDX1 expression in node-negative and early-stage patients with breast cancer [
33]. However, the role of DDX1 in HCC is not yet determined. Therefore, in this study, we focused on determining the potential prognostic value of DDX1 in HCC.
We acquired HCC samples from the TCGA database to explore the role of DDX1 in HCC progression. The results reveal that high DDX1 expression is associated with a poor prognosis of HCC. On analyzing the relationship between DDX1 and clinicopathologic features we found that DDX1 may serve as an adverse prognostic factor in HCC. Patients with high DDX1 expression may present an advanced T stage, histologic grade, and pathological stage.
According to molecular profiles and clinical outcomes, HCC has a highly heterogeneous nature, which present a formidable challenge to an accurate diagnosis and treatment [
34]. Luckily, previous studies have proved that molecular subtype stratification could overcome the hurdles caused by tumor heterogeneity [
35,
36]. In this research, 240 HCC samples downloaded from ICGC dataset were divided into high and low DDX1-expression subgroups, the KM survival analysis revealed that patients with high levels of DDX1 are at higher risk of suffering a poor prognosis. This results were consistent with the findings from TCGA dataset.
As liver carcinoma is typically associated with high malignancy, it is usually diagnosed at a late stage, posing a challenge for radical surgery. In order to improve the outcome of HCC, early diagnosis and treatment are crucial. As we known, tumorigenesis has been demonstrated to result from multiple gene mutations, patients with gene mutations are known to have a poor prognosis, In this research,we accessed the alteration percentage of DDX1 in HCC. After analyzing three independent datasets, we found the percentage of DDX1 alteration to be 1.9%. Furthermore, no significant differences was observed in OS between the mutation and unmutation DDX1 groups. In addition to gene mutations, numerous epigenetic changes, such as DNA methylation and histone modifications, contribute to tumor development [
37‐
39]. Nowadays, since DNA methylation is tissue-specific and stable, detecting abnormal DNA methylation in liquid biopsies has been shown to be a promising biomarker for cancer diagnosis [
40,
41]. A previous study indicated that HCC is caused by DNA methylation [
42]. For example, Kuo et al. found that patients with higher IRAK3 methylation had worse prognosis [
43]. However, no study has examined the relationship between DDX1 methylation and the oncologic outcome of HCC. Hence, in this study, we carried out methylation analysis and evaluated the correlation of methylation level with prognosis. We identified six CpG sites and two of them with hypermethylation were associated with poor prognosis. Besides, a previous study demonstrated that the hypomethylation status of oncogenes could also promote tumor development [
44]. In this study, using UALCAN we found that HCC tissues had lower levels of DDX1 promoter methylation than normal tissues (
P < 0.05). Further analysis indicated that the high tumor grade was linked to low promoter methylation levels. Thus, DDX1 methylation examination has the potential to be developed as a screening tool for predicting tumor status and progression; however, further in vivo and in vitro experiments are needed.
TME is a complicated assembly of the tumor, immune, stromal, and extracellular components [
45]. Previous studies revealed that TME facilitates the progression of HCC, thus indicating that it could be exploited as a potential therapeutic target [
46]. The importance of immune cell infiltration in the TME has been recognized for various solid tumors [
47,
48]. In this study, we determined the association between DDX1 expression and immune cell infiltration in HCC through the TIMER database. The results revealed that DDX1 is positively associated with dendritic cells, B cells, macrophages, and T cells. In addition, we verified the positive correlation between DDX1 expression and immune cell infiltration using the QUANTISEQ algorithm. M2 macrophages and Tregs were higher in the high DDX1 expression group compared with the low DDX1 expression group. Several studies indicate that macrophages are essential components of the TME and play key roles in tumor progression [
49]. Liu et al. [
50] found that M2 macrophages target miR-149-5p/MMP9 signaling pathway thereby facilitating HCC progression. Yeung et al. demonstrated that M2 macrophages are related to the adverse prognosis of liver cancer and promote HCC invasion by promoting Epithelial mesenchymal transformation(EMT) [
51]. Another research revealed that M2-polarized macrophages promote EMT of HCC cells and accelerate tumor progression through the TLR4/STAT3 signaling pathway [
52]. Similarly, Tregs are also critical components of TME. Jiang et al. proved that Tregs are closely associated with the prognosis of HCC. They secrete TGF-β1 which triggers EMT, thereby enhancing tumor invasiveness [
53]. Shen et al. [
54] put forth that TGF-β could drive Treg cell polarization to promote the progression of HCC. These results suggest that DDX1 plays a vital role in the TIME of HCC. However, additional experimental and theoretical studies are needed to validate the relationship between DDX1 and tumor infiltration.
To explore the potential biological functions of DDX1, we identified the DEGs in HCC samples and conducted GO and KEGG enrichment analyses. The results showed that DDX1 was mainly related to cell cycle and RNA editing. In addition, since DEAD box proteins function by interacting with other proteins, they are susceptible to being regulated by their partners and their microenvironment [
55]. Hence, using STRING and GEPIA database, we extracted 50 DDX1-binding proteins and the top 100 DDX1-related genes. Intersection analysis of the two groups identified HNRNPU, TARDBP, and HNRNPK as the common hits. TARDBP is an RNA-binding protein involved in the cell cycle of HCC tumor cells, and its expression level is related to an advanced stage and high grade of HCC [
56]. The CDK2 protein is known as an essential role in cell cycle regulation [
57]. Liang et al. have found HNRNPU enhances CDK2 transcription, thereby promoting HCC development [
58]. GSEA analysis also identified significant KEGG pathways associated with the cell cycle.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.