LncRNA DIRC1 is a novel prognostic biomarker and correlated with immune infiltrates in stomach adenocarcinoma : Medicine

Journal Logo

Research Article: Observational Study

LncRNA DIRC1 is a novel prognostic biomarker and correlated with immune infiltrates in stomach adenocarcinoma

Lin, Yuning Master’s degreea; Zhang, Zhongying Bachelor’s degreea,*; Li, Ying Master’s degreeb; Chen, Yongquan Master’s degreea; Su, Meiying Bachelor’s degreea; Zhao, Wenzhen Bachelor’s degreea

Author Information
doi: 10.1097/MD.0000000000031839

Abstract

1. Introduction

Stomach adenocarcinoma (STAD) is a common malignant tumor in the world and poses a serious threat to human health,[1] and STAD is 1 of the world’s top 5 cancers and a leading cause of cancer-related deaths, regardless of country development.[2] Since STAD mostly exists as nonspecific symptoms in the early stage of the disease, STAD is usually discovered after the disease progresses to a more serious state, and the high mortality and poor prognosis are due to metastasis, intra-tumor heterogeneity and chemotherapy resistance.[3,4] STAD has high mortality and poor prognosis, and the prognosis of STAD is closely related to clinical stage. The 5-years survival rate of early-stage STAD is 90%, while the 5-years survival rate of middle-advanced STAD is <30%.[5] Due to China’s dietary characteristics, lack of screening awareness and other reasons, the disease may have developed to an advanced stage by the time it is tested.[6]

At present, the early diagnosis of gastric cancer mainly relies on the pathological diagnosis of gastroscopy and biopsy tissue. The procedure is complicated and expensive, and the invasive procedure brings great pain to the subject, thus limiting the application of gastric cancer screening and early diagnosis in large-scale community population. Previous chemotherapy-based treatments only extended the median overall survival of patients with advanced STAD by 7 to 11 months.[7] Patients with early STAD have a good prognosis, but patients with advanced STAD have a poor prognosis due to the lack of effective targeted drugs and the tendency to develop drug resistance. Currently, only trastuzumab, ramucirumab, apatinib and Papalizumab have been approved for the targeted treatment of advanced STAD. The clinical application of these targeted agents is challenging.[8] In recent years, with the development of molecular biology, genetics and other disciplines and high-throughput omics technology, it is possible to explore the biomarkers related to the early diagnosis of gastric cancer from the multi-omics level, providing important evidence for the prevention of gastric cancer.

It was found that in the mammalian genome, < 2% of the transcripts have protein coding functions, and the remaining 98% are non coding RNAs. According to the length of nucleotide sequence, non coding RNAs can be divided into short chain ncRNAs and long chain ncRNAs.[9,10] Long non-coding RNAs (LncRNAs) are ≥ 200 nucleotides in length and do not encode proteins. According to its position and background in the genome, lncRNA can be divided into 5 main types: intergenic lncRNAs, intragenic lncRNAs, bidirectional lncRNAs, sense lncRNAs and antisense lncRNAs.[11] The mechanisms of lncRNA regulating gene expression mainly include transcriptional repression, RNA–DNA interaction (chromatin remodeling), nuclear RNA–RNA interaction and cytoplasmic RNA–RNA interaction. Their functions are to regulate a series of cellular biological processes, including chromatin remodeling, transcriptional and post-transcriptional events.[12,13] The most recognized molecular mechanism of lncRNAs is to act as a miRNA “sponge” to regulate downstream target genes.[14,15] LncRNAs is abnormally expressed in various types of cancer cells and plays an important role in several common hallmarks of cancer.[16]Wu et al demonstrated that the upregulation of lncRNA MER52A is associated with poor prognosis in hepatocellular carcinoma which can serve as a therapeutic target.[17] Lang et al demonstrated that the lncRNA PCAT7 functions as an oncogenic lncRNA associated with bone metastatic status and as a potential therapeutic target for prostate cancer bone metastasis.[18] In addition, some limited studies have shown that different lncRNAs are involved in human immune processes against cancer, such as immune activation, immune cell infiltration, and immune escape.[19–23] Currently, many studies have shown that some lncRNAs can serve as potential biomarkers for the diagnosis, treatment or prognosis of STAD.[24–26]

Disrupted in renal cancer 1 (DIRC1), initially characterized as a chromosome 2q33 breakpoint-spanning gene in a chromosomal translocation, has been closely associated with the development of renal cancer previously. Up to now, there are few studies on DIRC1.The study of Li et al showed that the knockdown of DIRC1 may reduce the carcinogenicity of Gastric cancer by inhibiting Akt/mTOR signal, and provide a potential therapeutic target for the treatment of Gastric cancer.[27] In the study of T druck et al, the low-level expression of DIRC1 was detected in adult placenta, testis, ovary and prostate, as well as fetal kidney, spleen and skeletal muscle through real time-polymerase chain reaction (RT-PCR), but the function of DIRC1 was not further explored.[28]

This study aimed to elucidate the association between lncRNA DIRC1 and STAD using the cancer genome atlas (TCGA) database. We found that the expression level of DIRC1 in STAD tissues was significantly higher than that in adjacent tissues. High expression of DIRC1 is associated with some clinicopathological features. Kaplan-Meier analysis showed that patients with high DIRC1 expression had lower overall survival than patients with low DIRC1 expression. These results suggest that lncRNA DIRC1 may be an independent biomarker for poor prognosis in STAD.

2. Methods

2.1. RNA-sequencing data and bioinformatics analysis

We used TCGA database (https://portal.gdc.cancer.gov/) to collect RNA-seq data and clinical information from 407 cases of STAD projects, including 27 cases with matched adjacent tissues. The downloaded data format was level 3 HTSeq- fragments per kilobase per million and then was converted into transcripts per million (TPM) format for subsequent analysis. We also download TPM format RNA-seq data in TCGA and genotype-tissue expression (GTEx) database that uniformly processed by Toil process from UCSC Xena (https://xenabrowser.net/datapages/).[29]All procedures performed in this study were in accordance with the Declaration of Helsinki (as revised in 2013). We used R package (DESeq2) to go differential analysis of DIRC1 expression, adjusted P value < 0.05 and |logFC| > 1.5 were consider as cut off criteria, the differentially expressed genes (DEGs) obtained were used for gene oncology (GO), Kyoto encyclopedia of genes and genomes (KEGG) analysis. And adjusted P < .05 and |logFC| > 1.0 were consider as another cut off criteria, the DEGs obtained were used for gene set enrichment analysis (GSEA). According to the default statistical method, the process was repeated 1000 times for each analysis and selected h.all.v7.2.symbols.gmt in MSigDB Collections as the reference gene collection, false discovery rate (FDR) q-value < 0.25 and adjusted P adjust < 0.05 were considered to be significantly enriched.

2.2. Immune infiltration analysis by ssGSEA

The immune infiltration analysis of STAD was performed by single sample GSEA (ssGSEA) method from R (v.3.6.3) package GSVA (version 1.34.0), and we quantified the infiltration levels of 24 immune cell types from gene expression profile in the literature. In order to discover the correlation between DIRC1 and the infiltration levels of 24 immune cells, P values were determined by the Pearson and Wilcoxon rank sum test.

2.3. Protein–protein interaction analysis

Search tool for the retrieval of interacting genes (STRING) is an online database that searches for known proteins and predicts protein interaction relationships, including direct physical interactions between proteins and indirect functional correlations. The STRING database collects, evaluates, and integrates all publicly available protein-protein interaction information and complements this information with computational predictions to build a protein-protein interaction network. The software analyzes all DEGs; the interaction score threshold was set at .400.

2.4. Total RNA extraction, reverse transcription and RT-PCR

Total RNAs from all samples were isolated with magnetic bead nucleic acid extraction reagent (TIB, Xiamen, China). RNA reverse transcription was performed using transcription kit (TIAGEN, Beijing, China).RT-PCR was done with Power SYBR Green RT-PCR reagent (Tiagen, Beijing, China) and ABI Quant Studio fluorescence quantitative PCR instrument (Applied Biosystems Inc., Foster City, CA).Primers for DIRC1 and β-actin were show as follows:

DIRC1-F: CCAGCCATGCATGTCCACTA

DIRC1-R: AGTGTCAGGGGGAAAGGAGT

β-actin-F: GTGGATCAGCAAGCAGGAGT

β-actin-R: CTCGGCCACATTGTGAACTT

2.5. Statistical analysis

All statistical analyses were performed using R (v.3.6.3). Wilcoxon rank sum test, chi-square test, Fisher exact test and logistic regression were used to analyze the relationship between clinical pathologic features and DIRC1. Kaplan-Meier method was used to calculate the overall survival rate and progression free interval of STAD patients from TCGA. Univariate and multivariate analysis were performed to estimate the association between clinical and genetic clinical characteristics. Overall survival (OS) were analyzed by Cox proportional hazard models. P values < 0.05 were considered statistically significant. area under curve value between 0.5 and 0.7 was of low accuracy; between 0.7 and 0.9, of medium accuracy; and above 0.9, of high accuracy.

3. Results

3.1. DIRC1 expression is correlated with poor clinicopathological features of STAD

Downloaded RNA-seq data in TPM format from TCGA and GTEx was processed uniformly using the Toil process from XENA (https://xenabrowser.net/datapages/) by the University of California, Santa Cruz.[29] As shown in Figure 1A, the Wilcoxon rank sum test was used to compare the expression of DIRC1 in GTEx and normal TCGA samples with corresponding TCGA tumor samples.DIRC1 expression is significantly different in the following cancers: bladder urothelial carcinoma, breast-infiltrating carcinoma, colon cancer (STAD), diffuse large B-cell lymphoma, esophageal cancer, glioblastoma multiforme, head and neck squamous cell carcinoma, kidney renal clear cell carcinoma, acute myeloid leukemia, brain lower grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, lung squamous carcinoma, ovarian serous cystadenocarcinoma, pancreatic cancer, rectal adenocarcinoma, skin melanoma, stomach adenocarcinoma (STAD), testicular germ cell tumors, thyroid carcinoma, thymoma, uterine carcinosarcoma.

F1
Figure 1.:
The expression of lncRNA DIRC1 in Genotype-Tissue Expression (GTEx) and normal The Cancer Genome Atlas (TCGA) samples with corresponding TCGA tumor samples(A);in 375 gastric cancer tissues and 32 normal tissues in TCGA database(B);in 27 pairs of gastric cancer tissues and non-cancerous adjacent tissues in TCGA database(C);in GTEx and normal TCGA samples with TCGA STAD(D);Volcano plot of differentially expressed genes, normalized expression levels are shown in descending order from green to red. There 75 differential molecules had log2FC > 1.5 and adjusted P < .05, and 132 differential molecules had log2FC < −1.5 and adjusted P < .05(E). DIRC1 = disrupted in renal carcinoma 1, GTEx = genotype-tissue expression, lncRNA = long non-coding RNA, STAD = stomach adenocarcinoma, TCGA = the cancer genome atlas.

In order to identify the difference of DIRC1 expression between STAD and normal tissues, we analyzed the expression level of DIRC1 in 375 STAD tissues and 32 adjacent normal tissues, and found that DIRC1 was highly expressed in STAD tissues (P < .001, Fig. 1B). Meanwhile, we also analyzed the expression of DIRC1 in 27 STAD tissues and their matched adjacent tissues. The results indicated that STAD tissues highly expressed DIRC1 (P < .001, Fig. 1C). Moreover, the Wilcoxon rank sum test was used to compare the expression of DIRC1 in normal GTEx samples and TCGA STAD samples (Fig. 1D).

3.2. Identification of DEGs

According to the expression of lncRNA DIRC1 in tumor samples, qualified HTSeq-counts format data were divided into high and low expression groups according to the median value. Then, 207 DEGs were obtained using the DESeq2 package. |log2FC| > 1.5 and adjusted P < .05 were used as the screening threshold for the DEGs. Among them, 75 were upregulated, and 132 were downregulated (Fig. 1E).

3.3. Clinical characteristics

The characteristics of patients with STAD in TCGA were collected. According to the median expression of lncRNA DIRC1, 188 patients were assigned to the low-expression group, and 187 patients were assigned to the high-expression group. The χ2 test or Fisher’s exact test determined that lncRNA DIRC1 expression was significantly associated with T stage (P = .004), race (P = .045), histologic grade (P = .029) and anatomic neoplasm subdivision (P = .034). No correlation existed between lncRNA DIRC1 expression and the other clinicopathological features, as shown in Table 1.

Table 1 - Association between lncRNA DIRC1 expression and clinicopathological features in the validation cohort.
Characteristic Low expression of DIRC1 High expression of DIRC1 P Statistic Method
n 187 188
T stage, n (%) .004 13.43 χ2 test
 T1 17 (4.6%) 2 (0.5%)
 T2 43 (11.7%) 37 (10.1%)
 T3 77 (21%) 91 (24.8%)
 T4 49 (13.4%) 51 (13.9%)
N stage, n (%) .83 .88 χ2 test
 N0 58 (16.2%) 53 (14.8%)
 N1 45 (12.6%) 52 (14.6%)
 N2 39 (10.9%) 36 (10.1%)
 N3 38 (10.6%) 36 (10.1%)
M stage, n (%) 1 0 χ2 test
 M0 163 (45.9%) 167 (47%)
 M1 12 (3.4%) 13 (3.7%)
Pathologic stage, n (%) .073 6.97 χ2 test
 Stage I 35 (9.9%) 18 (5.1%)
 Stage II 50 (14.2%) 61 (17.3%)
 Stage III 77 (21.9%) 73 (20.7%)
 Stage IV 17 (4.8%) 21 (6%)
Primary therapy outcome, n (%) .132 Fisher test
 PD 40 (12.6%) 25 (7.9%)
 SD 11 (3.5%) 6 (1.9%)
 PR 1 (0.3%) 3 (0.9%)
 CR 113 (35.6%) 118 (37.2%)
Gender, n (%) .884 .02 χ2 test
 Female 68 (18.1%) 66 (17.6%)
 Male 119 (31.7%) 122 (32.5%)
Race, n (%) .045 6.2 χ2 test
 Asian 37 (11.5%) 37 (11.5%)
 Black or African American 9 (2.8%) 2 (0.6%)
 White 106 (32.8%) 132 (40.9%)
Age, n (%) .899 .02 χ2 test
 <=65 82 (22.1%) 82 (22.1%)
 >65 101 (27.2%) 106 (28.6%)
Histological type, n (%) .109 9 χ2 test
 Diffuse Type 27 (7.2%) 36 (9.6%)
 Mucinous Type 8 (2.1%) 11 (2.9%)
 Not Otherwise Specified 101 (27%) 106 (28.3%)
 Papillary Type 1 (0.3%) 4 (1.1%)
 Signet Ring Type 6 (1.6%) 5 (1.3%)
 Tubular Type 44 (11.8%) 25 (6.7%)
Residual tumor, n (%) .913 .18 χ2 test
 R0 155 (47.1%) 143 (43.5%)
 R1 7 (2.1%) 8 (2.4%)
 R2 8 (2.4%) 8 (2.4%)
Histologic grade, n (%) .029 Fisher test
 G1 4 (1.1%) 6 (1.6%)
 G2 80 (21.9%) 57 (15.6%)
 G3 97 (26.5%) 122 (33.3%)
Anatomic neoplasm subdivision, n (%) .034 10.43 χ2 test
 Antrum/Distal 61 (16.9%) 77 (21.3%)
 Cardia/Proximal 26 (7.2%) 22 (6.1%)
 Fundus/Body 63 (17.5%) 67 (18.6%)
 Gastroesophageal Junction 27 (7.5%) 14 (3.9%)
 Other 4 (1.1%) 0 (0%)
Reflux history, n (%) .334 .93 χ2 test
 No 90 (42.1%) 85 (39.7%)
 Yes 24 (11.2%) 15 (7%)
Antireflux treatment, n (%) .584 .3 χ2 test
 No 71 (39.7%) 71 (39.7%)
 Yes 21 (11.7%) 16 (8.9%)
Hpylori infection, n (%) .464 .54 χ2 test
 No 90 (55.2%) 55 (33.7%)
 Yes 9 (5.5%) 9 (5.5%)
Barretts esophagus, n (%) .933 .01 χ2 test
 No 112 (53.8%) 81 (38.9%)
 Yes 8 (3.8%) 7 (3.4%)
CR = complete response, DIRC1 = disrupted in renal carcinoma 1, PR = partial response, SD = stable disease, PD = progressive disease.

Univariate logistic regression revealed that the increased lncRNA DIRC1 expression was related to poor prognostic clinicopathological characteristics, including a greater primary tumor extent (odds ratio [OR] = 1.734; 95% CI, 1.088–2.787) for T3 and T4 stages versus T1 and T2 stages (P = .021), histological type (OR = 2.187; 95% CI, 1.124‐4.321) for diffuse type & signet ring type versus tubular type (P = .022), histologic grade (OR = 1.677; 95% CI, 1.102‐2.563) for G3 versus G1 & G2 (P = .016), as shown in Table 2.The results showed that Stomach adenocarcinoma with increased expression of lncRNA DIRC1 is prone to adverse clinicopathological factors.

Table 2 - Association of lncRNA DIRC1 expression with clinical pathological characteristics by logistic regression.
Characteristics Total (N) OR P value
T stage (T3&T4 vs T1&T2) 367 1.734 (1.088‐2.787) .021
N stage (N1&N2&N3 vs N0) 357 1.112 (0.710‐1.744) .642
M stage (M1 vs M0) 355 1.057 (0.466‐2.417) .893
Pathologic stage (Stage III&Stage IV vs Stage I&Stage II) 352 1.076 (0.708‐1.637) .732
Primary therapy outcome (PD&SD&PR vs CR) 317 0.626 (0.376‐1.032) .068
Histological type (Diffuse Type&Signet Ring Type vs Tubular Type) 143 2.187 (1.124‐4.321) .022
Anatomic neoplasm subdivision (Fundus/Body&Gastroesophageal Junction&Other vs Antrum/Distal&Cardia/Proximal) 361 0.757 (0.500‐1.145) .188
Antireflux treatment (Yes vs No) 179 0.762 (0.363‐1.574) .465
Hpylori infection (Yes vs No) 163 1.636 (0.604‐4.436) .326
Barretts esophagus (Yes vs No) 208 1.210 (0.409‐3.502) .723
Histologic grade (G3 vs G1&G2) 366 1.677 (1.102‐2.563) .016
Reflux history (Yes vs No) 214 0.662 (0.320‐1.334) .254
Residual tumor (R1&R2 vs R0) 329 1.156 (0.549‐2.445) .701
DIRC1 = disrupted in renal carcinoma 1, lncRNA = long non-coding RNA, PD = progressive disease, PR = partial response, OR = odds ratio, SD = stable disease.

3.4. ROC differentiates normal tissue from tumor tissue

The data from para-carcinoma tissue of patients and carcinoma tissue of patients were applied to draw the receiver operating characteristic curve and evaluate the diagnostic value of lncRNA DIRC1. Its area under curve was .779, predicting a very efficient discrimination value for Stomach adenocarcinoma (Fig. 2A).

F2
Figure 2.:
(A) Receiver operating characteristic (ROC) analysis of lncRNA DIRC1 expression showing promising discrimination power between non-tumor and tumor tissues. (B)Prognostic value of DIRC1 in Overall Survival of STAD patients. Statistically significant subgroups were (C) T3; (D) N2; (E) M0; and (F) Histologic grade G2. DIRC1 = disrupted in renal carcinoma 1, lncRNA = long non-coding RNA, ROC = receiver operating characteristic, STAD = stomach adenocarcinoma.

3.5. Role of lncRNA DIRC1 in gastric adenocarcinoma survival (Including clinicopathological subgroup analysis)

Kaplan-Meier survival analysis showed that a high expression level of lncRNA DIRC1 was significantly correlated with a poorer OS of patients hazard ratio (HR) = 1.63; 95% confidence interval (CI), 1.17‐2.27; P = .004) (Fig. 2B).In these subgroup analysis, the T3 subgroup of DIRC1 for the T stage was statistically significant (HR = 1.67; 95% CI, 1.04–2.68; P = .033) (Fig. 2C), the N2 subgroup for the N stage was statistically significant (HR = 3.01; 95% CI, 1.33–6.83; P = .004) (Fig. 2D), and the M0 subgroup for the M stage was statistically significant (HR = 1.79; 95% CI, 1.25–2.57; P = .002) (Fig. 2E). Furthermore, the subgroups of G2 histologic grade had statistical significance (HR = 1.95; 95% CI, 1.08–3.54; P = .028) (Fig. 2F).

3.6. LncRNA DIRC1 related GSEA analysis

GSEA was used to identify DIRC1-related signaling pathways. GSEA revealed significant differences (Padj < 0.05, FDR < 0.25) in enrichment of MSigDB Collection (h.all.v7.2.symbols.gmt [Hallmarks]).[30] A total of 27 data sets met the requirements of FDR < 0.25 and adjusted P < .05. This analysis revealed that, in the lncRNA DIRC1 high-expression phenotype, 19 pathways were significantly differentially enriched. In addition, 8 pathways in the lncRNA DIRC1 low-expression phenotype were recognized (Table S1, Supplemental Digital Content, https://links.lww.com/MD/H952).We selected the top 6 data sets with high value of normalized enrichment score in the lncRNA DIRC1 high-expression phenotype and top 3 data sets with high value of normalized enrichment score in the lncRNA DIRC1 low-expression phenotype (Fig. 3A‐I).

F3
Figure 3.:
Functional enrichment analysis of 207 differentially expressed genes (DEGs) between high and low expression of lncRNA DIRC1 in patients with STAD in TCGA. The data set was on the left significantly enriched in red area (A-F, DIRC1 high expression group). The data set was on the right significantly enriched in blue area (G-I,DIRC1 low expression group).NES, normalized NS; Padj, adjust P value; FDR, false discovery rate. DEGs = differentially expressed genes, DIRC1 = disrupted in renal carcinoma 1, FDR = false discovery rate, lncRNA = long non-coding RNA, NES = normalized enrichment score, STAD = stomach adenocarcinoma, TCGA = the cancer genome atlas.

3.7. DIRC1 related GO and KEGG analysis

To estimate the potential functions of DEGs in high-risk versus (vs) low-risk groups, we identify DEGs of DIRC1 in TCGA-STAD data under cutoff criteria of adjusted P value < 0.05 and |logFC|>1.5. KEGG pathway and GO annotation were performed by R package cluster Profiler (3.14.3). GO reveals the catalogs of biological process, cellular component, and molecular function. After multiple-test correction, KEGG pathways and GO terms with corrected P (P adjust) value < 0.05 were considered to be prominently enriched in DEGs. We selected top 9 of the lowest adj. P value of GO and KEGG pathway enrichment analysis of 207 DEGs related to DIRC1 in TCGA-STAD data (Fig. 4).

F4
Figure 4.:
Functional enrichment analysis of 207 differentially expressed genes (DEGs) between high and low expression of lncRNA DIRC1 in patients with STAD in TCGA. (A) Enriched Gene Ontology (GO) terms in the biological process category.(B) Enriched GO terms in the molecular function category.(C) Enriched GO terms in the cellular component category. (D) Enriched GO terms in the Kyoto Encyclopedia of Genes and Genomes (KEGG) category. DEGs = differentially expressed genes, DIRC1 = disrupted in renal carcinoma 1, GO = gene oncology, KEGG = Kyoto encyclopedia of genes and genomes, STAD = stomach adenocarcinoma, TCGA = the cancer genome atlas.

3.8. Immune infiltration analysis by ssGSEA

The immune infiltration analysis of STAD was performed by single sample GSEA (ssGSEA) method from R (v.3.6.3) package GSVA (version 1.34.0),[31] and we quantified the infiltration levels of 24 immune cell types from gene expression profile in the literature.[32] In order to discover the correlation between DIRC1 and the infiltration levels of 24 immune cells (Fig. 5A), P values were determined by the Spearman and Wilcoxon rank sum test LncRNA DIRC1 expression was significantly positively correlated with Macrophages, dendritic cells (DC), Th1 cells, Tem, immature dendritic cells and so on. Helper T17 (Th17) cells and nature kill (NK) CD56bright cells were negatively correlated with lncRNA DIRC1 expression (P < .05). As show in Figure 5B‐G, macrophages (R = 0.535, P < .001), DC (R = 0.392, P < .001), Th1 cells (R = 0.367, P < .001), T effector memory cells (R = 0.337, P < .001) showed positive association with DIRC1. However, NK CD56bright cells (r = −0.166, P = .001) and Th17 cells (r = −0.126, P < .014) showed negative association with DIRC1.

F5
Figure 5.:
The expression level of lncRNA DIRC1 was associated with immune infiltration in the tumor microenvironment. (A) Correlation between the relative abundances of 24 immune cells and lncRNA DIRC1 expression level. The size of dots shows the absolute value of Spearman. Correlation between the relative enrichment score of Macrophages (B), DC (C), Th1 cells (D), Tem (E), NK CD56bright cells (F), Th17 cells (G) and the expression level (TPM) of lncRNA DIRC1. DC = dendritic cells, DIRC1 = disrupted in renal carcinoma 1, NK = nature kill, TPM = transcripts per million.

3.9. PPI network and hub gene identification

The search tool (STRING, https://string-db.org/) is a database for finding interacting genes, with which we constructed a protein–protein interaction (PPI) network.[33] 145 intersection target genes were imported into the STRING database and the species were selected as “Homo sapiens” to obtain the interaction relationship between the targets. Genes with significant interactions were identified based on a confidence level≥0.4, other irrelevant genes were excluded. The screening results were input into Cytoscape software 3.8.2[33] for network visualization (Fig. 6).

F6
Figure 6.:
145 intersection target genes were imported into the STRING database and the species were selected as “Homo sapiens” to obtain the interaction relationship between the targets. STRING = search tool for the retrieval of interacting genes.

3.10. Verification of the prognostic value of 10 hub genes

The maximum clique centrality (MCC) value of each node is calculated by the CytoHubba plugin in Cytoscape. In this study, the 10 genes with the highest MCC values were selected as hub genes. In co-expression networks, the MCC algorithm is considered to be the most efficient method to identify central nodes.[34]

After the identification of 10 hub genes (ALB, ORM1, APOH, FGA, ITIH2, PAH, AHSG, FGG, TTR, ORM2) using the CytoHubba plug-in, we verified the prognostic value of these 10 hub genes with survival-related data in TCGA. The OS analysis of the 10 hub genes was performed with the Kaplan–Meier plotter using the R survival package. The analysis showed the expression level of ALB, ITIH1, APOH and FGG were significantly correlated with OS in STAD patients (P < .05; Fig. 7). Our results showed that the high expression of ALB, ITIH1, APOH and FGG indicate worse prognosis of patients with STAD.

F7
Figure 7.:
(A) Identification of the hub genes from the PPI network by the maximum clique centrality (MCC) algorithm, the red nodes represent genes with the highest MCC sores. (B) Overall survival of ALB associated with STAD. (C) Overall survival of ITIH1 associated with STAD. (D) Overall survival of APOH associated with STAD. (E) Overall survival of FGG associated with STAD. MCC = maximum clique centrality, PPI = Protein-protein interaction, STAD = stomach adenocarcinoma.

3.11. Relative expression of DIRC1 in the serum of gastric cancer patients and normal people

PT-PCR analysis showed that, the relative expression level is expressed as N (N = 2-Ct, Ct = Ct (DIRC1) – Ct (β-actin),the relative expression of DIRC1 in the serum of Gastric Cancer patients is was higher than that in normal people (P = .027, Fig. 8).

F8
Figure 8.:
The relative expression of DIRC1 in the serum of Gastric Cancer patients is was higher than that in normal people (P = .027). DIRC1 = disrupted in renal carcinoma 1.

4. Discussion

More recently, the understanding of lncRNAs has evolved to identify new insights into their involvement in disease pathogenesis. LncRNAs regulate gene expression through a variety of mechanisms, such as interactions with RNA or protein molecules. At present, many lncRNAs have been identified as important biomarkers of STAD. In general, lncRNAs exert regulatory functions at different levels of gene expression, including chromatin modification, transcription, and post-transcription.[35] lncRNAs can interact with chromatin remodeling complexes to induce heterochromatin formation at specific genomic sites and reduce gene expression. In addition, lncRNAs interact with RNA-binding proteins and transcription factor co-activators, or regulate transcription by regulating the main promoters of their target genes. Mechanically, lncRNAs can communicate with DNA, mRNAs, ncRNAs and proteins and play cancer-related regulatory roles, such as, such as signals, decoys, scaffolds and guidelines.[36,37]

In this study, we collected and organized STAD data using high-throughput RNA sequencing from TCGA database, and we verified that lncRNA DIRC1 was significantly upregulated in STAD tissues compared with in adjacent normal or normal tissues.

Moreover, analyzing the relationship between the clinicopathological features of STAD and the dichotomy of high and low DIRC1 levels by using the logistic regression method, we showed that DIRC1 was also significantly correlated with T stage, histological grade, race and Anatomic neoplasm subdivision.

Upregulated lncRNA DIRC1 in STAD tissues was positively correlated with higher T stage; advanced histological grade; and poorer overall survival. Elevated lncRNA DIRC1 was related to advanced clinicopathological features. These results suggest that lncRNA DIRC1 might be an independent biomarker of poor outcomes for STAD.

We also used GSEA to study the function of lncRNA DIRC1 in STAD tissues, and the results showed that in the lncRNA DIRC1 high-expression phenotype,19 pathways were significantly differentially enriched, including the E2F target, G2M checkpoint, MYC Target, INF-α reaction, INF-γ reaction, MTORC1 signal and so on. In addition, 8 pathways in the lncRNA DIRC1 low-expression phenotype were recognized, including Epithelial mesenchymal transition, early response to estrogenas, down-regulated in response to ultraviolet (UV) radiation and so on. Epithelial mesenchymal transformation (EMT) is essential in development, wound healing, and stem cell behavior, and contributes pathologically to fibrosis and cancer progression.[38]Entry into mitosis is regulated by checkpoints at the G2 and M (G2/M) boundaries of the cell cycle, and misregulation of entry into mitosis usually leads to tumorigenesis or cell death.[39]

In this study, ssGSEA and Spearman correlation were used to reveal the relationship between lncRNA DIRC1 expression and immune infiltration level in STAD. We found the strongest relationship between lncRNA DIRC1 and Macrophages, DC, Th1 cells and Tem. In contrast, the levels of NK CD56 bright cells and Th17 cells were negatively correlated with lncRNA DIRC1 expression. Therefore, LncRNA DIRC1 may play a major role in immune cell infiltration and serve as a prognostic biomarker of STAD.

To discover the molecular significance of DIRC1, coregulatory proteins were included in the PPI network analysis. According to the MCC scores of the CytoHubba plug-in in Cytoscape, the first 10 genes (ALB, ORM1, APOH, FGA, ITIH2, PAH, AHSG, FGG, TTR, ORM2) related to DIRC1 were screened. Among the 10 hub genes associated with DIRC1, the high expression of 4 genes (ALB, ITIH1, APOH and FGG) are related to THE OS of STAD. Therefore, we believe that DIRC1 has a good prognostic value for STAD.

To confirm the relationship between lncRNA DIRC1 and overall STAD survival, Kaplan–Meier survival analysis was performed for stratified clinicopathological features. Kaplan-Meier survival analysis showed a significant correlation between lncRNA DIRC1 expression level and overall survival of T3, N2, M0 and G2 histological grade, suggesting that lncRNA DIRC1 expression level remains a strong predictor of prognosis in these subpopulations.

Although this study improves our understanding of the association between lncRNA DIRC1 and STAD, some limitations remain. First, in order to fully clarify the special role of lncRNA DIRC1 in the development and progression of STAD, all clinical factors, such as details of the patient’s treatment process, such as surgical treatment, chemotherapy or radiotherapy, should be included. However, such information is lacking or inconsistently processed in public databases. Second, this study only provides biological information analysis and a small amount of experimental verification, and in 18 Gastric cancer samples, it does not distinguish the types of Gastric cancer such as squamous cell carcinoma and adenocarcinoma, which has certain limitations. More blood samples, cell samples and tissue samples are needed for RT-PCR and immunohistochemical analysis to further study the function and mechanism of lncRNA DIRC1 from multiple perspectives. Third, the understanding of gene function by single omics is not comprehensive, so it should be extended to multi-omics research, especially the study of protein level and its functional mechanism. Fourth, the lack of external data set validation can lead to bias. Finally, retrospective study has its limitations. Prospective studies must be conducted in the future. In this study, we found that lncRNA DIRC1 was an independent predictor of lower overall STAD survival.

5. Conclusion

LncRNA DIRC1 expression was significantly correlated with poor survival and immune infiltrations in STAD, and it may be a promising prognostic biomarker in STAD. Furthermore, these findings provide clues to further explore the possible role of lncRNA DIRC1 in STAD.

Acknowledgments

Xiantao Academic provides technical support for R analysis.

Author contributions

Conceptualization: Yuning Lin, Ying Li.

Data curation: Yuning Lin, Ying Li.

Formal analysis: Yuning Lin, Ying Li, Yongquan Chen.

Funding acquisition: Yuning Lin.

Investigation: Yuning Lin, Ying Li.

Methodology: Yuning Lin, Ying Li.

Project administration: Yuning Lin.

Resources: Yuning Lin, Zhongying Zhang, Yongquan Chen.

Software: Yuning Lin, Zhongying Zhang.

Supervision: Yuning Lin, Zhongying Zhang, Yongquan Chen.

Validation: Yuning Lin, Zhongying Zhang, Yongquan Chen, Meiying Su, Wenzhen Zhao.

Visualization: Yuning Lin, Zhongying Zhang, Yongquan Chen.

Writing – original draft: Yuning Lin, Zhongying Zhang, Ying Li.

Writing – review & editing: Yuning Lin, Zhongying Zhang.

    References

    [1]. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–49.
    [2]. Pearce A, Sharp L, Hanly P, et al. Productivity losses due to premature mortality from cancer in Brazil, Russia, India, China, and South Africa (BRICS): a population-based comparison. Cancer Epidemiol. 2018;53:27–34.
    [3]. Allemani C, Matsuda T, Di Carlo V, et al. Global surveillance of trends in cancer survival 2000-14 (CONCORD-3): analysis of individual records for 37 513 025 patients diagnosed with one of 18 cancers from 322 population-based registries in 71 countries. Lancet. 2018;391:1023–75.
    [4]. Petryszyn P, Chapelle N, Matysiak-Budnik T. Gastric cancer: where are we heading? Dig Dis. 2020;38:280–5.
    [5]. Isobe Y, Nashimoto A, Akazawa K, et al. Gastric cancer treatment in Japan: 2008 annual report of the JGCA nationwide registry. Gastric Cancer. 2011;14:301–16.
    [6]. Ferlay J, Colombet M, Soerjomataram I, et al. Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods. Int J Cancer. 2019;144:1941–53.
    [7]. Sano T. [Evaluation of the gastric cancer treatment guidelines of the Japanese gastric cancer association]. Gan To Kagaku Ryoho. 2010;37:582–6.
    [8]. Catenacci DVT, Tebbutt NC, Davidenko I, et al. Rilotumumab plus epirubicin, cisplatin, and capecitabine as first-line therapy in advanced MET-positive gastric or gastro-oesophageal junction cancer (RILOMET-1): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2017;18:1467–82.
    [9]. Qi P, Du X. The long non-coding RNAs, a new cancer diagnostic and therapeutic gold mine. Mod Pathol. 2013;26:155–65.
    [10]. Song H, Yu X, Tian X, et al. Associations between long non-coding RNAs and tumors, and their clinical values. Chin J Cell Biol. 2012;34:704–12.
    [11]. Izant JG, Weintraub H. Inhibition of thymidine kinase gene expression by anti-sense RNA: a molecular approach to genetic analysis. Cell. 1984;36:1007–15.
    [12]. Li Y, Zhou L, Lu C, et al. Long non-coding RNA FAL1 functions as a ceRNA to antagonize the effect of miR-637 on the down-regulation of AKT1 in Hirschsprung’s disease. Cell Prolif. 2018;51:e12489.
    [13]. Raveendra BL, Swarnkar S, Avchalumov Y, et al. Long noncoding RNA GM12371 acts as a transcriptional regulator of synapse function. Proc Natl Acad Sci USA. 2018;115:E10197–205.
    [14]. Dong Z, Zhang A, Liu S, et al. Aberrant methylation-mediated silencing of lncRNA MEG3 functions as a ceRNA in esophageal cancer. Mol Cancer Res. 2017;15:800–10.
    [15]. Liu T, Chi H, Chen J, et al. Curcumin suppresses proliferation and in vitro invasion of human prostate cancer stem cells by ceRNA effect of miR-145 and lncRNA-ROR. Gene. 2017;631:29–38.
    [16]. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100:57–70.
    [17]. Wu Y, Zhao Y, Lin H, et al. An LTR retrotransposon-derived long noncoding RNA lncMER52A promotes hepatocellular carcinoma progression by binding p120-catenin. Cancer Res. 2020;80:976–87.
    [18]. Lang C, Dai Y, Wu Z, et al. SMAD3/SP1 complex-mediated constitutive active loop between lncRNA PCAT7 and TGF-β signaling promotes prostate cancer bone metastasis. Mol Oncol. 2020;14:808–28.
    [19]. Meng Z, Zhao H, Xu W, et al. Discovery and validation of immune-associated long non-coding RNA biomarkers associated with clinically molecular subtype and prognosis in diffuse large B cell lymphoma. Mol Cancer. 2017;16:16.
    [20]. Zhang P, Cao L, Zhou R, et al. The lncRNA Neat1 promotes activation of inflammasomes in macrophages. Nat Commun. 2019;10:1495.
    [21]. Wang CJ, Zhu CC, Xu J, et al. The lncRNA UCA1 promotes proliferation, migration, immune escape and inhibits apoptosis in gastric cancer by sponging anti-tumor miRNAs. Mol Cancer. 2019;18:115.
    [22]. Zheng Y, Tian X, Wang T, et al. Long noncoding RNA Pvt1 regulates the immunosuppression activity of granulocytic myeloid-derived suppressor cells in tumor-bearing mice. Mol Cancer. 2019;18:61.
    [23]. Wang Z, Cao L, Zhou S, et al. Construction and validation of a novel pyroptosis-related 4-lncRNA prognostic signature related to gastric cancer and immune infiltration. Front Immunol. 2022;13:854785.
    [24]. Zha Z, Zhang P, Li D, et al. Identification and construction of a long noncoding RNA prognostic risk model for stomach adenocarcinoma patients. Dis Markers. 2021;2021:8895723.
    [25]. Zhang C, Ren X, Zhang W, et al. Prognostic and clinical significance of long non-coding RNA SNHG12 expression in various cancers. Bioengineered. 2020;11:1112–23.
    [26]. Nie K, Deng Z, Zheng Z, et al. Identification of a 14-lncRNA signature and construction of a prognostic nomogram predicting overall survival of gastric cancer. DNA Cell Biol. 2020;39:1532–44.
    [27]. Li Z, Yang AJ, Wei FM, et al. Significant association of DIRC1 overexpression with tumor progression and poor prognosis in gastric cancer. Eur Rev Med Pharmacol Sci. 2018;22:8682–9.
    [28]. Druck T, Podolski J, Byrski T, et al. The DIRC1 gene at chromosome 2q33 spans a familial RCC-associated t(2;3)(q33;q21) chromosome translocation. J Hum Genet. 2001;46:583–9.
    [29]. Vivian J, Rao AA, Nothaft FA, et al. Toil enables reproducible, open source, big biomedical data analyses. Nat Biotechnol. 2017;35:314–6.
    [30]. Liberzon A, Birger C, Thorvaldsdóttir H, et al. The molecular signatures database (MSigDB) hallmark gene set collection. Cell Syst. 2015;1:417–25.
    [31]. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinf. 2013;14:7.
    [32]. Bindea G, Mlecnik B, Tosolini M, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity. 2013;39:782–95.
    [33]. Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47:D607–d613.
    [34]. Chin CH, Chen SH, Wu HH, et al. cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol. 2014;8(Suppl 4):S11.
    [35]. Wang L, Cho KB, Li Y, et al. Long noncoding RNA (lncRNA)-mediated competing endogenous RNA networks provide novel potential biomarkers and therapeutic targets for colorectal cancer. Int J Mol Sci. 2019;20:5758.
    [36]. Ørom UA, Derrien T, Beringer M, et al. Long noncoding RNAs with enhancer-like function in human cells. Cell. 2010;143:46–58.
    [37]. Wang KC, Chang HY. Molecular mechanisms of long noncoding RNAs. Mol Cell. 2011;43:904–14.
    [38]. Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial-mesenchymal transition. Nat Rev Mol Cell Biol. 2014;15:178–96.
    [39]. Yasutis KM, Kozminski KG. Cell cycle checkpoint regulators reach a zillion. Cell Cycle. 2013;12:1501–9.
    Keywords:

    biomarker; DIRC1; LncRNA; stomach adenocarcinoma

    Supplemental Digital Content

    Copyright © 2022 the Author(s). Published by Wolters Kluwer Health, Inc.