LncRNA DLGAP1-AS2 overexpression associates with gastric tumorigenesis: a promising diagnostic and therapeutic target

Aberrant expression of long noncoding RNAs (lncRNAs) is associated with the progression of human cancers, including gastric cancer (GC). The function of lncRNA DLGAP1-AS2, as a promising oncogene, has been identified in several human cancers. Therefore, this study was aimed to explore the association of DLGAP1-AS2 with gastric tumorigenesis, as well. The expression level of DLGAP1-AS2 was initially pre-evaluated in GC datasets from Gene Expression Omnibus (GEO). Moreover, qRT-PCR experiment was performed on 25 GC and 25 adjacent normal tissue samples. The Cancer Genome Atlas (TCGA) data were also analyzed for further validation. Consistent with data obtained from GEO datasets, qRT-PCR results revealed that DLGAP1-AS2 was significantly (p < 0.0032) upregulated in GC specimens compared to normal samples, which was additionally confirmed using TCGA analysis (p < 0.0001). DLGAP1-AS2 expression level was also correlated with age (p = 0.0008), lymphatic and vascular invasion (p = 0.0415) in internal samples as well as poor survival of GC patients (p = 0.00074) in GEO datasets. Also, Gene Ontology analysis illustrated that DLGAP1-AS2 may be involved in the cellular process, including hippo signaling, regulated by YAP1, as its valid downstream target, in GC samples. Moreover, ROC curve analysis showed the high accuracy of the DLGAP1-AS2 expression pattern as a diagnostic biomarker for GC. Our findings indicated that DLGAP1-AS2 might display oncogenic properties through gastric tumorigenesis and could be suggested as a therapeutic, diagnostic, and prognostic target.


Introduction
Gastric cancer (GC), also known as stomach cancer, is one of the major human health problems, with over one million new cases reported worldwide each year [1,2]. Although GC incidence has been decreased recently, it remains the fifth most frequently diagnosed malignancy and accounts for the third cause of cancer-related death in the world [3][4][5]. Patients suffering GC are often diagnosed at advanced stages, in which surgery, as the main treatment option, is not recommended. Besides, in such cases, the efficacy of other treatment options, including chemotherapy and radiotherapy, are limited [6][7][8]. Therefore, more investigations are needed to further illustrate novel molecular targets for better management of GC.
GC is a multifactorial disease that genetic, epigenetic, and environmental factors have been identified to participate 1 3 in its development [9,10]. As one of the important epigenetic factors, long non-coding RNAs (lncRNAs) are transcripts with 200 nucleotides length, playing important roles in various biological processes by multiple mechanisms, including gene expression regulation [11]. A growing body of evidence has revealed that lncRNAs participates in tumorigenesis, progression, and metastasis in numerous cancers, including GC [12][13][14][15]. Some lncRNAs, such as PCAT18 and lINC01133, are downregulated in GC patients and function as tumor suppressors; whereas other lncRNAs like H19, CCAT1, and GHET1, are abundantly expressed through gastric tumorigenesis, showing oncogenic properties [16][17][18][19][20][21]. Besides, the expression patterns of lncRNAs have been illustrated to possess great potential as diagnostic and prognostic targets for a wide array of human cancers, including GC [22][23][24].
Cytoplasmic lncRNA DLGAP1 antisense RNA 2 (DLGAP1-AS2), localized on human chromosome 18p11.31, is transcribed in the antisense of the DLGAP1 gene. The DLGAP1-AS2 function, as a promising oncogene, has been recently identified in glioma, hepatocellular carcinoma (HCC), and cholangiocarcinoma (Freddie Bray et al.) progression. However, little is known about the tumorigenesisrelated function of DLGAP1-AS2 [25][26][27]. Therefore, this study was aimed to explore the association of DLGAP1-AS2 with gastric tumorigenesis. In this regard, we initially analyzed microarray datasets from the Gene Expression Omnibus database (GEO) to evaluate the DLGAP1-AS2 expression level in GC samples. Subsequently, the outcomes were validated internal GC samples and The Cancer Genome Atlas-Stomach Adenocarcinoma (TCGA-STAD) dataset.
Also, the correlations between DLGAP1-AS2 expression level and clinicopathological features and patients' survival were analyzed. A workflow chart of the theoretical strategy is presented in Fig. 1.

Material and methods
In-silico investigation of DLGAP1-AS2 expression using the GEO database We first analyzed microarray expression datasets from the GEO (http:// www. ncbi. nlm. nih. gov/ geo/) database to identify DLGAP1-AS2 expression levels in GC compared to normal tissues. Three datasets were retrieved, including GSE79973 (10 GC and 10 paired normal tissues), GSE19826 (12 GC and 12 adjacent noncancer tissues), and GSE54129 (111 GC tissues and 21 noncancerous gastric tissues). Differentially expressed genes between gastric tumors and noncancerous gastric samples were obtained using GEO2R available at http:// www. ncbi. nlm. nih. gov/ geo/ geo2r/.

Sample preparation
Twenty-five GC samples and 25 paired non-tumoral adjacent tissue samples were obtained from the Iranian tumor bank (Tehran, Iran). All patients who participated in the current study had given written informed consent. Tissue samples were stored in liquid nitrogen until RNA extraction. The participants received no radiotherapy or chemotherapy before Fig. 1 The flowchart of theoretical strategy in the present study the surgery. All clinicopathological and epidemiological features, including gender, age, tumor size, primary tumor site, lymphatic invasion, histological grade, perineural, serosal, and vascular invasion, clinical-stage, smoking status, and GC family history, were provided in patients' profiles.

RNA extraction and cDNA synthesis
Tissue specimens were ground in liquid nitrogen using a mortar and pestle, afterward transferred into the lysis buffer, and homogenized with a needle and syringe. The total RNA was isolated using the AllPrep DNA/RNA Kit (Germany, Qiagen) according to the manufacturer's protocol. The concentration and quality of the RNA samples were calculated by NanoDrop spectrophotometer (ThermoFisher Scientific Life Sciences, USA). The 1% agarose gel electrophoresis was carried out to assess the integrity of RNA. cDNA was synthesized from 1 μg of total RNA using the Prime-ScriptTM RT Reagent Kit (TaKaRa Bio, Japan) in a final volume of 20 µl according to the manufacturer's instruction.

Quantitative real-time PCR (qPCR)
qPCR was performed using the BioFACT™ 2X Real-Time PCR Master Mix (Korea) and gene-specific primer sets in a 10 μl total reaction volume. qPCR was done by Step One Plus Real-Time PCR System (Applied Biosystems, USA) in following steps: initial denaturing at 95 °C for 15 min, 45 cycles of denaturation at 95 °C for 10 s; primer annealing temperature at 60 °C for 30 s; and elongation at 72 °C for 20 s. Melting curves were obtained at the end of each run. GAPDH gene was used as a reference gene for data normalization. The 2 −ΔΔCt method was used to calculate the relative expression level of target genes between groups. The primers were designed using NCBI primer blast, an online primer designing tool (https:// www. ncbi. nlm. nih. gov/ tools/ primerblast/) and are presented in Table 1. It should be mentioned that the expression primers for DLGAP1-AS2 were designed outside of any overleaping regions with the protein-coding gene in the sense strand.

Data validation using TCGA datasets
To confirm the results obtained from GEO datasets and qPCR analysis in GC internal samples, The Cancer Genome Atlas-Stomach Adenocarcinoma (TCGA-STAD) database was analyzed. RNAseq gene expression data were retrieved from the UCSC Xena Functional Genomics Explorer (https:// xenab rowser. net/) and analyzed. Furthermore, using TCGA-STAD dataset and GEO datasets, Kaplan-Meier survival analysis was performed to determine the relevance of DLGAP1-AS2 expression with the overall survival (OS) of GC patients. Also, a receiver operating characteristic (ROC) curve analysis was performed to evaluate the potential of DLGAP1-AS2 expression pattern as a diagnostic biomarker for GC. Then, the values for tumor samples (N = 375) and normal samples (N = 32) for DLGAP1-AS2 expression retrieved from TCGA-STAD were considered as patient and control values. Next, using GraphPad 6 Prism software ROC curve analysis was performed and the area under curve (AUC), the measure of the ability of a biomarker to distinguish between groups, was evaluated at a confidence interval equal to 95%.

Protein-protein interaction (PPI) and gene ontology (GO) analysis
To further explore the possible function of DLGAP1-AS2 through tumorigenesis, PPI and GO molecular function and biological process analysis were performed using STRING online application (https:// string-db. org/), which predicts functional protein association networks. The minimum required interaction score was set at 0.400. GraphPad 6 Prism was applied to analyze the qPCR results and datasets and draw the graphs. Paired and unpaired t-test, Wilcoxon matched-pairs signed-rank test, and Mann-Whitney test were performed to statistically analyze differences of DLGAP1-AS2 expression level between groups. All data are presented as mean ± standard deviation (SD). A p value less than 0.05 was considered to be significant.

DLGAP1-AS2 high levels in GC samples retrieved from GEO
We initially analyzed microarray datasets from the GEO to evaluate the DLGAP1-AS2 expression levels in GC samples CCT TTC CTT AAC AGT GGC ACC GGT TCG AGG GAC ACT GTA GC GAPDH AAG GTG AAG GTC GGA GTC AAC GGG GTC ATT GAT GGC AAC AA compared to normal gastric tissues. As shown in Fig. 2, all the GSE79973, GSE19826, and GSE54129 datasets were in line with each other and revealed that DLGAP1-AS2 is significantly upregulated (p = 0.0371, p = 0.0015, and p < 0.0001, respectively) in GC samples compared to normal ones; indicating its significance through gastric tumorigenesis. The related raw data was provided in Supplementary File 1.

DLGAP1-AS2 upregulation in internal GC samples
We performed a qPCR experiment using the 25 GC tissue samples and 25 paired noncancerous gastric tissues. The qPCR analysis data revealed that DLGAP1-AS2 was significantly upregulated (p < 0.0032) in GC tissue specimens compared to paired normal samples (Fig. 3), which confirmed the results obtained from GEO datasets. The correlation between the expression level of DLGAP1-AS2 and clinicopathological features in GC patients was analyzed. There was no correlation between DLGAP1-AS2 expression level and gender, tumor size, primary tumor site, histological grade, perineural and serosal invasion, clinicalstage, smoking status, and GC family history (Table 2 and Supplementary File 2). Interestingly, the relative expression level of DLGAP1-AS2 was significantly correlated with age (p = 0.0008) (Fig. 4a), and lymphatic and vascular invasion (p = 0.0415) (Fig. 4b). As shown in Supplementary File 2, interestingly, three of the patients with high levels of DLGAP1-AS2, were 45 years old or younger, which is related to early-onset GC.

Validation of DLGAP1-AS2 overexpression in TCGA-STAD dataset
The TCGA-STAD database was also analyzed to confirm the results obtained from the GEO datasets and qPCR experiment. Overall, DLGAP1-AS2 expression was analyzed in 375 GC and 32 normal samples. The data from TCGA-STAD analysis demonstrated that DLGAP1-AS2 is significantly overexpressed (p < 0.0001) in GC samples compared to normal specimens (Fig. 5a). Moreover, as illustrated in Fig. 5b, DLGAP1-AS2 expression possesses the potential as a diagnostic target for distinguishing GC and normal samples with the AUC up to 0.89 (p < 0.0001). Supplementary File 3 provides the related raw data. Furthermore, Kaplan-Meier survival analysis was performed to evaluate the association between GC patients' survival with the DLGAP1-AS2 expression level. Despite that, no significant relationship was observed between DLGAP1-AS2 levels and survival rates in TCGA-STAD cohort containing data for 350 patients (Fig. 6a), analysis of Kaplan-Meier Plotter data (https:// kmplot. com) favoring a larger number of GC cohorts in GEO datasets (N = 631) showed that the survival probability of patients with high levels of DLGAP1-AS2 (35.9 months) was significantly (p = 0.00074) lower than those patients exhibiting the low levels of DLGAP1-AS (80.7 months) (Fig. 6b).

DLGAP1-AS2 involvement in biological processes through modulating YAP1 in GC
To further reveal DLGAP1-AS2 function through gastric tumorigenesis, initially the association between its expression and YAP1 (Yes1 Associated Transcriptional Regulator) levels, as its valid downstream mediator [27], was first evaluated in GC patient's cohort from TCGA dataset. As shown in Fig. 7, Pearson's correlation analysis revealed a significant (p < 0.0001) positive correlation between DLGAP1-AS2 and YAP1 expression in GC samples retrieved from TCGA (r = 0.2623). This result suggests that DLGAP1-AS2 may partially exert its effects through modulating YAP1 expression. Then, PPI analysis using STRING application established that DLGAP1-AS2 through regulating YAP1 in cooperation with other proteins, including SMAD-7, ERBB4, CTGF, AMOLT1, LATS1-2, and TEAD1-4, is enriched in molecular function such as transcription factor binding, and may participate in the regulation of biological processes, including hippo signaling, signal transduction and transcription initiation from RNA polymerase II promoter.
To further evaluate DLGAP1-AS2 participation in gastric tumorigenesis through modulating YAP1, their correlation was also investigated in internal samples. As seen in Fig. 8a, the obtained results from qRT-PCR showed that YAP1 is significantly (p = 0.0007) overexpressed in GC samples compared to normal adjacent normal samples. Besides, a significant positive correlation (p = 0.25, r = 0.45) was found between DLGAP1-AS2 and YAP1 expression levels in internal samples as well (Fig. 8b).

Discussion
Despite the decreased incidence of GC, it is still a major human health concern. Exploration of pathogenic molecular mechanisms underlying GC is still needed for its early detection and lowering mortality rates [28]. Since lncR-NAs participate in various molecular and cellular processes through multiple mechanisms, including gene regulation, they have received increasing attention recently [29]. Additionally, aberrant expression of lncRNAs has been reported Fig. 4 a The correlations between DLGAP1-AS2 expression and GC patients' age (***p = 0.0008). b Lymphatic and vascular invasion status and DLGAP1-AS2 expression in GC tissue specimens (*p = 0.0415). The statistical analysis was performed using the unpaired t-test  The correlation between DLGAP1-AS2 and YAP1 expression levels was analyzed using TCGA-STAD patient cohort. b PPI analysis was performed. c GO molecular function analysis. d GO biological process analysis to be associated with the development and progression of numerous human cancers, including GC [24,29] Importantly, DLGAP1-AS2, as a promising oncogene, was shown to be overexpressed in various human cancers, including glioma, and to be involved in tumorigenesis [25,27]. Accordingly, the current study was aimed to investigate the association of DLGAP1-AS2 with gastric tumorigenesis. Then, microarray expression datasets from the GEO database were first analyzed to evaluate the DLGAP1-AS2 expression level in GC and normal tissues. DLGAP1-AS2 expression was found to be elevated in GC cases compared to normal samples. Also, qPCR results revealed that DLGAP1-AS2 was significantly upregulated in GC tissue specimens compared to paired normal samples. Another layer of confirmation came from analyzing the TCGA-STAD dataset showing high levels of DLGAP1-AS2 in GC samples. Also, the relative expression level of DLGAP1-AS2 in GC tissue samples was significantly correlated with lymphatic and vascular invasion. An interesting finding was the negative correlation between DLGAP1-AS2 expression level and the patients' age. It was found that, in internal samples, patients that were 45 years old or younger had the highest expression levels of DLGAP1-AS2. Consequently, it was suggested that DLGAP1-AS2 might be involved in early-onset GC which is known, as a type of GC presenting at the age of 45 or younger and accounting for approximately 10% of GC cases [30]. Therefore, our findings indicated that DLGAP1-AS2 overexpression might be associated with GC progression, suggesting its possible oncogenic role. Besides, despite the TCGA dataset, Kaplan Meier analysis illustrated a significant correlation between DLGAP1-AS2 overexpression and patients' poor survival in GEO datasets favoring the larger size of samples. Considering that and based on ROC curve analysis (p < 0.0001, AUC = 0.8920), it was suggested that DLGAP1-AS2 expression status may be suggested as a potential diagnostic and prognostic target for GC.
In line with our findings, Liu et al. previously observed the up-regulation of DLGAP1-AS2 in cholangiocarcinoma (CCA) cell lines, which contributes to CCA progression by modulating the miR-505/GALNT10 cascade [26]. Recently, it was shown that overexpression of DLGAP1-AS2 is correlated with down-regulation of miR-154-5p in HCC patients and its knockdown leads to decreased methylation and overexpression of miR-154 which in turn inhibits HCC cell invasion and migration [25]. Miao et al. also found that DLGAP1-AS2 was overexpressed in glioma patients. They revealed that DLGAP1-AS2 depletion in glioma cells promoted cell apoptosis and inhibited cell proliferation and migration, which consequently attenuated the progression of glioma. This effect was shown to be induced through DLGAP1-AS2-mediated YAP1 overexpression. Thereby, YAP1 was considered a downstream target of DLGAP1-AS2 that appears to play a biological role in glioma progression. [27]. Several lines of evidence from different studies have identified YAP1 as an oncogene in numerous cancers, including GC [31]. Further investigations have revealed that up-regulation of YAP1 could promote cell proliferation, growth, and migration in GC, showing an association with tumor progression and lymph node metastasis [23,[31][32][33][34][35]. It is noteworthy that a significant correlation was found in the current study between DLGAP1-AS2 overexpression and YAP1 high levels in TCGA dataset and internal samples by Pearson's correlation analysis. Then, it was suggested that YAP1 might be the downstream target of DLGAP1-AS2 in GC progression and invasion as well. Subsequently, PPI and GO molecular functions and biological processes showed that DLGAP1-AS2 may be involved in the regulation of Hippo signaling which is hyperactivated through gastric tumorigenesis. YAP1 was previously reported to function as one the key effectors in hippo pathway, that is translocated to the nucleus and through binding with TEAD1-4 transcription factors, induces transcriptional activity, leading to cell proliferation and tumor progression [36]. However, further investigations are required to identify the molecular mechanisms by which DLGAP1-AS2 contributes to GC progression through regulating this pathway.

Conclusion
In conclusion, our results revealed that the aberrant overexpression of DLGAP1-AS2 was associated with GC progression, invasion, and poor survival in patients, suggesting its possible oncogenic role through gastric tumorigenesis. Also, YAP1 was evidenced to be one of the downstream effectors of DLGAP1-AS2 in GC progression. Also, according to ROC curve analysis, DLGAP1-AS2 may be suggested as a potential diagnostic target for GC. However, further investigations are needed to elucidate the molecular mechanisms by which DLGAP1-AS2 participates in gastric tumorigenesis to provide clues that may help its clinical applications.