The diagnostic value of LGALS1 in esophageal cancer and its potential molecular pathways via bioinformatic analysis

Background: Esophageal cancer (ESCA) was one of the most common malignant tumors. The purpose of this study was to reveal the role and potential regulatory mechanism of LGALS1 in the progression of ESCA. Methods: Oncomine, Timer and TCGA databases were used to analyze the expression level of LGALS1 and its value in the diagnosis of ESCA. The correlation between LGALS1 expression and the clinicopathological characteristics in ESCA patients was analyzed through the Ualcan database. In the TCGA database, we performed Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) to elucidate the potential mechanism underlying the role of LGALS1 in ESAC pregression. Besides, LGALS1 co-expressed genes were imported into the STRING database to build a protein-protein interaction (PPI) network. Hub genes were screened by the CytoHubba plug-in in Cytoscape and was veried in Gene Expression Proling Interactive Analysis (GEPIA) and TCGA databases. Results: LGALS1 was highly expressed in ESCA tissues. Increased LGALS1 expression was related to age, ethnicity, clinical stage, tumor grade, and histological subtype of ESCA patients, and receiver operating characteristic (ROC) analysis showed that the area under the curve (AUC) was 0.8511 (P < 0.001). GO and KEGG showed that LGALS1 co-expressed genes, in which 482 genes were included, were mainly involved in endothelial cell differentiation, transforming growth factor receptor signaling pathway, epithelial cell proliferation, ECM receptor interactions, leukocyte migration, transcriptional disorders during cancer, PI3K/AKT signaling pathway and so on. Moreover, GSEA showed that elevated expression of LGALS1 was mainly enriched in ECM receptor interaction, cancer pathway, and TGF beta signaling pathway. The hub genes (COL1A1, FN1, COL1A2, COL3A1, COL5A1, COL5A2, COL4A2, COL18A1, and COL6A1) were highly expressed and had diagnostic signicance in ESCA. Conclusion: LGALS1 expression

LGALS1, also known as Galectin-1 (Gal-1), was increased expression in human malignant tumors such as hepatocellular carcinoma (HCC), ovarian cancer and gastric cancer [8][9][10][11]. And LGALS1 was associated with the prognosis of patients with malignant tumors [11][12][13]. Some researches have reported that LGALS1 was highly expressed in cervical cancer tissues and cells. Increasing the expression of LGALS1 could signi cantly promote cancer cell proliferation, inhibite cell apoptosis, and enhance cell migration and invasion capabilities. Interference with LGALS1 expression would lead to opposite effects [14].
Besides, Galectin-1 was highly expressed in the stroma of pancreatic ductal adenocarcinoma. And Galectin-1 overexpression promoted cancer cell proliferation, migration, and invasion [15]. Galectin-1 expression was increased in gastric cancer tissues. Increased Galectin-1 expression was associated with poor prognosis in patients with gastric cancer. Galectin-1 overexpression could induce migration and invasion of gastric cancer cells via Epithelial-mesenchymal transition (EMT) [11]. In addition, increased Galectin-1 expression might be an independent factor for the prognosis of bladder cancer [12]. In summary, LGALS1 expression was elevated in many malignancies, and the expression level of LGALS1 was related to the occurrence, development and poor prognosis of malignant tumors, which indicated that LGALS1 might be a potential target in tumor diagnosis and treatment. However, the role and potential value of LGALS1 in ESCA had not been reported in the literature. Therefore, the purpose of this study was to analyze the expression of LGALS1 and its clinical signi cance in ESCA, and to explore the role of LGALS1 in the development of ESCA and its possible regulatory mechanisms.

Methods
Oncomine database Oncomine (https://www.oncomine.org/) database was a gene chip-based database and integrated data mining platform [16]. The Oncomine database was used to set screening conditions to explore the expression of single gene in cancer. In this study, we set the screening thresholds as followed: 1) Gene: LGALS1; 2) Comparison type: Cancer vs Normal; 3) Date type: mRNA; 4) Threshold setting conditions (P < 0.001, Fold change 1.5, Gene rank = All).

TCGA database data download
Gene expression data of ESCA HTSeq-FPKM type from the Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/) website was downloaded, including 11 cases of normal esophageal tissues and 160 cases of ESCA tissues. The data was sorted and extracted to verify the expression of LGALS1 and hub genes in the TCGA database and we further explored its value in the diagnosis of ESCA patients.

Ualcan and GEPIA databases
According to Ualcan (http://ualcan.path.uab.edu/index.html) RNA-seq and ESCA type clinical data, the expression level of LGALS1 mRNA and its relationship with the clinicopathological characteristics of ESCA patients were analyzed [17]. Gene Expression Pro ling Interactive Analysis (GEPIA) database was a database that organized TCGA data and performed secondary analysis. The correlation between LGALS1 and Hub genes was veri ed in GEPIA (http://gepia.cancer-pku.cn/) [18].
Screening for LGALS1 co-expressed genes In this study, the Pearson correlation was used to analyse the association between genes. The Pearson correlation coe cient could be used to represent the biological relationship between two genes in numerical form. R was used to screen LGALS1 co-expressed genes in the TCGA database. Screening thresholds were |r| > 0.5 and P < 0.001.

GO and KEGG analysis
In order to explore the biological functions and mechanisms of LGALS1 that regulated ESCA progression, we performed Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis on LGALS1 co-expressed genes. GO annotations included biological processes (BP), molecular functions (MF), and cellular components (CC). GO annotation and KEGG analysis of LGALS1 co-expressed genes were performed. We set the screening thresholds as correlation coe cient and P < 0.001. Gene Set Enrichment Analysis (GSEA) was a method to reveal genomic expression data through basic knowledge.
The TCGA gene expression data set was divided into high and low expression groups based on the median expression level of LGALS1.The effect of high and low expression levels of LGALS1 on each gene was analyzed using GSEA version 4.01, and the related mechanism of LGALS1 participation in ESCA progress was further analyzed. The genome was permuted 1000 times per analysis. Moreover, LGALS1 expression level was used as a phenotypic marker. Nominal p-values and normalized enrichment score (NES) were used to classify enrichment pathways in each phenotype [19].

PPI network construction and hub genes analysis
The online STRING (https://string-db.org/) database was used to analyze the PPI network relationship of LGALS1 co-expressed genes to show the role of LGALS1 co-expressed genes in ESCA. The combined score > 0.7 was considered a statistically signi cant difference. The obtained PPI network was imported into Cytoscape 3.6.1 software. Highly connected nodes were very important to maintain the stability of the entire PPI network. Therefore, the top 10 genes screened by the CytoHubba plug-in were de ned as the hub genes in the PPI network [20], and we further analyze the clinical value of the Hub genes in the TCGA database to explore the role of LGALS1 in ESCA.

Statistical analysis
Data processing was performed using perl and R (V.3.5.2), Wilcoxon test was used to detect the expression of LGALS1 in ESCA. ROC was used to analyze the diagnostic value of LGALS1 and its hub genes in ESCA. Pearson correlation analysis was used to screen LGALS1 co-expressed genes.

Results
LGALS1 was abnormally expressed in cancer tissues According to the Oncomine database, LGALS1 expression was increased in tissues of nervous system tumors, breast cancer, cervical cancer, colorectal cancer, ESCA, gastric cancer, head and neck cancer, kidney cancer, and leukemia, while the expression in bladder cancer, cervical cancer, leukemia, lung cancer, prostate cancer and other tissues were decreased (Fig. 1A). In addition, LGALS1 expression was up-regulated in cholangiocarcinoma ( Expression of LGALS1 and its diagnostic value in ESCA In addition, LGALS1 was elevated in ESCA and its subtypes ( Table 1 and Fig. 2). In the Oncomine database, Hao and Kim found that LGALS1 expression was elevated in esophageal adenocarcinoma compared with normal esophageal tissues (P < 0.05). Hu and Su found that LGALS1 expression was elevated in esophageal squamous cell carcinoma compared with normal esophageal tissues (P < 0.05).
Kim also found that LGALS1 expression was higher in Barrett's esophagus than in normal esophageal tissues (P < 0.05). In the TCGA database, LGALS1 expression was elevated in ESCA tissues compared to normal esophageal tissues, and ROC analysis showed AUC = 0.8511 ( Fig. 2, P < 0.001). LGALS1 co-expressed genes 482 genes of LGALS1 with moderate or higher levels were screened, of which 349 were positively correlated ( Table 2) and 133 were negatively correlated ( Table 3). The top 10 genes of positively and negatively correlated with LGALS1 co-expressed genes are shown in Fig. 4. Table 2 LGALS1 positively related genes.

GO and KEGG analysis
To further understand the potential function of LGALS1 in ESCA progression, we performed GO and KEGG on LGALS1 co-expressed genes. GO annotation found that LGALS1 co-expressed genes were mainly involved in endothelial cell differentiation, transforming growth factor receptor signaling pathway, epithelial cell proliferation, and ECM receptor interaction ( Fig. 5A and online Table S1). KEGG found that LGALS1 co-expressed genes were mainly involved in leukocyte migration, transcriptional dysregulation during cancer, and PI3K/AKT signaling pathways ( Fig. 5B and Table 4). GSEA showed that elevated LGALS1 was mainly enriched for ECM receptor interactions, cancer pathways, TGF beta signaling pathways, and cytokine-to-cytokine interactions ( Fig. 6 and Table 5). Note: GSEA, Gene Set Enrichment Analysis

Hub genes expression and clinical signi cance analysis in PPI network
The potential biological function of LGALS1 was speculated by identifying the functions of LGALS1 coexpressed genes. The PPI network relationship was shown in Fig. 7A. The hub genes were COL1A1, FN1, COL1A2, COL3A1, and COL5A1, COL5A2, COL4A2, COL18A1, COL6A2, and COL6A1 ( Fig. 7B and Table 6).
In the GEPIA database, we found that LGALS1 expression level was related to hub genes expression level in ESCA tissues (Fig. 8). In the TCGA database, we found that COL1A1, FN1, COL1A2, COL3A1, COL5A1, COL5A2, COL4A2, COL18A1, and COL6A1 were abnormally expressed in ESCA tissues (Fig. 9) and had diagnostic signi cance (Fig. 10, P < 0.05), while the expression and diagnostic value of COL6A2 in ESCA was not statistically signi cant.

Discussion
Studies had shown that LGALS1 expression was elevated in cancer tissues such as gastric cancer, ovarian cancer, and neuroblastoma, and was associated with poor prognosis in gastric cancer, ovarian cancer, and neuroblastoma patients [11,[21][22][23]. Besides, we found a strange phenomenon using the Oncomine and Timer databases. For example, there were 5 data sets in the Oncomine database showed that the expression of LGALS1 was increased in breast cancer tissues, while the expression of LGALS1 was decreased in BRCA tissues via the Timer database. The reason for the con icting expression might be that the data was got from different research centers. However, the expression of LGALS1 in ESCA tissues showed signs of increase. In addition, the expression and potential clinical value of LGALS1 in ESCA had not been reported in the literature. We found that LGALS1 expression was elevated in ESCA tissues via the Oncomine, Timer, and TCGA databases, which was consistent with the expression in gastric, ovarian, and neuroblastoma. Moreover, the expression level of LGALS1 was related to age, race, weight, smoking history, cancer stage, tumor grade, histological subtype and lymph node metastasis in ESCA patients. In addition, ROC analysis showed that LGALS1 had an AUC of 0.8511 in ESCA and it was statistically signi cant. These results indicated that LGALS1 was involved in ESCA progress and expected to become a potential new target for diagnosis and treatment of ESCA.
Currently, LGALS1 acted as an oncogene in tumors to promote cancer progression. For example, Galectin-1 was upregulated in HCC cells and could regulate HCC cell adhesion, polarization, and tumor growth in vivo [24]. Galectin-1 overexpression could also activate FAK/PI3K/AKT signaling pathway to induce the resistance of EMT and sorafenib in HCC [25]. What's more, Galectin-1 could cause β-catenin nuclear translocation in HCC, as well as increased TCF4/LEF1 transcriptional activity, cyclin D1 and protooncogene expression [26]. Toll-like receptor (TLR) could mediate PI3K activation regulation via Galectin-1 production, and then participate in regulating ovarian cancer cell invasion and metastasis [27]. Galectin-1 overexpression promoted pancreatic cancer cells proliferation and metastasis, and could promote the expression of bronectin, collagen type I, α-SMA, MMP-2 and TIMP-1 through the TGF-β1/Smad signaling pathway [28]. In gastric cancer, overexpression of Galectin-1 could enhance TGF-β signaling through positive feedback, reduce cancer cell apoptosis, and promote cancer cell migration and invasion [29]. Most cytokines were synthesized by immune cells and secrete a small class of proteins with a wide range of biological activities. Interfering with LGALS1 expression could down-regulate M2 macrophages and myeloid-derived suppressor cells (MDSCs), and suppress immunosuppressive cytokines, and immunosuppressive microenvironment of glioma [30]. Galectin-1 could inhibit the viability, proliferation, and Th1 cytokine production of non-malignant T cells in patients with leukemia skin T-cell lymphoma [31]. These results indicated that LGALS1 could participate in tumor progression through PI3K/AKT signaling pathway, TGF-β signaling pathway, and regulation of cytokines secreted by immune cells. GO showed that LGALS1 co-expressed genes were mainly involved in endothelial cell differentiation, transforming growth factor receptor signaling pathway, epithelial cell proliferation, and ECM receptor interaction, and might be involved in the progression of ESCA including leukocyte migration, transcriptional disorders during cancer, PI3K/AKT signaling pathway, ECM receptor interaction, cancer pathway, TGF beta signaling pathway, cytokine and cytokine interaction, etc. Studies had shown that interferenced with LGALS1 expression could inhibit ESCA cell cycle transition and cell migration, and promote p27 and p21 protein expression, and inhibit cdk2 and MMP-14 protein expression [32]. This further con rmed that LGALS1 could participate in cancer progression via migration, cytokine and cytokine interaction, cancer pathway, etc. However, the regulation mechanism of LGALS1 in ESCA had not been con rmed yet further research.
The top 10 hub genes in the PPI network were COL1A1, FN1, COL1A2, COL3A1, COL5A1, COL5A2, COL4A2, COL18A1, COL6A2, and COL6A1. The Hub genes had extremely important clinical signi cance in the progress of cancer. COL1A1 and FN1 were highly expressed in esophageal squamous cell carcinoma (ESCC) tissues and both of them could promote ESCC cell proliferation and invasion [33,34]. COL1A2 was signi cantly down-regulated in colorectal cancer (CRC) tissues, and COL1A2 mRNA expression levels was related to tumor differentiation, invasion, and lymph node metastasis in CRC patients. Overexpressed COL1A2 inhibited CRC cell proliferation, migration, and invasion [35]. COL3A1 expression level was elevated in CRC tissues and cells and increased COL3A1 expression level was related to T stage, Dukes stage, grade, recurrence, etc. Patients with elevated COL3A1 expression had poorer OS and disease-free survival. Silencing COL3A1 could inhibit CRC cell proliferation [36]. Interfering the expression of COL5A1 in metastatic LUAD cells inhibited cell growth and invasion, and induced ce11 apoptosis [37]. Increased IDO1 expression promoted gastric cancer cell migration. Interfering with IDO1 expression in GC cells reduced LOXL2, COL6A1, COL6A2, and COL12A1 mRNA and protein expressions [38]. These results indicated that these hub genes were related to tumorigenesis and development.
We found that LGALS1 expression level was related to hub genes expression level in ESCA tissues in the GEPIA database, and COL1A1, FN1, COL1A2, COL3A1, COL5A1, COL5A2, COL4A2, COL18A1, and COL6A1 of Hub genes were elevated in ESCA tissues and had diagnostic signi cance in the TCGA database.
LGALS1 was involved in the progression of ESCA and was a new diagnostic target for ESCA patients. In other tumors, studies had reported that Hub genes and PI3K/AKT signaling pathways, TGF-β signaling pathways, and cytokine regulatory mechanisms had extremely important roles, which further con rmed that LGALS1 and its co-expressed genes had a important biological role in the progression of ESCA.
In summary, we showed that the increased expression of LGALS1 was related to the clinicopathological characteristics of ESCA patients and could be used as a potential target for ESCA diagnosis. Moreover, LGALS1 might promote the development of ESCA via PI3K/AKT signaling pathway, TGF beta signaling pathway, cancer pathway, cytokine and cytokine interaction.

Conclusion
LGALS1 expression was elevated in ESCA tissues and might be a potential diagnostic marker for ESCA patients. PI3K/AKT signaling pathway, TGF beta signaling pathway, cancer pathway, cytokine and cytokine interaction might be an important pathway for regulation in occurrence and development of ESCA.

Declarations
Funding This work was supported by the scienti c and technological research and development project plan of Shiyan city, Hubei Province (Number: 16Y10). GJL was responsible for experimental design and implementation of the scheme and was responsible for data veri cation and proofreading.
Availability of data and materials Not applicable.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.