Distal gastric cancer (DGC) is a subgroup of gastric cancer (GC), which has different molecular characteristics from proximal gastric cancer (PGC). These differences result in different overall survival (OS) rates; however, data pertaining to the survival rate in PGC or DGC are contradictory. This suggests that the location of GC is not the unique cause of the different survival rates, while the molecular characteristics might be more important factors determining the prognosis of DGC. Therefore, the aim of this study was to discover key prognostic factors in DGC using bioinformatic methods and to explore the potential molecular mechanism. The Cancer Genome Atlas (TCGA) public database was employed to screen data relating to DGC, and we conducted a weighted gene co-expression network analysis (WGCNA) on DGC patient samples to establish co-expression modules. High-weight genes (hub genes) in a dominant color module were identified. In vitro experiments and gene set enrichment analyses (GSEA) were carried out to elucidate the potential molecular mechanism. In this study, 139 DGC samples were enrolled to perform a co-expression analysis. According to the correlation between gene modules and clinical characteristics, the royal blue module related to stage M of DGC was screened, and a survival analysis was conducted to show that high-coagulation-factor V (F5) expression was related to the short OS of patients with GC. In vitro experiments confirmed that F5 could promote the migration of GC cells. GSEA suggested that F5 might have affected the prognosis of GC by modulating the activities of the Wnt and/or the TGF-β signaling pathways. Our results indicated that high F5 expression predicts poor prognosis of patients with DGC, and it functions probably by promoting cell migration through the Wnt and/or the TGF-β signaling pathways.
Gastric cancer (GC) is a high-risk tumor with the highest mortality and morbidity rates worldwide. GC is often diagnosed at an advanced stage when patients have a median overall survival (mOS) of about 10–12 months. According to the location of the disease, GC can be divided into proximal gastric cancer (PGC), middle stomach cancer, and distal gastric cancer (DGC). Epidemiological studies have shown that the biological and pathological characteristics of DGC are different from PGC (Devesa et al., 1998). For example, DGC tends to have poorly differentiated tumors compared with PGC, and DGC is associated with more advanced tumor stage and older age compared with PGC (Wang et al., 2019). Different biological and pathological characteristics result in different overall survival (OS) rates. However, the data pertaining to rates of survival of DGC or PGC are contradictory. A study (Higuchi et al., 2004) has reported shorter survival in patients with DGC compared to those with PGC, while some reports (Pacelli et al., 2001; Petrelli et al., 2017; Yu et al., 2018) have shown longer overall survival (OS) in DGC patients. Even though, another study (Costa et al., 2016) has shown no obvious difference in the prognosis of DGC and PGC. Therefore, the location of GC is not the unique cause of the different survival rates. Investigation of the molecular characteristics might be more important, which may be one of the decisive factors determining clinical biological behavior and prognosis of DGC.
Currently, the molecular mechanism that underlies the development of GC remains to be fully elucidated. A study has shown that cell invasion and migration in GC can be promoted by homeobox C10 (HOXC10) through the upregulation of pro-inflammatory cytokines (Li et al., 2020). Moreover, proliferation and migration in GC could be promoted via activation of the Wnt signaling pathway by LncRNA HLA complex group 11 (HCG11) (Zhang et al., 2019). However, there are no reports on the biomarkers related to DGC.
Weighted gene co-expression network analysis (WGCNA) is an approach used in the co-expression module correlation analysis from microarray samples (Langfelder and Horvath, 2008). WGCNA has been used in various biological scenarios and is useful in the exploration of therapeutic targets or potential biomarkers (Ivliev et al., 2010).
In this study, we aimed to reveal potential molecular mechanisms leading to the development of DGC using WGCNA. Our results suggested that high-coagulation-factor V (F5) expression results in a poor prognosis of patients with DGC by promoting cell migration, and which might function by regulating the Wnt and/or the TGF-β signaling pathways.
Materials and MethodsData source and data processingData processing
RNA-sequencing data of gastric cancer samples and relevant clinical data were downloaded from the database of TCGA (http://cancergenome.nih.gov/). It was used as data to perform our study.
Data inclusion and exclusion criteria
(1) Data which was at least 18 years of age were eligible for enrollment; (2) The sample must have been pathologically confirmed as gastric cancer; (3) The gastric cancer data in TCGA contains six aspects: Gastroesophageal Junction, Cardia/Proximal, Fundus/Body, Antrum/Distal, Stomach (NOS) and Other (please specify). We chose the “Antrum/Distal” part as DGC to perform our study.
Gene selection
For the selected distal gastric cancer data, the genes were sorted by the median absolute deviation variance size, and the front 3600 genes were extracted. The clinical characteristics of the sample contained nine aspects: histological grade, stage T, stage N, stage M, gender, race, tumor stage, age, and morphology.
Construction of co-expression network
WGCNA package was used to build a co-expression network (139 samples of DGC were used) in R after the 3600 most variant genes were tested. The adjacency matrix Amn was defined, and the soft-thresholding parameter β = 4 was chosen. Then, we constructed the topological overlap matrix (TOM) to counter the effects of missing or spurious connections between network nodes.
TOMm,n=∑k=1NAm,k.Ak,n+Am,nmin(Km,Kn)+1−Am,n
Next, the average linkage hierarchical clustering was conducted to classify genes into gene modules with high absolute correlations. The minimum size was 30.
Correlation between clinical characteristics and different modules
With the purpose to identify modules associated with clinical characteristics (distal gastric cancer), the correlation between clinical characteristics and module eigengenes was calculated by the Spearman correlation analysis and a p-value of <0.05 as statistically significant. Meanwhile, we tested the module significance defined as the average gene significance of each gene in the linear regression between the clinical characteristics and gene expression. In general, the correlation between clinical characteristics and module eigengenes tended to be related to the module significance.
Identification of hub genes
Genes within the co-expression module are highly connected and have similar effects. The hub genes we filtered in each module based on the connectivity within the module and the correlation with the module’s characteristic genes. The cytoHubba plugin was used to extract the hub genes in every module with the purpose of obtaining a balance between the core genes and avoid missing any key gene based on Cytoscape (Chin et al., 2014).
Survival analysis
With the purpose to evaluate each hub gene as a prognostic marker of distal gastric cancer, we conducted survival analyses. We divided DGC samples into 2 groups based on the median expression values of the gene. The R package “survival” was used to create K-M survival curves to assess the prognostic value, and the log-rank test was used to evaluate differences between groups. After this, the p-values were produced. On the other hand, the prognostic value of hub gene expression in GC was analyzed again by using Kaplan–Meier Plotter on-line database (https://kmplot.com/analysis/).
Differential expression of hub genes and PPI network constructions
The on-line database Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancer-pku.cn) was used to identify the differential expression of target molecules between the gastric cancer tissue and gastric tissue. GEPIA is an interactive web that includes 8587 normal and 9736 tumors samples from TCGA and the GTEx projects (Tang et al., 2017). Besides, the Protein-Protein Interaction (PPI) Network Construction was performed using the STRING web (https://string-db.org/cgi/input.pl) method to analyze the function of the protein encoded by the target gene.
Cytological experiments verify the effect of selected hub genes on gastric cancer cellsCell culture and transfection
We obtained GC cells lines MGC803, BGC823, SGC7901 and HGC27 from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China), which were cultured with RPMI-1640 medium containing 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin-streptomycin. They were cultured in a humidified incubator at 37°C containing 5% CO2, digested, and passaged with 0.25% trypsin digest, and passaged once every 2–3 days. We used logarithmic growth phase cells in all experiments.
The specific siRNAs targeted to F5 and corresponding negative control (NC), were compounded by RiboBio (Guangzhou, China). SGC7901 cells (2 × 105) were transfected with siRNAs Lipofectamine 2000 reagent (Invitrogen) Transfection Reagent based on the manufacturer’s instructions.
RT-qPCR detection of mRNA expression levels
Total RNA was isolated with Trizol reagent (Invitrogen, USA) and quantified by measuring the absorbance at 260 nm by nanodrop 2000 (Thermo Fisher Scientific, USA). For mRNA detection, we used the One Step PrimeScript mRNA cDNA Synthesis Kit (Takara, Japan) to carry out reverse transcription. SYBR Premix EX TaqTM II (Perfect Real Time) (Takara, Japan) was used to generate cDNA from 1000 ng total RNA. The Applied Biosystems 7500 Real-Time PCR Systems (Thermo Fisher Scientific, Waltham, MA, USA) was used to run the RT-qPCR. The conditions of PCR were 30 s at 95°C, followed by 45 cycles at 95°C for 5 s, and at 58°C for 25 s. The Applied Biosystems 7500 software program (version 2.3) was used to analyze the data. The fold change of the RNA expression was calculated by the 2−ΔΔCt method.
The PCR primers used are as follows:
F5: Forward (5’-ACCACAATCTACCATTTCAGGAC-3’) and Reverse (5’-CTTCTCCGCAGGGAATGTGT-3’).
18S: Forward (5’-CCCGGGGAGGTAGTGACGAAAAAT-3’) and Reverse (5’-CGCCCGCCCGCTCCCAAGAT-3’).
Cell viability assays
Cell viability was determined using MTT assay. Briefly, the 96-well plates with a cell density of 4 × 103 cells/well for 72 h were used to incubate SGC7901 cells, which were transfected with the F5 siRNA or NC siRNA. After treatment of cells with various conditions, the medium was removed, and 20 μL of MTT (5 mg/mL; Sigma Chemical Co., St Louis, MO, USA) was added to each well for 4 h incubation at 37°C. Then 200 μL of DMSO was added to each well. After shaking for 5 min, a microplate reader (model 550; Bio-Rad Laboratories, Inc., Hercules, CA, USA) was used to measure the absorbance at 570 nm.
Transwell migration assay
We carried out the migration assays in a 24-well chamber and used 8-μm pore size membranes (Corning, USA) to inserted polycarbonate. For migration assay, 3 × 104 cells were plated within 200 μL serum-free medium onto the upper chamber, and 500 μL medium with 2.5% FBS was added to the lower chamber. After incubating for 24 h, the chambers were fixed with methanol and then stained with 0.1% Wright-Giemsa dye. We captured 5 different fields and counted at 20X magnification per well. Fluorescence microscopy (BX53, Olympus, Japan) was used to visualize the cells. Each experiment was carried out at least three times.
GSEA pathway enrichment
Gene Set Enrichment Analysis (GSEA) was exploited to study the interpretation and analysis of the long lists of genes generated from high-throughput transcriptomic experiments (Mootha et al., 2003; Subramanian et al., 2005). Each hub genes of 443 GC samples in TCGA were divided into two groups according to the median expression values. We performed GSEA using the Java GSEA implementation to seek the potential function of targeted genes. We selected annotated gene set c2.cp.kegg.v5.0.symbols.gmt (Version 5.0 of the Molecular Signatures Database) as the reference gene set. The information on the signaling pathways was found by on-line web analysis (https://www.kegg.jp/kegg/pathway.html).
Statistical methods
For continuous variables, mean ± SD was used to express the normal distribution, and the median (range) was used to express the non-normal, whereas count (percentage) was applied to the categorical variables. We chose a p-value of <0.05 as statistically significant (two-tailed). We used SPSS Version 22.0 and R (version 3.5.3) to carry out these statistical analyses.
ResultsData processing
We obtained 139 tissue sample raw files from TCGA (Tab. 1). All selected expression datasets were log2-transformed, then standardized.
Clinical characteristics of DGC samples from TCGA
Groups
Number (n = 139)
Percentage (%)
Age (years)
≥65
78
56.1
<65
61
43.9
Stage T
T1
2
1.4
T2-4
137
98.6
Stage N
N0
37
26.6
N1-3
102
73.4
Stage M
M0
128
92.1
M1
11
7.9
Grade
G0-2
44
31.7
G3
95
68.3
Note: Clinical characteristics of the data used in this study. Although we divided the data into nine parts, there are some unknown data in four parts that do not show influence on the following studies.
Construction of co-expression networks
A network was constructed from the filtered probes which identified 21 modules. The soft-thresholding power 4 was selected to specify the adjacency matrix based on the standard of approximate scale-free topology (Fig. 1). We chose the module detection sensitivity deepSplit2, minimum module size 30, and cut height for merging of modules 0.25, which implied that the modules with an eigengene that has a correlation with different clinical characteristics of DGC higher than 0.75 have to be merged (Fig. 2).
Network topology for different soft-thresholding powers.
The numbers in the plots indicate the corresponding soft-thresholding powers. The approximate scale-free topology can be attained at the soft-thresholding power of 4.
Gene modules identified by WGCNA.
A gene dendrogram, obtained by clustering dissimilarity based on consensus topological overlap matrix with the corresponding module colors, is indicated by the colored row. Each colored row represents a color-coded module that contains a group of highly connected genes. A total of 21 modules were identified.
Correlation between different modules and clinical characteristics of DGC
Similar expression profiles in some modules were found in the analysis. We analyzed the connectivity of eigengenes to understand interactions amongst the 21 co-expressed modules and performed a cluster analysis. In conjunction with Fig. 3, significant differences were detected amongst the 21 modules, which may be due to different clinical parameters in the development of GC. For example, royal blue and green-yellow patches may be associated with the stage M of DGC, black and pink patches may be associated with tumor grade, and yellow patches may be negatively correlated with tumor grade.
Relationships between consensus module eigengenes and different clinical characteristics of DGC.
Each row in the table corresponds to a consensus module, and each column corresponds to a clinical characteristic. The module name is shown on the left side of each cell. Numbers in the table report correlations between the corresponding module eigengenes and characteristics, with the p-values printed below the correlations in parentheses. The table is color-coded by correlation according to the color legends. Direction and intensity of correlations are indicated on the right side of the heatmap (green, negatively correlated; red, positively correlated).
Identification of hub genes
We found a significant correlation between module membership and clinical characteristics in the royal blue module (Fig. 4(A)) and established a co-expression network of the hub genes, as shown in Fig. 4(B). The identification of 25 genes, which included F5, Wnt family member 11 (Wnt11), and testis associated actin remodeling kinase 1 (TESK1) in the royal blue module, may have a significant function in stage M of DGC.
Identification of hub genes in the royal blue module.
(A) Correlation between the module membership and clinical characteristics in the royal blue module. (B) The density of the circle corresponds to module membership values. The color of the circle represents the core degree of the gene.
Survival analysis
Survival analysis was conducted to verify the correlation between the expression of hub genes and the survival time of patients with DGC: high F5 expression was found to be significantly related to a short OS of DGC (p < 0.05, [HR 95%CI]: 1.466 (1.056–2.034), Fig. 5(A)). The Kaplan–Meier Plotter on-line database showed that overexpression of F5 was related to poor prognosis in GC (p < 0.05, HR = 1.66 (1.11–2.49), Fig. 5(B)).
Survival analysis and PPI network construction of <italic>F5</italic>.
(A) Kaplan–Meier curves of gene groups (F5) in the TCGA DGC dataset based on SurvExpress (N = 139). The horizontal axis represents time (year) to the event. Outcome event, time scale, concordance index (CI) and p-values of the log-rank test are shown. Red and blue curves represent high- and low-risk groups, respectively. (B) The analysis in the Kaplan–Meier Plotter on-line database of F5 in GC. Red and black curves represent high- and low-risk groups, respectively. (C) GEPIA was used for the analysis of F5 expression in stomach adenocarcinoma (STAD), and the boxplot was plotted. The red and grey boxes represent STAD (N = 408) and normal gastric tissues (N = 211), respectively. *p < 0.05. (D) PPI networks of F5 using the STRING tool. Genes are represented as nodes in the plot, and their interactions are denoted by lines.
Differential expression of hub genes and PPI network constructions
The encoded product of the F5 gene is a coagulation factor V, which is a basic cofactor in the coagulation cascade. It has been reported (Vossen et al., 2011) that polymorphisms in F5, such as F5 Leiden, were related to increased risk of colorectal cancer. On-line data analysis revealed a significant overexpression level of F5 in GC tissues compared to gastric tissues (Fig. 5(C)). The PPI network indicated that the protein encoded by F5 interacted with bone morphogenetic protein 4 (BMP4) and amphiregulin (AREG) and interacted indirectly with Wnt11 (Fig. 5(D)).
Cytological experiments to verify the effect of selected hub genes on GC cells
We verified F5 mRNA expression levels in four GC cell lines including MGC803, BGC823, HGC27, and SGC7901. Fig. 6(A) shows that F5 was expressed to varying degrees in the four GC cell lines. The SGC7901 cell line was used to further test the function of F5. Fig. 6(B) illustrates the efficiency of F5 knockdown in the SGC7901 cell line. MTT assays confirmed that inhibition of F5 slightly suppressed the proliferation of GC cells to 25% (p < 0.01, Fig. 6(C)). At 24-h, a Transwell migration assay indicated that when the F5 gene was silenced, the migration of GC cells was significantly inhibited from 383 to 19 (cell number/field, p < 0.001, Figs. 6(D)–6(E)).
<italic>In vitro</italic> experiments.
(A) Expression of F5 in the MGC803, BGC823, HGC27, and SGC7901 cell lines. (B) Expression of F5 in SGC7901 cell line after F5 gene knockdown. (C) The outcome of MTT assay. (D–E) The outcome of the Transwell migration assay. **p < 0.01; ***p < 0.001.
GSEA pathway enrichment
In this study, GSEA was used to explore the possible molecular mechanism of F5 in the prognosis of GC using 443 samples from the TCGA database. We identified two pathways that were clearly associated with the development of GC which were “basal cell carcinoma” (NES = 1.72, NOM p = 0.002) (Fig. 7(A)) and “TGF-β signaling pathway” (NES = 1.51, NOM p = 0.044) (Fig. 7(B)). Information pertaining to the basal cell carcinoma signaling pathway was shown in Fig. 8 and Tab. 2. The impact of F5 on the survival and prognosis of DGC may be mediated through a link with the above pathways.
GSEA analysis of <italic>F5</italic> in gastric cancer patients using the TCGA database.
(A–B) Two pathways ranked top in the F5-related predictive outputs.
Information pertaining to the basal cell carcinoma pathway from the KEGG on-line database.
Information of basal cell carcinoma signaling pathway
Name
Probe
Gene title
Rank in gene list
Rank metric score
Running es
Core enrichment
Row_0
WNT7B
wingless-type MMTV integration site family, member 7B
58
0.25326744
0.055809945
Yes
Row_1
FZD9
frizzled homolog 9 (Drosophila)
226
0.20891821
0.09464167
Yes
Row_2
FZD6
frizzled homolog 6 (Drosophila)
257
0.20379044
0.14055689
Yes
Row_3
FZD2
frizzled homolog 2 (Drosophila)
331
0.19461304
0.18172245
Yes
Row_4
AXIN1
axin 1
343
0.19329599
0.22632872
Yes
Row_5
WNT5A
wingless-type MMTV integration site family, member 5A
522
0.17658667
0.25692296
Yes
Row_6
FZD1
frizzled homolog 1 (Drosophila)
871
0.15791881
0.2728651
Yes
Row_7
DVL1
dishevelled, dsh homolog 1 (Drosophila)
1000
0.15183835
0.3006866
Yes
Row_8
SHH
sonic hedgehog homolog (Drosophila)
1274
0.14071888
0.31713563
Yes
Row_9
LEF1
lymphoid enhancer-binding factor 1
1647
0.12838289
0.3247089
Yes
Row_10
WNT8B
wingless-type MMTV integration site family, member 8B
1982
0.11861961
0.33229342
Yes
Row_11
BMP2
bone morphogenetic protein 2
2263
0.11144068
0.341462
Yes
Row_12
DVL2
dishevelled, dsh homolog 2 (Drosophila)
2314
0.11014112
0.3642344
Yes
Row_13
WNT4
wingless-type MMTV integration site family, member 4
wingless-type MMTV integration site family member 2
3116
0.09167766
0.45300686
Yes
Row_19
WNT6
wingless-type MMTV integration site family, member 6
3510
0.08424388
0.44897255
Yes
Row_20
WNT11
wingless-type MMTV integration site family, member 11
3599
0.08273254
0.46302778
Yes
Row_21
FZD10
frizzled homolog 10 (Drosophila)
3616
0.08251304
0.48138544
Yes
Row_22
GLI2
GLI-Kruppel family member GLI2
4139
0.0728891
0.46689105
Yes
Row_23
SUFU
suppressor of fused homolog (Drosophila)
4625
0.06510936
0.45281202
Yes
Row_24
WNT10A
wingless-type MMTV integration site family, member 10A
4766
0.06254979
0.4589958
Yes
Row_25
AXIN2
axin 2 (conductin, axil)
4800
0.06199568
0.47152016
Yes
Row_26
WNT8A
wingless-type MMTV integration site family, member 8A
5028
0.058139
0.47140998
Yes
Row_27
WNT9B
wingless-type MMTV integration site family, member 9B
5039
0.05796774
0.48438177
Yes
Discussion
The molecular mechanism involved in the development of DGC remains unclear. In this study, we attempted to identify a possible molecular mechanism through WGCNA and found that F5 may be related to stage M of DGC. GSEA indicated that F5 may not only have an effect on the development of DGC but may also be important in all GCs. The cytological experiments in this study confirmed that F5 can significantly affect cell migration of GC.
The encoded product of the F5 is the coagulation factor V, which is known as ‘clotting factor’. The protein circulates in plasma and participates in thrombin activation with factor Xa. Most existing reports focus on the coagulation function of F5, including the role of F5 and its protein in thrombosis (Zhang et al., 2018a; Zhang et al., 2018b), and DIC (Kou et al., 2019) in patients with hematological malignancies.
F5 has been reported in various tumor types as a marker. In breast cancer, the expression of F5 may inform clinical treatment decisions and prognosis of invasive breast cancer, indicating F5 as a potential biomarker of invasive breast cancer (Tinholt et al., 2018). However, another study (Ivancic et al., 2014) revealed that the increased risk of developing colorectal cancer is associated with F5 in ApcPirc/+ rats. Analysis using on-line databases showed significant over-expression level of F5 in GC tissues, and that high F5 expression can predict poor prognosis in both DGC and GC.
This study showed that the two pathways obtained from GSEA were closely related to F5. The basal cell carcinoma pathway includes the Wnt and TGF-β pathways. As these pathways were screened based on data of GC, the function of F5 in GC may be similar to basal cell carcinoma. According to GSEA analysis, we found that the core molecules which may have interaction with F5 in the basal cell carcinoma signaling pathway belong to the Wnt signaling pathway. In previous PPI analysis, we found that F5 is indirectly related to Wnt11. Furthermore, a study (Katoh and Katoh, 2009) also indicated that the canonical Wnt-to-Wnt11 signaling pathway is involved in cellular migration in tumor invasion during carcinogenesis. Therefore, F5 may participate in interactions with the Wnt signaling pathway.
Numerous studies have reported that the TGF-β and Wnt signaling pathways exert effects on the development of GC. Epithelial to mesenchymal transition (EMT) is a central biological process in which tumor cells lose epithelial characteristics and acquire mesenchymal features that make cancer cells more migratory and invasive (Thiery et al., 2009). Gastric epithelial cells acquire mesenchymal markers, become more invasive, and show stemness and metastasis during EMT (Thiery, 2002; Ye and Weinberg, 2015). The process of EMT is associated with certain signaling pathways, including the Wnt and TGF-β pathways. A study (Huang et al., 2015) revealed that Wnt signaling can promote progression in GC cells by EMT. Activation of the TGF-β/Smad signaling pathway can also induce EMT and then promote the metastasis in GC (Zhang et al., 2018c).
Previous studies (Luo et al., 2019) have suggested that GC patients have higher TGF-β levels in serum compared to healthy individuals. Elevated TGF-β levels are closely related to poor prognosis and shorter OS in GC patients (Hu et al., 2014). Furthermore, it has been confirmed (Chiurillo, 2015) that specific mechanisms up-regulate components of the Wnt signaling pathway. Also, the inactivation of inhibitors of Wnt signaling plays an important role in GC. Other studies (Boussioutas et al., 2003; Kurayoshi et al., 2006) have found that Wnt-5a (a molecule on the Wnt signaling pathway) was up-regulated in all types of GC and acted to promote invasion, migration and poor prognosis of patients with GC. Furthermore, suppression of the Wnt signaling pathway can inhibit the growth and migration of GC (Gao et al., 2018; Liu et al., 2018). The above studies are consistent with our cytological experiments and bioinformatics analysis. Based on this evidence, we hypothesize that the effect of F5 on the survival and prognosis of DGC might be through a link with the Wnt and/or the TGF-β signaling pathways.
In this study, the F5 gene was screened from samples of DGC. We found that F5 not only has an effect on DGC but also on GC based on PPI networks, cytological experiments, and GSEA studies. Cytological experiments confirmed that F5 significantly affected the migration of GC cells, which is consistent with our discovery by WGCNA. F5 is associated with stage M of DGC. There are many unsolved problems concerning the mechanism of F5 in GC, which warrant further investigation through cytological and zoological experiments.
Conclusions
This study suggested that high F5 expression was related to poor prognosis in patients with DGC and GC by promoting cell migration and this might participate in the regulation of the Wnt and/or the TGF-β signaling pathways. This provides a basis for further analysis of the prognosis and treatment of DGC and GC.
Acknowledgement: We would like to acknowledge the provider of the space and equipment for conducting the experiments. They were the Ministry of Education (China Medical University, Shenyang, China), Key Laboratory of Precision Diagnosis and Treatment of Gastrointestinal Tumors.
Availability of Data and Materials: The datasets analyzed during the current study are available from the corresponding author on reasonable request.
Author Contribution: The authors confirm contribution to the paper as follows: Study conception and design: Mengyi Tang, Bowen Yang, Yunpeng Liu, Zhi Li, Xiujuan Qu; data collection: Mengyi Tang, Bowen Yang, Chuang Zhang, Chaoxu Zhang, Dan Zang; analysis and interpretation of results: Mengyi Tang, Bowen Yang, Zhi Li, Xiujuan Qu; draft manuscript preparation: Mengyi Tang, Bowen Yang, Libao Gong. All authors reviewed the results and approved the final version of the manuscript.
Funding Statement: We would also like to thank these funds, including the National Natural Science Foundation of China (Nos. 81972331, 81972751, 81572374), The National Key Research and Development Program of China (2017YFC1308900), Technological Special Project of Liaoning Province of China (2019020176-JH1/103), Science and Technology Plan Project of Liaoning Province (NO. 2013225585) and The General Projects of Liaoning Province Colleges and Universities (LFWK201706).
Conflicts of Interest: In respect of the authorship, research, and/or publication of this article, all the authors declared no potential conflicts of interest.
ReferencesBoussioutasA, LiH, LiuJ, WaringP, LadeS, HollowayAJ, TaupinD, GorringeK, HavivI, DesmondPV, BowtellDD (
2003).
Distinctive patterns of gene expression in premalignant gastric mucosa and gastric cancer.
Cancer Research63:
2569–
2577.ChinCH, ChenSH, WuHH, HoCW, KoMT, LinCY (
2014).
cytoHubba: identifying hub objects and sub-networks from complex interactome.
BMC Systems Biology8:
S11. DOI
10.1186/1752-0509-8-S4-S11.ChiurilloMA (
2015).
Role of the Wnt/β-catenin pathway in gastric cancer: an in-depth literature review.
World Journal of Experimental Medicine5:
84–
102. DOI
10.5493/wjem.v5.i2.84.CostaLB, TonetoMG, MoreiraLF (
2016).
Do proximal and distal gastric tumours behave differently?Arquivos Brasileiros de Cirugia Digestiva29:
232–
235. DOI
10.1590/0102-6720201600040005.GaoJ, ZhaoC, LiuQ, HouX, LiS, XingX, YangC, LuoY (
2018).
Cyclin G2 suppresses Wnt/beta-catenin signaling and inhibits gastric cancer cell growth and migration through Dapper1.
Journal of Experimental & Clinical Cancer Research37:
317. DOI
10.1186/s13046-018-0973-2.DevesaSS, BlotWJ, FraumeniJF Jr. (
1998).
Changing patterns in the incidence of esophageal and gastric carcinoma in the United States.
Cancer83:
2049–
2053.HiguchiK, KoizumiW, TanabeS, SaigenjiK, AjaniJA (
2004).
Chemotherapy is more active against proximal than distal gastric carcinoma.
Oncology66:
269–
274. DOI
10.1159/000078326.HuWQ, WangLW, YuanJP, YanSG, LiJD, ZhaoHL, PengCW, YangGF, LiY (
2014).
High expression of transform growth factor beta 1 in gastric cancer confers worse outcome: results of a cohort study on 184 patients.
Hepatogastroenterology61:
245–
250.HuangL, WuRL, XuAM (
2015).
Epithelial-mesenchymal transition in gastric cancer.
American Journal of Translational Research7:
2141–
2158.IvancicMM, IrvingAA, JonakinKG, DoveWF, SussmanMR (
2014).
The concentrations of EGFR, LRG1, ITIH4, and F5 in serum correlate with the number of colonic adenomas in ApcPirc/+ rats.
Cancer Prevention Research7:
1160–
1169. DOI
10.1158/1940-6207.CAPR-14-0056.IvlievAE, 't HoenPAC, SergeevaMG (
2010).
Coexpression network analysis identifies transcriptional modules related to proastrocytic differentiation and sprouty signaling in glioma.
Cancer Research70:
10060–
10070. DOI
10.1158/0008-5472.CAN-10-2465.KatohM, KatohM (
2009).
Integrative genomic analyses of WNT11: transcriptional mechanisms based on canonical WNT signals and GATA transcription factors signaling.
International Journal of Molecular Medicine24:
247–
251. DOI
10.3892/ijmm_00000227.KouHM, ZhangXP, WangMZ, DengJ, MeiH, HuY (
2019).
Diagnostic and prognostic value of plasma factor V activity and parameters in thrombin generation for disseminated intravascular coagulation in patients with hematological malignancies.
Current Medical Science39:
546–
550. DOI
10.1007/s11596-019-2072-9.KurayoshiM, OueN, YamamotoH, KishidaM, InoueA, AsaharaT, YasuiW, KikuchiA (
2006).
Expression of Wnt-5a is correlated with aggressiveness of gastric cancer by stimulating cell migration and invasion.
Cancer Research66:
10439–
10448. DOI
10.1158/0008-5472.CAN-06-2359.LangfelderP, HorvathS (
2008).
WGCNA: an R package for weighted correlation network analysis.
BMC Bioinformatics9:
559. DOI
10.1186/1471-2105-9-559.LiJ, TongG, HuangC, LuoY, WangS, ZhangY, ChengB, ZhangZ, WuX, LiuQ, LiM, LiL, NiB (
2020).
HOXC10 promotes cell migration, invasion, and tumor growth in gastric carcinoma cells through upregulating proinflammatory cytokines.
Journal of Cellular Physiology235:
3579–
3591. DOI
10.1002/jcp.29246.LiuW, ChenY, XieH, GuoY, RenD, LiY, JingX, LiD, WangX, ZhaoM, ZhuT, WangZ, WeiX, GaoF, WangX, LiuS, ZhangY, YiF (
2018).
TIPE1 suppresses invasion and migration through down-regulating Wnt/beta-catenin pathway in gastric cancer.
Journal of Cellular and Molecular Medicine22:
1103–
1117.LuoJ, ChenXQ, LiP (
2019).
The role of TGF-β and its receptors in gastrointestinal cancers.
Translational Oncology12:
475–
484. DOI
10.1016/j.tranon.2018.11.010.MoothaVK, LindgrenCM, ErikssonKF, SubramanianA, SihagS, LeharJ, PuigserverP, CarlssonE, RidderstraleM, LaurilaE, HoustisN, DalyMJ, PattersonN, MesirovJP, GolubTR, TamayoP, SpiegelmanB, LanderES, HirschhornJN, AltshulerD, GroopLC (
2003).
PGC-1α-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes.
Nature Genetics34:
267–
273. DOI
10.1038/ng1180.PacelliF, PapaV, CaprinoP, SgadariA, BossolaM, DogliettoGB (
2001).
Proximal compared with distal gastric cancer: multivariate analysis of prognostic factors.
American Surgeon67:
697–
703.PetrelliF, GhidiniM, BarniS, SteccanellaF, SgroiG, PassalacquaR, TomaselloG (
2017).
Prognostic role of primary tumor location in non-metastatic gastric cancer: a systematic review and meta-analysis of 50 studies.
Annals of Surgical Oncology24:
2655–
2668. DOI
10.1245/s10434-017-5832-4.SubramanianA, TamayoP, MoothaVK, MukherjeeS, EbertBL, GilletteMA, PaulovichA, PomeroySL, GolubTR, LanderES, MesirovJP (
2005).
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles.
Proceedings of the National Academy of Sciences of the United States of America102:
15545–
15550. DOI
10.1073/pnas.0506580102.TangZ, LiC, KangB, GaoG, LiC, ZhangZ (
2017).
GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses.
Nucleic Acids Research45:
W98–
W102. DOI
10.1093/nar/gkx247.ThieryJP (
2002).
Epithelial-mesenchymal transitions in tumour progression.
Nature Reviews Cancer2:
442–
454. DOI
10.1038/nrc822.ThieryJP, AcloqueH, HuangRY, NietoMA (
2009).
Epithelial-mesenchymal transitions in development and disease.
Cell139:
871–
890. DOI
10.1016/j.cell.2009.11.007.TinholtM, GarredO, BorgenE, BerakiE, SchlichtingE, KristensenV, SahlbergKK, IversenN (
2018).
Subtype-specific clinical and prognostic relevance of tumor-expressed F5 and regulatory F5 variants in breast cancer: the CoCaV study.
Journal of Thrombosisand Haemostasis16:
1347–
1356. DOI
10.1111/jth.14151.VossenCY, HoffmeisterM, Chang-ClaudeJC, RosendaalFR, BrennerH (
2011).
Clotting factor gene polymorphisms and colorectal cancer risk.
Journal of Clinical Oncology29:
1722–
1727. DOI
10.1200/JCO.2010.31.8873.WangX, LiuF, LiY, TangS, ZhangY, ChenY, KhanSA (
2019).
Comparison on clinicopathological features, treatments and prognosis between proximal gastric cancer and distal gastric cancer: a national cancer data base analysis.
Journal of Cancer10:
3145–
3153. DOI
10.7150/jca.30371.YeX, WeinbergRA (
2015).
Epithelial-mesenchymal plasticity: a central regulator of cancer progression.
Trends in Cell Biology25:
675–
686. DOI
10.1016/j.tcb.2015.07.012.YuX, HuF, LiC, YaoQ, ZhangH, XueY (
2018).
Clinicopathologic characteristics and prognosis of proximal and distal gastric cancer.
Onco Targets and Therapy11:
1037–
1044. DOI
10.2147/OTT.S157378.ZhangH, HuangH, XuX, WangH, WangJ, YaoZ, XuX, WuQ, XuF (
2019).
LncRNA HCG11 promotes proliferation and migration in gastric cancer via targeting miR-1276/CTNNB1 and activating Wnt signaling pathway.
Cancer Cell International19:
350. DOI
10.1186/s12935-019-1046-0.ZhangCL, LiZM, SongZH, SongT (
2018a).
Coagulation factor V gene 1691G>A polymorphism as an indicator for risk and prognosis of lower extremity deep venous thrombosis in Chinese Han population.
Medicine97:
e10885. DOI
10.1097/MD.0000000000010885.ZhangS, TaylorAK, HuangX, LuoB, SpectorEB, FangP, RichardsCS (
2018b).
Venous thromboembolism laboratory testing (factor V Leiden and factor II c.*97G>A), 2018 update: a technical standard of the American College of Medical Genetics and Genomics (ACMG).
Genetics in Medicine20:
1489–
1498. DOI
10.1038/s41436-018-0322-z.ZhangX, ZhangP, ShaoM, ZangX, ZhangJ, MaoF, QianH, XuW (
2018c).
SALL4 activates TGF-β/SMAD signaling pathway to induce EMT and promote gastric cancer metastasis.
Cancer Management and Research10:
4459–
4470. DOI
10.2147/CMAR.S177373.