Assessing the genetic impact of Enterococcus faecalis infection on gastric cell line MKN74

Abstract

Purpose

Enterococcus faecalis (E. faecalis) is an important commensal microbiota member of the human gastrointestinal tract. It has been shown in many studies that infection rates with E. faecalis in gastric cancer significantly increase. It has been scientifically proven that some infections develop during the progression of cancer, but it is still unclear whether this infection factor is beneficial (reduction in metastasis) or harmful (increase in proliferation, invasion, stem cell-like phenotype) of the host. These opposed data can significantly contribute to the understanding of cancer progress when analyzed in detail.

Methods

The gene expression data were retrieved from Array Express (E-MEXP-3496). Variance, t test and linear regression analysis, hierarchical clustering, network, and pathway analysis were performed.

Results

In this study, we identified altered genes involved in E. faecalis infection in the gastric cell line MKN74 and the relevant pathways to understand whether the infection slows down cancer progression. Twelve genes corresponding 15 probe sets were downregulated following the live infection of gastric cancer cells with E. faecalis. We identified a network between these genes and pathways they belong to. Pathway analysis showed that these genes are mostly associated with cancer cell proliferation.

Conclusion

NDC80, NCAPG, CENPA, KIF23, BUB1B, BUB1, CASC5, KIF2C, CENPF, SPC25, SMC4, and KIF20A genes were found to be associated with gastric cancer pathogenesis. Almost all of these genes are effective in the proliferation of cancer cells, especially during the infection process. Therefore, it is hypothesized that downregulation of these genes may affect gastric cancer pathogenesis by reducing cell proliferation. And, it is predicted that E. faecalis infection may be an important factor for gastric cancer.

Introduction

According to the 2018 GLOBOCAN data, gastric cancer has a death rate of 8.2% among ten common cancers and a 5.7% new case rate among five common cancers (Bray et al. 2018). Risk factors include genetics, gastric ulcer, diet, smoking, and alcohol use, but the main risk factor is bacterial infections (Rawla and Barsouk 2019). Infections are very important in the chemotherapeutic response because of their effects on various mechanisms. Due to weakened immune systems and various treatment procedures, cancer patients are more likely to become infected from the flora.

Most cases of stomach cancer are associated with bacterial infections, particularly Helicobacter pylori gastritis, and have been found to increase the likelihood of stomach cancer by about 6 times within 10 years after infection. However, the data on the roles of bacterial infections in cancer are definitely in two different directions. Some researchers suggest that bacterial infections play a protective role, while others highlight the harmful effects associated with the infection.

Enterococcus is a part of the normal intestinal flora of humans and animals (Dubin and Pamer 2014). Enterococcus faecalis (E. faecalis) has a fundamental role in the development of intestinal immunity in the early stages of life in the human gastrointestinal system and plays a role as a protective agent in the regulation of colon homeostasis in newborn babies (Fanaro et al. 2003; Are et al. 2008; Wang et al. 2014). Moreover, due to its anti-inflammatory potential, it is considered to be probiotic in the treatment of certain diseases, such as chronic sinusitis, bronchitis, or infant acute diarrhea (Habermann et al. 2001; Huycke et al. 2002; Nueno-Palop and Narbad 2011; Gong et al. 2017). Besides its beneficial effects it is also known to play important roles in the formation of systemic infections and translocations (Khan et al. 2018; Ben Braïek and Smaoui 2019). It can easily colonize in extra-intestinal and unusual areas and can cause infections throughout certain predisposing conditions such as long-term hospitalizations and some immune compromised conditions.

It has been reported in many studies that the increase of E. faecalis infections in various cancer types is very common. An increase in E. faecalis-related infections has been noted in cancerous lesions of patients with oral and oropharyngeal squamous cell carcinoma (Boonanantanasarn et al. 2012). The increased levels of E. faecalis infection and leading to chromosomal imbalance have been shown in colon cancer (Khan et al. 2018). It triggers poor prognosis by causing invasion in bladder cancer cells (Horsley et al. 2013). It has been shown that cells that interact with E. faecalis trigger the production of reactive oxygen species (ROS), and the mitochondrial genome is impaired in gastric cancer (Strickertsson et al. 2013).

However, uncontrolled overproduction of ROS can lead to carcinogenic, cytotoxic results. On the contrary, it was found that the colorectal cancer cell line HCT-116-cocultured with E. faecalis downregulated the expression of the fasting-induced adipose factor (FIAF) gene, which is normally associated with the development of some cancer types (Grootaert et al. 2011). It has been suggested that E. faecalis strains isolated from human breast milk may be candidates for the prevention and treatment of cancer when they inhibit the proliferation of breast cancer cells (Hassan et al. 2016).

These opposed data can contribute significantly to understanding of cancer progression when analyzed in more detail whether the interaction of the E. faecalis will benefit or harm the host. Although there are studies reporting genetic alterations in gastric cancer caused by E. faecalis that may affect the chemotherapeutic response and patients’ prognosis, withal their roles, related pathways, and mechanisms remains unclear.

In the current study, we aimed to determine the E. faecalis-related altered genes, the network relations of these genes with each other and to relevant pathways using all genome profiles of gastric cancer cells. The results can provide important contribution how prognosis can be affected in the presence of infection.

Materials and methods

Microarray gene expression data

The gene expression data was obtained from the Cancer Genome Project (CGP) database (http://www.cancerrxgene.org/). Transcription profile data of human MKN74 gastric epithelial cells treated with Enterococcus faecalis were obtained from Array Express (E-MEXP-3496) (Strickertsson et al. 2014).

Processing and normalization of data

The raw data from Array Express were normalized with the Affy package in the R software. Normalized transcription profile data consists of 23,344 different genes/54,675 probe sets. The data contains control day 1, control day 5, E. faecalis infected day 1, and E. faecalis infected day 5 groups’ whole genome expression data, and they were triplicated.

Variance, t test, and linear regression analysis

Among the groups, significant genes with a standard deviation value above 1 were identified. In order to group the identified genes more specifically, Pearson’s correlation absolute P value based on the correlation coefficient was calculated and genes above 0.05 were selected. In addition, Pearson correlation coefficient value (R value) was calculated and genes above 0.99 were selected. Genes with a P value less than 0.01 were selected by comparing the expression values of the genes with the control groups by two-tailed t test analysis.

Analyses were done using GraphPad Prism 5.0 (Graphpad Prism 5 Software, San Diego, CA, USA). To determine the difference between groups two tail t test were performed. Genes with a t test P value less than 0.01 were selected. P value less than 0.05 were selected in Pearson analysis.

Hierarchical clustering

Genes determined in linear regression analysis were hierarchically clustered with mean standardized gene expression values with the Euclidean Gene Cluster 3.0 program. The data was standardized after cluster analysis, and the standardized data were viewed using Treeview. Hierarchical clustering was performed by clustering both genes and arrays using Euclidian distance as similarity metric and complete linkage as clustering method.

Network analysis

To generate a network based on coexpression and genetic interactions, the GeneMANIA software were used and genes in similar pathways were identified with Cytoscape (Shannon et al. 2003). Twelve differentially expressed genes from the 138 genes/165 probe sets were used as input. The datasets have been integrated, analyzed, and visualized to find out if they have functionally similar genes associated with each other and to identify related functions for different gene groups in the network. Thus, the network relationship of these genes was determined. The software scores each gene (node). The size of nodes directly correlated with the score. For example, NDC80 gene has virtually 0.56 score and CENPT has 0.05 score. This directly correlated with the size of the nodes. Higher scores indicate genes that are more likely to be functionally related and more likely has co-expression potential.

Pathway enrichment analysis

To understand biological linkage behind these genes, “Database for Annotation, Visualization and Integrated Discovery” (DAVID) software was used. The pathways associated with our genes were identified.

Gene set enrichment analysis (GSEA)

The gene set enrichment analysis (GSEA) was carried out in concordance with GSEA guideline procedure (http://software.broadinstitute.org/gsea/docGSEAUserGuideFrame.html). E-MEXP-3496 data was used in order to perform the analysis. This data includes 54,675 probe sets (23,344 different genes). Analysis was performed between the control day 5 group and the infected day 5 group to understand the pattern of E. Faecalis infection among these groups. The main purpose of this analysis is to determine which gene significantly enriched in which gene set that belongs to the GSEA as well as to understand which gene set enriched in which groups.

GSEA calculate the enrichment score (ES), normalized enrichment score (NES), nominal P value (NOM P value), false discovery rate q value (FDR q value), and familywise error rate P value (FWER). The ES value indicates gene’s maximum deviation in gene sets; in other words, this score helps to find most upregulated genes. NES value represents the connection or difference between gene sets and gene expression. The higher NES value shows the elevation of permutations. Hence, higher NES value increase significance of gene sets. In addition to ES and NES values, NOM P value evaluates the importance of ES calculation. Therefore, NOM P value directly correlated with ES as well as NES value. Increase of NOM P value show critical role of ES. On the other hand, FWER P value indicates false positives probability of NES and so, lower FWER P value directly and significantly correlated with correctness of NES calculation. Moreover, FDR q value is the most vital value of this analysis. This value needs to be lower than 0.25 and even, when this value become smaller, the enrichment of gene sets is more meaningful.

Results

Time-dependent gene expression alterations in the MKN74 gastric cancer cell line treated with E. faecalis

Whole genome expression data were analyzed by linear regression to determine gene expression alterations between control day 1, infected day 1, and infected day 5 groups of human gastric carcinoma MKN74 cells. According to the results, 138 genes corresponding to 165 probe sets with a standard deviation value above 1.0, P value below 0.05, and Pearson R value above 0.99 showed statistically significant expression alteration. For further analyses, we focused on these genes that altered expression diversity between groups. In the time-dependent E. faecalis-treated groups, 10 probe sets were positively correlated and upregulated, and 155 probe sets were negatively correlated and downregulated (Supplementary Table 1).

Hierarchical cluster analysis demonstrated gene alterations between 3 certain groups; control day 1, infected day 1, and infected day 5. Results defining as 155 of the probe sets were negatively correlated, highly expressed in the control day 1 group, and expression decreased at infected day 1 group and so on at infected day 5 group. And conversely 10 of the probe sets were positively correlated, low expressed in the control day 1 group, and expression increased at infected day 1 and so on at infected day 5 group. Top 50 genes were visualized in the figure and the rest were given as supplementary data (Fig. 1) (Supplementary Figures 1, 2, 3). So, there were significantly altered genes between control and treated groups.

Fig. 1

Hierarchical clustering of 138 statistically significant genes/165 probe sets in the three groups specified. The analysis reveals sensitive low expressions (green), intermediate (black), and high expressions (red) of 138 genes for control day 1, infected day 1, and infected day 5. Designated groups are clearly classified. Top 50 genes were given in the figure

Gene alterations due to treatment with E. faecalis

In order to determine whether this expression change was caused by E. faecalis or cancer itself, variance analyses and t test were performed between control day 5 groups’ and infected day 5 groups’ 138 genes/165 probe set expression data. To better analyze gene expression alteration, long range group “infected day 5” group was selected for variance analysis and t test. Thus, 12 statistically significant genes (NDC80, NCAPG, CENPA, KIF23, BUB1B, BUB1, CASC5, KIF2C, CENPF, SPC25, SMC4, KIF20A) corresponding 15 probe sets with a standard deviation above 1.0 and a t test P value less than 0.01 were determined (Table 1) (Supplementary Figure 4).

Table 1 The list of 12 genes (15 probe sets) which have the most alterations in expression. These genes have standard deviation above 1.0 and t test value less than 0.01 between 5 days control group and 5 days treated group. These significant values indicate that the change occurred due to infection.

In linear regression analysis, all of these 12 genes were found to be negatively correlated and downregulated. Figure 2 shows the expression alterations of these 12 genes time-based between control day 1, infected day 1, and infected day 5 (Fig. 2). Gene expressions have been shown to decrease as the infection process progresses in all these 12 genes.

Fig. 2

Twelve genes/15 probe sets that show significant downregulation after treatment with live infection. Statistically significant alterations were detected between control day 1, infected day 1, and infected day 5 for all probe sets

Network construction and identification of key candidate genes using gene-gene interaction

Network analysis was performed with Cytoscape to better demonstrate the biological linkage of these 12 genes whose expression decreased from day 1 to day 5 during infection. As can be seen from the figure, it has been shown that the 12 genes we identified have a strong network relationship between each other and with other candidate genes (Fig. 3). The linking line between genes illuminates the network of these genes. The thickness of the linking line determines the power of connection of the related genes. For instance, the linking line between NCAPG and SMC4 represent one of the thickest lines. This indicates the link formed between these genes has been determined to be stronger by studying more clearly. Additionally, the black nodes indicate the target genes giving by authors. On the other hand, the gray nodes demonstrate the genes which associated genes determined by GeneMANIA application.

Fig. 3

Network analysis of 12 genes that are statistically significant. The figure shows that these 12 genes have a strong network connection (Cytoscape). The linking line between genes illuminates the network of the genes. The thickness of the linking line determines the power of connection of the related genes. This indicates the link formed between these genes has been determined to be stronger by studying more clearly. Additionally, the black nodes indicate the target genes giving by authors. On the other hand, the gray nodes demonstrate the genes which associated genes determined by GeneMANIA application. The size of nodes directly correlated with score

Functional enrichment of genes and correlations with pathways

Pathway analysis of biological processes was done with DAVID software to reveal the relationship of these 12 genes with cellular functions and pathways and to understand their new meaning. The four important pathways have been identified such as cell cycle, pyrimidine metabolism, mismatch repair, and P53 signaling pathway which are associated with 12 genes during the exposure to E. faecalis. The relationship of these pathways with cancer development, such as the importance of cell cycle in gastric cancer cells, is very clear (Table 2).

Table 2 Pathways related to the genes are linked. It is seen that important pathways in cancer progression are related to our genes. Most of these genes are linked to the cell cycle pathway from the database for annotation, visualization, and integrated discovery (DAVID)

The significantly enriched gene sets and their ES, NES, NOM P value, FWER P value, and FDR q value represented in Supplementary Table 2. According to GSEA, the cell cycle checkpoint as well as the mitotic cell cycle checkpoint gene set was found as significantly correlated with the genes that are present in the data. These gene sets even correlated with pathway analysis in Table 2 and other emphasized studies. The enriched genes were NCD80, BUB1, BUB1B, and CENPF in these two gene sets.

Also, Fig. 4 demonstrated the cell cycle checkpoint gene set plot. This graph showed genes with enriched NCD80, BUB1, BUB1B, and CENPF in the control day 5 group, also called positively correlated. On the other hand, the same genes were downregulated and negatively correlated in the infected day 5 group. The same results were obtained for the mitotic cell cycle checkpoint gene set (Fig. 5).

Fig. 4

The cell cycle checkpoint enrichment plot was represented. The black straight line refers the enriched genes in the groups. Red part contains the genes that positively correlated in control day 5 group and were upregulated in control day 5. On contemporarily, the blue line includes downregulated or in other words negatively regulated genes that belong to infected day 5

Fig. 5

The mitotic cell cycle checkpoint enrichment plot was represented. The black straight line refers the enriched genes in the groups. Red part contains the genes that positively correlated in control day 5 group and were upregulated in control day 5. On contemporarily, the blue line includes downregulated or in other words negatively regulated genes that belongs to infected day 5

Discussion

In recent years, the association between intestinal flora-related microbiota and gastric cancer has increased the interest to its roles in human health with crucial findings. E. faecalis infections have become important due to their potential to affect cell cycle which can alter the response to chemotherapy. E. faecalis is a part of the normal intestinal flora that can pass through the gut and cause systemic infection (Khan et al. 2018). If the barrier is violated, the intestinal flora such as E. faecalis can translocate and cause infection (Ramphal 2004). In this study, it is aimed to identify E. faecalis-related gene expression alterations, their network relationships, and associated pathways using all genome expression data of MKN74 gastric cancer cells.

NDC80, NCAPG, CENPA, KIF23, BUB1B, BUB1, CASC5, KIF2C, CENPF, SPC25, SMC4, and KIF20A gene expression alterations were identified depending on time after treatment with E. faecalis of gastric cell line MKN70. These genes were identified as downregulated in gastric cancer cells, and the treated and untreated groups (P < 0.05) were hierarchically forming a very distinct cluster, as expected.

Gene set enrichment analysis supported the results that some crucial genes function effectively in the cell cycle as a result of E. Faecalis infection. NDC80, BUB1B, BUB1, and CENPF genes showed a statistically significant difference in cell cycle checkpoint and mitotic cell cycle checkpoint gene sets associated with cell cycle. Genes found statistically significant as a result of the comparison between the control day 5 and the infected day 5 were enriched in the specified biological gene sets. This enrichment supports that these genes have important effects on the cycles of cells. In addition, these effects trigger cancer development by causing critical changes in cell function, as indicated.

NDC80 is a kinetocol protein that combines with other proteins to form a microtubule binding site. It has important roles in the cell cycle and an abnormality can lead to apoptotic cell death. Thus, downregulation of this gene may cause abnormalities in the cell cycle of gastric cells after E. faecalis infection. In this case, cells can escape from proliferation and go into apoptosis. High expression of the NDC80 gene in pancreatic and breast cancer cells is associated with poor prognosis and plays important roles in tumor formation in pancreatic, breast, and stomach cancer (Meng et al. 2015; Tang and Toda 2015; Liu et al. 2018). NCAPG is the mitotic gene that plays important roles in the cell cycle. High expression of NCAPG is associated with proliferation in gastric cancer (Song et al. 2018). A downregulation of this gene may cause a decrease in proliferation of gastric cancer cells. CENP-A and CENP-F are histone variants that ensure kinetochores and centromeres to form and function properly. High expressions are related with cancer progress and have an important role in cell division (Sun et al. 2016). High expression can be predicted as a biomarker in the malignant progression and poor prognosis of many types of cancer. The decrease in the expression of this gene can cause irregularity in the regulation of the centromere, causing genome imbalance in gastric cancer. Kinesin superfamily proteins (KIFs) are microtubule-dependent motor proteins and function as oncogenes in cancer cells. It was shown that KIF overexpression is associated with poor prognosis and plays important roles in cancer progression and metastasis (Yang et al. 2019). BUB1 is a mitotic spindle checkpoint protein (budding uninhibited by benzimidazoles 1). It has been suggested that high expression of BUB gene family members in gastric cancer correlates strongly with tumor proliferation (Stahl et al. 2017). Its expression was inversely correlated with the residual tumor stage, and low BUB1 expression was associated with shorter survival (Grabsch et al. 2003; Stahl et al. 2017) (P < 0.001). CASC5 (cancer susceptibility candidate 5) is required for creation of kinetochore-microtubule attachments and chromosome segregation. It has been shown to play important roles in various types of cancer and has been shown to be a new oncogene in lung adenocarcinoma (Cui et al. 2020). Its downregulation in gastric cancer may be important. Structural maintenance of chromosome 2 (SMC2) gene encodes condensin complexes and responsible from chromosomal stability (Je et al. 2014). It is showed that SMC2 gene has low expression in gastric cancer, and it is a risk loci for breast and ovarian cancer (Je et al. 2014; Murakami-Tonami et al. 2014; Kar et al. 2016). The SPC25 gene encodes a protein that may be involved in kinetochore-microtubule interaction, and it is shown that upregulation of SPC25 promotes breast cancer (Wang et al. 2019).

It seems that all genes often play a role in the progression of cancer by affecting cell proliferation and especially the cell cycle. An irregularity in the cell cycle pathway is an important factor that could lead gastric tumor formation. Some cyclin-related kinases with bacterial infection are known to upregulate gastric cancer cells. Activation and regular functioning of these cyclin-depended kinases occur due to the progress of the cell cycle (Molaei et al. 2018). Cyclin-depended kinases play an important role in the G1/S phase transition in the cell cycle by targeting E2F transcription factors. An abnormality occurring in this molecular mechanism causes damage to this connection and causes the development of gastric cancer cells.

It has been reported that most of identified genes in this study are highly expressed in various cancers and are associated with a poor prognosis. Thus, cancer progression may also be expected to decrease due to E. faecalis infection. The relationship of some of these genes with gastric cancer was first shown in this study. As a result of network and pathway analysis, it was determined that these genes are associated with important pathways in cancer progression. These pathways are cell cycle, pyrimidine metabolism, mismatch repair, P53 signaling pathway, endocytosis, and spliceosome. It is obvious that most of these pathways are associated with cancer progression and proliferation. And this allows us to positively intervene in the progression of gastric cancer. The limitation of this study is lack of incorporation of in vivo experiments.

Conclusion

In this study, it has been shown that E. faecalis infection may have important effects on the proliferation of cells in gastric cancer. The results are important indicators that infections caused by E. faecalis directly or indirectly have an effect on the progression of gastric cancer, and this effect may favor the host in fighting with cancer. Most of the genes we identified have functions that slow down cell proliferation. More in vitro and in vivo studies are needed to reveal which structural or functional components are responsible for this adverse effect on gastric cancer cells.