Neutrophil Transcriptional Deregulation by the Periodontal Pathogen Fusobacterium nucleatum in Gastric Cancer: A Bioinformatic Study

Background Infection with the periodontal pathogen Fusobacterium nucleatum (F. nucleatum) has been associated with gastric cancer. The present study is aimed at uncovering the putative biological mechanisms underlying effects of F. nucleatum–mediated neutrophil transcriptional deregulation in gastric cancer. Materials and Methods A gene expression dataset pertaining to F. nucleatum-infected human neutrophils was utilized to identify differentially expressed genes (DEGs) using the GEO2R tool. Candidate genes associated with gastric cancer were sourced from the “Candidate Cancer Gene Database” (CCGD). Overlapping genes among these were identified as link genes. Functional profiling of the link genes was performed using “g:Profiler” tool to identify enriched Gene Ontology (GO) terms, pathways, miRNAs, transcription factors, and human phenotype ontology terms. Protein-protein interaction (PPI) network was constructed for the link genes using the “STRING” tool, hub nodes were identified as key candidate genes, and functionally enriched terms were determined. Results The gene expression dataset GEO20151 was downloaded, and 589 DEGs were identified through differential analysis. 886 candidate gastric cancer genes were identified in the CGGD database. Among these, 36 overlapping genes were identified as the link genes. Enriched GO terms included molecular function “enzyme building,” biological process “protein folding,'” cellular components related to membrane-bound organelles, transcription factors ER71 and Sp1, miRNAs miR580 and miR155, and several human phenotype ontology terms including squamous epithelium of esophagus. The PPI network contained 36 nodes and 53 edges, where the top nodes included PH4 and CANX, and functional terms related to intracellular membrane trafficking were enriched. Conclusion F nucleatum-induced neutrophil transcriptional activation may be implicated in gastric cancer via several candidate genes including DNAJB1, EHD1, IER2, CANX, and PH4B. Functional analysis revealed membrane-bound organelle dysfunction, intracellular trafficking, transcription factors ER71 and Sp1, and miRNAs miR580 and miR155 as other candidate mechanisms, which should be investigated in experimental studies.


Introduction
Gastric cancer is considered the sixth most common cancer globally [1]. A majority of gastric cancer cases occur in developing nations, and it is one of the chief causes of cancer-related morbidity and mortality [2]. Microbial factors are understood to play a central role in gastric cancer pathogenesis, and the best established among these is Helicobacter pylori (H. pylori) infection [3,4]. An increasing number of studies have shown an association of several specific microbial species and the gastric microbial community or microbiome's composition with gastric cancer [5][6][7][8].
Recently, a meta-analysis of gastric mucosa and associated microbiota demonstrated the periodontal pathogens Fusobacterium nucleatum (F. nucleatum), Parvimonas micra, and Peptostreptococcus stomatis as interacting and hub nodes associated with other gastric cancer-associated species and tumor status [9]. The periodontal pathogen F. nucleatum has been most strongly implicated in colorectal cancer (CRC) and is known to induce inflammation and suppress anticancer immune responses in CRC. F. nucleatum infection of neutrophils is known to induce NETosis [10]. In CRC, the circulatory transmission of F. nucleatum is the dominant mechanism [11], which suggests that systemic F. nucleatum and its immune signatures may be similarly relevant in other associated cancers. In particular, some F. nucleatum strains are shown to impede neutrophil-mediated oxidative killing [12], which could be implicated in its role in gastric cancer pathogenesis. In case of H. pylori, also a gram-negative pathogen, infection is also shown to promote N1 neutrophil subtype marked by nuclear hypersegmentation [13] but such mechanisms in case of F. nucleatum stimulated neutrophils are not yet investigated. As neutrophils play a central role in the tumor microenvironment [14], the role of F. nucleatum-induced neutrophil deregulation in gastric cancer merits further investigation. Tumor-activated neutrophils infiltrate the lesion and play a key role in the progression of gastric cancer via STAT3related mechanisms [15], and the interaction of gastric cancer cells with tumor neutrophils promotes their migration, epithelial-mesenchymal transformation (EMT), and invasion [16]. Considering the paucity of research in this domain, bioinformatic approaches may reveal neutrophil transcriptional mechanisms relevant to gastric cancer. Therefore, the present study focused on uncovering neutrophil-related genes and molecular factors, which could be considered candidate mechanisms in gastric cancer via bioinformatic investigation.   [17] describing F. nucleatum-mediated regulation of neutrophil genes was downloaded from the Gene expression omnibus (GEO). Differential gene expression (DEG) analysis was performed using the GEO2R tool. Data were log transformed and normalized, and limma precision weights were applied. A significance level cut-off of p = 0:05 with Benjamini and Hochberg (false discovery rate) correction was used to screen DEGs. Candidate human genes associated with gastric cancer from all available studies in the database were downloaded from the "Candidate Cancer Gene Database (CCGD)" [18]. The DEGs and candidate gastric cancer genes identified in the earlier step were overlapped using a Venn diagram, and shared genes were identified as "link" genes between F. nucleatum-mediated neutrophil transcriptome alteration and gastric cancer.

Functional
Profiling of Link Genes. The link genes list was subjected to functional profiling analysis using the web-based tool "Gprofiler" [19]. Here, the organism of interest was selected as "Human," only annotated genes were used as input, and the customized algorithm g:SCS significance threshold set at 0.05 was used for identification of enriched terms that was used.

Protein-Protein Interaction (PPI) Network and
Functional Enrichment Analysis. PPI network construction with the link gene list as input was done using the STRING webtool [20]. A full STRING network with interaction sources including text mining, experiments, databases, coexpression, neighborhood, gene fusion, and co-occurrence was constructed. A minimum required interaction score was set as 0.15, and network edges represented the confidence measure. Network characteristics, "hub" genes, and functionally enriched terms in the network were determined.

Link Gene
Identification. The analysis of the gene expression dataset GEO20151 identified 589 annotated DEGs (Table S1). Table 1 displays the top 20 DEGs ranked by the adjusted p value.
Using the CCGD database, 886 annotated candidate gastric cancer human genes were identified (Table S2). Table 2 shows the top 20 candidate gastric cancer genes ranked by the number of supporting studies.
A Venn diagram was constructed, and the overlapping genes were identified, which showed 36 link genes ( Figure 1). The 36 link genes are listed in Table 3.

PPI Network and Functional Enrichment
Functional enrichment analysis depicted multiple terms related to Extracellular exosomes, extracellular organelle, extracellular vesicle and membrane protein complex and tissues including blood cells and digestive glands (Table 4).

Discussion
The present identified key molecular mechanisms, which may link F. nucleatum-stimulated neutrophil transcriptomic alterations with the development of gastric cancer. Among the DEGs in F. nucleatum-stimulated neutrophils, 36 genes 3 Disease Markers were documented as gastric cancer candidate genes. The most significant genes among these included DNAJB1, EHD1, and IER2. DnaJ/Hsp40 (heat shock protein 40) proteins are key proteins for protein biology via stimulation of ATPase and are shown to play a role in p53 ubinquination to promote cancer cells in vitro [21]. EHD1 (Eps15 homology (EH) domain-containing protein 1) plays an important role in receptor-mediated endocytic recycyling [22], shows to promote tumor growth, and is implicated in resistance to cisplatin in case of non-small-cell lung cancer [23].
Human immediate early response 2 (IER2) is a nuclear protein that is implicated in cancer via transcriptional regulation of endothelial motility and adhesion via a FAKdependent mechanism [24], thereby regulating tumor angiogenesis. Apart from DNAJB1 and EHD1, the PPI network analysis showed CANX and PH4B as the top hub genes. Calnexin or CANX is an ER stress chaperone transmembrane protein involved in glycoprotein folding, is considered a prognostic indicator and therapeutic target in CRC [25], and is found to restrict antitumor CD4+ and CD8+ T cells
Functional enrichment analysis of the link genes and PPI network was conducted, and consistency in the findings was evident. Several extracellular processes including exosome, membrane protein complex, vesicles, and intracellular membrane-bound organelle were seen as enriched components in the PPI network. Protein folding and associated cel-lular components were evident as enriched, underscoring the potential relevance of the ER stress response as a linkage mechanism [37]. The 2 enriched transcription factors included ER71 and Sp1. The Ets transcription factor Er71 is a key regulator in endothelial and hematopoietic stem cell development [38] and recently has been reported as a valuable target to block tumor angiogenesis [39]. SP1 is shown to transcriptionally regulate oncostatin M receptor in gastric cancer and thereby contribute to cancer progression [40]. SP1 is also implicated in neutrophil elastase-mediated increase in mucin gene receptors [41] and thus may play a role in stimulated neutrophil-mediated deregulation of the mucous barrier [42].  Disease Markers The role of F. nucleatum in CRC is well studied. It has multiple adhesins, and Fap2-mediated adhesion of F. nucleatum to epithelial cells is shown to induce a proinflammatory cascade, whereas Fap2-independent mechanisms are demonstrated in CRC neutrophils and macrophages, which together increase proinflammatory signaling to increase tumor invasion, seeding, and metastatsis [43]. In the colon, F. nucleatum is shown to disrupt epithelial barrier integrity by damage to tight junctions and induction of cytokines of helper T cells [44]. Pathogenic strains of F. nucleatum are shown to induce MUC2 and TNF secretion from colonic cells [45]. The interaction of F. nucleatum with mucins warrants further investigation in the context of gastric cancer. The 2 enriched miRNAs included miR 580 and miR 155. miR 580 has been shown to inhibit chemokine ligand 2 (CCL2) production in the hepatocellular carcinoma tumor microenvironment [46]. miR-155 is involved in neutrophil NETosis [47] and is considered a key factor interlinking inflammation with cancer [48]. miR-155 was found to play a tumor suppressor role in gastric cancer [49]. The enriched GO terms and compartments in the PPI network supported the role of intracellular membrane trafficking as a key cancer mechanism harnessed by F. nucleatum stimulation of neutrophils [50].
Taken together, the findings of this bioinformatic analysis revealed several possible molecular mechanisms by F. nucleatum-induced neutrophil gene deregulation that may promote gastric carcinogenesis. At the same time, these findings are limited by the small sample number in the analyzed gene expression dataset and the lack of validation experiments to support the relevance of the highlighted candidate genes, transcription factors, cellular processes, and miRNAs. Furthermore, the effects of F. nucleatum are likely to be subspecies or strain-specific and should be investigated in future research. F. nucleatum strains with higher invasive capacity  [51], which raises the need for phylotype and functional characterization in context of its role gastric cancer. The present findings should be verified in experimental research models that investigate the candidate link genes and functional mechanisms involved in F. nucleatum-mediated neutrophil plasticity relevant to gastric cancer pathogenesis. Cell model experiments, animal experiments, and clinical examination of the theoretical premises established in this study are warranted. The present investigation focused on the role of F. nucleatum-stimulated neutrophils alone in gastric cancer but the tumor microenvironment constitutes of varied immune cell populations that may be deregulated by F. nucleatum and also warrant deeper investigation.

Conclusion
F nucleatum-induced neutrophil transcriptional activation may be implicated in gastric cancer via several candidate genes including DNAJB1, EHD1, IER2, CANX, and PH4B among the top genes of interest. Putative key functional mechanisms included membrane-bound organelle dysfunction and intracellular trafficking along with the modulation of transcription factors ER71 and Sp1 and miRNAs miR580 and miR155.

Data Availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest
The authors declare that they have no competing interests.

Authors' Contributions
Ting Zhou (email: ting.zhou@xs.ustb.edu.cn) as the first and corresponding author conceptualized the research idea and study design, performed the bioinformatic analyses, wrote the manuscript, and administered and supervised the whole research project. XM, DW, WF, and XL reviewed and edited the manuscript. All coauthors read and approved the whole manuscript. Table S1: list of significant DEGs in the gene expression dataset GEO20151. Table S2: 886 annotated candidate gastric cancer human genes identified in the CGGD database. Table S3: functional enrichment analysis results from "G:profiler." (Supplementary Materials)