Abstract

Background

   Advances in science and medicine have led to the identification of
   bacterial virulence factors (including lipopolysaccharide, LPS) and
   their key role in the occurrence and outcome of tumors. However, the
   effect of LPS on oral squamous cell carcinoma (OSCC) has yet to be
   fully understood.

Objective

   Hence, based on host genes related to bacterial LPS, the study
   investigated the potential role and mechanism of oral bacteria in OSCC
   via bioinformatics analysis and experimental validation.

Methods

   The sequencing datasets of OSCC were screened using the GEO database
   and the bacterial LPS-related genes were searched in the GeneCards
   database to identify the LPS-related differentially expressed genes
   (LR-DEGs) in OSCC. The molecular mechanism of bacteria affecting OSCC
   was explored through GO and KEGG enrichment analysis, as well as
   protein-protein interaction (PPI) network and module analysis.
   Subsequently, seven algorithms were integrated to identify the
   LPS-related hub genes (LRHGs), and their diagnostic specificities were
   explored by receiver operating characteristic (ROC) and transcription
   levels were verified by qRT-PCR. Immune infiltration was then analyzed.

Results

   We found a total of 345 LR-DEGs. GO and KEGG enrichment analysis
   demonstrated that the LR-DEGs were mainly enriched in
   inflammation-related pathways including cytokine-cytokine receptor
   interaction and IL-17 signaling, suggesting that bacteria may promote
   the development of OSCC through LPS-related gene-mediated inflammatory
   response. PPI and module analysis results revealed the presence of a
   complex regulatory network involving LR-DEGs. Totally, five LRHGs
   (including Cxcl8, Cxcl10, Il-1β, Il-6 and Mmp9) were screened out.
   Based on ROC analysis, the five LRHGs represented potential diagnostic
   biomarkers for OSCC (AUC > 0.7). The results of qRT-PCR, WB, ELISA and
   IF indicated that all LRHGs were upregulated in OSCC (P < 0.05). Immune
   infiltration analysis showed that LRHGs were closely related to the
   immunocyte infiltration level, suggesting a potential target for OSCC
   immunotherapy. In this study, 345 LR-DEGs and 5 LRHGs were identified
   in bacterial LPS-regulated OSCC progression. Importantly, the 5 LRHGs
   may mediate the OSCC progression in the host through
   inflammation-related pathways. These findings suggest that bacterial
   LPS plays a vital role in OSCC.

Conclusion

   Our study provides novel insights into the pathogenesis and development
   of oral bacteria in OSCC. The LRHGs identified in this study are
   crucial for the diagnosis of OSCC, and also provide new insights into
   the molecular mechanisms and targeted therapies of OSCC.

1. Introduction

   Oral squamous cell carcinoma (OSCC) is one of the most common head and
   neck malignant tumors encountered globally, with high rates of
   incidence and mortality [[28]1,[29]2]. In recent decades, the number of
   cases in Asia has gradually increased [[30]3]. Although the etiology of
   OSCC is reported to be related to smoking, alcohol consumption and
   chewing tobacco and betel nuts, the main risk factors are still
   unknown. Surgery combined with chemotherapy and radiotherapy has been
   widely used during the past four decades, with only a very slight
   improvement in patient survival [[31]4–[32]6]. The prognosis of
   patients with OSCC is still poor, with a 5-year survival rate of < 50%
   in most parts of the world [[33]7]. These challenges encourage
   scientists to search for new risk factors and pathogenic mechanisms.

   Since the first report of the relationship between gastric
   carcinogenesis and Helicobacter pylori published in the 1990s, studies
   investigating the relationship between bacteria and tumors have
   attracted increased attention [[34]8]. Ample evidence suggests that
   bacterial infections are strongly associated with a variety of cancers,
   accounting for about 13% of the global incidence [[35]9,[36]10]. Most
   cancers, such as oral, prostate, breast, and ovarian types exhibit
   unique bacterial profiles and functional relationships [[37]11–[38]13].
   The imbalance between oral bacteria and host can also contribute to
   oral and systemic diseases, as well as tumorigenesis [[39]14]. However,
   the mechanism underlying the role of oral bacteria in OSCC occurrence
   and outcome remains to be explored.

   Actually, inflammation has been considered as one of the dominant
   factors contributing to bacterial carcinogenesis [[40]15]. Bacteria
   induce tumors by creating pro-inflammatory microenvironment or
   interfering with normal immune response [[41]16–[42]18]. The imbalance
   in microbial ecology is closely related to host inflammation and
   occurrence of various cancers, such as colon and gastric cancers
   [[43]19,[44]20]. Additionally, epidemiological studies have found that
   the ecological imbalance of saliva microbiome and the excessive release
   of pro-inflammatory cytokines contribute to OSCC via inflammation and
   related signaling pathways [[45]21]. Lipopolysaccharide (LPS), a major
   cell-wall components of gram-negative bacteria, is a key exogenous
   virulence factors involved in oral bacterial metabolism [[46]22]. LPS
   triggers strong immune responses and leads to severe infections,
   especially inducing inflammation and ferroptosis in periodontal tissue
   [[47]23,[48]24]. LPS alters the host gene expression profile to
   regulate the occurrence and progression of cancers [[49]25].
   Additionally, LPS is an activator of toll-like receptor 4 (TLR4), and
   the activation of downstream signaling pathways (including nuclear
   factors-κB (NF-κB) pathway and MAPK pathway) enhances the invasion and
   migration of cancer cells [[50]26,[51]27]. In addition, LPS and
   LPS-induced pro-inflammatory cytokines promote the expression of
   adhesion molecules on cancer and endothelial cells, thereby promoting
   the metastasis of cancer cells outside normal tissues [[52]21]. A
   comparative analysis of bacterial composition of OSCC and normal
   tissues revealed that the proportion of Fusobacterium nucleatum and
   Pseudomonas aeruginosa was significantly increased in OSCC. Functional
   prediction revealed the enrichment of pro-inflammatory LPS in tumor
   tissue [[53]28]. Thus, exploring the regulatory pattern of oral
   bacteria in OSCC provides insights into disease pathogenesis and
   identification of novel therapeutic targets.

   In this study, we explored the potential mechanisms by which oral
   bacteria affect the occurrence and development of OSCC. Given that LPS
   is the most important bacterial virulence factor that could lead to the
   occurrence and development of various oral cancers, this study first
   combined GEO and GeneCards databases to screen LPS-related
   differentially expressed genes. Gene Ontology (GO) and Kyoto
   Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were then
   conducted to explore the potential role of oral bacteria in the
   occurrence and outcome of OSCC. Subsequently, based on seven integrated
   algorithms, hub genes were screened out to investigate immune
   infiltration as well as the impact of oral bacteria on the occurrence,
   development, and treatment of OSCC. The study investigates the
   relationship between OSCC and oral bacteria provides a standard of
   reference for OSCC prevention and treatment.

2. Materials and methods

2.1. Acquisition and preprocessing of public data

   The expression profiles of OSCC microarrays containing control samples
   and OSCC samples were acquired from the Gene Expression Omnibus
   ([54]http://www.ncbi.nlm.nih.gov/geo/) database, including
   [55]GSE37991, [56]GSE74530 and [57]GSE30784, which were then merged
   using the R package “inSilicoMerging” (version 1.17.0.1). The batch
   effect was eliminated via the “ComBat” method in R package “sva”
   (version 3.20.0).

2.2. Identification of LPS-related differentially expressed genes (LR-DEGs)
in OSCC

   R package “limma” (version 3.40.6) was employed to select
   differentially expressed genes (DEGs) in each OSCC dataset. Genes with
   P < 0.05 and |log fold change (FC)| ≥ 1 were considered as DEGs.
   GeneCards database was used to investigate genes related to bacterial
   LPS. Common genes of the two databases were identified as LR-DEGs.

2.3. Functional and pathway enrichment analysis of LR-DEGs

   In order to explore the potential molecular mechanism of LPS affecting
   OSCC, DAVID ([58]https://david.ncifcrf.gov/) was undertaken to conduct
   GO and KEGG enrichment analyses of the LR-DEGs. GO enrichment analysis
   contained biological processes (BP), molecular functions (MF) and
   cellular components (CC). The cut-off criteria were P < 0.05 and a
   false discovery rate (FDR) < 5%.

2.4. Protein-protein interaction network (PPI) construction, module analysis
and identification of LPS-related hub genes (LRHGs)

   The PPI network of LR-DEGs was constructed in STRING
   ([59]http://www.string-db.org/). Both direct and indirect
   protein-protein functional interaction scores were predicted via this
   online database. The LR-DEGs with interaction combined score ≥ 0.4 were
   extracted for the construction of the PPI network. Cytoscape (version
   3.9.1) was then utilized to visualize the network. Subsequently, module
   analysis was investigated via molecular complex detection (MCODE).
   Moreover, seven algorithms (including MCC, MNC, Degree, Closeness,
   Radial, EPC and stress) in CytoHubba were used to comprehensively
   identify LRHGs. As a result, hub genes overlapping in the seven
   algorithms were identified as LRHGs, and the co-expression network of
   LRHGs was illustrated using GeneMANIA ([60]http://www.genemania.org/).

2.5. Specific evaluation and diagnostic validation of LRHGs

   To determine the diagnostic specificity of 5 LRHGs in OSCC, the R
   package pROC (version 1.17.0.1) was used to perform receiver operating
   characteristic (ROC) curve analysis. The confidence interval was set as
   95%. The area under the ROC curve (AUC) ≥ 0.7 was considered
   significant (Li et al., 2021).

2.6. Clinical samples and ethics Statement

   To experimentally validate these LRHGs, five patients with OSCC who
   never received chemotherapy or radiotherapy were enrolled in this
   study. OSCC tissues and paired adjacent para-cancerous tissue located 5
   cm away from the tumor margin were obtained and immediately frozen in
   liquid nitrogen for subsequent experiments [[61]27]. All participants
   were informed of the experimental procedure and signed the informed
   consent form. The experimental samples were obtained in this study from
   November 25,2023 to January 24,2024. The study was approved by the
   Ethics Committee of Affiliated Hospital of North Sichuan Medical
   College2023ER433−1.

2.7. Validation of LRHGs by qRT-PCR, WB, ELISA and IF

   The tissues were sufficiently grinded, as described in the instruction
   manual. RNA extraction kit (Accurate Biotechnology, China) was used to
   extract total RNA from the OSCC tissues and control tissues. Then, RT
   Mix Kit with gDNA Clean (Accurate Biotechnology, China) was utilized to
   synthesize DNA. The transcription of mRNA was detected using SYBR green
   (Baoguang Biotechnology, China). The relative mRNA expression level of
   LRHGs (Cxcl8, Cxcl10, Il-1β, Il-6, and Mmp9) was normalized with
   beta-actin and determined via 2-ΔΔCt method. The primer sequences of
   the five LRHGs and beta-actin were displayed in Supplementary [62]S1
   Table.

   For western blot (WB) analysis, total protein was extracted from
   tissues using RIPA lysis buffer, and protein concentrations were
   measured using a BCA protein assay kit (Beyotime, China). Equal amounts
   of protein were separated by SDS-PAGE and transferred to PVDF membranes
   (Millipore, USA). The membranes were blocked with 5% non-fat milk and
   incubated overnight at 4°C with primary antibodies specific to MMP9
   (CST, America), and beta-actin (internal control). After incubation
   with HRP-conjugated secondary antibodies, protein bands were visualized
   using an enhanced chemiluminescence (ECL) detection system.

   For ELISA, tissue samples were processed using an ELISA kit
   (Servicebio, China) according to the manufacturer’s protocol to
   quantify the protein levels of CXCL8, CXCL10, IL-1β, and IL-6.

   For immunofluorescence (IF) staining, OSCC and control tissues were
   embedded in OCT compound and snap-frozen in liquid nitrogen. Frozen
   tissue sections (8 µm thick) were cut using a cryostat and mounted on
   glass slides. The sections were fixed in cold acetone for 10 minutes
   and air-dried. After blocking with 5% BSA for 1 hour at room
   temperature, the sections were incubated overnight at 4°C with primary
   antibodies specific to CXCL8, CXCL10, IL-1β, IL-6, and MMP9(antibody
   details are provided in Supplementary [63]S2 Table). Following primary
   antibody incubation, the sections were washed with PBS and incubated
   with fluorophore-conjugated secondary antibodies for 1 hour at room
   temperature in the dark. Nuclei were counterstained with DAPI, and the
   slides were mounted with antifade mounting medium. Fluorescent images
   were captured using a fluorescence microscope.

2.8. Evaluation of immunocytes infiltration levels

   To further explore the differences in immune characteristics between
   OSCC and control samples, Cell-type Identification by Estimating
   Relative Subsets of RNA Transcripts x (CIBERSORTx) was employed to
   evaluate the infiltration level of immunocytes between OSCC group and
   control group by transforming the gene expression matrix into the
   composition of infiltrating immunocyte. Immunocytes with P < 0.05 were
   considered as statistically significant and selected for subsequent
   correlation analysis with LRHGs.

2.9. Statistical analyses

   The statistical analysis commenced with F-tests to assess data
   distribution normality. For data following a normal distribution, t
   tests were conducted. Conversely, the Mann‒Whitney test was employed
   for nonnormally distributed data. To guarantee result reliability and
   reproducibility, each experiment was performed a minimum of three times
   with biological replicates. The results are presented as
   means ± standard errors of the means (SEMs). Statistical evaluation was
   carried out using Prism 8.1 software (GraphPad Software, USA), with P
   values below 0.05 deemed statistically significant.

3. Results

3.1. Identification of LR-DEGs in OSCC

   The detailed flowchart of this study is illustrated in [64]Fig 1. To
   identify novel genes in OSCC associated with LPS, we selected three
   OSCC datasets, including [65]GSE37991, [66]GSE74530, and [67]GSE30784,
   and the detailed information of these datasets is presented in
   [68]Table 1. The sequencing data of all patients in the three datasets
   were then processed following the procedure mentioned above. Compared
   with the data before processing, the results after processing
   demonstrated that the samples of each dataset were clustered and
   intertwined, which indicated favorable batch removal effects ([69]Fig
   2A-[70]F). Subsequently, 5125 DEGs of OSCC were screened out ([71]Fig
   2G) and a total of 1014 LPS-related genes with a relevance score > 0.4
   were obtained from GeneCards database. The DEGs overlapping LPS-related
   genes were identified as LR-DEGs. Totally, 345 common LR-DEGs were
   acquired ([72]Fig 2H). The LR-DEGs were listed in Supplementary [73]S3
   Table.

Fig 1. The flowchart outlining the study design.

   [74]Fig 1
   [75]Open in a new tab

Table 1. Detailed information about the three datasets used in the study.

   ID Author Publish
   Year Country Platform Sample type OSCC samples Normal samples
   [76]GSE30784 Chen C 2011 USA [77]GPL570 Human tissues 167 45
   [78]GSE37991 Lee CH 2013 China [79]GPL6883 Human tissues 40 40
   [80]GSE74530 Oghumu S 2017 USA [81]GPL570 Human tissues 6 6
   [82]Open in a new tab

Fig 2. Data processing and LR-DEGs identification.

   [83]Fig 2
   [84]Open in a new tab

   (A-C) GEO microarray datasets [85]GSE37991, [86]GSE74530 and
   [87]GSE30784 before data processing. (D-F) The GEO datasets microarray
   after data processing. (G) The volcano map of OSCC gene expression
   profile. Red indicated upregulated DEGs; blue indicated DEGs. (H) The
   Venn diagram illustrated LR-DEGs of OSCC.

3.2. Functional and pathway enrichment analysis of LR-DEGs

   In order to explore the potential biological mechanisms of the 345
   LR-DEGs, GO and KEGG pathway enrichment analyses were conducted. The
   results of GO analysis revealed that changes in BP of LR-DEGs included
   signal transduction, inflammatory response and immune response ([88]Fig
   3A). Changes in CC of LR-DEGs were mainly enriched in the basic
   cellular structure, such as plasma membrane, extracellular region and
   extracellular space ([89]Fig 3B). Within the MF, the LR-DEGs were
   significantly enriched in protein binding, identical protein binding
   and receptor binding ([90]Fig 3C). The KEGG pathway analysis indicated
   that the pathways involving cytokine-cytokine receptor interaction,
   pathways in cancer, PI3K-Akt and IL-17 signaling were significantly
   enriched ([91]Fig 3D). Among them, IL-17 signaling pathway was reported
   to be the predominant pathway in host immune inflammatory response
   [[92]28]. In addition to the biological mechanisms associated with
   cancer, these results also highlighted the critical role of LPS-induced
   inflammation in OSCC.

Fig 3. Functional and pathway enrichment analysis of LR-DEGs. (A) Top 10 BP
terms. (B) Top 10 CC terms. (C) Top 10 MF terms. (D) Top 10 KEGG terms.
P < 0.05 was considered statistically significant.

   [93]Fig 3
   [94]Open in a new tab

3.3. PPI network, module analysis and LRHG selection

   PPI networks of LR-DEGs identified in the analysis above were
   constructed by STRING website, and the results were visualized in
   Cytoscape. Furthermore, three closely connected gene modules were
   identified as significant using a graph clustering algorithm known as
   MCODE. Briefly, module 1 comprised 35 nodes and 1070 edges ([95]Fig
   4A); module 2 contained 39 nodes and 504 edges ([96]Fig 4B); and module
   3 included 26 nodes and 146 edges ([97]Fig 4C). The findings suggested
   that there were complex interaction regulatory networks between LR-DEGs
   and they might share potential molecular mechanisms of regulation in
   OSCC.

Fig 4. Significant modules of PPI network and LRHGs identification.

   [98]Fig 4
   [99]Open in a new tab

   (A) Significant module 1 (containing 34 nodes and 1038 edges). (B)
   Significant module 2 (containing 39 nodes and 504 edges). (C)
   Significant module 3 (containing 26 nodes and 146 edges). (D) The five
   overlapping LRHGs based on seven integrated algorithms. (E)
   Co-expression analysis of LRHGs.

   To make the results more accurate and credible, seven algorithms of
   CytoHubba were integratedly applied to identify common hub genes. By
   overlapping the results of seven algorithms, five LRHGs (including
   Cxcl8, Cxcl10, Il-1β, Il-6, and Mmp9)) were selected ([100]Fig 4D and
   [101]Table 2). Subsequently, [102]Table 3 presented the full names and
   partial functional annotation information of the five LRHGs. Besides,
   the co-expression networks and associated functions of LRHGs were
   determined using the GeneMANIA database ([103]Fig 4E). The functional
   enrichment results of LRHGs suggested the mechanism of bacterial LPS
   and the response to immune regulation, which encouraged us to further
   explore the potential association between immune infiltration and OSCC,
   as well as the correlation between immunocytes and LRHGs.

Table 2. The hub genes were obtained by seven algorithms respectively.

   algorithms hub genes
   MCC        IL6 IL1B CXCL8 CXCL10 MMP9 CCL3 CXCL1 CXCL9 ICAM1 CCL20
   MNC        IL6 IL1B CXCL8 CXCL10 MMP9 FN1  PTGS2 STAT1 ICAM1 CCL5
   Degree     IL6 IL1B CXCL8 CXCL10 MMP9 FN1  PTGS2 STAT1 ICAM1 CCL5
   Closeness  IL6 IL1B CXCL8 CXCL10 MMP9 FN1  PTGS2 STAT1 ICAM1 CCL5
   Radiality  IL6 IL1B CXCL8 CXCL10 MMP9 FN1  PTGS2 STAT1 ICAM1 CCL5
   EPC        IL6 IL1B CXCL8 CXCL10 MMP9 FN1  CXCR4 STAT1 ICAM1 CCL5
   Stress     IL6 IL1B CXCL8 CXCL10 MMP9 FN1  PTGS2 STAT1 PPARG LTF
   [104]Open in a new tab

Table 3. Detailed information of the 5 LRHGs.

   Gene Full Name Function
   CXCL8 C-X-C motif chemokine ligand 8 It functions as a chemotactic
   factor by guiding the neutrophils to the site of infection.
   CXCL10 C-X-C motif chemokine ligand 10 Binding of CLCL10 protein to
   CXCR3 can stimulate migration of monocytes, natural killer and T-cell,
   and modulation of adhesion molecule expression.
   IL-1β interleukin 1 beta This cytokine is an important mediator of the
   inflammatory response, and is involved in a variety of cellular
   activities, including cell proliferation, differentiation, and
   apoptosis.
   IL-6 interleukin 6 The protein is primarily produced at sites of acute
   and chronic inflammation, where it is secreted into the serum and
   induces a transcriptional inflammatory response through interleukin 6
   receptor, alpha.
   MMP9 matrix metallopeptidase 9 It involved in the breakdown of
   extracellular matrix in normal physiological processes and disease
   processes
   [105]Open in a new tab

3.4. Evaluation of the diagnostic role of LRHGs in OSCC

   The ROC curve analysis was undertaken to evaluate the diagnostic
   efficacy of LRHGs in OSCC. Based on the mRNA expression profile of
   sequencing data in GEO database, the AUC values of Cxcl8, Cxcl10,
   Il-1β, Il-6, and Mmp9 were 0.88, 0.85, 0.79, 0.77 and 0.90,
   respectively ([106]Fig 5, P < 0.05). These data demonstrated that the
   five LRHGs represented promising diagnostic biomarkers of OSCC.

Fig 5. The ROC analysis and AUC of CXCL8, CXCL10, IL1β, IL6 and MMP9,
respectively.

   [107]Fig 5
   [108]Open in a new tab

3.5. Experimental validation of LRHGs by qRT-PCR, WB, ELISA and IF

   The expression of LPS-related hub genes (LRHGs) was comprehensively
   analyzed in oral squamous cell carcinoma (OSCC) tissues and
   corresponding normal controls at both the mRNA and protein levels.
   Quantitative real-time PCR (qRT-PCR) revealed a statistically
   significant upregulation of all LRHGs, including Cxcl8, Cxcl10, Il-1β,
   Il-6, and Mmp9, in OSCC tissues compared to controls ([109]Fig 6A,
   P < 0.05). At the protein level, both enzyme-linked immunosorbent assay
   (ELISA) and Western blot (WB) analyses corroborated these findings,
   demonstrating a significant increase in the protein expression of
   CXCL8, CXCL10, IL-1β, IL-6 and MMP9 in OSCC tissues relative to normal
   controls ([110]Fig 6B-[111]D, P < 0.05). Additionally,
   immunofluorescence (IF) staining exhibited markedly enhanced
   fluorescence intensity for CXCL8, CXCL10, IL-1β, IL-6 and MMP9 in OSCC
   tissues, further validating the elevated protein expression observed in
   both ELISA and WB experiments ([112]Fig 6E). These data collectively
   suggest that the upregulation of LRHGs is closely linked to OSCC
   progression, likely mediated through the pro-inflammatory effects of
   lipopolysaccharide (LPS), implicating LRHGs as potential contributors
   to OSCC development via LPS-induced inflammatory pathways.

Fig 6. Experimental validation of CXCL8, CXCL10, IL1β, IL6 and MMP9,
respectively.

   [113]Fig 6
   [114]Open in a new tab

   (A) qRT-PCR analysis of differential mRNA expression of LRHGs between
   the OSCC and control groups. (B) ELISA was performed to assess the
   protein expression levels of CXCL8, CXCL10, IL-1B, and IL-6 between the
   OSCC and control groups. (C-D) Western blot analysis of MMP9 protein
   expression between the OSCC and control groups, with quantitative
   analysis conducted using Image-J. (F) Immunofluorescence detection and
   localization of LRHGs protein in the OSCC and control groups. P < 0.05,
   P < 0.01, P < 0.001.

3.6. LRHG expression correlates with immunocytes infiltration in OSCC

   [115]Fig 7A illustrated the infiltration levels of immunocytes
   expression in each sample. Totally, 22 immunocytes subtypes were
   investigated. All immunocytes except for B cell naive, T cell CD4
   naive, resting NK cells, resting CD4 memory T cells, follicular helper
   T cells, activated NK cells, activated dendritic cells, eosinophils and
   neutrophils showed significantly different levels of infiltration
   between OSCC group and control group ([116]Fig 7B). These immunocytes
   with significant differences were selected for subsequent analysis.
   Notably, the four types of immunocytes (including active CD4 memory T
   cells, M0 macrophages, M1 macrophages and active mast cells)
   demonstrated higher infiltration levels of immunocytes in OSCC than in
   control group ([117]Fig 7B). The distinct immunocytes infiltration
   pattern between OSCC group and control group indicated the vital role
   of immune regulation in the pathogenesis of OSCC.

Fig 7. Different immunocyte infiltration in OSCC and their association with
LRHGs.

   [118]Fig 7
   [119]Open in a new tab

   (A) Immunocyte proportions in OSCC and control samples. (B) Evaluation
   of immunocytes infiltration levels between OSCC group and control
   group. Control samples were shown in blue and OSCC samples in red. (C)
   Correlation analysis between LRHGs and immunocytes. *P < 0.05,
   **P < 0.01, ***P < 0.001.

   In order to explore the potential relationship between LPS-related
   genes and immune mechanisms in OSCC, we analyzed the correlation
   between the 13 immunocytes and the LRHGs. The heatmap illustrated that
   the Cxcl8 expression was strongly positively correlated with activated
   mast cells (r = 0.83, [120]Fig 7C). Mmp9 showed strong correlation with
   M0 macrophages (r = 0.78, [121]Fig 7C), while Il-1β was strongly
   correlated with mast cells (| r | ≥ 0.70, [122]Fig 7C). Cxcl10
   exhibited the strongest correlation with M1 macrophages (r = 0.84,
   [123]Fig 7C). The results of correlation analysis suggested that LRHGs
   might regulate the infiltration levels of immunocytes into OSCC tumor
   tissues. Overall, these findings indicated that bacterial LPS was
   closely associated with immune response in patients with OSCC.

4. Discussion

   OSCC is one of the most invasive epithelial malignancies that adversely
   affect patients’ appearance and quality of life [[124]29,[125]30]. The
   incidence of OSCC is notably high, especially in Asian countries
   [[126]31]. However, the pathogenic mechanism is still unclear. The role
   of bacteria in tumor development and progression has attracted
   increased attention, and more than 15% of cases are related to specific
   pathogenic bacteria [[127]32]. The types of oral bacteria and their
   biofunction in OSCC are diverse and contribute to various stages of
   tumor occurrence, development, metastasis, and immune response.
   Therefore, to further explore the potential relationship between OSCC
   and bacteria, bioinformatics and experimental validation were applied
   to explore the bacterial regulatory mechanisms in OSCC.

   Among the multitudinous pathogenic factors of bacteria, LPS is a key
   component of the cell wall of most gram-negative bacteria and induces
   cytotoxicity [[128]33]. It is also the main bacterial constituent
   inducing host immune response and typically plays an important role in
   bacterial infections. LPS is released into the surrounding medium after
   cell division, detachment of outer membrane vesicles, or bacterial cell
   death. It induces pro-inflammatory immune response in host, resulting
   in sepsis and even death [[129]34]. Chronic infection increases the
   risk of malignant tumors [[130]35]. It has been reported that multiple
   pro-inflammatory cytokines in the tumor microenvironment promoted tumor
   growth rather than tumor inhibition [[131]36,[132]37].

   In this study, we identified DEGs of OSCC using the GEO database, and
   combined with the LPS related genes of the GeneCards database to obtain
   LR-DEGs of OSCC. Then, a combination of experimental validation,
   functional enrichment, protein interaction co-expression networks, hub
   gene selection, and immune infiltration evaluation was employed to
   investigate the molecular mechanisms of the regulation of oral
   bacterial on OSCC. Finally, we identified 345 LR-DEGs with significant
   differences from 3 GEO datasets of OSCC. GO functional enrichment
   analysis revealed that DEGs were mainly enriched in inflammation
   related biological processes involving signal transduction,
   inflammatory response, and immune response. Besides, cell composition
   enrichment analysis results (such as extracellular region and plasma
   membrane) and molecular functions enrichment analysis results (such as
   protein binding and receiver binding) were also screened out.
   Importantly, KEGG pathway enrichment analysis results included
   cytokine-cytokine receptor interaction, pathways in cancer, the
   PI3K-Akt signaling pathway and IL-17 signaling, which were mainly
   enriched and strongly related to inflammatory response. Collectively,
   these above results indicated that bacterial inflammation mediated by
   LPS-related genes facilitates the development of OSCC.

   This study ultimately screened out five potential OSCC diagnostic
   biomarkers with high sensitivity and specificity via seven integrated
   algorithms, including Cxcl8, Cxcl10, Il-1β, Il-6, and Mmp9. Cxcl8 is a
   typical chemokine belonging to the CXC family, responsible for
   recruiting and activating neutrophils and granulocytes to the
   inflammatory sites [[133]38]. It is almost undetectable under
   physiological conditions, but the levels of CXCL8 are rapidly induced
   by pro-inflammatory cytokines [[134]39]. Changes in the CXCL8 signaling
   pathway inhibited cellular apoptosis and promoted multidrug resistance,
   thereby reducing the sensitivity of cancer cells to chemotherapy
   [[135]40,[136]41]. Upregulation of CXCL8 expression exhibited an
   increased risk of cancer and poor prognosis of normal patients
   [[137]42]. CXCL10, also known as human interferon-inducible protein 10
   (IP-10), is an intrinsic cytokine involved in regulating adaptive
   immune response and plays a vital role in combating infections and
   malignant tumors [[138]43]. Overexpression of CXCL10 in HeLa cells
   could activate TP53-mediated apoptosis [[139]44]. Previous studies
   reported that CXCL10 had a negative correlation with the survival time
   of the patients, suggesting it could be a poor prognostic indicator of
   pancreatic adenocarcinoma [[140]45].

   IL-1β is an important pro-inflammatory cytokine in the immune
   inflammatory response and a predominant mediator of pyroptosis and
   apoptosis in cells [[141]46]. Increased expression and secretion of
   IL-1β by cancer stem cells might activate the autocrine signaling loop
   by stimulating signaling pathways (such as NF-κB and cAMP-response
   element binding protein pathways) and promoted epithelial mesenchymal
   transformation, cell invasion and metastasis [[142]47]. Besides, the
   overexpression of IL-1β shown poor prognosis of patients with oral and
   nasopharyngeal squamous cell carcinoma [[143]48]. IL-6 is a
   multifunctional cytokine mainly produced in the tumor microenvironment
   and is a key factor driving tumor growth and metastasis [[144]49]. It
   was reported that IL-6 promoted distant metastasis of tumors by
   promoting epithelial mesenchymal transformation of cancer cells
   [[145]50]. IL-6 promoted the proliferation of head and neck carcinoma
   cells in vitro. Moreover, inhibiting IL-6 signaling significantly
   reduced the growth rate of cancer cells [[146]51]. MMP9 is a
   zinc-dependent proteolytic enzyme that degrades extracellular matrix
   (ECM) collagen, which can facilitate tumor invasion and metastasis
   [[147]52]. MMP9 not only plays a role in angiogenesis, metastasis, and
   cancer invasion, but also contributes to the survival and spread of
   cancer cells. Many clinical and experimental studies reported that the
   expression of MMP9 increased with cancer progression [[148]53]. The
   overexpression of MMP9 was associated with poor prognosis of patients
   diagnosed with oral or breast cancer [[149]54,[150]55].

   In this study, all proteins encoded by LRHGs were closely related to
   inflammation of cancer, suggesting that bacterial LPS could regulate
   the progression of OSCC through inflammation. The results of qRT-PCR,
   WB, ELISA and IF demonstrated that LRHG transcriptions were upregulated
   in OSCC, compared with control group. The ROC curve analysis
   illustrated that the five LRHGs had favorable diagnostic value for
   OSCC. LPS is an important factor associated with enhanced host
   inflammatory response. It activated the intracellular signal
   transduction mechanism related to NF-κB and the cell surface receptor
   TLR4, and further upregulated the downstream pro-inflammatory factors,
   culminating in the aggravation of cancer inflammation [[151]56].
   Following LPS stimulation, P65, the downstream NF-κB subunit, was
   phosphorylated after nuclear translocation, inducing other
   pro-inflammatory signaling factors, such as interleukins IL-1β, IL-6
   and tumor necrosis factor-α (TNF-α). Induction of additional chemokines
   and pro-inflammatory factors further exacerbated the inflammatory
   reactions and the destruction of cancer tissue integrity [[152]57]. The
   upregulation of pro-inflammatory cytokines, including CXCL8, CXCL10,
   IL-1β, IL-6 and MMP9 was associated with the progression of host
   bacterial infection. Therefore, the study findings suggested that
   inflammation in cancer wass a key mechanism linking bacterial LPS and
   OSCC, and LPS increased the risk and adverse outcomes of OSCC.

   Immunotherapy is currently an effective strategy for treating advanced
   cancers. However, the effectiveness of immunotherapy depends largely on
   the patients’ tumor microenvironment (TME) [[153]58]. TME refers to the
   presence of non-cancer cells and components in tumors, including the
   molecules synthesized and released, which can be used to predict
   disease outcomes and evaluate immunotherapeutic effects [[154]59]. OSCC
   has characteristic TME, and changes in OSCC TME balance due to altered
   immunocytes populations, immune checkpoints and tumor microenvironment
   factors in cancers are likely to facilitate the escape of cancer cells
   from immune surveillance [[155]60]. t is well known that immunocytes
   infiltration is a key index of the prognosis of OSCC and other cancers
   [[156]61]. Our results demonstrated that LRHGs promoted immunocytes
   infiltration and eventually contributed to the morbidity and metastasis
   of OSCC. Therefore, bacterial LPS was considered to be associated with
   the regulation of OSCC immune response and represents a promising
   target for immunotherapy in patients with OSCC.

   The role of oral bacteria in OSCC pathogenesis has been established
   initially. However, the dominant pathogenic mechanism is still unknown.
   Our study investigated the potential relationship between oral bacteria
   and OSCC based on bioinformation analysis combined with experimental
   validation. In ongoing work, we plan to analyze NGS datasets to provide
   a deeper understanding and explore additional LPS-related pathways.
   Collectively, LPS was found to be a potential etiological factor and a
   promising immunotherapeutic target of OSCC. Further experimental
   evidence is needed to validate the diagnostic role of bacteria in OSCC,
   as well as its prevention and immunotherapeutic effects.

5. Conclusion

   The findings indicated that LPS was a key component regulating the
   interaction between oral bacteria and OSCC. The results of qRT-PCR
   revealed that five LRHGs were upregulated in OSCC. The validation of
   LRHGs and ROC analysis suggested that the five LRHGs (Cxcl8, Cxcl10,
   Il-1β, Il-6, and Mmp9) were potential diagnostic molecular biomarkers
   of OSCC. The results of immune infiltration analysis illustrated that
   immunotherapy targeting LRHGs might improve the prognosis of patients
   with OSCC. In general, our study explored a complex regulatory
   mechanism involving oral bacteria, which affected the progression and
   outcome of OSCC via LPS and inflammation. The findings broadened the
   understanding of the potential molecular mechanisms of OSCC and
   provided novel insights into the development of OSCC treatment.

Supporting information

   S1 Table. Primer sequence for qRT-PCR of human samples.

   (DOCX)
   [157]pone.0329231.s001.docx^ (15.5KB, docx)
   S2 Table. Primary antibodies used in this study.

   (DOCX)
   [158]pone.0329231.s002.docx^ (16.1KB, docx)
   S3 Table. LR-DEGS.

   (XLSX)
   [159]pone.0329231.s003.xlsx^ (13.6KB, xlsx)

Acknowledgments