Conversion from epithelial to partial-EMT phenotype by Fusobacterium nucleatum infection promotes invasion of oral cancer cells

The ability of cancer cells to undergo partial-epithelial mesenchymal transition (p-EMT), rather than complete EMT, poses a higher metastatic risk. Although Fusobacterium nucleatum mainly inhabits in oral cavity, attention has been focused on the F. nucleatum involvement in colorectal cancer development. Here we examined the p-EMT regulation by F. nucleatum in oral squamous cell carcinoma (OSCC) cells. We cultured OSCC cells with epithelial, p-EMT or EMT phenotype with live or heat-inactivated F. nucleatum. Expression of the genes involved in epithelial differentiation, p-EMT and EMT were examined in OSCC cells after co-culture with F. nucleatum by qPCR. Cell growth and invasion of OSCC cells were also examined. Both live and heat-inactivated F. nucleatum upregulated the expression of p-EMT-related genes in OSCC cells with epithelial phenotype, but not with p-EMT or EMT phenotype. Moreover, F. nucleatum promoted invasion of OSCC cells with epithelial phenotype. Co-culture with other strains of bacteria other than Porphyromonas gingivalis did not alter p-EMT-related genes in OSCC cells with epithelial phenotype. F. nucleatum infection may convert epithelial to p-EMT phenotype via altering gene expression in OSCC. Oral hygiene managements against F. nucleatum infection may contribute to reduce the risk for an increase in metastatic ability of OSCC.

www.nature.com/scientificreports/ detailed mechanism of Fusobacterium in OSCC cells remains unclear. The fact that F. nucleatum is abundant in the oral cavity of OSCC patients seems to be important in tumorigenesis and progression of oral cancer. Epithelial mesenchymal transition (EMT) is a process by which epithelial cells lose their cell polarity and cell-cell adhesion, and gain mesenchymal properties. It can be seen during development, wound healing, inflammation, and cancer development. EMT contributes to cancer progression by imparting the mesenchymal phenotypes associated with the cells of highly aggressive tumors 15 . EMT is not binary process and cells move through between full epithelial and mesenchymal states, which is called the intermediate state of EMT, partial EMT (p-EMT) 16 . Cancer cells with p-EMT state act like cancer cells with mesenchymal state, but they do not completely lose their epithelial features 17 . We previously showed that p-EMT genes such as SERPINE1, ITGA5, TGFBI, P4HA2, CDH13, and LAMC2 are potent poor prognostic markers in head and neck squamous cell carcinoma (HNSCC) patients by bioinformatic analysis 18 . Thus, understanding the state of p-EMT is crucial in finding a target cancer potential including cancer development and malignancy.
It recently has been reported that F. nucleatum is associated with EMT in cancer development. F. nucleatum decrease the levels of epithelial marker including E-cadherin and increase the levels of EMT-associated transcription factors including Snail and Slug in colorectal cancer progression 19 . Moreover, ZEB1 level is upregulated by F. nucleatum exposure in oral cancer cells 20 . However, the contribution of F. nucleatum to the regulation of p-EMT in OSCC remains unclear.
For F. nucleatum inactivation, the bacterial cells were heat-inactivated at 100 °C for 15 min. Killing was confirmed by absence of growth on sheep blood agar plate (Becton Dickinson, Sparks, MD, USA). Other bacteria, P. gingivalis, S. mutans, P. aeruginosa, and E. coli were also heat-inactivated at 100 °C for 15 min.
To analyze the presence of bacteria, F. nucleatum were stained with CytoTell UltraGreen (AAT Bioquest, CA, USA) in accordance with the manufacturer's instructions. Briefly, the bacterial cells were treated with CytoTell UltraGreen for 15 min at 37 °C, and then treated to HOC621 cells at the concentration of 100 MOI for 3 h. The cells were fixed with 4% formalin for 30 min and then permeabilized with 0.1% Triton X-100 in PBS for 2 min on ice, followed by Hoechst 33,342 for 30 min. The samples were mounted and observed by a Nikon A1 laser fluorescence confocal microscopy (A1 R HD25, Nikon, Tokyo). Images were acquired with the NIS-Elements software (Nikon, Tokyo, Japan).

Colony-forming unit (CFU) determination.
To quantify the viability of bacteria within host cells, colony-forming units were quantified as follows. OSCC cells were cultured on a 48-well plate, followed by treatment with F. nucleatum for 3 h, and then treated with antibiotics (200 μg/mL metronidazole and 300 μg/mL gentamicin) for 60 min to kill the F. nucleatum on the cell surface. The cells were lysed by adding 200 μL of H 2 O and the extract was spread on a sheep blood agar using a spiral plater (EDDY-Jet, IUL Instruments, Königswinter, Germany). Agar plates inoculated were incubated for 5 days in an anaerobic jar at 37 °C, and then, formed colonies were counted according to the manufacturer's instructions.
Cell proliferation assay. Cells were suspended 5,000 cells/well in 96-well plate. After 24, 48, and 72 h, proliferating cells were determined by using cell-counting kit 8 (CCK-8) solution (Dojindo Molecular Technique, Inc.). 10 µl of CCK-8 solution was added in each well and incubated for 1 h in CO 2 incubator at 37 °C. The signal of proliferating cells was measured absorbance at 450 nm using microplate reader model 680 (BIORAD). qPCR. Total RNA was isolated from the cells by using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. RNA concentration and purity were determined by standard spectrophotometric methods using a NanoDrop ND-2000 (NanoDrop Technologies). cDNA was synthesized from 1 µg total RNA using the PrimeScript II reverse transcriptase (Takara Bio Inc.). Expression of p-EMT, EMT and epithelial differentiation-relatedgenes were measured using TB Green Premix Ex Taq II (Takara Bio Inc.) with a LightCycler 96 system (Roche). The primers used in this study are shown in Supplementary Table S1. Relative mRNA expression of each transcript was normalized to the endogenous control GAPDH mRNA and were calculated using the comparative cycle threshold method.
Heatmap was generated by R package pheatmap. Expression values were centered and scaled in the row direction, each gene. www.nature.com/scientificreports/ In vitro invasion assay. For in vitro invasion assay, we used a 24-well cell culture insert with 8 μm pores (3097, Falcon, Becton Dickinson). The membrane of cell culture insert was coated with 20 μg of matrigel (Becton Dickinson) for reconstituting basement membrane substance. 0.5 ml of serum-free medium were added in the lower compartment. 1.5 × 10 5 cells were resuspended in 100 μl of serum-free medium and placed in the upper compartment of the cell culture insert for 12-24 h. After incubation, the cells that penetrated the membrane into the lower side of the filter through the pores were fixed with formalin and then were stained by hematoxylin. The invasion ability of the cells was determined by counting the penetrating cells onto the lower side. We assayed 3 times. The penetrating cells in randomly selected 3 fields were counted for each assay under a microscope at × 100 magnification.
Statistical analysis. For cell proliferation and qRT-PCR analyses, the t-test was used to compare results between two groups. A P-value of < 0.05 was considered significant.

Results
It recently has been shown that F. nucleatum plays a role in EMT induction of cancer cells. Moreover, the ability of cancer cells to undergo p-EMT poses a higher metastatic risk rather than complete EMT. We previously classified the OSCC cell lines as "epithelial", "p-EMT", and "EMT" phenotype by gene expression profiles (Fig. 1A). In this study, four subspecies of F. nucleatum (JCM8532, ATCC10953, JCM11024, and ATCC23726) and five OSCC cell lines (EMT phenotype, HOC313 and HOC719NE; p-EMT phenotype, HOC719-PE; epithelial phenotype, HOC621 and HSC2) were used (Fig. 1B). In our preliminary experiments, the effects of F. nucleatum at MOI = 10 and 100 on cell proliferation and gene expression was similar. Therefore, we used F. nucleatum at MOI = 10 in the following experiments. We cultured OSCC cells with live or heat-inactivated F. nucleatum for 4 days. As shown in Fig. 2A, we confirmed that bacterial cells after heating for 15 min at 100 °C did not grow on blood agar. We first examined the attachment between OSCC cells and bacterial cells. HOC621 cells were co-cultured with green fluorescent-labeled live or heat-inactivated F. nucleatum (JCM8532 and ATCC10953) for 3 h, and then the cells were washed three times by PBS. After co-culture, live or heat-inactivated F. nucleatum was observed on the surface of HOC621 cells (Fig. 2B). We also examined the invasive potential of F. nucleatum (JCM8532, ATCC10953, JCM11024, and ATCC23726) into OSCC cells. After treatment with antibiotics, intracellular live bacterial cells were cultured on blood agar. Bacterial culture revealed that some bacterial cells invaded into OSCC cells (Fig. 2B).
Then, we compared the effects of live or heat-inactivated F. nucleatum on cell morphology, proliferation, and gene expression in HOC313 cells (EMT phenotype) and HOC621 cells (epithelial phenotype). After 48 and 72 h, co-culture with live or heat-inactivated F. nucleatum did not change the morphology of OSCC cells (Fig. 3A). Co-culture with live F. nucleatum, except heat-inactivated F. nucleatum (ATCC10953), did not affect cell proliferation in both HOC313 and HOC621 cells (Fig. 3B). We also examined the cell proliferation of HOC621 cells after co-culture with other subspecies of F. nucleatum (JCM11024 and ATCC23726). Among them, co-culture with live or heat-inactivated F. nucleatum, except heat-inactivated F. nucleatum (ATCC23726), did not affect cell proliferation in HOC621 cells ( Supplementary Fig. S1). Thus, co-culture with most F. nucleatum subspecies did not affect cell morphology and cell proliferation of OSCC cells.
Although F. nucleatum is associated with EMT in cancer development 19 , the involvement of p-EMT in F. nucleatum infection remains unclear. Therefore, we examined the involvement of F. nucleatum in the regulation of p-EMT of OSCC. Expression of epithelial differentiation-related genes, p-EMT-related genes, and EMTrelated genes was examined by qPCR in HOC313 and HOC621 cells with live F. nucleatum (JCM11024 and ATCC23726). In HOC621 cells, but not HOC 313 cells, p-EMT-related and EMT-related genes were remarkably upregulated by co-culture with live F. nucleatum (Fig. 4A,B). Moreover, we confirmed similar tendency in other subspecies of F. nucleatum (JCM11024, and ATCC23726) in HOC621 cells ( Supplementary Fig. S2). The upregulated levels of p-EMT-related genes depended on the bacterial subspecies, indicating that pathogenicity may vary depending on the bacterial subspecies. These findings suggest that live F. nucleatum may upregulate p-EMT genes in OSCC cells with epithelial phenotype.
Next, we examined whether the phenotype brought about by live F. nucleatum were also caused by heatinactivated F. nucleatum. The heat-inactivated F. nucleatum (JCM8532 and ATCC10953) did not alter the expression of p-EMT-related genes in HOC313 cells (Fig. 5A). Interestingly, heat-inactivated F. nucleatum as well as live F. nucleatum upregulated p-EMT-related genes in HOC621 cells (Fig. 5B and Supplementary Fig. S2). Thus, both live and heat-inactivated F. nucleatum upregulated p-EMT-related genes in OSCC cells with epithelial phenotype, but not with EMT phenotype. To confirm these findings, we examined the gene expression in other OSCC cells with epithelial and EMT phenotype after co-culture with live or heat-inactivated F. nucleatum (Fig. 6). Alteration of p-EMT-related genes in epithelial phenotype, HSC2 cells ( Supplementary Fig. S3) was remarkable after co-culture with live or heat-inactivated F. nucleatum, compared with EMT phenotype, HOC719-NE cells (Supplementary Fig. S4). We also examined the effect of live or heat-inactivated F. nucleatum in p-EMT phenotype, HOC719-PE cells. As expected, both live and heat-inactivated F. nucleatum did not alter the expression of p-EMT-related genes in HOC719-PE cells ( Fig. 6 and Supplementary Fig. S5).
It has been shown that p-EMT program in cancer cells enhances the invasive properties. Therefore, we examined the invasion ability of HOC621 cells after treatment with 4 subspecies of F. nucleatum (JCM8532, ATCC10953, JCM11024, and ATCC23726) by in vitro invasion assay. Most live or heat-inactivated F. nucleatum subspecies statistically promoted the invasion ability (Fig. 7A,B). Enhanced invasion ability by F. nucleatum was also observed in HSC2 cells (Supplementary Fig. S6). These findings indicate that F. nucleatum may enhance the invasion ability via upregulation of p-EMT-related genes in OSCC cells with epithelial phenotype. www.nature.com/scientificreports/ To know whether upregulation of p-EMT-related genes in OSCC cells with epithelial phenotype is specific event by F. nucleatum, we examined the gene expression in HOC621 cells after treatment with other strains, such as P. gingivalis (ATCC33277), S. mutans (UA159), P. aeruginosa (PAO1), and E. coli (ATCC23511). These bacteria did not change the morphology of HOC621 cells (Supplementary Fig. S7A). E. coli and S. mutans did not affect cell proliferation, but P. gingivalis and P. aeruginosa suppressed cell proliferation ( Supplementary  Fig. S7B). Most bacterial cells, except P. gingivalis, did not affect the expression of p-EMT-related genes ( Fig. 8A and Supplementary Fig. S8).

Discussion
Bacteria, fungi, viruses, and archaea naturally colonize in the oral cavity. The oral microbiota refers to a highly varied and complicated ecosystem of these organisms. Over 700 bacterial species are endemic to the oral cavity, and oral commensal bacteria play a critical role in the maintenance of a normal oral physiological environment and the development of systemic diseases 23,24 . It recently has been shown that oral commensal bacteria may be involved in the pathogenesis of OSCC 25,26 . The proportion of anaerobes within the biofilms on the surfaces www.nature.com/scientificreports/ of OSCC tissues is higher than that in the healthy control, indicating that OSCC surfaces provide an important reservoir for anaerobic bacteria 27 . There are many reports showing that Fusobacterium were detected in OSCC tissues 28,29 . Moreover, Fusobacterium was enriched in OSCC biopsy compared to fibroepithelial polyp as a control 13 . Thus, distribution of bacteria including Fusobacterium in OSCC tissues may be distinct from that in healthy oral mucosal tissues. As shown in Fig. 2B, Fusobacterium can invade into OSCC cells. Previous report also shows that F. nucleatum adheres to and invades into human gingival epithelial cells (HGECs) and OSCC cells 30 . Interestingly, the spontaneous mutant of F. nucleatum isolated as defective in autoagglutination was unable to invade into HGECs and OSCC cells, suggesting the requirement for bacterial components to their invasion 30 . Moreover, glucose inhibition assay shows that lectin-like interactions are involved in the attachment   30 . In the present study, both live and heat-inactivated F. nucleatum enhanced the invasion ability of OSCC cells via upregulation of p-EMT genes. Although some bacterial cells directly invaded into OSCC cells, most bacterial cells attached on the surface of the cells (Fig. 2B). Notably, heat-inactivated F. nucleatum as well as live F. nucleatum upregulated p-EMT-related genes and enhanced the invasive ability. These findings suggest that heat-resistant bacterial cell components may be involved in promoting invasion ability via upregulation of p-EMT genes. However, further experiments will be required for clarifying this mechanism. In recent years, attention has been focused on the involvement of F. nucleatum in colorectal cancer 8,9,31 . In colon tissue of patients with colorectal cancer, increased amount of F. nucleatum was significantly detected, compared with healthy people 32,33 . Moreover, F. nucleatum colonies in primary colorectal cancer tissues were detected in metastatic liver tissues 34 . Thus, cumulative clinical evidence indicates that the enrichment of F. nucleatum may correlate with colorectal cancer metastasis 32,34-37 . Interestingly, the outer membrane proteins of Figure 8. Remarkable upregulation of p-EMT-related genes by F. nucleatum in epithelial phenotype, HOC621 cells. (A) Heatmap shows the expression levels of epithelial differentiation-related, p-EMT-related, and EMTrelated genes after co-culture with heat-inactivated F. nucleatum, P. gingivalis, S. mutans, P. aeruginosa, and E. coli in epithelial phenotype, HOC621 cells. Gene expression data is shown in Supplementary Fig. S8.  www.nature.com/scientificreports/ F. nucleatum, Fusobacterium adhesion A (FadA) and Fibroblast activation protein 2 (Fap2) have been focused as a pathogen of colorectal cancer. FadA is required for bacterial attachment and invasion of gingival epithelial cells and endothelial cells 30 . FadA-mediated interaction with E-cadherin is involved in attachment and invasion into colorectal cancer cells 8 . Fap2 also plays a critical function in the development of colorectal cancer. Fap2 is involved in the inactivation of immune cells including T cells and NK cells via an immunoreceptor, TIGIT in colorectal cancer 31 . Moreover, Fusobacteria use a hematogenous route to reach colon adenocarcinomas via interaction between Fap2 and Gal-GalNAc 9 . We will examine the involvement of outer membrane proteins of F. nucleatum, FadA and Fap2 in the promotion of invasion via upregulation of p-EMT genes in OSCC cell with epithelial phenotype. Recently, it has been shown that exposure of primary epithelial cell cultures to heat-inactivated F. nucleatum induces downregulation of E-cadherin and upregulation of N-cadherin, vimentin, Snail, matrix metalloproteinase-2 and toll-like receptor 4 (TLR-4) and promotes migration 38 . Another study shows that heat-inactivated F. nucleatum upregulates TLR-4 in oral keratinocyte 39 . TLR-4 is known to be responsible for recognition of LPS and activation of this pathway results in increased production of inflammatory cytokines 40 . Indeed, we showed that heat-inactivated F. nucleatum promoted invasion via upregulation of p-EMT genes in OSCC cells with epithelial phenotype. These findings suggest that F. nucleatum LPS may be involved in p-EMT and/or EMT induction via TLR signaling. Interestingly, previous studies show that LPS promotes EMT via activation of TLR4/ JNK signaling 41,42 . Importantly, the effect of F. nucleatum on upregulation of the genes involved in invasion, survival, and EMT is comparable or greater than other periodontal bacteria, P. gingivalis, Treponema denticola, and Tannerella forsythia. In the present study, S. mutans, P. aeruginosa, and E. coli did not affect the expression of p-EMT-related genes. Interestingly, P. gingivalis, which is also known as periodontal pathogen, upregulated p-EMT genes in HOC621 cells (Fig. 8A). Cumulative evidences suggest that F. nucleatum and/or P. gingivalis among periodontal bacteria may have greater effects of promoting invasion, survival, and p-EMT/EMT during OSCC progression.
It is recognized that the ability of cancer cells to undergo p-EMT, rather than complete EMT, poses a higher metastatic risk 16 . Recent single cell transcriptome analysis in HNSCC identifies several genes involved in p-EMT program 43 . Among these p-EMT-related genes, we identified the prognosis-related genes, such as SERPINE1, TGFBI, ITGA5, CDH13, P4HA2, and LAMC2 18 . Thus, p-EMT program is correlated with poor prognosis of HNSCC. Importantly, the enrichment of F. nucleatum is frequently observed in OSCC tissues compared to healthy oral mucosal tissues. Conversion from epithelial to p-EMT phenotype via alteration of gene expression by F. nucleatum infection may lead to increase a metastatic ability of OSCC (Fig. 8B). As F. nucleatum is well known oral commensal bacterium involved in periodontal diseases, oral hygiene managements for reducing the amount of F. nucleatum may contribute to reduce the risk for an increase in metastatic ability of OSCC.