Host-microbe computational proteomic landscape in oral cancer revealed key functional and metabolic pathways between Fusobacterium nucleatum and cancer progression

Abstract

Oral squamous cell carcinoma (OSCC) is the most common manifestation of oral cancer. It has been proposed that periodontal pathogens contribute to OSCC progression, mainly by their virulence factors. However, the main periodontal pathogen and its mechanism to modulate OSCC cells remains not fully understood. In this study we investigate the main host-pathogen pathways in OSCC by computational proteomics and the mechanism behind cancer progression by the oral microbiome. The main host-pathogen pathways were analyzed in the secretome of biopsies from patients with OSCC and healthy controls by mass spectrometry. Then, functional assays were performed to evaluate the host-pathogen pathways highlighted in oral cancer. Host proteins associated with LPS response, cell migration/adhesion, and metabolism of amino acids were significantly upregulated in the human cancer proteome, whereas the complement cascade was downregulated in malignant samples. Then, the microbiome analysis revealed large number and variety of peptides from Fusobacterium nucleatum (F. nucleatum) in OSCC samples, from which several enzymes from the L-glutamate degradation pathway were found, indicating that L-glutamate from cancer cells is used as an energy source, and catabolized into butyrate by the bacteria. In fact, we observed that F. nucleatum modulates the cystine/glutamate antiporter in an OSCC cell line by increasing SLC7A11 expression, promoting L-glutamate efflux and favoring bacterial infection. Finally, our results showed that F. nucleatum and its metabolic derivates promote tumor spheroids growth, spheroids-derived cell detachment, epithelial-mesenchymal transition and Galectin-9 upregulation. Altogether, F. nucleatum promotes pro-tumoral mechanism in oral cancer.

Fig. 1

Computational proteomic analysis of host-microbe interactions in oral cancer. a Gene ontology functional enrichment analysis in biological process (GO:BP) and b Upstream regulator analysis in Ingenuity Pathways Analysis (IPA) of activated (red) and inhibited (blue) pathways identified in human malignant samples in comparison with healthy samples. c Venn diagram of peptides from healthy samples triplicates and the identification of unique healthy-samples proteins (pink circle) with the corresponding peptides (white circle) from specific bacteria. d Venn diagram of peptides from malignant samples triplicates and the identification of unique malignant-samples proteins (pink circle) with the corresponding peptides (white circle) from specific bacteria

Fig. 2

Distribution of proteins among bacterial species identified in OSCC secretome. Sankey diagram representing the protein distribution between bacterial species identified in oral cancer explants

Fig. 3

Metabolic pathways of L-Glutamate degradation are associated to F. nucleatum pepetides. a Bar chart of unsupervised and reproducible analysis of 25 metabolic pathways from MetaCyc containing 135 unique protein-encoding genes (in gray), from which the presence of protein-encoding genes derived from the oral malignant microbiome from F. nucleatum (in red) were present in metabolic pathways associated with L-Glutamate degradation, but not in L-Glutamate biosynthesis pathways. b Schematic representation of F. nucleatum enzymes found in the proteome of malignant samples (red) within the L-glutamate degradation pathway

Fig. 4

The L-Glutamate degradation pathway shows an association with bacterial peptides derived from F. nucleatum. a Scatter plots of quantitative value of SLC3A2 from the human proteomic dataset in Fig. 1. b Confocal image of the cystine/glutamate antiporter (System x_c^-) showing the expression and colocalization of SLC3A2 (green) and SLC7A11 (red) in HSC3 cells. c Scatter plots of extracellular L-glutamate from HSC3 cell cultures in the presence or absence of the Imidazole Ketone Erastin (IKE) and d in HSC3 cells uninfected or infected with F. nucleatum for 24 h. e Western blot of SLC7A11 in the HSC3 cells uninfected or infected with F. nucleatum for 24 h. f Scatter plots and representative images or dot plots of bacterial infection in the absence or presence of IKE by Incucyte and g Flow cytometry. h Schematic overview of F. nucleatum influencing the System xc- and L-Glutamate efflux. For all statistical analysis T test was used, ****P < 0.000 1, ***P < 0.001, **P < 0.01 and *P < 0.05 were considered significant

Fig. 5

Confocal images of Oral cancer cells after 6-hour infection with F. nucleatum. Confocal images of HSC3 cells infected with F. nucleatum (6 h). Bacteria was stained with CSFE (red) previous the infection. HSC3 were stained with Hoechst (blue), plasma membrane cell mask deep red (green) and propidium iodide (magenta) to confirm bacteria viability

Fig. 6

F. nucleatum infection drives tumoral growth and enhances the migratory behavior of OSCC cells. a Representative schema and scatter plot of cell counting of uninfected or F. nucleatum-infected HSC3 cells in monolayer (24 h). b Representative schema and point-&-connection line plot of tumourspheres area of uninfected or F. nucleatum-infected HSC3 cells at day 3, 6 and 10. c Representative schema and point-&-connection line plot of isolated cells from tumorsphere of uninfected or F. nucleatum-infected HSC3 cells at day 3, 6 and 10. d Representative schema and scatter plot of isolated HSC3 cells in supernatants from tumourspheres of uninfected or F. nucleatum-infected HSC3 cells after 6 days of infection. e Representative schema and scatter plot of isolated HSC3 cells in supernatants from de lower chamber of transwell from tumourspheres of uninfected or F. nucleatum-infected HSC3 cells after 6 days of infection. Data are presented as individual symbols with paired lines (Paired t test). For statistical analysis, Two way ANOVA and T test were used,****P < 0.000 1, ***P < 0.001, **P < 0.01 and *P < 0.05 were considered significant

Fig. 7

F. nucleatum promotes ETM markers expression in OSCC. a Symbols and line plots of E-cadherin and MMP9 mRNA expression uninfected or F. nucleatum-infected HSC3 cells in monolayer (6 h). Data are presented as individual symbols with paired lines (Paired t test). b Differential protein expression heatmap of EMT markers from lysate of uninfected or F. nucleatum-infected HSC3 cells in monolayer (24 h). c Differential protein expression heatmap of EMT markers from supernatants of uninfected or F. nucleatum-infected HSC3 cells (24 h). d, e Scatter plots of quantitative value of E-cadherin, MMP9, EGFR and Cathepsin S from the human proteomic dataset in Fig. 1. For statistical analysis, Sidaks multiple comparisons and T test were used, ****P < 0.000 1, ***P < 0.001, **P < 0.01 and *P < 0.05 were considered significant

Fig. 8

Butyrate intermediate concentrations encourage HSC3 tumorsphere growth. a Representative images by live cell microscopy using the IncuCyte system to determine HSC3 tumorsphere growth, after being challenged with different concentrations of butyrate. b Time plot of HSC3 tumorsphere size after butyrate treatment for 7 days. c HSC3 tumorsphere size after butyrate treatment at day 7. Data represent the mean ± SEM of at least 3 independent biological replicates. Data represent the mean ± SEM of at least 3 independent biological replicates. t test student, **P < 0.01; ***P < 0.001

Fig. 9

GAL-9 is induced on OSCC cells after F. nucleatum infection. Histograms and total percentage of cells expressing CD39, CD73, CD155, PDL-1 and GAL-9 from uninfected or F. nucleatum-infected HSC3 cells (24 h). For all statistical analysis, T test was used, ****P < 0.000 1, ***P < 0.001, **P < 0.01 and *P < 0.05 were considered significant

Fig. 10

Complement cascade proteins identified on OSCC secretome. a Plots, schematic representation, and differential protein expression heatmap of proteins from the complement cascade protein from the proteomic dataset in Fig. 1. b Scatter plots of soluble C4a, C3a and C5a in control and OSCC secretome. For all statistical analysis, T test was used, ****P < 0.000 1, ***P < 0.001,**P < 0.01 and *P < 0.05 were considered significant