Microbiome, metabolome, and transcriptome analyses in esophageal squamous cell carcinoma: insights into immune modulation by F. nucleatum

Graphical abstract

Dear Editor,

Esophageal cancer is the seventh most common malignant tumor in China and has the third highest fatality rate worldwide (Siegel et al., 2022). The most common type in China is esophageal squamous cell carcinoma (ESCC), which has shown a promising response to immune checkpoint inhibitors (Doki et al., 2022; Sun et al., 2021). However, challenges remain in effectively using immunotherapy due to varying patient responses and difficulties in identifying suitable candidates.

The esophagus serves as the vital link between the oral cavity and the stomach. Its unique anatomical position renders it susceptible to various microorganisms. This exposure significantly heightens the risk of intra-tumoral microbiota colonization in ESCC. Long-term dental caries and chronic periodontitis have been linked to immunotherapy resistance in ESCC. This resistance could be attributed to persistent streptococcal infections in the oral and pharyngeal regions (Gao et al., 2018). Furthermore, metabolites originating from gut microbiota play a critical role in mediating communication between microbiota and the host immune system (He et al., 2021; Mager et al., 2020; Wang et al., 2022). Microorganisms possess the ability to modulate the sensitivity of ESCC to immunotherapy (Mager et al., 2020; Matson et al., 2018). Recent research unequivocally demonstrated that the intra-tumoral microbiota plays a crucial role in influencing tumor responses to immunotherapy (Lo et al., 2021; Wu et al., 2023). However, the mechanistics are still not fully elucidated.

We conducted a comprehensive analysis of microbiome, metabolome, and transcriptome analyses of 52 ESCC patients, including 43 untreated and 9 treated patients (Fig. 1A; Supplementary Materials and Methods). Figure S1A displayed preoperative gastroscopy images and corresponding postoperative hematoxylin-eosin staining images, and 16S rRNA microbiota analysis of cohort 1 at the genus level revealed distinct differences between untreated ESCC and paired normal tissues, including Fusobacterium (Fig. S1B and S1C). LEfSe analysis identified p_Fusobacteria, c_Fusobacteriia, and o_Fusobacteriales as dominant germs in ESCC tissues, with untreated tissues showing higher levels of f_Fusobacteriaceae than treated ones (Figs. 1B and S1C). Higher levels of Fusobacterium were observed in the advanced stage (Fig. S1D and S1E) but decreased significantly in the treated cases (Fig. 1D), particularly in patients undergoing chemo-immunotherapy compared with chemotherapy alone (Fig. 1E). In our previous studies, oral mucosal and saliva samples were collected from ESCC patients undergoing immunotherapy (Zuo et al., 2023). Further analysis revealed that patients who achieved clinical benefit had lower levels of Fusobacterium in their oral samples compared with those who with no clinical benefit (Fig. S1F). Analyzing Fusobacterium alone reaffirmed its significant overexpression in ESCC tissue (Fig. 1F). Elevated Fusobacterium abundance correlated with worse TNM and N stages, exhibiting significant downregulation post-treatment (Fig. 1G–I).

ESCC tissues in cohort 2 were categorized into Fu-Low group and Fu-High group based on Fusobacterium abundance. There were four mRNAs not only upregulated in ESCC compared with para-cancer tissues but also in Fu-High compared with Fu-Low: C-Type Lectin Domain Family 12 Member A (CLEC12A), Glutamate Metabotropic Receptor 4 (GRM4), Protein Arginine Methyltransferase 8 (PRMT8), and Peptide YY 2 (PYY2) (Fig. 1J). A correlation analysis between Fusobacterium abundance and the target gene expression was conducted in ESCC samples (Fig. S2A). Gene Ontology analysis on differentially expressed genes (DEGs) based on Fusobacterium abundance revealed associations with inflammatory response pathways (Fig. S2B). Human multi-database association annotation analysis on Fu-High and Fu-Low DEGs provided insights into immune-related characteristics (Fig. S2C). Subsequently, we conducted an immune correlation analysis with data sourced from TIMER database. The corresponding clinical information of Asian ESCC patients was showed in Table S1. Results indicated a notable positive association between CLEC12A and macrophages (Fig. 1K), as well as the respective genetic markers (Fig. S2D and S2E). Immunofluorescence (IF) staining indicated that compared with para-cancer tissues, the presence of tumor-associated macrophages (TAMs) in ESCC tissues increased. Moreover, we observed an interesting phenomenon that the number of TAMs in treated ESCC tissue increased compared with untreated ESCC tissue (Fig. S2F). Besides, there were significant downregulation of M1-type macrophage-associated genes (IL-12A, CD68, CXCL12, IRF5) and upregulation of M2-type macrophage-associated genes (TGFB1, TGFB2, IL10, CCL18) in ESCC (Fig. S2G). The joint analysis revealed that chemokines associated with Fusobacterium abundance include CCR8, CXCL1, and CXCL8 (Fig. S2H). We hypothesized that Fusobacterium’s target genes are intricately associated with immune responses and may exhibit a specific relationship with TAMs.

Fusobacterium nucleatum was the represented species in Fusobacterium genus. To explore whether F. nucleatum plays a pivotal role as well as Fusobacterium, a quantitative real-time polymerase chain reaction (qRT-PCR) analysis was performed on 28 ESCC tissues of cohort 1. The results revealed that the expression pattern of F. nucleatum closely mirrored the abundance trend within the Fusobacterium genus (Fig. 1L). Besides, it demonstrated significantly higher expression levels of F. nucleatum in ESCC tissues compared with para-cancer tissues (Fig. 1M). We performed qRT-PCR on a gene chip containing 67 ESCC tissues and their corresponding 28 adjacent normal tissues. Figure S3A displayed the clinical and pathological characteristics. It was found that F. nucleatum infection is linked to worse TNM staging (Fig. S3B), and a poorer prognosis than non-F. nucleatum infection (log‐rank P = 0.086, Fig. 1N), which was consistent with previous research (Ding et al., 2023; Yamamura et al., 2016). However, the difference was not statistically significant, likely due to limited cases.

In order to elucidate the association between F. nucleatum and immune cells, fluorescent probes of F. nucleatum and IF staining of immune cells were performed on paraffin-embedded specimens. It was found that in areas with F. nucleatum infection, the number of TAMs was higher (Fig. 2A) and the number of CD8⁺ T cells is relatively low, while no significant change in CD4⁺ T cells (Fig. S3C). Database analysis also showed a significant positive correlation between CLEC12A and TAMs but no significant correlation with other immune cells (Fig. S3D). Dual IF results from ESCC paraffin tissue revealed differences in TAMs polarization based on F. nucleatum infection states (Figs. 2B and S3E). Despite the limited number of specimens, these results supported our hypothesis that F. nucleatum may exhibit a specific relationship with TAMs.

CLEC12A belongs to the myeloid C-type lectin receptor, which is a pattern recognition receptor for bacteria and is widely expressed on immune cells, including macrophages. Multiple IF staining reveals co-localization of CLEC12A and CD68 (Fig. 2C). Notably, F. nucleatum was present both inside and outside TAMs (Fig. 2A). Subsequent electron microscopy analysis of samples from co-cultured THP1 macrophages with F. nucleatum at a ratio of 1:100 revealed a limited presence of the bacterium within cells (Fig. S4A). This suggested that F. nucleatum may impact TAMs either directly or indirectly, with metabolites potentially playing a significant role in this interaction. Then, we conducted non-targeted metabolomics on cohort 3. There was a significant difference in metabolites between ESCC and normal tissues, encompassing untreated and treated patients (Fig. S4B). For untreated patients, 66 upregulated and 27 downregulated metabolites were observed in ESCC tissue, while for treated patients, 3 upregulated and 22 downregulated metabolites were identified, indicating treatment-associated changes (Fig. S4C). Kyoto Encyclopedia of Genes and Genomes enrichment analysis highlighted the involvement of numerous metabolites in linoleic acid metabolism which is a kind of free fatty acids (FFA) (Fig. S4D and S4E). Cluster analysis disclosed 12 Fusobacterium-associated metabolites, also including FFA (Fig. S4F). OPLS-DA analysis demonstrated eight upregulated and four downregulated metabolites in the Fu-High group (Fig. S4G). Integrating metabolite and microbiological data also identified FFA correlating with Fusobacterium (Fig. S4H), suggesting that Fusobacterium is crucial in FFA metabolism. Subsequently, a targeted FFA metabolomics analysis was performed on the cell supernatant of KYSE-30 and ECA109, which confirmed elevated levels of metabolites in F. nucleatum co-cultured groups (Table S2). Finally, 4 candidate metabolites were selected: Tridecanoic acid (TDA), Eicosapentaenoic acid (EPA), 7-Ketolithocholic acid (7-KLCA), and Phenyllactic acid (PLA).

Then, we co-cultured F. nucleatum with THP1 macrophages for wet experiments. There was a significant increase of CLEC12A after co-culturing with F. nucleatum (Fig. 2D). The activity of THP1 macrophages changed following the interventions of TDA, EPA, 7-KLCA, and PLA (Fig. S5A). Then, the intervention concentration was set at 10 μmol/L for EPA, and 500 nmol/L for 7-KLCA, PLA, and TDA. Specifically, the intervention of THP1 macrophages with PLA showed upregulation of CLEC12A, which was consistent with F. nucleatum (Fig. 2E). To investigate the interaction between PLA and the membrane protein CLEC12A, we obtained structural models of PLA and the CLEC12A protein from the PubChem and UniProt databases, respectively. Docking experiments were conducted using the Rosetta software. The target model exhibited stability in Rosetta scoring. Subsequently, the model was uploaded to the PLIP network server. The results revealed a well-defined region of strong hydrogen bonding interaction between protein side chain atoms (Num: 2350) and ligand atoms (Num: 4324). The experiment ultimately confirmed that PLA interacts with the 143rd asparagine residue of the A chain of the CLEC12A protein (Fig. 2F). Subsequently, we conducted an analysis of specific markers associated with M1 and M2 macrophages, utilizing LPS combined with IFN-γ as the positive control for M1 polarization and IL-4 as the positive control for M2 polarization in THP-1 macrophages. The results indicated that treatment with F. nucleatum or PLA significantly enhanced M1 polarization, as evidenced by a marked increase in the expression levels of the IL-6, TNF-α, NOS2, CXCL9, and CXCL10 genes. Meanwhile, both F. nucleatum and PLA interventions in THP-1 macrophages enhanced M2 polarization, as evidenced by a marked increase in the expression levels of the TGF-β1, IL-4, IL-10, CD206, CD163, and Arg1 genes, with PLA having a stronger effect. Notably, the M2 marker CD206 gene showed M2 anti-inflammatory activity (nearly 30-fold increase) significantly more than M1 pro-inflammatory activity (about 8-fold increase) (Fig. 2G). Flow cytometry results indicated a significant increase in M2-type THP-1 macrophages (both CD206 and CD163) after co-culturing with F. nucleatum, whereas no such change was detected when co-cultured with Escherichia coli (Figs. 2H and S5B). Furthermore, PLA, in contrast to the other three metabolites, also induced an upward trend in M2 macrophage levels (Fig. S5C).

Subsequently, we devised a rescue experiment to investigate the effects of F. nucleatum and CLEC12A on the polarization of THP1 macrophages. Our findings indicated that CLEC12A knockdown not only resulted in a reduction in the proportion of M2 macrophages but also partially mitigated the increase in M2 macrophage levels induced by F. nucleatum infection (Figs. 2I and S5D). The ELISA assay revealed that F. nucleatum infection lead a significant increase in IL4 and IL10, and knocking down CLEC12A significantly reduced IL4 and IL10 compared with the control group. Interestingly, THP1 macrophages with low CLEC12A expression did not demonstrate an elevation in IL4 following F. nucleatum infection. Furthermore, CLEC12A knockdown resulted in an enhanced secretion of pro-inflammatory cytokines IL-6 and TNF-α in THP1 macrophages (Fig. 2J). These findings indicated that CLEC12A potentially plays a pivotal role in the F. nucleatum-mediated inflammatory response of macrophages.

In conclusion, this study presented a novel integration of microbiome, metabolome, and transcriptome to investigate the metabolites and target genes associated with F. nucleatum. The findings suggested a correlation between F. nucleatum infection and immune characteristics in ESCC, with particular emphasis on TAMs. Fusobacterium nucleatum not only upregulated M1 markers in THP-1 macrophages but also significantly elevates M2-specific markers, with the most notable increase seen in the anti-inflammatory marker CD206. Fusobacterium nucleatum and its associated metabolite, PLA, upregulated the expression levels of CLEC12A in THP-1 macrophages, which is essential for the induction of the CD206 and CD163 phenotypes by F. nucleatum. We hypothesize that F. nucleatum may activate and polarize macrophages, affecting the tumor immune microenvironment. Further research is needed to understand these mechanisms and develop new immunotherapy strategies.

References

[1] Ding N, Cheng Y, Liu H et al. Fusobacterium nucleatum infection induces malignant proliferation of esophageal squamous cell carcinoma cell by putrescine production. Microbiol Spectr 2023;11:e0275922.

[2] Doki Y, Ajani JA, Kato K et al. Nivolumab combination therapy in advanced esophageal squamous-cell carcinoma. N Engl J Med 2022;386:449-462.

[3] Gao L, Xu T, Huang G et al. Oral microbiomes: more and more importance in oral cavity and whole body. Protein Cell 2018;9:488-500.

[4] He Y, Fu L, Li Y et al. Gut microbial metabolites facilitate anticancer therapy efficacy by modulating cytotoxic CD8(+) T cell immunity. Cell Metab 2021;33:988-1000.e7.

[5] Lo CH, Kwon S, Wang L et al. Periodontal disease, tooth loss, and risk of oesophageal and gastric adenocarcinoma: a prospective study. Gut 2021;70:620-621.

[6] Mager LF, Burkhard R, Pett N et al. Microbiome-derived inosine modulates response to checkpoint inhibitor immunotherapy. Science 2020;369:1481-1489.

[7] Matson V, Fessler J, Bao R et al. The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients. Science 2018;359:104-108.

[8] Siegel RL, Miller KD, Fuchs HE et al. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33.

[9] Sun JM, Shen L, Shah MA et al. KEYNOTE-590 Investigators et al. Pembrolizumab plus chemotherapy versus chemotherapy alone for first-line treatment of advanced oesophageal cancer (KEYNOTE-590): a randomised, placebo-controlled, phase 3 study. Lancet 2021;398:759-771.

[10] Wang H, Rong X, Zhao G et al. The microbial metabolite trimethylamine N-oxide promotes antitumor immunity in triple-negative breast cancer. Cell Metab 2022;34:581-594.e8.

[11] Wu H, Leng X, Liu Q et al. Intratumoral microbiota composition regulates chemoimmunotherapy response in esophageal squamous cell carcinoma. Cancer Res 2023;83:3131-3144.

[12] Yamamura K, Baba Y, Nakagawa S et al. Human microbiome fusobacterium nucleatum in esophageal cancer tissue is associated with prognosis. Clin Cancer Res 2016;22:5574-5581.

[13] Zuo J, Liu Y, Lv W et al. The host microbiota is associated with the occurrence and development of esophageal squamous cell carcinoma. Protein Cell 2023;14:294-298.

RIGHTS & PERMISSIONS

The Author(s) 2024. Published by Oxford University Press on behalf of Higher Education Press.