Gastric cancer-associated bacteria exhibit distinct butyrate and pyruvate metabolism: a metabolomic analysis of F. nucleatum, N. subflava, and H. pylori

Abstract

Background

Butyrate and pyruvate are central metabolites in anaerobic microbial metabolism, with key roles in gastrointestinal physiology. We recently identified Fusobacterium nucleatum (F. nucleatum) and Neisseria subflava (N. subflava) as non-Helicobacter pylori (H. pylori) bacteria potentially involved in gastric cancer development. However, the metabolic pathways distinguishing F. nucleatum, N. subflava, and H. pylori remain poorly characterized.

Methods

We performed capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS) to profile the metabolic outputs of these three species under isolated culture conditions.

Results

Distinct metabolic signatures were observed: F. nucleatum predominantly synthesized butyrate via the acetyl-CoA pathway, whereas N. subflava produced high levels of pyruvate and employed a cyclical route regenerating pyruvate from acetyl-CoA. In contrast, H. pylori lacked significant production of either metabolite.

Conclusions

This study delineates species-specific metabolic programs among gastric cancer-associated bacteria and highlights unique butyrate and pyruvate metabolism as a potential axis of microbe–host and microbe–microbe interaction in the gastric environment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-025-04466-7.

Background

Butyrate and pyruvate are central metabolites in anaerobic microbial metabolism, playing essential roles in both health-related and environmental systems. Butyrate, a short-chain fatty acid (SCFA), is primarily produced through microbial fermentation of organic substrates in gut and environmental microbiomes. It is critically important due to its roles in maintaining colonic epithelial cell integrity, modulating inflammatory responses, and serving as a key energy source for host tissues [1]. Pyruvate, a pivotal metabolic intermediate, bridges glycolysis with various fermentative pathways, including those leading to the synthesis of butyrate and other SCFAs [2].

Accumulating evidence suggests that butyrate also exerts antitumor effects. Several studies have demonstrated that butyrate modulates immune responses, regulates cell proliferation, and induces apoptosis in tumor cells, thereby contributing to tumor suppression [3–6]. These findings highlight the clinical importance of butyrate beyond its metabolic functions. Microbial interactions play a crucial role in butyrate biosynthesis. Within complex microbial communities, non-butyrogenic microorganisms supply essential precursor metabolites such as acetate, lactate, and succinate, which specialized butyrogenic bacteria subsequently convert into butyrate. Members of bacterial families such as Clostridiaceae, Lactobacillaceae, and Prevotellaceae are well-recognized butyrate producers. Comprehensive metabolomic investigations are needed to clarify these metabolic interdependencies and their implications for health and disease [7].

Fusobacterium nucleatum (F. nucleatum) and Neisseria subflava (N. subflava) are notable bacterial species in human-associated microbiomes. F. nucleatum, an anaerobic bacterium prevalent in the oral cavity and gastrointestinal tract, is implicated in periodontal disease [3], colorectal cancer [4], and notably gastric cancer following Helicobacter pylori (H. pylori) eradication [5]. Fusobacterium species can activate pro-oncogenic pathways, induce inflammation, and promote DNA damage, thus facilitating gastric carcinogenesis. Similarly, N. subflava, a facultative anaerobe associated with gastric inflammation, is believed to influence metabolic networks that support butyrate production, potentially impacting gastric cancer risk [5].

Metabolomics analysis represents a powerful approach for exploring butyrate and pyruvate metabolism within these bacteria [6]. Capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS) enables precise profiling of ionic metabolites such as acetate, lactate, and succinate, pivotal precursors in butyrate biosynthesis. These advanced metabolomic techniques allows detailed mapping of metabolic routes and identification of key intermediates facilitating butyrate production in F. nucleatum and N. subflava. Such approaches provide valuable high-resolution insights into microbial biochemical networks and interactions underlying butyrate metabolism, enhancing our understanding of microbial cooperation and its implications for gastric carcinogenesis.

In this study, we aimed to elucidate butyrate and pyruvate metabolism in F. nucleatum, N. subflava, and H. pylori using metabolomics analysis. Our specific objectives were to: (i) identify the key metabolic pathways contributing to butyrate and pyruvate production, (ii) determine the metabolic interactions between these species that facilitate butyrate biosynthesis, and (iii) utilize CE-TOFMS to comprehensively profile relevant metabolic intermediates and end-products. This research expands our understanding of microbial interactions in butyrate metabolism and highlights its broader significance for host health, particularly regarding the risk of gastric cancer development post-H. pylori eradication.

Methods

Bacterial strains

The Helicobacter strain used in the present study was H. pylori TN2, which potentiates active CagPAI and type IV secretion systems. H. pylori cells were cultured in Brucella broth culture medium (Becton Dickinson) containing 2.5% FBS (Cansera International, Inc.) in a microaerophilic atmosphere generated by CampyPak-Plus (Becton Dickinson) at 37 °C for 24 h [8]. The following bacterial strains were obtained from American Type Culture Collection: F. nucleatum 25,586 (which contains the fap2 and fadA genes [9, 10]) and N. subflava 49,275. F. nucleatum was cultured in Columbia broth under anaerobic conditions. N. subflava were cultured in trypticase soy medium with defibrinated sheep blood (BBL). Prior to use, these bacterial strains were washed with phosphate-buffered saline (PBS), and their concentrations were estimated considering an optical density of 0.1 at 560 nm being equivalent to 4 × 10⁷ colony-forming units of each bacterial strain. To minimize the loss of volatile metabolites including SCFAs, culture tubes were tightly sealed, headspace was minimized, and all handling was performed at pH 7.0 under cold conditions.

Metabolite extraction

Microbial culture was cooled on ice to quench enzymatic activity, and then centrifuged at 5,800 ×g, 4 °C for 5 min to pellet the cells. Upon removing the supernatant, the cell pellet was washed twice with 10 mL of Milli-Q water, to remove residual medium components while minimizing loss of intracellular metabolites; this washing step was applied uniformly across all bacterial groups to maintain comparability. The pellet was then treated with 1,600 µL of methanol, followed by ultrasonication for 30 s to dissolve the pellet. Next, 1,100 µL of Milli-Q water containing internal standards (H3304-1002, Human Metabolomics Technologies, Inc. (HMT), Tsuruoka, Yamagata, Japan) was added to the cell extract, which was left at room temperature for another 30 s. The mixture was then cooled down on ice and centrifuged at 2,300 ×g, 4 °C for 5 min. Subsequently, 700 µL of the supernatant was centrifugally filtered through a 5-kDa cutoff filter (UltrafreeMC-PLHCC, HMT) at 9,100 ×g, 4 °C to remove macromolecules. To avoid evaporative loss of volatile metabolites including SCFAs, no vacuum concentration or nitrogen blow-down was applied prior to CE-TOFMS. The filtrate was evaporated to dryness under vacuum and reconstituted in 50 µL of Milli-Q water for metabolomics analysis at HMT.

CE-TOFMS

Metabolomics analysis was conducted according to HMT’s Dual Scan package, using CE-TOFMS based on the methods described previously [11, 12]. Briefly, CE-TOFMS analysis was carried out using an Agilent CE capillary electrophoresis system equipped with an Agilent 6230 time-of-flight mass spectrometer (Agilent Technologies, Inc., Santa Clara, CA, USA). The systems were controlled by Agilent MassHunter Workstation Data Acquisition (Agilent Technologies).

The spectrometer was scanned from m/z 50 to 1,000 for CE-TOFMS analyses. Peaks were extracted using MasterHands, automatic integration software (Keio University, Tsuruoka, Yamagata, Japan) in order to obtain peak information including m/z, peak area, migration time (MT) for CE-TOFMS and retention time (RT) [13]. Signal peaks corresponding to isotopomers, adduct ions, and other product ions of known metabolites were excluded, and the remaining peaks were annotated according to HMT’s metabolite database based on their m/z values and MTs or RTs. Areas of the annotated peaks were then normalized to internal standards and sample amount in order to obtain relative levels of each metabolite. Primary 110 metabolites were absolutely quantified based on one-point calibrations using their respective standard compounds. Hierarchical cluster analysis (HCA) and principal component analysis (PCA) [14] were performed by HMT’s proprietary MATLAB and R programs, respectively. Detected metabolites were plotted on metabolic pathway maps using VANTED software [15]. For heatmap visualization (Fig. 2B), metabolite levels were further standardized by z-score transformation (mean-centered and scaled to unit variance across samples) to emphasize relative patterns rather than absolute concentrations.

Fig. 2.

Multivariate analysis of metabolite profiles. (A) PCA distinguishes metabolic profiles of H. pylori, F. nucleatum, and N. subflava. (B) Heatmap clustering highlights group-specific metabolite signatures

Statistical analysis

For the quantitative analyses of metabolites, Welch’s t tests were used to evaluate the association between the common identified metabolites in F. nucleatum and N. subflava referred as H. pylori group. Statistical significance was defined at p < 0.05. To visualize these results of differential metabolites analyses, a heatmap was generated using the GraphPad prism (version 9), based on metabolites with log₂ fold change ≥ 1. The H. pylori group was used as the reference, and all metabolite values in this group were set to a ratio of 1. When converted to log₂ scale, these reference values become log₂(1) = 0, representing the baseline (no difference). Thus, positive and negative log₂ fold change values indicate relative increases or decreases compared with H. pylori. For pathway analysis, we plotted identified butyrate and pyruvate associated metabolites in F. nuclatum and N. subflava group using KEGG map.

Results

CE-TOFMS identified the distinct metabolites of each bacterium

The base peak electropherograms of cation and anion metabolites for the F. nucleatum, N. subflava, and H. pylori groups are shown in Fig. 1. Each electropherogram had similar profiles across the bacterial samples. PCA revealed distinct clustering, clearly differentiating the F. nucleatum, N. subflava, and H. pylori groups from each other (Fig. 2A). Additionally, a heatmap clustering further confirmed unbiased grouping based on metabolite expression levels among these bacterial groups.

Fig. 1.

Base peak electropherograms of metabolites in three bacterial groups. (A–C) Cation mode for H. pylori, F. nucleatum, and N. subflava. (D–F) Anion mode for the same groups. Each panel shows three biological replicates

Quantitative analysis identified specific metabolites of F. nuclatum and N. subflava

Quantitative analysis identified a total of 110 metabolites (56 anion and 54 cation metabolites) across the three bacterial groups. Among these, 14 metabolites were specific to individual bacterial groups, as shown in Fig. 3A–C. The F. nucleatum group specifically showed high levels of hydroxybutyric acids (2-hydroxybutyric acid and 3-hydroxybutyric acid) and glyceraldehyde 3-phosphate, while the N. subflava group was characterized by high levels of pyruvic acid, divalent glutathione, 2-oxoisovaleric acid, divalent CoA, and glutathione. In contrast, no specific metabolites were uniquely identified in the H. pylori group.

Fig. 3.

Group-specific metabolites identified by quantitative analysis. (A) No unique metabolites in H. pylori. (B) Unique metabolites in F. nucleatum. (C) Unique metabolites in N. subflava

Furthermore, 46 common metabolites (11 anion and 35 cation metabolites) were identified in all three bacterial groups. Of the 35 cation metabolites, 26—including Met, Cytosine, Thymidine, Citrulline, Ornithine, Choline, Arg, Asp, Adenosine, Glu, S-Adenosylmethionine, Leu, Pro, Ile, Uracil, Gly, Phe, Cytidine, GABA, Adenine, Guanine, Tyr, Lys, Val, Guanosine, and Ala—showed significantly higher levels in the F. nucleatum group compared to the H. pylori group (Supplementary Table 1). Similarly, 32 metabolites—including Adenine, Adenosine, Ala, Arg, Asn, Asp, Choline, Citrulline, Cytidine, Cytosine, GABA, Glu, Gly, Guanine, Guanosine, His, Hypoxanthine, Ile, Inosine, Leu, Lys, Met, Ornithine, Phe, Pro, S-Adenosylmethionine, Ser, Thr, Trp, Tyr, Uracil, and Val—were significantly elevated in the N. subflava group compared to the H. pylori group (Supplementary Table 1). Heatmap visualization highlighted that F. nucleatum showed broad elevations in amino acids and nucleotide-related metabolites, whereas N. subflava exhibited even stronger and more widespread increases, particularly in amino acids such as serine, glycine, threonine, and aromatic amino acids, suggesting distinct but partially overlapping cationic metabolite enrichment patterns in the two species (Fig. 4A).

Fig. 4.

Heatmap of significantly altered metabolites. Heatmap shows metabolites with significant differences among bacterial groups. Rows represent metabolites, which were arranged according to metabolic pathways (glycolysis/pyruvate metabolism, TCA cycle/urea cycle, branched-chain amino acid metabolism, aromatic amino acid metabolism, sulfur/methyl cycle, energy-related metabolites, and nucleotide metabolism) and columns represent bacterial groups. Colors indicate log₂ fold changes relative to H. pylori. The H. pylori group was used as the reference (ratio = 1), and therefore all its metabolite values were converted to log₂(1) = 0, displayed as the baseline (no change)

In the analysis of 11 common anion metabolites, 10 metabolites (ADP, AMP, citric acid, GDP, glycerol 3-phosphate, lactic acid, NADP+, ribulose 5-phosphate, succinic acid, and UMP) were significantly higher in the F. nucleatum group compared to H. pylori (Supplementary Table 1). Notably, all 11 anion metabolites exhibited significantly higher concentrations in the N. subflava group compared to H. pylori (Supplementary Table 1). Heatmap visualization further indicated that F. nucleatum was characterized by elevated glycolytic and nucleotide-related metabolites, whereas N. subflava showed enrichment of TCA cycle intermediates such as citric acid, succinic acid, and malic acid, suggesting distinct utilization of central carbon metabolism between the two species (Fig. 4B).

Associations between the butyrate metabolism of F. nucleatum and the pyruvate metabolism of N. subflava

Quantitative and pathway analyses of butyrate and pyruvate metabolism revealed clear differences between the F. nucleatum and N. subflava groups. In F. nucleatum, the acetyl-CoA pathway initiated from pyruvate actively proceeded toward acetoacetyl-CoA, subsequently feeding into the butyrate metabolic pathway. This was supported by the enriched presence of metabolites associated with butyrate metabolism, indicative of enhanced butyrate synthesis through acetyl-CoA intermediates (Fig. 5). Conversely, the N. subflava group showed a distinct metabolic pattern in which the acetyl-CoA pathway functioned cyclically. Acetyl-CoA was metabolized via succinate and fumarate intermediates, ultimately cycling back to pyruvate. This cyclic activity suggests a continuous metabolic loop in N. subflava, potentially supporting sustained pyruvate regeneration and enhancing metabolic activity (Fig. 5).

Fig. 5.

Butyrate and pyruvate metabolism across bacterial groups. Quantitative comparison of metabolites related to butyrate (F. nucleatum) and pyruvate (N. subflava) pathways among the three groups

Discussion

Based on our study findings, CE-TOFMS analysis for F. nucleatum, N. subflava, and H. pylori identified the profiles of cationic and anionic metabolites produced by each bacterial species. The results demonstrated distinct metabolic profiles, with butyrate specifically detected in the F. nucleatum group and pyruvate prominently observed in the N. subflava group. In addition, our findings suggest potential interactions between butyrate and pyruvate within their metabolic pathways, highlighting a previously undetected connection.

Previous research, using methodologies such as 16 S ribosome RNA sequencing and genomic analyses, has reported interactions between the metabolic pathways of butyrate and pyruvate [2]. Our study provides novel evidence through direct measurement of metabolites, reinforcing the possibility of these interactions. Enzymes involved in these metabolic pathways—such as 4-Hydroxybutyryl-CoA dehydratase, butyrate kinase, acetate kinase, fumarate reductase, succinyl-CoA synthetase, and lactate dehydrogenase—have traditionally been associated with bacterial families like Ruminococcaceae, Clostridiaceae, Prevotellaceae, and Lachnospiraceae [2]. However, our results indicate that F. nucleatum and N. subflava may also participate in these metabolic pathways, extending our understanding of microbial metabolism and interactions.

These findings are clinically relevant because F. nucleatum and N. subflava have been implicated in promoting inflammation, carcinogenesis, and modulating immune responses [16]. Their characterization may provide further insights into bacterial competition, symbiosis, and the roles of microbial metabolites within the tumor microenvironment. Importantly, their behavior contrasts with that of H. pylori. Whereas H. pylori typically establishes a persistent gastric infection once acquired, F. nucleatum and N. subflava are oral commensals that may be continuously supplied from the oral cavity. Following H. pylori eradication, these bacteria may increase in prevalence, reflecting differences in adaptation to acidic gastric conditions. These distinctions underscore the importance of understanding microbial shifts in the stomach and their implications for gastric carcinogenesis.

Butyrate is a well-studied short-chain fatty acid with potential roles in host–microbe interactions. Although antitumor effects of butyrate have been reported, our study was not designed to evaluate these effects directly. Moreover, metabolite production in isolated bacterial cultures may differ from that in complex tumor microenvironments. Therefore, the functional implications of butyrate observed here should be interpreted with caution and warrant further investigation.

The major strength of this study is the direct measurement of multiple samples of F. nucleatum, N. subflava, and H. pylori using CE-TOFMS, providing comprehensive metabolomic profiles under isolated bacterial culture conditions. These metabolomic profiles provide valuable benchmarks for future research. A limitation of our study is that metabolomics analyses under co-culture conditions or in human pathological samples were not performed, potentially missing metabolites that might directly correlate with disease pathogenesis. In addition, the bacteria were cultured under neutral conditions (pH 7.0), which differs from the acidic gastric environment in vivo and may influence metabolic outputs. This difference should be considered when interpreting our results. Moreover, co-culture experiments of these bacteria were not conducted in this study. Future studies incorporating co-culture and in vivo models will be essential to clarify the interactions among bacterial species and their clinical significance. Finally, it should be noted that some metabolites may have been below the detection limit of CE-TOFMS (approximately several µM) rather than truly absent, and this technical limitation may have influenced the observed metabolite　profiles.

In this study, we clarified species-specific metabolite profiles using H. pylori as a reference. While our workflow was designed to minimize volatilization (neutral pH handling, rapid quenching on ice, sealed low-headspace processing, and no evaporative concentration), a minor underestimation of highly volatile SCFAs cannot be entirely excluded. Further investigations should extend these findings by applying CE-TOFMS to a broader range of samples and by performing co-culture experiments or analyses of combined culture supernatants. Such approaches will be essential to validate and expand the interspecies metabolic interactions suggested in our current work.

Although previous studies have shown that F. nucleatum and N. subflava are notable bacterial species in human-associated microbial diseases, including gastrointestinal cancers, the understanding of their metabolites involved in carcinogenesis remains limited. Our CE-TOFMS analysis of F. nucleatum, N. subflava, and H. pylori identified distinct profiles of cationic and anionic metabolites produced by each bacterial species, providing a foundational index of metabolites. Notably, butyrate was specifically detected in the F. nucleatum group, while pyruvate was prominently observed in the N. subflava group. Furthermore, our findings suggest potential interactions between butyrate and pyruvate within their individual metabolic pathways, highlighting a previously undetected connection. In conclusion, these results highlight the potential significance of the metabolic association between F. nucleatum and N. subflava in relation to bacteria-associated gastrointestinal oncogenesis.

Abbreviations

F. nucleatum

Fusobacterium nucleatum

N. subflava

Neisseria subflava

H. pylori

Helicobacter pylori

CE-TOFMS

Capillary electrophoresis time-of-flight mass spectrometry

PCA

Principal component analysis

SCFA

Short-chain fatty acids

HCA

Hierarchical cluster analysis

Authors’ contributions

R. Niikura contributed to data analysis and manuscript preparation, and Y. Hayakawa contributed to experiments. R. Niikura and Y. Hayakawa conducted the overall study design. T. Mukai assisted with experiments. Y. Kato and N. Yoshida supervised data analyses, and M. Fujishiro provided critical advice as experts and contributed to writing and editing the manuscript. All authors read and approved of the final manuscript.

Funding

This work was supported by the NIPRO funding.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to restrictions imposed by the sponsoring company but are available from the corresponding author on reasonable request and with permission from the sponsor.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Research Ethics Guidelines of the University of Tokyo. This study did not involve human participants, human tissue, animals, or genetically modified organisms, including genetically modified cells or animals. Therefore, no specific ethical approval was required according to the policy of the Institutional Review Board of the University of Tokyo.

Consent for publication

Not applicable.

Competing interests

R. Niikura and Y. Hayakawa received the NIPRO funding and T. Mukai, Y. Kato, and N. Yoshida are employees of Nipro Corporation.