Exopolysaccharides of Streptococcus mutans Enhance Oral Colonisation and Drug Tolerance in Helicobacter pylori

Abstract

Introduction and aims

The oral cavity is an extragastric reservoir for Helicobacter pylori (H. pylori), preventing its gastric eradication and causing recurrence. However, the oral environment is not ideal for H. pylori, and the factors facilitating oral H. pylori survival and colonisation are unclear. We explored how exopolysaccharides (EPSs), the fundamental building blocks of Streptococcus mutans (S. mutans) biofilms, affect H. pylori colonisation and drug tolerance in the oral cavity.

Methods

The EPS contents of 3 S. mutans biofilms (UA159, UA159ΔgtfB, and UA159ΔgtfBC) were determined via confocal laser scanning microscopy (CLSM). A coculture system of the H. pylori Sydney strain (SS1) and the 3 biofilms was established. The bacterial morphology and adhesive forces of the coculture systems were detected by scanning electron microscopy (SEM) and atomic force microscopy (AFM), respectively. CLSM observations confirmed the ability of glucosyltransferases (Gtfs) from S. mutans to form EPSs around SS1. The tolerance of SS1 to 2 representative antibiotics in the presence of EPSs was confirmed via colony-forming unit (CFU) counting and in vivo experiments.

Results

As the EPS content of the S. mutans biofilms increased, both the adhesive forces between SS1 and the S. mutans biofilms and the number of colonised SS1 cells increased. We extracted active Gtfs from UA159 and confirmed that Gtfs synthesised EPSs around SS1. The EPSs gradually accumulated over time, eventually encapsulating SS1. Both in vitro and in vivo experiments confirmed that S. mutans EPSs help protect SS1 against amoxicillin or clarithromycin, potentially confounding gastric lesion outcomes.

Conclusions

S. mutans biofilms mediate the adhesion, colonisation, and antibiotic tolerance of oral H. pylori via Gtf-driven EPS biosynthesis.

Introduction

The oral cavity is an extragastric reservoir for Helicobacter pylori (H. pylori).¹ Given the similarity and homology between oral and gastric H. pylori,^2,3 the oral cavity is recognised as a potential source of gastric reinfection following eradication therapy.⁴ Compared with oral H. pylori-negative patients, oral H. pylori-positive patients had significantly lower gastric eradication success at 4 weeks post-treatment (52.2% vs 91.6%, P = .0028).⁵ Furthermore, gastric infections are more severe when oral H. pylori is present; case–control research with 567 patients revealed that oral H. pylori was associated with an increased incidence of duodenitis, esophageal sphincter relaxation, and grade II gastroesophageal reflux.⁶ Therefore, oral H. pylori is a risk factor for suboptimal eradication and gastric reinfection.^7,8

H. pylori is disadvantaged in the oral cavity because of its low propensity for oral biofilm formation and microaerophilic metabolism.^9,10 However, the detection rate of H. pylori in dental plaque biofilms—with exopolysaccharides (EPSs) primarily comprising the matrix structure—is relatively high and far exceeds that in saliva.^11,12 A cross-sectional study of 443 Chinese dyspeptic patients revealed 59.4% (263/443) H. pylori positivity in first molar plaques, which is comparable to the 61.6% (273/443) gastric detection rate.¹³ Another study of 841 Chinese adults revealed a significantly greater dental plaque index in 574 gastric H. pylori-positive individuals than in H. pylori-negative individuals (P < .05).¹⁴ Collectively, these studies indicate that oral biofilms with EPSs as the main matrix support the persistent presence of H. pylori in the oral cavity.

Streptococcus mutans (S. mutans), a major producer of the oral biofilm matrix,¹⁵ synthesises glucosyltransferases (Gtfs) to catalyse the conversion of sucrose to EPSs.^16,17 Within the Gtf family, GtfB primarily generates water-insoluble α-1,3-linked glucans, whereas GtfC produces both glucan types (predominantly insoluble).¹⁸ Soluble glucans mediate the initial adhesion of oral bacteria, whereas insoluble glucans form the 3-dimensional scaffold of the biofilm, supporting microbial colonisation and cohesion.^19,20 The robust EPS-synthesising capacity of S. mutans benefits the resident bacteria within its biofilm through multiple mechanisms: EPSs can promote microbial aggregation,²¹ forming a diffusion-limiting barrier that protects encapsulated bacteria from damage,^15,22 resisting mechanical clearance (eg, chewing and saliva flushing),²³ and impeding antibiotic penetration.²⁴ EPSs can also maintain a heterogeneous microenvironment within biofilms (such as local hypoxia and nutrient gradients) and induce the formation of persister cells, thereby reducing bacterial sensitivity to antibiotics.25, 26, 27, 28 Furthermore, EPSs can mask bacterial surface antigens, helping them evade immune recognition and phagocytosis.²⁹ Despite these research advances regarding the protection and support provided by S. mutans-secreted EPSs to bacteria in oral biofilms, whether and through what mechanisms S. mutans-derived EPSs can facilitate the oral colonisation of H. pylori remain unclear.

In this study, we explored how variations in EPS content affect H. pylori adhesion, colonisation, and drug tolerance by comparing the biofilms formed by UA159 wild-type and UA159 gtfB/gtfC-deficient strains. Our findings clarify the synergistic pathogenic mechanism in dual-species biofilms: S. mutans biofilms promote the establishment of an H. pylori reservoir in the oral cavity, facilitating its gastrointestinal translocation. These findings validate the link between oral and systemic health.

Materials and methods

Bacterial strains and culture conditions

S. mutans UA159 and its mutants (ΔgtfB and ΔgtfBC) were obtained from the State Key Laboratory of Oral Diseases, Sichuan University. S. mutans was cultured anaerobically (10% H₂, 10% CO₂, and 80% N₂) at 37°C in brain heart infusion (BHI; OXOID) broth. H. pylori SS1 was grown in Brucella broth (BD, Sparks) supplemented with 10% fetal bovine serum (FBS) under microaerophilic conditions (5% O₂, 75% N₂, and 15% CO₂) at 37°C.

Dual-species biofilm formation

S. mutans UA159, UA159ΔgtfB, and UA159ΔgtfBC (OD₆₀₀_nm = 0.1) were diluted 1:100 in 1% sucrose-BHI and then randomly allocated to a 12-well plate. Each strain was set up with 3 replicate wells, labeled only with random codes (eg, “A1–A3,” “B1–B3,” and “C1–C3”). After 6 hours of culture to form biofilms, the S. mutans biofilms were rinsed twice with phosphate-buffered saline (PBS). H. pylori SS1 (OD₆₀₀_nm = 0.5) was diluted 1:20 in BHI supplemented with 10% FBS—which was used as the common culture medium for the dual-species system—and inoculated into a prelabeled 12-well plate containing S. mutans biofilms, initiating coculture of the 2 species in fresh 10% FBS–BHI medium. For subsequent assays, only the random codes were retained, and analysts remained blinded to the original group assignments until data collection was completed.

Colony-forming unit (CFU) counting

H. pylori-selective Columbia blood agar plates containing 9.75 g of Columbia blood agar base (OXOID) in 250 mL of ddH₂O were autoclaved and cooled to 30 to 45°C. Subsequently, 18.75 mL of defibrinated sheep blood with H. pylori-selective Dent antibiotics (Thermo Fisher) was added to each plate, and the plates were stored at 4°C. Mitis–salivarius–bacitracin (MSB) agar plates were used to isolate streptococci: 15 g of mitis–salivarius agar (Topbio, Shandong, China) was added to 200 mL of ddH₂O, autoclaved, and cooled to 45 to 50°C. Subsequently, 20 mg of 1% potassium tellurite and 40 U of mitis bacitracin (Topbio, Shandong, China) were added to each plate, and the plates were stored at 4°C. Dual-species biofilms resuspended in PBS were serially diluted. For the same sample at the same dilution, 3 parallel plates were prepared, spread-plated, and cultivated; the average number of colonies counted was calculated after 48 hours.

Measurement of the adhesive force

Six-hour biofilms of UA159, UA159ΔgtfB, and UA159ΔgtfBC were rinsed twice. The 3 types of biofilm samples were randomly shuffled and then labelled with numbers. A tipless cantilever (TL-CONT; NanoSensors) was dipped in poly-D-lysine (Solarbio) for 3 min, air-dried, and then immersed in a 10⁵ CFU/mL H. pylori SS1 suspension for 1 min to immobilise a layer of H. pylori on the tip. These freshly prepared bacterial-modified atomic force microscopy (AFM) probes were used to measure the adhesive force between H. pylori and S. mutans biofilms using a SHIMADZU SPM-9600 system (Shimadzu Corp., Kyoto, Japan). After data recording was completed, the sample information was decoded according to the labels.

Scanning electron microscopy (SEM) and environmental scanning electron microscopy (ESEM)

Dual-species biofilms and rat molars were fixed with glutaraldehyde and serially dehydrated in ethanol (50%, 70%, 80%, 90%, and 100%). The dual-species biofilms were observed via SEM (Hitachi Regulus8100), while the rat molars were examined using ESEM (Quattro ESEM, Thermo Fisher Scientific).

Biofilm imaging by confocal laser scanning microscopy (CLSM)

EPSs were stained with Alexa Fluor 647-labeled dextran conjugate (Invitrogen), live bacterial cells were stained with SYTO9 (Invitrogen), and H. pylori was stained using a 2-step immunofluorescence protocol: first, the bacteria were incubated with a Helicobacter pylori monoclonal antibody (Invitrogen), followed by a goat anti-mouse IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor Plus 405 (Invitrogen). Emission filters were set at 655 to 690 nm for Alexa Fluor 647, 495 to 515 nm for SYTO9, and 400 to 420 nm for Alexa Fluor Plus 405. Confocal microscopy (Olympus) was performed.

Gtf extraction

UA159 cultures were centrifuged (13,000 rpm, 4°C, 10 min) to collect the supernatants. A third volume of absolute ethanol was added, and the mixtures were incubated at -80°C for 30 min. After thawing, centrifugation (13,000 rpm, 4°C, 20 min) yielded precipitates, which were resuspended in 200 μL of PBS and purified in 100 kDa ultrafiltration centrifuge tubes (Merck Millipore, Billerica, MA, USA) to obtain Gtf suspensions.

Coomassie blue staining

Isolated Gtfs were mixed with 5X protein loading dye (Sangon Biotech, Shanghai, China), heated at 95°C for 5 min, and separated by SDS–PAGE. After electrophoresis, the gel was stained with Coomassie blue R-250 (Sigma, St. Louis, MO, USA) and destained until the background was clear.

Gtf activity assay

Gtf suspensions were quantified using a BCA kit (Epizyme Biomedical, Shanghai, China) and normalised to 1 μg/μL. Aliquots (0, 5, 10, 15, 20, 25, and 30 μL) were added to a 96-well plate and then mixed with 0.2 M sodium phosphate buffer containing 2.5% sucrose and 0.2% T70 dextran (Solarbio, Beijing, China) to a total volume of 200 μL. After 50 μL of mineral oil was added, the reactions were incubated at 37°C, and the OD_540nm was monitored continuously for 24 h to assess the Gtf activity.

Animal experiments

Three-week-old pathogen-free Sprague–Dawley rats of both sexes (Beijing HFK Bioscience Co., Ltd.) were randomly divided into 8 groups using a random number table method. Specifically, all the rats were first labeled with unique integer numbers (1–64). A total of 64 nonrepetitive random integers ranging from 1 to 64 were generated using the RANDBETWEEN function in Excel to form a random sequence. The rats with original numbers of 1 to 64 were then reranked according to the order of this random sequence and sequentially allocated to the 8 experimental groups (n = 8) on the basis of their reranked order. A blinded protocol was implemented throughout the experiment: animal caretakers and researchers responsible for drug administration and modeling operated on the basis of only the ear tag numbers (1–64) and had no access to the correspondence between numbers and groups. Following 3 days of ampicillin feeding and 2 days of distilled water administration, artificial caries-like lesions were created in the mandibular molars. Groups received daily injections of PBS, 10⁸/mL UA159, UA159ΔgtfBC (once/day) or 1 μg/μL Gtfs (twice/day) for 5 days, with only 5% sucrose water available for drinking. All the groups subsequently received 10⁸/mL SS1 in sterile water daily for 10 days. After incubation, 0.6 μg/mL amoxicillin or 0.15 μg/mL clarithromycin was injected into the artificial cavities for 5 days. Oral swabs from 8 rats in each group were used for CFU analysis. The rats were euthanised via CO₂ inhalation, and their mandibles and stomachs were harvested. For subsequent analyses, the collected tissues were labeled only with the numbers of the rats, and personnel performing pathological assays were blinded to group allocations.

Warthin–Starry staining

A Warthin–Starry staining kit was purchased from BaSO Co., Ltd.. Deparaffinised and rehydrated paraffin sections were stained with 1% silver nitrate in a 56°C water bath for 45 minutes. A developer solution (gelatin, glacial acetic acid, and hydroquinone) was added to the sections until they turned yellow, after which they were rinsed with distilled water at 56°C for 30 seconds.

Hematoxylin–eosin (HE) staining

Deparaffinised, rehydrated sections were hematoxylin-stained (5 minutes), differentiated in 1% acid alcohol (5–10 seconds), blued in running water (5–10 minutes), and then stained with Y-stained (1–3 minutes). The sections were subsequently briefly rinsed in distilled water, followed by dehydration, clearing, and mounting in neutral resin.

Statistical analyses

The data were processed using SPSS Statistics 26.0. At least 3 biological and technical replicates were included for each sample. Continuous data with a normal distribution were expressed as the mean ± standard deviation. For continuous variables, one-way analysis of variance (ANOVA) was used for comparisons among multiple groups, and an independent-samples t test was applied for 2 groups (both for normally distributed data). For nonnormal continuous data (eg, adhesive forces) or ordinal data, the Kruskal–Wallis nonparametric test was used for multiple-group comparisons, with Dunn's test for post hoc pairwise comparisons; and for 2-group comparisons, the Wilcoxon rank-sum test was applied. The threshold for statistical significance was set at *P < .05, **P < .01, and ***P < .001.

Results

EPS alterations in S. mutans biofilms modulate H. pylori adhesion

Gtfs, which are essential for EPS production and biofilm matrix formation in S. mutans,³⁰ enabled us to construct biofilms with varying EPS contents using UA159 and its gtfB- and gtfBC-deficient mutants. CLSM was used to examine the EPS content and 3-dimensional structure of 6-hour biofilms of UA159, UA159ΔgtfB, and UA159ΔgtfBC. Representative 3-dimensional pictures of the EPSs and live bacteria are displayed in Figure 1A. COMSTAT software was used to quantify bacterial and biofilm biomass and characterise the 3-dimensional vertical distribution of bacteria and EPSs (Figure 1B and C). The results demonstrated that compared with the UA159 wild-type biofilm, the UA159ΔgtfB and UA159ΔgtfBC biofilms exhibited a sequential reduction in EPS content (P < .001), with EPSs being nearly undetectable in UA159ΔgtfBC biofilms. Notably, the bacterial biomass did not significantly differ among the 3 biofilms.

Fig. 1.

Effect of the EPS content on the adhesion of S. mutans biofilms to SS1. (A) Confocal laser scanning microscopy (CLSM) images of the bacteria (green) and EPSs (red) of 6-h S. mutans biofilms. Scale bar = 100 μm. (B) Quantification of bacterial and EPS biomass using COMSTAT. The results are presented as the mean ± SD (^⁎⁎⁎P < .001). (C) Quantification of the vertical distribution of bacteria and EPSs within the biofilm structure using COMSTAT. (D) The adhesive force between SS1 and the 3 types of biofilms was measured by atomic force microscopy (AFM).

Successful adhesion and colonisation are critical prerequisites for H. pylori pathogenesis.³¹ To evaluate how EPS alterations in S. mutans biofilms affect H. pylori adhesion, force–distance curves between SS1-coated probes and 6-hour biofilms of UA159, UA159ΔgtfB, and UA159ΔgtfBC were generated by AFM. During the adhesive force measurement, the repulsive force of the SS1-coated probe approaching the biofilm surface increased exponentially with decreasing proximity. Upon withdrawal, the probe must displace sufficient matter to overcome bacterial–biofilm adhesion, causing bond rupture and maximal elastic deformation of the cantilever beam, which is defined as the adhesive force. Random 10 μm × 10 μm regions on each biofilm surface were scanned (64 × 64 lattice), yielding 4096 force–distance curves per region. The results demonstrated that the adhesive forces of the SS1-coated probes to the 3 biofilms decreased sequentially with decreasing EPS content (Figure 1D). Using SS1-coated probes, for the UA159 biofilm dataset (4096 observations), 55–70 nN adhesive forces were most frequent (1497 events, 36.55%); for the UA159ΔgtfB biofilm dataset (4096 observations), 35–50 nN forces were most frequent (1654 events, 40.38%); and for the UA159ΔgtfBC biofilm dataset (4096 observations), 10–25 nN forces were most frequent (1787 events, 43.63%). Following the adhesive force measurements, the bacterial-modified SS1-coated AFM probe was subjected to SEM for observation and verification (Figure S1). Owing to the large sample size of the adhesive force data, normality was assessed using the Kolmogorov–Smirnov test, which showed significant deviations from normality in both the overall and subgroup analyses (P < .05). Therefore, the Kruskal–Wallis nonparametric test with post hoc pairwise comparisons via Dunn's test was used for analysis. The results indicated significant differences: the adhesive force between SS1 and the UA159 biofilms was significantly greater than that between SS1 and the UA159ΔgtfB biofilms (P < .001), and the adhesive force between SS1 and the UA159ΔgtfB biofilms was significantly greater than that between SS1 and the UA159ΔgtfBC biofilms (P < .001).

EPS alterations in S. mutans biofilms modify the colonisation of cocultured H. pylori

To investigate the impact of altered EPS production in S. mutans biofilms on the colonisation dynamics of H. pylori, we compared SS1 monoculture, SS1 cocultured with UA159 biofilm, SS1 cocultured with UA159ΔgtfB biofilm, and SS1 cocultured with UA159ΔgtfBC biofilm (Figure 2A). SS1 monocultures showed a low propensity for structured biofilm formation in vitro because of their fastidious nature. When monocultured for 24 hours, SS1 showed no biofilm formation. SS1 was sparsely distributed on glass slides, with all cell lengths confined to less than 5 μm, and some were even spherical. Conversely, the coculture of SS1 with UA159 biofilm resulted in the development of a dense, 3-dimensional architecture encapsulated within a polysaccharide matrix, with bacterial microcolonies embedded throughout the EPS framework. Compared with those in the SS1 monoculture group, the number of colonies in the SS1 with UA159 biofilm group was significantly greater. Moreover, the SS1 bacterial cells were significantly elongated, reaching more than 10 μm in length, and intertwined with UA159 to form a stable community. Compared with the coculture of SS1 with the UA159 biofilm group, the coculture of SS1 with UA159ΔgtfB biofilm group resulted in a relatively spongy, thin, and porous structure, with decreased amounts of water-insoluble EPSs. Compared with those in the coculture of SS1 with UA159 biofilm group, the adhesion and colonisation numbers in the coculture of SS1 with UA159ΔgtfB biofilm group were significantly lower. In the coculture of SS1 with UA159ΔgtfBC biofilm group, water-insoluble glucan-mediated “scaffold” and “cross-linked network” formation was essentially absent, the number of SS1 colonies in the biofilm was markedly low, and the morphology of SS1 reverted to the short morphology observed in the monoculture (Figure 2B). Electron microscopy results across the 4 groups demonstrated that the biofilm matrix synthesised by the Gtfs of S. mutans UA159 not only facilitated SS1 colonisation but also promoted morphological elongation, which may indicate a more invasive phenotype.

Fig. 2.

Effects of S. mutans biofilms with different EPS contents on H. pylori colonisation. (A) Schematic diagram of the establishment of the SS1 monoculture and SS1–UA159 biofilm, SS1–UA159ΔgtfB biofilm, and SS1–UA159ΔgtfBC biofilm coculture systems. (B) Scanning electron microscopy (SEM) images of the SS1 monoculture and SS1–UA159 biofilm, SS1–UA159ΔgtfB biofilm, and SS1–UA159ΔgtfBC biofilm coculture systems, with yellow arrows indicating SS1. (C) Confocal laser scanning microscopy (CLSM) images of SS1 (blue) and EPSs (red) in the SS1 monoculture and SS1–UA159 biofilm, SS1–UA159ΔgtfB biofilm, and SS1–UA159ΔgtfBC biofilm coculture systems. Scale bar = 100 μm. (D) Quantification of SS1 and EPS biomass using COMSTAT. The results are presented as the mean ± SD (^⁎⁎P < .01, ^⁎⁎⁎P < .001). (E) Quantification of the vertical distribution of SS1 and EPSs within the biofilm structure using COMSTAT.

Subsequently, we acquired 3-dimensional CLSM images of the 4 groups, with red fluorescence used to label biofilm EPSs and blue fluorescence to label H. pylori SS1 (Figure 2C). COMSTAT software was employed to quantify the biomass of biofilm EPSs and SS1, as well as to characterise the 3-dimensional vertical distribution of EPSs and SS1 (Figure 2D and E). The results showed that compared with the SS1 monoculture group, the coculture of SS1 with the UA159 biofilm and the coculture of SS1 with UA159ΔgtfB biofilm groups exhibited significantly higher EPS content and SS1 colonisation levels (P < .05). However, the coculture of SS1 with UA159ΔgtfBC biofilm group showed almost no EPS formation, and the SS1 colonisation level in this group was similar to that in the SS1 monoculture group (P > .05).

Active S. mutans Gtfs can produce EPSs around H. pylori

The above findings reveal that the presence or absence of S. mutans Gtfs significantly affects the adhesion, colonisation, and morphology of its coculture partner, H. pylori. Given the pivotal role of Gtfs, we quantified the enzymatic activity of Gtfs secreted by S. mutans and verified that S. mutans can produce EPSs around H. pylori. The Gtfs from UA159 were analysed by Coomassie blue staining (Figure 3B), and the 2 bands on the left represent GtfB (166 kDa) and GtfC (153 kDa). The enzymatic activity of Gtfs was quantified via colorimetric measurement of EPS production (Figure 3A) using the method of Xu et al. as follows:³² key substrates (sucrose and dextran T-70) were added to phosphate buffer and coincubated with standardised Gtfs for 24 hours to catalyse glucan formation, and the absorbance served as a quantitative indicator of Gtf activity. Comparisons of microplates before and after glucan synthesis by 0–30 μg of Gtfs demonstrated concentration-dependent glucan production (Figure 3C). A bar chart (Figure 3D) based on the Gtf concentration and OD_540nm values, along with correlation analysis (Figure 3E), confirmed a dose-dependent relationship between the Gtf concentration and relative glucan formation content. To investigate whether Gtfs form EPSs around SS1, CLSM was used for comparative analysis. Two SS1 culture groups were established in hydroxyapatite sheets: a negative control (no exogenous Gtfs) and an experimental group (15 μL of 1 μg/μL standardised Gtfs per well). Both were cultured in Brucella broth containing sucrose and dextran T-70 for 6, 12, and 18 hours (Figure 3F). With respect to Gtfs, the EPSs around SS1 gradually increased over time until SS1 was completely encased; without Gtfs, no EPSs formed during the observation period (Figure 3G). Notably, as active Gtfs synthesised EPSs, the fluorescence intensity of live SS1 cells increased concurrently (Figure 3H). These observations suggested that EPS biosynthesis promoted the aggregation of H. pylori and drove the formation of larger bacterial microcolonies of H. pylori. Notably, compared with the Gtf enzyme-free control group, the experimental group treated with exogenous Gtf enzymes exhibited a marked elongation in the cellular morphology of SS1, which is consistent with the results shown in Figure 2B.

Fig. 3.

Glucosyltransferases (Gtfs) mediate the EPS encapsulation of H. pylori SS1. (A) Rapid detection method for Gtf activity. (B) Validation of Gtfs by Coomassie brilliant blue staining. (C) Microplate images before and after Gtf-mediated glucan synthesis. (D) Correlations between the Gtf content and OD₅₄₀ₙₘ in the reaction system. (E) Analysis of the correlation between the relative glucan concentration and the Gtf content. (F) Confocal laser scanning microscopy (CLSM) images showing live SS1 (green) and EPSs (red) in SS1 cultures with or without Gtf supplementation. Scale bar = 10 μm. (G) The relative fluorescence intensity of EPS formation. (H) The relative fluorescence intensity of live SS1. The results are presented as the mean ± SD (^⁎⁎⁎P < .001).

S. mutans EPSs improve H. pylori tolerance to representative antibiotics

To investigate whether the physical encapsulation of SS1 by Gtf-synthesised EPSs protects SS1 against amoxicillin and clarithromycin—the 2 most commonly used antibiotics for clinical H. pylori eradication—we designed 4 experimental groups (Figure 4A) and determined the SS1 survival rate for each group. The first 3 groups involved SS1 cocultured with biofilms of 3 S. mutans strains, followed by 12 hours of antibiotic intervention. In the fourth group, S. mutans was omitted; instead, in vitro-generated EPSs were added to the SS1 cultures before 12 hours of antibiotic intervention. We performed CFU counting to determine the survival rate of SS1 in the presence of amoxicillin and clarithromycin. SS1 cocultured with UA159ΔgtfB biofilm responded similarly to SS1 cocultured with UA159ΔgtfBC biofilm when challenged with amoxicillin and clarithromycin. Compared with the survival rate of SS1 cocultured with the UA159 biofilm, the survival rates of SS1 in cocultures with Gtf-deficient strain biofilms were significantly lower under amoxicillin (P < .001) and clarithromycin (P < .001) stress. Notably, the addition of exogenous EPSs to the biofilms of SS1 cocultured with UA159ΔgtfBC restored SS1 survival (P < .05), confirming that S. mutans EPSs are a key factor in the protection of SS1 against antibiotics in dual-species biofilms (Figure 4B and C).

Fig. 4.

SS1 tolerance to antibiotics varies depending on the presence of EPSs. (A) Schematic diagram illustrating the construction of 4 experimental groups, with a colony-forming unit (CFU) counting was performed for each group. (B) Survival rate of SS1 under amoxicillin pressure. (C) Survival rate of SS1 under clarithromycin pressure.

S. mutans EPS-derived plaque biofilms enable H. pylori to resist antibiotics and colonise the oral cavity

Next, we explored the role of S. mutans EPSs in mediating H. pylori tolerance to amoxicillin and clarithromycin antibiotics within mixed-species biofilms using in vivo models. Through ESEM and CFU assays, we compared how UA159, UA159ΔgtfBC, and exogenous Gtfs affect SS1 colonisation under amoxicillin or clarithromycin treatment. Artificial cavities were drilled in rat molars to simulate caries morphology, with bacterial suspensions and subsequent antibiotics precisely injected into these cavities. The rats were randomly allocated to 8 experimental groups (Figure 5A). SS1 group: PBS injection for 5 days, followed by SS1 injection for 10 days and local amoxicillin or clarithromycin injections for 5 days. UA159 + SS1 group: UA159 injection for 5 days, followed by SS1 injection for 10 days and local amoxicillin or clarithromycin injections for 5 days. UA159ΔgtfBC + SS1 group: UA159ΔgtfBC injection for 5 days, followed by SS1 injection for 10 days and local amoxicillin or clarithromycin injections for 5 days. Gtfs + SS1 group: Pretreatment with Gtf protein (1 μg/μL) twice daily for 5 days, followed by SS1 injection for 10 days and local amoxicillin or clarithromycin injections for 5 days. Dental plaque biofilms on the tooth samples were visualised using a disclosing agent (Figure 5B), and the stained areas were quantified via ImageJ (Figure 5C). Consistent trends were observed for both antibiotics. Plaque areas in the UA159 + SS1 and Gtfs + SS1 groups were significantly larger than those in the SS1 and UA159ΔgtfBC + SS1 groups (P < .001). The plaque area in the UA159ΔgtfBC + SS1 group was comparable to that in the SS1 group. Compared with the UA159 + SS1 group, the UA159ΔgtfBC + SS1 group exhibited a significantly smaller plaque area (P < .001), which is consistent with the role of Gtfs in biofilm formation. CFU quantification (Figure 5D) revealed that compared with the other 2 groups, the UA159 + SS1 and UA159ΔgtfBC + SS1 (with exogenous S. mutans) groups had significantly greater S. mutans colonisation on tooth surfaces (P < .001). Under both antibiotics, the degree of SS1 colonisation on teeth was significantly greater in the UA159 + SS1 and Gtfs + SS1 groups than in the SS1 and UA159ΔgtfBC + SS1 groups (P < .05), with no significant difference between the UA159ΔgtfBC + SS1 and SS1 groups. Notably, SS1 colonisation in the Gtfs + SS1 group was even greater than that in the UA159 + SS1 group (P < .05) under both antibiotics, indicating that exogenously added active Gtfs synthesise EPSs in vivo to promote SS1 colonisation. Consistent with the in vitro findings, alterations in Gtfs (deletion or exogenous supplementation) affect SS1 colonisation in rat dual-species biofilms. ESEM analysis of bacterial composition and abundance in dental plaque across the 8 groups (Figure 5E and F) revealed consistent trends for both antibiotics: In the UA159 + SS1 group, numerous UA159 and SS1 cells were entangled and interwoven, with extensive plaque biofilms observed on the tooth surfaces. In the Gtfs + SS1 group, many SS1 cells adhered to the cavity walls and were encased in an EPS-like matrix. In the SS1 group, SS1 formed clustered aggregates on the cavity walls but failed to develop biofilms of a comparable scale. In contrast, SS1 was barely detectable in the UA159ΔgtfBC + SS1 group, where biofilms were extremely sparse and lacked a structured architecture.

Fig. 5.

Effects of EPSs and consequent dental plaque on oral colonisation by SS1 under antibiotic pressure. (A) Flow chart of the animal experiments, including 8 groups. Each of the 4 base groups (SS1, UA159 + SS1, UA159ΔgtfBC + SS1, and Gtfs + SS1) was treated with either amoxicillin or clarithromycin. (B) Representative images of dental plaque staining on molars from model rats. (C) Quantification of dental plaque on the model molars using ImageJ. (D) Colony-forming unit (CFU) quantification of dental plaque on the model molars. (E) Representative environmental scanning electron microscopy (ESEM) micrographs of dental plaque under amoxicillin intervention in the SS1, UA159 + SS1, UA159ΔgtfBC + SS1, and Gtfs + SS1 groups. (F) Representative ESEM micrographs of dental plaque under clarithromycin intervention in the SS1, UA159 + SS1, UA159ΔgtfBC + SS1, and Gtfs + SS1 groups, with S. mutans indicated by yellow arrows and H. pylori indicated by white arrows. The results are presented as the mean ± SD (^⁎⁎P < .01 and ^⁎⁎⁎P < .001 compared with the SS1 group; ^#P < .05, ^##P < .01, and ^###P < .001 compared with the UA159 + SS1 group; and ^$$$P < .001 compared with the UA159ΔgtfBC + SS1 group).

Effects of varying degrees of oral colonisation by H. pylori on gastric tissues in rats

Warthin–Starry staining (Figure 6A and C) was performed on rat gastric tissues across all 8 groups, with blue arrows indicating SS1 invasion. HE staining (Figure 6B and D) revealed mucosal inflammatory foci with neutrophil infiltration (blue boxes) and hemorrhagic sites (red arrows). SS1 colonisation scoring and pathological scoring were performed on 3 microscopic fields of gastric tissues from 8 rats per group (Figure 6E and F). For SS1 colonisation, among the amoxicillin-treated groups, colonisation in the UA159ΔgtfBC + SS1 group was significantly lower than that in the UA159 + SS1 group (P < .01). Among the clarithromycin-treated groups, the UA159ΔgtfBC + SS1 group had the lowest mean score, although this score did not differ significantly from those of the other 3 clarithromycin-treated groups. Under both antibiotics, the Gtfs + SS1 group had lower colonisation than the UA159 + SS1 group but significantly higher colonisation than the UA159ΔgtfBC + SS1 group, confirming that oral Gtfs facilitate SS1 colonisation. In terms of pathology, mucosal erosion severity showed no significant intergroup differences under either antibiotic. Gastric mucosal inflammation was greatest in the UA159 + SS1 groups under both antibiotics, with significant differences only between the UA159 + SS1 and UA159ΔgtfBC + SS1 groups under clarithromycin (P < .05).

Fig. 6.

Effects of varying degrees of oral colonisation by H. pylori on gastric tissues. (A) Representative Warthin–Starry-stained images of gastric tissues from the SS1, UA159+SS1, UA159ΔgtfBC+SS1, and Gtfs+SS1 groups under amoxicillin treatment; H. pylori is marked by blue arrows. (B) Representative hematoxylin–eosin (HE)-stained images of gastric tissues from the above groups under amoxicillin treatment. Inflammatory regions are marked by blue boxes, and hemorrhagic sites are marked by red arrows. (C) Representative Warthin–Starry-stained images of gastric tissues from the SS1, UA159+SS1, UA159ΔgtfBC+SS1, and Gtfs+SS1 groups under clarithromycin treatment; H. pylori is marked by blue arrows. (D) Representative HE-stained images of gastric tissues from the above groups under clarithromycin treatment. (E) Histopathological scoring of gastric tissues from the above groups under amoxicillin treatment. (F) Histopathological scoring of gastric tissues from the above groups under clarithromycin treatment. The results are presented as the median (interquartile range, IQR), ^⁎⁎P < .01.

Discussion

Whether H. pylori can survive and colonise the human oral cavity and whether it is a permanent resident or temporary dweller remain controversial.³³ The literature has employed diverse methodologies—including immunological, biochemical, and molecular biological approaches—to detect H. pylori in oral specimens, with all techniques reporting notably high detection rates.34, 35, 36 However, H. pylori is widely regarded as nonculturable from oral samples,^34,37,38 and only one study has documented its cultivation from primary tooth root canal specimens rather than adjacent dental plaque.³⁹ Despite culturing being the gold standard for clinical H. pylori detection,⁴⁰ the failure to isolate the bacterium from oral samples may be due to the lower load of H. pylori in the oral cavity than in the stomach,⁴¹ a rod-to-sphere morphological transition that inhibits growth in culture media,⁴² loss of viability during sample processing,⁴³ or inadequacy of existing culture media to support H. pylori growth amid complex oral flora.⁴⁴ To investigate whether H. pylori can survive in the human oral cavity, a recent study reported that planktonic H. pylori strain SS1 can survive in human saliva, even with and in some cases favoured by the presence of certain oral microorganisms (including planktonic S. mutans).⁴⁵ However, in the oral cavity, microorganisms typically exist as organised biofilms covering the surfaces of teeth and dental restorations, which is one of the core focuses of our study.

In the oral cavity, S. mutans leverages its robust EPS biosynthesis to serve as the primary producer of the biofilm matrix.⁴⁶ The EPSs secreted by S. mutans aid various bacteria within the biofilm in resisting adverse oral environments through the following mechanisms: (1) EPSs promote microbial aggregation,²¹ forming a barrier with spatial heterogeneity and diffusion-limiting properties that protects encapsulated bacteria;^15,22 this physical barrier also impedes antibiotic penetration.²⁴ (2) EPSs maintain a heterogeneous microenvironment within the biofilm²⁵—characterised by local hypoxia, nutrient gradients, and acid–base differences—and induce some bacteria to enter a metabolically inactive “persister cell” state.²⁶ These persister cells exhibit significantly reduced antibiotic sensitivity, enabling them to evade eradication.^27,28 (3) EPSs mask bacterial surface antigens to prevent antibody recognition and inhibit the phagocytic and bactericidal activities of immune cells, thus interfering with host immune responses.²⁹ EPSs play similar critical roles in H. pylori. EPSs provide stable adhesion sites, enabling H. pylori to resist mechanical clearance (eg, salivary flushing and masticatory forces) and facilitating long-term colonisation.⁴⁷ Additionally, EPSs contribute to H. pylori tolerance to amoxicillin and clarithromycin antibiotics. Notably, H. pylori is an obligate microaerophile, and a marked difference in oxygen partial pressure exists between the oral cavity and stomach; the stomach maintains a microaerobic milieu, whereas most oral regions remain normoxic because of atmospheric exposure.⁴⁸ However, EPSs help sustain a stratified gas concentration gradient within oral biofilms, creating a gaseous microenvironment conducive to the survival of microaerophiles/anaerobes—a feature that strongly promotes H. pylori colonisation and adaptation.^49,50 Collectively, these mechanisms constitute key nodes of the synergistic interaction between S. mutans biofilms and H. pylori, providing a novel framework for understanding the oral reservoir of H. pylori.

The animal experimental design of this study was as follows: artificial cavities were drilled in rat molars to simulate caries morphology, with bacterial suspensions precisely injected into the cavities using syringes; antibiotics were also locally injected into the cavities during intervention, establishing an in vivo dental dual-species biofilm colonisation model. This study examined group differences in oral H. pylori colonisation and whether such differences in plaque biofilms induce differential gastric lesions to clarify the effects of oral H. pylori on gastric health. Notably, despite fine-needle intracavity inoculation, bacterial suspension ingestion via oral movements (eg, swallowing) during injection could not be fully excluded. Therefore, the observed gastric H. pylori colonisation may have resulted from oral colonisation or suspension swallowing during modelling, but not solely the former. Although this limitation was recognised, no optimised modelling method exists, potentially confounding gastric lesion outcomes.

The in vitro experiments in this study failed to fully recapitulate the physiological characteristics of the human oral cavity and thus could not simulate key elements of the oral environment, including food intake, mechanical stimuli induced by toothbrushing and speaking, and the dynamic secretion and flow of saliva, as well as the regulatory effects of components such as mucins, lactoferrin, and lysozymes.⁵¹ Additionally, the experiments lacked the replication of circadian rhythm-driven periodic fluctuations in salivary secretion volume, compositional changes, the pH value, and oxygen supply. These unmodelled factors may lead to discrepancies between the observed bacterial interaction patterns and biofilm phenotypes and their actual in vivo states. Therefore, future studies could optimise the experimental design by developing a bionic oral food processing system, leveraging fluid dynamics simulation technologies, and establishing dynamic saliva simulation systems as well as rhythmic culture systems.^52,53

Epidemiological studies have confirmed 2 key associations: a positive correlation between the dental plaque index and oral H. pylori detection rate and a close link between dental caries and H. pylori infection. The former, the focus of this paper, is detailed in the Introduction. For the latter, a cross-sectional study of 841 Chinese adults revealed a 73.52% (422/574) prevalence of caries in 574 H. pylori-positive individuals, which was significantly greater than the 35.21% (94/267) in H. pylori-negative individuals, as well as a higher average plaque index in H. pylori-positive individuals.¹⁴ Another cross-sectional study of 1,050 Chinese adults revealed that the oral H. pylori detection rate was significantly higher in caries patients (66.91%) than in caries-free individuals (54.07%, P < .05).⁵⁴ This association also exists in children; studies have shown a positive correlation between the severity of dental caries (by the decayed–missing–filled teeth index) and the H. pylori infection rate.^55,56 As a key cariogen, S. mutans forms biofilms via EPS production and induces caries through acidogenesis.⁵⁷ Notably, H. pylori can acquire complementary resistance by taking up L-lactic acid, thus facilitating gastrointestinal colonisation.⁵⁸ Intriguingly, S. mutans generates large amounts of L-lactic acid via the glycolysis of glucose.^59,60 Therefore, the acid production of S. mutans may represent an additional synergistic mechanism (beyond the biofilm EPSs explored herein) promoting H. pylori colonisation.

Clinical first-line H. pylori eradication therapies (eg, quadruple therapy) target gastric mucosal-colonising strains.⁶¹ The often overlooked “oral H. pylori reservoir” is likely a key driver of post-eradication gastric reinfection,⁶² contributing to eradication failure or short-term reinfection in ∼13% of patients.⁶³ A major clinical challenge persists: systemic triple or quadruple therapies exhibit differences in eradication efficacy for gastric vs oral H. pylori (85.8% vs 5.7%, OR 55.59, P < .00001),⁶⁴ highlighting the need for adjuvant local interventions to eliminate the oral reservoir.⁶⁵ For example, during systemic eradication, basic periodontal therapy (eg, subgingival scaling and root planing), enhanced oral hygiene (toothbrushing and oral hygiene education), mouth rinsing with 0.02% tinidazole plus 0.12% chlorhexidine, and ultrasonic scaling reduce oral H. pylori load and blocks its orogastric translocation.66, 67, 68, 69

Conclusion

In this study, we compared the effects of EPS content variations on H. pylori adhesion, colonisation, and drug tolerance between S. mutans UA159 wild-type and gtfB/gtfC-deficient strains. Our findings reveal that these H. pylori phenotypes are mediated by S. mutans biofilms through Gtf-driven EPS biosynthesis, revealing a dual-species synergistic pathogenic mechanism: S. mutans biofilms facilitate the establishment of oral reservoirs for H. pylori, which may in turn promote its translocation to the gastrointestinal tract. This discovery reinforces the intimate link between oral health and systemic health.

Author contributions

X.J. Huang and Y.Q. Li contributed to the conception and revision of the manuscript. Y.J. Hao contributed to the experiment design, data acquisition, and manuscript drafting. Y.Q. Huang and R.N. Chen contributed to data analysis.

Ethics approval

All experiments complied with animal welfare guidelines and were approved by Fujian Medical University IACUC (No. IACUC FJMU 2024-0382).

Conflict of interest

None disclosed.