A sucrose-rich diet promotes Clostridioides difficile carriage without prior antibiotics and significantly exacerbates both acute disease and long-term colonization

Abstract

Clostridioides difficile is a major cause of healthcare-associated diarrhea, with rising rates of community-acquired infections and asymptomatic carriage. While antibiotic exposure is a well-established risk factor, the role of diet in modulating susceptibility remains underexplored. Here, we demonstrate that a high-sucrose diet profoundly alters host susceptibility to C. difficile in a murine model. Mice consuming sucrose-rich chow exhibited exacerbated disease severity, characterized by increased weight loss, elevated clinical scores, heightened toxin burden, and persistent intestinal inflammation. Mice fed a high-sucrose diet failed to clear C. difficile and remained colonized long-term remaining susceptible to recurrent disease. Critically, high-sucrose-diet mice were susceptible to asymptomatic C. difficile carriage without prior antibiotic treatment, which progressed to overt CDI upon antibiotic exposure. Microbiome and metabolome profiling revealed that consumption of a sucrose-rich diet reshaped the gut microbiota, marked by blooms of Enterococcus and Akkermansia, a reduction in beneficial taxa, and remodeled the metabolome to favor C. difficile germination and growth. These findings establish dietary sucrose as a modulator of colonization resistance and identify a novel model of diet-induced asymptomatic carriage, with implications for the rising burden of community-associated C. difficile infection.

Introduction

Clostridioides difficile is a spore-forming, obligate anaerobe and the leading cause of healthcare-associated infectious diarrhea in the United States. Despite recent declines in hospital-acquired infections due to improved antibiotic stewardship and infection control, the incidence of community-acquired C. difficile infection (CDI) and asymptomatic carriage is rising.^1,2 Notably, C. difficile is detectable in up to 15% of healthy adults and as high as 50% in elderly populations.^1,2 While antibiotic exposure remains the most well-established risk factor for CDI, the mechanisms underlying asymptomatic carriage in otherwise healthy individuals remain poorly defined.^3,4 Currently, all established animal models of CDI require antibiotic pretreatment to facilitate colonization, and no model exists for asymptomatic carriage in the absence of antibiotics.

Modern Western diets, characterized by high intake of refined sugars and fats, are known to promote intestinal inflammation and microbial dysbiosis.^5,6 Americans consume, on average, over 100 g of added sugar daily, far exceeding recommended limits.⁷ This threshold is easily met; a single 12-ounce can of common sugar-sweetened beverages typically contains ~34 g of sugar. Previous studies have shown that dietary components can modulate CDI severity. In mouse models, high-protein and high-fat diets exacerbate CDI,^8–11 whereas fiber-rich diets confer protection by enhancing the production of microbiota-derived metabolites.^12–16 Specific sugars such as trehalose, sorbitol, glucose, and fructose have been implicated in promoting C. difficile virulence, spore production, and colonization.^17–20 A high-carbohydrate diet – comprising of simple carbohydrates (corn starch, 43.5% [w/v]; maltodextrin, 14.4% [w/v]; and sucrose, 11.0% [w/v]) – was protective against severe disease. However, mice on this diet became long-term asymptomatic carriers of C. difficile, suggesting that high-carbohydrate intake may create a gut environment conducive to colonization.⁹

C. difficile is a carbohydrate generalist, capable of metabolizing a variety of simple sugars via phosphotransferase systems.²¹ These include cellobiose, fructose, glucose, mannitol, mannose, melezitose, sorbitol, trehalose, and mucin derivatives.^22–26 However, sucrose is not directly metabolizable by C. difficile.

Here, we investigate the impact of a high-sucrose diet (HSD) on C. difficile colonization, disease severity, and microbiome-metabolome dynamics in a murine model. We demonstrate that a HSD exacerbates acute CDI, impairs bacterial clearance, and promotes long-term colonization. Strikingly, we show that consumption of a HSD alone or sucrose supplemented in drinking water is sufficient to permit asymptomatic C. difficile carriage in the absence of antibiotics, which can progress to fulminant disease upon subsequent antibiotic exposure. Through integrated microbiome and metabolome analyzes, we reveal that a HSD induces a dysbiotic gut environment depleted in microbes antagonistic to C. difficile, enriched in Enterococcus and Akkermansia, and metabolites that favor C. difficile germination and growth. These findings establish a novel model of diet-induced asymptomatic carriage and highlight the underappreciated role of dietary sugars in shaping susceptibility to C. difficile.

Materials and methods

A comprehensive list of reagents, consumables, and equipment used in this study is provided in Table S1.

Bacterial growth conditions

C. difficile strain CD2015 (ribotype 027) was maintained in 10% DMSO stocks at −70 °C. For experimental use, cultures were grown anaerobically at 37 °C in brain heart infusion broth supplemented with 0.5% yeast extract (BHIS).

Single sugar growth assays were performed using defined minimal media (CDMM), as described previously.²⁷ Overnight cultures grown in BHIS were washed once in reduced CDMM and diluted 1:20 into CDMM supplemented with 0.5% sucrose or glucose in a 96 well plate in technical triplicate, and the OD600 was monitored every 10 min for 24 h. The experiment was independently repeated 3−4 times.

Spore preparation

Spore preparation was performed as previously described.²⁸ Briefly, overnight BHIS cultures were diluted 1:10 in sterile, pre-reduced PBS and plated onto pre-reduced 70:30 agar to induce sporulation.²⁹ Plates were incubated anaerobically at 37 °C for 4–5 d. Spores were harvested by washing the agar surface with sterile water, followed by incubation at 4 °C for 24 h to facilitate spore release. The suspension was layered over 50% sucrose and centrifuged at 2680 × g for 20 min to separate spores from debris. Spores were washed five times in sterile water and stored at 4 °C in 1% BSA. Prior to use, spore stocks were enumerated on both BHIS + 0.1% sodium taurocholate (to facilitate spore germination) and BHIS + 0.1% sodium taurocholate supplemented with moxalactam (32 µg ml⁻¹) and norfloxacin (12 µg ml⁻¹) to ensure stock purity. Equivalent CFUs between non-selective and selective media confirmed spore stock purity. All spore stocks used in animal experiments were confirmed to be free of microbial contaminants.

Animal ethics and housing

This study was conducted in strict accordance with the recommendations outlined in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee of the University of Louisville (IACUC 21899).

Male and female C57BL/6J mice (6–8 weeks old) were obtained from Jackson Laboratories. Mice were housed in specific pathogen free (SPF) conditions with autoclaved or irradiated food, bedding, and water, and animal handling procedures were performed in sterile biosafety cabinets. Humane endpoints included >20% weight loss, lack of ambulation, or unresponsiveness. Euthanasia was performed via CO₂ asphyxiation followed by cervical dislocation.

Diets

Standard laboratory rodent chow (LD.5010) was used as the control diet. AIN-93M, a commonly used purified control diet, was not employed in this study due to its compositional similarity to high-carbohydrate diets; it contains a high proportion of simple carbohydrates (primarily corn starch, maltodextrin, and sucrose), making it an unsuitable comparator. The test diet, TD.98090 (Inotiv), was used as a high-sucrose diet. This formulation contains 70% carbohydrate by weight, of which 94.3% is sucrose, with the remainder derived from corn starch and maltodextrin. Both diets were approximately isocaloric, with energy densities of 4.17 kcal/g (control) and 4.00 kcal/g (sucrose). Diet macronutrients are outlined in Table 1. Mice were maintained on their respective diets ad libitum throughout the experimental period.

Table 1.

Diet macronutrient composition by % weight. Remaining components derived from vitamins, minerals, moisture and ash. NFE = nitrogen free extract. NDF = neutral detergent fiber.

Component	TD.98090	LabDiet 5010
Protein	17.7%	24.6%
Fat	5.2%	5.0%−6.5%
Carbohydrate	70.0%	49.9% (NFE)
Fiber	0.989% (cellulose)	4.2% (crude), 14.9% (NDF)

For additional experiments modeling asymptomatic colonization, all groups were fed control chow (LD.5010), high-sucrose chow (TD.98090), or control chow (LD.5010) plus sucrose-supplemented drinking water (110 mg/mL), a concentration comparable to that found in sugar-sweetened beverages.

Mouse model of C. difficile infection

Each mouse experiment was performed at least twice with approximately equal numbers of male and female mice. For diet switching experiments female mice were used to enable random switching of mice to new groups. Prior to starting, all mice were tested for C. difficile via selective plating. Mice were fed their respective diets for 14 d prior to infection with C. difficile. Where applicable, dietary fiber (inulin or polydextrose) was added to drinking water (5 mg mL⁻¹) starting 14 d before infection. Where applicable, mice were administered a single IP injection of clindamycin (10 mg kg⁻¹) 24 h before infection to induce CDI susceptibility. Uninfected control mice received diets, antibiotics, and were gavaged with sterile water.

Mice were infected via oral gavage with 10⁵ CD2015 (RT027) spores. Clinical signs were monitored daily and scored by a blinded observer using a validated scoring system that accounted for behavior, stool consistency, and weight loss.³⁰ Each category was scored from 0 to 4 (where a score of 0 indicates normal behavior, formed stool, and no change in weight, and a score of 4 indicates inability to ambulate, unresponsiveness, mucous stool, and significant weight loss), and the individual values were added to provide an overall severity score. Stool samples were collected at d = −14, −7, −1, 0, and every 2−3 d post-infection for bacterial enumeration, toxin quantification, Lipocalin-2 measurement, metabolomics, and 16S rDNA sequencing. All fecal pellets were collected fresh by scruffing the mice and inducing defecation. Fecal pellets for C. difficile enumeration were used immediately, while other samples were stored at −70 °C until needed. In experiments where diets were switched at d = 33 post-infection, mice were randomly selected to remain on their respective diets or be switched from a control diet to a HSD or vice versa.

C. difficile enumeration

Fresh fecal pellets were collected into pre-weighed 1.5 ml microcentrifuge tubes and re-weighed to determine the pellet weight. Pellets were transferred to an anaerobic chamber within 1 h of collection and diluted 1:10 w/v in sterile, pre-reduced, PBS. Following homogenization, supernatants were serially diluted into a 96 well PCR plate. Total C. difficile was enumerated by spot-plating onto pre-reduced CDMN agar³¹ containing 0.1% sodium taurocholate (CDMNT), spore enumeration was conducted by heat-killing the vegetative cells at 65 °C for 20 min prior to plating. The vegetative CFU/g was calculated by subtracting the spore CFU/g from the total CFU/g.

Gut permeability assay using FITC-dextran

Intestinal permeability was assessed using fluorescein isothiocyanate-dextran (FITC-dextran, 4 kDa). Mice were fasted for 4 h with free access to water prior to gavage. Each mouse was administered FITC-dextran at a dose of 600 mg/kg body weight, dissolved in phosphate-buffered saline (PBS), via oral gavage.

After 4 hours, mice were euthanized, and blood was collected via cardiac puncture. Serum was separated by centrifugation at 10,000 × g for 10 min at 4 °C. Serum FITC-dextran concentrations were measured using a fluorescence plate reader (excitation: 485 nm; emission: 528 nm). A standard curve was generated using serial dilutions of FITC-dextran in PBS. Gut permeability was quantified by comparing serum fluorescence values to the standard curve.

Toxin and lipocalin-2 quantification

Toxins A and B were quantified using a commercial ELISA kit (Eagle Biosciences) and normalized to stool mass. Lipocalin-2 levels were measured using a mouse-specific ELISA (R&D Systems). Fecal samples were homogenized in diluent (25 mg mL⁻¹), centrifuged, and supernatants were diluted and analyzed per manufacturer instructions.

Fecal metabolomics

Fecal samples for metabolomic analysis were collected on days −7 (pre-infection) and 14 or 21 (post-infection). Three to five fecal pellets were pooled within the diet groups at each time point to ensure that a dry weight of ~50 mg was obtained following lyophilization. Metabolites were extracted using 800 μl of ice-cold 1:1 chloroform–methanol solution, homogenized (Omni International Bead-ruptor 96 for 2 min at 30 Hz), and centrifuged at 15,000 × g for 5 min at 4 °C. An additional 800 µL of chloroform-methanol was added to the pellet, and the homogenization and centrifugation steps were repeated. Ice-cold HPLC-grade water (600 µL) was added to the chilled supernatants, and the samples were vortexed for 30 s. The samples were centrifuged at 3140 × g for 8 min at 4 °C and the aqueous phase was filtered through a 0.2 µm filter and concentrated using 3 kDa filters. Samples were stored at −80 °C until ready for use.

Sample analysis was carried out at Texas A&M University as follows. Untargeted liquid chromatography high resolution accurate mass spectrometry (LC-HRAM) analysis was performed on a Q Exactive Plus orbitrap mass spectrometer (Thermo Scientific, Waltham, MA) coupled to a binary pump HPLC (UltiMate 3000, Thermo ScientificFull MS spectra were obtained at 70,000 resolution (200 m/z) with a scan range of 50−750 m/z. Full MS followed by ddMS2 scans were obtained at 35,000 resolution (MS1) and 17,500 resolution (MS2) with a 1.5 m/z isolation window and a stepped NCE (20, 40, 60). Samples were maintained at 4 °C before injection. The injection volume was 10 µL. Chromatographic separation was achieved on a Synergi Fusion 4 µm, 150 x 2 mm reverse phase column (Phenomenex, Torrance, CA) maintained at 30 °C using a solvent gradient method. Solvent A was water (0.1% formic acid). Solvent B was methanol (0.1% formic acid). Sample acquisition was performed Xcalibur (Thermo Scientific). Data analysis was performed with Compound Discoverer 3.3 (Thermo Scientific).

Ex-vivo germination assay

Cecal contents from mice fed either a control or HSD were weighed, diluted 1:2 w/v in pre-reduced carbon-free defined media,³² and pelleted by centrifugation. Supernatants were filter-sterilized and 10³ CD2015 spores added to 100 µL of supernatant in a sterile tube. Spores were incubated at 37 °C for 30 min under anaerobic conditions. Germination was assessed by plating on BHI agar. Taurocholate-supplemented media served as the positive control, while carbon-free defined media was used for the negative control.

16S rDNA sequencing

16 s rDNA sequencing was performed as described by Kozich et al.³³ Briefly, DNA was extracted using the Qiagen DNeasy PowerSoil HTP 96 Kit. The V4 region was amplified with barcoded primers, checked via gel electrophoresis, and normalized using the SequelPrep Normalization Kit. Amplicons were pooled and sequenced on an Illumina MiSeq platform. Data were processed using the MOTHUR pipeline as described previously.³³

For 16 s profiling of uninfected control samples, stool pellets were stored at −70 °C until required. Sample extraction and sequencing was performed by Zymo Microbiome Sequencing Services.

Absolute abundance quantification

The number of genome copies per microliter DNA sample was calculated by dividing the gene copy number (determined by real-time PCR using a standard curve of 16S gene) by an assumed number (n = 4) of gene copies.

Statistics

Statistical analyzes were performed using R (v4.4.1). Welch's t-test or Wilcoxon rank-sum test were used, as appropriate, with Holm or FDR correction. When required, data normalization was performed using the bestNormalize package³⁴ and validated with visual inspection (histograms and Q–Q plots) and statistical tests (Shapiro–Wilk and Kolmogorov–Smirnov tests). For data with multiple groups (>2) across time series and/or multiple experiments, linear mixed models (LMMs) were fitted using the lmer function from the lme4 package³⁵ with group and time as fixed effects and mouse and experiment as random. Estimated marginal means were calculated using the emmeans package. For example, the following model was fitted to assess mouse weights between diet groups with fiber supplementation:

model = lmer(Weight ~ Group * factor(DPI) + (1|Experiment) + (1|mouse), data = weight_data)

The reference level for Group was set to ‘Control’ to facilitate comparisons of all test groups against the control group at each time point and estimated marginal means were calculated for Group within each level of DPI:

emms = emmeans(model, ~ Group | factor(DPI))

Pairwise comparisons were then performed:

pairs(emms, by = ‘DPI’, adjust = ‘BH’)

This approach allowed us to isolate the effects of each test group relative to the control group at each specific time point, ensuring that comparisons were not confounded by differences across time points or experiments.

Microbiome differential abundance was assessed using microViz,³⁶ MaAsLin2,³⁷ Corncob,³⁸ and Limma-Voom with TMM normalization.³⁹

Results

A high-sucrose diet exacerbates CDI severity and impairs clearance

Diets rich in simple sugars promote intestinal inflammation and microbial dysbiosis - conditions that favor the proliferation of enteric pathogens.^40–42 While sucrose is the most widely consumed sugar globally, C. difficile cannot directly metabolize it (Figure 1A). Therefore, we investigated whether the indirect effects of high sucrose consumption influenced C. difficile infection and carriage.

Figure 1.

A high-sucrose diet increases gut inflammation and permeability. (A) C. difficile cannot metabolize sucrose as a carbon source. C. difficile CD2015 was grown for 24 h in either minimal media without a carbon source (CDMM) or supplemented with 0.5% glucose or sucrose. (B) Macromolecule constituents of the control and HSD by % kcal. Both diets have a similar overall energy density of 4.17 kcal/g for the control diet and 4.0 kcal/g for the HSD. (C) Change in weight following 2 weeks on the diet. Four male and four female mice were fed the respective diets for 2 weeks, and their weights were monitored. There was no significant difference in weight gain between the two groups after 2 weeks of feeding. (D) HSD mice had significantly higher fecal lipocalin levels after 2 weeks of feeding as determined by ELISA. (E) HSD mice had significantly higher intestinal permeability after 2 weeks of feeding as determined by FITC-Dextran Intestinal Permeability Assay.

A high-sucrose diet increases intestinal inflammation and permeability

To assess the impact of dietary sucrose, mice were fed either a control or HSD ad libitum for 2 weeks prior to infection. Diets were approximately isocaloric, with energy densities of 4.17 kcal/g (control) and 4.0 kcal/g (sucrose), representing a <5% difference and containing comparable macronutrient compositions (Table 1 and Figure 1B). Although overall weight gain did not differ significantly, HSD mice exhibited significantly greater variability in weight (p = 0.018, Levene's test for homogeneity of variance). Mice fed the HSD had elevated fecal lipocalin-2 (367.0 ± 253.0 pg/mg vs. 94.2 ± 89.0 pg/mg, p = 0.008) and increased intestinal permeability, as measured by FITC-dextran translocation (0.42 ± 0.05 µg/ml vs. 0.32 ± 0.03 µg/ml, p = 0.002), consistent with heightened basal inflammation and compromised epithelial barrier integrity (Figure 1C-E).

High-sucrose diet mice fail to clear C. difficile following infection

To evaluate the effect of dietary sucrose on the course of CDI, we set up a model that incorporates acute disease, C. difficile clearance, the effect of dietary switching, and disease relapse (Figure 2A).

Figure 2.

Impact of a high-sucrose diet on CDI disease progression. (A) Model of acute CDI followed by long-term C. difficile carriage. (B–D) A HSD results in an inability to clear C. difficile. Points represent total C. difficile burden (vegetative cells + spores). (B) Mice on the control diet cleared C. difficile below the limit of detection (LOD) by ~2 weeks while the C. difficile burden remained high in HSD mice. (C) Mice were randomly assigned a new diet 33 d post infection. Switching from a HSD to a control diet was insufficient to enable clearance. Conversely, switching to a HSD significantly increased C. difficile burden back above the LOD. (D) Clindamycin induced relapse resulted in a spike in C. difficile burden for all groups except those that remained on the control diet. (E–G) Effects of diet on body weight changes during (E) initial infection, (F) Diet switch (33 DPI), and (G) Relapse induced by clindamycin IP. (H) A HSD resulted in significantly higher clinical scores during the initial infection period (p < 0.001, Wilcoxon test of area under the curve [AUC]). (I) HSD mice harbored significantly more C. difficile toxin during acute infection. (J) HSD mice had significantly elevated lipocalin levels both pre- and post-infection. (K) Mice switched from a control diet to a HSD experienced significantly higher clinical scores during relapse compared to mice that remained on the control diet (p = 0.036, Wilcoxon test of area under the curve [AUC]). Boxplots show the medians (middle line) and interquartile ranges, points indicate biological replicates, lines are mean, error bars are 95% CI. All p values are controlled for multiple testing where appropriate. (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). Female mice were used for diet switching experiments to enable random regrouping of mice.

Mice were pre-fed either control or HSD chow for 2 weeks, followed by a single intraperitoneal dose of clindamycin and oral challenge with 10⁵ C. difficile CD2015 spores. In control-fed mice, C. difficile was cleared to below the limit of detection by day 18 post-infection. In contrast, HSD mice maintained high bacterial burdens (1.48 × 10⁸ CFU/g on day 18 and 5.77 × 10⁷ CFU/g on day 30; Figure 2B). To determine whether the elevated burden was influenced by a change in the spore-to-vegetative cell ratio, we examined the first 10 d of infection using a linear mixed model with diet group as the fixed effect, during which both groups were shedding C. difficile. The ratio of spores was not significantly altered by diet (β = 0.09, SE = 0.02, t = 5.38, p = 0.67), suggesting that persistence was not due to changes in sporulation dynamics (Figure S1).

Dietary sucrose worsens acute CDI

Mice consuming the HSD exhibited significantly greater weight loss on days 4–6 post-infection (p = 0.0394, p < 0.001, and p < 0.001, respectively; Figure 2E). There was also a delay in maximal weight loss between the groups. Peak weight loss occurred later and was more severe in HSD mice (11.6% ± 5.5% on day 5) than in the controls (7.8% ± 4.3% on day 2; p = 0.035). Clinical scores were significantly elevated in the HSD group from days 4–8 (p < 0.001, <0.001, <0.001, <0.001, 0.024, days 4–8, respectively; Wilcoxon test with Holm correction for multiple comparisons, Figure 2H). As with weight loss, the clinical scores peaked later and were more severe in the HSD group (3 ± 1.33 on day 2 vs. 4.78 ± 1.4 on day 5; p = 0.001), and the cumulative disease burden, as measured by the area under the curve (AUC), was also significantly higher (p < 0.001, Wilcoxon test of AUC). Antibiotic-treated but uninfected mice did not develop signs of disease (Figure S2A-B).

Fecal toxin levels were significantly higher in HSD mice on days 1, 3, and 7 post-infection (p < 0.001, p < 0.001, and p = 0.004, respectively; Figure 2I), likely contributing to the observed increase in disease severity. Luminal lipocalin−2 levels were elevated at baseline and remained significantly higher in HSD mice throughout the acute phase of infection (Figure 2J). By day 33 post-infection, inflammation had resolved in both groups; however, basal lipocalin-2 levels remained elevated in the HSD group (p < 0.01), suggesting persistent low-grade inflammation. Collectively, these results demonstrate that a HSD enhances the severity of acute C. difficile infection.

Switching diet does not rescue the sustained-carriage phenotype but induces outgrowth in previously cleared mice

To determine whether the persistent C. difficile carriage observed in HSD mice could be reversed, we performed a dietary crossover at 33 d post-infection. Mice previously fed a HSD and switched to control chow maintained a substantial C. difficile burden (4.8 × 10⁶ ± 4.6 × 10⁶ CFU/g) throughout the 39-d diet switch period (Figure 2C). Similarly, mice that remained on the HSD continued to harbor high bacterial loads (3.1 × 10⁸ ± 1.7 × 10⁸ CFU/g).

All mice fed the control diet throughout the initial infection cleared C. difficile to below the LOD by day 18 post-infection (15 d prior to switching diets). Notably, when these mice were switched to a HSD on day 33, C. difficile re-emerged above the limit of detection within five days and persisted at an average burden of 1.0 × 10⁵ ± 1.3 × 10⁵ CFU/g stool throughout the period (Figure 2C). This resurgence occurred without significant weight loss (Figure 2F). These findings suggest that a HSD creates a permissive gut environment that supports C. difficile expansion and colonization without additional antibiotic perturbment.

Diet affects susceptibility to CDI relapse

Recurrent CDI (rCDI) occurs in 10%–20% of patients globally.⁴³ To assess the impact of diet on recurrent C. difficile infection, mice were administered a second intraperitoneal dose of clindamycin 72 d after the initial infection. Mice that had consumed a HSD at any point during the experiment exhibited a significant increase in C. difficile burden following antibiotic re-exposure (Figure 2D). In contrast, mice maintained exclusively on the control diet remained below the limit of detection and showed no signs of relapse.

Although weight loss did not differ significantly between the groups during relapse (Figure 2G), the clinical scores were significantly elevated in mice that transitioned from a control to a HSD (p = 0.036, Figure 2K). These findings indicate that prior or ongoing consumption of a HSD sensitizes the host to relapse, even in the absence of overt weight loss, and that dietary context plays a role in determining the outcome of antibiotic re-exposure.

Dietary fiber is insufficient to overcome the effects of a high-sucrose diet

Increases in short chain fatty acids (SCFA), whether through consumption of soluble dietary fiber or exogenous administration of butyrate, have been shown to attenuate CDI in mouse models.^13,14,44,45 To test whether dietary fiber could mitigate the adverse effects of the HSD, mice were supplemented with either inulin or polydextrose (5 mg ml⁻¹ ad libitum in drinking water) and subjected to the CDI model. Despite fiber supplementation, mice fed the HSD exhibited significantly greater weight loss (p < 0.05, Figure 3A), higher clinical scores (p < 0.05, Figure 3B), and sustained C. difficile burden compared to control-fed mice (Figure 3C). No significant differences were observed between the HSD and HSD + fiber-supplemented groups.

Figure 3.

Fiber supplementation does not rescue the high-sucrose phenotype despite significantly reducing C. difficile toxin burden. (A) Effects of diet on body weight changes during infection. Mice fed a HSD experienced significantly more bodyweight loss compared to control on days 4−7 and 12 post infection regardless of fiber supplementation. (B) A HSD resulted in significantly higher clinical scores compared to Control regardless of fiber supplementation on days 3−7 post infection and had a significantly higher overall clinical burden as determined by AUC of clinical scores. (C) Fiber supplementation does not promote C. difficile clearance. All mice on the control diet cleared C. difficile to below the limit of detection by day 16. In contrast, all HSD mice maintained a high C. difficile burden throughout the experiment, regardless of fiber supplementation. (D) Supplementation with either inulin or polydextrose correlates with a reduction in C. difficile toxin levels compared to HSD alone. Boxplots show the medians (middle line) and interquartile ranges of 3−4 biological replicates. Dashed and shaded gray line indicates the mean and 95% CI of 4 uninfected negative controls. Individual points are biological replicates. Panel A–C: Ο = male, ∆ = female mice. Asterisks indicate statistical significance calculated using a linear mixed-effects model followed by pairwise comparisons with a Benjamini–Hochberg adjustment (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

We next assessed whether soluble fiber altered luminal toxin levels. Toxin concentrations were not significantly different between groups at day 1 post-infection. However, by day 3, all HSD groups exhibited significantly elevated toxin levels, with both inulin and polydextrose supplementation resulting in higher toxin levels than HSD alone (p ≤ 0.05; Figure 3D). At later time points (days 7 and 14), both fiber-supplemented groups showed significantly reduced toxin levels compared to unsupplemented HSD mice (p ≤ 0.05; Figure 3D). However, toxin levels remained significantly higher than control in all HSD groups at day 14 despite the resolution of clinical disease.

A high-sucrose diet significantly alters the gut microbiome

Colonization resistance provided by a eubiotic microbiota is the primary barrier to CDI. As the host diet is a key factor that defines the composition of the gut microbiome, we monitored changes due to diet and infection using 16S rDNA.

A high-sucrose diet significantly alters gut microbial diversity

To assess the impact of diet on microbial diversity, we measured the bacterial abundance and composition before and after dietary intervention. There was no significant difference in Shannon alpha diversity at baseline (p = 0.625). Mice fed a HSD exhibited a significant reduction in alpha diversity throughout the diet only and infection periods (Figures 4A and S3). Similar trends were observed in mock-infected controls, which also showed significant reductions in microbial richness (Figure S4). Absolute bacterial abundance was lower in HSD-fed mice after 2 weeks, although this difference did not reach statistical significance (1.1 × 10⁹ ± 3.4 × 10⁸ vs. 5.8 × 10⁸ ± 1.7 × 10⁸, p = 0.144, paired t-test).

Figure 4.

A high-sucrose diet significantly alters microbial diversity. (A) Alpha diversity across diet and experimental timeline. Baseline = −14, Diet only = −7 & −1, day of infection = 0, post infection/recovery = 5, 14, 33. Mice received a single dose of clindamycin on day −1, resulting in reduced alpha diversity on day 0 (day of infection). A subset of HSD mice also received either inulin or polydextrose in drinking water (5 mg/ml). (B) Effect of diet switch on alpha diversity. On day 33 post infection, a subgroup of mice were randomly switched to the opposite diet. Points represent individual samples. T-tests with Holm correction for multiple comparisons; 0.001 < p ≤ 0.01 (**), 0.01 < p ≤ 0.05 (*). (C–E) PCoA of Bray–Curtis dissimilarities of C) diet only. (D) day of infection, and (E) post infection/recovery. Permutational multivariate analysis of variance (PERMANOVA) was used with 9999 permutations to calculate significant differences between groups. (F–H) Circular compositional bar plots sorted by ordination angle showing the top 10 genera during (F) diet only, (G) day of infection, and (H) post infection/recovery. The outer rings indicate the diet groups.

Dietary switching at 33 DPI partially reversed these effects. Mice transitioned from control to HSD chow showed a significant reduction in alpha diversity by 60 DPI (p ≤ 0.05, Figure 4B), whereas those switched from HSD to control chow exhibited a partial recovery (p ≤ 0.05, Figure 4B). However, as previously shown, this recovery in diversity was insufficient to reduce the C. difficile burden (Figure 2C).

Beta diversity analysis using Bray–Curtis dissimilarities and principal coordinates analysis (PCoA) revealed consistent separation between dietary groups across all time points, in both infected and mock-infected mice (Figures 4C–H, S5, and S6). Permutational multivariate analysis of variance (PERMANOVA) confirmed that a HSD was a significant determinant of community structure (p < 0.001, all time points), indicating that consuming a HSD induced a distinct and persistent shift in gut microbiota composition. Disaggregation of samples was observed within groups following antibiotic administration (Figure 4D). Further analysis showed that this separation was due to samples from one of the three independent experiments clustering together.

A high-sucrose diet depletes protective taxa and enriches potential facilitators of C. difficile

Given the absence of a universally accepted method for differential abundance testing in microbiome datasets, we employed a multi-method approach to identify taxa significantly altered by diet. Specifically, we applied four independent statistical frameworks: Wilcox test with FDR control for multiple comparisons, Microbiome Multivariable Associations with Linear Models (MaAsLin 2), count regression for correlated observations with the beta-binomial distribution (Corncob), and Limma-Voom with TMM normalization. Genera were considered significantly altered only if identified by all four methods.

Using this stringent consensus approach, we identified 21 genera that were significantly altered by the HSD prior to infection (Figures 5A, S5−6, and Table S2). HSD-fed mice exhibited significant reductions in genera associated with colonization resistance,^46–50 including Lactobacillus (16.2 ± 8.0 vs. 5.4% ± 4.8% abundance), Alistipes (8.0 ± 2.5 vs. 0.1% ± 0.2%), Muribaculaceae (12.2 ± 5.3 vs. 1.1% ± 0.5%), Bifidobacterium (0.5 ± 0.7 vs. 0.05% ± 0.1%), and Turicibacter (8.0 ± 3.4 vs. 0.01% ± 0.02%). In contrast, Akkermansia abundance experienced a substantial and significant increase (0.7 ± 1.1 vs. 27.5% ± 8.6%).

Figure 5.

High-sucrose diet mice had differentially abundant genera. Taxonomic association tree plots of differentially abundant genera. Trees show the strength, direction, and statistical significance of the independent metadata-to-microbe associations modeled at all ranks from phylum (innermost ring) to genera (outer-most ring). The color intensity of each node represents the regression coefficient for the association between the HSD and microbial taxa. The node size indicates the prevalence of that taxon. Nodes overlaid with *(p < 0.001), × (p < 0.01) and Ο (p < 0.05) represent statistically significant associations after BH correction. Data were parsed into (A) diet only, DPI = −7, −1, (B) day of infection, DPI = 0, and (C) post-infection/recovery, DPI = 5, 14, and 33). Venn diagrams indicate the number of statistically different genera between control and HSD mice. Statistical analyzes were performed using the two-sample Wilcoxon test, Limma-Voom with TMM normalization, count regression for correlated observations with the beta-binomial (Corncob), and Microbiome Multivariable Associations with Linear Models (MaAsLin2). The covariates DPI and experiment were controlled for where possible.

Following clindamycin treatment, the number of differentially abundant genera was reduced (n = 7, Figures 5B, S7−8, and Table S2), suggesting that antibiotic exposure temporarily homogenized the microbial communities. However, during the post-infection recovery period, 19 genera were significantly altered in HSD mice (Figure 5C, Figure S9−10, and Table S2). Many of the pre-infection changes persisted, including reductions in Alistipes (10.94 ± 7.63 vs. 0.18% ± 0.66%), Muribaculaceae (7.49 ± 10.97 vs. 0.04% ± 0.14%), and Turicibacter spp. (7.71 ± 6.52 vs. 0.12 ± 0.32), and sustained enrichment of Akkermansia spp. (9.06 ± 8.33 vs. 36.86% ± 13.91%).

Additionally, Lachnospiraceae was significantly depleted in HSD mice post-antibiotic treatment (DPI = 0, 12.01 ± 17.11 vs. 0.83% ± 0.90%), and this depletion persisted post-infection (27.56 ± 15.70 vs. 5.40% ± 5.04%). Enterococcus was highly enriched in HSD mice following infection compared to control mice (0.06 ± 0.09 vs. 11.72% ± 15.01% abundance).

To delineate the effects of diet and antibiotics from those of CDI, we profiled uninfected mice under the same conditions (Figures S12-S13). Only HSD mice exhibited an Enterococcus bloom following antibiotic treatment (0.01 ± 0.01 vs. 15.92% ± 0.61%, p < 0.05, Figure S12), indicating that both a HSD and antibiotics are required to drive this expansion.

A high-sucrose diet creates a distinct gut metabolome

To investigate how a HSD alters the gut metabolic environment, we performed untargeted metabolomic profiling of fecal samples collected pre- and post-infection. Principal component analysis revealed significant differences in metabolite composition between the control and HSD mice at both time points (pre-infection p = 0.031, post-infection p = 0.026; Figure 6A), indicating that consumption of a HSD induces a distinct and persistent shift in the gut metabolome. Furthermore, while the metabolic profile of control-fed mice returned to baseline following infection, HSD mice had a distinct metabolome following infection compared to the baseline (p = 0.029), potentially due to persistent C. difficile colonization.

Figure 6.

A High-sucrose diet significantly alters the gut metabolome. Fecal pellets were obtained from mice at pre- (DPI = −7) and post-infection (DPI = 14 & 21) time points. (A) Principal component analysis of metabolites. Ellipses represent 95% confidence interval of Barrycenter. (B–H) Box plots of significantly changed individual metabolites. T-tests with control for multiple comparisons were used to detect significant differences between the groups.

A high-sucrose diet alters amino acid metabolism in a manner that supports C. difficile growth

Stickland metabolism, particularly proline reduction, is a key energy-generating pathway in C. difficile. Prior to infection, control-fed mice exhibited significantly higher levels of 5-aminovaleric acid (5-AV), a byproduct of proline Stickland fermentation (Log2FC = 3.08, p = 0.048, Figure 6B, left). Sequencing data revealed a significantly higher abundance of the Stickland fermenter Clostridium sensu stricto in the control-fed mice (0.7 ± 0.6 vs. 0.03 ± 0.04%, p < 0.001, Figure 5A). Following infection, 5-AV levels were significantly elevated in HSD mice compared to both control-fed mice (Log2FC = 3.81, p = 0.04) and their own pre-infection levels (Log2FC = 6.16, p = 0.004, Figure 6B, right).

HSD mice had elevated levels of arginine (Log2FC = 1.48, p = 0.005, Figure 6C) and tryptophan (Log2FC = 1.41, p = 0.023, Figure 6D). Metabolism of tryptophan in the gut occurs via one of three pathways. The Kynurenine and Serotonin pathways represent the host-derived pathways for tryptophan metabolism, while members of the gut microbiota metabolize free tryptophan into various indole derivatives. Kynurenic acid, a host-derived tryptophan metabolite involved in immune regulation, was significantly reduced in HSD mice after infection (Log2FC = −9.13, p < 0.001, Figure 6E). 5-Hydroxyindole-3-acetic acid (5-IAA), a metabolite of the host-derived serotonin pathway, was significantly reduced in HSD mice both pre- and post-infection (Log2FC = −4.73, p = 0.003 and Log2FC = −7.02, p = 0.006, respectively, Figure 6F). In contrast, two microbial-derived indole derivatives of the tryptophan metabolism pathway, indolacrylic acid (IA) and indole−3-acetic acid (IAA), were significantly enriched in HSD mice post-infection (Log2FC = 1.43, p = 0.023, and Log2FC = 4.55, p < 0.001, respectively; Figure 6G–H).

Together, these findings demonstrate that a HSD reshapes the gut metabolome in a manner that may favor C. difficile.

High-sucrose diet mice are susceptible to C. difficile spore germination and carriage without prior antibiotic exposure

Given the persistent colonization observed in HSD mice, we hypothesized that a HSD may create a gut environment permissive to C. difficile germination and colonization even in the absence of antibiotic-induced dysbiosis.

To test this, we performed an ex vivo germination assay using sterile cecal supernatants from control and HSD mice that had not received prior antibiotics (Figure 7A). Significantly more C. difficile spores germinated in the presence of supernatants from HSD mice than in those from control mice (p = 0.009; Figure 7B), indicating that HSD-induced changes in the gut metabolome promote spore germination.

Figure 7.

High-sucrose diet mice become susceptible to C. difficile spore germination and asymptomatic carriage. (A) Schematic of ex vivo germination and outgrowth experiment. (B) A significant increase in germination was observed when C. difficile spores were incubated with cecal contents from HSD fed mice. (C) Schematic of asymptomatic carriage experiment. (D–I) Mice were fed a Control diet, HSD or Control + Sucrose water and challenged with 10⁵ spores without antibiotics (green). At 11 DPI mice were administered a single dose of Clindamycin (10 mg kg⁻¹). (D–E) Mice consuming sucrose were detectably colonized by C. difficile without the use of antibiotics. Following antibiotic treatment (red), significant C. difficile outgrowth was observed in only in mice consuming high-sucrose diets. (F–G) No weight loss was observed in either diet group prior to antibiotic treatment (green, weights normalized to 0 DPI). Significant weight loss was observed in HSD and sucrose water mice following antibiotic treatment (red, weights normalized to 11 DPI). (H–I) Clinical scores were significantly elevated in high-sucrose fed mice following antibiotic treatment. Panel D-I: Ο = male, ∆ = female mice Error bars are 95% CI. Non overlapping error bars indicate significant differences between groups confirmed using a linear mixed-effects models followed by pairwise comparisons with a Benjamini-Hochberg adjustment.

As primary bile acids are strong spore germination factors, we examined the metabolomics data for changes in bile acid composition. Although several bile acid species were detected, including the secondary bile acid derivative 7-Ketodeoxycholic Acid (elevated in HSD mice, Log2FC = 2.05, p = 0.038) the extraction protocol used was not optimized for bile acid recovery. Therefore, while these findings suggest potential shifts in bile acid metabolism, they were not used as the basis for mechanistic interpretation.

Next, we evaluated whether these effects were translated in vivo. Mice were fed either a control, HSD, or a control diet supplemented with sucrose in drinking water (110 mg/mL, mimicking sugar-sweetened beverage consumption) for 19 d and then challenged with 10⁸ C. difficile spores without prior antibiotic treatment (Figure 7C). In contrast to the HSD mice, mice supplemented with sucrose in the drinking water had a significant increase in the absolute bacterial abundance following 2 weeks of supplementation (6.75 × 10⁸ ± 1.65 × 10⁸ vs. 1.32 × 10⁹ ± 2.25 × 10⁸, p < 0.01). Mice on the HSD became colonized and maintained detectable C. difficile burdens for approximately 1-week post-infection (Figure 7D, green shaded), despite the absence of antibiotics. Mice receiving sucrose in water exhibited transient colonization, whereas control mice remained uncolonized (Figure 7E, green shaded). No clinical symptoms or weight loss were observed in any group during this period (Figure 7F–I, green shaded).

On day 11 post-infection, all mice received a single dose of clindamycin. This triggered rapid C. difficile outgrowth in both HSD and sucrose water groups but not in the control-fed mice (Figure 7D–E, red shaded). Outgrowth was accompanied by significant weight loss and elevated clinical scores in the high-sucrose groups (Figure 7F–I, red shaded), confirming the progression from asymptomatic carriage to symptomatic CDI. Together, these studies show a specific effect of sucrose, as opposed to contributions from other dietary components.

To investigate the specific effects of sucrose consumption on the microbiota, we sequenced stool samples from mice before and after 14 d of ad libitum sucrose water consumption. Despite consuming a high-fiber, nutritionally complete control diet, mice consuming sucrose water exhibited many of the microbiota changes observed in HSD mice (Figure S16). Sucrose consumption alone resulted in significant reductions in Alistipes (1.70 ± 0.11 vs. 1.01% ± 0.34%, p = 0.033), Lactobacillus (7.48 ± 2.05 vs. 2.51% ± 0.64%, p = 0.019), Clostridium (8.56 ± 0.99 vs. 0.58% ± 0.22%, p < 0.001), and Turicibacter (6.47 ± 0.53 vs. 0.24% ± 0.10%, p < 0.001) abundance, with a corresponding increase in Akkermansia (1.83 ± 0.88 vs. 4.57% ± 1.54%, p = 0.040). We also observed a large decrease in Bacteroidales (37.43 ± 2.07 vs. 25.21% ± 2.85%, p = 0.001) and a significant increase in Allobaculum (14.41 ± 1.48 vs. 35.93% ± 5.54%, p = 0.005), which, like Akkermansia, is an intestinal mucin degrader.⁵¹

These findings demonstrate that a diet high in sucrose is sufficient to permit C. difficile colonization in the absence of antibiotics and that this colonization can transition to fulminant disease upon antibiotic exposure.

Discussion

The ability of C. difficile to invade the host microbiota and cause disease is highly dependent on colonization resistance, the collective capacity of the resident gut microbiota to prevent pathogen establishment and expansion. Colonization resistance is maintained through a combination of microbial competition for nutrients and niche space, production of inhibitory metabolites such as short-chain fatty acids and secondary bile acids, maintenance of epithelial barrier integrity, and modulation of host immune responses.^16,52–54 Disruption of these protective mechanisms creates a permissive niche for C. difficile colonization and toxin-mediated disease. While broad-spectrum antibiotic use is the most well-characterized and potent disruptor of colonization resistance, other factors, including dietary perturbations, can also compromise this barrier in ways that alter the susceptibility to enteric pathogens.^55–58

In human populations, microbiomes of asymptomatically colonized individuals are enriched in sucrose degradation pathways, and increased carbon metabolism genes have been linked to recurrent CDI,^59,60 potentially reflecting higher dietary sugar intake in these groups.

Studies examining the impact of refined carbohydrates on CDI have yielded conflicting results. One study reported that mice fed a carbohydrate-rich diet developed only mild disease and were protected relative to those on a high-fat/high-protein diet.⁹ Interestingly, this high-carbohydrate diet was associated with persistent colonization, a phenotype we also observed. Other studies have similarly reported prolonged dysbiosis and delayed disease onset in the context of high-carbohydrate feeding.⁶¹ Hryckowian et al. (2018) demonstrated that diets lacking soluble fiber but enriched in simple sugars promoted sustained C. difficile carriage, a phenotype reversible by inulin supplementation.¹³ This model more closely parallels our findings, though supplementation with either inulin or polydextrose did not restore colonization resistance in our hands. These discrepancies may reflect differences in animal models, microbial communities, or specific dietary constituents.

We observed a significant effect of the HSD on C. difficile pathogenesis, despite sucrose’s metabolic inaccessibility to C. difficile (Summary, Figure 8). One caveat is that mice, like humans, possess a brush border sucrase which can cleave the glycosidic bond in sucrose releasing glucose and fructose at the intestinal brush border, enabling absorption of the resulting monosaccharides across the enterocyte membrane and their metabolism by other microbes. In a conventional CDI model, HSD mice exhibit worsened disease severity, characterized by increased weight loss, elevated clinical scores, higher toxin burden, and prolonged inflammation. Unlike control-fed mice, which cleared C. difficile within 18 d, HSD mice maintained high bacterial burdens for over 70 d. Notably, this phenotype was not reversed by switching to a control diet. Conversely, switching from a control to a HSD was sufficient to induce C. difficile outgrowth in the absence of antibiotics. To our knowledge, this is the first report of diet-induced C. difficile expansion, a finding that may have important implications for understanding and managing recurrent CDI.

Figure 8.

Summary of key findings.

Groups fed a HSD exhibited significantly higher toxin levels than controls, even after the resolution of clinical disease. Several factors may explain this paradoxical outcome, including the adaptive immune response and the production of metabolites that interfere with toxins. Antibodies IgA and IgG can mitigate toxin-induced damage by neutralizing toxins both locally at the mucosal intestinal surface and more broadly.⁶² However, 14 d post-infection might be too soon for a substantial production of specific antibodies. Bile acids can also directly bind and neutralize the TcdB toxin, with secondary bile acids being more potent than their primary bile acid precursors.⁶³ Additionally, while the SCFA butyrate can enhance initial toxin production, it also promotes epithelial barrier integrity and immune modulation, which can indirectly suppress toxin-mediated damage.⁶⁴ Data from HSD mice supplemented with soluble fiber support this hypothesis. Although supplementation initially increased toxin production, it was reduced at later time points, possibly due to fermentation into SCFAs, including butyrate. Despite the decrease in toxin levels, fiber supplementation did not restore colonization resistance or reduce the C. difficile burden in the context of a HSD. Further investigation is needed to clarify the relative contributions of toxin load and clinical disease symptoms in this context.

The HSD remodeled the gut microbiota and metabolome in ways that appear to favor C. difficile persistence. Consistent with prior studies, a HSD led to a reduction in microbial diversity and depletion of protective taxa such as Lactobacillus, Bifidobacterium, Alistipes, and Muribaculaceae.^{46–49,65–67}These taxa inhibit C. difficile through competitive exclusion, bacteriocin and SCFA production, and nutrient competition.^{46–49,65–67} Following antibiotic administration, and throughout the active disease and recovery periods, Turicibacter and Lachnospiraceae were significantly reduced in HSD mice. A reduction in Turicibacter abundance has been correlated with increased CDI severity in mouse models.⁵⁰ In human patients, successful fecal microbial transplantation (FMT) is associated with enrichment of Lachnospiraceae, whereas CDI relapse is linked to its depletion.^68,69 Moreover, monocolonization with Lachnospiraceae confers protection against CDI in germ-free mice.^70,71

Simultaneously, HSD mice exhibited enrichment of Akkermansia and Enterococcus, both frequently elevated in human CDI cases.^72,73 These taxa are capable of cross-feeding C. difficile via mucin degradation and fermentable amino acids, respectively.^74–78 The observed enrichment of C. difficile cross-feeders, along with the depletion of protective taxa, mirrors the microbiome signatures reported in human CDI cases and may underlie the prolonged colonization and asymptomatic carriage observed in HSD mice.

Metabolomic profiling further supports the establishment of a niche favorable to C. difficile. Prior to infection, HSD mice had reduced levels of 5-aminovaleric acid (5-AV), a byproduct of proline Stickland fermentation, potentially indicating a loss of commensals that compete with C. difficile for proline. In support of this hypothesis we observed the loss of Clostridium sensu stricto, a genus of known Stickland fermenters, in HSD mice. Following infection, 5-AV levels increased significantly in the HSD group, consistent with proline metabolism by C. difficile, which preferentially utilizes proline as an energy source.^79,80 Additionally, levels of arginine, a known co-germinant for C. difficile spores,⁸¹ and tryptophan, an essential nutrient for C. difficile,⁸² were elevated in HSD mice post-infection but not prior, suggesting these changes are driven by host and microbial metabolic responses rather than dietary content.

Tryptophan can be metabolized by the host through the kynurenine and serotonin pathways, the latter primarily occurring via enterochromaffin cells in the gut epithelium. In HSD mice, levels of 5-hydroxyindoleacetic acid (5-HIAA), a serotonin catabolite, were significantly reduced both pre- and post-infection. Concurrently, kynurenic acid, a host-derived immunomodulatory metabolite,⁸³ was significantly decreased following infection, indicating a dysregulation of the host tryptophan pathways. This dysregulation may account for the elevated tryptophan levels. Conversely, microbial indole pathways appeared to be upregulated, with indole derivatives such as indole-3-acetic acid (IAA) and 3-indoleacrylic acid (IA) exhibiting increased levels. Indoles accumulate in the stool of patients with C. difficile infection and may facilitate C. difficile persistence by inhibiting the growth of protective commensals.⁸⁴ The balance between host- and microbiota-driven tryptophan metabolism could have functional implications in the context of CDI, as mice deficient in kynurenic acid production exhibit exacerbated CDI symptoms.⁸⁵ Collectively, these shifts may impair host immune regulation and promote microbial interactions that favor C. difficile colonization and persistence in the gut.

Incubating spores with sterile cecal supernatant from HSD mice effectively induced robust germination, and mice fed either the HSD or supplementary sucrose water became colonized without prior antibiotic exposure. While recent studies in animal and in vitro models have shown that bile acids are not essential for CDI,^79,86 primary bile acids remain potent factors for spore germination. Our metabolite extraction focused on targeting amino acids and sugars rather than bile acids, and thus, it did not capture significant changes in these biomolecules. However, dietary changes may influence the production of bile acids and shifts in the microbiota can alter the composition of primary and secondary bile acid pools.

Dietary sugar intake remains high across Western populations, raising important questions about its role in shaping gut microbial ecology and susceptibility to enteric pathogens. In this study, we demonstrate that a HSD profoundly alters host susceptibility to C. difficile, impairing colonization resistance and enabling both prolonged colonization and re-emergence of infection in a conventional model. Notably, sucrose supplementation in drinking water, at concentrations comparable to those found in sugar-sweetened beverages, was sufficient to promote asymptomatic carriage and progression to disease following antibiotic exposure, even in the context of a fiber-rich, nutritionally complete diet. This model recapitulates the clinical trajectory from asymptomatic carriage to symptomatic infection and, to our knowledge, represents the first demonstration of diet-induced C. difficile colonization in the absence of prior antibiotic exposure. As such, it provides a valuable platform for mechanistic studies of colonization dynamics and disease progression.

Several limitations should be considered. The high-sucrose chow used in our primary model contains levels of refined sugar that exceed typical human consumption, which may limit direct translational relevance. Additionally, the control and test diets differ in multiple components beyond sucrose content, complicating direct comparisons. Although the use of a defined control diet might seem preferable, most commonly used formulations (e.g., AIN-93M) are themselves enriched in simple carbohydrates and lack soluble fiber, rendering them unsuitable comparators for studies of refined sugar intake. Furthermore, while we observed persistently elevated fecal lipocalin-2 levels in HSD mice, the immunological mechanisms underlying increased disease severity remain undefined. Finally, this study utilized a single C. difficile strain; future work should expand to include strains representing the full metabolic and virulence diversity of C. difficile to better understand how dietary factors interact with pathogen heterogeneity.

Despite these limitations, our findings establish a novel model of diet-induced C. difficile carriage and underscore the potential for dietary modulation to influence colonization resistance. These insights provide a foundation for future mechanistic studies and support the integration of dietary strategies into CDI prevention and management frameworks.

Supplementary Material

Supplementary material

Figure S1: C. difficile carriage by morphotype. Control and HSD mice were monitored for C. difficile burden for 90 DPI. Both vegetative cells (■, solid lines) and spores (●, dashed lines) were measured during (A) acute infection, (B) diet switch, and C) Relapse. Points represent biological replicates. Lines represent means.Figure S2: Uninfected mice do not develop signs of disease. Mice were fed either a control (n = 4) or HSD (n = 4) diet for 2 weeks, then given a single dose of clindamycin (10 mg kg⁻¹) but not challenged with C. difficile. (A) Weights and (B) clinical scores were monitored and mice did not develop disease in the absence of C. difficile challenge, regardless of diet. Lines represent means.Figure S3. Mice fed a high-sucrose diet showed significantly altered microbial evenness and richness. All mice were initially fed a control diet (blue) (n = 4−22 per group per day). On day −14 half of the mice were switched to a HSD (orange) (n = 4−22 per group per day). At the time of the diet switch, there was no difference in alpha diversity between the groups. Mice received a single dose of clindamycin on day −1, resulting in reduced alpha diversity observed on day 0 (day of infection). Post-infection, alpha diversity increased in both groups but remained significantly lower in the HSD group. A subset of mice on the HSD also received either inulin (green, n = 3−4 per day) or polydextrose (purple, n = 3−4 per day) in drinking water (5 mg/ml), which did not significantly affect alpha-diversity compared to HSD alone. (A) Observed sequence richness is shown. (B) Chao 1 estimator – richness. (C) Inverse Simpson – diversity. Welch's T-test with Holm correction for multiple comparisons. Statistical significance is indicated as follows: p < 0.0001 (****), 0.0001 < p ≤ 0.001 (***), 0.001 < p ≤ 0.01 (**), 0.01 < p ≤ 0.05 (*).Figure S4. Uninfected control mice fed a HSD showed significantly altered microbial richness. All mice were initially fed a control diet (blue) (n = 4 per group per day). On day −14 half of the mice were switched to a HSD (orange) (n = 4 per group per day). At the time of the diet switch, there was no difference in alpha diversity between the groups. Mice received a single dose of clindamycin on day −1, and were mock infected on day 0 with sterile water. (A) Observed sequences. (B) Chao 1 estimator. (C) Inverse Simpson Diversity. (D) Shannon Diversity. Welch’s T-test with Holm correction for multiple comparisons. Statistical significance is indicated as follows: p ≤ 0.001 (***), 0.001 < p ≤ 0.01 (**), 0.01 < p ≤ 0.05 (*).Figure S5. Microbiomes of HSD mice rapidly diverge from control-fed mice and remain distinct. PCoA of Bray Curtis dissimilarities of all 16S sequencing. Ellipse are 95% CI of group centroid, individual points indicate samples, shaded lines indicate direction centroids between time points.Figure S6. Microbiomes of mock infected HSD mice rapidly diverge from mock infected control-fed mice and remain distinct. PCoA of Bray Curtis dissimilarities of all 16S sequencing (n = 4 per group and timepoint). Ellipse are 95% CI of group centroid, individual points indicate samples, shaded lines indicate direction centroids between time points.Figure S7. Heatmap of taxa in pre-infection (diet-only, DPI −7 and −1) samples. OTUs were aggregated at the genus level, and abundances were standardized using the robust centered log ratio (CLR) transformation from the Vegan R package. Columns represent individual mice on different days (n = 8−10 per group per day).Figure S8. Significantly changed genera from pre-infection (DPI = −1 and −7) timepoints. Percentage abundance of significantly differentially expressed genera. All genera were found to be significantly different (p < 0.05) in all statistical models after controlling for multiple comparisons. Triangles = DPI −1 samples and circles = DPI −7 samples (n = 8−10 per group per day).Figure S9. Heatmap of taxa on the day of infection (post-antibiotic, DPI = 0) samples. OTUs were aggregated at the genus level, and abundances were standardized using the robust centered log ratio (CLR) transformation from the Vegan R package. Columns represent individual mice on different days (n = 20−22 per group).Figure S10. Significantly changed genera from day of infection (post-antibiotic, DPI = 0). Percentage abundance of significantly differentially expressed genera. All genera were found to be significantly different (p < 0.05) in all statistical models after controlling for multiple comparisons (n = 20−22 per group).Figure S11. Heatmap of taxa post infection (DPI = 5, 14, and 33). OTUs were aggregated at the genus level and abundances standardized with the robust centered log ratio (CLR) transformation from the Vegan R package. OTUs were aggregated at the genus level and abundances standardized with the robust centered log ratio (CLR) transformation from the Vegan R package. Columns represent individual mice on different days (n = 6−11 per group per day).Figure S12. Significantly changed genera from post-infection (DPI = 5, 14, 33) timepoints. Significantly changed genera from post-infection (DPI = 5, 14, 33) timepoints. Percent abundance of significantly differential genera. All genera were found to be significantly different (p < 0.05) in all statistical models after controlling for multiple comparisons. Triangles = DPI 33, squares = DPI 14, and circles = DPI 5 (n = 6−11 per group per day).Figure S13: Significantly changed genera from day of antibiotic treatment (DPI = −1) in uninfected control mice. Significantly changed genera from post-dietary-intervention (DPI = −1) timepoint. Percent abundance of significantly differential genera. All genera were found to be significantly different (p < 0.05) by Wilcoxon test after controlling for multiple comparisons. n = 4 per group.Figure S14: Significantly changed genera from following antibiotic treatment and mock infection (DPI = 5) in uninfected control mice. Significantly changed genera following antibiotic treatment and mock infection (DPI = 5). Percent abundance of significantly differential genera. All genera were found to be significantly different (p < 0.05) by Wilcoxon test after controlling for multiple comparisons. n = 4 per group.Figure S15: Significantly changed genera following recovery from antibiotic treatment (DPI = 14) in uninfected control mice. Significantly changed genera following recovery from antibiotic treatment (DPI = 14). Percent abundance of significantly differential genera. All genera were found to be significantly different (p < 0.05) by Wilcoxon test after controlling for multiple comparisons. n = 4 per group.Figure S16: Significantly changed genera following consumption of sucrose drinking water ad libitum for 14 d. Percent abundance of significantly differential genera. All genera were found to be significantly different (p < 0.05) by Wilcoxon test after controlling for multiple comparisons. n = 4 per group.Table S1. Reagents and Resources used.Table S2. Differential abundance of genera. Statistically different genera between control and HSD mice (FDR < 0.05 after BH correction for multiple hypothesis testing) are shown as ● (Diet only, no antibiotics), ● (Day of infection, 24 h post clindamycin IP), and ● (recovery post infection). Statistical analyses were performed using the two-sample Wilcoxon test, Limma-Voom with TMM normalization, Count Regression for Correlated Observations with the Beta-Binomial (Corncob), and Microbiome Multivariable Associations with Linear Models (MaAsLin2). The covariates DPI and experiment were controlled for where possible.

Funding Statement

This work was supported by the Centers of Biomedical Research Excellence (CoBRE) Grant P20GM125504 and the Jewish Heritage Fund for Excellence Research Recruitment Grant Program at the University of Louisville, School of Medicine. Sequencing was performed with the assistance of the University of Louisville Genomics Training and Education Center, which is supported by NIH P20GM103436 (KY IDeA Networks of Biomedical Research Excellence), the Brown Cancer Center, and user fees. (National Institute of General Medical Sciences) (P20GM125504)

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/19490976.2025.2566302.

Disclosure of potential conflicts of interest

No potential conflict of interest was reported by the author(s).

Acknowledgments

This work was supported by the Centers of Biomedical Research Excellence (CoBRE) Grant P20GM125504 and the Jewish Heritage Fund for Excellence Research Recruitment Grant Program at the University of Louisville, School of Medicine.

Sequencing was performed with the assistance of the University of Louisville Genomics Training and Education Center, which is supported by NIH P20GM103436 (KY IDeA Networks of Biomedical Research Excellence), the Brown Cancer Center, and user fees.

Author contributions

DE, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review and editing. MJC, Data curation, Investigation, Methodology, Writing – review, and editing. KS, Investigation, Writing – review and editing. LH, Investigation, Writing – review and editing. JC, Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing.

Data availability statement

The raw reads of all 16S amplicons were deposited in the NCBI Sequence Read Archive (SRA) with the Bioproject ID: PRJNA1274008.