Dietary fiber reduces mortality from secondary sepsis in a murine model of Clostridioides difficile infection

Madison Apgar; Elena S Wall; Monica Midura-Kiela; Charles P Neff; David J Orlicky; Laurie M Lyon; Claudio Bernardazzi; Casey G Martin; Oren M Gordon; Issa Jimenez-Espinoza; Jacob K Flynn; Catherine A Lozupone; Keith Z Hazleton

doi:10.1016/j.isci.2026.115258

. 2026 Mar 6;29(4):115258. doi: 10.1016/j.isci.2026.115258

Dietary fiber reduces mortality from secondary sepsis in a murine model of Clostridioides difficile infection

Madison Apgar ¹, Elena S Wall ¹, Monica Midura-Kiela ², Charles P Neff ³, David J Orlicky ⁴, Laurie M Lyon ¹, Claudio Bernardazzi ⁵, Casey G Martin ¹, Oren M Gordon ^1,⁶, Issa Jimenez-Espinoza ², Jacob K Flynn ¹, Catherine A Lozupone ^3,^8,^9,^∗, Keith Z Hazleton ^7,^8,^∗∗

PMCID: PMC13018909 PMID: 41907412

Summary

Despite antibiotic stewardship and improved sanitation, Clostridioides difficile infection (CDI) remains a major public health problem. Dietary modification is promising for prevention. Using a murine model of CDI, we show that mice-fed low-fiber diets with both high- and low-fat content had high mortality, but this was ameliorated with fiber supplementation. We found increased presence of intestinal pathobionts, including Escherichia-Shigella, Proteus, and Enterococcus, in blood, liver, and spleen in mice on low-fiber diets. Elevated proinflammatory cytokines and blood urea nitrogen in mice that succumbed to illness before the experimental endpoint indicate sepsis as the cause of death. Despite higher mortality, mice fed a high-fat/low-fiber diet did not have higher cecal TcdA/B toxin levels, displayed less cecal and colonic inflammation 3 days post-C. difficile inoculation, and had a delayed onset of morbidity compared to other diets. Our data suggest dietary fiber is associated with reduced CDI morbidity, potentially via mediation of intestinal barrier function and an appropriate immune response to CDI, thereby protecting against secondary sepsis.

Subject areas: Health sciences, Biological sciences, Systems biology, Diet

Graphical abstract

Highlights

•
Low-fiber diets increase mortality from sepsis secondary to C. difficile infection
•
Gut pathobionts that bloom post-antibiotics disseminate to the blood and organs
•
Dietary fat and fiber alter the kinetics of C. difficile infection progression
•
Fat and fiber interact to shape microbiome resilience after antibiotics

Health sciences; Biological sciences; Systems biology; Diet

Introduction

Clostridioides difficile infection (CDI) is responsible for 350,000 cases and 15,000 deaths per year in the United States.¹^,² The gut microbiome is protective against CDI when there is a high level of biodiversity, which is associated with general good health.³ Patients on broad-spectrum antibiotics or with diseases characterized by reduced gut microbiome diversity, such as inflammatory bowel disease, have increased CDI risk.⁴^,⁵^,⁶ A number of recent publications have demonstrated that diet can affect CDI in mice and humans. Diets with low soluble fiber or high amino acids demonstrate increased C. difficile colonization⁷^,⁸ and higher mortality,⁹ while diets high in fat have increased disease severity in mouse models of antibiotic-induced CDI.⁷^,¹⁰^,¹¹ Furthermore, in a randomized, controlled clinical trial, the prebiotic oligofructose prevented recurrent CDI.¹²

Commensal gut microbes can metabolize dietary components for the benefit of the host and themselves, which may be how diet influences CDI. For example, dietary fiber consumption can reduce the amount of gut microbiome disturbance caused by antibiotics.¹³^,¹⁴ Dietary fiber can also promote the production of short-chain fatty acids (SCFAs), including butyrate, acetate, and propionate, by commensal microbes.¹⁵ In vitro and murine model experiments have demonstrated butyrate’s ability to inhibit C. difficile growth¹⁶^,¹⁷ and protect against CDI by enhancing barrier function¹⁸ while acetate supplementation protected mice from CDI on a low-fiber diet by enhancing interleukin-22 (IL-22) production by intestinal type 3 innate lymphoid cells (ILC3s) and modulating MHC-II expression in the large intestines.¹⁹ Dietary fat influences bile acid pools by stimulating the production of primary bile acids that serve as germination factors for C. difficile spores.²⁰ Commensal gut microbes can transform primary bile acids to secondary bile acids such as deoxycholate (DCA), an inhibitor of C. difficile growth.²¹ In vivo studies have found that mice on a low-fiber diet have an increased risk of CDI associated with increased levels of primary bile acids after antibiotic treatment,²² which fits with existing literature showing that patients with recurrent CDI tend to have more primary than secondary bile acids.²³

While prior studies have investigated the effects of dietary fiber and fat independently in murine models,⁷^,⁹^,¹⁶^,¹⁷^,²² how the two components work in tandem is not well understood. Given the complexity of human dietary composition and the potential application of fiber supplementation to a baseline high-fat/low-fiber Western diet to treat microbiome-related disease, we sought to understand how fiber and fat ratios differentially affect the outcome of CDI. To accomplish this, we compared CDI outcomes in diets with high versus low dietary fat content, with and without the addition of fibers. We hypothesized that low-fat/high-fiber diets would protect against CDI-induced mortality by limiting antibiotic-induced microbiome disruption, reducing the loss of protective SCFAs, and decreasing the ratio of primary:secondary bile acids. While we observed a protective effect of low-fat/high-fiber diets, this protection was associated with an appropriate inflammatory response to CDI and a reduction of secondary disseminated bacterial infections and sepsis, rather than by directly reducing C. difficile toxin or host inflammatory response in the early phase of infection.

Results

Dietary fiber protects against mouse mortality in antibiotic-associated CDI

Previously, we found that adding dietary fat to a low fiber defined diet increased mortality in an established mouse model of antibiotic-induced CDI.¹¹^,²⁴ To expand on these findings, we sought to understand how dietary fat and fiber combined influence mortality. Six-week-old female C57BL/6 mice were placed on one of the four defined diets with varying fat and fiber content or standard mouse chow (Figure 1B). The high-fat/low-fiber “Western” diet¹¹ contains cellulose as the only fiber source with added lard, beef tallow, and milk fat, resulting in a fat content ∼2x that of chow (Table S1), as used previously.¹¹ The high-fat/low-fiber diet has similar macronutrient totals to a typical US diet²⁵ with 34.5% of calories from fat, of which ∼36% was saturated, ∼41% monounsaturated, and ∼21% polyunsaturated. As our control diet, we utilized a low-fat/low-fiber diet, identical to our previous work.¹¹ This diet is similar to the high-fat/low-fiber diet but with fat content equivalent to chow (Figure 1B); with 17.2% of the calories from fat, of which ∼19% was saturated, ∼41% monounsaturated, and ∼39% poly-unsaturated. The high fiber diets were created by adding a cocktail of soluble and insoluble fibers to both low-fiber diets. This cocktail included inulin, pectin, resistant starch (raw potato starch), and hemicellulose (psyllium husk) in a 1:1:1:1 ratio, which were chosen for their differing chemical properties and to better mimic the many fiber sources of a varied diet. Hemicellulose and cellulose are both insoluble stool bulking agents,²⁶ and hemicellulose can also be metabolized by gut microbes to butyrate.²⁷ Pectin forms a highly viscous solution, slows intestinal transit times, and can support commensal intestinal bacteria.²⁶^,²⁸ Resistant starches and inulin are soluble fibers that can be metabolized by the microbiome to SCFAs.²⁶^,²⁹ Standard mouse chow was included as well to compare the difference between a defined purified diet and standard unpurified laboratory diets used in much of the literature.

Study design, diet composition information, survival results, and weight change

(A) Study design and timeline.

(B) Macronutrient information for tested diets.

(C) Survival curves for the five diets (shaded area shows 95% CI).

(D) Log-normal pairwise survival probability comparisons between the diets with the Kaplan-Meier test. P-value significance (∗∗∗∗: p < 0.0001, ∗∗∗: p < 0.001, ∗∗: p < 0.01, and ∗: p < 0.05, ns: p ≥ 0.05).

(E) Average percent weight change of the survival cohort between the day of infection (0) and all subsequent days.

Data are represented as mean ± one standard deviation.

Mice were randomized to one of the five diets, and after one week, were treated with an antibiotic cocktail for 5 days. This was followed by an injection of clindamycin and gavage with C. difficile VPI 10463 spores (Figure 1A). In the first set of experiments (“survival cohort”; Figure 1A), mice were observed for 14 days after infection to evaluate survival. In subsequent experiments (“day 3 cohort”; Figure 1A), mice were sacrificed at three days post-CDI to collect tissue samples. Day 3 was selected because this was 1 day prior to the onset of mortality in any diet in our previously published study.¹¹

Mice on the low-fat/high-fiber diet had the highest survival among the four defined diets (Figure 1C), which was not significantly different from mice-fed chow (Figure 1D). Consistent with our previously published results,¹¹ mice fed a high-fat/low-fiber diet had poor survival compared to chow (hazard ratio = 11, p-value = 0.001; CoxPH test). Unlike our previous study, mice fed a low-fat/low-fiber diet also had poor survival compared to chow (hazard ratio = 24, p-value = <0.001; CoxPH test).¹¹ We explore the inconsistency between survival in the low-fat/low-fiber cohorts later in the article.

Defined diets with added fiber improved survival, regardless of fat content (Figure 1D). Though overall survival was improved in the high fiber cohorts, interestingly, mice fed a high-fat/low-fiber diet had a significantly later onset of mortality compared to mice fed a low-fat/high-fiber diet (p-value=<0.01; Kruskal-Wallis/Dunn’s post hoc test), with mortality occurring between days 4 and 9 in the high-fat/low-fiber group and between days 2 and 5 in the low-fat/high-fiber group (Figure 1C). Similarly, a decrease in mouse weight occurred in the same window as the onset of mortality, following the trends defined above. (Figure 1E).

Intestinal inflammation correlates with C. difficile toxin and is lowest in the high-fat/low-fiber diet at 3 days post infection

We hypothesized that the higher mortality observed in the low-fiber diets would relate to the increased production of C. difficile toxins TcdA and TcdB, and associated inflammation. However, we did not find that cecal levels of TcdA or TcdB differed significantly between diets 3 days post-infection (Figure 2A). Surprisingly, even though they had higher mortality than the high-fiber diets, mice fed the high-fat/low-fiber diet had the lowest inflammation in both the cecum and the colon based on the histologic scoring of H&E-stained tissue (Figures 2B–2D and 2E). Furthermore, while mice on a high-fat/high-fiber diet had improved survival over the low-fiber cohorts, these mice had significantly increased colonic and cecal inflammation at day 3 post infection (Figure 2B). Here, colonic and cecal inflammation are characterized by worse submucosal edema and epithelial integrity in the cecum, and higher levels of inflammation/injury and worsened epithelial repair in the colon (Figure S1). High-fiber diets, regardless of fat content, displayed elevated cecal and colonic inflammation. Additionally, increased cecal and colonic inflammation was observed in the low-fat/low-fiber diet but not in the high-fat/low-fiber diet. Despite unexpected histopathological findings between diets, there was a significantly positive correlation between cecal levels of TcdA/TcdB and cecal inflammation across diets, suggesting that C. difficile toxins contribute to the cecal inflammation observed here (Figure 2C).

Measures of *C. difficile* toxin concentration, gut inflammation, and the relationship between the two

(A) *C. difficile* TcdA and TcdB toxin concentrations in mouse cecal contents, determined by ELISA.³⁰

(B) Histologic inflammation scoring of the cecum (Barthel) and transverse colon (Dieleman) three days post *C. difficile* infection. Boxplot lines (from top to bottom) depict the 75^th, 50^th (median), and 25^th percentiles, with lines extending from the top/bottom of the boxplot indicating the largest/smallest observation within ±1.5∗IQR (inter-quartile range). P-value significance (∗∗∗∗: p < 0.0001, ∗∗∗: p < 0.001, ∗∗: p < 0.01, and ∗: p < 0.05, ns: p ≥ 0.05). Pairwise p-values were calculated via Kruskal-Wallis and Dunn’s post-hoc tests, with FDR correction for both panels a and b.

(C) Linear regression of *C. difficile* TcdA and TcdB cecal content concentration to cecal histologic inflammation scores. Linear regression models included a diet interaction term.

(D and E) Histologic imaging sections at 200x of the (D) cecum and (E) colon to show the range of low (left) to high (right) inflammation. The yellow and red brackets indicate the width of the mucosa and submucosal layers, respectively. The cecum has a swollen submucosa in the right versus left panel due to edema, indicative of higher inflammation. The blue arrow in the “cecum high inflammation” image indicates a fibrous cap sitting on top of a place in the epithelium where a breach has occurred that allowed cecal contents access to the tissue. In the colon (e), there is no inflammation in either the mucosa or submucosa layers of the low inflammation tissue. In the mucosa and submucosa of the high inflammation tissue there are abundant inflammatory cells, predominantly neutrophils. The submucosa has swollen in size to a small degree due to edema in response to the neutrophils. The scale bar in the bottom left corner of each micrograph represents 100 μm.

Cecal levels of C. difficile toxins and inflammation have a complex relationship with the microbiome-produced metabolites, DCA and butyrate, across diets

Bile acids are known to play important roles in the growth and inhibition of C. difficile,³^,²⁰^,³¹ so we investigated how bile acid concentrations may differ between diets. We predicted that bile acid pools would differ across diets because 1) dietary fat can promote higher levels of primary bile acids⁷^,¹¹ and 2) dietary fiber can promote the survival of obligate anaerobes that produce secondary bile acids in the context of antibiotic disturbance.¹³^,¹⁴ To test this, we quantified the levels of bile acids previously described to affect C. difficile growth, characterized as either growth-promoting (taurocholate (TCA), cholate (CA))²⁰^,³²^,³³ or growth-inhibiting (α-muricholate (α-MCA), β-muricholate (β-MCA), lithocholate (LCA), DCA).²⁰^,³¹^,³⁴ Inhibitory bile acids include the secondary bile acids DCA and LCA that are generated by obligate gut anaerobes.⁷^,¹¹^,¹³^,¹⁴^,³³^,³⁴ The total concentration of the four C. difficile-inhibiting bile acids was significantly lower for all the defined diets except the low-fat/high-fiber diet, compared to chow (Figure 3A), and negatively correlated with toxin production and inflammation in the cecum at 3 days post-infection (Figure S2a). We also separately evaluated DCA, since it was the most abundant bacterially produced secondary bile acid we measured. DCA levels were significantly reduced in the low-fiber diets compared to chow, and levels trended higher with fiber supplementation, but statistical power was challenged because we had limited data on cecal bile acids in the high-fiber-fed mice (Figures 3A; Table S2). Even with the challenges arising from uneven bile acid sampling, we were able to observe a significant negative correlation between DCA, TcdA/TcdB levels, and cecal inflammation (Figure 3C), supporting the previously described protective effect of DCA.²⁰^,³¹^,³⁴

Measures of bile acid pools and SCFA concentration and their relation to *C. difficile* toxin and inflammation measures

(A) Cecal content bile acid pool composition of deoxycholic acid (*DCA*, *main microbiome-produced inhibitor*), *C. difficile* growth inhibitors (*a-MCA*, *b-MCA*, *DCA*, *LCA*), and *C. difficile* growth promoters (*TCA*, CA) across diets.

(B) Ratio of *C. difficile* growth promoters:inhibitors across diets.

(C) Linear regression of cecal content DCA concentration and *C. difficile* toxin concentration/histopathology scores.

(D) Cecal content SCFA levels (acetate, butyrate, and propionate) three days post *C. difficile* infection.

(E) Linear regression of cecal content SCFA concentration and *C. difficile* toxin concentration/histopathology scores. Boxplot lines (from top to bottom) in a, b, and d depict the 75^th, 50^th (median), and 25^th percentiles, with the lines extending from the top/bottom of the boxplot indicating the largest/smallest observation within ±1.5∗IQR (inter-quartile range). Linear regression models are performed between the x and y axes, accounting for an interaction with diet. Pairwise comparisons shown in a, b, and d were calculated via Kruskal-Wallis and Dunn’s post hoc tests with p-value corrections via FDR.³⁰P-value significance (∗∗∗∗: p < 0.0001, ∗∗∗: p < 0.001, ∗∗: p < 0.01, and ∗: p < 0.05, ns: p ≥ 0.05).

Levels of C. difficile-promoting bile acids did not differ across diets (Figure 3A); however, previous studies suggest that since C. difficile-promoting and -inhibiting bile acids can interact competitively, the ratios of promoter:inhibitor bile acids may be more informative than either group alone.⁷^,¹¹ Mice fed the high-fat/low-fiber diet had the highest C. difficile promoter:inhibitor ratio (Figure 3B), indicating a more pro-C. difficile bile acid pool. However, the implications of this are unclear because the ratio did not correlate with cecal TcdA or TcdB levels (Figure S2A). Furthermore, the promoter:inhibitor ratio had a complex relationship with cecal histopathology, having the expected positive trend in the chow and high-fiber diets and the opposite trend in the low-fiber diets (Figure S2A).

We also hypothesized that dietary fiber would exert an effect by promoting SCFA production by the gut microbiome, and thus measured cecal levels of butyrate, propionate, and acetate at day 3 post-infection. We found that levels of all 3 SCFAs were generally reduced in the defined diets compared to chow, but particularly highly depleted in the high-fat/low-fiber diet, suggesting that our fiber cocktail had some positive effect on SCFA production (Figure 3D). Additionally, we found a concentration-dependent relationship between SCFA levels, toxin levels, and inflammation. In mice-fed chow diets, we observed the expected negative association between butyrate and acetate (but not propionate) and TcdA/TcdB and/or cecal inflammation (Figure 3E). However, in high-fat/low-fiber-fed mice with low SCFA concentrations, there was an unexpected positive correlation between SCFA levels and both toxin and inflammation (Figures 3D and 3E). Butyrate’s relationship with C. difficile is complicated: C. difficile is a butyrate producer, and butyrate can induce toxin production; however, it can also directly inhibit C. difficile growth.¹⁶^,¹⁷ One potential explanation for a positive association between butyrate and inflammation and toxin in the high-fat/low-fiber-fed mice is that butyrate is being produced by C. difficile directly; however, C. difficile relative abundance was not a significant mediator of the positive relationship between butyrate and inflammation in statistical mediation analyses. However, the positive relationship between butyrate and inflammation was related to C. difficile relative abundance indirectly; the interaction term in linear regression models between butyrate levels and C. difficile relative abundance was significant (p-value = 0.005; linear regression), and close inspection showed that butyrate and inflammation were highly positively correlated when C. difficile relative abundance was high (top 50%), but unrelated when the C. difficile relative abundance was low (bottom 50%) (Figure S2C). Taken together, this suggests that butyrate is only related to increased inflammation in the context of an active C. difficile infection in the high-fat/low-fiber diet.

Mice on low-fiber diets demonstrate greater loss of diversity compared to mice on chow and high-fiber diets

We next used 16S rRNA-targeted sequencing to evaluate fecal microbiome composition across the time-course and cecal contents collected upon sacrifice in the day 3 cohort, allowing for direct comparison to C. difficile toxin levels and microbial metabolites measured in the same samples. Cecal microbiome data from day 3 were combined with fecal microbiome data from earlier in the time course in longitudinal microbiome analyses (Figures 4 and 5). Based on prior studies,¹³^,¹⁴ we hypothesized that dietary fiber would promote resilience of the microbiome to antibiotic-induced disturbance, thereby preserving the ability of the microbiome to produce key metabolites that might protect against CDI morbidity, including butyrate and DCA.

Gut microbiome alpha diversity measurements (Faith’s PD) over time between tested diets

(A) Faith’s phylogenetic diversity (alpha diversity) measures. Boxplot lines (from top to bottom) depict the 75^th, 50^th (median), and 25^th percentiles, with the lines extending from the top/bottom of the boxplot indicating the largest/smallest observation within ±1.5∗IQR (inter-quartile range).

(B) Pairwise comparisons were calculated via Kruskal-Wallis and Dunn’s post hoc tests, with p-value corrections via FDR. The color gradient indicates the difference in means, with blue meaning group 2 > group 1 and pink meaning that group 1 > group 2.

(C) Linear regression of Faith’s phylogenetic diversity with deoxycholic acid (DCA; secondary bile acid) and butyrate (SCFA). Linear regression models are done between the x and y axes, accounting for an interaction with diet. P-value significance (∗∗∗∗: p < 0.0001, ∗∗∗: p < 0.001, ∗∗: p < 0.01, and ∗: p < 0.05, ns: p ≥ 0.05).

Changes in the relative abundance of obligate anaerobes and potential pathogens over time between the tested diets

(A and B) Relative abundances throughout the experimental timeline of the (A) obligate anaerobe families *Lachnospiraceae* and *Ruminococcaceae* and (B) four potential pathogen genera. Pairwise comparisons for the top five most abundant genera in *Lachnospiraceae* and *Ruminococcaceae,* and four potential pathogen genera are included in Figure S4. Boxplot lines (from top to bottom) depict the 75^th, 50^th (median), and 25^th percentiles. Vertical lines extending from the top/bottom of the boxplot are the largest/smallest observation within ±1.5∗IQR (inter-quartile range).

(C) Linear regression between microbe relative abundances in the cecum at day 3 and microbiome-produced metabolites, deoxycholic acid (secondary bile acid) and butyrate (SCFA). Linear regression models are performed between the x and y axes, accounting for an interaction with diet.

The 16S rRNA data from the four defined diets had a large percentage of reads assigned to the Lactococcus genus (Figures S3A and S3B). DNA from dead Lactococcus cells can be present in diets containing casein due to the use of Lactococcus in its production.³⁵ Indeed, in the 16S rRNA targeted sequencing of our defined diets, Lactococcus made up 100% of the reads. Lactococcus relative abundance increased with oral antibiotic treatment, with low-fiber diets showing a far larger increase in Lactococcus than the high-fiber diets (peak relative abundance was 79% and 64% in high-fat/low-fiber and low-fat/low-fiber diets vs. 13% and 5% in the respective high-fiber diets (Figures S3A and S3B)). Since the diets all contained similar levels of casein, the higher Lactococcus levels in low-versus high-fiber-fed antibiotic-treated mice are likely due to the lower abundance of intestinal bacteria in low-fiber-fed mice after antibiotic challenge, suggesting a protective effect of fiber on total bacterial load. For all subsequent analyses, Lactococcus reads were removed.

Mice on the chow diet had significantly higher alpha diversity as measured by Faith’s PD³⁶ than mice on the four defined diets at all time points (except for the mice on the low-fiber diets at day −3) (Figures 4A and 4B). Mice on the chow diet began to show a rebound in alpha diversity after antibiotics were stopped (day −3 to 3) that was stronger than the rebound observed in mice on any of the four defined diets (Figures 4A and 4B). Mice fed the high-fat/low-fiber diet had significantly worse recovery of alpha diversity than all other tested diets at sacrifice (day 3) (Figure 4B). Overall, adding the fiber cocktail to a defined diet had only a minor beneficial impact on alpha diversity recovery, which was much less than that observed with the chow diet (Figure 4). Alpha diversity of the cecal microbiome at day 3 was highly positively correlated with both butyrate and DCA levels measured from the same cecal samples (Figure 4C). We saw similar trends of disturbance and recovery in beta diversity analyses using the unweighted UniFrac algorithm³⁷ (see Supplemental text and Figures S3C–S3G).

We next investigated whether the five diets demonstrated differences in the relative abundances of key obligate and facultative anaerobe taxa. Oral antibiotic administration (days −8 to −3) was associated with a marked decrease in Lachnospiraceae and Ruminococcaceae families in all five diets, and mice on all diets had some recovery of these groups at day 3 post-CDI (Figure 5A); however, the recovery was least pronounced in the high-fat/low-fiber-fed mice for Lachnospiraceae. The relative abundance of Lachnospiraceae and Ruminococcaceae in the cecum correlated positively with both DCA and butyrate (Figure 5C), which is consistent with a subset of organisms in this family being producers of these metabolites.³⁸^,³⁹^,⁴⁰ Thus, our data are consistent with the hypothesis that the high-fat/low-fiber diet may have had detrimental effects via the lack of preservation of bacteria that produce butyrate and DCA upon antibiotic exposure.

As expected, mice in all 5 diets showed an increase in the genus Clostridioides following infection with C. difficile spores (days 0–3) (Figure 5B). On day 3, the relative abundance of Clostridioides positively correlated with butyrate in the high-fat/low-fiber diet, and negatively in the chow diet (Figure 5C), consistent with patterns observed with C. difficile toxins (Figure 3E).

We also observed increases in other pathobionts across the time-course with varied patterns based on diet. Specifically, we observed that mice on low-fiber but not high-fiber diets had an increase in the relative abundance of Enterococcus after a week on the new diet (days −15 to −8), and its abundance continued to increase throughout the rest of the experiment (Figure 5B). Enterococcus was high in all diets following clindamycin injection, and only the chow-fed mice showed a reduction at day 3 post-infection. Enterococcus relative abundance was negatively correlated with butyrate, suggesting that butyrate might mediate a protective effect of dietary fiber on Enterococcus outgrowth, although the association could also be from indirect factors (Figure 5C).

Enterobacteriaceae unclassified at the genus level also displayed a sharp increase following the clindamycin injection in all diets (days −3 to 0), a reduction at day 3 only in mice-fed chow and the low-fat/high-fiber diet, and a negative correlation with butyrate in chow and high-fiber-fed mice. Staphylococcus increased in relative abundance following oral antibiotics (days −8 to −3) only in the low-fiber diet mice (Figure 5B). Since genera that have been associated with bacteremia were increasing in these mice following CDI and varying across diets, we began to suspect extra-intestinal causes of mortality.

Mortality was accompanied by disseminated infection by gut-colonizing pathobionts

We conducted follow-up experiments to further investigate the cause of mortality in these mice. These experiments were also designed to test underlying drivers of inconsistent mortality results in the low-fat/low-fiber diet, because in our previously published work, mice fed the low-fat/low-fiber diet had low mortality,¹¹ but they showed high mortality here (Figure 1). One source of variability we considered was the mode of C. difficile challenge, because our earlier work used live cultured bacteria, whereas the current experiments used spores. Upon repeating the survival experiments with live C. difficile cultures, we again observed higher mortality in both low-fiber diets compared to chow (Figure S5A).

Other sources of variability included the mouse vendor and the location where experiments were conducted. Our previous publication used Taconic mice housed in Aurora, Colorado (Anschutz Medical Campus—primary experiment; AMC-P), while the experiments described earlier in this article used Charles River mice housed in Tucson, Arizona (University of Arizona; UA). Our additional series of mouse experiments in Aurora, CO, thus used mice from both Taconic and Charles River housed in the same facility (Anschutz Medical Campus—follow-up; AMC-FU). These experiments used the same protocol as the survival cohort (Figures 1A and S6A) but had added assays for investigating whether the cause of mortality was disseminated infections leading to sepsis. Specifically, at the time of sacrifice, we collected blood, spleen, and liver for bacterial culture and blood to assay for indicators of systemic exposure to bacterial antigens (sCD14),⁴¹^,⁴² immune activation (diverse cytokine panel), and sepsis (blood urea nitrogen (BUN); a marker of kidney injury).

Mortality observed was comparable to our first set of experiments (Figure 1), with both low-fiber diets showing high mortality, and fiber supplementation or chow conferring protection (Figure S6B). Low-fat/low-fiber mice again had poor survival (Figure S6B), and mice-fed chow or high-fiber diets had an earlier onset of mortality than those on low-fiber defined diets, regardless of mouse vendor (Figures S6B and S6C; p ≤ 0.05; Kruskal-Wallis with Dunn’s post hoc test). The onset of mortality coincided with sharp decreases in body weight, following the same trend across diets (Figure S6). Since these experiments were carried out at two different institutions, varied chows were used as well. We did not observe differences in the effects in our chow-fed mice, but recent data have demonstrated differences in CDI severity with different chow formulations.⁴³

To understand how vendor and diet shaped the microbiome in these experiments, we examined fecal microbiomes at baseline (day −15) and day 3 post-CDI using principal coordinate analysis (PCoA) of unweighted UniFrac distances (Figure 6A). Days relative to infection accounted for the largest amount of variation (sum of squares = 23.36, p = 0.001, sequential PERMANOVA), corresponding to separation along PC1. Vendor differences in the baseline microbiome were captured along PC2 (sum of squares = 4.35, p = 0.001), whereas dietary differences at day 3 were apparent along PC3 (sum of squares = 1.70, p = 0.001) (Figure 6A). The Taconic and Charles River mice had pronounced differences in baseline microbiomes and some persistent differences at day 3 post-CDI, when antibiotic treatment and infection produced low-diversity communities (Figure 6A, PC2).

Microbiome taxa contributions in follow-up experiments, cultured bacteria from spleen, liver, and blood of mice just before sacrifice, and changes in the relative abundance of cultured bacteria over time between all conducted experiments

(A) Microbiome unweighted UniFrac PCoA taxa biplots for PC1/PC2 and PC2/PC3 depicting taxa contributions to trends in diet (colors), vendor (shapes), and time point (transparency). Vectors show the correlation of each taxon with PC1/PC2/PC3, with the length indicating the relative strength of the correlation. Significant variance explained was determined by sequential adonis2 PERMANOVA (999 permutations).

(B) Whether mice reached criteria for their humane endpoint prior to the study completion, or survived the experiment, and the number of colony forming units (CFUs) cultured from the blood, spleen, and liver. The type of bacteria cultured from the samples is indicated with different colors. Bacteria were identified by 16S rRNA-targeted sequencing of DNA extracted from scrapings of the culture plate. c) Relative abundance of potentially pathogenic genera across all conducted experiments (AMC-P, UA, AMC-FU) at baseline and 3 days post-infection. Point colors indicate mouse diet, and shapes depict vendors; the lines are colored by diet and connect two samples from the same mouse between time points. Boxplot lines (from top to bottom) depict the 75^th, 50^th (median), and 25^th percentiles. The lines extending from the top/bottom of the boxplot are the largest/smallest observation within ±1.5∗IQR (inter-quartile range). Pairwise comparisons are included in Figures S7B and S7C.

To identify taxa driving these separations, we overlaid taxon biplot vectors onto the PCoA using 16S rRNA-based assignments. At baseline, mice from the different vendors were colonized by distinct Lactobacillus and Bacteroides species. At day 3 post-CDI, Taconic mice were more heavily colonized with Escherichia-Shigella and Bacteroides, while Charles River mice had higher levels of Enterococcus and Proteus (PC2). Across both vendors, low-fiber diets at day 3 were associated with higher abundances of Proteus and Escherichia-Shigella, while high-fiber diets were associated with Akkermansia, Bacteroides, and Erysipelatoclostridium (PC2/PC3). Of note, at day 3, high-fiber diets were in an area of the PCoA plot characterized by higher levels of C. difficile (Figure 6A). All high-fiber diets in the AMC-FU experiment had a significantly higher relative abundance of Clostridioides at day 3 post-infection as compared to the low-fiber diets. While this difference trended toward significance in the AMC-P experiment, it was not statistically significant in the UA experiment (Figures S7A and S7C), although this did appear to be the case in a subset of the high-fat/low-fiber-fed mice (Figure 5B). This increase in C. difficile in the high-fiber diets may be one reason why we observed an earlier onset of weight loss and mortality despite lower overall mortality in the AMC (Figure S6B) and UA (Figure 1) experiments.

To investigate whether opportunistic gut colonizers disseminate following CDI, we aerobically plated whole blood, spleen, and liver collected at sacrifice for bacterial colony-forming units (CFUs). Total CFUs recovered from these sites were significantly higher in mice that reached the humane endpoint due to clinical sickness compared to mice that survived to the end of the time course, regardless of vendor or diet (p-value = 0.003, linear mixed effects model) (Figure 6B). Mice on high-fat and/or low-fiber diets were more likely to have positive cultures from blood and organs than mice on chow or high-fiber diets (Figure 6B). Cultured bacteria included Enterococcus, Escherichia-Shigella, Proteus, and diverse other bacteria (e.g., Lactobacillus, Streptococcus, Staphylococcus, and Bacillus) (Figure 6B). Regardless of diet, Escherichia-Shigella was most often isolated from the blood of Taconic mice, whereas disseminated bacteria from Charles River mice were dominated by Proteus, Enterococcus, or a mixture of Proteus, Enterococcus, and Escherichia-Shigella (Figure 6B). These patterns paralleled the fecal microbiome differences along PC2 at day 3 (Figure 6A), supporting that these bacteria disseminated from the gut following CDI.

We next asked whether fecal relative abundances of the main disseminated genera predicted which mice would succumb early. Using multivariate binomial generalized linear models that accounted for diet and vendor, we tested whether relative abundances of Escherichia-Shigella, Enterococcus, Proteus, or Lactobacillus at baseline and day 3 were associated with reaching the clinical sickness threshold before the experimental endpoint. Baseline Lactobacillus relative abundance was protective, increasing the probability of surviving to the end of the time course (odds ratio of survival to experiment end = 5.86, p = 0.005). In contrast, higher Proteus relative abundance at day 3 post-CDI was strongly associated with early mortality (odds ratio of reaching humane endpoint = 8.2 x 10⁴, p = 0.026), identifying Proteus as a strong risk-associated taxon in this model.

Although we did not collect blood and organ cultures in the earlier AMC-P and UA experiments, we could still evaluate whether key disseminated genera in AMC-FU—Escherichia-Shigella, Proteus, and Enterococcus—showed consistent patterns across studies. Clostridioides was included to evaluate C. difficile colonization status and possible diet-dependent differences in CDI (Figures 6C and S7A). Staphylococcus was also included because of its high relative abundance observed in the UA study (Figure 5B). We therefore compared relative abundances of these taxa at baseline and day 3 across all three experiments, stratifying by vendor and diet (Figure 6C). In the AMC-FU study, Taconic mice had higher Escherichia-Shigella and lower Proteus at both baseline and day 3 than Charles River mice, which had higher Proteus and Enterococcus, consistent with the vendor-specific dissemination patterns observed in the CFU cultures (Figures 6B, 6C, and S7B). Taconic mice from the previously published AMC-P experiments had significantly lower Escherichia-Shigella and significantly higher Proteus at baseline and day 3 compared with Taconic mice in the AMC-FU experiments, and lower Enterococcus at day 3 than mice in either of the more recent studies (Figures 6C and S7B). Additionally, the low-fat/low-fiber mice from the AMC-P experiments had significantly less Proteus at day 3 compared to the high-fat/low-fiber mice in those same experiments (Figures 6C and S7C). Proteus and Escherichia-Shigella were not detected by 16S sequencing in the Charles River mice from the UA experiments; however, those mice had higher levels of Enterococcus and Staphylococcus relative to Taconic mice from the AMC experiments at day 3 (Figures 6C and S7B). All statistics for relative abundances between experiments and stratified by vendor are shown in Figures S7B and S7C.

Taken together, these cross-experiment comparisons indicate that our earlier AMC-P study¹¹ was conducted in mice with a lower baseline and post-CDI prevalence of potential pathogens such as Escherichia-Shigella, Proteus, Enterococcus, and Staphylococcus compared with both UA and AMC-FU experiments. In the more recent experiments, low-fiber diets consistently prompted blooms of these taxa, which were subsequently detected in the blood and organs of moribund mice. Thus, while vendor and location influence which specific gut colonizers predominate, the combination of low-fiber diets, blooms of opportunistic pathogens, and their dissemination provides a plausible explanation for the higher mortality observed in the current experiments compared with our previous work.

Plasma markers of inflammation and kidney damage point to sepsis as a plausible cause of mortality post-CDI

Given that low-fiber diets promoted blooms of gut colonizers that were recovered from blood and organs in moribund mice, we next asked whether mortality was accompanied by systemic inflammation and organ injury consistent with sepsis in the context of CDI. To address this, we quantified BUN and a panel of plasma cytokines from samples collected at the time of sacrifice in the most recent AMC-FU study. BUN is an indicator of kidney function that is frequently elevated in sepsis. The immune panel included proinflammatory cytokines (IL-1β, IL-6, and TNF-α), soluble CD14 (sCD14) as a marker of monocyte activation, which is an indirect measure of bacterial translocation due to intestinal barrier dysfunction,⁴⁴ the neutrophil-attracting chemokine CXCL1, the regulatory cytokine IL-10, and Th1/Th2-associated cytokines (IFN-γ, IL-4, IL-5, and IL-12p70). IL-4 and IL-p70 were also measured but were below or near the limit of detection, so they were excluded from further analysis.

Canberra distances for all remaining markers were calculated to construct a multivariate plasma profile for each mouse at sacrifice, with distances reflecting similarity (points closer together) or dissimilarity (points further apart) of plasma profiles. PC1 accounted for 53.67% of the variance and primarily separated mice that reached the humane endpoint from those that survived to the end of the time course (Figure 7A). Vectors were then calculated based on correlations between each marker and the principal component axes, and a permutation test was used to assess statistical significance. Of the vectors that significantly correlated with the axes (adj. p < 0.05), seven components - BUN, IL-1β, CXCL1, sCD14, TNF-α, IL-10, and IL-6 - point toward the cluster of mice that reached the humane endpoint, indicating that moribund mice had higher levels of these markers (Figure 7A).

Plasma markers of systemic inflammation and kidney injury in CDI mice across diets

(A) PCoA of Canberra distances of mean-normalized sepsis/immune marker concentrations. Points are colored by diet, and shapes indicate whether mice were humanely euthanized due to clinical sickness or survived until the experimental endpoint. Vectors show the correlation of each measured factor with PC1 and PC2, with the vector length indicating the relative strength of the correlation. Only statistically significant vectors are shown (multiple regression with Benjamini-Hochberg multiple test correction, p.adj. < 0.05).

(B) Plasma concentration of each marker that significantly differed between diets (blood urea nitrogen (BUN), and immune factors sCD14, CXCL1, IL-10, IL-1B, IL-6, and TNF-a) at sacrifice. Pairwise comparisons of concentrations between diets were calculated using Kruskal-Wallis and Dunn’s post hoc tests, with p-value corrections conducted via Benjamini and Hochberg. Boxplot lines (from top to bottom) depict the 75^th, 50^th (median), and 25^th percentiles, with lines extending from the top/bottom of the boxplot indicating the largest/smallest observation within ±1.5∗IQR (inter-quartile range). P-value significance (∗∗∗∗: p < 0.0001, ∗∗∗: p < 0.001, ∗∗: p < 0.01, and ∗: p < 0.05).

We next compared concentrations of each measured factor across diets and found that the seven components significantly correlated with early mortality were also the only markers that differed significantly between diets (Figure 7B). Mice in the low-fat/low-fiber group had significantly elevated levels of all seven factors compared to the low-fat/high-fiber group, and elevated levels of all factors except IL-10 compared to the high-fat/high-fiber group. High-fat/low-fiber mice also displayed elevated levels relative to high-fiber groups, but only CXCL1 and IL-6 reached statistical significance (Figure 7B). In the context of CDI and with the culture-confirmed dissemination of bacteria (Figure 6B), these plasma profiles indicate that low-fiber diets are associated with systemic inflammatory and kidney injury signatures consistent with sepsis in moribund mice, whereas high-fiber diets are associated with lower levels of these markers and increased odds of survival.

Discussion

Over the past decade, several studies have supported the impact of diet on CDI in mouse models, including showing that high dietary fat promotes CDI pathogenicity,⁷^,¹⁰^,¹¹ while dietary fiber has protective effects.⁹^,¹⁶^,¹⁷ Since various components contribute to dietary intake, we specifically sought to study the interaction of fat and fiber in CDI. Prior studies using antibiotic-induced murine CDI models to study diet have varied in whether the diet switch occurred before or after antibiotic/C. difficile challenge, the antibiotic treatment used to predispose the mice to infection, and the components used to manipulate dietary fiber and fat in defined diets. Our study used a fiber cocktail of pectin, inulin, resistant starch, and hemicellulose, an aggressive antibiotic regimen to model strong challenges to the microbiome, and a diet switch prior to antibiotics to model the prevention of C. difficile infection as well as recovery post-infection. Our results suggest that the combination of high fat and low fiber in the diet can have particularly detrimental effects on CDI outcomes.

Our data support that high mortality in the context of low-fiber diets was related to secondary bacteremia and sepsis following C. difficile infection, as indicated by higher plasma cytokines and BUN in mice upon reaching their humane endpoint. Levels of pathobionts in the gut at baseline and 3 days post-infection were associated with higher CFU counts of these bacteria in blood and tissue and likely responsible for the variability in mortality in low-fat/low-fiber diets observed between this and our previously published work.¹¹ The lower levels of Enterococcus in the baseline microbiomes of mice from our previously published study compared to current work may be particularly relevant, since Enterococcus has also been shown to enhance C. difficile pathogenesis through metabolic cross-feeding and providing metabolic cues for virulence,⁴⁵ which is particularly detrimental to the host in CDI and may also increase mortality via mechanisms that are independent of fiber intake.

Understanding the mechanisms underlying a higher incidence of bacteremia following CDI in low-fiber diets has clinical significance in humans. Several studies have shown higher mortality in patients with CDI when followed by a bloodstream infection, although this finding is not universal.⁴⁶ The severity of CDI in humans has also been shown to increase the odds of subsequent bloodstream infection with etiologic agents found to be Enterobacteriaceae, Enterococcaceae, and mixed infections, consistent with those observed in this study, as well as Candida.⁴⁷

The SCFA butyrate, a product of dietary fiber metabolism by commensal gut microbes,¹⁵ was depleted in all 4 defined diets compared to chow. Cecal butyrate levels were only slightly higher with our fiber cocktail. However, it is important to note that butyrate is consumed by intestinal epithelial cells, so measured levels in cecal content might not fully represent the amount produced. Butyrate is known to have a particularly complex relationship with CDI since C. difficile is a butyrate producer, but exogenous butyrate can both inhibit C. difficile growth and promote toxin production.¹⁶^,¹⁷^,⁴⁸^,⁴⁹ Our observation that butyrate negatively correlates with C. difficile toxins in chow-fed mice points to a protective mechanism of C. difficile growth inhibition by butyrate at higher concentrations. Butyrate, positively correlating with toxin in the high-fat/low-fiber-fed mice, suggests that measured butyrate levels are driven by production from C. difficile itself, although C. difficile relative abundance was not a significant mediator of the positive relationship between butyrate and inflammation in statistical mediation analyses. An alternative is that exogenous butyrate promotes toxin production at these lower butyrate concentrations. This is supported by our finding that butyrate levels positively correlated with cecal inflammation when the relative abundance of genus Clostridioides was high (top 50%) but did not at low (bottom 50%) abundances (Figure S2B). In low-fiber diets, butyrate concentrations do not show the positive correlation with Lachnospiraceae or Ruminococcaceae that is observed in the high-fiber diets, again suggesting an alternative source of butyrate, such as C. difficile.

Along with the presence of SCFAs, bile acid pool composition can influence C. difficile disease progression. Prior studies have shown that increased primary bile acids promote C. difficile growth, while increased secondary bile acids inhibit growth.²⁰^,²³^,³³^,³⁴ Our results suggest that dietary fiber supplementation promotes the retention of intestinal secondary bile acid-producing bacteria upon microbiome disturbance induced by antibiotic/C. difficile challenge, although we do note that statistical significance was challenged by low sample numbers for bile acids in the high-fiber diets due to limited cecal content for toxin, microbiome, and metabolomic assays (Table S2). Murine models of CDI are somewhat confounding due to the presence of muricholic acids that are primary bile acids but inhibit C. difficile growth.³¹ We thus focused our analysis on the most abundant microbially produced C. difficile inhibitor, DCA. In agreement with both in vitro and clinical data regarding the effects of DCA,³³ we found that C. difficile toxin levels and cecal inflammation decreased with increasing DCA concentrations, supporting the protective effect of DCA in the colon during CDI. Additionally, we observed that cecal microbiome alpha diversity was positively correlated with cecal DCA concentrations, supporting that robust secondary bile acid metabolism requires a diverse microbiome. Although we did find that the high-fat/low-fiber diet had a more pro-C. difficile bile acid pool, measured as the ratio of bile acids with known C. difficile promoting-to-inhibiting activity, the implications of this are unclear. The ratio of bile acids did not correlate with cecal TcdA or TcdB levels at 3 days post-infection and had a complex relationship with cecal histopathology, having the expected positive trend in the chow and high-fiber diets, and the opposite trend in the low-fiber diets. Furthermore, bile acid promoters did not correlate positively with inflammation or C. difficile toxin at 3 days post-infection. One prior study has shown that the treatment of high-fat diet-fed obese mice with obeticholic acid, an agonist of the intestinal farnesoid X receptor, resulted in decreased primary bile acid synthesis, fewer C. difficile bacteria, and better CDI outcomes.⁷ Another study showed disrupted bile acid metabolism in human subjects administered antibiotics while on a low-fiber diet, while mice on a low-fiber diet had prolonged susceptibility to CDI that was associated with similar bile acid changes.²² We thus do not rule out a potential role of primary bile acids here despite the lack of correlation with inflammation or C. difficile toxin levels at day 3 post-infection.

Our results regarding the protective effects of dietary fiber are consistent with prior studies. In one study that used a similar mouse model, supplementing a low-fat/low-fiber diet with pectin but not inulin two weeks before antibiotic/C. difficile challenge protected mice from mortality, suggesting that pectin may have been a more active protective component in our cocktail.⁹ In contrast, studies that began fiber-supplementation after antibiotic/C. difficile challenge with a less broad-spectrum antibiotic, did not see mortality at all, and found that inulin and resistant maltodextrin (but not various other fibers) led to rapid C. difficile clearance.¹⁶^,¹⁷ The protective effect of inulin supplementation after, but not before antibiotics/C. difficile challenge suggests that inulin may not provide sufficient protection for the microbiome from antibiotic disturbance to prevent C. difficile, but could play a role in clearance following an infection.

Studies with chow-fed knockout mice have elucidated immune mechanisms that can underlie disseminated bacterial infections following antibiotic-induced CDI. In one study, mice lacking ASC, a mediator of IL-1β and IL-18 secretion, had impaired CXCL1 production and neutrophil recruitment to the colonic epithelium induced by C. difficile toxin TcdA/TcdB and increased bacterial translocation and bacteremia.⁵⁰ Mice lacking the cytokine IL-22 also had increased susceptibility to bacteremia following CDI in an antibiotic-induced murine model, due to deficits in bacterial phagocytosis mediated by the complement system.⁵¹ Interestingly, the 3 major SCFAs (acetate, propionate and butyrate) have all been shown to promote IL-22 production by CD4⁺ T cells and ILCs,⁵² and were all particularly highly depleted in the cecum of mice fed a high-fat/low-fiber diet compared to chow, suggesting that a lack of SCFA-induced IL-22 production in low-fiber diets could be a driver of the susceptibility to secondary infections following CDI.

Another murine study showed that a low-fiber diet worsened CDI by promoting the development of pathogenic CD4⁺ intraepithelial lymphocytes, and that this was through a reduction in acetate and IL-22 production by ILC3s, which then increase MHC-II expression on the colonic epithelium.¹⁹ Consistent with this, expression levels of IL-22 in colon tissues were lower in mice on a low-fat/low-fiber diet and increased with pectin supplementation in Wu et al.⁹ These findings support the role of SCFAs from dietary fiber in maintaining an appropriate inflammatory response that bolsters intestinal barrier function in response to CDI, ultimately protecting from secondary bacteremia. Additional experiments addressing the effects of SCFAs from dietary fiber in mediating IL-22 production by ILC3s and secondary bacteremia prevention in this context would be beneficial.

Dietary fat has also been shown to have detrimental effects on innate immune activity. Bone marrow and blood neutrophils in mice fed a high-fat diet demonstrated impaired bacterial phagocytosis and extra/intracellular killing, reducing their ability to fight systemic infections.⁵³ Dietary fat-related immune mechanisms also have the potential to influence outcomes in high-fat defined diets.

Differences in intestinal inflammation, along with cecal levels of toxins and metabolites across diets, are likely affected by our choice to sacrifice the mice on day 3 post-infection, which we chose since it was just before the onset of mortality in any diet in the Survival cohort. In the study from Wu et al. described above, mice were subjected to the same model of antibiotic-induced C. difficile infection and diets like our low-fat/low-fiber and low-fat/high-fiber diets, but sacrificed at 6 instead of 3 days post-CDI. Unlike our study, where mice-fed low-fat/low-fiber and low-fat/high-fiber diets did not have significant differences in intestinal inflammation despite differences in survival, Wu et al. did show decreased inflammatory factors (including IL-6, IL-1β, and TNF-α) and protection of intestinal epithelial permeability at 6 days post-infection with added dietary fiber, in line with their mortality results.⁹ It is important to note that infection progression rates in our study were diet-dependent, with the chow- and high-fiber-fed mice having an earlier weight loss and average onset of mortality, and the high-fat/low-fiber-fed mice having the most protracted response. Sacrificing at day 6 was not possible since most low-fat/low-fiber-fed mice had already been euthanized. By sacrificing earlier, we were able to evaluate the early immune response to infection. The diminished inflammation observed at day 3 in the high-fat/low-fiber diet may be driven by either a diminished early immune response or reduced Clostridioides relative abundance in the low-fiber diets. The delayed colonization of C. difficile in the low-fiber diets could have been driven by the nutritional requirements of C. difficile or by cross-feeding relationships. Bacteroides also showed higher relative abundance in the high-fiber-fed mice at day 3, and Bacteroides have previously been shown to support C. difficile through sialic acid metabolism of host mucins.⁵⁴

Both low-fiber diets displayed a greater systemic immune response upon meeting their humane endpoint, as evidenced by higher levels of plasma cytokines TNF-α, IL-6, CXCL1, IL-10, and IL-1β, as well as sCD14. This humane endpoint was on average closer to day 6 post-infection, reflecting the results of Wu et al. above. However, measuring immune responses throughout the time course instead of just on day 3 or at the humane endpoint will be important to investigate in future studies.

Taken together, this work builds upon a growing body of literature showing that diet plays a complicated, but profound, role in CDI. Fiber has been shown to catalyze clearance of C. difficile, while our current work adds that low-fiber diets lead to the dysfunction of the intestinal barrier and dissemination of intestinal pathogens to extra-intestinal locations in the context of CDI. While we did not show clear associations between mortality and bacterial products, including TcdA/TcdB, SCFAs, and DCA, our experiments do provide in vivo evidence of previously described in vitro interactions of toxin production with these bacterial metabolites. This work further supports the idea that dietary therapy can contribute to the prevention of CDI morbidity and protection from secondary bloodstream infections.

Limitations of the study

Although we have demonstrated that dietary fiber is protective against mortality by sepsis secondary to C. difficile infection, the mechanisms underlying this difference are still unclear, as many of the findings in this study are observational. Associations that we observed with SCFAs and relationships with cytokines such as IL-22 in intestinal tissue in other studies suggest a potential mechanism, but we did not measure IL-22 in intestinal tissue. Further experiments that supplement SCFAs into the low-fiber diet-fed mice would also help to specifically implicate the role of SCFAs in the observed differences. Also, studies that independently alter bile acid concentrations from diet would also help to explore whether diet has a mechanistic effect on the bile acid pool. The variable timeline of C. difficile infection/morbidity across the different diets posed temporal challenges in selecting relevant timepoints for sacrifice across all diets for mechanistic assays, without too many mice dropping out of high-mortality diet cohorts in later infection. Our selection of day 3 led to a focus on differences across diets early in disease progression. We could then compare those results to assays conducted at the time of sacrifice in experiments where mice were followed for 15 days post-infection and often sacrificed early because they reached our humane endpoint. However, we still had a limited ability to understand dynamic processes such as how inflammation and toxin production are related to microbial shifts over time. In our later mortality experiments, we used sCD14 as an indirect measure of gut barrier function, but future studies utilizing a more direct metric would be helpful to establish a direct relationship between CDI-induced barrier disruption and secondary sepsis.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Catherine Lozupone (catherine.lozupone@cuanschutz.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

•
The 16S rRNA data are publicly available on QIITA: 16008 and EBI: ERP173015.
•
Additional information on the data analysis process and code associated with it can be found at the study repository on GitHub: https://github.com/madiapgar/diet_mouse_cdiff.
•
Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We would like to thank Dr. James Colbert for giving us guidance on evaluating sepsis in mice. This work was funded by U01AI150589. ESW was funded by an NSF GRFP award (DGE 1938058). LML is supported by the NLM Biomedical Informatics Institutional Training Program (T15LM009451). KZH was supported by the National Center for Advancing Translational Science (NCATS) of the National Institutes of Health under award numbers KL2TR001854 and UL1TR001855 (for administrative/co-curricular activities). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author contributions

M. A.: conceptualization, data curation, formal analysis, software, validation, visualization, writing – original draft, and writing – review and editing. E. S. W.: data curation, investigation, methodology, and writing – review and editing. M. M-K.: investigation. C. P N.: data curation, investigation, methodology, validation, and writing – review and editing. D. J O.: investigation, visualization, and writing – review and editing. L. M L.: formal analysis, visualization, and writing – review and editing. C. B.: formal analysis and investigation. C. G M.: formal analysis and software. O. M G.: conceptualization, data curation, formal analysis, and writing – review and editing. I. J-E.: investigation. J. K F.: investigation, methodology, and writing – review and editing. C. A L.: conceptualization, funding acquisition, methodology, project administration, supervision, writing – original draft, and writing – review and editing. K. Z H.: conceptualization, funding acquisition, investigation, methodology, project administration, supervision, validation, writing – original draft, and writing – review and editing.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains

C. difficile VPI 10463	ATCC	43255-FZ

Chemicals, peptides, and recombinant proteins

Vancomycin	Henry Schein	1474505
Metronidazole	Henry Schein	1455501
Gentamicin Sulfate	Henry Schein	2481007
Kanamycin Sulfate	SigmaAldrich	K1637
Colistin Sulfate	SigmaAldrich	C4461

Critical commercial assays

TcdA and TcdB ELISA	tgcBiomics	TGC-E002-1
Quantikine ELISA Mouse Immunoassay – sCD14	RnD Systems	MC140
QuantiChrom Urea Assay Kit	BioAssay Systems	DIUR-100
V-PLEX Proinflammatory Panel 1 Mouse Kit	Meso Scale Diagnostics	K15048D-1

Deposited data

Publicly available 16S rRNA sequencing data	QIITA	QIITA: 16008
Publicly available 16S rRNA sequencing data	EBI	EBI: ERP173015
Analyzed data and associated code	This study	https://github.com/madiapgar/diet_mouse_cdiff/tree/master

Experimental models: Organisms/strains

C57/bl6 female mice	Taconic Bioscience	C57BL/6NTac
C57/bl6 female mice	Charles River Laboratories	C57BL/6NCrl

Software and algorithms

QIIME2 v2023.5	Bolyen, E. et al.	RRID:SCR_021258
DADA2	Callahan, B. J. et al.⁵⁵	RRID:SCR_023519
SEPP	Janssen, S. et al.⁵⁶	RRID:SCR_024327
RDP Classifier	Wang, Q. et al.⁵⁷	RRID:SCR_022773
R v4.4.1	The R Foundation	RRID:SCR_001905
RStudio v2024.04.1 + 748	Posit Software, PBC formerly RStudio, PBC (2025)	RRID:SCR_000432
tidyverse⁵⁸	v2.0.0	RRID:SCR_019186
survminer⁵⁹	v0.5.0	RRID:SCR_021094
survival⁶⁰	v3.8-3	RRID:SCR_021137
readr	v2.1.5	https://cran.r-project.org/web/packages/readr/index.html
ggplot2⁶¹	v3.5.1	RRID:SCR_014601
ggsignif⁶²	v0.6.4	RRID:SCR_023047
ggh4x	v0.3.0	https://cran.r-project.org/web/packages/ggh4x/index.html
ggpubr	v0.6.0	RRID:SCR_021139
cowplot⁶³	v1.1.3	RRID:SCR_018081
rstatix	v0.7.2	RRID:SCR_021240
qiime2R	v0.99.6	https://github.com/jbisanz/qiime2R
apppleplots	v1.1.3	https://github.com/madiapgar/apppleplots
viridis	v0.6.5	RRID:SCR_016696
stats	v4.4.0	RRID:SCR_025968
vegan	v2.7.1	RRID:SCR_011950
ape	v5.8.1	RRID:SCR_017343
Snakemake v7.32.3	Mölder, F. et al.⁶⁴	RRID:SCR_003475
Anaconda v25.3.1	Anaconda Software⁶⁵	RRID:SCR_025572
Mambaforge (now Miniforge3) v2.0.8	QuantStack and mamba contributors (2024)	https://mamba.readthedocs.io/en/latest/installation/mamba-installation.html
Zotero	v6.0.37	RRID:SCR_013784
GitHub	GitHub, Inc (2025)	RRID:SCR_002630
git	v2.39.5	RRID:SCR_003932

Other

silva-138-99-515-806-nb-classifier taxonomic database (taxonomic classification)	Yilmaz, P. et al.⁶⁵ Quast, C. et al.⁶⁶	RRID:SCR_006423
sepp-refs-silva-128 taxonomic database (tree generation)	Janssen, S. et al.⁵⁶	RRID:SCR_024327
Alpine HPC	University of Colorado Boulder Research Computing⁶⁶	https://curc.readthedocs.io/en/latest/clusters/alpine/index.html

Open in a new tab

Experimental model and study participant details

Murine model of antibiotic-associated CDI

University of Arizona mouse experiments

Experiments took place at The University of Arizona College of Medicine-Tucson with six-week-old female C57BL/6 mice from Charles River Laboratories and diets with varying fat and fiber composition as well as standard mouse chow. Mice were infected following a widely used murine CDI model.²⁴ We chose to only use female mice to maintain consistency with our previous work and due to the female predominance of CDI in humans.¹¹^,⁶⁷ We determined the group sizes for the mouse experiments from power calculations using effect size for TCA and DCA concentrations from our previous work.¹¹ We calculated that groups of 25 would have a power of 95% to detect differences with half of the magnitude we observed previously with a p-value of 0.0125 giving the studies robust statistical power. Mice arrived on day −15 of the experiment and were cohoused with 4–5 mice per cage. For stool microbiome profiling, cecal metabolite, and toxin analysis, (Day 3 cohort), 7 independent experiments were conducted with separate starting dates. 24h after arrival, mouse feed was switched to one of five diets: standard chow, high-fat/high-fiber, high-fat/low-fiber, low-fat/high-fiber, and low-fat/low-fiber diets (7 batches total; groups 1, 2, 5, and 6: n = 15, groups 3 and 4: n = 20, group 7: n = 21). After 7 days on the new diet (day −8), mice were placed on a five-antibiotic oral cocktail (kanamycin (0.4 mg/mL), gentamicin (0.035 mg/mL), colistin (850 U/ml), metronidazole (0.215 mg/mL), and vancomycin (0.045 mg/mL)) via their drinking water for 5 days (days −8 to −3). 48h after the oral antibiotics were stopped (day −1), we administered an intraperitoneal injection of clindamycin in normal saline (10 mg/kg mouse body weight). 24h after mice were injected with clindamycin (day 0), mice were gavaged with 2 × 10² CFU of C. difficile VPI 10463 spores. Mice in the Day 3 Cohort were sacrificed 72h after infection (day 3) and cecal contents were collected for SCFA, bile acid, and toxin quantification. Cecal and colon contents were used for histopathology and hypoxia measurements as well. Fecal pellets were collected at arrival (day −15), 7 days after the diet change (day −8), after the end of oral antibiotic administration (day −3), on the day of infection (day 0), and 3 days post infection (day 3). In a separate set of experiments (Survival Cohort), mice were also obtained from Charles River Laboratories and the same experimental protocol was performed on 160 mice (chow = 40, high-fat/high-fiber = 20, high-fat/low-fiber = 40, low-fat/high-fiber = 20, low-fat/low-fiber = 40) over 4 different batches, but mice were not sacrificed 72h after infection (day 3). Instead, mice were weighed daily after the end of oral antibiotic administration (at day −3) and were euthanized if they lost >15% of their body weight or were moribund. The Institutional Animal Care and Use Committee approved all mouse experiments which complied with their guidelines and the NIH Guide for the Care and Use of Laboratory Animals (University of Arizona IACUC protocol #2021-0716).

AMC mouse experiments

Experiments that took place at the University of Colorado Anschutz Medical Campus were similar to those undertaken at the University of Arizona. Briefly, six-week-old female C57BL/6 mice from both Charles River and Taconic Laboratories arrived on day −15 being fed a chow diet and were cohoused with 5 mice per cage from same facility. For stool microbiome profiling, extraintestinal bacterial assessment, blood analysis, and tissue analysis of intestine, kidney, liver and spleen, 2 independent experiments were conducted with separate starting dates. Experimental protocol was identical to that used in Arizona described above. Upon reaching endpoint criteria (mice weights and temperatures were collected daily and mice were euthanized if they lost >15% of their body weight, reached a temperature of 82.13°F, or were moribund). Euthanized mice were dissected, blood was collected by cardiac bleed into dipotassium EDTA tubes (BD, 365974), with heparinized capillary tubes (Fisher, 22–260950). The spleen, liver, and one kidney were collected and either stored in 10% formalin for histology or were collected into a pre-weighed screw top cryotube (Fisher, 15-340-161) with tissue emulsifying beads (Fisher, 15-340-159) and 500 mL PBS (Gibco, 20012-027). Cecal contents were collected for SCFA, bile acid, and toxin quantification. Cecal and colon were butterflied and fixed in 10% formalin. For each mouse, fecal pellets were collected at arrival (day −15), 7 days after the diet change (day −8), after the end of oral antibiotic administration (day −3), on the day of infection (day 0), and everyday post infection that a mouse survived.

The Institutional Animal Care and Use Committee approved all mouse experiments which complied with their guidelines and the NIH Guide for the Care and Use of Laboratory Animals (University of Colorado Anschutz Medical Campus IACUC protocol 0001365).

Method details

Mouse diets

All diets were from ENVIGO Teklad. Standard mouse chow was item 7913 – NIH-31 Teklad Irradiated Modified Open Formula Mouse/Rat Diet (https://insights.inotivco.com/hubfs/resources/data-sheets/7913.pdf). ENVIGO Control Diet (TD.180811) was used for the low-fat/low-fiber diet and as the base for the low-fat/high-fiber diet. ENVIGO New Total Western Diet IV (TD. 110919) was used for the high-fat/low-fiber diet and as the base for the high-fat/high-fiber diet. The high fiber diets were generated by adding a cocktail of inulin, pectin, resistant starch (raw potato starch), and hemicellulose (psyllium husk) in a 1:1:1:1 ratio to both the Control Diet (produced low-fat/high-fiber) and New Total Western Diet IV (produced high-fat/high-fiber). Additional information on defined diet composition is in Table S1.

C. difficile growth

All experiments were done with C. difficile strain VPI 10463 (ATCC, Manassas Virginia). Spores were prepared according to methods described by Edwards and McBride.⁶⁸ Briefly, in anaerobic conditions we plated frozen glycerol stock of C. difficile on to brain-heart infusion supplement (BHIS) agar and grew at 37^oC overnight. A single colony was inoculated into BHIS liquid media with 0.1% taurocholate and 0.2% fructose and grown overnight at 37^oC in an anaerobic chamber. The next morning, we diluted culture in BHIS to an OD₆₀₀ of 0.5. We then plated 250 μL of culture onto sporulation agar medium (see reference for recipe) and incubated for 72 h at 37^oC in an anaerobic chamber. After 72 h, we scraped the bacterial lawn into PBS, which we subsequently centrifuged and washed once with PBS. To kill non-spore forms of the bacteria, the pellet was treated with 95% ethanol for 1 h at room temperature. We then washed the treated cell with PBS three times before storing in PBS +0.1% w/v BSA at 4^oC. We quantified the colony forming units by plating serial dilutions on to BHIS agar +0.1% TCA. If spores were more than 3 months old, we incubated them after dilution at 55^oC for 20 min to reactivate prior to infection.

DNA extraction and sequencing

Total genomic DNA was extracted from fecal pellets and cecal contents collected from the Day 3 Cohort mice using the DNeasy PowerSoil Kit (Qiagen, Germantown, MD). Modifications from the manufacturer’s protocol were a 10-min incubation at 65^oC immediately after the addition of the lysis buffer along with the use of a bead mill homogenizer at 4.5 m/s for 1 min. The V4 variable region of the 16S rDNA gene was the target for sequencing (515F: GTGCCAGCMGCCGCGGTAA, 806R: GGACTACHVGGGTWTCTAAT). Target DNA was amplified via AccuStart II PCR SuperMix (Quantabio, Beverly, MA). Construction of primers and amplification procedures follow the Earth Microbiome Project guidelines (www.earthmicrobiome.org). Amplified DNA was quantified in a PicoGreen (ThermoFisher Scientific) assay and equal quantities of DNA from each sample were pooled. Pooled DNA was sequenced on the Illumina MiSeq platform (San Diego, CA) using 2x250 run with a V2 500 cycle kit at the Anschutz Center for Microbiome Excellence.

Toxin quantification

C. difficile TcdA and TcdB concentrations were determined from the cecal contents collected from the Day 3 Cohort mice by comparison to a standard curve using ELISA (TGC-E002-1) (tcgBiomics, Germany) (chow = 13, high-fat/high-fiber = 3, high-fat/low-fiber = 15, low-fat/high-fiber = 5, low-fat/low-fiber = 13) and a Molecular Devices kinetic microplate reader. The protocol provided by the manufacturer was followed with the modification of the wavelength at which we blanked against air being 650 nm instead of 620 nm, which was an acceptable substitute according to the manufacturer. Samples that were too small to weigh accurately were assigned a mass of 5 mg for concentration calculations. 5 mg was selected since it was the lowest weight that could be accurately determined.

Histopathologic evaluation of large intestinal tissue

Cecum and transverse colon were harvested from the Day 3 Cohort mice on all diets (chow = 13, high-fat/high-fiber = 6, high-fat/low-fiber = 18, low-fat/high-fiber = 9, low-fat/low-fiber = 16). Tissue was fixed in 10% formalin in PBS, paraffin embedded, and sections cut before hematoxylin and eosin staining by the either the University of Arizona Cancer Center Tissue Acquisition Core or the University of Colorado Anschutz Medical Campus Histopathology Core. Inflammation was assessed in the cecum using the Barthel⁶⁹ scoring system and in the colon using the Dieleman⁷⁰ scoring system by a trained histologist. The Barthel system scores damage to the cecum using 0–3 scores for submucosal edema, neutrophil infiltration, number of goblet cells, and epithelial integrity for a composite score of 0–12. The Dieleman system scores colonic damage from 0 to 3 for inflammation and extent of injury, plus, scores from 0 to 4 for epithelial regeneration and crypt damage. Each score is multiplied by a factor from 0 to 4 accounting for percentage of involvement (0 = 0% and 4 = 100%) for a composite score from 0 to 56. For additional information, please see the original references.

SCFA quantification

SCFAs butyrate, propionate, and acetate were analyzed by stable isotope GC/MS as previously described⁵⁹ (chow = 11, high-fat/high-fiber = 3, high-fat/low-fiber = 15, low-fat/high-fiber = 4, low-fat/low-fiber = 13). Cecal samples were collected from the Day 3 Cohort mice directly into pre-weighed sterile cryovials and flash-frozen in liquid nitrogen and stored at −80°C until processing. Samples were then subject to an alkylation procedure in which sample and alkylating reagent were added, vortexed for 1 min, and incubated at 60°C for 25 min. Following cooling and addition of n-hexane to allow for separation, 170 μL of the organic phase was transferred to an autosampler vial and analyzed by GC/MS. Results were quantified in reference to the stable isotope standard and normalized to sample weight.

Bile acid quantification

Reagents

LC/MS grade methanol, acetonitrile, and isopropanol were obtained from Fisher Scientific (Fairlawn, New Jersey). HPLC-grade water was obtained from Burdick and Jackson (Morristown, New Jersey). Acetic acid, cholic acid (CA), chenodeoxycholic acid (CDCA), lithocholic acid (LCA), taurocholic acid (TCA), and deoxycholic acid (DCA) were obtained from Sigma Aldrich (St. Louis, Missouri). Taurodeoxycholic acid (TCDCA), taurochenodeoxycholic acid (TCDCA), taurolithocholic acid (TLCA), alpha-muricholic acid (a_MCA), and beta-muricholic acid (b_MCA) were obtained from Cayman Chemical (Ann Arbor, Michigan). Chenodeoxycholic acid-d4 (CDCA) and glycochenodeoxycholic acid-d4 were obtained from Cambridge Isotope labs (Tewksbury, Massachusetts).

Standards preparation

An internal standard containing 21 μM of chenodeoxycholic acid-d4 and 21 μM of glycochenodeoxycholic acid-d4 was prepared in 100% methanol. A combined stock of all bile acid standards was prepared at 0.5 mM in 100% methanol. Calibration working standards were then prepared by diluting the combined stock over a range of 0.05–50 μM in methanol. A 20 μL aliquot of each calibration working standard was added to 120 μL of methanol, 50 μL of water, and 10 μL of internal standard (200 μL total) to create ten calibration standards across a calibration range of 0.005–5 μM.

Sample preparation

Cecal samples were collected from the Day 3 Cohort mice (chow = 11, high-fat/high-fiber = 3, high-fat/low-fiber = 15, low-fat/high-fiber = 4, low-fat/low-fiber = 13) and prepared using the method described by Sarafian et al.⁶⁰ with modifications. Briefly, 15–30 mg of cecal sample were weighed in a tared microcentrifuge tube and the weight was recorded. 140 μL of methanol, 15–30 μL of water, and 10 μL of the internal standard were added. The sample was vortexed for 5 s and then incubated in a −20°C freezer for 20 min. The sample was then centrifuged at 3500 × g for 15 min at 4°C. 185–200 μL of the supernatant was transferred to an RSA autosampler vial (Microsolv Technology Corporation, Leland, NC) for immediate analysis or frozen at −70°C until analysis.

High-performance liquid chromatography/quadrupole time-of-flight mass spectrometry (HPLC/QTOF)

HPLC/QTOF mass spectrometry was performed using the method described by Sarafian et al.⁶⁰ with modifications. Separation of bile acids was performed on a 1290 series HPLC from Agilent (Santa Clara, CA) using an Agilent SB-C18 2.1 × 100 mm 1.8 μm column with a 2.1 × 5 mm 1.8 μm guard column. Buffer A consisted of 90:10 water:acetonitrile with 1 mM ammonium acetate adjusted to pH = 4 with acetic acid, and buffer B consisted of 50:50 acetonitrile:isopropanol. 10 μL of the extracted sample was analyzed using the following gradient at a flow rate of 0.6 mL/min: Starting composition = 10% B, linear gradient from 10 to 35% B from 0.1 to 9.25 min, 35–85% B from 9.25 to 11.5 min at 0.65 mL/min, 85–100% B from 11.5 to 11.8 min at 0.8 mL/min, hold at 100% B from 11.8 to 12.4 min at 1.0 mL/min, 100–55% B from 12.4 to 12.5 min 0.85 mL/min, followed by re-equilibration at 10% B from 12.5 to 15 min. The column temperature was held at 60°C for the entire gradient.

Mass spectrometric analysis was performed on an Agilent 6520 quadrupole time-of-flight mass spectrometer in negative ionization mode. The drying gas was 300°C at a flow rate of 12 mL/min. The nebulizer pressure was 30 psi. The capillary voltage was 4000 V. Fragmentor voltage was 200 V. Spectra were acquired in the mass range of 50–1700 m/z with a scan rate of 2 spectra/sec.

Retention time and m/z for each bile acid were determined by injecting authentic standards individually. All of the bile acids produced a prominent [M-H]⁻ ion with negative ionization. The observed retention time and m/z were then used to create a quantitation method. Calibration curves for each calibrated bile acid were constructed using Masshunter Quantitative Analysis software (Aligent Technologies). Bile acid results for feces in pmol/mg were then quantitated using the following calculation:

C = (Xs × Vt × D) ÷ (Vi × Ws)

C = Concentration in pmol/mg.

Xs = pmol on column.

Vt = total volume of concentrated extract (in μL)

D = dilution factor if sample was extracted before analysis. If no dilution, D = 1.

Vi = volume of extract injected (in μL)

Ws = weight of sample extracted in mg.

Extraintestinal bacterial culturing

An aliquot of blood from mice was diluted 1:10 in sterile PBS and 100uL was plated on BHI agar (BD, 241830). The tubes containing a portion of the liver and spleen were weighed to calculate the weight of the tissue. Screw-top tubes with tissues were placed in Omni Bead Mill 24 and emulsified for 30 s 100uL of the tissue solution was plated on BHI agar. All plates were incubated aerobically for 24 h at 37°C. After 24 h, the plates were checked and for growth, if no growth was detected plates continued incubating for 48 h. CFUs were counted and scraped off plates for sequencing with 1 mL of PBS and L shaped cell spreaders (Fisher, 14-665-230).

Plasma biomarker quantification

Terminal plasma samples were collected from mice via cardiac puncture at the time of euthanasia and stored at −80°C until analysis. Soluble CD14 (sCD14) levels were quantified using the Quantikine Mouse sCD14 ELISA Kit (Catalog No. MC140, R&D Systems, Minneapolis, MN, USA) following the manufacturer’s instructions. Blood urea nitrogen (BUN) concentrations were measured using the QuantiChrom Urea Assay Kit (Catalog No. DIUR-100, BioAssay Systems, Hayward, CA, USA) according to the manufacturer’s protocol. Cytokine and chemokine concentrations were determined using the V-PLEX Proinflammatory Panel 1 (Mouse) (Catalog No. K15048D-1, Meso Scale Diagnostics, Rockville, MD, USA) following the manufacturer’s instructions. The V-PLEX plate was read using a Meso Scale Discovery QuickPlex SQ 120 imager, and analyte concentrations were calculated using the accompanying MSD Discovery Workbench software.

Quantification and statistical analysis

Data analysis

16S rDNA sequencing data

Raw paired-end FASTQ files for both the fecal and cecal samples were processed with QIIME2 version 2023.5.⁷¹ Sequence denoising was performed using DADA2 (trim-left-forward and -reverse: 13//trunc-len-forward: 230//trunc-len-reverse: 160). SEPP⁵⁶ was used to build a phylogenetic tree, and the RDP classifier⁵⁷ assigned taxonomic classifications to amplicon sequence variants (ASVs) after being trained on the Silva-138-99-515-806-nb-classifier taxonomic database⁶¹^,⁶² using QIIME2.⁷¹ The data was normalized via total sum scaling in R version 4.4.0 “Puppy Cup”, following the removal of Lactococcus contamination (see Lactococcus Contamination Filtering Steps below). Alpha diversity was measured using Faith’s Phylogenetic Diversity³⁶ and beta diversity was calculated using unweighted UniFrac distances.⁶³ PCoA of unweighted UniFrac plots were constructed in QIIME2. Software was installed using Anaconda version 25.3.1⁶⁵/Mambaforge version 2.0.8 and streamlined via Snakemake⁶⁴ version 7.32.3 on Alpine⁶⁶ compute cluster at the University of Colorado Boulder.

Lactococcus contamination filtering steps

Modified diets and chow were sequenced as previously described and trimmed/truncated to same levels as stool/cecal samples. QIIME2 v. 2023.5 was used to look directly at ASVs annotated as Lactococcus. Filtering out Lactococcus ASVS resulted in 2,642,189 features filtered out of all samples. The Lactococcus ASVs in the modified diets were compared to the ASVs in the mouse samples and 21 out of 21 Lactococcus annotated sequences were identical from both sample types. The 21 Lactococcus ASVs were then matched up to their respective ASVs in the mouse sample BIOM table and filtered out prior to total sum scaling. Sample taxonomic classification was filtered using the new total sum scaled BIOM table and the metadata file prior to alpha and beta diversity generation.

Statistics and plots

Statistical analysis and plot generation was performed in R version 4.4.0 “Puppy Cup” and RStudio version 2024.04.1 + 748. Data pre-processing occurred using “tidyverse”⁵⁸ and “qiime2R” packages. Survival data was processed using “survminer” (specifically the “survival” library),⁷²^,⁷³ “tidyverse”, and “readr” packages. Plasma sepsis/immune marker distance matrix and variable correlations were calculated using “vegan”⁷⁴ and PCoA performed using “ape”.⁷⁵ All other data was plotted using “ggplot2”, “ggh4x”, “ggpubr”, “ggsignif”, “cowplot”, “viridis”, and “apppleplots” packages.⁷⁶^,⁷⁷^,⁷⁸ All statistical tests were performed using “stats” and “rstatix”⁷⁹ packages and were two-tailed with measurements from distinct samples. In all boxplots, the center line represents the median, boxes span the 25th–75th percentiles, and whiskers extend to the largest/smallest observation within ±1.5× IQR. Where means are presented, error bars represent ± one standard deviation unless otherwise noted. Sample sizes (n) represent individual mice and are listed per diet group in Tables S2 and S3. Statistical details for each experiment, including test used, n values, and p-value thresholds, are reported in the corresponding figure legends.⁸⁰^,⁸¹^,⁸²^,⁸³^,⁸⁴^,⁸⁵ All outputs were version-controlled via git version 2.39.5 and are located in a public GitHub repository.

Additional resources

There are no additional resources to report.

Published: March 6, 2026

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.115258.

Contributor Information

Catherine A. Lozupone, Email: catherine.lozupone@cuanschutz.edu.

Keith Z. Hazleton, Email: khazleton@chla.usc.edu.

Supplemental information

Document S1. Figures S1–S7, Tables S1–S3

mmc1.pdf^{(3.1MB, pdf)}

References

1.Leffler D.A., Lamont J.T. Clostridium difficile Infection. N. Engl. J. Med. 2015;372:1539–1548. doi: 10.1056/NEJMra1403772. [DOI] [PubMed] [Google Scholar]
2.Guh A.Y., Mu Y., Winston L.G., Johnston H., Olson D., Farley M.M., Wilson L.E., Holzbauer S.M., Phipps E.C., Dumyati G.K., et al. Trends in U.S. Burden of Clostridioides difficile Infection and Outcomes. N. Engl. J. Med. 2020;382:1320–1330. doi: 10.1056/NEJMoa1910215. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Buffie C.G., Bucci V., Stein R.R., McKenney P.T., Ling L., Gobourne A., No D., Liu H., Kinnebrew M., Viale A., et al. Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile. Nature. 2015;517:205–208. doi: 10.1038/nature13828. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Adams D.J., Eberly M.D., Rajnik M., Nylund C.M. Risk Factors for Community-Associated Clostridium difficile Infection in Children. J. Pediatr. 2017;186:105–109. doi: 10.1016/j.jpeds.2017.03.032. [DOI] [PubMed] [Google Scholar]
5.Bignardi G.E. Risk factors for Clostridium difficile infection. J. Hosp. Infect. 1998;40:1–15. doi: 10.1016/s0195-6701(98)90019-6. [DOI] [PubMed] [Google Scholar]
6.Nylund C.M., Goudie A., Garza J.M., Fairbrother G., Cohen M.B. Clostridium difficile Infection in Hospitalized Children in the United States. Arch. Pediatr. Adolesc. Med. 2011;165:451–457. doi: 10.1001/archpediatrics.2010.282. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Jose S., Mukherjee A., Horrigan O., Setchell K.D.R., Zhang W., Moreno-Fernandez M.E., Andersen H., Sharma D., Haslam D.B., Divanovic S., Madan R. Obeticholic acid ameliorates severity of Clostridioides difficile infection in high fat diet-induced obese mice. Mucosal Immunol. 2021;14:500–510. doi: 10.1038/s41385-020-00338-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Moore J.H., Pinheiro C.C.D., Zaenker E.I., Bolick D.T., Kolling G.L., Van Opstal E., Noronha F.J.D., De Medeiros P.H.Q.S., Rodriguez R.S., Lima A.A., et al. Defined Nutrient Diets Alter Susceptibility to Clostridium difficile Associated Disease in a Murine Model. PLoS One. 2015;10 doi: 10.1371/journal.pone.0131829. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Wu Z., Xu Q., Wang Q., Chen Y., Lv L., Zheng B., Yan R., Jiang H., Shen J., Wang S., et al. The impact of dietary fibers on Clostridioides difficile infection in a mouse model. Front. Cell. Infect. Microbiol. 2022;12 doi: 10.3389/fcimb.2022.1028267. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Mefferd C.C., Bhute S.S., Phan J.R., Villarama J.V., Do D.M., Alarcia S., Abel-Santos E., Hedlund B.P. A High-Fat/High-Protein, Atkins-Type Diet Exacerbates Clostridioides (Clostridium) difficile Infection in Mice, whereas a High-Carbohydrate Diet Protects. mSystems. 2020;5:10–1128. doi: 10.1128/mSystems.00765-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Hazleton K.Z., Martin C.G., Orlicky D.J., Arnolds K.L., Nusbacher N.M., Moreno-Huizar N., Armstrong M., Reisdorph N., Lozupone C.A. Dietary fat promotes antibiotic-induced Clostridioides difficile mortality in mice. Npj Biofilms Microbiomes. 2022;8:15. doi: 10.1038/s41522-022-00276-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Lewis S., Burmeister S., Brazier J. Effect of the prebiotic oligofructose on relapse of Clostridium difficile-associated diarrhea: a randomized, controlled study. Clin. Gastroenterol. Hepatol. 2005;3:442–448. doi: 10.1016/s1542-3565(04)00677-9. [DOI] [PubMed] [Google Scholar]
13.Tanes C., Bittinger K., Gao Y., Friedman E.S., Nessel L., Paladhi U.R., Chau L., Panfen E., Fischbach M.A., Braun J., et al. Role of dietary fiber in the recovery of the human gut microbiome and its metabolome. Cell Host Microbe. 2021;29:394–407.e5. doi: 10.1016/j.chom.2020.12.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Penumutchu S., Korry B.J., Hewlett K., Belenky P. Fiber supplementation protects from antibiotic-induced gut microbiome dysbiosis by modulating gut redox potential. Nat. Commun. 2023;14:5161. doi: 10.1038/s41467-023-40553-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Cummings J.H., Pomare E.W., Branch W.J., Naylor C.P., Macfarlane G.T. Short chain fatty acids in human large intestine, portal, hepatic and venous blood. Gut. 1987;28:1221–1227. doi: 10.1136/gut.28.10.1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Hryckowian A.J., Van Treuren W., Smits S.A., Davis N.M., Gardner J.O., Bouley D.M., Sonnenburg J.L. Microbiota-accessible carbohydrates suppress Clostridium difficile infection in a murine model. Nat. Microbiol. 2018;3:662–669. doi: 10.1038/s41564-018-0150-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Pensinger D.A., Fisher A.T., Dobrila H.A., Van Treuren W., Gardner J.O., Higginbottom S.K., Carter M.M., Schumann B., Bertozzi C.R., Anikst V., et al. Butyrate Differentiates Permissiveness to Clostridioides difficile Infection and Influences Growth of Diverse C. difficile Isolates. Infect. Immun. 2023;91:e00570-22. doi: 10.1128/iai.00570-22. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Fachi J.L., Felipe J.D.S., Pral L.P., Da Silva B.K., Corrêa R.O., De Andrade M.C.P., Da Fonseca D.M., Basso P.J., Câmara N.O.S., de Sales E Souza É.L., et al. Butyrate Protects Mice from Clostridium difficile-Induced Colitis through an HIF-1-Dependent Mechanism. Cell Rep. 2019;27:750–761.e7. doi: 10.1016/j.celrep.2019.03.054. [DOI] [PubMed] [Google Scholar]
19.Fachi J.L., De Oliveira S., Trsan T., Penati S., Gilfillan S., Cao S., Ribeiro Castro P., Fernandes M.F., Hyrc K.L., Liu X., et al. Fiber- and acetate-mediated modulation of MHC-II expression on intestinal epithelium protects from Clostridioides difficile infection. Cell Host Microbe. 2025;33:235–251.e7. doi: 10.1016/j.chom.2024.12.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Sorg J.A., Sonenshein A.L. Bile salts and glycine as cogerminants for Clostridium difficile spores. J. Bacteriol. 2008;190:2505–2512. doi: 10.1128/JB.01765-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Theriot C.M., Bowman A.A., Young V.B. Antibiotic-Induced Alterations of the Gut Microbiota Alter Secondary Bile Acid Production and Allow for Clostridium difficile Spore Germination and Outgrowth in the Large Intestine. mSphere. 2016;1:e00045-15. doi: 10.1128/mSphere.00045-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Hewlett K.K., PeBenito A., Hecht A.L., Tiffany C., Tanes C., Glover R.C., Fawad J.A., Friedman E.S., Reynolds J.C., Bittinger K., et al. Dietary Fiber Modulates the Window of Susceptibility to Clostridioides difficile Infection. Gastroenterology. 2025;169:970–982. doi: 10.1053/j.gastro.2025.04.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Allegretti J.R., Kearney S., Li N., Bogart E., Bullock K., Gerber G.K., Bry L., Clish C.B., Alm E., Korzenik J.R. Recurrent Clostridium difficile infection associates with distinct bile acid and microbiome profiles. Aliment. Pharmacol. Ther. 2016;43:1142–1153. doi: 10.1111/apt.13616. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Chen X., Katchar K., Goldsmith J.D., Nanthakumar N., Cheknis A., Gerding D.N., Kelly C.P. A Mouse Model of Clostridium difficile–Associated Disease. Gastroenterology. 2008;135:1984–1992. doi: 10.1053/j.gastro.2008.09.002. [DOI] [PubMed] [Google Scholar]
25.Hintze K.J., Benninghoff A.D., Ward R.E. Formulation of the Total Western Diet (TWD) as a basal diet for rodent cancer studies. J. Agric. Food Chem. 2012;60:6736–6742. doi: 10.1021/jf204509a. [DOI] [PubMed] [Google Scholar]
26.Mudgil D., Barak S. Composition, properties and health benefits of indigestible carbohydrate polymers as dietary fiber: A review. Int. J. Biol. Macromol. 2013;61:1–6. doi: 10.1016/j.ijbiomac.2013.06.044. [DOI] [PubMed] [Google Scholar]
27.Jana U.K., Kango N., Pletschke B. Hemicellulose-Derived Oligosaccharides: Emerging Prebiotics in Disease Alleviation. Front. Nutr. 2021;8 doi: 10.3389/fnut.2021.670817. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Jiang T., Gao X., Wu C., Tian F., Lei Q., Bi J., Xie B., Wang H.Y., Chen S., Wang X. Apple-Derived Pectin Modulates Gut Microbiota, Improves Gut Barrier Function, and Attenuates Metabolic Endotoxemia in Rats with Diet-Induced Obesity. Nutrients. 2016;8:126. doi: 10.3390/nu8030126. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Teferra T.F. Possible actions of inulin as prebiotic polysaccharide: A review. Food Front. 2021;2:407–416. doi: 10.1002/fft2.92. [DOI] [Google Scholar]
30.Benjamini Y., Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B Stat. Methodol. 1995;57:289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
31.Francis M.B., Allen C.A., Sorg J.A. Muricholic Acids Inhibit Clostridium difficile Spore Germination and Growth. PLoS One. 2013;8 doi: 10.1371/journal.pone.0073653. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Francis M.B., Allen C.A., Shrestha R., Sorg J.A. Bile acid recognition by the Clostridium difficile germinant receptor, CspC, is important for establishing infection. PLoS Pathog. 2013;9 doi: 10.1371/journal.ppat.1003356. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Thanissery R., Winston J.A., Theriot C.M. Inhibition of spore germination, growth, and toxin activity of clinically relevant C. difficile strains by gut microbiota derived secondary bile acids. Anaerobe. 2017;45:86–100. doi: 10.1016/j.anaerobe.2017.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Sorg J.A., Sonenshein A.L. Inhibiting the initiation of Clostridium difficile spore germination using analogs of chenodeoxycholic acid, a bile acid. J. Bacteriol. 2010;192:4983–4990. doi: 10.1128/JB.00610-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Bisanz J.E., Upadhyay V., Turnbaugh J.A., Ly K., Turnbaugh P.J. Meta-Analysis Reveals Reproducible Gut Microbiome Alterations in Response to a High-Fat Diet. Cell Host Microbe. 2019;26:265–272.e4. doi: 10.1016/j.chom.2019.06.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Faith D.P. Conservation evaluation and phylogenetic diversity. Biol. Conserv. 1992;61:1–10. doi: 10.1016/0006-3207(92)91201-3. [DOI] [Google Scholar]
37.Lozupone C., Lladser M.E., Knights D., Stombaugh J., Knight R. UniFrac: an effective distance metric for microbial community comparison. ISME J. 2011;5:169–172. doi: 10.1038/ismej.2010.133. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Sinha S.R., Haileselassie Y., Nguyen L.P., Tropini C., Wang M., Becker L.S., Sim D., Jarr K., Spear E.T., Singh G., et al. Dysbiosis-Induced Secondary Bile Acid Deficiency Promotes Intestinal Inflammation. Cell Host Microbe. 2020;27:659–670.e5. doi: 10.1016/j.chom.2020.01.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Kakiyama G., Pandak W.M., Gillevet P.M., Hylemon P.B., Heuman D.M., Daita K., Takei H., Muto A., Nittono H., Ridlon J.M., et al. Modulation of the fecal bile acid profile by gut microbiota in cirrhosis. J. Hepatol. 2013;58:949–955. doi: 10.1016/j.jhep.2013.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Marion S., Studer N., Desharnais L., Menin L., Escrig S., Meibom A., Hapfelmeier S., Bernier-Latmani R. In vitro and in vivo characterization of Clostridium scindens bile acid transformations. Gut Microbes. 2019;10:481–503. doi: 10.1080/19490976.2018.1549420. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Lakatos P.L., Kiss L.S., Palatka K., Altorjay I., Antal-Szalmas P., Palyu E., Udvardy M., Molnar T., Farkas K., Veres G., et al. Serum lipopolysaccharide-binding protein and soluble CD14 are markers of disease activity in patients with Crohn’s disease. Inflamm. Bowel Dis. 2011;17:767–777. doi: 10.1002/ibd.21402. [DOI] [PubMed] [Google Scholar]
42.Pastor Rojo O., López San Román A., Albéniz Arbizu E., de la Hera Martínez A., Ripoll Sevillano E., Albillos Martínez A. Serum lipopolysaccharide-binding protein in endotoxemic patients with inflammatory bowel disease. Inflamm. Bowel Dis. 2007;13:269–277. doi: 10.1002/ibd.20019. [DOI] [PubMed] [Google Scholar]
43.Denny J.E., Flores J.N., Mdluli N.V., Abt M.C. Standard mouse diets lead to differences in severity in infectious and non-infectious colitis. mBio. 2025;16 doi: 10.1128/mbio.03302-24. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Tabung F.K., Birmann B.M., Epstein M.M., Martínez-Maza O., Breen E.C., Wu K., Giovannucci E.L. Influence of Dietary Patterns on Plasma Soluble CD14, a Surrogate Marker of Gut Barrier Dysfunction. Curr. Dev. Nutr. 2017;1 doi: 10.3945/cdn.117.001396. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Smith A.B., Jenior M.L., Keenan O., Hart J.L., Specker J., Abbas A., Rangel P.C., Di C., Green J., Bustin K.A., et al. Enterococci enhance Clostridioides difficile pathogenesis. Nature. 2022;611:780–786. doi: 10.1038/s41586-022-05438-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Ulrich R.J., Santhosh K., Mogle J.A., Young V.B., Rao K. Is Clostridium difficile infection a risk factor for subsequent bloodstream infection? Anaerobe. 2017;48:27–33. doi: 10.1016/j.anaerobe.2017.06.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Trunfio M., Scabini S., Rugge W., Bonora S., Di Perri G., Calcagno A. Concurrent and Subsequent Co-Infections of Clostridioides difficile Colitis in the Era of Gut Microbiota and Expanding Treatment Options. Microorganisms. 2022;10:1275. doi: 10.3390/microorganisms10071275. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Pensinger D.A., Dobrila H.A., Stevenson D.M., Hryckowian N.D., Amador-Noguez D., Hryckowian A.J. Exogenous butyrate inhibits butyrogenic metabolism and alters virulence phenotypes in Clostridioides difficile. mBio. 2024;15:e02535-23. doi: 10.1128/mbio.02535-23. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Gregory A.L., Pensinger D.A., Hryckowian A.J. A short chain fatty acid–centric view of Clostridioides difficile pathogenesis. PLoS Pathog. 2021;17 doi: 10.1371/journal.ppat.1009959. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Hasegawa M., Kamada N., Jiao Y., Liu M.Z., Núñez G., Inohara N. Protective Role of Commensals against Clostridium difficile Infection via an IL-1β–Mediated Positive-Feedback Loop. J. Immunol. 2012;189:3085–3091. doi: 10.4049/jimmunol.1200821. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Hasegawa M., Yada S., Liu M.Z., Kamada N., Muñoz-Planillo R., Do N., Núñez G., Inohara N. Interleukin-22 regulates the complement system to promote resistance against pathobionts after pathogen-induced intestinal damage. Immunity. 2014;41:620–632. doi: 10.1016/j.immuni.2014.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Yang W., Yu T., Huang X., Bilotta A.J., Xu L., Lu Y., Sun J., Pan F., Zhou J., Zhang W., et al. Intestinal microbiota-derived short-chain fatty acids regulation of immune cell IL-22 production and gut immunity. Nat. Commun. 2020;11:4457. doi: 10.1038/s41467-020-18262-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Bowers E., Entrup G.P., Islam M., Mohan R., Lerner A., Mancuso P., Moore B.B., Singer K. High fat diet feeding impairs neutrophil phagocytosis, bacterial killing, and neutrophil-induced hematopoietic regeneration. J. Immunol. 2025;214:680–693. doi: 10.1093/jimmun/vkaf024. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Ng K.M., Ferreyra J.A., Higginbottom S.K., Lynch J.B., Kashyap P.C., Gopinath S., Naidu N., Choudhury B., Weimer B.C., Monack D.M., Sonnenburg J.L. Microbiota-liberated host sugars facilitate post-antibiotic expansion of enteric pathogens. Nature. 2013;502:96–99. doi: 10.1038/nature12503. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Callahan B.J., McMurdie P.J., Rosen M.J., Han A.W., Johnson A.J.A., Holmes S.P. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods. 2016;13:581–583. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Janssen S., McDonald D., Gonzalez A., Navas-Molina J.A., Jiang L., Xu Z.Z., Winker K., Kado D.M., Orwoll E., Manary M., et al. Phylogenetic Placement of Exact Amplicon Sequences Improves Associations with Clinical Information. mSystems. 2018;3:e00021-18. doi: 10.1128/mSystems.00021-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Wang Q., Garrity G.M., Tiedje J.M., Cole J.R. Naïve Bayesian Classifier for Rapid Assignment of rRNA Sequences into the New Bacterial Taxonomy. Appl. Environ. Microbiol. 2007;73:5261–5267. doi: 10.1128/AEM.00062-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Wickham H., Averick M., Bryan J., Chang W., McGowan L., François R., Grolemund G., Hayes A., Henry L., Hester J., et al. Welcome to the Tidyverse. J. Open Source Softw. 2019;4:1686. doi: 10.21105/joss.01686. [DOI] [Google Scholar]
59.Kelly C.J., Alexeev E.E., Farb L., Vickery T.W., Zheng L., Eric L C., Kitzenberg D.A., Battista K.D., Kominsky D.J., Robertson C.E., et al. Oral vitamin B 12 supplement is delivered to the distal gut, altering the corrinoid profile and selectively depleting Bacteroides in C57BL/6 mice. Gut Microbes. 2019;10:654–662. doi: 10.1080/19490976.2019.1597667. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Sarafian M.H., Lewis M.R., Pechlivanis A., Ralphs S., McPhail M.J.W., Patel V.C., Dumas M.-E., Holmes E., Nicholson J.K. Bile Acid Profiling and Quantification in Biofluids Using Ultra-Performance Liquid Chromatography Tandem Mass Spectrometry. Anal. Chem. 2015;87:9662–9670. doi: 10.1021/acs.analchem.5b01556. [DOI] [PubMed] [Google Scholar]
61.Yilmaz P., Parfrey L.W., Yarza P., Gerken J., Pruesse E., Quast C., Schweer T., Peplies J., Ludwig W., Glöckner F.O. The SILVA and “All-species Living Tree Project (LTP)” taxonomic frameworks. Nucleic Acids Res. 2014;42:D643–D648. doi: 10.1093/nar/gkt1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Quast C., Pruesse E., Yilmaz P., Gerken J., Schweer T., Yarza P., Peplies J., Glöckner F.O. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–D596. doi: 10.1093/nar/gks1219. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Lozupone C.A., Hamady M., Kelley S.T., Knight R. Quantitative and Qualitative β Diversity Measures Lead to Different Insights into Factors That Structure Microbial Communities. Appl. Environ. Microbiol. 2007;73:1576–1585. doi: 10.1128/AEM.01996-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Mölder F., Jablonski K.P., Letcher B., Hall M.B., van Dyken P.C., Tomkins-Tinch C.H., Sochat V., Forster J., Vieira F.G., Meesters C., et al. Sustainable data analysis with Snakemake. F1000Res. 2021;10:33. doi: 10.12688/f1000research.29032.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Anaconda Software (2012). (Anaconda Inc.).
66.University of Colorado Boulder Research Computing Alpine. 2023. [DOI]
67.Centers for Disease Control and Prevention . Centers for Disease Control and Prevention; 2024. Emerging Infections Program, Healthcare-Associated Infections – Community Interface Surveillance Report: Clostridioides Difficile Infection (CDI), 2022. [Google Scholar]
68.Edwards A.N., McBride S.M. Isolating and Purifying Clostridium difficile Spores. Methods Mol. Biol. 2016;1476:117–128. doi: 10.1007/978-1-4939-6361-4_9. [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Barthel M., Hapfelmeier S., Quintanilla-Martínez L., Kremer M., Rohde M., Hogardt M., Pfeffer K., Rüssmann H., Hardt W.-D. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host. Infect. Immun. 2003;71:2839–2858. doi: 10.1128/IAI.71.5.2839-2858.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Dieleman L.A., Palmen M.J., Akol H., Bloemena E., Peña A.S., Meuwissen S.G., Van Rees E.P. Chronic experimental colitis induced by dextran sulphate sodium (DSS) is characterized by Th1 and Th2 cytokines. Clin. Exp. Immunol. 1998;114:385–391. doi: 10.1046/j.1365-2249.1998.00728.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Bolyen E., Rideout J.R., Dillon M.R., Bokulich N.A., Abnet C.C., Al-Ghalith G.A., Alexander H., Alm E.J., Arumugam M., Asnicar F., et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 2019;37:852–857. doi: 10.1038/s41587-019-0209-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Kassambara A., Kosinski M., Biecek P. Survminer: Drawing Survival Curves Using Ggplot2. Version 0.4.9. 2020. https://rpkgs.datanovia.com/survminer/index.html [Google Scholar]
73.Therneau T.M. Survival: Survival Analysis. Version 3.5-8. 2024. https://github.com/therneau/survival [Google Scholar]
74.Oksanen J., Simpson G.L., Blanchet F.G., Kindt R., Legendre P., Minchin P.R., O’Hara R.B., Solymos P., Stevens M.H.H., Szoecs E., et al. Vegan: Community Ecology Package. 2025. https://vegandevs.github.io/vegan/ [Google Scholar]
75.Paradis E., Schliep K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics. 2019;35:526–528. doi: 10.1093/bioinformatics/bty633. [DOI] [PubMed] [Google Scholar]
76.Wickham H. Second edition. Springer; 2016. ggplot2: Elegant Graphics for Data Analysis. [DOI] [Google Scholar]
77.Ahlmann-Eltze C., Patil I. ggsignif: R Package for Displaying Significance Brackets for “ggplot2”. 2021. (PsyArXiv) [DOI]
78.Wilke C. cowplot: Streamlined Plot Theme and Plot Annotations for “ggplot2” Version R package version 1.1.3. 2024. https://wilkelab.org/cowplot/ [Google Scholar]
79.Kassambara A. rstatix: Pipe-Friendly Framework for Basic Statistical Tests. Version R package version 0.7.2. 2023. https://rpkgs.datanovia.com/rstatix/ [Google Scholar]
80.Li X., Zheng R., Zhang T., Zeng Z., Li H., Liu J. Association between blood urea nitrogen and 30-day mortality in patients with sepsis: a retrospective analysis. Ann. Palliat. Med. 2021;10:11653–11663. doi: 10.21037/apm-21-2937. [DOI] [PubMed] [Google Scholar]
81.Castro M., Silver H.J., Hazleton K., Lozupone C., Nicholson M.R. The Impact of Diet on Clostridioides difficile Infection: A Review. J. Infect. Dis. 2025;231:e1010–e1018. doi: 10.1093/infdis/jiaf233. jiaf233. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Falcone M., Russo A., Iraci F., Carfagna P., Goldoni P., Vullo V., Venditti M. Risk Factors and Outcomes for Bloodstream Infections Secondary to Clostridium difficile Infection. Antimicrob. Agents Chemother. 2016;60:252–257. doi: 10.1128/AAC.01927-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Falcone M., Venditti M., Sanguinetti M., Posteraro B. Management of candidemia in patients with Clostridium difficile infection. Expert Rev. Anti Infect. Ther. 2016;14:679–685. doi: 10.1080/14787210.2016.1197118. [DOI] [PubMed] [Google Scholar]
84.Ianiro G., Murri R., Sciumè G.D., Impagnatiello M., Masucci L., Ford A.C., Law G.R., Tilg H., Sanguinetti M., Cauda R., et al. Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients With Recurrent Clostridioides difficile Infection Treated With Fecal Microbiota Transplantation or Antibiotics. Ann. Intern. Med. 2019;171:695–702. doi: 10.7326/M18-3635. [DOI] [PubMed] [Google Scholar]
85.Thompson L.R., Sanders J.G., McDonald D., Amir A., Ladau J., Locey K.J., Prill R.J., Tripathi A., Gibbons S.M., Ackermann G., et al. A communal catalogue reveals Earth’s multiscale microbial diversity. Nature. 2017;551:457–463. doi: 10.1038/nature24621. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S7, Tables S1–S3

mmc1.pdf^{(3.1MB, pdf)}

Data Availability Statement

•
The 16S rRNA data are publicly available on QIITA: 16008 and EBI: ERP173015.
•
Additional information on the data analysis process and code associated with it can be found at the study repository on GitHub: https://github.com/madiapgar/diet_mouse_cdiff.
•
Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

[bib1] 1.Leffler D.A., Lamont J.T. Clostridium difficile Infection. N. Engl. J. Med. 2015;372:1539–1548. doi: 10.1056/NEJMra1403772. [DOI] [PubMed] [Google Scholar]

[bib2] 2.Guh A.Y., Mu Y., Winston L.G., Johnston H., Olson D., Farley M.M., Wilson L.E., Holzbauer S.M., Phipps E.C., Dumyati G.K., et al. Trends in U.S. Burden of Clostridioides difficile Infection and Outcomes. N. Engl. J. Med. 2020;382:1320–1330. doi: 10.1056/NEJMoa1910215. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.Buffie C.G., Bucci V., Stein R.R., McKenney P.T., Ling L., Gobourne A., No D., Liu H., Kinnebrew M., Viale A., et al. Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile. Nature. 2015;517:205–208. doi: 10.1038/nature13828. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Adams D.J., Eberly M.D., Rajnik M., Nylund C.M. Risk Factors for Community-Associated Clostridium difficile Infection in Children. J. Pediatr. 2017;186:105–109. doi: 10.1016/j.jpeds.2017.03.032. [DOI] [PubMed] [Google Scholar]

[bib5] 5.Bignardi G.E. Risk factors for Clostridium difficile infection. J. Hosp. Infect. 1998;40:1–15. doi: 10.1016/s0195-6701(98)90019-6. [DOI] [PubMed] [Google Scholar]

[bib6] 6.Nylund C.M., Goudie A., Garza J.M., Fairbrother G., Cohen M.B. Clostridium difficile Infection in Hospitalized Children in the United States. Arch. Pediatr. Adolesc. Med. 2011;165:451–457. doi: 10.1001/archpediatrics.2010.282. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Jose S., Mukherjee A., Horrigan O., Setchell K.D.R., Zhang W., Moreno-Fernandez M.E., Andersen H., Sharma D., Haslam D.B., Divanovic S., Madan R. Obeticholic acid ameliorates severity of Clostridioides difficile infection in high fat diet-induced obese mice. Mucosal Immunol. 2021;14:500–510. doi: 10.1038/s41385-020-00338-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Moore J.H., Pinheiro C.C.D., Zaenker E.I., Bolick D.T., Kolling G.L., Van Opstal E., Noronha F.J.D., De Medeiros P.H.Q.S., Rodriguez R.S., Lima A.A., et al. Defined Nutrient Diets Alter Susceptibility to Clostridium difficile Associated Disease in a Murine Model. PLoS One. 2015;10 doi: 10.1371/journal.pone.0131829. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Wu Z., Xu Q., Wang Q., Chen Y., Lv L., Zheng B., Yan R., Jiang H., Shen J., Wang S., et al. The impact of dietary fibers on Clostridioides difficile infection in a mouse model. Front. Cell. Infect. Microbiol. 2022;12 doi: 10.3389/fcimb.2022.1028267. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Mefferd C.C., Bhute S.S., Phan J.R., Villarama J.V., Do D.M., Alarcia S., Abel-Santos E., Hedlund B.P. A High-Fat/High-Protein, Atkins-Type Diet Exacerbates Clostridioides (Clostridium) difficile Infection in Mice, whereas a High-Carbohydrate Diet Protects. mSystems. 2020;5:10–1128. doi: 10.1128/mSystems.00765-19. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Hazleton K.Z., Martin C.G., Orlicky D.J., Arnolds K.L., Nusbacher N.M., Moreno-Huizar N., Armstrong M., Reisdorph N., Lozupone C.A. Dietary fat promotes antibiotic-induced Clostridioides difficile mortality in mice. Npj Biofilms Microbiomes. 2022;8:15. doi: 10.1038/s41522-022-00276-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Lewis S., Burmeister S., Brazier J. Effect of the prebiotic oligofructose on relapse of Clostridium difficile-associated diarrhea: a randomized, controlled study. Clin. Gastroenterol. Hepatol. 2005;3:442–448. doi: 10.1016/s1542-3565(04)00677-9. [DOI] [PubMed] [Google Scholar]

[bib13] 13.Tanes C., Bittinger K., Gao Y., Friedman E.S., Nessel L., Paladhi U.R., Chau L., Panfen E., Fischbach M.A., Braun J., et al. Role of dietary fiber in the recovery of the human gut microbiome and its metabolome. Cell Host Microbe. 2021;29:394–407.e5. doi: 10.1016/j.chom.2020.12.012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Penumutchu S., Korry B.J., Hewlett K., Belenky P. Fiber supplementation protects from antibiotic-induced gut microbiome dysbiosis by modulating gut redox potential. Nat. Commun. 2023;14:5161. doi: 10.1038/s41467-023-40553-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] 15.Cummings J.H., Pomare E.W., Branch W.J., Naylor C.P., Macfarlane G.T. Short chain fatty acids in human large intestine, portal, hepatic and venous blood. Gut. 1987;28:1221–1227. doi: 10.1136/gut.28.10.1221. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] 16.Hryckowian A.J., Van Treuren W., Smits S.A., Davis N.M., Gardner J.O., Bouley D.M., Sonnenburg J.L. Microbiota-accessible carbohydrates suppress Clostridium difficile infection in a murine model. Nat. Microbiol. 2018;3:662–669. doi: 10.1038/s41564-018-0150-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Pensinger D.A., Fisher A.T., Dobrila H.A., Van Treuren W., Gardner J.O., Higginbottom S.K., Carter M.M., Schumann B., Bertozzi C.R., Anikst V., et al. Butyrate Differentiates Permissiveness to Clostridioides difficile Infection and Influences Growth of Diverse C. difficile Isolates. Infect. Immun. 2023;91:e00570-22. doi: 10.1128/iai.00570-22. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Fachi J.L., Felipe J.D.S., Pral L.P., Da Silva B.K., Corrêa R.O., De Andrade M.C.P., Da Fonseca D.M., Basso P.J., Câmara N.O.S., de Sales E Souza É.L., et al. Butyrate Protects Mice from Clostridium difficile-Induced Colitis through an HIF-1-Dependent Mechanism. Cell Rep. 2019;27:750–761.e7. doi: 10.1016/j.celrep.2019.03.054. [DOI] [PubMed] [Google Scholar]

[bib19] 19.Fachi J.L., De Oliveira S., Trsan T., Penati S., Gilfillan S., Cao S., Ribeiro Castro P., Fernandes M.F., Hyrc K.L., Liu X., et al. Fiber- and acetate-mediated modulation of MHC-II expression on intestinal epithelium protects from Clostridioides difficile infection. Cell Host Microbe. 2025;33:235–251.e7. doi: 10.1016/j.chom.2024.12.017. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] 20.Sorg J.A., Sonenshein A.L. Bile salts and glycine as cogerminants for Clostridium difficile spores. J. Bacteriol. 2008;190:2505–2512. doi: 10.1128/JB.01765-07. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Theriot C.M., Bowman A.A., Young V.B. Antibiotic-Induced Alterations of the Gut Microbiota Alter Secondary Bile Acid Production and Allow for Clostridium difficile Spore Germination and Outgrowth in the Large Intestine. mSphere. 2016;1:e00045-15. doi: 10.1128/mSphere.00045-15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Hewlett K.K., PeBenito A., Hecht A.L., Tiffany C., Tanes C., Glover R.C., Fawad J.A., Friedman E.S., Reynolds J.C., Bittinger K., et al. Dietary Fiber Modulates the Window of Susceptibility to Clostridioides difficile Infection. Gastroenterology. 2025;169:970–982. doi: 10.1053/j.gastro.2025.04.027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.Allegretti J.R., Kearney S., Li N., Bogart E., Bullock K., Gerber G.K., Bry L., Clish C.B., Alm E., Korzenik J.R. Recurrent Clostridium difficile infection associates with distinct bile acid and microbiome profiles. Aliment. Pharmacol. Ther. 2016;43:1142–1153. doi: 10.1111/apt.13616. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Chen X., Katchar K., Goldsmith J.D., Nanthakumar N., Cheknis A., Gerding D.N., Kelly C.P. A Mouse Model of Clostridium difficile–Associated Disease. Gastroenterology. 2008;135:1984–1992. doi: 10.1053/j.gastro.2008.09.002. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Hintze K.J., Benninghoff A.D., Ward R.E. Formulation of the Total Western Diet (TWD) as a basal diet for rodent cancer studies. J. Agric. Food Chem. 2012;60:6736–6742. doi: 10.1021/jf204509a. [DOI] [PubMed] [Google Scholar]

[bib26] 26.Mudgil D., Barak S. Composition, properties and health benefits of indigestible carbohydrate polymers as dietary fiber: A review. Int. J. Biol. Macromol. 2013;61:1–6. doi: 10.1016/j.ijbiomac.2013.06.044. [DOI] [PubMed] [Google Scholar]

[bib27] 27.Jana U.K., Kango N., Pletschke B. Hemicellulose-Derived Oligosaccharides: Emerging Prebiotics in Disease Alleviation. Front. Nutr. 2021;8 doi: 10.3389/fnut.2021.670817. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Jiang T., Gao X., Wu C., Tian F., Lei Q., Bi J., Xie B., Wang H.Y., Chen S., Wang X. Apple-Derived Pectin Modulates Gut Microbiota, Improves Gut Barrier Function, and Attenuates Metabolic Endotoxemia in Rats with Diet-Induced Obesity. Nutrients. 2016;8:126. doi: 10.3390/nu8030126. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Teferra T.F. Possible actions of inulin as prebiotic polysaccharide: A review. Food Front. 2021;2:407–416. doi: 10.1002/fft2.92. [DOI] [Google Scholar]

[bib30] 30.Benjamini Y., Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B Stat. Methodol. 1995;57:289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]

[bib31] 31.Francis M.B., Allen C.A., Sorg J.A. Muricholic Acids Inhibit Clostridium difficile Spore Germination and Growth. PLoS One. 2013;8 doi: 10.1371/journal.pone.0073653. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] 32.Francis M.B., Allen C.A., Shrestha R., Sorg J.A. Bile acid recognition by the Clostridium difficile germinant receptor, CspC, is important for establishing infection. PLoS Pathog. 2013;9 doi: 10.1371/journal.ppat.1003356. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Thanissery R., Winston J.A., Theriot C.M. Inhibition of spore germination, growth, and toxin activity of clinically relevant C. difficile strains by gut microbiota derived secondary bile acids. Anaerobe. 2017;45:86–100. doi: 10.1016/j.anaerobe.2017.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Sorg J.A., Sonenshein A.L. Inhibiting the initiation of Clostridium difficile spore germination using analogs of chenodeoxycholic acid, a bile acid. J. Bacteriol. 2010;192:4983–4990. doi: 10.1128/JB.00610-10. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Bisanz J.E., Upadhyay V., Turnbaugh J.A., Ly K., Turnbaugh P.J. Meta-Analysis Reveals Reproducible Gut Microbiome Alterations in Response to a High-Fat Diet. Cell Host Microbe. 2019;26:265–272.e4. doi: 10.1016/j.chom.2019.06.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] 36.Faith D.P. Conservation evaluation and phylogenetic diversity. Biol. Conserv. 1992;61:1–10. doi: 10.1016/0006-3207(92)91201-3. [DOI] [Google Scholar]

[bib37] 37.Lozupone C., Lladser M.E., Knights D., Stombaugh J., Knight R. UniFrac: an effective distance metric for microbial community comparison. ISME J. 2011;5:169–172. doi: 10.1038/ismej.2010.133. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Sinha S.R., Haileselassie Y., Nguyen L.P., Tropini C., Wang M., Becker L.S., Sim D., Jarr K., Spear E.T., Singh G., et al. Dysbiosis-Induced Secondary Bile Acid Deficiency Promotes Intestinal Inflammation. Cell Host Microbe. 2020;27:659–670.e5. doi: 10.1016/j.chom.2020.01.021. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] 39.Kakiyama G., Pandak W.M., Gillevet P.M., Hylemon P.B., Heuman D.M., Daita K., Takei H., Muto A., Nittono H., Ridlon J.M., et al. Modulation of the fecal bile acid profile by gut microbiota in cirrhosis. J. Hepatol. 2013;58:949–955. doi: 10.1016/j.jhep.2013.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] 40.Marion S., Studer N., Desharnais L., Menin L., Escrig S., Meibom A., Hapfelmeier S., Bernier-Latmani R. In vitro and in vivo characterization of Clostridium scindens bile acid transformations. Gut Microbes. 2019;10:481–503. doi: 10.1080/19490976.2018.1549420. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib41] 41.Lakatos P.L., Kiss L.S., Palatka K., Altorjay I., Antal-Szalmas P., Palyu E., Udvardy M., Molnar T., Farkas K., Veres G., et al. Serum lipopolysaccharide-binding protein and soluble CD14 are markers of disease activity in patients with Crohn’s disease. Inflamm. Bowel Dis. 2011;17:767–777. doi: 10.1002/ibd.21402. [DOI] [PubMed] [Google Scholar]

[bib42] 42.Pastor Rojo O., López San Román A., Albéniz Arbizu E., de la Hera Martínez A., Ripoll Sevillano E., Albillos Martínez A. Serum lipopolysaccharide-binding protein in endotoxemic patients with inflammatory bowel disease. Inflamm. Bowel Dis. 2007;13:269–277. doi: 10.1002/ibd.20019. [DOI] [PubMed] [Google Scholar]

[bib43] 43.Denny J.E., Flores J.N., Mdluli N.V., Abt M.C. Standard mouse diets lead to differences in severity in infectious and non-infectious colitis. mBio. 2025;16 doi: 10.1128/mbio.03302-24. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] 44.Tabung F.K., Birmann B.M., Epstein M.M., Martínez-Maza O., Breen E.C., Wu K., Giovannucci E.L. Influence of Dietary Patterns on Plasma Soluble CD14, a Surrogate Marker of Gut Barrier Dysfunction. Curr. Dev. Nutr. 2017;1 doi: 10.3945/cdn.117.001396. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] 45.Smith A.B., Jenior M.L., Keenan O., Hart J.L., Specker J., Abbas A., Rangel P.C., Di C., Green J., Bustin K.A., et al. Enterococci enhance Clostridioides difficile pathogenesis. Nature. 2022;611:780–786. doi: 10.1038/s41586-022-05438-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] 46.Ulrich R.J., Santhosh K., Mogle J.A., Young V.B., Rao K. Is Clostridium difficile infection a risk factor for subsequent bloodstream infection? Anaerobe. 2017;48:27–33. doi: 10.1016/j.anaerobe.2017.06.020. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib47] 47.Trunfio M., Scabini S., Rugge W., Bonora S., Di Perri G., Calcagno A. Concurrent and Subsequent Co-Infections of Clostridioides difficile Colitis in the Era of Gut Microbiota and Expanding Treatment Options. Microorganisms. 2022;10:1275. doi: 10.3390/microorganisms10071275. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] 48.Pensinger D.A., Dobrila H.A., Stevenson D.M., Hryckowian N.D., Amador-Noguez D., Hryckowian A.J. Exogenous butyrate inhibits butyrogenic metabolism and alters virulence phenotypes in Clostridioides difficile. mBio. 2024;15:e02535-23. doi: 10.1128/mbio.02535-23. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] 49.Gregory A.L., Pensinger D.A., Hryckowian A.J. A short chain fatty acid–centric view of Clostridioides difficile pathogenesis. PLoS Pathog. 2021;17 doi: 10.1371/journal.ppat.1009959. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib50] 50.Hasegawa M., Kamada N., Jiao Y., Liu M.Z., Núñez G., Inohara N. Protective Role of Commensals against Clostridium difficile Infection via an IL-1β–Mediated Positive-Feedback Loop. J. Immunol. 2012;189:3085–3091. doi: 10.4049/jimmunol.1200821. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] 51.Hasegawa M., Yada S., Liu M.Z., Kamada N., Muñoz-Planillo R., Do N., Núñez G., Inohara N. Interleukin-22 regulates the complement system to promote resistance against pathobionts after pathogen-induced intestinal damage. Immunity. 2014;41:620–632. doi: 10.1016/j.immuni.2014.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] 52.Yang W., Yu T., Huang X., Bilotta A.J., Xu L., Lu Y., Sun J., Pan F., Zhou J., Zhang W., et al. Intestinal microbiota-derived short-chain fatty acids regulation of immune cell IL-22 production and gut immunity. Nat. Commun. 2020;11:4457. doi: 10.1038/s41467-020-18262-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib53] 53.Bowers E., Entrup G.P., Islam M., Mohan R., Lerner A., Mancuso P., Moore B.B., Singer K. High fat diet feeding impairs neutrophil phagocytosis, bacterial killing, and neutrophil-induced hematopoietic regeneration. J. Immunol. 2025;214:680–693. doi: 10.1093/jimmun/vkaf024. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib54] 54.Ng K.M., Ferreyra J.A., Higginbottom S.K., Lynch J.B., Kashyap P.C., Gopinath S., Naidu N., Choudhury B., Weimer B.C., Monack D.M., Sonnenburg J.L. Microbiota-liberated host sugars facilitate post-antibiotic expansion of enteric pathogens. Nature. 2013;502:96–99. doi: 10.1038/nature12503. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] 55.Callahan B.J., McMurdie P.J., Rosen M.J., Han A.W., Johnson A.J.A., Holmes S.P. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods. 2016;13:581–583. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib56] 56.Janssen S., McDonald D., Gonzalez A., Navas-Molina J.A., Jiang L., Xu Z.Z., Winker K., Kado D.M., Orwoll E., Manary M., et al. Phylogenetic Placement of Exact Amplicon Sequences Improves Associations with Clinical Information. mSystems. 2018;3:e00021-18. doi: 10.1128/mSystems.00021-18. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib57] 57.Wang Q., Garrity G.M., Tiedje J.M., Cole J.R. Naïve Bayesian Classifier for Rapid Assignment of rRNA Sequences into the New Bacterial Taxonomy. Appl. Environ. Microbiol. 2007;73:5261–5267. doi: 10.1128/AEM.00062-07. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] 58.Wickham H., Averick M., Bryan J., Chang W., McGowan L., François R., Grolemund G., Hayes A., Henry L., Hester J., et al. Welcome to the Tidyverse. J. Open Source Softw. 2019;4:1686. doi: 10.21105/joss.01686. [DOI] [Google Scholar]

[bib59] 59.Kelly C.J., Alexeev E.E., Farb L., Vickery T.W., Zheng L., Eric L C., Kitzenberg D.A., Battista K.D., Kominsky D.J., Robertson C.E., et al. Oral vitamin B 12 supplement is delivered to the distal gut, altering the corrinoid profile and selectively depleting Bacteroides in C57BL/6 mice. Gut Microbes. 2019;10:654–662. doi: 10.1080/19490976.2019.1597667. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib60] 60.Sarafian M.H., Lewis M.R., Pechlivanis A., Ralphs S., McPhail M.J.W., Patel V.C., Dumas M.-E., Holmes E., Nicholson J.K. Bile Acid Profiling and Quantification in Biofluids Using Ultra-Performance Liquid Chromatography Tandem Mass Spectrometry. Anal. Chem. 2015;87:9662–9670. doi: 10.1021/acs.analchem.5b01556. [DOI] [PubMed] [Google Scholar]

[bib61] 61.Yilmaz P., Parfrey L.W., Yarza P., Gerken J., Pruesse E., Quast C., Schweer T., Peplies J., Ludwig W., Glöckner F.O. The SILVA and “All-species Living Tree Project (LTP)” taxonomic frameworks. Nucleic Acids Res. 2014;42:D643–D648. doi: 10.1093/nar/gkt1209. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib62] 62.Quast C., Pruesse E., Yilmaz P., Gerken J., Schweer T., Yarza P., Peplies J., Glöckner F.O. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–D596. doi: 10.1093/nar/gks1219. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib63] 63.Lozupone C.A., Hamady M., Kelley S.T., Knight R. Quantitative and Qualitative β Diversity Measures Lead to Different Insights into Factors That Structure Microbial Communities. Appl. Environ. Microbiol. 2007;73:1576–1585. doi: 10.1128/AEM.01996-06. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib64] 64.Mölder F., Jablonski K.P., Letcher B., Hall M.B., van Dyken P.C., Tomkins-Tinch C.H., Sochat V., Forster J., Vieira F.G., Meesters C., et al. Sustainable data analysis with Snakemake. F1000Res. 2021;10:33. doi: 10.12688/f1000research.29032.2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib65] 65.Anaconda Software (2012). (Anaconda Inc.).

[bib66] 66.University of Colorado Boulder Research Computing Alpine. 2023. [DOI]

[bib67] 67.Centers for Disease Control and Prevention . Centers for Disease Control and Prevention; 2024. Emerging Infections Program, Healthcare-Associated Infections – Community Interface Surveillance Report: Clostridioides Difficile Infection (CDI), 2022. [Google Scholar]

[bib68] 68.Edwards A.N., McBride S.M. Isolating and Purifying Clostridium difficile Spores. Methods Mol. Biol. 2016;1476:117–128. doi: 10.1007/978-1-4939-6361-4_9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib69] 69.Barthel M., Hapfelmeier S., Quintanilla-Martínez L., Kremer M., Rohde M., Hogardt M., Pfeffer K., Rüssmann H., Hardt W.-D. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host. Infect. Immun. 2003;71:2839–2858. doi: 10.1128/IAI.71.5.2839-2858.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib70] 70.Dieleman L.A., Palmen M.J., Akol H., Bloemena E., Peña A.S., Meuwissen S.G., Van Rees E.P. Chronic experimental colitis induced by dextran sulphate sodium (DSS) is characterized by Th1 and Th2 cytokines. Clin. Exp. Immunol. 1998;114:385–391. doi: 10.1046/j.1365-2249.1998.00728.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib71] 71.Bolyen E., Rideout J.R., Dillon M.R., Bokulich N.A., Abnet C.C., Al-Ghalith G.A., Alexander H., Alm E.J., Arumugam M., Asnicar F., et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 2019;37:852–857. doi: 10.1038/s41587-019-0209-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib72] 72.Kassambara A., Kosinski M., Biecek P. Survminer: Drawing Survival Curves Using Ggplot2. Version 0.4.9. 2020. https://rpkgs.datanovia.com/survminer/index.html [Google Scholar]

[bib73] 73.Therneau T.M. Survival: Survival Analysis. Version 3.5-8. 2024. https://github.com/therneau/survival [Google Scholar]

[bib74] 74.Oksanen J., Simpson G.L., Blanchet F.G., Kindt R., Legendre P., Minchin P.R., O’Hara R.B., Solymos P., Stevens M.H.H., Szoecs E., et al. Vegan: Community Ecology Package. 2025. https://vegandevs.github.io/vegan/ [Google Scholar]

[bib75] 75.Paradis E., Schliep K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics. 2019;35:526–528. doi: 10.1093/bioinformatics/bty633. [DOI] [PubMed] [Google Scholar]

[bib76] 76.Wickham H. Second edition. Springer; 2016. ggplot2: Elegant Graphics for Data Analysis. [DOI] [Google Scholar]

[bib77] 77.Ahlmann-Eltze C., Patil I. ggsignif: R Package for Displaying Significance Brackets for “ggplot2”. 2021. (PsyArXiv) [DOI]

[bib78] 78.Wilke C. cowplot: Streamlined Plot Theme and Plot Annotations for “ggplot2” Version R package version 1.1.3. 2024. https://wilkelab.org/cowplot/ [Google Scholar]

[bib79] 79.Kassambara A. rstatix: Pipe-Friendly Framework for Basic Statistical Tests. Version R package version 0.7.2. 2023. https://rpkgs.datanovia.com/rstatix/ [Google Scholar]

[bib80] 80.Li X., Zheng R., Zhang T., Zeng Z., Li H., Liu J. Association between blood urea nitrogen and 30-day mortality in patients with sepsis: a retrospective analysis. Ann. Palliat. Med. 2021;10:11653–11663. doi: 10.21037/apm-21-2937. [DOI] [PubMed] [Google Scholar]

[bib81] 81.Castro M., Silver H.J., Hazleton K., Lozupone C., Nicholson M.R. The Impact of Diet on Clostridioides difficile Infection: A Review. J. Infect. Dis. 2025;231:e1010–e1018. doi: 10.1093/infdis/jiaf233. jiaf233. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib82] 82.Falcone M., Russo A., Iraci F., Carfagna P., Goldoni P., Vullo V., Venditti M. Risk Factors and Outcomes for Bloodstream Infections Secondary to Clostridium difficile Infection. Antimicrob. Agents Chemother. 2016;60:252–257. doi: 10.1128/AAC.01927-15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib83] 83.Falcone M., Venditti M., Sanguinetti M., Posteraro B. Management of candidemia in patients with Clostridium difficile infection. Expert Rev. Anti Infect. Ther. 2016;14:679–685. doi: 10.1080/14787210.2016.1197118. [DOI] [PubMed] [Google Scholar]

[bib84] 84.Ianiro G., Murri R., Sciumè G.D., Impagnatiello M., Masucci L., Ford A.C., Law G.R., Tilg H., Sanguinetti M., Cauda R., et al. Incidence of Bloodstream Infections, Length of Hospital Stay, and Survival in Patients With Recurrent Clostridioides difficile Infection Treated With Fecal Microbiota Transplantation or Antibiotics. Ann. Intern. Med. 2019;171:695–702. doi: 10.7326/M18-3635. [DOI] [PubMed] [Google Scholar]

[bib85] 85.Thompson L.R., Sanders J.G., McDonald D., Amir A., Ladau J., Locey K.J., Prill R.J., Tripathi A., Gibbons S.M., Ackermann G., et al. A communal catalogue reveals Earth’s multiscale microbial diversity. Nature. 2017;551:457–463. doi: 10.1038/nature24621. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Dietary fiber reduces mortality from secondary sepsis in a murine model of Clostridioides difficile infection

Madison Apgar

Elena S Wall

Monica Midura-Kiela

Charles P Neff

David J Orlicky

Laurie M Lyon

Claudio Bernardazzi

Casey G Martin

Oren M Gordon

Issa Jimenez-Espinoza

Jacob K Flynn

Catherine A Lozupone

Keith Z Hazleton

Summary

Graphical abstract

Highlights

Introduction

Results

Dietary fiber protects against mouse mortality in antibiotic-associated CDI

Figure 1.

Intestinal inflammation correlates with C. difficile toxin and is lowest in the high-fat/low-fiber diet at 3 days post infection

Figure 2.

Cecal levels of C. difficile toxins and inflammation have a complex relationship with the microbiome-produced metabolites, DCA and butyrate, across diets

Figure 3.

Mice on low-fiber diets demonstrate greater loss of diversity compared to mice on chow and high-fiber diets

Figure 4.

Figure 5.

Mortality was accompanied by disseminated infection by gut-colonizing pathobionts

Figure 6.

Plasma markers of inflammation and kidney damage point to sepsis as a plausible cause of mortality post-CDI

Figure 7.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Murine model of antibiotic-associated CDI

University of Arizona mouse experiments

AMC mouse experiments

Method details

Mouse diets

C. difficile growth

DNA extraction and sequencing

Toxin quantification

Histopathologic evaluation of large intestinal tissue

SCFA quantification

Bile acid quantification

Reagents

Standards preparation

Sample preparation

High-performance liquid chromatography/quadrupole time-of-flight mass spectrometry (HPLC/QTOF)

Extraintestinal bacterial culturing

Plasma biomarker quantification

Quantification and statistical analysis

Data analysis

16S rDNA sequencing data

Lactococcus contamination filtering steps

Statistics and plots

Additional resources

Footnotes

Contributor Information

Supplemental information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles