After Institutional Review Board (IRB) protocol approval (Protocol number: 21079), deidentified intestinal surgical resection specimens in paraffin blocks were obtained from the biorepository and tissue research facility of the University of Virginia (UVA, USA).

Fecal samples from patients with diarrhea caused by CDI and non-CDI were obtained from the clinical microbiology laboratory of University of Virginia after Institutional Review Board (IRB) protocol approval (Protocol number: 20813).

A total of 48 male C57BL/6 mice (Jackson Laboratory, Farmington, USA, 8 weeks of age) were housed in temperature-controlled rooms under 12-h light–dark cycles. The animals received water and food ad libitum. All surgical procedures and treatments performed with C57BL/6 mice were conducted in accordance with the Guidelines for Institutional and Animal Care and Use of the University of Virginia, Charlottesville, USA. The protocol has been approved by the committee on the Ethics of Animal Experiments of the University of Virginia (Protocol number: 4096).

Our murine CDI model was established as previously described (Chen X. et al., 2008; Moore et al., 2015; Shin et al., 2017; Broecker et al., 2019; Schmidt et al., 2020). This model is broadly used due to the ability to mimic clinical symptoms, such as severe diarrhea, presented by humans with CDI (Chen X. et al., 2008; Moore et al., 2015; Shin et al., 2017; Broecker et al., 2019; Schmidt et al., 2020). C57BL/6 mice (n = 5 for each group) received an antibiotic treatment in the drinking water for 3 days. The antibiotic cocktail contained gentamicin (0.035 mg per ml), colistin (850 U per ml), metronidazole (0.215 mg per ml), and vancomycin (0.045 mg per ml). After 1 day off antibiotics, an intraperitoneal injection of clindamycin (32 mg per kg) was given 1 day before C. difficile challenge. Then, 10⁵ CFU (in 100 μl of chopped meat broth, a pre-reduced medium) of the vegetative C. difficile strain VPI10463 (ATCC 43255, tcdA+tcdB+cdtB-) was administered by oral gavage. Control mice received chopped meat broth (100 μl). Mouse weights and the development of disease symptoms were monitored daily. Animals that became moribund or lost >20% of their body weight were euthanized.

For the S100B inhibitor treatment, mice were injected intraperitoneally daily once a day, starting 1 h prior C. difficile inoculation, with pentamidine (4, 10, or 40 mg/kg, Sigma-Aldrich, P0547) or saline solution.

The immortalized rat enteric glial cell (EGC) line PK060399egfr (ATCC CRL-2690, VA, USA) was cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) and supplemented with 10% fetal bovine serum, 1% antibiotics (100 μg/ml penicillin and 100 μg/ml streptomycin, Gibco), and 1 mM sodium pyruvate (Gibco) at 37°C in a humidified incubator under 5% CO₂ for no more than 16 passages. For all experiments, EGCs were released using 0.05% trypsin-EDTA for 5 min.

Cells were incubated with TcdA (10, 50, or 100 ng/ml) or TcdB (0.1, 1, or 10 ng/ml) for 0.5, 2, 6, 12, 18, 24, and 48 h to establish the best time point to study each parameter. For some mechanistic studies, cells were incubated with 30 μM FPSZM1 (TOCRIS, 6237), a high-affinity antagonist of RAGE, and a PI3K inhibitor (1 and 10 μM LY294002, TOCRIS, 1130) 1 h before incubation with TcdA or TcdB. All drug concentrations used were based on MTT assay results, as shown in Additional file (Additional File: Figure S1).

Purified TcdA and TcdB, produced by C. difficile strain VPI10463, were obtained from TechLab (VA, USA).

The EGC lineage (EGC/PK060399) has been shown to exhibit similar morphology and functional properties to primary enteric glial cells (Ruhl et al., 2001). This EGC lineage expresses GFAP and S100B, which are enteric glial factors, and has been applied in a variety of mechanistic studies (Bauman et al., 2017; Fettucciari et al., 2017; Macchioni et al., 2017; Chen et al., 2018; Fettucciari et al., 2018; Kimono et al., 2019).

Sections (4 µm thick) were prepared from paraffin-embedded mouse and human colonic tissues. After deparaffinization, antigens were recovered by incubating the slides in citrate buffer (pH 6.0) for 20 min at 95°C. Endogenous peroxidase was blocked with 3% H₂O₂ for 10 min to reduce non-specific binding. The sections were then incubated with an S100B antibody (NBP2-54426, Novus Biologicals, 1:1,000) overnight. The sections were then incubated for 30 min with polymer (K4061, Dako). The antibody-binding sites were visualized by incubating the samples with diaminobenzidine–H₂O₂ (DAB, Dako) solution. Sections incubated with antibody diluent without a primary antibody were used as negative controls. Antibody specificity was evaluated using positive controls for S100B in the mouse cerebellum (data not shown). The immunostaining for S100B in human samples was performed at the UVA Biorepository and Tissue Research Facility.

The amounts of DAB products after immunostaining were estimated from at least 6–8 and 15–20 digital images from different areas of each section (from four specimens per group) for mouse and human samples, respectively, at ×400 magnification using Adobe Photoshop software. The percentage of immunopositive area was calculated by dividing the number of DAB-positive pixels (immunostaining-positive pixels) by the number of pixels per total tissue image and multiplying the result by 100, as previously described (Costa et al., 2019).

Proteins from colon tissue from CDI-infected and control mice at days 1 and 3 p.i. were extracted by lysing the tissues using RIPA lysis buffer (supplemented with complete EDTA-free protease inhibitor cocktail and PhosSTOP, Sigma-Aldrich), followed by a step of centrifugation (17 min, 4°C, 13,000 rpm) and collection of the supernatant. Protein concentrations were determined through the bicinchoninic acid assay according to the manufacturer’s protocol (Thermo Fisher Scientific). Reduced 40-µg protein samples (prepared with a sample reducing agent—Invitrogen—and a protein loading buffer—LI-COR) were denatured at 99°C for 5 min, separated on NuPAGE 4%–12% BIS-Tris gel (Invitrogen), and transferred to nitrocellulose membranes (Life Technologies) for 2 h. The membranes were then immersed in iBind fluorescent detection solution (Life technologies) and placed in an iBind automated Western device (Life Technologies) overnight at 4°C for blocking and incubating with primary antibodies (mouse anti-α-tubulin, 1:2000, Sigma-Aldrich; rabbit anti-S100B, 1:500, Novus Biologicals, NBP2-54426) and secondary antibodies (Cy3-conjugated AffiniPure donkey anti-rabbit, 711-165-152, 1:1000, Jackson ImmunoResearch, and Cy5-conjugated AffiniPure donkey anti-mouse, 715-175-150, 1:1000, Jackson ImmunoResearch). Then, the membranes were immersed in ultrapure water and fluorescent signal was detected using the Typhoon system (GE Healthcare). Densitometric quantification of bands was performed using ImageJ software (NIH, Bethesda, MD, USA).

For stool shedding of C. difficile, DNA was extracted from stools using the QIAamp DNA Stool Mini Kit (Qiagen) according to the manufacturer’s instructions. The tcdB (C. difficile toxin B) gene was used as a specific target for detecting C. difficile in stools after CDI. Primer sequences included tcdB 5′-AATGCATTTTTGATAAACACATTG-3′ (forward) and 5′-AAGTTTCTAACATCATTTCCAC-3′ (reverse). Real-time PCR was performed using Bio-Rad CFX under the following conditions: 95°C for 3 min, followed by 40 cycles of 15 s at 95°C, 60 s at 55°C, and lastly 20 s at 72°C.

Diarrhea, an important CDI outcome, was assessed at days 2 and 3 p.i. Diarrhea scores were based on the following 0 to 3: 0—well-formed pellets; 1—stick stools adhering in microtube wall or color change (yellow); 2—pasty stools with or without mucus; and 3—watery stools, as previously described (Warren et al., 2013) with some modifications.

Mouse cecum and colon tissues from day 3 p.i. were fixed in 10% neutral buffered formalin for 20 h, dehydrated, and embedded in paraffin. Cecal and colonic sections (5 µm) were then stained with hematoxylin and eosin staining (H&E) and examined using light microscopy. Histopathological scores were performed by a blinded investigator, using a previously described method (Ledwaba et al., 2020) with some modifications. Histopathological scores were determined by quantifying the intensity of epithelial tissue damage (0–3, 0—no damage, 1—mild, 2—moderate, 3—extensive), edema in mucosa/submucosa layer (0–3), and cell infiltration (0–3). The total histological damage score was measured by the sum of the three parameters evaluated.

Protein lysates were extracted from the cecum contents, cecum, and colon tissues using radioimmunoprecipitation assay (RIPA) buffer (20 mM Tris, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1 mM EDTA, 0.1% SDS, adjusted to pH 7.5) containing protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitors (Sigma-Aldrich). Lysates were centrifuged at 13,000 rpm for 15 min, and the supernatant was used to perform the protein assay using the bicinchoninic acid assay (Thermo Fisher Scientific). Inflammatory biomarkers (MPO, IL-23, IL-22, IL-17, GMCSF, and IL-33) were measured using a commercial ELISA kit (R&D Systems) according to the manufacturer’s instructions. The absorbance (450 nm) was determined using an Epoch plate reader (BioTek). Interleukin-6 (IL-6), IL-1β, TNF-α, IL-18, and IL-2 were measured using a ProcartaPlex multiplex immunoassay (Invitrogen) by Luminex (Bio-Rad). Levels of these cytokines were measured as picograms per milligram of protein.

The isolation of total RNA from colon tissues of CDI-infected and control mice was performed by using a Qiagen RNeasy Mini Kit and QIAcube. cDNA was synthetized from 1 µg of total RNA, quantified by Qubit 3 Fluorometer 3000 (Invitrogen), and purified by deoxyribonuclease I (Invitrogen) treatment, with the iScript cDNA (Bio-Rad) as described by the manufacturer’s instructions. qPCR was performed with 50 ng of cDNA in each well and SensiFAST Probe No-ROX Mix (Bioline) using a CFX Connect system (Bio-Rad) with the following conditions: 95°C for 2 min, 40 cycles of 95°C for 10 s, and 60°C for 50 s. A predesigned TaqMan array mouse immune fast 96-well plate (Applied Biosystems) was used to assess the expression of all the genes shown in Figures 4A, B. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a reference gene. All fold changes were determined using the ΔΔC_t method (Livak and Schmittgen, 2001).

EGC lines (5×10³ cells/well) were seeded in 96-well plates and treated with TcdA or TcdB for different incubation periods. Then, the cells were incubated with thiazolyl blue tetrazolium bromide (MTT, 0.5 mg/ml reconstituted in supplemented DMEM, Sigma-Aldrich, M2128) for 2 h at 37°C in a humidified incubator under 5% CO₂. After removal of the MTT solution, 150 µl of dimethylsulfoxide was added to each well. The plates were then shaken for 2 min at room temperature, and the absorbance of the reaction at 570 nm was measured using an ELISA reader.

EGC lines (5×10³ cells/well) were seeded in 96-well plates and treated with TcdA or TcdB for different incubation periods. The percentage of rounded cells was measured among 100 cells from the center of the well, and rounded versus non-rounded cells were discriminated using contrast microscopy.

Fecal samples were homogenized with RIPA buffer (1:2), and the proteins were obtained by centrifuging (10,000 rpm, 10 min, 4°C). Levels of S100B were measured with a DuoSet S100B kit (R&D Systems) by ELISA according to the manufacturer’s protocol. The absorbance (450 nm) was determined using an Epoch plate reader (BioTek). The range of S100B detection was 46.9–3,000 pg per mg of protein.

EGCs line (6×10⁵ cells/well) were seeded in six-well plates and treated with TcdA or TcdB for different times. EGC supernatants were collected (stored at -80°C until use) and centrifuged (10,000 rpm, 10 min, 4°C), and secreted S100B was measured with a DuoSet S100B kit (R&D Systems) by ELISA according to the manufacturer’s protocol. The absorbance (450 nm) was determined using an Epoch plate reader (BioTek). The range of S100B detection was 46.9–3,000 pg/ml.

EGC lines (6×10⁵ cells/well) were seeded in six-well plates and treated with TcdA or TcdB and pharmacologic modulators. After incubation, total RNA was extracted with an RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) using QIAcube (Qiagen). RNA was quantified with a Qubit 3.0 fluorometer (Life Technologies) using a Qubit RNA BR Assay Kit (Invitrogen, Q10211). After DNA contamination was removed by RNA treatment with DNase I (Invitrogen, 18068-015), a total of 600 ng of RNA was then reverse transcribed using an iScript cDNA Synthesis Kit (Bio-Rad, 1708891) according to the manufacturer’s protocol. qPCR amplification of S100B, IL-6, RAGE, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) in cell samples was performed in a CFX Connect system (Bio-Rad) with the following conditions: 95°C for 30 s, 40 cycles of 95°C for 5 s and 60°C for 30 s, and melt curve analysis from 65 to 95°C in 0.5°C increments for 2 s each. All PCRs were performed with iTaq Universal SYBR Green Supermix (Bio-Rad, 172-5124). The primer sets are listed in Additional File: Table S1.

EGC lines (4×10⁴ cells/well) plated on eight-chamber glass tissue culture slides in a polystyrene vessel and treated with TcdA or TcdB for 18 h were fixed in 4% PFA solution (Alfa Aesar) in PBS for 30 min at room temperature and permeabilized with 0.5% Triton X-100 (Sigma-Aldrich) and 3% bovine serum albumin (BSA, Sigma) in PBS for 10 min at 4°C. After blocking with 5% normal donkey serum (Jackson ImmunoResearch, 017-000-121) in PBS for 30 min at room temperature, the cells were incubated with anti-NFκBp65 (Santa Cruz Biotechnology, sc-372) overnight at 4°C. After three washes with washing buffer (0.01% Tween 20 in PBS), the cells were incubated for 1 h with Cy3-conjugated AffiniPure donkey anti-rabbit (Jackson ImmunoResearch, 711-165-152) antibody, washed with PBS, and mounted with ProLong Gold antifade reagent containing DAPI (Thermo Scientific, P36931). The samples were visualized by fluorescence microscopy (Zeiss). For each experimental condition, 100 cells were counted, and the percentage of cells (%) with positive nuclear staining for NFκBp65 was determined.

EGC lines (6×10⁵ cells/well) were seeded in six-well plates and treated with TcdA or TcdB and pharmacologic modulators. After incubation, the supernatant was removed, and the nuclear extract was obtained by using a Nuclear Extract Kit (Thermo scientific) according to the manufacturer’s protocol. Protein concentrations were determined through the bicinchoninic acid assay according to the manufacturer’s protocol (Thermo Fisher Scientific). Reduced 15-µg protein samples (previously prepared with a sample reducing agent—Invitrogen NP0009—and a protein loading buffer—LI-COR 928-40004) were denatured at 99°C for 5 min, separated on NuPAGE 10% BIS-Tris gel (Invitrogen, NP0322BOX), and transferred to nitrocellulose membranes (Life Technologies, LC2000 or LC2006) for 2 h. After blocking with 5% non-fat dry milk at 4°C for 1 h, the membranes were incubated overnight with primary antibodies (mouse PCNA, 1:200, Cell Signaling; rabbit anti-NFκBp65, sc-372, 1:500, Santa Cruz Biotechnology; mouse laminin B1, sc-365962, 1:100, Santa Cruz Biotechnology; rabbit anti-phosphorylated NFκBp65, PA5-37718, 1:500, Thermo Scientific) and secondary antibodies (Cy3-conjugated AffiniPure donkey anti-rabbit, 711-165-152, 1:1000, Jackson ImmunoResearch; Cy5-conjugated AffiniPure donkey anti-mouse, 715-175-150, 1:1000, Jackson ImmunoResearch) for 1 h and 30 min. The membranes were washed in Tris-buffered saline (TBS) containing 0.05% Tween 20 (TSB-T), and the fluorescent signal was detected using the Typhoon system (GE Healthcare). Densitometric quantification of bands was performed using ImageJ software (NIH, Bethesda, MD, USA).

Analyses were performed using GraphPad software 9.0 (San Diego, CA, USA). The data are presented as the mean ± standard error of the mean (SEM). Student’s t test or one- or two-way analysis of variance (ANOVA) followed by the Bonferroni test was used to compare means. p < 0.05 was considered to indicate significance.

Increased S100B immunostaining was observed in colon mucosal, submucosal, and myenteric plexus tissues from patients with CDI (Figure 1A), and these results were verified by increased percentages of S100B-immunopositive cells in patients with CDI compared to control subjects (p < 0.0001, Figure 1B). In addition, we found increased S100B in fecal supernatants from patients with diarrhea associated with CDI compared to diarrhea not associated with CDI (Figure 1C). When age and gender were analyzed, we found that age 40–59 years and female gender had the statistically significant increases in levels of S100B among these patients with diarrhea associated with CDI compared to diarrhea not associated with CDI (Additional File: Figures S2A, B). In addition, no statistical difference was found between females and males with diarrhea associated with CDI (Additional File: Figure S2B).

Given that CDI murine models (Figure 1D) are important for investigating the pathogenesis of C. difficile-induced disease, we analyzed colonic tissues for S100B protein expression in this experimental model on days 1 and 3 postinfection (p.i.). We found that CDI increased the levels of colonic S100B on day 1 (p = 0.009) and day 3 (p = 0.02) p.i. compared to uninfected mice (Figures 1E, F).

Like patients with CDI, we found increased S100B in the colonic mucosa, submucosa, and myenteric plexus in mice with CDI on day 3 p.i. (Figure 1G). Immunostaining showed that CDI markedly enhanced the S100B protein expression in the colon of mice with CDI compared to the control group (p < 0.0001, Figure 1H).

These data demonstrated the association of elevated S100B expression with CDI.

Since S100B levels are increased during CDI in mice and humans, we next sought the influence of this mediator in C. difficile shedding, diarrhea, and intestinal damage. We used pentamidine, a known S100B inhibitor (McKnight et al., 2012; Costa et al., 2019), as a pharmacologic blocker to inhibit S100B activity. Pentamidine at 40 mg/kg was selected based on a dose-ranging study (Additional File: Figure S3). As shown in Figure 2A, S100B inhibition did not affect C. difficile shedding in stools from infected mice. As we have previously shown (Costa et al., 2019), pentamidine treatment does not prevent the initial weight loss from acute infection (Additional File: Figure S4A). However, S100B inhibition decreased diarrhea severity on days 2 (p = 0.01) and 3 (p = 0.04) p.i. (Figure 2B and Additional File: Figure S4B), as well as on days 4 (p = 0.007) and 6 (p = 0.01) p.i. (Additional File: Figure S4B). Of note, all non-treated infected mice developed severe diarrhea whereas most of S100B inhibitor-treated infected mice exhibited only mild to moderate diarrhea (Figure 2B).

By histopathology, CDI resulted in an extensive damage of the epithelial layer, edema, and intense inflammatory cell infiltration in the cecum and colon (Figure 2C), resulting in higher histopathologic scores (median of score = 9, Figure 2D). S100B inhibition decreased intestinal tissue damage in both cecum and colon during CDI, resulting in reduction of histopathologic scores (median of score = 1, Figures 2C, D).

In addition, we found that inhibition of S100B decreased the MPO levels in cecum content (p = 0.001), cecum (p = 0.001), and colon (p = 0.01) tissues during CDI compared to non-treated infected mice (Figure 2E), indicating a reduction in neutrophil recruitment during CDI.

Taken together, these findings indicate that the S100B inhibition-mediated protective effect in preventing diarrhea and intestinal damage occurs via controlling host inflammatory response and not by decreasing pathogen burden.

S100B is an important glial factor that exhibits a dual effect in the gut, stimulating the survival or death of enteric neurons and functioning as a proinflammatory cytokine during the inflammatory response. We investigated whether S100B regulates the synthesis of IL-1β, IL-18, IL-6, GMCSF, TNF-α, IL-17, IL-23, IL-2, and IL-22, all mediators involved in CDI pathogenesis, by pharmacologic blockage using pentamidine. We found that S100B inhibition decreased the colonic levels of IL-1β, IL-18, IL-6, GMCSF, TNF-α, IL-17, IL-23, and IL-2, but not IL-33, during CDI (p < 0.05, Figures 3A–I). On the other hand, S100B inhibition increased the colonic levels of IL-22 during infection (p = 0.008, Figure 3J).

Using a TaqMan qPCR, we found that blockage of S100B activity downregulated the expression of proinflammatory mediators (IL-1α, IL-1β, IL-6, TNF-α, and iNOS), chemokines (CCL2 and CCL3), chemokine receptors (CCR2 and CCR7), and cellular recruitment-related molecule (SELP, selectin P) transcripts and upregulated the anti-inflammatory (SOCS2, suppressor of cytokine signaling 2) and antiapoptotic (Bcl-2, B-cell lymphoma 2) mediators in infected mice (p < 0.05, Figures 4A–G).

Taken together, these findings suggest that S100B has a modulatory role in the intestinal inflammatory response during CDI.

Because S100B is predominantly secreted by EGCs during normal conditions (Rao et al., 2015), we investigated the effect of C. difficile toxins, TcdA and TcdB, in a rat EGC line (EGC/PK060399), which has been largely used (Bauman et al., 2017; Fettucciari et al., 2017; Macchioni et al., 2017; Chen et al., 2018; Fettucciari et al., 2018; Kimono et al., 2019). We tested varying concentrations of TcdA and TcdB and analyzed cell viability and morphology. We selected TcdA 50 ng/ml and TcdB 1 ng/ml as the optimal concentrations for the succeeding in vitro studies (Additional File: Figure S5).

Given that we found increased S100B in C. difficile-infected human and mice, we investigated whether TcdA and TcdB stimulate the expression and secretion of S100B in the EGC lineage (EGC/PK060399). We found that TcdA and TcdB increased S100B release by EGC (EGC/PK060399) in a time-dependent manner (p < 0.0001, Figures 5A, B). In addition, we found that TcdA and TcdB upregulated S100B mRNA in EGCs after 12 h of incubation (p < 0.0001, Figures 5C, D).

Since most inflammatory mediators have been shown to be increased by C. difficile toxins in several types of cells, such as epithelial cells, neutrophils, macrophages, and T cells (Ng et al., 2010; Cowardin et al., 2015), we investigated whether TcdA and TcdB induce IL-6 transcript expression in EGC (EGC/PK060399). TcdA and TcdB upregulated the expression of IL-6, a severity marker of CDI in humans and mice (Yu et al., 2017), at 12 h of incubation (p < 0.0001, Figures 5E, F).

These findings indicate that C. difficile toxins upregulate the expressions of both IL-6 and S100B in EGC (EGC/PK060399).

To study the signaling pathway involved in S100B-induced IL-6 expression in EGCs, we asked whether the receptor for advanced glycation end products (RAGE), a known receptor for S100B extracellular function, is activated during intoxication. While EGC/PK060399 expressed RAGE, we found that TcdA and TcdB did not alter RAGE expression in this cell line (Figure 6A). However, inhibition of RAGE activation using a RAGE antagonist, FPSZM1, reduced the IL-6 expression induced by TcdA and TcdB (p < 0.0001, Figure 6B). These results suggest that the TcdA- and TcdB-induced IL-6 upregulation in EGC (EGC/PK060399) occurs through RAGE activation.

In vascular inflammatory conditions, RAGE activation leads to NFκB activation, which in turn regulates the transcription of proinflammatory cytokines, such as IL-6 (Brasier, 2010). We found that TcdA and TcdB increased the nuclear translocation of NFκBp65 in EGC (EGC/PK060399) at 18 h of incubation (p < 0.0001, Figures 6C–E). TcdA or TcdB increased nuclear phosphorylated NFκBp65 even at 2 and 4 h in EGCs (Additional File: Figures S6A, B). Since RAGE activation is known to initiate PI3K/AKT signaling, which acts upstream of the NFκB pathway, we inhibited PI3K using LY294002, which accentually decreased nuclear phosphorylated NFκBp65 in EGCs challenged with TcdA or TcdB at 4 h of incubation (Additional File: Figure S7). LY294002 notably reduced the TcdA- and TcdB-induced IL-6 upregulation in EGC (EGC/PK060399) (p < 0.0001, Figure 6F).

Taken together, these results indicate that the RAGE/PI3K/NFκB pathway is involved in the induction of IL-6 expression by TcdA and TcdB in EGC (EGC/PK060399).