Clostridioides difficile major toxins remodel the intestinal epithelia, affecting spore adherence/internalization into intestinal tissue and their association with gut vitronectin

Abstract

The most common cause of healthcare-associated diarrhea and colitis in the U.S., is Clostridioides difficile, a spore-forming pathogen. Two toxins, TcdA and TcdB, are major virulence factors essential for disease manifestations, while C. difficile spores are essential for disease transmission and recurrence. Both toxins cause major damage to the epithelial barrier, trigger massive inflammation, and reshape the microbiome and metabolic composition, facilitating C. difficile colonization. C. difficile spores, essential for transmission and recurrence of the disease, persist adhered and internalized in the intestinal epithelia. Studies have suggested that toxin-neutralization in combination with antibiotic during CDI treatment in humans significantly reduces disease recurrence, suggesting a link between toxin-mediated damage and spore persistence. Here, we show that TcdA/TcdB-intoxication of intestinal epithelial Caco-2 cells leads to remodeling of accessible levels of fibronectin (Fn) and vitronectin (Vn) and their cognate alpha-integrin subunits. While TcdB-intoxication of intestinal tissue had no impact in accessible levels of Fn and Vn, but significantly increased levels of intracellular Vn. We observed that Fn and Vn released to the supernatant readily bind to C. difficile spores in vitro, while TcdB-intoxication of intestinal tissue led to increased association of C. difficile spores with gut Vn. Toxin-intoxication of the intestinal tissue also contributes to increased adherence and internalization of C. difficile spores. However, TcdB-intoxicated ligated loops infected of mice treated with Bezlotoxumanb (monoclonal anti-TcdB antibodies) did not prevent TcdB-mediated increased spore adherence and internalization into intestinal tissue. This study highlights the importance of studying the impact of C. difficile toxins of host tissues has in C. difficile interaction with host surfaces that may contribute to increased persistence and disease recurrence.

Introduction

Frequently related to antibiotic-associated diarrhea (AAD) is Clostridioides difficile, which accounts for approximately 30% of cases¹. C. difficile infections (CDI) have emerged as a significant concern for healthcare systems globally, occasionally leading to severe outbreaks with mortality rates as high as 20%. Most patients typically respond favorably to medical treatments like vancomycin and metronidazole^{1, 2}. However, in contrast to other gastrointestinal infections, CDI exhibits an exceptionally high recurrence rate, ranging from 20% after the first episode, to 40% after the second, and even 60% after the third, potentially leading to the development of more severe symptoms^{1, 3, 4}. In Europe and the USA alone, CDI results in substantial economic losses estimated at around US$ 5 billion annually for the healthcare system^{1, 5}. Addressing recurrent CDI presents a considerable clinical challenge.

In response to a dysbiotic microbiota (commonly caused by antibiotics), ingested or indigenous spores will germinate leading to C. difficile colonization and infection⁶. Two main virulence factors are responsible for the clinical manifestations of the disease, toxins TcdA and TcdB^{7, 8}. Both toxins monoglucosylate host GTPases of the Rho family resulting in downstream cellular changes as tight junction collapse, detachment of cells, and impairment of the intestinal epithelial barrier^{7, 8}. In addition, exposure to toxins triggers intestinal epithelial cells to release several pro-inflammatory cytokines, resulting in attraction of neutrophils and acute mucosal inflammation^{7, 8}. These events culminate in further activation of immune cells, exacerbating inflammation and mucosal damage. In addition to these events, C. difficile initiates a sporulation cycle during infection, resulting in a metabolically dormant spore that is impervious to all known antibiotics and regarded as essential for recurrence of the infection⁹.

Concatenated to the loose and tight mucosal layers in the gastrointestinal tract, is an array of potential interaction sites commonly exploited by enteric pathogens¹⁰. Of these, extracellular matrix (ECM) proteins are widely exploited as binding molecular bridges by pathogens, specifically fibronectin (Fn) and vitronectin (Vn)^{11, 12}. Fn, is a 450 kDa dimeric, multi-modular glycoprotein which can exist in two different forms, plasma and cellular Fn. Plasma Fn is synthetized by hepatocytes and secreted to the blood plasma to levels of 300 – 400 μg/ml¹³. By contrast, cellular Fn is synthetized by various cell types, including fibroblasts, intestinal epithelial cells, endothelial cells, chondrocytes, synovial cells and myocytes^{14, 15}, and in the intestinal epithelial crypt cells¹³. Vn is a multidomain glycoprotein of 75 kDa mainly synthetized in the liver and secreted in the plasma, and to a lesser extent produced in platelets and macrophages¹⁶. Vn is highly abundant in the plasma (0.2 – 0.4 mg/ml) and in the extracellular matrix¹⁶, it interacts with several other molecules to regulate various functions including wound healing, the complement system, cell migration, and adhesion^17–19.

Both, Fn and Vn are implicated in tissue repair. Inflammatory stimuli and intestinal injury derived from chemical-colitis leads to increase expression of intestinal Fn in the mucosal layers^{20, 21}, primarily by intestinal epithelial cells²². Notably, patients with inflammatory bowel disease also exhibit increased Fn levels²³. Although, in experimental-colitis expression of Vn increases and is associated with protection against colitis²⁴. Evidence of Fn-remodeling by C. difficile binary toxin has been documented^{25, 26}; however, it is unclear how both C. difficile major toxins impact levels of Fn and Vn in intestinal epithelial cells.

C. difficile toxins, TcdA and TcdB, have traditionally been considered primary virulence factors essential for clinical manifestation of the disease²⁷. However, the most recently FDA approved intravenous CDI therapeutic, Zinplava, which neutralizes C. difficile TcdB, was found to reduce the rate of recurrence of CDI²⁸. Zinplava is an intravenously administered anti-TcdB monoclonal antibody (Bezlotoxumab) intended to be used in combination with antibiotics^29–32. Bezlotoxumab clinical trials results demonstrate that administration of anti-toxin antibodies correlates with lower rates of CDI recurrence^29–32, suggesting that toxin-mediated damage could be contributing indirectly to C. difficile persistence. Although TcdA/TcdB-intoxication of intestinal epithelial cells results in increased adherence and internalization of C. difficile spores in vitro³³, and increased levels of Fn in the colon during infection^{34, 35}, the underlying mechanisms remain unknown. Consequently, in this work we assessed how TcdA/TcdB-intoxication remodels Fn and Vn, and spore-epithelial cell interactions in vivo, and whether Bezlotoxumab can prevent toxin-mediated remodeling and spore-binding in the intestinal mucosa. Results demonstrate that while TcdA/TcdB-intoxication of intestinal epithelial cells in vitro increased levels of accessible Fn and Vn, this was not the case in an ileal loop murine model. Moreover, confocal micrographs show that while a slight increase in Fn- and Vn-spore interactions were observed at the mucosal layer, these were a small proportion of analyzed spores. Interestingly, TcdB-mediated intoxication led to increased adherence and internalization of C. difficile spores to the intestinal epithelial layer in vivo, interactions that were reduced by administration of Bezlotoxumab.

Results

TcdA/TcdB-intoxication increases accessible Fn and Vn in intestinal epithelial cells.

In prior work, we demonstrated that Fn and Vn are required for spore-entry into intestinal epithelial cells (IECs)³⁶. TcdA/TcdB-intoxication of epithelial cells leads to opening of adherent junctions and increased spore internalization³³. Moreover, since intestinal inflammation and epithelial damage leads to increased intestinal Fn and Vn^{20, 21}, we explored how TcdA/TcdB-intoxication impacted redistribution of both molecules. For this, we quantified accessible and total levels of protein in differentiated Caco-2 cells (8 days post-confluence) intoxicated with TcdA and TcdB for 3, 6 and 8 hours. When compared to control cells treated with DMEM FBS-free alone, the immunostaining of unpermeabilized cells highlighted an increase in accessible Fibronectin (acc Fn) in the apical side of intoxicated cells, as depicted by the green fluorescence (Fig. 1A). Additionally, permeabilized cells were stained for total Fibronectin (total Fn) and nuclei (blue) (Fig. 1A). Representative confocal microscopy images, including 3D projections and magnified views, showcased an equal distribution of total Fn in both conditions (Fig. 1A) Interestingly, there is an evident accumulation of acc Fn around intoxicated cells when compared to control (Figure 1A). Quantitative analysis of relative fluorescent intensity (RFLi) showed no significant change in accessible and total Fn during intoxication (Fig. S1A, Fig. 1C,D). However, when observing the basal and apical side of cells individually, quantitative analysis showed that after 8 h of TcdA/TcdB-intoxication, there was a significant increase in acc Fn at the apical side of IECs (Fig. 1C). Yet, total abundance of Fn remains unchanged throughout the 8 hours of intoxication (Fig. S1A). In the basal side of intoxicated IECs, no significant change in accessible or total Fn was observed during the 8-hour intoxication (Fig. 1D). Altogether, these findings suggest that during intoxication, Fn is redistributed towards the apical side of IECs. Control western blots demonstrate that TcdA/TcdB-intoxication results in glucosylated Rac-1 (Fig. 1G).

Fig. 1 |. Effect of TcdA/TcdB intoxication of intestinal epithelial Caco-2 cells in redistribution of Fibronectin and Vitronectin.

Differentiated Caco-2 cells intoxicated with TcdA and TcdB for 3, 6, or 8h in DMEM FBS-free. As a control, cells were treated with DMEM FBS-free. Unpermeabilized cells were stained for accessible Fibronectin or Vitronectin (acc Fn or acc Vn; green), permeabilized, and stained total Fibronectin or Vitronectin (total Fn or total Vn; red) and nuclei (blue). a, b Representative confocal microscopy images 3D projection of control cells (left) and intoxicated cells for 8h (right) immunostained for acc Fn or Vn and total Fn or Vn, below a magnified slide (XY), and the orthogonal view (XZ). Relative fluorescence intensity measured as the sum of raw intensity density/area for each z-step of accFn and total Fn, its abundance in the c apical side or d in the basal side of the cell; in the same way, the relative fluorescence intensity of Vn, its abundance in the e apical side and f the basal side of the cell. g, immunoblotting of anti-nonglucosylated Rac1 and total Rac1 of cell lysates of differentiated Caco-2 cells intoxicated with TcdA and TcdB for 3, 6, or 8 h. Nonglucosylated Rac1 was evaluated with corresponding antibodies, then the membrane was stripped, and subsequently tested for total Rac1. Western blotting is representative of 3 independent experiments. Controls were set at 100%. Error bars indicate the mean ± SEM from at least 9 fields (n = 3). Statistical analysis was performed by Two-Way ANOVA post-Bonferroni; ns, p > 0.05; * p < 0.05. Scale bar, top panels 20μm; bottom panels 5μm.

For Vn, we observed that TcdA/TcdB-intoxication of differentiated Caco-2 cells displayed distinctive changes in Vn distribution. In comparison to control cells, immunostaining of unpermeabilized cells revealed an increase in accessible Vn (acc Vn) on the basal side of intoxicated cells, as indicated by accumulation of the green fluorescence (Fig. 1B). Additionally, permeabilized cells were stained for total Vitronectin (total Vn) and nuclei (blue) (Fig. 1B). Representative confocal microscopy images, along with 3D projections and magnified views, displayed equal distribution of total Vn surrounding the nucleus of the cells (Fig. 1B). Interestingly, there is an increase in green fluorescence which is accumulated in portions where cells have detached from the wells resulting in evident clusters of acc Vn (Fig. 1B). Quantitative analysis of RFLi showed a significant increase in accessibility but not in total Vn after 8 h of incubation (Fig. S1B). Further analysis shows that accessible Vn increase was on the basal side rather than apical, with increased of accessible basal levels of Vn in as early as 6 h post-intoxication (Fig. 1E,F). Despite these increments in Vn, no changes in total Vn were detected either in the apical or basal layer of monolayers of Caco-2 cells (Fig. 1E,F). These results suggest that changes in accessible Vn, are not due to increased apical levels of Vn, but due to detachment of Caco-2 intoxicated cells that lead to increased basal accessible vitronectin.

TcdA/TcdB-intoxication of intestinal epithelial cells leads to increased accessibility of α₅β₁ and α_vβ₁ integrins.

In prior work, we demonstrated that C. difficile spores bind to Fn and Vn to gain intracellular access through their cognate integrin receptors, α₅β₁ and α_vβ₁, respectively³⁶. Consequently, we also explored whether TcdA/TcdB-intoxication of IECs would lead to increased accessible α₅β₁ and α_vβ₁ integrin receptor. Representative confocal micrographs show that there is an evident increase in accumulation of accessible α_5, but this was observed to happen in clusters around cells that seem to be surrounding potential sites of cellular apoptosis because of intoxication (Fig. 2A). Quantitative analysis unveiled a significant increase in RFLi of accessible ɑ₅ (acc ɑ₅) and total ɑ₅ in cells intoxicated with TcdA/B compared to controls (Fig. S2A,B). Closer quantification of ɑ₅ abundance in the apical side of IECs revealed a significant increase in accessible but not total ɑ₅ integrin subunit (Fig. 2D). Conversely, a significant increase in accessible and total ɑ₅ integrin was observed at the basal side of IECs after 8 h of toxin-intoxication (Fig. 2E).

Fig. 2 |. TcdA and TcdB increase accessible α₅ and α_V but no β₁ integrins in intestinal epithelial cells.

Differentiated Caco-2 cells intoxicated with 600pM of TcdA and TcdB for 8h in DMEM FBS-free. As a control, cells were treated with DMEM FBS-free. Unpermeabilized cells were stained for accessible a, ɑ5 integrin; b, α_V integrin, and c, β1 integrin, (shown in green), permeabilized, and stained total ɑ5, α_V or β1 integrin respectively (shown in red) and F-actin (grey). a-c, Representative confocal microscopy images 3D projection of control cells (left) and intoxicated cells for 8h (right); below a magnified slide (XY), and the orthogonal view (XZ). d-i, Quantification of relative fluorescence intensity based on raw intensity density per area for each individual cell generated from the microscopy images using the 3D Surface Plotter plug-in of ImageJ. Relative fluorescence intensity measured as the sum of raw intensity density/area for each z-step of accessible and total α₅ located in the d, apical, and e, basal side of the cell. Relative fluorescence intensity measured as the sum of raw intensity density/area for each z-step of accessible and total α_V located in the f, apical and g, basal side of the cell. Relative fluorescence intensity measured as the sum of raw intensity density/area for each z-step of accessible and total β₁ located in the h, apical, and i, basal side of the cell. Controls were set 100%. Error bars indicate the mean ± S.E.M from at least 9 fields (n = 3). Statistical analysis was performed by unpaired Student’s t test, ns, p > 0.05; * p < 0.05; ** p < 0.01. Bars, top panels 20 μm; bottom panels 5μm.

For the integrin subunit α_v, representative confocal micrographs show that there is an evident increase in accumulation of accessible α_v distributed equally in the observed tissue (Fig. 2B). Quantitative analysis revealed a notable increase in RFLi of accessible ɑ_V (acc ɑ_V) but not total ɑ_V in cells intoxicated with TcdA/TcdB when compared to controls (Fig. S2B). Assessment of ɑ_V abundance in the apical side of IECs showed a significant increase in accessible but not in total ɑ_V integrin subunit (Fig. 2F). Similarly, a significant increase in accessible but not in total ɑ_V was observed on the basal side of IECs during intoxication (Fig. 2G).

In striking contrast to the redistribution of alpha-integrin subunits, no changes in the levels of accessible and total integrin subunit β₁ was observed upon intoxication of Caco-2 cells with TcdA/TcdB (Fig. 2C). Representative confocal micrographs show that there is an equal distribution of β₁ in both intoxicated and non-intoxicated cells (Fig. 2C). Quantitative analysis revealed no changes in the RFLi of accessible β₁ (acc β₁) as well as no significant change in total β₁ levels in cells intoxicated with TcdA/B compared to controls (Fig. S2C). Examination of β₁ abundance in the apical side of IECs revealed no significant changes in accessible and total β₁ integrin (Fig. 2H). Likewise, there was no significant change in the abundance of accessible and total β₁ integrin on the basal side of IECs during intoxication (Fig. 2I). These findings emphasize that TcdA/B doesn’t affect accessible or total β₁ levels within IECs on both the apical and basal sides. Overall, these results indicate that TcdA/TcdB-intoxication of IECs increases accessible levels of ɑ₅ and ɑ_V in both, apical and basal sides, while no change in accessible levels of β₁ where observed. Intriguingly, TcdA/TcdB-intoxication increased the levels of total ɑ₅ in the basal side, which is consistent with repair-functions of Fn-ɑ₅ signaling³⁷. Collectively, these data indicates that TcdA/TcdB-intoxication of IECs increases accessible ɑ₅ and ɑ_V integrin receptors but does not affect availability of β₁.

Effect of TcdB-intoxication on the redistribution of intestinal Fn and Vn in vivo.

To explore whether C. difficile toxin-intoxication would also lead to increase accessible Fn and Vn in vivo, we utilized an ileal ligated loop mouse model injecting various doses of TcdB (with 0.1, 0.5, 1 and 5 μg of TcdB) for 5 h prior to confocal analysis. Upon staining for Fn, we observed in non-intoxicated ileal loops sites with accessible and total Fn at various sites of the villus (Fig. 3A and S3A). Interestingly, holes at the dome of the villus, which appear empty in control samples, had a notorious increase in accessible Fn upon intoxication with 5 μg of TcdB (Fig. 3A and Fig. S3A). These holes are notoriously known as goblet cells as they do not stain for actin-cytoskeleton³⁸. However, upon analyzing how accessible and total fluorescence intensity per defined area changed upon TcdB-intoxication, we observed a slight increase in accessible and total fibronectin with 0.1 μg of TcdB (Fig. 3C,D), yet levels subsequently decreased by 40% when intoxicating with 5 μg of TcdB (Fig. 3C,D).

Fig. 3 |. Effect of TcdB in the redistribution of Fibronectin and Vitronectin a ligated ileal loop mouse model.

Ileal ligated loops were intoxicated for 5 h with 0.1, 0.5, 1, or 5μg of TcdB or saline as control. Then loops were removed, washed, fixed, and subjected to immunofluorescence. Unpermeabilized tissues were stained for accessible Fibronectin or Vitronectin (acc Fn or acc Vn; green), and then permeabilized and stained total Fibronectin or Vitronectin (total Fn or total Vn; red) and F-actin (grey). a-b, Representative confocal microscopy images 3D projection of control cells (left) or intoxicated loops with 5μg TcdB (right) immunostained for accessible and total Fn or Vn; right bottom, a magnified 3D projection, next to a z-stack (XY), and then magnified orthogonal view (XZ). Quantification of c, e, accessible or d, f, total Fn or Vn fluorescence intensity per cell measured in the z-projection (sum). For acc Fn or Vn, the analyzed area was Ctrl of 170,360 μm²; 0.1 μg TcdB of 340,720 μm²; 0.5 μg TcdB of 340,720 μm²; 1 μg TcdB of 340,720 μm² and 5 μg TcdB of 511,080 μm². n = 3 animal per group. In scatter plots, each dot corresponds to one independent cell. Dots in colors correspond to the average of each analyzed mice/field. Error bars indicate mean or mean ± SEM. Statistical analysis was performed by unpaired Student’s t test; ns, p > 0.05; * p < 0.05; **p < 0.01; **** p < <0.0001. Scale bar 20 μm.

In contrast to Fn redistribution, healthy un-intoxicated ileal tissue had significant accessible Vn in various sites in the villi surface, and particularly in goblet-like cells (Fig. 3B and Fig. S3B). Strikingly, upon intoxication, no qualitative changes where evident with 0.5 or 1.0 μg of TcdB (Fig. S3B). However, intoxication with 5 μg of TcdB led to a massive increase in Vn throughout the entire intestinal epithelial layer, and this increase was primarily in total intracellular Vn (Fig. 3B). Upon quantitative analysis of changes in Vn, while accessible Vn decreased in a TcdB-concentration dependent manner (Fig. 3E), we observed a striking TcdB-concentration dependent increase in total Vn was observed, with a ~ 3-fold increase in the presence of 5 μg of TcdB (Fig. 3F). In summary, these results indicate that TcdB-intoxication of intestinal tissue in vivo leads to overall decrease in accessible and total Fn, accessible increase is evident in goblet-like cells. By contrast, TcdB-treatment while decreasing accessible Vn, led to substantial increase in total intracellular Vn.

Impact of C. difficile toxin-intoxication in spore-Fn and -Vn associations.

Despite in vitro work showing that Fn and Vn interact with C. difficile spores with strong K_D,³⁹ and that these molecules increase during TcdA/TcdB-intoxication of IECs as shown in Fig. 1A,B, it is unclear whether intestinal Fn and Vn interacts with C. difficile spores. First, we tested whether TcdA/TcdB-intoxicated IECs release Fn and Vn to bind to C. difficile spores. For this, C. difficile spores were incubated with DMEM (control), supernatant of untreated or 8 h TcdA/TcdB-treated differentiated monolayers of Caco-2 cells. Upon immunostaining of C. difficile spores for Fn and Vn, while no immunoreactivity was observed in DMEM-incubated spores, significant Fn ad Vn were detected in spores incubated with untreated supernatant (Fig. 4A,B). Notably, TcdA/TcdB-intoxication lead to a significant increase in supernatant Fn and Vn binding to C. difficile spores (Fig. 4A,B). Single spore quantification of Fn and Vn binding confirms these results and demonstrate a nearly two-fold increase in Fn and Vn binding when comparing supernatants from untreated cells with TcdA/TcdB-intoxicated cells (Fig. 4C,D). These results suggest that C. difficile spores might interact with intestinal Fn and Vn in an intoxicated mucosal layer.

Fig. 4 |. Intoxication increases association of C. difficile spores with Fn and Vn.

a-d, Differentiated Caco-2 cells were intoxicated with 600pM of TcdA and TcdB for 8h in DMEM FBS-free media. Controls include non-intoxoicated cells in DMEM FBS-free. Supernatant was collected from untreated and intoxicated cells and subsequently utilized to resuspend C. difficile spores and incubate for 1 h at 37 °C. Spores resuspended in DMEM alone were also included as a control. Spores were washed and strained for immunofluorescence anti- fibronectin and - vitronectin. a-b, Micrographs show representative phase-contrast (phase), fibronectin and vitronectin specific immunofluorescence and fluorescence intensity profiles (Fl. int.). Representative Fl. int. were provided using 3D Surface plotter function of Fiji. c-d, Quantitative analysis of the fluorescence Fl. int. of Fn and Vn in spores Fl. Int of 600 spores. Mean ± SEM are denoted. e-g, Ileal ligated loops were intoxicated with 5μg of TcdB and 5 × 10⁸ C. difficile R20291 spores for 5 h. Then loops were removed, washed, fixed, and subjected to immunofluorescence. Unpermeabilized tissues were stained for accessible Fibronectin or Vitronectin (acc Fn or acc Vn; green), and then permeabilized and stained total Fibronectin or Vitronectin (total Fn or total Vn; red) and F-actin (grey). e, g, representative 3D confocal micrograph projection reconstruction of fixed whole-mount small intestine tissue, and magnification of C. difficile spores associated with Fn or Vn. Plot profiles of fluorescence intensity of C. difficile spores (red line) and accessible Fn or Vn (green lines). f,h, quantification of spores that were positive (Fn+) or negative (Fn-) for Fn fluorescence signal in f, or positive (Vn+) or negative (Vn-) for Vn fluorescence signal in h. The average of associated and non-associated spores with f Fn or h Vn for each field. A total of ~ 500 spores were counted per mice (n = 5 per group). GRUBB’s test was performed to identify outliers, and one point was removed in Vn. Error bars indicate mean ± S.E.M. Statistical analysis was performed by two-tailed unpaired Student’s t test; ns indicates non-significant differences. Scale bar, 20 μm.

To explore the hypothesis that C. difficile spores interact in vivo with intestinal Fn and Vn, ileal loops injected with 5 μg of TcdB and 5 × 10⁷ spores, incubated for 5 h prior to immunostaining for accessible Fn and Vn and for total C. difficile spores. In the absence of TcdB, very few of adhered C. difficile spores (~ 2 % of total spores) interacted with Fn (Fig. 4E,F). In striking contrast to in vitro results (Fig. 4A,C), TcdB-intoxication had no impact in the fraction of C. difficile spores associated with Fn, which remained similar to that of untreated ileal loops (Fig. 4E,F). The analysis of C. difficile spore association with Vn revealed significant differences between unintoxicated and TcdB-intoxicated ileal loops. In unintoxicated ileal loops, approximately 3% of total spores were found to be associated with Vn(Fig. 4G,H). However, this percentage increased markedly in TcdB-intoxicated ileal loops, where about 10% of total spores showed association with Vn. This increase was statistically significant, indicating a notable effect of TcdB intoxication on the interaction between C. difficile spores and Vn (Fig. 4G,H). Collectively, these results support the notion that C. difficile toxin-intoxication leads to increase association of C. difficile spores mainly with intestinal Vn, and to a lesser extent with intestinal Fn.

TcdB increases C. difficile spore adherence and internalization to intestinal epithelial cells.

In prior work, we demonstrated that toxin-intoxication of Caco-2 cells leads to increase spore adherence and internalization³³, yet whether this also occurred in vivo remains unclear. Therefore, to address this question, ileal ligated loops were injected with 5 × 10⁸ C. difficile R20291 spores and 1 μg, or 5 μg of TcdB for 5h, or saline alone as control. Representative confocal images indicate a notable spore accumulation following intoxication, with clusters displaying a distinct affinity for specific regions within the cells, rather than an even distribution (Fig. 5A). Quantitative analysis of the number of spots (spores) adhered per 10⁵ μm² relative to the control or internalized spores in the ileum in relation to the total spores (Fig. 5B,C), showed that while no changes in adhered and internalized spores where observed during intoxication with 1 μg of TcdB, a there was a nearly 4-fold increase in adhered spores (Fig. 5B), and 10-fold increase in internalized spores (Fig. 5C) were observed in the presence of 5 μg of TcdB. Overall, these results confirm that C. difficile toxin-intoxication of the intestinal epithelial layer increases adherence and internalization of C. difficile spores in vivo.

Fig. 5 |. TcdB-intoxication increases C. difficile spore adherence and internalization to the intestinal barrier in vivo.

Intestinal loops, of approximately ~1.5 cm, were injected with 5 × 10⁸ C. difficile R20291 spores and 1 or 5 μg of TcdB for 5h (or saline alone as control). a, 3D projections of representative confocal micrographs. C. difficile spores are shown in green, and F-actin is shown in grey (fluorophores colors were digitally reassigned for a better representation). b,c, Quantification of b adhered or c internalized spots (spores) per 10⁵ μm² is expressed in relative values to the unintoxicated control. Controls were set 100%. GRUBB’s test was performed to identify outliers, and one point was removed in b for group intoxicated with 1 μg TcdB and in c for control group. Error bars indicate the mean ± S.E.M. Statistical analysis was performed by unpaired Mann-Whitney test, ns, indicates non- significant differences, * p < 0.05. Scale bar 100μm. n = 5 mice per group.

Protective effect of Bezlotoxumab in the redistribution of fibronectin and vitronectin during intoxication of intestinal epithelial cells.

Clinical data demonstrates that administration of Bezlotoxumab correlates with a reduction in the rates of recurrent CDI^{40, 41}. In prior work, we demonstrated that Fn and Vn contributes to spore-internalization into IECs, and that intracellular spores are implicated in CDI recurrence³⁶. Hence, we sought to assess the impact of Bezlotoxumab in the TcdB-intoxication mediated redistribution of Fn and Vn in the intestinal mucosa. Although the reported dose utilized in mice has been reported to be ~50 mg/kg^{29, 42}, in our preliminary results we found that for our batch of Bezlotoxumab that dose proved to be lethal upon intraperitoneal injection, while 5 mg/kg yielded 100% survival (data not shown). Consequently, 5 mg/kg of Bezlotoxumab was utilized in downstream experiments. For this, 24 h prior to ileal loop surgery, mice were given an intraperitoneal injection of 5 mg/kg Bezlotoxumab or of saline (PBS). Next, ligated loops where injected either PBS (control) or 5μg TcdB. Additionally, representative confocal micrographs show that accessible Fn seems to decrease upon intoxication (Fig. 6A), similarly as in Fig. 3A, this trend is also observed in the presence of Bezlotoxumab (Fig. 6A). Upon quantitative analysis of abundance of Fn, a significant decrease in accessible and total Fn was observed upon intoxication with TcdB. While presence of Bezlotoxumab in intoxicated ileal loops did not impact the levels of accessible when compared to intoxicated ileal loops alone, it caused a significant reduction in total Fn to ~25 % of total when compared to unintoxicated ligated loops (Fig. 6C).

Fig. 6 |. Effect of Bezlotoxumab in TcdB-intoxication induced redistribution of Fibronectin and Vitronectin in vivo.

Prior to ileal surgeries intraperitoneal injection of 5mg/kg of Bezlotoxumab or saline solution was administered as indicated. Ligated loops were injected with C. difficile Spores and 5μg of TcdB or saline as control. Mice were let for recovery for 5 h before euthanasia. a,b Representative 3D confocal micrograph projection reconstruction of fixed whole-mount small intestine tissue. Unpermeabilized tissues were stained for Acc Fn or Vn shown in green, and then permeabilized and stained for total Fn or Vn shown in red, and F-actin in gray and cell DNA with Hoechst (blue), (some fluorophores colors were digitally reassigned for a better representation). Panels c,d show the fluorescence intensity immunodetected in cells for acc and total c Fn or d Vn. Nearly 2000 – 3000 cells were counted per field. Quantification of total or accessible Vn, fluorescence intensity per cell measured in the z-projection (sum). Error bars indicate mean ± S.E.M. Statistical analysis was performed by two-tailed unpaired Student’s t test, ns indicates non-significant differences; * p < 0.05; **p < 0.01; **** p < <0.0001. Colored dots represent the average normalized fluorescence intensity of each independent mice in the group. n = 5 mice per group. Scale bar, 30 μm.

In the case of Vn, representative confocal micrographs show that cell-types with increased accessible and total Vn became evident in TcdB-intoxicated ligated loops when compared to Bezlotoxumab treatment alone (Fig. 6B). Notably, intoxication of ligated loops in the presence of Bezlotoxumab lead to a substantial increase in these cell types expressing increased accessible and total Vn (Fig. 6B). Quantification analysis of the fluorescence intensity of accessible and total Vn reveals a significant increase in accessible Vn, but not total, in TcdB-treated ligated loops (Fig. 6D). Strikingly, in TcdB-intoxicated ligated loops in the presence of Bezlotoxumab, both, accessible and total Vn increased significantly (Fig. 6D), consistent with some cells having higher levels of accessible and total Vn. Taken together, these results indicate that while administration of Bezlotoxumab has no impact in levels of accessible and total Fn during TcdB-intoxication, it dramatically increases both accessible and total Vn in a subpopulation of intestinal epithelial cell types.

Bezlotoxumab reduces spore adherence in TcdB-intoxicated mucosal epithelium layer.

The increase in adhered and internalized C. difficile spores in a TcdB-intoxicated ligated loop, led us to hypothesize that prior administration of Bezlotoxumab could protect from increased adherence and internalization of spores into IECs in vivo. For this, we treated mice with an i.p. of 5 mg/kg of Bezlotoxumab or saline (NaCl) 24 h before surgery. Ligated loops where then injected with C. difficile spores in the presence or absence of 5 μg of TcdB (n = 4–5 per group). Interestingly, representative confocal micrographs show that during TcdB-intoxication, C. difficile spores adhered to IECs in clusters, as opposed to what was observed in the control treated with Bezlotoxumab alone (Fig. 7A). Upon analysis of micrographs of TcdB-intoxicated and Bezlotoxumab-treated ligated loops, there was an evident decrease in the levels of spores associated with the intestinal mucosa in the Bezlotoxumab treated tissue(Fig. 7A). Upon performing quantitative analysis, we observed that in tissues intoxicated with TcdB, spore adherence increases significantly by ~400 % when compared to Bezlotoxumab control (Fig. 7B). In animals pre-treated with Bezlotoxumab and intoxicated with TcdB, spore adherence increased significantly by 280 % when compared to Bezlotoxumab control (Fig. 7B). Quantification of intracellular spores demonstrated that while there was an increase in intracellular spores in TcdB-intoxicated ileal loops, which decreased in the presence of Bezlotoxumab, these changes where not significant (Fig. 7C). Collectively, these results indicate that Bezlotoxumab might prevent TcdB-mediated enhanced adherence and internalization into the intestinal mucosa.

Fig. 7 |. Bezlotoxumab neutralizes TcdB- increased adherence of C. difficile spores caused by TcdB.

A) Mice were treated with an intraperitoneal injection of 5mg/kg of Bezlotoxumab or saline solution 24h prior to ileal loops surgery. Loops where then injected with 5 × 10⁸ C. difficile R20291 spores with or without 5μg of TcdB. Representative confocal micrographs of ileal loops are shown, with C. difficile spores stained in red, F-actin is shown in green, and nuclei in blue. b,c, Quantification of b adhered and c internalized spots (spores) per 10⁵ μm² is shown relativized to the non-intoxicated Bezlotoxumab control. GRUBB’s test was performed to identify outliers, and one point was removed in group Bezlotoxumab + spore and one in TcdB + spore. Error bars indicate the mean ± S.E.M. Statistical analysis was performed by Mann-Whitney test, ns, p > 0.05, * p < 0.05. Scale bar 20μm. n = 5 mice per group.

Discussions

During CDI, toxin production and spore formation play crucial roles in disease manifestation and recurrence, respectively.^{27, 43}. Recent research has significantly advanced our understanding of how C. difficile interacts with host cells^{33, 36}. C. difficile spores utilize E-cadherin for adherence and Fn and Vn integrin-dependent pathways for internalization into IECs^{33, 36}, with bothspore-adherence and -internalization contributing to disease recurrence³⁶. However, the impact of C. difficile toxins in spore adherence and internalization remains unclear. This study reveals that C. difficile toxin-intoxication of IECs contributes to remodeling of Fn, Vn, and their cognate integrins. Specifically, TcdA/TcdB intoxication led to increased binding of C. difficile spores with released Fn and Vn from Caco-2 cell monolayers, but only with Vn in intoxicated ligated loops. Importantly, TcdB-intoxication increased spore adherence and internalization into intestinal tissue. Using Bezlotoxumab, a monoclonal antibody that neutralizes TcdB, we observed that neutralization of TcdB had no impact on redistribution of Fn and Vn, however there was a reduction in spore adherence and internalization into IECs in vivo. These findings provide valuable insights into the role of C. difficile toxins TcdA and TcdB in spore interactions with the intestinal tissues.

Our work expands our understanding of the roles of C. difficile major toxins, TcdA and TcdB, beyond their essential function in causing clinical symptoms and inflammation during infection.⁸. Our data suggests that TcdA/TcdB-intoxication enhances the levels of accessible bioavailable Fn and Vn from human IECs. While Fn and Vn are primarily located in the basal and basolateral membrane of IECs and contribute to their polarity,^{11, 12} accessible Fn and Vn are commonly identified in certain epithelial folds and extrusion sites that undergo adherent junction reorganization^{36, 44–46}.

Both TcdA and TcdB glucosylate host Rho family of GTPases⁴⁷; which blocks the exchange of GDP for GTP and prevents RhoA, Rac1 and Cdc42 from functioning. This leads to cell rounding, opening of tight and adherent junctions and cell death⁴⁸. The toxin-mediated cell-intoxication likely results in a remodeling of the basolateral membrane, increasing accessibility of Fn and Vn and causing extrusion of Caco-2 cells. This is supported by the observed increase in apical and basal levels of the α₅ and α_v integrins. Both toxins also elicit a robust proinflammatory response through multiple pathways, including mitogen-activated protein kinases that activate the transcription nuclear factor κB (NF-κB)^{49, 50}. They also trigger the ASC-containing inflammasome, leading to secretion of various cytokines such as IL-8, tumor necrosis factor alpha (TNF- α), and IL-6, and IL-1B^51–53. Additionally, these toxins also trigger a series of cell death pathways including apoptosis, necrosis and pyknosis^54–57. Under the experimental conditions tested, the observed changes in levels of accessible and total Fn and Vn are likely due to a combination of intoxication, rupture of tight and adherent junctions and to some extent to cell death and detachment. Prior work by Schwan et al. demonstrated that TcdB, but not TcdA, caused a slight change in apical Fn²⁵. However, the C. difficile binary toxin rapidly reroutes Fn-containing vesicles from the basal to the apical side of Caco-2 cell monolayers within 1 hour of intoxication²⁵. This Fn redistribution occurs through binary toxin-induced protrusions, primarily involving Rab11 vesicle traffic to the apical surface, colocalizing Fn with α₅ and β₁ integrins²⁵. Notably, accumulations of laminin and Vn were also observed, albeit to a lesser extent than Fn²⁵.

Given the extended time (~8 hours) required in our study to observe these changes, it is likely that alterations in accessible Fn and Vn are due to secondary cell death and/or opening of the basolateral junctions, rather than a specific pathway. However, early work shows that DSS-induced inflammation of IECs also leads to increased Fn production²². Therefore, it is plausible that TcdA/TcdB toxin-mediated inflammation also leads to increased Fn production. Further work to elucidate the specific signaling pathways involved in the increased Fn and Vn levels, as well as the individual roles of each toxin is warranted.

This work also demonstrates that, at least TcdB-intoxication of intestinal tissue increases total gut Vn in intestinal tissue. Surprisingly, in contrast to the increase of both accessible Fn and Vn in Caco-2 cells, only total Vn levels changed during TcdB-ileal loop intoxication. This discrepancy might reflect the differences between experimental platforms and/or additional pathways activated in the different environments. The inherent higher variability of immunofluorescence of these molecules in ligated loops compared to tissue monolayers could be a limitation of our technique. Despite these inconsistencies, the finding that TcdB-intoxication leads to an increase in total Vn is novel. Confocal micrographs clearly show that the main source of total Vn was intracellular, primarily in the monolayers of IECs. Our understanding of gut Vn is limited, but it is known to be produced by IECs during early stages of development⁵⁸, and its increased expression during chronic intestinal inflammation has been linked to protection against colitis²⁴.

Both TcdA and TcdB toxins are known to induce inflammasome activation^{51, 53}, but whether this impacts intracellular Vn production remains unclear and is a subject of ongoing research in our lab. Although TcdA and TcdB have similar functional domains and cause comparable changes in collapsing tight and adherent junctions and in cell physiology, recent studies show that they distinctively alter host pathways⁵⁹. For instance, TcdB, but not TcdA, affects stem cells, impacting tissue repair and disease recurrence prevention⁶⁰. TcdA and TcdB toxins also have different receptor requirements. TcdA binds to glycoproteins like sucrose isomaltose (SI) and glycoprotein 96 (gp96)^{61, 62}, sulfated glycosaminoglycans (sGAGs) and/or membrane low-density lipoprotein receptor (LDLR) family⁶³, with direct binding to LRP1⁶⁴. In contrast, TcdB has three reported protein receptors: chondroitin sulfate proteoglycan 4 (CSPG4), Frizzled 1 (FDZ1), FZD2, FZD7 and Nectin 3 (also called “polivirus receptor-like 3” or PVRL3)^65–67. The distribution of these receptors in the murine GI tract remains unclear, but it’s possible that TcdA also impacts Vn accessibility, and perhaps Fn. Furthermore, toxin-mediated epithelial cell death initiates a MAP kinase signaling cascade, triggering recruitment and infiltration of granulocytes and activation of innate immune cells. This complex interplay between toxins, host cells, and immune responses contributes to the pathogenesis of C. difficile infection and highlights the need for further research into the specific mechanisms involved.

Previous work demonstrated that Fn and Vn bind to C. difficile spores in a concentration-dependent manner through solid-phase binding and pulldown assay³⁹. This current study expands on those findings by showing that C. difficile spores bind to Fn and Vn released from TcdA/TcdB-intoxicated cells. Additionally, we observed increased spore-Vn association in TcdB-intoxicated intestinal loops, proving the first evidence of C. difficile spores associating with Vn in vivo. This aligns with the increased levels of total gut Vn observed in TcdB-intoxicated intestinal tissue.

The specific spore surface ligand responsible for Fn and Vn binding remains unclear, with no obvious candidates identified in the C. difficile spore exosporium proteome⁶⁸, current matter of investigation in our lab. In contrast, several Fn-binding proteins have been identified on the surface of C. difficile vegetative cells. C. difficileś genome encodes at least two cell-surface proteins for which experimental evidence supports their role as Fn-binding proteins (i.e., FbpA and Fbp68)^{69, 70}, and at least FbpA is required for normal colonization of germ-free mice⁶⁹. While we have previously demonstrated that C. difficile spores use Fn and Vn as molecular bridges to internalize into human intestinal epithelial Caco-2 cells³⁶, the implications of these interactions during C. difficile infection, particularly in terms of persistence and disease recurrence, remain to be fully elucidated.

A major conclusion provided by this work is that TcdB-intoxication of intestinal tissue enhances the adherence and internalization C. difficile spores in vivo. This finding extends previous research that demonstrated that TcdA/TcdB-intoxication of Caco-2 cells monolayers led to higher levels of spore adherence and internalization³³, by showing that this phenomenon also occurs in vivo. Although the study focused on TcdB, substantial evidence suggests that both toxins A and B, significantly disrupt the epithelial barrier^71–73, indicating that these results may also extrapolate to TcdA. Previous work showed that TcdA/TcdB-intoxicated monolayers of Caco-2 cell monolayers underwent adherent junction remodeling, increasing accessible E-cadherin and its association with C. difficile spores³³. Moreover, E-cadherin was identified as a spore-adherence receptor that contributed to spore entry into IECs.

While changes in accessible E-cadherin were not quantified in this work, this and additional mechanisms, including enhanced interactions between C. difficile spores and intestinal Vn, may contribute to increased adherence and internalization. These results also support the hypothesis that during infection, C. difficile toxins A and B contribute to intestinal epithelial remodeling, potentially increasing spore persistence and, consequently, disease recurrence after vancomycin treatment. Our lab is currently investigating the specific impact of each toxin in C. difficile persistence and disease recurrence.

A final contribution of this work pertains the impact of Bezlotoxumab on TcdB-mediated remodeling of Fn and Vn, and consequently, on adherence and internalization of C. difficile spores to intestinal tissue. Initial clinical trials by Merck, which included anti-TcdA and anti-TcdB monoclonal antibodies (i.e., Actoxumab and Bezlotoxumab), showed that Bezlotoxumab significantly reduced CDI recurrence when combined with antibiotics^{41, 74, 75}. In preclinical animal models, both, Actotoxumab and Bezlotoxumab together protect mice from severe C. difficile disease^{76, 77}. However, these previous studies used an i.p. dose of 50 mg/kg^{76, 77}, which is 10-fold higher than the dose used in this study (5 mg/kg). In preliminary work, we attempted to use 50 mg/kg, however, this dose proved lethal in our murine model system. Consequently, we identified 5 mg/kg as a non-lethal dose that could be administered via i.p. 24 h prior to the ileal loop surgery. Our data suggests a trend in preventing TcdB-mediated adherence and internalization into intestinal tissue, but higher levels of Bezlotoxumab might be necessary to fully prevent this effect. Similarly, TcdB-mediated increased total gut Vn did not decrease upon administration of Bezlotoxumab. Overall, while our data strongly supports that TcdB increases spore adherence and internalization into intestinal tissue, the low Bezlotoxumab concentration used failed to fully protect against TcdB-mediated damage. These findings emphasize the need for further research to optimize therapeutic strategies for preventing CDI recurrence. The study also underscores the importance of dosage considerations in translating preclinical findings to clinical applications.

Based on our findings our proposed model (Fig. 8) illustrates the dual impact of C. difficile toxins and Bezlotoxumab treatment on intestinal epithelial cells. In untreated CDI, TcdA and TcdB intoxication leads to increased Acc Fn, Vn, and their associated integrins α₅ and α_v, ultimately resulting in enhanced spore adherence and internalization, alongside epithelial cell apoptosis. When Bezlotoxumab is administered, it partially neutralizes TcdB, reducing epithelial damage and spore adherence/internalization. However, the treatment still results in increased total Vn while decreasing total Fn, suggesting a complex modulation of these extracellular matrix proteins. This model supports our findings that low-dose Bezlotoxumab provides partial protection against TcdB-mediated damage while highlighting the intricate relationship between toxin activity and spore-host interactions.

Fig. 8 |. Model of C. difficile toxin-mediated spore interactions and Bezlotoxumab effects on intestinal epithelial cells.

Left panel (CDI): During C. difficile infection, toxins A and B (TcdA/TcdB) cause increased accessible (Acc) fibronectin (Fn), vitronectin (Vn), and their associated integrins (α5 and αv) on both apical and basal surfaces of intestinal epithelial cells. This leads to enhanced spore adherence and internalization, ultimately resulting in epithelial cell apoptosis. Right panel (CDI + Bezlotoxumab): Treatment with Bezlotoxumab partially neutralizes TcdB, reducing epithelial damage and spore adherence/internalization. While total Vn levels increase, total Fn levels decrease. The presence of Bezlotoxumab antibodies results in reduced spore adherence and internalization compared to untreated conditions, demonstrating partial protection against TcdB-mediated damage. Arrows indicate increased (↑) or decreased (↓) levels of proteins. A and B represent TcdA and TcdB toxins, respectively. The apical and basal sides of the epithelium are indicated.

Materials and Methods

Data reporting

Statistical methods were not employed for predefining the sample size. In the case of animal experiments, mice were assigned to various groups through random allocation. The researchers did not employ blinding during the conduct of the animal experiments to prevent cross-contamination among animals from distinct treatment groups.

Animals used in this study.

Male or female C57BL/6 mice, aged 6–8 weeks, were sourced from the breeding colony at the Departamento de Ciencias Biológicas of the Universidad Andrés Bello, which was established using animals acquired from Jackson Laboratories. The mice were individually housed in cages maintained under uniform conditions, with autoclaved water, bedding, and cages. A 12-hour light-dark cycle was maintained, and the mice were kept at a temperature of 20–24°C with a humidity range of 40–60%. All experimental procedures strictly adhered to and were formally approved by the Institutional Animal Ethics Committee of the Universidad Andrés Bello (Protocol #015/2019) and the Animal Ethics Committee of the Facultad de Ciencias Biológicas of the Universidad Andrés Bello (Protocol #015/2019).

Recombinant C. difficile toxin purification

TcdA and TcdB of C. difficile were produced in Bacillus megaterium carrying the plasmid pHIS1622, which harbors the cloned genes tcdA or tcdB from C. difficile VPI 10463, as previously documented⁷⁸.

Transformed B. megaterium cells were cultivated in Luria-Bertani medium (BD, USA), supplemented with 10 μg/mL tetracycline. An overnight culture (35 mL) was used to inoculate 1 L of medium, and the bacteria were cultured at 37°C with shaking at 220 rpm. To induce toxin expression, 0.5% D-xylose was added once the culture reached an optical density at 600 nm of 0.5. After 4 hours, cells were harvested and resuspended in 200 mL of binding buffer (20 mM Tris [pH 8.0], 100 mM NaCl for TcdA and 20 mM Tris [pH 8.0] or 500 mM NaCl for TcdB), supplemented with 0.16 μg/mL DNase, 10 mg/mL lysozyme, and protease inhibitors (P8849; Sigma). Cell lysis was achieved using an Emulsiflex homogenizer, and lysates were centrifuged at 48,000 × g for 30 minutes. The toxins were purified from the supernatant through Ni-affinity, anion exchange, and size exclusion chromatography. The purified toxins were eluted and stored in 20 mM HEPES (pH 7.0), 50 mM NaCl.

Purification of C. difficile spores

Spore preparation was conducted following previously established procedures⁷⁹. In summary, a 1:1,000 dilution of an overnight culture in BHIS was plated (100 μL) on 70:30 agar plates (6.3% weight/vol BD, USA; 0.35% weight/vol protease peptone BD, USA; 0.07% ammonium sulfate (NH4)2SO4 Merck USA; 0.106% weight/vol Tris base Omnipur, Germany; 1.11% weight/vol brain heart infusion extract BD, USA; 0.15% weight/vol yeast extract BD, USA; 1.5% weight/vol Bacto agar BD, USA). Next, 70:30 agar plates were incubated for 7 days in an anaerobic chamber (Bactron III-2, Shellab USA) at 37°C. Subsequently, colonies were scraped from the plates with ice-cold sterile Milli-Q water. The sporulated culture underwent five washes with ice-cold Milli-Q water in a micro-centrifuge at 18,400×g for 5 minutes each. Spores were purified by density using 45% weight/vol autoclaved Nycodenz (Axell USA) and centrifugation at 18,400×g for 40 minutes. The bacterial debris and non-spore residuals were removed, and the spore pellet was separated and washed five times at 18,400×g for 5 minutes with ice-cold sterile Milli-Q water to eliminate Nycodenz. Spore concentration was determined in the Neubauer chamber, adjusted to 5 × 10⁹ spores/mL, and stored at −80°C until use.

Immunofluorescence of E-cadherin, Fn, Vn, a₅, a_v, and β₁ in TcdA and TcdB intoxicated Caco-2 cells.

Caco-2 cells, cultured on glass coverslips in 24-well plates for 8 days post-confluence, underwent two washes with DPBS before intoxication with 600 pM of TcdA and TcdB (161,83pg/mL and 184.83 pg/mL) in DMEM FBS-Free or DMEM alone (as a control) for 3, 6, or 8 hours. Following treatment, cells were rinsed with PBS, fixed with PBS-4% paraformaldehyde, and further washed with PBS.

For staining surface-accessible proteins, cells were incubated with primary antibodies (1:200 dilution) including mouse monoclonal antibodies against human fibronectin (sc8422, Santa Cruz Biotechnologies, USA), vitronectin (sc74484, Santa Cruz Biotechnologies, USA), integrin α₅ (ab78614, Abcam, USA), integrin α_v (ab16821, Abcam, USA), and integrin β₁ (MAB1959Z, Millipore, USA). Incubation was carried out either overnight at 4 °C or for 1 hour at room temperature, followed by secondary antibody treatment using donkey anti-mouse IgG Alexa Fluor 488 (ab150109, Abcam, USA) at a 1:400 dilution.

To detect total protein, samples were permeabilized with PBS-0.2% Triton X-100 for 10 minutes, blocked with PBS-1% BSA, and subjected to the same primary antibody as used for surface staining. After washing, cells were incubated with donkey anti-mouse IgG Alexa Fluor 568 (ab175700, Abcam, USA). Finally, samples were mounted using Dako Fluorescent Mounting Medium (Dako) and visualized using confocal microscopy.

Binding of Fn and Vn released from intoxicated cells to C. difficile spores

Differentiated Caco-2 cells were exposed to 600 pM of TcdA and TcdB for 8 hours at 37 °C in DMEM FBS-free. As a control, cells were incubated in DMEM FBS-free without toxins. Subsequently, the supernatant was meticulously purified of cellular debris through two centrifugation cycles: first at 760×g for 5 minutes, followed by collection of the supernatant, and then at 18,400×g for 10 minutes. This final supernatant was utilized in the subsequent experiment.

Next, 4 × 10⁷ C. difficile R20291 spores were incubated for 1 hour at 37 °C with 100 μL of the obtained supernatant from TcdA and TcdB-intoxicated cells or from cells incubated in DMEM FBS-free or fresh DMEM as a control. Unbound molecules were removed by washing the spores three times through cycles of centrifugation at 18,400×g for 5 minutes and resuspension in PBS.

The spores were then fixed on coverslips coated with poly-L-lysine, treated with PBS-4% PFA for 15 minutes at room temperature, and blocked with PBS-1% BSA for 1 hour at room temperature. Subsequently, coverslips were incubated with 1:200 dilution of mouse monoclonal anti-Fn (sc-8422, Santa Cruz Biotechnologies, USA) or mouse monoclonal anti-Vn (sc-74484, Santa Cruz Biotechnologies, USA) for 1 hour at room temperature. Following this, coverslips were incubated with 1:400 chicken anti-mouse IgG conjugated to Alexa 488 (A21200 ThermoFisher, USA) for 1 hour at room temperature. Finally, cells were mounted with Dako Fluorescent Mounting Medium (Dako). The samples were observed on an Olympus BX53 fluorescence microscope with UPLFLN 100× oil objective (numerical aperture 1.30), and images were captured using the microscope camera for fluorescence imaging, Qimaging R6 Retiga.

Quantitative analysis of Fn and Vn binding to C. difficile spores was conducted using ImageJ (NIH, USA), as previously described (see references Mora Uribe 2016 and Castro Cordova 2020). Relative Fn and Vn spore fluorescence intensity was measured by outlining around 600 spores per experimental condition, along with several adjacent background readings. The data presented represent [spore (Fl. int / area) – background (Fl. int / area)]. Fluorescence intensity profiles were generated from the microscopy images using the 3D Surface Plotter plug-in of ImageJ.

Association of C. difficile spore with cellular Fibronectin or Vitronectin in differentiated Caco-2 cells intoxicated with TcdA and TcdB.

To assess the interaction between C. difficile spores and cellular fibronectin or vitronectin during exposure to C. difficile toxins, differentiated Caco-2 cells were treated with 600 pM of TcdA and TcdB for 3, 6, or 8 hours at 37°C in DMEM FBS-free. As a control, cells were incubated in DMEM FBS-free. Subsequently, intoxicated cells were washed twice with PBS and then exposed to C. difficile spores for 3 hours at 37°C. Unbound spores were removed by three washes with PBS, and the cells were fixed with 4% PFA, followed by three additional washes with PBS.

In non-permeabilized monolayers, extracellular spores were labeled with 1:1,000 chicken IgY anti-C. difficile spore⁸⁰ in PBS–1% BSA for 1 hour at room temperature. After two washes, cells were incubated with 1:400 goat anti-chicken IgY conjugated with Alexa Fluor 488 (ab150173, Abcam, USA) in PBS–1% BSA for 1 hour at room temperature. Following two washes, cells were stained with mouse monoclonal antibodies against human fibronectin (sc8422, Santa Cruz Biotechnologies, USA) or vitronectin (sc74484, Santa Cruz Biotechnologies, USA). Subsequently, samples were washed twice with PBS, and fibronectin or vitronectin were detected with 1:400 donkey anti-mouse IgG Alexa Fluor 568 (ab175700, Abcam, USA) for 1 hour at room temperature. After three washes with PBS and one with sterile distilled water, samples were dried at room temperature for 30 minutes, and coverslips were mounted using Dako Fluorescence Mounting Medium (Dako, Denmark) and sealed with nail polish. Samples were observed using confocal microscopy, as described below.

To determine the association of C. difficile spores with fibronectin and vitronectin released from differentiated Caco-2 cells intoxicated with TcdA and TcdB, cells were intoxicated for 8 hours in DMEM FBS-free. As a control, cells were treated with DMEM FBS-free. The supernatant was collected, carefully centrifuged at 760×g for 5 minutes, and then centrifuged at 18,400×g for 10 minutes. The new supernatant was collected, and 4 × 10⁷ C. difficile R20291 spores were incubated with 100 μL of the supernatant from intoxicated cells, healthy cells, or fresh DMEM for 1 hour at 37°C. Unbound molecules were washed off three times by cycles of centrifugation at 18,400×g for 5 minutes and resuspension in PBS. Spores were fixed on coverslips coated with poly-L-lysine, treated with PBS-4% PFA for 15 minutes at room temperature, and blocked with PBS-1% BSA for 1 hour at room temperature. Afterward, C. difficile spores and fibronectin or vitronectin were immunodetected as described above. Samples were observed on an Olympus BX53 fluorescence microscope with UPLFLN 100× oil objective (numerical aperture 1.30), and images were captured using the microscope camera for fluorescence imaging, Qimaging R6 Retiga. Quantitative analysis of Fn and Vn binding to C. difficile spores was conducted using ImageJ (NIH, USA), as previously published (Mora-Uribe et al., 2016). To measure the relative Fn and Vn spore fluorescence intensity, an outline was drawn around 600 spores per experimental condition and several adjacent background readings. The data shown correspond to [spore (Fl. int / area) – background (Fl. int / area)]. Fluorescence intensity profiles were generated from the microscopy images using the 3D Surface Plotter plug-in of ImageJ.

Similarly, to investigate whether the association of Fn or Vn with C. difficile spores is time-dependent, differentiated Caco-2 cells were infected with C. difficile spores for 0, 15, 30, 60, 120, and 180 minutes at 37°C in DMEM FBS-free. Subsequently, cells were washed three times with PBS and fixed with 4% PFA. Immunostaining of spores and Fn or Vn was performed as described above. Samples were visualized in confocal microscopy, and the fluorescence intensity was quantified for each spore in the channels of anti-spore and anti-Fn or anti-Vn.

Ileal loop assay

C57BL/6 mice were initially anesthetized in an isoflurane chamber (RWD USA) using 4% vol/vol isoflurane (Baxter USA) and maintained at 2% vol/vol during surgery administered by air. The established intestinal loop model was implemented as per previously referenced protocol^{36, 81, 82}. Briefly, a midline laparotomy was performed, involving a 1-cm incision in the abdomen. The ileum and proximal colon (1.5 cm each, positioned 1.0 – 1.5 cm from the cecum as a reference) were ligated using silk surgical suture.

To assess the impact of TcdB on the cellular distribution of Fn and Vn in the ileum mucosa, C57BL/6 mice underwent ileum ligation and were injected with varying doses (0.1, 0.5, 1, or 5μg) of TcdB in 0.9% NaCl (saline), with a saline-only control (n = 3 for each treatment). For investigating whether TcdB influences the adherence of C. difficile spores in the ileum mucosa, ligated loops were injected with 5 × 10⁸ C. difficile R20291 spores and 1μg or 5μg of TcdB for 5 hours in saline (or saline alone as control) (n = 5 each treatment).

To explore the potential protective role of Bezlotoxumab in spore persistence in TcdB-intoxicated ileum, C57BL/6 mice were intraperitoneally injected with 5mg/kg of Bezlotoxumab (n = 10) 24 hours before surgery, with saline as a control. Subsequently, Bezlotoxumab-treated mice were divided into two groups: one injected with 5 × 10⁸ C. difficile R20291 spores (n = 5), and the other with both spores and 5μg of TcdB (n = 5). For animals injected with saline, loops were injected with 5 × 10⁸ C. difficile R20291 spores and 5μg of TcdB (n = 5). Following the procedures, the intestine was returned to the abdomen, the incision was closed, and the animals were allowed to regain consciousness. Mice were kept for 5 hours before euthanasia. The ligated loops were carefully removed and washed in PBS before subsequent immunostaining, as detailed below.

Immunostaining of ileal and colonic loops

Initially, we longitudinally cut the extracted and washed intestinal tissues from the ileal and colonic loops. Subsequently, these tissues underwent a triple immersion in PBS at room temperature (RT) for thorough washing. For enhanced visualization, the tissues were flattened at RT. This involved fixing the tissues flat over a filter paper saturated with a solution containing 30% sucrose (Winkler, Chile) in PBS–4% paraformaldehyde (Merck, USA) for a minimum of 15 minutes. Following this, the tissues were transferred to a microcentrifuge tube containing the same fixing solution and were left to incubate at 4 °C overnight.

Given that fixing mucus with cross-linking agents like paraformaldehyde can lead to the collapse and shrinkage of the mucus layer in the colon⁸³, it’s noteworthy that we did not observe a mucus layer in our ileal and colonic loops. In preparation for immunostaining, the intestinal and colonic tissues were subsequently cut into approximately ~5 × 5 mm fragments.

In the TcdB-intoxicated ileum, for immunostaining of both luminally accessible and total fibronectin or vitronectin, ileum fragments underwent an incubation step with a primary antibody at a dilution of 1:150 rabbit pAb anti-fibronectin (sc-9068, Santa Cruz Biotechnology, USA) or rabbit pAb anti-vitronectin (sc-15332, Santa Cruz Biotechnology, USA) in PBS-3% BSA overnight at 4 °C. Subsequent to washing, the fragments were incubated with a secondary antibody at a dilution of 1:400 donkey anti-rabbit IgG Alexa 568 (ab175692, Abcam, USA) for 3 hours at RT. To stain total protein, the tissues were permeabilized through a 2-hour incubation with PBS-0.2% Triton X-100 at RT and then blocked with 3% BSA for 3 hours at RT. Following this, the tissues were incubated overnight at 4 °C with the same primary antibody used earlier (1:150). F-actin was stained using a 1:150 dilution of fluorescently labeled phalloidin Alexa Fluor 647 (A22287, ThermoFisher, USA). On the subsequent day, the tissues were washed and incubated with a mixture of 1:400 donkey anti-rabbit IgG Alexa 488 (ab150061, Abcam, USA) and Hoechst at a 1:1,000 dilution for 3 hours at room temperature.

For the quantification of C. difficile spore adherence and internalization in both colonic and ileum mucosa, tissues underwent permeabilization through incubation with PBS–0.2% Triton X-100 (Merck, USA) and were subsequently blocked with PBS–3% BSA (Sigma–Aldrich, USA) for 3 hours at room temperature (RT). The same buffer was utilized for the subsequent antibody incubation. The tissue was then exposed to a primary polyclonal antibody, specifically a 1:1,000 dilution of anti-C. difficile spore IgY batch 7246 antibodies (Aveslab USA) in PBS–3% BSA. Notably, these antibodies do not react with epitopes of vegetative cells or murine microbiota⁸⁰. Concurrently, the tissue was also incubated with a 1:50 dilution of phalloidin Alexa-Fluor 568 (#ab176753 Abcam, USA) in PBS–3% BSA overnight at 4 °C to stain the actin cytoskeleton. Following a PBS wash, the samples were incubated with a secondary antibody, specifically a 1:400 dilution of goat anti-chicken IgY Alexa-Fluor 488 (#ab150173 Abcam USA) in PBS–3% BSA at RT. After three PBS washes, cellular nuclei were stained with a 1:1,000 dilution of Hoechst (ThermoFisher, USA) for 15 minutes at RT. Subsequently, the immune-stained tissues were mounted with the luminal side facing up, following the protocol previously described⁸¹.

Confocal and Epifluorescence analysis of immunostained tissued or intestinal epithelial cells.

For confocal imaging, the Leica SP8 system was employed, utilizing an HPL APO CS2 40× oil objective with a numerical aperture of 1.30. Detection of signals was facilitated through three PMT spectral detectors: PMT1 (410–483) for DAPI, PMT2 (505–550) for Alexa-Fluor 488, and PMT3 (587–726) for Alexa-Fluor 555. Emitted fluorescence was separated using dichroic mirrors DD488/552.

To quantify the fluorescence intensity of fibronectin, vitronectin, α5, αv, and β1 images (1,024×1,024 pixels) were acquired with a 0.5-μm Z step. Fluorescence intensity measurements were conducted for channels representing both accessible and total protein. The z-stack analysis involved an equal number of slides. To assess fluorescence intensity in the basal and apical halves, z-stacks were divided into halves with an equal number of slides, and the fluorescence intensity was then summed. Six microscopy fields were analyzed for each experimental condition.

For the evaluation of luminally accessible fibronectin and vitronectin in the ileum mucosa, images of 1,024×1,024 pixels were captured with a 2-μm Z step size, followed by filtering with Gaussian Blur 3D (sigma x: 0.6; y: 0.6; z: 0.6). Three-dimensional reconstructions of the ileum mucosa were performed using ImageJ software (NIH, USA). Villi were visualized through Hoechst and phalloidin signals. Each animal contributed two pictures (n = 3 independent mice), and the total intestinal epithelium area analyzed for each mouse was 0.5 mm². Representative 3D projections were generated using the 3D projection plug-in of ImageJ.