Saccharomyces boulardii protease inhibits Clostridium difficile toxin A effects in the rat ileum.

Saccharomyces boulardii, a nonpathogenic yeast, is effective in treating some patients with Clostridium difficile diarrhea and colitis. We have previously reported that S. boulardii inhibits rat ileal secretion in response to C. difficile toxin A possibly by releasing a protease that digests the intestinal receptor for this toxin (C. Pothoulakis, C. P. Kelly, M. A. Joshi, N. Gao, C. J. O'Keane, I. Castagliuolo, and J. T. LaMont, Gastroenterology 104: 1108-1115, 1993). The aim of this study was to purify and characterize this protease. S. boulardii protease was partially purified by gel filtration on Sephadex G-50 and octyl-Sepharose. The effect of S. boulardii protease on rat ileal secretion, epithelial permeability, and morphology in response to toxin A was examined in rat ileal loops in vivo. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified S. boulardii protease revealed a major band at 54 kDa. Pretreatment of rat ileal brush border (BB) membranes with partially purified protease reduced specific toxin A receptor binding (by 26%). Partially purified protease digested the toxin A molecule and significantly reduced its binding to BB membranes in vitro (by 42%). Preincubation of toxin A with S. boulardii protease inhibited ileal secretion (46% inhibition, P < 0.01), mannitol permeability (74% inhibition, P < 0.01), and histologic damage caused by toxin A. Thus, S. boulardii protease inhibits the intestinal effects of C. difficile toxin A by proteolysis of the toxin and inhibition of toxin A binding to its BB receptor. Our results may be relevant to the mechanism by which S. boulardii exerts its protective effects in C. difficile infection in humans.     

Clostridium difficile vaccine and serum immunoglobulin G antibody response to toxin A

There is a strong association between serum antibody responses to toxin A and protection against Clostridium difficile diarrhea. A parenteral C. difficile toxoid vaccine induced very-high-level responses to anti-toxin A immunoglobulin G (IgG) in the sera of healthy volunteers. After vaccination, the concentrations of anti-toxin A IgG in the sera of all 30 recipients exceeded the concentrations that were associated with protection in previous clinical studies. Furthermore, the median concentration of serum anti-toxin A IgG in the test group was 50-fold higher than the previous threshold. These findings support the feasibility of using a vaccine to protect high-risk individuals against C. difficile-associated diarrhea and colitis.     

Inhibition of in vitro cell adherence of Clostridium difficile by Saccharomyces boulardii

The influence on the adherence of Clostridium difficile to Vero cells of the yeast Saccharomyces boulardii, the yeast fractions (cytoplasm and cell wall) and the culture supernatant was investigated in vitro. C. difficile adherence was significantly inhibited when bacteria were pre-incubated with the whole yeast and the cell wall fraction; this adherence inhibition was dose-dependent. The cell wall fraction also acts upon the target cultured cells inasmuch as the level of adherence was significantly decreased when Vero cells were preincubated with it. The same experiments carried out in the presence of an inhibitor of serine proteases resulted in no inhibition of bacterial adherence. These results suggest that the yeast could inhibit adherence of C. difficile to cells thanks to its proteolytic activity but also through steric hindrance.     

Fixation of Clostridium difficile toxin A and cholera toxin to intestinal brush border membranes from axenic and conventional mice.

We have tested the in vitro binding of Clostridium difficile toxin A (enterotoxin) and cholera toxin to intestinal brush border membranes prepared from either conventional or axenic mice. Membranes from axenic mice were shown to be saturated at a lower toxin A concentration (at least 2.5 times lower). Because there were no significant differences between membranes from axenic and conventional mice in binding at low toxin A concentrations, the presence of the normal microflora seems to increase the number but not the affinity of brush border membrane receptors on the enterocyte surface. Corroborating the in vitro results, we observed that conventional mice were more sensitive to the pathological effects of toxin A given intragastrically than were axenic mice. In contrast, there was no difference in the binding characteristics of cholera toxin between membranes from conventional and axenic mice. We conclude that the presence of the mouse intestinal bacteria increases the number of C. difficile toxin A intestinal receptors but does not influence cholera toxin receptors.     

Cytokine response by human monocytes to Clostridium difficile toxin A and toxin B.

Clostridium difficile toxins A and B isolated from strain VPI 10463 were tested for induction of cytokine release by human monocytes. Toxin B at 10(-12) M activated human monocytes as measured by release of interleukin-1 (IL-1), tumor necrosis factor (TNF), or IL-6. These effects of toxin B were heat labile (51 degrees C, 30 min). Toxin B was as effective as bacterial lipopolysaccharides in inducing IL-1 beta but less effective in inducing TNF or IL-6. Toxin B and lipopolysaccharides were synergistic in induction of IL-1 beta, TNF, and IL-6. The toxin A preparation used was 1,000-fold less active than toxin B. Apart from the difference in activity, the two toxins showed identical patterns of reaction and there was no synergism between them. A short pulse with toxin B was sufficient to trigger IL-1 release. Toxin B was also extremely toxic for monocytes. The toxicity and the induced proinflammatory monokines (IL-1 and TNF) may contribute to the pathogenic mechanisms of C. difficile infection and pseudomembranous colitis.     

Saccharomyces boulardii Protease Inhibits the Effects of Clostridium difficile Toxins A and B in Human Colonic Mucosa

Saccharomyces boulardii is a nonpathogenic yeast used in the treatment of Clostridium difficile diarrhea and colitis. We have reported that S. boulardii inhibits C. difficile toxin A enteritis in rats by releasing a 54-kDa protease which digests the toxin A molecule and its brush border membrane (BBM) receptor (I. Castagliuolo, J. T. LaMont, S. T. Nikulasson, and C. Pothoulakis, Infect. Immun. 64:5225–5232, 1996). The aim of this study was to further evaluate the role of S. boulardii protease in preventing C. difficile toxin A enteritis in rat ileum and determine whether it protects human colonic mucosa from C. difficile toxins. A polyclonal rabbit antiserum raised against purified S. boulardii serine protease inhibited by 73% the proteolytic activity present in S. boulardii conditioned medium in vitro. The anti-protease immunoglobulin G (IgG) prevented the action of S. boulardii on toxin A-induced intestinal secretion and mucosal permeability to [³H]mannitol in rat ileal loops, while control rabbit IgG had no effect. The anti-protease IgG also prevented the effects of S. boulardii protease on digestion of toxins A and B and on binding of [³H]toxin A and [³H]toxin B to purified human colonic BBM. Purified S. boulardii protease reversed toxin A- and toxin B-induced inhibition of protein synthesis in human colonic (HT-29) cells. Furthermore, toxin A- and B-induced drops in transepithelial resistance in human colonic mucosa mounted in Ussing chambers were reversed by 60 and 68%, respectively, by preexposing the toxins to S. boulardii protease. We conclude that the protective effects of S. boulardii on C. difficile-induced inflammatory diarrhea in humans are due, at least in part, to proteolytic digestion of toxin A and B molecules by a secreted protease.     

Human antibody response to Clostridium difficile toxin A in relation to clinical course of infection.

This study investigated whether differences in fecal and serum antitoxin A antibody levels may account for the duration of Clostridium difficile-associated diarrhea (CDAD) and the occurrence of relapses. By an enzyme linked-immunosorbent assay, we tested 40 patients with CDAD including 25 patients without immunodeficiency and 15 patients receiving antineoplastic drugs. Two hundred eighty serum samples and 80 normal stool samples were investigated as controls. In nonimmunocompromised patients, serum immunoglobulin (IgG) and fecal IgA antitoxin A antibody titers were significantly higher in patients who suffered a single episode (n = 21) than in those with relapsing CDAD (n = 4) whose titers were at control levels. Of these 25 patients, eight suffered from diarrhea which lasted for more than 2 weeks. These patients had significantly lower serum- and feces-specific antibody levels than the others who presented symptoms of shorter duration. In cytostatic-treated patients, antitoxin A antibody levels were similar to controls, but relapses occurred in a single case. These data suggest an association between a defective humoral response to toxin A and a more severe form of C. difficile infection. They also indicate that other host-related factors control the severity of CDAD and remain to be elucidated.     

Transcutaneous immunization with Clostridium difficile toxoid A induces systemic and mucosal immune responses and toxin A-neutralizing antibodies in mice

Clostridium difficile is the leading cause of nosocomial infectious diarrhea. C. difficile produces two toxins (A and B), and systemic and mucosal anti-toxin A antibodies prevent or limit C. difficile-associated diarrhea. To evaluate whether transcutaneous immunization with formalin-treated C. difficile toxin A (CDA) induces systemic and mucosal anti-CDA immune responses, we transcutaneously immunized three cohorts of mice with CDA with or without immunoadjuvantative cholera toxin (CT) on days 0, 14, 28, and 42. Mice transcutaneously immunized with CDA and CT developed prominent anti-CDA and anti-CT immunoglobulin G (IgG) and IgA responses in serum and anti-CDA and anti-CT IgA responses in stool. Sera from immunized mice were able to neutralize C. difficile toxin A activity in an in vitro cell culture assay. CDA itself demonstrated adjuvant activity and enhanced both serum and stool anti-CT IgA responses. Our results suggest that transcutaneous immunization with CDA toxoid may be a feasible immunization strategy against C. difficile, an important cause of morbidity and mortality against which current preventative strategies are failing.     

Adaptive immune response to Clostridium difficile infection: A perspective for prevention and therapy

Clostridium difficile infection (CDI) is one of the most important nosocomial illnesses and a major cause of morbidity and mortality. While initial treatment of CDI is usually successful, unprovoked relapses remain an important and frustrating problem. This review examines the literature describing the natural immune response to CDI, and to what extent it can explain the propensity for relapses. In particular, we discuss studies on antibody and, to a lesser extent, B cell and T cell responses in CDI. Despite years of study, there remains incomplete understanding of the natural antibody response to the major pathogenic toxins, TcdA and TcdB, and other bacterial antigens, in CDI. Recent literature suggests that a specific subset of neutralizing antibodies that target the putative carbohydrate‐binding domains of TcdB and possibly TcdA have the greatest protective ability. This is further supported by recent successful clinical trials of a humanized monoclonal antibody to the major toxin TcdB. A better understanding of how and why the most protective adaptive immune response develops may lead to improved vaccine and therapeutic targets for recurrent CDI.     

Sporogenesis and toxin A production by Clostridium difficile

The kinetics of spore production by Clostridium difficile were not paralleled by release of C. difficile toxin A in vitro. Toxin A was not found to be associated with either purified whole spores or spore coats. Residual traces of toxin A detected in spore contents were almost certainly derived from contaminating vegetative cell debris. Thus, toxin A is unlikely to be a spore constituent or associated with sporogenesis.     

A non-haemagglutinating form of Clostridium difficile toxin A.

Analysis of crude culture filtrate of Clostridium difficile by Mono Q-anion exchange fast protein liquid chromatography (FPLC) demonstrated that toxin A had distinct peaks of activity for cytotoxicity and haemagglutination, as also did highly purified toxin A obtained by thyroglobulin affinity chromatography (TG) followed by two sequential anion-exchange chromatographic steps with Q-Sepharose FF and Mono Q. From TG unbound fractions a highly cytotoxic but weakly haemagglutinating variant (toxin A') of toxin A was obtained by Q-Sepharose FF and Mono Q chromatography. Analysis of toxins A and A' from cultures of C. difficile in a chemically defined medium, and of toxin A dialysed against brain heart infusion broth, indicated that A' was not merely toxin A coupled to a component of the growth medium. Polyacrylamide gel electrophoresis under non-denaturing conditions showed that toxins A and A' had the same Mr. Immunoblotting with mouse monospecific A antitoxin showed that five bands larger than the major 240-Kda band were more strongly developed in toxin A than in A' in denaturing but non-reducing conditions, and in reducing conditions eight bands (38-175 Kda) were seen in toxin A but not A'. Immunoblotting with a monoclonal antibody (PCG-4) showed that, in both reducing and non-reducing conditions, two bands of 160 and 155 Kda were more prominent in toxins A and A' respectively, and four bands (195, 180, 175 and 125 Kda) were detected only in toxin A'.     

Clostridium difficile toxins A and B inhibit human immune response in vitro.

Two Clostridium difficile toxins isolated from strain VPI 10463 were tested for their effect on different human T-cell proliferation systems. In mitogen- and antigen-driven T-cell proliferation systems, toxins inhibited the proliferative response in a dose-dependent fashion. In interleukin-2-driven culture systems, no effect of toxins could be found on preactivated T cells. We suspected that monocytes were the influenced cells, since in antigen- and mitogen-driven systems monocytes were necessary for the proliferative response, whereas the interleukin-2-driven system was independent of monocytes. To prove this concept, purified monocytes were treated with toxins. The treatment was found to markedly reduce the capacity of monocytes to stimulate T-cell proliferation. No inhibition of the proliferative response was measured when, instead of monocytes, resting or preactivated T cells were treated with toxins. These experiments clearly show that C. difficile toxins interact with monocytes and not with T cells. The effect of toxins on cells of the immune system might be one factor in the development of pseudomembranous colitis.     

Relationship between levels of Clostridium difficile toxin A and toxin B and cecal lesions in gnotobiotic mice.

Various Clostridium difficile strains were studied with respect to their pathogenicity in monoassociated mice in relation to levels of toxin A and toxin B in vivo and in vitro. Two strains which were the most potent toxin producers in vitro induced mortality (100%); mice monoassociated with these strains were found to have high levels of both toxins in their ceca and an intense cecal epithelial ulceration together with a severe inflammatory process. No mortality was observed with the other strains. Strains which were moderately toxinogenic in vitro induced inflammation of the cecum but no ulceration, and no toxin A was found. Inflammation intensity was not related to toxin B levels. After 3 weeks, ceca returned to normal in spite of a chronic cytotoxin production. When compared with in vitro results, which showed a good correlation between the levels of the two toxins, toxin A amounts in vivo were found to be lowered relative to toxin B levels. The lack of detectable toxin A levels in animals infected with all but the two most highly toxinogenic strains prevented death. This work points out the importance of investigation of toxin A for the understanding of C. difficile pathogenicity.     

Humoral immune response as predictor of recurrence in Clostridium difficile infection

Low serum concentrations of antibodies directed against the toxins TcdA and TcdB have been associated with a higher risk of recurrence of Clostridium difficile infection (CDI) after successful antibiotic treatment. However, there are conflicting reports. Herein, we compared serum levels of antibodies of patients with a single episode of CDI with those of patients who subsequently suffered a recurrence. We used a serum bank from patients who received an experimental whey protein product following successful antibiotic treatment for CDI. We determined levels of IgA and IgG directed against TcdA, TcdB and non-toxin cell surface antigens in serum collected directly and 3 weeks after completing a 10-day course of antibiotic treatment for CDI. We also developed an objective flow cytometry-based assay to determine the proportion of cells exhibiting cytopathic effect after exposure to TcdB. Using this method, we measured the TcdB-neutralizing capacity of sera. We compared the results for patients without a subsequent recurrence with those of patients who suffered a recurrence within 60 days after completing the antibiotic treatment. Advanced age, comorbidity other than immunocompromised state and low serum levels of anti-TcdA and anti-TcdB antibodies were associated with recurrence, whereas serum levels of antibodies directed against cell surface antigens were not. Serum TcdB-neutralizing capacity, which correlated only weakly with serum IgG anti-TcdB, was not significantly associated with recurrence.     

Enzyme-linked immunosorbent assay for Clostridium difficile toxin A.

Antibodies against Clostridium difficile toxin A were purified by affinity chromatography from antiserum prepared against crude C. difficile toxin preparations. The affinity-purified antibody preparation was free of detectable amounts of antibodies to other C. difficile antigens, as demonstrated by crossed immunoelectrophoresis, and specifically neutralized the cytotoxicity of toxin A. An indirect enzyme-linked immunosorbent assay (ELISA) was subsequently developed using the antibody preparation for the specific detection of toxin A. The ELISA, which could detect 1 ng (5 ng/ml) of toxin A, was used to quantitate the toxin in the culture supernatant fluids of strains of C. difficile. The ELISA values for toxin A closely correlated with the toxin A and B cytotoxic titers of the supernatant fluids. In addition, toxin A was detected by ELISA in human fecal specimens from persons with antibiotic-associated colitis, demonstrating that this toxin is produced during C. difficile colitis.     

Rapid detection of Clostridium difficile toxin A in stool specimens

Objective: To evaluate a rapid (15-min) enzyme immunoassay in the format of an individual cassette (ImmunoCard toxin A, Meridian, BMD, Marne-la-Vallée, France) for the detection of Clostridium difficile toxin A in stool specimens.
Methods: We compared this new test with the cytotoxicity assay using MRC-5 cells, the ToxA test (TechLab, BioWhittaker, Fontenay-sous-bois, France) and toxigenic culture for the diagnosis of C. difficile -associated diseases (CDAD). A total of 236 stool specimens collected from 220 patients was simultaneously tested with the four methods. Discordant results were resolved by reviewing patients' clinical records.
Results: The prevalence of CDAD was 13.9%. Test sensitivities and specificities were 100% and 99% respectively for the cytotoxicity assay, 87.5% and 100% for ImmunoCard toxin A, 77.4% and 100% for the ToxA test and 100% and 98% for toxigenic culture.
Conclusions: The ImmunoCard Toxin A is a very rapid, individual and easy-to-perform test for the diagnosis of CDAD. It provides same-day results and may be useful for both guiding appropriate treatment and controlling nosocomial spread of C. difficile.     

Molecular characterization of the Clostridium difficile toxin A gene

The gene encoding the toxin A protein of Clostridium difficile (strain VPI 10463) was cloned and sequenced. The coding region of 8,133 base pairs had a mol% G + C of 26.9 and encodes 2,710 amino acids. The deduced polypeptide has a molecular mass of ca. 308 kilodaltons. Nearly a third of the gene, at the 3' end, consists of 38 repeating sequences. The repeating units were grouped into two classes, I and II, on the basis of length and the low levels of DNA sequence similarities between them. There were seven class I repeating units, each containing 90 nucleotides, and 31 class II units, which, with two exceptions, were either 60 or 63 nucleotides in length. On the basis of DNA sequence similarities, the class II repeating units were further segregated into subclasses: 7 class IIA, 13 class IIB, 5 class IIC, and 6 class IID. The dipeptide tyrosine-phenylalanine was found in all 38 repeating units, and other amino acid sequences were unique to a specific class or subclass. This region of the protein has epitopes for the monoclonal antibody PCG-4 and includes the binding region for the Gal alpha 1-3Gal beta 1-4GlcNAc carbohydrate receptor. Located 1,350 base pairs upstream from the toxin A translation start site is the 3' end of the toxin B gene. Between the two toxin genes is a small open reading frame, which encodes a deduced polypeptide of ca. 16 or 19 kilodaltons. The role of this open reading frame is unknown.     

Genetic control of immune response to staphylococcal exfoliative toxin A in mice.

Different inbred and congenic resistant strains of mice were immunized with staphylococcal exfoliative toxin A (ETA). In antibody responses measured in sera of mice by a passive hemagglutination technique, A/J, DBA/2, BALB/c, B10A, B10D2, B10S, and A.SW were high responders. C57BL/10 (B10), A.BY, and DBA/1 were low responders. The congenic C3H/HeJ and C3H.SW mice were, respectively, high and low responders. The observation that the immune responses of the mice to ETA were closely linked with the haplotypes of their H-2 complexes suggests the existence of an H-2-linked immune response (Ir) gene coding for the production of humoral antibodies to ETA. Four B10A recombinants were used to map this gene within the H-2 complex. The finding that B10A(2R) and B10A(4R) were high responders, whereas B10A(3R) and B10A(5R) were low responders, indicates that the gene controlling antibody response to ETA is located in the I-A subregion or the H-2K end within the H-2 complex. We wish to propose the name Ir-ETA for this gene. The function of Ir-ETA seems to be at least related to antigen recognition at the T-lymphocyte level. Neonatal mice are generally susceptible to ETA regardless of their H-2 haplotypes. However, the neonatal mice born to a high-responder mother immunized with ETA were resistant to the subcutaneous challenge of ETA, but those born to an immunized low-responder mother were susceptible to the challenge. This result suggests that if the mother is a high responder to ETA and is effectively immunized with ETA, the maternal immunity makes it possible to neutralize this toxin in neonatal mice.