Clostridium difficile in gnotobiotic mice.

Germfree mice associated with Clostridium difficile developed intestinal disease characterized by polymorphonuclear cell infiltration of the lamina propria, diarrhea, and cecal cytotoxin concentrations positive at a 10(-6) dilution. The numbers of viable bacteria never exceeded 10(10) colony-forming units per g (dry weight). Despite the high toxin levels and chronic inflammation over a 30-day period, the mortality rate was low (less than 2%). Daily treatment of these animals with two oral doses of 2 mg of vancomycin resulted in stool levels of greater than 200 micrograms/ml, well in excess of the minimum inhibitory concentration for C. difficile. This therapy decreased viable cell density by 2 to 3 logs and increased the spore counts from 10(5.8) to 10(7.8) colony-forming units per g (dry weight) by day 7, and animals were free of detectable toxin. However, once therapy was stopped, viable bacteria and spore counts and cytotoxin concentrations returned to previous levels. Treatment of mice with concentrations of clindamycin shown to be inhibitory in vitro had no effect on C. difficile toxin titers or bacterial counts, although the appearance of a clindamycin-resistant population was noted. These data indicate that vancomycin, given orally, decreases the concentration of toxin, but C. difficile survive as spores. By contrast, large populations of vegetative cells and high cytotoxin levels persist when clindamycin is used, even at an inhibitory concentration.     

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.     

Protection against experimental pseudomembranous colitis in gnotobiotic mice by use of monoclonal antibodies against Clostridium difficile toxin A.

The pathogenicity of Clostridium difficile is due to the production of two toxins (toxins A and B). We prepared monoclonal antibodies against toxin A and determined whether axenic mice passively immunized with the monoclonal antibodies were protected against C. difficile disease. The mice were kept in an isolator and were given ascites fluid intravenously prior to challenge with a toxinogenic strain of C. difficile. Control mice and mice receiving ascites fluid devoid of toxin antibody died within 2 days and had high levels of toxins A and B in their feces. Mice that received ascites fluid containing high amounts of toxin A monoclonal antibodies directed against the repeating units of the toxin survived. In protected mice, toxin B levels were similar to those in dying mice, but toxin A levels were greatly reduced. These data show that passive immunity induced by monoclonal antibodies against toxin A was effective against pseudomembranous cecitis.     

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.     

Criteria for detection of Clostridium difficile toxin production by counterimmunoelectrophoresis.

Counterimmunoelectrophoresis (CIE) was compared with a cytotoxicity assay and isolation of toxin-producing Clostridium difficile for the diagnosis of antibiotic-associated diarrhea. Standardized criteria for CIE positivity were established as two major precipitin lines with specific C. difficile antitoxin. Using these criteria on 425 fecal specimens, we determined the sensitivity of CIE (80%) to be comparable to that of the cytotoxicity assay (82%). Both methods were highly specific. With carefully controlled techniques, CIE can be used in routine diagnostic laboratories to screen for toxigenic C. difficile.     

Detection of Clostridium difficile toxin in various tissue culture monolayers.

Thirty stool filtrates known to contain Clostridium difficile toxin based on previous testing on McCoy cells were tested for toxicity on primary African green monkey kidney (AGMK), McCoy, MRC-5, primary rhesus monkey kidney (RMK), and Vero cells. All 30 filtrates showed cytotoxic effect at greater than or equal to 1:100 dilution on McCoy and Vero cells. A total of 22 filtrates were positive on MRC-5 monolayers, while only 16 and 10 filtrates showed positive cytotoxic effect on AGMK and RMK cells, respectively. Another 630 stool specimens were tested on McCoy and Vero cells only. Of these stool filtrates, 70 were positive and 560 were negative with both cell lines, which thus gave 100% agreement. Vero cells can be used interchangeably with McCoy cells for the detection of C. difficile toxin in stool filtrates.     

Neutralization of Clostridium difficile toxin by Clostridium sordellii antitoxins.

Neutralization of Clostridium difficile toxin by Clostridium sordellii antitoxin was studied by cytotoxicity assay in tissue culture. The sources of toxin were stools from two patients with pseudomembranous colitis and a culture filtrate of C. difficile isolated from one of the patients. C. sordellii antitoxin was available either in monovalent form or as gas gangrene polyvalent antitoxin. The potency of antitoxins against C. difficile determined by cytotoxicity assay did not correlate with the established values reported for mouse protection tests against C. sordellii toxin. An equivalent zone of optimal neutralization was demonstrated for stool toxin, and a slightly different one for culture toxin. The rate of neutralization appeared to be instantaneous, either at 24 or at 37 degrees C. The efficacy of antitoxin in preventing cytotoxicity in cultured cells preexposed to toxin decreased rapidly with preexposure time. The union between toxin and antitoxin could be readily dissociated by simple dilution or by ammonium sulfate precipitation followed by dissociated by simple dilution or by ammonium sulfate precipitation followed by dilution. Continued incubation of toxin-antitoxin mixture did not increase the firmness of the union; on the contrary, more dissociation occurred. The unusual looseness of the toxin-antitoxin union is probably relatd to lack of serological specificity or affinity. Based on these observations, a practical diagnostic method for antibiotic-induced colitis is outlined.     

Identification of toxigenic Clostridium difficile strains by using a toxin A gene-specific probe.

A 4.5-kilobase PstI fragment encoding part of the toxin A gene was isolated and used as a DNA probe in colony hybridization studies with 58 toxigenic and 17 nontoxigenic Clostridium difficile strains. All 58 toxigenic strains showed positive hybridization, in contrast to the 17 nontoxigenic strains. Southern blot analysis with the toxin A gene probe showed hybridization to a single fragment of equal intensities for HindIII-digested genomic DNAs isolated from C. difficile strains of wide-ranging toxin production. The positive hybridization signals were due to fragments of heterogeneous lengths (9 to 13 kilobases) for toxigenic strains of different types but were absent for the nontoxigenic strains. These results suggest the presence of a single copy of the toxin A gene on the genome of C. difficile strains, and the wide variation of toxin expression is not a reflection of gene copy number. The lack of toxin activity for nontoxigenic strains can be explained by the absence of at least part of the toxin A gene. The toxin A gene probe was tested against clostridial strains from 18 other species, of which only toxigenic C. sordellii strains showed positive hybridization. The specificity of the toxin A gene probe for toxigenic strains may lead to improved methods for the specific identification of toxigenic C. difficile strains from clinical specimens.     

Linkage between toxin production and purine biosynthesis in Clostridium difficile.

The production of toxins A and B by Clostridium difficile was greatly enhanced under biotin-limited conditions, in which a 140-kDa protein was expressed strongly. Gene cloning revealed that this protein was a homologue of formylglycinamidine ribonucleotide synthetase (FGAM synthetase, EC 6.3.5.3), which is known as PurL in Escherichia coli and catalyses the fourth step of the de novo purine biosynthesis pathway. This enzyme consisted of a single polypeptide, although FGAM synthetases of gram-positive bacteria usually consist of two subunits. Inhibition of the enzymic activity of C. difficile PurL by O-diazoacetyl-L-serine (azaserine) resulted in enhanced toxin B production even in biotin-sufficient conditions. In contrast, blockade of the preceding step of the PurL catalysing step by sulfamethoxazole inhibited toxin B production almost completely. These results suggest that accumulation of formylglycinamide ribonucleotide (FGAR), a substrate of FGAM synthetase, enhances toxin production by C difficile and depletion of FGAR reduces toxin production.     

Interrelationships between digestive proteolytic activities and production and quantitation of toxins in pseudomembranous colitis induced by Clostridium difficile in gnotobiotic mice.

Clostridium difficile pathogenicity is related to in vivo production of toxins, and it is of great interest to detect toxins produced in biological samples. Several reports have shown that proteases in stools interfere with immunological methods for quantitation of toxin A. The purpose of this work was to estimate the relationship between the proteases and the C. difficile toxins produced in a gnotobiotic mouse model of pseudomembranous cecitis. Cecal proteolytic activities hydrolyzed toxin A, and immunoglobulin G bound to the microtiter plate used in immunoassays. This interference could be blocked by the addition of trypsin inhibitor to the samples. The ability of soluble toxin A to bind to bound antibodies in an enzyme-linked immunosorbent assay was not affected by the proteases, but the biological activity was reduced 100-fold. The cytotoxicity of toxin B was not modified by proteolytic activity treatment. Mice inoculated with a low toxin A-producing strain of C. difficile did not died, and no modulation of proteolytic activities occurred. After inoculation with the lethal VPI strain of C. difficile, toxins A and B reached maximum levels in the ceca at 12 h postinfection. At this time, the proteolytic activities did not decrease from the levels seen at zero time. Mice died within 2 days. At this time (about 32 postinfection), proteolytic activities were sharply decreased in the lower parts of the digestive tracts. The findings that serum inhibited the proteases and that there was a 100-fold increase in serum-derived mouse immunoglobulins in the lumen as the C. difficile infection progressed suggest that the decrease in protease activity in the lower digestive tract may be related to the exudation of serum from the inflammation process.     

Experimental reproduction of neonatal diarrhea in young gnotobiotic hares simultaneously associated with Clostridium difficile and other Clostridium strains.

Clostridium difficile, C. perfringens, and C. tertium are very often present simultaneously in the feces of conventional diarrheic young hares, whereas these three bacterial species are rarely encountered and never present simultaneously in the feces of healthy young hares. When a strain of each of the three bacterial species was monoassociated with axenic young hares, the appearance of pathological disorders was only observed in animals monoassociated with C. difficile, when the number of C. difficile exceeded 10(8) per g of fresh feces. When a strain of C. perfringens or a strain of C. tertium, or both, was associated with C. difficile, diarrhea and death occurred more rapidly than in hares monoassociated with C. difficile. C. difficile and C. perfringens became established more rapidly when disassociated than when monoassociated with axenic hares. The association of C. perfringens and C. tertium with axenic hares did not bring about any pathological disorders. It may be concluded that C. difficile is the causal agent of neonatal diarrhea in conventional and gnotobiotic young hares and that other strains of Clostridium enhance its pathogenic effect. C difficile alone or associated with C. perfringens or C. tertium does not play any pathogenic role in young rats, mice, or rabbits.     

Role of volatile fatty acids in colonization resistance to Clostridium difficile in gnotobiotic mice.

Clostridium difficile is an agent involved in the development of antibiotic-associated pseudomembranous colitis. The purpose of this work was to investigate the role of volatile fatty acids (VFAs) in resistance to colonization by C. difficile by using a gnotobiotic animal model. Accordingly, germfree mice were associated with different hamster flora, and the VFAs in their cecal contents were measured by gas chromatography. The results showed that VFAs were produced mainly by the intestinal flora, especially by the strictly anaerobic bacteria. In these associated mice, the concentrations of acetic, propionic, and butyric acids were higher than those of other acids, but at pH 6.8 the MICs of these three acids in vitro for C. difficile were more than 200 mu eq/ml. In gnotobiotic mice monoassociated with C. difficile and in the isolated ceca of these mice, VFAs did not inhibit the growth of C. difficile. In gnotobiotic mice which were diassociated with C. difficile and C. butyricum and given drinking water with a lactose concentration of 20%, the cecal contents included about the same amount of butyric acid as did those of the monoassociated mice, although the population of C. difficile remained the same. Therefore, it is suggested that VFAs alone cannot inhibit intestinal colonization by C. difficile and that, consequently, other inhibitory mechanisms are also present.     

Emergence in gnotobiotic mice of nontoxinogenic clones of Clostridium difficile from a toxinogenic one.

In previous studies, we showed that diet composition or Saccharomyces boulardii ingestion could protect gnotobiotic mice against lethal Clostridium difficile infection. Using an original method, we detected nontoxinogenic clones from feces of protected mice challenged with a toxinogenic clone of C. difficile. These clones became established at the same level as the toxinogenic one after about 30 days. In these protected mice bearing nontoxinogenic clones, no enterotoxin production could be detected and cytotoxin titers were highly reduced. These nontoxinogenic clones were genetically stable because nontoxinogenic clones and clones that produce intermediate levels of toxins in vivo did not revert to toxin production, even after repeated culture in vitro. Furthermore, the nontoxinogenic clones were shown to arise from a single toxinogenic clone and were identical to that clone in metabolic patterns and antibiotic sensitivity tests. When mice fed a nonprotective diet were challenged with a nontoxinogenic or intermediate clone, they remained healthy and no toxin production could be detected in their feces. Moreover, these mice were protected against further infections with toxinogenic strains of C. difficile, and a strong antagonism between nontoxinogenic and toxinogenic clones was observed.     

Rapid identification of Clostridium difficile by toxin detection.

Rapid identification of Clostridium difficile in a stool specimen could be accomplished within 24 h by detection of toxin elaborated in an agar or broth culture containing cycloserine and cefoxitin. Broth culture seemed to give a more rapid and sensitive result than the agar plate culture. For cultivation of C. difficile in stool, we recommend the use of chopped meat broth and blood agar plate, the former for toxin detection in 1 to 2 days and the latter for colonial morphology and isolation of a pure culture.     

Nontoxigenic strains of Clostridium difficile lack the genes for both toxin A and toxin B.

A total of 39 toxigenic and 20 nontoxigenic strains of Clostridium difficile were tested for the presence of either toxin A or toxin B by the polymerase chain reaction (PCR). All toxigenic strains produced cytotoxin as assayed by using highly sensitive fetal lung fibroblasts and were positive for toxin A as well as toxin B in the PCR assay. All nontoxigenic strains failed to produce toxin and were negative in the PCR assay. This study shows that nontoxigenic strains of Clostridium difficile lack the toxin A as well as the toxin B gene.     

Localization of two epitopes recognized by monoclonal antibody PCG-4 on Clostridium difficile toxin A.

The toxin A gene of Clostridium difficile contains a 2.5-kb region encoding a series of contiguous repeating units located at the COOH terminus of the molecule. We previously showed that the monoclonal antibody (MAb) PCG-4, which neutralizes the enterotoxic activity of toxin A, binds to epitopes located within these repeating units. In the present study, we subcloned a series of fragments from this portion of the gene. The recombinant peptides expressed from the gene fragments were examined for reactivity with MAb PCG-4 to identify the epitopes involved in binding. Our results showed that MAb PCG-4 recognizes epitopes in amino acid residues 2097 through 2141 and amino acid residues 2355 through 2398.     

Serogrouping of Clostridium difficile strains by slide agglutination.

Six different agglutinating antisera were obtained by immunizing rabbits with Formol-treated strains of Clostridium difficile. After appropriate absorption, these antisera were used to define six serogroups designated by the letters A, B, C, D, F, and G. Altogether, 315 strains of C. difficile from various origins were tested for slide agglutination by these antisera; 312 (99%) of them were agglutinated by one of these antisera. A and C were the most common serogroups. An excellent correlation, ranging from 85 to 100%, was found between the serogroup and the toxigenicity of the strains. The correlation between serogroup and sorbitol fermentation was higher, ranging from 89 to 100%. The results of this typing were compared with the clinical origin of the strains. Only strains of serogroups A, C, and D were isolated in 153 cases of antibiotic-associated diarrhea. This series included strains from three outbreaks; all the strains in two of the outbreaks belonged to serogroup C, and in the third, all the strains belonged to serogroup A. Strains of serogroups B, F, and G were only found in the stools of asymptomatic neonates or young children. In the latter samples, strains of serogroups A and D were found in the same ratio as in adults with antibiotic-associated diarrhea, but strains of serogroup C were seldom isolated. In patients treated with antineoplastic drugs and suffering from diarrhea, the distribution of the strains was the same as in cases of antibiotic-associated diarrhea.     

Interaction of Clostridium difficile and Escherichia coli with microfloras in continuous-flow cultures and gnotobiotic mice.

We studied the interactions between the entire cecal flora of hamsters and the pathogens Clostridium difficile and Escherichia coli in gnotobiotic mice and in a continuous-flow (CF) culture system in which the growth medium consisted of an extract of fecal pellets from germfree mice. CF cultures and germfree mice were colonized first with C. difficile and E. coli and then with the cecal flora of hamsters. Both in vivo and in vitro hamster flora markedly suppressed the potential pathogens. Contents of CF cultures inoculated with hamster flora were introduced into gnotobiotic mice previously colonized with C. difficile and E. coli. These mice were compared with mice given homogenates of hamster ceca. In both groups, the C. difficile population decreased by a factor of more than 10(6) and the E. coli population decreased by a factor of 10(4) to 10(5). CF culture contents also reduced the size of the dilated germfree mouse cecum to normal. When veal infusion broth was used as a medium, contents of CF cultures colonized with hamster flora failed to eliminate C. difficile from mice. Thus, the extract of fecal pellets appeared to contain a substance important for sustained colonization by important components of the cecal flora. We also studied the ability of collections of isolates to suppress the potential pathogens in both gnotobiotic mice and CF cultures. A total of 150 isolates obtained from predominant hamster flora at the ecologic climax stage (C flora) suppressed C. difficile and E. coli to 10 and 1 to 3%, respectively, of the population sizes attained in monoassociated mice. A total of 67 isolates obtained during ecologic succession combined with a C flora consisting of 100 isolates suppressed the potential pathogens to 0.3 and 0.03% of their original levels, respectively. Similar degrees of suppression occurred in CF cultures, further indicating that anaerobic CF cultures are promising models for investigation of the microbial ecology of C. difficile.