Clostridium difficile: Its Disease and Toxins - Clinical Microbiology

CLINICAL MICROBIOLOGY REVIEWS, Jan. 1988, p. 1-18 Vol. 1, No. 10893-8512/88/010001-18$02.00/0Copyright © 1988, American Society for Microbiology

Clostridium difficile: Its Disease and ToxinsDAVID M. LYERLY,* HOWARD C. KRIVAN, AND TRACY D. WILKINS

Department ofAnaerobic Microbiology, Virginia Polytechnic Institute and State University,Blacksburg, Virginia 24061

INTRODUCTION................................................................................................................... 1HISTORICAL PERSPECTIVES ...................................................... 2C. DIFFICILE DISEASE IN EXPERIMENTAL ANIMALS ...................................................... 2C. DIFFICILE DISEASE IN HUMANS...................................................... 3CLINICAL DIAGNOSIS

Pathology........................................................................................................................... 4Isolation and Identification of the Organism ...................................................... 4Detection of Toxin ...................................................... 5Problems Associated with Clinical Diagnosis of the Disease ...................................................... 6

TOXINS A AND BRole of Toxins A and B in PMC ...................................................... 7Factors Affecting Toxin Production ...................................................... 7Purification........................................................................................................................ 7Physicochemical Properties ...................................................... 9Biologic Activities ...................................................... 9Immunochemical Properties ...................................................... 10Toxin Receptors ...................................................... 10Cloning of the Toxin A Gene...................................................... 11

OTHER VIRULENCE FACTORS ...................................................... 11PROPERTIES OF C. SORDELLII ENTEROTOXIN AND CYTOTOXIN...........................................12FUTURE DIRECTIONS ...................................................... 12ACKNOWLEDGMENTS ...................................................... 12LITERATURE CITED ...................................................... 12

INTRODUCTION

In the 1960s, anaerobic bacteria began to be recognized asthe predominant organisms in the human large bowel andtheir roles as agents of disease in humans were established.Escherichia coli was removed from its false position as themost numerous organism in the human colon, and many"sterile abscesses" were found to be caused by anaerobicbacteria. Most antibiotics used at that time to sterilize thebowel were quite active against facultative bacteria but hadlittle effect on the anaerobic population. For example, someof the more commonly used antibiotics such as kanamycinand other aminoglycosides were not active on anaerobes. Asa result, pharmaceutical companies began to search foragents with improved activity against anaerobes. One of theearly successes was clindamycin. This antibiotic was aderivative of a successful old-line antibiotic, lincomycin, butit was much more active against anaerobes and reached thecolon via excretion in the bile. In the early 1970s, clindamy-cin began to be used for treating serious anaerobic infec-tions, and it proved to be a significant improvement overprevious types of therapy. Unfortunately, reports on thedeaths of patients treated with clindamycin appeared as theantibiotic became widely used (90, 106, 190, 222, 226, 227).

Diarrhea had been a common side effect of therapy withthe parent compound, lincomycin. Patients treated withclindamycin, however, sometimes progressed beyond diar-rhea to a severe inflammation of the colonic mucosa with theelaboration of pseudomembranes composed of fibrin, mu-

* Corresponding author.

cous, necrotic epithelial cells, and leukocytes. In somepatients this pseudomembrane formed a sheath over theentire colonic mucosa. Such sheaths were seen most com-monly at autopsy, and the disease, pseudomembranouscolitis (PMC), was thought to have a high mortality rate. Theincidence of PMC varied widely between different hospitalsand even between different wards in the same hospital. Someinvestigators reported rates as high as 10% in patientstreated with clindamycin (227), but most later studies re-ported much lower rates (90, 190). The higher incidence insome studies probably resulted from transmission within thehospital prior to the recognition that this was a nosocomialinfection. The increase in the number of cases of PMC inpatients treated with clindamycin is the factor which stimu-lated interest in the disease, but it soon became apparent thatpatients treated with other antibiotics developed this dis-ease. The early designation for the disease, "clindamycincolitis," thus was inappropriate, much to the relief of TheUpjohn Company, who marketed the drug. PMC createdquite a stir among investigators because it represented a"new" disease that was caused by the treatment given forother diseases and because it killed patients.PMC actually had been described in patients prior to the

start of the antibiotic era, but the number of cases increaseddramatically after antibiotics began to be used (86). It wasspeculated up until the late 1970s that the disease might becaused by Staphylococcus aureus. This organism, however,was not the cause of the numerous cases of PMC reported inthe 1970s and 1980s, and there are doubts as to whether S.aureus actually has caused any cases of the disease. It has

1

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

2 LYERLY ET AL.

now been established that the disease is almost alwayscaused by Clostridium difficile.

C. difficile produces two potent lethal toxins, an entero-toxin (toxin A) and a cytotoxin (toxin B) (12, 13, 220, 224). Itis the production of these toxins by C. difficile in the gutwhich ultimately leads to the disease. As a result, mucheffort has been made to develop detection methods for thesetoxins and to learn how they act. In the first portion of thisreview, we describe some of the clinical aspects ofPMC andevaluate some of the methods which have been used todetect the toxins of C. difficile in clinical specimens. Thesecond portion of the article covers the biochemical infor-mation available on the toxins and discusses their unusualcharacteristics.

HISTORICAL PERSPECTIVES

The disease PMC has been recognized for quite sometime. One of the earliest descriptions of the disease wasreported in 1893 by Finney (71), who noted diphtheriticpseudomembranes and hemorrhagic diarrhea in a youngwoman following surgery. The number of reported cases ofPMC increased dramatically, however, following the wide-spread use of broad-spectrum antibiotics.The causative agent of PMC, C. difficile, also has been

recognized for some time. The organism was initially iso-lated from the stools of healthy newborn infants by Hall andO'Toole in 1935 (93). These investigators referred to theorganism as Bacillus difficilis because of the difficulty theyencountered in the isolation of the organism. These investi-gators were also the first to show that the organism istoxigenic. This observation was based on the finding thatbroth cultures and culture filtrates of the organism causedlesions, respiratory arrest, and death when injected intorabbits and guinea pigs. They proposed the interesting ideathat the toxin was a neurotoxin because it caused convul-sions in animals.

C. difficile was not known to be a pathogen so the "toxin"of the organism was not studied in detail until the late 1970s,when the association of C. difficile with PMC becameapparent. At that time, it was shown that stools frompatients with PMC contained high levels of cytotoxic activ-ity. Initially, it was suspected that the activity was due toviruses or perhaps mycoplasma, but tests for these agentswere consistently negative. Investigators then began tosuspect a toxin-producing bacterium, and the competition toidentify the etiologic agent became even more intense.Results from several laboratories showed that the cytotoxicactivity was neutralized by gas gangrene antiserum. Gasgangrene is a disease caused by several species of clostridia,and the antitoxin is a mixture of antisera made to crude toxinpreparations from several clostridial species. When theindividual antisera were tested, only the antiserum againstClostridium sordellii neutralized the cytotoxicity. This ledinvestigators to believe that this organism might be causingPMC. C. sordellii, unfortunately, was almost never isolatedfrom patients with the disease. Soon it was demonstratedthat C. difficile was present in high numbers in patients withPMC and that the organism produces a cytotoxin which isneutralized by C. sordellii antiserum (17, 19, 20, 79, 82,125-127, 198). We now know that C. difficile and C. sordelliiproduce toxins that are almost identical.Once these findings were reported, researchers began to

provide evidence documenting the role of C. difficile in thepathogenesis of PMC. One of the key contributions was thedevelopment of a suitable animal model for studying the

disease. It had been shown by Small in 1968 that hamstersinjected with lincomycin developed severe enterocolitis anddied (214), but these observations were not extended toantibiotic-associated disease in humans. In the late 1970s, itwas shown that hamsters that died from cecitis after treat-ment with antibiotics contained high numbers of toxigenic C.difficile in their stools and that the toxin in the stools of theseanimals was very similar to the toxin found in stools fromPMC patients (6, 18, 22, 196, 197). On the basis of thesefindings, the hamster model began to be used extensively inthe study of the disease and is still in use today.

Until 1980, it was believed that the cytotoxic factor wasthe only toxin produced by C. difficile. There were reportsdescribing enterotoxic activity in fecal extracts from ham-sters with antibiotic-associated cecitis (99, 197), but therewas no indication that this activity and the potent cytotoxicactivity represented two different toxins. Consequently,much of the early literature is rather confusing because somegroups were working primarily with the cytotoxin whereasothers were working mostly with the enterotoxin. Then,within a short time, Taylor et al. (224; N. S. Taylor, G. M.Thorne, and J. G. Bartlett, Clin. Res. 28:285, 1980) andBanno et al. (13) demonstrated that the cytotoxic activitycould be separated from the enterotoxic activity by anion-exchange chromatography. These findings, along with re-sults from other laboratories, provided conclusive evidencethat the organism produces at least two distinct toxins.The cytotoxin (i.e., the toxin detected in fecal specimens)

is referred to as toxin B. The enterotoxin (i.e., the toxinresponsible for the fluid response in the animal ileal loopmodel) is referred to as toxin A. The designations A and Brefer to the elution pattern of the toxins on anion-exchangeresins. Toxin A binds less tightly to the resin than toxin Band elutes before toxin B. The designations toxin D-1 andtoxin D-2 have been used to denote the enterotoxin andcytotoxin, respectively, in the articles by Banno et al. (12,13).

C. DIFFICILE DISEASE IN EXPERIMENTALANIMALS

C. difficile causes antibiotic-associated disease in a num-ber of animal species, including hamsters, guinea pigs, andrabbits, but most of the work has been done with hamsters(1, 44, 51, 53, 68, 100, 104, 116, 141, 166, 168, 188, 192, 193,219, 245). As with PMC in humans, the hamster disease canbe initiated with a variety of antibiotics (1, 44, 68, 141, 188).If toxigenic strains of the organism are present in theenvironment, the animal usually develops diarrhea and diesfrom severe enterocolitis. The disease in hamsters is local-ized in the cecum (proximal colon) with some involvement ofthe ileum. PMC in humans, on the other hand, occursprimarily in the distal colon. Coverings of inflammatorydebris consisting of erythrocytes, inflammatory cells, bacte-ria, and sloughed mucosal cells are sometimes present in thececa of hamsters with the disease, and these coverings arevery similar to the pseudomembranes observed in the colonof many PMC patients. The hamster model did not meet withuniversal acceptance; some pathologists argued that thedifferences between the hamster and human disease meantthat there was a different cause. Fortunately, the investiga-tors working with hamsters did not believe this.One of the most interesting observations with C. difficile

infections in humans is that infants can have high numbers oftoxigenic C. difficile and high levels of toxins A and B in theirstools and be completely asymptomatic (see next section).

CLIN. MICROBIOL. REV.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

C. DIFFICILE TOXINS 3

This unusual phenomenon is also seen in infant hamsters (88,101, 203). The reasons for this are not known, but the resultsobtained with this model should provide some insight intowhy most human infants colonized with toxigenic C. difficiledo not become ill.

In addition to the results obtained with the conventionalhamster model described above, studies on the disease havebeen done in germfree mice and rats (51, 53, 166, 219, 245).Neither of these animal species is as susceptible to thedisease as hamsters. Gnotobiotic animals of each speciescan, however, be colonized with the organism. These mo-

noassociated animals develop some intestinal pathology, butcompared with conventional hamsters, only a low percent-age die from the infection. Unlike the disease in hamsters,the disease in germfree mice and rats involves the colon, andpseudomembranes are evident along the colonic wall in theinfected animals. The disease in these animals developsmore slowly than that in hamsters, and perhaps the chronicnature of the disease results in the production of the pseu-

domembranes. In this regard, it has been suggested thatthese animal models may be better suited than the hamstermodel for the study of PMC. However, these models are

considerably more expensive to use and the fact that no

other organisms are present makes the situation unlike PMCin humans. Gnotobiotic hamsters would be interesting tostudy, but unfortunately hamsters are almost impossible toraise in the germfree state.

C. DIFFICILE DISEASE IN HUMANS

The normal bacterial flora in the gut serves as the majorbarrier against colonization by pathogens. When the flora isdisturbed in some manner, the host becomes susceptible tocolonization or overgrowth by the pathogen. Antibiotics, theprimary predisposing agent of PMC, act in this fashion byupsetting the normal flora of the bowel. The numerousreports of PMC which appeared following the use of clinda-mycin led investigators to believe that this antibiotic was theonly one responsible for causing the disease. Actually,ampicillin and cephalosporins cause more cases than clinda-mycin (177, 202, 213). These antibiotics are widely pre-

scribed, and that is why they are associated with PMC moreoften than other antibiotics. We now know that almost anyantibiotic can cause the disease (14, 25, 32, 39, 49, 55, 56, 76,80, 189, 213, 231) and that very rarely the disease can even

occur without prior antibiotic therapy (63, 157, 170, 239). Inaddition, treatment with antineoplastic agents which haveantibacterial activity can cause the disease (52).Most cases of PMC result from nosocomial infections.

This is apparent from the increasing number of reportsdescribing outbreaks of the disease and from the fact that C.difficile is part of the normal flora in only a low percentage ofhealthy adults (57, 91, 113, 115, 122, 151, 159-161, 183). Theclustering of cases in hospitals and even within hospitalwards is the reason why such a wide range of incidence ofthe disease (0.01 to 10%) has been reported (90, 190, 227).PMC cases in hospitals can be especially difficult to controlbecause these patients have diarrhea. This mode of trans-mission increases the release of the organism into theenvironment, and the organism can be isolated from theclothing and room fixtures of the patient. In addition, theorganism is transmitted by hospital personnel. Once in theenvironment, the organism can persist for months since itproduces spores (115, 161). Efforts should be made to isolatepatients and minimize cross-contamination between pa-

tients. Compromised elderly patients undergoing antibiotic

therapy are especially at risk. The organism is present inhospitals not only from patients but also because it is presentas "normal" flora of the infants in many hospital nurseries.

Infants, for some unknown reason, are refractory to PMCeven though they carry high numbers of the organism andhigh levels of toxins A and B in their stools (5, 27, 30, 31, 50,58, 89, 98, 124, 136, 152, 195, 212, 216, 221, 233, 235-237). Infact, it has been estimated that 50% or higher of infants arecolonized with toxigenic C. difficile and are asymptomatic.The rate of colonization of infants parallels the incidence ofoutbreaks and cross-infections within hospitals, indicatingthat most infants acquire the organism nosocomially (30,176). That so many infants are colonized with toxigenic C.difficile and that so few develop the disease indicate that themanner in which they are protected is quite effective;otherwise, more infants would die from the infection. Noone knows why infants are protected, but there are a numberof hypotheses. The protection may well result from a com-bination of several factors. Colostrum contains substances(perhaps secretary antibody) which neutralize toxins A andB (114, 238), and these substances probably help to protectthe infant from the toxins, but it is more complicated thanthis since infants who are not breast-fed do not get thedisease. Fetal intestinal cells are reported to be much lesssensitive to the toxins than adult intestinal cells (46), and thismay contribute to the resistance.Another hypothesis is that infants may lack the toxin

receptors in their intestines. The receptor to which toxin Abinds (the Galal-3Gal,1-4GlcNAc trisaccharide; see sub-section, "Toxin Receptors") is similar to the I,i antigenswhich are developmentally regulated on erythrocytes. Thecarbohydrate chains on infant intestinal cells may exist in animmature form which is not recognized by toxin A. Thecarbohydrate sequences would subsequently develop intothe active receptor containing the Galal-3GalI1-4GlcNActrisaccharide in adults. Alternatively, the receptors on theinfant cells may be covered by a thicker layer of mucin thanin adults, and this could prevent the toxin from binding to itsreceptor. Determining the exact mechanism of infant resis-tance is one of the more exciting research areas, but weshould not assume that, because many asymptomatic infantshave the toxin present, infants are not affected. Infants whoare compromised by other disease or who have had intestinalsurgery may be at high risk (2).

Cystic fibrosis patients who are colonized with toxigenicC. difficile also are refractory to the development of PMC(169, 241, 249). It is not clear whether these patients areasymptomatic because they are colonized with nontoxigenicstrains or whether the manner in which the toxins areproduced and act in these patients is different from themanner in which the disease develops in other adults. It hasbeen suggested that the intestinal environment in thesepatients may be similar to the environment in the infantintestine and that the manner of protection is similar in thesetwo populations, but this has not been demonstrated.Although the role of C. difficile in PMC has been clearly

established, its role in other less severe gastrointestinalillnesses is not as well defined. C. difficile and its toxins arefound in patients with symptoms ranging from simple diar-rhea all the way to PMC, but the role of the organism inantibiotic-associated diarrhea still is not understood. Evenwhen the toxins are present, they may not be causing thesymptoms, but currently patients with antibiotic-associateddiarrhea who test positive for C. difficile toxins usually areconsidered to have C. difficile disease. It has been estimatedthat the organism may cause approximately 25% of the

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

4 LYERLY ET AL.

reported cases of antibiotic-associated diarrhea and that thenumber of cases of C. difficile diarrhea is second only to thenumber caused by Campylobacterjejuni, which is believedto be the most frequent cause of bacterial diarrhea (38, 85).More clinical information is needed, however, to get a betterunderstanding of the importance of the organism in diarrheicpatients.

In addition to its involvement in gastrointestinal illnesses,C. difficile can cause disease in other parts of the body,including abscesses, wound infections, osteomyelitis, pleu-ritis, peritonitis, septicemia, and urogenital tract infections,but these are rare (92, 133, 207). There was an initial reportthat C. difficile was a common isolate from urogenitalinfections, but this has been refuted. The organism has beenisolated from neonates with necrotizing enterocolitis andfrom patients with inflammatory bowel disease and Crohn'sdisease, but it appears that the organism does not play an

active role in these diseases except as a complication (29, 43,59, 87, 94, 109, 123, 155, 215, 234).A number of methods have been used to treat patients

with PMC. In many instances, just the discontinuation of theoffending antibiotic is sufficient to resolve the symptoms.The most common method of treating the disease, ironically,is by the use of antibiotics. Vancomycin was one of theinitial antibiotics used for the treatment of PMC, based on

the findings that: (i) vancomycin was effective against gram-

positive bacteria and had been used for "staphylococcalcolitis" (107); (ii) clinical isolates of C. difficile were suscep-

tible in vitro to the antibiotic (42, 61, 67); (iii) vancomycinwas effective against C. difficile disease in hamsters (21, 40);and (iv) very high concentrations could be achieved in theintestine because it was not absorbed. PMC patients treatedwith vancomycin showed prompt improvement, and this isthe most common treatment used in the United States.Vancomycin is expensive, and the high doses used in theearly trials have now been replaced by lower doses that are

equally as effective. Even the lower doses are more expen-sive than some other antimicrobial agents that have beenused extensively for treatment in European countries. Met-ronidazole is widely used in Europe for treatment of anaer-

obic infections and is an effective method of treatment forPMC (28, 225). Its advantage over vancomycin is its lowercost. Bacitracin also has been used successfully (60, 252).Any antibiotic that kills C. difficile should be effective ifconcentrations above the inhibitory concentration can beachieved in the colon. These same antibiotics can cause thedisease when the concentrations in the colon are not abovethe inhibitory level, and they can cause the disease whentherapy is stopped and the concentrations decline. Thus,vancomycin and metronidazole can cause the disease as wellas cure it. Many patients treated for PMC or minor diarrhearelapse several days after they are taken off the antibioticused for therapy. Relapse rates of 20% are reported forvancomycin and some patients "relapse" repeatedly (15, 16,24).Treatment with anion-exchange resins such as cholestyra-

mine and cholestipol has been used with some success. Theresin binds the toxins produced by the organism and mini-mizes the tissue damage which occurs during the disease.The alleviation of the symptoms, however, is more variablecompared with the use of antibiotics, and some patients donot respond at all to this type of treatment (117, 228).Treatment of an "antibiotic-associated" disease with an-

tibiotics obviously is not an ideal solution. Another form oftreatment currently being examined is the reestablishment ofthe gut flora, using normal fecal flora. This approach has

been successful in treating hamsters with antibiotic-associ-ated enterocolitis and has recently been shown to be effec-tive in treating PMC in humans (33, 34, 37, 210, 243, 246).Another approach is to give patients nontoxigenic strains ofC. difficile so that the toxigenic strains cannot get establishedin the bowel. This has also been used successfully inhamsters and is now being tried in human patients. The yeastSaccharomyces boulardii has been used successfully toprotect hamsters from clindamycin-associated enterocolitisand relapse following vancomycin treatment (64, 154, 232),but clinical trials in PMC patients have not been reported.

Certain prostaglandins protect the stomach and smallintestine from mucosal necrosis caused by harsh agents suchas alkaline and acidic solutions and alcohol. The manner bywhich this protection occurs is not clearly understood, butthese findings led investigators to test whether these agentswould protect hamsters against the extensive intestinalmucosal damage which occurs in clindamycin-associatedcecitis. The results show that prostaglandins (PG), 16,16-di-methyl-PGE2, 15(R)-15-methyl-PGE2, and 2-acetyl-2-decar-boxy-15(S)-15-methyl-PGF2a, as well as castor oil and veg-etable oil (which contain the prostaglandin precursor linoleicacid), prevent the disease in hamsters (201, 213). It has beensuggested that the prostaglandins may affect the release ofthe toxins by the organism (205), but the protection also mayresult from the effect of the prostaglandins on the tissue.These findings are exciting since the protection observedwith prostaglandins is the most dramatic protection yetobserved with pharmacologic agents.

CLINICAL DIAGNOSIS

Pathology

The most definitive diagnosis ofPMC is by the endoscopicdetection of pseudomembranes or microabscesses in antibi-otic-treated patients with diarrhea who have C. difficile toxinin their stools. The diarrhea may be watery or bloody andmay be accompanied by abdominal cramps, leucocytosis,and fever (70, 81, 228). Often the microabscesses andplaques are isolated in only certain parts of the colonic wall,with other areas remaining normal in appearance. Whycertain areas are affected while other adjacent areas are notis unknown. The diagnosis of C. difficile disease in patientswho develop antibiotic-associated diarrhea is considerablymore difficult. In these cases, toxigenic C. difficile actuallymay become established in the patient as a result of thediarrhea and play no role in the disease (see subsection,"Problems Associated with Clinical Diagnosis of C. difficileDisease").

Isolation and Identification of the Organism

The isolation of toxigenic C. difficile from the stool hasbeen used for the presumptive diagnosis of PMC becausepatients with the disease contain high numbers (107 orgreater) of C. difficile in their stools (23). The most widelyused medium for the isolation of C. difficile is the cyclose-rine-cefoxitin-fructose-egg yolk medium developed byGeorge et al. (83). The medium serves as a selective anddifferential medium for C. difficile and reportedly can detectas few as 2,000 organisms in a total count of 6 x 1010 bacteriaper g (wet weight) of feces.

Several modifications of the medium which affect theisolation rate of the organism have been described. Forexample, the incorporation of sodium taurocholate in place


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


TABLE 1. Evaluation of clinical tests for detection of C. difficile and its toxinsTest Antigen detected Sensitivity Evaluation

Tissue culture assay Toxin B 1 pg (50 pg/ml) The test is based on the detection of cytotoxic activity in stool specimensas noted by rounding of the tissue culture cells and neutralization of theactivity by C. djficile of C. sordelifi antitoxin. The assay shows a goodcorrelation with the disease and serves as the "gold standard" by whichto measure other tests for the disease. The test is extremely sensitive,with its primary disadvantages being the assay time and expense.

Counterimmunoelec- Not specific for 10 ng of antigen The test is not recommended as a clinical assay for C. difficile or its tox-trophoresis Tox+ isolates (0.5 pug/ml) ins. The C. difficile antiserum which has been used in the assay in clini-

cal trials cross-reacts with C. sordellii and C. bifermentans and givesfalse-positive reactions.

ELISA Toxin A or B 0.1 to 1 ng The ELISA is still in the research and development stage. The toxin Adepending on (1-10 ng/ml) ELISA is close to the sensitivity needed for the test but it needs to bespecific anti- shortened to <1 h to make it more suitable as a clinical test.body

Latex agglutination Not specific for Not reported The Culturette Brand Rapid Latex Test for C. difficile has created muchtest Tox+ isolates excitement because it is rapid (ca. 3 min) and simple. The assay is not

specific for Tox+ isolates; it reacts with Tox- isolates as well as C.sporogenes, P. anaerobius, and B. asaccharolyticus.

of the egg yolk apparently improves the recovery rate of theorganism (41, 165, 244). In addition, the concentrations ofcycloserine and cefoxitin and the type of basal medium affectthe recovery rate and the phenotypic properties (e.g., theintensity of the fluorescence) exhibited by the organism (130,131).Some of the unusual metabolic products (e.g., p-cresol

and isocaproic acid) produced by C. difficile have beensuggested as markers for the organism (54, 84, 129, 134, 150,171, 175, 182). The detection of these substances is notspecific enough to be used as a marker for clinical diagnosis.

Several microsystems, the API ZYM Micro-System, (APILaboratory Products, Ltd.), the Minitek Anaerobe II (BBLMicrobiology Systems), the API An-Ident System (AnalytabProducts), and the RapID ANA System (Innovative Diag-nostic Systems) are now available for the identification ofanaerobic bacteria. The organism must be isolated initially,however, before it can be identified by these test systems.These systems have only recently become available, andtheir specificity for C. difficile has not been critically evalu-ated (4, 26, 121, 132, 153).

Detection of ToxinA description of the various clinical assays described for

the toxins of C. difficile is given in Table 1. The tissue cultureassay has been used extensively for the detection of C.difficile toxin in stool specimens. This is based on the findingthat >90% of PMC patients have cytotoxic activity in theirstools (14, 19, 23, 66). The cytotoxic titer does not closelycorrelate with the severity of the disease, although there issome indication that the titer is higher in severe cases ofPMC. In the test, dilutions of the stool specimen are addedto tissue culture cells and a positive reaction is noted byrounding of the cells. Many stool specimens contain othertoxins or nonspecific interfering substances which causerounding of the cells; therefore, the specificity must beconfirmed by neutralizing the cytotoxic activity with antise-rum against C. difficile or C. sordelifi (62).Both toxins A and B have cytotoxic activity and are

present in stool specimens from PMC patients. Toxin B,however, is a much more potent cytotoxin than toxin A andmasks the weak cytotoxic activity of toxin A unless the

toxins are first separated (146). In addition, the toxins arecoproduced in complex media and no toxA+/toxB isolateshave been identified. Therefore, the cytotoxic activity de-tected in the stool specimens ofPMC patients is due to toxinB.The major advantage of the tissue culture assay for the

detection of toxigenic C. difficile is its extreme sensitivity; 1pg of toxin B is sufficient to cause rounding of the cells. Thisis the reason the test correlates well with the disease andwhy the test represents the "gold standard" for C. difficiletoxin tests. There are several possible problems, however,with the assay as it is currently used. Toxin B, as well astoxin A, is active against a variety of mammalian cells; infact, no cell lines have been described which are not sensi-tive to the toxins, although some cell lines are less sensitivethan others. As a result, clinical laboratories use differentcell lines, making- it difficult to compare clinical studies. Inaddition, the assay procedure has not been standardized.There are variations in the dilutions of stool specimenstested, the ahtiserum used, and the time of recording theresults, and this inhibits the direct comparison of clinicaldata. Another problem is that tissue culture assays requirespecialized equipment and personnel, making the assayexpensive. Recently, the Tox-Titer microtiter plate system(Bartels Immunodiagnostic Supplies, Inc.), which uses hu-man foreskin cells, became available for the detection of C.difficile toxin (11, 162, 248). The system is easy to use andallows persons not equipped for standard tissue culture toutilize the assay, but it also is expensive.A number of antibody-based tests have been described.

Counterimmunoelectrophoresis was the first test of this typeto be used for the detection of C. difficile, primarily becauseit is considerably faster than tissue culture assays (1.5 hcompared to overnight). However, the antiserum used inclinical trials was prepared against culture filtrates of C.difficile and the immunoprecipitin bands detected in moststool specimens do not represent the toxins. The amount oftoxin necessary to give a band is in the range of 10 to 100 ng,indicating that concentrations of micrograms per gram stillare needed. Most stool specimens from PMC patients do nothave this concentration of toxin. Nontoxigenic strains of C.difficile, as well as strains of C. sordellii and Clostridium

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

6 LYERLY ET AL.

bifermentans, react with the antiserum and give false-posi-tive reactions in the assay (102, 103, 135, 184, 208, 242, 247).The cross-reactions are most likely due to the cell surfaceantigens shared by these species (185-187). Some prelimi-nary work has been done with monospecific toxin antibodyin counterimmunoelectrophoresis tests, but the antibodydoes not provide the sensitivity needed for a clinical test(144).

Several variations of microtiter plate enzyme-linked im-munosorbent assays (ELISA) have been developed for thedetection of toxins A and B (10, 118, 128, 144, 147, 240).These tests are based on the detection of either toxin A or Bwith specific toxin antibody. At the present time, these testsare still in the research and development stage. A clinicalELISA which detects either toxin A or toxin B or bothwould be suitable since both toxins are present in the stoolsof PMC patients. The toxin A ELISA, however, looks morepromising for several reasons. Toxin A is a better "marker"antigen for toxigenic C. difficile since it is more stable thantoxin B. In addition, it is easier to purify, making it easier toobtain toxin A antibodies.

Stool specimens from patients with PMC almost alwayshave cytotoxic titers (toxin B) of 102 or greater. Specimenswith cytotoxic titers of 102 have a toxin A concentration ofabout 1 nglml, so the ELISA needs to be capable of detectingthis concentration of toxin A. The toxin A ELISA testswhich have been developed approach this level of sensitiv-ity, but they require at least 4 to 5 h to complete. Theincubation times can be shortened and direct instead ofindirect procedures can be used, but these modificationsdecrease the sensitivity of the test.

Latex agglutination tests have proved to be very popularbecause they are rapid and simple. For these reasons, manyclinicians have been interested in the Culturette Brand RapidLatex Test for C. difficile (Marion Scientific, Inc.). The testwas initially marketed for the detection of toxin A in stoolspecimens and consisted of a latex reagent supposedlycoated with the immunoglobulin G fraction of monospecificantiserum against toxin A. The development of the test wasbased on the characterization studies done by Banno et al.(12, 13) on toxin A. Our initial studies, however, showed thatour highly purified toxin A preparations did not react in thetest (148). The reasons for these findings were puzzling sincethe properties of toxin A reported by Banno et al. werealmost identical to the properties we observed with toxin A.We now suspect that the toxin A used to prepare theantiserum for the commercial test contained low amounts ofa contaminating protein (the "latex-reactive" antigen) whichis highly antigenic. When the toxin A preparation wasinjected into animals, antibody against the contaminatingantigen was produced. Latex reagent coated with this anti-serum was most likely used to screen antigen fractions whenadditional antiserum was prepared, resulting in the selectionof fractions containing the contaminating antigen. Each timethe cycle was repeated they would have had less toxin andmore of the contaminant.Nontoxigenic strains of C. difficile produce high amounts

of the latex-reactive antigen and react strongly in the test.The latex-reactive antigen is quite distinct from toxin A. It issmaller in size and binds more tightly to anion-exchangegels, but, most important, it is nontoxic and shows noimmunological relatedness to toxin A.

In addition to detecting toxigenic and nontoxigenic iso-lates of C. difficile, the Culturette Brand Rapid Latex Testreacts with strains of proteolytic Clostridium botulinum andstrains of Clostridium sporogenes (35, 108), Peptostrep-

tococcus anaerobius (S. Allen, submitted for publication),and Bacteroides asaccharolyticus (35). We have found thatC. sporogenes and P. anaerobius produce an antigen whichcross-reacts with the latex-reactive antigen of C. difficile.The antigen from each of these species contains a subunitwith an Mr of 43,000, indicating that the antigens are verysimilar. Analysis by immunodiffusion, however, shows thateach antigen contains some unique epitopes (D. Lyerly, D.Ball, J. Toth, and T. Wilkins, submitted for publication). Thefunction of the antigen is not known, but there is no evidencelinking it with the virulence of the organism.

Problems Associated with Clinical Diagnosis of the Disease

The question of how to diagnose PMC seems, at firstglance, to be relatively straightforward: it would seem that apositive diagnosis involves the isolation of toxigenic C.difficile or the detection of toxin in the stool of a patienttreated with antibiotics. There are a number of factors,however, which complicate the clinical picture. The numberof cases of PMC reported in the past several years hasdecreased considerably. This is because in most cases, oncethe patient develops antibiotic-associated diarrhea and C.difficile toxin is detected in the stool, the patient is switchedto vancomycin or some other type of therapy; the patient isnot allowed to progress past this stage of the disease. Thatthe patient responds to vancomycin treatment does notconfirm that C. difficile caused the disease. Cases of antibi-otic-associated diarrhea from which this organism or toxinscannot be demonstrated often improve following treatmentwith vancomycin. In addition, vancomycin has a broadspectrum of activity, and at the high concentrations achievedin the bowel it inhibits most pathogens. Therefore, resolu-tion of the disease by vancomycin does not prove that thedisease was caused by C. difficile. The development of thediarrhea in the first place indicates that the normal flora ofthe intestine has been altered; as a result, the patient couldbe colonized with any number of pathogens, including C.difficile. C. difficile may happen to colonize the patient oncethe flora has been changed by the antibiotic and may not beresponsible for the diarrhea. As an example, E. coli ispresent in stools from patients with diarrhea, but this in noway confirms that toxigenic E. coli caused the diarrhea. Itshould be remembered that toxin B of C. difficile is ex-

tremely potent and low numbers of toxigenic C. difficile inthe gut which do not pose a problem may produce enough ofthe toxin to be detected in tissue culture assay.Even the isolation of toxigenic isolates of C. difficile from

patients with antibiotic-associated diarrhea is not a definitivediagnosis. Toxigenic isolates vary in their toxin productionover a range of at least 6 logs. This means that a "highly"toxigenic strain may produce 106 times more toxin than a

"weakly" toxigenic strain. This unusual range of expressionof toxin is not seen in other bacteria. The manner in which C.difficile regulates this wide variation in toxin production isnot known. That C. difficile has this unusual regulation oftoxin production raises questions which have to be ad-dressed. For example, do the highly toxigenic isolates cause

PMC whereas weakly toxigenic isolates cause diarrhea or

are weakly toxigenic isolates not important in the disease?Another area which has caused considerable confusion

includes the relative sensitivities and specificities of thevarious toxin assays. It has been reported that stool speci-mens which are cytotoxic but not enterotoxic containtoxA/toxB+ isolates. However, it should be remembered thatthe tissue culture assay for toxin B is extremely sensitive


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


(10-12 g/ml of toxin detected) and that enterotoxic assays(10-6 g/ml of toxin detected) are 106-fold less sensitive. Inaddition, there are descriptions of toxA'ItoxB isolates byinvestigators who used the Culturette Brand Rapid LatexTest. We now know that the latex test does not detect toxinA and that these isolates were nontoxigenic isolates.

Questions then arise as to how to establish C. difficile asthe causative agent of the diarrhea and which toxin testsshould be used. Unfortunately, there are no straightforwardanswers. The identification of toxigenic C. difficile as thecause of the disease has to be established based on thephysician's experience. The tissue culture assay for thedetection of cytotoxic activity that is neutralized by C.difficile or C. sordelifi antitoxin still represents the primarytoxin test and should continue to be used until more clinicaldata are available regarding the other tests. The clinicalresults with the Culturette Brand Rapid Latex Test for C.difficile look promising (110, 173, 174, 206). The test issimple and the results can be obtained within several min-utes. It should be cautioned, however, that nontoxigenicisolates of C. difficile produce spores and can exist in theenvironment as well as toxigenic strains. Thus, nontoxigenicisolates can be spread nosocomially, and as a result "out-breaks" of nontoxigenic strains can occur which may resultin local high rates of false-positive reactions. At the presenttime, the latex assay may be suitable as an initial screeningtest for the organism and positive results can be confirmedwhen warranted for clinical reasons.

TOXINS A AND B

Role of Toxins A and B in PMC

The extensive tissue damage and fluid response whichtoxin A causes in experimental animals have led manyinvestigators to speculate that toxin A, and not toxin B, isthe toxin primarily responsible for the symptoms associatedwith PMC. The severity of the symptoms in antibiotic-associated cecitis appears to be more closely correlated withthe in vivo production of toxin A than with toxin B (36), andthese findings also support the idea that toxin A causes mostof the symptoms.There are several observations, however, indicating that

toxin B is also important in the disease. Hamsters must bevaccinated against both toxin A and toxin B to be completelyprotected against the disease; hamsters vaccinated againsttoxin A are not completely protected (69, 137; P.-H. Kimand R. D. Rolfe, Abstr. Annu. Meet. Am. Soc. Microbiol.1987, B222, p. 62). Like toxin A, toxin B is also highly lethalwhen injected into animals. In fact, the lethal dose of toxin Bis approximately the same as that of toxin A when the toxinsare injected intraperitoneally or intravenously (9, 145, 220).This indicates that toxin B (and toxin A for that matter) canact on target tissue outside of the intestines. We have foundthat if low doses of toxin A (i.e., doses too low to give adiscernible response) are given intragastrically along withtoxin B to hamsters, the animals die. The only pathologyconsists of some mucus and edema in the small intestine andhemorrhage in the lungs (146). Thus, toxin A apparentlyfacilitates the exit of toxin B from the gut. Both of the toxinsare present in stools of PMC patients, so the synergisticaction seen in the oral hamster model could also occur in thehuman disease. These findings also raise the possibility thatasymptomatic infants who have high levels of the toxins intheir intestines may be at risk if they are exposed toconditions (e.g., intestinal surgery) which result in the dis-

semination of the toxins into the systemic circulation. On thebasis of these observations, we feel that both of the toxinsare involved in the disease.

Factors Affecting Toxin Production

The incorporation of antibiotics such as clindamycin intothe growth medium has been reported to increase the levelsof cytotoxic activity in cultures of C. difficile by four- toeightfold (78, 163). This increase in toxin production doesnot occur with all isolates of toxigenic C. difficile, and thismay explain why some investigators have not observed thisstimulation in toxin production in cultures grown with clin-damycin (167).The medium used for the growth of the organism also

affects the amounts of toxins produced. The organism growswell in a complex medium such as brain heart infusion brothand produces high levels of both toxin A and toxin B (147,220). Under these conditions, the production of the toxinsappears to be coregulated. Highly virulent isolates producehigh levels of both toxins and weakly virulent isolatesproduce low levels of the toxins under these conditions.Avirulent strains do not produce either of the toxins. Under"starved" conditions, however, the coproduction of thetoxins does not appear to be as tightly regulated. Forexample, it has been reported that some isolates produceonly one of the toxins when grown in media deficient incertain amino acids (96). The level at which this inhibitionoccurs in the cell is not known. It is unlikely that this type ofinhibition of toxin A or B production occurs in the intestineofPMC patients since the organism is in an environment richin nutrients and relatively free of competing organisms.Some clostridial toxins such as the C2 enterotoxin of C.

botulinum and the enterotoxin of Clostridium perfringenshave been shown to be associated with spore production (77,251). Neither of the C. difficile toxins, however, has beenshown to be associated with sporulation (111; L. D. Bobo,A. L. Delisle, B. E. Laughon, and J. G. Bartlett, Abstr.Annu. Meet. Am. Soc. Microbiol. 1986, B67, p. 35). This isbased on the findings that nonsporulating mutants producelevels of toxins A and B comparable to the sporulatingparental strain and that neither toxin can be extracted fromspores.

Like almost all other bacterial species, C. difficile hasplasmids of various sizes (97, 158, 223). In C. difficile,plasmids with molecular weights ranging from 2.7 x 106 to100 x 10' have been detected. The presence or absence ofany of these plasmids has not been correlated with toxige-nicity. In addition, bacteriophage specific for C. difficile andC. sordellii have been isolated and characterized (149, 209,211), but, again, there is no evidence that the toxin genes arelocated on phage.

PurificationToxins A and B are present in the supernatant fluids of

cultures of C. difficile and can be purified from the culturefiltrates. The manner by which the toxins are liberated by thecells is not clear. We have found that the concentration ofthe toxins in the supernatant fluid parallels the growth of theorganism and that most of the toxin is released in latelogarithmic or early stationary phase, suggesting that thetoxins may be actively secreted by the cells. Ketley et al.(111), on the other hand, have reported that toxins A and Bare released after the organism reaches stationary phase andhave suggested that the toxins are released after the cellslyse.

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

8 LYERLY ET AL. CLIN. MICROBIOL. REV.

TABLE 2. Purification and properties of toxins A and B of C. difficile

Reference Purification method' Toxin Mr Mr subunit pH stability Heat stability Inactivation by PIproteasesBanno et al. (NH4)2SO4 ppt, Sephacryl S- A 550,000- 190,000- b Stable at 370C; Trypsin, pronase

(12, 13) 300, DEAE-Sephadex A-25 600,000 200,000 inactivatedat 600C

B 450,000- 190,000- Stable at 370C; Trypsin, pronase500,000 200,000 inactivated

at 600C

Krivan and Thermal affinity chromatog- A 440,000-Wilkins raphy 500,000(120)

Lonnroth (NH4)2SO4 ppt, Sephacryl S- Aand Lange 300, DEAE-Sephadex A-(139) 25, affinity chromatogra-

phy on agarose

Pothoulakis (NH4)2SO4 ppt, DEAE-Se- B 50,000 - 4.3et al. pharose CL-6B, HPLC on(180) Mono Q

Rihn et al. Ultrafiltration, fast liquid an- A 52,000 41,500 3.5(200) ion-exchange chromatogra- and

phy, chromatofocusing, 16,000TSK SW-3000

Stephen et Electrophoresis, DEAE- Aal. (217) Sepharose CL-6B

Sullivan et Ultrafiltration or (NH4)2SO4 A 440,000- 250,000- Stable at pH Stable at 37°C; Trypsin and chy- 5.2-5.7al. (220); ppt, DEAE-Sepharose CL- 500,000 300,000 4 and 10 inactivated motrypsinLyerly et 6B, ppt at pH 5.6 (toxin at 560Cal. (143, A), immunoadsorption B 360,000- 250,000- Inactivated Stable at 37°C; Trypsin and chy- 4.1-4.5145) (toxin B) 470,000 300,000 at pH 4 inactivated motrypsin

and 10 at 56°C

Taylor et al. Ultrafiltration, (NH4)2SO4 A Stable at pH Inactivated at(224) ppt, Sephadex G-200, 4 and 10 560C

DEAE-Sepharose CL-6B B Inactivated Inactivated atat pH 4 560Cand 10

DEAE, Diethylaminoethyl; ppt, precipitation; HPLC, high-pressure liquid chromatography.b-, Not reported.

The most commonly used medium for the large-scaleproduction of the toxins is brain heart infusion broth. Theorganism grows well in the broth, and many investigatorspurify the toxins from brain heart infusion broth cultures. Amodification of this approach is the brain heart infusiondialysis, sac flask in which the organism grows within a sacsuspended in the broth. In the dialysis flask procedure, thelow-molecular-weight nutrients in the brain heart infusiondiffuse into the sac, resulting in the slow growth of theorganism. The slow growth of the organism, in turn, resultsin the production of more toxin. In this procedure, the toxinsare not contaminated with the high-molecular-weight com-ponents from the brain heart infusion. This modification wasoriginally utilized by Sterne and Wentzel (218) for theproduction of botulinum toxin and works quite well for theproduction of C. difficile toxins.A number of purification methods have been reported for

preparing toxins A and B (Table 2). Most of the purificationprocedures which have been described utilize an initialconcentration step such as ammonium sulfate precipitationor uitrafiltration, followed by gel filtration and ion-exchangechromatography. Ion-exchange chromatography has beenused extensively because it separates toxins A and B very

effectively. The simplest and most gentle procedure devel-oped thus far for the purification of toxin A is the thermalaffinity method (120). The method is based on the specificbinding of toxin A to the carbohydrate sequence Gala1-3Gal11-4GicNAc on bovine thyroglobulin at 4°C and therelease of the toxin at 37°C (see below). Thus, complexmixtures containing toxin A can be mixed with thyroglobulingel at 4°C to bind the toxin. The gel is then washed and thetoxin is eluted at 37°C. The eluted toxin has biological andphysicochemical properties similar to those of toxin Apurified by conventional methods.

In many of the purification studies, the purity of the toxinwas either not demonstrated or analyzed at most by only onecriterion. We routinely use crossed immunoelectrophoresisin conjunction with polyacrylamide gel electrophoresis toanalyze our toxin preparations because each of these meth-ods is highly resolving and picks up contaminants notdetected by the other method. In addition, we have shownthat the biological activity present in the preparation (e.g.,the enterotoxic or cytotoxic activity) comigrates with theimmunoprecipitin arc or band. This has proved to be espe-cially important in the purification of toxin B since the toxinis extremely potent and amounts which have high levels of

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


TABLE 3. Biological activities of toxins A and B of C. difficile

Tissue culture Enterotoxic dose in Enterotoxic dose in Enterotoxic dose in VascularReference Toxin dose rabbit ileal loop oral hamster model suckling mice Lethal dose in micea permeability(Ig) (pLg) (Og) dose

Banno et al. (12) A Low activity 3 b 26 ng (LD50) 1 ngB 1 pg Negative 1.5 [ig (LD5o) 6 ng

Pothoulakis et al. B 5.1 fg(180)

Sullivan et al. (220) A 10 ng 1 10 0.25 50 ng (minimum dose) 1-10 ngLyerly et al. (142, B 0.2-1 pg Negative Negative >2.5 50 ng (minimum dose) 1-10 ng

145), Lima et al.(submitted forpublication)

Taylor et al. (224) A 40 ng 100- Negative 10 ng (LD50) 2 jigB 1 pg Negative Negative 4.4 pug (LD50) 1 jig

LD50, 50% lethal dose.b-, Not reported.' Minimum dose not reported.

cytotoxic activity may not be sufficient to give a band ongels.

Physicochemical Properties

Toxin A purified by Banno et al. (12, 13), Taylor et al.(224), and our group (120, 143, 145, 220) exhibits similarphysicochemical properties (Table 2). One of the mostunusual properties of the toxin is its extremely large size.The native toxin has an Mr in the range of 400,000 to 600,000.Even under denaturing conditions, the toxin has an Mr inexcess of 250,000 and does not dissociate to subunits (12,143, 191). These findings, along with the observation that thetoxin A gene is a continuous open reading frame of morethan 6.1 kilobases (see below), support the hypothesis thatthe toxin consists of a large polypeptide with an Mr of>250,000. It has been reported that toxin A binds to agarose(139), has an estimated Mr of 52,000 (200), and containsalmost 30% glycine (217). These properties have not beenconfirmed and the small size which was reported does notagree with the estimated toxin size determined from thetoxin A deoxyribonucleic acid sequence (see below).Toxin B purified by Banno et al. (12, 13), Taylor et al.

(224), and our group (143, 145, 220) also exhibits similarproperties (Table 2). Like toxin A, toxin B is extremely largeand has a native Mr in the range of 360,000 to 500,000. Underdenaturing conditions, the toxin behaves similarly to toxin Awith an Mr of at least 250,000 and does not dissociate (12,143). Thus, it appears that toxin B may also consist of asingle large polypeptide. It has been reported that toxin Bconsists of subunits which have an estimated Mr of 50,000(180), but these findings have not been confirmed and do notagree with findings by other investigators (12, 143).Toxin B is less stable than toxin A and is more susceptible

to pH extremes and proteases (12, 220, 224). Both of thetoxins are inactivated by oxidizing agents and can be pro-tected by the addition of a reducing agent such as dithio-threitol (145). The oxidizing agents affect a number of aminoacids, so it is difficult to determine the specific amino acidswhich are modified. Reducing agents such as 2-mercaptoeth-anol and dithiothreitol and sulfhydryl-inactivating agentssuch as p-hydroxymercuribenzoate, N-ethylmaleimide, andiodoacetate do not inactivate the toxins, indicating that thetoxins are not sensitive to reduction and that sulfhydryl

groups are not involved in the binding or toxic moieties ofeither toxin. Both toxins contain high amounts of the aminoacids Asp, Glu, and Gly, and low amounts of His andsulfur-containing amino acids (12, 143, 145).

Biologic Activities

Some of the biologic activities of toxins A and B are listedin Table 3. Toxin A is a potent enterotoxin with slightcytotoxic activity, whereas toxin B is an extremely potentcytotoxin. On a molar basis, toxin A is as active as choleratoxin in the rabbit ileal loop assay for enterotoxic activity.Its action, however, is quite different from cholera toxin.Cholera toxin stimulates adenylate cyclase, resulting in anincrease in intracellular cyclic adenosine 5'-monophosphate.This change, in turn, eventually leads to the rice-water fluidaccumulation which is seen in rabbit ileal loops injected withthe toxin. There is little or no tissue damage observed withcholera toxin. Toxin A, on the other hand, causes anextensive amount of tissue damage to the gut mucosa (142,156; A. Lima, D. Innes, Jr., K. Chadee, D. Lyerly, T.Wilkins, and R. Guerrant, Abstr. Annu. Meet. Am. Soc.Microbiol. 1987, B219, p. 61). Initially, the toxin damagesthe villous tips, and this is followed by the disruption of thebrush border membrane. Eventually the mucosa becomes"eroded." This tissue damage apparently results in the fluidresponse. The hemorrhagic fluid which develops in ilealloops treated with toxin A contains high levels of albumin,indicating that the fluid results from vessel leakage andedema fluid in the area surrounding the damaged gut mu-cosa. Toxin A also causes a fluid response when injected intorabbit colonic loops, but the fluid is considerably less hem-orrhagic than that seen in ileal loops (156).Both of the toxins cause the same type of rounding effect

on tissue culture cells, and both are active against all of themammalian cells which have been tested. This suggests thepossibility that the toxins may act by the same or a similarmechanism on cells. Most of the research on the cytotoxicactivity of the toxins has centered on toxin B because it ismuch more potent. Toxin B causes a number of nonspecificresponses in mammalian cells, including the loss of intracel-lular potassium, decrease in protein synthesis, and decreasein synthesis of ribonucleic and deoxyribonucleic acids (72,181, 199, 204). These probably occur as secondary responses

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

10 LYERLY ET AL.

following the intoxication process. There appears to be aneffect on the microfilament system of the cell since there isdisorganization of the actin filaments and an increase in theintracellular amount of globular actin (180, 229, 230). Thiseffect most likely is responsible for the "rounding" whichoccurs in toxin-treated cells. This rounding effect also isobserved in cells treated with cytochalasin B, a fungalmetabolite which acts on microfilaments. The action ofcytochalasin B, however, is reversible, whereas the round-ing effect caused by toxin B is not (229). Toxin B also causespolarization of the nucleus in cells, and it has been suggestedthat this effect is indirectly due to the disruption of themicrofilament system (3). It has been reported that in tissueculture cells the mechanism of action of toxin B is verysimilar to that of diphtheria toxin (73-75, 230). Additionalwork is needed, however, to confirm these findings and tofurther compare the similarities of toxin B and diphtheriatoxin.Both toxin A and toxin B are lethal toxins. Mice, ham-

sters, rats, and rhesus monkeys injected intraperitoneally orsubcutaneously with small amounts of either toxin die (9, 12,13, 145, 220, 224). There are no obvious signs in animalsinjected with the toxins, making it difficult to determine themanner in which the toxins act. Toxin A is also lethal whengiven intragastrically to hamsters (146). In this case, theanimals develop intestinal pathology and diarrhea. Toxin Aexerts some "chronic" type of effect on the intestine as well.This is based on the observation that hamsters given smallamounts of toxin A at weekly intervals develop diarrhea anddie. This effect may contribute to the severity of relapsesoften seen in PMC patients. Highly buffered solutions oftoxin B do not cause any response when given intragastri-cally to hamsters unless the intestine is damaged in somemanner, either by toxin A or by manipulation of the gut. Iftoxin B is given intragastrically along with low amounts oftoxin A, there is no significant damage to the intestines andthe animals die with no apparent symptoms. This is similarto what is seen in animals injected with toxin B intraperito-neally, suggesting that the toxin exits through the damagedgut mucosa and acts distally to the intestine. These resultsraise the possibility that any manipulation of the intestine(e.g., intestinal surgery) in patients colonized with toxigenicC. difficile may facilitate the exit of toxin B from the gut,possibly resulting in the death of the patient. This may beespecially of concern in infants colonized with toxigenic C.difficile who undergo some type of intestinal trauma.

Immunochemical Properties

Monoclonal antibodies against the toxins have recentlybeen developed by several laboratories and are being used tostudy the toxins (143, R. Rolfe, J. Szkudlarek, K. Follis, andC. Faust, Abstr. Annu. Meet. Am. Soc. Microbiol. 1986,B62, p. 34; S. Rothman, J. Brown, D. Foret, J. Gemski, M.Gentry, and P. Stricker, Abstr. Annu. Meet. Am. Soc.Microbiol. 1987, B214, p. 60). One of the monoclonal anti-bodies we have characterized (PCG-4 antibody) exhibits theunusual property of precipitating toxin A. The toxin containsrepeating amino acid sequences (see below), and it is likelythat this monoclonal antibody binds to these repeats,thereby explaining why the antibody precipitates toxin A.Another unusual property of the PCG-4 antibody is that itneutralizes the enterotoxic activity of toxin A but not thecytotoxic activity of the toxin. The antibody blocks thebinding of the toxin to the Galod-3Gal,1-4GlcNAc receptor,indicating that the antibody recognizes the binding and not

the active site on toxin A. The finding that the antibodyneutralizes the enterotoxic, but not the cytotoxic, activityindicates that the enterotoxic and cytotoxic activities oftoxin A are distinct. If this proves to be the case, it may bepossible to chemically or enzymatically cleave the toxin intodistinct enterotoxic and cytotoxic fragments. It is possiblethat the cytotoxic activity is on another subunit; if so, thesubunit has not been detected on polyacrylamide gels.Further sequencing of the toxin A gene should answer thisquestion.

In addition to monoclonal antibodies which react specifi-cally with toxin A, monoclonal antibodies which cross-reactwith toxins A and B have been described (143; Rolfe et al.,Abstr. Annu. Meet. Am. Soc. Microbiol. 1986; Rothman etal., Abstr. Annu. Meet. Am. Soc. Microbiol. 1987). Theseantibodies do not neutralize the biological activities of eithertoxin. The findings are surprising in light of the results fromprevious studies showing that polyclonal antibodies andantisera against the toxins do not show significant levels ofcross-reaction or cross-neutralization with the heterologoustoxin (13, 138, 224). Apparently, the polyclonal antibodiesused in these studies did not contain sufficient quantities ofthe cross-reacting antibodies.

Toxin Receptors

Toxin A has been shown to bind to brush border mem-branes from hamsters, indicating that these membranescontain the receptor needed for expression of the entero-toxic activity (119). The role of this binding component asthe physiological receptor is supported by the finding thattoxin A does not bind readily to brush border membranesfrom mice and rats and that these animals are much lesssensitive to the enterotoxic activity of toxin A. The bindingto the membranes is temperature dependent. At 40C, largeamounts of toxin A bind to the membrane. The toxin doesnot bind as well at higher temperatures, but the amount oftoxin A which binds to membranes at 370C is comparable tothe amount of cholera toxin which binds to the GM1 ganglio-side at that temperature.Toxin A also binds to rabbit erythrocytes and, when

present in high concentrations (micrograms per milliliter),agglutinates the cells (119). The binding to erythrocytesexhibits the same characteristics as the receptor in hamsterbrush border membranes, indicating that the binding com-ponents are probably identical. The receptor in each in-stance involves the trisaccharide Gala1-3GalP1-4GlcNAc.The terminal a-galactose residue is very important. Toxin Adoes not bind to erythrocytes treated with either ot-galacto-sidase, which enzymatically cleaves the terminal a-ga-lactose, or with the lectin from Bandeirea simplicifolia,which binds specifically to terminal a-galactose (172). Theentire trisaccharide sequence is needed for recognition bythe toxin. Hemagglutination by toxin A cannot be inhibitedby simple sugars such as galactose, by methyl-a-D-ga-lactoside, by stachyose and raffinose, both of which containGalal-6Gal linkages, or by the disaccharide Galal-3Gal. Thespecificity of the binding is further demonstrated by theobservation that toxin A does not bind to Salmonellamilwaukee, which expresses nonreducing Galal-3Gal41-3GalNAc on its lipopolysaccharide (140), or to Galal-3GalP1-4Glc or Galal-4GalI1-4GlcNAc (48). Bovine thyro-globulin is the only glycoconjugate which has been shown tocompetitively inhibit the binding of toxin A to rabbit eryth-rocytes. Thyroglobulin also contains the trisaccharideGa1oL-3Gal1P1-4GlcNAc and has been used to purify the


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

VO.,198C.DIFFICILE TOXINS 11

toxin by thermal affinity chromatography (see subsection,"Purification").One difference we have observed in the binding of toxin A

to rabbit erythrocytes and brush border membranes is thatthe low amount of binding of toxin A which occurs at 370Cwith membranes does not occur with erythrocytes. Essen-tially all of the toxin can be recovered from the erythrocytes.This probably reflects the unusual binding portion of toxin Aand the manner in which the receptor is expressed on brushborder membranes and rabbit erythrocytes. The bindingportion of toxin A consists of a number of repeats. This isbased on the sequence data obtained from a portion of thetoxin A gene which codes for a nontoxic, hemagglutinatingfragment of the toxin (see next subsection). On erythrocytes,the Galal-3Gal31-4GlcNAc trisaccharide is present singly oras branched structures containing two of the trisaccharideunits (47, 65, 95). In this instance, the toxin binds to thetrisaccharide by one or perhaps two of the numerous repeatswhich it contains. There are probably more of the trisaccha-ride structures on hamster brush border membranes, andthis increased number and clustering of these structuresmost likely results in binding through more of the repeats.Thus, this increased receptor density results in a tighterbinding of the toxin to the cell. As a result, the toxin cannotbe completely dissociated from the membrane receptors athigher temperatures.The multiple repeated binding sites on toxin A and the

apparent need for a high density of receptors for binding atbody temperature suggest the interesting possibility that thetoxin has evolved in such a way as to discriminate betweenthe receptor on the cell surface and free receptors (e.g.,soluble receptors and receptors on mucin). If the toxin had asingle high-affinity binding site, soluble receptors could bindand inactivate the toxin. Patients with diarrhea have largeamounts of mucin and partially digested cell membranes intheir colons, indicating that the toxin would have to discrim-inate between these receptors and receptors clustered on thecell surface. By binding only to areas of high receptordensity at body temperature, the toxin is able to do this.There is very little information on the receptor(s) needed

for the expression of the cytotoxic activity of toxins A andB. In both instances, the toxins bind to the cell and cluster incoated pits, indicating that receptor-mediated endocytosis isinvolved in the intoxication process (230; V. Kushnaryov,personal communication). The nature of the receptor in-volved in the binding, however, is not known. It has beenreported that tissue culture cells exposed to cholesterol,gangliosides, lecithin, and phospholipase C are still highlysensitive to the toxin, suggesting that the receptor is not alipid (229). Wheat germ agglutinin, concanavalin A, and ricinhave been reported to delay or reduce the cytotoxic effect,suggesting that the receptor on tissue culture cells may be acarbohydrate (229; N. J. Levy and A. B. Onderdonk,Program Abstr. 21st Intersci. Conf. Antimicrob. AgentsChemother. abstr. no. 719, 1981). Toxin B also has beenreported to bind to human erythrocyte ghosts and to rabbiterythrocytes (45, 230), but we have been unable to confirmthese findings. In any event, the toxin receptor(s) involved inthe cytotoxic activity will most likely prove to be some typeof ubiquitous molecule since both toxins are extremelyactive against such a wide variety of cell types.

Cloning of the Toxin A Gene

A 2.1-kilobase segment of the toxin A gene has recentlybeen cloned and sequenced (S. B. Price, C. J. Phelps, T. D.

Pst I

BCCIABBBC IABCCI ABBC IABBCI ABCI AB

F -_m- -_-_--_ _ _ _ _ _ _ _ _ _ _ _ _ _ _

4

FIG. 1. Diagram of the 2.1-kilobase toxin A gene insert. Theinsert represents a PstI restriction digest fragment of deoxyribonu-cleic acid from C. difficile VPI 10463. At least five large blocks ofrepeats (1, 2, 3, 4, and 5) are evident. Each of these blocks startswith the consensus sequence designated I consisting of about 90base pairs. The I sequence is followed by three, four, or five smallerrepeats (A, B, C), each about 60 base pairs. (Reprinted withpermission from J. Johnson, Department of Anaerobic Microbiol-ogy, Virginia Polytechnic Institute and State University, Blacks-burg).

Wilkins, and J. L. Johnson, Cuff. Microbiol., in press; C. H.Dove, S. B. Price, C. J. Phelps, and J. L. Johnson, Abstr.Annu. Meet. Am. Soc. Microbiol. 1987, D88, p. 86). Thetoxin A fragment coded for by the gene fragment reacts withpolyclonal and monoclonal antibodies against toxin A andshows a reaction of partial immunological identity withnative toxin A. In addition, it possesses the hemagglutinatingactivity, but not the enterotoxic, cytotoxic, or lethal activi-ties, associated with toxin A. Thus, either the portion of thegene coding for the toxic activities is not encoded within thisgene fragment or the gene product has not assumed thecorrect conformation for the expression of the toxic activity.The structure of the cloned gene fragment is very com-

plex. There are a number of blocks of repeating sequences.Within each of these larger blocks, there are four or fiverepeats which show decreasing levels of homology (Fig. 1).These smaller repeats consist of about 60 base pairs; thus,each of the repeats codes for a segment of about 20 aminoacids.The function of these repeating segments is not known,

but their presence helps to explain some of the unusualproperties of toxin A. For example, for a monoclonal anti-body to precipitate an antigen, the epitope recognized by theantibody must be present in multiple copies (either multiplesubunits or multiple internal repeats within a polypeptide).Therefore, these repeats in toxin A most likely represent theepitopes recognized by the PCG-4 monoclonal antibody thatprecipitates the toxin. Molecules which exhibit hemagglu-tinating activity must also be multivalent, and it is likely thatthe repeats on toxin A are involved in the recognition of theGalal-3GalI1-4GlcNAc trisaccharide receptor on rabbiterythrocytes and bovine thyroglobulin.The 2.1-kilobase fragment is located at the 3' end of the

gene, indicating that the binding portion of the toxin is at thecarboxylterminus. More than 4 kilobases of the gene havebeen sequenced upstream of this fragment (J. Johnson,personal communication). This entire 6.1-kilobase segmenthas an open reading frame, indicating that the toxin Apolypeptide has an Mr in excess of 250,000.

OTHER VIRULENCE FACTORS

A motility-altering factor which causes altered motoractivity in the intestine has been described in culture filtratesof C. difficile (105). This factor does not cause secretion andtissue damage, and it appears to be distinct from toxins Aand B. Nontoxigenic strains which are very similar to highlytoxigenic strains with the exception of toxin production donot cause disease in hamsters, suggesting that either the roleof the motility-altering factor in disease is minor or it is only

VOL. 1 a 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

12 LYERLY ET AL.

produced by toxigenic strains. Additional studies are neededto confirm the production of this factor and to learn moreabout its relevance to the disease.

In addition to the hemorrhagic enterotoxic activity causedby toxin A, a second nonhemorrhagic enterotoxic activityhas been described (12). This activity has an Mr of about200,000 and is not stable. These findings have not beenconfirmed, and it is not known whether this activity is due toa fragment of toxin A or whether it represents a distinctactivity.

PROPERTIES OF C. SORDELLII ENTEROTOXINAND CYTOTOXIN

C. sordeffli is not a common inhabitant of the humanintestine and is not recognized as an important humanpathogen. In animals, however, the organism causes enter-itis and enterotoxemia (7, 8, 178, 194). It has been known forsome time that this organism produces two toxins which arevery similar to toxins A and B of C. difficile. In fact, theidentification of C. difficile as the agent of PMC came aboutbecause investigators used C. sordeliji antitoxin to neutralizethe cytotoxic activity in stool specimens from PMC patients.

Like toxins A and B, the enterotoxin and cytotoxin of C.sordellii can be separated by ion-exchange chromatography(164, 250). The cytotoxin of C. sordellii has been purifiedby diethylaminoethyl-Trisacryl, gel filtration on UltrogelAcA34, and hydroxyapatite adsorption chromatography andhas an Mr of about 250,000 and a pI of 4.5 (179). Thecytotoxin has a 50% lethal dose in mice of about 3 ng,indicating that it is much more potent than toxin B. The 50%tissue culture infective dose of the cytotoxin, however, isabout 16 ng, indicating the toxin is much less active thantoxin B in tissue culture assays.The enterotoxin of C. sordellii has been purified by

ultrafiltration and immunoaffinity chromatography, using amonoclonal antibody against toxin A (R. D. Martinez andT. D. Wilkins, Abstr. Annu. Meet. Am. Soc. Microbiol.1987, B213, p. 60). The toxin has an Mr of about 500,000 anda pI of 5.8. It shows partial immunological identity to toxinA and, like toxin A, exhibits hemagglutinating and cytotoxicactivities.

It is not known whether the genes for the enterotoxin andcytotoxin of C. sordellii are located on the chromosome orare extrachromosomal. It is interesting that both of thetoxins are produced by strains of C. sordellii and that theyshare many properties with toxins A and B. This suggeststhat the toxin genes of C. difficile and C. sordellii evolvedfrom common ancestral genes.

FUTURE DIRECTIONS

Almost 10 years have elapsed since the discovery of C.difficile as the agent of PMC. During this time, we havelearned much about the organism and its toxins, but thereare many questions which remain to be answered. In theclinical aspects of the disease, we still do not have a clearpicture of the role of this organism in antibiotic-associateddiarrhea. There is some evidence that C. perfringens causesclinical illnesses similar to that caused by C. difficile, and wesuspect that other clostridia play a role in these types ofdiseases. Therefore, it is important that studies on theepidemiology of the disease continue and that clinicians beaware that other clostridia may be involved.There are other clinical areas which need to be addressed.

We have known for several years that many infants colo-

nized with toxigenic C. difficile are refractory to the disease,but we still do not know why this occurs. Research in thisarea may give us clues on better ways to treat the patient,especially those persons who relapse repeatedly with PMC.The toxins produced by C. difficile evidently lead to the

progression of the disease, but the manner in which thisoccurs appears to be much more complex than just the directtoxic action on the gut mucosa. The initial reaction appearsto involve the binding of toxin A to its receptor (possibly theGalal-3GalI1-4GlcNAc trisaccharide glycoconjugate) in thebrush border membrane, but the mechanism by which thetoxin kills mucosal cells and causes the enterotoxic responseis not known. The extensive damage which occurs from theaction of toxin A and the intense inflammatory infiltrateprobably results in the dissemination of both toxins A and Binto the systemic circulation. Toxin A which is no longerenterotoxic (i.e., neutralized with monoclonal antibody thatneutralizes the enterotoxic activity) is still lethal wheninjected. This raises the intriguing possibility that toxin Aacts by a different mechanism once it leaves the intestine.Toxin B also acts on tissues outside of the intestine, but thereceptor for the toxin and the target tissues have not beenidentified.Much work remains to be done on the structures of toxins

A and B. It has been suggested that the enterotoxic andcytotoxic activities of toxin A reside in small distinct sub-units complexed with a large hemagglutinin (230). The largesize of the toxin A gene indicates, however, that theseactivities reside in a single large polypeptide rather than inindividual subunits. Additional studies are needed to exam-ine these possibilities and to determine which structuralfeatures are shared by toxins A and B. Perhaps these studieswill give us some insight on their cytotoxic and lethalactivities.Not only are toxins A and B important, but it is also

important to learn how the organism becomes established inthe intestine and whether the organism produces othervirulence factors which are involved in the disease. Forexample, it is important to determine whether the labile,nonhemorrhagic enterotoxic activity which has been de-scribed is due to a fragment of toxin A or represents adistinct toxic activity. We hope that answers to these ques-tions will direct us to better ways of recognizing and treatingthe disease.

ACKNOWLEDGMENTS

Much of the work presented in this article was supported byPublic Health Service grant Al 15749 from the National Institutes ofHealth and State Support grant 2124520 from the Commonwealth ofVirginia.We thank Roger Van Tassell and Robert Carman for their

comments and suggestions.

LITERATURE CITED1. Abrams, G. D., M. Allo, G. D. Rifkin, R. Fekety, and J. Silva,

Jr. 1980. Mucosal damage mediated by clostridial toxin inexperimental clindamycin-associated colitis. Gut 21:493-499.

2. Adler, S. P., T. Chandrika, and W. F. Berman. 1981. Clostri-dium difficile associated with pseudomembranous colitis. Oc-currence in a 12-week-old infant without prior antibiotic ther-apy. Am. J. Dis. Child. 135:820-822.

3. Ahlgren, T., I. Florin, C. Jarstrand, and M. Thelestam. 1983.Loss of surface fibronectin from human lung fibroblasts ex-posed to cytotoxin from Clostridium difficile. Infect. Immun.39:1470-1472.

4. Al-Jumaili, I. J., and A. J. Bint. 1981. Simple method ofisolation and presumptive identification of Clostridium difficile.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt 1Orig. 250:142-146.

5. Al-Jumaili, I. J., M. Shibley, A. H. Lishmann, and C. 0.

Record. 1984. Incidence and origin of Clostridium difficile inneonates. J. Clin. Microbiol. 19:77-78.

6. Allo, M., J. Silva, Jr., R. Fekety, G. D. Riflin, and H. Waskin.1979. Prevention of clindamycin-induced colitis in hamsters byClostridium sordellii antitoxin. Gastroenterology 76:352-355.

7. Al Mashat, R. R., and D. J. Taylor. 1983. Bacteria in entericlesions of cattle. Vet. Rec. 112:5-10.

8. Al Mashat, R. R., and D. J. Taylor. 1983. Clostridium sordelliiin enteritis in an adult sheep. Vet. Rec. 112:19.

9. Arnon, S. S., D. C. Mills, P. A. Day, R. V. Henrickson, N. M.Sullivan, and T. D. Wilkins. 1984. Rapid death of infant rhesusmonkeys injected with Clostridium difficile toxins A and B:physiologic and pathologic basis. J. Pediatr. 101:3440.

10. Aronsson, B., M. Granstrom, R. Mollby, and C. E. Nord. 1985.Enzyme immunoassay for detection of Clostridium difficiletoxins A and B in patients with antibiotic-associated diarrhoeaand colitis. Eur. J. Clin. Microbiol. 4:102-107.

11. Bachamkin, I., L. Lotz-Nolan, and D. Skalina. 1986. Evaluationof a commercial cytotoxicity assay for detection of Clostridiumdifficile toxin. J. Clin. Microbiol. 23:954-955.

12. Banno, Y., T. Kobayashi, H. Kono, K. Watanabe, K. Ueno, andY. Nozawa. 1984. Biochemical characterization and biologicactions of two toxins (D-1 and D-2) from Clostridium difficile.Rev. Infect. Dis. 6:S11-S20.

13. Banno, Y., T. Kobayashi, K. Watanabe, K. Ueno, and Y.Nozawa. 1981. Two toxins (D-1 and D-2) of Clostridium difficilecausing antibiotic-associated colitis: purification and somecharacterization. Biochem. Int. 2:629-635.

14. Bartlett, J. G. 1979. Antibiotic-associated pseudomembranouscolitis. Rev. Infect. Dis. 1:530-539.

15. Bartlett, J. G. 1984. Treatment of antibiotic-associated pseu-domembranous colitis. Rev. Infect. Dis. 6:S235-S241.

16. Bartlett, J. G. 1985. Treatment of Clostridium difficile colitis.Gastroenterology 89:1192-1194.

17. Bartlett, J. G., T. W. Chang, M. Gurwith, S. L. Gorbach, andA. B..Onderdonk. 1978. Antibiotic-associated pseudomembra-nbus colitis due to toxin-producing clostridia. N. Engl. J. Med.298:531-534.

18. Bartlett, J. G., T. W. Chang, N. Moon, and A. B. Onderdonk.1978. Antibiotic-induced lethal enterocolitis in hamsters: stud-ies with eleven agents and evidence to support the pathogenicrole of toxin-producing clostridia. Am. J. Vet. Res. 39:1525-1530.

19. Bartlett, J. G., T. W. Chang, N. S. Taylor, and A. B. Onder-donk. 1979. Colitis induced by Clostridium difficile. Rev.Infect. Dis. 1:370-378.

20. Bartlett, J. G., N. Moon, T. W. Chang, N. Taylor, and A. B.Onderdonk. 1978. Role of Clostridium difficile in antibiotic-associated pseudomembranous colitis. Gastroenterology 75:778-782.

21. Bartlett, J. G., A. B. Onderdonk, and R. L. Cisneros. 1977.Clindamycin-associated colitis in hamsters: protection withvancomycin. Gastroenterology 73:772-776.

22. Bartlett, J. G., A. B. Onderdonk, R. L. Cisneros, and D. L.Kasper. 1977. Clindamycin-associated colitis due to a toxin-producing species of Clostridium in hamsters. J. Infect. Dis.136:701-705.

23. Bartlett, J. G., N. S. Taylor, T. Chang, and J. Dzink. 1980.Clinical and laboratory observations in Clostridium difficilecolitis. Am. J. Clin. Nutr. 33:2521-2526.

24. Bartlett, J. G., F. J. Tedesco, S. Shull, B. Lowe, and T. Chang.1980. Symptomatic relapse after oral vancomycin therapy ofantibiotic-associated pseudomembranous colitis. Gastroenter-ology 78:431-434.

25. Bartlett, J. G., S. H. Willey, T. W. Chang, and B. Lowe. 1979.Cephalosporin-associated pseudomembranous colitis due toClostridium difficile. J. Am. Med. Assoc. 242:2683-2685.

26. Bate, G. 1986. Comparison of Minitek Anaerobe Il®, APIAn-Identg, and RapID ANAO Systems for identification ofClostridium difficile. Am. J. Clin. Pathol. 85:716-718.

27. Boenning, D. A., G. C. Fleisher, J. M. Campos, C. W.Hulkower, and R. W. Quinlan. 1982. Clostridium difficile in apediatric outpatient population. Pediatr. Infect. Dis. 1:336-338.

28. Bolton, R. P., and M. A. Culshaw. 1986. Faecal metronidazoleconcentrations during oral and intravenous therapy for antibi-otic associated colitis due to Clostridium difficile. Gut 27:1169-1172.

29. Bolton, R. P., R. J. Sherriff, and A. E. Read. 1980. Clostridiumdifficile associated diarrhoea: a role in inflammatory boweldisease? Lancet i:383-384.

30. Bolton, R. P., S. K. Tait, P. R. F. Dear, and M. S. Losowsky.1984. Asymptomatic neonatal colonisation by Clostridium dif-ficile. Arch. Dis. Child. 59:466-472.

31. Borriello, S. P. 1979. Clostridium difficile and its toxin in thegastrointestinal tract in health and disease. Res. Clin. Forums1:33.

32. Borriello, S. P. 1984. Antibiotic-associated diarrhoea and co-litis. Martinus Nijhoff Publishers, Boston.

33. Borriello, S. P., and F. E. Barclay. 1984. Colonization resis-tance to Clostridium difficile infection. Microecol. Ther. 14:75-87.

34. Borriello, S. P., and F. E. Barclay. 1986. An in-vitro model ofcolonisation resistance to Clostridium difficile infection. J.Med. Microbiol. 21:299-309.

35. Borriello, S. P., F. E. Barclay, P. J. Reed, A. R. Welch, J. D.Brown, and D. W. Burdon. 1987. Analysis of latex agglutina-tion test for Clostridium difficile toxin A (D-1) and differentia-tion between C. difficile toxins A and B and latex reactiveprotein. J. Clin. Pathol. 40:573-580.

36. Borriello, S. P., F. E. Barclay, A. R. Welch, J. M. Ketley, T. J.Mitchell, J. Stephen, and G. E. Griffin. 1985. Host and micro-bial determinants of the spectrum of Clostridium difficilemediated gastrointestinal disorders. Microecol. Ther. 15:231-236.

37. Bowden, T. A., A. R. Mansberger, and L. E. Lykins. 1981.Pseudomembranous enterocolitis: mechanism of restoring flo-ral homeostasis. Am. Surg. 47:178-183.

38. Brettle, R. P., and E. Wallace. 1984. Clostridium difficile-associated diarrhoea. J. Infect. 8:123-128.

39. Brook, I. 1980. Isolation of toxin producing Clostridium diffi-cile from two children with oxacillin- and dicloxacillin-associ-ated diarrhea. Pediatrics 65:1154-1156.

40. Browne, R. A., R. Fekety, Jr., J. Silva, Jr., D. I. Boyd, C. 0.Work, and G. D. Abrams. 1977. The protective effect ofvancomycin on clindamycin-induced colitis in hamsters. JohnsHopkins Med. J. 141:183-192.

41. Buggy, B. P., K. H. Wilson, and R. Fekety. 1983. Comparisonof methods for recovery of Clostridium difficile from an envi-ronmental surface. J. Clin. Microbiol. 18:348-352.

42. Burdon, D. W., J. D. Brown, D. J. Youngs, Y. Arabi, N.Shinagawa, J. Alexander-Williams, and M. R. B. Keighley.1979. Antibiotic susceptibility of Clostridium difficile. J. Anti-microb. Chemother. 5:307-310.

43. Cashore, W. J., G. Peter, M. Lavermann, B. S. Stonestreet, andW. Oh. 1981. Clostridial colonization and clostridial toxin inneonatal necrotizing enterocolitis. J. Pediatr. 98:308-311.

44. Chang, T., J. G. Bartlett, S. L. Gorbach, and A. B. Onderdoihk.1978. Clindamycin-induced enterocolitis in hamsters as amodel of pseudomembranous colitis in patients. Infect. Im-mun. 20:526-529.

45. Chang, T.-W., S. L. Gorbach, and J. G. Bartlett. 1980. Inhibi-tion of binding of Clostridium difficile toxin by steroids. J.Infect. Dis. 142:113.

46. Chang, T.-W., N. M. Sullivan, and T. D. Wilkins. 1986.Insusceptibility of fetal intestinal mucosa and fetal cells toClostridium difficile toxins. Acta Pharmacol. Sin. 5:448-453.

47. Chien, J. L., S. C. Li, and Y. T. Li. 1978. Major glycosphingo-lipids of bovine-erythrocyte membranes, p. 241-254. In R. E.Harmon (ed.), Cell surface carbohydrate chemistry. AcademicPress, Inc., New York.

48. Clark, G. F., H. C. Krivan, T. D. Wilkins, and D. F. Smith.1987. Toxin A from Clostridium difficile binds to rabbit eryth-

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

14 LYERLY ET AL.

rocyte glycolipids with terminal Galal-3Galp1-4G0cNAc se-quences. Arch. Biochem. Biophys. 257:217-229.

49. Colardyn, F., G. Verschraegen, G. Claeys, and D. Vogelaers.1984. Clostridium dijsile-associated diarrhoea during treat-ment with imipenem. Eur. J. Clin. Microbiol. 3:565-566.

50. Collignon, A., J. Cotte-Laffitte, A. M. Quero, and J. C. Torlo-tin. 1986. Clostridium difficile and its cytotoxin in stools ofhospitalized infants. Influence of antibiotic treatment. Pathol.Biol. 34:977-982.

51. Corthier, G., F. Dubos, and P. Raibaud. 1985. Modulation ofcytotoxin production by Clostridium difficile in the intestinaltracts of gnotobiotic mice inoculated with various humanintestinal bacteria. Appl. Environ. Microbiol. 49:250-252.

52. Cudmore, M. A., J. Silva, R. Fekety, M. K. Liepman, andK.-H. Kim. 1982. Clostridium difficile colitis associated withcancer chemotherapy. Arch. Intern. Med. 142:333-335.

53. Czuprynski, C. J., W. J. Johnson, E. Balish, and T. Wilkins.1983. Pseudomembtranous colitis in Clostridium difficile-mono-associated rats. Infect. Immun. 39:1368-1376.

54. D'Ari, L., and H. A. Barker. 1985. p-Cresol formation bycell-free extracts of Clostridium difficile. Arch. Microbiol. 143:311-312.

55. Decoux, G. M., aqd C. Devroede. 1978. Acute colitis related tospiramycin. Lancet ii:993.

56. Deery, H. G., P. G. Jones, C. A. Kauffman, R. Fekety, J. Silva,T. L. Sell, and R. B. 5upena. 1984. Effect of therapy withlatamoxef (moxalactam) on carriage of Clostridium difficile. J.Antimicrob. Chemother. 13:521-524.

57. Delmee, M., G. Bulliard, and G. Simon. 1986. Application of atechnique for serogrouping Clostridium difficile in an outbreakof antibiotic-associated diarrhea. J. Infect. 13:5-9.

58. Donta, S. T., and M. G. Meyers. 1982. Clostridium difficiletoxin in asymptomatic neonates. J. Pediatr. 100:431-434.

59. Dorman, S. A., E. Liggoria, W. C. Winn, Jr., and W. L.Beeken. 1982. Isolation of Clostridium difficile from patientswith inactive Crohn's disease. Gastroenterology 82:1348-1351.

60. Dudley, M. N., J. C. McLaughlin, G. Carrington, J. Frick,C. H. Nightingale, and R. Quintiliani. 1986. Oral bacitracin vsvancomycin therapy for Clostridium difficile-induced diarrhea.Arch. Intern. Med. 146:1101-1104.

61. Dzink, J.,'and J. G. Bartlett. 1980. In vitro susceptibility ofClostridium difficile isolates from patients with antibiotic-associated diarrhea or colitis. Antimicrob. Agents Chemother.17:695-698.

62. Ehrich, M., R. L. Van Tassell, J. M. Libby, and T. D. Wilkins.1980. Production of Clostridium difficile antitoxin. Infect.Immun. 28:1041-1043.

63. Ellis, M. E., B. M. Watson, P. J. Milewski, and G. Jones. 1983.Clostridium difficile colitis unassociated with antibiotic ther-apy. Br. J. Surg. 70:242-243.

64. Elmer, G. W., and L. V. McFarland. 1987. Suppression bySaccharomyces boulardii of toxigenic Clostridium difficileovergrowth after vancomycin treatment in hamsters. Antimi-crob. Agents Chemother. 31:129-131.

65. Eto, T., Y. Ichikawa, K. Nishimura, S. Ando, and T. Yama-kawa. 1968. Chemistry of lipid of the posthemolytic residue orstroma of erythrocytes. XVI. Occurrence of ceramide penta-saccharide in the membrane of erythrocytes and reticulocytesof rabbit. J. Biochem. 64:205-213.

66. Fekety, R. 1979. Prevention and treatment of antibiotic-asso-ciated colitis, p. 276-279. In D. Schlessinger (ed.), Microbiol-ogy-1979. American Society for Microbiology, Washington,D.C.

67. Fekety, R., J. Silva, J. Armstrong, M. Allo, R. Browne, J.Ebright, R. Lusk, G. Rifkin, and R. Toshniwal. 1981. Treat-ment of antibiotic-associated enterocolitis with vancomycin.Rev. Infect. Dis. 3:S273-S281.

68. Fekety, R., J. Silva, R. Toshniwal, M. Allo, J. Armstrong, R.Browne, J. Ebright, and G. Rifkin. 1979. Antibiotic-associatedcolitis: effects of antibiotics on Clostridium difficile and thedisease in hamsters. Rev. Infect. Dis. 1:386-397.

69. Fernie, D. S., R. 0. Thomson, I. Batty, and P. D. Walker. 1983.Active and passive immunization to protect against antibiotic

associated caecitis in hamsters. Dev. Biol. Stand. 53:325-332.70. Finegold, S. M. 1986. Clinical considerations in the diagnosis of

antimicrobial agent-associated gastroenteritis. Diagn. Micro-biol. Infect. Dis. 4:87S-91S.

71. Finney, J. M. 1893. Gastro-enterostomy for cicatrizing ulcer ofthe pylorus. Johns Hopkins Hosp. Bull. 11:53-55.

72. Florin, I., and M. Thelestam. 1981. Intoxication of culturedhuman lung fibroblasts with Clostridium difficile toxin. Infect.Immun. 33:67-74.

73. Florin, I., and M. Thelestam. 1984. Polyphosphate-mediatedprotection from cellular intoxication with Clostridium difficiletoxin B. Biochim. Biophys. Acta 805:131-136.

74. Florin, I., and M. Thelestam. 1986. ADP-ribosylation in cul-tured cells treated with Clostridium difficile toxin B. Biochem.Biophys. Res. Commun. 139:64-70.

75. Florin, I., and M. Thelestam. 1986. Lysosomal involvement incellular intoxication with Clostridium difficile toxin B. Microb.Pathogen. 1:373-385.

76. Fournier, G., J. Orgiazzi, B. Lenoir, and M. Dechavanne. 1980.Pseudomembranous colitis probably due to rifampicin. Lanceti:101-102.

77. Frieben, W. R., and C. L. Duncan. 1973. Homology betweenenterotoxin protein and spore structural protein in Clostridiumperfringens type A. Eur. J. Biochem. 39:393-401.

78. George, R. H., M. Johnson, D. Youngs, and D. W. Burdon.1980. Induction of Clostridium difficile toxin by antibiotics, p.955-956. In J. D. Nelson and C. Grassi (ed.), Current chemo-therapy and infectious diseases, vol. 2. American Society forMicrobiology, Washington, D.C.

79. George, R. H., J. M. Symonds, F. Dimock, J. D. Brown, Y.Arabi, N. Shinagawa, M. R. B. Keighley, J. Alexander-Wil-liams, and D. W. Burdon. 1978. Identification of Clostridiumdifficile as a cause of pseudomembranous colitis. Br. Med. J.1:695.

80. George, W. L. 1980. Antimicrobial agent-associated colitis anddiarrhea. West. J. Med. 133:115-123.

81. George, W. L. 1984. Antimicrobial agent-associated colitis anddiarrhea: historical background and clinical aspects. Rev.Infect. Dis. 6:S208-S213.

82. George, W. L., E. J. C. Goldstein, V. L. Sutter, S. L. Ludwig,and S. M. Finegold. 1978. Etiology of antimicrobial-agent-associated colitis. Lancet i:802-803.

83. George, W. L., V. L. Sutter, D. Citron, and S. M. Finegold.1979. Selective and differential medium for isolation of Clostri-dium difficile. J. Clin. Microbiol. 9:214-219.

84. Gianfrilli, P., A. Pantosti, and I. Luzzi. 1985. Evaluation ofgas-liquid chromatography for the rapid diagnosis of Clostri-dium difficile associated disease. J. Clin. Pathol. 38:690-693.

85. Gilligan, P. H., L. R. McCarthy, and V. M. Genta. 1981.Relative frequency of Clostridium difficile in patients withdiarrheal disease. J. Clin. Microbiol. 14:26-31.

86. Goulston, S. J. M., and V. J. McGovern. 1965. Pseudo-membranous colitis. Gut 6:207-212.

87. Greenfield, C., J. R. Aguilar Ramirez, R. E. Pounder, T.Williams, M. Danvers, S. R. Marper, and P. Noone. 1983.Clostridium difficile and inflammatory bowel disease. Gut24:713-717.

88. Griffin, G. E., A. McDougall, S. A. Ash, and S. P. Borriello.1984. Age related susceptibility of hamsters to clindamycininduced Clostridium difficile caecitis. Microecol. Ther. 14:269-270.

89. Gurwith, M. J., C. Langston, and D. M. Citron. 1981. Toxin-producing bacteria in infants. Am. J. Dis. Child. 135:1104-1106.

90. Gurwith, M. J., H. R. Rabin, and K. Love. 1977. Diarrheaassociated with clindamycin and ampicillin therapy: prelimi-nary results of a cooperative study. J. Infect. Dis. 135:S104-S110.

91. Hachler, H., and J. Wust. 1984. Reexamination by bacterio-phage typing of Clostridium difficile strains isolated during anosocomial outbreak. J. Clin. Microbiol. 20:604.

92. Hafiz, S., R. S. Morton, M. G. McEntegart, and S. A. Waitkins.1975. Clostridium difficile in the urogenital tract of males and


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


females. Lancet i:420-421.93. Hall, I. C., and E. O'Toole. 1935. Intestinal flora in new-born

infants. Am. J. Dis. Child. 49:390-402.94. Han, V. K. M., H. Sayed, G. W. Chance, D. G. Brabyn, and

W. A. Shaheed. 1983. An outbreak of Clostridiulm difficilenecrotizing enterocolitis: a case for oral vancomycin therapy.Pediatrics 71:935-941.

95. Handfland, P., H. Egge, U. Dabrowski, S. Kuhn, D. Roelcke,and J. Dabrowski. 1981. Isolation and characterization of anI-active ceramide decasaccharide from rabbit erythrocytemembranes. Biochemistry 20:5310-5319.

96. Haslam, S. C., J. M. Ketley, T. J. Mitchell, J. Stephen, D. W.Burdon, and D. C. A. Candy. 1986. Growth of Clostridiumdifficile and production of toxins A and B in complex anddefined media. J. Med. Microbiol. 21:293-297.

97. Hayter, P. M., and J. W. Dale. 1984. Detection of plasmids inclinical isolates of Clostridium difficile. Microbios Lett. 27:151-156.

98. Holst, E., I. Helin, and P. A. Mardh. 1981. Recovery ofClostridium difficile from children. Scand. J. Infect. Dis.13:41-45.

99. Humphrey, C. D., C. W. Condon, J. R. Cantey, and F. E.Pittman. 1979. Partial purification of a toxin found in hamsterswith antibiotic-associated colitis. Gastroenterology 76:478-486.

100. Humphrey, C. D., W. B. Lushbaugh, C. W. Condon, J. C.Pittman, and F. E. Pittman. 1979. Light and electron micro-scopic studies of antibiotic associated colitis in the hamster.Gut 20:6-15.

101. laconis, J. P., and R. D. Rolfe. 1986. Clostridium difficile-associated ileocecitis in clindamycin-treated infant hamsters.Curr. Microbiol. 13:327-332.

102. Jarvis, W., 0. Nunez-Monteil, F. Thompson, V. Dowell, M.Towns, G. Morris, and E. Hill. 1983. Comparison of bacterialisolation, cytotoxicity assay, and counterimmunoelectropho-resis for the detection of Clostridium difficile and its toxin. J.Infect. Dis. 147:778.

103. Jarvis, W., 0. Nunez-Monteil, F. Thompson, V. Dowell, M.Towns, G. Morris, W. Vogler, E. Winton, and E. Hill. 1983.False-positive counterimmunoelectrophoresis tests for Clostri-diium difficile: the role of Clostridiurm bifermentans and Clo-stridiurn sordellii. J. Infect. Dis. 148:1168-1169.

104. Jones, R. L., W. S. Adney, and R. K. Shideler. 1987. Isolationof Clostridium difficile and detection of cytotoxin in the fecesof diarrheic foals in the absence of antimicrobial treatment. J.Clin. Microbiol. 25:1225-1227.

105. Justus, P. G., J. L. Martin, D. A. Goldberg, N. S. Taylor, J. G.Bartlett, R. W. Alexander, and J. R. Mathias. 1982. Myoelec-tric effects of Clostridinm difficile: motility-altering factorsdistinct from its cytotoxin and enterotoxin in rabbits. Gastro-enterology 83:836-843.

106. Kabins, S. A., and T. J. Spira. 1975. Outbreak of clindamycin-associated colitis. Ann. Intern. Med. 83:830-831.

107. Kahn, M. Y., and W. H. Hall. 1966. Staphylococcal enteroco-litis-treatment with oral vancomycin. Ann. Intern. Med.65:1-8.

108. Kamiya, S., S. Nakamura, K. Yamakawa, and S. Nishida. 1986.Evaluation of a commercially available latex immunoaggluti-nation test kit for detection of Clostridium difficile D-1 toxin.Microbiol. Immunol. 30:177-181.

109. Keighley, M. R. B., D. Young, M. Johnson, R. N. Allen, andD. W. Burdon. 1982. Clostridium difficile toxin in acute diar-rhoea complicating inflammatory bowel disease. Gut 23:410-414.

110. Kelly, M. T., S. G. Champagne, C. H. Sherlock, M. A. Noble,H. J. Freeman, and J. A. Smith. 1987. Commercial latexagglutination test for detection of Clostridium diffiule-asso-ciated diarrhea. J. Clin. Microbiol. 25:1244-1247.

111. Ketley, J. M., S. C. Haslam, T. J. Mitchell, J. Stephen, D. C. A.Candy, and D. W. Burdon. 1984. Production and release oftoxins A and B by Clostridiurm difficile. J. Med. Microbiol.18:385-391.

112. Ketley, J. M., T. J. Mitchell, S. C. Haslam, J. Stephen, D. C. A.

Candy, and D. W. Burdon. 1986. Sporogenesis and toxin Aproduction by Clostridiun difficile. J. Med. Microbiol. 22:33-38.

113. Kim, K., H. L. Dupont, and L. K. Pickering. 1983. Outbreaksof diarrhea associated with Clostridium difficile and its toxin inday-care centers: evidence of person-to-person spread. J.Pediatr. 102:376-382.

114. Kim, K., L. K. Pickering, H. L. Dupont, N. Sullivan, and T.Wilkins. 1984. In vitro and in vivo neutralizing activity ofhuman clostrum and milk against purified toxins A and B ofClostridium difficile. J. Infect. Dis. 150:57-62.

115. Kim, K.-H., R. Fekety, D. H. Batts, D. Brown, M. Cudmore, J.Silva, Jr., and D. Waters. 1981. Isolation of Clostridiumdifficile from the environment and contacts of patients withantibiotic-associated colitis. J. Infect. Dis. 143:42-50.

116. Knoop, F. C. 1979. Clindamycin-associated enterocolitis inguinea pigs: evidence for a bacterial toxin. Infect. Immun.23:31-33.

117. Kreutzer, E. W., and F. D. Milligan. 1978. Treatment ofantibiotic-associated pseudomembranous colitis with cholesty-ramine resins. Johns Hopkins Med. J. 143:67-72.

118. Krishnan, C. 1986. Detection of Clostridiurm difficile toxins byenzyme immunoassay. J. Hyg. 96:5-12.

119. Krivan, H. C., G. F. Clark, D. F. Smith, and T. D. Wilkins.1986. Cell surface binding site for Clostridium difficile entero-toxin: evidence for a glycoconjugate containing the sequenceGalcl-3Galp1-4GlcNAc. Infect. Immun. 53:573-581.

120. Krivan, H. C., and T. D. Wilkins. 1987. Purification of Clostri-dium difficile toxin A by affinity chromatography on immobi-lized thyroglobulin. Infect. Immun. 55:1873-1877.

121. Kudhair, A., M. Delmee, and G. Wauters. 1986. Rapid identi-fication of Clostridinm difficile by determination of leucinearylamidase activity. Eur. J. Clin. Microbiol. 5:160-161.

122. Kupier, E. J., J. H. Oudbier, W. N. H. M. Stuifbergen, A.Jansz, and H. C. Zanen. 1987. Application of whole-cell DNArestriction endonuclease profiles to the epidemiology of Clo-stridiurm difficile-induced diarrhea. J. Clin. Microbiol. 25:751-753.

123. LaMont, J. T., and Y. M. Trnka. 1980. Therapeutic implica-tions of Clostridimn difijcile toxin during relapse of chronicinflammatory bowel disease. Lancet i:381-383.

124. Larson, H. E., F. E. Barclay, P. Honour, and I. D. Hill. 1982.Epidemiology of Clostridiulm difficile in infants. J. Infect. Dis.146:727-733.

125. Larson, H. E., P. Honour, A. B. Price, and S. P. Borriello. 1978.Clostridinm difficile and the etiology of pseudomembranouscolitis. Lancet i:1063-1066.

126. Larson, H. E., J. V. Parry, A. B. Price, D. R. Davies, J. Dolby,and D. A. J. Tyrrell. 1977. Undescribed toxin in pseudomem-branous colitis. Br. Med. J. 1:1246-1248.

127. Larson, H. E., and A. B. Price. 1977. Pseudomembranouscolitis: presence of clostridial toxin. Lancet. ii:1312-1314.

128. Laughon, B. E., R. P. Viscidi, S. L. Gdovin, R. H. Yolken, andJ. G. Bartlett. 1984. Enzyme immunoassays for detection ofClostridinm difficile toxins A and B in fecal specimens. J.Infect. Dis. 149:781-788.

129. Levett, P. N. 1984. Detection of Clostridiulm dificile in faecesby direct gas liquid chromatography. J. Clin. Pathol. 37:117-119.

130. Levett, P. N. 1985. Effect of antibiotic concentration in aselective medium on the isolation of Clostridiurm difficile fromfaecal specimens. J. Clin. Pathol. 38:233-234.

131. Levett, P. N. 1985. Effect of basal medium upon fluorescence ofClostridium dfficile. Lett. Appl. Microbiol. 1:75-76.

132. Levett, P. N. 1985. Identification of Clostridium difficile usingthe API ZYM system. Eur. J. Clin. Microbiol. 4:505-507.

133. Levett, P. N. 1986. Clostridium difficile in habitats other thanthe human gastrointestinal tract. J. Infect. 12:253-263.

134. Levett, P. N., and K. D. Phillips. 1985. Gas chromatographicidentification of Clostridiuln difficile and detection of cytotoxinfrom a modified selective medium. J. Clin. Pathol. 38:82-85.

135. Levine, H. G., M. Kennedy, and J. T. LaMont. 1982. Counter-immunoelectrophoresis versus cytotoxicity assay for the de-

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

16 LYERLY ET AL.

tection of Clostridium difficile toxin. J. Infect. Dis. 145:398.136. Libby, J. M., S. T. Donta, and T. D. Wilkins. 1983. Clostridium

difficile toxin A in infants. J. Infect. Dis. 148:606.137. Libby, J. M., B. S. Jortner, and T. D. Wilkins. 1982. Effects of

the two toxins of Clostridium difficile in antibiotic-associatedcecitis in hamsters. Infect. Immun. 36:822-829.

138. Libby, J. M., and T. D. Wilkins. 1982. Production of antitoxinsto two toxins of Clostridium difficile and immunological com-parison of the toxins by cross-neutralization studies. Infect.Immun. 35:374-376.

139. Lonnroth, I., and S. Lange. 1983. Toxin A of Clostridiumdifficile: production, purification and effect in the mouse intes-tine. Acta Pathol. Microbiol. Immunol. Scand. 91:395-400.

140. Luderitz, O., D. A. R. Simmons, and 0. Westphal. 1965. Theimmunochemistry of Salmonella chemotype VI 0-antigens.Biochem. J. 97:820-826.

141. Lusk, R. H., R. Fekety, J. Silva, R. A. Browne, D. H. Ringler,and G. D. Abrams. 1978. Clindamycin-induced enterocolitis inhamsters. J. Infect. Dis. 137:464 475.

142. Lyerly, D. M., D. E. Lockwood, S. H. Richardson, and T. D.Wilkins. 1982. Biological activities of toxins A and B ofClostridium difficile. Infect. Immun. 35:1147-1150.

143. Lyerly, D. M., C. J. Phelps, J. Toth, and T. D. Wilkins. 1986.Characterization of toxins A and B of Clostridium difficile withmonoclonal antibodies. Infect. Immun. 54:70-76.

144. Lyerly, D. M., C. J. Phelps, and T. D. Wilkins. 1985. Mono-clonal and specific polyclonal antibodies for immunoassay ofClostridium difficile toxin A. J. Clin. Microbiol. 21:12-14.

145. Lyerly, D. M., M. D. Roberts, C. J. Phelps, and T. D. Wilkins.1986. Properties of toxins A and B of Clostridium difficile.FEMS Microbiol. Lett. 33:31-35.

146. Lyerly, D. M., K. E. Saum, D. MacDonald, and T. D. Wilkins.1985. Effect of toxins A and B given intragastrically to animals.Infect. Immun. 47:349-352.

147. Lyerly, D. M., N. M. Sullivan, and T. D. Wilkins. 1983.Enzyme-linked immunosorbent assay for Clostridium difficiletoxin A. J. Clin. Microbiol. 17:72-78.

148. Lyerly, D. M., and T. D. Wilkins. 1986. Commercial latex testfor Clostridium difficile toxin A does not detect toxin A. J.Clin. Microbiol. 23:622-623.

149. Mahoney, D. E., P. D. Bell, and K. B. Easterbrook. 1985. Twobacteriophages of Clostridium difficile. J. Clin. Microbiol.21:251-254.

150. Makin, T. 1984. Rapid identification of Clostridium difficile bydirect detection of volatile organic acids from primary isolationmedia. J. Clin. Pathol. 37:711-712.

151. Malamou-Ladas, H,, S. O'Farrell, J. Q. Nash, and S. Tabaq-chali. 1983. Isolation of Clostridium difficile from patients andthe environment of hospital wards. J. Clin. Pathol. 36:88-92.

152. Mardh, P.-A., I. Helin, I. Colleen, M. Oberg, and E. Hoist.1982. Clostridium difficile toxin in faecal specimens of healthychildren and children with diarrhoea. Acta Pediatr. Scand.71:275-278.

153. Marler, L., S. Allen, and J. Siders. 1984. Rapid enzymaticcharacterization of clinically encountered anaerobic bacteriawith the API ZYM system. Eur. J. Clin. Microbiol. 3:294-300.

154. Massot, J., 0. Sanchez, R. Couchy, J. Astoin, and A. L. Parodi.1984. Bacterio-pharmacological activity of Saccharomycesboulardii in clindamycin-induced colitis in the hamster. Arz-neim. Forsch. 34:794-797.

155. Meyers, S., L. Mayer, E. Bottone, E. Desmond, and H. D.Janowitz. 1981. Occurrence of Clostridium difficile toxin duringthe course of inflammatory bowel disease. Gastroenterology80:697-700.

156. Mitchell, T. J., J. M. Ketley, S. C. Haslam, J. Stephen, D. W.Burdon, D. C. A. Candy, and R. Daniel. 1986. Effect of toxin Aand B of Clostridium difficile on rabbit ileum and colon. Gut27:78-85.

157. Moskovitz, M., and J. G. Bartlett. 1981. Recurrent pseudo-membranous colitis unassociated with prior antibiotic therapy.Arch. Intern. Med. 141:663464.

158. Muldrow, L. L., E. R. Archibald, 0. L. Nunez-Montiel, andR. J. Sheehy. 1982. Survey of the extrachromosomal gene pool

of Clostridium difficile. J. Clin. Microbiol. 16:637-640.159. Mulligan, M. E. 1984. Epidemiology of Clostridium difficile-

induced intestinal disease. Rev. Infect. Dis. 6:S222-S228.160. Mulligan, M. E., S. Halebian, R. Y. Y. Kwok, W. C. Cheng,

S. M. Finegold, C. R. Anselmo, D. N. Gerding, and L. R.Peterson. 1986. Bacterial agglutination and polyacrylamide gelelectrophoresis for typing Clostridium difficile. J. Infect. Dis.153:267-271.

161. Mulligan, M. E., R. D. Rolfe, S. M. Finegold, and W. L.George. 1979. Contamination of a hospital environment byClostridium difficile. Curr. Microbiol. 3:173-175.

162. Nachamkin, I., L. Lotz-Nolan, and D. Skalina. 1986. Evaluationof a commercial cytotoxicity assay for detection of Clostridiumdifficile toxin. J. Clin. Microbiol. 23:954-955.

163. Nakamura, S., M. Mikawa, N. Tanabe, K. Yamakawa, and S.Nishida. 1982. Effect of clindamycin on cytotoxin productionof Clostridium difficile. Microbiol. Immunol. 26:985-992.

164. Nakamura, S., N. Tanabe, K. Yamakawa, and S. Nishida. 1983.Cytotoxin production by Clostridium sordellii strains. Micro-biol. Immunol. 27:495-502.

165. O'Farrell, S., M. Wilks, J. Q. Nash, and S. Tabaqchali. 1984. Aselective enrichment broth for the isolation of Clostridiumdifficile. J. Clin. Pathol. 37:98-99.

166. Onderdonk, A. B., R. L. Cisneros, and J. G. Bartlett. 1980.Clostridium diffidile in gnotobiotic mice. Infect. Immun. 28:277-282.

167. Onderdonk, A. B., B. R. Lowe, and J. G. Bartlett. 1981. Effectof environmental stress on Clostridium difficile toxin levelsduring continuous cultivation. Appl. Environ. Microbiol. 38:637-641.

168. Orchard, J. L., R. Fekety, and J. R. Smith. 1983. Antibiotic-associated colitis due to Clostridium difficile in a Kodiak bear.Am. J. Vet. Res. 44:1547-1548.

169. Peach, S. L., S. P. Borriello, H. Gayna, F. E. Barclay, and A. R.Welch. 1986. Asymptomatic carriage of Clostridium difficile inpatients with cystic fibrosis. J. Clin. Pathol. 39:1013-1018.

170. Peikin, S. R., J. Galdibini, and J. G. Bartlett. 1980. Role ofClostridium difficile in a case of non-antibiotic-associated pseu-domembranous colitis. Gastroenterology 79:948-951.

171. Pepersack, F., M. Labbe, C. Nonhoff, and E. Schoutens. 1983.Use of gas-liquid chromatography as a screening test fortoxigenic Clostridium difficile in diarrhoeal stools. J. Clin.Pathol. 36:1233-1236.

172. Peters, B. P., and I. J. Goldstein. 1979. The use of fluorescein-conjugated Bandeirea simplicifolia B4-isolectin as a histochem-ical reagent for the detection of aL-D-galactopyranosyl groups.Exp. Cell Res. 120:321-334.

173. Peterson, L. R., J. J. Holter, C. J. Shanholtzer, C. R. Garrett,and D. N. Gerding. 1986. Detection of Clostridium difficiletoxin A (enterotoxin) and B (cytotoxin) in clinical specimens.Am. J. Clin. Pathol. 86:208-211.

174. Peterson, L. R., J. J. Holter, C. J. Shanholtzer, C. R. Garrett,and D. N. Gerding. 1987. Detection of C. difficile toxin A(enterotoxin). Am. J. Clin. Pathol. 87:298-299.

175. Phillips, K. D., and P. A. Rogers. 1981. Rapid detection andpresumptive identification of Clostridium difficile by p-cresolproduction on a selective medium. J. Clin. Pathol. 34:6424644.

176. Phua, T. J., T. R. Rogers, and A. P. Pallett. 1984. Prospectivestudy of Clostridium difficile colonization and para-cresoldetection in the stools of babies on a special care unit. J. Hyg.93:17-25.

177. Pierce, P. F., Jr., R. Wilson, J. Silva, Jr., V. F. Garagusi, G. D.Riflin, R. Fekety, 0. Nunez-Montiel, V. R. Dowell, Jr., andJ. M. Hughes. 1982. Antibiotic-associated pseudomembranouscolitis: an epidemiologic investigation of a cluster of cases. J.Infect. Dis. 145:269-274.

178. Popoff, M. R. 1984. Bacteriological examination in enteroto-xaemia of sheep and lamb. Vet. Rec. 114:324.

179. Popoff, M. R. 1987. Purification and characterization of Clo-stridium sordellii lethal toxin and cross-reactivity with Clostri-dium difficile cytotoxin. Infect. Immun. 55:35-43.

180. Pothoulakis, C., L. M. Barone, R. Ely, B. Faris, M. E. Clark,C. Franzblau, and J. T. LaMont. 1986. Purification and prop-


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.


erties of Clostridium difficile toxin B. J. Biol. Chem. 261:1316-1321.

181. Pothoulakis, C., G. Triadafilopoulos, M. Clark, C. Franzblau,and J. T. LaMont. 1986. Clostridium difficile cytotoxin inhibitsprotein synthesis in fibroblasts and intestinal mucosa. Gastro-enterology 91:1147-1153.

182. Potvliege, C., M. Labbe, and E. Youassowsky. 1981. Gas liquidchromatography as screening test for Clostridium difficile.Lancet ii:1105.

183. Poxton, I. R., B. Aronsson, R. Mollby, C. E. Nord, and J. G.Collee. 1984. Immunochemical fingerprinting of Clostridiumdifficile strains isolated from an outbreak of antibiotic-associ-ated colitis and diarrhoea. J. Med. Microbiol. 17:317-324.

184. Poxton, I. R., and M. D. Byrne. 1981. Detection of Clostridiumdifficile toxin by counterimmunoelectrophoresis: a note ofcaution. J. Clin. Microbiol. 14:349.

185. Poxton, I. R., and M. D. Byrne. 1981. Immunological analysisof the EDTA-soluble antigens of Clostridium difficile andrelated species. J. Gen. Microbiol. 122:41-46.

186. Poxton, I. R., and M. D. Byrne. 1984. Demonstration of sharedantigens in the genus Clostridium by an enzyme-linked im-munosorbent assay. J. Med. Microbiol. 17:171-176.

187. Poxton, I. R., and T. D. I. Cartmill. 1982. Immunochemistry ofthe cell-surface carbohydrate antigens of Clostridium difficile.J. Gen. Microbiol. 128:1365-1370.

188. Price, A. B., H. E. Larson, and J. Crow. 1979. Morphology ofexperimental antibiotic-associated enterocolitis in the hamster:a model for human pseudomembranous colitis and antibiotic-associated diarrhoea. Gut 20:467-475.

189. Prigogine, T., C. Potvliege, A. Burette, T. Verbeet, and J.Schmerber. 1981. Pseudomembranous colitis and rifampicin.Chest 80:766-767.

190. Ramirez-Ronda, C. H. 1974. Incidence of clindamycin-associ-ated colitis. Comments and corrections. Ann. Intern. Med.81:860.

191. Rautenberg, P., F. Stender, and U. Ullman. 1986. Detection oftoxin A by immunoblotting. Zentralbl. Bakteriol. Parasitenkd.Infektionskr. Hyg. Abt. 1 Orig. 261:29-42.

192. Rehg, J. E. 1980. Cecal toxin(s) from guinea pigs with clinda-mycin-associated colitis, neutralized by Clostridium sordelliiantitoxin. Infect. Immun. 27:387-390.

193. Rehg, J. E., and S. P. Pakes. 1982. Implication of Clostridiumdifficile and Clostridium perfringens iota toxins in experimentallincomycin-associated colitis of rabbits. Lab. Anim. Sci.32:253-257.

194. Richards, S. M. 1982. Clostridium sordellii in lambs. Vet. Rec.111:22.

195. Rietra, P. J. G. M., K. W. Slaterus, H. C. Zanen, and S. G. M.Meuwissen. 1978. Clostridial toxins in faeces of healthy infants.Lancet ii:319.

196. Rifkin, G. D., R. Fekety, and J. Silva. 1978. Neutralization byClostridium sordellii antitoxin of toxins implicated in clinda-mycin-induced cecitis in the hamster. Gastroenterology 75:422-424.

197. Rifkin, G. D., J. Silva, Jr., and F. R. Fekety. 1978. Gastroin-testinal and systemic toxicity of fecal extracts from hamsterswith clindamycin-induced colitis. Gastroenterology 74:52-57.

198. Rifkin, G. D., J. Silva, Jr., F. R. Fekety, and R. B. Sack. 1977.Antibiotic-induced colitis. Implication of a toxin neutralised byClostridium sordellii antitoxin. Lancet ii:1103-1106.

199. Rihn, B., G. Beck, H. Monteil, F. Lecuf, and R. Girardot. 1985.Effect of the cytotoxin of Clost'ridium difficile on culturedhepatoma cells. Biol. Cell 53:23-32.

200. Rihn, B., J. M. Scheftel, R. Girardot, and H. Monteil. 1984. Anew purification procedure for Clostridium difficile entero-toxin. Biochem. Biophys. Res. Commun. 124:690-695.

201. Robert, A., G. L. Bundy, S. 0. Field, J. E. Nezamis, J. P.David, A. J. Hanchar, C. Lancaster, and M. J. Ruwart. 1985.Prevention of cecitis in hamsters by certain prostaglandins.Prostaglandins 29:961-980.

202. Rocca, J. M., A. S. Peiterse, R. Rowland, R. Hecker, and G. E.Rich. 1984. Clostridium difficile colitis. Aust. N. Z. J. Med.14:606-610.

203. Rolfe, R. D., and J. P. laconis. 1983. Intestinal colonization ofinfant hamsters with Clostridium difficile. Infect. Immun.42:480-486.

204. Rothman, S. W., and J. E. Brown. 1981. Inhibition of mem-brane functions in intact HeLa cells by Clostridium difficilecytotoxic culture filtrates. Curr. Microbiol. 6:221-224.

205. Ruwart, M. J., H. E. Renis, A. L. Erlandson, R. M. DeHaan,J. E. Nezamis, C. Lancaster, and J. P. Davis. 1983. 16,16-Dimethyl PGE2 inhibits toxin release from Clostridium difficile,p. 93-98. In B. Samuelsson (ed.), Advances in prostaglandins& thromboxane research, vol. 12. Raven Press, New York.

206. Ryan, R. W. 1986. Considerations in the laboratory diagnosisof antibiotic-associated gastroenteritis. Diagn. Microbiol. In-fect. Dis. 4:79S-86S.

207. Saginur, R., R. Fogel, L. Begin, B. Cohen, and J. Mendelson.1983. Splenic abscess due to Clostridium difficile. J. Infect.Dis. 147:1105.

208. Sands, M., M. Yungbluth, and H. M. Sommers. 1983. Thenon-value of counterimmunoelectrophoresis for the directrapid detection of Clostridium difficile toxin in stool filtrates.Am. J. Clin. Pathol. 79:375-377.

209. Schallehn, G. 1985. Clostridium sordellii bacteriophage activeon Clostridium difficile. Lett. Apple. Microbiol. 1:71-74.

210. Schwan, A., S. Sjolin, U. Trottestam, and B. Aronsson. 1984.Relapsing Clostridium difficile enterocolitis cured by rectalinfusion of normal faeces. Scand. J. Infect. Dis. 16:211-215.

211. Sell, T. L., D. R. Schaberg, and F. R. Fekety. 1983. Bacterio-phage and bacteriocin typing scheme for Clostridium difficile.J. Clin. Microbiol. 17:1148-1152.

212. Sherertz, R. J., and F. A. Sarubbi. 1982. The prevalence ofClostridium difficile in a nursery population: a comparisonbetween patients with necrotizing enterocolitis and a asymp-tomatic group. J. Pediatr. 100:435-441.

213. Silva, J., R. Fekety, C. Werk, J. Ebright, M. Cudmore, D.Batts, C. Syrjamaki, and J. Lukens. 1984. Inciting and etiologicagents of colitis. Rev. Infect. Dis. 6:S214-S221.

214. Small, J. D. 1968. Fatal enterocolitis in hamsters given lin-comycin hydrochloride. Lab. Anim. Care 18:411-420.

215. Smith, M. F., S. P. Borriello, G. S. Clayden, and M. W.Casewell. 1980. Clinical and bacteriological findings in necro-tizing enterocolitis: a controlled study. J. Infect. Dis. 7:23-31.

216. Stark, P. L., A. Lee, and B. D. Parsonage. 1982. Colonizationof the large bowel by Clostridium difficile in healthy infants:quantitative study. Infect. Immun. 35:895-899.

217. Stephen, J., S. C. Redmond, T. J. Mitchell, J. Ketley, D. C. A.Candy, D. W. Burdon, and R. Daniel. 1984. Clostridiumdifficile enterotoxin (toxin A): new results. Biochem. Soc.Trans. 12:194-195.

218. Sterne, M., and L. M. Wentzel. 1950. A new method for thelarge-scale production of high-titre botulinum formol-toxoidtypes C and D. J. Immunol. 65:175-183.

219. Sugiyama, T., M. Mukai, R. Yamashita, and K. Sunakawa.1985. Experimental models of Clostridium difficile enterocolitisin gnotobiotic mice, p. 203-206. In Germfree research: micro-flora control and its application to the biomedical sciences.Alan R. Liss, Inc., New York.

220. Sullivan, N. M., S. Pellett, and T. D. Wilkins. 1982. Purificationand characterization of toxins A and B of Clostridium difficile.Infect. Immun. 35:1032-1040.

221. Svedhem, A., B. Kaijser, and I. MacDowall. 1982. Intestinaloccurence of Campylobacter fetus subspecies jejuni and Clo-stridium difficile in children in Sweden. Eur. J. Clin. Microbiol.1:29-32.

222. Swartzberg, J. E., R. M. Maresca, and J. S. Remington. 1977.Clinical study of gastrointestinal complications associated withclindamycin therapy. J. Infect. Dis. 135:S99-S103.

223. Swindlehurst, J., T. G. Tachovsky, and M. Knoppers. 1986.Antibiotics are necessary for plasmid isolation from Clostri-dium difficile, p. 451-452. In D. L. Lark (ed.), Protein-carbo-hydrate interactions in biological systems. The molecular biol-ogy of microbial pathogenesis. Academic Press, Inc., NewYork.

224. Taylor, N. S., G. M. Thorne, and J. G. Bartlett. 1981. Com-

VOL. 1, 1988

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

18 LYERLY ET AL.

parison of two toxins produced by Clostridium difficile. Infect.Immun. 34:1036-1043.

225. Teasley, D. G., D. N. Gerding, M. M. Olson, L. R. Peterson,R. L. Begbard, M. J. Schwartz, and J. T. Lee, Jr. 1983.Prospective randomized trial of metronidazole versus vanco-mycin for Clostridium difiile-associated diarrhoea and colitis.Lancet ii:1043-1046.

226. Tedesco, F. J. 1976. Clindamycin-associated colitis. Review ofthe clinical spectrum of 47 cases. Digest. Dis. 21:26-32.

227. Tedesco, F. J., R. W. Barton, and D. H. Alpers. 1974. Clinda-mycin-associated colitis: a prospective study. Ann. Intern.Med. 81:429-433.

228. Tedesco, F. J., J. Napier, W. Gamble, T. W. Chang, and J. G.Bartlett. 1979. Therapy of antibiotic-associated pseudomem-branous colitis. J. Clin. Gastroenterol. 1:51-54.

229. Thelestam, M., and M. Bronnegard. 1980. Interaction of cyto-pathogenic toxin from Clostridium difficile with cells in tissueculture. Scand. J. Infect. Dis. 22:16-29.

230. Thelestam, M., and I. Florin. 1984. Cytopathogenic action ofClostridium difficile toxins. J. Toxicol. Toxin Rev. 3:139-180.

231. Toffler, R. B., E. G. Pingoud, and M. I. Burrell. 1978. Acutecolitis related to penicillin derivatives. Lancet ii:707-709.

232. Toothaker, R. D., and G. W. Elmer. 1984. Prevention ofclindamycin-induced mortality in hamsters by Saccharomycesboulardii. Antimicrob. Agents Chemother. 26:552-556.

233. Torres, J. F., R. Cedillo, J. Sanchez, C. Dillman, S. Giono, and0. Munoz. 1984. Prevalence of Clostridium difficile and itscytotoxin in infants in Mexico. J. Clin. Microbiol. 20:274-275.

234. Trnka, Y. M., and J. T. LaMont. 1981. Association of Clostri-dium difficile toxin with symptomatic relapse of chronic inflam-matory bowel disease. Gastroenterology 80:693-696.

235. Vesikari, T., E. Isolauri, M. Maki, and P. Gronroos. 1984.Clostridium difficile in young children. Acta Pediatr. Scand.73:86-91.

236. Viscidi, R., S. Wiley, and J. G. Bartlett. 1981. Isolation ratesand toxigenic potential of Clostridium difficile isolates fromvarious patient populations. Gastroenterology 81:5-9.

237. Viscidi, R. P., and J. G. Bartlett. 1981. Antibiotic-associatedpseudomembranous colitis in children. Pediatrics 67:381-386.

238. Wada, N., N. Nishida, S. Iwaki, H. Ohi, T. Miyawaki, N.Taniguchi, and S. Migita. 1980. Neutralizing activity againstClostridium difficile toxin in the supernatants of culturedcolostral cells. Infect. Immun. 29:545-550.

239. Wald, A., H. Mendelow, and J. B. Bartlett. 1980. Nonantibioticassociated pseudomembranous colitis due to toxin producingclostridia. Ann. Intern. Med. 92:798-799.

240. Walker, R. C., P. J. Ruane, J. E. Rosenblatt, D. M. Lyerly,

C. A. Gleaves, T. F. Smith, P. F. Pierce, Jr., and T. D. Wilkins.1986. Comparison of culture, cytotoxicity assays, and enzyme-linked immunosorbent assay for toxin A and toxin B in thediagnosis of Clostridium difficile-related enteric disease.Diagn. Microbiol. 5:61-69.

241. Welkon, C. J., S. S. Long, C. M. Thompson, Jr., and P. H.Gilligan. 1985. Clostridium difficile in patients with cysticfibrosis. Am. J. Dis. Child. 139:805-808.

242. West, S. E. H., and T. D. Wilkins. 1982. Problems associatedwith counterimmunoelectrophoresis assays for detecting Clo-stridium difficile toxin. J. Clin. Microbiol. 15:347-349.

243. Wilson, K. H., and R. Freter. 1985. Suppression of Clostridiumdifficile by isolates from indigenous hamster flora. Microecol.Ther. 15:323-327.

244. Wilson, K. H., M. J. Kennedy, and F. R. Fekety. 1982. Use ofsodium taurocholate to enhance spore recovery on a mediumselective for Clostridium difficile. J. Clin. Microbiol.15:443-446.

245. Wilson, K. H., J. N. Sheagren, R. Freter, L. Weatherbee, andD. Lyerly. 1986. Gnotobiotic models for study of the microbialecology of Clostridium difficile and Escherichia coli. J. Infect.Dis. 153:547-551.

246. Wilson, K. H., J. Silva, and F. R. Fekety. 1981. Suppression ofClostridium difficile by normal hamster cecal flora and preven-tion of antibiotic-associated cecitis. Infect. Immun. 34:626-628.

247. Wu, T. C., and J. C. Fung. 1983. Evaluation of the usefulnessof counterimmunoelectrophoresis for diagnosis of Clostridiumdifiile-associated colitis in clinical specimens. J. Clin. Micro-biol. 17:610-613.

248. Wu, T. C., and S. M. Gersch. 1986. Evaluation of a commercialkit for the routine detection of Clostridium difficile cytotoxinby tissue culture. J. Clin. Microbiol. 23:792-793.

249. Wu, T. C., V. P. McCarthy, and V. J. Gill. 1983. Isolation rateand toxigenic potential of Clostridium difficile isolates frompatients with cystic fibrosis. J. Infect. Dis. 148:176.

250. Yamakawa, K., S. Nakamura, and S. Nishida. 1985. Separationof two cytotoxins of Clostridium sordellii strains. Microbiol.Immunol. 29:553-557.

251. Yamakawa, K., S. Nishida, and S. Nakamura. 1983. C2 toxicityin extract of Clostridium botulinum type C spores. Infect.Immun. 41:858-860.

252. Young, G. P., P. B. Ward, N. Bayley, D. Gordon, G. Higgins,J. A. Trapani, M. I. McDonald, J. Labrooy, and R. Hecker.1985. Antibiotic-associated colitis due to Clostridium difficile:double-blind comparison of vancomycin with bacitracin. Gas-troenterology 89:1038-1045.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/c

mr

on 2

1 D

ecem

ber

2021

by

183.

162.

224.

24.

Documents

Clostridium difficile: Its Disease and Toxins - Clinical Microbiology