|Home | About | Journals | Submit | Contact Us | Français|
We have examined the piglet model for CDI in humans since swine are naturally susceptible to C. difficile. The piglet is a reproducible model for acute or chronic CDI with characteristic pseudomembranous colitis (PMC). Germfree piglets were consistently and extensively colonized when orally challenged with the human 027/BI/NAP1 strain, establishing an infectious dose-age relationship. This allowed a demarkation between acute fatal and chronic models. The clinical manifestations of disease inclusive of gastrointestinal and systemic symptoms and characteristic mucosal lesions in the large bowel including PMC are described. Additionally, we demonstrate the presence of toxins in feces, body fluids and serum, and a significant elevation of IL-8 in animals with severe disease. We conclude that piglets infected with C. difficile mimic many of the key characteristics observed in humans with CDI, and are suitable animals in which to investigate the role of virulence attributes, drug efficacy and evaluation of vaccine candidates.
Clostridium difficile is a gram positive, anaerobic, spore forming bacterium and a major cause of antibiotic-associated diarrhea in many countries worldwide [1–3]. It is the etiologic agent of pseudomembranous colitis in humans, but infection can result in a range of sequelae, from asymptomatic carriage to toxic megacolon and death. C. difficile illness (CDI) has reached epidemic proportions in several countries since the year 2000, and emerging hypervirulent strains are causing increased morbidity and mortality among patients. One group of these hypervirulent strains has been characterized as ribotype 027, restriction enzyme analysis type BI, North American Pulsed Field Type 1 (027/BI/NAP1). These strains produce the three known toxins: toxin A (TcdA), toxin B (TcdB) and binary toxin (CDT). TcdA and TcdB are known to be important virulence factors, affecting the intestinal epithelial cells directly and promoting inflammatory reactions which lead to the recognized signs of disease.
In addition to gastrointestinal pathology, systemic complications of infection such as ascites[6, 7], pleural effusion[8, 9], cardiopulmonary arrest[10, 11], hepatic abscess, abdominal compartment syndrome, acute respiratory distress syndrome, multiple organ dysfunction syndrome, and renal failure have been reported in human cases. The mechanisms by which C. difficile causes these systemic effects are not entirely understood, but the toxins produced by the bacterium, especially toxins A and B, are likely involved. A greater understanding of the systemic effects of C. difficile infection and why they occur in some patients, but not others, is important because these effects are often life-threatening in nature.
Many species have been evaluated as models for CDI, but the hamster has been the classic model because of extreme sensitivity to infection following antibiotic administration. Hamsters develop clinical signs of severe diarrhea, weakness and lethargy, and death usually occurs within 2–3 days of infection. Other laboratory animals such as mice, rats and rabbits have also been used, but are not as sensitive to infection as hamsters [5, 17]. While hamsters do provide a valuable model of acute CDI, the model does have limitations. Few commercial assays and immune reagents are available for them, and their extreme sensitivity precludes studies on many of the clinical and pathological conditions observed in humans with CDI.
C. difficile infection commonly occurs in swine, and in piglets it causes enteritis during the first week of life[18–20]. C. difficile outbreaks on swine farms usually include pasty yellowish diarrhea, sometimes with respiratory distress and death, and C. difficile has become the most commonly diagnosed cause of enteritis in neonatal pigs. CDI has been reproduced in pigs inoculated with pure cultures, however questions still remain regarding pathogenesis, immune response to infection, and treatment and prevention strategies. The similarities to human disease and availability of reagents make pigs an attractive model for C. difficile studies, and here we describe the development and characterization of the gnotobiotic piglet as a model of acute or chronic CDI.
Gnotobiotic piglets, derived by Cesarean section, were housed inside sterile isolators and fed Similac milk replacer three times daily . Thirty-five piglets derived from 9 litters were divided into 8 uneven groups and inoculated as summarized in Table 1. Eleven piglets from one litter were used for evaluation of the relationship between systemic manifestations of disease and toxemia. Two piglets from this litter were inoculated with a nontoxigenic C. difficile strain as controls, and the remaining 9 were inoculated with 1 × 105 spores of a toxigenic strain.
Fecal samples were collected daily for the duration of each experiment. After inoculation, piglets were monitored for signs of disease including diarrhea, dehydration, dyspnea, weakness, lethargy or anorexia. Piglets were euthanized at a pre-determined experimental endpoint at post-inoculation day (PID) 15 or 21 or sooner if they displayed severe symptoms such as weakness, lethargy, or anorexia. Blood was collected following deep sedation before euthanasia. Gross gastrointestinal and systemic lesions were noted during necropsy, and tissues were collected for histologic examination. Tissue sections were collected from the duodenum, jejunum, ileum, cecum, colon, mesenteric lymph nodes, pancreas, spleen, kidneys, liver, and lungs and fixed in formalin. If present, pleural effusion and ascites were also collected using aseptic technique. This study received IACUC approval.
The nontoxigenic strain CD37 was used to inoculate the two control piglets. For all other animals C. difficile strain UK6, a type 027/BI21/NAP1, which produces tcdA, tcdB, and binary toxins was used. Vegetative cells for inocula were grown anaerobically overnight in pre-reduced brain heart infusion (BHI) broth at 37 °C. Concentration was adjusted to contain 108 CFU per 2 ml per piglet.
Spores were grown on pre-reduced BHI agar plates anaerobically at 37 °C for 48 hours. Colonies, scraped off the plates, were suspended in BHI broth, left in flasks for 7–10 days in an anaerobic chamber at 37 °C to induce sporulation. The suspension was centrifuged, supernatant discarded, and washed with sterile PBS twice. The suspension was then heated at 70 °C for 20 minutes to kill vegetative cells. The spore suspension was stored at 4 °C, and spore concentration was determined by serial dilution before each experiment.
Daily fecal and necropsy samples from the gut, blood, pleural effusion and ascites were cultured for bacterial growth immediately following necropsy. Samples were streaked on C. difficile TCCFA selective media (taurocholate-cefoxitin-cycloserine-fructose agar) plates, and incubated anaerobically at 37 °C for 48 hours. C. difficile was confirmed using Pro Disk test (Remel). Samples were also streaked on MacConkey agar plates and incubated aerobically at 37 °C for 48 hours to determine presence of contaminant bacteria.
The presence of C. difficile toxins in samples was measured using murine macrophage RAW264.7 cells incubated overnight in a 96 well plate before addition of samples or recombinant toxins , and then incubated overnight before evaluation for cell rounding. Samples tested for cytotoxicity included feces, serum, pleural effusion, and ascites. In some cases, the standard assay was inadequately sensitive to detect toxin in sera, and consequently we used an ultrasensitive assay recently developed in this lab  for the detection of toxin in serum and other body fluid. In this assay, the mRG1-1 cell line, which expresses the FCγR1-α-chain, was used . The cells were incubated overnight in 96 well plates before sample or toxin addition. Samples were passed through a 0.45μm syringe filter before addition to cell culture in serial dilutions. In addition to sample dilutions alone, samples were also mixed with a saturating dose of A1H3 antibody before addition to cell culture. A1H3 is a mouse anti-toxin A monoclonal antibody of IgG2a isotype generated by our laboratory. A1H3 increases the sensitivity of cells to TcdA , allowing detection of low concentrations of TcdA in samples. Goat antiserum against toxins A and B was used for blocking toxin activity in the assays (TechLab, Inc). Recombinant TcdA was used as a positive control. After addition of samples or toxins, the cells were incubated overnight at 37 °C before evaluation for cell rounding.
Intestinal contents were collected from the small intestine (jejunum) and large intestine (cecum and colon) to determine regional bacterial counts. Contents were collected during necropsy, and serial 10-fold dilutions were plated in quadrants on BHIS (BHI + 5 g/L yeast extract + 0.1% L-cysteine) + 0.1% taurocholate agar plates or TCCFA plates and grown anaerobically for 24–48 hours at 37C to determine bacterial counts.
Cytokine concentration was determined in the large intestinal contents for IL-1β, IL-4, IL-6, IL-8, IL-10, IL-12, TNF-α, TGF-β, and IFN-γ using commercially available porcine cytokine ELISA kits (Invitrogen and R&D). Samples were stored at −20 °C until use. Contents were diluted 1:2 to 1:10 with sterile PBS, depending on the consistency of the sample, thoroughly mixed using a vortex, then centrifuged, and the supernatant was added to reagent wells in the assay. The assay was performed following the manufacturer’s instructions, and cytokine concentration determined based on the standard curve generated from absorbance measured at 450nm. Statistical analysis of cytokine measurements among piglet groups was performed with SPSS version 16.0.
Piglets inoculated with the nontoxigenic strain displayed no signs of disease, having normal feces throughout the experiment. Regardless of dose, all piglets inoculated with the toxigenic strain developed diarrhea within 48 hours of inoculation. Diarrhea typically progressed from yellow-brown and pasty to yellow and watery, typical of CDI. Overall severity of disease depended upon the type (vegetative or spore forms), dose, and age at inoculation. Higher doses of spores given to younger piglets generally produced more severe and often fatal disease. By altering the dose and age, we were able to induce in piglets either acute and severe systemic disease, or milder chronic gastrointestinal disease (Table 1).
The most severe disease occurred in piglets inoculated with 108 vegetative cells, 24 hours after birth, displaying serious signs of dyspnea. Severely affected piglets began to show signs of weakness, lethargy and anorexia from PID 3-7, and quickly progressed to near death within 24 hours, when they were euthanized. In the groups receiving the lowest inoculum dose of 105 spores, 8/14 piglets (57%) developed a clinically mild to moderate chronic diarrhea, lasting until the experiment was terminated at PID 15. One piglet inoculated with 108 vegetative cells at 5 days of age also developed chronic diarrhea lasting until the end of the experiment PID 21.
Piglets inoculated with the nontoxigenic strain had no gross gastrointestinal or systemic lesions (Fig 1-a). Piglets with chronic diarrhea had mild to moderate mesocolonic edema, and inflammation and dilation of the large intestine, mostly confined to the spiral colon (Fig 1-b). These piglets had no lesions of the small intestine or apparent systemic lesions. Piglets with severe, acute disease had more profound large intestinal lesions including extensive mesocolonic edema, severe dilation and inflammation of the large intestine, pseudomembrane and colonic mucosal hemorrhages, extending from the ileocecocolic junction to the rectum (Fig 1-c and d). Additional gastrointestinal lesions in the most severely affected piglets were profound thickening of the wall of the descending colon and rectum (Fig 1-d), and one piglet had a perforation of the spiral colon. In some cases, gross lesions of the gastrointestinal tract from the acutely affected piglets appeared similar to those from the chronically affected animals. However, those with severe clinical disease also developed the extra-intestinal lesions of ascites, pleural effusion, and cranial ventral lung consolidation, which we attribute to the toxins since no bacteria were ever cultured from any of these sites.
Piglets inoculated with the nontoxigenic strain had no microscopic gastrointestinal or systemic lesions (Fig 2-a and b). Histologic examination provided the best way to fully differentiate the effects of infection on the gastrointestinal tract in the chronically versus acutely affected piglets. Those which were chronically affected had extensive submucosal and mesenteric edema, but had only mild focal neutrophilic inflammation and mucosal erosions (Fig 2-c and d). These piglets had no pulmonary lesions, and some had nonspecific vacuolar hydropic change in the liver. The piglets with acute, critical symptoms had extensive and severe large intestinal lesions. Severe typhlocolitis was present with massive neutrophilic inflammation in the mucosa and submucosa (Fig 2-e), and extensive submucosal and mesenteric edema were present from the ileocecocolic junction to the rectum. Mucosal lesions ranged from severe erosions and ulcerations to nearly complete loss of mucosal lining with exposed submucosa in the most severe cases (Fig 2-f). The colonic lumen was filled with a combination of neutrophils, bacteria, and necrotic debris, which formed a pseudomembrane over the surface of the mucosa in many areas of the colon and cecum (Fig 2-f and g). In the most severe cases, neutrophilic infiltration was present in the mesenteric lymph nodes. Severely affected piglets also had systemic lesions in the lungs consisting of regional atelectasis, occasional macrophage infiltrate, alveolitis and interstitial thickening (Fig 2-h). No evidence of pneumonia, such as neutrophilic infiltrate or bacteria, was noted in any of the piglet lung sections. Some severely affected piglets also had nonspecific lesions including vacuolar hydropic changes in the liver and reduced periarteriolar lymphoid sheath diameter in the spleen.
All fecal cultures collected daily, including the control, excreted the respective strain between 24 and 48 hours after oral challenge. There was no bacterial growth on any of the aerobic fecal cultures on MacConkey agar, indicating the absence of contaminants. Cultured pleural effusion, ascites or blood on either TCCFA or MacConkey agar yielded no bacterial growth. The blood culture from one piglet was positive on TCCFA, with growth of 3 colonies, which is suspected to be due to skin contamination.
Feces typically became positive for presence of toxin 24 hours after the first positive fecal culture. All of the pleural effusion and ascites samples from the severely affected piglets were positive on the standard cytotoxicity assay. Several of the serum samples were also positive, and those which were negative using the standard assay were positive on the ultrasensitive assay described above. None of the serum samples from the mildly affected piglets or controls caused any degree of cell rounding on either the standard or ultrasensitive assay.
Bacterial counts in the large intestine were greater than in the small intestine for all animals, but there was no significant difference in the numbers between groups of piglets based on the size of the inoculum and severity of disease (Table 2). For the control piglets, small intestinal counts ranged from 104–105 CFU/ml, and large intestinal counts were 1012 CFU/ml. Piglets which developed a chronic course of disease had small intestinal counts ranging from no growth to 107 CFU/ml, and large intestinal counts ranging from 107 to 1012 CFU/ml. Piglets which developed an acute, severe course of disease had small intestinal counts ranging from no growth to 1011 CFU/ml, and large intestinal counts ranging from 107 to 1012 CFU/ml.
The cytokine measurements for all piglets were organized into groups based on severity of disease: acute, chronic, or control. The mean concentration was compared using the Kruskall Wallis test to evaluate significant differences between groups for each cytokine. Only IL-8 achieved a statistical level of significance (p=0.036), with acutely affected piglets having significantly greater levels than either chronically affected or control piglets (Fig 3). There was a similar trend but statistically not significant with regard to TNF-α.
We describe the response of gnotobiotic piglets orally challenged with C. difficile in which a spectrum of clinical symptoms and pathological abnormalities largely depending on age of the animal and the size of the infectious dose, were induced. The nature and outcome of the disease in these animals mimic many of the characteristics observed in human patients with CDI. Using the hypervirulent strain 027/BI/NAP1, this animal model offers reproducible results, with 100% colonization within 48 hours of inoculation, and a disease severity that can be tailored according to need. A disease, ranging from profoundly acute and lethal to chronic diarrhea, was readily induced under controlled laboratory settings. The range of clinical signs, including systemic complications, are similar to those observed in human cases [1, 6–8, 11–13, 15, 26], making gnotobiotic piglets an attractive alternative model to the hyperacute hamster model. The piglet model offers flexibility to meet specific research requirements such as studies on pathogenesis, evaluation of virulence attributes, testing the efficacy of therapeutics, and the evaluation of vaccine candidates for eliminating existing infections or protecting against infections. For this purpose, animals can be immunized orally, intranasally, or systemically at one-week of age followed by repeated boosting at ~2 weeks thereafter. Immunized animals can then be monitored for side effects/symptoms in the case of live attenuated vaccines. Sera and feces can be analyzed for bacterial excretion and toxins, protective mucosal and systemic antibodies, T cell responses, cytokine responses, etc. After the immunization, animals can then be challenged orally with wild type strains. To test vaccine candidates the model offers several important parameters for measurements before and after challenge including clinical symptoms, degree of mucosal injury, if any, extent of bacterial colonization in the gut, level of toxin production, cytokine responses, etc. Essentially, the model can provide most of the required information for preclinical evaluation of vaccine candidates.
Though pigs are naturally susceptible to C. difficile infection and conventional pigs could be used, gnotobiotic piglets were chosen for the models due to several advantages over conventional animals. Gnotobiotic piglets, delivered via Cesarean section, do not nurse their dam, therefore absent are interfering maternal antibodies against C. difficile in studies involving evaluation of the immune response. They lack normal as well as potentially pathogenic gut flora and consequently do not require starvation and treatment with antibiotics before inoculation to enhance susceptibility, as is required with other animals. The absence of normal gut microflora as a consequence of prolonged antibiotic treatment is the hallmark of CDI, and is considered as one of the most important risk factors for this disease[1, 2, 4]. Gnotobiotic piglets mimic this state without a need for antibiotic pre-conditioning. Since gnotobiotic piglets are maintained in sterile isolators for the duration of the study, the possibility for introduction of other pathogens from the sow or human caretakers, including C. difficile which is common in swine, is also eliminated. Because of these traits, inoculation of gnotobiotic piglets produces very consistent and reliable results, which can be modified according to need by manipulating the dose and/or the age of the animal. This is particularly useful in the study of vaccine candidates where several immunizations, with monitoring of immune and clinical responses followed by challenge can be accomplished independently on each animal.
The major objective of this study was to establish a dose and age relationship for the model with the 027/BI/NAP1 strain. In general, younger piglets inoculated with higher doses experience more severe clinical signs of disease. In addition to the range of clinical signs induced by varying the dose and age at inoculation, we also observed that differences in disease severity could be observed even between individuals from the same litter given the same dose at the same age. For the strain used in these experiments, a dose of 105 spores given at 5 days of age induced acute disease in approximately half of the piglets and chronic disease in the other half. Those experiencing acute disease developed systemic lesions of ascites, pleural effusion and lung consolidation, while those with chronic disease developed only gastrointestinal lesions.
The finding that the severity of the disease varies among the piglets resembles the situation in human cases. Humans too develop a range of systemic consequences of C. difficile infection such as ascites, pleural effusion, cardiopulmonary arrest, liver abscess, and multiple organ dysfunction syndrome[6–12, 15], which result in severe and even fatal disease. The reasons for the case differences observed among human patients, as well as those in the piglets in this study, are not well understood. Hopefully, the piglet model will prove to be a useful tool to delineate the relative role of each virulence attribute in contributing to the systemic and to the gastrointestinal disease observed in humans. Immune response is likely to play a deciding role in disease severity, and in this study, we analyzed cytokine levels in the large intestinal contents. IL-8 concentration was significantly elevated in the piglets which developed acute, severe disease compared to the chronically affected or control piglets. IL-8 is a component of the inflammatory response, mediating neutrophil migration, and is also elevated using in vitro experiments with human cells[27, 28], and in feces of human patients. Our findings suggest that IL-8 may be a detrimental component of the inflammatory response, as significantly elevated levels of this cytokine correlate with more severe disease.
This study supports the hypothesis that C. difficile toxins, rather than the bacteria, are responsible for the systemic complications we have observed. The presence of toxin in the serum, pleural effusion and ascites of the systemically affected piglets demonstrates the ability of toxins to reach circulation and to disseminate to extra-intestinal sites and suggests that they play a systemic role, which hopefully the piglet model will help address more precisely in the future. We were unable to culture bacteria from body fluids or sera of severely affected animals, and no bacteria were noted on histologic examination of tissues outside the gastrointestinal tract, indicating that dissemination of bacteria through damaged gut mucosa was not responsible for the systemic effects. The finding of toxin in the serum of systemically affected piglets, but not in the mildly affected or control piglets is especially important, because to our knowledge, toxemia has not been previously documented in human or animal cases. While toxin concentration in feces of infected individuals is quite high and may cause cultured cell rounding in a matter of hours, the concentration in body fluids is much lower and requires a more sensitive assay for detection. The ultrasensitive cytotoxicity assay developed at our laboratory  has increased considerably the ease and speed of toxin detection in body fluids, especially in the serum.
We would like to give special thanks to our animal care technicians, Patricia Boucher and Catherine Lane and to Don Girouard for their many hours of help with this study.
This project has been funded in whole with Federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, under Contract Number N01-AI-30050.
The authors of this manuscript do not have a commercial or other association that may pose a conflict of interest.
Part of this information was previously presented at the Diagnosis and Treatment of Clostridium difficile Infection conference in Bethesda, MD, November 17, 2008, sponsored by the NIAID, the CDC and Tufts University.