C. difficile isolates 101 and 115 of equine origin were used for this study [34]. Isolate 115 was used to prepare the toxoid because this strain produces a high level of toxin. Isolate 101 was used to challenge the mice. Bacteria were cultivated in pre-reduced anaerobically sterilized peptone yeast extract broth with 1% glucose (Anaerobe Systems, USA) at 37℃ for 48 h under anaerobic conditions.

4 weeks old female C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, USA) for experiments. All mice used in the experiments were housed in groups of five per cage under the same environmental and husbandry conditions. The animals were fed a standard rodent diet (Harlan Laboratories, USA). Food, water, bedding, and cages were all autoclaved. Experimental procedures commenced after 1 week of receipt. All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee (IACUC) at Cornell University.

Sporulation of C. difficile was induced as previously described [35]. Briefly, an overnight culture of C. difficile was diluted in fresh brain heart infusion (BHI) broth (BD Diagnostics, USA) to an optical density measured at 600 nm (OD₆₀₀) of 0.2. Next, 150 µL of this suspension was spread on 5 mL BHI agar (BD Diagnostics) in each well of a six-well tissue culture dish (Corning, USA). The culture was incubated anaerobically at 37℃ for 7 days. Each well of the six-well dish containing mixed populations of spores and vegetative cells was flooded with ice-cold sterile water. The vegetative cell-spore mixture was collected, washed five times with ice-cold water, and heated to 60℃ for 30 min to kill vegetative cells. The resulting solution contains the finial spores. 1 mL of the finial spores was diluted from 1 : 10³ to 1 : 10⁹ with PBS, plated on BHI agar plates containing 0.1% taurocholate (Sigma, USA), incubated for 2 days and then the resulting colonies were counted. The spores were store at -80℃ until use.

A crude culture filtrate of C. difficile toxoid was used in this study. After C. difficile was cultured inside a dialysis tubing (Thermo Scientific, USA) containing BHI broth for 3 days at 37℃, the supernatant was harvested by centrifuging twice at 10,000 × g for 10 min at 4℃ and concentrated by ultrafiltration through a 50 kDa membrane (Thermo Scientific) using an Amicon concentrator (EMD Millipore, USA). The concentrate was re-suspended in 200ml PBS and concentrated three more times. The last volume of the concentrate was then lyophilized using Freeze Dry System (Labconco, USA) and resuspended in 5 mL PBS. The resulting re-suspended lyophilized culture supernatant was treated with 1% formaldehyde for 10 days at 37℃ and stored at 4℃ until use. Toxin inactivation was identified by the absence of cell culture cytotoxicity relative to treatment with the native toxin as a positive control which caused CACO-2 cells become round while the toxoid didn't.

Bacterin was prepared as previously reported [37]. Briefly, C. difficile was grown in BHI broth at 37℃ for 24 h under anaerobic conditions. The culture was centrifuged at 8,000 × g for 10 min at 4℃ and washed three times with PBS. The pellet was then re-suspended in 500 mL PBS containing 1% formaldehyde and incubated at 40℃ for 24 h. The excess formaldehyde was removed by washing with PBS three times. After the final wash and re-suspension in 100 mL PBS, the number of bacterial cells was determined at OD₅₅₀ using. The bacterin was cultivated in BHI broth at 37℃ for 24 h under anaerobic conditions to check for live organisms and then stored at -4℃ until use.

Two healthy adult draft horses (1 and 2) were immunized with bacterin and toxoid. Each of the preparations used for immunization were combined with either complete Freund's adjuvant (CFA; Sigma) or incomplete Freund's adjuvant (IFA; Sigma) at the minimum ratio (generally 1 : 1 [v : v]) to create a stable emulsion. The first two rounds of immunization with either bacterin or toxoid were mixed with CFA. All other immunizations with toxoid and bacterin were mixed with IFA. The dose and frequency of each antigen administration is outlined in Table 1. All immunizations were administered via an intradermal (ID) or subcutaneous (SQ) route at multiple sites with no less than 0.1~0.2 mL per site.

To measure the production of antibodies against the toxins, sera collected from the immunized horses were subjected to a Western blot analysis using recombinant toxin A and B fragments prepared in our lab (unpublished) as antigens. These fragments include the C-terminal of TcdA (residues 1820-2710, 127 kDa), the N-terminal of TcdA (residues 1-566, 89 kDa), the C-terminal of TcdB (residues 1820-2366, 89 kDa), and the N-terminal of TcdB (residues 1-566, 90 kDa). The methods used for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting have been previously described [40]. Briefly, 100 ng the recombinant toxin fragments were boiled for 5 min in SDS sample buffer (0.15 M Tris [pH 6.8], 1.2% SDS, 0.3% glycerol, 0.15% β-mercaptoethanol, 0.018% [w/v] bromophenol blue; Sigma) and loaded onto a 12% SDS-PAGE gel to separate the proteins. After electrophoresis, sera collected from the horses were used to detect the proteins transferred to nitrocellulose membranes (Schleicher and Schuell Biosciences, USA) using a semi-dry method [40]. Basically, the membrane was blocked using 5% blotting grade non-fat dry milk (Bio-Rad, USA) in 0.1 M 1× PBS for 60 min at 37℃. The proteins were then detected by incubation with antiserum collected from an immunized animal at a 1 : 200 dilution in blocking buffer at 37℃ for 60 min. The membrane was then incubated at 37℃ for 60 min with 2 µL mouse anti-horse IgG diluted 1 : 3, 000 in blocking buffer as the secondary antibody. Finally, proteins were visualized by Novex AP Chromogenic Substrate (BCIP/NBT; Life Technologies, NY) according to the manufacturer's instructions.

An ELISA was used to measure the levels of IgG against the toxins and bacterin. Blood samples were collected before (pre-immune) immunization as well as 11 and 17 weeks after the first immunization. Purified TcdA (List Biological Laboratories, USA), the C-terminal end of TcdB, or a suspension of formalin-killed C. difficile were used to coat 96-well microtiter plates (100 µL/well; Corning) at a concentration of 1 g/well in Bicarbonate/carbonate coating buffer (100 mM). The plates were then blocked with 1% bovine serum albumin (Sigma) in PBS at 37℃ for 2 h. The pre-immune and immune sera were diluted at 1 : 500 in PBS containing 1% bovine serum albumin and 0.05% Tween 20, added to the wells at 100ul/well, and incubated for 1 h at 37℃. IgG reactivity was detected with 1 : 2,500 diluted peroxidase-labeled anti-horse IgG (KPL, USA) for 1 h at 37℃ and then 100 µL of freshly prepared tetramethylbenzidine (TMB) 2-component microwell peroxidase substrate (KPL) were added to each well. Finally, 100 µL of TMB stop solution (KPL) was added to the wells and absorbance of the plates was read at OD450 with a microtiter plate reader (BioTek Instruments, USA).

To assess in vitro neutralizing activities of the prepared horse serum, a cytotoxicity inhibition assay was performed as previously reported [10]. Briefly, 200 ng of purified C. difficile TcdA in modified Eagle's minimum essential medium (Sigma) containing 2 mM L-glutamine, 0.1 mM nonessential amino acids, 1.5 g/L sodium bicarbonate, and 1.0 mM sodium pyruvate was mixed with an equal volume of prepared immune serum or pre-immune serum and incubated for 1 h at 37℃. Caco-2 cells (American Type Culture Collection, USA), which were grown in monolayers, were then incubated with 5 µL mixture at 37℃ for 18 h, and the cytopathic effects were analyzed under a phase-contrast microscope (Nikon, Japan). Untreated cells were used as the negative control and cells treated with 1ug purified C. difficile TcdA prepared by our lab (unpublished) were used as the positive control.

A mouse model of CDI was established as previously described [8]. Briefly, C57BL/6 mice were given with water mixed with a mixture of kanamycin (0.4 × 10^-3 mg/L; Sigma), gentamicin (0.035 × 10^-3 mg/L; Sigma), colistin (850 U/mL; Sigma), metronidazole (0.215 × 10^-3 mg/L; Sigma), and vancomycin (0.045 × 10^-3 mg/L; Sigma) for 3 days. All mice were then given regular autoclaved water for 2 days and received a single dose of clindamycin (10 mg/kg; Sigma) intraperitoneally 1 day before C. difficile challenge. After treatment, all mice were challenged with 10⁷ of prepared toxigenic C. difficile spores by gavage feeding. The animals were observed daily after challenge for the presence and severity of diarrhea and other general symptoms of illness or mortality. Body weight was measured once a day at the same time each day. Mice judged to be in a moribund state were euthanized by CO₂. Tissue samples including intestine, live and kidney were collected for histopathologic analysis after euthanasia.

To confirm C. difficile as the causative agent of disease, perianal swabs were taken randomly 2 days after challenge from representative symptomatic mice and cultured anaerobically at 37℃ for 2 days on BHI agar plates containing 5 × 10⁵ mg/L clindamycin. Stool specimens were also collected 2 days after challenge and tested using a C. difficile toxin A/B microplate assay (Remel; Thermo Fisher Scientific, USA) to confirm the presence of toxins A and B. In addition, tissue samples including intestine, live and kidney were taken from representative symptomatic mice, fixed in 10% neutral buffered formalin, sectioned (4-µm-thick) by tissue section machine (Leica, German), stained with hematoxylin and eosin (Sigma), and examined by light microscopy (Nikon, Japan) to confirm the presence of typical CDI-associated lesions.

Normal pre-immune or immune serum from the horses was administered intravenously into the tail vein of different groups of mice 1 day before C. difficile challenge. For experiment 1, the protective properties of antisera prepared from the two different horses (1 and 2) named serum 1 and serum 2, respectively, were tested in the mouse model. Mice receiving the immune serum were divided into two groups of five mice each. One group of mice was treated with 100ul immune serum prepared from horse 1 while another was given 100ul immune serum prepared from horse 2. One group of five mice was given normal pre-immune serum (100 µL) from horse 1 alone as a negative control.

For experiment 2, the efficacy of serum from horse 1 was evaluated at two different doses. Two groups of five mice received either 50 µL or 100 µL of the hyperimmune serum prepared from horse 1. Another group of five mice was treated with normal pre-immune serum (100 µL) from horse 1 alone as a negative control. Additionally, one more group of five mice was treated with antibiotics and not challenged to determine whether the antibiotic treatment caused diarrhea or weight loss. Experiment 2 was repeated in duplicate. All the other mice were challenged with C. difficile spores as described above (see "Mouse model of CDI").

Differences in mean relative weight loss between the experimental and control mice were evaluated with a two-tailed Student's t test using GraphPad Prism (GraphPad Software, USA). A p value < 0.05 was considered statistically significant.

No adverse effects associated with the immunization were observed in the horses. Western blot analysis of postimmunization serum samples from horse 1 showed that the antiserum produced by the immunized horse recognized all recombinant toxin A and B fragments (Fig. 1).

Immune serum samples were collected 11 and 17 weeks after the first immunization from each horse and evaluated with an ELISA. The results showed that the OD values for different sera against purified toxins A and B as well as whole-cell C. difficile continually increased. At the end of immunization protocol, the OD values ranged from 1043 to 1246 for TcdA, 1313 to 1417 for TcdB, and 1730 to 1815 for the whole-cell bacterin. However, the OD values for pre-immune serum were only 65 to 176 for TcdA, 167 to 246 for TcdB, and 534 to 596 for C. difficile bacterin (Figs. 2A-C).

The ability of the prepared serum to neutralize C. difficile toxin was assessed with a cytotoxicity assay. Cytotoxicity was observed by rounding of Caco-2 cells grown in monolayers after exposure to C. difficile toxins. Controls for this assay included cells treated with the pre-immune serum that lacked specific antibodies against C. difficile toxin A or B as determined by an ELISA. Results of the experiment showed that cells exposed to TcdA after being pretreated with normal pre-immune horse serum became rounded after 24 h of incubation (Fig. 3A). In contrast, pretreatment with antitoxin serum prior to TcdA exposure inhibited cytotoxicity (Fig. 3B). The cells treated with purified TcdA only displayed typical signs of cytotoxicity while the untreated cells showed no signs of toxicity (data not shown).

In experiment 1, the protective effects of antisera prepared from sera of two different horses (serum 1 and 2) were tested in the mouse model. Both serum 1 and 2 given by intravenous injection 1 day before C. difficile challenge protected the mice against severe clinical disease. The group that only received pre-immune serum lost more than 12% body weight on day 3 post-challenge and all animals in this group were visibly sick with diarrhea. In contrast, the mice that received prophylactic antitoxin serum only lost 2% body weight without any signs of illness (Fig. 4A). On day 3, the pre-immune serum control group had lost a significantly greater amount of weight compared to mice treated with serum 1 or 2 (p < 0.05).

After the preliminary test in experiment 1, the efficacy of serum 1 delivered at two different doses was evaluated in experiment 2. All animals challenged with C. difficile lost weight. The loss of body weight reached a maximum at day 2 and gradually recovered thereafter in all mice. By day 2, animals in the groups that received normal pre-immune serum had lost the most weight (> 10%). The animals that received antitoxin serum lost significantly less weight (p < 0.05) compared to the group that received pre-immune serum alone. Furthermore, the level of protection was dose-dependent. Mice receiving 100 µL of antiserum lost a minimum amount of weight (3.0%) compared to the animals treated with pre-immune serum (p < 0.01) after C. difficile challenge without developing diarrhea. The body weight of these mice also returned to a normal range first (between days 4 and 5 after infection). The group that received 50 µL of antiserum lost more weight (3.5%) compared to mice treated with 100 µL of antiserum. However, this weight loss was still significantly less (p < 0.01) compared to the mice treated with pre-immune serum. Furthermore, it took 2 more days for the body weight of animals treated with 50 µL of the antiserum to return to the normal range than the 100 µL-treated group. By day 7 after infection, the weight of all animals intravenously administered the hyperimmune serum returned to a normal range while the group that received pre-immune serum alone still had a 1.9% weight loss. However, diarrhea in all animals belonging to both the control and antiserum-treated groups had resolved at the time of euthanasia (7 days post-infection). No typical gross or microscopic lesions associated with C. difficile enteritis were observed at the end of experiment (data not shown). The control group that was not challenged with C. difficile consistently maintained their body weight within normal ranges and did not show any sign of disease (Fig. 4B).

Cultures of perianal swabs from representative symptomatic mice produced large, slightly flat, circular colonies with a filamentous edge. Under the microscope, long, thin, straight Gram-positive rods with typical C. difficile morphology were observed. Additionally, the fecal specimens were positive in the C. difficile toxin A/B microplate assay that confirmed the presence of toxins A and B (data not shown). Histopathological results (Fig. 5) for representative symptomatic mice from experiment 1 showed that these animals suffered from C. difficile-mediated enterocyte apoptosis and/or necrosis with villous atrophy and inflammation based on characteristic changes in the intestine. The presence of microvesicular hepatic lipidosis indicated either acute liver injury or rapid mobilization of peripheral fat stores. Changes in the spleen likely reflected breakdown of the intestinal mucosal barrier and delivery of antigens to the white pulp with early secondary follicle formation. The apparent depletion of zymogen granules in the pancreas may have been reflecting atrophy secondary to anorexia.