We hypothesize that humans are exposed to CD from food animals via consumption of foods of animal origin and test this by work under the following specific aims: <ol> <LI>Obtain isolates of CD from foods of animal origin in the southwestern US, to complement our extensive collection of food and food animal isolates. We propose a food sampling plan that will yield isolates of any genotype present as ? 3% of the population. <LI>Determine the genotype of each isolate and assess its relationship to food or human case type (CA- or HA-CDAD) of origin. We will determine ribotype, toxinotype, and PFGE type, and detect binary toxin gene cdtA and deletions in tcdC, characteristics that comprise the standard fingerprint of CD strains. Data will be analyzed to determine the association of the entire fingerprint and its individual components with specimen of origin.
Non-Technical Summary: Clostridium difficile-associated disease (CDAD) is traditionally a nosocomial infection, characterized by low mortality and a high attack rate in the elderly. Over time, at least two variant clones have emerged and spread worldwide. A toxin-defective clone not producing TcdA is of greater virulence than historical strains (which produce both major toxins), but a recent epidemic, or "hypervirulent," strain is of greatly-increased virulence; disease occurs in a much younger average population, with a more severe clinical course and much higher case fatality rate. CD is not part of the normal microbiota of humans (or domestic animals), in that it is not found uniformly in most members of the population. However, domestic animals have a relatively relaxed view of sanitation, and it is not difficult to envision maintenance of CD in a population by a combination of asymptomatic carriers and environmental contamination. Human colonization in healthcare facilities has often been related to persistent contamination of those environments due to active CDAD cases. Strains causing CA-CDAD (e.g., ribotype 078) must ultimately come from some yet-unidentified source. Microarray-based comparative phylogenomic analysis of human and food animal strains (see Introduction and Preliminary Data; Appendix) strongly suggests both close genetic relationships between food animal and human ribotype 078 strains and directionality of flow [i.e., pig strains appear to be genetic predecessors of human strains. It is in the public interest that this be monitored and pathways of dissemination traced. Furthermore, given the precedent set by emergence and transcontinental dissemination of ribotype 027 and toxin-variant strains, it seems likely that further highly-transmissible, hypervirulent strains will emerge. Ribotype 078 is a candidate threat, and if so, we need to be alert to minimize the human health impact. Finally, the importance of this work goes beyond the importance of CD alone; the results will contribute to generic understanding of other emerging infectious diseases. We propose to assemble a collection of isolates with representatives of the common genetic types in the population of food contaminants and of the main genotypes in humans with HA- or CA-CDAD. We have an archived collection of strains from piglets and calves. We hypothesize that analysis of the population structure and the characteristics of individual strains will reveal relationships by source. <P> Approach: SPECIFIC AIM1. Solid meats will be removed aseptically from packaging and placed in a sterile plastic bag containing 200 ml sterile BHI enrichment medium. The bag will be agitated manually for 3 min and the rinse solution decanted into a sterile 250 ml screw-capped Erlenmeyer flask. Flasks, with lids loosened, will be incubated at 37 degrees Celsius in the anaerobe chamber for 120 h, followed by subculture to TCCFA. The remaining enrichment broth will be centrifuged (10,000 x g, 20 min) and the pellet subcultured to TCCFA. Isolates will be identified as above. Bacteria and spores in milk (100 ml aliquots) will be concentrated by filtration (200 nm pore diameter). Filters will be placed in enrichment broth (50 ml amounts in 100 ml flasks) and incubated and subcultured as above. Multiple suspect colonies from each specimen will be confirmed as CD by detection of L-proline aminopeptidase activity (Remel), and by PCR assays for tcdA, tcdB, and a CD-specific gene (11). Strains will be stored in chopped meat medium at room temperature and in 25% glycerol at -80 degrees Celsius for further characterization. SPECIFIC AIM2. Ribotyping will be as described (79) and used extensively by us (34). Template will be prepared by boiling colonies for 10 min in 100 microliters of 5% Chelex 100 (Bio-Rad) and cell debris removed by centrifugation (10 min at 17,000 g). Each 100 microliters reaction mixture will contain 5 U of Taq polymerase (Promega), 1.5 mM MgCl2, 10 microliters template, and 100 pmol of each primer (RiboF and RiboR). PCR products separated by electrophoresis in a 3% Nusieve agarose gel (with 100 bp ladder at 5 lane intervals) will be photographed with UV transillumination and patterns compared to ribotype controls by use of GelCompar image analysis software (Applied Maths). Toxinotyping will be according to established methods (22). Restriction fragments will be separated by agarose gel electrophoresis and patterns compared to toxinotype controls (via GelCompar). Strains without PaLoc-specific fragments will be confirmed as nontoxigenic by PCR with Lok1 and Lok3 primers (22). Deletions in tcdC and presence of binary toxin gene cdtA will be detected by PCR. Pulsed field gel electrophoresis (PFGE) will be by a standard protocol. Data already in hand will be combined with data collected here, and then the groups (food and food animals) will be compared. Our study design will not definitively link food source and food animal, but the data will provide convincing evidence of strain dominance (expressed as prevalence) in various food niches (expressed as proportional morbidity).