Recipient Organization
COLORADO STATE UNIVERSITY
(N/A)
FORT COLLINS,CO 80523
Performing Department
Clinical Science
Non Technical Summary
Mitigating Salmonella colonization is one way of reducing the risk for colitis in adult horses, but doing so requires a greater understanding of the niche requirements for Salmonella colonization in the colon, the microbiota changes within an equine host with colitis, and horse factors significantly associated with microbial diversity. To achieve this, we propose to characterize the fecal microbiome in Salmonella shedding and non-shedding horses with and without colitis using 16S rRNA sequencing. Fecal samples will be collected from healthy horses and horses with colitis and subjected to culture and PCR to determine presence or absence of Salmonella. Total DNA will be extracted for 16S rRNA sequencing. Metagenomic sequence data will be analyzed and microbial abundances and diversity will be compared between horse groups. A final modeling step will identify horse factors that are risk factors for lower microbial diversity as it relates to Salmonella shedding and colitis. We hypothesize that when Salmonella is present in horses with colitis, the richness of fecal microbiota will be significantly lower compared to both shedding and non-shedding horses without colitis. Horses with colitis and not shedding Salmonella will also have reductions in diversity and shifts in the relative proportions of different microbial taxa, when compared to non-shedding horses without colitis. Additionally, in non-shedding horses, we hypothesize that there will be a greater abundance of microbial genera shown to inhibit or compete with Salmonella growth, and this abundance will be greatest in healthy horses compared to horses with colitis. This work is critical to establishing techniques to suppress Salmonella colonization, and ultimately reduce the risk for colitis and promote the welfare and gastrointestinal health of horses.
Animal Health Component
25%
Research Effort Categories
Basic
50%
Applied
25%
Developmental
25%
Goals / Objectives
Mitigating Salmonella colonization is one way of reducing the risk for colitis in adult horses, but doing so requires a greater understanding of the equine microbiota as it relates to both Salmonella shedding and colitis. The overarching goal of this work is to better understand the niche requirements for Salmonella colonization in the colon and the microbiota changes within an equine host with colitis. To achieve this, we propose to characterize the fecal microbiome in Salmonella shedding and non-shedding horses with and without colitis using 16S ribosomal ribonucleic acid (rRNA) sequencing. This characterization of the fecal microbiota is critical to establishing techniques to suppress Salmonella colonization, and ultimately promote the welfare and gastrointestinal health of horses.We hypothesize that the microbial community will differ between Salmonella-shedding horses with colitis compared to non-shedding horses without colitis. When Salmonella is present in horses with colitis, the richness of fecal microbiota will be significantly lower compared to both shedding and non-shedding horses without colitis. Horses with colitis and not shedding Salmonella will also have reductions in diversity and shifts in the proportions of different microbial taxa, when compared to non-shedding horses without colitis. In non-shedding horses, we hypothesize that there will be a greater abundance of microbial genera shown to inhibit or compete with Salmonella growth in vitro, and this abundance will be greatest in healthy horses compared to horses with colitis. As a way of testing these hypotheses, the following objectives will be pursued:Objective 1: Characterize fecal samples of horses that are either asymptomatic or have clinical evidence of colitis for the presence of Salmonella and the composition of the microbial community. We will collect admission fecal samples from horses at the Colorado State University (CSU) Veterinary Teaching Hospital (VTH) for reasons unrelated to disease of the gastrointestinal tract, and from horses with clinical colitis. Samples will be initially defined by the presence or absence of Salmonella using culture and polymerase chain reaction (PCR). Total DNA will be extracted from each fecal sample and will undergo 16S rRNA sequencing.Objective 2: Calculate differential abundance of bacterial taxa among shedders/non-shedders and diseased and non-diseased, and assess the influence of metadata on microbial phylogenetic shifts and diversity. A questionnaire regarding husbandry practices will be administered to horse owners. Other pertinent history and information about clinical findings and treatment during hospitalization will be obtained from individual case records. Metagenomic sequence data will be analyzed and abundances of families, genera, and species will be compared between groups and metadata parameters. The four groups to be compared include healthy horses shedding Salmonella, healthy non-shedders, diseased shedders, and diseased non-shedders, where healthy horse = no colitis and diseased = colitis. Further comparisons between groups will be made for specific bacterial genera known to competitively inhibit Salmonella growth.The long-term goal of this work is to improve gastrointestinal health in adult horses. We hope to uncover potential treatment targets and management practices that may manipulate the equine gastrointestinal microbiome and mitigate Salmonella presence to reduce the risk of colitis, and improve the overall health of the horse.
Project Methods
Objective 1: Characterize fecal samples of horses that are asymptomatic or have clinical evidence of colitis for the presence of Salmonella and the composition of the microbial community. Experimental Approach. We will collect admission fecal samples from horses (2-15 years old) at the CSU VTH. Our case definition for controls will include horses presenting for elective procedures with no history of GI disease or systemic inflammation. The case definition for colitis will include horses presenting with diarrhea and one or more of the following suspect criteria: fever, leukopenia, and ultrasonographic evidence of colon inflammation. Any horse currently on or receiving antimicrobials in the past 30 days will be excluded from the study. The presence/absence of Salmonella in fecal samples collected from cases and controls will be determined using culture and PCR, and total DNA will be extracted from each fecal sample for 16S rRNA sequencing.Experimental Design and Methods. Sample collection: From August, 2018 through February, 2019, feces from cases and controls will be gathered from the stall floor immediately following defecation, and split into three equal samples. A simple questionnaire will be administered to horse owners to collect metadata information regarding diet, housing, animal use, and exercise routine (see attached questionnaire in Addendum B). Other pertinent history and information about clinical findings and treatment during hospitalization will be obtained from individual case records. From the three fecal aliquots, one will be submitted to the CSU Veterinary Diagnostic Laboratory (VDL) for culture as part of the VTH's Salmonella biosecurity screening procedure. The other two will be submitted to our laboratory where one will be immediately frozen at -80°C for future DNA extraction. Salmonella detection: The second fecal aliquot will be pre-enriched in buffered peptone water for 24 hours (h) at 37°C, followed by selective enrichment and incubation at 37°C for 24 h. Enriched cultures will then be frozen at -80°C for future PCR testing. Salmonella presence will be assessed through PCR using selective primers for the ompC gene, specific for over 40 serovars.3 DNA extraction and 16S rRNA sequencing: Prior to DNA extraction, a sedimentation step will be used on the third fecal sample prior to DNA extraction to allow us to process a greater volume and to remove additional PCR inhibitors for a more complete representation of bacterial DNA presence.25 According to the manufacture's recommendations and our previous work,25 the MoBio PowerMax Soil DNA Isolation Kit (Mo Bio Laboratories, Inc.) will be used to extract DNA from 10 g/fecal sample. DNA library preparation and 16S rRNA sequencing will be performed following the Earth Microbiome Project suggested protocols19, earthmicrobiome.org by Novogene (Chula Vista, CA). Briefly, using the Illumina HiSeq (Illumina, Inc.) instrument, paired-end sequencing to amplify the V4 hypervariable region of the 16S rRNA gene, with defined barcodes, Illumina flowcell adapter sequences, and the 600 cycle version3 reagent cartridge (Illumina, Inc.) will be performed.7,33 Sequencing will be done at a minimum depth of 100,000 single-end reads per sample. Additionally, PCR conditions, library quantification, integrity, clean up, and pooling to achieve equal representation will be performed as previously described19, earthmicrobiome.org. Quality score and error rate will be controlled for with the inclusion of the internal control, PhiX. Sample size: A sample size of n=30 fecal samples (n=10 positive and n=20 negative for Salmonella) was calculated with 80% power to detect a 75% differential relative abundance in the 2 most abundant taxa between positive/negative fecal shedders, and 50% in the next 2 abundant taxa with 0.05% significance. This was calculated using the Power & Sample Sizes Tool for Case-Control Microbiome Studies Shiny application.34Objective 2: Calculate differential abundance of bacterial taxa among shedders/non-shedders and diseased and non-diseased, and assess the influence of metadata on microbial phylogenetic shifts and diversity.Experimental Approach: This will be accomplished by assessing the differential abundance of taxa or OTUs within families, genera, and species, as compared between groups (healthy horses shedding Salmonella, healthy non-shedders, diseased shedders, diseased non-shedders) while controlling for factors previously shown to be associated with Salmonella shedding. Further comparisons between groups will be made for specific bacterial genera known to competitively inhibit Salmonella growth. To summarize, differential microbial abundances and diversity values among the four distinct groups will first be analyzed to tease out the role of microbiome taxa in the +Salmonella/+colitis status. If determined that a particular microbial species is associated with +Salmonella/+colitis, linear regression modeling will be used to further identify associations (odds ratios) between metadata (horse factors) and the count measurement of that particular microbial species. This modeling approach will also be applied using sample diversity values. These final modeling steps will identify horse factors that are risk variables for influential microbial taxa and different values of total community diversity.Experimental Design and Methods: Pre-processing: Sequence read quality with a maximum barcode error of 0 and pre-processing into organizational taxonomic units (OTUs) will be analyzed using the Quantitative Insights Into Microbial Ecology (QIIME) software package,6,12,13,23,35 Statistics and Epidemiologic modeling: Salmonella shedding and microbial community data will be analyzed to determine 1) if microbial diversity correlates with sample presence of Salmonella, 2) which bacterial taxa are associated with presence/absence of Salmonella, 3) if these taxa change in abundance between horses with and without colitis, and 4) associations between horse metadata, microbial diversity, and specific influential taxa. QIIME and R statistical software will be used to determine species present, abundances of species, and overall diversity of species in each sample.24,26 Using R, two different statistical approaches will be used to analyze data from microbial counts: 1) measurement of taxa differential abundance and 2) epidemiologic modeling of microbial alpha diversity and specific taxa counts. Briefly, alpha diversity metrics quantified will include richness, Shannon's index, and Pielou's measure of species evenness.24,26 Data will be normalized using cumulative sum scaling to correct for sequencing depth bias.27 Multinomial regression analysis will be used to determine associations between microbial diversity with colitis and Salmonella shedding (p<0.05). Separate horse-level covariates that may modify the risk of shedding8,16,17,20 will be included to evaluate their role in the association between microbial diversity and colitis with Salmonella shedding. Odds ratios and confidence intervals will be calculated with standard errors using the interquartile range of diversity measures. Briefly, to identify taxa that may drive associations highlighted during modeling, microbiota counts will be normalized and aggregated to the family, genus, and species level for measurement in horse fecal samples. Counts will be compared between groups (healthy horses shedding Salmonella, healthy non-shedders, diseased shedders, diseased non-shedders) using ZIG and negative binomial modeling strategies.21,27 Modeling the counts of specific taxa found to be differentially abundant will follow the procedure proposed for diversity values above.