• View in gallery

    Box-and-whisker plots of the bacterial species richness (ie, number of observed species; A through C) and diversity (ie, Shannon diversity index; D through F) for samples extracted from the center (A, B, D, and E) or surface (A, C, D, and F) of fecal balls selected at various times after removal from the rectum of 6 healthy adult nonpregnant mixed-breed mares. From each horse, feces were obtained from the rectum by use of a small amount of lubrication and a nonsterile plastic rectal sleeve. One fecal ball (minimum, 100 g) was selected to represent the rectum sample, and the remainder of the feces were placed on the floor of a clean, straw-bedded box stall in a temperature-regulated environment. A fecal ball (minimum, 100 g) was selected for analysis from the stall floor at 0 (immediately after placement on stall floor), 2, 6, 12, and 24 hours after collection from the rectum. For each plot, the lower and upper borders of the box delimit the first and third quartiles, the horizontal line within the box represents the median, the whiskers delimit the range, and circles represent outliers. Values of P ≤ 0.05 were considered significant.

  • View in gallery

    Principle coordinate plots of the weighted paired UniFrac distance for the fecal samples (ie, bacterial communities) described in Figure 1 on the basis of sample location and collection time (A) and sample location and horse (B). The UniFrac distance is a measure of β diversity, or the extent of the relationship between bacterial communities. Percentages in parentheses in the axis titles represent the percentage of variation explained by the given principle coordinates. See Figure 1 for remainder of key.

  • View in gallery

    Mean relative abundances of the predominant bacterial phyla (A), families (B), and genera (C) identified in the fecal samples described in Figure 1. *Genera within the given order. †Genera within the given family. ‡Genera within the given class. See Figure 1 for remainder of key.

  • 1. McNulty NP, Yatsunenko T, Hsiao A, et al. The impact of a consortium of fermented milk strains on the gut microbiome of gnotobiotic mice and monozygotic twins. Sci Transl Med 2011;3:106ra106.

    • Search Google Scholar
    • Export Citation
  • 2. Hartman AL, Lough DM, Barupal DK, et al. Human gut microbiome adopts an alternative state following small bowel transplantation. Proc Natl Acad Sci U S A 2009;106:1718717192.

    • Search Google Scholar
    • Export Citation
  • 3. Rup L. The human microbiome project. Indian J Microbiol 2012;52:315.

  • 4. The Human Microbiome Project Consortium. A framework for human microbiome research. Nature 2012;486:215221.

  • 5. Qin J, Ruiqiang L, Raes J, et al. A human gut microbial gene catalog established by metagenomic sequencing. Nature 2010;464:5965.

  • 6. Blackmore TM, Dugdale A, Argo CM, et al. Strong stability and host specific bacterial community in faeces of ponies. PLoS One 2013;8:e75079.

    • Search Google Scholar
    • Export Citation
  • 7. Proudman CJ, Hunter JO, Darby AC, et al. Characterisation of the faecal metabolome and microbiome of Thoroughbred racehorses. Equine Vet J 2015;47:580586.

    • Search Google Scholar
    • Export Citation
  • 8. Weese JS, Holcombe SJ, Embertson RM, et al. Changes in the faecal microbiota of mares precede the development of post partum colic. Equine Vet J 2015;47:641649.

    • Search Google Scholar
    • Export Citation
  • 9. Steelman SM, Chowdhary BP, Dowd S, et al. Pyrosequencing of 16S rRNA genes in fecal samples reveals high diversity of hindgut microflora in horses and potential links to chronic laminitis. BMC Vet Res 2012;8:231.

    • Search Google Scholar
    • Export Citation
  • 10. Costa MC, Arroyo LG, Allen-Vercoe E, et al. Comparison of the fecal microbiota of healthy horses and horses with colitis by high throughput sequencing of the V3–V5 region of the 16s rRNA gene. PLoS One 2012;7:e41484.

    • Search Google Scholar
    • Export Citation
  • 11. Eckburg PB, Bik EM, Bernstein CN, et al. Diversity of the human intestinal microbial flora. Science 2005;308:16351638.

  • 12. Durbán A, Abellán JJ, Jiménez-Hernández N, et al. Assessing gut microbial diversity from feces and rectal mucosa. Microb Ecol 2011;61:123133.

    • Search Google Scholar
    • Export Citation
  • 13. Milinovich GJ, Burrell PC, Pollitt CC, et al. Microbial ecology of the equine hindgut during oligofructose-induced laminitis. ISME J 2008;2:10891100.

    • Search Google Scholar
    • Export Citation
  • 14. Pitta DW, Kumar S, Vecchiarelli B, et al. Temporal dynamics in the ruminal microbiome of dairy cows during the transition period. J Anim Sci 2014;92:40144022.

    • Search Google Scholar
    • Export Citation
  • 15. Wu S, Baldwin RL 6th, Li W, et al. The bacterial community composition of the bovine rumen detected using pyrosequencing of 16S rRNA genes. Metagenomics (Cairo) 2012;1:235571.

    • Search Google Scholar
    • Export Citation
  • 16. Hoffmann C, Dollive S, Grunberg S, et al. Archaea and fungi of the human gut microbiome: correlations with diet and bacterial residents. PLoS One 2013;8:e66019.

    • Search Google Scholar
    • Export Citation
  • 17. Caporaso JG, Kuczynski J, Stombaugh J, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010;7:335336.

    • Search Google Scholar
    • Export Citation
  • 18. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 2010;26:24602461.

  • 19. Caporaso JG, Bittinger K, Bushman FD, et al. PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics 2010;26:266267.

    • Search Google Scholar
    • Export Citation
  • 20. Price MN, Dehal PS, Arkin AP. FastTree 2—approximately maximum-likelihood trees for large alignments. PLoS One 2010;5:e9490.

  • 21. McDonald D, Price MN, Goodrich J, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J 2012;6:610618.

    • Search Google Scholar
    • Export Citation
  • 22. Wang Q, Garrity GM, Tiedje JM, et al. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 2007;73:52615267.

    • Search Google Scholar
    • Export Citation
  • 23. Lozupone C, Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol 2005;71:82288235.

    • Search Google Scholar
    • Export Citation
  • 24. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecol 2001;26:3246.

  • 25. Oksanen J, Blanchet FG, Friendly M, et al. Vegan: community ecology package version 2.4–6. Available at: cran.r-project.org/package=vegan. Accessed May 11, 2015.

    • Search Google Scholar
    • Export Citation
  • 26. Bates D, Maechler M, Bolker B, et al. lme4: linear mixed-effects models using S4 classes, version 1.1–6. Available at: cran.r-project.org/package=lme4. Accessed May 11, 2015.

    • Search Google Scholar
    • Export Citation
  • 27. Julliand V, Grimm P. Horse species symposium: the microbiome of the horse hindgut: history and current knowledge. J Anim Sci 2016;94:22622274.

    • Search Google Scholar
    • Export Citation
  • 28. Wang W, Cao J, Yang F, et al. High-throughput sequencing reveals the core gut microbiome of Bar-headed goose (Anser indicus) in different wintering areas in Tibet. Microbiologyopen 2016;5:287295.

    • Search Google Scholar
    • Export Citation
  • 29. Guo Y, Li SH, Kuang YS, et al. Effect of short-term room temperature storage on the microbial community in infant fecal samples. Sci Rep 2016;6:26648.

    • Search Google Scholar
    • Export Citation
  • 30. Venable EB, Bland SD, McPherson JL, et al. Role of the gut microbiota in equine health and disease. Anim Front 2016;6:4349.

  • 31. Shobar RM, Velineni S, Keshavarzian A, et al. The effects of bowel preparation on microbiota-related metrics differ in health and in inflammatory bowel disease and for the mucosal and luminal microbiota compartments. Clin Transl Gastroenterol 2016;7:e143.

    • Search Google Scholar
    • Export Citation
  • 32. Durso LM, Harhay GP, Smith TP, et al. Animal-to-animal variation in fecal microbial diversity among beef cattle. Appl Environ Microbiol 2010;76:48584862.

    • Search Google Scholar
    • Export Citation
  • 33. Choo JM, Leong LE, Rogers GB. Sample storage conditions significantly influence faecal microbiome profiles. Sci Rep 2015;5:16350.

  • 34. Costa MC, Silva G, Ramos RV, et al. Characterization and comparison of the bacterial microbiota in different gastrointestinal tract compartments in horses. Vet J 2015;205:7480.

    • Search Google Scholar
    • Export Citation
  • 35. Dougal K, Harris PA, Edwards A, et al. A comparison of the microbiome and the metabolome of different regions of the equine hindgut. FEMS Microbiol Ecol 2012;82:642652.

    • Search Google Scholar
    • Export Citation

Advertisement

Characterization of the fecal microbiota of healthy horses

View More View Less
  • 1 Equine Orthopaedic Research Center, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.
  • | 2 Department of Clinical Studies–New Bolton Center, University of Pennsylvania, Kennett Square, PA 19348.
  • | 3 Department of Clinical Studies–New Bolton Center, University of Pennsylvania, Kennett Square, PA 19348.
  • | 4 Department of Clinical Studies–New Bolton Center, University of Pennsylvania, Kennett Square, PA 19348.
  • | 5 Department of Clinical Studies–New Bolton Center, University of Pennsylvania, Kennett Square, PA 19348.
  • | 6 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Kennett Square, PA 19348.
  • | 7 Department of Clinical Studies–New Bolton Center, University of Pennsylvania, Kennett Square, PA 19348.

Abstract

OBJECTIVE To characterize the fecal microbiota of horses and to investigate alterations in that microbiota on the basis of sample collection site (rectum vs stall floor), sample location within the fecal ball (center vs surface), and duration of environmental exposure (collection time).

ANIMALS 6 healthy adult mixed-breed mares.

PROCEDURES From each horse, feces were collected from the rectum and placed on a straw-bedded stall floor. A fecal ball was selected for analysis immediately after removal from the rectum and at 0 (immediately), 2, 6, 12, and 24 hours after placement on the stall floor. Approximately 250 mg of feces was extracted from the surface and center of each fecal ball, and genomic DNA was extracted, purified, amplified for the V1-V2 hypervariable region of the 16S rDNA gene, and analyzed with a bioinformatics pipeline.

RESULTS The fecal microbiota was unique for each horse. Bacterial community composition varied significantly between center and surface fecal samples but was not affected by collection time. Bacterial community composition varied rapidly for surface fecal samples. Individual bacterial taxa were significantly associated with both sample location and collection time but remained fairly stable for up to 6 hours for center fecal samples.

CONCLUSIONS AND CLINICAL RELEVANCE Results indicated that, for horses, fecal samples for microbiota analysis should be extracted from the center of fecal balls collected within 6 hours after defecation. Samples obtained up to 24 hours after defecation can be analyzed with the realization that some bacterial populations may deviate from those immediately after defecation.

Abstract

OBJECTIVE To characterize the fecal microbiota of horses and to investigate alterations in that microbiota on the basis of sample collection site (rectum vs stall floor), sample location within the fecal ball (center vs surface), and duration of environmental exposure (collection time).

ANIMALS 6 healthy adult mixed-breed mares.

PROCEDURES From each horse, feces were collected from the rectum and placed on a straw-bedded stall floor. A fecal ball was selected for analysis immediately after removal from the rectum and at 0 (immediately), 2, 6, 12, and 24 hours after placement on the stall floor. Approximately 250 mg of feces was extracted from the surface and center of each fecal ball, and genomic DNA was extracted, purified, amplified for the V1-V2 hypervariable region of the 16S rDNA gene, and analyzed with a bioinformatics pipeline.

RESULTS The fecal microbiota was unique for each horse. Bacterial community composition varied significantly between center and surface fecal samples but was not affected by collection time. Bacterial community composition varied rapidly for surface fecal samples. Individual bacterial taxa were significantly associated with both sample location and collection time but remained fairly stable for up to 6 hours for center fecal samples.

CONCLUSIONS AND CLINICAL RELEVANCE Results indicated that, for horses, fecal samples for microbiota analysis should be extracted from the center of fecal balls collected within 6 hours after defecation. Samples obtained up to 24 hours after defecation can be analyzed with the realization that some bacterial populations may deviate from those immediately after defecation.

The gastrointestinal microbiome is a diverse colony of microorganisms that has a synergistic and specifically adapted relationship to the digestive function of the host and may be altered by changes in the diet or environment, administration of medications (including antimicrobials), stress, or disease.1,2 Large-scale studies3–5 involving humans have been conducted to investigate the relationship between the microbiome and diseases such as obesity and inflammatory bowel disease. This area of research has since expanded to include veterinary species including equids. Results of a pilot study6 indicate that the fecal microbiome of ponies maintained with similar diets and housing conditions is fairly stable, but the fecal microbiome of individual ponies tends to have unique characteristics. In a study7 of Thoroughbred racehorses, a dietary change had a significant effect on the fecal microbial community structure and intestinal metabolites. Alterations in the fecal microbiome are associated with postpartum colic,8 chronic laminitis,9 and colitis,10 conditions that have high morbidity and mortality rates in horses.

Equine fecal samples used for DNA extraction and analysis in previous studies7–10 were obtained from the ground or stall floor, and sample collection following defecation ranged from immediately to within 24 hours after hospital admission. Those inconsistencies indicate lack of a standardized protocol for collection and handling of fecal samples for microbiome analysis. Although research on the equine microbiome is ever increasing, to our knowledge, the effect of sample collection protocol, including the portion (surface vs center) of the fecal ball from which the sample is obtained and the duration between defecation and sample collection, on the composition of the fecal microbiome has not been investigated. Protocols of fecal microbiome studies11,12 involving dogs and humans also vary in terms of sample collection, handling, and processing, but the various techniques used in those studies have been more comprehensively studied and validated than those used in studies involving horses.13 In regard to horses, because the fecal surface is generally exposed to air and the stall or pasture environment, it seems reasonable to postulate that environmental exposure may produce temporal differences in the absolute number or relative proportions of certain aerobic and anaerobic bacteria between the surface and center of fecal balls, which may occur within minutes or hours after defecation. Moreover, inconsistencies in fecal sample collection, processing, and handling, such as freezing and thawing, may contribute to variation in the microbial composition of the microbiome among horses.6

The purpose of the study reported here was to characterize the fecal microbiota of healthy horses with access to pasture and fed a diet of free-choice timothy hay and to investigate alterations in that microbiota on the basis of sample collection site (rectum vs stall floor), sample location in the fecal ball (surface vs center), and duration of environmental exposure (collection time). Our hypotheses were that the microbiota composition would differ significantly between fecal samples collected per rectum and those collected from the stall floor and between samples extracted from the surface and those extracted from the center of a fecal ball and that the fecal microbiota would change rapidly over time owing to environmental exposure.

Materials and Methods

Animals

All study protocols were approved by the University of Pennsylvania Institutional Animal Care and Use Committee. Six university-owned, adult mixed-breed nonpregnant mares with ages ranging from 5 to 16 years and weights ranging from 450 to 575 kg were enrolled in the study. All horses were determined to be healthy and free of gastrointestinal parasites on the basis of results of a physical examination and fecal flotation. Horses were housed on pasture with ad libitum access to water and timothy hay.

Fecal sample collection

From each horse, feces were obtained from the rectum by use of a small amount of lubrication and a nonsterile plastic rectal sleeve.a One fecal ball (minimum, 100 g) was selected to represent the rectum sample, and the remainder of the feces were placed on the floor of a clean, straw-bedded box stall (4.3 × 4.3 m) in a temperature-regulated environment where the ambient temperature ranged from 10° to 20°C during the experimental period. A fecal ball (minimum, 100 g) was collected from the stall floor at 0 (immediately after placement on stall floor), 2, 6, 12, and 24 hours after removal from the rectum. Each fecal sample was placed in a sterile container.b Samples were stored at 4°C during the 24-hour sample collection period, frozen, and stored at −80°C until processing.

16S rDNA pyrosequencing

Frozen fecal samples were thawed on ice. Approximately 250 mg of feces was scraped from the surface of each fecal ball to represent the surface fecal sample. Then, the fecal ball was split, and approximately 250 mg of feces was extracted from the center to represent the center fecal sample. Genomic DNA was extracted from each sample by use of a commercially available kit.c The extraction process involved a bead-beating step prior to initiation of the manufacturer's guidelines. The extracted DNA was quantified and purified as described14 and then amplified with bacterial-specific primers BSF8 (5′-GAGTTTGATCCTGGCTCAG-3′) and BSR357 (5′-CTGCTGCCTYCCGTA-3′) that annealed to the V1-V2 hypervariable regions of the 16S rDNA bacterial gene. A PCR assayd was performed in accordance with the manufacturer's instructions. Briefly, the thermal cycling conditions included an initial denaturing step at 95°C for 5 minutes; followed by 25 cycles of denaturing at 95°C for 30 seconds, annealing at 56°C for 30 seconds, and extension at 72°C for 90 seconds; and then a final extension step at 72°C for 8 minutes.15 The PCR product was bead-purifiede as described.16 The purified PCR products were pooled in equal concentrations and sequenced with a commercial platformf at the DNA Sequencing Facility, Perelman School of Medicine, University of Pennsylvania.

Data analysis

The 16S rDNA sequences obtained were decoded and analyzed with an open-source bioinformatics pipeline as described.17 Reads were eliminated if they did not match the sample-specific barcode and amplified sequences were < 200 bp or > 1,000 bp or if they contained homopolymer sequences > 6 bp. The possibility of chimeric sequences in the analysis was diminished by removing sequences of low quality (ie, < 200 bp) and unidirectional sequencing. Operational taxonomic units were formed at 97% similarity by use of a clustering, alignment, and search algorithm,18 and representative sequences from each OTU were aligned to 16S rDNA reference sequences.19 A phylogenetic tree was constructed with open-source softwareg as described.20 Taxonomic assignments within a 16 rRNA gene database21 were generated with a naïve Bayesian classifer.22 To perform α (comparison of OTUs within a sample) and β (comparison of OTUs between samples) diversity analyses, OTUs were rarified at a depth of 1,540 sequences/sample. Alpha diversity was measured by calculating both the number of observed species as well as the Shannon diversity index, which was used to measure microbial species richness and diversity. The β diversity, or extent of the relationship between bacterial communities, was quantified by calculation of the weighted pairwise UniFrac distance as described.23 The UniFrac distances were then plotted and evaluated by use of a principal coordinate analysis.

Statistical analysis

All statistical analyses were performed with commercially available software.h The Wilcoxon rank sum test was used to compare the measured α diversity matrices between surface and center fecal samples. The Kruskal-Wallis test was used to compare α diversity matrices over time within each location (surface or center) of the sampled fecal balls. A nonparametric permutational multivariate ANOVA24 was implemented as described25 to evaluate the effects of individual horse, sample location (surface or center of fecal ball), and collection time on the overall microbiota community composition as measured by the weighted pairwise UniFrac distance. To test for differences in taxon abundance, the abundances were normalized to the total number of reads in each sample (relative abundance). Operational taxonomic units that appeared in at least 75% of the bacterial communities were considered for further analysis. All fecal samples were analyzed, but OTUs that appeared in < 75% of the bacterial communities were removed from the statistical analysis. A generalized linear mixed model was constructed as described26 with a binomial link function. The model included a random effect for horse to account for evaluation of multiple samples from each horse. The input data for the mixed model consisted of a 2-column matrix; one column contained the number of reads assigned to a given OTU, and the other column contained the number of reads assigned to other OTUs. The false discovery rate method was used to adjust P values and control for multiple testing; values of P ≤ 0.05 were considered significant.

Results

Sequencing information

A total of 72 fecal samples were collected (6 horses × 2 locations [surface and center] for each fecal ball collected × 6 collection times). Two samples (a surface fecal sample and a center fecal sample collected from the stall floor at 0 and 24 hours after removal from the rectum, respectively) contained < 200 reads and were removed from analysis. Thus, 70 fecal samples were analyzed.

Approximately 351,176 reads were analyzed from the 70 fecal samples (ie, bacterial communities) analyzed. Those bacterial communities had a mean of 5,016 reads/sample and a minimum of 1,540 reads/sample. Approximately 38,546 OTUs were identified by clustering at 97% sequence similarity. Representative sequences from those OTUs were assigned to 19 bacterial phyla. A total of 115 known genera were observed in this study. Two-thirds of the OTUs were assigned to known genera, and the remaining OTUs were assigned the taxonomic rank to which they were identified.

Bacterial community comparisons

Surface fecal samples had significantly greater bacterial species richness (P = 0.032) and diversity (P = 0.018) than center fecal samples (Figure 1). However, the species richness and diversity did not differ significantly over time within a sample location.

Figure 1—
Figure 1—

Box-and-whisker plots of the bacterial species richness (ie, number of observed species; A through C) and diversity (ie, Shannon diversity index; D through F) for samples extracted from the center (A, B, D, and E) or surface (A, C, D, and F) of fecal balls selected at various times after removal from the rectum of 6 healthy adult nonpregnant mixed-breed mares. From each horse, feces were obtained from the rectum by use of a small amount of lubrication and a nonsterile plastic rectal sleeve. One fecal ball (minimum, 100 g) was selected to represent the rectum sample, and the remainder of the feces were placed on the floor of a clean, straw-bedded box stall in a temperature-regulated environment. A fecal ball (minimum, 100 g) was selected for analysis from the stall floor at 0 (immediately after placement on stall floor), 2, 6, 12, and 24 hours after collection from the rectum. For each plot, the lower and upper borders of the box delimit the first and third quartiles, the horizontal line within the box represents the median, the whiskers delimit the range, and circles represent outliers. Values of P ≤ 0.05 were considered significant.

Citation: American Journal of Veterinary Research 79, 8; 10.2460/ajvr.79.8.811

The similarity between bacterial communities (β diversity) was significantly associated with sample location (P < 0.001). However, within a sample location (surface or center), the similarity between bacterial communities was not associated with collection time (Figure 2). Bacterial communities were clustered by horse (P < 0.001). Moreover, within each horse, bacterial communities were clustered on the basis of sample location, and the clusters for surface fecal samples tended to be distinct from the clusters for center fecal samples. Clustering of bacterial communities on the basis of collection time was not observed. In fact, within a horse and sample location, the bacterial community composition remained fairly stable throughout the 24-hour observation period.

Figure 2—
Figure 2—

Principle coordinate plots of the weighted paired UniFrac distance for the fecal samples (ie, bacterial communities) described in Figure 1 on the basis of sample location and collection time (A) and sample location and horse (B). The UniFrac distance is a measure of β diversity, or the extent of the relationship between bacterial communities. Percentages in parentheses in the axis titles represent the percentage of variation explained by the given principle coordinates. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 79, 8; 10.2460/ajvr.79.8.811

Taxonomic composition of bacterial communities

Fecal microbial communities at each sample location and collection time were characterized at the level of bacterial phylum, family, and genus (Figure 3). The 16S rDNA-encoding gene sequences were dominated by Bacteroidetes (53% to 58%) and Firmicutes (31% to 35%), which collectively comprised 90% of the total bacterial abundance. Other phyla including Spirochaetes (2% to 4%), Fibrobacteres (1.5% to 4.2%), Proteobacteria (0.8% to 2.2%), and Cyanobacteria (0.5% to 1.5%) were present in low proportions across all samples. For most phyla, the relative abundance differed significantly between center and surface fecal samples (Supplementary Table S1; available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.8.811). The respective relative abundances of Fibrobacteres, Spirochaetes, and Tenericutes in center fecal samples were significantly (P < 0.001) greater than those in surface fecal samples. Conversely, the respective relative abundances of Bacteroidetes, Proteobacteria, Cyanobacteria, SR-1, and Verrucomicrobia in center fecal samples were significantly (P < 0.05) lower than those in surface fecal samples. The relative abundance of Firmicutes did not differ significantly between center and surface fecal samples. There was a significant interaction between sample location and collection time for Bacteroidetes (P < 0.001), Proteobacteria (P < 0.01), and Spirochaetes (P < 0.05).

Figure 3—
Figure 3—

Mean relative abundances of the predominant bacterial phyla (A), families (B), and genera (C) identified in the fecal samples described in Figure 1. *Genera within the given order. †Genera within the given family. ‡Genera within the given class. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 79, 8; 10.2460/ajvr.79.8.811

The predominant families identified within the bacterial communities were unclassified Bacteroidales (identified only at the order taxonomic rank; up to 57%) followed by Paraprevotellaceae, Prevotellaceae, Bacteroidaceae, and several families with relative abundances < 1% (Figure 3; Supplementary Table S2; available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.8.811). Among Firmicutes, Lachnospiraceae and Ruminococcaceae were the most abundant families followed by unclassified Clostridiales. At all collection times, the relative abundance of Lachnospiraceae in center fecal samples was greater than that in surface fecal samples, whereas the relative abundance of Ruminococcaceae in center fecal samples was less than that in surface fecal samples. Six families of Proteobacteria were identified; however, the collective relative abundance was < 3% for those 6 families at each sample location and collection time. In fact, in center fecal samples, 4 (Alcaligenaceae, Burkholderiales, Desulfovibrionaceae, and RF32) of those 6 families were not detected at all, and the remaining 2 families (Alphaproteobacteria and GMD14H09) were detected only at extremely low levels (< 1%). Spirochaetes and Fibrobacteres were the predominant families identified from the Spirochaetaceae and Fibrobacteraceae phyla, respectively. The respective relative abundance for both of those families was significantly greater in center fecal samples than in surface fecal samples.

A total of 15 genera within the Bacteroidetes phylum were identified, and all 15 genera were present in approximately 70% of the samples evaluated (Supplementary Table S3; available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.8.811). Of those 15 genera, the most predominant was an unclassified genus (Bacteroidales identified only to the order rank) followed by Prevotella (5% to 7%); the relative abundance of all other genera was < 2% (Figure 3). The relative abundances of Prevotella and YRC22 in center fecal samples were significantly greater than those in surface fecal samples. The relative abundances of 2 genera (Coprococcus and an unclassified genus) in the Lachnospiraceae family were consistently greater in center fecal samples than in surface fecal samples at all collection times. The relative abundance of an unclassified genus of the Ruminococcaceae family was consistently greater in surface fecal samples than in center fecal samples, whereas the relative abundance of Ruminococcus was consistently greater in center fecal samples than in surface fecal samples. Among genera of the Firmicutes phylum, Anaerostipes, Bulleidia, an unclassified genus in the Clostridiaceae family, Mogibacterium, and Pseudoramibacter were not detected in center fecal samples. Four of the 5 genera of the Proteobacteria phylum were unclassified, and 3 of those genera were not detected in center fecal samples. The predominant genus detected from the Spirochaetes phylum was Treponema, and the relative abundance of Treponema in center fecal samples was consistently greater than that in surface fecal samples at all collection times. Fibrobacter was the only genus detected from the Fibrobacteres phylum, and the relative abundance of Fibrobacter in center fecal samples was approximately twice that in surface fecal samples at all collection times. In general, sample location had a greater effect than collection time on bacterial community composition. Center fecal samples had fewer variations in individual bacterial taxa over time than did surface fecal samples.

Within a sample location, changes in the relative abundances of bacterial phyla, families, and genera were observed between the initial rectum samples and samples collected at other times (Supplementary Tables S2–S4, available at http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.8.811). For center fecal samples, Bacteroidetes and Firmicutes were the only 2 phyla that differed significantly between the initial rectum samples and samples collected at 0 hours. By 24 hours after removal from the rectum, the relative abundances for 6 of the 11 most predominant bacterial phyla detected differed significantly from the corresponding relative abundances for the initial rectum samples. In contrast, for surface fecal samples, several bacterial phyla varied significantly with time, and the magnitude of those differences tended to increase over time. In general, the relative abundances of bacterial phyla, families, and genera in surface fecal samples tended to change with a high frequency, whereas those in center fecal samples were fairly stable until 6 hours after removal from the rectum. Moreover, the magnitude of the changes in the relative abundances of bacterial phyla, families, and genera in surface fecal samples was typically greater than those in center fecal samples.

Discussion

Results of the present study indicated that the fecal microbiota of healthy horses did not vary significantly on the basis of collection site (rectum vs stall floor), but did vary significantly on the basis of sample location (surface vs center) in a fecal ball. Although the overall relative abundance of bacteria in fecal samples was not significantly associated with collection time, individual microbial populations at the bacterial phylum, family, and genus levels varied on the basis of both sample location and collection time (ie, exposure to environment). Those changes occurred fairly rapidly and frequently for surface fecal samples but typically did not occur until 6 hours after removal from the rectum for center fecal samples. Additionally, the magnitude of the compositional changes in the bacterial community for surface fecal samples tended to be greater than that for center fecal samples. In fact, the microbiota of center fecal samples remained fairly stable for the duration of the 24-hour observation period.

Advantages associated with the use of feces rather than digesta from the hindgut or other intestinal compartments for evaluation of the microbiota of horses have been described elsewhere.27 However, in regard to horses, the optimum strategy for collection of fecal samples from a stall floor, the ideal location within a fecal ball for sample extraction, and the duration that a fecal sample can remain on a stall floor or be exposed to the environment without adversely or substantially altering microbial analysis have not been described. Results of the present study indicated that small changes in individual bacterial taxa were detectable in both surface and center fecal samples as soon as the sample was placed on the stall floor (0 hours), which suggested that collection of fecal samples per rectum was the preferred sampling method for microbiota analysis. However, collection of fecal samples directly from the rectum may not be feasible in many situations, and the surface of fecal balls that have been exposed to the environment for any amount of time is likely to be contaminated and not representative of the fecal microbiota.28 Therefore, the purpose of the present study was to compare the microbiota between fecal samples obtained from the surface and center of fecal balls and also over time.

In the present study, both the number of bacterial species and bacterial diversity were greater for surface fecal samples than for center fecal samples. Additionally, for each of the 6 study horses, the composition of the bacterial community for center fecal samples was distinct from the composition of the bacterial community for surface fecal samples. Differences in the composition of the bacterial community from the surface to the center of the same fecal ball are most likely a function of the transition of feces from an anaerobic to aerobic environment during defecation (or removal from the rectum). The gut is an anaerobic environment; thus, gut microbes consist of both obligatory and facultative anaerobes. When feces are exposed to aerobic conditions, the growth of obligatory anaerobes slows while that of facultative anaerobes, such as Streptococcus spp, Enterococcus spp, and Enterobacteriaceae, increases.29 Although the methods (PCR assay) used in the present study detected both viable and unviable bacteria, the microbial composition of surface fecal samples was more dynamic than that of center fecal samples. In particular, populations of obligate anaerobes such as Prevotella, CF231, Cytophaga, YRC22, Coprococcus, unclassified genera of Clostridiales, Lachnospiraceae, Lactobacillaceae, Veillonellaceae, and Fibrobacter were found in lower abundance in surface fecal samples relative to center fecal samples, which suggested that the growth of those organisms slowed when exposed to oxygen. In contrast, populations of facultative anaerobes such as Proteobacteria, most genera of Bacteroidetes, Ruminococcaceae, Erysipelotrichaceae, and Eubacterium were found in greater abundance in surface fecal samples than in center fecal samples. In the present study, alterations in the bacterial population dynamics of fecal samples occurred fairly rapidly, with substantial shifts in bacterial communities identified between 2 and 6 hours after fecal sample removal from the rectum. We believe these results indicate that the environment required for sustained growth of obligate anaerobes was preserved in the center but not on the surface of fecal balls. Therefore, the bacterial community composition at the center of a fecal ball is likely to be most representative of the current microbiome, and fecal samples extracted from the center of fecal balls are recommended for microbiota analysis.

The effect of collection time on bacterial community composition was small, compared with that of sample (center or surface) location. Except for Bacteroidetes and Firmicutes immediately after placement on the stall surface (0 hours), individual bacterial taxa remained unaltered for center fecal samples for up to 6 hours after removal from the rectum, whereas most bacterial populations shifted noticeably over time for surface fecal samples. Those findings further supported our recommendation that the best samples for microbiota analysis are obtained from the center of fecal balls, especially within 6 hours after defecation.

The relative distribution of the various bacterial phyla comprising the fecal microbiota of the healthy horses of the present study was comparable to that described in a recently published review article.30 In the present study, the dominant bacterial phylum identified was Bacteroidetes (> 50%) followed by Firmicutes (30% to 35%). Minor phyla included Fibrobacteres, Spirochaetes, Proteobacteria, Tenericutes, and Cyanobacteria, which collectively comprised about 10% of bacterial abundance. At the genus level, unclassified genera among the Bacteroidales, Lachnospiraceae, and Ruminococcaceae families were fairly abundant, compared with the abundance of known genera such as Prevotella, Butyrivibrio, Clostridium, Bacillus, and Lactobacillus, which comprised a substantial proportion of the fecal microbiome of horses described in another review article.27 Additionally, the microbiota composition was unique for each horse, as evidenced by the fact that the intersubject variation was greater than the intrasubject variation. This substantial intersubject variation corroborates findings of other fecal microbiome studies involving ponies,6 humans,29,31 and cattle.32 Similar to the study32 involving cattle, significant differences in the fecal microbiota composition were detected among horses, despite the fact that all study subjects were healthy adult mixed-breed nonpregnant mares maintained at the same geographic location, fed the same diet, and used for the same general purposes (eg, teaching). These findings suggested that other factors contribute to each animal's unique microbiota, such as differences in digestive capabilities, genetics, prior diets, and use. Identification of those factors was beyond the scope of the present study and will require a larger study population and greater resolution of sequencing information than was available for this study to investigate individual host effects on microbiota composition.

The present study was conducted to determine the respective effects of sampling location and collection time on the fecal microbiota composition of healthy horses. The most notable limitation of this study was the small sample size, which limited the conclusions that could be drawn from the data. A larger study population would have increased bacterial coverage and sequencing depth and may have further reinforced the microbial profiles for samples extracted from the center and surface of fecal balls; however, the focus of the study was to monitor changes in the fecal microbiota over time for each horse. The fact that all fecal samples analyzed were obtained from healthy adult mares housed at the same location, fed the same diet, and used for the same general purposes at 1 point in time was essential for the development of a standardized technique. The fecal microbiota may change on the basis of temperature, exposure to natural elements in the local environment (ie, pasture vs straw-bedded stall), and geographic region. Also, in this study, changes in the fecal microbiota were monitored for only 24 hours after feces were removed from the rectum. Additional research is necessary to characterize changes in the fecal microbiota that occur over a longer duration. In this study, all fecal samples were handled similarly, and although different handling techniques were not specifically evaluated, researchers of another study29 reported that exposure of fecal samples to air at room temperature for 2 hours did not have a significant effect on α diversity. For that reason, all fecal samples obtained during the present study were refrigerated immediately after collection, frozen within 24 hours after collection, and stored at −80°C until analysis. Results of another study33 indicate that the duration the fecal samples are stored at 4°C does not affect microbial analysis results. However, freezing and thawing the samples might have affected microbiota composition, particularly of the surface fecal samples. Future studies should investigate how the fecal microbiota changes before and after samples are frozen and thawed, between samples that are refrigerated versus those that are frozen immediately after collection, and among samples collected at various times of the year. The physiologic temporal formation of fecal balls could not be controlled in the present study. For example, fecal matter at the center of fecal balls might have been organized from a more proximal region of the gastrointestinal tract than fecal matter at the surface of fecal balls, which could have affected the microbial composition. Regional differences in the microbiota of the gastrointestinal tract of horses have been identified34,35 and should be an area of continued investigation.

In the present study, the composition of the fecal microbiota varied significantly among 6 healthy adult mares maintained at the same location and fed the same diet. Although the fecal microbial composition was unique for each horse, it appeared to remain fairly stable within the center of fecal balls for 24 hours after removal of feces from the rectum. Therefore, we concluded that fecal samples collected from a stall floor were acceptable for microbiota analysis as long as the samples analyzed were extracted from the center of fecal balls. Ideally, fecal samples submitted for microbiota analysis should be extracted from the center of fecal balls collected within 6 hours after defecation, although samples extracted from the center of fecal balls collected for up to 24 hours after defecation can be analyzed with the realization that some bacterial populations may deviate from those immediately after defecation. The establishment of a standardized method for fecal sample collection is essential for continued meaningful research into the equine fecal microbiome. Nevertheless, the findings of the present study provided a better understanding of the effects of sample location and collection time on the fecal microbiota of healthy horses.

Acknowledgments

All research was performed at New Bolton Center, University of Pennsylvania, Kennett Square, Pa.

Supported by the Raymond Firestone Trust Research Grant. The authors declare that there were no conflicts of interest.

Data availability

The raw sequence data, corresponding metadata, quality-filtered reads, representative sequences, and taxonomy assignments to the representative sequences for fecal samples evaluated in this study have been deposited in Figshare (figshare.com/s/0b27b5f92be617ca6857).

ABREVIATIONS

OTU

Operational taxonomic unit

Footnotes

a.

Ag-Tek, Neogen, Lansing, Mich.

b.

Amsino International Inc, Pomona, Calif.

c.

PSP Spin Stool DNA Plus Kit, Invitek, Berlin, Germany.

d.

Accuprime Taq DNA Polymerase System, Invitrogen, Carlsbad, Calif.

e.

Agencourt AMPure XP Beads, Beckman-Coulter, Fort Collins, Colo.

f.

Ion Torrent, Thermo Fisher Scientific, Waltham, Mass.

g.

FastTree, version 2. Available at: www.microbesonline.org/fasttree/#Install. May 11, 2015.

h.

R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: www.R-project.org. Accessed May 11, 2015.

References

  • 1. McNulty NP, Yatsunenko T, Hsiao A, et al. The impact of a consortium of fermented milk strains on the gut microbiome of gnotobiotic mice and monozygotic twins. Sci Transl Med 2011;3:106ra106.

    • Search Google Scholar
    • Export Citation
  • 2. Hartman AL, Lough DM, Barupal DK, et al. Human gut microbiome adopts an alternative state following small bowel transplantation. Proc Natl Acad Sci U S A 2009;106:1718717192.

    • Search Google Scholar
    • Export Citation
  • 3. Rup L. The human microbiome project. Indian J Microbiol 2012;52:315.

  • 4. The Human Microbiome Project Consortium. A framework for human microbiome research. Nature 2012;486:215221.

  • 5. Qin J, Ruiqiang L, Raes J, et al. A human gut microbial gene catalog established by metagenomic sequencing. Nature 2010;464:5965.

  • 6. Blackmore TM, Dugdale A, Argo CM, et al. Strong stability and host specific bacterial community in faeces of ponies. PLoS One 2013;8:e75079.

    • Search Google Scholar
    • Export Citation
  • 7. Proudman CJ, Hunter JO, Darby AC, et al. Characterisation of the faecal metabolome and microbiome of Thoroughbred racehorses. Equine Vet J 2015;47:580586.

    • Search Google Scholar
    • Export Citation
  • 8. Weese JS, Holcombe SJ, Embertson RM, et al. Changes in the faecal microbiota of mares precede the development of post partum colic. Equine Vet J 2015;47:641649.

    • Search Google Scholar
    • Export Citation
  • 9. Steelman SM, Chowdhary BP, Dowd S, et al. Pyrosequencing of 16S rRNA genes in fecal samples reveals high diversity of hindgut microflora in horses and potential links to chronic laminitis. BMC Vet Res 2012;8:231.

    • Search Google Scholar
    • Export Citation
  • 10. Costa MC, Arroyo LG, Allen-Vercoe E, et al. Comparison of the fecal microbiota of healthy horses and horses with colitis by high throughput sequencing of the V3–V5 region of the 16s rRNA gene. PLoS One 2012;7:e41484.

    • Search Google Scholar
    • Export Citation
  • 11. Eckburg PB, Bik EM, Bernstein CN, et al. Diversity of the human intestinal microbial flora. Science 2005;308:16351638.

  • 12. Durbán A, Abellán JJ, Jiménez-Hernández N, et al. Assessing gut microbial diversity from feces and rectal mucosa. Microb Ecol 2011;61:123133.

    • Search Google Scholar
    • Export Citation
  • 13. Milinovich GJ, Burrell PC, Pollitt CC, et al. Microbial ecology of the equine hindgut during oligofructose-induced laminitis. ISME J 2008;2:10891100.

    • Search Google Scholar
    • Export Citation
  • 14. Pitta DW, Kumar S, Vecchiarelli B, et al. Temporal dynamics in the ruminal microbiome of dairy cows during the transition period. J Anim Sci 2014;92:40144022.

    • Search Google Scholar
    • Export Citation
  • 15. Wu S, Baldwin RL 6th, Li W, et al. The bacterial community composition of the bovine rumen detected using pyrosequencing of 16S rRNA genes. Metagenomics (Cairo) 2012;1:235571.

    • Search Google Scholar
    • Export Citation
  • 16. Hoffmann C, Dollive S, Grunberg S, et al. Archaea and fungi of the human gut microbiome: correlations with diet and bacterial residents. PLoS One 2013;8:e66019.

    • Search Google Scholar
    • Export Citation
  • 17. Caporaso JG, Kuczynski J, Stombaugh J, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010;7:335336.

    • Search Google Scholar
    • Export Citation
  • 18. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 2010;26:24602461.

  • 19. Caporaso JG, Bittinger K, Bushman FD, et al. PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics 2010;26:266267.

    • Search Google Scholar
    • Export Citation
  • 20. Price MN, Dehal PS, Arkin AP. FastTree 2—approximately maximum-likelihood trees for large alignments. PLoS One 2010;5:e9490.

  • 21. McDonald D, Price MN, Goodrich J, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J 2012;6:610618.

    • Search Google Scholar
    • Export Citation
  • 22. Wang Q, Garrity GM, Tiedje JM, et al. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 2007;73:52615267.

    • Search Google Scholar
    • Export Citation
  • 23. Lozupone C, Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol 2005;71:82288235.

    • Search Google Scholar
    • Export Citation
  • 24. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecol 2001;26:3246.

  • 25. Oksanen J, Blanchet FG, Friendly M, et al. Vegan: community ecology package version 2.4–6. Available at: cran.r-project.org/package=vegan. Accessed May 11, 2015.

    • Search Google Scholar
    • Export Citation
  • 26. Bates D, Maechler M, Bolker B, et al. lme4: linear mixed-effects models using S4 classes, version 1.1–6. Available at: cran.r-project.org/package=lme4. Accessed May 11, 2015.

    • Search Google Scholar
    • Export Citation
  • 27. Julliand V, Grimm P. Horse species symposium: the microbiome of the horse hindgut: history and current knowledge. J Anim Sci 2016;94:22622274.

    • Search Google Scholar
    • Export Citation
  • 28. Wang W, Cao J, Yang F, et al. High-throughput sequencing reveals the core gut microbiome of Bar-headed goose (Anser indicus) in different wintering areas in Tibet. Microbiologyopen 2016;5:287295.

    • Search Google Scholar
    • Export Citation
  • 29. Guo Y, Li SH, Kuang YS, et al. Effect of short-term room temperature storage on the microbial community in infant fecal samples. Sci Rep 2016;6:26648.

    • Search Google Scholar
    • Export Citation
  • 30. Venable EB, Bland SD, McPherson JL, et al. Role of the gut microbiota in equine health and disease. Anim Front 2016;6:4349.

  • 31. Shobar RM, Velineni S, Keshavarzian A, et al. The effects of bowel preparation on microbiota-related metrics differ in health and in inflammatory bowel disease and for the mucosal and luminal microbiota compartments. Clin Transl Gastroenterol 2016;7:e143.

    • Search Google Scholar
    • Export Citation
  • 32. Durso LM, Harhay GP, Smith TP, et al. Animal-to-animal variation in fecal microbial diversity among beef cattle. Appl Environ Microbiol 2010;76:48584862.

    • Search Google Scholar
    • Export Citation
  • 33. Choo JM, Leong LE, Rogers GB. Sample storage conditions significantly influence faecal microbiome profiles. Sci Rep 2015;5:16350.

  • 34. Costa MC, Silva G, Ramos RV, et al. Characterization and comparison of the bacterial microbiota in different gastrointestinal tract compartments in horses. Vet J 2015;205:7480.

    • Search Google Scholar
    • Export Citation
  • 35. Dougal K, Harris PA, Edwards A, et al. A comparison of the microbiome and the metabolome of different regions of the equine hindgut. FEMS Microbiol Ecol 2012;82:642652.

    • Search Google Scholar
    • Export Citation

Contributor Notes

Address correspondence to Dr. Stewart (holly.stewart@colostate.edu).