Evaluation of changes in microbiota after fecal microbiota transplantation in 6 diarrheic horses

Abstract

The purpose of this study was to characterize the fecal microbiota of horses with acute and chronic diarrhea before and after fecal microbiota transplantation (FMT). Six client-owned horses with acute and chronic diarrhea received FMT from 2 healthy donor horses. Microbiota analysis using next-generation sequencing was performed on fecal samples collected before and 2 and 7 d after FMT. Signs of diarrhea improved in 4 horses, whereas the remaining 2 horses did not survive. There was a significant difference in the number of bacterial species between donors and recipients (P < 0.05). The Order Lactobacillales and the genera Lactobacillus, Intestinimonas, and Streptococcus were increased in the microbiota of diarrheic horses, and Saccharofermentans genus increased in healthy donors. The results suggest that FMT from the healthy donors was not effective over a 7-day period as it did not change the fecal microbiota of the diarrheic horses. Further research to improve the efficacy of FMT in horses is needed.

Introduction

The bacteria residing in the gastrointestinal tract (GIT), known as the gut microbiota, are important contributors to maintaining the health of the host, such as in nutrition, energy metabolism, immune development, and host defense against harmful pathogens (1,2). The development of next-generation sequencing has made it possible for in-depth characterization of the bacterial communities present in the intestinal microbiota (3). The equine gut microbiota in the large intestine has a critical role in cellulose fermentation and short-chain fatty acid (SCFA) production, the latter providing them with their main energy sources (4). Colitis (5), colic (6), diet change (7), and antimicrobial administration (8) are associated with dysbiosis (imbalances) of the horse gut microbiota. Diseases affecting the GIT are the leading causes of morbidity and mortality in horses (9).

Horses with colitis experience a mortality rate of 40% (10) and disease etiology remains undetermined in > 50% of cases (10). Common infectious agents include Clostridioides difficile, Neorickettsia risticii, and Salmonella spp., which are commonly diagnosed through histopathology, bacterial cultures, polymerase chain reaction (PCR), and toxicology, whereas common non-infectious causes are antibiotic-associated diarrhea, sand impaction, dietary imbalances, and inflammatory bowel disease (10). Current treatments include fluid therapy, nonsteroidal anti-inflammatory drugs, and antimicrobials (10), but studies investigating manipulation of the gut microbiota in horses are scarce.

Microbiota manipulation has been used to treat GIT diseases, including the use of probiotics and fecal microbiota transplantation (FMT). Probiotics are defined as live microorganisms which when administered in adequate amounts confer a health benefit on the host (11). The use of probiotics to treat humans and horses with dysbiosis remains controversial, as many studies fail to show concrete evidence of clinical and microbiological benefits. Fecal microbiota transplantation involves administration of stool from a healthy donor to a patient with clinical signs of GIT disease in an attempt to correct dysbiosis (12). Fecal microbiota transplantation has had a 90% success rate in treating humans with recurrent Clostridioides difficile infection as a last-resource therapy (13). In humans, FMT is most successful to treat colitis when performed via enema compared to gastric administration, mainly because of the influence of gastric pH, digestive enzymes, and time of transit to reach the colon (14). FMT has been recommended for horses (15), but if delivered by enema, their long small colon will likely preclude the bacteria from reaching the compartments mainly affected in cases of diarrhea (i.e., cecum and large colon) (15). However, to the authors’ knowledge, this has not been investigated. In addition to the aforementioned factors decreasing efficiency of gastric administration of FMT in horses, bacteria must also pass through the fermentation and increased transit time occurring in the cecum, before arriving at the large colon.

McKinney et al (16) evaluated the effect of FMT in 5 geriatric horses with colitis or with diarrhea that developed inhospital, and reported that the fecal microbiota of 3 horses returned to a healthy state 24 h following the last FMT. Also, Dias et al (17) reported clinical recovery of 4 horses with postoperative acute colitis that had received 1 FMT, but no microbiota analyses were performed. Another limitation of these studies is that neither included a control group of diarrheic horses not receiving FMT, which precludes attribution of improvements to FMT. Furthermore, a recent study by McKinney et al (18), in which horses with colitis were treated with FMT, reported improvements in diarrhea scores and increased alpha-diversity compared to a control group of horses with colitis treated with standard care. However, the influence of location was present, as the control group was treated at a separate hospital (18). Although FMT has been empirically used by equine practitioners for decades, controlled studies demonstrating faster clinical recovery and correction of dysbiosis associated with the procedure remain to be performed. Therefore, there is a current gap in understanding the pathophysiology of equine undifferentiated colitis (cases of diarrhea without an established etiological agent or cause) and studies investigating microbiota manipulation are essential to improve current treatment options (10).

The objective of this study was to characterize the fecal microbiota of horses with diarrhea before and after FMT. We hypothesized that the microbiota of horses with diarrhea would change and would resemble the microbiota of the donor 48 h after receiving FMT.

Materials and methods

Selection of recipient horses

Horses with acute and chronic diarrhea admitted to the Universidade Estadual de Londrina and to the Universidade Filadelfia (Unifil), both in the city of Londrina, Paraná State, Brazil between 2013 and 2016 were included in the study. All horse owners provided written informed consent for participation in the study. Inclusion criteria included presence of diarrhea. No exclusion criteria were used. Before administration of FMT, diarrhea persisted regardless of treatment performed, such as fluid therapy, antibiotics, probiotics, and anthelmintic administration. Table 1 summarizes the studied population including previous treatments received, as well as information about the donor horses.

Table 1.

Description of the 6 horses receiving FMT as part of treatment for diarrhea.

Recipient	1	2	3	4	5	6
Age (y)	6	7	7	20	8	8
Sex	Male	Male	Female	Male	Male	Male
Breed	Quarter Horse	Quarter Horse	Quarter Horse	Mangalarga	Quarter Horse	Quarter Horse
Hospital of admission	UELVH	UELVH	UELVH	UVH	UELVH	UELVH
Reason for hospitalization	Decubitus after 36 h of transportation	History of intermittent chronic diarrhea for 1 mo	History of intermittent chronic diarrhea for 6 mo	1 wk	6 mo	3 wk Difficulty in locomotion, loose stools, dark urine
Clinical examination	– Rectal temperature 40.2°C – Apathy – Intestinal hypomotility – Black and fetid profuse diarrhea – Digital pulse in forelimbs	– Apathy – Congested mucous membranes – Intestinal hypermotility – Liquid and fetid feces	– Rectal temperature 38.5°C – Tachycardia – Apathy – Liquid feces – Lean body condition	– Dark and liquid feces – Lean animal – Periodontal disease – Lethargic with anorexia	– Intestina hypermotility – Pasty feces	– Rectal temperature 37°C – Apathy – Intestinal hypermotility in left and hypomotility in right quadrants – Lean body condition – Dark urine – Dark and fetid feces
Treatments upon hospitalization	– Fluid therapy – Hydroxyethyl starch – Flunixin meglumine – Metronidazole – Procaine penicillin	– Fluid therapy – Flunixin meglumine – Ceftiofur – Omeprazole – Fenbendazole	– Fluid therapy – Flunixin meglumine – Fenbendazole	– Oral antiparasitic (ivermectin and praziquantel) – Oxytetracycline	– Fenbendazole – Dexamethasone	– Fluid therapy – Hydroxyethyl starchb – Flunixin meglumineb – Phenylbutazoneb – Ceftiofurb – Metronidazoleb – Gentamicinb – Omeprazoleb – Ranitidineb
Days of hospitalization	6	11	11	7	20	18
Date of donor feces collection	2015-02-27	2013-01-17	2013-02-06	2015-04-18	2016-06-10	2016-04-20
Number of administrations	1	1	1	1	7	5
Outcomea	– Death 4 d after FMT – Diarrhea episodes intensified	– Survived – Improved stool consistency for minimum 5 d	– Survived – Improved stool consistency for minimum 9 d	– Survived – Improved stool consistency for minimum 5 d	– Survived – Improved stool consistency for minimum 3 mo	– Death 2 wk after FMT – Severe abdominal discomfort 6 d after FMT – Septic peritonitis – Cecal compression and rupture

UELVH — UEL Veterinary Hospital; UVH — Unifil Veterinary Hospital.

^aConsistency of feces improved until the end of hospitalization for Recipients 2, 3, and 4. Telephone contact with the client for Recipient 5, 3 mo after hospitalization, reported normal fecal consistency.

^bFlunixin meglumine [1.1 mg/kg body weight (BW), IV, q12h]; phenylbutazone (2.2 mg/kg BW, IV, q12h); ceftiofur (5 mg/kg BW, IM, q12h for 8 d); metronidazole (15 mg/kg BW, IV, q12h); gentamicin (6.6 mg/kg BW, IV, q24h, for 6 d); omeprazole (94 mg/kg BW, PO, during the entire hospital stay); ranitidine (1.5 mg/kg BW, IV, q8h for 2 d).

Selection of donor horses

Two clinically healthy horses (1 teaching animal at each institution) that were used as fecal donors had no history of intestinal disease and did not receive antimicrobials during the 6 mo before the study nor during the study period. The donor from the Universidade Estadual de Londrina was an ~ 10-year-old male mixed Paint horse and the donor from Unifil was a 2-year-old female Quarter Horse. Both donors were dewormed every 6 mo with ivermectin and praziquantel. Donor horses were on pasture for the duration of the study period. Feces from the donors were negative for Salmonella enterica, Clostridium perfringens, and Clostridioides difficile by culture and negative for parasitic eggs once at the beginning of the study period. Feces were collected from the donor before the first FMT was performed for each recipient.

Fecal microbiota transplantation and sample collection

The FMT protocol used in this study is similar to 1 previously described (15). Feces (1 kg) obtained from the donor per rectum was vigorously mixed in 5 L of warm water and filtered through a mesh. The fecal suspension was prepared approximately 15 min before each transplant, and was delivered by nasogastric tube, every 24 h for a maximum of 7 administrations, as indicated in Table 1.

Fecal samples from recipients were collected per rectum before FMT (Day 0), and 2 and 7 d after FMT. The amount of feces collected depended on the stool consistency of the recipient (fecal balls versus liquid feces). Feces were homogenized in a container and stored until further use. A fecal sample in the form of a fecal ball from the donor horse was also collected and homogenized before preparation of FMT for each recipient. Samples were stored in plastic sterile containers and frozen at −80°C within 2 h after collection until DNA extraction, for which 200 mg of homogenized feces were used.

Microbiota analysis

Total DNA was extracted using a commercial kit (E.Z.N.A. Stool DNA Kit; Omega Bio-Tek, Norcross, Georgia, USA) following manufacturer’s instructions. Polymerase chain reaction amplification of the V4 region of the 16S rRNA gene was performed using the primers forward S-D-Bact-0564-a-S-15 (5′-AYTGGGYDTAAAGNG-3′) and reverse S-D-Bact-0785-b-A-18 (5′-TACNVGGGTATCTAATCC-3′)(19). Sequencing was performed using an Illumina MiSeq platform (Illumina, Foster City, California, USA) for 250 cycles from each end at the University of Guelph Genomics Facility, following reported methods (8).

Sequence analysis and statistical analysis

Bioinformatic analysis was performed using the software mothur following the Standard Operating Procedure previously described (20). Statistical analyses were based on normality of data distribution (Shapiro-Wilk test). Sequencing reads were clustered in Operational Taxonomic Units (OTUs) at 97% similarity. The total number of OTUs per sample and the Simpson’s index were used respectively as measures of richness (number of species present in a community) and diversity. Those indices were compared between donor horses and FMT recipients using an unpaired 2-tailed t-test. For characterization of beta-diversity (comparison of taxonomic composition between each sample), the Jaccard index and the Yue and Clayton index were used to evaluate community membership (that considers presence or absence of each taxa) and structure (that considers also how often that bacteria appeared in the analysis), respectively. A 2-dimensional Principal Coordinate Analysis (PCoA) plot was generated to visualize the similarity between samples. Analysis of molecular variance (AMOVA) was used to determine significance of clustering between healthy donor horses and FMT recipients.

Community structure was further visualized through bar charts representing the relative abundance of the main genera in each horse. The linear discriminant analysis effect size (LEfSe), which uses a non-parametric factorial Kruskal-Wallis with subsequent unpaired Wilcoxon test was used to detect significant differences in relative abundances with respect to each group of interest (donors and patients), followed by a linear discriminant analysis (LDA cut-off = 2.0) to estimate the effect size of each differentially abundant group (21). A P-value < 0.05 was used to determine significance.

Results

Horses

Recruited horses were between 6 and 20 y of age with a history of acute and chronic diarrhea (1 wk to 6 mo). Five horses (Recipients 1, 2, 3, 5, and 6) were admitted to the UEL Veterinary Hospital and received an FMT originating from the donor housed at the same institution. One horse (Recipient 4) was admitted to the Unifil Veterinary Hospital (UVH) and received an FMT from the other donor, housed at the UVH institution. Signs of diarrhea were improved in 4 out of 6 horses within the time of hospitalization, whereas 2 out of 6 horses did not survive (Table 1). Feces were collected from the donor before the first FMT was performed for each recipient to obtain a representative sample for taxonomic profiling. Therefore, a total of 5 fecal samples were collected from the donor at the UEL Veterinary Hospital (1 fecal sample before the first FMT administration to each of the 5 recipient horses), and 1 fecal sample was collected from the donor at the UVH (1 fecal sample before the first FMT administration to the recipient horse at UVH). Fecal samples from recipients were collected before FMT (Day 0), and 2 and 7 d after FMT, for a total of 3 fecal samples collected from each recipient.

Challenges associated with unawareness of staff resulted in a missing sample (Day 7) from Recipients 1 and 4, whereas fecal samples from Recipient 6 were taken on Days 3 and 9 (instead of 2 and 7).

Microbiota analysis

A total of 114 015 reads from 21 samples passed all quality filters and were assigned into OTUs. To normalize the number of reads across all samples, a subsample of 1046 reads per sample was used for analysis (coverage average and standard deviation: 66.34% and 0.07%, respectively), which resulted in the exclusion of the sample from Donor 2 due to low yield of reads. The Donor 2 sample was therefore removed from the analysis.

There was a significant difference in richness (observed OTUs) between donors and FMT recipients (P < 0.05, unpaired t-test of donors versus recipients on all days, Figure 1A). Fecal microbiota transplantation was not associated with an increase in richness (P = 0.96, paired t-test of recipients on D0 versus D2, P = 0.77, D2 versus D7, P = 0.73, D0 versus D7) in diarrheic patients. The Simpson’s index was not significantly different when comparing donors to recipients, or when comparing recipients to each other (Figure 1B).

Figure 1.

Alpha diversity indices. A — Observed OTUs of healthy donor horses and FMT recipients. B — Simpson’s index of healthy donor horses and FMT recipients. Statistical analysis was performed using Student’s t-tests. Bars represent mean and SD.

* P ≤ 0.05.

There was no significant difference in beta-diversity membership (P = 0.15) or structure (P = 0.20) between donors and recipients, with the microbiota of each recipient remaining clustered together before and after FMT (Figure 2). In addition, the donors had similar microbiota membership (Figure 2A) and clustered with the microbiota of Recipient 3, which likely precluded achieving statistical significance.

Figure 2.

Principal coordinate analysis (PCoA) of bacterial communities in the feces of healthy donor horses and FMT recipients. Bidimensional representation of the principal coordinate analysis of bacterial communities’ membership addressed by the Classic Jaccard analysis (A) and structure addressed by the Yue and Clayton analysis (B). D0 represents the bacterial community before FMT administration, and D2 and D7 represent the bacterial community 2 and 7 d after FMT administration, respectively. Donor samples 1, 3, 5, and 6 represent samples from the donor at the UEL Veterinary Hospital (UVH), whereas donor sample 4 represents the donor at UVH. Samples from FMT recipients cluster together in membership and structure.

The relative abundances at the phylum and genus levels in each group at the various sampling times are shown in Supplementary Figure 1, which is available from the authors upon request. Firmicutes was the most abundant phylum among recipient horses (52.7%), followed by unclassified bacteria (23.0%), Verrucomicrobia (5.8%), Bacteroidetes (5.8%), Spirochaetes (5.0%), and Proteobacteria (3.3%). The most abundant taxa classified at lower taxonomic levels include unclassified Ruminococcaceae, unclassified Lachnospiraceae, unclassified Firmicutes, unclassified Clostridiales, Treponema, and unclassified Verrucomicrobiaceae.

Over-representations in taxa between donors and recipient horses were compared using LEfSe (Figure 3). There was overrepresentation in diarrheic horses of the genus Intestinimonas, unclassified Lactobacillales, Lactobacillus, and Streptococcus, when compared to the donors, which had an increased relative abundance of the genus Saccharofermentans.

Figure 3.

Linear discriminant analysis effect size (LEfSe) analysis. LEfSe analysis indicating genera that were signifi cantly differentially abundant between healthy donor horses (class: Donor) and FMT recipients (class: Diarrhea).

Discussion

This study failed to detect significant changes in the fecal microbiota of 6 horses with acute and chronic diarrhea after treatment with FMT, suggesting that the most abundant bacteria present in the fecal suspension of the healthy donor horses failed in colonizing the lower GI tract of the recipient horses during the study period. Generally, a decrease in alpha-diversity (richness and diversity) is observed in horses with gastrointestinal diseases (22,23). A previous study reported that FMT administered to 5 geriatric horses with diarrhea for 3 consecutive days led to a trend for increased alpha-diversity and to compositional similarities in 3 of these horses to the microbiota of healthy horses 24 h after the last FMT (16). No controls were used, however, precluding conclusions from being made. In addition, another study in which horses with colitis were treated with FMT reported a greater improvement in diarrhea score and greater alpha-diversity compared to control horses with colitis treated with standard care (18). However, the control group was treated at a separate hospital, making location an unaccounted variable (18). In contrast, the alpha-diversity in the fecal microbiota of the recipient horses in this present study remained unchanged after FMT. Furthermore, the microbiota of each recipient remained clustered together, indicating perpetuation of the dysbiosis and a lack of effect of FMT on altering the composition of the fecal microbiota in the recipients. Although the PCoA representing similarities in microbiota compositions demonstrated a clear separation between donors and diarrheic patients, the lack of statistical difference was likely due to the membership of Recipients 3 and 4 resembling the donor’s microbiota membership even before FMT treatment. This finding was unexpected because marked differences are present in microbiota composition between healthy and diarrheic horses (5). Normal microbial profiles in humans with chronic intestinal inflammation has been demonstrated (24), but this remains speculative in horses and other environmental factors (i.e., diet and housing) should be considered to explain this microbiota similarity (25).

The composition of the microbiota of Recipient 1 and of Recipient 6 were different from that of other recipients (Figure 2). Interestingly, those were the only horses that died, but also that had been treated with metronidazole, suggesting a more severe manifestation of the disease. Recipient 1 was treated with metronidazole on the same day of FMT treatment, and was euthanized 4 d later, whereas Recipient 6 was treated with metronidazole 1 wk after FMT treatment and was euthanized 2 wk later. Antimicrobials cause changes in the gut microbiota of horses. For instance, treatment with various antimicrobials caused a significant decrease in richness and diversity along with significant compositional changes (8,26). However, in this study it is not possible to attribute the differences in the microbiota of those 2 patients to the use of metronidazole because they also had more acute and severe episodes of diarrhea, and neither survived. The presence of specific fecal microbiota changes and detection of markers able to predict mortality or to be used as prognostic tools deserve further attention and require large multicenter studies.

Microbiota composition was similar in samples obtained from the same donor horse (donor samples 1, 3, 5, and 6), indicating that the microbiota was stable during the study period. Although the fecal sample from Donor 2 was excluded due to a low number of reads, the sample originated from the same donor as cases 1, 3, 5, and 6 which varied little over time, and therefore it would be expected that the microbiota would cluster along with those other samples. The microbiota of healthy horses might have little variation over a 1-year interval (27), which is especially important considering the lack of information regarding selection of donors for FMT in equine patients. One study comparing the fecal microbiota of healthy young horses to that of geriatric horses did not report significant differences between the 2 groups, possibly excluding age as a factor for donor selection (28). In humans, the high variation in patient response to FMT treatment for chronic diseases has led many to suggest that treatment success may rely on the diversity and composition of the donor’s stool (29). Some human studies have also alluded to the presence of a “super donor” effect, in which some stool donors are more successful in treating recipients than others, suggesting that some microbiota are more likely to successfully colonize recipients (30), which may have an increased chance at being satisfactorily colonized through pooling donor stools (31). For instance, successful donors in treating human patients with ulcerative colitis had high relative abundances of Clostridium clusters IV and XIVa (31), which are known to be important for intestinal homeostasis by producing large amounts of SCFAs (32). Interestingly, these taxa have been suggested to be part of the core microbiota in horses and are important contributors to gut health (33). Therefore, future studies evaluating the impact of pooling feces from multiple donors are required.

The use of the oral route of administration for FMT has been shown to limit its effectiveness in humans (14), and the same is likely true in horses. Bacteria given through a nasogastric tube must survive the acidic stomach, enzymatic actions in the small intestine and fermentation in the cecum before reaching the large colon. Nevertheless, this is the current method of choice for FMT delivery in horses (16,17). Another factor possibly influencing FMT efficacy in horses is exposure of the donor stool to oxygen. Human fecal bacterial viability decreased to 19% when exposed to oxygen during homogenization (34), whereas bacterial viability remained at 50% when prepared under anaerobic conditions (34).

The search for markers that can be associated with intestinal dysbiosis or health has been the focus of many researchers in human (35) and veterinary medicine (36). In this study, Streptococcus was associated with diarrheic horses, which was also observed in another study (37). Contrary to this study, Saccharofermentans was reported to be associated with healthy horses (22). Although Intestinimonas was associated with diarrheic horses in this study, to the authors’ knowledge, this has not been reported in any study of horses to date. Of interest, the present study reports that Lactobacillus spp. is associated with horses with colitis, consistent with previous findings (5). Similarly, a study investigating the luminal and mucosal microbial communities also reported Lactobacillus spp. as being associated with horses with colitis when compared to clinically healthy horses (38) and the Lactobacillales order was increased in diarrheic horses (37). This genus is normally associated with health and even used as probiotics in horses, but one study reported that Lactobacillus pentosus WE7 increased severity of diarrhea in foals (39). It is not clear if this Lactobacillus increase in diarrheic horses relates to environmental changes or to a recovery process to return to a healthy state.

The main limitation of this study relates to the small number of samples and no long-term evaluation. In addition, the group of recipients in this study was heterogenous, as clinical manifestations, age and treatment with antimicrobials were variable, as was the number of FMT administrations. Furthermore, only fecal samples were evaluated with the assumption that those are good representations of more distal changes occurring in the intestinal tract (i.e., cecum and colon). Although this has been shown in healthy horses, 1 study demonstrated no difference between the cecal and colonic microbiota of horses with colitis, possibly due to inflammation-driven changes and increased motility (38). In addition, the exclusion of the fecal sample from Donor 2 limited a comparison with the fecal microbiota of its corresponding recipient (Recipient 2). Nevertheless, important information is presented in this study, such as the clustering of samples according to the individual and, most importantly, the lack of changes after FMT.

In conclusion, the fecal microbiota of diarrheic horses was stable over a 7-day interval and not detectably affected by FMT. This study adds to the current knowledge and highlights the need for more science-based evidence to support the use of FMT to treat equine diarrhea. Larger multicenter controlled studies are necessary to demonstrate the clinical benefits of FMT in horses. Further research should focus on improving the FMT protocol to be used in horses by using anaerobic conditions for FMT preparation and by investigating the influence of the donors’ microbiota composition (or the use of multiple donors) on the ability to successfully colonize diarrheic horses. CVJ