Edited by: Guillermo Tellez, University of Arkansas, United States
Reviewed by: Victor Manuel Petrone-García, National Autonomous University of Mexico, Mexico; Roberto Senas Cuesta, University of Arkansas, United States
This article was submitted to Animal Nutrition and Metabolism, a section of the journal Frontiers in Veterinary Science
†These authors share last authorship
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Obesity is endemic in human populations in the western society, and with mounting evidence that the intestinal ecological environment plays a major role in its pathogenesis, identification of therapies based on intestinal microbiota modulation are gaining attention. Obesity in companion animals is also a common clinical problem. We set out using a multidimensional approach, to determine the effectiveness and safety of a weight loss program for horses incorporating diet restriction and exercise. In addition, we aimed to investigate the effect of this program on the overall intestinal health of overweight sedentary horses. The investigation comprised of a randomized, controlled, 6-week study of 14 overweight sedentary horses and ponies who were blocked for age, gender, and breed (controls
Human obesity is recognized as endemic worldwide (
Obesity is associated with further chronic health problems, including metabolic syndrome. In human metabolic syndrome, obesity and associated increased blood pressure, hyperglycemia, excess abdominal fat, and abnormal blood cholesterol and triglyceride concentrations result in an increased risk of cardiovascular disease and diabetes type-2 (
In order to combat obesity and its associated comorbidities, investigations are ongoing into predisposing factors and potential novel treatments. In humans, this has led to an increasing focus over the past decade on the possible influence of the intestinal microbiota on obesity (reviewed in Castaner et al. (
The current advice for treatment of EMS is similar to that provided for humans, i.e., diet restriction and controlled exercise (
We hypothesized that a weight loss program, through restricted diet and exercise, could not only induce weight loss and improve cardiovascular function in overweight horses, but could also alter the intestinal microbiota to improve gastrointestinal health. The objective of this study was to investigate effects of weight loss on morphological measurements, cardiovascular parameters, fecal microbiota, and fecal metabolome in a cohort of clinically normal but overweight horses and ponies. This was to be achieved using a multiomics approach combined with morphological assessment, cardiovascular diagnostics, and clinical observations.
This was a longitudinal study using client-owned, rather than experimental animals; therefore, the number of animals available were constrained. As we intended to measure a range of parameters related to body condition, physiological parameters, and gut health, we chose to use an anticipated change in body weight as a significant parameter for estimation of required group size.
We estimated that for a study with power = 90%, we could reliably measure an incremental body weight reduction of 5 kg, i.e., a weekly reduction of 1% in an average horse of 500 kg, and that the SD of the reduction would be no more than 4 kg. Hence, the SD units would be 1.25, and the required group size is
The study was carried out on a total of 14 horses. The study population was recruited from a rescue charity (
Outline of study population of 14 horses including horse number, sex, breed, and age.
5 | M | Cob pony | 5 | 1 | M | Cob | 10 |
7 | M | Cob pony | 7 | 2 | M | Cob pony | 4 |
8 | M | SB | 15 | 3 | M | Pony | 17 |
9 | M | Pony | 5 | 4 | M | SB | 5 |
11 | F | Cob | 10 | 13 | M | Pony | 5 |
12 | F | Cob pony | 4 | 6 | F | Cob pony | 5 |
14 | F | SB | 15 | 10 | F | Cob pony | 4 |
The 12 horses recruited from the rescue charity were treated with anthelmintic (Equest, Pramox 0.4 mg/kg PO, Zoetis Broomhill Rd, Tallaght, Dublin, Ireland) and vaccinated against equine influenza virus and tetanus (Equilis Prequenza TE, MSD Red Oak North, South County Business Park Leopardstown, Dublin 18, Ireland) 2 weeks prior to their arrival on the farm, as per the UCD research farm regulations for any new horses entering the farm.
For the week leading into the trial, the horses were housed on site for acclimatization purposes. The horses were fed 2% of their body weight (BWT) with fresh weight grass hay that was soaked for 1 h prior to administration. They were also made accustomed to an automated horse walker, together with sand paddocks that would be used for turn out purposes during the trial.
On arrival, all horses underwent a full clinical examination and lameness examination. Throughout the study, they were monitored daily for food and water intake, fecal consistency, urine output, demeanor, exercise tolerance, and any other observations. They were visually assessed for abnormalities and monitored while exercising on the horse walker for any development of an injury. A trot up examination, which included examination on the turn, was performed weekly to ensure soundness throughout the study. For any horse showing clinical abnormalities, exercise was stopped, and a full clinical examination was performed.
The horses were randomly assigned to control (C) and treatment (T) groups, blocked for age, gender, and breed. The herd remained on site for 45 days incorporating the acclimatization and study periods. The horses were stabled in individual loose boxes on sawdust bedding for the duration of the study. All horses were turned out in pairs for at least 25 min each day, in small sand paddocks to avoid interference with dietary restriction and enable natural social behavior.
The C group were kept on a maintenance diet of 2% BWT, as fresh weight hay, which was equivalent to 1.66% dry matter intake (DMI) of pre-soaked grass hay based on a DM of 83%. The T group was restricted to 1.4% BWT as fresh weight grass hay, as per guidelines outlined in the recent EMS consensus statement (
All horses were weighed weekly, and their DMI was adjusted accordingly. The target weight loss for the T group was 1% BWT per week, while the C group was to maintain initial BWT; therefore, the diet was modified weekly in line with weight fluctuations.
The daily DMI was divided into two feeds to be given morning and evening. The hay was placed in slow-feeder hay nets (small gauged hay nets) and weighed individually using hanging scales to calculate DM weight. The filled hay nets were then soaked in individual containers, with weights keeping the nets submerged, for an hour, and allowed to drain for 30 min before feeding. To compensate for loss of nutrients associated with soaking the hay, a multivitamin, mineral, and amino-acid supplement (EquitonTM Troytown Grey-Abbey Pharmacy, Kildare, Ireland) was administered, as per manufacturer's instructions (2 ml/50 kg), in 300 g of high-fiber chaff mix (Hi-Fi lite, Dengie Crops Ltd, Essex, England) once daily and incorporated into the total calculated DMI.
Exercise was carried out on an automated horse walker under continuous supervision with alternating directions every second day. The C group was exercised daily at walk to mimic foraging conditions ensuring that the total time in the walker matched that of the T group. The T group was enrolled in a 5-day weekly walk–trot exercise program, with two rest days. The T group was divided into two subgroups created to accommodate for the length of stride, i.e., horses under 148 cm (T-pony) and horses over 148 cm (T-horse). Exercise intensity was determined by observation, and increases in the exercise program were made in accordance with improved fitness levels and exercise tolerance (
Outline of the exercise regime of the treatment (T) group in minutes on automated walker.
Warm up | 5 | 5 | 5 | 5 | 5 | 5 |
Brisk walk | 15 | 15 | 15 | 15 | 15 | 15 |
Trot | 0 | 5 | 15 | 15 | 15 | 15 |
Warm down | 5 | 5 | 5 | 5 | 5 | 5 |
Physical measurements were assessed and recorded before the commencement of treatment and again at the end of the study, as appropriate, by a European Board of Veterinary Specialisation (EBVS) equine internal medicine specialist who was blinded to the group allocation. These included body weight (BWT), body length (BL), height at the withers (HW), body condition score (BCS), cresty neck score (CNS), and measurements of the neck (NC), girth (GC), and waist (WC) circumference. Measurements were taken in triplicate, and the mean value was calculated.
BWT was calculated using an electronic weighing system and was recorded weekly to allow for diet adjustment. BL was calculated from the point of the shoulder to the point of the ischium using a soft tape with 1-cm increments. HW was calculated using a wooden horse measuring stick. BCS was calculated by visual assessment and palpation using a nine-point scale, i.e., 1 (poor) to 9 (extremely fat) (
Resting blood pressure (BP) was measured non-invasively using a portable veterinary blood pressure monitor (Cardell Veterinary Monitor 9401; Sharn Veterinary Inc., Florida, USA). SV 8 (9 cm) and SV 10 (12 cm) cuffs were used, as appropriate for the size of the horse/pony to provide a snug fit with the hook and loop sections fully engaged. The cuff was applied to the base of the tail in the standing horse or pony to approximate the level of the right atrium. Mean blood pressure measurement was repeated five times at each recording time, and the average value was calculated. Clear outliers were discarded.
Subcutaneous fat thickness (SFT) measurements were made by a European College of Veterinary Diagnostic Imaging (ECVDI®) specialist who was blinded to the group allocation. If the horse's demeanor allowed, consistent with handler safety, the area was clipped, sprayed with 70% alcohol solution, and transmission gel applied. When clipping was not possible, large amounts of the alcohol solution were used to obtain images of the diagnostic value. The transcutaneous B-mode ultrasound examinations were performed using a GE Logiq e ultrasound machine (IVM imaging Ireland, Gormanston, Co. Meath), equipped with a linear probe (12L-RS 5–13 MHz). A technique described in a previous study (
Fecal samples were collected as per recommendations for microbiota analysis outlined by (
Initial processing of samples was carried out within 6 h of collection. The fecal ball was removed from the container, and a 10-g sample was taken from the center of the ball (
Fecal samples for microbiota analysis were processed as described previously (
The sequences obtained were filtered on the basis of quality (removal of low-quality nucleotides at the 3′ end and removal of window 10 nt with low average quality) and length [removal of sequences with <200 bp with prinseq as per the Schmider and Edwards protocol (
Fecal water was extracted from fecal samples as previously described (
The SPSS statistical software was used to estimate normality using the Shapiro–Wilk test. Two-tailed Student
Metabolome data analysis was performed using both SIMCA (multivariate) and MetaboAnalyst (
Microbiota analysis was performed using downstream analyses, and graphical outputs of the 16S rRNA results were generated with various packages in R (
All 14 horses recruited to the study, completed the trial.
Over the 6-week period, there were four horses, Horse 1 (C group) and Horses 7, 11, 13 (T group), that presented with signs of clinical disease. These individuals were treated by changes in environmental management, dietary management, or exercise regime (see
A summary of the horses who presented with signs of clinical disease during the 6-week trial period.
1 | Control | 1 | Increased respiratory effort at rest | Moved to outdoor stable |
7 | Treatment | 2 | Acute moderate weight loss, lethargy, and transient soft but formed fecal consistency | Increased feed intake to 2% BWT DMI |
11 | Treatment | 1 | 2/5 grade (AAEP) lameness right fore associated with poor foot balance | Decreased level of exercise to control group level |
12 | Treatment | 3 | 3/5 grade (AAEP) lameness right fore with tissue swelling and heat associated with the medial mid cannon area | Rehabilitation program combined with daily local hydrotherapy |
DMI was adjusted weekly in both groups. This ensured that there was no change in BWT of the control group and that the target of 1% BWT loss was obtained weekly in the treatment group. Therefore, the range of restriction on DMI in the treatment group ranged from 1.3 to 2% BWT DMI in accordance with weight loss or gain. Hay from a second source was introduced incrementally into the diet starting on day 18 with the inclusion of 10% of the “new” hay in the total DMI, increased by 10% increments every 2 days so as to reduce the risk of intestinal upset.
In the treatment group, five of the seven horses participated in the prescriptive exercise program. As mentioned above, two horses sustained injuries and, thus, had modified exercise regimes. Horse 11 joined the control exercise program. Horse 12 was enrolled in a rehabilitation program that began as 10 min walking in hand twice daily on day 15 with a gradual increase in time to include trotting for 10 min twice daily by day 25 (
There was a significant loss of BWT in the treatment group (
A summary of the significant morphological parameter findings between week 1 and week 6 of the study.
1 | 640 | 635 | −0.8% | 226.33 | 226.67 | 0.15% | 92.2 | 86.4 | −6.29% |
2 | 378 | 382 | 1.1% | 182.67 | 186.33 | 0.02 | 111 | 82 | −26.13% |
3 | 320 | 320 | 0.0% | 189.67 | 196.00 | 3.34% | 112.6 | 86.4 | −23.27% |
4 | 496 | 499 | 0.6% | 198.00 | 205.33 | 3.70% | NT | NT | NT |
6 | 568 | 576 | 1.4% | 215.33 | 222.00 | 3.10% | 99.4 | 87.8 | −11.76% |
10 | 408 | 408 | 0.0% | 194.00 | 197.00 | 1.55% | 108.4 | 85.2 | −21.04% |
13 | 372 | 373 | 0.3% | 189.33 | 192.67 | 1.76% | 91.2 | 91 | −0.22% |
5 | 453 | 430 | −5.1% | 187.67 | 183.67 | −2.13% | 116.4 | 90.6 | −22.16% |
7 | 358 | 338 | −5.6% | 192.67 | 185.00 | −3.98% | 116.8 | 56.2 | 51.88% |
8 | 528 | 511 | −3.2% | 209.67 | 209.33 | −0.16% | 107.4 | 75 | −30.17% |
9 | 260 | 249 | −4.2% | 171.33 | 165.33 | −3.50% | 92.2 | 85.8 | −6.94% |
11 | 656 | 633 | −3.5% | 228.00 | 225.00 | −1.32% | 130 | 117 | −10.31% |
12 | 484 | 470 | −2.9% | 209.00 | 202.33 | −3.19% | 109.2 | 91.6 | −16.12% |
14 | 608 | 586 | −3.6% | 237.67 | 226.00 | −4.91% | 82.8 | 67.2 | 18.84% |
There were no significant differences found between week 1 and 6 in subcutaneous fat measurements at any location in either treatment or control group (
Fecal samples were collected in weeks 1, 3, and 6 at the same time on the same day (Wednesday between 8 a.m. and 10:30 a.m.), with the exception of one sample from horse 9 in week 1, which was collected at 4 p.m. on Wednesday afternoon in week 1.
The analysis of the fecal microbiota showed general similarities at the phylum, family and genus level between the treatment and control groups across all time points (
The mean percentage relative abundance of the major phyla
However, when comparing diversity within groups over time, a significant increase in alpha diversity, Fisher Alpha index, and Richness index, between TP1 and TP6 in the T group (
Alpha diversity was measured by five indices of diversity: Richness, Fisher Alpha, Simpson, Shannon, and Pielou Evenness. The treatment group (
Significant differences (
The beta diversity illustrated using an NMDS graph; the treatment and control groups are shown at each timepoint and are color coded. The treatment group (
Assessment with linear discrimination analysis (LDA) effect size >3 and LefSe (Wilcoxon test) showed statistically significant differences when respective groups at specific time points were directly compared. More specifically, there was a significant reduction in the relative abundances of the families Eubacteriaceae (
Linear discriminant analysis effect size (LEfSe) was performed in order to discover specific bacterial biomarkers associated with control and treatment groups. The treatment group (
The NMR metabolomics data was successfully acquired from all fecal samples at each sampling timepoint. A representative NMR spectrum is displayed in
The multivariate analysis revealed that there was no major impact on the overall fecal metabolomic profile that could be attributed to treatment, i.e., the exercise and diet program (see
PCA plot of the fecal water NMR data color coded according to group (treatment and control). The treatment group (
Univariate data revealed changes in the metabolome profile over the study period irrespective of group, and a wo-way ANOVA analysis revealed that there were specific metabolites (dimethylsulfone and methylamine) with a significant factor effect (control/treatment) (
A two-way ANOVA analysis of metabolites taking into account the sample timepoint (Time) and the effect of the group (Factor), i.e., control vs. treatment group.
3.1405 | Dimethylsulfone | 0.006449 | 0.41649 | 2.39E−07 | 0.000835 | 0.30133 |
3.1415 | Dimethylsulfone | 0.008492 | 0.41649 | 9.72E−07 | 0.001501 | 0.39066 |
3.1395 | Dimethylsulfone | 0.010088 | 0.41649 | 1.29E−06 | 0.001501 | 0.42122 |
3.1425 | Dimethylsulfone | 0.011623 | 0.41649 | 8.11E−06 | 0.006362 | 0.54783 |
3.1435 | Dimethylsulfone | 0.019075 | 0.41649 | 6.61E−05 | 0.017793 | 0.69828 |
3.1385 | Dimethylsulfone | 0.048665 | 0.41649 | 4.58E−05 | 0.014563 | 0.54441 |
2.7765 | Unknown | 0.05342 | 0.41649 | 0.000239 | 0.049286 | 0.44754 |
2.7755 | Unknown | 0.058283 | 0.41649 | 0.000185 | 0.04324 | 0.49582 |
2.4135 | Unknown | 0.16904 | 0.41649 | 0.00022 | 0.048163 | 0.37797 |
2.5945 | Methylamine | 0.36387 | 0.52756 | 1.70E−05 | 0.008492 | 0.57962 |
2.5955 | Methylamine | 0.39142 | 0.54966 | 2.39E−05 | 0.009409 | 0.64331 |
2.5935 | Methylamine | 0.47239 | 0.61876 | 1.09E−05 | 0.006362 | 0.64923 |
2.5965 | Methylamine | 0.51512 | 0.65026 | 3.00E−05 | 0.010497 | 0.74757 |
2.5905 | Methylamine | 0.64991 | 0.74924 | 0.000141 | 0.035276 | 0.75996 |
2.5975 | Methylamine | 0.72159 | 0.7947 | 5.70E−05 | 0.016639 | 0.78131 |
2.5925 | Methylamine | 0.73149 | 0.79781 | 9.65E−06 | 0.006362 | 0.8052 |
2.5915 | Methylamine | 0.90805 | 0.92766 | 2.42E−05 | 0.009409 | 0.83179 |
Heatmap analysis of the treatment and control groups at three timepoints during the study. The treatment group (
The relationship between microbiota and metabolites in the feces, in the control and treatment groups over the three different time points, was visualized through Pearson network images of correlation. The total NMR profile and OTU readings for each sample was used in this analysis, to identify any significant correlations between fecal microbiota and metabolites in both groups at each time point. A correlation analysis cutoff was used (
A circos plot with a correlation cutoff of
The significant correlations between OTU 3778 and associated metabolites were only found in the T group at TP 3. The correlation between the other three OTUs (OTU 998, OTU 164, and OTU 748) was found at all timepoints in both groups; however, the association was more evident in the T group.
Obesity is endemic in human populations in the developed world, and this has spread into our domesticated animal population with reports of over 50% of the pet population being overweight or obese (
Dietary restriction and controlled exercise are the foundations of treatment of obesity in both humans and horses. In this study, significant weight loss was achieved in a group of overweight horses. This is in line with similar studies that implemented both diet and exercise as part of a weight loss program (
Although significant weight loss and decreased WC were recorded in the treatment group, these changes were not reflected in the remaining morphological parameters measured including BCS or the subcutaneous fat ultrasound measurements. This is consistent with another study of 12 overweight/obese horses that were submitted to exercise-induced weight loss but had no change in subcutaneous fat measurements (
It has been shown in both human (
Although the recommendations for dietary restriction in overweight horses is 1.4% BWT (
Due to the complexity of bacterial and host interactions, this study used a multiomic approach broadening the investigation to incorporate not only the structure of the microbial community but also the metabolic functions. This study analyzed the microbiota and metabolites found in the feces of overweight, sedentary horses undergoing a strict weight loss program compared with a control cohort maintained at their presenting weight.
There are a few studies investigating the effect of obese phenotype on the intestinal environment in horses, often using one sample point and a heterogenous population (
The composition of the major phyla found in equine feces in this study is consistent with other studies, dominated by Firmicutes, Bacteroidetes, and then the less dominant phyla of Spirochaetes and Fibrobacteres (
The taxonomic composition at the family and genus levels were broadly the same across the most abundant bacteria (
It was notable that there was an increased relative abundance of the families Eubacteriaceae and Pseudomondaceae in the C group. Eubacteriaceae has been associated with obesity in obese adults (
It is recognized that diet is a driving factor of microbial community composition in the animal kingdom (
In humans, high-level microbial diversity in the gut is frequently regarded as desirable, and it has been proposed that obese and overweight phenotypes are associated with decreased intestinal diversity (
The benefits of physical exercise are well-documented and have been proven in humans (
The metabolome analysis allows investigation of the functional changes associated with bacterial community and the potential changes that a weight loss program may have not only on the microbiota composition but also the associated metabolome. Although we did not observe a significant difference between the treatment and the control groups, both groups showed significant changes in metabolomic profile between the start and end of study period. This is most likely attributable to adjustment to the fiber-based diet over time.
To get a more nuanced view of microbe–metabolite–host interactions, this study used network analysis of correlations between the metabolites and OTUs found. The family Rikenellaceae was associated with urocanic acid at TP 3 in the T group. There were three other OTUs, which the analysis showed as linked to specific metabolites found, the most interesting being the family Ruminococcaceae, which was correlated to propionate.
Rikenellaceae is a saccharolytic bacteria that can ferment glucose to acid byproducts. In mouse studies, it has been shown to increase in obesity inducing high-fat diets and would be considered an obesity-related bacteria (
Although Ruminococcaceae is highly specialized in the degradation and fermentation of cellulose in short-chain fatty acids (SCFAs), they are more commonly associated with butyrate production (
In summary, the change in diversity of the samples in the T group between the start and end of the successful weight loss program is indicative of the effect of treatment on the fecal microbiota composition of horses. Moreover, the increased alpha diversity over time is linked with increase abundance of genus
In this study group, there was a mix of gender, breed, size, and place of origin, and although there was an acclimatization period, all of these factors can affect the fecal microbiota. The sample size was small, and further larger studies would be required to confirm the microbiota and metabolome findings. Although target weights were achieved, two of the seven horses in the treatment group were unable to participate fully in the prescribed exercise program, which may have affected the results. Further investigation into cardiovascular parameters through echocardiology as done by Heliczer (
As we reach for novel treatments to combat the ongoing upsurge of equine obesity, the manipulation of the intestinal microbiota is potentially a key target for future therapeutics. Therefore, we suggest longitudinal studies evaluating larger sample groups investigating the systemic effects of a weight loss program using a multiomic approach in tandem with investigation of insulin dynamics and adipokine levels. This approach could offer insights into the interaction between ongoing weight loss, increased insulin sensitivity, and the intestinal microbiome. Potentially, highlighting fecal biomarkers associated with weight loss resistance and insulin resistance could enable therapeutic modulation of the equine intestinal microbiota as has been shown to be effective in humans, such as diet supplementation (
In this study, it was shown that a controlled weight loss program using diet restriction and exercise can be successful and may contribute to an improvement in the intestinal health of the horses. The insights gained will be useful in designing further studies aimed at prevention and treatment of obesity and obese-associated disease in horses. With the overarching aim of the research to unravel the connections of a range of chronic inflammatory conditions linked to gut health in the horse, and, in line with the One Health paradigm, across species.
The data presented in the study are deposited in the ENA repository, accession number ERP127102.
The animal study was reviewed and approved by UCD Animal Research Ethics Subcommittee AREC-E-19-26-Mulcahy. Written informed consent was obtained from the owners for the participation of their animals in this study.
NW participated in the study design, execution, data analysis, and interpretation and preparation of the manuscript. RC participated in the study design and execution of the study. VG participated in the study design and execution of the study. VD and AP participated in the study design, performed the ultrasound examination of the horses, and preparation of the manuscript. GM participated in the study design, execution, data analysis and interpretation, preparation of the manuscript, and approved the final version of the manuscript. LB performed the metabolome analysis and contributed to the preparation of the manuscript. PC, FC and RC-R carried out the 16S rRNA gene sequence analysis, bioinformatic analysis, and contributed to the preparation of the manuscript. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We would like to thank My Lovely Horse Rescue for volunteering their horses for this study and their ongoing support of our research endeavors. We would like to say a special thank you to Caroline Barry, the equine manager at UCD research farm, for all her help throughout the study. We would like to thank Carel Le Roux and Neil Docherty of the Diabetes Complications Research Centre UCD for their helpful discussions and expertise.
The Supplementary Material for this article can be found online at: