This research was approved by the II Local Ethical Committee of the Wrocław University of Environmental and Life Science (Poland).

A total of 16 Welsh Ponies (eight mares and eight geldings) owned by private individuals were used for the experiment. The age of the animals ranged from 7 to 15 (10 ± 3) years and resided in two provinces of Poland: Lower and Upper Silesia. All animals were able to go on outdoor walks, spending most days outside the stables and exercising with moderate intensity.

Classification of the ponies was performed in two stages. Initially, 30 horses were included in the investigation. The ponies were screened for the study using the following steps: (i) extensive interview with the owners, (ii) measurement of body weight, (iii) estimation of body condition score (BCS) and crest neck scoring system (CNS) number, (iv) palpation and visual assessment the hoof capsule, (v) X-ray examination, (vi) measurement of resting insulin levels, (vii) combined glucose-insulin test (CGIT), and (viii) measurement of LEP concentration. During preliminary selection, all 30 animals were clinically examined (i-iv). Exclusion criteria of the ponies included: (a) normal body weight (x = 260 kg), (b) BCS < 6, CNS 0~1, (c) absence of divergent growth rings on the hoof (a symptom of chronic laminitis), and (d) pregnant mares, mares with foals, or young horses undergoing growth and development. Inclusion criteria were as follows: (a) overweight (x = 290 kg) or obese (x = 325 kg) animals, (b) BCS 7~8 (overweight~obese) or 8~9 (obese~extremely obese), (c) CNS 1~2 points or 3~5 points, (d) obese ponies with divergent growth rings on the hoof and (e) overweight animals with healthy hooves (no symptom chronic laminitis). Ponies with CNS 1 and that were overweight (BSC 7) were also incorporated into the study. Basis on the above criteria, 10 ponies were excluded (six with normal body weight, no divergent growth rings on the hoof, and CNS 1; one with normal body weight, CNS 1, and evidence of chronic laminitis; one pregnant mare; one 2-year-old horse, and one 8-month-old foal). Ultimately, 20 ponies were selected for further study (v-viii).

An X-ray examination of the 10 obese ponies enabled characterization of the coffin bone rotation and sinking that was associated with laminitis. The others ponies that were overweight were characterized by proper orientation of the coffin bone within the hoof capsule. Measurement of resting insulin levels allowed for further division of the animals into the appropriate groups. Ponies classified as healthy had insulin levels 6~20 uU/mL while animals with insulin concentrations 60~100 uU/mL were assigned to the experimental group. Despite the fact that the reference value of resting insulin is < 20 µU/mL, animals with the borderline result (i.e., 22~23 µU/mL) were excluded from the study. Next, 10 ponies with negative CGIT results were assigned 7~8 points on the BSC scale and 1~2 points on the CNS scale. LEP concentrations were categorized as normal (3~4 ng/mL) or high (> 7 ng/mL). Ponies with concentrations between 5.0~6.7 ng/mL were excluded from the study. Based on the above criteria, four ponies were excluded. Ultimately, the ponies were divided into experimental (group A, EMS ponies, n = 8) and control (group B, healthy ponies, n = 8) groups.

Body weight was measured using an electronic mobile Bosh equine weight scale (Bosch, Germany). The BCS was determined using a system developed by Henneke et al. [12] in which a numerical value for fat deposition ranging from 1 (poor) to 9 (extremely obese) is assigned. Additionally, all horses were evaluated using the CNS proposed by Carter et al. [4] for which numbers ranging from 0 (no palpable crest) to 5 (enormous neck) are given. An X-ray examination was performed to assess the gait and hoof capsule of the horses. Furthermore, palpation and visual assessments were conducted by veterinarians who could recognize subclinical and/or clinical signs of laminitis. Clinical examinations also included analysis of resting insulin levels and a CGIT. At 16 h prior to performing the CGIT, the horses received only 4 kg of timothy hay. Blood for analysis was collected from the jugular vein in order to measure the levels of resting insulin and glucose. Next, a 50% dextrose solution (150 mg/kg body weight [bw]) was injected intravenously and followed immediately by an insulin bolus of 0.10 U/kg bw (Humulin R; Eli Lilly and Company, USA) delivered intravenously. Glucose concentrations were measured at 1, 5, 15, 25, 35, 45, 60, 75, 90, 105, 120, 135, and 150 min after insulin administration using a glucometer (Glucosens 1040; ARKRAY, Japan) as previously described [6,21].

Fasting blood samples (10 mL) were collected from the external jugular vein of all ponies. Sterile techniques were performed throughout the collection process. The samples were transferred to polypropylene tubes (Monovette; Sarstedt, Germany) with serum clotting activator. Next, the samples were centrifuged at 1,137 × g for 10 min and temperature 4℃ (MPW 54; MPW Med. instruments, Poland). The serum was collected, and four aliquots were prepared and frozen in liquid nitrogen for further analysis. To measure glucose concentrations, blood was collected into tubes containing fluoride as a glycolysis inhibitor and EDTA as an anticoagulant (S-Monovette; Sarstedt). The preserved blood was analyzed within 24 h of collection. Adipose tissue samples (2 g) were collected from the base of the mane from each horse under local anesthesia (2% Lignocainum; Polfa Warszawa, Poland). The tissue samples were incubated in 10% buffered formalin for 24 h and temperature 4℃ and washed by Phosphate Buffered Saline-PBS (Sigma-Aldrich, Germany). All samples were dehydrated by increasing gradients of 50%, 60%, 70%, 80%, 96%, 100% and 100% ethanol. The samples were then imbued in ethanol/xylene 1 : 1 and clear xylene, both for 1 h at room temperature. After that the samples were embedded in paraffin.

Aliquots of serum were analyzed to determine the insulin, LEP, IL-6, and TNF-α concentrations. Measurement of each factor was performed using specific ELISA kits (insulin kit; Mercodia, Sweden; LEP kit; USBiological, US; IL-6 and TNF-α kits; Genorise Scientific, US) according to the manufacturers' instructions. Before the assays all samples were briefly centrifuged (1,137 × g; 5 min; temperature 4℃). The serum samples did not required dilution prior to the assays and were therefore transferred directly to plates pre-coated with a specific primary antibody. The samples were incubated at room temperature in the presence of the capture antibody for 2 h with the exception of the TNF-α kit for which the samples were incubated for 1 h. Measurement of insulin required addition of an enzyme conjugate immediately after the samples were applied. After incubation with the primary antibody, the serum samples were aspirated from each well and the plates were washed four to six times with Wash Buffer (supplied with the assay kits). For the IL-6 assay and TNF-α secondary detection antibody were applied, and the samples were incubated with detection antibody for 20 min at room temperature, while samples with anti-LEP secondary detection antibody were incubated for 30 min at 37℃. Next, the plates were washed as described above. Stop solution included to the assay kits was used to inhibit the reaction. Optical density of each well was immediately measured at 450 nm using a microplate reader (BMG LABTECH, Germany). The insulin kit used for the experiment had a sensitivity of 0.01 µg/L, while the precision of the assay expressed as co value (CV) was < 10% and < 12% within and between assays, respectively. Sensitivity of the test used for LEP determination was 0.242 ng/mL, while the intra-assay and inter-assay CVs were lower than 4% and 5%, respectively. The tests used for IL-6 and TNF-α measurements had a sensitivity equal to 0.8 pg/mL and an intra-assay CV of 6%. The inter-assay CVs for both the IL-6 and TNF-α tests was 9%.

Adipose tissue samples were prepared as 5 µm-thick sections cut with a Microm HM 340E microtome (Carl Zeiss, Germany) and placed on histological slides (Dako, Denmark). The sections were subsequently deparaffinized with xylene and rehydrated in ethanol (decreasing concentrations from 100 to 50%), and washed with distilled water. Next, the slides were stained with hematoxylin (Shandon; Thermo Scientific, USA) for 8 min, rinsed in running tap water for 10 min at room temperature, and stained with eosin (Shandon; Thermo Scientific) for 5 min at room temperature. The sections were then dehydrated by washing with ethanol (increasing concentrations from 50 to 100%) followed by xylene, and sealed with DPX mounting medium (AquaMed, Poland). The slides were viewed with a light microscope (Axio Imager A1; Carl Zeiss).

For immunohistochemistry, the tissue samples were cut into sections 3 µm-thick, deparaffinized in xylene, and rehydrated in alcohols with decreasing concentrations (decreasing concentrations from 100 to 50%). To visualize the antigen-antibody reaction was used EnVision System (Dako). Immunoperoxidase labeling was performed using polyclonal antibodies against IL-6 (Genorise Scientific) and TNF-α (R&D Systems, USA). Antigen heat-induced retrieval was performed by incubating the slides with target retrieval solution (pH 9.0; Dako) for 20 min at 96℃. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide and the slides were then washed with Tris-buffered saline (TBS) for 5 min at room temperature. The tissue sections were next labeled with primary antibodies for 20 min at 20℃. The antibodies were diluted to 1 : 10. The sections were counterstained with Mayer's hematoxylin for 1 min, washed by tapping water, dehydrated in ethanol (increasing concentrations from 50 to 100%), and closed in mounting medium with coverslips. Analysis was carried out with an optical microscope (Axio Imager A1; Carl Zeiss).

Normality of the data was determined using a Shapiro-Wilk test while equality of variances was assessed using Levene's test. Differences between the experimental groups were analyzed using parametric (Student's t) or non-parametric (Mann Whitney U) tests. All analyses were performed with STATISTICA 10.0 software for Windows (StatSoft, USA). P values < 0.05 were considered significant.

Characteristics of both experimental groups that were established based on clinical data are presented in Table 1. Group A (n = 8) was classified as EMS ponies. This group was characterized by divergent growth rings indicative of chronic laminitis. Ponies from this group were also characterized by high BCS (8.37 ± 051) and CNS (4.0 ± 0.76) values. A standard 95% confidence interval (95% CI) estimated for BSC was 7.92 ~8.82 while that for the CNS was 3.31~4.69. Moreover, horses from group A had a distribution of body fat typical for animals with EMS (i.e., particularly around the tail base, mane, and eyes). All horses from group A had elevated levels of resting insulin (74.75 ± 13.95 mU/mL; 95% CI = 62.29-87.21). Additionally, glucose concentrations did not return to baseline levels within 45 min as represented by positive CGIT results (130.75 ± 10.33 mg/dL; 95% CI = 121.50-140). The mean concentration of LEP was 7.72 ± 0.17 ng/mL (95% CI = 7.57-7.87) and the mean body weight was 325 ± 2.44 kg (95% CI = 322.82-327.17).

Ponies assigned to the control group (B, n = 8) did not show clinical signs of laminitis. The mean BCS value of the control horses was 7.25 ± 0.46 (95% CI = 6.88-7.68) while the mean CNS value was 1.62 ± 0.51 (95% CI = 1.17-2.07). In contrast to animals in group A, the control ponies did not have atypical body fat distribution typical for EMS horses. Ponies from group B had a normal resting insulin concentration (10.25 ± 4.68 mU/mL; 95% CI = 6.07-14.43). CGIT results were negative. Additionally, glucose concentrations decreased to a normal level (73.75 ± 9.61 mg/dL; 95% CI = 65.17-82.33) 45 min after insulin injection. The mean level of serum LEP was 3.97 ± 0.05 ng/mL (95% CI = 3.93-4.02) while the mean body weight was 290 ± 2.39 kg (95% CI = 287.87-292.12).

Comparison of adipose tissue samples from EMS ponies and healthy animals showed that tissues from horses in group A were highly infiltrated by macrophages and lymphocytes that were mainly localized in the intercellular space between adipocytes (panel A in Fig. 1). Infiltration of inflammatory cells was not observed in tissue samples from group B (panel B in Fig. 1). Additionally, adipose tissue samples obtained from the EMS group showed signs of fibrosis while samples from the healthy ponies lacked these features.

In the adipose tissue samples, both IL-6 and TNF-α were detected with immunohistochemical staining while serum levels of these factors were measured with ELISAs. IL-6 expression was higher in adipocytes from the EMS ponies (panel E in Fig. 1) compared to healthy individuals (panel F in Fig. 1). However, distribution of IL-6 in the tissue samples was similar for both groups. High accumulation of IL-6 was observed around the cell membrane and nuclei of adipocytes. Blood vessels in samples from group A were found to be infiltrated by macrophages and lymphocytes (panel G in Fig. 1). In contrast, no signs of inflammation were observed in the sections derived from group B (panel H in Fig. 1). Analysis of IL-6 concentrations in serum showed that EMS ponies had higher levels of the cytokine (x = 1.905 µg/mL; 95% CI =1.902-1.907). The mean IL-6 concentration of the healthy animals was 1.886 µg/mL (95% CI = 1.884-1.887). These differences were statistically significant (p = 0.0000; Fig. 2). TNF-α levels in adipocytes from EMS and healthy ponies were comparable. The protein was mainly located in the area around adipocyte nuclei (panels C and D in Fig. 1). EMS ponies had high levels of serum TNF-α with a mean concentration of 1.970 µg/mL (95% CI = 1.960-1.980). Healthy ponies had a mean serum TNF-α concentration of 1.250 µg/mL (95% CI = 1.220-1.280; Fig. 3). The observed differences were statistically significant (p = 0.0007).

Statistical analysis also revealed differences in serum LEP concentrations (p = 0.0007). The mean LEP level of group A was 7.72 ng/mL. On the other hand, the mean LEP concentration for group B was 3.97 ng/mL (Fig. 4).