This work was completed at the Middleburg Agricultural Research and Extension Center in Middleburg, VA 20117-3521. Presented at the 28th Forum of the American College of Veterinary Internal Medicine, Anaheim, CA, June 2010.
Corresponding author: B.A. Corl, Department of Dairy Science, Virginia-Maryland Regional College of Veterinary Medicine, Virginia Tech, Blacksburg, VA 24060-0315; e-mail: firstname.lastname@example.org.
Background: Obesity and insulin resistance increase the risk of laminitis in horses. Pioglitazone (PG) is an insulin-sensitizing drug used in humans that is absorbed after oral administration to horses.
Hypothesis: PG treatment will increase insulin sensitivity and transcript abundance of glucose and lipid transporters in adipose and skeletal muscle tissues.
Animals: Sixteen lean, healthy horses.
Methods: Eight horses were administered PG (1 mg/kg bodyweight PO) for 12 days before induction of insulin resistance through IV administration of lipopolysaccharide (LPS). Treated and untreated controls (CN; n = 8) were subjected to testing of peripheral insulin sensitivity and biopsies of both subcutaneous (nuchal ligament) adipose tissue and skeletal muscle before and after treatment, and 24 hours after LPS administration.
Results: PG treatment did not improve basal insulin sensitivity (CNs: 1.4 ± 0.3, PG-treated: 1.9 ± 1.3; P > .4) or mitigate LPS-induced insulin resistance (CNs: 0.4 ± 0.3, PG-treated: 0.4 ± 0.3); however, transcript abundance of glucose and lipid transporters was altered in both skeletal muscle and subcutaneous adipose tissue.
Conclusions and Clinical Importance: Either a higher dose or longer treatment period might be required for physiological effects to be observed. PG is a novel therapeutic agent requiring further investigation in horses in order to determine treatment efficacy.
Insulin maintains blood glucose concentrations by facilitating glucose uptake into insulin responsive tissues such as adipose, skeletal muscle, and liver. Insulin sensitivity is defined as the insulin concentration that results in half-maximal responsiveness of a tissue.1 Different physiological states can alter tissue responsiveness to insulin, with a decrease in insulin sensitivity being termed insulin resistance. Thus, insulin resistance is defined as the physiological state in which normal concentrations of insulin produce a less than normal biological response.
Reduced insulin sensitivity in horses is associated with obesity, and increases the risk of laminitis.2–4 Although the mechanisms of insulin resistance in horses have not been identified, hypotheses extrapolated from human medicine include reduced protein and mRNA levels of the insulin responsive glucose transporter (GLUT4).5 Treatment strategies to improve insulin sensitivity in horses include exercise, weight loss, and human-medicine based drugs such as metformin.6,7 Pioglitazone (PG) improves insulin sensitivity in other species.8,9 Its efficacy has not been evaluated in horses; however, it is absorbed after oral administration.10 PG is a synthetic ligand for peroxisome proliferator-activated receptor gamma (PPARγ), an intracellular receptor and transcription factor.11,12 Transcription of genes, controlled by PPARγ, directly increases the efficiency of insulin stimulated glucose clearance from blood by increasing the cellular content of glucose and lipid transporter proteins.13
Independent of obesity, intravenous administration of lipopolysaccharide (LPS) induces insulin resistance.14 The purpose of this study was to determine the effect of short-term PG treatment on peripheral insulin sensitivity and expression of glucose and lipid transporters in adipose tissue and skeletal muscle in horses.
Materials and Methods
Horses, Treatments, and Experimental Design
Use of animals for this study was approved by Virginia Tech's Institutional Animal Care and Use Committee. Sixteen nonpregnant Thoroughbred mares (8–21 years) with a mean ± SD body weight (BW) of 589 ± 40 kg and body condition score of 5.9 ± 0.8 were blocked by age and BW, and, within a block, randomly assigned to the PG (n = 8) or control (CN; n = 8) treatment group.15 Two animals from each treatment were randomly assigned to 1 of the 4 starting blocks and experiment start dates of these blocks were staggered. Horses allocated to PG received a 14-day dose (1 mg/kg BW/day, PO at 0700 hours crushed and mixed into unflavored gelatin) of PG (Fig 1).a Intravenous jugular catheters were placed aseptically and used for all blood sampling.b On day 13, LPS (Escherichia coli O55:B5, 35 ng/kg BW; diluted in 500 mL saline) was infused IV. To demonstrate that horses responded to LPS infusion, rectal temperatures were obtained every 30 minutes for 4 hours after the commencement of infusion. Frequently sampled intravenous glucose tolerance (FSIGT) tests were administered as described previously, and adipose and skeletal muscle tissue samples were collected on day 0, 12, and 14.c,d,16
Blood samples (10 mL) collected into EDTA or sodium heparin coated blood collection tubes before each FSIGT test were analyzed for plasma glucose, insulin, nonesterified fatty acids (NEFA), and triglyceride (TG) concentrations as described previously.e,16 Values for insulin sensitivity (Si), acute insulin response to glucose (AIRg), glucose effectiveness (Sg), and disposition index (DI) were calculated.f
Tissue Collections and RNA Extraction
After completion of FSIGT tests, horses were walked to the adjoining clinic and restrained in standing stocks. Horses were sedated with IV xylazine (300 mg) or detomidine (5 mg) before percutaneous skeletal muscle (400–500 mg) and adipose tissue (500 mg) biopsies. Samples were collected under aseptic conditions after desensitization of the overlying skin and subcutaneous tissues with up to 20 cm3 lidocaine (2%). Adipose tissue was taken from the nuchal crest region approximately 2 cm lateral to the base of the mane. Skeletal muscle was taken from the middle gluteal muscle approximately 20 cm lateral to the dorsal midline at a depth of 8 cm, by a needle biopsy technique. Tissue samples were flash-frozen in liquid nitrogen and stored at −80°C until analysis. Total RNA was isolated from tissue samples with a phenol extraction.g Samples (0.5 μg RNA from each sample) were assessed for RNA degradation with a 1% agarose gel including 1% SYBR Safe.h Any samples with evidence of degradation were re-extracted.
Quantification of mRNA
Primer-probe sets (Table 1) for detection of equine β-actin, long-chain fatty acid transporter (CD36), fatty acid transporter protein (FATp), insulin receptor (INSR), GLUT4, noninsulin-dependent glucose transporter 1 (GLUT1), and PPARγ were purchased.i Adipose and muscle tissue gene expression was determined by a 1-step PCR reaction with 30 ng of RNA (in a 3.76 μL total volume) and a 1-step Master Mix solution (5.00 μL Master Mix, 0.24 μL enzyme mix, 0.5 μL primer-probe mix, and 0.5 μL water) in a total volume of 10 μL.j Each sample was analyzed in duplicate and reaction mixtures were incubated at 48°C for 30 minutes followed by 10 minutes at 95°C before 40 cycles of 15 seconds at 95°C and 1 minute at 60°C in a thermocycler.k
Western blot analysis of total GLUT4 and INSR protein content in adipose and skeletal muscle was performed. Approximately 200–300 mg of tissue was homogenized in 300 μL of lysis buffer (58.3 mM Tris-HCl, pH 7.4; 0.5% [v/v] Triton X-100; 300 mM NaCl; 2 mM EDTA, pH 8.0; and 20% [v/v] Protease Inhibitor Cocktail).l After homogenization, samples were incubated on ice for 30 minutes and then centrifuged at 14,000 ×g for 15 minutes (4°C). The resulting supernatant was retained and used to quantify protein concentrations.m Protein concentrations were adjusted to 2 μg/μL with cold lysis buffer, combined with an equivalent volume of 2 × Laemmli Sample Buffer, and boiled at 95°C for 10 minutes.l
Proteins (20 μg per sample) were resolved on a 10% SDS-polyacrylamide gel before transfer to a PVDF membrane.n,o Membranes were blocked with Tris Buffered Saline that included 0.1% Tween-20 and 5% dried nonfat milk for 1 hour. Membranes were then incubated overnight with an antibody directed against GLUT4 (1 : 200) before incubation with an appropriate secondary antibody and visualization of protein bands with a chemiluminescent detection system.p,q,r After GLUT4 visualization, membranes were stripped and reprobed with an antibody directed against INSR (1 : 1000).s After visualization, membranes were washed and incubated with an antibody directed against β-actin (1 : 2000).t For all antibodies, bands were visualized and quantitated.u Detection of β-actin in skeletal muscle was lower than expected and thus bands were only visually assessed.
Differences in basal plasma insulin, glucose, NEFA, and TG concentrations on day 0 were analyzed by 1-way analysis of variance (ANOVA) with starting block included as a blocking factor. After the demonstration that no differences were detected, the effects of PG and LPS were analyzed by repeated measures using fixed effects of treatment and day of study, starting block as a blocking factor, and horse within treatment as the repeated term. Because mean plasma insulin concentrations were different between treatment groups on day 0, these values were included as a covariate. Values for NEFA and TG were natural log transformed in order to meet the assumption that residuals are normally distributed. The effects of treatment and time on rectal temperatures were assessed by repeated measures ANOVA, with horse within treatment included as the repeated term and starting block included as a random effect. Differences in minimal model parameters between treatment groups at day 0 were assessed by 1-way ANOVA with starting block as a blocking factor. After the determination that treatment groups were not different before treatment, a repeated measures ANOVA with day 0 values being used as a covariate and fixed effects of treatment and day was used to investigate the interactions and effects of PG treatment and LPS infusion on the minimal model parameters. When significant treatment by time effects were detected, Tukey's test was used to identify simple effect differences.
Gene expression data were analyzed using ΔCT values (where CT is the cycle threshold at which fluorescence rose above background, and ΔCT=CT [gene of interest] −CT [endogenous normalizer]).17 Differences in gene transcript abundance between treatment groups at baseline (day 0) were analyzed by 1-way ANOVA. After assertion that no differences existed between treatment groups before treatment, the effects and interactions of PG and LPS administration were analyzed in the day 12 and day 14 samples by repeated measures ANOVA with starting block as a blocking variable, fixed effects of treatment and day, and horse within treatment as the repeated term. The effects and interactions of treatment and LPS on GLUT4 and INSR protein abundance in adipose and muscle tissue were assessed by repeated measures ANOVA with fixed effects of treatment and day.
Plasma Glucose, Insulin, NEFA, and TG Concentrations
Before the beginning of the treatment period, plasma glucose, NEFA, and TG concentrations were not different between treatment groups (P > .2; data not shown). Plasma insulin concentrations were greater in the CN (24.3 ± 3.8 μIU/mL) versus PG (12.5 ± 4.0 μIU/mL) treated horses (P= .028; data not shown). In both PG and CN horses, there was increased plasma TG (P < .001) concentrations compared with day 12 values (Fig 2). However, there was no effect of LPS on plasma NEFA (P= .085). No effect of LPS administration was detected for the other variables. Plasma glucose concentrations were increased in PG-treated animals regardless of LPS administration (P= .037; Fig 2). Plasma insulin concentrations did not differ between PG and CN horses after either the treatment period (P > .1) or LPS administration (P > .6; Fig 2). LPS infusion increased (P < .001) rectal temperatures from 99.8 ± 0.3 to 102.0 ± 0.3°F with no difference between PG- and CN-treated animals (P > .8).
Before the start of the study, treatment groups were similar for all minimal model parameters (P > .3; data not shown). With pretreatment values used as a covariate, infusion of LPS reduced Si by 72% (P < .001), DI by 82% (P < .001), and Sg by 32% (P= .006), while not influencing the AIRg (P > .5) (Fig 3). PG treatment had no detectable effect on any of the parameters (P > .4 for all variables).
Transcript and Protein Abundance
Twelve days of PG treatment increased skeletal muscle tissue transcript abundance of both CD36 (P= .035) and GLUT1 (P= .050) regardless of LPS administration (Fig 4), but neither PG treatment nor LPS administration influenced transcript abundance of FATp or PPARγ (Table 2). LPS administration did not reduce transcript abundance of GLUT4 (P= .059; Fig 5B), while neither PG (P > 0.3) nor LPS (P > 0.5) influenced GLUT4 protein abundance (Fig 5A). Further, PG treatment increased transcript abundance of INSR (P= .018; Fig 5C) while not increasing INSR protein abundance (P > .7; Fig 5A).
Table 2. Skeletal muscle gene expression in untreated controls and horses treated for 12 days with pioglitazone (1 mg/kg PO) before (post-TRT) and 24 hours after administration of lipopolysaccharide (post-LPS; 35 ng/kg of bodyweight, IV).
In contrast to the effects seen in skeletal muscle, neither PG treatment nor LPS administration influenced adipose tissue transcript abundance of FATp, CD36, or GLUT1 (Table 3). PG did not enhance protein abundance of GLUT4 (P= .083; Fig 6A, B), while infusion of LPS reduced transcript abundance of GLUT4 in the CN horses only (P= .026; Fig 6C). Neither PG nor LPS influenced protein (Fig 6A, D) or transcript (Fig 6E) abundance of INSR.
Table 3. Adipose tissue gene expression in untreated controls and horses treated for 12 days with pioglitazone (1 mg/kg PO) before (post-TRT) and 24 hours after administration of lipopolysaccharide (post-LPS; 35 ng/kg of bodyweight, IV).
In order to investigate insulin sensitivity, horses underwent FSIGT testing. In diabetic, hyperglycemic humans and rodents, PG treatment increases insulin sensitivity by increasing glucose and lipid uptake into insulin-sensitive tissues.18 In this study, no parameter of insulin sensitivity, either minimal model or plasma variables, was altered after the PG treatment period. In contrast, a number of gene and protein targets had altered abundance after treatment. Alterations in gene and protein expression without a correlation to physiological changes could suggest that adequate plasma concentrations or duration of treatment was not achieved. Average peak plasma PG concentrations in horses were determined to be 448 ng/mL after 1 mg/kg dosing PO.10 This value was considerably lower than the reported concentration observed in human plasma (1482–1587 ng/mL) after a smaller daily dose of 45 mg (∼0.6 mg/kg for average 70 kg subject).19 Higher concentrations in horses could be attained through increasing dose or dose frequency. Alternatively, use of nonobese horses (likely not insulin-insensitive) might have prevented observing effects of greater magnitude. The determination of treatment efficacy in humans and rodents was made through observations of increased insulin sensitivity in insulin insensitive animals, and further increasing insulin sensitivity in insulin sensitive horses might be physiologically impossible.
In addition to evaluating insulin sensitivity posttreatment, the ability of PG to mitigate LPS-induced insulin resistance was investigated.14,20 In this study, administration of LPS-reduced insulin sensitivity by 75% at 24 hours regardless of PG pretreatment. Potentially, the acute induction of insulin resistance after LPS administration is elicited through a different mechanism than that which occurs in obesity. If this is the case, then it is possible that PG would be ineffective in preventing LPS-induced changes in insulin sensitivity. Alternatively, LPS induces insulin resistance more abruptly than PG is capable of counteracting. Determining the efficacy of PG treatment for obesity-induced insulin resistance will likely require testing in that specific population of horses or ponies.
PG improves insulin sensitivity in humans and laboratory animals by increasing mechanisms for glucose and lipid clearance from blood.21 While increased glucose uptake is primarily a result of increased insulin sensitivity, reduced plasma NEFA might actively increase insulin sensitivity by decreasing NEFA mediated impairment of insulin signaling.18 Increased NEFA uptake into adipose tissue also promotes adipogenesis and TG synthesis.22,23 An increase in TG synthesis necessitates an increase in glucose transport for provision of glycerol units.24 Our observation of increased GLUT4 protein levels in adipose tissue after 12 days of PG treatment would indicate a potential for increased glucose availability for TG synthesis. In insulin-resistant humans, GLUT4 mRNA and protein levels in adipose tissue were increased after 3 weeks of PG treatment.9 Interestingly, in the present study, there were no alterations to GLUT4 transcript, and as PG increases GLUT4 transcript half-life in adipocytes,25 it is possible that a longer half-life increased the efficiency of protein translation. Further, neither GLUT1 nor INSR transcript or protein abundance was altered. In regard to NEFA availability, plasma NEFA concentrations did not decrease and adipose tissue lipid transporter gene expression was not increased, which may indicate that adipose tissue lipid metabolism was not altered in this experiment. Increased plasma lipid concentrations are common in obese horses,26 but were not present before treatment in this study. Thus, whether PG can increase adipose tissue uptake of NEFA should be investigated in obese, hyperlipidemic horses.
Skeletal muscle is a major contributor to peripheral insulin sensitivity, accounting for up to 80% of plasma glucose disposal.27 We noted a minimal increase (0.3-fold) in INSR transcript abundance after PG treatment; however, this did not result in increased protein levels. In obese rats, PG treatment (10 days at 3 mg/kg) improved insulin signaling without altering INSR, potentially indicating that PG influences insulin sensitivity through insulin signaling cascade intermediates rather than through the receptor.28 While the lack of effect of PG on INSR abundance is in agreement with previous literature, the lack of influence of PG on either transcript or protein abundance of GLUT4 is inconsistent with research in other species. PG increased GLUT4 protein in rodents after higher dose administration (20 mg/kg) and a longer treatment period (28 days).8 We also noted an increase in GLUT1 transcript expression, potentially indicating an increased capacity for noninsulin-mediated glucose disposal. The lack of change in Sg suggests that this small increase was too minor to facilitate an alteration in corresponding protein levels and glucose disposal.
Plasma NEFA can directly impair insulin signaling, and PG treatment has been shown to influence transcript abundance of both CD36 and FATp.29,30 While PG did not alter plasma lipid concentrations, it did increase skeletal muscle transcript abundance of CD36. This increase, coupled to an increase in protein, could contribute to increased NEFA uptake into muscle. In diet-induced obese rats, 3 weeks of PG treatment (3 mg/kg) increased both muscle membrane CD36 content and uptake of NEFA from plasma.31 Potentially, the minor increase in mRNA did not translate into increased protein. Alternatively, as mentioned previously, the lack of basal hyperlipidemia could indicate that any increase in NEFA uptake was matched by an increase in adipose tissue release.
Administration of LPS tended to reduce adipose tissue GLUT4 mRNA abundance in the CN horses while having a more significant effect in 3T3-L1 adipocytes.32 PG-treated animals did not show a decrease in adipose GLUT4 mRNA, and this indicates the potential for maintenance of transcription in spite of the inhibitory effects of LPS. But it is unknown why this effect would be limited to adipose tissue, as LPS reduced GLUT4 mRNA transcript levels in skeletal muscle in both treatment and CN groups. In addition to reduced insulin sensitivity, LPS administration decreased Sg. This potentially suggests reduced GLUT1 concentrations; while mRNA levels did not decrease in either adipose or skeletal muscle, protein was not measured.
In addition to glucose metabolism and insulin sensitivity, LPS alters lipid metabolism.33 In this experiment, infusion of LPS increased plasma TG concentrations, which could indicate an increase in either adipose tissue lipolysis, because of adipose tissue insulin resistance, or decreased clearance. LPS increases adipose tissue lipolysis in vivo and in cultured adipocytes and decreases skeletal muscle lipid oxidation via decreased oxidative gene expression.34,35 The influence of either mechanism, increased secretion or decreased oxidation, remains to be investigated in horses.
a Actos, Takeda Pharmaceuticals North America Inc, Deerfield, IL
b Milacath, MILA International Inc, Erlanger, KY
c 50% w/v dextrose; 0.9% saline, Baxter, Deerfield, IL
d Novolin R, Novo Nordisk Inc, Princeton, NJ
e Vacutainer, Beckton-Dickson, Franklin Lakes, NJ
f MinMod Millenium, V 5.15, University of Pennsylvania, Kennett Square, PA
g TRIzol, Molecular Research Center Inc, Cincinnati, OH
h SYBR Safe DNA gel stain, Invitrogen, Carlsbad, CA
i Assays-By-Design software system, Applied Biosystems, Foster City, CA