Departments of Animal and Poultry Science, †Pathobiology and ‡Biomedical Science, University of Guelph, Guelph, Ontario, Canada.
Present addresses: S. Pratt-Phillips, Department of Animal Science, North Carolina State University, Raleigh, North Carolina, 27608, USA; L. J. McCutcheon, Department of Pathobiology and Diagnostic Investigation, College of Veterinary Medicine, Michigan State University, Lansing, Michigan 48910, USA; R. J. Geor, Department of Large Animal Clinical Science, College of Veterinary Medicine, Michigan State University, Lansing, Michigan 48910, USA. Email: firstname.lastname@example.org
Reasons for performing study: Starch rich (S) feeds reduce insulin sensitivity in untrained horses when compared to high fat (F) feeds, but insulin sensitivity is not affected when S or F are fed during exercise training. The effects of S vs. F on training-associated alterations in skeletal muscle glucose metabolism are unknown.
Objectives: To determine the effects of dietary energy source on training-associated changes in insulin sensitivity, skeletal muscle GLUT4 protein and hexokinase (HK) and glycogen synthase (GS) activities in horses.
Methods: After a baseline period on an all forage diet (Phase 1), horses were adapted to high starch (S) or high fat (F) diets (n = 7/group) for 6 weeks (Phase 2) and then completed 7 weeks of exercise training (Phase 3) on the same diets. To measure insulin sensitivity (SI), minimal model analysis of a frequently-sampled i.v. glucose tolerance test was performed at the end of each phase. Middle gluteal muscle biopsies to measure GLUT-4 protein content, muscle glycogen and HK and GS activities were taken before and after euglycaemic-hyperinsulinaemic clamps administered after each phase. Data were analysed by repeated measures ANOVA.
Results: In S, SI was 36% lower (P<0.05) after Phase 2 when compared to Phase 1 but was unchanged in F. After Phase 3, SI was increased (P<0.01) in S and F compared to Phase 2 and did not differ (P>0.05) between diets. Middle gluteal muscle GLUT-4 protein and post clamp HK activity were increased (P<0.05) in S after Phase 3, with higher (P<0.01) GLUT4 in S than in F. GS activities were unchanged in both diets.
Conclusions: Adaptation to S resulted in decreased SI mitigated by moderate physical conditioning. Increased GLUT-4 protein content and HK activity in S may have contributed to higher SI after training.
Recent studies have reported that adaptation to concentrate feeds with high (>40%) nonstructural carbohydrates (NSC; starch and sugars) reduces whole-body insulin sensitivity in untrained horses when compared to a feed utilising fat as the primary energy source (Hoffman et al. 2003; Treiber et al. 2005; Pratt et al. 2006). However, this difference in insulin sensitivity was not apparent when starch-rich or fat-rich concentrate feeds were fed to horses during exercise training (Treiber et al. 2006). These observations suggest that physical conditioning modifies the effect of dietary energy source on insulin sensitivity in horses, perhaps via training-associated alterations in insulin-mediated glucose disposal in muscle. To date, however, no study in horses has evaluated the impact of dietary energy source on training-associated alterations in skeletal muscle glucose metabolism together with measures of insulin sensitivity.
The objective of this study was to determine the effects of dietary energy source on insulin sensitivity and measures of skeletal muscle glucose metabolism in Standardbred horses, both in the sedentary state and after a subsequent period of physical conditioning. It was hypothesised that physical conditioning would mitigate insulin resistance induced by the feeding of a high NSC diet in association with increased GLUT-4 protein and activities of glycogen synthase and hexokinase in skeletal muscle. The present report is a follow-up to the study by Pratt et al. (2006) in which the same experimental design was used to examine the effects of dietary energy source on insulin sensitivity, measured by use of the euglycaemic-hyperinsulinaemic clamp (EHC) and oral glucose tolerance.
Materials and methods
All procedures were approved by the institution's Animal Care Committee. The horses used were 14 mature Standardbreds (6 geldings, 8 mares, 3–4 years old, mean [±s.d.] bodyweight [bwt] 424 ± 8 kg). All horses were paddock rested and maintained on mixed grass hay for at least 3 months prior to the study.
Study design and diets
The study was divided into 3 phases; Baseline (3 weeks), Diet (6 weeks), and Diet x Exercise (7 weeks) referred to as Phases 1, 2 and 3, respectively. An insulin-modified frequently sampled i.v. glucose tolerance test (FSIGTT) was administered 72 h following the end of each phase to assess insulin sensitivity and glucose dynamics. A middle gluteal muscle biopsy was obtained both before and after a 3 h euglycaemic-hyperinsulinaemic clamp (EHC) administered 48 h after the end of each phase. This procedure provided a sustained hyperinsulinaemic stimulus for assessment of muscle glycogen synthase and hexokinase activity. Measures of insulin sensitivity obtained from the EHC method have been reported elsewhere (Pratt et al. 2006). Before and after Phase 3, horses completed an incremental treadmill1 exercise test for measurement of the peak rate of oxygen consumption (VO2peak), as previously described (Pratt et al. 2006).
During Phase 1, all horses were fed mixed timothy-alfalfa forage cubes2 with the ration divided into 2 meals. At the end of Phase 1, the horses were randomly assigned to one of 2 groups of 7 for the remainder of the study. One group was fed forage cubes and a sweet feed mix rich in NSC (S diet), the other forage cubes and a fat-supplemented concentrate with low NSC content (F diet). The switch from forage only to concentrate : forage was made abruptly. Samples of concentrates and forage cubes were analysed in a commercial laboratory3 for nutrient composition (Table 1). Total NSC was the sum of starch and water-soluble carbohydrates. Mean daily intake (as fed basis) throughout the study was 18.4 ± 0.4 g/kg bwt and did not differ between S and F in Phases 2 and 3. In Phases 1, 2 and 3 mean DE intakes (on a dry matter basis) were, respectively, 36.1 ± 1.1, 45.2 ± 2.0 and 46.1 ± 1.9 kcal/kg bwt per day. In Phases 1 and 2 DE intake was, respectively, approximately 109 and 129% of published requirements, while in Phase 3 DE intake matched recommendations for horses in moderate exercise training (National Research Council 2007). In the S and F treatments, approximately 65% digestible energy (DE) was provided by concentrate and 35% DE from forage cubes. Mean dietary NSC in S and F were, respectively, 6.9 ± 0.1 and 1.8 ± 0.1 g/kg bwt per day. The estimated percentages of the total daily DE intake provided by NSC and fat were 56.7 and 1.1% in S and 14.1 and 36.2% in F.
Table 1. Chemical analysis (on a dry matter basis) of the high starch (S) and high fat (F) concentrates (n = 5) and the forage cubes (n = 6)
Data are presented as arithmetic means ± s.d. Analyses were performed by Dairy One DHIA Forage Testing Laboratory, Ithaca, NY. ADF, acid detergent fibre; NDF, neutral detergent fibre; NFC, nonfibre carbohydrates = 100 − CP − fat − (NDF + neutral detergent insoluble CP) – ash; NSC, nonstructural carbohydrate = measured starch + water soluble carbohydrate; DE, digestible energy. *Significant difference (P<0.05) between S and F.
In Phase 3, horses were exercised in an automated exercise machine4 at approximately 4 m/s for 15–20 min 2 times per week and on the treadmill (3° incline) 3 times per week. Treadmill training during weeks 1–2 consisted of trotting for 12 min at 35% VO2peak and 4 min at 50% VO2peak. An additional 2 min at 75% VO2peak was added during weeks 3–4 and in weeks 5–7 the first 14 min was completed at 50% VO2peak followed by 3 min at 75% VO2peak. The running speeds that elicited 35, 50 and 75% VO2peak were determined by linear regression analysis from the initial incremental exercise test.
Feed was withheld approximately 10 h prior to the procedure. On the morning of the experiment catheters (14 g × 5¼″)5 were inserted into each jugular vein. The FSIGTT procedure was performed as previously described (Hoffman et al. 2003). In brief, a 300 mg/kg bolus of 50% (wt/vol) dextrose was infused via the right jugular catheter at 0 min. Blood samples were taken from the left jugular catheter at 15 and 5 min before and 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16 and 19 min after dextrose injection. At 20 min, 30 mU/kg insulin6 was administered and further blood samples were collected at 22, 23, 24, 25, 27, 30, 35, 40, 50, 60, 70, 80, 90, 100, 120, 150 and 180 min after glucose dextrose administration. Blood samples were transferred to tubes containing sodium heparin as anticoagulant or no additive and were centrifuged for 15 min (1600 g), after which plasma and serum were harvested and stored at -20°C until analysis. Plasma glucose concentration was measured in triplicate spectrophotometrically by use of the hexokinase method7. Serum immunoreactive insulin was measured in duplicate by radioimmunoassay using a commercially available kit8 validated for equine insulin (Freestone et al. 1991). Glucose and insulin values were analysed by the minimal model method9 to provide measures of insulin sensitivity (SI), glucose effectiveness (SG), acute insulin response to glucose (AIRg) and disposition index (DI, the product of SI and AIRg).
Middle gluteal muscle samples were collected by the needle biopsy technique (Lindholm and Piehl 1974). Samples were obtained 5 min before and immediately after 3 h of insulin and glucose infusion during the EHC. Biopsy specimens were collected following aseptic preparation and desensitisation of the overlying skin with 2% lidocaine10. Samples were quickly blotted to remove excess blood, flash frozen in liquid nitrogen and stored at -80°C until analysed.
For measurement of muscle glycogen, frozen muscle samples (50–60 mg) were lyophilised, pulverised and dissected free of any visible blood, connective tissue and fat. Glycogen content (as glucosyl units) was then determined in duplicate after acid hydrolysis following the procedures described by Passonneau and Lauderdale (1974). Western immunoblot was used to assess GLUT-4 protein abundance, as previously described (McCutcheon et al. 2002; Stewart-Hunt et al. 2006). In brief, the total protein concentration of muscle lysates was determined by use of a commercially available kit11 and equal quantities of protein (40 µg) were separated by sodium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes12. After blocking, the membranes were incubated with a polyclonal, rabbit anti-GLUT-413 at a dilution of 1:1000 followed by an anti-rabbit horseradish peroxidase-linked immunoglobulin-G antibody14 applied for 2 h at a dilution of 1:2000. Protein bands were visualised after application of a chemiluminescence reagent15 and signals quantified by densitometry16. Intensity of GLUT-4 protein bands were corrected to tubulin17 protein bands imaged on the same membrane.
Glycogen synthase (GS) activity was determined as previously described (Stewart-Hunt et al. 2006). Duplicate homogenates of frozen muscle were incubated at 37°C with media containing either 0 or 10 mmol/l glucose-6-phosphate (G-6-P) for determination of active (GS0) and total (GS10) GS activity, respectively. Uridine diphosphate (UDP)-glucose was added to the samples to catalyse the incorporation of glucose from UDP-glucose to glycogen, in turn releasing UDP. Fluorescence readings were taken in the presence of UDP reagent prior to and following addition of pyruvate kinase. GS0 and GS10 activity was determined by the difference in NAD+ fluorescence measured in the samples before and after addition of pyruvate kinase. GS fractional velocity (GSFV) was calculated as GS0 divided by GS10. The intra- and interassay CVs were, respectively, 5.5 and 6.6%.
Hexokinase (HK) activity was assayed as previously described (Stewart-Hunt et al. 2006). In brief, muscle homogenates were incubated for 1 h at room temperature (20°C) after addition of G-6-P dehydrogenase. The reaction was then stopped, phosphogluconate dehydrogenase added to the homogenates and fluorescence readings taken following 15 min incubation at room temperature. The intra- and interassay CVs were, respectively, 1.9 and 2.2%.
Descriptive statistics of continuous variables are expressed as mean ± s.e. A mixed model ANOVA with repeated measures was used for analysis of the effects of phase and diet on minimal model parameters and muscle variables. Pair-wise comparisons were performed by the Bonferroni procedure. Significance level was set at P<0.05. Sigmastat 3.0 and Systat 11.018 software were used for statistical computations.
No adverse clinical effects were observed in association with the abrupt change in diet between Phases 1 and 2. Mean bodyweight for all horses at the end of Phases 1, 2 and 3 was 424 ± 8 kg, 439 ± 8 kg and 446 ± 10 kg, respectively. Mean bodyweight was higher (P<0.01) in Phases 2 and 3 compared to Phase 1; however, there was no effect of diet on bodyweight in any phase. Only 13 horses completed Phase 3 as one horse from the S diet was removed from the training protocol due to lameness. Physical conditioning during Phase 3 resulted in a 9.9% increase (P<0.05) in mean VO2peak from 112.9 ± 3.1 ml O2/kg/min at the beginning of Phase 3 to 124.1 ± 3.5 ml O2/kg/min at the end of Phase 3. There was no difference between the change in VO2peak for S and F diets.
Values for minimal model parameters are reported in Table 2. There was a time by diet interaction (P<0.01) for SI such that mean values were lower in S than in F after Phase 2. After physical conditioning (Phase 3), mean SI did not differ between diets, but within F and S was higher (P<0.05) when compared to Phases 1 and 2. In S, mean AIRg was higher (P<0.01) in Phases 2 and 3 than in Phase 1. In addition, mean AIRg was higher (P<0.01) in S than in F in Phase 2. Mean DI did not differ between diet treatments in any phase, but within F was higher (P<0.05) in Phase 3 when compared to Phases 1 and 2. Mean SG was unaffected by diet or phase.
Table 2. Results of minimal model analysis of the frequent-sample i.v. glucose tolerance test (FSIGT) in Standardbred horses fed the S and F concentrates
Data are mean ± s.e. for 7 horses in Phases 1 and 2 and for 6 horses in S and 7 horses in F in Phase 3. AIRg, acute insulin response to glucose; SI, insulin sensitivity index; SG, glucose effectiveness; DI, disposition index.
Significant difference (P<0.05) from Phases 1 and 2 within diet.
Significant difference (P<0.05) from Phase 1 within diet.
a Significant difference (P<0.05) between S and F within Phase.
In S, muscle glycogen content was increased (P<0.001) after Phase 3 when compared to Phases 1 and 2 (Table 3). Muscle glycogen was higher (P<0.01) in S than in F in Phases 2 and 3. In both diet treatments, there was no difference in muscle glycogen content between Pre- and Post EHC samples at any phase of the study.
Table 3. Muscle glycogen content and activities of hexokinase and glycogen synthase (GS) in samples of middle gluteal muscle collected before (Pre) and after (Post) 3 h of insulin and glucose infusion
Data are mean ± s.e. for 7 horses in Phases 1 and 2 and for 6 horses in S and 7 horses in F in Phase 3. GS0, active form of glycogen synthase; GS10, total glycogen synthase activity; GSFV, glycogen synthase fractional velocity. *Significant difference (P<0.05) from Pre within diet. †Significant difference (P<0.05) between S and F at corresponding time within Phase.
Total glycogen synthase activity (GS10) was unchanged throughout the study and did not differ between diet treatments (Table 3). Similarly, Pre-EHC GS0 activity was not different among phases or between diets. For F in Phase 1 and S in Phase 3, mean GS0 activity was higher (P<0.05) in Post EHC when compared to Pre-EHC samples. With the exception of Phase 1 in S, mean GSFV was higher (P<0.05) in Post when compared to Pre-EHC samples. In Phase 3, Post EHC GSFV was higher (P<0.05) in S than in F. Mean HK activity in Pre-EHC samples was higher (P<0.05) in S than in F in Phase 1. In S, HK activity in Post EHC samples in Phase 3 was higher (P<0.05) when compared to F and to the corresponding time point in Phases 1 and 2. Within F, HK activity was unaffected by phase or induced hyperinsulinaemia (Pre- vs. Post EHC samples).
GLUT-4 protein was unchanged in F but there was a significant increase in GLUT-4 protein in S following Phase 2, with a further increase (P<0.01) after physical conditioning (Phase 3) in this diet group (Fig 1). Muscle GLUT-4 protein did not differ between diet groups in Phases 1 and 2, but was higher (P<0.01) in S than in F in Phase 3.
The present study demonstrated a decrease in insulin sensitivity in Standardbred horses adapted for 6 weeks to a diet in which approximately 56% of DE was from NSC when compared to horses fed an isoenergetic diet that provided 36% DE from fat and 14% DE from NSC. The decrease in SI in S appeared to be compensated by increased insulin secretion (AIRg) so that DI did not differ between diet groups. After a further 7 weeks of adaptation to the S diet, during which the horses underwent moderate physical conditioning, SI did not differ from baseline (Phase 1) although the increased pancreatic β-cell response (AIRg) persisted. These results suggest that physical conditioning reversed the effects of the S diet on SI. In addition, a moderate increase in skeletal muscle GLUT-4 protein content in horses fed the S but not F diets, higher muscle glycogen content and Post EHC HK activity in S than in F after physical conditioning (Phase 3). It is possible that the increased GLUT-4 protein content and HK activity may have contributed to the improvement in insulin sensitivity evident in S after physical conditioning.
Minimal model parameters
The decrease in insulin sensitivity evident in S after Phase 2 is consistent with the findings of previous studies of mature horses (Hoffman et al. 2003; Pratt et al. 2006) and weanlings (Treiber et al. 2005) adapted to feeds rich in NSC. Adaptation to high carbohydrate diets in man also has been associated with decreased insulin sensitivity (Jenkins et al. 1987; Samaha et al. 2003) and may be due to changes in insulin signalling and/or intracellular metabolism of glucose and fatty acids (Saltiel and Kahn 2001). One hypothesis is that greater insulin responses following consumption of high vs. low glycaemic meals may result in insulin resistance via down-regulation of the insulin receptor or any of its downstream effector molecules (Daly 2003). However, direct evidence in support of this hypothesis is lacking.
Lower insulin sensitivity after 6 weeks adaptation to S was accompanied by increased AIRg. One interpretation is that enhanced insulin response represents compensation for the decreased efficiency of insulin to stimulate glucose uptake in peripheral tissues. Indeed, the DI, which is a measure of the combined effect of insulin secretion (AIRg) and insulin efficiency (or SI) to prevent hyperglycaemia, did not differ between diet groups. These results are in agreement with the findings of Treiber et al. (2005) who found compensated insulin resistance with unchanged DI in weanling Thoroughbreds fed a concentrate rich in starch and sugar (i.e. NSC) in comparison to a fat and fibre (low NSC) feed. Conversely, Hoffman et al. (2003) noted a decrease in AIRg and DI in conjunction with lower SI in mature Thoroughbreds fed the same starch and sugar feed for 8 weeks.
In several species (man, rats and horses), there is robust evidence that exercise training enhances insulin sensitivity and glucose tolerance (Tokuyama and Suzuki 1998; Powell et al. 2002; Houmard et al. 2004; Stewart-Hunt et al. 2006). Therefore, the increase in SI and DI evident in F after Phase 3 may be attributed to the effects of physical conditioning, which resulted in a 10% increase in aerobic capacity (VO2peak). Likewise, it is possible that exercise training contributed to the reversal of insulin resistance in S. However, a limitation of this study was the absence of nonexercised groups on each diet so that the effects of longer term (>6 weeks) dietary adaptation in the sedentary state could be assessed. Therefore, the apparent mitigation of insulin resistance may be attributed to further adaptation to the S diet rather than physical conditioning. Nonetheless, Treiber et al. (2006) found that SI was not different between groups of exercise trained horses adapted to low or high NSC concentrates. Furthermore, physical conditioning of rats moderated the adverse effects of a high (32% of metabolisable energy) starch or sugar diet on insulin sensitivity, potentially due to exercise training associated enhancements of insulin signalling and glucose utilisation in skeletal muscle (Wright et al. 1983).
The improvement in insulin sensitivity after physical conditioning in the present study is in agreement with the observations of Powell et al. (2002) and Stewart-Hunt et al. (2006) who found a significant increase in insulin sensitivity (measured by the EHC method) in horses following 7 consecutive days of exercise. However, the increase in SI observed after Phase 3 is in contrast to the results of the companion study by Pratt et al. (2006), who failed to detect an effect of physical conditioning (Phase 2 vs. Phase 3) on insulin sensitivity in the same horses, measured using the EHC technique. The reason for this discrepancy is unclear; however, it has been argued that the minimal model method provides a more physiological and sensitive assessment of insulin sensitivity when compared to the EHC (Kronfeld et al. 2005), perhaps allowing for detection of small changes in insulin sensitivity that are not evident under the conditions of the EHC. It also warrants mention that the magnitude of change in insulin sensitivity after physical conditioning in the horses of the present study was smaller when compared to previous reports (Powell et al. 2002; Stewart-Hunt et al. 2006). Differences in study design, including the timing of insulin sensitivity measurements relative to the last bout of exercise (24 h in previous studies vs. 72 h in the present study), method for assessment (EHC vs. minimal model), and dietary management (hay only vs. hay plus concentrate in the present study) could account for this apparent discrepancy.
Skeletal muscle GLUT-4
A notable finding of the present study was the increase in GLUT-4 protein content in horses fed the S concentrate but not in horses adapted to the F feed. In rats, dietary carbohydrate content has been reported to affect the training-associated modifications in muscle GLUT-4 protein expression (Lee et al. 2002). When trained rats were fed a high carbohydrate diet, GLUT-4 protein and mRNA were increased in type I (soleus) and type II (extensor digitorum longus[EDL]) skeletal muscle. In contrast, in rats fed a low carbohydrate diet GLUT-4 protein and mRNA were increased in the soleus but not in the EDL after 8 weeks of endurance training. These findings suggested that dietary carbohydrate availability altered GLUT-4 gene expression and translation, but only in type II muscle fibres. Interestingly, equine skeletal muscle GLUT-4 is expressed in a fibre specific manner with highest expression in glycolytic type IIx fibres and minimal expression in type I fibres (van Dam et al. 2004). Other studies in rats have shown that glucose (Kuo et al. 1999) and insulin (Kuo et al. 2004) availability are important in the regulation of the exercise-associated increase in GLUT-4 protein expression in fast-twitch (type II) muscle. We hypothesise that higher glycaemic and insulinaemic responses to the S vs. F diet contributed to the observed differences in GLUT-4 protein content observed in the present study.
Although we did not detect a significant correlation between muscle GLUT-4 protein content and SI in either dietary group, it is well recognised that glucose transport capacity in muscle is strongly associated with GLUT-4 protein content (Wojtaszewski and Richter, 1998). Therefore, in S the increase in GLUT-4 protein content may represent a functional adaptation that partially compensated for the lower insulin sensitivity in this diet group. The higher GLUT-4 in S also may have facilitated increased glycogen storage when compared to F.
Muscle enzyme activities
Changes in the intracellular metabolism of glucose through the action of glycogen synthase and hexokinase may have contributed to the changes in SI following Phase 3 in S. In man, increases in total GS activity and increased activation of GS in response to hyperinsulinaemia contribute to training-associated improvements in insulin action (Ebeling et al. 1993; Christ-Roberts et al. 2004). Although total glycogen synthase activity did not change throughout the study, GS0 and GSFV (i.e. the proportion of the enzyme in the active form) were higher in S than in F following sustained hyperinsulinaemia in Phase 3. Similarly, hexokinase activity was increased following hyperinsulinaemia in Phase 3 for the S diet. These alterations in GS and HK activity evident in S after Phase 3 may have contributed to increase in SI in this diet group.
Muscle glycogen content was lower in F when compared to S in Phases 2 and 3 (Table 3). Furthermore, in S muscle glycogen content was increased after physical conditioning (Phase 3 vs. Phases 1 and 2), indicating that level of dietary NSC affects muscle glycogen storage in horses. Ribeiro et al. (2004) reported that feeding a high starch diet (15–21% DE vs. 4–8% DE) for 6 weeks tended to result in greater muscle glycogen storage in horses with polysaccharide storage myopathy (PSSM). One possible explanation for the difference in muscle glycogen content between the 2 diet groups in the present study is the lower availability of substrate (glucose) for glycogen synthesis in the F vs. S diet. On the other hand, higher muscle GLUT-4 and HK activity in S may have contributed to enhanced glycogen storage.
In summary, adaptation of Standardbred horses to a concentrate rich in NSC was associated with development of a compensated insulin resistance that was partially mitigated by a subsequent period of physical conditioning. Increased GLUT-4 protein content and HK activity after adaptation to S but not F may have contributed to the improvement in insulin sensitivity evident following a period of regular physical conditioning.
Conflicts of interest
The authors have not declared any conflicts.
1 Sato, Uppsala, Sweden.
2 Ontario Dehy Inc. Ontario, Canada.
3 Dairy One DHIA Forage Testing Laboratory, Ithaca, New York, USA.