| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL GROWTH, PHYSIOLOGY, AND REPRODUCTION |
* Department of Animal and Poultry Sciences, Virginia Polytechnic Institute and State University, Blacksburg, 24061-0306;
and
Department of Clinical Studies, New Bolton Center, University of Pennsylvania, Kennett Square 19348; and
and
Equine Studies Group, Waltham Centre for Pet Nutrition, Melton Mowbray, U.K.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
Key Words: Glycemic Meals • Horse • Insulin Sensitivity • Minimal Model
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
). Insulin resistance has been observed in fasting, obese, and inactive horses (Argenzio and Hintz, 1971; Powell et al., 2002; Hoffman et al., 2003) and may affect reproduction and exercise performance (Fitzgerald et al., 2002; Powell et al., 2002).
Early studies observed lower oral glucose tolerance in horses fed grain diets compared with strictly forage-fed animals (Jacobs and Bolton, 1982). Modern tests, such as the minimal model, now allow for more comprehensive representations of the glucose and insulin system and affects of diet on glucose regulation. Decreased insulin sensitivity has now been quantified in nonobese horses adapted to high-glycemic grain concentrates (Hoffman et al., 2003). The objective of the present study was to further test whether insulin sensitivity would be altered in rapidly growing Thoroughbred weanlings chronically adapted to meals of a high-glycemic feed rich in starch and sugar (SS), compared with meals abundant in fat and fiber (FF), with glycemic and insulinemic responses resembling that of pasture.
We applied the minimal model of glucose-insulin dynamics (Bergman et al., 1979; Boston et al., 2004) to data from a frequently sampled i.v. glucose tolerance test (FSIGT). The minimal model is based on the physiological glucose and insulin system and has been applied extensively in human medicine as well as to cattle (Stanley et al., 2002), sheep (Williams et al., 2002), miniature pigs (Behme, 1996), and the horse (Hoffman et al., 2003) to evaluate insulin sensitivity and pancreatic ß-cell response.
Insulin resistance has been identified as a risk factor or component of several equine diseases, such as some forms of laminitis, hyperlipidemia (Jeffcott et al., 1986), pituitary adenoma (Garcia and Beech, 1986), and osteochondrosis (Ralston, 1996; Henson et al., 1997). As a secondary objective, we tested factors associated with insulin resistance that might contribute to laminitis, dyslipidemia, or osteochondrosis by measuring plasma concentrations of cortisol, triglyceride (TG), and IGF-I.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
Twelve Thoroughbred foals were raised on two adjacent, matched 12-ha pastures of mixed grass-legume (bluegrass, fescue, clover) with dietary groups in separate enclosures to allow for daily supplementation of the feed groups in their natural herd environment with minimal disturbance at feeding times. The adjacent pastures were similar in contents of DE and major nutrients according to equivalence tests (Byrd et al., 2005), and groups were rotated between pastures each month to minimize the potential effects of confounding dietary treatment and group. The horses received meals twice daily from birth, six receiving the SS feed (four fillies, two geldings), and six receiving the FF feed (three fillies, three geldings). Feeds were offered to groups of weanlings in individual pans, each containing a single portion. The pans were placed in a circular pattern approximately 6 m apart to minimize competitive behavior, and the weanlings were observed at each meal to ensure that each consumed their allotted supplement. Before weaning, foals shared meals with their dams. After weaning, the subjects had access to 1.6 kg of feed per meal, with two meals per day providing approximately 50% of their recommended DE (NRC, 1989). All weanlings were observed regularly by an experienced equine veterinarian and were considered to be in good health. All had similar BCS of 5 or 6 (Henneke et al., 1983). Before the study, the weanlings were gathered in the mornings to acclimate them to the stalls. The study was approved by the Institutional Animal Care and Use Committee.
Dietary Treatments and Experimental Design
The supplementary feeds were formulated (as-fed basis) to be isocaloric (8.37 ± 0.07 Mcal/d) and isonitrogenous (13.2 ± 0.7% CP), with vitamin and mineral contents designed to complement the pasture and fulfill present recommendations (NRC, 1989; Hoffman et al., 2001). Feed compositions were reported previously (Hoffman et al., 2003b). Samples of feeds and pasture were submitted monthly to a commercial laboratory for proximate and mineral analysis (Table 1). In a preliminary study on 12 yearlings, glycemic and insulinemic responses to a 2.0-kg meal of FF were 59 and 63% less than corresponding responses to SS (P = 0.030 and 0.020, respectively). Glycemic indices of approximately 129% (high) and 53% (moderate) were assigned to SS and FF, respectively, in relation to a standard of 100% for oats (Kronfeld et al., 2004).
|
). Three horses were assigned randomly to each of 4 d for sample collection. On the day before each test, the three subjects were weighed on an electronic scale (Tyrel Platform, model TC-10S; Allweights Hamilton Scale Corp., Richmond, VA). Catheters were inserted into a jugular vein, and the horses were kept in stalls overnight. The catheters were flushed with heparin in isotonic saline (10 U of heparin/mL of saline) that evening and again in the morning. To approximate a non-fasting grazing state, the stalled horses had access to orchardgrass hay and water ad libitum throughout the study. The nonstructural carbohydrate content of the hay was similar (P = 0.88) to that of the pasture. Hay consumption was of short duration and not expected to cause perturbations in plasma glucose or insulin (Stull and Rodiek, 1987; Pagan et al., 1999), while helping to avoid increases in insulin resistance possibly contributed by fasting or increased stress levels (Forehead and Dobson, 1997). Blood Sample Collection
The modified FSIGT was initiated at 0900 with a bolus of 300 mg/kg BW of glucose (Dextrose Solution 50%; Phoenix Pharmaceutical, Inc., St. Joseph, MO) rapidly administered (within 2 min) through the catheter. Twenty minutes after the glucose bolus, an insulin bolus (Humulin R, Eli Lilly and Co., Indianapolis, IN) of 1.5 mIU/kg BW was rapidly administered (within 30 sec) through the catheter.
Thirty-six venous samples were collected from each horse over the 6-h FSIGT. Basal samples were taken 60, 45, and 0 min before the glucose dose. Blood samples were drawn at 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16, 19, 22, 23, 24, 25, 27, 30, 35, 40, 50, 60, 70, 80, 90, 100, 120, 150, 180, 210, 240, 270, 300, 330, and 360 min after the glucose bolus. This schedule provides comprehensive curves for mathematical analysis, with more frequent sampling when rapid changes in glucose and insulin are expected. The blood samples were immediately transferred to heparinized sample tubes (Vacutainer; Fisher Health Care, Chicago, IL) and placed in ice water until centrifuged (3,000 x g for 10 min). Plasma was removed within 30 min of collection and frozen at –4°C until analysis.
Analysis of Glucose, TG, and Insulin in Blood
Plasma glucose and TG samples were analyzed by enzymatic assay (Beckman Instruments, triglyceride GPO reagent, Glucose Procedure No. 16-UV, Sigma Diagnostics, St. Louis, MO). Insulin was determined using an RIA (Coat-A-Count Insulin, Diagnostic Products Corp., Los Angeles, CA) previously validated for equine insulin (Freestone et al., 1991). Cortisol was determined using an RIA (Coat-A-Count Cortisol, Diagnostic Products) previously validated for equine cortisol (Alexander and Irvine, 1998). Insulin-like growth factor-I was extracted with acid-ethanol to remove binding proteins and analyzed by double-antibody RIA (Berry et al., 2001), as modified for the horse (Staniar, 2002). The intraassay CV of duplicate samples was <1% for glucose, 5% for insulin, 13% for IGF-I, and 5% for TG.
Minimal Model Analysis
The glucose and insulin curves were interpreted according to the minimal model of glucose and insulin dynamics as described by the following equations (Bergman et al., 1981):
 |
where G'(t) is the net rate (mg·dL–1·min–1) of change in plasma glucose. Glucose effectiveness (Sg) describes one component of this plasma disposal rate (min–1), which is the capacity of the cells to take up glucose without insulin mediation. The plasma glucose concentration (mg/dL) at time = t is G(t); Gb is the basal glucose concentration (mg/dL), maintained primarily by hepatic production. Insulin action, X(t), represents the insulin mediated component (min–1) of the plasma glucose disposal rate via the acceleration of glucose uptake in response to an increment change in the insulin concentration. This component is further described by:
 |
where X'(t) is the rate of change of the insulin action, p3 describes delivery of insulin to the interstitium, and p2 describes the disposal of insulin from the interstitial fluid, possibly reflecting hepatic extraction of plasma insulin. Insulin sensitivity (SI, L·min–1·mIU–1) is the ratio of these parameters: SI = p3/p2, and represents the efficiency of insulin to accelerate glucose uptake by the cells.
Responsiveness of ß-cells to the glucose load is described by the acute response of insulin to glucose (AIRg, mIU/[L·min]), which is the increase in plasma insulin above basal concentration integrated from 0 to 10 min after the glucose dose (Bergman, 1997). The product of AIRg and SI determines the disposition index (DI) or the appropriateness of the ß-cell response relative to the degree of insulin resistance in the tissue.
The minimal model was applied to the data using MinMod Millenium (Boston et al., 2003). This software iteratively fits the above equations to the sampling curves for plasma glucose and insulin, initially solving for the unknown parameters Sg, p2, and p3.
Basal Proxies
Basal values of insulin and glucose were used to calculate simple estimates of insulin sensitivity and ß-cell responsiveness for comparison with results from the minimal model (Treiber et al., 2005). The reciprocal of the insulin square-root index (RISQI) was calculated as insulin–0.5 and estimates insulin sensitivity as being relative to the amount of insulin compensation required to chronically maintain basal glucose homeostasis. The modified insulin response to glucose (MIRG) was calculated as (800 – 0.30[insulin – 50]2)/(glucose – 30), and estimates the capacity of the pancreatic ß-cells to increase insulin secretion and compensate for exogenous glucose. This capacity is limited by chronic basal insulin secretion, while chronic decompensation is indicated by increasing basal glucose levels.
Statistics
Statistics were performed using Intercooled Stata 8.0 (Stata Corp., College Station, TX). Assumption of normality was tested by the Shapiro-Wilkes statistic. For comparisons when normality was rejected (i.e., basal glucose, SI, and TG), the feed groups were analyzed by the nonparametric Kruskal-Wallis test, and data were reported as medians and 95% confidence intervals. All other comparisons were based on two-sample t-tests, and data were reported as arithmetic means and standard errors. Outliers were identified by Grubb’s test. Associations were determined by robust linear regression. Insulin secretion was calculated as the total area under the curve for insulin from 0 to 360 min and was calculated by trapezoidal approximation.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
Results for basal plasma metabolites and hormones are reported in Table 2. There was no difference between SS and FF treatments for basal concentrations of glucose, cortisol, or triglyceride. Higher basal IGF-I concentrations (P = 0.007) and a trend for higher insulin (P = 0.065) were observed in SS weanlings (Table 2). Values for solved minimal-model parameters and proxies are reported in Table 3.
|
|
). Glucose peaked at 1 min (350 ± 6 mg/dL) and decreased rapidly, returning to baseline at approximately 180 min. Insulin concentrations during the FSIGT were greater (P < 0.055) in the SS weanlings at 11 of the 36 time points (Figure 1B). Insulin concentrations did not increase (P = 0.17) immediately following the insulin dose, as the dose was small relative to physiologic concentrations of insulin. Total insulin secretion, the area under the curve, was 44% (P = 0.044) greater in SS than in FF weanlings (Table 3
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
The minimal model fit the data from all 12 weanlings, despite the low insulin dose, demonstrating the robustness of the minimal model (Boston et al., 2003). The minimal model FSIGT was originally designed for use without an insulin dose and only later modified to improve precision, primarily for diabetic patients (Pacini et al., 1998). Previous administration of 30 mIU/kg insulin to mature geldings obscured the secondary insulin response (Hoffman et al., 2003). Small insulin doses therefore allow for better characterization of endogenous metabolite and hormone responses to the glucose challenge. Nonetheless, in subjects with lower insulin sensitivity or insulin response than those in this study, small insulin doses may be inadequate to provide significant responses or return glucose concentrations to baseline, both of which are necessary to solve for minimal model parameters. For future minimal model studies in horses, we generally recommend an insulin bolus of 20 mIU/kg.
An intraclass correlation of 0.91 demonstrated the repeatability of the minimal model on eight Standard-bred horses (Geor et al., 2005). In addition, a correlation of 0.87 was found between the different indices of insulin sensitivity measured by the minimal model and another quantitative method, the euglycemic-hyperinsulinemic clamp (Pratt et al., 2005). As the first specific quantitative method for determining insulin sensitivity, the euglycemic-hyperinsulinemic clamp has been regarded as a historical "gold standard." However, the minimal model is now being widely accepted based on its merits as a more physiological test of the glucose-insulin regulatory system (Bergman, 1989). The correlation of 0.87 between clamp- and minimal model-derived insulin sensitivities in horses is comparable to correlations of 0.84, 0.82, and 0.91 between the two methods in humans, dogs, and cats, respectively (Finegood et al., 1984; Beard et al., 1986; Petrus et al., 1998).
A trend towards decreased SI was demonstrated previously in mature Thoroughbred geldings with normal BCS when adapted to the SS diet (Hoffman et al., 2003a). Insulin sensitivity also has been observed to decrease in humans after 4 mo of adaptation to high-glycemic diets and to increase in adaptation to low-glycemic diets (Wolever and Mehling, 2002). Decreased SI observed in SS-adapted weanlings indicates a change in insulin signaling affecting glucose entry into the cell. This change may involve interactions involving the surface of the cell, signal transduction pathways, or intracellular metabolism of glucose and fatty acids (Saltiel and Kahn, 2001; Kronfeld et al., 2005).
Insulin resistance is a normal response of the metabolic system to decreased energy availability or increased energy demand (Brand-Miller and Colagiuri, 1999). This adaptive strategy may be involved in the increased insulin resistance observed during fasting and pregnancy (Fowden et al., 1984; Forehead and Dobson, 1997). Rapid decreases in circulating glucose and insulin following meals high in sugar and starch could trigger similar energy-conserving regulation (Jenkins et al., 1987). These signals may affect glucose regulation and decrease the insulin sensitivity of the tissue (Yalow et al., 1969). The combination of low insulin sensitivity with high carbohydrate availability has been shown to elicit a variety of metabolic abnormalities, with implications to numerous disease states in many species (Shafrir and Ziv, 1998). Animal models similar to the one in the present study provide a useful window into the effects of diet on the well-conserved and complex system of glucose regulation, with implications to both animal and human health and welfare (Storlien et al., 2000).
In the present study, glucose effectiveness did not differ between dietary groups. This similarity suggests that Sg, the glucose-mediated component of glucose uptake, is not regulated by adaptation to a feed high in glucose equivalents. Similarly, Holstein or Jersey calves showed no difference in Sg when fed milk replacer as once- or twice-daily meals (Stanley et al., 2002). Humans adapted to diets with different glycemic indices also showed no significant change in Sg from baseline (Wolever and Mehling, 2002). Nevertheless, noninsulin-dependent GLUT-1 glucose transporters may represent a component of Sg and are capable of compensatory changes when nutrition is inadequate (Sadiq et al., 1999; Flanagan, 2000). In cases of extreme insulin resistance, Sg also may be required to compensate, so as to sufficiently clear glucose from the plasma (Hoffman et al., 2003a). Although consistent for all weanlings, Sg was lower in these weanlings than values reported for older horses and other species (Feldhahn et al., 1999; Stanley et al., 2002; Hoffman et al., 2003a).
Thoroughbred weanlings adapted to the SS meal had lower RISQI than weanlings adapted to the FF meal. These findings further indicate lower insulin sensitivity in the SS group and, along with strong correlations, demonstrate the capacity of basal proxies to estimate properties of the dynamic system. A number of studies of proxies or surrogates based on basal glucose and insulin values have been useful in public health (Legro et al., 1998; Gungor et al., 2004).
Lower SI in SS foals tended to be associated with increased AIRg, which represents compensation for the decreased efficiency of insulin to stimulate cellular responses. Similar compensation has been reported in apparently healthy mature Thoroughbred geldings, calves, and humans (Welch et al., 1990; Stanley et al., 2002; Hoffman et al., 2003a). Excessive compensation and elevated circulating insulin may be a precursor to metabolic disorders, such as laminitis and hyperlipidemia or osteochondrosis (Jeffcott and Field, 1985; Ralston, 1996).
Because of the inverse relationship of AIRg and SI, their product, DI, did not differ between dietary groups, suggesting that each group’s insulin response was appropriate to its insulin sensitivity. Similarly, final DI values did not differ between two human groups fed low or high-glycemic diets, indicating appropriate compensation in the high-glycemic group (Wolever and Mehling, 2002).
Consistent basal glucose concentrations across all weanlings suggests that hay consumption was either not variable before the study or had little effect on plasma glucose concentrations, as was predicted. No difference in circulating cortisol was observed between dietary groups, indicating that insulin resistance in Thoroughbred weanlings was not a result of changes in glucocorticoids, as have been associated with insulin resistance, obesity, and laminitis (Forehead and Dobson, 1997; Reeves et al., 2001; Johnson, 2002). Plasma TG concentrations showed large variations among individuals and were not different between dietary groups; however, plasma IGF-I concentrations were increased in SS-adapted weanlings, possibly reflecting altered insulin signaling on the somatotropic axis (Bereket et al., 1994; Smith et al., 1997) and additional metabolic regulatory hormones, such as growth hormone (Roth et al., 1963; Champion et al., 2000; Yakar et al., 2001). Increased IGF-I may be a risk factor for osteochondrosis and could explain the association between osteochondrosis, insulin resistance, and high-carbohydrate diets observed in growing horses (Ralston, 1996).
A recent study on quarter horse yearlings compared response to meals of a high-glycemic feed and a feed with 10% fat after 30- and 60-d adaptation (Ropp et al., 2003). Results were inconsistent, showing a tendency for a lower glucose response on the fat-added feed at 30 d only, greater insulin in the high-glycemic group at 60 d only, differences in GH secretion at 30 and 60 d, but no significant differences in IGF-I. Emphasis was placed on the IGF-I result, and it was concluded that added fat did not decrease the risk of skeletal abnormalities as has been suggested by others (Saastamoinen, 1996; Harris et al., 2004). This conclusion may be questioned, however, because of the similar glycemic responses to the feeds, probably due in part to the high starch content (24%) of the 10% fat diet. In addition, there were significant, albeit inconsistent, changes in responses of plasma insulin and growth hormone, which, like IGF-I, are components of the somatotropic axis regulating chondrocyte function and maturation (Thissen et al., 1994) and may therefore play a role in the development of osteochondrosis.
In summary, adaptation to a high-glycemic diet is associated with increased insulin resistance and a compensatory increase in insulin secretion. Increased insulin secretion maintains glucose homeostasis, but may alter insulin signaling in other systems, such as the somatotropic axis and increase the risk of metabolic disorder. This condition may be partially amenable to dietary management (Kronfeld et al., 2005).
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
|
Footnotes |
|---|
2 Correspondence: Virginia Tech MARE Center, 5527 Sullivans Mill Road, Middleburg 20117 (phone: 540-687-3521; fax: 540-687-5362; e-mail: ktreiber{at}vt.edu).
Received for publication April 13, 2005. Accepted for publication June 9, 2005.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionImplicationsLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Van Weyenberg, M. Hesta, I. D. Kalmar, J. Vandermeiren, and G. P. J. Janssens Nutritional management of laminitis in a horse Vet Rec., May 30, 2009; 164(22): 694 - 695. [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. Vasconcelos, J. E. Sawyer, L. O. Tedeschi, F. T. McCollum, and L. W. Greene Effects of different growing diets on performance, carcass characteristics, insulin sensitivity, and accretion of intramuscular and subcutaneous adipose tissue of feedlot cattle J Anim Sci, April 1, 2009; 87(4): 1540 - 1547. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Vervuert, K. Voigt, T. Hollands, D. Cuddeford, and M. Coenen Effects of processing barley on its digestion by horses Vet Rec., May 24, 2008; 162(21): 684 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. H. Treiber, R. J. Geor, R. C. Boston, T. M. Hess, P. A. Harris, and D. S. Kronfeld Dietary Energy Source Affects Glucose Kinetics in Trained Arabian Geldings at Rest and during Endurance Exercise J. Nutr., May 1, 2008; 138(5): 964 - 970. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Ordakowski-Burk, R. W. Quinn, T. A. Shellem, and L. R. Vough Voluntary intake and digestibility of reed canarygrass and timothy hay fed to horses J Anim Sci, November 1, 2006; 84(11): 3104 - 3109. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Kronfeld, K. H. Treiber, T. M. Hess, R. K. Splan, B. M. Byrd, W. B. Staniar, and N. W. White Metabolic Syndrome in Healthy Ponies Facilitates Nutritional Countermeasures against Pasture Laminitis J. Nutr., July 1, 2006; 136(7): 2090S - 2093S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. H. Treiber, D. S. Kronfeld, and R. J. Geor Insulin Resistance in Equids: Possible Role in Laminitis J. Nutr., July 1, 2006; 136(7): 2094S - 2098S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Harris, S. R. Bailey, J. Elliott, and A. Longland Countermeasures for Pasture-Associated Laminitis in Ponies and Horses J. Nutr., July 1, 2006; 136(7): 2114S - 2121S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or fat and fiber (FF, •). The insets within each of the large panels provide details of the first 40 min of glucose (A) or insulin (B) changes on an expanded scale. Asterisks indicate differences between dietary groups, P < 0.055.