Evaluation of beta-cell sensitivity to glucose and first-phase insulin secretion in obese dogs

Kurt R. Verkest Centre for Companion Animal Health, School of Veterinary Science, The University of Queensland, St Lucia, QLD 4072, Australia.

Search for other papers by Kurt R. Verkest in
Current site
Google Scholar
PubMed
Close
 BVSc
,
Linda M. Fleeman Centre for Companion Animal Health, School of Veterinary Science, The University of Queensland, St Lucia, QLD 4072, Australia.

Search for other papers by Linda M. Fleeman in
Current site
Google Scholar
PubMed
Close
 BVSc, PhD
,
Jacquie S. Rand Centre for Companion Animal Health, School of Veterinary Science, The University of Queensland, St Lucia, QLD 4072, Australia.

Search for other papers by Jacquie S. Rand in
Current site
Google Scholar
PubMed
Close
 BVSc, DVSc
, and
John M. Morton Centre for Companion Animal Health, School of Veterinary Science, The University of Queensland, St Lucia, QLD 4072, Australia.

Search for other papers by John M. Morton in
Current site
Google Scholar
PubMed
Close
 BVSc, PhD

Abstract

Objective—To compare beta-cell sensitivity to glucose, first-phase insulin secretion, and glucose tolerance between dogs with naturally occurring obesity of > 2 years' duration and lean dogs.

Animals—17 client-owned obese or lean dogs.

Procedures—Frequently sampled IV glucose tolerance tests were performed with minimal model analysis on 6 obese dogs and matched controls. Glucagon stimulation tests were performed on 5 obese dogs and matched controls.

Results—Obese dogs were half as sensitive to the effects of insulin as lean dogs. Plasma glucose concentrations after food withholding did not differ significantly between groups; plasma insulin concentrations were 3 to 4 times as great in obese as in lean dogs. Obese dogs had plasma insulin concentrations twice those of lean dogs after administration of glucose and 4 times as great after administration of glucagon. First-phase insulin secretion was greater in obese dogs.

Conclusions and Clinical Relevance—Obese dogs compensated for obesity-induced insulin resistance by secreting more insulin. First-phase insulin secretion and beta-cell glucose sensitivity were not lost despite years of obesity-induced insulin resistance and compensatory hyperinsulinemia. These findings help explain why dogs, unlike cats and humans, have not been documented to develop type 2 diabetes mellitus.

Abstract

Objective—To compare beta-cell sensitivity to glucose, first-phase insulin secretion, and glucose tolerance between dogs with naturally occurring obesity of > 2 years' duration and lean dogs.

Animals—17 client-owned obese or lean dogs.

Procedures—Frequently sampled IV glucose tolerance tests were performed with minimal model analysis on 6 obese dogs and matched controls. Glucagon stimulation tests were performed on 5 obese dogs and matched controls.

Results—Obese dogs were half as sensitive to the effects of insulin as lean dogs. Plasma glucose concentrations after food withholding did not differ significantly between groups; plasma insulin concentrations were 3 to 4 times as great in obese as in lean dogs. Obese dogs had plasma insulin concentrations twice those of lean dogs after administration of glucose and 4 times as great after administration of glucagon. First-phase insulin secretion was greater in obese dogs.

Conclusions and Clinical Relevance—Obese dogs compensated for obesity-induced insulin resistance by secreting more insulin. First-phase insulin secretion and beta-cell glucose sensitivity were not lost despite years of obesity-induced insulin resistance and compensatory hyperinsulinemia. These findings help explain why dogs, unlike cats and humans, have not been documented to develop type 2 diabetes mellitus.

Obese dogs, like obese humans, have components of the metabolic syndrome such as insulin resistance.1 However, type 2 diabetes mellitus, an obesity-associated disease, has not been rigorously documented in dogs.2 Type 2 diabetes mellitus occurs when insulin secretion fails to compensate for obesity-induced insulin resistance.3,4 Obesity-induced insulin resistance occurs in humans,3,4 cats,5 and dogs,1 and an associated failure of insulin secretion has been reported in humans and cats5 but not dogs.2 Well-documented studies revealing obesity-induced insulin resistance with failure of insulin secretion in dogs are therefore lacking.

Loss of the first phase of insulin secretion is an important early marker of beta-cell failure.6 In normal lean humans7 and dogs,8 the first phase of insulin secretion begins within 2 minutes of IV administration of glucose and continues for approximately 10 minutes. The first phase of insulin secretion rapidly increases insulin concentration in the interstitial space, where insulin acts,9 and inhibits hepatic glucose output.8 If first-phase insulin secretion is impaired, blood glucose is higher after a meal or glucose challenge,10,11 and the second phase of insulin secretion is increased.11 Therefore, when evaluating beta-cell function by use of an IV glucose tolerance test, it is important to obtain sufficient early samples to document first-phase insulin secretion.

Although the insulin secretion patterns of clinically normal dogs are documented,12 the insulin secretion patterns of dogs with naturally occurring obesity have not been described adequately. Previous studies of insulin secretion in obese dogs have not included enough early samples to assess first-phase insulin secretion,13,14 involved dogs with induced short-term obesity15–17 or used sexually intact bitches.5,13,14 Dogs with induced obesity might not be an adequate model of naturally occurring obesity because the high-fat diets used to induce obesity result in hypertriglyceridemia, which can impair insulin secretion.17 Because obesity is usually induced relatively rapidly, there is less time for beta cells to compensate for chronic insulin resistance than with naturally occurring obesity. Sexually intact bitches have a prolonged luteal phase that can result in a form of acromegaly.18 Although bitches in the luteal phase are insulin resistant, their insulin secretion patterns might not be similar to obese dogs in other ways.

It is important to assess insulin secretion in response to nonglucose secretagogues when evaluating beta-cell function. In humans with type 2 diabetes, nonglucose secretagogues such as glucagon can stimulate insulin secretion even when there is no insulin secretion in response to hyperglycemia.19 This indicates that beta cells are present but lack glucose sensitivity in individuals with type 2 diabetes.20 Glucagon stimulation tests have been validated to characterize insulin secretion in dogs,21 but glucagon-stimulated insulin secretion has not been described in obese dogs. Glucagon stimulation tests in obese dogs might help to identify whether inadequate insulin secretion, if present, is the result of inadequate beta cells or beta cell glucose insensitivity.

First-phase insulin secretion and insulin secretion in response to both glucose and nonglucose secretagogues have not been reported to date in dogs with naturally occurring obesity. The purpose of the study reported here was to compare insulin secretion and glucose tolerance following IV administration of glucose, and insulin secretion following IV administration of glucagon, in dogs with naturally occurring obesity with that in lean dogs.

Materials and Methods

Animals—Two cohort studies were conducted in which insulin secretion was compared between matched obese (n = 6) and lean (6) client-owned dogs by use of an IV glucose tolerance test and a glucagon stimulation test (5 obese dogs, 5 lean dogs; Table 1). Seventeen dogs were enrolled; 2 obese and 3 lean dogs were enrolled in both studies, separated by 12 months. Dogs were selected from clients of a private veterinary practice in Brisbane, Australia. Obese dogs were selected from those reported by the owners or documented in clinical records as having been obese for at least 2 years and whose body condition score was 8 or 9 on a 9-point scale.22 Lean dogs were selected from the same client population, had body condition scores of 4 or 5 on a 9-point scale, and were 1:1 matched within each study for sex, neuter status, and age category. Age categories were young adult (2 to 3.9 years), middle-aged (4 to 6.9 years), and older (7 to 10 years). Dogs were excluded if they had clinical signs or history of current or recent illness, were > 10 years of age, weighed < 15 kg (for those that received glucose tolerance tests) or < 10 kg (for those that received glucagon stimulation tests), or were sexually intact females or neutered < 12 months previously. A physical examination was performed and detailed history taken for each dog, but no laboratory testing was performed prior to recruitment. The project was approved by the University of Queensland Animal Ethics Committee, and informed consent was obtained from each dog's owner at the time of recruitment.

Table 1—

Variables in dogs evaluated in glucose tolerance tests (n = 6/group) and glucagon stimulation tests (n = 5/group).

 Glucose tolerance testGlucagon stimulation tests  
VariableObeseLeanObeseLean
BCS8.54.58.54.0
Age (y)778.58.5
Weight (kg)45 ± 18.422 ± 4.436 ± 14.524 ± 6.6
Lean weight (kg)29 ± 12.118 ± 3.723 ± 10.321 ± 6.0

Body condition score (BCS) was assessed on a 9-point scale.22 Values are median or mean ± SE. Estimated lean weight was calculated by subtracting the mean percentage body fat mass expected for dogs given their sex and BCS.

Glucose tolerance tests—Food was withheld from dogs overnight before a jugular cathetera was placed by use of a modified Seldinger technique23 under general anesthesia. The glucose tolerance test was performed no less than 24 hours after anesthetic recovery. Blood samples of 1.5 mL were collected 15, 10, 5, and 1 minute before and 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 19, 22, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 110, 120, 140, 160, and 180 minutes after IV administration of glucoseb (1 g/kg of body weight). Glucose intolerance was considered to be present when plasma glucose concentration was greater than baseline by more than the 90% range of differences24 at 2 hours during the glucose tolerance test. The 90% range of differences is a measure of variability within an individual. Glucose-intolerant dogs were further assessed for causes of insulin resistance and impaired insulin secretion.

Glucagon stimulation tests—Food was withheld for 12 hours. Then, blood samples were collected via a cephalic vein catheter23,c immediately before and 5, 10, 20, 30, and 60 minutes after IV injection of 1 unit of glucagon.d

Sample handling and assays—Blood samples were collected into tubes containing EDTA,e chilled immediately, and centrifuged within 5 minutes. Plasma aliquots were kept at 4°C until the conclusion of the test and then stored at −80°C until assayed.

Samples were assayed in 2 batches: after completion of all glucose tolerance tests and after completion of all glucagon stimulation tests. Plasma insulin concentrations were determined via radioimmunoassayf validated for use in dogsg with inter- and intra-assay coefficients of variation of 16% and 5%, respectively, at 22.5 μU/mL and < 23% and 5% at 66 μU/mL. Plasma glucose concentrations were determined by use of the glucose oxidase methodh with inter- and intra-assay coefficients of variation of < 1% at 4.8 mmol/L (86 mg/dL) and 3% and < 1% at 30 mmol/L (540 mg/dL).

Additional testing of glucose-intolerant dogs—Insulin secretion in glucose-intolerant dogs was evaluated with a glucagon stimulation test and compared with the results of glucagon stimulation tests in the other dogs. Exocrine pancreatic disease was assessed by determination of serum pancreatic lipase immunoreactivityi and trypsin-like immunoreactivityj performed after food withholding in baseline samples from glucose tolerance and glucagon stimulation tests. Hypothyroidism and hyperadrenocorticism were tested for by assay of free thyroxine concentration via equilibrium dialysis on 2 occasions and an ACTH stimulation test, respectively. A screen for concurrent systemic disease was performed by use of a CBC and serum biochemical panel. Plasma triglyceride concentrations were measured in baseline samples from the glucose tolerance tests. The medical records and history were examined for access to drugs known to affect insulin sensitivity or secretion.

Derived variables and statistical analyses—Mean plasma glucose and insulin concentrations were calculated as the area under the time versus concentration curve > 0 μU/mL (by use of the trapezoidal method) divided by the time interval.24,25 For each dog, measurements that differed significantly from baseline were identified by use of the 90% range of differences. Assuming each individual dog had the same underlying variance among measurements as the pooled variance estimate based on all dogs, and that differences among measurements were normally distributed, for any 1 dog, the 90% range of differences for glucose and log-transformed insulin concentrations was calculated separately for lean and obese dogs by use of the following formula24:

article image

where 1.734 is the 2-tailed critical value of the t-distribution at the 0.10 level of significance and 18 df (24 baseline samples from 6 dogs in each group). Within-dog variance used to calculate 90% range of differences was estimated as the residual mean square from ANOVA of the 4 baseline values obtained for the glucose tolerance test after accounting for between-dog variation in the ANOVA model.

Times taken to exceed and to return to baseline plasma glucose and insulin concentrations were defined as the times taken for the glucose and insulin concentrations to exceed and return to less than the mean baseline concentration for each dog plus the 90% range of differences, respectively. When log-transformed data were used, the back-transformed 90% range of differences was multiplied by the mean baseline concentration rather than added.

Glucose metabolism was assessed during the glucose tolerance test by calculating the insulin sensitivity index,k glucose half-life between 10 and 45 minutes,26 and time to return to baseline plasma glucose concentration. Glucose half-life was calculated from the inverse of the absolute value of the slope of the least squares regression line of log-transformed plasma glucose concentrations between 10 and 45 minutes after glucose administration.26

Insulin secretion during the glucose tolerance tests was assessed by calculation of the acute insulin response to glucose,k the disposition index,k time of onset of insulin secretion, time to return to baseline plasma insulin concentration, and mean plasma insulin concentration25 for each dog during the first (0 to 5 and 0 to 10 minutes) and second (10 to 30 and 30 to 60 minutes) phases of insulin secretion. The acute insulin response to glucose assessed first-phase insulin secretion and was calculated as the area under the time versus plasma insulin concentration curve that was greater than baseline for the first 10 minutes after glucose administration. Adequacy of insulin secretion for the prevailing insulin sensitivity was assessed by use of the disposition indexk (product of the insulin sensitivity index and acute insulin response to glucose). The association and 95% prediction bands between insulin sensitivity and first-phase insulin secretion were assessed by use of nonlinear correlationl to the hyperbolic equation:

article image

in which C is the constant and the regression coefficient. Glucose-intolerant dogs were not included when the prediction bands were calculated.

The glucose and insulin secretory responses to glucagon were assessed by use of the mean and peak plasma glucose and insulin concentrations for each dog from 0 to 60 minutes, respectively. The times of the plasma glucose and insulin peak concentrations were determined individually for each dog.

Kaplan-Meier survivor functions for times to peak and return to baseline were compared between obese and lean dogs by use of log-rank tests stratified on pair. Data from dogs that did not return to baseline during the glucose tolerance tests were censored at 180 minutes when times to return to baseline were analyzed. Time versus plasma glucose concentration curves were compared between obese and lean dogs by use of random-effects linear regression27 with dog fitted as a random effect and time point fitted as a fixed effect. Because the dose of glucose administered to each dog was calculated on the basis of actual body weight, peak plasma glucose concentrations were higher for the obese than for the lean dogs. Therefore, 3-minute plasma glucose concentration was fitted as a continuous variable in the linear regression model to account for this. Predicted means and associated SEs from the model were calculated for each time point separately for obese and lean dogs. To allow valid comparisons between obese and lean dogs, predicted mean plasma glucose concentrations at each time point after 3 minutes were calculated separately for the obese and lean groups and adjusted for 3-minute plasma glucose concentration, which was set at the mean value for the lean group.

Differences between obese and lean dogs in all other variables were assessed by use of ANOVA of untransformed data (for glucose tolerance tests: glucose half-life, baseline, mean, and peak plasma insulin concentrations, insulin sensitivity, and disposition index; for glucagon stimulation tests: baseline, mean, and peak plasma glucose concentrations, 60-minute plasma glucose and insulin concentrations, and peak plasma insulin concentrations), and ANOVA of log-transformed data if untransformed data did not appear to be normally distributed (for glucose tolerance tests: baseline and peak plasma glucose concentrations; for glucagon stimulation tests: baseline and mean plasma insulin concentrations), with group (obese or lean) and pair fitted as fixed effects. Ninety-five percent confidence limits for the estimated differences between means were calculated by use of the SE of those estimates. For log-transformed variables, 95% confidence intervals for differences between means (obese group minus lean group) were estimated by use of the transformed data; then the limits of the confidence intervals were back-transformed, and each limit was multiplied by the arithmetic mean of that variable in lean dogs.28 All statistical calculations were performed by use of proprietary software.l,m Results are presented as mean ± SEM or as median and range. Values of P < 0.05 were considered significant.

Results

Plasma glucose concentrations did not differ significantly between obese and lean dogs (Tables 2 and 3). After glucose administration, obese dogs had higher unadjusted plasma glucose concentrations at peak glucose concentrations. This was because the glucose dose was not adjusted for differences in the volume of distribution of glucose associated with adiposity. The volume of distribution of glucose does not increase linearly with increasing adiposity, so obese dogs were relatively overdosed with glucose. After accounting for the difference in peak plasma glucose concentrations by adjusting for 3-minute plasma glucose concentration, there were no significant differences in plasma glucose concentrations between obese and lean dogs at any subsequent time point (Figure 1). Times to return to baseline plasma glucose concentration and glucose half-life were not significantly different between obese and lean dogs.

Figure 1—
Figure 1—

Mean ± SEM plasma glucose concentrations in obese (solid line) and lean (dashed line) dogs before and after administration of glucose at 1 g/kg of body weight at time 0 (A; n = 6 dogs/group), predicted mean plasma glucose concentrations after glucose administration after adjusting plasma glucose concentrations in obese dogs for the difference in peak plasma glucose concentrations (B; 6 dogs/group), and mean plasma glucose concentrations before and after administration of 1 unit of glucagon at time 0 (C; 5 dogs/group).

Citation: American Journal of Veterinary Research 72, 3; 10.2460/ajvr.72.3.357

Table 2—

Variables (mean ± SEM or median [range]) in lean (n = 6) and obese dogs (6) during glucose tolerance tests.

VariableObese dogsLean dogsDifference (95% Cl)P value
Baseline plasma glucose concentration (mg/dL)92.5 ± 588.4 ± 4.54.2 (−8.1 to 14)0.44
Peak (3 min) plasma glucose concentration (mg/dL)920 ± 68684 ± 41222 (−2.2 to 519)0.05
Glucose half-life (min)27 ± 1420 ±47 (−5 to 19)0.22
Time to return to baseline plasma glucose concentration (min)70 (50 to > 180)65 (50−90)ND1.00
Baseline plasma insulin concentration (μU/mL)24 ± 2.46 ± 1.319 (12 to 26)0.01
Time to onset of insulin secretion (min)< 2 (< 2)< 2 (< 2)NDND
Peak plasma insulin concentration (μU/mL)174 ± 26101 ± 3073 (−23 to 168)0.11
Time to peak plasma insulin concentration (min)10 (3 to 25)15 (2−40)ND0.10
Time to return to baseline plasma insulin concentration (min)80 (60 to > 180)85 (70−110)ND0.41
Insulin sensitivity (× 10−4L·mU−1· min−1)3.0 ± 0.86.5 ± 1.5−3.5 (−6.6 to −0.3)0.04
Disposition index2,893 ±7112,431 ± 418462 (−792 to 1,716)0.91

CI = Confidence interval. ND = Not determined (time to onset of insulin secretion was < 2 minutes in all dogs, so differences between obese and lean dogs could not be tested).

Table 3—

Variables (mean ± SEM or median [range]) in obese (n = 5) and lean (n = 5) dogs during glucagon stimulation tests.

VariableObese dogsLean dogsDifference (95% CI)P value
Baseline plasma glucose concentration (mg/dL)99 ± 2.395 ± 34 (−4 to 12)0.26
Mean plasma glucose concentration (mg/dL)214 ± 17.2172 ± 14.843 (−36 to 121)0.21
Peak plasma glucose concentration (mg/dL)253.8 ± 15.7206.3 ± 19.648 (−40 to 135)0.21
Time to peak plasma glucose concentration (min)30 (10–30)30 (30–60)ND0.16
Plasma glucose concentration 60 min after glucagon (mg/dL)182 ± 33130 ± 1153 (−60 to 165)0.28
Baseline plasma insulin concentration (μU/mL)35 ± 3.111 ± 3.029 (12 to 57)0.01
Mean plasma insulin concentration (μU/mL)116 ± 9.329 ± 4.590 (66 to 121)0.005
Peak plasma insulin concentration (μU/mL)182 ± 23.345 ± 7.4137 (88 to 186)> 0.001
Time to peak plasma insulin concentration (min)20 (10–20)20 (20–20)ND0.32
Plasma insulin concentration 60 min after glucagon (μU/mL)74 ± 2114 ± 3.460 (−1.8 to 122)0.05

See Table 2 for key.

After glucagon administration, peak and mean plasma glucose concentrations were higher in obese dogs than in lean dogs, but the differences were not significant (P = 0.21; Figure 1; Table 3). Peak plasma glucose concentrations were attained after a median of 30 minutes in both groups.

Several differences between lean and obese dogs were identified in the insulin variables. After food withholding, plasma insulin concentrations were 3 to 4 times as great in the obese dogs as in the lean dogs (Figure 2; Tables 2 and 3). In the glucose tolerance tests, mean first-phase plasma insulin concentrations in obese dogs were more than twice those in lean dogs (Figure 3). Plasma insulin concentrations in the obese dogs were not significantly higher during the second phase of insulin secretion from 10 to 60 minutes.

Figure 2—
Figure 2—

Mean ± SEM plasma insulin concentrations in the same dogs as in Figure 1. A—Glucose administered at time 0. B—Glucagon administered at time 0.

Citation: American Journal of Veterinary Research 72, 3; 10.2460/ajvr.72.3.357

Figure 3—
Figure 3—

Mean ± SEM plasma insulin concentrations at baseline and during selected time intervals after glucose administration (1 g/kg) in lean (n = 6; white bars) and obese (6; black bars) dogs. P values indicate differences between groups.

Citation: American Journal of Veterinary Research 72, 3; 10.2460/ajvr.72.3.357

After glucagon administration, obese dogs attained peak and mean plasma insulin concentrations that were 4 times those of lean dogs (Figure 2; Table 3). Peak plasma insulin concentrations were attained at 20 minutes in all dogs except 1 obese dog in which concentration peaked at 10 minutes. In obese and lean dogs, peak plasma insulin concentrations were attained earlier than peak plasma glucose concentrations (median difference, 10 minutes). Plasma insulin concentrations returned to baseline values in all lean dogs but in only 2 of the 5 obese dogs by 60 minutes after glucagon administration. Plasma insulin concentrations were typically twice those of baseline values in the obese dogs at 60 minutes after glucagon administration.

Obese dogs were less than half as insulin sensitive as lean dogs (Table 2). The disposition index, a measure of how well first-phase insulin secretion compensates for prevailing insulin sensitivity in the glucose tolerance test, was not reduced in obese dogs. When insulin sensitivity was plotted against first-phase insulin secretion, all glucose-tolerant obese and lean dogs were within the 95% prediction band (Figure 4). The 95% prediction band encloses the area around the best-fit curve that is predicted to contain 95% of data points. One glucose-intolerant obese dog had insulin concentrations that were a mean of 370 μU/mL less than the 95% prediction band throughout first-phase insulin secretion.

Figure 4—
Figure 4—

First-phase insulin secretion plotted against insulin sensitivity in lean dogs (squares), obese dogs (circles), and a single obese dog (triangle). The shaded area between the dashed lines indicates the 95% prediction band around the curve of best fit (solid line).

Citation: American Journal of Veterinary Research 72, 3; 10.2460/ajvr.72.3.357

Only 1 obese dog was glucose intolerant during the glucose tolerance test (2-hour plasma glucose concentration, 247 mg/dL). That dog's glucose intolerance was associated with extreme insulin resistance (insulin sensitivity index was 0.35, compared with a mean value of 3.5 ± 0.7 × 10−4L·mU−1·min−1 for all other obese dogs). In addition, insulin secretion was not proportionally increased (first phase mean insulin concentration was 98 μU/mL, compared with a mean value of 129 ± 22 μU/mL for all other obese dogs), so that the disposition index was less than that of all other dogs (271, compared with a mean value of 2,879 ± 367 for all other dogs). The glucose tolerance test for that dog had a pattern of rapid first-phase secretion and a late second phase of insulin secretion that was maintained beyond 2 hours after the glucose dose was administered, with peak plasma insulin concentrations attained at 3 minutes. That dog had plasma insulin concentrations that were similar to those of other obese dogs during the first 8 minutes after the glucose dose was administered, were the lowest or second-lowest of the obese dogs from 8 to 50 minutes, and were less than the mean value for the obese dogs during the first hour after the glucose dose was administered. After 100 minutes, that dog's plasma insulin concentrations increased and were greater than those of all the other dogs. Insulin secretion in response to glucagon administration was less than all other obese dogs and similar to lean dogs. Follow-up investigation of this glucose-intolerant dog did not identify any underlying endocrinopathies or other conditions that might have accounted for these differences. The ACTH-stimulated plasma cortisol concentration (412 nmol/L; reference range, < 470 nmol/L) and plasma concentrations of free thyroxine (18.6 and 22.2 pmol/L; reference range, 10 to 45 pmol/L), pancreatic lipase immunoreactivity (7.6 and 24.5 μg/L; reference range, < 100 μg/L), and food-withheld triglyceride concentrations were within reference ranges, as were routine serum biochemical and hematologic values (data not shown). Plasma trypsin-like immunoreactivity concentrations were less than the reference range initially but not subsequently (4.2 and 9.6 μg/L; reference range, > 5 μg/L) and were not in the range diagnostic of exocrine pancreatic insufficiency (< 2.5 μg/L). There was no history of access to progestagens, serotonin reuptake inhibitors or other antipsychotic medications, cyclosporine, or corticosteroids. Acromegaly, subclinical urinary tract infection, pancreatic cancer, type 1 diabetes mellitus, and pancreatic amyloidosis were not specifically tested for, but acromegaly is rare in male dogs, and no evidence of these diseases was identified during 6 years of follow-up.

Discussion

The principal finding of this study was that unlike humans and cats, obese dogs did not lose first-phase insulin secretion in response to IV administration of glucose. Despite years of obesity, the first phase of insulin secretion was exaggerated. In contrast, obese cats have glucose intolerance and absent or reduced first-phase insulin secretion.29 Loss of first-phase insulin secretion is one of the early markers of the onset of type 2 diabetes mellitus. Although dogs are used as models of the human metabolic syndrome,1,12 the present results indicate that spontaneously obese dogs, including the dog that was glucose intolerant, do not lose first-phase insulin secretion.

Adequate compensation for obesity-induced insulin resistance in 5 of the 6 dogs in this study was indicated by glucose tolerance and the disposition index. The disposition index, which is decreased if insulin secretion is inadequate for prevailing insulin sensitivity, was not reduced in the glucose tolerant, insulin-resistant obese dogs, and all but one of the obese dogs had plasma glucose concentrations that had returned to baseline within 2 hours. The obese dog that had glucose intolerance and a decreased disposition index retained first-phase insulin secretion. Like obese cats29 and humans,20 the obese dogs in the present study were insulin resistant, but unlike obese humans and cats, most of the obese dogs compensated adequately.

In contrast to humans and cats, dogs have not been convincingly found to develop beta-cell failure leading to type 2 diabetes associated with obesity-induced insulin resistance. Type 2 diabetes involves acquired failure of beta-cell function with or without loss of beta-cell mass.20 Proposed mechanisms contributing to beta-cell failure in type 2 diabetes include exposure to increased glucose concentrations leading by various pathways to glucose toxicosis,30 exposure to excessive triglycerides or free fatty acids and their toxic by-products leading to lipotoxicosis,30 and combined glucolipotoxicosis.30 Canine beta cells are highly sensitive to loss of function and cell mass associated with glucose toxicosis,31 and therefore resistance to this mechanism is unlikely to be the explanation for the species difference in susceptibility to type 2 diabetes. Lipotoxicosis has not been tested as a cause of beta-cell failure in dogs, has been found not to mediate beta-cell failure in cats,32 and likely only causes beta-cell failure in humans when combined with glucose toxicosis.30 Nonimmune inflammation mediated by cytokines that circulate in increased concentrations in obesity has been proposed as a mechanism leading to beta-cell loss and type 2 diabetes,33 but has not been tested in dogs. Another mechanism postulated to cause loss of beta cells in type 2 diabetes is toxicosis from intracellular formation of amylin oligomers.34 Amylin, also called islet amyloid polypeptide, is a hormone secreted with insulin by beta cells. Increased demand for insulin also increases secretion and misfolding of amylin.35 Misfolded amylin oligomers are highly toxic to beta cells and trigger apoptosis when intracellular; they aggregate extracellularly as histologically visible islet amyloid deposits. Importantly, islet amyloid deposition has not been detected in diabetic dogs.36–38 Canine amylin does not form toxic intracellular oligomers,39 likely because the amino acid sequence of canine amylin does not favor beta-pleated sheet formation, whereas humans and cats are 2 of the few known species in which the amino acid sequence of amylin favors misfolding and oligomer deposition. This species difference in susceptibility to formation of toxic oligomers might be 1 difference predisposing humans and cats, but not dogs, to type 2 diabetes. Other possible factors affecting beta-cell failure associated with obesity involve interaction between the hormonal response to obesity and potential beta-cell toxins. For example, leptin protects beta cells from lipid accumulation and lipotoxicosis.40 It is possible that species differences in obesity-related changes in adipokines like leptin41 or adiponectin,42 to which beta cells have receptors, protect against beta-cell failure in dogs but not humans and cats. Adipokines have not been studied extensively in dogs, and there are no publications on their effects on beta-cell function. It is possible that adipokines like leptin and adiponectin preserve beta-cell function in obese dogs, but not in some obese humans and cats. In summary, the only mechanism of beta-cell failure reported to occur in humans and cats, but not in dogs, is intracellular formation of toxic oligomers of islet amyloid polypeptide. Clearly there is a need for further research to understand the species differences in susceptibility to obesity-induced beta-cell failure. This research could help to identify strategies to help protect feline and human beta cells from failure associated with obesity-induced insulin resistance.

One dog in the present study was glucose intolerant but had intact first-phase insulin secretion and beta-cell glucose sensitivity. That dog's plasma insulin concentration peak occurred in the first 10 minutes after glucose administration and was twice that in lean dogs even though glucagon-stimulated insulin secretion in that dog was similar to that of lean dogs, indicating that first-phase insulin secretion was not selectively impaired in a dog with impaired insulin secretion. In contrast, glucose-intolerant humans43 have first-phase insulin responses that are less than those of lean, insulin-sensitive individuals. The 5-fold increase from baseline in plasma insulin concentration during the first-phase insulin response was similar to that of other obese dogs in the first few minutes of the glucose tolerance test, indicating that beta-cell glucose sensitivity was also intact. Our results suggest that obesity-induced insulin resistance in dogs was not associated with impaired first-phase insulin secretion or loss of beta-cell glucose sensitivity, unlike in humans.7,10,19 This suggests that naturally occurring obesity in dogs might not progress from insulin resistance to an equivalent of the metabolic syndrome and type 2 diabetes mellitus. It follows that dogs with naturally occurring obesity do not have identical features of development of type 2 diabetes mellitus as in humans.

Obese dogs other than the glucose-intolerant dog in the present study retained insulin secretion in response to administration of a nonglucose secretagogue. The insulin response to glucagon was exaggerated in obese dogs, and the difference between lean and obese dogs was more pronounced after glucagon administration than after glucose administration. Glucagon triggers insulin release by a mechanism that bypasses the glucose-sensing apparatus10,19 and acts on a different intracellular pool of insulin than does glucose.39,44 Therefore, glucagon stimulation tests appear to reflect beta-cell secretory capacity, defined as the ability of beta cells to increase plasma insulin concentrations, regardless of beta-cell glucose sensitivity. Our results indicated that glucose-tolerant obese dogs had a 4-fold increase in beta-cell secretory capacity to compensate for a 2-fold decrease in insulin sensitivity, compared with lean dogs.

It is not clear why dogs in the present study had a peak plasma insulin concentration response to glucagon administration that was later than that in dogs in a previous study.21 In the present study, lean and obese dogs had peak insulin concentrations at approximately 20 minutes after glucose administration, whereas healthy lean dogs in a previous study had peak concentrations at 10 minutes and insulin-resistant dogs with hyperadrenocorticism had peak concentrations at 20 minutes.21 In humans, a single sample collected at 6 minutes after glucagon administration is commonly used to assess beta-cell function10,45 and adequately reflects the insulin secretory response to glucagon.46 The disparity between our results and those of a previous study21 in the time of the plasma insulin concentration peak suggests that, in dogs, a single sample time might not adequately reflect the insulin secretory response to glucagon administration, and multiple samples should be collected.

A potential bias in this study was the difference in peak plasma glucose concentration associated with glucose being administered on a per-kilogram basis in lean versus obese dogs. However, the difference in first-phase insulin secretion between lean and obese dogs in the present study was unlikely to be attributable to the difference in peak plasma glucose concentrations. The dose of glucose used for the IV glucose tolerance tests was chosen to induce maximal insulin secretion.47 The only study47 that has tested the effects of different glucose doses on insulin secretion found that first-phase insulin secretion was maximal with a glucose dose of 1,000 mg/kg and a mean peak plasma glucose concentration of 560 mg/dL. Higher glucose doses did not increase the first-phase insulin response. Lean and obese dogs in the present study had peak glucose concentrations > 560 mg/dL, so all would be expected to have maximal first-phase insulin secretion. Therefore, further increases in insulin secretion would not have been induced by higher plasma glucose concentrations in the obese dogs, so the difference in first-phase insulin secretion between lean and obese dogs in this study was unlikely to be attributable to the difference in peak plasma glucose concentrations. Furthermore, obese dogs also secreted more insulin than lean dogs after glucagon administration. The glucagon dose given was independent of body weight, so obese dogs received a proportionately lower dose than did lean dogs. This provided further evidence that the increased insulin secretion in obese dogs was not attributable to the higher plasma glucose concentrations in the obese dogs. It is possible that glucose clearance was affected by the different peak glucose concentrations between obese and lean dogs and by the fact that obese dogs also had higher estimated lean mass. This might have artifactually changed the insulin sensitivity, which is calculated in part on the basis of the glucose clearance kinetics. However, this would not have altered any of the key findings of this study: that obese dogs generally compensated well for insulin resistance, and if they did not, they still maintained beta-cell glucose sensitivity and first-phase insulin secretion.

This study had several limitations. The principal limitation was that there was no laboratory screening prior to study entry for subclinical causes of insulin resistance other than obesity, such as inflammatory diseases (including urinary tract infection) and endocrinopathies (including hyperadrenocorticism, acromegaly, and hypothyroidism), so it cannot be stated with complete certainty that the difference in insulin sensitivity between lean and obese dogs was entirely obesity related. Many of the laboratory abnormalities associated with endocrinopathies associated with obesity (eg, hypertriglyceridemia and hypercholesterolemia) are also present in obese dogs without endocrinopathies. Excluding dogs on the basis of abnormalities that are expected to occur in obese dogs might also have created bias in favor of metabolically normal dogs. Hyperadrenocorticism and hypothyroidism were ruled out in the 1 glucose-intolerant dog with the lowest insulin sensitivity, and acromegaly is extremely rare in dogs other than sexually intact females, but urinary tract infection was overlooked as a potential contributor to insulin resistance. However, none of the conditions that might contribute to insulin resistance would be expected to improve compensatory insulin secretion. The principal findings of this study involved the ability of obese dogs to compensate for insulin resistance and maintain glucose tolerance or, in the case of 1 glucose-intolerant dog, to preserve first-phase insulin secretion and glucose sensitivity. These findings remained valid if the dogs happened to have concurrent conditions other than obesity; in fact, the findings are strengthened if dogs can compensate simultaneously for obesity and other causes of insulin resistance. Another limitation was that the number of dogs used was small, and only 1 dog was glucose intolerant. The results, particularly those that related to glucose-intolerant obese dogs, might be altered by further work in larger numbers of dogs.

Results of the present study indicated that dogs with naturally occurring obesity appeared to preserve beta-cell glucose sensitivity and first-phase insulin secretion even in the presence of long-term obesity, glucose intolerance, and insulin resistance. Insulin secretion during food withholding and first-phase insulin secretion were increased to compensate for obesity-induced insulin resistance. These findings help explain why type 2 diabetes mellitus has not been definitively diagnosed in dogs. In contrast, cats lose first-phase insulin secretion in association with obesity-induced insulin resistance29 and develop a form of diabetes analogous to human type 2 diabetes.2 Differences between species might help clarify why some individuals develop type 2 diabetes while others do not. The mechanism that protects canine beta cells from failure in the face of chronic obesity-induced insulin resistance is unknown. Better understanding of these mechanisms might have important implications for prevention of type 2 diabetes mellitus in susceptible species, including cats and humans.

a.

Internal Jugular Puncture Kit AK-04050, Arrow International Inc, Reading, Pa.

b.

Glucose Intravenous Infusion BP 50%, CSL Ltd, Parkville, VIC, Australia.

c.

BD Insyte I.V 22-gauge, 1-inch catheter, Becton Dickinson Infusion Therapy Systems Inc, Sandy, Utah.

d.

Glucagen Hypokits, Novo-Nordisk, Baulkham Hills, NSW, Australia.

e.

Vacuette EDTA K2, 2 mL, Greiner Bio-one, Tokyo, Japan.

f.

DSL-1600 RIA insulin kit, Diagnostic System Laboratories, Webster, Tex.

g.

Nachreiner R, Department of Small Animal Clinical Sciences, Michigan State University, East Lansing, Mich: Personal communication, 2002.

h.

YSI 2300 Stat Plus, Yellow Springs Instruments, Yellow Springs, Ohio.

i.

Spec cPL, IDEXX Laboratories Inc, Westbrook, Me.

j.

Double Antibody Canine TLI, Siemens Medical Solutions, Los Angeles, Calif.

k.

MINMOD-Millennium, version 6.02, MinMod Inc, Los Angeles, Calif.

l.

Prism 5 for Windows, GraphPad Software Inc, San Diego, Calif.

m.

Stata, version 8.2, StataCorp, College Station, Tex.

References

  • 1.

    Bergman RN, Kim SP, Catalano KJ, et al. Why visceral fat is bad: mechanisms of the metabolic syndrome. Obesity (Silver Spring) 2006; 14(suppl):16S19S.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2.

    Rand JS, Fleeman LM, Farrow HA, et al. Canine and feline diabetes mellitus: nature or nurture? J Nutr 2004; 134(suppl):2072S2080S.

  • 3.

    Polonsky KS. Evolution of beta-cell dysfunction in impaired glucose tolerance and diabetes. Exp Clin Endocrinol Diabetes 1999; 107(suppl):s124s127.

    • Search Google Scholar
    • Export Citation
  • 4.

    Wallace TM, Matthews DR. The assessment of insulin resistance in man. Diabet Med 2002; 19:527534.

  • 5.

    Appleton DJ, Rand JS, Sunvold GD. Insulin sensitivity decreases with obesity, and lean cats with low insulin sensitivity are at greatest risk of glucose intolerance with weight gain. J Feline Med Surg 2001; 3:211228.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6.

    Gerich JE. Is reduced first-phase insulin release the earliest detectable abnormality in individuals destined to develop type 2 diabetes? Diabetes 2002; 51(suppl):S117S122.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Cavaghan MK. The beta cell and first-phase insulin secretion. Medscape Diabetes Endocrinol [serial online] 2004;6. Available at: www.medscape.org/viewarticle/483307. Accessed July 30, 2004.

    • Search Google Scholar
    • Export Citation
  • 8.

    Cherrington AD, Sindelar D, Edgerton D, et al. Physiological consequences of phasic insulin release in the normal animal. Diabetes 2002; 51(suppl 1):S103S108.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Getty L, Hamilton-Wessler M, Ader M, et al. Biphasic insulin secretion during intravenous glucose tolerance test promotes optimal interstitial insulin profile. Diabetes 1998; 47:19411947.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10.

    Gredal C, Rosenfalck AM, Dejgaard A, et al. Impaired first-phase insulin response predicts postprandial blood glucose increment in patients with recently diagnosed type 2 diabetes. Scand J Clin Lab Invest 2007; 67:327336.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11.

    Leahy JL. Natural history of beta-cell dysfunction in NIDDM. Diabetes Care 1990; 13:9921010.

  • 12.

    Toffolo G, Bergman RN, Finegood DT, et al. Quantitative estimation of beta cell sensitivity to glucose in the intact organism: a minimal model of insulin kinetics in the dog. Diabetes 1980; 29:979990.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13.

    Larson BT, Lawler DF, Spitznagel EL Jr, et al. Improved glucose tolerance with lifetime diet restriction favorably affects disease and survival in dogs. J Nutr 2003; 133:28872892.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14.

    Mattheeuws D, Rottiers R, Baeyens D, et al. Glucose tolerance and insulin response in obese dogs. J Am Anim Hosp Assoc 1984; 20:287293.

  • 15.

    van Citters GW, Kabir M, Kim SP, et al. Elevated glucagon-like peptide-1-(7–36)-amide, but not glucose, associated with hyperinsulinemic compensation for fat feeding. J Clin Endocrinol Metab 2002; 87:51915198.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16.

    Bailhache E, Ouguerram K, Gayet C, et al. An insulin-resistant hypertriglyceridaemic normotensive obese dog model: assessment of insulin resistance by the euglycaemic hyperinsulinaemic clamp in combination with the stable isotope technique. J Anim Physiol Anim Nutr (Berl) 2003; 87:8695.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17.

    Kaiyala KJ, Prigeon RL, Kahn SE, et al. Reduced beta-cell function contributes to impaired glucose tolerance in dogs made obese by high-fat feeding. Am J Physiol 1999; 277:E659E667.

    • Search Google Scholar
    • Export Citation
  • 18.

    Eigenmann JE, Eigenmann RY, Rijnberk A, et al. Progesterone-controlled growth hormone overproduction and naturally occurring canine diabetes and acromegaly. Acta Endocrinol (Copenh) 1983; 104:167176.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19.

    Yoneda H, Ikegami H, Yamamoto Y, et al. Analysis of early-phase insulin responses in nonobese subjects with mild glucose intolerance. Diabetes Care 1992; 15:15171521.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20.

    Porte D Jr. Banting lecture 1990. Beta-cells in type II diabetes mellitus. Diabetes 1991; 40:166180.

  • 21.

    Montgomery TM, Nelson RW, Feldman EC, et al. Basal and glucagon-stimulated plasma C-peptide concentrations in healthy dogs, dogs with diabetes mellitus, and dogs with hyperadreno-corticism. J Vet Intern Med 1996; 10:116122.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 22.

    LaFlamme D. Development and validation of a body condition score system for dogs. Canine Pract 1997; 22(4):1015.

  • 23.

    Fleeman L, Rand J. Efficacy of intermittent low-dose heparinised saline flushes for maintaining indwelling peripheral and central intravenous catheters in diabetic dogs. Aust Vet Pract 2001; 31:126134.

    • Search Google Scholar
    • Export Citation
  • 24.

    Marshall RD, Rand JS, Morton JM. Glargine and protamine zinc insulin have a longer duration of action and result in lower mean daily glucose concentrations than lente insulin in healthy cats. J Vet Pharmacol Ther 2008; 31:205212.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25.

    Rowland M, Tozer TN. Assessment of AUC. In: Clinical pharmacokinetics: concepts and applications. Baltimore: Williams & Wilkins, 1995; 469472.

    • Search Google Scholar
    • Export Citation
  • 26.

    Link KR, Rand JS. Reference values for glucose tolerance and glucose tolerance status in cats. J Am Vet Med Assoc 1998; 213:492496.

  • 27.

    Stata longitudinal/panel-data reference manual. 10th ed. College Station, Tex: Stata Press, 2007.

  • 28.

    Bland JM, Altman DG. Transformations, means, and confidence intervals. BMJ 1996; 312:1079.

  • 29.

    Hoenig M, Alexander S, Holson J, et al. Influence of glucose dosage on interpretation of intravenous glucose tolerance tests in lean and obese cats. J Vet Intern Med 2002; 16:529532.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30.

    Poitout V, Robertson RP. Glucolipotoxicity: fuel excess and β-cell dysfunction. Endocr Rev 2008; 29:351366.

  • 31.

    Imamura T, Koffler M, Helderman JH, et al. Severe diabetes induced in subtotally depancreatized dogs by sustained hyperglycemia. Diabetes 1988; 37:600609.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32.

    Zini E, Osto M, Franchini M, et al. Hyperglycaemia but not hyperlipidaemia causes beta cell dysfunction and beta cell loss in the domestic cat. Diabetologia 2009; 52:336346.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33.

    Donath MY, Storling J, Berchtold LA, et al. Cytokines and β-cell biology: from concept to clinical translation. Endocr Rev 2008; 29:334350.

  • 34.

    Haataja L, Gurlo T, Huang CJ, et al. Islet amyloid in type 2 diabetes, and the toxic oligomer hypothesis. Endocr Rev 2008; 29:303316.

  • 35.

    Scheuner D, Kaufman RJ. The unfolded protein response: a pathway that links insulin demand with β-cell failure and diabetes. Endocr Rev 2008; 29:317333.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36.

    Cotton RB, Cornelius LM, Theran P. Diabetes mellitus in the dog: a clinicopathologic study. J Am Vet Med Assoc 1971; 159:863870.

  • 37.

    Dixon J, Sanford J. Pathological features of spontaneous canine diabetes mellitus. J Comp Pathol 1962; 72:153166.

  • 38.

    Watson PJ, Roulois AJA, Scase T, et al. Prevalence and breed distribution of chronic pancreatitis at post-mortem examination in first-opinion dogs. J Small Anim Pract 2007; 48:609618.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 39.

    Gilead S, Gazit E. The role of the 14–20 domain of the islet amyloid polypeptide in amyloid formation. Exp Diabetes Res 2008; 2008:256954.

    • Search Google Scholar
    • Export Citation
  • 40.

    Shimabukuro M, Koyama K, Chen G, et al. Direct antidiabetic effect of leptin through triglyceride depletion of tissues. Proc Natl Acad Sci U S A 1997; 94:46374641.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 41.

    Shimabukuro M, Wang MY, Zhou YT, et al. Protection against lipoapoptosis of beta cells through leptin-dependent maintenance of Bcl-2 expression. Proc Natl Acad Sci U S A 1998; 95:95589561.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 42.

    Kharroubi I, Rasschaert J, Eizirik DL, et al. Expression of adiponectin receptors in pancreatic beta cells. Biochem Biophys Res Commun 2003; 312:11181122.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 43.

    Byrne MM, Sturis J, Sobel RJ, et al. Elevated plasma glucose 2 h postchallenge predicts defects in β-cell function. Am J Physiol Endocrinol Metab 1996; 270:E572E579.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 44.

    Friedenberg R, Metz R, Mako M, et al. Differential plasma insulin response to glucose and glucagon stimulation following ethanol priming. Diabetes 1971; 20:397403.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 45.

    Zangeneh F, Arora PS, Dyck PJ, et al. Effects of duration of type 2 diabetes mellitus on insulin secretion. Endocr Pract 2006; 12:388393.

  • 46.

    Scheen AJ, Castillo MJ, Lefebvre PJ. Assessment of residual insulin secretion in diabetic patients using the intravenous glucagon stimulatory test: methodological aspects and clinical applications. Diabetes Metab 1996; 22:397406.

    • Search Google Scholar
    • Export Citation
  • 47.

    Rottiers R, Mattheeuws D, Kaneko JJ, et al. Glucose uptake and insulin secretory responses to intravenous glucose loads in the dog. Am J Vet Res 1981; 42:155158.

    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 89 0 0
Full Text Views 1377 1076 344
PDF Downloads 358 204 19
Advertisement