Hyperinsulinemic Compensation for Insulin Resistance Occurs Independent of Elevated Glycemia in Male Dogs

Abstract

Insulin resistance engenders a compensatory increase in plasma insulin. Inadequate compensation is a primary element in the pathogenesis of type 2 diabetes. The signal that heralds developing insulin resistance and initiates hyperinsulinemic compensation is not known. It has often been assumed to be increased glucose. We tested this assumption by determining whether development of fasting and/or glucose-stimulated hyperinsulinemia with diet-induced insulin resistance occurs because of concomitant elevation of glycemia. Male dogs (n = 58) were fed a hypercaloric, fat-supplemented diet for 6 weeks. Dogs underwent magnetic resonance imaging to quantify total and regional (visceral, subcutaneous) adiposity as well as euglycemic hyperinsulinemic clamps. A subset of animals also underwent an insulin-modified intravenous glucose tolerance test to assess insulin sensitivity, acute insulin response (AIRg), and glucose effectiveness. Fat feeding caused modest weight gain, increased visceral and subcutaneous fat, and insulin resistance at both peripheral and hepatic levels. Hyperinsulinemic compensation was observed in fasting levels as well as increased AIRg. However, we observed absolutely no increase in carefully measured fasting, evening (6 to 8 pm) or nocturnal glycemia (2 to 4 am). Insulin resistance and hyperinsulinemia occurred despite no elevation in 24-hour glucose. Compensatory development of hyperinsulinemia during diet-induced insulin resistance occurs without elevated fasting or 24-hour glycemia. These data refute the idea that glucose itself is a requisite signal for β-cell upregulation. Alternative feedback mechanisms need to be identified.

Insulin resistance is a hallmark of type 2 diabetes, with excessive hyperglycemia resulting from inadequate compensatory hyperinsulinemia to counter reduced insulin action. In contrast, nondiabetic individuals who develop chronic resistance are able to mount an adequate physiologic insulin response, by increased β-cell secretion and reduced insulin clearance, such that glucose tolerance is maintained and hyperglycemia is avoided. Failure to mount an adequate hyperinsulinemic response in the face of insulin resistance is a critical step in the pathogenesis of type 2 diabetes. However, while attention has focused on the role of increasing β-cell mass and/or function (1, 2), and to a lesser extent, the contributions of reduced insulin clearance to resistance-associated hyperinsulinemia (3, 4), the identity of the feedback signal(s) generated by the insulin resistant state to which islets and the liver clearance mechanisms normally respond remains unknown (5-7). Clearly, it is important to identify said signals to understand the pathogenesis of diabetes, and to identify targets for treatment.

Development of insulin resistance induced by a high-fat diet elicits a cascade of changes in multiple organ systems which could play a role in compensatory hyperinsulinemia. For a candidate to be considered a signal for resistance-associated hyperinsulinemia, concentration of the variable must be elevated in insulin resistant states, and such elevation must precede the onset of hyperinsulinemia. We previously attempted to identify the feedback signal responsible for insulinemic compensation after diet-induced resistance (8, 9), but the sample size in these studies was relatively small. As such, our prior studies could not absolutely rule out the possible role of plasma glucose per se, the most likely signal candidate for the pancreatic hypersecretion characteristic of insulin resistance. Significantly, no measurable change in glycemia was observed in prior small studies (3, 8, 9). Moreover, β-cell upregulation subsequent to development of insulin resistance has been reported to occur in the absence of hyperglycemia (6, 7). However, Weir and others (10, 11) have championed the hypothesis that small increases in glucose develop during insulin resistance and are responsible for stimulation of compensatory β-cell secretion. If glucose were the primary signal, one should be able to detect elevations of fasting and/or postprandial glycemia before and during the development of insulin resistance. To test the hypothesis that plasma glucose is a critical signal for hyperinsulinemia in the insulin resistant state, we employed a relatively large canine cohort. After characterizing obesity and insulin resistance induced by a high-fat diet, we carefully measured circulating insulin and glucose as resistance developed. The results shown herein fail to support a critical role for glycemia per se as a signal mediating the compensatory hyperinsulinemia required for the physiologic response to insulin resistance. Alternative feedback signals other than glucose per se (such as nocturnal free fatty acids [FFA]) must therefore be important.

Methods

Animals

Procedures were performed on 58 male mongrel dogs (28.8 ± 0.4 kg), individually housed under controlled environmental conditions. Data were pooled from previous published studies (12-14). At least 7 days prior to baseline (pre-fat feeding) testing, dogs were surgically outfitted with a chronic catheter (n = 17) or sampling port (n = 29) in the jugular vein, advanced into the right atrium for sampling of mixed central venous blood. Blood sampling for remaining dogs (n = 12) was drawn from intracatheters inserted percutaneously into a limb vein. All procedures were approved by the University of Southern California Institutional Animal Care and Use Committee.

Diet

Dogs were fed a standard weight-maintaining diet consisting of 3885 kcal/day, derived from 38% carbohydrates, 26% protein, and 35% fat. After 2 to 3 weeks of weight stabilization and baseline metabolic testing (see below), animals were fed a hypercaloric high-fat diet in which the standard diet was supplemented with 6 g/kg of the pre-diet body weight of either cooked bacon grease or lard, representing a hypercaloric diet enriched in saturated fat. This high-fat diet contained 5236 kcal/day, derived from 28% carbohydrates, 19% protein, and 51% fat. Throughout the study, dogs were presented with daily meals at 9 am and were given 1 to 3 hours to consume available ration, after which remaining food (if present) was removed.

Metabolic Assessment

Comprehensive assessment was performed prior to and after a 6-week period of high-fat feeding. During each period, insulin sensitivity was measured both by the hyperinsulinemic euglycemic clamp (EGC) and by minimal model analysis of the intravenous glucose tolerance test (IVGTT). Development of diet-induced obesity was confirmed by magnetic resonance imaging (MRI). Hyperinsulinemic compensation was assessed from fasting insulin as well as the acute insulin response to glucose (AIRg) from the IVGTT. To examine the influence of glucose to trigger the hyperinsulinemic response after fat feeding, glucose was measured in the fasting state and under nonfasting conditions during nighttime blood sampling in a subset of animals (n = 24). All experimental procedures were performed after overnight fasting and all sensitivity tests occurred while dogs were conscious, resting comfortably, and given free access to water.

Euglycemic clamp

Prior to EGCs, 2 intracatheters were inserted percutaneously in each saphenous vein for infusions of glucose and 3-³H-glucose (“tracer”). At t = −180 minutes, a primed infusion of tracer (25 µCi + 0.25 µCi/min; Perkin-Elmer NEN) was initiated and continued for a 90-minute equilibration period. After basal blood sampling, somatostatin (1 μg/min per kg; Bachem) was infused to suppress endogenous insulin secretion. Hyperinsulinemia was induced by intravenous insulin infusion (regular purified pork; Lilly) from t = 0 to 180 minutes at either 0.75 mU/min per kg (n = 41) or 1.15 mU/min per kg (n = 17). Euglycemia was maintained by variable rate 50% dextrose infusion, spiked with 3-³H-glucose (specific activity: 2.2 μCi/g). Blood samples were drawn every 10 minutes from t = 10 to 60 minutes, every 15 minutes from t = 75 to 150 minutes, and every 10 minutes from 160 to 180 minutes, and samples were assayed for glucose, insulin, and tracer.

IVGTT

After basal blood sampling, glucose (0.3 g/kg) was injected at time 0 over 30 seconds. Blood samples were drawn at t = 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, and 19 minutes. At t = 20 minutes, insulin (0.03 U/kg; Lilly) was injected, and followed by blood sampling at t = 22, 23, 24, 25, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 140, 160, and 180 minutes. Samples were assayed in duplicate for glucose and insulin (see below).

MRI

Dogs were pre-anesthetized and sedated. Thirty 1-cm axial abdominal images were obtained using a GE 1.5T Horizon magnet and analyzed with ScionImage software to quantify fat tissue and nonfat tissue in each slice. Total trunk fat volume (cm³) was estimated as the integrated fat across all 30 slices. Visceral fat was defined as fat within the peritoneal cavity at ± 5 image “slices” from the level where the left renal artery branches from the abdominal aorta (ie, from 11-cm abdominal range).

Blood Sampling and Assays

Blood samples were collected in lithium and heparin-coated tubes containing EDTA, centrifuged immediately, and plasma stored at −80 °C. Glucose was assayed in duplicate by the glucose oxidase technique on an automated analyzer (YSI Model 2300), with intra-assay coefficient of variation (CV) < 1%. Insulin was measured in duplicate by enzyme-linked immunosorbent assay (Linco Research; St. Charles, MO; RRID:AB_2800327), with a detection limit of 5 pM, and intra- and inter-assay CV of 2 ± 1% and 5 ± 1%, respectively. In our lab, glucose and insulin assays were highly reproducible and stable over time. Re-assay after 5-year sample storage yielded highly consistent results (r > 0.9, slope ~1.0). Assay of 3-³H-glucose was performed after sample deproteinization (15).

Calculations and Data Analysis

Using data pooled from multiple studies with similar feeding regimens, diet-induced insulin resistance was quantified by EGC and IVGTT. Compensatory hyperinsulinemia was assessed from fasting insulin and the acute insulin response to glucose during the IVGTT.

Insulin sensitivity

Euglycemic clamp.

EGCs were used to quantify whole-body insulin sensitivity (SI_CLAMP), as well as sensitivity of peripheral tissues (SIp_CLAMP) and liver (SIHGO_CLAMP), as previously described (16). Whole-body insulin sensitivity was calculated as:

$${\text{SI}}_{\text{CLAMP}}=\Delta \text{GINF}/(\Delta \text{INS}\times {\text{GLU}}_{\text{ss}})$$

where ΔGINF and ΔINS are the respective increments in glucose infusion rate and insulin measured during exogenous insulin infusion (ie, steady state minus basal), and GLU_ss is the steady state glucose concentration. In a subset of animals (n = 57), rates of glucose uptake (R_d) and hepatic glucose output (HGO) were calculated after data smoothing (17), using Steele’s equations modified for use with labeled glucose infusion (18). Peripheral insulin sensitivity (SIp_CLAMP) was defined as:

$${\text{SIp}}_{\text{CLAMP}}=\Delta {\text{R}}_{\text{d}}/(\Delta \text{INS}\times {\text{GLU}}_{\text{ss}})$$

where ΔR_d is the change in R_d from basal to steady state. Finally, hepatic insulin sensitivity was calculated as:

$${\mathit{\text{SIHGO}}}_{\mathit{\text{CLAMP}}}=\text{|}\Delta \mathit{\text{HGO}}/(\Delta \mathit{\text{INS}}\times {\mathit{\text{GLU}}}_{\mathit{\text{ss}}})\text{|}$$

where ΔHGO is the observed change in HGO (basal minus steady state). Steady state was defined as the final 30 minutes of the clamp procedure.

IVGTT.

A comparable measure of whole-body insulin sensitivity (insulin sensitivity index, S_Imm) was obtained from minimal model analysis of the IVGTT (MINMOD Millennium, ver. 6.02 (19)).

Glucose-stimulated insulin response

Animals were tested before and after fat feeding to assess development of hyperinsulinemia. In addition to fasting hormone levels, glucose-stimulated insulin response was calculated as the acute insulin response to glucose injection (AIR_G) from the IVGTT, defined as the incremental insulin area under the curve from 0 to 10 minutes after glucose injection.

Nighttime glucose sampling

To assess the possibility that feeding animals a diet enriched in fat induces an elevation in nonfasting glucose, we obtained plasma samples during evening and nocturnal periods in animals before and 6 weeks after fat feeding (n = 41). On the day of nighttime sampling, animals were fed from 9 am until either 10 am (n = 8) or noon (n = 33), when any uneaten food was withdrawn. Dogs were provided free access to water. Venous samples were drawn at 30-minute intervals from 6 to 8 pm (“evening” period) and from 2 to 4 am of the following day (“nocturnal” period).

Statistics

Basic statistics (t test and ANOVA, with Tukey’s post hoc analysis when overall significance was detected) were performed using MINITAB statistical software (ver. 13.32; State College, PA; RRID:SCR_014483). Data are reported as mean ± standard error. Statistical significance was set at P ≤ 0.05.

Results

Baseline Body Composition and Fasting Values

Prior to feeding the fat-enriched diet, dogs weighed 28.8 ± 0.4 kg and exhibited a wide range of adiposity, as measured by MRI. Total trunk fat mass varied nearly 5-fold, from the leanest animal (278 cm³) to the most obese (1363 cm³). This variability was explained by both visceral and subcutaneous adiposity, which varied over 4.4-fold and 7.5-fold, respectively, in this cohort of normal healthy animals. Fasting glucose (92.3 ± 1.0 mg/dL) and insulin (10.0 ± 0.7 µU/mL) were normal.

Diet-Induced Obesity and Insulin Resistance

As expected, we observed a consistent profile of modest obesity in dogs that were provided a saturated fat-supplemented diet for 6 weeks (Fig. 1). Animals gained an average of 2.5 ± 0.2 kg (P < 0.0001), reflecting an increase of 8.6% over baseline weight (Fig. 1A). Total adiposity increased 60% ± 5% overall, reflecting moderate increase in visceral fat mass (43% ± 4%; Fig. 1B) and a doubling of fat mass in the subcutaneous depot (100% ± 9% increase; Fig. 1C; P < 0.0001 for all). Increased adiposity was observed in 57 of 58 animals.

Figure 1.

Changes in body composition induced by high-fat diet (n = 58). (A) mean body weight; (B) and (C) individual trajectories of diet-induced changes in visceral and subcutaneous fat, respectively. Each line represents a single animal, tested prior to and after 6 weeks of fat-supplemented diet. ‡ P < 0.0001.

As expected, diet-induced obesity was accompanied by development of insulin resistance, as quantified by euglycemic clamp. Whole-body insulin sensitivity (SIginf) declined 32%, from 36.4 ± 1.9 to 24.6 ± 1.4 dL/min per kg per µU/mL (P < 0.0001), and resistance developed regardless of initial degree of insulin sensitivity (Fig. 2A). Tissue-specific changes in insulin sensitivity were quantified in 53 of 58 animals (5 experiments had technical problems in either baseline or post-fat assessment). Consumption of high-fat diet induced both hepatic (Fig. 2B) and peripheral (Fig. 2C) insulin resistance (P < 0.0001 for each variable). Development of diet-induced insulin resistance was confirmed by minimal model analysis of IVGTTs performed in a subset of animals (n = 36; P < 0.0001; data not shown).

Figure 2.

Development of insulin resistance in fat feeding in normal dogs. (A) changes in whole-body insulin sensitivity in individual dogs (n = 58); (B) and (C) reductions in mean hepatic and peripheral insulin sensitivity (n = 53), respectively. ‡ P < 0.0001.

Compensatory Hyperinsulinemia

Development of insulin resistance is known to engender an elevation in circulating insulin levels in nondiabetic subjects, which is viewed as a response compensatory to the insulin resistance. Indeed, such apparent compensation was observed in our cohort of normal animals (Fig. 3). Fasting insulin increased after dietary intervention (Fig. 3A), with mean plasma levels rising from 10.0 ± 0.7 to 12.8 ± 0.9 µU/ml (Fig. 3B; P < 0.0001), and changes were independent of baseline adiposity. In addition to fasting hyperinsulinemia, diet-induced resistance was accompanied by dynamic hyperinsulinemia, as revealed by the insulin profile during the IVGTT in a subset of animals (n = 36; Fig. 3C). The insulin response to glucose injection (AIRg) increased from 609 ± 38 to 850 ± 52 µU/mL after fat feeding (P < 0.0001). The increase in AIRg was sufficient to compensate for diet-induced insulin resistance, as the disposition index (SI × AIRg) was unchanged (P = 0.24; data not shown).

Figure 3.

Diet-induced compensatory hyperinsulinemia in fasting state (A, B; n = 58) and acute response to glucose injection (C, D; n = 36). (A) Individual and (B) mean changes in fasting insulin (n = 58); (C) Insulin dynamics during insulin-stimulated IVGTT; (D) Mean changes in AIRg. ‡ P < 0.0001.

Effects of Fat Feeding on Plasma Glucose

We examined the hypothesis that the hyperinsulinemia evident after feeding a fat-supplemented diet was due to elevation of plasma glucose (Fig. 4). Despite very careful and repeated assays of glucose, we observed absolutely no trend toward higher glycemia after 6 weeks of high-fat diet (Fig. 4A). There was no influence of baseline adiposity on diet-induced changes in glycemia. Mean fasting glucose was absolutely unchanged between pre-fat and post-fat periods (92.3 ± 1.0 vs 92.2 ± 1.0 mg/dL, respectively; P = 0.86; Fig. 4B).

Figure 4.

Effect of high-fat diet on fasting glycemia (n = 58). (A) Changes in individual animals; (B) Mean fasting glucose during each testing period.

In addition to monitoring fasting glucose, we investigated whether nonfasting glucose—specifically glucose levels obtained in the evening and nighttime periods—was increased after feeding animals a fat-supplemented diet (Fig. 5). Blood was sampled in a subset of conscious animals (n = 41) over two 2-hour periods. We observed no significant increase in plasma glucose after fat feeding in either the evening period (6 to 8 pm; 91.2 ± 1.5 vs 93.7 ± 1.0 mg/dL; Fig. 5A; P = 0.08) or nocturnal period (2 to 4 am; 92.1 ± 1.2 vs 92.7 ± 1.0 mg/dL; Fig. 5B; P = 0.58). The absence of elevated glucose in diet-induced obese animals that display insulin resistance and hyperinsulinemia, rejects the hypothesis that fasting or 24-hour glucose per se could have been the signal for hyperinsulinemic upregulation under conditions of fat feeding–induced insulin resistance.

Figure 5.

Absence of changes in nonfasting glucose after fat feeding (n = 41). (A) glucose levels during evening (6-8 pm) period; (B) plasma glucose during nocturnal (2-4 am) phase.

Discussion

Chronic insulin resistance is a known risk factor for type 2 diabetes, yet the majority of resistant individuals do not present with overt disease. In the absence of full-blown diabetes, development of resistance triggers a physiologic compensatory feedback mechanism, involving both insulin secretory and hepatic clearance machinery (3, 20, 21), that results in hyperinsulinemia sufficient to counter the impairment of insulin action. When this response is inadequate, risk for diabetes increases substantially. Thus, it is of critical importance to reveal the signal(s) generated during development of insulin resistance that initiate the compensatory response by the pancreatic β-cells and the liver. It has long been considered, in fact assumed, that glucose itself is the signal to enhance β-cell function. However, the latter hypothesis has never been adequately tested. Therefore, we carefully measured circulating glucose levels as a primary candidate for the feedback signal.

To address the importance of glucose as a signal, we examined data generated by our laboratory using the canine model of obesity (12-14). We reasoned that use of dietary intervention, specifically supplementing daily food allotment with 6 g/kg per day of saturated fat, would allow us to monitor development of insulin resistance, characterize diet-induced weight gain and obesity, and quantify both fasting and dynamic changes in insulin and circulating signal candidates, including glucose.

Our results demonstrate our ability to generate a predictable, consistent model of diet-induced insulin resistance and obesity that reflects the common forms evident in Western society. Prior to dietary intervention, dogs exhibited a wide range of baseline adiposity, as measured by MRI, with variations in both visceral and subcutaneous fat depots. Not surprisingly, clamp- and IVGTT-based measures of insulin sensitivity varied considerably among animals, with similar range and negative correlation with adiposity. Dietary treatment–induced insulin resistance elicited a compensatory increase in both fasting and glucose-stimulated insulin release (Fig. 3). Thus, the fat-fed dog model was well-suited to identify possible bloodborne signals for this hyperinsulinemic response.

The primacy of glucose as a signal for upregulation, eg, increased insulin secretion and β-cell mass, has been championed by Weir and his colleagues (10). Recent evidence to support this contention was reported by Terauchi et al (11), who reported that mice with β-cell specific knockout of glucokinase, an enzyme considered a “glucose sensor” for glucose-stimulated insulin release, showed reduced β-cell hyperplasia in response to a high-fat diet. In our studies, if glucose were the primary signal for hyperinsulinemic upregulation, we anticipated that insulin resistance would elicit a measurable increase in circulating glucose, which would be expected to stimulate β-cells (and perhaps decrease clearance) whereby elevated insulin would lower glycemia to pre-resistance levels. The system would thus be “re-regulated” such that insulin resistance would be precisely compensated by hyperinsulinemia. However, despite insulin resistance and hyperinsulinemia, absolutely no increase in fasting glucose was detected in our large cohort of animals (Fig. 4). Fasting glucose was unchanged by fat feeding (pre: 92.3 ± 1.0; post: 92.2 ± 1.0 mg/dL). Although there was variability in the diet-induced glucose changes across animals, there was no relationship between the change of glucose and the hyperinsulinemic response to diet-induced obesity (P = 0.57). In addition, there was no relationship between the diet-induced change in fasting glucose and either nonfasting glucose at the 6 to 8 pm (P = 0.77) or 2 to 4 am periods (P = 0.20). Similar findings have been reported by groups studying the metabolic changes of pregnancy. In nondiabetic women, pregnancy is characterized by progressive development of insulin resistance, maximal by the third trimester (22), which is accompanied by compensatory hyperinsulinemia (23). This feedback occurs in the absence of elevated fasting glucose—in fact, fasting glucose levels actually decline during pregnancy. These data deny the role of fasting glucose as a signal for hyperinsulinemic compensation in pregnancy-associated insulin resistance, and other signals have been proposed (5, 24, 25).

We went on to consider the possibility that while fasting glucose may not be a primary signal, perhaps elevations in 24-hour glucose concentrations would develop in animals fed a fat-enriched diet, which would suggest that nonfasting glucose may play a role in upregulation. In a majority of our animals fed the high-fat diet (41 of 58 animals), we measured glucose from blood samples drawn during evening (6-8 pm) and nocturnal (2-4 am) phases. In fact, nonfasting glucose levels from either period were also not measurably affected by fat feeding (Fig. 5). While it is possible that other dietary interventions could have elicited both hyperinsulinemia and elevated glucose, it remains true that hyperinsulinemia occurs in the absence of elevated glucose during insulin resistance induced by a wide variety of conditions and interventions such as pregnancy, saturated fat diet, genetic manipulation, or sleep restriction (5-9). Taken together, available data clearly demonstrate that glucose per se cannot be not the signal by which diet-induced insulin resistance is “sensed” by tissues responsible for the physiologic hyperinsulinemic response.

In the face of data denying a role for glucose as the signal for insulin upregulation, we considered alternative candidates. Evidence from rodent studies suggests that islet cell mitogenic factors may be a signal for β-cell hyperplasia in mouse models of insulin resistance (6). We observed elevated leptin levels after fat feeding, an expected result of diet-induced increases in fat mass. After 6 weeks of fat feeding, leptin levels were elevated throughout the 24-hour sampling period, with no greater increase observed in either the 6 to 8 pm or 2 to 4 am time intervals (data not shown). However, leptin is generally considered to have an inhibitory effect on insulin secretion (26), thus making it an unlikely mediator of compensatory hyperinsulinemia in the fat-fed dog. The extent to which leptin resistance exists or contributes to the hyperinsulinemia is unknown and beyond the scope of the current study. Previous studies in our laboratory of diet-induced obesity and insulin resistance (with attendant compensatory hyperinsulinemia) failed to detect changes in fasting levels of key metabolites (glucose, FFA). No change in 24-hour profiles of other likely candidates (growth hormone, cortisol, glucagon-like peptide-1) were observed in animals fed a fat-enriched diet. In sharp contrast, however, we observed that FFA are markedly increased in the middle of the night (2-4 am), and nocturnal FFA were further elevated after dogs were made insulin resistant by high-fat diet (9). These previous data are consistent with our current results, in which a subset of animals that underwent nighttime sampling exhibited elevated nocturnal (2-4 am) FFA levels after fat feeding (0.68 ± 0.07 vs 0.57 ± 0.06 mM; P = 0.04), but no significant change in either fasting or evening (6-8 pm) concentrations. Further confirmatory results have been reported by Broussard et al (27) in healthy human subjects who develop resistance after sleep restriction. To further test whether elevated nocturnal FFA may be a possible signal for β-cell upregulation for insulin resistance, studies were previously performed in which hyperinsulinemic compensation was measured after blocking the nocturnal rise in FFA by an adenosine agonist (8). When nocturnal elevation of FFA was blocked, animals did not exhibit the expected hyperinsulinemic upregulation induced by high-fat diet. Taken together, these previous data support a role of nocturnal FFA as a signal for β-cell upregulation for diet-induced insulin resistance. Further studies are necessary to quantify the contribution of reduced insulin clearance to the hyperinsulinemic compensation for insulin resistance, and the possible role of nocturnal FFA in the observed changes in clearance.

While we failed to detect any increase in fasting or nonfasting glucose in dogs tested prior to or after 6 weeks of receiving a high-fat diet, it is possible that small increments in glycemia may have been present prior to the 6-week endpoint period. This would manifest as development of insulin resistance within several weeks of consuming a high-fat diet, leading to glucose levels increasing transiently to stimulate β-cell secretion and/or reduce insulin clearance, then returning to pre-insulin-resistant levels. Under such a scenario, one would assume elevated glucose would act as a “pulse signal,” where the physiologic response (hyperinsulinemia) would be sustained for weeks despite the disappearance of the earlier increase of the signal (ie, memory effect). While such data were not collected in the present animal cohort, prior studies by our group which employed identical dietary intervention revealed no change in glucose measured at 2-week intervals of fat feeding (pre-fat: 93.8 ± 1.9; week 2: 94.3 ± 1.3; week 4: 93.0 ± 1.4; week 6: 94.4 ± 1.3 mg/dL; (28)). These data are consistent with earlier findings from our group that no upward trend of fasting glycemia occurs during 6-week high-fat diet ((3, 9)). These data do not support even a transient role of glucose to mediate hyperinsulinemia during fat feeding. It should also be noted, however, that when insulin resistance develops after short-term (4-day) sleep restriction in healthy human subjects, the hyperinsulinemic feedback occurs without an increase in fasting glucose (27). Moreover, nocturnal elevations of FFA are increased under these conditions, supporting the role of FFA as a signal for hyperinsulinemic upregulation. This latter hypothesis warrants further testing.

Conclusion

In summary, we demonstrated that fat feeding in healthy canines induces a reproducible model of hepatic and peripheral insulin resistance as well as obesity in both subcutaneous and visceral depots. This model shows characteristic hyperinsulinemic compensation, both of fasting hormone levels and dynamic response to injected glucose. Under such conditions, we failed to detect any significant elevation of either fasting or nonfasting glucose, strongly demonstrating that glucose is not a primary signal for hyperinsulinemic compensation during insulin resistance. We posit that other signals, perhaps nocturnal FFA or a signal yet unidentified, play a more dominant role in the initiation of the physiologic compensation that is so vital for maintaining glucose homeostasis in the face of developing resistance. It is of great interest to search for important signals, whether humoral or neural, which may account for hyperinsulinemic upregulation.

Glossary

Abbreviations

AIRg

acute insulin response

CV

coefficient of variation

EGC

euglycemic clamp

FFA

free fatty acids

HGO

hepatic glucose output

IVGTT

intravenous glucose tolerance test

MRI

magnetic resonance imaging

R_d

glucose uptake

Additional Information

Disclosures: The authors declare no competing interest.

Data Availability

Some or all data generated or analyzed during this study are included in this published article or in the data repositories listed in References.