Long-Term Treatment With Dipeptidyl Peptidase IV Inhibitor Improves Hepatic and Peripheral Insulin Sensitivity in the VDF Zucker Rat: A Euglycemic-Hyperinsulinemic Clamp Study

The DP IV inhibitor P32/98 (di-[2S,3S] -2-amino-3-methyl-pentanoic-1,3-thiazolidine fumarate) was synthesized as previously described (24). Glucose for intravenous infusion was obtained from Abbott Laboratories (Montreal, PQ, Canada). All other chemicals were obtained from Sigma Canada (Toronto, ON, Canada) unless otherwise stated.

Eleven pairs of male VDF (fa/fa) littermates were randomly assigned to a control or treatment (P32/98) group at 400 g body weight (10 ± 0.5 weeks of age). Animals were housed on a 12-h light/dark cycle (lights on at 6:00 a.m.) and allowed ad libitum access to standard rat food and water. The treatment group received P32/98 (10 mg/kg) by oral gavage twice a day (8:00 a.m. and 5:00 p.m.) for 12 weeks, whereas the control animals received concurrent doses of 1% cellulose vehicle. The techniques used in this study were in compliance with the guidelines of the Canadian Council on Animal Care and were approved by the University of British Columbia Council on Animal Care (certificate A99-0006).

After 11 weeks of treatment, an OGTT (1 g/kg) was performed after a 16-h fast and complete drug washout (∼11 circulating half-lives for P32/98). No dose of inhibitor was administered on the morning of the OGTT. Blood samples (250 μl) were collected from the tail vein using heparinized capillary tubes (Fisher Scientific, Pittsburgh, PA) and centrifuged, and the plasma was stored at −20°C. Plasma insulin was measured by radioimmunoassay using a guinea pig anti-insulin antibody (GP-01) as previously described (25). Whole blood glucose was measured using a hand-held glucose monitor (Lifescan Canada, Burnaby, BC, Canada).

Estimation of insulin sensitivity from OGTT blood glucose and plasma insulin profiles was performed using the composite insulin sensitivity index (CISI) proposed by Matsuda and DeFronzo (26). Calculation of the index was made according to the following equation:

where FBG is the fasting blood glucose concentration, FPI is the fasting plasma insulin concentration, MG is the mean glucose concentration, and MI is the mean insulin concentration (area under the curve [AUC]/120 min) over the course of the OGTT.

After 12 weeks of treatment, animals were anesthetized with sodium pentobarbital (Somnotol; 36 mg/kg) and a midline incision was made on the ventral aspect of the neck. Chronic cannulas were then inserted into the left carotid artery (PE-50 cannula; Clay Adams) and the right jugular vein (dual PE-10 cannula encased in Silastic tubing; Dow Corning), brought around the neck subcutaneously, and passed through a small skin incision at the base of the neck. After ≥4 days of recovery with at least 2 days of consecutive weight gain, the animals underwent a euglycemic-hyperinsulinemic clamp. Treatment was discontinued the day of catheter implantation and re-initiated 2 days later; the final bolus of inhibitor was given at 5:00 p.m. the day before the clamp. The protocol consisted of three sequential 90-min periods: priming, which involved a bolus (2 μCi) followed by continuous infusion of d-[3-³H]glucose (0.03 μCi/min; Amersham), and two insulin infusion steps (HumulinR; Eli Lilly Canada, Montreal, PQ, Canada) of 5 and 15 mU · kg⁻¹ · min⁻¹, respectively (30 and 90 pmol · kg^{− 1} · min⁻¹). Blood samples (30 μl) were taken every 5 min and centrifuged briefly, and the plasma was analyzed for glucose using the glucose oxidase method (Beckman Glucose Analyzer 2; Beckman Instruments, Palo Alto, CA). During the final 30 min of each period, three 150-μl blood samples were taken at 15-min intervals for determination of plasma tracer and insulin levels. Samples were deproteinated (BaOH₂/ZnSO₄) and evaporated to dryness (to remove [³H]H₂O), and tracer levels were measured using a liquid scintillation counter after resuspension in distilled H₂O.

Samples of epididymal adipose tissue (∼3 cm³) were obtained under anesthesia 3 days after the clamp procedure. After a 16-min collagenase digestion (0.5 mg/ml), recovered adipocytes were washed three times and allowed to stabilize for 1 h in 37°C Krebs buffer repetitively gassed with 95% O₂, 5% CO₂. Aliquots of adipocyte suspension (2 ml) were then prestimulated for 10 min with 0, 0.3, 0.6, 1.5, or 4.75 nmol/l insulin, after which lipolysis was stimulated with a maximally stimulating dose of isoproterenol (10⁻⁷ mol final concentration). The reaction was allowed to proceed for 30 min at 37°C, after which the samples were boiled for 10 min and centrifuged at 4°C for 15 min (12,000g). The aqueous phase of the supernatant was recovered and stored at −70°C. Glycerol determinations were made using a colorimetric glycerol kit (Boehringer Mannheim).

PEPCK activity was measured as follows: liver samples (0.5 g) were homogenized in 1 ml homogenization buffer (10 mmol/l Tris-HCl, 1 mmol/l EDTA, 0.25 mol/l sucrose, and 50 mmol/l KCl, pH 7.2) and centrifuged for 30 min at 10,000g (4°C). After protein determination (BCAprot; Pierce, Rockville, MD), samples containing 100 μg protein were combined with reaction buffer (50 mmol/l Tris-HCl, 2 mmol/l MnCl₂, 2.5 mmol/l phosphoenolpyruvate, 10 mmol/l NaHCO₃, 5 units/ml malate dehydrogenase, and 0.15 mmol/l β-NADH) in a 96-well plate. The reaction was initiated with the addition of 0.4 mmol/l dGDP (final concentration) and followed at 340 nm on a microtiter plate reader. One unit of PEPCK activity corresponds to the conversion of 1 μmol β-NADH to NAD in 1 min. Plasma glucagon levels were measured using a COOH-terminally directed glucagon radioimmunoassay kit (Ab 1032K; Linco Research, St. Charles, MO).

Hepatic glucose output (HGO) and glucose disposal rate (GDR) were calculated according to the method of Steele (27). In summary, HGO was calculated by subtracting the glucose infusion rate (GIR) from the tracer-determined rate of glucose appearance into the plasma compartment. Similarly, at steady state, the GDR is equal to the sum of the rates of endogenous (HGO) and exogenous (GIR) glucose entry into the plasma compartment.

Data (means ± SE) were compared using Prism 3.02 data analysis software (GraphPad Software, La Jolla, CA) within groups using a paired t test and between groups using a Student’s t test (P < 0.05).

In the current study, two groups of 11 VDF rats were treated with the DP IV inhibitor P32/98 for 12 weeks, after which they underwent a euglycemic-hyperglycemic clamp. Fourteen animals (seven per group) were successfully clamped (the remainder failed due to loss of aortic cannula patency). The data presented below are compiled from these 14 animals.

One week before surgical preparation (week 11), an OGTT was performed on all animals. Animals were fasted, and inhibitor dosing was discontinued for 16 h, an interval sufficient to ensure complete drug washout. Plasma DP IV activity measurements made to confirm inhibitor washout showed a 41% increase in DP IV activity in the treated animals versus controls (33.3 ± 0.9 and 23.6 ± 0.6 mU/ml, respectively), a finding consistent with our previous study. Fasting plasma glucose levels in the treated animals were ∼2 mmol/l lower than in control littermates (6.5 ± 0.3 vs. 8.6 ± 0.4 mmol/l), despite comparable fasting insulin levels (1.71 ± 0.23 and 1.69 ± 0.19 nmol/l, respectively) (Fig. 1). The difference in fasting blood glucose increased to nearly 4 mmol/l over the 120-min course of the OGTT, with glucose levels peaking at 10.4 ± 0.4 mmol/l in the treated group compared with 14.2 ± 0.7 mmol/l in the control group. Significant insulin responses were elicited in both groups despite marked hyperinsulinemia; however, the early-phase insulin response in the treated animals measured 260% of that of the controls (Fig. 1). Calculation of AUC revealed no difference in total insulin secretion between the two groups (data not shown). Analysis of these data according to a composite insulin sensitivity index (ISI) revealed mean ISI scores (arbitrary units) of 0.93 ± 0.07 and 1.46 ± 0.13 (P < 0.05) for the control and treated animals, respectively, suggesting enhanced insulin sensitivity after DP IV inhibitor treatment (Fig. 1, inset).

Baseline glucose, insulin, and DP IV activity (treated 35.6 ± 1.7, control 24.4 ± 1.0 mU/ml) values measured at the outset of the clamp were comparable to those observed during the OGTT ∼10 days earlier; the treated group exhibited reduced fasting plasma glucose concomitant with unchanged plasma insulin values (Fig. 2). Fasting plasma glucose levels appeared slightly elevated under clamp conditions, a phenomenon attributable to the difference in measurement technique, and in keeping with our own observations and with the literature on measurement of whole blood versus plasma glucose levels (28). Figure 2 clearly shows that steady-state insulin levels between groups were comparable, rising ∼0.7 nmol/l during the first insulin step and a further ∼2 nmol/l during the second. These values, although hyperinsulinemic, correspond closely to plasma values measured both during the OGTT (Fig. 1) and during normal feeding (21). P32/98-treated animals displayed an immediate requirement for exogenous glucose infusion (GIR) in response to 5 mU · kg⁻¹ · min^{− 1} insulin; under the same conditions, control animals showed a reduced and significantly delayed (∼25 min) response to the same stimulus (Fig. 2). Elevation of the insulin infusion rate to 15 mU · kg⁻¹ · min⁻¹ elicited a further increase in GIR in both groups, with plateau levels reaching 39.2 ± 5.3 and 26.8 ± 4.5 μmol · kg⁻¹ · min⁻¹ in the treated and control groups, respectively.

HGO, calculated from plasma tracer levels, was significantly reduced in the treated animals compared with controls during each stage of the clamp (Fig. 3A and B). Basal HGO averaged 6.5 ± 0.9 μmol · kg⁻¹ · min⁻¹ less in the treated than in the control group (12.7 ± 0.9 μmol · kg⁻¹ · min⁻¹). A difference of similar magnitude was found at each insulin infusion level (Fig. 3A and B). Calculation of GDR revealed little or no response to insulin in the control animals, consistent with previous clamp studies of obese Zucker rats (29). GDR in the treated group, however, showed a return of insulin responsiveness, with steady-state levels 105 and 216% above basal during each of the two insulin steps, respectively (Fig. 4). Basal GDR in the treated animals was reduced relative to controls (Fig. 4).

To further investigate the changes in hepatic insulin sensitivity, liver PEPCK levels and fasting plasma glucagon levels were measured. Measurement of PEPCK activity in liver samples obtained at termination showed no significant difference between the treated and untreated animals, suggesting a non- PEPCK-dependent pathway for the left shift in insulin sensitivity. PEPCK activity averaged 25.9 ± 2.9 and 31.5 ± 4.3 mU/mg tissue in the treated and control groups, respectively (Fig. 3C). Values for the treated animals were comparable to those obtained in control Wistar rats (24.7 ± 0.2 mU/ mg). No difference was detected in fasting plasma glucagon levels either with time or as a result of the treatment (Fig. 3D).

Using glycerol release as an indicator, inhibition of isoproterenol-stimulated lipolysis by insulin was examined in isolated epididymal adipocytes (Fig. 5). Concentration response of inhibition by insulin showed a left shift in adipocytes isolated from treated animals compared with controls; estimated half-maximal concentrations (EC₅₀ values) were 0.29 ± 0.01 and 1.11 ± 0.01 nmol/l, respectively (P < 0.01).

The protection of full-length GIP_1-42 and GLP-1_{7-36 amide} (incretins) in the circulation using DP IV inhibitors represents a significant advancement in the search for new and effective alternative treatments for diabetes. Since the early 1990s, numerous studies have revealed a pleiotropy of antidiabetic effects triggered by interaction of the incretins with their respective cell-surface receptors. Among them are stimulation of β-cell glucose competence, proliferation, differentiation, growth, and cell survival; inhibition of glucagon secretion; and several reports indicating stimulation of glucose uptake in muscle cells (2–7, 30,31). Recently, we showed that long-term DP IV inhibitor therapy caused marked improvements in glucose tolerance, hyperinsulinemia, and β-cell function in the VDF rat, findings that highlight the potential utility of these compounds in diabetes therapy. In the same study, OGTT data and an in vitro determination of insulin-stimulated glucose uptake in isolated muscle provided indirect evidence for a treatment-induced improvement in insulin sensitivity. The present study was carried out to characterize the effects of DP IV inhibitor therapy on insulin sensitivity in the same model, using a euglycemic-hyperinsulinemic clamp. The data collected constitute the first conclusive demonstration that DP IV inhibitor therapy improves both hepatic and peripheral insulin sensitivity.

Extensive literature exists on the obese Zucker rat, including numerous studies examining insulin resistance by means of euglycemic-hyperinsulinemic clamp (32). In one such study, Terrettaz et al. (29) tested HGO and GDR responses in obese Zucker rats over a wide range of insulin concentrations, allowing a comprehensive evaluation of both responsiveness (the efficacy or magnitude of the response to insulin) and sensitivity (the potency or EC₅₀ of the response to insulin) to the hormone. The authors concluded that the obese animals demonstrate marked hepatic insulin resistance (characterized by a fully responsive yet right-shifted HGO and severely impaired glucose disposal (characterized by a total lack of responsiveness over a wide range of insulin concentrations) compared with lean littermates (29). As in these early reports, the control VDF rats in the current study showed a responsive yet right-shifted hepatic glucose response to insulin and an extremely blunted peripheral response (Figs. 3 and 4). It is apparent that long-term treatment with the DP IV inhibitor P32/98 partially reversed both of these functional pathologies, causing a left shift in the HGO response to insulin (Fig. 3) and a partial restoration of peripheral insulin-stimulated glucose uptake (Fig. 4). The primary physiological consequence of these improvements appears to be a reduction in fasting plasma glucose and a marked enhancement of glucose tolerance. Similar changes in HGO and GDR have been shown for a number of classes of oral antidiabetic therapies in obese Zucker rats (33–35).

The finding that P32/98 treatment improves hepatic insulin sensitivity supports our previous work showing reduced hyperinsulinemia in the basal fed state and reduced fasting plasma glucose in the same model (21). Similar reductions in insulinemia and fasting plasma glucose have been shown in a number of studies using long-term GLP-1 (or GLP-1 mimetic) treatment (36–38). In keeping with the previous discussion of HGO sensitivity to insulin, the defining feature of P32/98 treatment on hepatic insulin sensitivity appeared to be a left shift in insulin responsiveness. Because fasting insulin levels did not differ between groups, a reduction in basal HGO in the treated animals is implicit. And because basal (fasting) glucose output from the liver is primarily determined by the rates of gluconeogenesis and glycogenolysis (both of which are potently stimulated by glucagon), a potential underpinning for these results might have been a reduction in glucagon levels in the treated group. Examination of fasting (complete drug washout) glucagon levels, however, although measured using an antibody blind to NH₂-terminal truncation, did not support such a hypothesis (Fig. 3D). Also, PEPCK activity was measured as an index of gluconeogenic enzyme expression, a group of enzymes that exhibit coordinate hormonal expression with respect to insulin and glucagon. Because PEPCK does not undergo short-term regulation via phosphorylation or allosteric effectors and because, physiologically, it is rate limiting for gluconeogenesis, PEPCK activity measurement provides an indicator of gluconeogenic potential. Although a nonsignificant (∼15%) decrease in PEPCK activity was demonstrated in the treated animals, the exact mechanisms responsible for the HGO sensitivity shift remain unclear and warrant further study (Fig. 3C).

Whereas treatment-induced alterations in HGO preferentially affected sensitivity to insulin rather than responsiveness, alterations in GDR appeared to comprise shifts in both responsiveness and sensitivity to insulin. Control animals displayed a basal GDR of 12.6 ± 1.1 μmol · kg⁻¹ · min⁻¹, nearly twofold that of the treated group (7.1 ± 1.2 μmol · kg⁻¹ · min⁻¹). Further, the P32/98-treated group displayed two- and threefold responses to 5 and 15 mU · kg^{− 1} · min⁻¹ insulin, respectively (Fig. 4B), whereas the control animals showed their first sign of peripheral insulin responsiveness (46%) only during the latter infusion step. These data suggest a marked left shift in insulin-stimulated peripheral glucose uptake and are consistent with our previous demonstration of increased glucose uptake in soleus muscle (21) and with several reports of incretin-stimulated increases in glucose uptake (30,39). A number of studies using GLP-1 receptor agonists over the long term have suggested similar improvements in insulin sensitivity (38). It is unclear whether direct acute incretin effects on glucose disposal, as reported by some groups (40,41) and refuted by others (42), play a role.

On that note, although the literature on DP IV inhibition focuses primarily on a mechanism of action involving the enhancement of circulating active GIP and GLP-1, the role of other peptide substrates of DP IV in the improvements evidenced here, and previously, should not be discounted. Natural substrates of DP IV include all tested members of the glucagon/vasoactive intestinal polypeptide (VIP) superfamily of polypeptides (43), including glucagon (44,45), as well as a number of neuroendocrine and immune factors (46). Many of these peptides play significant roles in the regulation of energy metabolism and are likely to contribute toward the improvements resulting from DP IV inhibitor treatment. For instance, recent work by Huypens et al. (47) has highlighted the importance of the counterregulatory hormone glucagon in the maintenance of glucose competence of the β-cell and in proper insulin secretion. Considering that blood glucose is maintained by circulating glucagon levels for approximately two-thirds of the 24-h cycle, DP IV inhibitor− induced enhancement of NH₂-terminally intact glucagon (active; glucagon_1-29) is likely to contribute toward the reported improvements in β-cell function.

In addition to the examination of HGO and GDR in vivo, an in vitro examination of adipocyte insulin sensitivity was performed 3 days after the clamp. The VDF Zucker rat displays excessive fat accumulation and pronounced hyperlipidemia (including elevated free fatty acids [FFAs]), pathologies intimately associated with human type 2 diabetes. It has been suggested that >50% of the insulin resistance in diabetic patients is FFA-induced, with elevated plasma FFAs stimulating insulin secretion, peripheral glucose underutilization, and hepatic glucose overproduction (increased gluconeogenesis) (48, 49). Physiologically, these FFA effects likely serve to preserve glucose stores when supply is limited. However, in times of plenty (e.g., a typical Western diet), these effects become counterproductive, inhibiting the utilization of glucose (48, 49). The demonstration that insulin-induced inhibition of lipolysis is sensitized in adipocytes from DP IV inhibitor-treated animals suggests a potential mechanism of action for the improvements in hepatic and peripheral insulin sensitivity discussed above. A reduction in FFA release from adipocytes, secondary to a sensitized inhibition of lipolysis by insulin, might attenuate the glucose sparing effects of plasma FFAs, thereby reducing the severity of insulin resistance. These findings warrant further investigation into the effects of DP IV inhibition on lipid metabolism.

The success of DP IV inhibitors as a therapeutic strategy in the treatment of diabetes is owed in great part to the pleiotropic nature of its primary effectors, the incretins GIP and GLP-1. Previously believed to be mere enhancers of β-cell function, the incretins are being shown to possess numerous non-insulin-dependent functions, including stimulation of cell survival and modulation of peripheral energy disposal (liver and muscle) (50). The findings of this study corroborate such reports and further exemplify the importance of the non-insulinotropic effects of GIP and GLP-1 in the regulation of glucose homeostasis. In conclusion, the addition of improved hepatic and peripheral insulin sensitivity to the list of beneficial metabolic effects of long-term DP IV inhibitor therapy provides strong support for the use of these compounds in the treatment of diabetes.

This study was provided for by grants from the Canadian Institute for Health Research (CIHR) and the Department of Science and Technology of Sachsen Anhalt (9704/00116 to H.-U.D.). J.A.P. is grateful to the CIHR, Michael Smith Foundation for Health Research, and the Killam Trust Foundation for scholarship support.

Further, we would like to thank Diane Finegood, Brian Topp, Horatio Vinerean, and Christian Lehn-Brand for sharing their expertise and advice on the art of glucose clamping. Last but not least, many thanks are in order to Madeleine Speck and Shalea Piteau for their excellent technical support.