Enhanced Energy Expenditure, Glucose Utilization, and Insulin Sensitivity in VAMP8 Null Mice

The control and VAMP8 knockout mice were maintained in a mixed C57BL/6J-129SvJ background (19) for over 11 generations and were housed in a temperature-controlled environment with a 12-h light/12-h dark cycle. Food intake and weights were determined daily, and all studies were performed on 12–14-week-old male mice. All animal protocols were performed in accordance with Albert Einstein College of Medicine of Yeshiva University IACUC (Institutional Animal Care and Use Committee) approval.

Metabolic measurement was performed using an Oxymax indirect calorimetry system (Columbus Instruments, Columbus, OH) (30). Mice were individually housed in the chamber with a 12-h light/12-h dark cycle in an ambient temperature of 22–24°C. VO₂ and VCO₂ rates were determined under Oxymax system settings as follows: air flow, 0.6 l/min; sample flow, 0.5 l/min; settling time, 6 min; and measuring time, 3 min. The system was calibrated against a standard gas mixture to measure O₂ consumed (VO₂, ml/kg/h) and CO₂ generated (VCO₂, ml/kg/h). Metabolic rate (VO₂), respiratory quotient (ratio of VCO₂/VO₂), and activity (counts) were evaluated over a 24-h period. Energy expenditure was calculated as the product of the calorific value of oxygen (3.815 + 1.232 × respiratory quotient) and the volume of O₂ consumed.

Five days before the experiment, the mice were anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg), and two indwelling catheters with prefilled heparin (1,000 units/ml) and polyvidone–povidone (0.5 g/ml; Sigma) were introduced in the right internal jugular vein and left carotid artery. The catheters were externalized through an incision in the skin flap behind the head, and the mice were returned to individual cages after the surgery. The mice were fully recovered from the surgery before the in vivo experiments, as reflected by their reaching preoperative weight.

After a 12-h fast, euglycemic–hyperinsulinemic clamps were conducted in conscious mice as previously described (31,32) with minor modification. The 2-h euglycemic–hyperinsulinemic clamp was conducted with a continuous infusion of human insulin (4 mμ/kg/min) and a variable infusion of 25% glucose to maintain glucose at 90 mg/dl. Insulin-stimulated whole-body glucose metabolism was estimated using continuous infusion of [3-³H]glucose (0.1 μCi/min; PerkinElmer Life Sciences). To determine the rate of basal glucose turnover, [3-³H]glucose (0.05 μCi/min) was infused for 2 h (basal period) with 5 μCi bolus before starting the euglycemic–hyperinsulinemic clamp, and a blood sample was taken at the end of this basal period. To assess insulin-stimulated tissue-specific glucose uptake, 2-deoxy-d-[1-¹⁴C]glucose 2-[¹⁴C]DG (PerkinElmer Life Sciences) was administered as a bolus (10 μCi) 75 min after the start of the clamp. Blood samples were taken at time 0 and every 10 min from 80–120 min through the carotid artery. To estimate basal muscle glucose uptake, 2-[¹⁴C]DG glucose was infused with isotonic saline. During the clamp, plasma glucose was monitored using 2 μl of blood by glucose meter (precision Xtra, Abbot, Bedford, MA). Plasma [3-³H]glucose, 2-[¹⁴C]DG glucose, and ³H₂O concentration were measured as described previously (31). After the euglycemic–hyperinsulinemic clamp, individual tissue samples were collected and immediately frozen at −80°C for the measurement of glucose uptake.

Mice fasted for 12 h were administrated glucose (1g/kg) by oral gavage. Plasma samples were collect at 0, 30, 60, and 120 min for glucose and insulin levels. Glucose was measured using a glucose meter (precision Xtra, Abbot, Bedford, MA). Insulin levels were determined by ELISA (Crystal Chem Inc., Downers Grove, IL).

The stable isotope-labeled intraperitoneal glucose tolerance test (SipGTT) was performed as previously described (33). Briefly, male wild-type and VAMP8 null mice were overnight fasted and were given a 1.0 g/kg intraperitoneal injection of glucose (containing 50% [6,6-²H₂]-glucose). The amount of the M+2 label in the C3–C6 fragment for [6,6-²H₂]glucose was determined by electron impact ionization mass spectrometry, after conversion of the glucose to a glucose aldonitrile pentaacetate derivative (34).

For acute insulin stimulation, the mice were fasted for 12 h and given either isotonic saline or human recombinant insulin (1U/kg) by intraperitoneal injection. Fifteen min later the animals were anesthetized with pentobarbital sodium (50 mg/kg), and tissue samples were collected and subjected to immunoblotting and immunofluorescence.

Tissues isolated from acute insulin stimulation were homogenized in ice-cold lysis buffer (50 mmol/l Tris, pH 7.5, 150 mmol/l NaCl, 1% Triton X-100, 1 mmol/l EDTA, 1 mmol/l phenylmethylsulfonyl fluoride, 0.25% sodium deoxycholate, 1 mmol/l NaF, 1 mmol/l Na₃VO₄, and 2 mmol/l Na₄P₂O₇) containing a protease inhibitor mixture (Roche Diagnostics). The resultant lysates were centrifuged at 16,000g for 60 min at 4°C, and protein concentrations were quantified using the bicinchoninic acid protein assays (Pierce). The protein samples (30 μg) were separated on a 4–12% gradient SDS-PAGE gel and transferred to nitrocellulose membranes using a semidry electroblotter (Owl Separation System, Portsmouth, NH). Membranes were immunoblotted with a VAMP8 polyclonal (18). Other antibodies were purchased from commercial suppliers: GLUT4 polyclonal antibody (East Acres Biologicals Inc.), GLUT1, -3, and -11 (Cell Signaling Technology, Inc.), UCP-1, UCP-2 (Abcam Inc.), and UCP-3 (Sigma). Quantification of all immunoblots was performed using NIH IMAGE software.

Tissues isolated from acute insulin stimulation were embedded in optimal cutting temperature compound. The frozen tissue cross sections (10 μm) were blocked with 3% BSA in PBS for 60 min at room temperature. Primary antibodies were used at the following dilutions: GLUT4 polyclonal antibody (1:100) and α-dystroglycan (α-DG) antibody (1:50) as previously described (35). Fluorescently conjugated secondary antibodies (1:100, Jackson ImmunoResearch Laboratories) were added to the sections for 30 min at room temperature. After extensive washing with PBS, the slides were mounted with Vectashield Mounting Medium (Vector Laboratories). The slides were visualized by confocal fluorescent microscopy (model LSM510; Carl Zeiss MicroImaging, Inc.). Quantification of cell surface immunofluorescence was performed using NIH IMAGE software.

Results are presented as the means ± SEM. Statistical significance was determined using an unpaired two-tailed Student t test, with P < 0.05 considered significant.

To examine the metabolic phenotype of the VAMP8 null mice, we first determine total body weight, total water, and lean and fat mass of male VAMP8 null mice compared with wild-type controls by quantitative nuclear magnetic resonance noninvasive imaging (Fig. 1). The VAMP8 null mice had reduced total body weight with no statistically significant difference in lean mass or total water content compared with control mice. In contrast, there was a marked reduction in total fat mass (Fig. 1,A). This occurred despite normal caloric intake during both the light and dark cycles, and if normalized for body weight the VAMP8 null mice in fact have an approximate 10% increase in food intake (Fig. 1,B). These observed changes in fat mass were confirmed by tissue dissection that confirmed no significant difference in gastrocnemius muscle (Fig. 1,C) or liver (Fig. 1,D) weight but with a marked reduction in epididymal adipose tissue mass (Fig. 1 E). The reduced adiposity resulted from a large reduction in adipocyte cell size as observed by histological imaging (supplementary Fig. 1, available in an online appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db10-0231/DC1).

Because changes in adipocyte cell size have been correlated with changes in the secretion and plasma levels of different cytokines, we next determined the plasma levels of leptin, resistin, retinol binding protein 4 (RBP4), and adiponectin in the fasted state (Fig. 2). Consistent with reduced adipocyte cell size in the VAMP8 null mice, leptin and RBP4 levels were significantly decreased whereas adiponectin levels were somewhat increased. However, there was no significant difference in plasma resistin levels between the VAMP8 null and wild-type control mice.

Because the VAMP8 null mice have reduced adiposity despite equivalent or slightly enhanced daily caloric intake, we hypothesized that these mice would display increased energy expenditure. This speculation would be consistent with increased levels of adiponectin. To determine whole-body energy expenditure, wild-type and VAMP8 null mice were subjected to indirect calorimetry (Fig. 3). Oxygen consumption was increased during both the light and dark cycles, resulting in a total 24-h increase corresponding to a 40% increase in total energy expenditure (Fig. 3,A and B). The increased energy expenditure was also apparent when corrected for lean body mass (supplementary Fig. 5). This increase in energy expenditure also occurred even though the VAMP8 null mice also displayed a decrease in locomotor activity during the dark cycle (Fig. 3 C).

Because increased energy expenditure can result from mitochondrial uncoupling, we examined the expression of UCP-1/2/3 in skeletal muscle and adipose tissue (supplementary Fig. 2, available in an online appendix). There was no significant difference in UCP-1 protein levels between wild-type and VAMP8 null mice in brown adipose tissue, white adipose tissue, or red gastrocnemius skeletal muscle, although we were unable to detect UCP-1 protein in white gastrocnemius skeletal muscle. Similarly, there was no change in UCP-3 levels in any of the tissues examined. In contrast, UCP-2 protein levels were increased in white gastrocnemius skeletal muscle of the VAMP8 null mice with no change in white adipose tissue or red gastrocnemius skeletal muscle.

In any case, the increase in energy expenditure and reduced adiposity resulted in marked enhancement of glucose tolerance along with a decrease in fasting plasma glucose levels (Fig. 4,A). Although insulin levels were somewhat reduced in both the fasting state and during an oral glucose tolerance test, the changes in insulin levels paralleled the changes in plasma glucose, suggesting that total pancreatic β-cell insulin secretion was not significantly affected (Fig. 4,B). Similarly, SipGTT demonstrated a greater clearance of glucose in the VAMP8 null mice compared with controls (Fig. 4,C). Quantification of the area under the curve showed that the VAMP8 null mice display a reduction in glucose excursion, increased glucose disposal with a concomitant reduction in total insulin secretion levels (Fig. 4 D).

To specifically determine whether the VAMP8 null mice display increased insulin sensitivity, we performed conscious nonstressed euglycemic–hyperinsulinemic clamps. As observed during the oral glucose tolerance test, in the basal or fasting level of plasma, glucose was reduced in the VAMP8 null mice compared with wild-type control and was normalized to the same level (∼90 mg/dl) during the euglycemic–hyperinsulinemic clamp (Fig. 5,A). Similarly, plasma insulin was normalized to the same hyperinsulinemic levels during the euglycemic–hyperinsulinemic clamp for both the wild-type and VAMP8 null mice (Fig. 5,B). Consistent with an increase in insulin sensitivity, the glucose infusion rate during the euglycemic–hyperinsulinemic clamp was greater in the VAMP8 null mice along with increased whole-body glucose uptake (Fig. 6,A and B). The overall rate of body glycogen/lipid synthesis was also increased with no statistically significantly change in total body glycolysis rate, although there was a consistent increased trend in the VAMP8 null mice (Fig. 6,C and D). Skeletal muscle accounts for the majority of glucose postprandial glucose uptake and was significantly increased in red type 1 skeletal muscle fibers (soleus and red gastrocnemius) in both the basal and insulin-stimulated states (Fig. 7,A). In parallel, insulin induced a greater suppression of hepatic glucose output in the VAMP8 null mice (Fig. 7 B).

Because the VAMP8 null mice displayed increased insulin-stimulated glucose uptake and fasting hypoglycemia, we speculated that the sarcolemma distribution of GLUT4 might be increased. To examine this possibility, red gastrocnemius muscle from overnight-fasted wild-type and VAMP8 null mice were fixed and subjected to confocal immunofluorescence for the endogenous GLUT4 protein (Fig. 8,A). The muscle fiber from the VAMP8 null mice displayed an increase in the cell surface distribution of GLUT4 as colocalized with the sarcolemma marker protein α-DG. The increased muscle cell surface distribution of GLUT4 in the VAMP8 null mice was even more pronounced in the insulin-stimulated state (Fig. 8,B). Quantification of these data demonstrated a statistically significant increase in the sarcolemma GLUT4 protein of the VAMP8 null mice in the basal and insulin-stimulated states (Fig. 8 C). Similar to that in the gastrocnemius muscle, confocal immunofluorescence of soleus muscle also demonstrated increased basal and insulin-stimulated sarcolemma GLUT4 localizations in the VAMP8 null mice compared with the control wild-type mice (supplementary Fig. 3, available in an online appendix). This change in GLUT4 subcellular distribution occurred with no significant change in the total expression levels of the GLUT4 protein or the total protein levels of several other facilitative glucose transporters GLUT1, GLUT3, and GLUT11 (supplementary Fig. 4, available in an online appendix).

Loss of VAMP8 protein levels results in the impairment of multiple distinct membrane-trafficking processes including the terminal membrane fusion event of cytokinesis, zymogen granule release in pancreatic acinar cells, degranulation of mast cells, transcytosis in polarized epithelial cells, secretory granule release from platelet and cytotoxic T-cells, and the plasma membrane exocytosis of the cystic fibrosis membrane conductance regulator and the aquaporin 2 water channel (21,23,25,26,28,29,36,,,,–41). In contrast, other studies have reported alternative roles for VAMP8 in membrane trafficking. For example, VAMP8 function is necessary for homotypic fusion of early and late endosomes, endosome–lysosome fusion, enhanced vesicle recycling between endocytic vesicles and the endosome in pancreatic β-cells, and GLUT4 endocytosis in adipocytes (15,17,20,21).

To examine the physiologic role of VAMP8 in glucose homeostasis in vivo, we took advantage of the VAMP8 null mice (19). Surprisingly, despite the alterations of different intracellular trafficking events, these mice displayed a marked reduction in adipocyte cell size with increased secretion levels of the adipokine adiponectin that favor enhanced glucose metabolism through increased fatty acid oxidation (42,,–45). More remarkably, the VAMP8 null mice have reduced spontaneous activity with identical caloric intake compared with the wild-type mice, yet rather than gaining adipose tissue mass, the VAMP8 null mice are leaner apparently due to increased total body energy expenditure. The increase in whole-body energy expenditure was not due to changes in adipose tissue or skeletal muscle UCP-1 or UCP-3 expression. Although UCP-3 is the predominant isoform present in skeletal muscle with lower levels of UCP-2 (46,47), we observed a significant increase in UCP-2 protein levels specifically in white gastrocnemius skeletal muscle of the VAMP2 null mice. Thus, it remains possible that the elevated UCP-2 levels account for the increase in energy expenditure, and the reduced size of adipocytes could result from an increased demand of skeletal muscle for fatty acid oxidation. Future studies will be necessary to directly assess this possibility and the mechanism responsible for the tissue-specific UCP-2 induction.

The enhanced metabolic balance of the VAMP8 null mice was also reflected in increased whole-body glucose tolerance and insulin sensitivity. Although there was a small enhancement in insulin suppression of hepatic glucose output, the major effect was a marked increase in skeletal muscle glucose uptake in both the basal and insulin-stimulated states. The increase in glucose uptake in the basal state is also consistent with the relative fasting hypoglycemia in the VAMP8 null compared with wild-type mice. Mechanistically, the loss of VAMP8 resulted in a greater distribution of the skeletal muscle GLUT4 protein at the sarcolemma membrane in the basal state that is consistent with the increase in basal glucose uptake. Similarly, the greater extent of insulin-stimulated GLUT4 translocation to the sarcolemma is also consistent with the increase in insulin-stimulated glucose uptake. The increase in cell surface GLUT4 levels occurred without any change in total GLUT4 protein levels. Thus, the loss of VAMP8 function could result in either an increase in the rate of GLUT4 exocytosis and/or a decrease in sarcolemma GLUT4 endocytosis. As we have previously observed that VAMP8 has no effect on GLUT4 exocytosis but is required for efficient GLUT4 endocytosis in 3T3L1 adipocytes (15), these data provide in vivo support for VAMP8 providing a necessary function in the cell surface endocytosis of the GLUT4 protein.

The glucose metabolic characteristics in the VAMP8 null mice are remarkably analogous to transgenic mice overexpressing the GLUT4 gene both in normal GLUT4 target tissues and specifically in skeletal muscle (48,,–51). In all these models, overexpression of GLUT4 resulted in increased cell surface levels of the GLUT4 protein that resulted in reduced fasting glycemia, increased skeletal muscle glycogen content, and improved glucose tolerance and insulin sensitivity in peripheral tissue. This enhanced glucose handling previously reported for GLUT4 transgenic mice parallels the same reduction in fasting glucose, increased glycogen synthesis, and increased insulin sensitivity and glucose tolerance observed in the VAMP8 null mice. Moreover, the reduction in insulin levels that follows the reduction in plasma glucose was accompanied by compensatory lipolysis that is also consistent with the reduction of adipocyte cell size observed in the VAMP8 null mice

Although all these metabolic data are consistent with VAMP8 playing a role in modulating skeletal muscle cell surface GLUT4 levels, how then do we also account for the observations that a variety of exocytotic secretory events are inhibited in VAMP8 null mice? One simple possibility is that VAMP8 plays different roles in the membrane trafficking of different cell types and/or for different cargo proteins. Alternatively, a more interesting possibility is the presence of a direct coupling between the endocytosis and exocytosis processes. Previously, it was reported that inhibition of chromaffin cell endocytosis by blocking dynamin function with a dynamin antibody resulted in a progressive decrease in catecholamine release (52). Similarly, blockade of dynamin function with a dominant-interfering mutant or reduction in dynamin protein levels by siRNA resulted in relatively normal initiation of first-phase secretion but that was subsequently inhibited along with the more prolonged inhibition of second-phase secretion in cultured β-cells (53). These data suggest the presence of a direct coupling between granule exocytosis (fusion pore opening) and endocytosis (fusion pore closure) events. Thus, another possibility is that the extent of membrane exocytosis is directly coupled to endocytosis in a manner necessary to prevent unrestricted expansion of the plasma membrane surface area. In this case, blockade of VAMP8 function also results in an inhibition of membrane exocytosis via a feedback mechanism arising from the prevention of normal membrane endocytosis and recycling.

In either case, the data presented demonstrate that the loss of VAMP8 expression has a marked effect on the integrative physiology of glucose metabolism and energy expenditure. These combined effects result in a leaner mouse with increased energy expenditure and glucose disposal and insulin sensitivity. Thus, the development of a therapeutic strategy to inhibit VAMP8 function in skeletal muscle may be a viable approach to enhance glucose homeostasis.

This study was supported by National Institutes of Health grants DK-033823, DK-082694, and DK-020541. No potential conflicts of interest relevant to this article were reported.

H.Z. performed experiments, contributed to discussion, and drafted the manuscript. C.-C.W., I.J.K., and W.H. contributed to discussion and reviewed and edited the manuscript. B.V. performed experiments, contributed to discussion, and reviewed and edited the manuscript. J.E.P. contributed to discussion and reviewed and edited the manuscript.