1 Departments of Nutrition
4 Surgery, University of Tennessee, Knoxville, Tennessee 37996
2 Zen-Bio, Inc., Research Triangle Park, North Carolina 27709
3 Pennington Biomedical Research Center, Baton Rouge, Louisiana 70808
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
calcium; obesity; pancreas
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The mechanism of agouti modulation of coat color development is largely understood. This regulation occurs via competitive receptor antagonism of the -melanocyte-stimulating hormone (
-MSH) at its receptor (melanocortin 1 receptor; MC1-R), resulting in an inhibition of cAMP production and a consequent switch from eumelanin (black pigment) to phaeomelanin (yellow pigment) production (19, 31). This competitive antagonism of melanocortin binding has served as a paradigm for the mechanism of the action of agouti in obesity, and it has been proposed that competitive antagonism of another melanocortin receptor may be responsible for the metabolic effects of this gene (18). However, the actual mechanism of agouti action in obesity and insulin resistance has not yet been elucidated. Indeed, we have recently demonstrated that agouti also regulates cellular functions, such as cell Ca2+ signaling, independently of melanocortin receptor antagonism (23, 50).
Intracellular Ca2+ concentration ([Ca2+]i) appears to play an important role in the metabolic derangements associated with obesity and insulin resistance (6, 10, 11). We have found that recombinant agouti protein causes a dose-dependent increase in [Ca2+]i in a variety of cells (23, 51), including both murine and human adipocytes. Agouti protein also stimulates both the expression and activity of fatty acid synthase (FAS), a key enzyme in de novo lipogenesis, and thereby increases triglyceride accumulation in a Ca2+-dependent manner (20). Moreover, this stimulation can be mimicked by stimulation of Ca2+ influx with KCl (52) and blocked by Ca2+ channel antagonism (22). Accordingly, because the human homologue of agouti is expressed in adipose tissue (28), the human agouti protein may similarly exert paracrine effects on adipocyte [Ca2+]i and thereby stimulate de novo lipogenesis and promote obesity.
Alternatively, it has been proposed that central effects of agouti, mediated by melanocortin receptor antagonism, result in hyperphagia and obesity. Huszar et al. (18) recently reported that targeted disruption of the murine melanocortin-4 receptor (MC4-R), which is expressed in brain, resulted in a syndrome of hyperphagia and hyperinsulinemia similar to that found in dominant agouti mutations. This suggested that chronic antagonism of the central nervous system MC4-R may be a primary cause of agouti-induced obesity (18). However, although MC4-R expression is considered to be primarily limited to most regions of the brain, MC4-R mRNA is also expressed in muscle and adipose tissue (7, 35). Consequently, the contribution of peripheral MC4-R antagonism to this hyperphagia-hyperinsulinemia syndrome is not clear.
We have recently reported that transgenic mice expressing agouti in adipose tissue under the control of the aP2 promoter become obese if supplemental insulin is concomitantly provided, whereas supplemental insulin does not produce this effect in their nontransgenic littermates (36). Because hyperplasia of pancreatic ß-cells precedes the development of obesity in agouti mutant mice (43), it is possible that hyperinsulinemia is a direct effect of agouti action on the pancreas and that this hyperinsulinemia, combined with adipocyte agouti expression, may be responsible for the obesity syndrome found in agouti mutants. In support of this concept, we have noted that the effects of insulin and agouti on FAS expression are additive and that the FAS promoter has an agouti response region distinct from its insulin-response element (9).
Accordingly, because we have previously shown agouti to regulate [Ca2+]i in several cell types (23), and because increasing [Ca2+]i is the proximate signal for insulin release (37), we have now evaluated the role of agouti in regulating pancreatic [Ca2+]i and, consequently, insulin release. We report here that agouti is expressed and stimulates Ca2+ signaling in human pancreatic islets, as well as in ß-cell lines, thereby serving as a potent insulin secretagogue. This suggests a potential role for agouti in the development of hyperinsulinemia in humans.
![]() |
MATERIALS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Cell culture.
RIN-5F pancreatic ß-cells (ATCC, Manassas, VA) and primary human pancreatic islets were cultured in RPMI-1640 medium supplemented with 5% FBS (vol/vol) and antibiotics (50 U penicillin/ml and 5 µg streptomycin/ml). HIT-T15 pancreatic ß-cells (ATCC) were cultured in F-12K nutrient mixture supplemented with 10% horse serum (vol/vol), 2.5% FBS (vol/vol), and antibiotics. Cells were maintained at 37°C in an atmosphere containing 5% CO2 and 90% humidity. Cell culture reagents (horse serum, trypsin-EDTA, antibiotics, Hank's balanced salt solution, DMEM, RPMI-1640, F-12K nutrient mixture) were obtained from Life Technologies (Grand Island, NY). FBS was obtained from Hyclone (Logan, UT).
Production and purification of recombinant agouti polypeptide.
A 614-bp Xba I-Pst I fragment of the full-length mouse agouti cDNA (5) or human agouti cDNA (28) was subcloned into a baculovirus expression vector and expressed in Trichiphisia ni cells. Medium was collected 48 h after infection and purified as previously described (45). The purified agouti protein was stored in PBS at -80°C before use. Previous data from our laboratory demonstrated that recombinant agouti protein dose dependently increased [Ca2+]i in several cell lines, including adipocytes, with an EC50 ranging from 18 to 62 nM, depending on cell type (23). Accordingly, a plateau dose of 100 nM agouti was used in the present study.
[Ca2+]i determination.
[Ca2+]i values in RIN-5F, HIT-T15 pancreatic ß-cells, and primary human pancreatic islets were determined fluorometrically as previously described (1, 24). Briefly, cells were preincubated in serum-free medium for 2 h, washed with Hank's balanced salt solution, and trypsinized. Cells were collected by centrifugation and washed with HEPES-buffered salt solution [HBSS, containing (in mM) 138 NaCl, 1.8 CaCl2, 0.8 MgSO4, 0.9 NaH2PO4, 4 NaHCO3, 0 or 5 glucose (referred to as glucose-free HBSS or HBSS, respectively), 6 glutamine, 20 HEPES, and 1% BSA]. Cells were then loaded with fura 2-AM (10 µM) in the same solution for 20 min at 37°C in the dark with continuous shaking, and [Ca2+]i was measured using dual-excitation (340 and 380 nm) and single-emission (510 nm) fluorometry. After the establishment of stable baseline (at 100 s), the response to agouti (0100 nM) was determined. Digitonin (25 µM) and Tris-EGTA (both 100 mM, pH 8.7) were added at 500 and 550 s to determine maximal and minimal fluorescent ratios, respectively, and [Ca2+]i was then calculated by the equation of Grynkiewicz et al. (16).
Ca2+ influx measurement.
Ca2+ influx was determined using the Mn2+-quench technique. Mn2+ enters cells via Ca2+ influx pathways and, once inside cells, quenches Ca2+-fura 2 fluorescence (8, 13, 34). When measured at 360 nm, i.e., the isosbestic wavelength of fura 2, the fall in fluorescence is indicative of Mn2+ influx irrespective of changes in [Ca2+]i (8, 34). Consequently, when excess Mn2+ is added to the medium, this Mn2+-quench technique serves as a surrogate measurement of Ca2+ influx.
To measure Mn2+ influx in cells, HIT-T15 ß-cells or human pancreatic islets were loaded with fura 2-AM as described above for [Ca2+]i determination. Fura 2-loaded cells were then washed with HBSS, and Mn2+ quench was measured fluorometrically by excitation at 360 nm and emission at 510 nm. After a stable baseline was established, MnCl2 (0.2 mM for HIT-T15 cells and 1 mM for human pancreatic islets) was added at 50 s. After the establishment of a new baseline (at 100 s), Mn2+ quench in response to agouti (0100 nM) was determined. To study the effect of agouti on KCl-induced Ca2+ influx, the Mn2+ tracing was continued after the addition of agouti until 400 s, and then KCl (550 mM) was added. Digitonin (50 µM) was added at the end of the experiment (at 500550 s), allowing rapid Mn2+ influx and quench of the fluorescence to basal level. The difference between this value and the initial value before the cells were exposed to Mn2+ was taken as 100 arbitrary fluorescence units (8, 21). Mn2+-quench rate was estimated by linear regression of data collected after addition of each agonist or vehicle control (33).
Insulin release.
Insulin release measurements were conducted by a modification of Shimizu et al. (40). Briefly, RIN-5F and HIT-T15 cells were seeded in six-well plates until confluent. Primary human pancreatic islets were maintained in six-well plates for 48 h after isolation. On the day of each experiment, cells were placed in glucose-free, serum-free medium for 2 h and then incubated with either 100 nM agouti or the indicated test agents in the absence or presence of glucose for 2 h. At the end of the experiment, the medium was removed and stored at -80°C. Cells were sonicated in homogenization buffer containing (in mM) 250 sucrose, 1 dithiothreitol, 0.1 phenylmethylsulfonyl fluoride, and 1 EDTA and stored at -80°C for subsequent total protein correction. Immunoreactive insulin levels in the medium were measured by insulin RIA kits (Linco, St. Charles, MO) for rat or human, as appropriate, and total protein was determined by a modified Bradford method using Coomassie blue dye (Pierce, Rockford, IL).
RT-PCR and Southern blot analysis.
Total RNA from human pancreas was isolated using CsCl2 density centrifugation. mRNA from human pancreas were isolated according to manufacturer instructions [Micro poly(A) pure kit, Ambion, Austin, TX]. RT-PCR was conducted essentially as described by Kwon et al. (28). Briefly, 500 ng of human pancreas mRNA were reverse-transcribed to first-strand cDNA using random hexamer and RT (Perkin Elmer, Norwalk, CT) and amplified by PCR (Perkin Elmer, Norwalk, CT). The conditions of PCR were 1 cycle (94°C, 5 min), 35 cycles (94°C, 30 s; 65°C, 1 min; 72°C, 1 min), and 1 cycle (72°C, 7 min) with 0.5 µM 5' primer (5'-ATGGATGTCACCCGCTTACTCCTGGCC-3') and 3' primer (5'-GCGCTCAGCAGTTGAGGCTGAGCACGC-3'), which corresponds to the 5' or 3' ends of the putative open reading frame of the agouti gene (28). The amplified PCR products were then visualized by 1.2% agarose gel electrophoresis, transferred to nylon membrane, and hybridized with human agouti cDNA as a probe at 68°C overnight. The membrane was washed with 2x standard saline citrate (SSC)-0.1% SDS at room temperature for 30 min and then washed with 0.1x SSC-0.1% SDS at 65°C for 1 h. The autoradiograph was exposed at -80°C for 30 min.
Statistics.
All data are expressed as means ± SE. Statistical analysis of the data was performed by Student t-test, paired Student t-test, or single-factor analysis of variance (ANOVA) where appropriate.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
To investigate whether agouti modulation of Ca2+ is due to stimulation of Ca2+ influx, Ca2+ influx was monitored by the Mn2+-quench method in HIT-T15 cells. Agouti (100 nM) induced a twofold increase in Mn2+-quench rate in these cells (Fig. 3A, P < 0.001). This indicates Ca2+ influx via Mn2+-permeable Ca2+ channels into HIT-T15 cells in response to 100 nM agouti.
|
The consequence of agouti potentiation of the effect of KCl on Ca2+ influx, as indicated by Mn2+ quench, was studied by measuring insulin release from HIT cells under the same conditions. As shown in Fig. 3C, 10 mM KCl alone elicited no increase in insulin release from HIT cells compared with 5 mM KCl. However, the addition of 100 nM agouti into either 5 or 10 mM KCl resulted in a 43 and 51% increase in insulin release, respectively (P < 0.05). KCl stimulated insulin release dose dependently from 15 to 50 mM. Agouti potentiated this KCl effect on insulin release in these cells, with more apparent effects at higher KCl concentrations (25 and 50 mM, Fig. 3C).
To evaluate the relevance of these data to humans, we determined the effect of agouti on both Ca2+ influx and insulin release in primary cultured human pancreatic islets. Ca2+ influx into pancreatic islets was measured using the Mn2+-quench method. Agouti (100 nM) induced a marked Mn2+ influx in primary cultured human pancreatic islets (Fig. 4A), indicating Ca2+ influx into cells through Mn2+-permeable Ca2+ channels, similar to that found in HIT-T15 cells. Agouti treatment of human pancreatic islets resulted in a fivefold increase in Mn2+-quench rate compared with vehicle control (Fig. 4B, P < 0.001). In addition, agouti treatment also potentiated the effect of KCl on Ca2+ influx in human pancreatic islets, as 25 mM KCl stimulated a more profound Mn2+ influx into agouti-treated islets than nontreated islets (Fig. 4A). This resulted in an increase in [Ca2+]i (58 ± 11 nM over baseline) that was blocked by nitrendipine. The effect of agouti in stimulating insulin release was also studied in human pancreatic islets. As shown in Fig. 5A, agouti (100 nM) and KCl (10 mM) induced a moderate increase in insulin release in the absence of glucose (P < 0.05). Under these conditions, agouti slightly potentiated KCl-induced insulin release, similar to that found in HIT-T15 cells. In the presence of 10 mM glucose, agouti stimulated a similar, albeit smaller, increase in insulin release (P < 0.05). As expected, KCl-induced insulin release was greatly potentiated by glucose (P < 0.05 vs. control). This effect was further significantly potentiated by agouti (P < 0.05 vs. control and KCl). Figure 5B demonstrates that agouti increases insulin release from human pancreatic islets over a broad range of glucose concentration (030 nM). This effect was blocked by 30 µM nitrendipine (Fig. 5C).
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In addition, insulin resistance is a marked feature in both type II diabetes and obesity and is a characteristic of agouti mutant mice (26, 49). Although this resistance may result from either insulin receptor or postreceptor defects, we have not found recombinant agouti protein to exert any effect on insulin signaling in skeletal muscle cells (unpublished data). However, it is well documented that hyperinsulinemia may directly downregulate the insulin receptor (27), thereby leading to insulin resistance. This suggests that agouti-mediated hyperinsulinemia may also contribute to the insulin resistance in this animal model.
Insulin release can be stimulated by both nutrient and nonnutrient secretagogues, and [Ca2+]i plays an important role in the signaling pathway for insulin release (30, 37). The primary physiological insulin secretagogue is glucose. Normally, glucose oxidation increases the ATP-to-ADP ratio and inhibits ATP-sensitive K+ (K+ATP) channels in ß-cells. Closing of K+ channels leads to depolarization of ß-cells and secondary activation of voltage-dependent Ca2+ channels and subsequent Ca2+ influx into pancreatic ß-cells, triggering insulin release (37). In addition, growing evidence indicates a K+ATP channel-independent mechanism in glucose-stimulated insulin release (14, 15). Gembal et al. (14, 15) reported that opening K+ATP channels by diazoxide blocked the effect of glucose on mouse pancreatic ß-cell insulin release but did not prevent glucose potentiation of 30 mM KCl-induced insulin release. This effect was not accompanied by changes in ß-cell membrane potential, 45Ca flux, or [Ca2+]i (14, 15). However, it did require an initial elevation of [Ca2+]i, although further changes in [Ca2+]i were not needed. This effect was not ascribed to phosphoinositide metabolism or activation of protein kinase A or C. Instead, it required glucose metabolism and was correlated with ß-cell ATP-to-ADP ratio (15). This suggests that, in addition to its primary signal in stimulating insulin release (i.e., closure of K+ATP channels-depolarization-Ca2+ influx), glucose can also enhance insulin release by modulating and amplifying the secretory response to an increased [Ca2+]i signal, whereas it did not further affect changes in membrane potential and [Ca2+]i. This mechanism appears to involve changes in ß-cell energy state (15).
Our data demonstrate that 100 nM agouti induced a gradual and sustained increase of [Ca2+]i in pancreatic ß-cells in the absence of glucose. Agouti also stimulated significant insulin release in these cells in the absence of glucose. This suggests that agouti stimulation of insulin release is not dependent on glucose. We also studied the effect of agouti on glucose-induced insulin release in HIT-T15 cells. Our data showed that the half-maximal effect of glucose in insulin secretion in HIT cells is 2.5 mM and reaches a plateau at 5 mM, similar to previous reports (38). The half-maximal effect of glucose in insulin release in human pancreatic islets is
7 mM. As the half-maximal dose of glucose is 8 mM for isolated pancreatic islets (47), the cell response to glucose is exaggerated in HIT-T15 cells (38). We showed that agouti stimulated a marked insulin release both in the absence and presence of different concentrations of glucose, suggesting that agouti is a potent insulin secretagogue.
However, we found that agouti induced a smaller increase in [Ca2+]i in the presence than in the absence of glucose. Several reports revealed that glucose has both a stimulatory and an inhibitory effect on both 45Ca efflux and insulin release (reviewed in Ref. 17). This inhibitory effect, however, is more obvious when Ca2+ entry into ß-cells is prevented by using diazoxide (2) or removal of extracellular Ca2+ (25). This has been attributed to the ability of glucose to promote intracellular Ca2+ sequestration. However, whether our observation also results from this effect of glucose is unclear, as the static incubation method used in our study made it impossible to further study this phenomenon.
The observed increase in [Ca2+]i may result either from the release of Ca2+ from intracellular stores or from Ca2+ influx. Previous data from this laboratory showed that agouti stimulation of lipogenesis was blocked by nitrendipine, suggesting a Ca2+ influx-mediated effect (22). Moreover, we have shown that the agouti-induced increase in [Ca2+]i is inhibited by Ca2+-channel antagonism in several cell types (23, 51). In addition, the structure of the COOH-terminal region of agouti protein exhibits structural homology to the -conotoxins and the plectoxins, two toxins produced by the cone snail Conus geographus and the hunting spider Plecteurys tristis, respectively, that primarily affect Ca2+-channel function (32). Further, our data using the Mn2+-quench technique clearly demonstrated agouti-stimulated Mn2+ influx into both HIT-T15 pancreatic ß-cells and human pancreatic islets, indicating Ca2+ influx into these cells in response to agouti stimulation.
This Ca2+ influx resulted in increases in both basal and glucose-stimulated insulin release in these cells, demonstrating that agouti is an insulin secretagogue in both pancreatic ß-cell lines and human pancreatic islets. Moreover, we have shown here that agouti is expressed in human pancreas. As agouti is a paracrine factor that does not enter general circulation (46), these data suggest that agouti may contribute to the development of hyperinsulinemia in humans in a paracrine fashion, similar to its paracrine actions in adipocytes (20) and hair follicles (19).
In addition to stimulation of basal insulin release, agouti also potentiated the effect of KCl on insulin release in both HIT-T15 cells and human pancreatic islets. Under basal (glucose free) conditions, agouti potentiation of KCl-induced insulin release is only apparent at high KCl concentrations. In the presence of 10 mM glucose, however, the effect of a low concentration of KCl (10 mM) on insulin release was significantly potentiated by agouti. This suggests that agouti potentiation of KCl-induced insulin release is dependent on glucose. As glucose has been shown to modulate ß-cell secretory response to increased [Ca2+]i stimulation (14, 15), the glucose dependence of agouti potentiation on KCl-induced insulin release may also involve mechanisms that require glucose to sensitize the ß-cell response to the primary stimulant. However, we also found that agouti potentiated KCl-induced Ca2+ influx into pancreatic ß-cells, suggesting that agouti may modulate Ca2+-channel activity, thereby sensitizing Ca2+-channel response to other stimuli. This mechanism, combined with glucose-dependent amplification of ß-cell secretory machinery to the primary [Ca2+]i signal, may contribute to agouti potentiation of KCl-induced insulin release in the presence of glucose.
Nonnutrient insulin secretagogues may modulate ß-cell ion channels (K+ channels or Ca2+ channels) by acting through cAMP or phospholipase pathways, activating protein kinase A or protein kinase C, or may increase the release of arachidonic acid, thereby sensitizing these channels to primary stimuli. Alternatively, they may phosphorylate target proteins involved in secretory machinery, thereby amplifying cell response to the increased [Ca2+]i signal (reviewed in Ref. 30).
It is not clear whether the effect of agouti is similarly mediated by such mechanisms or is a direct effect on Ca2+ channels. However, the homology between the COOH-terminal region of agouti and -conotoxins and plectoxins suggests that agouti may act directly on Ca2+ channels and thereby modify channel activity (32).
It is well documented that agouti antagonizes melanocortin receptors and inhibits the generation of cAMP during its regulation of pigment production (19, 31). Whether agouti also inhibits cAMP production in pancreatic ß-cells is not clear, as neither cAMP nor other second messengers have been determined in our study. Consequently, the mechanisms involved in agouti potentiation of KCl-induced insulin release remains to be further explored.
Both central and peripheral effects of agouti in agouti-induced obesity have been extensively explored. It was recently reported that MC4-R-knockout mice developed an obesity/hyperinsulinemia syndrome similar to that found in agouti mutants (18). In addition, intracerebroventricular administration of the MC4-R agonist MTII inhibited feeding, whereas this effect was prevented by coadministration of SHU-9119, a potent MC4-R antagonist (12). These data indicate an important role for central melanocortin receptors in regulating feeding behavior, suggesting that central MC4-R antagonism may be one of the mechanisms involved in agouti-induced obesity.
Recent data showed that central administration of SHU-9119 prevented the anorexic effect of leptin, suggesting that MC4-R may be involved in leptin signaling (39). However, Boston et al. studied the effect of agouti and leptin deficiency on weight gain in agouti mutant (lethal yellow, Ay/a), leptin-knockout (lepob/lepob), and double-mutant (Ay/a lepob/lepob) mice. Agouti produced similar weight gain in both the wild type and lepob/lepob mice, demonstrating that the effects of agouti and leptin deficiency in inducing obesity are independent (4). Further, the leptin resistance seen in obese yellow mice is due to desensitization of the leptin receptor, not due to a defect in leptin signaling, as removal of leptin from these mice restored leptin sensitivity (4). These data suggest that agouti-induced obesity is independent of leptin action and that the increased level of leptin in agouti mutant mice may serve to attenuate agouti-induced obesity.
Alternatively, data in this report and in our previous reports suggest an important role for the peripheral effect of agouti in agouti-induced obesity. Insulin is well known to stimulate lipogenesis and inhibit hormone-induced lipolysis. Similarly, we have demonstrated that agouti not only promotes adipocyte lipogenesis (20, 52) but also inhibits lipolysis (48). Both murine and human adipocytes express high levels of MC2-R (3, 7), whose natural ligand is the potent lipolytic agonist, adrenocorticotropic hormone. Recent data showed that increasing [Ca2+]i in rat adipocytes dose dependently inhibited hormone-stimulated lipolysis (41). In addition, we have recently demonstrated that recombinant agouti protein inhibited lipolysis in human adipocytes via a Ca2+-dependent mechanism (48). Thus agouti appears to exert insulin-like effects on adipocytes. These effects, combined with agouti-induced insulin release, may then act synergistically on adipocytes and thereby contribute to the agouti-induced obesity syndrome. In addition, hyperinsulinemia may also contribute to insulin resistance by direct downregulation of the insulin receptor (27). This indicates that agouti stimulation of insulin release may also play an important role in the development of insulin resistance in agouti mutant mice.
In summary, our data indicate that agouti acts directly in pancreatic ß-cells and may thereby play an important role in the hyperinsulinemia of dominant agouti mutations. This hyperinsulinemia may act in an additive or synergistic manner with adipocyte agouti expression to increase adipocyte triglyceride accumulation; it may also contribute to the insulin resistance seen in the agouti-induced obesity syndrome.
![]() |
ACKNOWLEDGMENTS |
---|
Address for reprint requests and other correspondence: M. B. Zemel, Univ. of Tennessee, 1215 W. Cumberland Ave., Rm. 229, Knoxville, TN 37996-1900 (E-mail: mzemel{at}utk.edu).
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Visit Other APS Journals Online |