Uncarboxylated osteocalcin (GluOC), a bone-derived hormone, regulates energy metabolism by stimulating insulin secretion, pancreatic β-cell proliferation, and adiponectin expression in adipocytes. Previously, we showed that long-term intermittent or daily oral administration of GluOC reduced the fasting blood glucose level, improved glucose tolerance, and increased the fasting serum insulin concentration as well as pancreatic β-cell area in female mice fed a normal or high-fat, high-sucrose diet. We have now performed similar experiments with male mice and found that such GluOC administration induced glucose intolerance, insulin resistance, and adipocyte hypertrophy in those fed a high-fat, high-sucrose diet. In addition, GluOC increased the circulating concentration of testosterone and reduced that of adiponectin in such mice. These phenotypes were not observed in male mice fed a high-fat, high-sucrose diet after orchidectomy, but they were apparent in orchidectomized male mice or intact female mice that were fed such a diet and subjected to continuous testosterone supplementation. Our results thus reveal a sex difference in the effects of GluOC on glucose homeostasis. Given that oral administration of GluOC has been considered a potentially safe and convenient option for the treatment or prevention of metabolic disorders, this sex difference will need to be taken into account in further investigations.
- insulin resistance
- sex difference
osteocalcin is a bone matrix protein synthesized by osteoblasts that also functions as a hormone in the uncarboxylated state to regulate various aspects of physiology, including glucose metabolism, brain function, and male fertility (14, 17, 30, 39, 40). Uncarboxylated osteocalcin (GluOC) constitutes ∼10% of total circulating osteocalcin in adult mice under normal physiological conditions (2, 10). However, the serum concentration of GluOC is increased in pathological conditions such as vitamin K deficiency and osteoporosis (2, 26, 31). Animal studies have shown that an increase in the circulating concentration of GluOC prevents obesity and glucose intolerance (10, 13, 28, 30). Mice genetically deficient in Esp (a protein tyrosine phosphatase expressed in osteoblasts), a gain-of-function model for osteocalcin, were thus protected from obesity, glucose intolerance, and insulin resistance induced by a high-fat diet (14, 30). Continuous administration of GluOC via a subcutaneous osmotic pump lowered blood glucose levels and increased pancreatic β-cell mass, insulin secretion, and insulin sensitivity in mice with high-fat diet- or gold thioglucose-induced obesity, with this latter effect being achieved via upregulation of expression of the gene for the insulin-sensitizing adipokine adiponectin in adipocytes (10). Moreover, daily intraperitoneal injection of GluOC for 14 wk in mice resulted not only in a substantial improvement in glucose tolerance and insulin sensitivity but also in full recovery from hepatic steatosis induced by a high-fat diet (13). Although many studies indicate that GluOC is the only form of osteocalcin that exerts favorable effects on metabolism, both GluOC and carboxylated osteocalcin (GlaOC) were found to modulate adipocyte and myoblast function in one study (19). However, the increased GluOC and reduced GlaOC levels in serum of mice with osteoblast-specific inactivation of the genes for γ-carboxylase or vitamin K epoxide reductase were associated with improved endocrine function (12).
Several clinical studies have also implicated circulating osteocalcin in the regulation of glucose metabolism in humans (11). Most such studies have found that the serum concentration of osteocalcin is inversely correlated with markers of metabolic syndrome such as glucose intolerance and insulin resistance (11, 24, 25, 36). However, the effects of osteocalcin on glucose metabolism in humans appear to be dependent on sex difference (4, 24, 45). The serum concentration of osteocalcin is thus correlated with that of adiponectin in women, but not in men, for example (4). A low serum osteocalcin level was also more likely to be associated with hyperglycemia and hyperinsulinemia during an oral glucose tolerance test in women than in men (24). However, the underlying nature of this sex difference has remained unclear.
Recently, we showed that intermittent or daily oral administration of GluOC improved glucose handling in female mice fed a normal or high-fat, high-sucrose (HFS) diet in large part by promoting glucagon-like peptide-1 (GLP-1) secretion (34). In the present study, we performed similar experiments with male mice and found that long-term oral administration of GluOC induced glucose and insulin intolerance in those fed an HFS diet. Our results thus indicate that the beneficial effects of GluOC administration on glucose metabolism are restricted to females, with such GluOC treatment actually being harmful in males, especially in the setting of a fat- and sugar-rich diet.
MATERIALS AND METHODS
Preparation of recombinant GluOC.
Recombinant mouse GluOC was prepared as described previously (33). In brief, a glutathione S-transferase-GluOC fusion protein was isolated from bacteria by consecutive exposure to a buffer containing 0.1% Triton X-114 (Sigma-Aldrich, St. Louis, MO) and the same buffer without detergent to remove endotoxin. The GluOC moiety was cleaved from glutathione S-transferase with the use of thrombin, which was then removed from the mixture with the use of a Benzamidine Sepharose 4 Fast Flow column (GE Healthcare, Little Chalfont, UK). The purity of the GluOC preparation was assessed by SDS-PAGE with a Tris-Tricine buffer system, followed by staining with Coomassie brilliant blue, and the concentration and integrity of the purified protein were determined with an enzyme-linked immunoassay kit for mouse GluOC (Takara Bio, Shiga, Japan).
Animals and GluOC administration.
All animal experiments were approved by the Animal Ethics Committee of Kyushu University (approval no. A27-138). Male or female C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME) were maintained in a specific pathogen-free facility under a 12:12-h light-dark cycle and fed normal chow (CRF-1; Oriental Yeast, Osaka, Japan) or an HFS diet (F2HFHSD; Oriental Yeast) ad libitum immediately after weaning. The composition of the diets is shown in Table 1, and the experimental schedules are presented in Fig. 1. Recombinant mouse GluOC (3 ng/g) in 100 μl of physiological saline was administered orally to intact mice three times/wk (at 0800 on Monday, Wednesday, and Friday) for 13 wk, beginning at 3 wk of age, with the use of a gastric gavage needle attached to a syringe and held at the entrance of the esophagus. Orchidectomy, ovariectomy, or sham surgery was performed under anesthesia with pentobarbital sodium (40 mg/kg) in combination with sevoflurane at 3 wk of age. For examination of the effects of estradiol administration, a β-estradiol 17-cypionate pellet (NE-291, 0.5 mg/pellet, 90-day release; Innovative Research of America, Sarasota, FL) was implanted subcutaneously at the anterior back during surgery. Mice were treated with GluOC as described above for 13 wk after surgery. On the other hand, for examination of the effects of testosterone administration, orchidectomy was performed at 6 wk of age. A 28-day miniosmotic pump (Alzet, model 1004; Durect, Cupertino, CA) was also implanted subcutaneously at the anterior back of orchidectomized (ORX) male mice or intact female mice at 6 wk of age. The pump was used to administer either testosterone (Sigma-Aldrich) at 1.1 ng/h or vehicle (10% ethanol, 90% propylene glycol), with the dose of testosterone being the same as that applied in a previous study (22). Mice were fed the HFS diet beginning 2 wk before, and GluOC was administered orally every day at a dose of 10 ng/g for 4 to 7 wk after the surgery. The GluOC dose was increased from 3 ng/g three times/wk for 10–13 wk to 10 ng·g−1·day−1 for 4–7 wk because of the delivery period for the available osmotic pump. Previously, we confirmed that a dose of 10 ng·g−1·day−1 for 4 wk produced results similar to those for a dose of 3 ng/g three times/wk for 10 wk (34).
An intraperitoneal glucose tolerance test (IPGTT) was performed in mice that had been deprived of food for 22 h. Glucose (2 or 1 g/kg for mice fed the normal or HFS diet, respectively) was administered intraperitoneally, and the blood glucose concentration was measured at various times thereafter with the use of FreeStyle Lite Blood Glucose test strips (Abbott Laboratories, Abbott Park, IL). An insulin tolerance test (ITT) was performed in animals that had been deprived of food for 4 h. Insulin (Humulin R; Eli Lilly, Indianapolis, IN) was injected intraperitoneally at 0.5 or 1.5 U/kg for mice fed the normal or HFS diet, respectively, and the blood glucose concentration was measured at the indicated times thereafter. Lower glucose and higher insulin doses were required for mice fed an HFS diet compared with those fed a normal diet to obtain an appropriate response, likely as a result of impairment of glucose tolerance and insulin sensitivity induced by HFS feeding. For serological analysis, mice deprived of food for 22 h were anesthetized with sevoflurane, and blood was collected from the orbital plexus for serum preparation. Thus serum prepared was assayed for insulin with an ELISA kit (Mercodia, Uppsala, Sweden) as well as for testosterone, adiponectin, and GLP-1 levels with ELISA kits (R & D Systems, Minneapolis, MN; Merck-Millipore, Billerica, MA; and Shibayagi, Gunma, Japan, respectively). Serum concentrations of triglyceride and nonesterified fatty acid were also analyzed by the Nagahama Institute for Biochemical Science (Oriental Yeast, Shiga, Japan).
Mice were deprived of food for 4 h, injected intraperitoneally with insulin (0.15 U/kg; Eli Lilly), and then euthanized by cervical dislocation 5 min thereafter. Liver, skeletal muscle, and gonadal white adipose tissue (WAT) were immediately isolated and snap-frozen in liquid nitrogen. The tissues were subsequently lysed in a solution containing 20 mM Tris·HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 1% sodium deoxycholate, 2.5 mM Na4P2O7, 1 mM β-glycerophosphate, 1 mM Na3VO4, 100 μM NaF, and a protease inhibitor cocktail consisting of pepstatin A (2.5 μg/ml), leupeptin (5 μg/ml), 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (25 μg/ml), and aprotinin (1.7 μg/ml). The lysates were centrifuged at 20,000 g for 30 min at 4°C, and the protein concentration of the resulting supernatants was determined with the use of a Protein Assay Rapid Kit (Wako, Osaka, Japan). Portions of the supernatants (10 μg of protein) were subjected to SDS-PAGE, the separated proteins were transferred to a polyvinylidene difluoride membrane (Merck-Millipore), and the membrane was then exposed to 5% dried skim milk in Tris-buffered saline containing 0.1% Tween-20 before incubation with primary antibodies to total or phosphorylated forms of Akt (1:1,000 dilution of no. 9272 or no. 4060, respectively; Cell Signaling Technology, Danvers, MA). WAT lysates were also analyzed with antibodies to the β-chain of the insulin receptor (1:1,000 dilution of no. 3025; Cell Signaling Technology). Immune complexes were visualized with horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence reagents (GE Healthcare). The intensity of immunoreactive protein bands was measured with the use of ImageJ software (National Institutes of Health).
Gonadal WAT and liver were excised, fixed in 10% neutral formalin, dehydrated with a series of ethanol solutions, embedded in paraffin, and sectioned at a thickness of 6 μm. The sections were then depleted of paraffin, rehydrated with PBS, and stained with Mayer's hematoxylin and EosinY (Muto Pure Chemicals, Tokyo, Japan). The number and size of adipocytes in sectional areas of gonadal WAT as well as the area of white spots (indicative of lipid droplets) in sections of the liver were evaluated with the use of a BZ-II Analyzer (Keyence, Osaka, Japan). Histomorphometric analysis of the pancreas was performed as described previously (34). In brief, dissected pancreatic tissue was fixed in 10% neutral formalin, embedded in paraffin, and sectioned at 6 μm for immunostaining with rabbit antibodies to insulin (1:1,000 dilution of no. 4590; Cell Signaling Technology). Immune complexes were visualized with biotinylated goat antibodies to rabbit IgG (1:1,000 dilution; Vector Laboratories, Burlingame, CA) and the use of an ABC Elite Kit plus ImmPact DAB (Vector Laboratories). The sections were counterstained with Mayer's hematoxylin (Muto Pure Chemicals) and analyzed with the use of a BZ-II Analyzer (Keyence). β-Cell area was measured as the surface positive for insulin immunostaining. At least three specimens of each tissue were analyzed for each mouse, and five to 15 mice per group were analyzed.
RT and real-time PCR analysis.
Total RNA was isolated from the liver with the use of an RNeasy Lipid Tissue Mini Kit (Qiagen, Hilden, Germany), and 1 μg of the isolated RNA was subjected to RT for 2 h at 37°C with the use of a High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA). The resulting cDNA was subjected to real-time PCR analysis with the use of a TaKaRa PCR Thermal Cycler Dice Gradient (Takara Bio) and KOD SYBR quantitative PCR Mix (Toyobo, Osaka, Japan). The PCR protocol included an initial incubation at 98°C for 2 min, followed by 40 cycles of 98°C for 10 s, 61°C for 10 s, and 68°C for 30 s. β-Actin cDNA was amplified as an internal control. PCR primer sequences were 5′-GTCAAAACCAGCCTCCCAAG-3′ (forward) and 5′-CAGTCCCCGTCCACAAAGA-3′ (reverse) for for sterol regulatory element-binding protein 1c (SREBP-1c) and 5′-ATGAAGAGGGCTGAGCGTAGGTAA-3′ (forward) and 5′-TGCCGTTGTCTGTCACTGTCTGAA-3′ (reverse) for peroxisome proliferator-activated receptor-α (PPARα). Primer sequences for β-actin were obtained from PrimerBank (6671509a1, http://pga.mgh.harvard.edu/primerbank/index.html).
Quantitative data are presented as means ± SE. Statistical analysis of measurements at single time points was performed with the unpaired, two-tailed Student's t-test. Two-way ANOVA for repeated measurements and the Bonferroni post hoc test were performed with the use of Prism software version 6.0 (GraphPad Software, La Jolla, CA). A P value of <0.05 was considered statistically significant.
Oral administration of GluOC induces glucose intolerance and insulin resistance in male mice fed an HFS diet.
Previously, we showed that long-term intermittent oral administration of GluOC improves glucose tolerance and lowers fasting blood glucose levels in female mice fed a normal or HFS diet (34). Performance of the same experiments (Fig. 1) with male mice revealed that such administration of GluOC had no effect on body weight gain in animals fed a normal or HFS diet (Fig. 2A), consistent with our previous results in female mice (34). Food consumption (4.7 ± 0.3 and 4.4 ± 0.3 g/day for mice administered saline and GluOC, respectively) as well as the weights of gonadal WAT and the liver (Fig. 2B) were also similar for the control and experimental animals fed the HFS diet. The serum concentration of endogenous GluOC in mice maintained on the HFS diet for 2 mo was reduced by ∼50% compared with that in those maintained on the normal diet (0.42 ± 0.05 vs. 0.83 ± 0.04 ng/ml, respectively, P < 0.01). An IPGTT and ITT revealed no effect of GluOC administration in mice fed the normal diet (Fig. 2C). However, for those fed the HFS diet, the IPGTT and ITT showed that GluOC administration induced intolerance to both glucose and insulin (Fig. 2D). Glucose-stimulated insulin secretion (data not shown), the serum GLP-1 level (data not shown), the fasting serum insulin concentration (Fig. 2E), and pancreatic β-cell area (Fig. 2E) were unaffected by GluOC administration in male mice fed the normal or HFS diet, although the latter two parameters tended to be reduced by GluOC in those fed the HFS diet; in contrast, all of these parameters were previously found to be increased by GluOC administration in female mice fed a normal diet (34). The smaller β-cell area of male mice fed the HFS diet compared with those fed the normal diet (Fig. 2E) was likely due to apoptosis induced by long-term consumption of the HFS diet (7, 8). We also examined intracellular insulin signaling in WAT, liver, and skeletal muscle to determine which organ might be responsible for the insulin resistance of male mice fed the HFS diet and administered GluOC. GluOC administration inhibited insulin-induced phosphorylation of the protein kinase Akt in WAT (Fig. 2F), but not in liver or skeletal muscle (data not shown), of mice fed the HFS diet. Immunoblot analysis with antibodies to the β-chain of the insulin receptor revealed that the total abundance of the receptor in WAT was not affected by GluOC administration in male mice fed the HFS diet (data not shown). The reason for the GluOC-induced impairment of insulin signaling specifically in adipocytes remains unknown. Together, these results thus indicated that, in contrast to its beneficial effects in female mice, long-term oral administration of GluOC impairs glucose utilization in male mice fed an HFS diet, likely in part as a result of attenuation of insulin signaling in WAT.
Oral administration of GluOC increases adipocyte size in male mice.
We next examined the size of adipocytes in gonadal WAT, given that we showed previously that long-term oral GluOC administration reduced the size of these cells in female mice fed a normal (37) or HFS (data not shown) diet. Morphometric analysis of adipocyte size distribution revealed that oral administration of GluOC increased the proportion of larger adipocytes in male mice fed a normal or HFS diet (Fig. 3, A and B).
We also measured the serum concentration of adiponectin in male and female mice fed a normal or HFS diet and treated with GluOC or saline, given that adiponectin is an insulin-sensitizing adipokine whose secretion is known to be downregulated in hypertrophic adipocytes (58, 59). Whereas long-term oral administration of GluOC had no effect on the serum adiponectin level in male mice fed the normal diet, it significantly reduced that in male mice fed the HFS diet (Fig. 3C). In contrast, the serum adiponectin concentration was not significantly affected by GluOC treatment in female mice fed either diet, although it showed a tendency to be increased in those fed the HFS diet (Fig. 3C). These results thus suggested that oral GluOC administration induced adipocyte hypertrophy and reduced the secretion of adiponectin in male mice, especially in those fed an HFS diet.
Oral administration of GluOC increases the serum testosterone level in male mice.
Most sex differences are attributable to steroid sex hormones. Furthermore, intraperitoneal injection of GluOC was recently shown to promote testosterone production and secretion by testicular Leydig cells in male mice (38, 40). Therefore, we examined whether testosterone might be responsible for the sex difference in the effects of oral GluOC administration on metabolic status in mice. HFS feeding itself reduced the serum concentration of testosterone, whereas long-term oral administration of GluOC induced a small but significant increase in this parameter in male mice fed either a normal or HFS diet (Fig. 4). GluOC had no such effect in females (data not shown).
Orchidectomy protects against glucose and insulin intolerance induced by oral GluOC administration in male mice.
To examine the role of testosterone in the effects of oral GluOC administration on glucose metabolism, we subjected male mice to orchidectomy at 3 wk of age. We confirmed that ORX mice showed a reduction in the serum testosterone level of ∼85% at 15 wk of age (data not shown). Long-term oral administration of GluOC had no effect on body or liver weight in ORX mice fed a normal (data not shown) or HFS (Fig. 5, A and B) diet, whereas it tended to reduce the weight of gonadal WAT in those fed the HFS diet (Fig. 5B). Food consumption was unaffected by GluOC in ORX mice fed the HFS diet (4.2 ± 0.2 and 4.4 ± 0.2 g/day for those administered saline and GluOC, respectively). Serum insulin and GLP-1 levels in the fed state tended to be increased by GluOC administration in ORX mice maintained on the normal or HFS diet (data not shown). The metabolic responses of ORX males to GluOC administration were similar to those of intact females (34), with the exception that glucose-stimulated insulin secretion was not significantly affected (data not shown). Oral GluOC treatment thus tended to improve glucose tolerance and significantly increased the fasting serum insulin concentration without affecting insulin sensitivity in ORX males fed the normal diet (Fig. 5C). It also significantly improved both glucose tolerance and insulin sensitivity and tended to ameliorate fasting hyperinsulinemia in ORX mice fed the HFS diet (Fig. 5D). On the other hand, ovariectomized (OVX) females showed no difference from control mice in the effect of oral GluOC administration on their IPGTT response (Fig. 5E), indicating that female sex hormones do not play a role in GluOC action.
Administration of GluOC significantly increased the percentage of smaller adipocytes in ORX males fed either a normal or HFS diet, although this effect was slight in the former animals (Fig. 6, A and B). Serum adiponectin levels were higher in ORX mice fed a normal or HFS diet compared with the corresponding control mice (Fig. 6, C and D), consistent with previous observations (35, 56, 60). Whereas GluOC significantly reduced the serum adiponectin concentration in sham-operated male mice fed an HFS diet, it had no effect in ORX mice (Fig. 6D).
Testosterone supplementation restores GluOC-induced insulin resistance in ORX mice fed an HFS diet.
We next examined the effects of testosterone supplementation on the action of GluOC in ORX mice fed an HFS diet. Male mice were subjected to orchidectomy at 6 wk of age, and a pump was implanted for continuous delivery of testosterone or vehicle. The HFS diet was initiated 2 wk before, and GluOC (10 ng/g) or saline was administered orally each day for 7 wk after the surgery (Fig. 1). Supplementation with testosterone resulted in a significant increase in the serum testosterone concentration, whereas GluOC treatment had no effect on this parameter (Fig. 7A). The serum adiponectin level was significantly reduced by testosterone supplementation in mice treated with saline, but it was not affected by testosterone in those treated with GluOC (Fig. 7B). Whereas GluOC administration improved glucose tolerance slightly without affecting insulin sensitivity in control ORX mice (Fig. 7C), it tended to trigger intolerance to both glucose and insulin in ORX mice supplemented with testosterone (Fig. 7D). Similar experiments performed with ORX males supplemented with estradiol revealed no differences in GluOC action on glucose metabolism compared with control ORX mice (Fig. 7E). These results suggested that the serum testosterone level plays a key role in the adverse effects of long-term oral GluOC administration on glucose tolerance and insulin sensitivity.
Testosterone supplementation reverses GluOC effects in female mice fed an HFS diet.
We also examined the effects of testosterone supplementation on GluOC action in female mice fed an HFS diet (Fig. 1). Supplementation with testosterone resulted in a significant increase in the serum testosterone level, whereas GluOC had no effect on this parameter (Fig. 8A). The serum adiponectin level was reduced by testosterone supplementation in mice treated with saline, but it was unaffected by testosterone in those treated with GluOC (Fig. 8B). GluOC administration improved glucose tolerance in control female mice (Fig. 8C), consistent with our previous results (34), whereas it induced glucose intolerance in females supplemented with testosterone (Fig. 8D). These results thus provided further evidence for the important role of testosterone level in the adverse effects of long-term oral GluOC administration on glucose tolerance.
Daily administration of GluOC induces lipid accumulation in the liver of testosterone-supplemented ORX mice fed an HFS diet.
Given that hepatic lipid accumulation is often associated with systemic insulin resistance (3, 61), we performed a histological analysis of the liver in testosterone-supplemented ORX mice fed an HFS diet. GluOC administration induced a marked increase in the extent of lipid accumulation in the liver of testosterone-supplemented mice, whereas it reduced that in the liver of control ORX mice (Fig. 9, A and B). Neither GluOC nor testosterone significantly affected serum triglyceride or nonesterified fatty acid levels in ORX mice (Fig. 9C). The concentrations of these analytes were also not significantly affected by GluOC administration in intact male or female mice fed an HFS diet (data not shown). Finally, RT and real-time PCR analysis revealed that the abundance of the mRNA for PPARα, a transcription factor responsible for the expression of genes related to β-oxidation of fatty acids (16, 49), was downregulated by GluOC in the liver of testosterone-supplemented ORX mice fed an HFS diet, whereas it was upregulated by GluOC in control ORX mice (Fig. 9D). Furthermore, GluOC had opposite effects on the abundance of the mRNA for SREBP-1c, a transcription factor responsible for the expression of genes related to fatty acid synthesis (21), in these groups of animals (Fig. 9D).
The notion that bone acts as an endocrine organ to regulate glucose and energy metabolism has become well established (10–14, 17, 19, 28, 30, 53), with GluOC having been administered by intraperitoneal injection as a therapeutic agent for diabetic mice (13). We have shown previously that the beneficial effects of GluOC on glucose metabolism are mediated in part by GLP-1 (33, 34) and that oral administration of GluOC is as effective as peritoneal injection (34). However, our previous studies were performed with female mice. Performance of similar experiments with male mice has now revealed that long-term oral GluOC administration induced intolerance to glucose and insulin in the setting of an HFS diet. In contrast, GluOC improved glucose and insulin tolerance in ORX males fed an HFS diet, similar to its effects in females. Supplementation with testosterone restored the deleterious action of GluOC in ORX males.
GluOC has been shown recently to stimulate testosterone production by Leydig cells in the testes of male mice (38, 40). The increased circulating concentration of GluOC in Esp-deficient mice or in wild-type mice injected intraperitoneally with GluOC thus resulted in increased testosterone production by Leydig cells (38, 40). We also observed that oral administration of GluOC increased the serum testosterone level in male mice in the present study. In addition, the circulating GluOC level has been positively associated with that of free testosterone in men with or without type 2 diabetes (4, 23). The deleterious effects of GluOC on glucose and energy metabolism in male mice are likely due, at least in part, to increased production of testosterone. On the other hand, estrogen does not appear to contribute to the beneficial effects of GluOC in female mice, given that 1) GluOC was previously found to have no effect on estrogen production by the ovaries (40), 2) GluOC was as effective at improving glucose tolerance in ovariectomized female mice as in sham-operated controls, and 3) supplementation of ORX male mice with estradiol instead of testosterone did not affect the metabolic responses of these animals to GluOC administration.
The serum concentration of GluOC itself was reduced in male mice fed an HFS diet, consistent with previous observations of obese humans (46, 50). Insulin resistance in osteoblasts generated by partial or total deletion of the insulin receptor gene was previously shown to reduce bone turnover via downregulation of the function of both osteoblasts and osteoclasts (17, 55). Given that HFS feeding triggers insulin resistance, reduced GluOC production by osteoblasts as a result of impaired insulin signaling might be responsible for the reduced serum level of GluOC in male mice fed an HFS diet. Such reduced GluOC levels might contribute to the lower serum concentration of testosterone in male mice fed an HFS diet than in those fed a normal diet, and therefore, they might be expected to be beneficial for glucose metabolism. However, intolerance to glucose and insulin triggered by a HFS diet is also caused by the release of inflammatory cytokines (adipokines) from hypertrophic adipocytes (47, 48) and downregulation of the receptors for insulin and adiponectin (52), effects that likely mask any action of a reduced serum testosterone concentration.
The expression of adiponectin appears to be stimulated by GluOC to a greater extent in females than in males (4, 24, 45). The plasma adiponectin level is lower in men than in women regardless of menopausal status (5, 36, 56). The serum levels of osteocalcin and adiponectin have also been found to be positively associated in postmenopausal women but not in men (25). Moreover, an adiponectin-dependent, insulin-sensitizing effect of osteocalcin has been observed only in women (4). Serum adiponectin concentrations were found to be higher in hypogonadal men than in eugonadal controls, and these increased adiponectin levels were reduced by testosterone replacement therapy (29). Similar findings have been obtained with mice. Adult female mice thus have higher adiponectin levels compared with males (6, 60); increased circulating testosterone levels in male mice were found to reduce the plasma adiponectin concentration without affecting the abundance of adiponectin mRNA in adipose tissue (35, 56), and the serum adiponectin level in androgen receptor-deficient male mice was higher than that in controls (9). These various observations support the notion that testosterone negatively regulates adiponectin secretion. We have now shown that the serum adiponectin concentration was reduced, whereas the serum testosterone level was increased, by GluOC administration in male mice on a HFS diet. The development of GluOC-induced insulin resistance in such mice was prevented by orchidectomy, and this effect of surgery was in turn reversed by testosterone supplementation. Furthermore, testosterone supplementation in female mice also led to a reduction in the serum adiponectin concentration accompanied by a change in the effect of GluOC on glucose tolerance from beneficial to adverse. It is thus possible that the increased testosterone levels induced by GluOC in male mice fed an HFS diet inhibit the secretion of adiponectin by adipocytes and thereby contribute to insulin resistance.
Serum adiponectin levels were not affected by diet type, whereas HFS feeding triggered adipocyte hypertrophy in male mice in the present study, and adiponectin secretion is known to be downregulated in hypertrophic adipocytes (58, 59). Serum adiponectin levels were also previously shown to be reduced in obese humans and animal models of obesity (27, 47), although other studies have found no relation between serum adiponectin concentration and obesity (1, 20, 51). On the other hand, the biological activity of adiponectin is dependent on its oligomerization and consequent ability to stimulate type 1 and 2 receptors and thereby activate AMP-activated protein kinase and PPARα, respectively (41, 54, 57). The oligomerization state of adiponectin in mice in the present study was not investigated.
Testosterone exerts many of its actions through binding to its nuclear receptor and the consequent transcriptional regulation of its target genes (18). In addition to this classical signaling mechanism mediated by nuclear receptors, however, recent studies have shown that steroid hormones, including testosterone, have rapid nongenomic effects that are initiated at the plasma membrane (15, 32). Multiple signal transduction pathways have been linked to the nongenomic actions of testosterone, with adenylyl cyclase, MAPKs, phosphatidylinositol 3-kinase, and intracellular calcium having been implicated (15, 42). Although it remains to be determined whether the suppressive effect of testosterone on adiponectin secretion is mediated by genomic or nongenomic mechanisms, it is likely a nongenomic action given that adiponectin secretion has been found to be reduced by testosterone without an affect on adiponectin mRNA abundance in adipose tissue (6, 35, 56). Indeed, the putative GluOC receptor GPRC6A (40, 44) is expressed in adipocytes (37, 42) and has been shown to mediate nongenomic effects of testosterone in vitro and in vivo (43). On the other hand, treatment of rat adipocytes with actinomycin D, an inhibitor of transcription, prevented the testosterone-induced inhibition of the release of high-molecular-weight adiponectin into the culture medium (56), suggesting that this effect of testosterone is mediated at the transcriptional level.
We found that long-term oral administration of GluOC appeared to induce a low level of insulin resistance in testosterone-supplemented ORX mice fed an HFS diet. Given that GluOC is not able to stimulate testosterone production in ORX mice because of the loss of Leydig cells, it appears to induce insulin resistance in these animals independently of its effect on testosterone synthesis. It is possible that GluOC administration induces liver steatosis in testosterone-supplemented ORX mice fed an HFS diet, with lipid accumulation in the liver often having been associated with insulin resistance (3, 61). Indeed, histological analysis of the liver of these mice revealed increased lipid accumulation compared with controls. The mechanism whereby GluOC might induce lipid accumulation in this setting remains unknown. Daily intraperitoneal injection of GluOC was previously shown to prevent the appearance of liver steatosis in mice fed a high-fat diet, but the sex of the animals was not identified (13). We found that GluOC prevented lipid accumulation in the liver of ORX mice fed an HFS diet but not supplemented with testosterone. It is thus possible that testosterone suppresses the ability of GluOC to reduce serum triglyceride levels and lipid accumulation that has been observed in obese female mice (10). Alternatively, GluOC may act directly on hepatocytes to block or promote the accumulation of lipids in the liver of female or male animals, respectively. Indeed, we found that GluOC had differential effects on the expression of genes related to lipid metabolism in the liver of HFS-fed ORX mice supplemented with testosterone or with vehicle. The expression of the PPARα gene was downregulated by GluOC in the liver of the testosterone-supplemented ORX mice, whereas it was upregulated in that of the control mice. Furthermore, the abundance of SREBP-1c mRNA was increased by GluOC in the liver of the testosterone-supplemented mice, whereas it was reduced in that of the control mice. We also confirmed the expression of GPRC6A as a potential mediator of these effects in the liver (data not shown).
Several studies have attempted to address the importance of osteocalcin in the regulation of glucose metabolism in humans (11). The serum osteocalcin level has thus been found to be negatively correlated with blood glucose concentration, insulin resistance, obesity, or markers of metabolic syndrome (24, 25, 46, 50). Moreover, an increase or decrease in circulating GluOC levels induced by treatment with parathyroid hormone or alendronate, respectively, in osteoporotic postmenopausal women was inversely associated with changes in body weight, fat mass, and the circulating adiponectin concentration (46). Given that osteocalcin-deficient mice are hyperglycemic and insulin resistant (14, 30), the physiological role of circulating GluOC may be to maintain basal levels of blood glucose and insulin secretion. In addition, the regulation of metabolism by GluOC is highly sensitive to exogenous application of the hormone (10, 13, 34, 37). Previously, we showed that oral GluOC administration for 4 wk or longer increased the serum GluOC concentration from 0.6 to up to 1 ng/ml (34), an increase that would be expected to be effective (34). Even a small increase in the circulating GluOC level would thus be expected to affect glucose handling in both animals and humans (10, 34, 46).
In conclusion, our results have revealed that, in contrast to its improvement of glucose handling in female mice, long-term oral administration of GluOC induced insulin resistance in male mice fed a HFS diet. Thus there is an important sex difference in the endocrine function of GluOC, which will need to be taken into account in further investigations into the potential of GluOC administration as an option for the prevention or treatment of metabolic diseases.
This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (KAKENHI Grants 24229009 to M. Hirata and 26861553 to A. Mizokami) as well as grants from Uehara Memorial Science Foundation, Takeda Science Foundation, and Shimabara Science Foundation (to A. Mizokami).
The authors declare no potential conflicts of interest, financial or otherwise, relevant to this study.
Y.Y., A.M., T.K.-Y., H.T., and M.H. conception and design of research; Y.Y., A.M., and S.C. performed experiments; Y.Y., A.M., and M.H. drafted manuscript; Y.Y., A.M., T.K.-Y., S.C., I.T., H.T., and M.H. approved final version of manuscript; A.M. and M.H. analyzed data; A.M. and M.H. interpreted results of experiments; A.M. and M.H. prepared figures; A.M., T.K.-Y., I.T., H.T., and M.H. edited and revised manuscript.
We thank K. Kaku (Kyushu University) for technical assistance as well as Dr. M. Nakatomi and T. Nakatomi (Kyushu Dental University) for help in preparing paraffin sections.
- Copyright © 2016 the American Physiological Society