Amylin (AMY) is a 37 amino acid peptide cosecreted with insulin (INS) by pancreatic β-cells and absent in type 1 diabetes, a condition frequently associated with osteopenia. AMY binds to calcitonin receptors, lowers plasma calcium concentration, inhibits osteoclast activity, and stimulates osteoblasts. In the present study, we examined the effects of AMY replacement on bone loss in a streptozotocin (STZ)-induced rodent model type 1 diabetes. Of 50 male Wistar rats studied, 40 were made diabetic with intraperitoneal STZ (50 mg/kg; plasma glucose concentrations >11 mM within 5 days). Ten nondiabetic control (CONT) rats received citrate buffer without STZ. Diabetic rats were divided into four groups (n = 10/group) and injected subcutaneously with rat AMY (45 mg/kg), INS (12 U/kg), both (same doses), or saline (STZ; diabetic controls) once per day. After 40 days of treatment and five 24-h periods of urine collection for deoxypyridinoline (DPD), the animals were killed, blood was sampled, and femurs were removed. The left femur was tested for mechanical resistance (three-point bending). The right femur was tested for total, diaphyseal (cortical bone), and metaphyseal (trabecular bone) bone densities using dual-energy X-ray absorptiometry (DXA). Bone was ashed to determine total bone mineral (calcium) content. None of the treatments had any significant effect on femoral length and diameter. Untreated diabetic rats (STZ; 145 ± 7N) had lower bone strength than did nondiabetic CONT (164 ± 38; p < 0.05). Total bone mineral density (BMD; g/cm2) was significantly lower in STZ (0. 2523 ± 0. 0076) than in CONT (0.2826 ± 0.0055), as were metaphyseal and diaphyseal densities. Diabetic rats treated with AMY, INS, or both had bone strengths and bone densities that were indistinguishable from those in nondiabetic CONT. Changes in bone mineral content paralleled those for total BMD (T-BMD). Plasma osteocalcin (OC) concentration, a marker for osteoblastic activity, was markedly lower in untreated diabetic rats (7. 6 ± 0. 9 ng/ml); p < 0. 05) than in nondiabetic CONT (29. 8 ± 1. 7; p < 0. 05) or than in AMY (20. 1 ± 0. 7; p < 0. 05). Urinary DPD excretion, a marker for bone resorption, was similar in untreated and AMY-treated diabetic rats (35.0 ± 3.1 vs. 35.1 ± 4.4 nmol/mmol creatinine), intermediate in rats treated with INS (49.9 ± 2.7), and normalized in diabetic rats treated with both agents (58.8 ± 8.9 vs. 63.2 ± 4.5 in CONT). Thus, in our STZ rat model of diabetic osteopenia, addition of AMY improved bone indices apparently by both inhibiting resorption and stimulating bone formation.
AMYLIN (AMY) is a 37 amino acid hormone secreted from pancreatic β-cells with insulin (INS) in response to nutrient stimuli.(1) AMY has a physiological role of coordinate regulation of nutrient uptake via several mechanisms, including inhibition of food intake, slowing of gastric emptying, regulation of acid and digestive enzyme secretion, and inhibition of nutrient-stimulated glucagon secretion.(2) These actions, which are centrally mediated,(3) complement the peripheral actions of INS to accelerate nutrient disposal. AMY receptors mediating this metabolic regulation recently have been characterized as being comprised of a calcitonin receptor isoform in association with a receptor activity modifying protein (RAMP).(4, 5)
Although it is only 13% homologous to mammalian calcitonins, AMY is surprisingly potent at inducing hypocalcemia in vivo (where the half effective concentration [EC50] with continuous infusion is 10 pM(6)), inhibiting resorption in osteoclast preparations,(7) and in stimulating osteoblasts number and function.(8) The receptors whereby AMY exerts these actions on bone are unclear. Although some aspects of AMY's action on plasma calcium and osteoclasts appear to be mediated via calcitonin receptors, motility of osteoclasts appears to be under the control of at least two signaling pathways.(9) The “quiescence” or Q effect on osteoclasts responds to AMY, calcitonin, and calcitonin gene-related peptide (CGRP), while the “retraction” or R effect is preferentially sensitive to calcitonin. However, osteoblasts are sensitive to CGRP and AMY but are unaffected by calcitonin.
Type 1 diabetes mellitus is associated with total or near-total absence of AMY.(10) In late type 2 diabetes, AMY is present at a low concentration but does not change in response to meals; while in INS-resistant states, AMY is present in excess. Type 1 diabetes (but not type 2 diabetes or INS resistance) is associated with osteopenia, and it has been proposed that the lack of bone-preserving actions of AMY may at least partly account for this disease association.(9)
Treatment of type 1 diabetic subjects for 1 year with the AMY agonist pramlintide was reported to have no effect on bone markers in one study(11) but improved these markers in postmenopausal women in another.(12) The aim of the present work was to investigate the effect of AMY replacement on bone loss, bone strength, and biochemical markers of resorption and formation in a commonly used model of (AMY deficient) type 1 diabetes mellitus(13) and diabetic osteopenia, the streptozotocin (STZ)-induced diabetic rat.
MATERIALS AND METHODS
Animals and treatments
These experiments were made in accordance with current legislation on animal experiments in France. Each rat was housed individually in a plastic cage allowing separation and collection of urine and feces. Ambient temperature was 21°C, with a 12 h/12 h light/dark cycle. Individual daily consumption of the diet containing 0.65% Ca and 0.73% phosphorus (UAR A03; Usine d'Alimentation Rationnelle, Villemoisson sur Orge, France) was measured. Each animal was weighed weekly. Fifty male Wistar rats weighing 440 ± 3g (mean ± SEM) were used at 4 months of age. Of these, 40 received a single intraperitoneal (ip) injection of STZ (Sigma, L'Isle d'Abeaux, France; 50 mg/kg in 0.1 M citrate buffer). The 10 other rats served as nondiabetic controls (CONT) and received citrate buffer only. Diabetes was confirmed by glycosuria 2 days later. The 40 diabetic rats were stratified by body weight into four treatments, which lasted for 40 days:
(1) AMY (n = 10), injected daily subcutaneously (sc) with 45 μg/kg synthetic rat AMY (Amylin Pharmaceuticals, Inc., San Diego, CA, USA)
(2) INS (n = 10), injected daily sc with 15 U/kg synthetic human INS (Ultratard HM GE; Novo Nordisk Pharmaceutique S.A., Boulogne Billancourt, France).
(3) AMY + INS (n = 10), injected daily sc with AMY and INS as separate injections, doses as mentioned previously.
Urine was collected during a 24 h-period on days 32, 34, 36, 38, and 40 to measure the excretion of calcium and of deoxypyridinoline (DPD), a marker for bone resorption.(14)
Rats were killed by cervical dislocation on day 41. Blood was collected by cardiac puncture. After centrifugation, plasma was harvested and frozen until analysis.
Femoral mechanical testing:
Immediately after collection, the length of the right femur and the mean diameter of the femoral diaphysis were measured using a caliper. Because of the irregular shape of the femoral diaphysis, the femoral diameter used in the calculation was the mean of the greatest and the smallest femoral diaphysis diameters. Each bone was stored in 0. 9% NaCl at 4°C before testing mechanical resistance 24 h later by using a three-point bending test. Each bone was secured on the two lower supports (diameter, 4 mm; separation, 20 mm) of the anvil of a Universal Testing Machine (Instron 4501; Instron, Canton, MA, USA). An upper crosshead roller (diameter, 6 mm) advanced at 0. 5 mm/minute until rupture was automatically determined by the Instron 4501 software. Load (Newtons) at rupture was recorded. The femur was mounted so that the crosshead was applied in the middle of the bone. This test method was previously validated by using Plexiglas standard probes.(15)
Bone mineral density:
Dual-energy X-ray absorptiometry (DXA) measurements were made with a Hologic QDR-4500 A X-ray bone densitometer (Hologic France, Massy, France). Bone mineral density (BMD) was determined for the total femur (T-BMD) and for two subregions; the distal metaphyseal zone (M-BMD), rich in trabecular bone; and the diaphyseal zone (D-BMD), rich in cortical bone.(16)
Bone mineral (calcium) content:
After DXA assessment, as mentioned previously, the right femur was ashed at 550°C for 16 h and a sample was diluted with 0.1% lanthanum chloride solution. Calcium content in this solution and of plasma and urine was determined by atomic absorption spectrophotometry (Perkin Elmer 401; Perkin Elmer, Courtaboeuf, France).
Plasma and urine osmolalities were measured by freezing point depression (Fiske OS Osmometer; Fiske Associates, Needam Heights, MA, USA).
Marker of osteoblastic activity:
Plasma osteocalcin (OC) concentrations were measured by homologous radioimmunoassay (RIA) using rat OC standard (Biochemical Technologies kit; Biochemical Technologies, Stoughton, MA, USA). The lowest limit of detection was 60 pg/ml. Intra- and interassay variations were 6.4% and 8.3%, respectively.
Marker of bone resorption:
DPD in urine was measured by a radioimmunoenzymatic assay using a Pyrilinks-D kit (Metra Biosystems, Mountain View, CA, USA). The assay coupling to an alkaline phosphatase reaction, which was read at a wavelength of 405 nm, was performed in association with the picric acid assay for creatinine(17) on 50-μl aliquots of 24-h urine collections. In our laboratory, the lowest limit of detection was 3 nM. The intra- and interassay variations were 6.5% and 8%, respectively. Results were expressed as nanomoles of DPD per millimole of creatinine.(14)
Plasma INS concentration was measured by homologous RIA using rat INS standard (Rat Insulin RIA kit; Linco Research, Inc., St. Charles, MO, USA). The assay is 100% cross-reactive with rat and human INS. The lowest limit of detection was 0.1 ng/ml. Intra- and interassay variations were 2.7% and 3.1%, respectively.
RIA of plasma leptin concentration used an homologous assay incorporating anti-rat leptin antibody and rat leptin as the standard (Rat Leptin RIA kit; Linco Research Inc.). In our experimental conditions, the lowest limit of sensitivity was 0.5 ng/ml, and the intra- and interassay variations were 1.5% and 2.5%, respectively.
Glucose concentration in plasma and urine was measured using the hexokinase/glucose-6-phosphate dehydrogenase method.(18)
Results are presented as means ± SEM. A parametric one-way analysis of variance (ANOVA) was used to test for any differences among the groups. If the result was found significant (p < 0. 05), the Student-Newman-Keuls multiple comparison test was then used to determine the specific differences between group means. If a parametric ANOVA was not feasible (when there were significant differences between the SD groups, tested by Kolmogorov-Smirnov test), a Kruskal-Wallis test followed by the Mann-Whitney Wilcoxon U test was used to compare differences between groups.
The diabetic status induced by STZ treatment alone (STZ group) persisted throughout the study period as evidenced by high water intake, high urine flow rate, low urine osmolality, and hyperglycemia. INS or INS + AMY decreased water intake, urine flow rate, and plasma glucose concentration and increased urine osmolality in diabetic rats; several of these measures were not different from CONT in these treatment groups (Table 1). Glycemia in AMY + INS (25. 8 ± 1. 9 mM) that was higher than in INS (8. 9 ± 1. 4 mM; p < 0.05) did not adversely affect the metabolic benefits that had been gained in the INS group. Diabetic rats treated with AMY alone remained in poor metabolic control as evidenced by levels of polyuria, polydipsia, and hyperglycemia similar to the STZ group (Tables 1 and 2).
Table Table 1.. Intake and Excretion
Table Table 2.. Metabolic Indices
Food intake and body weight
Although daily food intake was greater in diabetic animals (STZ) than in nondiabetic CONT, body weight and daily weight gain were lower. INS treatment increased body weight and weight gain toward normal. In animals receiving AMY (AMY + INS), body weight and weight gain were lower than those receiving INS alone (Table 1).
Bone size and strength
At necropsy, there were no significant differences in femoral length or mean femoral diameter between treatment groups. In diabetic rats treated with AMY, INS, or AMY + INS, femoral failure load was higher than in the diabetic STZ group and was not different from nondiabetic rats (Table 3).
Table Table 3.. Physical Bone Indices
Compared with nondiabetic control rats, STZ-treated rats showed a decrease in T-BMD (−10.3 ± 0.7% vs. CONT; p < 0.05). In contrast, BMD in diabetic rats treated with AMY, INS, or both agents was greater than in the STZ group and was not different from that in nondiabetic CONTs (Table 3). Metaphyseal (trabecular) BMD was lowest in untreated diabetic rats and was nearly completely corrected by AMY, INS, or AMY + INS treatment (Table 3). Similarly, D-BMD (cortical) was lower in untreated diabetic rats than in nondiabetic rats while values in diabetic rats treated with AMY, INS, or both agents did not differ from nondiabetic CONTs (Fig. 1).
Changes in total femoral calcium content paralleled changes in T-BMD (Fig. 1), with a positive linear relationship (r = 0.9871; p < 0.01). Plasma calcium concentration did not differ between groups. The highest daily urinary calcium excretion was observed in untreated diabetic rats. Diabetic rats treated with AMY, INS, or both agents did not differ from nondiabetic CONTs (Table 1).
Biochemical bone markers
Plasma OC concentration, a marker of osteoblastic activity, was depressed substantially in diabetic rats (STZ). In diabetic rats treated with AMY, INS, or both agents, it was increased toward values observed in nondiabetic (CONT) rats (Table 4). Urinary DPD excretion, derived from bone collagen and indicative of bone turnover was higher in CONT than in STZ. Treatment with INS or AMY + INS partially restored DPD excretion (Table 4).
Table Table 4.. Biochemical Markers
The diabetic state induced with STZ in this study was maintained throughout the experimental period (Tables 1 and 2). STZ animals were characterized by extreme hyperglycemia, low circulating INS and leptin concentrations associated with marked weight loss in spite of increased food intake, and polydipsia and polyuria. A low-turnover osteopenia was evidenced in this group by a reduced femoral failure load, decreased BMD (especially in trabecular bone; Fig. 1), reduced OC and DPD markers of formation and resorption, and calciuria (Tables 3 and 4). These data concur with previous reports of decreased bone mass in INS-deficient animals and humans.(19–22)
INS treatment was associated with a partial normalization of body weight gain; a correction of hyperglycemia, food, and water intake; urine production; plasma INS and leptin concentrations (Tables 1 and 2). There was no evidence of osteopenia in these rats; femoral failure load (Table 3) and measures of bone density (Fig. 1) were normal, and plasma OC concentration and urinary DPD excretion were partly normalized (Table 4).
The present results confirm those from previous animal studies showing reduced skeletal mass associated with insulopenic diabetes and partial reversal of bone abnormalities with INS treatment.(22, 23) A role for INS in determining maintenance of bone density is suggested by an association between circulating levels and BMD in normal postmenopausal women.(24) Anabolic effects of INS on bone(25) include effects on osteoblasts, which have INS receptors.(26) INS stimulates proliferation of osteoblast-like cells,(27) increases the steady-state level of a1-procollagen messenger RNA (mRNA) in fetal rat calvaria(28) and collagen synthesis.(29) These effects of INS on osteoblasts could be mediated through insulin-like growth factor 1 (IGF-1) receptors(30); INS displaces IGF-1 from its receptors in rat chondrocytes but with a potency of only 10−4 that of IGF-1 or IGF-2.(31) Thus, a direct anabolic effect of INS on bone in the present study would account for the increased bone density, increased bone strength, and increased plasma OC concentrations in INS-treated diabetic rats (Tables 3 and 4; Fig. 1).
Recent data also support a possible effect of INS on bone resorption. Immunocytochemical studies indicate INS receptor expression by mature mono- and multinucleated murine osteoclast-like cells generated in vitro and in primary neonatal rat and mouse osteoclasts.(32) An antiresorptive effect mediated via direct actions on osteoclasts, for example, could explain why urinary DPD excretion was lower in INS than in CONT (Table 4).
Some aspects of the bone response to INS administration may be accounted for via leptin, a protein hormone secreted by adipocytes in response to repletion, especially that mediated by INS. Leptin administration is reported to modulate body weight,(33) gonadal function,(34) and bone growth.(35) Strong up-regulation of leptin and leptin receptors has been recently shown in pluripotent stem cells that can be differentiated fully along the osteoblastic lineage on appropriate culture conditions, suggesting a direct role of leptin in osteoblast proliferation and differentiation.(36) Osteoclasts differentiate from hemopoietic precursors; in a model employing peripheral blood mononuclear cells, murine leptin (10−9-10−7 M) inhibited osteoclastogenesis in a dose-dependent manner.(37) Thus, leptin expression by marrow adipocytes may contribute to linkage of bone formation and bone resorption; leptin may increase bone mass both by stimulating osteoblastic differentiation and by inhibiting osteoclastogenesis. In the present experiments, leptin concentrations were lower in diabetic rats treated with AMY and AMY + INS (Table 2), in which weight gain was not as fast as with INS alone (Table 1). High plasma leptin concentrations were associated with partial restoration of OC in INS-treated diabetic rats, but OC concentrations were just as high in both groups of animals treated with AMY (Table 4), although their leptin concentrations were somewhat lower (Table 2). These data suggest that leptin was not the major determinant of osteoblast stimulation.
The dose of AMY used in this study (45 μg/kg) is calculated to result in 1000-fold supraphysiological plasma concentrations.(6) However, with such excess and with a t1/2 of 10-17 minutes,(6) it is likely that physiological AMY concentrations would persist for only 1.5-3 h after dosing. Thus, the high doses used here do not preclude a physiological effect of AMY on bone.
In the present experiments, treatment with AMY alone normalized or partly normalized several of the bone indices that were disturbed in the untreated diabetic rats. These indices included bone strength (Table 3), BMDs (Table 3; Fig. 1), calciuria (Table 1), and plasma OC (Table 4). Treatment with AMY only at doses used here did not markedly improve glycemic control, water intake, urine production, or plasma osmolality (Table 1). Therefore, it appears that improvement in indices of osteopenia with AMY was not obligatorily linked to improvements in glycemic control. This improvement was not caused by a change in the acid-base status of the animals (a major contributor of the osteopenic bone observed in diabetic animals) because AMY and AMY agonists do not change acid-base status in rats.(6)
In vivo, AMY causes hypocalcemia in rats and rabbits,(7, 10) consistent with the lowered plasma calcium observed in the present experiment (Table 2). A calciuric effect of AMY has been reported in dogs(38) and rats(39) at high doses. A similar action has been described with calcitonins but not with CGRP in humans.(40) In assessing whether a calciuric effect of AMY was pertinent to its hypocalcemic effect, it was calculated that at doses of 1 g/h and 20 g/h, the shift of calcium from the extracellular space over 60 minutes was 1 mg and 2 mg, respectively. Because only a small fraction of these shifts (6.6% and 2.1%, respectively) could be accounted for by urinary losses, it was concluded that the hypocalcemic effect of AMY was, instead, likely to represent movements from plasma into other compartments (e.g., bone).(39) In vitro, AMY inhibits osteoclastic activity.(7, 10) It also stimulates proliferation of osteoblasts in a dose-dependent manner at concentrations as low as 10 pM. Histomorphometric indices of bone formation reportedly are increased in vivo after local injection.(8, 41, 42) In adult male mice given daily sc injections of AMY (10.5 μg for 4 weeks), histomorphometric indices of bone formation increased 30-100%, whereas resorption indices were reduced by 70%.(43)
A significant effect of AMY on bone observed in the present work contrasts with a previous report in which no effect other than a small increase in bone mass was observed and then only in nondiabetic animals.(44) Differences may lie in the 15-fold higher dose used in the present study (45 μg/kg vs. 3 μg/kg) and a longer period of treatment (40 days vs. 18 days). Although AMY receptors recently have been identified as resulting from the membrane association of calcitonin receptors and RAMPs,(5) the receptors at which AMY may mediate effects in bone are as yet undetermined. Although AMY is an agonist at calcitonin receptors, as are found in osteoclasts, there is evidence for both calcitonin-selective and AMY-selective effects in those cells. Calcitonin is ineffective on osteoblasts, although AMY does stimulate such cells, possibly via a receptor that also is activated by CGRP.(45) AMY may be unique in simultaneously activating several receptor classes in bone and may have synergistic effects with INS coreleased from β-cells on skeletal maintenance.(10)
INS + AMY group
Although INS receptors do exist in osteoblasts,(25, 28) it is likely that AMY acts at IGF-1 receptors in these cells. Experiments in osteoblast-like cell lines with or without IGF-1 receptors showed that IGF-1 stimulation was necessary to permit a proliferative response to AMY, although only low concentrations of each were required.(46) In the present experiments, the bone-sparing effect of AMY and INS in STZ-induced diabetic rats was not additive in that improvements in indices of osteopenia achieved with either agent alone were not further improved with the combination. This may be because several of these measures had been normalized by maximally activating pathways so that incremental benefit was impossible. Other more fundamental reasons could underlie why maximal achievable effect results in less than normalization, for example, the duty-cycle effect, in which because of kinetics (t1/2 about 15 minutes for AMY in rats(6)) the duration of effect of administered peptide is for less than the dosing interval. Thus, although maximally effective at administration, average effect was less than full (Fig. 1; Tables 3 and 4).
In conclusion, the present study indicates that in rats made diabetic by STZ, treatment with AMY, INS, or both peptides improved or normalized many of the indices of osteopenia present in untreated diabetic rats. Improvements associated with AMY treatment appeared unassociated with changes in glycemic control or plasma leptin concentration. AMY effects appear to include actions on both osteoclasts, as evidenced by a hypocalcemic effect and inhibition of trabecular bone loss, and stimulation of osteoblastic activity, as shown by increased plasma OC concentration. AMY (absent or reduced in insulopenic diabetes) or AMY agonists may be potentially useful in the treatment of diabetic osteopenia.
We gratefully acknowledge the expert assistance of M. Balage and D. Dardevet.