JOURNAL OF CELLULAR PHYSIOLOGY 153:6-14 (1992)
Mechanism of Action of Amylin in Bone TATSUYA TAMURA, CHISATO MIYAURA, ICHIRO O W A N , AND TATSUO SUDA* Department of Biochemistry, Schooi of Dentistry, Showa University, I-5-8 1 latanodai, Shinagawd-ku, Tokyo 142 (T.T., C.M., /.0., T.S.), Fuji Cotemha Rcscarch Laboratories, Cbugdi Pharmaceutical Co., Lld., Shiruokd 4 12 IT.T.1, and Department of Orthopedics, Faculty of Medicine, University of h e Ryukyus, Okindwd, 903- I 11.0.1, japan Amylin is a 37 amino acid peptide produced mainly by p-cells of the endocrine pancreas. Human amylin has 4 3 % homology with human calcitonin gene-related peptide (CGRP) and 13% homology with human calcitonin (CT). Amylin and CGRP have been reported to have CT-like hypocalcemic activity in vivo. To investigate the role of amylin in bone, we examined the mechanisms of action of human drnylin, CGRP, and CT in osteoclasts and osteoblasts. Both human amylin and CGRP inhibited Ia,25-dihydroxyvitamin D, [la,25(0H),D,]-induced bone resorption in an organ culture system, and the potencies of the two peptides were similarly -60-fold lower than that of human CT. Using a recently developed procedure for preparing large numbers of osteoclast-like multinucleated cells (MNCs) formed in co-cultures of mouse osteoblasts and bone marrow cells in the presence of la,25(OH),D,, we found that both humdn amylin and CGRP stimulated CAMP production in osteoclast-like MNCs, but only at 60-fold higher concentrations than human CT. Specific binding of ['"I]-human CT to osteoclast-like MNCs was detected (dissociation constant, 3 x lo-" M; binding sites, 3 x l o 7 per cell). To displace the bound ['"l]-human CT from osteoclast-like MNCs, about 170-fold higher concentrations of human amylin and CGRP were required. No specific bindings of ['251]-amylin and 1'251]-CC;Kt' to osteoclast-like MNCs could be detected. Human CGRP stimulated CAMPproduction both in established mouse osteoblast-like cells (KS-4) and in mouse primary osteoblast-like cells. Arnylin was a weak agonist for CAMP produLtion in KS-4 cells. The increment in CAMP production induced by CGKP and amylin was abolished by the addition of human CGRP(8-37), a selective antagonist for CGRP receptors. CT did not stimulate CAMP production in KS-4 cells. Amylin, but not CT, displaced the bound ['L51]-humanCGRP from rat brain membranes. These results indicate that amylin binds not only to CT receptors in osteoclast-like MNCs but also to CGRP receptors in osteoblasts. The relative potencies of these compounds to induce CAMP production was CT > a m y l i n k CGRP in osteoclast-like MNCs and CGRP > amylin >> CT in osteoblast-like cells. o 1992 ~ i l e y - ~ i s Iric. i,
Amylin is a 37 amino acid peptide originally isolated from the islet amyloid of patients with type I1 diabetes as a n islet amyloid polypeptide (IAPP) (Westermark et al., 1986; Cooper et al., 1987, 1988). The gene for amylin has been cloned, and the sequence of the peptide has been reported to be relatively conserved in several species (Sanke et al., 1988; Lefert et al., 1989; Nishi et al., 1990). The sequence of the corresponding cDNA indicated that human amylin has 43% homology with human calcitonin gene-related peptide (CGRP) and 13%homology with human calcitonin (CT). Like CGRP and CT, amylin is amidated a t the carboxy-terminal and has a disulfide bridge between positions 2 and 7 (Nishi et a]., 1990). Recent studies have indicated that amylin is mainly produced by p-cells of the normal endocrine pancreas, and there is no detectable amylin mRNA in portions of the nervous system in which CGRP mRNA is present in abundance (Westermark et al., 1987; Ferrier et al., 1989; Lefert et al., 1989). With regard to the biological activity of amylin, Leighton and Cooper (1988) have shown that amylin $3 1992 WILEY-LISS, INC
inhibits glucose uptake and glycogen synthesis in muscle in vitro. This suggests that amylin plays a role in the pathogenesis of the insulin resistance seen in type I1 diabetes. In addition, Ohsawa et al. (1989) reported that amylin inhibits insulin secretion from rat pancreatic islets. The effects of amylin on carbohydrate metabolism in vivo, however, have not been clarified and the biological significance of' this peptide remains controversial. Zaidi and his co-workers (Datta et al., 1989; Zaidi et al., 1990a) have shown that amylin has serum calcium-lowering activity in vivo and that it inhibits the bone-resorbing activity of isolated osteoclasts in vitro. It has also been reported that CGRP shows CTlike activity in vivo and in vitro (Tippins et al., 1984; D'Souza e t al., 1986; Zaidi et al., 1987). It remains to be determined whether the effects of amylin and CGRP in
Received February 5,1992; accepted April 22, 1992.
*To whom reprint requestsicorrespondence should be addressed.
ACTION OF AMYLIN IN BONE
7
osteoclasts are mediated via CT receptors or amylinl sham (Bucks, UK). These peptides have been radioiodinated by the chloramine-T method and purified by CGRP receptor(s1. Osteoclasts have abundant CT receptors. However, high-performance liquid chromatography (HPLC). the number of osteoclasts isolated from mammalian la,25(OH),D, was supplied by Chugai Pharmaceutical bone tissues is not sufficient for reliably measuring the Co. (Tokyo). Collagen gel solutions (Cellmatrix, Type binding activity of [12sI]-labeled CT to isolated osteo- I-A) were obtained from Nitta Gelatin Co. (Osaka).Bacclasts. Recently, we developed a new method of prepar- terial collagenase was obtained from Wako Pure Cheming large numbers of osteoclast-like multinucleated ical Co. (Osaka). Other chemicals and reagents used in cells (MNCs) formed in co-cultures of mouse osteoblasts this study were of analytical grade. and bone marrow cells in the presence of la,25-dihyBone resorption a s s a y droxyvitamin I), [ la,25(OH)ZD,l (Akatsu et al., 1991). The osteoclast-like MNCs formed satisfied most of the Bone resorption assay was performed using a modifimajor criteria of osteoclasts, such as tartrate-resistant cation of Raisz's organ culture method as previously acid phosphatase (TRAP) activity, acid production, CT described (Raisz, 1965; Sat0 et al., 1986). In short, 16receptors, and bone-resorbing activity (Akatsu et al., day-old pregnant mice were injected subcutaneously 1991). with 25 yCi of 45CaC1, (New England Nuclear ProdIt has been reported that CGRP, but not CT, stimu- ucts, Boston, MA). The radii and ulnae were excised 24 lates cAMP production in osteoblast-like cells (Mich- h later and cultured in minimum essential medium elangeli et al., 1986, 1989). Recently, we cloned a n os- (MEM) for 48 h a t 37°C in a humidified atmosphere of teoblast-like cell line (KS-4) from fetal mouse calvarial 5% CO, in air. After being precultured for 48 h, each cells, which also had the ability to support osteoclast- bone was transferred t o fresh medium containing 0.2% like MNC formation in co-cultures with spleen cells bovine serum albumin (BSA) with or without assay (Yamashita et al., 1990). The KS-4 cells produced a samples and cultured for a n additional 72 h. At the end significant amount of CAMP in response to CGRP of the culture period, 45Ca was counted separately in (Thiebaud et al., 1991). Amylin shares some biological the medium and bone, and 4"Ca release (%) was calcuactivities with CGRP (Nishi et al., 1990), suggesting lated using the following formula: the possibility that amylin exerts a CGRP-like effect on 4"Ca in medium osteoblasts. In this study, we examined the mecha45Ca release (7~) = 45c x 300 a in medium nisms of action of amylin, CGRP and CT in this re+ 45Ca in bone cently developed osteoclast-like MNC preparation and in various osteoblast-like cells including KS-4 cells. We The 4sCa release induced by 2 x M la,25 report here that amylin binds to CT receptors in osteo- (OH),D, in the absence of peptides was expressed a s clast-like MNCs and inhibits bone resorption. Amylin la,25(0H),D3-induced 45Carelease. 45Carelease in the also stimulates cAMP production in osteoblasts via absence of both la,25(OH),D3 and peptides was exCGRP receptors. pressed as spontaneous 45Ca release. The inhibitory effect of the peptides on osteoclastic bone resorption was MATERIALS AND METHODS expressed as the percentage of la,25(0H),D3-induced Animals and drugs 45Ca release, and calculated using the following forPregnant ddy mice, 6- to 9-week-old male ddy mice, mula: and Wistar rats (150-200 g) were obtained from Shi45Ca release induced by zuoka Laboratories Animal Center (Shizuoka, Japan). Synthetic human amylin, human calcitonin gene-re1aj25(OH)ZD, and a peptide (96.70) lated peptide-1 (CGRP), human CGRP(8-37), and huspontaneous 4 5 ~ a man calcitonin (CT) were purchased from Peninsula 45Ca release release (%) Laboratories (Belmont, CA) and the Peptide Institute of la,25(OH),D,- = x 100 (Osaka, Japan). Porcine insulin (40 Uiml) was pur- (%induced la,25(0H),D3-induced chased from Yamanouchi Pharmaceutical Co. (Tokyo, 4"Ca release (%) Japan). Human amylin was dissolved in dimethyl sulfspontaneous 45Ca oxide and diluted with distilled water to prepare lop4 release (%I M solution. Other peptides were dissolved in 1 mM Preparation of osteoclast-like acetic acid containing 0.1% BSA. These peptide solumultinucleated cells . tions were stored at -80°C and used within 3 weeks. [12sIIJ-human CT (2,000 Ciimmol) and [l2"1]-human Osteoclast-like multinucleated cells (MNCs) were CGRP (2,000 Ciimmol) were purchased from by Amer- prepared by the method of Akatsu et al. (1991) using ddy mice. In short, primary osteoblast-like cells (5 x 1O5/dish)obtained from mouse calvaria and nucleated marrow cells (6 x lO'/dish) were co-cultured in Abbreviations a-MEM containing 10% fetal bovine serum (FBS) in the presence of lo-' M la,25(OH),D3. Culture dishes (@10 MEM minimum essential medium cm, Corning 25020, Corning, NY) previously coated BSA bovine serum albumin CGRP calcitonin gene-related peptide with 4 ml of 0.2% collagen gel matrix were used. After lo1,25(OH),D31a,25-dihydroxyvitaminD, culturing for 5-6 days, culture media were discarded CT calci lonin and the dishes were treated with 4 ml of 0.2% bacterial FBS fetal bovine serum TRAP tartrate-resistant acid phosphatase collagenase for 20 min at 37°C in a shaking water bath -
8
TAMURA ET AL
(60 cycleshin). The recovered cell suspensions were gently pipetted 10-15 times with a wide bore plastic pipette and centrifuged at 250 g for 5 min. The cell pellet collected from a 10 cm-dish was resuspended in 1 ml of Tyrode's solution, overlaid on 4 volumes of 35% percoll solution, and centrifuged a t 250 g for 20 min. The top and interface layers were collected, washed with a-MEM, and resuspended in a-MEM containing 10% FBS. An aliquot (1125) of osteoclast-like MNCs preparations from a 10 cm-dish was inoculated into 24-well culture dishes (Corning 25820) and settled for 3 h. The resulting adherent cells were used for the following experiments. The purity of osteoclast-like MNCs in the preparations was 5.6% of total cells and 32% of the total nuclei counted. The contaminating cells consisted of mainly fibroblastic cells and macrophage-like cells.
TR AP-staining Osteoclast-like MNCs were stained for TRAP as described previously (Udagawa et al., 1989). Briefly, cells were washed with a phosphate-buffered saline solution (PBS, pH 7.4) and fixed with 10% formalin in PBS for 10 min. After treatment with ethanol-acetone (50:50, viv) for 1min, the well surface was dried and stained for TRAP, employing to a modification of the original method of Burstone, in the presence of 25 mM sodium tartrate. TRAP-positive multinucleated cells containing three o r more nuclei were counted a s osteoclast-like MNCs.
tured in a-MEM containing 10%FBS for 48 h a t 37°C in a humidified atmosphere of 5% CO, in air. After the cells had been cultured, cAMP production by cells treated with or without peptides was tested a s described above.
Binding assays in osteoclast-like MNCs For binding studies of [12511-humanCT, osteoclastlike MNCs (-1,000 cells) were placed in 24-well culture dishes and incubated for 1h a t 30°C in 0.3 ml of a-MEM containing 0.1% BSA and 0.1% bacitracin with M [12511-humanCT and increasing amounts 6x of unlabeled human CT or other peptides. Nons ecific binding was determined by measuring bound ["'I]-human CT in the presence of 10 M unlabeled human CT. At the ends of the incubations, cells were washed twice in cold PBS, solubilized in 0.2N NaOH, and radioactivity in the preparations was counted in a gammacounter.
'
Preparation of rat brain membranes and binding assays A rat brain membrane fraction was prepared from male Wistar rats by the method of Nakamuta et al. (1986, 1990). The brainstem region including the hypothalamus, midbrain-thalamus, and medulla oblongata and pons was used for the preparation. The brain membrane preparation (0.2 mg protein? was incubated with M) in 0.4 ml of 50 mM [1"51]-human CGRP (4 x Tris HCl buffer (pH 7.4) containing 1%BSA, 0.1% bacitracin, 0.025% lactose, and 2 Uiml of aprotinin with and without various unlabeled peptides. Nonspecific binding was determined by measuring bound [ 1"51]-human CGRP in the presence of lop7 M unlabeled human CGRP. Incubations were carried out a t 4°C for 2 h and stopped by filtration (Whatman GFiB glass filters precoated with polyethylenimine). The radioactivity trapped in the filters was measured with a gammacounter.
Preparation of primary osteoblast-like cells and osteoblast-like cell lines Primary osteoblast-like cells were isolated from 1- to 2-day-old mouse calvaria. A total of 20-30 calvaria was collected and subjected to routine 5 min-sequential digestions, using 0.1% collagenase and 0.28 dispase (Godo Shusei, Tokyo). Osteoblast-like cells isolated from fractions 2 - 4 were combined and cultured in a-MEM supplemented with 10% FBS. A clonal osteoblast-like cell line, KS-4, was established from fetal ddy Statistical analysis mouse calvaria (Yamashita et al., 1990). MC3T3-E1, a n osteoblast-like cell line established from newborn C571 Data are expressed as means * standard error of the EL6 mouse calvaria, was generously provided by Dr. H. mean (SEM). The statistical significance of the differKodama, Ohu University (Fukushima, Japan). ence between the control and the experimental group ROS1712, a rat osteoblastic osteosarcoma cell line, was was determined by Student's t-test. kindly provided by Dr. I. Yamamoto, Shiga Medical RESULTS University (Shiga, Japan). Figure 1 shows the effects of human amylin, CGRP, Measurement of cAMP production by and CT on 45Ca release from prelabeled fetal mouse osteoclast-like MNCs and osteoblast-like cells long bones. The long bones were incubated for 72 h with The osteoclast-like MNCs (-1,500 cells) were placed graded concentrations of either peptide in the presence in 24-well culture dishes, preincubated in 0.5 ml of of 2 x lop8 M la,25(OH),D la,25(OH),D3 had a a-MEM containing 0.1% BSA and 1 mM isobutyl- strong stimulatory effect on "Ca release a s compared methylxanthine for 20 min a t 37°C and treated with to the control culture (treatedicontrol ratio: 5.0 -t 0.6, increasing amounts of the peptides for 15 min. The cells n = 7), and the increase was dose-dependently abolwere washed in ice-cold PBS and extracted with 95% ished by the addition of 10p8-10p6M human amylin. ethanol containing 3 mM HC1. The ethanol extracts Human amylin and CGRP similarly inhibited were dried in a boiling water bath for 5 min and dis- la,25(0H),D3-induced bone resorption, and their posolved in 3 mM HC1. The amount of cAMP in each tencies were -60-fold lower than that of human CT extract was measured with a radioimmunoassay kit for (Fig. 1).Similar inhibitory effects of human amylin and cAMP (Yamasa Shoyu Co., Chiba, Japan). Primary os- CGRP were obtained in PTH- and IL-1-induced bone teoblast-like cells and three clonal osteoblast-like cell resorption (data not shown). To examine the effects of CT, it is essential to prepare lines (KS-4, MC3T3-El and ROS 1712) were placed in 24-well culture dishes (2 x lo4 cells per well) and cul- a fraction containing large quantities of mammalian
ACTION OF AMYLIN IN BONE
9
T
1001
10
Peptide concentrations (M) Fig. 1. I.M'ects of human amylin, CGRP, and CT on la,25(OH),D,induced bone resorption in mouse long bone cultures. Fetal long bones were cultured with increasing amounts of human amylin (*), CGRP (o), or CT ( 0 )in the presence of 2 X l V H M la,25(OH),D3 for 72 h, and
45Ca release in individual incubations was assayed as described in Materials and Methods. Data are expressed a s percentage of '%a release induced by 1fi$5!OH&D3; and means i SEM of three to six separate experiments in quadruplicate cultures.
osteoclasts. We recently developed a new procedure for preparing large numbers of osteoclast-like MNCs formed in co-cultures of mouse primary osteoblasts and bone marrow cells in the presence of lcu,25(OH)zD3 (Akatsu et al., 1991). Using these osteoclast-like MNC preparations, the effects of human amylin, CGRP, and CT on cAMP production were examined. Human CT strongly and dose-dependently stimulated CAMP production by osteoclast-like MNCs (Fig. 2). Human amylin and CGRP also stimulated cAMP production by osteoclast-like MNCs, but only at 60-fold higher concentrations (Fig. 2 ) . Using [12511-salmon CT, we have demonstrated that osteoclast-like MNCs possess large numbers of CT receptors. Scatchard analysis of the saturation binding experiments and the competitive binding experiments using salmon CT indicated that the dissociation constant (Kd) was 9 x 10-lo M and the number of binding sites was 2 x l o 7 per cell (Akatsu et al., 1991). To examine specific bindings of human amylin and CGRP to CT receptors in osteoclastlike MNCs, binding analysis was performed using [lZ511-humanCT as well. Specific binding of T1"Il-human CT to osteoclast-like MNCs was also detected. The Kd value (3 x lop8M) was much greater than that for salmon CT, but the number of binding sites (3 x lo7per cell) was almost the same (Fig. 3A, B). The difference of the Kd values between salmon and human CT was consistent with that of their potencies in biological activities. Human amylin and CGRP displaced [1251J-human CT binding in a concentration-dependent manner (Fig. 3A). The binding affinity of human amylin and CGRP was 170-fold weaker than that of human CT. Displacement of bound [1251]-human CT by human amylin and CGRP was somewhat different from the dose-response curve for cAMP production shown in Figure 2. It has been reported that displacement by cold CT
of bound [1Z51]-CTto human breast cancer cells (T47D) and isolated osteoclasts dose not necessarily correlated to dose-response curves for CAMP production. Larger amounts of cold CT were required to displace bound ['"I]-CT than to induce cAMP production (Nicholson et al., 1986). This would be one of the reasons for the phenomenon. Porcine insulin failed to displace the bound [1251]-human CT in osteoclast-like MNCs (Fig. 3). No specific bindings of [l"I]-rat amylin and [12 11-human CGRP to osteoclast-like MNCs could be detected a t concentrations up to 2 X M (data not shown). Figure 4 compares the stimulation of cAMP production by various osteoblast-like cells in response to human amylin, CGRP and CT. As reported previously (Thiebaud et al., 19911, KS-4 cells produced substantial amounts of CAMPin response to CGRP. Mouse primar osteoblasts also generated cAMP in response to 10- M human CGRP, but the response was smaller than that of KS-4 cells. Human amylin also stimulated cAMP production by both KS-4 cells and mouse primary osteoblasts, but only a t 100-fold higher concentrations (Fig. 4). MC3T3-El and ROS 1712 cells produced no CAMP in response to human CGRP and amylin (Fig. 4).No osteoblast-like cells tested produced CAMP in response t o human CT. Chiba et al. (1989) have shown that human CGRP(8-371, a fragment of human CGRP, acts as a specific CGRP receptor antagonist. Simultaneous addition of human CGRP(8-37) dose-dependently abolished cAMP production by KS-4 cells in response to not only human CGRP but also human amylin (Table 1).Human CGRP(8-37) per se had no effect on CAMPproduction by KS-4 cells. To determine the binding affinities of a m lin and CT to CGRP receptors, the binding affinity of ['"I]-human CGRP was examined using rat brain membranes in
z
10
TAMURA E T AL
J I
0
lo-"
10-l0
lo-'
10.'
10'~
Peptide concentrations (M) Fig. 2. Effects of human amylin, CGRP and CT on cAMP production by mouse osteoclast-like MNCs. Osteoclast-like MNCs formed in co-culturcs (1583 -C 65 cells per well) were incubated with increasing amounts of human amylin (-1, CGRP (c)or CT (r)for 15 min, and the amount of cAMP produced in individual incubations was determined as described in Materials and Methods. Points without x r t i c a l bars represent means with smaller SEM than the size of each symbol.
which CGRP receptors are present in abundance. The addition of graded concentrations of unlabeled human CGRP displaced the binding of [1251]-human CGRP from brain membranes in a dose-dependent manner i . 5). Human amylin also displaced the binding of 12* Ikhuman CGRP, but the potency was much weaker than that of human CGRP (Fig. 5). CT failed to displace I1"1]-human CGRP (Fig. 5).
ing that amylin decreases serum calcium levels by inhibiting osteoclastic bone resorption. Studying the mechanisms of action of amylin and CGRP in osteoclasts requires preparation of large numbers of mammalian osteoclasts. We have developed a new method for preparing large quantities of mouse osteoclast-like MNCs formed in co-cultures of osteoblasts and bone marrow cells in the presence of 1a,25(OH),D3 (Akatsu et al., 1991).The binding analyDISCUSSION sis by [l"I]-salmon CT showed that these osteoclastIt has been reported that both amylin and CGRP like MNCs possessed specific CT receptors with the Kd have hypocalcemic actions in vivo (Datta et al., 1989; value of 9 x lo-'' M and the binding sites of 2 x lo7 Tippins et al., 1984).In the present study, we examined per cell, and a n excess amount of unlabeled human CT the mechanisms underlying the hypocalcemic actions displaced bound [ 1L51]-salmonCT from osteoclast-like of these two peptides. The potencies of human amylin MNCs (Akatsu et al., 1991). In the present study, [ 1251]and CGRP were -100-fold lower than that of human human CT also specifically bound to CT receptors CT, when serum calcium was measured 1 h after intra- present in osteoclast-like MNCs. The number of bindvenous injection of these peptides into SD rats (data not ing sites ( 3 x lo7 per cell) for human CT was similar to shown). In vitro, CGRP inhibited the osteoclastic bone that for salmon CT, but the Kd value ( 3 x M) was resorption induced by PTH, PGE,, and la,25(OH),D3 in much greater than that for salmon CT (Fig. 3). The rat long bone cultures (D'Souza et al., 1986) and de- difference of the Kd values is consistent with the fact creased the bone-resorbed area excavated by isolated that the biological potency of human CT is -50- to rat osteoclasts (Zaidi et al., 1987). These effects of 100-fold lower than that of salmon CT (Nakamuta CGRP were much weaker than those of CT. Recently, et al., 1990; Akatsu et al., 1991). Both human amylin Zaidi et al. (1990a) reported that amylin also inhibits and CGRP were capable of displacing [12511-humanCT bone resorption by isolated rat osteoclasts. The present from its receptors in the osteoclast-like MNCs (Fi . 31, study confirmed that amylin inhibits the osteoclastic but no specific bindings of [l2'I1-rat amylin and ["'I]bone resorption induced by la,25(OH)2D, in 45Ca-la- human CGRP to osteoclast-like MNCs could be debeled mouse long bones. The potency of human amylin tected. These results suggest that the effects of amylin in suppressing bone resorption was similar to that of hu- and CGRP on bone resorption occur via CT receptors man CGRP, which was -60-fold lower than that of hu- present in osteoclasts. It has been well established that CT stimulates man CT (Fig. 1).The in vitro inhibitory effect of amylin on osteoclastic bone resorption had a significant corre- cAMP production by osteoclasts. The cAMP acts a s a lation with its hypocalcemic activity in vivo, suggest- possible second messenger in mediating the inhibitory
iF 8
11
ACTION OF AMYLIN IN BONE
TABLE 1 EKects of human CGRP(8-37) on human amylin- and CGRP-induccd CAMPproduction by KS-4 cells' CGRP(8-371 (Mi 0
Amount of cAMP produced (pmol/well) Human CGRP Human amylin (10 7Mi None (5 x 10 ' M i 4.9 I 0 . 3 1
(1.0) 10-8
10 10
Peptide concentrations (M)
-
5.8 & 0.34 (1.2) 5.4 -C 0.08 (1.1)
10.7 i 0.69* 12.2) 6.0 ? 0.40'* (1.2) 6.1 2 0.142% (1.3) 4.3 0.31"s: (0.9)
*
36.3
L
0.51*
(7.4)
30.6 ? 0.60*
19.0
*(6.2) 1,21*
x9:ii
(3.9) 9.3 -C 0.13*,*"" (1.9)
M) in the M) or CGRP 'KS-4 cells were incubated with human amylin I5 x prcsence or absence of incrcasing amounts of human CGRP(A-37) for 15 min. The amount of cAMP produced in individual incubations was then determined as described in Materials and Methods. Data are expressed as mean i SEM of triplicate cultures. The valuer- in parentheses indicate the mean treated t o control ratio in triplicate cultures "Significantly different from control, P
Mechanism of Action of Amylin in Bone TATSUYA TAMURA, CHISATO MIYAURA, ICHIRO O W A N , AND TATSUO SUDA* Department of Biochemistry, Schooi of Dentistry, Showa University, I-5-8 1 latanodai, Shinagawd-ku, Tokyo 142 (T.T., C.M., /.0., T.S.), Fuji Cotemha Rcscarch Laboratories, Cbugdi Pharmaceutical Co., Lld., Shiruokd 4 12 IT.T.1, and Department of Orthopedics, Faculty of Medicine, University of h e Ryukyus, Okindwd, 903- I 11.0.1, japan Amylin is a 37 amino acid peptide produced mainly by p-cells of the endocrine pancreas. Human amylin has 4 3 % homology with human calcitonin gene-related peptide (CGRP) and 13% homology with human calcitonin (CT). Amylin and CGRP have been reported to have CT-like hypocalcemic activity in vivo. To investigate the role of amylin in bone, we examined the mechanisms of action of human drnylin, CGRP, and CT in osteoclasts and osteoblasts. Both human amylin and CGRP inhibited Ia,25-dihydroxyvitamin D, [la,25(0H),D,]-induced bone resorption in an organ culture system, and the potencies of the two peptides were similarly -60-fold lower than that of human CT. Using a recently developed procedure for preparing large numbers of osteoclast-like multinucleated cells (MNCs) formed in co-cultures of mouse osteoblasts and bone marrow cells in the presence of la,25(OH),D,, we found that both humdn amylin and CGRP stimulated CAMP production in osteoclast-like MNCs, but only at 60-fold higher concentrations than human CT. Specific binding of ['"I]-human CT to osteoclast-like MNCs was detected (dissociation constant, 3 x lo-" M; binding sites, 3 x l o 7 per cell). To displace the bound ['"l]-human CT from osteoclast-like MNCs, about 170-fold higher concentrations of human amylin and CGRP were required. No specific bindings of ['251]-amylin and 1'251]-CC;Kt' to osteoclast-like MNCs could be detected. Human CGRP stimulated CAMPproduction both in established mouse osteoblast-like cells (KS-4) and in mouse primary osteoblast-like cells. Arnylin was a weak agonist for CAMP produLtion in KS-4 cells. The increment in CAMP production induced by CGKP and amylin was abolished by the addition of human CGRP(8-37), a selective antagonist for CGRP receptors. CT did not stimulate CAMP production in KS-4 cells. Amylin, but not CT, displaced the bound ['L51]-humanCGRP from rat brain membranes. These results indicate that amylin binds not only to CT receptors in osteoclast-like MNCs but also to CGRP receptors in osteoblasts. The relative potencies of these compounds to induce CAMP production was CT > a m y l i n k CGRP in osteoclast-like MNCs and CGRP > amylin >> CT in osteoblast-like cells. o 1992 ~ i l e y - ~ i s Iric. i,
Amylin is a 37 amino acid peptide originally isolated from the islet amyloid of patients with type I1 diabetes as a n islet amyloid polypeptide (IAPP) (Westermark et al., 1986; Cooper et al., 1987, 1988). The gene for amylin has been cloned, and the sequence of the peptide has been reported to be relatively conserved in several species (Sanke et al., 1988; Lefert et al., 1989; Nishi et al., 1990). The sequence of the corresponding cDNA indicated that human amylin has 43% homology with human calcitonin gene-related peptide (CGRP) and 13%homology with human calcitonin (CT). Like CGRP and CT, amylin is amidated a t the carboxy-terminal and has a disulfide bridge between positions 2 and 7 (Nishi et a]., 1990). Recent studies have indicated that amylin is mainly produced by p-cells of the normal endocrine pancreas, and there is no detectable amylin mRNA in portions of the nervous system in which CGRP mRNA is present in abundance (Westermark et al., 1987; Ferrier et al., 1989; Lefert et al., 1989). With regard to the biological activity of amylin, Leighton and Cooper (1988) have shown that amylin $3 1992 WILEY-LISS, INC
inhibits glucose uptake and glycogen synthesis in muscle in vitro. This suggests that amylin plays a role in the pathogenesis of the insulin resistance seen in type I1 diabetes. In addition, Ohsawa et al. (1989) reported that amylin inhibits insulin secretion from rat pancreatic islets. The effects of amylin on carbohydrate metabolism in vivo, however, have not been clarified and the biological significance of' this peptide remains controversial. Zaidi and his co-workers (Datta et al., 1989; Zaidi et al., 1990a) have shown that amylin has serum calcium-lowering activity in vivo and that it inhibits the bone-resorbing activity of isolated osteoclasts in vitro. It has also been reported that CGRP shows CTlike activity in vivo and in vitro (Tippins et al., 1984; D'Souza e t al., 1986; Zaidi et al., 1987). It remains to be determined whether the effects of amylin and CGRP in
Received February 5,1992; accepted April 22, 1992.
*To whom reprint requestsicorrespondence should be addressed.
ACTION OF AMYLIN IN BONE
7
osteoclasts are mediated via CT receptors or amylinl sham (Bucks, UK). These peptides have been radioiodinated by the chloramine-T method and purified by CGRP receptor(s1. Osteoclasts have abundant CT receptors. However, high-performance liquid chromatography (HPLC). the number of osteoclasts isolated from mammalian la,25(OH),D, was supplied by Chugai Pharmaceutical bone tissues is not sufficient for reliably measuring the Co. (Tokyo). Collagen gel solutions (Cellmatrix, Type binding activity of [12sI]-labeled CT to isolated osteo- I-A) were obtained from Nitta Gelatin Co. (Osaka).Bacclasts. Recently, we developed a new method of prepar- terial collagenase was obtained from Wako Pure Cheming large numbers of osteoclast-like multinucleated ical Co. (Osaka). Other chemicals and reagents used in cells (MNCs) formed in co-cultures of mouse osteoblasts this study were of analytical grade. and bone marrow cells in the presence of la,25-dihyBone resorption a s s a y droxyvitamin I), [ la,25(OH)ZD,l (Akatsu et al., 1991). The osteoclast-like MNCs formed satisfied most of the Bone resorption assay was performed using a modifimajor criteria of osteoclasts, such as tartrate-resistant cation of Raisz's organ culture method as previously acid phosphatase (TRAP) activity, acid production, CT described (Raisz, 1965; Sat0 et al., 1986). In short, 16receptors, and bone-resorbing activity (Akatsu et al., day-old pregnant mice were injected subcutaneously 1991). with 25 yCi of 45CaC1, (New England Nuclear ProdIt has been reported that CGRP, but not CT, stimu- ucts, Boston, MA). The radii and ulnae were excised 24 lates cAMP production in osteoblast-like cells (Mich- h later and cultured in minimum essential medium elangeli et al., 1986, 1989). Recently, we cloned a n os- (MEM) for 48 h a t 37°C in a humidified atmosphere of teoblast-like cell line (KS-4) from fetal mouse calvarial 5% CO, in air. After being precultured for 48 h, each cells, which also had the ability to support osteoclast- bone was transferred t o fresh medium containing 0.2% like MNC formation in co-cultures with spleen cells bovine serum albumin (BSA) with or without assay (Yamashita et al., 1990). The KS-4 cells produced a samples and cultured for a n additional 72 h. At the end significant amount of CAMP in response to CGRP of the culture period, 45Ca was counted separately in (Thiebaud et al., 1991). Amylin shares some biological the medium and bone, and 4"Ca release (%) was calcuactivities with CGRP (Nishi et al., 1990), suggesting lated using the following formula: the possibility that amylin exerts a CGRP-like effect on 4"Ca in medium osteoblasts. In this study, we examined the mecha45Ca release (7~) = 45c x 300 a in medium nisms of action of amylin, CGRP and CT in this re+ 45Ca in bone cently developed osteoclast-like MNC preparation and in various osteoblast-like cells including KS-4 cells. We The 4sCa release induced by 2 x M la,25 report here that amylin binds to CT receptors in osteo- (OH),D, in the absence of peptides was expressed a s clast-like MNCs and inhibits bone resorption. Amylin la,25(0H),D3-induced 45Carelease. 45Carelease in the also stimulates cAMP production in osteoblasts via absence of both la,25(OH),D3 and peptides was exCGRP receptors. pressed as spontaneous 45Ca release. The inhibitory effect of the peptides on osteoclastic bone resorption was MATERIALS AND METHODS expressed as the percentage of la,25(0H),D3-induced Animals and drugs 45Ca release, and calculated using the following forPregnant ddy mice, 6- to 9-week-old male ddy mice, mula: and Wistar rats (150-200 g) were obtained from Shi45Ca release induced by zuoka Laboratories Animal Center (Shizuoka, Japan). Synthetic human amylin, human calcitonin gene-re1aj25(OH)ZD, and a peptide (96.70) lated peptide-1 (CGRP), human CGRP(8-37), and huspontaneous 4 5 ~ a man calcitonin (CT) were purchased from Peninsula 45Ca release release (%) Laboratories (Belmont, CA) and the Peptide Institute of la,25(OH),D,- = x 100 (Osaka, Japan). Porcine insulin (40 Uiml) was pur- (%induced la,25(0H),D3-induced chased from Yamanouchi Pharmaceutical Co. (Tokyo, 4"Ca release (%) Japan). Human amylin was dissolved in dimethyl sulfspontaneous 45Ca oxide and diluted with distilled water to prepare lop4 release (%I M solution. Other peptides were dissolved in 1 mM Preparation of osteoclast-like acetic acid containing 0.1% BSA. These peptide solumultinucleated cells . tions were stored at -80°C and used within 3 weeks. [12sIIJ-human CT (2,000 Ciimmol) and [l2"1]-human Osteoclast-like multinucleated cells (MNCs) were CGRP (2,000 Ciimmol) were purchased from by Amer- prepared by the method of Akatsu et al. (1991) using ddy mice. In short, primary osteoblast-like cells (5 x 1O5/dish)obtained from mouse calvaria and nucleated marrow cells (6 x lO'/dish) were co-cultured in Abbreviations a-MEM containing 10% fetal bovine serum (FBS) in the presence of lo-' M la,25(OH),D3. Culture dishes (@10 MEM minimum essential medium cm, Corning 25020, Corning, NY) previously coated BSA bovine serum albumin CGRP calcitonin gene-related peptide with 4 ml of 0.2% collagen gel matrix were used. After lo1,25(OH),D31a,25-dihydroxyvitaminD, culturing for 5-6 days, culture media were discarded CT calci lonin and the dishes were treated with 4 ml of 0.2% bacterial FBS fetal bovine serum TRAP tartrate-resistant acid phosphatase collagenase for 20 min at 37°C in a shaking water bath -
8
TAMURA ET AL
(60 cycleshin). The recovered cell suspensions were gently pipetted 10-15 times with a wide bore plastic pipette and centrifuged at 250 g for 5 min. The cell pellet collected from a 10 cm-dish was resuspended in 1 ml of Tyrode's solution, overlaid on 4 volumes of 35% percoll solution, and centrifuged a t 250 g for 20 min. The top and interface layers were collected, washed with a-MEM, and resuspended in a-MEM containing 10% FBS. An aliquot (1125) of osteoclast-like MNCs preparations from a 10 cm-dish was inoculated into 24-well culture dishes (Corning 25820) and settled for 3 h. The resulting adherent cells were used for the following experiments. The purity of osteoclast-like MNCs in the preparations was 5.6% of total cells and 32% of the total nuclei counted. The contaminating cells consisted of mainly fibroblastic cells and macrophage-like cells.
TR AP-staining Osteoclast-like MNCs were stained for TRAP as described previously (Udagawa et al., 1989). Briefly, cells were washed with a phosphate-buffered saline solution (PBS, pH 7.4) and fixed with 10% formalin in PBS for 10 min. After treatment with ethanol-acetone (50:50, viv) for 1min, the well surface was dried and stained for TRAP, employing to a modification of the original method of Burstone, in the presence of 25 mM sodium tartrate. TRAP-positive multinucleated cells containing three o r more nuclei were counted a s osteoclast-like MNCs.
tured in a-MEM containing 10%FBS for 48 h a t 37°C in a humidified atmosphere of 5% CO, in air. After the cells had been cultured, cAMP production by cells treated with or without peptides was tested a s described above.
Binding assays in osteoclast-like MNCs For binding studies of [12511-humanCT, osteoclastlike MNCs (-1,000 cells) were placed in 24-well culture dishes and incubated for 1h a t 30°C in 0.3 ml of a-MEM containing 0.1% BSA and 0.1% bacitracin with M [12511-humanCT and increasing amounts 6x of unlabeled human CT or other peptides. Nons ecific binding was determined by measuring bound ["'I]-human CT in the presence of 10 M unlabeled human CT. At the ends of the incubations, cells were washed twice in cold PBS, solubilized in 0.2N NaOH, and radioactivity in the preparations was counted in a gammacounter.
'
Preparation of rat brain membranes and binding assays A rat brain membrane fraction was prepared from male Wistar rats by the method of Nakamuta et al. (1986, 1990). The brainstem region including the hypothalamus, midbrain-thalamus, and medulla oblongata and pons was used for the preparation. The brain membrane preparation (0.2 mg protein? was incubated with M) in 0.4 ml of 50 mM [1"51]-human CGRP (4 x Tris HCl buffer (pH 7.4) containing 1%BSA, 0.1% bacitracin, 0.025% lactose, and 2 Uiml of aprotinin with and without various unlabeled peptides. Nonspecific binding was determined by measuring bound [ 1"51]-human CGRP in the presence of lop7 M unlabeled human CGRP. Incubations were carried out a t 4°C for 2 h and stopped by filtration (Whatman GFiB glass filters precoated with polyethylenimine). The radioactivity trapped in the filters was measured with a gammacounter.
Preparation of primary osteoblast-like cells and osteoblast-like cell lines Primary osteoblast-like cells were isolated from 1- to 2-day-old mouse calvaria. A total of 20-30 calvaria was collected and subjected to routine 5 min-sequential digestions, using 0.1% collagenase and 0.28 dispase (Godo Shusei, Tokyo). Osteoblast-like cells isolated from fractions 2 - 4 were combined and cultured in a-MEM supplemented with 10% FBS. A clonal osteoblast-like cell line, KS-4, was established from fetal ddy Statistical analysis mouse calvaria (Yamashita et al., 1990). MC3T3-E1, a n osteoblast-like cell line established from newborn C571 Data are expressed as means * standard error of the EL6 mouse calvaria, was generously provided by Dr. H. mean (SEM). The statistical significance of the differKodama, Ohu University (Fukushima, Japan). ence between the control and the experimental group ROS1712, a rat osteoblastic osteosarcoma cell line, was was determined by Student's t-test. kindly provided by Dr. I. Yamamoto, Shiga Medical RESULTS University (Shiga, Japan). Figure 1 shows the effects of human amylin, CGRP, Measurement of cAMP production by and CT on 45Ca release from prelabeled fetal mouse osteoclast-like MNCs and osteoblast-like cells long bones. The long bones were incubated for 72 h with The osteoclast-like MNCs (-1,500 cells) were placed graded concentrations of either peptide in the presence in 24-well culture dishes, preincubated in 0.5 ml of of 2 x lop8 M la,25(OH),D la,25(OH),D3 had a a-MEM containing 0.1% BSA and 1 mM isobutyl- strong stimulatory effect on "Ca release a s compared methylxanthine for 20 min a t 37°C and treated with to the control culture (treatedicontrol ratio: 5.0 -t 0.6, increasing amounts of the peptides for 15 min. The cells n = 7), and the increase was dose-dependently abolwere washed in ice-cold PBS and extracted with 95% ished by the addition of 10p8-10p6M human amylin. ethanol containing 3 mM HC1. The ethanol extracts Human amylin and CGRP similarly inhibited were dried in a boiling water bath for 5 min and dis- la,25(0H),D3-induced bone resorption, and their posolved in 3 mM HC1. The amount of cAMP in each tencies were -60-fold lower than that of human CT extract was measured with a radioimmunoassay kit for (Fig. 1).Similar inhibitory effects of human amylin and cAMP (Yamasa Shoyu Co., Chiba, Japan). Primary os- CGRP were obtained in PTH- and IL-1-induced bone teoblast-like cells and three clonal osteoblast-like cell resorption (data not shown). To examine the effects of CT, it is essential to prepare lines (KS-4, MC3T3-El and ROS 1712) were placed in 24-well culture dishes (2 x lo4 cells per well) and cul- a fraction containing large quantities of mammalian
ACTION OF AMYLIN IN BONE
9
T
1001
10
Peptide concentrations (M) Fig. 1. I.M'ects of human amylin, CGRP, and CT on la,25(OH),D,induced bone resorption in mouse long bone cultures. Fetal long bones were cultured with increasing amounts of human amylin (*), CGRP (o), or CT ( 0 )in the presence of 2 X l V H M la,25(OH),D3 for 72 h, and
45Ca release in individual incubations was assayed as described in Materials and Methods. Data are expressed a s percentage of '%a release induced by 1fi$5!OH&D3; and means i SEM of three to six separate experiments in quadruplicate cultures.
osteoclasts. We recently developed a new procedure for preparing large numbers of osteoclast-like MNCs formed in co-cultures of mouse primary osteoblasts and bone marrow cells in the presence of lcu,25(OH)zD3 (Akatsu et al., 1991). Using these osteoclast-like MNC preparations, the effects of human amylin, CGRP, and CT on cAMP production were examined. Human CT strongly and dose-dependently stimulated CAMP production by osteoclast-like MNCs (Fig. 2). Human amylin and CGRP also stimulated cAMP production by osteoclast-like MNCs, but only at 60-fold higher concentrations (Fig. 2 ) . Using [12511-salmon CT, we have demonstrated that osteoclast-like MNCs possess large numbers of CT receptors. Scatchard analysis of the saturation binding experiments and the competitive binding experiments using salmon CT indicated that the dissociation constant (Kd) was 9 x 10-lo M and the number of binding sites was 2 x l o 7 per cell (Akatsu et al., 1991). To examine specific bindings of human amylin and CGRP to CT receptors in osteoclastlike MNCs, binding analysis was performed using [lZ511-humanCT as well. Specific binding of T1"Il-human CT to osteoclast-like MNCs was also detected. The Kd value (3 x lop8M) was much greater than that for salmon CT, but the number of binding sites (3 x lo7per cell) was almost the same (Fig. 3A, B). The difference of the Kd values between salmon and human CT was consistent with that of their potencies in biological activities. Human amylin and CGRP displaced [1251J-human CT binding in a concentration-dependent manner (Fig. 3A). The binding affinity of human amylin and CGRP was 170-fold weaker than that of human CT. Displacement of bound [1251]-human CT by human amylin and CGRP was somewhat different from the dose-response curve for cAMP production shown in Figure 2. It has been reported that displacement by cold CT
of bound [1Z51]-CTto human breast cancer cells (T47D) and isolated osteoclasts dose not necessarily correlated to dose-response curves for CAMP production. Larger amounts of cold CT were required to displace bound ['"I]-CT than to induce cAMP production (Nicholson et al., 1986). This would be one of the reasons for the phenomenon. Porcine insulin failed to displace the bound [1251]-human CT in osteoclast-like MNCs (Fig. 3). No specific bindings of [l"I]-rat amylin and [12 11-human CGRP to osteoclast-like MNCs could be detected a t concentrations up to 2 X M (data not shown). Figure 4 compares the stimulation of cAMP production by various osteoblast-like cells in response to human amylin, CGRP and CT. As reported previously (Thiebaud et al., 19911, KS-4 cells produced substantial amounts of CAMPin response to CGRP. Mouse primar osteoblasts also generated cAMP in response to 10- M human CGRP, but the response was smaller than that of KS-4 cells. Human amylin also stimulated cAMP production by both KS-4 cells and mouse primary osteoblasts, but only a t 100-fold higher concentrations (Fig. 4). MC3T3-El and ROS 1712 cells produced no CAMP in response to human CGRP and amylin (Fig. 4).No osteoblast-like cells tested produced CAMP in response t o human CT. Chiba et al. (1989) have shown that human CGRP(8-371, a fragment of human CGRP, acts as a specific CGRP receptor antagonist. Simultaneous addition of human CGRP(8-37) dose-dependently abolished cAMP production by KS-4 cells in response to not only human CGRP but also human amylin (Table 1).Human CGRP(8-37) per se had no effect on CAMPproduction by KS-4 cells. To determine the binding affinities of a m lin and CT to CGRP receptors, the binding affinity of ['"I]-human CGRP was examined using rat brain membranes in
z
10
TAMURA E T AL
J I
0
lo-"
10-l0
lo-'
10.'
10'~
Peptide concentrations (M) Fig. 2. Effects of human amylin, CGRP and CT on cAMP production by mouse osteoclast-like MNCs. Osteoclast-like MNCs formed in co-culturcs (1583 -C 65 cells per well) were incubated with increasing amounts of human amylin (-1, CGRP (c)or CT (r)for 15 min, and the amount of cAMP produced in individual incubations was determined as described in Materials and Methods. Points without x r t i c a l bars represent means with smaller SEM than the size of each symbol.
which CGRP receptors are present in abundance. The addition of graded concentrations of unlabeled human CGRP displaced the binding of [1251]-human CGRP from brain membranes in a dose-dependent manner i . 5). Human amylin also displaced the binding of 12* Ikhuman CGRP, but the potency was much weaker than that of human CGRP (Fig. 5). CT failed to displace I1"1]-human CGRP (Fig. 5).
ing that amylin decreases serum calcium levels by inhibiting osteoclastic bone resorption. Studying the mechanisms of action of amylin and CGRP in osteoclasts requires preparation of large numbers of mammalian osteoclasts. We have developed a new method for preparing large quantities of mouse osteoclast-like MNCs formed in co-cultures of osteoblasts and bone marrow cells in the presence of 1a,25(OH),D3 (Akatsu et al., 1991).The binding analyDISCUSSION sis by [l"I]-salmon CT showed that these osteoclastIt has been reported that both amylin and CGRP like MNCs possessed specific CT receptors with the Kd have hypocalcemic actions in vivo (Datta et al., 1989; value of 9 x lo-'' M and the binding sites of 2 x lo7 Tippins et al., 1984).In the present study, we examined per cell, and a n excess amount of unlabeled human CT the mechanisms underlying the hypocalcemic actions displaced bound [ 1L51]-salmonCT from osteoclast-like of these two peptides. The potencies of human amylin MNCs (Akatsu et al., 1991). In the present study, [ 1251]and CGRP were -100-fold lower than that of human human CT also specifically bound to CT receptors CT, when serum calcium was measured 1 h after intra- present in osteoclast-like MNCs. The number of bindvenous injection of these peptides into SD rats (data not ing sites ( 3 x lo7 per cell) for human CT was similar to shown). In vitro, CGRP inhibited the osteoclastic bone that for salmon CT, but the Kd value ( 3 x M) was resorption induced by PTH, PGE,, and la,25(OH),D3 in much greater than that for salmon CT (Fig. 3). The rat long bone cultures (D'Souza et al., 1986) and de- difference of the Kd values is consistent with the fact creased the bone-resorbed area excavated by isolated that the biological potency of human CT is -50- to rat osteoclasts (Zaidi et al., 1987). These effects of 100-fold lower than that of salmon CT (Nakamuta CGRP were much weaker than those of CT. Recently, et al., 1990; Akatsu et al., 1991). Both human amylin Zaidi et al. (1990a) reported that amylin also inhibits and CGRP were capable of displacing [12511-humanCT bone resorption by isolated rat osteoclasts. The present from its receptors in the osteoclast-like MNCs (Fi . 31, study confirmed that amylin inhibits the osteoclastic but no specific bindings of [l2'I1-rat amylin and ["'I]bone resorption induced by la,25(OH)2D, in 45Ca-la- human CGRP to osteoclast-like MNCs could be debeled mouse long bones. The potency of human amylin tected. These results suggest that the effects of amylin in suppressing bone resorption was similar to that of hu- and CGRP on bone resorption occur via CT receptors man CGRP, which was -60-fold lower than that of hu- present in osteoclasts. It has been well established that CT stimulates man CT (Fig. 1).The in vitro inhibitory effect of amylin on osteoclastic bone resorption had a significant corre- cAMP production by osteoclasts. The cAMP acts a s a lation with its hypocalcemic activity in vivo, suggest- possible second messenger in mediating the inhibitory
iF 8
11
ACTION OF AMYLIN IN BONE
TABLE 1 EKects of human CGRP(8-37) on human amylin- and CGRP-induccd CAMPproduction by KS-4 cells' CGRP(8-371 (Mi 0
Amount of cAMP produced (pmol/well) Human CGRP Human amylin (10 7Mi None (5 x 10 ' M i 4.9 I 0 . 3 1
(1.0) 10-8
10 10
Peptide concentrations (M)
-
5.8 & 0.34 (1.2) 5.4 -C 0.08 (1.1)
10.7 i 0.69* 12.2) 6.0 ? 0.40'* (1.2) 6.1 2 0.142% (1.3) 4.3 0.31"s: (0.9)
*
36.3
L
0.51*
(7.4)
30.6 ? 0.60*
19.0
*(6.2) 1,21*
x9:ii
(3.9) 9.3 -C 0.13*,*"" (1.9)
M) in the M) or CGRP 'KS-4 cells were incubated with human amylin I5 x prcsence or absence of incrcasing amounts of human CGRP(A-37) for 15 min. The amount of cAMP produced in individual incubations was then determined as described in Materials and Methods. Data are expressed as mean i SEM of triplicate cultures. The valuer- in parentheses indicate the mean treated t o control ratio in triplicate cultures "Significantly different from control, P
