Interleukin-4 Inhibits Bone Resorption and Acutely Increases Cytosolic ...

Val. 269, No. 19, Issue of May 13, pp. 13817-13824,1994 Printed i n U.S.A.

THE JOURNAL OF BIOLWXCAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Interleukin-4 Inhibits Bone Resorption and Acutely Increases Cytosolic Ca2+in Murine Osteoclasts* (Received for publication, November 22, 1993, and in revised form, February 2, 1994)

Cinzia BizzarriSQ,Atsushi Shioin, Steven L. Teitelbaumn, Jun-ichi Oharall, Vijay A. Harwalkar"", Jeanne M. Erdmannn, David L. Lacen, and RobertoCivitelli""S8 From the **Division of Endocrinology and Bone and Mineral Diseases, the Wepartment of Pathology, the Jewish Hospital at Washington University Medical Center, St. Louis, Missouri 63110, the lllivision of Gastroenterology, University of Colorado Health Sciences Center, Denver, Colorado 80262, and SConsorzio Biolaq. L'Aquila, I-67100, Italy

Interleukin-4 (IL-4) is an immune cytokine recently 20-kDa glycoprotein is a well-recognized growth and/or differshown to inhibit bone resorption. To determine whether entiation factor for a wide variety of cells of hematopoietic of its IL-4 directly acts on osteoclasts, we have analyzed its lineage. IL-4 was originally characterized on the basis effect on cytosoliccalcium concentration [Ca2+l,and capacity to co-stimulateB cell growth with anti-IgM antibodies bone resorptive function of murine osteoclastic cells (3,4), but i t i snow knownthat this cytokine can affect antibody generated from bone marrow/stromal cell co-cultures. isotype and antibody production, T cell activation, mast cell IL-4 exposure induced an immediate and sustained in- growth, monocyte activation, and hematopoiesis (seeRef. 5 for crease in [Ca2+], that remained elevated for at least 10 review). min. ThisIL-4 effect wasdose-dependent,with the maxiThe potential for IL-4 to affect bone resorption was initially mal effect (209 * 15%of baseline, n = 16) at 200 unitdm1 documented by Watanabe et al.(11, who demonstrated that IL-4 and an apparent ED,,5 of 60 unitdml. The IL-4-induced inhibited in vitro bone resorption induced by a wide variety of [Ca2+],rise required extracellular Ca2+ influx,since the agents, suchas parathyroid hormone (PTH),' PTH-related pepresponse was prevented by LaCl,,and voltage-gated Ca2+ tide (PTHrP), 1,25(OH),D,, interleukin l a a n d +, and prostachannel blockers, although the IL-4 effect was more sensitive to nicardipine and nifedipine than to diltiazem. glandin E, (6).These observations were subsequently extended Depolarization by high extracellular K+ concentration to animal models; in particular, in mice IL-4 antagonizes bone also raised [Ca2+],, and, under these conditions, oste- resorption stimulated by either infusion of PTHrP or transoclasts failed to respond to IL-4.On the other hand, plants of tumors expressing PTHrP and IL-la (2). While these observations established IL-4 as a potent antiwhen intracellular Ca2+stores were depleted by thapsigargin, IL-4 still induced an increase in [Ca2+li, although osteolytic factor, the mechanism underlying its effects remain the action of such a smaller in amplitude and transient. Calcitonin also pro- poorly characterized. Because IL-4 inhibits duced [Ca2+l, increases in osteoclasts, yet it only slightly diverse group of bone resorptive factors (11, we reasoned that desensitized these cells to IL-4. Furthermore, IL-4 was this cytokine may impact essential steps of the resorptive procmuch less effective onosteoclasts pretreated (5-10 min) ess. In support of this hypothesis, we and others have docuwith either forskolin or 8-bromo-CAMP. Both IL-4and mented that IL-4 significantly antagonizes osteoclast generacalcitonin were effective even when[Ca"], had been in- tion in vitro ( 7 , 8 )a n d that receptors forthis cytokine are extant creased by exposure to high extracellular Ca2+. Finally, on osteoblasts (9). Furthermore, this cytokine affects the exIL-4 dose dependently inhibited the bone-resorptive ac- pression of both alkaline phosphatase(9), a n d colony stimulattivity of mature osteoclasts.Therefore, IL-4 signal trans- ing factor 1 (10) by osteoblasts. These findings suggest that duction in osteoclasts involves a rapid and sustained IL-4 may regulate the function of many cells operative in the elevation of [Ca2+1,mediated by a voltage-dependent resorptive process. Ca2+ influx, in combination with Ca2+release from intraWhile osteoblasts mediate the osteoclast-activating effects of cellular stores. Modulation of osteoclast [Ca2+],represents apotential mechanism by which IL-4 inhibits bone a number of cytokines (11, 121, the capacity for osteotropic factors to directly regulate osteoclast activity appears to be resorption. uncommon, at least in vitro. Except for prostaglandin E, a n d that other physiologic calcitonin (131, there is limited evidence Interleukin 4 (IL-4) is an immunoregulatory cytokine recently found to influence skeletal metabolism (1,2). Produced and secreted by activated T lymphocytes and mast cells, this *This work has been supported in part by National Institutes of Health Grants AR32087 and AR41255 (to R. C.), AR42356 (to D. L.), DE05413 (to S. L. T.)and grants from the Monsanto-Washington University Collaborative Agreement (to D. L.) and the Shriners Hospitals for CrippledChildren (to S. L. T.). The costsof publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely toindicate this fact. 8 Current address: Domp6, S.p.A., Localitll Campo di Pile, 1-67100 L'Aquila, Italy. On leave from Domp6 S.p.A., L'Aquila, Italy. $$ To whom correspondence and reprint requests should be addressed: Division of Endocrinologyand Bone and Mineral Diseases, The Jewish Hospital of St. Louis, 216 S. Kingshighway Blvd., St. Louis, MO 63110. Tel.: 314-454-7765;Fax: 314-454-5047.

mediators act directly on the osteoclast to reduce its resorptive functions. Furthermore, both prostaglandin E, a n d calcitonin directly inhibit osteoclasts through mechanisms that involve CAMP and/or [Ca"], as second messengers (14). that IL-4 directly influThis report addresses the possibility ences osteoclasts, the cells that ultimately degrade bone matrix. Because osteoclast activity is dependent on intracellular calcium concentration [Ca2+l, (15), we have examined IL-4 effect on[Ca"], i n osteoclast-like cells generated in bone marrowstromal cell co-cultures. We found that IL-4 induces a rapid, sustained elevation in [Ca2+l,in these osteoclastic cells, a n effect initiated and maintained by a voltage-dependent Ca2+ influx, in combination with release of Ca2+ from intracellular The abbreviations used are: PTH, parathyroid hormone; MEM, Eagle's modified Eagle's medium; BSS, balanced salt solution.

13817

13818

IL-4 Increases [Ca2+Iiin Murine Osteoclastic Cells

stores. Thesefindings widen the scope of IL-4's cellular targets in skeletal tissue to include,inaddition to osteoblasts and osteoclast precursors, mature osteoclasts, and provide another potential mechanism by which this immune cytokine inhibits bone resorption. MATERIALSANDMETHODS I'

Reagents a n d Chemicals-Recombinant IL-4 was expressed and purified from baculovirus-transfected SF-9 (Spodoptera frugiperda) cell culture supernatants as described (16, 17) and was stored a t -80 "C. IL-4's biologic potency was determined using the HT-2 (murine T cell line) proliferation assay (18) with the amount of material stimulating half-maximal proliferation equal to 1 unit. Stage 1 colony-stimulating factor 1 (CSF-1) was prepared from supernatants of L929 cells a s described (19). Salmon calcitonin was obtained from Peninsula Laboratories, dissolved in 1.5 mM acetic acid, and storeda t -80 "C. 1,25-(OH),D, was kindlyprovided by Dr. Uskokovic (Hoffman-La Roche). Fura-2 acetoxymethylester (fura-2AM) was purchased from Molecular Probes. The FIG.1. Phase-contrast micrograph of d a y 10 osteoclast-ST-2 stockwas dissolved in acetone, aliquoted, desiccated, and stored a t cell c o - c u l t u r e sd, e m o n s t r a t i n g the presence of a giant -20 "C in the dark. Thedye was redissolved in dimethyl sulfoxide to a multinucleated osteoclastic cell, surrounded by smaller, polyconcentration of 5 mhf, just before use. Ionomycin was obtained from m o r p h i c s t r o m a l and other hematopoietic cells. Magnification x Calbiochem, and thapsigargin from LC Laboratories (Woburn, MA). A 400. balanced salt solution (BSS: 127 mM NaCI, 3.8 mM KCl, 1.2 mM KH,PO,, 1.2 mhl CaCI,, 0.8 mM MgCI,, 5 mM glucose, and 10mM HEPES, adjusted [Ca2+l,by applying the relationship (Equation l), proposed by Grynkto pH 7.4 with 1 N NaOH) was used as bathing medium during the iewicz et al. (23): [Ca2+],experiments.TheCa2+-freebufferwasobtained by omitting CaCI, in the BSS formulation. 5 mM EDTA (Na' salt) was also added, (Eq. 1) and theionic strength of the solution was balancedby decreasing NaCl to 125 mM. The buffer containing a high concentration of K' [K'] was Rmi,and R,,, are the values of R at very made by replacing NaCl withKC1 in theBSS formulation, to give a final where R is the measured ratio, low and saturatingconcentrations of [Ca"],, and P is the ratio of emisconcentration of 80 mM K+ and 50.8 mM Na'. All chemicals, unless sion intensitiesa t 380 nm excitation in these two sets of conditions. The otherwise indicated, and the tissue culture media werefrom Sigma. Cell Cultures-Murine osteoclastic cells were prepared by co-cultur- dissociation constant (K,) for fura-2 + Ca" was assumed to be 224 n# ing nonadherent bone marrowcells with ST-2 cells as described previ- (23). Calcium saturation and zero Ca2+ wereachieved by adding 5 pu ously (7). Briefly, the bone marrow of femurs and tibiae obtained from ionomycin and 2.5 mM EGTA, respectively. In experiments with osteC3WHEN mice (Harlan Sprague-Dawley) was flushed with ice-cold oclasts, the average Rmi, was 1.23 k 0.11, R,,, was 4.88 lr 0.40, and cells, the average a-minimal essential medium (a-MEM). Thecells derived from 8 bones was 2.36 k 0.10 (n = 76). For stromal and fibroblastic Rmi, was 0.97 -c 0.02, R,,, was 4.53 2 0.17, and P was 3.38 * 0.16 (n = (corresponding to 2 mice) were collected, pelleted, resuspended in 50 ml of a-MEM supplemented with 10% heat-inactivated fetal calf serum 21). Measurement of Cyclic AMP-Production of cyclic AMP (CAMP)was containing 500 unitdm1 of stage 1 CSF-1, plated in a 150-mm tissue measured a s described (24). Briefly, day 10 osteoclast cultures were culture dish, and subsequently incubated for 24 h at 37"C in theprespreincubated with a-MEM bovine serum albumin containing 0.1% boence of5%CO,. The nonadherent cells were then collected, pelleted vine serum albumin and1 mM isobutylmethylxanthine for 10 min. The (1,000 rpm, 7 min, 4"C), and resuspended (1 x lo7 cells/ml) in pronase experimental compounds were then added to the medium, andcells the solution (0.02% Pronase (Boehringer Mannheim),1.5 mM EDTAin phoswere incubated for the different time periodsa t 37 "C. Thereafter, the phate-buffered saline). After a 15-min incubation at 37"C, reaction was medium wasremoved and thecell layer containingCAMPwas extracted stopped by adding 200 p1of horse serum, and the suspension was of these with 500 pl of 5% trichloroacetic acid. One-hundred-pl aliquots layered onto 15ml of heat-inactivated horse serum and sedimentedat samples were washed three times with 5 volumes of water-saturated 1 x g for 15 mina t 4 "C. The cell suspension from the top ofthe gradient ethyl ether and then dried. The extract was then analyzed for CAMP was carefully transferred onto 15ml of another heat-inactivated horse utilizing a CAMPradioimmunoassay kit (DuPont NEN). serum gradient and centrifuged for 10 minat 2,000 rpma t 4 "C, and the Bone Resorption Assay-Bone resorption was assessedas the amount cell pellet was suspended in a-MEM containing 10% heat-inactivated of ["Hlproline releasedfrom prelabeled bone particles (25).["Hlprolinefetal calf serum. These fractionated nonadherent bone marrow cells (3 labeled bone particles (25-63 pm in size, 4-6 x 10' cpd100 pg) were x loficells/well) were co-cultured withST-2 cells (3 x lo5cells/well) (20) prepared as described (25). The co-culture was performed a s detailed on 31-mm coverslip diameter for photon counting and placed in 6-well above in 48-well plates (bone marrowcells (5 x 10" cells/well); ST-2 cells tissueculturedishes(3mlperwell)inthepresence of lo-' M (5 x lo4 cells/well)) for 8-10 days until osteoclasts became apparent 1,25(OH),D3and M dexamethasone.Theculturesweremaintained within the ST-2 cell layer. After the co-culture period, culture media for 8-10 days until osteoclastic cells became apparentby phase contrast were removed and 100 pg of ["Hlproline-labeled bone particles in 1 ml microscopy underthe ST-2 cell layer. Medium waschanged twice of fresh a-MEM supplemented 10% heat-inactivated fetal calf serum weekly. containing 1,25(OH),D, (lo-' M) and dexamethasone(10" M)were added Measurement of Cytosolic Calcium-This was performed according toto the experimental wells, along with IL-4 or its vehicle. After a 48-h a method previously described (21,22). In brief, osteoclasticcells devel- incubation, aliquotsof the supernatant were removed, and the released oped from co-cultures grown on round coverslips (6-8 days)were isotope wasassessed by scintillation spectroscopy. Thequantity of washed with serum-free MEM, supplemented with 10 mM HEPES (pH resorbed bone was then calculated and expressed a s micrograms/48 h 7.4), and thenloaded with 10 p~ fura-2AM for 2 ha t room temperature. (25). The coverslips were then washed three times withBSS, mounted on a Statistical Analysis-Because of the high variability of the results temperature-controlled tissue culture chamber (Biophysica Technolo- when expressed as percent changesof [Ca"],, in some cases (when the gies Inc., Baltimore, MD), covered with 1 ml of BSS, and fitted on a coefficient of variation was higher than 15%) the median value and the custom-made aluminum holder. The assembly was positioned in the range are reported as an indication of central tendency. In all other stage of a Nikon Diaphot invertedepifluorescence microscope, equipped cases, data are expressed a s averages k S.E. Group means were comwith Nikon phase-contrast oil-immersionobjectives. pared by either Student's t test or Wilcoxon rank test for unpaired Fluorescence measurements were performed using a dual-excitation samples, as appropriate. spectrofluorometricsystem (CM1111, Spex Industries,Edison,NJ), coupled to the microscope. Fura-2 emission wasrecorded in single cells, RESULTS using photon counting, by alternating between340 and 380 nm excitaAs shown in the phase contrast photomicrograph in Fig. 1, tion every second. Fluorescence intensities emitted between 510 and 520 nm a t each excitation wavelength were ratioed and converted into the osteoclasts utilized in these experiments were morphologi-

Osteoclastic Cells IL-4 Increases [Ca2+Iiin Murine

13819 EGTA 2.SmM

A

1

I 200 150

FIG.2. Effect of IL-4 on [Ca2'l, of single osteoclasts. Co-cultures of bone marrow-ST-2stromal cells were grown on glass coverslips and incubated until osteoclasts developed ( 6 - 8 days), as described under "Materials and Methods."Cells were then loaded with fura-2, and fluorescence was monitored by quantitative fluorometry in single murine osteoclasts. Zkucings represent continuous recording of fura-2 fluorescence (3401380nm excitation ratio; scale on the right), as a direct measure of [Ca"], (scale on the left). IL-4 (200 unitdml) was added when indicated, either in the absence (A) or in the presence ( E )of an anti-IL-4 receptor antibody. Ionomycin (3 p ~ and ) EGTA(2.5mM) were addedwhen indicated for calibration of the fluorescent signal into [Ca"],. The increase in [Ca2+liinduced by different concentrations of the cytokine was calculated when a plateau was reached and expressed as percent of baseline (C). Averages were derived from at least six measurements per each concentration.

l o50 o

2

I

IL4 200U/ml

1

t-"'

lonomycin 3pM

I

5

EGTA 2.5mM

"

B +

anti-IL4receptor

Ab

lonomycln 3pM IL4

50U/ml

150 100

50 I

I

I

1

I

I

Time (min) 250

100 I

C

I

I

I

I

I

30

50

70

100

300

[IL-41

cally identified as large, polymorphic multinucleated cells that wereeasily distinguishable from smaller mononuclear cells also present in our cultures (7). These multinucleated cells express calcitonin receptors (not shown) and respond to calcitonin with an increase of CAMP (7, see below). The basal[Ca"], level in resting murine osteoclasts was 108 2 5 nM, n = 96. Addition of IL-4 to the bathing medium was immediately followed by a sudden increase of [Ca"],, which rapidly reached a plateau phase (Fig. 2 A ) . Osteoclast [Ca''], remained elevated for at least 10 min after IL-4 exposure. Monoclonal anti-IL-4 receptorantibodies (clone M1 (10)) blocked IL-4's effect on [Ca"], (Fig. 2B), indicating that this action of the cytokine is mediated by binding to the previously characterized IL-4 receptor that we have detected in osteoclasts by both Northern analysis and ligand binding.' As illustrated in Fig. 2C, IL-4's effect on [Ca"], was dose-dependent with an apparent ED,,, of 60 unitdml, and maximal activity was observed at 200 units/ ml, a dose that induced an average[Ca"], increase of 209 * 15% of baseline ( n = 16). Because the osteoclast cell preparations were heterogeneous, containing both mononuclear bone marrow cells and residual stromal cells, we next determined thespecificity of the effect of 'A. Shioi, S. L. Teitelbaum, F. P. Ross, J. Ohara, and D. L.Lacey, manuscript in preparation.

(units/rnI)

IL-4 on [Ca"],.No significant changes of fura-2 fluorescence were observed ineitherthesmall mononuclear or inthe spindle-like stromal cells (basal [Ca2+l,= 66 * 5 nM, n = 19) after exposure t o the cytokine (112 2 1%of baseline, n = 19; not shown), indicating that IL-4's capacity to increase [Ca'+], was restricted to osteoclasts. Next, we explored the mechanisms by which IL-4 increased [Ca2+l,in osteoclasts using antagonists of plasma membrane Ca'' channels and inhibitors of intracellular Ca'' release. To verify that theabsence of a n IL-4-induced [Ca"], response was actually due to the inhibitor and not t o IL-4-insensitive cells, IL-4 responsiveness was tested beforehand in each cell preparation by exposing at least three osteoclasts to maximal IL-4 concentrations. Only preparations in which all osteoclasts (three or more) responded to IL-4 with a rise in [Ca2+l,were used. This occurred in more than 80% of all cell preparations. As shown in Fig. 3B, 2-min pretreatment with 100 p~ La3+,an aspecific inhibitor of Ca2+ influx, completely abolished IL-4induced increases in [Ca2+],,suggesting that IL-4 stimulates Ca2+influx as the initiating event of its action on [Ca"],. To explore whether a plasma membrane CaZ+channel was involved, we then studied the sensitivity of this effect to Ca" channel blockers. Bothnifedipine (Fig. 3C) and nicardipine (not shown) at 11.1~concentrations completely abolished IL-4's

13820

IL-4 Increases [Ca2+liin Murine Osteoclastic Cells 250

FIG.3. Effect of Ca2+channel andrelease inhibitors on IL-4 effect on osteoclast [Caa+lp Cells were prepared as described under "Materials and Methods."A, effect of 200 unitslml IL-4 shown as control. 100 PM La3+( B ) , 1 nifedipine ( C ) ,or 100 diltiazem (D) was applied shortly before IL-4, as indicated by the arrows. E, the medium was rapidly exchanged with a buffercontaining 80 mM K+ before addition of IL-4. F , 3 PM thapsigargin was added to the bathing medium before exposure to IL-4.

t

La3+ l O O f i M

B

lo 50

-

250

I C

200

U

.-

+n N

150

U 0 U

50

i

100 50

i t

Thapslgargln 3pM

250 200 150

150

100

50 50

Time effect on [Ca'+],. In some instances, nifedipine or nicardipine addition was followed by a small riseof fura-2 fluorescence (see Fig. 3C). This phenomenon, representing a true [Ca"], increase, hasalso been observed in osteoblastic cells (26), but its physiologic basis remain unknown. Although the IL-4 effect was sensitive to dihydropyridines, the agonist Bay-K8466 did not produce detectable changes in [Ca2+l,in our murine osteoclasts, thus arguing against the hypothesis that the Ca2+influx triggered by IL-4 is mediated by the dihydropyridine-sensitive Ca2+channel described in other systems(27). To verify that a Ca2+ channel wasinvolved in the action of IL-4, osteoclasts were preincubated with diltiazem (1-10 PM),a Ca'' channel blocker of the benzothiazepine class (27). Diltiazem partially, but significantly, inhibited the rise of[Ca'+], induced by the cytokine (median increase 128% of baseline (114-167), n = 7; Fig. 30). We next analyzed the impact of membrane depolarization on IL-4-induced increase of osteoclast [Ca2+],.Depolarization was obtained by rapid medium exchange to a buffer with ahigh K+concentration (80 m).This treatment caused a slow but progressive increase of[Ca"], (median increase 166.5% of baseline (49-320) n = 10; Fig. 3E), which was reversible upon repolarization(not shown). Addition of maximal concentrations of IL-4 to depolarized cells did not elicit any effect on fura-2 fluorescence, indicating that thecytokine's effect on [Ca"], was voltage-dependent (Fig. 3E). We then tested the sensitivity of IL-4 action on osteoclast [Ca2+],t o thapsigargin, an inhibitorof intracellular Ca'' store refilling (28). Addition of thapsigargin was immediately followed by a rapid rise of [Ca'+], (Fig. 3F), probably due to its

(min)

rapid effect on a n intracellular Ca2+pool (28). Cell exposure to IL-4 5 min after thapsigargin addition caused an immediate but transient furtherincrease of [Ca2+li(Fig. 3F).The average increase of [Ca2+l,induced by IL-4 in thepresence of thapsigargin was 141 2 9% of pretreatment values, corresponding to a 63% inhibition. Similar resultswere obtainedby incubating the osteoclast cultures for 30 min with 100 J ~ Mdantrolene, an antagonist of sarcoplasmic Ca2+release (29, 30). Dantrolene decreased the amplitudeof IL-4's effect on [Ca2+l,(not shown) by 57% compared to control experiments. Calcitonin has also been shown to induce increases of [Ca2+], in osteoclasts (31, 32). In line with these observations, the osteoclastic cells in our culturesalso responded to calcitonin (1 nM), with a median [Ca'+], increase of 158% of baseline (136216, n = 5) (Fig. 4B). The calcitonin-induced [Ca2+l,increase was also very rapid, but had a more transient duration than that observed for IL-4 (Fig. 4, A and B ) . Calcitonin also induced CAMP production in the same cells, but with a slower time course compared t o its effect on [Ca''], (Fig. 5). A small but significant production of CAMP was observed after a 5-min exposure, but its maximal effect wasobtained only after a 15-min treatment. These calcitonin responses are a further indication of the osteoclastic nature of the multinucleated cells utilized in our experiments. As shown in Fig. 4C, IL-4 was still able to elevate [Ca2+l, when added to the bathing medium after calcitonin had produced its maximal effect on [Ca2+l,,although 1L-4's effect was slightly but not significantly smaller in amplitudewhen compared to control conditions (average increase of 173 2 32% of

13821

IL-4 Increases [Ca2+Iiin Murine Osteoclastic Cells

F

250

L

100

-

A

250

F

Calcitonin 1 nM

6

n

2 C W

FIG.4. Continuous recording of [Ca2+liof single murine osteoclasts during stimulation with 200 unitdml IL-4 (A), 1 MI calcitonin ( B ) ,calcitonin followed by IL-4 ( C ) ,or forskolin followed by IL-4 ( D l , as indicated by the arrows. Cells were prepared as described under “Materials and Methods.”

.-

+n CY

r

IL4 200U/rnl

250

CI 0

IL4 200U/rnl

250 - C a l c i t o n i n 1 nM

Y

200 150

l o50 o

i

100

50

Time

(m in)

FIG.5. Time course of calcitonin effect on C A M P production by osteoclastic cells. Co-cultures of bone marrow-ST-2 stromal cells prepared as explained under “Materials and Methods” were exposed to either1nM calcitonin ( 0 ) or its vehicle (01and , cellular CAMPwas assessed at the time points illustrated. Results are the average of three experiments. n e

”

0

I

I

I

I

I

I

I

0

10

20

30

40

50

60

Time

(min)

baseline; n = 4). The capacity of IL-4 to increase [Ca2+],in cells pre-exposed to calcitonin was essentially independent of the length of time extant between calcitonin and IL-4 addition. Analogous results were obtained when the order of addition was reversed, ie. calcitonin could induce an increase in [Ca2+li in cells pretreated withIL-4 (not shown). Increasing the intracellular CAMPconcentration by forskolin inhibited the action of IL-4 on [Ca2+],.As shown in Fig. 4 0 , forskolin induced a transient [Ca2+l,spike, followed by a second, slower rise, to a median percent increaseof 180.5 (109-414, n = 10). IL-4 addedt o the osteoclast cultures during the second phase rise caused a detectable increase of [Ca2+],,with a median 112.5% increase above preaddition levels (80-120, n = 8). Similar results were obtained using the CAMPanalog 8-Br-CAMP(not shown). Osteoclasts have been shown to regulate [Ca”], in response to changes in extracellular Ca2+ concentrations ([Ca2+],)(15, 33). In avian osteoclasts, increasing [Ca”], leads to elevated [Ca2+],via a mechanism involving Ca2+ influxthat is preceded by an initial Ca2+release from intracellular stores (15). To

determine whether theaction of IL-4 on osteoclast [Ca2+l,was modulated by [Ca2’l,, the response to thecytokine was assessed after osteoclasts were exposed to high external CaZ+levels. As was observed in freshly isolated avian osteoclasts (151, rapid buffer change to a medium containing 3 mM CaC1, triggered an of [Ca2+li(median increase 416% of acute and marked increase baseline (219-7361, n = 71, followedby a sustained plateau phase at a higher level than resting [Ca2+li(Fig. 6A). Neither depolarization nor IL-4 exposure altered the osteoclastic cell response to high [Ca2+],(Fig. 6B). When IL-4 was added to the bathing medium during the plateau phase following the high [Ca2’l, treatment, it retained the capacity to elevate [Ca2+], above plateau levels (Fig. 6C), although the average increase induced by the cytokine in theseconditions was smaller thanin the presence of standard [Ca2+l,(161 f 11, n = 3). Similar to IL-4, calcitonin also increased [Ca2+liwhen added to the bathing medium after exchange to high [Ca2+l,buffer (Fig. 6D). Thus, the action of IL-4 and calcitonin on osteoclast [Ca2+l,is mostly independent of [Ca2+l,levels.

13822

IL-4 Increases [Ca2+liin Murine Osteoclastic Cells 300 250

FIG.6. Effect of increasing extracellular Caz+ concentration on osteoclast[Caz+],and on osteoclastresponse to IL-4. Cells prepared as described under to high "Materials were CaCI, and by addition Methods" exposed of

10 1.11 of a 300 mM stock to the bathing buffer (A). B , cells were depolarized by rapid exchange to a buffer containing 80 mM K', then exposed, in succession, t o IL-4 and 3 mM CaCI,, when indicated by the arrows. After addition of 3 mM CaCI,, the cells were exposed to IL-4 ( C ) or 100 n~ calcitonin ( D ) .Dacings are representative of at leastthree experiments for each

I C

.-

::lIlII!5

200

150

100

100

n

+

N

CoCI2 3mM

0

300

0

-

1

Calcitonin lOOnM

D

250

IL4 200U/ml

250

-

200

200

300-

150

150

100 L

100

II

Tim e

( 5 min)

Quantitative data on the experiments described above are summarized in Fig. 7. In essence, blockade of Ca" entry from Nifedipine 1pM the extracellular milieu and membrane depolarization abolished IL-4 action, while depletion of intracellular Ca2+stores Nicardipine 1pM * (6) only partially affected the cytokine's effect. Increasing [Ca2+], Diltiazern 1 OpM *' (7) did not significantly alter IL-4's action on [Ca"],. T h a p s i g a r g i n 3pM (3) Finally, we examined the effects of IL-4 on bone resorption. In order t o study IL-4's effect on terminally differentiated osteoclasts, the bone resorption assays were performed at a time CaCI, 3rnM 3) when the osteoclast number had maximized. In these conditions, the inhibitoryaction of IL-4 on osteoclastformation Control (10) should not interfere with the evaluation of the cytokine's effect I 0 50 100 on mature osteoclasts. As shown in Fig. 8, IL-4 inhibited bone resorption in a dose-dependent manner, withdetectable effects Percent o f maximal IL-4 effect observed with 10 unitdm1 IL-4. At 200 units/ml, IL-4 inhibited FIG.7. Effect of calcium transport inhibitorson IL-Cinduced bone resorption by 39%. increase of [Caz+lj. Experiments were performedas described in Figs.

B

f

2, 3, 4, and 7, and the results were expressed as the amplitude of the effect of IL-4 on [Ca2+l,in the presence of the antagonist, normalized to the response obtained in the absence of inhibitors (Control).The conThe present studies demonstrate that IL-4 rapidly increases centration of IL-4 was 200 units/ml in all experiments. The number of [Ca"], in murine osteoclasts, via a mechanism involving both experiments for each condition is shown inparentheses. * , p < 0.01; **, extracellular CaZ+ influx and release of Ca2+from intracellular p < 0.05; p values were obtained by unpaired Student's t test, except for the experiments with diltiazem and high K', for which the Wilcoxon stores. The initiating event appears be to the opening of a Ca2+ rank test was performed. DISCUSSION

channel, as indicated by the sensitivity to dihydropyridines and membrane depolarization. The partial inhibitory action of diltiazem on IL-4's effect may be explained by a lower sensitivity of osteoclast Ca2+ channels to antagonists of the benzothiazepine class, as previously observed in avian osteoclasts (15). The slow and progressive [Ca"], increase induced by cell depolarization in high K+-containingbuffers would suggest thepresence of voltage-dependent Ca'+ channels on murine osteoclasts. On the other hand, the lack of Bay-K8466 effect argues against the hypothesis that these channels are the same as the dihydropyridine Ca'' channels described in other systems(27). Although theevidence for voltage-dependent Ca'+ channels isstill only pharmacologic, they have been identified in avian osteoclasts as well (15).The voltage-gated Ca" channel observed in that study produced sustained elevations of [Ca2+l,that lasted for at least 300 s following depolarization with high [K+l(151, a time course virtually identical with thatobtained with themurine osteoclastic cells employed in the presentstudy. Although

our results point to a Ca" channel-operated phenomenon, the possibility that the IL-4 effect on Ca2+influx may occur via other mechanisms, such as the Na+/Ca'+ exchanger, alsoa voltage-dependent Ca2+ transporter(341, cannot be excluded. While Ca2+influx from the extracellular milieu is absolutely required for the IL-4 effect t o occur, Ca2+release from internal stores also appears to contribute t o the full response. The sustained phaseof IL-4's effect was notobserved after depletion of intracellular Ca" stores, and the amplitude of the [Ca''], increase wasreduced in theseconditions. Thus, the transient rise of [Ca2+],induced by the cytokine when internal stores are depleted mayreflect the opening of a membrane channel.However, the integrityof intracellular pools is essential to produce the tonic elevation of [Ca"], after the early rapid increase. The transient effect of IL-4 observed after osteoclasts are exposed to high [Ca2+],may also reflect a similar process, since [Ca'+I, also

IL-4 Increases [Ca2+liin Murine Osteoclastic Cells

13823

are not conclusive as to the role of CAMP in regulating the [Ca"], response to calcitonin and IL-4, because both forskolin and 8-Br-cAMP, which produce high intracellularCAMPlevels, also increased [Ca''], by themselves, thus making the interpretation of the results very difficult. Nevertheless, the present study demonstrates that IL-4 and calcitonin, two factors that inhibit bone resorption, differ in the intracellular signal-transducing mechanisms they activate in osteoclasts. Ourresults extendpreviousobservations inavian osteoclasts (15,33) to mammalian osteoclastic cells by establishing a linkage between changes in [Ca'+], with homologous changes in [Ca''],. In our hands, the rapid [Ca"], rise that follows an increase of [Ca"], is not voltage-dependent, sinceit occurred in depolarizingmedia.Interestingly,both IL-4 and calcitonin wereable to augment [Ca"], after a new steady state was Control I 10 50 200 I reached following cell exposure to high [Ca"],. Therefore, these [IL-4] (U/ml) two factors may still be effective on osteoclasts in conditions FIG.8. Effect of IL-4 on bone resorption. The co-cultures were that already dampen their activity (15). On the other hand, the established and incubated for 8-10 days until osteoclastic cells became modest reduction of the amplitude of IL-4's effect compared to confluent as described under "Materials and Methods." After this pecontrol conditions indicates that a feedback inhibition of high riod, 100 pg of [3Hlproline-labeledbone particles were added to each well, along with IL-4 at different concentrations, as indicated. The [Ca"], levels on the channel operated by IL-4 may occur. isotope released form the resorbed bone particles was assessed after a Alterations of [Ca"], levels appear t o affect osteoclast activ48-h incubation by scintillation spectroscopy. The quantity of bone ity. Agonists or factors that augment [Ca"],, such ashigh conresorbed was then determined. Data are presented as average * S.E. of four replicates. The average resorbed bone in the cultures treated with centrations of either [K'], or [Ca2+le,as well as calcitonin and 10,50,and 200 unitdm1 IL-4 was significantly different from controlat now IL-4, lead to reduced bone resorption (15). In contrast, thep level of 0.016, 0.014, and 0.003, respectively (single-tailed t test). extracellular acidification, which decreases [Ca'+], ( E ) , stimulates bone resorption. The meansby which high [Ca"], leads t o induces discharge of Ca'' from intracellular pools (15).In this reduced osteoclast bone-resorbing activityare still under invesscenario, the action of IL-4 on osteoclast [Ca''], may occur tigation. However, the leading hypothesis envisions osteoclast through a Ca'+-induced intracellular Ca'' release triggered by polarization and adherence t o bone matrix as the primary the initial, rapidCa'+ influx. This mechanism accounts for the event. Osteoclast adherence to bone, mediated by specific adrapid [Ca"], spiking following depolarization in muscle cells, hesion structures called podosomes (40, 411, is followed by seand it is mediated by the ryanodine receptor. The sameprocess cretion of protons and lysosomal enzymes intothe sealed may also be operative in other cells expressing the ryanodine resorbing compartment under the ruffled border (25, 40). Poreceptor (35).Interestingly, ryanodine receptors are present in dosome expression and/or formation is inhibited by elevated osteoclasts and are thought tobe involved in [Ca2+l,regulation [Ca''], (15, 42) through a mechanism involving gelsolin, a calby [Ca"], (36,371. Release of Ca'' from intracellular storesmay cium-dependent, actin-regulating protein. In the presence of in turn trigger Ca2+ influx,perhaps via the concomitant release high [Ca"],, gelsolin disrupts themicrofilament architecture of of diffusible factor(s) present in the stores (38, 391, thus pro- the podosomes (4345). The sustained increase of [Ca"], genviding the basisfor an autoregenerating system that maintains erated by IL-4, along with other anti-resorptive agents, could a prolonged [Ca''], elevation. theoretically cause podosome disassembly leading to impaired The effect of IL-4 on [Ca"], was specific for osteoclasts and osteoclast function by inhibiting attachmentt o bone. appeared to be similar, but distinct, from that of calcitonin. In In summary, we have demonstrated that IL-4, a cytokine fact, theeffect of IL-4 on [Ca"], was very rapid and sustained, with potent antiresorptive activity, rapidly increases [Ca"], in with [Ca"], remaining elevated for more than 10 min. On the murine osteoclasts, via Ca2' influx, probably through opening other hand, thechanges in [Ca2+l,produced by calcitonin were of a Ca'' channel and Ca'' release from intracellular stores. transient, witha return towardbaseline values withinapproxi- These findings provide a potential mechanism by which IL-4, mately 5 min. Such observations closely reproduce previous acting directly on terminally differentiated osteoclastic cells, results obtained in chick (31)and rat (31,321osteoclasts. In the could dampen their osteolytic activity. studies of Moonga et al. (321, the rapid [Ca"], increase induced REFERENCES by calcitonin occurred also in a Ca'+-free medium, but the decline to baseline [Ca"], wasfaster,suggesting abiphasic 1. Watanabe, K., Tanaka, Y., Morimoto, I., Yahata, IC, Zeki, K., Fujihira, T., Yamashita, U., and Eta, S. (1990) Biochem. Biophys. Res. Commun. 172, mechanism of calcitonin action, involving an initial rapid re1035-1041 lease from intracellular Ca'' stores, followed bylater activation 2. Watanabe, K., Sato, K., Kasono, K., Nakano, Y., and Eto,S. (1991) J. Bone Min. Res. 6, 5288 (abstr.) of a Ca'' channel. This sequence of events proposed for calci3. Howard, M., Farrar, J., Hilfiker, M., Johnson, B., Takatsu, K., Hamaoka, T., suggest for tonin is fundamentally different from what our data and Paul, W. E. (1982) J. Exp. Med. 155,914-923 IL-4. 4. Rabin, E. M., Ohara, J., and Paul,W. E. (1985) Proc. Natl. Acad. Sci. U. S. A . 82,2953-2959 The effect of IL-4 on [Ca"], was only partially prevented by 5 . Boulay, J. L., and Paul, W. E. (1992) J. B i d . Chem. 267,20525-20528 high intracellular concentrations of CAMP. It is interesting to 6. Mundy, G . R., Ibbotson, K. J., DSouza, S. M., Simpson, E.L., Jacobs, J. W., and Martin, T. J. (1984) N. Engl. J . Med. 310, 1718-1727 note that, while IL-4 does not increase CAMP in these cells A., Teitelbaum, S. L., Ross, F. P., Welgus, H. G., Suzuki, H., Ohara, J., (data not shown), calcitonin increases both CAMPand [Ca2+],. 7. Shioi, and Lacey, D. L. (1991) J . Cell. Biochem. 47,272-277 However, the [Ca"], rise induced by this hormone is not as 8. Kasono, K., Sato, K., Sato, Y., Tsushima, T., Shizume, K., and Demura, H. (1993) Bone Min. 21, 179-188 sustained as that driven by IL-4, suggesting that the CAMP 9. Lacey, D. L., Erdmann, J. M., and Ohara, J. (1993) J. Cell. Biochem. 63, increase may act to lessen calcitonin's effect on [Ca''],, by de122-134 activating the membrane channel. This hypothesis contrasts 10. LaceY, D. L., Erdmann, J. M., Shima, M., Kling, S., Ohara, J., and Perkins,S. L. (1992) J. Bone Min. Res. 7, S95 (abstr.) with that proposed by Malgaroli et al. (31), who theorize that 11. Thornson, B. M., Saklatvala, J., and Chambers, T. J. (1986) J. Exp. Med. 164, the effect of calcitonin on [Ca"], is driven by CAMP.Our results 1OP112 80

[A

13824

IL-4 Increases [Ca2+liin Murine Osteoclastic Cells

12. Thomson, B. M., Mundy, G. R., and Chambers, T. J. (1987) J. Zmmunol. 138, 775-779 13. Chambers, T. J., Fuller, K., and McSheehy, P. M. J. (1985) J. Pathol. 145, 297-305 14. Chambers, T. J., and Hall, T. J. (1991) Vitam. Horm. 46, 41-86 15. Miyauchi, A., Hruska, K. A,, Greenfield, E. M., Duncan, R. L., Alvarez, J. I., Barattolo, R., Colucci, S., Zambonin-Zallone,A,,andTeitelbaum, S. L. (1990) J. Cell Bid. 111, 2543-2552 16. Ohara, J., Colian, J. E., Zoon, K., Maloy, W. L., and Paul, W. E. (1987) J. Zmmunol. 139, 1127-1134 17. Ohara, J. (1989) in TheYear in Immunology: Zmmunoregulatory Cytokines and Cell Growth (Cruse, J. M., and Lewis, R. E. J., eds) pp. 126-159, Karger, Base1 18. Lacey, D. L., Axelrod, J., Chappel, J. C., Kahn, A. J., and Teitelbaum, S. L. (1987) J. Zmmunol. 138, 1680-1686 19. Stanley, E.R., and Heard, P. M. (1977) J . Biol. Chem. 252, 43054312 20. Udagawa, N., Takahashi, N., Akatsu, T., Sasaki, T., Yamaguchi, A,, Kodama, H., Martin, T. J., and Suda, T. (1989) Endocrinology 125, 1805-1813 21. Fujimori,A., Cheng, S.,Anoli,L. V., and Civitelli, R. (1991)Endocrinology128, 3032-3039 22. Civitelli, R., Miyauchi,A., and Hruska,K. A. (1991) J. Tissue Culture Methods 13,217-228 23. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260,34403450 24. Takahashi, N., Akatsu, T., Sasaki, T., Nicholson, G. C., Moseley, J. M., Martin, T. J., and Suda, T. (1988) Endocrinology 123, 1504-1510 25. Blair, H. C., Kahn, A. J., Crouch, E. C., Jeffrey, J. J., and Teitelbaum, S. L. (1986) J. Cell Biol. 102, 1164-1172 26. Civitelli, R., Kim, Y. S., Gunsten, S. L., Fujimori, A,, Huskey, M., Avioli, L. V., and Hruska, K. A. (1990) Endocrinology 127,2253-2262 27. Catterall, W. A., Seagar, M. J., and Takahashi, M. (1988) J . Biol. Chem. 263, 3535-3538 28. Takemura, H., Hughes, A. R., Thastrup, O., and Putney, J. W., Jr. (1993) J.

Biol. Chem. 264, 12266-12271 29. Kojima, I., Kojima, K., and Rasmussen, H. (1985) J. Biol. Chem. 260, 91779184 30. Reid, I. R., Civitelli, R., Halstead, L. R., Avioli, L. V., and Hruska, K. A. (1987) Am. J. Phvsiol. 253. E45-E51 31. Malgaroli, A,”, Meldolesi, J., Zambonin-Zallone, A,, and Teti,A. (1989) J. Biol. Chem. 264, 14342-14347 32. Moonga, B. S., Towhidul Alam, A. S. M., Bevis, P. J. R., Avaldi, F., Soncini, R., Huang, C. L., and Zaidi, M. (1992) J . Endocrinol. 132, 241-249 33. Zaidi, M., Alam, A. S., Huang, C. L., Pazianas, M., Bax, C.M., Bax, B. E., Moonga, B. S., Bevis, P. J., and Shankar, V. S. (1993) Cell Calcium 14, 271-277 34. Eisner, D. A,, and Lederer, W. J. (1985) Am. J. Physiol. 248, C189-C202 35. Berridge, M. J. (1993) Nature 361,315-325 36. Zaidi, M., Shankar, V. S., Towhidul Alam, A. S., Moonga, B. S., Pazianas, M., and Huang, C. L. (1992) Biochem. Biophys. Res. Commun. 188, 1332-1336 37. Shankar, V. S., Bax, C. M., Bax, B. E., Alam, A. S., Moonga, B. S., Simon, B., Pazianas, M., Huang, C. L., andZaidi, M. (1993) J. Cell Physiol. 155, 120-129 38. Randriamampita, C., and Tsien, R.Y.(1993) Nature 364,809-813 39. Parekh, A. B., Terlau, H., and Stuhmer, W. (1993) Nature 364,814-818 40. Marchisio, P. C., Cirillo, D., Naldini, L., Primavera, M. V., Teti, A,, and Zambonin-Zallone, A. (1984) J. Cell Biol. 99, 1696-1705 41. Zambonin-Zallone, A,, Teti, A,, Carano,A., and Marchisio, P. C. (1988)J . Bone Min. Res. 3, 517-523 42. Teti, A,, Blair, H. C., Schlesinger, P., Grano, M., Zambonin-Zallone, A,, Kahn, A. J., Hruska, K. A,, and Teitelbaum, S. L. (1989) J . Clin. Znuest. 84, 773780

43. Marchisio, P. C., Cirillo, D., Teti, A,, Zambonin-Zallone, A,, and Tarone, G. (1987) Exp. Cell Res. 169, 202-214 44. Bryan, J., and Coluccio, L. M. (1985) J . Cell Biol. 101, 1236-1244 45. Chaponnier, C., Yanmey, P.A., and Yin, H. L. (1986) J. Cell B i d . 103, 14731481