Expression of Annexin I, II, V, and VI by Rat Osteoblasts In

Expression of annexin I, II, V, and VI by rat osteoblasts in primary culture: Stimulation of annexin I expression by dexamethasone Françoise Suarez, Bernard Rothhut, Christine Coméra, Lhousseine Touqui, Françoise Russo Marie, Caroline Silve

To cite this version:

Françoise Suarez, Bernard Rothhut, Christine Coméra, Lhousseine Touqui, Françoise Russo Marie, et al.. Expression of annexin I, II, V, and VI by rat osteoblasts in primary culture: Stimulation of annexin I expression by dexamethasone. Journal of Bone and Mineral Research, American Society for Bone and Mineral Research, 2009, 8 (10), pp.1201-1210. 10.1002/jbmr.5650081007. hal-02656032

HAL Id: hal-02656032 https://hal.inrae.fr/hal-02656032 Submitted on 29 May 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. JOURNAL OF BONE AND MINERAL RESEARCH Volume 8, Number 10, 1993 Mary Ann Liebert, Inc., Publishers

Expression of Annexin I, 11, V, and VI by Rat Osteoblasts in Primary Culture: Stimulation of Annexin I Expression by Dexamethasone

FRANCOISE SUAREZ,' I3ERNARD ROTHHUT,2 CHRISTINE COMERA,'.* LHOUSSEINE TOUQUI,3 FllANCOISE RUSSO MARIE,' and CAROLINE SILVE'

ABSTRACT To determine whether rat osteoblasts synthesize proteins of the annexin family and to evaluate the extent to which glucocorticoids modulate the expression of annexins by these cells, osteoblasts were grown in primary cultures in the absence or presence of dcxamethasone, and the expression of annexins was evaluated by immunoblotting using polyclonal antibodies against human annexins. Four different annexins (I, 11, V, and VI) were found to be expressed by rat osteoblasts. The expression of annexin I, but not the other annexins studied, was increased in osteoblasts cultured in the presence of dexamethasone (173 f 33% increase comparing untreated cells and cells treated for 10 days with 5 x lo-' M dexamethasone). Increased expression of annexin I was observed after the third day of exposure to dexamethasone and rose thereafter until day 10; annexin I expression increased with dexamethasone concentrations above M throughout the range of concentrations studied. The increase in annexin I protein was associated with an increase in annexin I mRNA and was completely blocked by the concomitant addition of the glucocorticoid receptor antagonist RU 38486. The increase in annexin I content following dexamethasone treatment was associated with an increase in alkaline phosphatase activity and PTH-induced CAMP stimulation, whereas phospholipase A, activity in the culture medium was reduced to undetectable levels. The finding that four annexins are expressed in rat osteoblasts in primary culture raises the possibility that these proteins could play an important role in bone formation by virtue of their ability to bind calcium and phospholipids, serve as Ca'+ channels, interact with cytoskeletal elements, and/or regulate phospholipase A1 activity. In addition, the dexamethasone-induced increase in annexin I may represent a mechanism by which glucocorticoids modify osteoblast function.

INTRODUCTION and I1 (cell attachment),"." bone morphogenic proteins (promotion of bone formation),"' alkaline phosphatase HE PRODUCTION OF A MINERALIZED ORGANIC MATRIX by and osteonectin (bone mineralization), Ib.') and osteocalcin Tthe osteoblasts depends on a series of complex events, (bone remodeling).1*)Although studies evaluating the spe- including the synthesis of matrix proteins, the specific or- cific function of these components in the production of the ganization of this matrix necessary to create a milieu pro- mineralized organic matrix have improved our understand- pitious for the deposition of mineral, and subsequent min- ing of this process, important gaps remain in our knowl- eral deposition.''.') A number of osteoblast products and edge of bone formation. In particular the definition of bone matrix components necessary for bone formation osteoblast products necessary for the initiation of bone have been characterized, including bone siiiloproteins I mineralization is far from complete.

. - - - . .. -~ ...... 'CNRS URA 583. Hapital Necker Enfants Malades, Paris, France. 'INSERM U 332, ICGM. Paris, France. 'Unite des Venins, lnstitut Pasteur, Paris, France. 'Present affiliation: Laboratoire de Toxicologie et Pharrnacologie, INRA, Toulouse. France. 1202 SUAREZ ET AL.

In recent years, independent approaches have led to the serum (FCS) and cultured at 37°C in 95% air and 5% identification of a family of proteins, “annexins,” which CO,. The medium was replaced 48 h later and every third are characterized by their ability to bind calcium and inter- day thereafter with complete medium containing 10% act with phospholipids in a calcium-dependent FCS. Cells reached confluence on day 4-5 of culture. The annexins (originally designated lipocortin, calpactin, endonexin, and chromobindin)‘12)are structurally related Pretreatment of the cells and appear to be derived from a common ancestral gene.‘’) Annexins are present in a number of different tis- Dexamethasone (Sigma; 5 pl/ml culture medium, 5 x sues, including cartilage and matrix vesicles isolated from lo-” to 5 x M final concentration) was added to chondrocytes.‘1’-15’Although the presence of annexins in osteoblast cultures for various times. Fresh dexamethasone bone or bone cells in culture has not been reported, these was added at each medium change. In some experiments, proteins could play a role in mineralization by virtue of 50 pl/ml culture medium of I x lo-‘ M RU 38486 (kindly their ability to bind to calcium and phospholipids. For ex- provided by Roussel Uclaf Laboratories, Romainville, ample, calcium acid phospholipid complexes have been France, 5 x M final concentration) was added to the found to be concentrated in mineralizing connective tissue cultures in addition to dexamethasone. Stock solutions of and have been found to promote hydroxyapatite deposi- dexamethasone and RU 38486 (lo-’ M in ethanol) were tion.(16-18)Other roles for annexins in the homeostasis of stored at -2O”C, and intermediate dilutions were per- bone are also possible. The phenotype of osteoblasts is formed in medium without serum. Control cultures, main- subject to regulation by glucocorticosteroids, which are tained in parallel, received 5 or 55 81 solvent alone (ethanol potent modulators of bone cell metabolism and function in diluted 1:100 in medium without serum). vivo(19) and in vitro.(20-22)One member of the annexin family (lipocortin I) was defined as a glucocorticoid-in- Sodium dodecy1:sulfate (SDS)-polyacrylamide gel ducible phospholipase A, inhibitor.(2’”5) In this context, electrophoresis and immunoblotling some of the effects of glucocorticoids on bone, including their effects on arachidonic acid metabolism by osteo- Cells were grown in 28 cm’ culture dishes and treated blasts, could be mediated through the regulation of such with dexamethasone as described. To collect the cells, cul- annexins. ture dishes were placed on ice, washed twice with 3 ml ice- The purpose of this study was threefold. First, we deter- cold Dulbecco’s phosphate-buffered saline, scraped in 200 mined whether osteoblasts in primary culture synthesize pl Laemmli SDS-sample buffer,‘”) and immediately boiled members of the annexin family. Second, we evaluated the for 10 minutes. Samples were stored at -20°C. The pro- extent to which glucocorticosteroids modulate the expres- tein content was determined on cells grown in dishes of the sion of annexins by these cells, and compared these same size and cultured in parallel. Protein (100 pg) was changes to those observed in two other characteristics of loaded on SDS-polyacrylamide gels. osteoblast activities, parathyroid hormone (PTH)-induced SDS-polyacrylamide gel electrophoresis was performed CAMP production and alkaline phosphatase activity. Fi- according to Laemmli‘17’on 12.5% acrylamide gels (1.5 nally, in an attempt to identify the mechanism by which mm thick). The molecular weight of proteins was deter- glucocorticosteroids modulate arachidonic acid metabo- mined by comparison to migration of standard proteins lism in osteoblasts, we sought evidence that annexins mod- (Pharmacia-LKB), which were electrophoresed in parallel. ify osteoblast phospholipase A, activity. The separated proteins were then transferred from the gel to nitrocellulose membranes (HAHY 304 FO, Millipore) according to Burnette(181using an electroblotting appara- MATERIALS AND METHODS tus (Pharmacia-LKB 21 17-250 Novablot) operated at con- Primary rat osteoblast cultures stant current (250 mA) for 2.5 h in 20 mM Tris-base, 150 mM glycine and 20% methanol. At the end of the transfer, Osteoblasts were isolated from the calvariae of newborn nitrocellulose membranes were stained with india ink (1 pl rats (Charles River, Saint Aubin les Elbeuf, France) by se- india ink per ml of 50 mM Tris and 150 mM NaCI) or with quential enzymatic digestion, as previously described for 0.2% Ponceau red (Serva, 33429) in 3% trichloroacetic mice,126)with slight modifications. The central parts of acid for 10 minutes and incubated in 50 mM Tris, 150 mM parietal bones were removed aseptically and the periosteal NaCI, and 5% skimmed powdered milk (TNM) for I h at layers carefully stripped off. Parietal bones were extracted 37°C with constant shaking. Antisera against annexins (see by stirring for 15 minutes at 37°C in Ca-free phosphate- later) were diluted (1:500) in 10 ml TNM and incubated buffered saline containing 5 mM EDTA and 0.05% tryp- with membranes for I5 h at 4°C with constant shaking. sin, followed by three incubations of 20 minutes in phos- Following five washes in TNM, the blot was incubated for phate-buffered saline containing 0.1 ‘To collagenase (Sigma, 30 minutes at room temperature with 1251-labeledprotein A St. Louis, MO). The cells released during the last two incu- diluted in 15 ml TNM (lo5cpm/ml, specific activity =24 bations were combined and cultured. Cells were plated at 2 mCi/mg), washed four more times with TNM, and dried x lo4 cells/cm’ in 0.25 ml/cm’ of Dulbecco’s modified between sheets of filter paper; autoradiography of the Eagle’s medium (Eurobio, les Ullis, France) supplemented dried blot was performed for 4-7 days at -80°C. Strips of with 100 U/ml of penicillin G, 50 pg/of streptomycin, 1 the blots that displayed bands on the autoradiograms were mM glutamine (complete medium), and 20% fetal calf excised and the amount of radioactivity was quantified ANNEXINS I, 11, V, and VI IN OSTEOBLASTS 1203

using a gamma counter. Radioactivity measxed in strips for I minute. The reactions were then initiated by the addi- of the blots not displaying bands was subtracted as back- tion of CaCl, at 10 mM (final concentration). The fluores- ground. cence measurements were performed with a Jobin el Yvon Polyclonal antibodies were raised in rabbits by popliteal JY3D spectrofluorometer equipped with a Xenon lamp. lymph node of annexins I, V, and VI purified The fluorescence intensity was monitored using excitation from human placenta.(30~31JCross-reactivity between the and emission wavelengths of 345 and 398 nm, respectively, different antisera was less than 5%. Antibody against an- and a slit width of 4 nm. The final ethanol concentration nexin I1 was a gift from J.C. Cavadore (INS,ERM U 249, was less than 0.1% and had no effect on the assay. Phos- CNRS LP 8402, Montpellier, France). pholipase A, activity measured in medium without serum and in medium containing 10% FCS was used as a blank Northern blot analysis for phospholipase A, activity in cells and medium, respectively. Results are expressed as pmol hydrolyzed substrate/ Cells were grown in 69 cm' culture dishes and treated with minute/mg protein. dexamethasone as described. Total cellular RNA was isolated with guanidinium thiocyanate followed by centrifugation in CAMPcellulur content cesium chloride gradient^.'^,) RNA (10 pg) was electrophoresed in a 1To agarose gel containing 0.45 M formaldehyde(321 Cells were grown in 24-well culture plates and treated in 20 mM MOPS (3-(N-morpholino) propane sulfonic acid) for 10 days with dexamethasone as before. Stimulation of buffer at 160 V for I h, transferred to Nytrari nylon mem- cAMP production was performed for 10 minutes at room branes (Schleicher and Schuell, NY), and dried for 2 h at temperature in the presence of 1 mM isobutyl-l-rnethyl- 80°C. Blots were hybridized with cDNA of human annexins I xanthine (Sigma) and extracted with ethanol as previously and V1 (kindly provided by Barbara Wallnei, Biogen Re- de~cribed."~)cAMP production was determined as previ- search Corporation, Cambridge, MA), which were labeled ously described(36)using a modification of the Gilman pro- by random priming (1.5 x 10' dpm/25 ng cDNA, lo6 cpm/ ml) and denatured by heating to 100°C for 10 minutes before I II V VI use. Prehybridization (30 minutes at 65°C) and hybridization kD (overnight, 65°C) were done in 50 mM PIPES, pH 6.5, 50 mM sodium phosphate, 1 mM EDTA, 100 mM NaCl and 5% 100 - SDS.'33)Blots were then washed in standard saline citrate 80 - (SSC) and 5% SDS, first in 20 ml for 20 mirutes at room temperature and again in 400 ml for 30 minutes at 65°C. To ensure that t he total amount of RNA transferred was similar 4 9.5 - for each sample, the blots were hybridized with 7" P-labeled 28s RNA probe as described by Barbu and Dautry.(34)Blots 3 2.5 _. were then dried and autoradiographed for 24 11. - 27.5 - Measurement of phospholipase A2 activity Cells were grown in 69 cm' culture dishes and treated with dexamethasone as indicated earlier. After 10 days in 18.6 - culture, medium was collected, cells were scraped in 2 mi culture medium without serum, and both medium and cells were stored at -20°C. On the day of the assay, cells were -+ -+ -+ -+ lysed at 4°C by ultrasonication (30 s, I50 W) using an DEXA MSE sonifier. Phospholipase A, activity was measured in FIG. 1. Expression of annexin 1, 11, V, and VI in osteo- culture medium and cell homogenates using a fluorometric blasts. Effect of dexamethasone. Western blots of annexin assay as described by Radvanyi et al.(35)The fluorescent 1, 11, V, and VI were detected by Western blotting in pro- substrate [ 1 -palmitoyl-2-( 10-pyrenyldecanoy1)-sn-glycero- tein extracts from osteoblasts cultured in the absence (-) 3-monomethyl phosphatidic acid; Interchim, Monthcon, or in the presence (+) of 5 x lo-' M dexamethasone for 10 France] was dried under nitrogen and suspended in ethanol days. Osteoblast cultures, dexamethasone treatment, SDS- at 0.2 mM. Vesicles were prepared by mixing the ethanol polyacrylamide gel electrophoresis, and Western blot anal- solution of fluorescent phospholipid with an aqueous me- ysis of annexins were performed as described in Materials dium for 2 minutes. The reaction solution was prepared by and Methods. Protein (100 pg) was electrophoresed in each introducing sequentially into a 4 x 10 mm disposable plas- lane. In addition to annexin V, material with a higher mo- tic cuvette 980 buffer (100 mM Tris-HCI, 1 M NaCI, and lecular weight (MW) was recognized by the annexin V anti- pI bodies. The MW of this cross-reactive material does not 1 mM EGTA, pH 7.9, 10 pl substrate (2 pM final concen- correspond to that of any annexin described, and we have tration), and 10 pl of a 10% fatty acid-free bovine serum no indication of its nature, The intensity of this band was albumin solution (0.1 To final concentration); aliquots of very low compared to that given by annexin V. The data cellular homogenates or medium (10-50 pl) were intro- reported are representative of two independent experi- duced into the cuvettes and allowed to equilibrate at 37°C ments. 1204 SUAREZ ET AL. tein binding assay"7) in which human erythrocytes are Western blots demonstrated a 173 f 33% increase com- used as the source of the CAMP binding protein.(381Re- paring dexamethasone-treated and untreated cells (n = 7). sults are expressed as picomoles cAMP/100 pg protein. In contrast, dexamethasone treatment did not significantly change expression of annexin II, V, or VI (Fig. I). Cellular alkaline phosphatase activity Time course and dose-response curve of the Cells were grown in 12-well culture plates and treated induction of annexin I by dexarnethasone with dexamethasone as described for 2-10 days. At the end of the culture period, medium was aspirated and the plates The characteristics of the dexamethasone-induced stimu- were stored at -20°C. On the day of the assay, cells from lation of annexin I were further investigated by determin- each dish were scraped into 1 ml freshly prepared solution ing the kinetics and the dose-response curve of the dexa- containing 8.5 g saccharose and 100 pl Triton X-100/100 methasone effect. An increase in annexin I expression was ml water. Cells were sonicated for 20 s on ice and centri- not observed during the first 3 days of exposure of osteo- fuged for 10 minutes at 2500 rpm at 4°C. and alkaline blasts to 5 x lo" M dexamethasone (Fig. 2). Increased ex- phosphatase activity was measured in the supernatant pression of annexin I was observed after the third day of using p-nitrophenyl phosphate as a substrate (Sigma). Ali- exposure and thereafter rose progressively until day 10, the quots (100 pl) of supernatant were incubated for 1 h at latest time point evaluated. 37°C in the presence of 1 pmol substrate, 0.625 pmol When osteoblasts were cultured in the presence of dexa- MgCI2, and 25 pmol glycine in a total volume of 500 pl at methasone at concentrations < M, no increase in an- pH 10.5. The reaction was stopped by adding 2.5 ml 1 N nexin I expression was observed after 10 days (Fig. 3). In NaOH and 10 mM EDTA. The amount of p-nitrophenol liberated was measured by absorbance at 405 nm. Results are expressed as nmol p-nitrophenol released/minute/ 100 pg protein. Preliminary experiments demonstrated that alkaline phosphatase activity could not be detected in the culture medium. 0234710 Protein content DAYS Protein content was estimated by the method of Lowry 2.5 c et al.'''' using bovine serum albumin as standard.

Presentation of data Results are expressed as mean * standard deviation un- less otherwise stated. All immunoblots and Northern blots were performed with equivalent amounts of cellular proteins and total RNA, 100 and 10 pg, respectively. bd 0.5 I I I I , I RESULTS 0 2 4 6 8 10 Expression of annexins I, II, V, and VI in DAYS osteoblasts: Effect of dexarnethasone FIG. 2. Time course of the stimulation by dexametha- Cellular proteins were extracted from rat osteoblasts and sone of annexin I content in osteoblasts. Western blot separated by electrophoresis and expression of annexins analysis of annexin I (top) was performed in protein ex- evaluated by immunoblotting. Proteins reactive with anti- tracts from osteoblasts cultured in the presence of 5 x lo-' sera specifically recognizing annexins I, 11, V, and VI were M dexamethasone for the indicated times. The bands of expressed by these cells (Fig. 1). In each case, the molecu- lower molecular weight are likely to be proteolytic frag- lar weights of these proteins corresponded closely to those ments of annexin I generated during the cell preparation previously reported for these four different annexins. In for the protein blotting. Regions of the blots correspond- some cases, fainter bands of lower molecular weight were ing to annexin I, including the lower molecular weight also reactive with these antibodies and presumably repre- bands, were excised, and the amount of radioactivity in each band was determined in a gamma counter (bottom). sented annexins that were partially degraded during the ex- Osteoblast cultures, dexamethasone treatment, SDS-poly- traction procedures. acrylamide gel electrophoresis, and Western blot analysis The expression of annexin 1 was significantly increased of annexin I were performed as described in Materials and in osteoblasts cultured for 10 days in the presence of 5 x Methods. Protein (100 pg) was electrophoresed in each lo-' M dexamethasone (Fig. I). Quantitation of annexin I lane. Similar results were obtained in two independent ex- expression by determining "'I-labeled protein A bound to periments. ANNEXINS I, 11. V, and VI IN OSTEOBLASTS 1205 general, annexin I expression increased with dexametha- Effect of dexarnethasone on other osleoblast sone concentrations above lO-'O M throughout the range of phenotypic characteristics concentrations studied (Fig. 3). Alkaline Phosphatase Activity: Treatment the cells In some experiments, the glucocorticoid receptor antag- of with 5 x IC7 M dexamethasone for 10 days resulted in an onist RU 38486 was added to the cultures in the presence increase in alkaline phosphatase activity from 0.94 0.23 or in the absence of dexamethasone to determine if the ef- to 2.21 0.10 nmol/minute/100 Fg protein (mean SD, fect of dexamethasone was mediated through the activa- + n = 6; Fig. 6). Like the increase in annexin I, maximal in- tion of glucocorticoid receptors. The induction of annexin crease in alkaline phosphatase was observed at 10 days, the I by dexamethasone (5 x M, 10 days) was totally latest time studied, whereas shorter exposures had little or blocked by the concomitant addition of 5 >: M RU no effect on alkaline phosphatase activity. 38486 (Fig. 4). RU 38486 alone had no effect on annexin I content (Fig. 4). PTH-Induced CAMP Production: Treatment of the cells with 5 x M dexamethasone for 10 days also Effect of dexarnethasone on annexin mRNA greatly enhanced the stimulation of CAMP production by PTH (14- and 106-fold stimulation in response to 240 nM To determine if the increase in annexin 1 protein induced PTH, respectively, without and with dexamethasone treat- by dexamethasone was associated with a similar increase in ment; Fig. 7). annexin I mRNA, cells were maintained for 7 and 10 days in the presence or absence of 5 x lo-' M dexamethasone Phospholipaw A, Acvivity in Osteoblast Cu1ture.s: Be- and the level of annexin I mRNA was examined by North- cause annexin was first identified by its ability to inhibit ern blot analysis. Results demonstrated that the increase in I phospholipase A,, it was interest to measure phospho- annexin I protein was associated with an increase in an- of nexin I mRNA, comparing dexamethasone-treated and un- lipase A, in osteoblasts cultured in the absence or presence of dexamethasone. Phospholipase A, activity was easily treated osteoblasts (Fig. 5). In agreement with the absence detectable in both culture supernatants and cell sonicates of effect of dexamethasone on the content of annexin VI of osteoblasts maintained the absence dexamethasone in osteoblasts, no effect of dexamethasone on annexin VI in of (Table I). Interestingly, treatment of cells with 5 x M mRNA was observed (Fig. 5).

2.5

c7 1 I 0 7

+RU +RU CONTROL +DW FIG. 4. Effect of RU 38486 on the stimulation by dexa- DEXAMETHASONE concentration (M) methasone of annexin I expression in osteoblasts. Western FIG. 3. Annexin I content in osteoblasts as a function of blot analysis of annexin I (top) was performed in protein increasing concentrations of dexamethasone. Western blot extracts from osteoblasts cultured in the absence (control) analysis of annexin I (top) was performed in protein ex- or in the presence of 5 x M dexamethasone (+Dexa) tracts from osteoblasts cultured in the presence of the indi- and 5 x M RU 38486 (+RU) for 10 days. Regions of cated dexamethasone concentrations for 10 days. Regions the blots corresponding to annexin 1 were excised, and the of the blots corresponding to annexin I were excised, and amount of radioactivity in each band was determined in a the amount of radioactivity in each band was determined gamma counter (bottom). Osteoblast cultures, dexametha- in a gamma counter (bottom). Osteoblast cultures, dexa- sone, and RU 38486 treatment, SDS-polyacrylamide gel methasone treatment, SDS-polyacrylamide gel electropho- electrophoresis, and Western blot analysis of annexin I resis, and Western blot analysis of annexin I were per- were performed as described in Materials and Methods. formed as described in Materials and Methods. Protein Protein (100 pg) was electrophoresed in each lane. The (100 pg) was electrophoresed in each lane. 'The data re- data reported are representative of two independent exper- ported are representative of two independent experiments. iments. 1206 SUAREZ ET AL.

0' I I I I 1 2 4 6 8 10 VI DAYS + DEXAMETHASONE FIG. 6. Effect of dexamethasone on cellular alkaline phosphatase activity in osteoblasts. Osteoblasts were cultured for 10 days without dexamethasone or in the pres- 28 S ence of 5 x M dexamethasone for the indicated times, and cellular alkaline phosphatase activity was measured as described in Materials and Methods. Alkaline phosphatase activity measured in osteoblasts cultured for 10 days without dexamethasone is the value given at time 0. Results are - +-+ the mean f SD, n = 6. 7 10 DAYS DISCUSSION FIG. 5. Northern blot analysis of RNA from osteoblasts probed with cDNA for human annexins I and VI and a 28s ribosomal RNA probe. Osteoblasts were cultured in This study demonstrates that four members of the an- the presence of 5 x M dexamethasone for 7 and 10 nexin family, annexins I, 11, V, and VI, are produced by days. Total cellular RNA was prepared using a cesium gra- rat osteoblasts maintained in primary cultures, and it is the dient, electrophoresed in a 1 To agarose gel and transferred first description of the expression of annexins by bone to Nytran nylon membranes, and blots were hybridized cells. Exposure of osteoblasts to 5 x M dexametha- with radiolabeled probes as described in Materials and sone results in an increase in the cellular content of an- Methods. RNA (10 pg) was electrophoresed in each lane. nexin I and mRNA coding for this protein, whereas the cellular content of the other annexins studied was not modified. The increase in annexin I content following dexamethasone treatment was not an early event, however, but was associated with an increase in other activities con- dexamethasone for 10 days suppressed phospholipase A, sidered typical of mature osteoblasts (alkaline phosphatase activity in the culture medium to undetectable levels, activity and PTH-induced CAMP stimulation). whereas no consistent effect on cellular phospholipase A, activity was observed (Table 1). Annexins and bone formation Total Cellular Protein: Dexamethasone treatment de- Although annexins share structural and functional fea- creased the amount of total cellular protein in the cultures tures,('+"' the presence of a unique N-terminal sequence (35 * 6 and 24 f 5 pg/cm', respectively, without and for each annexin suggests that each annexin has a specific after dexamethasone treatment; n = 5, p < 0.05). The role. In view of the previously described properties of dif- electrophoretic pattern of proteins obtained from osteo- ferent members of the annexin family, several possible blasts cultured in the absence of dexamethasone and after mechanisms exist by which annexins could contribute to treatment with 5 x M dexamethasone for 10 days ap- bone formation, none of which is mutually exclusive. peared similar both before and after transfer to nitrocellu- First, annexins share a 17 amino acid residue consensus lose membranes, as demonstrated by the Coomassie blue sequence that is probably involved in Ca'+-dependent bind- staining of the SDS-polyacrylamide gels and the india ink ing to anionic phospholipids.(lO)The ability of annexins to staining of nitrocellulose membrane after transfer (Fig. 8). bind calcium and phospholipids could contribute to miner- ANNEXINS I, II, V, and VI IN OSTEOBLASTS 1207

.- kD A B C m

0, 2 0 200 0 40.5

3 2.5 100 2 7.6

0 1 10 100 1000 18.5 hPTH (1-34) ng/ml FIG. 7. Stimulation of cAMP production by increasing PTH concentrations in osteoblasts treated or not with dexamethasone. Osteoblasts were cultured iri the absence -+ -+ -+ or in the presence of 5 x M dexamethasone for 10 DE XA days, and stimulation of cAMP production by the indicated concentration of PTH was performed a:; described in FIG. 8. Distribution of cellular proteins obtained from Materials and Methods. Results are the mean * SD of one untreated (-) and dexamethasone-treated (+) osteoblast representative experiment, performed in triplicate: un- cultures. Proteins were separated by SDS-polyacrylamide treated cells (open squares); dexamethasone-treated cells gel electrophoresis. (A) Pattern of proteins as revealed (closed circles). Similar results were obtained in a second by staining proteins in the gel with Coomassie blue R-250 experiment. and (B) proteins after transfer to nitrocellulose membranes and stained with india ink. (C) Western blots of annexin I from the corresponding protein extracts. Osteoblast cultures, dexamethasone treatment, SDS-polyacrylamide gel TABLE1. PHOSPHOLIPASE A, ACTIVITYIN THE MEDIUM AND electrophoresis, and Western blot analysis of annexin I CELL PELLET OF OSTEOBLASTS CULTURED IN THE ABSENCEOR were performed as described in Materials and Methods. PRESENCEOF 5 X M DEXAMETHASONE~ Protein (100 pg) was electrophoresed on each lane. A simi- Medium Cells lar pattern was obtained in all experiments.

Experiment Conrrol + Dexa Control + Dexa

1 3.9 <0.1 3.3 1.1 blast-collagen interactions, (‘O) implying that it could bind 2 1.67

lipase A, activity and to evaluate the consequences of their biochemical properties and their specific regulation, dexamethasone-induced changes in phospholipase A, ac- this unique group of membrane-directed regulatory pro- tivity on osteoblast function. teins are likely to play a significant role in the osteoblast function and bone metabolism. The demonstration that glucocorticoids strongly inhibit phospholipase A1 activity Dexamethasone and annexin expression I release in the medium and increase the expression of an- We found that the addition of dexamethasone to osteo- nexin 1 may provide new insights into the mechanisms by blast cultures increased the expression of annexin I, but which glucocorticoids modify osteoblast biology. not of the other annexins, in a time- and concentration-dependent manner. The induction of annexin 1 by dexamethasone was blocked by the concomitant addition of the glucocorticoid receptor antagonist RU 38486, demonstrating REFERENCES that dexamethasone-induced annexin I RNA and protein synthesis occurred via the activation of the glucocorticoid I. Marks SC Jr, Popoff SN 1988 Bone cell biology: The regula- receptor. tion of development, structure, and function in the skeleton. Am J Anat 183:l-44. Conflicting data exist in the literature regarding the in- 2. Raisz LG, Kream BE 1983 Regulation of bone formation. ducibility of annexin 1 by glucocorticoids. (2324) Steroid in- Med Prog 309:29-35, 83-89. duction of annexin I has been observed consistently in 3. Somerman MJ. Fisher L.W. Foster RA, Sauk JJ 1988 Human monocytes incubated in vitro or recovered from animals or bone sialoproteins I and I1 enhance fibroblast attachment in humans treated with steroids. In contrast, no induction of vitro. Calcif Tissue Int 43:SO-53. annexin I mRNA or protein has been demonstrated in 4. Oldber GA, Franzen A, Heinegard D 1986 Cloning and se- most cell lines, including in U937 cells'42)and primary cul- quencing of rat bone sialoprotein (osteopontin) cDNA re- tures of human endothelial cells,(43)although dexametha- veals an Arg-Gly-Asp cell binding sequence. Proc Natl Acad sone was found to increase both annexin 1 mRNA and pro- Sci USA 83:88 19-8823. 5. Wang EA. Rosen V, D'Alessandro JS. Bauduy M, Cordes P, tein in 3T3 fibroblasts.(44)The discrepancy observed be- Harada T. Israel DI, Hewick RM. Kerns KM. LaPan P, tween these studies may be explained in part by the state of Luxenberg DP, McQuaid D, Moutsatsos IK, Nove J. Woz- differentiation of the cells'4') and/or by the experimental ney JM 1990 Recombinant human bone morphogenetic pro- design used to characterize the effect of dexamethasone. tein induces bone formation. Proc Natl Acad Sci USA 87: For instance, the effect of dexamethasone on annexin 1 2220- 2224. synthesis in human endothelial cells was studied after 3 h 6. Beerstsen W, van den Bos T 1992 Alkaline phosphatase in- of incubation of the cells in presence of the but duces the mineralization of sheets of collagen implanted sub- the effect of dexamethasone on 3T3 fibroblasts was ob- cutaneously in the rat. J Clin Invest 89:1974-1980. served after 4 days of culture in the presence of dexameth- 7. Termine JD, Kleinman HK, Whitson SW, Conn KM. Mc- asone.'4') Consequently the term "induction" covers two Garvey ML., Martin GR 1981 Osteonectin. a bone-specific protein linking mineral to collagen. Cell 26:99-105. different processes, a rapid induction, as described in 8. Price PA, Parthemore JG, Deftos LJ 1980 New biochemical monocytes, and a relatively slow induction generally asso- marker for bone metabolism. Measurements by radioimmu- ciated with cellular differentiation, as described in 3T3 fi- noassay of bone gla protein in the plasma of normal subjects broblasts. The induction of annexin I content by dexa- and patients with bone disease. J Clin Invest 66:878-883. methasone in osteoblasts was never detected before 3 days 9. Barton GJ, Newman RH. Freemont PS, Crumpton MJ 1991 of exposure to dexamethasone. The increase in annexin I Amino acid sequence analysis of the annexin super-gene fam- was associated with an increase in cellular alkaline phos- ily. Eur J Biochem 198:749-760. phatase activity and with a potentiation of the stimulation 10. Geisow MJ. Fritsche U, Hexham JM. Dash B, Johnson T of cAMP production by PTH, as described for human 1986 A consensus amino-acid sequence repeat in Torpedo bone cells.(46.4')rodent osteoblastic cells,'4*)and rat osteo- and mammalian Cat+-dependent membrane-binding proteins. Nature 320:636-638. sarcoma cells. (49-51) Stimulation of PTH-induced cAMP II. Pepinsky RB, Tizard R, Mattaliano RJ, Sinclair KL., Miller production and alkaline phosphatase activity are two hall- GT, Browning JL, Chow EP, Burne C. Huang KS, Pratt D, marks of the osteoblastic phenotype. Such an effect of Wachter L, Hession C, Frey AZ, Wallner BP 1988 Five dis- dexamethasone on osteoblast differentiation has been ob- tinct calcium and phospholipid binding proteins share ho- served in studies in which dexamethasone has been shown mology with lipocortin 1. J Biol Chem 263:10799-10811. to stimulate osteogenesis(20.22)and the expression of bone 12. Crumpton MJ, Dedman JR 1990 Protein terminology tangle. matrix proteins associated with mineral tissue forma- Nature 345:212. tion.(," Thus, the induction of annexin l expression by 13. Genge BR, Wu LNY, Wuthier RE 1990 Differential fraction- dexamethasone may be part of a general effect of dexa- ation of matrix vesicle proteins. Further characterization of methasone on osteoblast differentiation. It will be interest- the acidic phospholipid-dependent Ca"-binding proteins. J Biol Chem 265:4703-4710. ing to compare the regulation of annexin metabolism in 14. Genge BR, Wu LNYH, Adkinsson HD, Wutheri RE 1991 cells cultured in the presence of factors thought to pro- Matrix vesicle annexins exhibit proteolipid-like properties. J mote matrix deposition and mineralization, such as ascor- Biol Chem 266:10678-10685. bic acid and P-glycerophosphate. IS. Pfaffle M, Ruggiero F, Hofmann H, Fernandez MP. Selmin In conclusion, we report the expression of annexins 1, 11, 0, Yamada Y, Carrone R, von der Mark K 1988 Biosynthe- V, and VI by rat osteoblasts in primary culture. Because of sis. secretion and extracellular localization of anchorin CII, a ANNEXINS I, 11, V, and VI IN OSTEOBLASTS I209

collagen-binding protein of the calpactin fami:y. EMBO J 7: 35. Radvanyi F, Jordan L, Russo-Marie F, Bon C 1989 A sensi- 2335-2342. tive and continuous fluorometric assay for phospholipase A, 16. Boskey AL, Posner AS 1977 The role of synthetic and bone using pyrene-labeled phospholipids in the presence of serum extracted Ca-phospholipid-PO, complexes in liydroxyapatite albumin. Anal Biochem 177:103-109. formation. Calcif Tissue Res 23:25 1-258. 36. Silve C, Grosse B, Tau C, Garabkdian M. Fritsch J, Delmas 17. Boyan BD 1985 Proteolipid-dependent calcification. In: But- PD, Cournot-Witmer G, Balsan S 1986 Response to parathy- ler WT (ed.) The Chemistry and Biology of Mineralized Tis- roid hormone and 1,25-dihydroxyvitamin D, of bone-derived sues. Ebsco Media, Birmingham, AL, pp. 12.5- I3 1. cells isolated from normal children and children with abnor- 18. Posner AS 1987 Bone mineral and the mine:-alization pro- malities in skeletal development. J Clin Endocrinol Metab cess. In: Peck WA (ed.) Bone and Mineral Research, Annual 62: 5 83- 5 90. 5. Elsevier, Amsterdam, pp. 65-1 16. 37. Gilman A 1970 A protein binding assay for 3’5’-cyclic adeno- 19. Peck WA 1984 The effects of glucocorticoidi on bone cell sine monophosphate. Proc Natl Acad Sci USA 67:305-312. metabolism and function. Adv Exp Med Biol 171:1I1-119. 38. lyengar R, Birnbaumer L 1981 Techniques in cyclic nucleo- 20. Bellows CG, Heersche JNM, Aubin JE 1990 Determination tides research. In: Schrader WT. O’Malley BW (eds.) Labo- of the capacity for proliferation and differentiaion of osteo- ratory Methods Manual for Hormone Action and Molecular progenitor cells in the presence and absence ‘of dexametha- Endocrinology, Vol. 9. Biological Associates, Houston. pp. sone. Dev Biol 140:132-138. 37-38. 21. Kasugai S. Todescan R Jr, Nagata T. Yao KI-, Butler WT, 39. Lowry OH, Rosebrough RJ, Farr AL, Randall RJ 1951 Pro- Sodek J 1991 Expression of bone matrix protcins associated tein measurement with the fohn phenol reagent. J Biol <:hem with mineralized tissue formation by adult rat bone marrow 193:265-275. cells in vitro: Inductive effects of dexamethasone on the 40. Mauch C. von der Mark K, Helle 0, Mollenhauer J, Pffaffle osteoblastic phenotype. J Cell Physiol 147:l II - 120. M, Krieg T 1988 A defective cell surface collagen-binding 22. Tenenbaum HC. Heersche JNM 1985 Dexameihasone stimu- protein in dermatosparactic sheep fibroblasts. J Cell Biol lates osteogenesis in chick periosteum in vitro. Endocrinol- 106:205-21 I. ogy 117:2211-2217. 41. Rojas E, Pollard HB, Haigler HT, Parra C. Burns AL 1990 23. Browning JL. Ward MP, Wallner BP, Pepinsky RB 1990 Calcium-activated endonexin 11 forms calcium channels Studies on the structural properties of lipocortin-l and the across acidic phospholipids bilayer membranes. J Biol Chem regulation of its synthesis by steroids. In: Melli M, Parente I. 265:21207-21215. (eds.) Cytokines and Lipocortins in Inflammation and Dif- 42. Bienkowski MJ, Petro MA, Robinson LJ 1989 Inhibition of ferentiation. Wiley L.iss. New York, pp. 27-45. thromboxane A synthesis in U937 cells by glucocorticoids. J 24. Flower RJ 1990 Lipocortin. In: Melli M, Parente L (eds.) Biol Chem 264:6536-6544. Cytokines and Lipocortins in Inflammation ard Differentia- 43. Hullin F, Raynal P, Ragab-Thomas JMI;. Fauvel J. Chap H tion. Wiley Liss, New York, pp. 11-25. 1989 Effect of dexamethasone on prostaglandin synthesis 25. Russo-Marie F 1991 Lipocortins: An update. Prostaglandins and on lipocortin status in human endothelial cells. Inhibi- Leukot Essent Fatty Acids 42:83-89. tion of prostaglandin 1, synthesis occurring without altera- 26. Wong G, Cohn DV 1974 Separation of parathyroid hormone tion of arachidonic acid liberation and of lipocortin synthe- and calcitonin-sensitive cells from non-responsive bone cells. sis. J Biol Chem M4:3506-3513. Nature 252:713-715. 44. Wong WT. Nick HS, Frost SC, 12992 Regulation of annexin 27. Laemmli UK 1970 Cleavage of structural proteins during the I in adipogenesis: CAMP-independent action of methyliso- assembly of the head of bacteriophage T4. Ntture 227:680- butylxanthine. Am J Physiol 262:C91-C97. 685. 45. Solito E, Raugei G, Melli M, Parente L 1991 Dexamethasone 28. Burnette WN 1981 “Western blotting”: Electrophoresis trans- induces the expression of the mRNA of lipocortin I and 2 fer of proteins from sodium dodecyl sulfate polyacrylamide and the release of lipocortin 1 and 5 in differentiated, but not gels to unmodified nitrocellulose and radiographic detection undifferentiated U-937 cells. FEBS Lett 291:238-244. with antibody and radioiodinated protein A. Anal Biochem 46. Beresford JN, Gallagher JA. Poser JW, Russell RGC; 1984 72~248-254. Production of osteocalcin by human bone cells in vitro. Ef- 29. Goudie RB, Horne CHW, Wilkinson PA 1’266 A simple fects of 1,25(OH),D,, parathyroid hormone and glucocorti- method for producing antibody to a single selected diffusible coids. Metab Bone Dis 5:229-234. antigen. Lancet 2:1224. 47. Silve C, Fritsch J, Grosse B, Tau C, Edelman A, Delmas P, 30. Comkra C. Rothhut B, Russo-Marie F 1990 Identification Balsan S, Garabkdian M 1989 Corticosteroid-induced and characterization of phospholipase A, inhibitory proteins changes in the responsiveness of human osteoblast-like cells in human mononuclear cells. Eur J Biochem 188:139-146. to parathyroid hormone. Bone Miner 6:65-75. 31. Rothhut B, Comkra C, Prieur B. Errasfa M. Minassian G, 48. Chen TL. Feldman D 1979 Glucocorticoid receptors in sub- Fusso-Marie F 1987 Purification and characterization of a populations of cultured rat bone cells. Mechanism of dexa- 32-kDa phospholipase A, inhibitory protein (lipocortin) from methasone potentiation of parathyroid hormone-stimulated human peripheral blood mononuclear cells. FI3BS Lett 219: cyclic AMP production. J Clin Invest 63:750-758. 169- 1 75. 49. Catherwood BD 1985 1.25-Dihydroxycholecalciferol and 32. Sambrook J, Fritsch EF, Maniatis T 1989 Molecular cloning. glucocorticosteroid regulation of adenylate cyclase in an In: Nolan C (ed.) A Laboratory Manual, 2nd ed. Cold osteoblast-like cell line. J Biol Chem 260:736-743. Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 50. Rizzoli R. von Tscharner V, Fleisch HF 1986 Increase of pp. 7-19-7-22. adenylate cyclase catalytic unit activity by dexamethasone in 33. Virka GD, Northemann W, Schieles BR, Widera G, Broome rat osteoblast like cells. Biochem J 2373447-454. S 1990 Simplified northern blot hybridization using 5% so- 51. Rodan SB, Fisher MK, Egan JJ, Epstein PM. Rodan GA dium dodecyl sulfate. Biotechniques 8:370-37 l. 1984 The effect of dexamethasone on parathyroid hormone 34. Barbu V, Dautry F 1989 Northern blot norma1ii:ation with 28 stimulation of adenylate cyclase in ROS 1712.8 cells. Endo- S rRNA oligonucleotide probe. Nucleic Acids Res 17:7115. crinology 115:951-958. 1210 SUAREZ ET AL.

52. Yamamoto I. Potts JT, Segre GV 1988 Glucocorticoids increase parathyroid hormone receptors in rat osteoblastic osteosarcoma cells (ROS 1712). J Bone Miner Res 3:707-712.

Address reprint requests 10: Caroline Silve CNRS URA 583 Tour Lavoisier #me etage H6pital Necker Enfants Malades 149 rue de Skvres 75743, Paris Cedex 15, France

Received in original form July 30, 1992; in revised form February 17. 1993; accepted February 18, 1993.