- Email: [email protected]
Distinct pathways of extracellular signal-regulated kinase activation by growth factors, fibronectin and parathyroid hormone 1–34
BBRC Biochemical and Biophysical Research Communications 305 (2003) 573–578 www.elsevier.com/locate/ybbrc
Distinct pathways of extracellular signal-regulated kinase activation by growth factors, fibronectin and parathyroid hormone 1–34q Edelgard Kaiser1 and Srinivasan Chandrasekhar* Gene Regulation, Bone and Inflammation Research, Lilly Research Laboratories, DC 0403, Eli Lilly and Company, Corporate Center, Indianapolis, IN 46285, USA Received 11 April 2003
Abstract Growth factors, hormones, and matrix proteins regulate osteoblast proliferation and differentiation, acting through cognate receptors. Since each of the receptors are coupled to a variety of distinct signal transduction pathways, in this report we evaluated whether there is a common convergent intermediate step that allows cross-talk among the various pathways. Since extracellular signal-regulated kinases 1 and 2 (Erk1/2) play a role in mitogenesis and differentiation processes, we evaluated the effects of various osteotrophic factors on Erk1/2 phosphorylation in osteoblasts. Osteoblasts isolated from the metaphyseal marrow (MM) and diaphyseal marrow (DM) of 4–6 week old male rat longitudinal bones were grown to confluency and Erk1/2-phosphorylation was evaluated using antibodies that recognized either the total or the phosphorylated form of the kinase. There was very little Erk1/2 phosphorylation in cells kept in suspension. Both MM and DM cells attached to fibronectin (FN), demonstrated Erk1/2 phosphorylation that persisted for at least up to 8 h. Platelet-derived growth factor AB (PDGF-AB) induced a transient and robust Erk1/ 2 phosphorylation that was attenuated by 2 h. Studies with specific inhibitors indicated that the effects of these factors were mediated by protein kinase C, by receptor tyrosine kinase, as well as by protein phosphatases. Parathyroid hormone (PTH 1–34), a bone anabolic agent however, caused a down-regulation of FN stimulated Erk1/2 phosphorylation in MM derived cells. The inhibitory effect of PTH was mediated through cAMP-dependent protein kinase A (PKA) activation. The data collectively suggest that a combination of diverse extracellular stimuli regulates Erk1/2 phosphorylation that may ultimately influence osteoblast proliferation and/or differentiation. Ó 2003 Elsevier Science (USA). All rights reserved.
Bone is a dynamic tissue that is continually repaired and remodeled in response to a variety of signals. The interaction between osteoblasts and osteoclasts is responsible for maintaining bone homeostasis. The regulation of osteoblast proliferation and differentiation is complex, and is subject to modulation by a variety of osteotrophic factors such as PTH, bone morphogenetic proteins, TGFb1, PDGF, and extracellular matrix proteins [1–10]. However, it is not clear how these multiple q
Abbreviations: PTH, parathyroid hormone; FN, fibronectin; MM, metaphyseal marrow; DM, diaphyseal marrow; Erk1/2, extracellular signal-regulated kinase 1/2; TGF-b, transforming growth factor b; PDGF, platelet-derived growth factor. * Corresponding author. Fax: 1-317-276-9722. E-mail address: [email protected] (S. Chandrasekhar). 1 Present address: Department of Bone Pathology, Center for Biomechanics/UKE, Hamburg University, Martinistr. 52, 20246 Hamburg, Germany.
signals are transduced and coordinated that ultimately result in osteoblast formation and differentiation. Mitogen-activated protein kinases, such as the p42/p44 serine/theronine extracellular signal-regulated kinases (Erk 1 and 2), play a central role in the proliferation and gene transcription in a variety of cells and in response to a variety of signals [11–15]. In this report, we evaluated whether Erk1/2 might be a key intermediary in osteoblast differentiation stimulated by various factors. Osteoprogenitor cells were isolated from the metaphyseal and diaphyseal marrow of young rats and were evaluated for Erk1/2 activation after in vitro treatment with PDGF-AB, PTH 1–34, and FN. The results indicate that FN stimulated Erk1/2 phosphorylation in osteoprogenitors was prolonged, while PDGF treatment resulted in a robust, but transient activation of Erk1/2. Erk1/2 phosphorylation, stimulated with PDGF or FN, was completely abolished by PTH 1–34. These results
0006-291X/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0006-291X(03)00820-9
574
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
suggest that the Erk1/Erk2 phosphorylation may serve as a converging point that links a variety of osteotrophic signals.
Methods Animals. All of the studies were done using 4–6 week old (75– 100 g), Sprague–Dawley male rats (Harlan, Indianapolis, IN). The rats were fed ad libitum with Purina chow (calcium 1%, phosphate 0.61%, and PMI Feeds, St. Louis, MO) and water. Rats were sacrificed and the bones were used for isolation of cells as described below. All animal protocols were approved by the Lilly Animal Care and Use Committee. Isolation of osteoblasts/osteoprogenitors. MM and DM-derived osteoprogenitor/osteoblasts were isolated from the femurs of rats as described [3]. In brief, freshly harvested bones were kept in a-minimal essential media (a-MEM) + 10% fetal bovine serum (FBS) + 5% penicillin (Pen) and streptomycin (Strep), rinsed for 10 s in 70% ethanol, followed by several rinses with HankÕs balanced salt solution (Mg2þ / Ca2þ -free). For the isolation of the MM-derived osteoblasts, about 3 mm of the metaphyses subjacent to the growth plate were cut off, minced, and digested for an hour at 37 °C with 0.6 ml of 0.25% trypsin in HankÕs balanced salt solution (Mg2þ /Ca2þ -free solution) for each femur, centrifuged, and resuspended in a-MEM + 10% FBS + 1% Pen/Strep. The debris was separated from the cells with a sieve and the cells were plated in a-MEM + 10% FBS + 1% Pen/Strep. For the isolation of the DM-derived osteoblasts, both ends of the femur were cut off and the bone marrow was flushed out with phosphate-buffered saline (Mg2þ / Ca2þ -free), using a syringe. The cells were centrifuged and plated in aMEM + 10% FBS + 1% Pen/Strep and incubated at 37 °C (5% CO2 + air). The non-adherent population was removed by aspiration after 24–48 h and the cultures were fed with a-MEM + 10% FBS + 1% Pen/Strep. The media were changed every fourth day thereafter. Only the first passage cells were used for all studies described. Reagents. Growth factors (PDGF-AB or TGF-b1) were obtained from R&D Systems (Minneapolis, MN) and rat PTH 1–34 and 3–34 were obtained from Bachem (Torrance, CA). The following (H-89, genistein, calphostin, staurosporine, forskolin, sodium orthovanadate, and okadaic acid) were purchased from Calbiochem (La Jolla, Ca). LY294002, a PI-3 kinase inhibitor was a gift from Dr. Chris Vlahos, Eli Lilly and Company. The proteins were solubilized in 4 mM HCl (for PTH, 1 mM) containing 1% bovine serum albumin (BSA) + 0.15 M NaCl, pH 7.2. The stock solutions of the compounds were prepared in dimethyl sulfoxide (1–10 mM) and then diluted into 4 mM HCl (or 1 mM HCl) containing 1% BSA + 0.15 M NaCl, pH 7.2, before addition. Treatments. Erk1/2 levels (both total and phosphorylated Erk1/2) were analyzed in cells treated with various factors at confluency or after attachment to FN coated onto tissue culture plates. Cells were grown to near confluency in six well plates and washed three times with PBS (Mg2þ /Ca2þ -free) and treated with growth factors or PTH 1–34 in a-MEM + 1% Pen/Strep + 0.1% BSA. For experiments involving inhibitors, cells were pre-treated for 15 min with various compounds, followed by the addition of growth factors or PTH 1–34 at indicated concentrations and time. For assays involving attachment to FN, cells grown to near confluency were first serum-deprived for 24 h with aMEM + 1% FBS + 1% Pen/Strep, were washed twice with PBS (Mg2þ / Ca2þ -free), and released by EDTA treatment (Versene 1:5000, GibcoBRL, Grand Island, NY). Cells were kept in suspension for 1 h at 37 °C in a-MEM containing 2% BSA, medium was changed to aMEM containing 0.1% BSA prior to plating of the cells onto tissue culture dishes, previously coated with FN (10 lg/ml in PBS at 4 °C for 16 h). Western blot analysis of Erk1/2. After treatments, the media were removed, the attached/confluent cells were scraped and lysed by boiling
in SDS buffer (10 mM Tris–HCl, pH 7.4, 1% SDS, and 1 mM Na3 VO4 ), boiled an additional 5 min, and then cooled down on ice. The volumes of all samples were kept constant and the lysates were then sonicated. Equal volume of the lysates was mixed with sample buffer containing dithiothreitol, the samples were subjected to electrophoresis on a denaturing SDS–polyacryamide gel (7% resolving/3% stacking), and transferred to 0.45 l nitrocellulose filter (Pharmacia, San Francisco, CA). The membrane was blocked with 5% blotto (nonfat dried milk in 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20) for 1 h at 25 °C, incubated with 1:1000 dilution of the rabbit primary antibody (phospho-specific P44/42 MAPK antibody, New England BioLabs, MA) for 1 h at room temperature, washed four times with 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20 for 15 min, incubated with 1:2000 dilution of horseradish peroxidaseconjugated secondary antibody (Southern Biotechnology, Birmingham, AL) for 1 h at 25 °C, and again washed four times as above. For the assessment of relative amounts of total Erk1 and two independent of their phosphorylation status, filters were stripped for 30 min at room temperature with a buffer containing 10% SDS, 62.5 mM Tris–HCl, pH 6.7, and 100 mM 2-mercaptoethanol, extensively washed with 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20, and blocked with 5% blotto. The primary antibody (rabbit anti rat antibody, New England BioLabs, MA) was used at 1:5000 dilution (5% blotto), while the anti-rabbit horseradish peroxidase-conjugated secondary antibody was used at 1:10,000 dilution. The immunoreactive bands were visualized using the ECL detection system (Amersham). The relative amount of the 42 kDa (Erk1) and 44 kDa (Erk2) was quantitated by densitometric scanning of the X-ray film.
Results PDGF transiently induces Erk1/2 phosphorylation in confluent metaphyseal and diaphyseal bone marrow cells Confluent MM-derived and DM-derived cells from rats were exposed to PDGF (10 ng/ml) for various time intervals (0–2 h). Cell lysates were evaluated by Western blot analysis, using two distinct polyclonal antibodies that detect either the total Erk-1 (42 KDa) and Erk-2 (44 KDa) or their phosphorylated forms. The results (Fig. 1) indicate that both DM and MM-derived cells treated with PDGF showed a significant increase in Erk1/2 phosphorylation, which peaks between 15 and 30 min and returns to basal levels by 2 h. The DM-derived osteoblasts treated with various concentrations of PDGF (0–30 ng/ml) for 30 min demonstrate a concentration-dependent increase in Erk1/2 phosphorylation (Fig 1B). Similar results were obtained for MM-derived cells (data not shown). Effect of FN on Erk1/2 phosphorylation in osteoblasts To determine if matrix molecules influenced Erk1/2 phosphorylation, and whether PDGF effects on Erk1/2 phosphorylation were dependent on cell matrix interactions, we next compared the Erk1/2 phosphorylation of DM-derived cells kept in suspension and of cells plated onto FN. Within 30 min of attachment to FN, a robust increase in Erk1/2 phosphorylation was observed that maintained for at least 8 h (Fig. 2A). There
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
575
Fig. 2. Effect of FN on Erk1/2 phosphorylation. (A) Time course of Erk1/2 phosphorylation. DM cells were grown to confluency, harvested by trypsin, and allowed to recover in suspension for 1 h in 2% BSA in a-MEM with 1% Pen/Strep. Cells were then plated on FN coated petri dishes in a-MEM containing bovine serum albumin (0.1% w/v) + Pen/Strep and Erk1/2 phosphorylation was evaluated at the indicated times. (B) Cell adhesion and Erk1/2 phosphorylation. In a separate experiment, cells were kept in suspension in 2% BSA for 1 h, were then plated on FN coated dishes for 1 h, then treated for an additional 1 h with PDGF (10 ng/ml), and the Erk1/2 phosphorylation was evaluated.
Fig. 1. PDGF leads to a strong and transient induction of Erk1/2 phosphorylation. (A) Effect of time. MM and DM cells were grown to confluency and treated with PDGF (10 ng/ml) o for indicated times, lysed by boiling in a buffer (1% SDS, 10 mM Tris–HCl, pH 7.4, and 1 mM Na3 VO4 ), sonicated, equal amounts of lysates were separated on a denaturing SDS–acrylamide gel, and analyzed by Western blot analysis, using antibodies that recognize the non-phosphorylated Erk or phosphorylated Erk as described (Materials and methods). The relative molecular weights were established by protein standards. (B) Effect of various PDGF concentrations. Erk1/2 phosphorylation was evaluated in confluent DM cells at various PDGF concentrations (0– 30 ng/ml) treated for 30 min. The autoradiographs were quantitated by densitometric scanning of the X-ray film. The results represent the average standard error of the mean from two independent experiments.
was no Erk1/2 phosphorylation in cells kept in suspension (Fig. 2B). PDGF elicited Erk1/2 phosphorylation response in both the cells attached to FN as well as in cells kept in suspension. The results suggest that while PDGF effects are transient (Fig. 1), FN effects were much more prolonged and persistent (Fig. 3A). These studies further demonstrate that PDGF stimulation of Erk1/2 phosphorylation is independent of cell matrix interaction and that the Erk1/2 activation by the two agents are likely to occur by independent pathways. Similar results were obtained for MM-derived cells (data not shown).
Fig. 3. Effect of kinase and phosphatase inhibitors on Erk1/2 phosphorylation. (A) Confluent layers of MM cells were pre-treated for 15 min with staurosporin (0.1 lM), calphostin (0.5 lM), genistein (50 lM), and LY294002 (40 lM), and continued with a further treatment of PDGF (10 ng/ml) for an additional 1.5 h. (B,C) Confluent layers of MM cells were pre-treated for 15 min with sodium 2 mM orthovanadate (B) or 0.5 lM okadaic acid (C), and further treated with PDGF (10 ng/ml) for 0.5 h or 1.5 h. The samples were analyzed by Western blot analysis as before.
Role of kinases and phosphatases in PDGF phosphorylation of Erk1/2 To further investigate whether the activation of Erk1/2 by PDGF involves kinase, phosphatase or both, DMderived cells were treated with PDGF in the presence of PKC inhibitors (0.1 lM staurosporine and 0.5 lM calphostin), a receptor tyrosine kinase inhibitor (50 lM genistein), a PI-3-kinase inhibitor (40 lM LY294002), and a protein phosphatase inhibitor (2 mM sodium orthovanadate). The results (Fig. 3A) show that PKC inhibitors and receptor tyrosine kinase inhibitors reduced Erk1/2 phosphorylation in confluent (or FN attached, data not shown), as well as PDGF-treated cells. Further, LY294002, a PI-3 kinase inhibitor was ineffective.
576
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
Finally, orthovanadate (Figs. 3B and C) alone elicited a strong Erk1/2 phosphorylation response and prevented the rapid dephosphorylation that occurred in response to PDGF (Fig. 3C). These results suggest that PKC and receptor tyrosine kinase activation play a key role in Erk1/2 activation by both PDGF and FN and that the phosphorylation levels are further regulated by serine– threonine protein phosphatases. Integrin induced Erk1/2 phosphorylation induction is reduced by PTH 1–34 We also evaluated the effect of PTH 1–34, a bone active agent, on Erk1/2 phosphorylation. The MM-derived cells were allowed to attach to FN for 3 h and were treated with PTH 1–34 (108 M) for 1 h. PTH 1–34 treatment resulted in time (Fig. 4A) and concentration (Fig. 4B) dependent reduction in Erk1/2 phosphorylation in MM-derived cells. The minimal effective PTH 1– 34 concentration was 108 M. PTH 1–34 had no effects on total Erk1/2 levels (Fig. 4A).
Fig. 5. cAMP mediates the downregulation of Erk1/2 phosphorylation by PTH (1–34). (A) MM cells attached to FN for 1 h were treated with PTH (1–34 or 3–34; 108 M), or forskolin (FSK) (105 M) for the indicated time intervals. (B) Cells were pre-treated in suspension with H89 (10 lM/DMSO) or DMSO alone for 30 min, plated on to FN coated petri dishes for 1 h, and were finally treated with vehicle or 108 M PTH (1–34) for 30 min. Lysates were evaluated by Western blot analysis.
PTH (1–34) downregulation of Erk1/2 phosphorylation involves cAMP mediated-PKA activation Since PTH 1–34 effects on osteoblasts are mediated through a specific GPCR that activates adenylate cyclase, we evaluated whether cAMP-mediated PKA activation plays a role on PTH inhibition of Erk1/2 phosphorylation. Confluent monolayer of MM cells were treated with PTH 1–34 (108 M), PTH 3– 34 (108 M), and forskolin (105 M) for 2 h, and Erk1/2 phosphorylation was evaluated. The results indicate (Fig. 5A) that agents capable of eliciting a cAMP response (PTH 1–34 and forskolin) reduced Erk1/2 phosphorylation within 15 min of incubation, whereas PTH 3–34, which does not elicit a cAMP response, had no effect on Erk1/2 phosphorylation. In order to further validate a role for cAMP in the inhibition of Erk1/2 phosphorylation, MM cells were preincubated for 30 min with 10 lM of H89 (an inhibitor of PKA activation) then treated for 30 min with PTH 1–34. The data (Fig. 5B) indicate that the PTH 1–34 mediated reduction of Erk1/2 phosphorylation was abolished by H89. Fig. 4. PTH (1–34) reduces the FN induced Erk1/2 phosporylation in metaphyseal marrow cells. (A) Effect of treatment time. Confluent layers of MM cells were released from the flasks, allowed to recover in suspension for 1 h in 2% BSA in a-MEM 1% Pen/Strep, were then plated on FN coated petri dishes for 3 h (time 0), and then treated with vehicle (4 mM HCl and 1 mg/ml BSA) or 108 M PTH (1–34), and at indicated time points, the samples were analyzed. (B) Effect of various concentrations. In a separate experiment, MM cells were treated with various concentrations (0.05–50 nM) of PTH (1–34). The phosphorylated Erk1/2 was quantitated by densitometric scanning of the X-ray film. The quantitative results are shown only for Erk1. Lane 1, control; lane 2, PTH.
Discussion Osteoblast proliferation and/or differentiation are influenced by a variety of factors that act via distinct receptors and corresponding signal transduction pathways [1–10]. We particularly focused on Erk1/2 activation in response to PDGF (receptor tyrosine kinase), PTH (GPCR), and FN (integrin) mediated signaling pathways and evaluated whether Erk1/2 activation can serve as common intermediate step among these agents.
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
Our results indicate that Erk1/2 phosphorylation might serve as a common intermediate among distinct signal transducing pathways that ultimately lead to osteoblast proliferation and differentiation. We utilized MM-derived and DM-derived osteoblasts/osteoprogenitor cells freshly isolated from young rats. These cells have been extensively characterized and display characteristics of bone forming cells [2,3]. In particular, MM-derived cells are responsive to various growth factors in vitro, and are the cells that are primarily responsive to PTH injection in vivo [2,3,16]. Erk1 (42 kDa) and Erk2 (44 kDa) activation was evaluated by Western blot analysis of the cell extracts, using specific antibodies that recognize either the total proteins or the phosphorylated forms. The total levels of Erk1/2 were also measured to determine whether a change in phosphorylation was due to a decrease in total protein levels. In all studies reported here there was very little change in the total Erk1/2 levels. Therefore, with the exception of the data shown in Figs. 1 and 5, the results are shown only for the phosphorylated Erk1/2. Finally, Erk1/2 phosphorylation was evaluated on cells that were either confluent (Figs. 1, 2, and 4) or on cells freshly plated onto FN coated dishes (Figs. 3 and 5). There was very little difference between the two conditions, perhaps because the confluent cells synthesize and secrete FN and other molecules into the matrix during the course of culture conditions (1 week in culture) and allow cells to respond in a manner similar to those freshly plated onto FN coated dishes. PDGF and FN both caused a dramatic increase in Erk1/2 phosphorylation. TGF-b that signals via a serine/threonine kinase receptor, was ineffective (data not shown) However, there was a clear difference in the duration of expression of phosphorylated Erk1/2. While the effects of PDGF were robust and transient lasting for only 1 h, the effects of FN were persistent for at least up to 8 h. These observations are similar to those made in PC12 cells, where epidermal growth factor and matrix proteins exert distinct time requirements for differentiation [17–19]. Although we do not know the reason for this difference, it is possible that differences in protein phosphatase levels may play a role in the phosphorylated Erk1/2 levels. It is also possible that PDGF receptor and integrin may be internalized at different times and rates. Irrespective of the reason, the transient activation by PDGF may primarily serve as a trigger for proliferative response, while matrix–integrin interactions may serve as an anti-apoptotic and differentiation signal as suggested for other systems [9,20]. Although there are areas of convergence, our results do indicate that Erk1/2 phosphorylation by PDGF and FN are additive (data not shown), and can occur independent of each other (Fig. 2). The results are consistent with previous observations on co-operativity between integrin and growth factors [21,22]. For example, cells in sus-
577
pension (without any exogenously added FN) still were able to demonstrate Erk1/2 phosphorylation, suggesting that PDGF stimulation of Erk1/2 phosphorylation does not require cell–matrix interaction. Other studies have suggested that PDGF receptor couples to grb2/SOS pathway to Ras, while integrin activation may additionally utilize shc/grb2/Ras activation, all leading to Erk1/2 activation [20,23,24]. The prolonged activation may also involve a suppression of MAPKP phosphatases, as described for vascular smooth muscle cells [25]. In contrast to PDGF and FN, PTH 1–34, an agent that is known to be anabolic as well as catabolic to bone in various animals and humans, reduced Erk1/2 phosphorylation elicited by FN within a short exposure time (Fig. 4) or PDGF (data not shown). In addition to these in vitro data, in an in vivo study not reported here, cells freshly isolated from PTH 1–34 treated rats failed to demonstrate Erk1/2 phosphorylation upon binding to FN (Kaiser and Chandrasekhar, unpublished data). The reduction in signal is mediated through the cAMP pathway since forskolin, a non-specific activator of adenylate cyclase, also caused a decrease in Erk1/2 phosphorylation. The effects of forskolin and PTH 1–34 were reversed in the presence of H-89, an agent that blocks cAMP-dependent PKA activation. Finally, PTH 3–34, an N-terminally truncated form that does not evoke a cAMP response in these cells was not effective. These results confirm previous observations of PTH and cAMP on Erk1/2 phosphorylation and further extend that PTH inactivation of Erk1/2 occurs independent of activation pathways [19,26,27]. The mechanism of cAMP-dependent Erk1/2 inactivation is not clear. It may involve blocking the Ras-dependent activation of Raf-1 [28,29] or induction of the dual specific phosphatase MAP kinase phosphatase-1 [30]. It is now well recognized that the proliferation and differentiation of cells is trans-regulated by a variety of factors. Although all three agents (PDGF, FN, and PTH) are recognized as osteotrophic factors, the effects on Erk1/2 phosphorylation exhibit similarities and clear differences. It is possible that while PDGF may primarily serve to regulate cell proliferation, FN may play a role either in prevention of apoptosis or differentiation or both. PTH 1–34 on the other hand may have a primary role in cell differentiation that ultimately is responsible for the organization of new bone. Although each of these agents signal through specific receptors, the ultimate cellular response may be a consequence of selective temporal regulation of a convergent intermediary signal, such as Erk1/2.
Acknowledgments The authors acknowledge Drs. Charles Frolik and Venkatesh Krishnan for critical evaluation of the manuscript.
578
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
References [1] C.S. Tam, J.N. Heersche, T.M. Murray, J.A. Parsons, Endocrinology 110 (1982) 506–512. [2] J.E. Onyia, J. Bidwell, J. Herring, J. Hulman, J.M. Hock, Bone 17 (1995) 479–484. [3] J.E. Onyia, B. Miller, J. Hulman, J. Liang, R. Galvin, C. Frolik, S. Chandrasekhar, A.K. Harvey, J. Bidwell, J. Herring, J.M. Hock, Bone 20 (1997) 93–100. [4] G.R. Mundy, Clin. Orthop. Relat. Res. (1996) 24–28. [5] T.A. Linkhart, S. Mohan, D.J. Baylink, Bone 19 (1996) 1S–12S. [6] E. Canalis, S. Varghese, T.L. McCarthy, M. Centrella, Growth Regul. 2 (1992) 151–155. [7] M.E. Antosz, C.G. Bellows, J.E. Aubin, J. Cell. Physiol. 140 (1989) 386–395. [8] X. Yu, S.C. Hsieh, W. Bao, D.T. Graves, Am. J. Physiol. 272 (1997) C1709–C1716. [9] A.M. Moursi, R.K. Globus, C.H. Damsky, J. Cell Sci. 110 (1997) 2187–2196. [10] D.R. Ganta, M.B. McCarthy, G.A. Gronowicz, Endocrinology 138 (1997) 3606–3612. [11] N.G. Ahn, R. Seger, E.G. Krebs, Curr. Opin. Cell Biol. 4 (1992) 992–999. [12] S.L. Pelech, J.S. Sanghera, Science 257 (1992) 1355–1356. [13] J. Blenis, Proc. Natl. Acad. Sci. (USA) 90 (1993) 5889–5892. [14] R.J. Davis, J. Biol. Chem. 268 (1993) 14553–14556. [15] M.H. Cobb, E.J. Goldsmith, J. Biol. Chem. 270 (1995) 14843– 14846.
[16] E.A. Kaiser, M. Sato, J.E. Onyia, S. Chandrasekhar, J. Cell. Biochem. 83 (2001) 617–630. [17] C.J. Marshall, Cell 80 (1995) 179–185. [18] S. Traverse, N. Gomez, H. Paterson, C. Marshall, P. Cohen, Biochem. J. 288 (1992) 351–355. [19] L.R. Chaudhary, L.V. Avioli, Mol. Cell. Biochem. 178 (1998) 59– 68. [20] A. Howe, A.E. Aplin, S.K. Alahari, R.L. Juliano, Curr. Opin. Cell Biol. 10 (1998) 220–231. [21] M. Schneller, K. Vuori, E. Ruoslahti, EMBO J. 16 (1997) 5600– 5607. [22] S. Miyamoto, H. Teramoto, J.S. Gutkind, K.M. Yamada, J. Cell Biol. 135 (1996) 1633–1642. [23] K.E. Bornfeldt, E.W. Raines, L.M. Graves, M.P. Skinner, E.G. Krebs, R. Ross, Ann. NY Acad. Sci. 766 (1995) 416–430. [24] D.D. Schlaepfer, K.C. Jones, T. Hunter, Mol. Cell. Biol. 18 (1998) 2571–2585. [25] J.L. Duff, B.P. Monia, B.C. Berk, J. Biol. Chem. 270 (1995) 7161– 7166. [26] M.H. Verheijen, L.H. Defize, J. Biol. Chem. 272 (1997) 3423– 3429. [27] S.R. Siddhanti, J.E. Hartle II, L.D. Quarles, Endocrinology 136 (1995) 4834–4841. [28] R. Plevin, K. Malarkey, D. Aidulis, A. McLees, G.W. Gould, Cell. Signal. 9 (1997) 323–328. [29] S.J. Cook, F. McCormick, Science 262 (1993) 1069–1072. [30] M. Kurino, K. Fukunaga, Y. Ushio, E. Miyamoto, J. Neurochem. 67 (1996) 2246–2255.
Distinct pathways of extracellular signal-regulated kinase activation by growth factors, fibronectin and parathyroid hormone 1–34q Edelgard Kaiser1 and Srinivasan Chandrasekhar* Gene Regulation, Bone and Inflammation Research, Lilly Research Laboratories, DC 0403, Eli Lilly and Company, Corporate Center, Indianapolis, IN 46285, USA Received 11 April 2003
Abstract Growth factors, hormones, and matrix proteins regulate osteoblast proliferation and differentiation, acting through cognate receptors. Since each of the receptors are coupled to a variety of distinct signal transduction pathways, in this report we evaluated whether there is a common convergent intermediate step that allows cross-talk among the various pathways. Since extracellular signal-regulated kinases 1 and 2 (Erk1/2) play a role in mitogenesis and differentiation processes, we evaluated the effects of various osteotrophic factors on Erk1/2 phosphorylation in osteoblasts. Osteoblasts isolated from the metaphyseal marrow (MM) and diaphyseal marrow (DM) of 4–6 week old male rat longitudinal bones were grown to confluency and Erk1/2-phosphorylation was evaluated using antibodies that recognized either the total or the phosphorylated form of the kinase. There was very little Erk1/2 phosphorylation in cells kept in suspension. Both MM and DM cells attached to fibronectin (FN), demonstrated Erk1/2 phosphorylation that persisted for at least up to 8 h. Platelet-derived growth factor AB (PDGF-AB) induced a transient and robust Erk1/ 2 phosphorylation that was attenuated by 2 h. Studies with specific inhibitors indicated that the effects of these factors were mediated by protein kinase C, by receptor tyrosine kinase, as well as by protein phosphatases. Parathyroid hormone (PTH 1–34), a bone anabolic agent however, caused a down-regulation of FN stimulated Erk1/2 phosphorylation in MM derived cells. The inhibitory effect of PTH was mediated through cAMP-dependent protein kinase A (PKA) activation. The data collectively suggest that a combination of diverse extracellular stimuli regulates Erk1/2 phosphorylation that may ultimately influence osteoblast proliferation and/or differentiation. Ó 2003 Elsevier Science (USA). All rights reserved.
Bone is a dynamic tissue that is continually repaired and remodeled in response to a variety of signals. The interaction between osteoblasts and osteoclasts is responsible for maintaining bone homeostasis. The regulation of osteoblast proliferation and differentiation is complex, and is subject to modulation by a variety of osteotrophic factors such as PTH, bone morphogenetic proteins, TGFb1, PDGF, and extracellular matrix proteins [1–10]. However, it is not clear how these multiple q
Abbreviations: PTH, parathyroid hormone; FN, fibronectin; MM, metaphyseal marrow; DM, diaphyseal marrow; Erk1/2, extracellular signal-regulated kinase 1/2; TGF-b, transforming growth factor b; PDGF, platelet-derived growth factor. * Corresponding author. Fax: 1-317-276-9722. E-mail address: [email protected] (S. Chandrasekhar). 1 Present address: Department of Bone Pathology, Center for Biomechanics/UKE, Hamburg University, Martinistr. 52, 20246 Hamburg, Germany.
signals are transduced and coordinated that ultimately result in osteoblast formation and differentiation. Mitogen-activated protein kinases, such as the p42/p44 serine/theronine extracellular signal-regulated kinases (Erk 1 and 2), play a central role in the proliferation and gene transcription in a variety of cells and in response to a variety of signals [11–15]. In this report, we evaluated whether Erk1/2 might be a key intermediary in osteoblast differentiation stimulated by various factors. Osteoprogenitor cells were isolated from the metaphyseal and diaphyseal marrow of young rats and were evaluated for Erk1/2 activation after in vitro treatment with PDGF-AB, PTH 1–34, and FN. The results indicate that FN stimulated Erk1/2 phosphorylation in osteoprogenitors was prolonged, while PDGF treatment resulted in a robust, but transient activation of Erk1/2. Erk1/2 phosphorylation, stimulated with PDGF or FN, was completely abolished by PTH 1–34. These results
0006-291X/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0006-291X(03)00820-9
574
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
suggest that the Erk1/Erk2 phosphorylation may serve as a converging point that links a variety of osteotrophic signals.
Methods Animals. All of the studies were done using 4–6 week old (75– 100 g), Sprague–Dawley male rats (Harlan, Indianapolis, IN). The rats were fed ad libitum with Purina chow (calcium 1%, phosphate 0.61%, and PMI Feeds, St. Louis, MO) and water. Rats were sacrificed and the bones were used for isolation of cells as described below. All animal protocols were approved by the Lilly Animal Care and Use Committee. Isolation of osteoblasts/osteoprogenitors. MM and DM-derived osteoprogenitor/osteoblasts were isolated from the femurs of rats as described [3]. In brief, freshly harvested bones were kept in a-minimal essential media (a-MEM) + 10% fetal bovine serum (FBS) + 5% penicillin (Pen) and streptomycin (Strep), rinsed for 10 s in 70% ethanol, followed by several rinses with HankÕs balanced salt solution (Mg2þ / Ca2þ -free). For the isolation of the MM-derived osteoblasts, about 3 mm of the metaphyses subjacent to the growth plate were cut off, minced, and digested for an hour at 37 °C with 0.6 ml of 0.25% trypsin in HankÕs balanced salt solution (Mg2þ /Ca2þ -free solution) for each femur, centrifuged, and resuspended in a-MEM + 10% FBS + 1% Pen/Strep. The debris was separated from the cells with a sieve and the cells were plated in a-MEM + 10% FBS + 1% Pen/Strep. For the isolation of the DM-derived osteoblasts, both ends of the femur were cut off and the bone marrow was flushed out with phosphate-buffered saline (Mg2þ / Ca2þ -free), using a syringe. The cells were centrifuged and plated in aMEM + 10% FBS + 1% Pen/Strep and incubated at 37 °C (5% CO2 + air). The non-adherent population was removed by aspiration after 24–48 h and the cultures were fed with a-MEM + 10% FBS + 1% Pen/Strep. The media were changed every fourth day thereafter. Only the first passage cells were used for all studies described. Reagents. Growth factors (PDGF-AB or TGF-b1) were obtained from R&D Systems (Minneapolis, MN) and rat PTH 1–34 and 3–34 were obtained from Bachem (Torrance, CA). The following (H-89, genistein, calphostin, staurosporine, forskolin, sodium orthovanadate, and okadaic acid) were purchased from Calbiochem (La Jolla, Ca). LY294002, a PI-3 kinase inhibitor was a gift from Dr. Chris Vlahos, Eli Lilly and Company. The proteins were solubilized in 4 mM HCl (for PTH, 1 mM) containing 1% bovine serum albumin (BSA) + 0.15 M NaCl, pH 7.2. The stock solutions of the compounds were prepared in dimethyl sulfoxide (1–10 mM) and then diluted into 4 mM HCl (or 1 mM HCl) containing 1% BSA + 0.15 M NaCl, pH 7.2, before addition. Treatments. Erk1/2 levels (both total and phosphorylated Erk1/2) were analyzed in cells treated with various factors at confluency or after attachment to FN coated onto tissue culture plates. Cells were grown to near confluency in six well plates and washed three times with PBS (Mg2þ /Ca2þ -free) and treated with growth factors or PTH 1–34 in a-MEM + 1% Pen/Strep + 0.1% BSA. For experiments involving inhibitors, cells were pre-treated for 15 min with various compounds, followed by the addition of growth factors or PTH 1–34 at indicated concentrations and time. For assays involving attachment to FN, cells grown to near confluency were first serum-deprived for 24 h with aMEM + 1% FBS + 1% Pen/Strep, were washed twice with PBS (Mg2þ / Ca2þ -free), and released by EDTA treatment (Versene 1:5000, GibcoBRL, Grand Island, NY). Cells were kept in suspension for 1 h at 37 °C in a-MEM containing 2% BSA, medium was changed to aMEM containing 0.1% BSA prior to plating of the cells onto tissue culture dishes, previously coated with FN (10 lg/ml in PBS at 4 °C for 16 h). Western blot analysis of Erk1/2. After treatments, the media were removed, the attached/confluent cells were scraped and lysed by boiling
in SDS buffer (10 mM Tris–HCl, pH 7.4, 1% SDS, and 1 mM Na3 VO4 ), boiled an additional 5 min, and then cooled down on ice. The volumes of all samples were kept constant and the lysates were then sonicated. Equal volume of the lysates was mixed with sample buffer containing dithiothreitol, the samples were subjected to electrophoresis on a denaturing SDS–polyacryamide gel (7% resolving/3% stacking), and transferred to 0.45 l nitrocellulose filter (Pharmacia, San Francisco, CA). The membrane was blocked with 5% blotto (nonfat dried milk in 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20) for 1 h at 25 °C, incubated with 1:1000 dilution of the rabbit primary antibody (phospho-specific P44/42 MAPK antibody, New England BioLabs, MA) for 1 h at room temperature, washed four times with 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20 for 15 min, incubated with 1:2000 dilution of horseradish peroxidaseconjugated secondary antibody (Southern Biotechnology, Birmingham, AL) for 1 h at 25 °C, and again washed four times as above. For the assessment of relative amounts of total Erk1 and two independent of their phosphorylation status, filters were stripped for 30 min at room temperature with a buffer containing 10% SDS, 62.5 mM Tris–HCl, pH 6.7, and 100 mM 2-mercaptoethanol, extensively washed with 10 mM Tris–HCl, pH 8, 130 mM NaCl with 0.1% Tween 20, and blocked with 5% blotto. The primary antibody (rabbit anti rat antibody, New England BioLabs, MA) was used at 1:5000 dilution (5% blotto), while the anti-rabbit horseradish peroxidase-conjugated secondary antibody was used at 1:10,000 dilution. The immunoreactive bands were visualized using the ECL detection system (Amersham). The relative amount of the 42 kDa (Erk1) and 44 kDa (Erk2) was quantitated by densitometric scanning of the X-ray film.
Results PDGF transiently induces Erk1/2 phosphorylation in confluent metaphyseal and diaphyseal bone marrow cells Confluent MM-derived and DM-derived cells from rats were exposed to PDGF (10 ng/ml) for various time intervals (0–2 h). Cell lysates were evaluated by Western blot analysis, using two distinct polyclonal antibodies that detect either the total Erk-1 (42 KDa) and Erk-2 (44 KDa) or their phosphorylated forms. The results (Fig. 1) indicate that both DM and MM-derived cells treated with PDGF showed a significant increase in Erk1/2 phosphorylation, which peaks between 15 and 30 min and returns to basal levels by 2 h. The DM-derived osteoblasts treated with various concentrations of PDGF (0–30 ng/ml) for 30 min demonstrate a concentration-dependent increase in Erk1/2 phosphorylation (Fig 1B). Similar results were obtained for MM-derived cells (data not shown). Effect of FN on Erk1/2 phosphorylation in osteoblasts To determine if matrix molecules influenced Erk1/2 phosphorylation, and whether PDGF effects on Erk1/2 phosphorylation were dependent on cell matrix interactions, we next compared the Erk1/2 phosphorylation of DM-derived cells kept in suspension and of cells plated onto FN. Within 30 min of attachment to FN, a robust increase in Erk1/2 phosphorylation was observed that maintained for at least 8 h (Fig. 2A). There
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
575
Fig. 2. Effect of FN on Erk1/2 phosphorylation. (A) Time course of Erk1/2 phosphorylation. DM cells were grown to confluency, harvested by trypsin, and allowed to recover in suspension for 1 h in 2% BSA in a-MEM with 1% Pen/Strep. Cells were then plated on FN coated petri dishes in a-MEM containing bovine serum albumin (0.1% w/v) + Pen/Strep and Erk1/2 phosphorylation was evaluated at the indicated times. (B) Cell adhesion and Erk1/2 phosphorylation. In a separate experiment, cells were kept in suspension in 2% BSA for 1 h, were then plated on FN coated dishes for 1 h, then treated for an additional 1 h with PDGF (10 ng/ml), and the Erk1/2 phosphorylation was evaluated.
Fig. 1. PDGF leads to a strong and transient induction of Erk1/2 phosphorylation. (A) Effect of time. MM and DM cells were grown to confluency and treated with PDGF (10 ng/ml) o for indicated times, lysed by boiling in a buffer (1% SDS, 10 mM Tris–HCl, pH 7.4, and 1 mM Na3 VO4 ), sonicated, equal amounts of lysates were separated on a denaturing SDS–acrylamide gel, and analyzed by Western blot analysis, using antibodies that recognize the non-phosphorylated Erk or phosphorylated Erk as described (Materials and methods). The relative molecular weights were established by protein standards. (B) Effect of various PDGF concentrations. Erk1/2 phosphorylation was evaluated in confluent DM cells at various PDGF concentrations (0– 30 ng/ml) treated for 30 min. The autoradiographs were quantitated by densitometric scanning of the X-ray film. The results represent the average standard error of the mean from two independent experiments.
was no Erk1/2 phosphorylation in cells kept in suspension (Fig. 2B). PDGF elicited Erk1/2 phosphorylation response in both the cells attached to FN as well as in cells kept in suspension. The results suggest that while PDGF effects are transient (Fig. 1), FN effects were much more prolonged and persistent (Fig. 3A). These studies further demonstrate that PDGF stimulation of Erk1/2 phosphorylation is independent of cell matrix interaction and that the Erk1/2 activation by the two agents are likely to occur by independent pathways. Similar results were obtained for MM-derived cells (data not shown).
Fig. 3. Effect of kinase and phosphatase inhibitors on Erk1/2 phosphorylation. (A) Confluent layers of MM cells were pre-treated for 15 min with staurosporin (0.1 lM), calphostin (0.5 lM), genistein (50 lM), and LY294002 (40 lM), and continued with a further treatment of PDGF (10 ng/ml) for an additional 1.5 h. (B,C) Confluent layers of MM cells were pre-treated for 15 min with sodium 2 mM orthovanadate (B) or 0.5 lM okadaic acid (C), and further treated with PDGF (10 ng/ml) for 0.5 h or 1.5 h. The samples were analyzed by Western blot analysis as before.
Role of kinases and phosphatases in PDGF phosphorylation of Erk1/2 To further investigate whether the activation of Erk1/2 by PDGF involves kinase, phosphatase or both, DMderived cells were treated with PDGF in the presence of PKC inhibitors (0.1 lM staurosporine and 0.5 lM calphostin), a receptor tyrosine kinase inhibitor (50 lM genistein), a PI-3-kinase inhibitor (40 lM LY294002), and a protein phosphatase inhibitor (2 mM sodium orthovanadate). The results (Fig. 3A) show that PKC inhibitors and receptor tyrosine kinase inhibitors reduced Erk1/2 phosphorylation in confluent (or FN attached, data not shown), as well as PDGF-treated cells. Further, LY294002, a PI-3 kinase inhibitor was ineffective.
576
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
Finally, orthovanadate (Figs. 3B and C) alone elicited a strong Erk1/2 phosphorylation response and prevented the rapid dephosphorylation that occurred in response to PDGF (Fig. 3C). These results suggest that PKC and receptor tyrosine kinase activation play a key role in Erk1/2 activation by both PDGF and FN and that the phosphorylation levels are further regulated by serine– threonine protein phosphatases. Integrin induced Erk1/2 phosphorylation induction is reduced by PTH 1–34 We also evaluated the effect of PTH 1–34, a bone active agent, on Erk1/2 phosphorylation. The MM-derived cells were allowed to attach to FN for 3 h and were treated with PTH 1–34 (108 M) for 1 h. PTH 1–34 treatment resulted in time (Fig. 4A) and concentration (Fig. 4B) dependent reduction in Erk1/2 phosphorylation in MM-derived cells. The minimal effective PTH 1– 34 concentration was 108 M. PTH 1–34 had no effects on total Erk1/2 levels (Fig. 4A).
Fig. 5. cAMP mediates the downregulation of Erk1/2 phosphorylation by PTH (1–34). (A) MM cells attached to FN for 1 h were treated with PTH (1–34 or 3–34; 108 M), or forskolin (FSK) (105 M) for the indicated time intervals. (B) Cells were pre-treated in suspension with H89 (10 lM/DMSO) or DMSO alone for 30 min, plated on to FN coated petri dishes for 1 h, and were finally treated with vehicle or 108 M PTH (1–34) for 30 min. Lysates were evaluated by Western blot analysis.
PTH (1–34) downregulation of Erk1/2 phosphorylation involves cAMP mediated-PKA activation Since PTH 1–34 effects on osteoblasts are mediated through a specific GPCR that activates adenylate cyclase, we evaluated whether cAMP-mediated PKA activation plays a role on PTH inhibition of Erk1/2 phosphorylation. Confluent monolayer of MM cells were treated with PTH 1–34 (108 M), PTH 3– 34 (108 M), and forskolin (105 M) for 2 h, and Erk1/2 phosphorylation was evaluated. The results indicate (Fig. 5A) that agents capable of eliciting a cAMP response (PTH 1–34 and forskolin) reduced Erk1/2 phosphorylation within 15 min of incubation, whereas PTH 3–34, which does not elicit a cAMP response, had no effect on Erk1/2 phosphorylation. In order to further validate a role for cAMP in the inhibition of Erk1/2 phosphorylation, MM cells were preincubated for 30 min with 10 lM of H89 (an inhibitor of PKA activation) then treated for 30 min with PTH 1–34. The data (Fig. 5B) indicate that the PTH 1–34 mediated reduction of Erk1/2 phosphorylation was abolished by H89. Fig. 4. PTH (1–34) reduces the FN induced Erk1/2 phosporylation in metaphyseal marrow cells. (A) Effect of treatment time. Confluent layers of MM cells were released from the flasks, allowed to recover in suspension for 1 h in 2% BSA in a-MEM 1% Pen/Strep, were then plated on FN coated petri dishes for 3 h (time 0), and then treated with vehicle (4 mM HCl and 1 mg/ml BSA) or 108 M PTH (1–34), and at indicated time points, the samples were analyzed. (B) Effect of various concentrations. In a separate experiment, MM cells were treated with various concentrations (0.05–50 nM) of PTH (1–34). The phosphorylated Erk1/2 was quantitated by densitometric scanning of the X-ray film. The quantitative results are shown only for Erk1. Lane 1, control; lane 2, PTH.
Discussion Osteoblast proliferation and/or differentiation are influenced by a variety of factors that act via distinct receptors and corresponding signal transduction pathways [1–10]. We particularly focused on Erk1/2 activation in response to PDGF (receptor tyrosine kinase), PTH (GPCR), and FN (integrin) mediated signaling pathways and evaluated whether Erk1/2 activation can serve as common intermediate step among these agents.
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
Our results indicate that Erk1/2 phosphorylation might serve as a common intermediate among distinct signal transducing pathways that ultimately lead to osteoblast proliferation and differentiation. We utilized MM-derived and DM-derived osteoblasts/osteoprogenitor cells freshly isolated from young rats. These cells have been extensively characterized and display characteristics of bone forming cells [2,3]. In particular, MM-derived cells are responsive to various growth factors in vitro, and are the cells that are primarily responsive to PTH injection in vivo [2,3,16]. Erk1 (42 kDa) and Erk2 (44 kDa) activation was evaluated by Western blot analysis of the cell extracts, using specific antibodies that recognize either the total proteins or the phosphorylated forms. The total levels of Erk1/2 were also measured to determine whether a change in phosphorylation was due to a decrease in total protein levels. In all studies reported here there was very little change in the total Erk1/2 levels. Therefore, with the exception of the data shown in Figs. 1 and 5, the results are shown only for the phosphorylated Erk1/2. Finally, Erk1/2 phosphorylation was evaluated on cells that were either confluent (Figs. 1, 2, and 4) or on cells freshly plated onto FN coated dishes (Figs. 3 and 5). There was very little difference between the two conditions, perhaps because the confluent cells synthesize and secrete FN and other molecules into the matrix during the course of culture conditions (1 week in culture) and allow cells to respond in a manner similar to those freshly plated onto FN coated dishes. PDGF and FN both caused a dramatic increase in Erk1/2 phosphorylation. TGF-b that signals via a serine/threonine kinase receptor, was ineffective (data not shown) However, there was a clear difference in the duration of expression of phosphorylated Erk1/2. While the effects of PDGF were robust and transient lasting for only 1 h, the effects of FN were persistent for at least up to 8 h. These observations are similar to those made in PC12 cells, where epidermal growth factor and matrix proteins exert distinct time requirements for differentiation [17–19]. Although we do not know the reason for this difference, it is possible that differences in protein phosphatase levels may play a role in the phosphorylated Erk1/2 levels. It is also possible that PDGF receptor and integrin may be internalized at different times and rates. Irrespective of the reason, the transient activation by PDGF may primarily serve as a trigger for proliferative response, while matrix–integrin interactions may serve as an anti-apoptotic and differentiation signal as suggested for other systems [9,20]. Although there are areas of convergence, our results do indicate that Erk1/2 phosphorylation by PDGF and FN are additive (data not shown), and can occur independent of each other (Fig. 2). The results are consistent with previous observations on co-operativity between integrin and growth factors [21,22]. For example, cells in sus-
577
pension (without any exogenously added FN) still were able to demonstrate Erk1/2 phosphorylation, suggesting that PDGF stimulation of Erk1/2 phosphorylation does not require cell–matrix interaction. Other studies have suggested that PDGF receptor couples to grb2/SOS pathway to Ras, while integrin activation may additionally utilize shc/grb2/Ras activation, all leading to Erk1/2 activation [20,23,24]. The prolonged activation may also involve a suppression of MAPKP phosphatases, as described for vascular smooth muscle cells [25]. In contrast to PDGF and FN, PTH 1–34, an agent that is known to be anabolic as well as catabolic to bone in various animals and humans, reduced Erk1/2 phosphorylation elicited by FN within a short exposure time (Fig. 4) or PDGF (data not shown). In addition to these in vitro data, in an in vivo study not reported here, cells freshly isolated from PTH 1–34 treated rats failed to demonstrate Erk1/2 phosphorylation upon binding to FN (Kaiser and Chandrasekhar, unpublished data). The reduction in signal is mediated through the cAMP pathway since forskolin, a non-specific activator of adenylate cyclase, also caused a decrease in Erk1/2 phosphorylation. The effects of forskolin and PTH 1–34 were reversed in the presence of H-89, an agent that blocks cAMP-dependent PKA activation. Finally, PTH 3–34, an N-terminally truncated form that does not evoke a cAMP response in these cells was not effective. These results confirm previous observations of PTH and cAMP on Erk1/2 phosphorylation and further extend that PTH inactivation of Erk1/2 occurs independent of activation pathways [19,26,27]. The mechanism of cAMP-dependent Erk1/2 inactivation is not clear. It may involve blocking the Ras-dependent activation of Raf-1 [28,29] or induction of the dual specific phosphatase MAP kinase phosphatase-1 [30]. It is now well recognized that the proliferation and differentiation of cells is trans-regulated by a variety of factors. Although all three agents (PDGF, FN, and PTH) are recognized as osteotrophic factors, the effects on Erk1/2 phosphorylation exhibit similarities and clear differences. It is possible that while PDGF may primarily serve to regulate cell proliferation, FN may play a role either in prevention of apoptosis or differentiation or both. PTH 1–34 on the other hand may have a primary role in cell differentiation that ultimately is responsible for the organization of new bone. Although each of these agents signal through specific receptors, the ultimate cellular response may be a consequence of selective temporal regulation of a convergent intermediary signal, such as Erk1/2.
Acknowledgments The authors acknowledge Drs. Charles Frolik and Venkatesh Krishnan for critical evaluation of the manuscript.
578
E. Kaiser, S. Chandrasekhar / Biochemical and Biophysical Research Communications 305 (2003) 573–578
References [1] C.S. Tam, J.N. Heersche, T.M. Murray, J.A. Parsons, Endocrinology 110 (1982) 506–512. [2] J.E. Onyia, J. Bidwell, J. Herring, J. Hulman, J.M. Hock, Bone 17 (1995) 479–484. [3] J.E. Onyia, B. Miller, J. Hulman, J. Liang, R. Galvin, C. Frolik, S. Chandrasekhar, A.K. Harvey, J. Bidwell, J. Herring, J.M. Hock, Bone 20 (1997) 93–100. [4] G.R. Mundy, Clin. Orthop. Relat. Res. (1996) 24–28. [5] T.A. Linkhart, S. Mohan, D.J. Baylink, Bone 19 (1996) 1S–12S. [6] E. Canalis, S. Varghese, T.L. McCarthy, M. Centrella, Growth Regul. 2 (1992) 151–155. [7] M.E. Antosz, C.G. Bellows, J.E. Aubin, J. Cell. Physiol. 140 (1989) 386–395. [8] X. Yu, S.C. Hsieh, W. Bao, D.T. Graves, Am. J. Physiol. 272 (1997) C1709–C1716. [9] A.M. Moursi, R.K. Globus, C.H. Damsky, J. Cell Sci. 110 (1997) 2187–2196. [10] D.R. Ganta, M.B. McCarthy, G.A. Gronowicz, Endocrinology 138 (1997) 3606–3612. [11] N.G. Ahn, R. Seger, E.G. Krebs, Curr. Opin. Cell Biol. 4 (1992) 992–999. [12] S.L. Pelech, J.S. Sanghera, Science 257 (1992) 1355–1356. [13] J. Blenis, Proc. Natl. Acad. Sci. (USA) 90 (1993) 5889–5892. [14] R.J. Davis, J. Biol. Chem. 268 (1993) 14553–14556. [15] M.H. Cobb, E.J. Goldsmith, J. Biol. Chem. 270 (1995) 14843– 14846.
[16] E.A. Kaiser, M. Sato, J.E. Onyia, S. Chandrasekhar, J. Cell. Biochem. 83 (2001) 617–630. [17] C.J. Marshall, Cell 80 (1995) 179–185. [18] S. Traverse, N. Gomez, H. Paterson, C. Marshall, P. Cohen, Biochem. J. 288 (1992) 351–355. [19] L.R. Chaudhary, L.V. Avioli, Mol. Cell. Biochem. 178 (1998) 59– 68. [20] A. Howe, A.E. Aplin, S.K. Alahari, R.L. Juliano, Curr. Opin. Cell Biol. 10 (1998) 220–231. [21] M. Schneller, K. Vuori, E. Ruoslahti, EMBO J. 16 (1997) 5600– 5607. [22] S. Miyamoto, H. Teramoto, J.S. Gutkind, K.M. Yamada, J. Cell Biol. 135 (1996) 1633–1642. [23] K.E. Bornfeldt, E.W. Raines, L.M. Graves, M.P. Skinner, E.G. Krebs, R. Ross, Ann. NY Acad. Sci. 766 (1995) 416–430. [24] D.D. Schlaepfer, K.C. Jones, T. Hunter, Mol. Cell. Biol. 18 (1998) 2571–2585. [25] J.L. Duff, B.P. Monia, B.C. Berk, J. Biol. Chem. 270 (1995) 7161– 7166. [26] M.H. Verheijen, L.H. Defize, J. Biol. Chem. 272 (1997) 3423– 3429. [27] S.R. Siddhanti, J.E. Hartle II, L.D. Quarles, Endocrinology 136 (1995) 4834–4841. [28] R. Plevin, K. Malarkey, D. Aidulis, A. McLees, G.W. Gould, Cell. Signal. 9 (1997) 323–328. [29] S.J. Cook, F. McCormick, Science 262 (1993) 1069–1072. [30] M. Kurino, K. Fukunaga, Y. Ushio, E. Miyamoto, J. Neurochem. 67 (1996) 2246–2255.