PTHrP Modulates Chondrocyte Differentiation through AP-1 and CREB Signaling*

During the process of differentiation, chondrocytes integrate a complex array of signals from local or systemic factors like parathyroid hormone-related peptide (PTHrP), Indian hedgehog, bone morphogenetic proteins and transforming growth factor β. While PTHrP is known to be a critical regulator of chondrocyte proliferation and differentiation, the signaling pathways through which this factor acts remain to be elucidated. Here we show that both cAMP response element-binding protein (CREB) and AP-1 activation are critical to PTHrP signaling in chondrocytes. PTHrP treatment leads to rapid CREB phosphorylation and activation, while CREB DNA binding activity is constitutive. In contrast, PTHrP induces AP-1 DNA binding activity through induction of c-Fos protein expression. PTHrP activates CRE and TRE reporter constructs primarily through PKA-mediated signaling events. Both signaling pathways were found to be important mediators of PTHrP effects on chondrocyte phenotype. Alone, PTHrP suppresses maturation and stimulates proliferation of the chondrocyte cultures. However, in the presence of dominant negative inhibitors of CREB and c-Fos, these PTHrP effects were suppressed, and chondrocyte maturation was accelerated. Moreover, in combination, the effects of dominant negative c-Fos and CREB are synergistic, suggesting interaction between these signaling pathways during chondrocyte differentiation.

During limb development, a cartilaginous template is formed from mesenchymal condensations that perform the skeletal elements (1). Subsequent longitudinal bone growth is dependent upon the process of endochondral ossification whereby chondrocytes sequentially proliferate and differentiate. Chondrocyte differentiation is marked by profound physical and biochemical changes, including a 5-10-fold increase in volume and expression of alkaline phosphatase and type X collagen (2,3). The process of chondrocyte differentiation culminates in calcification of the matrix and cellular apoptosis (4 -6). The calcified matrix subsequently serves as the template for primary bone formation.
Endochondral ossification is regulated by local growth factors including PTHrP 1 , which has been defined as a critical regulator of the rate of endochondral ossification. Mice null for either PTHrP (7) or its receptor (8) display accelerated chondrocyte differentiation and therefore abnormal endochondral bone formation. In contrast, animals that overexpress PTHrP exhibit delay in chondrocyte terminal differentiation (9). Humans with an activating mutation in the PTH/PTHrP receptor have Jansen's metaphyseal chondrodysplasia, characterized by disorganization of the growth plate and delayed chondrocyte terminal differentiation (10). Thus, PTHrP signaling is critical for normal growth plate morphology and function.
PTHrP receptor activation stimulates both protein kinase A (PKA) and protein kinase C (PKC) signaling. Activation of these signaling pathways mediates important effects and in chondrocytes PKA signaling is associated with proliferation, stimulation of proteoglycan synthesis, and inhibition of alkaline phosphatase activity (11). PTHrP also stimulates phospholipase C activity with metabolism of membrane phospholipids and production of diacyl glycerol and inositol phosphate (12,13). Inositol phosphate stimulates the release of intracellular calcium stores while the regeneration of diacylglycerol leads to activation of protein kinase C. While much is known about the upstream signaling events immediately following receptor ligation, there is little information on the downstream transcription factors involved in mediating PTHrP effects in chondrocytes and their relationship with PKA and PKC signaling.
One of the potential downstream signaling targets of PKA is the cyclic AMP response element-binding protein (CREB) (14). This transcription factor is a CREB/ATF family member that is constitutively present in the nucleus. CREB binds to a DNA consensus cAMP response element (CRE) primarily as a homodimer, via a leucine zipper domain. CREB is activated by PKA-mediated phosphorylation within its P-box domain (15), which permits its interaction with p300/CBP and other coactivators, leading to gene transcription. Thymocytes and T cells from transgenic mice expressing a dominant-negative form of CREB specifically in these cells display a profound proliferative defect and G 1 cell-cycle arrest in response to a number of different activation signals (16). ATF-2 deficiency, another member of the CREB/ATF family, has been shown to induce chondrodysplasia and neurological abnormalities in mutant mice (17). Although the role of CREB in cartilage development or bone formation has not been investigated, the finding that CREB KO mouse has a dwarf phenotype (18) combined with the established important role of CREB in PKA mediated events in other cells, suggests that this transcription factor may be involved in regulating some of the critical events in chondrocyte differentiation.
PTHrP also activates the transcription factor AP-1, a complex formed through interactions between Fos and Jun family members (19,20). These interactions are also mediated by leucine zipper domains to form Fos/Jun heterodimers or Jun/ Jun homodimers (21). The protein complex binds to the phorbol 12-myristate 13-acetate response element (TRE), a specific cisacting DNA consensus sequence in the promoter region of target genes. AP-1-regulated genes appear to have an important role in skeletal physiology. Transgenic mice overexpressing c-Fos develop bone and cartilaginous tumors (22), while Fos-Jun double transgenic mice develop osteosarcomas at a higher frequency (23). PTHrP stimulates c-Fos mRNA and protein synthesis in osteoblasts (24), and in chondrocytes PTHrP was found to increase c-Jun and JunD mRNA and protein levels (25). AP-1 activation has been shown to be mediated by several different signaling pathways including PKA and PKC (26,27).
In this work we have characterized CREB and c-Fos as two important regulators of cartilage development and have investigated their activation by PKA and PKC signaling. We have shown that their activation occurs rapidly in response to PTHrP stimulation and that interference with their function results in alteration of cartilage phenotype produced by PTHrP. This represents the first report documenting a critical role for CREB in cartilage development.

MATERIALS AND METHODS
Chondrocyte Cell Culture-Embryonic cephalic sternal chondrocytes (day 13) were prepared and cultured as described (28). After isolation and primary culture for 5-7 days, floating cells were plated in secondary cultures at 2.5 ϫ 10 5 cells/cm 2 in Dulbecco's modified Eagle's medium containing 10% NuSerum IV (Collaborative Biomedical, Bedford MA), 4 units/ml hyaluronidase (Sigma), and 2 mM L-glutamate (Sigma). After 6 days, upper sternal chondrocytes (USC) were harvested and plated in 6-well plates for the transient transfection assay and alkaline phosphatase or 60-mm dishes for Northern analysis. PTHrP (10 Ϫ7 M) was added to the cultures in some experiments.
Western Blotting-Cell were washed with cold phosphate-buffered saline and lysed on ice in Golden lysis buffer (29) supplemented with protease inhibitor mixture tablets (Roche Molecular Biochemicals), 1 mM sodium orthovanadate, 1 mM EGTA, 1 mM NaF, 1 M microcysteine. Insoluble material was removed by centrifugation at 12,000 ϫ g. The protein concentration of the soluble material was estimated using Coomassie Plus Protein Assay kit (Pierce, Rockford, IL). 25 g of extracts was assayed by SDS-polyacrylamide gel electrophoresis. After transfer to a nitrocellulose membrane (Schleicher and Schuell), the blots were probed with the following antibodies: anti-CREB or anti-phospho-CREB antibody (Upstate Biotechnology, Lake Placid, NY) at a dilution 1:500, anti-c-Fos, anti-ATF-2, anti-CREM, and anti-ATF-1 (Santa Cruz) at a dilution 1:1000. In one case, the membrane was striped and reprobed with anti-␤-actin (Sigma) in a dilution 1:8000 to serve as a loading control. Horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse polyclonal antibodies (Bio-Rad) were used as secondary antibody. The immune complexes were detected using ECL (Amersham Pharmacia Biotech).
Gel-shift and Supershift Assay-Nuclear protein extracts were prepared as previously described (30) were analyzed by electrophoretic mobility shift assay. 5 g of this nuclear extracts from upper sternal chondrocytes treated with PTHrP for 30 min, 1 h, 2 h, or 4 h were incubated with 1 g of poly(dI-dC) (Amersham Pharmacia Biotech) and 2 ng of 32 P-end-labeled CRE or TRE consensus oligos (Santa Cruz) and run on a 5% polyacrylamide gel. Supershifts were performed with antibodies specific for CREB/ATF and Fos/Jun family members. The incubation with antibodies was 20 min prior to the addition of oligonucleotides.
Transfections and Luciferase Assay-For this set of experiments we have used Stratagene's Path detect in vivo signal transduction pathway reporting systems, which are a series of inducible reporter vectors that contain the firefly luciferase reporter gene driven by a basic promoter (TATA box) plus a defined inducible enhancer element: four repeats of CRE (CRE-Luc) or seven repeats of the TRE (TRE-Luc). The transfection efficiency was controlled for by co-transfecting pRL vector from Promega and determining the Renilla uniformis luciferase activity. The upper sternal chondrocytes were transiently transfected in serum-free media (supplemented with insulin and triiodothyronine) (31) with these plasmids using the transfection reagent Fugene-6 (Roche Molecular Biochemicals). Twelve hours after transfection, a PTHrP (10 Ϫ7 M) treatment was given to the cells for 24 h. Two plasmids expressing the PKA catalytic subunit and 360 -672 amino acids from MEKK were used as positive controls for the CRE/TRE-Luc reporting systems, respectively. The following inhibitors purchased from Calbiochem were used: H-89, 10 M for PKA inhibition, GO6976, 12 M for PKC inhibition, 1 M KN-93 for CaM kinase II, and 50 M PD98059 for MEKK. We have also used a dominant negative CREB and c-Fos in the pRC-CMV vector, as previously described (32). In the co-transfection experiments 1 g of reporter and 1 g of dominant negative (or empty vector for control) were co-transfected in a ratio 1:3 with the transfection reagent. Other reporters used were c-Fos promoter and several cyclin D1 promoter constructs from Dr. R. Pestell (Albert Einstein College of Medicine). Constructs employed were the following: 1745-bp full-length cyclin D1 promoter, 66 bp of the promoter encompassing the CRE site situated at 52 bp and 1745 mut and 66 mut that have the CRE site mutated as previously described (33). Luciferase activity in cell lysate was measured using Luciferase assay system (Promega) and an Optocomp luminometer (MGM Instruments).
Expression of ACREB and AFOS using Retroviral Systems-Chick embryonic fibroblasts grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 0.2% fetal chick serum and penicillin/streptomycin were transfected with the replication competent avian sarcoma virus RCASBP(A) alone and RCASBP(A) containing dominant negative c-Fos and CREB inserts. Cells were passaged three times to allow the spreading of the virus. At confluence, growth media was changed to low serum (Dulbecco's modified Eagle's medium with 10% NuSerum) for collection of virus. Viral supernatants were collected at 24-h intervals for 3 days. At the time of secondary plating, chondrocytes were incubated with fresh viral supernatant for 48 h followed by addition of PTHrP in some of the samples.
Northern Blot Analysis-Total RNA was extracted from cultures using the RNAeasy kit (Qiagen, Valencia, CA). 5 g of the total RNA was run on a 1.2% agarose gel containing 17.5% formaldehyde and transferred to a GeneScreen Plus membrane (PerkinElmer Life Sciences). The RNA was UV cross-linked to the membrane. Northern analysis was performed using denaturing formaldehyde/agarose gels as described (34). A synthetic type X oligonucleotide was end labeled as previously described (34). Prehybridization was performed in QuickHyb solution (Stratagene, La Jolla, CA) for 20 min at 68°C. Hybridization was done at 73°C for 1 h. The blot was washed twice for 15 min with 2 ϫ SSC and 0.1% SDS, followed by a 30-min wash with 0.1 ϫ SSC and 0.1% SDS. The blot was exposed to X-Omat AR film (Kodak, Rochester, NY) for autoradiography.
Alkaline Phosphatase Activity-Alkaline phosphatase activity was measured as previously described (35). Culture medium was aspirated from chondrocytes cultured in 6-well plates, which then were rinsed with 150 mM NaCl. One ml of reaction buffer containing 0.25 M 2-methyl-2-aminopropanol, 1 mM MgCl 2 , and 2.5 mg/ml p-nitrophenyl phosphate (Sigma) at pH 10.3 was added to the wells at 37°C. The reaction was stopped after 30 min by the addition of 0.5 ml of 0.3 M Na 3 PO 4 (pH 12.3). The alkaline phosphatase activity was determined spectrophotometrically at 410 nm by comparison with standard solutions of pnitrophenol and an appropriate blank.
Fluorescence-activated Cell Sorter Analysis-Flow cytometry was performed using a fluorescence-activated cell sorter Calibur cytometer and the Cell Quest (Becton-Dickinson, Franklin Lakes, NJ) plotting program. 2-3 ϫ 10 6 cells were resuspended in 70% EtOH at 4°C for at least 12 h. The cells were then resuspended in 1 ϫ phosphate-buffered saline with 1 mg/ml RNase and incubated for 30 min at room temperature. Finally cells were stained with propidium iodide (10 g/ml in phosphate-buffered saline) and run through the flow cytometer.
[ 3 H]Thymidine Incorporation-Chondrocytes plated at 200,000 cells/ well in a 6-well plate were treated with PTHrP for 24 h and incubated with 2 Ci/ml [ 3 H]thymidine (PerkinElmer Life Sciences) in the last 4 h. Cells were washed with cold 5% trichloroacetic acid two times and then the cells were left in 0.5 ml of 1 N NaOH, 0.1% Triton X-100 overnight at 4°C. 100-l aliquots were added to 4 ml of Ecoscint fluid and assayed for 3 H incorporation.
Statistics-Statistical comparisons were made between the groups using a way analysis of variance (ANOVA). Significance was considered present when the p value was less than 0.05 and is denoted in each of the figures.

Activation of Signaling Pathways by PTHrP-The effects of
PTHrP stimulation on CREB/ATF activation were investigated in sternal chondrocytes at various times following treatment over a 2-h time course experiment. The level of CREB protein remained constant as determined by Western blotting (Fig.  1A). However, phosphorylated CREB (pCREB) levels were markedly increased following this treatment. The increase in pCREB was detectable within 5 min, maximal by 30 min, and returned to baseline levels 2 h after PTHrP stimulation (Fig.  1A). Following PTHrP stimulation, there was also an increase in c-Fos protein levels that was observed as early as 30 min, was maximal by 1 h, and returned to basal levels by 4 h (Fig.  1B). To determine the effect of PTHrP stimulation on CREB DNA binding activity in chondrocytes, gel mobility shift assays were performed. These experiments revealed minimal differences over a 4-h time course following PTHrP stimulation ( Fig.  2A). To characterize this protein-DNA complex, supershift experiments using specific antibodies to members of the CREB/ ATF family were performed. In this experiment, only CREB binding to this consensus sequence was detected (Fig. 2B), although ATF-2 and CREM proteins could be detected in these extracts by Western blot (Fig. 2C). These findings suggest that CREB protein levels and DNA binding activity remain constant following PTHrP treatment. However, PTHrP may modulate gene expression in sternal chondrocytes by inducing CREB phosphorylation and therefore increasing its transcriptional activity.
To assess AP-1 activation following PTHrP stimulation, gel mobility shift assays were performed using a probe containing the TRE consensus sequence. PTHrP stimulated a marked increase in DNA binding ( Fig. 2A). The increase was detectable at 30 min, peaked at 1 h, and remained elevated 4 h following PTHrP treatment. Addition of an antibody directed against c-Fos protein inhibited DNA binding ( Fig. 2A), showing the presence of c-Fos protein in the gel mobility shift complex, consistent with the increase in c-Fos protein expression observed on Western blot. In contrast, incubation with antibodies directed against Fra1, ATF-2, and CREM did not affect DNA binding (Fig. 2B).
To determine whether this apparent stimulation of CREB and AP-1 signaling is associated with an increase in gene transcription, sternal chondrocytes were transfected with CRE and TRE luciferase reporter constructs and treated with PTHrP (Fig. 3). PTHrP (10 Ϫ7 M) treatment resulted in transcriptional activation of both the TRE (23-fold) and CRE (103fold) constructs. As a positive control we co-transfected the reporter constructs with either activated MAPK kinase kinase (MEKK) or PKA constructs. In these experiments the MEKK construct was a more potent stimulator of the TRE reporter, while the PKA construct caused about a 40-fold stimulation of both the CRE and the TRE reporters (Fig. 3A). These latter findings indicate the presence of cross-talk between these signaling pathways in chondrocytes following PTHrP stimulation. Furthermore, PKA signaling in these cells results in activation of both CRE and TRE mediated transcription.
To determine the relative effects of the PKA, PKC, mitogenactivated protein kinase, and calmodulin kinase signaling pathways on CRE activation, the CRE reporter was transfected FIG. 2. PTHrP induces AP-1, but not CREB DNA binding activity. Nuclear extracts (5 g/lane) obtained from USC stimulated with PTHrP for various time points, were incubated with or without anti-c-Fos antibody, followed by addition of a radioactive labeled consensus AP-1 probe (A). Nuclear extracts (5 g/lane) from USC were also incubated with radioactively labeled CRE oligonucleotides and then gel shift assay was performed (A). Nuclear extracts (5 g/lane) from USC were preincubated with antibodies against different CREB/ATF or Fos family members, before addition of radioactive labeled CRE or AP-1 oligonucleotides and then gel shift assay was performed (B). Western blots with antibodies against different CREB/ATF family members were done with whole cell lysates from USC (C).
into sternal chondrocytes and the cultures were treated with PTHrP in the presence and absence of pharmacological inhibitors of these various pathways (Fig. 3B). Inhibition of PKA signaling with H89 resulted in a 95% inhibition of CRE activation, consistent with the role of PKA as a potent inducer of CREB phosphorylation. Inhibitors of PKC (GO6976, 45% inhibition), mitogen-activated protein kinase (PD98059, 27% inhibition), and calmodulin kinase (KN93, 17% inhibition) all had less potent effects. PKA inhibition also had the greatest effect on the TRE reporter (80% inhibition), while inhibition of MAP kinase signaling resulted in a 37.7% decrease in luciferase activity. In contrast, inhibition of PKC and calmodulin kinase did not affect PTHrP-mediated activation of the TRE reporter (Fig. 3C).
To confirm our findings regarding CREB and AP-1 transcriptional activation following PTHrP treatment, constructs expressing dominant negative c-Fos (AFOS) and CREB (ACREB) were each co-transfected into sternal chondrocytes with reporter constructs for the CRE and TRE (Fig. 4). The ACREB construct completely inhibited CRE activation by PTHrP, while the AFOS construct did not inhibit PTHrP activation of the CRE (Fig. 4A). Moreover, the AFOS construct induced an in-crease in the PTHrP stimulation of CRE reporter. A possible explanation for this finding is that AP-1 inhibition may result in greater availability of the transcriptional coactivator, CBP (CREB-binding protein), for CREB-mediated signaling. In contrast, both ACREB and AFOS constructs inhibited activation of the TRE reporter following stimulation by PTHrP (Fig. 4B). Co-transfection of both constructs reduced the detectable luciferase activities to below basal levels. These findings further confirm PTHrP activation of PKA and phosphorylation of CREB, leading to CRE dependent transcription.
In contrast, activation of AP-1 by CREB is likely indirect and may be mediated by stimulation of c-Fos transcription. To test this, a c-Fos promoter-reporter construct was transfected into sternal chondrocytes. PTHrP treatment resulted in a 91-fold stimulation in luciferase activity (Fig. 4C). Co-transfection of these cultures with the ACREB construct blocked this PTHrP stimulation of the cFos promoter, while the AFOS construct had no effect, further supporting a model of PTHrP activation of AP-1 through CREB-dependent transcription of c-Fos.
PTHrP Effects on Maturation Mediated by CREB and AP-1-PTHrP regulates the rate of chondrocyte terminal differentiation. To see if PTHrP inhibits terminal differentiation through CREB or AP-1 signaling, we investigated the expression of markers of chondrocyte maturation, type X collagen (colX) mRNA expression, and alkaline phosphatase activity, in cell cultures following PTHrP treatment. Short-term experiments were performed using transiently transfected chondrocytes (2 days), while longer term experiments (7 days) were performed with chondrocytes infected with a replication competent retrovirus that permitted sustained expression in a larger percentage of cells. In the absence of PTHrP, ACREB, and AFOS had minimal effects in both the transiently transfected and retrovirally infected cultures (Fig. 5, A and B). However, in the presence of PTHrP, transient transfection of ACREB and AFOS constructs stimulated type X collagen (colX) mRNA expression, 1.6-and 1.3-fold, respectively (Fig. 5A), compared with vector-transfected control chondrocytes. These signaling pathways appeared to be synergistic, since co-transfection of both ACREB and AFOS resulted in even greater colX mRNA expression (5.2-fold). Effects of the dominant negative signaling molecules were much greater when the constructs were introduced by retroviral infection. Alone, both ACREB and AFOS infection resulted in a larger induction of colX in PTHrP-treated cultures compared with minimal increases in retroviral infected control samples (Fig. 5B).
In PTHrP-treated cultures, transient transfection with both ACREB and AFOS resulted in only small increases in alkaline phosphatase activity (data not shown). Much larger effects on alkaline phosphatase activity were observed when the cultures were infected with retroviral constructs, consistent with the findings observed with the other maturation marker, type X collagen. In the PTHrP-treated cultures, ACREB doubled and AFOS tripled alkaline phosphatase activity (Fig. 6) compared with smaller increases in untreated samples (1.3-and 1.7-fold increases). Collectively, these findings demonstrate that PTHrP effects on maturation are mediated by both CREB and AP-1 signaling. Furthermore, while the findings suggest basal AP-1 and CREB signaling, these transcription factors are much more relevant in the presence of PTHrP.
CREB and AP-1 Effects on Chondrocyte Proliferation-To evaluate the effects of PTHrP on proliferation, sternal chondrocytes were treated with PTHrP and analyzed by fluorescence-activated cell sorter 24 h latter (Table I) ϩ G 2 ϩ M). Since cyclin D1 is a required for the transition from the G 1 to S phase, we next investigated the effect of PTHrP and downstream signals on the cyclin D1 promoter (Fig. 7). PTHrP treatment caused a 5.6-fold increase in luciferase activity by the full-length (1745 bp) cyclin D1 promoter (Fig. 7A). A point mutation at the CRE site (position 52 bp) results in a 30% reduction in PTHrP mediated stimulation. Similar experiments with a truncated 66-bp promoter construct that includes the CRE-binding site, results in a 2.1fold stimulation following PTHrP treatment. This induction was lost in a mutant construct containing a point mutation that disrupts the CRE consensus sequence. Co-transfection of the cultures with the AFOS or ACREB constructs, either alone or in combination, further confirmed a role for the CRE site in PTHrP-mediated cyclin D1 promoter activation. Transfection of the dominant negative ACREB construct completely inhibited the PTHrP induction of both the fulllength and truncated 66-bp cyclin D1 promoter. In contrast, transfection of the AFOS construct did not significantly alter the induction of the full-length construct and caused only a partial reduction in the truncated promoter (Fig. 7B). These findings confirm the effects of PTHrP on proliferation and suggest that the effects on cyclin D1 are primarily related to CREB signaling.
To further define the role of AP-1 and CREB on PTHrP mediated proliferative effects, chondrocytes were treated with PTHrP in cultures transfected with AFOS or ACREB alone or in combination (Fig. 8). ACREB resulted in a 20% reduction in thymidine incorporation and 25% AFOS reduction. A statistically significant 30% reduction in thymidine incorporation (p Ͻ 0.05) was observed when cultures were co-transfected with both constructs. Thus, similar to the effects observed on differentiation, the regulation of chondrocyte proliferation by PTHrP is mediated by both AP-1 and CREB signaling, with interactive and synergistic effects observed between these pathways.

DISCUSSION
While PTHrP is known to be a critical regulator of chondrocyte proliferation and differentiation, the signaling pathways through which this factor acts remain to be elucidated. The current article demonstrates that both CREB and AP-1 activation are critical to PTHrP signaling in chondrocytes. In these chondrocytes, CREB DNA binding activity is constitutive. However, PTHrP treatment leads to CREB phosphorylation within 5 min. In contrast AP-1 DNA binding activity is induced by PTHrP, with effects initially observed 1 h following treatment. Consistent with these findings PTHrP activated CRE and TRE reporter constructs, and both signaling pathways were found to be important mediators of PTHrP effects on chondrocyte phenotype. Alone, PTHrP suppressed maturation and stimulated proliferation of the chondrocyte cultures. However, in the presence of dominant negative inhibitors of CREB and c-Fos, these PTHrP effects were suppressed and chondrocyte maturation was accelerated. Moreover, in combination, the effects of dominant negative cFos and CREB were synergistic, suggesting interaction between these signaling pathways during chondrocyte differentiation.
Prior studies have demonstrated that CREB binds constitutively to the CRE consensus sequence and is activated following phosphorylation by upstream kinases (36,37). The cAMP-activated protein kinase A (PKA) dependent pathway is the classic activator of CREB and acts through phosphorylation of a critical serine residue located at position 133 (15,38). However, other signals, including the calcium/calmodulin signaling (CaM kinases II and IV) (39), PKC signaling (40), and Ras signaling mediated by the serine/threonine kinase RSK2 also regulate CREB through phosphorylation at Ser 133 or other sites (36,41). In our pharmacological studies with inhibitors of these kinases, we found that the PKA pathway is the one through which the most potent signal is transduced.
Similarly, AP-1 activation is also stimulated by numerous pathways, but here we found that following PTHrP stimulation, PKA signaling is most critical for its transcriptional activity. Inhibition of PKA signaling with the drug H-89, or co-transfection of ACREB, resulted in nearly complete inhibition of AP-1 transcriptional activation, as measured by reporter assay. Maximal transcriptional activation of AP-1 regulated genes is mediated by c-Fos/c-Jun heterodimers, and is highly dependent upon c-Fos protein expression. Previously, it has been shown that c-Fos transcription is induced by CREB activation (42)(43)(44). Here we find evidence that this mechanism is also operative in chondrocytes following PTHrP stimulation. Our transient transfection experiments here demonstrated the induction of a c-Fos promoter by CREB as previously described (43,45,46). Consistent with these findings was the increase in c-Fos protein and the delay in AP-1 DNA binding activity, which was maximally increased at 1 h following PTHrP stimulation, while an increase in CREB phosphorylation could be detected within 5 min. Thus, our data support a mechanism whereby PTHrP rapidly activates CREB through PKA signaling, and results in the subsequent transcriptional activation of PKA/CREBdependent genes. One of these genes is likely to be c-Fos, which following its de novo synthesis, enhances AP-1 signaling (43,45,46). These events result in a delayed enhancement of AP-1 signaling through c-Fos transcription, and are consistent with PTHrP effects reported in other cell systems (43,44).
The primary role for cyclic AMP/CREB signals in the regulation of chondrocyte differentiation is supported by genetic experiments. CREB knockout mice has a dwarf phenotype and die in the neonatal period of respiratory distress, similar to PTHrP knockout animals. Recently, a mutant PTH receptor with normal phospholipase C signaling, but deficient G␣ s signaling has been expressed in chimeric mice (47). Cells with deficient G␣ s signaling underwent premature maturation in the growth plate, while wild type cells had a normal rate of differentiation. In contrast, mice that express a mutant PTHrP receptor with normal G␣ s signaling, but deficient phospholipase C signaling, have normal size, are fertile, and do not have reduced rates of chondrocyte differentiation (48). These genetic experiments support our findings in that cyclic AMP/PKA signal transduction regulates the activation of the two important transcription factors, CREB and AP-1, that mediate PTHrP effects in chondrocytes. In contrast, PKC signaling appears less relevant for these events both in vivo and in vitro.
To establish whether either CREB or AP-1 transcriptional regulation is important for the effect of PTHrP on chondrocyte phenotype, experiments were performed using ACREB and AFOS. In the absence of PTHrP, the effects of these signaling inhibitors was minimal, suggesting relatively low basal activation of these signaling pathways. However, large effects were observed in the presence of PTHrP, consistent with the activation of CREB and induction of c-Fos by PTHrP. Although the magnitude of the individual signaling effects was greater in retroviral infected cultures, the transient transfection experiments demonstrated interactive effects between CREB and AP-1 signaling in chondrocyte maturation.
Chondrocyte maturation was determined by the expression of type X collagen and alkaline phosphatase, both of which are markedly elevated during chondrocyte differentiation and are important for the normal process of endochondral ossification (49,50). As a suppresser of chondrocyte differentiation, PTHrP inhibited both type X collagen and alkaline phosphatase activity. Here we used two strategies to examine the role of CREB and AP-1 in this process: use of transient transfection and long-term retroviral expression vectors. Although transient transfection allows for modest expression for a shorter period of time, it also permits use of both dominant-negative constructs in combination. Thus, the effects we observed in these experiments were smaller but we were able to define synergistic effects and show interdependence between AP-1 and CREB mediated effects. In contrast, the effects observed in the retrovirally infected cultures were greater and further confirmed the importance of these transcription factors in PTHrP mediated signaling events. The phenotype that we obtained by inhibiting CREB and AP-1 signaling is similar to the phenotype of the PTHrP knockout mouse; accelerated chondrocyte differentiation of chondrocytes, resulting in premature endochondral ossification at many sites (7,51).
Although, PTHrP has an accepted role as an inhibitor of maturation, its role in proliferation is controversial. While Lee et al. (52), observed similar expression of H4-histone mRNA (marker for S-phase of the cell cycle) in wild-type and PTHrPdeficient animals, others have shown diminished [ 3 H]thymidine incorporation and a marked reduction in the number of proliferating chondrocytes in the knockout animals (51,53). In vitro studies have supported an important role for PTHrP in chondrocyte proliferation (54 -57). In the current experiments, PTHrP doubled the number of cells entering mitosis, and stimulated [ 3 H]thymidine incorporation and cyclin D1 promoter activity in chondrocytes. This increase was diminished by inhibiting AP-1 and CREB signaling. One crucial step for G 1 progression through the cell cycle appears to be induction of cyclin D1 expression during growth factor stimulated cell cycle re-entry (33). Consistent with reports in the literature (58), our results point out a role of CREB/ATF and AP-1 family members in the activation of cyclin D1 following PTHrP stimulation.
Chondrocytes integrate a complex array of signals during the process of differentiation. PTHrP is one of the major regulators but other factors such as Indian hedgehog, bone morphogenetic proteins, and transforming growth factor-␤ also affect the differentiation through overlapping or complementary pathways (54,59,60). Our experiments show that both CREB and AP-1 transcription factors are important regulators of the rate of chondrocyte differentiation. Additional studies will be required to define how other growth factors interact with CREB and AP-1 signaling and thereby modulate PTHrP effects during chondrocyte differentiation.