Preclinical and clinical evidence from megakaryocyte (MK)-related diseases suggests that MKs play a significant role in maintaining bone homeostasis. Findings from our laboratories reveal that MKs significantly increase osteoblast (OB) number through direct MK-OB contact and the activation of integrins. We, therefore, examined the role of Pyk2, a tyrosine kinase known to be regulated downstream of integrins, in the MK-mediated enhancement of OBs. When OBs were co-cultured with MKs, total Pyk2 levels in OBs were significantly enhanced primarily because of increased Pyk2 gene transcription. Additionally, p53 and Mdm2 were both decreased in OBs upon MK stimulation, which would be permissive of cell cycle entry. We then demonstrated that OB number was markedly reduced when Pyk2−/− OBs, as opposed to wild-type (WT) OBs, were co-cultured with MKs. We also determined that MKs inhibit OB differentiation in the presence and absence of Pyk2 expression. Finally, given that MK-replete spleen cells from GATA-1–deficient mice can robustly stimulate OB proliferation and bone formation in WT mice, we adoptively transferred spleen cells from these mice into Pyk2−/− recipient mice. Importantly, GATA-1–deficient spleen cells failed to stimulate an increase in bone formation in Pyk2−/− mice, suggesting in vivo the important role of Pyk2 in the MK-induced increase in bone volume. Further understanding of the signaling pathways involved in the MK-mediated enhancement of OB number and bone formation will facilitate the development of novel anabolic therapies to treat bone loss diseases.
In the last decade, platelet-producing megakaryocytes (MKs) have been shown to play a role in regulating bone mass. Myeloproliferative diseases in which increases in MKs are accompanied by osteosclerosis have been reported,[1-3] and several mouse models have been described in which increased numbers of MKs correlate with increased bone mass. These mouse models, as well as relevant in vitro data, were recently reviewed. Three key findings from these data provide the rationale for our current studies. First, MKs stimulate osteoblast (OB) proliferation and bone formation in vivo. This was demonstrated in four distinct mouse models in which MK number was increased and a significant increase in bone volume was observed.[5-12] This was further illustrated by adoptive transfer studies in which irradiated wild-type (WT) mice were reconstituted with spleen cells from mice lacking either NF-E2 or GATA-1, transcription factors required for normal MK differentiation. NF-E2 and GATA-1–deficient mice have significantly increased numbers of MKs, reduced platelet numbers, and a two- to threefold increase in bone mass.[10, 11, 13] Both the hematologic phenotype and the high bone mass phenotype can be adoptively transferred, suggesting a role for hematopoietic cells, most likely MKs, in this mechanism. Second, recent studies by Dominici and colleagues described how, in mice treated with lethal total body irradiation, surviving MKs migrated to endosteal bone surfaces (in close contact with OBs) and stimulated a twofold increase in OB number. Third, in vitro studies demonstrate that MKs enhanced OB proliferation up to sixfold by a mechanism that required direct MK-OB cell-cell contact and the engagement of integrins.[10, 15-18] Taken together, these observations suggest that MKs, via a cell-cell contact mechanism mediated in part by integrins, stimulate an increase in OB number, which in turn results in an increase in bone formation.
The primary goal of our study was to determine the cellular mechanisms by which MKs regulate OB proliferation. We show for the first time to our knowledge an important role for proline-rich tyrosine kinase 2 (Pyk2), a tyrosine kinase involved in signaling downstream of activated integrins and other key signaling pathways in OBs, in regulating MK-mediated enhancement of OB number, and the importance of Pyk2 expression in regulating MK-mediated bone formation in vivo.
Materials and Methods
C57BL/6 mice were originally obtained from Jackson Laboratories (Bar Harbor, ME, USA). GATA-1–deficient mice were generously provided by Dr. Stuart Orkin. Generation and breeding of GATA-1–deficient mice was described previously.[13, 19] These mice have a selective loss of MK GATA-1 expression but with sufficient erythroid cells to avoid lethal anemia. GATA-1–deficient mice have been backcrossed onto the C57BL/6 background for more than seven generations. Pyk2−/− mice were described previously and were generously provided by Pfizer (Groton, CT, USA). Pyk2 mice were backcrossed for greater that seven generations and maintained on a C57BL/6 background. All procedures were approved by the Institutional Animal Care and Use Committee of the Indiana University School of Medicine and followed NIH guidelines.
Preparation of neonatal calvarial cells (OBs)
Murine calvarial cells were prepared as previously described.[15, 21] Our technique was a modification of the basic method described by Wong and Cohn. Briefly, calvaria from mice less than 48 hours old were pretreated with EDTA in PBS for 30 minutes. The calvaria were then subjected to sequential collagenase digestions. Cells were collected after incubation in collagenase. Fractions 3 to 5 were used as the starting population. Although heterogeneous, cells collected from these fractions are enriched for osteogenic potential. These cells were ∼90% OBs or OB precursors by a variety of criteria[21, 24-26] and are referred to as OBs throughout this article. We recently published the characterization of these cells. Here, we provide new flow cytometric evidence that 83% ± 2% of 2-day calvarial OBs were CD45-CD31-Ter119-Sca1- and of these cells, 90% ± 4% were also OPN +  (data are presented as the average ± standard deviation of the mean from four separate experiments each with three replicate samples). OBs were seeded at 2 × 104 cells/mL in both 6-well plates (3 mL/well) and 10-cm dishes (10 mL/well) (optimal pretested).
Preparation of fetal liver–derived MKs
Murine MKs were prepared as previously described.[10, 29] In brief, fetuses were dissected from pregnant mice at E13 to 15. The livers were removed and single-cell suspensions made by forcing cells through sequentially smaller gauge needles (18G, 20G, 23G). Cells were washed twice with DMEM + 10% fetal calf serum (FCS) and then seeded (5 fetal livers/100-mm dish) in 100-mL culture dishes, in DMEM + 10% FCS + 1% murine thrombopoietin (TPO). After 3 to 5 days, when the cells became confluent, MKs were separated from lymphocytes and other cells using a one-step albumin gradient to obtain a 95% pure MK population. The bottom layer was 3% albumin in PBS (Bovine Albumin, protease free, fatty acid poor, Serologicals Proteins Inc., Kankakee, IL, USA), the middle layer was 1.5% albumin in PBS, and the top layer was media containing the cells to be separated. All of the cells sedimented through the layers at 1 g for approximately 40 minutes at room temperature. The MK fraction was collected from the bottom of the tube. MKs were seeded at 1 × 105 cells/well in 6-well plates and 2 × 105 cells/dish in 10-cm dishes (optimal pretested).
Preparation of bone marrow (BM) stromal cells
BM was isolated from the tibias and femurs of 6- to 10-week-old C57BL/6 or Pyk2−/− mice and seeded at 5 × 106 cells/well into a 24-well plate and MKs were titrated into wells. To provide osteogenic conditions, cells were cultured in αMEM supplemented with 10% FCS, which was further supplemented with ascorbic acid (50 μg/mL added on day 0 and at all medium changes) and β-glycerophosphate (5 mM added starting on day 7 and all subsequent medium changes). On day 14, cultures were stopped and OB function was analyzed as detailed below.
OBs (2 × 105 OBs/mL) were cultured in the presence or absence of MKs (2.5 × 105 MKs/mL) for 14 days. Alternatively, OBs were cultured with MKs for the first 4 hours of co-culture and then MKs were removed by washing. Co-cultures were incubated in osteogenic media as described above. The media was changed at day 7, and at this time point fresh MKs were added as nonadherent MKs were removed. On day 14, cultures were assessed for calcium deposition (Alizarin Red S staining) as a functional measure of OB differentiation and mineralization (as previously described[15, 24, 31]) and mRNA expression of alkaline phosphatase, type I collagen, and osteocalcin was assessed by real-time PCR as detailed below.
Preparation of cell lysates for Western blot analysis
Cell lysates were prepared by one of two methods. In the first method, OBs were cultured alone or co-cultured with MKs for 5 days. After respective treatments, MKs were removed by washing (4 washes with PBS) and cells were lysed in a nonionic detergent buffer (40 mM Tris, 150 mM NaCl, 1% Igepal) supplemented with protease inhibitors and sodium orthovanadate. Protein concentrations were measured using the Bradford Reagent (Bio-Rad Laboratories, Hercules, CA, USA). Lysates were boiled in laemmli buffer for 10 minutes, fractionated using SDS-PAGE, and transferred to PVDF membrane (GE Healthcare, Piscataway, NJ, USA). Western blots were performed with antibodies to Mdm2 (2A10, 4B11) and GAPDH (6C5) (Calbiochem, San Diego, CA, USA), and p53 (FL-393) and Vinculin (N-19) (Santa Cruz Biotechnology, Santa Cruz, CA, USA).
The second method was used for examining expression of Pyk2. In brief, OBs were cultured alone, co-cultured with MKs for 5 days, or co-cultured with MKs for 4 hours and then MKs removed by washing (4 washes with PBS) and OBs remained in culture for the remaining 5-day period. Alternatively, for short stimulation studies, OBs were cultured until they were ∼70% confluent, serum starved overnight (0.1% FCS), and then stimulated for 1, 2, or 4 hours with MKs, where appropriate. For all cultures, MKs were removed by washing (4 washes with PBS). Cell lysates were collected in SDS-PAGE lysis buffer containing urea. Protein concentrations were measured using the amido black method. The proteins were fractionated via SDS-PAGE and transferred to nitrocellulose. They were then probed for Pyk2 (BD Transduction Labs, San Jose, CA, USA; 610548) using the Fujifilm LAS-3000 imager. The HRP-tagged secondary antibodies were purchased from Jackson ImmunoResearch (West Grove, PA, USA). The blots were stripped and reprobed for vinculin (Sigma, St. Louis, MO, USA; V 4505) to demonstrate equal protein loading.
Quantitative real-time PCR
After respective treatments, all cultures were washed 4 times with PBS to ensure removal of MKs from MK-OB co-cultures. Total RNA was isolated from OBs using Trizol reagent (Invitrogen Corporation, Carlsbad, CA, USA). Total RNA from OBs was treated with DNAse I (Qiagen, Valencia, CA, USA) and used to generate cDNAs by reverse transcription according to the manufacturer's instructions (First Strand cDNA Synthesis Kit; Roche Applied Science, Indianapolis, IN, USA). QPCR reactions were performed in an MX3000 detection system using SYBR green PCR reagents as described by the manufacturer (Stratagene, La Jolla, CA, USA). For each primer pair, a calibration curve was performed and all oligonucleotides were tested to ensure specificity and sensitivity. For each OB sample, arbitrary units obtained using the standard curve and the expression of GAPDH was used to normalize the amount of the mRNA transcript. The following primer pairs were used:
- GAPDH: 5'CGTGGGGCTGCCCAGAACAT; 5'TCTCCAGGCGGCACGTCAGA
- Pyk2: 5'GGGACACTACCTGGAACGAA; 5'CCAGCTTCACACACTCAGGA
- Runx2: 5'CGACAGTCCCAACTTCCTGT; 5'CGGTAACCACAGTCCCATCT
- Alkaline phosphatase: 5'GCTGATCATTCCCACGTTTT; 5'CTGGGCCTGGTAGTTGTTGT
- Type I collagen: 5'CAGGGAAGCCTCTTTCTCCT; 5'ACGTCCTGGTGAAGTTGGTC
Adoptive transfer studies
Ten-week-old female C57BL/6 and Pyk2−/− mice were lethally irradiated (1100 cGy split dose) and used as recipient mice. Within 1 hour of the second dose, mice were injected intravenously with 107 spleen cells from either C57BL/6 or GATA-1–deficient mice. It should be noted that identical protocols were followed in three separate experiments and all surviving mice were pooled and data analyzed as outlined below (three experiments, n = 3–4/experiment with final n values 9–12/group; see Tables 1 and 2). Briefly, mice were bled at 4 and 8 weeks post-adoptive transfer and platelet numbers determined using a validated HEMAVET 950FS Hematology System (Drew Scientific, Waterbury, CT, USA). Mice were sacrificed at 8 weeks, spleens were weighed, and the distal femurs were analyzed by standard micro-computed tomography (µCT; Skyscan 1172, Bruker-microCT, Kontich, Belgium) using the nomenclature recommended by Bouxsein and colleagues. In brief, the trabecular bone compartment of each femur was sliced into 50 segments from the cortical shell in a region ∼0.5 mm below the most distal portion of the growth plate. The X-ray source was set at 60 keV and 167 µA over an angular range of 180 degrees (rotational steps of 0.40 degrees) with a 12-µm pixel size, and projection images were reconstructed using standard Skyscan software. Images were binarized (threshold of 100 on a 0 to 255 scale), and three-dimensional bone volume parameters were calculated: trabecular bone volume fraction (BV/TV), bone surface density (BS/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp). To convert grayscale values to density (mg/cm3), a phantom standard was scanned and reconstructed using the above parameters. Two densities were assessed: the entire trabecular bone compartment and the bone alone. Although the trabecular bone compartment includes the marrow and thus provides the average density of the bone and marrow, the density of the bone alone provides an index of the degree of bone mineralization. Static histomorphometric analysis of trabecular bone was performed on femurs as previously described using the nomenclature recommended by Parfitt and colleagues.[10, 11, 33]
|Recipient mice||Donor cells||No. of mice||Spleen weight (mg)||Platelets (k/µL)||BV/TV||BS/TV||Tb.Th||Tb.N||Tb.Sp|
|C57BL/6||C57BL/6||12||74 ± 7||884 ± 23||3.1 ± 0.5||3.2 ± 0.3||0.042 ± 0.003||0.70 ± 0.07||0.27 ± 0.01|
|C57BL/6||GATA-1||9||204 ± 5*||247 ± 14*||6.5 ± 1.4*||5.5 ± 1.4*||0.044 ± 0.003||1.43 ± 0.25*||0.29 ± 0.02|
|Pyk2−/−||C57BL/6||11||72 ± 7||866 ± 67||5.8 ± 0.8||5.2 ± 0.5||0.047 ± 0.002||1.2 ± 0.1||0.23 ± 0.01|
|Pyk2−/−||GATA-1||11||207 ± 5*||239 ± 22*||6.6 ± 1.2||5.4 ± 0.8||0.049 ± 0.005||1.2 ± 0.2||0.28 ± 0.01|
|Recipient mice||Donor cells||No. of mice||BV/TV||N.Ob/T.Ar||N.Oc/T.Ar||N.Ob/BS||Oc.S/BS|
|C57BL/6||C57BL/6||12||4.4 ± 0.5||42 ± 4||28 ± 3||1.1 ± 0.5||9.9 ± 0.8|
|C57BL/6||GATA-1||9||6.8 ± 0.9*||66 ± 6*||53 ± 10*||0.8 ± 0.2||13.1 ± 0.9*|
|Pyk2−/−||C57BL/6||11||6.9 ± 1.2||72 ± 12||41 ± 5||0.9 ± 0.2||11.0 ± 2.3|
|Pyk2−/−||GATA-1||11||6.0 ± 0.8||71 ± 12||49 ± 8||0.7 ± 0.1||11.3 ± 1.8|
Unless otherwise stated, all data are presented as the mean ± SD and a Student's t test was used to determine significant differences, with p < 0.05 (Systat 6.0 for Microsoft Windows, SPSS Inc., Chicago, IL, USA). For the adoptive transfer studies, a sample size estimate was performed using the standard deviation of the initial Pyk2−/− recipient data sets, an 80% power, and an alpha of 0.05. This yielded a sample size of 10 for each arm of the study. We therefore completed additional experimentation to give an anticipated combined n of 10/experimental group (to account for loss in mice because of death resulting from transplantation failure or anesthetic complications).
MKs upregulate Pyk2 expression in OBs
Because we previously demonstrated that MK-induced OB proliferation is mediated in part by α3 and α5 integrins, we sought to determine whether Pyk2 was involved in the integrin-mediated regulation of OB proliferation by MKs. As shown using Western blot analysis (Fig. 1A), we observed a marked upregulation of Pyk2 expression in OBs after stimulation of OBs with MKs (greater than twofold increase after 1 hour of stimulation). Pyk2 expression was similarly elevated in OBs co-cultured with MKs for 2 and 4 hours.
We next examined whether the elevated Pyk2 protein levels were the result of changes in Pyk2 gene transcription or protein translation. For these studies, we used the chemical inhibitors actinomycin D (ActD, 5 µg/mL, optimal pretested) and cycloheximide (Chx, 10 µM, optimal pretested), which inhibit RNA synthesis or mRNA translation, respectively. OBs were pretreated with Chx or ActD for 1 or 3 hours, respectively, and then cultured in the presence or absence of MKs for 4 hours (plus inhibitors). Cells were then lysed and proteins were prepared for detection of Pyk2 by Western blotting (Fig. 1B, C). Consistent with our previous studies, Pyk2 protein levels increased in untreated OBs co-cultured with MKs compared with untreated OBs cultured alone. We also found that Chx reduced Pyk2 levels in OBs cultured alone or in the presence of MKs, and that the percentage decrease of Pyk2 was similar in both culture conditions (24% and 25%, respectively). ActD treatment also led to a decrease in Pyk2 protein levels in OBs cultured alone or in the presence of MKs. However, although Pyk2 protein levels in OBs were reduced by 29% in the presence of ActD, Pyk2 levels were reduced by 38% in OBs co-cultured with MKs. This finding suggested that the increase in Pyk2 protein levels in response to MKs was likely owing to an increase in transcription of the Pyk2 gene. To confirm the effect of MKs on Pyk2 mRNA levels, we cultured OBs in the presence or absence of MKs as above, isolated RNA from OBs, and then examined Pyk2 mRNA expression via real-time PCR. As illustrated in Fig. 1D, Pyk2 mRNA expression was markedly upregulated in OBs co-cultured with MKs. As expected, ActD treatment significantly reduced Pyk2 mRNA expression in OBs as well as in OB + MK cultures. Together, these findings suggest that MKs increase Pyk2 mRNA expression, leading to increased Pyk2 protein levels in OBs.
Pyk2 expression is necessary for the MK-mediated increase in OB number
Next, we examined the involvement of Pyk2 in the MK-mediated enhancement of OB number. As illustrated in Fig. 2, Pyk2−/− OB numbers were essentially identical to WT OBs at each time point (Pyk2−/− OB versus WT OB, days 1 to 5). As would be expected based on our previously published results,[10, 15-17] by day 3, OB number was significantly increased when WT OBs were co-cultured with MKs compared with cultures in which WT OBs were cultured alone (WT OB + MK versus WT OB, p < 0.01). In contrast, when Pyk2−/− OBs were co-cultured with MKs, OB number was not significantly different from that measured in cultures containing WT or Pyk2−/− OBs cultured alone, even at day 5, when the greatest increase in OB number in WT OB-MK co-cultures was observed (Pyk2−/− OB + MK versus WT OB or Pyk2−/− OB). In fact, on day 5, OB number in Pyk2−/− OB cultures containing MKs was significantly lower than that observed in WT OB cultures containing MKs (Pyk2−/− OB + MK versus WT OB + MK, p = 0.01), suggesting that Pyk2 was necessary for OBs to respond maximally to the presence of MKs. We are unable to determine if longer time points would promote Pyk2−/− OB proliferation because the duration of experiments may not be extended past 5 days as OB cultures become confluent. We also examined Annexin V expression as a marker for apoptosis. In all cases, less than 2% of the OBs were Annexin V positive, suggesting that apoptosis was not significant (data not shown). Therefore, our results suggest that Pyk2 is a necessary component of the MK-induced signaling pathway leading to OB proliferation, although additional proteins are also likely to be involved in this process.
Role of β1 integrin signaling in MK-mediated expression of Pyk2 in OBs
We previously demonstrated that the MK-mediated increase in OB proliferation requires direct MK-OB cell-cell contact. We further demonstrated that the response was partially mediated by fibronectin/RGD-binding integrins (α3β1 and αvβ1) because titration of the integrin-blocking peptide RGDS into OB-MK co-cultures caused a dose-dependent decrease in MK-mediated OB proliferation without affecting OBs cultured alone. Therefore, in the current study, we determined whether the addition of RGDS to co-cultures would impact Pyk2 expression. Unexpectedly, as shown in Fig. 3, treatment with RGDS did not alter Pyk2 levels regardless of the presence or absence of MKs, which indicates that the MK-mediated effects on Pyk2 levels are likely to be regulated by an alternative signaling pathway. However, an effect of MKs on the activation of Pyk2's catalytic activity cannot be ruled out.
Short-term priming of OBs with MKs increases OB number
To evaluate at which point the proliferative signal is initiated in OBs in response to MKs, we determined whether short-term stimulation of OBs with MKs impacts OB number long term. As detailed in Fig. 4A, washing had no impact on OB number (OB versus OB washed). Again, as expected, OB number was significantly elevated in cultures continuously stimulated with MKs for 5 days compared with OBs cultured alone (OB + MK versus OB, p < 0.001). Interestingly, whereas continuous stimulation of OBs with MKs resulted in the highest number of OBs, OBs primed with MKs for 4 hours also exhibited a significant increase in OB number 5 days later compared with WT OBs cultured alone or washed WT OBs cultured alone (OB + MK washed versus OB or OB washed, p < 0.001 and p < 0.01, respectively). These data suggest that priming of OBs with MKs initiates a series of signaling events, which persist for several days, leading to a significant increase in OB number by 5 days.
MKs enhance long-term Pyk2 in OBs
We then determined whether priming OBs by short-term stimulation followed by washing to remove MKs resulted in a persistent increase in Pyk2 in OBs. As detailed in Fig. 4B, long-term culture of OBs with MKs resulted in a greater than fivefold increase in total Pyk2 compared with OBs cultured alone (OB + MK versus OB, p < 0.001). OBs co-cultured with MKs for 4 hours before MK removal also exhibited an approximate twofold increase in Pyk2 compared with OBs cultured alone (OB + MK washed versus OB or OB washed, p < 0.05). Interestingly, the twofold increase in Pyk2 levels observed after priming of OBs with MKs for 4 hours was similar to the increase in Pyk2 observed when OBs were harvested after 1 to 4 hours of MK stimulation (Fig. 1A). These results suggest that MKs stimulate the persistent upregulation of Pyk2 in OB cultures, which is consistent with transcriptional regulation of the Pyk2 gene that we observed (Fig. 1C, D). Furthermore, our findings suggest that important signal transduction events are initiated within 4 hours of MK-OB co-culture, which lead to an increase in Pyk2 expression and to a corresponding increase in OB proliferation.
Pyk2 is integrated in the p53-Mdm2 pathway in OBs
Next, we examined whether short-term stimulation (priming) of OBs with MKs resulted in alterations in p53 levels, a key protein involved in cell cycle regulation. As shown in Fig. 5, OBs co-cultured with MKs for 4 hours before MK removal (OB + MK washed) exhibited a marked reduction in p53 levels compared with OBs cultured alone (OB and OB washed). Co-culture of OBs with MKs for 5 days (OB + MK) also resulted in a marked reduction in p53 levels compared with OBs cultured alone (OB) (Fig. 5). These data suggest that both continuous and short-term stimulation of OBs with MKs results in a significant reduction of p53 in OBs after 5 days.
To determine whether Pyk2 was obligate in mediating an increase in OB number through Mdm2 and p53 expression, WT and Pyk2−/− OBs were either cultured alone or with WT MKs for 5 days and the levels of Mdm2 and p53 were assessed. As shown in Fig. 6, Mdm2 and p53 levels were markedly reduced in WT OBs after incubation with MKs for 5 days, although both Mdm2 and p53 levels remained largely unchanged in Pyk2−/− OBs co-cultured with MKs. When activated, Mdm2 destabilizes p53 and is then itself degraded. Thus, our data suggest that Pyk2 is involved in the downregulation of Mdm2 and p53 and suggest a novel Pyk2-regulated pathway that is activated in OBs in response to MKs.
MKs reduce OB differentiation in the presence or absence of Pyk2
Because we previously demonstrated that MKs inhibit OB differentiation after 14 days of co-culture, we next wanted to determine if short-term stimulation of OBs by MKs is sufficient for MK-mediated reductions in OB differentiation. Furthermore, we examined if Pyk2 expression is important in MK-mediated reductions in OB differentiation. The addition of MKs to WT and Pyk2−/− OBs significantly reduced the expression of alkaline phosphatase, type I collagen, osteocalcin, and calcium deposition (Fig. 7A–D). These data support the idea that MKs inhibit OB differentiation in the presence or absence of Pyk2, albeit the reduction in OB differentiation markers (with the possible exception of type I collagen) was far greater in Pyk2−/− OBs than WT OBs. Additionally, MKs inhibited OB differentiation when co-cultured with OBs for the entire 14-day period or when co-cultured for the first 4 hours (Fig. 7A–D). Finally, we examined the effects of MKs on the differentiation of stromal-derived osteoprogenitors by culturing BM cells from WT and Pyk2−/− mice under osteogenic conditions in the presence or absence of MKs for 14 days (Fig. 7E). As expected, Pyk2−/− BM cultures were significantly more mineralized than WT BM cultures. Similar to our findings with 2-day calvarial OBs, co-culture of both WT and Pyk2−/− BM cells with MKs resulted in a significant reduction in mineralization on day 14 compared with WT and Pyk2−/− BM cells cultured alone, respectively.
Pyk2 is important for MK-induced bone formation in vivo
Although our in vitro findings clearly demonstrate the involvement of Pyk2 in the MK-stimulated increase in OB number, MKs blocked OB differentiation in the presence or absence of Pyk2. To obtain further functional in vivo evidence for the role of Pyk2 in the MK-mediated increase in OB number, we employed a method known as adoptive transfer, which involves the transfer of bone marrow or spleen cells from one mouse into a lethally irradiated recipient mouse. We previously demonstrated that GATA-1–deficient mice have a significant greater than twofold increase in bone volume and that the high bone mass phenotype of these mice can be adoptively transferred into C57BL/6 recipient mice with GATA-1–deficient spleen cells, suggesting a role for hematopoietic cells in this mechanism, most likely MKs (and data not shown). Because the marrow cavity in GATA-1–deficient mice is filled with bone, spleen cells rather than bone marrow cells were used for the adoptive transfer. Spleen cells are an excellent source of immature MKs (up to 100-fold increase) and hematopoietic progenitor cells because of the extra-medullary hematopoiesis. For the current studies, C57BL/6 or Pyk2−/− mice were lethally irradiated and used as recipient mice. The hematopoietic system of these recipients was reconstituted with 10 million spleen cells derived from either C57BL/6 or GATA-1–deficient mice. Eight weeks post-transfer, mice were bled and platelet numbers determined. As shown in Table 1, all mice reconstituted with GATA-1–deficient cells had significantly fewer platelets compared with those reconstituted with C57BL/6 cells, which demonstrates the transfer of the hematologic phenotype of GATA-1–deficient mice. The transfer of the hematologic phenotype was further confirmed by increased spleen weight 8 weeks post-transfer, whereby mice receiving GATA-1–deficient donor cells had a three- to fourfold increase in spleen weight compared with those receiving C57BL/6 donor cells.
To investigate differences in the bone phenotypes of recipient mice at 8 weeks post-transfer, the distal femur was analyzed by µCT and histomorphometric analyses. These data sets are reported in Tables 1 and 2, respectively. With respect to µCT analyses as detailed in Table 1 and consistent with our previously published studies, reconstituting C57BL/6 mice with GATA-1–deficient spleen cells resulted in a statistically significant, greater than twofold increase in bone volume fraction (BV/TV, p < 0.01) compared with C57BL/6 mice reconstituted with C57BL/6 cells. The effect size was a robust 1.16. Likewise, C57BL/6 mice reconstituted with GATA-1–deficient spleen cells had significant elevations in bone surface density (BS/TV, p < 0.01), and trabecular number (Tb.N, p < 0.001) compared with those reconstituted with WT cells. As predicted based on the high bone mass phenotype previously reported in Pyk2−/− mice,[20, 39] most of the bone parameters measured in Pyk2−/− recipient mice reconstituted with C57BL/6 cells were higher than C57BL/6 recipient mice reconstituted with C57BL/6 cells (BV/TV, p < 0.01; BS/TV, p < 0.01; Tb.N, p < 0.001). Importantly, although some of the bone parameters measured by µCT in Pyk2−/− recipient mice reconstituted with GATA-1–deficient spleen cells were slightly elevated compared with those measured in Pyk2−/− recipients reconstituted with C57BL/6 cells, none of them were significantly different and the effect size was a modest 0.26 (BV/TV). We next conducted more rigorous bone histomorphometric analyses, which are recorded in Table 2 (recall that each femur was first scanned for µCT analysis and then processed for histomorphometric analysis). Like µCT measurements (Table 1), histomorphometric analyses (Table 2) demonstrated that reconstituting C57BL/6 mice with GATA-1–deficient spleen cells resulted in a significant increase in BV/TV (p < 0.05). Similarly, Pyk2−/− mice reconstituted with C57BL/6 cells had a significantly higher BV/TV than did C57BL/6 mice reconstituted with C57BL/6 cells (p < 0.05). Importantly, similar to our µCT analyses, no difference in BV/TV was observed when Pyk2−/− mice were reconstituted with spleen cells from GATA-1–deficient mice compared with spleen cells from C57BL/6 control mice (p = 0.5). In addition to measuring BV/TV, the numbers of OBs and osteoclasts in the tissue area were also examined. Consistent with our previously published adoptive transfer findings, when C57BL/6 mice were reconstituted with GATA-1–deficient cells, significantly higher numbers of OBs and osteoclasts were observed compared with C57BL/6 mice reconstituted with C57BL/6 cells (p = 0.001 and p < 0.01, respectively). However, when Pyk2−/− mice were reconstituted with GATA-1–deficient cells, no significant differences were detected compared with mice reconstituted with C57BL/6 cells (p = 0.6 and 0.4, respectively). Of interest, when OB number was normalized for bone surface rather than the tissue area, no significant differences were detected (p = 0.2 for C57BL/6 recipients and p = 0.4 Pyk2 recipients). However, analysis of osteoclast surface/bone surface resulted in a similar trend as observed for number of osteoclasts/tissue area. That is, osteoclast surface/bone surface was significantly higher when C57BL/6 mice were reconstituted with GATA-1 cells compared with reconstitution with C57BL/6 cells (p < 0.01). In addition, no significant difference was detected between Pyk2−/− mice reconstituted with GATA-1 cells or C57BL/6 cells (p = 0.9).
Finally, it should be noted that the bone phenotype of nonirradiated C57BL/6 and Pyk2−/− mice was also examined as additional baseline controls. Using µCT, the BV/TV was 4.2 ± 0.3% for C57BL/6 controls and was 17.4 ± 1.0% for Pyk2−/− mice (p < 0.001), again confirming the previously reported high bone mass phenotype.[20, 39] Comparing these measurements to those in Table 1, there was a significant increase in BV/TV in untreated C57BL/6 mice compared with C57BL/6 mice reconstituted with C57BL/6 cells (35% increase). Likewise, there was a dramatic increase in BV/TV in untreated Pyk2−/− mice compared with Pyk2−/− mice reconstituted with C57BL/6 cells (197% increase). The striking decreases in BV/TV in reconstitution studies (likely owing to radiation-induced bone loss) do not lend nonirradiated mice as a useful control for these studies. Of note, virtually identical data were observed by histomorphometric analyses for these baseline control mouse bone phenotypes (data not shown). Taken together, the robust bone phenotype observed when C57BL/6 recipients received GATA-1–deficient cells versus the minimal difference, lacking significance, observed when Pyk2−/− recipients received GATA-1 cells suggests that Pyk2 expression facilitates the MK-mediated enhancement of bone formation in vivo. However, additional de novo pathways are likely able to help in the maintenance of skeletal homeostasis.
Integrins play a central role in regulating tissue homeostasis, and our recent work demonstrates that integrins are also important for MK-mediated regulation of OB proliferation. Although in vitro studies have demonstrated an important interaction between MKs and OBs, the specific details regarding the downstream signaling pathways and the effector molecules involved in this process are unknown. Pyk2 is a nonreceptor protein tyrosine kinase, which is closely related to the focal adhesion kinase (FAK) family. FAK and Pyk2 function in signal transduction pathways that are stimulated by integrin activation. Pyk2 is known to be activated in a Ca2+-dependent manner after integrin engagement and has been linked to the proliferation, migration, and activity of a variety of mesenchymal, epithelial, and hematopoietic cell types.
Our studies and those of others have reported the pivotal role of Pyk2 in bone mass and bone cell function.[35-39] Of importance, Pyk2−/− animals have a significant increase in bone mass and bone formation,[20, 40] and in vitro studies by Buckbinder and colleagues, which were confirmed here (Fig. 7), demonstrated that OBs differentiated from Pyk2−/− mice showed a dramatic increase in alkaline phosphatase activity and mineralization compared with WT controls. Taken together, these data suggest that Pyk2 negatively regulates bone mass in part by reducing osteoblastogenesis.
On the other hand, MKs consistently stimulate WT OB proliferation and we show that MKs stimulate the rapid and sustained increase in total Pyk2 levels by promoting Pyk2 gene transcription in WT OBs (Fig. 1). Also, in the absence of Pyk2, OBs are less responsive to MK-stimulated proliferation (Fig. 2). Furthermore, examination of the role of Pyk2 as an integration point for integrin-mediated signaling in OBs revealed that the MK-mediated increase in Pyk2 levels in OBs was unaltered by treatment of cultures with integrin-blocker RGDS at doses we have previously shown to be efficacious for reducing MK-mediated OB proliferation (Fig. 3). This finding was unexpected because Pyk2 is known to be activated by phosphorylation downstream of integrin engagement in other cell types, including OCs, MKs express and secrete fibronectin,[41-43] a β1-binding integrin ligand, and titration of RGDS into MK-OB cultures resulted in a significant reduction in OB proliferation. It should be noted that our findings do not, however, exclude the possibility that integrin activation in response to MKs results in an increase in the phosphorylation and therefore catalytic activity of Pyk2 in OBs. Additionally, it is also important to note that in our previous studies RGDS blocked MK-induced OB proliferation by approximately 50%, suggesting that a signaling pathway(s) other than β1 integrin signaling may also be involved in regulating MK-mediated OB proliferation. In support of this, Pyk2 is also known to be regulated by G-protein-coupled receptors, inflammatory cytokines, and stress signal activation.[34, 44] In other studies, we are investigating whether Pyk2 activity in OBs is regulated by MK stimulation.
To begin to further dissect the signaling pathways by which MKs regulate OB proliferation, we examined whether stimulation of OBs by MKs required chronic exposure or whether acute exposure (4 hours) could also enhance OB proliferation. Here, we show that Pyk2 expression and OB number are significantly increased in OBs co-cultured with MKs after acute or chronic exposure to MKs. We demonstrated that stimulation of OBs with MKs for 1 hour results in a significant (greater than twofold) increase in total Pyk2 protein levels (Fig. 1A). We also found that enhanced Pyk2 levels are sustained and even elevated with long-term MK stimulation (5 days, Fig. 4B). Additionally, we showed that priming OBs for 4 hours with MKs and then removing the MKs still results in a greater than twofold increase in total Pyk2 protein expression 5 days later (Fig. 4B). Interestingly, we also observed that the increase in Pyk2 levels in OBs stimulated with MKs paralleled the MK-induced increase in OB number. Our data further illustrate the requirement of Pyk2 in the observed reduction of known cell cycle regulator Mdm2 and p53. Thus, the reduction in p53 likely results in the release of cycle arrest mediated by p53 and thereby stimulates proliferation to increase OB number in response to MK. We are currently investigating the mechanism of interaction of Pyk2 with Mdm2 and p53 in OBs and the role of these proteins in OB cell cycle progression.
Although a major goal of this study was to determine whether Pyk2 expression was important in the MK-mediated increase in OB proliferation, to conduct a more rigorous study, we also examined the importance of Pyk2 expression on MK-mediated changes in OB differentiation. We previously reported that the co-culture of MKs with OBs for 14 days dramatically inhibited OB differentiation. Here, we confirm these findings in WT OBs and extend these data by demonstrating that Pyk2−/− OB differentiation, as assessed by OB expression of type I collagen, osteocalcin, and alkaline phosphatase as well as bound calcium as a measure of mineralization, are all significantly reduced with 14 days of co-culture with MKs (Fig. 7A–D). We and others have shown that Pyk2 has a negative effect on OB differentiation. Interestingly, although Pyk2−/− OBs have increased alkaline phosphatase and mineralization compared with WT OBs, MKs reduced OB differentiation in Pyk2−/− OBs to the same extent as that observed in WT OBs. This suggests that MKs can suppress OB differentiation in the absence of Pyk2 and irrespective of the differentiation status of OBs. Moreover, our findings that MKs increase Pyk2 levels in OBs and that Pyk2 expression is important for the MK-mediated increase in OB number are consistent with the idea that MKs act via Pyk2 to increase OB number. Furthermore, as discussed below, it appears that in vivo the MK-mediated increase in OB number may be able to compensate for the reduction in differentiation, resulting in a high turnover state with a net gain in bone mass.
To address the role of Pyk2 in MK-mediated OB differentiation, we further demonstrate that acute exposure of both WT and Pyk2−/− OBs to MKs results in a dramatic reduction in OB differentiation (Fig. 7A–D). In addition to examining markers of OB differentiation in 2-day calvarial OBs cultured in the presence and absence of MKs, we also tested the effects of MKs on osteogenic progenitors. We found that co-culture of both WT and Pyk2−/− BM in the presence of MKs resulted in a significant reduction in mineralization compared with that observed in WT and Pyk2−/− BM cultures alone. Thus, it appears that the effects of MKs on both OB proliferation and differentiation are rapidly initiated upon co-culture and that MKs reduce OB and BM differentiation in co-cultures in the presence or absence of Pyk2. Taken together, our findings that MKs upregulate Pyk2 mRNA and protein levels in OBs soon after cell-cell contact suggests that Pyk2 may play a role in the proliferative effect of MKs on OBs, rather than at later stages of OB mineralization. This finding is also consistent with our recently published study demonstrating that MKs stimulate OB proliferation via Pyk2.
Although our studies begin to unravel the mechanisms by which MKs mediate OB proliferation and differentiation in vitro, it is of clinical importance to determine whether these observations can be replicated in vivo. To address this, we used an adoptive transfer model whereby C57BL/6 and Pyk2−/− recipient mice are reconstituted with C57BL/6 donor cells or GATA-1–deficient donor cells. The GATA-1–deficient mice contain numerous MKs, which allows us to assess whether MKs differentially regulate WT and Pyk2−/− bone mass. As detailed in Tables 1 and 2, consistent with our previous findings, reconstitution of C57BL/6 mice with GATA-1–deficient cells resulted in transfer of both the hematologic and bone phenotype observed in GATA-1–deficient mice.[10, 13, 19] Also, as detailed in Table 2 and consistent with our previously published studies,[10, 11] reconstitution of C57BL/6 mice with GATA-1–deficient cells resulted in what appears to be a high turnover state with a significant increase in both OB and osteoclast number, resulting in a net gain in bone volume. Further, as shown in Tables 1 and 2, Pyk2−/− recipient mice reconstituted with C57BL/6 donor cells exhibited an increase in bone volume compared with C57BL/6 recipient mice reconstituted with C57BL/6 donor cells, as would be predicted based on previous reports of a high bone mass phenotype in Pyk2−/− mice.[20, 40] Of interest, baseline bone measurements of C57BL/6 and Pyk2−/− mice of the same age and sex showed that both mouse strains experienced significant radiation-induced bone loss (compared with C57BL/6 and Pyk2−/− recipient mice, respectively, reconstituted with C57BL/6 cells), but the relative loss was more pronounced for Pyk2−/− mice. This may be a result of the impairment of MK-mediated OB proliferation observed in Pyk2−/− OBs, especially considering the findings of Dominici and colleagues, who demonstrated that lethally irradiated mice (as was done for adoptive transfer studies described here), surviving MKs migrated to endosteal bone surfaces (in close contact with OBs) and stimulated a twofold increase in OB number. Likewise, when Pyk2−/− mice were reconstituted with GATA-1–deficient cells containing significantly increased numbers of MKs, the MK-mediated OB proliferation resulted in only a marginal increase in bone volume compared with the significant increase observed when C57BL/6 recipients were reconstituted with GATA-1-deficient cells rather than C57BL/6 cells. These in vivo data suggest that in Pyk2-deficient recipients, MKs are unable to stimulate a significant increase in bone volume as is observed in WT recipient mice. This suggests that Pyk2 plays a significant role in the OB response to MKs. The small, but not significant, rise in BV/TV in the Pyk2−/− recipient mice transplanted with GATA-1–deficient cells may indicate that other proteins in OBs are also mediated in part by MKs.
Although in the current studies we did not conduct dynamic histomorphometry or measure biochemical markers of bone turnover, our previous investigations demonstrated significant increases in all bone formation indices measured, including an increase in bone formation rate, mineral apposition rate, and osteocalcin levels in the bone and serum after adoptive transfer. Additionally, we observed no differences in urinary Dpd.[10, 11] Thus, there appears to be a paradox between the high bone mass phenotype of Pyk2−/− mice and our findings that MKs stimulate a rapid increase in Pyk2 expression, OB proliferation, and bone formation in vivo. However, it must be remembered that global deletion of Pyk2 results in a significant impairment in osteoclast activity along with enhancement of bone formation.[20, 40] In the current study, we investigated the specific MK-mediated increases in OB Pyk2 expression, and we demonstrate using both in vitro and in vivo approaches that Pyk2−/− OBs are less responsive to MK stimulation than WT OBs.
In conclusion, we demonstrate that Pyk2 expression is important for MK-mediated OB proliferation. We also show that Pyk2 regulates Mdm2 and p53, which are known regulators of the cell cycle. Finally, we demonstrate that Pyk2 expression is positively associated with the MK-induced increase in bone mass. Although in normal BM MKs tend to be localized away from endosteal surfaces, it is known that OBs stimulate megakaryopoiesis by expressing cytokines and growth factors important for MK proliferation, including membrane-bound stem cell factor,[46-48] which is suggestive of paracrine effects. Furthermore, there are a number of human diseases characterized by abnormally elevated numbers and/or distribution of MKs, in which osteosclerosis is evident[1-3] and in which closer proximity of MKs to endosteal surfaces is observed. Perhaps the most convincing evidence of the proximity of MKs to endosteal surfaces is seen in recent radiation studies in which MKs were observed to migrate within close proximity of endosteal surfaces and stimulate a twofold increase in OB number, thus augmenting the so-called endosteal hematopoietic stem cell niches. This in turn enhances hematopoietic stem cell recovery. Thus, the MK-stimulated increase in OB proliferation could be of clinical importance not only for radiation-induced bone loss and bone loss diseases such as osteoporosis but also for cytotoxic regimens required for bone marrow transplantation. Of importance, these studies have begun to dissect the mechanism by which MKs regulate OB proliferation and bone formation and may provide valuable insights into therapeutic interventions in which OBs could be stimulated via a mechanism similar to that mediated by MKs.
All authors state that they have no conflicts of interest.
This work was sponsored in part by the Department of Orthopaedic Surgery at Indiana University School of Medicine, the Department of Oral Biology at Indiana University School of Dentistry, a Biomedical Research Grant and Pilot Funding for Research Use of Core Facilities Award both from Indiana University School of Medicine (MAK), and by the following NIH grants: R03 AR055269 (MAK), R01 AR060332 (MAK, AB), R01 HL55716 (EFS), R01 AR052682 (FMP), and R01 CA109262 (LDM). We sincerely thank Dr. Charles H Turner (posthumously) for helpful discussions and Pfizer for providing us with the Pyk2-deficient mice. We also thank Drs. Stuart Orkin and Ramesh Shivdasani for providing us with the GATA-1–deficient mice. We are thankful to Dr. Matthew Allen for training us in µCT analyses. Finally, we thank the operators of the Indiana University Melvin and Bren Simon Cancer Center Flow Cytometry Resource Facility for their technical help and support. The FCRF is partially funded by P30 CA082709.
Authors' roles: Study conception and design: COO, EFS, LDM, FMP, AB, and MAK. Acquisition of data: Y-HC, RAH, KN, RG-O, DLW, BRC, TEM, HLC, APP, and MAK. Analysis and interpretation of data: Y-HC, RAH, RG-O, DLW, BRC, TEM, EFS, LDM, FMP, AB, and MAK. Drafting of manuscript and revising it critically for important intellectual content: Y-HC, RAH, KN, RG-O, DLW, BRC, TEM, HLC, APP, COO, EFS, LDM, FMP, AB, and MAK. Approving final version of manuscript: Y-HC, RAH, KN, RG-O, DLW, BRC, TEM, HLC, APP, COO, EFS, LDM, FMP, AB, and MAK. MAK and AB take responsibility for the integrity of the data analysis.