Osteoblast-specific inactivation of p53 results in locally increased bone formation

Inactivation of the tumor suppressor p53 (encoded by the Trp53 gene) is relevant for development and growth of different cancers, including osteosarcoma, a primary bone tumor mostly affecting children and young adolescents. We have previously shown that deficiency of the ribosomal S6 kinase 2 (Rsk2) limits osteosarcoma growth in a transgenic mouse model overexpressing the proto-oncogene c-Fos. Our initial aim for the present study was to address the question, if Rsk2 deficiency would also influence osteosarcoma growth in another mouse model. For that purpose, we took advantage of Trp53fl/fl mice, which were crossed with Runx2Cre transgenic mice in order to inactivate p53 specifically in osteoblast lineage cells. However, since we unexpectedly identified Runx2Cre-mediated recombination also in the thymus, the majority of 6-month-old Trp53fl/fl;Runx2-Cre (thereafter termed Trp53Cre) animals displayed thymic lymphomas, similar to what has been described for Trp53-deficient mice. Since we did not detect osteosarcoma formation at that age, we could not follow our initial aim, but we studied the skeletal phenotype of Trp53Cre mice, with or without additional Rsk2 deficiency. Here we unexpectedly observed that Trp53Cre mice display a unique accumulation of trabecular bone in the midshaft region of the femur and the humerus, consistent with its previously established role as a negative regulator of osteoblastogenesis. Since this local bone mass increase in Trp53Cre mice was significantly reduced by Rsk2 deficiency, we isolated bone marrow cells from the different groups of mice and analyzed their behavior ex vivo. Here we observed a remarkable increase of colony formation, osteogenic differentiation and proliferation in Trp53Cre cultures, which was unaffected by Rsk2 deficiency. Our data thereby confirm a critical and tumorigenesis-independent function of p53 as a key regulator of mesenchymal cell differentiation.


Introduction
The p53 tumor suppressor is known to regulate the transcription of various genes, which eventually causes specific cellular responses, such as inhibition of proliferation to prevent detrimental effects of genotoxic stress [1]. Mutations of p53 were identified in more than 50% of primary tumors, and heterozygous germline mutations in the p53-encoding gene (TP53) are the cause of Li-Fraumeni syndrome, a disease characterized by increased cancer incidence with a 10% risk to develop osteosarcomas [2]. The tumor suppressor function of p53 has also been demonstrated by targeted deletion of the murine p53-encoding gene (Trp53), albeit the incidence of specific tumors was slightly different when compared to Li-Fraumeni syndrome [3]. More specifically, the majority of Trp53-deficient mice develop lymphomas, which also reduces their lifespan and the development of other cancer types such as osteosarcoma. However, although the tumor spectrum associated with p53 inactivation was remarkably different in Trp53-heterozygous mice, our previous skeletal analysis of 12-month-old Trp53 +/animals revealed that osteosarcoma development was only detectable, when the gene encoding the protein tyrosine phosphatase Rptpz was deleted simultaneously [4]. Therefore, we took advantage of mice expressing the Cre recombinase under the control of the Runx2 promoter [5] with the aim to delete p53 selectively in osteoblast progenitor cells to enable osteosarcoma development. The initial purpose of our study was to utilize this conditional p53 inactivation model, to combine it with a deficiency of Rsk2 and to monitor osteosarcoma incidence and growth.
Rsk2 is a broadly expressed serine/threonine kinase, which in response to growth factor stimulation phosphorylates specific proteins, including p53, to influence various cellular functions [6,7]. Inactivating mutations of human RSK2, which is located on the X chromosome, cause Coffin Lowry syndrome, a rare disease characterized by mental and psychomotoric retardation, but also by skeletal abnormalities [8,9]. The impact of Rsk2 inactivation on the skeleton was also evidenced by the generation and analysis of Rsk2 -/0 mice, which display low bone mass due to a cell-autonomous osteoblast defect, but also dental abnormalities, such as alveolar bone loss and impaired cementum formation [10][11][12]. With respect to osteosarcoma development we have previously shown that Rsk2 deficiency has a remarkable protective influence on tumor growth in c-Fos-transgenic mice. More specifically, while c-Fos-transgenic mice develop osteosarcomas in different skeletal locations, which strongly increase in size from 1 to 7 months of age, the tumor area did not increase in Rsk2-deficient c-Fos-transgenic mice [11,13]. Although these remarkable findings suggested that Rsk2 is essential for osteosarcoma progression, they were not fully unexpected, since c-Fos is one of the known substrates of the Rsk2 kinase activity [14,15]. Therefore, it was important to address the question if Rsk2 deficiency can also limit osteosarcoma progression in other mouse models.
Here we describe that the majority of 6-month-old Trp53 fl/fl;Runx2-Cre (thereafter termed Trp53 Cre ) mice displayed thymic lymphomas, due to additional Trp53 recombination in the thymus, whereas osteosarcomas were not detected. Although this unexpected finding did not allow us to analyze the influence of Rsk2 deficiency on osteosarcoma development, we unexpectedly observed a remarkable accumulation of trabecular bone in the femur and humerus midshaft region of Trp53 Cre mice, which did not appear in Rsk2-deficient Trp53 Cre mice. Analysis of cultured bone marrow cells from mice of the different genotypes further revealed that Trp53 Cre cultures displayed increased osteogenic differentiation and proliferation, regardless of their Rsk2 genotype. These data identify p53 as a key regulator of mesenchymal cell differentiation, whose expression in osteoblast progenitors is required to limit physiological bone formation.

Animals
All mice used in this study were male littermates from matings of Trp53 fl/fl (thereafter termed Trp53 fl ) males and Trp53 Cre females, the latter being heterozygous for the Rsk2-deficient allele (Rps6ka3 tm1.1Kry Trp53 tm1Brn Tg(Runx2-icre)1Jtuc/Uke). Mice of all genotypes were available and studied previously by our group [5,11,16]. Initially, genotyping of Trp53 fl mice was performed with the primers 5'-AGC ACA TAG GAG GCA GAG AC-3' and 5'-CAC AAA AAC ACG TTA AAC CCA G-3', amplifying a 370 bp and 288 bp fragment for floxed and wildtype allele of Trp53, respectively. Genotyping for Trp53 recombination at the loxP-sites was performed with primers 5'-CAC AAA AAC ACG TTA AAC CCA G-3' and 5'-GAA GAC AGA AAA GGG GAG GG-3', amplifying a 612 bp fragment for the recombined Trp53 allele. Genotyping for the Runx2-Cre transgene was performed using the primers 5'-TGG CTT GCA GGT ACA GGA-3', 5'-CCA GGA AGA CTG GAA GAA GC-3' and 5'-GGA GCT GCC GAG TCA ATA AC-3', which amplify a 780 bp fragment of the endogenous Runx2 gene and a 600 bp fragment of the Runx2-Cre transgene, respectively [5]. Genotyping for Rsk2 was performed using the primers 5'-TTG TTG GTT TAC TTT CTT TCG GTC TG-3', 5'-AAG ATG ATT GCT TTG CTT AGT TTA-3', amplifying a 230 bp and 320 bp for the wildtype and mutant Rsk2 allele, respectively [11]. All mice were kept with a 12-hour light/dark cycle and had access to tap water and standard rodent chow (1328P, Altromin Spezialfutter GmbH & Co. KG, Germany) ad libitum. Mice were analyzed at a maximum age of 6 months, mostly before the occurrence of any visible symptoms. Mice were anesthetized with an 80% (v/v) CO 2 /O 2 gas mixture before switching to 100% CO 2 . Sample size was determined by availability, however, with the exception of FACS analysis (S1 Fig), at least 6 mice were analyzed per genotype. Therefore, in this study, 75 mice were used. All animal experiments were approved by the animal facility of the University Medical Center Hamburg-Eppendorf and by the "Behörde für Soziales, Familie, Gesundheit und Verbraucherschutz" (Org869 and 984).

μCT analysis
For μCT analysis the right femur and the right humerus were dissected from the mice fixed in 3.7% PBS-buffered formaldehyde for 24 h and fixed in the holder (F20mm � H75mm). The holder was placed in a μCT 40 desktop cone-beam microCT (Scanco Medical, Swizerland). Anatomical separation of the skull was performed in the same way, but a different holder was used (F30mm � H75mm). μCT scanning and analysis were performed with a voxel resolution of 10 μm for bone analysis and 15 μm for dental analysis as previously described using a μCT 40 desktop cone-beam microCT (Scanco Medical, Swizerland) according to standard guidelines [17,18]. Sample was analyzed in the distal or mid metaphysis in a volume situated 2500 μm to 500 μm proximal of the distal growth plate. A threshold value of 250 was implemented.

Histology
Skeletons were fixed in 3.7% PBS-buffered formaldehyde for 24 h and subsequently stored in 80% ethanol. For bone histology, the lumbar vertebral bodies L1 to L4 and the right tibia of each mouse were dehydrated in ascending alcohol concentrations and then embedded in methylmetacrylate. Sections of 4 μm thickness were cut in the sagittal plane on a Microtec rotation microtome (Techno-Med GmbH, Germany) and stained by von Kossa/van Gieson, as described previously [17]. Paraffin histology was performed on the skull. The skulls were placed into the 10% EDTA (pH7.4) for at least 4 weeks, followed by embedding in paraffin. Sections of 3.5 μm were cut on a semi-automated rotary microtome (Leica Biosystems Nussloch GmbH) and stained by toluidine blue, as described previously [17]. Histomorphometry was performed according to the ASBMR guidelines [19] using the OsteoMeasure histomorphometry system (Osteometrics Inc., USA).

Ex vivo osteogenic differentiation of bone marrow cells
Bone marrow cells were collected by centrifugation from long bones of mice with the four different genotypes. The cells were then plated in 12-well plates at a density of 5x10 6 cells/well in α-MEM (Sigma-Aldrich, USA) supplemented with 10% (v/v) FBS (American Type Culture Collection, USA) and 100 U/ml penicillin/streptomycin (Life Technologies, USA). To induce osteoblastic differentiation the medium was additionally supplemented with 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate for 10 days. Alizarin red staining and quantification of mineralization was performed as described previously [17].

Colony formation assay
Bone marrow cells were collected as described above, and 9 x 10 6 cells were plated into a Ø15 cm dish. Cells were cultured with α-MEM medium (Sigma-Aldrich, USA) supplemented with 15% (v/v) FBS (American Type Culture Collection, USA) and 100 U/ml penicillin/streptomycin (Life Technologies, USA). After 14 days the cultures were washed with PBS and fixed with 3.7% PBS-buffered formaldehyde for 90 seconds. Staining was performed with alkaline phosphatase substrate solution (0.6 mg/ml fast blue RR salt, 0.25% (w/v) naphthol AS-MX phosphate (both Sigma-Aldrich, USA), pH 8.6) for 30 min. All colonies with more than 50 cells were counted. There further, hematoxylin (Sigma-Aldrich, USA) was used to determine the total colony number.

Proliferation and gene expression
Bone marrow cells were plated in a 96-well plate for proliferation analysis and in 12-well plate for expression analysis. Cells were cultured with α-MEM medium (Sigma-Aldrich, USA) supplemented with 10% (v/v) FBS (American Type Culture Collection, USA) and 100 U/ml penicillin/streptomycin (Life Technologies, USA). For proliferation analysis, the Amersham cell proliferation biotrak Elisa kit (GE Healthcare, UK) was used and measured with a microplate reader at an absorbance of 450 nm within 5 minutes. For gene expression analysis, RNA was isolated using the peqGOLD TriFast™ (Darmstadt, DE), according to manufacturer's instructions. Concentration and quality of RNA were measured using a NanoDrop ND-1000 system (NanoDrop Technology). Complementary DNA synthesis was performed using the Verso cDNA Synthesis Kit (Thermo Fisher). Expression analysis by qRT-PCR was performed using a StepOnePlus system and predesigned TaqMan gene expression assays (Applied Biosystems). Gapdh expression was used as an internal control, and presented data are given as relative expression towards Gapdh.

FACS analysis
For fluorescence activated cell sorting (FACS) we isolated bone marrow from long bones of Trp53 fl and Trp53 Cre mice. Cells were stained with fluorescent conjugated antibodies against CD45, CD34, CD11b, CD11c, F4/80, CD3e, CD4 and CD45R (BD Biosciences, San Jose, CA). Analyses were performed using a FACS Aria Fusion (BD Biosciences, San Jose, CA) and data were analyzed using FlowJo software (Tree Star Inc, Ashland, OR).

Statistical analysis
All data in the manuscript are presented as means ± standard deviations. Data were analyzed using GraphPad Prism Software (Graphpad Software Inc.). Statistical significance for the comparison of four groups was calculated using one-way ANOVA followed by a post hoc comparison with Bonferroni's method. For the comparison of two groups data were analyzed by twotailed Student's t-test using Excel software (Microsoft Corp., Redmond, WA, USA). p-values below 0.05 were considered as statistically significant.

Trp53 fl/fl;Runx2-Cre mice develop thymic lymphomas, but no osteosarcomas
For this study we generated and analyzed male littermates with four different genotypes (Trp53 fl;Rsk2+/0 ; Trp53 Cre;Rsk2+/0 ; Trp53 fl;Rsk2-/0 ; Trp53 Cre;Rsk2-/0 ). When we analyzed these mice at the age of 6 months we found, by contact Xray, no indication of osteosarcoma development in any genotype ( Fig 1A). Unexpectedly however, we identified thymic lymphomas in the majority of Trp53 Cre mice, whereas this incidence was apparently reduced in Rsk2-deficient Trp53 Cre littermates (Fig 1B). We next analyzed the specificity of Cre-mediated recombination in different tissues by genomic PCR. We hereby observed that Trp53 Cre mice displayed recombination within the Trp53 locus, not only in calvaria and femur (after removal of the bone marrow by centrifugation), but also in skin and thymus (Fig 1C).
Although this finding precluded to study the influence of Rsk2 deficiency on osteosarcoma formation in the Trp53 Cre model, we took of advantage of these animals to analyze a potential impact of p53 inactivation on differentiation of osteoblast lineage cells. In a first step we determined body weight and the weights of different organs (spleen, liver, heart, fat), but we did not detect statistically significant differences between the four groups of mice ( Fig 1D). These data suggested that Trp53 Cre mice, despite having thymic lymphomas, do not display a multi-organ pathology, at least not at 6 months of age.

Trp53 fl/fl;Runx2-Cre mice display increased trabecular bone mass in the femur and humerus diaphysis
We next analyzed the femur bone architecture in the respective animals by micro computed tomography (μCT) (Fig 2A). Here we identified a remarkable accumulation of trabecular bone in the mid diaphysis specifically in Trp53 Cre mice. Moreover, while the bone volume per tissue volume (BV/TV) was significantly reduced in Rsk2-deficient mice, the additional Rsk2 deficiency also corrected the high trabecular bone mass phenotype of Trp53 Cre mice (Fig 2B). There were also subtle changes in the cortical bone compartment observed in Trp53 Cre mice (Fig 2C). More specifically, these animals displayed moderately reduced cortical thickness and increased cortical porosity, and again this was not observed in Trp53 Cre;Rsk2-/0 littermates (Fig 2D).
At that point we decided to focus on the Trp53 Cre model in order to support the μCT findings. Here we first analyzed undecalcified femur sections from Trp53 fl and Trp53 Cre littermate mice (Fig 3A). We identified increased trabecular bone mass in the mid diaphysis of Trp53 Cre mice, but no significant difference towards controls in the distal diaphysis (Fig 3B). To substantiate these findings at the level of gene expression we dissected the femora and removed the epiphysis as well as the bone marrow by centrifugation. We then prepared RNA from the

PLOS ONE
p53 inhibits excessive bone formation femoral bones for qRT-PCR expression analysis. As expected, we observed increased expression of osteoblast markers (Alpl, Bglap and Col1a1) in femora of Trp53 Cre mice (Fig 3C). Consistent with a previous report on Trp53-deficient mice and experiments with primary calvarial osteoblasts [20], we also observed that Sp7 (encoding the transcription factor Osterix) was expressed at higher levels in Trp53 Cre mice. We additionally identified significantly higher expression of Tnfsf11 (encoding RANKL), whereas expression of Tnfrsf11b (encoding osteoprotegerin) was not different between Trp53 fl and Trp53 Cre mice, unlike it was reported by others for p53-deficient mice [21]. We further used the bone marrow for FACS analysis, where we analyzed the CD45-positive cell populations. Despite a moderate increase of CD11c/CD45 + dendritic cells however, we did not observe significant differences for the most abundant cell types (S1 Fig). To address the question, if the increased trabecular bone mass is only observed in femora or also in other skeletal elements of Trp53 Cre mice, we additionally analyzed spine and tibia sections. Whereas no differences in trabecular bone parameters were observed between Trp53 fl and Trp53 Cre littermate mice for the spine (Fig 4A), there was significantly increased trabecular bone mass in the tibia of Trp53 Cre animals, albeit this phenotype did not extend into the mid diaphysis ( Fig 4B). Therefore, we additionally performed μCT scanning of the humerus, which can be regarded as the forearm equivalent of the femur. Here we again identified an increased amount of trabecular bone in the mid diaphysis of Trp53 Cre mice ( Fig 4C). Finally, we measured serum levels of the bone turnover biomarkers PINP and CTX, as well as RANKL and OPG concentrations, but here we did not detect significant differences between Trp53 fl and Trp53 Cre mice (Fig 4D). Taken together, these findings demonstrated that Trp53 Cre mice do not display a systemic high bone mass phenotype and suggested that p53 is physiologically required to limit trabecular bone formation in specific anatomical locations.

Trp53 Cre bone marrow cells display increased osteogenesis, colony formation and proliferation
We next isolated bone marrow cells from long bones of Trp53 fl and Trp53 Cre mice and studied their behavior ex vivo. Here we first assessed their osteogenic differentiation capacity by alizarin red staining of mineralized matrix formed after 10 day of culture in the presence of ascorbic acid and ß-glycerophosphate (Fig 5A). We hereby identified more than two-fold increase of mineralization in Trp53 Cre cultures. We additionally plated the bone marrow cells at lower density to let them establish colonies. By staining the osteogenic colonies for alkaline phosphatase we again observed a remarkable increase in Trp53 Cre cultures (Fig 5B). Importantly however, the increased colony forming capacity of Trp53 Cre bone marrow cells was not restricted to osteogenic colonies, since the total colony number was also substantially increased ( Fig 5C).
This finding led us to analyze the proliferation of bone marrow cells immediately after plating. Here we identified, using BrdU incorporation assay, a more than 2-fold increased proliferation rate in Trp53 Cre cultures (Fig 5D). Consistently, expression of the p53 target gene Cdkn1a, encoding the cell cycle inhibitor p21 [22], was strongly repressed in Trp53 Cre cultures, and the same was the case for Casp9 [23], which is required for p53-dependent apoptosis ( Fig  5E). We additionally performed the same set of experiments with bone marrow cells from Rsk2-deficient Trp53 fl and Trp53 Cre mice. Here we observed the same differences of osteogenesis, colony formation, proliferation and gene expression between bone marrow cultures from Trp53 Cre;Rsk2-/0 and Trp53 fl;Rsk2-/0 mice (S2 Fig). Taken together, these findings provided further evidence for a role of p53 as a negative regulator of osteoblastogenesis. They additionally suggested that the correction of the Trp53 Cre high bone mass phenotype by Rsk2 deficiency is not explained by the lack of a direct molecular interaction between p53 and Rsk2.

Runx2-Cre-mediated p53 inactivation
Since it was reported that Rsk2 phosphorylates p53 at Ser15, thereby acting upstream of p53 to promote a DNA damage response [24], it was not expected that the lack of such phosphorylation would affect the phenotype caused by p53 deficiency. On the other hand, it would have been possible that an alteration of p53 activity in Rsk2-deficient mice would contribute to their phenotypes. Since Trp53 Cre mice displayed a Rsk2-independent trabecular bone phenotype, we therefore focused on another pathologically affected tissue in Rsk2-deficient mice, i.e. the dental apparatus [12]. By assessing the dental phenotypes in the four groups of mice (Trp53 fl;Rsk2+/0 ; Trp53 Cre;Rsk2+/0 ; Trp53 fl;Rsk2-/0 ; Trp53 Cre;Rsk2-/0 ) by μCT imaging we identified alveolar bone loss, indicated by the visible root area, in Rsk2-deficient animals, regardless of their Trp53 status ( Fig  6A). Similarly, both Rsk2-deficient groups displayed a reduced area of cellular cementum ( Fig  6B), which is a specialized hard tissue maintaining the teeth in the alveolar bone. These data indicate that the dental pathologies of Rsk2-deficient mice do not dependent on the presence of the Rsk2 substrate p53.

Discussion
Oncogenic transformation of mesenchymal cells can cause osteosarcoma, the most prevalent primary bone tumor primarily affecting children or adolescents [25]. Osteosarcoma development is also a rare event in mice, but there is one particular mouse model displaying high incidence of osteosarcomas in different skeletal locations with 100% penetrance. More specifically, these mice carry a transgene causing ubiquitous over-expression of the oncogene c-Fos [13]. We have previously found that deficiency of the ribosomal S6 kinase 2 (Rsk2) strongly reduces osteosarcoma growth in cFos-transgenic mice [11]. However, since Rsk2 is known to phosphorylate two serine residues (Ser 362 and 374) of c-Fos, thereby stabilizing the protein to allow subsequent phosphorylation events, this remarkable influence of Rsk2 deficiency was to some extent expected. Therefore, it was important to address the question, if Rsk2 deficiency would also prevent osteosarcoma development in other mouse models.

Rsk2-/0
) with the aim to analyze osteosarcoma progression. Unexpectedly, due to Trp53 recombination in the thymus, we identified a high incidence of thymic lymphomas in Trp53 Cre mice, which did not allow us to let them get older in order to develop osteosarcomas. Importantly, however, when we analyzed the different groups of mice we identified another fully unexpected pathology in Trp53 Cre mice, i.e. increased trabecular bone formation in the midshaft of femur and humerus. Although such a phenotype is in principal agreement with previous evidence showing that p53 deficiency in mice enhances proliferation and osteogenic differentiation of mesenchymal stem cells to cause an increased bone mass phenotype [20,21,[26][27][28][29][30], we are not aware of a conditional p53 inactivation model with such a specific localized influence on trabecular bone formation.
Skeletal development, growth and remodeling are highly complex processes, which are regulated by a variety of cell types and specific molecules [31]. Moreover, even after development and growth are completed, there is a steady remodeling of the bone matrix, which is a prerequisite for life-long skeletal stability [32]. This process is mediated by two antagonistically acting cell types, i.e. bone-resorbing osteoclasts and bone-forming osteoblasts [33,34]. Importantly, these two cell types are fundamentally different, not only in terms of progenitor cells, morphology and mode of action but also with respect to specific molecules regulating their differentiation and function [35]. It is also noteworthy that there are two distinct compartments in adult bones, i.e. cortical and trabecular bone, the latter being remodeled at higher rate [36]. This architecture is of key importance, as it provides non-mineralized space for the bone marrow, which does not only contain hematopoietic cells, but also mesenchymal stromal cells, the latter representing a pool of osteoblast progenitor cells [37]. Especially in the long bones of the extremities, such as the femora or the humeri, there is essentially no trabecular bone in the midshaft region of the marrow cavity under physiological circumstances. Therefore, it is certainly possible to state that the phenotype observed in Trp53 Cre mice is an example of non-physiological or may be even ectopic bone formation. This implies that, at least in mice, p53 expression in osteoprogenitor cells is required to limit their bone-forming capacity at specific skeletal locations.
Another surprising observation was that additional Rsk2 deficiency apparently prevented the increased trabecular bone phenotype of Trp53 Cre mice. However, since Rsk2-deficient mice display a low bone mass phenotype, this finding did not specifically demonstrate a genetic interaction between p53 and Rsk2 in controlling bone formation. Therefore, it was important to compare the osteogenic potential of bone marrow cells derived from the four groups of mice. The results we obtained here were truly informative, since there was a remarkable difference between Trp53 fl and Trp53 Cre cultures, but no normalization of this difference by additional Rsk2 deficiency. Vice versa, we did not observe a correction of the dental phenotype in Rsk2-deficient mice by conditional inactivation of p53. From these latter findings it is certainly possible to conclude that the dental abnormalities caused by Rsk2 deficiency are not explained by the lack of Rsk2-mediated p53 phosphorylation.
Despite the serendipity of our findings, we recognize the limitations of the present study. First, as already stated above, we mainly confirm but also extent previous findings reported for p53-deficient mice or their respective osteoblast cultures. Second, we only used one Cre-expressing mouse line, which does not allow conclusions about the stage of osteoblast differentiation that is most critical for the function of p53. Third, we only analyzed unsorted bone marrow cells in ex vivo experiments, which did not allow to address the question, if p53 inactivation in lymphoid cells is a major driver of the increased osteogenesis we observed. Therefore, our findings provide the basis for future experimental approaches that should not only include the use of additional Cre-deleters to inactivate p53 at different stages of osteoblast differentiation, but also co-cultures of sorted bone marrow cell populations from Trp53 fl and Trp53 Cre mice. Especially since there is hallmark evidence for a crosstalk between bone and immune cells [38], such future experiments are certainly required to fully understand the complexities of p53 action in the bone marrow, but exceeds the scope of this manuscript.
Regardless of these limitations, and despite the fact that our findings did not allow us to study the impact of Rsk2 deficiency on osteosarcoma development, it is important that we observed a fully unexpected and highly specific skeletal phenotype in Trp53 Cre mice. Our collective data demonstrate that p53, in a Rsk2-independent manner, is required to limit trabecular bone accumulation in the appendicular skeleton, at least in mice. qRT-PCR analysis for p21 and Casp9 expression, relative to Gapdh, in undifferentiated bone marrow cells from mice of the indicated genotypes. Data represent mean ± SD (n � 6). Asterisks indicate statistically significant differences ( � p<0.05, �� p<0.005, ��� p<0.0005). (TIF)