Mitogen-activated protein kinase 2 regulates physiological and pathological bone turnover


  • Tobias Braun,

    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    Search for more papers by this author
  • Johannes Lepper,

    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    Search for more papers by this author
  • Gisela Ruiz Heiland,

    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    Search for more papers by this author
  • Willy Hofstetter,

    1. Group for Bone Biology and Orthopedic Research, Department of Clinical Research, University of Bern, Bern, Switzerland
    Search for more papers by this author
  • Mark Siegrist,

    1. Group for Bone Biology and Orthopedic Research, Department of Clinical Research, University of Bern, Bern, Switzerland
    Search for more papers by this author
  • Patrick Lezuo,

    1. Association for the Study of Osteosynthesis Research Institute, Davos, Switzerland
    Search for more papers by this author
  • Matthias Gaestel,

    1. Institute of Biochemistry, Hannover Medical School, Hannover, Germany
    Search for more papers by this author
  • Monika Rumpler,

    1. Ludwig Boltzmann Institute of Osteology at the Hanusch Hospital of Wiener Gebietskrankenkasse and Allgemeine Unfallversicherungsanstalt Trauma Centre Meidling, 1st Medical Department, Hanusch Hospital, Vienna, Austria
    Search for more papers by this author
  • Roman Thaler,

    1. Ludwig Boltzmann Institute of Osteology at the Hanusch Hospital of Wiener Gebietskrankenkasse and Allgemeine Unfallversicherungsanstalt Trauma Centre Meidling, 1st Medical Department, Hanusch Hospital, Vienna, Austria
    Search for more papers by this author
  • Klaus Klaushofer,

    1. Ludwig Boltzmann Institute of Osteology at the Hanusch Hospital of Wiener Gebietskrankenkasse and Allgemeine Unfallversicherungsanstalt Trauma Centre Meidling, 1st Medical Department, Hanusch Hospital, Vienna, Austria
    Search for more papers by this author
  • Jörg HW Distler,

    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    Search for more papers by this author
  • Georg Schett,

    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    Search for more papers by this author
  • Jochen Zwerina

    Corresponding author
    1. Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nurnberg, Germany
    2. Ludwig Boltzmann Institute of Osteology at the Hanusch Hospital of Wiener Gebietskrankenkasse and Allgemeine Unfallversicherungsanstalt Trauma Centre Meidling, 1st Medical Department, Hanusch Hospital, Vienna, Austria
    • Department of Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nuremberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany
    Search for more papers by this author


The objective of this study was to investigate the role of the serine-threonine kinase mitogen-activated protein kinase 2 (MK2) in bone homeostasis. Primary bone cell cultures from MK2+/+ and MK2–/– mice were assessed for osteoclast and osteoblast differentiation, bone resorption, and gene expression. Bone architecture of MK2+/+ and MK2–/– mice was investigated by micro–computed tomography and histomorphometry. Ovariectomy was performed in MK2+/+ and MK2–/– mice to assess the role of MK2 in postmenopausal bone loss. Osteoclastogenesis, bone resorption, and osteoclast gene expression were significantly impaired in monocytes from MK2–/– compared to MK2+/+ mice. Mechanistically, loss of MK2 causes impaired DNA binding of c-fos and nuclear factor of activated T cells cytoplasmic 1 (NFATc1) to tartrate-resistant acid phosphatase (TRAP) and the calcitonin receptor gene promoter. In addition, MK2–/– mice showed an age-dependent increase in trabecular bone mass and cortical thickness, fewer osteoclasts, and lower markers of bone resorption than MK2+/+ mice. Furthermore, MK2–/– mice were protected from ovariectomy-induced bone loss. Osteoblastogenesis and bone formation were unchanged in MK2–/– mice, whereas osteoblast expression of osteoprotegerin (OPG) and serum levels of OPG were higher in MK2–/– than in MK2+/+ mice. Loss of MK2 effectively blocks bone resorption and prevents the development of postmenopausal bone loss. Small-molecule inhibitors of MK2 could thus emerge as highly effective tools to block bone resorption and to treat postmenopausal bone loss. © 2013 American Society for Bone and Mineral Research.


Development and maintenance of bone mass and bone quality is a continuously ongoing process during life. Bone loss and impaired bone quality cause osteoporosis, which is a major health problem in the aging population and is associated with high morbidity and mortality.1 Bone loss is the result of a negative net balance of bone resorption and bone formation. Enhanced bone resorption, which is not fully compensated by bone formation, is a common mechanism for the induction of osteoporosis. Osteoclasts are the sole bone-resorbing cells and stem from the monocyte lineage. Their function is tightly regulated by cytokines and growth factors, reflecting close interaction between the immune system and the bone.

Over the past decade, pathways involved in the immune system have been recognized as important mediators of bone metabolism, giving rise to the term osteoimmunology.2 The first molecular link was established when receptor activator of NF-κB ligand (RANKL) was discovered as a crucial osteoclast-inducing molecule.3 Thus, many proinflammatory molecules such as tumor necrosis factor (TNF) and interleukin 1 (IL-1) can increase expression of RANKL, thereby directly acting on osteoclast formation.4, 5 Moreover, osteoblast formation is concomitantly reduced by the same cytokines aggravating bone loss.6 Thus, chronic inflammatory diseases such as rheumatoid arthritis and ankylosing spondylitis can directly cause osteoporosis.7, 8 Interfering with the immune system may thus have direct effects on bone metabolism.9

Mitogen-activated protein kinases (MAPKs) are a large group of intracellular signal transduction molecules that integrate a variety of cellular signals such as growth, inflammation, and danger signals. There are three major families of MAPK: the extracellular signal-regulated kinases (ERKs), c-Jun N-terminal kinases (JNKs), and p38MAPK; and each family of MAPK is organized as a cascade in which an upstream kinase phosphorylates and thus activates the downstream kinase.10 The p38MAPK family consists of four different isoforms, termed p38MAPKα, -β, -γ, and -δ, which face dual phosphorylation at threonine 180 and tyrosine 182 residues upon activation. MAPK-activated protein kinase-2 (MK2) and -3 (MK3) are the main targets of p38MAPKα and β.11 MK2 and MK3 are both Ser/Thr protein kinases, but MK2 seems to be of superior physiological importance compared to MK3.12 Activation of MK2 is involved in cellular stress responses, nuclear transport, cell proliferation, gene expression, and cytokine signaling.13 MK2 also positively regulates upstream signaling of p38MAPK via feedback mechanisms. Consequently, loss of MK2 renders mice resistant to an established model of chronic destructive arthritis.14 Further, MK2-deficient mice are protected from allergen-induced airway inflammation and even sustain less damage in rodent models of neuroinflammation.15, 16

Based on its involvement in cytokine synthesis, we hypothesized that MK2 has an important role of MK2 in bone metabolism. We thus investigated of the role of MK2 on skeletal architecture and bone metabolism under physiological conditions as well as during postmenopausal bone loss by using mice deficient in MK2.

Materials and Methods


MK2-deficient mice (C57BL6 genetic background) lacking functional activity of MK2 have been described.13 Male MK2–/– and wild-type mice at indicated ages were used to determine the bone phenotype (n = 8/group and age). Ovariectomy was performed in 10-week-old female wild-type and MK2–/– mice (n = 8/group). Mice were anesthetized with ketamine (2 mg) and xylazine (0.2 mg). Mice were either sham operated or received ovariectomy via flank surgery.17 The removed ovaries were analyzed histologically (2-µm paraffin sections stained with hematoxylin and eosin) to confirm correct ovariectomy. Eight weeks after ovariectomy, mice were euthanized and analyzed. The local ethics committee approved all animal procedures.

Histological analysis and bone histomorphometry

Tibias were fixed overnight in 4% formalin and then decalcified in 14% EDTA until bones were pliable. Decalcified tibias were embedded in paraffin and serial sections of 2 µm were stained with tartrate-resistant acid phosphatase (TRAP; Sigma, St. Louis, MO, USA) as described17 before. The other tibia was fixed overnight in 4% formalin and then transferred to 70% ethanol. Prior to histology, tibias were measured by micro–computed tomography (µCT). After CT scans, undecalcified tibias were embedded in methylmethacrylate (Technovit; Heraeus Kulzer, Wehrheim, Germany), 3- to 4-µm sections were prepared, deplastinated in methoxymethylmetacrylate (Merck, Darmstadt, Germany) and stained using the Goldner staining method.

Bone histomorphometry was performed using a microscope (Nikon, Japan) equipped with a digital camera and an image analysis system (OsteoMeasure; OsteoMetrics, Decatur, GA, USA). The following parameters were measured according to international standards18: fraction of bone volume of the total sample volume (bone volume/tissue volume [BV/TV]), trabecular number (Tb.N), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th), osteoclast surface per bone surface (OC.S/BS), number of osteoclasts per bone perimeter (OC.N/BP), osteoblast surface per bone surface (OB.S/BS), and number of osteoblasts per bone perimeter (OB.N/BP). In vivo bone formation was assessed by dynamic histomorphometry: calcein (30 mg/kg; Sigma) was injected subcutaneously 7 and 2 days before mice were euthanized. Mineral apposition rate (µm/d) was determined in undecalcified tibia sections using Osteomeasure.

µCT imaging

µCT measurements were performed using the µCT40 scanner (Scanco, Bruettisellen, Switzerland).19 Tibias were placed in sample holders and placed perpendicularly to the X-ray beam. A scout view scan was performed to define the region of interest (ROI), which was subsequently scanned at 12-µm resolution. The X-ray tube with a 5-mm focal spot was set at 70 kV and 114 mA, and the integration time was 200 ms. After reconstruction of the 3D image of the ROI, the thresholds allowing the separation of mineralized and soft tissues were defined.

Biomechanical testing

Isolated femoral bones from wild-type and MK2–/– mice were tested ex vivo in four-point bending experiments according to international standards. Bones were kept hydrated with Ringer's lactate solution during experiments. The crosshead speed of the material testing system (ElectroForce, Model 3220; Bose EFSG, Eden Prairie, MM, USA) was 2.1 degrees/min. The linear portion of the measured and acquired curve was used to calculate the bending stiffness for each femur; for noise reduction of the acquired data (500 Hz rate) a Savitzky-Golay filter was implemented in a custom-written Matlab program. (Matlab R14SP1; Mathworks, Bern, Switzerland).

Ex vivo osteoclastogenesis assay

Bone marrow cells were isolated from tibias and femurs of 8-week-old wild-type and MK2–/– mice. Cells were cultured in α modified essential medium (α-MEM) containing 10% fetal bovine serum (FBS) and macrophage colony-stimulating factor (M-CSF) (30 ng/mL) after red blood cell lysis. After 24 hours, nonadherent cells were seeded onto 48-well plates (1 × 106 cells per well) and stimulated with M-CSF (30 ng/mL) and RANKL (10 ng/mL) (both from R&D Systems, Minneapolis, MN, USA). Medium was changed on day 3. After 4 days, TRAP staining was performed to evaluate osteoclast differentiation. Osteoclasts were identified by the presence of three or more nuclei and purple stain.

For in vitro bone resorption assays, bovine cortical bone slices (300-µm slices) were cut with a diamond saw and sterilized in 70% ethanol for 5 minutes before use. Osteoclasts (5 × 105 cells/cm2) were seeded onto bones and cultured for 7 days with M-CSF (30 ng/mL) and RANKL (10 ng/mL). After 7 days, bone slices were washed, sonicated for 10 minutes to remove cells, and air dried. Images were obtained by reflection light microscopy from the entire surface and analyzed with standard image analysis software (ImageJ; NIH) to calculate the resorbed area (%).

Primary osteoblast culture

Calvaria were dissected from newborn mice and digested with trypsin (1 mg/mL; Invitrogen, Grand Island, NY, USA), followed by dispase (2 mg/mL; Roche Diagnostics, Indianapolis, IN, USA) and then three times with collagenase A (2 mg/mL; Roche Diagnostics). Supernatant was collected and cells (5000 cells/cm2) were cultured in α-MEM containing 10% FBS and stimulated with ascorbic acid (50 µg/mL) and β-glycerol-phosphate (10 mM) for 3 weeks after confluence. Medium was changed every 7 days. Bone nodule formation was detected using the alizarin red staining.

Alkaline phosphatase (ALP) activity was measured using an established enzymatic assay. Briefly, alkaline phosphatase (ALP) substrate was prepared from solution A (8 ng naphtol-AS-TR phosphate/300 µL N,N-dimethylformamide; both from Sigma) and solution B (24 mg fast blue BB salt/30 mL of 100 mmol/L Tris HCl, pH 9.6). The two solutions were combined, 10 mg of MgCl2 was added, and after sterile filtration the resulting solution was used immediately. Enzymatic activity was determined in cell lysates that were solubilized with 0.1% Triton X-100. Aliquots (20 µL) of each sample were incubated with 100 µL of AP substrate buffer (100 mmol/L diethanolamine, 150 mmol/L NaCl, 2 mmol/L MgCl2 p-nitrophenylphosphate at 2.5 µg/mL) for 5 to 30 minutes at room temperature. Total cellular protein was determined using the bicinchoninic method (Pierce Chemical Co., Rockford, IL, USA). ALP activity is expressed as U/mg protein with 1 U defined as enzymatic activity that released 1 µmol of p-nitrophenol/min.

Coculture experiments

Primary osteoblasts (0.5 × 106 per well) were cultured with bone marrow cells (1 × 106 cells per well) in α-MEM containing 10% FBS, 1α25(OH2) vitamin D3 (1 × 10−8 M), and dexamethasone (1 × 10−7 M) in 24-well plates. Cells were stained for TRAP to identify osteoclasts after 9 days of culture.

Colony-forming unit assays

Granulocyte-macrophage–colony-forming unit assay

Bone marrow cells were isolated as described in “Ex vivo osteoclastogenesis assay.” A total of 1 × 107 nonadherent cells were cultured in α-MEM containing 1.2% methylcellulose, 30% fetal calf serum (FCS), 1% bovine serum albumin, and granulocyte-macrophage (GM)-CSF (2 ng/mL) in a 136-mm Petri dish. Colonies per square centimeter (cm2) were counted after 7 days using an inverted microscope.

Colony-forming unit–osteoblast assay

A total of 1 × 107 nonadherent bone marrow cells were cultured in α-MEM and 10% FCS in a 136-mm Petri dish. After 7 days, osteogenic medium supplemented with β-glycerophosphate (10 mmol/L) and L-ascorbic acid (50 µg/mL) was added and then changed every 7 days. Von Kossa staining was performed using a standard protocol after 28 days of culture.

Real-time RT-PCR

RNA was isolated from cell cultures using Trizol (Invitrogen). RNA was reverse transcribed using the Gene Amp RNA PCR kit (Applied Biosystems, Foster City, CA, USA). Quantitative real-time RT-PCR was performed using LightCycler technology (Roche Diagnostics) and the Fast Start SYBR Green I kit for amplification and detection using a standardized program. Expression of the target gene was normalized to the expression of β-actin. Primer sequences are available upon request.

Fluorescence-activated cell sorting analysis

Bone marrow cells were incubated with anti-CD11b fluorescein isothiocyanate (FITC)-labeled antibody in fluorescence-activated cell sorting (FACS) buffer (PBS + 2% FCS). Cells were analyzed using a FACScalibur (Becton-Dickinson, Franklin Lakes, NJ, USA) after removal of the unbound antibody.


Serum and cell culture supernatant levels of RANKL, osteoprotegerin (OPG), C-terminal telopeptides of type I collagen (RatLaps), and osteocalcin were detected using ELISA according to the manufacturers' recommendations (RANKL and OPG from R&D Systems; RatLaps from Immunodiagnostic Systems [IDS], Scottsdale, AZ, USA; and osteocalcin from Biomedical Technologies Inc., Stoughton, MA, USA).

c-Fos and nuclear factor of activated T cells cytoplasmic 1 chromatin immunoprecipitation

After isolation of bone marrow cells from tibias and femurs of 10-week-old wild-type and MK2–/– mice, cells were cultured for 3 days in α-MEM containing 10% fetal bovine serum and M-CSF (30 ng/mL) after red blood cell lysis. Thereafter, osteoclast precursors were plated at a density of 3 × 106 cells per 10-cm dish and treated with M-CSF (30 ng/mL) and RANKL (10 ng/mL) for 3 days. Subsequently, cells were cross-linked with 1% final concentration of formaldehyde at 37°C for 10 minutes before harvest. Cell lysis, chromatin shearing, c-fos and nuclear factor of activated T cells cytoplasmic 1 (NFATc1) immunoprecipitation, and DNA cleanup were performed using the ChampionChip 1-day kit (SABiosciences, Fredrick, MD, USA) following the manufacturer's instructions. For immunoprecipitation 5 µg of anti-c-fos (sc-413X), 5 µg of anti NFATc1 antibody (sc-7294, both from Santa Cruz Biotechnology, Santa Cruz, CA, USA), or 5 µg of nonimmune serum (as negative control) were used. Before immunoprecipitation, 1% (10 µL) of the chromatin was saved and stored at 4°C for further use as reference. For quantitation of the quantitative real-time RT-PCR values, for each sample DNA signal of the precipitated chromatin was normalized to the nonprecipitated chromatin. c-Fos and NFATc1 binding was assessed on Calcr and Trap promoters. Promoter sequences (NW_001030802 for Calcr and CH466522 for Trap) were analyzed by Genomatix online software (Genomatix GmbH, Munich, Germany; For Calcr, the proximal 391 bp preceding the transcriptional start site (TSS) containing a putative NFATc1 and a putative activator protein 1 (AP-1) binding site and promoter P3 were analyzed.20 For Trap, promoter P1 and P2, as described by Reddy and colleagues,21 were analyzed. Primer sequences for chromatin immunoprecipitation (ChIP) promoters are available upon request.

Statistical analysis

Data are shown as the mean ± SD. Group mean values were compared by paired Student's t test.


MK2 regulates osteoclastogenesis in vitro

Bone marrow mononuclear cells were isolated and stimulated with M-CSF and RANKL in vitro. MK2 deficiency significantly impaired the ability of osteoclast precursors to differentiate into osteoclasts (osteoclasts/well: MK2+/+ 535 ± 66 versus MK2–/– 249 ± 55; p < 0.001). Moreover, MK2–/– osteoclasts were much smaller and contained significantly fewer nuclei than MK2+/+ osteoclasts (p < 0.001; Fig. 1A). Osteoclast formation was determined at different time points to exclude delayed osteoclastogenesis in MK2–/– cells. However, there were significantly fewer osteoclasts in MK2–/– cultures than in MK2+/+ cultures at every time point investigated (p < 0.03; Fig. 1B). We also assessed in vitro bone resorption of MK2+/+ and MK2–/– osteoclasts; thus, bone resorption was significantly impaired in MK2-deficient osteoclasts compared to MK2+/+ osteoclasts (resorbed bone surface: MK2+/+ 1.86% ± 0.43%; MK2–/– 0.41% ± 0.57%; p < 0.05; Fig. 1C). To exclude the possibility that the observed difference in osteoclast differentiation is caused by decreased numbers of osteoclast precursors in the bone marrow of MK2–/– mice, we isolated bone marrow mononuclear cells and assessed the number of CD11b-positive cells. There was no significant difference in the percentage of CD11b-positive cells in the bone marrow of MK2–/– mice (mean, 79.6%) and MK2+/+ controls (mean, 79.4%; p = not significant [ns]; Fig. 1D). Further, we stimulated bone marrow cells with GM-CSF and detected colony forming units (CFUs) after 7 days. MK2+/+ and MK2–/– cells showed no significant difference in the number of CFUs (Fig. 1D).

Figure 1.

Impaired osteoclastogenesis in MK2-deficient mice. (A) TRAP staining after stimulation of bone marrow mononuclear cells from wild-type (WT) MK2+/+ mice and MK2–/– mice with M-CSF and RANKL for 4 days. Representative image (scale bar = 200 µm). Graphs: Mean ± SD number of osteoclasts per well and number of nuclei per osteoclast in osteoclast cultures of WT mice and MK2–/– mice (n = 6 independent experiments performed in triplicates; ***p < 0.001). (B) Time course of osteoclastogenesis. (Means from three independent experiments performed in triplicates (p < 0.03). (C) Ex vivo bone resorption (n = 3 independent experiments performed in triplicates; *p < 0.05). (D) Number of CD11b-positive bone marrow mononuclear cells detected by flow cytometry (means from five independent experiments) and CFU-GM assay (means from three independent experiments). (E) Relative mRNA expression of TRAP, RANK, OSCAR, c-fos, NFATc1, and MMP9 assessed by quantitative real-time PCR after 4 days stimulation with M-CSF and RANKL (n = 6 independent experiments; ***p < 0.001, **p < 0.01, *p < 0.05). (F) cfos and NFATc1 chromatin immunoprecipitation (ChIP) for Trap P2 promoter, Calcr P3 promoter, and Calcr proximal promoter in preosteoclasts stimulated with RANKL (10 ng/mL). (n = 3 independent experiments; ***p < 0.001, **p < 0.01, *p < 0.05). (G) Osteoclasts in coculture of WT MK2+/+ or MK2–/– osteoblasts and WT MK2+/+ or MK2–/– monocytes stained for TRAP (n = 3 independent experiments; *p < 0.05).

Next, we determined RNA expression of osteoclast-associated genes in MK2+/+ and MK2–/– osteoclasts. As shown in Fig. 1E, relative mRNA expression of TRAP (–27% ± 8%, p < 0.01), RANK (–22% ± 10%, p < 0.05), osteoclast-associated receptor (OSCAR) (-47 ± 10%, p < 0.001) and matrix metalloproteinase 9 (MMP9) (mean ± SD: –15% ± 10%; p < 0.05) was significantly reduced in MK2–/– osteoclasts as compared to wild-type controls. There was no difference in the relative mRNA expression of c-fos and NFATc1. There was also no significant difference in c-fos and NAFTc1 mRNA expression in preosteoclasts at 1, 2, 4, 8, 12, and 24 hours after stimulation with RANKL (data not shown). c-Fos and NFATc1 are key transcription factors involved in osteoclastogenesis. They were described19, 20 to bind on the promoters of Calcr and Trap, two main osteoclastic genes. As shown in Fig. 1F, knockdown of the MK2 gene clearly reduced NFATc1 binding on the Calcr P3 promoter, Calcr proximal promoter, and Trap P2 promoter in preosteoclastic monocytes stimulated with M-CSF and RANKL using ChIP experiments. Similar results were obtained when c-fos binding to the three promoters were analyzed (Fig. 1F).

Increased OPG production in MK2-deficient osteoblasts

To determine whether osteoblasts could also contribute to a bone phenotype, we performed osteoblast-monocyte coculture experiments. There was no significant difference in the number of osteoclasts depending on the genotype of the osteoblasts. These data suggest that impaired osteoclastogenesis is most likely due to an intrinsic osteoclast defect (Fig. 1G).

Osteoblast precursor cells were isolated from the calvaria and stimulated with ascorbic acid and β-glycerol-phosphate for 3 weeks to generate osteoblasts. We could not detect any differences in cell proliferation (data not shown) and differentiation. As shown in Fig. 2A, there was no significant difference in bone nodule formation or ALP activity. However, OPG production by osteoblasts from MK2–/– mice (mean ± SD: 7.5 ± 0.6 ng/mL) was significantly higher than by MK2+/+ osteoblasts (5.4 ± 1.8 ng/mL, p < 0.05). MK2-deficient osteoblasts did also produce higher amounts of RANKL (MK2+/+: 23.09 ± 2.9 pg/mL versus MK2–/– 32.8 ± 2.6 pg/mL, p < 0.01; Fig. 2B). There was no significant difference in the ratio of RANKL to OPG (data not shown). We next assessed mRNA expression of various osteoblast genes such as OPG, RANKL, Runt-related transcription factor 2 (RUNX2), osterix (OSX), osteocalcin, type I collagen and transforming growth factor (TGF)-activated kinase 1 (TAK1) by RT-PCR. Relative mRNA expression of OPG (p < 0.05) and RANKL (p < 0.05) was significantly higher in osteoblasts from MK2–/– mice than from wild-type controls. Furthermore, there was no significant difference in mRNA expression of the other tested genes (Fig. 2C). Moreover, we cultured bone marrow cells with osteogenic medium and determined no significant difference in the number of CFUs (mean ± SD, wild-type: 58 ± 8 CFU/plate versus MK2–/–: 54 ± 10 CFU/plate; p = 0.32).

Figure 2.

Effects of MK2 on osteoblasts. (A) Bone nodule formation detected by alizarin red staining and ALP activity of osteoblasts from wild-type (WT) MK2+/+ mice and MK2–/– mice (means from four independent experiments performed in duplicates). (B) Level of OPG and RANKL in cell culture supernatants of osteoblasts after stimulation with ascorbic acid and β-glycerol-phosphate for 3 weeks (n = 4 independent experiments performed in duplicates; **p < 0.01, *p < 0.05). (C) Relative mRNA expression of OPG, RANKL, RUNX2, osterix, osteocalcin, type I collagen, and TAK1 detected by quantitative real-time RT-PCR (n = 4 independent experiments; *p < 0.05).

Loss of MK2 leads to increased bone mass and mechanical stability

To determine the consequences of MK2 deficiency on bone in vivo, we performed µCT analysis of tibial bones from 11-week-old male MK2+/+ and MK2–/– mice. In accordance with data from cell culture experiments, bone mass increased by 125% in MK2–/– mice (286 ± 38 mg hydroxyapatite [HA]/cm3) as compared to MK2+/+ controls (127 ± 11 mg HA/cm3; p < 0.001). BV/TV was also significantly higher in MK2-deficient mice (22% ± 3.5%, p < 0.001) than in the controls (9% ± 0.4%). The observed increase in bone mass was caused by increased numbers of bony trabeculae (MK2+/+ 4.9 ± 0.4/mm versus MK2–/– 7.3 ± 0.7/mm; p < 0.001) and thicker trabeculae (MK2+/+ 30 ± 0 µm versus MK2–/– 40 ± 0 µm). Tb.Sp was consequently decreased in MK2-deficient mice (MK2+/+: 196 ± 17 µm versus MK2–/–: 124 ± 15 µm, p < 0.001). There were also differences in cortical bone structure. Ct.Th (MK2+/+ 0.08 ± 0.01 mm versus MK2–/– 0.12 ± 0.02 mm) and cortical bone fraction (MK2+/+ 0.49% ± 0.03% versus MK2–/– 0.56% ± 0.02%) were increased in MK2-deficient mice (Fig. 3A).

Figure 3.

Higher trabecular bone mass and mechanical stability in MK2-deficient mice. (A) Representative 3D µCT images of tibias from 11-week-old wild-type (WT) MK2+/+ mice and MK2–/– mice. Scale bar = 1 mm. Bone mass, trabecular volume, trabecular number, trabecular thickness, trabecular separation, and cortical parameters of the tibial metaphysis of WT MK2+/+ mice and MK2–/– mice as assessed by µCT (n = 5 mice per group; ***p < 0.001). (B) Four-point-bending experiments with femoral bones from WT MK2+/+ mice and MK2–/– mice (n = 5 mice; *p < 0.05).

To test biomechanical properties of bones from MK2+/+ and MK2–/– mice, we isolated femoral bones and performed four-point-bending experiments. Interestingly, bending stiffness was significantly increased in MK2-deficient mice (MK2+/+ 1.1 ± 0.1 Newton millimeter (Nmm)/degree versus MK2–/– 1.4 ± 0.4 Nmm/degree, p < 0.05; Fig. 3B). Thus, the loss of MK2 increases bones mass and biomechanical properties in vivo.

High bone mass due to lower osteoclast numbers in mice deficient for MK2

To determine whether the increase of bone mass in MK2–/– mice is age-dependent, we performed bone histomorphometry of tibial bones from 4-, 11-, and 26-week-old mice. Interestingly, we could not detect any significant differences in bones from 4-week-old mice. However, we observed an increase in BV/TV in 11-week-old MK2–/– mice (MK2+/+ 8.9% ± 2.4%, MK2–/– 12.5% ± 1.3%, p < 0.01). As observed in µCT analysis, this increase was due to a higher number of bony trabeculae (MK2+/+ 2.3 ± 0.6/mm, MK2–/– 3.1 ± 0.3/mm, p < 0.05) and reduced Tb.Sp (MK2+/+ 404 ± 92 µm, MK2–/– 281 ± 34 µm, p < 0.05), whereas trabecular thickness was not altered (MK2+/+ 38 ± 8 µm, MK2–/– 40 ± 3 µm, p = ns). This increase in BV/TV in MK2–/– mice was preserved when analyzing the 26-week-old MK2+/+ and MK2–/– mice (Fig. 4A).

Figure 4.

Low osteoclast numbers in MK2-deficient mice. Histomorphometry of tibial bones of wild-type (WT) MK2+/+ mice and MK2–/– mice. Parameters include: (A) trabecular volume, number, separation, and thickness; (B) osteoclast surface/bone surface, osteoclast number/bone perimeter; and (C) osteoblast surface/bone surface and osteoblast number/bone perimeter. The results represent means of 5 mice (***p < 0.001, **p < 0.01, *p < 0.05). (D) Mineral apposition rate determined by calcein labeling. The results represent means of 5 mice (*p < 0.05).

To investigate the cause for the increased bone mass in MK2–/– mice, we analyzed bone resorption and bone formation in vivo. Thus, we determined whether osteoclast and osteoblast numbers changed and performed dynamic histomorphometry to assess bone formation. As expected from the BV/TV data, we could observe no difference between 4-week-old MK2+/+ and MK2–/– mice. However, a significant decrease of the number osteoclasts per bone perimeter (OC.N/BP) was found in 11-week-old (MK2+/+ 4.8 ± 1.0/mm versus MK2–/– 2.6 ± 0.7/mm, p < 0.01) and 26-week-old (MK2+/+ 5.6 ± 0.4/mm versus MK2–/– 3.8 ± 0.1/mm, p < 0.001) MK2-deficient mice. Similar changes were found when we analyzed OC.S/BS (Fig. 4B). With respect to osteoblasts, there was no significant difference in osteoblast number per bone perimeter and OB.S/BS between MK2–/– and controls at all ages investigated (Fig. 4C). There was, however, a tiny increase in mineral apposition rate in 11-week-old MK2–/– mice compared to the MK2+/+ controls but not in the other age groups investigated (Fig. 4D).

Low systemic bone resorption in MK2-deficient mice

We next analyzed serum levels of bone resorption as well as bone formation in MK2+/+ and MK2–/– mice at different time points (Fig. 5). Collagen type I cleavage products (RatLaps) indicating osteoclastic activity were significantly reduced in MK2–/– mice compared to controls at the age of 4 weeks (MK2+/+ 85.6 ± 11.8 ng/mL versus MK2–/– 60.7 ± 8.9 ng/mL, p < 0.01) and 11 weeks (MK2+/+ 65.7 ± 16.8 ng/mL versus MK2–/– 38.0 ± 13.2 ng/mL, p < 0.01), but did not show a significant difference after 26 weeks. In accordance, serum OPG levels were elevated in MK2–/– mice at the age of 4, 11, and 26 weeks, but not at the age of 2 weeks. Serum RANKL levels were not different among MK2+/+ and MK2–/– mice. Serum levels of osteocalcin were significantly reduced in MK2–/– mice at 11 weeks (MK2+/+ 52.9 ± 12.0 ng/mL versus MK2–/– 21.4 ± 11.3 ng/mL, p < 0.01), but they were unchanged at 4 and 26 weeks of age.

Figure 5.

Low bone resorption markers in MK2-deficient mice. Serum levels of: (A) RANKL and OPG; and (B) type I collagen cleavage products (RatLaps) and osteocalcin in wild-type (WT) MK2+/+ mice and MK2–/– mice, as detected by ELISA. The results represent mean values from 5 mice (***p < 0.001, **p < 0.01, *p < 0.05).

MK2 deficiency protects from bone loss due to estrogen deficiency

To determine whether MK2 is involved in bone loss due to estrogen deficiency, we performed either ovariectomy or sham surgery in 10-week-old female MK2+/+ and MK2–/– mice. Bone was assessed 8 weeks after surgery. As expected, control MK2+/+ mice showed significantly reduced BV/TV (p < 0.01) and Tb.N (p < 0.01) after ovariectomy, whereas Tb.Sp (p < 0.01) was increased. In contrast, MK2-deficient mice did not lose trabecular bone after ovariectomy, as determined by µCT (Fig. 6A-B). Serum levels of collagen type I cleavage products were significantly elevated in ovariectomized MK2+/+ mice (p < 0.05), whereas ovariectomized MK2–/– mice showed no increase in bone resorption parameters (Fig. 6C). Serum levels of osteocalcin were significantly higher in MK2+/+ mice than in MK2–/– mice after sham surgery or ovariectomy. RANKL was significantly increased after ovariectomy both in MK2+/+ and MK2–/– mice (p < 0.05), whereas OPG levels did not change after ovariectomy but were higher in MK2–/– mice as mentioned above (Fig. 5A).

Figure 6.

Protection from ovariectomy-induced bone loss in MK2-deficient mice. µCT with representative images (scale bar = 1 mm) (A) and histomorphometry (B) of tibial bones of in wild-type (WT) MK2+/+ mice and MK2–/– mice after sham surgery (CTL) or ovariectomy (OVX). Trabecular volume, number, separation, and thickness was assessed 8 weeks after ovariectomy (n = 4 mice per group; ***p < 0.001, **p < 0.01, *p < 0.05). (C) Serum levels of RANKL, OPG, type I collagen cleavage products (RatLaps), and osteocalcin were detected by ELISA 8 weeks after ovariectomy (n = 4 mice per group; **p < 0.01, *p < 0.05).


In this study, we show a specific role of the MAPK MK2 in bone cells and the skeletal system in vitro and in vivo. We show that MK2-deficient cells fail to develop properly into osteoclasts. Osteoclast differentiation and expression of key osteoclast genes is impaired in the absence of MK2. In vivo, MK2 deficiency leads to lower numbers of osteoclasts, lower biochemical markers of bone resorption, higher bone mass, and increased mechanical properties of the bones. Moreover, mice lacking MK2 showed protection to ovariectomy-induced bone loss.

MAPKs are important regulators of both osteoblastogenesis and osteoclastogenesis in vitro. Previous studies suggested that all three MAPK pathways, p38MAPK, ERK, and JNK, are involved in bone cell physiology. For instance, pharmacologic inhibition of the ERK upstream activator MEK by PD 98059 potently inhibits osteoclast differentiation and, more potently, survival.22, 23 Confirmatory, genetic disruption of ERK1 (p44) reduced osteoclast differentiation, migration, and activity, whereas deletion of ERK2 (p42) had only minor effects in vitro.24 Importantly, ERK1 deletion in hematopoietic cells in vivo caused higher bone mass. However, ERK is also important for the growth, early differentiation, and function of osteoblasts.25 The net effect of a combined deletion of ERK in both osteoblasts and osteoclasts is still unclear. JNK1, but not JNK2, regulates RANKL-induced osteoclastogenesis but JNK1-deficient mice have no apparent bone phenotype.26 This might be a result of the importance of JNK for late-stage differentiation of osteoblasts.27

The p38 MAPK pathway is regulated upstream by MKK3 and MKK6.28 In mammalian cells, four p38 isoforms (α, β, γ, and δ) have been identified. p38α and β are widely expressed, whereas tissue expression of p38γ and δ is restricted.29 MK2 is an important downstream kinase of p38α and β.30 p38 MAPK has long been implicated as important for osteoclast biology. Earliest data arose from the use of the p38 inhibitor SB203580, which showed potent inhibition of RANKL-induced osteoclastogenesis.31 Also, TNF-mediated osteoclastogenesis can be prevented by addition of p38 inhibitors.32 Moreover, monocyte-specific deletion of p38α inhibited osteoclastogenesis in vitro and in vivo.33 In contrast, Greenblatt and colleagues34 identified TAK1 as an upstream kinase regulating MKK3/6 and p38α and β activity in osteoblasts. Osteoblast-specific deletion of TAK1 and global deletion of MKK3, MKK6 or p38β caused impaired bone formation in vivo.

Contrasting the effects of kinases upstream of MK2 on bone formation, MK2-deficient primary osteoblasts develop normally. Also, osteoblast-specific gene expression such as osteocalcin and runx2 were similarly expressed in wild-type and MK2–/– cells. Consequently, osteoblasts of MK2-deficient mice support osteoclastogenesis in vitro, similar to wild-type cells. Further, in vivo analysis of osteoblast function by dynamic bone histomorphometry did not indicate any effect of MK2 on bone formation. These data point to substantial differences in the role of individual MAPK family members in the regulation of bone mass and reflect earlier observations of differences in the role of p38 and MK2 in skin inflammation.35 Importantly, however, MK2-deficient osteoblasts produced increased amounts of OPG and the serum levels of OPG were consistently higher in MK2-deficient mice. In contrast, osteoblasts secreted also more RANKL in vitro whereas serum levels were not altered; the reason for this is unclear. Thus, the low bone resorption phenotype of MK2-deficient mice is partially based on increased synthesis of OPG and disruption of RANKL-RANK.

In contrast, we could clearly find an intrinsic phenotype in osteoclasts derived from MK2–/– mice. First, monocytes fail to develop properly into osteoclasts in vitro. This was associated with decreases in gene expression of osteoclast-associated molecules such as TRAP, OSCAR, and RANK. Mechanistically, we found that DNA binding of c-fos and NFATc1 to TRAP and calcitonin receptor are important for osteoclastogenesis, was significantly reduced in MK2-deficient preosteoclasts stimulated with RANKL. Interestingly, the phenotype in vivo evolves over time and is not present after birth. We could detect the first biochemical changes at 4 weeks of age, whereas bone mass increases until 11 weeks of age. The reasons for these findings are currently unclear, but might be based on a mild osteoclast defect allowing resorption of primary spongiosa and thus preventing development of full-blown osteopetrosis. Alternatively, age-associated changes in pathways affecting bone development may cause the delayed in vivo effect.

Bone histomorphometry and µCT analyses revealed a profound effect of MK2 deficiency on bone architecture. We found a consistent increase in bone mineral density in MK2–/– mice. Further, we could also detect an increase in cortical thickness in these mice. Both may together contribute to the increased biomechanical properties we found using four-point bending experiments. Thus, our findings may suggest a significant effect on MK2 on bone health. Because MK2 is an important inflammatory mediator, these effects may be even more significant when MK2 is inhibited in chronic inflammatory disease associated with bone loss such as rheumatoid arthritis.

In conclusion, MK2 is involved in the regulation of osteoclast differentiation and activity. The observation that absence of MK2 protects from ovariectomy-induced bone loss is particularly interesting because it indicates that small-molecule–based inhibition of MK2 could be a strategy to treat postmenopausal osteoporosis by mitigating RANKL-RANK–mediated osteoclastogenesis by inducing endogenous OPG production. Small-molecule inhibitors specifically targeting MK2 inhibitors have been developed but have so far not been tested in models of bone loss.36 The observations that (1) postmenopausal bone loss is associated with increased secretion of pro-osteoclastogenic cytokines and (2) neutralization of these cytokines can protect from ovariectomy-induced bone loss supports the concept that MK2, based on its role in integrating proinflammatory and pro-osteoclastogenic signals, represents an interesting target for protecting bone by small-molecule signaling inhibitors.


All authors state that they have no conflicts of interest.


This study was supported by the Deutsche Forschungsgemeinschaft (FG 661/TP4 and SPP1468-IMMUNOBONE), the Bundesministerium für Bildung und Forschung (BMBF; project ANCYLOSS), the MASTERSWITCH project of the European Union, the Austrian Science Fund (P21524-B12), and the IMI-funded project BTCure.

Authors' roles: Study design: TB, MG, KK, JHWD, GS and JZ. Study conduct: TB, JL, GRH, WH, MS, PL, MR, RT. Data collection: TB, JL. GRH, WH, MS, PL, MR, RT. Data analysis: TB, JL, WH, KK, JHWD, GS, JZ. Data interpretation: TB, JL, WH, KK, JHWD, GS, JZ. Drafting manuscript: TB, GS, JZ. Revising manuscript content: TB, JL, WH, KK, JHWD, GS, JZ. Approving final version of manuscript: all authors. TB, GS and JZ take responsibility for the integrity of the data analysis.