The Role of Dot1l in Prenatal and Postnatal Murine Chondrocytes and Trabecular Bone

ABSTRACT Osteoarthritis and osteoporosis are widely prevalent and have far‐reaching public health implications. There is increasing evidence that epigenetics, in particular, histone 3 lysine 79 methyltransferase DOT1L, plays an important role in the cartilage and bone biology. In this study, we evaluated the role of Dot1l in the articular cartilage, growth plate, and trabecular bone utilizing conditional KO mouse models. We generated chondrocyte‐specific constitutive and inducible conditional Dot1l KO mouse lines using Col2a1‐Cre and Acan‐CreER systems. Prenatal deletion of Dot1l in mouse chondrocytes led to perinatal mortality, accelerated ossification, and dysregulation of Col10a1 expression. Postnatal deletion of Dot1l in mouse chondrocytes resulted in trabecular bone loss decreased extracellular matrix production, and disruption of the growth plate. In addition, pharmacological inhibition of DOT1L in a progeria mouse model partially rescued the abnormal osseous phenotype. In conclusion, Dot1l is important in maintaining the growth plate, extracellular matrix production, and trabecular bone. © 2019 The Authors. JBMR Plus published by Wiley Periodicals, Inc. on behalf of American Society for Bone and Mineral Research.


Introduction
O steoarthritis is the most common arthritis affecting over 30 million people in the United States. (1) The estimated direct medical cost of osteoarthritis is $185.5 billion per year (2) and is expected to increase as the population ages and becomes more obese. With the exception of weight loss, (3,4) there is currently no convincing disease-modifying therapy to halt the progression of osteoarthritis, and patients eventually need joint replacement. Although overall very safe and effective, up to 3% of patients require critical care services after elective arthroplasty. (5) Additionally, complicated revision surgery may be necessary in up to 10%-to 12% of patients within 10 years of hip or knee arthroplasty. (6,7) These statistics are especially problematic for younger patients, who have longer life expectancies and also have higher activity levels.
Similarly, osteoporosis affects 5.1% of men and 24% of women over age 65 in the United States and is an important public health issue in the elderly. (8) Resulting fractures, in particular hip fractures, can be devastating with up to 13% mortality in the first 30 days after surgery and up to 33% mortality within the first year after surgery. (9) Although bisphosphonates, teriparatide, and denosumab are effective treatments, these medications do carry risks including osteonecrosis, atypical fractures, and in the case of denosumab, increased infection. (10,11) Recently, there is growing evidence that epigenetics has an important role in cartilage and bone biology. Disruptor of telomeric silencing 1 (Dot1) and its mammalian homolog DOT1L (DOT1-Like) are a novel class of non-redundant histone 3 lysine 79 (H3K79) methyltransferases. (12,13) It has been shown that human Dot1l polymorphism rs12982744 is associated with increased risk of osteoarthritis in European and Chinese populations in genome-wide association studies. (14,15) Dot1l is expressed in the normal growth plate and articular cartilage of prenatal and skeletally mature mice, (16) and may preserve cartilage health by preventing the hyperactivation of Wnt signaling, (17) inhibiting osteoclastogenesis, (18) and preventing age-related and posttraumatic osteoarthritis. (19) This study further assesses the role of Dot1l in prenatal and postnatal chondrocytes and trabecular bone in vivo using conditional KO mouse models. USA). Animals were housed in a standard animal facility maintained by the Animal Resource Center at the University of Chicago. A conditional Dot1l KO mouse line with loxp sites around the second exon was generated previously utilizing the Knockout Mouse Project repository. (20) For evaluation of prenatal Dot1l deletion, the Dot1l KO mouse line was crossed with a Col2a1-Cre mouse line (21) until Dot1l KO allele homozygosity and Col2a1-Cre heterozygosity (Dot1l Δ/Δ ; Col2-Cre) were obtained. For control animals, littermates including Dot1l Δ/Δ , Dot1l wt/Δ , Dot1l wt/Δ ; Col2a1-Cre genotypes were used.

Dot1l deletion and evaluation of cell proliferation
Three-week-old weaning age mice with genotype Dot1l Δ/Δ ; Acan-CreER, and gender-matched littermate Dot1l Δ/Δ were used for Dot1l-deletion experiments unless otherwise specified. Dot1l was deleted by intraperitoneal injection of two doses of 150 mg/ kg of tamoxifen (Millipore-Sigma T5648) dissolved in corn oil. (22) Deletion of Dot1l was assessed with chondrocyte genomic DNA PCR and chondrocyte RNA qPCR of Dot1l.
For cell proliferation evaluation, Dot1l was deleted from 3-week-old mice for 4 weeks, and then intraperitoneally injected with 75 mg/kg of bromodeoxyuridine (BrdU). Mice were euthanized 48 hours after BrdU injection for tissue harvest.
Chondrocyte harvest and genomic DNA and RNA extraction Femoral head (4-week-old mice and 7-week-old mice) and xiphoid process (15-week-old mice) cartilage were dissected, placed in QuickExtract (Lucigen) for DNA or TRIzol (Invitrogen) for RNA, and homogenized. DNA or RNA was extracted per manufacturer instructions. DNA samples were used for genomic DNA PCR to confirm Dot1l second exon excision. RNA samples were reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and used for qPCR of Dot1l, Col2a1, and Acan expression. qPCR was performed with iTaq Universal SYBR Green Supermix (Bio-Rad). qPCR primers are listed in Supplemental Table S2.

Whole-mount Alcian Blue-Alizarin Red staining
Postnatal day 2 (P2) mice were fixed in 95% ethanol for 24 hours with agitation. Skin was removed and organs were eviscerated. The mice were placed in Alcian Blue 8GX (Millipore-Sigma) cartilage staining solution for 1 week and dehydrated in 95% ethanol for 1 week. Soft tissues were cleared with 1% potassium hydroxide (KOH) solution for 3 to 4 days and placed in Alizarin Red (Millipore-Sigma) bone-staining solution for several days until staining was complete. Soft tissues were further cleared in graded KOH solutions. Samples were stored in glycerolformaldehyde solution before acquiring images. mRNA in situ hybridization P2 mice were prepared for in situ hybridization as described previously. (23) Briefly, dissected limbs were sunk in 20% sucrose-10% formalin in PBS, embedded in gelatin, and cut on a sledge microtome. Sections were permeabilized and hybridized with digoxigenin-(DIG-) labeled riboprobes, and DIG-labeled RNA duplexes were detected with alkaline phosphatase/anti-DIG Fab conjugate. Sense controls were always negative. The colorimetric reaction was carried out using 4-nitro blue tetrazolium-5-bromo-4-chloro-3-indolyl phosphate (NBT-BCIP; Roche Molecular Diagnostics) as substrates.
μCT image acquisition and image analysis Knee joints of 15-week-old mice were harvested after 12 weeks of Dot1l deletion. Mouse knee joint trabecular bone architecture at the metaphysis was assessed using μCT (μCT50; Scanco Medical) at the Rush University MicroCT Core in Chicago (https://www. rushu.rush.edu/research/rush-core-laboratories/rush-microct-andhistology-core). Samples were scanned at 55 kVp and 145 μA, with a 500-ms integration time and 6-μm isotropic voxel size. 3D images of knee joint and distal femoral and proximal tibial quantitative trabecular and cortical bone morphology variables were obtained using associated manufacturer's software. Because of inherent differences between male and female mice morphology such as bone volume/total volume (BV/TV) and trabecular thickness (Tb.Th), representative data from male mice are shown.

Histological evaluation
Knee joint tissue was harvested after mice were euthanized and fixed overnight in 4% paraformaldehyde. Tissues were subsequently washed and stored in 70% ethanol solution. After the samples were scanned with μCT, they were decalcified for 2 days at room temperature with Osteosoft (Millipore-Sigma 101728). Tissues were embedded in paraffin and sectioned by the University of Chicago Histology Core (https://htrc.uchicago.edu/ services.php?service=1).
H&E stain was performed by the University of Chicago Histology Core. For Safranin O staining, sections were deparaffinized, incubated in Safranin O solution for 5 min, and counterstained with hematoxylin. For immunohistochemistry, sections were incubated with primary antibodies for chondroitin sulfate 1-500 (monoclonal CS-56; Millipore-Sigma) or BrdU 1:100 (BD Biosciences) overnight and appropriate secondary antibodies. Negative control slide for BrdU was from P5 mouse femur without BrdU injection stained with BrdU antibody. Sections were developed with the DAB kit (Pierce).
Quantifications of articular cartilage and growth plate thickness and stain intensity were performed using ImageJ software (NIH, Bethesda, MD, USA; https://imagej.nih.gov/ij/). Length was measured on the image and adjusted according to a scale bar. The ratio of intensity density (IntDen) of the region of interest over background was reported as stain intensity.

Statistical analysis
Results were expressed as the mean AE SD for bar graphs. Student's t test (two-tailed, unpaired) was used for bar graphs and scatter plots. p < 0.05 was considered statistically significant per convention.
DOT1L-inhibitor-treated Zmpste24-deficient mice Mice forelimb samples were generous gifts from Dr López-Otín from Universidad de Oviedo, Spain. These samples were from 6-week-old Zmpste24 KO mice intraperitoneally injected with EPZ-5676 inhibitor (2.5 mg/kg body weight; Selleck Chemicals) or vehicle once per day, every day for 12 weeks. Age-matched untreated WT mice served as control. There were two forelimbs per each group.
Once formalin-fixed forelimb samples were received, they were washed in 70% ethanol and images were acquired with University of Chicago PaleoCT (http://luo-lab.uchicago.edu/ paleoCT.html). Samples were scanned at a voltage of 70 kV, current of 115 μA, exposure time of 200 ms, and isotropic voxel size of 4.7 μm. Raw data were processed by datos|x software (GE Healthcare) reconstruction program using the Feldkamp algorithm. The projections underwent automatic geometric calibration, and were cropped and sliced. Resulting images were analyzed with ImageJ software using BoneJ plugin.

Results
Altered ossification in prenatal Dot1l-deleted mice To understand the role of Dot1l in the development of prenatal cartilage, we generated a Dot1l KO model under the control of the constitutive Col2a1 promoter (Dot1l Δ/Δ ; Col2-Cre). Interestingly, the number of live births of homozygous Dot1l KO mice was substantially lower than expected: Of the 181 animals born from 27 litters, only 8 animals were homozygous Dot1l KO, whereas the expected Mendelian number is 58. Furthermore, the pups needed to be genotyped early between P0-2. Genotyping at conventional age P11-14 revealed only one homozygous pup, which survived to adulthood, but displayed major growth retardation (data not shown). Genotype was verified using skin and tail-tip samples (data not shown). We performed whole-animal Alcian Blue-Alizarin Red staining of the P2 homozygous Dot1l KO mice to investigate phenotypes that could account for the poor survival. Homozygous Dot1l KO pups did not exhibit gross difference in size (Fig. 1A). However, with dissection, several elements showed accelerated ossification compared with control. Most notable accelerated ossification was observed in hyoid, skull base, iliac crest, and lumbar vertebrae elements (Fig. 1B). The hyoid bone showed increased mineralization in the central section as well as incipient calcification of the long horn. At the base of the skull, after the mandible was removed to enhance the view of the cranial base, increased ossification was noticed between the basioccipital and the basisphenoid bone. The iliac crest was smaller, indicative of advanced ossification of the ilium. Similarly, the vertebrae elements of the lumbar spine showed advance mineralization as evidenced by larger Alizarin Red stain at the expense of Alcian Blue stain. In some cases, precocious mineralized vertebrae were observed between the ossified centrum and neural arch.
Given the accelerated ossification, we next evaluated if Dot1l affects hypertrophic chondrocyte differentiation and subsequent endochondral ossification. mRNA in situ hybridization of Col10a1, a marker of hypertrophic chondrocytes, (24) was performed in the femur of P2 pups. Although there was comparable level of Col10a1 expression in the Dot1l-deleted and control growth plates, there was ectopic expression of Col10a1 in epiphyseal chondrocytes (Fig. 1C, arrows). μCT evaluation of knee joint in postnatal Dot1l-deleted mice Because of the difficulty obtaining sufficient number of animals for experimentation, we developed a chondrocyte-specific inducible conditional Dot1l KO mouse line using Acan-CreER system to evaluate the role of Dot1l in postnatal mice. Three-weekold weaning age mice were intraperitoneally injected with tamoxifen. After 4 weeks of tamoxifen treatment, a 50% reduction of Dot1l expression was observed by qPCR ( Supplemental  Fig. S1). Although there was no gross difference in the body size between the control and Dot1l-deleted mice at 15 weeks of age, 12 weeks after Dot1l deletion (data not shown), there was mild decrease in body weight in Dot1l-deleted mice (Supplemental Fig. S2).
For more detailed evaluation, mouse knee joints were harvested and μCT was performed. With Dot1l deletion, there was qualitative and quantitative decrease of trabecular bone in the distal femur and proximal tibia ( Fig. 2A,B). Quantitative evaluation showed statistically significant decreases in BV/TV, Tb.Th, connectivity density, and apparent density of trabecular bone in the Dot1l-deleted mice compared with age-matched controls. There was also increased BS/BV consistent with decreased Tb.
Th. (25) The decreased apparent density may account for the mild decreased body weight in Dot1l-deleted mice without a gross difference in size. Next, we examined the cortical bone at the metaphysis. Notably, there was decreased medullary space and increased cortical thickness of femur and tibia both qualitatively and quantitatively (Fig. 2C,D). These findings suggest that Dot1l loss in chondrocytes led to weakened trabecular bone with likely adaptive response of the cortical bone.
Reduction of extracellular matrix in articular cartilage and growth plate in postnatal Dot1l-deleted mice H&E stain of mouse knee joint growth plate and articular cartilage was performed in 4-week-old mice, 1 week after Dot1l deletion in Dot1l Δ/Δ and Dot1l Δ/Δ ; Acan-CreER mice (Fig. 3A). Growth plate showed decreased length with irregular cell stacking of the hypertrophic zone, shortening of the proliferative zone, and slightly enlarged hypertrophic cells. On the other hand, articular cartilage showed mild decreased articular cartilage thickness without surface fibrillation or cleft. We next evaluated the growth plate and articular cartilage in 15-week-old mice after 12 weeks of Dot1l deletion (Fig. 3B). More marked shortening of the articular cartilage and growth plate were observed, but again no articular cartilage damage was apparent. Quantification of articular cartilage and growth plate lengths noted in Fig. 3C.
To establish whether these changes were in part related to reduction in extracellular matrix deposition, Safranin O staining of the same samples was performed. Reduced Safranin O staining indicating loss of extracellular matrix in articular cartilage and growth plate was observed as early as 1 week after Dot1l deletion (Fig. 3D). Particularly, there was loss of matrix deposition in the articular cartilage superficial zone and growth plate hypertrophic zone. Reduction of matrix deposition in the articular cartilage and growth plate became even more extensive after 12 weeks of Dot1l deletion (Fig. 3E). Quantification of stain intensity noted in Fig. 3F.
Because we observed increased ossification of the lumbar spine with prenatal Dot1l deletion, we also examined L3/4 discs 1 week and 12 weeks after postnatal Dot1l deletion using H&E and Alcian Blue stains. We observed similar hypertrophic zone disruption and decreased extracellular matrix deposition in vertebral discs of Dot1l-deleted mice (Supplemental Fig. S3). Although these findings indicated that Dot1l is essential in maintaining extracellular matrix production and growth plate differentiation, Dot1l loss alone did not result in rapid articular cartilage damage.
Next, we performed Acan and Col2a1 qPCR of 4-week-old Dot1l Δ/Δ and Dot1l Δ/Δ ; Acan-CreER mice femoral head chondrocytes 1 week after Dot1l deletion. There was 62% and 60% reduction in Acan and Col2a1 expression, respectively (Fig. 4A), which correlated with the reduced Safranin O staining. There was also decreased chondroitin sulfate expression, determined by immunohistochemistry with the CS-56 antibody 1 week (Fig. 4B) and 12 weeks (Fig. 4C) after Dot1l deletion, reflecting reduced expression of the ACAN core protein. Quantification of stain intensity noted in Fig. 4D. These results confirm that Dot1l is an important modulator of the proteoglycan component of the extracellular matrix.

Decreased cell proliferation in postnatal
Dot1l-deleted mice Dot1/DOT1L is a crucial regulator of the cell cycle and loss of Dot1/DOT1L leads to impaired cell division. (26)(27)(28) Beacuse shortening of the proliferative zone was observed in growth plates of Dot1l deletion in Dot1l Δ/Δ ; Acan-CreER mice (Fig. 3A,B), we evaluated whether Dot1l deletion led to altered proliferation of chondrocytes. Seven-week-old Dot1l Δ/Δ and Dot1l Δ/Δ ; Acan-CreER mice (4 weeks after Dot1l deletion) were injected with BrdU and tissues were harvested after 48 hours. Detection with a BrdU antibody showed decrease in BrdU-positive cells in Dot1ldeleted growth plate (Fig. 5), corroborating the shortened growth plate observed with H&E staining after Dot1l deletion. Immunohistochemistry with cleaved Caspase 3 was performed on the same samples, but there was no appreciable difference between Dot1l-deleted mice and control mice (data not shown).
Phenotype of Dot1l loss in bone morphology is dependent on nuclear lamina integrity As osteoarthritis and osteoporosis are strongly correlated with aging, we next turned to progeria model Zmpste24-deleted mice. ZMPSTE24 is a prelamin-processing zinc metalloproteinase and deficiency of this enzyme in humans results in premature aging syndrome. (29) Previous studies have shown that in Zmpste24deleted mice, alterations in histone modification affects chromatin organization, (30) and in particular, depletion of H3K9 methylation in these mice can rescue bone mineral density (BMD) loss. (31) Based on these studies, we evaluated the effects of DOT1L inhibitor EPZ-5676 in Zmpste24-deleted mouse forelimbs using μCT. Forelimb samples of WT mice, Zmpste24-deleted mice, and Zmpste24-deleted mice treated with DOT1L inhibitor were gifts from Dr López-Otín. Six-week-old Zmpste24-deleted mice were treated with either vehicle or DOT1L inhibitor for 12 weeks. Age-matched WT mice without treatment were used as reference. In Zmpste24-deleted mice, there was marked decrease in BV/TV and Tb.Th, which was partially rescued in mice treated with the DOT1L inhibitor (Fig. 6). These results correlate with the existing evidence that alterations in histone modification can rescue bone loss in progeria models.

Discussion
Our results showed that Dot1l is important in mouse prenatal and postnatal articular cartilage, growth plate, and trabecular bone. Dot1l deletion in chondrocytes during embryonic development led to embryonic and perinatal mortality, altered ossification, and dysregulation of Col10a1 expression. Postnatal chondrocyte Dot1l deletion led to trabecular bone loss including decreased BV/TV, Tb.Th, connectivity density, and apparent density, and increased BS/BV. This was likely caused by disorganization of the growth plate hypertrophic zone with Dot1l deletion seen in histology. On the other hand, there was decreased medullary space and increased cortical bone thickness, likely compensatory. There was also decreased extracellular matrix production and corresponding decrease in Acan, Col2a1, and chondroitin sulfate expression in articular cartilage and the growth plate as early as 1 week after Dot1l deletion that persisted to 12 weeks. Dot1l deletion also led to decreased growth plate cell division, which resulted in shortening of the proliferative zone.
Previously reported cross of the conditional Dot1l KO line with the Col2-Cre line was more successful in generating mice that survived Dot1l deletion, (17) possibly because of decreased penetration of the deletion than in the present study; however, profound growth retardation at 1 month of age was reported. The previous study did not go into details on the number of pups born, but a different Col2-Cre mouse line in B6SJL/F1 background was used, which possibly contributed to the difference in observed phenotype. In our study, a single Dot1l Δ/Δ ; Col2-Cre mouse that survived to adulthood showed major growth retardation as well (data not shown), suggesting that the acceleration of ossification observed at P2 eventually extended to all elements. One possible explanation of the perinatal mortality is that early ossification of the hyoid bone and skull base had adverse effects on development or exerted mass effect on the trachea and brainstem, respectively. Although increased ectopic bone formation around the knee joint was noted in 16-month-old Dot1l heterozygote mouse, (19) this likely reflects osteophyte formation from accelerated osteoarthritis and is unlikely to affect survival. Further studies are needed to delineate how prenatal Dot1l loss in Col2a1-expressing cells effects survival, but it is important to point out that KO models of extracellular matrix  components such as Acan (32,33) and Col2a1 (34) are also embryonic lethal. Furthermore, accelerated hypertrophy of chondrocytes was reported in the Acan KO mice. (32,33) We also observed ectopic expression of Col10a1 in the P2 epiphysis with prenatal Dot1l deletion. As Col10a1 expression reflects cell cycle exit, this result is congruent with decreased BrdU-positive cells in growth plate seen with postnatal Dot1l deletion, and together suggest that Dot1l is required to maintain chondrocytes in the proliferative state. No previous studies have specifically examined cell cycle in Dot1l-deleted growth plates. However, regulation of cell cycle and differentiation by Dot1l has been reported in other cell types, including hematopoietic cells (20) and cortical neurons. (35) Dot1l loss in postnatal mouse chondrocytes also led to weakened trabecular bone. In particular, there was decrease in total trabecular bone (BV/TV) as well as changes in bone morphology (BS/BV), both of which are correlated with callus strength of healing fracture. (36) In contrast, there was increased cortical bone thickness and decreased medullary space, which may be compensatory. The trabecular bone loss was consistent with a recent study of DOT1L-inhibitor treatment of ovariectomized mice, (18) and increasing DOT1L activity may be therapeutic in human osteoporosis, particularly in postmenopausal women. The mechanism of bone morphologic change after Dot1l deletion in chondrocytes remains unknown. Altered growth plate integrity by reducing matrix deposition and proliferation could affect the bone formation by affecting the osteoblast, osteocyte, and osteoclast genesis. In fact, it has been reported that a proportion of hypertrophic chondrocytes can differentiate into osteoblast and osteocytes during endochondral bone formation and bone repair. (37,38) Involvement of Dot1l in this process remains to be explored. In addition, in human osteoarthritis, cartilage degenerative changes are associated with increased turnover in the adjoining bone, but reduced turnover in the nonadjoining bone, (39) indicating interaction between cartilage integrity and surrounding bone.
Despite significant decrease in the extracellular matrix and thinning of the growth plate, we did not see fibrillation or cleft of the articular cartilage. This is likely related to the sedentary laboratory mice setting and a relatively short observation period (12 weeks of Dot1l deletion). A recent study showed that with an extended observation period (14 and 16 months) and under stress (surgically induced osteoarthritis model), Dot1l loss led to accelerated osteoarthritis in mice. (19) These findings suggest that increasing DOT1L activity may be protective in human osteoarthritis.
We also observed partial rescue of the bone loss phenotype in Zmpste24-deleted mice treated with the DOT1L inhibitor. This finding is consistent with H3K9 methylation depletion in Zmpste24-deleted mouse leading to improved BMD, (31) but it is somewhat contrary to the known histone methylation code. H3K9 methylation is associated with transcription suppression and chromatin silencing, whereas H3K79 methylation is associated with transcription activation and euchromatin. (26,31) The result is also somewhat divergent from our observation of trabecular bone loss with postnatal Dot1l deletion. One can speculate on several explanations for these findings. First, H3K79 residue methylated by DOT1L is located near the central nucleosome, unlike H3K9 residue that is located in the histone tail. (40) This difference may result in H3K79 methylation loss affecting chromatin structure differently in WT mice compared with Zmpste24-deleted mice, which have compromised nuclear membrane. Alternatively, there may be yet undelineated interactions (either direct or indirect) with Dot1l/H3K79 methylation and Zmpste24/LAMIN. Prior studies show that LAMIN and LAMINbinding proteins at the inner nuclear membrane associate with and regulate chromatin conformation, (41) and that Dot1l deletion led to reduced heterochromatic marks in centromeres and telomeres. (42) Given these results, H3K79 methylation and inner nuclear membrane LAMIN may interact to control euchromatin/heterochromatin structure. Another possibility is confounding cross-inhibition of other methyltransferases with DOT1L inhibitor EPZ-5676, but this is less likely given reported 37,000-fold selectivity of the inhibitor. (43) Also, there is technical difference between the use of chondrocyte-specific genetic Dot1l deletion and systemic chemical inhibition of DOT1L. The latter affects not only chondrocytes but all cells, including osteoblast, osteocyte, osteoclast, and synovial cells, which may have contributed to differences in phenotype.
In conclusion, Dot1l is important in prenatal and postnatal chondrocyte development as well as trabecular bone maintenance. Dot1l prevents the premature cell cycle exit of proliferating chondrocytes and supports chondrocyte matrix production. Further studies are necessary to delineate the complex Dot1l regulation of chondrocyte-osteoblast/osteocyte homeostasis and its therapeutic potential in human osteoarthritis and osteoporosis.

Disclosures
Authors have nothing to disclose.