While the major source of vitamin D is D₃ from ultraviolet exposure, some supplements supply D₂. The relative potency of vitamin D₂ versus vitamin D₃ remains controversial. The aims of the current study were, 1. To optimize the in vitro model, including use of cell lines, vitamin D concentrations, and outcome biomarkers. 2. To compare the potency of vitamin D₂ and D₃ metabolites on mouse and human bone cellular activity. 3. To explore the expression of VDR in osteoarthritic (OA) bone tissues as well as cellular responses to vitamin D₂ and D₃ metabolites ex-vivo. In mouse 2T3 osteoblasts, at physiological doses, both vitamin D₂ and D₃ metabolites increased ALP activity and mineralisation and up-regulated osteoblastic signature genes and proteins. At supra-physiological doses D₃ metabolites were more potent inhibitors of 2T3 function than D₂ metabolites. Although hBMS cell proliferation was inhibited by both 25(OH)D₂ and D₃, ALP activity was enhanced by both metabolites. However, 25(OH)D₃ was a more potent stimulator of ALP and mineralisation of hBMSCs. D₂ and D₃ equally stimulated expression of CX43 and PHEX markers in osteocytic cell lines. Immunohistochemistry of femoral heads showed much reduced VDR expression in OA osteocytes and osteoclasts, yet both 25(OH)D2 and D₃ increased OA-hBMSCs mineralisation more than non-OA-hBMSCs ex-vivo. While vitamin D₂ or D₃ increased mouse 2T3 osteoblastic activity at physiological doses, OA and non-OA hBMSCs differentiation was more responsive to 25(OH)D₃. Key bone cells such as osteocyte and osteoclasts expressed less VDR in OA. For the first time vitamin D₂ metabolites have been thoroughly examined and emerged as a potent stimulator of bone cell differentiation, at least in vitro. Vitamin D₃ in contrast is confirmed as highly potent in bone cells, but with toxicity at much lower doses than D₂.

Fibroblast growth factor 2 (FGF-2) is known to be released from cartilage upon injury and is able to influence chondrocyte gene expression in vitro. In cartilage, FGF-2 regulates E11/podoplanin expression in murine joints following surgical destabilisation (DMM model of osteoarthritis (OA)), and in cartilage explant injury models. In bone, E11 is critical for the early stages of osteocytogenesis and is responsible for the acquisition of the osteocyte dendritic phenotype. This dendritic phenotype is dysregulated in OA and given the known role of the osteocyte in controlling bone remodelling, this may contribute to the subchondral bone thickening observed in OA. Hence, the aim of this study was to elucidate the nature of FGF-2- mediated E11 expression and osteocytogenesis in skeletal health and disease. This thesis has shown that FGF-2 dose-dependently increased E11 mRNA expression in MC3T3 cells, primary osteoblasts and in primary calvaria organ cultures, which was confirmed by E11 protein western blotting data. The FGF-2 induced changes in E11 expression were accompanied by significant increases in the mRNA expression of the osteocyte markers Phex and Dmp1, and significant decreases in the mRNA expression of the osteoblast markers Col1a1, Postn, Bglap and Alpl expression. This thus supports the hypothesis that FGF-2 drives osteocytogenesis. The acquisition of osteocyte phenotype involves the re-organisation of the cytoskeleton, such as F-actin. This step is important for the transition of cuboidal-shaped osteoblasts to the stellate-shaped osteocyte phenotype. FGF-2 stimulation of MC3T3 cells and primary osteoblasts revealed more numerous and longer dendrites, as visualised by phalloidin staining for F-actin and indicative of the acquisition of the osteocyte phenotype. In contrast, control cells had a typical rounded morphology with fewer and shorter dendrites. Furthermore, immunofluorescence labelling for E11 in control cells revealed uniform distribution throughout the cytoplasm, especially in the perinuclear region. In contrast, FGF-2 treated cells showed a modified distribution where E11 was negligible in the cytoplasm, but concentrated in the dendrites. The use of siRNA knockdown of E11 achieved a 70% reduction of basal E11 mRNA expression. This knockdown also effectively abrogated FGF-2-related changes in E11 expression and dendrite formation as disclosed by mRNA and protein expression, immunofluorescence and F-actin staining with phalloidin. Despite these FGF-2 driven increases in E11 and osteocyte dendrite formation in vitro, immunohistochemical labelling revealed no differences in E11 expression in subchondral, trabecular and cortical osteocytes from naïve Fgf-2 deficient mice in comparison to wild-type mice. Similar results were observed upon sclerostin immunolabelling. FGF-2 stimulation of MC3T3 cells elicited activation of ERK1/2, Akt and p38 MAPK. However, inhibition of the aforementioned pathways failed to reduce FGF-2- mediated E11 expression and as such, the specific signalling pathway responsible remains unclear. Upstream, the expression of Fgfr1 was increased (>10-fold) over 24 h time point, while a reduction was seen in Fgfr2/3 expression over same time point especially in the FGF-2 treated cultures. This suggests that increased E11 expression and the acquisition of the osteocyte phenotype may be speculatively though upregulation of Fgfr1. The expression of E11 and sclerostin in OA pathology in mice, human and dogs were investigated. Initially sequence homology using the Clustal Omega alignment program showed both proteins to be homologous in the domestic animals under study. A comparative study using canine subchondral bone osteocytes revealed increased E11 expression in the OA samples relative to the control. This feature may be related to newly embedded osteocytes during sclerosis. However, E11 and sclerostin were unchanged in both murine (DMM) and human OA subchondral bone osteocytes in comparison to controls. In mice, this may be due to limited OA development; whilst in humans the sample size, age, stage of the disease and sourcing from same diseased joint may be important in the interpretation of the results. The expression of E11 and sclerostin during OA pathology was also investigated in Fgf-2 deficient mice in which OA was induced using the DMM model. There was no difference in E11 expression between the OA and control (sham-operated) samples, suggesting that compensation of E11 expression may be mediated by growth factors from the FGF family. Surprisingly, increased E11 expression was observed in the control Fgf-2 deficient mice, in comparison to the wild-type control mice. This suggests a potential adjustment to loading by the contralateral knee, as this was not observed in naïve mice from both groups. Together, these data show that FGF-2 promotes the osteocyte phenotype, and that this is mediated by increased E11 expression. These results may help explain (1) the altered osteocyte phenotype and (2) increased subchondral bone thickening observed in OA. This knowledge will be of interest in the search for disease modifying therapeutics for skeletal health, including OA and osteoporosis.

Vascular calcification is prevalent in ageing, in atherosclerosis, and especially in patients with Chronic Kidney Disease (CKD), with associated increased morbidity and mortality. The phenotypic transition of Vascular Smooth Muscle Cells (VSMCs) into osteoblastic/chondrogenic-like cells is critical for the development of calcification in CKD patients. Osteocytes, terminally differentiated osteoblasts, have recently emerged as major regulators of calcification in bone. Recently, osteocytelike cells have been observed in human peripheral arteries with medial vascular calcification. However, it remains undetermined as to whether VSMCs can undergo osteocytic differentiation within a calcifying environment and the functional role of osteocyte formation in the development of medial vascular calcification. Initial studies have characterised the ectonucleotide pyrophosphatase/phosphodiesterase 1 knockout (Enpp1-/-) mouse as a valid model of medial vascular calcification, which is employed throughout this thesis. This thesis has compared VSMCs to osteoblasts undergoing osteocytic differentiation in vitro. VSMC in vitro calcification was accompanied by up-regulated expression of osteocyte markers, including Sost, E11, Dmp1, Phex, Mepe and Fgf23. Immunohistochemistry confirmed the appearance of sclerostin and E11 in calcified aortae from the Enpp1-/- mouse. Further studies have identified a direct inhibitory role for the osteocyte specific gene FGF23 in modulating vascular calcification. The inhibitory effect of FGF23 on VSMC calcification was mediated through the MAPK/ERK signalling pathway. This thesis has also determined the role of BMP9, a highly osteogenic bone morphogenic protein, in vascular calcification, which induces VSMC calcification through a Smad signalling mechanism. Furthermore, VSMC expression of the osteocytic marker Sost was markedly increased following BMP9 treatment. Intriguingly, BMP9 was markedly elevated in serum from dialysis patients and a significant correlation was observed between dialysis time and BMP9 concentration in patients receiving haemodialysis. The work described herein has demonstrated that vascular calcification is associated with an osteocyte phenotype, and reports a direct inhibitory effect of the osteocyte specific gene FGF23 on vascular calcification. Furthermore, this thesis has shown that BMP9 induces the expression of the osteocytic marker Sost in VSMCs, and appears to play a critical role in vascular calcification.

Bone changes composition and structure to adapt to its mechanical environment. Osteocytes are putative mechanosensors responsible for orchestrating the bone remodeling process. Recent in vitro studies showed that osteocytes could sense and respond to substrate strain and fluid shear. However the capacity of osteocytes to sense cyclic hydraulic pressure (CHP) associated with physiological mechanical loading is not well understood. In this study, osteocyte-like MLO-Y4 cells were subjected to CHP of 68 kPa at 0.5 Hz, and the effects of CHP on intracellular calcium concentration, cytoskeleton organization, mRNA expression of genes related to bone remodeling, and osteocyte apoptosis were investigated. The results indicate that osteocytes could sense CHP and respond by increased intracellular calcium concentration, altered microtubule organization, an increase in COX-2 mRNA level and RANKL/OPG mRNA ratio, and decreased apoptosis. Therefore cyclic hydraulic pressure in bone a mechanical stimulus to osteocytes and may play a role in regulating bone remodeling.

Bone changes composition and structure to adapt to its mechanical environment. Osteocytes are putative mechanosensors responsible for orchestrating the bone remodeling process. Recent in vitro studies showed that osteocytes could sense and respond to substrate strain and fluid shear. However the capacity of osteocytes to sense cyclic hydraulic pressure (CHP) associated with physiological mechanical loading is not well understood. In this study, osteocyte-like MLO-Y4 cells were subjected to CHP of 68 kPa at 0.5 Hz, and the effects of CHP on intracellular calcium concentration, cytoskeleton organization, mRNA expression of genes related to bone remodeling, and osteocyte apoptosis were investigated. The results indicate that osteocytes could sense CHP and respond by increased intracellular calcium concentration, altered microtubule organization, an increase in COX-2 mRNA level and RANKL/OPG mRNA ratio, and decreased apoptosis. Therefore cyclic hydraulic pressure in bone a mechanical stimulus to osteocytes and may play a role in regulating bone remodeling.

Indiana University-Purdue University Indianapolis (IUPUI) / Old age is associated with increased bone fragility and risk of fracture as a result of skeletal alterations, including low bone density and cortical thinning. Further, apoptotic osteocytes accumulate in old mice and humans. We have previously shown that mice lacking osteocytic connexin (Cx) 43 (Cx43ΔOt) exhibit a phenotype similar to that of the aging skeleton, with elevated osteocyte apoptosis and an associated increase in osteoclastogenesis. These findings suggest that osteocyte apoptosis results in the release of factors that recruit osteoclasts to bone surfaces close to areas that contain apoptotic osteocytes. However, the specific chemotactic signals, the events mediating their release, and the mechanisms of their action remain unknown. Consistent with this notion, we also found that HMGB1 released by Cx43-deficient (Cx43def) MLO-Y4 osteocytic cells enhances osteoclastogenesis in part by increasing osteocytic RANKL, which promotes osteoclastogenesis, and, at the same time, directly stimulating osteoclastogenesis. Further, expression of the pro-survival microRNA (miR), miR21, is low in Cx43def cells and bones from old female mice, and low miR21 levels increase osteocyte apoptosis. However, surprisingly, mice lacking miR21 (miR21ΔOt) have decreased osteoclast number and activity even under conditions of elevated osteocyte apoptosis; suggesting that osteocytic miR21 may mediate osteoclast precursor recruitment/survival induced by apoptotic osteocytes. However, whether HMGB1/miR21 are released by osteocytes, and if the HMGB1 receptors, receptor for advanced glycation end products (RAGE) and/or tolllike receptor (TLR4) are involved in osteoclast recruitment in Cx43ΔOt and old mice is unknown. The overall objectives of this series of studies were to elucidate the mechanisms

Previous research has provided evidence in support of apoptotic osteocytes playing a role in the bone remodeling process. In this study, we examined the regional and quadrantal variations of apoptotic and viable osteocytes in cyclically loaded and unloaded samples. Left tibias of C57 Black 6 Taconic mice (C57Bl/6) were cyclically loaded for either 2 weeks or 5 weeks, with the right tibias being used as controls. After loading, tibias were resected, processed, and then stained using either a TUNEL stain, to show apoptotic osteocytes, or a 2.0% methyl green solution, to reveal viable cells. Cross-sectional images from each tibia were then captured and analyzed in each region (distal, midshaft and proximal) and quadrant (cranial, lateral, caudal, and medial) by counting the number of osteocytes, both apoptotic and viable, and subsequently calculating the percentages and densities of those osteocytes. Individual analysis of each sample group showed that the 5 week loaded bones, with the most statistically significant p-values, had the most regional variations within the samples, specifically showing decreased apoptotic and viable osteocytes in the lateral quadrants. Comparative analysis revealed a statistically significant higher percentage and density of apoptotic osteocytes in 5 week loaded samples compared to all other samples. This provides further quantitative evidence in support of apoptotic osteocytes playing a role in bone remodeling.

Osteocytes (resident bone cells) are believed to sense loading-induced interstitial fluid flow in bone and transduce the signals to osteoclasts (bone resorption cells) and osteoblasts (bone formation cells) to regulate bone remodeling. Recent studies have shown that bone disuse causes osteocyte apoptosis, which precedes osteoclast activity at the local remodeling site. Although osteoclast precursors are known to travel via the circulation, the specific mechanism by which they are transported to the remodeling site is unclear. We hypothesized that lack of fluid flow induces osteocyte apoptosis. Furthermore, we hypothesized that osteocyte populations containing apoptotic osteocytes secrete cytokines that: 1) promote angiogenesis, and 2) activate the endothelium to promote osteoclast precursor adhesion to the endothelium such that osteoclast precursors can be delivered closer and directly to the remodeling site. In our in vitro studies, we found that lack of oscillatory fluid flow (mimicking mechanical disuse) promotes osteocyte apoptosis. In addition, osteocyte populations containing apoptotic cells promote endothelial cell proliferation, migration, and tubule formation. Inhibition of the potent angiogenic cytokine, vascular endothelial growth factor (VEGF), abrogated osteocyte apoptosis-mediated angiogenesis. Furthermore, we found that osteocyte populations containing apoptotic cells secrete cytokines that promoted osteoclast precursor adhesion. Upon further investigation, we found that apoptotic osteocytes secreted elevated levels of inflammatory cytokine interleukin 6 (IL-6), and its soluble receptor, sIL-6R. We demonstrated that both IL-6 and sIL-6R are required to activate the endothelium to express ICAM-1. Inhibition of ICAM-1 and IL-6 by blocking antibodies abolished apoptotic osteocyte-mediated osteoclast precursor adhesion. Our findings suggest for the first time that osteocytes communicate to endothelial cells directly to mediate angiogenesis and osteoclast precursor adhesion. Results from this study may assist in a better understanding of osteoclast precursor recruitment at the initial onset of bone resorption.

Osteocytes, the most abundant cell within bone, have been linked to the processes of mechanosensation and transduction. Based upon relatively limited empirical evidence, variations in their abundance and morphology have been linked to sex, age, biomechanics and disease. In order to better elucidate lacunar variation within a healthy cohort, samples from 30 women aged 20-86 and 36 men aged 18-92 were studied utilizing synchrotron radiation micro-CT. Initial studies of normal variation within the femoral proximal shaft cross-section found high variation in lacunar density (up to ~54%) and associated morphological differences linked to biomechanical regions. In women, a non-significant trend in lacunar density reduction was apparent with age; however, a significant reduction in lacunar volume with age (~30%) was observed. Also noted were differences in lacunar morphology, with the lacunae of younger women characterized as flatter and less equant than their older counterparts. The males, who demonstrated lacunar density decline with age and a tendency towards more equant and less elongate lacunae, did not share these characteristics. Intriguingly, the previously noted reductions in lacunar volume were not observed in males. The results of this research indicate that normal variation in osteocyte lacunar parameters is high. To our knowledge the observation that lacunar volume differs in women with age is novel, potentially resulting from preferential surface infilling within the extracellular space. The functional impact of this infilling is unclear but such a change in scale likely impacts the mechanosensing function of the osteocyte network. This hypothesis warrants further investigation as, if confirmed, it would represent a profound negative impact on the osteocyte network and may provide new insights into age-related bone loss.

Osteocytes are believed to be the mechanosensory cells that detect and respond to mechanical loading. Physiological loading by oscillatory fluid flow (OFF) activates osteocytes to increase intracellular calcium concentration and release prostaglandin E2 (PGE2). Osteocytes are also sensitive to chemical stimulations such as serotonin, which can also increase PGE2 release. However, it is unclear whether mechanical stimulation can influence osteocyte sensitivity towards serotonin. In this thesis, MLO-Y4 osteocyte-like cells were subjected to serotonin with or without precondition by OFF, the responses of intracellular calcium and PGE2 release were measured. Serotonin increased intracellular calcium and PGE2 release in osteocytes. The effects were significantly reduced by OFF precondition, suggesting mechanical precondition by OFF can reduce osteocyte sensitivity towards serotonin. In terms of mechanisms, OFF- and serotonin-induced calcium responses depended on intra- and extracellular calcium stores. ATP was found to partially mediate OFF modulation of serotonin-induced PGE2 release but not calcium.