ATP Release and Purinergic Mechanotransduction in Bone Cells
Open Access
- Author:
- Genetos, Damian C.
- Graduate Program:
- Physiology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 18, 2005
- Committee Members:
- Henry Joseph Donahue, Committee Chair/Co-Chair
Randall L Duncan, Committee Member
Blaise Peterson, Committee Member
David A Antonetti, Committee Member
Shao Cong Sun, Committee Member
Neil Sharkey, Committee Member - Keywords:
- bone
osteoblast
osteocyte
ATP
hemichannel
prostaglandin
mechanical loading
fluid flow - Abstract:
- Osteoporosis is a disease of the skeleton that is characterized by fragile bones, which increases the risk of skeletal fracture. According to the latest report from the Surgeon General of the United States, 10 million Americans over the age of 50 have osteoporosis, and 34 million are at risk [1]. Osteoporosis causes 1.5 million skeletal fractures per year at an annual cost of $12.2–17.9 billion per year [2]. As the population continues to age [3], this annual expenditure will surely rise as well. Consequently, osteoporosis prevention is at the forefront of orthopaedic research. One method to prevent osteoporosis is to increase the formation of new bone while preventing the resorption of older bone. A wealth of reports have demonstrated that mechanical loads can regulate skeletal architecture by increasing the activity of bone-forming osteoblasts and inhibiting the activity of bone-resorbing osteoclasts. This would strengthen the skeleton and thereby reduce the risk of osteoporosis. What continues to perplex orthopaedic investigators are the cellular mechanisms whereby external mechanical forces are perceived by cells of the osteoblastic lineage and transduced into an anabolic response (i.e., formation of new bone by osteoblasts). Movement of pericellular fluid, or fluid flow, induced by mechanical loads appears to be the most likely localized signal perceived by osteoblasts and osteocytes. The application of fluid flow to these cell lineages induces a rapid, yet transient increase in cytosolic calcium (Ca2+i) that requires both calcium entry into the cell and calcium release from cytoplasmic stores [4]; changes in Ca2+i are implicated in the regulation of a multitude of cellular responses ranging from acute (such as kinase activation [5]) to trophic (i.e., proliferation [6-8]). One signaling molecule that is able to induce both calcium entry and calcium release in osteogenic cells is adenosine triphosphate (ATP) which can activate both ionotropic P2X and metabotropic P2Y receptors [9]. Indeed, the importance of P2 receptors in the calcium response of osteoblasts to fluid flow was recently demonstrated by You et al. [10]. However, the mechanism(s) whereby cytosolic ATP is released from an osteoblast or osteocyte in response to fluid flow has not been described. The overall aims of this thesis were to address how ATP is released from osteoblasts and osteocytes and to further elucidate the importance of P2 receptor activation by ATP in bone cell mechanotransduction. We first examined whether osteoblasts, as the cells directly responsible for the formation of new bone, release ATP in response to fluid flow. Using a steady, laminar flow system with a flow rate that induced a shear rate of 12 dynes/cm2, we found that conditioned media from osteoblasts exposed to fluid flow for 5 minutes contained approximately 10-fold more ATP than did conditioned media from static osteoblasts (59.8+15.7 vs 6.2+1.8 nM). We next used a Harvard pump one-pass system, which perfused fresh media across the osteoblasts, to examine the time course of ATP release in response to fluid flow. We found that fluid flow induced a rapid release of ATP within one minute of the onset of flow that returned to pre-flow levels with prolonged fluid flow. That ionomycin, a calcium ionophore, increased ATP release in static osteoblasts not exposed to fluid flow, suggested that changes in Ca2+i regulated ATP release. Because the L-type VSCC (Cav1.2) and the MSCC have been implicated in the Ca2+i response to mechanical load, we hypothesized that these channels may be involved in flow-induced ATP release. We found that inhibition of the L-type VSCC with nifedipine or verapamil significantly inhibited flow-induced ATP release, whereas inhibition of the MSCC had no effect on ATP release. Inhibition of gap junctional intercellular communication (GJIC) or hemichannel activation with either 18-glycyrrhetinic acid or 18-glycyrrhetinic acid failed to inhibit flow-induced ATP release, suggesting that neither GJIC nor hemichannels are involved in osteoblastic ATP release. Immunolocalization of cytosolic ATP revealed punctate, granular stores of ATP, possibly within cytosolic vesicles. As exocytosis of such ATP-filled vesicles has been implicated in endothelial cells exposed to fluid flow, we examined the effect of pharmacologic antagonists of vesicle formation, release, and fusion on flow-induced ATP release. Inhibition of each of these exocytotic steps significantly attenuated flow-induced ATP release. Taken as a whole, these data suggest that ATP is localized within vesicles in osteoblasts, and that fluid flow promotes vesicle exocytosis through a mechanism requiring calcium entry through the Cav1.2 calcium channel. Wherein load-induced osteogenesis requires prostaglandin synthesis, we demonstrated that activation of P2 purinoceptors by ATP mediates prostaglandin E2 release. Having reported that fluid flow induces ATP secretion that regulates PGE2 release, we sought to refine the signaling between P2 receptor activation and PGE2 release. Flow-induced PGE2 synthesis has been shown to be due solely to cyclooxygenase-2 (COX-2), which is itself induced in response to flow. Further, we have previously demonstrated that maximal COX-2 induction in response to flow requires the activation and translocation of the transcription factor NF-B. As such, we sought whether the effect of purinoceptor activation on PGE2 release involved NF-B activation and translocation. Wherein fluid flow for 1 hour at a shear stress of 12 dynes/cm2 induced robust nuclear staining for the p65 subunit of NF-B, osteoblasts flowed in the presence of an ATPase, apyrase, demonstrated cytosolic NF-B localization. Similarly, flow-induced degradation of IB was impaired in the presence of apyrase, suggesting that purinoceptor activation occurs proximal to IB phosphorylation. Intriguingly, pharmacologic inhibition of the P2X7R also demonstrated impaired IB degradation, implicating this purinoceptor in flow-induced NF-B translocation. Although many investigators examine the effect of fluid flow on osteoblast mechanotransduction, there is disagreement whether osteoblasts in vivo experience magnitudes of fluid flow used in vitro. As such, we chose to confirm our results using osteocytes which, because of their localization in lacunae and canaliculi in vivo, are speculated to be the primary mechanosensor in bone. As with osteoblasts, the conditioned media from osteocytes exposed to fluid flow contained significantly higher levels of ATP than did the conditioned media from static osteocytes. In contrast to osteoblasts, however, osteocytic ATP release was inhibited in the presence of 18-glycyrrhetinic acid, suggesting that either GJIC or connexon hemichannels were responsible for this effect. Because of the culture conditions used, GJIC was physically impaired and these results instead implicate connexon hemichannels. Using the uptake of fluorescent dyes as a marker for hemichannel activation, we reported that fluid flow increased hemichannel activity in MLO-Y4 osteocytes, but not in MC3T3-E1 osteoblasts. Pharmacologic inhibition of protein kinase C inhibited both dye uptake and ATP release in response to flow, strengthening the hypothesis that flow-induced ATP release occurs via connexon hemichannels. The addition of siRNA directed against connexin43 similarly impaired both dye uptake and ATP release, suggesting that hemichannels formed by Cx43 are the primary mechanism for flow-induced ATP release in MLO-Y4 osteocytes. Given the surprisingly different mechanism for ATP release between osteoblasts and osteocytes, we speculated whether purinoceptor activation differentially mediated PGE2 release among the two cell types. As with osteoblasts, the addition of ATP to static osteocytes significantly increased PGE2 release, suggesting that, despite being released by two different mechanisms, ATP activates P2 purinoceptors in both osteoblasts and osteocytes to increase PGE2 release. In summary, these data suggest that mechanical load can directly regulate skeletal architecture by increasing the release of a local factor (ATP) that directly affects osteoblasts and osteoclasts, and indirectly through a mechanism involving increased prostaglandin synthesis and release.