The potentiating effects of cell stress upon osteoclast differentiation

Osteoclast formation is a highly regulated multi-step process, which involves the differentiation of haemopoietic progenitor cells into mature, active osteoclasts. This involves complex interactions between the progenitor haematopoietic cells and osteoblast lineage cells that produce RANKL, the key osteoclast differentiation factor. RANKL binding to its receptor RANK on progenitor cells activates signalling cascades, causing activation of key transcription factors and signalling molecules including NFκB, NFATc1, c-FOS, MITF and p38, which are required for osteoclast differentiation. Dysregulation of these molecules can cause an increase in osteoclast formation, a symptom present in many pathological bone conditions. Stressed cells i.e. under pathological conditions, increase the transcription of molecular chaperone, HSP90. Previous studies by our group found the anti-cancer N-terminal HSP90 inhibitor, 17-AAG, increased osteoclast numbers and tumour growth in bone in vivo of MDA-MB-231 cancer cells. The 17-AAG-mediated increase in osteoclast numbers was also reproduced in the absence of tumour load. The increase in osteoclast numbers was not attributed to osteoblast activity i.e., the production of RANKL, but rather a direct increase of osteoclast differentiation in response to the HSP90 inhibitor. However, why 17-AAG affects osteoclasts in this way and whether other anti-cancer agents act similarly was unknown. In this thesis, 17-AAG and the more recently developed N-terminal HSP90 inhibitors CCT018159 and NVP-AUY922 were shown to increase osteoclast formation. This study identified that 17-AAG did not affect the transcriptional activity nor protein levels of transcription factors NFκB, NFATc1, and c-FOS. However, both 17-AAG and NVP-AUY922 increased levels of the transcription factor MITF, not often studied since its activation is downstream of the other factors. These N-terminal HSP90 inhibitors not only inhibit HSP90 but also indirectly cause HSF-1 mediated cell stress response, suggesting cell stress may affect osteoclast differentiation. The 17-AAG- and NVP-AUY922-mediated increase in osteoclast formation was found to be dependent on HSF1-mediated cell stress. These N-terminal inhibitors also activated another stress-associated factor, p38, which is essential for osteoclast differentiation and phosphorylates MITF. These findings suggested other cell stressors might increase MITF levels and osteoclastogenesis. Chemotherapeutics, doxorubicin, MG132, cisplatin and bortezomib do not inhibit HSP90 but act to cause cancer cell death through various cytotoxic actions including proteasome inhibition and DNA damage. These compounds induced an HSF1-mediated cell stress response, and increased NFATc1 and MITF protein expression. These therapeutics also increased osteoclast formation in a HSF-1 dependent manner with pharmacological inhibition of HSF-1 abolishing the increase. These data indicate that cell stress inducing chemotherapeutic drugs act on osteoclast formation at least partly in a HSF-1 dependent manner and show parallels with HSP90 inhibitors although NFATc1 was also activated by these compounds. Likewise ethanol, which induces a classic oxidative stress also acted in the same manner. The novel findings in this thesis suggest stress-inducing compounds including clinically used cancer therapeutics may have bone-damaging effects by increasing osteoclastogenesis in a HSF-1 dependent manner. Furthermore, this work suggests a more general effect of cell stress in regulating osteoclast differentiation and indicates that any therapeutic that induces cell stress in bone may increase osteoclastogenesis and bone loss.