The heat shock proteins (HSP) and the ubiquitin-proteasome system (UPS) contribute to maintain protein homeostasis in eukaryotic cells. They have been recently proposed as therapeutic targets in a wide variety of cancers [31
]. Preclinical studies have demonstrated the antitumour activity of proteasome inhibitors in childhood sarcomas, including rhabdomyosarcoma, Ewing sarcoma (EWS) and osteosarcoma (OS) [7
]. In addition, when used in combination with standard chemotherapeutic agents or targeted therapeutics, proteasome inhibitors can increase anti-tumor activity, suggesting that combination therapy can achieve better results than single-agent treatments [37
]. Because several Hsp90 client proteins are relevant for tumour growth and survival, inhibition of Hsp90 has also been investigated in cancer, including rhabdomyosarcomas, demonstrating that targeting Hsp90 may reduce tumour cell proliferation and migration, both in vivo
and in vitro
]. Targeting the Hsp90 molecular chaperone has shown anti-tumor efficacy in Ewing sarcomas, whereas the efficacy in childhood osteosarcoma is lower and limited to specific treatment modalities [12
]. Similarly to proteasome inhibitors, preclinical and clinical studies in haematological malignancies have demonstrated that Hsp90 inhibitors are effective as anti-cancer drugs especially when combined with conventional chemotherapy or targeted therapeutics [41
]. Based on these observations, we investigated the antitumour activity of Bortezomib and the Hsp90 inhibitor 17-DMAG in rhabdomyosarcoma cells, comparing single-agent exposures with combination treatments. Consistent with previous findings [7
], both Bortezomib and 17-DMAG were highly efficient at inhibiting growth and survival of RMS cells. Embryonal RD cells were more sensitive than alveolar RH30 cells to single-agent exposure, although such a different sensitivity was not justified by the expression of specific target proteins, like heat shock proteins, cell cycle inhibitors and pro-apoptotic factors [45
]. Conversely, effectiveness of the combined treatment was comparable among cell lines, including RH30 cells that responded poorly to Bortezomib alone. Combination of Bortezomib and 17-DMAG induced autophagy in addition to apoptosis, and this was confirmed by the concurrent cleavage of LC3-II and PARP proteins.
Consistent with these findings, inhibition of proteasome activity causes the accumulation of misfolded proteins inside cells, which bind to heat shock proteins in discrete structures known as aggresomes and are subsequently degraded by lysosomes [47
]. Attachment of ubiquitin to proteins, in fact, not only constitutes a degradation signal for the proteasome, but also serves for removal of proteins by lysosome-mediated autophagy [48
]. In these conditions, unfolded proteins are delivered to lysosomes by heat shock proteins, in what is known as chaperone-mediated autophagy (CMA) [23
]. Macroautophagy, the most important lysosome-mediated type of autophagy, does not usually need shuttling substrate proteins by heat shock proteins, but recent evidences suggest that in some cases heat shock proteins translocate ubiquitinated proteins into lysosomes and assist their autophagic degradation as well [48
]. Collectively, these findings indicate that autophagy may be activated by proteasome inhibitors, likely as a response mechanism that alleviates from stress and protects cells from apoptosis [19
]. Consistent with this scenario we demonstrated that autophagy is activated in rhabdomyosarcoma cells to withstand drug-induced cytotoxicity, as suggested by LC3-I activation and intensive cytoplasmic vacuolization. As expected, the inhibition of autophagy by chloroquine increases caspase-dependent PARP cleavage in rhabdomyosarcoma stressed cells, whereas its induction by rapamycin partially rescued cells from drug-induced apoptosis. In particular, inhibition of apoptosis occurs when cells are pretreated with rapamycin prior to administering Bortezomib and/or 17-DMAG, whereas the inhibition of autophagy increases cell death when induced together with both proteasome and Hsp90 inhibition. This suggests that RMS cells may activate autophagy as cytoprotective response to drug treatment, and the inhibition of autophagy enhances sensitivity of RMS cells to anti-cancer drugs, including Hsp90 and proteasome inhibitors. Of note, inhibition of autophagy is more effective at early onset of stress response than following apoptosis induction, providing evidence that autophagy occurs before cell death and it functions primarily as a cell survival mechanism.
The ubiquitin-proteasome and autophagy-lysosome are often considered distinct degradation systems. However, recent studies suggest that these two pathways are mechanistically linked [53
], as proteasome inhibition induces autophagy when removal of toxic polyubiquitinated aggregates is necessary for cell survival, while proteasome activity is induced when formation and activity of lysosomes are impaired. Autophagy and apoptosis are events regulated by common survival pathways, including the JNK1, Bcl-2 and the PI3K/AKT signaling pathway. It has been shown that JNK-dependent phosphorylation of Bcl-2 promotes cell survival by disrupting Bcl-2 binding to Beclin-1 and activates autophagy, whereas sustained Bcl-2 phosphorylation blocks Bcl-2 anti-apoptotic activity and apoptosis overwhelms autophagy [57
]. AKT phosphorylates and prevents Bad pro-apoptotic activity and inhibits autophagy by impairing TSC1TSC2 tumor suppressor proteins activity [59
]. Of note, JNK and AKT are, among survival proteins, mostly affected in rhabdomyosarcoma cells treated with proteasome and Hsp90 inhibitors, but their involvement in drug-induced autophagy have not been investigated yet [7
Nevertheless, our findings suggest that combination treatment with Bortezomib and 17-DMAG can overcome autophagy, a mechanism protecting rhabdomyosarcoma cells from drug-induced cytotoxicity. Further studies are warranted on the use of low concentrations of proteasome inhibitors in combination with Hsp90 inhibitors, both in vitro and in vivo, as they might represent a tool capable of counteracting protective mechanisms, such as autophagy, that may affect treatment efficacy and, ultimately, the outcome of RMS patients.