Autophagy and UPS coordinately sustain cellular homeostasis through protein quality control and clearance. A wide range of neurodegenerative diseases (often described as proteinopathies) are characterized by accumulation of aggregate-prone proteins that are not efficiently removed by proteasome. Compensatory autophagy may alleviate disease pathology in these cases, as numerous studies showed that substrates of autophagy and UPS overlap [1, 4, 5]. Indeed, genetic or pharmacological activation of autophagy facilitates aggregate clearance and neuronal function in disease models [1, 5]. Chronic block of proteasomal degradation also leads to toxicity, ER stress and cell death in various mammalian and Drosophila cells, and proteasome inhibitors are in clinical use to trigger apoptosis of cancer cells in multiple myeloma patients [31–34]. While we found no evidence of cell death in larval Drosophila fat body upon silencing of genes encoding essential subunits of the proteasome or genetic activation of hypoxia signaling, these genetic manipulations strongly enhanced both basal and starvation-induced autophagy in these cells, and this effect was dependent on canonical autophagy genes Atg1, Vps34, Atg9, Atg4 and Atg12.
An important aspect of our study is that overexpressed tagged Atg8 is not ideal for analyzing autophagy in proteasome RNAi cells, as excess p62-containing aggregates incorporate these reporter molecules, likely due to the physical interaction of p62 and Atg8. This artefact may then lead to false conclusions regarding autophagic activity, similar to the case of co-overexpressed p62 and Atg8 reporters .
How does proteasome impairment lead to compensatory autophagy? Multiple scenarios are possible, and these are not mutually exclusive. One hypothesis would be that accumulation of excess cargo triggers the formation of autophagosomes. However, overexpression of specific autophagy substrates p62 or blue cheese (the fly homolog of human Alfy that is also involved in the clearance of ubiquitinated proteins) did not result in an obvious increase in autophagy in Drosophila, unlike proteasome impairment [21, 35]. Moreover, p62 was found to be dispensable for both viability and starvation-induced autophagy in knockout mice . Multiple selective autophagy receptors have been identified in mammalian cells, but these are relatively uncharacterized in Drosophila, as only HDAC6 and blue cheese have been analyzed so far besides p62 [13, 35, 36]. Further studies are necessary to find out whether any of the selective adaptors is involved in autophagy induced by proteasome inactivation, or by starvation. After all, autophagy is thought to be responsible for breakdown and recycling of bulk cytoplasm upon nutrient limitation , so it is possible that none of the receptors are actually required for these large-scale autolysosomal degradation processes.
A second hypothesis is that proteasomal degradation is essential to maintain free intracellular amino acid levels and compensatory autophagy is induced by amino acid shortage, as suggested in a recent publication . In this paper, addition of exogenous amino acids was reported to attenuate proteasome inactivation-induced autophagy and partially rescue lifespan reduction in flies treated with a proteasome inhibitor . In this scenario, loss of autophagy in cells with impaired proteasome activity would further reduce amino acid levels as both recycling pathways are inhibited, presumably resulting in a stronger block of cell growth. However, blocking the function of core autophagy genes did not further decrease the growth of proteasome RNAi cells in fat body clones, indicating that reduced amino acid levels are likely not the only link between these processes in this setting.
A third hypothesis is that stabilization of certain proteasome substrate(s) positively regulates autophagy. The levels of HIF-1α are normally kept low through ubiquitination mediated by VHL, which is followed by proteasomal degradation. Hypoxia leads to stabilization of HIF-1α, resulting in its translocation to the nucleus and upregulation of hypoxia-inducible genes that mediate survival at low oxygen concentrations . Cytoprotective autophagy is also induced by hypoxia, and hypoxia-induced autophagy is necessary for survival of cancer cells in the central, poorly vascularized regions of solid tumors . We found that genetic activation of hypoxia signaling induced autophagy in Drosophila, and silencing of the fly HIF-1α ortholog sima attenuated proteasome impairment-induced autophagy. These results suggest that autophagy is enhanced by proteasome inhibition at least in part through hypoxia signaling. A recent genome-wide association study identified mutations affecting the UPS pathway in renal cell carcinoma, which were associated with increased levels of HIF-1α in these tumors . Interestingly, cells derived from many established cancers show elevated levels of autophagy, and multiple oncogenes and tumor suppressors (such as VHL) encode enzymes of the UPS involved in ubiquitin conjugation or deconjugation . Thus, based on these and our results, activation of hypoxia signaling by UPS impairment may be responsible for increased autophagy in the absence of hypoxic conditions in a subset of established cancers.