The first description of autophagy as

The first description of autophagy as a tumour suppression process arises from the observation that the initial step regulatory gene, Atg6/BECN1, was monoallelically lost in 40% to 75% of human prostate, breast, and ovarian cancers [31]. However, while BECN1 heterozygous mutant mice develop, with long latency, lymphomas, liver and lung tumours, Splitomicin of other essential autophagy genes, such as ATG5 or ATG7 in mice, specifically in the liver and pancreatic tissues, produces only benign adenoma (Fig. 1). The concept that BECN1 could be a tumour suppressor (based on the studies demonstrating its allelic loss in tumour tissues) could be confused by the presence, in the same locus on human chromosome 17q21, of another tumour suppressor, BRCA1 (breast and ovarian tumour suppressor breast cancer 1). BRCA1 mutations are involved in the early onset of human breast and ovarian cancers. In fact, the current explanation is that the driver mutation in hereditary ovarian and breast cancers are dependent from germline BRCA1 missense mutation and the subsequent somatic loss of wild-type allele, that sometimes include (or not include) BECN1 deletion. BECN1 mutation can be considered in this scenario only a passenger mutation. To confirm this observation, exploring the large-scale genomic analysis of human cancers (Cancer Genome ATLAS, TCGA), no recurrent mutations in BECN1 or other essential autophagy genes are found, with few exceptions [32], [33]. Sequencing over 10.000 human cancers and matching the DNA sequence of normal tissues in TCGA database, large deletions were found on chromosome 17 comprising BRCA1 and BECN1 sequences in breast and ovarian tumours or deletion of only BRCA1, not BECN1, confirming that the driver mutation is BRCA1. The DNA sequencing data indicate that the loss of BECN1 in human cancers is strictly related to the loss of BRCA1 and, finally, that BECN1 could not be a tumour suppressor in most human cancers [33]. Nevertheless, autophagy could be still considered a tumour suppressor process in specific tissues, such as liver and pancreas, because autophagy-deficient mice (systemic mosaic deletion of ATG5 and liver-specific ) develop benign liver and pancreas adenomas [34] (Fig. 1). This could be explained considering autophagy an essential process to suppress the initial stages of liver or pancreas carcinogenesis for its role in controlling organelle integrity and protein quality and to suppress inflammation [29]. To reinforce this hypothesis, another suppressor gene strictly correlated to autophagy is Parkin, an E3 ubiquitin ligase, essential for the clearance of damaged mitochondria (a process called “mitophagy”). This gene is frequently deleted in human cancer on chromosome 6q25-q26, even if another Parkin function is related to cyclin D and E stability, in the context of cell cycle control, so this latter function could also explain its role in tumour suppression [35]. An important autophagy receptor is p62/SQSTM1 (p62, a complex “adapter” protein characterized by a multi-domain structure) which includes a LIR domain (LC3 Interacting Domain) essential to drive cargo and lipidated LC3 to autophagosome [36] (Fig. 2). This function explains why p62 accumulates in cells when autophagy flux is blocked or in autophagy-deficient cells. In animal models lacking p62, this genetic alteration inhibited the development of lung cancer, while its gain of function (p62 amplification on chromosome 5q) is linked to renal cancer tumorigenesis [37]. In addition to its role in autophagy, p62 can activate two transcription factors: Nrf2 (through KIR domain binding with its negative regulator Keap-1) and NF-κB (through TB domain binding with TRAF6). The final effect is a ROS (reactive oxygen species)-mediated stress response which enhances cell proliferation and induces stress resistance in cancer cells, favouring carcinogenesis [36]. These data reinforce the vision of autophagy as an essential process to block cancer formation at its initial stage (Fig. 1).