Autophagy and cancer stem cells: molecular mechanisms and therapeutic applications

Abstract

Autophagy and mitophagy act in cancer as bimodal processes, whose differential functions strictly depend on cancer ontogenesis, progression, and type. For instance, they can act to promote cancer progression by helping cancer cells survive stress or, instead, when mutated or abnormal, to induce carcinogenesis by influencing cell signaling or promoting intracellular toxicity. For this reason, the study of autophagy in cancer is the main focus of many researchers and several clinical trials are already ongoing to manipulate autophagy and by this way determine the outcome of disease therapy. Since the establishment of the cancer stem cell (CSC) theory and the discovery of CSCs in individual cancer types, autophagy and mitophagy have been proposed as key mechanisms in their homeostasis, dismissal or spread, even though we still miss a comprehensive view of how and by which regulatory molecules these two processes drive cell fate. In this review, we will dive into the deep water of autophagy, mitophagy, and CSCs and offer novel viewpoints on possible therapeutic strategies, based on the modulation of these degradative systems.

Fig. 1

Roles of autophagy in cancer stem cells (CSCs). Tumor cells are heterogeneous and include cancer stem cell (CSC) populations. CSCs are often characterized by high levels of autophagy that (1) maintain pluripotency; (2) cope with low nutrients and oxygen levels (hypoxia) in the tumor microenvironment; (3) regulate CSCs migration and invasion; (4) promote resistance to chemotherapy, (5) help to escape immunosurveillance; (6) support oncovirus capability to infect, replicate in and kill them. In this scenario, autophagy manipulation is found to be crucial for the effective targeting of cancer cells. Aa: Amino acids; Glc: Glucose; NK: Natural Killer; O₂: Oxygen

Fig. 2

Role of autophagy in breast cancer stem cells (CSCs). a In human breast cancer, two different populations of cancer cells co-exist, the breast and non-breast CSCs. While the first population is characterized by high autophagy levels, increased resistance to chemotherapy and mesenchymal phenotype, the latter shows decreased autophagy, higher sensitivity to drugs and an epithelial, rather than mesenchymal, phenotype. Also, breast and non-breast CSCs differentially express cell surface markers (CD24, CD44, and ALDH). Importantly, autophagy inhibition results into enhanced migration and stemness and alteration in IL-6 secretion. b. In a mouse model of breast tumor, two different populations of breast CSCs have been isolated: a luminal one (in pink) and a mesenchymal one (in light blue). Intriguingly, all stemness markers (ALDH, CD29, and CD61) are downregulated in both populations upon FIP200 depletion, and this event correlates with Stat3 or TGFβ2/3 signaling downregulation, respectively. Also in this case, the two different populations of CSCs are distinguished by differential expression of cell surface markers (ALDH, CD29, and CD61)

Fig. 3

Mitophagy in CSCs, a working model. In brain tumor-initiating cells (BTICs), CDK5-dependent DRP1 activation promotes mitochondrial fission, and (hypothetically) mitophagy, to sustain self-renewal and growth. In leukemia stem cells (LSCs), activated AMPK induces FIS1-mediated mitophagy to guarantee removal of damaged mitochondria, keeping ROS levels under tight control, and, thus, contributing to LSC proliferation. In fact, AMPK or FIS1 loss leads to accumulation of damaged mitochondria and increase of ROS production, which promote GSK3 activation that drives cell cycle arrest and differentiation. Furthermore, in hepatic CSCs, PINK1 on the one hand activates p53 through phosphorylation on the mitochondrial membrane and, on the other hand, mediates mitophagy-dependent p53 degradation, thus favouring NANOG expression and hepatic CSCs proliferation. The suppression of mitophagy efficiency entails the accumulation of activated p53 that translocates to the nucleus, where it inhibits NANOG expression, hindering hepatic CSC proliferation. Moreover, BNIP3L-mediated mitophagy contributes to doxorubicin resistance in colorectal CSCs. A hypothetical model upon hypoxia conditions: hypoxia activates HIF-1α that drives the CSCs metabolic reprogramming. We can hypothesize that HIF-1α endorses BNIP3 and BNIP3L expression and, in turn, these factors mediate mitochondrial degradation and contribute to the switch from oxidative to glycolytic metabolism