Metabolic axis of autophagy: A key player in tumor maintenance and opportunities for therapeutic exploitation

Abstract

Autophagy is an evolutionary conserved catabolic process that recycles the metabolites which serve as building blocks of biomolecules and support cellular growth. Dysregulation of selective and bulk autophagy is associated with many diseases including cancer and plays dual role. The present review examines tumor protective and supportive functions of autophagy-dependent metabolic rewiring and explores potential therapeutic targets to overcome drug resistances in anti-cancer treatments. Many recent studies argue that disruption of core autophagic genes or use of autophagic inhibitors can significantly inhibit release of nutrients, impair metabolism and suppress tumor proliferation but may potentiate genomic instability under metabolic stress conditions and drive tumorigenesis. Coordinated upregulation of autophagy by modulating energy metabolism via either putting cells on short-term fasting, or treating cells with autophagy inducing caloric restriction mimetics/fasting mimicking diets followed by sensitization of tumor cells by targeting them with potential anticancer agents may hold great therapeutic potential. Robust autophagic induction results in differential stress sensitization wherein healthy cells enter a protective ‘low-power mode’ whereas tumor cells due to oncogenic mutations become more vulnerable to anticancer agents and undergo apoptotic cell death. Large scale randomized trials are required to establish the treatment efficacy of supplemental approach in clinical setting.

Key Words: Anti-cancer therapies; Autophagy activation and inhibition; Paradoxical role of autophagy; Metabolic rewiring; Therapy resistance

Core Tip: Autophagy is an evolutionary conserved catabolic process that ensures cell adaptability under stress conditions and maintains cellular homeostasis. Dysregulation of autophagy supports tumor growth where its paradoxical role (protective and supportive) appears to depend on tumor stage, driver mutations, tissue type and metabolic sensitivity. Production of metabolic fuel sources and extracellular energy in an autophagy dependent manner contribute to autophagic switch. Pharmacological inhibition of autophagy or depletion of core autophagic genes have been significantly shown to impair metabolism and suppress tumor proliferation. Many recent studies argue that disruption of autophagy/core autophagic genes may potentiate the genomic instability under metabolic stress conditions and drive tumor development. Paradoxical change in the functions of autophagy-dependent metabolic crosstalk in tumor maintenance and disease aggressiveness emphasizes the therapeutic exploitation of supplemental approaches.

INTRODUCTION

Bulk autophagy/macroautophagy is an evolutionary conserved catabolic process that captures the intracellular components (cellular organelles and other macromolecules) in a double-membraned autophagosomes/vesicles. Autophagolysosomes are formed upon fusion of these vesicles with lysosomes which degrade the cellular components into the metabolites[1]. Although low level of basal autophagy is present in most of the tissues, it is significantly induced during cellular stress and maintains cell survival. There are currently more than 40 proteins known which are encoded by autophagy-related protein (ATG). These proteins orchestrate the various steps of autophagosome formation, including initiation, nucleation, membrane expansion, closure and fusion with lysosome and execute the process of autophagy in a tightly regulated manner[2,3]. The other two forms of autophagy include microautophagy and chaperone mediated autophagy. Microautophagy is based on the engulfment of cellular components into the lysosomal membrane for its degradation by lysosomal hydrolases. Chaperone mediated autophagy involves the identification, tagging and translocation of damaged proteins by the chaperone proteins across the lysosomal membrane for its degradation[4]. Recent studies suggest that multiple selective autophagy pathways are constitutively active at basal level, however, specific stimuli activate them significantly against the particular stress to maintain cellular homeostasis. Selective autophagy pathways are named based on the cargo destined for degradation via selective autophagy receptor protein. Further mechanism includes the recruitment of autophagosomal machinery for delivery to lysosomes for degradation. Selective autophagy pathways include mitophagy, aggrephagy, lipophagy, ferritinophagy, xenophagy, and endoplasmic reticulum (ER)/reticulophagy, for mitochondrial, protein aggregates, lipids, ferritin, bacteria, and ER degradation respectively[5].

Autophagy is induced by cellular stresses like nutrient starvation, hypoxia, oxidative stress and infection to ensure cell adaptability to stressed conditions and thus maintains the cellular homeostasis. The activities of four main signal-sensing kinases, mechanistic target of rapamycin (mTOR) complex 1/2, adenosine monophosphate-activated protein kinase (AMPK), and unc-51 like autophagy activating kinase 1/2 (ULK1/2), and protein kinase B (AKT) are the key regulators of autophagy and are tightly controlled by various macro nutrients and micro nutrients. These kinases inhibit/activate the process of autophagy to maintain the normal physiology of cells. Metabolites including sugars, fatty acids, amino acids and nucleotides thus formed are recycled to the cytoplasm and serve as the building blocks for the biosynthesis of macromolecules and cellular growth.

Dysregulation of selective and bulk autophagy has been associated with many diseases including neurodegenerative disorders, metabolic diseases, cardiovascular diseases, infectious diseases and cancer. Ongoing studies examine the dual functions of autophagy in cancer which appears to depend on tumor stage, driver mutations, tissue type and metabolic sensitivity. Tumor suppressive functions of autophagy by sustaining the metabolic demands under stressful conditions, maintaining the genomic integrity and preventing proliferation are widely accepted. Autophagy generates extracellular adenosine triphosphate (ATP) which is thought to be responsible for the recruitment of immune cells (dendritic cells and cytotoxic T lymphocytes) to tumor bed and stimulate anti-cancer immune responses[6,7]. Autophagy inhibition results in ATP degradation to adenosine and facilitates accumulation of anti-immune T-regulatory cells, escape from immune detection and tumor development[8]. Tumor cells exploit autophagy to survive cellular stresses and proliferate in the hostile tumor microenvironment. Production of metabolic fuel sources and extracellular energy under cellular stress in an autophagy dependent manner provide tumors with metabolic plasticity and allow them to grow in amenable tumor microenvironment.

Paradoxical change in functions from tumor suppressor/cytotoxic to tumor promoter/cytoprotective also known as ‘autophagic switch’ within the tumor cells (intrinsic) and in the surrounding stroma (extrinsic) is context-dependent and controversial. This mini-review comprehensively summarizes the tumorigenic effects of the crosstalk between autophagy and metabolic pathways (glucose, amino acid and lipid metabolisms) across the tumor cells and the surrounding stroma. Understanding the tumor protective and supportive functions of autophagy-dependent metabolic rewiring would certainly help to explore potential therapeutic targets to overcome drug resistances in anti-cancer treatments.

AUTOPHAGY: UPSTREAM SIGNALS AND MECHANISTIC REGULATION

Macroautophagy, the principal autophagic pathway typically constitutes five stages including initiation, nucleation, membrane expansion, fusion with lysosome, and degradation of the autophagy cargo. Serine kinase activity of mTORC1, AMPK and ULK1/2 is required for the formation of pre-initiation complex. The p53 signaling activates tuberous sclerosis complex 1/2 (TSC1/TSC2) which in turn inhibits mTORC1, a complex that regulates cell growth and metabolism. Ras homologue enriched in brain, a guanosine triphosphate (GTP)-binding protein, extracellular signal regulated kinases 1/2, mitogen-AMPKs, phosphoinositide-dependent kinase-1 and AKT activate mTORC1 and thus inhibit autophagy[9]. Formation of pre-initiation complex comprising of ATG13, focal adhesion kinase-family interacting protein 200 kDa (FIP200), ATG101 and ULK1/ULK2, is regulated by AMPK under nutrient stress conditions. Direct phosphorylation of the regulatory-associated protein of mTOR complex 1 (raptor) protein or indirect phosphorylation of TSC2 by AMPK inhibits mTOR activity. Phosphorylation of ULK1 (ser637 and ser757) and ATG13 (ser258) by mTORC1 under nutrient rich conditions disrupts AMPK-ULK1 complex and suppresses autophagy[10]. On the other hand, separation of ULK1 from mTORC1 complex followed by its autophosphorylation at threonine 180^th position brings about the phosphorylation of other members of ULK1 complex and activates the process of autophagy[11].

Lipid kinase activity of vacuolar sorting protein-34, a class III phosphatidylinositol 3-kinase (PI3K) is important for the formation of initiation complex [ATG14 L, vacuolar protein sorting (VPS) 15 and other regulatory factors, in addition to Beclin1] during biogenesis of autophagosomes and protein localization. Binding of B-cell lymphoma 2 (Bcl2) to N-terminal Bcl2 homology 3 domain of Beclin1 inhibits the formation of autophagosomes[12]. ULK1 phosphorylates Beclin1/ATG6, activates lipid kinase activity of VPS34, increases the production of phosphatidylinositol-3-phosphate from phosphatidylinositol-2-phosphate and recruits ATG5-ATG12/ATG16 L conjugation complex to the growing phagophore. Ligase activity of ATG5-ATG12/ATG16 complex is needed for pulling of the processed microtubule-associated protein 1 light chain 3 (LC3)/ATG8 to nascent phagophores and its conjugation to phosphatidylethanolamine and mark cargo for degradation in an ubiquitin-dependent autophagy[13,14]. An autophagy adaptor protein known as sequestosome 1 (SQSTM1) or p62 after binding to cargo interacts with LC3-II via LC3-interacting region and facilitates the delivery of cargo to autophagosomes. Specific soluble N-ethylmaleimide-sensitive factor attachment protein receptor like Syntaxin 17 and synaptosomal-associated protein 29 on the autophagosome and vesicle-associated membrane proteins 7/8 or synaptobrevin homolog (YKT6) on the lysosome are essential for the membrane fusion of autophagosome and lysosome. Lysosome-associated membrane proteins 1 and 2 (LAMP1 and LAMP2) ensure proper lysosome docking and its fusion with autophagosome. Fusion of autophagosome with lysosomes triggers the breakdown of cellular debris and recycling of breakdown products (nucleotides, fatty acids and amino acids) to the cytosol to promote cellular homeostasis (Figure 1)[15-17].

Figure 1 The macroautophagy pathway and recycling of macromolecules. AKT: Protein kinase B; AMPK: Adenosine monophosphate-activated protein kinase; ATG: Autophagy-related protein; ATP: Adenosine triphosphate; LC3: Light chain 3; mTORC: Mechanistic target of rapamycin complex; PI3KC1: Phosphatidylinositol 3-kinase complex 1; PE: Phosphatidylethanolamine; ULK1: Unc-51 like autophagy activating kinase 1.

Microautophagy is characterized by the autonomous modification of lysosomal membrane via invagination and formation of lipid enriched autophagic tubes mediated by ATG7-dependent ubiquitin-like conjugation or via vacuolar transporter chaperone molecular complex. This is followed by the vesicle enlargement and engulfment of cellular components for its degradation by lysosomal hydrolases. It is primarily involved in the maintenance of the organellar size and the composition of biological membrane. It degrades the lipids by engulfing them into the lysosomal vesicles and thus keeps a check on the lysosomal/vacuolar membrane composition. It allows the cells to survive under nitrogen-limited conditions and maintains the membrane proteins turnover[18]. Process of mitophagy is regulated by phosphatase and tensin homolog-induced putative kinase 1/Parkin pathway (ubiquitin-dependent pathway). It is associated with changes in mitochondrial membrane potential and is triggered by inhibitors of the respiratory chain, protein toxicity, and mitochondrial reactive oxygen species (ROS)[19]. As a result, accumulation of putative kinase 1 on the outer mitochondrial membrane and subsequent recruitment and activation of E3 ubiquitin ligase Parkin to the mitochondrial surface cause polyubiquitination of various mitochondrial outer membrane proteins. These proteins are recognized by autophagy adaptor proteins with LC3-interacting region motifs, including SQSTM1/p62, optineurin (OPTN), nuclear dot protein 52, and neighbor of breast cancer 1 gene 1 (NBR1). This is followed by mature autophagosome formation and fusion with lysosomes for degradation of damaged mitochondria[19]. ER-phagy brings about the selective degradation of redundant ER membranes and insoluble or toxic protein aggregates by direct binding of specific receptor proteins including SQSTM1/p62, Bcl2 interacting protein 3 (BNIP3), family with sequence similarity 134 member B, reticulon-3 L, SEC62 homolog, preprotein translocation factor (SEC62), cell cycle progression protein 1, testis expressed 264, and calcium binding and coiled-coil domain protein 1 with ATG8/LC for transportation of damaged components/intracellular protein clumps to lysosome for degradation[19]. Degradation of protein aggregates/misfolded proteins by aggrephagy depends on the interaction of cargo receptors including SQSTM1/p62, NBR1, toll interacting protein, tax1 binding protein 1, and OPTN with LC3 on autophagosomes. Lipophagy mediated degradation of lipids by lysosomal enzymes into free fatty acids and its oxidation regulates lipid metabolism. Direct interaction of lipid transfer protein oxysterol-binding protein-related protein 8 with LC3 promotes envelopment of lipid droplets by autophagosomal membranes and its degradation[19]. Degradation of harmful pathogens by xenophagy is used by host system to protect from invading pathogens and maintain cellular homeostasis. Recruitment of cargo to the autophagosomal membrane depends on selective interaction of receptors (SQSTM1/p62, nuclear dot protein 52, OPTN, and NBR1) with ATGs[19].

Prolonged starvation/oxidative stress conditions promote the degradation of damaged/altered proteins directly by the process of chaperone mediated autophagy. Target substrate proteins containing KFERQ motif form a complex with heat shock protein of 70 kDa (cytosolic chaperone protein) and is sequestered to the cytoplasmic tail of the LAMP2A. Following the formation of translocation complex via LAMP2A multimerization, the substrate proteins are directed to lysosomal matrix for its degradation[20-23].

PARADOXICAL FUNCTIONS OF AUTOPHAGY

Role of autophagy in cancer is complex and has been extensively investigated over the past few years. Paradoxical functions of autophagy are influenced by the cancer stage, tumor suppressor gene constellation of a tumor, nutrient availability, state of immune system, pathogenic conditions, microenvironmental stress and the model systems used to investigate its function.

TUMOR SUPPRESSIVE EFFECTS OF AUTOPHAGY

The basal autophagy eliminates excessive production and accumulation of ROS, reduces the emergence of mutagenic factors, establishes immune surveillance, provides cells the protective milieu and keeps a check on tumor development[24]. Being the major homeostatic mechanism, it maintains the cellular integrity, redox balance and proteostasis. Tumor suppressive mechanism/function of autophagy has been well examined and supported by various recent studies. Tumor suppressive function of autophagy was initially supported and based on two major evidences. First, the autophagy inhibition was examined to be associated with activation of oncogenes e.g., phosphatidylinositol-4, 5-bisphosphate 3-kinase catalytic subunit alpha or inactivation of phosphatase and tensin homolog tumor suppressor gene. Second, defective autophagy due to deletion of autophagy genes was predicted to result tumor initiation in mouse models in a tissue-specific and gene-specific manner. Frequent allelic loss of Beclin1 and mice hemizygosity of Beclin1 in breast cancer cell lines and primary mammary tumor tissues are often associated with tumorigenesis. Tumor formation in lung, liver and lymphatic tissues in mice is observed to be due to Beclin1 hemizygosity[25,26]. Another study examined the effect of ATG7 gene deletion alone without any other genetic event on tumor formation in liver[27]. Cycles of tissue destruction and regeneration due to defective autophagy in the liver resulted in the emergence of hepatocyte-derived progenitor cells and thus promoted tumor initiation[28]. Frameshift mutations in ATG2B, ATG5, ATG9B and ATG12 were reported in gastrointestinal and liver cancers whereas downregulation of ATG5 and ATG7 was observed in melanoma[29,30]. Deletion of the mitophagy receptors BNIP3 or BNIP3 L in mouse models promoted the development of pancreatic and breast cancers[31,32]. Activation of nuclear factor kappa B cells and nuclear factor erythroid 2-related factor 2 pathways mediated by p62 supported tumor development. Concomitant deletion of p62 is shown to reverse the tumor development (caused due to deletion of key autophagy genes) in the liver of mouse models[27]. Further, owing to the involvement of p62 in suppressing the degradation of Twist1, a key epithelial-mesenchymal transition-activating transcription factor (ATF), its overexpression promoted mesenchymal differentiation and increased metastasis in vivo[33]. Several studies spotted the important role of autophagy in maintaining the disseminated tumor cells in dormant phase and thereby suppressing metastatic colonization and outgrowth. Knockdown of ATG3 in mammary cancer cells made the tumor cells to exit dormancy and enter into highly proliferative metastatic phase with increased cancer stem cell like properties[34]. Escape from dormancy and early metastatic recurrence was observed in dormant breast cancer models (dormancy induced via doxorubicin treatment) upon ATG5 knockdown compared to autophagy proficient control cells[35].

TUMOR PROMOTING EFFECTS OF AUTOPHAGY

Autophagy eliminates the defective proteins and organelles, helps the cells to overcome stressful conditions including nutrient deficiency and hypoxia and inhibits apoptosis during later stages of tumor development. Besides, it also helps the cells to survive therapy stresses which could be due to chemotherapy, radiotherapy and targeted agents, thereby confers therapeutic resistances[36-38]. Genetic or pharmacological inhibition of autophagy in various model systems is shown to slow down the growth of tumor cells. Studies on genetically engineered mouse models of cancer driven by oncogenic Ras examine the critical role of autophagy in the pathogenesis of pancreatic ductal adenocarcinomas[39]. Recent study on in vivo model of hepatocellular carcinoma suggested the involvement of autophagy in anoikis resistance and metastatic dissemination[40]. In another study, genetic deletion of FIP200 in polyoma middle T oncogene-driven (PyMT) mammary tumor model demonstrated autophagy inhibition and thus reduced primary tumor growth as well as metastasis to lung[41]. Further, autophagy inhibition and reduced tumor growth was noted in PyMT cells genetically deficient for either ATG12 or ATG5[42]. One of the key roles of autophagy in tumor maintenance is to fulfil the high metabolic demand of the proliferating cells, mitigate the limited availability of external nutrients and thus maintains the viability and survival of cancer cells[43,44].

METABOLIC PATHWAYS REGULATING AUTOPHAGY

Metabolic stresses are built up when rate of growth of tumor cells exceeds the rate of angiogenesis. Metabolic rewiring and autophagy in such hostile condition ensure the supply of energy and building blocks for tumor growth and survival. Metabolism in tumor cells becomes more anabolic and undergoes a switch from oxidative phosphorylation to glycolysis. Glycolytic intermediates are subsequently diverted to biosynthetic pathways to synthesize metabolites which are essentially required for tumor growth[45]. The following section discusses the mechanistic functions of metabolic pathways and metabolites in tumor cells in the regulation of autophagy.

GLUCOSE METABOLISM AND AUTOPHAGY

The Warburg effect was observed by Otto Warburg in the 1920s and is a hallmark of cancer. Tumor cells experience a metabolic shift from oxidative phosphorylation to aerobic glycolysis and lactic acid fermentation for energy production even in the presence of oxygen. Tumor cells being more sensitive to low levels of glucose compared to normal cells induce autophagy by activating AMPK and inactivating mTOR. Glucose starvation leads to the binding of hexokinase-II [catalyses the conversion of glucose to glucose-6-phosphate (first step in glycolysis)] to mTOR complex 1 and inactivates mTOR[46]. Decrease in glycolysis rate lowers the levels of fructose 1,6-bisphosphate (glycolysis intermediate) and makes the aldolase enzyme available to interact with and inhibit ER localized transient receptor potential channel subfamily V to suppress calcium release. Decrease in calcium levels at ER/Lysosome contact site allows the binding of inactive transient receptor potential channel subfamily V to lysosomal vacuolar-type adenosine triphosphatase or vesicular-type adenosine triphosphatase. As a result, Axin and liver kinase B1 are recruited for AMPK activation and autophagy induction[47,48]. Besides, lower levels of glucose are also responsible for increasing the intracellular levels of adenosine monophosphate/adenosine diphosphate/ATP for AMPK activation (Figure 2).

Figure 2 Regulation of autophagy by glucose metabolism. AMPK: Adenosine monophosphate-activated protein kinase; ATP: Adenosine triphosphate; FBP: Fructose 1,6-bisphosphate; LKB1: Liver kinase B1; mTORC: Mechanistic target of rapamycin complex; TRPV: Transient receptor potential channel subfamily V; ULK: Unc-51 like autophagy activating kinase.

Aerobic glycolysis in tumor cells produces lactate and is transported to neighbouring oxidative tumor cells to maintain metabolic symbiosis. Protons are generated during the conversion of lactate to pyruvate catalysed by lactate dehydrogenase B. Protons promote vesicular-type adenosine triphosphatase-dependent lysosome acidification and autophagosome maturation (Figure 2)[49].

ACETYL COENZYME A, NICOTINAMIDE ADENINE DINUCLEOTIDE (PLUS) HYDROGEN/NICOTINAMIDE ADENINE DINUCLEOTIDE⁺ AND AUTOPHAGY

Amino acids, lipids and glucose catabolisms generate acetyl coenzyme A which acts a cofactor for histone acetyltransferases p300 (EP300 acetyltransferases). This enzyme catalyzes the acetylation of LC3, VPS4 and raptor component of mTOR to negatively regulate autophagy by affecting the autophagosome maturation and formation processes[50-52]. Sirtunin family deacetylases require nicotinamide adenine dinucleotide⁺ as a cofactor, deacetylate several ATG proteins and activate autophagy[53]. High nicotinamide adenine dinucleotide (plus) hydrogen/nicotinamide adenine dinucleotide⁺ is reported to inhibit autophagic activity via suppressing Sirtunin (Figure 3)[54].

Figure 3 Acetyl coenzyme A, nicotinamide adenine dinucleotide (plus) hydrogen/nicotinamide adenine dinucleotide⁺ in autophagy regulation. Acetyl CoA: Acetyl coenzyme A; ATG: Autophagy-related protein; LC3: Light chain 3; NAD⁺: Nicotinamide adenine dinucleotide⁺; VPS: Vacuolar protein sorting.

AMINO ACID METABOLISM AND AUTOPHAGY

Mechanisms adapted by amino acids to stimulate mTOR-mediated signaling and inhibit autophagy have been intensively studied (Figure 4). Various cytosolic sensors including Sestrin, cytosolic arginine sensor for mTORC1 subunit 1 (CASTOR1 and CASTOR2) allow these proteins to interact with GTPase-activating protein activity towards recombination activating genes (Rags)-2 (GATOR2) and activate mTORC1. S-adenosylmethionine sensor upstream of mTORC1 (SAMTOR) activates mTOR in the presence of methionine. Upon binding of respective amino acids, cytosolic sensors dissociate from GATOR2 towards Rags. The released GATOR2 inhibits the guanosine triphosphatase (GTPase)-activating protein activity of GATOR1, thereby activates mTOR through Rag GTPases. Secretion associated Ras related GTPase1B is identified as another sensor for leucine. Upon its dissociation from leucine, it binds to GATOR2 and regulates mTOR activity[54]. Leucyl-transfer RNA synthetase senses cytosolic leucine, regulates Rag and stimulates mTOR activity[55].

Figure 4 Regulation of autophagy by amino acid metabolism. AMPK: Adenosine monophosphate-activated protein kinase; ATF: Activating transcription factor; ATP: Adenosine triphosphate; elf2α: Eukaryotic initiation factor-2 alpha; tRNA: Transfer RNA; ULK: Unc-51 like autophagy activating kinase.

Lysosomal arginine-dependent Rag GTPase nucleotide state switching by solute carrier family 38 member 9 is reported to activate mTORC1 pathway[56]. ADP-ribosylation factor 1 is a family of small GTPases that play crucial role in vesicular trafficking within the cells and activates mTOR in the presence of glutamine[57]. Leucine, glutamine, and glutamate contribute to the generation of acetyl coenzyme A via anaplerotic metabolism and acetylation of Raptor, thereby enhance mTOR activity[50].

Amino acid deficiency results in an increase in uncharged transfer RNA species and activates the serine/threonine-protein kinase general control nonderepressible 2. Activated control nonderepressible 2 phosphorylates eukaryotic initiation factor 2 alpha subunit and reduces the translation of proteins preferentially of ATF. ATF now acts as a transcription factor and activates the expression of autophagic target genes thereby accelerates autophagy[58]. In another study, generation of ammonia during glutaminolysis (conversion of glutamine to α-ketoglutarate) to maintain tricarboxylic acid cycle ATP and nicotinamide adenine dinucleotide phosphate production is shown to stimulate autophagy via AMPK activation but independent to ULK1 activation or mTOR inhibition. Nevertheless, very high concentration of ammonia (more than 20 mmol/L) is examined to affect lysosomal function and inhibit autophagic flux[59].

LIPID METABOLISM AND AUTOPHAGY

Impact of altered lipid metabolism on autophagy in tumor cells has been examined over the recent years (Figure 5). One of the most abundant saturated fatty acid, palmitate, activates autophagy in a protein kinase C and Jun N-terminal kinase 1 dependent mechanism in a cell specific manner[60,61]. Oleate, another monosaturated fatty acid elevates autophagy by increasing the ROS levels[60]. Enzymes involved in lipid metabolism including fatty acid synthase (catalyzes the de novo synthesis of fatty acids) inhibit autophagic flux whereas stearoyl coenzyme A desaturase 1 which catalyzes the conversion of saturated fatty acids into unsaturated fatty acids, stimulate starvation-induced autophagy[62,63]. However, another study reports the effect of stearoyl coenzyme A desaturase 1 inhibition in inducing autophagic cell death via stimulating the nuclear translocation of forkhead box O 1 in Tsc2 knockout mouse embryonic fibroblasts[64].

Figure 5 Regulation of autophagy by lipid metabolism. JNK: C-Jun N-terminal protein kinase; PKC: Protein kinase C; ROS: Reactive oxygen species; SCD: Stearoyl-coenzyme A desaturase.

AUTOPHAGY-DEPENDENT METABOLIC CROSS TALK IN TUMOR MAINTENANCE

Autophagy-mediated release of breakdown products (sugar, lipids, nucleotides and amino acids) in stromal cells (endothelial cells, immune cells, fibroblasts in a tumor microenvironment) serve as the building blocks for macromolecular synthesis. Energy and macromolecules thus produced are supplied to tumor cells to support their growth and survival in the epithelium. Increase in autophagy is reported in pancreatic stellate cells (part of tumor stroma) in the transplantation models of pancreatic ductal adenocarcinomas. Induced autophagy generates and secretes nonessential amino acid extracellularly which is later utilized by pancreatic tumor cells for their growth and survival in adverse microenvironments[65]. It is also reported that autophagy in pancreatic stellate cells contributes to the secretion of alanine locally which is taken up by the tricarboxylic acid cycle of tumor cells as an alternative carbon source for the synthesis of macromolecules and facilitate tumor growth[66]. Increase in the demand of nonessential amino acid, arginine during tumor development renders tumor cells auxotrophic for this amino acid[67]. Release of the arginase I (arginine-degrading enzyme) from the liver into the blood during whole-body or liver-specific deletion of autophagy has been studied for its poor association with a distant primary tumor growth in the lung[68]. This makes autophagy inhibition an important intervention for the cessation of tumor growth. Another study on the effect of systemic inhibition of stromal cell autophagy on reduced host-tumor metabolite transfer has been demonstrated in mice models with lung tumor. Autophagy inhibition via transient ATG5 knockdown is shown to suppress the uptake of glucose and lactate into Kras^G12D/+; p53^-/- lung tumors and impaired tumor growth[69]. The hypoxic tumor microenvironment induced autophagy via hypoxia-inducible factor 1 in cancer-associated fibroblasts allows the release of nutrients (glutamine, glycolytic intermediates, and ketone bodies), supports mitochondrial metabolism, tumor growth and metastasis[70,71].

THERAPEUTIC INTERVENTIONS BY TARGETING AUTOPHAGY-DEPENDENT METABOLIC CROSS-TALK IN CANCER TREATMENT

Recent studies collectively demonstrate the versatile role of autophagy in providing cancer cells with metabolic plasticity and supporting the established cancers to proliferate unconstrainedly in mouse models. Pharmacological inhibition of autophagy or depletion of core autophagic genes have been significantly shown to impair metabolism and suppress tumor proliferation. Commonly used autophagic inhibitors include antimalarial drug chloroquine (CQ) and its derivative hydroxy CQ (inhibit final step of autophagy and limits acidification of lysosomes); and 3-methyladenine (targets class-III PI3K). Table 1 lists clinical studies on the use of autophagy inhibitor as monotherapeutic in human cancer types. Reduced glycolysis flux and suppressed mammary tumor initiation and progression is reported in mouse mammary tumor virus-polyoma middle T driven breast cancer mouse model post deletion of FIP200. Deletion of FIP200 resulted in accumulation of ubiquitinated protein aggregates and p62/SQSTM1, deficient LC3 conversion, reduced expression of cyclin D1 and increased number of mitochondria with abnormal morphology in tumor cells[72]. Tumor harboring Ras and Raf mutations exhibit increased dependency on autophagic process both in vitro and in vivo. These mutations result in the metabolic depletion of cellular energy but autophagy helps in tumor cell survival by preserving mitochondrial integrity and providing metabolic substrates required for growth. Recent studies on conditional deletion of ATG5 or ATG7 in tumor cells in mouse models of Ras-driven cancers observed although an increase in incidence of premalignant lesion but cessation of growth of established malignant tumors and therefore, increase in the survival of mice[73,74]. Reversion of carcinomas to benign oncocytoma-like tumors upon autophagy inhibition was observed in Kras-driven lung cancer models[73-75]. Post several months of whole-body genetic inhibition of autophagy through ATG7 deletion, adult mice were observed to develop number of metabolic disorders including starvation intolerance, liver glycogen, muscle mass and gradual loss of white adipose tissue. This study noticed greater degree of regression of Ras driven tumors upon acute and systemic deletion of ATG7/autophagy in all the cells rather than deletion only in tumor cells[76]. Faster rate of tumor regression compared to systemic metabolic and neurologic deterioration in mice models opens up a new therapeutic window of autophagy inhibition in cancer treatment. Deletion of ATG5 or ATG7 in the host cells was found to be associated with the release of arginine degrading enzyme arginase 1 by the liver into the circulation. This enzyme is responsible for the degradation of serum arginine and thus limits the growth of arginine auxotrophic tumors[68]. Dominant-negative ATG4b mutant under the control of a tetracycline-inducible promoter in both the tumor and host cells of fully formed Kras-driven pancreatic tumor models resulted in acute inhibition of autophagy and tumor suppression[77]. Study by Yang et al[78] reported the effect of genetically induced inhibition of autophagy across the normal mouse pancreas on absence of metaplasia or growth of benign lesions. Raf-mutated tumors harbouring V600E mutations, characteristic of melanoma, were examined to depend on autophagy for maintaining mitochondrial integrity and reduced oxidative stress, thereby contributing to poor prognosis due to their aggressive nature[78].

Table 1 List of preclinical/clinical studies on use of autophagy inhibitors as monotherapeutic agent and their mechanism of action in human cancer types.

Cancer type(s)	Pre-clinical/clinical phase	Trial identifier
Melanoma (stage III/IV)	0 (pilot study)	NCT00962845
Renal cell carcinoma	I	NCT01144169
Chronic lymphocytic leukemia	II	NCT00771056
Metastatic pancreatic adenocarcinoma	II	NCT01273805
Breast ductal carcinoma in situ	I/II	NCT01023477
Brain metastases with whole brain radiation therapy	II	NCT01894633
DCC-3116 (ULK1 inhibitor): Targets ULK1 and blocks early autophagy initiation
Kras-mutated or Raf-mutated advanced solid tumors	I (first in human)	NCT04892017
SBI-0206965 (ULK1 inhibitor): Targets ULK1 kinase and inhibits early phase autophagy initiation
Non-small cell lung cancer cells, neuroblastoma	Preclinical studies on cell lines and mouse models	Not available
Verteporfin (photoactivatable autophagy inhibitor): Inhibits autophagosome formation and lysosomal degradation
Glioblastoma, pancreatic ductal adenocarcinomas	I/II	NCT03033225
3-Methyladenine: Targets class III phosphatidylinositol 3-kinase (vacuolar protein sorting 34), thereby blocks autophagosome formation (early-stage autophagy)
Cervical cancer, breast cancer, glioblastoma, hepatocellular carcinoma, colon cancer, ovarian cancer, pancreatic cancer, leukemia, mantle cell lymphoma	Preclinical studies on cell lines and murine models	Not available
Lys05 (lysosome deacidifier): Late-stage autophagy inhibitor
Glioblastoma, colorectal cancer, melanoma, acute myeloid leukemia	Preclinical studies on cell lines and murine models	Not available

ULK: Unc-51 like autophagy activating kinase.

Autophagy inhibition was examined to make fatty acid oxidation pathway defective in tumor cells with deleted p53[75]. These findings explain the anti-cancer therapeutic functions of autophagy inhibition or dual inhibition of autophagy and fatty acid oxidation contingent on the oncogene activation or tumor suppressor inactivation. Autophagy knockout studies conclude the effect of autophagy inhibition on preferential growth of benign tumors or only in those cells which harbour additional oncogenic insults. Tumor suppressive functions of docosahexaenoic acid, and eicosapentaenoic acid (polyunsaturated fatty acids) by phosphorylating AKT and activating mTOR and attenuating autophagosome formation in A549 lung cancer cells have been investigated[79].

Studies on Ras-driven tumor model in Drosophila melanogaster emphasize the importance of autophagy in the maintenance of tumor growth. Blunted tumor growth and invasion upon chemical or genetic inhibition of autophagy was observed in flies. Besides, resumption of exponential growth of stunted tumors post transplantation of autophagy into autophagy-deficient flies supports the momentousness of systemic inhibition of autophagy to limit the cancer growth[80].

COORDINATED AUTOPHAGY MODULATION, ANTICANCER AGENTS AND METABOLIC REWIRING: SUPPLEMENTAL THERAPEUTIC APPROACH

Many recent studies argue that disruption of core autophagic genes may potentiate the genomic instability under metabolic stress conditions and thereby drive tumorigenesis. Degradation of autophagic cargo receptor p62 serves as another important mechanism for tumor suppressive effects of autophagy. Absence of autophagy leads to the accumulation of p62 which in turn activates the expression of transcription factors, nuclear factor kappa B and nuclear factor erythroid 2-related factor 2 and promotes inflammation, angiogenesis and survival[81,82]. Autophagy promotes the lysosomal degradation of pro-tumor factors and mediates a senescence transition under oncogene-stimulated conditions. Study by Hao et al[83] examined the tumorigenic effect of autophagy inhibition by disrupting FIP200 in mammary tumor virus-Neu mouse breast cancer model. Disruption of autophagy promoted tumor development by switching of the post-golgi trafficking of human epidermal growth factor receptor 2 from plasma membrane to the endocytic compartment, thereby allowing its release from tumor cells through small extracellular vesicles[84]. Autophagy is examined to restrict the tumor formation. However, autophagy promotes tumorigenesis once tumor is formed. Owing to the tumor suppressive role of autophagy via establishing immune surveillance and maintaining genomic integrity, its induction has been examined to be helpful in eradicating the tumor cell population. Treatment with mTOR inhibitor, rapamycin resulted in tumor regression in tobacco carcinogen [4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone]-induced lung carcinoma in a murine model[83].

Recently investigated studies underpin the autophagy induction via modulating the autophagic core genes as one of the key mechanisms of resistance during cancer treatment (cytotoxic chemotherapy, targeted therapy, and radiotherapy) in multiple cancer types. Metabolic and therapeutic stress signaling in cancer cells and malignant stroma is amplified and in turn activate autophagy. Autophagy promotes the survival of targeted tumor cells by delaying the apoptosis and DNA damage mechanisms. Autophagy induction via AKT/mTOR pathway made anaplastic lymphoma kinase-positive lung cancer cells resistant to crizotinib therapy[85]. Non-small cell lung cancer cells H1650 when treated with Erlotinib (tyrosine kinase inhibitor) exhibited therapeutic resistance and increased tumor cell survival owing to globin transcription factor binding protein 6-upregulating autophagy[86]. Autophagy activation via PI3K/AKT/mTOR pathway resulted in the decline in the uptake of glucose as measured by 18[F] fluorodeoxyglucose positron emission tomography, acquired resistance and survival in the presence of cabozantinib (an inhibitor of receptor tyrosine kinases)[87]. Development of paclitaxel resistance has been observed to be associated with increased Beclin1 production and autophagy induction. Use of microRNA (miR), miR-216b, post transcriptionally downregulated Beclin1 by specifically targeting the 3’UTR of Beclin1 and helped in inhibiting autophagy and improving the paclitaxel treatment in non-small cell lung cancer cells therapy[88]. Overexpression of miR-224-3p (targets ATG5) or knockdown of ATG5 is shown to inhibit cell mobility with increased chemosensitivity of temozolomide by suppressing hypoxia-induced autophagy in glioblastoma and astrocytoma[89]. Autophagy enables the drug targeted tumor cells to enter into a dormant state which leads to the relapse of cancer cells and aggravates disease aggressiveness.

Coordinated upregulation of autophagy by modulating the energy metabolism via either putting cells on short-term fasting (STF), or treating cells with autophagy inducing caloric restriction mimetics (CRMs)/fasting mimicking diets (FMD) followed by the sensitization of tumor cells by targeting them with potential anticancer agents may hold great therapeutic potential (Figure 6). Metabolic stresses [STF or supplying the cells with non-nutrient small molecules (spermidine, resveratrol, hydroxycitrate, metformin) which mimic fasting signals without reducing calorie intake] induce autophagy. Energy metabolism gets reprogrammed [impaired aerobic glycolysis, mitochondrial bioenergetics (decreased electron transfer system and oxidative phosphorylation capacity, mitochondrial fission and fusion dynamics)] upon autophagy induction[90]. Robust autophagic induction is examined to result in differential stress sensitization wherein healthy cells enter a protective ‘low-power mode’ whereas tumor cells due to oncogenic mutations become more vulnerable to anticancer agents and undergo apoptotic cell death. It is hypothesized that STF or treating cells with CRMs/FMD initially upregulate autophagy but in combination with anticancer agents results in downregulation of autophagic markers as well as enhanced apoptotic and cytotoxic responses.

Figure 6 Two different therapeutic approaches to target tumor cells. A: Differential stress resistance; B: Differential stress sensitization. CRMs: Caloric restriction mimetics; FMD: Fasting mimicking diets; STF: Short-term fasting.

Study by Buono et al[91] examined the synergistic increase in the toxicity to pre-B-cell acute lymphoblastic leukemia cells and higher cancer cell death following the application of FMD in combination with vincristine. The cycles of a 3-day FMD given to high-fat-diet-fed mice once a week increased the efficacy of vincristine, decreased the expression of multiple autophagy markers (ULK1, Beclin1, LC3B), activated p53, increased cluster of differentiation 8⁺ (CD8⁺) T cell toxicity and improved survival from breakpoint cluster region-abelson B-ALL[91]. In another study, combination of chemotherapy and FMD cycles is shown to increase the levels of bone marrow common lymphoid progenitor cells, cytotoxic CD8⁺ tumor-infiltrating lymphocytes and down-regulation of the stress-responsive enzyme heme oxygenase-1, leading to a major delay in breast cancer and melanoma progression[91]. FMD is studied to restore the sensitivity of cyclin-dependent kinase 4/6 inhibitors (CDK4/6i)-resistant cells to abemaciclib and potentiate the anti-tumor activity of CDK4/6i in triple-negative breast cancer mouse models. The anti-tumor effects of FMD and/or CDK4/6i were accompanied by the inhibition of mTORC1 signaling, downregulation of S6 phosphorylation, lowered levels of insulin-like growth factor 1 and Ras[92]. Fasting switched the colorectal cancer cells from an active proliferative to a slow-cycling state. Fasting-induced quiescent cells were found to be more prone to develop drug-tolerant persister tumor responsible for cancer relapse and metastasis. Combining fasting with ferroptosis inducer treatment lead to tumor inhibition and eradication of quiescent cells by modulating autophagy[93]. Fasting cycles prior to high-dose etoposide reduced toxicity of high-dose etoposide in mice but did not reduce etoposide activity profile against neuroblastoma allografts in metastatic neuroblastoma mouse models. Fasting was observed to sensitize human mesothelioma and lung cancer xenografts to cisplatin whereas complete remissions were observed in only the combination treatments with cytotoxic agents. Many preclinical studies observed the effect of fasting and CRMs on increased clinical efficacy of anticancer agents in an immune system dependent and autophagy dependent fashion in metastatic models of melanoma, neuroblastoma, fibrosarcoma, sarcoma, lung, colorectal, breast, colon and pancreatic cancer. Clinical phase trials report the success of combinational treatment based on the autophagy induction during fasting/FMD/CRMs on enhanced clinical efficacy of chemotherapy, immunotherapy or endocrine therapy via promoting cancer cell death, protecting healthy cells and activating immune system[94]. Table 2 enlists and provides the details of clinical phase trials in humans. Small number of clinical trials often phase I/II/III are reported. They are designed for safety, feasibility, assessing biological marker changes but lack tumor efficacy endpoints. Large scale randomized trials are required to establish the treatment efficacy of supplemental approach based on combinational approach of drug use (cytotoxic chemotherapy, targeted therapy, and radiotherapy) and autophagy modulation by inducing metabolic stress in clinical setting.

Table 2 List of preclinical/clinical studies on autophagy modulation followed by application of anticancer drug and their key endpoints in human cancer.

Cancer types	Autophagic inducer	Anticancer therapy	Clinical phase	Key endpoints	Trial identifier
Advanced solid/hematologic malignancies	FMD repeated every 3-4 weeks, up to a maximum of eight cycles	Carboplatin-based chemotherapy, pembrolizumab, abraxane-gemcitabine, and XELIRI-bevacizumab	I/II	Safe and feasible; enhanced systemic and intratumor antitumor immunity; exceptional tumor responses in approximately 5 patients in sub-analysis	NCT03340935
HER2-negative breast cancer	STF (approximately 24 hours before and after chemo)	Docetaxel/doxorubicin/cyclophosphamide	II/III	Reduced hematologic toxicity and DNA damage in peripheral blood mononuclear cells vs non-fasting	NCT01304251
Breast and ovarian cancer	STF (36 hours before, 24 hours after chemo; total approximately 60 hours)	Adjuvant chemotherapy	III	Better quality of life and reduced fatigue during chemo cycles with fasting; no serious adverse effects	NCT01954836
HER2-negative stage II/III breast cancer	Cyclic FMD (3 days prior + during chemo) vs standard diet	Anthracycline-taxane	II/III	Radiological and pathological responses (miller and payne 4/5) more frequent with FMD; reduced DNA damage in T-lymphocytes; toxicity not significantly different despite no dexamethasone in FMD	NCT02126449
Solid and hematologic malignancies	Cyclic 5-day FMD administered every approximately 3 weeks; supplemented with nutritional and muscle training during refeeding	Chemotherapy, endocrine, targeted, checkpoint inhibitors, radiotherapy	I/II single arm	Lower IGF-1, leptin; better body composition; feasible	NCT03595540
Stage I-III triple negative breast cancer	Eight triweekly cycles of 5-day FMD	Metformin with standard preperative anthracycline-taxane chemotherapy	II	The pCR rate; plasma glucose, insulin, IGF-1, lipid and amino acid shifts; radiologic response, distant metastasis free survival, recurrence free survival, overall survival up to 5 years; adverse effects incidence/severity, adherence to regimen, dosage modifications; tumor/metabolic gene expression; association with pCR; biomarker discovery	NCT04248998
Gynecologic cancers (breast and ovarian cancer)	Intermittent fasting of 72-84 hours parallel to the application of the chemotherapy	Breast cancer: Epirubicin + cyclophosphamide + paclitaxel and doxorubicin + cyclophosphamide + docetaxel. Ovarian cancer: Paclitaxel + carboplatin	Not assigned	Lowered fatigue and improved quality of life. Evaluated chemotherapy side effects, tumor response, metabolic changes (e.g., weight, glucose, insulin), and long-term outcomes including recurrence and biomarker trends	NCT03162289
Advanced metastatic prostate cancer	60 hours fasting-mimicking diet (36 hours before + 24 hours after chemo)	Docetaxel in some cases; abiraterone (combined with prednisone)	Not assigned	Quality-of-life at 3 months and 6 months, hospital anxiety and depression scale score changes, blood counts, chemo side-effects	NCT02710721
Metastatic non-small cell lung cancer	FMD 3 days prior	Carboplatin + pemetrexed + pembrolizumab	II	Patients did not experience serious adverse events; DNA damage in blood cells; circulating tumor cell spheroid formation; changes in immune biomarkers	NCT03700437
Metastatic castratesensitive prostate adenocarcinoma	FMD 5 days/month for 6 months	Abiraterone, apalutamide, enzalutamide, or darolutamide	II	Metabolic outcomes, including weight, blood glucose, insulin levels; biomarkers of treatment response, such as changes in prostate-specific antigen and inflammatory markers	NCT05832086
Lung cancer	FMD	Cisplatin, carboplatin, and agents like pemetrexed, alongside metformin	II	Progression-free survival; feasibility and safety of the FMD + metformin combination; immune and metabolic biomarker changes (e.g., peripheral blood mononuclear cells, DNA damage markers, metabolic health); overall response rate and tolerability of adjunctive interventions	NCT03709147

FMD: Fasting mimicking diets; HER2: Human epidermal growth factor receptor 2; IGF1: Insulin-like growth factor 1; pCR: Pathologic complete response; STF: Short-term fasting.

CONCLUSION

Given the significant role of autophagy in the metabolism of various cancers, researches in the field of autophagy-dependent metabolic cross talk between tumor and tumor microenvironment (stromal cells) and its context dependent cytotoxic and cytoprotective effects have been exponentially expanded over the last two decades. Autophagy-mediated release of energy and breakdown products (glycolytic intermediates, ketone bodies, nucleotides and amino acids) in hypoxic tumor microenvironment feed into nearly every pathway in central carbon metabolism and beyond, support mitochondrial metabolism, and allow tumor growth and its dissemination.

Despite the intense efforts in pharmacologically inhibiting autophagy using autophagic inhibitors including antimalarial drug CQ and its derivative hydroxy CQ, and 3-methyladenine or conditionally depleting core autophagic genes (ATG5, ATG7, ATG4b) in cancer clinical trials, there is yet no definitive therapeutic edge. Pharmacological inhibitors or depletion of core autophagic genes as part of the treatment regimen are shown to suppress the preferential growth of benign tumors or only in those cells which harbor additional oncogenic insults. Owing to the lack of specificity in autophagy inhibition and inefficacy of these autophagic inhibitors at clinically permitted doses for variety of cancer histologies, there is a need to design and explore therapeutics for its greater impact on tumor regression by targeting tumor metabolic pathways.

Considering the benefits of autophagy in preventing tumor development and its fundamental role in maintaining the cellular homeostasis in normal tissues, elimination of excessive production and accumulation of ROS, reducing the emergence of mutagenic factors, establishing the immune surveillance, providing cells the protective milieu and maintaining the cellular integrity, the timing of therapeutic intervention based on autophagy induction/modulation is very crucial. Recently investigated studies vouch the autophagy induction via modulating the core autophagic genes as one of the key mechanisms of resistance during cancer treatment (cytotoxic chemotherapy, targeted therapy, and radiotherapy) in multiple cancer types. Autophagy is observed to promote the survival of drug-targeted tumor cells by delaying the apoptosis and DNA damage mechanisms and thus aggravates disease aggressiveness.

Preclinical and clinical phase (I/II/III) studies provide the early evidences of the therapeutic efficacy of novel supplemental approach based on the coordinated upregulation of autophagy by modulating the energy metabolism via either putting cells on STF, or treating cells with autophagy inducing CRMs/FMD, followed by the sensitization of tumor cells by targeting them with potential anticancer agents may hold great therapeutic potential. Robust autophagic induction is examined to result in differential stress sensitization wherein healthy cells enter a protective ‘low-power mode’ whereas tumor cells due to oncogenic mutations become more vulnerable to anticancer agents and undergo apoptotic cell death. Large number of randomized trials may be designed to utilize therapeutic combinations of more potent anticancer agents and autophagic activators via inducing metabolic stress that directly hit the core tumor metabolic pathways to overcome resistance to current oncologic treatments in clinical setting.