Contrasting Views on the Role of AMPK in Autophagy

Abstract

Efficient management of low energy states is vital for cells to maintain basic functions and metabolism and avoid cell death. While autophagy has long been considered a critical mechanism for ensuring survival during energy depletion, recent research has presented conflicting evidence, challenging the long-standing concept. This recent development suggests that cells prioritize preserving essential cellular components while restraining autophagy induction when cellular energy is limited. This essay explores the conceptual discourse on autophagy regulation during energy stress, navigating through the studies that established the current paradigm and the recent research that has challenged its validity while proposing an alternative model. This exploration highlights the far-reaching implications of the alternative model, which represents a conceptual departure from the established paradigm, offering new perspectives on how cells respond to energy stress.

Graphical Abstract

Newly discovered functions of AMPK are transforming our understanding of its role in autophagy, highlighting its dual role as an inhibitor and a safeguard. This dual functionality prepares cells for long-term resolution of energy stress, fostering efficient recovery. AMPK distinctly regulates the mTORC1 pathways associated with cell growth and autophagy.

Introduction

Eukaryotic cells, which possess ample energy reserves, can mobilize intracellular resources for survival when external energy supply is limited. However, when cells lack adequate intracellular energy stores, allocating their limited resources becomes critical for survival. It is widely recognized that cells adhere to evolutionarily-conserved principles when navigating energy stress. These principles involve prioritizing energy allocation for essential metabolic processes to maintain basic cellular functions. Concurrently, cells engage in macroautophagy, often referred to as autophagy, to break down cellular components for energy. Notable but less emphasized is that cells employ strategies to protect critical cellular components from degradation. This protection potentially supports cells in sustaining their ability to survive and recover from challenging conditions. Disruptions in these mechanisms can have profound consequences in various physiological and pathological situations.

Depletion of cellular energy, specifically adenosine triphosphate (ATP), can result from impaired metabolism or disturbances in the external energy supply. Glucose serves as a primary source of cellular energy, and its shortage can profoundly impact cell survival. Traditionally, glucose starvation has been associated with inducing autophagy, an evolutionarily-conserved catabolic process through which cells break down cytoplasmic components to retrieve essential building blocks and energy sources, such as amino acids, fatty acids, and glucose. The importance of autophagy for cell and organism survival during energy stress has been extensively studied using genetic knockout models. However, these investigations have yielded conflicting results, leading to controversy regarding the effects of glucose starvation on autophagy. In contrast, amino acid starvation robustly induces autophagy compared to glucose starvation. Autophagy induction is primarily mediated by the mechanistic target of rapamycin complex 1 (mTORC1)^1–7, a central negative regulator of autophagy. During amino acid starvation, reduction in the protein kinase activity of mTORC1 leads to the alleviation of its inhibitory effect on autophagy initiation. Glucose starvation can also reduce mTORC1 activity by enhancing the activity of AMPK (adenosine monophosphate-activated protein kinase), a central energy sensor that negatively regulates mTORC1^8–10. The AMPK-mTORC1 link is widely known to serve as a mechanism by which glucose starvation induces autophagy.

The autophagy-promoting role of AMPK has long been central to the prevailing paradigm regarding how autophagy is regulated during energy stress. It has served as the conceptual framework for interpreting numerous studies investigating the link between energy stress and autophagy^4,11 (Figure 1a). However, a growing body of evidence has challenged this conventional view^12–21. Notably, comprehensive biochemical analysis has revealed that AMPK does not promote but rather negatively regulates autophagy initiation²² (Figure 1b). In this evolving view, cells do not immediately trigger autophagy in response to energy depletion. Instead, cells prioritize the preservation of essential autophagy components from degradation. This preservation ensures the availability of the autophagy machinery when needed during the recovery from stress. This recent development has represented a conceptual departure from the current paradigm, demanding a reevaluation of the long-held concept.

Figure 1. Contrasting models of AMPK-mediated regulation of autophagy in response to energy stress.

(a) Prevailing model: AMPK promotes autophagy during energy stress. (b) Revised model: AMPK exhibits a dual role in autophagy control. It inhibits prompt induction of autophagy while supporting the stability of the autophagy machinery for long-term survival of cells.

In light of these recent advancements, this essay provides a comparative overview of two contrasting models regarding the role of AMPK in autophagy. This overview seeks to explore how the current paradigm has been established as the conventional concept in the field and how it has been questioned by recent studies. This discussion highlights the limitations of the current paradigm and emphasizes how a new model derived from recent studies provides a compelling alternative explanation for autophagy regulation during energy stress. The main purpose of this essay is to inspire further research toward a better understanding of the mechanisms that govern how cells adapt to energy stress.

Long-standing paradigm: energy depletion promotes autophagy

The concept that energy depletion induces autophagy has a rich historical background, initially established in yeast models^23–33. Studies employing gene knockout approaches have revealed the crucial role of autophagy in cell and organismal survival during energy depletion^23,34. Investigating the underlying mechanisms led to the identification of AMPK, or its yeast orthologue Snf1, as a key mediator of autophagy induction during energy depletion^{4,11,30,31,35,36}. AMPK, a primary energy sensor^37,38, plays a central role in governing various cellular functions related to metabolism and cell survival, which has been extensively discussed in numerous review papers^{39 40 37 38}. AMPK becomes activated in response to conditions that reduce cellular energy levels, such as glucose starvation and mitochondrial dysfunction, in a manner dependent on liver kinase B1 (LKB1)^41–43, and its activation can also be influenced by cellular calcium levels^44–46 (Figure 1).

AMPK negatively regulates mTORC1, a protein kinase complex that plays a central role in promoting protein synthesis and inhibiting autophagy initiation in response to amino acid levels (Figure 2a). AMPK achieves this by phosphorylating Tsc2 (tuberous sclerosis complex 2)⁹ and raptor (regulatory associated protein with mTOR)⁴⁷. Other key roles of AMPK involve the inhibition of lipogenesis by phosphorylating and inhibiting acetyl-CoA carboxylase (ACC)¹⁰. Furthermore, AMPK phosphorylates mitochondrial fission factor (MFF)⁴⁸, a protein that promotes mitochondrial fragmentation, thereby enhancing mitochondrial fission and integrity. This process contributes to diminishing the generation of reactive oxygen species (ROS) and facilitates the repair of damaged mitochondria. In addition to these roles, AMPK governs various transcriptional and epigenetic factors essential for metabolic control^49–53. Through these multifaceted mechanisms, AMPK ensures cellular adaptation to energy stress, contributing to the maintenance of metabolic homeostasis.

Figure 2. The role of AMPK in the prevailing model.

(a) Upon glucose starvation or energy stress, AMPK activation results in the suppression of mTORC1 activity. This action enhances ULK1 activity and autophagy. AMPK also directly activates ULK1 through phosphorylation. Additionally, AMPK phosphorylates the Atg14-containing Vps34 complex, enhancing its lipid kinase activity necessary for autophagy initiation. (b) In conditions of inactive mTORC1, such as during glucose or amino acid starvation, AMPK forms a stable interaction with ULK1. This interaction enables AMPK to phosphorylate ULK1, stimulating its kinase activity.

While AMPK could promote autophagy by inhibiting mTORC1, the prevailing model emphasizes a more direct role of AMPK in promoting autophagy initiation. According to the model, AMPK initiates autophagy by phosphorylating and activating ULK1 (UNC-51 like kinase 1)^4,11 (Figure 2a). ULK1 is a protein kinase responsible for triggering key molecular events involved in autophagy initiation^44–46, such as the activation of the Atg14-associated Vps34 complex^54–57. When both amino acids and glucose are abundant in cells, mTORC1 phosphorylates ULK1, disrupting the interaction between AMPK and ULK1 (Figure 2b). In this state, AMPK loses its ability to phosphorylate and activate ULK1. During amino acid starvation or mTORC1 inhibition, AMPK forms a stable interaction with ULK1 and robustly phosphorylates and activates ULK1. Similarly, glucose starvation enhances the interaction, enabling AMPK to activate ULK1. The model also proposes that AMPK directly phosphorylates and activates the Vps34 complex⁵⁸.

Paradigm in doubts

The prevailing model, established in the early 2010s, has provided the conceptual foundation for interpreting results in numerous studies. However, the model has encountered challenges from a growing body of research. Studies utilizing AMPK activators have shown evidence that AMPK activation inhibits or fails to induce autophagy^15,18–21. The primary event highlighted in the model, the phosphorylation of ULK1 at Ser556 (or mouse ULK1 Ser555) by AMPK², has been found to be abolished rather than promoted during conditions of mTORC1 inhibition or amino acid starvation^14,15,22. This phosphorylation was discovered to negatively regulate ULK1 activity²², contrary to the model. Notably, the interaction between AMPK and ULK1 was disrupted, rather than stabilized, upon mTORC1 inhibition or amino acid starvation^16,22. Furthermore, AMPK was proposed to inhibit ULK1 by phosphorylating Atg13, a binding partner of ULK1⁵⁹. Even before the model was proposed, some studies showed that glucose starvation suppresses, rather than promotes, autophagy^12,13. Additionally, the role of autophagy in cell survival during glucose starvation has not been entirely clear in the literature^13,60,61.

The emergence of the conflicting findings has cast doubts on the validity of the paradigm model. However, those publications have not gained enough momentum to drive a paradigm shift without presenting an alternative model. Nevertheless, the persistent challenges have demanded continued scrutiny to clarify the precise role of AMPK in autophagy. Given this circumstance, it is worthy of note why the field has predominantly adhered to the paradigm despite the accumulating contradictory evidence. Numerous studies leaned on the model, occasionally leading to misinterpretations as they sought to fit their findings within its framework. This issue might be related at least in part to the intrinsic complexity of autophagy regulation and the dynamic variations among different cell types and experimental conditions. These complexities might have inclined investigators to rely on the model rather than seek alternative explanations when interpreting results.

Often overlooked in many publications is the inadequate distinction made between short-term and long-term effects of glucose starvation on autophagy. When examining dynamic cellular processes, such as autophagy, it is important to differentiate between immediate and sustained responses to stimuli. Numerous studies have investigated the role of AMPK under prolonged glucose starvation over extended periods, such as overnight or even days. Such prolonged deprivation can profoundly alter cellular physiology, impacting metabolism and complicating the direct assessment of glucose starvation effects on autophagy. This metabolic alteration adds complexity to interpreting the regulatory role of AMPK as a signaling mediator. Such extended periods would involve multiple rounds of autophagy, which typically lasts just 1 to 4 hours each. Therefore, it is ideal to conduct analyses within this time frame to elucidate immediate molecular changes influenced directly by AMPK, avoiding secondary effects from prolonged autophagy activation or depletion. Another pertinent issue concerns the use of knockout cells and animals. While AMPK knockout cells have provided valuable insights, permanently eliminating AMPK can disrupt the entire signaling system and metabolism, introducing complexities in discerning genuine signaling events controlled by AMPK.

The observed discrepancy could also be attributed to insufficiently robust autophagy assays in prior publications. Numerous studies have heavily relied on markers such as autophagosome accumulation or LC3 II levels to assess autophagy, often overlooking the essential evaluation of autophagy flux. This omission could have led to conflicting interpretations of actual phenomena. Additionally, many publications have relied on monitoring the AMPK-mediated phosphorylation of ULK1 Ser556 as a simplified autophagy measure without any additional assessments. However, this phosphorylation has been observed to suppress ULK1 activity²². Notably, replacing this residue with alanine destabilized ULK1²², suggesting that the reduced autophagy observed in cells expressing the alanine mutant¹¹ might have stemmed from its destabilizing effect on ULK1. It is also important to note that the widespread practice of overexpressing biomarkers to assay autophagy might have contributed to the discrepancy. Overexpressed biomarkers can potentially introduce artifacts by disturbing the natural balance of autophagy components, thereby influencing the autophagy process.

AMPK role in autophagy being reevaluated

Despite its widespread acceptance in the field, the paradigm model has never presented evidence demonstrating that AMPK promotes ULK1 activity within cells. When the paradigm model was introduced, measuring ULK1 activity within cells was challenging due to the lack of any clearly defined ULK1 substrates. Thus, the prevailing model largely relied on in vitro kinase assays that might not accurately reflect ULK1 activity within cells or in physiologically relevant contexts. Over the past decade, multiple ULK1 substrates, including Atg14 (Autophagy-related protein 14), Beclin 1, and Vps15 (Vacuolar protein sorting-associated protein 15), have been identified^{54–57,62–67}. Their phosphorylations have been shown to play important roles in driving amino acid starvation-induced autophagy^54,55,57,66. Additionally, the phosphorylations of Vps34, Atg13, FUNDC1 (FUN14 domain-containing protein 1), and Atg16L1 by ULK1 have been linked to different types of selective autophagy, such as mitophagy^62,67,68 and xenophagy⁶⁴. These findings have allowed for a thorough investigation into how AMPK regulates ULK1 activity within cells, leading to the demonstration that AMPK inhibits ULK1 activity. This discovery has facilitated the development of an alternative model depicting AMPK as a negative regulator of autophagy^22,69 (Figure 3a).

Figure 3. The role of AMPK in the new model.

(a) During glucose starvation or energy deprivation, AMPK inhibits ULK1 activity, resulting in the suppression of autophagy. (b) When cellular amino acid levels are high, AMPK forms a stable interaction with ULK1, which is facilitated by mTORC1-mediated phosphorylation of ULK1. In the absence of glucose, AMPK activation triggers ULK1 phosphorylation. This phosphorylation prevents the dephosphorylation of ULK1 at the mTORC1-targeted site (Ser758), hindering AMPK dissociation from ULK1. This results in a locked and inactive state of ULK1, resistant to activation in response to amino acid starvation.

At the core of the alternative model lies the AMPK-mediated phosphorylations of ULK1 at Ser556 and Thr660. These phosphorylations play a key role in suppressing rather than promoting ULK1 activation during glucose starvation. The phosphorylations also play a role in protecting ULK1 Ser758 phosphorylation, mediated by mTORC1, from dephosphorylation when mTORC1 activity or amino acid levels decline^4,22 (Figure 3b). While the paradigm model suggests that Ser758 phosphorylation disrupts the AMPK-ULK1 interaction, the new model proposes the opposite—that Ser758 phosphorylation stabilizes the interaction. As a result, AMPK maintains its stable association with ULK1 during glucose starvation, resistant to disruption by amino acid starvation or mTORC1 inhibition. This stable association enables AMPK to persistently maintain ULK1 in the inactive state.

The new model presents a distinct perspective on autophagy regulation during energy stress, departing from the current paradigm. In the new model²², energy depletion does not trigger autophagy. Instead, it suppresses its immediate induction. This suppression serves a beneficial purpose, enhancing the cellular ability for survival and recovery. When faced with energy depletion, cells confront uncertainty regarding the stress duration. In temporary energy shortages, autophagy might be unnecessary. In prolonged energy shortages, cells might prioritize non-autophagic mechanisms, such as lipolysis and fatty acid oxidation, to secure energy, delaying autophagy as a last-resort response. The key aspect of this model is that prompt induction of autophagy in energy-deprived cells may lead to unnecessary depletion of essential cellular constituents. Autophagy induced by starvation could either selectively or non-selectively break down essential cellular substrates^70–73, posing a risk to essential components necessary for cell survival. Consequently, prioritizing autophagy as an immediate response to energy depletion might impede vital processes necessary for survival and recovery.

While autophagy is recognized for its roles in supporting cellular renewal, detoxification, and providing building blocks for recovery, its contribution to energy production during energy depletion requires further clarification. The primary cellular strategy in response to energy depletion might involve extending cell survival until external energy sources become available. Relying solely on internal energy resources is not a sustainable strategy to prolong cell survival. External energy sources are ultimately necessary for continued viability. Consequently, cells experiencing energy depletion may minimize energy expenditure and allocate available energy resources to essential cellular processes rather than actively engaging in autophagy, which could potentially deplete critical intracellular constituents necessary for recovery. During energy scarcity, cells may prioritize maintaining a minimal energy threshold necessary for basic metabolic functions rather than striving for high energy production.

Autophagy needs energy or functional mitochondria

The new model is in a stark contrast from the traditional view that autophagy primarily serves to supply energy to energy-deprived cells. Often underappreciated is that autophagy inherently demands energy supply^74–77. Autophagy entails extensive membrane fusion and rearrangement processes that require energy. Furthermore, maintaining the acidic pH within lysosomes, crucial for successful autophagy completion, relies on ATP⁷⁷. A recent study has shown that DFCP1 (Double FYVE-Containing Protein 1), a protein involved in forming initial membrane structures for autophagosomes^78,79, possesses ATPase activity⁸⁰. The need for energy in autophagy is further supported by recent research demonstrating that functional mitochondria are necessary for autophagy induction^22,81 or for maximizing autophagy⁸². This emphasizes the crucial role of mitochondria in facilitating autophagy initiation. This interplay suggests that mitochondria have a broader role beyond their previously recognized function in supplying membrane lipids for autophagosome formation^83–85. The connection between mitochondria and autophagy is particularly compelling, as the mitochondrial activity is ultimately necessary for producing ATP from autophagy-derived products, such as fatty acids. If the mitochondrial function is compromised, autophagy would serve no purpose in providing energy to cells.

AMPK: stimulator versus sensitizer for autophagy induction

Another key aspect of the new model pertains to how AMPK regulates autophagy in nutrient- and energy-rich environments. In such conditions, AMPK maintains the autophagy initiation machinery, specifically ULK1 and Atg14-Vps34 complexes, in highly sensitive states ready for activation²². During this phase, the phosphorylation of ULK1 Ser556 persists moderately due to basal AMPK activity. While this phosphorylation reduces basal ULK1 activity, this reduction amplifies the contrast in ULK1 activity between its resting state and full activation (Figure 4a). Thus, the basal AMPK activity contributes to the heightened sensitivity of ULK1 activity to amino acid starvation. In contrast, AMPK deficiency results in increased basal ULK1 activity, reducing the sensitivity of ULK1 activity to amino acid starvation.

Figure 4. AMPK sensitizes autophagy machinery under basal conditions.

(a) AMPK reduces the basal activity of ULK1, enhancing the sensitivity of ULK1 for activation upon amino acid starvation. Through a stable interaction, AMPK phosphorylates ULK1 to a moderate degree. This phosphorylation is sufficient to keep ULK1 activity low. This basal low activity amplifies the difference in ULK1 activity upon amino acid starvation. (b) In nutrient-rich basal conditions, AMPK phosphorylates ULK1 and Atg14-Vps34 complex, sensitizing them to be ready for activation upon amino acid starvation.

While basal autophagy in nutrient-rich environments is widely perceived as essential for cellular homeostasis and survival, proving its specific significance is technically challenging. Our understanding heavily relies on studies using cultured cells lacking specific autophagy-related genes. These cells typically thrive under normal nutrient conditions but exhibit reduced viability when exposed to stress, such as amino acid or energy deficiency^{22,43,86–89}. Caution is needed when interpreting results from such autophagy-defective cells, as observed effects might not solely be attributed to defects in basal autophagy but could involve secondary consequences resulting from the permanent depletion of autophagy-related genes, creating an artificial scenario. This parallels the perception of AMPK as a positive regulator of autophagy, based on studies using AMPK-deficient cells. AMPK is necessary for the expression of key autophagy-related genes, as discussed below. This role of AMPK may contribute to cellular physiology by maintaining an environment favorable for autophagy induction upon stress. However, this role, largely inferred from knockout cell studies, might not necessarily indicate that AMPK activation triggers autophagy.

The newly understood role of AMPK also suggests that its basal activity plays a role in maximizing the activation of the Atg14-associated Vps34 complex, a key player in autophagy initiation²². This heightened sensitivity may be due to the increased sensitivity of ULK1 activity to starvation, as ULK1 phosphorylates and stimulates the Vps34 complex^54,55,57 (Figure 4b). An alternative mechanism might involve direct phosphorylation and regulation of the Vps34 complex by AMPK^58,90. However, this possibility requires further investigation²². Song et al. showed that Beclin 1 phosphorylation by AMPK increases apoptosis and reduces autophagy⁹¹, contributing to uncertainty regarding its role in Vps34 complex activation. Nonetheless, it remains plausible that AMPK phosphorylates Beclin 1 or other components of the Vps34 complex not to directly stimulate the activity but to maximize its activation by ULK1 during amino acid starvation. AMPK also targets several other proteins involved in autophagy, such as Atg9, RACK1 (Receptor for Activated C Kinase 1), and PAQR3 (Progestin and AdipoQ Receptor Family Member 3)^92–94. While those phosphorylations alone might not be sufficient to induce autophagy, they could prime these molecules to facilitate autophagy upon initiation. Through this intricate regulation, AMPK could prepare various steps of autophagy initiation, ensuring their synchronized and coordinated engagement.

Two distinct pools of AMPK-mTORC1 for cell growth and autophagy

An apparent inconsistency might immediately be noticed with the new model when attempting to understand how AMPK suppresses autophagy while inhibiting mTORC1^9,47. This seemingly paradoxical role of AMPK could be reconciled by considering two distinct functions of AMPK in regulating cell growth and autophagy (Figure 5). This distinction became apparent when comparing mTORC1-mediated phosphorylation events on S6K1 and ULK1 following AMPK activation or knockout: while AMPK activation suppressed mTORC1-mediated S6K1 Thr389 phosphorylation, linked to cell growth regulation, it did not have the same effect on mTORC1-mediated ULK1 phosphorylation at Ser758²². Conversely, AMPK knockout enhanced the S6K1 phosphorylation but reduced the ULK1 phosphorylation. These differing roles of AMPK in cell growth and autophagy pathways might be attributed to distinct AMPK-mTORC1 pools, each influencing either cell growth or autophagy. This mirrors how mTORC1 selectively modulates different targets depending on various stimuli⁹⁵. A key takeaway is that AMPK primarily impacts autophagy by directly interacting with ULK1 rather than through mTORC1 inhibition, whereas its inhibition of mTORC1 primarily regulates cell growth.

Figure 5. Two distinct populations of mTORC1 in regulating cell growth and autophagy.

There may exist two distinct pools of mTORC1, one for cell growth regulation (^GmTORC1) and the other for autophagy regulation (^AmTORC1). The superscripts G and A represent growth and autophagy, respectively. AMPK negatively regulates mTORC1 signaling in cell growth regulation but not in autophagy regulation.

Nonetheless, the inhibitory role of AMPK in autophagy requires mTORC1. AMPK reinforces the inhibitory influence of mTORC1 on ULK1 activity (Figure 3b). In nutrient-rich conditions, mTORC1 phosphorylates ULK1 at Ser758, stabilizing the AMPK-ULK1 interaction and ensuring stringent inhibition of ULK1 by AMPK. As amino acid levels decrease, mTORC1 activity diminishes. Consequently, the inhibitory phosphorylation of ULK1 Ser758 is removed, leading to ULK1 activation. In the event of either glucose deprivation or simultaneous decreases in both amino acid and glucose levels, AMPK becomes activated and phosphorylates ULK1, stabilizing mTORC1-mediated ULK1 Ser758 phosphorylation. This prevents the dissociation of AMPK from ULK1, even when amino acid levels decline. While the exact mechanism behind this stabilization remains uncertain, this model highlights the intricate interplay among three kinases—mTORC1, AMPK, and ULK1—in orchestrating autophagy initiation in response to cellular nutrient and energy conditions.

AMPK: breakdown for immediate energy retrieval versus preservation to prepare for recovery from energy stress

Another critical role of AMPK portrayed in the new model is its contribution to preserving ULK1. During prolonged periods of energy depletion, ULK1 becomes susceptible to caspase-mediated degradation²². AMPK intervenes in this degradation by phosphorylating ULK1, thereby safeguarding it from breakdown²² (Figure 6). This phosphorylation occurs at Ser556 and Thr660, the same sites involved in suppressing ULK1 activity. While either Ser556 or Thr660 phosphorylation is essentially required, it is possible that other AMPK-mediated phosphorylations of ULK1^4,11,17 may also contribute to enhancing ULK1 stability. AMPK also contributes to the stabilization of the Atg14-associated Vps34 complex during glucose starvation²². The precise mechanism of this stabilization, particularly whether it involves direct phosphorylation by AMPK, requires further investigation. Notably, a study by Song et al. showed that AMPK-mediated phosphorylation of Beclin 1 at Ser93/96, previously associated with enhanced Vps34 complex activity⁵⁸, triggers caspase-mediated degradation of Beclin 1⁹¹, emphasizing the need for additional research to clarify the exact function of AMPK in regulating the Vps34 complex.

Figure 6. AMPK preserves autophagy machinery for long-term resolution of energy stress for survival.

(a) Cells rely on AMPK to exhibit two distinct responses to glucose starvation or energy depletion. AMPK rapidly suppresses ULK1 activity while preserving key autophagy components, such as ULK1 and Atg14-Vps34 complex. AMPK-mediated phosphorylation stabilizes ULK1 and the Atg14-Vps34 complex, preventing their degradation by caspases and contributing to cell survival. In conditions with abundant amino acid levels, AMPK forms a stable interaction with ULK1, phosphorylating and inhibiting ULK1 and Vps34 activity. Upon glucose deprivation, AMPK activation further increases the phosphorylation of ULK1 and the Vps34 complex, rendering them inactive and resistant to amino acid starvation-induced activation. Under amino acid starvation, AMPK dissociates from ULK1, leading to ULK1 activation. The active state of ULK1 phosphorylates and stimulates the Atg14-Vps34 complex. (b) Revised model for AMPK role in autophagy, highlighting its dual function as an inhibitor and preserver.

The safeguarding role of AMPK emphasizes the cellular strategy to ensure long-term survival during energy stress. As proposed by the new model²², when faced with energy depletion, cells prioritize preserving critical cellular elements, such as ULK1 and Vps34 complexes, over immediately initiating autophagy to break down intracellular components. In this context, AMPK restrains autophagy initiation during energy stress while maintaining the autophagy machinery prepared for efficient activation once the stress subsides (Figure 7). The preservation of essential cellular components enables cells to maintain the integrity of vital cellular functions or the cellular competence for autophagy, thereby contributing to recovery and cell survival after extended periods of energy stress.

Figure 7. AMPK maintains autophagy competence during energy stress.

(a) Dynamic changes of AMPK activity and autophagy competence during glucose starvation. Early in glucose starvation, AMPK activity increases, phosphorylating ULK1 and the Atg14-Vps34 complex, which stabilizes them and preserves cellular autophagy competence. As glucose starvation continues, AMPK activity decreases, resulting in the dephosphorylation of ULK1 and the Vps34 complex, which destabilizes them. The red arrows indicate cellular treatment with a medium containing high glucose but lacking amino acids, designed to test autophagy competence. (b) AMPK activating agents sustain cellular autophagy competence. Treatment with AMPK activating agents, such as A769662, MK-8722 or phenformin, prolongs AMPK activity, preventing the dephosphorylation of ULK1 and the Atg14-Vps34 complex. Consequently, ULK1 and the Vps34 complex remain stable during glucose starvation, sustaining cellular autophagy competence.

This protective role aligns with the well-documented function of AMPK in upregulating various autophagy-related genes, enhancing the cellular capacity for autophagy. AMPK controls several transcription and epigenetic factors, including FOXO3 (Forkhead box O3), TFEB (Transcription Factor EB) and CARM1 (Coactivator-ssociated Arginine Methyltransferase 1), to stimulate the expression of autophagy-related genes^49–51. Furthermore, it promotes the expression of genes involved in mitochondrial biogenesis^96,97, enhancing the cellular capacity to generate energy. AMPK also regulates Nrf2 (Nuclear factor erythroid 2-related factor 2), a transcription factor for genes related to antioxidant defenses and repair processes^98–102. However, caution is needed when interpreting these roles of AMPK during glucose starvation or energy depletion. Gene expression requires an adequate supply of nucleotides and energy. Thus, testing the effects of AMPK activators or AMPK knockout in normal energy conditions may not accurately reflect the role of AMPK during energy depletion. In energy-depleted cells, gene transcription and translation might be hindered because those processes entail high energy demands. Therefore, AMPK might not directly promote gene or protein expression in energy-depleted cells. Instead, AMPK might prepare transcription and epigenetic factors via phosphorylation to ensure their readiness to regulate gene expression when energy stress subsides. This hypothesis warrants experimental validation.

AMPK role in physiology and disease may need reevaluation

Understanding how autophagy responds to energy stress in physiological contexts is a challenging task, due to the diverse responses of different cell types to reduced energy levels and the intricate crosstalk between tissues. During fasting, our body initiates compensatory hormonal responses to balance energy needs across tissues. Some cell types, such as hepatocytes, adipocytes, and myocytes, can harness energy reserves such as lipids and glycogen without immediately resorting to autophagy during energy depletion. In contrast, many cell types depend on other tissues, such as the liver and adipose, for energy resources and might not promptly engage in autophagy to resolve energy stress. The coordination of energy utilization among tissues relies on systemic hormonal actions that manage the energy requirements of the entire body during fasting. Such systemic responses complicate the direct assessment of the impact of AMPK or energy depletion on autophagy in specific cell types in vivo. Furthermore, the hormonal system plays a key role in maintaining stable glucose levels in the bloodstream, with any temporary decrease during fasting usually short-lived. Hence, it is conceivable that hormonal responses play a significant role in regulating autophagy induction in tissues.

During fasting, our body swiftly adjusts its hormonal balance by reducing insulin levels and elevating glucagon levels in the bloodstream. This shift results in diminished insulin signaling and increased glucagon signaling across various tissues and cell types. Glucagon triggers the activation of cAMP (cyclic adenosine monophosphate) signaling pathways in tissues such as the liver, facilitating the conversion of glycogen into glucose^103,104. While the precise mechanisms governing autophagy in response to glucagon remain subjects of ongoing investigation^105–107, recent studies have provided insights. Ruan et al. showed that glucagon induces autophagy in hepatocytes¹⁰⁸, whereas Johanns et al. suggested that AMPK counters hepatic glucagon-stimulated cAMP signaling¹⁰⁹, potentially suppressing glucagon-induced autophagy. Further research is necessary to precisely delineate the impact of glucagon on autophagy.

The reduction in tissue insulin action during fasting can trigger the upregulation of autophagy through inhibiting mTORC1, a downstream component of insulin receptor signaling^110–113. This upregulation of autophagy might support energy production. Alternatively, reduced insulin signaling can lead to increased lipolysis and gluconeogenesis independently of autophagy^114,115, generating fatty acids and glucose as energy sources in the liver or other tissues. The autophagy-independent processes to break down lipid or glycogen could serve as an immediate response to energy depletion, as they do not involve the extensive membrane processes and energy expenditure associated with autophagy. However, autophagy induction later during prolonged fasting could be beneficial, potentially offering improved efficiency in energy production compared to lipolysis or glycogen hydrolysis. The timing of such a delayed induction of autophagy would likely rely on hormonal responses and specific regulatory factors in tissues. The preserving role of AMPK, as proposed by the new model, might support such a delayed autophagy induction.

An extensively studied area relevant to this topic is exercise-induced autophagy. Several studies have reported that exercise induces autophagy^116–120, linked to AMPK activation^116,118. However, uncertainties remain regarding whether activated AMPK immediately initiates autophagy, as suggested by the paradigm model, or adopts the preservation roles, as proposed by the new model. This uncertainty mainly stems from challenges in precisely assessing autophagy flux in muscle tissues. Indicators like LC3B or ULK1 Ser555 phosphorylation indirectly hint at autophagy processes but do not directly measure autophagy flux. Furthermore, understanding the energy contribution of exercise-induced autophagy remains elusive. Autophagy-defective mouse models cannot clarify this, as they affect not just exercise-regulated autophagy but also basal autophagy that might be responsible for detoxification and muscle homeostasis. Thus, while AMPK activation during exercise might contribute to muscle physiology through enhancing various processes, such as glucose uptake, mitochondrial biogenesis, and insulin sensitivity, it might not engage in immediate induction of autophagy but instead restrain its initiation while preserving the autophagy machinery. Autophagy might play a crucial role in post-exercise muscle recovery rather than serving as an energy source, aligning with the preserving role of AMPK. With the evolving understanding of the role of AMPK in autophagy, reassessing its role in exercise-regulated autophagy in muscle is warranted.

Conclusion

This essay has explored key concerns about the paradigm model, highlighting the need for continued scrutiny to discern which models offer a more effective explanation for autophagy regulation during energy stress. Challenges are notable in this pursuit due to the inherent complexity of the dynamic autophagy process or from premature conclusions drawn based on inadequate autophagy assays. This issue can be especially challenging when evaluating the contrasting models in physiological and pathological contexts, spanning areas such as aging, muscle exercise, ischemia, neurodegeneration, and cancer, for which the precise roles of autophagy demand explicit clarification. Future investigations employing both the alternative and paradigm models may offer opportunities to unveil previously overlooked aspects regarding the role of AMPK in various physiological and pathological contexts through autophagy. Such studies should encompass the intricate interplay among nutrient availability, signaling pathways, and hormonal responses, utilizing more rigorous measurement parameters. Given the central role of AMPK in numerous pathological processes, insights garnered from the alternative model may redirect the course of therapeutic development efforts.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this review.