Autophagy begins when a double-membraned vesicle, the autophagosome, engulfs cytoplasmic cargo and delivers it to the lysosome for degradation. The process is orchestrated by a cascade of autophagy-related genes, or ATG proteins, including Atg5, Atg7, and Atg12, which conjugate lipid to LC3/GABARAP proteins to build the vesicle. Upstream, two nutrient sensors act as master switches: mTOR, which suppresses autophagy when nutrients are plentiful, and AMPK, which triggers it through direct phosphorylation of the ULK1 complex when energy runs low. Sirtuin 1 adds another layer of control by deacetylating autophagy components. In most tissues this system is a survival mechanism; in the hypothalamus, the review contends, it has been co-opted into the circuitry of appetite itself.

The arcuate nucleus contains two opposing neuronal populations that sit at the heart of energy balance. Pro-opiomelanocortin, or POMC, neurons produce alpha-melanocyte-stimulating hormone, a signal that suppresses food intake and raises energy expenditure through the central melanocortin system. Agouti-related peptide, or AgRP, neurons do the opposite: they release AgRP and neuropeptide Y, powerful orexigenic signals that ramp up during fasting and drive feeding. According to the review, autophagy pushes these two populations in opposite directions, and understanding that divergence is essential to any therapeutic logic.

In POMC neurons, autophagy is portrayed as a guardian of metabolic health. Experiments in which autophagy genes were deleted from POMC cells showed impaired lipolysis in peripheral fat tissue, disrupted alpha-MSH production, and diminished leptin and insulin responsiveness. Work by Kaushik and colleagues published in EMBO Reports in 2012 demonstrated that losing autophagy in POMC neurons impairs fat breakdown, while a study in Cell Metabolism by Coupé and colleagues found that autophagy-deficient POMC neurons fail to extend axons properly during development, causing lasting metabolic dysregulation. An ATF4-ATG5 signaling axis has also been implicated in regulating obesity through these neurons. Disrupting the process, in short, predisposes animals to hyperphagia, leptin resistance, and the metabolic dysfunction that accompanies obesity.

The story in AgRP neurons is nearly inverted. Here, autophagy appears to promote feeding and energy conservation: it maintains AgRP expression, integrates ghrelin signals from the stomach, and participates in sensing circulating fatty acids. The landmark study by Kaushik and colleagues in Cell Metabolism in 2011 showed that autophagy in AgRP neurons regulates food intake and energy balance, and later work by Chen and colleagues in 2023 revealed that nutrient-sensing AgRP neurons can even relay control of liver autophagy during energy deprivation — a striking example of central-to-peripheral command. Consistent with this picture, inhibiting autophagy in AgRP neurons alleviates diet-induced obesity in mice, suggesting that damping down the hunger program could, in principle, be protective.

Yet the review is candid about contradictions in the literature, and this critical appraisal is arguably its most valuable contribution. Deleting Atg12 from POMC neurons exacerbates diet-induced obesity, but deleting Atg5 does not — an unsettling discrepancy given that both genes participate in the same LC3 lipidation cascade. The authors attribute such inconsistencies to several variables: the specific autophagy gene targeted, whether the deletion occurs during development or in adulthood, the dietary paradigms used, and a widespread reliance on static markers such as LC3 puncta or p62 aggregates, which reflect the accumulation of autophagosomes rather than the true rate of autophagic flux. The fourth-edition guidelines for monitoring autophagy emphasize exactly this distinction, and many hypothalamic studies fall short of the dynamic measurements needed to resolve it.

The framework also extends beyond the arcuate nucleus. The review situates POMC and AgRP autophagy within broader hypothalamic and extra-hypothalamic networks, including newly identified basonuclin 2 neurons that acutely suppress food intake when activated by leptin, steroidogenic factor 1 neurons of the ventromedial hypothalamus — where defective autophagy perturbs the metabolic response to fasting and causes mitochondrial dysfunction — and downstream paraventricular pathways projecting to the brainstem, liver, and adipose tissue. Glial cells and tanycytes, the specialized ependymal cells lining the third ventricle, add further complexity: tanycytic TSPO inhibition has been shown to induce lipophagy and improve energy balance, and adenosine transmission from tanycytes to AgRP neurons regulates energy homeostasis. Obesity-induced blood-brain barrier dysfunction may further modulate how circulating nutrients and hormones reach these circuits.

A particularly intriguing thread involves chaperone-mediated autophagy, or CMA, a distinct form of lysosomal degradation in which the chaperone Hsc70 delivers specific proteins across the lysosomal membrane through the LAMP2A receptor. Studies in POMC-like neurons in vitro show that palmitic and stearic acids — saturated fatty acids abundant in Western diets — inhibit CMA, and lipid excess has been shown to impair CMA in the whole hypothalamus. Because CMA degrades lipid droplet-associated proteins and thereby facilitates lipolysis, its suppression by saturated fat could plausibly link dietary lipid exposure, hypothalamic proteostasis, and insulin responsiveness. Recent work even suggests that central activation of CMA reduces appetite by fine-tuning hypothalamic amino acid pools. The review stresses, however, that direct evidence for this pathway in vivo remains limited, and it should be treated as a hypothesis rather than a mechanism.

The translational picture is equally restrained. No study has yet demonstrated altered autophagic flux in human POMC or AgRP neurons, although a comprehensive spatio-cellular map of the human hypothalamus published in 2025 and region-specific transcriptomic data from macaques now make such investigation conceivable. Existing obesity therapies that touch the hypothalamus — setmelanotide for acquired hypothalamic obesity, GLP-1 receptor agonists, and metformin’s effects on hypothalamic inflammation — were not designed around autophagy. The review concludes that future therapies will likely require pathway-biased, neuron-type-specific, and circuit-aware modulation rather than global activation or inhibition of autophagy, since wholesale suppression could harm POMC neurons while selective inhibition of AgRP autophagy might help. A drug that broadly drenches the hypothalamus in an autophagy modifier would almost certainly deliver both benefit and harm to adjacent cell types.

For a field long dominated by accounts of hypothalamic inflammation and endoplasmic reticulum stress in obesity, the reframing is provocative: the same intracellular recycling pathway that keeps neurons alive may also calibrate the signals that make organisms hungry or satisfied. The authors acknowledge that the framework is best viewed as a mechanistic and translational research program, not a clinical endpoint. But with human hypothalamic atlases, flux reporters, and single-cell tracing of POMC subtypes now available, the tools to test whether autophagic flux in appetite neurons is genuinely altered in human obesity are, for the first time, within reach.

Subject of Research: The role of autophagy in hypothalamic POMC and AgRP neurons in regulating appetite and obesity

Article Title: Autophagy in hypothalamic POMC and AgRP neurons: mechanisms and therapeutic implications for obesity

Article References: Jing, D., Dai, R., Peng, S., Zeng, L., & Cai, J. (2026). Autophagy in hypothalamic POMC and AgRP neurons: mechanisms and therapeutic implications for obesity. International Journal of Obesity. https://doi.org/10.1038/s41366-026-02223-y

Image Credits: AI Generated

DOI: 10.1038/s41366-026-02223-y

Keywords: autophagy, hypothalamus, POMC neurons, AgRP neurons, obesity, leptin resistance, arcuate nucleus, chaperone-mediated autophagy, appetite regulation, Atg genes, energy homeostasis, melanocortin system

Macroautophagy, the best-studied form of autophagy, is an intracellular degradation system that is essential for cellular homeostasis. In this process, a double-membraned vesicle called an autophagosome forms around cytoplasmic cargo, including damaged organelles and aggregated proteins, and delivers the cargo to the lysosome or vacuole for destruction and recycling. Building an autophagosome requires the concerted action of around twenty autophagy-related genes, known as ATG genes, together with components of the cytoskeleton and the endomembrane systems. Because this machinery involves membrane deformation on a massive scale, its origin poses a striking evolutionary puzzle: how did a process requiring such elaborate coordination arise in the first eukaryotic cells?

The review argues that the autophagy pathway likely already existed in the last eukaryotic common ancestor, often abbreviated LECA, the hypothetical organism from which all living eukaryotes descend. This conclusion rests on the observation that core components of the autophagy machinery are found across the major eukaryotic supergroups, suggesting that the pathway was present before those groups diverged. Rather than being an invention of any single lineage, autophagy appears to be an ancestral eukaryotic innovation that was subsequently modified, expanded or trimmed as different branches of the eukaryotic tree pursued their own evolutionary paths.

Interestingly, the pathway itself does not exist in prokaryotes. Bacteria and archaea lack autophagosomes, lysosomes and the endomembrane systems on which eukaryotic autophagy depends. Yet when researchers search prokaryotic genomes for sequences resembling ATG genes, they find remote homologs of several of them. These distant relatives of eukaryotic autophagy genes hint that the raw materials from which the autophagy machinery was built were available long before eukaryotic cells existed, and that the pathway was assembled by repurposing pre-existing protein modules for new cellular functions.

Zhang and Mizushima organize their review around three major themes. The first concerns the evolutionary context surrounding the emergence of the autophagy pathway during eukaryogenesis, the series of events by which complex eukaryotic cells arose from prokaryotic ancestors. The second addresses how ATG genes diversified in different eukaryotic lineages through gene duplications, gene losses, and lineage-specific gains, producing the varied complements of autophagy genes observed in fungi, plants, animals and protists today. The third explores the functions that remote homologs of ATG genes carried out in prokaryotes and the phylogenetic and evolutionary relationships linking the eukaryotic and prokaryotic versions of these genes.

The eukaryogenesis story is central to understanding autophagy’s origin. The autophagosome depends on membranes, vesicle trafficking and cytoskeletal dynamics, all hallmarks of the eukaryotic cell plan. As the ancestors of eukaryotes acquired endomembrane compartments and an elaborate cytoskeleton, the molecular tools needed for autophagosome formation became available. The review suggests that the pathway’s emergence was tied to this broader cellular transformation, with ancestral protein families that carried out simpler membrane-related or stress-related functions in prokaryotes being recruited and elaborated into the coordinated autophagy machinery of early eukaryotes. In this view, autophagy is not an isolated invention but an integrated product of the cellular revolution that created the eukaryotic cell.

The diversification of ATG genes across eukaryotic lineages reveals how evolution both conserves and reshapes essential systems. Core autophagy genes have been retained across vast evolutionary distances, reflecting the fundamental importance of intracellular degradation to cell survival. At the same time, gene duplications have produced expanded families in some lineages, allowing functional specialization of related proteins, while gene losses have streamlined the machinery in others, and lineage-specific gains have introduced novel components not found elsewhere. These patterns mean that different organisms deploy partially distinct sets of ATG proteins to accomplish a fundamentally conserved process, and they complicate the task of identifying autophagy genes by sequence similarity alone, particularly in poorly studied protist groups.

The prokaryotic homologs of ATG genes add an intriguing dimension to this picture. Although bacteria and archaea do not perform macroautophagy, several of their proteins are recognizably related in structure or sequence to eukaryotic autophagy factors. Understanding what these remote homologs do in prokaryotic cells, whether they handle membrane remodeling, stress responses, protein trafficking or other tasks, offers clues about the ancestral functions from which autophagy components were drawn. Tracing the phylogenetic relationships between the eukaryotic and prokaryotic versions of these genes helps distinguish scenarios in which eukaryotes inherited the genes vertically from archaeal ancestors from those in which lateral gene transfer or other routes contributed to the modern complement of autophagy factors.

The implications of this evolutionary perspective extend well beyond comparative genomics. Autophagy is implicated in human health in contexts ranging from neurodegenerative disease to cancer and immunity, and much of what is known about the pathway comes from studies of yeast and mammalian cells. Recognizing which components are ancestral and universal, which are lineage-specific innovations, and which features of the machinery have been reshaped by evolution can guide researchers in choosing appropriate model systems and in interpreting similarities and differences between organisms. The review’s evolutionary framework provides a map for predicting where the pathway may vary and which of its elements represent the ancient core.

Zhang and Mizushima also look forward, outlining possible future directions for the field. These include filling in the distribution of autophagy genes across under-sampled eukaryotic lineages to better resolve when and how duplications and losses occurred, deepening the functional characterization of prokaryotic homologs to illuminate the pre-eukaryotic roles of ancestral proteins, and refining models of eukaryogenesis to place autophagy’s emergence within a firmer chronology. As genome data from diverse organisms continue to accumulate, the authors anticipate that the boundary between what is known about autophagy in model organisms and what can be inferred across the tree of life will continue to blur, yielding a more complete account of how one of the cell’s most sophisticated degradation systems came to be.

The review, which is open access, was supported by a Grant-in-Aid for Specially Promoted Research from the Japan Society for the Promotion of Science awarded to Noboru Mizushima. By gathering the threads of genomics, phylogenetics and cell biology into a single evolutionary narrative, the work underscores a theme that resonates throughout modern biology: even the most intricate molecular machines of complex cells are built from parts with far older histories, and reconstructing those histories is essential for understanding how the machinery works today.

Subject of Research: Evolutionary origins and diversification of autophagy-related genes across prokaryotes and eukaryotes

Article Title: Autophagy genes across the tree of life: from prokaryotic homologs to eukaryotic complexity

Article References: Zhang, S., & Mizushima, N. (2026). Autophagy genes across the tree of life: from prokaryotic homologs to eukaryotic complexity. Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-026-06423-7

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06423-7

Keywords: autophagy, ATG genes, macroautophagy, eukaryogenesis, last eukaryotic common ancestor, prokaryotic homologs, gene duplication, genome evolution, phylogenetics, autophagosomes, cellular degradation, evolution