Hyperactive neuronal autophagy depletes BDNF and impairs adult hippocampal neurogenesis in a corticosterone-induced mouse model of depression

Stress hormones eat your brain's growth factor. Not metaphorically — literally, through a cellular recycling process that destroys the very protein keeping new neurons alive. A team led by Kuo Zhang and colleagues traced the full chain of causation, step by step: from a stress hormone to a molecular shredder, to a missing growth factor, to a brain that can no longer grow new cells, and finally, to depression. Here's why the hippocampus is the right place to look. Patients with depression consistently show elevated cortisol, the human equivalent of the stress hormone corticosterone. The hippocampus — specifically a region called the dentate gyrus — is one of the only places in the adult brain where new neurons are born continuously. That process, adult hippocampal neurogenesis, is tightly linked to mood and stress reactivity. Antidepressants and exercise both boost it. Block it experimentally, and some antidepressant effects vanish entirely. The fuel for that process is a protein called brain-derived neurotrophic factor, or BDNF, which drives proliferation, survival, and synapse formation in newborn neurons. Research has long shown that chronically elevated corticosterone reduces hippocampal BDNF. The question Zhang's team set out to answer was: by what mechanism, exactly? To model chronic stress, they gave adult male mice corticosterone dissolved in their drinking water at a concentration of 0.1 milligrams per milliliter for eight weeks. Then they ran a battery of behavioral tests. In the tail suspension test and the forced swim test — standard measures of behavioral despair — corticosterone-treated mice showed significantly more immobility than controls, with p-values of 0.003 and 0.0002 respectively. In the elevated plus maze and open field test, they spent less time in exposed areas, a signature of anxiety. In a novel object recognition task, their preference index dropped significantly. Crucially, the animals moved the same total distance as controls, ruling out simple fatigue or motor impairment as an explanation. The cellular picture matched. Using BrdU labeling, a technique that tags dividing cells so you can track their fate, the team found that newborn cell survival in the dentate gyrus was dramatically reduced in corticosterone-treated mice, with p-values below 0.0001. The cell-cycle kinetics were disrupted too: fewer cells were exiting the cycle normally, and more were abnormally re-entering it. The migration of newborn mature neurons into the granule cell layer was also impaired. By this point in the study, the behavioral and cellular profiles of depression were both clearly present, and both pointed at the dentate gyrus. The next question was what was actually going wrong inside the neurons. The answer turned out to be autophagy. Autophagy, from the Greek for self-eating, is the cell's internal recycling system. It wraps unwanted or damaged material in a membrane, fuses that package with a lysosome, and digests the contents. Under normal conditions, autophagy is essential. But in corticosterone-treated mice, it was running far too hot. Zhang's team found elevated LC3-II protein in the dentate gyrus, a marker of autophagosome formation, along with reduced levels of p62, which accumulates when autophagy is inhibited and falls when autophagy is active. The core autophagy regulator ATG5 was also upregulated by corticosterone. Electron microscopy and a fluorescent reporter system using a tandem mCherry-GFP-LC3 construct, which distinguishes autophagosomes from autolysosomes by color, confirmed the finding at the ultrastructural level: both autophagosomes and autolysosomes increased significantly, and autophagic flux was enhanced. Then comes the connection that makes this more than a story about recycling machinery. The number of LAMP2-positive lysosomal puncta inside BDNF-producing neurons — lysosomes actively sitting inside the very cells that make BDNF — was significantly elevated. BDNF protein in the dentate gyrus was substantially reduced. But here's the telling detail: BDNF messenger RNA did not change. The gene was being transcribed. The protein was being made. It was being destroyed after synthesis. The hyperactive autophagy was consuming BDNF protein before it could do anything useful — the neuron was eating its own fuel supply. The causal chain is now complete. Corticosterone upregulates ATG5, ATG5 drives hyperactive autophagy, hyperactive autophagy degrades neuronal BDNF through lysosomal processing, and without BDNF, neurogenesis collapses. That's the mechanism. And it's testable — because if the chain is real, breaking one link should reverse the whole cascade. That's exactly what the rescue experiment did. Zhang and colleagues used stereotactic injections of an adeno-associated virus, an AAV vector, designed to knock down the Atg5 gene selectively in dentate gyrus neurons. The construct was driven by the human Synapsin promoter, which is active only in neurons, so this was not a blunt systemic intervention. It was targeted to a specific cell type in a specific brain region. Control mice received a scrambled version of the same construct. The biochemical results came first. LC3-II fell and p62 rose in the knockdown mice — autophagy was being dampened. Autophagosomes and autolysosomes both decreased. LAMP2 puncta inside BDNF-positive neurons dropped sharply. Most strikingly, neuronal BDNF protein recovered strongly. The number of BDNF-positive cells increased significantly, with p-values below 0.0001. Again, BDNF messenger RNA was unchanged between the two groups. The rescue was happening at the protein level, not the transcriptional level. The neurogenesis results followed. Neural stem cells, identified by the markers GFAP and SOX2, increased after Atg5 knockdown, as did the proliferating subset that also expresses MCM2. Doublecortin-positive neuroblasts recovered. Newborn neurons, tracked by BrdU co-labeling, increased significantly across both immature and mature stages. The dentate gyrus was growing new cells again. And the behavior reversed. Immobility in the tail suspension test fell significantly in Atg5 knockdown mice compared to corticosterone controls, with a p-value of 0.04. The forced swim test showed the same pattern with a p-value of 0.02. The novel object recognition preference index improved dramatically, with a p-value below 0.0001. Three independent behavioral readouts, all pointing in the same direction. The mechanism that caused the depression could be reversed by silencing the one molecular gear that drove it. What does this actually open up? The neurogenesis hypothesis of depression — the idea that lost or impaired neurogenesis underlies depressive states — has circulated for decades. What it has lacked is a clean molecular mechanism connecting the stress axis to the death of newborn neurons. Zhang and colleagues have now supplied one: ATG5-driven neuronal autophagy in the dentate gyrus is the link between chronically elevated corticosterone and the BDNF depletion that starves neurogenesis. That makes ATG5, and neuronal autophagy more broadly, a named and targetable node in the depression circuit. The authors honestly acknowledge what remains unresolved. The upstream question — exactly how corticosterone activates ATG5, and whether the glucocorticoid receptor is the intermediary — is not yet answered. And the distance from a mouse model of corticosterone-induced depression to human clinical depression is real. Whether this pathway operates in patients is unknown. But what the study has done is convert a vague hypothesis into a mechanism with a name, a molecular address, and a point of intervention: neuronal autophagy in the dentate gyrus, driven by ATG5, depleting BDNF after it's made, not before, and reversible by targeted gene silencing. That's not a hypothesis anymore — that's a mechanism. And mechanisms, unlike hypotheses, can be drugged. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.