Stress can linger long after the triggering event is over. A person who has experienced a traumatic incident may continue to avoid certain places, certain situations, a persistent internal state that colors their behavior for days, weeks, or longer. Biologists have long understood that this phenomenon is not unique to humans, but its underlying mechanisms have remained frustratingly opaque.
A new study published in PNAS by researchers at the Hong Kong University of Science and Technology (HKUST), Tohoku University, and the Tokyo Metropolitan Institute of Medical Science has identified the neural and molecular machinery that produces precisely this kind of persistent, stress-induced internal state, in fruit flies.
The discovery is a significant step toward understanding the fundamental biology of anxiety and phobic states.
What the flies showed
The research team, led by Yukinori Hirano of HKUST, designed a simple but clever behavioral assay. Flies were placed in a two-arm maze: one arm 4 millimeters wide, the other just 2 millimeters wide, both only slightly wider than a male fly’s body. Naive flies explored both arms equally. But flies that had received a 5-minute pulsed electric shock (60 volts, 1.5 seconds on, 3.5 seconds off) before entering the maze showed a marked avoidance of the narrow arm, spending significantly more time in the wider space.
The researchers call this “claustrophobia-like behavior” (CLB), and it is remarkably persistent: a single 5-minute shock produced avoidance that lasted up to 7 days. Individual flies showed consistent levels of avoidance over time, suggesting a stable change in internal state rather than simple habituation.
Importantly, the behavior is not classical associative memory. Flies with mutations that block the formation of aversive memories, the dumb² mutation, which disables dopamine receptor function in the mushroom body, the insect brain’s center for associative learning, showed normal CLB. This dissociation is a central finding of the paper: generalised, persistent stress-induced states are biologically distinct from cue-specific fear memories.
The effect is also not specific to electric shock. Heat shock (40 degrees Celsius, or 104 degrees Fahrenheit, for 10 minutes) produced the same behavior. But other stressors — vibration and restraint — did not, suggesting that the internal state mechanism is selectively activated by certain types of intense stimuli.
Two pathways, one internal state
The researchers traced the phenomenon to two independent molecular pathways, each sufficient on its own to produce CLB.
The first involves a neuropeptide called allatostatin-A (AstA), the fruit fly equivalent of the mammalian neuropeptide galanin, which has been implicated in anxiety and stress responses. Using genetic tools, the team identified a single pair of AstA-producing neurons in the subesophageal zone of the fly brain that became active during electric shock. Silencing these neurons, either permanently or just during the shock period, prevented the development of CLB. Knocking down the AstA receptor (AstA-R1) after the shock also suppressed the behavior, demonstrating that the receptor is required not just for the induction of the internal state but for its maintenance over time.
The second pathway is surprising: it operates through immune signaling in the blood-brain barrier. Transcriptomic profiling of individual fly heads revealed that genes involved in the Toll innate immune pathway — the fly equivalent of mammalian TLR2/4 signaling — were upregulated in perineurial glia, the outermost layer of the insect’s blood-brain barrier. Genetically activating Toll signaling in these barrier cells was sufficient to trigger CLB even without any stress exposure. This suggests that the blood-brain barrier is not merely a passive filter but an active participant in shaping behavioral state.
“We identified at least two independent molecular routes that can generate the same phobia-like internal state,” Hirano said. “This may help explain why anxiety disorders are so heterogeneous and why different patients respond to different treatments.”
Implications for anxiety research
The findings resonate with a growing body of literature linking blood-brain barrier dysfunction and neuroinflammation to stress, anxiety, and post-traumatic stress disorder in humans. The AstA-galanin connection is particularly intriguing: galanin has been studied as a potential target for anxiolytic drugs, and the current study’s finding that the receptor is required for state maintenance — not just induction — suggests that blocking galanin signaling after a traumatic event could potentially prevent the development of persistent phobic states.
“There is a conceptual gap between fear memories and anxiety states,” Hirano said. “Our work provides a biological basis for that distinction.”
Several caveats apply. The “claustrophobia” observed in flies is an analogy — the insects lack the amygdala, claustrum, and other brain structures associated with human anxiety. The behavioral readout is a simple preference for a wider arm, and whether flies experience anything akin to human subjective fear cannot be determined. Only male flies were tested, leaving sex differences unexplored. And the translational gap from fly glia to the mammalian blood-brain barrier — which involves different cell types (endothelial cells, pericytes, and astrocytic endfeet) — is significant.
Nonetheless, the study provides a molecularly detailed account of how a brief stressor can create a lasting, generalized change in behavior, and it opens the door to new strategies for understanding and treating anxiety. The researchers have deposited their transcriptomic data in the Gene Expression Omnibus (accession GSE294159) and made all fly lines available for further study.
Source: Alia, A.G., Hu, X., Gu, Y. et al. “Neuropeptide signaling and the blood–brain barrier generate a persistent stress-induced internal state in Drosophila.” PNAS 123(27), e2517987123 (2026). DOI: 10.1073/pnas.2517987123