Hepatocellular carcinoma, the most common form of primary liver cancer, is a disease in which psychological state and clinical outcome have repeatedly been correlated but rarely connected mechanistically. Patients with liver cancer experience depression at strikingly high rates, and depression, in turn, has been associated with poorer survival in multiple cancer types. What has been missing is a concrete anatomical and molecular route by which a mental state could change the biology of a tumor. The new work addresses that gap by combining circuit neuroscience, single-cell transcriptomics and tumor biology in a single experimental framework, asking not just whether stress matters for cancer, but which neurons carry the signal and which molecules deliver it.

The team began with a broad observation: in male mice, chronic stress, systemic inflammation and neuropathic pain were all associated with faster hepatocellular carcinoma progression. Rather than treating stress as a diffuse hormonal phenomenon, the researchers focused on a precisely defined pathway. Chronic stress, they found, hyperactivates glutamatergic neurons in the paraventricular nucleus of the hypothalamus, a master hub of the stress response. These neurons project to the dorsal motor nucleus of the vagus, where they drive cholinergic vagal motor neurons whose fibers reach the liver. The result of this hypothalamus-to-vagus-to-liver relay is a measurable increase in acetylcholine inside the tumor microenvironment, a neurotransmitter arriving not from the bloodstream but along a dedicated neural line from the brain.

To find out what that acetylcholine actually does inside the tumor, the researchers turned to single-cell transcriptomics, profiling the gene-expression identities of individual cells within hepatocellular carcinomas. Among the acetylcholine-responsive population, one cell type stood out: epithelial cells carrying the receptor subunit CHRNA9, also known as the nicotinic acetylcholine receptor subunit alpha-9. These Chrna9-positive epithelial cells were not acting alone. The single-cell data revealed that they communicate with tumor-associated macrophages through a signaling axis involving the collagen proteins COL1A1 and COL1A2, the adhesion molecule CD44, and the transcription factor STAT3. That interaction prompts the macrophage compartment to produce CXCL1, a chemokine best known for recruiting neutrophils and amplifying inflammatory signaling.

The chemokine turned out to be a pivotal node in the pathology. CXCL1 did more than shape the immune landscape of the tumor; acetylcholine also acted directly on tumor cells through CHRNA9 to promote their proliferation, giving the neurotransmitter both a direct growth-promoting role and an indirect inflammatory one. More surprising still was where some of the CXCL1 ended up. The team found that circulating CXCL1 can cross the blood-brain barrier and activate glutamatergic neurons in the hypothalamus, the very region where the circuit begins. In the animals, this liver-to-brain leg of the loop manifested as exacerbated depression-like behaviors, closing a feedback circle in which stress promotes tumor growth, the tumor broadcasts inflammatory signals back to the brain, and those signals deepen the depressive state that started the cycle.

Perhaps the most clinically provocative part of the study is what happened when the researchers interrupted the loop at different points. Silencing the hypothalamic vagal circuit restrained tumor progression. So did knocking down Chrna9 specifically in tumor cells, and so did systemic neutralization of CXCL1 with blocking antibodies. Crucially, these interventions did not merely slow the cancer; they also alleviated the depression-like behaviors in the mice. A single molecular target, in other words, sat at the intersection of two seemingly separate diseases, suggesting that the comorbidity of liver cancer and depression may not be a coincidence of suffering but a shared physiology with shared points of attack.

The technical achievement underlying these conclusions is considerable. Mapping a hypothalamic circuit requires the ability to identify, manipulate and monitor specific neuronal populations, and the study’s use of circuit-tracing and silencing approaches allowed the authors to establish causality rather than mere correlation. The single-cell transcriptomic analysis then bridged scales, moving from the whole-animal level of behavior and tumor volume down to the molecular conversations between individual epithelial cells and macrophages. By demonstrating that the COL1A1/2-CD44-STAT3 axis mediates the epithelial-macrophage crosstalk, the researchers provided a mechanistic handle on how a neural signal, acetylcholine, is transduced into an inflammatory output, CXCL1, that reshapes both tumor and brain.

Several caveats deserve emphasis. All of the findings come from male mice, and the authors are explicit about this sex limitation in the title of the paper itself; whether the same hypothalamic vagal circuit operates in female animals, in which both stress physiology and liver cancer biology can differ, remains an open question. Mouse models of hepatocellular carcinoma, moreover, do not fully recapitulate the human disease, which typically arises on a backdrop of cirrhosis, viral hepatitis or metabolic dysfunction. The link between chronic stress and tumor progression in humans is supported by epidemiological associations but has never been demonstrated through a defined neural circuit, and translating circuit-level interventions from mice to patients would face formidable technical and ethical hurdles. The blood-brain barrier crossing of CXCL1, while demonstrated in this model, would also need careful validation in human cancer patients with depression.

Even with those limitations, the conceptual implications are substantial. The study reframes the relationship between mental state and cancer as a bidirectional physiological loop rather than a one-way influence. Stress activates a brain-to-liver neural pathway; the tumor responds by growing and by releasing inflammatory mediators; those mediators return to the brain and deepen the behavioral state that drove the circuit in the first place. Each leg of the loop offers a potential intervention point, from neuromodulation of hypothalamic output to receptor-targeted drugs acting on CHRNA9 to cytokine-neutralizing therapies aimed at CXCL1. The finding that blocking any of these nodes improved both tumor outcomes and mood-related behaviors hints at a future in which oncology and psychiatry share therapeutic targets.

For the millions of people living with hepatocellular carcinoma, and for the large fraction of them who also experience depression, the work offers something more than a molecular diagram: it offers a rationale. If chronic stress genuinely accelerates liver cancer through a named circuit and named molecules, then managing stress, inflammation and neural signaling ceases to be a matter of comfort and becomes a matter of tumor biology. The Guangzhou team’s circuit, running from the paraventricular hypothalamus through the vagus into the tumor and back again, is likely to inspire a wave of studies asking whether similar brain-to-organ loops govern other cancers, and whether the liver-brain axis they uncovered can be safely pharmacologically severed in patients. The feedback loop between despair and disease, this research suggests, may be far more literal than medicine has assumed.

Subject of Research: Neural and immune mechanisms linking chronic stress to hepatocellular carcinoma progression and depression-like behavior in male mice

Article Title: Chronic stress activates a hypothalamic vagal circuit to promote hepatocellular carcinoma progression in male mice

Article References: Zhang, K., Zhang, J., Chi, D., He, Z., Wang, S., Zheng, X., Guo, R., Zeng, W., Ouyang, H., Ou, C., & Bai, X. (2026). Chronic stress activates a hypothalamic vagal circuit to promote hepatocellular carcinoma progression in male mice. Nature Communications. https://doi.org/10.1038/s41467-026-77923-0

Image Credits: AI Generated

DOI: 10.1038/s41467-026-77923-0

Keywords: chronic stress, hepatocellular carcinoma, hypothalamus, vagus nerve, acetylcholine, CHRNA9, CXCL1, liver-brain axis, depression, tumor microenvironment, macrophages, single-cell transcriptomics

The research, led by Lubin Gou and Junqiang Lei of the First Hospital of Lanzhou University and Lanzhou University’s First Clinical Medical College, focused on patients with liver cirrhosis without overt hepatic encephalopathy, a stage abbreviated LC-nOHE. This is the window in which the disease has already compromised liver function but has not yet produced the confusion, disorientation, and personality changes that define overt hepatic encephalopathy. Understanding what happens in the brain during this silent phase is critical, because it may reveal the earliest opportunities for intervention before irreversible damage takes hold.

The team recruited 86 patients with liver cirrhosis but no overt encephalopathy, along with 62 healthy controls, and acquired high-resolution three-dimensional T1-weighted magnetic resonance images of every participant’s brain. From these scans, the researchers extracted the volume of the thalamus and three distinct measures of cortical shape: cortical thickness, sulcal depth, and fractal dimension, a measure of the complexity of the brain’s folded surface. Rather than examining these features in isolation, the investigators constructed structural covariance networks, mathematical maps in which brain regions are connected if their anatomical features covary across individuals. Such networks are widely used as proxies for coordinated maturational and degenerative processes, offering a window into how brain regions change together as a system.

The results were striking at multiple levels. Compared with healthy controls, patients with cirrhosis showed enlargement of the thalamus alongside a broad range of cortical morphological abnormalities. At the level of whole-network organization, the thalamo-cortical structural covariance networks of patients displayed reduced segregation, meaning the brain’s specialized modules were less clearly differentiated, and decreased integration, meaning efficient communication across the network was impaired. These two properties, segregation and integration, are hallmarks of a healthy, well-organized brain, and their simultaneous deterioration suggests a fundamental disruption of the architecture that supports cognition.

Zooming in on individual network nodes, the researchers found that the centrality of three regions was significantly reduced in patients: the thalamus itself, the supramarginal gyrus, and the insula. Each of these regions plays a recognizable role in the syndrome. The thalamus relays sensory and motor signals to the cortex and regulates consciousness and alertness; the supramarginal gyrus contributes to language and spatial cognition; and the insula supports interoception and awareness of the body’s internal state. Reduced centrality in these hubs indicates that they had lost influence within the network, becoming less connected to the rest of the brain’s structural architecture.

Perhaps the most clinically significant finding was the relationship between brain changes and liver function. The degree centrality of the thalamus was negatively correlated with liver function measures, meaning that the worse the liver was performing, the more the thalamus had lost its position within the brain’s structural network. This correlation ties the brain imaging directly to the severity of liver disease and supports the idea that the thalamus sits at the front line of the liver-brain axis, the pathway by which hepatic dysfunction, circulating toxins such as ammonia, and systemic inflammation are translated into neural injury.

To move beyond correlation and probe causality, the team applied a technique called causal analysis of structural covariance networks, or CaSCN. This approach examines how changes in one brain region’s morphology relate to changes in others across the progression of disease, allowing researchers to ask which region leads and which follows. The analyses demonstrated that thalamus volume had causal effects on the alterations of cortical morphology as liver dysfunction progressed. In other words, the data are consistent with a model in which the thalamus is not merely another victim of cirrhosis but an active driver, propagating structural changes outward to the cortical regions with which it is connected.

The study then took an unusual additional step: linking the brain imaging to genomics through imaging transcriptomics. Using normative gene-expression profiles from the Allen Human Brain Atlas, the researchers evaluated whether the spatial pattern of causal effects across the cortex overlapped with the spatial distribution of specific genes. It did. The pattern of causal path coefficients was spatially correlated with the expression of particular genes in the normative atlas, suggesting that the cortical regions most vulnerable to thalamus-driven change are also those with distinctive molecular signatures. Gene ontology analyses pointed toward enrichment in biological processes, molecular functions, and cellular components that may help explain why some cortical areas are preferentially affected while others are relatively spared.

Taken together, the findings provide a comprehensive, multilevel view of how the thalamo-cortical circuit becomes progressively vulnerable in liver cirrhosis before overt encephalopathy appears. The enlargement of the thalamus, consistent with processes such as edema or glial changes reported in prior literature on hepatic encephalopathy, appears to initiate a cascade that erodes the segregation and integration of the entire structural network, strips key hubs of their centrality, and reshapes the cortex in patterns governed partly by underlying gene expression. Because the thalamus’s network position tracks liver function, measures of thalamo-cortical network integrity could potentially serve as imaging biomarkers for identifying patients at risk of progressing to overt hepatic encephalopathy, enabling earlier monitoring and treatment.

The authors emphasize that these findings offer a potential mechanism-driven framework for understanding the earliest brain consequences of liver disease, one that connects organ function, network neuroscience, and transcriptomics within a single analytical pipeline. The work was supported by the Lanzhou University First Affiliated Hospital Foundation and the Science and Technology Department of Gansu Province, and it was approved by the Institutional Ethics Committee of the First Hospital of Lanzhou University. As cirrhosis continues to affect millions worldwide, studies like this one bring clinicians closer to detecting, and perhaps preventing, the neurological toll of liver disease before it announces itself in the clinic.

Subject of Research: Structural covariance network alterations of the thalamo-cortical circuit in liver cirrhosis patients without overt hepatic encephalopathy

Article Title: Multilevel structural covariance network alterations of thalamo-cortical circuit in liver cirrhosis patients without overt hepatic encephalopathy: associations with liver function and imaging transcriptomics

Article References: Gou, L., Ren, H., Xu, W., Gao, Y., Wang, S., Zhang, Y., Dou, Y., & Lei, J. (2026). Multilevel structural covariance network alterations of thalamo-cortical circuit in liver cirrhosis patients without overt hepatic encephalopathy: associations with liver function and imaging transcriptomics. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02790-6

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02790-6

Keywords: liver cirrhosis, hepatic encephalopathy, thalamus, thalamo-cortical circuit, structural covariance network, MRI, liver-brain axis, imaging transcriptomics, cortical thickness, liver function, neuroimaging, Allen Human Brain Atlas