A body–brain circuit that regulates body inflammatory responses

The study addresses how the brain, via the body–brain axis, monitors and regulates innate immune responses to maintain a balance between pro-inflammatory and anti-inflammatory states. Overactive pro-inflammation can lead to autoimmune and inflammatory diseases, underscoring the need to understand regulatory mechanisms. Although prior work has shown that infection activates neural circuits related to fever, malaise, and feeding changes, and that broad vagus nerve stimulation can attenuate inflammation, the specific neuronal circuit elements that link peripheral immune signaling to brain control remained unclear. The authors hypothesize that peripheral cytokines engage specific vagal afferents that signal to defined brainstem neurons in the caudal nucleus of the solitary tract (cNST), enabling the brain to modulate peripheral immunity homeostatically.

Prior studies have established extensive neural control of physiology, including organ function, metabolism, and internal state. Infection and sickness responses engage brain circuits (for example, preoptic and parabrachial circuits controlling fever and sickness behaviors). Foundational work by Tracey and colleagues demonstrated that electrical stimulation of the vagus nerve can blunt TNF-mediated inflammation and shock, highlighting neural modulation of immunity. Additional reviews and studies suggested multiple brain-immune communication pathways and cytokine-to-brain routes. However, the precise sensory neurons, central nodes, and logic by which the brain concurrently modulates both pro- and anti-inflammatory arms were largely unresolved. This study builds on that background by identifying discrete cytokine-responsive vagal populations and their brainstem targets.

• Immune activation and brain mapping: Mice received intraperitoneal lipopolysaccharide (LPS) or saline; circulating cytokines (IL-6, IL-1β, TNF, IL-10) were measured by ELISA over time. Neural activation was mapped by FOS immunostaining across brain regions, focusing on cNST and area postrema.
• Fibre photometry and necessity of vagus: Vglut2-cre mice received AAV9-Syn-Flex-GCaMP6s in cNST; bulk calcium signals were recorded in awake mice following LPS vs saline. Bilateral sub-diaphragmatic vagotomy assessed the necessity of vagal inputs for cNST activation by LPS.
• Genetic access to LPS-activated cNST neurons: TRAP2 mice were LPS-TRAPed (LPS then 4-OHT) to label LPS-activated cNST neurons. Co-labelling with FOS after a second LPS confirmed targeting specificity (>80% overlap).
• Chemogenetic manipulation of cNST TRAPed neurons: AAV9-DIO-hM4Di (inhibitory DREADD) or AAV9-DIO-hM3Dq (excitatory DREADD) was bilaterally injected into cNST of TRAPed mice. Clozapine-N-oxide (CNO) was administered prior to saline or LPS, and cytokines were quantified 2 h post-LPS.
• scRNA-seq of cNST and LPS-TRAPed neurons: Single-cell suspensions from cNST were processed with 10x Chromium to catalog neuronal clusters (glutamatergic and GABAergic). Separately, FACS-isolated tdTomato+ (TRAPed) and unlabeled neurons underwent single-cell RNA-seq; datasets were integrated (Seurat) to map TRAPed cells onto cNST clusters.
• Identification and manipulation of cNST effector neurons: Candidate marker Dbh was identified for key glutamatergic clusters. Dbh-cre mice received cNST AAV-DIO-hM3Dq for activation; cytokines were measured after LPS. Anti-DBH–saporin was used to ablate DBH cNST neurons, followed by LPS and cytokine assays.
• Vagal cytokine sensing: In vivo calcium imaging of nodose ganglia recorded responses to intraperitoneal or extraintestinal perfusion of cytokines (pro: IL-6, IL-1β, TNF; anti: IL-10), LPS, and nutrient controls (glucose or fat). Vglut2-cre labeled all vagal sensory neurons with GCaMP6s.
• Chemogenetic activation of vagal subpopulations: Using scRNA atlases, AAV-DIO-hM3Dq was bilaterally injected into nodose ganglia of Trpa1-cre and Calca-cre mice (and other Cre lines as controls). Following CNO and LPS, cytokines were assayed. TRPA1 neuron responses to IL-10 were imaged; TRPA1 neurons were ablated (Trpa1-cre; Rosa-DTR + diphtheria toxin) to test necessity for anti-inflammatory signaling.
• Circuit tracing: In Dbh-cre mice, cNST neurons were infected with AAV-Flex-TVA and AAV-Flex-G, followed by EnvA-pseudotyped G-deleted rabies (RABV-GFP) to map monosynaptic retrograde inputs from nodose neurons expressing TRPA1 or CALCA (RNAscope).
• Disease models and physiological outcomes: Lethal endotoxaemia model (12.5 mg kg−1 LPS) assessed survival with repeated CNO-driven activation of TRPA1 vagal or DBH cNST neurons. DSS-induced colitis model evaluated colon integrity, CXCL1 levels, and fecal occult blood with TRPA1 vagal activation. Salmonella Typhimurium infection tested consequences of sustained anti-inflammatory bias. Circulating corticosterone was measured to test HPA axis involvement.
• Statistics: Mann–Whitney U-test, Wilcoxon test, ANOVA, and log-rank tests were used; data are mean ± s.e.m. Sample sizes specified per experiment.

• Peripheral immune challenge engages a vagal-cNST axis:
  • LPS induced robust increases in circulating IL-6, IL-1β, TNF, and IL-10 peaking ~2 h post-injection.
  • Strong bilateral FOS induction in cNST and area postrema after LPS; minimal after saline. Myd88−/− mice lacked cNST FOS induction and showed blunted cytokine responses.
  • Fibre photometry revealed cNST activation tracking immune response; bilateral vagotomy reduced LPS-evoked cNST activity by ~80% (Wilcoxon saline vs LPS P=0.03; Mann–Whitney LPS vs vagotomy P=0.004).
• cNST neurons homeostatically regulate inflammation:
  • Chemogenetic inhibition (hM4Di) of LPS-TRAPed cNST neurons exacerbated inflammation: pro-inflammatory cytokines increased >300% (e.g., IL-1β ~200 to ~800 pg ml−1) and anti-inflammatory IL-10 decreased (~750 to ~250 pg ml−1). LPS group P-values: IL-6 P=0.004; IL-1β P=0.004; IL-10 P=0.002.
  • Chemogenetic activation (hM3Dq) of TRAPed cNST neurons suppressed pro-inflammatory cytokines by ~70% and increased IL-10 nearly tenfold (all LPS comparisons P=0.002). Activation without immune challenge did not alter cytokines.
• scRNA-seq identified key cNST populations:
  • cNST comprised 14 glutamatergic and 6 GABAergic clusters. LPS-TRAPed neurons mapped mainly to glutamatergic clusters 7, 10, 12 (and some cluster 2) and a GABAergic cluster 15.
  • Activating glutamatergic (Vglut2) cNST neurons suppressed inflammation (IL-6 P=0.002; TNF P=0.004; IL-10 P=0.02). Activating GABAergic (Vgat) cNST neurons had no effect.
  • DBH marked a key excitatory cNST subset. Activating DBH cNST neurons decreased IL-6, IL-1β, TNF and increased IL-10 (all LPS comparisons P=0.03). Ablating DBH neurons increased IL-6 and TNF and reduced IL-10 (P=0.02–0.04).
• Vagal sensory neurons encode cytokine classes:
  • Two non-overlapping nodose populations responded to pro-inflammatory vs anti-inflammatory cytokines. Among 423 imaged vagal neurons: 21 responded to pro-inflammatory (13 TNF, 8 IL-1β) and 11 responded to IL-10. LPS did not directly activate vagal neurons.
  • TRPA1-expressing vagal neurons responded selectively to IL-10 (27/189 TRPA1 neurons) and not to pro-inflammatory cytokines; IL-10-responsive TRPA1 neurons were distinct from fat-responsive TRPA1 neurons.
  • CALCA-expressing vagal neurons responded to pro-inflammatory cytokines (35/211 CALCA neurons).
• Vagal-to-brain circuit and functional impact:
  • Chemogenetic activation of TRPA1 vagal neurons dramatically suppressed pro-inflammatory cytokines (>80% decrease) and increased IL-10 ~6-fold (P<0.01 for IL-6, IL-1β, IL-10).
  • Activation of CALCA vagal neurons reduced IL-6 (P<0.01) and IL-1β (P=0.001) without changing IL-10 (P=0.88).
  • Pro-inflammatory “clamping” (maintaining high IL-6/IL-1β/TNF) did not prevent TRPA1-driven IL-10 enhancement, indicating independent amplification of anti-inflammatory signaling.
  • TRPA1 vagal neuron ablation reduced cNST activation by IL-10 and halved IL-10 levels after LPS; IL-6 and IL-1β were unaffected.
  • Monosynaptic retrograde rabies tracing showed direct nodose inputs from TRPA1 and CALCA neurons to DBH cNST neurons; activating TRPA1 vagal neurons induced FOS in ~40% of DBH cNST neurons.
• Therapeutic potential and physiological trade-offs:
  • Repeated activation of TRPA1 vagal or DBH cNST neurons rescued survival in lethal LPS sepsis (~90% survival vs 0% in controls; log-rank P<0.001 for both).
  • In DSS colitis, activating TRPA1 vagal neurons preserved distal colon integrity, reduced CXCL1 (P=0.03), and eliminated fecal occult blood (P=0.03).
  • Sustained activation during Salmonella infection increased bacterial burden by nearly 2 logs in spleen and lymph nodes (P=0.03) and caused severe weight loss, indicating that over-suppression of inflammation compromises host defense.
  • Circuit activation did not alter LPS-induced corticosterone levels, suggesting effects are independent of HPA axis changes.

The findings demonstrate a defined body–brain circuit in which peripheral cytokines engage distinct vagal sensory populations—TRPA1 neurons for anti-inflammatory IL-10 signals and CALCA neurons for pro-inflammatory cytokines—that converge onto DBH-expressing neurons in the cNST. This circuit functions as a homeostatic rheostat, exerting positive feedback to enhance anti-inflammatory responses and negative feedback to restrain pro-inflammation. Disrupting cNST activity causes runaway inflammation, while targeted activation rebalances cytokine profiles and improves outcomes in sepsis and colitis models. The absence of changes in corticosterone suggests modulation via dedicated brainstem-autonomic pathways rather than generalized stress responses. The delineation of specific afferent (TRPA1/CALCA nodose) and central (DBH cNST) elements provides mechanistic insight into neuroimmune regulation and supports therapeutic strategies that selectively engage this axis to treat hyperinflammatory states. However, the Salmonella results caution that excessive or prolonged anti-inflammatory drive can impair pathogen clearance, emphasizing the need for calibrated intervention.

This study identifies and functionally validates a vagal–brainstem (nodose–cNST) circuit that monitors peripheral cytokines and bidirectionally modulates innate immune responses. Pro- and anti-inflammatory cytokines activate separate vagal afferent populations (CALCA and TRPA1, respectively), which monosynaptically target DBH-expressing cNST neurons. Manipulating this circuit can suppress pro-inflammatory cytokines while boosting IL-10, rescuing survival in endotoxaemia and protecting against colitis. These discoveries open avenues for precise neuromodulatory or pharmacologic interventions to manage autoimmune and hyperinflammatory conditions. Future work should map downstream efferent pathways and effectors, define how TRPA1/CALCA inputs integrate within cNST microcircuits, identify additional participating neuronal populations and signals, and assess translational applicability and safety in chronic disease and infection contexts.

• The work is performed in mouse models; generalizability to humans remains to be established.
• Chemogenetic approaches (DREADDs/CNO) and viral targeting may have off-target effects; although controls were used, pharmacokinetics and specificity in clinical settings differ.
• Sample sizes are modest and experiments were not randomized or blinded (as noted), which may introduce bias.
• The study identifies key afferent and central nodes but does not fully delineate descending efferent pathways and peripheral effectors mediating immune modulation.
• Sustained activation experiments reveal trade-offs in infection control; optimal dosing/timing parameters for therapeutic modulation are not defined.