The research, led by Haipeng Lu, Mingqiao Ying and Chuanchen Hu of the Department of Neurology at Affiliated Jinhua Hospital, Zhejiang University School of Medicine, set out to identify which inflammatory mediators drive barrier breakdown after ischemic stroke. Their attention settled on GPR84, a G protein-coupled receptor long associated with immune activation but poorly understood in the context of cerebral blood vessels. The team hypothesized that this receptor, when switched on during the inflammatory aftermath of ischemia, might not merely accompany barrier failure but actively cause it.

To test the idea, the researchers first looked at what happens to GPR84 in the injured brain. In a mouse model of experimental cerebral ischemia, they found a marked increase in GPR84 expression in the cortex, the outer layer of the brain where stroke damage is often most severe. The pattern was not limited to animals. When the team exposed human cerebral microvascular endothelial cells, the specialized cells that form the inner wall of brain capillaries, to oxygen-glucose deprivation followed by reoxygenation, a laboratory procedure that mimics the metabolic stress of stroke and reperfusion, those cells also ramped up GPR84 production. The receptor, in other words, is switched on by ischemic stress in both living brains and in the human vessel cells that build the barrier itself.

The decisive experiment came from genetics. The researchers studied mice engineered to lack a functional Gpr84 gene entirely. When these animals were subjected to ischemic insult, the difference was striking. Compared with normal mice, the GPR84-deficient animals developed smaller areas of tissue infarction, meaning less brain tissue died outright. Their brains swelled less, a critical measure because edema inside the rigid skull drives much of the acute damage after stroke. On standardized neurological assessments, which test motor coordination, reflexes and overall function, the knockout mice performed significantly better. Most importantly for the barrier hypothesis, vascular leakage, the pathological escape of blood components into brain tissue, was significantly reduced.

Underlying these functional improvements, the team observed preservation of the barrier’s molecular architecture. Two structural proteins, Claudin-5 and ZO-1, are essential components of tight junctions, the zipper-like seals that knit adjacent endothelial cells together and make the blood-brain barrier impermeable. In normal mice after stroke, these proteins degrade and disappear from the junctions, opening gaps in the wall. In the GPR84-deficient mice, Claudin-5 and ZO-1 remained intact at the neurovascular unit, the anatomical complex of vessel cells, supporting cells and neural tissue that together maintain the barrier. The genetic deletion had effectively protected the physical seals of the brain’s vascular lining.

To confirm that GPR84 acts directly on endothelial cells rather than through some indirect immune mechanism, the researchers turned to cell culture. They applied selective GPR84 agonists, molecules that artificially activate the receptor, to monolayers of endothelial cells grown in the laboratory. The result was a recapitulation of barrier failure: the cell layers, which normally form tight, impermeable sheets, became leaky. This experiment demonstrated that switching on GPR84 is sufficient by itself to break down endothelial barrier function, without any contribution from immune cells or other brain cell types.

The team then probed the signaling pathway downstream of the receptor, and this is where the study delivers its most mechanistically interesting finding. GPR84 activation disrupted the cAMP pathway, a ubiquitous intracellular signaling system in which the messenger molecule cyclic AMP regulates countless cellular processes. When the researchers co-administered a cell-permeable cAMP analogue, a synthetic version of the messenger that can enter cells and restore signaling, the barrier failure caused by GPR84 activation was abrogated. The protective effect of restoring cAMP signaling pointed to a specific cascade: GPR84 suppresses cAMP, and the loss of cAMP drives the breakdown of tight junctions.

Further experiments refined the pathway. When the researchers silenced GPR84 expression in endothelial cells, the cells became resistant to barrier dysfunction under ischemic stress. Critically, this resistance was linked to maintained phosphorylation of CREB, a transcription factor that acts downstream of cAMP and regulates the expression of numerous genes, including those involved in maintaining tight junctions. In cells with active GPR84, CREB phosphorylation drops, and the transcriptional programs that keep tight junction proteins in place presumably falter. In cells where GPR84 is silenced, CREB phosphorylation is preserved, and the structural integrity of the barrier holds. The authors describe this as a cell-autonomous GPR84-cAMP-CREB signaling cascade that directly impairs brain endothelial barrier function, meaning the entire pathway operates within the endothelial cell itself rather than requiring signals from neighboring cells.

The authors are careful to note an important limitation. While the cell culture experiments establish that GPR84 on endothelial cells is sufficient to cause barrier failure, the integrated contributions of GPR84 expressed on immune cells within the neurovascular unit in vivo remain to be defined. In the living brain, GPR84 is also present on microglia, macrophages and other immune cells that swarm into the injured area after stroke, and these cells may contribute to barrier disruption through additional mechanisms that the current study does not fully disentangle. The endothelial-autonomous pathway is clearly established, but the full in vivo picture likely involves multiple cell types acting in concert.

Nevertheless, the findings position GPR84 as a candidate mechanistic target for barrier preservation strategies after ischemic stroke. Current stroke therapies remain limited, with thrombolysis and mechanical thrombectomy addressing the clot but doing little to prevent the secondary inflammatory damage that unfolds over hours and days. If pharmacological inhibition of GPR84, or interventions that restore cAMP-CREB signaling in endothelial cells, can be developed into viable treatments, they could offer a way to keep the blood-brain barrier intact during the critical window after reperfusion, potentially reducing edema, limiting hemorrhagic transformation and improving neurological outcomes. The study, supported by Affiliated Jinhua Hospital, Zhejiang University School of Medicine, and published open access, provides a detailed molecular map of one pathway by which stroke damages the brain’s most important protective boundary, and a clear starting point for therapeutic exploration.

Subject of Research: The role of the inflammatory receptor GPR84 in blood-brain barrier disruption after ischemic stroke

Article Title: GPR84 exacerbates blood-brain barrier disruption after ischemic stroke via cAMP/CREB-dependent tight junction suppression

Article References: Lu, H., Ying, M., & Hu, C. (2026). GPR84 exacerbates blood-brain barrier disruption after ischemic stroke via cAMP/CREB-dependent tight junction suppression. Translational Psychiatry. https://doi.org/10.1038/s41398-026-04378-8

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04378-8

Keywords: GPR84, blood-brain barrier, ischemic stroke, tight junctions, Claudin-5, ZO-1, cAMP, CREB, endothelial cells, cerebral ischemia, neurovascular unit, Translational Psychiatry

GPR84 belongs to a subgroup of GPCRs that respond to medium-chain fatty acids, the metabolic fragments released when fats are broken down. Because these fragments accumulate in tissues under metabolic stress, GPR84 is thought to act as a metabolic sensor for the immune system, translating changing lipid levels into altered cellular behavior. Decades of pharmacological studies have linked the receptor to macrophages, neutrophils and other innate immune cells, where its activation has been associated with amplified production of inflammatory signaling molecules. That association made GPR84 an attractive drug target for companies pursuing anti-inflammatory therapies, and it also made the receptor a cautionary tale: several clinical candidates targeting it were halted, in some cases because the biology proved more complicated than early animal studies suggested.

The central problem has been that GPCRs are not simple on-off switches. They are dynamic molecular machines that sample a range of conformations, and different ligands can stabilize different subsets of those conformations. A ligand that locks the receptor into a fully active shape will recruit signaling proteins robustly, while a ligand that favors inactive shapes will silence the pathway. But many ligands do something subtler: they stabilize partially active conformations, or they favor active shapes in one signaling branch while leaving others untouched. Without structural information, medicinal chemists were essentially adjusting molecular shapes blindly, hoping that small changes in a ligand scaffold would produce predictable changes in receptor behavior.

The new study addresses that gap by capturing GPR84 in two distinct functional states, allowing the researchers to compare the receptor’s architecture when it is being activated against its architecture when it is being blocked or modulated. Structures determined in multiple states are technically demanding, because a membrane-embedded receptor must be stabilized in each conformation long enough to be imaged at atomic resolution. Achieving this typically requires engineered variants, stabilizing antibodies or nanobodies, and carefully chosen ligands that preferentially hold the receptor in the desired state. The resulting paired structures function like two frames of a molecular movie, revealing which helices shift, which side chains rotate, and which structural water molecules rearrange as the receptor transitions between resting and signaling-competent forms.

Comparing the two states highlights the allosteric heart of the receptor. In GPCRs, ligand binding at a pocket nestled among the transmembrane helices is transmitted through a conserved relay of hydrogen bonds, salt bridges and hydrophobic contacts to the intracellular face, where G proteins and other effectors dock. The dual-state GPR84 structures delineate how agonist binding contracts this relay into the canonical active arrangement, with an inward movement of the extracellular portion of a key transmembrane helix and a corresponding outward swing on the intracellular side that opens the effector-binding cavity. In the inactive or antagonist-bound state, that same helix relaxes outward, the intracellular cavity collapses, and the network of polar interactions reconfigures into a pattern incompatible with effector coupling. The residue-by-residue map of these changes gives chemists concrete positions to target when they want to bias the equilibrium toward one state or the other.

Crucially, the structural work is not left to stand alone. The authors couple it with functional assays performed in immune cells, measuring how well each ligand drives or suppresses downstream signaling and, importantly, how it affects inflammatory outputs such as cytokine release. This combination is what elevates the study from structural description to pharmacological instruction. A structure can suggest that a compound should be an agonist, but only cellular assays reveal the potency, the efficacy ceiling, and whether the compound behaves as a full agonist, a partial agonist or a biased ligand in a physiologically relevant context. By testing ligands across the spectrum, the researchers could correlate specific structural features of the binding pocket with specific functional consequences, effectively building a translation table between atomic geometry and immune cell behavior.

One of the more interesting implications concerns ligand bias, the phenomenon in which a receptor signals preferentially through one intracellular pathway over another. For GPR84, whose activation intersects with inflammatory programs in macrophages, a biased ligand could in principle dampen harmful signaling while preserving useful functions, or vice versa. The dual-state structures provide a mechanistic handle on bias: ligands that engage only part of the activation relay, or that fail to complete certain helical movements, may recruit one effector but not another. The immune functional assays then quantify what that partial engagement means for cytokine production, giving drug developers a rational framework for designing compounds with tailored signaling profiles rather than accepting whatever profile a scaffold happens to produce.

The study also speaks to the thorny issue of species differences, which has plagued GPR84 drug development. Synthetic agonists developed against the human receptor have often shown sharply different potency in mouse models, complicating the interpretation of preclinical efficacy studies and contributing to the sector’s clinical disappointments. Although the structures presented are of the human receptor, the detailed pocket architecture allows researchers to pinpoint which residues differ across species and to predict, before expensive animal work, whether a given compound is likely to translate. That kind of structural triage could save years of effort and redirect resources toward candidate molecules with a realistic chance of reproducing human biology in animal systems.

Beyond immediate drug design, the work contributes to a broader shift in GPCR pharmacology toward state-based thinking. Rather than classifying ligands simply as agonists, antagonists or inverse agonists, the field increasingly describes them by the conformational ensembles they stabilize and the signaling outputs those ensembles produce. GPR84, with its metabolic ligands, its immune cell expression and its checkered clinical history, is an ideal test case for this framework. The demonstration that paired structures plus immune assays can jointly explain and predict pharmacological behavior provides a template that other lipid-sensing and metabolite-sensing receptors could follow, particularly those where immune modulation is the therapeutic goal.

For patients, the significance lies in what this molecular clarity could eventually enable. Chronic inflammatory and metabolic diseases, including conditions involving macrophage-driven tissue damage, remain areas of substantial unmet need, and receptors that sense the metabolic environment of inflamed tissue are logical points of intervention. By showing exactly how ligands of different pharmacological classes occupy the GPR84 binding pocket and drive or block its conformational changes, the study turns a previously opaque target into an addressable one. The next steps, testing structure-guided ligand designs and validating their immune effects in disease models, will determine how quickly this structural knowledge moves from the pages of a journal toward the clinic, but the roadmap for modulating GPR84 rationally is now drawn.

Subject of Research: Structural and pharmacological modulation of the lipid-sensing immune receptor GPR84

Article Title: Pharmacological modulation of GPR84 revealed by dual states structures and immune functional assays

Article References: Choi, M. K., Park, D. J., Kim, P., Choi, H. S., Myung, S., Yoo, Y., Chang, N., Yoon, G.-Y., Kang, H. J., Ha, S.-J., & Cho, H.-S. (2026). Pharmacological modulation of GPR84 revealed by dual states structures and immune functional assays. Experimental & Molecular Medicine. https://doi.org/10.1038/s12276-026-01841-w

Image Credits: AI Generated

DOI: 10.1038/s12276-026-01841-w

Keywords: GPR84, G protein-coupled receptor, dual-state structures, immune functional assays, ligand bias, inflammatory signaling, macrophages, medium-chain fatty acids, structural biology, drug discovery, Pharmacological, modulation