The formula, known as Zang-Si-Wei Qingfei Mixture, or ZSW, is a simplified version of a centuries-old Tibetan prescription called Shiwei Longdanhua, which was traditionally prepared from ten plant species native to the Qinghai-Tibet Plateau and used for respiratory ailments ranging from acute bronchitis to chronic obstructive pulmonary disease. The modern ZSW mixture contains four of these plants: Rhododendron anthopogonoides, Gentiana straminea, Corydalis yanhusuo, and Berberis kansuensis. Clinically, the formula has been reported to relieve symptoms such as dyspnea, cough, and sputum production in chronic lung diseases, and earlier mechanistic work suggested it could suppress particulate matter-induced pulmonary inflammation and fibrosis through the TGF-β/SMAD pathway, as well as alleviate lung injury via the ACE2/NF-κB pathway. What remained elusive was which of its many chemical constituents actually drive these effects.

To answer that question, the team, led by Doudou Tang and corresponding author Shangjie Wu of the Second Xiangya Hospital at Central South University, combined computational screening with laboratory experiments. Their starting point was a previous serum pharmacochemical analysis using liquid chromatography-tandem mass spectrometry, which had identified 58 chemical components circulating in rat blood after oral administration of the formula. Filtering these through the TCMSP database and the SwissADME platform narrowed the list to four primary bioactive candidates: palmatine, (+)-erythraline, ellagic acid, and 3′-desmethylpapaverine. Cross-referencing the predicted targets of these compounds with 1,493 acute lung injury-related genes drawn from GeneCards, OMIM, the Therapeutic Target Database, and DisGeNET produced 139 potential therapeutic targets. A protein-protein interaction network built in the STRING database, analyzed with twelve CytoHubba algorithms, highlighted ten hub genes, including EGFR, PIK3CA, PIK3R1, AKT1, SRC, and STAT3, with the PI3K/AKT signaling pathway emerging prominently from enrichment analysis.

Molecular docking then tested how tightly each of the four compounds binds the hub proteins. Binding energies ranged from −6.4 to −10.3 kilocalories per mole, and ellagic acid stood out with a binding energy of −10.3 kilocalories per mole against AKT1, alongside strong affinity for PIK3CA and PIK3R1, two key subunits of the PI3K enzyme. One hundred nanosecond all-atom molecular dynamics simulations run with GROMACS confirmed the stability of the ellagic acid complexes with AKT1, PIK3CA, and PIK3R1; the EA-AKT1 complex converged to a root-mean-square deviation of 0.30 ± 0.05 nanometers after 30 nanoseconds, and free energy landscape plots revealed a single low-energy potential well, indicating restricted conformational rearrangement of the signaling components. These computational results, together with prior literature showing that ellagic acid can alleviate hydrochloric acid-induced lung injury, positioned it as the leading candidate for the formula’s activity.

The researchers then moved to animal experiments, approved by the Animal Ethics Committee of the Second Xiangya Hospital, using male C57BL/6J mice in which acute lung injury was induced by intratracheal instillation of lipopolysaccharide, a component of the outer membrane of Gram-negative bacteria widely used to provoke inflammatory lung damage. In the protective model, 28 mice were divided into four groups, and ellagic acid was given by tail vein injection at 25 or 50 milligrams per kilogram daily for three days before the LPS challenge, a route chosen because the compound has poor oral bioavailability. Seventy-two hours after LPS exposure, high-dose ellagic acid pretreatment had significantly attenuated pulmonary edema, alveolar wall thickening, and inflammatory cell infiltration seen on histology, and it significantly reduced Evans blue dye extravasation, lung coefficients, and lung wet-to-dry weight ratios, all standard measures of barrier breakdown and fluid accumulation.

Inflammation in the airways told a parallel story. Differential cell counts in bronchoalveolar lavage fluid showed that the LPS-driven surges in neutrophils, lymphocytes, and macrophages were significantly suppressed by high-dose ellagic acid, and the fluid’s levels of the pro-inflammatory cytokines TNF-α, IL-6, and IL-1β were markedly reduced. A second experiment with 42 mice added the whole ZSW formula given intragastrically twice daily, along with the corticosteroid dexamethasone and the PI3K/mTOR dual inhibitor BEZ235 as comparator arms. Both ZSW and high-dose ellagic acid significantly improved Evans blue extravasation, lung coefficients, and wet-to-dry ratios, reduced total protein and cytokine concentrations in the lavage fluid, and, on immunohistochemical staining of lung tissue, significantly decreased the levels of phosphorylated PI3K and phosphorylated AKT, the activated forms of the pathway’s core kinases. NLRP3 and its adaptor protein ASC, both induced by LPS, were also significantly inhibited by the treatments.

The cellular arm of the study used MLE-12 mouse lung epithelial cells stimulated with LPS, with cells pretreated for 48 hours either with 10 percent ZSW-containing serum or 10 micromolar ellagic acid. Both interventions significantly reduced the LPS-induced release of TNF-α, IL-6, and IL-1β, and reversed the LPS-driven phosphorylation of PI3K and AKT. The team then probed pyroptosis, a form of programmed inflammatory cell death mediated by the gasdermin D protein, which ruptures the cell membrane after being cleaved into its pore-forming N-terminal fragment. LPS markedly induced NLRP3 expression and the cleavage of pro-Caspase-1 into active Caspase-1 in these cells; ZSW and ellagic acid significantly reduced cleaved Caspase-1 levels, mature IL-1β and IL-18, lactate dehydrogenase release, a marker of membrane rupture, and the upregulation of GSDMD and its N-terminal fragment. BEZ235 produced similar effects, consistent with the idea that pharmacological modulation of PI3K/mTOR signaling can attenuate injury in this model.

Perhaps the most clinically relevant finding came from a therapeutic rather than preventive design. In a model with 20 mice, treatment with ZSW or ellagic acid began only six hours after the LPS challenge, mimicking the reality that patients usually arrive at the hospital after the inflammatory insult has already begun. Even with this delayed start, both treatments significantly improved lung pathology, reduced cell counts and protein concentrations in the bronchoalveolar lavage fluid, and downregulated the mRNA levels of TNF-α, IL-6, and IL-1β in lung tissue. Western blotting showed decreased phosphorylated PI3K, phosphorylated AKT, and NLRP3. Interestingly, in this late therapeutic window, LPS induced full-length GSDMD but not a significant increase in the cleaved GSDMD-N fragment, a discrepancy the authors attribute to the transient nature of GSDMD cleavage, possible degradation of the fragment by 72 hours, and the heterogeneous mix of cell types in lung tissue compared with a synchronized pure cell culture.

The authors are careful about what their data do and do not prove. Because BEZ235 is a dual PI3K/mTOR inhibitor and no genetic or rescue experiments were performed, the findings support the involvement of PI3K/AKT signaling in the effects of ZSW and ellagic acid, but do not establish it as the direct upstream regulator of the inflammatory events observed, nor do they prove a direct causal link between PI3K/AKT signaling and NLRP3 activation. Other limitations remain: constituents such as palmatine and (+)-erythraline may contribute additively or synergistically, the four constituent medicinal materials were not analyzed separately so the origin of each circulating component could not be traced, and validation in additional models and ultimately clinical studies will be needed to determine translational relevance. Ellagic acid did show comparable efficacy to dexamethasone in reducing cytokine levels in this model, which the authors note as grounds for further investigation rather than a claim of clinical superiority. Within those bounds, the study offers a concrete pharmacological account of how a traditional Tibetan formula, and one of its circulating constituents, may act on inflammatory lung injury, and it demonstrates that the benefit is not confined to pretreatment but extends to a post-injury therapeutic window that mirrors real-world clinical timing.

Subject of Research: Preclinical evaluation of the Tibetan herbal formula Zang-Si-Wei and ellagic acid in lipopolysaccharide-induced acute lung injury

Article Title: Zang-Si-Wei and its core constituent ellagic acid alleviate LPS-induced acute lung injury by modulating PI3K/AKT signaling and NLRP3-associated pyroptosis

Article References: Tang, D., Song, B., Wang, H., Wang, Y., Chen, M., & Wu, S. (2026). Zang-Si-Wei and its core constituent ellagic acid alleviate LPS-induced acute lung injury by modulating PI3K/AKT signaling and NLRP3-associated pyroptosis. Frontiers in Pharmacology, 17, 1966973. https://doi.org/10.3389/fphar.2026.1966973

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1966973

Keywords: acute lung injury, Zang-Si-Wei, ellagic acid, NLRP3 inflammasome, pyroptosis, PI3K/AKT signaling, network pharmacology, LPS mouse model, MLE-12 cells, Tibetan medicine, GSDMD, molecular docking