The investigation was divided deliberately between a multicomponent plant fraction and a purified compound. Researchers first tested an ethyl acetate fraction of S. flavescens in a mouse model of carbon tetrachloride-induced liver fibrosis. They then used purified kurarinol A, designated KA, for functional and mechanistic experiments in activated LX-2 cells. TGF-β1 stimulation provided a model of stellate-cell activation for measuring fibrosis-associated behavior, while lipopolysaccharide was used to stimulate TLR4/MyD88/NF-κB signaling. This separation allowed the researchers to assess the fraction in an intact animal while avoiding the claim that the activity of the entire extract could be attributed directly to KA. The contribution of KA to the fraction’s effects in vivo remains inferred rather than directly demonstrated.

For the animal experiment, male C57BL/6 mice were assigned to normal-control, fibrosis-model, silibinin-control and two EAF-treatment groups, with eight animals per group. Except for controls, the mice received carbon tetrachloride twice weekly for two weeks. They were then treated by gavage for another two weeks with vehicle, silibinin at 100 mg/kg or EAF at 50 or 100 mg/kg. Hematoxylin and eosin staining showed inflammatory infiltration, hepatocyte injury and disrupted liver architecture in untreated model animals. Both EAF doses reduced these lesions, with the 100 mg/kg group showing the greatest improvement. Blinded investigators assessed the histology and quantified Masson-stained collagen, while serum alanine aminotransferase, aspartate aminotransferase and hepatic hydroxyproline served as complementary indicators of injury and fibrosis.

The EAF-treated mice had significantly less collagen-positive area than model animals, and the reduction was dose-dependent. Liver hydroxyproline and serum ALT and AST levels also fell after treatment, while hepatic α-smooth muscle actin and fibronectin expression was lower in the EAF groups. The high dose produced the strongest changes in these measures. These observations show that the ethyl acetate fraction alleviated liver injury and fibrotic changes in this short-term carbon tetrachloride mouse model. They do not establish the effect of purified KA in animals, nor do they show that KA caused the fraction’s benefits. The experiment also used only males and a model designed to represent subacute toxic injury rather than the full progression from chronic fibrosis to cirrhosis.

To identify active constituents, the researchers screened 35 purified compounds from the fraction alongside EAF in LX-2 cells. Most compounds left viability above 80% at the initial screening concentration of 40 μM, indicating limited cytotoxicity under those conditions. Eleven compounds inhibited the growth of TGF-β1-activated cells by more than 50% in the initial screen, with half-maximal inhibitory concentrations ranging from 4.80 to 40.35 μM. KA ranked third among these hits, with an IC50 of 12.65 μM from three independent experiments. Because EAF was tested by mass concentration and purified compounds by molar concentration, the authors treated that comparison as a qualitative ranking rather than a direct potency comparison. KA was selected for follow-up because its lavandulyl-flavanone scaffold differed from those of the other active molecules and because earlier work from the group had suggested effects on inflammatory signaling.

Functional tests in activated LX-2 cells showed that KA reduced viability and fibronectin messenger RNA in a concentration-dependent manner. A scratch assay indicated reduced migration at both 24 and 48 hours. Flow cytometry detected more apoptosis, rising from 10.3% at 2.5 μM to 19.9% at 10 μM, compared with 6.8% in the activated model control and 13.6% in the silibinin control. After 48 hours of treatment, KA also lowered messenger RNA and protein levels of fibronectin, α-smooth muscle actin and collagen I. These measurements connect the compound’s cellular effects with markers of stellate-cell activation, although they were obtained in an LX-2 cell line rather than in primary cells or animals. The reported apoptosis value is a cell-population measurement and does not establish that KA selectively kills activated cells.

Transcriptome sequencing was used to explore how KA altered the activated-cell state. Each of the untreated, TGF-β1-stimulated and 10 μM KA-treated groups had four biological replicates. Relative to untreated cells, TGF-β1 stimulation upregulated 444 genes and downregulated 2,947. KA treatment changed 216 genes that had been upregulated in the model group and 119 that had been downregulated, reversing a modest and specific part of the stimulation-associated profile. Thirty-four overlapping genes were subjected to Gene Ontology and KEGG enrichment analysis, which linked them with several processes, including Toll-like receptor signaling. The authors cautioned that the 34-gene input was small and that enrichment results based on such a list require careful interpretation. The transcriptomic data therefore supported pathway involvement but did not, on their own, establish MyD88 as the direct target.

Several complementary experiments pointed to MyD88 engagement. Docking placed KA in a predicted pocket on the TIR domain of human MyD88 with a score of −8.573 kcal/mol and identified hydrogen bonds involving the backbone groups of Arg251, Leu252 and Phe270, alongside multiple hydrophobic contacts. A 100-nanosecond molecular-dynamics simulation indicated that the complex stabilized after roughly 20 nanoseconds, with persistent contacts involving residues such as Arg251, Leu252 and Ser244. Because this was a single trajectory, the computational result provides support rather than definitive proof of binding or its thermodynamic properties. In the cellular thermal shift assay, MyD88 retained more protein after heat exposure in KA-treated cells than in vehicle-treated controls, and the effect at 55 °C increased with KA concentration. These observations were consistent with KA altering MyD88’s thermal stability and, together with the computational results, supported the proposed interaction.

Genetic manipulation tests then examined whether MyD88 was necessary for KA’s cellular effects. Knockdown of MyD88 reduced the fibrosis-associated proteins α-smooth muscle actin, fibronectin and collagen I. KA did not further suppress these fibrosis markers or the downstream proteins TAK1, phosphorylated p65 and TNF-α in MyD88-knockdown cells, suggesting that its additional effect required the adaptor. In the reciprocal experiment, MyD88 overexpression increased fibrosis-associated markers, while adding KA reduced α-smooth muscle actin, fibronectin and collagen I relative to cells with elevated MyD88 alone. Together, the knockdown and overexpression results support a functional role for MyD88 in mediating KA’s anti-fibrotic activity in LX-2 cells. They do not show that MyD88 is the only target, that KA selectively inhibits it, or that the compound acts through it in humans.

Among the study’s limitations was the distance between cell-culture concentrations and exposure in a target organ. KA was effective in vitro at 2.5 to 10 μM in several assays, but the researchers did not measure whether oral administration of EAF produces comparable hepatic concentrations, and no pharmacokinetic study established absorption, distribution, metabolism or excretion of purified KA. The animal findings also cannot separate KA from other constituents of the fraction. Key binding residues require confirmation through site-directed mutagenesis, while direct biophysical methods such as microscale thermophoresis, surface plasmon resonance or isothermal titration calorimetry were not performed. Dedicated counterscreens for assay interference, quantitative purity assessment of the KA batch and selectivity comparisons against other adaptor proteins also remain outstanding. The study used only male mice, and its carbon tetrachloride model did not cover spontaneous regression of fibrosis or progression to cirrhosis. Taken together, the work identifies KA as a mechanistically promising preclinical compound whose effects require pharmacokinetic studies, direct binding validation and testing across additional fibrosis models before it could be considered for therapeutic development.

Subject of Research: Preclinical study of kurarinol A, MyD88 signaling, hepatic stellate-cell activation and liver fibrosis

Article Title: Kurarinol A from Sophora flavescens suppresses hepatic stellate cell activation via the TLR4/MyD88/NF-κB pathway

Article References: Chen, X.-J., Wang, Q., Wang, M.-Z., Zhang, X., Liao, S.-G., Zeng, Z., Li, X.-P., Wang, L., Tu, B., & Lin, Y. (2026). Kurarinol A from Sophora flavescens suppresses hepatic stellate cell activation via the TLR4/MyD88/NF-κB pathway. Frontiers in Pharmacology, 17, Article 1948257. https://doi.org/10.3389/fphar.2026.1948257

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1948257

Keywords: kurarinol A, Sophora flavescens, liver fibrosis, hepatic stellate cells, MyD88, TLR4 signaling, NF-κB, natural products, preclinical research, anti-fibrotic activity, Kurarinol, Sophora