Liver fibrosis, characterized by excessive extracellular matrix (ECM) deposition resulting from sustained hepatic injury and activation of hepatic stellate cells (HSCs), is a common outcome of chronic liver diseases and a precursor to cirrhosis, liver failure, and hepatocellular carcinoma with substantial global morbidity and mortality (Kisseleva and Brenner, 2021; Akkız et al., 2024). Although cessation of liver injury may allow partial regression of fibrosis, there are currently no approved anti-fibrotic drugs with proven clinical efficacy (Zhao et al., 2022). Many agents that show efficacy in rodent models fail in clinical trials, underscoring the limited translational relevance of conventional liver models. Recent advances in human induced pluripotent stem cell (hiPSC)-derived organoid technology have provided a more predictive platform for identifying and validating novel anti-fibrotic targets and agents (Verstegen et al., 2025; Huang et al., 2025).

Multilineage hepatic organoids generated from hiPSCs, containing parenchymal hepatocyte-like cells alongside non-parenchymal lineages such as HSC, have emerged as physiologically relevant models capable of mimicking key aspects of liver development, homeostasis, and injury responses (Wu et al., 2019; Shi et al., 2022; Guan et al., 2021; Kim et al., 2023). However, most reported multilineage organoids systems typically rely on separately differentiating each cell lineage to maturity before combining them, a strategy that often results in marked heterogeneity in cellular maturation and functionality. Additionally, such models suffer from considerable batch-to-batch variability and a limited ability to consistently induce and maintain fibrotic responses, which constrains their utility in mechanistic studies and anti-fibrotic drug screening. Based on our previous work (Wu et al., 2019; Shi et al., 2022), we developed multilineage hepatobiliary organoids (mHBOs) construction approach in which multiple cell lineages undergo synchronous maturation and self-organization during co-differentiation, thereby enhancing intercellular communication, establishing physiologically relevant architecture, and improving reproducibility.

Lycii Radicis Cortex (LRC, named Digupi in Chinese), derived from the root bark of Lycium barbarum L. or related species in the Solanaceae family, has long been used in traditional Chinese medicine for treating chronic inflammatory and liver disease (Yang et al., 2017; Chen et al., 2021; Anwar et al., 2020; Xie et al., 2014). Nevertheless, most current evidence is derived from animal models or conventional cell cultures, which fail to recapitulate the complex human hepatic fibrotic microenvironment and multicellular interactions. The lack of validation in organoid-based models limits precise mechanistic elucidation and hinders clinical translation. Therefore, integrating LRC pharmacological investigation with mHBOs technology could enable systematic evaluation of its anti-fibrotic effects under physiologically relevant conditions that closely mimic human disease progression.

In this study, we sought to bridge these gaps by integrating LRC pharmacological evaluation with a physiologically relevant human mHBO fibrosis model. We combined in vivo assessment of a carbon tetrachloride (CCl₄)-induced mouse model with in vitro mHBOs modeling of transforming growth factor (TGF)-β driven fibrosis to systematically investigate the anti-fibrotic potential of LRC. Furthermore, using network pharmacology analyses, we identified putative bioactive compounds, molecular targets, and signaling pathways mediating its effects. Mechanistic studies focused on the cyclic adenosine monophosphate (cAMP)-dependent protein kinase A (PKA) axis revealed that LRC attenuates fibrosis by suppressing parenchymal apoptosis in mHBOs. This work establishes a translational pipeline from animal models to human organoids, providing both mechanistic insight and methodological advancement for anti-fibrotic drug discovery.

Human iPSC line UC (Wu et al., 2019; Shi et al., 2022) was obtained from the Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences. Passages of the cell lines ranged from 40 to 50. Human iPSC line Q-iPS-10 was purchased from the Institute of Zoology, Chinese Academy of Sciences. Passages of the cell lines ranged from 18 to 23. Cells were cultured on Matrigel-coated plates (Corning, 354,277) in mTeSRplus medium (STEMCELL Technologies, 100–0276) at 37 °C in a humidified incubator (ESCO, CLM-240B-8-CN) with 5% CO₂. The medium was refreshed daily, and cells were passaged when they reached approximately 80% confluency. The transmission ratio is in the region of 1:6 to 1:8.

Upon reaching approximately 80% confluence, the medium was aspirated and cells were washed once with PBS. Accutase was added, and the cells were incubated for approximately 5 minutes. The digestion was terminated with DMEM/F12. Following centrifugation, the cells were resuspended in CryoStor CS10 cryopreservation medium (STEMCELL Technologies, 100–1061), transferred to freezing tubes tube and a record made. The tubes were placed in a freezing container at −80 °C overnight and subsequently transferred to liquid nitrogen for long-term storage.

We generated mHBOs following previously described protocols (Wu et al., 2019), in which a small fraction of mesodermal cells is intentionally retained during definitive endoderm specification. Under directed induction, endodermal cells differentiate into hepatic parenchymal cells, whereas the mesodermal fraction gives rise to non-parenchymal lineages, including hepatic stellate cells, as described previously.

On day 25 of culture, mHBOs were allocated to four groups: Control, Model, LRC, and H89+LRC. Fibrogenesis was induced in the Model and LRC groups by a 96-h continuous exposure to transforming growth factor-β (TGF-β, 20 ng/mL; NovoProtein, CA72); the Control group received an equal volume of phosphate-buffered saline (PBS). From 48 h onward, the LRC group received LRC (50 μg/mL) until endpoint; the Control and Model groups received vehicle (DMSO) on the same schedule. The 50 μg/mL concentration was selected with reference to prior in vitro studies across multiple cell types, where it is commonly used and non-cytotoxic (Anwar et al., 2020; Mughal et al., 2022; Son et al., 2024; Lee et al., 2021; Park et al., 2019). In the absence of mHBO-specific reports, this choice follows established cell-based practice. To interrogate involvement of the cAMP–PKA pathway, a fourth experimental group was included that received identical LRC treatment but was pretreated with the PKA inhibitor H89 for 30 min prior to LRC exposure. At the end of treatment, organoids and conditioned media were collected for downstream analyses, including qPCR, immunofluorescence, and ELISA.

Total RNA was isolated from cultured cells using the High-Purity RNA Rapid Extraction Kit (MIKX, MKG-865). cDNA was synthesized from total RNA with the PrimeScript RT Reagent Kit with gDNA Eraser (Takara, RR037A) according to the manufacturer’s protocol. qPCR was performed using TB Green Premix Ex Taq II (Takara, RR820A) on a q900 thermal cycler. Relative transcript abundance was calculated using the 2^−ΔΔCT method with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal reference. Primer sequences are provided in Supplementary Table S2.

Cultures were first exposed to 4% paraformaldehyde (PFA; Solarbio, P1110) at room temperature for 30 min, after which membranes were permeabilized for 1 h in DPBS containing 0.3% Triton X-100 (Sigma-Aldrich, T9284). Samples were rinsed three times with DPBS (5 min each) and then placed in 5% donkey serum (Solarbio, SL050) in DPBS for at least 1 h. Primary antibodies diluted in 3% donkey serum were applied for 1 h at room temperature or overnight at 4 °C, followed by secondary antibodies prepared in 1% donkey serum for 30 min at room temperature. Where nuclear labeling was required, DAPI was introduced for 3 min. After the primary, secondary, and DAPI steps, cultures were washed four times with DPBS at room temperature (5 min per wash). At least three samples per group were stained, with three non-overlapping fields of view captured per sample. Images were acquired using a confocal laser scanning microscope (LSM800; Carl Zeiss). The list of antibodies used in this study were provided in Supplementary Table S3.

Cultures were dissociated into single cells with trypsin at 37 °C and pelleted by centrifugation at 350 g for 3 min. After three washes with DPBS, 250 µL of Fixation/Permeabilization solution (BD, 554,714) was added, and cells were incubated for 20 min at 4 °C. Cells were then washed twice with BD Perm/Wash™ buffer (BD, 554,714) and resuspended thoroughly in 100 µL of BD Perm/Wash™ containing fluorochrome-conjugated antibodies at empirically determined working dilutions or appropriate isotype/negative controls. Incubations were performed for 30 min at 4 °C in the dark. Following two additional washes with Perm/Wash™ (1 mL per tube each), cells were resuspended in DPBS and subjected to flow-cytometric acquisition. Data were collected on a BD FACS LSRFortessa flow cytometer and analyzed using FlowJo (v10.8.1).

Intracellular cAMP and p-CREB in mHBOs were measured using ELISA kits according to the manufacturer’s instructions (cAMP: MM-61018H2; p-CREB: MM-0006H2; Jiangsu Meimian industrial Co., Ltd.). Absorbance at 450 nm was recorded with a microplate reader (Labsystems Multiskan MS, 352), and concentrations were calculated from standard curves.

Cultures were exposed to 4% PFA at room temperature for 30 min to fix the cells, followed by permeabilization in DPBS containing 0.3% Triton X-100 for 10 min. After two DPBS washes, samples were incubated with 100 µL TUNEL detection solution (Beyotime, C1088) for 30 min at 4 °C in the dark. Subsequently, they were washed twice with DPBS (5 min per wash), and imaging was performed using a confocal laser scanning microscope (LSM800; Carl Zeiss).

Data were analyzed using GraphPad Prism (version 9.5.1). Group differences were evaluated with Student’s t-test for two-group comparisons and one-way ANOVA for multiple groups. Statistical significance was indicated by asterisks: *p < 0.05, **p < 0.01 and ***p < 0.001. Error bars denote standard deviations (SD). Full statistical details are reported in the corresponding figure legends.

Liver fibrosis represents a major clinical challenge with no approved anti-fibrotic drugs currently available, making discovery of effective agents an urgent need (Parola et al., 2024; Juanola et al., 2025; Mohammed et al., 2023). A notable strength of this study is the use of a human mHBOs model incorporating mesoderm-derived HSCs generated via synchronous co-differentiation from a single hiPSC source (Figure 2). This design preserves developmental lineage-lineage interactions and yields organoids capable of exhibiting robust and sustained fibrotic responses upon TGF-β stimulation (Figures 2A,B). Importantly, most reported multilineage hepatic organoid systems typically rely on separately differentiating each cell lineage to maturity before mixing—a strategy that often results in marked heterogeneity in cell maturation and functional status (Kim et al., 2023; Akbari et al., 2019; Moss et al., 2025; Chi et al., 2025; Thompson and Takebe, 2020). Additionally, such assembled models suffer from considerable batch-to-batch variability and limited ability to consistently induce and maintain fibrotic phenotypes, thereby constraining their utility in mechanistic studies and anti-fibrotic drug screening. In the present study, we combined in vivo and in vitro approaches to provide comprehensive evidence that Lycii Radicis Cortex (LRC) exerts potent anti-fibrotic effects. Our findings expand the pharmacological profile of LRC in hepatic fibrosis, bridging a translational gap between conventional rodent models and human-relevant systems.

In a CCl₄-induced murine model, LRC administration ameliorated liver injury indicated by improved body weight gain, reduced serum ALT and AST, and attenuation of histological markers of fibrosis including collagen deposition (Figures 2B–F). Extending this observation to a human multilineage hepatobiliary organoids (mHBOs) model incorporating hepatic stellate cells (HSCs), we demonstrated that LRC significantly mitigated TGF-β–induced ECM accumulation and HSC activation, paralleled by decreased organoid structural disruption and improved biochemical injury markers (Figures 3B–F; Supplementary Figure S2A,B). To strengthen methodological rigor and translational relevance, we replicated all key mHBO experiments in an independent hiPSC line under identical protocols; results were consistent across lines, reducing the influence of cell line–specific and batch-specific effects and supporting LRC’s antifibrotic activity (Supplementary Figure S5).

Mechanistically, network pharmacology implicated 13 bioactive compounds, with beta-sitosterol, acacetin, and stigmasterol emerging as potential lead constituents via high target connectivity (Figures 4A–E). GO and KEGG enrichment analyses converged on pathways related to receptor-mediated cAMP signaling and apoptosis regulation (Figures 4F,G), indicating that modulation of cyclic nucleotide signaling cascades could be a central axis through which LRC mediates its anti-fibrotic effects. Notably, cAMP is recognized as a second messenger capable of integrating upstream receptor activation with downstream transcriptional programs, including those governing cell survival and matrix turnover (Musheshe et al., 2018). Beta-sitosterol and stigmasterol have been reported to mitigate hepatic fibrosis via modulation of MAPK, NF-κB, and NLRP3 pathways (Devaraj et al., 2020; Wang X. et al., 2025; Han et al., 2024) and intersect with cAMP-related signaling in other systems (beta--sitosterol: cAMP/NF-κB in osteoclasts; stigmasterol: cAMP-CREB in the nervous system) (Guo et al., 2025; Haque et al., 2018), supporting a mechanistic link to the cAMP–PKA–CREB activation observed here. Acacetin, a hepatoprotective flavone, may further augment this axis (Zhou et al., 2021; Chen and Gao, 2024), as flavone scaffolds can elevate intracellular cAMP via phosphodiesterase inhibition (Cormier et al., 2018; Hamsalakshmi et al., 2022). Although constituent levels and exposure were not quantified, phytochemical studies identify sterols and flavonoids as measurable LRC components (Yang et al., 2017). Together, these observations provide a focused molecular rationale for LRC’s antifibrotic activity.

Mechanistic validation in the mHBO model confirmed that LRC activates the cAMP–PKA pathway, as evidenced by increased PRKACA transcription, elevated intracellular cAMP levels, and enhanced phosphorylation of CREB (Figure 5; Supplementary Figure S3A). Importantly, pharmacological inhibition of PKA with H89 abolished these effects and reversed LRC-mediated protection against fibrosis and parenchymal apoptosis (Figure 6; Supplementary Figure S3B,C). This is consistent with literature demonstrating that cAMP–PKA signaling can suppress HSC activation, promote matrix degradation, and inhibit pro-fibrotic gene expression (Wójcik-Pszczoła et al., 2020; Weng et al., 2015), partly via phosphorylated CREB’s transcriptional regulation of apoptosis-related genes such as BCL2 and BAX (Wahlang et al., 2018; Li et al., 2025; Wang Q. et al., 2025; Zhang et al., 2024; Yang et al., 2015). These effects are thought to occur in part through phosphorylated CREB’s transcriptional regulation of apoptosis-related genes such as BCL2 and BAX, thereby re-balancing cell fate toward resolution of fibrotic injury. Complementing this, GO/KEGG enrichment included G protein-coupled receptor (GPCR) signaling and adenylyl cyclase (AC)–related terms, and the PPI network featured ADRB1/ADRB2/CHRM1, suggesting potential receptor-level input into the cAMP-PKA-CREB axis (Ripoll et al., 2025). While our data demonstrate PKA dependence (H89-sensitive), GPCR/AC involvement was not directly tested.

This study has several limitations. We used only male mice, precluding assessment of sex-dependent variability and limiting generalizability across sexes. The in vivo design lacked a non-fibrotic LRC-only group, limiting assessment of baseline effects of LRC on hepatic morphology, metabolic parameters, and signaling (e.g., cAMP–PKA) in the absence of injury. Finally, although the data implicate the cAMP–PKA axis, contributions from upstream or parallel pathways (e.g., GPCR–adenylyl cyclase) cannot be excluded. Future work will include age-matched female cohorts, add an LRC-only group, and test receptor-mediated inputs with targeted pharmacologic and genetic perturbations.

In conclusion, this study demonstrates that LRC attenuates hepatic fibrosis by suppressing parenchymal apoptosis via activation of the cAMP–PKA signaling pathway, substantiated through both murine and human organoid models. Beyond revealing a novel anti-fibrotic mechanism for a traditional medicinal material, our work highlights the utility of mHBOs as an effective translational platform for natural product research, mechanistic studies, and drug discovery in chronic liver disease.