Suppression of the STAT3/MYC/MCL-1 pathway by the fixed Jianpi Huoxue Formula relieves early biochemical abnormalities in liver fibrosis

Abstract

BACKGROUND

Liver fibrosis (LF), a core pathological process driving chronic liver disease, has no ideal clinical intervention approaches. A fixed Jianpi Huoxue Formula (FJHF), a traditional Chinese medicine (TCM) prescription derived from the Sijunzi Decoction and Taohong Siwu Decoction, has shown clinical benefits against LF.

AIM

To investigate the mechanism by which FJHF attenuates LF via STAT3/MYC/MCL-1 pathway regulation.

METHODS

Three strategies were integrated: (1) Clinical single-cell RNA sequencing (scRNA-seq) was performed on liver tissues from three untreated cirrhosis patients and three FJHF-treated patients to characterize hepatic cell heterogeneity and FJHF-responsive gene expression; (2) A randomized controlled trial was conducted in 40 cirrhosis patients with hypersplenism (20 each in the Jianpi Huoxue Formula and control groups) to assess a hepatocellular injury indicator alanine aminotransferase (ALT), hepatic reserve function indicators [total bilirubin (TBIL) and prothrombin time], and fibrosis indicators [hyaluronic acid (HA), laminin, type III procollagen N-terminal peptide (PIIINP), and type IV collagen (CIV)]; and (3) In vitro experiments were performed in human hepatic stellate cells (HSCs) LX-2 using cell counting kit-8, transmission electron microscopy, western blotting, and quantitative real-time PCR to verify the regulatory role of FJHF in the STAT3/MYC/MCL-1 pathway.

RESULTS

FJHF modulated the proportion of HSC populations in a small-sample scRNA-seq cohort. According to clinical data, FJHF treatment significantly decreased ALT (P < 0.01), TBIL (P < 0.05), and the fibrosis markers HA, PIIINP, and CIV (P < 0.05) compared to the control group. In vitro, 20% FJHF-containing serum suppressed HSC activation, as indicated by reduced α-SMA, collagen I, and TIMP1 levels and elevated MMP1 levels. Notably, FJHF inhibits the STAT3/MYC/MCL-1 pathway to ameliorate LF.

CONCLUSION

Short-term FJHF intervention improves early biochemical abnormalities in LF by inhibiting the STAT3/MYC/MCL-1 pathway and suppressing HSC activation. This work provides a scientific basis for the application of FJHF and identifies the STAT3/MYC/MCL-1 axis as a potential target for TCM intervention in early LF metabolic abnormalities.

Key Words: Liver fibrosis; STAT3/MYC/MCL-1 pathway; Network pharmacology; Single-cell RNA sequencing; Hepatic stellate cells

Core Tip: This study comprehensively explored the pharmacological mechanism of a fixed Jianpi Huoxue Formula (FJHF). We integrated single-cell transcriptomics, clinical observations, and in vitro experiments to confirm that FJHF inhibits the STAT3/MYC/MCL-1 pathway, suppresses hepatic stellate cell activation, and ameliorates early biochemical abnormalities in liver fibrosis.

INTRODUCTION

Chronic liver disease (CLD) represents a significant global health issue, with cirrhosis accounting for the majority of CLD-related mortality. In 2016, cirrhosis accounted for 2.2% of global deaths, ranking as the 11^th leading cause of mortality and the 15^th leading cause of morbidity worldwide; however, specific antifibrotic therapies remain largely unavailable[1,2]. Despite advances in modern medical management of CLD, effective pharmacotherapy for liver fibrosis (LF) remains limited[3]. In contrast, traditional Chinese medicine (TCM) has demonstrated unique clinical benefits in alleviating LF[4].

The fixed Jianpi Huoxue Formula (FJHF) used in this study was specially developed with fixed herbal components and dosage, modified from the Sijunzi and Taohong Siwu decoctions. FJHF, developed by Professor Yusheng Yu, consists of ten herbs: Panax ginseng (Renshen), Poria cocos (Fuling), Atractylodes macrocephala (Baizhu), Glycyrrhiza uralensis (Gancao), Angelica sinensis (Danggui), Ligusticum chuanxiong (Chuanxiong), Paeonia lactiflora (Baishao), Rehmannia glutinosa (Shudihuang), Prunus persica kernel (Taoren), and Carthamus tinctorius (Honghua). All the herbal names were verified using the MPNS database (https://mpns.science.kew.org/) (Table 1). Clinically, FJHF has been proven to alleviate hepatocellular injury and enhance the hepatic reserve function, thereby attenuating fibrosis. Previous studies have validated the antifibrotic effects of several components or herb pairs in FJHF. For instance, ginseng regulates the TGF-β1/Smad pathway[5], the Taoren-Honghua pair modulates oxidative stress via the PI3K/Akt pathway[6], Chuanxiong extract alleviates LF via the CTCF/c-MYC/H19 pathway[7], and the Danggui-Chuanxiong pair inhibits collagen deposition and inflammation[8].

Table 1 Details of the herbs in the Jianpi Huoxue Formula.

Chinese name	Latin name	Part(s) used	Amount (g)	Source (manufacturer)	Batch number	Processing specification (per ChP 2020)
Renshen	Panax Ginseng C. A. Mey.	Roots and rhizomesroots	10	Anguo Kangtai Pharmaceutical Co., Ltd.	20230115	Unprocessed (selected for intact roots, no insect damage)
Fulin	Poria Cocos (Schw.) Wolf.	Sclerotium	10	Anhui Huayuan Chinese Herbal Pieces Co., Ltd.	20230208	Decorticated (removed outer bark, cut into 1-2 cm thick slices)
Baizhu	Atractylodes Macrocephala Koidz.	Roots and rhizomesroots	10	Bozhou Huichuntang Chinese Herbal Pieces Co., Ltd.	20230312	Stir-fried (stir-fried with wheat bran until slightly yellow, to enhance spleen-invigorating effect)
Gancao	licorice	Roots and rhizomesroots	10	Ningxia Qiyuan Chinese Herbal Pieces Co., Ltd.	20230120	Honey-roasted (stir-fried with mature honey until evenly coated, to strengthen qi-tonifying effect)
Danggui	Angelicae Sinensis Radix	Roots	10	Gansu Minxian Zhongtian Pharmaceutical Co., Ltd.	20230225	Unprocessed (cut into 2-4 mm thin slices, with strong aromatic odor)
Chuanxiong	Chuanxiong Rhizoma	Roots and rhizomesroots	10	Sichuan Chuanxiong Pharmaceutical Co., Ltd.	20230305	Unprocessed (cut into irregular thin slices, brownish-yellow cross-section)
Baishao	Paeoniae Radix Alba	Roots	10	Anhui Bozhou Baicao Chinese Herbal Pieces Co., Ltd.	20230130	Boiled (boiled in water until the cross-section is white, peeled, cut into slices, dried; raw Baishao, to nourish blood and soothe liver)
Shudihuang	Rehmanniae Radix Praeparata	Roots	10	Henan Huaiqing Pharmaceutical Co., Ltd.	20230218	Steamed (steamed with rice wine, then sun-dried repeatedly until black and glossy, to tonify kidney-yin)
Taoren	Persicae Semen	Seed	10	Shaanxi Weinan Chinese Herbal Pieces Co., Ltd.	20230320	Blanched (blanched in boiling water to remove seed coat, dried, crushed slightly)
Honghua	Carthami Flos	Flower	10	Xinjiang Hotan Red Flower Co., Ltd.	20230205	Unprocessed (selected for bright red florets, no impurities)

The STAT3 signaling pathway plays a pivotal role in LF pathogenesis. Phosphorylated STAT3 directly targets downstream molecules, including MYC and MCL-1, which are strongly involved in LF progression[9-11]. STAT3 activation drives hepatic stellate cell (HSC) activation and aggravates LF, whereas its suppression effectively improves early metabolic abnormalities in hepatic fibrosis. Although the individual herbal components of FJHF are known to act on multiple molecular pathways, the integrated mechanism of this multicomponent formula remains insufficiently understood. Therefore, we hypothesized that FJHF inhibited the STAT3/MYC/MCL-1 pathway to suppress HSC activation, thereby alleviating LF. To test this hypothesis, we combined clinical single-cell RNA sequencing (scRNA-seq), a randomized controlled trial, network pharmacology analysis, and in vitro cellular experiments to elucidate the mechanism by which FJHF improves early metabolic disorders associated with LF, providing both theoretical rationale and experimental evidence to support its clinical application.

MATERIALS AND METHODS

General information

In this randomized controlled trial, patients with cirrhosis and hypersplenism were consecutively recruited at the First Affiliated Hospital of Anhui University of Chinese Medicine from January 2023 to December 2024. Finally, 40 patients were included in the statistical analysis (Figure 1). A statistician, independent of recruitment, generated a 1: 1 random sequence (1-40) via SPSS 21.0. Inclusion criteria: Definite cirrhosis diagnosis per the EASL Clinical Practice Guidelines[12], confirmed by: (1) Clinical evidence of portal hypertension (splenomegaly, jaundice, ascites, varices, or encephalopathy); (2) Biochemical markers: Albumin < 35 g/L, bilirubin > 1.2 mg/dL; and (3) Imaging: Liver stiffness (LS) ≥ 17.5 kPa by FibroTouch or nodular hepatic morphology on contrast-enhanced computed tomography/magnetic resonance imaging.

Figure 1  Flow diagram of the participant inclusion process.

Exclusion criteria were: (1) Incomplete clinical data (missing more than two key parameters); (2) Refusal to provide informed consent; and (3) Early hospital discharge.

Patients were assigned to the control (n = 20, odd numbers) or Jianpi Huoxue Formula (JHF) (n = 20, even numbers) group using sealed opaque envelopes stored at the clinical research center, ensuring allocation concealment until baseline data collection. The participants in the control group received conventional Western medicine for liver protection, jaundice reduction, and other supportive symptomatic therapies. The JHF group received conventional western medicine combined with FJHF, and the formula was administered twice a day orally, with a continuous 21-day medication period as one treatment course. Blood samples for hepatocellular injury [serum alanine aminotransferase (ALT)], hepatic reserve function [total bilirubin (TBIL) and prothrombin time (PT)], and fibrosis markers [hyaluronic acid (HA), laminin (LN), type III procollagen N-terminal peptide (PIIINP), and type IV collagen (CIV)] were anonymized and tested by a third-party laboratory (Anhui Provincial Clinical Laboratory Center). Clinical data of patients, including demographic characteristics, hepatocellular injury indicator (ALT), hepatic reserve function parameters (TBIL, PT), and LF marker results (HA, LN, PIIINP, and CIV), were collected by two nurses. Following the end of treatment, three patients were randomly selected from both the control and JHF groups among the 40 enrolled patients who underwent liver biopsy, and their liver tissue samples were analyzed via scRNA-seq.

The sample size was determined based on primary outcome measures, including serum ALT and LF markers (HA, PIIINP, and CIV). Assuming a mean difference of 10-15 U/L in ALT and 20-30 ng/mL in fibrosis markers between the two groups, with a type I error (α) of 0.05, a power (1-β) of 0.80, a minimum sample size of 20 patients per group was determined to be sufficient to detect statistically significant differences. The sample size was reviewed, and the study was approved by the Institutional Review Board of the First Affiliated Hospital of Anhui University of Chinese Medicine. The total sample size was set to 40 patients (20 in each group). This study was conducted from January 2023 to December 2024. Forty cirrhotic patients with hypersplenism were enrolled, including 24 males and 16 females, with a mean age of 37.88 ± 13.16 years. The higher proportion of male patients was consistent with the epidemiological features of liver cirrhosis and CLD, in which males have a higher incidence because they are more exposed to risk factors such as alcohol consumption, viral hepatitis infection, and unhealthy lifestyles.

Ethics statement

Prior to the study implementation, all participants were provided with a complete explanation of the research objectives and detailed protocols, and written informed consent was obtained from every subject or their authorized legal representative. This research was conducted in compliance with the ethical guidelines of the Declaration of Helsinki of the World Medical Association (revised in Fortaleza, Brazil, October 2013). Ethical approval for this study was granted by the Institutional Review Board of the First Affiliated Hospital of the Anhui University of Chinese Medicine (No. 2024AH-30). All patient information was anonymized and stored in a password-protected database that adheres to HIPAA requirements, with secure data management performed using the Research Electronic Data Capture platform. All procedures involving animals were reviewed and approved by the Institutional Animal Care and Use Committee of the Anhui University of Chinese Medicine (No. 20240030).

Preparation and quality control of the FJHF

All 10 medicinal herbs included in FJHF were obtained from the Bozhou TCM Market located in Bozhou, Anhui Province, China. All herbs were authenticated by Prof. Qing-Lin Li from the Anhui University of Chinese Medicine, in strict accordance with the requirements specified in the 2020 Edition of the Chinese Pharmacopoeia (2020 Edition, Volume I). The FJHF decoction was prepared in the GMP-qualified TCM Decoction Room of the First Affiliated Hospital of Anhui University of Chinese Medicine following a standardized manufacturing procedure to ensure consistency and reproducibility. Each dose of medicinal material was added to 500 mL of water, decocted for 30 minutes, filtered, and concentrated to a crude drug decoction at a concentration of 1.0 g/mL. The decoction was stored at 4 °C and rewarmed before administration, and the entire preparation process was completed with assistance from the institutional preparation center. This formula adopts a fixed herbal species, dosage, and preparation procedure in the present study, which is a specific modified prescription rather than a general Jianpi Huoxue therapeutic principle, and the research conclusions are only applicable to this fixed formulation.

ScRNA-seq

Liver tissue samples were collected from cirrhotic patients (n = 3) and FJHF-treated cirrhotic patients (n = 3) for scRNA-seq analysis. Tissue samples were dissociated into single-cell suspensions (QC: Viability > 80%, no aggregation) and processed for library construction using the 10 × Genomics Chromium platform, followed by mRNA release, reverse transcription to cDNA, PCR amplification, and library construction using (10 × Genomics Single Cell 3′ Reagent Kit v3.1). Validated libraries (Agilent 2100, Qubit 4.0, insert size approximately 300 bp, adapter dimer < 5%) were sequenced using Illumina NovaSeq 6000 (PE150 mode). For bioinformatics, 10 × Genomics Cell Ranger (v6.0) was used for read filtering (Q30 > 85%), genome alignment (GRCh38/GRCm38), and UMI-based quantification. Seurat (v4.0) filtered low-quality cells and integrated samples (Canonical Correlation Analysis), performed PCA-tSNE, and conducted Louvain clustering/differential gene analysis (adjusted P < 0.05, log2FC > 1). SingleR (v1.0) annotated cell types, CellChat (v1.1.0), and inferred ligand-receptor interactions (all via R v4.5.1). We used the VlnPlot function to draw violin plots and applied the feature plot function to generate feature plots for data visualization. Experimental analyses were performed by Nanjin Personal Gene Technology Co., Ltd (Nanjing, China). Baseline characteristics of the six enrolled patients are listed in Supplementary Table 1. Given the limited sample size (n = 3 per group), our scRNA-seq results should be interpreted with caution and cannot be generalized to all patients with cirrhosis.

Network pharmacology analysis of FJHF in LF

Active compounds and their targets in FJHF were retrieved from TCMSP (https://www.91tcmsp.com/#/home) and filtered by pharmacokinetic properties (oral bioavailability ≥ 30%, drug-likeness ≥ 0.18)[13]. The corresponding target genes were obtained from UniProt (https://www.uniprot.org/) and validated using UniProt and GeneCards[14]. LF-related targets were collected from five public databases (Online Mendelian Inheritance in Man (OMIM, https://omim.org/), GeneCards (https://www.genecards.org/), DrugBank (https://go.drugbank.com/), Pharmacogenetics and Pharmacogenomics Knowledge Base (PharmGKB, https://www.pharmgkb.org/), Therapeutic Target Database (TTD, https://db.idrblab.net/ttd/) using the keyword “liver fibrosis”. The intersection targets of FJHF and LF were used to construct drug-compound-gene (Cytoscape 3.8.0) and protein-protein interaction (PPI) (STRING 11.0, confidence = 0.9) networks[15]. Hub genes were selected using the CytoNCA plugin based on six topological indicators: Betweenness, closeness, degree, eigenvector, local average connectivity, and network centrality. Genes with values above the median of all six indices were defined as key hub targets[16]. Functional enrichment analyses, including Gene Ontology (GO), terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, were performed using the clusterProfiler package in R software (version 4.5.1)[17]. Molecular docking (AutoDockTools v1.5.6) was performed using compound structures (PubChem, https://pubchem.ncbi.nlm.nih.gov/) and hub target structures (PDB, http://www.rcsb.org/pdb/home/home.do), with the predicted binding affinity used to evaluate interactions at the computational simulation level[18]. Molecular docking was performed as a computational bioinformatics simulation to predict the potential binding affinity between the active compounds and core targets, without experimental validation. All detailed docking results are provided in the Supplementary Figure 1 section of the manuscript.

Drug-containing serum preparation and quality control

Healthy SD rats were acquired from the Animal Experiment Center of Anhui Medical University [production license No. SCXK(Anhui)20170002; quality certificate No. 202017532]. After one week of adaptive feeding, the rats were randomly divided into two groups (n = 10 per group). Rats in the experimental group were intragastrically administered the FJHF decoction at a daily dose of 10 g/kg for seven consecutive days, while control rats received an equal volume of normal saline via the same route. One hour after the last administration, abdominal aortic blood was collected and centrifuged at 3000 g for 15 minutes to separate the serum. Serum was inactivated in a water bath at a temperature of 56 °C for 30 minutes, and then filtered and sterilized through a 0.22 μm filter to prepare FJHFcontaining serum and blank serum. The sera were aliquoted and stored at -20 °C for later use. The rat dosage was converted from the clinical human equivalent dose, according to the body surface area (BSA) normalization method[19]. The theoretical conversion factor calculated by standard Km values (human Km = 37, rat Km = 6) is 37/6 = 6.17 (rat dose = human dose × 6.17). This BSA-based conversion is the most widely accepted standard for translating clinical human doses to experimental rat doses in pharmacological studies. To guarantee the stability and bioactivity of major constituents in the drug-containing serum, UPLC was employed for quantitative determination to satisfy experimental criteria.

Cell co-culture and experimental grouping

LX-2 (source: CCTCC, catalog No.: GNHu58, species: Human) cells were cultured in DMEM with 10% FBS at 37 °C in a 5% CO₂ atmosphere and sub-cultured (80%-90% confluence) via 0.25% trypsin digestion. LX-2 cells in the logarithmic growth phase were prepared as cell suspensions and co-seeded into 96-well culture plates. LX-2 cells were cultured in FJHF drug-containing serum for 24 hours, after which the supernatant cells were collected for detection.

To investigate the regulatory effects of FJHF on HSCs, the study sample was randomly divided into five groups: (1) Control group: LX-2 cells treated with blank serum for 24 hours[20]; (2) JHF group: LX2 cells treated with JHF drug-containing serum for 24 hours; (3) STAT3 activator group: LX-2 cells treated with the STAT3 activator colivelin (10 μM) for 24 hours[21]; (4) STAT3 inhibitor group: LX-2 cells treated with the STAT3 activator colivelin (10 μM) and STAT3 inhibitor Stattic (10 μM) for 24 hours[22]; and (5) STAT3 activator + JHF group: LX-2 cells treated with JHF drug-containing serum and the STAT3 activator colivelin (10 μM) for 24 hours.

Determination of cell viability via the cell counting kit-8 assay

LX-2 cells were seeded in 96-well plates and treated with blank serum, control serum, or FJHF-containing serum at gradient concentrations of 2%, 5%, 10%, 20%, and 40%. After 24 hours of incubation at 37 °C, the culture medium was decanted, and the cells were rinsed twice. Subsequently, 100 μL of fresh medium supplemented with 10 μL of the cell counting kit-8 (CCK-8) solution was added to each well. After 2 hours of incubation, the optical density at 450 nm was measured using a microplate reader.

Observation by transmission electron microscopy

HSCs were fixed (2.5% glutaraldehyde/osmium tetroxide), dehydrated, embedded, sectioned, stained, and observed using transmission electron microscopy (TEM) (JEM-1400)[23].

Western blotting and the quantitative real-time PCR assay

Subsequently, 10% or 12% gels were prepared. Protein samples (20 μg) were boiled in loading buffer. Electrophoresis was performed at 80 V for 20 minutes for the stacking gel and at 120 V for 60 minutes to separate the gel. The semi-dry transfer was performed at 25 V for 30 minutes, followed by membrane activation. Membranes were blocked and incubated with primary and secondary antibodies before imaging with enhanced chemiluminescence using a gel scanner. The antibodies used for immunoblotting included GAPDH (Proteintech, 60004-1-AP), α-SMA (Proteintech, 14395-1-AP), Collagen I (Abcam, Ab138492), MMP1 (Abcam, Ab134184), TIMP1(Abcam, Ab211926), STAT3 (Abcam, ab68153), MYC (Abcam, ab32), MCL-1 (Abcam, ab32087), and a goat anti-mouse HRP-conjugated secondary antibody (Zhongshan Goldenbridge, ZB-2305). GAPDH was used as the reference band.

Total RNA was extracted from adherent cells using a reagent with chloroform and isopropanol steps, followed by washing with 75% ethanol. cDNA was synthesized from RNA. qPCR was performed in triplicate using standard cycling[24]. Relative expression was calculated via the 2^-ΔΔCt method.

All western blotting and quantitative real-time PCR (qRT-PCR) assays were carried out in three independent biological replicates, and each biological replicate was further examined in triplicate to guarantee the accuracy and dependability of the results.

Statistical analysis

Statistical evaluation was performed using SPSS 21.0. All experiments were performed in triplicate. Quantitative data were expressed as mean ± SD. Differences in quantitative variables between the two groups were analyzed using the independent-sample t-test, and comparisons among multiple groups were performed using one-way analysis of the relative protein/mRNA expression levels was conducted using one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test. Categorical data are presented as n (%) and analyzed using the χ² test; Fisher’s exact test was used when the expected frequency was < 5 or the total sample size was small (n < 40). Statistical ANOVA, followed by Tukey’s post-hoc test. A P value of < 0.05 was considered statistically significant. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001.

RESULTS

In total, 40 cirrhotic patients with hypersplenism were enrolled, including 24 males and 16 females, with a mean age of 37.88 ± 13.16 years.

ScRNA-seq analysis

To investigate changes in cell types during the progression of LF following FJHF treatment, scRNA-seq was performed on six liver tissue samples (three from the control group and three from the JHF treatment group). After quality control, 56550 cells, including 32151 cells in the control group and 24399 cells in the JHF group, were retained for analysis. Cells from both groups were integrated into a batch effect-free standardized dataset via Canonical Correlation Analysis. Dimensionality reduction was conducted using principal component analysis, and the resulting low-dimensional data were visualized via Uniform Manifold Approximation and Projection (Figure 2A). Fourteen major liver cell types were identified using cell type-specific marker genes, and their proportional distribution was significantly altered post-JHF treatment (Figure 2B). The top 20 differentially expressed genes in HSC subtypes were subjected to GO and KEGG analyses, which indicated significant enrichment in inflammation-related pathways and cytokinereceptor interaction signaling (Figure 2C). FJHF modulated cytokine expression in HSCs (Figure 2D and E). (PRJNA1304907, https://www.ncbi.nlm.nih.gov/sra/PRJNA1304907, the Control Group (SAMN50575320, SAMN50575321, and SAMN50575322), and the JHF Group (SAMN50575323, SAMN50575324, and SAMN50575325).

Figure 2 Single-cell RNA sequencing of liver tissues from cirrhotic patients. Six liver samples, divided into the control group (n = 3) and the Jianpi Huoxue Formula (JHF) group (n = 3), were analyzed using single-cell RNA sequencing. A: Diagram showing the experimental procedure in this study; B: Fourteen major liver cell types were identified based on cell-type-specific marker genes after JHF treatment; C: Gene Ontology and terms and Kyoto Encyclopedia of Genes and Genomes analyses of the top 20 differentially expressed genes in human hepatic stellate cells (HSCs) subtypes; D: Violin plots and Uniform Manifold Approximation and Projection (UMAP) plots showing the expression levels of STAT3, MYC, and MCL-1 across HSC subpopulations after JHF treatment; E: Violin plots and UMAP plots showing the expression levels of ACTA2, COL1A1, COL1A2, and MMP1 in the major cell population.

Clinical results of FJHF on hepatocellular injury, hepatic reserve function, and LF markers

No significant differences in baseline characteristics prior to treatment, including general clinical indices, baseline hepatocellular injury, hepatic reserve function, fibrosis markers, LS, and shear wave velocity, were observed between the two groups (all P > 0.05) (Table 2).

Table 2 Baseline characteristics of the included patients.

Variable	JHF group (n = 20)	Control group (n = 20)	P value
Sex (male/female)	10:10	14:6	0.167
Age	36.25 ± 12.53	39.5 ± 13.89	0.442
Etiology			0.673
HLD	17	15
HBV	2	4
Autoimmune hepatitis	1	1
Pre-treatment
ALT (U/L)	66.04 ± 11.72	67.09 ± 12.43	0.786
TBIL (μmol/L)	61.49 ± 18.97	64.29 ± 24.26	0.687
PT (s)	13.75 ± 1.72	13.16 ± 1.46	0.248
HA (ng/mL)	187.38 ± 67.98	171.84 ± 49.02	0.412
LN (ng/mL)	441.74 ± 455.09	404.33 ± 404.86	0.785
PIIINP (ng/mL)	157.95 ± 57.07	153.03 ± 82.53	0.828
CIV (ng/mL)	89.19 ± 30.76	65.75 ± 56.59	0.112
Liver stiffness (kPa)	14.12 ± 4.43	14.50 ± 5.02	0.802
Shear wave velocity (m/s)	2.10 ± 0.33	2.13 ± 0.39	0.760
Post-treatment
ALT (U/L)	30.65 ± 11.63	40.00 ± 9.27	0.008
TBIL (μmol/L)	22.47 ± 9.73	31.01 ± 10.33	0.032
PT (s)	11.25 ± 1.08	12.28 ± 1.80	0.033
HA (ng/mL)	132.09 ± 66.54	171.74 ± 52.42	0.043
LN (ng/mL)	86.44 ± 27.76	167.35 ± 178.08	0.052
PIIINP (ng/mL)	15.87 ± 5.04	76.34 ± 80.59	0.002
CIV (ng/mL)	21.83 ± 12.67	39.44 ± 32.35	0.029
Liver stiffness (kPa)	12.50 ± 3.61	13.60 ± 3.93	0.367
Shear wave velocity (m/s)	1.98 ± 0.28	2.08 ± 0.29	0.315

JHF: Jianpi Huoxue Formula; HLD: Hepatolenticular degeneration; HBV: Hepatitis B virus; ALT: Alanine aminotransferase; TBIL: Total bilirubin; PT: Prothrombin time; HA: Hyaluronic acid; PIIINP: Procollagen N-terminal peptide; CIV: Type IV collagen; LN: Laminin.

Following the FJHF intervention, the level of ALT (the hepatocellular injury indicator) was significantly lower in the JHF group than in the control group (P < 0.01). Meanwhile, the levels of PT and TBIL (hepatic reserve function indicators) also significantly decreased (both P < 0.05). Similarly, the concentrations of HA, PIIINP, and CIV were significantly lower in the JHF group than in the control group (P < 0.05). While LN levels did not differ significantly between the groups, a downward tendency was detected in the JHF group, supporting the antifibrotic potential of FJHF. Additionally, no significant between-group differences were observed in LS (12.50 ± 3.61 kPa vs 13.60 ± 3.93 kPa, P = 0.367) or shear wave velocity (1.98 ± 0.28 m/second vs 2.08 ± 0.29 m/second, P = 0.315) after treatment.

Network pharmacology analysis of FJHF in LF

Disease-related genes were screened using GeneCards, TTD, DrugBank, OMIM, and PharmGKB (Supplementary Figure 1A). In total, 223 intersecting genes were identified between 248 active compound targets and 8968 disease targets (Supplementary Figure 1B). A “herbal-compound-gene target” network (212 nodes, 415 edges) was constructed to analyze the relationships between herbs, ingredients, targets, and diseases (Supplementary Figure 1C). Inputting 223 intersection genes into STRING 11.0 yielded a PPI network (222 nodes, 679 edges, interaction score ≥ 0.9). We analyzed the topological properties of the PPI network in Cytoscape using a network analysis tool based on six indicators. Nodes with values above the median across all parameters were defined as core nodes, yielding 20 hub nodes and 217 edges for the final hub gene network (Supplementary Figure 1D). The enrichment analysis revealed 2409 GO biological processes, 107 cellular components, 253 molecular functions, and 191 KEGG pathways (Supplementary Figure 1E and F). Molecular docking simulation using AutoDockTools 1.5.6 predicted that Licochalcone A luteolin has binding affinities of -6.0 and -7.3 kcal/mol for STAT3 and MCL-1, respectively (Supplementary Figure 1G and H).

Screening of drug-containing serum concentrations

LX-2 cells were incubated with gradient concentrations of FJHF-containing serum (2%, 5%, 10%, 20%, and 40%) with blank and control sera as negative controls. After incubation, the CCK-8 reagent was added to measure the absorbance and calculate viability. As shown in Figure 3, 20% FJHF-containing serum was optimal (appropriate biological effect and no significant cytotoxicity) for subsequent experiments.

Figure 3 Screening of Jianpi Huoxue Formula drug-containing serum concentrations and detection of LX-2 cell proliferation via the CCK-8 assay. CCK-8: Cell counting kit-8; JHF: Jianpi Huoxue Formula.

TEM observations of HSC ultrastructure

The control and JHF groups had HSCs with lipid droplets, sparse short rough endoplasmic reticulum (rER), simple Golgi-scarce mitochondria, small nuclei with indistinct nucleoli, and no collagen. The STAT3 activator group showed HSCs without lipid droplets but with densely distributed and dilated rER, developed Golgi, more mitochondria, large nuclei, and dense collagen deposition. The STAT3 inhibitor and STAT3 activator + JHF groups had HSCs with reduced rER, simplified Golgi, fewer mitochondria, shrunken nuclei, and no lipid droplets or collagen (Figure 4).

Figure 4 Transmission electron microscopy observations of human hepatic stellate cells ultrastructure. A: Transmission electron microscopy (TEM) observation results of the control group; B: TEM observation results of the Jianpi Huoxue Formula (JHF) group; C: TEM observation results of the STAT3 activator group; D: TEM observation results of the STAT3 inhibitor group; E: TEM observation results of the STAT3 activator + JHF group.

FJHF inhibits the activation of HSCs

To explore the effect of FJHF on collagen I, MMP1, α-SMA, and TIMP1 expression, western blotting and qRT-PCR were performed. Relative to the control group, the STAT3 activator group exhibited notably upregulated protein and mRNA expression of collagen I, α-SMA, and TIMP1, along with downregulated MMP1 expression (P < 0.05). The STAT3 inhibitor and STAT3 activator + JHF groups not only showed reduced levels of collagen I, α-SMA, and TIMP1 but also increased levels of MMP1 (Figure 5).

Figure 5 Western blot and quantitative real-time PCR for detecting α-SMA, Collagen I, MMP1, and TIMP1. A: Western blot results of α-SMA, Collagen I, MMP1, and TIMP1; B: Α-SMA western blot result; C: Collagen I western blot result; D: MMP1 western blot result; E: TIMP1 western blot result; F: Α-SMA quantitative real-time PCR (qRT-PCR) result; G: Collagen I qRT-PCR result; H: MMP1 qRT-PCR result; I: TIMP1 qRT-PCR result. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001.

FJHF inhibits the STAT3/MYC/MCL-1 pathway and modulates early fibrotic metabolic abnormalities

To investigate FJHF’s regulatory effects on MCL-1, MYC, and STAT3 expression, western blotting and qRT-PCR analyses were performed. Compared with the control group, the STAT3 activator group showed significantly upregulated protein and mRNA levels of these factors (P < 0.05), whereas the STAT3 inhibitor and STAT3 activator + JHF groups significantly reversed this effect and decreased their expression (P < 0.05) (Figure 6).

Figure 6 Western blot and quantitative real-time PCR for detecting STAT3, MYC, and MCL-1. A: Western blot results of STAT3, MYC, and MCL-1; B: STAT3 western blot result; C: MYC western blot result; D: MCL-1 western blot result; E: STAT3 quantitative real-time PCR (qRT-PCR) result; F: MYC qRT-PCR result; G: MCL-1 qRT-PCR result. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001.

DISCUSSION

This study integrated clinical scRNA-seq, network pharmacology, and in vitro experiments to elucidate the antifibrotic mechanisms of FJHF and its monomeric compounds. The principal finding of this study was that FJHF inhibits the STAT3/MYC/MCL-1 pathway to suppress HSC activation, thereby ameliorating early metabolic abnormalities in LF. Human liver scRNA-seq revealed that FJHF regulates the proportion and expression of cytokines in HSCs. A clinical study revealed that FJHF treatment could alleviate hepatocellular injury (ALT), enhance hepatic reserve function (TBIL and PT), and reduce the levels of early fibrosis markers (HA, PIIINP, and CIV) in the short term. Network pharmacology and computational molecular docking prediction indicated that the main active monomers (luteolin and licochalcone A) of FJHF predicted the binding potential toward STAT3 and MCL-1 at the simulation level. In vitro experiments confirmed that FJHF suppresses HSC activation and downregulates the expression of proteins and mRNAs associated with the STAT3/MYC/MCL-1 pathway. Suppression of HSC activation was evidenced by reduced expression of α-SMA, collagen I, and TIMP1. These findings were consistent with the western blotting, qRT-PCR, and TEM observations.

Collectively, these findings indicate that FJHF exerts antifibrotic effects through the modulation of the STAT3/MYC/MCL1 signaling axis. In addition to consolidating the scientific basis for the therapeutic potential of FJHF against LF, this study identified the STAT3/MYC/MCL-1 axis as a promising therapeutic target. Activation of STAT3 (phosphorylation at Tyr705) upregulates the transcription of MYC and MCL-1, thereby promoting LF[25,26]. Per our findings, FJHF markedly reduced the expression of total STAT3 and its downstream genes, MYC and MCL-1. Further detection of p-STAT3 is needed to confirm its inhibitory effect on STAT3 phosphorylation.

The therapeutic efficacy of FJHF is likely attributed to the synergistic interactions between its herb pairs. For example, the Angelica-Ligusticum pair suppresses collagen deposition, whereas the Prunus-Carthamus pair modulates oxidative stress. These herb-pair-mediated effects complement FJHF’s inhibitory role in the STAT3/MYC/MCL-1 pathway, collectively reflecting the “multicomponent, multi-target” advantage of TCM over single-target western antifibrotic drugs.

A discrepancy was observed between the non-significant change in LS (P = 0.367) and shear wave velocity (P = 0.315), and the significant reductions in the levels of fibrosis markers (HA, PIIINP, and CIV) after FJHF treatment. This inconsistency may be attributed to the following factors. First, LS and shear wave velocity measurements are sensitive to hepatic inflammation and edema, and HA, PIIINP, and CIV are direct biomarkers of extracellular matrix (ECM) degradation, which more specifically reflects fibrotic activity. Second, the short treatment duration (21 days) may be insufficient to induce detectable changes in established hepatic scar tissue, whereas fibrosis markers can reflect early changes in ECM metabolism. Consistent with the clinical findings, 21-day FJHF treatment significantly reduced the levels of serum biochemical fibrosis markers, whereas the levels of LS and shear wave velocity did not differ significantly between the groups. This phenomenon suggests that FJHF primarily modulates early ECM turnover, rather than reversing mature hepatic scar formation. The short intervention duration was insufficient to induce measurable improvements in liver histology and elastography. Therefore, the current results only support short-term biochemical improvement of fibrotic markers and cannot be interpreted as a histological or elastographic reversal of LF. Future studies with longer intervention periods and stratification by inflammatory status are needed to confirm whether FJHF can reduce LS in patients with controlled hepatic inflammation.

The findings of this study provide a theoretical and scientific basis for the clinical application of FJHF in patients with cirrhosis complicated by hypersplenism. Short-term (21-day) FJHF treatment alleviated hepatocellular injury (ALT), enhanced hepatic reserve function (TBIL and PT), and reduced levels of early fibrosis markers (HA, PIIINP, and CIV). Furthermore, the identification of the STAT3/MYC/MCL1 pathway as a key target provides potential for biomarker development, which can be employed as a non-invasive biomarker for assessing the clinical therapeutic efficacy of FJHF.

Nevertheless, this study had the following noteworthy limitations: (1) ScRNA-seq employed only six human samples (three per group). Such a minute sample cannot fully reflect the individual heterogeneity of cirrhotic patients and may introduce potential selection bias, which limits the generalizability of our single-cell transcriptomic findings; (2) This study entailed only in vitro cell experiments and clinical observational analysis, and there were no in vivo functional verification using classic LF animal models such as CCl-induced or bile duct ligation models. The absence of histological and pathological evaluation of LF in animal models weakens the in vivo evidence for the regulatory effect of FJHF on the STAT3/MYC/MCL-1 pathway; (3) We focused on the STAT3/MYC/MCL-1 pathway, without examining potential crosstalk with other pathways, which may also contribute to FJHF’s effects; (4) The short treatment duration (21 days), despite confirming the short-term benefits of FJHF, cannot fully reflect its efficacy in retarding or reversing established LF, thus precluding the ability of the study to assess long-term safety and efficacy; (5) This study only detected the protein and mRNA expression of total STAT3, but did not examine the phosphorylation level of STAT3 (p-STAT3) or calculate the p-STAT3/total STAT3 ratio, which weakens the evidence for the activation status of the STAT3 signaling pathway. Furthermore, we did not perform STAT3 overexpression rescue experiments to further confirm the specific regulatory mechanism of FJHF on the STAT3/MYC/MCL-1 axis; and (6) The molecular docking analysis conducted in this study is merely a bioinformatics computational prediction, and it lacks in vitro or in vivo experimental verification of actual protein binding. Future research will focus on six key directions: First, to compensate for this limitation, we have supplemented detailed clinical baseline information of the six patients enrolled for scRNA-seq in the Supplementary Table 1 to confirm the baseline balance between the two groups; second, we will adopt classic CCl-induced and bile duct ligation LF animal models to further verify the in vivo regulatory effect of FJHF on the STAT3/MYC/MCL-1 pathway, and systematic histological staining and fibrosis stage grading will be performed to evaluate the pathological changes in LF and confirm the in vivo efficacy of FJHF; third, to explore the crosstalk between the STAT3/MYC/MCL-1 pathway and other fibrotic pathways (such as the TGF-β pathway) using proteomic techniques; fourth, future work will detect p-STAT3 and total STAT3 protein expression to calculate their ratio and verify FJHF inhibition on STAT3 phosphorylation, and conduct STAT3 overexpression rescue assays in LX-2 cells to elucidate how FJHF modulates the STAT3/MYC/MCL-1 pathway and represses HSC activation; fifth, to enroll a larger cohort of cirrhotic patients, extend the treatment course to 3-6 months, and conduct follow-up assessments at 6 months, 12 months, and 24 months post-treatment; and sixth, further experimental validation is required to confirm the real molecular interaction.

CONCLUSION

In conclusion, short-term FJHF therapy improves early biochemical abnormalities of LF by inhibiting the STAT3/MYC/MCL-1 pathway and suppressing HSC activation without reversing mature hepatic scar tissue or improving LS in the short term. This study not only establishes a robust theoretical foundation for the clinical use of FJHF in patients with cirrhosis but also identifies the STAT3/MYC/MCL-1 axis as a potential target for TCM intervention in early LF metabolic abnormalities.