Modulation of gut microbiota and its metabolite Equol by Huaier granule suppresses hepatocellular carcinoma via the gut-liver axis

Abstract

Hepatocellular carcinoma (HCC) is a frequently seen malignant tumor globally. Huaier is the dried fruiting body of the fungus Trametes robiniophila Murr. Huaier granule (HEG), formulated from the Huaier extract, is a Class I innovative anti-cancer drug in China and exhibits significant anti-HCC effects in clinical settings. Nevertheless, the specific mechanisms underlying its efficacy remain incompletely understood. This research demonstrated that HEG effectively suppressed tumor development in the orthotopic HCC mouse model in a gut microbiota-dependent manner and modified the gut microbiota composition. Notably, the primary differential bacterial genera between the Model group and the HEG group included Adlercreutzia. HEG exerted anti-HCC effects by repairing the intestinal barrier, improving colon immunity, and ameliorating the immune microenvironment by suppressing the MAPK signaling pathway via the gut microbiota-gut-liver axis. By integrating 16S rRNA sequencing with metabolomics data, supplemented by literature mining and in vitro validation, Equol, produced by specific gut microbiota Adlercreutzia, was identified as a key metabolite through which HEG exerted its anti-HCC effects by modulating gut microbiota. Moreover, Equol was essential for the anti-HCC effects of HEG. Additionally, Equol ameliorated the immune microenvironment through inhibiting the MAPK signaling pathway, while concurrently inhibiting the growth of HCC cells by inducing the G₀/G₁ phase blockade through suppression of Cyclin E1-CDK2/Rb signaling pathway. This study provided a robust scientific foundation for the clinical use of HEG, with Equol emerging as a promising candidate for HCC treatment.

Introduction

Liver cancer is a frequently seen malignant tumor worldwide. In 2022, it ranked sixth in global incidence and third in mortality rates, resulting in over 750,000 deaths¹. Epidemiological predictions indicate that by 2040, 1.4 million individuals will suffer from liver cancer, and 1.3 million will succumb to the disease worldwide². Among its diverse subtypes, hepatocellular carcinoma (HCC) shows the highest prevalence. For early-stage HCC, the main treatment strategies consist of surgical resection, liver transplantation, and radiofrequency ablation. Conversely, systemic therapy and arterial chemoembolization are the primary treatments for intermediate and advanced stages, respectively. Approved clinical agents for HCC treatment include sorafenib, levatinib, atezolizumab, and bevacizumab³. However, many patients struggle to complete the recommended treatments due to compromised immune function or the inability to tolerate the toxic side effects associated with long-term drug use. Consequently, it is crucial to identify highly effective and low-toxicity drugs for the treatment of HCC.

The human gut is home to around 1 × 10¹⁴ microorganisms, predominantly made up of bacteria. These microorganisms possess more than 3 million genes, generating thousands of metabolites⁴. The interaction between gut microbiota and HCC occurs via the gut-liver axis. When the intestinal barrier is weakened, the gut microbiota and its metabolites move from the intestine to the liver via the portal vein and are later secreted back into the intestine through bile, which therefore affects the development of HCC^5,6. Equol, a metabolite derived from soy isoflavones through the action of specific gut microbiota such as Adlercreutzia, exhibits both estrogenic and antioxidant activities, contributing to various beneficial effects on human health⁷. However, the precise role and underlying mechanisms of Equol in the progression of HCC remain inadequately understood.

Huaier, the dried fruiting body of the fungus Trametes robiniophila Murr., exhibits anticancer and hemostatic properties⁸. Currently, solid fermentation processes can be employed to produce Huaier fungus, from which Huaier extract can be extracted using water or ethanol. The Huaier extract can subsequently be re-prepared into Huaier granule (HEG)⁹. HEG is a Class I innovative anti-cancer drug in China and has been incorporated into the “Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2022 Edition)”. Both clinical and basic research have demonstrated that HEG possesses significant anti-HCC effects¹⁰. An authoritative multicenter randomized controlled study has confirmed that HEG significantly reduces the recurrence rate in patients following liver cancer resection¹¹. However, there is currently no existing research that demonstrates the anti-HCC effects of HEG through the modulation of gut microbiota, and the exact mechanisms involved remain ambiguous and require further exploration.

In the present study, we demonstrated that HEG modulated the gut microbiota composition and elevated the level of its metabolite Equol, and exerted anti-HCC effects by repairing the intestinal barrier, enhancing colonic immunity, and improving the immune microenvironment of HCC by suppressing the MAPK signaling pathway through the gut microbiota-gut-liver axis. The differential metabolite Equol was essential for the anti-HCC properties exhibited by HEG. It enhanced the immune microenvironment by weakening the MAPK signaling pathway, while also suppressing the growth of HCC cells by inducing the G₀/G₁ phase blockade via inhibition of the Cyclin E1-CDK2/Rb signaling pathway. Thus, this study established the solid scientific foundation for applying HEG in the clinic and laid an important theoretical groundwork for considering Equol in treating HCC.

Results

Inhibition of orthotopic HCC development by HEG depended on the presence of gut microbiota

The H22-luc orthotopic HCC mouse model was developed and subjected to HEG (Fig. 1A). Compared to the Model group, the tumor weight was significantly reduced in mice treated with HEG at both the medium (M) and high (H) doses, which were 1-fold and 2-fold of the clinical equivalent dose, respectively. The tumor inhibition rates were 47.75% and 56.49%, respectively (Fig. 1B). Furthermore, the administration of HEG did not significantly impact the weight of mice (Fig. 1C) and did not cause notable toxic injury to vital organs, including the heart, liver, spleen, lungs, and kidneys (Fig. 1D). Thus, HEG effectively suppressed tumor development in the H22-luc orthotopic HCC mouse model while maintaining a favorable safety profile. The bacterial community depletion rate exceeded 90% following treatment of HCC model mice with antibiotic (ABX), confirming the successful establishment of the pseudo-sterile mouse model (Supplementary Fig. S1A and 1B). Results from the bacterial community depletion experiments (Fig. 1E) indicated that, compared to the Model group, both tumor fluorescence intensity and tumor weight in the HEG group were significantly reduced, yielding a tumor inhibition rate of 50.34%. No significant differences in tumor fluorescence intensity or weight were observed when comparing the ABX group to the ABX + HEG group (Fig. 1F–1H). H&E staining results demonstrated that a significant clustering of HCC cells was observed within the tumor tissue of Model group, with nuclear morphology of the tumor cells markedly differing from that of normal hepatocytes, an effect that was mitigated by the administration of HEG. However, no marked variations were observed in the ABX + HEG group in comparison with the ABX group. As revealed by immunohistochemical staining results, the proliferation-related markers Ki67 and PCNA had considerably reduced levels in tumor tissues of mice in HEG versus the Model groups. While Ki67 and PCNA levels were not significantly different in ABX + HEG versus ABX groups (Fig. 1I and 1J). The experimental results of fecal microbiota transplantation (FMT) (Fig. 1K) indicated that, in comparison with the FMT-Model group, FMT-HEG-treated mice exhibited both notable reductions in tumor fluorescence intensity and substantial decreases in tumor weight, resulting in a tumor inhibition rate of 33.63%, which was comparable to the anti-HCC effect observed in the HEG (L) group (Fig. 1L–1N). H&E staining results revealed that the aggregation of tumor cells was diminished in the FMT-HEG group compared to the FMT-Model group, and the nuclear morphology was improved. Additionally, immunohistochemical staining revealed marked reductions in the expression of the proliferation markers Ki67 and PCNA within the FMT-HEG group versus the FMT-Model group (Fig. 1O and 1P). Therefore, the inhibitory effect of HEG against HCC can be significantly diminished under conditions of bacterial community depletion. Moreover, the transplantation of gut microbiota regulated by HEG exerted certain anti-HCC effects. In conclusion, the gut microbiota was one of the contributors to HEG’s inhibitory effect against HCC.

Fig. 1. Inhibition of orthotopic HCC development by HEG depended on the presence of gut microbiota.

A Experimental design of the anti-HCC efficacy experiment of HEG. B Liver tumor weights of mice in the Model (double-distilled water, once a day, i.g.), HEG low-dose (4 g/kg, once a day, i.g.), HEG medium-dose (8 g/kg, once a day, i.g.), HEG high-dose (16 g/kg, once a day, i.g.), and positive drug cisplatin (2 mg/kg, once every three days, i.p.) groups, one-way ANOVA. C Body weights of mice in Sham, Model, Cisplatin, HEG (L), HEG (M), and HEG (H) groups. D H&E staining results of HCC tissues of mice in Sham, Model, Cisplatin, HEG (L), HEG (M), and HEG (H) groups, scale bar: 50 μm. E Experimental design of bacterial community depletion. F, G In vivo imaging results of mice in Model, HEG, ABX, and ABX + HEG groups, one-way ANOVA. H Liver tumor weights of mice in Model, HEG, ABX, and ABX + HEG groups, one-way ANOVA. I, J H&E staining results of HCC tissues and immunohistochemical staining results of proliferation markers Ki67 and PCNA in mice of Model, HEG, ABX, and ABX + HEG groups, scale bar: 50 μm, one-way ANOVA. K Experimental design of the fecal microbiota transplantation (FMT) experiment. L, M Results of in vivo imaging of mice in FMT-Model and FMT-HEG groups, unpaired Student t test. N Liver tumor weight of mice in FMT-Model and FMT-HEG groups, unpaired Student t test. O, P H&E staining results for HCC tissues and immunohistochemical staining results of proliferation markers Ki67 and PCNA in FMT-Model and FMT-HEG mice, scale bar: 50 μm, unpaired Student t test. n = 6. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. ABX antibiotic, FMT fecal microbiota transplantation, HCC hepatocellular carcinoma, HEG Huaier granule, PCNA proliferating cell nuclear antigen, Luc luciferase.

Regulation of gut microbiota by HEG in mice with orthotopic HCC model

16S rRNA sequencing was carried out for further explore the modulatory impacts of HEG on gut microbiota. From Fig. 2A, the number of mouse-specific amplicon sequence variants (ASVs) was significantly higher in HEG versus Model groups, revealing that HEG treatment successfully altered gut microbiota community in orthotopic HCC mice, introducing or amplifying specific ASVs. At the phylum level, the gut microbiota of mice was predominantly composed of Bacteroidota and Bacillota. Compared with the Sham group, the abundances of Desulfobacterota and Patescibacteria were significantly reduced in the Model group, and this alteration was reversed by HEG intervention. Furthermore, HEG effectively reduced the relative abundance of Campilobacterota. At a genus level, compared with the Sham group, the Model group showed a reduction in genera such as Odoribacter, Rikenella, Alistipes, and Mucispirillum. Administration of HEG reversed these changes, and reduced the proportional abundance of Helicobacter and elevated the proportional abundance of Akkermansia (Fig. 2B). Additionally, at a phylum level, the Bacteroidota to Bacillota ratio was significantly higher in HEG group (Fig. 2C). At a genus level, the Faecalibaculum was markedly lowered in Model group; nevertheless, HEG slightly overturned this change, helping to maintain the balance of gut microbiota. Furthermore, Parabacteroides was markedly predominant in HEG group, whereas Desulfovibrio and Enterorhabdus were substantially reduced (Fig. 2D). The genus-level ternary phase diagram indicated that the HEG group was predominantly enriched in Alloprevotella, Akkermansia, Clostridia_UCG-014, and Lachnospiraceae_NK4A136_group (Fig. 2E). The alpha diversity results demonstrated that the richness and diversity of the microbiota in HEG group exhibited a closer resemblance to that of Sham group, suggesting that the administration of HEG could help regulate the balance of gut microbiota, although no significant differences were observed (Fig. 2F). PCoA demonstrated substantial disparities in microbial composition across the three sample groups (Fig. 2G). LEfSe analysis, with a threshold of LDA > 3, showed that the dominant genera in the HEG group included Coriobacteriia, Coriobacteriales, Coriobacteriaceae_UCG-002, Butyricoccaceae, Clostridium_sp, and Clostridiales_bacterium (Fig. 2H). MetaStat analysis of the Model and HEG groups identified six main differential genera: Adlercreutzia, Clostridia-UCG-014, Coriobacteriaceae-UCG-002, Eubacterium-xylanophilum-group, Bilophila, and Allobaculum. Notably, the Adlercreutzia and Clostridia_UCG-014 were significantly increased, while the Bilophila were significantly decreased (Fig. 2I). Therefore, HEG optimized the gut microbiota community toward a beneficial and stable equilibrium.

Fig. 2. Regulation of gut microbiota by HEG in mice with orthotopic HCC model.

A Venn diagram of ASVs of gut microbiota of HCC mice in Sham, Model, and HEG groups. B Bar charts depicting microbial abundance at phylum and genus taxonomic levels for intestinal microbiota of HCC mice in Sham, Model, and HEG groups. C Histogram of proportional abundance of Bacteroidota and Bacillota, along with their ratio, in the intestinal microbiota of HCC-bearing mice in Sham, Model, and HEG groups, one-way ANOVA. D Histogram of relative abundance of Faecalibaculum, Parabacteroides, Desulfovibrio, and Enterorhabdus in gut microbiota of HCC mice in Sham, Model, and HEG groups, one-way ANOVA. E Ternary phase diagram showing gut microbiota of HCC mice from Sham, Model, and HEG groups at a genus level. F Alpha diversity analysis results of gut microbiota of HCC mice from Sham, Model, and HEG groups, including chao 1, simpson, and shannon indices, Kruskal-Wallis H test. G PCoA analysis of gut microbiota of HCC mice from Sham, Model and HEG groups. The ellipses represented 95% confidence intervals based on principal coordinate analysis. H LEfSe analysis of the gut microbiota of HCC mice in Sham, Model, and HEG groups, including histograms of LDA value distribution and evolutionary branching diagrams. I Relative abundance of differential genera in MetaStat analysis of gut microbiota in HCC mice in Model and HEG groups, including Adlercreutzia, Clostridia_UCG-014, Coriobacteriaceae-UCG-002, Eubacterium-xylanophilum-group, Bilophila, and Allobaculum, Mann-Whitney U test. n = 6. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. ASVs ampliconsequence variants, LDA linear discriminant analysis, LEfSe linear discriminant analysis effect size, PCoA principal co-ordinates analysis.

HEG repaired the intestinal barrier and improved colon immunity

H&E staining of mouse colon tissues revealed that the intestinal epithelial cells in the Model group exhibited detachment and loosened cellular junctions relative to the Sham group, indicating damage to the intestinal barrier. Additionally, the number of mucus-secreting goblet cells was reduced, and their morphology was altered. Administration of HEG alleviated these changes in the colon. In contrast, neither cell morphology nor intercellular junction was significantly changed in the colon of mice from ABX + HEG versus ABX groups. Furthermore, the colon cell and intercellular junction morphologies in the FMT-HEG group more closely resembled normal conditions than those in the FMT-Model group (Fig. 3A). Immunohistochemical staining demonstrated that the abundance of intestinal barrier-associated proteins ZO-1 and Claudin-1 of the HEG group apparently elevated relative to the Model group. Compared with the ABX group, the expression levels of ZO-1 and Claudin-1 showed no significant change in ABX + HEG group. Additionally, their protein levels in the FMT-HEG group were dramatically up-regulated in comparison with FMT-Model group (Fig. 3B, 3C). Therefore, HEG may exert anti-HCC effects by repairing the intestinal barrier, potentially involving intestinal microbiota.

Fig. 3. HEG repaired the intestinal barrie and improved colon immunity.

A H&E staining results of mouse colon in Sham, Model, HEG, ABX, ABX + HEG, FMT-Model, and FMT-HEG groups, scale bar: 500 µm, 200 µm. n = 3. B, C Immunohistochemical staining results of ZO-1 and Claudin-1 in mouse colon tissues in the Sham, Model, HEG, ABX, ABX + HEG, FMT-Model, and FMT-HEG groups, scale bar: 200 µm, one-way ANOVA. n = 3. D Cytometric bead array results for IL-6, IL-10, IL-1β, TNF-α, M-CSF, IL-17/17 A, and IFN-γ within colon tissues of Model, HEG, FMT-Model and FMT-HEG groups, unpaired Student t test or Mann-Whitney U test. n = 6. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. ZO-1: zonula occludens-1.

IL-6, IL-1β, TNF-α, and IL-17/17A are recognized as pro-inflammatory factors, and their reduction can inhibit the inflammatory response and slow the progression of HCC^12,13. Conversely, IL-10 may facilitate immune escape in HCC by suppressing immune cell activation¹⁴. M-CSF is implicated in inflammation and immune responses, and the inhibition of CSF-1R has become a focal point in numerous preclinical cancer studies^15,16. IFN-γ modulates neutrophil activity and consequently protects inflammatory tissues^17,18. A flow multifactor assay conducted on mouse colon tissues revealed that the administration of HEG significantly decreased IL-6, IL-10, IL-1β, TNF-α, M-CSF, and IL-17/17 A levels while elevating the level of IFN-γ relative to the Model group (Fig. 3D). Additionally, the cytokine concentrations in colonic tissues of the FMT-HEG group exhibited similar changes to HEG group when compared with the FMT-Model group (Fig. 3D). Therefore, HEG may exert anti-HCC effects by enhancing colon immunity, in which gut microbiota has an important function.

HEG enhanced the immune microenvironment in HCC by inhibiting the MAPK signaling pathway

Next, we utilized flow cytometry to demonstrate that the administration of HEG substantially enhanced CD4⁺ T cell, CD8⁺ T cell, and macrophage abundances in mouse HCC tissues, while concurrently reducing the percentage of M2 macrophages (Fig. 4A, 4B). Furthermore, results from multiple fluorescence immunohistochemical assays indicated the elevated CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, and M1 macrophage abundances, alongside a reduction in M2 macrophages within HCC tissues of HEG versus Model groups (Fig. 4C, 4D). Additionally, using cytometric bead array analysis, we found that concentrations of IL-6, IL-10, TNF-α, IL-17A, IL-13, and CCL2 were notably reduced in HEG group HCC tissues compared to Model group. Moreover, in the FMT-HEG group, the levels of IL-6, IL-10, TNF-α, IL-17/17 A, IL-1β, and M-CSF drastically decreased, while the quantities of IFN-γ was notably enhanced when compared to FMT-Model group (Fig. 4E). Notably, IL-6, TNF-α, IL-17A, CCL2, IL-1β, and M-CSF are known to promote tumor cell proliferation and metastasis^19–23. Conversely, IL-10 and IL-13 are implicated in HCC progression and are associated with tumor immune evasion^24,25. In contrast, IFN-γ is recognized for its role in inhibiting tumor cell proliferation and enhancing immune system recognition and clearance of tumor cells²⁶. Inhibition of the MAPK signaling pathway influences the immune cell quantity and functionality within the immune microenvironment of HCC, as well as regulate cytokine levels located within the neoplastic microenvironment. This immunomodulation potentiated the anti-tumor immune response^27,28. As illustrated in Fig. 4F and 4G, the administration of HEG significantly decreased p-JNK, p-ERK, and p-p38 protein levels, thereby inhibiting the MAPK signaling pathway in mouse HCC tissues. Furthermore, fecal transplantation experiment demonstrated that the gut microbiota, regulated by HEG, also facilitated the suppression of the MAPK signaling cascade in mouse HCC tissues (Fig. 4F and 4G). Correlation analysis between differential flora and various indices, including tumor weight, tumor fluorescence intensity, cytokine levels, and immune cell content, revealed significant associations between certain indices and the differential flora (Fig. 4H). This further underscored the relationship of anticancer efficiency of HEG with the gut microbiota modulation. Therefore, HEG inhibited the MAPK signaling pathway by modulating gut microbiota, subsequently affecting immune cell content and cytokine levels in the tumor immune microenvironment, thereby exerting an anti-HCC effect.

Fig. 4. HEG enhanced the immune microenvironment in HCC by inhibiting the MAPK signaling pathway.

A, B Detection of CD4⁺ T cells, CD8⁺ T cells, M1 and M2 macrophages in mouse HCC tissues from Model and HEG groups through flow cytometry, unpaired Student t test. n = 3. C, D CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, M1 and M2 macrophages were identified using multiple fluorescence immunohistochemistry for HCC tissues from Model and HEG groups, scale bar: 50 μm, unpaired Student t test. n = 3. E Concentrations of IL-6, IL-10, TNF-α, IL-17/17 A, IL-13 and CCL2 were detected with the cytometric bead array within HCC tissues from Model and HEG groups, and the abundances of IL-6, IL-10, TNF-α, IL-17/17 A, IL-1β, M-CSF and IFN-γ inside HCC tissues of FMT-Model and FMT-HEG groups, unpaired Student t test or Mann-Whitney U test. n = 6. F–G MAPK pathway-related protein levels (p-JNK, p-ERK and p-p38) in HCC tissues of mice from Model, HEG, FMT-Model and FMT-HEG groups was detected by immunoblotting, unpaired Student t test. n = 3. H Spearman correlation analysis of differential colonies at phylum and genus levels with tumor weight, tumor fluorescence intensity, cytokine levels and immune cell content. n = 6. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. ERK: extracellular regulated protein kinase; JNK: c-Jun N-terminal kinase.

HEG enhanced Equol content in fecal metabolites of mice with orthotopic HCC model

To elucidate the key metabolic mediators through which HEG exerted its anti-HCC effects by modulating gut microbiota, we integrated 16S rRNA sequencing with metabolomics data, combined with literature mining and in vitro functional validation, to systematically screen for critical differential metabolites (Fig. 5A). Mouse fecal metabolomics indicated that the overall metabolite compositions were significantly different in Model versus HEG groups under both positive and negative ion modes (Fig. 5B). In positive ion mode, 143 differential metabolites were identified (108 up-regulated and 35 down-regulated) (Fig. 5C), while 75 were identified in negative ion mode (57 up-regulated and 18 down-regulated) (Fig. 5D). KEGG pathway analysis indicated that these differential metabolites were primarily involved in pathways such as nicotinate and nicotinamide metabolism, unsaturated fatty acid biosynthesis, and tryptophan metabolism (Fig. 5E, 5F). Further Spearman correlation analysis between six gut bacterial genera significantly regulated by HEG and the above metabolites revealed significant associations (Fig. 5G, 5H), suggesting that the anti-HCC effect of HEG may be mediated via microbiota-metabolite axis. Among these genera, the abundances of Adlercreutzia (which possesses anti-inflammatory and immunomodulatory functions) and Clostridia_UCG-014 (which promotes short-chain fatty acids (SCFAs) production) were increased, while Bilophila (which produces H₂S) was significantly reduced. Based on literature review, among the 25 core differential metabolites, 4 have demonstrated clear anti-tumor activity^29–32, and 9 have been confirmed to be associated with gut microbiota abundance^33–41. Only Equol and 5-Deoxy-5-(methylthio) adenosine (MTA) satisfied both criteria, and their levels in feces were significantly elevated after HEG administration (Fig. 5I and Supplementary Fig. S2A). Considering that impaired intestinal barrier function might facilitate the direct action of metabolites on the liver via the gut-liver axis, we evaluated their in vitro inhibitory activity against human HCC HepG2 cells. The results showed that Equol significantly inhibited HepG2 cell proliferation in a dose-dependent manner, whereas MTA exhibited only a weak inhibitory effect within the same concentration range (Supplementary Fig. S2B). Therefore, Equol was selected as the key effector metabolite. Equol exists as two enantiomers, R and S³⁰. We separately tested the inhibitory effects of racemic Equol, S-equol, and R-equol on the proliferation of human HCC HepG2 and Bel7402 cells. The results demonstrated that all three significantly inhibited the proliferation of both cell lines in a concentration-dependent manner, with comparable potency (Supplementary Fig. S2C). Given that the S-equol is the predominant form of Equol in the human body⁴², subsequent research focused on the anti-HCC effects of S-equol.

Fig. 5. HEG enhanced Equol content in fecal metabolites of mice with orthotopic HCC model.

A Flowchart for Screening Key Differential Metabolites. B PLS-DA analysis results of overall metabolite composition in the Model and HEG groups under positive and negative ion modes. C, D Heatmaps of differential metabolites in the Model and HEG groups in positive and negative ionization modes. E, F Pathway enrichment analysis via KEGG was conducted on metabolites with significant abundance changes in Model and HEG groups, in both positive and negative ion detection modes. G, H Correlation analysis of differential genera and differential metabolites in positive and negative ion modes. I Structural formula of the key differential metabolite Equol, and its content in the feces of mice in the Model and HEG groups, unpaired Student t test. n = 6. ^*P < 0.05 and ^**P < 0.01. PLS-DA partial least squares discrimination analysis.

Differential metabolite Equol enhanced the immune microenvironment of HCC by suppressing the MAPK signaling cascade

Next, the effect of the differential metabolite Equol on the anti-HCC efficacy of HEG was analyzed. As illustrated in Fig. 6A, we constructed an H22-luc orthotopic HCC model and administered HEG along with S-equol. After 12 days of treatment, S-equol significantly suppressed tumor development in orthotopic HCC model mice. As indicated by results from small animal live imaging, tumor fluorescence intensity of S-equol group remarkably declined in a manner dependent on dose when compared with Model group (Fig. 6B and 6C). The tumor inhibition rates of low, medium, and high dose S-equol groups were 11.37%, 47.24%, and 53.33%, respectively (Fig. 6D). H&E staining results demonstrated a reduction in tumor cell aggregation and an improvement in nuclear morphology in the tumor tissues of mice receiving S-equol treatment relative to Model group (Fig. 6E). Furthermore, S-equol administration made no significant difference to mouse body weight (Supplementary Fig. S3A), nor did it induce notable toxic damage to vital organs, like the heart, liver, spleen, lungs, and kidneys (Supplementary Fig. S3B). It is important to note that mice fed AIN 93 M diet exhibit impaired endogenous Equol synthesis due to a lack of raw materials necessary for its production. Thus, a mouse model with deficient endogenous Equol synthesis can be established by feeding mice AIN 93 M diet⁴³. Utilizing the aforementioned mouse model (Fig. 6A), we further explored the impact of Equol on the anti-HCC efficacy of HEG. As illustrated in Fig. 6B–6E, compared to normal diet, the AIN 93 M diet significantly mitigated the anti-HCC effects of HEG on tumor progression in murine HCC model. Consequently, Equol demonstrated notable in vivo anti-HCC efficacy and was involved in the anti-HCC benefits of HEG. Furthermore, flow cytometry analyses revealed that S-equol administration elevated the CD4⁺ T, CD8⁺ T cell populations, upregulated M1 macrophages, and reduced the M2 macrophage proportion inside the mouse HCC tissues (Fig. 6F and 6G). Results from multiple fluorescence immunohistochemistry assays indicated an increase in CD3⁺ T, CD4⁺ T, CD8⁺ T cells, as well as M1 macrophages, alongside a decrease in M2 macrophages in the HCC tissues of the S-equol group versus Control group (Figs. 6H and 6I). Immunoblotting experiments demonstrated that S-equol administration downregulated the abundances of phosphorylated JNK, ERK, and p38 proteins, thereby inhibiting the MAPK signaling pathway in mouse HCC tissues (Figs. 6J and 6K). Collectively, as a participant in the anti-HCC effects of HEG, Equol enhanced HCC treatment by inhibiting the MAPK signaling pathway and improving the immune microenvironment.

Fig. 6. Differential metabolite Equol enhanced the immune microenvironment of HCC by suppressing the MAPK signaling cascade.

A Experimental design of the S-equol anti-HCC efficacy trial and the AIN 93 M diet intervention trial. B, C Small animal in vivo imaging results of mice in Model (AIN 93 M) group (AIN 93 M diet; double-distilled water, once a day, i.g.), HEG (AIN 93 M) group (AIN 93 M diet; HEG 8 g/kg, once a day, i.g.), Model (Normal) group (Normal diet; double-distilled water, once a day, i.g.), HEG (Normal) group (Normal diet, HEG 8 g/kg,, once a day, i.g.), S-equol (L) group (AIN 93 M diet, S-equol 10 mg/kg, once a day, i.g.), S-equol (M) group (AIN 93 M diet, S-equol 20 mg/kg, once a day, i.g.), S-equol (H) group (AIN 93 M diet, S-equol 40 mg/kg, once a day, i.g.), and cisplatin group (2 mg/kg, once every three days, i.p.), one-way ANOVA, n = 6. D The liver tumor weight of mice in Model (AIN 93 M), HEG (AIN 93 M), Model (Normal), HEG (Normal), S-equol (L), S-equol (M), S-equol (H) and Cisplatin groups, one-way ANOVA. n = 6. E H&E staining results of HCC tissues of mice in Model (AIN 93 M), HEG (AIN 93 M), Model (Normal), HEG (Normal) and S-equol (H) groups, scale bar: 50 μm. n = 6. F, G Detection of CD4⁺ T cells, CD8⁺ T cells, M1 and M2 macrophages in mouse HCC tissues from Control (Model (AIN 93 M)) and S-equol (S-equol (H)) groups by flow cytometry, unpaired Student t test. n = 3. H, I Measurement of CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, macrophages, M1 and M2 macrophages in HCC tissues of mice in Control and S-equol groups by multiplex fluorescence immunohistochemistry, scale bar: 50 μm, unpaired Student t test or Mann-Whitney U test. n = 3. J, K Expression of MAPK pathway-related proteins (p-JNK, p-ERK and p-p38) in HCC tissues of mice from Control and S-equol groups was measured through immunoblotting, unpaired Student t test. n = 3. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001.

Differential metabolite Equol induced G₀/G₁ phase arrest of HCC cells by inhibiting the Cyclin E1-CDK2/Rb signaling pathway

Immunohistochemical staining of mouse HCC tissues revealed that Ki67 and PCNA expression, key indicators of cell proliferation, was markedly reduced in the S-equol group versus Model (AIN 93 M) group (Fig. 7A, 7B). This indicated that S-equol effectively inhibited the proliferative capacity of mouse HCC tissues. Furthermore, the immunohistochemical staining results for Ki67 and PCNA were consistent with the findings of previous experiments (Fig. 7A, 7B), further demonstrating that the anti-HCC effect of HEG was closely correlated with Equol. Additionally, S-equol significantly reduced the proliferative capacity of human HCC HepG2 and Bel7402 cells concentration- and time-dependently (Fig. 7C). IC₅₀ values of S-equol for HepG2 and Bel7402 cells were 14.28 μM and 13.03 μM, respectively, following treatment for 48 h. As demonstrated by EdU proliferation assay, S-equol markedly reduced HepG2 and Bel7402 cell proliferation (Fig. 7D, 7E). Moreover, S-equol significantly inhibited the colony-forming ability of HepG2 and Bel7402 cells (Fig. 7F, 7G) and also significantly reduced ATP production in these cells (Fig. 7H). The results indicated that S-equol significantly inhibited the growth of human HCC cells. Additionally, RNA sequencing results suggested that the differential gene KEGG enrichment analysis primarily focused on the cell cycle (Fig. 7I). We analyzed the effect of S-equol on cell cycle of human HCC cells using flow cytometry. S-equol induced a G₀/G₁ phase arrest in human HCC HepG2 and Bel7402 cells (Fig. 7J, 7K). The Cyclin E1-CDK2/Rb signaling pathway is important for G₀/G₁ phase⁴⁴. Then, the influence of S-equol on the Cyclin E1-CDK2/Rb signaling cascade in human HCC cells was explored. As shown in Fig. 7L and 7M, S-Equol effectively downregulated the abundances of Cyclin E1, CDK2, and p-Rb proteins in HepG2 and Bel7402 cells, as well as in mouse HCC tissues (Fig. 7N, 7O). Additionally, to evaluate the selectivity of Equol, we examined its effects on the growth of human normal liver LO2 cells. The results indicated that Equol had no significant impact on the proliferation and cell cycle progression of LO2 cells (Supplementary Fig. S4). Therefore, Equol induced G₀/G₁ phase block in human HCC cells by inhibiting the Cyclin E1-CDK2/Rb signaling pathway, demonstrating favorable selective specificity.

Fig. 7. Differential metabolite Equol induced G₀/G₁ phase arrest of HCC cells by inhibiting the Cyclin E1-CDK2/Rb signaling pathway.

A, B Immunohistochemical staining results of proliferation markers Ki67 and PCNA in mouse HCC tissues from Model (AIN 93 M), HEG (AIN 93 M), Model (Normal), HEG (Normal) and S-equol (H) groups, scale bar: 50 μm, one-way ANOVA. n = 3. C After treatment using 0, 5, 10, 15, 20, 25 μM S-equol, viability of human HCC HepG2 and Bel7402 cells was analyzed through CCK8 assay at 24, 48, and 72 h. D, E HepG2 and Bel7402 cells were exposed to 48 h of 0, 10, 20 μM S-equol treatment. Beyo Click^TM EdU-488 was added for cell labeling, scale bar: 200 μm, one-way ANOVA. F, G HepG2 and Bel7402 cells were exposed to 8 days of 0, 5, 10 μM S-equol treatment to observe colony formation, one-way ANOVA. H HepG2 and Bel7402 cells were exposed to 24 h of 0, 10, 20 μM S-equol treatment to measure ATP content, one-way ANOVA. I HepG2 cells were exposed to 24 h of 0, 20 μM S-equol treatment to perform RNA sequencing and KEGG enrichment to analyze differential pathways. J, K HepG2 and Bel7402 cells were exposed to 24 h of 0, 10, 20 μM S-equol treatment to detect cell cycle by flow cytometry, one-way ANOVA. L, M HepG2 and Bel7402 cells were exposed to 48 h of 0, 10, 20 μM S-equol treatment to detect protein levels of Cyclin E1, CDK2 and p-Rb by immunoblotting, one-way ANOVA. N, O Protein levels of Cyclin E1, CDK2 and p-Rb were detected with the aid of immunoblotting within hepatic cancer tissues of murine models in Control (Model (AIN 93 M)) and S-equol (S-equol(H)) groups, unpaired Student t test. n = 3. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. CDK2 cyclin dependent kinase 2, Rb retinoblastoma.

Discussion

Most Chinese herbal medicines are administered orally, which inevitably leads to interactions with gut microbiota⁴⁵. Therefore, the investigation of how Chinese medicines exert anti-HCC effects by regulating gut microbiota is of considerable importance. These medicines alter the community composition and structural architecture of the intestinal microbiota, coupled with the abundances of host-derived endogenous metabolites are modulated, thereby regulating the gut barrier function, immune system, and systemic metabolism, ultimately resulting in significant anticancer effects⁴⁶. HEG is a Class I innovative anti-cancer drug in China, having been included in the “Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2022 Edition)”. Both clinical and preclinical studies have demonstrated that HEG posses notable anticancer properties¹⁰. HEG enhanced the immune function of patients by modulating cytokine levels, thereby achieving improved therapeutic effects against HCC⁴⁴. However, research on the anti-HCC effects of HEG mediated through gut microbiota modulation is currently limited. Herein, we observed that HEG significantly suppressed neoplastic progression in an orthotopic HCC murine model implanted with H22-luc cells while demonstrating a favorable safety profile. Bacterial community depletion and fecal microbiota transplantation experiments demonstrated that the gut microbiota was one of the factors through which HEG exerted its anti-HCC effects. It is important to note that in most current animal FMT studies, the preparation of donor microbiota cannot fully eliminate potential residues of the drug or its metabolites^47–49. Furthermore, the in vivo pharmacokinetic profile of the main active component of HEG, proteoglycan, remains unclear. Whether it undergoes biliary excretion, leading to drug residues that may subsequently affect FMT outcomes, also requires further elucidation. Despite these limitations, in conjunction with the results from the bacterial community depletion experiments, this study supported the conclusion that the gut microbiota was one of the contributing factors to the anti-HCC effects of HEG.

Subsequently, using 16S rRNA sequencing, we observed that the HEG group exhibited a greater number of unique ASVs. Compared to the Model group, HEG significantly increased the Bacteroidota/Bacillota ratio and the relative abundances of Akkermansia and Faecalibaculum, while reducing the relative abundances of Campilobacterota and Helicobacter. It is known that transplantation of a microbiota with a reduced Bacteroidota/Bacillota ratio exacerbates liver injury and fibrosis in recipient mice⁵⁰. Both Akkermansia and Faecalibaculum have been reported to exert anti-tumor efficacy^51,52. The abundances of Campilobacterota and Helicobacter are positively correlated with the severity of HBV-induced liver disease^53,54. Notably, Akkermansia was not detected in the Sham group, suggesting that its abundance may be highly dependent on the host’s physiological or pathological state. Under conditions of intestinal barrier impairment induced by HCC, the emergence of Akkermansia may represent an adaptive repair response by the host, a process that was further enhanced by HEG intervention. This finding aligns with the results reported by Cao et al.⁵⁵. Furthermore, although Helicobacter is generally considered a pathogenic bacterium, its abundance did not differ between the Model and Sham groups in this study, reflecting the baseline characteristics of the experimental animals rather than acting as a disease driver⁵⁶. Additionally, the decrease in the abundance of Lactobacillus in the HEG group may stem from competitive niche shifts during microbial restructuring, reflecting a specific modulation of the microbiota composition by HEG rather than a generalized promotion of all probiotic bacteria⁵⁷. Furthermore, butyrate-producing bacteria such as Butyricicoccaceae, Clostridium_sp, and Clostridiales_bacterium were markedly enriched in HEG group. Further analysis revealed that the primary differential genera between Model and HEG groups included Adlercreutzia, Clostridia_UCG-014, and Bilophila. It is known that Adlercreutzia exhibits significant anti-inflammatory properties both in vivo and in vitro by inhibiting the NF-κB signaling pathway and reducing the production of the inflammatory cytokine IL-6⁵⁸. Moreover, it is a key bacterial genus involved in the synthesis of Equol⁵⁹, which provides an important clue for subsequent analysis of the metabolic mechanisms. Additionally, Clostridia_UCG-014 is associated with the production of SCFAs⁶⁰, while Bilophila can produce potentially toxic H₂S⁶¹. Therefore, HEG intervention did not simply restore the basal microbial state of Sham group, but rather optimized the gut microbiota community to establish a stable community, thereby influencing host physiology.

Communication of gut microbiota with HCC is mediated via gut-liver axis⁶. When the bowel barrier is compromised due to dysbiosis, an influx of bacteria and their metabolites into the lamina propria disrupts the status and function of intestinal immune cells⁶². Subsequently, these microbial communities and their metabolic byproducts migrate to the hepatic organ through the portal venous circulation, activating specific immune cells, promoting inflammatory factor production, and accelerating HCC progression⁶³. In this study, we found that HEG alleviated intestinal barrier damage in mice with orthotopic HCC, up-regulated the expression levels of tight junction proteins, and modulated the levels of various cytokines, such as IL-6, IL-10, and IFN-γ, in colon tissues. Furthermore, HEG improved the HCC immune microenvironment by increasing the infiltration of CD4⁺ T cells, CD8⁺ T cells, and M1 macrophages, while reducing the levels of M2 macrophages and cytokines, including IL-6, IL-10, and TNF-α in tumor tissues. Notably, although M1 macrophages increased, the levels of the pro-inflammatory cytokines IL-6, TNF-α, and IL-17A decreased, suggesting that HEG may preferentially inhibit other pro-inflammatory signaling pathways, highlighting the complexity of its immunomodulatory effects. Previous studies have shown that activation of the MAPK signaling pathway promotes tumor immune escape from CD8⁺ T cells by regulating molecules such as CXCL8 and PD-L1, thereby suppressing CD4⁺ T cell infiltration^64,65. MAPK inhibitors promote the transition of tumor-associated macrophages from the M2 to the M1 phenotype and activate CD4⁺ T cells, CD8⁺ T cells, and the IFN-γ pathway, thereby enhancing anti-tumor immunity^27,28,66. Our results demonstrated that HEG significantly inhibited the MAPK signaling pathway in mouse HCC tissues. Additionally, FMT from HEG-treated donors exhibited anti-HCC mechanisms similar to direct HEG administration. A correlation was found between cytokine levels and differential flora following the administration of HEG. Therefore, via gut microbiota-gut-liver axis, HEG regulated the community of gut microbiota, hence exerting an anti-HCC effect through the repair of the intestinal barrier, enhancement of colonic immunity, and enhancing the immune microenvironment by inhibiting MAPK signaling pathway.

Metabolomic analysis revealed that following HEG intervention, differential metabolites were primarily enriched in pathways such as unsaturated fatty acid biosynthesis, tryptophan metabolism, as well as nicotinate and nicotinamide metabolism. By integrating 16S rRNA sequencing with metabolomics data, along with literature mining and in vitro validation, we identified Equol as a key metabolite through which HEG exerted anti-HCC effects by modulating the gut microbiota. Equol is produced by specific gut microbiota (e.g., Adlercreutzia) from soybean isoflavones⁷, and its level positively correlated with the abundance of Adlercreutzia. In an endogenous Equol-deficient model induced by AIN 93 M diet⁴³, the anti-HCC effect of HEG was significantly attenuated, indicating that Equol served as an important mediator of HEG’s efficacy. However, the multiple nutritional differences between AIN 93 M and normal diet may influence tumor progression or microbiota composition. Further studies, including Equol supplementation experiments and quantification of Equol levels in plasma, liver, and feces, are needed to elucidate its pharmacokinetics and dose-response relationship, thereby more rigorously establishing causality and systematically investigating the mechanisms underlying the HEG-microbiota-Equol axis. In vivo experiments further confirmed that Equol exhibited potent anti-HCC activity. Moreover, Equol not only enhanced the immune microenvironment of HCC by blocking the MAPK signaling cascade, but also suppressed HCC cell proliferation by inhibiting the Cyclin E1-CDK2/Rb signaling pathway and inducing G₀/G₁ phase blockage. Notably, Equol selectively suppressed the proliferation of HCC cells in vitro without exhibiting significant toxicity toward human normal liver cells, suggesting favorable selective specificity. This may be attributed to inherent differences between cancer cells and normal cells in terms of receptor expression, drug uptake efficiency, or intracellular redox status. Currently, the Equol-MAPK pathway-immune microenvironment axis in HCC was primarily supported by correlative analyses. Future studies should employ gain-of-function experiments to further validate its causal relationship and utilize methods such as molecular docking and surface plasmon resonance to identify the direct targets of Equol. In addition, the direct regulatory effect of Equol on macrophage polarization and its influence on other immune components should also be key focuses of follow-up research. To comprehensively elucidate the long-term efficacy and mechanisms of HEG, subsequent studies will employ spontaneous HCC models and germ-free animals to evaluate its impact on HCC progression and recurrence. Metagenomic and metatranscriptomic technologies will be applied to precisely identify the key bacterial species and functional genes responsive to HEG. Additionally, influencing factors such as diet, sex, and host genetics should be systematically investigated to facilitate its clinical translation.

In summary, via the gut microbiota-gut-liver axis, HEG regulated gut microbiota composition and elevated the level of the metabolite Equol, thereby exerting an anti-HCC effect through repairing the intestinal barrier, enhancing colon immunity, and improving the immune microenvironment of HCC by inhibiting the MAPK signaling pathway. Notably, the differential metabolite Equol demonstrated significant in vivo anti-HCC efficacy and was actively involved in the anti-HCC effects of HEG. It enhanced the immune microenvironment of HCC by inhibiting the MAPK signaling pathway, while also suppressed HCC cell growth by inducing the G₀/G₁ phase blockade through inhibition of the Cyclin E1-CDK2/Rb signaling pathway (Fig. 8). This study established the solid scientific foundation for applying HEG in the clinic, identifying Equol as a promising agent for HCC treatment.

Fig. 8. Schematic diagram of the anti-HCC effect of HEG via modulation of gut microbiota and its metabolite Equol through the gut-liver axis.

HEG exerted its anti-HCC effects by modulating the composition of the gut microbiota, which subsequently repaired the intestinal barrier, enhanced colonic immunity, and improved the immune microenvironment of HCC by inhibiting the MAPK signaling pathway. By integrating 16S rRNA sequencing with metabolomics data, alongside literature mining and in vitro validation, Equol was identified as a key differential metabolite through which HEG suppressed HCC via gut microbiota regulation. Equol was essential for the anti-HCC effects of HEG. Equol ameliorated the immune microenvironment by inhibiting the MAPK signaling pathway, while concurrently inhibited the growth of HCC cells by inducing the G₀/G₁ phase blockade through suppression of Cyclin E1-CDK2/Rb signaling pathway.

Methods

Reagents, antibodies, and drugs

Roswell Park Memorial Institute (RPMI-1640) (30-002-CI), Dulbecco’s Modified Eagle’s Medium (DMEM) (10-013-CV), Fetal bovine serum (FBS) (35-010-CV) were provided by Corning Life Sciences (Steuben County, New York, USA). Ampicillin sodium (A105484), neomycin sulfate (N109017), metronidazole (M109874) were obtained from Aladdin Biochemical Technology (Shanghai, China). Vancomycin hydrochloride (BN20412), 4% paraformaldehyde (BN20094)were obtained from Biorigin (Beijing, China). FastPure Bacteria DNA Isolation Mini Kit-BOX2 (DC103-01), Equalbit dsDNA HS Assay Kit (EQ111-01), Rapid Taq Master Mix (P222-01) were obtained in Vazyme Biotech (Nanjing, China). GAPDH (HRP conjugated) (ZB15004-HRP) was obtained from Servicebio (Wuhan, China). β-actin (sc-47778) antibody was provided by Santa Cruz Biotechnology (Santa Cruz, CA, USA). p-JNK (4668T), JNK (9252T), p-ERK (4370T), ERK (4695T), p-p38 (4511T), p38 (8690T) were obtained from Cell Signaling Technology (Danvers, MA, USA). APC anti-mouse CD45 (103112), PerCP/Cyanine5.5 anti-mouse CD3 (100217), PE/Dazzle™ 594 anti-mouse CD4 (100456), Brilliant Violet 650™ anti-mouse CD8a (100742), APC/Cyanine7 anti-mouse/human CD11b (101225), PE/Cyanine7 anti-mouse F4/80 (123113), PE anti-mouse CD86 (105007), FITC anti-mouse CD206 (MMR) (141703) were provided by Biolegend (CA, USA). Huaier granule (BC18) was purchased from Qidong Gaitianli Pharmaceutical (Jiangsu, China). S-equol (HY-100583) was provided by MedChem Express (Monmouth Junction, New Jersey, USA).

H22-luc orthotopic HCC model

Mouse HCC H22-luc cells (FH0473) were obtained from FuHeng Cell Center (Shanghai, China). The 6–8-week-old male C57BL/6 J mice were obtained in SPF Biotechnology (Beijing, China). Following induction of anesthesia with 3% isoflurane (inhaled) and maintenance with 2% isoflurane, a 10 μL suspension containing 1 × 10⁶ H22-luc cells was slowly injected into the liver⁶⁷. Six mice were allocated to each group, with treatments consisting of cisplatin (2 mg/kg, once every three days, i.p.)⁶⁸ and HEG (4 g/kg, 8 g/kg, 16 g/kg, once a day, i.g.). The medium dose of HEG (8 g/kg) was determined as the clinically equivalent dose based on the clinical dosage and the equivalent dose conversion coefficient method. The low and high doses were set at 0.5-fold and 2-fold of this medium dose, respectively⁶⁹. Mouse body weights of different groups were recorded, and small animal live imaging was conducted at the conclusion of the treatment period. For the euthanasia of mice, compressed CO₂ was administered into the chamber at a flow rate that displaced 30% to 70% of the chamber volume per minute. Subsequently, various tissues, including the tumor, colon, and organs, including the heart, liver, spleen, lungs, and kidney, were harvested. This animal experiment was conducted according to the ethical policies and procedures approved by the ethics committee of Beijing University of Chinese Medicine, China (Approval no. BUCM-2023031603-1039).

Bacterial community depletion experiment

The Qubit method was employed to quantitatively assess the DNA concentration in the feces of HCC model mice, while gel electrophoresis was utilized for the qualitative evaluation of bacterial community depletion, which was expected to reach a depletion rate of 90%^70,71. After the successful creation of the orthotopic HCC model, a suspension containing various antibiotics, specifically 200 mg/kg ampicillin, 200 mg/kg neomycin, 200 mg/kg metronidazole, and 100 mg/kg vancomycin, was administered for three consecutive days to create a pseudo-bacteria-free HCC mouse model⁷². Subsequently, HEG (8 g/kg, once a day, i.g.) and antibiotic cocktail was continuously administered. After 14 days of administration, small animal live imaging was performed, followed by euthanasia of the mice via CO₂ asphyxiation and collection of tissues such as tumor, and colon. This animal experiment was conducted according to the ethical policies and procedures approved by the ethics committee of Beijing University of Chinese Medicine, China (Approval no. BUCM-2023101001-4018).

Fecal microbiota transplantation (FMT)

Every day at 8:00 a.m., donor mice with liver cancer were administered either double-distilled water or HEG (8 g/kg). Fresh fecal samples from the respective mice were collected at 4:00 p.m. on the same day, which served as the fecal microbiota transplantation materials for the FMT-Model and FMT-HEG groups. A total of 100 mg of feces from each group was subjected to suspension within 1 mL saline, vortex, and centrifugation to collect the supernatant as the transplantation material. This material was then gavaged to the pseudo-bacteria-free HCC mice at a dosage of 0.2 mL/10 g of body weight every day^47–49. Ten days following FMT, small animal live imaging was conducted, followed by euthanasia of the mice via CO₂ asphyxiation and subsequent collection of various tissues, including tumor and colon, for analysis. This animal experiment was conducted according to the ethical policies and procedures approved by the ethics committee of Beijing University of Chinese Medicine, China (Approval no. BUCM-2023101001-4018).

Equol in vivo efficacy experiment

A total of 10 μL of a mouse HCC H22-luc cell suspension, containing 1×10⁶ H22-luc cells, was slowly injected into the livers of C57BL/6 J male mice. As a result, 48 mice were successfully established as orthotopic HCC model, with drug administration commencing in groups 7 days post-injection. The experimental groups included Model (AIN 93 M) group (AIN 93 M diet, double-distilled water, once a day, i.g.), HEG (AIN 93 M) group (AIN 93 M diet, HEG 8 g/kg, once a day, i.g.), Model (Normal) group (Normal diet, double-distilled water, once a day, i.g.), HEG (Normal) group (Normal diet, HEG 8 g/kg, once a day, i.g.), S-equol groups (AIN 93 M diet, S-equol 10 mg/kg, 20 mg/kg, 40 mg/kg, once a day, i.g.)^73,74, and cisplatin group (2 mg/kg, once every three days, i.p.). Mouse body weights were measured. At the conclusion of drug administration, small animal live imaging was conducted, followed by euthanasia of the mice via CO₂ asphyxiation. Subsequently, tumors, colon, and tissues from the heart, liver, spleen, lungs, and kidneys were collected. This animal experiment was conducted according to the ethical policies and procedures approved by the ethics committee of Beijing University of Chinese Medicine, China (Approval no. BUCM-2024051006-2109).

16S rRNA sequencing

Genomic DNA from mouse feces was isolated, and its concentration and purity were assessed with 1% agarose gel electrophoresis. Then DNA dilution to 1 ng/μL was completed. Primers targeting the 16S V4 region (515 F and 806 R) were designed, and a PCR amplification system was established to amplify the genomic DNA. The NEB Next Ultra™ II FS DNA PCR-free Library Prep Kit was employed for constructing a library, which was later quantified with Qubit and qPCR, followed by sequencing on the NovaSeq 6000 platform with a PE250 configuration. Subsequently, bioinformatics analyses were performed, which encompassed quality control of data, reduction of noise from amplified sequence variants, annotation of species, analysis of sample complexity, comparative analysis across samples, community divergence assessments, functional predictions, and association analyses. For data visualization of the PCoA plot, the R-ggplot2 was employed. The confidence ellipses in the scatter plot were generated using the default parameters of the stat_ellipse function to illustrate the distribution range of data within each group.

Non-targeted metabolomics

Mouse feces were ground into liquid nitrogen and subsequently centrifuged at low temperature. Thereafter, the supernatant was gathered and analyzed using LC-MS. Data files from downstream processing were leveraged to achieve both metabolite identification and relative quantification results by predicting molecular formulas from molecular ion peaks and fragmentation ions, comparing the outcomes against existing databases. The metabolites identified were annotated with the aid of KEGG, HMDB, and LIPIDMaps databases. The data underwent transformation via metaX, after which principal component analysis and partial least squares discriminant analysis (PLS-DA) were conducted, and later variable importance in projection (VIP) values for diverse metabolites were calculated. Through t-tests, significance levels for metabolites (P-value) were computed, while fold change (FC) for these metabolites was also assessed. The standard criteria used for screening differential metabolites were defined as follows: VIP > 1, P-value < 0.05, and FC ≥ 2 or FC ≤ 0.5.

Cytometric bead array

Weigh 0.1 g of tumor tissue or 0.04 g of colon tissue, and add 500 μL of PBS. Afterwards, proceed to break and lyse the tissue for a duration of 10 min, and then centrifugation at 4 °C at 5000 g for another 10 min to retrieve the supernatant. The protein concentration is then measured using the BCA method. First, 6–7 standard wells were prepared, then 50 μL standards or samples were added to each well, and later 50 μL particulate mixture was added. Incubate the wells on a horizontal orbital microtiter plate shaker at 800 ± 50 rpm and 4°C overnight, protected from light. After incubation, use a microplate magnetic device to adsorb the particles, and wash each well thrice with 100 μL wash buffer. Next, add 50 μL of biotin antibody cocktail to each well and incubate for 1 h at room temperature on a shaker, again protected from light, followed by three washes. Subsequently, introduce the diluted streptavidin-PE (50 μL) into every well and incubate at room temperature on a light-proof shaker for 30 min, followed by three washes. Finally, introduce washing buffer (100 μL) into every well, re-suspend particles, and incubate onto the shaker under ambient temperature for a 2-min duration. The test should be performed within 90 min, with the particles being re-suspended 2 min prior to reading.

Flow cytometry method

To prepare HCC tissue for analysis, the tissue was grinded and 4 mL of red blood cell lysate was added. After 5 min of centrifugation at 500 g, the supernatant was discarded. Cells were resuspended in PBS and 100 μL designated surface antibody premixes were added for 15 min of incubation in dark under ambient temperature: PBS, 100 μL; APC anti-mouse CD45, 3 μL; PerCP/Cyanine5.5 anti-mouse CD3, 3 μL; PE/Dazzle™ 594 anti-mouse CD4, 3 μL; Brilliant Violet 650™ anti-mouse CD8a: 3 μL; APC/Cyanine7 anti-mouse/human CD11b: 3 μL; PE/Cyanine7 anti-mouse F4/80: 3 μL; PE anti-mouse CD86: 3 μL. Upon completion of the incubation, PBS (1 mL) was applied to each well, followed by 5 min of centrifugation at 500 g to remove the supernatant. Following this, introduce 500 μL of 1× Fixation Buffer and incubate under ambient temperature for a 30-min duration away from light, followed by 5 min of centrifugation at 500 g to remove the supernatant. Introduce 1 mL membrane-breaking solution, and incubate under ambient temperature for a 10-min duration, and remove the supernatant after 5 min of centrifugation at 500 g. Then, introduce 100 μL the intracellular antibody premix, mix thoroughly, and incubate for 30 min at room temperature, protected from light. The components of the intracellular antibody premix include 100 μL of 1× membrane-breaking solution and 3 μL of FITC anti-mouse CD206 (MMR). Finally, introduce the 1× membrane-breaking solution (1 mL), remove the supernatant after 5 min of centrifugation at 500 g, and introduce 200 μL PBS before transferring the sample to an EP tube for online detection.

H&E staining

The tissues underwent immersion in a 4% paraformaldehyde solution, followed by sequential ethanol dehydration (50%, 70%, 80%, 95%, and 100%, 30 min each). After dehydration, all tissues were permeabilized with the 1:1 mixture of xylene and anhydrous ethanol for 1.5 h, then transferred to xylene for a total of two permeabilization cycles, lasting 3 h. The permeabilized tissues were then immersed in the 1:1 mixture of xylene and paraffin wax and dried at 40 °C for 40 min. Next, these tissues were immersed within paraffin I and paraffin II, each dried at 55 °C for 30 min. The wax-impregnated tissues were embedded in pure wax three times, drying at 60 °C for 1 h each time, before being placed on a microtome for slicing at a thickness of 5 μm. A slide coated with mucoadhesive was prepared, and the slices were floated on this mucoadhesive before drying at 40 °C overnight. The tissue sections were subjected to deparaffinization into xylene for 20 min, then treated for 10 min with a 1:1 mixture of xylene and anhydrous ethanol. They were subsequently washed with 100%, 95%, 85%, 70%, and 50% ethanol, each wash lasting 5 min, before being stained with hematoxylin for a duration of 10 min. Dehydration resumed in ethanol concentrations of 85% and 95% for another 5 minutes each, followed by a 5-minute staining with eosin. The sections were dehydrated again in 95% ethanol and anhydrous ethanol for 5 min each, and permeabilized with xylene twice for 5 min each, followed by neutral gum sealing. The tissue morphology was observed under a microscope, and images were recorded⁷⁵.

Immunohistochemical analysis

The sections were re-immersed in xylene two times, each immersion lasting 10 min. Next, the tissues were placed in 100%, 95%, and 70% ethanol solutions for 5 min each and subsequently washed by distilled water. The sections were subsequently transferred to a repair cassette filled with boiling sodium citrate buffer; the heating process was halted after 8 min, prior to heating for additional 5 min. After cooling, these sections were rinsed by PBS thrice, each wash lasting 5 min, then incubated in a 3% H₂O₂ solution, protected from light for a 20-min duration under ambient temperature, and rinsed by PBS thrice (5 min each). A sealing solution of 3% bovine serum albumin (BSA) was carefully poured for coating these tissues evenly and allowed to rest for a 30-min duration under ambient temperature. After aspiration of sealing solution, the primary antibody was applied dropwise to incubate the sections at 4 °C overnight. The sections were rinsed thrice by PBS, each wash lasting 2 min. The secondary antibody was then added dropwise to incubate these sections for 1 h at ambient temperature. Thereafter, all sections were rinsed with PBS thrice, each time for 2 min. The fresh DAB color development solution was introduced dropwise, followed by monitoring of color development time through microscopic observation. The color development process was concluded by rinsing the sections with distilled water. Hematoxylin staining was performed for 10 min, followed by washing with distilled water, soaking for 10 s in a 1% hydrochloric acid ethanol solution, rinsing again with distilled water, and then staining dropwise with 0.6% ammonia. Finally, dehydration was performed to seal the sections, and the tissue morphology was observed under the microscope, with images captured⁷⁵.

Multiplex fluorescence immunohistochemistry

Liver cancer paraffin sections were baked at 60 °C for 30 min, deparaffinized using xylene twice for 10 min each, and then hydrated using graded ethanol concentrations of 100%, 95%, 80%, and 70%, each time for 5 min. PBS was employed to rinse the sections repeatedly. They underwent heat-induced antigen retrieval by placing them in sodium citrate antigen retrieval solution, heating at 95 °C for 10 min, cooling to room temperature, and rinsing with PBS thrice (5 min each). These sections were enclosed at room temperature using 5% BSA for a duration of 30 min. Afterward, the prepared primary antibody was applied dropwise, allowed to incubate overnight at 4 °C, and three individual 5 min cycles of rinsing with PBS were carried out for each sample. The secondary antibody, which was fluorescently labeled, was also introduced dropwise to these sections for 1 h of incubation under ambient temperature, avoiding exposure to light. The labeling process was repeated with another fluorescent dye as per the experimental requirements, ensuring thorough washing after each staining to prevent cross-reactivity. DAPI was added for nuclear staining, followed by 10 min of incubation under ambient temperature and three washes using PBS for 5 min each. After drying under ambient temperature, the sections were sealed with an anti-fluorescence quenching sealer and subsequently stored at 4 °C. Ultimately, images were obtained utilizing a fluorescence microscope.

Cell culture

Human HCC HepG2, Bel7402 cells, and human normal liver LO2 cells were obtained from the Cell Culture Center of the Institute of Basic Medical Sciences of the Chinese Academy of Medical Sciences. HepG2 and LO2 cells were cultured in DMEM medium supplemented with 10% FBS and 1% penicillin-streptomycin, while Bel7402 cells were maintained in RPMI-1640 medium with 10% FBS and 1% penicillin-streptomycin. All cells were incubated at 37°C with 5% CO₂.

CCK-8 cell proliferation assay

Human HCC HepG2 and Bel7402 cells were collected, with their densities being adjusted to 2.0 × 10⁴ cells/mL in 96-well plates. Various concentrations of S-equol (0, 5, 10, 15, 20, and 25 μM) were applied to the cells for durations of 24, 48, and 72 h. Afterward, a CCK-8 working solution (Meilunbio, MA0218) was added to incubate cells under 37°C for a 2-h duration. Absorbance readings were taken at 450 nm, allowing for the calculation of cell viability.

EdU-488 cell proliferation assay

Human HCC HepG2 and Bel7402 cells were inoculated into 12-well plates at suitable densities and exposed to 0, 10, and 20 μM of S-equol for 48 h. The protocol outlined in the BeyoClick™ EdU-488 cell proliferation detection kit (Beyotime, C0071s) was meticulously followed. Subsequently, the 12-well plate was examined under an inverted fluorescence microscope for observation and imaging.

Cell colony formation experiment

Human HCC HepG2 and Bel7402 cells were seeded into 6 cm dishes at 800 cells/dish. The cells were exposed to 0, 5, and 10 μM S-equol treatment for a duration of 8 days. After the treatment period, all cells were washed, fixed, and stained using a crystal violet solution (10 g/L) for 30 min. Next, the cells were photographed and counted.

ATP detection

Human HCC HepG2 and Bel7402 cells were inoculated into 24-well plates at 1 × 10⁵ cells/well and incubated using 0, 10, and 20 μM S-equol for a 24-h duration. Thereafter, the cells underwent lysis, and the supernatant was harvested and subsequently centrifuged. Subsequently, 100 μL of the ATP assay working solution (Beyotime, S0026) was dispensed in the black 96-well plate with a transparent bottom. After this, 10 μL of either the sample or standard was introduced into every well, with RLU values being detected with the microplate reader.

Detection of cell cycle by flow cytometry

Human hepatocellular carcinoma HepG2 and Bel7402 cells were inoculated into 6 cm dishes at 2 × 10⁵ cells/dish and exposed to 24 h of 0, 10, and 20 μM S-equol treatment. Subsequently, these cells were collected into 15 mL centrifuge tubes, and centrifuged to discard the supernatant, followed by resuspension with 3 mL PBS. After another centrifugation to remove the supernatant, a droplet of 3 mL of 70% ethanol was added, and the samples were placed at −20°C for overnight fixation. The following day, the cells were centrifuged again to discard the supernatant, and 1 mL of PBS was added to each tube and the supernatant was removed by another centrifugation. Subsequently, 300 μL propidium iodide staining solution (Becton, Dickinson and Company, 550825) was introduced to each tube and maintained at 37°C for 30 min, protected from light. The analysis of the cells was performed using a flow cytometer, with resultant results being manipulated and quantitatively examined with Flow-Jo software.

RNA sequencing analysis

Human HCC HepG2 cells were inoculated into 6 cm dishes at 2 × 10⁵ cells/dish and subsequently exposed to 24 h of 0 or 20 μM S-equol treatment. Later, these cell samples were harvested and sent to Beijing Prheath Biotechnology Co. for RNA sequencing analysis. The Illumina HiSeq × Ten platform was used to conduct the process of RNA sequencing. The results obtained from the sequencing were analyzed via data mining techniques.

Immunoblotting

Tumor or colon tissues were homogenized in RIPA lysis solution, which was prepared with the addition of 1% PMSF and 1% phosphatase inhibitor, for fragmentation and lysis over a duration of 10 min. The overlying liquid was isolated via centrifugation and stored appropriately. Furthermore, human HCC HepG2 and Bel7402 cells were propagated into 6-well plates prior to 48 h of incubation using 0, 10, or 20 μM S-equol. After treatment, cell lysates were prepared and stored. After 10 min of heating at 99°C for determining protein concentration using the BCA method, electrophoresis was performed using 10 μL of each sample under the following conditions: 110 V, 300 mA for 90 min. This was followed by membrane transfer at 200 V, 400 mA for 45 min. The transferred PVDF membranes were then placed in a sealing solution for 2 h of incubation under ambient temperature, followed by overnight primary antibody (1:1000) incubation under 4 °C and an additional 2 h of incubation using a secondary antibody (1:10,000) under ambient temperature. At last, these membranes were developed and the results quantitatively analyzed with Image J software. When the target protein and the internal reference protein have similar molecular weights, the membrane was first incubated with the primary antibody against the target protein, and the signal was detected. Subsequently, antibodies were completely stripped from the membrane using the stripping solution. Finally, the same membrane was re-incubated with the antibody against the internal reference protein, followed by signal detection⁷⁶.

Statistical analysis

All data were expressed as mean ± SD. Statistical analysis was performed using IBM SPSS Statistics 20 software. Normality was confirmed by the Shapiro-Wilk test (P > 0.05), and homogeneity of variance was verified by the Levene test (P > 0.05). Comparisons between two groups were conducted using the unpaired Student t test. For comparisons among three or more groups, one-way analysis of variance (ANOVA) was applied, followed by Tukey’s HSD test for all pairwise comparisons. If the data did not meet the normality assumption, non-parametric tests were used: the Mann-Whitney U test for two-group comparisons and the Kruskal-Wallis H test for three or more groups. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Author contributions

X.W.: Conceptualization, Methodology, Investigation, Data curation, Writing–original draft. H.H.: Visualization, Investigation. F.W.: Visualization, Investigation. P.T.: Methodology, Validation. Z.W.: Methodology, Validation. X.Q.: Validation, Formal Analysis. R.Z.: Validation, Formal Analysis. Y.G.: Validation, Formal Analysis. P.F.T.: Methodology, Formal Analysis. Z.H.: Conceptualization, Supervision, Project administration, Funding acquisition, Writing–review and editing.

Data availability

The raw data generated in this study are stored in the National Center for Biotechnology Information (NCBI), with their project numbers being PRJNA1282939 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1282939), PRJNA1283998 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1283998), and PRJNA1285111 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1285111), corresponding to the raw data of 16S rRNA sequencing, non-targeted metabolomics, and RNA sequencing, respectively. Additionally, all other data is contained within the main manuscript and supplemental files.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41522-026-00919-7.