Ginsenoside Rh4 Triggers Ferroptosis in Lung Cancer: Targeting // and Remodeling Gut Microbiota for Butyrate-Mediated Activation.

1. Introduction
Lung cancer has a high global incidence and malignancy, seriously endangering human health [1]. With approximately 85% of lung cancer cases being classified as such, non-small cell lung cancer (NSCLC) is the predominant histological subtype [2]. Major risk factors for NSCLC include smoking, genetic predisposition, and chronic lung infections [3]. However, traditional chemotherapy for NSCLC faces challenges like high drug toxicity and resistance. As a result, natural products—because of their lower toxicity and ability to target multiple pathways—are increasingly studied for preventing and treating NSCLC.
Ferroptosis involves iron-driven lipid peroxidation and oxygen species (ROS) accumulation, distinguishing it from apoptosis, necrosis, and autophagy, and highlighting its therapeutic potential in cancer [4]. NSCLC cells, characterized by high iron dependence and low ROS tolerance, are particularly vulnerable to ferroptosis [5]. The hallmark features of ferroptosis are the excessive buildup of lipid peroxides (LPO) and Fe2+ overload, regulated by antioxidant pathways and iron homeostasis [4]. Antioxidant defense mainly depends on glutathione peroxidase 4 (GPX4), whose activity requires glutathione (GSH) synthesized from cysteine imported via the cystine/glutamate antiporter (SLC7A11/xCT). Inhibiting the SLC7A11/GSH/GPX4 pathway triggers ferroptosis. Regarding iron balance, transferrin receptor (TFRC) promotes iron uptake, which drives ferroptosis, while solute carrier family 40 member 1 (SLC40A1) counteracts this process by exporting iron [6]. Notably, the Kelch-like ECH-associated protein 1/nuclear factor erythroid 2-related factor 2 (KEAP1/NRF2) antioxidant axis is a key negative regulator of ferroptosis. Frequent loss-of-function mutations in KEAP1 in NSCLC lead to constitutive activation of NRF2, which upregulates a suite of cytoprotective genes (including SLC7A11 and FTH1), thereby conferring resistance to ferroptosis and posing a significant clinical challenge [7]. Therefore, targeting the NRF2/GPX4 pathway to induce ferroptosis represents a new therapeutic approach. However, the efficacy of ferroptosis induction can be profoundly influenced by systemic factors, including the gut microbiota and its metabolites, which have recently emerged as master regulators of this cell death process.
Recent studies indicate that the gut microbiota significantly influences NSCLC development. Gut microbes affect NSCLC progression through mechanisms including host immune response modulation, inflammation suppression, and metabolic regulation [8]. Furthermore, the gut microbiota has been found to inhibit tumor growth by remodeling the microbial community to mediate immune cell infiltration into tumors and to regulate ferroptosis [9]. Conversely, specific pathogenic gut bacteria, such as Fusobacterium nucleatum, have been demonstrated to inhibit ferroptosis and confer chemotherapy resistance in colorectal cancer, highlighting the dual role of microbiota in regulating this cell death process [10]. Notably, during their metabolic processes, gut microorganisms produce various metabolites that interact with host cells. Emerging evidence indicates that gut microbiota-derived metabolites are involved in the occurrence of ferroptosis and participate in the progression of tumor diseases. For instance, microbial metabolites such as short-chain fatty acids (SCFAs) can enhance tumor cell susceptibility to ferroptosis [11]. Butyrate, a major SCFA, possesses anti-inflammatory, immunomodulatory, and anticancer activities. Research has shown that butyrate can inhibit tumor growth and induce cell death in multiple malignancies, such as lung cancer. Importantly, butyrate has been demonstrated to indirectly inhibit the expression of SLC7A11, which promotes ferroptosis in endometrial cancer [12]. These findings collectively establish gut microbiota-derived metabolites as key regulators of disease progression and therapeutic response by mediating ferroptosis. However, their function in lung cancer remains unclear. Currently, it remains poorly understood whether alterations in the gut microbiota and butyrate directly contribute to ferroptosis resistance in NSCLC cells and what their specific roles are in NSCLC progression.
Ginsenoside Rh4, a rare type of protopanaxatriol saponin found in the plant Panax ginseng, has anti-inflammatory, anticancer, and immunomodulatory activities [13]. The anti-tumor activity of ginsenoside Rh4 has been demonstrated in several malignancies, such as colon cancer, myeloma, and breast cancer. Research on lung cancer has demonstrated that ginsenoside Rh4 inhibits the JAK2/STAT3 axis, hence suppressing the spread of lung adenocarcinoma [14]. Notably, ferroptosis has been identified as a key mechanism underlying ginsenoside Rh4’s anti-tumor activity, involving pathways such as SIRT2 inhibition or p53 activation and autophagy [15,16]. Given its multi-target potential, we hypothesized that ginsenoside Rh4 might exert its anti-NSCLC effects not only by directly inducing tumor cell ferroptosis but also by potentially modulating the gut microbiota-butyrate axis to create a tumor microenvironment more permissive to ferroptosis. However, the specific mechanism by which ginsenoside Rh4 inhibits lung cancer through ferroptosis induction and whether it involves gut microbiota remodeling remains unclear. Thus, the current work assessed the effect of ginsenoside Rh4 on ferroptosis-related proteins and genes as well as cell survival in mouse Lewis lung cancer (LLC) and A549 cells in vitro. In parallel, an LLC tumor-bearing mouse model was created to evaluate Ginsenoside Rh4’s in vivo anti-tumor activity and its capacity to trigger ferroptosis via the KEAP1/NRF2/HO-1 axis. Additionally, 16S rRNA analysis assessed alterations in the gut microbiota of tumor-bearing mice, and butyrate concentration among SCFAs was measured. This further explores how butyrate enhances tumor cell sensitivity to ginsenoside Rh4-induced ferroptosis by activating transcription factor 3 (ATF3) and suppressing GPX4 expression.

2. Results

2.1. Ginsenoside Rh4 Promotes Ferroptosis-Associated Changes in LLC and A549 Cells
To evaluate the anti-tumor effects of ginsenosides Rg1 and Rh4, we first assessed their inhibitory activity on the proliferation of LLC and A549 cells. As shown in Figure S1A,B, treatment with ginsenoside Rh4 for 24 h inhibited the proliferation of both LLC and A549 cells in a dose-dependent manner within the tested concentration range. In comparison, ginsenoside Rh4 exhibited significantly stronger inhibitory effects compared to the unfermented ginsenoside Rg1. The IC50 values of Rh4 in LLC and A549 cells were 54.61 μg/mL and 55.75 μg/mL, respectively. To investigate the underlying mechanism, we examined markers of ferroptosis. As demonstrated in Figure 1A,B, compared with the control group, ginsenoside Rh4 increased the Fe2+ concentration in LLC and A549 cells from 2.39 ± 0.13 nmol/107 cells and 3.53 ± 0.37 nmol/107 cells to 7.10 ± 1.07 nmol/107 cells and 5.90 ± 0.31 nmol/107 cells, respectively. Similarly, the LPO level rose from 0.20 ± 0.06 μmol/L and 0.30 ± 0.07 μmol/L to 0.55 ± 0.07 μmol/L and 0.59 ± 0.04 μmol/L in the two cell lines. In contrast, the Rg1 group showed no significant differences compared with the control. The MDA content in the Rh4-treated group was significantly higher than that in the control, with increases of 44.00 ± 14.2% and 49.83 ± 4.10% in LLC and A549 cells, respectively (Figure 1C). Both Rg1 and Rh4 treatment resulted in significantly reduced glutathione levels compared with the control; however, the decrease was markedly greater in the Rh4 group than in the Rg1 group (Figure 1D). Notably, the ferroptosis inhibitor Ferrostatin-1 (Fer-1) largely reversed these Rh4-induced changes, confirming the involvement of ferroptosis in its anti-tumor activity. Furthermore, immunofluorescence analysis revealed that compared with the strong GPX4 fluorescence observed in the control and Rg1 groups, ginsenoside Rh4 treatment significantly reduced GPX4 signal intensity. Conversely, the TFRC signal was markedly enhanced in the Rh4 group (Figure 2A–C). Quantitative analysis showed that after treatment with ginsenoside Rh4, the proportion of GPX4-positive cells decreased to 38.93 ± 5.8%, while the proportion of TFRC-positive cells increased by 40.29 ± 2.0% relative to the control group. Importantly, these Rh4-induced changes were substantially attenuated by co-treatment with the ferroptosis inhibitor Fer-1, confirming the specific role of ferroptosis in this process.

2.2. Ginsenoside Rh4 Affects the Expression of Ferroptosis-Related Proteins and Genes in LLC Cells
To further elucidate the molecular mechanism underlying ginsenoside Rh4-induced ferroptosis, we analyzed the expression of key ferroptosis-regulatory proteins in LLC (Figure 3A–F) and A549 (Figure 3G–M) cells by Western blotting. In LLC cells, ginsenoside Rh4 treatment significantly upregulated TFRC while downregulating FTH1, SLC40A1, SLC7A11, and GPX4. These changes were substantially reversed by co-treatment with the ferroptosis inhibitor Fer-1. A consistent expression profile was observed in A549 cells. Ginsenoside Rh4 similarly increased TFRC and decreased FTH1, SLC40A1, SLC7A11, and GPX4, effects that were again attenuated by Fer-1. We next investigated whether Rh4 regulates ferroptosis at the transcriptional level. qRT-PCR analysis of key ferroptosis-related genes in LLC cells revealed that Rh4 significantly downregulated mRNA expression of FTH1 (p < 0.05), SLC40A1 (p < 0.01), SLC7A11 (p < 0.01), and GPX4 (p < 0.05), while upregulating TFRC (p < 0.01) compared with the control (Figure 4A–E). These transcriptional changes were largely reversed upon co-treatment with Fer-1.

2.3. Ginsenoside Rh4 Inhibits Tumor Growth in LLC Tumor-Bearing Mice
As shown in Figure 5A, the body weight of mice in the Rh4 group was significantly lower than that in the model group (p < 0.05). Concurrently, Figure 5B demonstrates that treatment with ginsenoside Rg1 and ginsenoside Rh4 reduced tumor volume compared to the model group, with the Rh4 group exhibiting the greatest reduction of 30.65% (p < 0.01). Consistent with the tumor volume changes, the Rh4 group also showed the lowest tumor weight (Figure 5C) and achieved a tumor inhibition rate of 34.32% (p < 0.01) relative to the model group. These results indicate that ginsenoside Rh4 demonstrates superior efficacy compared to ginsenoside Rg1. Furthermore, compared to the model group, H&E staining results (Figure 5D) demonstrated disrupted tumor architecture, visible necrotic areas, loosely arranged cells, and evident pyknosis in the Rh4 group. Therefore, ginsenoside Rh4 effectively inhibited the growth of LLC cells in vivo.

2.4. Ginsenoside Rh4 Modulates Changes in Iron Homeostasis and Ferroptosis in LLC Tumor-Bearing Mice
Subsequently, levels of LPO, iron concentration, and ferroptosis-related proteins were measured in mouse tumor tissues. As shown in Figure 6A,B, compared to the model group, the Rh4 group exhibited increased LPO levels (1.63 ± 0.08 vs. 0.80 ± 0.15 μmol/gprot) and elevated iron concentration (13.73 ± 1.45 vs. 8.10 ± 0.64 μmol/gprot) in tumor tissues. Furthermore, Western blotting (Figure 6C–H) revealed that compared to the model group, ginsenoside Rh4 treatment significantly reduced the levels of FTH1, SLC40A1, SLC7A11, and GPX4 by 40.96%, 52.71%, 17.55%, and 21.53%, respectively, and increased TFRC levels by 58.27%. Compared with ginsenoside Rg1, the effects of ginsenoside Rh4 were more pronounced. These findings are consistent with the in vitro results.

2.5. Ginsenoside Rh4 Downregulates KEAP1/NRF2/HO-1 Axis in LLC Tumor-Bearing Mice
The NRF2 antioxidant pathway is aberrantly activated in LLC cells and promotes the malignant progression of lung cancer through multiple mechanisms. Since ferroptosis can be suppressed by the NRF2/HO-1 pathway, we investigated whether ginsenoside Rh4 could inhibit the NRF2 pathway to induce ferroptosis. As shown in Figure 7A–D, compared to the model group, the Rh4 group exhibited significantly increased MDA content and decreased considerably activities of GSH, CAT, and SOD in tumor tissues. Concurrently, Western blot analysis (Figure 7E–H) revealed that Rh4 intervention significantly increased KEAP1 protein levels and significantly decreased NRF2 and heme oxygenase-1 (HO-1) protein levels compared to the model group. In tumor cells, NRF2 is typically localized and highly expressed in the nucleus. Immunofluorescence analysis of LLC cells demonstrated that ginsenoside Rh4 significantly reduced overall NRF2 expression and inhibited its nuclear translocation (Figure 7I). Molecular docking studies further indicated that ginsenoside Rh4 binds directly to KEAP1 with a docking score of −9.755 kcal/mol. PyMOL (version 3.1) visualization revealed that Rh4 forms hydrogen bonds with key residues of KEAP1, including ARG-326, VAL-561, ILE-559, and VAL-420 (Figure 7J).

2.6. Ginsenoside Rh4 Reverses Gut Microbiota Dysbiosis in LLC Tumor-Bearing Mice
The effect of ginsenoside Rh4 on gut microbiota was evaluated through 16S rRNA sequencing, with sample abundance rank curves presented in Figure 8A. The gentle slope and broad horizontal range of the curves indicate high evenness and low variation in abundance among species within the samples. Alpha diversity analysis (Figure 8B) revealed that, compared to the model group, the Rh4 group exhibited significantly increased Chao 1 (p < 0.05), Simpson (p < 0.05), and Shannon (p < 0.05) indices. Venn analysis identified a total of 7744 OTUs, with 548 OTUs shared between the Rh4 and control groups, indicating high similarity between these groups (Figure 8C). Beta diversity analysis, represented by PCA and PCoA, demonstrated that the microbial community structure in the Rh4 group shifted closer to that of the control group after ginsenoside Rh4 treatment (Figure 8D,E). Following treatment with ginsenoside Rh4, there was an increase in Bacteroidota and a decrease in Firmicutes and Proteobacteria at the phylum level (Figure 8F). At the genus level, intervention with ginsenoside Rh4 increased the relative abundances of beneficial bacteria, such as Muribaculum, Duncaniella, CAG-485, Dubosiella, Paramuribaculum, and UBA3282, while decreasing the relative abundances of deleterious bacteria, including Lactobacillus, Ligilactobacillus, and Limosilactobacillus, bringing them closer to the levels observed in the control group (Figure 8G).

2.7. Ginsenoside Rh4 Reshapes Gut Microbiota Functionality and Intestinal SCFAs
To identify key bacterial species involved in ginsenoside Rh4-mediated lung cancer suppression, we investigated the effect of ginsenoside Rh4 administration on the gut microbiota in mice. As shown in Figure 9A, compared to the model group, ginsenoside Rh4 treatment significantly increased the relative abundances of Muribaculaceae, Dubosiella, and Lachnospiraceae (p < 0.05). This finding was corroborated by Linear Discriminant Analysis Effect Size (LEfSe) analysis (Figure 9B). Furthermore, functional profiling based on KEGG pathway abundance indicated that the primary pathways modulated by ginsenoside Rh4 were associated with fatty acid and lipid biosynthesis, fermentation, and glycolysis (Figure 9C). Subsequently, the levels of SCFAs in mouse colonic contents were quantified. As shown in Table 1, compared with the model group, the Rh4 group exhibited significantly elevated levels of SCFAs, with butyric acid levels increasing dramatically by 70.50% following ginsenoside Rh4 administration (p < 0.05).

2.8. Gut Microbiota, SCFAs, and Cell Death Factors from Multifactorial Correlation Analysis
Correlation analyses were performed between gut microbiota and SCFAs. As shown in Figure 10A, at the phylum level, Proteobacteria and Firmicutes were negatively correlated with SCFAs, while Actinobacteriota was positively correlated. Specifically, Proteobacteria was negatively correlated with butyric acid, and Bacteroidota was positively correlated. At the genus level (Figure 10B), Alistipes_A was negatively correlated with butyric acid, whereas UBA7173, Dubosiella, and Muribaculum showed significant positive correlations. Subsequently, correlations between oxidative factors, ferroptosis-related indicators, and SCFAs were examined. Heatmap analysis (Figure 10C) revealed that the oxidative damage markers MDA and Keap1 were negatively correlated with the anti-ferroptosis factors FTH1, SLC7A11, SLC40A1, and GPX4. Moreover, butyric acid among the SCFAs exhibited significant correlations with ferroptosis-related factors, including iron, LPO, SLC7A11, and GPX4 in both LLC cells and mice.

2.9. Butyrate and Ginsenoside Rh4 Activate Ferroptosis in LLC Cells in Vitro via the ATF3/SLC7A11/GPX4 Pathway
The effect of butyrate on ferroptosis in LLC cells was subsequently investigated in vitro. CCK-8 assays revealed that butyrate exerted cytotoxic effects on LLC cells at approximately 1.5 μm (Figure S2A). Furthermore, co-treatment with butyrate and ginsenoside Rh4 significantly enhanced the inhibition rate of cell viability (Figure S2B). Furthermore, as shown in Figure 11A–E, Western blot and RT-qPCR analyses revealed that butyrate treatment alone increased both mRNA and protein levels of ATF3, while it decreased protein expression of SLC7A11 and GPX4 compared with the control group. In comparison to the control, combined treatment with butyrate and ginsenoside Rh4 resulted in a 6.61% increase in ATF3 protein expression, a 27.10% decrease in SLC7A11 protein expression, and a 20.03% reduction in GPX4 protein expression.

3. Discussion
Ferroptosis, a newly recognized form of programmed cell death, has attracted considerable attention in cancer research. Ginsenosides can influence tumor cells by modulating ferroptosis. The SLC7A11/GSH/GPX4 signaling pathway is a key regulator of ferroptosis. SLC7A11 imports extracellular cystine into cells for GSH synthesis, and GPX4 utilizes GSH to reduce lipid peroxides, thereby inhibiting ferroptosis [17]. In this study, ginsenoside Rh4 was found to downregulate SLC7A11 and GPX4 protein expression, reduce cellular GSH levels, promote the accumulation of LPO and MDA, and consequently induce ferroptosis in lung cancer cells. Critically, the ferroptosis inhibitor Fer-1 significantly rescued Rh4-induced cell death, confirming the specific role of ferroptosis in this process. Previous studies indicate that ginsenoside CK also induces ferroptosis and inhibits tumor growth in liver cancer cells by suppressing the SLC7A11/GPX4 pathway, which aligns with our findings [18]. However, not all ginsenosides act via identical mechanisms. For instance, ginsenoside Rh2 was reported to indirectly suppress SLC7A11 by upregulating IRF1 expression to induce ferroptosis, while ginsenoside F2 directly reduces cellular GSH levels and impairs GPX4 function without significantly affecting SLC7A11 [19,20]. Notably, excessive intracellular Fe2+ accumulation is a key trigger for ferroptosis, making the regulation of iron metabolism-related proteins crucial in ferroptosis-related diseases [21]. Furthermore, this study revealed that ginsenoside Rh4 upregulates the iron uptake receptor TFRC, while downregulating the iron storage protein FTH1 and the iron export transporter SLC40A1, thereby promoting intracellular free iron accumulation and further activating ferroptosis in lung cancer cells. Additionally, ginsenoside Rg5 can induce ferroptosis by inhibiting HSPB1 and upregulating NCOA4 expression, which promotes FTH1 degradation via the ferritinophagy pathway, increasing intracellular free iron levels and oxidative stress [22].
This study demonstrated that ginsenoside Rh4 treatment significantly reduced the activity of antioxidant enzymes, including SOD and CAT, and effectively suppressed lung cancer cell growth by targeting the KEAP1/NRF2 signaling pathway. Specifically, ginsenoside Rh4 stabilized KEAP1 protein levels by inhibiting its ubiquitin–proteasome pathway degradation. The enhanced KEAP1/NRF2 interaction effectively prevented NRF2 from escaping degradation and subsequent nuclear translocation. This mechanism aligns with reported findings where ginsenoside Rd inhibited both protein and mRNA expression of NRF2 and its target genes (NQO1, HO-1, GCLC), exerting anticancer effects in lung cancer cells [23]. Notably, the effect of ginsenoside Rh4 on the NRF2 pathway may differ across cell types and involve multiple mechanisms. In tumor cells, ginsenoside Rh4 stabilizes KEAP1, thereby downregulating NRF2 activity to inhibit growth [16]. Conversely, in normal cells, it promotes NRF2 nuclear translocation via Akt signaling activation, exerting cytoprotective effects [24]. Furthermore, the NRF2 signaling pathway is a key negative regulator of ferroptosis. Its activation transcriptionally upregulates multiple critical anti-ferroptosis proteins, including HO-1, SLC7A11, FTH1, and SLC40A1 [25]. This study further revealed that ginsenoside Rh4 treatment significantly decreased HO-1 protein levels and increased intracellular iron content in lung cancer cells, thereby promoting ferroptosis. HO-1 downregulation promotes ferroptosis through a dual mechanism: it weakens cellular antioxidant capacity, reducing ROS scavenging efficiency, and decreases the chelation of redox-active iron, leading to free iron accumulation and ferroptosis activation [26]. Studies have shown that ginsenoside Rh3 can mediate ferroptosis in colorectal cancer cells by activating the Stat3/p53/NRF2/HO-1 signaling pathway [27].
Recent studies suggest that ginsenosides can modulate gut microbiota composition and abundance, and gut microbiota dysbiosis is associated with an increased risk of lung cancer and other diseases [28,29]. At the phylum level, the Firmicutes/Bacteroidota (F/B) ratio is closely associated with lung cancer development [30]. Our results show that ginsenoside Rh4 significantly reduced this ratio and decreased the abundance of the potentially harmful phylum Proteobacteria. Consistent with these findings, Bai et al. reported that ginsenoside Rk3 inhibited tumor growth by restoring microbiota homeostasis through reducing the F/B ratio [31]. At the genus level, ginsenoside Rh4 significantly increased the abundance of beneficial bacteria Duncaniella and Paramuribaculum, which are known for their anti-tumor activities [32]. In addition, this study observed that ginsenoside Rh4 treatment significantly reduced Lactobacillus abundance, weakened the AhR/Nrf2/GPX4 signaling pathway, and promoted the activation of ferroptosis to inhibit tumor growth [33]. In contrast, ginsenoside CK treatment led to increased Lactobacillus abundance, regulated the intestinal barrier, and enhanced immunity [34]. This study further revealed that ginsenoside Rh4 significantly increased the abundances of Muribaculum, Dubosiella, and Lachnospiraceae, which are known to ferment complex carbohydrates into SCFAs, resulting in elevated overall SCFA levels. Notably, fecal butyrate levels were significantly elevated in ginsenoside Rh4-treated mice. Butyrate, a key SCFA, exhibits anti-inflammatory, immunomodulatory, and anti-tumor effects. Correlation analysis further revealed that increased butyrate levels were positively correlated with higher abundances of Muribaculum and Lachnospiraceae, consistent with previous reports identifying these microbes as major butyrate producers in the gut [35,36]. Additionally, earlier reports have shown that ginsenoside Rk3 can promote butyrate synthesis by enriching specific bacterial genera such as Bacteroides and Alloprevotella, thereby contributing to its anti-tumor effects [37]. These findings reveal gut microbiota alterations induced by ginsenoside Rh4, with potential mechanisms involving microbial metabolites or host immune responses warranting further investigation. Our in vitro experiments demonstrated that exogenous butyrate significantly enhanced ginsenoside Rh4-induced ferroptosis in LLC cells. Butyrate promoted this process by activating the ATF3/SLC7A11/GPX4 signaling pathway (Figure 12). Additionally, other studies report that butyrate can inhibit the SLC7A11/GPX4 pathway via upstream proteins (e.g., c-Fos) or directly promote GPX4 ubiquitination and degradation to induce cancer cell ferroptosis [38,39].

4. Materials and Methods

4.1. Ginsenoside Rh4 Preparation
Ginsenoside Rh4 was prepared by biotransformation of ginsenoside Rg1 using Lactiplantibacillus plantarum TRG22, following the method reported by Shen et al. [40]. A total of 1 g of ginsenoside Rg1 (purchased from Shanghai Yuanye Biotechnology Co., Ltd., Shanghai, China) was dissolved in 1 L of sterile fermentation medium containing 30 g of glucose and 10 g of soybean oligopeptide. After filtering the solution using a 0.22 μm microporous filter, L. plantarum TRG22 was added to the sterilized medium and allowed to ferment for 21 days at 37 °C. Centrifugation was used to gather the precipitate following fermentation, and it was then loaded onto a D101 macroporous resin column for purification. The column was eluted with gradient ethanol (0%, 30%, 55%, and 70% v/v), and the fractions eluted with 70% ethanol were concentrated under vacuum and subsequently freeze-dried. High-performance liquid chromatography (HPLC) revealed a single peak at 73.54 min, matching the retention time of reference standards for ginsenoside Rh4 (Shanghai Yuanye Biotechnology Co., Ltd., Shanghai, China) (Figure S3). The purity of the isolated product was determined to be 95.07% by an external standard method using a ginsenoside Rh4 reference standard calibration curve (y = 4948.8x − 6, R2 = 0.999). The isolated product was further confirmed to be ginsenoside Rh4 through liquid chromatography–mass spectrometry (LC-MS) as well as 1H and 13C nuclear magnetic resonance (NMR) spectroscopy (Figure S4). All spectroscopic data were in full agreement with published values for authentic ginsenoside Rh4 [41].

4.2. Cell Culture and Grouping
The LLC cells (Jiangsu Keygen BioTECH, Nanjing, China) and A549 cells (Jiangsu Keygen BioTECH, Nanjing, China) were kept in DMEM or RPMI 1640, respectively, supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. Cells were maintained in a humidified incubator at 37 °C with 5% CO2. Tour groups of LLC and A549 cells were created by seeding them in 6-well plates at a density of 1 × 105 cells each: the Control group, the Rg1 group, the Rh4 group, and the Rh4 + Fer-1 group. A total of 100 μg/mL of ginsenoside Rg1 and ginsenoside Rh4 were incubated with the Rg1 and Rh4 groups for 24 h, while the control group was cultured with complete media. For the Rh4 + Fer-1 group, cells were pretreated with the ferroptosis inhibitor Ferrostatin-1 (Fer-1, 3 μM) for 6 h prior to the addition of 100 μg/mL ginsenoside Rh4, followed by co-incubation for 24 h as a rescue experiment.

4.3. Immunofluorescence Staining
LLC cells were harvested from the culture medium, fixed for 20 min in 4% paraformaldehyde, permeabilized for 15 min in 0.5% Triton X-100, and blocked for one hour in 3% bovine serum albumin. The cells were then incubated overnight with primary antibodies GPX4 (Servicebio, Wuhan, China), TFRC (Proteintech, Wuhan, China) and NRF2 (Proteintech, Wuhan, China) at 4 °C. The following day, after three rinses with phosphate-buffered saline (PBS), the cells were incubated with secondary antibodies Cy3-conjugated goat anti-rabbit IgG (Servicebio, Wuhan, China) and FITC-conjugated goat anti-rabbit IgG (Servicebio, Wuhan, China) for 1 h. The cells then underwent DAPI nuclear staining. Three repetitions of each experiment were made.

4.4. Quantitative Real-Time PCR (RT-qPCR)
Total RNA was extracted from LLC cells with Trizol reagent (Invitrogen, Carlsbad, CA, USA), followed by cDNA synthesis with a reverse transcription kit (Servicebio, Wuhan, China). The cDNA was then subjected to PCR amplification using specific primers, and the PCR products were detected with SYBR Green fluorescent dye in a CFX96 real-time quantitative PCR system (Servicebio, Wuhan, China). GAPDH was used as the reference gene, and analyzed by the 2−ΔΔCT method. The primers used for the analysis were as follows: FTH1 (forward: 5′-GATGTGGCTCTGAAGAACTTTGC-3′, reverse: 5′-CAGTCATCACGGTCTGGTTTCT-3′), SLC40A1(forward: 5′-GAGACAAGTCCTGAATCTGTGCC-3′, reverse: 5′-TTCTTGCAGCAACTGTGTCACAG-3′), TFRC(forward: 5′-TTAGTGATTGTTAGAGCAGGGGA-3′, reverse: 5′-GGCGGAAACTGAGTATGATTGA-3′), SLC7A11(forward: 5′-GCTATCATCACAGTGGGCTACG-3′, reverse: 5′-TAGAATAACCTGGAGACAGCGAAC-3′), GPX4(forward: 5′-GAGGCAGGAGCCAGGAAGTAA-3′, reverse: 5′-CACCACGCAGCCGTTCTTAT-3′).

4.5. Animals and Experimental Design
A total of 48 male C57BL/6 mice (20–22 g) were acclimated for 7 d in SPF-grade facilities. Experimental protocols were approved by the Institutional Animal Care Committee of Changchun University of Chinese Medicine (Ethics reference number: 2024881). The LLC tumor-bearing mouse model was established using the concentration reported by Zheng et al. through a subcutaneous injection of 1 × 106 viable LLC cells suspended in 0.2 mL sterile PBS into the right axillary region [42]. When tumors reached approximately 500 mm3 (calculated as V = 0.5 × length × width2), mice were randomly allocated into four groups (n = 12/group): Control, Model, Rg1, and Rh4. Randomization was performed using a computer-generated random number sequence to ensure comparable baseline tumor volumes across groups. The Rg1 and Rh4 groups were gavaged with 100 mg/kg ginsenoside Rg1 and ginsenoside Rh4, while the control group and model group were gavaged with saline, respectively. Mice were treated for 21 d with daily monitoring of health and food intake. Body weight and tumor size were recorded every 2 d by investigators blinded to group allocation to minimize measurement bias. After the last administration, the mice were executed to collect the tumors for weighing, which was used to calculate the tumor inhibition rate (TIR), and spleen, lung, thymus, and fecal samples were harvested and stored at −80 °C. Sample processing and subsequent biochemical and molecular analyses were also performed in a blinded manner where feasible.

4.6. Biochemical Index Assay
LLC and A549 cells, as well as mouse tumor tissues, were lysed using RIPA lysis solution that contained protease inhibitors. A bicinchoninic acid (BCA) assay kit (Beyotime, Shanghai, China) was used to measure the protein content in the resultant supernatants. Using spectrophotometry, the amounts of Fe2+ and LPO in cells and tumor tissues were identified (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Using ELISA (Sangon Biotech, Shanghai, China), the amounts of malondialdehyde (MDA), GSH, catalase (CAT), and superoxide dismutase (SOD) in cells and tumor tissues were measured.

4.7. Western Blotting
LLC and A549 cells, as well as mouse tumor tissues, were lysed with RIPA lysis solution containing protease inhibitors (Solarbio, Beijing, China) and the supernatant was collected. The protein content of the supernatant was measured and modified using the BCA kit (Beyotime, Shanghai, China). Protein samples were resolved on 10% SDS–polyacrylamide gels (UEIandy, Suzhou, China), transferred to PVDF membrane (Keygen BioTECH, Nanjing, China), and blocked with 5% bovine serum albumin. Primary antibodies against FTH1 (Sangon Biotech, Shanghai, China), GPX4 (Sangon Biotech, Shanghai, China), HO-1 (Proteintech, Wuhan, China), KEAP1 (Proteintech, Wuhan, China), NRF2 (Proteintech, Wuhan, China), SLC40A1 (Proteintech, Wuhan, China), SLC7A11 (Sangon Biotech, Shanghai, China), TFRC (Proteintech, Wuhan, China), and β-actin (Proteintech, Wuhan, China) were incubated with the membrane at 4 °C overnight. Following an hour of room temperature incubation with an HRP-conjugated goat anti-rabbit secondary antibody (Proteintech, Wuhan, China), protein bands were identified using a ChemiScope 5300 (Clinx Science Instruments Co., Ltd, Shanghai, China). Signal intensity was quantified using ImageJ software (version 1.53).

4.8. Hematoxylin and Eosin (H&E) Staining
Tumor tissues underwent 24 h fixation in 4% paraformaldehyde at ambient temperature, followed by paraffin embedding and microtome sectioning. Sections were deparaffinized with xylene, soaked in an ethanol gradient, and washed with ultrapure water. They were stained with H&E, sealed with neutral adhesive, and visualized with a light microscope.

4.9. Gut Microbiota Analysis
To extract total DNA from the contents of the mouse colon, the QIAamp Fast Stool DNA Mini Kit (QIAGEN, Hilden, Germany) was utilized. Using primers 8F: 5′-ACTCCTACGGGAGGCAGCA-3′ and 149R: 5′-GGACTACHVGGGTWTCTAAT-3′, the contents of the colon was denatured at 98 °C for 30 s, annealed at 54 °C for 5 s, and extended at 54 °C for 15 s; this was then repeated 25–30 times, followed by extension at 72 °C for 2 min and the amplification of the highly variable regions V3-V4 of bacterial 16S rRNA. PCR products were sequenced on the Illumina MiSeq platform, and data were processed using the QIIME2 pipeline for sequence denoising or operational taxonomic unit (OTU) clustering. The organization and composition of the microbial community were then evaluated using α and β diversity analyses.

4.10. SCFAs Analysis
The quantities of SCFAs in the contents of the mouse colon were ascertained using gas chromatography–mass spectrometry (GC-MS). An Agilent HP-INNOWAX capillary column (30 m × 0.25 mm ID × 0.25 μm film thickness) combined with a 1 μL injection at a 10:1 split ratio was used for the analyses. The GC program set the inlet at 250 °C, the ion source at 300 °C, and the transfer line at 250 °C. A 90 °C temperature ramp was used to begin the temperature program. The heating protocol began at 90 °C and increased by 10 °C each minute to 120 °C, then gradually increased to 150 °C and then quickly to 250 °C. The helium carrier gas flow rate was kept constant throughout the chromatographic separation process at 1.0 mL/min. Mass spectrometric detection was conducted using a Thermo ISQ LT system (Thermo Fisher Scientific, Waltham, MA, USA) operated in electron impact ionization mode under specific conditions: 70 eV ionization energy with selective ion monitoring acquisition. Pure standards of acetic, butyric, propionic, isobutyric, isovaleric, and valeric acids (Sigma-Aldrich, St. Louis, MO, USA) were used for the identification and quantification of SCFAs.

4.11. Evaluating the Potentiating Effects of Butyrate on Ferroptosis in LLC Cells

4.11.1. Cell Grouping and Treatment
LLC cells were plated in 6-well plates at 1 × 105 cells per well and divided into six groups: Control, Rg1, Rh4, butyrate, butyrate + Rg1, and butyrate + Rh4. The control group was maintained in complete culture medium. The Rg1 and Rh4 groups were incubated with 100 μg/mL ginsenoside Rg1 and ginsenoside Rh4 for 24 h. The butyrate group was treated with butyrate (1.0 μM) for 24 h. The butyrate + Rg1 and butyrate + Rh4 groups were pretreated with butyrate (1.0 μM) for 24 h, followed by the addition of 100 μg/mL of ginsenoside Rg1 or ginsenoside Rh4, respectively, for an additional 24 h.

4.11.2. RT-qPCR
The expression levels of selected genes were analyzed using RT-qPCR, following the method described in Section 4.4. The following primers were used for the amplification of target genes: ATF3 (forward: 5′-CCTCGTCCCGTAGACAAAATG-3′, reverse: 5′-TTCTTGTTTCGACACTTGGCA-3′). The relative expression levels of these genes were quantified using the 2−ΔΔCt method, with GAPDH serving as the internal control.

4.11.3. Western Blotting
According to the Western blotting method in 2.7, the primary antibodies were ATF3 (ImmunoWay Biotech, Plano, TX, USA), SLC7A11 (Sangon Biotech, Shanghai, China), GPX4 (Sangon Biotech, Shanghai, China), and β-actin (Proteintech, Wuhan, China). Protein expression levels were normalized to β-actin as a loading control.

4.12. Molecular Docking
The comprehensive chemical information of ginsenoside Rh4 was provided by the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). The target protein KEAP1 (PDB ID: 6TYM) was downloaded from the PDB database (http://www.wwpdb.org/). Prior to molecular docking, the protein was preprocessed using Schrodinger’s “Protein Preprocessing Wizard” tool, which added hydrogen atoms and missing side chains and ring structures while removing water molecules and ligands bound via hydrogen bonds. The grid was set to a 62 × 62 × 62 square grid. Finally, the treated proteins were fed into the optimized ligand ginsenoside Rh4 and XP was selected for docking to predict the docking binding results.

4.13. Statistical Analysis
GraphPad Prism 8.0.1, IBM SPSS Statistics 20.0, and OriginPro 2022 (OriginLab Corporation, Northampton, MA, USA) were used for all statistical analyses. Experimental results were presented as means ± standard deviation. Fisher’s LSD post hoc comparisons were used in one-way ANOVA to assess parametric datasets that confirmed normality. The thresholds for statistical significance were set at p-values < 0.05.

5. Conclusions
This study demonstrates that ginsenoside Rh4 exerts anti-tumor effects by inducing ferroptosis through modulating the KEAP1/NRF2/HO-1 pathway, downregulating the SLC7A11/GPX4 axis, and disrupting iron homeostasis. Simultaneously, ginsenoside Rh4 remodels the gut microbiota, increasing the abundance of butyrate-producing bacteria and elevating butyrate levels, which in turn enhances Rh4-induced ferroptosis through ATF3-mediated inhibition of GPX4. Collectively, these findings reveal a novel anti-tumor mechanism of ginsenoside Rh4 involving gut microbiota remodeling and the butyrate-mediated augmentation of ferroptosis.