TMG were obtained through L. plantarum MB11 fermentation of the major ginsenosides in De Man, Rogosa, and Sharpe (MRS) broth at 37°C for 21 days. Based on previous research, the fermentation broth was freeze-dried, subjected to D101 macroporous resin column chromatography, and eluted with water and 30%, 50%, and 70% (v/v) ethanol (Shen et al., 2023). The eluted fraction with 70% ethanol contained the desired ginsenosides. This fraction was then vacuum-concentrated and lyophilized to obtain a powdered product, referred to as TMG, which served as the test sample. TMG were examined using high-performance liquid chromatography (HPLC) and liquid chromatography mass spectrometry (LC-MS) techniques to determine its composition, which is principally composed of the following ginsenosides: Rg6 (3.2 ± 0.4 mg/g), F4 (4.4 ± 0.3 mg/g), Rh4 (43.3 ± 0.8 mg/g), Rk3 (80.2 ± 1.2 mg/g), 20(R)-Rg3 (185.5 ± 3.5 mg/g), 20(S)-Rg3 (55.1 ± 0.6 mg/g), CK (37.6 ± 0.7 mg/g), and Rh2 (165.2 ± 4.5 mg/g).

The CT26.WT cells were procured from Wuhan Servicebio Technology Co., Ltd., Wuhan, China. The cells were maintained in the RPMI-1640 medium supplemented with 10% fetal bovine serum, 100 IU/mL penicillin, and 100 μg/mL streptomycin. They were cultured in a humidified environment at 37°C with 5% CO₂ for 24 h, with passaging performed every 2–3 days.

Cell viability analyses were performed using the CCK-8 assay. (US Everbright Inc.). CT26.WT cells were seeded in 6-well plates at a density of 1.0 × 10⁴ cells/well and allowed to incubate for 24 h before the treatment. Subsequently, the cells were exposed to TMGs at concentrations of 0, 25, 50, 100, and 200 μg/mL, all dissolved in PBS. The cell culture was then incubated in a CO₂ incubator for 48 h at 37°C and 5% CO₂. Subsequently, the cells were incubated in a 10% (v/v) CCK-8 solution for 3 h. The absorbance was measured at 450 nm using a SpectraMax ABS Plus Microplate Reader from Molecular Devices, Shanghai, China (Kong et al., 2021).

The cell viability was calculated using the following formula: Cell viability  % = A test A control × 100 % .

The apoptosis state of CT26.WT cells was evaluated by performing the TUNEL assay. In a 6-well plate, 2 × 10⁵ CT26.WT cells were cultured with sterile coverslips and exposed to control (0 μg/mL), cyclophosphamide (CTX) (100 μg/mL), and TMG (200 μg/mL) for 24 h; in this experiment, CTX served as the positive control. After reaching the desired size, cells were washed thrice with PBS and fixed with 4% paraformaldehyde for 30 min. Following fixation and another round of washing with PBS, the cells were treated with 0.3% Triton X-100 for 20 min at room temperature, followed by another three PBS washes. Next, the cells were covered with TUNEL kit buffer (provided by Wuhan Service Biotechnology Co., Ltd.) and incubated at room temperature for 10 min. A mixture of TDT enzyme: dUTP: buffer (1:5:50) was applied to the cells and incubated at 37°C for 2 h. After three 5-min PBS washes, the PBS was removed, and the cells were stained with the DAPI dye. The samples were incubated in the dark at room temperature for 10 min, and the images were captured under a fluorescence microscope for observation. The number of TUNEL-positive cells was determined using ImageJ software (Wang et al., 2022).

A total of 30 BALB/c mice (female, weighing 20–22 g) were obtained from Liaoning Changsheng Biotechnology Co., Ltd. (Liaoning, China) and acclimated for 1 week prior to the study’s start. All animal procedures followed the criteria established by the Jilin Academy of Agricultural Sciences’ Animal Care and Ethics Committee (Ethical reference number: 57/2021). The mice’s right flanks were subcutaneously implanted with a suspension of 1 × 10⁷ live CT26.WT cells in 0.2 mL PBS for inducing tumors. Once the tumors reached a size of approximately 0.5 cm (Liu et al., 2005), the mice were randomly assigned to three groups: the model group, the CTX group, and the TMG group. In the TMG group, mice received a dose of 100 mg/kg/day of TMG administered via gavage. The CTX group received intraperitoneal injections of CTX at a dose of 25 mg/kg/day via gavage. The model group was administered 0.9% NaCl via gavage. All mice received 21 days of drug administration before being euthanized to collect feces and tumor tissues, which were immediately frozen in liquid nitrogen and stored at −80°C for further analysis.

The transfected cells were collected from each group after 48 h of cultivation. The CT26.WT cells and tumor tissues were lysed with the RIPA lysis buffer, followed by centrifugation to collect the supernatant. The protein concentration was determined using a BCA Kit. Protein samples were resolved on 8%, 10.0%, or 12% SDS-PAGE by boiling with denaturation buffer [1.5 mol Tris-HCl, 10% sodium dodecyl sulfate (SDS), TEMED, and 10% ammonium persulfate]. After that, the samples were incubated for 40 min on a polyvinylidene fluoride (PVDF) membrane obtained from Millipore (Massachusetts, United States). A measure of 3% BSA in Tris-buffered saline with 0.05% Tween-20 (TBST) was used for the membrane blocking. Next, the membrane was treated with the various primary antibodies at a temperature of 4°C for 12 h. The primary antibodies used were anti-β-actin (GeneTex, GTX629630), anti-Raf (CST, #9422), anti-phosphorylated (p)-Raf (CST, #9421), anti-MEK (Bioss, bs-1433R), anti-p-MEK (Bioss, bs-3270R), anti-ERK1/2 (GeneTex, GTX134462), anti-p-ERK1/2 (GeneTex, GTX24819), anti-Bcl-2 (Bioss, bs-4563R), anti-caspase-3 (CST, #9661), anti-cleaved-caspase-3 (CST, #9661), anti-Bax (Proteintech, 50599-2-Ig), and anti-GAPDH (Proteintech, 60004-1-Ig). After this, the membrane was exposed to horseradish peroxidase (HRP)-conjugated secondary antibody at 37°C for 1 h. The quantification of the immunoreactive bands was carried out by utilizing Image Quant LAS 4000 (Fuji Film, Tokyo, Japan) software by using β-actin and GAPDH as the loading controls.

The tumors were fixed in 4% paraformaldehyde before being implanted in the paraffin block. Thereafter, the slides with 5-μm-thick sections were made. These sections were dewaxed with xylene and rehydrated progressively with graded alcohol solutions. They were then incubated with 3% hydrogen peroxide for 25 min at room temperature away from light. Following that, antigen retrieval was carried out in PBS (pH = 7.4). The sections were then exposed to different primary antibodies, including anti-cleaved-caspase-3 (1:200, GB11532, Servicebio), anti-Bcl-2 (1:500, GB113375, Servicebio), and anti-Bax (1:400, GB114122, Servicebio). After blocking with 10% goat serum for an hour, the sections were incubated with the primary antibody at 4°C overnight. The slices were then exposed to an HRP-labeled secondary antibody. DAB staining was employed, followed by counterstaining with hematoxylin. The slides were visualized using a bright-field microscope (Nikon ECLIPSE E100, Nikon, Tokyo, Japan).

The mouse feces were collected aseptically, and the samples were sent to Sangon Biotech Co., Ltd. (Shanghai, China) for paired-end sequencing on the Illumina MiSeq platform, PCR amplification of the V3–V4 hypervariable region, and the total DNA extraction of the fecal microbiota. The sequencing information has been uploaded to the National Center for Biotechnology Information database, and the accession number is Project No. PRJNA1003433. In addition, for alpha (α)-diversity analysis, Chao1, Shannon, Simpson, and Goods indexes were used to evaluate the microbial diversity in each group. Beta (β)-diversity was measured using the principal coordinates analysis. The unweighted UniFrac distance was employed to determine the between-group differences. Relative abundances at the phylum and genus levels were analyzed to understand the microbial community structure. The potential association between the gut microbiota and apoptosis was evaluated by microorganisms, and the correlation between apoptotic variables and microorganisms was examined by Spearman’s analysis.

Each data displays the mean value and standard deviation (SD). We used one-way ANOVA and Tukey’s multiple comparison test to examine whether there were any significant differences. Additionally, the data were collected and made visually appealing by creating bar and line charts using the Origin 8.0 program. The threshold for statistical significance was p < 0.05.

Figure 1 illustrates that TMG inhibited the proliferation of CT26.WT cells in a dose-dependent manner (at concentrations of 0, 25, 50, 100, and 200 μg/mL). Significant inhibition was observed at a concentration of 200 μg/mL after 24 h of TMG treatment, with the inhibition rate reaching 43.97%. A concentration of 25 μg/mL showed no significant inhibitory effect on CT26.WT cells, with a cell viability rate of 94.36% ± 1.78%. In contrast, 50 μg/mL concentration exhibited a modest inhibitory effect, resulting in a slightly reduced cell viability rate of 88.31% ± 1.63% (p < 0.05).

To further characterize the apoptotic cell death by TMG, we performed the TUNEL staining assay (Figure 2A). TMG-treated cells were stained with TUNEL, and fluorescence intensity was found to be increased. The apoptotic cell number of CT26.WT colon cancer cells reached 54.01% ± 5.14% (p < 0.01) after treatment with TMG for 24 h (Figure 2B). Therefore, TMG could exhibit substantial anti-tumor effects on CT26.WT cell by inducing apoptosis.

The protein expression of apoptosis-related markers, including Bax, Bcl-2, caspase-3, and cleaved-caspase-3, was analyzed by Western blotting on CT26.WT cell lines (Figure 3A). When compared to the control group, CT26.WT cells treated with TMG showed a significant reduction in Bcl-2 protein expression by 32.45% (Figure 3B, p < 0.05). In contrast, untreated control cells exhibited lower levels of Bax, caspase-3, and cleaved-caspase-3 proteins. Specifically, the protein levels of Bax, caspase-3, and cleaved-caspase-3 increased by 38.46%, 27.88%, and 67.41%, respectively, in the treated cells, respectively (Figures 3C–E, p < 0.05).

The results indicated that TMG at a dose of 25 mg/kg significantly increased (p < 0.05) the expression of cleaved-caspase-3 (Figure 4B) and Bax (Figure 4C) compared to the model. Moreover, TMG had a significant inhibitory effect (p < 0.05) on the expression of Bcl-2 in the transplanted mouse CT26 colorectal cancer tissues, as shown in Figure 4D. Upon treatment with TMG, the percentage of cells positive for cleaved-caspase-3 increased significantly to 35.02%. Similarly, the percentage of cells positive for Bax also increased significantly, reaching 46.49%. On the contrary, the percentage of positive cells for Bcl-2 decreased to 27.53% following treatment with TMG.

To further characterize the mechanism underlying the inhibition of cell proliferation by TMG, we examined whether it affected Raf, MEK, and ERK expressions in the tumor tissues using the Western blotting analysis (Figure 5A). The study found that compared with the treatment group, the model group had higher levels of p-Raf, p-MEK, and p-ERK1/2, but the expression levels of these proteins were significantly inhibited by TMG by 15.37%, 27.41%, and 33.1%, respectively (p < 0.01, Figures 5B–D).

To further illustrate the holistic mechanisms underlying the effects of TMG on CT26-bearing mice, we explore the influence of TMG on the gut microbiota by high-throughput DNA sequencing analysis of 16S rRNA in fecal bacteria across all groups of mice. First, we performed α-diversity analysis, in which the Chao1 index reflects the community richness, the Shannon and Simpson indexes visualize the species diversity, and the Goods index represents the coverage of the sequencing libraries for each individual sample. Compared with the model group, the treatments in both the CTX and TMG groups effectively improved the Chao1 index by 161.17% and 59.74%. (Figure 6A, p < 0.05), respectively. For the Shannon index, a significant difference was observed between the CTX and model groups (Figure 6B, p < 0.05), but there was no significant difference between the TMG and model groups. Additionally, in the context of the Simpson index, compared with the model group, both the CTX and TMG groups displayed insignificant differences (Figure 6C). Furthermore, compared to the model group, both the CTX and TMG groups showed significant differences in the Goods index (p < 0.05). Notably, the Goods index for each group was greater than 97%, indicating that a substantial majority of sequences within each group were accurately identified. This high percentage emphasizes that the sequencing results can reflect the composition and diversity of the microorganisms in the samples (Figure 6D). To clarify which bacteria were altered by TMG, the top 10 species of the highest abundance at various taxonomic levels were analyzed (Figures 6E,F). At the phylum level, the probiotic phylum Firmicutes composed the majority in all samples, and it was downregulated in CT26-bearing mice, whereas TMG evidently reversed the decline in Firmicutes abundance (Figure 6E). At the genus level, the probiotic phylum Lactobacillus constituted the majority in all samples. In CT26-bearing mice, it was observed that Lactobacillus abundance was diminished; however, TMG were found to effectively reverse this decline (Figure 6F). Meanwhile, TMG decreased the fecal abundances of Proteobacteria in CT26-bearing mice. β-diversity at the operational taxonomic unit (OTU) level was analyzed by principal component analysis (PCA), whose scatter plot showed that the model group mice were far different from those in the other two groups, which were clustered, indicating that the principal component of fecal bacteria in CT26-bearing mice was altered, and TMG decreased the β-diversity of the gut microbiota in CT26-bearing mice (Figure 6G). We compared the groups by plotting a heatmap of species abundance at the genus level (Figure 6H). In the model group, Helicobacter, Lactobacillus, and Alistipes, which are beneficial strains, were less abundant, while Enterococcus, Proteus, and Blautia, which are harmful strains, were more abundant. However, TMG treatment significantly improved the abundance of these floras.

Notably, Proteobacteria emerged as a marker species for discernible distinctions between groups at the phylum level, as revealed by the forest randomization analysis (Figure 6I). Furthermore, the possible connection between gut microbiota and apoptosis factors was further verified through Spearman analysis (Figure 6J). This analysis indicated a potential positive correlation between Proteus and the levels of Bcl-2 (p = 0.002 and R = 0.73), p-Raf/Raf (p = 0.003 and R = 0.72), p-MEK/MEK (p = 0.017 and R = 0.61), and p-ERK/ERK (p = 0.002 and R = 0.72), but it exhibited a negative correlation with the levels of Bax (p = 0.005 and R = −0.69) and caspase-3 (p = 0.004 and R = −0.33) apoptosis factors. These findings collectively suggested that TMG may potentially affect cancer-associated apoptosis by modulating factors such as Lactobacillus and Proteus (p < 0.01).