Epigenetic activation of NK-cell effector programs and caspase-8-dependent apoptosis mediates the antitumor activity of LGP in NSCLC

Abstract

Background

Effective activation of natural killer (NK) cell cytotoxicity and caspase-8-dependent extrinsic apoptosis remains a major challenge in non-small cell lung cancer (NSCLC). Epigenetic mechanisms regulating NK cell function within the tumor microenvironment are poorly understood and rarely targeted therapeutically.

Methods

The antitumor activity of Li-Ginseng Powder (LGP), a specifically processed Panax ginseng formulation enriched in rare ginsenosides (Rh4, Rg3, Rg5, Rk1, and Rk3), was evaluated in human lung cancer A549 cells and A549 xenograft mouse models. NK cell infiltration and activation were assessed by flow cytometry, immunoblotting, and immunohistochemistry. Whole-genome bisulfite sequencing (WGBS) was performed to analyze DNA methylation changes. The effects of LGP ginsenosides (LGG) on tumor cell apoptosis and death receptor signaling were examined in vitro.

Results

LGP significantly suppressed tumor growth and enhanced systemic and intratumoral NK cell activation. Promoter demethylation of NK cell effector genes, including Ncr1, Gzmb, Nktr, and Itgal, was associated with increased NK cell infiltration and activation, elevated granule-mediated cytotoxicity, and enhanced IFN-γ signaling. In parallel, LGP treatment induced caspase-8-dependent apoptosis associated with increased expression of membrane death receptors, their ligands, FADD, and procaspase-8 in tumor tissues. In vitro, LGG upregulated these apoptosis-initiating proteins and triggered caspase-8 activation in A549 cells independent of promoter methylation changes. Collectively, these immune-associated and tumor-intrinsic responses contributed to robust tumor suppression with a favorable systemic safety profile.

Conclusions

LGP exerts dual antitumor effects characterized by enhanced NK-cell activation and increased sensitivity of tumor cells to caspase-8–dependent extrinsic apoptosis. These coordinated immune-associated and apoptosis-sensitizing effects underscore the therapeutic potential of LGP for the treatment of NSCLC.

Graphical abstract

1. Introduction

According to the World Health Organization's 2022 report, lung cancer caused about 2.48 million new cases and 1.82 million deaths globally in 2022, remaining the top cause of cancer-related deaths [1]. Non-small cell lung carcinoma (NSCLC) is the most prevalent form of lung cancer, comprising 85% of all lung cancer cases [2].

Immunotherapy has emerged as a promising anti-cancer strategy and is now an integral component of the standard treatment regimen for NSCLC [3,4]. Natural killer (NK) cells, a subset of cytotoxic innate lymphocytes, play a crucial role in anti-tumor immunity by regulating tumor growth and metastasis [5]. Recognized as a preferred cell type for cellular immunotherapy, NK cells have been clinically utilized to treat various cancers, including NSCLC [6]. However, the efficacy of immunotherapy in solid tumors remains significantly lower than in hematological malignancies, primarily due to the limited infiltration of immune cells into tumor tissues [7]. Therefore, enhancing NK cell infiltration may represent a promising strategy for both the prevention and treatment of lung cancer [8].

Several genes are central to NK-cell effector function. Ncr1 encodes the activating receptor NKp46, whose upregulation is commonly associated with heightened NK cell activation and enhanced cytotoxic and immunoregulatory functions [[9], [10], [11]]. Gzmb encodes granzyme B, a key cytotoxic effector indicative of increased NK cell killing capacity [[12], [13], [14]]. Nktr has been implicated in NK cell activation and signaling in certain cancer types, such as colorectal cancer [15]; however, its role in NK cells more broadly is not well-characterized, whereas Itgal is essential for NK cell adhesion, migration, and tumor infiltration and has been linked to patient prognosis [16]. Collectively, these genes govern NK cell infiltration, activation, and NK cell trigered cytotoxicity.

Rare ginsenosides, secondary metabolites derived from protopanaxadiol and protopanaxatriol, exhibit significantly enhanced biological activity and are more efficiently absorbed and utilized by the human body than their parent compounds [17]. Studies have reported that rare ginsenosides Rg3 and Rh2 enhance NK cell activity in animal models and prolong the postoperative lifespan of patients with NSCLC [[18], [19], [20], [21]]. Furthermore, rare ginsenosides Rk3, Rk1, Rg3, and Rg5 induce apoptosis via both the death receptor and mitochondria-mediated pathways [[22], [23], [24], [25], [26], [27]]. Our previous research has shown that (20S) G-Rh2 upregulates the expression of Fas, FasL, TNF-α, and TNFR1, thereby inducing death receptor-mediated apoptosis in HeLa cells [28].

In this study, we investigated the anti-NSCLC activity of Li-Ginseng Powder (LGP), a specially processed ginseng product enriched with rare ginsenosides (Rh4, Rg3, Rg5, Rk1, and Rk3) (Fig. S1A; Table S1). Acute toxicity studies in animal models indicate that LGP is non-toxic, with the LD50 remaining undetermined even at the maximum tested dose of 22.9 g/kg body weight. We demonstrate that LGP exerts sustained anticancer activity in human NSCLC A549 xenograft mouse models by epigenetically enhancing NK cell effector gene expression, promoting NK cell infiltration and activation within tumor tissues to trigger the perforin/granzyme B cytotoxic pathway, while concurrently sensitizing tumor cells to death receptor–mediated apoptosis through upregulation of key components of the death-inducing signaling complex (DISC).

2. Materials and methods

2.1. Contents of ginsenoside in LGP

LGP was provided by ADKH Technology Development Co., Ltd. HPLC analysis of the Li-ginseng ginsenoside fraction (LGG) revealed that Rg5 accounted for 22.57% of total saponins, followed by 20(S)-Rg3 (7.27%), 20(R)-Rg3 (7.94%), Rh4 (12.03%), Rk1 (16.12%), and Rk3 (5.65%) [29] (Fig. S1A; Table S1).

2.2. Animals and environmental conditions

Specific pathogen-free (SPF) male BALB/c-nu mice (6 weeks old, 22 ± 2 g) were obtained from Vital River Laboratory Animal Technology (SCXK(Jing) 2021-0006, RRID: IMSR_RJ: BALB-C-NUDE, Beijing, China) and quarantined and acclimated for one week prior to the initiation of experiments. All animals were housed under standardized conditions (temperature, 24 ± 2 °C; humidity, 50 ± 10%; 12 h light/dark cycles; 13–18 air changes/h) with ad libitum access to water and a standard SPF diet for rats and mice (Keao xieli feed CO., LTD., Tianjin, China). All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Jilin University and conducted in accordance with the Guidelines for Animal Experiments of Jilin University (Approval Number: YNPZSY2024025).

2.3. Transplantation tumor modeling and experimental grouping

Human non-small cell lung cancer (NSCLC) A549 cells were subcutaneously injected into the right flanks of mice at a concentration of 5 × 10⁶ cells/100 μL in PBS. Tumor growth was monitored, and treatment was initiated when the tumor volume reached approximately 100 mm³. Twenty tumor-bearing mice were randomly assigned to the following groups: (1) Model group (oral vehicle + intraperitoneal injection of vehicle); (2) PTX group (oral vehicle + intraperitoneal injection of paclitaxel [20 mg/kg/week]); (3) LGP-L and LGP-H groups (oral administration of LGP at 100 mg/kg every two days and 200 mg/kg every two days, respectively, + intraperitoneal injection of vehicle). Additionally, five healthy mice were included as a control group. Paclitaxel (PTX; Yangtze River Pharmaceutical) was administered for 4 weeks, while LGP treatment continued for 8 weeks. The study was terminated when the mean tumor volume in the model group reached approximately 1500 mm³.

2.4. Tumor volume and tumor growth inhibition

The length (a) and width (b) of the subcutaneous tumors in Balb/c-nu mice were measured weekly using vernier calipers, and the tumor volume (V) was calculated using the formula: V = $\frac{1}{2}$ ab² (mm³) [30].

Tumor Growth Inhibition (TGI) was calculated using the following formula: TGI (%) = (1 - $\frac{\text{Tt}-\mathrm{T}0}{\text{Ct}-\mathrm{C}0}$) × 100. Tt, Tumor volume in the treatment group at a specific time point after treatment initiation. T0, Tumor volume in the treatment group at baseline (start of the experiment). Ct, Tumor volume in the control group at the same time point. C₀, Tumor volume in the control group at baseline [31].

2.5. Whole-genome bisulfite sequencing (WGBS)

2.5.1. DNA extraction, bisulfite conversion, library preparation, and sequencing

Genomic DNA was extracted from cell pellets using the MagPure Bacterial DNA Kit (Magen, Guangzhou, China). DNA concentration was measured with the Qubit dsDNA HS Assay Kit (Sangon, Shanghai, China), and integrity was confirmed via 1% agarose gel electrophoresis. Bisulfite conversion and library preparation were performed using the DNA bisulfite conversion kit (TIANGEN, Beijing, China) according to the manufacturer's instructions. Briefly, genomic DNA was treated with bisulfite to convert unmethylated cytosines to uracil while leaving methylated cytosines unchanged. The DNA fragments were then amplified and converted into sequencing libraries. Library quality and quantity were evaluated using Qubit 4.0 (Thermo Fisher, USA) and 2% agarose gel electrophoresis. Libraries were pooled and sequenced on an Illumina NovaSeq 6000 or DNBseq-T7 (BGI, Shenzhen, China) platform with 2 × 150 bp paired-end reads.

2.5.2. Data processing, methylation analysis, DMR identification, and functional enrichment

Raw reads were trimmed to remove adapters and low-quality bases using Fastp. Lambda DNA was used to calculate bisulfite conversion efficiency, which exceeded 99%. Clean reads were aligned to the reference genome using Bismark (v0.22.3), and uniquely mapped reads were used to extract cytosine methylation information. Methylation motifs were visualized using the ggseqlogo R package. The methylation level of each region was calculated as the fraction of methylated cytosines over total cytosines. Differentially methylated regions (DMRs) were identified using the DMRcaller R package and annotated with SnpEff (v4.3T). Promoter regions were defined as 2000 bp upstream of the transcription start site (TSS). Significant promoter methylation changes were determined based on the following criteria: (i) an average methylation difference ≥0.1 between groups, (ii) at least five cytosine sites within the region, (iii) region length ≥50 bp, and (iv) adjusted P values < 0.05 in both the Mann–Whitney U test (MWU-test) and the two-dimensional Kolmogorov–Smirnov test (2D KS-test).

2.6. Flow cytometry

Spleen and tumor tissues were collected, and single-cell suspensions were prepared and adjusted to 1 × 10⁷ cells/mL in cold RPMI-1640 containing 10% fetal bovine serum (FBS). Cells were stained with the indicated antibodies according to the manufacturers’ instructions and incubated at 4 °C for 1 h in the dark. The detailed information for all antibodies used in flow cytometry is listed in Supplementary Table S2. Samples were acquired on a BECKMAN COULTER CytoFLEX flow cytometer and analyzed using CytExpert software.

2.7. Histopathological examination

Tumor tissues and major organs, including the heart, liver, spleen, lung, and kidney, were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 4 μm. Paraffin sections were deparaffinized and rehydrated using standard procedures, followed by hematoxylin and eosin staining (H&E; Servicebio, Wuhan, China). Histopathological features were examined under a light microscope (Nikon ECLIPSE Ni-U; RRID: SCR_024835).

2.8. Terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL)

Apoptotic cells in tumor tissues were detected using a fluorescein-based TUNEL assay kit (FITC TUNEL Cell Apoptosis Detection Kit; Servicebio, G1501) according to the manufacturer's instructions. TUNEL-positive cells were visualized by FITC fluorescence, and nuclei were counterstained with DAPI (Beyotime, P0131) before mounting.

2.9. Immunohistochemistry

For the immunohistochemistry (IHC) assay, tissue sections were deparaffinized and rehydrated using standard protocols [32]. Antigen retrieval was performed using Tris-EDTA buffer (pH 9.0) or citrate buffer (pH 6.0), followed by blocking with 3% hydrogen peroxide (H₂O₂) and 10% goat serum. The sections were then incubated with the indicated primary antibodies at 4 °C for at least 16 h. Detailed information on all antibodies used for immunohistochemistry is listed in Supplementary Table S2. After incubation with the corresponding polydimerized secondary antibodies, positive signals were visualized using DAB chromogen (ZSGB-BIO, ZLI-9018). The nuclei were counterstained with hematoxylin prior to coverslip mounting. Images were acquired using a Nikon ECLIPSE Ni-U microscope (RRID: SCR_024835; Tokyo, Japan) at 100× or 200× magnification, and positive staining was quantified using ImageJ software.

2.10. Western blotting

Western blotting assay was carried out according to previous description [28]. Briefly, 30 μg of total protein was separated by 10% or 12% SDS-PAGE and transferred onto PVDF membranes (Millipore, Bedford, MA, USA; IPVH00010). After blocking with 5% skim milk at room temperature for 1 h, the membranes were incubated with the indicated primary antibodies at 4 °C for at least 16 h. Detailed information on the primary antibodies used for western blotting is listed in Supplementary Table S2. The membranes were then incubated with anti-rabbit IgG HRP-linked secondary antibody (1:5000; Cell Signaling Technology, 7074, RRID: AB_2099233) for 1 h at room temperature. Protein bands were detected using Super ECL reagent and quantified with ImageJ software.

2.11. Cell culture and processing

Human non-small cell lung cancer A549 cells (CL-0016, RRID: CVCL_0023, Pricella) were cultured in DMEM/high glucose medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 μg/mL streptomycin in a 37 °C incubator with 5% CO₂. A total of 2 × 10⁶ A549 cells were seeded in 100 mm culture dishes. The cells were treated with LGG at concentrations of 7.5 μg/mL, 15 μg/mL, and 30 μg/mL for 8 h. Cells were subsequently collected for sample preparation.

2.12. Quantitative real-time PCR

Total RNA from cells or tissue samples was isolated using TRIzol reagent (15596018CN, Invitrogen) following the manufacturer's protocol. RNA quality and concentration were determined using a spectrophotometer. Reverse transcription was carried out using 1 μg of RNA to generate cDNA with a commercial reverse transcription kit (A240-10, GenStar).

Real-time PCR was conducted using a SYBR Green–based detection system on a quantitative PCR platform (QuantStudio 5). The thermal cycling program consisted of an initial denaturation step at 95 °C, followed by 40 amplification cycles of denaturation at 95 °C and annealing/extension at 60 °C. Melting curve analysis was performed to verify amplification specificity.

Relative mRNA expression levels were normalized to GAPDH and calculated using the 2^−ΔΔCt method. Each experiment was performed with at least three independent replicates. Primer sequences are provided in Supplementary Table S3.

2.13. Statistical analyses

GraphPad Prism 8.0 (RRID: SCR_002798) software was used for statistical analysis. Data are presented as mean ± standard deviation (SD). Statistical significance was determined using an unpaired, two-tailed Student's t-test. A P value of ≤0.05 was considered statistically significant.

3. Results

3.1. LGP produces long-lasting inhibition of tumor growth even after treatment cessation

We evaluated the inhibitory effects of LGP in human NSCLC A549 xenograft mouse models (Fig. 1A), using Paclitaxel (PTX) as the positive control. The study was terminated when the mean tumor volume in the model group reached approximately 1500 mm³. In accordance with the standard treatment protocols for PTX, administration of PTX was discontinued at the end of the fourth week, while LGP therapy continued until the completion of the eighth week. By the end of the eighth week, both PTX (after a four-week cessation period) and LGP (at doses of 100 mg/kg and 200 mg/kg) demonstrated significant inhibition of tumor growth in comparison to the model group (Fig. 1C). The tumor growth inhibition (TGI) rate in the PTX group was observed to decrease by the end of the eighth week relative to the fourth week, indicating a resurgence in tumor growth (Fig. 1B–C, Fig. S2A–B). By week 12 (after 8 weeks of discontinuation), the TGI in the PTX group further decreased compared to week 8. Surprisingly, at the conclusion of the 12-week period (after 4 weeks of discontinuation) the TGI in the LGP group was significantly enhanced compared to both week 4 and week 8 (Fig. 1B–D, Fig. S2A–C) and the tumor mass continued to exhibit a significant decrease (Fig. 1C). Overall, the TGI in the PTX group declined after treatment cessation, whereas the TGI in the LGP groups was sustained or increased over time. Additionally, hematoxylin and eosin (H&E) staining revealed extensive cellular necrosis and a significantly disorganized and relaxed arrangement of tumor cells in the LGP-treated groups (100 mg/kg and 200 mg/kg). Conversely, the tumor architecture was relatively well-preserved in both the model group and the paclitaxel (PTX) group (Fig. 1E). Notably, unlike conventional chemotherapy, LGP treatment resulted in prolonged and extensive cell death even for a period of four weeks following the cessation of treatment.

Fig. 1.

LGP exhibits sustained antitumor efficacy that persists even after treatment cessation. A Schematic of experimental design. B Tumor growth inhibition (TGI) percentage was measured at Week 4, 8, and 12 post-treatment (n = 5). C Tumor volume progression over 12 weeks (n = 5). D Representative images of tumors collected at Week 12 from each group. E Hematoxylin and eosin (H&E) staining of tumor tissues. The dashed line delineates the boundary between the normal tumor cell region and the necrotic tissue area. Scale bar: 100 μm.

3.2. LGP activates caspase-8–dependent extrinsic apoptosis pathway in tumor cells

To elucidate the sustained shrinkage of tumor mass after LGP treatment, we assessed cell proliferation and apoptosis in the tumor tissues. The results showed that LGP (100 mg/kg and 200 mg/kg) treatment significantly decreased the proportion of Ki67-positive cells (proliferating cells) compared to the model group, while PTX exhibited similar levels of Ki67-positive cells as the model group (Fig. 2A and B). This indicates that cell proliferation rebounded after PTX treatment was discontinued, whereas the cessation of LGP treatment maintained a sustained anti-proliferative effect, consistent with the changes in tumor size (Fig. 1D). Furthermore, a significantly higher number of TUNEL-positive cells (indicative of apoptosis) were observed in the tumors of mice treated with LGP compared to both the model group and the PTX group (Fig. 2C and D). These findings indicate that substantial apoptosis was induced in the LGP-treated tumors even four weeks after the cessation of treatment. The analysis of PARP cleavage demonstrates the intense activation of caspase-3/-7 in the LGP group (Fig. 2E and F). Caspase-8 was prominently activated in LGP-treated tumors, whereas only minimal cleaved caspase-9 was detected in the LGP-H group, suggesting that apoptosis was predominantly associated with activation of the caspase-8-dependent extrinsic pathway (Fig. 2E and F). Importantly, procaspase-8 expression was markedly upregulated in the LGP-treated groups (100 mg/kg and 200 mg/kg) compared with both the model and PTX groups, as demonstrated by immunoblotting and immunohistochemical staining (Fig. 2A–E and G). The increased abundance of procaspase-8 expands the pool of activatable caspase-8, thereby sensitizing tumor cells to caspase-8–dependent apoptosis and promoting activation of the extrinsic apoptosis pathway.

Fig. 2.

LGP reduces cell proliferation and induces apoptosis via a caspase-8–dependent extrinsic pathway. A Ki67 and Caspase-8 immunohistochemical staining of tumor tissues. Scale bar: 100 μm. B Quantification of Ki67-positive cells (n = 3). C TUNEL staining of tumor tissues to detect apoptosis. Scale bar: 250 μm. D Quantification of TUNEL-positive cells across groups (n = 3). E Western blot analysis of apoptosis-related proteins, including PARP, cleaved PARP, Procaspase-8, Cleaved caspase-8, Procaspase-9, and Cleaved caspase-9, in tumor tissues. F Relative protein expression levels normalized to GAPDH (n = 3). G Quantification of caspase-8 immunohistochemical staining (n = 3). Data are presented as mean ± SD. Statistical significance was determined using Student's t-test. ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

3.3. LGP upregulates death receptor pathway proteins without altering promoter methylation

Given that tumor cell apoptosis was predominantly mediated through the extrinsic apoptosis pathway, we examined key regulators of death receptor–mediated apoptosis signaling. Notably, death receptor–mediated apoptosis regulating proteins including Fas, FasL, DR4, DR5, TRAIL, TNF-α, TNFR1, and FADD were significantly upregulated (Fig. 3A and B), paralleling the expression pattern observed for procaspase-8 (Fig. 2E). These expression changes were unlikely to result solely from the pharmacokinetic presence of LGP and may instead reflect stable epigenetic regulation, because these were occurred four weeks after cessation of LGP administration. To explore this possibility, whole-genome bisulfite sequencing (WGBS) was performed on xenograft tumor tissues. Given that the tumor samples comprised both human cancer cells and mouse stromal/immune components, sequencing reads were independently aligned to human and mouse reference genomes for parallel analyses.

Fig. 3.

LGP upregulates key proteins in the death receptor-mediated apoptosis pathway in tumor tissues. A Immunoblot analysis of death receptor-mediated apoptosis regulating proteins, including Fas, FasL, DR4, DR5, TRAIL, TNF-α, TNFR1, and FADD, in tumor tissues from Model, PTX, LGP-L, and LGP-H groups. GAPDH served as a loading control. B Quantification of protein expression levels corresponding to panel A(n = 3). C Quantification of protein expression levels corresponding to panel A (n = 3). Data are presented as mean ± SD, statistical significance was determined using Student's t-test. ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

Methylation profiling aligned to the human genome showed predominant CG methylation [33], with comparatively low mCHH and mCHG levels (Fig. S3A). Manhattan plot analysis showed the chromosomal distribution of differentially methylated regions (DMRs) identified by the Mann–Whitney U test, with significant DMRs distributed across multiple chromosomes and enriched at specific loci (Fig. S3B). Regional profiling of CpG methylation across genomic features (upstream, exon, intron, and downstream regions) revealed distinct distribution patterns, with particular enrichment in upstream regulatory regions (Fig. S3C). In total, 669 DMRs were identified, including 53 hypermethylated (gain) and 616 hypomethylated (loss) regions, while 2287 regions showed no differential methylation. Notably, no significant differences in promoter methylation were observed in genes, including CASP8, FAS, TNFRSF1A, TNFRSF10A, TNFRSF10B, and FADD in the human-aligned dataset (Fig. S3D), nor Fasl, Tnf-α, and Trail in the mouse-aligned dataset (Fig. 4D) in tumor tissues of model and LGP group. These findings suggest that the sustained upregulation of death receptor–associated apoptotic proteins is unlikely to be directly attributable to promoter demethylation of the corresponding genes. Other regulatory mechanisms may contribute to these expression changes and warrant further investigation.

Fig. 4.

LGP induces promoter demethylation of NK-cell effector genes and promotes NK-cell infiltration and activation in tumors. A Distribution of cytosine methylation contexts showing proportions of mCG, mCHH, and mCHG. B Manhattan plot of differentially methylated regions (DMRs) across chromosomes identified by Mann–Whitney U test; the dashed line indicates the significance threshold. C Average CpG methylation levels across gene regions (upstream, exon, intron, downstream) in the Model and LGP groups. D Volcano plot of DMRs showing hypermethylated and hypomethylated regions in LGP versus Model tumors; NK cell–related genes (Ncr1, Nktr, Gzmb, Itgal) and membrane death ligand (Fasl, Tnfsf10, tnf) are highlighted. Promoter regions were defined as 2000 bp upstream of the transcription start site (TSS). DMRs were defined as regions with an average methylation difference ≥0.1, at least five cytosine sites, a region length ≥50 bp, and adjusted P < 0.05 in both the Mann–Whitney U test and the 2D Kolmogorov–Smirnov test. E Immunoblot analysis of NKp46, GZMB, NKTR and ITGAL proteins in tumor tissues from control, model, PTX, and LGP-treated groups, with GAPDH used as a loading control. F Quantification of protein expression levels corresponding to panel E (n = 3). G Flow cytometry analysis used to determine the proportion of CD49b ⁺ CD69⁺ NK cells within tumor-infiltrating lymphocytes across treatment groups. H Quantification of CD49b ⁺ CD69⁺ NK-cell subsets based on flow cytometry measurements (n = 3). I Immunoblot analysis of CD16, IFN-γ, and Perforin in tumor tissues across the indicated groups, with GAPDH as a loading control. J Quantification of CD16, IFN-γ, and Perforin protein expression levels shown in panel I (n = 3). Data are presented as mean ± SD, statistical significance was determined using Student's t-test. ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

3.4. LGP causes demethylation of the promoter regions (2 kilobases upstream of the transcription start site) of the Ncr1, gzmb, nktr and itgal genes, thereby increasing the expression of these genes and enhancing the infiltration, activation and cytotoxicity of NK cells

Using the mouse reference genome for alignment and analysis, we observed a methylation landscape consistent with typical mammalian methylomes. Alignment to the human reference genome likewise revealed a typical mammalian methylation pattern, predominantly in the CG context, with relatively low levels of mCHH and mCHG methylation [33] (Fig. 4A). Manhattan plot analysis illustrated the chromosomal distribution of differentially methylated regions (DMRs) identified by the Mann–Whitney U test (Fig. 4B). Regional profiling of CpG methylation across genomic features (upstream, exon, intron, and downstream regions) revealed distinct distribution patterns (Fig. 4C). In total, 32 DMRs were identified, including 19 hypermethylated (gain) and 13 hypomethylated (loss) regions, while 75 regions showed no differential methylation. Interestingly, LGP treatment induced upstream demethylation of key NK-cell effector genes, including Ncr1, Nktr, Gzmb, and Itgal (Fig. 4D).

The observed promoter demethylation of these loci suggests increased transcriptional potential and supports enhanced NK cell–mediated antitumor immunity. Expectedly, Western blot and immunohistochemistry analysis showed that the protein levels of NCR1 (NKp46), GZMB, NKTR, and ITGAL were markedly increased by LGP treatment at both 100 mg/kg and 200 mg/kg (Fig. 4E and F). These findings suggest that LGP preferentially remodels the epigenetic landscape of tumor-infiltrating NK cells thereby promoting the survival and activation of NK cells in tumor tissues.

Flow cytometry analysis was conducted to quantify NK cell infiltration and activation status in tumor tissues. LGP treatment resulted in a significantly higher proportion of CD69⁺ and CD49b⁺ NK cells compared to model group (Fig. 4G and H). Consistent with these quantitative data, immunohistochemical analysis confirmed increased expression of NK effector molecules (NKp46, NKTR, and ITGAL) within the tumor, further supporting enhanced NK-cell infiltration and activity in the tumor microenvironment (Fig. S4A–D). The increased NK-cell infiltration and activation observed in LGP-treated tumors, suggest the activation of cytotoxic mechanisms through which NK cells eliminate tumor cells, including granule-dependent cytotoxicity (perforin/GZMB) [34,35], death receptor–mediated apoptosis (FasL/TRAIL–Caspase-8 axis) [36,37], and antibody-dependent cell-mediated cytotoxicity (ADCC, CD16) [38].

Indeed, the increased cleavage of caspase-8 (Fig. 2E), the expression of membrane death receptors, their ligands, and FADD was elevated following LGP treatment demonstrate the activation of death receptor–mediated apoptosis (Fig. 3A and B). Granule-mediated cytotoxic effectors and ADCC-related proteins, including granzyme B (GZMB), perforin, and CD16, were also significantly increased in LGP-treated tumors compared with the model group, whereas paclitaxel did not induce appreciable changes (Fig. 5I and J). These alterations were accompanied by increased IFN-γ levels, supporting enhanced NK-cell effector function and recruitment within the tumor microenvironment [39,40].

Fig. 5.

LGG upregulates key death receptor pathway components in A549 lung cancer cells. A Immunoblot analysis of Fas, DR4, DR5, TNFR1, FADD, Procaspase-8 and Cleaved caspase-8 in A549 cells treated with increasing concentrations of LGG (0, 7.5, 15, 30 μg/mL). GAPDH served as a loading control. B-C Densitometric quantification of the protein expression levels shown in panel A, presented as fold change relative to the untreated group (n = 3). Data are presented as mean ± SD, statistical significance was determined using Student's t-test. ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

Collectively, these results demonstrate that LGP promotes robust NK-cell infiltration and activation within tumor tissues and enhances multiple NK cell–mediated cytotoxic mechanisms. This includes activation of death receptor–dependent apoptosis, granule-mediated cytotoxicity, and ADCC-related pathways, accompanied by increased IFN-γ production. These coordinated changes indicate that LGP elicits a strengthened NK cell–driven antitumor immune response in the tumor microenvironment.

3.5. LGG activates death receptor–mediated apoptosis in A549 cells in vitro

Given the observed upregulation of death receptor components and procaspase-8 in tumor tissues, we investigated whether LGG (LGP-derived ginsenosides), a major active constituent of LGP, contributes to this effect. A549 cells were treated with increasing concentrations of LGG (0, 7.5, 15, and 30 μg/mL) for 8 h. LGG treatment significantly elevated the protein levels of Fas, TNFR1, DR4, DR5, FADD, and procaspase-8 in A549 cells (Fig. 5A–C). These findings are consistent with our previous report demonstrating that 20(S)-ginsenoside Rh2 upregulates Fas, FasL, TNF-α, and TNFR1, thereby inducing death receptor–mediated apoptosis in HeLa cells (18).

To determine whether the upregulation of membrane death receptors, FADD, and caspase-8 was associated with promoter demethylation, whole-genome DNA methylation profiling was performed in LGG-treated A549 cells. Alignment to the human reference genome showed predominant CG methylation and lower CHH/CHG levels, consistent with mammalian patterns (Fig. S5A). Manhattan plot analysis demonstrated the genome-wide distribution of differentially methylated regions (DMRs) identified by the Mann–Whitney U test (Fig. S5B). Distinct CpG methylation patterns were observed across genomic regions (upstream, exon, intron, and downstream) (Fig. S3C). A total of 499 DMRs were identified, comprising 94 hypermethylated (gain) and 405 hypomethylated (loss) regions, while 2425 regions exhibited no significant methylation changes. Notably, LGG treatment did not result in detectable alterations in methylation levels within the upstream regulatory regions of apoptosis-related genes, including CASP8, FAS, TNFRSF1A, TNFRSF10A, TNFRSF10B, and FADD, compared with untreated controls (Fig. S5C). These findings are consistent with the protein upregulation observed in tumor tissues, indicating that the increased expression of these death receptor pathway components is not attributable to these gene promoter demethylation but rather to other regulatory mechanisms.

3.6. LGP and LGG selectively activate apoptotic signaling in tumor cells while sparing normal tissues

It is important whether LGP also induces apoptosis in normal organs of experimental mice, as observed in tumor tissues. We then examined the expression of those key components of the membrane death receptor pathway in lung and liver tissues from tumor-bearing mice. In lung tissues, LGP treatment slightly increased the protein levels of several death receptor–associated components, including DR5, TRAIL, TNF-α, FADD (Fig. 6A–C). However, TUNEL staining revealed no detectable apoptosis in lung tissues following LGP treatment (Fig. 6G). Similarly, several proteins involved in the death receptor pathway were modestly upregulated in liver tissues after LGP administration, yet no evident apoptotic signals were observed by TUNEL staining (Fig. 6D–G). Consistently, treatment of normal human cell lines, Human Umbilical Vein Endothelial Cells (HUVEC) and Medical Research Council cell strain-5 (MRC-5) with graded concentrations of LGG under serum-free conditions for 8 h did not induce detectable changes in the expression of these apoptosis-related proteins (Fig. S6).

Fig. 6.

LGP does not activate death receptor–mediated apoptosis in lung and liver tissues. A Immunoblot analysis of death receptor–associated proteins, including Fas, FasL, DR4, DR5, TRAIL, TNF-α, TNFR1, FADD, Procaspase-8 and Cleaved caspase-8 in lung from Model, PTX, LGP-L, and LGP-H groups. GAPDH served as a loading control. B Quantification of protein expression levels corresponding to panel A (n = 3). C Quantification of protein expression levels corresponding to panel A (n = 3). D Immunoblot analysis of death receptor–associated proteins, including Fas, FasL, DR4, DR5, TRAIL, TNF-α, TNFR1, FADD, Procaspase-8 and Cleaved caspase-8 in liver from Model, PTX, LGP-L, and LGP-H groups. GAPDH served as a loading control. E Quantification of protein expression levels corresponding to panel D (n = 3). F Quantification of protein expression levels corresponding to panel D (n = 3). G TUNEL staining of normal lung and liver tissues showing no detectable apoptosis after LGP treatment. TUNEL-positive cells are shown in green, and nuclei are counterstained with DAPI (blue). Scale bar, 100 μm. Data are presented as mean ± SD, statistical significance was determined using Student's t-test. ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

LGP and PTX do not induce overt toxicity in major organs (heart, liver, lungs, spleen, and kidneys) of tumor-bearing mice, as evidenced by unchanged organ indices, H&E staining, and body weight measurements in the heart, liver, lungs, and kidneys following LGP administration (Fig. S7A–F). In contrast, the model group exhibited marked splenomegaly and structural disruption in the spleen, which were significantly alleviated by LGP treatment (Fig. S7H–G). Importantly, the flow cytometry showed that LGP increased both the total NK-cell population and the proportion of activated NK cells in the spleen compared with the model group (Fig. S8), indicating enhanced systemic NK-cell activity without apparent toxicity to normal tissues or cells.

Collectively, these findings demonstrate that LGP and its ginsenoside extract LGG exert tumor-selective proapoptotic activity. They markedly upregulate components of the death receptor signaling pathway and induce apoptosis in tumor tissues and cancer cells, while causing no overt toxicity in major organs and no significant activation of apoptotic signaling in normal tissues or non-malignant cells.

4. Discussion

Despite substantial progress in targeted therapy and immune checkpoint blockade, many patients with NSCLC derive limited benefit due to a persistently immunosuppressive tumor microenvironment that constrains innate immune effector function. Among these, NK-cell dysfunction has emerged as a key barrier to effective antitumor immunity [41,42]. The present study identifies LGP as a potential immunomodulatory agent associated with enhanced NK-cell activity and increased tumor cell sensitivity to apoptosis, thereby establishing a coordinated antitumor effects.

A central finding of this study is that LGP treatment was associated with reduced promoter methylation of key NK effector genes, including Ncr1, Gzmb, Nktr, and Itgal, together with enhanced NK-cell survival, activation, and cytotoxic mediator expression. These results support a model in which LGP may help restore NK-cell transcriptional competence and functional activity within the tumor microenvironment. Given that epigenetic constraints are increasingly recognized as drivers of NK-cell exhaustion [[43], [44], [45]], our findings highlight a previously underappreciated strategy for reinvigorating innate immunity in NSCLC.

Beyond immune modulation, LGG directly increased tumor susceptibility to apoptosis. Upregulation of death receptors, FADD, and caspase-8 in tumor cells indicates activation of the extrinsic apoptotic pathway. Notably, these changes occurred in the absence of promoter demethylation, suggesting that LGG regulates apoptotic signaling through transcriptional or post-transcriptional mechanisms distinct from those associated with NK-cell effector gene activation. Increased mRNA expression following LGG treatment supports transcriptional activation (Fig. S9), although the upstream regulatory pathways remain to be elucidated.

Importantly, these immune and tumor-intrinsic effects appear to converge in vivo. NK cells in LGP-treated tumors exhibited enhanced cytotoxicity-related features, including granzyme/perforin expression, death ligand signaling, and CD16-dependent ADCC, while tumor cells displayed heightened sensitivity to caspase-8–mediated apoptosis. This coordinated interaction suggests that LGP-associated immune activation and LGG-mediated tumor cell apoptosis may jointly contribute to antitumor efficacy.

From a translational perspective, the ginsenoside-based origin of LGP provides additional therapeutic relevance. Natural product–derived immunomodulators have gained increasing attention due to their multi-target activity and potential tolerability [[46], [47], [48]]. In this study, LGP demonstrated a favorable safety profile, with no overt toxicity in major organs and no activation of apoptotic pathways in normal human cells. Restoration of splenic immune architecture and NK-cell activity further supports systemic immune modulation within a manageable therapeutic window.

Several limitations should be acknowledged. The upstream signaling pathways responsible for LGG-mediated activation of the death receptor axis remain undefined. Furthermore, while xenograft models allowed mechanistic dissection, studies in immunocompetent and syngeneic models are required to assess interactions with adaptive immunity and the durability of NK-cell responses. Broader validation across NSCLC subtypes and other malignancies will also be necessary to determine generalizability.

In summary, our study demonstrates that LGP exerts multifaceted antitumor activity associated with reduced promoter methylation of selected NK-cell effector genes, enhanced NK-cell function, immune-associated amplification of tumor cell apoptosis, and direct activation of apoptotic signaling in tumor cells, all while maintaining a favorable safety profile. These findings position LGP as a promising immunomodulatory strategy and provide a strong preclinical rationale for further development of ginsenoside-based therapeutics in NSCLC.

5. Conclusion

In summary, LGP suppresses NSCLC through coordinated immune-associated and tumor-intrinsic mechanisms. LGP treatment is associated with enhanced NK-cell recruitment and cytotoxicity, accompanied by promoter demethylation of Ncr1, Gzmb, Nktr, and Itgal. Concurrently, LGP and its ginsenoside extract increase tumor cell susceptibility to apoptosis, with increased expression of death receptors, FADD, and caspase-8. These effects occur with minimal toxicity, supporting LGP as a promising immunomodulatory strategy for NSCLC treatment.

Institutional review board statement

This study was approved by the Institutional Animal Care and Use Committee (IACUC) of Jilin University (Approval Number: YNPZSY2024025).

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this manuscript, the authors used ChatGPT to improve the clarity, readability, and overall language quality of the text. All AI-assisted suggestions were carefully reviewed and revised by the authors where appropriate. The authors take full responsibility for the final content of the manuscript.

Conflict of interest

All authors declare that they have no conflict of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article.