Background

Colon adenocarcinoma (COAD) represents a major global health challenge, accounting for approximately 880,000 deaths and over 1.85 million new cases annually [1]. Standard treatments for COAD include surgery, radiation, chemotherapy and immunotherapy [2]. However, despite ongoing advancements in treatment strategies, metastasis remains a critical issue for some patients, significantly impacting prognosis [2,3,4]. Furthermore, oxaliplatin (L-OHP)-based regimens have resulted in no more than a 20% reduction in disease recurrence in advanced COAD [5]. Thus, identifying the underlying causes and mechanisms driving COAD metastasis and L-OHP resistance is essential.

Cigarettes contain at least 93 harmful substances, including nicotine, nitrosamines and benzene, which significantly elevate the risk of various cancers, including lung cancer and colorectal cancer (CRC) [6,7,8]. A study demonstrated a dose-dependent increase in COAD risk with cigarette consumption, revealing a 7.8-fold increase in risk for each additional 10 cigarettes smoked per day [9]. In addition, smoking has been linked to poorer overall survival (OS) and COAD-specific survival compared to non-smokers [10, 11]. While the molecular mechanisms underlying cigarette-induced carcinogenesis have been extensively explored [6, 12], scientific evidence detailing how smoking specifically promotes COAD development remains sparse.

Growth differentiation factor 15 (GDF15) has recently been recognized as a key member of the glial cell-derived neurotrophic factor (GDNF) family [13]. Under normal physiological conditions, GDF15 expression remains low across most tissue cells; however, its expression significantly increases in response to pathological conditions, such as malignant tumors and inflammation [14]. Research has shown that GDF15 contributes to tumor growth and progression by modulating the phosphorylation of the MAPK and PI3K/AKT signaling pathways [15]. In addition, studies have demonstrated that GDF15 plays a role in promoting tumor chemoresistance [16,17,18]. Despite this, the role and underlying mechanisms of GDF15 in COAD remain underexplored, and the relationship between cigarette smoke and GDF15 in COAD has not been elucidated.

This study initially investigated the impact of cigarette exposure on the prognosis of patients with COAD. Subsequently, a long-term cigarette smoke extract (CSE)-exposed cell model was established to examine the specific mechanisms through which cigarette smoke influences COAD progression. Our findings revealed that prolonged CSE exposure enhanced both metastatic potential and L-OHP resistance in COAD. Mechanistically, cigarette smoke increased GDF15 levels, and GDF15 expression was correlated with poorer COAD prognosis. Inhibition of GDF15 mitigated cigarette-induced L-OHP resistance and metastasis in both in vitro and in vivo models by suppressing the phosphorylation of ERBB2 and AKT. In conclusion, this study identified a specific mechanism by which cigarette exposure accelerates COAD progression.

Materials and methods

Clinical data analysis

Categorical variables (presented as counts/percentages, n/%) were analyzed via χ2 or Fisher's exact tests to evaluate associations between smoking history and clinicopathological parameters (age, gender, TNM stage, differentiation grade, tumor size, and anatomical location). Univariable and multivariable COX regression models were applied to assess smoking status as a prognostic risk factor. Survival outcomes were analyzed by Kaplan–Meier curves with log-rank testing. Approval for the research protocol was granted by the Ethics committee at the First Affiliated Hospital of Anhui Medical University.

CSE preparation

CSE was prepared as previous study [19, 20]. Briefly, the smoke of 25 commercially available Huangshan-brand cigarettes (Tobacco Industrial Corporation, China) in a controlled system, with smoke constituents dissolved in 250 mL DMEM culture medium. The solution was adjusted to a physiological pH of 7.4 and sterilized by filtration through 0.22 μm microporous membrane to eliminate particulate contaminants and microbial agents. The stock solution was defined as 100% CSE, aliquoted, and cryopreserved at − 80 °C until use. Prior to experiments, the stock was serially diluted to desired concentrations for downstream applications.

Cell culture and transfection

HT29 and HCT116 human COAD cell lines were obtained from the American Type Culture Collection (ATCC). The cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Thermo Fisher Scientific) with an addition of 10% fetal bovine serum (FBS; HyClone) at 37 °C under 5% CO₂ humidified atmosphere. According to the manufacturer's guidelines, stable cell lines were created using Lipofectamine 2000 from Thermo Fisher Scientific. Following transfection, cells were subjected to puromycin selection at optimized concentrations to isolate stably transfected clones. Puromycin-resistant colonies were expanded and validated for target gene expression prior to subsequent experiments. The following short hairpin RNA (shRNA) sequences were used for gene silencing: GDF15 shRNA: 5-GAGAGTTGCGGAAACGCTACGTTCAAGAGACGTAGCGTTTCCGCAACTCTC-3. ERBB2 shRNA:5′-GGTGGTCATTGGAATCCTATCAAGAGTAGGATTCCAATGACCACC-3′.

Cell viability assay

The assessment of cell viability was conducted using a Cell Counting Kit-8 (CCK8; Beyotime). In short, cells were placed in 96-well plates with 100 μL of culture medium, either with or without CSE or L-OHP and incubated at 37 °C under 5% CO₂. After adding 10 μL of CCK-8 reagent to each well, the plates were incubated at 37 °C for 2 h, and absorbance was read at 450 nm using a microplate reader (Bio-Tek).

Migration and invasion assays

Experiments on cell migration and invasion were performed using 24-well transwell chambers (Costar 3422; Corning Inc.) equipped with polycarbonate membranes with 8-μm pores. Around 40,000 cells, suspended in DMEM without serum, were placed in the upper chamber. For invasion assays, the membrane was pre-coated with Matrigel from BD Biosciences. DMEM with 10% FBS was used as a chemoattractant in the lower chamber. After being incubated at 37 °C with 5% CO₂ for 24–48 h, cells that did not migrate or invade were wiped off the upper membrane surface using a cotton swab. After reaching the lower surface, the cells were fixed with methanol and stained with 0.1% crystal violet. These migrated or invaded cells were then visualized under an optical microscope (Olympus).

Wound healing assay

HT29 and HCT116 cells were seeded in 6-well plates at a density of 5 × 104 cells per well and cultured in complete DMEM (10% FBS) at 37 °C in a 5% CO2 incubator until reaching confluence. Prior to wounding, the monolayers were washed twice with warmed PBS (pH 7.4) to remove residual serum. A sterile 200 μL pipette tip was used to create uniform linear wounds by applying consistent pressure in a perpendicular direction across the well diameter. Cellular debris was removed through three sequential PBS washes (5 min each) with gentle agitation. Phase-contrast images (Olympus, Tokyo, Japan) were acquired at 0 h and 24 h post-wounding.

Flow cytometry analysis of cell apoptosis

Apoptosis induced by drug treatment was quantified using the Annexin V-Alexa Fluor660/7-AAD Apoptosis Detection Kit (BestBio). Following 48-h exposure to L-OHP (with or without treatment), HT29 and HCT116 cells were processed according to the manufacturer’s protocol for dual staining with Fluor660-conjugated Annexin V and 7-AAD. Apoptotic cells were quantified via flow cytometry analysis. Flow cytometry data were processed using Cytexpert software (Beckman Coulter).

Western blot analysis

The extraction of proteins was performed using RIPA lysis buffer supplemented with 100 μg/mL PMSF (Beyotime), followed by 5-min denaturation at 100 °C. The levels of protein were measured using a bicinchoninic acid (BCA) assay kit from Beyotime. The samples were run on SDS–PAGE gels from Beyotime and then transferred onto PVDF membranes from Merck Millipore. The membranes underwent a 2-h blocking with 5% non-fat milk in TBST (TBS containing 0.1% Tween-20), followed by an overnight incubation at 4 °C with primary antibodies: anti-GDF15 (27455-1-AP, 1:1000; Proteintech, Wuhan, China), anti-E-cadherin (20874-1-AP, 1:1000; Proteintech, Wuhan, China), anti-N-cadherin (22018-1-AP, 1:1000; Proteintech, Wuhan, China), anti-ERBB2 (18299-1-AP, 1:1000; Proteintech, Wuhan, China), anti-p-ERBB2 (#2243, 1:1000; CST, Boston, USA), anti-AKT (#9272, 1:1000; CST, Boston, USA), anti-p-AKT (#4060, 1:1000; CST, Boston, USA), anti-SLC7A11 (DF12509, 1:1000; Affinity, JiangSu, China) and anti-β-Tublin (GB15140-100, 1:1000; Servicebio, Wuhan, China). The membranes were washed three times with TBST before being incubated with HRP-conjugated secondary antibodies for 1.5 h at room temperature. The visualization of protein bands was achieved using enhanced chemiluminescence (ECL; Thermo Scientific), and their quantification was done with ImageJ software.

Glutathione (GSH) assay

Intracellular GSH levels were assessed using a commercial assay kit (Catalog #A006-2-1, Nanjing Jiancheng Bioengineering Institute), following standardized protocols. The enzymatic reaction produced 5-thio-2-nitrobenzoic acid, a chromogenic product with maximal absorbance at 415 nm. Optical density (OD) values were recorded using a microplate reader, with the reaction rate being linearly correlated with GSH concentrations. Final GSH levels were normalized to total protein content, determined using the BCA assay (Beyotime).

Immunohistochemical (IHC) staining

Formalin-fixed, paraffin-embedded (FFPE) tumor samples (human clinical samples and murine xenografts) were sectioned at 4 μm. After deparaffinization and heat-induced antigen retrieval in citrate buffer (pH 6.0), sections were incubated overnight with primary antibodies at 4 °C. The primary antibodies used were anti-GDF15 (1:200; Cat# 27455-1-AP, Proteintech Group), anti-N-cadherin (1:100; Cat# 22018-1-AP, Proteintech Group), and anti-SLC7A11 (1:100; Cat# DF12509, Affinity Biosciences). Immunoreactivity was detected using a polymer-based HRP-conjugated secondary antibody system (Zhongshan Jinqiao, China) followed by chromogenic development with 3,3′-diaminobenzidine (DAB) (Zhongshan Jinqiao, China). Sections were counterstained with Mayer’s hematoxylin, dehydrated through graded alcohols, and mounted with neutral balsam. Digital images were captured using a microscope (Olympus).

Co-immunoprecipitation (Co-IP) assays

To examine GDF15-ERBB2 interactions, cells were lysed in RIPA buffer (50 mM Tris–HCl pH 7.4; 150 mM NaCl; 1% NP-40; 0.5% sodium deoxycholate) containing protease/phosphatase inhibitors. Lysates were pre-cleared with Protein A/G magnetic beads (Thermo Fisher) at 4 °C for 1 h, followed by overnight incubation with 2 μg anti-GDF15 antibody (27455-1-AP, Proteintech, Wuhan, China) or species-matched IgG control (Proteintech, Wuhan, China) for immunoprecipitation. Immune complexes were captured using fresh Protein A/G beads for 2 h at 4 °C, Immune complexes were detected by SDS–PAGE with anti-GDF15 antibody and anti-ERBB2 antibody (18299-1-AP, 1:1000; Proteintech, Wuhan, China).

Animal study

Male BALB/c nude mice (4 weeks) were purchased from SLAC Laboratory Animal (Shanghai). To establish the cell-derived xenograft (CDX) model, HT29-P30 cells or HT29-CSE-P30 cells (2.5 × 106) with or without GDF15 knockout were suspended in 100 µL of 50% Matrigel (35,624, Corning, USA) and inoculated subcutaneously under the left axilla. One week later, PBS or L-OHP (5 mg/kg) was administered intraperitoneally every 3 days. Tumor volume (mm3) was recorded every 3 days using the formula: volume = length × width2 × 1/2. On day 20, all mice were euthanized, tumor weights were measured, and tissues were collected for IHC staining. All animal studies were approved by the Institutional Animal Care and Use Committee of Anhui Medical University, China.

Bioinformatics analysis and RNA-sequence

Gene expression matrices from colon tissues were retrieved from the TCGA and GEO databases. Both data sets underwent log2 transformation of transcripts per kilobase million (TPM) values with a pseudo-count of 0.01 for normalization. GSE174650 included 37 healthy colon organoids exposed to carcinogens from smoking and 37 control colon organoids. Differentially upregulated genes between cancerous and normal tissues from two tumor-related data sets (TCGA and GSE180440) were considered tumor-related genes (logFC > 1, FDR < 0.05). The predictive accuracy of GDF15 for colon cancer was assessed using receiver operating characteristic (ROC) analysis. GSE69657, a matrix data set measuring global gene expression in patients with primary advanced CRC post-chemotherapy, was used to compare GDF15 expression between chemotherapy non-responders and responders via the Wilcoxon test. GSE39582 was used to analyze the impact of GDF15 and SCL7A11 expression on OS and progression-free survival (PFS) in patients with colon cancer. The Tumor Immune Estimation Resource (TIMER) was used to explore the correlations between GDF15 expression and immune cell infiltration levels. Pearson correlation analysis was performed to assess the relationship between GDF15 and SCL7A11 in COAD.

RNA-sequence analysis was conducted as previously described [21].

Statistical analysis

All statistical analyses were performed using R v4.0.5 and GraphPad Prism 7.0. Continuous variables with a normal distribution (presented as mean ± SD) were compared using the Student's t test. Non-normally distributed variables (presented as median [IQR]) were analyzed using the Mann–Whitney U test, with statistical significance set at P < 0.05.

Results

Cigarette was associated with poor prognosis in patients with COAD

Cigarette smoke has been identified as a negative prognostic factor for COAD. A retrospective analysis was performed on the survival and prognostic outcomes of patients with COAD (n = 914) at the Cancer Center of the First Affiliated Hospital of Anhui Medical University over a 5-year period (June 2014–June 2019). Table S1 presents the correlation between smoking history and clinicopathological characteristics in patients with COAD. Smoking history was significantly associated with age, gender and alcohol consumption history, but not with pathological stage (Table S1). To assess the impact of smoking history on COAD prognosis, its role as an independent risk factor for adverse outcomes was evaluated using univariate and multivariate Cox regression analyses. In Models I–IV, other risk factors, including gender, age, alcohol consumption, tumor stage, grade, size and location, were progressively controlled. The results indicated that smoking history was a prognostic factor for poor outcomes in Models I–III. Although Model IV suggested a worse prognosis associated with smoking history, it did not reach statistical significance (Table 1). The result demonstrated that smoking was associated with poor prognosis of COAD. Furthermore, the impact of smoking history on OS was analyzed in patients with COAD (Fig. 1a) and performed stratified analyses based on stages, age and alcohol consumption history (Fig. S1a–f). The results indicated that patients with COAD with a smoking history had significantly poorer overall prognosis compared to non-smokers. However, the mechanisms underlying the negative impact of smoking on COAD prognosis remain unclear.

Table 1 Prognostic value of history of smoking by univariate and multivariate cox regression analysis
Full size table
Fig. 1
figure 1

Cigarette promotes the metastasis and L-OHP resistance of COAD. A Effect of smoking history on OS of COAD patients. B CCK-8 assay to detect the cell viability of HT29 and HCT116 after 48 h treatment with different concentrations of CSE. C Difference in proliferative capacity between HT29, HCT116 cells subjected to long-term CSE stimulation and their control group via CCK8 assay. D Cell scratch assay to detect the difference in cell migration ability between HT29, HCT116 cells subjected to long-term CSE stimulation and their control group. E Transwell assay to detect the difference in cell invasion and migration ability between HT29, HCT116 cells subjected to long-term CSE stimulation and their control group. F, G Effects of long-term CSE stimulation on E-cadherin and N-cadherin protein in HT29, HCT116. Levels of the E-cadherin and N-cadherin proteins were determined using WB assay and quantified using gray scale analysis. H CCK-8 assay to detect the cell viability in HT29, HCT116 cells subjected to long-term CSE stimulation and their control group after treatment with L-OHP of different concentrations for 48 h. I Flow cytometry revealed the apoptosis levels of HT29, HCT116 cells subjected to long-term CSE stimulation and their control group after treatment with/without L-OHP for 48 h (HT29:10 µM; HCT116: 5 µM), *P < 0.05, **P < 0.01, ***P < 0.001. CSE cigarette smoke extract; COAD colon adenocarcinoma; L-OHP oxaliplatin

Full size image

CSE promoted the malignant potential of COAD

To investigate the effects of smoking on the malignant progression of COAD cells, this study utilized two COAD cell lines, HT29 and HCT116, and employed CSE to simulate smoking's impact on COAD cells in vitro. Cells were exposed to varying concentrations of CSE in medium for 48 h, followed by cell viability assessment using the CCK-8 assay. The results showed that cell viability was significantly inhibited in both HT29 and HCT116 cells when exposed to CSE concentrations exceeding 1% and 0.5%, respectively (Fig. 1b). Consequently, long-term treatment was conducted using 1% CSE for HT29 cells and 0.5% CSE for HCT116 cells, with each passage cultured for 48 h over 30 consecutive generations, leading to the establishment of long-term CSE-exposed COAD cells (HT29-CSE-P30 and HCT116-CSE-P30) and control COAD cells (HT29-P30 and HCT116-P30).

A series of malignant phenotypes were subsequently evaluated. Although CCK8 assays indicated that long-term CSE exposure did not promote colon cancer cell proliferation (Fig. 1c), Transwell and wound-healing assays demonstrated that long-term CSE exposure enhanced the invasive potential of colon cancer cells (Fig. 1d, e). To investigate the association between metastatic potential and epithelial–mesenchymal transition (EMT), the expression of EMT markers was assessed. Western blot (WB) analysis revealed that prolonged CSE exposure significantly decreased E-cadherin protein levels in COAD cells, while simultaneously increasing N-cadherin expression (Fig. 1f, g). This study further explored whether long-term CSE exposure could induce L-OHP resistance in COAD cells. HT29-CSE-P30, HCT116-CSE-P30, and control cells (HT29-P30 and HCT116-P30) were treated with L-OHP. CCK8 and apoptosis assays demonstrated that long-term CSE-exposed COAD cells exhibited greater resistance to L-OHP (Fig. 1h–j).

CSE-induced GDF15 upregulation was associated with poor prognosis in COAD

To further investigate the molecular mechanisms underlying the increased malignant potential of colon cancer induced by cigarette exposure, a common data set (GSE174650) containing 37 healthy colon organoids exposed to smoking-related carcinogens and 37 control organoids was analyzed. A total of 114 upregulated genes were identified in GSE174650 (log2FC > 0.5, FDR < 0.05, Supplementary Fig. S2a), with these genes primarily enriched in pathways related to chemical carcinogenesis (Supplementary Fig. S2b). Among the 114 upregulated genes, 8 common genes were found to overlap with tumor-related genes (Fig. 2a). To assess the relevance of these genes to tumor progression, an ROC curve analysis was performed. Of these eight genes, GDF15 exhibited the highest area under the curve (AUC) in both TCGA and GSE180440 data sets (AUC of GDF15 in TCGA: 0.969, AUC of GDF15 in GSE180440: 0.975; Fig. 2b, c). GDF15, a member of the TGF-β superfamily, plays a critical role in tumor progression, metastasis and immune evasion by modulating the tumor microenvironment (TME) and enhancing cancer cell survival. Pan-cancer analysis revealed that GDF15 was overexpressed in a variety of digestive tract cancers, including cholangiocarcinoma, hepatocellular carcinoma, gastric carcinoma, colon carcinoma and rectal carcinoma (Supplementary Fig. S2c). In addition, Fig. 2d illustrates that GDF15 expression was upregulated in colon organoids exposed to carcinogens from smoking.

Fig. 2
figure 2

CSE-induced GDF15 upregulation was associated with poor prognosis in COAD. A Venn diagram to identify key genes by taking the intersection among DEGs in GSE174650 and tumor-related genes. B, C Tumor-predictive capacity of GDF15 in the TCGA and GSE180440 data sets via ROC curves. D Differential GDF15 expression between smoking-related carcinogen-exposed colon tissue samples and control samples in the GSE174650 data set. E Expression of GDF15 in tumor tissues of COAD patients with/without smoking history was detected via IHC assay. F Effects of long-term CSE stimulation on GDF15 protein in HT29, HCT116 were detected by WB. G Comparison of GDF15 expression between chemotherapy response group and chemotherapy no-response group. H OS of COAD patients with high and low GDF15 expression was compared via Kaplan–Meier curves in GSE39582. I PFS of COAD patients receiving chemotherapy with high and low GDF15 expression was compared via Kaplan–Meier curves in GSE39582. J GDF15 expression levels have substantial associations with infiltration levels of CD8+ T lymphocytes, CD4+ T lymphocytes macrophage and macrophage M1. CSE cigarette smoke extract; COAD colon adenocarcinoma; DEGs differentially expressed genes; ROC receiver operating characteristic; IHC immunohistochemistry; OS overall survival; PFS progression-free survival; WB western blotting

Full size image

The association between GDF15 expression and smoking in colon cancer was further validated. IHC results indicated that GDF15 expression was significantly higher in COAD tissues from smokers compared to non-smokers (Fig. 2e). WB analysis revealed a substantial increase in GDF15 levels in COAD cells following long-term exposure to CSE (Fig. 2f, Supplementary Fig. S2d). In addition, the relationship between GDF15 expression and chemotherapy sensitivity was explored. Analysis of GDF15 levels in advanced CRC tissues from chemotherapy responders and non-responders in the GSE69657 data set showed that although not statistically significant, non-responders exhibited higher GDF15 expression (Fig. 2g). Survival analyses further demonstrated that patients with COAD exhibiting elevated GDF15 levels had worse OS compared to those with lower GDF15 expression (Fig. 2h). Furthermore, patients with COAD undergoing chemotherapy with elevated GDF15 levels exhibited significantly shorter PFS (Fig. 2i). Utilizing the TIMER database, the association between GDF15 expression and immune cell infiltration dynamics in COAD was systematically evaluated. Quantitative analysis revealed that elevated GDF15 expression was inversely correlated with the infiltration densities of CD8+ T lymphocytes, CD4+ T lymphocytes, and macrophages (Fig. 2j). Interestingly, reduced infiltration of these immunologically active subsets—a characteristic of immunosuppressive TMEs—has been extensively linked to advanced disease stages and poor clinical outcomes in colorectal malignancies [22]. These results further suggest that GDF15 plays a critical role in contributing to the poor prognosis observed in COAD.

CSE-induced GDF15 upregulation drives metastatic progression and L-OHP resistance in COAD

To investigate the mechanistic role of GDF15 in CSE-induced COAD metastasis and L-OHP resistance, genetically modified cell models were established: GDF15 overexpression in HT29-P30 and HCT116-P30 cells, and GDF15 knockdown in long-term CSE-exposed HT29-CSE-P30 and HCT116-CSE-P30 cells (Fig. 3a, b). CCK8 assays revealed that neither GDF15 overexpression nor knockdown impacted the proliferative capacity of COAD cells (Supplementary Fig. S3a, b). Transwell migration and wound healing assays demonstrated that GDF15 overexpression significantly enhanced metastatic potential, whereas GDF15 knockdown effectively reversed the pro-metastatic effects of chronic CSE exposure (Fig. 3c, d). WB analysis revealed that GDF15 overexpression induced EMT, characterized by a reduction in E-cadherin and an increase in N-cadherin levels. In contrast, GDF15 knockdown restored the expression of epithelial markers in CSE-exposed cells (Fig. 3a). In addition, the impact of smoking-induced GDF15 on L-OHP resistance was explored. Cell viability assays and flow cytometry showed that GDF15 overexpression significantly reduced drug sensitivity to L-OHP, while GDF15 knockdown enhanced apoptosis in CSE-exposed cells upon L-OHP treatment (Fig. 3e, f, Supplementary Fig. S3c–e).

Fig. 3
figure 3

CSE-induced GDF15 upregulation drove metastatic progression and L-OHP resistance in vitro. A, B Overexpression of GDF15 in HT29-P30 cells and HCT116-P30 cells, and knockdown of GDF15 in HT29-CSE-P30 cells and HCT116-CSE-P30 cells. Levels of the GDF15, E-cadherin and N-cadherin proteins were determined using WB assay and quantified using gray scale analysis. C Wound healing assay demonstrated migration capacity (0/24 h imaging). D Transwell invasion/migration assay with crystal violet staining. E, F Flow cytometry analysis of apoptosis inducted by L-OHP (HT29:10 µM; HCT116: 5 µM) for 48 h. *P < 0.05, **P < 0.01, ***P < 0.001. CSE cigarette smoke extract; L-OHP oxaliplatin

Full size image

GDF15 influenced AKT signaling and increased GSH levels, contributing to L-OHP resistance

Subsequently, RNA sequencing analysis of HT29-CSE-P30 cells with GDF15 overexpression and knockdown identified 155 differentially expressed genes (DEGs), including 42 upregulated and 113 downregulated genes in HCT116-CSE-P30 shGDF15 cells (Supplementary Fig. S4a). Enrichment analysis revealed significant pathways involving AKT signaling, GSH metabolism, and cell adhesion processes (Fig. 4a). Given the established role of intracellular GSH metabolism in chemotherapeutic resistance [23], the regulatory role of GDF15 on GSH levels was assessed. Overexpression of GDF15 significantly increased intracellular GSH content in COAD cells, while GDF15 knockdown markedly reduced GSH levels in CSE-exposed cells (Fig. 4b). SLC7A11, a cystine/glutamate antiporter critical for intracellular GSH elevation and chemoresistance in tumor cells [23], was further examined. Bioinformatics analysis revealed a strong positive correlation between SLC7A11 and GDF15 expression in COAD tissues (Fig. 4c), with elevated SLC7A11 expression linked to poorer patient prognosis (Fig. 4d). WB analysis confirmed that GDF15 overexpression upregulated SLC7A11 expression, while GDF15 knockdown in long-term CSE-exposed cells downregulated SLC7A11 expression (Fig. 4f, g). Moreover, previous research has indicated that GDF15 can activate ERBB2, a member of the EGFR family that regulates cellular metastasis via AKT signaling, consistent with our pathway enrichment findings [15]. Co-IP experiments confirmed a physical interaction between GDF15 and ERBB2 (Fig. 4e), while WB analysis demonstrated that GDF15 significantly promoted ERBB2 and AKT phosphorylation (Fig. 4f, g). Given the established relationship between AKT phosphorylation, SLC7A11 expression, and L-OHP resistance, it is hypothesized that GDF15 enhances SLC7A11-mediated L-OHP resistance in colon cancer cells through the activation of the ERBB2/AKT phosphorylation cascade.

Fig. 4
figure 4

GDF15 influenced AKT signaling and increased GSH levels contributing to L-OHP resistance. A Enrichment analysis of DEGs from transcriptome sequencing in HT29-CSE-P30 and HT29-CSE-P30-sh-GDF15 cells. B Effect of GDF15 on intracellular glutathione (GSH) levels in COAD cells. C Correlation analysis between GDF15 and SLC7A11 expression in colon cancer tissues from the GSE39582 data set, assessed using Pearson’s correlation coefficient. D OS of COAD patients stratified by high vs. low SLC7A11 expression in the GSE39582 cohort, analyzed by Kaplan–Meier curves. E Interaction between GDF15 and ERBB2 proteins in HT29 cells, evaluated by co-immunoprecipitation (Co-IP). F, G WB analysis and quantification of ERBB2, p-ERBB2, AKT, p-AKT, and SLC7A11 protein levels under GDF15 modulation. *P < 0.05, **P < 0.01, ***P < 0.001. L-OHP oxaliplatin; DEGs differentially expressed genes; OS overall survival; WB western blotting

Full size image

CSE promoted metastasis and L-OHP resistance of COAD through the GDF15/ERBB2/AKT pathway

To investigate whether CSE promotes metastatic progression and L-OHP resistance in COAD through the GDF15/ERBB2/AKT pathway, long-term CSE-exposed COAD cells were treated with the highly selective AKT allosteric inhibitor MK-2206, the AKT-specific agonist SC79, and subjected to ERBB2 or GDF15 knockdown. WB analysis revealed that MK-2206 treatment significantly increased E-cadherin protein levels, while decreasing N-cadherin, SLC7A11 protein levels, and AKT phosphorylation. Similar effects were observed in cells with ERBB2 or GDF15 knockdown, with a notable decrease in ERBB2 phosphorylation levels in GDF15-knockdown cells. Conversely, addition of SC79 to GDF15-knockdown cells led to a reduction in E-cadherin expression, along with a recovery of N-cadherin, SLC7A11 protein levels, and AKT phosphorylation (Fig. 5a, b). Wound healing and transwell assays demonstrated that treatment with MK-2206 or knockdown of ERBB2/GDF15 significantly impaired the migration and invasion capabilities of long-term CSE-exposed COAD cells, while SC79 partially restored these abilities (Fig. 5c, d). CCK-8 assays and flow cytometry showed that MK-2206 treatment or ERBB2/GDF15 knockdown significantly sensitized long-term CSE-exposed COAD cells to L-OHP, whereas SC79 supplementation promoted chemoresistance (Fig. 5e, f, Supplementary Fig. S5a, b). Intracellular GSH levels were significantly reduced upon MK-2206 treatment or ERBB2/GDF15 knockdown, but were partially restored by AKT agonist addition (Fig. 5g). These results suggest that CSE upregulates GDF15 expression to enhance intracellular GSH biosynthesis via the ERBB2/AKT/SLC7A11 axis, thereby contributing to reduced L-OHP sensitivity.

Fig. 5
figure 5

CSE promoted metastasis and L-OHP resistance through GDF15/ERBB2/AKT pathway in vitro. A Long-term CSE-exposed colon cancer cells were treated with the highly selective AKT allosteric inhibitor MK-2206, AKT-specific agonist SC79, and knocked down ERBB2 or GDF15. B Levels of the GDF15, E-cadherin, N-cadherin, ERBB2, P-ERBB2, AKT, P-AKT and SLC7A11 proteins were determined using WB assay and quantified using gray scale analysis. C Wound healing assay demonstrated migration capacity (0/24 h imaging). D Transwell invasion/migration assay with crystal violet staining. E CCK-8 assay to detect the cell viability in HT29, HCT116 cells after treatment with L-OHP for 48 h. F Flow cytometry analysis of apoptosis inducted by L-OHP (HT29:10 µM; HCT116: 5 µM) for 48 h. G Levels of GSH were evaluated in HT29, HCT116 cells. * P < 0.05, ** P < 0.01, *** P < 0.001. CSE cigarette smoke extract; L-OHP oxaliplatin

Full size image

CSE promoted L-OHP resistance in COAD through GDF15 in vivo

To explore whether CSE promotes L-OHP resistance in COAD cells through GDF15 expression in vivo, mice were divided into three groups and subcutaneously implanted with HT29-P30, HT29-CSE-P30, and HT29-CSE-P30-sh-GDF15 cells in the axillary regions. Seven day post-implantation, mice were intraperitoneally administered 5 mg/kg L-OHP or an equivalent volume of saline every 3 days until day 19, when the experiment was terminated and tumors were dissected. After euthanasia, mice were photographed (Fig. 6a), and representative image of tumors was captured (Fig. 6b). Tumor volume was measured every 3 days during the tumor-bearing phase (Fig. 6c), and tumor weights were recorded (Fig. 6d). No significant differences in tumor size or weight were observed among control groups. However, mice transplanted with HT29-P30 cells and treated with L-OHP showed significantly smaller tumor volumes and weights compared to the HT29-P30 group without L-OHP. In contrast, mice transplanted with HT29-CSE-P30 cells and treated with L-OHP exhibited markedly larger tumor volumes and weights than the HT29-P30 group with L-OHP, suggesting that long-term CSE exposure significantly reduced in vivo sensitivity to L-OHP. Tumor sensitivity to L-OHP was partially restored in the GDF15-knockdown group. Subsequent IHC and WB analyses of tumor tissue revealed protein expression profiles consistent with in vitro observations (Fig. 6e, f). Long-term CSE exposure significantly upregulated GDF15, N-cadherin and SLC7A11 protein levels in COAD cells, whereas GDF15 knockdown markedly reduced their expression. These in vivo results corroborated our in vitro findings, demonstrating that chronic CSE exposure promotes GDF15 expression, reduces tumor sensitivity to L-OHP, and enhances EMT capacity.

Fig. 6
figure 6

CSE promoted L-OHP resistance through GDF15 in vivo. A Representative images of mice at the experimental endpoint. B Gross morphology of resected subcutaneous tumors. C Tumor growth curves over time post-transplantation of HT29-P30, HT29-CSE-P30, and HT29-CSE-P30-shGDF15 with/without L-OHP. D Final tumor weights after being executed. E WB analysis of GDF15, E-cadherin, N-cadherin, and SLC7A11 expression in xenografts. F IHC staining of GDF15, N-cadherin, and SLC7A11 in tumor tissues. CSE cigarette smoke extract; L-OHP oxaliplatin; EMT epithelial–mesenchymal transition; IHC Immunohistochemistry

Full size image

Discussion

Cigarette smoking is a well-established environmental risk factor for colon cancer. A meta-analysis by Edoardo Botteri reveals that smoking increases the risk of colon cancer, with a combined relative risk of 1.18 (95% CI, 1.11–1.25) [9]. Previous studies have also demonstrated a significant association between smoking history and COAD prognosis, with current smokers showing worse OS compared to never-smokers, consistent with our findings [10, 11]. Moreover, numerous studies have investigated the relationship between smoking status and COAD-specific survival [11, 24, 25]. In addition, Fujiyoshi et al. found a consistent interaction between cumulative smoking pack-years at the time of colorectal cancer diagnosis and both colorectal cancer-specific survival and overall survival [26]. Although some studies have suggested that current smoking history is not statistically significant in relation to poor COAD-specific survival [11], the correlation between current smoking and worse COAD-specific survival is generally strongest [24, 25, 27].

Extensive research has shown that smoking significantly accelerates the onset and progression of COAD [6, 28]. According to Kim et al., CSE promotes cell metastasis by altering the expression of EMT markers [28, 29]. Cigarette smoke enhances docetaxel resistance in esophageal adenocarcinoma [30]. However, the precise mechanisms by which cigarette smoke influences L-OHP resistance in COAD remain unclear. To better understand the role of cigarette-induced COAD progression, COAD cells were treated with CSE for 60 days. The results revealed significant enhancement in migration, invasion, and EMT markers, while long-term CSE-exposed cells exhibited notably reduced sensitivity to L-OHP.

For patients with advanced COAD, chemotherapy remains a primary treatment modality. L-OHP-based regimens such as XELOX and FOLFOX are first-line chemotherapy options for COAD [31]. However, resistance to L-OHP significantly impairs treatment efficacy. Chemoresistance in tumor cells is often associated with elevated GSH levels. GSH, through its redox cycle between the reduced form (GSH) and oxidized form (GSSG), directly scavenges reactive oxygen species (ROS) induced by chemotherapeutic agents, thereby protecting tumor cells from oxidative damage [23, 32]. In addition, SLC7A11, a cystine/glutamate transporter, plays a pivotal role in GSH synthesis by regulating cystine uptake, positioning it as a key player in tumor antioxidant defense and chemoresistance. In addition, targeting SLC7A11 reduces GSH levels and enhances chemosensitivity [23, 33]. In the present study, both GSH and SLC7A11 levels were significantly elevated in cells chronically exposed to CSE, suggesting that this mechanism may be a critical factor contributing to smoking-induced L-OHP resistance in COAD.

GDF15, a member of the transforming growth factor-β (TGF-β) superfamily, plays a pivotal role in the progression of metabolic disorders and various cancers [14]. In the present study, cigarette smoking significantly increased GDF15 expression in colon cancer tissues in both experimental and in vivo settings. As a stress-responsive cytokine, GDF15 was influenced by a range of cellular stress signals, including inflammation, hypoxia, tissue damage, ischemia, and diverse cancers [14, 34, 35]. For instance, inflammation can activate the integrated stress response (ISR), in which phosphorylation of eIF2α acts as a central regulatory node. This post-translational modification enhanced the translation of activating transcription factor 4 (ATF4), which in turn directly bind to and activated the GDF15 promoter [36]. In addition, stress-responsive transcription factors such as early growth response protein 1 (EGR1) also play key roles in modulating GDF15 expression [37]. Under conditions of oxidative stress, NRF2 served as a major regulator and transcriptionally controls GDF15 expression via direct promoter binding [38]. Cigarette exposure itself served as an inflammatory stimulus. Chronic exposure to conventional cigarette smoke has been shown to alter intestinal renewal, immune function, and barrier integrity in the mouse ileum and colon, thereby promoting inflammatory responses in the gut [39, 40]. Furthermore, cigarette exposure may also activate other transcription factors, including NRF2, EGR1 and ATF4 [41,42,43]. These previous studies verified the possibility that tobacco can promote the expression of GDF15.

Furthermore, elevated GDF15 levels have been observed in several cancers, including pancreatic cancer and CRC [14]. In CRC, GDF15 binds to TGF-β receptors, promoting EMT in CRC cells and facilitating metastasis [44]. Lin et al. demonstrated that GDF15 induces chemotherapy resistance by modulating the NRF2 feedback loop [45]. Clinical data have shown significantly higher GDF15 levels in the tissues and serum of patients with CRC, with elevated GDF15 expression correlating with reduced OS. Moreover, we found the expression of GDF15 was associated with the inhibition of immune cell infiltration. Previous studies demonstrated that tumor-derived GDF-15 blocks T cell recruitment and suppresses responses to anti-PD-1 treatment [46]. In addition, GDF15 can directly promote macrophage M2 type activation and inhibit macrophage M1 type polarization [47]. These studies suggested GDF15 potential as a novel prognostic marker for CRC [48]. Furthermore, in this study, GDF15 significantly enhanced metastasis and L-OHP resistance in COAD. Next-generation sequencing revealed that knockdown of GDF15 in long-term CSE-stimulated cells led to significant enrichment of DEGs in the AKT signaling pathway and GSH metabolism. WB analysis confirmed that GDF15 substantially enhanced AKT phosphorylation. In addition, previous studies found that GDF15 can be directly or indirectly combined with ERBB2, which promotes AKT phosphorylation by enhancing ERBB2 phosphorylation [15, 49]. Moreover, activation of the AKT pathway was associated with elevated SLC7A11 expression. In this study, Co-IP and WB experiments confirmed that GDF15 interacts with ERBB2 to promote AKT phosphorylation, which subsequently upregulates SLC7A11 expression, contributing to L-OHP resistance.

Several limitations exist in this study. First, the clinical aspect of the research was based on a single-center retrospective analysis. Through the extraction of case files and subsequent follow-ups, we still failed to obtain the specific smoking history of many patients, including the number of packs they smoked each day. This limited our ability to investigate the dose-dependent effects of smoking on COAD progression and prognosis. Second, due to the COVID-19 pandemic, many patients were unable to undergo regular follow-up, restricting our ability to evaluate the impact of smoking history on disease-free survival and chemotherapy sensitivity in patients with COAD. Third, previous literature indicated that GDF15 can exert oncogenic effects by activating MAPK pathways [50]. However, due to the scope and depth of this study, we only selected the most significantly enriched pathway—the AKT pathway—for further investigation. Finally, although the study convincingly demonstrates that chronic CSE exposure leads to increased GDF15 expression in COAD cells and tissues, the precise molecular mechanisms driving this upregulation remain unclear.

A significant association was identified between cigarette smoking and poor prognosis by analyzing the clinical outcomes of patients with colon cancer treated in our hospital. In addition, long-term CSE exposure markedly enhanced GDF15 expression, which promoted ERBB2/AKT phosphorylation, thereby increasing metastatic potential and reducing L-OHP sensitivity in COAD cells. To further validate the GDF15/ERBB2/AKT signaling axis, mice were subcutaneously injected with COAD cells subjected to long-term CSE stimulation or GDF15 knockdown. These findings provide crucial insights into strategies aimed at reversing metastasis and chemotherapy resistance in COAD.

Conclusion

Smoking was identified as a clinical risk factor for adverse outcomes in COAD. Cell and animal experiments demonstrated that CSE promotes metastasis and L-OHP resistance through the GDF15/ERBB2/AKT pathway in COAD (Fig. 7).

Fig. 7
figure 7

Overall mode diagram of CSE promoting the metastasis and L-OHP resistance in COAD. CSE significantly promoted the expression of GDF15 in COAD. GDF15 can promote the phosphorylation of ERBB2 and AKT. Phosphorylated AKT drives colon cancer metastasis by promoting EMT, while simultaneously upregulating SLC7A11 expression to increase intracellular GSH levels, thereby inducing chemoresistance. CSE cigarette smoke extract; L-OHP oxaliplatin; COAD colon adenocarcinoma; EMT epithelial–mesenchymal transition; GSH glutathione

Full size image