METTL3-driven m⁶A modification promotes tumor progression and immune evasion in hypoxic pancreatic cancer by enhancing ADAM10-mediated MICA shedding

Abstract

Background

Pancreatic cancer is characterized by asymptomatic early stages, fast progression, and dismal survival rates. Therefore, a better understanding of immune evasion is required to create more effective treatments. m⁶A methylation is linked to pancreatic cancer, particularly in hypoxia-induced immune escape.

Results

GEO microarray datasets identified pancreatic cancer differentially expressed genes (DEGs) and m⁶A modification regulators. Compared with controls, pancreatic cancer samples showed increased METTL3 levels. METTL3 levels are strongly associated with HIF1A, a key component of the hypoxic tumor microenvironment, suggesting its involvement in tumor development and immune evasion. HIF1A expression levels divided pancreatic tumor tissues into high- and low-HIF1A groups. The high-HIF1A group showed higher METTL3 and MICA levels. Further histopathological, immunohistochemical, and m⁶A modification studies indicated that METTL3 expression and m⁶A modification increased, especially under hypoxic conditions. In normoxic pancreatic cancer cell lines, METTL3 overexpression boosted cell survival, migration, reduced apoptosis, and facilitated immune evasion by lowering MICA-positive cells and increasing sMICA, decreasing NKG2D-positive NK92 cells, and impairing NK92 cell death. Hypoxic METTL3 knockdown reduced cell survival and migration, enhanced apoptosis, increased MICA-positive cells and decreased sMICA levels, and boosted NKG2D-positive NK92 cells and pancreatic cancer cell death, suggesting reduction of immune evasion. Moreover, the in vivo xenograft model further confirmed that METTL3 knockdown reduced tumor growth and increased NK cell infiltration. METTL3 stabilizes ADAM10 mRNA through m⁶A modification. Under hypoxic conditions, ADAM10 overexpression had effects opposite to those of METTL3 knockdown in pancreatic cancer cells and NK92 cells. It partially abolished METTL3’s effects, suggesting that m⁶A modification might mediate METTL3’s effects on hypoxia-related pancreatic cancer immune evasion.

Conclusions

In conclusion, targeting METTL3 or ADAM10, or intervention in the m⁶A modification pathway, may provide novel therapeutic avenues, particularly for pancreatic cancer cells’ immune-evasive tactics.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-08238-3.

Introduction

Pancreatic cancer is characterized by asymptomatic early stages, rapid progression, and poor survival rates. It is among the most aggressive forms of solid malignancies, and its high mortality rate is primarily due to late diagnosis and resistance to conventional therapies [1, 2]. A critical aspect of its pathobiology is immune evasion, a process by which cancer cells evade host immune surveillance [3, 4]. This evasion is mediated by multiple mechanisms, including alterations in major histocompatibility complex (MHC) molecules, downregulation of adhesion molecules and co-stimulatory factors, antigenic variation, apoptosis of immune cells, and secretion of immunosuppressive cytokines [3, 4]. These adaptations enable tumor cells to proliferate and metastasize, underscoring the need for a deeper understanding of immune escape mechanisms in pancreatic cancer to develop more effective therapeutic strategies.

The molecules MICA (MHC class I chain-related protein A) and its soluble form, sMICA, play pivotal roles in pancreatic cancer immune evasion [5]. The shedding of membrane-bound MICA from tumor cell surfaces, a process exacerbated within the tumor microenvironment, leads to sMICA accumulation [6, 7]. sMICA exerts a significant impact on immune surveillance, particularly by engaging and inhibiting the NKG2D receptors on natural killer (NK) cells. The decrease in NKG2D impairs NK cell cytotoxicity, a key line of defense against malignancies, thereby facilitating tumor cell escape from immune eradication [8]. Furthermore, the hypoxic conditions often prevalent in solid tumors such as pancreatic carcinoma play a pivotal role in this context [9]. Hypoxia primarily acts through the transcription factor hypoxia-inducible factor-1α (HIF-1α), which upregulates processes that enhance MICA shedding [10–12]. This hypoxia-driven increase in sMICA production further impairs NK cell-mediated cytotoxicity. It contributes to the broader paradigm of immune escape, thereby underlining the complex interplay between tumor microenvironmental factors and immune evasion mechanisms. Therefore, targeting the hypoxic microenvironment and HIF-1α represents a promising strategy for developing therapies aimed at restoring effective immune surveillance in pancreatic cancer.

N6-methyladenosine (m⁶A) methylation is the most prevalent internal modification within eukaryotic mRNA [13]. Recent research has established a link between m⁶A methylation and immune escape, particularly through the FTO-mediated demethylation of m⁶A sites on PD-L1 mRNA within colorectal tumor cells [14]. This modification alters PD-L1 expression, which is a mechanism now recognized as a key factor in enabling cancer cells to evade immune surveillance. The exploration of m⁶A methylation’s effect on pancreatic carcinoma, particularly in the context of hypoxia-induced immune escape, presents a promising avenue for understanding and potentially disrupting these complex molecular pathways.

This study aims to elucidate the intricate molecular mechanisms underlying immune escape in pancreatic carcinoma, particularly under hypoxic conditions, with a specific focus on the effects of m⁶A methylation.

Materials and methods

Clinical sample collection

A total of 30 paired pancreatic ductal adenocarcinoma (PDAC) tumor tissue samples and the matched adjacent non-cancerous tissue samples were harvested from subjects undergoing surgery at the Xiangya Hospital, Central South University (Changsha, China), and then subjected to liquid nitrogen storage at -80 °C. All procedures were approved by the Ethics Committee of the Xiangya Hospital, Central South University (approval no.: 2022020460).

qRT-PCR

TRIzol reagent (Invitrogen, Carlsbad, USA) was used according to established protocols to extract total RNA from tissues and cells. RNA was reverse-transcribed into cDNA using the FastKing cDNA Synthesis Kit (TIANGEN, Beijing, China). The mRNA levels were detected using the SYBR Green PCR Master Mix from Qiagen according to the manufacturer’s instructions. Using the 2^−ΔΔCT method, the expression of the target gene was calculated by standardizing to that of ACTB. The primer sequences used in qRT-PCR are listed in Table S1.

Histopathological examination

Dehydrated and paraffin-embedded tissues were sliced into 5-µm-thick slices. Xylene (Sigma-Aldrich, St. Louis, USA) was used to dewax the slices, and a graded ethanol series was subsequently used to rehydrate them. For staining, the sections underwent a 5-minute treatment with hematoxylin (Sigma-Aldrich), followed by differentiation in hydrochloric acid alcohol for 30 s. They were then rinsed in tap water for 15 min to ensure thorough bluing. The slices were then stained using eosin (Sigma-Aldrich) for 3 min. After staining, the slices were dehydrated with 75%, 80%, 95% and 100% ethanol and rinsed with xylene. Lastly, slices were mounted using neutral resin and examined under a microscope for histopathological evaluation.

Immunohistochemical staining (IHC staining)

For IHC staining, formalin-fixed, paraffin-embedded sections were deparaffinized in xylene and rehydrated through a series of graded ethanol solutions. For antigen retrieval, slices were incubated for 20 min in an EDTA buffer (pH 8.0) using a microwave heating method. To inhibit endogenous peroxidase activity, the slices were treated with a solution of 100% methanol supplemented with 0.3% hydrogen peroxide (H₂O₂) at room temperature (RT) for 25 min. Subsequently, slices were subjected to an overnight incubation at 4 °C using anti-METTL3 antibody (dilution 1:500, 15073-1-AP, Proteintech, Wuhan, China) and anti-HIF1A antibody (dilution 1:100, 20960-1-AP, Proteintech). The next step involved incubating the sections with an HRP anti-rabbit secondary antibody. The staining process was completed by incubating the sections in DAB solution for 5–10 min at room temperature. METTL3 and HIF1A protein expression was evaluated under light microscopy. The relative protein contents were determined from average optical density (AOD) values obtained using ImageJ software (National Institutes of Health, Bethesda, USA).

Total m⁶A levels

The EpiQuik m⁶A RNA Methylation Quantitation Kit (Epigentek, Farmingdale, USA) was used to quantify total m⁶A levels within total RNA isolated from target tissues or cells. This colorimetric assay was conducted strictly as directed by the manufacturer. The optical density (OD) value at 450 nm of each sample was determined. Subsequently, m⁶A levels were accurately determined by plotting the OD values against a standard curve prepared as directed in the kit.

Cell lines and cultivation

Human pancreatic cancer cell lines PANC-1 (CRL-1469) and MIA PaCa-2 (CRM-CRL-1420) were procured from ATCC (Manassas, USA) before being cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM; 30-2002, ATCC) containing 10% Fetal Bovine Serum (FBS; GIBCO, Grand Island, USA). All cell types were cultivated at 37 °C in a humidified atmosphere containing 5% CO₂.

Cell infection

METTL3 overexpression or ADAM10 overexpression was achieved by transfecting cells using lentivirus overexpressing METTL3 (lv-METTL3) or lv-ADAM10, respectively; lv-NC was used as the negative control. Cells were transfected using lentivirus containing two short hairpin RNAs (shRNAs) against METTL3 (lv-sh-METTL3-1/-2) or one shRNA of ADAM10 (lv-sh-ADAM10) to achieve METTL3 or ADAM10 knockdown, respectively; lv-sh-NC was used as the negative control. Infection was performed using OriTrans-LV (Ori-Bio, Changsha, China), which was used as directed by the manufacturer for transfection. The sequences of overexpressing vectors and shRNA vectors are listed in Table S1.

Hypoxia exposure

Untransfected or transfected cells were plated into 6-well plates and incubated under normoxia (20% O₂, 5% CO₂, 37 °C) or hypoxia (1% O₂, 5% CO₂, 37 °C) for 24 h using a controlled hypoxia incubator [15], then harvested for subsequent analysis.

Immunoblotting

The ice-cold radioimmunoprecipitation assay (RIPA) buffer, composed of 150 mM NaCl, 10 mM Tris (pH 8.0), 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, and 5 mM EDTA, was used to lyse cells. Concurrently, the tissue samples were homogenized within ice-cold Dulbecco’s Phosphate Buffered Saline (DPBS) containing protease and phosphatase inhibitors to extract proteins from target tissues or cells. The BCA Protein Assay kit (Pierce, Rockford, USA) was used to quantify the protein contents in the lysates. Following electrophoresis by 12% SDS-PAGE, the separated 50 µg protein was electroblotted from the gel to PVDF membranes (Merck Millipore, Billerica, USA), followed by an overnight incubation at 4 °C using primary antibodies specifically targeting METTL3 (15073-1-AP, Proteintech), HIF1A (20960-1-AP, Proteintech), ADAM10 (ab39177, Abcam, Cambridge, UK), p-PI3K (AF3242, Affinity Bioscience), PI3K (60225-1-Ig, Proteintech), AKT (10176-2-AP, Proteintech), p-AKT (66444-1-Ig, Proteintech), STAT3 (10253-2-AP, Proteintech), p-STAT3 (28945-1-AP, Proteintech) and β-actin (4967S, Cell Signaling Technology, Danvers, USA). The blots were then incubated with HRP-conjugated secondary antibody (1:5000). Signal visualization was performed using ECL Substrates (Merck Millipore), normalizing to endogenous β-actin.

Cell counting kit-8 (CCK8) assay

The CCK8 assay was conducted to assess pancreatic tumor cell viability. After transfection and exposure to either normoxic or hypoxic conditions, cells were plated into 96-well plates (5 × 10³ cells/well) and incubated at 37 °C for 24 h. Each well was supplemented with 1 10 µl of CCK-8 reagent (Beyotime, Shanghai, China) and incubated for 2 h. The optical density (OD) of each well, indicative of cell viability, was determined with a microplate reader (BioTek, Winooski, USA) at 450 nm.

Migration ability assessment

The transwell chamber (8-µm pore size, Corning Incorporated, Corning, USA) was used to assess cell migration. After transfection and exposure to either normoxic or hypoxic conditions, 1 × 10⁴ pancreatic tumor cells in 200 µL serum-free DMEM were seeded on the upper chamber. The complete medium (600 µL) was added to the lower chamber. 24 h later, cells that migrated to the bottom surface of the upper chamber were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, and observed under a microscope (Olympus, Tokyo, Japan).

A wound healing assay was performed to evaluate the migratory capacity of pancreatic tumor cells. The cells were seeded into 6-well plates and allowed to grow to confluence. A uniform wound was then introduced into the cell layer using a sterile 200-µl pipette tip. Cell migration was monitored by capturing images of the wound area at 0- and 24-hours post-wounding under a microscope. For quantitative analysis, wound areas were measured using ImageJ software (NIH, Bethesda, USA). The wound closure rate was calculated as: Wound healing rate (%) = (initial wound area − remaining wound area) / initial wound area × 100%.

Xenograft animal model

All animal experiments were approved by the Animal Ethics Committee of Xiangya Hospital, Central South University (approval no.: 2022020460). Male BALB/c nude mice (6–8 weeks old) were used for the study. PANC-1 cells were transfected with lv-sh-METTL3 or lv-sh-NC. A total of 5 × 10⁶ transfected cells were suspended in 100 µL of PBS and subcutaneously injected into the right flank of each mouse. The mice were randomly assigned to three groups: Control (uninjected or PBS), lv-sh-NC, and lv-sh-METTL3. Tumor volume was monitored every five days and calculated using the formula: Volume = (length × width²)/2. After 28 days, the mice were euthanized, and the tumors were excised, weighed, and photographed. A portion of each tumor was fixed in formalin for histopathological analysis, while the remaining part was processed for flow cytometry.

Flow cytometry for tumor-infiltrating NK cells

To produce single-cell suspensions, fresh tumor tissue was minced into small pieces and digested in DMEM medium containing collagenase IV (300 U/mL) and hyaluronidase (200 U/mL) at 37 °C for 1–2 h. The cell suspension was then centrifuged at 350 × g for 5 min at 4 °C. The resulting cells were blocked with FACS Buffer containing mouse Fc block for 30 min and then stained with fluorescently-conjugated antibodies against CD3 (APC labeled, E-AB-F1013E, Elabscience, Wuhan, China) and NKp46 (PE labeled, E-AB-F1182D, Elabscience) to identify tumor-infiltrating natural killer (NK) cells (CD3⁻NKp46⁺). The analysis was performed using a NovoCyte flow cytometer (Agilent Technologies, Santa Clara, USA).

NK cell killing assay

PANC-1 and MiA PaCa-2 cells were transfected, followed by a 24-h incubation in either normoxia or hypoxia, respectively. The NK92 cells (Procell, Wuhan, China) were then incubated with pancreatic cancer cells at 2.5:1 (Effector/Target E/T) for 24 h. After the co-culture period, the supernatant was removed, and the cells were subjected to a 4-h incubation using 20 µL of CCK-8 solution. The optical density (OD) was then measured at 450 nm. For reference, the survival rate of the PANC-1/MIA PaCa-2 cell line under normoxic conditions, without NK cell incubation, was set as the 100% baseline.

Cell apoptosis determination

The Annexin V-FITC Apoptosis Detection Kit (Beyotime) was used in conjunction with flow cytometry to determine the apoptotic rate of pancreatic cancer cells. After different treatments, pancreatic cancer cells were further incubated with 1 µM gemcitabine (Sigma-Aldrich) for 24 h. Upon completion of the incubation period, the collected cells were resuspended in 200 µL of binding buffer and then stained concurrently with 10 µL of Annexin V-FITC and Propidium Iodide (PI). A flow cytometer (Novocyte) was subsequently used to analyze the stained cells. The proportion of cells located in both the upper and lower right quadrants of the flow cytometry scatter plot was calculated to quantify the percentage of apoptotic cells.

Flow cytometry for surface antigen determination

Pancreatic cancer cells were subjected to transfection followed by exposure to either normoxia or hypoxia. Following harvesting, these cells were subjected to 30-min incubation at 4 °C in light-deprived conditions using an Alexa Fluor 488-conjugated antibody (anti-human monoclonal MIC A/B, Biolegend, San Diego, USA). NK92 cells were co-cultured with conditioned medium from pancreatic tumor cells under different conditions for 24 h. Post co-culture, NK92 cells were collected, followed by 30-min incubation at 4 °C in light-deprived conditions using a FITC-conjugated antibody (anti-human monoclonal NKG2D, Biolegend). The cells were subsequently rinsed in PBS and determined using flow cytometry (Novocyte) to assess MICA expression on pancreatic cancer cells and NKG2D expression on NK92 cells.

ELISA

The supernatant from the cell culture was collected to quantify soluble MICA (sMICA) concentrations using a sandwich enzyme-conjugated immunosorbent assay (ELISA) as directed by the manufacturer (Multi Sciences, Hangzhou, China). A microplate reader (Roche, Nutley, USA) was used to measure the OD value of each sample at 450 nm. Based on these OD values, sMICA concentration within the culture media was calculated, with the results expressed in terms of sMICA concentration (pg/mL) present in the culture medium.

Bioinformatic analysis

The microarray expression data have been deposited in the Gene Expression Omnibus (GEO) database with accession numbers GSE15471 [16], GSE43797 [17], and GSE56560 [18]. The GSE15471 dataset includes the gene expression profiling of 36 PDAC tumors and matching normal pancreatic tissue samples. The GSE43797 dataset includes the gene expression profiling of 5 PDAC tissues as well as 5 non-neoplastic pancreas tissue samples. The GSE56560 dataset includes the gene expression profiles of 28 PDAC tissues and 4 adjacent pancreatic tissues. Differentially expressed genes (DEGs), including differentially expressed m⁶A genes, were screened using Student’s t-test (adj.p.val < 0.05) accompanied by |log₂ (fold change)| > 0.4 using the DESeq2 tool. Gene set enrichment analysis (GSEA) was performed to evaluate the enrichment of m⁶A-related pathways across three pancreatic cancer cohorts (GSE15471, GSE43797, and GSE56560) using the clusterprofiler R package. The expression data of ADAM10 in pancreatic cancer and normal tissues were analyzed using the GEO datasets GSE15471, GSE43797, and GSE56560. To predict potential m⁶A modification sites on the ADAM10 transcript, the full-length mRNA sequence of human ADAM10 was obtained from the NCBI GenBank database. The sequence was then analyzed using the SRAMP (Sequence-based RNA Adenosine Methylation Site Predictor) online tool, and potential methylation sites were identified based on the comprehensive prediction score.

ADAM10 mRNA stability

PANC-1 and MIA PaCa-2 cell lines were subjected to transfection with lv-METTL3/lv-sh-METTL3. Post-transfection, these cells were subjected to either normoxic or hypoxic conditions. Concurrently, PANC-1 and MIA PaCa-2 cell lines were subjected to treatment using actinomycin D (ActD, 5 µg/ml, Sigma-Aldrich) to inhibit transcription. At predetermined time points of 0, 2, 4, and 6 h following ActD treatment, total RNA was isolated from both PANC-1 and MIA PaCa-2 cell lines. The reverse transcription quantitative PCR (RT-qPCR) assays were then performed to quantify ADAM10 mRNA levels.

MeRIP assay

The level of m⁶A modification in ADAM10 mRNA was evaluated using the Magna MeRIP m⁶A Kit (Merck Millipore) as directed by the manufacturer. Total RNA was initially extracted, purified, and then fragmented. This fragmented RNA was subsequently incubated with magnetic beads conjugated to an m⁶A-specific antibody, thereby facilitating selective enrichment of m⁶A-modified RNA. The methylated RNA was eluted from the beads and further purified. The levels of ADAM10 mRNA in these enriched fractions were quantified by reverse transcription quantitative PCR (RT-qPCR). Equal amounts of total RNA from each sample were used for these assays to ensure consistency.

Colony formation assay

After transfection, cells were seeded into 6-well plates at a density of 500 cells per well and cultured in complete DMEM medium. The medium was replaced every three days. After 14 days of incubation, the cells were washed with PBS, fixed with 4% paraformaldehyde for 15 min, and then stained with 0.1% crystal violet solution for 20 min. The plates were then washed, air-dried, and photographed. Colonies containing more than 50 cells were counted for quantitative analysis.

Statistical analysis

The cell experiments were performed for three biological replicates. The animal experiments were performed for six biological replicates. Experimental data from individual experiments were presented as the mean ± standard deviation (SD). SPSS 21.0 (IBM, Armonk, USA) was used for data analysis. Comparisons between groups were conducted using the Student’s t-test. Comparisons among groups were conducted using one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. P < 0.05 was deemed a statistically significant difference.

Results

Selection of deregulated m6A regulators in pancreatic cancer

By using pancreatic cancer-related expression profile datasets from the GEO database (GSE15471, GSE43797, and GSE56560), differential gene analysis was conducted to identify m⁶A modification regulators. Fig. S1A-C shows DEGs between pancreatic tumor and normal samples according to GSE15471, GSE43797, and GSE56560, which were obtained 9423, 9795, and 9372 differential genes, respectively. GSEA was subsequently performed to evaluate whether m⁶A-related pathways were consistently dysregulated across multiple pancreatic cancer cohorts. Using ranked expression profiles from three datasets (GSE15471, GSE43797, and GSE56560), an m⁶A-associated gene set was observed to be significantly enriched in tumor samples compared with controls in all datasets (Fig. S1D). Then, overlapping differentially expressed m⁶A regulators in pancreatic cancer samples were analyzed according to GSE15471, GSE43797, and GSE56560 datasets, and only METTL3 was identified (Fig.S1E). According to the GSE15471, GSE43797, and GSE56560 datasets, METTL3 expression was markedly upregulated in pancreatic tumor tissues compared to non-cancerous normal controls (Fig. S1F). Consistently, in collected pancreatic cancer samples, METTL3 level was significantly upregulated compared to normal controls (Fig. S1G). Based on GSE15471, GSE43797, and GSE56560 datasets, HIF1A, a key factor in the hypoxic microenvironment, was also significantly upregulated in pancreatic cancer samples (Fig. S1H). According to the GSE15471 dataset, METTL3 was positively correlated with HIF1A (Fig. S1I). Therefore, METTL3 was selected for further investigation.

METTL3 and H1F1A expression in pancreatic cancer

The histopathological alterations in collected pancreatic cancer tissues were evaluated with H&E staining (Fig. 1A). IHC staining revealed that both METTL3 and HIF1A proteins were markedly elevated in PDAC tissues compared with adjacent non-tumor pancreatic tissues (Fig. 1B), with stronger staining intensity observed in tumor samples (Fig. 1C). Moreover, Western blot analysis from five paired PDAC and adjacent normal tissues demonstrated significantly higher levels of METTL3 and HIF1A in tumor specimens (Fig. 1D). To confirm the correlation of METTL3 and hypoxia and immune evasion, qRT-PCR was performed to determine HIF1A expression in pancreatic tumor tissues, and pancreatic tumor tissues were separated into two groups (high-HIF1A, H-HIF1A; low-HIF1A, L-HIF1A) using the median expression of HIF1A mRNA as a cutoff (Fig. 2E). In H-HIF1A group, METTL3 and MICA mRNA levels were higher than the L-HIF1A pancreatic cancer samples (Fig. 1E). Furthermore, IHC staining results revealed METTL3 protein level was markedly increased in H-HIF1A group compared to L-HIF1A group (Fig. 1F). Within tissues, METTL3 exhibited a positive correlation with both HIF1A and MICA levels (Fig. 1G). Moreover, the m⁶A modification levels of H-HIF1A samples were significantly higher than in the L-HIF1A pancreatic cancer samples (Fig. 1H). Next, the PANC-1 and MIA PaCa-2 cell lines were exposed to normoxic or hypoxic conditions, and METTL3, HIF1A, m⁶A modification levels and cell viability were determined. Compared with normoxia, the protein levels of METTL3 and HIF1A were significantly increased under hypoxic conditions (Fig. 1I). Consistent with the tissue sample results, hypoxic exposure significantly increased m⁶A modification levels (Fig. 1J). Moreover, hypoxic stimulation notably promoted cell viability of PANC-1 and MIA PaCa-2 cells (Fig. 1K).

Fig. 1.

HMETTL3 expression in pancreatic cancer. (A) Histopathological alterations in pancreatic cancer patients and adjacent controls’ pancreatic tissue samples were evaluated using H&Estaining. Scale bar = 50 μm. (B) The protein levels and distribution of METTL3 and HIF1A in pancreatic cancer and adjacent control tissues were examined using Immunohistochemical (IHC) staining. Scale bar = 50 μm. (C) Quantification of METTL3 and HIF1A protein levels across groups by average optical density (AOD) using ImageJ software. (D) The protein levels of METTL3 and HIF1A in pancreatic cancer and adjacent control tissues were detected by Immunoblotting. (E) Pancreatic cancer samples were divided into high-HIF1A (H-HIF1A) and low-HIF1A (L-HIF1A) groups using the median expression of HIF1A as a cutoff. Then, the mRNA expression of HIF1A, METTL3, and MICA in H-HIF1A and L-HIF1A groups was determined using qRT-PCR. (F) The protein levels and distribution of METTL3 in H-HIF1A and L-HIF1A pancreatic cancer samples were examined using IHC staining. Scale bar = 50 μm. (G) The correlation of METTL3 with HIF1A and MICA expression in tissue samples was analyzed using Pearson’s correlation analysis. (H) The m6A modification levels of H-HIF1A and L-HIF1A pancreatic cancer samples were examined using the EpiQuik m6A RNA Methylation Quantitation Kit. (I-J) PANC-1 and MIA PaCa-₂ cells were exposed to normoxic (20% O₂) or hypoxic (1% O₂) conditions and examined for (I) HIF1A and METTL3 protein levels by Immunoblotting; (J) the m6A modification levels using the EpiQuik m6A RNA Methylation Quantitation Kit; (K) cell viability using CCK-8 assay. N = 5 (biological replicates) for (A, B, C, D, F), 15 (biological replicates) for (E, G, and H), and 3 (biological replicates) for (I, J and K). Data are analyzed using an unpaired Student’s t-test. ** P < 0.01 compared to adjacent or L-HIF1A or 1% O₂ groups

Fig. 2.

METTL3 promotes pancreatic cancer cell viability and inhibits apoptosis under normoxic and hypoxic conditions. (A) METTL3 overexpression or knockdown were achieved in PANC-1 and MIA PaCa-2 cells by transfecting lentivirus overexpressing METTL3 (lv-METTL3) or lv-sh-METTL3-1/-2, respectively. METTL3 expression in PANC-1 and MIA PaCa-2 cells transfected with lv-METTL3 under normoxia, and with lv-sh-METTL3-1/-2 under hypoxia were detected by Immunoblotting. The lv-sh-METTL3-2 was selected for subsequent experiments. (B) Cell viability was assessed by CCK-8 assay in PANC-1 and MIA PaCa-2 cells under normoxia with METTL3 overexpression and under hypoxia with METTL3 knockdown. (C) The m⁶A modification levels in PANC-1 and MIA PaCa-2 cells under normoxia (lv-NC or lv-METTL3) and hypoxia (lv-sh-NC or lv-sh-METTL3) were determined using the EpiQuik m⁶A RNA Methylation Quantitation Kit Relative m⁶A levels. (D) Cell apoptosis was detected by flow cytometry in gemcitabine-treated cells under normoxia or hypoxia with corresponding METTL3 modulation. N = 3 (biological replicates) for each group. Data are analyzed using one-way ANOVA followed by Tukey’s post hoc test. ** P < 0.01 compared to Normoxia + lv-NC or Normoxia + Gem+lv-NC groups; ## P < 0.01 compared to Hypoxia + lv-sh-NC or Hypoxia + Gem+lv-NC groups

METTL3 promotes pancreatic cancer cell viability and inhibits apoptosis under normoxia and hypoxia

To investigate the role of METTL3 in pancreatic cancer progression, its expression was modulated in PANC-1 and MIA PaCa-2 cells under both normoxic and hypoxic conditions. Western blot analysis confirmed effective METTL3 overexpression (lv-METTL3) under normoxia and efficient knockdown (lv-sh-METTL3-1/-2) under hypoxia, with lv-sh-METTL3-2 showing the highest silencing efficiency (Fig. 2A). Functional assays revealed that METTL3 overexpression significantly enhanced cell viability under normoxia, whereas METTL3 knockdown under hypoxia markedly reduced cell viability compared with corresponding controls (Fig. 2B). Moreover, m⁶A levels were markedly increased following METTL3 overexpression under normoxia, whereas METTL3 knockdown under hypoxia significantly reduced m⁶A levels compared with corresponding controls (Fig. 2C). Notably, under gemcitabine treatment, compared with normoxia, hypoxia significantly inhibited cell apoptosis; and METTL3 overexpression suppressed apoptosis in normoxic cells, while METTL3 depletion under hypoxia significantly increased apoptotic rates (Fig. 2D). Collectively, these results demonstrated that METTL3 promoted pancreatic cancer cell survival and conferred resistance to apoptosis across both oxygen conditions, likely through its regulation of m⁶A modification.

METTL3 promotes migration in pancreatic cancer cells

Given the role of METTL3 in cell viability and apoptosis, its effect on the migratory capacity of pancreatic cancer cells was investigated. The impact of METTL3 was assessed using both Transwell and wound healing assays. Under normoxic conditions, the overexpression of METTL3 in PANC-1 and MIA PaCa-2 cells resulted in a significant increase in cell migration compared to the negative control group (Fig. 3A-B). To understand its function in a simulated tumor microenvironment, METTL3 was knocked down in cells cultured under hypoxia. The results showed that while hypoxia itself enhanced cell migration, knockdown of METTL3 significantly attenuated this hypoxia-induced migratory ability (Fig. 3A-B). These findings indicate that METTL3 plays a crucial role in promoting pancreatic cancer cell migration.

Fig. 3.

METTL3 promotes migration in pancreatic cancer cells. The effect of METTL3 on the migration of PANC-1 and MIA PaCa-2 cells was assessed under normoxic (METTL3 overexpression) and hypoxic (METTL3 knockdown) conditions. (A) Cell migration was measured using a Transwell assay. (B) A wound healing assay was performed to evaluate cell migratory potential. N = 3 for each group. Data are analyzed using one-way ANOVA followed by Tukey’s post hoc test. ** P < 0.01 compared to Normoxia + lv-NC group; ## P < 0.01 compared to Hypoxia + lv-sh-NC group

METTL3 knockdown suppresses tumor growth and enhances immune infiltration in vivo

To validate the functional role of METTL3 in vivo, a xenograft model was established using PANC-1 cells transfected with lv-sh-METTL3 or lv-sh-NC. As shown in Fig. 4A-B, mice in the METTL3 knockdown (lv-sh-METTL3) group exhibited significantly reduced tumor growth, with smaller tumor volumes and lower final tumor weights compared to the lv-sh-NC and control groups. Histopathological analysis by H&E staining revealed the tissue structure of the tumors (Fig. 4C). To confirm the target knockdown in vivo, IHC staining was performed, which showed that METTL3 protein levels were substantially lower in tumors from the lv-sh-METTL3 group than in the control groups (Fig. 4D-E).

Fig. 4.

METTL3 knockdown inhibits tumor growth and enhances immune infiltration in a xenograft model. PANC-1 and MIA PaCa-2 cells transfected with lv-sh-METTL3 or lv-sh-NC were subcutaneously injected into nude mice. (A) Representative images of tumors excised from mice in the Control, lv-sh-NC, and lv-sh-METTL3 groups at the end of the experiment. (B) Quantification of tumor volume and final tumor weight. (C) Representative H&E staining of tumor tissue sections. (D-E) Representative images of immunohistochemical (IHC) staining for METTL3 in tumor tissues. Quantification of METTL3 protein expression by average optical density (AOD) using ImageJ software. (F) Flow cytometry analysis showing the percentage of tumor-infiltrating NK cells. N = 6 (biological replicates) for each group. Data are analyzed using one-way ANOVA followed by Tukey’s post hoc test. ** P < 0.01 compared to lv-sh-NC group

To investigate whether METTL3 knockdown affects the tumor immune microenvironment, tumor-infiltrating NK cells were analyzed by flow cytometry. The results demonstrated that tumors from the lv-sh-METTL3 group had a significantly higher proportion of infiltrating NK cells than those from the control and lv-sh-NC groups (Fig. 4F). These in vivo findings suggest that inhibiting METTL3 suppresses pancreatic tumor growth and promotes an anti-tumor immune response by enhancing NK cell infiltration.

METTL3 modulates pancreatic cancer immune evasion via the MICA/NKG2D axis

To determine how METTL3 affects the immune evasion of pancreatic cancer cells, key molecules in the natural killer (NK) cell recognition pathway were examined. First, under normoxic conditions, METTL3 overexpression in PANC-1 and MIA PaCa-2 cells significantly decreased the percentage of cells positive for surface MICA (MICA⁺) (Fig. 5A) while increasing the concentration of soluble MICA (sMICA) in the culture medium (Fig. 5B). Conditioned medium from these METTL3-overexpressing cells significantly reduced the expression of the activating receptor NKG2D on NK92 cells (Fig. 5C). Consequently, the cytotoxic ability of NK92 cells to kill pancreatic cancer cells was significantly impaired (Fig. 5D).

Fig. 5.

METTL3 modulates pancreatic cancer cell immune evasion. A-D: Effects of METTL3 overexpression under normoxia. PANC-1 and MIA PaCa-2 cells overexpressing METTL3 were analyzed. (A) The percentage of MICA-positive cells was determined by flow cytometry. (B) sMICA concentration in the culture supernatant was measured by ELISA. (C) The percentage of NKG2D-positive NK92 cells was measured by flow cytometry after incubation with conditioned medium (CM) from cancer cells. (D) The killing rate of NK92 cells against pancreatic cancer cells was determined by CCK-8 assay. E-H: Effects of METTL3 knockdown under hypoxia. PANC-1 and MIA PaCa-2 cells with METTL3 knockdown were analyzed. (E) The percentage of MICA-positive cells was determined by flow cytometry. (F) sMICA concentration was measured by ELISA. (G) The percentage of NKG2D-positive NK92 cells was measured after incubation with CM. (H) The NK cell killing rate was determined. N = 3 (biological replicates) for each group. Data are analyzed using an unpaired Student’s t-test. ** P < 0.01 compared to Normoxia + lv-NC, Normoxia + lv-NC-CM, Hypoxia + lv-sh-NC or Hypoxia + lv-sh-NC-CM groups

These effects were reversed under hypoxic conditions by knocking down METTL3. In the hypoxic environment, METTL3 knockdown led to an increased percentage of MICA-positive pancreatic cancer cells (Fig. 5E) and a decrease in sMICA levels (Fig. 5F). Furthermore, conditioned medium from METTL3-knockdown cells significantly increased the number of NKG2D-positive NK92 cells (Fig. 5G). Consistent with these changes, the knockdown of METTL3 under hypoxia significantly enhanced the killing effects of NK92 cells on both PANC-1 and MIA PaCa-2 cells (Fig. 5H).

METTL3 enhances ADAM10 expression and activates PI3K/AKT/STAT3 signaling

To uncover the mechanism by which METTL3 promotes immune evasion, its downstream target was investigated. ADAM10 is a protein known to be involved in MICA shedding. The analysis of datasets GSE15471, GSE43797, and GSE56560 revealed that ADAM10 expression was significantly upregulated in pancreatic tumor samples compared to normal controls (Fig. 6A-C). This clinical relevance was confirmed in our collected tissue samples, where qRT-PCR showed significantly higher ADAM10 mRNA levels in pancreatic cancer tissues compared to non-cancerous samples (Fig. 6D).

Fig. 6.

METTL3 regulates ADAM10 expression via m⁶A modification and activates the PI3K/AKT/STAT3 pathway. (A-C) ADAM10 expression in pancreatic cancer and normal samples from GEO datasets GSE15471, GSE43797, and GSE56560. (D) qRT-PCR analysis of ADAM10 mRNA in collected pancreatic ductal adenocarcinoma (PDAC) (N = 30) and non-cancerous tissues (N = 30). (E-F) qRT-PCR analysis of ADAM10 mRNA in cells with METTL3 overexpression (normoxia) or knockdown (hypoxia). (G) Immunoblotting analysis of ADAM10 protein levels after METTL3 modulation. (H) ADAM10 mRNA stability was measured by qRT-PCR at different time points after treatment with actinomycin D. (I) MeRIP-qPCR showing m⁶A enrichment of ADAM10 mRNA following METTL3 modulation. (J) Prediction of the m⁶A modification site in ADAM10 mRNA by the SRAMP tool. Three very high-confidence m⁶A sites were identified at nucleotide positions 794, 7821, and 7955; position 794 is located in the coding sequence (CDS) of ADAM10, while positions 7821 and 7955 reside in the 3′ untranslated region (3′ UTR). (K) Immunoblotting for key proteins in the PI3K/AKT/STAT3 pathway in cell lysates after METTL3 modulation. (L) Immunoblotting for the PI3K/AKT/STAT3 pathway in tumor tissues from the xenograft model. N = 3 (biological replicates) for (E-L). Data are analyzed using unpaired Student’s t-test for (A-F) and one-way ANOVA followed by Tukey’s post hoc test for (G-L). ** P < 0.01 compared to Normoxia + lv-NC group; ## P < 0.01 compared to Hypoxia + lv-sh-NC or lv-sh-NC groups

To confirm direct regulation, METTL3 expression was modulated in PANC-1 and MIA PaCa-2 cells. Under normoxia, METTL3 overexpression significantly increased both the mRNA and protein levels of ADAM10 (Fig. 6E, G). Conversely, under hypoxic conditions, METTL3 knockdown resulted in a significant decrease in ADAM10 mRNA and protein levels (Fig. 6F, G). These findings suggest METTL3 positively regulates ADAM10 expression.

To determine if this regulation occurs via m⁶A modification, the bioinformatic analysis was performed using the SRAMP tool, which predicted three very high-confidence m⁶A modification sites in the ADAM10 mRNA sequence (position 794 in the coding sequence [CDS] and positions 7821 and 7955 in the 3′ untranslated region [3′ UTR]) (Fig. 6J), supporting the likelihood of m⁶A modification affecting ADAM10 post-transcriptional regulation. This assumption was tested by assessing ADAM10 mRNA stability. An actinomycin D chase assay revealed that METTL3 overexpression increased the stability of ADAM10 mRNA, while METTL3 knockdown decreased its stability (Fig. 6H). Finally, a MeRIP-qPCR assay confirmed that METTL3 overexpression enhanced the m⁶A modification of ADAM10 mRNA, whereas METTL3 knockdown reduced it (Fig. 6I). To further clarify the directionality between m⁶A modification and ADAM10, we examined whether ADAM10 suppression could reciprocally affect m⁶A levels. ADAM10 knockdown in PANC-1 and MIA PaCa-2 cells under hypoxia was confirmed at the protein level (Fig. S2A). Notably, ADAM10 depletion did not significantly alter m⁶A enrichment on ADAM10 mRNA compared with control cells (Fig. S2B). These findings suggest that ADAM10 does not regulate m⁶A modification in a feedback manner, supporting a unidirectional regulatory axis in which m⁶A modification functions upstream of ADAM10.

To explore the signaling pathways downstream of the METTL3/ADAM10 axis, the PI3K/AKT/STAT3 pathway was examined. In vitro, METTL3 overexpression increased the phosphorylation of PI3K, AKT, and STAT3, while its knockdown under hypoxia suppressed their hypoxia-induced phosphorylation (Fig. 6K). This finding was validated in vivo, with tumors from the METTL3 knockdown group showing markedly reduced levels of phosphorylated PI3K, AKT, and STAT3 compared with control tumors (Fig. 6L). Collectively, these results indicate that METTL3 regulates ADAM10 expression by enhancing its mRNA stability through m⁶A modification, thereby promoting the activation of the PI3K/AKT/STAT3 signaling pathway.

ADAM10 overexpression rescues the effects of METTL3 knockdown on proliferation and apoptosis

To determine if ADAM10 is a key functional mediator of METTL3’s pro-tumorigenic effects, a rescue experiment was performed under hypoxic conditions. PANC-1 and MIA PaCa-2 cells were co-transfected with lv-sh-METTL3 and a lentivirus overexpressing ADAM10 (lv-ADAM10). The efficiency of ADAM10 overexpression was first confirmed by immunoblotting (Fig. 7A).

Fig. 7.

Overexpression of ADAM10 partially rescues the effects of METTL3 knockdown on pancreatic cancer cell proliferation and apoptosis. PANC-1 and MIA PaCa-2 cells were co-transfected with the indicated lentiviruses and cultured under hypoxia. (A) Representative immunoblot confirming the overexpression of ADAM10. (B) Cell viability of the different groups was measured by a CCK-8 assay. (C) Long-term proliferation was assessed by a colony formation assay, with representative images and quantification shown. (D) Cell apoptosis was measured by flow cytometry after gemcitabine treatment. The groups are: lv-NC + lv-sh-NC (Control), lv-NC + lv-sh-METTL3 (METTL3 knockdown), lv-ADAM10 + lv-sh-NC (ADAM10 overexpression), and lv-ADAM10 + lv-sh-METTL3 (rescue). N = 3 (biological replicates) for each group. Data are analyzed using unpaired Student’s t-test for (A) and one-way ANOVA followed by Tukey’s post hoc test for (B-D). ** P < 0.01 compared to Hypoxia + lv-NC or Hypoxia + lv-NC + lv-sh-NC groups; ## P < 0.01 compared to Hypoxia + lv-ADAM10 + lv-sh-METTL3 group

As previously established, METTL3 knockdown suppressed cell proliferation under hypoxic conditions. The CCK-8 assay and a colony formation assay showed that METTL3 knockdown significantly reduced cell viability and colony formation, whereas simultaneous ADAM10 overexpression partially but significantly rescued this phenotype, restoring proliferative capacity (Fig. 7B-C). Similarly, the effect on apoptosis was assessed. In cells treated with gemcitabine, METTL3 knockdown led to a significant increase in the apoptotic rate. However, co-transfection with lv-ADAM10 protected the cells against gemcitabine, significantly reducing apoptosis to control levels (Fig. 7D). These results strongly suggest that METTL3 promotes the survival and proliferation of pancreatic cancer cells under hypoxia, at least in part, by regulating its downstream target, ADAM10.

ADAM10 overexpression reverses the inhibition of migration caused by METTL3 knockdown

To further confirm that ADAM10 is the key downstream effector of METTL3 in regulating cell motility, cell migration was investigated under hypoxic conditions. As shown by both Transwell assays (Fig. 8A) and wound healing assays (Fig. 8B), the knockdown of METTL3 significantly impaired the migratory capabilities of PANC-1 and MIA PaCa-2 cells. However, when ADAM10 was simultaneously overexpressed in these METTL3-knockdown cells, their migratory potential was significantly restored (Fig. 8A-B). These results demonstrate that the promotion of pancreatic cancer cell migration by METTL3 is mediated through its downstream regulation of ADAM10.

Fig. 8.

ADAM10 overexpression rescues the inhibition of cell migration caused by METTL3 knockdown. PANC-1 and MIA PaCa-2 cells were co-transfected with the indicated lentiviruses and cultured under hypoxia to assess changes in cell migration. The groups are: lv-NC + lv-sh-NC, lv-NC + lv-sh-METTL3, and lv-ADAM10 + lv-sh-METTL3. (A) Representative images and quantification of Transwell migration assays. (B) Representative images and quantification of wound healing assays. N = 3 (biological replicates) for each group. Data are analyzed using one-way ANOVA followed by Tukey’s post hoc test. ** P < 0.01 compared to Hypoxia + lv-NC + lv-sh-NC group; ## P < 0.01 compared to Hypoxia + lv-ADAM10 + lv-sh-METTL3 group

ADAM10 mediates the effects of METTL3 on hypoxia-related immune evasion

To further establish that METTL3 exerts its influence on immune evasion through ADAM10, the dynamic effects were conducted under hypoxic conditions. PANC-1 and MIA PaCa-2 cells were co-transfected to knock down METTL3 while simultaneously overexpressing ADAM10. As expected, METTL3 knockdown increased the percentage of MICA-positive cells, while ADAM10 overexpression alone had the opposite effect. Crucially, the concurrent overexpression of ADAM10 significantly eliminated the effect of METTL3 knockdown, reducing the percentage of MICA-positive cells (Fig. 9A). Consistent with this, METTL3 knockdown decreased the levels of soluble MICA (sMICA); however, this effect was reversed by the overexpression of ADAM10 (Fig. 9B).

Fig. 9.

ADAM10 mediates the effects of METTL3 on hypoxia-induced immune evasion. PANC-1 and MIA PaCa-2 cells were co-transfected as indicated and cultured under hypoxia. (A) Flow cytometry analysis of the percentage of MICA-positive cells on the cancer cell surface. (B) ELISA measurement of soluble MICA (sMICA) concentration in the culture supernatant. (C-D) Flow cytometry analysis of the percentage of NKG2D-positive NK92 cells after incubation with conditioned medium (CM) from the indicated cancer cell groups. (E) The killing rate of NK92 cells against the pancreatic cancer cells was determined using a CCK-8 assay. N = 3 (biological replicates) for each group. Data are analyzed using one-way ANOVA followed by Tukey’s post hoc test. ** P < 0.01 compared to Hypoxia + lv-NC + lv-sh-NC group; ## P < 0.01 compared to Hypoxia + lv-ADAM10 + lv-sh-METTL3 group

These changes directly affected interactions with NK cells. Conditioned medium from METTL3-knockdown cells increased the percentage of NKG2D-positive NK92 cells and enhanced the NK cell-mediated killing of cancer cells. In contrast, conditioned medium from ADAM10-overexpressing cells exerted opposite effects. More importantly, the effects of METTL3 knockdown were abolished; conditioned medium from cells overexpressing ADAM10, even with METTL3 knocked down, failed to increase NKG2D expression and resulted in a lower cancer cell-killing rate (Fig. 9C-E). Taken together, these results demonstrate that ADAM10 is a critical downstream mediator of METTL3’s role in promoting immune evasion in pancreatic cancer cells under hypoxia.

Discussion

Herein, microarray datasets from the GEO database were evaluated to identify DEGs and m⁶A modification regulators within pancreatic carcinoma. A significant upregulation of METTL3 was observed in pancreatic cancer samples compared to non-cancerous controls. Notably, METTL3 exhibited a positive correlation with HIF1A, a key factor within the hypoxic tumor microenvironment. This study demonstrates that METTL3 promotes cell viability, migration, and survival while inhibiting apoptosis. Crucially, these findings were validated in vivo, demonstrating that METTL3 knockdown suppresses tumor growth and enhances NK cell infiltration. Mechanistically, METTL3 was found to increase ADAM10 mRNA stability in a m⁶A modification-related manner, which in turn activates the PI3K/AKT/STAT3 signaling pathway. Rescue experiments confirmed that ADAM10 is a key mediator of METTL3’s effects on proliferation, apoptosis, migration, and immune evasion.

Hypoxia is a common feature in the tumor microenvironment and is known to increase cell survival, facilitate immune evasion and aggressive tumor behavior, primarily through hypoxia-inducible factors such as HIF1A [19–21]. Herein, hypoxic stimulation notably promoted pancreatic cancer cells viability and inhibited cell apoptosis. Moreover, the revealed positive correlation between METTL3 and HIF1A underscores METTL3’s critical role in tumor progression, especially under hypoxic conditions often prevalent in solid tumors such as pancreatic cancer [22–24]. This correlation indicates that METTL3 could play a key mediatory role in the hypoxia-induced pathways that promote tumor growth and survival. The differential expression of METTL3 in high-HIF1A versus low-HIF1A groups further accentuates its potential role in modulating the tumor’s response to hypoxia. In the meantime, the expression of MICA, a stress-induced ligand for NK cells, was also found to be upregulated within the high-HIF1A group. Decreased cell-surface MICA expression, as indicated by increased soluble MICA (sMICA) levels, is a well-documented mechanism of immune evasion that impairs NK cells’ ability to identify and destroy cancer cells [25, 26]. Further corroborating these findings, METTL3 expression and m⁶A modification levels were elevated in pancreatic cancer tissues, particularly under hypoxic conditions. This elevation in m⁶A modification, known to affect various aspects of RNA metabolism, aligns with recent studies highlighting its role in tumor development and metastasis [27] and suggests that METTL3 may regulate hypoxia-related immune evasion in pancreatic cancer.

METTL3 is an m⁶A writer that could influence the stability or translation of target RNAs, thereby affecting gene expression profiles conducive to tumor survival and immune evasion. As previously reported, METTL3 can exert complex effects on modulating immune escape in different cancers by affecting multiple mRNAs, including PD-L1 mRNA in bladder cancer [28] and NLRC5 mRNA in endometrial cancer [29]. In this study, METTL3 had a significant impact on cell viability, migration, and immune escape under both normoxia and hypoxia. Under normoxic conditions, METTL3 overexpression in pancreatic tumor cells enhanced viability and migration, reduced apoptosis, and reduced MICA-positive cells and elevated sMICA levels. These changes suggest an enhanced capacity for immune evasion, likely through reduced NK cell-mediated cytotoxicity [25, 26]. Under hypoxic conditions, METTL3 knockdown reversed these phenotypes, manifested as inhibited cell viability and migration, promoted cell apoptosis, and enhanced NK92 cell cytotoxicity. These in vitro functions were confirmed in the xenograft model, where METTL3 knockdown not only inhibited tumor growth but also significantly enhanced the infiltration of NK cells, creating a more immune-active microenvironment. These findings are consistent with previous studies demonstrating the role of hypoxia in modulating immune surveillance mechanisms in the tumor microenvironment [30, 31]. Therefore, by modulating immune recognition, METTL3 emerges as a potential therapeutic target to restore immune surveillance and disrupt tumor progression in pancreatic cancer.

Regarding the underlying mechanism, ADAM10’s overexpression in pancreatic cancer cells suggests its substantial impact on immune evasion. ADAM10, a disintegrin and metalloproteinase domain-containing protein, is implicated in the shedding of cell-surface proteins, including MICA, thereby reducing NK cell-mediated tumor cell lysis [7, 32]. In our study, ADAM10 was positively regulated by METTL3 in pancreatic cancer cells under normoxia or hypoxia, suggesting that ADAM10 may serve as a mediator of METTL3’s regulatory functions in hypoxia-related immune escape within pancreatic tumor cells. Previously, Xia et al. [33] reported that METTL3 knockdown in MIA PaCa-2 and BxPC-3 cell lines reduced RNA m⁶A modifications and reduced the capacity of cancer cells to proliferate, invade, and migrate. Our findings confirm this and further demonstrate that METTL3 influences ADAM10 mRNA stability through its m⁶A methyltransferase activity. Bioinformatic prediction using SRAMP identified a high-confidence m⁶A site on ADAM10 mRNA, strengthening this mechanistic link. Critically, ADAM10 overexpression partially abolished the effects of METTL3 knockdown on immune evasion but also on proliferation, apoptosis, and migration, suggesting its role in mediating the effects of METTL3.

In this study, our findings suggest that METTL3, through its m⁶A methyltransferase activity, influences the stability of ADAM10 mRNA. M⁶A modifications are known to impact RNA stability and translation, thereby regulating gene expression [34]. In the context of cancer, m⁶A modifications could alter the expression levels of genes involved in crucial pathways such as proliferation, apoptosis, and immune evasion [35, 36]. The modulation of ADAM10 mRNA stability by METTL3 suggests that METTL3 m⁶A modification of ADAM10 might be critical for understanding the adaptability of cancer cells in the hypoxic tumor microenvironment. As expected, under hypoxia, ADAM10 overexpression exerted opposite effects to those of METTL3 knockdown upon pancreatic cancer cell lines and NK92 cells; moreover, ADAM10 overexpression partially abolished the effects of METTL3 knockdown on pancreatic cancer cells and NK92 cells, suggesting that ADAM10 might mediate the effects of METTL3 on hypoxia-related pancreatic cancer immune evasion in a m⁶A modification-related manner.

Furthermore, a key signaling pathway downstream of the METTL3/ADAM10 axis was identified. Our results demonstrate that METTL3-mediated upregulation of ADAM10 leads to the phosphorylation and activation of the PI3K/AKT/STAT3 pathway, a critical cascade in cancer cell survival and proliferation [37]. This activation was observed in vitro and was validated in vivo, where METTL3 knockdown suppressed the phosphorylation of these signaling proteins in tumor tissues, providing a more complete molecular narrative for how METTL3 exerts its pro-tumorigenic functions.

In conclusion, the interplay between METTL3 and ADAM10, mediated through m⁶A modification, underscores the complexity of regulatory mechanisms that govern cancer progression and immune evasion. By demonstrating that METTL3-dependent stabilization of ADAM10 enhances MICA shedding and suppresses NK-cell recognition, our study reveals a mechanistic axis that contributes to the immunosuppressive tumor microenvironment in hypoxic pancreatic cancer. These insights provide a direct rationale for therapeutic strategies aimed at reprogramming the immunosuppressive microenvironment by targeting the m⁶A regulatory machinery or its downstream effectors, such as ADAM10, to restore anti-tumor immune surveillance. Moreover, this mechanism highlights the potential of microenvironment-specific combination therapies, where modulation of m⁶A modification could synergize with existing immunotherapies to overcome resistance and improve clinical outcomes in pancreatic cancer. Given the emerging evidence that m⁶A regulators influence immune cell infiltration, immune checkpoint expression, and responsiveness to immunotherapy across cancers, these findings support the translational potential of m⁶A-targeted interventions to reshape the tumor immune landscape.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Linwei Wang: Conceptualization, Data Curation, Writing - Original Draft; Fan Zhang: Methodology, Investigation, Writing - Original Draft; Qizhen Chen: Software, Visualization; Shuai Zhu: Formal analysis, Resources; Xuejun Gong: Validation, Writing - Review & Editing; Yebin Lu: Project administration, Supervision, Writing - Review & Editing.

Funding

None.

Data availability

Please contact the authors for data requests.

Declarations

Ethics approval and consent to participate

All procedures performed in studies involving human participants were in accordance with the ethical standards of Xiangya Hospital, Central South University (approval no.: 2022020460), and with the 1964 Helsinki declaration. Informed consent to participate in the study has been obtained from participants. The guidelines for the care and use of animals were approved by the Medicine Animal Welfare Committee of Xiangya Hospital, Central South University (approval no.: 2022020460).

Consent for publication

Consent for publication was obtained from the participants.

Competing interests

The authors confirm that there are no conflicts of interest.