PUM2 inhibits ferroptosis and enhances oxaliplatin resistance in COAD via the NEDD4L/NRF2 axis Authors: Jiacheng Xie, Kai Dong, Jialin Wu, Xinsheng Cheng, Shikai Wang, Hongzhen Zhang Published: 2026 ABSTRACT Highlights•PUM2 drives oxaliplatin resistance in COAD by directly binding to and suppressing NEDD4L mRNA.•PUM2 inhibits NEDD4L, leading to Nrf2 accumulation and ferroptosis blockade.•Combining oxaliplatin with Nrf2 inhibitor enhances antitumor efficacy against resistant tumors in mouse model.Keywords: COAD, Oxaliplatin resistance, Ferroptosis, RNA binding protein Introduction Colon adenocarcinoma (COAD) remains one of the most prevalent and deadly malignancies worldwide, accounting for approximately 0.88 million deaths and over 1.85 million new cases annually [1]. The progression of COAD is often characterized by clinically silent invasion and metastasis, which pose major challenges for diagnosis and treatment, and present a serious threat to public health [2,3]. The treatment for COAD includes surgery, radiotherapy, chemotherapy and immunotherapy [[4], [5], [6]]. Currently, combination chemotherapy regimens based on oxaliplatin (L-OHP) remain the first-line treatment option for patients with advanced colorectal cancer, effectively inhibiting tumor cell proliferation, delaying disease progression, and prolonging patient survival [7,8]. However, the widespread occurrence of oxaliplatin resistance in clinical practice often leads to diminished treatment efficacy, tumor recurrence, and ultimately treatment failure, posing a significant challenge to improving the prognosis of COAD patients [9,10]. Therefore, an in-depth exploration of the underlying regulatory mechanisms is essential to identify effective intervention strategies. Ferroptosis is a recently identified form of regulated cell death which depends on iron ions [11,12]. Ferroptosis-inducing factors can directly or indirectly affect glutathione peroxidase through various pathways, leading to reduced intracellular antioxidant capacity and the accumulation of lipid reactive oxygen species (ROS), ultimately contributing to cell death [13,14]. Recent studies have indicated a close association between ferroptosis and cancer progression [15]. On the one hand, tumor cells may develop resistance to ferroptosis by activating specific protective pathways, thereby promoting their progression and metastasis [16,17]. On the other hand, ferroptosis can suppress tumor growth and enhance sensitivity to chemotherapy [18,19]. Zeng and colleagues demonstrated that inhibition of ferroptosis enhanced oxaliplatin resistance in colorectal cancer [20]. However, the regulatory mechanisms of ferroptosis that mediate oxaliplatin resistance in COAD require further investigation. PUM2 belongs to the pumilio family, an evolutionarily conserved family of RNA-binding protein [21,22]. Its core function involves post-transcriptional regulation of target gene expression by specifically recognizing and binding to the 3′ untranslated region (UTR) of target mRNA, thereby participating in various critical biological processes including cell proliferation, differentiation, and apoptosis [23,24]. It has been reported that PUM2 exhibits abnormal expression in multiple malignant tumors, showing strong correlations with tumorigenesis, progression, invasion, metastasis, and patient prognosis, including ovarian, breast, and liver cancers [[25], [26], [27]]. This suggests that PUM2 may be a key regulator of the malignant phenotype in tumors. Additionally, PUM2 is also associated with chemotherapy resistance in pancreatic cancer [28]. However, the role of PUM2 in COAD resistance and its underlying mechanisms remain unclear. In this study, we found that PUM2 was highly expressed in oxaliplatin-resistant cells and resistant xenografts tumor tissues. Knocking down PUM2 reverses oxaliplatin resistance by reducing resistant cell viability, promoting ROS accumulation and inducing ferroptosis. Mechanistically, PUM2 binds NEDD4L mRNA to inhibit its expression, which attenuates NEDD4L-mediated ubiquitination and degradation of NRF2, leading to NRF2 accumulation and subsequent suppression of ferroptosis, thereby enhancing oxaliplatin resistance. Materials and methods Bioinformatics analysis The mRNA levels of PUM2 were analyzed using the UALCAN database (http://ualcan.path.uab.edu). Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia.cancer-pku.cn) was used to analyze the pathological stage of patients with COAD. Kaplan-Meier Plotter was used to analyze the overall survival of patients with COAD (https://kmplot.com/analysis/).Gene Expression Omnibus Series (GSE) (GSE35279, GSE212156) were obtained from the Gene Expression Omnibus database (GEO; http://www.ncbi.nlm.nih.gov/geo). Cell culture The COAD cell lines SW480, HCT116, LOVO and human normal colonic epithelial cells (HIEC) were purchased from the American Type Culture Collection (ATCC). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Hyclone), and 1% penicillin/streptomycin (Gibco) in an atmosphere with 5% CO2 at 37°C. Two oxaliplatin‑resistant colorectal cancer cell lines (SW480‑OR and HCT116‑OR) were established through chronic exposure of parental SW480 and HCT116 cells to gradually increasing oxaliplatin concentrations over 3 months. During induction, the drug concentration was escalated 1.5–2‑fold every 2 weeks, starting at 1 μM for SW480 and 0.5 μM for HCT116. Dose escalation was delayed when cell mortality exceeded 50% until stable proliferation was restored. Resistance was confirmed when cells maintained > 90% viability and normal proliferation kinetics in complete medium containing 16 μM (SW480‑OR) or 12 μM (HCT116‑OR) oxaliplatin. Quantitative real-time PCR (RT-qPCR) Cellular RNA was extracted using an RNA extraction kit (Yeason). The isolated total RNA was then subjected to reverse transcription with a reverse transcription kit (Vazyme). For quantitative PCR analysis, Hieff qPCR SYBR Green Master Mix (Yeason) was utilized. Gene expression levels were normalized to the internal control GAPDH and calculated using the 2−ΔΔCt method. All experiments were performed in triplicate. Primer sequences used in this study are listed in the Supplementary Table 1. Western blot Total protein was extracted via RIPA lysis buffer (Beyotime, China), and a BCA kit (Beyotime, China) was used to determine the protein concentration. SDS–PAGE was used to separate proteins, and they were transferred onto PVDF membranes (Millipore, USA), which were then blocked with 5% skim milk powder, incubated with primary and secondary antibodies sequentially, and finally the signal was captured using the Bio-Rad Image Lab system. The antibodies used in this study were as follows: PUM2 (1:2000, Abcam, #ab92390, Clone No.: EPR3813), ACSL4 (1:3000, Proteintech, #66617-1-Ig, Clone No.: 1H5D3), GPX4 (1:1000, Proteintech, # 67763-1-Ig, Clone No.: 3F5G5), SLC7A11 (1:2000, Proteintech, #82115-3-RR, Clone No.: 252789G7), NEDD4L (1:1000, Proteintech, # 67276-1-Ig, Clone No.: 2F6H6), NRF2 (1:1000, Cell Signaling Technology, #12721, Clone No.: D1Z9C), GAPDH (1:5000, Proteintech, #60004-1-Ig, Clone No.: 1E6D9). Immunofluorescence After fixation with 4% paraformaldehyde and permeabilization with 0.1% Triton X-100, the cells were blocked with 5% BSA, incubated with primary antibody overnight at 4°C, followed by a fluorescent secondary antibody, and finally the cell nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) before imaging under a fluorescence microscope. Cell counting kit-8 (CCK-8) assay Cell proliferation ability was detected using a CCK-8 Kit (Beyotime, China). Cells were seeded in 96-well plates and after the treatment period, 10 μL of CCK-8 reagent was added to each well, followed by incubation for 2 hours at 37°C. The absorbance was then measured at 450 nm using a microplate reader to determine cell viability. Colony formation assay Cells were seeded at a density of 1000 cells per well in 6-well plates and cultured for 7-14 days to allow colony formation. The resulting colonies were then fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. Colony quantification was conducted using ImageJ software. Flow cytometry detection of ROS The intracellular ROS levels were measured using a DCFH-DA probe, for which the harvested cells were incubated with 10 μM DCFH-DA (Beyotime, China) at 37°C for 30 minutes, followed by analysis with a flow cytometer, and the fluorescence intensity was detected in the FITC channel. RNA immunoprecipitation (RIP) The RNA immunoprecipitation (RIP) assay was performed using the Magna RIP Kit (Millipore, USA) according to the manufacturer's instructions. Briefly, cell lysates were incubated with protein A/G magnetic beads conjugated to anti-PUM2 antibody (Abcam, USA) or normal rabbit IgG beads overnight at 4°C, followed by RNA purification from the immunoprecipitated complexes, and the co-precipitated RNA was subsequently analyzed by RT-qPCR. Co-immunoprecipitation (Co-IP) Cells were lysed, and the supernatants were incubated with the target primary antibody overnight at 4°C, followed by the addition of Protein A/G beads to capture the immune complexes. After extensive washing, the bound proteins were eluted and analyzed by Western blot. Animal experiment All animal experiments were approved by the Animal Ethics Committee of Guangzhou Seyotin Biotechnology Co., Ltd (approval number: SYT2025033). NOD-scid mice (6 weeks old) were procured from Ruige Biological Technology Co., Ltd. (Guangzhou, China) and maintained in a specific pathogen-free (SPF) barrier facility. To establish the subcutaneous xenograft model, 1 × 106 SW480-OR cells were injected into the right armpit of the mice. After tumor formation, the mice received oxaliplatin (5 mg/kg) every two days for treatment (intraperitoneal injection). Following the final treatment, the mice were euthanized, and tumors were excised, weighed, and processed for further analysis. Statistical analysis All statistical analyses were performed using GraphPad Prism 9.0, and data are presented as the mean ± SD from at least three independent experiments. The Shapiro–Wilk test was used to assess the normality of data distribution. Group comparisons were performed using an unpaired two-tailed Student's t-test for two groups, or one-way ANOVA followed by Tukey’s post-hoc test for multiple groups. A p-value < 0.05 was considered statistically significant. Results PUM2 is upregulated and associated with poor prognosis in COAD To explore the role of PUM2 in the progression and prognosis of COAD. we assessed the expression level of PUM2 in public datasets of COAD patients. Based on data obtained from the UALCAN database, PUM2 was significantly upregulated in COAD tissues compared to normal groups (Fig. 1A). GSE32323, GSE35279, GSE212156 were selected as COAD datasets from the NCBI-GEO database, and the expression of PUM2 was analyzed. The results showed that PUM2 was highly expressed in cancer samples (Fig. 1B). Moreover, GEPIA database analysis demonstrated that PUM2 expression levels varied significantly across different pathological stages (Fig. 1C). Analysis of the Kaplan-Meier Plotter database revealed that high PUM2 expression was associated with poor overall survival (Fig. 1D). Additionally, qPCR and immunofluorescence assays demonstrated that both mRNA and protein expression levels of PUM2 were significantly higher in COAD cell lines (SW480, LOVO, HCT116) compared with the normal colon epithelial cell line HIEC (Fig. 1E-F). These findings collectively indicate that PUM2 may play a role in COAD progression. PUM2 is upregulated in colorectal cancer and associated with poor prognosis. (A) The expression of PUM2 in COAD was analyzed in the UALCAN database. (B) The expression of PUM2 in colorectal cancer was analyzed in GEO databases (GSE35279 and GSE212156). (C) The correlation between PUM2 expression and cancer pathological stage was analyzed in GEPIA databases. (D) Kaplan-Meier analysis of the association between PUM2 expression and overall survival. (E) The mRNA expression of PUM2 in different cell lines was detected using qPCR (n = 3). (F) Immunofluorescence staining of PUM2 in different cell lines. Representative images (left) and quantitative analysis of fluorescence intensity (right), scale bar = 100 μm, (n = 3). For (B) significance was determined by Student’s t-test. For (E) and (F), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. PUM2 plays a crucial role in the sensitivity of oxaliplatin in COAD Previous studies have revealed that PUM2 is closely related to cancer proliferation, however, its relationship with oxaliplatin (L-OHP) sensitivity remains unknown. Therefore, we further explored the association between PUM2 and L-OHP sensitivity in COAD. We first treated COAD cell lines with increasing concentrations of oxaliplatin and found that PUM2 protein expression increased in a dose-dependent manner (Supplemental Fig. 1). We established oxaliplatin-resistant COAD cell models (Fig. 2A), with successful resistance induction confirmed by CCK-8 assays that exhibited a significantly higher half-maximal inhibitory concentration (IC50) in the resistant versus parental cells (Fig. 2B). Meanwhile, PUM2 expression was markedly upregulated at both the mRNA and protein levels in L-OHP-resistant cell lines relative to their parental counterparts (Fig. 2C–D). To evaluate the functional consequence of this association in vivo, we established mouse xenograft models using parental and resistant cells (Fig. 2E). Tumors derived from oxaliplatin-resistant cells demonstrated significantly increased growth, as evidenced by larger tumor volume and greater weight compared to those from parental cells (Fig. 2F). Consistent with our in vitro findings, immunohistochemical (IHC) analysis confirmed markedly higher PUM2 expression in the resistant tumor tissues (Fig. 2G). Collectively, this in vivo evidence reinforces the conclusion that PUM2 is associated with oxaliplatin resistance in COAD. PUM2 plays a critical role in regulating oxaliplatin sensitivity in colorectal cancer. (A) Schematic diagram of the establishment of oxaliplatin-resistant cell models SW480-OR and HCT116-OR (n = 3). (B) The half-maximal inhibitory concentration (IC50) of L-OHP was determined in sensitive and resistant cell pairs using the CCK-8 assay (n = 3). (C) RT-qPCR analysis of PUM2 mRNA expression in parental and resistant cell lines (n = 3). (D) Western blot analysis of PUM2 protein expression in parental and resistant cell lines. Representative blots are shown (left). The bar graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). (E) Schematic of the subcutaneous tumor model established in mice using sensitive and resistant cells (n = 3). (F-H) Representative images of tumors (F), quantitative analysis of tumor volume (G), and tumor weight (H) from the xenograft models (n = 5/group). (I) Immunohistochemical (IHC) staining of Ki67 and PUM2 in tumor tissues from the xenograft models. Representative images are shown (left). The graph (right) represents the quantitative analysis of the relative positive area ratio, scale bar = 100 μm (n = 3). For (C), (D), (G) and (H), significance was determined by Student’s t-test. For (I), significance was determined by two-way ANOVA. Data are presented as mean ± SD of three independent experiments, *p < 0.05, **p < 0.01, ***p < 0.001. PUM2 knockdown overcomes L-OHP resistance in COAD cells by promoting ferroptosis To further investigate the impact of PUM2 on oxaliplatin sensitivity in resistant Ncells, we knocked down PUM2 in resistant cell lines (Fig. 3A-B and Supplementary Fig. 2A-B). It was found that PUM2 knockdown significantly reduced cell viability, colony formation capacity, and migration ability in resistant cells, compared to the control group (Fig. 3C-E and Supplementary Fig. 2C-E). This enhanced their sensitivity to L-OHP. Previous studies have reported that oxaliplatin resistance is closely associated with suppressed ferroptosis in tumors. Therefore, this study further investigated whether PUM2 participates in the L-OHP resistance process of COAD-resistant cells by regulating ferroptosis. The results demonstrated that PUM2 knockdown resulted in a significant increase in ROS levels (Fig. 3F and Supplementary Fig. 2F), a substantial decrease in SOD and GSH concentrations, and a considerable elevation in MDA levels (Fig. 3G and Supplementary Fig. 2G) in COAD-resistant cell lines. Furthermore, ACSL4 expression, a marker of ferroptosis, was found to be increased, while the expression of ferroptosis-suppressing markers GPX4 and SLC7A11 were decreased (Fig. 3H and Supplementary Fig. 2H). Treatment with the ferroptosis inhibitor Fer-1 and the apoptosis inhibitor Z-VAD-FMK in resistant cells with PUM2 knockdown showed that only ferroptosis inhibition significantly restored cell viability and migration (Supplementary Fig. 3). These suggest that PUM2 knockdown reverses L-OHP resistance in COAD-resistant cells by promoting ferroptosis. Downregulation of PUM2 promotes ferroptosis and increases SW480-resistant cells sensitivity to oxaliplatin. (A) PUM2 mRNA expression levels were assessed by RT-qPCR (n = 3). (B) PUM2 protein expression levels were assessed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). (C) Cell viability was measured by the CCK-8 assay in resistant cells under different treatments (n = 3). (D) Representative images (left panels) and corresponding quantitative analysis (right panels) of colony formation and Transwell migration assays (n = 3). (E) The concentrations of superoxide dismutase (SOD), glutathione (GSH) and malondialdehyde (MDA) were measured by ELISA (n = 3). (F) Intracellular reactive oxygen species (ROS) levels were detected by flow cytometry (n=3). (G) Quantitative analysis of ROS levels (n = 3). (H) The protein expression of ferroptosis markers (ACSL4, GPX4, and SLC7A11) was analyzed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization(n = 3). For (A), (B), (C), (D), (E), (G) and (H), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments, n = 3, *p < 0.05, **p < 0.01, ***p < 0.001. PUM2 knockdown reduces L-OHP resistance by promoting ferroptosis in vivo To validate the effects of PUM2 inhibition on ferroptosis and L-OHP resistance in COAD-resistant cells in vivo, subcutaneous tumor models were established in mice using different treated resistant cell lines (Fig. 4A). The results demonstrated that the PUM2-knockdown combined with L-OHP treatment resulted in reduced tumor volumes and diminished tumor weights, compared to the control group (Fig. 4B). Furthermore, IHC analysis revealed reduced PUM2 expression and significantly decreased Ki67 expression, a marker of cell proliferation, in tumors from PUM2-knockdown mice (Fig. 4C). Concurrently, the levels of ROS increased, while the concentrations of SOD and GSH decreased, and the levels of MDA increased in the tumor tissues following PUM2 knockdown (Fig. 4D-E). Western blot analysis revealed that PUM2 knockdown led to an increase in ACSL4 expression and decreases in GPX4 and SLC7A11 expression in tumor tissues (Fig. 4F). These findings, which are consistent with in vitro results, confirm that in vivo suppression of PUM2 promotes ferroptosis in resistant cells and reduces their resistance to L-OHP. PUM2 knockdown reduces L-OHP resistance by promoting ferroptosis in vivo. (A) Representative images of tumors in each group. (B) Quantitative analysis of tumor volume and tumor weight (n = 5/group). (C) Representative images and quantitative analysis of IHC staining for Ki67 in tumor tissues from the xenograft models, scale bar = 100 μm (n = 3). (D) The concentrations of SOD, GSH and MDA were measured by ELISA (n = 3). (E) The protein expression of ferroptosis markers (ACSL4, GPX4, and SLC7A11) in tumor tissue was analyzed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). For (B), (C), (D) and (E), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments, *p < 0.05, **p < 0.01, ***p < 0.001. PUM2 inhibit ferroptosis by regulating NEDD4L-mediated NRF2 ubiquitination To further elucidate the mechanisms by which PUM2 regulates ferroptosis and L-OHP resistance in COAD, we first identified differentially expressed genes (DEGs) from the GEO database. These were then intersected with the ferroptosis-inhibiting gene set from the Ferroptosis Database (Ferr Db), identifying four candidate genes: NR5A2, NEDD4L, MT1G, and MUC1 (Fig. 5A). RT-qPCR analysis confirmed that NEDD4L exhibited the most significant negative correlation with PUM2 expression, suggesting NEDD4L to be a highly probable key potential target of PUM2 in regulating ferroptosis and L-OHP resistance in COAD (Fig. 5B). Consequently, we conducted an in-depth investigation of NEDD4L. Given that the core function of PUM2 is as an RNA-binding protein that regulates target gene by binding to its mRNA, we performed RNA immunoprecipitation (RIP) to validate its interaction with NEDD4L, and the results showed that the PUM2 group significantly enriched NEDD4L mRNA(Fig. 5C). Subsequently, we predicted the binding sites of PUM2 in the 3′-UTR of NEDD4L mRNA using the RBPmap tool, selected the top-scoring site (P < 0.05), and confirmed that PUM2 specifically binds to the 3′-UTR of NEDD4L mRNA through dual-luciferase reporter assay (Supplementary Fig. 4A-B). PUM2 inhibit ferroptosis by regulating NEDD4L-mediated Nrf2 ubiquitination. (A) Venn diagram of the candidate genes obtained from public databases: GEO (GSE35279 and GSE32323) and Ferr Db. (B) Heatmap of qPCR analysis showing the mRNA expression patterns of overlapping genes. (C) The interaction between PUM2 and NEDD4L was validated in SW480-OR cells by RNA immunoprecipitation (n = 3). (D) Nrf2 expression was evaluated using RT-qPCR (n=3). (E) Western blot analysis of Nrf2 protein expression. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). (F) Immunoprecipitation confirming the interaction between NEDD4L and Nrf2 (n = 3). (G) Effects of NEDD4L on Nrf2 ubiquitination (n = 3). (H) The concentrations of SOD, GSH and MDA were measured by ELISA (n = 3). (I) Representative immunofluorescence images and quantitative analysis of ROS (n = 3). (J) The protein expression of ferroptosis markers (ACSL4, GPX4, and SLC7A11) was analyzed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). For (C), significance was determined by Student’s t-test. For (B), (D), (E), (H), (I) and (J), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments, *p < 0.05, **p < 0.01, ***p < 0.001. NRF2 is a pivotal transcription factor in the ferroptosis regulatory pathway, exerting its function by modulating the expression of ferroptosis-inhibitory genes, such as GPX4 and SLC7A11 [29]. Previous studies have demonstrated that dysregulation of NRF2 signaling is closely associated with the development of oxaliplatin resistance in cancer [30]. RT-qPCR and Western blot analysis showed that mRNA and protein expression levels of NRF2 were significantly reduced in the NEDD4L overexpression group, compared with the control group (Fig. 5D-E). The results of the CO-IP confirmed the interaction between NEDD4L and NRF2 (Fig. 5F). In consideration of the established function of NEDD4L as an E3 ubiquitin ligase, subsequent ubiquitination experiments provided further evidence that NEDD4L significantly promotes NRF2 ubiquitination (Fig. 5G and Supplementary Fig. 4C). This observation suggests that NEDD4L regulates NRF2 expression through the ubiquitination pathway. To determine whether NEDD4L regulates ferroptosis through the NRF2 signaling pathway. We found that concurrent overexpression of NRF2 in NEDD4L-overexpressing resistant cells markedly reversed the ferroptosis effect of NEDD4L. This reversal was evidenced by decreased intracellular ROS, elevated SOD and GSH concentrations, and reduced MDA levels (Fig. 5H-I). Additionally, the ferroptosis marker ACSL4 was downregulated, while GPX4 and SLC7A11 were upregulated (Fig. 5J). NRF2 is essential in PUM2-mediated ferroptosis Based on the finding that PUM2 regulates NRF2 via NEDD4L, we next sought to establish the functional necessity of NRF2 in the PUM2-mediated ferroptosis pathway. Consistent with its proposed role as a downstream effector, the protein level of NRF2 was significantly elevated in L-OHP-resistant cells compared to their parental counterparts, while its upstream regulator NEDD4L showed an opposite expression pattern (Fig. 6A). Furthermore, knockdown of PUM2 led to a marked increase in NEDD4L and a concomitant decrease in NRF2 protein expression (Fig. 6B), reinforcing the PUM2/NEDD4L/NRF2 regulatory axis. In order to definitively determine whether NRF2 is the critical functional target, rescue experiments were performed by overexpressing NRF2 in the PUM2-knockdown background. It is noteworthy that the ectopic expression of NRF2 (sh-PUM2+OE-NRF2) effectively reversed the inhibitory effect on migration caused by PUM2 silencing (Fig. 6C). Concurrently, the increase in intracellular reactive oxygen species (ROS) levels observed upon PUM2 knockdown was also rescued by NRF2 re-expression (Fig. 6D). The ferroptosis molecular profile induced by PUM2 knockdown, characterized by upregulation of ACSL4 and downregulation of GPX4 and SLC7A11, was reversed when NRF2 was co-overexpressed (Fig. 6E). The results of this study demonstrate that NRF2 is an indispensable downstream executor, and that its stabilization is essential for PUM2 to suppress ferroptosis and promote oxaliplatin resistance. Nrf2 is essential in PUM2-mediated ferroptosis. (A) Western blot analysis of NEDD4L and Nrf2 protein expression in parental and resistant cells. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). (B) Western blot (left) and quantitative analysis (right) of NEDD4L and Nrf2 protein expression in sh-NC and sh-PUM2 cells (n = 3). (C) Representative images (left) and corresponding quantitative analysis (right) of transwell migration assays (n = 3). (D) Representative immunofluorescence images and quantitative analysis of ROS. Scale bar=100 μm (n = 3). (E) The protein expression of ferroptosis markers (ACSL4, GPX4, and SLC7A11) was analyzed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). For (A) and (B), significance was determined by two-way ANOVA. For (C), (D) and (E), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments, *p < 0.05, **p < 0.01, ***p < 0.001. NRF2 inhibitor enhances the sensitivity of COAD to L-OHP To investigate the role of the NEDD4L/NRF2 axis in PUM2-mediated L-OHP resistance, a combined regimen of the NRF2 inhibitor ML385 with oxaliplatin was used to determine whether inhibition of this pathway could potentiate ferroptosis and suppress tumor growth in mouse. In comparison with mice receiving monotherapy, those administered combination therapy exhibited diminished tumor fluorescence signals (Fig. 7A), reduced tumor volumes, and decreased tumor weights (Fig. 7B-C). IHC results revealed significantly reduced expression of the cell proliferation marker Ki67 in tumor tissues from the combination therapy compared to the monotherapy group (Fig. 7D-E). Furthermore, serum analysis revealed significantly greater reductions in SOD and GSH concentrations in the combination group relative to the monotherapy group, while MDA concentrations were markedly higher in the combination group (Fig. 7F). Western blot analysis similarly revealed that the combination therapy exhibited a greater upregulation of ACSL4 expression and a greater downregulation of GPX4 and SLC7A11 expression in tumor tissues compared to the monotherapy group (Fig. 7G). The findings of this study suggest that NRF2 inhibitors have the potential to enhance the sensitivity of COAD to oxaliplatin. Combination of ML385 (Nrf2 inhibitor) with oxaliplatin enhances the sensitivity of resistance cells. After tumor formation, the mice were randomly divided into four groups (n = 5/group). The vehicle group received an equal volume of solvent control. The L-OHP group was treated with oxaliplatin (5 mg/kg), the Nrf2 inhibitor group with ML385 (30 mg/kg), and the combination group with both oxaliplatin (5 mg/kg) and ML385 (30 mg/kg). All treatments were administered by intraperitoneal injection every two days for three weeks. (A) Representative images of living imaging of xenograft mouse model and the fluorescence intensity were quantified. (B) Representative images of tumors in each group after surgical resection (n = 3). (C) Quantitative analysis of tumor volume and tumor weight (n = 5/group). (D) Representative images of IHC staining for Ki67 in tumor tissues from the xenograft models, scale bar = 100 μm (n = 3). (E) Quantitative analysis of Ki67 (n = 3). (F)The concentrations of SOD, GSH and MDA were measured by ELISA (n = 3). (G) The protein expression of ferroptosis markers (ACSL4, GPX4, and SLC7A11) in tumor tissue was analyzed by Western blot. Representative blots are shown (left). The graph (right) represents the quantitative analysis performed by densitometry normalization (n = 3). For (A), (C), (E), (F) and (G), significance was determined by one-way ANOVA. Data are presented as mean ± SD of three independent experiments, *p < 0.05, **p < 0.01, ***p < 0.001. Discussion Colon adenocarcinoma continues to be a major cause of cancer-related mortality worldwide, with oxaliplatin-based chemotherapy being the primary treatment for advanced cases [31]. Nevertheless, roughly 50% of patients develop oxaliplatin resistance following first-line chemotherapy, a development that directly drives disease progression, shortens survival, and drastically increases the difficulty of subsequent treatment [32,33]. Recent studies have highlighted ferroptosis, an iron-dependent, lipid ROS-driven form of regulated cell death, as a critical mediator of chemotherapy sensitivity, and promoting ferroptosis can enhance drug sensitivity [34,35]. However, the post-transcriptional regulatory circuits that link ferroptosis suppression to oxaliplatin resistance remain incompletely defined, particularly for RNA-binding proteins (RBPs) that orchestrate mRNA stability and translation to fine-tune cell death pathways. To date, an increasing amount of evidence suggests that tumor resistance to oxaliplatin is closely associated with the regulatory mechanisms of ferroptosis. Research on ferroptosis-mediated oxaliplatin resistance currently focuses on three main aspects: Firstly, transcriptional regulation. For example, Shu and colleagues found in gastric cancer that JMJD3 drives demethylation of the ALOX5 promoter, leading to its overexpression, which helps maintain gastric cancer stem cell characteristics and oxaliplatin resistance [36]. Secondly, post-transcriptional regulation is mediated by non-coding RNAs. For instance, the long non-coding RNA BASP1-AS1 can inhibit ferroptosis by promoting the lactylation modification of PCBP2, thereby conferring oxaliplatin resistance to gastric cancer cells [37]. Thirdly, classical oncogenic signaling pathways, such as the MAPK and NF-κB pathways [38,39]. However, although RNA-binding proteins play a key regulatory role in mRNA stability and translation, their role in regulating ferroptosis and oxaliplatin resistance remains largely unexplored, suggesting that our understanding of the mechanisms underlying chemotherapy resistance is incomplete. In this study, we explored the relationship between PUM2, ferroptosis, and oxaliplatin resistance in COAD using public datasets, oxaliplatin-resistant cell lines and xenograft models. We found that PUM2 was significantly upregulated in COAD tissues, and its high expression correlated with advanced pathological stages, poor overall survival, and oxaliplatin resistance. Functionally, knocking down PUM2 in oxaliplatin-resistant cells not only suppressed cell viability and cell migration but also promoted ferroptosis, thereby restoring oxaliplatin sensitivity. This effect was corroborated in vivo. Mechanistically, we demonstrated that PUM2, directly binds to NEDD4L mRNA and represses its expression. As an important member of the E3 ubiquitin ligase family, NEDD4L has been associated with several types of cancer and has been shown to be involved in regulating the stability of proteins in some oncogenic signaling pathways [40,41]. Downregulation of NEDD4L, as regulated by PUM2, attenuates the ubiquitination of the transcription factor NRF2. Consequently, the accumulated NRF2 protein upregulates the expression of key ferroptosis suppressors, GPX4 and SLC7A11, ultimately dampening ferroptosis and enhancing oxaliplatin resistance in colon cancer. Notably, PUM2 is an important RNA-binding protein in colorectal cancer and exerts its pro-tumor effects far beyond the regulation of the NEDD4L/NRF2 axis reported in this study. Previous studies have demonstrated that PUM2 acts in synergy with PUM1 in colorectal cancer to induce the G1/S phase transition in tumor cells, thereby promoting the malignant proliferation of colorectal cancer cells [42]. Additionally, our results further demonstrated that combining an NRF2 inhibitor with oxaliplatin synergized to activate ferroptosis and suppress tumor growth in xenograft models, achieving a superior anti-tumor effect compared to oxaliplatin monotherapy. As a commonly used platinum-based chemotherapy drug in clinical practice, oxaliplatin is associated with numerous adverse reactions in patients with colorectal cancer. Among these, neuropathy, diarrhea, vomiting, and allergic reactions are the primary reasons for treatment discontinuation during oxaliplatin therapy [43,44]. Its toxic mechanisms are closely linked to drug-induced cellular oxidative stress imbalance and DNA damage [45]. NRF2, as a regulator of the body’s antioxidant stress pathway, is extensively involved in the defense against oxidative damage in normal organs such as the liver, kidneys, and nervous tissue, and plays a crucial role in maintaining the homeostasis of normal tissue cells [46]. Although current studies on the NRF2 inhibitor ML385 have not focused on its toxic effects on cells and tissues, when used in combination with oxaliplatin, it may indirectly affect the tolerance of normal tissues to oxaliplatin toxicity due to pathway inhibition, thereby increasing the risk of toxicity. Therefore, this study provides only preliminary proof of concept for this combination regimen. Further evaluation is needed in animal models that more closely mimic clinical conditions to assess acute and chronic toxicity, maximum tolerated dose, and histopathological damage. Combined with pharmacokinetic analysis, this will optimize the dosing regimen to reduce potential adverse effects while ensuring antitumor efficacy, thereby providing a more comprehensive safety basis for subsequent preclinical studies. Collectively, our findings establish a crucial role for PUM2 in promoting oxaliplatin resistance in COAD by inhibiting ferroptosis through the NEDD4L/NRF2 axis, providing a new insight into the molecular mechanism of COAD chemoresistance (Fig. 8). The Schematic diagram of mechanism. PUM2 upregulates NEDD4L, which in turn promotes NRF2 ubiquitination and degradation, leading to reduced ferroptosis. This work has several limitations. The mechanistic analysis was primarily conducted using two oxaliplatin-resistant colorectal cancer cell lines and a subcutaneous xenograft model. Although these models are widely used and have enabled the preliminary elucidation of oxaliplatin resistance mechanisms, they do not fully capture the complexity of the human tumor microenvironment, tumor heterogeneity, or clinical dosing regimens. Accordingly, this study remains at the proof-of-concept stage, and the findings cannot be directly extrapolated to clinical translation. Compared with cell lines and subcutaneous xenograft models, patient-derived organoids (PDOs) can largely retain the histomorphology, genetic mutation profiles, and drug response characteristics of the original tumor, while patient-derived xenograft (PDX) models can more comprehensively simulate the tumor microenvironment and clinical resistance phenotypes. Both are considered ideal models for bridging basic research and clinical translation. The absence of validation using PDOs/PDX models derived from patients with oxaliplatin-resistant colorectal cancer in this study somewhat undermines the clinical relevance and translational potential of the findings. Therefore, subsequent studies will incorporate clinical samples for validation by establishing and utilizing PDOs and PDX models of oxaliplatin-resistant colorectal cancer to further verify the role of PUM2 in resistant patients and the efficacy of targeted interventions. Concurrently, we will analyze the correlation between PUM2 and oxaliplatin resistance, progression-free survival, and overall survival using clinical treatment and prognostic data. This could help extend the current findings beyond cell line and xenograft models, thereby strengthening the preclinical rationale for subsequent investigation. Funding This work was supported by Shenzhen Nanshan District Health System Science and Technology Major Project (No. NSZD2023045, No. NSZD2023001, NO.NSZD2024008). Shenzhen Basic Research Project (No.JCYJ20230807115823048). Ethic statement All procedures were approved by the Animal Ethics Committee of Guangzhou Seyotin Biotechnology Co., Ltd (SYT2025033). CRediT authorship contribution statement Jiacheng Xie: Writing – original draft, Visualization, Data curation. Kai Dong: Formal analysis, Data curation. Jialin Wu: Formal analysis, Data curation. Xinsheng Cheng: Visualization, Methodology. Shikai Wang: Writing – review & editing, Visualization, Methodology. Hongzhen Zhang: Writing – review & editing, Formal analysis. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.tranon.2026.102842. Contributor Information Shikai Wang, Email: wangshikai007@126.com. Hongzhen Zhang, Email: 18194005997@163.com. Supplementary materials Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.