NUAK1 silencing enhances radiotherapy-induced ferroptosis in locally advanced rectal cancer by impairing Nrf2-driven transcription of GPX4.

Introduction
Locally advanced rectal cancer (LARC) is one of the most challenging gastrointestinal malignancies worldwide [1]. The standard treatment approach for LARC is neoadjuvant chemoradiotherapy (nCRT) based on radiotherapy and combined with total mesorectal excision (TME). This approach reduces tumor volume and downstages the disease, thereby improving rates of radical resection, anal sphincter preservation, and local disease control [2]. A critical indicator of nCRT efficacy is pathological complete response (pCR), defined by the absence of viable tumor cells in the surgical specimen. Achieving pCR is strongly associated with superior long-term disease-free survival (DFS) and overall survival (OS), representing the optimal therapeutic goal [3]. Nonetheless, the clinical reality is challenging, as only approximately 15–20% of patients achieve pCR. This statistic implies that over 80% of patients experience varying degrees of radioresistance, resulting in poor outcomes, including local recurrence and distant metastasis [4]. Therefore, elucidating the molecular mechanisms that contribute to radioresistance and developing predictive biomarkers are crucial for addressing existing therapeutic limitations and enhancing patient prognosis.
A primary strategy to overcome radioresistance is to enhance tumor cell sensitivity to ionizing radiation. Traditional research has focused predominantly on increasing irreparable DNA double-strand breaks or inducing apoptosis [5]. However, tumor cells exhibit high heterogeneity and plasticity, often bypassing these mechanisms through compensatory pathways, which limits the efficacy of single-target strategies [6]. In recent years, with the deepening understanding of cell death, ferroptosis—an iron-dependent, lipid peroxidation-driven form of regulated cell death—has been identified as a key effector pathway for eliminating tumor cells under oxidative stress conditions such as radiotherapy [7]. Radiotherapy generates abundant reactive oxygen species (ROS) through water radiolysis; these ROS attack polyunsaturated fatty acids in cell membranes, triggering a lethal chain reaction of lipid peroxidation. To survive, tumor cells have evolved sophisticated antioxidant defense systems, in which GPX4 plays a vital role by reducing lipid peroxides and suppressing ferroptosis [8]. Nevertheless, the upstream signaling events driving GPX4 upregulation in response to radiotherapy remain poorly understood in LARC. Elucidating this regulatory mechanism would enable disruption of tumor self-defense at its source, rather than merely targeting the terminal effector.
NUAK1, a member of the AMPK-related kinase family, is widely recognized as a crucial survival hub for cells under energy stress, hypoxia, and metabolic challenge [9]. Substantial evidence indicates that NUAK1 is aberrantly overexpressed in various solid tumors, including gastric, liver, esophageal, and breast cancers [10–13], and promotes tumor cell proliferation, invasion, and metastasis by phosphorylating downstream substrates such as MYPT1 [14–16]. In colorectal cancer, NUAK1 has been identified as an oncogene-driven protein associated with poor prognosis [17, 18]. However, whether NUAK1 regulates radiosensitivity in LARC via crosstalk with the newly recognized process of ferroptosis remains unknown. Noting a potential link between NUAK1 activation and cellular antioxidant stress responses, we hypothesized that NUAK1 might serve as a critical molecular bridge, connecting radiotherapy stress to the cellular ferroptosis defense system as a sensor and signal amplifier.
This study aims to systematically investigate the biological function and precise molecular mechanisms of NUAK1 in LARC radioresistance. By integrating in vitro cell models, in vivo animal experiments, and clinical LARC tissue samples, we unveil a novel signaling axis: Radiotherapy activates NUAK1, which promotes the nuclear translocation of the master antioxidant transcription factor Nrf2 and enhances its binding to antioxidant response elements (ARE), thereby transcriptionally upregulating its key target gene, GPX4. This molecular cascade is essential for tumor cells to counteract radiotherapy-induced ferroptosis. Using the specific Nrf2 inhibitor ML385, we further confirmed that Nrf2 is an indispensable core mediator within this pathway. Moreover, our clinical analysis indicates that NUAK1 protein level is a potential biomarker for predicting radiosensitivity in LARC patients. In summary, our research not only deepens the understanding of radioresistance mechanisms by uncovering the NUAK1-Nrf2-GPX4 axis as a new functional unit, but also provides a potential combination strategy and a predictive biomarker for overcoming clinical radioresistance in LARC.

Materials and methods

Patient samples
This study encompassed rectal cancer patients who underwent nCRT at the First Affiliated Hospital of Fujian Medical University between January 2020 and January 2022. Detailed inclusion and exclusion criteria for patient selection are provided in Supplementary Material 1. The study ultimately included a total of 17 tumor specimens collected before radiotherapy, along with the corresponding clinical data. All participants received standard long- course neoadjuvant radiotherapy prior to surgery. TME was consistently performed by the same experienced colorectal surgeon 6 to 8 weeks following the completion of radiotherapy. The study adhered to the Helsinki Declaration and was approved by the Ethics Committee of First Affiliated Hospital of Fujian Medical University (NO. [2021]274). All patients provided written informed consent.

Radiotherapy sensitivity evaluation
This study focused on the short-term effects of preoperative radiotherapy in rectal cancer, using pathological response as an indicator of radiosensitivity, rather than local recurrence or survival rates. Experienced pathologists assessed all available tumor samples to confirm histological type and evaluated them using the tumor regression grade (TRG) criteria from the 7th edition of the AJCC manual. TRG0 represented complete response with no viable tumor cells; TRG1 represented moderate response with rare tumor cells; TRG2 represented minimal response with residual tumor outgrowing fibrosis; and TRG3 represented poor response with extensive residual tumor. Tumors with TRG0-1 were categorized as radiosensitive, whereas those with TRG2-3 were categorized as radioresistant.

Immunohistochemical (IHC) staining
Immunohistochemistry analysis was performed as previously described [12]. Sections were incubated overnight at 4 °C with anti-NUAK1 (1:200 dilution, 22723-1-AP, Proteintech) and anti-4-HNE (1:400 dilution, ab46545, Abcam) antibodies, with PBS as the negative control. Finally, DAB development was performed using a DAB kit (Servicebio, China). Images of five random fields per section were captured using an Olympus MVX10 microscope (Olympus, Japan). The staining results were assessed by two independent, experienced pathologists. The immunohistochemical score was quantified by calculating the product of the staining intensity and the percentage of positive cells. The percentage of positively stained cells was scored on a scale from 0 to 5 (0 = 0–5%; 1 = 6–25%; 2 = 26–50%; 3 = 51–75%; 4 = 76–100%), and the staining intensity was rated from 0 to 3 (0 = negative staining; 1 = weak; 2 = moderate; 3 = strong).

Cell culture and lentiviral transfection
Colorectal cancer cell lines (HCT116, SW620, SW837, and DLD-1) were purchased from GeneChem (Shanghai, China). All cell lines were authenticated by short tandem repeat (STR) profiling and confirmed to be free of mycoplasma contamination. Cells were maintained at 37 °C in a 5% CO2 incubator using RPMI 1640 medium with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Lentiviral vectors for NUAK1 and GPX4 overexpression and NUAK1 knockdown were purchased from Zolgene Biotechnology (Fuzhou, China). Cells were infected with the viral supernatant and then selected with puromycin (2 µg/mL). Empty vector was used as a negative control. The primer sequences used for lentiviral transfection in this study were listed in Table S1 (Supplementary Material 2). Transfection efficiency was verified by Western blot and Quantitative real-time polymerase chain reaction (qPCR).

Cell irradiation
During the logarithmic growth phase, cells were irradiated with X-rays at a dose rate of 3 Gy/min using a medical electron linear accelerator (Varian, USA). Following ionizing radiation (IR), the external surfaces of the culture flasks were wiped with 75% ethanol and returned to the incubator for continued cultivation.

CCK-8 assays
Cell viability was assessed using the Cell Counting Kit-8 (APExBIO, USA) according to the manufacturer’s instructions. Cells were seeded at 3000 per well in 96-well plates. In specified experiments, cells were pre-treated with cell death inhibitors for 24 h before exposure to X-ray irradiation. Optical density (OD) at 450 nm was measured using an enzyme reader (Bio-Rad, USA). A blank containing only RPMI 1640 medium was used to correct the OD values.

Colony formation assay
The radiosensitivity of CRC cells was determined using a colony formation assay. Cells (200–2000/well) were seeded in 6-well plates and treated with different doses of IR (2, 4, 6, 8 Gy) after attachment. In specified experiments, cells were pre-treated with Ferrostatin-1 (2 µM, HY-100579, MedChemExpress) or ML385 (10 µM, HY-100523, MedChemExpress) for 24 h before exposure to X-ray irradiation. Following irradiation, the medium was replaced with fresh medium containing the same concentration of the respective inhibitor. The inhibitors were replenished with every medium change (every 2 day) and maintained throughout the subsequent colony formation period. After 14 days, colonies were fixed with 4% paraformaldehyde for 15 min and stained with 0.2% crystal violet for 30 min. Colonies containing > 50 cells were counted to determine the survival fraction, and the survival curves were fitted using the single-hit multi-target model.

ROS and lipid peroxidation assay
In experiments involving Nrf2 inhibition, cells were pre-treated with ML385 (10 µM) or vehicle control (DMSO) for 24 h before exposure to irradiation. At 24 h post-irradiation, cells were harvested for assessment. Total ROS levels were measured using 10 µM DCFH-DA (Beyotime, China), while lipid peroxidation was assessed with 5 µM BODIPY 581/591 C11 dye (Invitrogen, USA). After a 30-minute dark incubation, cells were trypsinized, washed with PBS to remove unbound dye, and resuspended in 0.5 ml PBS for analysis using a flow cytometer (Bio-Rad, USA).

Measurement of intracellular MDA and Fe2+ levels
Cells were exposed to 6 Gy X-rays and then cultured for an additional 48 h. The cells were then rinsed with PBS, and the intracellular MDA or Fe2+ levels were measured using the MDA Assay kit (Nanjing Jiancheng Institute, China) and Ferrous Iron Colorimetric Assay Kit (Elabscience, China) according to the manufacturer’s instructions.

Transmission electron microscopy (TEM)
Mitochondrial morphology was examined using a transmission electron microscope. Briefly, cells cultured in 6-well plates were fixed with buffer containing 2.5% glutaraldehyde and 0.1 M phosphate. The samples were then washed, dehydrated, and embedded in resin according to standard procedures. The resin blocks were sectioned into ultrathin slices, which were then observed using a transmission electron microscope (Hitachi, Japan).

RNA extraction and qPCR assay
Total RNA was extracted from CRC cells using the RNAeasy™ Kit (Beyotime, China) and cDNA was synthesized with the RevertAid Kit (ThermoFisher, USA). Qpcr was performed using PowerUp™ SYBR™ Green Master Mix (ThermoFisher, USA) on an ABI QuantStudio 5 system (Applied Biosystems, USA), with β-Actin as an internal control. Gene expression levels were calculated using the 2 − ΔΔCt method. The primer sequences used for qPCR in this study were listed in Table S2 (Supplementary Material 2).

Western blot assay
Cells were lysed in buffer containing protease inhibitor and phosphatase inhibitor cocktails. BCA Protein Assay Kits (Beyotime, China) were used to determine protein concentrations. Total proteins were separated by SDS-PAGE and transferred onto PVDF membranes. The membranes were then incubated overnight at 4 °C with primary antibodies against NUAK1 (22723-1-AP, Proteintech), GPX4 (ab125066, Abcam), β-actin (bs-0061R, Bioss), Nrf2 (16396-1-AP, Proteintech), Lamin B1 (12987-1-AP, Proteintech), MYPT1 (22117-1-AP, Proteintech), Phospho-MYPT1 (Ser445) (CABT-BL6360, Creative Diagnostics), and NQO1 (11451-1-AP, Proteintech). The membranes were washed three times and then incubated with a secondary antibody for 1 h at room temperature. After incubation, the membranes were washed three times again. Finally, the antibody-antigen complex signal was detected using the ECL chemiluminescence kit (Beyotime, China), according to the manufacturer’s instructions.

Co-immunoprecipitation (Co-IP) assay
Co-IP was performed to assess protein interactions. Cell lysates prepared in RIPA buffer with protease inhibitors were incubated overnight at 4 °C with antibodies against NUAK1 (22723-1-AP, Proteintech), Nrf2 (16396-1-AP, Proteintech), or control IgG, followed by capture with Protein A/G magnetic beads. Beads were washed and bound proteins were eluted in SDS loading buffer. Co-precipitated proteins and input lysates were analyzed by Western blotting as described above, using antibodies against NUAK1, Nrf2, and MYPT1 (22117-1-AP, Proteintech).

Dual-luciferase reporter assay
The dual-luciferase reporter assay was performed to investigate the impact of NUAK1 depletion on the activity of both the GPX4 promoter and the ARE. A reporter plasmid containing the GPX4 promoter was synthesized and constructed by Zolgene Biotechnology (Fuzhou, China). The ARE-Luc reporter plasmid was purchased from Yeasen Biotechnology (Shanghai, China). Cells were co-transfected with the respective reporter plasmids and the Renilla luciferase plasmid (pRL-TK) as an internal control. At 48 h post-transfection, cells were harvested, and the firefly and Renilla luciferase activities were sequentially measured using the Dual Luciferase Reporter Assay Kit (Vazyme Biotech, China) following the manufacturer’s instructions. The relative luciferase activity was calculated by normalizing the firefly luminescence to the Renilla luminescence for each sample.

Xenograft tumor mouse model
Male BALB/c nude mice (4–6 weeks old, n = 18) were purchased from Shanghai Slake Laboratory Animal Co., LTD. Nude mice were randomly assigned to six groups of more than three mice each. Cells were irradiated with 4 Gy, and within 2 hours, 3 × 10⁶ irradiated cells were resuspended and injected subcutaneously into mice to establish xenograft models. A caliper was used to measure the length and width of each tumor every other day. After 2 weeks of injection, mice were euthanized, and xenograft specimens were surgically excised. The tumor volume was calculated using the following formula: Tumor volume (mm3) = (length×width2) / 2. All animal procedures were approved by the Animal Ethics Committee of Fujian Medical University (NO. 2023-Y-0548).

Bioinformatic analyses
The GSE123924 dataset from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) was used to explore the potential mechanism underlying NUAK1 knockdown. This dataset comprises three samples of control cells and three samples of NUAK1-knockdown cells. The ‘limma’ R package was used to identify differentially expressed genes (DEGs) between them, with a threshold adjusted P value < 0.05. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses of the DEGs were performed using the ‘clusterProfiler’ R package. Separately, to evaluate the clinical predictive value of key pathway genes, pre-treatment expression data from LARC patient cohorts (GSE35452, GSE45404, and GSE119409) were used for analyzed. For each dataset, receiver operating characteristic (ROC) curves were constructed based on the expression levels of NUAK1, GPX4, and Nrf2 to assess their ability to distinguish between patients with poor versus good response to neoadjuvant chemoradiotherapy.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 10 software (GraphPad, USA). Bioinformatics analyses were performed using R software (Version 4.3.0). Two-group comparisons were analyzed using the Student’s t-test (two-tailed). One-way analysis of variance (ANOVA) was used for multiple group comparisons. The Chi-square test was used to compare categorical variables between different groups. When the assumptions of the Chi-square test were not met, Fisher’s exact test was used for the comparison. The gene expression correlations were computed using the Pearson correlation test. Quantitative data are presented as mean ± standard deviation (SD). The number of biologically independent replicates (n) for each experiment is specified in the corresponding figure legends. A P value of ≤ 0.05 was considered statistically significant.

Result

Ferroptosis correlates with radiosensitivity in CRC cells
To investigate the association between ferroptosis and radiosensitivity, we identified SW620 as a radiosensitive cell line and DLD-1 as a radioresistant cell line using clonogenic survival assays (Fig. 1A). As shown in Fig. 1B and C, ROS levels were significantly elevated at 24 h after 6 Gy irradiation across all four cell lines. Furthermore, the level of MDA, a terminal product of lipid peroxidation, was markedly increased in three cell lines (HCT116, SW620, SW837) following radiation (P < 0.05, Fig. 1D), whereas no significant change was observed in the radioresistant DLD-1 cells (P > 0.05). PTGS2, a recognized ferroptosis marker, was significantly upregulated in SW620 cells after irradiation (P < 0.01), but remained unchanged in DLD-1 cells (P > 0.05, Fig. 1E). Transmission electron microscopy revealed that irradiated SW620 cells displayed the characteristic morphological features of ferroptosis, including shrunken mitochondria with condensed membrane density and reduced or lost cristae (Fig. 1F). To determine whether ferroptosis serves as a major contributor to radiation-induced cell death, we pretreated cells with ferroptosis inhibitors (Ferrostatin-1, Deferoxamine), apoptosis inhibitor (Z-VAD-FMK), or necroptosis inhibitor (Necrostatin-1) for 24 h prior to radiation, followed by CCK-8 assay. Pretreatment with these inhibitors significantly enhanced cell viability compared to radiation alone (Fig. 1G). Notably, ferroptosis inhibitors markedly rescued cell viability, with recovery rates of 32.9% (Ferrostatin-1) and 32.1% (Deferoxamine), suggesting a dominant role for ferroptosis in radiation-induced cell death. Finally, we translated these findings to clinical relevance. Immunohistochemical staining for 4-HNE, a lipid peroxidation marker, was significantly enhanced in post-radiotherapy samples. Critically, tumors from radiotherapy-sensitive patients exhibited more intense 4-HNE staining than those from resistant patients, indicating that radiation-induced lipid peroxidation may underlie differential treatment sensitivity (Fig. 1H).

NUAK1 deficiency sensitizes CRC cells to radiation
To select a suitable cell line with high baseline NUAK1 expression for knockdown experiments, we measured the mRNA and protein levels of NUAK1 in four CRC cell lines. Results showed that the radioresistant DLD-1 cell line exhibited the highest NUAK1 expression (Fig. 2A, B), and was therefore selected for subsequent functional studies. Lentiviral transduction efficiency in DLD-1 cells was confirmed by qPCR and Western blot, with shNUAK1-2 and shNUAK1-3 constructs demonstrating the highest knockdown efficiency (Fig. 2C, D). These two knockdown constructs were used in parallel for the following experiments.
Assessment of cell viability revealed no significant difference in growth among the shNC, shNUAK1-2, and shNUAK1-3 groups in the absence of irradiation. However, by day 7 post-irradiation, both shNUAK1-2 and shNUAK1-3 groups showed significantly reduced viability compared to the shNC control group (Fig. 2E, F). Furthermore, the single-hit multi-target model based on clonogenic survival assays demonstrated that NUAK1 knockdown cells (shNUAK1-2 and shNUAK1-3) exhibited markedly enhanced radiosensitivity relative to shNC controls (Fig. 2G, H). The tumorigenic potential was further evaluated in vivo using a xenograft mouse model. Both shNUAK1-2 and shNUAK1-3 groups displayed significantly inhibited tumor growth rates following radiation compared to the shNC group (Fig. 2I, J), and ultimately formed much smaller tumors (Fig. 2K). These findings collectively indicate that NUAK1 knockdown enhanced the radiosensitivity of DLD-1 cells.

NUAK1 knockdown enhances radiation-induced ferroptosis
We next investigated whether NUAK1 knockdown influences radiation-induced ferroptosis. Intracellular ROS and lipid peroxidation levels were measured in the three cell groups following irradiation. Results showed that ROS levels were significantly higher in both shNUAK1-2 and shNUAK1-3 groups compared to the shNC group after radiation (Fig. 3A, B). Similarly, lipid peroxidation levels were markedly elevated in NUAK1-knockdown cells post-irradiation (Fig. 3C, D). Immunofluorescence analysis revealed that the intensity of green fluorescence (oxidized BODIPY™ 581/591 C11) was substantially stronger in shNUAK1-2 and shNUAK1-3 groups than in shNC controls after radiation (Fig. 3E), indicating enhanced lipid peroxidation. We further assessed MDA levels after radiation. Consistent with the above findings, MDA content was significantly higher in shNUAK1-2 and shNUAK1-3 groups compared to the shNC group following radiation (Fig. 3F). Surprisingly, intracellular Fe²⁺ levels showed no significant differences among the groups (Fig. 3G), suggesting that NUAK1 may not be involved in iron metabolism. To establish whether ferroptosis is necessary for the enhanced radiosensitivity observed upon NUAK1 knockdown, we performed a rescue experiment. Cells were pretreated with the ferroptosis inhibitor Ferrostatin-1 for 24 h prior to irradiation, followed by clonogenic survival assay. Results showed that Ferrostatin-1 pretreatment significantly restored the colony-forming ability of NUAK1-knockdown cells after radiation compared to the irradiation-only group (Fig. 3H, I), indicating that inhibition of ferroptosis can reverse the radiosensitizing effect caused by NUAK1 depletion.

NUAK1 knockdown suppresses GPX4 expression
We analyzed the GSE123924 dataset from the GEO database to investigate the downstream biological processes affected by NUAK1 knockdown. The results revealed a distinct gene expression profile between the two groups, leading to the identification of 5,354 DEGs (Fig. 4A). KEGG pathway enrichment analysis of these DEGs highlighted significant alterations in several pathways, including colorectal cancer, platinum drug resistance in colorectal cancer, p53 signaling, fatty acid metabolism and glutathione metabolism (Fig. 4B). Notably, all these enriched pathways have established associations with ferroptosis. Given that GPX4 is a central regulator of ferroptosis and is known to be inactivated by RSL3, we assessed the functional consequence of NUAK1 knockdown on GPX4 activity by determining the IC₅₀ of RSL3. As shown in Fig. 4C, both shNUAK1-2 and shNUAK1-3 cells exhibited significantly lower IC₅₀ values (5.54 µM and 5.14 µM, respectively) compared to shNC controls (7.27 µM).
We therefore investigated the correlation between NUAK1 and GPX4 expression. The qPCR analysis demonstrated that GPX4 mRNA was most highly expressed in DLD-1 cells and showed the lowest expression level in SW620 cells (Fig. 4D), mirroring the expression pattern of NUAK1. Pearson correlation analysis confirmed a strong positive correlation between NUAK1 and GPX4 mRNA levels across the four CRC cell lines (r = 0.92, P < 0.001, Fig. 4E). This positive correlation was further validated in a rectal adenocarcinoma (READ) cohort from the GEPIA database (Fig. 4F). Consistently, both qPCR and Western blot analyses confirmed that GPX4 expression was significantly downregulated at both the mRNA and protein levels upon NUAK1 knockdown (Fig. 4G, H). These results demonstrate that NUAK1 positively regulates GPX4 expression.

GPX4 overexpression reverses the radiosensitizing effect of NUAK1 knockdown
To determine whether GPX4 serves as the critical downstream effector responsible for the radioresistance conferred by NUAK1, we conducted a rescue experiment by overexpressing GPX4 in the shNUAK1-3 cell line, which was selected as a representative model for NUAK1 knockdown. Remarkably, the restoration of GPX4 expression almost completely reversed the radiosensitization phenotype induced by NUAK1 deficiency. We first confirmed the efficiency of GPX4 overexpression in NUAK1-knockdown DLD-1 cells via Western blot analysis (Fig. 5A). We then found that GPX4 overexpression substantially attenuated the key ferroptotic hallmarks observed in irradiated shNUAK1 cells, including elevated lipid peroxidation (Figs. 5B, C), characteristic mitochondrial shrinkage with heightened membrane density (Fig. 5D), and increased MDA levels (Fig. 5E). At the functional level, this molecular rescue translated into a significant recovery of clonogenic survival in GPX4-overexpressing cells following radiation treatment (Figs. 5 F, 5G). The critical role of this regulatory axis was further solidified in a xenograft mouse model, where GPX4 overexpression effectively suppressed the enhanced tumor growth inhibition resulting from NUAK1 knockdown in vivo (Figs. 5 H, 5I). Collectively, our multi-faceted data robustly demonstrate that GPX4 is the indispensable functional mediator through which NUAK1 protects colorectal cancer cells from radiotherapy-induced ferroptosis.

NUAK1 regulates GPX4 expression and ferroptosis via Nrf2
To elucidate how NUAK1 regulates GPX4 expression, we first examined GPX4 promoter activity. Radiation significantly enhanced promoter activity in control cells, but this induction was markedly attenuated upon NUAK1 knockdown (Fig. 6A), indicating that NUAK1 regulates GPX4 transcriptionally. Given Nrf2’s central role in antioxidant responses, we hypothesized it mediates NUAK1’s effect. Nuclear-cytoplasmic fractionation revealed that radiation induced Nrf2 nuclear translocation in control cells, but this process was substantially impaired in NUAK1-knockdown cells, which retained more Nrf2 in the cytoplasm (Fig. 6B). This demonstrates NUAK1 is required for radiation-induced Nrf2 nuclear translocation. We next assessed Nrf2 transcriptional activity using an ARE-driven luciferase reporter. Radiation strongly activated ARE reporter in control cells, but this response was significantly blunted in NUAK1-knockdown cells (Fig. 6C). These results collectively establish that NUAK1 is indispensable for both the nuclear translocation and transcriptional activation of Nrf2 following radiation exposure.
To determine whether this regulation involves a direct physical interaction, we performed Co-IP assays. As shown in Fig. S1, while NUAK1 efficiently co-precipitated its established substrate MYPT1, it did not form a stable complex with Nrf2. This finding suggests that NUAK1 regulates Nrf2 activation through an indirect mechanism.
To confirm Nrf2 as the essential mediator of NUAK1 in radioresistance, we employed the Nrf2 inhibitor ML385. Western blot analysis (Fig. 6D) demonstrated that NUAK1 overexpression enhanced the phosphorylation of its substrate MYPT1 and upregulated Nrf2 targets NQO1 and GPX4. ML385 treatment completely abolished this upregulation despite NUAK1 activation. Combined NUAK1 knockdown and Nrf2 inhibition showed no additive effect, confirming their operation within the same pathway. Functional assessment (Figs. 6E, F) revealed NUAK1 knockdown significantly reduced clonogenic survival after radiation, while NUAK1 overexpression conferred strong radioprotection. Notably, ML385 treatment completely abolished the survival advantage provided by NUAK1 overexpression, establishing that NUAK1’s pro-survival function is Nrf2-dependent. Corresponding lipid peroxidation assays (Figs. 6G, H) showed NUAK1 knockdown exacerbated radiation-induced lipid peroxidation, whereas NUAK1 overexpression suppressed it. This antioxidant protection was nullified by Nrf2 inhibition, with lipid peroxidation returning to levels observed in Nrf2-inhibited controls. No additive effect was observed with combined NUAK1 knockdown and Nrf2 inhibition. These results provide definitive evidence that Nrf2 serves as the essential effector through which NUAK1 exerted radioprotection by suppressing radiation-induced ferroptosis.

NUAK1 expression predicts clinical radiosensitivity in LARC patients
To demonstrate the correlation between NUAK1 expression and clinical pathological features in LARC patients, immunohistochemical analysis was performed to assess the protein expression level of NUAK1 in 17 cases of LARC tissues before nCRT. Baseline characteristics of patients as shown in Table 1, Most rectal cancer patients receiving preoperative neoadjuvant therapy were male (70.6%) and aged 56 to 67. Tumors were typically located within 5 cm of the anal margin, averaging 4.8 cm in length on MRI. The overall effectiveness of neoadjuvant therapy was moderate, with 58.8% of patients showing moderate tumor regression (TRG2), 17.6% achieving pCR, and a radiotherapy sensitivity rate of 41.1%.

To establish clinical relevance, we assessed NUAK1 protein expression in LARC patients. IHC revealed heterogeneous NUAK1 expression (Fig. 7A), with significantly higher scores in radioresistant versus sensitive patients (Fig. 7B). Stratification by mean IHC score (6.98) showed high NUAK1 expression was predominant in the radioresistant cohort, while most sensitive patients exhibited low levels (Fig. 7C). ROC analysis showed predictive power (AUC = 0.978), suggesting NUAK1 as a potential biomarker for distinguishing radioresistant LARC (Fig. 7D). Clinically, high NUAK1 correlated with higher pre-nCRT T-stage, inferior tumor regression, and significantly lower pCR rates (Table 2). To independently validate the clinical relevance of our findings, we also analyzed the pre-treatment mRNA expression of NUAK1, GPX4, and Nrf2 in multiple publicly available LARC cohorts treated with nCRT. In the GSE35452 dataset, NUAK1 expression was significantly associated with pathological response, with an AUC of 0.769 (Fig. 7E). Conversely, GPX4 and Nrf2 showed limited predictive value (AUC = 0.555 and 0.498, respectively). We further expanded our analysis to two additional cohorts (GSE45404 and GSE119409), in which NUAK1 consistently demonstrated predictive power (AUC = 0.808 and 0.698), while GPX4 and Nrf2 again showed inconsistent and poor performance (Fig. S2). Collectively, these multi-cohort bioinformatics analyses nominate NUAK1, but not GPX4 or Nrf2, as a promising transcriptional biomarker for nCRT response.

Discussion
Although radiotherapy is a standard treatment for LARC, a subset of patients derives limited benefit due to radioresistance, instead experiencing radiation injury, increased surgical complexity, and the declining quality of life [19, 20]. Our study elucidates a novel adaptive mechanism underlying this clinical challenge, demonstrating that tumors exploit the NUAK1-Nrf2-GPX4 signaling axis to reinforce their antioxidant defenses and thereby conferring radioresistance by evading radiotherapy-induced ferroptosis.
In recent years, ferroptosis has been established as a significant contributor to radiation-induced cell death [21–23]. In our study, radiosensitive SW620 cells exhibited significantly more ferroptosis after radiation than radioresistant DLD-1 cells. Moreover, ferroptosis inhibitors were more effective than other cell death inhibitors in rescuing cells from radiation-induced cell death, confirming ferroptosis as a key mechanism of radioresistance in colorectal cancer. We further identified NUAK1 as a critical regulator of this process. NUAK1 deficiency significantly enhanced the radiosensitivity of DLD-1 cells both in vitro and in vivo. Notably, unlike its reported role in cancer cell proliferation [10, 11], NUAK1 knockdown itself had a minimal impact on cell viability, strongly suggesting its function is specific to radiation-induced stress rather than general growth inhibition. Mechanistically, NUAK1 deficiency significantly exacerbated radiation-induced ferroptosis. This finding is consistent with other studies, for example, Heng et al. reported that ACAT2 knockdown significantly increased radiation-induced ferroptosis in esophageal squamous cell carcinoma, suggesting ACAT2 as a potential therapeutic target [24]. Mao et al. found that SCARB1 inhibits ferroptosis and promotes radioresistance by regulating cholesterol metabolism [25]. Crucially, the radiosensitization caused by NUAK1 deficiency was significantly reversed by ferroptosis inhibitors, demonstrating that ferroptosis is the necessary pathway for NUAK1 to regulate radiosensitivity.
To delineate the downstream molecular targets of NUAK1, we analyzed the NUAK1-associated dataset GSE123924 from the GEO database. Enrichment analysis revealed significant involvement of pathways related to colorectal cancer and ferroptosis. Notably, our enrichment analysis highlighted p53 signaling, which aligns with the established role of p53 in promoting lipid peroxidation and ferroptosis by antagonizing SLC7A11 expression and inhibiting glutathione synthesis following ionizing radiation [26]. Fatty acid metabolism also plays a critical role in ferroptosis, as lipid peroxidation of polyunsaturated fatty acids (PUFAs) contributes to ferroptosis [27]. Furthermore, the glutathione pathway was among the first identified as crucial for ferroptosis regulation [28]. Given that GPX4 serves as a central effector for multiple ferroptosis-related pathways, we sought to investigate its relationship with NUAK1. RSL3, a direct GPX4 inhibitor, covalently binds to the selenocysteine residue in the GPX4 active site, inhibiting its enzymatic activity and thereby inducing ferroptosis [29]. We found that NUAK1-deficient cells exhibited significantly increased sensitivity to RSL3, indicating that NUAK1 loss reduces intracellular GPX4 levels. Moreover, a strong positive correlation between NUAK1 and GPX4 mRNA expression was observed across four CRC cell lines and was further validated in the rectal cancer cohort from GEPIA. Ultimately, qPCR and Western blot analyses directly confirmed that NUAK1 regulates GPX4 expression at the molecular level. Through a series of functional rescue experiments, we demonstrated that GPX4 downregulation is the core driver of both radiosensitization and exacerbated ferroptosis, rather than a concomitant phenomenon of NUAK1 deficiency. Overexpression of GPX4 significantly suppressed the enhanced radiation-induced lipid peroxidation, the destruction of mitochondrial ultrastructure, and the elevated MDA levels caused by NUAK1 knockdown. Most importantly, re-expression of GPX4 effectively restored the clonogenic survival and in vivo tumorigenic capacity of NUAK1-deficient cells, demonstrating that GPX4 is the essential downstream mediator through which NUAK1 loss activates ferroptosis and ultimately confers radiosensitization.
To investigate the molecular mechanism by which NUAK1 regulates GPX4 expression, we first examined the impact of NUAK1 deficiency on the GPX4 promoter activity. Dual-luciferase reporter assays revealed that NUAK1 knockdown significantly attenuated the radiation-induced enhancement of GPX4 transcriptional activity. This finding prompted us to explore transcriptional regulation. Given that the transcription factor Nrf2 serves as a master regulator of the antioxidant response, controlling a network of genes involved in glutathione metabolism and ROS clearance [30], we hypothesized that NUAK1 mediates its transcriptional control of GPX4 through Nrf2. To test this hypothesis, we first assessed the effect of NUAK1 on Nrf2 subcellular localization. Nuclear-cytoplasmic fractionation experiments demonstrated that NUAK1 deficiency effectively suppressed radiation-induced Nrf2 nuclear translocation. Furthermore, an ARE-driven reporter assay showed that NUAK1 knockdown significantly impaired Nrf2 transcriptional activity. These results establish NUAK1 as a positive regulator of both Nrf2 nuclear translocation and its transcriptional function.
To further confirm Nrf2 as an indispensable downstream mediator in the NUAK1 signaling pathway, we employed the Nrf2-specific inhibitor ML385 [31]. In cells with functional NUAK1, ML385 treatment not only completely blocked Nrf2 transcriptional activity after radiation but also effectively suppressed the expression of its downstream target, GPX4. The combination of ML385 and NUAK1 knockdown showed no additive inhibition on Nrf2 transcriptional activity. More importantly, in NUAK1-overexpressing cells, the introduction of ML385 likewise inhibited the NUAK1-mediated upregulation of GPX4 following radiation. This indicates that even under conditions of NUAK1 activation, its ability to regulate GPX4 strictly depends on Nrf2 activity.
While our data clearly place NUAK1 upstream of Nrf2 activation, the precise biochemical link remained to be defined. Co-IP experiments revealed that NUAK1 does not form a stable complex with Nrf2, arguing against a direct kinase-substrate relationship. This finding logically points to an indirect regulatory mechanism. Interestingly, our data and existing literature suggest a plausible signaling cascade: NUAK1 activation leads to phosphorylation of its downstream substrate MYPT1, and phosphorylated MYPT1 is known to negatively regulate the activity of Glycogen Synthase Kinase 3 Beta (GSK3β) [15]. GSK3β, in turn, is a well-established kinase that phosphorylates Nrf2, promoting its nuclear export and degradation [32]. Therefore, a coherent model emerges whereby NUAK1, through p-MYPT1, may inhibit GSK3β activity, thereby stabilizing Nrf2 and facilitating its nuclear accumulation. This cascade explains how a cytosolic kinase can regulate a nuclear transcription factor without direct binding, revealing a novel kinase-based activation mechanism beyond the classical KEAP1-dependent degradation pathway, and expands our understanding of how cancer cells dynamically fine-tune their antioxidant capacity under stress.
Analysis of tissue samples from LARC patients revealed that NUAK1 protein levels were significantly elevated in radioresistant tissues compared to their radiosensitive counterparts. Moreover, high NUAK1 expression was significantly correlated with inferior pathological response and advanced tumor stage, aligning with both its role in promoting tumor aggressiveness [33] and prior studies that have linked Nrf2 pathway activation to poor treatment response [34, 35]. ROC curve analysis of our cohort further indicated that NUAK1 protein level possessed significant predictive power for radiotherapy response. Importantly, a notable and informative contrast emerged from analysis of independent public transcriptomic datasets: the predictive power of GPX4 and Nrf2 mRNA expression was inconsistent and generally poor. This observation aligns with and strengthens our findings. First, as a downstream effector, GPX4 expression is subject to multiple layers of regulation beyond the NUAK1-Nrf2 axis, which may obscure a direct correlation with response in heterogeneous patient tumors. Second, Nrf2 activity is predominantly controlled post-translationally (e.g., via KEAP1), meaning its mRNA level is a poor surrogate for its functional state. Thus, the stable predictive performance of NUAK1 across cohorts underscores its position as the upstream driver of this radioprotective pathway, whose expression level may more reliably capture the pathway’s activation state and the tumor’s intrinsic resistance potential.
These cumulative clinical observations suggest that IHC-based detection of NUAK1 could be developed into a companion diagnostic tool to identify LARC patients with radioresistance. For these high-risk patients, enrollment in clinical trials testing NUAK1 inhibitors in combination with radiotherapy is warranted.
Although our study establishes the critical role of the NUAK1-Nrf2-GPX4 signaling axis in radioresistance of LARC, some limitations should be acknowledged. Mechanistically, while our data rule out a direct interaction, the precise molecular intermediaries such as GSK3β linking NUAK1 activation to Nrf2 nuclear translocation remain to be fully validated, and the evidence of Nrf2’s transcriptional control over GPX4 needs to obtain direct through approaches like ChIP and ARE-mutant reporter assays. It should also be noted that radiation is known to trigger a mixture of cell death modalities. While our data establish ferroptosis as the dominant pathway through which NUAK1 loss exerts its radiosensitizing effect, we cannot exclude the possibility that apoptosis or other forms of cell death contribute to the overall radiation response in a minor or context-dependent manner.
From a therapeutic perspective, future efforts will focus on employing selective NUAK1 inhibitors to pharmacologically validate the NUAK1–Nrf2–GPX4 axis, and to evaluate their efficacy as radiosensitizers in patient-derived organoid models and subsequent preclinical trials. Furthermore, although our preliminary data from a limited size reveal a significant correlation between NUAK1 expression and radiotherapy response, these findings need verification in larger, multi-center patient cohorts to establish their clinical applicability (Fig. 8).

Conclusion
This study systematically uncovers a novel role for NUAK1 in regulating radiotherapy-induced ferroptosis in LARC. We found that NUAK1 deficiency led to the downregulation of GPX4 expression and increased radiosensitivity, a phenotype that could be reversed by restoring GPX4. More importantly, using the Nrf2-specific inhibitor ML385, we demonstrated that Nrf2 serves as an indispensable core molecule for NUAK1’s function. In summary, we propose a novel signaling pathway (radiotherapy→NUAK1→Nrf2→GPX4→ferroptosis), which not only reveals a new survival mechanism in colorectal cancer but also provides innovative combination therapeutic strategies and predictive biomarkers for overcoming clinical radioresistance.

Supplementary Information
Below is the link to the electronic supplementary material.