Exosomal lncRNA NORAD drives oxaliplatin resistance in AEG via the miR-433-3p/autophagy axis: a novel mechanism and potential biomarker

Abstract

Background

Oxaliplatin resistance poses a substantial therapeutic challenge in adenocarcinoma of the esophagogastric junction (AEG). Although the lncRNA NORAD has been implicated in chemoresistance, its specific role, underlying mechanism, and clinical significance in AEG remain unclear. We hypothesize that exosome-derived NORAD promotes oxaliplatin resistance via a competing endogenous RNA network that modulates autophagy and may serve as a potential circulating biomarker.

Methods

The expression levels of NORAD and miR-433-3p were analyzed in 56 paired AEG tumor and adjacent normal tissues, as well as in AEG cell lines (OE19, PDC) and their corresponding oxaliplatin-resistant derivatives (OE19-R, PDC-R), using quantitative real-time polymerase chain reaction. The direct interaction between NORAD and miR-433-3p was validated by dual-luciferase reporter assay. Functional consequences of NORAD knockdown were assessed: cell viability was determined by CCK-8 assay to calculate the half-maximal inhibitory concentration (IC₅₀); autophagic flux was evaluated by western blotting for the LC3B-Ⅱ/LC3B-I ratio and p62 degradation; and apoptosis was quantified by flow cytometry. Serum exosomal NORAD levels were measured in AEG patients and healthy controls.

Results

In AEG tissues, NORAD was significantly upregulated, whereas miR-433-3p was downregulated, and a strong inverse correlation was observed between them (r = −0.864, p < 0.001). NORAD directly bound to miR-433-3p, acting as a molecular sponge. Oxaliplatin-resistant cells exhibited elevated NORAD levels, enhanced autophagic flux—characterized by an increased LC3B-Ⅱ/I ratio and decreased p62 levels—and a higher IC₅₀ value. Silencing NORAD resensitized the resistant cells to oxaliplatin (e.g., IC₅₀ in PDC-R decreased from 22.27 to 2.51 µg/mL), inhibited autophagy, and promoted apoptosis. Clinically, serum exosomal NORAD levels were significantly elevated in AEG patients, and a strong positive correlation was observed between serum exosomal NORAD levels and tumor NORAD expression (r = 0.886, p < 0.001).

Conclusion

Our study identifies a novel NORAD/miR-433-3p/autophagy axis driving oxaliplatin resistance in AEG. Targeting NORAD restores chemosensitivity, highlighting its therapeutic potential. Moreover, circulating exosomal NORAD serves as a promising non-invasive biomarker, supporting its application in liquid biopsy for AEG management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12935-026-04340-0.

Introduction

Adenocarcinoma of the esophagogastric junction (AEG), a clinically aggressive malignancy, exhibits distinct molecular and clinicopathological characteristics compared with distal gastric cancers [1]. Over the past few decades, its incidence has risen globally [2]. According to data from the World Health Organization (WHO), the global age-standardized incidence rate of AEG is approximately 1.5 per 100,000, with a corresponding mortality rate of about 1.2 per 100,000 [3]. The burden of AEG varies significantly across regions, reflecting disparities in risk factors, screening practices, and access to treatment advances [4]. East Asia is a high-burden region: in 2021, age-standardized incidence rates were reported as 29.05 per 100,000 in China, 25.54 per 100,000 in Japan, and 25.76 per 100,000 in South Korea [5]. In contrast, Australia reported a considerably lower rate of 3.3 per 100,000 [6]. The prognosis of AEG is influenced by multiple factors, including tumor subclassification (e.g., Siewert type), pathological stage, treatment response, and the systemic immune-inflammation index (SⅡ). Survival rates differ notably among Siewert classifications; for example, the 5-year overall survival rate is 54.5% for Siewert type I and 64.8% for Siewert type Ⅱ [7]. Achieving a pathological complete response after neoadjuvant therapy is strongly associated with improved overall survival and disease-free survival [8]. Research by Pinar Peker et al. has clearly demonstrated that an elevated SII serves as an independent prognostic factor for unfavorable outcomes—both in general and specifically following perioperative chemotherapy [9]. For patients with locally advanced AEG, the combination of perioperative chemotherapy and surgical resection remains the standard treatment, conferring a substantial survival advantage compared with surgery alone [10, 11]. Oxaliplatin-based regimens play a central role in these therapeutic strategies [12]. However, the development of intrinsic and acquired resistance to oxaliplatin represents a primary cause of treatment failure and disease progression, posing a major challenge to improving long-term patient outcomes [13].

A key mechanism through which cancer cells evade chemotherapy-induced cell death is the activation of pro-survival pathways, such as autophagy. This evolutionarily conserved cellular process maintains cellular homeostasis by degrading damaged organelles and misfolded proteins. A growing body of evidence suggests that excessive autophagy activation contributes to chemoresistance in various malignancies, including gastric [14], hepatocellular [15], and colorectal cancers [16]. In the context of cancer therapy, elevated autophagic activity enables tumor cells to mitigate oxidative stress and repair DNA damage induced by chemotherapeutic agents such as oxaliplatin, thereby enhancing cell survival and promoting therapeutic resistance [17]. Accurate assessment of autophagy levels often relies on key molecular markers. Among these, microtubule-associated protein 1 light chain 3β (LC3B) is widely regarded as the canonical marker for autophagy. LC3B is initially synthesized in the cytoplasm as the precursor form, LC3B-I. Subsequently, after cleavage and lipidation by the Atg4 protease, it conjugates with phosphatidylethanolamine to form the membrane-bound LC3B-Ⅱ. LC3B-Ⅱ is specifically localized to the autophagosome membrane and directly participates in autophagosome formation and maturation. Therefore, the LC3B-Ⅱ/LC3B-I ratio is widely used to quantitatively assess autophagic activity [18, 19]: an increase in this ratio generally reflects enhanced autophagic flux and increased autophagosome number, whereas a decrease indicates autophagy inhibition. The molecular regulation of such complex processes—autophagy and chemoresistance—is increasingly linked to long non-coding RNAs (lncRNAs) [20]. These RNA transcripts, typically longer than 200 nucleotides and possessing limited protein-coding capacity, act as essential epigenetic regulators. A well-established regulatory mechanism is their function as competitive endogenous RNAs (ceRNAs)—often termed “molecular sponges”—which sequester microRNAs (miRNAs) and thereby inhibit miRNA-mediated suppression of target messenger RNAs [21]. This ceRNA network constitutes a crucial layer of post-transcriptional regulation in cancer pathogenesis and treatment response. Notably, AEG exhibits distinct lncRNA expression profiles compared with distal gastric cancer [22]. Previous research has shown that upregulation of LINC00626 promotes metastasis in AEG via activation of the JAK1/STAT3/KHSRP signaling pathway. In contrast, LINC00626 is expressed at relatively low levels in distal gastric cancer and does not undergo significant upregulation [23].

Among the multitude of oncogenic lncRNAs, the lncRNA activated by DNA damage (NORAD, also known as LINC00657) has emerged as a crucial regulator of tumorigenesis and therapeutic resistance [24]. NORAD is persistently upregulated in multiple cancer types and drives tumor progression, metastasis, and therapy resistance through diverse molecular mechanisms [25]. Its functional roles are highly context-dependent and are typically mediated by acting as a ceRNA that sequesters distinct miRNAs in different malignancies. For instance, in gastric cancer, NORAD enhances oxaliplatin resistance by promoting autophagic flux via sponging miR-433-3p [26]. Notably, NORAD contributes to chemotherapy resistance in a wide range of cancers—including non-small cell lung cancer [25], neuroblastoma [27], osteosarcoma [28], bladder cancer [29], colorectal cancer [30], and esophageal squamous cell carcinoma [31]—through diverse molecular pathways. Xuefei Zhang et al. demonstrated that NORAD inhibits ferroptosis by modulating the miR-144-3p–mTOR–ferritinophagy axis, thereby increasing tumor cell resistance to chemotherapeutic agents such as cisplatin, doxorubicin, and paclitaxel [32]. In contrast, Yongjun Hu et al. revealed that the NORAD/miR-378c regulatory axis suppresses malignant phenotypes in gastric adenocarcinoma both in vitro and in vivo [33]. Furthermore, quantitative proteomic analyses using Tandem Mass Tag labeling have indicated that NORAD regulates autophagic flux in gastric cancer, underscoring its role in cellular stress responses [26]. Importantly, a pivotal study confirmed that chemotherapeutic agent–induced oxidative stress triggers NORAD expression via histone modification—specifically, H3K27 acetylation (H3K27ac) [26]. The subsequent upregulation of NORAD enhances autophagic flux by acting as a ceRNA for miR-433-3p, ultimately conferring oxaliplatin resistance to gastric cancer cells.

Exosomes, extracellular vesicles with a size range of 30–150 nm, have emerged as crucial mediators of intercellular communication by transferring bioactive molecules—including proteins, mRNAs, and lncRNAs [34]. Their lipid-bilayer membrane protects encapsulated cargo from RNase-mediated degradation, conferring high stability in circulation and making them ideal candidates for liquid biopsy applications [35]. Tumor-derived exosomes can horizontally deliver oncogenic lncRNAs to recipient cells, thereby promoting the spread of drug-resistant phenotypes. The use of serum-derived exosomal lncRNAs as non-invasive biomarkers is increasingly supported by evidence from multiple cancer types. A growing body of research indicates that exosomal lncRNAs play a significant role in chemotherapy resistance development through intercellular transfer [36]. For example, specific exosomal lncRNA signatures in serum demonstrate high diagnostic and prognostic accuracy in gastric cancer [35] and colorectal cancer [37].

As previously noted, AEG demonstrates a distinct lncRNA expression profile compared with distal gastric cancers, suggesting divergent regulatory mechanisms underlying its aggressive phenotype and differential treatment response [22]. Although emerging evidence implicates NORAD in the pathogenesis of gastric cancer, its expression pattern, functional significance, and clinical relevance in AEG remain poorly characterized. Notably, the role of exosomal NORAD in mediating oxaliplatin resistance in AEG has yet to be established. Based on bioinformatic predictions of a binding site between NORAD and miR-433-3p—and supported by existing evidence linking NORAD to autophagy regulation—we hypothesize that NORAD contributes to oxaliplatin resistance in AEG by acting as a ceRNA for miR-433-3p. This sequestration derepresses downstream targets and promotes a pro-survival autophagic response. Furthermore, we propose that serum-derived exosomal NORAD may serve as a novel, non-invasive biomarker for AEG. This study aims to systematically: (1) validate the differential expression of NORAD and miR-433-3p in clinical AEG samples and cell lines; (2) confirm the direct molecular interaction between NORAD and miR-433-3p; (3) investigate the functional impact of NORAD knockdown on oxaliplatin sensitivity and autophagic flux; and (4) evaluate the correlation between exosomal NORAD levels in serum and NORAD expression in matched primary AEG tumors. By elucidating the NORAD/miR-433-3p/autophagy regulatory axis, this work seeks to provide mechanistic insights into chemoresistance in AEG and identify potential diagnostic and therapeutic targets for this clinically challenging malignancy.

Materials and methods

Patients and tissue samples

A total of 56 patients diagnosed with locally advanced AEG and 56 age- and sex-matched healthy controls were enrolled in this study. Paired tissue samples—namely, primary tumor tissue and adjacent normal tissue (ANT), defined as tissue collected from a site at least 5 cm away from the tumor margin—were obtained from each AEG patient during curative surgical resection at Anyang Tumor Hospital between January and September 2024. All tumor specimens were histopathologically confirmed as adenocarcinoma according to the WHO classification criteria. None of the participants had received neoadjuvant chemotherapy or radiotherapy prior to surgery. The study protocol was approved by the Ethics Committee of Anyang Tumor Hospital (approval number: AZLL0220240103008) and conducted in strict accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants before enrollment.

Blood collection and serum isolation

Peripheral blood (5 mL) was collected from each participant using EDTA-containing vacuum tubes (BD Biosciences, USA). The samples were promptly placed on ice and centrifuged within one hour at 1000 × g for 20 min at 4 °C to separate cellular components. The resultant plasma was carefully aliquoted and stored at −80 °C until analysis, preserving exosome integrity and RNA stability.

RNA extraction and lncRNA microarray

Total RNA was extracted from tissue samples using TRIzol® Reagent (Invitrogen, USA) according to the manufacturer’s instructions. RNA quality and concentration were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, USA) and the Agilent 2100 Bioanalyzer (Agilent Technologies, USA). LncRNA and mRNA expression profiles were generated using the ArrayStar Human LncRNA/mRNA Expression Microarray V3.0 (ArrayStar, USA), following the manufacturer’s instructions. Microarray annotations were derived from public databases including NCBI RefSeq, UCSC, RNAdb, and literature-validated lncRNAs [38]. Hybridization and initial data processing were performed by Kangchen Biotech (Shanghai, China). Differentially expressed transcripts were identified using thresholds of |fold change| ≥ 2.0 and an adjusted *p*-value < 0.05.

Cell lines and culture conditions

The human gastric mucosal epithelial cell line GES-1 was obtained from the American Type Culture Collection (ATCC, USA) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, USA) and 1% penicillin/streptomycin (HyClone, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. The OE19 cell line was established from primary tumor tissue isolated from a patient with Siewert type Ⅱ AEG. Histopathological analysis confirmed a moderately differentiated adenocarcinoma, corresponding to UICC stage ⅡI. The OE19 cell line was obtained from the Shanghai Cell Bank, Chinese Academy of Sciences (China), and cultured under identical conditions. Cells were passaged every 3–5 days using 0.25% trypsin–EDTA (Sigma-Aldrich, USA) and routinely tested for mycoplasma contamination.

Patient-derived cell (PDC) culture

The human gastric mucosal epithelial cell line GES-1 was obtained from the American Type Culture Collection (ATCC, USA) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, USA) and 1% penicillin/streptomycin (HyClone, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. The OE19 cell line was established from primary tumor tissue isolated from a patient with Siewert type Ⅱ AEG. Histopathological analysis confirmed a moderately differentiated adenocarcinoma, corresponding to UICC stage ⅡI. The OE19 cell line was obtained from the Shanghai Cell Bank, Chinese Academy of Sciences (China), and cultured under identical conditions. Cells were passaged every 3–5 days using 0.25% trypsin–EDTA (Sigma-Aldrich, USA) and routinely tested for mycoplasma contamination.

Generation of oxaliplatin-resistant cells

Oxaliplatin-resistant sublines (PDC-R and OE19-R) were established by stepwise exposure to incrementally increasing concentrations of oxaliplatin (2–10 µg/mL; Sigma-Aldrich, USA) over multiple passages spanning approximately six months. Resistance development was confirmed by sustained proliferative capacity in the presence of oxaliplatin and by calculation of resistance indices (RI). To preserve the resistant phenotype, all resistant sublines were routinely maintained in complete culture medium supplemented with 2 µg/mL oxaliplatin.

Bioinformatic analysis

Putative interactions between NORAD and miR-433-3p were predicted using StarBase v3.0 (http://starbase.sysu.edu.cn) and DIANA-LncBase v3.0 (http://diana.e-ce.uth.gr/lncbasev3) [39]. In StarBase v3.0, the miRNA–lncRNA module was employed with “hsa-miR-433-3p” specified as the miRNA and “NORAD” as the target lncRNA. To ensure high confidence, predictions were filtered under stringent criteria: only interactions supported by at least two pan-cancer CLIP-seq datasets for Argonaute (Ago) binding *and* corroborated by degradome-seq evidence of RISC-mediated cleavage were retained. In DIANA-LncBase v3.0, the “Predicted Targets” module was used to analyze the “NORAD” transcript and “hsa-miR-433-3p”, applying a microT-CDS score threshold > 0.8 for high-confidence predictions. Binding sites exhibiting cross-species conservation and those previously experimentally validated were prioritized. High-confidence putative binding sites were defined as those consistently identified by both databases under the specified filtering criteria, thereby establishing a robust set of candidate interactions for downstream experimental validation. The predicted binding sites were visualized using RNAhybrid (http://bibiserv.cebitec.uni-bielefeld.de/rnahybrid). Additionally, potential mRNA targets of miR-433-3p were identified through an integrated analysis leveraging miRDB (http://mirdb.org) and TargetScan (http://www.targetscan.org/vert_80/).

Dual-luciferase reporter assay

NORAD overexpression plasmids—containing either the wild-type NORAD binding site for miR-433-3p (NORAD-WT) or a mutated version thereof (NORAD-Mut)—and miR-433-3p overexpression plasmids were constructed using the GV272 vector (GeneChem, China) [26]. Cells were co-transfected with these reporter constructs and either a miR-433-3p mimic or a negative control (NC) using Lipofectamine 3000 (Invitrogen, USA). Luciferase activity was measured 48 h post-transfection using the Dual-Luciferase® Reporter Assay System (Promega, USA) on a GloMax® 20/20 Luminometer (Promega, USA). Firefly luciferase activity was normalized to Renilla luciferase activity for each sample.

Lentiviral transduction and knockdown

Lentiviral constructs encoding short hairpin RNA (shRNA) targeting NORAD, constructs for miR-433-3p overexpression, and their respective negative controls (NCs) were synthesized by GeneChem (Shanghai, China; Cat. #222,962). The shRNA sequences were designed using a modern deep neural network model that integrates guide RNA sequence features, target site contextual information, and processing efficiency predictions to ensure high silencing potency and minimize off-target effects. Cells were transduced at a multiplicity of infection (MOI) of 20 in the presence of Polybrene (8 µg/mL; Sigma-Aldrich, USA), and stable transductant pools were established by puromycin selection.

Cell viability and IC₅₀ assay

Cells were seeded into 96-well plates at a density of 3 × 10³ cells per well and treated with various concentrations of oxaliplatin (1–64 µg/mL) for 48 h. Cell viability was assessed using the CCK-8 assay (ApexBio, USA) according to the manufacturer’s instructions [29]. Absorbance was measured at 490 nm using a microplate reader (BioTek, USA). The half-maximal inhibitory concentration (IC₅₀) values were calculated by nonlinear regression analysis using GraphPad Prism version 9.0 (GraphPad Software, USA). The resistance index (RI) was defined as the ratio of the IC₅₀ values between the resistant and parental cell lines.

Quantitative real-time PCR (qRT-PCR)

Total RNA was reverse-transcribed using the PrimeScript RT Reagent Kit (Takara, Japan). For miRNA-specific analysis, the Mir-X miRNA First-Strand Synthesis Kit (Takara, Japan) was used separately. Quantitative PCR was performed on a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, USA) using SYBR Premix Ex Taq Ⅱ (Tli RNase H Plus) (Takara, Japan; Cat. # RR820A). The 20 µL qPCR reaction mixture consisted of 10 µL of 2 × SYBR Premix Ex Taq Ⅱ, 1 µL of each forward and reverse primer (10 µM), 2 µL of cDNA template, and nuclease-free water to bring the final volume to 20 µL. Thermal cycling conditions were as follows: initial denaturation at 95 °C for 3 min; 40 amplification cycles of 95 °C for 15 s (denaturation) and 58 °C for 15 s (annealing/extension); followed by melting curve analysis to confirm primer specificity. GAPDH and U6 snRNA were used as endogenous controls for NORAD and miR-433-3p, respectively. Relative gene expression levels were calculated using the 2^−ΔΔCt method. The primer sequences were as follows: NORAD forward, 5′-ACAGCCAAATCACAGCCACA-3′; reverse, 5′-TGGGCTCTGTGTTTGCATCC-3′; miR-433-3p forward, 5′-ATCACAUUGGGCUCCUCGGUGU-3′; reverse, universal primer provided in the Mir-X kit; GAPDH forward, 5′-GGAGCGAGATCCCTCCAAAAT-3′; reverse, 5′-GGCTGTTGTCATACTTCTCATGG-3′; U6 forward, 5′-CTCGCTTCGGCAGCACA-3′; reverse, 5′-AACGCTTCACGAATTTGCGT-3′.

Protein extraction and western blotting

GO pathway analysis demonstrated a potential role of NORAD in autophagosome assembly. The LC3B-Ⅱ/LC3B-I ratio is widely regarded as the most reliable indicator of autophagic flux in tumor cells [18, 19]. Moreover, p62 protein levels exhibit a negative correlation with autophagic flux activity. Total proteins were extracted using RIPA lysis buffer (Beyotime, China) supplemented with protease inhibitors (Roche, Switzerland). Protein concentrations in the lysates were determined using a bicinchoninic acid (BCA) protein assay kit (Sigma-Aldrich, USA). Equal amounts of protein (20–30 μg) were separated by 10% SDS-PAGE and transferred onto PVDF membranes (Millipore, USA). Membranes were blocked with 5% non-fat dry milk in TBST to minimize nonspecific binding. Primary antibodies against LC3B (1:1000; Novus Biologicals, USA, Cat. #HZ-R4843), p62 (1:800; Santa Cruz Biotechnology, USA, Cat. #Hz-R55207), and GAPDH (1:5000; Proteintech, USA, Cat. #WK352) were incubated overnight at 4 °C. After primary antibody incubation, membranes were incubated with HRP-conjugated secondary antibodies (1:5000; Cell Signaling Technology, USA, Cat. #C31430100), and immunoreactive bands were visualized using an ECL detection reagent (Bio-Rad, USA). Band intensities were analyzed and quantified using ImageJ software (NIH, USA).

Flow cytometry

Flow cytometry analysis was performed on PDC-R cells and shNORAD PDC-R cells 48 h after oxaliplatin treatment (5 μg/mL) to assess apoptosis. Cells (4 × 10⁵) were washed twice with PBS and resuspended in 100 μL of 1 × binding buffer containing 5 μL of Annexin V–FITC (Transgen Biotech, Guangzhou, China). After incubation for 15 min at room temperature in the dark, 5 μL of propidium iodide (PI) was added to each sample. Flow cytometric analysis was performed immediately using a Beckman CytoFLEX flow cytometer (Brea, CA, USA). Data were analyzed using CytExpert 2.0 software (Beckman Coulter, Brea, CA, USA).

Exosome isolation and characterization

Serum exosomes were isolated using the exoRNeasy Serum/Plasma Midi Kit (Qiagen, Germany) according to the manufacturer’s instructions. The size distribution and concentration of exosomes were assessed by nanoparticle tracking analysis (NTA; Malvern NanoSight LM10), and their morphology was evaluated by transmission electron microscopy (TEM; Philips Tecnai 12). Exosomal purity was confirmed by Western blotting detection of canonical markers CD63, CD81, and TSG101. Total RNA was extracted from the isolated exosomes using the same kit and subsequently reverse-transcribed using the miScript Ⅱ RT Kit (Qiagen) for downstream analysis [40].

Statistical analysis

Data are presented as mean ± standard deviation (SD). Two-group comparisons were performed using two-tailed Student’s *t*-tests, whereas one-way analysis of variance (ANOVA) with Tukey’s post hoc test was used for multiple-group comparisons. Pearson’s correlation coefficient was calculated to assess correlations. Statistical analyses were conducted using SPSS version 22.0 (IBM, USA) and GraphPad Prism version 9.0. A *p*-value < 0.05 was considered statistically significant. All experiments were independently repeated at least three times.

Results

Identification of differentially expressed lncRNAs in AEG

We analyzed lncRNA expression profiles in primary tumor tissues and paired adjacent non-tumor tissues (ANTs) from four treatment-naïve patients with locally advanced AEG using the ArrayStar Human LncRNA/mRNA Microarray V3.0. Among the 503,687 lncRNAs profiled, several differentially expressed transcripts were identified (Fig. 1A). The inclusion criteria for this study required histologically confirmed moderately or poorly differentiated adenocarcinoma. Samples predominantly consisted of these subtypes, with minimal representation of other pathological types—including well-differentiated adenocarcinoma, signet-ring cell carcinoma, mucinous adenocarcinoma, neuroendocrine carcinoma, undifferentiated carcinoma, and squamous cell carcinoma. Notably, the lncRNA NORAD was significantly upregulated in AEG tissues compared with ANTs (fold change ≥ 2.0, p < 0.05; Fig. 1B). This finding was further validated by qRT-PCR in an independent cohort of 56 AEG patients, which consistently demonstrated NORAD overexpression in tumor tissues (t = 21.213, p < 0.001; Fig. 1C).

Fig. 1.

Identification and validation of differentially expressed long non-coding RNAs (lncRNAs) in AEG. A Volcano plot illustrating the differential expression of lncRNAs between AEG tumor tissues and ANTs, based on a threshold of |fold change|≥ 2.0 and an adjusted p-value < 0.05. Significantly upregulated lncRNAs are indicated by red dots, downregulated lncRNAs by blue dots, and non-significant transcripts by gray dots. B Heatmap displaying the expression profiles of differentially expressed lncRNAs from the same analysis, with NORAD highlighted as one of the most significantly upregulated lncRNAs. Upregulated expression is shown in red and downregulated expression in blue. Expression levels are presented as Z-scores. C Validation of NORAD expression using quantitative qRT-PCR in a cohort of 56 AEG patient samples. NORAD expression was significantly elevated in tumor tissues compared to matched ANTs (t = 21.213, p < 0.001). Data are presented as mean ± standard deviation (SD)

NORAD and miR-433-3p are differentially expressed in AEG tissues and show inverse correlation

Bioinformatic analysis using StarBase V3.0 and DIANA-LncBase V3.0 predicted miR-433-3p as a direct target of the lncRNA NORAD, suggesting a potential ceRNA regulatory mechanism. qRT-PCR validation in tumor tissues from 56 AEG patients revealed that NORAD was significantly upregulated (t = 21.213, p < 0.001), whereas miR-433-3p was markedly downregulated (t = 19.463, p < 0.001) relative to ANTs (Table 1). Furthermore, a strong inverse correlation was observed between NORAD and miR-433-3p expression levels in AEG tissues (r = −0.864, p < 0.001; Fig. 2).

Table 1.

Expression levels of NORAD and miR-433-3p in tumor tissues compared to adjacent non-tumor tissues from 56 patients with AEG

Group	NORAD (mean ± SD)	P-value	miR-433-3p (mean ± SD)	P-value
AEG	2.245 ± 0.376	< 0.001	0.444 ± 0.143	< 0.001
ANT	1.005 ± 0.265		1.039 ± 0.161

Gene expression levels of NORAD and miR-433-3p were determined by quantitative RT-PCR in paired tumor (AEG) and adjacent non-tumor (ANT) tissues from 56 patients. Statistical significance was assessed using a paired t-test. NORAD was significantly upregulated (t = 21.213, p < 0.001), while miR-433-3p was significantly downregulated (t = 19.463, p < 0.001) in AEG tissues compared to ANT tissues

Fig. 2.

NORAD is upregulated and exhibits an inverse correlation with miR-433-3p in AEG tissues. qRT-PCR was used to analyze the expression levels of NORAD and miR-433-3p in tumor tissues relative to ANTs from 56 AEG patients. A scatter plot illustrates a significant inverse correlation between NORAD and miR-433-3p expression levels in this cohort (Pearson correlation coefficient r = −0.864, p < 0.001)

NORAD directly binds to miR-433-3p

Dual-luciferase reporter assays further confirmed that miR-433-3p directly targets the NORAD 3′-UTR. As shown by a significantly reduced luciferase activity in PDCs and PDC-Rs co-transfected with the wild-type NORAD 3′-UTR reporter construct and miR-433-3p mimics (approximately 60% reduction compared to the control, p < 0.001; Fig. 3), this interaction is functionally significant. Mutation of the predicted miR-433-3p binding site within the NORAD 3′-UTR completely restored luciferase activity, confirming the specificity of the interaction.

Fig. 3.

NORAD is a direct target of miR-433-3p, as confirmed by dual-luciferase reporter assay. A, B Luciferase activity was measured in oxaliplatin-sensitive patient-derived cells (PDCs) (A) and oxaliplatin-resistant PDCs (PDC-Rs) (B) after co-transfection with either a wild-type (WT) NORAD reporter construct or a mutant (MUT) construct, along with miR-433-3p mimics or a negative control (NC) mimic. Renilla luciferase activity served as the internal control for normalization. Data are expressed as mean ± SD from at least three independent experiments. Statistical significance was assessed using Student’s t-test (p < 0.001)

Expression of NORAD and miR-433-3p in AEG cell lines

Oxaliplatin-resistant sublines (PDC-R and OE19-R) were established from their parental PDC and OE19 cell lines through prolonged, stepwise exposure to oxaliplatin. NORAD-knockdown cells (shNORAD) were generated via lentiviral transduction, achieving knockdown efficiency exceeding 80% in all cell types (p < 0.001). qRT-PCR analysis revealed that NORAD expression was significantly upregulated in both PDC and PDC-R cells compared with that in the non-tumoral GES-1 cell line (p < 0.001 for both), with a further significant increase in PDC-R relative to PDC cells (t = 19.432, p < 0.001; Fig. 4A). In contrast, miR-433-3p expression was markedly downregulated in all AEG-derived cell lines compared with that in GES-1 (F = 147.567, p < 0.001). NORAD knockdown significantly increased miR-433-3p levels in both PDC and PDC-R cells (Fig. 4A). Similar trends were observed in the OE19-based cell line model (Fig. 4B).

Fig. 4.

NORAD knockdown upregulates miR-433-3p expression in AEG cell lines, including oxaliplatin-resistant models. A, B The relative expression levels of NORAD and miR-433-3p were assessed by qRT-PCR in A patient derived cell lines (PDC, PDC-R) and B OE19 and OE19-R cell lines following transfection with NORAD-specific shRNA (shNORAD) or a non-targeting control (shNC). The human gastric epithelial cell line GES-1 was included as a non-malignant control. Data are representative of three independent experiments and are presented as mean ± SD. Statistical significance was evaluated using one-way ANOVA (p < 0.001)

NORAD knockdown sensitizes AEG cells to oxaliplatin

CCK-8 assays revealed that oxaliplatin-resistant cell lines (PDC-R and OE19-R) exhibited significantly higher IC₅₀ values than their parental counterparts (PDC-R: 22.27 μg/mL vs. PDC: 14.28 μg/mL, resistance index = 1.559; OE19-R: 25.43 μg/mL vs. OE19: 15.16 μg/mL, resistance index = 1.677). Knockdown of NORAD markedly reduced oxaliplatin resistance, decreasing the IC₅₀ to 2.00 μg/mL in shNORAD-transduced PDC cells and to 2.51 μg/mL in shNORAD-transduced PDC-R cells, yielding resistance indices of 0.140 and 0.113, respectively (Fig. 5A). A comparable resensitization effect was observed in OE19-derived cell lines (Fig. 5B).

Fig. 5.

Knockdown of NORAD restores oxaliplatin sensitivity in AEG cell lines. A Oxaliplatin sensitivity was evaluated in PDCs and PDC-R following transduction with shNORAD using CCK-8 assays. NORAD knockdown significantly decreased the half-maximal inhibitory concentration (IC₅₀), thereby reducing the resistance index (RI). B A similar resensitization effect was observed in OE19 and OE19-R cell lines upon NORAD knockdown. Data are presented as mean ± SD from three independent experiments. Statistical significance was assessed by two-way ANOVA (p < 0.001). Resistance indices were calculated as the ratio of IC₅₀ values in resistant cells to those in their corresponding parental sensitive cells

NORAD regulates autophagic activity and apoptosis in AEG cells

To investigate the role of autophagy in acquired oxaliplatin resistance, we first compared the expression of key autophagy markers between oxaliplatin-resistant (PDC-R and OE19-R) cells and their parental, oxaliplatin-sensitive (PDC and OE19) counterparts. Western blot analysis revealed that resistant cells exhibited a significantly increased LC3B-Ⅱ/LC3B-I ratio and a concomitant decrease in p62 levels—findings indicative of enhanced autophagic flux (Fig. 6). This established an autophagy-prone phenotype in the resistant cell lines. Next, we examined the functional role of the lncRNA NORAD in this context. Notably, NORAD knockdown in resistant cells reversed the autophagic phenotype, as evidenced by a marked reduction in the LC3B-Ⅱ/LC3B-I ratio and restoration of p62 protein levels (Fig. 6A, B). These results suggest that NORAD acts as a critical upstream regulator sustaining elevated autophagic activity in oxaliplatin-resistant AEG cells. To evaluate the functional consequences of NORAD depletion, we measured apoptosis in PDC-R cells with or without NORAD knockdown following oxaliplatin treatment. Although oxaliplatin induced apoptosis in both groups, cell death was significantly more pronounced in NORAD-silenced cells (17.80 ± 3.08% vs. 11.77 ± 1.99%, p < 0.05) (Fig. 7A, B). This suggests that NORAD suppression not only inhibits pro-survival autophagy but also resensitizes resistant cells to oxaliplatin-induced apoptosis. To determine whether miR-433-3p mediates these effects, we transfected miR-433-3p mimics into PDC-R cells—either alone or in combination with NORAD knockdown. Transfection of the mimics alone significantly increased the LC3B-Ⅱ/LC3B-I ratio and significantly decreased p62 levels compared with control PDC-R cells, confirming that miR-433-3p overexpression enhances autophagic flux (Fig. 6C). Importantly, in cells overexpressing miR-433-3p, NORAD knockdown conferred no additional significant effect on autophagic markers, indicating that miR-433-3p acts downstream of NORAD. Similarly, after oxaliplatin treatment, PDC-R cells transfected with miR-433-3p mimics exhibited a significantly higher apoptotic rate than non-transfected controls (16.42 ± 2.27% vs. 11.77 ± 1.99%, p < 0.05) (Fig. 7B, D). Notably, NORAD knockdown did not significantly alter the apoptosis level in the context of miR-433-3p overexpression (17.11 ± 2.28% vs. 16.42 ± 2.27%, p = 0.141) (Fig. 7C, D).

Fig. 6.

NORAD knockdown and miR-433-3p mimic attenuate autophagic activity in AEG cells. A, B Western blot analysis of the autophagy markers LC3B and p62 in A PDC, PDC-R and B OE19 and OE19-R cell lines following NORAD knockdown. Resistant cell lines exhibited increased LC3B-II conversion and reduced p62 levels, indicative of elevated autophagic flux, both of which were reversed upon shNORAD transfection. C Compared with the PDC-R cell line, the content of LC3B-II was significantly increased in cells transfected with miR-433-3p mimics, whereas the level of p62 decreased. Whether NORAD was knocked down or not had a relatively minor impact on the autophagic flux of PDC-R cells transfected with miR-433-3p mimics. GAPDH served as a loading control. Representative blots are from three independent experiments. Densitometric quantification of protein levels relative to GAPDH is shown as mean ± SD. These findings indicate that NORAD plays a role in sustaining heightened autophagic flux in oxaliplatin-resistant AEG cells

Fig. 7.

Knockdown of NORAD and overexpression of miR-433-3p mimic enhance oxaliplatin-induced apoptosis in chemoresistant AEG cells. Apoptosis was evaluated in oxaliplatin-resistant PDC-R cells transduced with shNORAD or a non-targeting control after a 48-h treatment with oxaliplatin (5 µg/mL) using Annexin V/PI staining and flow cytometry. Simultaneously, we also examined the apoptosis level of PDC-R cells transfected with miR-433-3p mimics. A PDC-R cells transduced with shNORAD. B PDC-R cells with non-targeting control. The proportion of apoptotic cells was significantly increased in the NORAD knockdown group compared to the control group. C, D Flow cytometry analysis results showed that, in contrast to PDC-R cells, the apoptosis level of PDC-R cells transfected with miR-433-3p mimics and treated with oxaliplatin was higher, and the differences were statistically significant. However, whether NORAD was knocked down or not had a relatively minor impact on the apoptosis level of PDC-R cells transfected with miR-433-3p mimics. Data are representative of three independent experiments and presented as mean ± SD. Statistical significance was determined by Student’s t-test (p < 0.01). These results suggest that the suppression of NORAD renders chemoresistant AEG cell more sensitive to oxaliplatin-induced apoptosis

Serum exosomal NORAD is elevated in AEG and correlates with tumor expression

Serum exosomes were isolated from patients with AEG and healthy volunteers, and characterized by NTA and TEM. The exosomes exhibited the typical cup-shaped morphology, with diameters ranging from 50 to 150 nm. Western blot analysis confirmed the enrichment of exosomal markers CD63, CD81, and TSG101, as well as the absence of the negative marker calnexin. Notably, exosomal NORAD levels were significantly higher in AEG patients than in healthy controls (t = 35.646, p < 0.001; Fig. 8A). Moreover, a strong positive correlation was observed between exosomal NORAD levels and NORAD expression in matched tumor tissues (r = 0.8858, p < 0.001; Fig. 8B).

Fig. 8.

Exosomal NORAD is elevated in the serum of AEG patients and correlates with tumor tissue expression. A Quantitative analysis of exosomal NORAD levels in serum samples from patients with AEG (n = 56) and healthy volunteers (n = 30), as determined by qRT-PCR. NORAD levels were significantly higher in AEG patients (t = 35.646, p < 0.001). Data are presented as mean ± SD. B Correlation between exosomal NORAD levels in serum and NORAD expression in matched AEG tumor tissues, assessed using Pearson correlation analysis (r = 0.886, p < 0.001), indicating a strong positive association

Discussion

This study precisely delineates a novel and specific mechanism underlying oxaliplatin resistance in advanced AEG, with a primary focus on the lncRNA NORAD. We demonstrate that NORAD is significantly upregulated in AEG tumor tissues and, more notably, in circulating serum-derived exosomes. Functionally, NORAD promotes a chemoresistant phenotype by acting as a molecular sponge for miR-433-3p, thereby derepressing pro-survival autophagy and inhibiting apoptosis. The strong correlation between tumor-derived and exosome-associated NORAD levels not only underscores its potential as a non-invasive biomarker but also highlights the emerging role of exosomal lncRNA transfer in mediating intercellular communication and therapy resistance [41].

The molecular landscape of AEG clearly diverges from that of distal gastric cancers [1, 22]. Our findings add a disease-specific dimension to the oncogenic function of NORAD. In other gastric cancer contexts, NORAD facilitates tumor progression via distinct miRNAs—such as miR-608 [42]. Nevertheless, we demonstrate that in AEG, its oncogenic activity is specifically mediated by sequestering miR-433-3p. This functional disparity underscores AEG as a distinct molecular entity and highlights the need for customized therapeutic strategies. By integrating bioinformatic prediction with subsequent experimental validation—including the strong inverse correlation observed in clinical samples (r = −0.864, p < 0.001) and direct binding confirmed by dual-luciferase reporter assay—we robustly substantiate a ceRNA mechanism wherein NORAD sequesters miR-433-3p to alleviate its inhibitory effects on downstream survival pathways.

Our functional experiments provide compelling evidence that upregulation of NORAD serves as a hallmark of acquired oxaliplatin resistance. The establishment of resistant sublines (PDC-R and OE19-R) was accompanied by a substantial increase in NORAD expression and a shift toward an autophagy-prone state, as evidenced by an elevated LC3B-Ⅱ/LC3B-I ratio and decreased p62 protein levels (Fig. 6). Notably, genetic silencing of NORAD reversed this phenotype, restored miR-433-3p expression, and markedly resensitized resistant cells to oxaliplatin—demonstrated by significantly reduced IC₅₀ values (e.g., IC₅₀ for PDC-R decreased from 22.27 µg/mL to 2.51 µg/mL). Mechanistically, this resensitization was associated with a switch from pro-survival autophagy to apoptotic cell death upon oxaliplatin treatment, as confirmed by flow cytometry (Fig. 7). These findings establish NORAD as a critical upstream regulator of a key resistance mechanism—consistent with the well-established roles of autophagy dysregulation and apoptosis evasion in chemotherapy failure across malignancies [14, 43].

The translational implications of our findings are twofold. First, the significant elevation of NORAD in serum exosomes from AEG patients—coupled with its strong positive correlation with tumor expression (r = 0.8858, p < 0.001; Fig. 8)—identifies exosomal NORAD as a promising liquid biopsy biomarker for noninvasive diagnosis and dynamic monitoring of therapeutic response. Second, our data strongly support targeting NORAD therapeutically. The substantial decrease in oxaliplatin IC₅₀ following NORAD knockdown indicates that pharmacological inhibition of NORAD could effectively re-sensitize tumors to chemotherapy.

Emerging RNA-targeting modalities—such as antisense oligonucleotides and nanoparticle-delivered siRNA—offer clinically viable strategies for the selective neutralization of these oncogenic lncRNAs. Future prognostic models for AEG can be significantly refined by integrating molecular drivers (e.g., exosomal NORAD) with well-established systemic inflammatory indices, such as the SII and the neutrophil-to-lymphocyte ratio [9]. This multi-parametric approach will enable more precise risk stratification and may identify patients who could benefit from intensified or alternative treatment regimens, including immunotherapy. Although this study focused on a specific lncRNA-mediated axis, the broader tumor microenvironment—including the prevalence of non-activated T-cell phenotypes in AEG [44]—undoubtedly contributes to therapeutic resistance and should be incorporated into future combinatorial treatment strategies.

Limitations

While this study provides robust proof-of-concept evidence, several limitations should be acknowledged. First, although the sample size was adequate for this initial exploration, the cohort was recruited from a single institution; validation in larger, independent, multicenter populations is essential to confirm the generalizability of our findings regarding NORAD’s expression and biomarker potential. Second, although we identified the NORAD/miR-433-3p/autophagy axis as a key pathway, NORAD is a multifunctional lncRNA that may interact with other miRNAs or signaling networks not examined in this study—interactions that could contribute to the resistant phenotype. Third, the molecular mechanisms governing the selective loading and secretion of NORAD into exosomes remain entirely unexplored and warrant dedicated future research. Fourth, the promising in vitro therapeutic potential of NORAD knockdown requires rigorous validation in in vivo models to assess efficacy, biodistribution, and potential off-target effects. Finally, all clinical samples and cellular models were derived from patients with locally advanced AEG; therefore, our findings may not be directly applicable to early-stage or metastatic disease, highlighting the need for future studies encompassing the full spectrum of AEG progression.

Conclusion

In summary, this study identifies the lncRNA NORAD as a critical determinant of oxaliplatin resistance in AEG. We elucidate a novel molecular pathway in which NORAD—frequently upregulated in tumors and detectable in serum-derived exosomes—functions as a ceRNA to sequester miR-433-3p. This interaction promotes protective autophagy and suppresses apoptosis, ultimately enabling tumor cell survival under chemotherapy-induced stress. The substantial chemosensitization observed upon NORAD knockdown underscores its potential as a promising therapeutic target. Moreover, the elevated and positively correlated levels of NORAD in tumor tissue and serum-derived exosomes highlight its dual role as both a key mechanistic driver and a potential non-invasive biomarker. Our findings suggest that targeting the NORAD/miR-433-3p/autophagy axis represents a viable strategy to overcome chemoresistance in AEG. Future efforts should focus on developing clinical-grade NORAD inhibitors and prospectively validating the diagnostic, prognostic, and predictive value of circulating exosomal NORAD to improve the management of this aggressive malignancy.

Author contributions

The final manuscript has undergone thorough review and approval by all authors, who disclose no conflicts of interest. The authors unanimously consent to the sharing of data and materials utilized in this study, as well as granting permission for publication of this research.

Funding

This research was funded by Henan Province Science and Technology Tackling Key Project (232102310090) and (No. 242102310125).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study was performed in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the Anyang Tumor Hospital. The patients informed consent was obtained according to regional regulations.

Institutional review board statement

Informed written consent was obtained from all patients. The present study was approved by the Ethics Committee of the Anyang Tumor Hospital (Anyang, China).

Competing interests

The authors declare no competing interests.