Arteannuin B Inhibits NSCLC Cells via Regulating miR-194-3p/CLDN2 Axis.

1
Introduction
Lung cancer remains the most lethal malignancy globally, with non‐small cell lung cancer (NSCLC) accounting for approximately 85% of cases and being characterized by aggressive progression, intrinsic therapeutic resistance and early metastatic dissemination [1, 2]. NSCLC is histologically classified into adenocarcinoma, squamous cell carcinoma and large cell carcinoma, each driven by distinct molecular alterations such as EGFR mutations and ALK rearrangements [3]. Despite advancements in targeted therapies and immunotherapies, a significant proportion of patients acquire resistance, which ultimately leads to treatment failure and poor prognosis [4]. This persistent challenge highlights the imperative to elucidate biomarkers and mechanisms governing NSCLC progression and therapeutic resistance.
Artemisinin derivatives, low‐toxicity sesquiterpene lactones isolated from
Artemisia annua
, are widely recognized for their antimalarial efficacy and exhibit broad therapeutic potential in oncology [5, 6, 7]. Beyond their antiparasitic applications, these compounds selectively impair cancer cell viability through mechanisms such as reactive oxygen species (ROS)‐mediated cytotoxicity, angiogenesis suppression via VEGF signaling pathway inhibition, and reversing multidrug resistance through inhibition of drug efflux transporters [8, 9, 10]. Among these derivatives, arteannuin B (Art B) has demonstrated anti‐inflammatory and anticancer efficacy in preclinical models, including osteoporosis and neuroinflammatory disorders [11, 12]. However, its molecular targets and functional roles in NSCLC, particularly in overcoming chemoresistance, remain uncharacterized, hindering its clinical translation due to undefined molecular targets [13, 14].
The claudin family, first identified in 1998, comprises tetraspan transmembrane proteins encoded by genes essential for maintaining epithelial barrier integrity and cell polarity [15, 16]. Claudin‐2 (CLDN2), significantly upregulated in NSCLC compared to adjacent normal tissues, has emerged as a pivotal oncogenic driver [17, 18, 19]. CLDN2 promotes tumor progression through ligand‐independent activation of Epidermal Growth Factor Receptor (EGFR)/Extracellular Signal‐Regulated Kinase (ERK) signaling pathway, disruption of epithelial‐mesenchymal transition (EMT) dynamics via E‐cadherin downregulation, and enhancement of chemoresistance through Sp1‐mediated upregulation of multidrug resistance‐associated protein 2 (MRP2/ABCC2) [20, 21, 22, 23]. Notably, CLDN2 knockdown restores chemosensitivity by increasing intracellular accumulation of platinum‐based agents, directly linking its expression to therapeutic failure [24]. Given its central role in driving proliferation and, most importantly, chemoresistance, CLDN2 represents a compelling therapeutic target for overcoming treatment failure in NSCLC.
Aberrantly expressed microRNAs (miRNAs) are increasingly implicated in NSCLC tumorigenesis and therapy resistance, with emerging roles in regulating claudin family members [25]. While miRNAs are known to regulate claudin family members in different malignancies, their roles in CLDN2‐driven NSCLC progression and chemoresistance remain unexplored [26]. This study aims to (1) elucidate CLDN2's functional contributions to NSCLC progression, (2) identify upstream miRNA regulators of CLDN2 and (3) evaluate the efficacy of Art B in suppressing CLDN2‐mediated chemoresistance. We hypothesize that the natural compound Arteannuin B (Art B) exerts its antitumor effects by modulating a specific miRNA to inhibit CLDN2, thereby suppressing proliferation and reversing cisplatin resistance in NSCLC.

2
Materials and Methods
2.1
Bioinformatic Analysis of
CLDN2
Expression

CLDN2 expression profiles and associated clinical metadata (tumor/normal status, sex) for lung adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC) cohorts were extracted from UCSC Xena‐processed TCGA data sets. Normalized RNA‐seq by expectation–maximization (RSEM) values (log2‐transformed) for CLDN2 were stratified by tissue type (tumor vs. adjacent normal) and sex. Batch effects were mitigated using ComBat (sva R package, v3.48.0), and samples with low CLDN2 expression (median transcripts per million (TPM) < 1) were excluded. Differential expression analysis between tumor and normal tissues was performed with limma (v3.56.2), incorporating sex as a covariate.

2.2
Cell Culture
Human NSCLC cell lines A549, HCC827, H1299, and H460 were acquired from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were propagated in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and maintained under standard culture conditions (37°C, 5% CO2, humidified atmosphere).

2.3
Plasmids and Transfection

CLDN2‐targeting siRNA and scrambled control siRNA were sourced from Sangon Biotech (Shanghai, China). Hsa‐miR‐194‐3p mimics, mimic negative controls (NC), inhibitors, and inhibitor NCs were acquired from RiboBio (Guangzhou, China). Plasmid transfections were performed using Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer's guidelines.

2.4
MTT Assay
Following transfection, NSCLC cells were plated in 96‐well plates (1 × 103 cells/well) and allowed to adhere for 24 h in complete medium. DDP and Art B were diluted to target concentrations, applied to respective wells and incubated for 48 h. MTT solution (5%, 20 μL) was then introduced, followed by a 4‐h incubation. Formazan crystals were solubilized with dimethyl sulfoxide (DMSO; 200 μL/well) under light‐protected conditions, and plates were agitated for 10 min. Absorbance at 570 nm was measured using a microplate reader, and cell viability was quantified relative to untreated controls.

2.5
Colony Formation Assay
Transfected NSCLC cells were plated in 6‐well plates at a density of 1 × 103 cells/well and maintained in culture medium for 14 days. Colonies were fixed with 4% paraformaldehyde (PFA; 30 min) and stained with 0.5% crystal violet (30 min, room temperature).

2.6
Flow Cytometry
Apoptosis was assessed using an Annexin V/PI detection kit (4A Biotech, Beijing) according to the manufacturer's protocol. Cells were detached with 0.25% EDTA‐free trypsin, washed twice with ice‐cold PBS and resuspended in binding buffer prior to staining. For intracellular DDP accumulation, cells underwent identical trypsinization and PBS washing steps, followed by analysis of DDP‐FITC fluorescence intensity. All samples were processed on a FACSCanto II flow cytometer.

2.7
Western Blot Analysis
Cells were lysed 48 h post‐transfection using RIPA buffer (Beyotime). Protein lysates were resolved on 10% SDS‐PAGE gels and transferred to nitrocellulose membranes (Millipore). Membranes were probed overnight at 4°C with primary antibodies: rabbit anti‐CLDN2 (Abcam), rabbit anti‐CTR1 (Abcam), rabbit anti‐MRP2 (Abcam), and mouse anti‐β‐actin (Cell Signaling Technology). After TBST washes, membranes were incubated for 1 h at room temperature with HRP‐conjugated secondary antibodies: goat anti‐rabbit IgG for CLDN2, CTR1, and MRP2 and goat anti‐mouse IgG for β‐actin. Signal detection was performed using a chemiluminescent substrate (ECL; Beyotime) following three 10‐min TBST washes.

2.8
Quantitative Real‐Time PCR (qRT‐PCR)
Total RNA was isolated from NSCLC cells using TRIzol reagent (Sangon Biotech, Shanghai, China) and reverse‐transcribed into cDNA with a RevertAid First Strand cDNA Synthesis Kit (Yeasen Biotech, Shanghai, China). qRT‐PCR was performed using SYBR Green Master Mix on a QuantStudio 6 Flex system (Thermo Fisher Scientific). Primer sequences for miR‐194‐3p and U6 (RiboBio, Guangzhou, China) are provided in Table S1. Relative gene expression was calculated via the 2−ΔΔCt method, normalized to U6 (for miRNA) and GAPDH (for mRNA).

2.9
Tumor Xenograft Model
Male BALB/c nude mice (n = 32, 3–4 weeks old, specific pathogen‐free) were acquired from the Institute of Laboratory Animal Sciences (Chinese Academy of Medical Sciences) and acclimatized under controlled environmental conditions. A549 cells (2 × 106 cells in 100 μL PBS) were subcutaneously injected into the left flank. Tumor volume was monitored twice weekly using calipers and calculated as V = length × width2 × 0.5. Drug administration commenced when tumors reached 70 mm3. After 28 days, mice were euthanized, and tumors were excised, weighed, and processed for downstream analysis.

2.10
Dual‐Luciferase Reporter Assay
Putative miR‐194‐3p binding sites within the CLDN2 3′ UTR were predicted using miRDB and TargetScan algorithms. Wild‐type CLDN2 3′ UTR sequences were cloned into pmirGLO vectors (Promega), while scrambled sequences served as controls. 293 T cells were co‐transfected with pmirGLO‐CLDN2 3′ UTR (or control vector), miR‐194‐3p mimic, or mimic negative control (NC) using Lipofectamine 3000. Luciferase activity was measured 48 h post‐transfection using a dual‐luciferase assay system (Yeasen Biotech), with Renilla luciferase normalized to firefly luciferase for each sample.

2.11
Statistical Analysis
All statistical analyses were conducted in R (v4.4.2). Continuous data are expressed as mean ± SD (n = 3). Differences between two groups were evaluated using unpaired two‐tailed Student's t‐tests. For multi‐group comparisons, two‐way ANOVA followed by Tukey's post hoc test was applied. A significance threshold of p < 0.05 was used for all analyses.

2.12
Data Visualization and Statistical Annotation
Data visualization was implemented using the ggplot2 package (v3.4.2) for layered plot construction. Color palettes were optimized for accessibility and perceptual uniformity via viridis (v0.6.2). Axes scaling and labeling were standardized using scales (v1.2.1). Statistical annotations (e.g., significance brackets, adjusted p‐values) were generated with ggpubr (v0.6.0) and validated through non‐parametric testing workflows in rstatix (v0.7.2).

3
Results
3.1
Cross‐Omics Profiling Identifies
CLDN2
Suppression by Art B Attenuates NSCLC Progression
To delineate the molecular targets of Art B, we performed transcriptomic profiling in NSCLC cell lines (A549, HCC827). RNA‐seq analysis of Art B‐treated A549 cells revealed 985 downregulated and 614 upregulated genes (Figure 1A). HCC827 cells showed a concordant suppression trend with 799 downregulated genes, though the magnitude of change was smaller than in A549 cells. Intersection analysis of co‐downregulated genes identified CLDN2 as the most significantly suppressed target in A549 cells (Figure 1B), exceeding its attenuation in HCC827 and ranking it within the top 1% of shared pharmacodynamic responders.

CLDN2 suppression displayed cell line‐specific dynamics, with pronounced knockdown efficiency in A549 cells compared to attenuated yet consistent downregulation in HCC827. Although inter‐sample variability persisted in HCC827 cohorts, replicate‐level analyses confirmed directional suppression aligned with negative Z‐score distributions (Figure 1C).
In clinical validation, TCGA analysis of LUAD/LUSC tissues confirmed elevated CLDN2 expression in tumors compared to matched normal tissues (Figure 1D). Sex‐stratified analysis, informed by CLDN2's X‐chromosome localization, revealed higher baseline expression in female NSCLC patients (Figure 1E), consistent with its pro‐tumorigenic function.
Collectively, our data identify CLDN2 as a key pharmacodynamic target of Art B in NSCLC, underscoring its promise as a therapeutic target given its elevated expression in tumors, particularly in females.

3.2
CLDN2
Mediates Art B's Suppression of NSCLC Across Models
Western blot analysis revealed elevated CLDN2 protein expression in A549 and HCC827 cells compared to H1299 and H460 (Figure 2A). Through screening of siRNA constructs targeting CLDN2, siRNA‐3 demonstrated the highest knockdown efficiency, achieving near‐complete CLDN2 suppression in both A549 and HCC827 cells (Figure 2B). Based on this efficacy, CLDN2‐siRNA 3 (si‐CLDN2) was selected for subsequent experiments.
Art B treatment caused dose‐dependent CLDN2 suppression. At 25 μM, CLDN2 protein was nearly undetectable in A549 cells but retained residual expression in HCC827 cells, attributable to higher basal CLDN2 levels in the latter cell line (Figure 2C). Immunofluorescence imaging revealed marked reduction of CLDN2 expression in Art B‐treated cells, concurrent with decreased cancer cell density, both inversely correlated to escalating drug concentrations (Figure 2D). Transcriptional suppression paralleled protein downregulation, with CLDN2 mRNA levels decreasing significantly in response to escalating Art B doses (Figure 2E).
si‐CLDN2 attenuated Art B's anti‐proliferative effects. In A549 cells, Art B's inhibitory potency was significantly reduced in si‐CLDN2‐treated groups, reflected by a substantial increase in IC
50 (Figure 2F). Colony formation assays revealed partial restoration of clonogenic capacity in si‐CLDN2‐transfected HCC827 cells under Art B treatment (Figure 2G). These results establish CLDN2 as a key mediator of Art B's anti‐NSCLC activity.
Our functional studies establish CLDN2 as a central executor of Art B's action. Art B transcriptionally and translationally represses CLDN2 in a dose‐dependent manner, while genetic silencing of CLDN2 diminishes its therapeutic efficacy, confirming a direct mechanistic link between target suppression and anti‐tumor outcome.

3.3
CLDN2
Depletion Potentiates Cisplatin Sensitivity in NSCLC

CLDN2 depletion induced sustained suppression of NSCLC cell proliferation. Compared to controls, si‐CLDN2‐treated A549 and HCC827 cells exhibited progressive viability loss and near‐complete elimination of clonogenic capacity (Figure 3A,B). Morphological analysis revealed marked cellular shrinkage and nuclear fragmentation, consistent with impaired proliferative potential (Figure 3B).
The combination of CLDN2 knockdown with cisplatin (DDP) demonstrated synergistic cytotoxicity, as shown by dose‐dependent reductions in viability and a pronounced shift toward late‐stage apoptotic populations (Figure 3C,D). CLDN2 depletion thus not only suppressed proliferation but also enhanced DDP sensitivity in NSCLCs. To elucidate the molecular basis of this chemosensitization, we evaluated two critical determinants of platinum pharmacokinetics: multidrug resistance‐associated protein 2 (MRP2), a drug efflux pump, and copper transporter 1 (CTR1), the primary platinum influx channel.
Mechanistic studies showed selective downregulation of MRP2 in CLDN2‐deficient cells, with CTR1 expression remaining unaffected (Figure 3E). Flow cytometry analysis revealed prolonged intracellular DDP retention in si‐CLDN2‐treated cells (Figure 3F), with spatially resolved fluorescence microscopy further confirming perinuclear drug aggregation—a spatial signature of efflux pathway disruption (Figure 3G). Therefore, we hypothesize that CLDN2 depletion enhances cisplatin sensitivity in NSCLC by suppressing MRP2‐mediated drug efflux, leading to prolonged intracellular cisplatin retention and synergistic cytotoxicity through impaired proliferation and enhanced apoptosis.
Together, these results demonstrate that CLDN2 loss exerts a dual anti‐tumor effect: it intrinsically suppresses NSCLC growth and extrinsically primes cells for cisplatin response. Mechanistically, this chemosensitization is achieved by selectively inhibiting the drug efflux pump MRP2, thereby promoting intracellular cisplatin accumulation and synergistic apoptosis.

3.4
CLDN2
Silencing Synergizes With DDP to Suppress the Progression and Toxicity of NSCLC

CLDN2 silencing significantly inhibited tumor growth and synergistically enhanced DDP efficacy in NSCLC xenograft models. Tumors in the untreated model group exhibited progressive increases in volume and mass, whereas the si‐CLDN2 + DDP combination group showed maximal suppression (Figure 4A,B). Longitudinal monitoring delineated divergent tumor dynamics across cohorts, with untreated controls exhibiting sustained growth, si‐CLDN2 monotherapy inducing partial suppression, DDP monotherapy driving transient stabilization prior to relapse, and the combination cohort achieving sustained regression (Figure 4C).
Systemic toxicity profiles differed significantly across treatment groups. Body weight remained stable in both the Model group and si‐CLDN2 monotherapy group, whereas DDP monotherapy and combination therapy groups induced progressive weight loss (Figure 4D). Organ index analysis revealed that DDP monotherapy and combination therapy groups exhibited reduced spleen indices (a marker of splenic toxicity) and elevated renal toxicity indices, consistent with the known nephrotoxic effects of DDP (Figure 4E).
This in vivo study confirms that CLDN2 silencing synergizes with cisplatin to achieve superior and sustained tumor regression in NSCLC. However, this combination retains the characteristic systemic toxicity profile associated with cisplatin monotherapy, indicating that its enhanced therapeutic window stems primarily from increased efficacy rather than reduced toxicity.

3.5
miR‐194‐3p Suppresses
CLDN2
Through Dual 3′ UTR Binding in NSCLC
Bioinformatic screening (miRDB, TargetScan) identified miR‐194‐3p as a putative regulator of CLDN2, with evolutionarily conserved binding motifs at positions 358–365 and 1232–1238 of the CLDN2 3′ UTR (Tables 1, 2). In 293 T cells, dual‐luciferase assays demonstrated that miR‐194‐3p overexpression significantly suppressed wild‐type CLDN2 3′ UTR reporter activity. Site‐directed mutagenesis at either predicted site (358–365 or 1232–1238) or concurrent mutation of both sites abrogated this suppression, rendering luciferase activity indistinguishable from controls (Figure 5A).
In NSCLC cell lines, transfection of a miR‐194‐3p mimic reduced CLDN2 protein abundance (Figure 5B) and transcript levels (Figure 5C). Confocal microscopy revealed attenuated CLDN2 membrane localization in mimic‐treated cells, consistent with functional downregulation (Figure 5D). Conversely, transfection of a miR‐194‐3p antisense inhibitor elevated CLDN2 protein and mRNA expression above baseline levels in both A549 and HCC827 cells (Figure 5B,C). Notably, CLDN2 membrane signal intensity in inhibitor‐treated cells exceeded that of untreated controls, indicating ectopic overexpression rather than restoration of physiological expression levels (Figure 5D). Notably, given the critical regulatory role of miR‐194‐3p on CLDN2 established above, we further investigated the effect of Art B on miR‐194‐3p expression in NSCLC cells. Our results demonstrate that Art B significantly upregulates the intracellular level of miR‐194‐3p (Figure S1).
This study identifies miR‐194‐3p as a key post‐transcriptional regulator of CLDN2 in NSCLC, directly suppressing its expression via two evolutionarily conserved binding sites in its 3′ UTR. The additional finding that Art B upregulates miR‐194‐3p suggests this microRNA as a potential upstream mediator of Art B's pharmacological action.

3.6
miR‐194‐3p Enhances Cisplatin Efficacy Through MRP2 Silencing in NSCLC
Overexpression of miR‐194‐3p suppressed NSCLC cell proliferation and potentiated DDP sensitivity. Longitudinal monitoring revealed sustained growth arrest over 96 h in A549 and HCC827 cells following miR‐194‐3p overexpression, whereas inhibition of endogenous miR‐194‐3p partially rescued proliferative capacity (Figure S2A). Morphological assessment demonstrated reduced cellular density and cytoplasmic condensation in miR‐194‐3p‐overexpressing cells, contrasting with maintained structural integrity in inhibitor‐treated groups (Figure S2B).
Dose–response cytotoxicity assays demonstrated that miR‐194‐3p overexpression enhanced DDP‐induced cell death across increasing concentrations, while inhibition of miR‐194‐3p attenuated chemosensitization (Figure 6A). Apoptotic profiling confirmed significantly increased cell death in miR‐194‐3p‐overexpressing groups, characterized by elevated proportions of cells in early and late apoptotic phases; this phenotype was reversed by miR‐194‐3p inhibitor co‐treatment (Figure 6B).
Mechanistically, miR‐194‐3p mediated post‐transcriptional repression of MRP2 (a canonical multidrug resistance efflux transporter) without affecting CTR1 expression (Figure 6C). This MRP2 downregulation correlated with increased intracellular DDP retention, evidenced by enhanced platinum‐DNA adduct formation in flow cytometric analysis (Figure 6D) and perinuclear DDP compartmentalization in confocal microscopy (Figure 6E).
These findings elucidate a functional role for miR‐194‐3p in enhancing cisplatin efficacy, mediated through its post‐transcriptional repression of the efflux transporter MRP2. This leads to increased intracellular drug accumulation and sensitizes NSCLC cells to apoptosis, positioning miR‐194‐3p as a key modulator of chemotherapeutic response.

3.7
The miR‐194‐3p/CLDN2 Axis Modulates DDP Sensitivity Through MRP2 in NSCLC
Combined CLDN2 silencing and miR‐194‐3p inhibition rescued proliferative capacity after CLDN2 depletion. Longitudinal monitoring showed restored growth kinetics in A549 and HCC827 cells over 96 h following dual intervention, contrasting with sustained suppression in CLDN2‐silenced groups (Figure S2C). Morphological analysis revealed preserved cell–cell adhesion and attenuated cytoplasmic condensation in co‐treated cells, whereas CLDN2 knockdown alone induced disrupted cellular architecture (Figure S2D).

CLDN2 silencing significantly enhanced DDP‐induced cytotoxicity, as shown by viability across increasing drug concentrations (Figure 7A). This hypersensitivity was partially rescued by miR‐194‐3p inhibition, restoring baseline chemoresistance. Apoptosis assays corroborated heightened cell death in CLDN2‐silenced groups, with expanded early and late apoptotic populations, whereas miR‐194‐3p inhibition attenuated this phenotype (Figure 7B). Mechanistically, miR‐194‐3p blockade abrogated CLDN2 knockdown‐mediated MRP2 suppression, restoring MRP2 to near‐physiological levels (Figure 7C).
This MRP2 recovery correlated with decreased intracellular DDP accumulation, quantified by reduced platinum‐DNA adduct formation (Figure 7D). Confocal imaging demonstrated distinct spatial redistribution patterns across cohorts, with CLDN2‐silenced cells exhibiting perinuclear DDP aggregation co‐localizing with nuclear DNA, whereas co‐treatment with miR‐194‐3p inhibitor showed diffuse cytoplasmic DDP distribution in vesicular compartments (Figure 7E). These spatial patterns aligned with MRP2's efflux function—perinuclear accumulation indicates impaired export, while cytoplasmic dispersion reflects restored transporter activity.
Our results position miR‐194‐3p as the principal mediator through which CLDN2 silencing enhances cisplatin sensitivity. Inhibition of miR‐194‐3p reverses the chemosensitizing effects of CLDN2 depletion by restoring MRP2 expression and its efflux function, thereby altering the spatial distribution and intracellular accumulation of cisplatin.

4
Discussion
Artemisinin derivatives hold significant therapeutic potential, demonstrating tumor‐selective cytotoxicity and favorable pharmacodynamic properties in preclinical evaluations [7]. Art B specifically shows robust antitumor efficacy in hepatocellular carcinoma, leukemia, and gastric cancer models, yet its precise molecular targets within NSCLC remain uncharted [27, 28]. Previous work by our group confirmed Art B's capacity to inhibit NSCLC proliferation, metastatic dissemination, and cisplatin chemoresistance; nevertheless, the mechanistic foundation underpinning these phenotypes was unresolved. Here, we delineate the miR‐194‐3p/CLDN2 axis as the primary effector pathway mediating Art B's anticancer activity. This discovery not only resolves a key uncertainty regarding Art B's mode of action but also establishes a novel molecular blueprint for overcoming chemoresistance in NSCLC treatment, moving beyond phenomenological observation to mechanistic understanding.
As a core constituent of leaky epithelial tight junctions, CLDN2 governs paracellular cation/water flux—a physiological function hijacked in malignancies to accelerate tumorigenesis [29]. Pathological CLDN2 overexpression defines aggressiveness in gastric, colorectal, lung, breast, renal carcinomas, and osteosarcoma, correlating with dismal prognosis [26, 30, 31]. In lung adenocarcinoma specifically, CLDN2 elevation versus normal bronchial epithelium drives oncogenesis; its ablation impedes proliferation, metastasis and chemoresistance in preclinical systems [17, 18, 32, 33]. Epigenetic modulators and flavonoids attenuate NSCLC progression by suppressing CLDN2, with phenotypic rescue upon CLDN2 restoration confirming its centrality [19, 34, 35]. Notably, ECL2‐mimetic peptides induce CLDN2 internalization and lysosomal degradation, provoking necrotic death—a promising therapeutic strategy [36]. Our chemoproteomic studies, validated via TCGA analysis and functional rescue experiments, establish that CLDN2 orchestrates cisplatin resistance through MRP2‐dependent efflux pumps, depleting intracellular drug accumulation. This finding is significant because it elucidates a non‐canonical, transporter‐mediated resistance mechanism for CLDN2, positioning it as a linchpin of chemoresistance and a compelling druggable target whose inhibition could prevent drug efflux.

CLDN2 expression is dynamically regulated by microRNAs (miRNAs), post‐transcriptional modulators that fine‐tune oncogenic networks. Established miRNAs including miR‐488, miR‐16, and the miR‐199a‐5p/214‐3p cluster engage complementary sequences within the CLDN2 3′ UTR to inhibit translation [19, 37]. We identify miR‐194‐3p as a previously uncharacterized CLDN2 repressor targeting dual loci (nt 358–365 and 1232–1238) in its 3′ UTR. This miRNA functions as a tumor suppressor in lung adenocarcinoma by impeding SLC12A5/TWIST1‐driven metastasis and in nasopharyngeal carcinoma by counteracting PTPRG‐AS1‐mediated invasion [38, 39]. Pharmacological miR‐194‐3p induction via CG200745 inhibits cholangiocarcinoma growth, while its restoration suppresses breast cancer progression across aggressive subtypes [40, 41, 42]. Crucially, we demonstrate that miR‐194‐3p directly suppresses CLDN2 in NSCLC at both transcriptional and translational levels, reversing proliferative capacity and cisplatin resistance—effects potentiated by Art B through specific miR‐194‐3p upregulation. This mechanistic linkage positions the miR‐194‐3p/CLDN2 axis as a druggable epigenetic switch governing chemoresistance. Although preclinical artemisinin data suggest favorable toxicity profiles, Art B's long‐term safety—particularly in cisplatin combination regimens—warrants rigorous clinical evaluation. The translational implication of our work lies in the potential of targeting the miR‐194‐3p/CLDN2 axis, either directly with Art B or through future miRNA‐based therapeutics, to resensitize resistant NSCLC tumors.

5
Conclusion
This work delineates CLDN2 as a molecular driver of NSCLC chemoresistance through integrated functional genomics and experimental validation, establishing its non‐canonical resistance mechanism via MRP2‐mediated cisplatin extrusion. We elucidate the miR‐194‐3p/CLDN2 axis as a master regulatory circuit governing malignancy and nominate artemisinin derivative Art B as its first‐in‐class modulator that enhances cisplatin cytotoxicity through specific miR‐194‐3p activation. These findings resolve persistent ambiguities regarding artemisinin's antitumor efficacy while positioning CLDN2 as a druggable target. Current conclusions, however, rely predominantly on in vitro models and lack validation in histological subtype‐specific patient‐derived systems. Future studies must prioritize translational development of Art B‐cisplatin combinations to counter evolved resistance, alongside advancing CLDN2‐targeted biologics or miRNA therapeutics. Critical implementation barriers remain: engineering tissue‐selective delivery platforms for miR‐194‐3p mimics and establishing predictive biomarkers for patient stratification.

Author Contributions

Ting‐Sha He: conceptualization (equal), formal analysis (lead), writing – original draft (lead). Rong‐Hui Chen: methodology (equal), investigation (equal), data curation (equal). Jing Feng: methodology (equal), validation (equal), writing – review and editing (supporting). Qiang Zhang: software (equal), formal analysis (supporting). Xin‐Ling Li: investigation (equal), resources (supporting). Jia‐Hui Han: visualization (equal), project administration (supporting). Tian‐Ze Chen: resources (equal), supervision (supporting). Rong‐Min Yu: validation (equal), writing – review and editing (equal), supervision (equal). Li‐Yan Song: conceptualization (equal), funding acquisition (lead), supervision (lead), writing – review and editing (equal). Wei‐Juan Huang: conceptualization (lead), funding acquisition (equal), project administration (lead), writing – review and editing (equal).

Funding
This work was supported by grants from the National Natural Science Foundation of China (82574486 and 82003774 to W.‐J Huang; 82174019 and 81673646 to L.‐Y Song), and Guangdong Provincial Natural Science Foundation (2023A1515012, and 2024A1515010185 to W.‐J Huang).

Ethics Statement
The mice used in this study were purchased from Beijing Huafukang Company and housed in the Experimental Animal Center of Jinan University. After a ten‐day quarantine period, the experiment was conducted. This study was reviewed and approved by the Animal Ethics Committee of Jinan University (Approval number: IACUC20201125‐06), following the principles of animal welfare and ethics.

Conflicts of Interest
These authors declare no conflicts of interest.

Supporting information

Figure S1: Bar graphs show the relative expression levels of miR‐194‐3p in A549 and HCC827 cells after treatment with negative control (NC) or Art B (6.25 μM). Data are expressed as mean ± SD (n = 3). Student's t‐test, **p < 0.01.

Figure S2: Time‐dependent viability and morphological changes under indicated treatments. (A) Viability curves of miR‐194‐3p mimic‐NC, miR‐194‐3p mimic, miR‐194‐3p inhibitor‐NC, and miR‐194‐3p inhibitor groups over 96 h. A549 is shown as circles, and HCC827 as triangles. Data are expressed as mean ± SD (n = 3). Two‐way ANOVA, ***p < 0.001. (B) Morphological assessment of miR‐194‐3p mimic‐NC, miR‐194‐3p mimic, miR‐194‐3p inhibitor‐NC, and miR‐194‐3p inhibitor‐treated cells. (C) Viability curves of NC, si‐CLDN2, and si‐CLDN2 + miR‐194‐3p inhibitor groups. A549 is shown as circles, and HCC827 as triangles. Data are expressed as mean ± SD (n = 3). Two‐way ANOVA, ***p < 0.001, ns: not significant. (D) Morphological changes in NC, si‐CLDN2, and si‐CLDN2 + miR‐194‐3p inhibitor groups.

Table S1: Primer sequences for qPCR amplification of CLDN2 and GAPDH.