The miR-200 Family in Non-Small-Cell Lung Cancer: Molecular Mechanisms, Clinical Applications, and Therapeutic Implications.

1. Introduction
Lung cancer represents a major global health burden, causing over 1.8 million deaths annually [1]. The microRNA-200 (miR-200) family has emerged as a critical regulator in non-small-cell lung cancer (NSCLC), orchestrating a molecular switch centered on a bistable feedback loop with ZEB1/ZEB2 transcription factors that governs epithelial–mesenchymal plasticity, therapeutic resistance, and immune evasion [2,3]. This regulatory network has actionable interfaces with key therapies; preclinical evidence links the miR-200c/LIN28B axis to acquired resistance to EGFR tyrosine kinase inhibitors (TKIs) [4], while the circuit’s control over PD-L1 expression modulates the tumor immune microenvironment [5]. Consequently, the miR-200 family holds significant promise as a multifaceted clinical biomarker, with evidence supporting its diagnostic, prognostic, and predictive value for both targeted therapy and immunotherapy responses [6,7,8,9].
In clinical practice, treatment decisions are guided by molecular drivers and PD-L1 expression, which are strong determinants of survival [10,11]. However, tissue-based PD-L1 assessment is limited by sample availability and spatiotemporal heterogeneity. This gap highlights the urgent need for minimally invasive surrogate biomarkers, such as circulating miR-200, to enable longitudinal monitoring and better inform therapeutic strategies.
The primary objective of this review is to synthesize the extensive evidence positioning the miR-200/ZEB axis as a central node that integrates the molecular mechanisms of both targeted therapy resistance and immune evasion in NSCLC. We aim to provide a comprehensive overview of the clinical applications of this axis as a multifaceted biomarker and to highlight emerging therapeutic strategies that target this network. Ultimately, we propose a framework for future biomarker-stratified trials designed to translate these mechanistic insights into clinical practice.
A structured literature search was conducted using the PubMed and Web of Science databases to identify relevant articles published through March 2025. This review included original research and review articles focusing on the miR-200 family in non-small-cell lung cancer (NSCLC). Conference abstracts and non-peer-reviewed articles were excluded to ensure a comprehensive and reproducible synthesis of the available evidence.
In the following sections, we first synthesise the core miR-200/ZEB circuitry that governs EMT plasticity, and then link these mechanisms to clinical phenotypes and therapeutic interfaces (EGFR-TKIs and immuno-oncology). We subsequently appraise the clinical utility of miR-200 as a biomarker and survey emerging therapeutic strategies. Finally, we outline concrete future directions to accelerate clinical translation.

2. The Core miR-200/ZEB Regulatory Circuit
Building on the clinical motivation introduced above, this section delineates the bistable, double-negative miR-200-ZEB1/2 loop and its epigenetic and signalling inputs that stabilise epithelial or mesenchymal states.
The miR-200 family, which comprises miR-200a, miR-200b, miR-200c, miR-141, and miR-429, orchestrates a bistable molecular switch through a double-negative feedback loop with the ZEB1 and ZEB2 transcription factors [12,13]. This circuit enforces mutually exclusive epithelial and mesenchymal cellular states. In the epithelial state, high levels of miR-200 suppress the translation of ZEB1/ZEB2 mRNA, thereby maintaining E-cadherin expression and preserving cell–cell adhesion. Conversely, in the mesenchymal state, ZEB1 and ZEB2 bind to E-box elements within the miR-200 gene promoters, leading to reciprocal transcriptional repression that stabilizes the mesenchymal phenotype [14,15].
This core loop integrates a multitude of upstream signaling inputs. Positive regulators that promote an epithelial state include p53, which directly transactivates miR-200 expression under cellular stress, as well as the transcription factors GRHL2, KLF4, and OVOL2 [16,17]. Negative regulators that induce a mesenchymal state include the canonical EMT-inducing transcription factors SNAIL and TWIST, along with pathway-specific inputs from TGF-β (via SMAD-mediated repression), Wnt (via β-catenin/LEF complexes), and Notch signaling (via HEY1/HEY2 effectors) [18,19].
Epigenetic modifications provide an additional layer of control. In mesenchymal cells, the miR-200 promoters are silenced by CpG hypermethylation and the deposition of repressive histone marks, such as H3K27me3 and H3K9me3 [20]. These epigenetic changes create a heritable silenced state that is maintained through cell divisions.
The downstream consequences of this circuit extend beyond ZEB proteins to influence a wide array of cellular functions. These include direct targeting of CTNNB1 (modulating Wnt signaling and cell junctions), regulation of RhoE and WAVE3 (controlling actin dynamics and motility), targeting of matrix metalloproteinases and fibronectin (reducing invasive capacity), and suppression of FLT1/VEGFR1 (limiting angiogenic responses), a pathway also implicated in modulating EMT-driven TKI resistance [21,22,23]. Furthermore, the miR-200/ZEB axis regulates key stemness factors (SOX2, BMI1), metabolic enzymes (PKM2, LDHA), and immune modulators (PD-L1, CXCL1/2) [24,25]. The bistable and hysteretic nature of this circuit means that a significant and sustained signal is required to flip the switch between states, a dynamic that has profound implications for therapeutic resistance (Figure 1).
These circuit properties provide a mechanistic substrate for heterogeneous EMT states observed in tumours; we next examine how such states map onto disease progression and stemness features in patients.

3. Role of the miR-200 Family in NSCLC Phenotypes and Disease Progression
We next integrate patient-level evidence demonstrating that low miR-200 expression aligns with mesenchymal features, advanced stage, and inferior survival, while single-cell and spatial analyses contextualise this along an EMT spectrum.
Evidence from patient cohorts consistently demonstrates that low miR-200 expression in tumors correlates with mesenchymal features, advanced disease stage, and poor survival outcomes [26,27]. Large-scale studies and meta-analyses have shown that the expression levels of the miR-200 family are independent predictors of overall survival after adjusting for stage, histology, and treatment modality [7]. This prognostic significance is observed across both adenocarcinoma and squamous cell carcinoma subtypes.
Single-cell RNA sequencing has illuminated the marked intratumoral heterogeneity of miR-200 expression, revealing that tumors are composed of cellular subpopulations existing along a spectrum of EMT states rather than in binary epithelial or mesenchymal phenotypes [28]. Spatial transcriptomic analyses have further shown that miR-200-low, mesenchymal-like cells are preferentially located at the invasive fronts of tumors and within sites of lymphovascular invasion, consistent with their role in metastatic dissemination [29]. These findings help explain the frequent discordance in EMT status between primary tumors and metastatic sites.
The miR-200/EMT axis is also intricately linked to cancer stem cell (CSC) properties. Low miR-200 expression is associated with an enrichment of CD133+/ALDH1+ populations, enhanced self-renewal capacity, and resistance to anoikis [30,31]. These CSC-like features are thought to contribute to metastatic colonization and therapeutic tolerance. It is important to note, however, that this connection is context-dependent, as some studies have identified epithelial stem cell populations that maintain high miR-200 expression [32]. The clinical implication of this heterogeneity is the need for multi-site and longitudinal sampling to accurately capture the dynamic EMT state of a patient’s tumor.
Given these clinicopathologic correlations, the question becomes how EMT-linked miR-200 loss interfaces with treatment response; we therefore turn to therapeutic contexts in EGFR-targeted therapy and immuno-oncology.

4. Therapeutic Interfaces
This section aggregates evidence that EMT-coupled miR-200 suppression shapes both primary and acquired resistance to EGFR-TKIs and wires immune evasion through PD-L1 and the tumour microenvironment.

4.1. EGFR Tyrosine Kinase Inhibitors (TKIs)
Primary resistance to EGFR-TKIs in patients with activating EGFR mutations is strongly associated with baseline mesenchymal features and low miR-200 expression [33]. Complex tri-culture organoid models further validated these findings, showing that EMT-linked TKI tolerance was attenuated by either VEGF blockade (bevacizumab) or miR-200c mimic transfection. Both interventions improved TKI sensitivity and reversed EMT markers, implicating VEGF signaling in this resistance mechanism. These data nominate VEGF blockade and miR-200c restoration as rational combinatorial strategies against EMT-driven EGFR-TKI resistance [23]. The mechanisms underlying EMT-mediated resistance are multifactorial and include the activation of alternative receptor tyrosine kinases, such as AXL and MET, which sustain PI3K/AKT signaling and bypass the dependency on EGFR [34,35].
Preclinical studies suggest that the miR-200c/LIN28B axis contributes to acquired resistance. In resistant cells, the miR-200c promoter undergoes hypermethylation, leading to its silencing. This, in turn, causes the upregulation of LIN28B, an RNA-binding protein that blocks the maturation of the let-7 family of miRNAs. The subsequent decrease in mature let-7 leads to increased expression of its targets, including RAS, MYC, and HMGA2, which collectively drive stemness and drug tolerance [35,36]. Critically, restoring miR-200c expression or knocking down LIN28B can resensitize cells to EGFR-TKIs in vitro, validating this pathway as a viable therapeutic target. Clinical studies have corroborated these findings, showing that circulating miR-200c levels decline in patients prior to radiographic progression on EGFR-TKI therapy, suggesting its potential utility for monitoring emergent resistance [37,38].
Interpretation and Clinical Implications: These findings suggest that while low miR-200 expression contributes to TKI resistance, its restoration alone may be insufficient due to co-occurring bypass pathway activation, such as MET or AXL. A more rational approach may involve dual-targeting strategies that combine miR-200 restoration or epigenetic reactivation with agents targeting the tumor microenvironment, like anti-VEGF therapy. Therefore, the critical next step is to design biomarker-stratified trials evaluating such combinations in patients with EMT-high tumors, using serial liquid biopsies to monitor response and confirm predictive value.

4.2. Immuno-Oncology
The miR-200/ZEB1 circuit directly controls Programmed Death-Ligand 1 (PD-L1) by regulating its transcription and targeting its 3′-UTR, thereby linking EMT to an immune-evasive phenotype [5]. Clinical validation for this axis comes from the work of Katakura et al., who demonstrated a significant inverse correlation between miR-200b and PD-L1 tumor proportion score (TPS) in both tumor cells and serum-derived exosomes from NSCLC patients. These findings were functionally corroborated in NSCLC cell lines, where miR-200b levels directly modulated PD-L1 expression. The data strongly support miR-200b as a minimally invasive surrogate biomarker for PD-L1, mechanistically connecting EMT-driven miRNA dysregulation to immune evasion [39].
The functional consequences of this regulation extend to the tumor immune microenvironment. Tumors characterized by low miR-200 and high ZEB1 expression exhibit reduced infiltration of CD8+ T cells, increased expression of T cell exhaustion markers (e.g., PD-1, TIM-3, LAG-3), and diminished cytotoxic effector function [40]. Mechanistically, ZEB1 upregulates immunosuppressive chemokines, such as CXCL1 and CXCL2, which recruit myeloid-derived suppressor cells, while simultaneously reducing signals that attract T cells [41]. This suggests a complex therapeutic implication: mesenchymal tumors with low miR-200 expression may, despite their poor prognosis, be preferentially responsive to PD-1/PD-L1 blockade due to high baseline PD-L1 levels. Conversely, epithelial tumors with high miR-200 expression may require alternative immune activation strategies to overcome their limited baseline immunogenicity [42].
Interpretation. While the inverse correlation between miR-200 and PD-L1 links EMT to immune evasion, this relationship is complex. PD-L1 expression is also dynamically induced by inflammatory signals like IFN-γ, potentially overriding miR-200-mediated repression. This interplay involves both cell-intrinsic PD-L1 derepression after miR-200 loss and cell-extrinsic mechanisms, where the EMT program reshapes the tumor microenvironment to suppress CD8+ T cell infiltration. Consequently, the predictive value of miR-200 for immune checkpoint inhibitor (ICI) therapy remains a hypothesis until its incremental benefit over established biomarkers like PD-L1 and TMB is clinically validated.
Because circuit-level regulation of PD-L1 and T cell infiltration is measurable in tissue and circulation, we proceed to evaluate miR-200 as a clinical biomarker across diagnostic, prognostic, and predictive settings.
A hypothesis-generating clinical workflow that integrates miR-200 status with driver mutations and PD-L1 expression is summarized in Figure 2.

5. Clinical Utility of miR-200 as a Biomarker
The clinical utility of the miR-200 family as a biomarker in NSCLC is supported by a growing body of evidence across diagnostic, prognostic, and predictive contexts. Key diagnostic, prognostic, and predictive studies of the miR-200 family in NSCLC are summarized in Table 1.
Diagnostic applications have been explored using both tissue and liquid biopsy samples. A comprehensive meta-analysis reported a pooled sensitivity of 73% and specificity of 85% for NSCLC detection, yielding an area under the curve (AUC) of 0.83 [6]. Performance was superior in tissue specimens compared to blood, and quantitative RT-PCR demonstrated better reproducibility than microarray platforms.
Prognostic value is robustly established, with studies demonstrating that miR-200 expression is an independent predictor of overall survival in multivariate analyses that account for key clinical variables such as stage and histology [7].
Predictive applications are context-dependent. In the setting of EGFR-TKI therapy for EGFR-mutant patients, high miR-200c expression correlates with improved response rates and longer progression-free survival [8]. For immunotherapy, emerging clinical data suggest that serum miR-200 levels may predict outcomes, although the relationship is complex and requires further validation [9].
Liquid biopsy approaches, particularly those analyzing miRNAs within extracellular vesicles (EVs), are promising due to the enhanced stability and tumor-specificity of these miRNAs compared to their free circulating counterparts, which are found in various body fluids [43,44]. Recent studies have validated the diagnostic potential of EV-derived miR-200, showing high accuracy for detecting early-stage disease. However, widespread clinical adoption is hindered by a lack of standardization in pre-analytical and analytical variables.
Evidence Grading and Limitations. Although the miR-200 family shows promise as a biomarker, the strength of the current evidence is uneven across different clinical applications. Its prognostic value is the most consistently reported, whereas its diagnostic accuracy is moderate, and its predictive capacity is more established for EGFR-TKI response than for immunotherapy. However, widespread clinical adoption is hindered by major limitations, including small study sizes, lack of external validation, and significant heterogeneity in assay platforms and pre-analytical methods. Establishing standardized protocols and performing large-scale validation studies are therefore essential prerequisites for its integration into clinical practice.
Taken together, assayable miR-200 signatures hold actionable predictive value and enable longitudinal monitoring; we therefore survey therapeutic strategies that either restore miR-200 activity or exploit vulnerabilities of the EMT state.

6. Therapeutic Strategies and Combinations
The central role of the miR-200/ZEB circuit in NSCLC progression makes it an attractive target for therapeutic intervention. Several strategies, overviewed in general reviews of miRNA therapeutics [45], are currently under investigation.
MiRNA Replacement Therapy: The most direct approach involves the use of synthetic miR-200 mimics. Preclinical studies using lipid nanoparticle formulations have shown tumor-selective delivery and sustained target engagement. Clinical translation will require optimization of dosing, delivery vehicles, and strategies to minimize off-target effects in normal epithelial tissues.
Clinical lessons: miRNA mimic trials have shown both therapeutic potential and significant safety concerns [46,47,48]. MRX34 (liposomal miR-34a) demonstrated target engagement and responses but was terminated after immune-mediated adverse events, including cytokine reactions and immune hepatitis, resulting in four treatment-related deaths. Similarly, TargomiRs (EGFR-targeted miR-16 mimic) showed activity but caused frequent infusion reactions and anaphylactoid events. These outcomes highlight risks from innate immune activation and delivery vehicle toxicity, emphasizing the need for improved safety strategies.
On-target/off-tumour considerations for the miR-200 family: The miR-200 family is physiologically enriched in epithelial tissues and enforces epithelial identity by antagonising EMT networks (e.g., ZEB1/2). While this epithelial “gate-keeping” is desirable in carcinoma cells, supraphysiologic replacement in normal epithelia could theoretically impede EMT-dependent re-epithelialisation during wound healing and mucosal repair [21,49]. Moreover, alveolar and airway epithelia express miR-200 family members and dynamically modulate them during injury and fibrosis, implying the need to avoid broad exposure in healthy epithelium.
Risk mitigation framework: To reduce innate immune activation and constrain on-target/off-tumour effects when developing miR-200 mimics, we propose: (i) chemistry that dampens TLR7/8 and other sensors (2′-O-methyl/2′-F/LNA bases and judicious phosphorothioate use) to reduce cytokine-driven reactions [50,51,52]; (ii) delivery vehicles that localise exposure—ligand-directed lipid nanoparticles (LNPs) (e.g., EGFR- or HER3-targeting, or optimised inhaled/nebulised LNPs for thoracic disease) [53], hepatocyte-specific GalNAc conjugation where confinement to liver is desired [54,55], and antibody-oligonucleotide conjugates (AOCs) that exploit tumour antigens [56]; and (iii) expression gating when using vectorised (expressed) miRNA systems—tumour-selective promoters (e.g., survivin/BIRC5, hTERT) and/or microRNA-response-element detargeting to silence expression in normal epithelia [57,58]. Collectively, these strategies aim to preserve anti-tumour efficacy while mitigating risk to epithelia that natively express miR-200.
Epigenetic Reactivation: This strategy aims to reverse the epigenetic silencing of miR-200 expression in mesenchymal tumors. DNA methyltransferase (DNMT) inhibitors (e.g., 5-azacytidine) and histone deacetylase (HDAC) inhibitors (e.g., vorinostat) have been shown to restore miR-200 expression, with synergistic effects observed when used in combination [59,60].
Downstream Targeting: While ZEB1 and ZEB2 are challenging to target directly, novel approaches such as proteolysis-targeting chimeras (PROTACs) offer new possibilities [61]. LIN28B represents a more druggable downstream node, and small-molecule inhibitors are in preclinical development for overcoming EGFR-TKI resistance [62,63,64].
Rational Combinations: Exploiting the circuit’s biology to create synthetic lethalities is a promising avenue. Combining EGFR-TKIs with miR-200 restoration has shown synergistic effects in organoid models [34]. Similarly, combining EMT reversal agents with anti-angiogenic therapies or with immune checkpoint inhibitors could simultaneously target tumor invasion, vascular supply, and immune recognition [65,66]. Finally, restoring an epithelial state can sensitize tumors to conventional chemotherapy by enhancing apoptotic sensitivity and reducing the activity of drug efflux pumps [67].
Translational Barriers and Development Pathway: Translating these therapeutic strategies into clinical practice faces significant barriers. For instance, miRNA replacement therapies must overcome challenges in targeted delivery, potential immunogenicity, and off-tumor effects, while epigenetic reactivation strategies raise concerns about non-specific toxicity. A pragmatic development path involves early-phase trials anchored by pharmacodynamic endpoints in patients selected for EMT-high signatures. These trials should test rational combinations, such as pairing EMT modulators with EGFR-TKIs or immune checkpoint inhibitors. Ultimately, demonstrating a clear clinical benefit over existing standards of care will be essential for regulatory approval.
These avenues motivate a translation roadmap that prioritises assay standardisation, biomarker-stratified trial designs, and drug discovery efforts against transcription factors—detailed in the subsequent Future Directions.

7. Future Perspectives and Unresolved Questions
In light of the above, we distil the principal clinical messages while avoiding repetition of the forthcoming future-facing priorities. Despite significant progress, several barriers must be overcome to translate our understanding of the miR-200 family into clinical practice.

7.1. Limitations and Caveats
Evidence synthesis is constrained by (i) spatiotemporal heterogeneity of EMT states and sampling discordance; (ii) variability in pre-analytical handling and miRNA quantification platforms; (iii) predominance of retrospective or cross-sectional designs with residual confounding; (iv) limited spatial and single-cell resolution linking miR-200 states to immune contexture [68]; and (v) uncertain generalizability across histologies and treatment lines. These limitations temper causal inferences and necessitate prospective, harmonized studies before clinical adoption [69].

7.2. Trial Design Priorities
We advocate biomarker-stratified, randomized designs that (a) pre-specify EMT-high/miR-200-low strata; (b) embed serial EV-miR-200 and spatial profiling for pharmacodynamics; (c) compare standard-of-care ± EMT-targeted add-ons (anti-VEGF, epigenetic priming, miR-200 mimic); (d) test incremental predictive value over PD-L1/TMB/IFN-γ scores using decision-curve analysis; and (e) adopt iRECIST/EMT-adapted endpoints for ICI-containing regimens. Falsifiable predictions (e.g., early EV-miR-200 rise → restored epithelial markers → improved TKI sensitivity) should be prospectively specified [70,71].

8. Conclusions
The miR-200/ZEB circuit integrates EMT, therapy tolerance, and immune contexture, positioning it as a central regulatory node with translational potential. Our synthesis highlights where evidence is strongest (prognostic tissue markers; EGFR-TKI contexts) and where it remains emergent (predicting ICI response; liquid-biopsy standardization). Near-term priorities are assay harmonization, spatial/longitudinal validation, and biomarker-stratified trials that demonstrate incremental clinical utility. Real-world implementation will likely require coordinated efforts across basic, translational, clinical, and regulatory domains.