Inhibition of LIMK by Cofilin-1 peptidomimetics enhances actin depolymerization and reduces metastasis of non-small cell lung cancer Authors: Min-Ying Lin, Jyh-Der Leu, Chun-Yi Wu, Liang-Cheng Chen, Yi-Jang Lee Published: 2026 ABSTRACT AbstractMetastasis is a major contributor to the mortality and morbidity of human non-small cell lung cancer (NSCLC), driven by actin cytoskeletal rearrangements that enhance cancer cell motility and invasiveness. Cofilin-1 (CFL-1), a key actin-binding protein regulated by LIM kinase (LIMK), controls actin turnover through phosphorylation at serine-3. However, the role of CFL-1 phosphorylation in lung cancer metastasis remains unclear. In this study, in silico analysis revealed that high LIMK/CFL-1 expression correlates with poor prognosis in lung cancer patients. Although total CFL-1 levels were elevated across NSCLC cell lines, its serine-3 phosphorylation showed a stronger association with cellular invasiveness. To therapeutically target this pathway, we developed a cofilin mimetic peptide (CMP) designed to competitively inhibit endogenous CFL-1 phosphorylation. CMP treatment selectively accumulated in LIMK-overexpressing lung cancer cells both in vitro and in vivo, and effectively suppressed actin stress fiber formation by enhancing actin depolymerization. Acting as a competitive decoy of LIMK, CMP reduced cancer cell motility and invasion, and significantly inhibited metastatic progression in orthotopic NSCLC mouse models. Importantly, CMP treatment prolonged survival without inducing systemic toxicity. These findings highlight CFL-1 phosphorylation as a critical driver of lung cancer metastasis and propose peptidomimetic inhibition of the LIMK/CFL-1 pathway as a promising therapeutic strategy for metastatic NSCLC. Subject terms: Phosphorylation, Cytoskeleton, Metastasis Introduction Lung cancer is the most common malignancy with high mortality worldwide. More than 85% of lung cancers are non-small cell lung cancer (NSCLC) [1]. This aggressive is associated with a high recurrence rate, with 70–80% of NSCLC patients experiencing disease relapse within two years after surgery, contributing to poor clinical outcomes [2]. Additionally, the 5-year survival rate for metastatic NSCLC is around 7%, with first-line treatment typically involves small molecular compounds targeting specific tyrosine kinase receptors and immune checkpoints [3]. However, the high redundancy of signaling pathways may limit the efficacy of these compounds and induce drug resistance of cancer cells [4]. Moreover, advanced NSCLC often shows high resistance to chemotherapy and radiation therapy, suggesting the need to develop alternative therapeutic approaches for effective treatment [5, 6]. Despite major progress in understanding NSCLC pathogenesis, the mechanisms by which cancer cells reprogram cytoskeletal dynamics and bioenergetics to support invasion and metastasis remain incompletely understood. Cell migration is a crucial and highly regulated process driving cancer metastasis, primarily controlled by the dynamic remodeling of the actin cytoskeleton and its regulatory proteins. The continuous assembly and disassembly of actin filaments generate the forces required for cell movement, allowing cancer cells to infiltrate surrounding tissues and disseminate to distant sites [7, 8]. Despite extensive research, targeting actin dynamics remains challenging due to the redundancy and adaptability of the signaling pathways involved [9]. Direct inhibition of actin polymerization using cytoskeletal destabilizers has shown anti-proliferative and anti-metastatic potential but is limited by high systemic toxicity, as the actin cytoskeleton is essential for all cell types and vital organs such as the liver and heart [10]. Consequently, recent strategies have shifted toward selectively targeting actin-associated proteins that are overexpressed or uniquely activated in tumor cells. For example, the anti-tropomyosin compound TR-100 suppresses melanoma and neuroblastoma growth in vivo with minimal toxicity by selectively disrupting tumor-specific actin filaments containing the Tm5NM1/2 isoform [11]. Therefore, identifying and selectively inhibiting key actin-regulatory proteins in cancer cells represents a promising approach to suppress metastatic dissemination while minimizing off-target effects in normal tissues. The actin depolymerizing factor (ADF)/cofilin family comprises key actin- binding proteins that modulate actin filament dynamics both in vitro and in vivo [12, 13]. Among them, cofilin-1 (CFL-1) is the most abundant isoform expressed in non-muscle tissues. CFL-1 accelerate actin filament turnover by approximately 200 times in vitro and plays a pivotal role in regulating cell motility, morphogenesis, and cytokinesis [14, 15]. CFL-1 is long to know a poor prognostic factor in human cancers [16, 17]. However, its expression and activity are malignancy remain unclear. The activity of CFL-1 is tightly controlled by phosphorylation at serine-3, a modification catalyzed by the Rho-ROCK-LIM kinase (LIMK) signaling cascade [18]. In cancer cells, the activation and deactivation cycle of the LIMK/CFL-1 axis is essential for coordinating actin remodeling, and even subtle alterations in this pathway can profoundly influence invasive behavior [19–22]. In line with this, inhibition of LIMK activity has been reported to suppress tumor progression across several cancer types [23, 24]. Furthermore, activation of LIMK by transforming growth factor-β (TGF-β) leads to CFL-1 phosphorylation, which inhibits its actin-depolymerizing function and promotes cytoskeletal reorganization during epithelial-mesenchymal transition (EMT) [25, 26]. Therefore, the development of low-toxicity and effective agents that disrupt the LIMK/CFL-1 signaling axis represents a promising therapeutic strategy. However, suitable molecular candidates for targeting this pathway remain underexplored. Given the central role of CFL-1 phosphorylation in actin cytoskeletal regulation and metastatic progression, we hypothesized that inhibiting CFL-1 phosphorylation could suppress the invasive and metastatic potential of NSCLC cells. To test this, we designed a synthetic cofilin mimetic peptide (CMP) to act as a competitive decoy, thereby blocking CFL-1 phosphorylation and disrupting LIMK-mediated signaling. In this study, we demonstrate that CMP effectively suppresses cytoskeletal remodeling and metastasis in vitro and in vivo, highlighting the therapeutic potential of targeting the LIMK/CFL-1 pathway for both imaging and treatment of metastatic lung cancer. Materials and methods In silico database analysis Data from clinical patients and genomic profiles available in The Cancer Genome Atlas (TCGA) database were retrieved and analyzed using the UCSC Xena browser [27]. Publicly accessible datasets were employed to examine gene expression patterns and their relationship to survival outcomes. The Kaplan–Meier methods were utilized to evaluate the correlation between mRNA expression levels and patient survival rates. Differential analysis of gene expression was performed using GEO2R on a previously published microarray data set (GSE42407). Immunohistochemical staining (IHC) Lung cancer tissue arrays (Cat. #BC041115e) from US Biomax were analyzed for CFL-1 expression using immunohistochemical staining. Tissue sections embedded in paraffin were initially heated at 60 °C for 1 h to facilitate the deparaffinization process. Deparaffinization was conducted with xylene, followed by rehydration through a graded ethanol series, beginning with 95% ethanol and gradually transitioning to 75%. Antigen retrieval was performed by heating the slides in 10 mM citric acid buffer with 0.05% Tween-20 (pH 6.0) at 121 °C for 3 min. Subsequently, the tissue sections were incubated overnight at 4 °C with anti‑CFL‑1 antibody (1:100). This was followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies, which amplify the signal for visualization. To detect the antibody-antigen complexes, the sections were rinsed thoroughly with PBST and treated with 3’,3’-diaminobenzidine (DAB) substrate chromogen, resulting in a brown precipitate marking positive signals. Finally, the tissue sections were counterstained with Mayer’s hematoxylin to enhance cellular detail and provide contrast for microscopic examination. The sections were scanned using Axioscan 7 Slide Scanner (Carl Zeiss Microscopy GmbH, Jena, Germany). Cell cultures The human bronchial epithelium cell line Beas2B (RRID: CVCL_0168) (American Type Culture Collation, Manassas, VA, USA) was cultured in keratinocyte-SFM medium (Life Technologies Inc., Carlsbad, CA, USA). The NSCLC cell lines H1299 (RRID: CVCL_0060), H441 (RRID: CVCL_1561), H292 (RRID: CVCL_0455) and H1355 (RRID: CVCL_1464) were obtained from ATCC and maintained in RPMI-1640 medium (Life Technologies Inc., Carlsbad, CA, USA) supplemented with 10% FBS and 1% penicillin-streptomycin. The human lung adenocarcinoma cell line A549 (RRID: CVCL_0023), obtained from ATCC, was maintained in DMEM (Life Technologies Inc., Carlsbad, CA, USA) supplemented with 10% FBS and 1% penicillin-streptomycin. These authenticated cell lines were obtained from the ATCC cell bank within the past three years, accompanied by a certificate of authentication. The NSCLC cell lines CL1-0 (RRID: CVCL_3871) and CL1-5 (RRID: CVCL_D521) were provided as a gift by Dr. Pan-Chyr Yang and grown in RPMI1640 medium (Life Technologies Inc., Carlsbad, CA, USA) supplemented with 10% FBS and 1% penicillin-streptomycin. These cell lines were originally developed by Dr. Chu et al. through in vitro selection for progressive invasiveness from a clonal human lung adenocarcinoma cell line (CL1) [28]. All cells were maintained at 37 °C in a humidified atmosphere with 5% CO₂ and confirmed as mycoplasma-free using a detection kit. Gene construction and lentivirus production The shRNA plasmids pLKO.1-shLuc and pLKO.1-shCFL1 were purchased from the RNAi core facility (Academia Sinica, Taiwan). For Flag-CFL WT, and S3A expression, Flag-CFL WT, and S3A cDNA were amplified by PCR and then subcloned into the lentiviral vector pCDH-CMV-MCS-EF1-puro. This construct was co-transfected with pMDG and pCMV-Δ8.91 plasmids into 293 T cells using the calcium phosphate transfection method. After 48 h of incubation, the supernatant was collected for cell infection in the presence of polybrene (10 μg/mL). The infected cells were then selected with puromycin (0.5 μg/mL) for one week. Western blot analysis Protein lysates were collected in EBC lysis buffer containing 2% protease inhibitor (Sigma-Aldrich Co., St. Louis, MO, USA). The lysate was then centrifuged at 12,000 g for 30 min at 4 °C, and the protein concentration was determined using Bradford protein-binding assay. The amounts of equal protein were performed with 10–15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then electrotransferred onto nitrocellulose membranes (Bio-Rad Laboratories Inc., Hercules, CA, USA). After blocking with 5% of bovine serum albumin (BSA, Sigma-Aldrich) or TBST with 4% skim milk for 1 h, the blots were hybridized with primary antibodies overnight at 4 °C. The membranes were incubated with a 1:10,000 dilution of anti-horseradish peroxidase conjugated anti-mouse/rabbit secondary antibody (#32430/#31460, Thermo Scientific). The signals were performed using Enhanced Chemiluminescence reagent (Bio-Rad Laboratories Inc., Hercules, CA, USA) and recorded on an Image-Quant LAS 4000 imaging system (GE Healthcare, Chicago, IL, USA). The quantification of immunoblots were measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA). All densitometric values represent the mean ± standard deviation (SD) from at least three independent experiments. The primary antibodies used in this study included anti-CFL1 (Cat. #GTX102156, dilution = 1:4000), anti-Destrin (Cat. #GTX11072, dilution = 1:1500), anti-E-cadherin (Cat. #GTX100443, dilution = 1:3000), anti-Vimentin (Cat. #GTX100619, dilution = 1:5000), anti-DDDDK tag (Cat. #GTX82562, dilution = 1:5000), anti-N-cadherin (Cat. #GTX127345, dilution = 1:5000), anti-MMP9 (Cat. #GTX100458, dilution = 1:1000), and anti-α-tubulin (Cat. #GTX112141, dilution = 1:5000) (GeneTex Inc., Alton Pkwy Irvine, CA, USA); anti-phospho (ser3) cofilin (Cat. #3313, dilution = 1:1000), and anti-LIMK2 (Cat. #3845, dilution = 1:1000) (Cell Signaling Technology, Danvers, MA, USA); anti-TBP (Cat. #ab63766, dilution = 1:1000) (Abcam, Cambridge, MA, USA); anti-GAPDH (Cat. #MA515738, dilution = 1:5000) (Thermo Fisher Scientific Inc., Waltham, MA, USA). In vitro cell invasion assay The polycarbonate filter was precoated with human fibronectin (F0395; Sigma-Aldrich Co., St. Louis, MO, USA) on its lower side and Matrigel (50% in Serum-free culture medium) on their upper side, and medium containing 10% FBS was added to the lower compartment of each Boyden chamber. 15,000 cells were mixed in serum-free medium and loaded into each well of the upper layer. Cells were incubated for 8–16 h at 37 °C in 5% CO2. The cells were fixed with 100% methanol and then stained with Giemsa solution. The cells on the lower side of the membrane were further counted by light microscopy. All experiments were carried out in triplicate. Phalloidin staining Cells were seeded onto the coverslips for 24 h and then fixed with 4% paraformaldehyde at 37 °C for 30 min. After fixation, cells were permeabilized for 5 min using Triton X-100 at a concentration of 0.1%. After permeabilized, cells were co-incubated with Alexa 488 fluorescein-conjugated phalloidin (Life Technologies Co, Grand Island, NY, USA) and 4’,6’-diamidino-2-phenylindole (DAPI) for actin filaments and nuclear staining, respectively. Cells were mounted by 90% glycerol in PBS and the images were acquired using a confocal fluorescence microscope (Zeiss LSM880, Carl Zeiss Microscopy GmbH, Jena, Germany). Design and treatment of CMP To specifically interfere with the LIMK/CFL-1 signaling axis, we designed a cofilin mimetic peptide (CMP) based on the unique structural recognition mechanism of the kinase. Structural analyses have demonstrated that the N-terminal Ser-3 residue of cofilin inserts into a specific catalytic “pit” on LIMK, a non-canonical interaction that is essential for phosphorylation [29]. Guided by this rationale, we designed a 16-amino acid sequence mimicking the N-terminal domain of human cofilin-1 (MASGVAVSDGVIKVFN). The length was selected to include the critical Ser-3 phosphoacceptor and preserve the adjacent structural determinants required for competitive LIMK recognition. The CMP and its reverse sequence control (RV-CMP; NFVKIVGDSVAVGSAM) were synthesized by BioTools Co., Ltd (New Taipei City, Taiwan) and conjugated with the cell-penetrating sequence (RQIKIWFQNRRMKWKK) to facilitate intracellular delivery. Purities of the purified peptides were over 95% and characterized by matrix assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Each synthetic peptide was dissolved in sterile double-distilled water to a concentration of 2.5 mM and added to the cultures to the indicated final concentrations. Cellular uptake assay studies Cells were seeded in a 6-well plate at a density of 1 × 106 cells/well and incubated overnight. Culture medium was replaced by fresh medium containing 111In-DTPA-CMP (1 μCi/mL). At different time points post incubation, the medium and washing PBS were collected in a counting tube. Furthermore, cells were trypsinized and collected by washing with PBS into another counting tube. Cellular uptake of 111In-DTPA-CMP was expressed as the percentage of administrated dose per one million cells (%AD/106 cells). Establishment of CL1-5 tumor-bearing animal model We used age-matched severe combined immunodeficient NOD/SCID (NOD.CB17-Prkdcscid/JNarl) male mice at 6-8 weeks old from National Laboratory Animal Center, Taiwan. To evaluate lung colony-forming capacity, 1 × 106 CL1-5 cells were resuspended in 100 μL of PBS and injected into the lateral tail vein. Lung nodule formation was quantified after H&E staining using Axioscan 7 Slide Scanner (Carl Zeiss Microscopy GmbH, Jena, Germany) at the endpoint. For the lung orthotopic xenograft model, we resuspended 1 × 106 CL1-5 cells with the luciferase gene in 10 μL PBS and mixed with 10 μL Matrigel and intrapulmonary injection transplanted into the right lung of mice. After tumor inoculation, we measured the body weights twice a week. Imaging of tumor growth and development under different experimental conditions were investigated using In vivo Imaging System Lumina X5 (IVIS Lumina X5, PerkinElmer, Inc., Hopkinton, USA). In brief, mice were intraperitoneal injected with 150 mg/kg D-luciferin (Caliper Co., Hopkinton, MA, USA), and anesthetized using 2–4% isoflurane in IVIS system for imaging acquisition. We stopped administering the treatment to the animals when the body weights decreased below 80% of the starting body weights. The mice were euthanized by CO2 filled in the air chamber. All animal experiments were conducted according to the Guide for the Use and Care of Laboratory Animals (Animal Research: Reporting of In Vivo Experiments Guidelines). Animal SPECT/CT imaging of 99mTc-DTPA-CMP After injection of 1 mCi/100 μL 99mTc-DTPA-CMP intravenously into mice, static images were taken with the nanoSPECT/CT imaging modality (Mediso, Hungary). SPECT/CT images were acquired at 3 h after injection in the lung cancer mice model. Animal SPECT/CT images were acquired at Cheng Gung Memorial Hospital and the National Atomic Research Institute, Taoyuan City, Taiwan. The analysis of standard uptake values (SUV) was calculated and performed by the AMIDE software (Version 1.0.5). Actin polymerization/depolymerization assay The actin polymerization and de-polymerization rates were measured using an actin polymerization assay kit (Abcam, Cambridge, MA, USA). To determine polymerization, ATP contained pyrene-labeled rabbit muscle actin mixture containing pyrene-labeled ATP was incubated with cell lysate harvested in cell lysis buffer with pH = 6.8. In this particular condition, CFL-1 exhibited binding capacity without displaying severing activity. For de-polymerization, ATP contained pyrene-labeled rabbit muscle actin mixture were incubated with solvent for 1 h at room temperature to allow pyrene F-actin formation. Cell lysate buffer with pH = 8 was added into the mixture and measure fluorescence signal was measured by Ex/Em: 365/410 nm in kinetic mode for 1 to 2 h with the M2000 PRO TECAN reader (TRCAN Group Ltd., Männedorf, Switzerland). Statistical analysis Statistical analysis was performed using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA). All data are presented as the mean ± standard deviation (SD) of independent experiments. Student’s t-test was used for statistical analysis. Two-way analysis of variance (ANOVA) was used to compare the actin polymerization/depolymerization rate. The survival rates were determined by the Kaplan–Meier method with the log-rank test. A p < 0.05 was considered significant. Results Aberrant activation of the LIMK/CFL-1 pathway correlates with poor prognosis in non-small cell lung cancer To elucidate the clinical relevance of ADF/cofilin families in patients with NSCLC, we first examined mRNA expression profiles of cofilin-1, cofilin-2 (CFL-2), and destrin (DSTN) from The Cancer Genome Atlas database through in silico analysis (Fig. 1A). The results of heatmap quantification performed the patient’s normal tissues and tumor pairing showed that CFL-1 was significantly upregulated in lung adenocarcinoma cohorts than in surrounding tissues, whereas CFL-2 and DSTN were not upregulated. Furthermore, the paired analysis of CFL-1 expression within the same patient’s normal and tumor tissues confirmed this consistent upregulation in lung adenocarcinoma (Fig. 1B). To further validate these findings, we examined CFL-1 expression in NSCLC tissue samples from clinical patients. IHC staining was performed to evaluate the distribution and intensity of CFL-1 protein expression (Fig. 1C). The results demonstrated that elevated CFL-1 protein levels were correlated with more severe clinicopathological parameters in the patients (Fig. 1D). Given that LIMK phosphorylates CFL-1 to regulate actin cytoskeletal dynamics, we next investigated their combined clinical relevance. Kaplan–Meier survival analysis showed that patients with elevated expression of both CFL-1 and LIMK had significantly reduced overall survival (OS) (Fig. 1E) and progression-free survival (PFS) (Fig. 1F) compared with those with lower expression. Collectively, these findings indicate that aberrant activation of the LIMK/CFL-1 signaling axis is associated with poor prognosis and may serve as a potential therapeutic target in NSCLC. High expression of LIMK/CFL-1 correlates with poor prognosis in NSCLC patients. A Heatmap showing the expression of the CFL-1, CFL-2 and DSTN genes in non-cancerous and tumor tissues from patients with lung adenocarcinoma, based on TCGA RNA sequencing data. B Expression levels of the CFL-1, CFL-2, and DSTN gene were analysis in paired or unpaired tumor tissues and corresponding adjacent normal tissue from lung adenocarcinoma patients using TCGA RNA sequencing data. *p < 0.05. C Representative IHC staining of the CFL-1 protein in NSCLC tissues from clinical patients. The scale bar is 50 μm lengths. D The IHC scores of CFL-1 in normal lung tissues and lung cancer tissues at different stages (stages I, II, and III) were analyzed using a tissue array. *p < 0.05. E Comparison of overall survival rates in NSCLC patients. The expression of CFL-1 combined with LIMK gene expression was stratified by three groups: CFL-1 high with LIMK high, CFL-1 low plus LIMK low and others. F Comparison of progression-free survival rates in NSCLC patients. The expression of CFL-1 combined with LIMK gene expression was stratified by three groups: CFL-1 high with LIMK high, CFL-1 low plus LIMK low and others. CFL-1 serine-3 phosphorylation is a crucial regulator of actin remodeling and invasiveness in non-small cell lung cancer To investigate the functional role of CFL-1 in promoting NSCLC cell invasiveness, we analyzed microarray data from Gene Expression Omnibus (GEO) dataset (GSE42407), which provides transcriptional profiles of CL1-0 cells and CL1-5 cells that represent poorly invasive and highly invasive human NSCLC cells, respectively. We compared the Rho-ROCK-LIMK-cofilin signaling pathway between CL1-0 and CL1-5, and found that CFL-1, DSTN (also known as destrin, or ADF), and LIMK2 were significantly up-regulated in CL1-5 cells compared to CL1-0 cells, as shown in the heatmap (Fig. 2A). Western blot analysis further confirmed these results, showing elevated levels of LIMK2, CFL-1, its phosphorylated form (pCFL-1), and destrin (ADF) in CL1-5 cells compared to CL1-0 cells (Fig. 2B). To verify this observation, we manipulated CFL-1 expression in these two lung cancer cell lines with different invasive capacities. In CL1-0 cells, which exhibit low endogenous CFL-1 levels, overexpression of CFL-1 markedly increased total CFL-1, pCFL-1, destrin and invasive ability. Conversely, silencing CFL-1 in the highly invasive CL1-5 cells, which express high basal levels of CFL-1, led to a reduction in CFL-1, pCFL-1, and destrin expression as well as diminished invasion in vitro (Fig. 2C, D). Furthermore, significantly fewer lung tumor nodules and reduced tumor growth were observed in lung tissues of mice injected with CFL-1-silenced CL1-5 cells compared to those injected with non-target control CL1-5 cells. Quantification of lung nodules indicated approximately a 6-fold reduction in pulmonary lesions in the CFL-1-sliencing group relative to the non-target control group (Supplementary S1). Moreover, we validated the endogenous protein level of CFL-1 and its serine-3 phosphorylation in non-tumor Beas2B cells and eight different of lung cancer cells. Our results revealed that both CFL-1 and pCFL-1 were up-regulated in H441 and CL1-5 cell lines compared to normal bronchial cells (Fig. 2E). Statistical analysis of the expression of the CFL-1 and pCFL-1 proteins in relation to the invasion ability of these lung cancer cells showed a positive correlation between CFL-1 and pCFL-1 and cellular invasiveness across the panel of lung cancer cell lines. Notably, the phosphorylation state of CFL-1 exhibited a stronger correlation with invasion ability than the expression of the total protein of CFL-1 (Fig. 2F and Supplementary S2). To clarify how CFL-1 phosphorylation regulates cytoskeletal dynamics in lung cancer, we expressed either wild-type CFL-1 or a non-phosphorylatable S3A mutant in CL1-0 cells. As expected, phosphorylation at serine-3 was markedly reduced in cells expressing CFL-1S3A (Fig. 2G), resulting in a visible reduction in stress fiber assembly across the central cell body (Fig. 2H). Functionally, CFL-1WT promoted greater invasive capacity, whereas CFL-1S3A expression was associated with a slightly reduction in cell invasion (Fig. 2I). These results suggest that serine-3 phosphorylation of CFL-1 plays a crucial role in the regulation of actin dynamics and promoting lung cancer metastasis. The serine-3 phosphorylation of CFL-1 is a crucial factor in driving alterations in cell morphology and cell motility. A Heatmap of microarray results (GSE42407) of the mRNA expression in CL1-0 (poorly invasive) and CL1-5 (highly invasive) lung cancer cells. B Comparison of LIMK2, pCFL-1, CFL-1 and destrin protein expression in CL1-0 and CL1-5 cells. GAPDH was used as an internal control. C Western blot analysis of the expression of the CFL-1, pCFL-1 and destrin proteins after CFL-1 overexpression with pCDH vector control or pCDH-CFL-1 in CL1-0 cells, or after CFL-1 knockdown with nontarget control (shLuc) in CL1-5 cells. GAPDH was used as an internal control. D Invasive abilities of CL1-0 cells following CFL-1 overexpression with pCDH vector control or pCDH-CFL-1, and CL1-5 cells following the elimination of CFL-1 with non-target control (shLuc). Scale bar: 50 μm. *p < 0.05. E Western blot analysis of the expression of CFL-1 and pCFL-1 (Ser3) in various lung cancer cells. GAPDH was used as an internal control for protein loading (Left). Densitometric quantification of total CFL-1 and pCFL-1 expression was performed using ImageJ software, and results are presented as mean ± SD. Statistical comparisons were made relative to the normal bronchial epithelial cell line Beas2B (Right). *p < 0.05. F Correlation between the expression of the CFL-1 or phosphorylated CFL-1 protein and cellular invasion ability in various lung cancer cells and Beas2B cells from the normal bronchial epithelium. The significance of the correlation was analyzed using linear regression method. G Western blot analysis of CFL-1 and pCFL-1 expression in CL1-0 cells transfected with lentiviral plasmids encoding vector control, wild-type CFL-1, or its mutant form CFL-1S3A. GAPDH was used as a loading control. H Fluorescein-conjugated phalloidin staining of actin filaments in CL1-0 cells expressing vector control, CFL-1WT, or CFL-1S3A. Fluorescence intensity was quantified by analyzing signal profiles along the white lines indicated in each image. Scale bar: 10 μm. I Invasive capacity of CL1-0 cells transfected with vector control, CFL-1WT, or CFL-1S3A, evaluated by Giemsa staining. Scale bar: 50 μm. *p < 0.05. Cofilin Mimetic Peptide (CMP) specifically inhibits CFL-1 phosphorylation in non-small cell lung cancer To therapeutically target the CFL-1 serine-3 phosphorylation in NSCLC, we employed a peptidomimetic approach to disrupt endogenous LIMK activity. We synthesized a cofilin mimetic peptide (CMP) designed to act as a competitive decoy for LIMK. The peptide comprises the N-terminal 16 residues of human cofilin-1, a region sufficient to mimic the specific kinase-substrate interface, fused to a penetratin sequence to enable efficient intracellular internalization. A control peptide containing the reverse cofilin sequence linked to the same penetrating sequence (RV-CMP) was also generated (Fig. 3A). Western blot analysis demonstrated that CMP treatment reduced endogenous pCFL-1 levels in CL1-5 cells in a dose-dependent manner, whereas RV-CMP had no effect, confirming the specificity of CMP in inhibiting cofilin-1 phosphorylation (Fig. 3B). To evaluate the pharmacological behavior and tumor-targeting potential of CMP, the peptide was conjugated at the N-terminus with DTPA and radiolabeled with either 111In or 99mTc. After purification, both 111In-DTPA-CMP and 99mTc-DTPA-CMP exhibited high radiochemical purity 95.88 ± 1.93% and 95.76 ± 2.63%, respectively (Supplementary S3). Cellular uptake studies revealed that both 111In-DTPA-CMP and 111In-DTPA-RV-CMP reached maximal accumulation at approximately 180 min post-incubation; however, intracellular levels of 111In-DTPA-CMP remained stable over 540 min, while those of 111In-DTPA-RV-CMP declined over time, indicating enhanced intracellular retention of CMP (Fig. 3C). In vivo pharmacokinetic analysis following intravenous injection with 1.85 MBq showed comparable systemic clearance between 99mTc-DTPA-CMP and 99mTc-DTPA-RV-CMP, with terminal half-lives of 3.82 h and 3.67 h, respectively (Supplementary S4). NanoSPECT/CT imaging revealed notable accumulation of 99mTc-DTPA-CMP in orthotopic lung tumors of CL1-5 mice, with a standardized uptake value (SUV) of 1.726 ± 0.8 at 3 h post-injection (Fig. 3D). Functionally, CMP treatment significantly inhibited the migration and invasion of CL1-5 cells in a dose-dependent manner, whereas RV-CMP exhibited no such effect (Fig. 3E, F). Together, these findings demonstrate that CMP selectively accumulates in NSCLC and effectively suppresses metastatic cell motility through inhibition of CFL-1 phosphorylation. CMP acted as a competitive decoy of endogenous CFL-1 in lung cancer cells. A Schematic representation of the CMP sequence and its reverse counterpart (RV-CMP), corresponding to amino acids 1–16 of the CFL-1 protein fused to a cell-penetrating sequence. B Western blot analysis of pCFL-1, CFL-1 and GAPDH expression following treatment with various concentrations of CMP or RV-CMP. Quantification of pCFL-1/CFL-1 ratio at different CMP or RV-CMP doses is shown. *p < 0.05. C Cellular uptake of 111In-DTPA-RV-CMP or 111In-DTPA-CMP at different incubation times in CL1-5 cells. *p < 0.05. D NanoSPECT/CT imaging of CL1-5 tumor-bearing mice injected with 99mTc-DTPA-CMP at 3 h post-injection (Sag. Sagittal plane, Cor. Coronal plane). E Representative images of the wound healing assay of CL1-5 cells treated with different doses of RV-CMP or CMP. The scale bar is 100 μm lengths. Quantification of migration abilities of CL1-5 cells was shown. *p < 0.05. F Representative images of the Matrigel transwell assay of CL1-5 cells treated with different dose of RV-CMP or CMP. The scale bar is 50 μm lengths. The quantification of the invasion abilities of CL1-5 cells were shown. *p < 0.05. CMP suppresses EMT-associated cytoskeletal remodeling through inhibition of CFL-1 phosphorylation in lung cancer To determine whether CMP modulates actin dynamics in lung cancer cells, we assessed actin polymerization and depolymerization using a pyrene-conjugated actin assay. The results showed a similar rate of actin polymerization rate between the lysates treated with CMP and those treated with solvent control. However, the actin depolymerization rate was significantly enhanced with CMP treatment (Fig. 4A), indicating that CMP promotes the instability of actin filaments. Importantly, CMP treatment did not induce cytotoxicity in either normal bronchial epithelial cells or several lung cancer cell lines across a range of concentrations (Fig. 4B). Mechanistically, we investigated the intracellular distribution of the target proteins. Immunofluorescence analysis revealed that CMP treatment markedly reduced the fluorescence intensity of pCFL-1 throughout the cell, whereas the distribution and intensity of total CFL-1 remained relatively stable (Fig. 4C). To quantify these changes in specific compartments, we separated cellular lysates into cytosolic and nuclear fractions. Western blot analysis revealed that CMP treatment significantly decreases the pCFL-1/CFL-1 ratio in both the cytosolic and nuclear compartments (Fig. 4D). We further examined the downstream effects on epithelial-mesenchymal transition (EMT). In CL1-5 cells, CMP treatment reduced endogenous pCFL-1 levels and reversed EMT-associated molecular changes (Fig. 4E). This reversal was characterized by the upregulation of the epithelial marker E-cadherin and the downregulation of mesenchymal markers, including N-cadherin, vimentin, and MMP-9 (Fig. 4F). To further explore the therapeutic potential of CMP in modulating TGF-β-induced EMT, CL1-0 cells were pre-treated with TGF-β followed by CMP administration. Western blot analysis demonstrated that CMP effectively inhibited the TGF-β-induced phosphorylation of CFL-1 (Fig. 4G). Phalloidin staining showed that CMP treatment significantly reduced actin filament organization and stress fiber formation, suggesting impaired cytoskeletal remodeling as a result of CFL ser3 phosphorylation inhibition (Fig. 4H). Moreover, CMP markedly suppressed TGF-β-induced cell invasion, indicating its potential to block EMT-associated invasive behavior (Fig. 4I). Together, these findings demonstrate that CMP effectively inhibits CFL-1 phosphorylation and EMT-associated cytoskeletal remodeling, thereby reducing the invasive potential of lung cancer cells without inducing cytotoxicity. CMP regulates actin dynamics to inhibit the ability to metastasis lung cancer cells. A Comparison of the actin polymerization rate and depolymerization rate between CL1-5 cells with CMP or solvent control treatment. *p < 0.05. B Comparison of cell viability in Beas2B, CL1-0, CL1-5, A549, and H1299 cells with different dose of CMP treatment using the MTT assay. C Immunofluorescence staining of CFL-1 and pCFL-1 with DAPI in CL1-5 cells treated with solvent or CMP. Scale bar: 20 μm. The fluorescence intensity was quantified. *p < 0.05. D Western blot analysis of fractionated lysates showing pCFL-1 and CFL-1 levels in the cytosol (α-tubulin as marker) and nucleus (TBP as marker). The pCFL-1/CFL-1 ratio for each fraction is quantified. *p < 0.05. E Western blot analysis for the expression of pCFL-1, CFL-1, E-cadherin, N-cadherin, Vimentin, MMP-9 and GAPDH with different doses of CMP treatment. F The corresponding densitometric quantification of pCFL-1/CFL-1 and EMT markers (E-cadherin, N-cadherin, Vimentin and MMP-9, each normalized to GAPDH) is shown. The data are represented as mean ± SD. *p < 0.05. G Western blot analysis of pCFL-1, total CFL-1, LIMK2, and GAPDH expression in CL1-0 cells treated with CMP, TGF-β, or their combination. H Immunofluorescence images of fluorescein-conjugated phalloidin staining in CL1-0 cells following treatment with CMP, TGF-β, or their combination. Scale bar: 10 μm. The fluorescence intensity of each group was calculated. *p < 0.05. I Representative images from the Matrigel Transwell invasion assay in CL1-0 cells treated with CMP, TGF-β, or their combination. Scale bar: 50 μm. The invasion abilities of invaded cells were calculated. *p < 0.05. CMP treatment suppresses lung cancer metastasis and prolongs survival without inducing systemic toxicity To examine the therapeutic efficacy of CMP treatment in lung cancer metastasis, we established an orthotopic xenograft lung cancer model in CL1-5 with luciferase gene (Supplementary S5). The tumor cells were administered via intrapulmonary injection into right lungs of NOD-SCID mice and randomly assigned to treatment and control groups. Mice were treated with 10 mmol/L/kg of CMP once every two days via intravenous injection (Fig. 5A). The IVIS spectrum showed that CMP could inhibit the lung cancer from right lobes to the left lobes of the lung and bioluminescent signals were significantly decreased after 20 days of CMP treatment. On the contrary, lung tumor cells were metastasis from right lung to the left lung with saline treatment only (Fig. 5B). The ex-vivo bioluminescence images also showed that CMP significantly reduced the signal of metastatic lung tumor to the left lung (Fig. 5C). The CMP treatment group significantly reduced the size and number of lung metastatic lesions (Fig. 5D). IHC staining also revealed a significant decrease in the intensity of endogenous CFL-1 serine-3 phosphorylation in lung tumors after 20 days of CMP treatment (Fig. 5E). To assess the toxicity of CMP treatment, histopathological examination of the liver, spleen, and kidneys showed no significant morphological changes following CMP treatment (Fig. 5F), further suggesting that CMP did not cause major organ toxicity. Furthermore, the changes in body weights showed no significantly different compared to the sham control groups and CMP-injected group (Fig. 5G). At the end of the experiment, blood samples were collected from each animal, and hematological and biochemical analyses were performed to assess toxicity. The hematology results did not show observable differences in the counts of white blood cells and red blood cells among the mice (Supplementary S6). Biochemical analysis did not reveal significant abnormalities, particularly in AST and ALT levels, indicating that high accumulation in the liver did not affect liver function after peptide treatment (Supplementary S7). After the treatment approach in the lung orthotopic mice model, CMP could prolong the mouse survival rate to the sham control group (Fig. 5H). Combining these results, the delivery of synthetic CMP efficiently inhibits lung cancer metastasis and does not affect the normal organ. CMP inhibits lung cancer metastasis in vivo. A The experimental scheme for orthotopic lung cancer model was treated by CMP intravenous injection. B Bioluminescence imaging of the response of lung cancer to CMP and solvent control after 20 days of treatment. Quantification of photon signaling by CMP treatment versus solvent control using the In Vivo Imaging System (n = 5). *p < 0.05. C Representation bioluminescence and microscopic images were evaluated using the IVIS system. Quantification of ex-vivo bioluminescence signal after CMP treatment versus solvent control using IVIS (n = 5). *p < 0.05. D Representative microscopic images of H&E staining and the number of metastatic tumors were counted (n = 5). *p < 0.05. E Representative microscopic images of IHC staining of lung tumor tissue after CMP treatment. The scale bar is 100 μm lengths. *p < 0.05. F Representative microscopic images of H&E staining of normal tissue organ after CMP treatment. G Body weight of the mice in CMP treatment group and the solvent control (saline) group during the experimental course (n = 5). ns not significant. H A Kaplan–Meier plot is displayed for solvent control (saline) or CMP treated mice, considering the time to death (n = 10). Discussion The actin cytoskeletal architectures of cancer cells are usually reconstructed via alterations of actin-binding proteins at the transcriptional or post-translational levels. Among these, CFL-1 and LIMK are key regulators whose upregulation has been associated with poor prognosis in advanced non-small cell lung cancer (NSCLC) [30, 31]. TGF-β has been shown to activate LIMK2, which subsequently phosphorylates CFL-1 at serine-3, thereby promoting actin cytoskeletal rearrangements that drive epithelial-mesenchymal transition [32]. Notably, excessive CFL-1 expression is often accompanied by an accumulation of its phosphorylated form in the serine-3, which has been linked to increased cell size and senescence [15]. In this study, we demonstrated that suppression of CFL-1 expression by shRNA can reduce the metastatic potential of NSCLC. However, our data also revealed that phosphorylated CFL-1, rather than total CFL-1 levels, correlates more strongly with invasive behavior, suggesting that CFL-1 phosphorylation serves as a crucial determinant of actin dynamics and metastatic progression. Interestingly, genetic manipulation of CFL-1 resulted in concomitant changes in endogenous destrin levels, a co-regulation that likely reflects a shared transcriptional response driven by actin cytoskeletal feedback. Previous studies have established that destrin can serve as a dynamic feedback regulator during cellular transitions, coupling its expression to cytoskeletal tension to maintain structural homeostasis [33]. The linked compensatory networks highlight the potential limitations of therapeutically targeting a single actin-binding isoform via genetic knockdown. On the other hand, LIMK is the first identified kinase targeting CFL-1, and overexpression of LIMK has been reported in various types of cancer [19, 34]. Instead of conventionally inhibiting LIMK activity, we considered a competitive substrate that may interfere with the LIMK/CFL-1 pathway in cancer progression. The use of CMP offers a direct approach to modulating actin dynamics. These peptides can mimic or enhance the natural function, leading to increased actin filament disassembly [35, 36]. Potential therapeutic applications of CMPs are promising, especially in conditions where enhanced actin turnover may be beneficial, such as in neurodegenerative diseases where cytoskeletal abnormalities are prevalent [37]. We also confirmed that CMP effectively inhibits actin reorganization in A549 cells, which exhibit well-characterized EMT responses upon TGF-β stimulation (Supplementary S8). To our best knowledge, this is the first study to use synthetic CMP for cancer treatment both in vitro and in vivo. Particularly, we found that CMP inhibited the local metastasis of human invasive NSCLC tumor model without significant toxicity and immune suppression. The survival rate of tumor-bearing mice was also increased by this compound, suggesting that the synthetic CMP is a potent anticancer agent for NSCLC treatment. Traditional chemotherapeutic agents like taxanes and vinca alkaloids, known for their anti-microtubule activity, disrupt cell division by affecting microtubule stability. However, their use is often associated with severe side effects and the potential for drug resistance in serval cancer cell lines [38, 39]. Given these challenges, targeting the LIMK/CFL pathway has gained attention as a promising alternative strategy. For, instant, LIMK inhibitors exhibit therapeutic potential by preventing CFL-1 phosphorylation and causing aberrant actin dynamics in tumor metastasis [40, 41]. Although the development of LIMK inhibitors was initiated in 2006, little success has been reported in acting as agents for anticancer or other diseases, such as glaucoma, by accessing clinical trials. However, the challenge remains to achieve specificity for LIMK isoforms and minimize side effects, as LIM kinases are involved in other cellular functions beyond actin regulation [42]. The development of CMP is believed to specifically intervene in the LIMK/CFL-1 regulatory pathway with less unknown biological effects than LIMK inhibitors. Structural analyses have demonstrated that LIM kinases exhibit exquisite substrate specificity for the Ser-3 residue and the N-terminal recognition interface, which are highly conserved across both CFL-1 and destrin [29, 43]. By acting as a decoy for LIMK, CMP effectively functions as a pan-inhibitor of this shared upstream regulatory node, preventing the phosphorylation of both overexpressed isoforms simultaneously. The design of other CMPs with a different motif and length near the Ser-3 residue may be interesting in discovering the best peptidomimetic inhibitor to block the pathophysiological functions of the LIMK-cofilin/ADF regulatory pathway for medical application. Uncovering the fundamental mechanisms holds promise as a potential venue for drug development and predicting patient prognosis. Peptide-based therapeutics have attracted attention in the field of tumor diagnosis and treatment because of their small size, high affinity, easy modification [44]. Furthermore, peptidomimetics are designed to mimic the specific conformations that bioactive peptides adopt when they bind to their targets. They are naturally occurring bioactive chemical tools that can mimic a protein surface, and that possess high selectivity, good efficacy, and low toxicity [45, 46]. However, most of peptide drug are large and hydrophilic, preventing them from crossing the plasma membrane directly and limiting their therapeutic applications. Cell-penetrating peptides (CPPs) offer a solution by enabling intracellular delivery [47]. CPPs containing the serine-3 phosphorylation site of CFL-1 can inhibit the kinase activity of LIMK1 in T-cell chemotaxis and semaphorin 3A-induced growth cone collapse [48, 49]. As we demonstrated that treatment of CMP in LIMK-overexpressing cancer cells suppressed the enhanced phosphorylation of endogenous CFL-1 and inhibits tumor metastasis, suggesting that the peptide effectively exhibits targets LIMK in advanced lung cancer. Although CPPs are widely used for intracellular delivery, previous studies have reported their limited specificity, with significant accumulation in off-target organs such as the liver, spleen, and kidneys. In addition, many CPPs exhibit rapid blood clearance, significantly limits their therapeutic effectiveness [50]. Notably, CMP includes a CFL-1 mimicking sequence linked to a penetratin, and similarly shows accumulation in these organs. This tumor selective accumulation may be attributed to the phosphorylation-dependent negative charge of CMP, which could enhance intracellular retention and improve selectivity within LIMK-active tumors in vivo. In the future, CMP can be conjugated with therapeutic radionuclides, such as lutetium-177 (177Lu) or Actinium-225 (225Ac), allowing for targeted delivery of radiotherapy [51, 52]. This strategy not only monitors the specificity of CMP on tumor targeting but also reduces systemic toxicity by confining the therapeutic radiation to cancer cells and minimizing damage to surrounding healthy tissues. Our results showed that CMP labeled with radionuclides could help us observe CMP accumulation in malignant NSCLC a both in vitro and in vivo. These findings may provide useful information on the feasibility of developing an inhibitor of CFL-1 phosphorylation in cancer therapy. In summary, the current study showed that serine-3-phosphorylated CFL-1 was more correlated with NSCLC compared to total CFL-1. We synthesized CMP, a peptidomimetic involving the serine-3 residue of CFL-1 not only modulated the rates of actin dynamics but also destabilizes the actin cytoskeleton of NSCLC cells with over-expressed LIMK. Additionally, we demonstrated that CMP can inhibit lung cancer metastasis by competing with endogenous CFL-1 phosphorylation. Furthermore, CMP labeled with medical radioisotopes was still to be phosphorylated in advanced lung cancer and would be important for theranostic purposes. Together, the regulation of cofilin activity by peptidomimetics provides a novel therapeutic approach to treat cancer malignancies. Supplementary information Supplementary Materials and Methods Supplementary S1-S8 Acknowledgements This paper is part of Min-Ying Lin Ph.D. dissertation. We thank Dr. Pan-Chyr Yang and Dr. Yu-Chan Chang for providing CL1-0 and CL1-5 cell lines. We also thank the Taiwan Mouse Clinic, Academia Sinica and Taiwan Animal Consortium for the technical supports in blood chemistry and complete blood count analysis. We would like to thank Laboratory Animal Center, Chang Gung Memorial Hospital (Linkou, Taiwan) and Department of Isotope Application Research, National Atomic Research Institute for the technical supports in nanoSPECT/CT imaging. These works were supported by the grants of National Science and Technology Council of Taiwan (114-2314-B-A49-065-MY3 and 111-2314-B-A49-037-MY3), Department of Health, Taipei City Government (11201-62-046), and grants from the Ministry of Education, Higher Education SPROUT Project for Cancer and Immunology Research Center (113W031101). Author contributions M-Y, Lin: writing—original draft, investigation, methodology, software, formal analysis, data curation, visualization; J-D, Leu: supervision, funding acquisition, project administration; C-Y, Wu: resources, validation, conceptualization; L-C, Chen: methodology, validation; Y-J, Lee: writing—review and editing, funding acquisition, supervision, conceptualization. Data availability Data sources and handling of the publicly available datasets used in this study are described in the Materials and Methods and Supplementary Materials. Publicly available datasets analyzed in this study were obtained from the Gene Expression Omnibus (GEO) under accession number GSE42407, and The Cancer Genome Atlas Lung Adenocarcinoma (TCGA-LUAD) dataset accessed via the UCSC Xena platform (https://xena.ucsc.edu/). The original Western blot imaging data generated in this study have been deposited in the Figshare repository (10.6084/m9.figshare.30415666). Competing interests The authors declare no competing interests. Ethical approval We confirm that all methods were performed in accordance with the relevant guidelines and regulations. The animal protocol was approved by the Institutional Animal Care and Use Committee, National Yang Ming Chiao Tung University, Taipei, Taiwan (IACUC No. 1110601). The human tissue microarrays (BC041115e) used in this study were commercially purchased from US Biomax. According to the vendor’s ethical policies, all tissue samples were obtained under proper local ethical and legal standards, guaranteeing the confidentiality and safety of the donors. Furthermore, all patient identities and collection dates have been de-identified to protect privacy. There was no identifiable images from human research participant in this study. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary information The online version contains supplementary material available at 10.1038/s41389-026-00624-2. References Associated Data Supplementary Materials Supplementary Materials and Methods Supplementary S1-S8 Data Availability Statement Data sources and handling of the publicly available datasets used in this study are described in the Materials and Methods and Supplementary Materials. Publicly available datasets analyzed in this study were obtained from the Gene Expression Omnibus (GEO) under accession number GSE42407, and The Cancer Genome Atlas Lung Adenocarcinoma (TCGA-LUAD) dataset accessed via the UCSC Xena platform (https://xena.ucsc.edu/). The original Western blot imaging data generated in this study have been deposited in the Figshare repository (10.6084/m9.figshare.30415666).