Dynamin 1 promotes colorectal cancer progression by activating the PI3K/Akt signaling pathway Authors: Rui Chen, Runqi Hong, Liang Chen, Qichang Liu, Zhipeng Cao, Huipeng Wang, Shanliang Han, Chongwei Ke Published: 2025 ABSTRACT BackgroundUncontrolled proliferation, local invasion, and distant metastasis are hallmarks of malignancies, and they play crucial roles in tumor progression. Dynamin 1 (DNM1) is a member of the dynamin guanosine-5′-triphosphatease (GTPase) subfamily that regulates membrane fission, cytokinesis, and vesicle secretion. However, its function in colorectal-cancer (CRC) progression remains unclear.MethodsWe used public datasets and a tissue microarray (TMA) cohort to analyze differential expression of DNM1 between CRC tissues and adjacent nontumor tissues (NTs). The role of DNM1 in CRC cells was determined by in vitro experiments, including colony formation, Cell Counting Kit-8 (CCK-8), and Transwell assays. Ribonucleic acid sequencing (RNA-seq) and bioinformatics analysis were performed to investigate mechanisms that DNM1 may be involved in. KLF5 was identified as an upstream transcriptional activator of DNM1 through a combination of bioinformatic screening and functional experiments. Xenograft and metastatic models were established to investigate the effects of DNM1 on tumor growth and metastasis.ResultsWe found that DNM1 expression was upregulated in CRC tissues, which significantly correlated with poor patient prognoses. DNM1 overexpression (OE) enhanced the proliferation, migration, and invasion capabilities of CRC cells both in vitro and in vivo, whereas DNM1 knockdown (KD) yielded the opposite results. Mechanistically, we identified Krüppel-like factor 5 (KLF5) as an upstream transcription factor that activates DNM1 transcription. Furthermore, DNM1 was shown to mediate the activation of epithelial–mesenchymal transition (EMT) and the phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt) signaling pathway.ConclusionDNM1 promotes CRC progression and malignant phenotypes, and targeting the KLF5/DNM1/Akt axis may represents a promising strategy to improve patient outcomes.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-025-07600-1. Introduction Colorectal cancer (CRC) ranks as the third most prevalent malignancy globally, accounting for approximately 9.4% of cancer-related deaths annually [1]. The majority of patients with distant metastases are ineligible for curative resection due to extensive dissemination, excessive tumor burden, or insufficient future liver remnants, resulting in limited therapeutic efficacy and poor outcomes [2]. Therefore, targeting tumor cell proliferation, invasion, and epithelial–mesenchymal transition (EMT) is a core strategy for suppressing metastasis of CRC [3]. At the molecular level, the transforming growth factor beta (TGF-β)/Smad and phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt) signaling pathways have been extensively validated as key regulators of EMT [4]. Although the contributions of these key biological processes and their regulatory pathways to CRC progression have been increasingly elucidated over time, our understanding of the precise molecular networks governing their activities remains incomplete. Dynamin 1 (DNM1), a member of the guanosine-5′-triphosphate (GTP)–binding dynamin family, was identified in 1989 as a mechanoenzyme [5, 6]. Its expression varies across tissues; it primarily orchestrates essential membrane remodeling processes, including membrane fission, cytokinesis, and vesicle secretion [7]. Relatively higher expression of DNM1 occurs in the nervous system [8], where functional aberrations to the gene are linked to developmental disorders and epileptic encephalopathy [9]. Studies have recently linked DNM1 to the onset and progression of several types of cancer, such as bladder cancer, non–small-cell lung cancer, pancreatic ductal adenocarcinoma, and hepatocellular carcinoma [10–13]. As a membrane-localized GTPase, DNM1 may influence signal transduction pathways by modulating receptor endocytosis, vesicular trafficking, and cytoskeletal dynamics. Nevertheless, its biological functions, clinical significance, and mechanistic roles in colorectal carcinogenesis remain unclear, and the upstream transcriptional mechanisms driving its dysregulation are largely unknown. In this study, we sought to examine the clinical correlations between DNM1 and CRC progression, as well as clarify the biological mechanisms of the former in the latter. Initially, we compared DNM1 expression in CRC tissues with that in nearby normal tissues and assessed its clinical prognostic significance. Through in vitro and in vivo functional experiments, we systematically investigated the effects of DNM1 on CRC cell proliferation, migration, and invasion as well as on EMT. Integrated ribonucleic acid sequencing (RNA-seq) analysis, bioinformatics analysis, and key-protein validation further revealed the downstream molecular mechanisms by which DNM1 regulated CRC progression. Our results demonstrated that DNM1 significantly promoted CRC progression and metastasis by activating the PI3K/Akt signaling pathway. Furthermore, by combining bioinformatic prediction and experimental validation, we identified Krüppel-like factor 5 (KLF5)—a zinc-finger transcription factor with established oncogenic roles in epithelial malignancies [14]—as a key upstream transcription factor that positively regulates DNM1 at the transcriptional level. These findings not only establish the critical pro-oncogenic role of DNM1 in CRC via KLF5-mediated transcriptional activation, but also nominate it as a valuable prognostic biomarker and potential therapeutic target for suppressing tumor progression and metastasis. Materials and methods Human-tissue specimens and tissue microarray Tumor tissues and corresponding adjacent normal samples were obtained from 15 CRC patients at the Fifth People’s Hospital of Shanghai, China. DNM1 expression was analyzed following approval by the Institutional Ethics Committee. From Mingyi Bio (Shanghai, China), we purchased a tissue microarray (TMA) containing 90 human-CRC tissues and paired adjacent normal tissues, accompanied by clinicopathological information such as tumor size, differentiation grade, tumor–node–metastasis (TNM) stage, distant metastasis, date of death, and overall survival (OS). We excluded 23 cases within the TMA from statistical analysis because of tissue slice detachment. Survival analysis was performed on a TMA cohort composed of the remaining 67 CRC patients; clinical characteristics of these patients are detailed in Table 1. Table 1Clinical and pathological characteristics of the colorectal cancer patientsa (n = 67)Clinicopathological characteristicsNo. of cases (%) Age (years) < 6034(50.7)≥ 6033(49.3) Gender Male36(53.7)Female31(46.3) Tumor size (cm) < 534(50.7)≥ 533(49.3) Pathologic T classification T1/213(19.4)T3/454(80.6) Pathologic N classification N037(55.2)N130(44.8) Pathologic M classification M047(70.1)M120(29.9) TNM stage I/II25(37.3)III/IV42(62.7) DNM1 expression Low28(41.8)High39(58.2)aDuplicated data may be found in other published literatures because of the same cohort Table 1Clinical and pathological characteristics of the colorectal cancer patientsa (n = 67)Clinicopathological characteristicsNo. of cases (%) Age (years) < 6034(50.7)≥ 6033(49.3) Gender Male36(53.7)Female31(46.3) Tumor size (cm) < 534(50.7)≥ 533(49.3) Pathologic T classification T1/213(19.4)T3/454(80.6) Pathologic N classification N037(55.2)N130(44.8) Pathologic M classification M047(70.1)M120(29.9) TNM stage I/II25(37.3)III/IV42(62.7) DNM1 expression Low28(41.8)High39(58.2)aDuplicated data may be found in other published literatures because of the same cohort Clinical and pathological characteristics of the colorectal cancer patientsa (n = 67) aDuplicated data may be found in other published literatures because of the same cohort Immunohistochemical and hematoxylin and eosin staining Tissues were fixed in fixation solution (Servicebio, Wuhan, China) overnight at 4 °C, after which they were dehydrated, embedded in paraffin, and sectioned into 5-µm sections. We then deparaffinized the sections in xylene, rehydrated them in graded ethanol, and subjected them to antigen retrieval using sodium citrate buffer (Beyotime Institute of Biotechnology, Shanghai, China). Endogenous peroxidase was quenched in 3% hydrogen peroxide for 20 min at room temperature (RT). After being blocked with 10% goat serum (Beyotime) for 1 h, slides were incubated overnight at 4 °C with a rabbit polyclonal anti-DNM1 antibody (Ab; 1:50; Cat. No. OM123624; Omnimabs, Alhambra, CA, USA). After a wash in phosphate-buffered saline (PBS; Servicebio), the sections were treated with a secondary Ab conjugated to horseradish peroxidase (HRP; 1:2000; goat anti-mouse; Cat. No. P0946; Beyotime) at RT for 1 h. Staining was performed using 3,3′-diaminobenzidine (DAB; Cat. No. GK500705; GeneTech, Shanghai, China), followed by hematoxylin counterstaining for 5 s. Finally, the sections were soaked in 1% ammonia water for 10 min, dehydrated through successive ethanol solutions with graded concentrations, and cleared in xylene. During hematoxylin and eosin (H&E) staining, we followed the aforementioned hydration method and then stained the sections with hematoxylin solution for 3 min. Following a rinse under running water, the slides were exposed to 1% hydrochloric acid alcohol for 2 s, followed by a 10-min rinse in running water. We then stained the samples with eosin solution for 2 min and subsequently dehydrated them through a graded ethanol series. Finally, the samples were cleared again in xylene. All slices were scanned, and images were captured using panoramic DESK and analyzed using CaseViewer version 2.4 (both 3D HISTECH Ltd., Budapest, Hungary). Cell lines and culture conditions The CRC cell lines HCT-116, SW620, HCT-15, LOVO, and CACO2 were obtained from Ke Lei Biological Technology (Shanghai, China). The HEK-293T cell line had been preserved in our laboratory after using it in previous research. All cell lines tested negative for Mycoplasma contamination, and short tandem repeat (STR) authentication was performed. Cells were cultured in high-glucose DMEM (Solarbio, Beijing, China) containing 10% fetal bovine serum (FBS; ExCell Bio, Suzhou, China) under standard conditions (37 °C, 5% CO2) in an incubator (Thermo Fisher Scientific, Waltham, MA, USA). RNA interference Small interfering RNA sequences targeting KLF5, GATA1, SP1, and THAP1 were obtained from GenePharma (Shanghai, China) and transfected sequentially using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s protocol. Infection of CRC cells with lentiviruses Sequences encoding short-hairpin RNAs (shRNAs) targeting human DNM1/KLF5 were cloned and inserted into the pLKO.1 plasmid, and the full length of human DNM1/KLF5 complementary deoxyribonucleic acid (cDNA) was cloned and inserted into Ubi–multiple cloning site (MCS)–3×FLAG–chicken β-actin hybrid (CBh)–guard cell green fluorescent protein (gcGFP)–internal ribosome entry site (IRES)–puromycin, by Shanghai GeneChem Co., Ltd. (Shanghai, China). Lentiviruses were subsequently generated for transduction. DNM1/KLF5-targeting shRNA sequences were as follows: shDNM1-1: 5′-GCTGGCTGACTATCAATAATA-3′; shDNM1-2: 5′-CCGGAATCTTGTGGACTCATA-3′; shKLF5-1:5′-GCTGTAATGTATATGGCTTTA-3′; shKLF5-2:5′-CGGATCTAGATATGCCCAGTT-3′. On reaching 30% confluence, the cells in six-well plates were exposed to 20 µL lentiviral suspension mixed with 1 mL complete medium containing polybrene (10 µg/mL). After 8 h, this medium was exchanged for fresh culture medium. Transduced cells were selected via stepwise increase in puromycin concentration (from 5 to 10 µg/mL) in the specified medium. Transcriptomic RNA sequencing Total RNA was isolated using TRIzol reagent (Invitrogen [Thermo Fisher]) for subsequent transcriptome RNA sequencing, which was performed at the Haplox Genomics Center (Shanghai, China). In brief, the cDNA library was assembled using an appropriate amount of RNA sample. After quality control was performed on an Agilent 4200 Tape Station (Agilent Technologies, Inc., Santa Clara, CA, USA), the libraries were sequenced on an Illumina PE150 system (Illumina, San Diego, CA, USA). RNA extraction and reverse transcription quantitative polymerase chain reaction Total RNA was isolated from cultured cells using RNAiso Plus (TaKaRa Bio, Kusatsu, Japan) as per the manufacturer’s instructions. Messenger RNA (mRNA) was then reverse transcribed using HiScript III RT SuperMix (Cat. No. R323-01; Vazyme, Jiangsu, China). Data were analyzed via the 2−ΔΔCT method; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as the internal control. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) protocol consisted of three steps: (1) an initial denaturation phase of 95 °C for 30 s; (2) 40 cycles of amplification, each having 10 s of denaturation at 95 °C and 30 s of annealing/extension at 60 °C; and (3) 1 cycle at 95 °C for 15 s, followed by 60 °C for 60 s and then 95 °C for 15 s. Primer sequences used in RT-qPCR are presented in Table s1. Protein extraction and western blotting To extract total protein from cells and tissues, samples were lysed in radioimmunoprecipitation assay (RIPA) buffer (Beyotime) supplemented with protease or phosphatase inhibitors (Sharebio, Shanghai, China). After sonication for 5 s, lysates were centrifuged and supernatants collected for protein quantification. Protein concentrations were determined using a bicinchoninic acid (BCA) assay kit (Sangon Biotech, Shanghai, China). Protein samples were mixed with 5× loading buffer and denatured at 100 °C for 10 min in a metal water bath. The treated samples were then subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under appropriate voltage and transferred onto 0.45-µm polyvinylidene difluoride (PVDF) membranes (Merck Millipore, Germany). The membranes were blocked with 5% skim milk or bovine serum albumin (BSA) at RT and incubated with primary Abs for 14–16 h at 4 °C. After a wash in tris-buffered saline + Polysorbate 20 (TBST), the membranes were incubated with HRP-linked secondary Abs for 2 h at RT. Protein bands were ultimately detected using an enhanced chemiluminescence (ECL) system (SQ101; Epizyme Biotech, Shanghai, China [Ipsen Global, Paris, France]), and band intensities were analyzed using ImageJ software (v1.54f; National Institutes of Health, Bethesda, MD, USA). The Abs employed are detailed in Table s2. Immunofluorescence assay Cells were plated at equal densities onto 24 mm × 24-mm cell culture inserts placed in six-well plates. We removed the medium after 24 h and fixed the cells with 4% formaldehyde solution for 15 min. Then, the samples were incubated in 0.5% Triton for 20 min. After a wash in PBS, the samples were blocked with blocking buffer (Beyotime) and then incubated overnight at 4 °C with rabbit anti–epithelial-cadherin (CDH1) monoclonal antibodies (mAbs; 1:1500; Cat. No. A20798; ABclonal Technology, Wuhan, China). Subsequently, the samples were incubated with secondary Abs (goat anti-rabbit; Cat. No. P0948; Beyotime) at RT for 1 h, followed by 4′,6-diamidino-2-phenylindole (DAPI; Solarbio) staining for 10 min. Finally, images were captured and analyzed under a fluorescence confocal microscope (Leica, Germany). Cell proliferation assay The proliferative capacity of stably transfected cells was assessed using a Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan) following the manufacturer’s protocol. The cells were seeded in 96-well plates at a density of 3 × 103 cells/well and cultured for 24, 48, 72, or 96 h; the reagent was then added; and the plates incubated at 37 °C for 1 h. Absorbance at 450 nm was then measured using a BioTek Synergy 2 microplate reader (BioTek Instruments, Inc. [Agilent]). For the colony formation assay (CFA), the same stably transfected cells were seeded in 12-well plates. HCT-116 and HCT-15 cells with DNM1 overexpression (OE) were seeded at 400 cells/well, while SW620 cells with DNM1 knockdown (KD) were plated at 800 cells/well. All cells were cultured in complete medium for 7–14 days. Next, the cells were washed three times in PBS, fixed with anhydrous methanol for 15 min, and then stained with 0.1% crystal violet solution (Sharebio) for 1 h. Finally, after a rinse in double-distilled water (ddH2O) and drying, the plates were photographed, and colonies with cell counts > 50 were counted manually. Cell migration and invasion assays Cells were suspended at 5 × 105 cells/mL in serum-free high-glucose Dulbecco’s Modified Eagle’s Medium (DMEM; Solarbio, China). We added 200 µL of the suspension to the upper chambers of Transwell inserts with or without Matrigel coating (both Corning Labware Products, Inc., Corning, NY, USA), while DMEM supplemented with 10% fetal bovine serum (ExCell Biotechnology Co., Ltd., China) was placed in the lower chambers. After 24 h of incubation at 37 °C, cells remaining on the upper surface of the membrane were removed using a cotton swab. Migrated or invading cells on the underside were fixed with methanol for 15 min, stained with 0.1% crystal violet for 1 h, imaged using a Leica DM IL LED microscope (Leica Biosystems, Nussloch, Germany), and quantified using ImageScope software (version 12.4.6, Leica Biosystems). Animal study Animal models of xenograft growth and lung metastasis were established in 6-week-old BALB/c nude mice (SPF Bioscience, Suzhou, China). In the xenograft model, a suspension of stably transfected HCT-116 or SW620 cells (5 × 106 cells/mouse) was administered via subcutaneous injection into the right flank. Tumors were harvested for imaging and immunohistochemistry (IHC) on euthanasia at the 4-week endpoint. Volumes were determined using the following formula: (tumor width2 × tumor length) × 0.52. In the metastasis model, 1 × 106 cells in 100 µL PBS were introduced into each mouse via slow tail vein injection. After 6 weeks, mouse lungs were excised, imaged, and processed for histological evaluation with H&E staining. The percentage of lung area occupied by metastatic tumors was measured using ImageJ. All procedures were reviewed and approved by the Animal Welfare and Ethics Committee of the Experimental Animal Science Department of Fudan University, Shanghai, China. Bioinformatics analyses Gene expression datasets with clinical features, including GSE37182 and GSE17536, were obtained from the Gene Expression Omnibus (GEO) database. The transcriptome datasets TCGA-COAD and TCGA-READ were downloaded from The Cancer Genome Atlas (TCGA) Program, as were their matched clinicopathologic characteristics. Based on TCGA-COAD and TCGA-READ datasets, correlation analyses were performed between DNM1 and all other genes. Next, a gene list was generated by ranking genes according to their correlation coefficients with DNM1 expression. Subsequently, Gene Set Enrichment Analysis (GSEA) was conducted using the clusterProfiler package in R software (R Foundation for Statistical Computing, Vienna, Austria). We used a Kaplan–Meier (KM) plotter to assess the correlation between DNM1 expression and patient survival using the survival and survminer packages with optimum cutoff values in R. Following data normalization and quality control, differential-expression analysis of genes detected in RNA-seq data was conducted using the DESeq2 package in R. Then, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed using the clusterProfiler package. Statistical analysis Data analyses were conducted using Microsoft Excel 2021 (Microsoft, Redmond, WA, USA), GraphPad Prism 10 (GraphPad Software, Inc., San Diego, CA, USA), SPSS version 22 (IBM Corp., Armonk, NY, USA), and R version 4.1.2. All experiments were independently repeated at least three times, and data are presented as mean ± standard deviation (SD) unless otherwise specified. For comparisons between two groups, paired or unpaired Student’s t tests were used for continuous variables and Wilcoxon signed-rank tests for non-normally distributed data. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was conducted, followed by Dunnett’s test for multiple comparisons. Correlations between continuous variables were determined using Pearson’s correlation coefficient (PCC). The relationship between DNM1 expression and clinical characteristics was evaluated using Pearson’s χ2 test and Fisher’s exact test. KM survival analysis and the log-rank test were used to evaluate survival differences. A multivariate Cox regression model was used based on the significant variables found in univariate analysis. Statistical significance was set at P < 0.05, with all tests being two-sided. Results DNM1 was aberrantly upregulated in CRC and associated with poor prognosis To compare DNM1 mRNA expression between CRC tissues and normal tissues, we retrieved gene transcription data from the TCGA and GEO databases. We found that the expression of DNM1 was significantly higher in tumor tissues than in nontumor tissues(Fig. 1A–B) and that it gradually increased as tumor stage advanced (Fig. 1C), indicating a potential correlation between DNM1 OE and CRC progression. To verify the in silico results, we then detected DNM1 expression via Western blotting (WB) and IHC in 15 pairs of clinical samples. We observed that DNM1 protein was located mainly in the cytoplasm and that its levels were robustly elevated in most CRC tumor tissues compared to that in adjacent nontumor tissues (Fig. 1D–E). Concurrent IHC analysis of DNM1 expression in a CRC TMA that included 90 paired CRC patient tissues with corresponding clinicopathological features (23 paired samples were excluded due to tissue slice escape) revealed upregulation of DNM1 expression in the paired tumor tissues(Fig. 1F). As shown in Table 2, patients with high DNM1 expression presented a higher T stage (P = 0.025), increased risk of distant metastasis (P = 0.004), and advanced TNM stage (P = 0.001). KM survival analysis of the TMA cohort stratified by DNM1 expression levels demonstrated that high DNM1 expression was significantly associated with poor prognosis(Fig. 1G). Furthermore, subgroup analysis based on the presence or absence of distant metastasis showed higher DNM1 expression levels in patients with metastasis than in those without(Fig. 1H). Univariate Cox regression analysis revealed that DNM1 expression, N classification, M classification, and TNM stage were significantly associated with CRC patient survival. In multivariate Cox regression analyses, DNM1 expression levels (hazard ratio [HR], 2.324; 95% confidence interval [CI], 1.005–5.370; P = 0.049), M classification (HR, 2.878; 95% CI, 1.150–7.203; P = 0.024), and TNM stage (HR, 9.903; 95% CI, 2.061–47.583; P = 0.004) were considered predictors of poor prognosis(Table 3). To further investigate the adverse effects of DNM1 on CRC patient prognosis, we analyzed the TCGA-COAD, TCGA-READ, and GSE17536 clinical datasets. Taken together, these data showed that patients with high expression of DNM1 had relatively shorter OS(Fig. 1I–J). The above-described evidence collectively suggested that DNM1 was associated with unfavorable prognosis and could be a molecular prognostic marker in CRCs. Fig. 1High DNM1 expression was associated with poor prognosis in colorectal-cancer patients. (A, B) Data from the TCGA and GEO databases showed that DNM1 was upregulated in CRC tissues compared with normal controls. (C) DNM1 expression discrepancies among different tumor stages in CRC patients in the TCGA-COAD and -READ datasets. (D, E) Relative protein expression of DNM1 in 15 paired CRC tissues and adjacent nontumor tissues was determined by WB and IHC staining. (F) Following IHC staining of the CRC TMA, DNM1 expression levels were quantified via histochemical score and compared between tumor tissues and adjacent normal tissues. (G) KM survival analysis was performed on the TMA patient cohort grouped by DNM1 expression levels. (H) Patients in the TMA cohort were grouped by the presence or absence of liver metastasis, and DNM1 expression levels were compared between these two groups. (I, J) Survival analyses of patients with relatively high or low DNM1 mRNA expression in public datasets (TCGA-COAD/READ and GSE17536). Data are presented as mean ± SD. *P < 0.05, **P < 0.01 versus the control group High DNM1 expression was associated with poor prognosis in colorectal-cancer patients. (A, B) Data from the TCGA and GEO databases showed that DNM1 was upregulated in CRC tissues compared with normal controls. (C) DNM1 expression discrepancies among different tumor stages in CRC patients in the TCGA-COAD and -READ datasets. (D, E) Relative protein expression of DNM1 in 15 paired CRC tissues and adjacent nontumor tissues was determined by WB and IHC staining. (F) Following IHC staining of the CRC TMA, DNM1 expression levels were quantified via histochemical score and compared between tumor tissues and adjacent normal tissues. (G) KM survival analysis was performed on the TMA patient cohort grouped by DNM1 expression levels. (H) Patients in the TMA cohort were grouped by the presence or absence of liver metastasis, and DNM1 expression levels were compared between these two groups. (I, J) Survival analyses of patients with relatively high or low DNM1 mRNA expression in public datasets (TCGA-COAD/READ and GSE17536). Data are presented as mean ± SD. *P < 0.05, **P < 0.01 versus the control group Table 2Associations between DNM1 expression and clinicopathological characteristics in CRC patients (n = 67)Clinicopathological characteristicsDNM1 expressionp-valueLow, no of cases (%)High, no of cases (%) Age (years) 0.018 < 6019 (28.4)15 (22.4)≥ 609 (13.4)24 (35.8) Gender 0.142Male18 (26.9)18 (36.9)Female10 (14.9)21 (31.3) Tumor size (cm) 0.375< 516 (23.9)18 (36.9)≥ 512 (17.9)21 (31.3) Pathologic T classification 0.025 T1/29 (13.4)4 (6.0)T3/419 (28.4)35 (52.2) Pathologic N classification 0.078N019 (28.4)18 (36.9)N19 (13.4)21 (31.3) Pathologic M classification 0.004 M025 (37.3)22 (32.8)M13 (4.5)17 (25.4) TNM stage 0.001 I/II17 (25.4)8 (11.9)III/IV11 (16.4)31 (46.3)Note: Bold format of p-values indicates statistical significance Table 2Associations between DNM1 expression and clinicopathological characteristics in CRC patients (n = 67)Clinicopathological characteristicsDNM1 expressionp-valueLow, no of cases (%)High, no of cases (%) Age (years) 0.018 < 6019 (28.4)15 (22.4)≥ 609 (13.4)24 (35.8) Gender 0.142Male18 (26.9)18 (36.9)Female10 (14.9)21 (31.3) Tumor size (cm) 0.375< 516 (23.9)18 (36.9)≥ 512 (17.9)21 (31.3) Pathologic T classification 0.025 T1/29 (13.4)4 (6.0)T3/419 (28.4)35 (52.2) Pathologic N classification 0.078N019 (28.4)18 (36.9)N19 (13.4)21 (31.3) Pathologic M classification 0.004 M025 (37.3)22 (32.8)M13 (4.5)17 (25.4) TNM stage 0.001 I/II17 (25.4)8 (11.9)III/IV11 (16.4)31 (46.3)Note: Bold format of p-values indicates statistical significance Associations between DNM1 expression and clinicopathological characteristics in CRC patients (n = 67) Note: Bold format of p-values indicates statistical significance Table 3Univariate and multivariate COX regression analyses of overall survival in colorectal cancer patients (n = 67)Clinicopathological characteristicsUnivariate analysisMultivariate analysisHR95% CIp-valueHR95% CIp-value Age (years) < 601≥ 601.6470.858–3.1650.134 Gender Male1Female1.3880.727–2.6490.320 Tumor size (cm) < 51≥ 50.8280.433–1.5820.567 Pathologic T classification T1/21T3/41.8290.712–4.6980.209 Pathologic N classification N01N12.1621.118–4.180 0.022 0.6620.260–1.6830.386 Pathologic M classification M01M18.7624.345–17.670 < 0.001 2.8781.150–7.203 0.024 TNM stage I/II1III/IV12.7013.871–41.675 < 0.001 9.9032.061–47.583 0.004 DNM1 expression Low1High4.311.958–9.487 < 0.001 2.3241.005–5.370 0.049 Note: Bold format of p-values indicates statistical significance. CI, confidence interval; HR, hazard ratio Table 3Univariate and multivariate COX regression analyses of overall survival in colorectal cancer patients (n = 67)Clinicopathological characteristicsUnivariate analysisMultivariate analysisHR95% CIp-valueHR95% CIp-value Age (years) < 601≥ 601.6470.858–3.1650.134 Gender Male1Female1.3880.727–2.6490.320 Tumor size (cm) < 51≥ 50.8280.433–1.5820.567 Pathologic T classification T1/21T3/41.8290.712–4.6980.209 Pathologic N classification N01N12.1621.118–4.180 0.022 0.6620.260–1.6830.386 Pathologic M classification M01M18.7624.345–17.670 < 0.001 2.8781.150–7.203 0.024 TNM stage I/II1III/IV12.7013.871–41.675 < 0.001 9.9032.061–47.583 0.004 DNM1 expression Low1High4.311.958–9.487 < 0.001 2.3241.005–5.370 0.049 Note: Bold format of p-values indicates statistical significance. CI, confidence interval; HR, hazard ratio Univariate and multivariate COX regression analyses of overall survival in colorectal cancer patients (n = 67) Note: Bold format of p-values indicates statistical significance. CI, confidence interval; HR, hazard ratio DNM1 potentiated the proliferation, migration, and invasion of CRC cells in vitro To investigate the role of DNM1 in CRC in vitro, we detected protein expression thereof in five human-CRC cell lines (Fig. S1A). The results showed that DNM1 expression was lower in the HCT-116 and HCT-15 lines but relatively higher in the SW620 line. Therefore, we overexpressed DNM1 in HCT-116 and HCT-15 cells and knocked it down in SW620 cells. WB and RT-qPCR analyses were conducted to verify DNM1 expression levels (Fig. 2A–B, D and E, S1B, S1C). CCK-8 (Figs. 2C, S1F) and CFA (Figs. 2G, S1D) experiments indicated that DNM1 OE potentiated cell proliferation and enhanced the migratory and invasive abilities of HCT-116 and HCT-15 cells (Figs. 2I, S1E). Conversely, DNM1 KD suppressed the proliferation (Fig. 2F and H), migration, and invasion of SW620 cells (Fig. 2J). Taken together, the results of these experiments demonstrated that DNM1 enhanced the malignant phenotypes of CRC cells. Fig. 2DNM1 promoted the proliferation, migration, and invasion of CRC cells in vitro. (A, B) WB and RT-qPCR confirmation of DNM1 overexpression in HCT-116 cells. (C) Proliferation of DNM1-OE and control HCT-116 cells was measured by CCK-8 assay. (D, E) WB and RT-qPCR confirmation of DNM1 KD in SW620 cells. (F) CCK-8 assay was employed to determine proliferation in SW620 cells transfected with control RNA and DNM1 shRNA. (G) CFA was conducted to determine cell growth in DNM1-OE and control HCT-116 cells. (H) CFA was employed to determine cell growth in SW620 cells transfected with control RNA and DNM1 shRNA. (I, J) Transwell assays were performed to examine the migration and invasion abilities of DNM1-OE and control HCT-116 cells, as well as SW620 cells transfected with control RNA and DNM1 shRNA. Data are presented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 versus the control group DNM1 promoted the proliferation, migration, and invasion of CRC cells in vitro. (A, B) WB and RT-qPCR confirmation of DNM1 overexpression in HCT-116 cells. (C) Proliferation of DNM1-OE and control HCT-116 cells was measured by CCK-8 assay. (D, E) WB and RT-qPCR confirmation of DNM1 KD in SW620 cells. (F) CCK-8 assay was employed to determine proliferation in SW620 cells transfected with control RNA and DNM1 shRNA. (G) CFA was conducted to determine cell growth in DNM1-OE and control HCT-116 cells. (H) CFA was employed to determine cell growth in SW620 cells transfected with control RNA and DNM1 shRNA. (I, J) Transwell assays were performed to examine the migration and invasion abilities of DNM1-OE and control HCT-116 cells, as well as SW620 cells transfected with control RNA and DNM1 shRNA. Data are presented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 versus the control group DNM1 promoted EMT of CRC cells In our TMA cohort, DNM1 expression was markedly higher in colorectal cancer tissues from patients with liver metastasis than in those without metastasis (Fig. 1H), suggesting a possible link between DNM1 and metastatic behavior or EMT. To further investigate the role of DNM1 in CRC progression, we next examined its association with metastatic and EMT-related phenotypes at both the transcriptomic and experimental levels. Specifically, we analyzed correlations between the expression of DNM1 and that of other genes in TCGA-COAD&READ tumors, after which we performed GSEA of gene order, in both cases as estimated by PCC. The results revealed that DNM1 might participate in the regulation of the HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION signaling pathway (Fig. 3A). Further analyses showed that high abundance of DNM1 transcripts was negatively correlated with transcriptional expression of epithelial markers such as CDH1 and positively correlated with that of other mesenchymal markers such as zinc finger E-box–binding homeobox 1 (ZEB1), neural cadherin (CDH2), vimentin (VIM), and Snail (Fig. 3B). These results were corroborated by decreased expression of E-cadherin and increased levels of ZEB1, N-cadherin, VIM, and Snail in DNM1-OE, HCT-116, and HCT-15 cells (Figs. 3C, S1G, S1H, S1J). The reverse effect was observed in DNM1-KD SW620 cells (Figs. 3C, S1I). Immunofluorescence (IF) assays further validated that DNM1-OE HCT-116 and HCT-15 cells expressed less E-cadherin on their cellular membranes than cells in the control group, while DNM1 KD increased E-cadherin expression on the membranes of SW620 cells (Figs. 3D, S1K). Taken together, these results suggested that DNM1 might regulate EMT and thereby enhance cell migration and invasion abilities in CRC. Fig. 3DNM1 promoted EMT of CRC cells. (A) GSEA of TCGA-COAD/READ data, with genes ranked by their correlation with DNM1 expression levels (as determined by PCC). (B) Gene–gene correlation analyses within the same datasets further indicated that DNM1 mRNA expression was negatively correlated with transcriptional expression of CDH1 and positively correlated with that of CDH2, ZEB1, Snail, and VIM. (C) Protein expression of EMT-related markers in DNM1-OE HCT-116 and DNM1-KD SW620 cells was analyzed via WB. (D) IF staining showed the expression of EMT-related marker E-cadherin on the cell membrane in DNM1-OE HCT-116 and DNM1-KD SW620 cells. Data are presented as mean ± SD from three independent experiments. **P < 0.01 DNM1 promoted EMT of CRC cells. (A) GSEA of TCGA-COAD/READ data, with genes ranked by their correlation with DNM1 expression levels (as determined by PCC). (B) Gene–gene correlation analyses within the same datasets further indicated that DNM1 mRNA expression was negatively correlated with transcriptional expression of CDH1 and positively correlated with that of CDH2, ZEB1, Snail, and VIM. (C) Protein expression of EMT-related markers in DNM1-OE HCT-116 and DNM1-KD SW620 cells was analyzed via WB. (D) IF staining showed the expression of EMT-related marker E-cadherin on the cell membrane in DNM1-OE HCT-116 and DNM1-KD SW620 cells. Data are presented as mean ± SD from three independent experiments. **P < 0.01 High DNM1 expression activated the PI3K/Akt signaling pathway in CRC To obtain molecular insights into the mechanisms engaged by DNM1, we used RNA-seq to analyze the transcript profiles of DNM1 KD and relative-control CRC cells. Based on a cut-off of an absolute log2 fold change > 1.0 and P < 0.05, differentially expressed genes (DEGs) were identified and are shown in the heatmap and volcano plots (Fig. 4A–B) and detailed in Supplementary Table S3. Sample separation between scramble and DNM1-KD groups was further assessed by principal component analysis (PCA) (Fig. S1L). KEGG enrichment analysis found significant enrichment of the PI3K/Akt and mitogen-activated protein kinase (MAPK) signaling pathways (Fig. 4C). Subsequently, we performed WB experiments to assess the biomarkers of the abovementioned pathways. The results revealed that DNM1 KD had no obvious effect on regulation of the MAPK signaling pathway in CRC cells (Fig. S2A). However, DNM1 KD impaired the PI3K/Akt pathway, while DNM1 OE obviously activated it (Figs. 4D, S2B). Overall, these results suggested that DNM1 might regulate the PI3K/Akt pathway and affect tumor progression in CRC. Fig. 4High DNM1 expression activated the PI3K/Akt signaling pathway in CRC. (A, B) Heatmap and volcano plots of DEGs (absolute log2FC > 1.0, P < 0.05) between scramble and DNM1-KD SW620 cells. (C) KEGG enrichment analysis of DEGs from RNA-seq suggested potential cellular pathways and functions regulated by DNM1. (D) WB showed PI3K/Akt signaling alterations in HCT-116 and SW620 cells with DNM1 OE or KD. Representative WB results from three independent experiments are shown High DNM1 expression activated the PI3K/Akt signaling pathway in CRC. (A, B) Heatmap and volcano plots of DEGs (absolute log2FC > 1.0, P < 0.05) between scramble and DNM1-KD SW620 cells. (C) KEGG enrichment analysis of DEGs from RNA-seq suggested potential cellular pathways and functions regulated by DNM1. (D) WB showed PI3K/Akt signaling alterations in HCT-116 and SW620 cells with DNM1 OE or KD. Representative WB results from three independent experiments are shown The functions of DNM1 relied on activation of the PI3K/Akt signaling pathway To further investigate the role of DNM1 in CRC progression, we performed rescue experiments using LY294002 (a PI3K inhibitor) and SC79 (a specific Akt agonist). Through dose gradient optimization, we identified 4 µg/mL SC79 and 40 µM LY294002 as the optimal concentrations for PI3K/Akt pathway activation and inhibition in CRC cells, respectively (Fig. S2C–E). Notably, PI3K/Akt pathway modulation (activation or inhibition) exhibited no regulatory effect on DNM1 expression in either CRC or 293T cells (Fig. S2F–G). We then pharmacologically manipulated PI3K/Akt pathway activity using LY294002 in DNM1-OE HCT-116 and HCT-15 cells and SC79 in DNM1-KD SW620 cells (Figs. 5A, S2H), followed by functional phenotypic assays. Our data demonstrated that PI3K/Akt inhibition significantly upregulated the expression of the epithelial marker E-cadherin and downregulated that of the mesenchymal markers ZEB1, N-cadherin, VIM, and Snail; conversely, PI3K/Akt signaling activation yielded the opposite effects (Figs. 5B, S2I). Furthermore, suppression of the PI3K/Akt pathway markedly attenuated CRC clonogenic capacity (Figs. 5C, S2J), viability (Figs. 5D, S2K), and migration and invasion abilities of CRC cells (Figs. 5E, S2L). SC79-mediated reactivation of PI3K/Akt signaling in DNM1-KD SW620 cells rescued the inhibitory effects on clonogenicity (Fig. 5F), cell viability (Fig. 5G), and migration and invasion abilities (Fig. 5H). Collectively, these findings showed that DNM1 drove colorectal-tumor progression in a mechanistically dependent manner by targeting PI3K/Akt signaling. Fig. 5The functions of DNM1 relied on activation of the PI3K/Akt pathway. (A, B) WB analysis revealed that LY294002 attenuated the effects of DNM1 OE on protein levels of the PI3K/Akt signaling pathway and EMT-related markers in HCT-116 cells, whereas SC79 rescued the effect of DNM1 KD on protein levels of PI3K/Akt signaling and EMT-related markers in SW620 cells. (C, D) CFA and CCK-8 assays showed the effect of LY294002 on the proliferation abilities of DNM1-OE HCT-116 cells. (E) Transwell assays were conducted to investigate the effect of LY294002 on the migration and invasion abilities of DNM1-OE HCT-116 cells. (F, G) CFA and CCK-8 assays showed the effect of SC79 on the proliferation abilities of DNM1-KD SW620 cells. (H) Transwell assay results showed the effect of SC79 on the migration and invasion abilities of DNM1-KD SW620 cells. Data are presented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 versus the control group The functions of DNM1 relied on activation of the PI3K/Akt pathway. (A, B) WB analysis revealed that LY294002 attenuated the effects of DNM1 OE on protein levels of the PI3K/Akt signaling pathway and EMT-related markers in HCT-116 cells, whereas SC79 rescued the effect of DNM1 KD on protein levels of PI3K/Akt signaling and EMT-related markers in SW620 cells. (C, D) CFA and CCK-8 assays showed the effect of LY294002 on the proliferation abilities of DNM1-OE HCT-116 cells. (E) Transwell assays were conducted to investigate the effect of LY294002 on the migration and invasion abilities of DNM1-OE HCT-116 cells. (F, G) CFA and CCK-8 assays showed the effect of SC79 on the proliferation abilities of DNM1-KD SW620 cells. (H) Transwell assay results showed the effect of SC79 on the migration and invasion abilities of DNM1-KD SW620 cells. Data are presented as mean ± SD from three independent experiments. **P < 0.01, ***P < 0.001 versus the control group DNM1 promoted tumor growth and metastasis in vivo Building on previous in vitro findings, we sought to investigate the in vivo effects of DNM1 on tumor growth and metastasis. Subcutaneous xenografts were established in nude mice using HCT-116 cells stably overexpressing DNM1 and SW620 cells with DNM1 KD, and tumor dimensions were measured every 4 days. Results demonstrated that DNM1 OE significantly accelerated colorectal-tumor growth in vivo (Fig. 6A–B), whereas tumor weight and volume were decreased in the DNM1-KD group compared to the control group (Fig. 6C–D). Subsequent IHC analyses of xenograft tissues confirmed a positive correlation between DNM1/antigen Kiel 67 (Ki-67) expression levels and in vivo tumor growth (Fig. 6E). To evaluate the metastatic potential of DNM1-OE or -KD CRC cells, lung metastatic models were generated in mice via tail vein injection. Six weeks post-injection, lung tissues were harvested for macroscopic imaging and H&E staining. Metastatic-burden analyses revealed that DNM1 OE potentiated CRC cell dissemination (Fig. 6F), whereas DNM1 KD dramatically attenuated distant metastasis (Fig. 6G). These in vivo findings conclusively demonstrated that DNM1 promoted colorectal-tumor growth and distant metastasis. Fig. 6DNM1 promoted tumor growth and metastasis in vivo. (A, B) DNM1-OE and control HCT-116 cells were subcutaneously implanted into BALB/c nude mice (n = 5 mice per group); four weeks later, tumors were harvested. Representative tumor images, tumor weight analysis results, and tumor growth curve are shown. (C, D) SW620 cells transfected with control RNA and DNM1 shRNA were subcutaneously implanted into BALB/c nude mice (n = 6 mice per group); tumors were harvested after 4 weeks. Representative tumor images, tumor weight analysis, and tumor growth curve are shown. (E) IHC staining revealed the expression levels of DNM1 and Ki-67 within indicated samples derived from xenograft tumors. (F) Representative images and H&E staining of mouse lungs 6 weeks after tail vein injection of DNM1-OE and control HCT-116 cells. Metastatic burden in the lungs was analyzed (n = 3 mice per group). (G) Representative images and H&E staining of mouse lungs 6 weeks after tail vein injection of SW620 cells transfected with control RNA and DNM1 shRNA. Metastatic burden in the lungs was analyzed (n = 3 mice per group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 versus the control group DNM1 promoted tumor growth and metastasis in vivo. (A, B) DNM1-OE and control HCT-116 cells were subcutaneously implanted into BALB/c nude mice (n = 5 mice per group); four weeks later, tumors were harvested. Representative tumor images, tumor weight analysis results, and tumor growth curve are shown. (C, D) SW620 cells transfected with control RNA and DNM1 shRNA were subcutaneously implanted into BALB/c nude mice (n = 6 mice per group); tumors were harvested after 4 weeks. Representative tumor images, tumor weight analysis, and tumor growth curve are shown. (E) IHC staining revealed the expression levels of DNM1 and Ki-67 within indicated samples derived from xenograft tumors. (F) Representative images and H&E staining of mouse lungs 6 weeks after tail vein injection of DNM1-OE and control HCT-116 cells. Metastatic burden in the lungs was analyzed (n = 3 mice per group). (G) Representative images and H&E staining of mouse lungs 6 weeks after tail vein injection of SW620 cells transfected with control RNA and DNM1 shRNA. Metastatic burden in the lungs was analyzed (n = 3 mice per group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 versus the control group KLF5 is a potential upstream transcription factor of DNM1 in CRC To identify potential upstream transcriptional regulators of DNM1, we employed an online transcription factor–target prediction tool that integrates data from multiple transcription factor databases [15]. Four candidate transcription factors—KLF5, GATA1, SP1, and THAP1—were identified across three databases (Fig. S3A). We then designed corresponding siRNAs and examined their effects on DNM1 mRNA expression. Among these candidates, only KLF5 knockdown led to a significant decrease in DNM1 mRNA levels (Fig. S3B-C), suggesting that KLF5 may positively regulate DNM1. To further validate the regulatory role of KLF5, we established stable KLF5 KD and KLF5 OE cell lines. Successful modulation of KLF5 expression was confirmed by RT-qPCR and WB. In SW620 cells, KLF5 OE increased DNM1 mRNA and protein expression (Fig. S3D-F), whereas KLF5 KD reduced DNM1 expression (Fig. S3G-I), further supporting KLF5 as a positive regulator of DNM1 at transcriptional level. In summary, these findings suggest that KLF5 may bind to the promoter of DNM1 and activate its transcription, thereby enhancing DNM1 expression in CRC. In conclusion, KLF5, as an upstream transcription factor, transcriptionally activated DNM1, which in turn drove CRC progression by regulating proliferation, migration, invasion, and EMT through the PI3K/Akt signaling pathway. This KLF5/DNM1/Akt axis underlines the potential of DNM1 as a promising therapeutic target(Fig. 7). Fig. 7Schematic representation of DNM1-driven CRC progression. KLF5 acts as an upstream transcription factor that positively regulates DNM1 expression. DNM1 activated the PI3K/Akt pathway, inducing EMT and oncogenic phenotypes (proliferation, migration, and invasion). Pharmacological inhibition (LY294002) or activation (SC79) of PI3K/Akt signaling reversed DNM1-mediated effects Schematic representation of DNM1-driven CRC progression. KLF5 acts as an upstream transcription factor that positively regulates DNM1 expression. DNM1 activated the PI3K/Akt pathway, inducing EMT and oncogenic phenotypes (proliferation, migration, and invasion). Pharmacological inhibition (LY294002) or activation (SC79) of PI3K/Akt signaling reversed DNM1-mediated effects Discussion Colorectal cancer ranks as the third most common malignancy globally, and it is the second-leading cause of cancer-related mortality [16, 17]. Improving the diagnosis and treatment of this cancer critically depends on the continuous exploration and discovery of novel biomarkers and therapeutic targets. Dynamin-1, a member of the GTP-binding dynamin subfamily [5, 7, 18], has been studied primarily for its role in regulating vesicle secretion and synaptic transmission within the nervous system [6, 19]. This study is the first to reveal a significant pro-tumorigenic function of DNM1 in CRC. Clinical-sample analysis demonstrated that DNM1 expression was significantly elevated in CRC tissues compared to adjacent normal tissues. Moreover, patients with high DNM1 expression presented with advanced TNM stages and significantly shorter OS. This finding not only expands the understanding of DNM1’s biological functions beyond the nervous system into oncology but also suggests its potential as an independent prognostic predictor of CRC. Epithelial–mesenchymal transition is a process in which cells lose their epithelial characteristics and acquire mesenchymal properties, enhancing their migratory and invasive capacities [20]. Most human-CRC cells exhibit partial EMT activation, a state implicated in metastasis [3, 21, 22]. Studies indicate that EMT not only facilitates tumor metastasis but is also closely associated with cancer stemness and therapeutic resistance [23]. Our study, for the first time, combined bioinformatics and experimental validation to reveal the molecular mechanism by which DNM1 regulated EMT to promote CRC metastasis. High DNM1 expression enhanced CRC cell migration and invasion, which was concomitant with the upregulation of the mesenchymal markers ZEB1, VIM, N-cadherin, and Snail and significant downregulation of the epithelial marker E-cadherin. Notably, DNM1 KD reversed these phenotypic changes. These results established DNM1 as a positive regulator of the EMT process, providing crucial experimental evidence for its role in driving CRC progression. Consistent with these in vitro findings, in vivo experiments further demonstrated that DNM1 overexpression promoted CRC metastasis, whereas DNM1 knockdown markedly suppressed metastatic dissemination, suggesting that DNM1 may serve as a biomarker for assessing metastatic risk in CRC. Moreover, given the close association between EMT and therapeutic resistance [24], it will be important to determine whether DNM1 contributes to the development of treatment resistance in CRC. In the future, the development of small-molecule inhibitors targeting DNM1 may provide a means to counteract its pro-EMT and pro-metastatic effects, although this hypothesis requires rigorous pharmacological validation. This study demonstrated that DNM1 activated core components of the PI3K/Akt pathway, as evidenced by increased activity of key molecules (PI3K, phosphorylated PI3K, pyruvate dehydrogenase lipoamide kinase isozyme 1 [PDK1]) and elevated phosphorylation of Akt at serine 473. Importantly, the PI3K/Akt pathway is a critical hub regulating EMT [25–28]. Our results further support the critical role of this pathway in CRC at a functional level. Rescue experiments confirmed that the PI3K inhibitor LY294002 (40 µM) partially blocked DNM1-induced cell proliferation, migration, invasion, and EMT progression without alteration of DNM1 expression. Conversely, the Akt activator SC79 (4 µg/mL) effectively rescued the suppression of these phenotypes caused by DNM1 KD. Collectively, these data demonstrate that the PI3K/Akt pathway constitutes a key downstream effector axis through which DNM1 exerts its pro-tumorigenic effects and provide experimental support for the development of CRC intervention strategies centered on this pathway. This study also identified KLF5 as a potential transcriptional activator of DNM1, revealing a previously unrecognized KLF5–DNM1 regulatory axis in CRC. Krüppel-like factor 5 (KLF5) is a zinc-finger transcription factor that is highly expressed in multiple epithelial-derived malignancies and is widely regarded as oncogenic [29–32]. By integrating transcription factor prediction with KLF5 knockdown and overexpression experiments, we demonstrated that alterations in KLF5 levels positively regulated DNM1 mRNA and protein expression, suggesting that KLF5 may activate DNM1 transcription by binding to its promoter region and thereby promote CRC progression. This finding not only provides a new transcriptional regulatory mechanism underlying DNM1 overexpression in CRC but also further reinforces the oncogenic role of KLF5 in CRC progression. At the clinical level, the concomitant upregulation of KLF5 and DNM1 suggests that their combined assessment may improve the precision of metastasis risk stratification and evaluation of therapeutic responses in patients with CRC, and that the KLF5–DNM1 axis may represent a promising therapeutic target for preventing CRC progression. In conclusion, this study established for the first time the oncogenic role of DNM1 in CRC and systematically elucidated the molecular mechanism by which it promoted tumor progression: through both activation of the PI3K/Akt signaling pathway and induction of EMT. These findings position DNM1 as a potential prognostic and therapeutic target. Nevertheless, this work has limitations that highlight important future research directions. First, this study was primarily based on cell lines and nude-mouse models to investigate the function and mechanisms of DNM1; however, these models cannot fully recapitulate the complexity of human tumors, particularly with respect to the tumor microenvironment and the immune system. Therefore, future studies will need to validate our findings in larger and more representative clinical cohorts in order to enhance their clinical relevance and translational potential. Second, although our omics and functional data support a role for DNM1 upstream of the PI3K/Akt pathway, the precise molecular mechanism remains unclear. In particular, future studies will be needed to determine whether DNM1 directly interacts with components of the PI3K complex or AKT, or instead modulates this pathway indirectly. Finally, therapeutic strategies targeting DNM1 or its critical downstream effectors must be developed and validated for efficacy and safety in clinically relevant models. Addressing these questions will significantly advance the translational potential of DNM1 as a novel diagnostic and therapeutic target in CRC. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2