Multi-omics analysis reveals the impact of YAP/TEAD4-mediated EIF5A1 expression on mitochondrial apoptosis and bladder cancer progression

Abstract

Background

Eukaryotic Initiation Factor 5A1 (EIF5A1) is a translation factor, and its pro-tumorigenic role has been extensively documented across various cancer types. However, its specific function in bladder cancer (BLCA) remains unclear.

Methods

We integrated proteomics and transcriptomics data with clinical data from BLCA patients to investigate the correlation between EIF5A1 expression and BLCA, as well as its potential clinical applications. Transcriptomic data were employed to explore the downstream signaling pathways regulated by EIF5A1. Furthermore, ChIP analysis and luciferase reporter assays were conducted to identify the upstream transcription factors regulating EIF5A1.

Results

EIF5A1 expression is significantly upregulated in cancer tissues and cells and is strongly associated with poor prognosis. Silencing EIF5A1 in BLCA cells significantly reduced invasiveness, and proliferative capacity. Mechanistic studies identified YAP/TEAD4 as a transcription factor that regulates EIF5A1, influencing mitochondrial-mediated apoptosis by activating the JAK2/STAT3 signaling pathway, thereby promoting BLCA progression.

Conclusion

Our research demonstrates that EIF5A1 is upregulated in BLCA and associated with poor prognosis. We identified TEAD4 as a potential transcriptional regulator of EIF5A1 and showed that EIF5A1 expression is associated with changes in JAK2/STAT3 signaling and mitochondrial apoptosis in BLCA.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-13522-4.

Introduction

Bladder cancer (BLCA) is widely recognized for its significant heterogeneity and is the ninth leading cause of cancer-related deaths globally. It is also one of the most common malignant tumors in the urinary tract [1]. Statistics indicate that BLCA ranks as the tenth most common cancer worldwide, with over 430,000 new cases diagnosed annually [2, 3]. The disease exhibits high heterogeneity and is primarily categorized into two main subtypes: non-muscle-invasive BLCA and muscle-invasive BLCA. Approximately 70–80% of BLCA patients are diagnosed with NMIBC (Non-Muscle Invasive Bladder Cancer) at their initial presentation [4]. Although the prognosis for NMIBC patients has significantly improved, more than 60% experience recurrence, and over 20% may progress to MIBC (Muscle Invasive Bladder Cancer), which is associated with higher mortality and metastasis rates [5]. The limited treatment options for invasive and metastatic BLCA present a significant clinical challenge. To date, the field of BLCA still lacks biomarkers that can effectively predict and guide treatment [6]. However, with the advancement of multi-omics technologies, research focused on the molecular characteristics of BLCA has shown great promise. Identifying prognostic biomarkers related to treatment is crucial for advancing the molecular subtyping of BLCA and improving therapeutic outcomes. Consequently, the quest for more reliable prognostic and therapeutic biomarkers has become a pressing priority in biomedical research.

Eukaryotic translation initiation factor 5A (EIF5A1) is a small yet important protein that is highly abundant and remarkably conserved across evolutionary lines [7]. In humans, EIF5A exists in two distinct isoforms, encoded by the EIF5A1 and EIF5A2 genes, located on chromosomes 17 and 3, respectively [8]. The coding sequences of EIF5A1 and EIF5A2 exhibit an 80% homology, with the proteins they produce demonstrating 84% sequence similarity and an overall similarity of 94% [9]. Notably, EIF5A1 and EIF5A2 are unique in their post-translational modifications, specifically the incorporation of a distinctive hypusine residue, which is critical for their functional efficacy [10]. Notably, elevated levels of these two isoforms are characteristic of various cancers and tumor-derived cell lines. Increasing evidence indicates that EIF5A1 is closely associated with cell proliferation, cancer progression, invasiveness, metastasis, and poor clinical outcomes [11]. Furthermore, current research has demonstrated that both the deficiency and overexpression of EIF5A1 can impact mitochondrial function, and disruption of EIF5A1 expression can reduce apoptosis. However, the specific mechanism remains unknown [12]. Additionally, the upstream transcription factors regulating EIF5A1 have yet to be definitively identified.

TEA domain (TEAD) transcription factors are pivotal in regulating tissue regeneration and cell proliferation [13, 14]. Within the TEAD family, TEAD4 contributes to these processes by engaging with transcriptional co-activators, notably YAP, to exert its biological effects [15]. The role of TEAD4 in tumorigenesis and cancer progression is increasingly being uncovered, making it a novel molecular marker for predicting cancer prognosis. The JAK/STAT signaling pathway is integral to the function of a diverse array of cell types, encompassing stem cells, ocular cells, hepatocytes, neurons, mammary cells, lymphocytes, adipocytes, and cardiomyocytes. Disruption of this pathway is strongly implicated in the development and progression of various diseases, including cancer. STAT proteins, as key regulators within this pathway, orchestrate critical cellular processes such as apoptosis, differentiation, and stem cell maintenance. Investigating the direct impact of JAK/STAT signaling on tumor cells, its underlying mechanisms, and its interactions with other signaling pathways offers promising avenues for the advancement of JAK/STAT-based therapeutic strategies in oncology [16, 17].

This study aims to investigate the role of EIF5A1 in the malignant progression of BLCA and to elucidate the underlying molecular mechanisms involved. Our research demonstrates that EIF5A1 is significantly overexpressed in BLCA tissues and cells, identifying it as a novel prognostic marker for disease progression and patient survival. We have shown that the transcription factor TEAD4 binds to the promoter region of EIF5A1, thereby enhancing its transcription. Furthermore, EIF5A1 plays a critical role in regulating the JAK2/STAT3 pathway, which is essential for modulating mitochondrial apoptosis in BLCA cells. These findings suggest that the YAP/TEAD4-EIF5A1-JAK2/STAT3 signaling axis is instrumental in the malignant transformation of BLCA. Consequently, EIF5A1 emerges as a promising biomarker and potential therapeutic target for BLCA.

Material and methods

Patient sample collection

Between May 2019 and November 2021, paired tumor and adjacent normal tissue samples were obtained from 50 patients diagnosed with BLCA at the Second Hospital of Lanzhou University. In January 2023, a follow-up evaluation was completed, with follow-up periods varying from 2 to 24 months. All tissue specimens were supplied by the Department of Pathology at the Second Hospital of Lanzhou University. The study protocol was reviewed and approved by the Ethics Committee of the same institution, and written informed consent was secured from all patients prior to surgical intervention.

Differential genes were identified through an integrated analysis of transcriptomics and urinary proteomics

We performed statistical analysis and visualization of urinary proteomics data from 5 BLCA patients and 5 healthy controls using R software (version 3.6.3). Transcriptomic data of BLCA patients (tumor = 414, normal = 19) were downloaded from the TCGA database, with data preprocessing and acquisition conducted using the “GEOquery” package. Gene volcano plots were visualized using the “ggplot2” package, and differential gene ranking plots were generated with the “limma” package to identify highly expressed genes. Subsequently, the urinary proteomics data were compared with the TCGA transcriptomic data to screen for differentially expressed genes that are highly expressed across both omics datasets. A Venn diagram was created using the “ggplot2” package to further filter these highly expressed intersecting genes. The identified highly expressed differential genes from the multi-omics analysis were then visualized using the “ComplexHeatmap” package. Finally, differential expression analysis of the target genes across various cancer types was visualized with the “ggplot2” package.

Clinical correlation and prognostic analysis of EIF5A1 in BLCA

Initially, we detected a notable differential expression of EIF5A1 in BLCA. To evaluate the diagnostic potential of EIF5A1, we conducted Receiver Operating Characteristic curve analysis using the “pROC” and “ggplot2” software packages. Furthermore, Kaplan–Meier survival curves were constructed to determine the prognostic significance of EIF5A1 in patients with BLCA.

Transcriptomic RNA sequencing

For transcriptomic RNA sequencing, we employed TRIzol reagent (Invitrogen) to process EIF5A1 knockout J82 cells and control J82 cells, with the processed samples immediately transferred to liquid nitrogen for storage. All subsequent experimental procedures were conducted by Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) detail steps (Supplementary Material 1).

Cell culture

This study was conducted with the support of the Gansu Provincial Clinical Research Center for Urological Diseases, which supplied the human cell lines. The SV-HUC-1 human ureteral epithelial cell line was cultured in Ham’s F12K medium, which was procured from Shanghai Yuanpei Biotechnology, China. BLCA cell lines, including T24, 5637, UMUC-3, and J82, were maintained in RPMI-1640 medium, also from the same supplier. The 293 T cell line was cultured in DMEM medium, which was sourced from Shanghai Yuanpei Biotechnology, China. All media were supplemented with 10% fetal bovine serum (PAN Biotech, Germany) and 5% antibiotics, consisting of 100 U/ml penicillin G and 100 µg/ml streptomycin (Solarbio Science & Technology Co., Ltd., Beijing, China). Cell lines were incubated in a humidified environment at 37 °C with 5% CO₂.

CRISPR/Cas9 gene editing technology and the construction of stable cell lines

To construct the EIF5A1 knockout UMUC-3 and J82 cell lines, we employed the Q5 Site-Directed Mutagenesis Kit (E0552S, New England Biolabs) to design and construct two sgRNA sequences targeting EIF5A1 (sg1-EIF5A1: 5′-CACCTCTCGACGATCTTACATGGCGTTT-3′; sg2-EIF5A1: 5′-CACCGGCTCGGGTCCTAATCACCCGTTT-3′). These sgRNAs were subsequently cloned into the lenti-CRISPR v2 vector (Addgene). Virus particles were produced by co-transfecting HEK293T cells with psPAX2 and pMD2.G vectors along with the target plasmid, using linear polyethylenimine (PEI) (40816ES02, China) as the transfection reagent. Viral supernatants were collected at 48 and 72 h post-transfection and used to infect UMUC-3 and J82 cells. To establish stable cell lines, 5 µg/ml puromycin was added 72 h after infection, with selection maintained until stable cell populations were established. Single clones were isolated from 96-well plates using the limiting dilution method, and the gene editing status of these cell clones was confirmed by Sanger sequencing and western blot analysis. Additionally, we purchased the PLV3-CMV-EIF5A1-puro plasmid from Miaoling Bio (Wuhan, China) and employed a similar method to construct stable cell lines.

Quantitative RT-PCR (qRT-PCR)

Total RNA was extracted from each cell line using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and RNA concentrations were normalized by spectrophotometry. Following the AJ reverse transcription kit protocol, 1 µg of total RNA was reverse transcribed into cDNA. qRT-PCR was then performed using the BIO-RAD CFX-96 system, with data analyzed via the 2-ΔΔCt method. The primers used in this study are listed in Supplementary Table S1.

Protein extraction and Western blot

Total proteins were extracted using RIPA lysis buffer supplemented with protease inhibitors (cat# P0013B, Beyotime). Protein concentrations were quantified using the bicinchoninic acid (BCA) assay, and the samples were subsequently subjected to electrophoretic separation. Proteins separated by SDS-PAGE were transferred onto PVDF membranes, which were then blocked with 6% non-fat milk. Primary antibody incubation was performed overnight at 4 °C.The western blot analysis was performed using an automated chemiluminescence imaging system (Tanon 5200 Multi) and the corresponding secondary antibody (cat#AC005, Abconal). The primary antibodies used in this study are listed as follows: β-actin (cat#66009-1-Ig, Proteintech); GAPDH (cat#60004-1-Ig, Proteintech); EIF5A1 (cat#11309-1-AP, Proteintech); Snail (cat#12129-1-AP, Proteintech); E-cadherin (cat#20874-1-AP, Proteintech); N-cadherin (cat#22018–1-AP, Proteintech); Vimentin (cat#10366-1-AP, Proteintech); Bcl-2 (cat#12789-1-AP, Proteintech); Cytochrome c (cat#10993-1-AP, Proteintech); cleaved-Caspase3 (cat#A25309, Abclonal); phospho-STAT3 (pY705) (cat#AP0705, Abclonal); STAT3 (cat#A1192, Abclonal); phospho-JAK2 (pY1007) (cat#AP0531, Abclonal); JAK2 (cat#AF6022, Affinity).

Colony formation assay and wound-healing assay

For each well of a 6-well plate, 2 mL of culture medium was added, and 1,000 cells were seeded per well. Following incubation at 37 °C in a 5% CO₂ atmosphere for 8 to 10 days, the cells were fixed with 4% paraformaldehyde (cat# BL539A, Biosharp) and stained with 0.1% crystal violet solution (cat# G1063, Solarbio). Additionally, approximately 6 × 10⁵ transfected cells were seeded into 6-well culture plates. Following cell adhesion, a scratch was created in each well using a 200 μL pipette tip, and the original medium was replaced with serum-free medium. Images were captured at 24-h intervals, and the extent of scratch closure was assessed over the subsequent 48 h.

Cell migration/invasion assay

In the migration assay, 100 μL of serum-free medium containing 1 × 10^5 BLCA cells was introduced into the upper chamber of a transwell apparatus with an 8 μm pore membrane (pore size: 8 µm; Corning Life Sciences, Tewksbury, MA, USA). The lower chamber was filled with 600 μL of RPMI-1640 medium supplemented with 20% fetal bovine serum. For the invasion assay, the membrane surface of the upper chamber was pre-coated with Matrigel (diluted at a ratio of 1:8; cat# 356234, Corning). The cells were then incubated at 37 °C in a 5% CO₂ atmosphere for 1 to 2 days. Cells that migrated or invaded to the underside of the membrane were fixed with 4% methanol, stained with crystal violet (cat# G1063, Solarbio), and examined under an inverted microscope.

3D matrigel drop invasion assay

Transfer 2 × 10⁴ cells into a 1.5 mL Eppendorf tube and centrifuge at 1000 g for 5 min at 4 °C. Carefully remove the supernatant. Slowly add thawed Matrigel to the cell pellet in the Eppendorf tube and mix thoroughly. Carefully dispense 10 μL of the Matrigel-cell suspension into the center of each well in a 24-well plate. Incubate the 24-well plate at 37 °C in a 5% CO₂ incubator for 15 min to allow the droplets to solidify. After solidification, gently add 1 mL of complete culture medium to each well by pipetting along the sidewall. Continue incubating the 24-well plate and monitor the development process daily under a microscope.

Cell cycle analysis and flow cytometry for apoptosis

In the cell cycle analysis, cells were prepared in accordance with the protocol provided by the cell cycle analysis kit (Yeasen). The stained cells were subsequently analyzed using a Beckman Coulter flow cytometer. For the apoptosis assay, procedures adhered strictly to the manufacturer’s instructions utilizing the Annexin V-FITC/PI apoptosis detection kit (Cat# AP101, Multi Sciences). Apoptotic cells were quantified with a CytoFLEX S flow cytometer (Beckman, USA). The overall apoptosis rate was determined as the aggregate of early and late apoptotic cells.

JC-1 and TMRE staining

Mitochondrial membrane potential was evaluated using the JC-1 staining kit (Beyotime Biotech, Nantong, China) following the manufacturer’s protocol. Cells were pretreated and then incubated with the JC-1 working solution (1 ×) at 37 °C for 20 min before being imaged with a confocal microscope. Additionally, mitochondrial membrane potential was assessed using the TMRE probe (Beyotime Biotech, Nantong, China). After seeding cells into confocal dishes and allowing them to adhere, they were incubated with TMRE at 37 °C for 20 min, followed by imaging with a confocal microscope.

Immunofluorescence assay

According to the experimental protocol, approximately 1 × 10⁵ cells were seeded into confocal microscopy dishes. Once the cells had adhered, they were fixed with 4% paraformaldehyde for 30 min. After fixation, the cells were permeabilized at room temperature for 5 min using 0.1% Triton X-100 in PBS. Blocking was performed with blocking buffer (Cat# G2052, Servicebio) for 20 min. Subsequently, the cells were incubated overnight at 4 °C with a primary antibody diluted 1:200. The following day, after washing with PBS, the cells were incubated with a secondary antibody for 1 h at room temperature in the dark. Finally, the nuclei were stained with DAPI (Cat# GDP1024, Servicebio) for 3 min, and the cells were observed and imaged using a confocal microscope.

Chromatin immunoprecipitation (ChIP) assay and Dual-luciferase reporter assay

Detection of TEAD4 Binding Sites in the EIF5A1 Promoter Region Using ChIP-qPCR: ChIP-qPCR experiments were conducted using J82 and UMUC-3 cells according to the standard protocol provided by the ChIP kit (cat#RK20258-1, ABclonal). Immunoprecipitation was performed using TEAD4 and IgG antibodies. Following the purification of the ChIP products, qRT-PCR was employed to assess the enrichment of specific sites within the EIF5A1 promoter region. The PCR primers for the three binding sites were designed as follows: TEAD4 Binding Site 1 (TEBS1): Forward (F)—AGCTCGTCTCCAGCTTCATG, Reverse (R)—AGGAGACACAGCGAGAGACT; TEAD4 Binding Site 2 (TEBS2): Forward (F)—TTGGTCGCTTGTTTCTGTCT, Reverse (R)—ATACAAAAATTAGCCGTGCCGT; TEAD4 Binding Site 3 (TEBS3): Forward (F)—CCCACCTTGACCTCCCAAAA, Reverse (R)—TGATGCCTGAACTTCCCAAT.

The EIF5A1 promoter reporter gene vector or the empty vector (pGL3B) was co-transfected into 293 T cells with or without the Overexpression YAP-expressing pcDNA3.1 plasmid using PEI as the transfection reagent. Forty-eight hours post-transfection, firefly and renilla luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Cat# 11402ES60, Yeasen). The relative activity of luciferase was determined using the luciferase assay.

Hematoxylin and Eosin (H&E) staining and Immunohistochemistry (IHC) experiment

In the H&E staining procedure, fresh tissue samples were initially fixed in 10% neutral buffered formalin for 48 h to ensure thorough protein denaturation and coagulation. Following fixation, the samples were dehydrated through an ethanol gradient, cleared with xylene for 30 min to achieve transparency, and then embedded in paraffin. The paraffin was subsequently removed from the sections using xylene, and the samples were rehydrated through an ethanol gradient before staining with H&E. After staining, the sections were treated with neutral resin and mounted for microscopic analysis and imaging.

In the IHC procedure, tissue specimens were fixed in formalin, embedded in paraffin, and sectioned into 4 μm thick slices. The processing adhered to the guidelines of the UltraSensitive™ S-P IHC Kit (Maixin Biotechnology Development Co., Ltd., #KIT-9710). This included deparaffinization of the sections with xylene, followed by rehydration through a series of ethanol solutions. The sections were then immersed in 0.01 M sodium citrate buffer, subjected to heat-induced epitope retrieval under pressure for 1 h, and allowed to cool before being washed with phosphate-buffered saline (PBS). The samples were incubated overnight at 4 °C with primary antibodies. After three washes with PBS, the sections were incubated at room temperature with a biotin-labeled anti-mouse secondary antibody and streptavidin for 30 min, followed by staining with freshly prepared DAB solution (Solarbio, China). Finally, the sections were counterstained with hematoxylin and examined under an optical microscope.

Xenograft tumor model

All procedures involving male nude mice in this study were conducted in accordance with the guidelines approved by the Ethics Committee of the Second Hospital of Lanzhou University (Approval No: D2024-736). Ten nude mice, aged four weeks, were randomly assigned to two groups, with five mice per group. Following a 7-day acclimatization period, each mouse was administered a subcutaneous injection into the right axillary region. The injection comprised 100 μL of serum-free medium containing 50 μL of Matrigel matrix and 5 × 10⁶ UMUC-3 cells, with the cells either corresponding to the NC-EIF5A1 or KO-EIF5A1 group. Throughout the experimental timeline, body weight and tumor volume were measured at 4-day intervals. On day 28, the mice were euthanized through the administration of an overdose of pentobarbital. Tumors were excised post-mortem, weighed, measured for volume, and photographed. Tumor volume was calculated using the formula: (length × width²) × 0.5.

Statistical analysis

Data analysis was conducted with GraphPad Prism version 9.0. For survival outcomes evaluation, we employed Kaplan–Meier methodology. All experimental procedures were independently repeated three times, with results expressed as mean values along with standard deviations. Comparative analyses between groups were performed utilizing Student’s t-test methodology. The relationships between different gene expression patterns were evaluated through Pearson’s correlation coefficients. In terms of significance levels, we adopted the following notation system: non-significant (n.s.); * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. P values < 0.05 were considered statistically significant.

Results

Identification of key genes in BLCA diagnosis and prognosis through integrated multi-omics analysis

Our research group successfully extracted protein from the urine of BLCA patients and healthy controls, followed by LC–MS analysis [18, 19]. Building on these data, we aim to explore non-invasive diagnostic biomarkers for BLCA and identify genes with differential expression. Initially, we utilized a volcano plot to screen for differentially expressed genes in TCGA-BLCA transcriptomic data (Fig. 1A). To increase the reliability of the selected biomarkers, we integrated urine proteomics data. Through the urine proteomics volcano plot (Fig. 1B) and the differential distribution map from the multi-omics analysis (Fig. 1C), we identified highly sensitive differential molecules. Given the study’s objective to identify molecules that can be effectively secreted into urine and used for early diagnosis, we focused on differentially expressed molecules that showed upregulation across various omics analyses. A Venn diagram was used to illustrate the overlap of highly expressed differential molecules between urine proteomics and TCGA-BLCA transcriptomics (Fig. 1D). The analysis revealed that 352 highly expressed differential molecules were identified in urine proteomics (Fig. 1D, left, and Table S2), while 3895 differentially expressed genes were identified in the TCGA-BLCA transcriptomic dataset (Fig. 1D, right), with an overlap of 32 molecules between these two datasets (Fig. 1D, center). To construct the heatmap visualization (Fig. 1D, right), we performed a analysis on the TCGA dataset, which comprised 414 tumor samples and 19 normal tissue samples. We strategically selected 13 most representative cases from each group (tumor and normal) based on their differential expression significance. The heatmap illustrates the expression patterns of the top 80 differentially expressed genes, which were identified through stringent statistical criteria. Notably, EIF5A1 exhibited significant differential expression across multiple omics analyses. To further investigate EIF5A1 expression across various cancers, we analyzed the TCGA database and found that EIF5A1 is highly expressed in the majority of cancer types (Fig. 1E). Additionally, pan-cancer paired expression analysis revealed significant upregulation of EIF5A1 (Fig. 1F). These findings suggest that EIF5A1 may play a crucial biological role in BLCA and other cancers, with potential as an early diagnostic biomarker.

Fig. 1.

Identification of Differentially Expressed Genes in Bladder Cancer (BLCA) Through Integrated Multi-Omics Analysis (A) Volcano plot displaying differentially expressed genes from the TCGA-BLCA gene expression profile (|Log2 fold change|≥ 1.0 and adjusted p-value < 0.05). B Volcano plot illustrating differentially expressed genes from the urine proteomics dataset (|Log2 fold change|≥ 0.5 and p-value < 0.05). C Curve demonstrating the relationship between the rank and expression abundance of differentially expressed genes across the integrated TCGA-BLCA and urine proteomics data (the y-axis represents the Log2 Fold Change values derived from the combined analysis of these two data sources). D Venn diagram highlighting the intersection of highly expressed differentially expressed genes in the urine proteomics data (left) and the TCGA-BLCA gene expression profile (right), along with their respective top-ranked differentially expressed genes. E Expression levels of EIF5A1 across various cancer types. F Paired expression levels of EIF5A1 in tumor versus normal tissues across different cancer types

EIF5A1 is overexpressed in BLCA and correlates with poor prognosis

Through a comprehensive analysis of the TCGA-BLCA dataset, we found that EIF5A1 expression is significantly elevated in BLCA tissues compared to normal tissues (Fig. 2A). This difference in expression remained statistically significant when comparing cancerous tissues to adjacent non-tumor tissues (Fig. 2B). To assess the potential diagnostic value of EIF5A1, we conducted a ROC curve analysis, which revealed that EIF5A1 exhibits high sensitivity and specificity for diagnosing BLCA, with an area under the curve (AUC) of 0.805 (Fig. 2C). Additionally, in the TCGA-BLCA dataset, patients with high EIF5A1 expression demonstrated significantly lower overall survival compared to those with low expression, indicating a poorer prognosis (Fig. 2D).

Fig. 2.

Differential Expression of EIF5A1 in BLCA and Its Correlation with Clinicopathological Features (A) Expression levels of EIF5A1 in BLCA tissues compared to normal tissues. B Paired comparison of EIF5A1 expression levels between BLCA and matched normal tissues. C ROC curve assessing the sensitivity and specificity of EIF5A1 as a predictive marker for BLCA. D Relationship between EIF5A1 expression levels and overall survival of BLCA patients (TCGA-BLCA dataset). E Immunohistochemical staining of EIF5A1 in cancerous and adjacent normal tissues. F Immunohistochemical staining of EIF5A1 in low-grade (Ta) and high-grade (T3) BLCA tissues. G Expression levels of EIF5A1 across different pathological stages in TCGA-BLCA dataset. H, I Kaplan–Meier survival analysis of 50 BLCA patients based on EIF5A1 expression levels, depicting overall survival (H) and recurrence-free survival (I). Statistical significance is denoted as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

IHC analysis was conducted on 50 BLCA specimens to delineate the expression patterns of EIF5A1 across cancer tissues, adjacent non-tumor tissues, as well as in low-grade and high-grade BLCA (Fig. 2E/F). The analysis revealed a notable upregulation of EIF5A1 expression in tumor tissues, with particularly high levels observed in high-grade BLCA compared to adjacent non-tumor tissues and low-grade cancers. Further analysis of EIF5A1 expression across different pathological stages revealed increased expression in advanced stages (II, III, and IV) compared to Stage I (Fig. 2G), providing additional evidence for its association with disease progression (TCGA dataset). Based on median EIF5A1 expression levels, patients were classified into low and high expression groups. Table 1 details the correlation between EIF5A1 expression and various clinical parameters, demonstrating an association with tumor grade and T stage. Kaplan–Meier survival analysis further indicated that patients with elevated EIF5A1 expression experienced markedly shorter overall survival and recurrence-free survival compared to those with lower EIF5A1 levels (Fig. 2H/I). These results highlight the potential of EIF5A1 as a prognostic biomarker, suggesting that increased EIF5A1 expression is closely linked to adverse prognosis and a higher risk of cancer recurrence.

Table 1.

Correlations between EIF5A1 and clinicopathological characteristics

Characteristics	EIF5A1 expression		p value
	High (25 cases)	Low (25 cases)
Age (years)			0.7820
≥ 65	13	14
≤ 65	12	11
Gender			0.6909
Male	21	22
Female	4	3
Tumor size (cm)			0.2572
≥ 3	12	8
≤ 3	13	17
Multiplicity			0.4507
Single	20	22
Multiple	5	3
Tumor grade			< 0.0001
High	20	2
Low	5	23
Tumor stage
T stage			< 0.0001
T0-T1	4	23
T2-T4	21	2
N stage			0.3223
N0	24	25
N1-N2	1	0
M stage			NA
M0	25	25
M1-M2	0	0

EIF5A1 expression is elevated in BLCA cells compared to normal cells

Initially, we analyzed the expression of EIF5A1 across five cell lines (SV-HUC-1, UMUC-3, J82, T24, 5637) using qRT-PCR and western blot. The qRT-PCR results demonstrated a significant upregulation of EIF5A1 in BLCA cell lines (Fig. 3A), which was further corroborated by western blot analysis (Fig. 3B). Subsequently, we conducted a comparative analysis of EIF5A1 expression in paired adjacent normal tissues (Pa) and tumor tissues (Ca) from five BLCA patients. The results revealed that EIF5A1 expression was higher in tumor tissues compared to corresponding adjacent normal tissues in all patients (Fig. 3C). To further investigate the functional role of EIF5A1, we successfully established EIF5A1 knockout stable cell lines in J82 and UMUC-3 using the Crispr-Cas9 technique, with knockout efficiency confirmed by western blot (Fig. 3D/E). Notably, despite the critical role of the EIFs family as translation initiation factors, EIF5A1 knockout cells remained viable, which is consistent with a previous study [20]. Lastly, we generated EIF5A1 overexpression stable cell lines in J82, UMUC-3, and T24, and validated the overexpression efficiency via western blot (Fig. 3F).

Fig. 3.

Expression of EIF5A1 in BLCA Cell Lines and Patient Tissues, and the Construction of Stable Cell Lines (A) qRT-PCR analysis showing the relative expression levels of EIF5A1 in five cell lines (SV, UMUC-3, J82, T24, and 5637). B Western blot analysis demonstrating the expression of EIF5A1 protein in these cell lines. C Comparative analysis of EIF5A1 expression between adjacent normal tissues (Pa) and tumor tissues (Ca) in five BLCA patients. D Transfection efficiency of lentivirus in J82 and UMUC-3 cells (left panel: bright field; right panel: GFP fluorescence). E Western blot analysis showing the expression of EIF5A1 in J82 and UMUC-3 cell lines after gene knockout using CRISPR-Cas9 technology. F Western blot analysis showing the protein expression levels of EIF5A1 in J82, UMUC-3, and T24 cell lines after overexpression of the EIF5A1 gene. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

EIF5A1 enhances proliferative and promotes cell cycle progression

The CCK-8 assay demonstrated that EIF5A1 knockout significantly inhibited cell proliferation in the J82 and UMUC-3 cell lines. In contrast, EIF5A1 overexpression markedly promoted cell proliferation in the J82, UMUC-3, and T24 cell lines (Fig. 4A/B). These results were further validated by the colony formation assay. EIF5A1 knockout notably reduced colony-forming ability in the J82 and UMUC-3 cell lines, whereas EIF5A1 overexpression significantly increased the number of colonies in the J82, UMUC-3, and T24 cell lines (Fig. 4C/D). Additionally, flow cytometry analysis revealed that EIF5A1 knockout led to a higher proportion of cells in the G1 phase and a lower proportion in the S phase, indicating a G1 phase cell cycle arrest (Fig. S1A/B).

Fig. 4.

The Critical Role of EIF5A1 in Cell Proliferation and Cell Cycle Regulation (A, B) The left two panels display cell proliferation in J82 and UMUC-3 cell lines following EIF5A1 knockout; the right three panels show cell proliferation in J82, UMUC-3, and T24 cell lines following EIF5A1 overexpression. C, D Results of the colony formation assay demonstrating the effects of EIF5A1 knockout and overexpression on the colony-forming ability of cells. Statistical significance is denoted as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

EIF5A1 knockout inhibits BLCA cell migration and invasion

The effect of EIF5A1 on cell migration was evaluated using a wound healing assay. The results showed that EIF5A1 knockout significantly inhibited the migration capacity of the J82 and UMUC-3 cell lines (Fig. 5A and Fig. S2A). Conversely, EIF5A1 overexpression markedly enhanced cell migration in the J82, UMUC-3, and T24 cell lines (Fig. 5B and Fig. S2B/C). To further validate these findings, transwell migration and invasion assays were conducted. EIF5A1 knockout significantly reduced both migration and invasion abilities in J82 and UMUC-3 cells (Fig. 5C and Fig. S2D).

Fig. 5.

The Role of EIF5A1 in Promoting Cell Migration, Invasion and EMT (A) Wound healing assay results showing changes in cell migration ability in J82 cells following EIF5A1 knockout. B Wound healing assay results showing changes in cell migration ability in J82 cells following EIF5A1 overexpression. C Transwell migration and invasion assay results showing that EIF5A1 knockout significantly inhibits cell migration and invasion abilities in J82 cells. D Transwell migration and invasion assay results showing changes in cell migration and invasion abilities in J82 cells following EIF5A1 overexpression. E Western blot analysis showing changes in the expression of EMT-related proteins (Snail, N-cadherin, E-cadherin, Vimentin) following EIF5A1 knockout. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

In contrast, the overexpression of EIF5A1 markedly enhanced the migration and invasion capabilities of J82, UMUC-3, and T24 cell lines (Fig. 5D and Fig. S2E/F). Additionally, a 3D Matrigel droplet invasion assay corroborated these findings, showing that EIF5A1 knockout substantially impeded the migration of J82 and UMUC-3 cell lines (Fig. S2G/H). Western blot analysis further revealed that EIF5A1 knockout resulted in a notable reduction in the expression of Snail, N-cadherin, and Vimentin, whereas E-cadherin expression was significantly elevated in J82 and UMUC-3 cell lines (Fig. 5E). These results suggest that EIF5A1 modulates the migration and invasion of BLCA cells, potentially through the regulation of epithelial-mesenchymal transition (EMT) markers.

The key role of EIF5A1 in regulating cell apoptosis suggests that EIF5A1 may affect apoptosis by regulating mitochondrial function

Based on transcriptome sequencing data, we conducted a functional analysis of differentially expressed genes, including GO annotations and KEGG pathway analysis, to identify genes associated with EIF5A1. The GO analysis revealed that a proportion of these genes are involved in biological regulation (Fig. 6A). Additionally, the KEGG pathway analysis highlighted the “Cell Growth and Death” pathway as being closely related to biological regulation (Fig. 6B). These findings suggest that EIF5A1 plays a pivotal role in regulating biological processes, particularly those related to cell growth and death in BLCA.

Fig. 6.

EIF5A1 Influences Cell Apoptosis by Regulating the Mitochondrial Apoptosis Pathway (A) GO annotations results of functional enrichment analysis for differentially expressed genes identified through transcriptome sequencing. B KEGG pathway analysis results for differentially expressed genes identified through transcriptome sequencing. C Flow cytometry analysis of apoptosis following EIF5A1 knockout. D Flow cytometry analysis of apoptosis following EIF5A1 overexpression. E, F Western blot analysis showing changes in the expression of apoptosis-related proteins (Bcl-2, cytochrome C, cleaved-Caspase-3, and Caspase-3) following EIF5A1 knockout and overexpression. G Representative images of mitochondrial membrane potential (Δψm) as indicated by JC-1 staining following EIF5A1 knockout. H Mitochondrial membrane potential changes as indicated by TMRE staining following EIF5A1 knockout. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

Building on these discoveries, we conducted further investigations. Flow cytometry analysis demonstrated that EIF5A1 knockout significantly increased the apoptosis rate in BLCA cells (Fig. 6C), while overexpression of EIF5A1 markedly reduced apoptosis (Fig. 6D). These observations were further validated by western blot analysis, which showed that EIF5A1 knockout significantly downregulated the expression of the anti-apoptotic protein Bcl-2 while upregulating the levels of apoptosis-related proteins, including cytochrome c and cleaved caspase-3 (Fig. 6E). Conversely, EIF5A1 overexpression enhanced Bcl-2 expression and reduced the levels of apoptosis-related proteins (Fig. 6F). Moreover, JC-1 staining assays revealed that EIF5A1 knockout decreased mitochondrial membrane potential in BLCA cells, indicating mitochondrial dysfunction and the induction of apoptosis (Fig. 6G). This result was further corroborated by TMRE staining assays, which demonstrated that EIF5A1 knockout led to a reduction in mitochondrial membrane potential, reflecting an increase in apoptosis (Fig. 6H).

The crucial role of EIF5A1 in regulating the JAK/STAT signaling pathway

Integration of transcriptome sequencing and database analysis has revealed the pivotal role of the EIF5A1 gene in regulating the JAK/STAT signaling pathway. KEGG enrichment analysis of differentially expressed genes obtained from transcriptome sequencing (Fig. 7A) and transcriptomic data from the TCGA BLCA cohort (Fig. 7B) both indicate a strong association between EIF5A1 and the downstream JAK/STAT pathway. Western blot analysis demonstrated that EIF5A1 knockout significantly reduced the phosphorylation levels of phospho-JAK2 (pY1007) and phospho-STAT3 (pY705) (Fig. 7C), suggesting inhibition of the JAK/STAT pathway. Conversely, overexpression of EIF5A1 markedly increased the phosphorylation of phospho-JAK2 (pY1007) and phospho-STAT3 (pY705), further confirming the regulatory role of EIF5A1 in the JAK/STAT signaling pathway (Fig. 7D). Immunofluorescence staining in J82 and UMUC-3 cell lines showed the localization and expression of EIF5A1, phospho-JAK2 (pY1007), and phospho-STAT3 (pY705). The reduction in phosphorylation levels of JAK2 and STAT3 upon EIF5A1 knockout was consistent with the results observed in the western blot analysis (Fig. 7E).

Fig. 7.

The Critical Role of EIF5A1 in Regulating the JAK-STAT Signaling Pathway (A) KEGG enrichment analysis of differentially expressed genes obtained from transcriptome sequencing. B KEGG enrichment analysis of transcriptome data from the TCGA database in the BLCA cohort. C Western blot analysis showing changes in the phosphorylation levels of JAK2 (pY1007) and STAT3 (pY705) following EIF5A1 knockout. D Western blot analysis showing changes in the phosphorylation levels of JAK2 (pY1007) and STAT3 (pY705) following EIF5A1 overexpression. E Immunofluorescence staining showing the localization and expression of phospho-JAK2 (pY1007) and phospho-STAT3 (pY705) in J82 and UMUC-3 cell lines following EIF5A1 knockout. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

EIF5A1 regulates mitochondrial apoptosis via the JAK2/STAT3 signaling pathway and influences BLCA progression

We further explored the role of EIF5A1 in regulating mitochondrial apoptosis and its impact on BLCA progression through the JAK2/STAT3 signaling pathway. Flow cytometry analysis revealed that EIF5A1 knockout, combined with the JAK/STAT pathway inhibitor Stattic, increased apoptosis rates in J82 and UMUC-3 cell lines (with 2 μmol/mL Stattic incubation for 24 h) (Fig. 8A). Conversely, in cells overexpressing EIF5A1, stattic treatment significantly restored the apoptosis rate (Fig. 8B), suggesting that EIF5A1 regulates apoptosis via the JAK2/STAT3 signaling pathway. Additionally, Additionally, mitochondrial membrane potential was restored in EIF5A1-overexpressing cells after stattic treatment (Fig. 8C/D), further supporting the role of EIF5A1 in modulating mitochondrial apoptosis through the JAK2/STAT3 signaling pathway, thereby affecting BLCA progression. Moreover, cell proliferation assays demonstrated that EIF5A1 overexpression promoted cell proliferation, while stattic effectively inhibited this proliferative effect (Fig. 8E). These findings collectively confirm that EIF5A1 regulates mitochondrial apoptosis via the JAK2/STAT3 signaling pathway, thereby influencing BLCA progression.

Fig. 8.

EIF5A1 Influences Mitochondrial Apoptosis and Regulates BLCA Progression Through the JAK2-STAT3 Signaling Pathway (A) Flow cytometry analysis of apoptosis in J82 and UMUC-3 cell lines after EIF5A1 knockout, followed by treatment with stattic (a JAK-STAT signaling pathway inhibitor, 2 μmol/mL, incubated for 24 h). B Flow cytometry analysis of apoptosis in J82 and UMUC-3 cell lines after EIF5A1 overexpression, followed by treatment with stattic. C, D Representative images of mitochondrial membrane potential (Δψm) as indicated by JC-1 staining in J82 and UMUC-3 cell lines after EIF5A1 overexpression and subsequent stattic treatment. E Cell proliferation assay showing the proliferation of cells after EIF5A1 overexpression and subsequent stattic treatment. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

YAP/TEAD4 transcriptionally upregulates EIF5A1 expression

To explore the binding relationship between TEAD4 and the promoter region of EIF5A1, we conducted an analysis using data from the KnockTF, ENCODE, and GTRD databases to identify potential transcription factors that could interact with the EIF5A1 promoter. As depicted in Fig. 9A, cross-referencing these three databases revealed 10 common transcription factors that may bind to the EIF5A1 promoter. Subsequent validation results indicated a significant positive correlation between TEAD4 and EIF5A1. Analysis of the TCGA and CCLE dataset revealed that, across various cancer types, the expression levels of TEAD4 and EIF5A1 exhibited a notable positive correlation. Specifically, in BLCA, the Pearson correlation coefficient was 0.456, with a p-value of less than 0.001, indicating a strong association between the two variables (Fig. 9B). Similarly, analysis of the CCLE dataset demonstrated a significant positive correlation between TEAD4 and EIF5A1 across multiple cancer types. In BLCA, the Pearson correlation coefficient reached 0.735, with a p-value also below 0.001, further supporting their correlation (Fig. 9C). These results suggest that the positive correlation between TEAD4 and EIF5A1 is widespread in various cancers, particularly BLCA. Additionally, analysis of the TCGA database showed that only TEAD4 expression levels were significantly elevated in BLCA patients compared to normal tissues (Fig. 9D). Survival analysis further revealed that high TEAD4 expression was significantly associated with poor prognosis in BLCA patients (Fig. 9E). Based on these findings, we selected TEAD4 as an upstream regulatory transcription factor of EIF5A1 for subsequent experiments.

Fig. 9.

TEAD4 is an upstream transcription factor of EIF5A1 (A) Venn diagram showing the intersection of predicted TEAD4 binding to the EIF5A1 promoter based on data from the KnockTF, ENCODE, and GTRD databases. B, C Pearson correlation analysis of TEAD4 and EIF5A1 expression levels in the TCGA and CCLE datasets. D Comparison of TEAD4 expression levels in normal tissues and BLCA patients from the TCGA database. E Kaplan–Meier survival curve illustrating the association between TEAD4 expression levels and overall survival in BLCA patients from the TCGA database. F Predicted TEAD4 binding motif sequence obtained from the JASPAR database. G Western blot analysis showing EIF5A1 expression levels in J82 and UMUC-3BLCA cell lines following treatment with verteporfin (2 μmol/mL and 4 μmol/mL) for 24 h. H qRT-PCR results demonstrating EIF5A1 expression in J82 and UMUC-3 BLCA cell lines after verteporfin treatment. I Luciferase reporter assay indicating that YAP overexpression significantly increased EIF5A1 promoter activity in 293 T cells. J, K Chromatin immunoprecipitation (ChIP)-PCR analysis showing significant enrichment of TEAD4 at the EIF5A1 promoter region in J82 and UMUC-3 cells. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

Using the JASPAR database, we predicted the binding motifs of TEAD4 within the EIF5A1 promoter region and identified potential binding sites (Fig. 9F). To verify the regulatory role of TEAD4 on EIF5A1 expression, we treated UMUC-3 and J82 cell lines with verteporfin, a compound known to disrupt the interaction between YAP and TEAD4. Western blot analysis revealed a reduction in EIF5A1 expression levels (Fig. 9G). qRT-PCR analysis further confirmed that EIF5A1 expression was significantly decreased in these two BLCA cell lines following verteporfin treatment (Fig. 9H). Moreover, a luciferase reporter assay demonstrated that YAP overexpression significantly enhanced the luciferase activity of the EIF5A1 promoter, indicating that YAP/TEAD4 positively regulates EIF5A1 expression through transcriptional activation (Fig. 9I). These results strongly suggest that EIF5A1 is a target gene of the YAP/TEAD4 co-transcriptional activation complex.

To further investigate the specific binding sites between TEAD4 and the EIF5A1 promoter, we utilized the JASPAR database to predict three potential binding sites within the EIF5A1 promoter region. Subsequently, ChIP-qPCR assays were conducted in J82 and UMUC-3 cells to validate the binding of TEAD4 at these sites. The results demonstrated significant enrichment of the TEBS1 sequence (AACATACCAC) in ChIP-derived DNA from both J82 and UMUC-3 cells (Fig. 9J/K). These findings provide compelling evidence for the critical role of TEAD4 in the transcriptional regulation of EIF5A1.

YAP/TEAD4 transcriptional activity affects EIF5A1 expression, regulating mitochondrial apoptosis and influencing BLCA progression

We investigated the role of YAP/TEAD4 in transcriptionally regulating EIF5A1 expression and its impact on mitochondrial apoptosis and BLCA progression. Flow cytometry analysis revealed that verteporfin treatment increased the apoptosis rate in the J82 and UMUC-3 cell lines. Similarly, in cells overexpressing EIF5A1, verteporfin treatment also increased the apoptosis rate, though to an extent (Fig. 10A/B). Cell proliferation assays demonstrated that verteporfin treatment significantly inhibited the proliferation of J82 and UMUC-3 cell lines, while in EIF5A1-overexpressing cells, verteporfin treatment restored proliferation (Fig. 10C). These findings suggest that YAP/TEAD4 influences BLCA progression by regulating EIF5A1 expression.

Fig. 10.

YAP/TEAD4 regulates EIF5A1 expression through transcriptional control, influencing mitochondrial apoptosis and promoting BLCA progression. A, B Apoptosis levels were assessed using flow cytometry in J82 and UMUC-3 cell lines following treatment with verteporfin. Apoptosis was also evaluated in EIF5A1-overexpressing J82 and UMUC-3 cell lines after verteporfin treatment. C Cell proliferation was analyzed in J82 and UMUC-3 cell lines after verteporfin treatment. Similarly, cell proliferation changes were evaluated in EIF5A1-overexpressing J82 and UMUC-3 cell lines following verteporfin treatment. D Western blot analysis showing changes in the expression of apoptosis-related proteins (Bcl-2, cytochrome C, cleaved Caspase-3, and Caspase-3) in J82 and UMUC-3 cell lines after verteporfin treatment. Changes in apoptosis-related protein expression were also assessed in EIF5A1-overexpressing J82 and UMUC-3 cell lines following verteporfin treatment. E Schematic diagram illustrating the regulatory role of YAP/TEAD4 in controlling EIF5A1 expression, activating the JAK2/STAT3 pathway, and modulating mitochondrial apoptosis, ultimately promoting BLCA progression. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

Furthermore, western blot analysis provided additional support for these observations. In the J82 and UMUC-3 cell lines, verteporfin treatment downregulated the expression of the anti-apoptotic protein Bcl-2 while upregulating the levels of apoptosis-related proteins, including cytochrome c and cleaved caspase-3. In cells overexpressing EIF5A1, verteporfin treatment restored Bcl-2 expression and enhanced the levels of cytochrome c and cleaved caspase-3, further emphasizing the critical role of YAP/TEAD4 in regulating mitochondrial apoptosis and influencing BLCA progression through the modulation of EIF5A1 expression (Fig. 10D). Finally, a schematic diagram summarizes these findings, illustrating how YAP/TEAD4 transcriptionally regulates EIF5A1 expression, modulating mitochondrial apoptosis via the JAK2/STAT3 signaling pathway, thereby influencing BLCA progression (Fig. 10E).

EIF5A1 regulates the JAK2/STAT3 signaling pathway, and its deletion in vivo significantly inhibits tumor growth

To investigate the role of EIF5A1 in BLCA progression, an in vivo xenograft experiment was conducted using the UMUC-3 cell line. The results indicated that EIF5A1 knockout (KO) significantly suppressed tumor growth. Compared to the NC-EIF5A1 group, the KO-EIF5A1 group exhibited markedly slower tumor growth, culminating in significantly reduced tumor volume by the end of the experiment (Fig. 11A-C). Furthermore, the tumor weight in the KO-EIF5A1 group was substantially lower than that of the control group (Fig. 11D). Western blot analysis revealed changes in the expression of apoptosis-related proteins, including Bcl-2, cleaved Caspase-3, and Caspase-3, in the tumor tissues of the KO-EIF5A1 group, compared to the NC-EIF5A1 group (Fig. 11E). IHC confirmed that the expression levels of EIF5A1, Ki67, Bcl-2, phosphorylated JAK2, and phosphorylated STAT3 were higher in the NC-EIF5A1 group than in the KO-EIF5A1 group (Fig. 11F). These findings suggest that EIF5A1 knockout effectively inhibits BLCA growth by modulating apoptosis-related signaling pathways.

Fig. 11.

Knockout of EIF5A1 in bladder cancer cells results in altered JAK2/STAT3 signaling pathways and affects the characteristics of xenografted tumors. A-D An in vivo xenograft experiment using the UMUC-3 cell line demonstrated that EIF5A1 knockout significantly reduced the rate of tumor growth. Tumor volume and weight were markedly lower in the KO-EIF5A1 group compared to the NC-EIF5A1 group. E Western blot analysis revealed significant changes in the expression of apoptosis-related proteins (Bcl-2, cleaved Caspase-3, and Caspase-3) in the KO-EIF5A1 group compared to the NC-EIF5A1 group. F Immunohistochemical staining showed that expression levels of EIF5A1, Ki67, Bcl-2, JAK2, and STAT3 were substantially higher in tumor tissues from the NC-EIF5A1 group than in the KO-EIF5A1 group. Statistical significance is indicated as follows: n.s., not significant; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001

Discussion

BLCA is a highly prevalent and lethal malignancy of the urinary system. It is the tenth most common cancer worldwide, with over 430,000 new cases diagnosed annually [2, 3]. Cystoscopy remains the gold standard for the initial diagnosis and monitoring of BLCA [21]. However, it is an invasive procedure that carries risks of complications and increases healthcare costs. In recent decades, urinary biomarkers have gained significant attention in both experimental and clinical research related to BLCA [22]. Advances in omics technologies and bioinformatics have broadened the scope of urinary biomarkers, enabling the detection of a wider range of markers for the diagnosis and monitoring of BLCA [23].

EIF5A2, another isoform of EIF5A, was not identified among the 352 upregulated differentially expressed proteins in our urinary proteomics analysis. Furthermore, analysis of TCGA database revealed that EIF5A2 was downregulated in bladder cancer patients and demonstrated no correlation with patient prognosis (Fig. S3). In recent years, the rapid advancement of RNA sequencing and proteomics technologies has led to the discovery and application of numerous biomarkers for tumor stratification and precision therapy [24, 25]. Notably, the integrative analysis of multi-omics data has significantly enhanced our understanding of the molecular dysfunctions driving tumorigenesis [26].

EIF5A1 plays a critical role in the initiation and progression of various cancers. Zhang et al. [27] reported that EIF5A1 is highly expressed in epithelial ovarian cancer (EOC) tissues, with elevated expression significantly associated with poor patient prognosis. Their experimental results demonstrated that EIF5A1 promotes cancer cell proliferation, migration, invasion, and EMT. The researchers suggested that EIF5A1 could serve as a potential biomarker and therapeutic target for EOC. Furthermore, in pancreatic cancer, EIF5A1 regulates YAP1/TAZ expression by modulating PEAK1, which in turn activates stem cell-associated transcription factors, including Oct4, Nanog, and c-Myc, thereby promoting cancer cell growth [28]. In cervical cancer, EIF5A1 was found to be overexpressed, and inhibition of EIF5A1 modifications led to altered expression of cancer-related genes and a significant reduction in cancer cell proliferation [11]. Collectively, these findings suggest that EIF5A1 is not only a potential prognostic biomarker for various cancers but also a key player in tumorigenesis. In our study, Kaplan–Meier survival analysis using clinical data from 50 BLCA patients revealed that high EIF5A1 expression is significantly associated with poor prognosis (Fig. 2H/I). This further supports the idea that elevated EIF5A1 expression may be closely linked to unfavorable outcomes in BLCA patients, positioning EIF5A1 as a potential prognostic biomarker and therapeutic target.

The overexpression of TEAD4 in BLCA cells and its regulatory role in cellular processes have attracted significant attention. Previous studies have demonstrated that the transcription factor TEAD4 is closely associated with poor prognosis in BLCA and promotes cancer cell migration by regulating EMT via the PI3K/AKT pathway [29]. These findings have preliminarily elucidated TEAD4’s role in BLCA progression through specific regulatory mechanisms. In our study, we identified a novel downstream target of TEAD4, EIF5A1. Using a dual-luciferase reporter assay, we confirmed that TEAD4 binds to the promoter sequence of EIF5A1 and enhances its expression. It is important to note that YAP, lacking a DNA-binding domain, requires association with transcription factors to exert its influence on DNA. Previous research has shown that YAP primarily mediates its biological functions through the TEAD family of transcription factors. When YAP is activated, it translocates to the nucleus, where it interacts with transcription factors such as TEAD, promoting tumor growth, metastasis, and cellular senescence in various cancer cell lines [30–33]. Our study is the first to demonstrate that TEAD4 specifically binds to the AACATACCAC sequence in the EIF5A1 promoter, thereby enhancing its transcriptional activity. The identification of this regulatory relationship between TEAD4 and EIF5A1 is a novel finding of our research. Through dual-luciferase reporter assays and ChIP experiments, we confirmed this interaction for the first time. Furthermore, we showed that treatment with the YAP/TEAD4 inhibitor verteporfin significantly inhibited BLCA cell proliferation and induced apoptosis via the mitochondrial pathway. These findings offer new insights into the mechanistic role of EIF5A1 in BLCA.

Various studies have reported that both the deficiency and overexpression of EIF5A1 have detrimental effects on mitochondrial function. For example, one study found that in rat cardiomyocytes, overexpression of EIF5A1 induced by the chemotherapeutic drug doxorubicin led to a gradual increase in mitochondrial reactive oxygen species and an associated increase in Ca2 + influx [34]. These changes were closely linked to a loss of mitochondrial membrane potential and the induction of apoptosis, while inhibition of EIF5A1 expression significantly reduced apoptosis. These results are consistent with findings by Sun et al., who demonstrated that EIF5A1 overexpression in human cells, stimulated by viral infection, produced similar outcomes. Similarly, overexpression of hypusination-deficient EIF5A1 mutants also induced apoptotic cell death through the intrinsic mitochondrial pathway. Elevated levels of EIF5A1 resulted in the loss of mitochondrial membrane potential, along with the translocation of the apoptosis marker Bax to the mitochondria, release of cytochrome c, and caspase activation. Moreover, proteomic analysis of HeLa cells overexpressing EIF5A1 revealed a significant upregulation of mitochondrial-associated proteins. Collectively, these findings suggest that abnormal regulation of EIF5A1 can lead to mitochondrial dysfunction and apoptosis [12, 35]. Based on transcriptome sequencing data, we performed a functional analysis of differentially expressed genes, including GO annotation and KEGG pathway analysis, which indicated that EIF5A1 plays a pivotal role in biological processes, particularly in cancer cell growth and death. Further flow cytometry analysis revealed that EIF5A1 knockout significantly increased apoptosis rates in BLCA cells. This was confirmed by Western blot analysis, which showed that EIF5A1 knockout significantly downregulated the anti-apoptotic protein Bcl-2 while upregulating apoptosis-related proteins, including cytochrome c and cleaved caspase-3. In addition, JC-1 and TMRE staining assays demonstrated that EIF5A1 knockout significantly reduced mitochondrial membrane potential in BLCA cells.

The JAK/STAT signaling pathway is essential for regulating cell apoptosis, differentiation, and stem cell maintenance [16, 17, 36]. Recent studies have highlighted the involvement of eukaryotic translation initiation factor 3F (EIF3F) in cell migration, invasion, and energy metabolism in lung cancer cells. Overexpression of EIF3F in these cells has been shown to enhance cell migration and invasion by interacting with STAT3 and modulating a series of genes associated with metastasis [37]. Our research, which integrates transcriptome sequencing with database analysis, underscores the critical role of EIF5A1 in regulating the JAK/STAT signaling pathway. Western blot analysis revealed that EIF5A1 knockout significantly reduced the phosphorylation levels of JAK2 and STAT3, indicating inhibition of the JAK/STAT pathway. Additionally, treatment with Stattic in EIF5A1-overexpressing cells restored apoptosis rates to normal levels. JC-1 staining further demonstrated that the mitochondrial membrane potential in EIF5A1-overexpressing cells was significantly restored, suggesting that EIF5A1 is crucial for regulating mitochondrial apoptosis. Collectively, these findings indicate that EIF5A1 regulates mitochondrial apoptosis through the JAK2/STAT3 signaling pathway, thereby influencing BLCA progression. Consequently, EIF5A1 may emerge as a promising new target for therapeutic intervention.

Conclusions

In summary, our research findings revealed the critical role of EIF5A1 in the progression of BLCA. While our findings indicate a possible regulatory mechanism by which YAP/TEAD4 transcriptionally regulates EIF5A1, and suggest that this regulation influences BLCA progression through the JAK2/STAT3 signaling pathway and mitochondrial apoptosis, further studies are needed to fully understand the molecular mechanisms underlying EIF5A1’s role in bladder cancer.

Authors’ contributions

PKL, SW, and HSL contributed to the conception and design of the study. PKL and SW were primarily responsible for implementing the experiments and analyzing the data. PKL, SW, YCW, YSC, and LW conducted several cell experiments and interpreted the results. LY and HSL provided essential reagents and assisted with experimental design and data analysis. PKL and SW drafted and carefully revised the manuscript. PKL and LY made critical revisions. All authors participated in manuscript revision and approved the final submitted version.

Funding

This study was supported by the Cuiying Scientifc and Technological Innovation Program of Lanzhou University Second Hospital (Grant numbers CY2022-MS-A09); Traditional Chinese Medicine Scientific Research Project of Gansu Province, China (GZKP-2023-36); Investigation of the Effects and Mechanisms of HTRA1 on Bladder Tumors Through Combined Non-invasive Urinary Metabolomics (Grant numbers 22YF7FA090); Research Based on Novel Optical Fiber Biosensors for Multi-Parameter High-Sensitivity Real-Time Detection of Bladder Cancer Urinary Biomarkers (Grant numbers 24YFFA049); The Special Fund Project for Central Guiding Local Science and Technology Development (Grant numbers 24ZYQA050).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Ethics approval and consent to participate Ethical approval was obtained from the Institutional Review Board and Human Ethics Committee of the The Second Hospital of Lanzhou University for this study (No. 2024A-946). Informed consent was obtained from all the participants and/or their legal guardians. In addition, the design and conduct of the animals for this study were approved (No. D2024-736).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.