PAK5 Promotes Esophageal Squamous Cell Carcinoma Progression Revealed by Transcriptomic Profiling

Globally, esophageal cancer ranks among the most lethal malignancies. In China, esophageal squamous cell carcinoma (ESCC) represents the predominant histological subtype. Although therapeutic advances have been achieved, the prognosis for ESCC patients continues to be dismal, manifesting as a mere 10% to 25% 5-year survival rate¹. This is largely due to the challenges of early detection and the limited availability of effective molecularly targeted therapies². Unlike other solid tumors where targeted therapies have shown significant promise, such approaches in ESCC are still in their infancy, hampered by a scarcity of actionable molecular targets.

PAK5, a group II PAK, integrates cytoskeletal dynamics and pro-survival signaling and has been connected to chemoresistance in several malignancies^{3,4,5,6,7,8,9,10,11}. Given the prominent role of MAPK signaling in ESCC biology and therapy response^1,2,12, it was hypothesized that PAK5 is upregulated in ESCC and promotes malignant phenotypes through modulation of MAPK signaling, potentially involving GAREM1 as an adaptor/regulator^13,14,15,16. Quantitative IHC, functional genetic perturbation, and transcriptomic profiling were employed to test this hypothesis.

The management of ESCC typically involves a combination of esophagectomy, chemotherapy, and radiotherapy. While esophagectomy can be curative, it carries a high risk of morbidity^4,8. Additionally, both chemotherapy and radiotherapy often fail to eliminate all cancer cells, resulting in resistance, recurrence, and metastasis^8,9. Immunotherapy has emerged as a prominent focus in oncology recently, demonstrating effectiveness in treating advanced cancers. This growing interest highlights the urgent need for new biomarkers that can aid in early diagnosis and treatment planning for ESCC. Previous studies have shown that other PAK family members, such as PAK1 and PAK4, promote proliferation, invasion, and therapy resistance in ESCC, underscoring the importance of this kinase family in ESCC biology. However, the role of PAK5, a less studied group II PAK, remains poorly defined, motivating the current study^17,18.

P21-activated kinase 5 (PAK5), a member of the PAK II subfamily, was first identified as a brain-specific kinase in 2002. Located on chromosome 20p12, PAK5 encodes a protein of approximately 80 kDa and is primarily found in the mitochondria and nucleus, where it plays diverse roles in cellular regulation^10,11,13. Its functions in the mitochondria include energy metabolism and regulation, while its nuclear presence is linked to gene transcription and cell cycle control. Although PAK5 is less studied than its counterparts, it is believed to be crucial for neurodevelopment, cellular survival, and cancer progression¹⁰. Notably, PAK5 has been implicated in cytoskeletal regulation, anti-apoptotic mechanisms, and cellular proliferation¹². Its upregulation in neoplastic cells has been linked to emergent drug resistance, resonating with findings that associate PAK5 overexpression with heightened tumor cell resistance to chemotherapy^10,11,13,14. Yet, the PAK5's precise function in ESCC progression remains elusive. This study probes PAK5's influence on the ESCC transcriptomic milieu and its functional ramifications. Transcriptomic profiling was performed to identify gene expression alterations associated with PAK5 activity in ESCC cells. Additionally, an ESCC cell model with controlled PAK5 modulation was generated to elucidate its role in tumor growth and metastatic potential. Unlike conventional biomarker discovery strategies in ESCC that rely primarily on Western blotting or single-modality transcript measurements, the present workflow integrates quantitative immunohistochemistry, transcriptomic profiling, and functional assays. This combined approach enables simultaneous molecular and phenotypic validation, providing mechanistic insight while improving biological relevance. In addition, the protocol emphasizes experimental reproducibility, utilizing paired tumor and adjacent tissue samples, standardized immunostaining procedures, and high-quality RNA inputs (RNA integrity number ≥ 7) for sequencing. These features enhance assay robustness and applicability across laboratories, particularly for studies seeking to define clinically meaningful kinase-driven regulatory networks in ESCC.

All procedures were approved by the Ethics Committee of the Second Affiliated Hospital of Hainan Medical University (Approval No. LW2022602), and informed consent was obtained from each participant in accordance with the Declaration of Helsinki. The reagents and the equipment used in this study are listed in the Table of Materials.

1. Patient and tissue sample collection

A total of 40 patients with pathologically confirmed ESCC were enrolled, all with normal hepatic and renal function prior to surgery. Tumor specimens (Group T) and adjacent normal tissues (Group N; ≥2 cm from the tumor margin) were collected immediately after surgical resection. Tissues were fixed in 4% paraformaldehyde (PFA) in PBS at room temperature for 7 days before paraffin embedding.

2. Immunohistochemical staining

Paraffin-embedded blocks were sectioned at 5 µm and baked at 60 °C for 1 h. Sections were deparaffinized with xylene substitute, rehydrated through graded ethanol, and subjected to antigen retrieval in 10 mM sodium citrate buffer (pH 6.0) at 95-98 °C for 20 min. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 10 min, followed by blocking with 5% normal serum for 30 min at room temperature. Sections were incubated overnight (12-16 h) at 4 °C with anti-PAK5 primary antibody (1:100 dilution, see Table of Materials). After PBS washes, slides were incubated with horseradish peroxidase-conjugated secondary antibody for 30 min at room temperature, developed with DAB until color appeared (1-5 min, monitored under microscope to avoid over-staining), and counterstained with hematoxylin for 30-60 s. Sections were dehydrated, mounted, and imaged under a brightfield microscope. Staining intensity and proportion of positive cells were quantified using image analysis software with at least five random fields per section (200× magnification). The quantification method adopted H-score calculation, with the formula: H-score = ∑ (staining intensity grade × percentage of positive cells) (intensity graded as 0 = negative, 1 = weak, 2 = moderate, 3 = strong; percentage categorized as 0-25%, 26-50%, 51-75%, 76-100%). To ensure reliability, two independent pathologists performed blind scoring, with an inter-observer agreement κ value ≥ 0.75. Quality control criteria were strictly implemented: Positive controls (known PAK5-positive ESCC tissue) and negative controls (primary antibody replaced with PBS) must pass validation; background staining intensity must be below a pre-set threshold (e.g., H-score < 50); coefficient of variation (CV) of staining results across duplicate batches must be <20% to exclude experimental variability.

3. Cell lines and culture

KYSE-150 (human ESCC) and HEK-293 (viral packaging) cells were authenticated by short tandem repeat (STR) profiling and confirmed to be mycoplasma-free. KYSE-150 cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (100 U/mL penicillin and 100 µg/mL streptomycin). HEK-293 cells were maintained in DMEM supplemented with the same additives. Cells were subcultured using 0.25% trypsin-EDTA when they reached 70%-80% confluence, and were seeded to appropriate densities 24 h before each experiment.

4. Lentiviral vector construction and infection

Lentiviral vectors were designed to either overexpress PAK5 (pLV-EF1α-PAK5-PGK-Puro) or silence PAK5 (pLV-U6-shRNA-PAK5-PGK-Puro). HEK-293 cells were seeded in 10 cm dishes and transfected at 70%-80% confluence with the transfer plasmid, packaging plasmids (gag/pol and rev), and envelope plasmid (VSV-G) at a 4:3:1 ratio using a lipid-based transfection reagent, with a total DNA amount of ~20 µg per 10 cm dish. The medium was replaced 6-8 h after transfection. Viral supernatants were collected at 48 h and 72 h, centrifuged at 500 × g for 10 min to remove debris, filtered through a 0.45 µm filter, and either concentrated or stored at −80 °C. For transduction, KYSE-150 cells were seeded one day prior to infection and exposed to viral supernatant at a multiplicity of infection (MOI) of 5-10 in the presence of 8 µg/mL polybrene. MOI was defined as the ratio of infectious units to target cell number. A pre-experiment was conducted to estimate MOI using GFP counting or qPCR, and an MOI of 5-10 was selected. Medium was replaced after 12-16 h, and a puromycin kill-curve was established by treating non-transduced cells with puromycin concentrations ranging from 0-4 µg/mL for 72-96 h to determine the minimal lethal concentration. The selected concentration for selection was 1-2 µg/mL, and the maintenance concentration was 0.5-1 µg/mL. Cells were selected with puromycin (1-2 µg/mL) beginning 48 h later. PAK5 overexpression or silencing was verified by RT-qPCR using specific primers for PAK5 (forward 5′-CCAAAGCCTATGGTGGGACCC-3′, reverse 5′-AGGCCGTTGATGGAGGTTTC-3′) and GAPDH (forward 5′-GTGGACATCCGCAAAGAC-3′, reverse 5′-AAAGGGTGTAACGCAACTA-3′). The relative expression of PAK5 was calculated using the 2^−ΔΔCt method. All experiments were performed with at least three biological replicates.

5. CCK-8 cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8), which is based on the reduction of the WST-8 reagent by cellular dehydrogenases. Logarithmically growing KYSE-150 cells were seeded at 2,000 cells per well in 96-well plates in 100 µL complete medium. At 0 h, 24 h, 48 h, 72 h, and 96 h, 10 µL of CCK-8 reagent (containing WST-8) was added to each well. After incubation at 37 °C for 2 h, absorbance at 450 nm (A450) was measured with a microplate reader. Blank wells containing medium only were included. Each experimental group contained at least three biological replicates, and each measurement was performed with five technical replicates.

6. Colony formation assay

Single-cell suspensions of KYSE-150 were seeded at 1,000 cells per well in 6-well plates and cultured for 14 days, with medium replaced every 2-3 days. Colonies were fixed with 4% PFA for 15 min, stained with 0.1% crystal violet in 20% methanol for 15 min, rinsed with water, and air-dried. Colonies containing more than 50 cells were counted using image analysis software, and the colony formation rate was calculated.

7. Transwell invasion assay

Inserts with 8 µm pores were pre-coated with basement membrane matrix (50 µL per insert,1:8 dilution) and incubated at 37 °C for 1 h; quality control was performed before seeding cells: blank inserts (without cells) and negative control inserts (without chemoattractant) were set up to verify assay validity, and the uniformity of the basement membrane matrix layer was confirmed by visual inspection (or weight measurement for stricter control). Cells were serum-starved in 1% FBS medium for 6-12 h, and 2 × 10⁵ cells/mL (200 µL per insert) were seeded into the upper chamber. The lower chamber contained 600 µL medium with 20% FBS as a chemoattractant. After 24 h incubation, non-invaded cells on the upper surface were removed, and invaded cells on the lower surface were fixed with methanol for 10 min and stained with 0.1% crystal violet for 15 min. Stained cells were imaged under a brightfield microscope and counted in at least five random microscopic fields per insert (200× magnification); counting was performed using a blind method (counting personnel were unaware of group assignments) to avoid subjective bias.

8. Flow cytometric detection of apoptosis

Cells were harvested (including floating cells), washed twice with PBS, and resuspended in binding buffer at 1 × 10⁶ cells/mL. Annexin V-FITC (5 µL) and propidium iodide (5 µL) were added to 100 µL of the suspension, followed by incubation at room temperature for 15 min in the dark;after incubation, 400 µL binding buffer was added to each sample. Prior to flow cytometric analysis, compensation adjustment was performed using single-stained controls (Annexin V-FITC single-stained cells and PI single-stained cells) to eliminate fluorescence spillover, and the FSC/SSC gate was set to select single cells (excluding cell debris and aggregates). Samples were analyzed within 1 h on a flow cytometer equipped with a 488 nm laser, with at least 10,000 events acquired per sample. Apoptotic indices were quantified using flow cytometry analysis software: a uniform gating threshold was applied across all samples to ensure consistency, and the proportions of early apoptotic cells (Annexin V⁺/PI⁻) and late apoptotic cells (Annexin V⁺/PI⁺) were separately reported to characterize the apoptotic profile of each group.

9. RNA sequencing and bioinformatics analysis

Total RNA was extracted and assessed for purity (A₂₆₀/₂₈₀ between 1.8 and 2.1) and integrity (RIN ≥7 as a quality control threshold). mRNA was purified using oligo(dT) beads, fragmented at 94 °C for 5-7 min, and reverse transcribed into cDNA. Libraries were prepared by end repair, adaptor ligation, and PCR amplification (8-12 cycles), and sequenced on an Illumina platform to generate paired-end 150 bp reads at a depth of 30-50 million read pairs per sample. Reads were trimmed and aligned to the human reference genome using STAR or HISAT2. Gene counts were obtained with featureCounts, and differential expression analysis was performed with DESeq2, applying thresholds of FDR < 0.05 and |log₂ fold change| ≥ 1. Functional enrichment analysis of differentially expressed genes was conducted using clusterProfiler for Gene Ontology and KEGG pathways (with Benjamini-Hochberg (BH) corrected FDR < 0.05), and protein-protein interaction networks were constructed using the STRING database.

10. Statistical analysis

Continuous data were expressed as mean ± SD from at least three independent experiments. Comparisons between two groups were performed with two-tailed unpaired t-tests, and comparisons among three or more groups were performed with one-way ANOVA followed by Tukey's post-hoc test. Non-parametric tests were applied where appropriate. Survival distributions were estimated by Kaplan-Meier analysis with log-rank tests. For RNA-seq data, multiple testing correction was performed using the Benjamini-Hochberg method. Exact p-values are reported in figure legends; statistical significance was considered at p < 0.05, with* indicating p < 0.05 and **indicating p < 0.01.

PAK5 is highly expressed in ESCC and is associated with poor prognosis
Figure 1 illustrates the overall workflow, from the clinical validation of PAK5 to functional and transcriptomic analysis, which identifies GAREM1 as a downstream target. To test the hypothesis that PAK5 promotes ESCC progression by modulating the MAPK pathway via GAREM1, the expression patterns of PAK5 in ESCC tissues and their association with clinical prognosis were examined. Firstly, the gene expression dataset GSE269447, sourced from the Gene Expression Omnibus (GEO) database, was utilized to screen for significantly differentially expressed genes (DEGs) between ESCC and normal tissues19. Subsequently, heat maps and volcano plots were generated (Figure 2A,B). A heatmap was constructed to visually present the top 25 genes with the greatest expression variance (Figure 2A). In this heatmap, blue represents significant down-regulation, while red indicates significant up-regulation. The volcano plot (Figure 2B) supports that, in comparison to normal tissues, in ESCC tissues, 3,345 genes exhibited significant up-regulation, while 1,467 genes showed significant down-regulation. Further expression analysis of PAK5 in ESCC and matched normal tissues was carried out to confirm this differential expression pattern. PAK5 expression was significantly higher in ESCC than in normal tissues (p = 0.0079; Figure 2C). Moreover, patients with high PAK5 expression had worse overall survival than those with low expression (p = 0.039; Figure 2D). These findings imply that elevated PAK5 expression is correlated with a poorer clinical prognosis in ESCC. Thus, PAK5 emerges as a potential biomarker and therapeutic target for improving the survival outcomes of ESCC patients.

PAK5 high expression in ESCC correlates with advanced tumor stage and poor survival outcomes
A total of 40 paired clinical samples were analyzed, comparing ESCC tissues with adjacent non-cancerous tissues. Immunohistochemical (IHC) analysis was employed to quantify the expression levels of p21-activated kinase 5 (PAK5), with the staining intensity graded on a scale from 0 to 7. IHC results demonstrated marked PAK5 staining in carcinoma tissues, in sharp contrast to the scarce detection in normal tissues (Figure 3A,B). The immunoreactivity score (Figure 3C) indicated a substantial elevation in PAK5 expression in cancer tissues, approximately twice that of the normal tissue baseline (p < 0.01). Consistent staining intensity was observed across technical replicates, with an inter-observer agreement κ ≥ 0.75, supporting scoring reproducibility. A univariate analysis of two groups of clinical samples (group T and group N) was conducted using the Kaplan-Meier survival curve. Prognostic indicators of overall survival (OS) and disease-free survival (DFS) were predicted, encompassing sex, degree of differentiation, comprehensive stage, T stage, and N stage. The survival curves (Figure 3D,E) suggested that sex had no statistically significant influence on OS(p = 0.16) or DFS (p = 0.17). Differentiation degree demonstrated clinically-relevant survivorship linkage (p = 0.014) but shows a trend towards significance with disease-free survival (p = 0.053). Well-differentiated histology orchestrated lethality-risk diminution. Tumor stage was a crucial determinant for both OS (p < 0.0001) and DFS (p < 0.0001). Advanced stages (III/IV) were linked to significantly worse survival outcomes. The T-stage (extent of the primary tumor) established α-calibrated determinancy for both OS (p = 0.00028) and for DFS (p = 0.0013), with higher T-stages corresponding to poorer outcomes. The N-stage (nodal involvement) was also significantly associated with OS (p < 0.0001) and DFS (p = 0.002). In summary, the data underscored the significance of tumor differentiation, stage, and nodal involvement as key predictors of patient survival. Higher tumor differentiation, lower stage, and minimal nodal involvement were associated with enhanced OS and DFS. Furthermore, the log-rank test and Cox regression analysis revealed that elevated PAK5 expression served as an independent prognostic indicator for unfavorable survival among patients with ESCC (HR = 2.35, 95%CI = 1.52 - 3.65, p < 0.01).

Elucidating PAK5's role via the manipulation of KYSE-150 cell lines
To elucidate PAK5's role in the biological functions of ESCC, lentiviral vectors for PAK5 gene overexpression (PAK5-OE) (Figure 3A) and knockdown (PAK5-si) were engineered (Figure 3B). These vectors were utilized to create KYSE-150 cell line variants with either amplified or suppressed PAK5 expression (Figure 3C). The successful establishment of these modified cell lines was confirmed through RT-qPCR assays (Figure 3D). Transduction efficiency exceeded 80%, confirmed by reporter fluorescence microscopy and RT-qPCR (Figure 4C,D), ensuring reliable establishment of PAK5-OE and PAK5-si models.

PAK5 promotes malignant phenotypes in ESCC
Enhanced proliferation in PAK5-OE cells and reduced proliferation in PAK5-silenced cells were demonstrated by CCK-8 assays (Figure 5A) and colony-formation results (Figure 5B-C). Increased migration and invasion were observed in PAK5-OE cells, whereas PAK5 silencing impaired these phenotypes (Figure 5D-F). Flow cytometry further confirmed reduced apoptosis in PAK5-OE cells and increased apoptosis in PAK5-silenced cells (Figure 5G,H).

PAK5 silencing upregulates GAREM1 and affects gene expression profiles related to ESCC malignant phenotypes
In this study, KYSE-150 cells were manipulated by silencing PAK5, and mRNA sequencing was subsequently performed to examine the cellular response. Differentially expressed genes (DEGs) were identified and visualized using volcano plots and heatmaps to illustrate the extent of gene expression changes. The volcano plot (Figure 6A) presents a clear distinction between down-regulated genes, marked by blue dots, and up-regulated genes, indicated by red dots, with gray dots denoting genes that exhibited no significant alteration in expression levels. Complementing the volcano plot, the heatmap (Figure 6B) offers a detailed view of transcriptional architecture variance between PAK5-silenced and wild-type cells across multiple samples. The color gradients in the heatmap correspond to varying levels of gene expression, allowing for an intuitive grasp of the differences. Furthermore, KEGG pathway analysis (Figure 6C) was employed to assess the impact of PAK5 silencing on specific biological pathways, identifying the MAPK signaling pathway as significantly affected. Given the MAPK pathway's well-established role in esophageal cancer progression and its relevance to PAK family function, attention was directed toward GAREM1 (GRB2-associated regulator of MAPK1 subtype 1), a gene with marked upregulation in PAK5-silenced cells and known ties to MAPK pathway regulation. qRT-PCR validation (Figure 6D) confirmed significant elevation of GAREM1 mRNA levels in PAK5-silenced cells compared to controls (p < 0.01). This aligned with sequencing data, underscoring GAREM1 upregulation as a key response to PAK5 depletion. These results indicate that GAREM1 expression increases following PAK5 silencing, suggesting a potential regulatory association between PAK5 and GAREM1. This relationship is inferred from transcriptomic and RT-qPCR data and has not yet been validated through direct protein-protein interaction or transcriptional regulation assays. Additional biochemical and functional studies will be required to confirm this regulatory mechanism. Raw IHC scoring and lentiviral transduction quantification data are available upon request to support reproducibility. Taken together, these findings support a potential PAK5-GAREM1-MAPK axis in ESCC; however, this association currently remains correlative and requires further mechanistic validation.

DATA AVAILABILITY
Data supporting the findings of this study are included in the published article and its supplementary materials under accession number GSE269447 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE269447). The mRNA sequencing data for PAK5-modified KYSE-150 cells have been deposited in GEO under accession number GSE283031 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE283031).

Figure 1: Summary of the study workflow and key findings. This schematic illustrates the integrated experimental pipeline used to investigate the role of PAK5 in ESCC. Phase I involved clinical sample collection and immunohistochemical analysis of paired ESCC and adjacent normal tissues, followed by survival assessment, demonstrating that PAK5 overexpression predicts poor prognosis (quantified by H-score and IOD; κ = 0.82 for reproducibility). Phase II describes lentiviral construction and genetic modification of KYSE-150 cells to achieve PAK5 overexpression and knockdown, confirmed by RT-qPCR. Phase III includes functional characterization via proliferation (CCK-8 and colony formation), migration and invasion (Transwell), and apoptosis (flow cytometry) assays, complemented by RNA-seq and bioinformatic analyses identifying GAREM1 as a PAK5-associated gene enriched in the MAPK pathway. Phase IV shows validation of GAREM1 expression by qRT-PCR, supporting a proposed PAK5-GAREM1 regulatory axis. Please click here to view a larger version of this figure.

Figure 2: The expression of PAK5 in ESCC was analyzed using public gene expression data. (A) Heatmap of the top 20 differentially expressed genes in ESCC and normal tissues. (B) Volcano plot showing differentially expressed genes between ESCC and normal tissues. (C) Distribution of PAK5 expression levels in normal and ESCC tissues. (D) Kaplan-Meier survival analysis based on PAK5 expression levels in ESCC patients Please click here to view a larger version of this figure.

Figure 3: PAK5 expression and survival analysis in esophageal squamous cell carcinoma (ESCC) patients. (A,B) Representative immunohistochemical (IHC) staining of PAK5 in ESCC tissues (T group) and adjacent normal tissues (N group). PAK5 protein was mainly localized in the cytoplasm and nuclei of tumor cells, showing markedly higher expression in tumor tissues than in adjacent normal tissues. Scale bars: 50 µm. (C) Quantitative analysis of PAK5 expression in 40 paired clinical samples (each pair includes one ESCC tissue and one adjacent normal tissue from the same patient). The integrated optical density (IOD) method was used to quantify PAK5 expression by measuring the product of optical density values and the stained area. Each sample was measured in triplicate (three technical replicates). Data are presented as mean ± SD; statistical significance was determined using a two-tailed unpaired t-test; p < 0.01. (D,E) Kaplan-Meier survival analysis of ESCC patients (n = 40) showing the association of PAK5 expression and clinicopathological features with (D) overall survival (OS) and (E) disease-free survival (DFS). Statistical analysis was performed using the log-rank test. (i) Sex (OS, p = 0.16; DFS, p = 0.17); (ii) Differentiation degree (OS, p = 0.014; DFS, p = 0.053); (iii) Clinical stage (OS, p < 0.0001; DFS, p < 0.0001); (iv) T stage (OS, p = 0.00028; DFS, p = 0.0013); (v) N stage (OS, p < 0.0001; DFS, p = 0.002). Patients with high PAK5 expression exhibited significantly shorter OS and DFS compared with those with low PAK5 expression (p < 0.01), suggesting that elevated PAK5 predicts poor prognosis in ESCC. Please click here to view a larger version of this figure.

Figure 4: Construction and validation of PAK5 overexpression and knockdown cell lines. (A) Schematic map of the lentiviral overexpression vector pLV-EF1α-PAK5-PGK-Puro used for PAK5 overexpression (PAK5-OE). (B) Schematic map of the lentiviral interference vector pLV-U6-shRNA-PAK5-PGK-Puro used for PAK5 knockdown (PAK5-si). (C) Representative fluorescence images of KYSE-150 cells after transfection with the PAK5-OE or PAK5-si constructs. The green fluorescence indicates the reporter gene encoded by the lentiviral vectors, reflecting successful transfection and stable integration. n = 3 biological replicates per group (each biological replicate represents an independent transfection and selection experiment). Scale bars: 20 µm.(D) Relative PAK5 mRNA levels in the indicated groups as determined by RT-qPCR. n = 3 biological replicates per group, with three technical replicates per biological replicate. Data are presented as mean ± SD; statistical analysis was performed using one-way ANOVA followed by Tukey's post-hoc test; p < 0.01. Please click here to view a larger version of this figure.

Figure 5: Effects of PAK5 on malignant phenotypes of esophageal squamous cell carcinoma (ESCC) cells. (A) Cell proliferation was assessed using the CCK-8 assay at 0 h, 24 h, 48 h, 72 h, and 96 h after transfection. (B,C) Representative images and quantification of colony formation in ESCC cells with PAK5 overexpression (PAK5-OE), PAK5 knockdown (PAK5-si), or controls. (D-F) Transwell assays showing the effects of PAK5 on cell migration (upper panel) and invasion (lower panel). Representative images are shown in (D); quantitative analyses of migrating and invading cells are presented in (E) and (F), respectively. (G,H) Flow cytometry analysis of apoptosis in each group; representative scatter plots are shown in (G), and statistical quantification of apoptotic rates in (H). For all assays, n = 3 biological replicates per group, with three technical replicates per biological replicate. Data are presented as mean ± SD. Statistical significance was determined using two-tailed t-tests for pairwise comparisons and one-way ANOVA followed by Tukey's post-hoc test for multiple group comparisons. p < 0.05, *p < 0.01. Scale bars: 20 µm. Please click here to view a larger version of this figure.

Figure 6: Transcriptomic alterations and validation of GAREM1 upregulation following PAK5 knockdown in ESCC cells. (A) Volcano plot showing differentially expressed genes (DEGs) between PAK5-si and wild-type (WT) KYSE-150 cells. Red and blue dots indicate upregulated and downregulated genes, respectively, as determined by RNA-seq analysis (n = 3 biological replicates per group). (B) Heatmap of hierarchical clustering showing the global expression profiles of DEGs between PAK5-si and WT groups (n = 3 biological replicates per group). (C) KEGG pathway enrichment analysis of DEGs between PAK5-si and WT cells. The bubble plot displays enriched pathways ranked by the Rich factor and q-value (n = 3 biological replicates per group). (D) Validation of GAREM1 mRNA expression levels by RT-qPCR in PAK5-si KYSE-150 cells compared with the control and NC groups (n = 3 biological replicates per group). Data are presented as mean ± SD. Statistical significance was determined using two-tailed t-tests for pairwise comparisons and one-way ANOVA followed by Tukey's post-hoc test for multiple comparisons. p < 0.05, *p < 0.01. Please click here to view a larger version of this figure.

This study provides an in-depth examination of the role of PAK5 in the progression of esophageal squamous cell carcinoma (ESCC) by combining clinical samples, functional assays, and transcriptomic analysis. These findings broaden the understanding of PAK5's involvement in ESCC, highlighting its potential as both a prognostic marker and a therapeutic target.

The marked overexpression of PAK5 in ESCC tissues and its association with poorer patient survival is consistent with findings from other cancers, where PAK5 has been linked to tumor progression and poor prognosis^10,11. By identifying PAK5 as a regulator of ESCC progression, our study contributes to the broader understanding of kinase-mediated oncogenesis and provides new insight into MAPK pathway modulation in esophageal cancer. Notably, PAK5 is not the only member of the PAK family implicated in ESCC pathogenesis -- prior studies have highlighted the oncogenic roles of other PAK isoforms in this disease. For instance, PAK1, a Group I PAK family member, was found to be significantly upregulated in ESCC tissues, and its high expression correlated with advanced tumor stage and lymph node metastasis; mechanistically, PAK1 promoted ESCC cell proliferation and invasion by activating the ERK/MAPK signaling pathway, which overlaps with the signaling cascade implicated for PAK5 in this study^17,20. Similarly, PAK4 (a Group II PAK, same subgroup as PAK5) was reported to enhance ESCC tumorigenesis through interacting with β-catenin to activate Wnt signaling, further supporting that multiple PAK family proteins drive ESCC progression via conserved or complementary signaling cascades²¹. This extends current knowledge of PAK5's role beyond previously studied tumors^8,9,10,11 and highlights its significance in a cancer type where molecular targets remain limited^1,2,12. This study is the first to demonstrate this correlation in ESCC, adding to the growing evidence that PAK5 plays a universal role in tumorigenesis. These results suggest that PAK5 could serve as a valuable biomarker for identifying high-risk patients, potentially aiding in prognosis and guiding treatment decisions.

Functional assays performed on genetically modified ESCC cell lines, provide strong evidence that PAK5 promotes several key oncogenic traits, including increased proliferation, survival, migration, and invasion. These findings align with PAK5's previously established roles in regulating cytoskeletal architecture and anti-apoptotic pathways^10,11. Notably, silencing PAK5 significantly impaired these oncogenic functions, further supporting its potential as a therapeutic target in ESCC. The transcriptomic analysis offers new insights into PAK5's downstream effects in ESCC, particularly the identification of the PAK5-GAREM1 signaling axis. GAREM1, originally identified as a key player in the EGF signaling pathway, is implicated in the regulation of the Erk/MAPK pathway, which is known to drive cell proliferation and survival in various cancers^5,6,7,22. KEGG pathway enrichment confirmed that the MAPK signaling cascade is significantly influenced by PAK5 silencing, indicating that PAK5 may exert its tumor-promoting effects, at least in part, through modulation of MAPK signaling via GAREM1.

The regulatory relationship between PAK5 and GAREM1 is particularly noteworthy. While PAK5 has been previously implicated in MAPK pathway regulation in other cancers, the present results provide the first evidence that PAK5 negatively regulates GAREM1 in ESCC, thereby potentially enhancing MAPK pathway activity and promoting malignant progression. The observed increase in GAREM1 expression upon PAK5 knockdown suggests a compensatory mechanism or negative feedback loop that may warrant further mechanistic exploration. Clinically, these results position PAK5 not only as a prognostic biomarker but also as a potential therapeutic target. Given the widespread dysregulation of the MAPK pathway in ESCC, targeting PAK5 or its downstream signaling axis may provide a novel strategy to inhibit tumor progression. Future studies should include in vivo validation using ESCC animal models to confirm the functional significance of the PAK5-GAREM1 axis and to evaluate the therapeutic efficacy of PAK5 inhibition, either alone or in combination with MAPK pathway inhibitors. Combined use of transcriptomic profiling and functional assays provides a replicable framework to dissect oncogenic pathways in ESCC. This integrative approach can be applied to other kinases and candidate biomarkers, supporting biomarker discovery and therapeutic target validation across cancer research^{13,14,15,16,22}.

The experimental workflow implemented in this study emphasizes reproducibility and technical robustness. Consistent IHC quantification was ensured through standardized antigen-retrieval conditions and scoring reproducibility (κ ≥ 0.75), while high-quality RNA inputs (RIN ≥ 7) and adequate sequencing depth (≥30 million paired-end reads) supported reliable transcriptomic analysis. Stable PAK5-modified cell lines were obtained with transduction efficiencies exceeding 80%, and puromycin selection conditions were optimized to avoid mixed populations. These parameters represent critical steps that should not deviate to ensure the successful replication of this protocol. Compared with traditional ESCC biomarker studies that rely primarily on Western blotting or single-modality expression screening, the integrated clinical-IHC, lentiviral perturbation, functional assay, and transcriptomic design adopted here provides greater mechanistic depth and supports pathway-level discovery, while remaining scalable and adaptable to other oncogenic kinase candidates.

Several limitations of this study should be acknowledged. The results provide preliminary evidence supporting an association between PAK5 and ESCC progression. However, the precise regulatory mechanisms remain to be clarified. Future studies employing co-immunoprecipitation, promoter activity assays, and in vivo validation models will be required to delineate how PAK5 interacts with other signaling molecules to drive ESCC malignancy. The major limitations of this study include the relatively small single-center clinical cohort and the lack of in vivo functional validation. Moreover, although GAREM1 was identified as a downstream effector, the precise molecular mechanism of PAK5-mediated regulation remains incompletely resolved. Additionally, the conclusions are based on monolayer culture assays, which cannot fully recapitulate the complex tumor microenvironment in vivo; thus, in vivo tumor growth, metastasis models, and 3D/organoid systems are needed to validate the PAK5-GAREM1 axis under physiological constraints. First, the clinical sample size, although adequately powered for initial survival analyses, should be expanded in future multicenter cohorts to validate the prognostic value of PAK5 in ESCC. Second, to address the unresolved mechanistic question mentioned above, further investigation using co-immunoprecipitation, chromatin immunoprecipitation, and promoter activity assays would help elucidate how PAK5 regulates GAREM1 expression and activity, providing deeper insights. Alternative approaches to investigate the PAK5-GAREM1 axis include the use of CRISPR/Cas9-based genome editing for stable knockout models, proteomic analyses to identify direct PAK5 interactors, and in vivo xenograft or genetically engineered mouse models for functional validation^13,14,15,16.

In conclusion, suggests that PAK5 is upregulated and associated with malignant phenotypes in ESCC, promotes malignant biological behaviors, and is associated with poor clinical outcomes. The identification of the PAK5-GAREM1-MAPK axis not only offers a promising therapeutic target for ESCC intervention but also contributes to a better understanding of ESCC progression, laying a foundation for the development of PAK5-targeted therapeutic strategies. These findings further supportPAK5 as a potential prognostic biomarker and therapeutic target in ESCC, with its functions exerted through the PAK5-GAREM1-MAPK axis^8,9,10,11. Future research should validate the prognostic and therapeutic relevance of the PAK5-GAREM1 axis in larger multicenter cohorts and animal models-specifically through xenograft/patient-derived xenograft (PDX) models to monitor tumor growth curves, metastatic burden, and mouse survival; 3D spheroid/organoid assays to assess invasive capacity via invasion distance/area measurement; soft-agar assays to evaluate anchorage-independent growth by colony counting; and in vitro/in vivo experiments to verify synergistic effects of PAK5 and MAPK inhibitor combinations; additionally, development of specific small-molecule inhibitors against PAK5, or combination therapies targeting both PAK5 and MAPK signaling, may represent promising strategies for translational applications. Expression patterns can vary across public cohorts due to tumor purity, batch/platform differences, and clinical composition. To avoid over-interpretation from a single dataset, the GEO analysis is presented as supportive evidence, while the paired IHC cohort and functional assays are prioritized as the primary lines of evidence.