Integrative analysis of Anoikis-related genes reveals that FASN is a novel prognostic biomarker and promotes the malignancy of bladder cancer via Wnt/β-catenin pathway

Abstract

Bladder cancer (BC) exhibits diversity in clinical outcomes and is characterized by heterogeneity. Anoikis, a form of programmed cell death, plays a crucial role in facilitating tumor invasion and metastasis. This study comprehensively investigated the genetic landscape of BC progression, identifying 300 differentially expressed Anoikis-related genes (DE-ARGs) through in-depth analysis of the GSE13507 datasets. Functional enrichment analysis revealed associations with diverse diseases and biological processes. Employing machine learning algorithms, a logistic regression model based on nine marker genes demonstrated superior accuracy in distinguishing BC from normal samples. Validation in TCGA datasets highlighted the prognostic significance of LRP1, FASN, and SIRT6, suggesting their potential as cancer biomarkers. Particularly, FASN emerged as an independent prognostic indicator, regulating BC cell proliferation and metastasis through the Wnt/β-catenin pathway. The study provides crucial insights into altered genetic landscapes and potential therapeutic strategies for BC, emphasizing the significance of FASN in BC prognosis and progression.

Keywords: Anoikis-related genes; Biomarker; Bladder cancer; FASN; Machine learning; Wnt/β-catenin pathway.

Fig. 1

Identification of the DE-ARGs in BC specimens and functional enrichment analysis. (A) The analysis of GSE13507 datasets (including 67 normal samples and 165 tumor samples) using Student's t-test revealed 300 DE-ARGs in BC samples, including 149 downregulated genes and 151 upregulated genes. (B) Disease Ontology (DO) Analysis of DE-ARGs. The distribution of 300 DE-ARGs across various diseases, as shown in the DO analysis. (C) Gene Ontology (GO) Analysis of DE-ARGs. Functional enrichment analysis depicting the biological processes associated with the 300 DE-ARGs. (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) Analysis of DE-ARGs. Pathway enrichment analysis indicating the main pathways associated with the 300 DE-ARGs.

Fig. 2

Identification of BC-Related biomarkers Using LASSO Algorithm. (A and B)Application of the LASSO logistic regression algorithm on the GSE13507 dataset, with penalty parameter tuning through 10-fold cross-validation, resulted in the selection of 28 BC-related features. (C) Expression Patterns of 28 Novel Diagnostic Genes. Visualization of the expression patterns of the 28 identified BC-related features, providing insights into their potential diagnostic significance.

Fig. 3

Selection of Optimal Feature Genes Using SVM-RFE Algorithm. (A and B) The SVM-RFE algorithm was applied to filter DE-ARGs, identifying an optimal combination of 13 feature genes. (C)Expression Patterns of 13 Optimal Feature Genes. Visual representation of the expression patterns of the 13 optimal feature genes, offering insights into their potential diagnostic utility.

Fig. 4

Identification of 9 Marker Genes and the development of a novel diagnostic model. (A) Intersection of marker genes obtained from LASSO and SVM-RFE models, resulting in the selection of 9 marker genes (PTGS2, CHEK2, BIRC3, PRKCQ, NRAS, CDH3, LRP1, FASN, and SIRT6) for further analysis. (B) Evaluation of Logistic Regression Model. Construction of a logistic regression model based on the 9 identified marker genes, demonstrating its effectiveness in differentiating between normal and BC samples with an AUC of 0.948 in ROC curves. (C) Validation in GSE3167 Datasets. Assessment of the diagnostic capability of the logistic regression model in GSE3167 datasets, confirming its robustness. (D and E)Expression patterns of 9 marker genes in BC Specimens. Visualization of the expression patterns of the 9 marker genes in BC specimens from both GSE13507 and GSE3167 datasets.

Fig. 5

Identification of the critical survival-related marker genes in BC based on TCGA datasets. (A) Expression Profiles of 9 Marker Genes in BC Specimens based on TCGA datasets (including 19 normal samples and 412 tumor samples). (B) Prognostic Value of LRP1, FASN, and SIRT6 in BC Patients. Survival analysis demonstrated a significant association between LRP1, FASN, and SIRT6 expression levels and the clinical outcomes of BC patients. (C–E) Pan-Cancer Analysis of SIRT6, LRP1, and FASN. Examination of dysregulated expression patterns of SIRT6, LRP1, and FASN across various tumor types. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 6

The prognostic values of FASN expressions in BC patients. (A) Univariate Analysis of Prognostic Factors in BC Patients. Univariate analysis using Cox's proportional hazard model illustrating the associations between overall survival in BC patients and key variables. (B) Multivariate Cox Regression Analysis of FASN Expression. Multivariate Cox regression analyses confirming the independent prognostic value of FASN expression for overall survival in BC patients. (C and D) Nomogram model for survival prediction in BC Patients. Development of a nomogram model integrating age, clinical stage, and FASN expression as factors to provide a comprehensive and individualized assessment of overall survival in BC patients. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 7

Association between FASN expression and clinical parameters in BC. (A) Exploration of the correlations between FASN expression levels and key clinical factors, including age, gender, stage and grade. (B) Heatmap of Clinical Variables and FASN Expression. Construction of a heatmap visualizing the distribution of clinical variables among individuals with high or low FASN expression. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 8

The Biological Functions of FASN in BC using Functional enrichment analysis. (A) Identification of DEGs Associated with FASN Levels. (B) DO Analysis of 1421 DEGs. (C) GO Analysis of 1421 DEGs. (D) KEGG Analysis of 1421 DEGs.

Fig. 9

Knockdown of FASN distinctly suppressed the proliferation of BC cells. (A) Elevated Expression of FASN in BC Cell Lines by RT-PCR and Western blot. The original Western blot was shown in Supplementary S1A. (B) Efficient Knockdown of FASN in RT4 and SW780 Cells. Implementation of shRNA plasmids to achieve a robust knockdown of FASN expression in RT4 and SW780 cells. The original Western blot was shown in Supplementary S1B. (C) CCK-8 assays. (D)Clonogenic assays. (E) TUNEL assays. Scale bars: 20μm. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 10

Knockdown of FASN distinctly suppressed the metastasis of BC cells via Wnt/β‐catenin/EMT Pathway. (A) Wound healing experiment highlighting a substantial reduction in the migratory capacity of BC cells following FASN knockdown. Scale bars: 200 μm. (B) Diminished Invasive Ability After FASN Knockdown by transwell assays. Scale bars: 100 μm. (C) Modulation of Wnt/β‐catenin/EMT Pathway proteins by FASN silencing using Western blot. The original Western blot was shown in Supplementary S1C. *p < 0.05, **p < 0.01, ***p < 0.001.