TCF/LEF family transcription factors: Molecular landscape and prognostic scoring in pan‐cancer
Zhengjun Yang, D. Q. Cai, Zhen Zhao, Zhixiang Jian, Mude Shi, Yajin Chen, Jueming Chen, Chunhua Qu, Diankui CaiAbstract
Background
TCF/LEF transcription factors (TFs) are indispensable for canonical Wnt signal transduction, acting through the recruitment of β‐catenin and its co‐activators to Wnt response elements at target gene promoters. Dysregulation of TCF/LEF family TFs is implicated in various cancers, yet their comprehensive molecular landscape and prognostic implications across cancers remain incompletely characterised.
Methods
We analysed data from The Cancer Genome Atlas and GTEx databases to assess the expression patterns and correlations of four TCF/LEF family TFs (TCF1/TCF7, LEF1, TCF7L1 and TCF7L2) across 33 cancer types, validating findings in tumour cell lines using the CCLE database. We conducted a systematic characterisation of the molecular landscape, encompassing copy number alterations, somatic mutations and DNA methylation patterns. In parallel, we built molecular regulatory networks leveraging motif‒TF relationships and mRNA‒miRNA interactions. Using gene set variation analysis, we calculated TCF/LEF scores and stratified pan‐cancer samples into high‐score ( H ‐score) and low‐score ( L ‐score) subgroups.
Results
Functional enrichment analysis confirmed the previously reported significant positive correlation between TCF/LEF gene expression and Wnt/β‐catenin signalling, as well as pathways involved in cell differentiation and development. Immune infiltration analysis showed strong associations between TCF/LEF scores and specific immune subsets: positive correlations with CAFs, NKT, MAIT, CD4 T cells (naïve and regulatory subtypes), dendritic cells (various subsets), B cells (naïve) and Tfh cells; and negative correlations with Th1, Th17, pDC, pro‐B cells, M2 macrophages, neutrophils and NK cells. We also examined seven metabolic features and uncovered significant associations between TCF/LEF scores and metabolic reprogramming. Drug sensitivity analysis using the CellMiner database and connectivity map identified potential compounds targeting the TCF/LEF score. Prognostic analysis demonstrated the efficacy of the TCF/LEF score in predicting patient survival (overall survival or progression‐free interval). We also constructed a THYM‐specific nomogram for prognostic prediction.
Conclusions
In summary, we developed a TCF/LEF family TF‐based pan‐cancer transcriptomic signature with promising potential for predicting prognosis and therapeutic response.