CAST as a Tumor-Specific Suppressor in Colorectal Cancer: Discovery Through Integrative Analysis of Unannotated Isoforms, Proteomics, and Epigenetic Regulation
These findings establish CAST as a tumor-specific suppressor in CRC and provide a generalizable framework for discovering therapeutic targets from unannotated isoforms.
Abstract
Colorectal cancer (CRC) remains a leading cause of cancer-related mortality worldwide, yet the molecular mechanisms driving tumor initiation and progression are incompletely understood. Here, we integrate single-cell long-read sequencing, proteomics, and DNA methylation data to identify tumor-specific targets from unannotated isoforms in CRC. Analysis of 29,429 isoforms across 3,262 single cells revealed 3,338 novel isoforms, comprising 2,779 Novel In Catalog (NIC) and 559 Novel Not in Catalog (NNC). Cross-referencing with 68 normal tissues from GTEx identified 3 tumor-specific candidates absent from normal colon. Among these, CAST (Calpastatin) emerged as a tumor-specific suppressor, with a 63.1-fold higher expression in tumor (2.62) versus normal colon (0.03). CAST protein was confirmed in 100 tumor samples by mass spectrometry (mean expression: 0.22). Mutational analysis of 100 CRC samples revealed 8 deleterious mutations, including 2 nonsense, 2 splice-site, 2 missense, 1 frameshift, and 1 in-frame variant, consistent with a tumor-suppressor role. Promoter methylation analysis of 393 tumor samples showed that 89.3% (351/393) were hypomethylated ({beta} < 0.1), whereas only 1.5% (6/393) were hypermethylated ({beta} > 0.3), indicating demethylation-mediated reactivation. Furthermore, CAST expression was negatively correlated with promoter methylation and regulated by the miR-200 family (miR-200c: 13.48, miR-141: 10.13, miR-429: 8.09). Our findings establish CAST as a tumor-specific suppressor in CRC and provide a generalizable framework for discovering therapeutic targets from unannotated isoforms.
Tetratricopeptide repeat domain protein 39B (TTC39B) is a lipid metabolism–related gene recently implicated in cholesterol regulation. However, its clinical and biological relevance in human cancers remains undefined.
Based on multi-omics data from the Cancer Genome Atlas (TCGA) pan-cancer dataset represen...
Hui-Juan Yi, Yan Wang, Chen-Hao Zhao et al.· BMC Medical Genomics· 0 citations
It is suggested that UPK1A-AS1 is a hypoxia‐inducible oncogenic lncRNA that plays dual roles in cancer, with cancer type‐dependent associations with progression and immune modulation across malignancies and mechanistically mediating hypoxia‐associated drug resistance in HCC.
Ze-Kai Li, Min Luo, Shu-Sen Fang et al.· Analytical Cellular Patholog...· 0 citations
Background Zinc finger CCCH-type protein 15 (ZC3H15) is a ribosome-associated eukaryotic protein involved in many biological processes, including gene expression regulation, translation, and immune modulation. Although ZC3H15 has been implicated in the development of certain tumors, its expression, functional relevance...
Bin Zhou, Jian Huang, Qi-Fei Tao et al.· Frontiers in Immunology· 0 citations
Multivariate exploratory cross-sectional analyses revealed that ACTB expression patterns in the tumor environment are independently associated with stage-related clinical variables, although its clinical utility as a biomarker requires further longitudinal validation.
It is demonstrated for the first time that LOY is significantly enriched in advanced-stage disease and the aggressive basal molecular subtype, establishing LOY as a causal genetic driver of PDAC progression, acting through epigenetic remodeling to promote EMT, basal subtype identity, metastasis, and therapy resistance.
M. Dorsch, Felix E. G. Beaudry, A. Baginska et al.· Cancer Research· 0 citations
BACKGROUND
This study aimed to investigate the expression patterns, immunological functions, and prognostic value of LINC00942 across multiple cancer types, and to further examine its role in chemoresistance.
METHODS
LINC00942 expression and its diagnostic and prognostic significance were analyzed across 33 cancer ty...
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.