A prognostic six-gene expression risk-score derived from proteomic profiling of the metastatic colorectal cancer secretome.

The necessity to accurately predict recurrence and clinical outcome in early stage colorectal cancer (CRC) is critical to identify those patients who may benefit from adjuvant chemotherapy. Here, we developed and validated a gene-based risk-score algorithm for patient stratification and personalised treatment in early stage disease based on alterations in the secretion of metastasis-related proteins. A quantitative label-free proteomic analysis of the secretome of highly and poorly metastatic CRC cell lines with different genetic backgrounds revealed 153 differentially secreted proteins (fold-change >5). These changes in the secretome were validated at the transcriptomic level. Starting from 119 up-regulated proteins, a six-gene/protein-based prognostic signature composed of IGFBP3, CD109, LTBP1, PSAP, BMP1, and NPC2 was identified after sequential discovery, training, and validation in four different cohorts. This signature was used to develop a risk-score algorithm, named SEC6, for patient stratification. SEC6 risk-score components showed higher expression in the poor prognosis CRC subtypes: consensus molecular subtype 4 (CMS4), CRIS-B, and stem-like. High expression of the signature was also associated with patients showing dMMR, CIMP status, and BRAF mutations. In addition, the SEC6 signature was associated with lower overall survival, progression-free interval, and disease-specific survival in stage II and III patients. SEC6-based risk stratification indicated that 5-FU treatment was beneficial for low-risk patients, whereas only aggressive treatments (FOLFOX and FOLFIRI) provided benefits to high-risk patients in stages II and III. In summary, this novel risk-score demonstrates the value of the secretome compartment as a reliable source for the retrieval of biomarkers with high prognostic and chemotherapy-predictive capacity, providing a potential new tool for tailoring decision-making in patient care.