Uncovering tumour heterogeneity through PKR and nc886 analysis in metastatic colon cancer patients treated with 5-FU-based chemotherapy
Ortega-García, M.B.; Mesa, A.; Moya, E.L.J.; Rueda, B.; Lopez-Ordoño, G.; García, J.Á.; Conde, V.; Redondo-Cerezo, E.; Lopez-Hidalgo, J.L.; Jiménez, G.; Peran, M.; Martínez-González, L.J.; Val, C.; Zwir, I.; Marchal, J.A.; García, M.Á.
Año de publicación: 2020
Colorectal cancer treatment has advanced over the past decade. The drug 5-fluorouracil is still used with a wide percentage of patients who do not respond. Therefore, a challenge is the identification of predictive biomarkers. The protein kinase R (PKR also called EIF2AK2) and its regulator, the non-coding pre-mir-nc886, have multiple effects on cells in response to numerous types of stress, including chemotherapy. In this work, we performed an ambispective study with 197 metastatic colon cancer patients with unresectable metastases to determine the relative expression levels of both nc886 and PKR by qPCR, as well as the location of PKR by immunohistochemistry in tumour samples and healthy tissues (plasma and colon epithelium). As primary end point, the expression levels were related to the objective response to first-line chemotherapy following the response evaluation criteria in solid tumours (RECIST) and, as the second end point, with survival at 18 and 36 months. Hierarchical agglomerative clustering was performed to accommodate the heterogeneity and complexity of oncological patients’ data. High expression levels of nc886 were related to the response to treatment and allowed to identify clusters of patients. Although the PKR mRNA expression was not associated with chemotherapy response, the absence of PKR location in the nucleolus was correlated with first-line chemotherapy response. Moreover, a relationship between survival and the expression of both PKR and nc886 in healthy tissues was found. Therefore, this work evaluated the best way to analyse the potential biomarkers PKR and nc886 in order to establish clusters of patients depending on the cancer outcomes using algorithms for complex and heterogeneous data.