In Search of Outliers. Mining for Protein Kinase Inhibitors Based on Their Anti-Proliferative NCI-60 Cell Lines Profile

Abstract

Protein kinases play a pivotal role in signal transduction, protein synthesis, cell growth and proliferation. Their deregulation represents the basis of pathogenesis for numerous diseases such as cancer and pathologies with cardiovascular, nervous and inflammatory components. Protein kinases are an important target in the pharmaceutical industry, with 48 protein kinase inhibitors (PKI) already approved on the market as treatments for different afflictions including several types of cancer. The present work focuses on facilitating the identification of new PKIs with antitumoral potential through the use of data-mining and basic statistics. The National Cancer Institute (NCI) granted access to the results of numerous previously tested compounds on 60 tumoral cell lines (NCI-60 panel). Our approach involved analyzing the NCI database to identify compounds that presented similar growth inhibition (GI) profiles to that of existing PKIs, but different from approved oncologic drugs with other mechanisms of action, using descriptive statistics and statistical outliers. Starting from 34,000 compounds present in the database, we filtered 400 which displayed selective inhibition on certain cancer cell lines similar to that of several already-approved PKIs.

Discovery of a benzimidazole-based dual FLT3/TrKA inhibitor targeting acute myeloid leukemia

“…This might be attributed to the inhibitory effects of sorafenib on other tyrosine kinase receptors as stem cell factor tyrosine kinase receptor (c-kit) which is overexpressed in GDM-1 cells 76 . Consistently, glioblastoma better selectivity and less off-target toxicity 77,78 . In particular, 4ACP was most potent against colon cancer KM12 cell line (IC50 = 358 nM).…”

Quantitative and Qualitative Analysis of the Anti-Proliferative Potential of the Pyrazole Scaffold in the Design of Anticancer Agents

“…Based on the values of A and SD, the outlier’s values were identified for each compound [ 26 ]. The variable O was designed to count the number of outliers for each compound.…”