Analysis of drug affected clusters in a multi-PTM correlation network
Mentor: Dr. Karen Ross, Department of Biochemistry and Molecular & Cellular Biology, Georgetown University Medical Center.
Date/Time: December 3, 2020 at 8:00am
Abstract: Tyrosine kinase inhibitors (TKIs) are used to treat cancers that are driven by hyperactive tyrosine kinase pathway signaling. These inhibitors are often effective in the short term but resistance inevitably develops either because of mutations in the target kinase that make it unresponsive to the drug or because other signaling pathways are activated that bypass the drug inhibited pathway (pathway crosstalk). The goal of this project is to identify these “crosstalking” pathways and identify specific proteins in these pathways that can be inhibited by drugs. These drugs could then potentially be used in combination with the TKI to prevent or overcome drug resistance.
One of the main ways that the activity of signaling pathways is regulated by post-translational modification (PTM) of proteins. A method was previously developed to identify groups of correlated PTMs (PTM clusters) given quantitative measurements of PTMs from several different cell line/drug treatment conditions. These clusters might contain PTMs that are co-regulated by a common upstream factor or that belong to the same signaling cascade. This PTM correlation network is called the Co-clustered Correlation Network (CCCN).
For each cell-line/drug combination, I identify clusters in the CCCN that are enriched for drug-affected PTMs. I examined five drugs (four tyrosine kinase inhibitors–erlotinib, crizotinib, afatinib, and desatinib–and one proteasome inhibitor–PR171) in nine different lung cancer cell lines with different cancer-driver mutations for a total of 25 drug/cell line combinations.
Several clusters showed interesting enrichment patterns. For example, one cluster (130.141.111) was enriched for drug-affected sites in the most experiments (20 experiments), and this cluster was affected my multiple drugs of different types (PR171, Dasatinib, Crizotinib, Erlotinib, and Afatinib). On the other hand, the 166.204.74 cluster has only one drug association (Crizotinib). The enriched clusters are likely to represent signaling modules that are targeted by the drugs.