Kinomic Profiling of Pediatric Acute Myeloid Leukemia for Detailed Cellular Insights

Acute myeloid leukemia (AML) remains a life threatening malignancy in children. Considerable progress has been made in elucidating the new diagnostic and prognostic markers over the past decades. The precise etiology remains unclear. Therefore, it is essential to evaluate the activation of the components of cellular signaling pathways to understand AML signaling and to design the most successful approach for combinational therapies and new kinase inhibitors.

In this study, we used a high-throughput PepChip^TMKinomics microarray system containing 976 different kinase substrates and assayed primary leukemic samples of 96 AML patients to produce an exceptionally detailed map of kinome enzymatic activities towards predefined peptide substrates. The generated profiles provide a comprehensive insight in signaling pathways active in AML patients.

As expected the activation of proteins belonging to MAPK signaling, PI3K/AKT signaling, cell cycle regulation, apoptosis and insulin signaling pathways along with the signaling receptors and immune system regulators were found. Unsupervised hierarchical cluster analysis separates the AML blast profiles based on 192 peptide activities into two clusters. Cumulative incidence of relapse (CIR) was significantly higher in the patients of cluster-2. Peptide activity patterns were independent of patient characteristics. In addition, with Gaussian network modeling, a total of 540 peptides (55%) showed at least one peptide-peptide association without a prior assumptions whereas 74 peptides (7.5%) had >39 nodes suggesting to be potential interesting signaling hubs. Among these 74 peptides, 10 peptides were identified in cluster-1 and 50 peptides were in cluster-2. Thus, this total analysis defined peptides correlated to low incidence for relapse, for examples AKT1, HGFR, RGS7 and to high incidence for relapse for instance, proteins involve in MAPK pathways (RAF1, RAC1,14-3-3 eta) and cell cycle regulation and cellular growth (c-Myc, FOXO3A, RBL1).

In conclusion, our study demonstrates the feasibility of peptide activity profiling to identify two active signaling network clusters in pediatric AML correlated to CIR. Highly correlated peptides belonging to cluster-2 provide stronger leads for selection of novel targets in future therapeutics.