Identification of Target Pathways Induced By the Multiple Myeloma Tumor Microenvironment Using Activity-Based Protein Profiling and Ex Vivo Protein Kinase Inhibitor Screening

Multiple myeloma (MM) is a heterogeneous plasma cell neoplasm that remains all but incurable despite recent advances in treatment. Indeed, nearly all patients eventually experience disease progression or relapse due to a reservoir of residual myeloma cells that appear to persist through pro-survival signaling from interactions with the tumor microenvironment (TME), leading to eventual clonal expansion. Thus, identifying targets that are induced in MM by the TME may reveal new and important targets amenable to therapeutic intervention. To develop a non-biased method to screen bone marrow specimens from myeloma patients for activated targets throughout the course of disease, we used a combination of activity-based protein profiling (ABPP) and a high-throughput protein kinase inhibitor (PKI) screen using a platform that recapitulates the TME. Target validation was then performed using ex vivo functional screens of pathways using MM patient specimens.

The MM cell lines MM1.S, H929, and OPM2 were grown in mono-culture or co-culture with HS5 bone marrow stroma cells for 24h and lysates were enriched for ATP binding proteins by affinity purification versus a chemical probe (ActivX, Thermo). Tryptic peptides were measured using discovery proteomics (nano-UPLC and QExactive Plus mass spectrometer). Using this method, 176, 136, and 85 kinases out of a total of 1511, 1409, and 1281 proteins were preferentially enriched by 2-fold change from MM1.S, H929, and OPM2 myeloma cells grown in co-culture conditions with HS5 bone marrow stroma, respectively. Of these, 42 kinases were common to all three and 87 were common to two of three MM cell lines. Kinases were chosen for target validation after pathway analysis using the Kyoto Encyclopedia of Genes and Genomes database to identify signaling networks.

To identify functionally relevant signaling networks identified via ABPP experiments, the same MM cell lines were simultaneously screened with 30 protein kinase inhibitors (PKIs) in a novel high throughput viability assay. This label-free method measures the viability of MM cells grown in a collagen matrix with bone marrow stroma cells in 384-well plates to simulate the TME by capturing brightfield images every 30 minutes for 96h using a motorized microscope equipped with an incubation chamber. Digital image analysis software measures live cell numbers by detecting membrane motion and generates viability curves as a function of drug concentration and exposure time (Khin et al. Cancer Research 2014). This functional screen confirming known MM survival networks, validated 12 kinases/PKIs in the context of the TME and highlighted novel targetable pathways.

To provide an additional level of screening, the same PKIs were tested in CD138-MACS-selected cells from 15 MM patient specimens in a high-throughput viability assay. Eight PKIs targeting IGFR, PLK1, Abl, mTOR, FAK/Pyk2, ALK, Akt, and Casein Kinase-1δ (CK1δ)/CK1ε also showed significant activity in the 15 primary MM specimens. Our three-tiered pharmaco-proteomic screen identified eight kinases critical to MM survival in the context of the TME. Notably, a highly specific in-house inhibitor of Casein Kinase 1δ/CK1ε, SR-3029, which targets the Wnt/β-catnenin pathway, was identified as the most effective compound assessed as a single agent in our ex vivo viability assay in all patients with an average 36h LD50 of 290nM. This compound is under further investigation in MM (Submitted Abstract: Burger, et al, ASH 2016). Additional studies are underway to functionally interrogate the pathways identified in this screen, including ErbB1/EGFR, EphA1 and AMPK. Future work will optimize this method for evaluation of primary bone marrow specimens with ABPP followed by functional validation to better predict potential clinical response at different disease stages. We anticipate that this iterative "at the moment of care" approach is critical because drug resistant tumor phenotypes fluctuate with therapy, and this strategy can track and define clinically relevant changes in tumor cells in situ after the selection pressures applied by exposure to therapy.