A Drug Screen in Zebrafish Identifies a New Therapeutic Agent Active in Acute Lymphoblastic Leukemia

T cell lineage is an independent high risk factor in acute lymphoblastic leukemia (ALL). T-ALL requires high dose multi-agent chemotherapy, conferring many toxic side effects. T-ALL treatment therefore needs new, targeted agents that preserve or improve current treatment efficacy, yet cause fewer side effects than existing chemotherapeutic regimens. To identify such drugs, we pioneered an in vivo screen using transgenic zebrafish with T cell-specific green fluorescent protein (GFP) expression. We reasoned that immature T cells in 5-day-old zebrafish larvae are developmentally analogous to T-ALL lymphoblasts, and likely rely upon similar signaling pathways. Hence, compounds that specifically eliminate T cells in zebrafish larvae might likewise selectively target T-ALL cells. An added benefit of our in vivo screen is that drugs added to the water housing fish larvae must cross an epithelial barrier, can be metabolized by the liver, and can be renally cleared, characteristics not assessed in in vitro-based screening strategies. In addition, by using early larvae as our subjects, drugs lacking T cell specificity will likely impair normal development and/or survival, which we postulate is a predictor of unwanted toxicities. In these ways, our screen mimicked the scenario encountered in patients. Our use of the transgenic p56^lck::EGFP zebrafish line facilitated rapid visual assessment of efficacy of a large number of compounds in 96-well format.

In a proof-of-principle experiment, we identified several known anti-T-ALL drugs, most prominently glucocorticoids, from the “Spectrum” library (MicroSource Discovery Systems, Inc., Gaylordsville, CT) of well-characterized compounds. We then screened 39,200 small molecules from the “ChemBridge DIVERSet” combinatorial chemistry library (ChemBridge Corp., San Diego, CA) for those active against zebrafish larval T cells. Of 20 novel “hits” identified, one compound, dubbed Lenaldekar (“LDK”), met additional prioritization criteria. LDK does not impair the cell cycle of developing zebrafish, is well tolerated and orally bioavailable with favorable pharmacokinetic properties in mice. In addition, pilot studies with LDK indicate it is efficacious in treating T-ALL in juvenile and adult fish from an established transgenic rag2::ER-human cMyc T-ALL model. LDK kills all of several murine T-lymphocytic malignancies, induces apoptosis in all human T-ALL lines tested, and shows some activity in human B-ALL lines. However, glioblastoma, colon carcinoma, melanoma, or immortalized human embryonic kidney cell lines are not affected by LDK, even at high concentration (up to 25μM). Using the recently established “phosflow” technique we measured phosphorylation status of key signaling proteins in permeabilized T- and B-ALL lines. Regardless of PTEN status, PDK1, AKT and mTOR downstream targets p4EBP1 and p70S6kinase were dephosphorylated by LDK treatment, as was the p65 subunit of NFκB on serine 529. Results were corroborated by conventional Western blots. However, phosphorylation of STAT3, STAT5, pERK1-2, and p38 were not affected by LDK. LDK's action is distinguished from other AKT/mTOR inhibitors by its lack of activity against AKT-dependent glioblastoma and melanoma cell lines, and its lack of effect on cell size. Finally, LDK decreased tumor burden of human T-ALL in murine xenografts.

In view of its apparent lymphocyte selectivity, we posit that LDK modulates a pathway relatively unique to ALL (and immature lymphoblast) cells. This suggests that LDK can serve as a novel molecular tool for studies of normal and malignant lymphocyte biology. Moreover, with its favorable pharmacokinetics, apparent lack of toxicity, and in vivo efficacy in two vertebrate ALL models, LDK is an attractive molecule for development into a targeted treatment for ALL and perhaps other lymphocytic malignancies.

No relevant conflicts of interest to declare.