Efficacy of JAK/STAT pathway inhibition in murine xenograft models of early T-cell precursor (ETP) acute lymphoblastic leukemia

Abstract

Early T-cell precursor (ETP) acute lymphoblastic leukemia (ALL) is a recently described subtype of T-ALL characterized by a unique immunophenotype and genomic profile, as well as a high rate of induction failure. Frequent mutations in cytokine receptor and Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signaling pathways led us to hypothesize that ETP-ALL is dependent on JAK/STAT signaling. Here we demonstrate aberrant activation of the JAK/STAT pathway in ETP-ALL blasts relative to non-ETP T-ALL. Moreover, ETP-ALL showed hyperactivation of STAT5 in response to interleukin-7, an effect that was abrogated by the JAK1/2 inhibitor ruxolitinib. In vivo, ruxolitinib displayed activity in 6 of 6 patient-derived murine xenograft models of ETP-ALL, with profound single-agent efficacy in 5 models. Ruxolitinib treatment decreased peripheral blast counts relative to pretreatment levels and compared with control (P < .01) in 5 of 6 ETP-ALL xenografts, with marked reduction in mean splenic blast counts (P < .01) in 6 of 6 samples. Surprisingly, both JAK/STAT pathway activation and ruxolitinib efficacy were independent of the presence of JAK/STAT pathway mutations, raising the possibility that the therapeutic potential of ruxolitinib in ETP-ALL extends beyond those cases with JAK mutations. These findings establish the preclinical in vivo efficacy of ruxolitinib in ETP-ALL, a biologically distinct subtype for which novel therapies are needed.

© 2015 by The American Society of Hematology.

Figures

Figure 1

Increased JAK/STAT activation in ETP-ALL relative to non-ETP T-ALL. (A) Levels of phosphophorylated (p) and total signaling pathway proteins in the spleens of non-ETP T-ALL xenograft models (ALL8, ALL16, ALL29, ALL30, ALL31, ALL33) and ETP-ALL xenograft models (ETP1, ETP5, ETP8, ETP12, ETP13, ETP14) by immunoblot. A CRLF2-overexpressing B-ALL xenograft with high levels of JAK/STAT pathway proteins was used as a positive control (Control). (B) Comparison of levels of pSTATs and Bcl-2 in non-ETP T-ALL and ETP-ALL xenografts. Protein levels by immunoblot of 3 replicates were quantitated relative to positive control (Control), graphed as means, and analyzed using the nonparametric Mann-Whitney test.

Figure 2

ETP-ALL hypersensitivity to IL7 stimulation. (A) Phosphoflow cytometry analysis of levels of pSTAT5 in xenografted non-ETP T-ALL and ETP-ALL leukemic blasts in the basal state (shaded) and in response to stimulation with 100 ng/mL IL7 for 15 minutes (unshaded). (B) Quantitation of mean fluorescence intensity (MFI) in the basal state and in response to IL7 exposure. (C) Change in pSTAT5 MFI with IL7 exposure compared with vehicle treatment of each xenograft. The nonparametric Mann Whitney test was used to evaluate statistical significance.

Figure 3

Expression of the IL7 receptor, CD127, correlates with IL7-induced STAT5 phosphorylation. Flow cytometry analysis of surface CD127 expression. Quantitation of MFI in non-ETP T-ALL samples that did or did not respond to IL7 and in ETP-ALL samples.

Figure 4

Ruxolitinib abrogates IL7-induced STAT5 phosphorylation. (A) ETP-ALL xenograft samples were treated for 30 minutes with vehicle (shaded) or 500 nM ruxolitinib (unshaded), and pSTAT5 levels were assessed by flow cytometry. (B) ETP-ALL xenograft samples treated as in A were subsequently exposed to vehicle or 100 ng/mL IL7 for 20 minutes, and pSTAT levels were measured.

Figure 5

Efficacy of ruxolitinib in xenograft models of ETP-ALL. Peripheral blood blast count over time and splenic blast count at death in ETP-ALL xenografts. Graphed are means and standard deviations with absolute blast count on the vertical axis and days of treatment on the horizontal axis.

Figure 6

Down-regulation of JAK targets with ruxolitinib treatment. Levels of phosphophorylated (p) and total STAT proteins in the spleens of ETP-ALL xenografts. Xenografts (2-3 mice per arm, replicates shown) were treated with ruxolitinib (rux) or vehicle (con) for 72 hours, spleens were harvested, and protein lysates were subjected to immunoblot. Protein levels by immunoblot of replicates were quantitated relative to actin and analyzed by the Welch-Satterthwaite t test. Levels of pSTAT1 were reduced by >50% in treated mice compared with control for all 6 samples (P < .05). There was a >50% reduction in pSTAT3, which was statistically significant (P < .05) for ETP1, 5, 8, and 14 but not significant for ETP12 and 13. There was a 70% or greater reduction in pSTAT5 for all 6 samples (P < .05).