Bcr-Abl Directly Activates Stat5 Independent of Jak2

Abstract 511

Persistent activation of the transcription factor Stat5 is a signaling hallmark of Chronic Myelogenous Leukemia (CML). In mouse models, Stat5 was required for initial myeloid and lymphoid transformation (by Bcr-Abl p210 or p185 and v-Abl). Most importantly, we and others recently showed that Stat5 was also required for maintenance of Bcr-Abl-dependent leukemia in vivo and for engraftment and reconstitution of Bcr-Abl p210-positive leukemia in secondary recipients. Therefore, Stat5 is of central functional importance in the Bcr-Abl signaling network and represents a possible critical vulnerable node in CML. In contrast to the well-studied functional role of Stat5 in disease models, the molecular mechanism of Bcr-Abl dependent Stat5 activation, including the tyrosine kinase(s) that phosphorylate and activate Stat5, remain only partially understood. In particular, conflicting views on the involvement of the Jak2 kinase and its potential role as a drug target in CML exist.

We used conditional Jak2 knock-out mice to study the contribution of Jak2 in Bcr-Abl-dependent transformation and leukemogenicity. Jak2 ablation did not compromise the Bcr-Abl p210-mediated transforming capability in primary murine bone marrow- or fetal liver-derived hematopoietic cells in vitro. In contrast, initial lymphoid transformation by v-abl and Bcr-Abl p185 was abolished in Jak2 knock-out mice. Jak2 deletion did not have an effect on maintenance of lymphoid leukemia cells in vitro, whereas deletion of Stat5 induced a G1 arrest and subsequent apoptosis. In line with this, ablation of Jak2 expression after leukemia induction did not alter disease latency or disease phenotype.

Consistently, we did not observe a decrease in Stat5 activation upon siRNA-mediated knock-down of Jak2 alone or all four Jak kinases (Jak1, Jak2, Jak3 and Tyk2) in CML cell lines. Using a panel of pharmacological inhibitors, we found that neither Jak2-selective, nor pan-Jak kinase inhibitors or Src family kinase-selective inhibitors led to a decrease in Stat5 phosphorylation, while the highly selective Bcr-Abl inhibitor nilotinib completely abrogated Stat5 phosphorylation. To study possible contributions of other tyrosine kinases in the Bcr-Abl dependent activation of Stat5, we used Ba/F3 cells expressing the TKI-resistant Bcr-Abl mutant T315I in combination with different broad-specificity tyrosine kinase inhibitors, like dasatinib. At dasatinib concentrations that inhibited several dozens of tyrosine kinases, Stat5 phosphorylation in Ba/F3 Bcr-Abl T315I cells was unaffected, excluding a role for most tyrosine kinases other than Bcr-Abl in Stat5 activation and pointing towards a direct phosphorylation of Stat5 by Bcr-Abl. Together, this data excludes a role of Jak and Src kinases in the activation of Stat5 in Bcr-Abl positive cell lines.

Finally, in comprehensive enzyme kinetic analysis experiments using recombinant kinase, Stat5 had a similar K_M value for Bcr-Abl as the canonical direct Bcr-Abl substrate CrkL and displayed only mildly lower kinase substrate parameters (v_max, k_cat) than CrkL, fully compatible with direct phosphorylation of Stat5 by Bcr-Abl.

Together with our earlier data on the pivotal role of Stat5 in the transcriptional and signaling network of Bcr-Abl, we propose a hypersensitive switch-like behavior of the Bcr-Abl-Stat5 kinase substrate pair that mechanistically rationalizes the central functional role of Stat5 in the signaling of CML cells.

In summary, we provide compelling evidence that activation of Stat5 by Bcr-Abl is likely to be direct and that targeting of Jak2 in CML may not be of therapeutic benefit, as Jak2 is not required for CML initiation or maintenance.