Divergent Evolutionary Trajectories of Erk- and Stat5-Activating Lesions in Acute Lymphoblastic Leukemia

Background and significance: Activation of STAT5- and ERK-signaling are segregated to early and later stages of normal B cell development, respectively. While pro-B cells depend on activation of Stat5 downstream of cytokine receptors, pre-B and immature B cells depend on Erk-activation downstream of pre-B cell receptor (pre-BCR). Since Stat5- (cytokine receptor) and Erk- (pre-BCR) signaling antagonize each other, we studied evolutionary trajectories of Stat5- and Erk-activating lesions in B-cell lineage ALL. We tested the hypothesis that an initial activating mutation in one of the two pathways represents a commitment step that makes the tumor clone non-permissive to activating lesions in the other pathway.

Results: Studying B-lineage ALL (n=578), we found that STAT5-activating (e.g. BCR-ABL1, JAK2, cytokine receptors) and ERK-activating (NRAS, KRAS, PTPN11, NF1) lesions were mutually exclusive, with only 9 cases (1.6%) carrying lesions in both pathways. Even in the few remaining cases, low allele frequencies of one of the two lesions suggests that they do not co-occur in the same cells. Reverse phase protein array measurements revealed that phosphorylation of ERK and its upstream kinase MEK were inversely correlated with STAT5-phosphorylation (MDACC, 1983-2007; P<0.001). These findings prompted us to study mechanisms of biochemical cross-inhibition between Erk and Stat5 pathways. In support of this notion, we found that inducible NRAS^G12D activated Erk at the expense of Stat5 phosphorylation. This was due to stabilization and increased activation of the Stat5-phosphatase Ptpn6 (SHP1). Inducible ablation of Ptpn6 elevated phospho-Stat5 levels, while genetic inactivation of Stat5 strongly increased Erk activity. Studying a matched patient-derived pre-B ALL sample at diagnosis and at relapse (acquired KRAS^G12V mutation) revealed activation of ERK in association with increased BCL6 and decreased Stat5 levels in the KRAS^G12V relapsed ALL sample. With engagement of BCL6 and the Stat5-inhibitory phosphatase Ptpn6 downstream of Erk signaling, these findings suggest that occupancy of either Erk or Stat5-pathway represents a commitment step that renders cells non-permissive to the respective alternative pathway. To test this hypothesis, we induced B cell transformation with either Erk or Stat5-pathway oncogenes and then subsequently transduced with either empty vectors or vectors carrying the alternative oncogene. Compared to empty vector controls, Erk- and Stat5-driven B-ALL cells did not tolerate transformation with the alternative oncogene. To test the antagonistic relationship between Stat5- and Erk-activating lesions in B-ALL, we transformed Stat5^fl/fl B-ALL cells with NRAS^G12D for Erk-activation and Erk2^fl/fl B-ALL with BCR-ABL1 for activation of Stat5. While Erk-driven B-ALL clones were non-permissive to transduction with Stat5-activating oncogenes (BCR-ABL1), Cre-mediated inducible ablation of Stat5 increased competitive fitness and colony formation ability. Likewise, Stat5-driven B-ALL clones were non-permissive to transduction with Erk-activating oncogenes (NRAS^G12D). However, Cre-mediated inducible ablation of Erk2 increased competitive fitness and colony formation ability.

Conclusion: These results support a model of clonal evolution that selects malignant B-ALL clones based on increased polarization in favor of one single dominant pathway of oncogenic signal transduction with concurrent silencing of alternative pathways. This concept suggests that malignant transformation is not only associated with progressively increased signaling strength of the central oncogenic driver but also concomitant silencing of cytokine and growth factor receptors that normal cells would respond to.