DUSP6-Mediated Negative Feedback to Oncogenic Tyrosine Kinase Signaling Prevents Excessive Accumulation of ROS and Enables Leukemia Cell Survival

Negative feedback signaling has been described as a mechanism to prevent oncogene-induced senescence in RAS-driven tumor cells. Here we studied mechanisms of negative feedback signaling in response to oncogenic tyrosine kinases in solid tumors (EGFR, Her2; n=5), myeloid leukemia (CML/BCR-ABL1, AML/FLT3-ITD; n=6) and B cell lineage leukemia (Ph+ ALL/BCR-ABL1; n=4).

Studying gene expression changes in response to tyrosine kinase inhibitor (TKI) treatment, we found that the DUSP6 (dual specificity phosphatase 6) represents an integral component of negative feedback signaling in a wide array of malignancies including solid tumors, myeloid and B cell lineage leukemia. As shown by Western blot, DUSP6 protein levels are high in all 10 cases of patient-derived Ph+ ALL studied. By contrast, normal pre-B cells (n=3) and B lymphoma cells lacking oncogenic tyrosine kinase activity (n=4) also lack expression of DUSP6 protein. A comprehensive CpG methylation analysis of the DUSP6 promoter region (HELP assay) revealed that CpG methylation levels observed in normal pre-B cells (n=12) are significantly increased in various types of B cell lymphomas (No tyrosine kinase; n=68) but drastically reduced in Ph+ ALL (BCR-ABL1 kinase; n=83; p=0.013).

To study the role of Dusp6 in a genetic experiment, we transformed bone marrow progenitor cells from Dusp6−/− and Dusp6+/+ mice with BCR-ABL1 to model human Ph+ ALL and CML. While Dusp6−/− leukemia cells show normal growth kinetics, they are prone to cellular senescence (11-fold increase of β-galactosidase+ cells; p=0.0001) and fail to form colonies in methylcellulose (p=0.0002). Strikingly, flow cytometry staining using DCF dye revealed drastic accumulation of ROS in Dusp6−/− leukemia cells. While protein levels of p27 and Arf were similar between Dusp6+/+ and Dusp6−/− B cell lineage leukemia cells, protein levels of p53 and p21 were significantly increased, which is consistent with high ROS levels, cellular senescence and failure to form colonies in methyl cellulose.

To test whether Dusp6-mediated negative feedback signaling represents a potential therapeutic target for the treatment of tyrosine kinase-driven leukemias, we tested the Dusp6 small molecule inhibitor 2-benzylidene-3-(cyclohexylamino)-1-Indanone hydrochloride (BCI). At 3 umol/l, BCI induces significant accumulation of ROS and cell death in Dusp6+/+ but not Dusp6−/− leukemia cells. The effect of BCI is dependent on Ncf1, the regulatory cytosolic p47 subunit of the NAPH oxidase, suggesting that ROS-accumulation represents the main pathway of cytotoxity when Dusp6 function is inhibited. We next studied the effect of BCI on patient derived Ph+ ALL cells from five patients, including two patients with T315I mutation. As expected, Imatinib had no measurable effect on patient-derived Ph+ ALL cells. In contrast, in all 5 cases Ph+ ALL cells were sensitive to BCI treatment (IC50 2.8 umol/l). In a panel of B cell lymphoma cell lines (No tyrosine kinase; n=11), BCI had no significant effect, indicating that the effects of pharmacological inhibition of Dusp6 are selective for tyrosine kinase-driven leukemia cells. To test in vivo efficacy of BCI, two patient-derived samples of Ph+ ALL carrying the T315I mutation were xenografted into sublethally irradiated NOD/SCID recipient mice. Mice were treated ten times with either vehicle, 75 mg/kg Nilotinib (oral gavage) or 25 mg/kg BCI (via tail vein injection). Repeated intravenous injection of BCI results in local toxicity and necrosis of tail tissue in a number of cases. As expected, Nilotinib-treatment had no effect on overall survival compared to vehicle. Treatment with BCI, by contrast resulted in significant prolongation of overall survival (BCI vs vehicle p=0.008; BCI vs Nilotinib p=0.01).

Our studies identify DUSP6-mediated negative feedback signaling in tyrosine kinase-driven leukemias as a novel therapeutic target. Inactivation of DUSP6-mediated negative feedback leads to massive accumulation of ROS, activation of p53, loss of leukemia self-renewal and propensity to cellular senescence. Importantly, pharmacological inhibition of Dusp6 is equally active on patient-derived Ph+ ALL that carry the T315I mutation, which evades treatment with all currently available TKI.

No relevant conflicts of interest to declare.