Calcium/Calmodulin Dependent Protein Kinase 2 Gamma Is Identified As a Potential Therapeutic Target of Myelofibrosis Using Disease-Specific Induced Pluripotent Stem Cells

Myelofibrosis (MF) is a myeloproliferative neoplasm characterized by megakaryocytic atypia, fibrosis in bone marrow, and extramedullary hematopoiesis. Although mutational analyses reveal that about 90% of patients with primary MF harbor JAK2, CALR or MPL mutation that constitutively activates JAK-STAT pathway, the outcome of JAK2 inhibitor treatment for MF patients still remains to be improved. However, only a limited number of humanized disease-models are currently available to develop a novel therapeutic strategy for MF. We have previously reported that induced pluripotent stem cells (iPSCs) established from MF patients (MF-iPSCs) by using retroviral vectors recapitulated the disease phenotype (Hosoi et al. 2014), indicating that MF-iPSCs are one of the useful humanized disease-models of MF.

We established iPSCs from three MF patients harboring a JAK2 V617F, CALR type 1, or CALR type 2 mutations by using integration-free episomal vectors and confirmed that established MF-iPSCs harbored the identical mutations to parental cells. After hematopoietic differentiation, we obtained hematopoietic progenitor cells (HPCs) immunophenotypically defined as CD34+/CD43+ cells. Viability assays revealed that JAK2 inhibitors (Ruxolitinib and HSP90) impaired cell survival of MF-HPCs compared to normal-HPCs, suggesting that MF-HPCs also reproduce drug-sensitivity.

To investigate therapeutic targets of MF, we optimized culture condition of HPCs to perform compound screening. By the use of small chemical library containing 192 compounds and two MF-iPSCs harboring a JAK2 or CALR mutation, we identified KN93, calcium/calmodulin dependent protein kinase (CAMK) 2 inhibitor, as a compound which inhibited the viability of MF-HPCs from all three patients in dose-dependent manner, compared to normal-HPCs. For validation, we used Trifluoperazine (TFP), another CAMK inhibitor and TFP had similar inhibitory effects on MF-HPCs. In Ba/F3 mouse cell line ectopically expressing MPL W515L mutant (Ba/F3 MPLmu ;IC50 of ruxolitinib = 112 nM) as other models mimicking MF than that using JAK2 mutant and CALR mutant, KN93 and TFP decreased cell growth, induced apoptosis, and suppressed the phosphorylation of Stat5, a major signaling event in MF, indicating that CAMK2 is a candidate of therapeutic target for MF.

To determine the critical subtype of CAMK2 in MF, we assessed the expression of CAMK2 subtypes. In Ba/F3 MPLmu, Camk2 gamma subtype (Camk2g) was detectable among all Camk2 subtypes in mRNA levels and phosphorylation of CAMK2G protein was enhanced at the absence of interluekin-3. Knock down of CAMK2G by short-hairpin RNA decreased cell growth of Ba/F3 MPLmu. Moreover, berbamine, a CAMK2G inhibitor, also inhibited the viability of Ba/F3 MPLmu and colony forming capacity of c-kit+ primary mouse bone marrow ectopically expressing MPLW515L compared to that of expressing wild type MPL at dose-dependent manner, suggesting that CAMK2G is a critical subtype of CAMK2 in MF.

To address whether CAMK2 inhibition overcome the JAK2 inhibitor-resistance, one of the reason why JAK2 inhibitor is not efficient in clinical setting, we established ruxolitinib-resistant Ba/F3_MPLmu (Ba/F3_MPLmu_R; IC50 of ruxolitinib = 642 nM) cells through one-month exposure to ruxolitinib. KN93 and TFP decreased cell growth, induced apoptosis, and suppressed the phosphorylation of Stat5 in Ba/F3_MPLmu_R and combination of CAMK2 inhibitors and ruxolitinib exhibited inhibitory effects more efficiently than single agent, suggesting that CAMK2 inhibitors show a cooperative effect with ruxolitinib against MF cells and have the potential to overcome JAK2 inhibitor-resistance.

To address the effectiveness of CAMK2 inhibition in primary samples, we used CD34+ cells isolated from peripheral blood mononuclear cells of MF patients. Although either KN93 or ruxolitinib inhibited the colony forming capacity of primary MF cells with similar efficiency, combination of KN93 with ruxolitinib showed cooperative inhibition. Taken together, these findings indicate that CAMK2G would be a therapeutic target of MF patients.

In conclusion, MF-iPSCs enable comparable compound screening assay and identify CAMK2G as a potential therapeutic target of MF. Our results revealed that CAMK2G inhibition could lead the novel therapeutic strategies to combine with JAK2 inhibitors and to overcome the resistance against JAK2 inhibitors.