The SMAC mimetic LCL-161 selectively targets JAK2V617F mutant cells

Background Evasion from programmed cell death is a hallmark of cancer and can be achieved in cancer cells by overexpression of inhibitor of apoptosis proteins (IAPs). Second mitochondria-derived activator of caspases (SMAC) directly bind to IAPs and promote apoptosis; thus, SMAC mimetics have been investigated in a variety of cancer types. particularly in diseases with high inflammation and NFĸB activation. Given that elevated TNFα levels and NFĸB activation is a characteristic feature of myeloproliferative neoplasms (MPN), we investigated the effect of the SMAC mimetic LCL-161 on MPN cell survival in vitro and disease development in vivo. Methods To investigate the effect of the SMAC mimetic LCL-161 in vitro, we utilized murine and human cell lines to perform cell viability assays as well as primary bone marrow from mice or humans with JAK2V617F–driven MPN to interrogate myeloid colony formation. To elucidate the effect of the SMAC mimetic LCL-161 in vivo, we treated a JAK2V617F–driven mouse model of MPN with LCL-161 then assessed blood counts, splenomegaly, and myelofibrosis. Results We found that JAK2V617F-mutated cells are hypersensitive to the SMAC mimetic LCL-161 in the absence of exogenous TNFα. JAK2 kinase activity and NFĸB activation is required for JAK2V617F-mediated sensitivity to LCL-161, as JAK or NFĸB inhibitors diminished the differential sensitivity of JAK2V617F mutant cells to IAP inhibition. Finally, LCL-161 reduces splenomegaly and may reduce fibrosis in a mouse model of JAK2V617F-driven MPN. Conclusion LCL-161 may be therapeutically useful in MPN, in particular when exogenous TNFα signaling is blocked. NFĸB activation is a characteristic feature of JAK2V617F mutant cells and this sensitizes them to SMAC mimetic induced killing even in the absence of TNFα. However, when exogenous TNFα is added, NFĸB is activated in both mutant and wild-type cells, abolishing the differential sensitivity. Moreover, JAK kinase activity is required for the differential sensitivity of JAK2V617F mutant cells, suggesting that the addition of JAK2 inhibitors to SMAC mimetics would detract from the ability of SMAC mimetics to selectively target JAK2V617F mutant cells. Instead, combination therapy with other agents that reduce inflammatory cytokines but preserve JAK2 signaling in mutant cells may be a more beneficial combination therapy in MPN.


Background
Evasion of programmed cell death is a common survival technique utilized by cancer cells and one way these cells can evade apoptosis is by upregulating anti-apoptotic proteins such as inhibitor of apoptosis proteins (IAPs) [1]. IAPs inhibit apoptosis by directly binding to and inhibiting pro-apoptotic caspases. However, when a cell is triggered to undergo apoptosis, second mitochondria-derived activator of caspases (SMAC) is released into the cytosol and bind IAP proteins, leading to their degradation. Degradation of IAP proteins induces apoptosis by freeing pro-apoptotic caspases and allowing them to induce cell death. Small molecules which mimic the antagonistic effects of SMAC and lead to degradation of IAP proteins, termed SMAC mimetics, are actively being investigated in both hematologic malignancies and solid tumors [2,3]. In preclinical studies, SMAC mimetics induce apoptosis in cancer cells directly or indirectly by priming these cells for killing by other cytotoxic agents. SMAC mimetics sensitize cells to TNFα-induced cell death [4]. TNFα is a master regulator of inflammation and induces NFĸB activation among other signaling pathways. However, NFĸB activation is critical for SMAC mimetic-induced apoptosis [5]. In a study involving primary AML samples, sensitivity to a SMAC mimetic correlated with activation of TNF signaling and low XIAP expression [6]. Therefore, SMAC mimetics may be particularly useful in cancers associated with high inflammation and NFĸB activation. Myeloproliferative neoplasm (MPN) is a chronic hematologic malignancy characterized by chronic inflammation [7][8][9], high TNFα levels in serum [10], activation of NFĸB [11] and reduced expression of XIAP [12]. MPN is characterized by the somatic acquisition of a mutation in either JAK2 (JAK2 V617F ) [13][14][15][16][17] or calreticulin (CALR) [18,19] in a hematopoietic stem cell. This mutant clone expands, producing excessive numbers of mature myeloid cells that also carry the driver mutation. The current goals of therapy in MPN are reduction of thrombotic risk, reduction in spleen volume, and alleviation of symptoms. There are no pharmacologic therapies currently in use apart from interferon-alpha that specifically deplete the neoplastic clone and lead to molecular remission [20]. Thus, there is an unmet need to identify therapeutics to successfully target JAK2 V617F cells in MPN. The SMAC mimetic LCL-161 is currently in a Phase 2 clinical trials for the MPN subtype myelofibrosis (NCT02098161) [21]. However, it still remains unclear how LCL-161 impacts JAK2 V617F or CALR mutant cells compared to wildtype counterparts.
Based upon the fact that there is hyperactivation of NFĸB signaling pathway [22] we reasoned that MPN cells would be more susceptible to killing by SMAC mimetics. To test this, we performed cell viability assays using murine and human cell lines, myeloid colony formation assays using primary MPN patient cells and a mouse model to test the effect of the SMAC mimetic LCL-161 on JAK2 V617F cells in the presence and absence of TNFα. Here, we find that murine and human JAK2 V617F cell lines are hypersensitive to the SMAC mimetic LCL-161 in vitro in the absence but not presence of TNFα. Furthermore, LCL-161 reduced myeloid colony formation to a greater extent in MPN mice and patients and reduced spleen size and possibly myelofibrosis in vivo.

Drug formulation
LCL-161 was provided by Novartis Pharmaceutical Corporation. For in vitro assays, LCL-161 was dissolved in DMSO at a concentration of 10 mM, stored in single use aliquots at − 80 °C, and dissolved directly in culture media fresh for each experiment. LCL-161 was formulated for oral gavage by dissolving in 0.1 N HCl to a final concentration of 10 mg/ml in sodium acetate buffer (100 mM, pH 4.6). LCL-161 was prepared by first wetting 10 mg powder stock with 30 μl water, and then dissolving in two equivalents (0.73 μl/mg of compound) of 6.0 N HCl. The resulting (clear) solution was brought up to 1 ml in pH 4.6 acetate buffer, and the resulting stock was stored frozen at − 20 °C until used for gavage treatment of mice.

Protein extraction and Western Blot
Following washes with PBS, cells were lysed in 1× Immunoprecipitation buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA pH 8.0, 1 mM EGTA, 1% Triton X-100, 10 mM sodium pyrophosphate, 10 mM α-glycerophosphate, 1 mM Na 3 VO 4 , and 50 mM NaF) by vortexing for 10 s and placing cells on ice for 10 min. Next, samples were centrifuged at 12,000×g for 15 min at 4° C. Supernatants were collected and 4× Lamelli buffer (Bio-Rad) was added to each sample prior to storage at − 20 °C. Protein concentration was determined by Bicinchoninic acid (BCA) assay at absorbance at 532 nm. Western blot analysis was performed using 30 µg for protein boiled for 10 min at 95 °C and electrophoresis was carried out on a 10% acrylamide gel. The protein was transferred onto a nitrocellulose membrane and blocked with 5% BSA in 0.1% PBS-T at 25 °C. The membrane was incubated for 1 h at room temperature or overnight at 4 °C with the primary antibodies for cIAP1 (ab2399, Abcam), cIAP2 (ab23423, Abcam), XIAP (ab28151, Abcam), and β-actin (ab8227, Abcam). Following three 15 min wash steps with 0.1% PBS-T, membranes were incubated with secondary antibody goat-anti-rabbit IgG for 1 h. Following an additional wash step, membranes were imaged on a Chemdoc luminescence reader (Bio-Rad).

Human cell lines and apoptosis assays
K562 and HEL cells (gift from Brian Druker Lab, Oregon Health & Science University) were maintained in RPMI 1640 (Corning) containing phenol red supplemented 100 U penicillin, 100 µg/ml streptomycin, 2 mM glutamine, and 10% FBS. Cells were cultured in a humidified tissue culture incubator at 37 °C and 5% CO 2 . For apoptosis assays, K562 and HEL cells were plated into 12-well plates at a density of 1 × 10 6 cells per well in RPMI+ 10% FBS + Pen/Strep/l-glutamine containing vehicle, LCL-161, or LCL-161 + TNFα. Following 16 h, flow cytometry for annexin-V and propidium iodide (PI) was performed.

Murine colony formation assays
Bone marrow was harvested from JAK2 V617F knock-in mouse model of MPN and cells were plated at a density of 10,000 cells/ml in Methocult M3230 (stem cell) supplemented 10 ng/ml mouse interleukin-3 (mIL-3), 3 U/ml human erythropoietin (hEPO), and 50 ng/ml mouse stem cells factor (mSCF), all purchased from Peprotech. The cells were plated in triplicate at a density of 10,000 cells/ ml and incubated at 37 °C in a 5% carbon dioxide humidified incubator for 5-7 days. Standard morphological criteria were used to score hematopoietic colonies using a light microscope.

Collection, isolation and culture of primary human samples
All participants gave their informed consent for the studies conducted in accordance with the Declaration of Helsinki. This study was approved by the Institutional Review Board of the University of California, Irvine (IRB HS #2014-9995). Peripheral blood samples were collected from MPN patients with Polycythemia Vera, Essential Thrombocytopenia, and Primary Myelofibrosis or normal controls. Density gradient centrifugation with Ficoll was performed to isolate blood mononuclear cells (PBMCs). Following PBMC isolation, a red blood cell lysis step was performed with ammonium chloride buffer (ACK).

Human hematopoietic colony assays
Isolated PBMCs were plated in Methocult H4320 (Stem Cell Technologies, Vancouver Canada) supplemented 10 ng/ml human interleukin-3 (hIL-3), 3 U/ml human erythropoietin (hEPO), and 50 ng/ml human stem cells factor (hSCF). The cells were plated in triplicate at a density of 100,000 cells/ml and incubated at 37 °C in a 5% CO 2 humidified incubator for 7-10 days. Standard morphological criteria were used to score hematopoietic colonies using a light microscope.

Mice
All mouse work was performed with approval from the IACUC committee at University of California, Irvine. Equal numbers of male and female mice were used for all experiments. The Jak2 V617F was a gift from Ann Mullally, Brigham and Women's Hospital. The transductiontransplantation mouse model of JAK2 V617F driven-MPN was established as previously described [10,23].

JAK2 V617F mutant cells are hyper-sensitive to the SMAC mimetic LCL-161 in the absence but not presence of TNFα
To investigate the sensitivity of JAK2 V617F mutant cells to LCL-161, we created cell lines with stable expression of JAK2 V617F , JAK2 WT , or empty GFP vector by retroviral transduction. In this assay, we utilized the mouse fibroblast L929 line which are extremely sensitive to TNFαinduced death. We measured the impact of increasing concentrations of LCL-161 on the viability of these cells using a resazurin based assay in the presence or absence of TNFα (Fig. 1a-c). Sub-micromolar concentrations of LCL-161 significantly reduced the viability of L929 JAK2 V617F cells compared to JAK2 WT or empty vector cells (Fig. 1a). We observed 50% cell viability of L929 JAK2 V617F cells with 0.45 µM LCL-161 as compared to similar cell viabilities with 5.5 µM and 9.5 µM LCL-161 for empty vector or JAK2 WT cells, respectively (P < 0.0001, 2way ANOVA; Fig. 1a). However, this differential sensitivity was lost when TNFα was added to culture (Fig. 1b). Interestingly, overexpression of JAK2 WT rendered L929 cells resistant to killing by higher concentrations of LCL-161 in the presence of TNFα, as these cells were 50% viable in the presence of 19 µM LCL-161 (P < 0.0001, 2way ANOVA; Fig. 1b).
Next, we assessed whether autonomous production of TNFα could explain the increased sensitivity of JAK2 V617F cells to LCL-161. To do this, we tested whether TNFα neutralizing antibody abrogates the increased sensitivity of JAK2 V617F mutant L929 cells to LCL-161. Addition of TNFα neutralizing antibody did not rescue the increased sensitivity of JAK2 V617F cells to LCL-161 (Fig. 1c) suggesting that the sensitivity of JAK2 V617F cells was not due to autonomous production of TNFα. Moreover, we could not detect any TNFα in the supernatant of L929 JAK2 V617F cells, either at baseline or with the addition of LCL-161 by ELISA (data not shown). As expected, treatment with LCL-161 decreased both cIAP1 and cIAP2 (Fig. 1d). We found no marked differences in expression of cIAP1, cIAP2, or XIAP in JAK2 V617F versus JAK2 WT or parental L929 cells under any conditions tested (Fig. 1d). However, we did find evidence that sensitivity to LCL-161 is mediated at least in part to NFĸB activation, the NFĸB inhibitor QNZ had a protective effect on empty vector cells in the presence but not absence of TNF, whereas QNZ had a protective effect on JAK2 V617F mutant cells only in the absence of TNF (Additional file 1: Figure S1).
This suggests that JAK2 V617F induces NFĸB activation in L929 cells and this mediates the sensitivity to LCL-161.
We also compared the ability of LCL-161 to induce apoptosis in the HEL human erythroleukemia cell line which harbors the JAK2 V617F mutation versus the K562 human erythroleukemia cell line which harbors the BCR-ABL translocation. LCL-161 induced apoptosis significantly more in HEL cells as compared to K562 cells in the absence (P < 0.001, unpaired t test; Fig. 1e) but not presence of TNFα (Fig. 1f, g). Likewise, expression levels of cIAP1, cIAP2, or XIAP did not correlate with sensitivity to LCL-161 in HEL human cells expressing JAK2 V617F as compared to K562 cells (Additional file 1: Figure S2). These data in human cell lines corroborate our findings in L929 cells that JAK2 V617F mutant cells are more sensitive to killing by LCL-161 alone, but when TNFα is added this differential sensitivity is lost.

Calreticulin mutated cells are not hyper-sensitive to LCL-161
To investigate the effect of LCL-161 on calreticulin mutant cells, we also created L929 cell ectopically expressing MPN associated calreticulin (CALR) mutations with and without its obligate cytokine receptor scaffold thrombopoietin receptor (MPL). CALR mutant L929 cells were not more sensitive to LCL-161 as compared to empty vector either with or without co-expression of MPL (Additional file 1: Figure S3). Interestingly, overexpression of wild-type CALR led to increased sensitivity to LCL-161 when co-expressed with MPL as compared to cells with empty vector, CALR deletion or CALR insertion (Additional file 1: Figure S3).

Sensitivity of JAK2 V617F cells to LCL-161 is dependent upon Janus kinase activity
To determine whether the increased sensitivity of JAK2 V617F cells to LCL-161 is dependent upon Janus kinase (JAK) activity, we utilized the JAK inhibitors ruxolitinib (JAK1/2 inhibitor) and pacritinib (JAK2/ Flt3 inhibitor) in a resazurin based cell viability assay. Sensitivity of JAK2 V617F L929 cells was dependent upon activation of JAK2 activity, as treatment with either ruxolitinib or pacritinib reversed the increased sensitivity of JAK2 V617F cells to LCL-161 (Fig. 2). This data suggests that the hyper-sensitivity of JAK2 V617F cells to LCL-161 is due to constitutive JAK2 activation.

knock-in mice and MPN patient cells in the absence of TNFα
To investigate the effect of LCL-161 on myeloid colony formation in primary mouse bone marrow cells, we performed colony formation assays from JAK2 V617F knock-in and wild-type mice with increasing concentrations of LCL-161. LCL-161 reduced colony formation in JAK2 V617F cells compared to wild-type cells in the presence of all doses tested, although only the 1 µM LCL-161 dose reached statistical significance (P < 0.001, unpaired t-test; Fig. 3a). However, the differential sensitivity of JAK2 V617F cells was lost in the presence of TNFα (Fig. 3b). This data further supports the model that JAK2 V617F cells have increased sensitivity to SMAC mimetics in the absence but not presence of TNFα. We next investigated the effect of LCL-161 on colony formation using primary cells from MPN patients compared to unaffected individuals. In the presence of LCL-161 alone, there was a reduction in colony formation from JAK2 V617F mutated MPN patients versus normal controls at 1 µM and 2.5 µM doses, though this trend did not reach significance (P = 0.05 and 0.07, respectively, unpaired t test; Fig. 3c). These same trends were observed when erythroid and G/M colonies were enumerated separately (Additional file 1: Figure S4). When TNFα was added to the culture, this diminished the apparent difference between MPN and normal controls in response to LCL-161 at all but the highest concentration (5 µM) of LCL-161 (P < 0.05, unpaired t test; Fig. 3d).

LCL-161 reduces spleen size and ameliorates fibrosis in a JAK2 V617F transduction-transplantation model
To test the efficacy of LCL-161 in vivo, we utilized a JAK2 V617F transduction-transplantation MPN mouse model that develops myelofibrosis around 1 year of age [23][24][25][26]. Treatment with LCL-161 was initiated 15 weeks after transplant with twice weekly oral administration. We did not detect changes in peripheral blood counts nor percentage of GFP+ cells in the peripheral blood in LCL-treated mice compared to untreated mice (n = 3-4 mice/group, Fig. 4a-e). However, at sacrifice we found that LCL-161 treatment significantly reduced spleen size (P < 0.05, unpaired t test; Fig. 4f ) and modestly reduced the percentage of GFP+ cells in the bone marrow and spleen of JAK2 V617F mice (n = 3-4 mice/group, Fig. 4g). Upon analysis of reticulin-stained bone marrows, we observed that fibrosis was reduced in mice that had been treated with LCL-161 (Fig. 4h). These data demonstrate that LCL-161 may have a beneficial effect on an in vivo model of JAK2 V617F mutated MPN.

Discussion
Here, we found that mouse cell lines ectopically expressing JAK2 V617F , the JAK2 V617F mutated human cell line HEL, JAK2 V617F knock-in mice, and primary MPN samples were more sensitive to killing by the SMAC mimetic LCL-161 compared to their JAK2 WT counterparts. Although addition of exogenous TNFα increased the sensitivity of both mutant and wild-type cells to LCL-161, the differential sensitivity of mutant cells was abolished when TNFα was added to culture. The increased sensitivity to LCL-161 was not due to autocrine production of TNFα, as pharmacologic blockade of TNFα did not rescue JAK2 V617F mutant cells from LCL-161. Activation of JAK2 signaling is critical for the increased sensitivity of JAK2 V617F mutant cells to LCL-161. Addition of JAK2 inhibitors ruxolitinib or pacritinib restored LCL-161 sensitivity of JAK2 V617F mutant to that of wild-type cells. This observation suggests JAK2 inhibitors may reduce the vulnerability of JAK2 mutant cells to be preferentially killed by SMAC mimetics and that therapeutic combinations of JAK2 inhibitors and SMAC mimetic in patients may not be an ideal combination. This is an important point given that it is common for new investigational agents in myelofibrosis to be tested in combination with ruxolitinib.
The ex vivo experiments with mouse and primary human myeloid colony formation corroborate our findings in cell lines that MPN mutant cells are more sensitive than wild-type cells in the absence but not presence of TNFα. Human MPN patients have variable frequencies of JAK2 V617F and so our methylcellulose colony formation assays with primary MPN patient cells assess both mutant and wild-type colonies. This may explain why the impact of LCL-161 on MPN patient colony formation was less robust than our observations in cell lines bearing the JAK2 V617F mutation.
Cells bearing the MPN associated CALR DEL mutation were not more sensitive to LCL-161 when expressed either with or without its obligate cytokine receptor scaffold MPL. This suggests that SMAC mimetics may be most beneficial in JAK2 V617F mutated MPN patients. CALR binds to and specifically activates MPL, whereas JAK2 V617F activates signaling of multiple cytokine receptor pathways. The sensitivity of JAK2 V617F to SMAC mimetics is likely mediated via activation of pathways distinct from MPL. Therefore, combination therapy with SMAC mimetic and drugs which target MPL signaling may be additive or even synergistic in JAK2 V617F mutated MPN. In our in vivo mouse model, we found that LCL-161 had modest effects on peripheral blood counts. Although LCL-161 did reduce the percentage of GFP+ cells in the spleen and bone marrow, these effects were modest and did not reach significance. The primary benefit of LCL-161 in this mouse model was reduction in splenomegaly and possible reduction of fibrosis. Future studies would require increasing numbers of mice to validate the effect of LCL-161 on splenomegaly and myelofibrosis in vivo. Regardless, our findings suggest that LCL-161 may not be an ideal single agent in MPN but may be most useful as an adjunct to other MPN therapies.
Approaches that reduce TNFα but maintain constitutive activation of JAK2 in mutant cells may be an ideal combination with SMAC mimetics in MPN. The TNFα decoy receptor etanercept has been investigated as a single agent in an open label pilot study in myelofibrosis and was found to alleviate symptoms [27]. Another potential approach as a combinatorial study with SMAC mimetics would be to combine with a JAK inhibitor that is devoid of JAK2 inhibitory activity, such as itacitinib (INCB039110). This would result in reduction of inflammatory cytokines including TNFα while preserving the vulnerability of JAK2 V617F cells to killing by SMAC mimetic.

Conclusion
SMAC mimetics may be useful in patients with JAK2 V617F mutated MPN, as JAK2 V617F mutant cells are more readily killed by SMAC mimetics as compared to wild-type cells. This vulnerability to SMAC mimetics is mediated