Search

Sequestration in the bone marrow niche may allow leukemic stem cells to evade exposure to drugs. Because the CXCR4/SDF-1 axis is an important mechanism of leukemic stem cell interaction with marrow stroma, we tested whether plerixafor, an antagonist of CXCR4, may dislodge chronic myeloid leukemia (CML) cells from the niche, sensitizing them to tyrosine kinase inhibitors. We initially treated mice with retrovirally induced CML-like disease with imatinib plus plerixafor. Plerixafor mobilized CXCR4+ cells, but no difference was observed in leukemia burden, possibly reflecting insufficient disease control by imatinib. In a second series of experiments, we tested the combination of plerixafor with dasatinib in the same as well as an attenuated CML model. Despite much improved leukemia control, plerixafor failed to reduce leukemia burden over dasatinib alone. In addition, mice receiving plerixafor had an increased incidence of neurologic symptoms in association with CNS infiltration by BCR-ABL–expressing cells. We conclude that plerixafor is ineffective in reducing leukemia burden in this model but promotes CNS infiltration. Beneficial effects of combining tyrosine kinase inhibitors with plerixafor may be observed in a situation of minimal residual disease, but caution is warranted when disease control is incomplete.

Recent studies have revealed that p27, a nuclear cyclin-dependent kinase (Cdk) inhibitor and tumor suppressor, can acquire oncogenic activities upon mislocalization to the cytoplasm. To understand how these antagonistic activities influence oncogenesis, we dissected the nuclear and cytoplasmic functions of p27 in chronic myeloid leukemia (CML), a well-characterized malignancy caused by the BCR-ABL1 tyrosine kinase. p27 is predominantly cytoplasmic in CML and nuclear in normal cells. BCR-ABL1 regulates nuclear and cytoplasmic p27 abundance by kinase-dependent and -independent mechanisms, respectively. p27 knockdown in CML cell lines with predominantly cytoplasmic p27 induces apoptosis, consistent with a leukemogenic role of cytoplasmic p27. Accordingly, a p27 mutant (p27(CK-)) devoid of Cdk inhibitory nuclear functions enhances leukemogenesis in a murine CML model compared with complete absence of p27. In contrast, p27 mutations that enhance its stability (p27(T187A)) or nuclear retention (p27(S10A)) attenuate leukemogenesis over wild-type p27, validating the tumor-suppressor function of nuclear p27 in CML. We conclude that BCR-ABL1 kinase-dependent and -independent mechanisms convert p27 from a nuclear tumor suppressor to a cytoplasmic oncogene. These findings suggest that cytoplasmic mislocalization of p27 despite BCR-ABL1 inhibition by tyrosine kinase inhibitors may contribute to drug resistance, and effective therapeutic strategies to stabilize nuclear p27 must also prevent cytoplasmic mislocalization.

Abstract 1308 Background: Despite the great strides that have been made in the treatment of acute myeloid leukemia (AML), one-third of patients are resistant to treatment. In many cases, disease-causing genetic targets still need to be elucidated. Cytokine and growth factor receptors contribute to cancer pathogenesis by regulating various downstream signaling cascades, including tyrosine kinase driven pathways, which represent amenable therapeutic targets. However, additional investigation is needed to understand the functional and therapeutic implications of cytokine and growth factor receptors in cancer pathogenesis. We have developed a novel RNAi-based screen to uncover the functional and prognostic relevance of growth factor/cytokine receptors for leukemia pathogenesis. This assay enables us to discover novel mechanisms of leukemogenesis involving various non-kinase receptors. Importantly, we show proof-of-principle data where IL2Rγ is found to be essential for JAK3 mutant mediated leukemogenesis. Methods and Results: To identify non-kinase cytokine and growth factor receptors that are critical for leukemia cell viability, we have designed a novel RNAi based functional screen targeting 188 growth factor receptors that were found to be highly expressed by gene microarray analysis of primary AML patient specimens. We have tested 40 AML patient samples and 5 AML cell lines for dependence on these receptors by electroporating cells with 188 individual siRNAs. After culturing cells for four days, effects on cell viability were measured using an MTS assay. Candidate targets found in primary patient samples include CD24, NCOA4, IL2Rα, IL15Rα and IL2Rγ. In many cases, we found that these targets have genetic abnormalities ranging from splice variation (IL2Rα) to intron retention (IL15Rα). In a few cases, the receptor was found to serve as a scaffold for regulating downstream oncogenic signaling and thus contributes to oncogenesis. For instance, we identified IL2Rγ as a potential target in the JAK3 A572V mutation positive AML cell line (CMK). Validation experiments demonstrated that knockdown of IL2Rγ significantly reduces the viability of CMK cells (90% decrease) and abrogates phosphorylation of JAK3 and downstream signaling molecules, STAT5, MAPK and the pS6 ribosomal protein. In a converse experiment, overexpression of IL2Rγ enhanced JAK3 A572V mediated signaling and increased its transformation potential in a ligand-independent manner. Similarly, IL2Rγ overexpression also increased the oncogenic potential of other JAK3 mutants such as M511I, which have previously been identified in AML patients. Intriguingly, the absence of IL2Rγ in murine bone marrow completely abrogated the clonogenic potential of JAK3 A572V as compared to IL2Rγ-wild type marrow. These effects can be rescued by co-expressing IL2Rγ with JAK3 A572V but not by co-expressing IL2Rγ with JAK3 Y100C, a FERM domain inactivating mutation. Additionally, overexpression of IL2Rγ made JAK3 A572V cells less sensitive to the JAK family inhibitor, JAK Inhibitor I (IC50is increased from 150 nM to 625 nM). Mechanistically, IL2Rγ contributes to constitutive JAK3 mutant signaling by increasing JAK3 protein levels and phosphorylation. In fact, mutant but not wild type JAK3 increased the expression of IL2Rγ, indicating IL2Rγ contributes to constitutive JAK3 signaling through a feedback mechanism. Overall these results demonstrate an oncogenic potentiating role of IL2Rγ. Conclusions: RNAi-based functional screening for AML cell dependence on non-kinase cytokine and growth factor receptors led to the identification of novel oncogenic etiologies for AML. These findings underscore the importance of cytokine and growth factor receptors in leukemia pathogenesis. This assay can identify genes that are crucial for malignant cell growth, regardless of the mutational status both in cell lines and in primary samples. Future studies integrating this RNAi screen with deep sequencing will lend additional power to this assay by accelerating our understanding of the genetic mechanisms underlying these functional gene targets such that these findings can be rapidly translated into novel therapeutic strategies. Disclosures: No relevant conflicts of interest to declare.

Abstract 3504 The imatinib paradigm established continuous BCR-ABL inhibition as a design principle for ABL tyrosine kinase inhibitors (TKIs). However, once-daily dasatinib (serum half-life: 3–5 h) is clinically effective despite only transient BCR-ABL inhibition, opening an opportunity for in-depth study of the mechanistic requirements for ABL TKI-induced CML cell death. Apoptosis commitment after potent, transient target inhibition is observed with ABL TKIs in vitro (Blood, 114, 2009, 3459–63), although variations in required TKI concentrations relative to their activity against BCR-ABL kinase suggest involvement of previously unrecognized factors. The “oncogenic shock” concept holds that temporary disruption of BCR-ABL-mediated prosurvival and proapoptotic signaling sets up a kinetic imbalance in favor of apoptosis. We have undertaken a comprehensive mechanistic exploration of this issue, wherein CML cells were transiently exposed to the ABL TKIs imatinib (50 and 500 nM), nilotinib (50 and 500 nM), dasatinib (10 and 100 nM), and ponatinib (AP24534; 10 and 100 nM) and then investigated with respect to pathways critical to drug efficacy and intracellular residence time. Cellular studies utilized multi-parameter intracellular FACS and immunoblot analysis, liquid chromatography-mass spectrometry, and high-throughput real-time qPCR expression analysis. Corresponding biochemical studies to determine ABL kinase/inhibitor dissociation parameters were also performed. All four ABL TKIs tested were capable of triggering apoptosis following transient exposure, although nilotinib and imatinib (which feature much narrower kinase target profiles than dasatinib and ponatinib) did so only at high concentrations. In contrast to potent, transient inhibition of BCR-ABL being necessary and sufficient for commitment of CML cells to apoptosis, we found that apoptosis could be reversed under conditions involving extensive additional TKI washout protocols. Consistent with the best indicator of apoptosis induction in our experiments being incomplete restoration of BCR-ABL signaling activity to pre-treatment levels, in all cases for which apoptosis commitment was irreversible, we identified a small, functionally important pool of intracellular TKI after washout of drug from the media. This property correlated with results of ABL kinase/inhibitor dissociation studies. For example, we found that ponatinib is a tight-binding inhibitor with a remarkably slow off-rate (t1/2 >95 h). Transient exposure to ponatinib followed by thorough washout committed CML cells to apoptosis despite very low intracellular concentrations that did not completely inhibit BCR-ABL. Based on these findings, we explored the possibility that effective TKIs are inhibiting additional targets involved in committing cells to an apoptotic fate, and used high-throughput qPCR assays to identify a preliminary profile of 30 apoptosis-related genes differentially expressed in TKI-treatment conditions that do and do not irrevocably commit CML cells to apoptosis. Taken together, our findings reveal that even slightly attenuated restoration of BCR-ABL signaling correlates with apoptosis commitment and that cryptic cellular retention of ABL TKIs is important in mediating this effect, potentially via sustained low-level inhibition of auxiliary targets. By extension, monitoring intracellular drug levels by LC-MS may be informative, especially for short serum half-life kinase inhibitors such as dasatinib. These studies further establish and refine the guiding principles of commitment of CML cells to apoptosis and improve our ability to design kinase inhibitors for CML and other malignancies. Disclosures: Riddle: Life Technologies Corporation: Employment, Equity Ownership. Apgar:BD Biosciences: Employment. Deininger:BMS: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding; Ariad: Consultancy, Membership on an entity's Board of Directors or advisory committees; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Celgene: Research Funding; Genzyme: Research Funding. Druker:Bristol-Myers-Squibb: OHSU has clinical trial contracts with Bristol-Myers-Squibb to pay for patient costs, nurse and data manager salaries, and institutional overhead. Dr. Druker does not derive salary, nor does his lab receive funds from these contracts.; Novartis: OHSU has clinical trial contracts with Novartis to pay for patient costs, nurse and data manager salaries, and institutional overhead. Dr. Druker does not derive salary, nor does his lab receive funds from these contracts.; MolecularMD: OHSU and Dr. Druker have a financial interest in MolecularMD. Technology used in this research has been licensed to MolecularMD. This potential COI has been reviewed and managed by the OHSU COI in Research Committee & Integrity Program Oversight Council.

Abstract 3757 Background: The majority of patients with chronic myeloid leukemia (CML) achieve durable remissions with ABL tyrosine kinase inhibitor (TKI) therapy. However, a subset of CML patients develop resistance to ABL TKIs, either through BCR-ABL kinase-dependent (e.g. kinase domain point mutations) or BCR-ABL kinase-independent (e.g. overexpression of LYN) mechanisms. Recent studies have demonstrated that the presence of BCR-ABL in CML cells result in overexpression of the protein phosphatase 2A (PP2A) inhibitor SET, which in turn prevents PP2A from performing its regulatory role in deactivating signaling proteins through dephosphorylation. Further, loss of PP2A function accounts for increased and sustained BCR-ABL activity in CML blast crisis cells and activation of PP2A via treatment with Forskolin or FTY720 reduces clonogenicity of CML cells, regardless of their sensitivity to ABL TKIs. Therefore, given the central role of PP2A in regulating both BCR-ABL kinase-dependent and kinase-independent downstream signaling, pharmacological activation of PP2A may provide a novel approach to the development of CML therapeutics, particularly for patients with imatinib-resistant CML. Results: Here, we report a novel, specific, cell penetrating peptide known as OP449 that binds to SET and antagonizes SET's inhibition of PP2A. Initial in vitro studies of OP449 in human (K562) and murine (32Dp210) CML cell lines revealed a robust (45%) increase in PP2A activity after treatment with 1μM OP449 when compared to vehicle treated cells. OP449 was more potent than the PP2A activator FTY720 (5μM; 20% increase in PP2A activity compared to vehicle treated control). In comparative cytotoxicity studies, treatment with OP449 consistently resulted in reduced proliferation and increased apoptosis compared to FTY720 in K562 (IC50: 2.5μM vs. 5μM, respectively) and LAMA (IC50: 1.25μM vs. 2.5μM, respectively) cells. Consistent with the role of OP449 as a PP2A activator, its growth inhibitory effects in CML cells were rescued by treatment with okadaic acid, a PP2A inhibitor, indicating that OP449 specifically activates PP2A. Furthermore, the cytotoxic effects of OP449 are more selective for BCR-ABL-positive cells (K562, LAMA, 32Dp210 IC50 range: 1.0 –2.5μM) relative to parental 32D cells (IC50: 5μM). Importantly, TKI-resistant mutant CML cells were also sensitive to OP449 at similar doses (32Dp210T315I cells IC50: 2.5μM). Similarly, in primary CML CD34+ cells, we noted CFU-GM colony formation was reduced by 75% and 98% with 2.5 μM and 5 μM OP449, respectively, while no toxicity was observed in normal human CD34+ cells at concentration of 5 μM. Intriguingly, the combination of OP449 with ABL TKIs (imatinib 1.0μM, nilotinib 50nM, or dasatinib 5nM) further reduced proliferation and increased apoptosis in K562 cells and in primary CD34+ CML cells combination of 2.5 μM OP449 with nilotinib (50nM) reduced colony formation by 98%, which was more than treatment with each drug alone. These data support the potential use of OP449 in combination with ABL TKIs. Mechanistically, the amount of SET protein remained unchanged with OP449 treatment, although levels of CIP2a, another PP2A inhibitor, were reduced significantly. OP449 also inhibited BCR-ABL-mediated downstream signaling, and induced dephosphorylation and degradation of BCR-ABL, STAT5, AKT, ERK, and p38/MAPK in CD34+ CML cells and K562 cells within 6 hr of treatment. Lastly, in vivo toxicity studies did not identify any toxicity of OP449 in non-leukemic mice up to 10mg/kg/day. In a 32Dp210 xenograft model, mice treated with OP449 survived slightly but significantly longer than vehicle-treated mice (p=0.05). In a murine CML bone marrow transplantation model, increasing doses of OP449 correlated with improved survival, reduced WBC count, and reduced GFP+ leukemic cells. Combination treatment of OP449 with nilotinib showed further reduced leukemia burden as compared to monotherapy. Conclusions: Taken together, these data demonstrate that OP449 is selectively cytotoxic to CML cells both in vitro and in vivo by antagonizing SET and increasing PP2A activity, thus restoring PP2A's tumor suppressor function. Furthermore, these data validate SET antagonism as a target for development of new CML therapeutics for treatment of CML patients with resistance to ABL TKIs either alone or in combination with ABL TKIs. Disclosures: Oddo: Oncotide: Employment, Research Funding. Vitek:Oncotide: Employment, Equity Ownership. Christensen:Oncotide: Employment, Equity Ownership. Druker:Bristol-Myers-Squibb: OHSU has clinical trial contracts with Novartis and Bristol-Myers-Squibb to pay for patient costs, nurse and data manager salaries, and institutional overhead.; Novartis: OHSU has clinical trial contracts with Novartis and Bristol-Myers-Squibb to pay for patient costs, nurse and data manager salaries, and institutional overhead.; MolecularMD: Equity Ownership, verbatim: OHSU and Dr. Druker have a financial interest in MolecularMD. Technology used in this research has been licensed to MolecularMD. This potential conflict of interest has been reviewed and managed by the OHSU Conflict of Interest in Research Commi; Cylene: Consultancy; Ariad pharmaceuticals: Consultancy.

Abstract 199 Background: The tyrosine kinase inhibitors (TKIs) imatinib, nilotinib and dasatinib are very effective for the treatment of chronic phase CML. However, the majority of these patients continue to have persistence of CML cells despite continued therapy, suggesting that TKIs fail to target leukemic stem cells (LSCs). There is increasing evidence that the bone marrow microenvironment provides a sanctuary to LSCs, thereby contributing to persistence. Results: We used the human stromal cell lines (HS-5, HS-23, HS-27a) to model the microenvironment. Conditioned media from HS-5, but not HS-23 or HS-27a, reduced apoptosis of CML cell lines treated with TKIs (K562, LAMA-84, KBM-5 and KYO-1), consistent with previous reports. Similarly, CML CD34+ cells were protected from 5 μ M imatinib in a 4 day co-culture with HS-5 cells, as assessed by CFU-GM colony survival (23% vs 9% when compared to untreated controls, N=6, p=0.018). We were also able to demonstrate protection from TKIs with transwells over HS-5 and with HS-5 conditioned media, which suggests that factors secreted by HS-5 cells protect CML cells from TKIs. Cytokine analysis of conditioned media revealed relatively higher concentrations of IL6, IL-8, MCP-1, MCP-3, G-CSF and GM-CSF from HS-5 as compared to HS-23 and HS-27a. Since IL-6, G-CSF and GM-CSF are known to signal via JAK2, we tested combinations of imatinib and JAK2 inhibitors (TG101209 or CYT387) using our in vitro assay. Combination treatment with imatinib and CYT387 or TG101209 abrogated the protective effects of HS-5 conditioned media in CML cell lines. Combination treatment of CML CD34+ in HS-5 co-culture assays also abrogated the protective effects of stroma on colony formation. However, we observed that both normal CD34+ and CML CD34+ colony formation was dramatically reduced by JAK2 inhibitors using our HS-5 co-culture system, particularly at higher doses. Thus, it was unclear if a potential therapeutic window existed in vivo. To test the potential of combination therapy in vivo, we infected marrow from Balb/c mice with a retrovirus that simultaneously expresses BCR-ABL and GFP, followed by transplantation into lethally irradiated syngeneic recipients. The mice were separated into five cohorts: vehicle control, TG101209 (200mg/kg/d), nilotinib (75mg/kg/d), nilotinib + low-dose TG101209 (50mg/kg/d) and nilotinib + high-dose TG101209 (200mg/kg/d). The vehicle-treated control group died rapidly of myeloproliferative disease (MPD) with a median survival of 15.5 days. The median survival of mice treated with TG101209 was slightly prolonged at 20.5 days (p=0.06); however, these mice also died of MPD with enlarged spleens/livers and lung hemorrhage. The survival curves of mice treated with nilotinib monotherapy and nilotinib + low-dose TG101209 were similar (median survival not reached at termination of experiment). Mice treated with nilotinib + high-dose TG101209 initially had minimal mortality, however on day 26 the mice began to die without signs of MPD (no definitive cause of death at autopsy) and thus the remaining cohort was sacrificed on day 27 for analysis. In stark contrast to the other cohorts, the spleens of these mice were very small, leukopenic, and largely devoid of normal follicles, with decreased spleen weight compared to mice treated with nilotinib monotherapy (0.025 vs. 0.072 gm, p Conclusions: (1) Factors secreted by human bone marrow stromal cells attenuate the effects of imatinib in CML cell lines and primary CML CD34+ cells in a JAK2-dependent fashion. (2) Simultaneous in vivo inhibition of BCR-ABL and JAK2 dramatically reduces BCR-ABL expressing cells, but at the cost of marrow toxicity. We speculate that this limitation may be overcome by intermittent rather than continuous JAK2 inhibition, a strategy that might avoid toxicity while reducing persistent BCR-ABL disease burden. Disclosures: Deininger: Novartis: Consultancy; Bristol-Myers Squibb: Consultancy; Genzyme: Research Funding; Ariad: Consultancy.

Abstract 3276 Poster Board III-1 Background: Several studies have indicated that BCR-ABL causes cell cycle defects by interfering with the cell cycle regulatory functions of p27, a Cyclin dependent kinase (Cdk) inhibitor and tumor suppressor. Studies in BCR-ABL positive cell lines have shown that BCR-ABL promotes proteasomal degradation of p27 in a pathway that involves the SCFSKP2 ubiquitin ligase, while cytoplasmic mislocalization has been described in primary CML cells. It has been suggested that the principal effect of this cytoplasmic mislocalization is to remove p27 from the nucleus, thereby relieving Cdks from p27 inhibition. However, recent studies have shown that a p27 mutant (p27CK-), that cannot bind to Cdks or Cyclins, actively contributes to oncogenesis. This raises the question as to whether cytoplasmic mislocalization of p27 in CML cells may in fact promote leukemogenesis rather than merely compromise Cdk inhibition. We therefore hypothesized that the net contribution of p27 in CML is to promote leukemogenesis due to the oncogenic activity of cytoplasmic p27. Experimental approach and results: We determined p27 localization in BCR-ABL positive cell lines and CD34+ progenitor cells from newly diagnosed chronic phase CML patients (N=7) and from CML patients in blast crisis (N=2) by immunoblotting of nuclear and cytoplasmic cellular fractions. We found that p27 is predominantly cytoplasmic in most CML cell lines and in CD34+ cells from 8/9 (89%) patient samples, including patients in blastic phase. Cytoplasmic localization of p27 in CD34+ cells from CML patients was also confirmed by immunofluorescence analysis. Further, we observed that inhibition of BCR-ABL kinase by imatinib, an Abl kinase inhibitor increased nuclear p27 in all cell lines tested and in 4/9 patient samples (3/7 chronic phase and 1/2 blastic phase samples). However, we did not observe a substantial change in the cytoplasmic p27 levels. Similar results were obtained in Ba/F3 and 32D murine hematopoietic cell lines expressing BCR-ABL when compared with the respective parental cells. Further, SKP2 was up-regulated in CD34+ cell from CML patients as compared to the normal patients consistent withSKP2 mediated down-regulation of nuclear p27. These data suggest that nuclear but not cytoplasmic p27 levels are predominantly regulated by BCR-ABL kinase activity. To test whether p27 is crucial for BCR-ABL-driven leukemia, we compared leukemogenesis between recipients of BCR-ABL transduced p27+/+ and p27-/- bone marrow. Mice transplanted with BCR-ABL infected p27-/- marrow had significantly longer median survival (70 days, range 48-150 days) compared to recipients of p27+/+ marrow (37 days, range 14-56 days) (p=0.0123). To exclude that this difference was related to the differences in homing and engraftment capabilities of p27+/+ and p27-/- bone marrow cells, we compared short term homing and long term engraftment of p27+/+ and p27-/- bone marrow cells transplanted into wild-type recipients and found no differences. These data suggest that the net contribution of p27 to BCR-ABL-mediated leukemogenesis is positive. Further, to investigate the contribution of nuclear p27 to leukemogenesis, we utilized marrow from p27S10A mice in the murine CML model. In p27S10A mice, p27 is nuclear to to abrogation of the phosphorylation site implicated in nuclear export. We injected BCR-ABL transduced bone marrow cells of p27S10A and p27+/+ mice into wild-type recipients and compared the disease progression. We observed that mice transplanted with BCR-ABL infected p27S10A marrow had significantly longer median survival (28 days, range 23-79 days) compared to the recipients of p27+/+ marrow (23 days, range 21-38 days) (p=0.0139). This data is consistent with nuclear tumor suppressor function of p27. Combined with the data above, this suggests that cytoplasmic p27 promotes BCR-ABL mediated leukemogenesis. Conclusions: Our data suggest that though nuclear p27 functions as a tumor suppressor, the net contribution of p27 in CML might be oncogenic due to an oncogenic role of the increased cytoplasmic p27. Restoring nuclear p27 or reducing cytoplasmic p27 may be therapeutically useful in malignancies with low nuclear and high cytoplasmic p27 expression. Disclosures: Druker: OHSU patent #843 - Mutate ABL Kinase Domains: Patents & Royalties; MolecularMD: Equity Ownership; Roche: Consultancy; Cylene Pharmaceuticals: Consultancy; Calistoga Pharmaceuticals: Consultancy; Avalon Pharmaceuticals: Consultancy; Ambit Biosciences: Consultancy; Millipore via Dana-Farber Cancer Institute: Patents & Royalties; Novartis, ARIAD, Bristol-Myers Squibb: Research Funding. Deininger:Genzyme: Research Funding; BMS: Consultancy; Novartis: Consultancy, Honoraria; Ariad : Research Funding.

Abstract 941 Defining extrinsic survival pathways in CML progenitor cells Ryan MacKenzie1, Jennifer Snead1, Jonathan VanDyke1, Brian Druker1,2, Michael Deininger1 1Center for Hematologic Malignancies, Oregon Health & Science University Cancer Institute, 2Howard Hughes Medical Institute, 3181 S.W. Sam Jackson Park Rd., Portland, Oregon 97239-3098, USA Background: The stem cell niche is thought to provide a protective microenvironment that allows for the persistence of malignant progenitors in various types of leukemia treated with chemotherapy. Tyrosine kinase inhibitor (TKI) treatment of chronic myeloid leukemia (CML) does not completely eliminate residual BCR/ABL positive progenitors, despite complete cytogenetic responses in the majority of patients, a phenomenon referred to as disease persistence. We hypothesize that niche factors may contribute to CML progenitor survival in the presence of imatinib and that the identification of defined survival factors may lead to therapeutic approaches to eliminate residual disease. Approach and results: In an initial series of experiments CD34+ CML progenitor cells were cultured in (a) serum free media supplemented with low concentrations of cytokines (=control), (b) in direct contact with M210B4 murine stroma or (c) separated from M210B4 murine stroma through a transwell membrane, in the presence or absence of TKIs (dasatinib, imatinib). After 96h, cells were plated in semisolid media and CFU-GM scored. Murine stroma consistently increased CFU-GM survival 4-5-fold compared to controls, with most of the effect attributable to soluble factors. Similar observations were made with human stroma (HS-5), although direct contact accounted for approximately 40% of the protective effect. Since integrin-mediated adhesion has been shown to protect CML cells lines from imatinib and chemotherapy, we hypothesized that direct contact protection of CML CD34+ cells may be related to integrin engagement. To test this hypothesis we cultured CML CD34+ cells on plates coated or not with fibronectin and in the presence or absence of integrin activating (B44) or blocking antibodies (6S6), with imatinib added at 1.5 and 5 uM or dasatinib added at 5 or 50 nM. However, in several experiments (N=3), we found that integrin-mediated adhesion to fibronectin failed to promote CFU-GM survival in the presence of TKIs. To identify soluble factors responsible for progenitor cell survival in the presence of TKIs we measured cytokine and chemokine concentrations in HS-5 stroma conditioned media (CM) and found measurable levels of several cytokines with JAK2-dependent signal transduction, including IL3, IL6, GM-CSF and G-CSF. Based on this we asked whether a JAK2 inhibitor (CYT387) would synergize with imatinib to inhibit CML CFU-GM survival. Thus, CD34+ cells were cultured for 96h with imatinib and CYT387 HS-5 stroma cells. Imatinib (5uM) reduced CFU-GM survival by 79.4%. Addition of 5 uM of CYT387, a JAK2 inhibitor (IC50 for growth inhibition of BAF3 cells expressing JAK2V617F: 1.5uM) reduced CFU-GM survival by 92.9%. Combination of imatinib with graded concentrations of CYT387 led to a further reduction of CFU-GM in a dose-dependent manner (Table 1). Conclusions: (a) Contact to and soluble factors produced by stromal cells appear to protect CD34+ CML progenitor cells in the presence of TKIs. (b) We find no evidence that integrin-mediated adhesion to fibronectin is involved in the protection of primary cells (unlike BCR-ABL expressing cell lines). (c) Combination of a JAK2 inhibitor with imatinib synergistically reduces colony formation, suggesting that this combination may be effective in targeting residual leukemia cells in vivo. Disclosures: Deininger: Novartis: Consultancy; Bristol-Myers Squibb: Consultancy; Calistoga: Research Funding; Genzyme: Research Funding.