Search

Acute myeloid leukemia (AML) is an aggressive malignancy of stem cell origin that contributes to significant morbidity and mortality. The long-term prognosis remains dismal given the high likelihood for primary refractory or relapsed disease. An essential component of relapse is resurgence from the bone marrow. To date, the murine hematopoietic stem cell (HSC) niche has been clearly defined, but the human HSC niche is less well understood. The design of niche-based targeted therapies for AML must account for which cellular subsets compete for stem cell occupancy within respective bone marrow microenvironments. In this review, we highlight the principles of stem cell niche biology and discuss translational insights into the AML microenvironment as of 2021. Optimization of competition for niche occupancy is important for the elimination of measurable residual disease (MRD). Some of these novel therapeutics are in the pharmacologic pipeline for AML and may be especially useful in the setting of MRD.

Background: The combination of HMA and venetoclax is now standard of care for patients with AML who are not candidates for intensive chemotherapy. Elderly patients are more likely to have secondary AML (sAML), although the presence of an antecedent hematologic malignancy is often not apparent by history. Lindsley et al (Blood, 2015) showed that a somatic mutation in SRSF2, SF3B1, U2AF1, ZRSR2, ASXL1, EZH2, BCOR, or STAG2 is >95% specific for sAML and associated with worse outcomes. While outcomes with HMA/ven in patients meeting standard criteria for sAML have recently been reported (Pullarkat, ASCO 2021), we set out to conduct a real-world analysis of sAML patients receiving HMA/ven, including those with a secondary mutation profile (SMP) as described by Lindsley et al. We hypothesized that-when treated with HMA/ven-outcomes of patients with SMP may be most similar to those with de novo AML. Methods: Patients diagnosed with AML at Stanford Cancer Institute from 4/2017-3/2021 and treated with front-line HMA/ven were retrospectively reviewed. These included patients previously treated with HMA monotherapy for an antecedent hematologic malignancy and those who had previously received ≤ 3 cycles of HMA monotherapy for AML. Responses were classified per the modified International Working Group response criteria. Overall survival (OS) was assessed for all patients, and for patients who had a complete response (CR) or CR with incomplete hematologic recovery (CRi), duration of response (DoR) was also assessed. Statistical analyses were performed in R using the logrank test, with hazard ratios (HR) computed using the Cox proportional hazards model. For multivariate analyses, p-values for a specific variable were calculated using Cox proportional hazards regression. Results: 82 patients met criteria for inclusion; 78 had valid response assessments and 49 (62.8%) had achieved a CR or CRi at first response assessment. Median age was 72 years, with 3 patients younger than 60. 62 patients were male, median ECOG performance status (PS) was 1, median Charlson Comorbidity Index (CCI) was 6, median time to death or end of follow-up from the start of treatment was 366 days, and 58% of patients had adverse risk AML per ELN guidelines. Fig 1a demonstrates demographics for de novo, sAML (excluding SMP), and patients with SMP AML. 13 patients met criteria for AML-MRC, 23 patients had prior history of antecedent hematologic malignancy (18 with MDS or CMML, 5 with MDS/MPN overlap or MPN), 12 had tAML, and 20 patients possessed a SMP and did not meet criteria for the other three categories of sAML. 14 patients with de novo AML were characterized by the absence of any of the above factors. Patients with de novo AML were less likely to have adverse risk disease (29% vs. 64% in others) and had lower CCI scores (mean 5.1 vs. 6.2) but had no significant differences in age, gender, follow-up time, or PS. There was no statistically significant difference in rates of CR/CRi between the different subgroups or the different types of sAML; 69% of patients with de novo AML, 79% of SMP patients, and 57% of patients with other types of sAML achieved a CR or CRi. However, SMP patients had response durations and OS patterns similar to patients with de novo AML (Fig 1b and 1c), and when grouped with de novo patients, both DoR (HR = 3.5, p = 0.047, Fig 1d) and OS (HR = 2.1, p = 0.042, Fig 1e) were significantly longer than those of the sAML patients. Neither DoR nor OS were significantly longer when the SMP patients were grouped with sAML patients (respectively: HR = 3.3, p = 0.22, Fig 1f; HR = 1.5, p = 0.37, Fig 1g). In multivariate Cox proportional regression adjusting for age, ELN risk category, CCI, and PS, worse OS for sAML patients was maintained relative to the SMP and de novo patients (HR 2.9, p = 0.036), although the difference in DoR was no longer significant (HR 4.4, p= 0.10). Conclusions: Patients meeting standard definitions of sAML had worse outcomes than those with de novo AML when treated with HMA/ven in a retrospective, real-world analysis. Although a secondary mutation profile as described by Lindsley et al may be helpful in identifying patients with sAML, when treated with HMA/ven, patients with this profile have outcomes that align more closely with those of patients with de novo AML. Figure 1 Figure 1. Disclosures Mannis: Astex, Forty Seven Inc/Gilead, Glycomimetics, and Jazz Pharmaceuticals: Research Funding; AbbVie, Agios, Astellas Pharma, Bristol Myers Squibb, Genentech, MacroGenics, Pfizer, and Stemline: Consultancy.

Introduction: The addition of gemtuzumab ozogamicin (GO) to 7+3 chemotherapy for newly diagnosed acute myeloid leukemia (AML) has been shown to significantly improve event-free survival (EFS) for cytogenetically favorable-risk AML, with marginal benefit for intermediate-risk AML, and no benefit for cytogenetically adverse-risk AML. Of note, with the exception of mutated FLT3-ITD, little is known about the impact of GO in ELN 2017-defined genotypically adverse-risk AML, and a recent randomized trial found no EFS benefit for 7+3+GO in patients (pts) with genotypically favorable-risk, NPM1-mutated AML. Since 2017, our institution incorporated GO into 7+3-based inductions for all "non-adverse" risk AML pts, as defined by wild-type FLT3 and no abnormalities on rapid FISH analysis for del(5q)/monosomy 5, del(7q)/monosomy 7, and del(20q). We report our experience treating all pts with "non-adverse" risk AML-as defined by this algorithm-with 7+3+GO. Methods: An institutional database was queried in order to identify all pts ≥18 years old who received 7+3-based chemotherapy for newly diagnosed AML between 2017 and 2020; pts who received the FDA-approved fractionated dose of GO were included in the analysis. Data collection included demographic variables, karyotype/FISH, targeted PCR analyses, and multigene NGS panels for AML-related mutations including, but not limited to, mutations in FLT3, NPM1, CEBPA, TP53, RUNX1, and ASXL1. Outcome data included response to induction, relapse, and death, as well as hematopoietic cell transplant (HCT) rates, conditioning regimens, and post-transplant complications. Results: Between January 2017 and July 2020, 96 pts received 7+3-based induction at our institution. Of these, 29 (30%) received 7+3 in combination with GO. Median age at diagnosis was 46 years (range 23-66), with 17 (59%) males. Sixteen (55%) pts had ELN favorable-risk AML (5 [31%] by cytogenetics and 11 [69%] by genotype), 6 (21%) pts had ELN intermediate-risk AML, and 7 (24%) pts had ELN adverse-risk AML (4 [57%] by cytogenetics and 3 [43%] by genotype). Median time from diagnosis to start of induction was 4 days (range 0-43). For cytogenetically adverse-risk pts, median time from diagnostic bone marrow biopsy to receipt of adverse karyotype results was 8 days (7-14). Median time from start of induction to receipt of multigene NGS results for all pts was 15 days (3-32). Overall, 22 (76%) pts achieved remission. All genotypically adverse-risk pts (1 with mutated TP53 and 2 with mutated RUNX1) were refractory to induction, while 3 of 4 (75%) cytogenetically adverse-risk pts (1 with t(6;9), 1 with monosomy 7, and 2 with 11q23 abnormalities) achieved remission. Eight of the 29 (28%) pts proceeded to HCT, including 4 adverse-risk pts. Of the adverse-risk pts, all received myeloablative conditioning prior to HCT and 3 (75%) developed veno-occlusive disease (VOD), with 2 (50%) requiring defibrotide therapy. In favorable/intermediate-risk pts, 4 (18%) proceeded to HCT (2 intermediate-risk pts in first remission and 2 favorable-risk pts in second remission). Of these, 2 (50%) received myeloablative conditioning and 1 (25%) developed VOD. At last follow-up, 23 of 29 pts (79%) remained alive, with a median overall survival not reached (range 1-29 months) and a median EFS of 20 months (9-31). The percentage of ELN favorable-, intermediate-, and adverse-risk pts who remained event-free at last follow-up was 75%, 33%, and 43%, respectively. Discussion: This single-center, retrospective cohort describes the outcomes of pts with "non-adverse" risk AML who received induction chemotherapy with 7+3+GO according to a pre-defined algorithm. Using this algorithm, 30% of all pts receiving 7+3-based inductions received GO. Of these, nearly 25% were ultimately found to have adverse-risk AML as defined by ELN 2017 criteria, largely driven by long turn-around times for karyotyping and NGS multigene panel results. No patient with genotypically adverse-risk AML by ELN criteria responded to induction chemotherapy, and 75% of cytogenetically adverse-risk pts who proceeded to HCT developed VOD. Routine use of 7+3+GO induction outside of the context of cytogenetically favorable-risk AML remains controversial, and further study is needed to define the role of GO, particularly for pts with ELN genotypically adverse-risk AML. Table Disclosures Gotlib: Blueprint Medicines Corporation: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Chair of the Response Adjudication Committee and Research Funding, Research Funding; Deciphera: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: co-chair of the Study Steering Committee and Research Funding. Liedtke:Jazz Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees; Janssen: Membership on an entity's Board of Directors or advisory committees; Pfizer: Honoraria; GSK: Membership on an entity's Board of Directors or advisory committees; Adaptive: Membership on an entity's Board of Directors or advisory committees; Caelum: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees. Muffly:Adaptive: Research Funding; Amgen: Consultancy; Servier: Research Funding. Mannis:AbbVie, Agios, Bristol-Myers Squibb, Genentech: Consultancy; Glycomimetics, Forty Seven, Inc, Jazz Pharmaceuticals: Research Funding.

Acute Myeloid Leukemia (AML) remains one of the most difficult cancers to treat, with a 30% 2-year survival rate. High-throughput sequencing of AML patients has identified mutations, including FLT3, IDH1, and IDH2, for which targeted therapies have been developed. Enasidenib is an FDA-approved, first-in-class agent that preferentially inhibits IDH2-mutant activity and reduces levels of the oncometabolite 2-HG, inducing differentiation of IDH2-mutated blasts. Interestingly, greater than 50% of enasidenib-treated patients who had no objective clinical response still demonstrated improvement in their peripheral blood counts and reached RBC transfusion independence. The mechanism underlying this phenomenon is unknown but is of great clinical relevance given the high transfusion dependence and anemia-associated complications universally associated with AML. Thus, we sought to investigate how enasidenib drives normal hematopoiesis to improve quality of life and reduce morbidity in AML patients. In this study, we demonstrate that enasidenib enhances erythropoiesis from normal CD34+ hematopoietic stem and progenitor cells (HSPCs) derived from cord blood (CB) and bone marrow. Enasidenib doubled the proportion and total number of mature CD71+/GPA+ erythroblasts after 8 days of culture with EPO, SCF, and IL-3. In the presence of EPO, enasidenib induced a gene signature characteristic of maturing erythrocytes, with increased expression of GATA1 (1.3 fold), EPOR (2 fold), and KLF1 (1.4 fold), and decreased PU.1 (0.5 fold) and GATA2 (0.7 fold). Enasidenib-treated progenitor cells further demonstrated increased hemoglobin production (1.9 fold) and morphologic characteristics of increased erythroid maturation. Next, we sought to determine if enasidenib augments erythroid differentiation through IDH2 and IDH2-dependent pathways. First, we found that other IDH inhibitors (AG-120, AGI-6780, and AG-881) did not increase erythropoiesis at doses ranging from 1-10μM. As expected for normal HSPCs, 2-HG was not present at detectable levels in either the DMSO or enasidenib-treated conditions, and addition of 2-HG (50, 200μM) did not affect the ability of enasidenib to increase the proportion of CD71+GPA+ cells. Because it is possible that enasidenib acts through inhibition of wild-type IDH2, we generated CRISPR-Cas9 engineered IDH2 knockout (KO) CD34+ cells and treated them with enasidenib. Similar to wildtype cells, IDH2 KO CB CD34+ cells demonstrated a 3.4-fold increase in %CD71+GPA+ erythroid cells. Thus, enasidenib augments erythropoiesis independently of both mutant and wildtype IDH2 pathways. We then investigated the progenitor population that enasidenib acts on to drive erythroid maturation. Enasidenib did not increase the number of BFU-E or CFU-E colonies or the proportion of BFU-E (IL3R-CD34+CD36-) and CFU-E (IL3R-CD34-CD36+) progenitors in colony forming or liquid culture assays, respectively, leading us to conclude that enasidenib acts on more mature erythroid progenitors. Indeed, treating sorted mature CD71+ erythroid progenitors with enasidenib increased %CD71+GPA+ cells compared to DMSO control, whereas enasidenib treatment of CD71- early erythroid progenitors showed no effect. These observations provide evidence that enasidenib acts on CD71+ erythroid progenitors to increase late-stage erythroid differentiation. Given that CD71 allows for iron uptake into erythroid progenitors, we hypothesized that enasidenib modulates the heme biosynthesis pathway. Enasidenib inhibited the ABCG2 transporter, which effluxes protoporphyrin IX (PPIX), the direct precursor to heme, from the mitochondria and cytosol. Inhibition of ABCG2 by enasidenib could lead to PPIX accumulation within the cell, driving increased heme synthesis. To investigate this hypothesis, we treated cells with 20μM Ko143, a potent ABCG2 inhibitor, and observed a similar increase in %CD71+GPA+ cells as seen with enasidenib. Measurement of PPIX autofluorescence by flow cytometry and microscopy revealed an increase of PPIX in enasidenib-treated cells by 1.2-fold. Together, our data suggests that enasidenib drives maturation of CD71+ erythroid precursors independently of wildtype or mutant IDH2. Our results position enasidenib as a promising therapy to stimulate erythropoiesis and provide the basis for a clinical trial using enasidenib to improve anemia in a wide array of clinical contexts. Disclosures Majeti: Forty Seven Inc.: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties; BioMarin: Consultancy.

Mutations in the BCR-ABL1 kinase domain are an established mechanism of tyrosine kinase inhibitor (TKI) resistance in Philadelphia chromosome-positive leukemia, but fail to explain many cases of clinical TKI failure. In contrast, it is largely unknown why some patients fail TKI therapy despite continued suppression of BCR-ABL1 kinase activity, a situation termed BCR-ABL1 kinase-independent TKI resistance. Here, we identified activation of signal transducer and activator of transcription 3 (STAT3) by extrinsic or intrinsic mechanisms as an essential feature of BCR-ABL1 kinase-independent TKI resistance. By combining synthetic chemistry, in vitro reporter assays, and molecular dynamics-guided rational inhibitor design and high-throughput screening, we discovered BP-5-087, a potent and selective STAT3 SH2 domain inhibitor that reduces STAT3 phosphorylation and nuclear transactivation. Computational simulations, fluorescence polarization assays and hydrogen–deuterium exchange assays establish direct engagement of STAT3 by BP-5-087 and provide a high-resolution view of the STAT3 SH2 domain/BP-5-087 interface. In primary cells from chronic myeloid leukemia (CML) patients with BCR-ABL1 kinase-independent TKI resistance, BP-5-087 (1.0 μM) restored TKI sensitivity to therapy-resistant CML progenitor cells, including leukemic stem cells. Our findings implicate STAT3 as a critical signaling node in BCR-ABL1 kinase-independent TKI resistance, and suggest that BP-5-087 has clinical utility for treating malignancies characterized by STAT3 activation.

Acute Myeloid Leukemia (AML) is an aggressive cancer resulting in severe cytopenias related to bone marrow (BM) failure. The common assumption for AML-induced BM failure is overcrowding due to clonal expansion of immature myeloid blasts, leading to failure of normal hematopoiesis. However, in a cohort of 293 AML patients, we found that disease burden (% of blasts determined on diagnostic BM aspirate) did not predict severity of cytopenia (hemoglobin rs=-0.053; p=0.49; WBC rs=-0.030, p=0.70; platelet rs=0.091, p=0.026), strongly arguing against simple crowding as the main mechanism underlying AML-induced BM failure. Thus, the goal of our study is to identify novel mechanism(s) associated with AML-induced BM failure, potentially enabling development of new therapies to improve AML management and reverse morbidity. Conventional xenograft models of human AML do not typically exhibit cytopenias, making them unsuitable to study AML-induced BM failure. We speculated that in the mouse, increased splenic extramedullary hematopoiesis compensated for failed intramedullary hematopoiesis due to AML. To test this hypothesis, we performed surgical splenectomy on NSG mice prior to their transplantation with human AML. Strikingly, splenectomized NSG mice engrafted with primary human AML at a 30-70% disease burden (n=8 for primary AML samples, n=5-10 for each group of splenectomized NSG mice) developed leukopenia and severe anemia compared to sham-operated AML-engrafted controls. AML-engrafted splenectomized NSG mice showed early mortality compared to AML-engrafted NSG mice with intact spleens (10.1 weeks vs. 34.2 weeks p Utilizing our model, splenectomized NSG mice engrafted with human AML demonstrated depletion of normal hematopoietic progenitors (HSPCs) including HSCs (11.1 fold p To explore mechanisms by which AML blasts inhibit normal HSPCs, we generated conditioned media (CM) from patient-derived AML blasts, and found that AML-CM suppressed BFU-E colony formation from normal HSPCs (3.1-5.1 fold). Furthermore, in a direct co-culture system with AML blasts and CD34+ HSPCs, AML blasts inhibited erythroid differentiation from the CFU-E to normoblast stage by 81%. Removing CD34+ cells from the AML co-culture allowed cells to resume differentiation to the proerythroblast stage. These experiments demonstrate that AML imparts a differentiation blockade along the MEP-proerythroblast axis in a cell non-autonomous, reversible fashion. Using cytokine array analysis, we identified elevated IL-6 levels in AML sample-derived CM (n=10, 2841±766.4 pg/ml, 7.80 fold increase, p=0.03) compared to CD34+-derived CM (n=5, 364±36.0 pg/ml). Increased IL-6 was also found in BM aspirates from human AML engrafted splenectomized NSG mice (715±125pg/ml, p=0.001) compared to mice engrafted with normal CD34+ HSPCs (undetectable). Thus, we hypothesized that IL-6 produced by AML blasts acts as a paracrine factor to suppress erythropoiesis. Consistent with this hypothesis, an IL-6 neutralizing antibody reversed the inhibition of BFU-E formation imparted by AML-CM. Furthermore, the addition of recombinant IL-6 to liquid cultures of erythroid differentiation resulted in a 23% reduction in proerythroblasts. Together, our data suggests that (1) overcrowding is not the primary mechanism resulting in BM failure in AML; (2) AML blasts play a previously unrecognized role in imparting a differentiation blockade along the MEP-proerythroblast axis, resulting in progressive anemia; and (3) this differentiation blockade is at least partially attributable to IL-6 secreted from AML blasts as a paracrine factor. Disclosures No relevant conflicts of interest to declare.

A β-catenin expression signature is associated with primary resistance to tyrosine kinase inhibitors (TKIs) (McWeeney et al. 2010), but its role in TKI resistance is not completely understood. To assess the role of β-catenin in TKI resistance, we used shRNA targeting β-catenin (shβcat) in in vitro models of BCR-ABL1 kinase-independent resistance. To model resistance in the absence of bone marrow (BM)-derived factors, we used TKI-resistant K562R cells that are adapted for growth in the presence of imatinib, as well as primary CD34+ progenitors from CML patients who had failed treatment with two or more TKIs. These cells are cultured in regular medium (RM), proliferate in 1.0-2.5 μM imatinib, and exhibit β-catenin expression that is independent of BCR-ABL1 kinase activity. To model resistance mediated by the BM microenvironment, we cultured TKI-sensitive parental K562 cells and CD34+ progenitors from newly diagnosed CML patients in direct contact (DC) with HS-5 BM stromal cells. HS-5 co-culture increases β-catenin protein levels and clonogenic potential by >3-fold despite continued suppression of BCR-ABL1 kinase activity. All cells lack BCR-ABL1 kinase domain mutations and undergo TKI-mediated kinase inhibition as detected by immunoblot analyses. CML cell lines and primary cells with BCR-ABL1-independent resistance were lentivirally transduced with shβcat or a scrambled control (shSCR), and knockdown was confirmed by immunoblot and/or qRT-PCR. In RM, shβcat reduced the clonogenicity of TKI-sensitive K562 cells and CD34+ cells from newly diagnosed CML patients by 49% and 39%, respectively, compared to shSCR controls. In TKI-resistant K562R cells and CD34+ cells from TKI-resistant CML patients, shβcat reduced clonogenicity by 60% and 50%, respectively, in the presence or absence of imatinib (0-10 μM), suggesting a role for β-catenin in the development or maintenance of TKI resistance. In contrast to cells grown in RM, clonogenicity of cell lines and patient samples cultured in HS-5 DC was unaffected by shβcat compared to imatinib alone. Immunoblot analyses revealed that β-catenin protein levels were fully restored in HS-5 DC, despite the continued presence of shβcat. qRT-PCR revealed that while cells in HS-5 DC have high amounts of β-catenin protein, the mRNA levels remained similar to shβcat-expressing cells cultured in RM, consistent with post-translational stabilization of β-catenin. Importantly, increased β-catenin was not observed when cells were cultured in HS-5 conditioned medium, indicating that stabilization requires DC with the bone marrow stroma. These data are consistent with a role for β-catenin in TKI resistance mediated by DC with the BM microenvironment, similar to a recent report (Zhang et al., 2013). To understand nuclear versus cytoplasmic distribution following HS-5 DC, CD34+ cells from newly diagnosed CML patients were cultured in RM or HS-5 DC for 36 hours and analyzed for β-catenin localization by immunfluorescence. As expected, cells cultured in RM had low levels of β-catenin in the nucleus and cytoplasm that decreased upon treatment with imatinib (2.5 μM). In contrast, cells cultured in HS-5 DC had a marked increase of β-catenin that was unaffected by treatment with imatinib. While detectable in the nucleus, the majority of β-catenin protein was localized in the cytoplasm and at the cell membrane, consistent with its role in cell-cell junctions. Accordingly, in CD34+ cells from newly diagnosed CML patients, an N-cadherin blocking antibody impaired the clonogenic potential of cells cultured in HS-5 DC, with no significant effect on cells grown in RM. Affymetrix Human Gene 1.0 ST arrays revealed high levels of genes encoding the CDH2 and CDH13 cadherins, which may be involved in N-cadherin-mediated β-catenin stabilization, even in the presence of shβcat. Preliminary data also suggests that HS-5 DC reduces luciferase reporter activity from a construct harboring sequential β-catenin binding elements (pGF1-Lef/Tcf-eFGP-luc), further supporting a role for cytoplasmic β-catenin in TKI resistance. These data demonstrate a critical role for β-catenin in BCR-ABL1 kinase-independent TKI resistance, and suggest new strategies for targeting TKI resistance in the absence of BCR-ABL1 kinase domain mutations. Disclosures: Deininger: Bristol-Myers Squibb: Advisory Boards Other, Consultancy, Research Funding; Ariad Pharmaceuticals: Advisory Boards, Advisory Boards Other, Consultancy; Novartis: Advisory Boards, Advisory Boards Other, Consultancy, Research Funding; Celgene: Research Funding; Gilead Sciences: Research Funding.

Mutations in the BCR-ABL1 kinase domain are a well-established mechanism of tyrosine kinase inhibitor (TKI) resistance, but fail to explain many cases of clinical TKI failure. In the remaining patients, resistance occurs via activation of alternative signaling pathways that maintain survival despite BCR-ABL1 inhibition (BCR-ABL1-independent resistance). STAT3 mediates TKI resistance in chronic myeloid leukemia (CML) cells cultured in the presence of bone marrow-derived factors (Bewry et al., 2008; Traer et al., 2012; Nair et al., 2012), and also plays a critical role in survival of CML cells with BCR-ABL1-independent resistance (Eiring et al. #31, ASH 2012). While targeting transcription factors is notoriously difficult, our combination of synthetic chemistry, in vitro reporter assays, and computational modeling has led to a low micromolar mechanism-based STAT3 inhibitor, which, in combination with TKIs, shows promise as a treatment for CML patients with BCR-ABL1-independent resistance. The original compound of the series, SF1-066 (10 µM; Fletcher et al., 2009), combines with TKIs to reduce survival of CML CD34+ cells exhibiting BCR-ABL1-independent resistance (Eiring et al. #31, ASH 2012). To improve the potency and selectivity of SF1-066, we synthesized successive STAT3 inhibitor libraries and ranked candidates by structure-activity relationship using a luciferase-based reporter screen (Kraft et al. #2445, ASH 2012). This reporter assay quantifies STAT3 transcriptional activity in TKI-resistant AR230R cells, which grow in the continuous presence of imatinib (1.0 µM), lack BCR-ABL1 kinase domain mutations, and exhibit high levels of pSTAT3Y705, thereby enabling convenient, high-throughput screening for potency and selectivity in the context of endogenous STAT3 activation. Among three sequential STAT3 inhibitor libraries, BP5-087 emerged as the new lead compound. Fluorescence polarization assays verified that BP5-087 was 5-fold more effective than SF1-066 in outcompeting an SH2 peptide probe, and computational simulations predicted better overall binding of BP5-087 (-9.6 kcal/mol) versus SF1-066 (-7.6 kcal/mol) to the STAT3 SH2 interface. In AR230R cell growth assays, BP5-087 was effective at a 5-fold lower dose compared to SF1-066, with minimal effects on TKI-sensitive parental controls. Therefore, we tested BP5-087 in the context of primary TKI resistance. BP5-087 (1 µM) in combination with imatinib (2.5 µM) reduced colony formation and increased apoptosis of CD34+ cells from CML patients with BCR-ABL1-independent resistance. These cells have no BCR-ABL1 kinase domain mutations and undergo BCR-ABL1 kinase inhibition as detected by immunoblot analyses. In contrast, BP5-087 had no effect on CD34+ cells from newly diagnosed CML patients or normal individuals. Immunofluorescence demonstrated that dual treatment of TKI-resistant CML CD34+ cells resulted in reduced levels of nuclear pSTAT3Y705, consistent with an inhibitor of STAT3 dimerization. In more primitive CML stem cells, long term culture-initiating cell (LTC-IC) assays revealed that neither inhibitor alone had any effect on colony formation of primitive LTC-IC progenitors, whereas imatinib (2.5 µM) in combination with BP5-087 (1.0 µM) reduced LTC-IC colony formation by 66%. Consistent with this observation, immunofluorescence showed high levels of pSTAT3Y705 in primitive TKI-resistant CD34+CD38- cells when cultured in the presence but not absence of TKIs. To test the feasibility of BP5-087 for in vivo use, we treated mice orally with BP5-087 (25 mg/kg/day) for 4 weeks and observed no changes in body weight, peripheral blood cellularity, or bone marrow colony forming ability. Mass spectrometry confirmed that BP5-087 is orally bioavailable. In summary, BP5-087 is a systematically-derived, direct inhibitor of STAT3 that, in combination with TKIs, reduces survival of CML cells with BCR-ABL1-independent resistance. Further rounds of structure-activity optimization may reveal an inhibitor with a clinically-relevant effective concentration. Disclosures: Deininger: Bristol Myers Squibb: Advisory Boards Other, Consultancy, Research Funding; Ariad Pharmaceuticals: Advisory Boards, Advisory Boards Other, Consultancy; Novartis: Advisory Boards, Advisory Boards Other, Consultancy, Research Funding; Celgene: Research Funding; Gilead Sciences: Research Funding.

Abstract 2778 In chronic myeloid leukemia (CML), imatinib and other tyrosine kinase inhibitors (TKIs) inhibit BCR-ABL1 tyrosine kinase activity but also target additional kinases including KIT. The role of KIT inhibition in the therapeutic efficacy of TKIs is controversial. We used TKIs with selective activity against ABL (PPY-A) or KIT (BAW667) and genetic tools to assess the role of KIT signaling for growth of CML cell lines and primary CML progenitor and stem cells. In Mo7eBCR-ABL1 or newly diagnosed CML CD34+ progenitor cells, immunoblotting confirmed that PPY-A (1 μM) suppresses BCR-ABL1 phosphorylation but not KIT tyrosine phosphorylation. In contrast, treatment of cells with a KIT-blocking antibody (K44.2, 200ng/mL), shRNA targeting KIT (shKIT), or the KIT selective inhibitor BAW667 (1 μM), suppressed KIT activity without affecting BCR-ABL1 kinase activity. Therefore, these systems are suitable to isolate the role of BCR-ABL1 vs. KIT inhibition. Treatment of Mo7eBCR-ABL1 cells with PPY-A resulted in suppression of growth by 91.7% (p To assess BCR-ABL1 vs. KIT inhibition in primary cells, CD34+ cells from newly diagnosed CML patients (n=4) and normal controls (n=3) were cultured in semisolid medium supplied with IL-3 and GM-CSF (no SCF), in the presence of 1 μM PPY-A combined with shKIT or 1 μM BAW667. KIT inhibition by shKIT or 1 μM BAW667 reduced CFU-GM formation by 40% compared to controls (p To assess the role of KIT vs. BCR-ABL1 inhibition on primitive CML cells we performed long-term culture-initiating cell (LTC-IC) assays on M2–10B4 murine stromal cells, using Lin− cells from newly diagnosed patients (n=3). Cultures were performed with K44.2, PPY-A, K44.2 plus PPY-A or 2 mM imatinib, with colonies plated at 1, 3, and 6 weeks. At 1 week colonies were reduced by 30% with K44.2 and 70% with PPY-A, but by 90% with the PPY-A / K44.2 combination or with imatinib. In contrast, at 6 weeks colony formation was unaffected by K44.2 but reduced by >95% with PPY-A, the PPY-A / K44.2 combination or imatinib. Week 3 colony growth was intermediate. Consistent with the LTC-IC assay, KIT inhibition with BAW667 enhanced PPY-A suppression of colony formation in Lin−CD34+CD38+ progenitor cells from newly diagnosed patients (n=3) by 18.7% (p Our findings suggest KIT inhibition is much more critical for suppression of mature progenitors compared to primitive CML cells. Since AKT is active in CML progenitors but suppressed by TGFβ in stem cells (Nature, 2010;463(7281):676; JCI, 2011;121(1):396), we speculate that upon BCR-ABL1 inhibition CML progenitors but not stem cells switch to an SCF-dependent mode of AKT activation, which renders these cells uniquely sensitive to dual inhibition of BCR-ABL1 and KIT signaling. Disclosures: No relevant conflicts of interest to declare.