Greenberg, Peter L.

Abstract Abstract 4013 Background: In recent years, azacytidine (AZA) has become the standard of treatment for high risk myelodysplasia (MDS). The approved schedule of AZA uses a 75mg/m2/d s.c. regimen for 7 days based on the CALGB-9221 (Silverman, JCO 2002) and the AZA-001 studies (Fenaux, Lancet Oncol 2009). The clinical response rates after AZA have been extensively presented but there is only limited data on the rates of cytogenetic response (CyR). Based on the review of the literature, there are no specific cytogenetic data published on prospective trials. Methods: We based our analysis on a randomized phase 2 study from the US Leukemia Intergroup (E1905 study, NCT00313586) testing 10 days of AZA (50mg/m2/d s.c.) vs 10 days of AZA+ the histone deacetylase inhibitor entinostat (4 mg/m2/d PO days 3 and day 10). MDS, CMML, and AML with myelodysplasia-related changes were included. This analysis includes all patients with cytogenetic abnormalities (at baseline or acquired following treatment) with available cytogenetic follow-up (cycle 6). Of 150 patients, 70 demonstrated baseline cytogenetic abnormalities. To date, forty patients (27 MDS and 13 AML) were evaluable for both time points. Karyotypes were performed at local laboratories, and reviewed centrally (RPK and GH). Cytogenetic response was assessed using IWG 2000 (Cheson et al, Blood 2000) criteria. Results: The clinical response rate (CR+PR+ trilineage HI) according to IPSS cytogenetic risk stratification were of 20%, 33%, and 35% for favorable, intermediate and poor cytogenetic risk groups respectively (p=NS). Patients with Chr 7 abnormalities (i.e. -7 or -7q, n=18) had a response rate of 28% including 17% CR. Of patients with complete cytogenetic data, the rate of overall CyR was 52% (n=21): 22% (n=9) complete CyR, 30% (n=12) partial CyR. This represents a complete CyR of 13% and a partial CyR of 23% as a proportion of all treated patients with initial cytogenetic abnormalities (including those who did not receive six cycles of therapy). To date, confirmatory FISH analyses were available for 4 patients with CyR (2 CCyR and 2 PCyR). All four had complete clearance of their cytogenetic clone. Among the cytogenetic responders, 15 had MDS and 6 had AML (p=NS). CyR did not differ between the two treatment arms. CyR and clinical response were highly correlated (p<0.001): all patients with complete CyR and 7/10 partial CyR had a clinical CR. Only 3 patients (1 CCyR and 2 PCyR) had a CyR without associated clinical response. Among 8 patients with cytogenetic progression, 3 pts showed increased frequency of a known aberration and 5 showed new or additional aberration including 2 cases involving Chr7. 65 AZA-treated patients from AZA001 had comparable cytogenetic results available. Only 3 patients experienced complete CyR 4.6% with 33.8% partial CyR (n=22). In contrast, the reported rate of clinical response was similar in the two studies. Complete CyR were more frequent in the E1905 10 days AZA schedule compared to the 7 days AZA schedule (p=0.007). There was no difference in overall CyR between the two studies. Conclusions: Both the standard AZA schedule and the 10 day schedule administered on E1905 lead to high rates of cytogenetic response. The novel schedule explored in E1905 demonstrated a particularly high rate of complete CyR (22%) with complete clearance of abnormal clone by FISH in 4 tested patients. Further prospective comparison of different AZA schedules will be needed to confirm this higher rate of complete CyR and determine its clinical significance. Disclosures: Beach: Celgene: Employment. Gore:Celgene: Consultancy, Research Funding, stock options; Syndax: Consultancy.

Abstract Background: MDSs are clonal hematopoietic stem cell malignancies characterized by cytopenias resulting from ineffective hematopoiesis. A phase 2 study of lenalidomide (MDS-002) demonstrated hematological improvement in patients with MDS without an associated deletion 5q cytogenetic abnormality (List AF, et al. Haematologica2006;91[suppl1]:379[abst 1032]). We now report multicenter results after 〉2 yrs follow-up. Methods: A total of 214 pts with transfusion-dependent anemia were categorized into the following IPSS MDS risk groups after central review of bone marrow morphology, karyotype, and peripheral blood findings: Low/Int-1 (168 [78%]); Int-2/High (8 [4%]); and unclassified (38 [18%]). Patients were unclassified for missing or inadequate marrow studies for central review (morphology [29] or cytogenetic [7]) or other diagnoses (AML [1], atypical CML [1]). Patients were evaluated for frequency and duration of red blood cell transfusion independence (RBC-TI) (IWG criteria), hemoglobin (Hgb) response, and safety. The starting dose of lenalidomide was 10 mg (daily or daily × 3 wks q28d cycle). Results: Overall, 56 (26%) of 214 enrolled patients achieved RBC-TI. Median time to RBC-TI was 5 wks, median duration of RBC-TI response was 41.0 wks (range, 8.0–136.4), and median Hgb increase achieved during RBC-TI was 3.2 g/dL (range, 1.0–9.8). An additional 36 patients experienced a ≥ 50% reduction in RBC transfusions (overall hematological improvement in 92 [43%] patients). Overall, 47 (22%) patients had an abnormal karyotype at baseline and 9 (19%) patients achieved a cytogenetic response (4 complete). The most common drug-related grade 3/4 adverse events (AEs) were neutropenia and thrombocytopenia (25% and 20%, respectively). Deep vein thrombosis occurred in only 2 (1%) patients. The duration of RBC-TI was ≥ 52 weeks in 21 (38%) patients. Among these 21 patients, 14 (67%) did not require a lenalidomide dose reduction during the first 52 weeks of treatment. Analyses of response variables will be presented. Dose-limiting neutropenia/thrombocytopenia was not reported after 52 weeks. The most commonly reported AEs after 52 weeks were mild-moderate diarrhea and fatigue (38% and 29%, respectively). Conclusion: Lenalidomide is an active, well-tolerated treatment in MDS patients with transfusion-dependent anemia that is not associated with a deletion 5q abnormality. The rate of erythroid hematologic improvement and duration of RBC-TI is encouraging and offers a possible alternative to cytokine therapy.

Abstract Background: Myelodysplastic syndromes (MDS) are hematological disorders characterized by chronic anemia, and are often refractory to cytokine therapy. Two phase 2 studies evaluating lenalidomide in MDS (MDS-001 and MDS-003), demonstrated a high frequency of erythroid response in patients with a chromosome 5q31 deletion (List AF et al., N Eng J Med2005; 352:549–557 and List AF et al., in press). Herein we report on the durability of response with long term follow-up approaching 4 years. Methods: All enrolled subjects were evaluated for frequency and duration of red blood cell transfusion independence (RBC-TI), change in hemoglobin (Hgb), pathologic and cytogenetic response, and safety. Results: Ten 5q31- patients achieved a cytogenetic and/or major erythroid response in MDS-001. Four of these patients have maintained a major response for at least 2 years (range 2.2 to 3.7 years). In MDS-003, 99 (67%) of 148 enrolled patients achieved RBC-TI. Of the 99 responders 83 were Low/Int-1, 3 were Int-2/High and 13 were not classified. Median increase in Hgb from baseline to maximum achieved during RBC-TI was 5.4 g/dL (range 1.1, 11.4). The estimated median duration of response was 115.9 weeks, and 52 (53%) of 99 responders remain on study with ongoing RBC-TI response as of 30 June 2006. Duration of response was at least 52 weeks for 63 (64%), 78 weeks for 52 (53%), and 104 weeks for 36 (36%) of 99 responding patients. Variables (or co-variates) associated with longer duration of TI in multivariate analysis included low baseline transfusion requirement (<4 units per 8 wks, p=0.002), 5q- syndrome (p=0.01) and low IPSS (p=0.02) classifications. Among 85 evaluable patients, major cytogenetic responses were achieved in 37 (44%) and minor responses in 24 (28%). Cytogenetic responses occurred in patients with complex karyotypes as well as those with an isolated 5q31 deletion. The frequencies of major and minor cytogenetic responses were similar among patients with and without 5q- syndrome. The most common drug-related adverse events (AE) were Grade 3/4 neutropenia and thrombocytopenia, reported in 87 (59%) and 74 (50%) of 148 patients, respectively although hematologic AEs led to study discontinuation in only 13 (9%) of patients. Pneumonia was the only serious adverse event that occurred in >10% of patients (15/148; 10%). Adverse events were manageable through supportive treatment and dose reduction. Conclusion: Lenalidomide is an effective treatment in MDS patients with an associated chromosome 5q31 deletion, with RBC-TI and major erythroid response maintained for median duration > 2 years.

Abstract Background: Rigosertib (RIG) is a RAS-mimetic interacting with the RAS-binding domains of RAF kinases, preventing their binding to RAS and inhibiting the RAS-RAF-MEK pathway. Rigosertib also binds to the RAS-binding domains of Ral-GDS and PI3Ks (Athuluri-Divakar, et al, Cell 165:643, 2016). Mutations of RAS are involved in proliferative processes important in neoplastic transformation (Gil-Bazo, et.al, Cancer Biology and Therapy 17:719, 2016). As of May 2016, 557 patients with MDS or AML received IV (N=335) or oral (N=222) rigosertib. All IV rigosertib was administered as monotherapy. Oral rigosertib was administered as monotherapy (N=168) or in combination with injectable azacitidine (AZA; N=54). Safety results across this rigosertib development program have not been presented. Methods: We reviewed all treatment-emergent adverse events (TEAEs) among patients with MDS/AML treated with rigosertib IV monotherapy, rigosertib oral monotherapy, or in combination therapy, overall and by grade of severity, and serious TEAEs and TEAEs leading to discontinuation. Due to the unexpected observation of urinary AEs, including occasional gross hematuria, we analyzed urinary AEs in greater detail. Results: Rigosertib IV monotherapy: The most common TEAEs were fatigue (33%), nausea (33%), diarrhea (27%), constipation (25%), anaemia (24%) and pyrexia (24%). The most common ≥Grade 3 AEs were anaemia (21%), febrile neutropenia (13%), pneumonia (12%) and thrombocytopenia (11%). The most common serious AEs were febrile neutropenia (10%), pneumonia (9%), and sepsis (7%). The most common AEs leading to discontinuation of IV rigosertib were sepsis and pneumonia (3% each) (Table 1). Rigosertib oral monotherapy: The most common TEAEs were pollakiuria (increased urinary frequency) (35-%), fatigue (32%), diarrhoea (26%), dysuria (29%) and haematuria (24%). The most common ≥Grade 3 AEs were anaemia (17%), thrombocytopenia (5%), haematuria (4%) and urinary tract infection (4%). The most common serious AE was pneumonia (6%). The most common AEs leading to discontinuation of patients receiving oral rigosertib as monotherapy were (dysuria [8%], urinary tract pain [7%], haematuria [5%] and urinary frequency [5%] (Table 2). Rigosertib oral in combination with azacitidine (AZA): The most common TEAEs were nausea (41%), fatigue (39%), diarrhea (37%), constipation (37%) and dysuria (28%). The most common serious AEs were pneumonia (11%) and febrile neutropenia (7%). The most common AEs leading to discontinuation were AML (4%) and pneumonia (4%) (Table 3). In patients fully evaluable for gross hematuria the cause of bleeding was bladder inflammation. Urinary and plasma concentration of RIG were examined to assess dose relationship. According to the azacitidine package insert, the incidence of gross hematuria of any grade with single agent azacitidine is 6.3 % and Grade 3-4 of 2.3%. Evaluations performed of hematuria among 5 studies of oral rigosertib (4 monotherapy and 1 in combination with AZA) in patients with MDS/AML (Table 4). Conclusions: Rigosertib IV or oral formulations were generally well tolerated in clinical trials in over 500 patients with MDS/AML. Gastrointestinal AEs were most frequently reported with IV rigosertib and genitourinary AEs were seen more often with oral than IV dosing. The greater rate of urinary AEs in the oral than in the IV studies indicates a need for close monitoring of these patients and proactive risk management, such as adequate hydration, timing of administration of the second dose, and bladder emptying prior to sleep. Studies are being designed to optimize dosing and schedule to maximize efficacy and minimize hematuria of single agent oral rigosertib. A randomized trial of oral rigosertib/azacitidine versus azacitidine will be required to optimize the clinical benefit and safety in the management of higher-risk patients with MDS. Disclosures Fenaux: Celgene: Honoraria, Research Funding; Janssen: Research Funding; Novartis: Honoraria, Research Funding; Astex: Research Funding; Teva: Honoraria, Research Funding. Al-Kali:Celgene: Research Funding; Onconova Therapeutics, Inc.: Research Funding. Navada:Onconova Therapeutics, Inc.: Research Funding. Roboz:Agios, Amgen, Amphivena, Astex, AstraZeneca, Boehringer Ingelheim, Celator, Celgene, Genoptix, Janssen, Juno, MEI Pharma, MedImmune, Novartis, Onconova, Pfizer, Roche/Genentech, Sunesis, Teva: Consultancy; Cellectis: Research Funding. Santini:Celgene: Consultancy, Honoraria, Research Funding; Janssen: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Onconova: Consultancy; Amgen: Consultancy; Astex: Consultancy. Platzbecker:Onconova, Teva, Celgene, Janssen, Novartis, Amgen: Honoraria, Research Funding. Petrone:Onconova Therapeutics, Inc.: Employment. Brownstein:Onconova Therapeutics, Inc.: Consultancy. Zbyszewski:Onconova Therapeutics, Inc.: Employment. Maniar:Onconova Therapeutics, Inc.: Employment. Silverman:Onconova Therapeutics, Inc.: Patents & Royalties: Co-Patent holder for the combination of azacitidine and rigosertib, Research Funding.

Abstract Background: Patients (pts) withHR-MDS have a median OS of 4 to 6 months (mo) after HMA failure (Prebet et al, J Clin Oncol 2011) and no approved salvage therapy. Development of new therapeutics for this population will benefit from the availability of surrogate endpoints and markers that can predict survival. Gore et al established response to azacitidine (Vidaza®) in first-line therapy for HR-MDS as a reasonable surrogate to predict survival (Gore et al, Haematologica 2013). Rigosertib, a novel dual PI3K/PLK pathway inhibitor, has been shown to reduce bone marrow blasts (BMBL) in these pts (Seetharam et al, Leuk Res 2012). Silverman et al described complete or partial bone marrow (BM) response, or stabilization after 4-8 weeks (wks) of treatment with rigosertib as a potential surrogate for predicting survival in pts with HR-MDS after failure of primary HMA therapy (Silverman et al, Hematol Oncol 2014). We tested this hypothesis in the context of a randomized Phase III trial. Methods:Pts with HR-MDS were randomly assigned 2:1 to rigosertib or best supportive care (BSC) after progressing on, failing to respond to, or relapsing after HMA treatment. BM aspirates were assessed pretreatment, at 4 weeks and at 8-week intervals thereafter. Central slide review was undertaken in a representative population of samples. The BMBL response at each time point was assessed using the following definitions: bone marrow complete response (mCR) = BMBL ≤ 5% and decrease of ≥ 50% from baseline; bone marrow partial response (mPR) = BMBL decrease from baseline of ≥ 50%, but BMBL still > 5%; stable disease (SD) = BMBL decrease or increase from baseline of < 50%; progressive disease (PD) = BMBL increase from baseline of ≥ 50% by an absolute minimum of 5%; Not evaluable (NE). Results:Bone marrow assessment was carried out in 156 patients (pts) on the rigosertib arm and 24 pts on the BSC arm at 4 wks after enrollment, and in 86 and 20 pts, respectively, at 12 wks. The invasive BM procedure was optional on the BSC arm, which accounts for the low number of assessments in this group. BM responses at the 2 time points are presented in Table 1. Since no difference in overall survival was noted between pts who had objective BM response and those who did not progress (ie, stable disease), a landmark analysis was conducted that separated pts who were alive at the 4-wk landmark time into two 4-wk response categories: BM response + SD vs. PD. Results of this analysis in rigosertib-treated patients were statistically significant at p = 0.011, with a hazard ratio (HR) of 0.62 and a median OS (from 4 wks onward) of 9.8 months in the mCR + mPR + SD group vs. 4.6 months in the PD group (Figure 1). Another landmark analysis was conducted at 12-wks. Results of this analysis were also significant (p < 0.001) in rigosertib-treated patients, with an HR of 0.39 and a median OS (from 12 wks onward) of 10.4 months in the mCR + mPR + SD group vs.7.5 months in the PD group (Figure 2). A time-dependent Cox regression of OS by 4-wk BMBL response reinforced the validity of the 4-wk and 12-wk BM assessments as surrogate biomarkers for survival (Table 2). Conclusions: These data suggest that BMBL response at 4 or 12 weeks was correlated with OS in this population of pts with HR-MDS treated with rigosertib after HMA failure and are consistent with previous observations in Phase II studies. Table 1 4- and 12-week Bone Marrow Blast Response (Intention-to-Treat Population) Number (%) of Patients 4-wk BMBL Response 12-wk BMBL Response Rigosertib N = 199 BSC N = 100 Rigosertib N = 199 BSC N = 100 Pts with BMBL assessment 156 (78) 24 (24)* 86 (43) 20 (20)* BM complete response (mCR) 22 4 11 5 BM partial response (mPR) 8 2 9 2 Stable disease (SD) 77 9 32 8 Progressive disease (PD) 49 9 34 5 * Bone marrow assessment was not required on the BSC arm. Figure 1 Figure 1. Figure 2 Figure 2. Table 2 Time-dependent Cox Regression of Overall Survival by Bone Marrow Blast Response Analysis Rigosertib BSC Wald P-value Hazard Ratio (95% Confidence Interval) Wald P-value Hazard Ratio (95% Confidence Interval) By 4-wk BMBL response 0.051 0.72 (0.51 - 1.00) 0.56 0.83 (0.45 - 1.54) By 12-wk BMBL response 0.0005 0.55 (0.39 - 0.77) 0.16 0.68 (0.39 - 1.17) *Stratified by pretreatment BMBL: 5%-19% vs. 20%-30% Disclosures Fenaux: Celgene: Research Funding; Janssen: Research Funding; Novartis: Research Funding. Sekeres:Celgene Corp.: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Boehringer Ingelheim: Membership on an entity's Board of Directors or advisory committees. Wilhelm:Onconova Therapeutics, Inc: Employment, Equity Ownership. Azarnia:Onconova Therapeutics, Inc: Employment.

Abstract Background: No approved treatment options are available to HR-MDS pts after HMA therapy. Study 04-21 (“ONTIME” trial) was a Phase III, randomized, controlled study of the efficacy and safety of rigosertib, a novel small molecule inhibitor of PI3-kinase and PLK pathways, in a heterogeneous population of MDS pts who had relapsed after, failed to respond to, or progressed during administration of HMAs. The study was conducted at 87 sites in the United States and 5 European countries. Methods:From Dec 2010 to Aug 2013, 299 HR-MDS pts [<30% bone marrow blasts (BMBL)] who had progressed on (37% of total enrollment), failed to respond to (25%), or relapsed after (38%) HMA treatment were stratified on BMBL count and randomized 2:1 to receive rigosertib (199 pts) or BSC (100 pts). Rigosertib was administered at 1800 mg/24 hr for 72-hr as a continuous intravenous (CIV) ambulatory infusion, every 2 weeks for the first 16 weeks, and then every 4 weeks. The primary endpoint was overall survival (OS), analyzed on an intention-to-treat (ITT) basis using the Kaplan-Meier method stratified on BMBL (5% to 19% vs. 20% to 30%). The trial had a 95% power to detect a 13-wk increase in median OS from 17 wks on BSC, with a 2-sided alpha = 0.05. The following results are based on 242 deaths: 161 in the rigosertib arm and 81 in the BSC arm. Results : Overall, the 2 arms were balanced in terms of baseline characteristics, with the majority of pts being male (66%), and White (82%). Age ranged from 50-90 yrs in the rigosertib arm and 55-86 years in the BSC arm (median, 74 yrs). The majority of pts (85%) had an Eastern Cooperative Oncology Group (ECOG) score of 0 or 1. The median duration of the last HMA therapy was 8.8 months (mo) in the rigosertib arm and 10.3 mo in the BSC arm; 127 (64%) of rigosertib pts and 57% of BSC pts were classified as “primary HMA failure” (ie, they failed to respond to or progressed during HMA therapy, as defined by Prebet et al, J Clin Oncol, 2011). A 2.3-mo improvement in median OS was found in the overall (ITT) population (8.2 mo rigosertib vs. 5.9 mo BSC) (Figure 1). The ITT survival for rigosertib was similar to that noted in Phase I/II studies (35 weeks). The stratified log-rank p-value was 0.33. The stratified hazard ratio was 0.87, which was quite different from the ratio of medians (5.9/8.2 = 0.72), due to the fact that the 2 survival curves converged at 15 mo. Notably, among the 184 patients with primary HMA failure, the median OS was 8.6 mo in the rigosertib arm (N = 127) vs. 5.3 mo in the BSC arm (N = 57), HR= 0.69, p= 0.040 (Figure 2). Multivariate Cox regression, adjusting for pretreatment prognostic factors, showed little change in the treatment effect. The following subgroups were correlated with better OS: pts with failure of/progression on HMA treatment, pts with duration of HMA treatment ≤ 9 mo, pts < 75 years of age, and pts with very high risk per IPSS-R (Figure 3). Rigosertib was well tolerated, with a median dose intensity of 92%. There were no significant compliance or operations issues related to ambulatory continuous infusion. Protocol-defined dose reductions were reported in 5% of pts, with 24% experiencing dose delays of >7 days, mostly due to unrelated adverse events (AEs). No obvious differences between rigosertib and BSC were found in the incidence of AEs (rigosertib, 99%; BSC, 85%) or of ≥ Grade 3 AEs (rigosertib, 79%; BSC, 68%). In the rigosertib arm, AEs reported by ≥ 20% of pts, irrespective of severity or causality, were nausea (35%), diarrhea (33%), constipation (31%), fatigue (30%), fever (27%), anemia (22%), and peripheral edema (21%). Rigosertib had low myelotoxicity, consistent with previous clinical experience. Conclusions:Although the primary endpoint in this Phase III study of rigosertib vs BSC in pts with HR-MDS did not reach statistical significance in the ITT population, encouraging rigosertib treatment-related improvement in OS was noted in several subgroups of MDS pts, including those with “primary HMA failure and in patients in the IPSS-R Very High Risk category. CIV therapy with rigosertib had a favorable safety profile in this orphan population of elderly pts with MDS. Figure 1 Figure 1. Figure 2 Figure 2. Figure 3 Figure 3. Disclosures Fenaux: Celgene: Research Funding; Janssen: Research Funding; Novartis: Research Funding. Sekeres:Celgene Corp.: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Boehringer Ingelheim: Membership on an entity's Board of Directors or advisory committees. Roboz:Novartis: Consultancy; Agios: Consultancy; Celgene: Consultancy; Glaxo SmithKline: Consultancy; Astra Zeneca: Consultancy; Sunesis: Consultancy; Teva Oncology: Consultancy; Astex: Consultancy. Wilhelm:Onconova Therapeutics, Inc: Employment, Equity Ownership. Wilhelm:Onconova Therapeutics, Inc: Employment. Azarnia:Onconova Therapeutics, Inc: Employment. Maniar:Onconova Therapeutics, Inc: Employment.

Abstract Abstract 1875 Recent experimental evidence suggests that acute myeloid leukemia (AML) originates from hematopoietic stem and progenitor cells (HSPC) following the acquisition of multiple genetic or epigenetic changes that initially give rise to pre-leukemic HSPC (pre-LSC) and then to fully transformed leukemia stem cells (LSC). Relapse continues to be the major cause of death in most subtypes of AML, suggesting that current therapies are largely ineffective in eliminating LSC and pre-LSC. Cellular heterogeneity and the recent observation that LSC are contained within different phenotypic cellular compartments are challenges for the identification of pathways contributing to the initiation and maintenance of AML. To address these challenges we employed a novel strategy of parallel transcriptional analysis of multiple phenotypic HSPC populations from individuals with AML with normal karyotype (N=5), -7/7q- (N=6) and complex karyotype (N=5), including long-term HSC, short-term HSC, and granulocyte-monocyte progenitors (GMP), and comparison to corresponding cell populations from age-matched healthy controls (HC) (N=6). Specifically, we sorted Lin−/CD34+/CD38−/CD90+ (LT-HSC), Lin−/CD34+/CD38−/CD90− (ST-HSC), and Lin−/CD34+/CD38+/CD123+/CD45RA+ (GMP), and hybridized RNA to Affymetrix GeneST 1.0 expression arrays. Differential gene expression was determined within each compartment by direct comparison of AML LT-HSC vs. HC LT-HSC, AML ST-HSC vs. HC ST-HSC, and AML GMP vs. HC GMP. Subsequent intersection of all differentially expressed genes revealed that only a relatively small number of 6 to 11 genes were consistently dysregulated in all examined leukemic stem and progenitor cell compartments. Interleukin 1 receptor accessory protein (IL1RAP) was one of the most significantly upregulated genes in LT-HSC, ST-HSC, and GMP in all examined subtypes of AML. IL1RAP is a transmembrane protein required for signaling through several receptors of the IL1 family, including IL-1R1 and ST2. We detected significant overexpression of IL1RAP protein on HSC and progenitor cells of AML patients. Interestingly, CD34+/Lin+ precursor cells showed only a marginal increase of IL1RAP at the protein level in AML, underscoring the importance of purifying HSPC with stringent lineage depletion. We performed fluorescence in situ hybridization in sorted IL1RAP+ and IL1RAP− cells from -7 AML. We observed that the -7 clone was restricted to IL1RAP+ cells, while IL1RAP- cells did not display monosomy 7, demonstrating that IL1RAP overexpression is a distinguishing feature of the cells of the -7 clone. Patients with normal karyotype AML showed a wider range of IL1RAP expression levels; some were as high as in -7 AML and others were as low as in HC. We asked whether IL1RAP expression levels were associated with known clinical or molecular parameters. We analyzed two published datasets of patients with normal karyotype AML (Metzeler, Blood. 2008;112:4193–4201; Tomasson, Blood. 2008;111:4797–4808). Patients with high IL1RAP levels showed inferior overall survival than patients with lower IL1RAP (p=2.2×10−7; median survival: 7.82 mo. for IL1RAP high, 20 mo. for IL1RAP low). Multivariate analysis using a Cox regression model showed that high IL1RAP status was an independent prognostic factor (p=0.002), and even stronger than FLT3 mutation status (p=0.006). In addition, we analyzed data from 183 patients with MDS and found IL1RAP expression to be specifically elevated in cases with RAEB-2, suggesting a role of IL1RAP in MDS and in the progression to AML. Downregulation of IL1RAP protein expression in 4 AML cell lines (THP1, OCI-AML3, HL60, HEL) led to a significant 45–98% decrease in clonogenic growth and increased apoptosis in vitro. To assess the effects of IL1RAP downregulation in vivo, we performed xenotransplants into immunodeficient NOD/SCID/IL2Rg-null mice. THP-1 AML cells showed a 92% reduced proliferation and infiltration of hematopoietic organs upon IL1RAP knockdown in comparison to a non-silencing control in vivo. Genetic studies to assess the role of IL1RAP in the initiation and maintenance of AML in an IL1RAP−/− mouse model are currently ongoing. In summary, our study provides a map of consistently dysregulated transcripts across multiple fractionated stem and progenitor cell types from patients with AML, and identifies IL1RAP as a putative new therapeutic and prognostic target in stem cells in AML and MDS. Disclosures: No relevant conflicts of interest to declare.

7056 Background: No therapies are approved for MDS after HMA failure. 04-24 was a single-arm study to evaluate best BMBL response as a potential surrogate for OS in higher-risk (HR) MDS pts who progressed on or after an HMA. Rigo is a Ras-mimetic that inhibits the RAS-RAF-MEK pathway, which is frequently activated in HR MDS (Athuluri-Divakar Cell 2016; Gil-Bazo Cancer Biol Ther 2016). Methods: Eligible MDS pts had 5-30% BMBL confirmed within 6 wks pre-study and progression per International Working Group (IWG) 2006 criteria on or after HMAs within 2 yrs. Rigo 1800 mg/24 hrs was continuously infused over 72 hrs q 2 wks × 8 cycles, then q 4 wks until progression or unacceptable toxicity. Primary endpoint was relationship of best BMBL IWG response to OS by Kaplan Meier method. Results: 64 pts were treated (median 5 cycles, range 1-32+), with 61% male, median age 73 (range 47-87), median prior HMA duration 10.8 mos (range 1.2-70.2). Revised International Prognostic Scoring System scores were low 2%, intermediate 11%, high 27%, very high 53%, and unknown 8%. ≥Grade 3 adverse events in ≥10% of pts were anemia 19%, thrombocytopenia 19%, and febrile neutropenia 16%. At the analysis time 40 pts (63%) had died. Best BMBL IWG response was marrow complete response (mCR) 14 pts (22%), stable disease (SD) 30 (47%), progressive disease (PD) 15 (23%), and failure (early death/withdrawal) 5 (8%); 2 mCR pts had transplant. Median OS was 7.0 mos (95% confidence interval 4.8-10.8). Landmark median OS (from day of best BMBL response) was mCR not reached; SD 6.3 mos; PD 3.3 mos. Median OS of mCR+SD was 8.5 mos, with log-rank p = 0.011 (mCR+SD OS to PD OS). Conclusions: BMBL response is a predictor of survival for MDS pts receiving Rigo after HMA failure, confirming findings in earlier Phase 1/2 studies (Silverman ASCO 2015 Abstr 7017). Based on earlier results identifying an MDS subset benefitting from Rigo (Garcia-Manero Lancet Oncol 2016; ASCO 2016 Abstr 165681), a randomized Phase 3 trial of Rigo vs physician’s choice (INSPIRE) is ongoing to determine if Rigo improves survival after HMA failure within 9 cycles. Clinical trial information: NCTO 1928537.