Biallelic CEBPA Mutations Research Articles

WT1 Mutations Are Secondary Events In AML and Show Varying Frequencies Within Genetic Subgroups and Different Impact On Prognosis

BackgroundMutations (mut) in the WT1 gene belong to the first genetic aberrations described in AML. In contrast to recurrent fusion genes or NPM1mut WT1mut do not seem to be disease defining. Also in contrast to other mutations in AML, for most of which a certain prognostic value has been established, the impact of WT1mut still is discussed controversially. AimAnalyze the frequency and prognostic impact of WT1 mutations in comparison to other genetic aberrations. Patients and Methods3,157 unselected AML patients (pts) were analyzed (de novo: n=2,699, s-AML: n=234, t-AML: n=224). 1,708 pts were male and 1,449 female. Median age was 67.1 years (y) (range: 17.8-100.4 y) with 1,108 pts <60 y and 2,049 ≥60 y. The mutational hot spot regions of WT1 (exons 7 and 9) were analyzed by direct Sanger sequencing with a sensitivity of ∼10%. Karyotype and WT1 mutation status was available in all cases. Other mutations were assessed in subsets: ASXL1 (n=1,951), CEBPA (n=2,670), DNMT3A (n=1,293), FLT3-ITD (n=3,149), FLT3-TKD (n=3,004), IDH1R132 (n=2,431), IDH2R140 (n=2,380), IDH2R172 (n=2,412), KRAS (n=1,409), NRAS (n=1,780), NPM1 (n=3,003), MLL-PTD (n=2,961), RUNX1 (n=2,390), TET2 (n=1,016) and TP53 (n=1,215). ResultsA total of 189 WT1 mutations were detected (exon 7: n=151, exon 9: n=38). The total frequency of WT1mut pts was 175/3,157 (5.5%). 11 pts were double to quadruple mutated. The frequency was heterogeneous with respect to AML subtypes. Compared to all others, significantly higher frequencies were detected in biallelic CEBPAmut (15/110; 13.6%; p=0.001), followed by t(15;17)/PML-RARA (18/164; 11.0%, p=0.004), and FLT3-ITD (58/682; 8.5%, p<0.001). Lower frequencies were observed in DNMT3Amut (18/412; 4.3%, p=0.014, ASXL1mut (6/355; 1.7%, p<0.001), IDH2R140 (5/286; 1.7%, p=0.001), and IDH1R132 (2/222; 0.9%, p<0.001). WT1mut were never detected in pts with complex karyotypes (0/175; p=0.047) or those with IDH2R172 (0/68; p=0.020).Further, WT1mut were more frequent in females (95/1,449, 6.6%) than in males (80/1,708, 4.7%) (p=0.014) and in younger pts (<60 y: 102/1,108, 9.2% vs ≥ 60 y: 73/2,049, 3.6%; p<0.001). Median age of pts with WT1mut was 55.5 y compared to 63.6 in WT1wt (p<0.001). Further, WT1mut were associated with lower platelet count (58.4 vs 84.7 x109/L; p<0.001) and lower hemoglobin level (8.8 vs 9.3 g/dL, p=0.001). There was no association to the history of the disease or white blood cell count.Stability of WT1mut was analyzed in 35 paired diagnostic and relapse samples (median time of relapse after diagnoses: 11.1 months (m); range: 2.6-60.6 m). In 23 cases (65.7%) the WT1mut was retained at relapse and in 12 cases (34.3%) it was lost. In 5 cases a sample at 2nd relapse was available (median time from 1st relapse: 8.5 m, range: 6.0-18.0 m). 3 of these cases retained and 2 lost the WT1mut.Analysis of prognostic impact was restricted to intensively treated pts (n=1,936, WT1mut: n=132, 6.8%). In the total cohort, there was no impact of WT1mut on prognosis. In pts ≥60 y there was a trend to shorter event free survival (EFS) for WT1mut (9.3 vs 12.3 m, p=0.052). In the two prognostically favorable groups with high WT1mut incidences (biallelic CEBPAmut and PML-RARA) no effect on outcome was seen. When restricting the analysis to normal karyotype AML (WT1mut: n=85, WT1wt: n=1,093) WT1mut pts had shorter EFS (10.8 vs 17.9 m, p=0.008). This was true for the younger (12.2 vs 29.0 m, p=0.007) as well as for the older pts (9.3 vs 13.9 m, p=0.016). In a multivariate analysis all parameters with significant impact on EFS in univariate analysis were included: age (p<0.001, HR: 1.24), ASXL1mut (p<0.001, HR: 1.36), FLT3-ITD (p<0.001, HR: 1.55), NPM1mut/FLT3-ITD wild-type (p<0.001, HR:1.55), RUNX1 (p=0.019, HR: 1.23, and WT1mut (p=0.009, HR: 1.64). In multivariate analysis WT1mut was found to have independent adverse impact on EFS (p=0.002, HR: 1.64) besides FLT3-ITD status (p<0.001, HR: 1.71) and age (p<0.001, HR: 1.28). ConclusionsWT1 mutations are 1) more frequent in females and younger AML, 2) more frequent in t(15;17)/PML-RARA, biallelic CEBPAmut, FLT3-ITD mutated AML, and nearly mutually exclusive of ASXL1, IDH1, IDH2 and complex karyotype. 3) The distribution pattern in different genetic subtypes and the instability during follow-up as shown by paired sample analyses clearly emphasize a secondary character of this mutation. 4) For AML with normal karyotype an independent adverse impact of WT1mut on EFS was shown. Disclosures:Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Eder:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Dicker:MLL Munich Leukemia Laboratory: Employment. Ulke:MLL Munich Leukemia Laboratory: Employment. Kohlmann:MLL Munich Leukemia Laboratory: Employment. Kuznia:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Coinhibitory Expression Profiling of AML Blasts Identifies HVEM as Prognostic Marker in AML

Acute myeloid leukemia (AML) is a highly heterogeneous disease. Numerous cytogenetic and molecular markers distinguish different biological and prognostic groups. Immunological effects play an important role for the treatment of AML, especially for the eradication of minimal residual disease, as proven by the success of allogeneic stem cell transplantation and some autologous immunotherapeutic strategies. However, various immunosuppressive factors dampen these potential effects. For several malignancies, the relevance of coinhibitory receptor ligand interactions that influence the interaction of tumor cells with specific immune effector cells such as T and NK cells has recently been proven by clinical studies with respective blocking antibodies.In order to investigate the potential significance of this mechanism in AML, we measured the surface expression of a broad panel of immune checkpoint proteins (CD80, CD86, CD273, CD274, CD275, CD276, B7-H4, HVEM, OX40L) on 209 AML samples at initial diagnosis by flow cytometry. The specific fluorescence intensity of these markers on the CD33 positive blast cell population was correlated with morphologic, cytogenetic and molecular characteristics of the disease. Subsequently, we analyzed the potential prognostic significance of immune checkpoint protein expression on overall and relapse-free survival after achievement of complete remission (median observation time: 225 days). All patients were included in the AML registry of the AML Cooperative Group and treated according to their treatment recommendations.A few of these molecules (CD80, CD273, CD275, B7-H4) were not detectable in the majority of cases and thus do not seem to play a role in the interaction of AML blast cells with immune effector cells, but others showed a significant and patient-specific expression. We found that AMLs of monocytic lineage (M4 or M5 according to FAB) expressed higher levels of CD86 (p < 0.001) and CD276 (p < 0.001), compared to the other morphologic subgroups. The FLT3-ITD correlated with lower expression of CD274 (p=0.001) and CD276 (p = 0.007), while the prognostically favorable subgroup with an isolated NPM1 mutation (without accompanying FLT3-ITD) and normal karyotype correlated with higher expression of CD86 (p = 0.049) and CD276 (p = 0.033) and lower expression of OX-40L (p = 0.039). The patients with very high expression of both CD86 and CD274 had a worse overall (p = 0.006) and relapse-free (p = 0.040) survival.Of particular interest was the analysis of HVEM, a member of the TNF-receptor superfamily, which has rarely been described in the context of AML so far. Its expression was reduced both in the groups with FLT3-ITD (p = 0.001) and with NPM1mut/FLT3wt/CN (p = 0.049), but highly increased in leukemias with normal karyotype and biallelic CEBPA mutation (p = 0.008). Of note, a high expression of HVEM correlated with significantly higher relapse-free survival within the prognostically intermediate group according to ELN (p = 0.005).We thus for the first time provide evidence that the profile of immune checkpoint molecules on CD33 positive leukemic cells correlates with molecular disease characteristics in AML and may even possess prognostic information, especially for relapse. Based on these data, we are currently evaluating different immune checkpoint molecules for their potential as therapeutic targets with respect to boosting antileukemic immune responses. Disclosures:No relevant conflicts of interest to declare.

A Combined Molecular and Clinical Prognostic Index For Relapse and Survival In Cytogenetically Normal AML (PINA)

BackgroundCytogenetically normal acute myeloid leukemia (CN-AML) is the largest and most heterogeneous cytogenetic AML subgroup. For the practicing clinician it is difficult to know how to use the prognostic information of the growing number of clinical and molecular markers. Our purpose was to develop a widely applicable prognostic model by combining well-established pre-treatment patient and molecular characteristics. Patients and methodsTwo prognostic indices for CN-AML, one with regard to overall survival (PINAOS) and the other regarding relapse-free survival (PINARFS) were derived based on a cohort of 669 CN-AML patients treated within the AML Cooperative Group 99 (AMLCG99) study. ResultsBased on age (median: 60 years [range: 17-85 years]), performance status, white blood count, and presence or absence of NPM1 mutation, biallelic CEBPA mutation, and FLT3-ITD, patients were classified into three risk groups according to PINAOS and PINARFS: 29% of all and 32% of responding patients had low risk (5-year OS 72%; 5-year RFS 55%), 56% and 39% intermediate risk (5-year OS 28%; 5-year RFS 27%), and 15% and 29% high risk disease (5-year OS 3%; 5-year RFS 8%) (Figure 1). PINAOS and PINARFS further subdivided the European LeukemiaNet (ELN) favorable-genetic group as well as the ELN intermediate-I-genetic group. Both, PINAOS and PINARFS were confirmed in a large, independent, and comparable CN-AML cohort of 529 patients from the Cancer and Leukemia Group B (CALGB/Alliance) trials (Figure 2). [Display omitted] [Display omitted] ConclusionsWe have developed and validated the first prognostic indices specifically designed for CN-AML patients of all ages combining well-established molecular and clinical variables easily applicable in routine clinical care. The integration of both clinical and molecular markers could provide a basis for individualized patient care by risk-adapted therapy of CN-AML. Disclosures:No relevant conflicts of interest to declare.

BAALC-Associated Mir-3151 Is An Independent Prognostic Factor In Younger Patients With Intermediate-Risk Cytogenetic Acute Myeloid Leukemia

IntroductionPatients with intermediate-risk cytogenetic AML (IR-AML) have a heterogeneous prognosis, and are currently further stratified based on determined gene mutations. However, the optimal post-remission therapy, especially in younger patients, is unclear. Recently, miR-3151, a novel microRNA (miRNA) located in intron 1 of the BAALC gene, has been identified. High miR-3151 expression –either alone or in combination with high BAALClevels– independently correlates with poor prognosis in patients ≥ 60 years with normal cytogenetics AML (CN-AML) (Eisfeld AK, et al. Blood 2012). However, the prognostic value of miR-3151 in younger patients (≤60 years) with IR-AML has not been examined. We hypothesized that miR-3151 expression could also be a prognostic marker in younger patients. AimTo analyze whether miR-3151 expression – either alone or in combination with BAALC– improved prognostic assessment in younger patients (up to 60 years) with IR-AML and to characterize in this subset of patients the miRNA signature associated with high miR-3151 expression. MethodsSamples were available from two separate cohorts of patients with IR-AML who had received intensive therapy: a training set of 76 patients from a single institution and a validation set of 108 patients from several centers who had been treated within the CETLAM AML-2003 protocol. Information on NPM1, FLT3-ITD and CEBPA was available for both patient cohorts. The expression levels of 670 miRNAs had previously been analyzed in the 76 patients in the training set. In the present study, the expression of miR-3151 and BAALC was analyzed using TaqMan® MicroRNA Assay and TaqMan® Gene Expression Assay (Applied Biosystems), respectively. Expression levels of miR-3151 and BAALC –both alone and in combination – were correlated with patient outcome. Statistical analyses were performed with SPSS v.15.0.1, R software v.2.9.0 and TIGR MultiExperiment Viewerv4.0. ResultsIn the training set, higher expression of miR-3151(dichotomized by its median value of normalized expression) correlated with a shorter 5-year overall survival (OS) (32%±17% vs. 61±17%, p=0.029) and 5-year leukemia-free survival (LFS) (29%±17% vs. 58±17%, p=0.036) in patients ≤ 60 yrs. When the analysis was restricted to patients with CN-AML, miR-3151 expression retained its prognostic significance (p= 0.016). In addition, increased BAALCexpression was associated with shorter OS (28%±20% vs. 58±14%, p=0.054) and LFS (17%±18% vs. 55±14%, p=0.039).In the multivariate analysis for OS and LFS, including age, WBC, NPM1mut, FLT3-ITD, BAALC and miR-3151 expression levels as covariates, miR-3151 showed independent prognostic significance for OS (p=0.016; HR=2.52, 95% CI: 1.2-5.4),and a statistical trend for LFS (p=0.09).Patients with low expression of both miR-3151 and BAALC had better prognosis (OS: 66%±18% vs. 34±16%, p=0.027; LFS: 67%±20% vs. 27±16%, p=0.009).Interestingly, the combination of both miR-3151 and BAALC retained a significant prognostic value for LFS within the favorable molecular subgroup/ ELN favorable subgroup (patients harboring NPM1mut without FLT3-ITD or biallelic CEBPAmut; LFS: 44%±30% vs. 100%, p=0.017) and showed a trend in the unfavorable molecular subgroup/ELN Intermediate I&II subgroups (patients lacking both NPM1 and CEBPA mutations and/or harboring FLT3-ITD; OS: p=0.064 and LFS: p=0.072).In the validation set, miR-3151 overexpression confirmed its prognostic impact in patients in the univariate analysis for OS (45%±12% vs. 26±19%, p=0.039) and LFS (51%±14% vs. 30±24%, p=0.034) and in the multivariate analysis for OS (OS: p=0.038; HR 1.88, IC 95%: 1.06-3.34) and LFS (p= 0.014; HR 2.411 (1.198-4.855)Finally, a supervised analysis revealed that samples exhibiting high levels of miR-3151 expression had a distinctive miRNA signature including miR-509, miR-135a, miR-100*, miR-186*, let-7a* and miR-501. ConclusionmiR-3151 is an independent prognostic marker in patients with IR-AML. The study of miR-3151 refines the molecular prognostic stratification of these patients and hence could be of help to guide therapy. AcknowledgmentsMarina Díaz-Beyá is supported by Fundación Española de Hematologia y Hemoterapia. This research is supported by Sociedad Española de Hematologia y Hemoterapia and by grants from Fondo de Investigaciones Sanitarias FIS-PI080158. Disclosures:No relevant conflicts of interest to declare.

Sole Trisomy 8 in AML: Concomitant Molecular Markers, Stability of Genetic Patterns and Impact On Outcome.

Abstract 2503 Background:Trisomy 8 belongs to the most frequent cytogenetic aberrations in AML and is classified to intermediate-risk karyotypes if not within favorable or complex karyotypes (Grimwade et al., Blood 2010). Some study groups showed pts with sole trisomy 8 (+8sole) respond poorly to cytarabine-based chemotherapies. In addition, there is dissent whether trisomy 8 is a primary event or a secondary hit in pathogenesis. Aim:Evaluation of pts with +8sole with respect to related molecular markers, stability of cytogenetic and molecular aberrations and impact on outcome. Patients and Methods:1,181 newly diagnosed adult AML pts with intermediate-risk karyotypes were distributed as follows: +8 (n=117), normal karyotype (NK) (n=801), other intermediate-risk abnormalities (n=263). In detail, 80/117 pts with trisomy 8 (68.4%) showed +8sole. All of these 80 pts were screened for the molecular markers: ASXL1, CEBPA, FLT3- ITD, FLT3- TKD, IDH1, MLL- PTD, NPM1, and RUNX1. WT1 was analyzed in 79 and IDH2 in 78 pts. For comparison we characterized 400 NK pts for all before mentioned molecular markers. Results:Comparing clinical features +8sole pts were older than NK (69.9 vs 63.6 years; p<0.001). No differences were seen for gender, Hb, WBC or PLT counts. Molecular Mutations:+8sole pts harbored more often ASXL1 mut than NK pts (47.5% (38/80) vs 15.0% (60/400); p<0.001) and also more frequently RUNX1 mut (36.3% (29/80) vs. 15.8% (63/400); p<0.001). In contrast, +8sole pts less frequently showed NPM1 mut (16.3% (13/80) vs 49.3% (197/400; p<0.001) or CEBPA mut (5.0% (4/80) vs 14.5% (58/400); p=0.018). All 4 cases with +8sole and CEBPA mut were monoallelic, whereas only 37.7% (23/58) of NK pts with CEBPA mut were monoallelic (p=0.031). Incidences for FLT3- ITD, FLT3- TKD, IDH1, IDH2, MLL- PTD, and WT1 did not differ between +8sole and NK. Genetic Stability:We analyzed relapse samples of 12 pts: 10/12 (83.3%) showed +8sole and their individual molecular marker pattern identical to diagnosis. However, 2/12 did show progress: one with +8sole and NPM1 mut at diagnosis gained FLT3- ITD in relapse. Case 2 habored an ASXL1 mut in addition to +8sole at diagnosis but relapsed with ASXL1 mut and a complex karyotype (including +8). Thus, +8sole as well as the corresponding mutations are stable through course of disease. Furthermore, in 5 pts we analyzed samples from MDS phase preceding AML. Two pts were genetically stable with either ASXL 1mut or MLL- PTD, and +8sole in MDS as well as in s-AML. One patient showed +8sole and ASXL1 mut in MDS and gained RUNX1 mut in s-AML. Interestingly, 2 pts harbored an ASXL1 mut and a NK at diagnosis of MDS, but an ASXL1 mut and +8sole at diagnosis of s-AML. This leads to the hypothesis of ASXL1 mut being first hit, whereas +8sole is gained during course of disease. Comparing overall survival (OS) for all pts with available follow-up data and intensive treatment regimes there was a significantly worse outcome for pts with +8sole (n=38) compared to NK (n=300; 13.0 vs 35.7 mo; p=0.003). Furthermore, we evaluated the total cohort (n=338) for impact of molecular markers. Pts with ASXL1 mut, FLT3-ITD, MLL-PTD, and RUNX1 mut showed worse outcome than pts without the respective marker mutated (p<0.001; <0.001; 0.009; <0.001, respectively). In contrast, pts with NPM1 mut, NPM1 mut/no FLT3-ITD or biallelic CEBPA mut showed better OS (p=0.003; <0.001; 0.044). FLT3- TKD, IDH1, IDH2, WT1, and CEBPA had no impact on OS. Combing markers associated with poor prognosis (ASXL1, FLT3-ITD, MLL- PTD, RUNX1) we detected at least one of these in 75% (60/80) within +8sole but only in 46.3% (185/400) within NK (p<0.001). Therefore we evaluated pts with at least one of these poor prognostic markers (n=147) and found median OS of 12.4 vs 49.6 months for pts with none of these (n=191; p<0.001). For multivariable Cox regression analysis age, WBC count, karyotype, biallelic CEBPA, NPM1 and the combined group of poor prognostic molecular markers were included (all being significant in univariable analysis). Age, WBC count and the combined group of poor prognostic molecular markers were significant (p<0.001, <0.001, 0.026, respectively). +8sole did not show an independent impact on OS. Conclusions:+8sole had poor outcome compared to NK. This is related to the concomitant presence of adverse molecular markers, particularly ASXL1, and RUNX1, or both. Thus, for risk stratification of pts with +8sole ASXL1 and RUNX1 mutation assessment should be considered. Disclosures:Alpermann:MLL Munich Leukemia Laboratory: Employment. Haferlach:MLL Munich Leukemia Laboratory: Equity Ownership. Eder:MLL Munich Leukemia Laboratory: Employment. Kohlmann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Equity Ownership.

GATA2 zinc finger 1 mutations associated with biallelic CEBPA mutations define a unique genetic entity of acute myeloid leukemia

AbstractCytogenetically normal acute myeloid leukemia (CN-AML) with biallelic CEBPA gene mutations (biCEPBA) represents a distinct disease entity with a favorable clinical outcome. So far, it is not known whether other genetic alterations cooperate with biCEBPA mutations during leukemogenesis. To identify additional mutations, we performed whole exome sequencing of 5 biCEBPA patients and detected somatic GATA2 zinc finger 1 (ZF1) mutations in 2 of 5 cases. Both GATA2 and CEBPA are transcription factors crucial for hematopoietic development. Inherited or acquired mutations in both genes have been associated with leukemogenesis. Further mutational screening detected novel GATA2 ZF1 mutations in 13 of 33 biCEBPA-positive CN-AML patients (13/33, 39.4%). No GATA2 mutations were found in 38 CN-AML patients with a monoallelic CEBPA mutation and in 89 CN-AML patients with wild-type CEBPA status. The presence of additional GATA2 mutations (n=10) did not significantly influence the clinical outcome of 26 biCEBPA-positive patients. In reporter gene assays, all tested GATA2 ZF1 mutants showed reduced capacity to enhance CEBPA-mediated activation of transcription, suggesting that the GATA2 ZF1 mutations may collaborate with biCEPBA mutations to deregulate target genes during malignant transformation. We thus provide evidence for a genetically distinct subgroup of CN-AML. The German AML cooperative group trials 1999 and 2008 are registered with the identifiers NCT00266136 and NCT01382147 at www.clinicaltrials.gov.

GATA2 mutations are frequent in intermediate-risk karyotype AML with biallelic CEBPA mutations and are associated with favorable prognosis

GATA2 mutations are frequent in intermediate-risk karyotype AML with biallelic CEBPA mutations and are associated with favorable prognosis

ID1 expression associates with other molecular markers and is not an independent prognostic factor in cytogenetically normal acute myeloid leukaemia

In acute myeloid leukaemia with normal karyotype (CN-AML), gene mutations (e.g. NPM1, FLT3, CEBPA) as well as deregulated gene expression affect outcome. High expression of ID1 was described as a negative prognostic factor. We have shown that CEBPA regulates ID1 expression. Therefore, we analysed the prognostic impact of ID1 expression in 269 patients (aged 16-60 years) with CN-AML in the context of other molecular markers, particularly CEBPA mutations. ID1(high) status was an independent negative prognostic factor for overall survival (OS) in multivariate analysis when analysed together with age, extramedullary disease, platelets, expression of BAALC and WT1, FLT3-internal tandem duplication, NPM1, WT1 single nucleotide polymorphism rs16754 and IDH1. ID1 expression was higher in CEBPA wildtype patients than in patients with monoallelic CEBPA mutations and these patients showed higher ID1 expression compared to patients with biallelic CEBPA mutations. Thus, when CEBPA mutations were considered, ID1 expression lost its prognostic impact. Likewise, the negative impact of ID1(high) status on relapse-free survival (RFS) was lost when CEBPA mutations were included in the analysis. In CEBPA wildtype patients, ID1 expression had no impact on complete remission-rate, OS or RFS. In conclusion, CEBPA mutations seem to deregulate ID1 expression. Therefore, ID1 expression is not an independent prognostic factor in CN-AML.

Abstract SY38-01: Genetic and epigenetic changes contributing to leukemogenesis and prognostic impact in acute myeloid leukemia

Abstract SY38-01: Genetic and epigenetic changes contributing to leukemogenesis and prognostic impact in acute myeloid leukemia

Molecular and cellular effects of oncogene cooperation in a genetically accurate AML mouse model

Biallelic CEBPA mutations and FMS-like tyrosine kinase receptor 3 (FLT3) length mutations are frequently identified in human acute myeloid leukemia (AML) with normal cytogenetics. However, the molecular and cellular mechanisms of oncogene cooperation remain unclear because of a lack of disease models. We have generated an AML mouse model using knockin mouse strains to study cooperation of an internal tandem duplication (ITD) mutation in the Flt3 gene with commonly observed CCAAT/enhancer binding protein alpha (C/EBPα) mutations. This study provides evidence that FLT3 ITD cooperates in leukemogenesis by enhancing the generation of leukemia-initiating granulocyte-monocyte progenitors (GMPs) otherwise prevented by a block in differentiation and skewed lineage priming induced by biallelic C/EBPα mutations. These cellular changes are accompanied by an upregulation of hematopoietic stem cell and STAT5 target genes. By gene expression analysis in premalignant populations, we further show a role of FLT3 ITD in activating genes involved in survival/transformation and chemoresistance. Both multipotent progenitors and GMP cells contain the potential to induce AML similar to corresponding cells in human AML samples showing that this model resembles human disease.

Evaluation of the New Genetic Risk Classification of the European LeukemiaNet Recommendations in 1,110 Patients with De Novo AML and Proposal of a Refined Version

Abstract 413 Background:The recently published recommendations for prognostication in AML (Döhner et al. for ELN, Blood 2010;115,453–474) were based on a review of the literature and included cytogenetics as well as NPM1, CEBPA and FLT3-ITD mutation status for risk stratification. We here aimed to evaluate the prognostic impact of this approach in an independent cohort. Patients: We started with a cohort of 1,428 adults with newly diagnosed AML, which were investigated by cytomorphology, immunophenotyping, cytogenetics, and molecular genetics. We first excluded patients with t(15;17) (n=59), therapy-associated AML (n=111) and secondary AML (n=148). Thus, 1,110 patients with de novo AML and cytogenetics available in all cases were further evaluated according to ELN criteria. The following molecular markers were investigated: NPM1 (1,064/1,110), FLT3-ITD (1,066/1,110), CEBPA (880/1,110), MLL-PTD (1,064/1,110) and RUNX1 (454/1,110). Results: Male/female ratio was 1.2 (598/512), median age was 66.6 years (range 18.3 – 100.4). According to the ELN proposal 297 (26.8%) pts were assigned to the favorable group (CBF leukemias, NPM1mut/without FLT3-ITD in normal karyotype (NK), or CEBPAmut in NK), 363 (32.7%) pts were classified as intermediate I (NPM1mut/FLT3-ITD+, or NPM1wt/FLT3-ITD+, or NPM1wt without FLT3-ITD; all NK), 249 (22.4%) as intermediate II (t(9;11) or cytogenetic abnormalities not classified as favorable or adverse), and 201 (18.1%) as adverse (inv(3)/t(3;3); t(6;9); t(v;11)(v;q23); −5 or del(5q); −7; abn(17p); complex karyotype, i.e. ≥ 3 chromosome abnormalities)). Evaluation according to these criteria revealed significant differences in overall survival between the favorable subgroup and all other subgroups (inter I p<0.001; inter II 0.008, adv <0.001). Also adverse vs all other subgroups (all p<0.001) differed significantly with respect to OS. However, no significant differences were observed between both large cohorts of inter I and inter II (together 55.1% of all pts). We therefore intended to revise the ELN criteria for better discrimination of the intermediate groups. In addition to ELN recommendations we considered a threshold of 0.5 for the FLT3-ITD ratio (mut/wt) which had been suggested more valid for prognostication than the mutation status per se. For the revised classification molecular markers were mandatory for all cases with intermediate risk cytogenetics. Therefore, 100 cases had to be excluded due to missing data. Thus, 1,010 pts were reclassified into our new subgroups defined as: favorable I: CBF leukemias; favorable II:NPM1mut or biallelic CEBPAmut (without any other molecular marker and no fav or adv cytogenetics); intermediate I:FLT3-ITD ratio <0.5 (without RUNX1 or MLL-PTD and no fav or adv cytogenetics); intermediate II:FLT3-ITD ratio ≥0.5 and/or RUNX1mut and/or MLL-PTD+ (and no fav or adv cytogenetics); adverse: as defined by ELN. Patients were distributed as follows: fav I: 68 (6.7%), fav II: 286 (28.3%), inter I: 157 (15.5%), inter II: 298 (29.5%), adv: 201 (19.9%). Fav I and fav II had no significant differences in OS (median n.r. vs 62.2 mo, n.s.) and therefore were grouped together as “favorable”. This finally leads to four different prognostic subgroups: favorable: CBF leukemias; NPM1mut or biallelic CEBPAmut, intermediate I:FLT3-ITD ratio <0.5, intermediate II:FLT3-ITD ratio ≥0.5 and/or RUNX1mut and/or MLL-PTD+, adverse. Patients were distributed as follows: fav: 354 (35.0%), inter I 157 (15.5%), inter II: 298 (29.5%), adv: 201 (19.9%). Median OS differed between all subgroups: fav 62.2, inter I 24.3, inter II 12.4, adv 8.7 mo. (fav vs inter I p=0.058, vs inter II <0.001, vs adv <0.001; inter I vs inter II 0.004, vs adv <0.001; inter II vs adv 0.039). Conclusion: The new ELN proposal for prognostication in de novo AML is based on cytogenetic and molecular genetic data. Based on this proposal we further improved prognostication in a series of 1,010 pts by integrating the following molecular markers besides cytogenetics: NPM1mut, biallelic CEBPAmut and FLT3-ITD ratio <0.5 for the favorable group and FLT3-ITD ratio ≥0.5, other CEBPAmut, MLL-PTD+, or RUNX1mut for the intermediate group, and adverse based on cytogenetics only. This refined version may contribute to a better risk assessment in de novo AML patients allowing to separate 4 subgroups with striking differences in OS. Disclosures:Alpermann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

MicroRNAs in Intermediate Risk Cytogenetic Acute Myeloid Leukemia Add Relevant Prognostic Information to Molecular Categorization

Abstract 235The prognosis of AML patients within the intermediate cytogenetics category is mainly determined by the mutational status of some relevant genes, such as NPM1 mutations (NPMmut), or biallelic CEBPA mutations (CEBPAmut), associated with a favorable outcome, and with the presence of FLT3 internal tandem duplication (FLT3-ITD), which correlates with an adverse prognosis. Nonetheless, additional biological features such as microRNA (miRNA) expression pattern might contribute to refine prognosis and guide therapy in this setting.The aim of the present study is to investigate whether miRNA expression is associated with molecular characteristics and clinical outcome in intermediate-risk AML patients (IR-AML).We have analyzed samples from 85 IR-AML patients (median age, 52 [range, 18–71]; 52% males) who received intensive therapy from 1994 to 2009. Forty-three patients (51%) harbored NPMmut, 37 (44%) harbored FLT3-ITD (including 23 with NPMmut), and 11 (13%) harbored CEBPAmut, including 7 with biallelic mutations. The expression of 670 mature miRNAs was analyzed by multiplex Real Time PCR using TaqMan Human MicroRNA Arrays (Applied Biosystems). All PCR reactions were performed using an ABI 7900 HT sequence detection system. miRNA expression data was analyzed by the 2−DDCt method, using RNU48 as endogenous control. Statistical analysis was performed with BRB Array Tools, SPSS version 15.0.1 and R software version 2.9.0.Supervised analysis by means of t-test based on multiplex permutations (class comparisons analysis, p<0.001) revealed a distinctive miRNA signature in patients with NPMmut, with overexpression of miR-10a, miR-10a*, miR-10b and miR-196b, and downregulation of miR-126, miR126*, miR-424, miR-424* and miR-335, as well as patients with biallelic CEBPAmut, characterized by downregulation of miR-196b and upregulation of miR-181a. Response rate in this series of patients was 84%, with 5-year survival of 43±11% and relapse incidence (RI) of 55±14%. Multivariate analysis for overall survival(OS) including NPM status, FLT3-ITD status, age, WBC, and Log Rank OS significant miRNAs (miR-632, miR-23b, miR-409-3p, let-7a*, miR-565 and miR-196b) identified age, absence of NPMmut, and FLT3-ITD as unfavorable variables together with low expression of miR-409-3p (p<0.001; HR=3.3, 95% CI: 1.7–6.4), and increased level of let-7a* (p=0.026; HR=5.1, 95% CI: 1.21–21.5) and miR-196b (p=0.056; HR=7.27, CI: 0.95–55.6). Concerning risk of relapse (RR), multivariate analysis including NPM status, age, FLT3-ITD, WBC, and Log Rank RR significant miRNAs (miR-632, miR-155*, miR-135a, miR-409-3p, miR-150, miR-23a* and miR-363) the absence of NPMmut, FLT3-ITD and increasing leukocyte count were associated with a higher RI. Remarkably, decreased miR-409-3p expression (p=0.011; HR=3.3, 95% CI: 1.3–8.2) and miR-135a (p=0.02; HR=4.2, 95% CI: 1.2–14.2), together with higher levels of miR-23a* (p<0.001; HR=6.2, 95% CI: 2.61–14.7) were independently associated with a higher relapse risk. Of note, a decreased miR-409-3p level retained its adverse prognosis value in the subgroup of patients without favorable molecular markers (i.e., wild-type NPM1 and CEBPA and/or FLT3-ITD;p=0.001) together with low miR-361-3p (p=0.013, HR= 2.4, CI: 1.2–5.1). On the contrary, let-7a* levels segregated subgroups of patients in the category of favorable genotype (i.e., mutated NPM1 without FLT3-ITD p=0.027).In this series of patients of intermediate-risk cytogenetic AML, measurement of expression levels of several miRNAs such as miR-409-3p, miR-135a, let-7a* or miR-23a* showed independent prognostic value, and contribute to predict the outcome within specific molecular subgroups. Nonetheless, confirmation of the prognostic impact of these miRNAs and investigation of possible underlying mechanisms account for this effect require future studies. Disclosures:No relevant conflicts of interest to declare.

High Frequency of GATA2 Mutations in Cytogenetically Normal Acute Myeloid Leukemia with Biallelic CEBPA Mutations Identified by Exome Sequencing

Abstract 72Cytogenetically normal acute myeloid leukemia (CN-AML) with biallelic CEBPA gene mutations (biCEPBA) represents a distinct genetic entity associated with a favorable clinical outcome (Dufour et al, JCO, 2010; Green et al, JCO, 2010; Pabst et al, Br J Cancer, 2009; Wouters et al, Blood, 2009). Furthermore, biCEBPA mutations are seldomly associated with other known prognostic mutations, like mutated NPM1 or FLT3-ITD. So far, it is not known if other alterations cooperate with the biCEBPA mutations in the process of leukemogenesis.To identify collaborating mutations, we performed whole exome sequencing in five biCEBPA mutated CN-AML patients. We generated at least 5 Gbp of exome sequence for each of the biCEBPA AML samples and for the corresponding remission samples. This allowed us to cover at least 80% of RefSeq coding exon positions with a minimum read depth of 10. Comparison of the AML exome sequence with the remission exome sequence and exclusion of annotated polymorphisms led to the identification of leukemia-specific variants. So far, we were able to confirm between 2 to 10 non-synonymous coding somatic mutations per patient in addition to the previously known biCEBPA mutations using Sanger sequencing. Thus, we detected tumor-specific mutations (nonsense and missense) in a total of 22 genes. Two genes were found recurrently mutated in 2 of the 5 biCEBPA samples: DNMT3A (2/5) and GATA2 (2/5).GATA2 is a zinc finger transcription factor important for haematopoietic stem cell proliferation and normal megakaryocytic development. GATA2 mutations have recently been associated with familial monocytopenia and familial myelodysplastic syndrome (Hsu et al, Blood, 2011; Scott et al, ASH abstract 2010). In the M5 subtype of AML, GATA2 mutations were found at a low frequency of 3.6% (Yan et al, Nature Genetics, 2011). Interestingly, GATA2 is a direct protein interactor and negative regulator of CEBPA. (Huang et al., MCB, 2009; Tong et al, MCB, 2005).Therefore, we determined the frequency of GATA2 mutations in 32 patients with biCEBPA mutant AML by screening all coding exons of GATA2 using high resolution melting curve analysis. Aberrant melting curves were subsequently confirmed by Sanger sequencing. Interestingly, 13 out of 32 (40.6%) biCEBPA patients carried heterozygous missense mutations in GATA2 and strikingly these mutations were all located in the highly conserved N-terminal zinc finger domain of GATA2. The missense mutations A318T and G320D surrounding the C319 which coordiates the zinc atom were recurrently detected in 6 out of 13 biCEBPA patients (3 with A318T and 3 with G320D). Two patients were found to carry each two different mutations in GATA2. 4 out of 13 biCEBPA patients with GATA2 mutations who could be analyzed during molecular remission had lost the GATA2 mutation at remisssion. Furthermore, no GATA2 mutations were found in 38 patients with a monoallelic CEBPA mutation and in 90 CN-AML patients with wildtype CEBPA. We are currently analyzing the functional consequences of these GATA2 mutations.In summary, we describe for the first time the specific association of mutations within the N-terminal zinc finger of GATA2 with biallelic CEBPA mutations in cytogenetically normal AML. Although high throughput sequencing so far has mainly revealed an increasing genetic heterogeneity in AML, our results suggest that there is an association of distinct mutations in defined genetic subgroups of AML. Disclosures:Krebs:Illumina: Honoraria. Greif:Illumina: Honoraria.

ID1 Expression Correlates with CEBPA Mutational Status and Is Not An Independent Risk Factor in Cytogenetically Normal AML,

Abstract 3554 Background:Acute myeloid leukemia with normal karyotype (CN-AML) is a heterogenous disease. During the last years, mutations in several genes (e.g. NPM1, FLT3, CEBPA, WT1, IDH1, IDH2) have been identified which are involved in the pathogenesis of AML and affect the prognosis of these patients. Moreover, deregulated expression of genes such as MN1, BAALC, ERG and WT1 was demonstrated to be predictive of outcome in CN-AML. Recently, high expression of the ID1 gene was described as a negative prognostic factor in AML (Tang et al. Blood 2009, 114:2993–3000). Aims:We have shown that C/EBPα, a transcription factor encoded by the CEBPA gene, binds to a regulatory element in the promoter region of the ID1 gene and regulates ID1 expression in leukemic cells (Wagner et al. Proc Natl Acad Sci USA 2006, 103:6338–6343). Therefore, we wanted to analyze the prognostic impact of ID1 expression in CN-AML in the context of other molecular markers, in particular CEBPA mutations. Methods:ID1 expression was quantified normalized to ABL by real time RT-PCR in 269 patients (age 16–60 years) with CN-AML treated with intensive double induction and consolidation therapy within the AMLSG 295 and 0199 trials (NCT00209833). The patients were also analyzed for mutations in the genes NPM1, FLT3, CEBPA, WT1, IDH1 and IDH2. Median follow up was 79 months. Results:Expression of ID1 varied over a 3-log range. High expression of ID1 (ID1high, defined as > median expression level) was significantly associated with the presence of a FLT3 -ITD or an IDH2 mutation and WT1 wildtype. Moreover, ID1 expression was closely associated with CEBPA mutational status. Altogether, 41 patients (15%) harboured a CEBPA mutation (24 monoallelic and 17 biallelic mutations). ID1 expression in the CEBPA wildtype patients was significantly higher than in patients with monoallelic CEBPA mutations and these patients had a significantly higher ID1 expression compared to patients with biallelic CEBPA mutations (p = 0.001). ID1high patients had a trend to a lower complete remission (CR) rate (74% vs. 84%; p = 0.07), but in multivariate analysis only blast clearance on day 15 after induction 1, age and WT1 SNP rs16754 were independent predictors for the achievement of CR. In univariate analysis, ID1high patients had an inferior overall survival (OS) compared to patients with low expression (median OS 29 vs. 78 months, 5 year OS 39% vs. 53%, p = 0.026). ID1high status was an independent negative prognostic factor in multivariate analysis when analyzed together with NPM1, FLT3 -ITD, WT1, IDH1, IDH2, extramedullary disease and platelet counts (HR 1.51; 95% CI 1.06–2.19). However, when also CEBPA mutational status was entered into the model, ID1 expression lost its prognostic impact and the only independent prognostic factors were age, platelets, CEBPA mutations, NPM1 /FLT3 -ITD risk group and WT1 SNP rs16754. Likewise, ID1high patients had a trend to an inferior relapse-free survival (RFS; HR 1.36, 95% CI 0.96–1.93, p = 0.086) in univariate analysis. However, in multivariate analysis including CEBPA mutational status, ID1 expression had no impact on RFS and the only prognostic factors for RFS were NPM1 and CEBPA mutations and WT1 SNP rs16754. In CEBPA wildtype patients, ID1 expression had no impact on CR-rate, OS or RFS in univariate or multivariate analysis. Conclusions:CEBPA mutations seem to deregulate ID1 expression in CN-AML. Therefore, ID1 expression is not an independent prognostic factor in CN-AML. Disclosures:No relevant conflicts of interest to declare.

Molecular and Cellular Effects of Oncogene Cooperation in a Genetically Adequate AML Model

Abstract 1379Biallelic CEBPA mutations and FLT3 length mutations are frequently identified in human acute myeloid leukemia (AML) with normal cytogenetics. However, the molecular and cellular mechanisms of oncogene cooperation remain unclear due to a lack of disease models.We have generated an AML mouse model using knock-in mouse strains to study cooperation of FLT3 ITD with commonly observed biallelic C/EBPa mutations which in this model are expressed from the original promoter and thereby at the right dose and at the correct level of cellular differentiation.Interestingly, Cebpa mutants with FLT3 ITD progressed to AML with shorter latency than cohorts with Cebpa mutations only. Before onset of leukemia, biallelic Cebpa mutations caused depletion of granulocyte-macrophage progenitors (GMPs), and expansion of megakaryocyte-erythroid progenitors (MkP, pMegE, pCFU-E) and HSCs. However, FLT3 ITD attenuated MegE progenitor frequencies while MPPs and myeloid progenitors (pGMP and GMP) were more frequently observed. These cellular changes are accompanied by an upregulation of hematopoietic stem cell and STAT5 target genes. Both MPP and GMP cells contain the potential to induce AML similar to corresponding cells in human AML samples showing that this AML model resembles human disease. By gene expression analysis in premalignant populations we further show a role of FLT3 ITD in activating genes involved in survival/transformation and chemoresistance.This murine AML model is perfectly suited to test the effects of chemotherapy and tyrosine kinase inhibitors blocking FLT3 ITD function on specific cellular populations to investigate their specific role in relapse of leukemia after complete remission. Disclosures:No relevant conflicts of interest to declare.

Multilineage Dysplasia Has No Prognostic Impact in 108 Adult Patients with CEBPA Mutated AML Supporting the WHO Proposal to Classify These Patients As A Molecularely Defined Unique Entity

Abstract 1443 Background:In the 2008 WHO classification, CEBPA and NPM1 mutated AML are considered as “provisional entities”. However, patients without “recurrent genetic abnormalities” are classified as “AML with MDS-related changes, AML-MRC” based on multilineage dysplasia (MLD), cytogenetics, clinical history, or “AML, not otherwise specified”. MLD+ patients may also carry CEBPA mutations but will still be classified as AML-MRC. Study Design: We investigated the prognostic impact of MLD in 108 CEBPA mutated AML (54 males, 54 females; median age 67.2; range 15.7–87.6 years) who were all characterized by cytogenetics and investigated for molecular markers in subcohorts. MLD was defined by ≥50% dysplastic cells in 2–3 hematopoietic lineages according to WHO. CEBPA mutation analysis was done by Sanger sequencing or deep-sequencing (454 Life Sciences, Branford, CT). 44 pts (40.7%) had mono-, 53 (49.1%) biallelic, and 11 (10.2%) homozygous CEBPA mutations. Frequencies of other markers were: FLT3 -ITD: 10/105 (9.5%); FLT3 -TKD: 2/79 (2.5%); NPM1: 15/105 (14.3%); MLL -PTD: 4/105 (3.8%); IDH1/2: 8/63 (12.7%); RUNX1: 8/90 (8.9%). Results: Dysplasia (≥50% of cells) was detected in granulopoiesis in 46/106 cases (43.4%), in erythropoiesis in 14/108 (13.0%), and in megakaryopoiesis in 34/90 (37.8%). MLD was found in 28/108 (25.9%) of bone marrow samples: two-lineage dysplasia in 26, trilineage dysplasia in 2 cases. According to MRC criteria (Grimwade et al., Blood, 2010), the majority had intermediate karyotypes (KTs) (101/108; 93.5%) mainly due to normal KTs (NK; n=80/101). This confirmed the strong association of CEBPA mutations with NK-AML. Adverse KTs were identified in 7 pts only (6.5%). MLD+ and MLD- CEBPA mutated pts had similar median age (MLD+: 68.7; MLD-: 66.7 yrs; p=n.s.) and male/female ratios (1.5 vs. 0.9; p=n.s.). MLD+ had lower mean WBC counts (13.7 vs. 38.6×10(9)/l; p=0.004), but mean platelet counts, Hb, and BM blasts did not differ significantly from MLD-. The frequency of MLD+ did not differ significantly between monoallelic (14/45; 31.1%), biallelic (11/53; 20.8%), or homozygous CEBPA mutations (3/10; 30.0%). Intermediate and adverse KTs were similarly distributed between MLD+ and MLD- cases (intermediate: 27/28; 96.4% vs. 74/80; 92.5%; adverse: 1/28; 3.6% vs. 6/80; 7.5%; p=n.s.). Also, additional NPM1 mutations showed a similar frequency in MLD+ (4/28; 14.3%) or MLD- (11/77; 14.3%) cases, which holds true as well for FLT3 -ITD (frequency in MLD+: 1/28; 3.6%; in MLD-: 9/77; 11.7%; p=n.s.) or FLT3 -TKD, RUNX1, MLL -PTD, IDH1/2, or NRAS mutations. Outcome did not differ significantly between MLD+ and MLD- (2-year OS rate: MLD+: 56.5%; MLD-: 65.5%; median EFS: MLD+: 13.8 months; MLD-: 16.3 months, p=n.s.). OS and EFS were independent from mono-, biallelic, or homozygous CEBPA mutations in the total cohort, but when only FLT3 -ITD negative pts were considered, biallelic as compared to monoallelic/homozygous CEBPA mutations had better median OS (not reached, n.r. vs. 23.3 months; p=0.060) and median EFS (22.8 vs. 14.4 months; p=n.s.). Adverse KTs had worse OS than intermediate KTs (median 8.4 months vs. n.r.; p=0.021), while EFS was not significantly influenced by MRC group. By univariable Cox analysis, gender (better for female; p=0.024), age (p=0.031), WBC count (p=0.003) (continuous), and MRC risk category (p=0.03) were prognostically relevant for OS, whereas presence of MLD, Hb, platelets, BM blasts (continuous), biallelic vs. monoallelic/homozygous CEBPA mutations, FLT3 -ITD, and NPM1 mutations had no impact on survival. MLD also had no significant impact on EFS in univariable analysis. By multivariable analysis, only male sex (p=0.01), higher WBC count (p<0.001), and unfavorable MRC category (p=0.045) had significantly shorter survival. Multivariable analysis showed an adverse impact on EFS for gender (p=0.014) and high WBC count (p=0.003). Conclusion: Presence of MLD has no impact in pts with CEBPA mutated AML with respect to biological characteristics, cytogenetic risk group, associated molecular markers, and prognosis. Thus, our data does not support to overrule the detection of CEBPA mutations for classification and prognosis by the presence of MLD as suggested by the WHO guideline. We therefore strongly suggest to categorize CEBPA mutated AML as a separate entity and further delineate patients only according to mutation status and additional FLT3 -ITD but not by dysplasia. Disclosures:Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Macijewski:MLL Munich Leukemia Laboratory: Employment. Grossmann:MLL Munich Leukemia Laboratory: Employment. Kohlmann:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Methylation of the Proximal, Distal and Core Promoter of CEBPA in 574 Cases with Normal Karyotype AML and 44 with t(8;21) Disclosed Different Frequencies but No Impact on Prognosis

Abstract 2560 Background:Loss of CEBPA function due to mutations is thoroughly explored in AML. Recently epigenetic modifications such as promoter methylation have gained increasing interest as additional mechanisms for transcriptional regulation of cancer related genes. In this context, the clinical impact of aberrant CEBPA promoter methylation (PM) in AML is controversially discussed. The aim of this study was to clarify the frequency and the significance of aberrant CEBPA PM with regard to clinical features in a large cohort of de novo AML patients. The study comprises the CEBPA core promoter as well as the distal and proximal CEBPA promoter region as it has been shown that these upstream located regions also bear promoter activity. Patients:CEBPA PM was analyzed in a cohort of 574 de novo AML patients with normal karyotype (NK-AML) and without CEBPA mutations. The cohort was composed of 268 females and 306 males. Age ranged from 20.0 to 89.6 years (median: 63.7). In addition, methylation status was assessed in 48 patients with biallelic (n=10) or monoallelic (n=38) CEBPA mutations to exclude coincidence with CEBPA PM. As the fusion transcript RUNX1-RUNX1T1 is known to down-regulate CEBPA mRNA and protein levels in AML, we also analyzed CEBPA distal PM status in 44 RUNX1-RUNX1T1 positive AML patients to evaluate a possible correlation between CEBPA distal PM and RUNX1-RUNX1T1 induced down-regulation of CEBPA expression. All patients were intensively treated using standard AML protocols. Methods: Proximal PM and distal PM were analyzed in the total cohort using Sanger sequencing. Core PM was screened by methylation specific PCR in a subcohort of 326 CEBPA unmutated cases. Methylation data was correlated to clinical outcomes and to the presence of FLT3-ITD (n=175/568), NPM1 (n=256/565), RUNX1 (n=91/275), MLL-PTD (n=76/567), IDH1G105 (n=47/353), IDH1R132 (n=38/371), IDH2R140 (n=65/332) and IDH2R172 (n=12/342) molecular mutations. In addition, CEBPA mRNA expression levels were assessed by quantitative real time PCR (Taqman®, Life Technologies, Carlsbad, CA) in 39 cases with CEBPA distal PM and in 8 cases of the NK-AML cohort tested negative for CEBPA distal PM. CEBPA expression was normalized against the expression of the control gene ABL1. Results: The CEBPA distal promoter was methylated in 54/574 cases (9.4%) whereas CEBPA proximal PM was found in none of the 574 cases. Methylation of the CEBPA core promoter was detected in only 8 of 326 cases (2.5%). As CEBPA proximal and core PM seem to be rare events in AML, they were excluded from further analysis. None of the 48 CEBPA mutated cases revealed any PM and thus aberrant CEBPA PM and mutation status were mutually exclusive. Surprisingly, analysis of CEBPA mRNA expression level revealed no difference between CEBPA distal PM positive and CEBPA distal PM negative cases (mean ± SD 145 ± 97.9 and range 2.7–474.2 vs. 141.4 ± 85.3 and 23.1–259.2, n.s.) suggesting that CEBPA distal PM has no influence on CEBPA mRNA expression in NK-AML. In contrast, we observed a significantly higher frequency of CEBPA distal PM in patients with RUNX1-RUNX1T1 positive AML (n=17/44; 38.6%) compared to the NK-AML cohort (n=55/572; 9.4%) (p<0.001) indicating a correlation between RUNX1-RUNX1T1 induced down-regulation of CEBPA expression and CEBPA distal PM. In the NK-AML cohort, there was no correlation between CEBPA distal PM and age, sex, white blood cell count or Hb levels at diagnosis compared to unmethylated cases. We also were not able to detect a significant correlation between the presence of CEBPA distal PM and the other molecular mutations except for the frequency of IDH2R140 mutations which was significantly lower in CEBPA distal PM positive compared to CEBPA distal PM negative cases (21/267; 7.9% vs. 13/65, 20%, p=0.010). In addition, CEBPA distal PM was not related to overall survival, event free survival or incidence of relapse in NK-AML. Also in subcohorts that were defined by specific molecular mutations (FLT3-ITD, NPM1, RUNX1, MLL-PTD, IDH1, or IDH2) no prognostic impact of CEBPA distal PM could be shown. Conclusion:CEBPA PM in NK-AML was detected in 11.9% of all cases and seems to be mainly restricted to the distal promoter (9.4%). In contrast to RUNX1/RUNX1T1 positive AML no impact of CEBPA PM on CEBPA mRNA expression levels was detected in NK AML. We also conclude that the presence of aberrant CEBPA PM has no clinical relevance in NK-AML and therefore is negligible as prognostic marker. Disclosures:Fasan:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Grossmann:MLL Munich Leukemia Laboratory: Employment. Eder:MLL Munich Leukemia Laboratory: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Specific Molecular Mutations with Prognostic Relevance in AML with Normal Karyotype Are Also Associated with Outcome in AML with Aberrant Karyotype: A Study on 1981 Cases,

Abstract 3517 Background and Aim:The karyotype and molecular mutations are well established prognostic parameters in AML. However, the impact on outcome of molecular mutations has been evaluated mainly in the subset of AML with normal karyotype (NK). The aim of this study was 1. to determine the frequency of NPM1 mutations (NPM1 mut), partial tandem duplications within the MLL gene (MLL -PTD), length mutations within the FLT3 gene (FLT3 -ITD) and CEPBA mutations (CEPBA mut) in distinct cytogenetic subgroups and 2. to evaluate the prognostic impact of these mutations in relation to chromosome abnormalities. Patients and Methods: 1981 patients with AML and evaluable cytogenetics were included. Mutation data was available in the majority of cases: NPM1: 1646, CEPBA: 1324, FLT3 -ITD: 1726 and MLL -PTD: 1656. Based on the karyotype the cohort was subdivided according to revised MRC criteria (Grimwade et al. Blood 2010) and in addition into distinct cytogenetic subgroups. Results: According to cytogenetics 170 cases were assigned to the favorable MRC class (MRCF), 1414 to the intermediate MRC (MRCI) and 397 to the unfavorable MRC subset (MRCU). The frequency of NPM1 mut, CEPBA mut, MLL -PTD and FLT3 -ITD differed significantly between MRC classes and distinct cytogenetic groups.NPM1mut (%)CEPBAmut (%)MLL-PTD (%)FLT3-ITD (%)MRCF0/87 (0)0/19 (0)0/91 (0)26/156 (16.7)MRCI511/1245 (41.0)99/1120 (8.8)106/1247 (8.5)346/1253 (27.6)MRCU10/314 (3.2)3/185 (1.6)17/318 (5.3)21/317 (6.6)t(15;17)0/4 (0)0/14 (0)0/5 (0)21/68 (30.9)t(8;21)0/42 (0)0/3 (0)0/43 (0)2/42 (4.8)inv(16)0/41 (0)0/2 (0)0/43 (0)3/46 (6.5)NK438/857 (51.1)72/789 (9.1)80/858 (9.3)274/864 (31.7)11q23/MLL0/57 (0)0/16 (0)0/58 (0)2/58 (3.4)t(6;9)0/4 (0%)0/5 (0)0/4 (0)4/6 (66.7)3q26/EVI10/18 (0)0/10 (0)0/18 (0)0/18 (0)+49/16 (56.3)0/14 (0)0/14 (0)5/15 (33.3)+816/82 (19.5)4/70 (5.7)6/82 (7.3)16/83 (19.3)+110/22 (0)1/21 (4.8)10/22 (45.5)6/22 (27.3)+131/23 (4.3)3/24 (12.5)3/23 (13.0)5/23 (21.7)+214/17 (23.5)6/18 (33.3)1/17 (5.9)3/17 (17.6)5q-0/22 (0)0/21 (0)5/23 (21.7)3/23 (13.0)7q−/−70/26 (0)2/28 (7.1)0/26 (0)0/26 (0)complex7/189 (3.7)1/81 (1.2)8/191 (4.2)8/190 (4.2)other abn.*46/226 (20.4)13/208 (6.3)10/227 (4.4)41/225 (18.2)*abnormalities other than the before mentioned.In the MRCI subset we evaluated overall survival (OS) between patients with normal or aberrant karyotype (AK) within the respective mutation groups. Within NPM1 mut, CEPBA mut, MLL -PTD+ and FLT3 -ITD+ patients no significant differences in OS were observed between patients with NK or AK. When separating CEPBA mut into biallelic (n=52) and monoallelic cases (n=40) also no difference in OS was observed between pats with NK or AK. Next we tested the prognostic impact of the respective molecular mutations within MRCI including 966 cases with NK and 448 cases with AK. OS was significantly longer in patients with NPM1 mut or CEPBA mut (median OS (mOS) 49.6 months (mo) vs 18.6 mo, p=0.003; not reached (n.r.) vs 21.1 mo, p=0.016) and significantly worse for patients with MLL -PTD or FLT3 -ITD (mOS 10.8 vs 23.0 mo, p=0.039; 13.8 vs 24.9 mo, p=0.003). Analyzing biallelic and monoallelic CEPA mut separately revealed that only biallelic CEPBA mut was associated with a longer OS (p=0.006). Restricting the analysis to MRCI patients with aberrant karyotype revealed a longer OS for patients with NPM1 mut/FLT3 -ITD-, and a shorter OS for patients with FLT3 -ITD (mOS n.r. vs 18.0 mo, p=0.033; 5.7 vs 23.0 mo, p=0.015). A trend towards better OS was observed for biallelic CEBPA mut. Conclusions: 1. The frequency of molecular mutations varies significantly between distinct cytogenetic subsets. They are particularly common in AK within MRCI. NPM1 mut, CEBPA mut and MLL -PTD were not observed in MRCF, in AML with 11q23/MLL -rearrangements, or in 3q26/EVI1 -rearrangements. Their frequency was below 5% in AML with complex karyotype. 2. The outcome of NPM1 mut, CEPBA mut and MLL -PTD+ cases was not different in AML with normal or aberrant karyotype within MRCI. 3. This data suggest to extend mutation screening for NPM1 mut, CEPBA mut, MLL -PTD and FLT3 -ITD to all AML with intermediate risk cytogenetics, as they are significantly associated with outcome not only in AML with normal karyotype but also in AML with cytogenetic abnormalities assigned to MRCI. This consequently will lead to better characterization of a reasonable number of cases from MRCI with important implication on treatment. Disclosures:Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Alpermann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Gene mutation patterns and their prognostic impact in a cohort of 1185 patients with acute myeloid leukemia

AbstractTo evaluate the prognostic value of genetic mutations for acute myeloid leukemia (AML) patients, we examined the gene status for both fusion products such as AML1 (CBFα)–ETO, CBFβ-MYH11, PML-RARα, and MLL rearrangement as a result of chromosomal translocations and mutations in genes including FLT3, C-KIT, N-RAS, NPM1, CEBPA, WT1, ASXL1, DNMT3A, MLL, IDH1, IDH2, and TET2 in 1185 AML patients. Clinical analysis was mainly carried out among 605 cases without recognizable karyotype abnormalities except for 11q23. Of these 605 patients, 452 (74.7%) were found to have at least 1 mutation, and the relationship of gene mutations with clinical outcome was investigated. We revealed a correlation pattern among NPM1, DNMT3A, FLT3, IDH1, IDH2, CEBPA, and TET2 mutations. Multivariate analysis identified DNMT3A and MLL mutations as independent factors predicting inferior overall survival (OS) and event-free survival (EFS), whereas biallelic CEBPA mutations or NPM1 mutations without DNMT3A mutations conferred a better OS and EFS in both the whole group and among younger patients < 60 years of age. The use of molecular markers allowed us to subdivide the series of 605 patients into distinct prognostic groups with potential clinical relevance.

The Role of Different Genetic Subtypes In CEBPA Mutated AML

AbstractAbstract 752To evaluate the role of CEBPA mutations (CEBPAmut) in the context of other molecular mutations and cytogenetic aberrations we have analyzed 1567 AML cases for CEBPAmut. The patients were selected according to cytogenetics excluding the following karyotypes: t(15;17)/PML-RARA, t(8;21)/AML1-ETO, inv(16)/t(16;16)/CBFB-MYH11, inv(3)/t(3;3)/EVI1, t(6;9)/DEK-CAN and 11(q23)/MLL, and complex aberrations. The cohort was composed of 697 females and 870 males. Age ranges from 16.7 to 88.3 years (y) (median: 71.0 y). CEBPAmut were detected in 126/1567 cases (8.0%). The biologic characteristics of the CEBPAmut patients (age range 16.7 to 87.6 y, median: 64.8 y) were further investigated. Three different CEBPAmut patterns were observed: 1) in 50/126 cases (39.7%) one mutation and one wildtype allele were detected (monoallelic pattern), 2) 61 cases (48.4%) had two different mutations (biallelic pattern), 3) 15 cases (11.9%) had one mutation without detectable wildtype allele due to loss of heterozygosity (LOH). Overall we found 186 different mutations of following types: 1) 108 led to a premature N-terminal stop of the protein (6 due to a nonsense and 102 due to a frameshift mutation), 2) 60 were inframe mutations in the b-ZIP region, 3) 8 were frameshifts in the b-ZIP region and, 4) 2 were frameshifts 3'of the b-ZIP region, and 8 were C-terminal point mutations. Correlation to cytogenetics shows a normal karyotype (NK) in 86 (68.3%) of the 126 CEBPAmut patients whereas in 40 pts (31.7%) at least one cytogenetic aberration was detected (-7: n=7; +8: n=7, 9q-: n=2; 11q-: n=3, other trisomies: n=11; other non recurrent translocations: n=4, all others: n=6). Cytogenetic aberrations were more frequent in the monoallelic group (55%) compared to the biallelic (35%) (p=0.001) and to cases with LOH (10%) (p=0.047). Interestingly in the total cohort of 13 pts with monosomy 7 seven pts (53.8%) were CEBPA mutated (53.8%) and of these 6 were in the monoallelic group. Additional mutations were detected in 48 cases (RUNX1: n=11, NPM1: n=10, FLT3-ITD: n=20, FLT3-TKD: n=3, MLL-PTD: n=5, NRAS: n=9, IDH1: n=2, IDH2: n=7; 15 pts showed 2 and 2 pts 3 of these mutations). Similar to the cytogenetic aberrations the molecular mutations were more frequent in the monoallelic group (61.9%) compared to the biallelic (31.0%) and the LOH group (7.1%) (p=0.001). NPM1 mutations were mutually exclusive of biallelic CEBPAmut. As previously described we also detected a significantly higher expression of CD7 in the CEBPAmut compared to the CEBPAwt group (71.2% vs. 18.9%, p<0.001). Furthermore, CD7 was higher expressed in biallelic cases as compared to the monoallelic ones (86.2% vs. 43.8%, p=0.011). It was similar in the LOH group (71.4%) compared to the biallelic group. There was no influence of cytogenetic aberrations or any additional mutation on EFS and OS. Solely the presence of high FLT3-ITD load (>0.5 FLT3-ITD/FLT3wt) was correlated with a shorter EFS (EFS at 2 y: 20.3% vs. 44.8%; p=0.020) and OS (OS at 2 y 54% vs 68%, p=0.047) when compared to the combined group of FLT3wt and those with an FLT3-ITD load of <0.5. Regarding the different CEBPA groups the biallelic cases had a slightly better OS compared to monoallelic cases (OS at 2 y: 75.0% vs. 60.8%; n.s.). The 2-year OS in the LOH group was significantly lower (33.8%; p=0.023, compared to the biallelic group; and p=0.043 compared to 69.7% in the combined biallelic + monoallelic group). In addition, the different functional mutation types were analyzed. Out of frame mutations in b-ZIP had no specific impact on survival within the CEBPAmut cohort. N-terminal stop mutations and in frame mutations in b-ZIP were associated with favourable outcome (OS at 2 y: 72.1% vs. 43.9% all others mutations, p=0.007 and 2 y OS at 2 y: 76.3% vs. 46.9%; p=0.043, respectively), whereas all other mutations were extremely unfavourable (OS at 2 y: 0% vs. 70.5% compared to N-terminal and b-ZIP mutations, p<0.001). In summary, 1) the biology and prognostic impact varies depending on distinct CEBPAmut patterns. 2) Cytogenetic and molecular alterations had no prognostic impact with the exception of FLT3-ITD with a mutation load of >0.5 FLT3-ITD/FLT3wt. 3) Our data for the first time demonstrate that CEPBAmut with LOH are associated with an even inferior outcome than monoallelic mutations. In conclusion, these data show that CEBPA should be analyzed in detail in all NPM1wt NK AML and in those with unfavourable but non-complex karyotypes. Disclosures:Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Alpermann:MLL Munich Leukemia Laboratory: Employment. Eder:MLL Munich Leukemia Laboratory: Employment. Schindela:MLL Munich Leukemia Laboratory: Employment. Grossmann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.