Search

Glioblastoma multiforme (GBM) is the most lethal primary brain tumor in adults. The epigenetically active ribonucleoside analog 5-azacitidine is a new therapy option that changes tumor cell chromatin, which is frequently modified by methylation and deacetylation in malignant gliomas. In vitro, we analyzed cell viability, cell apoptosis, and migration of human GBM cells. In vivo, we established subcutaneous and intracerebral GBM mouse models originating from U87MG, U373MG, and primary GBM cells as well as one patient-derived xenograft. Xenografts were treated with 5-azacitidine as well as valproic acid, bevacizumab, temozolomide, and phosphate buffered saline. The tumor sizes and Ki67 proliferation indices were determined. Glioma angiogenesis was examined immunohistochemically by expression analysis of endothelial cells (CD31) and pericytes (PDGFRβ). In vitro, 5-azacitidine treatment significantly reduced human glioblastoma cell viability, increased cellular apoptosis, and reduced cellular migration. In vivo, 5-azacitidine significantly reduced growth in two intracerebral GBM models. Notably, this was also shown for a xenograft established from a patient surgery sample; whereas, epigenetically acting valproic acid did not show any growth reduction. Highly vascularized tumors responded to treatment, whereas low-vascularized xenografts showed no response. Furthermore, intracerebral glioblastomas treated with 5-azacitidine showed a clearly visible reduction of tumor angiogenesis and lower numbers of endothelial cells and tumor vessel pericytes. Our data show significant growth inhibition as well as antiangiogenic effects in intracerebral as well as patient-derived GBM xenografts. This encourages to investigate in detail the multifactorial effects of 5-azacitidine on glioblastomas.

Low-grade gliomas (LGG) are brain tumors with a low or intermediate biological aggressiveness. According to histopathological features, they are further specified as grade I or II by WHO criteria. Diffuse astrocytomas, oligodendrogliomas, and mixed gliomas are the most common LGG. They mainly affect young patients in their 3rd to 5th decade and often manifest with epileptic seizures. A macroscopically complete or near-complete tumor resection that does not induce additional neurological deficits, is recommended as first line therapy in surgically accessible tumors, as a significant benefit for overall survival has been demonstrated. The indication for adjuvant chemo- or radiotherapy must be discussed interdisciplinary in each case. MGMT promotor methylation, LOH 1p/19q, as well as the status of somatic mutations within IDH1/2 gene constitute biomarkers that may predict response to adjuvant therapy and may correlate to overall survival. These and other biomarkers could be of benefit in future managing plans to offer patients with LGG an individually tailored, optimal treatment.

Single-nucleotide polymorphism arrays allow for genome-wide profiling of copy-number alterations and copy-neutral runs of homozygosity at high resolution. To identify novel genetic lesions in myeloproliferative neoplasms, a large series of 151 clinically well characterized patients was analyzed in our study. Copy-number alterations were rare in essential thrombocythemia and polycythemia vera. In contrast, approximately one third of myelofibrosis patients exhibited small genomic losses (less than 5 Mb). In 2 secondary myelofibrosis cases the tumor suppressor gene NF1 in 17q11.2 was affected. Sequencing analyses revealed a mutation in the remaining NF1 allele of one patient. In terms of copy-neutral aberrations, no chromosomes other than 9p were recurrently affected. In conclusion, novel genomic aberrations were identified in our study, in particular in patients with myelofibrosis. Further analyses on single-gene level are necessary to uncover the mechanisms that are involved in the pathogenesis of myeloproliferative neoplasms.

Since the discovery of the gain-of-function mutation JAK2 V617F in 2005, an increasing number of mutations have been described in myeloproliferative neoplasms (MPN) and related myeloid malignancies.1, 2, 3, 4 For instance, use of array-based techniques such as comparative genomic hybridization and single-nucleotide polymorphism (SNP) analysis led to the identification of genes involved in epigenetic regulation such as tet-oncogene family member 2 (TET2), additional sex combs like 1 (ASXL1) and enhancer of zeste 2 (EZH2).5, 6, 7 Inactivating mutations in TET2, ASXL1 and EZH2 are considered to promote myeloid tumorigenesis because of epigenetic modulation of target genes. More recently, a whole-genome sequencing study in acute myeloid leukemia (AML) uncovered recurrent mutations in 22% of AML patients in another epigenetic regulator, the DNA methyltransferase 3A gene DNMT3A.8 In this study, DNMT3A mutations were associated with poor outcome and, thus, are of clinical relevance. However, exact mechanisms of action of DNMT3A mutations are still unclear because global methylation patterns and 5-methylcytosine content in AML genomes were not significantly altered. Nevertheless, mutations of TET2, ASXL1, EZH2 and DNMT3A occur in a significant number of patients with myeloid malignancies underlining the pathogenic relevance of epigenetic changes.

Glioblastomas are neuroepithelial tumors with lost cellular differentiation and tenfold increased growth rates compared to low-grade gliomas. Despite of very aggressive treatment options based on surgery, irradiation, and chemotherapy, the prognosis of affected patients has remained poor and showed only slight improvements during the last 30 years. Research on glioblastoma border zone was hindered by the tumor's intense invasion into the brain parenchyma and the lack of suitable tumor cell markers. Nevertheless, the compact tumor mass and tumor invasion zone are composed of distinct cell types that need to be distinguished from each other to be addressed selectively. As the isoform 140 of the neural cell adhesion molecule (NCAM-140) was recently demonstrated to be lost in human gliomas with rising WHO grade, human multiform glioblastomas were characterized as a NCAM-140 negative entity displaying three main distinct invasion patterns. Evaluation of putative therapy targets within the tumor tissue and tumor invasion zone has been made possible through NCAM-140 negativity. In the present study, brain tissue controls and human glioblastoma samples with compact tumor mass and invasion areas were analyzed for their vascularization at the tumor border and the expression of thrombin receptor protease-activated receptor type 1 (PAR-1) within tumor tissue and vascular vessel walls. Use of NCAM-140 enabled the identification of the tumor invasion zone and its experimental investigation. Tissue vascularization was found to be significantly increased in the compact tumor mass of glioblastomas compared to their invasion zone and tumor-free controls with a significantly high and specific overexpression of PAR-1 within tumor cells and within tumor blood vessels depending upon the tumor area. This suggests thereby a functional role of the thrombin receptor PAR-1 in glioma cell malignancy and glioblastoma neoangiogenesis.

The expression of functional GABA(A)-receptors in glioma cells correlates with low malignancy of tumours and cell lines from glioma lack these receptors. Here we show that contact with neurons induces the expression of functional GABA(A)-receptors. C6 and F98 glioma cell lines were labelled by recombinant expression of enhanced green fluorescent protein injected into rat brain and studied in acute slices after two to three weeks of tumour growth. The cells responded to GABA or the specific agonist, muscimol with a current typical for GABA(A)-receptors, as studied with the patch-clamp technique. To get insight into the mechanism of GABA(A) receptor induction, the C6 or F98 cells were co-cultured with neurons, astrocytes, oligodendrocytes and microglia. Glioma cells expressed functional GABA(A) receptors within 24 h only in cultures where physical contact to neurons occurred. Activation of GABA(A)-receptors in the co-cultures attenuated glioma cell metabolism while blockade of the receptors increased metabolism. We conclude that with this form of interaction, neurons can influence tumour behaviour in the brain.

The most common malignant brain tumor in adults is glioblastoma multiforme (GBM) showing a very heterogeneous, diffuse infiltrative and aggressive growth. The mean survival is between 8 and 18 months. The standard therapy for glioma consists of neurosurgery, radiation, and chemotherapy with temozolomide (SOC), but frequently resistance to temozolomide or irradiation will be developed. Translational research is focusing on the molecular mechanisms of glioma formation and development of resistance to identify new therapeutic targets. We have transplanted more than 50 glioma tissue samples to immunodeficient mice and were able to establish 13 new PDX models (engraftment rate 25%). Identical morphology of the s.c. growing PDX (pleomorphic cells with marked anisocytosis/ anisonucleosis; solid, picket similar growth, atypical mitoses) compared with the original glioma was proven by histological and immunohistochemical staining. Heterogeneity of the tumor cells in their size, shape and nuclear irregularities were observed. Glioma PDX models with stable growth were screened for sensitivity towards several targeted and cytotoxic drugs (everolimus, sorafenib, bevacicumab, irinotecan, salinomycin, temozolomide). Strong initial response to the treatment with bevacicumab, irinotecan, or temozolomide was observed. From our panel, (10/13) PDX were sensitive to temozolomide-treatment, (9/13) to irinotecan, and (7/13) to bevacicumab. The glioma PDX were resistant to most of the other targeted drugs, however in some single models sensitivity was observed. The frequency of common “onco-mutations” was analyzed using the Illumina TrueSeq Cancer panel sequencing. Although some frequent mutations were detected, i.e. in KDR, FGFR3, PIK3CA, PTEN, P53 and NOTCH1, no correlation with drug sensitivity have be identified. Extended correlation between drug sensitivity, gene expression profiles, and further mutations are still under analysis. The available data demonstrate, that our glioma PDX model panel has retained the original tumor biology and reflect the heterogeneity of the disease, ensuring a high similarity to the clinical situation. Our approach can not only be used for testing of established and new drugs, but also offers an individualized treatment of patients. For large scale screening the s.c. transplantation of glioma PDX is still the method of choice. However, for some drugs it may become important to implement a functional blood-brain barrier (BBB) in the model system. We have established a technology for orthotopic transplantation of glioma cells using the cell lines U373 and U87. Cells are transplanted into the left hemisphere of immunodeficient mice (NMRI nu/nu). Both cell lines showed good take rates and thus can be used for evaluation of systemic therapy approaches where an intact BBB is required. This technology has been further developed for orthotopic transplantation of glioma PDX and as surrogate models for brain metastases from breast and lung cancer (Ref: Hoffmann et al., Neurooncology, 2009). Summary: Our new clinically relevant glioma models enable a better understanding of tumor biology, allow correlations between biomarkers, growth and drug sensitivity and have an essential role to improve individual glioma therapy. Citation Format: Andrea Orthmann, Annika Hoffmann, Reiner Zeisig, Amin EL-Heliebi, Christina Ernst, Johannes Haybäck, Andreas Jödicke, Susanne Kuhn, Michael Linnebacher, Jens Hoffmann, Iduna Fichtner. Therapeutic response to bevacicumab, irinotecan, and temozolomide of PDX from human glioma cannot be correlated to common mutations as routinely identified by panel sequencing. [abstract]. In: Proceedings of the AACR Special Conference: Advances in Brain Cancer Research; May 27-30, 2015; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2015;75(23 Suppl):Abstract nr B24.

TPS2079 Background: Venous thromboembolism (VTE) is a preventable complication in cancer and represents the second most frequent death cause, despite widely available anticoagulants. Further, VTE w...

Invasion of tumor cells into adjacent brain areas is one of the major problems in treatment of glioma patients. To identify genes that might contribute to invasion, fluorescent F98 glioma cells were allowed to invade an organotypic brain slice. Gene expression analysis revealed 5 up-regulated and 14 down-regulated genes in invasive glioma cells as compared to non-invasive glioma cells. Two gene products, ferritin and cyclin B1, were verified in human gliomas by immunohistochemistry. Ferritin exhibited high mRNA levels in migratory F98 cells and also showed higher protein expression in the infiltrating edge of human gliomas. Cyclin B1 with high mRNA expression levels in stationary F98 cells showed marked protein expression in the central portions of gliomas. These findings are compatible with the concept of tumor cells either proliferating or migrating. Our study is the first to apply brain slice cultures for the identification of differentially regulated genes in glioma invasion.

gamma-Aminobutyric acid (GABA) can act as a neuroprotective agent besides its well-established role as the main inhibitory neurotransmitter in the CNS. Here we report that microglial cells express GABA(B) receptors indicating that these prominent immunocompetent cells in the brain are a target for GABA. Agonists of GABA(B) receptors triggered the induction of K(+) conductance in microglial cells from acute brain slices and in culture. Both subunits of GABA(B) receptors were identified in cultured microglia by Western blot analysis and immunocytochemistry, and were detected on a subpopulation of microglia in situ by immunohistochemistry. In response to facial nerve axotomy, we observed an increase in GABA(B) receptor expressing microglial cells in the facial nucleus. We activated microglial cells in culture with lipopolysaccharide (LPS) to induce the release of interleukin-6 and interleukin-12p40. This release activity was attenuated by simultaneous activation of the GABA(B) receptors indicating that GABA can modulate the microglial immune response.

e13016 Background: Ets-1 regulates migration & angiogenesis and is active in malignomas. Methods: We investigated Ets-1 in 60 glioblastomas WHO IV° & 40 astrocytomas WHO II° and evaluated its progn...

Background In recent phase-I/II studies, pomalidomide (POM) has shown activity in myelofibrosis (MF) patients (pts) with anemia and/or thrombocytopenia. In the first cohort of the German multicenter MPNSG 01-09 phase-II trial, 38 MF pts were treated with POM 2 mg QD +/- prednisolone (PRED); 34% responded according to IWG-MRT criteria or became RBC-transfusion independent (Schlenk et al., ASH 2012). Additional 58 MF pts were treated in the second study cohort with POM 0.5 mg QD +/- PRED. Main inclusion criteria were age >50y, RBC-transfusion-dependency or Hb Aims To explore the presence of copy number alterations (CNAs), uniparental disomies (UPDs) and MPN-related gene mutations in the 01-09 study cohort and to correlate genetic parameters with treatment response. Methods Affymetrix 6.0 SNP arrays were used for genome-wide screening of CNAs and UPDs in 91% (87/96) of pts. Frequently mutated genes were analyzed in all pts using PCR-based techniques and sequence analysis of coding exons: JAK2 V617F, MPL W515L, ASXL1 (exon 12), EZH2 (exon 2-20), IDH1/2 (exons 4), SRSF2 (exon 1-2), and TP53(exon 4-10). Results In total, 78 CNAs were identified by SNP array analysis. The overall frequency of CNAs was 52% (45/87 pts); 24% (21 pts) showed ≥2 CNAs, while 7% (6 pts) had a complex genotype defined by ≥3 genomic aberrations. Recurrent large (>5 Mb) gains were trisomy 8 (7%; n=6), gain 1q, and trisomy 9 (5% each; n=4), whereas recurrent large losses were identified in 20q (11%; n=10), 5q (8%; n=7), 13q (7%; n=6), 7q/7 (6%; n=5), and 12p/12 (2%; n=2). Moreover, 20% (17/87) of pts showed 19 informative micro-deletions ( Large UPDs (>10 Mb) affecting a terminal end of the chromosome were present in 32% (28/87) of pts. The most frequent UPDs occurred in 9p including JAK2 (17%; n=15). Other recurrent UPDs mapped to 1p (5%; n=4), 4q, and 7q (2% each; n=2). UPDs in 9p were associated with JAK2 V617F in all pts, whereas 75% (3/4) of pts with UPD in 1p were MPLW515L mutated. Mutations in JAK2 (V617F) and MPL (W515L) were present in 55% (53/96) and 6% (6/96) of pts, respectively. The mutation frequencies for the remaining genes were 30% (ASXL1; n=29/96), 9% (SRSF2; n=9/96), 5% (EZH2; n=5/96), 2% (TP53; n=2/96), and 1% (IDH2; n=1/96). Taken together, at least one mutation was found in 79% (76/96) of pts, whereas 21% (20/96) of pts lacked any of these mutations; 28% (27/96) of pts showed ≥2 concurrent mutations. To evaluate genetic differences and predictive factors, pts were separated into a responder (R, n=23) and a non-responder (Non-R, n=73) group. Regarding DIPSS, Non-R pts were not associated with higher risk compared to R pts: 27% (20/73) high risk vs. 17% (4/23) (p=0.334), 61% (44/73) intermediate-II risk vs. 79% (18/23), and 12% (9/73) intermediate-I vs. 4% (1/23) (p=0.275). However, the overall frequency of CNAs was in trend higher in the Non-R group: 56% vs. 39% (p=0.159). Since the total number of genomic losses was similarly distributed (27% vs. 29%), the difference was mainly due to the higher frequency of large genomic gains/trisomies: 4% (R) vs. 22% (Non-R) (p=0.083). UPDs were similar frequent in R and Non-R pts (26% vs. 34%) (p=0.466). The total number of gene mutations was not different between both groups (n=25 in 23 R pts vs. n=77 in 73 Non-R pts) (p=0.475). However, genetic alterations being associated with unfavourable outcome were enriched in the Non-R group: ASXL1 mutation (39% vs. 22%), EZH2 mutation (5% vs. 0%), loss in 5q (30% vs. 4%), trisomy 8 (9% vs. 0%), 12p / ETV6 deletion (6% vs. 0%), 17p deletion / TP53 mutation (3% vs. 0%), and complex genotype (7% vs. 0%). Taken together, adverse genetic alterations were significantly more frequent in Non-R pts compared to R pts: 48% (35/73) vs. 22% (5/23) (p=0.026). Of note, adverse genetic alterations indicated high risk according to DIPSS in no more than 50% (12/24) of the pts. Conclusion Our study on a well-defined patient cohort with advanced MF revealed a high frequency of genetic alterations reflecting the molecular complexity of the disease. Pts with adverse genetic alterations identified by SNP-array and mutation analyses were not sufficiently represented by DIPSS and showed inferior response to POM +/- PRED. Disclosures: Reiter: Sanofi: Honoraria. Gattermann:Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. Platzbecker:Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. Hochhaus:Novartis: Consultancy, Honoraria, Research Funding, Travel Other; BMS: Consultancy, Honoraria, Research Funding; Pfizer: Consultancy, Honoraria; Ariad: Consultancy, Honoraria. Schlenk:Amgen: Research Funding; Pfizer: Research Funding; Novartis: Research Funding; Chugai: Research Funding; Ambit: Honoraria.

2038 Background: Glioblastoma patients suffer from thromboembolism, aggravating the disease. Recently, increased activity was shown for coagulation factors II, VIII, IX, X, XI, and XII in patient peripheral blood. In vitro, the high potential of FXa-PAR1-loop as growth factor system and significant effects of its blockade were proven. Methods: Patient samples (n=108) were analyzed for FXa and PAR-1. In immunodeficient mice, glioblastoma xenografts were established and treated with low-molecular weight heparins (LMWH) (30mg/kg SC daily for 3 wks). Phosphate buffered saline served as control. Tumor growth rates, final tumor size, tumor proliferation (Ki67), and tumor vascularization (CD31) were determined. Results: Human glioblastomas overexpressed FXa and PAR-1 in the compact tumor center (FXa: p

2077 Background: Targeted therapies are increasingly important in oncology. Axitinib is a novel orally available tyrosine kinase inhibitor which is mainly directed against the VEGFR family. VEGF-antibody bevacizumab is the acting drug of running clinical trials in glioblastoma. Methods: In immunodeficient mice, cell line- and patient-derived glioblastoma xenografts were established and treated with axitinib. Temozolomide, bevacizumab, and phosphate buffered saline served as controls. Tumor size, vascularization (CD31), vascular pericyte coverage (NG-2), apoptosis in tumor and endothelial cells as well as expression of EGFR, VEGFR1, VEGFR2, PDGFRβ, VEGF, and PlGF were determined. Experiments were approved by local ethics committee. Results: Positive control temozolomide always caused growth inhibition whereas primary resistances to bevacizumab as primary drug were observed. As most delighting result, axitinib caused a delay of tumor growth in a glioblastoma xenograft with primary resistance to bevacizumab. Tumor cell proliferation (Ki67) was significantly lower than in controls (p

Abstract 4111 Background: Although an increasing number of gene mutations provide new insights into the pathogenesis of myeloproliferative neoplasms (MPN), the biological origins of myelofibrosis (MF) remain poorly understood. Most likely, to date unknown genetic lesions contribute to disease phenotype and influence clinical outcome. Recently, mutations of EZH2 (enhancer of zeste 2) in 7q36.1 encoding the catalytic subunit of the histone methyltransferase PRC2 (polycomb repressive complex 2) have been identified in myeloid malignancies including MF. Functional studies indicate that EZH2 plays an important role in stem cell renewal by epigenetic repression of target genes. Mutations of EZH2 are considered to result in loss of tumor suppressor activity. These data indicate that EZH2 mutations may be important in the pathogenesis of myeloid malignancies such as MF. Aim: To explore the frequency of EZH2 mutations in a cohort of genetically and clinically well-characterized MPN patients with MF and acute myeloid leukemia (AML) secondary to MPN with 7q alteration. Methods: All coding exons of EZH2 (2-20) were analyzed for mutations in 89 MPN patients using DNA sequencing: primary MF (PMF), n=62; secondary MF (SMF), n=21; AML secondary to MPN (sAML) with 7q alteration, n=6. Data on the mutation status of IDH1/IDH2 (Exon 4), JAK2 (V617F), and MPL (W515L) were available in all cases. In addition, all cases were analyzed on Affymetrix 250K (n=66) or 6.0 (n=23) SNP-arrays allowing for genome-wide screening of copy-number alterations (CNA) and uniparental disomies (UPD) at high resolution. Results: In total, 10 EZH2 mutations were identified in 8 patients (PMF, n=6; SMF, n=1; sAML, n=1) resulting in an overall mutation frequency of 9% (8/89); two PMF cases showed two EZH2 mutations each. The mutation frequency was not significantly different between PMF (6/62, 10%) and SMF (1/21, 5%) patients (p=0.67). EZH2 alterations identified in our study consisted of 8 missense mutations in exons 3, 9, 14–17, and 19, one insertion in exon 17 (c.1957_1958insT), and one deletion in exon 10 (c.1160_1161delAA). Eight of the 10 mutations were heterozygous; in one PMF and one post-PMF AML patient homozygous missense mutations were detected in exon 15 (c.G1720C) and 17 (c.A1947C), respectively; only the mutation in the latter patient was associated with UPD in 7q. None of the remaining 9 patients (PMF, n=4; sAML, n=5) with UPD (n=3) or deletion (n=6) in 7q had EZH2 mutations. EHZ2 mutations occurred in JAK2 mutated and unmutated patients (n=4, each) and none of the cases harbored additional mutations in IDH1/IDH2 or MPL. Furthermore, there was no correlation of the EZH2 mutation status with distinct CNA and/or UPD identified by SNP-arrays: in 5/8 patients, CNA such as trisomy 8, gains in 3q (29 Mb) and 6p (28 Mb), as well as loss in 13q (63 Mb), 18p (13.5 Mb), and 20q (22 Mb) were found. Four patients had large UPD (28-101 Mb) in 7q (n=1), 9p (n=2), or 11q (n=1). In total, only 2 of 8 cases lacked genomic aberrations; with regard to clinical characteristics, EZH2 mutations were associated with marked splenomegaly (20-28 cm) and leukocytosis (median white blood cell counts: 13.8 × 109/L vs. 7.3 × 109/L; p Conclusion: Our study on 89 MF/sAML patients revealed EZH2 mutations as a frequent genetic lesion in these disease entities. However, EZH2 mutations were not associated with specific genetic or genomic aberrations. Further analyses comprising larger patient cohorts are necessary to explore the clinical role of EZH2 in MF. Disclosures: No relevant conflicts of interest to declare.

Abstract 2608 Poster Board II-584 Background: Myeloproliferative neoplasms (MPN) represent a heterogeneous group of acquired hematopoietic stem cell disorders. Clonality leads to exceeding production of myeloid cells resulting in an inherent tendency for thrombotic and hemorrhagic complications as well as transformation into acute myeloid leukemia (sAML). While vascular complications predominantly account for the morbidity in essential thrombocythemia (ET) and polycythemia vera (PV), the mortality of MPN is significantly related to leukemic transformation. Secondary AML occurs more frequently in primary and secondary myelofibrosis (PMF and SMF) than in ET and PV, and the risk for leukemic transformation increases with the duration of the disease. The molecular basis underlying the progression of MPN is poorly understood. Clonal evolution due to genomic instability is considered to play an important role. Aim: To identify genomic lesions associated with leukemic transformation, we applied 250K single-nucleotide polymorphism (SNP) arrays that allow for genome-wide screening of both copy-number alterations (CNAs) and copy-neutral runs of homozygosity (ROH) at high resolution. Method: An unpaired SNP-array analysis of 23 sAML samples was performed [former diagnosis: ET, n=5; PV, n=7; PMF, n=9; SMF, n=2;]. An own set of 30 reference samples was used for normalization. CNAs and ROH were analyzed by CNAG 3.0 software. Aberrations were compared with the 250K SNP-array dataset of 151 MPN patients [ET, n=45; PV, n=45; PMF, n=47, SMF, n=14]. In one sAML patient corresponding SNP-array data from the time of ET diagnosis were available. Results: CNAs were present in 15 of 23 (65%) sAML patients. Thirty-five percent of cases (n=8) exhibited complex genomic aberrations with up to 20 CNAs in one patient (range 5–20). The most frequent larger (>5 Mb) CNAs were trisomy 8 (n=7), gain of 1q, loss of 5q, and deletion in 6p25-pter (16.2–26.7 Mb) and 20q11-q13 (13.6–16.9 Mb) in three cases each, followed by gain in 3q24-qter (51.7 and 54.1 Mb) in two patients. Of note, one case with deletion in 17p12-pter (64.4 Mb) encompassing TP53 and a second with gain in 21q22.12-qter (11.8 Mb) were identified; in the latter one the proximal breakpoint of the gain was located at RUNX1. Smaller CNAs ( Conclusion: In summary, our data on a large series of well-defined sAML cases that evolved from MPN demonstrate that 250K SNP-array profiling is an excellent tool to identify genomic aberrations. In contrast to MPN, genomic alterations in sAML are characterized by a marked complexity reflecting both genomic instability and genetic heterogeneity of sAML. However, in our study several regions of interest including recurrently affected candidate genes such as TCF1, RUNX1, and TP53 were identified that need to be further investigated on a single gene level. Disclosures: No relevant conflicts of interest to declare.