Search

We investigated genome-wide DNA methylation patterns in 64 pediatric patients with acute myeloid leukemia (AML). Based on unsupervised clustering with the 567 most variably methylated cytosine guanine dinucleotide (CpG) sites, patients were categorized into 4 clusters associated with genetic alterations. Clusters 1 and 3 were characterized by the presence of known favorable prognostic factors, such as RUNX1-RUNX1T1 fusion and KMT2A rearrangement with low MECOM expression, and biallelic CEBPA mutations (all 8 patients), respectively. Clusters 2 and 4 comprised patients exhibiting molecular features associated with adverse outcomes, namely internal tandem duplication of FLT3 (FLT3-ITD), partial tandem duplication of KMT2A, and high PRDM16 expression. Depending on the methylation values of the 1243 CpG sites that were significantly different between FLT3-ITD+ and FLT3-ITD− AML, patients were categorized into 3 clusters: A, B, and C. The STAT5-binding motif was most frequently found close to the 1243 CpG sites. All 8 patients with FLT3-ITD in cluster A harbored high PRDM16 expression and experienced adverse events, whereas only 1 of 7 patients with FLT3-ITD in the other clusters experienced adverse events. PRDM16 expression levels were also related to DNA methylation patterns, which were drastically changed at the cutoff value of PRDM16/ABL1 = 0.10. The assay for transposase-accessible chromatin sequencing of AMLs supported enhanced chromatin accessibility around genomic regions, such as HOXB cluster genes, SCHIP1, and PRDM16, which were associated with DNA methylation changes in AMLs with FLT3-ITD and high PRDM16 expression. Our results suggest that DNA methylation levels at specific CpG sites are useful to support genetic alterations and gene expression patterns of patients with pediatric AML.

Recent advances in the genetic understanding of acute myeloid leukemia (AML) have improved clinical outcomes in pediatric patients. However, ∼40% of patients with pediatric AML relapse, resulting in a relatively low overall survival rate of ∼70%. The objective of this study was to reveal the comprehensive genetic background of pediatric AML. We performed transcriptome analysis (RNA sequencing [RNA-seq]) in 139 of the 369 patients with de novo pediatric AML who were enrolled in the Japanese Pediatric Leukemia/Lymphoma Study Group AML-05 trial and investigated correlations between genetic aberrations and clinical information. Using RNA-seq, we identified 54 in-frame gene fusions and 1 RUNX1 out-of-frame fusion in 53 of 139 patients. Moreover, we found at least 258 gene fusions in 369 patients (70%) through reverse transcription polymerase chain reaction and RNA-seq. Five gene rearrangements were newly identified, namely, NPM1-CCDC28A, TRIP12-NPM1, MLLT10-DNAJC1, TBL1XR1-RARB, and RUNX1-FNBP1. In addition, we found rare gene rearrangements, namely, MYB-GATA1, NPM1-MLF1, ETV6-NCOA2, ETV6-MECOM, ETV6-CTNNB1, RUNX1-PRDM16, RUNX1-CBFA2T2, and RUNX1-CBFA2T3. Among the remaining 111 patients, KMT2A-PTD, biallelic CEBPA, and NPM1 gene mutations were found in 11, 23, and 17 patients, respectively. These mutations were completely mutually exclusive with any gene fusions. RNA-seq unmasked the complexity of gene rearrangements and mutations in pediatric AML. We identified potentially disease-causing alterations in nearly all patients with AML, including novel gene fusions. Our results indicated that a subset of patients with pediatric AML represent a distinct entity that may be discriminated from their adult counterparts. Based on these results, risk stratification should be reconsidered.

TO THE EDITOR: Acute myeloid leukemia (AML) is a complicated disease characterized by the uncontrolled proliferation of hematopoietic precursors and the loss of differentiation ability caused by various genetic alterations. Recent advances in massively parallel sequencing technologies have

[Background] Customized gene panel sequencing is commonly used in cancer research to detect point mutations and small insertions or deletions. This technique can also be used to detect copy number alterations (CNAs) in tumor samples based on the comparison of sequence depth of targeted regions between samples and variant allele frequencies of germline single nucleotide polymorphisms (SNPs). CNAs in TP53 and CDKN2A/2B are frequently found in various types of leukemias. TP53 loss was reported to be a frequent CNA with poor prognosis in adult AML, and CDKN2A/2B alterations were found to be frequent in adult and pediatric acute lymphoblastic leukemia. However, the molecular and clinical profiles of the CNAs of these genes remain unclear in pediatric AML. [Patients and Methods] We analyzed clinical samples from 331 pediatric AML patients in the AML-05 trial, which was conducted by the Japanese Pediatric Leukemia/Lymphoma Study Group. Targeted sequencing was performed using a 343-gene custom panel and next-generation sequencer. These 343 genes are reportedly associated with hematopoietic malignancy or solid tumor pathogenesis. Copy number analysis was performed using the in-house pipeline CNACS. Additionally, a SNP array was performed for 40 of the 331 patients, including 34 patients with complex karyotype, to detect CNAs of TP53. [Results] TP53 alterations were identified in 7 (2.1%) of 331 patients. TP53 losses and mutations were detected in 7 and 4 patients, respectively, and all the patients with TP53 mutations concurrently had TP53 losses. Seven patients with TP53 losses were identified by CNACS, although only 4 patients with TP53 losses were found by SNP array. As for the cytogenetics of 7 patients with TP53 losses, six carried a complex karyotype, lacking FAB-M7 morphology. Accordingly, 6 (28.6%) of 21 non-FAB-M7 patients with complex karyotype had TP53 losses. Fusion genes and gene mutations were rare in patients with TP53 losses (NUP98-KDM5A, n=1; KIT, n=1; NRAS, n=1; CEBPA, n=1; ASXL1, n=1), and no patients with core-binding factor AML (CBF-AML) or KMT2A-rearrangements had TP53 losses. In survival analyses, patients with TP53 losses had significantly worse overall survival, event-free survival, and cumulative incidence of relapse rates than patients with wildtype TP53 (14.3% vs 71.4%, p CDKN2A/2B alterations were identified in 10 patients (3.0%) (9 losses and 1 mutation). Three patients with CDKN2A/2B losses had simultaneous TP53 alterations. Unlike TP53 alterations, CDKN2A/2B alterations were found in patients with CBF-AML and those with FAB-M7. In survival analyses, no significant difference was observed between patients with or without CDKN2A/2B alterations. [Discussion] TP53 losses were identified in approximately 2% of patients with pediatric AML. However, these alterations were found to be frequent in patients with complex karyotypes, particularly in those with non-FAB-M7 (28.6%). Furthermore, patients with TP53 losses had very poor prognosis. These findings indicate the necessity of risk stratification using TP53 status for pediatric AML. The efficacy of hematopoietic stem cell transplantation (HSCT) for the treatment of patients with TP53 losses remains unclear because 6 of 7 patients with TP53 losses received HSCT in this study. Moreover, because some of the identified TP53 mutations were possible germline mutations, HSCT may increase the incidence rate of the secondary cancer in patients with TP53 mutations. Thus, the efficacy of new agents (e.g., azacitizine and gemtuzumab ozogamicin) should be evaluated in future clinical trials for the intensification of chemotherapy for these patients. Meanwhile, the clinical and molecular profiles of CDKN2A/2B alterations were not fully identified in this study. However, the high incidence of CDKN2A/2B alterations in patients with TP53 losses indicates that the correlation between these gene alterations is essential for oncogenesis in a subset of pediatric AML patients. Figure Disclosures Ogawa: ChordiaTherapeutics, Inc.: Consultancy, Equity Ownership; Kan Research Laboratory, Inc.: Consultancy; RegCell Corporation: Equity Ownership; Asahi Genomics: Equity Ownership; Qiagen Corporation: Patents & Royalties; Dainippon-Sumitomo Pharmaceutical, Inc.: Research Funding.

Introduction Treatment outcomes of pediatric acute myeloid leukemia (AML) have improved via stratification therapy. However, relapse and mortality are still observed in 40% and 30% patients, respectively. Recently, comprehensive analyses using next-generation sequencing have revealed novel genetic alterations in adult AML patients, but insufficient analyses have been performed in pediatric AML patients. Further studies are warranted to evaluate the relationship between a more detailed genetic background and prognosis in pediatric AML patients. Methods We performed targeted sequencing using a 343-gene custom panel and next-generation sequencer in 302 pediatric patients with de novo AML who participated in the Japanese AML-05 study conducted by the Japanese Pediatric Leukemia/Lymphoma Study Group from 2006 to 2010. One hundred and twenty patients with core binding factor (CBF)-AML, 52 with KMT2A rearrangements, 56 with normal karyotype (NK), 18 with complex karyotype (CK), and 57 with other karyotypes except for patients with acute megakaryocytic leukemia were enrolled. Correlations among gene mutations, other cytogenetic alterations, and clinical characteristics were investigated. Results and Discussion We found the following 6 recurrent gene mutations, which have not been frequently reported in pediatric AML: KMT2C (n = 13, 4.3%), PHF6 (n = 6, 2.0%), MGA (n = 13, 4.3%), TET2 (n = 11, 3.6%), JAK3 (n = 12, 4.0%), and GATA2 (n = 11, 3.6%). KMT2C mutations have been reported in adult and pediatric patients with CBF-AML and in a few patients with other types of AML; however, their clinical significance has not been revealed. Six of 13 patients with KMT2C mutations were found in CBF-AML, whereas the other 7 patients were detected in non CBF-AML (1 patient in KMT2A-rearrangement, 1 in CK, and 5 in other karyotypes). Loss of function mutations in KMT2C have also been reported to be associated with chemotherapy resistance in pediatric AML patients. Although no significant differences were found, 4 of 13 patients with KMT2C mutations did not achieve complete remission (p = .072), and 3 patients relapsed in In adult AML, PHF6 mutations have been detected in 3% of patients and have been reported as poor prognostic factors. In 6 patients with PHF6 mutations, 3 patients had a normal karyotype, and 2 patients had a complex karyotype. No PHF6 mutations were detected in AML patients with CBF or KMT2A rearrangements. PHF6 mutations were significantly frequently detected in AML patients with myelodysplasia-related changes (4/6, p = .008). Although 5 of 6 patients with PHF6 mutations archived complete remission, finally, 5 patients relapsed or died. There were significant differences in EFS between patients with or without PHF6 mutations (3-year EFS 16.7% vs. 58.2%; p = .041). MGA mutations have been reported thus far in a few AML patients. MGA is a negative regulator of MYC signaling; thus, loss of functions of MGA is considered to contribute to tumorigenesis. However, the clinical significance of MGA mutations has not been reported. MGA mutations were significantly frequently detected in AML patients with FAB M2 (7/13, p = .020). In our cohort, 2 of 3 AML patients with DEK-NUP214 had MGA mutations (p = .004). We did not find correlations between MGA mutations and prognosis (3-year OS 83.3% vs. 75.5%; p= .488, 3-year EFS 66.7% vs. 56.7%; p = .507) TET2 mutations have been found in 8%-19% of adult AML cases and are considered poor prognostic factors. In previous reports, TET2 mutations were also detected in 3.8% of pediatric AML patients, but because of the small number of patients, the association between TET2 mutations and prognosis has not been revealed. TET2 mutations were significantly frequently detected in older AML patients (median age 13.95 vs. 9.0 years; p = .022) and in AML patients with NK (5/11, p = .035). Contrary to previous reports on adult AML, TET2 mutations were not associated with prognosis (3-year OS 77.9% vs. 75.9%; p= .688, 3-year EFS 54.5% vs. 57.4%; p = .907). This suggested that TET2 mutations have a different clinical significance between adult and pediatric AML patients. In conclusion, mutations in the KMT2C and PHF6 genes were associated with poor prognosis, but TET2 mutations were not. Further analysis should be needed about TET2 mutations. Disclosures Ogawa: ChordiaTherapeutics, Inc.: Consultancy, Equity Ownership; Kan Research Laboratory, Inc.: Consultancy; RegCell Corporation: Equity Ownership; Asahi Genomics: Equity Ownership; Qiagen Corporation: Patents & Royalties; Dainippon-Sumitomo Pharmaceutical, Inc.: Research Funding.

[Introduction] Infant AML has been defined as AML in patients aged [Methods] This study enrolled patients aged [Results] We first analyzed the distribution of fusion genes and FAB classification in 503 patients with clinical samples. Fusion genes were found in >50% of the patients in each age group. RUNX1-RUNX1T1, KMT2A-R, and CBFB-MYH11 were present in almost all age groups. RUNX1-RUNX1T1 was the most frequently identified fusion gene in all age groups excluding those Furthermore, we analyzed the prognosis of KMT2A-R and CBFB-MYH11 in all 723 patients and confirmed that these fusions had significantly different prognoses between patients In patients with KMT2A-R (n = 131), no molecular markers were identified as prognostic markers in combination with age; however, high WBC counts were related to poor prognosis regardless of age. Particularly, among patients 3-10,000/µL (n = 37) had a significantly poorer prognosis than those with WBC In patients with CBFB-MYH11 with clinical samples (n = 39), NRAS-mutated patients (n = 13) had significantly better 5y EFS and CIR than NRAS-wild-type patients (n = 26) (100% vs 60.5%, p = 0.012, and 0% vs 35.7%, respectively). Moreover, the combination of NRAS status with age identified the worst prognosis in NRAS-wild-type patients [Conclusions] Although age has not been used for the risk stratification of recent pediatric AML clinical trials, it might be a useful prognostic maker of this disease in combination with some molecular makers. Disclosures No relevant conflicts of interest to declare.

Introduction Pediatric acute myeloid leukemia (AML) comprises less than 20% of pediatric leukemia, representing one of the major therapeutic challenges in pediatric oncology. Approximately 40% of patients still have a relapse after first-line therapies, and the expected long-term survival rate decreases following relapse. Stem cell transplantation (SCT) was a conclusive strategy for de novo AML patients with a high risk and relapsed or refractory patients with standard and intermediate risk in a recent clinical trial. AML is a molecularly and clinically heterogeneous disease caused by various genetic alterations. Thus, it is difficult to accurately evaluate risk stratification even if known representative molecular markers, including KIT, FLT3-ITD, t(8;21)/RUNX1-RUNX1T1, and KMT2A (also known as MLL)-rearrangements, are used. Methods We investigated differences in the genetic background between SCT and non-SCT groups in participants of the Japan Pediatric Leukemia/Lymphoma Study Group (JPLSG) AML-05 trial. Among 369 patients with de novo AML, 175 patients received SCT. A standardized form was used to record clinical variables, including patient demographic information. The clinical data of patients in each risk group were followed for 3 years after the date of final registration. JPLSG performed a central review of morphologic classification and karyotyping based on the World Health Organization Classification, French-American-British classification, and cytogenetic analysis using conventional G-banding. Molecular characterization included mutational analyses of KIT (exons 8 and 17), N- and K-RAS (exons 1 and 2), NPM1 (exon 12), CEBPA (exon 12), FLT3-ITD, NUP98-NSD1,and CBFA2T3-GLIS2 gene rearrangement, as well as KMT2A- partial tandem duplication (MLPA methods). We also evaluated the gene expression of MECOM (also known as EVI1) and PRDM16 (also known as MEL1) because their high expressions are known poor prognostic markers. Overall survival (OS) was defined as the time from AML diagnosis to death or censorship at the last follow-up. Event-free survival (EFS) was defined as the time from AML diagnosis to treatment failure, relapse, death, or last follow-up. Results The 3-year OS among SCT patients (n = 175) was approximately 50%. It was significantly worse than that of non-SCT patients (n = 194, 90%; P < 0.001). Among 137 CBF-AML patients, 44 patients (32%) received SCT and their 3-year OS was 80%. On the other hand, among 232 non-CBF-AML patients, 131 patients (57%) received SCT, and their 3-year OS was only 35%. This result indicated that SCT is beneficial for relapsed CBF-AML, whereas most non-CBF patients who received SCT did not obtain the clinical benefit. In terms of the molecular characteristics, among 58 patients with high EVI1 expression, 39 patients (67%) received SCT, and their 3-year OS was approximately 35%. EVI1 expression was especially useful when using with MLL rearrangement because prognosis of patients with both MLL rearrangement andhigh EVI1 expression were extremely poor. On the other hand, among 84 patients with high PRDM16 expression, 63 patients (75%) received SCT, and their 3-year OS was approximately 20%. Although all AML patients with FLT3-ITD were assigned to receive SCT in the AML-05 trial, FLT3-ITD(+) patients withlow PRDM16 expression had a better outcome than FLT3-ITD(+) patients with high PRDM16 expression (3-year OS: 70% vs. 21%; P < 0.001). This result indicated that FLT3-ITD itself might not necessarily be associated with poor prognosis. PRDM16 gene expression is a useful marker to select patients needing SCT, particularly for patients with FLT3-ITD. Conclusion SCT was beneficial for patients with CBF-AML, whereas SCT was insufficient to rescue patients with non-CBF AML who relapsed, particularly patients with both FLT3-ITD(+) and high PRDM16 expression. New strategies, such as gemtuzumab ozogamicin or haploidentical SCT, are urged to rescue high risk/refractory patients. Disclosures No relevant conflicts of interest to declare.

Background: Pediatric acute myeloid leukemia (AML) comprises approximately 20% of pediatric leukemia, representing one of the major therapeutic challenges in pediatric oncology with a current overall survival rate of less than 70%. The pathogenesis of AML is heterogeneous and can be caused by various chromosomal aberrations, gene mutations/epigenetic modifications, and deregulated/overregulated gene expressions, leading to increased proliferation and decreased hematopoietic progenitor cell differentiation. Recurrent chromosomal structural aberrations [e.g., t(8;21), inv(16), and MLL-rearrangements] have been well established as diagnostic and prognostic markers of AML. Furthermore, recurrent mutations in FLT3, KIT, NPM1, and CEBPA have been reported in both adult and pediatric AML. Recently, massively parallel sequencing enabled the discovery of recurrent mutations in DNMT3A, TET2, and IDH, which are clinically useful for the prediction of the prognosis. However, these mutations are rare in pediatric AML, suggesting that other genetic alterations exist in pediatric AML. In contrast, recent reports have described NUP98-NSD1 fusion as an adverse AML prognostic marker and PRDM16 (also known as MEL1) as the representative overexpressed gene in patients harboring NUP98-NSD1 fusion. Intriguingly, PRDM16 overexpression occurs in nearly one-quarter of all children, with AML involving NUP98-NSD1-negative patients. Moreover, this overexpression is enriched in specimens with other high-risk lesions (e.g., FLT3-ITD, NUP98-NSD1, and MLL-PTD). Patients and Methods: To reveal a complete registry of gene rearrangements and other genetic lesions in pediatric AML with a normal karyotype, we performed transcriptome analysis (RNA sequencing) of 61 of 70 de novo pediatric AML patients with a normal karyotype using Illumina HiSeq 2000. We could not perform RNA sequencing in nine patients because of a lack of RNA quantity or quality. Among the 70 AML patients with a normal karyotype, 33 patients overexpressed PRDM16, which was found to be strongly associated with a poor prognosis in our previous studies. All patients were enrolled and treated with AML-05 in the study conducted by the Japan Pediatric Leukemia/Lymphoma Study Group (JPLSG). We also analyzed the known genetic mutations associated with these patients using the data derived from RNA sequencing. Results: A total of 144 candidate gene rearrangements, which were not observed in normal samples, were identified in 51 of 61 samples. Many of the recurrent gene rearrangements identified in this study involved previously reported targets in AML, including NUP98-NSD1, NUP98-JARID1A, CBFA2T3-GLIS2, MLL-MLLT10, and MLL-MLLT3. However, several gene rearrangements were newly identified in the current study, including MLL-SEPT6, HOXA10-HOXA-AS3, PRDM16-SKI, and CUL1-EZH2. We have also performed the validation of these novel gene rearrangements using Sanger sequencing. Most of these gene rearrangements were found in patients with a high expressionof PRDM16. In contrast, CEBPA mutations were frequently observed in patients with a low expression of PRDM16. Known gene alterations, such as FLT3-ITD and MLL-PTD, and mutations of the RAS, KIT, CEBPA, WT1, and NPM1genes were also detected using RNA sequencing. Conclusion: RNA sequencing unmasked a complexity of gene rearrangements and mutations in pediatric AML genomes. Our results indicate that a subset of pediatric AML represents a discrete entity that could be discriminated from adult counterparts, regarding the spectrum of gene rearrangements and mutations. In the present study, we identified at least one potential gene rearrangement or driver mutation in nearly all AML samples, including some novel fusion genes. These findings suggest that gene rearrangements in conjunction with mutations also play essential roles in pediatric AML. Disclosures Ogawa: Kan research institute: Consultancy, Research Funding; Takeda Pharmaceuticals: Consultancy, Research Funding; Sumitomo Dainippon Pharma: Research Funding.

Background Acute myeloid leukemia (AML) is a complex disease caused by various genetic alterations. Some prognosis-associated cytogenetic aberrations or gene mutations such as FLT3-internal tandem duplication (ITD), t(8;21)(q22;q22)/RUNX1-RUNX1T1, and inv(16)(p13q22)/CBFB-MYH11 have been found and used to stratify the risk. Numerous gene mutations have been implicated in the pathogenesis of AML, including mutations of DNMT3A, IDH1/2, TET2 and EZH2 in addition to RAS, KIT, NPM1, CEBPA and FLT3in the recent development of massively parallel sequencing technologies. However, even after incorporating these molecular markers, the prognosis is unclear in a subset of AML patients. Recently, NUP98-NSD1 fusion gene was identified as a poor prognostic factor for AML. We have reported that all pediatric AML patients with NUP98-NSD1 fusion showed high expression of the PR domain containing 16 (PRDM16; also known as MEL1) gene, which is a zinc finger transcription factor located near the breakpoint at 1p36. PRDM16 is highly homologous to MDS1/EVI1, which is an alternatively spliced transcript of EVI1. Furthermore, PRDM16 is essential for hematopoietic stem cell maintenance and remarkable as a candidate gene to induce leukemogenesis. Recent reports revealed that high PRDM16 expression was a significant marker to predict poor prognosis in pediatric AML. However, the significance of PRDM16 expression is unclear in adult AML patients. Methods A total of 151 adult AML patients (136 patients with de novo AML and 15 patients with relapsed AML) were analyzed. They were referred to our institution between 2004 and 2015 and our collaborating center between 1996 and 2013. The median length of follow-up for censored patients was 30.6 months. Quantitative RT-PCR analysis was performed using the 7900HT Fast Real Time PCR System with TaqMan Gene Expression Master Mix and TaqMan Gene Expression Assay. In addition to PRDM16, ABL1 was also evaluated as a control gene. We investigated the correlations between PRDM16 gene expression and other genetic alterations, such as FLT3-ITD, NPM1, and DNMT3A, and clarified the prognostic impact of PRDM16 expression in adult AML patients. Mutation analyses were performed by direct sequence analysis, Mutation Biased PCR, and the next-generation sequencer Ion PGM. Results PRDM16 overexpression was identified in 29% (44/151) of adult AML patients. High PRDM16 expression correlated with higher white blood cell counts in peripheral blood and higher blast ratio in bone marrow at diagnosis; higher coincidence of mutation in NPM1 (P = 0.003) and DNMT3A (P = 0.009); and lower coincidence of t(8;21) (P = 0.010), low-risk group (P = 0.008), and mutation in BCOR (P = 0.049). Conversely, there were no significant differences in age at diagnosis and sex distribution. Patients with high PRDM16 expression tended to be low frequency in M2 (P = 0.081) subtype, and the remaining subtype had no significant differences between high and low PRDM16 expression. Remarkably, PRDM16 overexpression patients were frequently observed in non-complete remission (55.8% vs. 26.3%, P = 0.001). Patients with high PRDM16 expression tended to have a cumulative incidence of FLT3-ITD (37% vs. 21%, P = 0.089) and MLL-PTD (15% vs. 5%, P = 0.121). We analyzed the prognosis of 139 patients who were traceable. The overall survival (OS) and median survival time (MST) of patients with high PRDM16 expression were significantly worse than those of patients with low expression (5-year OS, 17% vs. 32%; MST, 287 days vs. 673 days; P = 0.004). This trend was also significant among patients aged Conclusions We investigated the correlations among PRDM16 expression, clinical features, and other genetic alterations to reveal clinical and prognostic significance. High PRDM16 expression was independently associated with non-CR and adverse outcomes in adult AML patients, as well as pediatric AML patients. Our finding indicated that the same pathogenesis may exist in both adult and pediatric AML patients with respect to PRDM16 expression, and measuring PRDM16 expression was a powerful tool to predict the prognosis of adult AML patients. Disclosures Inokuchi: Bristol-Myers Squibb: Honoraria, Research Funding; Novartis: Honoraria; Celgene: Honoraria; Pfizer: Honoraria.

[Background] Pediatric acute myeloid leukemia (AML) remains to be a disease with a poor prognosis, with a long-term survival rate of approximately 60%. We have recently reported that high PRDM16 expression (high-PRDM16) was common in t(5;11)(q35;p15.5)/NUP98-NSD1-positive patients, and that high-PRDM16 was found to be a novel poor prognostic factor in this disease. However, this finding was not observed in FAB-M7. The presence of inv(16)(p13.3q24.3)/CBFA2T3-GLIS2 was postulated to be a reason. This fusion was reported as the most frequent fusion gene associated with a poor prognosis in FAB-M7, and we found that all CBFA2T3-GLIS2-positive patients had a low PRDM16 expression (low-PRDM16). Furthermore, some studies reported that bone morphogenetic protein 2 (BMP2) was frequently and highly expressed in CBFA2T3-GLIS2-positive patients, indicating the importance of BMP2 in pathogenesis in a subset of AML. In this study, we analyzed the expression status of BMP2 and PRDM16 and their relationship to prognosis in pediatric AML. [Patients and Methods] A total of 369 AML patient samples, who were enrolled in the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG) AML-05 trial, were analyzed. Expression level of BMP2 and PRDM16 is analyzed by real-time quantitative PCR (Taqman Gene Expression Assay), using ABL1 as a control gene. The cutoff value for high-BMP2 was defined as the lowest expression level among 11 CBFA2T3-GLIS2-positive patients, and that for high-PRDM16 was defined as the lowest expression level among 10 NUP98-NSD1-positive patients. We also screened fusion genes and gene mutations of CBFA2T3-GLIS2, NUP98-JARID1A, RBM15-MKL1, MLL rearrangements, NUP98-NSD1, FUS-ERG, DEK-NUP214, RUNX1-RUNX1T1, CBFB-MYH11, and gene mutations of KIT, NRAS, KRAS, WT1, NPM1, MLL-PTD, and FLT3-ITD. [Results] Out of 369 patients, high-BMP2 and high-PRDM16 were identified in 47 (12.7%) and 44 (11.9%) patients, respectively. Only three (0.8%) of 369 patients had both high-BMP2 and high-PRDM16 (high-BMP2/PRDM16). In 44 high-BMP2 patients and 41 high-PRDM16 patients (3 high-BMP2/PRDM16 patients were excluded), MLL rearrangements were only found in high-BMP2 patients (25.5% vs 0%, p < 0.001), whereas FLT3-ITD was significantly much more frequent in high-PRDM16 patients (4.5% vs 56.1%, p < 0.001). Core-Binding-Factor (CBF)-AML was rare in both groups (4.5% and 0%, p = 0.495). Kaplan|Meier analyses revealed that high-BMP2 patients and high-PRDM16 patients had a low 5-year overall survival (OS) (46.3% and 29.3%, respectively). In molecular subgroup analyses, the prognosis of MLL-rearranged patients, including 12 high-BMP2 patients (21.8%), was not influenced by BMP2 expression (5-year OS: high-BMP2 66.7% vs low-BMP2 64.9%, p = 0.488), suggesting that the prognosis of MLL-rearranged patients might primarily be defined by their fusion partner genes and other gene aberrations as previously reported (e.g., EVI1 expression). Considering this result, further survival analyses of the patients lacking CBF-AML or MLL-rearrangements were performed (n = 174, 5-year OS: 50.5%). In this subgroup, 5-year OS of high-BMP2 patients (n = 30) and high-PRDM16 patients (n = 41) were significantly lower than that of low-BMP2/low-PRDM16 patients (36.7% vs 67.2%, p < 0.001, and 29.3% vs 67.2%, p < 0.001, respectively). [Conclusions] High-BMP2 and high-PRDM16 could identify distinct subgroups with poor prognosis in pediatric AML. Further investigation of molecular-targeted therapy against these genes may contribute to the improvement of clinical outcome in these subgroups. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures No relevant conflicts of interest to declare.

Background Several molecular markers, such as FLT3-internal tandem duplication (ITD), NPM1, CEBPA are well known to correlate with mortality in patients with acute myeloid leukemia (AML). Recently, a number of gene mutations have been implicated in the pathogenesis of AML, including mutations of DNMT3A, IDH1/2, TET2 and EZH2 in addition to RAS, KIT and FLT3. However, DNMT3A, IDH1/2, and TET2 are rare in pediatric patients with AML, thus accurate risk evaluation remained challenging even after incorporating these molecular markers. On the other hand, overexpression of the EVI1 gene is reported to be associated with adverse outcome in pediatric AML. Moreover, we have previously reported that measuring of PRDM16 gene expression was a powerful tool to predict the prognosis of pediatric AML. PRDM16 gene is highly homologous to the MDS1/EVI1 gene, which is an alternatively spliced transcript of the EVI1 gene. In this study, we investigated EVI1 gene expression to verify the prognosis of EVI1 gene expression and the relationship between EVI1 and PRDM16 gene expression. Methods Between 2006 and 2010, 485 de novo pediatric AML patients participated in the Japanese AML-05 study conducted by the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG). Among them, 116 patients were excluded from the study because of misdiagnosis and unavailability of their RNA samples. Therefore, 369 patients were analyzed. Quantitative RT-PCR analysis was performed in these patients using the 7900HT Fast Real Time PCR System with TaqMan Gene Expression Master Mix and TaqMan Gene Expression Assay. In addition to EVI1 and PRDM16, ABL1 was also evaluated as a control gene. We investigated the correlations between these gene expressions and other genetic alterations, and clarified the prognostic impact of EVI1 and association between EVI1 and PRDM16 genes. Results A total of 58 of 369 patients (15.7%) showed high expression of EVI1 gene. Overexpression of EVI1 gene was strongly associated with dismal prognosis; low risk (LR; 1 of 123 patients, or 0.8%); intermediate risk (IR; 38 of 147 patients, or 25.9%); high risk (HR; 6 of 50 patients, or 12%); and non-complete remission (Non-CR; 13 of 49 patients, or 26.5%), (P < 0.001). Overexpression of EVI1 correlated with the following characteristics: younger age at diagnosis; M4, M5, and M7 subtype; higher coincidence of MLL-rearrangement; and lower coincidence of t(8;21), and inv(16). EVI1 overexpression was very frequent among patients with de novo pediatric AML and IR/non-CR groups. Furthermore, more than half of patients in M6 (5 of 8 patients, or 62.5%) were EVI1 high expression. Interestingly, no patients with EVI1 high expression in M7 had a fusion of CBFA2T3-GLIS2. Patients with EVI1 overexpression also more frequently harbored a complex karyotype and monosomy 7. The frequencies of patients with high or low EVI1 expression differed widely with respect to each genetic alteration, as follows: t(8;21), 1% vs 99%, P < 0.001; inv(16), 0% vs 100%, P < 0.001; NUP98-JARID1A, 83% vs 17%, P < 0.001; OTT-MAL, 100% vs 0%, P = 0.02; and KIT, 5% vs 95%, P = 0.003. The overall survival (OS) and event-free survival (EFS) among patients with EVI1 high expressions were significantly lower than that among patients without such gene aberrant expression (3-year OS 54% vs. 77%, P=0.008G3-year EFS: 34% vs 58%, P On the other hand, a total of 84 of 369 patients (22.8%) showed high expression of PRDM16 gene. The OS and EFS among PRDM16 overexpressing patients were significantly worse than those among low expression group (3-year OS: 51% vs 81%, P Conclusions We investigated EVI1 and PRDM16 gene expression in de novo pediatric AML patients, and their high expression was associated with inferior survival, respectively. We suggest that high EVI1 and/or PRDM16 expression is useful marker for adverse outcome. On the other hand, low EVI1 and PRDM16 expressions are useful to elucidate low risk patients. Disclosures No relevant conflicts of interest to declare.

Background Pediatric acute myeloid leukemia (AML) comprises approximately 20% of pediatric leukemia cases. AML is a major therapeutic challenge in pediatric oncology, and the current overall survival rate is In addition, recent studies have reported that the NUP98-NSD1 fusion is an adverse AML prognostic marker and that PRDM16 (also termed MEL1) is a representative gene that is overexpressed in patients who have the NUP98-NSD1 fusion. PRDM16 overexpression occurs in nearly a quarter of pediatric AML patients who are NUP98-NSD1 negative, and this overexpression is increased in specimens with other high-risk lesions (e.g., FLT3-ITD, NUP98-NSD1,and MLL-PTD). Patients and Methods To obtain a complete overview of gene rearrangements and other genetic lesions, we performed RNA sequencing of samples from 47 de novo pediatric AML patients using Illumina HiSeq 2000, including 39 patients with normal karyotypes and 6 patients with Trisomy 8. Among these 47 patients, 35 patients overexpressed PRDM16, which was strongly associated with a poor prognosis in our previous studies. As a control, we selected 12 patients with low PRDM16 expression levels. All patients were enrolled and treated as part of the AML-05 study conducted by the Japan Pediatric Leukemia/Lymphoma Study Group. We determined the known gene mutations present in these patients using the RNA sequencing data. Results Approximately 300 candidate gene rearrangements were identified in 46/47 samples, including 26 in-frame and 78 out-of-frame gene rearrangements. Several recurrent gene rearrangements identified in this study involved previously reported targets in AML, such as FUS-ERG, NUP98-NSD1,and MLL-MLLT3. However, several novel gene rearrangements were identified in the current study, including HOXA10-HOXA-AS3, PRDM16-XXX, CUL1-YYY, and DAZAP1-ZZZ. At present, we are validating these novel gene rearrangements using Sanger sequencing. Known gene alterations, such as FLT3-ITD, MLL-PTD, and mutations of RAS, KIT, CEBPA, WT1, and NPM1 genes, were detected by RNA sequencing. Conclusion In the present study, RNA sequencing was employed to elucidate the complexity of gene rearrangements/mutations in pediatric AML genomes. Our results indicate that a subset of pediatric AML represents a discrete entity that can be discriminated from adult AML in terms of the spectrum of gene rearrangements/mutations. We identified at least one potential gene rearrangement or driver mutation in nearly all AML samples, including various novel fusion genes. Thus, our results suggest that gene rearrangements and mutations play essential roles in pediatric AML. Disclosures No relevant conflicts of interest to declare.

Background; Acute myeloid leukemia (AML) is a molecularly and clinically heterogeneous disease. Currently, a number of gene mutations have been implicated in the pathogenesis of both adult and pediatric AML, including mutations of CEBPA,NPM1, DNMT3A, IDH1/2, TET2 and EZH2 in addition to RAS, KIT and FLT3, because the recent development of massively parallel sequencing technologies. We have performed whole-exome sequencing of paired tumor-normal DNA from 19 patients, and identified 80 somatic mutations or 4.2 mutations per sample. Many of the recurrent mutations identified in this study involved previously reported targets in adult AML, such as FLT3, CEBPA, KIT, CBL, NRAS, WT1, BCORL1, EZH2, and major cohesin components including SMC3 and RAD21. In addition to these mutations, we also identified disease-associated candidate genes of ASXL2, PAX5 and others. Recently, recurrent somatic mutations in ASXL1 have been reported to occur in patients with adult AML, and to be associated with adverse outcome. Another study suggested that ASXL1 or ASXL2 mutations were associated with a high incidence of relapse. To reveal the significance of these mutations, we performed mutational analysis of ASXL1 and ASXL2 in 184 pediatric AML patients. Methods; Between 2006 and 2010, 485 de novo pediatric AML patients aged To estimate the frequency and prognostic impact of the ASXL1 and ASXL2 mutations in pediatric AML, we performed targeted sequencing of ASXL1 (exon 12) and ASXL2 (exon 12) genes using next-generation sequencer in 184 de novo AML patients including 51 patients with t(8;21). We validated the mutations in ASXL1 and ASXL2 by Sanger sequencing. Furthermore, we investigated the correlation among these mutations, other cytogenetic alterations and clinical characteristics. Results and Discussion; ASXL1 mutations were identified in 4 of 184 de novo pediatric AML patients (2.2 %) and all 4 ASXL1 mutation positive patients harbored t(8;21). Two of them relapsed, and one died after relapsed. On the other hand, ASXL2 mutations were identified in 10 of 184 de novo pediatric AML patients (5.4%) and 6 of them harbored t(8;21). Five of these 10 patients relapsed, and 2 of them died after relapsed. Especially, all of 4 patients without t(8;21) relapsed (2 in M5a and 2 with CBFA2T3-GLIS2 in M7), and 2 died after relapsed. Although only one ASXL2 patients with t(8;21) relapsed (1/6 or 17%), no ASXL2 positive patients with t(8;21) died. ASXL2 mutations were more observed in AML patients with t(8;21) (11.8%, 6/51), but not in 13 patients with inv(16). Overall survival of the patients with or without ASXL2 mutations were 80% and 66.7% (p=0.54), respectively. ASXL1 and ASXL2 mutations were mutually exclusive in this study. Conclusion; Ten of 184 patients (5.4%) had mutations of ASXL2 in pediatric AML, and the outcome of ASXL2 mutant patients with t(8;21) was favorable. Among the 51 pediatric AML patients with t(8;21), ASXL2 mutations were detected in 6 (11.8%) patients. All of them have been survived, suggesting that ASXL2 mutations in patients with t(8;21) may be associated with favorable prognosis in pediatric AML in contrast to adult AML. On the other hand, 4 (2.2%) of 184 patients had ASXL1 mutations, and all of them were t(8;21). In these 4 patients, 2 of them relapsed, and one died after relapsed. Although the number of patients is too small, ASXL1 mutations were not considered to be associated with favorable outcome. Both ASXL1 and ASXL2 mutations were detected at high frequency among pediatric AML patients with t(8;21) and mutual exclusive. As we consider that further study will be needed to clarify the significance of these mutations, we are now analyzing mutations in other exons of ASXL2, and would like to report these data in the annual meeting. Disclosures No relevant conflicts of interest to declare.

Background Acute myeloid leukemia (AML) is a complex disease caused by mutations, epigenetic modifications, and deregulated expression of genes, leading to increased proliferation and decreased differentiation of hematopoietic progenitor cells. Although, many prognosis-associated gene alterations have been identified in adult AML, genetic alterations responsible for an adverse outcome in most patients with pediatric AML have remained obscure.In our previous study, we have identified NUP98-NSD1 fusion gene in 6 of 124 pediatric AML patients. As a result of gene expression profile using microarray, all these 6 patients have showed high expression of HOXA9, HOXA10, HOXB3, HOXB5, HOXB6, and a zinc finger transcription factor gene MDS1-EVI1-like-1 (MEL1), and clustered a distinct poor prognostic subgroup [4-year-overall survival (OS); 33%]. Furthermore, this subgroup has also formed the bigger poor prognostic subgroup involving the 18 patients without NUP98-NSD1 fusion gene. We defined this subgroup as NUP98-NSD1 signature consisting of 24 of 124 patients (19%), 4-year-OS; 37.5%. MEL1 gene expression represented the specific characteristics of this subgroup.To verify the significance of MEL1 gene expression and the relationship with other poor prognostic markers such as FLT3-ITD, MLL-PTD, and NPM1, we investigated MEL1 expression by real-time PCR in 369 pediatric AML cases other than acute promyelocytic leukemia (APL) and Down syndrome-associated AML, because these subtypes formed genetically different unique entities and treated by other protocols. Interestingly, MEL1 is highly homologous to MDS1/EVI1 gene, which is an alternatively spliced transcript of the EVI1 genes. As EVI1 high expression has been reported as a poor prognostic marker in pediatric AML patients, especially in patients with MLL rearrangement, we also measured and examined the EVI1 gene expression. Methods Between 2006 and 2010, 485 de novo pediatric AML patients participated in the Japanese AML-05 study conducted by the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG). Among them, 369 samples were available in this study. Quantitative RT-PCR analysis was performed in these patients using the 7900HT Fast Real Time PCR System with TaqMan Gene Expression Master Mix and TaqMan Gene Expression Assay. In addition to EVI1 and MEL1, ABL1 was also evaluated as a control gene. We investigated the correlations between these gene expressions and other genetic alterations, and clarified the prognostic impact of MEL1 gene. Results A total of 84 of 369 patients (22.8%) showed high expression of MEL1 gene. The number of patients with MEL1 high expressions were gradually enriched in order of the intermediate risk (IR; 34 of 142 patients, or 23.9%), high risk (HR; 21 of 48 patients, or 43.8%), and non-complete remission (non-CR; 20 of 39 patients, or 51.3%), but were almost all absent in patients with a low risk (LR) cytogenetic profile (4 of 120 patients, or 3.3%) consisting of t(8;21) and inv(16). The absolute values of MEL1 gene expression have also gradually increased in order of LR < IR < HR < non-CR. The overall survival among patients with MEL1 high expressions was significantly lower than that among patients without such gene aberrant expression (50% vs. 82%, P Conclusions MEL1 high expression was highly recurrent in de novo pediatric AML patients with high/intermediate-risk cytogenetic profiles and was independently associated with a poor outcome. Combined with MEL1expression and genetic alterations enable to clarify the genetic background of pediatric AML profiles. The data in this study showed a way in which integrated mutational profiling of a clinical trial cohort can advance our understanding of the biologic characteristics of AML, improve current prognostic models, and inform prospective therapeutic decisions. Disclosures No relevant conflicts of interest to declare.

Abstract 2536 Background Acute myeloid leukemia (AML) is a complex disease caused by mutations, epigenetic modifications, and deregulated expression of genes leading to increased proliferation and decreased differentiation of hematopoietic progenitor cells. Although many important molecular markers have already been discovered in AML, no known prognosis-associated cytogenetic aberrations or mutations were detectable in a subset of AML patients. In this regard, recent reports of somatic mutations affecting the C-terminal zinc finger (ZF) 2 of GATA2 are intriguing, because these GATA2 mutations were associated with the progression of chronic myeloid leukemia, whereas hereditary ZF2 of GATA2 mutations predispose to AML and myelodysplastic syndrome. GATA2 mutations were also reported as a predisposing gene of the monocytopenia, mycobacterial infection (MonoMAC) syndrome and the dendritic cell, monoctye, B and NK lymphoid deficiency (DCML) syndrome. GATA2 belongs to a family of zinc finger transcription factors, and is important for hematopoietic stem cell proliferation and normal megakaryocytic development. These findings prompted us to search for possible GATA2 mutations in pediatric AML. Methods To explore the frequency and clinical impact of GATA2 mutations, we examined 157 Japanese pediatric AML patients, including 13 with FAB-M3 and 10 with Down syndrome (DS) who were treated on different treatment protocol, by PCR following direct sequencing. As GATA2 mutations thus far reported almost exclusively involved exons 4–6 that encode zinc finger 1 and 2 domain, we confined our analysis to these exons. Results GATA2 missense mutations were found in 7 out of 157 patients (4.5%). Notably, All of GATA2 mutations were located in ZF2 in this study, although almost all of GATA2 mutations in adult AML were located in ZF1. Wild type GATA2 were found in 3 of 3 AML patients with GATA2 mutation. The other 4 patients had no history of Mono MAC syndrome, suggesting that these mutations were acquired. The zinc finger region of GATA2 is required for binding to promyelocytic leukemia zinc finger (PLZF) protein can interact with GATA2 and can modify its transactivation capacity. Interestingly, 2 GATA2 mutations were found in FAB-M3 in this study, GATA2 mutations also may be associated with acute promyelocytic leukemia. Clinical and molecular features between patients with and without GATA2 mutations were not significantly different in the clinical parameters (WBC, age, sex, etc.), and the outcome of GATA2 mutation positive patients was not poor when compared to GATA2 mutation negative patients: Two of the 7 patients received allogeneic-stem cell transplantation (Allo-SCT) and one of them died of gastrointestinal hemorrhage after SCT. The other 5 patients who did not receive the SCT were still alive. Conclusion GATA2 is a new predisposition gene for pediatric AML and shows functional changes caused by mutations within a highly conserved threonine repeat located in ZF2. Interestingly, MonoMAC and DCML syndromes were not observed in de novo AML patients with GATA2 mutations. Although further investigation is needed, our results indicated that GATA2 mutations were associated with a favorable outcome in pediatric AML. Most of the patients with GATA2 mutations have been classified into an intermediate risk group in our study, however, their favorable outcome suggests that less aggressive treatment strategy without SCT might be appropriate for AML patients carrying GATA2 mutations. Disclosures: No relevant conflicts of interest to declare.

Abstract 1391 Background Acute myeloid leukemia (AML) is a complex disease caused by mutations, epigenetic modifications, and deregulated expression of genes leading to increased proliferation and decreased differentiation of hematopoietic progenitor cells. Many important molecular markers have already been discovered in AML that have not only helped to better characterize patients, but also helped to improve risk stratification. However, in a subset of AML patients, no known prognosis-associated cytogenetic aberrations or mutations were detectable. In hematological malignancies, 11p15 translocations involving the nucleoporin 98-kDa (NUP98) protein gene are relatively rare, but more than 20 different chromosomal rearrangements have been identified. The cryptic t(5;11)(q35;p15.5), which is frequently accompanied by deletion of the long arm of chromosome 5, del(5q), creates a fusion gene between NUP98 and the nuclear receptor-binding SET-domain protein 1 (NSD1) gene. Recently, Hollink et al identified NUP98-NSD1 gene fusion by high-resolution genome-wide copy number analysis and reverse transcription (RT)-PCR in pediatric and adult AML patients, and reported that NUP98-NSD1 gene fusion was a predictor of poor prognosis. Thus, we explored the frequency, gene expression pattern, and clinical impact of NUP98-NSD1 in a large cohort of Japanese pediatric AML patients. Methods From January 2000 to December 2002, 318 patients were newly diagnosed with de novo AML. Of these, samples from 124 patients whose gene expression profiling data were available were analyzed, including 10 patients with FAB-M3 and 6 patients with Down syndrome (DS) who were treated on different treatment protocols. For the detection of NUP98-NSD1 gene fusion, direct sequencing was performed using an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). Gene expression profiling data, which were performed before, were available for these 124 patients. Hierarchical clustering analysis was performed using Cluster and Tree View software. Result NUP98-NSD1 fusion transcript was detected in 6 (4.8%) of 124 pediatric AML patients and a characteristic gene expression pattern related to these patients was also detected. Further hierarchical clustering analysis using the 87 probe sets which were differentially expressed in 6 NUP98-NSD1-positive patients did not detect a simple cluster of 6 NUP98-NSD1-positive patients but a relatively large cluster including 18 NUP98-NSD1-negative patients. When compared with 100 NUP98-NSD1 signature-negative patients, the 24 NUP98-NSD1 signature-positive patients frequently had normal karyotype (62.5%), and del(9q) (8.3%), but did not have any favorable chromosomal translocations such as t(8;21), inv(16) and t(15;17). The frequencies of M4 and M5 subtypes (14/24, P < 0.001) were also higher than those in NUP98-NSD1 signature-negative patients. In addition, they frequently had FLT3-ITD (12/24, P < 0.001), MLL-PTD (7/24, P = 0.124), NPM1 (3/24, P = 0.007) and WT1 (5/24, P = 0.037) mutations. In total, NUP98-NSD1 related gene expression signature (NUP98-NSD1 signature) represented 19% (24/124) of total patients and 58% (15/26) of cytogenetically normal patients in our cohort. Their 4-year overall survival (OS) and event-free survival (EFS) were poor when compared to 100 NUP98-NSD1 signature-negative patients (37.5% vs. 86.0% and 37.5% vs. 72.0%). Conclusion Our results indicated that Patients with NUP98-NSD1 related gene expression signature had extremely poor clinical outcome, irrespective of the presence of NUP98-NSD1 gene fusion. Most of the patients displaying the NUP98-NSD1 signature have been classified into an intermediate risk group, but their unfavorable outcome suggests that a high risk group is a more suitable stratification. These results suggest the presence of a new poor prognostic subgroup which revealed the value of the NUP98-NSD1 signature as prognostic markers in pediatric AML. Although further investigation is necessary, we believe that our work contributes to improve the risk stratification in pediatric AML. Disclosures: No relevant conflicts of interest to declare.

Abstract 1475 Background: Acute myeloid leukemia (AML) is a complex disease caused by mutations, deregulated gene expression, and epigenetic modifications of genes leading to increased proliferation and decreased differentiation of hematopoietic progenitor cells. In hematological malignancies with 11p15 translocations, the nucleoporin 98-kDa (NUP98) protein regulates nucleocytoplasmic transport of protein and RNA. More than 20 different chromosomal rearrangements involving NUP98 have been identified. The cryptic translocation t(5;11)(q35;p15.5), which creates a fusion gene between the NSD1 and the NUP98. The NUP98-NSD1 fusion gene has been associated with a deletion of the long arm of chromosome 5, del(5q), in pediatric AML patients. Recently, NUP98-NSD1 was identified in childhood AML by fluorescence in situ hybridization (FISH) with subtelomeric probes. However, the frequency and clinical impact of NUP98-NSD1 in pediatric AML remain uncertain. Methods: To estimate the frequency and clinical impact of the NUP98-NSD1 in pediatric AML, we examined the NUP98-NSD1 fusion transcript in a consecutive series of 157 newly diagnosed AML patients enrolled in the AML99 protocol in Japan clinical trial. Total RNA extracted from the leukemia cells of the patient at diagnosis was reverse transcribed to cDNA. The PCR was performed using the primer sets for detection of NUP98-NSD1. The PCR amplification was performed with a DNA thermal cycler followed by direct sequencing with an ABI PRISM 310 Genetic Analyzer. We also examined somatic mutations of the FLT3, KIT, WT1, NPM1, NRAS, KRAS and MLL genes, which are prevalent in AML, in these AML patients with NUP98-NSD1. Results: In the current study, we identified the NUP98-NSD1 by RT-PCR followed by direct sequencing in 6 (3.8%) out of 157 AML patients, and clustered significantly in the cytogenetically normal AML subgroup (6/33, 18.2%). Significant differences between patients with and without NUP98-NSD1 were observed in 4-year OS (33.3% versus 80.0%), and EFS (33.3% versus 69.2%). NUP98-NSD1 was highly seen in males (5 males and 1 female). The statistical differences between patients with and without NUP98-NSD1 was not significant in age (median: 7 years [range: 2–15 years] versus 6 years [range: 0–15 years]), and the initial WBC count (median: 115.9 × 109/L [range: 9.0–329 × 109/L] versus 51.8 × 109/L [range: 1.0–621 × 109/L]). Interestingly, NUP98-NSD1 was observed in 4 patients with normal karyotype, and was observed in 2 patients with del(9q). NUP98-NSD1 could not be identified by conventional G-banding technique in any cases. Patients with NUP98-NSD1 had any mutation of FLT3 -ITD (66.7%), WT1 (50%), KIT (16.7%) or RAS (33.3%). Furthermore, NUP98-NSD1 was frequently found in M4 or M5 patients according to FAB classification. Conclusions: We identified the NUP98-NSD1 in 6 out of 157 AML patients, and clustered significantly in the cytogenetically normal AML subgroup. Patients with NUP98-NSD1 had any FLT3 -ITD, WT1, KIT or RAS mutations. These results suggest that simultaneous occurrence of FLT3 -ITD, WT1, KIT mutations and RAS aberrations in NUP98 -related leukemia may be associated with poor prognosis. The NUP98-NSD1 was not identified by conventional G-banding technique, confirming that this translocation can exist as the only cytogenetic alteration. Therefore, FISH and RT-PCR analyses for NUP98-NSD1 fusion gene is recommended at the onset of disease in larger series of AML patients presenting an apparently normal karyotype, del(9q) or del(5q) to gain insight into the actual incidence and clinical features of patients with t(5;11)(q35;p15.5). Disclosures: No relevant conflicts of interest to declare.

Abstract 5106 NADP+dependent isocitrate dehydrogenase 1 (IDH1) and 2 (IDH2) catalyzes the oxidative carboxylation of isocitrate to α-ketoglutarate (α-KG) in citric acid cycle. Recently, recurrent somatic missense mutations in IDH1 at codon R132, as well as IDH2 at codon R172 have been identified in low-grade gliomas/secondary glioblastoma by high throughput sequencing. Subsequent studies also revealed that acquired somatic mutations in IDH1 and IDH2 frequently occurred in adult hematological malignancies, such as acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS), and that these mutations were associated with older age, poor prognosis, cytogenetically normal AML, and the genotype of mutated NPM1 without FLT3-internal tandem duplication (ITD). Tumor-derived IDH1 and IDH2 mutations impair the affinity of the enzymes for the substrates and dominantly inhibit wild-type IDH1 and IDH2 activities through the formation of catalytically inactive heterodimers. However, little is known about the incidence and prognostic values of IDH1 and IDH2 mutations in pediatric hematological malignancies. Here, we analyzed mutations that involve the activation sites of IDH1 and IDH2 using polymerase chain reaction amplification/sequencing in a total of 244 samples of pediatric myeloid malignancies as well as infantile leukemia including 17 AML-derived cell lines, 115 primary cases of AML, 28 primary cases of MDS, 15 primary cases of juvenile myelomonocytic leukaemia (JMML), 6 chronic myeloid leukemia (CML)-derived cell line, 18 primary cases of CML and 45 infantile leukemia(6 AML and 39 acute lymphoblastic leukemia (ALL) patients). Moreover, to assess whether IDH1 and IDH2 mutations overlap with known gene abnormalities, such as FLT3, c-KIT, and NPM1 mutations, mutational analyses of FLT3, c-KIT, and NPM1 were also performed. The common IDH2 R140Q mutation was detected in a single AML case, whereas no IDH1 mutation was detected in samples of myeloid malignancies. Although no IDH2 mutation was detected in infantile leukemias, novel P127S, H133I and I130V of IDH1 mutations were detected in 4 of 45 (8.9%) infantile ALL cases. No IDH1 and IDH2 mutations were detected in the JMML, MDS, or CML samples examined. Six AML samples including one cell line had c-KIT mutations (D816V, N822K, or D419fs), 12 AML cases had FLT3-ITD and 10 infantile leukemia cases had FLT3 mutations (D835E or I836). The NPM1 mutation was detected in 2 of 132 AML samples. The AML case harboring the IDH2 mutation, Case 39 was a 12-year-old boy diagnosed as AML-M2 according to the French-American-British cooperative group classification system having t(8;21)(q22;q22), showed no abnormalities of NPM1, c-KIT, and FLT3. Remarkably, among 4 infantile ALL cases with IDH1 mutations, 3 cases showed mixed lineage leukemia(MLL) rearrangements with t(4;11). The FLT3-D835 mutation was found in 1 of 4 patients with IDH1 mutations. These results suggested that IDH1 mutations are one of the second genetic events in infant ALL with MLL rearrangements; however, it was concluded that the involvements of IDH1 and IDH2 mutations in the pathogenesis of pediatric myeloid malignancies are extremely rare. Disclosures: No relevant conflicts of interest to declare.

Abstract 2149 Recent reports of somatic mutations of the CBL proto-oncogene in myeloid neoplasms are intriguing, because these CBL mutations were shown to results in aberrant tyrosine kinase signaling, which would lead also to activation of RAS signaling pathways. We and others reported that CBL mutations occurred in a variety of myeloid neoplasms, including de novo acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and MDS/myeloproliferative neoplasm, especially in chronic myelomonocytic leukemia (CMML) and juvenile myelomonocytic leukemia (JMML). The importance of CBL mutations concerning leukemogenesis is substantially increased. We investigated CBL mutations in 20 therapy-related leukemia/MDS (t-Leuk/MDS) cases and 26 infant leukemia cases. Homozygous mutation of theCBL gene (P417R), which was located in the RING finger domain, was identified in one out of 20 (5%) t-Leuk/MDS cases. This patient was a 5 year-old boy, whose biopsied specimen of the buccal lymph node showed malignant lymphoma (diffuse large T cell type, MT1(+), MB1(-), UCHL1(+)). No pathogenic nucleotide changes were identified in the CBL gene in the initial sample. Subsequently, the patient was treated with chemotherapy including VP-16 (200 mg/m2) given twice weekly. Nineteen months after initial diagnosis, he was diagnosed as having therapy-related leukemia with t(5;21) and MLL gene rearrangement due to VP-16. Furthermore, CBL gene mutations were found in 3 of 26 (12%) infant leukemia cases with 11q23 translocation/MLL gene rearrangement. CBL gene mutations were located in splice site in intron 8 and the RING finger domain (Y371H, in 2 cases), SNP array analysis (Affymetrix, GeneChip) of these cases with mutated CBL gene disclosed 11q-acquired uniparental disomy in all cases, but not in cases with wild-type CBL. CBL mutation has not been reported in acute leukemia with 11q23 translocation/MLL gene rearrangement. To our knowledge, these are the first t-Leuk/MDS case and infant cases with 11q23 translocation/MLL rearrangement, suggesting that CBL is mutated in a unique subset of t-Leuk/MDS and infant leukemia which is considered to play a pathogenic role in the development of t-Leuk/MDS and infant leukemia. Further accumulation of cases with t-Leuk/MDS and infant leukemia having the CBL mutation is needed. Disclosures: No relevant conflicts of interest to declare.

Activating mutations of the NOTCH1 gene have been reported in about 50% of T cell acute lymphoblastic leukemia (T-ALL). We performed mutation analysis of the NOTCH1, FLT3 and p53 genes by polymerase chain reaction followed by direct sequence. Mutations of the NOTCH1 were identified in 24 (30%) of 80 fresh samples and 10 (71.4%) of 14 T-ALL cell lines. Six missense mutations and 2 insertion in HD domain, 2 nonsense mutations and 6 insertions in PEST domains were found in 14 cell lines. Eight missense mutations, 9 insertions and one deletion in HD domains, 5 missense mutations, 3 nonsense mutations and 3 deletions in PEST domain were found in 80 fresh samples. The incidence of the NOTCH1 mutations is less frequent than that of previous reports. We observed about 95% of single nucleotide polymorphisms (5097 C/T) in HD domain, which is more frequent than 30% of previous report, possibly due to the racial difference. FLT3 internal tandem duplication, which is known as the poor prognostic factor in acute myeloid leukemia, were not identified in any T-ALL cell lines or fresh samples. Mutations of the p53 gene were found in 5 of 8 cell lines and 5 of 50 fresh samples. Mutations of both NOTCH1 and p53 genes were identified 3 of 8 cell lines and one of 50 fresh samples. In our study NOTCH1 mutations were not significantly associated with high WBC count and prognosis. Further studies are needed to find the association between novel genes and clinical features of pediatric T-ALL.