Search

Supplemental Material and Methods

Purpose:In chronic lymphocytic leukemia (CLL), TP53 mutations are associated with reduced survival and resistance to standard chemoimmunotherapy (CIT). Nevertheless, the clinical impact of subclonal TP53 mutations below 10% to 15% variant allele frequency (VAF) remains unclear.Experimental Design:Using a training/validation approach, we retrospectively analyzed the clinical and biological features of TP53 mutations above (high-VAF) or below (low-VAF) the previously reported 10.0% VAF threshold, as determined by deep next-generation sequencing. Clinical impact of low-VAF TP53 mutations was also confirmed in a cohort (n = 251) of CLL treated with fludarabine-cyclophosphamide-rituximab (FCR) or FCR-like regimens from two UK trials.Results:In the training cohort, 97 of 684 patients bore 152 TP53 mutations, while in the validation cohort, 71 of 536 patients had 109 TP53 mutations. In both cohorts, patients with the TP53 mutation experienced significantly shorter overall survival (OS) than TP53 wild-type patients, regardless of the TP53 mutation VAF. By combining TP53 mutation and 17p13.1 deletion (del17p) data in the total cohort (n = 1,220), 113 cases were TP53 mutated only (73/113 with low-VAF mutations), 55 del17p/TP53 mutated (3/55 with low-VAF mutations), 20 del17p only, and 1,032 (84.6%) TP53 wild-type. A model including low-VAF cases outperformed the canonical model, which considered only high-VAF cases (c-indices 0.643 vs. 0.603, P < 0.0001), and improved the prognostic risk stratification of CLL International Prognostic Index. Clinical results were confirmed in CIT-treated cases (n = 552) from the retrospective cohort, and the UK trials cohort.Conclusions:TP53 mutations affected OS regardless of VAF. This finding can be used to update the definition of TP53 mutated CLL for clinical purposes.

Purpose: In chronic lymphocytic leukemia (CLL), TP53 mutations are associated with reduced survival and resistance to standard chemoimmunotherapy (CIT). Nevertheless, the clinical impact of subclonal TP53 mutations below 10% to 15% variant allele frequency (VAF) remains unclear. Experimental Design: Using a training/validation approach, we retrospectively analyzed the clinical and biological features of TP53 mutations above (high-VAF) or below (low-VAF) the previously reported 10.0% VAF threshold, as determined by deep next-generation sequencing. Clinical impact of low-VAF TP53 mutations was also confirmed in a cohort (n = 251) of CLL treated with fludarabine-cyclophosphamide-rituximab (FCR) or FCR-like regimens from two UK trials. Results: In the training cohort, 97 of 684 patients bore 152 TP53 mutations, while in the validation cohort, 71 of 536 patients had 109 TP53 mutations. In both cohorts, patients with the TP53 mutation experienced significantly shorter overall survival (OS) than TP53 wild-type patients, regardless of the TP53 mutation VAF. By combining TP53 mutation and 17p13.1 deletion (del17p) data in the total cohort (n = 1,220), 113 cases were TP53 mutated only (73/113 with low-VAF mutations), 55 del17p/TP53 mutated (3/55 with low-VAF mutations), 20 del17p only, and 1,032 (84.6%) TP53 wild-type. A model including low-VAF cases outperformed the canonical model, which considered only high-VAF cases (c-indices 0.643 vs. 0.603, P < 0.0001), and improved the prognostic risk stratification of CLL International Prognostic Index. Clinical results were confirmed in CIT-treated cases (n = 552) from the retrospective cohort, and the UK trials cohort. Conclusions: TP53 mutations affected OS regardless of VAF. This finding can be used to update the definition of TP53 mutated CLL for clinical purposes.

KRAS mutations account for the most frequent mutations in human cancers, and are generally correlated with disease aggressiveness, poor prognosis, and poor response to therapies. KRAS is required for adult hematopoiesis and plays a key role in B cell development and mature B cell proliferation and survival, proved to be critical for B cell receptor-induced ERK pathway activation. In mature B cell neoplasms, commonly seen in adults, KRAS and RAS-MAPK pathway aberrations occur in a relevant fraction of patients, reaching high recurrence in some specific subtypes like multiple myeloma and hairy cell leukemia. As inhibitors targeting the RAS-MAPK pathway are being developed and improved, it is of outmost importance to precisely identify all subgroups of patients that could potentially benefit from their use. Herein, we review the role of KRAS and RAS-MAPK signaling in malignant hematopoiesis, focusing on mature B cell lymphoproliferative disorders. We discuss KRAS and RAS-MAPK pathway aberrations describing type, incidence, mutual exclusion with other genetic abnormalities, and association with prognosis. We review the current therapeutic strategies applied in mature B cell neoplasms to counteract RAS-MAPK signaling in pre-clinical and clinical studies, including most promising combination therapies. We finally present an overview of genetically engineered mouse models bearing KRAS and RAS-MAPK pathway aberrations in the hematopoietic compartment, which are valuable tools in the understanding of cancer biology and etiology.

not available

In chronic lymphocytic leukemia (CLL), the mechanisms controlling cell growth and proliferation in the presence of NOTCH1 mutations remain largely unexplored. By performing a gene expression profile of NOTCH1-mutated (NOTCH1-mut) versus NOTCH1 wild-type CLL, we identified a gene signature of NOTCH1-mut CLL characterized by the upregulation of genes related to ribosome biogenesis, such as nucleophosmin 1 (NPM1) and ribosomal proteins (RNPs). Activation of NOTCH1 signaling by ethylenediaminetetraacetic acid or by coculture with JAGGED1-expressing stromal cells increased NPM1 expression, and inhibition of NOTCH1 signaling by either NOTCH1-specific small interfering RNA (siRNA) or γ-secretase inhibitor reduced NPM1 expression. Bioinformatic analyses and in vitro activation/inhibition of NOTCH1 signaling suggested a role of MYC as a mediator of NOTCH1 effects over NPM1 and RNP expression in NOTCH1-mut CLL. Chromatin immunoprecipitation experiments performed on NOTCH1 intracellular domain (NICD)-transfected CLL-like cells showed the direct binding of NOTCH1 to the MYC promoter, and transfection with MYC-specific siRNA reduced NPM1 expression. In turn, NPM1 determined a proliferation advantage of CLL-like cells, as demonstrated by NPM1-specific siRNA transfection. In conclusion, NOTCH1 mutations in CLL are associated with the overexpression of MYC and MYC-related genes involved in protein biosynthesis including NPM1, which are allegedly responsible for cell growth and/or proliferation advantages of NOTCH1-mut CLL.

// Elena Vendramini 1, 2, 3 , Marco Giordan 1 , Emanuela Giarin 1 , Barbara Michielotto 1 , Grazia Fazio 4 , Gianni Cazzaniga 4 , Andrea Biondi 4 , Daniela Silvestri 4 , Maria Grazia Valsecchi 5 , Martina U. Muckenthaler 6 , Andreas E. Kulozik 6 , Valter Gattei 7 , Shai Izraeli 2, 3 , Giuseppe Basso 1 and Geertruy te Kronnie 1 1 Department of Women’s and Children’s Health, University of Padova, Padova, Italy 2 Edmond and Lily Safra Children’s Hospital, Sheba Medical Center, Ramat Gan, Israel 3 Tel Aviv University, Tel Aviv, Israel 4 Centro Ricerca Tettamanti, Clinica Pediatrica, University of Milano-Bicocca, Monza, Italy 5 School of Medicine and Surgery, University of Milano-Bicocca, Milano, Italy 6 Department of Pediatric Oncology Hematology, University of Heidelberg, Heidelberg, Germany 7 Clinical and Experimental Onco-Hematology Unit, Centro di Riferimento Oncologico, I.R.C.C.S., Aviano (PN), Italy Correspondence to: Elena Vendramini, email: elena.vendramini@gmail.com Keywords: B cell precursor acute lymphoblastic leukemia, ERG aberrations, noncoding RNAs, miR-125, SNORD116 Received: June 29, 2016 Accepted: March 04, 2017 Published: March 21, 2017 ABSTRACT ERG-related leukemia is a B cell precursor acute lymphoblastic leukemia (BCP ALL) subtype characterized by aberrant expression of DUX4 and ERG transcription factors, and highly recurrent ERG intragenic deletions. ERG-related patients have remarkably favorable outcome despite a high incidence of inauspicious IKZF1 aberrations. We describe clinical and genomic features of the ERG-related cases in an unselected cohort of B-other BCP ALL pediatric patients enrolled in the AIEOP ALL 2000 therapeutic protocol. We report a small noncoding RNA signature specific of ERG-related group, with up-regulation of miR-125b-2 cluster on chromosome 21 and several snoRNAs in the Prader-Willi locus at 15q11.2, including the orphan SNORD116 cluster.

Children with Down syndrome (DS) are at increased risk for acute myeloid leukemias (ML-DS) characterized by mixed megakaryocytic and erythroid phenotype and by acquired mutations in the GATA1 gene resulting in a short GATA1s isoform. The chromosome 21 microRNA (miR)-125b cluster has been previously shown to cooperate with GATA1s in transformation of fetal hematopoietic progenitors. In this study, we report that the expression of miR-486-5p is increased in ML-DS compared with non-DS acute megakaryocytic leukemias (AMKLs). miR-486-5p is regulated by GATA1 and GATA1s that bind to the promoter of its host gene ANK1. miR-486-5p is highly expressed in mouse erythroid precursors and knockdown (KD) in ML-DS cells reduced their erythroid phenotype. Ectopic expression and KD of miR-486-5p in primary fetal liver hematopoietic progenitors demonstrated that miR-486-5p cooperates with Gata1s to enhance their self renewal. Consistent with its activation of AKT, overexpression and KD experiments showed its importance for growth and survival of human leukemic cells. Thus, miR-486-5p cooperates with GATA1s in supporting the growth and survival, and the aberrant erythroid phenotype of the megakaryocytic leukemias of DS.

Despite the current risk-based stratification protocol, 15% of pediatric patients with B-Cell Precursor Acute Lymphoblastic Leukemia (BCP-ALL) still experience relapse. In the large subset of 'B-other' patients (negative for common fusion transcripts, non-high hyperdiploid and non-Down Syndrome), 'Ph-like' is a high risk subgroup with high incidence of relapses, which represents 30% of B-others or 10-15% of BCP-ALL patients. The PAX5 gene, encoding for a B-cell related transcription factor, is frequently involved in several translocations in Ph-like patients, determining the formation of fusion genes encoding for aberrant proteins. Our previous studies showed that PAX5 fusions sustain survival of leukemic cells by Lymphocyte kinase (LCK) hyperactivation, which can be targeted by the kinase inhibitor Nintedanib/BIBF1120. This study aims (i) to unravel the functional features of PAX5 fusion genes, elucidating the involved signaling pathways; (ii) to develop new pre-clinical strategies to target PAX5 fusion genes, testing the efficacy of the LCK inhibitor BIBF1120. We selected a cohort of 138 B-other cases among 400 childhood BCP-ALL cases enrolled in Italy in AIEOP-BFM ALL2000/R2006 protocols; by gene expression analysis (HG-U133Plus2.0 Affymetrix array), 59/138 presented as Ph-like, and by NGS, a driver fusion gene was identified in 34/59 (58%). Beside known genetic lesions, such as P2RY8/CLRF2 (N=9), EBF1 fusions (N=3), TCF3/HLF (N=1), IKZF1-fusion (N=1) and BCL9/MEF2D (N=1), and single cases with novel fusions, seven cases were carrying a PAX5 fusion gene, representing the most frequent rearrangement. Ex-vivo inhibition with BIBF1120 was setup on primary cells from 5 PAX5-fusion BCP-ALL samples, demonstrating significant efficacy in inducing leukemic cell apoptosis, both as single agent and in combination with standard chemotherapeutic agents (Annexin V viability assay of leukemic cells in co-culture on human bone marrow stroma). Strikingly, in 3/5 cases dexamethasone and BIBF1120 had a synergistic effect and were further tested in in vivo assays. A daily treatment was performed at bulk disease detection (BM aspiration, mean engraftment 20% CD10/CD19+) in patient-derived xenotransplanted NSG mice from 2 different PAX5 fusion cases. At the endpoint (after two weeks), in the PAX5/AUTS2 PDX mice we detected a mild effect in the BM by BIBF1120 alone (disease reduction 24%, p=0.057), further enhanced in combination with dexamethasone (-49%, p=0.005, with a mean engraftment in vehicle mice of 82.6%). In the spleen, the efficacy was highly significant both for BIBF1120 (-52%, p=0.025) and the combination (-91%, p=0.015, mean engraftment vehicle 69.5%). A similar statistical significant effect was observed also in peripheral blood, whilst, BIBF1120 alone showed a specific significant efficacy in CNS meninges. Analogous results have been observed in PDX from the PAX5/DACH2 PDX mice; BM leukemia decreased 47% using BIBF1120 alone (p=0.004), further diminished by the combination (-65%, p=0.0004, with a mean engraftment in vehicle mice of 65%). In the spleen, the efficacy was highly significant both with BIBF1120 (-45.6%, p=0.04) and for the combination (-96.3%, p=0.0008, mean engraftment vehicle 72.4%). Strikingly, BIBF1120 treatment alone showed analogous efficacy in PB and CNS, with leukemia decrease as low as -45% (p=0.04) and -76% (p=0.007), respectively, and the combination nearly achieved remission in PB (-94%, p=0.0001) and it was significant in CNS (-81.2%, p=0.03). Dexamethasone alone was not effective in the BM and spleen, whereas it decreased the leukemia bulk both in PB (-65%, p=0.0004) and CNS (-52.8%, p=0.03). Overall, BIBF1120 treatment was more effective than dexamethasone. Interestingly, phosphoflow analysis showed a marked inhibition by BIBF1120 of pAkt-Thr308 and its downstream effectors, such as pS6 and 4pEBP1, in ex vivo BM and spleen cells. In conclusions, PAX5 fusion genes are highly recurrent among Ph-like patients and can potentially be targeted by Nintendanib/BIBF1120, that showed a significant effect ex vivo and in vivo, even as a single agent. Disclosures No relevant conflicts of interest to declare.

In chronic lymphocytic leukemia (CLL), the mechanisms controlling cell growth and proliferation in the presence of NOTCH1 mutations remain largely unexplored. By performing a gene expression profile of NOTCH1-mutated (NOTCH1-mut) versus NOTCH1 wild-type CLL, we identified a gene signature of NOTCH1-mut CLL characterized by the upregulation of genes related to ribosome biogenesis, such as nucleophosmin 1 (NPM1) and ribosomal proteins (RNPs). Activation of NOTCH1 signaling by ethylenediaminetetraacetic acid or by coculture with JAGGED1-expressing stromal cells increased NPM1 expression, and inhibition of NOTCH1 signaling by either NOTCH1-specific small interfering RNA (siRNA) or γ-secretase inhibitor reduced NPM1 expression. Bioinformatic analyses and in vitro activation/inhibition of NOTCH1 signaling suggested a role of MYC as a mediator of NOTCH1 effects over NPM1 and RNP expression in NOTCH1-mut CLL. Chromatin immunoprecipitation experiments performed on NOTCH1 intracellular domain (NICD)-transfected CLL-like cells showed the direct binding of NOTCH1 to the MYC promoter, and transfection with MYC-specific siRNA reduced NPM1 expression. In turn, NPM1 determined a proliferation advantage of CLL-like cells, as demonstrated by NPM1-specific siRNA transfection. In conclusion, NOTCH1 mutations in CLL are associated with the overexpression of MYC and MYC-related genes involved in protein biosynthesis including NPM1, which are allegedly responsible for cell growth and/or proliferation advantages of NOTCH1-mut CLL.

Background Extensive sequencing data highlighted the recurrence of mutations resulting in constitutive MAP kinase signaling in chronic lymphocytic leukemia (CLL), and proposed these mutations as late drivers of CLL progression (X.S. Puente, Nature, 2015; D.A. Landau, Nature, 2015). Although some preliminary studies suggested KRAS, NRAS and BRAF mutations as possibly associated with trisomy 12 (TRI12) and/or unmutated (UM) IGHV genes (C.D. Herling, Blood, 2016; K. Takahashi, Blood, 2018), dedicated analysis of these aspects is still missing. Aim To correlate the occurrence of KRAS, NRAS and BRAF mutations with specific clinico-biological features in CLL. Methods The study included 534 primary CLL purposely enriched with 300 TRI12 cases: 190 cases with TRI12 as sole chromosomal aberration (TRI12only) and 110 cases harboring TRI12 plus other chromosomal aberrations (TRI12plus). The cohort was also enriched in cases with NOTCH1 aberrations (n=214) to evaluate their possible role in RAS-RAF mutations incidence. Amplicon-based targeted next-generation sequencing assay was performed using Miseq Illumina on DNA from purified CLL samples. Analyzed amplicons mapped in the hotspot regions of KRAS, NRAS (exons 2, 3 and 4) and BRAF (exons 11 and 15) genes. Variant allele frequency (VAF) was obtained by Miseq report. CLL cases were characterized for IGHV mutational status, main cytogenetic abnormalities, NOTCH1 and TP53 aberrations. Time to first treatment (TTFT) data were correlated with molecular findings. Results We found 90 missense point mutations (Fig.1) in 64 CLL cases, with prevalence of KRAS (44 mutations in 38 patients), followed by BRAF (31 mutations in 24 patients) and NRAS (15 mutations in 13 patients). Nearly all mutations have been previously associated with gain-of-function phenotype and increased RAS/ERK downstream signaling. The co-occurrence of 2 mutated genes (mainly KRAS and BRAF) were observed in 11 patients. Mutations were mainly subclonal (median VAF 6.15%, range 1.3%-61.6%) with only one third of mutations (27/90) above 15% VAF. A high association between the presence of KRAS/NRAS/BRAF mutations and UM IGHV and the presence of TRI12 was observed. Overall, 87.3% of KRAS/NRAS/BRAF mutated cases had a UM IGHV (p Furthermore, a higher prevalence of KRAS/NRAS/BRAF mutations was found in NOTCH1 wild type (wt) (22.1%) when compared to NOTCH1 mutated (11.2%) cases, in the context of the UM IGHV group (p=0.008), pointing to a mutual exclusivity of these mutations in the pathogenesis of disease. No other significant associations with clinical variables as RAI stage at diagnosis, presence of TP53 mutations or gender were observed. We finally evaluated if KRAS/NRAS/BRAF mutations had an impact on TTFT in the context of UM IGHV/TRI12only/NOTCH1wt, that showed the highest incidence of mutations (26/70, 37%). Both KRAS mutations alone (p=0.005) and KRAS/NRAS mutations (p=0.04) were associated with shorter TTFT (Fig.3). Conclusions Taken together our data show that KRAS, NRAS and BRAF mutations are almost exclusively found in UM IGHV/TRI12/NOTCH1wt CLL and, in the context of this CLL group the presence of KRAS and NRAS mutations is associated with unfavorable prognosis. The high incidence of KRAS, NRAS and BRAF activating mutations in TRI12 CLL together with the observation of higher level of ERK phosphorylation (S. Decker, Blood, 2012) and higher sensitivity to MEK/ERK inhibitors (S. Dietrich, J Clin Invest, 2018) described in the TRI12, reinforce the evidence of an essential role for MEK/ERK signaling in TRI12 CLL, pointing to this pathway as the most appropriate therapeutic target. Disclosures Zaja: Novartis: Honoraria, Research Funding; Janssen: Honoraria; Abbvie: Honoraria; Celgene: Honoraria, Research Funding; Amgen: Honoraria; Takeda: Honoraria; Sandoz: Honoraria.

Background Chronic lymphocytic leukemia (CLL) is characterized by a CD20 expression lower than other lymphoproliferative disorders, hampering treatment efficacy of anti-CD20 antibodies. Although most mechanisms behind this phenomenon are still obscure, we demonstrated that mutations of NOTCH1 associate with a lower CD20 expression, ascribed to a NOTCH1 mutation-driven repression of CD20 transcription by hystone deacetylases (Pozzo et al. Leukemia 2016). This may explain the association between NOTCH1 mutations and clinical resistance to anti-CD20 immunotherapy in CLL patients treated with FCR (Stilgenbauer et al. Blood 2014). SF3B1 mutations affect ~10% CLL cases at diagnosis; a recent characterization of their functional meaning highlighted transcriptome-wide alterations of splicing patterns affecting several pathways, including the NOTCH1 pathway (Wang et al. Cancer Cell 2016). In this context, one of the most aberrantly-spliced genes is DVL2, a key component of the Wnt pathway and negative regulator of the NOTCH1 pathway. Aim To investigate CD20 expression in SF3B1-mutated CLL and the correlation with a putative secondary activation of the NOTCH1 pathway through splicing alterations of the NOTCH1 negative regulator DVL2. Methods Mutational status of NOTCH1 and SF3B1 was evaluated by next generation sequencing. Mutations with VAF-variant allele frequency below 2% were discarded. CD20 expression was evaluated by flow cytometry on the CD19+5+ population. Percentage of low-CD20 expressing (CD20dim) population was defined as %P1-(100-%P1), P1 being a linear gate spanning from zero to CD20 mode. Transcript levels of MS4A1 (encoding for CD20 protein) and spliced DVL2 were evaluated by QRT-PCR in cases with at least 75% of CD19+5+ cells. Statistical analyses were performed using Mann-Whitney rank-test. Results In a cohort of 537 unselected CLL cases, 93 cases (17.3%) carried NOTCH1 mutations: 62 delCT, 20 other truncating, 11 3′UTR (VAF range 3-84%, mean 32%; NOTCH1-mut); 46 cases carried SF3B1 mutations (8.5%, VAF range 5-53%, mean 32%; SF3B1-mut), the hotspot surrounding K700 being the most frequent (50%) followed by the hotspot surrounding G742 (30%); 7 cases (1,3%) carried both NOTCH1 and SF3B1 mutations. As trisomy 12 CLL is characterized by increased expression of CD20 and mutation of SF3B1 co-occurred in only 5/121 cases, the trisomy 12 subset was excluded from further analyses. Confirming previous findings, NOTCH1-mut cases were characterized by a lower CD20 mean fluorescence intensity (MFI) compared to WT cases (p=0.0154). In agreement with the hypothesis of a more active NOTCH1 pathway, CD20 expression was found diminished also in those cases carrying SF3B1 mutations, compared to WT cases (p=0.0059). Since CD20 expression is highly variable between cases, spanning a 2/3-log range of fluorescence intensity, we evaluated the percentage of the CD20dim population, this quantity being independent from the actual MFI. This analysis confirmed that CLL cases with NOTCH1 or SF3B1 mutations show a greater proportion of CD20dim cells (NOTCH1-mut p Presence of the alternatively spliced isoform of DVL2 (altDVL2) was identified only in those CLL cases with SF3B1 mutations (p35%) to isolate CD20high and CD20low populations. MS4A1 expression (higher in the CD20high subset) inversely correlated with altDVL2 expression (Figure 1c). As DVL2 can act as negative regulator of NOTCH1, we followed the hypothesis that loss of WT DVL2 due to alternative splicing may result in a more active NOTCH1 pathway. To simulate loss of DVL2 in the SF3B1-wild type CLL-like MEC1 cell line, cells were trasfected with siRNA for DVL2 itself and screened for CD20 expression. After 24 hours from transfection, CD20 expression was found downregulated at both protein (p=0.0062; Figure 1d) and transcript (p=0.0129) level. Conclusions In addition to NOTCH1-mut cases, also CLL cases bearing SF3B1 mutations are characterized by a lower CD20 expression, allegedly through a more active NOTCH1 pathway, potentially resulting in clinical resistance to anti-CD20 monoclonal antibodies. Disclosures Zaja: Amgen: Honoraria; Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Janssen: Honoraria; Takeda: Honoraria; Sandoz: Honoraria; Abbvie: Honoraria.

Background Despite the current risk-based stratification and therapies, up to 20% of pediatric patients with B-cell precursor acute lymphoblastic leukemia (BCP-ALL) experience relapse. A major research effort is dedicated to identifying poor prognostic subgroups in the larger subset of patients with intermediate-risk disease, where most relapses occur. Recently, two groups in the US and the Netherlands independently identified a novel BCP-ALL subtype negative for the BCR/ABL fusion gene but with a gene-expression profile similar to Philadelphia chromosome-positive (Ph+) ALL. This ‘Ph-like’ (or ‘BCR/ABL-like') subtype encompasses 10-15% of BCP-ALL patients, predicts high incidence of relapses and defines a candidate subgroup for targeted treatment with tyrosine kinase inhibitors and alternative drugs. Aim Aim of this project was to identify BCP-ALL Ph-like cases and their prognosis in patients treated in Study Protocols of the Italian Association of Pediatric Hematology and Oncology (AIEOP) Methods In the context of the Microarray Innovations in LEukemia (MILE) study, gene expression profiling was successfully performed on 400 Italian childhood BCP-ALL cases enrolled in AIEOP-BFM ALL2000/R2006 protocols, from whom material was available. The MILE classifier recognized the most common genetic subgroups, namely t(4;11), t(12;21), t(1;19) or t(9;22) positive, or high hyperdiploid (DNA index≥1.16). Out of 400 BCP-ALL cases, 143 negative for common fusion transcripts, non-high hyperdiploid (DNA index Results Using signatures of known ALL subgroups, the likelihood to be close to one of the ‘known’ classes was defined for each of the B-others samples. Specifically, 43 B-other cases (43/143=30.1%, about 10% of total BCP-ALL cohort) presented as a cluster with a gene expression signature close to the BCR/ABL signature, therefore referred to as ‘Ph-like’. Among B-others, Ph-like cases had a significantly increased proportion of males, age>10 years and WBC>20x109/L. The 5y event-free survival (EFS) of Ph-like patients was 54.8% (SE 8.2) vs 83.1% (3.9) in the remaining B-others patients (p Conclusions and perspectives We identified a ‘Ph-like’ subgroup in the Italian cohort of children with BCP-ALL, associated to a poor outcome. Ph-like patients, independently of the other known risk features, could be considered eligible for alternative treatments. The genetic characterization of this subgroup is ongoing within the AIEOP-BFM research trial cooperative group, which is aimed at further defining the pathogenetic mechanisms and identifying the therapeutic translation. Disclosures: No relevant conflicts of interest to declare.

Pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL) has achieved an 80% cure rate as a result of a risk-adapted therapy largely based on minimal residual disease (MRD) monitoring. However, relapse is still the most frequent adverse event, occurring mainly in the patients with intermediate MRD levels (intermediate risk, IR), emphasizing the need for new prognostic markers. We analyzed the prognostic impact of cytokine receptor-like factor 2 (CRLF2) over-expression and P2RY8-CRLF2 fusion in 464 BCP-ALL patients (not affected by Down syndrome and BCR-ABL negative) enrolled in the AIEOP-BFM ALL2000 study in Italy. In 22/464 (4.7%) samples, RQ-PCR showed CRLF2 over-expression (≥20 times higher than the overall median). P2RY8-CRLF2 fusion was detected in 22/365 (6%) cases, with 10/22 cases also showing CRLF2 over-expression. P2RY8-CRLF2 fusion was the most relevant prognostic factor independent of CRLF2 over-expression with a threefold increase in risk of relapse. Significantly, the cumulative incidence of relapse of the P2RY8-CRLF2 + patients in the IR group was high (61.1% ± 12.9 vs 17.6% ± 2.6, P

SummaryWe investigated the engraftment properties and impact on patient outcome of 50 pediatric acute lymphoblastic leukemia (ALL) samples transplanted into NOD/SCID mice. Time to leukemia (TTL) was determined for each patient sample engrafted as weeks from transplant to overt leukemia. Short TTL was strongly associated with high risk for early relapse, identifying an independent prognostic factor. This high-risk phenotype is reflected by a gene signature that upon validation in an independent patient cohort (n = 197) identified a high-risk cluster of patients with early relapse. Furthermore, the signature points to independent pathways, including mTOR, involved in cell growth and apoptosis. The pathways identified can directly be targeted, thereby offering additional treatment approaches for these high-risk patients.

Abstract 3518 Introduction. In the AIEOP-BFM ALL 2000 study, the MRD-based risk groups stratification allowed to achieve more than 80% cure rate. However, relapse is still the most frequent adverse event, occurring mainly in the largest subgroup of non-high risk (HR) patients. This emphasizes the need for new prognostic markers for upfront identification of patients with a high risk of relapse. Recently, new genomic abnormalities of the CRLF2 and Ikaros (IKZF1) genes have been reported in the subset of BCP-ALL patients without known chromosomal aberrations. We recently analyzed the poor prognostic impact of P2RY8-CRLF2 fusion in a representative cohort of 464 non-DS Ph- BCP-ALL patients enrolled in Italy into AIEOP-BFM ALL2000 study from February 2003 to July 2005. Aim. In order to estimate the incidence and prognostic value of IKZF1 deletions, alone or in combination with CRLF2 alterations, we initiated the screening of the full cohort of patients. A pilot test of a subset of 154 non-DS, Ph-, BCP-ALL patients treated in Italy according to the AIEOP-BFM ALL 2000 protocol from February 2003 to June 2004 has been completed so far. Methods. Ig/TcR PCR-MRD and RT-PCR screening for known chromosomal translocation were routinely performed; P2RY8-CRLF2 fusion was detected by RT-PCR. Multiplex Ligation-dependent Probe Amplification (MLPA) (Salsa MLPA P335-A3 ALL-IKZF1 kit; MRC-Holland, Amsterdam, NL) was performed to identify IKZF1 deletions together with deletions in additional B-cell development genes. Patients positive for IKZF1 deletions were further analyzed by Salsa MLPA P202-A1 IKZF1 kit to confirm and better define the extension of the IKZF1 gene deletion. Results. IKZF1 deletions were detected in 22/154 cases (14.3%), in keeping with incidence data reported in the literature. In 9 cases (5.8%) the deletion was intra-genic, involving only some exons of the IKZF1 gene, while in 13 cases (8.4%) the deletion was encompassing the whole IKZF1 gene. Three out of 16 patients (18.8%) carried both IKZF1 and CRLF2 deletions. Overall, the 5-year Cumulative Incidence of Relapse (median follow-up: 5.8 years) was 23±9% for patients carrying IKZF1 deletions and 12.2±2.9% for patients who did not (p=0.22). Interestingly, out of 13 cases with complete deletion of IKZF1, only one case (1/13, 7.7%) relapsed, vs 4/9 cases (44.4%) positive for IKZF1 intra-gene deletion. Conclusions. IKZF1 intra-gene deletion, but not whole gene deletion, showed a tendency to be associated with poor prognosis in childhood BCP-ALL treated according to the AIEOP-BFM ALL 2000 protocol. Whether these results will be confirmed at the completeness of this study, the assessment of IKZF1 status might serve as a new stratification marker. It will be of particular interest to verify whether different IKZF1 deletions within non-HR patients by MRD will have a prognostic impact. Disclosures: No relevant conflicts of interest to declare.

We report gene expression and other analyses to elucidate the molecular characteristics of acute lymphoblastic leukemia (ALL) in children with Down syndrome (DS). We find that by gene expression DS-ALL is a highly heterogeneous disease not definable as a unique entity. Nevertheless, 62% (33/53) of the DS-ALL samples analyzed were characterized by high expression of the type I cytokine receptor CRLF2 caused by either immunoglobulin heavy locus (IgH@) translocations or by interstitial deletions creating chimeric transcripts P2RY8-CRLF2. In 3 of these 33 patients, a novel activating somatic mutation, F232C in CRLF2, was identified. Consistent with our previous research, mutations in R683 of JAK2 were identified in 10 specimens (19% of the patients) and, interestingly, all 10 had high CRLF2 expression. Cytokine receptor-like factor 2 (CRLF2) and mutated Janus kinase 2 (Jak2) cooperated in conferring cytokine-independent growth to BaF3 pro-B cells. Intriguingly, the gene expression signature of DS-ALL is enriched with DNA damage and BCL6 responsive genes, suggesting the possibility of B-cell lymphocytic genomic instability. Thus, DS confers increased risk for genetically highly diverse ALLs with frequent overexpression of CRLF2, associated with activating mutations in the receptor itself or in JAK2. Our data also suggest that the majority of DS children with ALL may benefit from therapy blocking the CRLF2/JAK2 pathways.

Abstract 271 Introduction. Although in AIEOP-ALL 2000 protocol, early MRD assessment allowed to better recognize high risk (HR) cases, still about 20% of children relapse, with more than half of relapses occurring in the largest subgroup of non-HR patients. Therefore, new prognostic markers are useful to adapt treatment. Recently, high-level expression of the CRLF2 gene was recognized to characterize a novel subgroup of BCP-ALL. Over-expression of CRLF2 can be the consequence of a translocation at the IGH@ locus, and/or an interstitial deletion, juxtaposing CRLF2 with the promoter of the P2RY8 gene. CRLF2 over-expression appeared to be the driver of STAT5 activation, either alone or in combination with gain-of-function mutations in Janus Kinases (JAKs), mainly JAK2 R683. Aim. In order to estimate the incidence of this abnormality, its prognostic value and the association with patients' biological features, we analyzed CRLF2 gene expression in 515 non-Down Syndrome BCP-ALL children, treated according to the ALL-AIEOP 2000 protocol. Methods. Ig/TcR PCR-MRD and RT-PCR screening for known chromosomal translocation were routinely performed. Gene expression profiling by Affymetrix HG U133 Plus 2.0 arrays; RQ-PCR for CRLF2 expression; PCR detection of P2RY8-CRLF2 fusion, FISH for CRLF2 translocations, High Resolution Melting (HRM) and sequencing for JAK2 mutations, gene Copy Number Analysis by Affymetrix Cyto 2.7M array were performed. CRLF2 expression was reported as a fold increase to the median of all values detected in the complete series. Results. Out of 515 samples, CRLF2 expression was 10 times higher than the overall median in 46 cases (8.9%), and 20 times more in 24 (4.7%). The P2RY8-CRLF2 rearrangement was detected in 15/168 cases overall (8.9%), in 13/43 tested as CRLF2-high(>10) samples (30%) and in 11/22 considering only the CRLF2-highest(>20) expression group (50%). JAK2 mutations were tested in the first 121 cases highly expressing CRLF2 (>3 times), and were observed in 4 cases (3.3%). All JAK2 mutated cases were over-expressing CRLF2 (19, 31, 43 and 58 times), and all were P2RY8-CRLF2 positive; 2 were R683G substitutions, and 2 cases presented 6 and 12 inserted nucleotides, respectively. Comparing CRLF2-highest(>20) and low expressing patients, none of the CRLF2-highest cases had recurrent chromosomal translocations (while 1 BCR/ABL and 3 TEL/AML1 positive cases were in the >10 times, and 20,000/μ l: 25% vs 35%) and a higher age at diagnosis (median age 67 months vs 56). CRLF2-highest and low had a similar incidence of prednisone poor response (9.2% vs 12.5%). Interestingly, among CRLF2-highest patients only 2 were in the PCR-MRD HR group. Remarkably, all 15 cases P2RY8-CRLF2-positive were non-HR by MRD (4 were SR, one of them in the CRLF2>20 group). Patients CRLF2-highest had a statistically significant higher 5-year Cumulative Incidence of Relapse (CIR) compared to CRLF2-low patients (37.5±9.8% vs. 17.7±1.7%; relapses=9/24 vs 93/491, p=0.00296). This difference was due to a higher CIR for P2RY8-CRLF2 positive cases; in fact, among CRLF2 >20 times cases, CIR for deleted cases was 60±15.5% compared to 16.6±10.7% for non-deleted (relapses=6/10 versus 3/14). The difference in CIR between CRLF2>20 times and CRLF2-low remains statistically different within the hyperdiploid subgroup (interestingly, none of the P2RY8-CRLF2 positive was hyperdiploid). Only 2 of the 4 JAK2 mutated (and CRLF2 deleted) cases relapsed. Conclusions. High levels of CRLF2 expression (>20 times the median) define a subgroup of childhood ALL (around 5%) without known high risk features, associated with a higher incidence of relapse if treated according to the AIEOP-ALL 2000 protocol. This effect is mainly related to the presence of the P2RY8-CRLF2 rearrangement. Once prospectively confirmed in large series, the assessment of CRLF2 status may serve as a new stratification marker. Disclosures: No relevant conflicts of interest to declare.

Abstract 1629 Poster Board I-655 Primary childhood acute lymphoblastic leukemia (ALL) samples are very difficult to culture in vitro and the currently available cell lines only poorly reflect the heterogeneous nature of the primary disease. Many groups therefore use mouse xenotransplantation models not only for in vivo testing but also as a means to amplify the number of leukemia cells to be used for various analysis. It remains unclear as to what extent the xenografted samples recapitulate their respective primary leukemia. It has been suggested for example that transplantation may result in the selection of a specific clone present only to a small amount in the primary diagnostic sample. We used a NOD/SCID xenotransplantation model and injected leukemia cells isolated from fresh primary diagnostic material of 4 pediatric ALL patients [2 pre-B-ALL, 1 pro-B-ALL (MLL/AF4}, 1 cortical T-ALL] intravenously into the lateral tail vein of unconditioned mice. As soon as the mice presented clinical signs of leukemia, leukemia cells were isolated from bone marrow and spleen. Isolated leukemia cells were retransplanted into secondary and tertiary recipients. RNA was isolated from diagnostic material and serial xenograft passages and gene expression profiles were obtained using a human whole genome array (Affymetrix U133 2.0). Simultaneously, immunophenotypic analysis via multicolor surface and cytoplasmatic staining by flow cytometry was performed for the diagnostic samples and respective serial xenograft passages. In an unsupervised clustering analysis the diagnostic sample of each patient clustered together with the 3 derived xenograft samples, although the 3 xenograft samples clustered stronger to each other than to their respective diagnostic sample. Comparison of the 4 diagnostic samples vs. all xenograft samples resulted in a gene list of 270 genes upregulated at diagnosis and 8 genes upregulated in the xenograft passages (Wilcoxon, p< .05). The high number of genes upregulated at diagnosis is most likely due to contamination of primary patient samples with normal peripheral blood and/or bone marrow cells as 15% of genes are attributed to myeloid cells, 7% to erythroid cells, 7% to lymphoid cells, 32% to bone marrow in general as well as to innate immunity, chemokines, immunoglobulins. The remaining genes can not be attributed to a specific hematopoetic cell lineage and are not known to be related to leukemia or cancer in general. Accordingly, there are no statistically significant differences between the primary, secondary and tertiary xenograft passages. The immunophenotype analysis are also in accordance with these findings, as the diagnostic blast population retains its immunophenotypic appearance during serial transplantation, whereas the contaminating CD45-positive non- leukemic cells disappear after the first xenograft passage. The few genes upregulated in xenograft samples compared to diagnosis are mainly involved in cell cycle regulation, protein translation and apoptosis resistance. Some of the identified genes have already been described in connection with cancer subtypes, their upregulation therefore being indicative of a high proliferative state in general and could argue towards a more aggressive potential of the engrafted leukemia cells but alternatively could also simply be due to the fact that the xenograft samples are pure leukemic blasts and are not contaminated with up to 15% of non-cycling healthy bone marrow cells as in the diagnostic samples. We conclude that the gene expression profiles generated from xenografted leukemias are very similar to those of their respective primary leukemia and moreover remain stable over serial retransplantation passages as we observed no statistically significant differences between the primary, secondary and tertiary xenografts. The differentially expressed genes between diagnosis and primary xenotransplant are most likely to be due to contaminating healthy cells in the diagnostic samples. Hence, the NOD/SCID-xenotransplantation model recapitulates the primary human leukemia in the mouse and is therefore an appropriate tool for in vivo and ex vivo studies of pediatric acute leukemia. Disclosures No relevant conflicts of interest to declare.

Abstract 744 Background: T-lineage Acute Lymphoblastic Leukemia (T-ALL) accounts for about 15% of pediatric ALL. Although the outcome of T-ALL has improved with current therapies, T-ALL pediatric patients remain at high risk for early relapse. It is therefore very important to identify new prognostic biomarkers and to develop new therapeutic approaches for these patients. More specific and less toxic therapies might be developed through the identification of molecular targets, such as aberrantly activated phosphoproteins, that offer the possibility to use specific kinase inhibitors. In this research we first explored the phosphoproteomic profile of pediatric T-ALL patients at diagnosis, through Reverse Phase Protein Arrays (RPPA), and we further investigated the origin and the functional significance of Protein Kinase C alpha (PKCα) downregulation in patients that are liable to relapse. Methods: We analyzed 98 pediatric T-ALL patients at diagnosis using RPPA. This innovative approach allows to profile phosphorylation modifications and to characterize the activation state of cellular protein networks. The whole proteome of each patient was immobilized on nitrocellulose coated glass slides and each slide was stained with one of 53 different antibodies included in the study. Protein expression/activation was compared between patients subgroups defined by clinical/molecular features through statistical analyses. Gene expression analysis was performed by means of HG-U133 Plus 2.0 array. Permission for this study was obtained following the tenets of the Declaration of Helsinki. A set of commercially available human T-ALL cell lines was studied by Western Blot and cells were treated with Ro-32-0432, a commercial PKC inhibitor. Results: The phosphoproteomic profile of T-ALL patients was compared between patients subgroups at diagnosis. Comparison between relapsed vs non relapsed patients samples revealed the downregulation of PKCα(S657) in relapsed patients (t test with BH multiplicity corrections p=0.02) and Relapse Free Survival analysis through Kaplan-Meier estimates showed that patients with low PKCα(S657) have a high probability to relapse. Notably, 54.5% of patients with low PKCα(S657) relapsed whereas no relapses were observed among patients with high PKCα activity. Multivariate analysis showed no correlation between PKCα(S657) levels and other parameters currently used for clinical characterization of T-ALL patients. Differences in PKCα(S657) between relapsed and non relapsed patients were confirmed by Western Blot in two independent sets of T-ALL patients. Analysis of PKCα total protein form revealed concordance between total protein expression and PKCα(S657) levels. Furthermore, gene expression analysis revealed a significant difference in PKCα gene expression levels, comparable with levels of protein expression among T-ALL patients studied by RPPA. Gene expression data were validated by Sybr Green Real-Time Quantitative PCR. Moreover, gene pathway analysis in a larger cohort of T-ALL pediatric patients, divided by high and low PKCα mRNA levels, revealed the phosphatidylinositol signalling pathway as the most important deregulated pathway. Of note, this pathway is involved in the production of DAG and Ca2+, the PKCα cofactors that play an essential role in mediating its translocation from the cytosol to the plasmatic membrane and subsequent activation. PKCα and PKCα(S657) were then analysed by Western Blot in five T leukemic cell lines (DND41, P12 ICHIKAWA, MOLT3, TALL1 and CEM). These cell lines were treated with a PKCα inhibitor and PKCα inhibition effects were studied by the means of MTT assay and Trypan Blue cell count. An increase in cell proliferation was observed in each treated cell line 9 hours after PKCα inhibition. Conclusions: PKCα is differentially expressed in T-ALL pediatric patients. Low levels of PKCα(S657) are associated with a higher probability to relapse and downregulation or absence of PCKα seems to be related to a more aggressive T-ALL phenotype. Ongoing studies are needed to better define its role as a new prognostic marker. Disclosures: No relevant conflicts of interest to declare.

Abstract 11 Children with Down's Syndrome (DS) have an increased risk for acute lymphoblastic leukemia (ALL). We have previously identified somatic mutations in R683 of JAK2 in 20% of DS-ALLs (Bercovich et al, Lancet 2008;372:1484). However the cooperating cytokine receptor and other molecular features have been obscured. Here we report that gene expression analysis reveals that DS-ALL is a highly heterogeneous disease not definable as unique ALL subtype. However, 62% (33 of 53) of the DS-ALL samples analyzed were characterized by aberrant expression of the type I cytokine receptor CRLF2 caused by two types of genomic aberrations. CRLF2 encodes one chain of the heterodimeric receptor of the TSLP cytokine involved in inflammation and lymphoid development and has been recently reported to be involved in translocations to the IgH locus or micro-deletions in 3% of ALLs in children without DS. We have confirmed its cell surface expression on primary DS-ALL cells by Flow cytometry analysis. Consisting with a possible role of CRLF2 as a cytokine receptor, all specimens with the R683 mutated JAK2 had high CRLF2 expression. We further show that CRLF2 and R683 mutated JAK2 cooperate in conferring cytokine independent growth of pro-B cells. Furthermore in 3 of 23 DS-ALL patients with wt. JAK2 we identified a novel activating somatic mutation in CRLF2 replacing the juxta-membrane F232 with Cystein. Mutated CRLF2 but not wild type CRLF2 conferred cytokine independent growth of BaF3 cells associated with phosphorylation of STAT5. Aberrant expression of CRLF2 was also associated with a younger age at diagnosis and worse prognosis. Intriguingly, bioinformatic analysis revealed that gene expression of DS-ALLs is enriched with DNA damage and BCL6 responsive genes suggesting the possibility that DS is associated with B-cell lymphocytic genomic instability leading to the genomic aberrations activating CRLF2. Thus DS confers increased risk for genetically highly diverse ALLs with frequent somatic anomaly in CRLF2, whose higher expression further selects for the activating mutations in itself or in JAK2 (figure). Our data also suggests that the majority of DS children with ALL may benefit from therapy blocking the CRLF2/JAK2 pathway. Disclosures: No relevant conflicts of interest to declare.

10042 Background: Acute lymphoblastic leukemia (ALL) is the most frequent malignant disease in childhood. Although advances in therapy have led to improved long term survival, relapse remains a major challenge. In a recent study we transplanted pediatric leukemia samples from newly diagnosed BCP-ALL patients into NOD/SCID mice. Time to leukemia (TTL) was analyzed for each patient sample as time from transplant to overt leukemia in the recipients. Patients whose leukemia cells engrafted rapidly showed a clearly inferior relapse free survival in contrast to patient samples with prolonged in vivo growth. Multivariate analysis showed an almost 45- fold increased risk for relapse in patients with short TTL. Methods: Gene expression profiles of ALL samples (N = 14) with short versus long TTL in the xenograft model were analyzed using a human whole genome array (Affymetrix U133 Plus 2.0) correlating gene expression values (relative expression) to the time from transplant to manifestation of leukemia in the NOD/SCID mice (TTL, in weeks) by quantitative traits analysis (QTA). Results: Among 5 genes significantly correlated (Spearman correlation, P < .0001) XIAP-associated factor 1 (XAF1) was found to be up-regulated in patients with long TTL. XAF1 abrogates the inhibitory effect of XIAP on caspase-3 thereby sensitizing for apoptosis. In accordance to this caspase-3 activation was also found to be up-regulated in patients with long TTL. Patient samples exhibiting a short time to overt leukemia in the xenotransplant model associated with poor relapse free survival showed down-regulated XAF1 and impaired caspase-3 activation leading to decreased apoptosis of the leukemia cells. Conclusions: Taken together, we used a novel approach directly correlating gene expression values to time from transplant to overt leukemia (TTL) identifying the apoptosis regulator XAF1 to be associated with poor outcome of patients. Small XIAP-inhibiting molecules can be used to substitute the lacking inhibitory effect of down-regulated XAF1 in these poor responding pediatric ALL patients. No significant financial relationships to disclose.

A number of clinical and biological prognostic factors are currently used for B-cell precursor Acute Lymphoblastic Leukemia (BCP-ALL) patient risk stratification into (standard risk) SR, (medium risk) MR and (high risk) HR treatment protocols. In spite of careful stratification patients stratified in the same risk class may respond differently to therapy suggestive of a need for additional prognostic markers in BCP-ALL. Here we analyzed gene expression profiles (GEP) of BCP-ALL patients at diagnosis in search for correlations with the diverse responses to therapy among these patients. We specifically searched for expression signatures that predict relapse or the absence of relapse in BCP-ALL. Among BCP-ALL 50% of patients are characterized by recurrent genomic aberrations that are strong indicators for risk stratification of which BCR/ABL and MLL rearrangements have a HR to relapse; TEL/AML1 and hyperdiploid have a standard risk to relapse. BCPALL patients without these aberrations form a heterogeneous risk group in which new prognostic markers will be very important to help stratify these patients. A GEP data set of 86 BCP-ALL patients (47 males; 39 females) without known aberrations was obtained using Affymetrix HG-U133Plus2.0 arrays that were analyzed using Partek and R software packages. Patients were treated according to AIEOP ALL 2000 protocols. Nine out of 86 patients relapsed of which 6 within 24 months; the minimum follow-up time was 24 months. For each probe set the variance among patients was computed and 10 % of the probe sets with the major variance were selected for further analysis. Unsupervised hierarchical cluster analysis separated patients in two groups: one group of 19 patients that were all in complete remission (CR) and a second group of 67 patients in which all patients that eventually relapsed (14%) and CR patients (86%) cluster together. The median follow-up time in the first group of patients was 40 months and the median patient age at diagnosis was 7 years. In the second group median follow-up was 29 months and median patient age at diagnosis was 5.5 years. The separation of patients into two groups was not related to risk group (SR, MR or HR), gender, age or response to induction therapy. A t-test (multiplicity corrections were adopted to control for false positives) was used to find probe sets that were differentially expressed between the two groups. Functional evaluation of the top 200 probe sets of this analysis established 3 major biological categories according to the GO database and revealed differential expression of genes involved in cell adhesion, signal transduction and immune response. We identified a subtype of patients with a gene expression signature that is correlated with excellent prognosis using GEP analysis on BCP-ALL patients that present no recurrent known genomic aberrancies. The signature of excellent prognosis involves about 20% of BCP-ALL patients without common genotypic aberrations and 8% of all BCP-ALL patients and is independent of currently applied prognostic risk factors.

Acute lymphoblastic leukemia (ALL) is the most frequent malignant disease in childhood. Advances in therapy, in particular stratification of patients to appropriate treatment according to risk factors improved long term survival attaining cure rates of up to 80 %. Despite the efforts achieved by the stratification strategies the majority of relapse patients are recruited from the low risk groups emphasizing the need for additional independent risk factors. In a recent study we transplanted primary leukemia samples obtained from pediatric patients with newly diagnosed B cell precursor ALL (BCP-ALL) into NOD/SCID mice. Time to leukemia (TTL) was analyzed for each patient sample transplanted from date of transplant to date of leukemia manifestation in the recipients. We demonstrated that patients whose leukemia cells engrafted rapidly leading to manifestation of the disease within 10 weeks (TTLshort) showed a clearly inferior relapse free survival in contrast to patient samples with prolonged in vivo growth (TTLlong). Interestingly, the same distinct difference in relapse free survival was observed in the low risk groups only. Multivariate analysis showed an almost 45- fold increased risk for relapse in TTLshort patients. In order to further characterize the biological properties of the leukemia cells in the two groups, gene expression profiles of samples with short versus long TTL in the xenograft model were analyzed using a human whole genome array (Affymetrix U133 Plus 2.0). Here, we used quantitative traits analysis (QTA) correlating gene expression values (relative expression) to the time from transplant to manifestation of leukemia in the NOD/SCID mice (TTL, in weeks). 14 different xenograft samples (TTLshort n= 7, mean TTL 8.14 weeks; TTLlong n= 7, mean TTL 19.71 weeks; T-test P = .03) isolated from leukemia bearing mice were investigated. All 14 patients included were stratified in low (standard or intermediate) risk groups. All patients were negative for TEL/AML1- fusion, BCR/ABL- fusion or MLL rearrangement. By QTA we identified 5 genes that were significantly correlated (Spearman correlation, for all 5 genes P < .0001) to time from transplant to leukemia manifestation in the recipient mice (TTL). Analysis of the 14 xenograft samples using the 5 genes identified showed clustering of all but one sample in one group. The 5 genes were than explored for their power to predict relapse in an independent cohort of pediatric ALL patients. For these patients expression profiles were analyzed in leukemia samples obtained at diagnosis. All patients in this cohort were also stratified in low risk groups and negative for the presence of TEL/AML1-, BCR/ABL- or MLL- fusion transcripts. 8 of the 38 patients encountered relapse, 5 at early and 3 at late time points (< or > 24 months after diagnosis). Clustering according to the 5 genes identified by QTA lead to separation of the patients into two groups. Interestingly, all relapsed patients except one clustered in the group associated with the signature characteristic for TTLshort. Most importantly, all 5 patients with early relapse gathered in this cluster. Taken together, we used a novel approach directly correlating gene expression values to the continuous variable time to leukemia (TTL) which might be reflected more accurately by this strategy than comparing two groups. We applied this gene signature characteristic for TTLshort, thereby identifying poor overall survival in a set of independent pediatric ALL patients and found clustering of all early relapse patients in one group. This new independent risk factor might identify patients with high risk for early relapse avoiding xenografting in the mouse model.