Search

The fusion gene MLL-AF4 defines a high-risk subtype of pro-B acute lymphoblastic leukaemia. However, relapse can be associated with a switch from acute lymphoblastic to acute myeloid leukaemia. Here we show that these myeloid relapses share oncogene fusion breakpoints with their matched lymphoid presentations and can originate in either early, multipotent progenitors or committed B-cell precursors. Lineage switching is linked to substantial changes in chromatin accessibility and rewiring of transcriptional programmes indicating that the execution and maintenance of lymphoid lineage differentiation is impaired. We show that this subversion is recurrently associated with the dysregulation of repressive chromatin modifiers, notably the nucleosome remodelling and deacetylation complex, NuRD. In addition to mutations, we show differential expression or alternative splicing of NuRD members and other genes is able to reprogram the B lymphoid into a myeloid gene regulatory network. Lineage switching in MLL-AF4 leukaemia is therefore driven and maintained by defunct epigenetic regulation.Statement of SignificanceWe demonstrate diverse cellular origins of lineage switched relapse within MLL-AF4 pro-B acute leukaemia. Irrespective of the developmental origin of relapse, dysregulation of NuRD and/or other epigenetic machinery underpins fundamental lineage reprogramming with profound implications for the increasing use of epitope directed therapies in this high-risk leukaemia.

Infant t(4;11) acute lymphoblastic leukemia (ALL) is associated with a high relapse rate with 4-year event-free survival (EFS) of only 36%. Relapse has been shown to be the major cause of death as 83% of relapsed infant t(4;11) ALL patients die within three years of diagnosis. Therefore, it is of utmost therapeutic interest to elucidate molecular mechanisms of relapse. Here we show that t(4;11) ALL cells upregulate the inducible T-cell costimulator ligand (ICOSLG) in an early growth response 3 (EGR3) dependent manner thereby promoting the development of regulatory T-cells (T regs). We propose that short EFS and high ICOSLG expression are causally linked, in that ICOSLG-mediated induction of T regs interferes with immune recognition of leukemia cells. According to that hypothesis, ICOSLG would not only be a novel and independent prognostic biomarker but a potential therapeutic target. We investigated the function of EGR3 in infant ALL since EGR3 has been described as an indirect target of the iroquois homeobox 1 (IRX1) transcription factor. For that purpose we created a stable sleeping beauty transposon-based SEM cell line expressing EGR3 in a Doxycycline-inducible manner. Gene expression and western blot analysis revealed strong upregulation of ICOSLG 48h after Doxycycline induction when compared to an empty vector control. Additionally, chromatin immunoprecipitation (ChIP) experiments depicted that EGR3 directly binds to the promoter of the ICOSLG gene. The expression of ICOSLG in mesenchymal stem cells has been shown to foster the induction of T regs and ICOSLG-mediated T reg expansion has been identified as a driver of acute myeloid leukemia and glioblastoma. In the bone marrow (BM) hematopoietic stem cells (HSC) colocalize with T regs which provide an immune privilege to the stem cell niche. Based on these observations, we hypothesize that ICOSLG expressing ALL cells could create an immune privileged niche appearing to be necessary for HSC maintenance and sanctuary from immune attack. This could create independence from the BM immune privilege enabling migration of the ALL cells. If so, this could contribute to minimal residual disease (MRD) formation after induction therapy and subsequently to higher probability of relapse. To evaluate our hypothesis, we cocultured the EGR3-SEM and control cells with primary human CD4 + and CD8 + T-cells. The T-lymphocytes were isolated from the peripheral blood mononuclear cells (PBMC) of healthy donors and stimulated with coated α-CD2, -CD3 and -CD28 beads. We observed 16% more CD4 +CD25 ++FOXP3 + T regs after 48h of coculture with EGR3-SEM cells compared to the control. The addition of a neutralizing monoclonal α-ICOSLG antibody to the coculture led to a reduction of the T reg frequency in EGR3-SEM coculture. Taken together, these results strongly suggest that ALL cells expressing EGR3 induce the formation of T regs via ICOSLG expression. To confirm our results with patient-derived material, we investigated the gene expression of 50 infant t(4;11) pro-B phenotypic ALL patients. Pearson correlation testing confirmed that ICOSLG expression strongly correlates with EGR3 and IRX1 expression. Furthermore, we were able to classify the patients considering their ICOSLG expression level into an ICOSLG-high (ICOSLG-hi) and an ICOSLG-low (ICOSLG-lo) group. Outcome data for a 5-year follow-up were available for 35 of 50 patients, n=5 ICOSLG-hi and n=30 ICOSLG-lo. 100% (5/5) of the ICOSLG-hi patients failed within 17 months of diagnosis whereas 53% (16/30) of the ICOSLG-lo patients failed within 56 months of diagnosis. The remaining 14 ICOSLG-lo patients were alive after 60 months. These data underscore the role of overexpressed ICOSLG in relapse development. However, a verification of these findings in a larger cohort is needed. In conclusion, our study identifies ICOSLG as a promising prognostic marker and novel therapeutic target in infant t(4;11) ALL. Furthermore, our findings implicate the interaction between T-cells and leukemia stem cells as contributory to disease progression. Disclosures Cario: Novartis: Other: Lecture Fee.

Chromosomal rearrangements of the KMT2A gene are associated with acute leukemias and myelodysplastic syndromes. The large number of known KMT2A fusions (>100) renders a precise diagnosis a demanding task. More than 50% of all KMT2A partner genes have been analyzed at the DCAL, including the novel partner genes BCAS4, FAM13A, RANBP3, and STK4. Even though all KMT2A rearrangements are associated with high-risk acute leukemia, the outcome (poor or very poor) is influenced by the partner gene. So far, we have analyzed more than 3,200 patients positive for a KMT2A rearrangement. The breakpoints of these cases are located mainly in the major breakpoint cluster region (bcr1) and to a small extent in the recently described minor bcr (bcr2). A small number of breakpoints were also found outside of these two bcrs. Most of these patients were analyzed by long distance inverse (LDI)- or multiplex-PCR which only cover bcr1. More recently, we used targeted KMT2A-NGS with whole gene coverage in over 450 patients, which was initially applied selectively in patients negative by LDI- and multiplex-PCR and then used more widely. Within the KMT2A-NGS group, 410 patients had bcr1 breakpoints mainly between the KMT2A exons 7 and 13, while 46 patients bcr2 breakpoints mainly between exons 20 and 24. Of note, five patients had their breakpoint outside of these two bcrs: three of them within intron 2 and no functional KMT2A rearrangement; the other two within intron 35 and intron 36, fusing almost the whole KMT2A gene in frame to the respective partner genes ARHGEF12 and MLLT4. These two breakpoints may define a third and rare bcr (bcr3), although further cases are needed to support this hypothesis. Interestingly, 70 patients displayed a 3'-KMT2A deletion, indicating that the number of terminal deletions is higher than described previously. Two patients had a 5'-KMT2A deletion. All deletions started or ended in bcr1 and bcr2. We also observed a striking difference in the distribution of partner genes between bcr1 and bcr2. The most frequent translocation partners fused to bcr1 sites are transcription factors, while the partner genes linked to bcr2 sites generally code for cytosolic proteins. In bcr1, the 4 most frequent partner genes AFF1, MLLT3, MLLT1, and MLLT10, found in 80% of cases, all code for transcription factors that are part of the super elongation complex (SEC). These fusions therefore all lead to disruption of the hematopoietic lineage commitment. In contrast in bcr2, 3 partner genes USP2, MLLT4, and USP8 account for 85% of the cases. USP2 and USP8 are ubiquitin specific peptidases involved in cell signaling and exclusively fused to bcr2 in KMT2A. While MLLT4 is found as a partner in bcr1, bcr2 and bcr3 fusions; unlike other recurrent KMT2A partners linked to bcr1, it is not a transcription factor and it exerts oncogenic potential via dimerization like other cytosolic partners. We hypothesize that the oncogenic properties of USP2 and USP8 are dependent on dimerization like MLLT4 and that the most frequent fusions involving at different bcrs favor different oncogenic mechanisms: bcr1 transactivation and bcr2 dimerization. Further studies are needed to explain why USP2 and USP8 are exclusively associated with bcr2, and why the most frequent partner genes AFF1 and MLLT3 of the bcr1 are less frequent in bcr2. In conclusion, targeted NGS combined with bioinformatic analysis has expanded our knowledge of the KMT2A recombinome to include more fusion partners and has generated new hypotheses for future research on oncogenic mechanisms. Disclosures No relevant conflicts of interest to declare.

Approximately 25% of childhood acute lymphoblastic leukemias carry the ETV6/RUNX1 fusion gene. Despite their excellent initial treatment response, up to 20% of patients relapse. To gain insight into the relapse mechanisms, we analyzed single nucleotide polymorphism arrays for DNA copy number aberrations (CNAs) in 18 matched diagnosis and relapse leukemias. CNAs were more abundant at relapse than at diagnosis (mean 12.5 vs 7.5 per case; P = .01) with 5.3 shared on average. Their patterns revealed a direct clonal relationship with exclusively new aberrations at relapse in only 21.4%, whereas 78.6% shared a common ancestor and subsequently acquired distinct CNA. Moreover, we identified recurrent, mainly nonoverlapping deletions associated with glucocorticoid-mediated apoptosis targeting the Bcl2 modifying factor (BMF) (n = 3), glucocorticoid receptor NR3C1 (n = 4), and components of the mismatch repair pathways (n = 3). Fluorescence in situ hybridization screening of additional 24 relapsed and 72 nonrelapsed ETV6/RUNX1-positive cases demonstrated that BMF deletions were significantly more common in relapse cases (16.6% vs 2.8%; P = .02). Unlike BMF deletions, which were always already present at diagnosis, NR3C1 and mismatch repair aberrations prevailed at relapse. They were all associated with leukemias, which poorly responded to treatment. These findings implicate glucocorticoid-associated drug resistance in ETV6/RUNX1-positive relapse pathogenesis and therefore might help to guide future therapies.

Chromosomal rearrangements of the MLL gene are responsible for 5-10% of all acute leukemias, biphenotypic leukemias and myelodysplastic syndromes. The large number of known MLL fusions (>80) renders a precise diagnosis a demanding task. Even though all MLL rearrangements are associated with high-risk acute leukemia, the outcome (poor or very poor) is influenced by the partner gene. The applied diagnostic methods (LDI-PCR and multiplex PCR) allows the identification of MLL fusion genes at the nucleotide level, providing important information on the genetics of leukemia patients, and patient-specific biomarkers. These biomarkers are used for monitoring of minimal residual disease in acute leukemia patients during and after therapy. Thus, the identification of MLL gene fusions is necessary for rapid clinical decisions to determine the best therapy regimen. We have developed a customized NGS panel for MLL diagnostics to utilize state of the art technology at DCAL. With this new tool, the whole MLL gene is analyzed in contrast to the LDI-PCR where only the main MLL breakpoint cluster region (BCR-1) is covered. The first results of the NGS analysis of 84 patients identified MLL breakpoints located outside the main BCR-1 of MLL. Furthermore, a novel MLL partner gene USP2 was identified in 16 patients. All MLL-USP2 positive patients had a breakpoint located outside BCR-1 and within a newly defined breakpoint cluster region BCR-2. The BCR-2 site was also used in 2 other patients with MLL-AFF1 and one patient with MLL-MLLT3. These findings reveal USP2 as a new entity for MLL rearrangements affecting indifferently children aged 3 months to 10 years old (mean 30 months) with no gender bias (M/F=1.3). Interestingly, only 5/16 affected children were below 1 year of age at diagnosis and thus treated according to the Interfant trial. Clinical presentation as well as outcome associated with this new entity deserves further investigation to define whether those patients should be allocated, as other MLL-rearranged ones, in high-risk treatment groups. More MLL patients should also be analyzed to get a better idea of the frequency of breakpoints within BCR-2, especially the frequency of MLL-USP2 fusions. Indeed, standard FISH analysis and CGH array do not permit reliable detection of this fusion, explaining why they remained undetected so far. The biology of this novel MLL rearrangement also deserves further investigation, considering that USP2 is the only MLL partner fused exclusively to BCR-2. Disclosures No relevant conflicts of interest to declare.

Rearrangements of the mixed lineage leukemia gene (MLL, re-named KMT2A) result in aggressive leukemia. Current risk stratification of MLL-rearranged (MLL-r) leukemia is directed by the fusion partner gene and, increasingly, by minimal residual disease (MRD) assessment after induction therapy. The clinical significance of quantifying fusion transcript levels in leukemia patients is firmly established in chronic myeloid leukemia and acute promyelocytic leukemia but is less well studied in MLL-r patients. Real-time quantitative PCR (RQ-PCR) is the standardized assay for molecular MRD monitoring in patients with MLL-rearranged leukemia. However, this method is less precise when few leukemic cells are present, thus limiting its application for highly sensitive MRD monitoring. Droplet digital PCR (ddPCR) allows for absolute quantification of fusion transcripts when multiple copies of fusion transcripts are present per cell. Therefore, we aimed to evaluate whether determining MLL fusion transcript levels by ddPCR could improve the sensitivity of MRD monitoring in MLL-r leukemia. A total of 44 diagnostic and follow-up samples obtained from paediatric MLL-r leukemia patients (26 ALL, 18 AML) were subjected to targeted next-generation sequencing to obtain patient-specific fusion sequences. MLL fusion transcripts were quantified by ddPCR in a total of 17 samples obtained from 4 paediatric AML patients with MLL fusions involving MLLT3 (n = 3) and MLLT10 (n = 1). Fusion-specific probe assays were designed from each of the patient specific fusion sequences for MRD assessment by ddPCR. To determine the detection limit of this method in quantifying MLL fusion transcripts, two MLL-r AML cell lines (MV4-11 and THP-1), and one MLL-wt cell line (Kasumi-1) were used. MLL fusion transcript level of detection of ddPCR was determined by serially diluting MLL-r cDNA into MLL-wt cDNA (Kasumi-1). Using 20ng of MLL-r cDNA in 200ng diluent as the highest concentration, a 10-fold dilution series was performed to make concentrations ranging from 10−2 to 10−7. Each ddPCR reaction mixture contained 11ul of cDNA mix as template with 1X Supermix no dUTP (Bio-Rad), 500 nM of both F/R primers and 250 nM of 5'-FAM labelled probe (IDT). Droplets were generated using a QX200 Droplet Generator (Bio-Rad). A general thermal cycler protocol with annealing at 61°C for 1 minute was performed and positive fluorescence droplets were read using QX200 Droplet Reader (Bio-Rad). MRD of patient samples, derived from ddPCR, was then compared to MRD derived from DNA-based RQ-PCR, following the guidelines established by the EuroMRD group. Our ddPCR method showed high reliability and sensitivity, with the detection limit determined to be 10-5 for a cell line with low MLL fusion transcript expression (THP-1), and 10-6 for a cell line with high MLL fusion transcript expression (MV4-11). Comparison of results obtained by RQ- PCR and ddPCR in a total of 17 diagnostic and follow-up samples from 4 AML patients showed excellent/good concordance between methods for 13 samples with moderate MRD levels. The 4 samples with low levels of MRD (10-4 to 10-5) below the quantitative range as defined by EuroMRD for RQ-PCR were all detectable by ddPCR, highlighting that ddPCR could provide robust and highly sensitive MRD assays compared to the standardized RQ-PCR assays. In conclusion, ddPCR is a promising technique that can reproducibly and reliably quantify MLL-r transcripts for MRD monitoring of MLL-r leukemia. Highly sensitive and robust molecular MRD monitoring by ddPCR holds promise for improving response-based therapeutic stratification and prediction of molecular relapse before overt hematological relapse. Disclosures No relevant conflicts of interest to declare.

Introduction Acute lymphoblastic leukaemia (ALL) in infants has poor overall survival despite being characterized by very few genetic aberrations per case. The most common genetic change, present in over 75% of cases, is the rearrangement of the mixed lineage leukaemia (MLL/KMT2A) gene (MLL-R) that also occurs in AML and mixed phenotype acute leukaemia (MPAL). Because infant ALL is rare and distinctive, most Australian and New Zealand patients are enrolled on specific clinical trials. In earlier infant ALL trials, MRD was not used for risk stratification firstly because of the difficulty of finding sensitive markers for specific immunoglobulin and T-cell receptor (Ig/TCR) gene rearrangements and the secondly because, in infant ALL, Ig/TCR markers are often present in sub-clones that can be lost at relapse due to clonal selection. MRD testing is performed in the current Interfant 06 trial preferentially using markers based on the genomic breakpoint sequences of MLL gene rearrangements. The treatment of infant ALL remains very challenging with relatively poor survival rates attributable to both toxicity and relapse. The objectives of this study were therefore to analyse MRD data for Interfant 06 patients enrolled at ANZCHOG centres and to perform a pilot experiment evaluating gene expression for key genes in infant ALL samples on a microfluidics platform, as a basis for identifying potential targeted therapies. Methods MRD was measured in bone marrow DNA from Interfant 06 patients enrolled since 2006, using sensitive Real-time Quantitative PCR (PCR-MRD) patient-specific assays to detect either MLL gene rearrangements (in 17) and/or conventional immunoglobulin and T-cell receptor (Ig/TCR) markers (21). Gene expression levels for 90 genes important in childhood cancers were measured by Taqman-based microfluidic assays in duplicate using cDNA from 2 micrograms of total RNA from 10 infant ALL samples (5 MLL, 5 non-MLL) at diagnosis (6) or relapse (4). Results Patient-specific PCR-MRD tests were developed for 27 out of 28 Australian and New Zealand infant ALL patients. 17/18 MLL-R ALL patients had MLL-R assays and 21/28 patients had Ig/TCR MRD tests, with only 1 (non-MLL) patient having no MRD markers. There was a wide range of MRD responses to induction therapy (Figure 1). Bone marrow MRD at the end of induction was high (>1x10-3) in 44% of MLL-R ALL infants compared to 25% of non-MLL ALL infants and 15% of older children enrolled on ANZCHOG ALL8. In 8/13 MLL-R patients who had both types of marker, MRD levels were higher when measured by their MLL-R marker than by their Ig/TCR marker. In a set of 90 genes selected for expression analysis, higher levels were found for 17 genes in 2 or more of the 10 infant ALL samples evaluated. These more highly expressed genes included potential or known drug targets BCL2, ERBB2, ERBB4, ILRA2, CSF1R and PARP1. Conclusions The quantitation of MRD based on MLL rearrangements in ALL is effective and can also be used to monitor response to therapy in infant ALL as well as MLL-R cases of AML and MPAL. The combined application of MLL-R and Ig/TCR markers allowed 97% of infant ALL patients to be MRD monitored with a sensitive marker. In most patients with both type of MRD marker, higher levels of MRD were detected in end of induction samples using the MLL versus Ig/TCR tests. One interpretation is that Ig/TCR genes are rearranged after the MLL rearrangement in ALL sub-clones that are both more mature and more chemo-sensitive. This finding also confirms the current consensus that disease-related MLL-R markers provide better risk assessment than Ig/TCR markers. Our Fluidigm analysis has shown that quantitative measurement of multiple gene expressions is feasible on small RNA samples and can be used to rapidly screen for specific expression of genes coding for drug targets in ALL patients. Support: NHMRC Australia APP1057746, Sporting Chance Cancer Foundation. Figure 1. Comparison of MRD response to induction therapy in MLL-R infant ALL compared with non-MLL infant ALL (Interfant 06) and older children (ANZCHOG ALL8). Figure 1. Comparison of MRD response to induction therapy in MLL-R infant ALL compared with non-MLL infant ALL (Interfant 06) and older children (ANZCHOG ALL8). Disclosures No relevant conflicts of interest to declare.

MLL translocations in adult B-cell precursor (BCP) acute lymphoblastic leukemia (ALL) are largely restricted to the immature CD10− immunophenotypes. MLL-AF4 is known to be the most frequent fusion transcript, but the exact frequencies of MLL aberrations in CD10− adult BCP-ALL are unknown. We present a genetic characterization of 184 BCR-ABL− CD10− adult ALL cases (156 cyIg−, 28 cyIg+) diagnosed between 2001 and 2007 at the central diagnostic laboratory of the GMALL study group. Patient samples were investigated by RT-PCR for MLL-AF4, MLL-ENL, and MLL-AF9 and by long-distance inverse polymerase chain reaction, thus also allowing the identification of unknown MLL fusion partners at the genomic level. MLL-AF4 was detected in 101 (54.9%) and MLL-ENL in 11 (6.0%) cases. In addition, rare MLL fusion genes were found: 2 MLL-TET1 cases, not previously reported in ALL, 1 MLL-AF9, 1 MLL-PTD, a novel MLL-ACTN4, and an MLL-11q23 fusion. Chromosomal breakpoints were determined in all 118 positive cases, revealing 2 major breakpoint cluster regions in the MLL gene. Characteristic features of MLL+ patients were significantly lower CD10 expression, expression of the NG2 antigen, a higher white blood count at diagnosis, and female sex. Proposals are made for diagnostic assessment. The clinical studies are registered at http://www.clinicaltrials.gov as NCT00199056 and NCT00198991.

Clonal expansion of mixed-lineage leukemia (MLL) rearrangements is highly indicative for acute leukemia.[1][1]-[5][2] Here, we report the unusual finding of an MLL rearrangement in the absence of leukemia. A 5-year-old girl was diagnosed with acute myeloid leukemia (AML) characterized by a variant

Background Infant acute leukemia is characterized by high incidence of MLL gene rearrangements. Purpose To evaluate the distribution of MLL genomic DNA breakpoints and their relation to several diagnostic parameters among infant acute leukemia. Methods 72 infants with MLL-rearranged acute lymphoblastic leukemia (ALL) (n=52), acute myeloid leukemia (AML) (n=19) and mixed phenotype acute leukemia (n=1) were included in this study based on the availability of DNA material at diagnosis. In the observed group there were 28 boys (39%) and 44 girls (61%) with median age of 4.9 mo (range 0.03-11.9). Genomic DNA breakpoint detection in MLL gene and translocation partner genes (TPG) was performed by long-distance inverse PCR (LDI-PCR). Exon-intron numbering of MLL gene was done according to I. Nilson et al, 1996. Results Majority of ALL cases (n=28; 54%) was characterized by presence of MLL-AF4 fusion gene (FG), less frequently MLL-MLLT1 (n=12; 23%), MLL-MLLT3 (n=7; 13%) and others were found (Table 1). The most common breakpoint location within MLL gene in ALL patients was intron 11, detected in 25 cases (48%). The highest variability of MLL breakpoints was found in MLL-AF4-positive patients: only 11 of 28 (39%) had breakpoints in intron 11. The most stable pattern of MLL genomic DNA breakpoints was observed in MLL-MLLT1-positive patients: 8 of 12 (67%) had breakpoints in intron 11. In AML patients two the most prevalent FGs were MLL-MLLT3 (n=7, 37%) and MLL-MLLT10 (n=5, 26%). The remaining ones are listed in Table 1. The most frequent breakpoints location was intron 8 (8 out of 19, 42%). The most stable pattern was revealed for MLL-MLLT10 FG: MLL breakpoints in 4 of 5 (80%) cases were found in intron 9 (Table 1). ALL patients who had breakpoints in intron 11 were significantly younger (median 3.0 mo, range 0.03-11.6) than all others (median 5.6 mo, range 0.7-11.9) (p=0.025) and than patients with MLL breakpoints in intron 9 (median 6.6 mo, range 3.1-11.9) (p=0.017). For AML cases we did not find any relation between age and breakpoints locations. Distribution of MLL DNA breakpoints was similar in boys and girls and did not depend on type of TPG. Genetic recombinations involving MLL gene predominantly resulted in reciprocal chromosomal translocations (n=62; 86%). Beside them, 6 (11%) insertions were identified in all MLL-MLLT10-positive cases and MLL-SEPT6-positive one. In 11 (15%) patients we found breakpoints within the regions located from 0.7 Kb to 25.4 Kb 3' of the first exon of TPGs (MLLT1 n=9; EPS15 n=1; MYO1F n=1), however fusion transcripts at cDNA level were identified and sequenced in all these cases, indicating a spliced fusion mechanism. 3-way translocations were found in 5 patients and in 1 case we found combination of insertion with interstitial deletion of chromosome 11. The list of reciprocal genes involved in these 6 cases was as follows: CEP164, DNAH6, DCPA1, MCL1 as well as non-coding regions of 2q21.2 and 2p21. We also analyzed breakpoints in TPGs. Except above mentioned spliced fusion cases, the remaining 3 breakpoints in MLLT1 as well as 3 of 4 breakpoints in EPS15 and all breakpoints in MLLT11 were within intron 1 of corresponding genes. In AF4 the major breakpoint region included intron 3 (n=19), intron 4 (n=6) and intron 5 (n=2). We also revealed 2 rare breakpoints in intron 6 and 10. In MLLT3 the most frequent breakpoint location was intron 5 (n=12), additionally 2 cases in intron 5 were identified. In MLLT10 two separate breakpoint locations were found: intron 3 (n=1) and intron 8 (n=3) in combination with intron 9 (n=1). We estimated prognostic significance of MLL breakpoint locations in 31 cases of infant ALL treated by MLL-Baby protocol. 3-year cumulative incidence of relapse was remarkably higher in patients with breakpoints in intron 11 (n=18) in comparison to patients with breakpoint localized from intron 7 to exon 11, inclusively (n=13) (0.85±0.01 and 0.57±0.02, respectively), although difference between these two groups did not achieve statistical significance (p=0.261). Median follow-up time in the observed group was 30 months (range 6–42). Conclusion In the current study we estimated clinical and prognostic significance of MLL and TPG genomic DNA breakpoints in infant acute leukemia. Our data provide additional information of molecular genetic features of MLL-rearranged infant acute leukemia. Disclosures: No relevant conflicts of interest to declare.

The regulation of hematopoietic lineage fate and commitment is fundamental to normal and malignant hematopoiesis. Switches between lymphoid and myeloid lineages in leukemia are rare and associated with poor clinical outcome, but potentially very informative regarding the regulation of hematopoietic lineage commitment. In contrast to therapy-related acute leukemia (AL) after a first primary leukemia, lineage-switch ALs arise from a common pre-leukemic or leukemic clone and share a founder mutation, most often rearrangement of MLL at 11q23. The majority of switches are from acute lymphoblastic leukemia (ALL) to acute myeloid leukemia (AML); however, conversions from myeloid to lymphoid and even oscillations between the two lineages have been observed, although the molecular mechanisms underlying lineage switch have not yet been identified. Here we describe a male patient who presented at 9 months of age with a t(4;11)-positive B-ALL and was subsequently treated according to the Interfant06 protocol. He achieved complete remission, but relapsed at the age of 4 years with a t(4;11)-positive AML. He underwent allogeneic BM transplantation and has remained in remission 13 months. Sanger sequencing revealed identical translocation breakpoints in the ALL and AML samples, demonstrating a lymphoid to myeloid lineage switch with a common pre-leukemic or leukaemic cell of origin for both ALs. Interestingly, whereas the AML shows no V(D)J rearrangements, we found incomplete rearrangements in the ALL cells indicating a ProB cell origin. In line with this observation, B-ALL cells expressed 6-fold and 120-fold higher levels of PAX5 and EBF1, respectively, compared to AML blast cells. Microsatellite instability measurements argued against a strong therapy-associated impairment of DNA mismatch repair in the AML. The translocation t(4;11) is the most frequently found chromosomal rearrangement in infant leukaemia and is almost exclusively associated with ALL at presentation, suggesting a strong instructive potential towards the lymphoid cell fate. However, the occurrence of lineage switch in t(4;11) AL demonstrates that this instruction can be overcome by as yet unknown mechanisms. Exome sequencing identified 16 and 98 novel somatic variants in the diagnostic ALL sample (0.23 mutations per Mb) and AML (1.4 mutations per Mb), respectively, of which 10 were shared. Of the total novel somatic mutations, there were 1 and 12 non-synonymous alterations in the B-ALL and AML, respectively, of which one was shared. RNA sequencing confirmed that all 12 genes with non-synonymous mutations were expressed in AML blast cells, and all belonged to the top 25% of expressed genes in both the AML and B-ALL. Genes carrying non-synonymous somatic AML-specific mutations include CHD4 (12p13, NuRD helicase, chromatin maintenance, DNA repair, lineage fidelity, part of MLL complex), NCOA2 (8q13, also known at TIF2 and part of the MOZ/TIF2 fusion gene, cofactor of nuclear receptors including VDR and NR3C1; control of myeloid differentiation, putative tumour suppressor), CEP164 (11q23, centrosome protein involved in microtubule organization, DNA damage response and chromosome segregation) and PPP1R7 (2q37, regulatory subunit of protein phosphatase 1, control of mitosis, regulator of AURKB). All amino acid residues predicted to be mutated are conserved between several species. Each of the identified mutations is located in functionally relevant regions and may, thus, interfere with protein function. Notably, exome and RNAseq showed that all 12 of the AML-specific non-synonymous mutations were found in at least 40% of the reads covering the corresponding genomic positions (the sample analysed constituted 80% blast material), thus suggesting heterozygosity for each mutation and that all mutations are common to the major leukemic clone. Taken together, these data suggest that the B-ALL and AML share a common ancestral pre-leukemic or leukemic cell of origin. Whilst the B-ALL revealed few novel somatic mutations, the change in lineage is associated with the acquisition of a substantial number of novel mutations indicating significant clonal evolution preceding the emergence of myeloid neoplasia. These data identify candidate mutations/genes which may overcome lineage instruction by a leukemic master regulator such as MLL/AF4 and which may therefore play an essential role in the control of hematopoietic lineage fate. Disclosures: No relevant conflicts of interest to declare.

Abstract 2537 Background. MLL gene rearrangements are the most common genetic events in infant leukemia. Up to date more than 100 various MLL rearrangements were described. Purpose. To evaluate the distribution of MLL rearrangements among infants (aged from 1 to 365 days) with both acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). Methods. 174 infants (117 ALL and 57 AML cases) were included in the current study. 11q23/MLL rearrangements were detected by chromosome banding analysis (CBA), fluorescence in-situ hybridization (FISH) and reverse-transcriptase PCR (RT-PCR). CBA was done according to standard procedure. FISH analysis using LSI MLL Dual Color, Break Apart Rearrangement Probe (Abbott Molecular, USA) was performed on at least 200 interphase nuclei and on all available metaphases. RT-PCR was performed as previously described (A. Borkhardt et al.,1994, N. Palisgaard et al., 1998, J. van Dongen et al., 1999). In 39 cases genomic DNA breakpoint was detected in MLL and translocation partner genes by long-distance inverse PCR (LDI-PCR). Exon-intron numbering of MLL gene was done according to I. Nilson et al, 1996. Results. 11q23/MLL rearrangements were revealed in 74 ALL patients (63.2%). Among this group MLL-AF4 was detected in the majority of cases (53.5%), less frequently were found MLL-MLLT1, MLL-MLLT3, MLL-MLLT10 and others (fig. 1a). Children with ALL under 6 months of age had significantly higher incidence of MLL rearrangements in comparison with older infants (84.0% vs. 47.8%, p Conclusion. In the current study we precisely characterized large cohort of MLL-rearranged infant acute leukemia patients. Combination of all available techniques, including cytogenetics, FISH, RT-PCR and LDI-PCR can lead to detailed verification of every single MLL rearrangement. Disclosures: No relevant conflicts of interest to declare.

Abstract 2498 Background. MLL gene rearrangements are associated with unfavorable outcome in infant acute lymphoblastic leukemia (ALL) and have intermediate prognosis in infant acute myeloid leukemia (AML). Application of fluorescence in-situ hybridization (FISH) allows detecting not only conventional MLL rearrangements, but also concurrent 3'-deletion of MLL gene. However, detailed characteristics of infant leukemia carrying 3' MLL deletion remain unclear. Aim. To investigate molecular genetic features of MLL-rearranged infant acute leukemia with concurrent 3' MLL deletion. Methods. 64 patients (27 boys and 37 girls) aged from 1 day to 11 months (median 6.6 months) including 44 ALL patients, 18 AML patients, 1 patient with acute bilineage leukemia and 1 patient with acute undifferentiated leukemia were enrolled in the current study. Chromosome banding analysis was done according to standard procedure. FISH analysis using LSI MLL Dual Color, Break Apart Rearrangement Probe (Abbott Molecular, USA) was performed on at least 200 interphase nuclei and on all available metaphases. Presence of MLL rearrangements was detected by FISH, reverse-transcriptase PCR. In 29 cases long-distance inverse PCR was additionally performed. In case of MLL rearrangement presence standard FISH pattern was defined as simultaneous detection of 3 different fluorescent signals: 1 fused (orange) signal, 1 green signal derived from 3' part of MLL gene, 1 red signal from 5' end of MLL (1F1G1R). MLL rearrangements with concurrent 3' MLL deletion led to 1F1R FISH pattern formation due to lack of green signal. Results. FISH revealed MLL rearrangements in 73% of ALL cases that was higher than frequency of 11q23 translocations detected by conventional cytogenetics — 55%. In MLL-positive cases we found 38 patients (81%) with standard FISH pattern, 7 ones (15%) with concurrent 3'-deletion of MLL gene and 2 (4%) with complex MLL rearrangements. Among patients with 3' MLL deletions there were 1 case with 5' MLL duplication (1F2R) and 1 case with 5' MLL triplication (1F3R). Frequency of 3'-deletions were similar in ALL and AML patients (13% and 15%, respectively). We did not find more than one FISH pattern in bone marrow blast cells of each patient with 3' MLL deletion. In this cohort of patients all blast cells carried concurrent 3'-deletion of MLL gene. Moreover, percentage of blast cells carrying MLL rearrangements did not differ significantly between patients with standard FISH pattern (median 97%, range 22–100%) and 3'-deletion (median 83%, range 13–99%) (p=0.206). 3'-deletion of MLL was not associated with breakpoint position in MLL gene and type of translocation partner gene. MLL translocation partner genes detected in patients with 3' deletions were as follows AF4(n=2), MLLT3(n= 3), MLLT10(n=2). None of the patients with 3'-deletions had reciprocal fusion gene. Initial patients' characteristics (age, sex, WBC count, immunophenotype, CNS-status, type of MLL partner gene) and treatment response parameters (day 8 peripheral blood blast cell count, day 15 bone marrow status, day 36 remission achievement, minimal residual disease status at time point 4) did not differ significantly between 2 groups. Although cumulative incidence of relapse was lower in patients with 3'-deletion as compared to patients with standard FISH pattern (0.31±0.04 and 0.55±0.01, respectively), difference between these two groups was not statistically significant (p=0.359). Conclusion. In our work we characterized rare subgroup of infant MLL-rearranged acute leukemia carrying concurrent 3' MLL deletion. Our data provide additional information of molecular genetic features of acute leukemia in children younger than one year. Disclosures: No relevant conflicts of interest to declare.

Abstract 1351 Chromosomal rearrangements of the MLL gene are associated with pediatric and adult de novo as well as therapy related acute myeloid leukemias, acute lymphoid leukemias, biphenotypic leukemias, and myelodysplastic syndromes. So far more than 70 MLL fusion partner genes have been characterized at the molecular level involving nearly all chromosomes. Though 11q23 rearrangements are associated with high-risk leukemias the clinical outcome of MLL rearrangements depends highly on the specific fusion partner involved. Nearly 40% of these partner genes have been identified at the Diagnostic Center of Acute Leukemia (DCAL) including the novel partner genes ABI2, PDS5A and TOP3A by analyzing more than 1400 MLL rearrangements. These rearrangements from 27 different countries include 52 different partner genes. This overview indicates all the specific introns of the translocation partner genes (TPGs) found to be involved in MLL translocations, their chromosomal locations and the type of genetic abnormality that is leading to the MLL fusion. More than 20% of all MLL rearrangements are complex ones. Within these complex rearrangements, the 5' part of the MLL gene is generally fused in frame to one of the most frequent partner genes. But the 3' part of the MLL gene is fused in many cases to a novel so-called “reciprocal MLL partner gene” like DENND4A or LRRTM4. To date more than 90 reciprocal MLL partner genes have been identified. In addition we present a recently analysed 4-way translocation where all four breakpoints could be identified by systematic breakpoint analysis using LDI-PCR. Also three novel “spliced MLL fusions”, a mechanism to generate functional chimeric fusion transcripts, involving MLLT4 (AF6), MYO1F, and CT45A2 have been identified at the DCAL. With these results, the number of genes involved in „spliced MLL fusions“ has increased from eight to eleven. Moreover we present the current breakpoint cluster region (bcr) for the 14 most frequent partner genes, namely AFF1 (AF4), MLLT3 (AF9), MLLT1 (ENL), MLLT10 (AF10), MLLT4 (AF6), ELL, EPS15, MLLT6 (AF17), SEPT6 (AF17), MLLT11 (AF1Q), SEPT9, AFF3 (LAF4), TET1, SEPT5 (PNUTL) and partial tandem duplications. For six patients, no partner gene could be identified at the molecular level and for 5 patients the identified fusion is out of frame. Unfortunetely all attemps to identify functional chimeric fusion transcripts by RACE and RT-PCR failed. These unusual MLL rearrangements probably represent a subclass of MLL abnormalities which have per se no or only a weak ability to transform hematopoeitic cells and are indentified only in the context of other hematopoeitic malignancies like the recently described MLL partner SACM1L. Finally, the determined patient-specific fusion sequences are succesfully used worldwide for minimal residual disease (MRD) monitoring to improve the treatment and outcome of acute leukemia patients. Disclosures: No relevant conflicts of interest to declare.

Abstract 3614 Pediatric MLL-rearranged acute monoblastic leukemia with t(9;11)(p22;q23) has a good outcome as compared to other MLL-rearranged AML. The biological background for this difference is unknown. Therefore, we compared gene expression profiles (GEP) of -t(9;11)(p22;q23) patients with other MLL-rearranged AML patients to identify differentially expressed genes. We performed GEP (Affymetrix HG U133 plus 2.0) in 245 pediatric AML patients (237 de novo and 8 secondary AML patients) and used RT-qPCR and Western Blot to validate expression. Methylation specific PCR was used to investigate epigenetic regulation. We tested the effect of the demethylating agent decitabine and the effect of knock-down by siRNA on both proliferation and drug sensitivity in AML cell lines. IGSF4, a cell-cell adhesion molecule, was highly expressed in AML-t(9;11). Expression within AML-t(9;11) was 18.5 times higher in FAB-M5 versus other FAB-types (p=0.013). RT-qPCR and Western Blot confirmed this. Methylation status investigation showed that high IGSF4 expressing AML-t(9;11) patients with FAB-M5 have no promoter hypermethylation, whereas all other cases do. This was also seen in cell lines. Cell line incubation with decitabine resulted in promoter demethylation and increased expression of IGSF4. Downregulation of IGSF4 by siRNA did not affect proliferation nor drug sensitivity in suspension culture. In conclusion, we identified IGSF4 overexpression to be discriminative for AML-t(9;11) with FAB-M5, regulated partially by promoter methylation. Disclosures: No relevant conflicts of interest to declare.

Abstract 365 Background: Translocations of the Mixed Lineage Leukemia (MLL) gene occur in a subset (5%) of acute myeloid leukemia (AML) and in mixed phenotype acute leukemia in infancy, a disease with extremely poor prognosis. Animal model systems show that MLL gain of function mutations may contribute to leukemogenesis. Wild-type MLL carries histone methyltransferase activity and affects specific target genes, such us HOXA cluster genes. While the more than three dozen MLL fusion proteins known today exert different specific functions, they finally induce transcription of individual target genes. Consequently, acute lymphoblastic leukemias (ALL) with MLL mutations (MLLmu) exhibit typical gene expression profiles including high-level expression of HOXA cluster genes. Aim of this study was to find a correlation between the MLL mutational status and tumor suppressor gene methylation/expression in acute leukemia cell lines. Results: Using MS-MLPA (methylation-specific multiplex ligation-dependent probe amplification assay), methylation of 24 different TSG was analyzed in 28 MLLmu and MLLwt acute leukemia cell lines. 1.8/24 TSG were methylated in MLLmu AML cells, 6.2/24 TSG were methylated in MLLwt AML cells. Hypomethylation and expression of the tumor suppressor genes (TSG) BEX2, IGSF4 and TIMP3 turned out to be characteristic of MLLmu acute myeloid leukemia (AML) cell lines. MLL wild-type (MLLwt) AML cell lines displayed hypermethylated TSG promoters resulting in transcriptional silencing. Demethylating agents and inhibitors of histone deacetylases restored expression of BEX2, IGSF4 and TIMP3 confirming epigenetic silencing of these genes in MLLwt cells. The positive correlation between MLL translocation, TSG hypomethylation and expression suggested that MLL fusion proteins were responsible for dysregulation of TSG expression in MLLmu cells. This concept was supported by our observation that Bex2 mRNA levels in MLL-ENL transgenic mouse cell lines required expression of the MLL fusion gene. Conclusion: These results suggest that the conspicuous expression of the TSG BEX2, IGSF4 and TIMP3 in MLLmu AML cell lines is the consequence of altered epigenetic properties of MLL fusion proteins. Disclosures: No relevant conflicts of interest to declare.

Abstract 4700 Statement WT1 gene expression (GE) is widely used for monitoring of minimal residual disease (MRD) in acute myeloid leukemia (AML) patients. Predictive value of WT1 GE in MLL-positive AML is not clear. Comparison of WT1 GE and fusion genes (FG), detected at RNA and DNA levels is rarely done. Here we present comparison of MRD data, assessed by WT1 GE, FG transcripts and MLL fusions alleles (MFA). Materials and methods 27 samples from 5 MLL-positive AML patients (pts) were analyzed. Median age was 13 years (range 0.8-58 years). 5 samples were obtained at diagnosis, 21 samples during follow-up treatment, 1 at diagnosis of relapse. Initial diagnostics included cytomorphology, immunophenotyping of blast cells, cytogenetics. Presence of MLL rearrangements was detected by FISH, reverse-transcriptase PCR (RT-PCR) and confirmed by long-distance inverse PCR. Among 5 pts there were two MLL-MLLT4-positive cases, two MLL-MLLT10-positive. One patient carried MLL-MLLT3 fusion gene. Quantification of MFA in genomic DNA by patient-specific primers-probe combinations and detection of FG transcript kinetics in RNA/cDNA during treatment were performed by RQ-PCR as previously described (T. Burmeister et al, Leukemia 2006, 20; J. Gabert et al, Leukemia, 2003, 17, respectively). Due to high concordance rate (96%) between qualitative results obtained by FG transcript and MFA detection, these results were combined into one group named as standard methods (SM) of MRD monitoring. Sample was defined as positive if only one result obtained by SM was positive. Sample was defined as negative if both results assessed by FG transcript and MFA were negative. Quantification of WT1 was performed in bone marrow (BM) samples by real-time quantitative PCR (RQ-PCR) with normalization to ABL gene. Additionally, BM samples taken from 14 pts without hematological malignancies were investigated. WT1 threshold level (TL) calculation was based on ROC curve analysis. TL was defined as value with maximal sensitivity and specificity. Results ROC analysis of WT1 GE revealed an area under curve (AUC) of 0.890 (95% CI 0.800-0.990). TL was set at 32 with sensitivity of 62% and specificity of 100%. Value of TL was equal to upper level of WT1 normalized copy number (NCN) in BM of 14 pts without hematological malignancies (median 10, range 0-32 WT1 copies/104ABL copies). In AML pts it was detected 13 WT1-positive samples (48%). 21 samples (78%) were positive by SM. 6 samples (22%) were negative both by SM and WT1 GE. Discrepant results were found in 8 samples (30%) (table 1). These discrepant results were observed in follow-up samples only. Qualitative concordance between SM and WT1 GE was 70% in our series. Conclusions Based on the ROC curves analysis it was not found particular TL of WT1 NCN that clearly distinguishes MRD-negative pts from those with positive MRD, detected by SM. Applicability of WT1 NCN quantification for MRD monitoring in the assessed MLL-rearranged AML samples appears to be limited because of relatively high percentage of false-negative samples. Discrepant results could be explained by high variability of WT1 GE during antileukemic therapy, and presence of WT1 GE in normal BM. Thus WT1-negative results should be considered as truly MRD-negative with great caution and have to be confirmed by FG transcript and/or MFA quantifications. Disclosures: No relevant conflicts of interest to declare.

Minimal residual disease (MRD) monitoring is an essential tool for current leukaemia therapy. The only standard method for MRD monitoring in childhood ALL is the quantitative detection of clonal immunoglobulin (Ig) and T-cell receptor (TCR) genes rearrangements. The quantitative detection of fusion genes or transcripts provides an alternative option for MRD monitoring. We aimed to compare the significance of these MRD methods by parallel monitoring of fusion transcripts/genes and Ig/TCR targets during the follow-up of children from the three most common ALL genotype groups. We analysed 117, 109 and 191 bone marrow samples from 28 TEL/AML1-positive, 7 MLL fusion-positive and 16 BCR/ABL-positive patients, respectively. To keep the comparability of different MRD approaches, we used qPCR detection systems with similar sensitivity (at least 10−4), we adopted ESG-MRD-ALL principles for MRD quantification and we related the MRD level in follow-up samples to the diagnostic level for all MRD methods. We found a very good correlation of fusion transcript- and Ig/TCR-based approaches (R2=0.903) with only 7% of samples differing by more than 1 log in a cohort of TEL/AML1-positive patients. A good correlation was also found between fusion transcript- and Ig/TCR-based MRD in MLL fusion-positive patients (R2=0.8419). Only 10% of samples differed by more than 1 log, being underestimated by Ig/TCR in 4.5% and by MLL-fusion transcript in 5.5%. For the follow-up of MLL-fusion-positive patients we further employed the monitoring of MLL-fusions on genomic level. The MRD based on genomic MLLfusion genes showed a very good correlation with Ig/TCR -based method (R2=0.9124) with only 5% of samples differing by more than 1 log, and it also closely correlated with MLL-fusion transcript levels (R2=0.9195). Strikingly, in BCR/ABL-positive patients we found a limited correlation of fusion transcript-based and Ig/TCR-based MRD (R2=0.6880) with 1/3 (34%) of samples differing by more than 1 log. In contrast to the MLL cases, the underestimation of MRD by individual methods was “asymmetrical”: 8% of the discordant samples had higher MRD measured by Ig/TCR and 26% by BCR/ABL transcript. Despite identical sensitivity of both methods, in 19% of samples the MRD positivity was revealed only by BCR/ABL approach while Ig/TCR approach gave negative results. Detailed analysis showed clinical significance of the discordant BCR/ABL vs. Ig/TCR MRD information. Altogether, 13 relapses occurred during the follow-up of our cohort. We compared number of BCR/ABL and Ig/TCR -positive samples among all BM specimens taken 6 and 12 months before relapse. While the majority of samples preceding relapse were BCR/ABL-positive (14/18 and 22/36 six and twelve months before relapse) only a minority of samples showed Ig/TCR positivity (7/18 and 12/36, respectively). The non-equal distribution of the BCR/ABL and Ig/TCR-positive samples was statistically significant (p=0.04 and p=0.03 for the two time-points, respectively). Our study shows, that in TEL/AML1 and MLL fusion-positive patients, fusion gene/transcript-based MRD monitoring provides information highly concordant to the standard Ig/TCR approach and thus it is useful as a complementary method in patients with absent or inadequate Ig/TCR targets (particularly in MLL cases where clonal Ig/TCR rearrangements are rare). The situation is different in BCR/ABL patients, where the MRD information from both approaches is discordant in a high subset of samples. This result probably reflects the dissimilar biology of this ALL subtype and the fact, that BCR/ABL-positive (prae-)leukaemic stem cell is different and multilineage involvement more common. Thus, in some cases, the fusion transcript monitoring reveals the existing pool of cells that increase the risk of relapse despite the Ig/TCR negativity. We conclude that MRD in all BCR/ABL–positive patients should be monitored not only by the standard Ig/TCR approach but in parallel also by the quantitative fusion transcript-based detection. Support: MSM0021620813, MZO00064203.

Objectives and aims of the study: The MLL oncogene on chromosome 11q23 undergoes various translocations in acute leukemias. MLL gene rearrangements are associated with worse outcome in infant acute lymphoblastic leukemia (ALL), while t(4;11) is a high-risk factor in children with ALL > 1 year of age. The prognostic value of MLL gene rearrangements in acute myeloid leukemia (AML) remains to be determined. In this study, we aimed at comprehensive analysis of the incidence and spectrum of MLL gene rearrangements in a large cohort of pediatric acute leukemias in Poland. Material and Methods: The study group comprised 355 children including 271 patients with de novo ALL, 24 children with relapsed ALL, 56 children with AML, three children with relapsed AML and a patient with acute bi-lineage leukemia. The presence of MLL rearrangements was determined with split-signal fluorescent in situ hybridization (FISH). Partner genes rearranged to MLL locus were identified with long-distance inverse PCR at the genomic DNA level. Results: MLL rearrangements were found in 18 patients with de novo ALL, 12 infants and six children > 1 year of age (6.6%). They included 12 t(4;11) with MLL-AFF1 fusion, two t(11;19) with MLL-MLLT1 fusion, one t(9;11) with MLL-MLLT3 fusion and one t(10;11) translocation with two gene fusions MLL-MLLT10 and PIWIL4-MLL. MLL rearrangements were also present in two patients with relapsed ALL. MLL rearrangement characterized 11 patients with de novo AML [four t(9;11) with MLL-MLLT3 fusion, one t(11;19) with MLL-ELL fusion, one (1;11) with MLL-EPS15 fusion, and a single t(1;11;17)] and one patient with secondary AML after ALL treatment [t(11;19)], which comprises 22% of all AML patients. This incidence is higher than usually described in literature (approximately 10–15%). Interestingly, in patient with t(1;11;17) two in-frame gene fusions were identified: MLL-MLLT11 and MYO18A-MLL, with the latter previously not reported. MLL gene rearrangement [t(9;11)], was also found in one of three relapsed AML cases. In a patient with acute bi-lineage leukemia both lymphoblasts and myeloblasts displayed t(4;11) translocation. Interestingly MLL-AFF1 fusion in this patient was accompanied by the fusion of the distal part of MLL to KIAA0999 gene on chromosome 11q23.3. Conclusions: In ALL patients MLL gene rearrangements are most frequent in infants (80% of cases) and very infrequent in older children (< 2%). Application of split-signal FISH as a screening for MLL gene rearrangements revealed unprecedentedly high incidence of these aberrations in childhood AML. Molecular analysis of MLL gene fusions and breakpoints shows several different mechanisms leading to these chromosome aberrations.

Statement. The MLL-MLLT4 (former MLL-AF6) fusion gene (FG) is relatively rare genetic abnormality, predominantly found in AML. It averages 3–5% among other MLL rearrangements. Here we present data of MRD monitoring in 2 patients with AML carrying an MLL-MLLT4 rearrangement by 2 approaches: using 1. FG transcript at RNA/cDNA level in comparison with 2. FG at genomic DNA level. Materials and methods. Patients (pts) were diagnosed according French-American-Britain (FAB) classification. Initial diagnostics included cytomorphology, immunophenotyping of blast cells, cytogenetics, FISH and reverse transcriptase PCR (RT-PCR). RT-PCR products were directly sequenced afterwards. In both cases identification of genomic chromosomal breakpoint sequences within MLL and MLLT4 genes was done by long-distance inverse PCR (LDI-PCR). MRD quantification in genomic DNA was performed using patient-specific primers and probes by real-time quantitative PCR (RQ-PCR). 500 ng of DNA was used per reaction. Standard curve was received by serial 10-fold dilutions of pts’ DNA into the DNA isolated from pooled lymphocytes of ten healthy donors. b-actin was used as DNA quality and quantity control. Detection of FG transcript kinetics during treatment was performed by RQ-PCR according to “Europe Against Cancer” recommendations for normalization by using control gene ABL (Gabert J. et al Leukemia, 2003, 17) and for using 10-fold dilutions of plasmids carrying MLL-MLLT4 fragment as source of standard curve. MRD value for cDNA targets were estimated as previously described (Beillard E. et al Leukemia, 2003, 17). Each sample was run in triplicates. According to the treatment design, time-points for MRD estimation were scheduled before each block of treatment. Totally 5 samples were evaluated in each patient (initial and 4 follow-ups). FLT3 -ITD status was estimated at the time of diagnosis. Informed consent was obtained in both cases. Patients’ characteristics, treatment and clinical outcome | | Case # 1 | Case #2 | |:-------------------------------------------------------------------:| ----------------------------------------------------------------------------------------- | ------------------------------- | | Age | 58 | 13 | | Sex | Male | Male | | Initial WBC*106/ml | 5.5 | 94.9 | | Immunophenotype | CD34+CD117+HLA-DR+ CD11c+CD13+CD33+CD65+ | CD34+CD117+CD13+ CD33+CD45+MPO+ | | Cytogenetics | 46, XY, del(5)(q?), der(5)t(5;6;11) (q22;q15q27;q23), der(6)t(5;6) (q22;q15), del der(11) | 46, XY, t(6;11)(q27;q23) | | FISH with LSI MLL | MLL deletion | MLL split | | RT-PCR | MLL-MLLT4 positive | MLL-MLLT4 positive | | MLL-MLLT4 FG transcript | exon 9-exon 2 | exon 9-exon 2 | | Localization of genomic chromosomal breakpoint within MLL and MLLT4 | intron 9-intron 1 | intron 9-intron 1 | | FLT3-ITD | Negative | Negative | | Induction treatment | 7+3 | AIE | | Consolidation therapy | 2× HAM 2× HDAC | 1× HAM 1× FLAG 1× HAE | | Maintenance | − | + | | Duration of therapy, months | 8 | 7 | | Achievement of CR | + | + | | OS, months | 8 | 7 | | EFS, months | 6 | 5 | | Current status | Alive in CR | Alive in CR | Results . Despite of achievement of CR, MLL-MLLT4 FG transcripts were detected in every sample tested after induction and consolidation chemotherapy by RQ-PCR. MRD value in case #1 in cDNA was fluctuated significantly within 2 log. Although in case #2 there was successive reduction from 260% at the beginning of treatment till 0.7% before maintenance therapy (after HAE block). Limited dilution series of a MLL-MLLT4-positive RNA into RNA of ten healthy donors showed a sensitivity limitation of 1E–05. For quantification of genomic chromosomal breakpoint sequences b-actin was amplified in each well. Deviation between Ct values of b-actin in different wells did not exceed ±2.0. In case # 1 the standard curve of the RQ-PCR assay for MLL-MLLT4 FG had slope of −3.19. Correlation coefficient was 0.987. Quantitative range of this assay was 1E–04 and sensitivity 1E–5. It was also observed a considerable variation of MRD levels in genomic DNA during treatment, like it was observed in MRD monitoring by FG transcripts. Fluctuations run up to 2.5 log. In case #2 the standard curve of the RQ-PCR assay for MLL-MLLT4 had a slope of −3.63 with correlation coefficient 0.992. Quantitative range of this assay was reached 1E–04 with sensitivity 1E–05. MRD level in this patient constantly decreased. Conclusions . The same tendency has been shown in each patient: fluctuations of MRD levels (2.5 log in case #1) and successive reduction (case #2). Results received at RNA/cDNA level and in genomic DNA cannot substitute each other, but they can be used as additives. It has been demonstrated that quantification of MLL-MLLT4 FG in genomic DNA is precise and suitable for MRD monitoring.

The pro B and the CD10-negative pre B immunophenotypes are found in approximately 12% and 2–3% of adult acute lymphoblastic leukemia cases, respectively. Both are characterized by a high prevalence of MLL aberrations. MLL fusion genes define a high risk group of ALL patients and they have proven useful as target structures for assessing minimal residual disease. It has been a matter of debate whether MLL aberrations are of prognostic significance within the group of pro B ALL patients. To better characterize the MLL recombinome in adult ALL we have prospectively and retrospectively investigated BCR-ABL-negative pro B (N=152) and CD10-negative pre-B ALL (N=31) patient samples obtained within the framework of the GMALL study group between 2001 and 2007. We used RT-PCR to detect MLL-AF4, MLL-ENL and MLL-AF9 and other MLL fusion genes. An inverse long PCR method was used to identify the exact chromosomal breakpoints in the MLL gene or unknown MLL fusions. Results: RT-PCR analysis revealed a MLL-AF4 fusion in 20/31 (64.5%) and a MLL-ENL fusion in 2/31 (6.5%) of CD10-negative pre-B cases. A MLL-AF4 fusion was detected in 101/151 (67.3%) and a MLL-ENL fusion in 9/151 (6.0%) of pro B cases. Additionally, two pro B patients showed a MLL-AF9 and MLL-CXXC6 fusion, respectively. The relative frequency of MLL aberrations was higher (50/68 = 73.5%) in those pro B patients with immunophenotypic coexpression of myeloid markers than in those without myeloid coexpression (41/84 = 48.8%). The chromosomal breakpoints in the MLL gene and the respective partner gene were determined on the genomic level in 91 patients. The distribution of chromosomal breakpoints in the MLL gene was the following: 1 in intron 7, 1 in intron 8, 33 in exon/intron 9, 27 in exon/intron 10, 28 in intron/exon 11, 1 in intron 12. These data illustrate the spectrum and relative frequency of MLL fusion genes in adult pro B and CD10-negative pre B ALL. The chromosomal breakpoint data provide insight into the molecular mechanisms leading to MLL aberrations in adult ALL. Chromosomal breakpoint sequences are powerful molecular markers for assessing MRD in individual patients as has been shown in the context of the GMALL MRD evaluation trials.

Chromosomal rearrangements of the human MLL gene are a genetic hallmark for aggressive acute leukemias. More than 60 partner genes have been characterized at the molecular level until now. The observed recombination mechanisms between MLL and its many recombination partner genes include reciprocal chromosomal translocations, insertions, and intrachromosomal 11q deletions and inversion. “Spliced MLL fusions”, have recently been described as a novel mechanism to generate functional chimeric fusion transcripts for 50% of t(11;19)(q23;p13.3) leukemia patients where the MLL gene is fused to the MLLT1 (ENL) gene and for two single leukemia patients. One patient had a t(11;15) involving the ZFYVE19 gene and the other patient had a 11q interstitial deletion involving the DCPS gene. Within these cases the disrupted MLL gene is fused only on RNA level to a functional transcription unit located in the vicinity of the breakpoint. Here we describe eight novel “spliced MLL fusions” involving the MLL partner genes ELL (1x), EPS15 (3x), MLLT3 (1x), and SEPT5 (1x). For ELL, MLLT3 and SEPT5 the identified spliced fusion seems to be a single or rare event. However 1/3 of the t(1;11)(p32;q23) leukemia cases exhibit “spliced MLL fusions”. I.e. approximately 1/3 of all breakpoints are located upstream of the EPS15 gene. Further studies will help to investigate the remaining question whether these two groups of t(1;11)(p32;q23) leukemia patients have any significant difference in outcome. Interestingly, for the t(11;22)(q23;q11.2) leukemia the major identified fusion transcript has an additional 57 bp exon derived from the 2 kb area up-stream of the SEPT5 gene. Furthermore this patient has a T-ALL and thus differs from all other cases where the patients show myeloid markers when the MLL gene is fused to one of the five members of the septine family. This is leading to the question whether or not this 57 bp big exon can be made responsible for the lineage shift. With these results the number of genes involved in „spliced MLL fusions” has increased from three to seven. Supported in part by grants Ma 1876/7-1, /8-1 and /9-1 from the DFG, grant N1KR-S12T13 from the BMBF and grant 102362 from the Deutsche Krebshilfe.

The MLL gene on chromosome 11q23 is translocated to a variety of fusion partners in about 5–10% of adults with acute myeloid leukemia (AML). The heterogeneity of these genetic aberrations hampers risk stratification of the patients. To identify prognostic parameters in AML patients up to 60 years with 11q23 aberrations, a pooled data analysis of 8 trials for the treatment of adults with AML was performed. All patients (pts) received double induction with araC and an anthracycline followed by intensive consolidation with either a high dose araC based chemotherapy or an autologous or allogeneic stem cell transplantation (alloSCT). 180 pts with 11q23 translocations were identified by cytogenetics and/or molecular techniques. 76 pts (42%) had a t(9;11), 35 (19%) a t(6;11), 17 (9%) a t(11;19), 14 (8%) a t(10;11) and 10 (6%) a t(11;17). 28 pts (16%) had other translocation partners. 77 pts (43%) had secondary chromosomal aberrations. Median age was 39 years (range 16–60). Median platelet and PB blast count was 45G/l and 7.5G/l. Median follow up is 53 months. Complete remission rate was 71%. Logistic regression identified t(9;11) (OR 2.6; 95% CI 1.2–5.9), secondary chromosomal aberrations (OR 0.4; 95% CI 0.2–0.8) and secondary leukemia (OR 0.3; 95% CI 0.1–0.6) as significant factors for induction success. Four year overall survival (OS) was 29% (95% CI 25–33%). Multivariate analysis identified presence of a t(9;11) (HR 0.6; 95% CI 0.4–0.9), platelets above the median (HR 0.6; 95% CI 0.4–0.9), secondary leukemia (HR 1.9; 95% CI 1.2–2.9) and PB blasts above the median (HR 1.9 95% CI 1.3–2.8) as significant factors for OS. Combination of these factors separated three risk groups for OS (p=0.0001): Good, pts with de novo AML, t(9;11) and PB blasts below median (4y OS 61%); intermediate, remaining pts with platelets above median (4y OS 32%); poor, remaining pts with platelets below median (4y OS 15%). Four year relapse-free survival (RFS) was 29% (95% CI 24–34%). Cox regression analysis identified the presence of a t(6;11) (HR 1.8; 95% CI 1.1–3.0) and PB blasts above the median (HR 1.7; 95% CI 1.0–2.6) as independent risk factors for RFS. By combination of these factors, three risk groups for RFS were defined: Good, pts without t(6;11) with PB blasts below median (4y RFS 63%); intermediate, pts without t(6;11) with PB blasts above median (4y RFS 23%); poor, pts with t(6;11) (4y RFS 9%). The effect of different consolidation strategies was analyzed “as treated” in CR-pts who had received at least one cycle of consolidation. In these pts, RFS was improved by an alloSCT (MRD or MUD) in first CR (p=0.04). However, this effect was restricted to the RFS-intermediate risk group (p=0.029), whereas pts in the good or poor risk group did not benefit from an alloSCT in first CR. In conclusion, the prognosis of patients with translocations of chromosome 11q23 is heterogenous. Risk stratification of these patients is feasible based on the MLL-fusion partner and clinical parameters. This prognostic model can be used as a basis for the selection of consolidation therapy.

The human MLL gene on 11q23 is frequently involved in chromosomal rearrangements including almost all human chromosomes. These rearrangements are associated with high-risk acute leukemias. Today more than 60 MLL fusion partner genes which represent the MLL recombinome have been characterized at the molecular level. Interestingly, more than 20% of all MLL translocations are highly complex. Here we distinguish between two types of complex rearrangements: Type I: A complex rearrangement is a genetic prerequisite due to the transcriptional orientation of the involved genes, for example t(10;11) with MLL and AF10, respectively; Type II: A complex rearrangement is not prerequisite, although identified in leukemia patients. More than 500 MLL rearrangements have been analyzed at the Diagnostic Center of Acute Leukemia so far and here, we present data on 60 complex translocations that were resolved at the molecular level. Three-way chromosomal translocations, insertions and combinations thereof are the molecular basis for most of these complex MLL rearrangements which in most cases create 3 or more fusions. However, type I is mainly based on insertions and type II is based on 3 way translocations. Within these complex rearrangements, the 5′ part of the MLL gene is in general fused in frame to one of the most frequent partner genes. But the 3′ part of the MLL gene is fused in 36 cases (60%) to a novel gene such as, for example, APBB1IP or CMAH. We defined this heterogeneous group of novel translocation partner genes as “reciprocal MLL fusion genes” as they had not been identified as fusion partners to the 5′ part of the MLL gene before. These fusions are only in frame in four cases; the remaining fusions are out of frame or the genes have an opposite transcriptional orientation. I.e. these genes show genetic haplo-insufficiency and thus may also cooperate in promoting leukemogenesis. For the remaining 40% the 3′ part of the MLL gene is fused to a chromosomal area which does not encode any gene. In many complex aberrations additional genes like for example RREB1 or OS9 are fused mostly out of frame to other genes or non-coding chromosomal areas of these complex rearrangements. These genes are also haplo-insufficient and may also contribute to the leukemogenic process. These results demonstrate that LDI-PCR is a potent tool to identify highly complex and cryptic rearrangements as both MLL fusion alleles are analyzed at the same time and reveal immediately if a balanced or more complex rearrangement is present. Whether the pathobiology of type II complex rearrangements is different in comparison to their not complex counterparts has still to be proven. Supported in part by grants Ma 1876/7-1, /8-1 and /9-1 from the DFG, grant N1KR-S12T13 from the BMBF and grant 102362 from the Deutsche Krebshilfe.

Aberrations of the MLL gene on chromosome 11q23 occur in about 5–10% of adult patients with acute myeloid leukemia (AML). Most frequently, the MLL gene is fused to the genes AF9 on chromosome 9 or AF6 on chromosome 6 resulting in the translocation t(9;11) and t(6;11), respectively. The remaining patients with aberrations of chromosome 11q23 constitute a heterogeneous group with numerous fusion partners. This heterogeneity hampers risk stratification of the patients. Accordingly, AML patients with 11q23 aberrations have varyingly been stratified as intermediate or high risk patients by different study groups. To analyze the prognostic impact of different 11q23 aberrations in AML patients up to 60 years, a pooled data analysis of 6 consecutive studies for the treatment of adult AML patients was performed. All patients received double induction treatment with araC and an anthracycline followed by an intensive consolidation with either a high dose araC based chemotherapy regimen or an autologous or allogeneic stem cell transplantation. In total, 150 patients with 11q23 aberrations were identified by cytogenetics and/or molecular techniques. 53 patients (35%) had a t(9;11), 27 patients (18%) a t(6;11), 10 patients (7%) a t(11;19), 8 patients (5%) a t(11;17) and 6 patients (4%) a t(10;11). 46 patients (31%) had other fusion partners or deletions of 11q23. Median age of all patients was 39 years (range 16–60). Overall complete remission rate was 73% with no significant difference between the groups (79% for t(9;11), 74% for t(6;11) and 66% for others). Altogether, 9% of the patients had treatment failure and 7% died during induction. Relapse-free survival (RFS) and overall survival (OS) at 5 years for the entire group was 24% and 26%, respectively. RFS and OS of the t(6;11) group was 6% and 13% and thus was significantly inferior to patients with t(9;11) (RFS: 48%, OS: 31%) and patients with other 11q23 translocations (RFS: 20%, OS: 18%) or 11q23 deletions (RFS: 38%, OS: 33%). Within the t(9;11) group there was no difference in RFS or OS between patients who had a t(9;11) as a sole aberration (n=36) and patients with additional aberrations (n=17) and between patients with de novo or secondary AML. Moreover, in t(9;11) no difference in RFS was observed between treatment with high dose araC, autologous or allogeneic stem cell transplantation as late consolidation. Regarding other translocations of 11q23, there was no difference in RFS and OS between patients with t(11;17), t(11;19), t(10;11) and patients with various other fusion partners. In conclusion, these data demonstrate that the prognosis of patients with 11q23 is heterogeneous. Patients with t(9;11) who enter CR have a relatively good outcome independent of the consolidation therapy used. In contrast, the prognosis of patients with t(6;11) is extremely poor. Therefore, patients with t(6;11) should be regarded as high risk and alternative treatment strategies for this subgroup are required.

CCAAT/enhancer binding proteins (C/EBPs) are transcription factors that play a role in myeloid differentiation and may be dysregulated in myeloid leukemias as targets of BCR/ABL, ETO, and FLT3 mutations. CUG triplet repeat binding protein-1 (CUGBP1) is a translational regulator of C/EBP-β that increases the translation of full length C/EBP-β (LAP) and a dominant negative isoform, LIP. Translocations involving the human mixed-lineage leukemia gene (MLL) on chromosome 11 are leukemogenic events that result in a characteristic profile of HOX gene expression and confer a poor prognosis. Expression of AF1q, a gene involved in normal hematopoiesis, correlates with increased risk in acute myeloid leukemia (AML) and myelodysplastic syndrome and is fused with MLL in translocation t(1;11)(q21;q23). Here we describe a complex rearrangement of chromosome 1 and 11 involving translocation t(1;11)(q21;q23) and a separate event disrupting the MLL gene by either insertion or translocation of the 3′ region into the p arm of chromosome 11 (p11). Both activation of HOX genes and inactivation of C/EBP result from this MLL rearrangement. In blasts obtained from an infant with newly diagnosed AML, G-band karyotyping and fluorescence in situ hybridization (FISH) for chromosome 1 and 11 telomeres revealed an apparent reciprocal translocation. FISH for MLL, however, identified the 3′ end of MLL on 11p rather than the expected 1q. Array comparative genomic hybridization (aCGH, SpectralChip 2600), consisting of BACs that provide an average resolution of 1.0 MB across the genome, revealed no genomic losses or gains. Genomic long distance inverse polymerase chain reaction (LDI-PCR) for the 5′ portion of MLL identified fusion of MLL with AF1q at introns 10/11 and 1/2, respectively. LDI-PCR for the 3′ portion of MLL identified a novel fusion with CUGBP1, located on 11p and consistent with the 3′ FISH probe for MLL. Reverse transcriptase polymerase chain reaction (RT-PCR) confirmed the predicted MLL-AF1q fusion transcript but did not identify a CUGBP1-MLL product. To confirm that MLL-AF1q was an activating fusion product, we assessed HOXA9 expression in four AML samples by RT-PCR and found HOXA9 in this AML sample to be highly elevated. We then hypothesized that interruption of the CUGBP1 gene could lead to an arrest in differentiation by a decrease in C/EBP expression. We found CUGBP1 expression by quantitative RT-PCR to be markedly reduced in this sample compared to all other AML samples tested. Western blot analyses confirmed the low expression of CUGBP1 and demonstrated marked reduction in C/EBP-β expression. These findings suggest a complex rearrangement event that caused (a) MLL activation by fusion with AF1q and (b) C/EBP suppression by loss of transcriptional enhancement from CUGBP1. This rearrangement could have resulted from two sequential events: independent translocation and deletion/insertion events or a single crossover event involving four recombination sites. The sequences obtained by LDI-PCR showed complete conservation of the genomic MLL sequence, suggesting the latter mechanism. This is the first report to suggest a role for CUGBP1 in leukemogenesis.

Chromosomal rearrangements of the MLL gene are responsible for 5–10% of all acute leukemias, biphenotypic leukemias and myelodysplastic syndromes. 5–10% of these MLL aberrations are therapy-related leukemias with the frequent fusion partners ELL, MLLT3, MLLT10 and MLLT2. The large number of known and unknown MLL fusions (>80) renders a precise diagnosis a demanding task. Even though all MLL rearrangements are associated with high risk acute leukemias, the outcome (favorable or poor) depends on the partner gene. Thus, the identification of MLL gene fusions is necessary for rapid clinical de-ci-sions resulting in specific therapy regimens. After their cytogenetic identification, only the most common MLL fusions MLLT2, MLLT3, MLLT1, MLLT4, ELL and MLLT10 (80%) are investigated by RT-PCR analysis, whereas infrequent or unknown MLL rearrangements (20%) are more or less excluded from further analysis. Therefore we established a LDI-PCR (long-distance inverse PCR) based method that uses small amounts of genomic DNA to determine any type of MLL associated chromosomal abnormality. With this diagnostic tool more than 400 prescreened und unscreened adult and pediatric samples (AML/ALL) from different European diagnostic centers have been analyzed. During this study more than 200 MLL rearrangements were characterized for their precise localization of genomic break-points. The identified MLL rearrangements consist of 25 different fusion genes including the following eight novel MLL partner genes: ACACA, ARGHEF17, BCL9L, MAML2, SELB, SMAP1, and TIRAP. The identified MLL aberrations are basically reciprocal translocations, but also deletions, inversions, insertions and interstitial duplications have been determined. Combining the data of our study and data retrieved from the literature, a total of 88 partner genes can now be annotated. 52 fusion partners (60%) have been characterized on the molecular level whereas at least 36 (40%) remain to be identified. These results demonstrate that this new diagnostic tool in combination with split-signal FISH is qualified for the rapid analysis of known as well as unknown MLL partner genes. Furthermore, the determined patient specific fusion sequences are useful for minimal residual disease (MRD) monitoring of MLL associated acute leukemias. The use of these MRD markers will contribute to improve the treatment and outcome of acute leukemia patients. A first prospective study was already initiated by the German ALL study group and verified the reliability of these genomic markers for MRD monitoring.

While there is enough convincing evidence in childhood ALL, the data on the pre-natal origin in childhood AML are less comprehensive. Our study aimed to screen Guthrie cards (neonatal blood spots) of childhood acute myeloid leukaemia (AML) and acute lymphoblastic leukaemia (ALL) patients for the presence of their respective leukaemic markers. For the analysis we used PML/RARa, AML1/ETO and CBFb/MYH11 fusion genes, translocations of MLL gene and internal tandem duplication of Flt3 gene (Flt3/ITD) in patients with AML and TEL/AML1 fusion gene and immunoglobulin (Ig) and/or T-cell receptor (TCR) gene rearrangements in patients with ALL. These molecular markers on the DNA level present suitable candidates for backtracking of leukaemic cells in newborn material, because they are clonotypic and altogether their incidence is ~40% and >90% in AML and ALL, respectively. We screened Guthrie cards from 13 AML patients (4x PML/RARa,, 3x CBFb/MYH11, 2x AML1/ETO, 2xFlt3/ITD, 3x MLL rearrangement - MLL/AF6, MLL/AF9 and MLL/AF10) and from 15 ALL patients (13x Ig/TCR rearrangements, 3x TEL/AML1). One AML patient from our group had PML/RARa fusion gene together with Flt3/ITD, in 1 ALL patient we screened the Guthrie card for both Ig/TCR and TEL/AML1. We designed patient-specific PCR primers and we established nested PCR assay for each clonotypic sequence prior to the analysis of the corresponding Guthrie card. Assay specificity was determined using serial dilutions of patient DNA into the DNA of a healthy donor. The sensitivity of PCR was >=10(−4) in all patients. Therefore, this approach allowed us to detect the pre-leukaemic clone provided 1–10 positive cells were present on the Guthrie card. In three patients with ALL we reproducibly detected identical rearrangements both in the presentation sample and on the Guthrie card. The first patient harboured two independent rearrangements: TCR-delta Vd2/Dd3 and IgH VH3/JH5 and both were detectable on Guthrie card. In the second patient two PCR systems were optimised with adequate identical sensitivity: Igkappa Vk1/Kde and TCRgamma VgI/Jg1.3-2.3. However, only the latter detected pre-leukaemic cells on Guthrie card. The third patient had TEL/AML1 and we found this translocation on her Guthrie card. We did not find patient-specific molecular markers in any patient with AML. In the present study we confirmed the prenatal origin in 20 % of ALL patients. The negative results in the AML group can not disprove the theory that some childhood AML cases are also initiated in-utero. The negative results might be caused also by the fact that the age at presentation was higher in our AML group compared to ALL patients (median 7 and 3 years, respectively), the size of pre-leukaemic clone in AML might be below the sensitivity threshold of our approach, and Flt3/ITD could be of post-natal origin as recently suggested. However, our present data based on the largest cohort examined so far show that there is much less evidence for the pre-natal origin of childhood AML. Support: grants #7436 (Czech Ministry of Health) and #0021620813 (Czech Ministry of Education).

Acute leukemias are frequently associated with specific chromosomal translocations of the human MLL gene. In general, MLL translocations define a distinct disease entity that needs to be diagnosed with precision to facilitate rapid clinical decisions. Here we present data about a new PCR based method that uses patient genomic DNA to identify any MLL fusion. Fourty different MLL translocations were successfully analyzed. We will present three novel MLL translocation partner genes and a new MLL deletion. The benefits of this novel technique for diagnosis and MRD analyses will be discussed. Supported by grant 2001.061.1 from the Wilhelm Sander foundation.