Letter to the Editor | Published:

MLL partner genes drive distinct gene expression profiles and genomic alterations in pediatric acute myeloid leukemia: an AIEOP study

Leukemia volume 25, pages 560563 (2011) | Download Citation

Childhood acute myeloid leukemia (AML) is a heterogeneous disease with an overall poorer outcome as compared with acute lymphoid leukemia. Marked differences in terms of outcome of AML patients have been confirmed to be related to the presence of specific genetic aberrations.1 A subset of high-risk AML patients is characterized by rearrangements, involving the mixed-lineage leukemia gene (MLL) located on chromosome 11q23. More than 40 different translocation fusion partners of MLL have been identified in AML at diagnosis; however, five partner genes account for over 90% of MLL-translocated pediatric AML (AF9, AF10, AF6, ENL and ELL).2 Recently, Balgobind et al.3 published results from a collaborative international study group describing different clinical outcomes for MLL-11q23 translocation partner genes.

Here, we report our investigation of the role of various MLL translocations in the Italian AML patients enrolled in recent AML protocols of the AIEOP (Italian Association of Pediatric Hematology and Oncology). As genomic aberrations are used to stratify patients, we used genome-wide technology in order to characterize MLL subgroups. Bone marrow samples of children aged <18 years, with newly diagnosed de novo AML, were collected from 2000 to 2008. The patients analyzed were found from the AIEOP registry that collects data from all children with cancer diagnosed in AIEOP centers in Italy. Patients diagnosed to be affected by acute promyelocitic leukemia, granulocytic sarcoma, secondary AML, myelodysplastic syndrome or Down syndrome, as well as patients with a pretreatment phase longer than 14 days, were excluded from this study.4 The initial diagnosis of AML was centrally established according to morphology (French–American–British (FAB) classification) and immunophenotypic analysis at the laboratory of Pediatric Hematology of the University Hospital in Padova. Chromosome analysis was performed using standard laboratory procedures. Multiplex reverse transcriptase PCR (RT-PCR) was used to detect nine MLL fusion transcripts, and 77 MLL-rearranged patients were found (10.4%) in the series of 741 Italian children with AML, consecutively diagnosed between 2000 and 2008. The outcome of patients with different MLL rearrangements varied significantly. In particular, the t(6;11)(q27;q23) group had a very poor prognosis with a 3-year probability of event-free survival (EFS) (23.3%). On the contrary, the t(9;11)(p22;q23) was intermediate, as previously described (55.4%, P<0.01, Figure 1a). The identification of a heterogeneous outcome associated with different MLL translocation partners suggested that specific biological features might have a distinct role in those subgroups. The GeneChip Human Genome U133 Plus 2.0 (Affymetrix, Santa Clara, CA, USA) was used for 42 of 77 MLL-rearranged patients (see Table 1) to generate gene expression profiles (GEP). Unsupervised analysis of GEP consistently separated MLL-rearranged leukemias with respect to partner gene and FAB subtypes (Figure 1b).5 MLL/AF9-positive samples were found to be divided into two major groups, which were related to distinct FAB variants, M7 and M5. We demonstrated through GEP that the t(9;11) with FAB M7 clustered tightly together with respect to the t(9;11) with FAB M5. The fact that a supervised algorithm assigned a specific gene expression signature common to all t(9;11)-positive patients despite FAB classification indicates that the translocation drives a specific t(9;11) genetic leukemia signature. Patients with MLL or other rearrangements (AF1, ENL, ELL, SEPT6) clustered together for the most part, probably related to outcome or for the shared M5 FAB subtype. The supervised analysis of variance (ANOVA) identified 229 probe sets that were differentially expressed among the considered MLL subgroups (Figure 1c). Among the ANOVA-predicted genes, we highlighted GAS1 expression associated to the two worst prognostic subgroups, MLL/AF6 and MLL/AF10, and FLT3 expression in t(9;11) cases, which was found lowered in patients with FAB M7, whereas it increased significantly for t(9;11) FAB M5. The supervised analysis of each MLL partner gene, compared with the rest of the MLL translocations, showed a specific and significant gene expression signature exclusively for MLL/AF6 (Figure 1d).6 This finding implies that this translocation is biologically different. In fact, clinical events in t(6,11)-translocated patients occurred within 1 year from diagnosis, suggesting a very aggressive behavior of the MLL/AF6 chimeric gene. Adequate investigation of its functional role is urgent in order to identify more effective therapy. By GEP, we found novel candidate genes with higher expression in MLL/AF6 patients (AF6, TANC1, IL12R2), which will be further investigated.7 To support the new biological interpretation of MLL-rearranged AML, the same cohort of patients was evaluated with standard cytogenetic analysis, which is currently used for risk stratification in Italian protocols, improved by the array comparative genomic hybridization (a-CGH) analysis using an Agilent Human Genome Microarray Kit 244A (Agilent Technologies, Santa Clara, CA, USA). Karyotypes showed recurrent abnormalities in 32 of 39 cases. MLL translocation as sole abnormality was seen in 17 of 32 cases (53%). The complex karyotype, defined as MLL translocation associated to additional cytogenetic abnormalities, was observed in 13 of 32 (41%). Array CGH was performed on samples of 28 MLL-rearranged patients. Of 28 MLL-rearranged patients, 19 (68%) showed genomic abnormalities in this analysis. Among the abnormal cases, the total genomic copy number alteration was 575.8 Mb; 75% were amplifications and 25% deletions. Gain and loss of chromosomes were mainly observed in t(6;11)-positive patients, whereas the t(9;11) cases never showed gain or loss of chromosomes. We identified two recurrent regions of deletion: one at chromosome 12p and the other at 6q27, found in 4/8 t(6;11)-translocated patients. One amplified region at 11q was found in five patients with t(6;11), one patient with t(10;11) and three patients with t(11;others). We also evaluated the prognosis of novel cytogenetic features described here, even if the patient number is too low to perform statistical analysis. Among MLL-rearranged patients with del(12p), 5 out of 6 patients relapsed at a median time of 10 months; 3 of these 5 patients died after relapse. If we exclude the t(6;11) patients with del(12p), the 3-year EFS of the t(6;11) group increased to 53.3% (compared with 23.3% reported above), indicating an adverse effect on prognosis of del(12p).

Figure 1
Figure 1

(a) EFS probability of 77 patients with MLL-rearranged AML. The group of t(11;other) represents t(11;19)(q23;p13.1), t(11;19)(q23;p13.3), t(x;11)(q24;q23) and t(11;17)(q23;q25) grouped together, all showing a favorable outcome (EFS 74.1%), whereas the t(6;11)(q27;q23) together with t(10;11)(p12;q23) showed a very poor clinical outcome (EFS 23.3%) (P<0.01). (b) Unsupervised hierarchical clustering analysis; GEP of t(6;11) (n=11)-, t(9;11) (n=11)-, t(10;11) (n=10)- and t(11-other) (t( 11;19) n=5, t(1;11) n=2 and t(x;11) n=3)-rearranged patients are shown. Dendrograms were generated to cluster patients using Ward's method and Euclidean distance, and heat maps were used to highlight associations between clusters of patients and clusters of genes. The dendrogram was obtained using expression data filtered by variance. (c) Supervised hierarchical clustering analysis using the 229 probe sets identified by ANOVA analysis among 42 MLL–AML patients. The genetic subtypes are indicated below the dendrogram with color codes. (d) Heat map of the most differentially expressed genes between MLL/AF6 and all MLL–AML rearrangements studied. The 59 probe sets differently expressed between the two groups were identified by gene selection based on Wilcoxon's test. The result was that 44 probe sets were upregulated and 15 probe sets downregulated in MLL/AF6 patients. The two analyzed groups are indicated with color codes. P-values (Wilcoxon's tests and ANOVA) were obtained using a permutation approach. To control the false discovery rate, multiplicity corrections were used. Probes with adjusted P-values <0.01 for Wilcoxon's tests, and <0.05 for ANOVA analysis were declared significant.

Table 1: MLL patients’ clinical features

Finally, we showed that the MLL-AML is a heterogeneous leukemia depending on the MLL partner gene for GEP and cytogenetic analysis, with novel features for each subgroup of patients. The aggressive behavior mediated by the MLL/AF6 chimeric gene calls for in-depth investigation of its functional role in order to identify more effective therapy. The MLL/AF6-positive patients showed the highest frequency of genomic imbalances along with new target genes aberrantly expressed, which require further study. We also described here for the first time that del(12p) influences the outcome of MLL–AML, in particular in t(6;11)-positive cases. The use of a-CGH in AML–MLL revealed that genomic gains were found more frequently than losses, suggesting a general role of proto-oncogene activation in this leukemia. The observation that 6 out of 39 patients had the del(12p) suggested a possible association between the del(12p) and pediatric MLL-rearranged AML (15.5% in comparison with 3.7%, recently described in a large overall AML cohort8, 9). The finding that MLL–AML encompasses distinct biological and clinical diseases within the high-risk forms of pediatric AML forces reconsideration of these patients for distinct, specific therapies.


  1. 1.

    , , , , , et al. Clinical features of childhood acute myeloid leukaemia with specific gene rearrangements. Leukemia 2004; 18: 1427–1429.

  2. 2.

    , , , , , et al. New insights to the MLL recombinome of acute leukemias. Leukemia 2009; 23: 1490–1499.

  3. 3.

    , , , , , et al. Novel prognostic subgroups in childhood 11q23/MLL-rearranged acute myeloid leukemia: results of an international retrospective study. Blood 2009; 114: 2489–2496.

  4. 4.

    , , , , , et al. Results of the AIEOP AML 2002/01 study for treatment of children with acute myeloid leukemia. Blood 2009; 114: 22.

  5. 5.

    , , , , , et al. Clinical utility of microarray-based gene expression profiling in the diagnosis and subclassification of leukemia: Report on 3334 cases from the international MILE study group. J Clin Oncol 2010; 28: 2529–2537.

  6. 6.

    , . Microarray-based identification of new targets for specific therapies in pediatric leukemia. Curr Drug Targets 2007; 8: 761–764.

  7. 7.

    , , , , . Self-association mediated by the Ras association 1 domani of AF6 activates the oncogenic potential of MLL-AF6. Blood 2010; 8: 63–70.

  8. 8.

    , , , , , et al. Cytogenetics of childhood acute myeloid leukemia: United Kingdom Medical Research Council Treatment trials AML 10 and 12. J Clin Oncol 2010; 28: 2674–2681.

  9. 9.

    , , , , , et al. Prognostic impact of specific chromosomal aberrations in a large group of pediatric patients with acute myeloid leukemia treated uniformly according to trial AML-BFM 98. J Clin Oncol 2010; 28: 2682–2689.

Download references


This study was supported by grants from the Fondazione Città della Speranza-Padova-Italy (to MP, GB, AZ, SB and LT), from AIRC (Associazione Italiana Ricerca sul Cancro, to GB and GteK), from PRIN/Programmi di ricerca di Rilevante Interesse Nazionale, Rome (to GteK, GB), and the Ministero della Salute (to GB, GteK). We thank MariaGrazia Giacometti and Silvia Disarò for sample collection; Dr Laura Sainati and Dr Samuela Francescato for morphology evaluation; Nancy Jenkins for paper editing.

Author information


  1. Department of Pediatrics, Laboratory of Hematology-Oncology, University of Padova, Padova, Italy

    • M Pigazzi
    • , S Bresolin
    • , A Beghin
    • , A Di Meglio
    • , S Gelain
    • , L Trentin
    • , E Baron
    • , M Giordan
    • , A Zangrando
    • , B Buldini
    • , A Leszl
    • , M C Putti
    • , G Te Kronnie
    •  & G Basso
  2. Department of Pediatrics, ‘Lalla Seràgnoli’, Hematology-Oncology Unit, Ospedale Sant’Orsola, University of Bologna, Bologna, Italy

    • R Masetti
    •  & A Pession
  3. Department of Pediatrics, Hematology-Oncology Unit, University of Milano-Bicocca, Hospital S. Gerardo, Monza, Italy

    • C Rizzari
  4. Onco-hematology Department, IRCCS Ospedale Bambino Gesu′, Roma, University of Pavia, Rome, Italy

    • F Locatelli


  1. Search for M Pigazzi in:

  2. Search for R Masetti in:

  3. Search for S Bresolin in:

  4. Search for A Beghin in:

  5. Search for A Di Meglio in:

  6. Search for S Gelain in:

  7. Search for L Trentin in:

  8. Search for E Baron in:

  9. Search for M Giordan in:

  10. Search for A Zangrando in:

  11. Search for B Buldini in:

  12. Search for A Leszl in:

  13. Search for M C Putti in:

  14. Search for C Rizzari in:

  15. Search for F Locatelli in:

  16. Search for A Pession in:

  17. Search for G Te Kronnie in:

  18. Search for G Basso in:

Corresponding author

Correspondence to M Pigazzi.

About this article

Publication history




Further reading