A strong correlation between trisomy 13 and AML1 mutation has been recently reported in AML, particularly in M0 FAB subtype (Dicker et al, Silva et al). Whereas the exact mechanism of this coincidence has not been completely clarified, this correlation suggests a probable cooperation of these two events in leukemogenesis. FLT3, located on the 13q12 region, has been proposed to be implicated by over expression induced by gene dosage effect. However, trisomy 13 remains a rare event in AML and its association with AML1 mutations is not constant. To better document this association, we studied, a cohort of patients with trisomy 13 by array-CGH, which is a powerful tool to detect cryptic alterations as small size gene copy number variations (CNV). In addition we searched for AML1 mutations and performed the correlation with FLT3 alterations. We studied 12 AML with trisomy 13 (7 of which were M0 FAB subtype) and 6 additional patients affected by non-AML hematological malignancies (2 T-ALL, 1 NHL and 3 MPD). We performed direct sequencing of the exons 3 to 8 of AML1. Array-CGH was performed on 10/12 AML patients using Agilent CGH-Array 185k with pooled normal DNA as reference. GEP was performed using HG-U133 plus 2.0 Affymetrix Array on 5 patients characterized by trisomy 13 and AML1 mutations. AML1 mutations were observed in 8/18 patients including 6 of the 7 M0 AML, which represents a particularly high rate even in this subtype (86%). On the contrary, none of the non-AML patients harbored AML1 gene mutation. FLT3 locus analysis in the 12 AML patients showed the following results: None FLT3-ITD was observed whereas 5 FLT3-D835 mutations were detected (42%). FLT3 over expression was observed by GEP on 4/5 analyzed patients all without FLT3-D835 mutation. The remaining patient presented a FLT3-D835 mutation. CGH study confirmed trisomy 13 in 9/10 AML cases with weak amplitude for 4 patients, strongly suggesting that trisomy 13 was a sub-clonal anomaly, in accordance with caryotype results. Only one other recurrent abnormality was observed by CGH in 3 AML patients: 2 gains and one deletion affecting chromosome 21, all including AML1 locus. Moreover, CNV incidence was strong in these patients traducing a great genomic instability: there were 34 CNV gains and 23 deletions equally distributed all along the genome with a size extending from 300kb to complete chromosome. Among the interstitial CNV loci were included the following genes (CD68, CDH1, CDX2, CTNNA3, DOCK2, EIF3S9, EIF4A1, ERG, GNA12, IGF2, KLRC3, MAD1L1, MAFA, MKL2, VAV2) that could perhaps play a role in leukemogenesis. Here we confirmed the strong correlation between AML1 mutation and trisomy 13. Trisomy 13 is often sub-clonal indicating this alteration could be an additional event contributing to leukemia progression. In accordance with previous reported data, FLT3 over expression or FLT3-D835 mutation could cooperate with AML1 mutation. However, alteration of genes located on other CNV could not be excluded.
Disclosure: No relevant conflicts of interest to declare.