T-cell acute lymphoblastic leukemia (T-ALL) is a heterogeneous disease comprising several biological and clinical entities. Based on recent studies and in agreement with the paradigm of multistep oncogenesis the notion that multiple genetic hits are cooperating in the generation of T-ALL has emerged. Genetic and functional data suggests that a stepwise alteration of at least four specific pathways is required before thymocytes become fully malignant:

  • cell cycle deregulation,

  • impaired differentiation,

  • proliferation and survival advantage, and

  • unlimited self-renewal capacity.

As the sequential order of the genetic events remains widely elusive, we aimed to determine their consecutive acquisition in the model of NUP214-ABL1 positive T-ALL, which is always accompanied by either a TLX3 or TLX1 rearrangement as well as frequent loss of CDKN2A. Array comparative genomic hybridization (aCGH) analysis of 102 childhood T-ALL patients and 12 matched relapse samples revealed a NUP214-ABL1 amplification in 5 samples, two of which were relapse samples and showed the aberration only at this stage of the disease. However, NUP214-ABL1-specific RT-PCR determined that the fusion gene was also present in those two diagnostic specimens in which aCGH failed to detect the amplification and FISH analysis showed ambiguous results with a minor fraction (5%) of positive cells. At relapse in one of these patients the NUP214-ABL1 fusion was found integrated and amplified at 7q, thus in contrast to the episomal form stably transmitted, and present in 65% of the cells. On the other hand, at diagnosis all five patients showed a TLX3 rearrangement by fluorescence in situ hybridization (FISH) in >60% of the cells. The data obtained by aCGH revealed also homozygous or heterozygous loss of CDKN2A (9p21) in four and one of the cases, respectively, and heterozygous deletion of WT1 (11p13) in three cases. FISH analysis confirmed the heterozygous loss of WT1 and inactivating mutations in the C-terminal zinc-finger motifs of the second allele were found in two of the patients. In one of these cases the WT1 mutation was readily detectable at the stage of relapse, but sensitive allele-specific PCR was required to trace the mutation back to the diagnostic sample, suggesting that - as typical for a tumor suppressor function - loss of the first allele was followed by complete inactivation of WT1 by mutation of the second allele during disease progression. Moreover, in this case monoallelic loss of WT1 was observed in 93% of the cells, whereas TLX3 was rearranged in only 63%. In contrast, in a second case exactly the opposite pattern was observed, namely monoallelic loss of WT1 in 42% and TLX3 rearrangement in 73% of the cells. In this sample, preliminary data obtained by consecutive FISH on the single cell level showed two copies of WT1 in TLX3 positive cells. Moreover, based on the aCGH data, which however still requires confirmation by sequential FISH experiments, CDKN2A was deleted in a very high percentage of the cells indicating that loss of this cell cycle regulating tumor suppressor is an early genetic event. Together our data suggest that rearrangement of TLX3 and loss of CDKN2A and WT1 precede the formation of the NUP214-ABL1 fusion and subsequent amplification. On the other hand loss of WT1 and TLX3 rearrangement may possibly occur in a changeable order, whereas the temporal occurrence of WT1 inactivating mutations remains to be determined.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author