Genetic studies of acute lymphoblastic leukemia

Abstract

Acute lymphoblastic leukemia (ALL) is the most common pediatric malignancy, comprising approximately 25% of all childhood cancers. A huge improvement in outcome of children with ALL during the last 20 years has resulted in remission rates exceeding 95% and a 5 year event-free-survival approaching 80%. Tailored and individualized treatment with prolonged regimens of multi-agent chemotherapy based on clinical and cytogenetic prognostic factors may explain this development. However, although ALL may be considered as the same disorder in adults and children, the improvement in outcome has been less prominent in adults, and relapse remains the most important cause of treatment failure in all age groups. Therefore, there is a need for more detailed and functional characterization of leukemic clones with already established recurrent cytogenetic abnormalities, as well as for detection and investigation of previously unidentified recurrent changes.

In the first study we used a whole genome expression array to study and compare gene expression profiles in children and adults in order to find an explanation for the difference in outcome. However, no specific expression signatures, characteristic for the different age groups, were observed. Surprisingly, we detected some similarities in the gene expression pattern between infants with MLL gene translocation and older adults. We also observed that prognostically important cytogenetic aberrations determined the pattern of the expression profile and defined genes in which the expression pattern was characteristic of cytogenetic subgroups and was independent of age. Gene expression profiling turned out to be a sensitive tool and helped us to refine the karyotype in a few patients. Some recurrent chromosomal aberrations in ALL remain hard to discover and their prognostic importance is still unclear.

Intrachromosomal amplification of 21q is often detected as multiple RUNX1 gene signals on interphase FISH analysis, but the size of the amplicon and the mechanism of the amplification have not been thoroughly studied. We investigated eight patients with 21q amplification using array-CGH. A step-wise pattern of amplification was found in all patients, and the amplification was followed by a terminal deletion of 21q in some of them. All imbalances at the genome level were confirmed by FISH. Breakage/fusion/bridge cycles were suspected as the mechanism of amplification, and this hypothesis was supported by the observation of anaphase bridge structures including DNA from chromosome 21 in two cases. The same approach was used to characterize seven cases with dic(9;20). Although no identical breakpoint was found, clustering of the breakpoints, covering a 1.5 Mb segment of 9p13.2 and a 350 kb segment of 20q11.2, was observed. All cases were unbalanced with loss of 9p13.2-pter and five had loss of 20q11.2-qter, whereas two displayed gain of 20cen-pter. This means that the cytogenetic description of this abnormality should be dic(9;20)(p13.2;q11.2). It remains to be elucidated whether the rearrangement leads to a chimeric gene or if the functionally important outcome is loss of 9p and/or 20q.

In the fourth study, 28 patients with ALL and normal or failed karyotype on G-banding were studied with a 33K, tiling BAC array. This approach allowed us to revise the karyotypes in 75%, and genetic changes were found in 86% of the patients. Most of these patients showed copy number alterations that were below the resolution of Gbanding. The frequency of ETV6 gene deletion was found to be underestimated. In three patients no aberrations were found, which may be explained by the limitation of the method, absence of copy number alterations in these samples, e.g. balanced chromosome rearrangements, or other genetic mechanisms responsible for the development of leukemia.