Functional validation of recurrent genomic alterations in pediatric acute leukemia

Members of the Downing laboratory previously used a combination of array-based comparative genomic hybridization and targeted gene re-sequencing to identify a number of recurrent mutations in pediatric acute lymphoid leukemia (ALL). The most commonly targeted genes were PAX5, EBF1 , and IKZF. A number of other rarer lesions were also identified, including focal deletions on chromosome 3 that resulted in the loss of two genes, CD200 and BTLA. By mapping the breakpoint of these chromosome 3 deletions in primary ALL blasts and cell lines, I demonstrated that the deletion resulted in the loss of both CD200 and BTLA in all but one case. Moreover, the results of analyzing the sequence at the breakpoints suggested that deletion was the result of aberrant RAG-mediated recombination. Importantly, no point mutations in CD200 or BTLA were detected in any of the analyzed pediatric ALL cases, suggesting that haploinsufficiency of these two genes was likely to contribute to leukemogenesis.;To explore the functional consequence of these deletions, I first analyzed the expression pattern of both genes as a function of normal B-cell development. In agreement with published studies, both genes showed low expression during early B-cell development but were markedly upregulated in mature B cells. To explore whether loss of both genes contributes to leukemogenesis, I crossed mice heterozygous for a Cd200-null allele (Cd200 +/-) with mice heterozygous for a Btla-null allele (Btla +/-) to generate mice that were heterozygous for both genes (Cd200+/-: Btla+/-). I then examined B-cell development in these mice. My analysis revealed no significant alteration in B-cell development in mice heterozygous for either Cd200 or Btla alone. In contrast, the Cd200+/-: Btla+/- mice had a subtle decrease in mature B cells and a modest increase in pre-B cells, suggesting that loss of both genes cooperated to dysregulate normal B-cell development. To determine if this defect could cooperate with other genetic lesions in leukemogenesis, bone marrow cells from wild-type, Cd200 +/-, Btla +/-, and Cd200+/-: Btla+/- mice were transduced with a retrovirus expressing BCR-ABL1 and GFP and then transplanted into lethally irradiated syngeneic mice. As previously published, BCR-ABL1 induces a pre-B cell acute lymphoblastic leukemia in wild-type mice. No significant difference was noted in the penetrance, latency, or disease phenotype of bone marrow cells obtained from mice heterozygous for Cd200 or Btla. In contrast, expression of BCR-ABL1 in marrow obtained from Cd200+/-: Btla+/- mice resulted in a significant decrease in latency and increase in penetrance of ALL without affecting the immunophenotype of the resultant leukemia. These data suggest that the co-deletion of CD200 and BTLA induces a subtle block in normal B-cell development and can cooperate with the expression of BCR-ABL1 in leukemogenesis.;In a second series of experiments, I focused on exploring the role of a focal recurrent amplification of chromosome 8q24.21 in AML that was initially identified in the Downing laboratory. The minimally altered region contains a single putative gene called CCDC26 that was initially identified as the target of retroviral integration in a forward genetic screen to identify genes whose alteration blocked retinoic acid-induced myeloid cell differentiation. This region is amplified in approximately 4% of AMLs and is also affected by focal amplifications or deletions in rare cases of pediatric ALL. To further characterize this deletion, I used a variety of approaches, including tiling array CGH, whole-genome sequencing, and transcriptome sequencing of mRNA and nuclear and cytoplasmic large-intervening non-coding RNAs (Linc-RNAs), to determine the target of these leukemia-associated copy-number alterations (CNAs). Our analysis revealed the expression of altered splice variants of CCDC26 in leukemia cells that lacked an open reading frame and showed no change in expression as a result of CNAs. These data suggest that CCDC26 is not the primary target of the CNAs. Our transcriptome analysis, coupled with detailed bioinformatic analysis of public databases of chromatin modification in leukemic cells, identified a 1.2-kb unspliced nuclear Linc-RNA whose expression was increased in leukemia cell lines and primary AML samples that contained a focal 8q24.21 amplification. Knockdown of this Linc-RNA with two independent antisense oligonucleotides resulted in a subtle decrease in cell viability. The complexity of the Linc-RNAs encoded by the genomic region, coupled with the lack of homology between mice and humans in this genomic region, makes it very challenging to define the role of this region in leukemogenesis. (Abstract shortened by UMI.)...