Focal Deletion of Genes Involved in the Control of Cell Cycle Progression Contributes to Growth Factor Independence in Acute Lymphoblastic Leukemia Cells

Abstract

B-lineage ALL, and chronic myeloid leukemia in B-lineage lymphoid blastic phase (CML-LBP), are characterized by deregulated proliferation of clonal B-precursor lymphoblasts. Cytogenetic aberrations, such as the t(9;22) translocation resulting in the bcr/abl fusion oncogene, play a critical role in leukemogenesis. However, full malignant transformation of B cell precursors likely requires secondary cytogenetic lesions. Recently, through array-based comparative genomic hybridization (array-CGH), recurrent submicroscopic cytogenetic deletions have been identified in the majority of ALL and CML-LBP. These deletions involve genes that control cell cycle progression such as BTG1 or RB1, or the p14ARF, p15INK4B or p16INK4A genes which are all encoded within the CDKN2A/B region. Alternative lesions involve genes that control lymphocyte development, such as IKAROS or PAX5. Little is known on the contribution of these deletions to the deregulated proliferation of ALL lymphoblasts, due to the limited availability of in vitro assays that allow manipulation of primary ALL blasts. We have established a serum- and growth factor-free in vitro system in which primary cells from 12 out of 34 ALL cases continuously proliferated for over 1 year. Leukemic cells from the other 22 cases survived in vitro for a significant period (>3 weeks) but did not divide significantly. Growth-factor independence was not restricted to a distinct cytogenetic subtype. Proliferating samples included 2 CML-LBP, 4 bcr/abl positive ALL, 1 etv6/abl positive ALL, 2 e2a-pbx1 positive ALL, 1 mll/enl positive ALL and 2 ALL cases with non-typical cytogenetics. To study whether growth factor independence correlated with submicroscopic lesions, we analyzed 10 in vitro proliferating and 10 non-proliferating samples on Agilent 44k CGH arrays. Seven of the 10 in vitro proliferating samples displayed a focal deletion (~500kb) of the CDKN2A/B locus at 9p21. Of these seven, three showed a focal (~150kb) deletion at the RB-1 locus on 13q14.2, two showed focal (~750kb) deletion at the BTG1 locus at 12q22, and one displayed focal deletions at both the RB1 and BTG1 loci. In the remaining three in vitro proliferating samples no submicroscopic deletions were detected. In the 10 non-proliferating samples, only 4 displayed deletions at the CDKN2A/B locus and no RB1 or BTG1 deletions were observed. Deletion of IKAROS was detected only in one of the proliferating samples that displayed RB1 deletion. No deletions at the PAX5 locus were detected. To confirm knock-out of the affected genes, and to study which of the three genes encoded by the CDKN2A/B locus were affected, we analyzed expression of full length transcripts in the primary blasts by RT-PCR. All cases that displayed deletions at the RB1 or BTG1 loci lacked RB1 or BTG1 transcripts, respectively, confirming homozygous deletion. The remaining cases expressed normal RB1 and BTG1 transcripts. Of the samples that showed deletions at the CDKN2A/B locus, two expressed p14, p15 and p16, suggesting hemizygous deletion, two only expressed p15 and p16, suggesting homozygous deletion of p14, and one only expressed p15, suggesting homozygous deletion of p14 as well as p16. Finally, two samples expressed none of the three transcripts, suggesting homozygous deletion of the entire CDKN2A/B region. Interestingly, homozygous deletion of RB1 and homozygous deletion of p15 or p16 were mutually exclusive, suggesting that either of these two events could suffice for deregulation of this pathway. After six months of continuous in vitro proliferation we again determined the status of BTG1, RB1, and the CDKN2A/B encoded genes in the 10 proliferating cell populations. No de novo RB1 or BTG1 deletions were observed. However, all 10 populations now lacked expression of one or more of the CDKN2A/B encoded genes. The majority of the new functional losses could be attributed to de novo deletions within the CDKN2A/B region, as determined by genomic PCR. In conclusion, our results provide evidence that deletion of genes that are involved in the control of cell cycle progression may decrease growth factor dependence of B lymphoblasts, and as such significantly contribute to leukemic transformation and/or clonal evolution.