Relapsed acute lymphoblastic leukemia (ALL) is one of the leading causes of death among children with cancer. While children who relapse late (≥ 3 years from initial diagnosis) fare better than those who relapse early in treatment, the prognosis for these children remains poor, even with aggressive treatment. Thus new approaches to prevent and treat relapsed ALL are needed. To discover the underlying biological pathways that may play a role in drug resistance and relapse, we integrated three high-throughput assays.
Using matched diagnosis/relapse bone marrow samples from children with early relapsed Pre B-ALL (n=10) and T-ALL (n=6) we evaluated gene expression, DNA methylation and DNA copy number alterations (CNAs) analysis. All diagnosis and relapsed samples had more than 80% bone marrow blast account and were initially treated on BFM protocol. As controls, we used flow-sorted normal B cell progenitor subpopulations and CD4+CD8+ T cell samples purified from thymus. Gene expression profile of RNA samples were detected by Illumina HumanHT-12v4 ExpressionChip and bisulfite converted DNA materials were hybridized to the Infinium HumanMethylation450K BeadChip. DNA copy number and LOH data were obtained using Illumina HumanCytoSNP-12. Analysis variant CpG sites in an unsupervised manner, we identified three clearly distinct DNA methylome profiles in samples. T-ALL cases clustered separately from B-ALL samples and B-ALL samples clustered two branch according to their structural variations. We performed supervised comparisons of methylation data between B-ALL and normal B cells, as well as comparing B-ALL with T-ALL cases and T-ALL cases with normal T cells. Differentially methylated regions were identified using limma and lumi packages. ALL cases did not show significant differences between paired diagnose/ relapse samples though they had hyper DNA methylation than control samples. Aberrant DNA methylation had been detected in negative regulator of cell cycle genes and DNA damage repair genes in ALL. Unsupervised expression analysis of the pairs, T-ALL cases clustered separately from B-ALL samples. We defined a signature of 5,762 probe sets that differentially expressed between samples and control. Matched-pair analyses revealed significant differences in the expression of genes involved in cell-cycle regulation, and apoptosis between diagnostic and early relapse samples. Copy number analysis of patients revealed varying numbers of genetic lesions ranging from 0 to 45 CNAs per sample. The vast majority of CNAs observed were shared between diagnosis and relapse in the same patients. For the B-ALL majority of the copy number changes were gross, beside that T-ALL samples mostly had cryptic lesions. One of the most frequent CNAs involved deletions of CDKN2A/B at 9p21.3, occurring in 8 patients, 2 of 8 patients lost the CDKN2A/B at relapse time. CDKN2A, a benchmark gene and had been found hypermethylated at relapse parallel to recent studies.
Early relapse samples were more likely to be similar to their respective diagnostic sample that suggests early-relapse results from the emergence of a related clone. Differential expression profile had been observed between relapse and diagnose samples that was independent from the methylation profile. By combining three high-throughput platforms, we demonstrated that both methylation and genomic alterations contribute to evolution of relapse and drug resistance.
Karakas:Novartis: Honoraria, Research Funding. Sayitoglu:Roche Diagnostics: Research Support Other.
Asterisk with author names denotes non-ASH members.