Targeting Genomic Heterogeneity of Adult Acute Lymphoblastic Leukemia

Introduction: In acute lymphoblastic leukemia (ALL) genetic diversity at diagnosis has been reported to be correlated with an increased likelihood of emerging drug resistance and an increased relapse risk.

Aim: To evaluate 1. the frequency of chromosome abnormalities at initial diagnosis and their acquisition during the course of the disease (clonal evolution), 2. the pattern of acquired genetic abnormalities.

Patients: We investigated 11 adult ALL patients (age 18.9-76.7 years, median 42.6 years) with paired samples at initial diagnosis and at relapse by chromosome banding analysis (CBA), array-based comparative genomic hybridization (aCGH) and molecular analysis. The cohort consisted of 7 B-ALL and 4 T-ALL cases. CBA was performed in both diagnostic and relapse samples in 10 cases. aCGH was performed in 6 cases (12x270 K microarray slides, Roche Nimblegen, Madison, Wl) both at initial diagnosis and at relapse. Molecular analysis was performed in all cases by 454 amplicon deep-sequencing (Roche, Branford, CT) in the following genes: AKT1, FBXW7, IL7R, JAK1, NOTCH1, PHF6, PTEN and WT1 in T-ALL cases, and in CRLF2, IL7R, IKZF1, JAK2, LEF1 and PAX5 in B-ALL cases at diagnosis and relapse.

Results: Median duration from diagnosis to relapse was 10 months (range 4-18 months). Regarding cytogenetics of B-ALL, 2 cases had t(9;22)(q34;q11), 1 case a MYC-rearrangement, 3 cases complex aberrations, 1 case a translocation t(5;16)(p15;q13) and 1 case had a normal karyotype (NK). In T-ALL, 2 cases had NK, 1 case had t(10;14) (q24;q11) and in 1 case no cytogenetic data was available.

At initial diagnosis, 8/10 cases (80.0%) with available CBA had one cytogenetic clone, while 2/10 (20.0%) cases had 2 cytogenetically related subclones. At relapse, one case retained both clones already detected at diagnosis. The second case showed loss of genetic aberrations in one subclone (clonal regression), while the second clone was stable. Evolution of genetically related clones was detected by CBA in 5/10 cases. The acquired abnormalities comprized unbalanced (n=8) or balanced (n=3) translocations, gains (n=3) or deletions (n=2) of whole or partial chromosomes. 3 of these 5 cases in addition showed clonal regression which may, however, be due to impaired in vitro proliferation of the subclone at relapse. The lost abnormalities comprized unbalanced (n=6) or balanced (n=1) translocations, gains (n=1) or deletions (n=1) of whole or partial chromosomes.

In all 6 cases analyzed by aCGH, clonal evolution was detected. In one case, without clonal evolution detected by CBA, aCGH revealed clonal evolution on the submicroscopic level. In detail, 92 copy number alterations (CNA) were observed (mean 8 per case, range: -33) at initial diagnosis and 105 CNA at relapse (mean 9 per case; range 2-36). Losses were more frequent than gains (79 vs 13 at initial diagnosis; 87 vs 18 at relapse). 53 CNA of initial diagnosis and 33 of relapse were larger than 10 Mbp and were already detected by CBA accordingly. 43 CNA at initial diagnosis (mean: 7 per case; range 4-11) and 40 CNA at relapse (mean 7 per case; range 3-11) were smaller than 10 Mbp and thus below the resolution of CBA. The most frequent CNA were loss of 9p21 (CDKN2A, n=5) and 7p12 (IKZF1, n=3). These stayed stable from initial diagnosis to relapse. Apart from these, no specific pattern of gains or losses of distinct chromosomal regions was observed. In 4/6 cases, clonal regression of 5 CNA was detected on submicroscopic level. By sequencing 17 mutations and 2 variants were observed in 9/11 patients without major changes between initial diagnosis and relapse. In B-ALL cases, mutations were detected in JAK2 (n=2) and IKZF1 (n=3). In 2 cases a variant in CRLF2 was observed, which cannot be assigned as mutation or SNP according to current knowledge. In T-ALL cases, mutations were detected in NOTCH1 (n=3), PHF6 (n=1), PTEN (n=1), WT1 (n=1) and IL7R (n=1). The pattern of mutations was stable in 7/9 cases. One case lost a mutation in NOTCH1 at relapse, while another patient gained a mutation in IKZF1. Two patients without mutations at diagnosis also showed no mutations at relapse.

Conclusions: 1. Adult ALL is characterized by multiple genetic aberrations a subset of which can be detected by aCGH only. 2. The genomic heterogeneity of adult ALL increases at relapse. 3. Based on this limited set of patients gene mutations seem to be more stable between diagnosis and relapse compared to cytogenetic aberrations.