Genome-Wide DNA Copy Number Analysis by SNP Arrays of B-Cell Chronic Lymphocytic Leukemia: Correlation with Known Biological and Molecular Prognostic Markers.

B-cell chronic lymphocytic leukemia (B-CLL) is a genetically heterogeneous disease with a variable clinical course. Chromosomal changes have been identified by FISH in approximately 80% of patients, and the presence of specific lesions, such as trisomy 12 and 13q14, 11q23, 17p13.1 and 6q23 deletions represent prognostic markers for disease progression and survival. In order to characterize further the complexity of B-CLL genomic lesions, we performed high density, single nucleotide polymorphism (SNP) array analysis in highly purified neoplastic cells (>92%) from a panel of 100 untreated, newly diagnosed patients (57 males and 43 females; age, median 63 years, range 30–87) in Binet stage A. All patients were investigated by FISH for the presence of trisomy 12 (21 cases); 13q14 deletion (44 cases, 34 as the sole abnormality); 11q22.3, 17p13.1 and 6q23 (15, 7 and 2 patients, respectively). In addition, ZAP-70 and CD38 expression resulted positive in 42 and 46 patients, whereas IgVH genes were mutated in 45 patients. Genome-wide DNA profiling data were generated on GeneChip^® Human Mapping 250K NspI arrays (Affymetrix); copy number alterations (CNA) were calculated using the DNA copy Bioconductor package, which looks for optimal breakpoints using circular binary segmentation (CBS) (Olshen et al, 2004). A total of 782 CNAs (ranging from 1 to 31 per sample, mean and median values 7.82 and 7, respectively) were detected; DNA losses (365/782=46.67% loss; 194/782=24.81% biallelic deletion) were found to be more frequent than gains (148/782=18.93% gain; 75/782=9.59% amplification). The most recurrent alterations detected by FISH were all confirmed by SNP array analysis, strengthening further the good reliability of such high-resolution technology. We identified 12 minimally altered regions (MARs) larger than 100 kb with a frequency higher than 5%. Among well known alterations, the largest was represented by chromosome 12 trisomy, followed by 6q, 17p and 11q23 deletions (32.87, 19.09 and 10.43 Mb, respectively) and 13q14 deletion (635 kb). Gain of 2p25.3 involves a common region of 4.39 Mb region in 7 patients, although it was extended to the whole short arm of chromosome 2 in 3 cases. Among those alterations previously described in B-CLL, we found losses at 14q32.33 (12 pts) and 22q11.2 (5 pts) involving the IGH and IGLλ loci, respectively. With regard to novel regions, we identified losses at 4q35.2 (5 pts) and 11q25 (6 pts). In addition we found a high frequency of losses/gains at 14q11.2 (42 pts) and 15q11.2 (33 pts), two genomic regions reported to be affected by DNA copy number variations in normal individuals. As regards correlations between CNAs and biological markers, we found that the number of CNAs is significantly higher in cases with unmutated IgVH (9.4; range 2–31) as compared with mutated IgVH (6; range 1–13) (p=0.002), while neither CD38 nor ZAP-70 expression showed significant correlation. In addition, a significant higher number of either CNAs (p=0.001), total MARs (p<0.0001) or even only novel MARs (p=0.009) was significantly associated with cases with 17p deletion or multiple cytogenetic aberrations as evaluated by FISH analysis. Our data indicate that genetic abnormalities involving chromosomal gains and losses are very common in early-stage B-CLL and further support the application of high resolution SNP array platforms in the characterization of genetic changes in the disease. In addition, we detected novel altered chromosomal regions that warrant further investigations to better define their pathogenetic and prognostic role in B-CLL.