Array CGH Analysis of Chronic Lymphocytic Leukemia Reveals Frequent Cryptic Monoallelic and Biallelic Deletions of Chromosome 22q11.

Bacterial artificial chromosome array-based comparative genomic hybridization (array CGH) was used to detect recurrent genomic imbalances in 187 CLL cases. Detection rates of clinically relevant chromosomal abnormalities involving losses of 11q23, 13q14, 17p13, and whole chromosome 12 gains were consistent with previously reported results. However, submicroscopic deletions of chromosome 22q11 were observed in 28 cases (15%), and the frequency of these deletions was second only to loss of the 13q14 region, the most commonly detected recurrent chromosomal aberration in CLL. The del(22)(q11) occurred as the sole genomic alteration in three cases, but was associated with additional copy number changes in 25 cases, most commonly with the 13q14 deletion, where it was found in 17 of the 28 cases (61%). Twenty-four cases exhibited monoallelic 22q11 losses, four cases showed biallelic losses, and two cases also exhibited concomitant gains of whole chromosome 22.

Breakpoint mapping using oligonucleotide-based array CGH analysis showed that the 22q11 deletions ranged in size from 0.34 Mb up to approximately 1 Mb, and all cases included a minimally deleted 340,279 bp region located within the lambda immunoglobulin light chain variable region locus. The minimally deleted region included four protein coding genes, including the closely related zinc finger protein genes ZNF280A and ZNF280B, the gamma-glutamyltransferase light chain 2 (GGTLC2) gene, and the preferentially expressed antigen in melanoma (PRAME) locus. Quantitative real-time PCR expression studies for these four genes revealed that the expression of ZNF280A, ZNF280B, and PRAME mRNA was significantly lower in the 22q11 deletion cases compared to non-deleted cases by both the two sample t-test and the Wilcoxon rank sum test methods. Mean expression of ZNF280A mRNA was 26-fold lower in the del(22) (q11) samples compared to non-deleted samples (t-test, p=0.02612; Wilcoxon rank sum test, p=0.002511). Mean expression of ZNF280B mRNA was decreased by 11-fold in the del(22)(q11) samples compared to non-deleted samples (t-test, p=0.01442; Wilcoxon rank sum test, p=9.768x10⁻⁶). Mean expression of PRAME mRNA was 11-fold lower in the del(22)(q11) cases compared to non-deleted cases (t-test, p=0.005517; Wilcoxon rank sum test, p=0.0317). However, mean expression of GGTLC2 mRNA was only 1.6-fold lower in the del(22)(q11) cases compared to non-deleted cases. This change was marginally significant by t-test (p=0.06718) and not significantly different by the Wilcoxon rank sum test (p=0.1329). CLL cases with a biallelic del(22)(q11) exhibited the lowest levels of expression of ZNF280A, ZNF280B, and PRAME mRNA.

The ZNF280A and ZNF280B genes are closely related zinc finger proteins which exhibit significant sequence similarity to the Drosophila Suppressor of Hairy-wing (Su[Hw]) protein, a leucine zipper protein that contributes to the regulation of gene expression in the Drosophila genome through the establishment of endogenous insulators (independent transcriptional domains). To date, there is no direct evidence linking these two zinc finger proteins with human cancers. However, it is conceivable that alterations in the copy number of these genes could have a significant effect on gene regulation in CLL, and possibly other human cancers.

The PRAME gene encodes a membrane- and cytoplasm-associated protein that was first described as a tumor antigen in melanoma that triggered autologous cytotoxic T-cellmediated immune responses. High levels of PRAME mRNA expression have been described in a number of human malignancies, including CLL. High PRAME expression is present in acute leukemias with a more favorable prognosis such as AML with the t(8;21), AML-M3 with the t(15;17), and childhood B-ALL with or without the t(12;21). These findings have made PRAME an attractive candidate for immunotherapy and a potentially useful marker for minimal residual disease monitoring. Our study is the first to correlate submicroscopic deletions of the PRAME locus with decreased expression of PRAME mRNA in CLL. The lack of reciprocal duplications and the presence of biallelic deletions also point to a possible role for loss of the PRAME gene locus in the pathogenesis of CLL.

Further studies are needed to assess the prognostic significance of submicroscopic 22q11 deletions and to determine whether these alterations contribute to the pathogenesis of CLL.