High Resolution Genomic Profiling Using Single Nucleotide Polymorphism Microarrays Reveals Novel Genomic Lesions in Hairy Cell Leukaemia and Hairy-Cell Leukaemia Variant

Hairy-cell leukaemia (HCL) is a rare B-cell disorder, with a variant form (HCL-v), which differs from the former in morphology, immunophenotype, clinical behaviour and response to treatment. Whilst HCL is highly sensitive to purine analogues, HCL-v patients do not respond to these treatments and have a median survival of only 7 years compared to over 20 years in HCL. The relationship between these two disorders is yet to be determined and the molecular mechanisms underlying them are still largely unknown. Genomic studies are limited due to the rarity of the disease and have mainly been performed at the cytogenetic level. We have used single nucleotide polymorphism (SNP) arrays to examine DNA copy number changes and loss of heterozygosity (LOH) in 14 HCL and 15 HCL-v cases with the aim of finding genetic markers that may distinguish both disease types and can explain their different clinical outcome. DNA extracted from PBMC samples enriched for tumour cells (HC >84%) was interrogated using the Affymetrix Human Mapping 500K Array Set. Data were analysed using CNAG 2.0 and dCHIP 2006 to identify regions of genomic imbalance and LOH by comparison to a set of 25 normal controls. DNA copy number abnormalities (excluding the IGH and IGL heterogeneic loci) were identified in 11/14 (78%) HCL and 13/15 (87%) HCL-v cases. The median number of lesions per patient in HCL (2.5, range=0–14) was half that observed for HCL-v (5, range=0–34) (p=0.1, not significant) with 3/15 cases of HCL-v exhibiting more than 20 aberrations. The most significant difference between HCL and HCL-v was the occurrence of deletions on 17p, involving monoallelic loss of TP53. Such losses were evident in 33% (5/15) of HCL-v cases compared to 0% of HCL cases (p=0.02). Sequencing analysis detected mutations in the non-deleted allele in 3/5 of these cases, suggesting biallelic loss of TP53 function. Other frequent events were present in both HCL and HCL-v and included gains on chromosome 5 (1/14 HCL and 5/15 HCL-v) and deletions on 7q (3/14 HCL and 3/15 HCL-v) and 14q (2/14 HCL and 1/15 HCL-v). Similar aberrations have been detected in previous cytogenetic studies of typical HCL but the high resolution of the SNP arrays has allowed us to further define these regions of imbalance in both disorders. Chromosome 5 abnormalities included trisomy 5 in two cases and a minimal region of gain, 18.9 Mb in size (5q34–q35.3), was present in 4/6 cases. The cases exhibiting 7q deletions allowed the minimal region to be further defined to 4.1 Mb (7q31.31–q31.33), a region containing approximately 25 known genes. Deletion of 14q24.1–q32.13 was seen in two HCL cases, therefore further minimising the 14q22–q32 region identified previously. One HCL-v case exhibited an interesting region of deletion at 14q32.32, consisting of a focal homozygous deletion flanked by hemizygous loss, containing candidate tumour suppressor genes that may be involved in HCL-v pathogenesis. In addition to these larger lesions, focal gains, harbouring only a single gene, were identified in both disorders at 19p13.2 (3/14 HCL, 3/15 HCL-v), 20p13 (4/14 HCL, 3/15 HCL-v) and 17q21.2 (4/16 HCL and 5/15 HCL-v). High-resolution copy number analyses allow the identification of novel genetic lesions in HCL and HCL-v and further define known cytogenetic aberrations. Our data show that HCL and HCL-v are generally similar at the genomic level but can be distinguished based on 17p deletions, suggesting that biallelic loss of TP53 may contribute to the greater genomic instability and to the aggressiveness of HCL-v. Other HCL-v cases may carry aberrations of other genes involved in DNA damage/repair pathways aside from TP53 and will be investigated.