High-Resolution SNP-Array Profiling Discloses Novel Genomic Aberrations in BCR/ABL-Negative Myeloproliferative Neoplasms with Myelofibrosis

Although the recent identification of the gain-of-function mutations JAK2 V617F and MPL W515L/K provided new insights into the pathogenesis of myeloproliferative neoplasms (MPN), the biological origins of myelofibrosis (MF) are complex and remain poorly understood. Until today, it is unclear whether JAK2 or MPL mutations are causative lesions of bone marrow fibrosis in MPN patients (pts). Approximately 50% of the pts lack evidence of clonality such as gene mutations or an abnormal karyotype. In addition, the clinical course of pts with MF differs from primary polycythemia vera (PV) or essential thrombocythemia (ET) pts. Therefore, it is most likely that other unknown genetic events contribute to the disease phenotype. To identify novel disease-related aberrations in this subset of MPN, we applied single nucleotide polymorphism (SNP) arrays [Affymetrix 250K Nsp arrays] that allow for genome-wide screening of both DNA copy number alterations (CNAs) and copy neutral losses of heterozygosity (uniparental disomies, UPDs) at high resolution. For SNP-array mapping, DNA from granulocytes of 61 clinically well characterized MPN pts in fibrotic stage was studied [primary MF (PMF), n=47; post-ET MF, n=6; post-PV MF, n=8]. Genotypes were analyzed using CNAG 2.0 software. Data were normalized against an own set of 30 reference samples. Genomic regions recently detected as copy number polymorphisms were excluded from data analysis. CNAs were identified in 55% of PMF, 33% of post-ET MF, and 62% of post-PV MF cases. The most frequent recurrent CNAs were losses of 20q11-q13 (n=7), followed by trisomy 9 (n=5), trisomy 8 (n=4), gains of 1q (n=2), and micro-deletions in 17q11.2 (n=2; 1.3 and 2.3 Mb, respectively). Deletions in 20q and trisomy 9 were detected in all MF subgroups, whereas gains of 1q and trisomy 8 were restricted to PMF. Furthermore, 17q11.2 micro-deletions occurred in secondary MF pts (one post-ET and one post-PV). In both cases, monoallelic loss of the tumor suppressor gene NF1 was confirmed by fluorescence in-situ hybridization (FISH). In addition, 18 more non-recurrent micro-deletions ranging from 0.1 to 1.9 Mb in size were identified in 14 PMF cases and one post-PV MF pt. For instance, one PMF case exhibited a micro-deletion in 13q12.12 (1.4 Mb) encompassing TNFRSF19 (TNF-receptor superfamily member 19) that encodes a receptor which is capable of inducing apoptosis by a caspase-independent mechanism. UPDs were detectable in 27% of PMF, 17% of post-ET MF, and 62% post-PV MF pts. The most frequent UPDs included the JAK2 locus in 9p24 (n=9; 11.6 to 38 Mb), followed by UPDs in 7q31.1-q33 (24.3 Mb), 7q34-q35 (5.6 Mb), and 17q23.3-q24.2 (4.8 Mb) in two cases, each. Interestingly, all pts of the post-PV MF cohort exhibited chromosome 9 abnormalities, either by trisomy 9 (n=3) or by 9p UPD (n=5). Furthermore, all pts with trisomy 9 were heterozygously JAK2 V617F mutated, whereas all cases with 9p UPD revealed homozygous JAK2 mutations. In conclusion, our data on a large series of well defined fibrotic MPN cases demonstrate that array SNP-mapping is an excellent tool to identify known and novel genomic aberrations, thereby pinpointing to regions that harbor potential disease-related genes. Furthermore, the large number of micro-deletions in single MF cases presume a broad genetic heterogeneity and complexity of fibrotic MPN that has to be further investigated in larger pts series.