Detection of Recurrent Uniparental Disomy and Cryptic Chromosomal Abnormalities in MDS/MPD-U and MDS/MPD-Derived Secondary AML

The WHO classification distinguishes MDS/MPD as a distinct entity. The JAK2 V617F mutation is present in a minority of these patients (pts). UPD9p characterizes pts homozygous for the V617F mutation. Other chromosomal abnormalities can also be detected in pts with typical MPD and MDS/MPD, and it is likely that cytogenetic methods with higher resolution could detect additional defects. We applied 250K SNP-arrays to examine genomic composition and identify previously cryptic defects and molecular abnormalities in pts with MDS/MPD-U and secondary AML (sAML) arising from MDS/MPD-U both in pts wild-type for V617F and those with the mutation. Any deletions, duplications, and/or UPD found by SNP-A in 76 controls or on available internet databases were considered copy number variants (CNV) and non-pathogenic. First, we used pts with typical MPD to assess the ability of SNP-A to detect UPD. All pts homozygous for V617F showed UPD9p by SNP-A. In pts with MDS/MPD, several additional cryptic lesions were detected, including segmental micro-deletions on chr 1, 5, 9, and 12. UPD was common, occurring on chromosomes other than 9 in 9/28 patients (32%, i.e. on chr 1,11,12). Shared/overlapping lesions (in >2 pts) included small segmental lesions on chr 7 (N=3), and a small cytoband (q14.1) of chr 11 (N=3). Overall, clonal lesions including segmental UPD were found in 23/28 (82%) pts by SNP-A in comparison to 17/28 (61%) by metaphase cytogenetics (MC). Pts with a history of MDS/MPD-U with (N=14) and without the JAK2 V617F mutation (N=14) were also analyzed. MC revealed chr aberrations in 10/14 (71%) of V617F+ pts, including common lesions such as +8 and del5q. With SNP-A, 1 additional pt with normal MC was found to have an abnormal karyotype (UPD 7 and 9), and 9/10 pts with abnormal MC had additional lesions previously undetected, including UPD on chrs other than 9 in 3/14 pts (21%). Examples of deleted regions include segmental losses within chrs 2, 7, 8 and 13, and UPD on chrs 1p, 7q, and 11. Likewise, additional lesions were identified in MDS/MPD-U pts negative for V617F. 8/14 pts (57%) showed abnormal MC; however SNP-A showed lesions in 12/14. In addition, 6/8 pts with abnormal MC had lesions in addition to those detected by MC. UPD was also common in V617F- pts, occurring in 6/14 (43%), predominantly on chr 11 (in 3/6 pts). No significant difference was found between the number or type of lesions found in pts with and without V617F mutation. SNP-A can also be used to identify lesions acquired during AML evolution. In 1 MPD pt at diagnosis, SNP-A showed UPD9p as a sole abnormality consistent with a homozygous V617F mutation. Upon transformation, repeated SNP-A showed a V617F- leukemic clone (normal chr 9) possessing microdeletions on chr 4 and 19. Similar evolution of a V617 negative leukemic clone was also observed in an MDS/MPD pt in whom a new UPD6 was detected. However, in 3 other MDS/MPD patients, SNP-A showed the presence of a V617+ leukemic clone (showing UPD9) in AML blasts. In summary, SNP-A-based karyotyping complements MC and allows for precise definition of chr aberrations in pts with MDS/MPD, including copy-neutral LOH. UPD is common in both JAK2 V617F+ and V617F- disorders, and is not restricted only to chr 9p, indicating other potential causative genes.