Polycythemia Vera: Redefinition in the Genomic Era

Abstract 1754

For more than 100 years since the initial descriptions, polycythemia vera (PV) was defined by an aggregation of clinical and laboratory features, reported to be more common in males than females, and diagnosed on average at age 60 (Modan Blood 1965, Berlin Sem Hematol 1975). The 2005 discovery of somatic mutations in the JAK2 gene introduced the molecular era of PV and required redefinition of this disease entity. We established a prospective, observational cohort of 556 patients evaluated in our center between 2005 and 2011, of which 273 had a PV phase at some point during their myeloproliferative neoplasm (MPN). Serial samples were obtained from each patient for genomic analyses, including neutrophil JAK2V617F allele burdens, clinical karyotypes, SNP-array karyotypes, JAK2 and ASXL1 sequencing and copy number variation, allele burden analysis in sorted hematopoietic stem cell (HSC) fractions, and whole exome sequencing. These data were used to define the relationship of genotype to clinical phenotype with regard to PV epidemiology, natural history and disease transformation. Thirty three percent of the cohort was evaluated within 1 year from PV diagnosis and the median MPN disease duration at the last update of the cohort was 9 years (range 1–52 years). As of 7/2012, of the 273 PV patient cohort, 47 had antecedent essential thrombocytosis (ET/PV), 176 had PV, 43 had developed post-PV myelofibrosis (PPVMF) and 7 had developed acute leukemia (AML) (PPVAML). 270 of the 273 PV patients had JAK2 mutations, either V617F (264, 97%) or exon 12 (6, 2%); the remaining 3 (1%) are molecularly undefined. Women outnumbered men (169/104; ratio 1.6), even when stratified by ET/PV (2.1), PV (1.6), PPVMF (1.4) and PPVAML (1.3). Age at PV diagnosis was significantly younger in women, 54 (range 8–88), compared to men, 56.5 (range 15–77) (p=0.022), and the proportion diagnosed before age 40 was 26% in women compared to 10.5% in men. PPVMF occurred on average after 9 years (range 2–53 years) of PV at a median age of 62.5 years. PPVAML occurred on average after 10 years (range 3–28 years) of PV, at a median age of 71 years, significantly higher than the age at PPVMF (p=0.038). Aside from JAK2V617F, acquired 9pUPD was the most common genomic lesion in PV, occurring in 57% within the first year after PV diagnosis, in 84% of PPVMF and 100% of PPVAML patients. Studied prospectively, the prevalence of 9pUPD increased from 0 to 40% in 11 patients transitioning from ET to PV, and increased from 59% to 75% in 30 PV patients from year 1 to year 6 after diagnosis, but stayed at 90% in 11 patients pre and post transformation to PPVMF. Chromosomal loss/gain was not highly prevalent during PV (2%) in contrast to PPVMF (64%) and PPVAML (100%). The most frequent chromosomal abnormalities in PPVMF were trisomy 9 (27%), 13q deletion (12%), 1q gain(12%), 20qdeletion (8%) and 11qdel (8%), whereas the most common chromosomal abnormalities in PPVAML were 5qdel or −5 (75%), and 7qdeletion (50%), both of which were often found in the setting of complex changes (75%). Genomic lesions identified in PV and PPVMF, including JAK2V617F, 9pUPD, 11qdel, and ASXL1 mutations, were detected at high allele burdens by quantitative allele assays in flow-sorted, pluripotent HSCs. We conclude that acquisition of a JAK2 mutation is implicated in the vast majority (99%) of PV patients, that PV occurs more often in women, and that younger women (<40) particularly are at higher risk than younger men. Genomic lesions in PV and PPVMF arise and accumulate in a primitive HSC population. 9pUPD is a common occurrence during transition from JAK2V617F+ ET to PV, and while highly prevalent, age and time dependent in PV, 9pUPD is not sufficient to generate PPVMF or PPVAML. In PPVMF, JAK2 mutations associate with specific recurrent chromosomal changes that are also found in normal individuals with advancing age (9pUPD, 13qdel, 20qdel, 11qdel; Nature Genetics 44, 2012). JAK2 mutations with 9pUPD enhance the acquisition of age-associated and therapy- associated genomic instability lesions, promoting the development of PPVMF and PPVAML. Given the molecular epidemiology of PV, it will be crucial recognize and reduce the risk factors that lead to the excess acquisition of PV in young women, to identify the risk factors that lead to 9pUPD, to study whether targeted therapy can prevent the development of 9pUPD, and to avoid genotoxic therapy that accelerates genomic instability in PV.