A Role for TET2 Mutations in Paroxysmal Nocturnal Hemoglobinuria (PNH)

Abstract 1262

PNH is characterized by circulating blood cells that are deficient in the surface expression of proteins that require glycosylphosphatidylinositol (GPI). The loss of the GPI-linked complement inhibitors CD55 and CD59 on the red cell results in complement mediated hemolysis, and this defect on the surface of platelets is likely responsible for the marked hypercoagulable state seen in PNH. The loss of GPI-linked proteins is due to the clonal expansion of a stem cell with an acquired somatic mutation in the PIG-A gene, which is responsible for an early step in the production of GPI. In PNH, as in aplastic anemia, there is a decreased stem cell and erythroid progenitor pool, there are oligoclonal T cell expansions, there is an HLA allele association, and cytopenias can occur, which respond to immunosuppression. Normal individuals harbor occult populations of cells with the PNH phenotype and genotype, and there is evidence that PIG-A mutations are growth neutral in animals. These findings are all supportive of the immune escape model, which posits that the abnormal environment associated with marrow injury represents the “second hit” which selects in favor of the PNH clone. However, there is a long-standing interest in finding second genetic hits. Indeed, others have reported rearrangements of HMGA2 in PNH; we have reported a series where 24% of patients had a cytogenetic abnormality and another series where 3 out of 29 patients with PNH had the JAK2V617F mutation and an MPN/PNH overlap syndrome. Here we have investigated a series of 17 patients with PNH to see whether we would find mutations in TET2, which is commonly mutated in myeloid disorders. We separated granulocytes from patients with at least 40% GPI-negative granulocytes, extracted DNA, which was subjected to whole genome amplification, followed by bi-directional sequencing using a dye terminator approach. In 11 out of 17 patients, we identified the 5284A>G, I1762V variant, with an allele frequency of 41% (95% confidence interval 25% to 58%) compared with 22% in the NCBI dbSNP database. In 4 of the patients we identified the 5162T>G, L1721W variant, with an allele frequency of 12%, which was not significantly different from the NCBI dbSNP database (9%). In 3 patients, we identified the 1088C>T, P363L variant, with an allele frequency of 9%, which also was not significantly different from the database (3%). However, one remarkable patient was heterozygous for all three of these SNPs– and was also heterozygous for a previously undescribed nonsense mutation, 2697T>A, Y899X. This mutation was seen in 3 separate sequencing reactions, and the alleles were present in approximately a 1:1 ratio. This mutation occurs at the 3' end of exon 3, a region where chain terminating mutations have been previously reported. Of note, it has been recently reported that heterozygous disruption of Tet2 in a mouse model results in an increase in the stem cell compartment as determined by a competitive repopulation assay, suggesting that this patient's TET2 mutation may have partly contributed to the advantage that the PNH clone demonstrated. Because 100% of her granulocytes were GPI–negative, it was not possible to be certain which abnormality came first. This patient's diagnosis of severe hemolytic PNH had been first confirmed 13 years ago, and she has a prior history of intra-abdominal thromboses and a DVT. She is now doing well on both coumadin and eculizumab. Unlike patients with a concurrent JAK2 mutation, she does not have features of an MPN, but her blood counts are higher than most of the other patients in this series, with a WBC of 7.7 compared with a median of 3.7 for the other patients, a platelet count of 321 compared with a median of 131, and a reticulocyte count of 247,000 compared with a median of 132,000. In conclusion, inactivating TET2 mutations, like activating JAK2 mutations, may contribute to the expansion of PNH clones in a subset of patients.