Introduction: Paroxysmal nocturnal hemoglobinuria (PNH) is caused by somatic mutations of the X-linked gene, PIG-A, in hematopoietic stem cells (HSCs). The product of this gene is necessary for the assembly of glycosylphosphatidylinositol (GPI) anchors. Consequently, PNH cells lack the expression of GPI-anchored proteins on their cell surface. PIG-A mutations have been found in granulocytes and T lymphocytes from most normal individuals. Although the significance of these mutations is unclear, it suggests that they are important in the pathogenesis of PNH.

Methods: We isolated CD34+ progenitors from 4 PNH patients, 18 healthy donors, and 9 non-PNH patients undergoing peripheral blood stem cell tranplantation. The frequency of PIG-A mutant progenitors was determined by assaying for colony forming cells (CFC) in methylcellulose containing toxic doses of aerolysin. Aerolysin is a pore-forming toxin that uses the GPI anchor as it receptor; hence, PNH cells are unique in their resistance to aerolysin. DNA was extracted from individual day 14 aerolysin resistant CFC and the PIG-A gene was sequenced to determine clonality. We performed a Poisson distribution of the mutational frequency to determine the probability that the PIG-A mutation in controls arose from HSC.

Results: In PNH patients, 67% of the CFC were aerolysin resistant. The frequency of aerolysin resistant CFC was 14.7 ± 4.0 x 10−6 in the bone marrow of healthy donors and was 57.0 ± 6.7 x 10−6 from mobilized peripheral blood. Aerolysin resistant CFC from PNH patients exhibited clonalPIG-A mutations, and thus, arose from HSC. In contrast, PIG-A mutations in the CFC from controls were polyclonal (up to 15 different mutations from one individual). Recent evidence suggests that humans and other large mammals possess only 10,000 primitive HSC, and that only 1000 of these cells are thought to contribute to hematopoiesis at any one time. Poisson statistics show that only 5% of normals would be expected to harbor a PIG-A mutation in 1000 HSC, and < 1 x 10−9 persons would harbor 10 or more different PIG-A mutations even if all 10,000 hematopoietic stem cells were contributing to hematopoiesis. Thus, the high frequency of PIG-A mutations in controls, coupled with their polyclonality, suggests that they do not arise at the level of HSC; rather, PIG-A mutations in normals appear to arise as a consequence of hematopoietic differentiation, between the level of an HSC and a CFC.

Conclusion: Our data confirm the findings that PIG-A mutations are relatively common in normal hematopoiesis. Although we cannot rule out that a rare PIG-A mutant blood cell in normals does in fact arise from a mutant HSC, our data suggest that most of the mutations occur with differentiation. Genetic fidelity can be lost with differentiation without consequence, as mutations in differentiated cells would not be propagated. These data also call into question the relevance of PIG-A mutations in normals to the pathogenesis of PNH.

Author notes

Corresponding author