In paroxysmal nocturnal hemoglobinuria (PNH), acquired somatic mutations in the PIG-A gene give rise to clonal populations of red blood cells unable to express proteins linked to the membrane by a glycosylphosphatidylinositol anchor. These proteins include the complement inhibitors CD55 and CD59, and this explains the hypersensitivity to complement of red cells in PNH patients, manifested by intravascular hemolysis. The factors that determine to what extent mutant clones expand have not yet been pinpointed; it has been suggested that existing PNH clones may have a conditional growth advantage depending on some factor (e.g., autoimmune) present in the marrow environment of PNH patients. Using f low cytometric analysis of granulocytes, we now have identified cells that have the PNH phenotype, at an average frequency of 22 per million (range 10-51 per million) in nine normal individuals. These rare cells were collected by f low sorting, and exons 2 and 6 of the PIG-A gene were amplified by nested PCR. We found PIG-A mutations in six cases: four missense, one frameshift, and one nonsense mutation. PNH red blood cells also were identified at a frequency of eight per million. Thus, small clones with PIG-A mutations exist commonly in normal individuals, showing clearly that PIG-A gene mutations are not sufficient for the development of PNH. Because PIG-A encodes an enzyme essential for the expression of a host of surface proteins, the PIG-A gene provides a highly sensitive system for the study of somatic mutations in hematopoietic cells.

In paroxysmal nocturnal haemoglobinuria (PNH), somatic mutation of the PIG-A gene is thought to result in altered expression of glycosylphosphatidylinositol (GPI)-anchored proteins. This study was performed to determine if there were any heterogeneities of cellular phenotypes between two major peripheral blood cells, erythrocytes and granulocytes. Using CD59-based immunocytometry, the patterns of CD59 expression were shown to be conserved in the circulating erythroid cells (reticulocytes and mature erythrocytes) in all 29 patients with PNH. Twenty-one patients had distinct combinations of PNH type I, II, and III cells in different lineages. Only eight patients exhibited similar patterns of CD59 expression between the two lineages. Approximately one third of the patients had PNH type II cells in either or both of the two lineages indicating variable lineage involvement. The proportion of abnormal granulocytes was higher than those of abnormal reticulocytes and erythrocytes. In patients with appropriate erythropoietic responses to haemolysis (RPI > 2.0), shift reticulocytes display predominantly PNH phenotypes. These immature erythroid cells with altered expression of GPI-anchored proteins may dominate the peripheral blood during periods of increased marrow activity resulting in greater phenotypic mosaicism in such patients. Discrepancies in expression of GPI-anchored proteins in PNH which are highly variable between the two lineages may be the result of their different life spans and the influence of complement-mediated cytolysis. The phenomena also indicated the possible occurrence of more than one PNH clones with variable clonal dominance.