Update on the diagnosis and management of paroxysmal nocturnal hemoglobinuria

Abstract

Once suspected, the diagnosis of paroxysmal nocturnal hemoglobinuria (PNH) is straightforward when flow cytometric analysis of the peripheral blood reveals a population of glycosyl phosphatidylinositol anchor protein-deficient cells. But PNH is clinically heterogeneous, with some patients having a disease process characterized by florid intravascular, complement-mediated hemolysis, whereas in others, bone marrow failure dominates the clinical picture with modest or even no evidence of hemolysis observed. The clinical heterogeneity is due to the close, though incompletely understood, relationship between PNH and immune-mediated bone marrow failure, and that PNH is an acquired, nonmalignant clonal disease of the hematopoietic stem cells. Bone marrow failure complicates management of PNH because compromised erythropoiesis contributes, to a greater or lesser degree, to the anemia; in addition, the extent to which the mutant stem cell clone expands in an individual patient determines the magnitude of the hemolytic component of the disease. An understanding of the unique pathobiology of PNH in relationship both to complement physiology and immune-mediated bone marrow failure provides the basis for a systematic approach to management.

Figure 1.

The molecular basis of PNH. Normal hematopoietic stem cells express both transmembrane and GPI-anchored proteins (top). PNH stem cells express transmembrane proteins normally but fail to express GPI-APs because the first step in synthesis of the anchor is inactivated because the gene (PIGA) that encodes the enzyme that is required for transfer of the nucleotide sugar (UDP-GlcNAc) to phosphatidylinositol is mutant (middle). Of the more than 25 genes involved in synthesis of the GPI anchor, only PIGA is located on the X-chromosome (all others are autosomal). Location on the X-chromosome accounts for the observation that essentially all cases of PNH are due to somatic mutation of PIGA because inactivation of only 1 allele is required to produce the PNH phenotype as males have 1 X-chromosome and in females only 1 of the 2 X-chromosomes is active in somatic tissues (bottom). UDP, uridine diphosphate.

Figure 2.

Complement and PNH. The hemolysis of PNH is due to aberrant regulation of the APC. The APC is a component of the innate immune system. Unlike the classical pathway of complement that requires a recognition factor such as antibody to activate the pathway, the APC is continuously active. Therefore safeguards have evolved to protect host cells against APC-mediated injury. In the case of erythrocytes, 2 GPI-APs, CD55 and CD59, serve this function. Two enzymatic convertases amplify the activity of the APC (top). The C3 convertase consists of activated C3 (C3b), activated factor B (Bb, the enzymatic subunit of the complexes that is proteolytically activated by factor D, a trace plasma protein that may be activated by 1 of the mannose-binding lectin-associated serine proteases), and factor P (formerly called properdin). Factor P stabilizes the C3 convertase, allowing each convertase to activate many molecules of C3, and in the process, generate the weak anaphylatoxin, C3a. The C5 convertase is similar in structure to the C3 convertase except that 2 molecules of C3b are required to position C5 for cleavage by activated factor B (Bb). Many molecules of C5 are cleaved by the C5 convertase, and this process generates many molecules of the potent anaphylatoxin and neutrophil chemo-attractant, C5a. Activated C5 (C5b) is the nidus for formation of the MAC of complement consisting of C5b, C6, C7, C8, and multiple molecules of C9. The MAC inserts into the lipid bilayer of the cell, forming a transmembrane torus that results in osmotic lysis. CD55 (DAF) blocks the formation and stability of both the C3 and C5 convertases, whereas CD59 (MIRL) blocks formation of the cytolytic MAC, primarily by inhibiting binding and multiplicity of C9. Eculizumab is a humanized monoclonal anti-C5 antibody that prevents activation of C5 by the C5 convertase. Consequently, the MAC cannot form (and C5a is not generated), accounting for the inhibition of the intravascular hemolysis of PNH. However, eculizumab does not inhibit formation of the C3 convertase, accounting for the opsonization by activation and degradation products of C3 observed in patients with PNH treated with eculizumab. Normal RBCs are protected against APC-mediated injury (black crosses represent APC C3 and C5 convertase formation; yellow stars represent MAC formation) by CD55 (blue ovals) and CD59 (green ovals) (bottom). PNH cells lacking the complement inhibitory proteins CD55 and CD59 undergo complement-mediated lysis, releasing cellular contends including hemoglobin (red circles) and LDH into the plasma.

Figure 3.

Flow cytometry of peripheral blood cells from patients with either subclinical PNH, PNH/bone marrow failure (BMF), or classic PNH. (A) In these examples, 2-color flow cytometry is used to analyze RBCs (top) and neutrophils (PMNs) (bottom). RBCs are gated on based on staining with phycoerythrin (PE)-conjugated anti-glycophorin, and GPI-APs are gated on using a combination of fluorescein isothiocyanate–conjugated anti-CD55 and anti-CD59. Neutrophils (PMNs) (bottom) are gated on using PE-conjugated anti-CD11b. (Left, top and bottom) Analysis of RBCs and PMNs from a normal volunteer (control). All the normal RBCs and PMNs express CD55 and CD59. Patients with subclinical PNH (middle) have very small clones, typically less than 1% (middle), and they have no biochemical evidence of hemolysis. Patients with PNH/BMF (right) have at least biochemical evidence of hemolysis, but clone size is generally relatively small. In the case illustrated, the clone size of 21% is at the lower limit for biochemical detection of hemolysis. (B) (Right, top and bottom) Analysis of RBCs and PMNs from a patient with classic PNH. The percentage of cells that are deficient in expression of GPI-APs is shown the inner-upper quadrant of the histograms. The percentage of deficient RBCs is invariably greater than the percentage of deficient PMNs because GPI-AP–deficient RBCs are selectively destroyed by complement-mediated lysis, whereas GPI-AP–deficient PMNs have a normal life span. For this reason, the size of the PNH clone is determined by the percentage of GPI-AP–deficient PMNs rather than by the percentage of GPI-AP–deficient RBCs.

Figure 4.

Management of PNH. A management scheme based on classification of PNH into 3 subcategories (subclinical, PNH in the setting of another bone marrow failure syndrome [PNH/BMF], and classic PNH. See Table 2 for characteristics of each category.