SEARCH

SEARCH BY CITATION

Keywords:

  • Paroxysmal nocturnal hemoglobinuria;
  • Anemia;
  • Hemolytic;
  • Glycosyl phosphatidylinositols;
  • Hematopoiesis;
  • Hematopoietic stem cell;
  • CD55;
  • CD59;
  • PIG-A;
  • X chromosome

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

The purpose of this review is to summarize recent studies that have led to a more complete understanding of the molecular basis of paroxysmal nocturnal hemoglobinuria (PNH). Somatic mutations of PIG-A arising in pluripotent hematopoietic stem cells are necessary for the development of PNH. PIG-A is an X-linked gene that is essential for synthesis of the glycosyl phosphatidylinositol (GPI) moiety that serves as a membrane anchor for a functionally diverse group of cell surface proteins. Consequently, the progeny of stem cells with mutant PIG-A are deficient in all GPI-anchored proteins (GPI-AP). Among the GPI-AP that are expressed on hematopoietic cells are two important regulators of the complement system, decay-accelerating factor, (CD55) and membrane inhibitor of reactive lysis, (CD59). It is the deficiency of erythrocyte CD55 and CD59 that accounts for the intravascular hemolysis and hemoglobinuria that are the clinical hallmarks of PNH. A remarkable feature of PNH is that the peripheral blood is a mosaic composed of variable proportions of GPI-AP+ and GPI-AP cells and that, in an individual patient, the GPI-AP cells can be derived from multiple mutant stem cells. Currently, however, there is no evidence that the PIG-A mutation per se provides a proliferative advantage. Thus, PNH is not a monoclonal disease with a malignant phenotype. Rather, the mutant stem cells appear to dominate hematopoiesis because under some pathological conditions, GPI-AP deficiency is advantageous. The close association of PNH with aplastic anemia suggests that the selection pressure arises as a consequence of a specific type of bone marrow injury.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired disease of the pluripotent hematopoietic stem cell. Although the pathogenesis is speculative, there is incontrovertible evidence that somatic mutation of a specific gene (PIG-A) is a necessary component of the disease process. Clinically, PNH is characterized by intermittent, recurrent episodes of hemoglobinuria, thrombophilia and abnormal hematopoiesis [1,, 2]. The hemoglobinuria is a consequence of intravascular hemolysis that is due to the abnormal sensitivity of PNH erythrocytes to the lytic action of complement. Compelling evidence indicates that the aberrant regulation of complement is caused by a deficiency of membrane proteins that restricts the activity of both the C3 convertase and the membrane attack complex [3]. The protein that modulates the C3 convertase is decay-accelerating factor (DAF), (CD55) [4], and the protein that regulates the membrane attack complex of complement is membrane inhibitor of reactive lysis (MIRL), (CD59) [5]. DAF (CD55) and MIRL (CD59) share the common biochemical feature of being anchored to the cell by a glycosyl phosphatidylinositol (GPI) moiety [6]. In addition to these two complement regulatory elements, the hematopoietic cells of PNH are deficient in at least 15 other membrane constituents [7,, 8]. These proteins are functionally diverse, but all are GPI-anchored. This observation supports the following paradigm: All proteins that are deficient in PNH are GPI-anchored, and all GPI-anchored proteins (GPI-AP) that are expressed by hematopoietic cells are deficient in PNH. Implicit in this paradigm is the suggestion that the gene that is mutant in PNH is essential for the normal biosynthesis of the GPI anchor.

A gene that complements the abnormal expression of GPI-AP in a PNH cell line was identified by Takeda et al. [9]. The gene was named PIG-A (for phosphatidylinositol glycan-class A), and it encodes a protein that is essential for the normal synthesis of N-acetylglucosaminyl-phosphatidylinositol (GIcNAc-PI), an early intermediate in the pathway of GPI anchor assembly [10]. Together, these studies defined both the biochemical and molecular bases of the deficiency of GPI-AP in PNH. Somatic mutations in PIG-A have been found in all cases (n = 72) of PNH reported to date [8]. Additional studies by Takeda et al. [9] localized PIG-A to the short arm of the X chromosome (Xp22.1). The molecular basis of PNH is illustrated in Fig. 1. Because males have one X chromosome and the somatic cells of females are functionally haploid due to X inactivation, mutation of only one gene is needed to produce the abnormal phenotype in hematopoietic tissues. As anticipated from these observations, females are affected by PNH as commonly as males [11,, 12]. The absence of inherited cases of PNH due to deficiency of all GPI-AP [13] implies that functionally significant germ line mutations of PIG-A are lethal. Because the somatic mutation in PNH affects the pluripotent stem cell, progeny that are deficient in GPI-AP can be identified in all hematopoietic lineages (i.e., granulocytes, monocytes, platelets, erythrocytes and lymphocytes).

thumbnail image

Figure Figure 1.. The PNH defect.Transmembrane proteins have three domains: an ectoplasmic domain (inline image), a transmembrane domain (), and a cytoplasmic domain (inline image). In contrast, GPI anchored proteins lack the cytoplasmic and transmembrane domains. This class of proteins is anchored to the cell by a GPI moiety consisting of phosphatidylinositol (inline image), glucosamine (inline image), and three mannose (inline image) molecules. The GPI moiety is linked to the COOH-terminus of the protein portion of the molecule by ethanolamine. The PIG-A gene product is essential for the transfer of the nucleotide sugar uridine diphosphate-N-acetylglucosamine (UDP-GIcNAc) to phosphatidylinositol (PI) to form GIcNAc-PI, the first intermediate in the synthesis of the GPI anchor. PIG-A is located on the X chromosome. Hematopoietic cells in PNH are deficient in all proteins that are GPI-anchored because a somatic mutation occurring in a pluripotent stem cell partially or completely inactivates the PIG-A gene product. Consequently, the GPI moiety is not synthesized.

Download figure to PowerPoint

Phenotypic Mosaicism in PNH

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Much Madness is divinest Sense -

To a discerning Eye -

   Emily Dickinson

   #435

A defining characteristic of PNH is phenotypic mosaicism. This remarkable feature was first recognized by Rosse in 1972 [14]. In 1966, he and Dacie [15] reported the development of a method for quantitating the sensitivity of human erythrocytes to complement-mediated lysis. Using this assay, Rosse [14] identified three discrete populations of erythrocytes in patients with PNH. Based on differences in complement sensitivity, the following three groups were defined: PNH I erythrocytes showed normal (or relatively normal) sensitivity to complement; PNH II erythrocytes were moderately sensitive to complement, requiring one-third to one-fourth as much serum (as the complement source) for an equal degree of lysis compared to normal erythrocytes; and PNH III erythrocytes were exquisitely sensitive to complement, requiring one-fifteenth to one-twenty-fifth the amount of serum needed to lyse an equal proportion of normal cells. Additional studies showed that essentially all patients had more than one population of PNH erythrocytes and that the percentage of PNH I, PNH II and PNH III cells varied greatly among patients [14,, 16,, 17]. Based on this classification, the following four patterns were recognized: Pattern 1, PNH I + PNH III; Pattern 2, PNH II + PNH III; Pattern 3, PNH I + PNH II + PNH III; Pattern 4, PNH I + PNH II [14]. Collectively, these experiments [14] provided the first evidence of phenotypic mosaicism in PNH.

Holguin et al. [18] defined the basis of the differences in sensitivity of PNH erythrocytes to complement-mediated lysis. Those investigators showed that PNH III erythrocytes are completely deficient in DAF (CD55) and MIRL (CD59), whereas PNH II cells are partially deficient and PNH I cells have normal expression. Thus, the variability in sensitivity to lysis is explained by quantitative differences in expression of the GPI-anchored complement regulatory proteins, DAF (CD55) and MIRL (CD59). Analysis by flow cytometry of the erythrocytes of a large group of PNH patients (n = 50) showed heterogeneous expression of GPI-AP in 72% [19], demonstrating that phenotypic mosaicism is common in PNH. Recently, we have undertaken studies designed to determine the molecular basis of this phenotypic mosaicism [20]. The results of those studies will be reviewed here because they have important implications about the origins of PNH.

By using flow cytometry (Table 1), a patient with PNH whose erythrocytes showed heterogeneity of expression of GPI-AP was identified (Fig. 2). T lymphocyte clones were isolated from the peripheral blood of the patient, and analysis of expression of GPI-AP revealed four distinct phenotypes (Fig. 3). Some of the clones had normal expression of both DAF (CD55) and MIRL (CD59), and this phenotype (represented by clone 2D10) was designated (++). Some clones showed slightly discordant expression of DAF (CD55) and MIRL (CD59), with DAF (CD55) expression being borderline positive and MIRL (CD59) expression being primarily negative. This phenotype (represented by clone 2Dl) was designated (±–). Some clones showed borderline positive expression of DAF (CD55) and MIRL (CD59), and this phenotype (represented by 1M9) was designated (±±). Finally, DAF (CD55) and MIRL (CD59) expression was negative for other clones, and this phenotype (represented by 2D39) was designated (– –).

Table Table 1.. PNH phenotypes based on sensitivity to complement and on flow cytometric analysis of GPI-AP expression
Phenotypic DesignationComplement SensitivityaGPI-AP Expression by Flow Cytometryb
  • a

    aBased on the complement lysis sensitivity assay of Rosse and Dacie [14,, 15].

  • b

    bPhenotypic characterization based on flow cytometry is illustrated in Figures 2 and 3, Figure 3..

PNH Inormalnormal
   
PNH IImoderately sensitive (3 to 4 times moredim positive (± ±)
 sensitive than normal) 
   
PNH IIImarkedly sensitive (15 to 20 times morenegative (– –) or (± –)
 sensitive than normal) 
thumbnail image

Figure Figure 2.. Phenotypic mosaicism in PNH.Erythrocytes from a patient with PNH and from a normal volunteer (NL) were analyzed by flow cytometry using anti-DAF (CD55) or anti-MIRL (CD59) as the primary antibody. The histogram of the erythrocytes from the normal donor shows uniformly positive staining with both antibodies. In contrast, the patient's histograms (PNH) suggest three discrete populations of cells (a negative population, a population with partial expression and a population with normal expression). Statistical analysis of the three groups of cells from the patient showed that the negative population contributed 14% to the total, the intermediate population contributed 75% and the normal population contributed 11%. (Modified from data published in [35], by copyright permission of Blackwell Science, Ltd.).

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Analysis by flow cytometry of PNH T cell clones.The T cell clones were derived from the peripheral blood cells of a PNH patient with phenotypic mosaicism (shown in Figure2). Based on expression of DAF (CD55) and MIRL (CD59), four distinct phenotypes were identified. The (+ +) phenotype indicates normal expression of both DAF and MIRL; (± –) indicates borderline positive expression of DAF and primarily negative expression of MIRL; (± ±) indicates borderline positive expression of both DAF and MIRL; (– –) indicates negative expression of both DAF and MIRL. (Modified from data published in [20], by copyright permission of W.B. Saunders Company).

Download figure to PowerPoint

The molecular basis of these phenotypes was determined by nucleotide sequencing of PIG-A cDNA and genomic DNA. No somatic mutations were found among the clones with the (++) phenotype, demonstrating that these cells were progeny of stem cells with normal PIG-A. In contrast, four discrete PIG-A mutations were found among the clones with the abnormal phenotypes (Figure 4). Clones with the (– –) phenotype were found to have a 737 bp deletion involving part of the promoter [21], all of exon I and part of intron 1. Clones with the (±±) phenotype had a missense mutation at nucleotide position 304 that changed threonine at codon 102 to proline. Presumably, this substitution affects the functional activity of the PIG-A protein. Support for this conclusion is provided by the observation that the threonine at this position is conserved in both murine PIG-A and SPT-14 from Saccharomyces cerevisiae, the putative yeast homolog of PIG-A [23]. Clones with the (±–) phenotype were found to have two distinct mutations. In one example, a thymine deletion at position 564 was identified. In the other example, a guanine deletion involving the first nucleotide of exon IV was observed. In both cases, these deletions produced frame shifts that introduced premature stop codons. These results demonstrate that the phenotypic mosaicism that is a characteristic feature of PNH is due to genotypic mosaicism (Figure 5).

thumbnail image

Figure Figure 4.. Schematic representation ofPIG-Amutations identified in T cell clones derived from the PNH patient with phenotypic mosaicism.The PIG-A exons are indicated by the Roman numerals. Introns are omitted, except for intron 1 (thin black line between exons I and II). Intron 1 is not drawn to scale. The approximate size of the intron is 3.1 kb [22]. The phenotype of the T cell clones (Figure3) in which each mutation was identified is shown in parentheses. The large deletional mutation that gives rise to cells with the (– –) phenotype deletes 137 bp of the promoter region [21], all of exon I (23 bp) and 577 bp of the 5′ end of intron 1. A missense mutation (A-304-C) gives rise to the (± ±) phenotype. That nucleotide substitution converts threonine at codon 102 to proline. Two clones with the same phenotype (± –) were found to have discrete mutations. In both cases, the single nucleotide deletion caused a frame shift that generated a premature stop codon.

Download figure to PowerPoint

thumbnail image

Figure Figure 5.. Schematic model of the molecular basis of phenotypic mosaicism in PNH.Cells with normal expression of GPI-AP (the PNH I phenotype, Table1) are derived from stem cells with normal PIG-A. In patients with heterogeneous expression of GPI-AP, more than one abnormal clone is present, and the progeny of some of the abnormal clones is phenotypically different because the PIG-A genotype is different. For example, PNH II cells (± ± phenotype) are the progeny of stem cells with mutations that incompletely inactivate the PIG-A protein, whereas PNH III cells (– – phenotype) are progeny of stem cells with mutations that completely inactivate PIG-A. Inasmuch as any mutation that completely inactives PIG-A produces PNH III cells, phenotypically identical cells can have different PIG-A genotypes.

Download figure to PowerPoint

Others have also observed more than one PIG-A mutation in cells from individual patients with PNH [24–, 26]. Further, Ostendorf et al. [27] reported two mutations involving the same PIG-A gene (A-98-G and an 11 bp deletion involving nucleotides 102-112). If these mutations arose independently, it would suggest that PIG-A was targeted for mutation. Because of their proximity within PIG-A, however, it seems more likely that they arose as a consequence of the same mutational event [28].

Evidence of Polyclonality in PNH

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Studies by others have suggested that PNH is a monoclonal disease [29–, 31]. If this interpretation is correct, all of the abnormal PNH T cell clones identified in the patient described above would have to arise from a common ancestor. To investigate this hypothesis, X chromosomal inactivation among the four PNH T clones was determined [20]. Those experiments demonstrated that the pattern of X chromosome inactivation was not the same for all of the clones; therefore, all of the T cell clones with mutant PIG-A could not have arisen from a common ancestor. Thus, at least in this patient, PNH is a polyclonal rather than a monoclonal disease. Further support for this conclusion is provided by the observation that since four separate PIG-A mutations were present in this patient, the mutational events must have occurred independently rather than by clonal evolution.

That PNH is not a monoclonal disease suggests that the PIG-A mutation does not endow the mutant stem cell with an absolute growth advantage. Thus, PNH differs from monoclonal hematopoietic stem cell disorders such as chronic myelogenous leukemia, in which the translocation that results in the formation of P210bcr-abl is sufficient to cause an otherwise normal hematopoietic stem cell to undergo transformation. The polyclonal nature of PNH suggests a process in which GPI-AP-deficient stem cells dominate hematopoiesis because they have a relative rather than an absolute proliferative or survival advantage. In this case, GPI-AP+ stem cells are the targets of a pathophysiological process. The ultimate consequence of this process is the “negative selection” of GPI-AP stem cells. This hypothesis will be discussed in more detail below.

Clonal Dominance in PNH

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Our studies also showed that not all of the abnormal stem cells contributed equally to the peripheral blood phenotype of the patient. For example, cells with the (±±) phenotype comprised approximately 75% of the circulating erythrocytes (Fig. 2), and by using restriction analysis of polymerase chain reaction (PCR)-amplified genomic DNA, we found that the vast majority of the neutrophils, monocytes and lymphocytes were heterozygous for the same PIG-A mutation that was identified in the T cell clone with the (±±) phenotype [20]. These results demonstrate clonal dominance by the stem cell with the (±±) phenotype (Fig. 6). Available evidence, however, supports the hypothesis that clonal dominance is determined not by the type of PIG-A mutation, but by the intrinsic proliferative properties of the affected stem cells. This conclusion is based on the observation that PNH patients can have two separate mutations that completely inactivate PIG-A, but the contribution to hematopoiesis of the two affected stem cells may be quantitatively very different [25,, 32]. Further, while in our patient, cells with the (±±) phenotype (PNH II) were dominant relative to cells with the (– –) phenotype (PNH III). In other patients with a mixture of PNH II and PNH III cells, the PNH III cells are the dominant phenotype [33]. Thus, while PIG-A genotype determines the PNH phenotype, it does not appear to determine the proliferative properties of the affected stem cell.

thumbnail image

Figure Figure 6.. Model of the clonal dominance hypothesis.Not all mutant stem cells contribute equally to the peripheral blood phenotype. For example, the stem cell with the missense mutation that results in progeny with the (± ±) phenotype contributes approximately five times as many cells to the peripheral blood than the other three mutant stem cells combined (Figure2). Currently there is no evidence that either the type of PIG-A mutation or the PNH phenotype determines the proliferative capacity of affected stem cells. Thus, clonal dominance appears to be a manifestation of differences in the intrinsic proliferative properties of the affected stem cells independent of the PIG-A mutation.

Download figure to PowerPoint

That the peripheral blood phenotype of patients with PNH remains consistent over years of observation [1,, 11] suggests that the mutant stem cells are capable of long-term self-renewal and proliferation. Whether the mutant stem cells are actually transformed is an unresolved issue. One of the difficulties in resolving this issue is that the life span, self-renewal, and proliferative properties of normal hematopoietic stem cells are incompletely understood. Available evidence, however, does not support the concept of clonal dominance as a component of normal hematopoiesis [34].

Only 18 of the 84 (21%) reported PIG-A mutations are missense mutations that could potentially cause partial inactivation of the protein product [8]. The majority of mutations (66 of 84, 79%) are small deletions or insertions, splice site mutants, nonsense mutants, insertion/deletions and large deletions that are expected to cause complete inactivation of the PIG-A protein [8,, 26]. Based on these findings, Nafa et al. [26] postulated that the preponderance of null mutations indicates that complete deficiency of GPI-AP is relatively more advantageous than a partial deficiency. The observation that PNH II cells can dominate hematopoiesis relative to PNH III cells (Fig. 2), however, argues against this hypothesis. Thus, it seems likely that the PNH III phenotype is more common than the PNH II phenotype, not because of a functional advantage derived from complete deficiency of GPI-AP, but rather because there are many types of mutations that can inactivate the gene product, whereas relatively few types of mutations result in partial inactivation of the PIG-A protein. For example, deletions, insertions and single nucleotide substitutions that introduce premature stop codons (nonsense mutations) or modify consensus splice sites produce null mutations. In contrast, mutations that partially inactivate the protein product are largely confined to single nucleotide substitutions that change an amino acid that is functionally important. Thus, we conclude that, with respect to hematopoietic stem cell survival in PNH, complete GPI-AP deficiency is not relatively more advantageous than partial GPI-AP deficiency (and vice versa). Clearly, this situation does not extend to survival of erythrocytes. In this case, partial expression of DAF (CD55) and MIRL (CD59) is sufficient to provide significant protection of PNH II cells from complement-mediated destruction in vivo [1,, 18,, 33].

Identification of a Rare Sequence Variant in PIG-A

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

In addition to the four different somatic mutations, we identified a constitutional PIG-A mutation in our patient with PNH [35]. In this case, a single nucleotide substitution (cytosine to thymine at position 55, C-55-T) was observed. This substitution caused replacement of arginine with tryptophan at codon 19. Restriction analysis of PCR-amplified genomic DNA using AvaI and direct genomic sequencing showed that both of the patient's PIG-A alleles were affected by this mutation. Inasmuch as the patient was homozygous, we postulated that the mutation was inherited rather than acquired. Further, because the C-55-T transition was observed in all of her T cell clones, including those with the (++) phenotype (Fig. 3), we hypothesized that the mutation was functionally silent. This hypothesis was supported by the observation that neither the patient's parents nor siblings had clinical evidence of PNH. AvaI restriction analysis of PCR-amplified genomic DNA and cDNA showed that the patient's mother was heterozygous for C-55-T, while the father was hemizygous. Based on these results, we concluded that the patient inherited both mutant alleles. Flow cytometric analysis of the peripheral blood cells of the patient's parents showed normal expression of GPI-AP, confirming that the mutation is functionally silent.

Interestingly, the C-55-T mutation has been reported in four other patients with PNH [26,, 36], bringing the total to 5 of 72 (7%). In each case where C-55-T was observed, additional somatic mutations were identified that accounted for the abnormal GPI-AP expression. Of the five reported cases of C-55-T in PNH, Nafa et al. [26] reported three. They also identified C-55-T in three males who had no clinical evidence of PNH (in one case, a negative Ham's Test and normal GPI-AP protein expression were documented). However, Nafa et al. [26] did not include information on how many volunteers were screened in order to identify the three C-55-T mutations in individuals without PNH. Thus, the frequency in the general population of this sequence variant is unknown. We have analyzed DNA from the neutrophils of 30 volunteers, and no C-55-T mutants have been identified. Additional volunteers and patients with PNH must be screened in order to determine the frequency of C-55-T in the general population, and whether there is linkage disequilibrium between C-55-T and PNH.

From the above discussion, it is clear that C-55-T alone does not produce the PNH phenotype; however, it is not clear whether C-55-T increases the probability of developing PNH. In our patient who is homozygous for C-55-T, somatic mutations involving PIG-A were identified on both alleles [20]. In one of the cases reported by Nafa et al. [26], the patient was a female who was heterozygous for C-55-T. In that case, the somatic mutation that produced the abnormal phenotype affected the PIG-A allele bearing C-55-T. In the other two PNH patients with constitutional C-55-T reported by Nafa et al. [26], information about gender was not included. Nagarajan et al. [36] reported a PNH patient with C-55-T and T-762-A involving the same PIG-A gene (the latter mutation introduced a premature stop codon at amino acid position 254). Those investigators did not include information about whether C-55-T was constitutional, although the probability seems high based on our results [35] and those of Nafa et al. [26]. Collectively these observations suggest that in most (if not all) cases, C-55-T is present on the same PIG-A gene as the somatic mutation, suggesting that the constitutional mutation may contribute to the development of PNH, perhaps by increasing the mutational frequency of PIG-A or by stabilizing a somatically mutated gene.

As discussed by Nafa et al. [26], C-55-T involves a CpG dinucleotide. Potentially, these dinucleotides are mutational hot spots because the cytosine is the substrate for DNA methyltransferase [37]. Methylcytosine is susceptible to deamination, and hence, conversion to thymine. If cytosine at position 55 of PIG-A is a mutational hot spot, it could explain the appearance of this sequence variant in the general population.

Aplastic Anemia (AA) and PNH

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Patients with PNH may be divided into the following two groups: those without a preceding history of AA (primary PNH), and those with an antecedent history of AA who subsequently develop PNH (AA/PNH). An association between AA and PNH has been long-recognized [38] and frequently confirmed [39–, 44]. Recently, Hillmen et al. [38] reported that 23 of 80 PNH patients (29%) had an antecedent history of AA. The proportion of patients with AA who subsequently develop PNH varies widely (13% to 57%) among studies, in part, because the criteria for diagnosis of PNH are not uniform (in some cases a positive Ham's test or sucrose lysis test was required, and in other cases, identification of a population of peripheral blood cells with GPI-AP deficiency was required). In two recent reports, 38% (11 of 29 patients) and 32% (12 of 37 patients), respectively, of patients with AA treated with immunosuppressive therapy developed laboratory evidence of PNH [43,, 44].

The time between the diagnosis of AA and the development of PNH varies from a few months to several years [36,, 45]. In most patients with AA/PNH, however, there is no clinical or laboratory evidence of PNH at the time the AA is diagnosed, suggesting that GPI-AP-deficient stem cells are not present initially, but develop during the recovery phase of the disease. This observation has important implications for the etiology of PNH because it suggests that stem cells with PIG-A mutations are absent from the marrow at presentation and develop from the injured stem cells during a period of proliferation (Fig. 7, Model B).

thumbnail image

Figure Figure 7.. Two models of PNH.In model A, stem cells with mutant PIG-A are present in the bone marrow prior to the injury process. The mutant stem cells survive the pathological event because GPI-AP deficiency provides a selective advantage. During the recovery phase, the abnormal stem cells expand, and clinical evidence of PNH becomes apparent. In model B, the injury process depletes the stem cell pool and renders the surviving stem cells genetically unstable. During the recovery phase, PIG-A mutations develop in the rapidly proliferating, genetically unstable stem cells. The GPI-AP-deficient stem cells subsequently dominate hematopoiesis because the underlying pathological process eliminates the GPI-AP+stem cells.

Download figure to PowerPoint

Although essentially all patients with AA who develop PNH are treated with immunosuppressive therapy (e.g., antithymocyte globulin) at the time the AA is diagnosed, there is no evidence that immunosuppression causes PNH. Further, for patients with AA who recover clinically following immunosuppressive therapy, the risk of developing myelodysplastic disease or acute leukemia, or both, is similar to that of developing PNH [46]. This observation supports the hypothesis that AA is the consequence of a process that permanently injures the hematopoietic stem cell and that the injury results in genetic instability. Of note, patients with AA who undergo allogeneic bone marrow transplantation do not have an increased risk of developing PNH or other clonal myelopathies [46], suggesting that ablation of the abnormal stem cells eliminates the mutagenic potential.

The basis of the relationship between PNH and AA is obscure. Most patients with PNH have some evidence of bone marrow failure (e.g., thrombocytopenia or leukopenia) during the course of their disease [1]. While it is possible that GPI-AP are essential for normal hematopoiesis, it seems more likely that bone marrow injury is essential for the development of PNH. Tichelli and colleagues [47] reported that, compared to patients with AA who did not develop PNH, patients with AA who subsequently developed PNH had a higher mean corpuscular volume, higher granulocyte count and more dyserythropoiesis at the time the AA was diagnosed, as well as lower hemoglobin and an increased proportion of erythroblasts in the bone marrow after clinical recovery. Those studies suggest that AA is a heterogeneous disease and that the AA/PNH syndrome may have a specific etiology. Currently, there is no evidence that the types of PIG-A mutations that occur in AA/PNH are different from those observed in primary PNH [36]. Further, a distinction between primary PNH and AA/PNH may be artificial, since the underlying pathophysiological process could be the same with the aplastic or hypoplastic component of primary PNH being subclinical, and the recovery phase being spontaneous.

Etiology of PNH

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

We will now discuss in a little more detail the Struggle for Existence.

   Charles Darwin

   On the Origin of Species

Despite the great progress that has been made in determining the basis of the abnormal sensitivity of the erythrocytes to complement-mediated lysis and the global deficiency of hematopoietic cell GPI-AP, an issue that is fundamental to a more complete understanding of PNH remains unresolved. In order for PNH to become clinically evident, the hematopoietic stem cells bearing the mutant PIG-A must expand so that progeny sufficient to produce symptoms and signs of the disease is generated. In most instances, GPI-AP cells dominate hematopoiesis in patients with PNH [10], suggesting that the mutant stem cell has either an abnormally high proliferative capacity or a relative survival advantage compared to GPI-AP+ stem cells. Conceivably, the PIG-A mutation per se is advantageous, although available evidence does not support this hypothesis. For example, hematopoietic stem cells from patients with AA/PNH have markedly abnormal growth characteristics in vitro [48]. However, when stem cells from these patients are separated into GPI-AP and GPI-AP+ populations, differences in the growth characteristics are not apparent [48]. Further, studies using chimeric mice with both GPI-AP and GPI-AP+ stem cells showed that the ratio of GPI-AP:GPI-AP+ peripheral blood cells did not increase during several months of observation (personal communication, Dr. Junji Takeda, Osaka University, Osaka, Japan). These results support the concept that in PNH, GPI-AP stem cells dominate hematopoiesis, not because they have an absolute growth advantage, but because they have a conditional survival or proliferative advantage relative to GPI-AP+ stem cells. Thus, available evidence suggests that PIG-A mutation is necessary, but not sufficient, for the development of PNH.

The mechanism by which the mutant stem cells come to dominate hematopoiesis in PNH is speculative; however, it seems plausible to suggest that under some conditions, GPI-AP deficiency is advantageous [34,, 49]. For example, marrow injury or infection could initiate an immune-mediated process (cellular or humoral) in which the target antigen is a GPI-AP (of either host or foreign origin) expressed on hematopoietic stem cells. Under those circumstances, PIG-A mutant stem cells would escape injury and consequently dominate hematopoiesis because they lack the target antigen. Hematopoietic cells express a relatively large number of functionally diverse GPI-AP [7]. Thus, it seems reasonable to suggest that if the target antigen is a host protein, the selective advantage is due to the absence of a single GPI-AP, and that the reason for the global deficiency of GPI-AP is that PIG-A is located on the X chromosome. According to this hypothesis, the gene that encodes the detrimental GPI-AP (the target antigen) is autosomal. Inasmuch as two alleles rather than one must be mutated, the probability of inactivating an autosomal gene through somatic mutagenesis is remote compared to the probability of inactivating an X-linked gene. Therefore, stem cells with a deficiency of the detrimental GPI-AP are most likely to arise as a consequence of PIG-A mutations. Assuming that the GPI-AP complement regulatory proteins are not the targets of the underlying pathological process, the hemolytic anemia that is the clinical hallmark of PNH may represent an epiphenomenon related to the chromosomal location of PIG-A [50].

Support for the hypothesis that an immune attack against a specific GPI-AP exerts sufficient selective pressure to cause the emergence of a population of GPI-AP cells was provided by a recent study [51] that showed that patients with lymphoma treated with anti-CD52 (the Campath 1 antigen, a GPI-AP) developed a population of GPI-AP-deficient peripheral blood lymphocytes and monocytes. Analysis of PIG-A from cloned GPI-AP T lymphocytes of one of these patients showed a nonsense mutation [52], demonstrating that a mutation involving PIG-A rather than the gene that encodes CD52 produced the population of cells with the selective advantage. This situation is not entirely analogous to PNH, however, in that the abnormal lymphocytes and monocytes did not arise from a mutant pluripotent hematopoietic stem cell, since erythrocytes from these patients expressed GPI-AP normally [51]. As expected, once the treatment with anti-CD52 was discontinued, the GPI-AP peripheral blood cells disappeared. Hillmen et al. [38] reported that of 35 PNH patients who survived for 10 years or more, 12 had spontaneous clinical recovery. These remissions could be due to resolution of the pathological process that selects for GPI-AP deficiency. If, however, the injury process depletes the number of GPI-AP+ stem cells, and if the residual GPI-AP+ stem cells have the same growth properties as the GPI-AP stem cells, clinical and laboratory evidence of improvement may not be observed despite alleviation of the selective pressure. This hypothesis may explain why most of the long-term survivors have persistent evidence of GPI-AP cells [38].

Another potential basis for selection of PIG-A mutant stem cells can be envisioned if a GPI-AP functions as a receptor for a factor that acts as a negative regulator of hematopoiesis. Under these circumstances, GPI-AP-deficient stem cells would have a growth advantage because they lack the receptor for the inhibitor. In support of this hypothesis, there is evidence that GPI-AP can signal through interactions with tyrosine kinases [53,, 54].

Although unlikely, it is conceivable that the GPI-AP deficiency is not functionally advantageous in PNH. According to this hypothesis, the PIG-A mutations occur in a setting of marrow injury where the pool of stem cells is depleted and the frequency of DNA damage is increased. There are at least two reasons why PIG-A mutations are likely to be observed in this setting. First, as noted above, since PIG-A is located on the X chromosome [9,, 50], only one allele must be mutated in order to produce the phenotype. Second, the clinical consequences (i.e., complement-mediated intravascular hemolysis and hemoglobinuria) of GPI-AP deficiency are readily apparent. Further, in the setting of stem cell injury, it is not necessary to evoke a proliferative advantage for the GPI-AP deficiency, since under these circumstances, the stem cell pool is depleted and hematopoiesis is maintained by a small number of cells. Thus, mutation of PIG-A affecting a pluripotent cell that dominates hematopoiesis would give the appearance of a proliferative advantage bestowed by the GPI-AP deficiency when, in fact, the GPI-AP deficiency might be functionally neutral with respect to its effects on growth or survival of the affected stem cell. An instance in which the PIG-A mutation provides no obvious advantage is when PNH arises in association with a clonal myelopathy [1,, 55]. For example, PNH has been reported in patients with myeloproliferative diseases [1,, 55]. Under this circumstance, the abnormal clone has an absolute rather than a relative proliferative advantage that is determined by the underlying neoplastic process, and the clinical manifestations of PNH become apparent because the PIG-A mutation affects a stem cell in which growth is unregulated. The major problem with the hypothesis that GPI-AP deficiency in PNH is not functionally advantageous is that it does not provide an explanation for the observation that patients can have multiple mutant stem cells [20,, 24–, 26]. For multiple mutant stem cells to dominate hematopoiesis without some selective pressure, it would be necessary to suggest that PIG-A is specifically targeted for mutation.

Based on available evidence, two models for the development of PNH are proposed (Fig. 7). The primary difference between these two schemes is that in Model A, the GPI-AP stem cells are a component of normal bone marrow that arise as a result of spontaneous mutation of PIG-A. In this model, the marrow injury process depletes the GPI-AP+ stem cells. During the recovery phase, the GPI-AP stem cells proliferate, and signs and symptoms of PNH become apparent. This model works best for primary PNH. It does not work well, however, for AA/PNH because clinical and laboratory evidence of GPI-AP-deficient stem cells is usually not apparent at the time the AA is diagnosed [54]. If the injury process selects for GPI-AP deficiency, stem cells with mutant PIG-A should be most evident at the time of diagnosis, rather than during the recovery period when the selective pressure is presumably ameliorated by immunosuppressive therapy.

Model B (Fig. 7) is tenable for both primary PNH and AA/PNH. In this model, injury to the bone marrow depletes the stem cell pool and renders the residual cells functionally abnormal and genetically unstable. During the recovery phase, GPI-AP-deficient stem cells appear as a consequence of spontaneous mutation of PIG-A, and subsequently dominate hematopoiesis because they have a relative proliferative or survival advantage due to persistence of the underlying pathological process that induced the initial marrow injury. While this model incorporates an element of genetic instability to account for the multiple PIG-A mutants that are observed in some patients with PNH, hypermutability may not be a necessary element because stem cell proliferation is enhanced during the recovery phase. Thus, multiple PIG-A mutants could accrue despite normal mutational frequency if the stem cell proliferative rate is high enough. In support of this hypothesis, the bone marrow is often hypercellular in PNH [1,, 47], and mutational hot spots on PIG-A have not been identified [8].

Future Directions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

To understand PNH more completely, the following two processes must be explained: (1) the basis of the selective advantage owing to GPI-AP deficiency and (2) the basis of the clonal dominance manifested by some GPI-AP-deficient stem cells. In addition to illuminating the underlying pathophysiology of PNH, understanding these processes may provide important new insights into the regulation of hematopoiesis. For example, clonally dominant PNH stem cells appear to combine a high proliferative rate with the capacity for long-term self-renewal, and yet, they lack other characteristics associated with hematologic malignancy. Thus, an understanding of the molecular basis of the clonal dominance may suggest methods for developing a nonmaligant stem cell with a high proliferative rate that is capable of long-term self-renewal. Such stem cells would be useful in gene therapy experiments.

PNH has been a source of fascination for generations of researchers dating back to its original descriptions by Gull in 1866 and Strübing in 1882 [56]. Despite the great progress that has been made in defining both the molecular basis of the hemolytic anemia and the genetic mutation that accounts for the deficiency of GPI-AP, there is still much to learn about this uncommon (but not rare) disease [8]. It seems almost certain that a more detailed understanding of PNH will provide unanticipated, remarkable new information about the mechanisms that regulate proliferation and differentiation of the hematopoietic stem cell.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References

Work that originated in the laboratory of the author was supported by PHS grants R55 DK35830, K04 DKO1942, and P50 DK49219. The author thanks Dr. E. M. Prystas (LDS Hospital, Salt Lake City, Utah) for critically reviewing the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Phenotypic Mosaicism in PNH
  5. Evidence of Polyclonality in PNH
  6. Clonal Dominance in PNH
  7. Identification of a Rare Sequence Variant in PIG-A
  8. Aplastic Anemia (AA) and PNH
  9. Etiology of PNH
  10. Future Directions
  11. Acknowledgements
  12. References
  • 1
    Rosse WF. Paroxysmal nocturnal hemoglobinuria. In: RosseWF, eds. Clinical Immunohematology: Basic Concepts and Clinical Applications. Boston: Blackwell Scientific Publications, 1990:593593.
  • 2
    Parker CJ. Paroxysmal nocturnal hemoglobinuria and complement-mediated red cell damage. Curr Opin Hematol 1994; 1:151156.
  • 3
    Wilcox LA, Ezzell JL, Bemshaw NJ, et al. Molecular basis for the enhanced susceptibility of the erythrocytes of paroxysmal nocturnal hemoglobinuria to hemolysis in acidified serum. Blood 1991; 78:820829.
  • 4
    Nicholson-Weller A. Decay accelerating factor. Curr Top Microbiol Immunol 1992; 178:730.
  • 5
    Holguin NM, Parker CJ. Membrane inhibitor of reactive lysis. Curr Top Microbiol Immunol 1992; 178:6185.
  • 6
    Lublin DM. Glycosyl-phosphatidylinositol anchoring of membrane proteins. Curr Top Microbiol Immunol 1992; 178:141162.
  • 7
    Rosse WF. The glycolipid anchor of membrane surface proteins. Semin Hematol 1992; 30:219231.
  • 8
    Rosse NW, Ware RE. The molecular basis of paroxysmal nocturnal hemoglobinuria. Blood 1995; 86:32773286.
  • 9
    Takeda J, Miyata T, Kawagoe K, et al. Deficiency of the GPI anchor caused by a somatic mutation of the PIG-A gene in paroxysmal nocturnal hemoglobinuria. Cell 1993; 73:703711.
  • 10
    Miyata T, Junji T, Iida Y, et al. The cloning of PIG-A, a component in the early step of GPI-anchor biosynthesis. Science 1993; 59:13181320.
  • 11
    Dacie JV, Lewis SM. Paroxysmal nocturnal hemoglobinuria; clinical manifestations, haematology and nature of disease. Ser Haematol 1972; 3:323.
  • 12
    Socie' G, Mary JY, de Gramont A, et al. Long term follow-up and prognostic factors in paroxysmal nocturnal hemoglobinuria (PNH): a study on 203 cases by the French Society of Hematology. Blood 1995; 86(suppl 1):130a.
  • 13
    Yamashina M, Ueda D, Kinoshita T, et al. Inherited complete deficiency of 20 kilodalton homologous restriction factor (CD59) as a cause of paroxysmal nocturnal hemoglobinuria. N Engl J Med 1990; 323:11851189.
  • 14
    Rosse WF. Variations in the red cell in paroxysmal nocturnal hemoglobinuria. Br J Haematol 1973; 24:327.
  • 15
    Rosse WF, Dacie JV. Immune lysis of normal human and paroxysmal nocturnal hemoglobinuria (PNH) red blood cells. I. The sensitivity of PNH red cells to lysis by complement and specific antibody. J Clin Invest 1966; 45:736.
  • 16
    Rosse WF. The complement sensitivity of PNH cells. Ser Haematol 1972; 3:101114.
  • 17
    Rosse WF. Paroxysmal nocturnal hemoglobinuria-present status and future prospects. West J Med 1980; 132:219228.
  • 18
    Holguin NM, Wilcox LA, Bemshaw NJ, et al. Relationship between the membrane inhibitor of reactive lysis and the erythrocyte phenotypes of paroxysmal nocturnal hemoglobinuria. J Clin Invest 1989; 84:13871394.
  • 19
    Rosse WF, Hall S, Campbell M, et al. The erythrocytes of paroxysmal nocturnal hemoglobinuria of intermediate sensitivity to complement lysis. Br J Haematol 1991; 79:99107.
  • 20
    Endo M, Ware RE, Vreeke TM, et al. Molecular basis of the heterogeneity of expression of glycosyl phosphatidylinositol anchored proteins in paroxysmal nocturnal hemoglobinuria. Blood 1996 (in press).
  • 21
    Iida Y, Takeda J, Miyata T, et al. Characterization of genomic PIG-A gene: a gene for GPI-anchor biosynthesis and paroxysmal nocturnal hemoglobinuria. Blood 1994; 83:31263131.
  • 22
    Bessler M, Hillmen P, Longo L, et al. Genomic organization of the X-linked gene (PIG-A) that is mutated in paroxysmal nocturnal haemoglobinuria and of a related autosomal pseudogene mapped to 12q21. Hum Mol Genet 1994; 3:751757.
  • 23
    Kawagoe K, Takeda J, Endo Y, et al. Molecular cloning of murine Pig-A, a gene for GPI-anchor biosynthesis, and demonstration of interspecies conservation of its structure, function, and genetic locus. Genomics 1994; 23:566574.
  • 24
    Bessler M, Mason P, Hillmen P, et al. Somatic mutations and cellular selection in paroxysmal nocturnal hemoglobinuria. Lancet 1994; 343:951953.
  • 25
    Yamada N, Miyata T, Maeda K, et al. Somatic mutations of PIG-A gene found in Japanese patients with paroxysmal nocturnal hemoglobinuria. Blood 1995; 85:885892.
  • 26
    Nafa K, Mason PJ, Hillmen P, et al. Mutations in PIG-A gene causing paroxysmal nocturnal hemoglobinuria are mainly of the frame shift type. Blood 1995; 86:46504655.
  • 27
    Ostendorf T, Nischan C, Schubert J, et al. Heterogeneous PIG-A mutations in different cell lineages in paroxysmal nocturnal hemoglobinuria. Blood 1995; 85:16401646.
  • 28
    Nafa K, Bessler M, Mason PJ, et al. The complex nature of somatic mutations in paroxysmal nocturnal hemoglobinuria. Blood 1995; 86(suppl 1):245a.
  • 29
    Oni SB, Osunkoya BO, Luzzatto L. Paroxysmal nocturnal hemoglobinuria: evidence of monoclonal origin of abnormal red cells. Blood 1970; 36:145152.
  • 30
    Josten KM, Tooze JA, Borthwick-Clarke C, et al. Acquired aplastic anemia and paroxysmal nocturnal hemoglobinuria: studies on clonality. Blood 1991; 78:31623167.
  • 31
    Ohashi H, Hotta T, Ichikawa A, et al. Peripheral blood cells are predominantly chimeric of affected and normal cells in patients with paroxysmal nocturnal hemoglobinuria: simultaneous investigation of clonality and expression of glycosylphosphatidylinositol-anchored proteins. Blood 1994; 83:853859.
  • 32
    Endo M, Beatty PG, Vreeke TM, et al. Syngeneic bone marrow transplantation without conditioning in a patient with paroxysmal nocturnal hemoglobinuria: in vivo evidence that the abnormal stem cells have a survival advantage. Blood 1996 (in press).
  • 33
    Ware RE, Rosse WF, Hall SE. Immunophenotypic analysis of reticulocytes in paroxysmal nocturnal hemoglobinuria. Blood 1995; 86:15861589.
  • 34
    Young NS. The problem of clonality in aplastic anemia: Dr. Dameshek's riddle, revisited. Blood 1992; 79:13851392.
  • 35
    Endo M, Ware RE, Vreeke TM, et al. Identification and characterization of an inherited mutation of PIG-A in a patient with paroxysmal nocturnal haemoglobinuria. Br J Haematol 1996 (in press).
  • 36
    Nagarajan S, Brodsky RA, Young NS, et al. Genetic defects underlying paroxysmal nocturnal hemoglobinuria that arises out of aplastic anemia. Blood 1995; 86:46564661.
  • 37
    Holliday R, Grigg GW. DNA methylation and mutation. Mutat Res 1993; 285:6167.
  • 38
    Hillmen P, Lewis SM, Bessler M, et al. Natural history of paroxysmal nocturnal hemoglobinuria. N Engl J Med 1995; 333:12531258.
  • 39
    de Planque MM, Bacigalupo A, Würsch A, et al. Long-term follow-up of severe aplastic anemia patients treated with antithymocyte globulin. Br J Haematol 1989; 73:121.
  • 40
    Schrezenmeier H, Marin P, Raghavachar A, et al. Relapse of aplastic anaemia after immunosuppressive treatment: a report from the European Bone Marrow Transplantation Group SAA Working Party. Br J Haematol 1993; 85:371377.
  • 41
    Tichelli A, Gratwohl A, Würsch A, et al. Late complications in severe aplastic anemia. Br J Haematol 1988; 69:413418.
  • 42
    Doney K, Pepe M, Storb R, et al. Immunosuppressive therapy of aplastic anemia: results of a prospective, randomized trial of antithymocyte globulin (ATG), methylprednisolone, and oxymetholone to ATG, very high-dose methylprednisolone, and oxymetholone. Blood 1992; 79:25662571.
  • 43
    Schubert J, Vogt HG, Zielinska-Skowronek M, et al. Development of glycosylphosphatidylinositol-anchoring defect characteristic for paroxysmal nocturnal hemoglobinuria in patients with aplastic anemia. Blood 1994; 83:23232328.
  • 44
    Griscelli-Bennaceur A, Gluckman D, Scrobohaci ML, et al. Aplastic anemia and paroxysmal nocturnal hemoglobinuria: search for a pathogenetic link. Blood 1995; 85:13541363.
  • 45
    Nissen C, Filipowicz A, Dalle Carbonare V, et al. Subclinical paroxysmal nocturnal hemoglobinuria (PNH) in aplastic anemia (AA): incidence and predictive factors for progression to clinical disease. Blood 1995; 86(suppl 1):130a.
  • 46
    Young NS, Barrett AJ. The treatment of severe acquired aplastic anemia. Blood 1995; 85:33673377.
  • 47
    Tichelli A, Gratwohl A, Nissen C, et al. Morphology in patients with severe aplastic anemia treated with antilymphocyte globulin. Blood 1992; 80:337345.
  • 48
    Maciejewski JP, Sloand E, Sato T, et al. Impaired hematopoiesis in PNH/aplasia is not associated with a selective proliferative defect in the GPI-protein deficient clone. Blood 1995; 86(suppl 1):589a.
  • 49
    Rotoli B, Luzzatto L. Paroxysmal nocturnal hemoglobinuria. Semin Hematol 1989; 26:201207.
  • 50
    Ware RE, Howard TA, Kamitani T, et al. Chromosomal assignment of genes involved in glycosyl-phosphatidylinositol anchor biosynthesis: implications for the pathogenesis of paroxysmal nocturnal hemoglobinuria. Blood 1994; 83:37533757.
  • 51
    Hertenstein B, Wagner B, Bunjes D, et al. Emergence of CD52, phosphatidylinositolglycan-anchor-deficient T lymphocytes after in vivo application of Campath-lH for refractory B-cell non-Hodgkin lymphoma. Blood 1995; 86:14871492.
  • 52
    Schrezenmeier H, Merk B, Ilijewa S, et al. Pig-A gene mutations are a necessary but not sufficient prerequisite for the occurrence of glycosylphosphatidylinositol (GPI)-anchor deficient blood cells. Blood 1995; 86(suppl 1):589a.
  • 53
    Robinson PJ. Phosphatidylinositol membrane anchors and T cell activation. Immunol Today 1991; 12:3541.
  • 54
    Stefanova I, Horejsi V, Ansotegui IJ, et al. GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science 1991; 254:10161019.
  • 55
    Graham DL, Gastineau DA. Paroxysmal nocturnal hemoglobinuria as a marker for clonal myelopathy. Am J Med 1992; 93:671674.
  • 56
    Crosby WH. Paroxysmal nocturnal hemoglobinuria: a classic description by Paul Strübing in 1882, and a bibliography of the disease. Blood 1951; 6:270284.