International Journal of

HEMATOLOGY

Clinical Manifestations of Paroxysmal Nocturnal Hemoglobinuria:Present State and Future Problems

Wendell F. Rosse,a lunichi Nishimurab

aDivisions of Hematology and bMedical Oncology-Bone Marrow Transplantation,Department of Medicine, Duke University Medical Cente1;

Durham, North Carolina, USA

Received October 11,2002; accepted October 29, 2002

AbstractThe clinical pathology of paroxysmal nocturnal hemoglobinuria (PNH) involves 3 complications: hemolytic anemia, throm-

bosis, and hematopoietic deficiency. The first 2 are clearly the result of the cellular defect in PNH, the lack of proteins anchoredto the membrane by the glycosylphosphatidylinositol anchor. The hemolytic anemia results in syndromes primarily related tothe fact that the hemolysis is extracellular. Thrombosis is most significant in veins within the abdomen, although a number ofother thrombotic syndromes have been described. The hematopoietic deficiency may be the same as that in aplastic anemia, aclosely related disorder, and may not be due to the primary biochemical defect. The relationship to aplastic anemia suggests anomenclature that emphasizes the predominant clinical manifestations in a patient. This relationship does not explain casesthat appear to be related t9 myelodysplastic syndromes or the transition of some cases of PNH to leukemia. Treatment, exceptfor bone marrow transplantation, remains n9Jlcurative and in need of improvement. lnt J Bernato/. 2003;77:113-120.@2003 The Japanese Society of Hematology

Key words: Paroxysmal nocturnal hemoglobinuria; Hemolysis; Thrombosis; Aplastic anemia; Myelodysplasia

1. Introduction hematopoiesis. To define the future solutions to the appro-priate treatment Qf the disease, we must consider each sepa-rately and all together.

2. Hemolytic Anemia

The cause of the hemolysis in PNH is well documented.Because of the defect in the construction of the glyco-sylphosphatidylinositol (GPI) anchor [2],2 proteins impor-tant in the defense of the red cell against the hemolyticaction of complement are missing-CD55 (decay accelerat-ing factor [DAF]) [3] and CD59 (protectin, membraneinhibitor of reactive lysis, etc) [4]. The result is that the redcells (as well as the other blood cells) are more vulnerable tothe hemolytic action of complement than are normal cells [5].

Of these 2 proteins, CD59 appears to be the moreimportant. Results of in vitro tests have shown that inhibi-tion of this molecule on normal erythrocytes by antibodyresults in a susceptibility to complement of those cellsnearly equal to the susceptibility of the most abnormalPNH cells (those completely lacking the GPI-anchoredproteins), whereas similar inhibition of CD55 results in amodest increase in the susceptibility of normal red cells tothe action of complement. The congenital absence of

Paroxysmal nocturnal hemoglobinuria (PNH) is anuncommon disorder that was first described unequivocallyin 1866 as present in a patient with bouts of dark urine,thought to be due to "hemitin," that were worse in theearly morning [1]. Since that time, the clinical and labora-tory manifestations have been recorded in numerousworks that now give a more complete account of the disor-der. In parallel, considerable basic scientific endeavorshave yielded information that explains many of these man-ifestations. Nevertheless, a great deal more needs to beknown about the clinical and basic manifestations of thisdisorder so that therapy, which is at present inadequate,can be improved.

The major clinical manifestations of paroxysmal noctur-nal hemoglobinuria can be classified into 3 categories: (1)hemolytic anemia, (2) venous thrombosis, and (3) deficient

Correspondence and reprint requests: Wendell F. Rosse, MD,4605 Timberly Dr, Durham NC, 27707; fax: 1-919-489-0275 (e-mail:

rosseOO1@mc.duke.edu).

113

CD55-the so-called Inab phenotype-results in little ifany hemolysis in vitro or in the patient [6], whereas thecongenital absence of CD59 results in a striking syndromeof hemolysis similar to that of PNH [7]. Thus if progress isto be made in the prevention of hemolysis, it should con-centrate on inhibiting that part of the complementsequence converging on the formation of the membraneattack complex or on functionally replacing CD59.

Hemolysis becomes most evident when complement isactivated. The nocturnal pattern of hemolysis from which thedisease acquires its name may be due to the nocturnalabsorption of lipopolysaccharide (LPS), a byproduct of thebacterial cell wall, from the gut. LPS markedly activates thecomplement system, possibly through the mannan-bindingprotein pathway [8]. LPS is normally bound by monocytesthrough a GPI-linked protein, CD14 [9], which is missing inPNH [10]. In many patients with PNH, this hemolysis is aminor inconvenience, whereas in others it is a major sourceof the loss of hemoglobin and other complications attendantupon it. Certainly the latter group would benefit from betterknowledge of LPS metabolism and perhaps better ways ofcontrolling its activation of complement.

Complement also is activated by concurrent inflammationor the immune reaction to infections. Most viral disorderswill result in a burst of hemolysis, which often becomes daylong rather than nocturnal. The most serious hemolysisresults from gastrointestinal inflammation, usually viral gas-troenteritis. This hemolysis may be due to increased absorp-tion of LPS and the effect of the immune reaction against thevirus. It is hemolysis in this setting that is most likely to resultin acute renal damage [11].

Immune reactions that ordinarily entail little hemolysismay cause serious hemolysis in PNH. For example, mostpatients with infectious mononucleosis from Epstein-Barrvirus infection generate a low level of cold agglutinin anti-bodies of anti-i specificity. Except in rare cases in which thetiter is very high, these antibodies are not hemolytic to nor-mal cells. InPNH, however, because of the characteristic sen-sitivity to complement, the presence of the antibodies mayresult in severe hemolysis (personal experience).

Immunizations and vaccinations are designed to causelimited immunological reactions, but these reactions mayactivate complement. Ordinarily, this activation is notenough to cause hemolysis of the patient's red cells; how-ever, in patients with PNH, because of the sensitivity of thecells to complement, rapid hemolysis may occur in somecases. Great care should be taken in giving repeated doses ofpolysaccharide-containing vaccine against Streptococcuspneumoniae, because this treatment has been known toresult in a severe hemolytic reaction (personal experience).

Administration of a transfusion to a PNH patient maycause a burst of hemolysis of the patient's red cells. This reac-tion often is alarming because the hemoglobinuria thatresults is taken as a sign of an incompatible transfusion reac-tion with hemolysis of the transfused cells. The incidence andcause of the reaction are not certain, but the preponderanceof evidence suggests that the reaction is due to activation ofcomplement by immune reactions involving leukocytes orplasma proteins [12]. The problem usually can be avoidedwith thorough washing of red cells before transfusion, but

this step probably is not necessary unless the patient has ahistory of such reactions [13].

The most direct and obvious result of intravascularhemolysis as seen in PNH is hemoglobinemia. The plasmahemoglobin level is nearly always elevated, and plasma hap-toglobin (its binding protein) level is diminished,even in aquiescent clinical state. Thus the binding capacity of hapto-globin is readily exceeded, and hemoglobin a~ dimers circu-late in the plasma and are filtered by the glomeruli of the kid-ney. The dimers are resorbed in the proximal tubules, wherethey are degraded, and the iron is stored in ferritin in theepithelium of the proximal tubule. This iron is detected asurinary hemosiderin in almost all patients with PNH almostall the time, as observed by Marchiafava and Micheli [14,15].

Considerable hemolysis may occur without evidence ofhemoglobinuria. As greater hemolysis occurs, the level of a~globin dimers in the glomerular filtrate will rise to exceed theresorption capacity of the proximal tubules. The result isexcretion of the dimers in the urine as hemoglobinuria. If thelevel of globin dimers becomes so great that the resorptioncapacity of the proximal tubule for other molecules normallyresorbed by this epithelium, such as glucose and small pep-tides, is impaired, the other small molecules may appear inthe urine in a form of temporary Fanconi syndrome of prox-imal tubular dysfunction.

In rare instances, the concentration of hemoglobin in thetubular filtrate becomes sufficiently high to impair renalfunction, and acute renal failure results [11,16,17]. This situa-tion usually is seen with gastrointestinal illnesses and is com-plicated by the inability of the patient to take in water.Although there usually is good recovery if the condition istreated with hydration and careful monitoring of blood pres-sure, there often is residual renal impairment.

Chronic damage to the kidneys takes 2 forms: proximalrenal tubular acidosis and chronic renal failure. The former israrely recognized (acidosis with low serum bicarbonate level,low potassium level, increased calciuria, etc). The latter usu-ally is slowly progressive and may result in death. To date, noeffective treatment has been found for progressive renal fail-ure, the cause of death in approximately 8% of patients.

Hemoglobin binds nitric oxide (NO) with considerableavidity. However, when it is confined to the red cells, hemo-globin is too distant from the site of NO formation to affectorgan function. When it is free in the plasma, on the otherhand, hemoglobin is able to diffuse into tissues and sop upNO, causing a local deficiency manifest in contraction ofsmooth muscle. This effect is most evident in the esophagus;patients with hemoglobinemia frequently have substernal"tightness," especially early in the morning. Manometricstudies have shown that the waves of muscular contractionare normal in initiation and propagation but are severaltimes as strong as normal (personal experience). With theclearing of the hemoglobinuria, the spasms subside. Manypatients gain benefit from sublingual or transdermal nitro-glycerine during periods of heavy hemoglobinuria.

Removal of NO has been implicated in a frequent com-plication among men with PNH-penile erectile dysfunctionduring periods of hemoglobinuria. This symptom may berelieved with relatively large doses of sildenafil (Viagra),although this treatment is not always successful. More infor-

Clinical Manifestations of PNH 115

mation about the role of hemoglobinemia and the metabo.lism of NO clearly is needed.

3.

Thrombosis

An excessive incidence of venous thrombosis has beennoted almost since the first clinical descriptions of PNH andhas been recognized both in Europe and in the United Statesas a leading cause of death since at least 1956 [18]. Results ofrecent epidemiological studies in Europe and in the UnitedStates suggest that the incidence is approximately 40% of allpatients [19,20] (J.N. et aI, unpublished data). It is curious

.that the incidence among East Asian patients (Chinese, Thai,and Japanese) and among Mexicans is considerably less thanthat in other groups. In a recent review, the incidence wasfound to be approximately 5% in a large Japanese popula-tion of PNH patients (J.N. et al, unpublished data).

Venous thrombosis can occur in almost any venous site.Hepatic veins and veins of the portal system are particularlyaffected [21,22]. Hepatic venous thrombosis can occur assudden obliteration of major hepatic veins (the classic Budd-Chiari syndrome), often in the setting of severe hemolysis;this form is easily demonstrated with radiographic and sono-graphic techniques. Clinical examination shows the liver islarge and painful to percussion. Ascites may accumulateacutely, and the patient may become jaundiced. Levels of theserum enzymes indicative of liver damage are elevated; caremust be used to distinguish the elevations of lactic acid dehy-drogenase and aspartate aminotransferase due to hemolysis(the level of alanine aminotransferase is not increased byhemolysis).

In other patients, hepatic venous thrombosis may gradu-ally involve the small radicals of the hepatic vein, a situationmuch more difficult to demonstrate with radiographic orother techniques [22]. Clinically this disorder may manifestas pain in the right upper quadrant, enlargement of the liver,and the gradual onset of ascites. Signs of portal hypertensionmay be seen. "Liver enzyme" levels, notably serum alkalinephosphatase, usually are elevated.

Hepatic venous thrombosis may be accompanied by infe-rior vena caval thrombosis or renal vein thrombosis, whichleads to lower body anasarca in the first case and renal dys-function (proteinuria, renal failure) in the second. In bothcases, the thrombus usually is easily visualized by radio-graphic or sonographic means.

Other veins of the abdomen may be affected with throm-bosis, including splanchnic veins [17]; the result is a syndromeof recurrent abdominal pain and sometimes bowel necrosis[23]. Involvement of the portal vein may result in a syndromeof portal hypertension (ascites, esophageal varices, caputmedusae, etc) [24]. Thrombosis of the splenic vein can causemassive splenic enlargement and even rupture [25].

Cerebral venous thrombosis is common among the sitesin which thrombosis occurs [26,27]. The sagittal sinus is prob-ably the most frequently involved. Thrombosis in this areamay present as a history of severe headaches and evidence ofincreased intracranial pressure [28]. Thrombosis of the veinscovering the cerebrum may cause neurological syndromesconsonant with the site affected (eg, hemiparesis if over themotor strip, Gerstmann syndrome if over the dominant pari-

116

is clearly usef\ll in acute thrombosis. Low-molecular-weightheparin (LMWH) is simpler to use than and probablyequally as effective as conventional heparin; however, suffi-ci~nt experience has not been gained with the use ofLMWH to lead to understanding of the advantages and lim-its of this agent.

It is important to instit\lte thrombolytic therapy whenacute thrombus formation is going on or is very recent. Inacute Budd-Chiari syndrome, the use of tissue plasminogenactivator has dramatically reduced the size of the liver andallowed the patient to survive without further episodes orprogression of thrombosis (personal experience). The use ofthrombolytic agents in chronic thrombosis has been sug-gested but not established [42].

The following suggestions have been made ilcbout thetreatment of thrombosis in PNH after a metanalysis of theproblem [31]. It should be noted that these suggestions havenot been tested prospectively,

1. Long-term anticoagulation with warfarin derivativesshould be considered for patients of Europ~an descent whoare not thrombocytopenic; alternatively, daily aspirin therapymay be considered.

2. Prophyla{{:is with heparin or LMWH should be used inthe perioperative period, during prolonged immobilization,or during the prolonged use of an intravenous catheter. Pro-phylaxis with heparin (7,500-10,000 units twice a day) orLMWH (75-100 anti-Xa units/kg) should be started in thefirst trimester of pregnancy and m~intained until 4 to 6 weeks

postpartum.3. Aggressive anticoagulation with heparin or LMWH

should be pursued during any acute thrombotic episode. Thistherapy probably should be maintained for a long termbecause recurrence is likely.

4. If the platelet count is less than 10 X 109/L, anticoagu-lation is contraindicated. When the platelet count is greaterthan 50 X 109/L, th~ usual anticoagulation may be used.Between those platelet counts, care must be taken, andplatelet transfusions may be needed.

It is clear from this discussion that the phenomenon ofexcessive thrombosis and its treatment are insufficientlyunderstood. This is an area in which gr~at aid could begiven to patients with PNH, particularly those whose ances-try is European, by improvements in u:nderstanding andtreatment.

4.

Hematopoietic Deficiency

The relationship between bone marrow hypoplasia andPNH has been documented for some time but is still poorly

understood. Lewis and Dacie pointed out 40 years ago thatmany patients with PNH had a history of aplastic anemia andthat hematopoiesis appeared to be diminished to a greater orlesser extent in all patients with the disease [43,44]. This asso-ciation became evident when it was realized that the survivalof platelets and granulocytes in PNH was normal [45,46]; thattwo thirds of patients had thrombocytopenia or granulocy-topenia or both [47] indicated there is commonly underpro-duction of these cells in these patients. As many as 50% ofpatients with aplastic anemia have a readily detectable pop-ulation of GPI- (PNH-like) hematopoietic cells, particularly

during and after recovery in response to antithymocyte glob-ulin [48-50]. Finally, many patients with PNH develop aplas-tic anemia as the final stage of the disease [19,20] (J.N. et aI,unpublished data).

When marrow culture methods became available, it wasshown that the cells of pati~nts with PNH did not grow wellin vitro [51]. This finding was true of both t!1e normal GPI+[52,53] and the abnormal GPI- precursors. This lack ofgrowth was shown not to be a fault of the stroma needed forlong-term culture because the PNH cells grew poorly onnormal or PNH stroma, and normal cells grew normally oneither type of stroma. In these respects, thf' growth of mar-row resembled that of the marrow in aplastic anemia [53].

These findings led to the current dominant hypothesis(the Young-Luzzatto hypothesis) for the explan~tion of thepathophysiology of PNH [54,55]. This hypothesis supposesthat GPI- hematopoietic stem cells exist in very small num-bers in the bone marrow of many healthy persons. Thissupposition has been indirectly confirmed by severallabora-tories [56]; approximately 1 in 106 granulocytes or lympho-cytes is found to lack the GPI anchor in many if not mosthealthy donors. These defective cells are thought have agrowth disadvantage in the pormal marrow environmentand therefore rem~in in very small numbers. The fundamen-tal and difficult challenge is to understand how the defectivecells come to be expressed in large enough numbers toresult in clinical PNH.

The hypothesis suggests that when the marrow is affectedby aplastic anemia (thought to be an autoimmune reactionagainst marrow pr~cursors) [57], these defective precursorsare thought to be less suppressed than the normal GPI+ pre-cursors and thus become, by Darwinian principles, the domi-nant source of hematopoiesis. The findings show that growthc!1aracteristics of marrow from patients with PNH are verysimilar to those of marrow from patients with aplastic ane-mia. There is marked diminution of colony formation in bothcases, and long-term marrow culture is difficult. It is espe-cially noteworthy that the GPI+ precursors in PNH marrowgrow no better than the GPI- precursors, a finding consistentwith the action of a suppressive element. However, the causeof selection and expansion of the PNH clone(s) is at the pres-ent time not at all clear.

An alternative hypothesis is that the dominance of theGPI- clone in PNH arises because of a decrease in apoptosisof the nuclear precursors in PNH marrow [58]. The specificityof this finding has been doubted, because the same phenom-enon is seen in other disorders of the bone marrow, includingaplastic anemia and myelodysplastic syndromes [59], and isnot limited to the GPI- cells [60]. More understanding of theinitiation and control of apoptosis is needed.

The GPI- precursors do not appear to have any prolifera-tive advantage in vivo. In experiments in which the PIG-Agene was knocked out, fetuses completely lacking the GPIanchor did not survive to birth. The chimeric animals with asmall proportion of GPI- cells did survive, but the proportionof those cells did not increase with time [61]. These facts sug-gest that the Young-Luzzatto hypothesis is, in large measure,correct.

PNH is related to another class of bone marrow disor-ders, the myelodysplastic disorders. The association with

Clinical Manifestations of PNH 117

myelofibrosis was made many years ago [62,63]. The occur-rence of myelodysplasia in the marrow of some patientswith PNH was noted at about the same time [64]. The asso-ciation with other myelodysplastic disorders has been madesince that time [65,66]. Cells lacking GPI-linked proteinshave been found in up to 20% of patients with myelodys-plastic disorders [67]. These facts are difficult to reconcilewith the Young-Luzzatto hypothesis on the basis of currentknowledge.

A rare complication of PNH is its evolution into acuteleukemia, first noted in 1969 [68-70]. Since that time, manymore cases have been reported [71]. With few exceptions, the.leukemia is acute nonlymphocytic in type. It is heralded bythe disappearance of the abnormal PNH red cells and oftenby a period of dyshematopoiesis lasting several months. Thisform of leukemia has occurred in patients who presentedwith the hemolytic form of PNH as well as in those who pre-sented with aplastic anemia [72,73] or myelodysplastic syn-drome. The leukemic cells are invariably lacking GPI-linkedproteins, a finding that indicates the origin of the cells in theabnormal clone. The only exceptions are cells seen inpatients with myelodysplasia and PNH, in whom theleukemic cells may have GPI-linked proteins. This findingsuggests that the cells arose from the myelodysplastic cloneof cells [56,74,75]. It is not at all clear how the occurrence ofleukemia in the abnormal clone fits with the Young-Luzzattohypothesis. This association is an area of fruitful and inter-esting research.

presence of PNH cells (>5% of the granulocytes) might leadto some clinical symptoms.

PNH-aplastic anemia (PNH-AA): The predominant clin-ical syndrome is that of PNH (hemolysis, thrombosis) butwith significant evidence of bone marrow hypoplasia, includ-ing granulocytopenia and/or thrombocytopenia.

PNH (classic PNH): The clinical syndrome is that bf PNHwithout clinical evidence of bone marrow hypoplasia.

The presence of myelodysplastic hematopoiesis could alsobe indicated, either as MDS-PNH when PNH cells are pres-ent in the patient with predominantly myelodysplastichematopoiesis or as PNH-MDS when the clinical syndromeis predominantly due to the abnormal cells of PNH but ele-ments of MDS are significantly present.

At the risk of making more complex the nomenclature ofthese diseases, this suggested classification would focus ther-apy on the predominant abnormality and would be helpful inunderstanding the natural history of PNH.

6. Treatment of PNH

5. Classification and Nomenclatnre

The treatment of PNH is aimed either directly at theabnormalities of the cells or of hematopoiesis or at theeffects of the defect. In any case, treatment may be of ben-efit but is curative in only 1 case-bone marrow transplan-tation.

In the general care of patients with PNH or PNH-AA,iron supplementation usually is recommended because alarge amount of iron is lost either as hemoglobin or as hemo-siderin [77]. This supplementation may be accompanied by aburst of hemoglobinuria as first described by Strubing [78];this effect was found to be due to the emergence of a cohortof defective red cells in the therapeutic response [79]. Sup-plementation with folic acid also is usually prescribed,although deficiency of this vitamin in this disease has notbeen reported. It is possible that the folic acid of red cells isreused when the red cells are lysed in the circulation.

The role of adrenocorticosteroids in PNH and PNH-AAis controversial. Reports of the utility of this form of treat-ment have circulated for many years; however, the dosesrequired are high and cannot be continued on a daily basis.Approximately 60% of patients derive some benefit as meas-ured by an increase in hemoglobin when steroids (0.3-0.5 mglkg) are administered on an alternate-day basis. With this reg-imen most patients have few side effects, and none of theseside effects is of a serious nature (personal experience).

Androgenic hormones have been advocated, presumably,as improving hematopoiesis [80]. This mode of treatment isnot without androgenizing side effects and, in some cases,liver dysfunction. Because the hormones are useful to aminority of patients, a short (1-:?- month) trial often is used. Ifno improvement in hemoglobin level is seen, the drug shouldnot be used.

Treatments that improve hematopoiesis in aplastic ane-mia have been used in the care of patients with PNH, partic-ularly those with a major aplastic component (AA-PNH andPNH-AA). Antithymocyte or antilymphocyte globulinaffects remission in approximately 40% to 60% of patientswith aplastic anemia. Remission is complicated in up to 10%of patients by the appearance or increase in the number of

With the development of newer diagnostic tools, particu-larly the detection of cells lacking GPI-linked proteins withthe use of monoclonal antibodies and flow cytometry, andthe recognition that the disease has 2 components-abnor-mal cells and abnormal hematopoiesis, the diagnostic defini-tion of PNH has become somewhat muddled. Does thepatient with aplastic anemia with a very small population ofGPI- cells have PNH? Does the patient with hemolytic ane-mia along with granulocytopenia and thrombocytopeniahave only PNH? Where does myelodysplasia fit in? Perhapsa new system of nomenclature is in order.

Clinically, PNH has 2 sets of symptoms and signs-thosedue to the abnormality of the cells (hemolysis, thrombosis)and those due to the insufficiency of hematopoiesis(hypoplasia or aplasia of the marrow, granulocytopenia,thrombocytopenia). Many patients have a predominance ofone or the other set of symptoms, and it is useful when treat-ing the patients to emphasize the predominant abnormality.For these reasons, we propose the following nomenclature.

Aplastic anemia with detectable PNH cells [AA-(PNH)]:The clinical syndrome is caused by bone marrow failure withthe detection of fewer than 5% PNH granulocytes in theperipheral blood with standard monoclonal antibodies andflow cytometry (note: detection of the abnormal cells in thegranulocytes is more reliable than detection in red cells orplatelets [76]). This classification would flag patients whomight develop overt PNH symptoms in the future.

Aplastic anemia-PNH (AA-PNH): The predominantclinical syndrome is that of bone marrow failure, but the

118 Rosse and ~ishimura / /nternationill Journal of Hematol"gy 77 (2003) /13-120

GPI- cells [81]. This immunotherapy improves hematopoiesisin approximately the same proportions of patients with AA-PNH and PNH-AA (personal experience); however, the pro-portion as well as the number of PNH erythrocytes oftenincreases, and hemolysis increases. Relapse can be treatedwith repeated administration of the drug, but great care mustbe taken to administer massive doses of prednisone (500 mg!d) or its equivalent to prevent anaphylaxis and massivehemolysis.

Cyclosporine also is immunosuppressive and has showngreat utility in the treatment of aplastic anemia [82]. Use ofcyclosporine in PNH has not been documented, but cyclo-sporine presumably would be useful in treating the hemato-poietic deficiency in the disease.

Immunosuppression and subsequent remission of aplasticanemia have been achieved with very high doses ofcyclophosphamide [83]. This treatment has been used in thecare of patients with AA-PNH and PNH-AA with improve-ment of hematopoiesis but persistence of the abnormal cloneof PNH cells.

Allogeneic and syngeneic bone marrow transplantationhas been used successfully in the treatment of PNH whenserious prognostic signs are present (onset of aplasia andthrombosis in particular) [84-87]. The complications encoun-tered appear to be those encountered in bone marrow trans-plantation for other diseases-nonengraftment, graft-versus-host disease, infection, and so on. In almost all cases, theabnormal clone was eliminated. Survival has increased astreatment of the complications has improved, and this formof therapy is becoming more useful to patients with PNH. Todate, the incidence of complications in transplantation of themarrow of HLA-identical but unrelated donors is several-fold higher than the incidence in transplantation fromrelated donors, and this therapy generally is reserved for des-perate situations. The use of nonablative regimens, which aremore easily tolerated, is under preliminary investigation [88].

Because the abnormal gene is known and can be cor-rected, the role of gene therapy should be considered. In genetherapy, the normalized gene is inserted into hematopoieticprecursors in the hope that normal cells or cells containing aspecific gene will proliferate. In PNH, many patients retain asignificant population of normal hematopoietic precursors sothat the generation of normal cells is not a problem for them.Techniques are now available for transferring the PIG-Agene relatively efficiently into hematopoietic precursors [89]with expression of the GPI-linked proteins. Thus in patientswho have essentially replaced the normal hematopoietic stemcells with the defective ones, a clone of PIG-A-replete cellscan be generated. It is not clear, however, that this gene trans-fer is all that need be done. If, as is supposed in the Young-Luzzatto hypothesis, a secondary insult to the marrow hasoccurred to suppress normal hematopoiesis, then the alteredcells would presumably be under the same repression. Thusrelief of the suppression of the normal marrow elements willbe necessary before gene therapy can be successful.

References

1. Gull WW. A case of intermittent haematinuria, with remarks. GuysHosp Rep. 1866;12:381-392.

2. Takeda J, Miyata T, Kawagoe K, et al. Deficiency of the GPI anchorcaused by a somatic mutation of the PIG-A gene in paroxysmalnocturnal hemoglobinuria. Cell. 1993;73:703-711.

3. Nicholson-Weller A, March JP, Rosenfeld SI, Austen KF. Affectederythrocytes of patients with paroxysmal nocturnal hemoglobin-uria are deficient in the complement regulatory protein, decayaccelerating factor. Proc Natl Acad Sci USA. 1983;80:5430.

4. Holguin MH, Wilcox LA, Bernshaw NJ, Rosse WF, Parker 0.Relationship between the membrane inhibitor of reactive lysis andthe erythrocyte phenotypes of paroxysmal nocturnal hemoglobin-uria. J Clin Invest. 1989;84:1387-1394.

5. Rosse WF, Dacie JV. Immune lysis of normal human and paroxys-mal nocturnal hemoglobinuria red blood cells, I: the sensitivity ofPNH red cells to lysis by complement and specific antibody. J ClinInvest. 1966;45:736-748.

6. Telen MJ, Green AM. The Inab phenotype: characterization of themembrane protein and complement regulatory defect. Blood.1989;74:437-441.

7. Yamashina M, Veda E, Kinoshita T, et al. Inherited complete defi-ciency of 20-kilodalton homologous restriction factor (CD59) as acause of paroxysmal nocturnal hemoglobinuria. N Engl J Med.1990;323:1184-1189.

8. Ohta M, Okada M, Yamashina I, Kawasaki T. The mechanism ofcarbohydrate-mediated complement activation by the serum man-nan-binding protein. J Bioi Chern. 1990;265:1980-1984.

9. Couturier C, Haeffner Cavaillon N, Caroff M, Kazatchkine MD.Binding sites for endotoxins (lipopolysaccharides) on humanmonocytes. J Irnrnunol. 1991;147:1899-1904.

10. Simmons DL, Tan S, Tenen DG, Nicholson-Weller A, Seed B.Monocyte antigen CDl4 is a phospholipid anchored membraneprotein. Blood. 1989;73:284-289.

11. Jose MD, Lynn KB. Acute renal failure in a patient with paroxys-mal nocturnal hemoglobinuria. Clin Nephrol. 2002;56:172-174.

12. Zupanska B, Vhrynowska M, Konopka L. Transfusion-relatedacute lung injury due to granulocyte-agglutinating antibody in apatient with paroxysmal nocturnal hemoglobinuria. Transfusion.1999;39:944-977.

13. Brecher ME, Taswell HF. Paroxysmal nocturnal hemoglobinuriaand the transfusion of washed red cells: a myth revisited. Transfu-sion. 1989;29:681-685.

14. Marchiafava E. Anemia emolitica con emosiderinuria perpetua.Policlinico fMedJ. 1992;35:109.

15. Micheli F. Anemia splenomegalia emolyitica con emoglobinuria-emosiderinuria tipo Marchiafava. Haernatologica. 1931;12:101.

16. Mooraki A, Boroumand B, Mohammad Zadeh F, Ahmed SH, Bas-tani B. Acute reversible renal failure in a patient with paroxysmalnocturnal hemoglobinuria. Clin Nephrol. 1998;50:255-257.

17. Sechi LA, Marigliano A, Pala A, Tedde R. Acute renal failure inparoxysmal nocturnal haemoglobinuria with splanchnic venousthrombosis. Clin Lab Haernatol. 1989;11:273-275.

18. Crosby WHo Paroxysmal nocturnal hemoglobinuria: relation of theclinical manifestations to underlying pathogenic mechanisms.Blood. 1953;8:769-812.

19. Socie G, Mary J- Y, De Gramont A, et al. Paroxysmal nocturnalhaemoglobinuria: long term follow-up and prognostic factors.Lancet. 1996;348:573-577.

20. Hillmen P, Lewis SM, Bessler M, Luzzatto L, Dacie JV. Natural his-tory of paroxysmal nocturnal hemoglobinuria. N Engl J Med. 1995;333:1253-1258.

21. Hartmann RC, Luther AB, Jenkins DE Jr, Tenorio LE, Saba HI.Fulminant hepatic venous thrombosis (Budd-Chiari syndrome) inparoxysmal nocturnal hemoglobinuria: definition of a medicalemergency. Johns Hopkins Med 11980;146:247-254.

22. Peytremann R, Rhodes RS, Hartmann RC. Thrombosis in paroxys-mal nocturnal hemoglobinuria (PNH) with particular reference toprogressive, diffuse hepatic venous thrombosis. Ser Haernatol.1972;5:115-136.

23. Blum SF, Gardner FH. Intestinal infarction in paroxysmal noctur-nal hemoglobinuria. N Engl J Med. 1966;274:1137-1138.

119

24. Grossman JA, McDermott WV Jr. Paroxysmal nocturnal hemoglo-binuria associated with hepatic and portal venous thrombosis. AmJ Surg. 1974;127:733-736.

25. Zimmerman D, Bell WR. Venous thrombosis and splenic rupturein paroxysmal nocturnal hemoglobinuria. Am J Med. 1980;68:275-279.

26. Donhowe Sp, Lazaro RP. Dural sinus thrombosis in paroxysmalnocturnal hemoglobinuria. Clin Neurol Neurosurg. 1984;86:149-152.

27. Omardeen F, Wharfe G, St Orner L, Richards JS, Morgan OS. Cere-bral sinovenous thrombosis in a patient with paroxysmal nocturnalhaemoglobinuria. West Indian Med J 1992;41:31-33.

28. Hauser D, Barzilai N, Zalish M, Oliver M, Pollack A. Bilateralpapilledema with retinal hemorrhages in association with cerebralvenous sinus thrombosis and paroxysmal nocturnal hemoglobin-uria. Am J Ophthalmol. 1996;122:592-593.

29. Payne PR, Holt JM, Neame PB. Paroxysmal nocturnal haemo-globinuria parturition complicated by venous thrombosis.J Obstet Gynaecol Br Commonw. 1968;75:1066-1068.

30. Wozniak AJ, Kitchens CS. Prospective hemostatic studies in apatient having PNH pregnancy and cerebral venous thrombosis.Am J Obstet Gynecol. 1982;142:591-593.

31. Ray JG, Burows RF, Ginsberg JS, Burrows EA. Paroxysmal noc-turnal hemoglobinuria and the risk of venous thrombosis: reviewand recommendations for management of the pregnant and non-pregnant patient. Hemostasis. 2000;30:103-117.

32. Hartmann RC, Bruce ASC. Fetomaternal outcomes of pregnancyin paroxysmal nocturnal hemoglobinuria. In: Bern MM, ed. Hema-tologic Disorders in Maternal-Fetal Medicine. New York: Wiley-Liss; 1990:261-281.

33. Rietschel RL, Lewis CW, Simmons RA, Phyliky RL. Skin lesions inparoxysmal nocturnal hemoglobinuria. Arch Dermatol. 1978;134:560-563.

34. Draelos ZK, Hansen RC. Hemorrhagic bullae in an anemicwoman: Paroxysmal nocturnal hemoglobinuria (PNH). Arch Der-matol.1986;122:1326-1327,1329-1330.

35. Hansen NE, Killmann SA. Paroxysmal nocturnal haemoglobin-uria: a clinical study. Acta Med Scand. 1968;184:525-541.

36. Rosse WF. Paroxysmal nocturnal hemoglobinuria as a moleculardisease. Medicine. 1997;76:63-94.

37. Hamilton KK, Hattori R, Esmon cr, Sims PJ. Complement pro-teins C5b-9 induce vesiculation of the endothelial plasma mem-brane and expose catalytic surface for assembly of the prothrom-binase enzyme complex. J BioI Chem. 1990;265:3809-3814.

38. Wiedmer T, Hall SE, Ortel TL, et al. Complement-induced vesicu-lation and exposure of membrane prothrombinase sites in plateletsof paroxysmal nocturnal hemoglobinuria. Blood. 1993;82:1192-1196.

39. Hugel B, Socie G, Vu T, et al. Elevated levels of circulating proco-agulant microparticles in patients with paroxysmal nocturnalhemoglobinuria and aplastic anemia. Blood. 1999;93:3451-3564.

40. Gilbert GE, Sims PJ, WiedmerT, et al. Platelet-derived microparti-cles express high affinity receptors for factor VIII. J BioI Chem.1991;266:17261-17268.

41. Ploug M, Pilesner T, Ronne E, et al. The receptor for urokinase-type plasminogen activator is deficient on peripheral blood leuko-cytes in patients with paroxysmal nocturnal hemoglobinuria.Blood. 1992;79:1447-1455.

42. McMullin MF, Hillmen P, Jackson J, Ganly P, Luzzatto L. Tissueplasminogen activator for hepatic vein thrombosis in paroxysmalnocturnal hemoglobinuria. J Intern Med. 1994;235:85-89.

43. Dacie N, Lewis SM. Paroxysmal nocturnal haemoglobinuria: vari-ation in clinical severity and association with bone marrowhypoplasia. Br J Haematol. 1961;7:442-457.

44. Lewis SM, Dacie JV. The aplastic anaemia-paroxysmal nocturnalhaemoglobinuria syndrome. Br J Haematol. 1967;13:236-251.

45. Devine DV, Siegel RS, Rosse WF. Interactions of the platelets inparoxysmal nocturnal hemoglobinuria with complement: relation-

ship to defects in the regulation of complement and to platelet sur-vival in vivo. J Clin Invest. 1987;79:131-137.

46. Brubaker L, Essig U, Mengel CEo Neutrophil life span in paroxys-mal nocturnal hemoglobinuria. Blood. 1977;50:657-662.

47. Dacie JV, Lewis SM. Paroxysmal nocturnal hemoglobinuria, clini-cal manifestations, hematology and nature of the disease. SerHaemataol. 1972;5:3-23.

48. Nissen C, Tichelli A, Gratwohl A, et al. High incidence of tran-siently appearing complement-sensitive bone marrow precursorcells in patients with severe aplastic anemia: a possible role of highendogenous IL-2 in their suppression. Acta Haematol. 1999;101:165-172. -

49. Tichelli A, Gratwohl A, Nissen C, Speck B. Late clonal complica-tions in severe aplastic anemia. Leuk Lymphoma. 1994;12:167-175.

50. Schrezenmeier H. Hertenstein B. Wagner B, Raghavachar A,Heimpel H. A pathogenetic link between aplastic anemia andparoxysmal nocturnal hemoglobinuria is suggested by a high fre-quency of aplastic anemia patients with a deficiency of phos-phatidylinositol glycan anchored proteins. Exp Hematol. 1995;23:81-87.

51. Sultan C, Marquet M, Joffroy Y. Etude de dysmyelopoiesesacquises idiopathiques en culture de moelle in vitro. Nouv Rev FrHematol. 1973;13:431-426.

52. Chen R, Nagarajan S, Prince GM, et al. Impaired growth and ele-vated fas receptor expression in PIGA( +) stem cells in primaryparoxysmal nocturnal hemoglobinuria. J Clin Invest. 2000;106:689-696.

53. Maciejewski JP, Sloand EM, Sato T, Anderson S, Young NS.Impaired hematopoiesis in paroxysmal nocturnal hemoglobinuria!aplastic anemia is not associated with a selective proliferativedefect in the glycosylphosphatidylinositol-anchored protein-defi-cient clone. Blood. 1997;89:1173-1181.

54. Young NS. The problem of clonality in aplastic anemia: Dr.Dameshek's riddle, restated. Blood. 1992;79:1385-1392.

55. Luzzatto L, Bessler M. The dual pathogenesis of paroxysmal noc-turnal hemoglobinuria. Curr Opin Hematol. 1996;3:101-110.

56. Araten DJ, Nafa K, Pakdeesuwan K, Luzzatto L. Clonal popula-tions of hematopoietic cells with paroxysmal nocturnal hemoglo-binuria genotype and phenotype are present in normal individuals.Proc Natl Acad Sci USA. 1999;96:5209-5214.

57. Young NS, Maciejewski J. The pathophysiology of acquired aplas-tic anemia. N Engl J Med. 1997;336:1365-1372.

58. Brodsky RA, Vala MS, Barber JP, Medof ME, Jones RJ. Resis-tance to apoptosis caused by PIG-A gene mutations in paroxys-mal nocturnal hemoglobinuria. Proc Natl Acad Sci USA. 1997;94:8756-8760.

59. Horikawa K, Nakakuma H, Kawagauchi T, et al. Apoptosis resist-ance of blood cells from patients with paroxysmal nocturnal hemo-globinuria, aplastic anemia, and myelodysplastic syndrome. Blood.1997;90:2716-2722.

60. Ware RE, Nishimura J, Moody MA, et al. The PIG-A mutation andabsence of glycosylphosphatidylinositol-linked proteins do notconfer resistance to apoptosis in paroxysmal nocturnal hemoglo-binuria. Blood. 1998;92:2541-2550.

61. Kawagoe K, Kitamura T, Okabe M, et al. Glycosylphosphatidyli-nositol-anchor-deficient mice: implications for clonal dominance ofmutant cells in paroxysmal nocturnal hemoglobinuria. Blood. 1996;87:3600-3606.

62. Lewis SM, Petit JE, Tattersall MH, Pepys MB. Myelosclerosis andparoxysmal nocturnal haemoglobinuria. Scand J Haematol. 1971;8:451-460.

63. Hansen NE, Killman SA. Paroxysmal nocturnal hemoglobinuria inmyelofibrosis. Blood. 1970;36:428-431.

64. Lewis SM, Verwilghen RL. Dyserythropoiesis and dyserythropoi-etic anemias. Prog Hematol. 1973;8:99-128.

65. Longo L, Bessler M, Beris P, Swirsky D, Luzzatto L. Myelodyspla-sia in a patient with pre-existing paroxysmal nocturnal haemoglo-binuria: a clonal disease originating from within a clonal disease. BrJ Haematol. 1994;87:401-403.

78. Strubing P. Paroxysmale Hamoglobinurie. Dtsch Med Wochenschr.1882;8:1-8.

79. Rosse WF, Gutterman LA. The effect of iron therapy in paroxys-mal nocturnal hemoglobinuria. Blood. 1970;36:559-565.

80. Hartmann RC, Jenkins DE Jr, McKee LC, Heyssel RM. Paroxys-mal nocturnal hemoglobinuria: clinical and laboratory studiesrelating to iron metabolism and therapy with androgen and iron.Medicine. 1966;45:331-363. .

81. Tichelli A, Gratwohl A, Nissen C, Speck B. Late haematologicalcomplications in severe aplastic anaemia. Br J Hacmatol. 1988;69:413-418.

82. Rosenfeld SJ, Kimball J, Vining D, Young NS. Intensive immuno-suppression with anti thymocyte globulin and cyclosporine as treat-ment for severe acquired aplastic anemia. Blood. 1995;85:3058-3065.

83. Brodsky RA, Sensenbrenner L, Jones RJ. Complete remission insevere aplastic anemia after high-dose cyclophosphamide withoutbone marrow transplantation. Blood. 1996;87:491-494.

84. Fefer A, Freeman H, Storb R, et al. Paroxysmal nocturnal hemo-globinuria and marrow failure treated by infusion of marrow froman identical twin, Ann Intern Med. 1976;84:692-695.

85. Antin JR, Ginsburg D, Smith BR, et aL Bone marrow transplanta-tion for paroxysmal nocturnal hemoglobinuria: eradication of thePNH clone and documentation of complete lymphohematopoieticengraftment. Blood. 1985;66:1247-1250.

86. Kawahara K, Witherspoon RP, Storb R. Marrow transplantationfor paroxysmal nocturnal hemoglobinuria. Am J Hematol. 1992;39:283-288.

87. Raiola AM, van Lint MT, Lamparelli T, et aL Bone marrow trans-plantation for paroxysmal nocturnal hemoglobinuria. Haematolog-ica. 2000;85:59-62.

88. Suenaga K, Kanda Y, Niiya H, et aL Successful application of non-myeloablative transplantation for paroxysmal nocturnal hemoglo-binuria. Exp Hematol. 2001;29:639-642.

89. Nishimura J, Phillips KL, Ware RE, et aL Efficient retrovirus-mediated PIG.A gene transfer and stable restoration of GPI-anchored protein expression in cells with the PNH phenotype.Blood. 2001;97:3004-3010.

66. Nagakura S, Kawaguchi T, Fujimoto K, et al. Sequential develop-ment of myelodysplasia and paroxysmal nocturnal hemoglobinuriain a patient with preceding aplastic anemia. Int J Hematol. 1997;65:187-189.

67. Dunn DE, Tanawattancharoen P, Boccuni P, et al. Paroxysmal noc-turnal hemoglobinuria cells in patients with bone marrow failuresyndromes. Ann Intern Med. 1999;21:467-468.

68. Holden D, Lichtman H. Paroxysmal nocturnal hemoglobinuriawith acute leukemia. Blood. 1969;33:283-286.

69. Jenkins DE Jr, Hartmann RC. Paroxysmal nocturnal hemoglobin-uria terminating in acute myeloblastic leukemia. Blood. 1969;33:274-282.

70. Kaufmann RW, Schechter G, McFarland W. Paroxysmal nocturnalhemoglobinuria terminating in acute granulocytic leukemia.Blood. 1969;22:287-291.

71. Harris JW, Koscick R, Lazarus HM, Eschleman JR, Medof ME.Leukemia arising out of paroxysmal nocturnal hemoglobinuria.Leuk Lymphoma. 1999;32:401-426.

72. Wasi P, Kruetrachue M, Na-Nakorn S. Aplastic anemia paroxysmalnocturnal hemoglobinuria syndrome- acute leukemia in the samepatients: the first record of such occurrence. J Med Assoc Thai.1970;53:656-663.

73. Hiroshige Y, Marsumoto N; Ha.rima K, et al. An autopsy case ofaplastic anemia-PNH syndrome terminating in acute granulocyticleukemia. Nippon Ketsueki Gakkai Zasshi. 1977;40:16-23,74. van Kamp H, Smit JW, van den Berg E, Roud-Halle M, Vellenga E.Myelodysplasia following paroxysmal nocturnal haemoglobinuria:evidence for the emergence of a separate clone. Br J Haematol.1994;87:399-400.

75. Jin J- Y, Tooze JA, Marsh JC, Mathhey F, Gordon-Smith EC.Myelodysplasia following aplastic anaemia-paroxysmal nocturnalhaemoglobinuria syndrome after treatment with immuno-suppres-sion and G-CSF: evidence for the emergence of a separate clone.Br J Haematol. 1996;94:510-512.

76. Naka.kuma H, Nagakura S, Iwamoto N, et al. Paroxysmal nocturnalhemoglobinuria clone in bone marrow of patients with pancytope-nia. Blood. 1995;85:1371-1376.

77. Sears DA, Anderson PR, Foy AL, Williams HL, Crosby WHo Uri-nary iron excretion and renal metabolism of hemoglobin inhemolytic diseases. Blood. 1966;28:708-725.