Autor: C. Thomas Nebe, Klinikum Mannheim, D-68135 Mannheim

The following text does not exactly match the published article:
Nebe, CT, Paroxysmal Nocturnal Hemoglobinuria: Clinical Aspects and Flow Cytometric Analysis
in: Aspects of the Flow-Cytometric Analysis of Red Blood Cells, K. Gutensohn, H.-H.
Sonneborn and P. Kühnl (eds.), 81-94, Clin. Lab. Publications (1997)
 

The disease called paroxysmal nocturnal hemoglobinuria (PNH, also known as Marchiafava-Micheli-syndrome) is a rare, acquired, chronic, intravasal, hemolytic anemia with letal outcome in most cases, caused by complement mediated lysis of red cells (1). It was first described in 1882 by Strübing, later in 1911 by Marchiafava and Nazari and in 1931 by Micheli. The term "paroxysmal nocturnal hemoglobinuria" is poor, as it describes only one feature of the illness and moreover, that occurs in less than one quater of the patients (1). By nature it is an acquired stem cell disease that affects all blood cell lineages.

The patients show up with the uncharacteristic picture of chronic anemia with fatigue, with dysphagia, abdominal pain or headache. The dark coloured urine may have gone unnoticed or was not appreciated by the physician. When the onset of hemolysis is more acute, the diagnosis is made more readily. Sometimes the thrombotic event is the first sign, what makes it even more difficult ((1,2,3), see table 1). The MRT with T1- and T2-weighed spin echo is the method of choice to prove for intraabdominal thrombosis (3a). The age at the time of diagnosis is between 16 and 75 years (median at 42) with a mean survival time of 8-10 years and both genders are affected (3,4). There exist no reliable data on the prevalence and incidence of the disease. Half of the patients die of venous thrombosis or thrombopenic hemorrhage and others by the complications of aplastic anemia (AA). Ca. 10% show spontaneous long term remission (4).

Allogeneic or autologous stem cell transplantation after purging of GPI-deficient precursor cells (4a) is the therapy of choice at the late stage of disease. If aplastic anemia and PNH occur at the same time Cyclosporin A can be used successfully like in cases with AA alone (5). Oral anticoagulation is ultimately required after thrombotic complications but several authors consider this option when diagnosis of PNH is made. In pregnancy when the risk of thrombosis is even higher heparin should be given as normal deliveries have bee reported under continous therapy. Complement activation at low dose heparin should be mentioned. Female patients should avoid oral contraceptives as they increase the risk for venous thrombosis. Supportive therapy consists of red cell or platelet transfusions and the replacement of iron as iron deficiency develops in most patients. The application of cortisone and androgens is under debate (4).

From a clinical point of view PNH can be divided into a primary and secondary form acquired during aplastic anemia. In almost all cases so far the defect could be attributed to a somatic mutation in the PIG-A gene (6,7). This prevents the synthesis of a glycolipid (glykosylphosphatidylinositol, GPI), that anchors several proteins in the outer membrane of the cell. The defect or deficient phosphatidylinositol glycan class A (PIG-A) protein does not allow the transfer of acetylglucosamin on phosphatidylinositol, for the synthesis of GPI (3). The enzyme UDP-GlcNAc:phosphatidylinositol-alpha-1,6-N-acetylglucosaminyltansferase may be the affected gene product (8). By this way GPI anchored proteins are synthesized, however they don´t become inserted into the cell membrane and therefore appear in part at higher levels in the plasma. The GPI anchor allows for a high lateral movement in the cell membrane and the protein can be easily cleaved from the membrane by a phopshoinositol specific phospholipase C. An transmembrane signalling independent from other proteins seems to be impossible for such type of surface proteins.

The proteins which are known as GPI anchored are listed in table 2, however additional ones will add soon. Two of them are involved in the inhibition of the complement cascade: CD55 is the decay accelereating factor and CD59 the membrane inhibitor of reactive lysis preventing the membrane attack by the terminal complex C5-C9. As these proteins are missing, the patient developes a complement mediated intravasal hemolysis that will lead above a certain threshold to hemoglobinuria (9, 9a). CD59 inhibits the polymerization of C9 what seems to be the most relevant step because a mutation in this gene shows a clinical picture similar to PNH. Hemolysis is facilitated by lowering the pH to 6.4 - 6.7 and temperature increase to 37-40°C, which is the basis for the laboratory test (Ham´s test, (1)).

The PIG-A defect appears on the stem cell level and affects all lineages: leucocytes, platelets and erythrocytes. As it is an acquired somatic mutation only some of the precursor cells carry the defect and therefore a mosaic of normal and affected cells occurs that shifts in time towards the GPI deficient cells. It was shown in vitro that the GPI deficient lymphocytes have a proliferation advantage (10) and in vivo reticulocytes contain a higher proportion of defect cells than the erythrocytes (11, 12). This probably explains the progress of the disease. The function of lymphocytes may be affected as well and proliferative defects in their mitogen response were shown(13), a finding that could not be confirmed by others (10). It was observed that the neutrophils have a defect in chemotaxis and phagocytosis that might explain the increased rate of infection in the late stage of PNH (1). However the increased tendency of platelets for venous thrombosis with increased rate of activation dependent antigens waits for a molecular explaination (14). It should also be mentioned that a defect sialylation has been described for glycophorin A in PNH (15).

A suspicion for PNH should always be considered when a patient presents with one of the following signs:

Tests based on hemolysis and esp. the classical laboratory findings may vary during the course of the disease while the membrane defect can always be shown by flow cytometry. Cytometrically all lineages can be analyzed and the early mosaic form be detected in a reliable manner. This encompasses the expression analysis of CD59 and CD55 with suited fluorescent monoclonal antibodies on erythrocytes, thrombocytes and leukocytes. Due to the severity of the diagnosis, other GPI anchored proteins might be included (see table 2). The following points need to be considered:

The primary form should show up with the classical laboratory parameters of a hemolytic anemia which may be missing in the secondary form (see table 3). Before starting the flow test, a complete blood count should be available. The flow cytometric approach has been published in various forms (22-26) and has been optimized in our lab. It consists of three test runs:

The reticulocyte enumeration using the thiazole orange (TO)/LDS method and whole blood is the first test (for details see BERTSCH and NEBE in this volume). Because of the spectral properties of the dyes it allows for an additional simultaneous staining with CD59 phycoerythrin (PE), that presents the fraction of defect reticulocytes compared to erythrocytes.

In a second whole blood assay, 10 µl of 1:10 diluted blood are incubated with the antibodies CD55 bzw. CD59 (eg. Cymbus Bioscience Ltd., Southhampton, UK). A healthy person serves as a positive control and isotype matched antibodies are the negative control. Using titrated and purified directly conjugated moabs, 15 minutes later the cells in the test tube are diluted with 500 µl of phosphate buffered saline (PBS, eg. CellWash, Becton Dickinson) and directly analysed by flow cytometry. Otherwise, afterwards the cells are washed once in PBS and 50,000 events are collected with log amplification for all parameters in list mode data storage. During analysis, a software gate is set around the red cell population or the platelets respectively and the fraction of CD59 or CD55 positive cells compared to the control tube are reported (see fig 1). The three colour fluorescence assay contains TO/CD59-PE/LDS751 and combines the first two tests.

In the third test 50 µl whole blood are incubated with CD59 moab and subsequently the red cells are lysed (eg. FACS-Lysing Solution, Becton Dickinson) and washed once with PBS. 20,000 cells are collected and the CD59 positive neutrophils and lymphocytes can be analysed separately because of their different side scatter properties. Positive and negative controls are done in a similar manner as in the red cell test. The determination of CD16 has to be done with caution as can be seen by the negative cells in the healthy control (see figure 2E). Three colour combinations are CD24-FITC/CD59-PE/CD45-PerCP and CD4-FITC/CD14-PE/CD16-PE-Cy5 where dim CD4 versus side scatter is used to gate on monocytes (FITC=fluoresceinisothiocyanate, PE=R phycoerythrin, PE-Cy5=phycoerythrin-cyanin5 tandem conjugate, PerCP=peridinium chlorophyll A protein).

Finally, the findings are summarized and checked for plausibility using the internal controls, other laboratory test results as mentioned above and clinical data. Primary PNH is diagnosed more easily compared to the secondary form esp. in the early phase. Sometimes consecutive measurements in the course of the disease clarify the situation.

In summary the rare disease PNH or the GPI deficiency is caused by somatic mutation in the PIG-A gene. By using fluorescent monoclonal antibodies against PI-anchored cell surface proteins (CD55 and CD59) and flow cytometric analysis as the method of choice (considering the pitfalls mentioned above), the diagnosis can be done early and reliably. However, because of the often unspecific symptoms, the clinical suspicion is difficult and therefore late. There is a continous transition of PNH towards aplastic anemia. Most of the clinical symptoms may be explained by the loss of the GPI anchored molecules.
 
 
 
 

Literature:

Clinical symptoms of 80 PNH patients at the time of diagnosis (after (2))

symptom frequency

symptoms of anemia 35 %

hemoglobinuria 26 %

hemorrhage 18 %

aplastic anemia 13 %

gastrointestinal symptoms 10 %

hemolytic anemia with jaundice 9 %

iron deficient anemia 6 %

thrombosis or embolism 6 %

infections 5 %

neurological signs or symptoms 4 %
 
 

Table 2

GPI anchored membrane proteins

CD design. distribution function / remark

CD14 monocytes +/++ receptor for the LPS-binding protein
neutrophils (+) expr. maturation- and activation dependent !

CD16 Fcg RIII neutrophils ++ low affinity receptor for IgG
monocytes +/- expr. maturation- and activation dependent !
NK cells on NK cells transmembraneously anchored !

CD24 neutrophils maturation dependent

CD48 Blast1 monocytes
lymphocytes

CD52 Campath-1 lymphocytes partially on granulocytes, eosinophils

CD55 DAF erythrocytes maturation independent, well suited
leukocytes

CD58 LFA3 erythrocytes expressed on many cells in the body
leukocytes

CD59 MIRL erythrocytes maturation independent, well suited
leukocytes

CD66b CD67 all granulocytes

CD73 lymphocytes ecto-5`-nucleotidase, membr. bound enzyme

C8bp erythrocytes complement factor C8 binding protein

Acetylcholinesterase erythrocytes membrane bound enzyme

Alkal. phosphatase neutrophils membrane bound enzyme

further CDw90 (Thy1), CDw108, CDw109

Table 3

laboratory parameters in the diagnosis of PNH

parameter material remark

increased

reticulocytes EDTA blood in primary PNH (> 5%), may be low in AA

bilirubin serum esp. indirect bilirubin
LDH esp. during hemolytic crisis
free Hb

urobilinogen urine
urobilin together > 3 mg/Tag

stercobilinogen stool
stercobilin together > 280 mg/Tag
 

decreased

erythrocytes EDTA blood
leukocytes esp. neutrophils, sometimes little
thrombocytes little decrease
hemoglobin often 7-8 g/dl

haptoglobin serum
 

normal

Coombs test whole blood direct test negative (no autoantibodies)
cold antibodies season !
osmot. resistance EDTA blood normal fragility
 
 

Legends to the figures
 
 
 
 

Figure 1

Figure 1 shows a typical mosaic pattern of a patient with PNH. Other hematological parameters were: Hb13,9 g/dl, Ery 4,2 x10⁶/µl, WBC 3980/µl, retics 3,5% = 145,000/µl

1A: Analysis of CD59 on erythrocytes (1B) after light scatter gating (1A). Figure 1C-F show that a higher proportion (more than 50%) of neutrophils are affected. Triple staining for red cells was CD55-FITC/CD59-PE/LDS751 and for leukocytes CD4-FITC/CD14-PE/CD16-PE-Cy5 and CD24-FITC/CD59-PE/CD45-PerCP respectively.

Figure 2

Figure 2 shows the same tests for a normal donor. Please note the negative dots for CD59 in 2A, CD16 in 2E and CD14 in 2F.