British Journal of Haernatology, 1990, 74, 320-329

Studies on a patient with thrombocytopenia, giant platelets and a platelet membrane glycoprotein Ib with reduced amount of sialic acid A. M. AAKHUS,P. STAVEM,* T. HOVIG,? T. M. PEDERSEN A N D N. 0. SOLUM Research Institutefor Internal Medicine, *Medical Department A, and TInstitute of Pathology, Rikshospitalet, University of Oslo, Norway Received 19 June 1989; accepted for publication 3 November 1989

Summary. A 70-year-old patient with a life-long bleeding tendency, giant platelets and thrombocytopenia (1040 x lo9 platelets/l) has been studied. This is a condition often associated with lack of platelet membrane glycoprotein Ib (GP Ib). Electron microscopy of fixed platelets incubated with monoclonal antibodies to GP Ib (AN 5 1, AP 1) and goldlabelled goat anti-mouse IgG. showed a distinct distribution of GP Ib on the patient’s platelets, however. Crossed immunoelectrophoresis and SDS-PAGE demonstrated a reduced mobility of the patient’s GP Ib which could be explained by absence of sialic acid. Blotting with peroxidase-conjugated peanut agglutinin confirmed this conclusion. This lectin binds to galactose-N-acetyl-galactosamine residues exposed

terminally when sialic acid is absent from the carbohydrate side-chains. Such binding could be seen with normal GP Ib only after neuraminidase treatment. Fluorescence studies with FITC-conjugated peanut agglutinin showed binding of the lectin to intact patient platelets, indicating that lack of sialic acid was not introduced during the platelet isolation procedure. Neither could the lack of sialic acid be attributed to increased neuraminidase activity as studied in vitro. Platelets treated with neuraminidase in vivo or in vitro are rapidly cleared from the circulation. Therefore the patient’s thrombocytopenia may be associated with the reduced amount of GP Ib sialic acid. As far as we know, similar cases have not been described previously.

Normal haemostasis requires an adequate number of functional platelets. Thrombocytopenia associated with giant platelets. reduced membrane sialic acid. and shortened platelet survival has been described (for review see George et al, 1984). The membrane defect concerns particularly the glycoprotein Ib (GP Ib) which exerts an essential receptor function in the initial phases of haemostasis. GP Ib is a carbohydrate-rich glycoprotein with a molecular mass of approximately 170000 Da. It is composed of an a-chain (M,=145000)anda~-chain(M,=25000)whicharelinked by disulphide bonds (Phillips & Agin. 1 977). Glycocalicin is a major proteolytic split product from the GP Ib a-chain (Solum et al, 1980a. b). It contains 40-60% (w/w) of carbohydrate and principal sugars are galactose, N-acetyl-glucosamine, Nacetyl-galactosamine, and N-acetyl-neuraminic acid (sialic acid) in a molar ratio of 2 : 1: 1:2 (Okumura et al, 1976). The sialic acid is located terminally linked to either ga1actose-Nacetyl-galactosamine or galactose-N-acetyl-glucosamine in the major carbohydrate side chains. The negative charge of

the sialic acid in GP Ib contributes significantly to the overall negative charge of the platelet. Removal of the sialic acid leads to a reduced electrophoretic mobility of GP Ib even in the presence of SDS (for reviews see Solum et al. 1980a; Clemetson et al, 1981; Tsuji et af. 1983). A lectin from Arachis hypogea, called peanut agglutinin (PNA), will bind to galactosyl residues terminally exposed as part of the galactose-N-acetyl-galactosamine disaccharide (Uhlenbruck et af, 1969; Presant & Kornfeld, 1972: Clemetson et al, 1981). The presence of giant platelets is a feature observed in the clinical conditions described as Bernard-Soulier syndrome (Bernard & Soulier, 1948).the May-Hegglin anomaly (Oski et al, 1962) as well as in the Montreal Platelet Syndrome (Lacombe & d’Angelo, 1963). In its classical form the Bernard-Soulier syndrome, which is the one best studied in terms of membrane glycoproteins. is a rare hereditary bleeding disorder characterized by thrombocytopenia. giant platelets, reduced content ofmembrane sialic acid, and lack of GP Ib (Groettum & Solum, 1969; Nurden & Caen. 1975; Jenkins et al, 1 976; Solum et al, 1977).Also the glycoproteins GP V and GP IX have been found to be absent or deficient in this disorder (Clemetson et al, 1982; Berndt et al, 1983).

Correspondence:Dr Anne M. Aakhus, Research Institute for Internal Medicine, Rikshospitalet, 0027 Oslo 1, Norway.

320

Reduced Sialic Acid on GPIb i n Giant Platelets We here present studies on a patient with a life-long bleeding tendency, thrombocytopenia, and giant platelets, with no clear classificationrelated to the classical forms of the above-mentioned anomalies. Our studies show the presence of GP Ib molecules with a reduced amount of sialic acid, as judged from electrophoretic mobility and binding of the peanut agglutinin. To our knowledge, similar cases have not been described previously.

MATERIALS AND METHODS

Polyclonal antisera. Rabbit antisera to human whole platelet proteins and to glycocalicin (and thus to GP tb) were prepared as described previously (Hagen et al, 1979: Solum et al, 1980a). Monoclonal antibodies. Monoclonal antibodies to the GP Ib cl-chain were generous gifts from Dr J. M. Wilkinson, London, U.K. (PM6 40. PM6 108),Dr J. L. McGregor, Lyon, France, (P 1).and Dr T. J. Kunicki, Milwaukee, Wisconsin, U.S.A. (AP I ) , or purchased from Dakopatts A/S, Glostrup, Denmark (AN 51). AP 1 and AN 51 are considered to be directed towards the von Willebrand factor binding area of the GP Ib %-chain. Monoclonal antibody to GP IX (FMC 25) was a generous gift from Dr M. Berndt, Westmead. Australia. Dynabeads M-450 (Dynal, Oslo, Norway),coated with sheep anti-mouse IgG (10pg/ml) and a monoclonal antibody to GP IIb-IIIa (IT1Pl-l), were a gift from Dr G. Gaudernack, Oslo, Norway. Btiffers. The following buffers were used throughout the study. Platelet washing solution: 148 mmol/l NaCl, 5 mmol/l glucose, 0.6 mmol/l Na-EDTA and 20 mmol/l Tris-HC1 (pH= 7.4), Triton X-100 extraction buffer: 38 mmol/l Tris, 100 mmol/l glycine (pH=8.7) containing 4.2 mmol/l leupeptin. 1.5 mg/ml soybean trypsin inhibitor (SBTI) and 1%(v/v) Triton X-100, Tris-saline buffer: 20 mmol/l Tris. 1 5 0 mmol/l NaCl containing 1.5 mg/ml SBTI (pH= 7.4). Transfer buffer: 2 5 mmol/l Tris, 192 mmol/l glycine and 10% methanol (pH= 8 . 3 ) . Blotting buffer: 0.25% gelatine in 200 mmol/l NaCI, 5 mmol/l Na-EDTA, 0.05% Triton X-100 (or Non Idet P-40) and 50 mmol/l Tris-HC1(pH = 7.5), Neuraminidase buffer: 50 mmol/l Na-acetate, 154 mmol/l NaC1, 9 mmol/l CaCI, (pH= 5.5). Platelet isolation. 36 ml venous blood obtained from the patient or normal donors was drawn into a mixture of 4 ml 54 mmol/l Na-EDTA and 120 mmol/l NaCl in conical plastic tubes after discarding the first 2 ml of blood. Due to the abnormally large platelets observed in vitro, the patient’s blood was centrifuged at low speed ( 5 0 9 for 10 min) to obtain platelet-rich plasma (PKP). The PKP was then centrifuged at 2000 g for 10 min, and the platelet pellet suspended in the platelet washing solution described above. The washing procedure was repeated once. A concentrated platelet suspension was centrifuged in a narrow plastic tube in a Beckman microfuge at 1 2 0 0 0 g for 1 min. The contaminating erythrocytes were sedimented in the bottom of the tube. and this lower part of the tube was cut off and discarded. This procedure was repeated once. The pellet was then either extracted in Triton X-100 extraction buffer at

321

1 x 1 O Yplatelets/ml for preparation of a Triton X-100 extract, or suspended in Tris-saline buffer at 1 x lo9platelets/ml and subjected to freezing and thawing three times. The soluble fraction (platelet lysate) obtained after centrifugation contains glycocalicin. Normal platelets were isolated in a standard platelet isolation procedure as described previously (Solum & Olsen, 1984) and adjusted to 5 x 10’ platelets/ml. Platelet isolation was performed at 20°C. Platelets were counted in Whole Blood Analyzer 148 C. Analys Instrument AB, Sweden. Size distribution of platelets in EDTA-PKP was analysed using the same apparatus. Platelet agglutination with bovine von Willebrand factor or ristocetin. A microtitre plate with stir bars in the wells was placed on a magnetic stirrer. EDTA-PKP (150 pl) was added and stirred at room temperature with bovine von Willebrand €actor(30pl)(Solum&Oisen,1985)orristocetin(30plof12 mg/ml dissolved in Tris-saline). To study the inhibitory effect of AN 5 1,EDTA-PRP ( 150 pl) was incubated with AN 51 (10 pl) for 1 0 min before addition of ristocetin. The agglutination response was described as positive or negative by visual inspection after 2 min. Platelet aggregation with ADP and collagen. Using the microtitre system described above, PKP from citrate anticoagulated blood was stirred at room temperature with ADP (93.7-1.9 pmol/l, final concentration) or collagen (8.41.05 pg/ml, final concentration) in microtitre wells. The aggregation response was described as positive or negative by visual inspection after 2 min. Initnunogold technique and preparation for transmission electron microscopy ( T E M ) . 3 ml blood were drawn directly into 7 ml of 0.2 5% glutaraldehyde in Gey’s buffer (Breton-Goriuset al, 1983), and fixed for 1 5 min at 20°C. The platelets were isolated by centrifugation and washed three times in Gey’s buffer. The immuno-labelling was carried out according to Breton-Gorius et a1 (1983). briefly as follows. The platelets were incubated with AP 1 or AN 51 for 60 min at 20°C and after being washed twice, incubated with goat anti-mouse IgG conjugated to gold particles with a diameter of 5 or 1 5 nm at 4°C for 60min. Thereafter the platelets were washed twice and fixed in 2% glutaraldehyde in 100 mmol/l cacodylate buffer for 30 min and post-fixed in 1%osmium tetroxide for 1 5 min. The specimens were embedded in Epon 812. The ultrathin sections were stained with uranyl acetate and lead citrate. The sections were examined in a Jeol 1200EX electron microscope. Control experiments were carried out by comparing the platelets with non-reacting cells (red blood cells and leucocytes) and by omitting the primary antibody in the reaction. Morphometric analysis was carried out according to M. Roeger (unpublished data), using a Joyce-Loebl Magiscan computerized image analyser system. lmrnunofluorescence. EDTA-PKP or isolated platelets ( 3 x loxplatelets/ml) were incubated in microtitre wells with monoclonal antibodies to GP Ib (AN 5 1,AP 1, P I ) for 30 min at room temperature. After washing, the platelets were incubated for 30 min with fluorescein isothiocyanate (FITC)conjugated goat anti-mouse IgG/M as secondary antibody.

322

A. M . Aakhus et a1

After washing, the supernatant was discarded and the pellet resuspended in washing solution before fluorescence microscopy in a Nikon microscope. Ascites from a non-producing mouse myeloma cell line (X-63) was used for negative control. The film used for photographs was Ektachrome 400. Crossed immunoelectrophoresis (CIE) was performed in the regular system as described by Hagen et a1 (1979) or in the Pharmacia PhastSystem as described in detail by Solum et a2 (1989). The gels were stained by 0.5% Coomassie Brilliant Blue R in 10%acetic acid and 45% ethanol. Passive imniunoblotting after CIE. Immunoblotting of monoclonal antibodies coprecipitated with their antigens during CIE was performed as described by Thorsen et a1 (1987). Sodium dodeeyl sulphate polyacrylaniide gel electrophoresis (SDS-PAGE). SDS-PAGE was performed as described by Solum & Olsen (198 5). Blotting procedures. Transfer of proteins from polyacrylamide gels to nitrocellulose membranes (Bio-Rad 0.45 pm) was done according to the method of Towbin et aZ(19 79) in the transfer buffer described. The transfer was controlled by staining separate lanes with 0.1% Amido Black in 25% isopropanol and 10%acetic acid. Blocking of unreacted sites was performed with 1% gelatine in the blotting buffer described above. Washing of the nitrocellulose membrane

was performed with the blotting buffer. The nitrocellulose membrane was incubated with either a monoclonal antibody for the GP Ib cc-chain as primary antibody, followed by rabbit anti-mouse IgG conjugated with horseradish peroxidase as secondary antibody, or directly with PNA-lectin conjugated with horseradish peroxidase (2 pg/ml in blotting buffer). Visualization of either the bound antibody or lectin was obtained in 0.005% o-dianisidin and 0.015% H L 0 2 . Isolation of leucocgtes. Lymphocytes were isolated on Lymphoprep according to Boyum (1968).Granulocytes were isolated on 6% dextran 500 as described by Boogaerts et a1 (1986). Lectin fluorescence. EDTA-PRP, isolated platelets ( 3 x lo8 platelets/ml) or isolated granulocytes (1-3 x lo5 cells/ml) (200 p1 of each), were incubated in microtitre wells with FITC-conjugated PNA-lectin or FITC-conjugated Helix poniatia-lectin (10p1 of 1mg/ml in 154 mmol/l NaCI) for 30 min at room temperature. To obtain positive controls, platelets were suspended in the neuraminidase buffer and incubated with neuraminidase (Vibrio cholerae, 0.15 U/5 x 1 O Yplatelets) at 37°C for 30 min. Samples were treated as described under immunofluorescence. Lectin Jluorescence on platelets isolated with magnPtic beads. EDTA-anticoagulated blood (50 p l ) was incubated with

Fig 1A. ELectron micrograph demonstrating large platelets (PLT) surface labelled with AP 1 and gold particles conjugated to goat anti-mouse IgG.

Xote that the red blood cell (RBC) shows no labelling.

Reduced Sialic Acid on GPlb in Giant Platelets magnetic beads coated with a monoclonal antibody to platelet membrane glycoprotein GP IIb-IIIa (IT1PI-1) ( 5 pl) in microtitre wells on ice for 20 min. After washing by centrifugation, the beads/platelets were incubated with FITCconjugated PNA-lectin (10 pl of 1 mg/ml in 154 mmol/l NaCI) and treated as described above. Determination of protein. Protein concentration was determined in lysates from frozen and thawed platelets or granulocytes suspended in 154 mmol/l NaCl using the BioRad protein assay kit based on Bradford (1976) with bovine serum albumin as standard. Determination of sialic acid and neuraminidase activity. Sialic acid was determined by the thiobarbituric acid method of Warren (1959). Neuraminidase activity in lysates from frozen and thawed platelets (5 x 10’ normal platelets/ml or 1 x loy patient platelets/ml), or granulocytes ( 3 x loi cells/ ml) was determined by the ability of these samples to split off sialic acid from crl-acid glycoprotein (1mg/ml in neuraminidase buffer) at 3 7°C. Free sialic acid was measured after 0 and 10 min of incubation. The concentration of sialic acid split off from ccl-acid glycoprotein (1mg/ml) was linear with Vibrio cholerae neuraminidase concentration up to 0.15 U/ml when incubated for 10 min at 37°C. The activity expressed in enzyme units, was assessed by direct comparison to commercial neuraminidase from Vibrio cholerae for which one unit is defined as the ability to split off N-acetyl-neuraminic acid

323

from human al-acid glycoprotein (10 mg/ml) in 1 min at 37°C and pH= 5.5 in neuraminidase buffer. Patient. The patient, a male born in 1919, has experienced a life-long bleeding tendency. Thrombocytopenia was first reported in 1941. Between 1978 and 1988 eight platelet counts were performed using a counting chamber. These gave a mean platelet count of 21 x loy platelets/l with a range of 12-33 x 10’ platelets/l. On two occasions previous to the present studies it was recorded that approximately 90% of the platelets were large (5-8 pm in diameter), leading to a tentative diagnosis of Bernard-Soulier’s syndrome. Between 1941 and 1979 he received about 150 blood transfusions. During the present investigation he has not had any transfusions. Blood group is 0 . Splenectomy in 1 9 5 5 did not improve his bleeding tendency, his bleeding time still exceeded 30 min when examined by us by the Ivy technique. Coagulation and fibrinolysis parameters were within normal ranges (thrombotest, normotest, cephotest, thrombin time, ethanol test, fibrin degradation products and fibrinogen). Megakaryocyte count and morphology was normal. No other relatives are reported to suffer from a bleeding tendency. The only relatives available for studies are three of the patient’s children. These are two females and one male born in 1946, 1955 and 1948, respectively. For all three, the platelet counts were in the range of 200-400 x 10’ platelets/l with a median platelet size in the range of 2-2.5 pm as tested on

+

Fig 1B. Higher magnification demonstrating extensive surface labelling. Dense tubular system (DTS), cc-granules (GR).

324

A. M . Aakhus et al

blood smears made from EDTA-anticoagulated blood. These are normal values. During the last years the patient has been treated for moderately elevated blood pressure and for ‘muscular rheumatism’ or fibrositis by daily doses of 2 x 1 mg prazosine chloride and 5 mg prednisone.

RESULTS The platelet studies reported here were performed on four occasions during the period September 1986 to January 1989. During this period the patient’s whole blood platelet count varied between 10 and 40 x 1Oy platelets/l. Due to the low platelet count, platelet aggregation could not be studied by aggregometry. However, the patient platelets in citrated PRP were aggregated by collagen and ADP when stirred in microtitre wells as observed visually. Also, a distinct agglutination of the patient platelets in EDTAPRP with bovine von Willebrand factor or ristocetin could be observed by this technique. The agglutination was inhibited by the monoclonal antibody AN 5 1 directed against the von Willebrand factor binding domain of GP Ib.

Platelet size The large platelet size previously reported was confirmed by electron microscopy (Fig 1A. B) combined with morphometric analyses. The platelets in blood drawn directly into fixative had a mean area of 4.32 (It 3.04) pm’ as compared to 2.19 ( 11.46) pm2for normal platelets. The values for the longestaxiswere4~20(+1~46)pmversus2~65 (f0.72)pm and for the shortest axis 2.66 (k1.10)pm versus 1.17 [ 1 0 . 5 8 ) pm, comparing patient and normal platelets. Size distribution for platelets in EDTA-PRP gave a range for patient platelets from 2 to 30 fl with a peak value at 1 5 fl. The

distribution of normal platelets tested simultaneously was determined to be 2-20 fl with a peak value at 6.5 fl.

Binding of monoclonal antibodies The presence of GP Ib on the patient’s platelets was further confirmed by immunofluorescence microscopy with nionoclonal antibodies directed against GP Ib (AN 51, P 1 or AP 1) as primary antibodies (data not shown), as well as by electron microscopy using the immunogold technique with the same antibodies. This latter technique demonstrated the presence of GP Ib on the patient’s giant platelets (Fig 1A, B). Crossed ininiunoelectrophoresis Crossed immunoelectrophoresis of Triton X-100 extracts of normal platelets run against antiserum to glycocalicin showed the classic immunoprecipitate pattern with separate GP Ib and glycocalicin peaks (Fig 2a). CIE of Triton X-100 extracts of patient platelets prepared in the presence of SBTI to inhibit proteolysis, gave an immunoprecipitate arc with a more cathodal position than that of regular GP Ib. whereas no peak was observed at the position of normal glycocalicin (Fig 2b). However, CIE of lysates (soluble fraction) from frozen and thawed patient platelets revealed a glycocalicinimmunoprecipitate (Fig 2d), with a reduced mobility in the first dimension electrophoresis compared to that of normal glycocalicin (Fig 2c). CIE of lysate from freeze-thawed patient platelets against antiserum to glycocalicin was combined with immunoblotting of a coprecipitated monoclonal antibody directed against GP Ib (AP 1),according to a technique preivously described at this institute (Thorsen et 01, 1987). This confirmed that the immunoprecipitate observed was undoubtedly derived from GP Ib (Fig 3 ) . The same technique performed with Triton X100 extracts using a monoclonal antibody to GP IX (FMC

Fig 2 . Crossed immunoelectrophoresis against antiserum to glycocalicin of Triton X - 1 0 0 platelet extracts (a and b) or lysates obtained by freezing and thawing of platelets (c and d) with both preparations made in the presence of SBTI (1.5 mgiml). a and c: 10 pl of normal platelet sample ( 5 x 10’ pIateIets/mI), b and d: 10 p1 of patient platelet sample (1x lo9 platelets/ml) stained by Coomassie Brilliant Blue R.

Reduced Sialic Acid on GPIb in Giant Platelets

Fig 3. Crossed immunoelectrophoresis combined with passive immunoblotting of a coprecipitated monoclonal antibody to GP Ib. Lysate obtained by freezing and thawing of patient platelets (5 pl of 1 x lo9 platelets/ml) was mixed with AP 1 (1 pl) and subjected to crossed immunoelectrophoresis against antiserum to glycocalicin. The coprecipitated monoclonal antibody was transferred to nitrocellulose membrane and immunostained with rabbit anti-mouse antiserum conjugated with horseradish peroxidase. (a) Gel after passive immunoblotting stained by Coomassie Brilliant Blue R (b) immunostained nitrocellulose memtirane.

25), demonstrated the presence of GP IX in association with GP Ib from the patient platelets (data not shown). CIE of Triton X-100 extracts of patient platelets against antiserum to whole platelet proteins clearly showed the presence of the major immunoprecipitates, i.e. those of GP IIb-IIIa, albumin and fibrinogen. The abnormal behaviour of GP Ib was also observed with this antiserum (data not shown).

SDS-PAGE with Western blotting and blotting with PNA-lectin Triton X-100 extracts of patient and normal platelets were subjected to SDS-PAGE combined with electrophoretic transfer of proteins to nitrocelluiose membranes. Immunovisualization using monoclonal antibodies to GP Ib as primary antibodies showed that the patient GP Ib had a significantly reduced mobility compared to normal GP Ib (Fig 4, lane e compared to lane d, and Fig 6, lane g compared to lanes e and f). Triton X-100 extracts of normal platelets treated with neuraminidase prior to extraction showed a reduced mobility of desialylated normal GP Ib (Fig 4, lane f), whereas this was not observed after incubation of normal platelets with neuraminidase buffer (Fig 4, lane g). Fig 4. lanes b and c show Coomassie Blue stained lanes of normal and patient platelet extracts, respectively, demonstrating

325

Fig 4. SDS-PAGE on a 7% gel combined with Western blotting of Triton X-100 extracts of normal and patient platelets using a monoclonal antibody to GP Ib (PM6 108) as a primary antibody. Triton X-100 extracts were prepared in the presence of leupeptin (4.2 mmol/l) and SBTI ( 1 . 5 mgiml), and diluted 1: 1in SDS-sample buffer before reduction with 3% 2-mercaptoethanol. The applied volume of all samples was 2 0 PI. Lanes a, b. c: gel stained by Coomassie Brilliant Blue G; lanes d, e. f, g: nitrocellulose membrane after electrophoretic transfer, immunostained with a monoclonal antibody to GP Ib (PM6 108) as primary antibody and rabbit anti-mouse antiserum conjugated with horseradish peroxidase, as secondary antibody. Lane a: molecular weight standards, lanes b and d: normal platelets (2.5 x 10’ platelets/ml). lanes c and e: patient platelets (0.5 x 10’ platelets/ml), lane f: normal platelets (2.5 x 10’ platelets/rnl) incubated with neuraminidase (0.15 U/ml at pH=5.5 at 37°C for 30 min) prior to extraction, and lane g: normal platelets (2.5 x 10’ platelets/ml) incubated at 37’C for 30 min with neuraminidase buffer.

fairly similar protein profiles in the patient and normal extracts in spite of differences in the staining intensity due to uneven loading of the gel. Immunovisualization using monoclonal antibodies to GP Ib as primary antibodies showed also that glycocalicin, from the soluble fraction of frozen and thawed patient platelets, had a reduced mobility in SDS-PAGE (Fig 5, lane b) compared to normal glycocalicin (Fig 5, lane a), as also observed with normal platelet lysates treated with neuraminidase (Fig 5,

326

A. M . Aakhus et a1

Fig 6. SDS-PAGE on a 7% gel combined with Western blotting of Triton X - 1 0 0 extracts from patient or normal platelets. using either a monoclonal antibody to GP Ib (PM6 40) or horseradish peroxidaseconjugated PNA-lectin. The extracts were prepared in the presence of leupeptin (4.2 mmol/l) and SBTI (1.5 mg/ml) and diluted 1 : I in SDS-sample buffer before reduction in 3% 2-mercaptoethanol. A sample of normal platelets was also dissolved directly in SDS-sample buffer and reduced with 3% 2-mercaptoethanol. Lane a: gel stained by Coomassie Brilliant Blue G. lanes b, c, d nitrocellulose membrane stained with horseradish peroxidase-conjugated PNA-lectin ( 2 big/ ml). lanes e. f. g: nitrocellulose membrane visualized with a monoclonal antibody to GP Ib (PM6 40) as described in Fig 4.10 pl of each sample was applied. Lane a: molecular weight standards, lanes b and e: normal platelets ( 5 x loyplatelets/ml) solubilized directly in SDS-sample buffer, lanes c and f: Triton X-100 extracts of normal platelets (2.5 x 10’ platelets/ml). lanes d and g: Triton X-100 extracts of patient platelets (0.5 x 10’ platelets/ml). Fig 5. SDS-PAGE on a 7% gel combined with Western blotting of lysates from normal and patient platelets using a monoclonal antibody to GP Ib (PM6 108)as primary antibody. Lysates obtained by freezing and thawing in the presence of SBTI ( 1 . 5 mg/ml). diluted 1: 1 in SDS-sample buffer before reduction with 3% 2-mercaptoethanol. The applied volume of all samples was 2 0 pl. Lane a: normal platelets ( 2 . 5 x 10’platelets/ml), lane b: patient platelets (0.5 x 10’ pIate!ets/mI). lane c: normal platelets (2.5 x 10y platelets/ml) preincubated with neuraminidase (as in Fig 4, lane 0 , and lane d: normal platelets (2.5 x 10’ platelets/ml) preincubated in buffer only (as in Fig 4, lane g). The nitrocellulose membrane was immunostained as in Fig 4.

lane c). Fig 5, lane d shows a normal platelet lysate incubated with neuraminidase buffer as control. Following SDS-PAGE, horseradish peroxidase-conjugated PNA-lectin was used to detect free galactosyl residues on glycoproteins after electrophoretic transfer to nitrocellulose membranes (Fig 6, lanes b, c and d). A positive reaction was observed with patient extracts (Fig 6, lane d )in the molecular weight area corresponding to patient GP Ib ct as detected by a monoclonal antibody to GP Ib (Fig 6, lane g). No staining was observed with normal platelet material whether this was obtained by dissolving the platelets directly in SDS-sample buffer (Fig 6, lane b) or prepared via extraction in Triton X100 buffer (Fig 6, lane c), unless these had been pretreated with neuraminidase (data not shown).

Fig 7. Fluorescence microscopy of platelet-rich plasma from the patient incubated with FITC-conjugatedPNA-lectin (50 pg/ml. final concentration). Magnification x 500.

Binding of PNA-lectin to blood cells Incubation of patient EDTA-PRP with FITC-conjugated PNAlectin showed a strong fluorescence of the platelets (Fig 7 ) , whereas practically no reaction was observed with normal platelets unless these had been preincubated with neuraminidase. Isolated granulocytes or erythrocytes from patient or normal blood did not react with FITC-conjugated PNA-lectin

Reduced Sialic Acid on GPlb in Giant Platelets

Fig 8. Fluorescence microscopy of patient platelets isolated by magnetic beads coated with a monoclonal antibody to GP IIb-IIIa (IT1 P1-1) and incubated with FITC-conjugated PNA-lectin (15 0 pg/ ml, final concentration). Magnification x 500.

(data not shown). Neither normal nor patient platelets gave a reaction with the Helix pomatia-lectin. Fig 8 shows a positive reaction with FITC-conjugated PNAlectin on patient platelets isolated from whole blood with magnetic beads conjugated with a monoclonal antibody to GP IIb-IIIa. Neuraminidase activity Neuraminidase activity in lysates obtained by freezing and thawing of platelets or granulocytes was practically unmeasurable both in patient and normal samples, i.e. less than 0.001 U/mg protein. DISCUSSION We have studied a patient with thrombocytopenia, giant platelets, and reduced amount of sialic acid in GP Ib. Qualitatively the platelets respond normally to ADP and collagen. and also to bovine von Willebrand factor or ristocetin. The presence of GP Ib was demonstrated by several immunological techniques including electron microscopy, which clearly showed its presence on the giant platelets. Therefore we do not regard the patient as a classical BernardSoulier syndrome patient. It is not possible from our studies to decide whether the patient’s platelets contain a normal surface density of GP Ib or not. Doehle bodies (Oski et al, 1962) were not present in the patient granulocytes, thus excluding the May-Hegglin anomaly, another macrothrombocytopenic bleeding disorder with the presence of platelet GP Ib (Coller & Zarrabi, 1981 ). Neither was there any indication of spontaneously formed platelet aggregates typical of the Montreal Platelet Syndrome patients (Milton et al, 1984).

327

A reduced electrophoretic mobility of patient GP Ib was observed on CIE for both GP Ib and glycocalicin immunoprecipitates. Coprecipitation of a monoclonal antibody to GP Ib (AP l ) ,confirmed that these really represented GP Ibrelated material. If the altered mobility of GP Ib were due to proteolytic action, one would expect to observe a degraded GP Ib with a lower molecular mass than normal by SDS-PAGE. Although this was occasionally observed, we found repeatedly that both GP Ib and glycocalicin from the patient migrated to a position of apparently higher molecular weight than normal GP Ib and glycocalicin. The reduced migration of GP Ib on SDS-PAGE could be explained by a reduced sialic acid content as treatment of normal platelets with neuraminidase also resulted in such a reduced migration. This has also been observed by others (Nurden, 1977; Clemetson et al, 1981). This apparent anomaly is explained by the fact that the high content of sialic acid contributes to the migration of GP Ib even in SDS-PAGE. Removal of sialic acid from human platelets by neuraminidase exposes the disaccharide galactose-N-acetyl-galactosamine (Uhlenbruck et a!, 1969: Presant & Kornfeld, 1972) which can be detected by the lectin peanut agglutinin (Gloeckneretal, 1978). Clemetsonet a1 (1981)andNaimeta1 (1982) showed that after treatment of normal platelets by neuraminidase. peanut agglutinin binds only to GP Ib. This is consistent with our observations. With untreated platelets from the patient, the GP Ib crchain reacted positively to peanut agglutinin. Normal untreated platelets did not react. Thiy indicated that the patient’s GP Ib ct oligosaccharide chains had terminal galactose-N-acetyl-galactosamine instead of the terminal sialic acid normally present. As the positive reaction to peanut agglutinin was observed in samples of GP Ib cr as well as in samples containing the water-soluble glycocalicin, we regard the side-chain abnormality to be located in the glycocalicin area of the GP Ib cr-chain. Fluorescence studies showed that peanut agglutinin also bound to the patient platelets in PRP, indicating that the modified GP Ib was present on the patient platelets prior to their isolation. This suggests that the carbohydrate anomaly may exist in vivo, either due to defective biosynthesis or to enzymatic removal of sialic acid. The use of magnetic beads coated with a platelet specific monoclonal antibody against the GPIIb-IIIa complex to isolate the patient platelets for the fluorescence studies, confirmed that the reaction of the lectin was directed towards platelets. Also, binding ofpeanut agglutinin was not observed with other blood cells. However, although the experiments with the magnetic beads were performed with whole blood, we can not exclude the possibility that preparations of PRP select a PNA-positive subpopulation of platelets. Considering the possibility that patient GP Ib might have been enzymatically desialylated by contaminating erythrocytes or leucocytes, we were never able to mimic the reaction to the peanut agglutinin by the addition of such cells to normal platelets prior to cell lysis (data not shown). Also, determination of neuraminidase activity in lysates of patient or normal platelets or granulocytes gave values considered as

328

A. M . Aakhus et a1

insignificant. Further, the fact that the observation interpreted as reduced sialic acid was done on three different occasions over a period of 18 months on the patient in an apparently healthy condition, practically rules out the possibility that these observations were due to viral or bacterial neuraminidase activities. Various acquired carbohydrate anomalies in platelet glycoproteins, including a reduced amount of sialic acid in GP Ib, have been described and related to malignant diseases (Clezardin et al, 1985). Our patient has shown no sign of malignant disease so far. Nurden et a1 (1982) described a patient with the Tnantigen expressed on platelet GP Ib. The Tn-antigen is represented by the N-acetyl-galactosamine group and can be detected by a lectin from Helix pomatia (Prokop et al, 1968). Our patient’s platelets did not react with the Helix pomatialectin. A genetically based polymorphism of GP Ib with variant mobilities in SDS-PAGE has been described (Moroi et al, 1984). as well as a double band of GP Ib in a patient with a bleeding tendency (Bolin et al, 1977). The present studies show that reduced sialic acid also has to be considered when :such variations occur. Platelets treated with neuraminidase in vivo (Choi et al, 19 72) and in vitro (Greenberg et al, 19 7 5) are rapidly cleared from the circulation. Therefore, whereas the presence of sialic acid is probably not a prerequisite for the receptor function of GP Ib (Korrel et al, 1988), this may be important in controlling the platelet life span. Thus, a relationship may exist between the thrombocytopenia and the observed sialic acid deficiency in the patient. ACKNOWLEDGMENTS This work was supported by the Norwegian Council of Cardiovascular Diseases, The Norwegian Research Council for Science and the Humanities, and by Anders Jahres fond ti1 vitenskapens fremme. REFERENCES Bernard, J. & Soulier. J.P. (1948) Sur une nouvelle variete de dystrophie thrombocytaire-hemoragipare congenitale. Semaine des Hdpitaux de Paris, 24, 3227-3223. Bemdt, M.C., Gregory, C.. Chong, B.H., Zola. H. & Castaldi, P.A. (198 3) Additional glycoprotein defects in Bernard-Soulier’s syndrome: confirmation of genetic basis by parental analysis. Blood, 62,800-807. Boyum, A. (1968) Isolation of mononuclear cells and granulocytes from human blood. Scandinavian Journal of Clinical and Laboratory Investigation, 21, 77-89. Bolin, R.B., Okumura. T. & Jamieson, G.A. (1977) New polymorphism of platelet membrane glycoproteins. Nature, 269, 69-70. Boogaerts, M.A., Vercelotti. G.. Roelant, C., Malbrain, S., Verwilghen, R.L. & Jacobs. H.S. (1986) Platelets augment granulocyte aggregation and cytotoxicity: uncovering of their effects by improved cell separation techniques using Percoll gradients. Scandinavian Journal of Haematology, 37, 229-236. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Analytical Biochemistry, 72, 248254.

Breton-Gorius, J.. Vanhaeke, D., Tabilio. A., Vainchenker, W.. Deschamps, J.F. & Xer. F.S. (1983) Glycoprotein I identification during normal and pathological megakaryocytic maturation. Blood Cells, 9, 275-291. Choi, S.4.. Simone, J.V. & Journey. L.J. (1972) Neuraminidaseinduced thrombocytopenia in rats. British Journal of Haematology. 22, 93-101. Clemetson, K.J.. McGregor, J.L. &James, E. (1982) Characterization of platelet membrane glycoprotein abnormalities in BernardSoulier syndrome and comparison with normal by surfacelabelling techniques and high-resolution two-dimensional gel electrophoresis. Journal of Clinical Investigation, 70, 304-31 1. Clemetson, K.J.. Naim. H.Y. & Luscher. E.F. (1981) Relationship between glycocalicin and glycoprotein Ib of human platelets. Proceedings of the National Academy ofsciences ofthe United States of America. 78, 2712-271 6. Clezardin. P.. McGregor, J.L., Dechavanne. M. & Clemetson. K.J. (198 5) Platelet membrane glycoprotein abnormalities in patients with myeloproliferative disorders and secondary thrombocytosis. British Journal of Haematology, 60, 331-344. Coller. B.S. & Zarrabi, M.H. (1981) Platelet membrane studies in the May-Hegglin anomaly. Blood, 58, 279-284. George, J.N., Nurden. A.T. &Phillips. D.R. (1984) Molecular defects in interactions of platelets with the vessel wall. New England Journal of Medicine, 31 1, 1084-1098. Gloeckner. W.M., Kaulen, H.D. & Uhlenbruck, G. (1978) Immunochemical detection of the Thomsen-Friedenreich antigen (Tantigen) on platelet plasma membranes. Thrombosis and Haemosfasis, 39, 186-192. Greenberg, J., Packham. M.A., Cazenave, J.-P.. Reirners. H.-J. 6; Mustard, J.F. (1975) Effects on platelet function of removal of platelet sialic acid by neuraminidase. Laboratory Investigation. 32, 476-484. Groettum. K.A. & Solum, N.O. (1969) Congenital thrombocytopenia with giant platelets: a defect in the platelet membrane. British Journal of Haematology, 16, 277-290. Hagen. I., Bjerrum. O.J. & Solum, N.O. (1979) Characterization of human platelet proteins solubilized with Triton X-100 and examined by crossed immunoelectrophoresis. Reference patterns of extracts from whole platelets and isolated membranes. European journal of Biochemistry. 99, 9-22. Jenkins, C.S.P., Phillips, D.R.. Clemetson. K.J.. Meyer. D., Larrieu, M.-J. & Luscher. E.F. (1976) Platelet membrane glycoproteins implicated in ristocetin-induced aggregation. Journal of Clinical Investigation, 57, 112-124. Korrel. S.A.M.. Clemetson, K.J., van Halbeek, H., Johannis, P.K., Sixma, JJ. & Vliegenthart. J.F.G. (1988) Identification of a tetrasialylated monofucosialylated tetraantennary N-linked carbohydrate chain in human platelet glycocalicin. FEBS Letters, 228, 321-326. Lacombe, M. & d’Angelo, G. (1963) Etudes sur une thrombopathie familiale. Nouvelle Revue Fmqaise d‘Himatologie, 3, 61 1-614. Milton, J.G., Frojmovic, M.M., Tang, S.S., & White, J.G. (1984) Spontaneous platelet aggregation in a hereditary giant platelet syndrome (MPS). American Journal of Pathology. 114, 33634s. Moroi, M., Jung. S.M. & Yoshida, N. (1984)Genetic polymorphism of platelet GP Ib. Blood, 64, 622-629. Naim. H.Y.. Clemetson, K.J. & Luscher, E.F. (1982) Effects of galactose-binding lectins on human blood platelets: identity of the peanut agglutinin receptor with the von Willebrand factor receptor. Thrombosis Research, 26, 431-441. Nurden, A.T. (1977) A peripheral high molecular weight glycoprotein located at the surface of human platelets. Experientia, 33, 331-332.

Reduced Sialic Acid on GPZb in Giant Platelets Nurden, A.T. & Cam, J.P. (1975) Specific roles for platelet surface glycoproteins in platelet function. Nature, 255, 720-722. Nurden. A.T.. Dupuis, D.. Pidard, D.. Kieffer. N.. Kunicki. T.J. & Cartron, ].-P. (1982) Surface modifications in the platelets of a patient with a-N-acetyl-D-galactosamine residues, the Tn-syndrome. Iourrral of Clinical Investigation, 70. 1281-1291. Okumura. T.. Lombart. C. & Jamieson. G.A. ( 1 976) Platelet glycocalicin. 11. Purification and characterization. Journal of Biological Chemistty, 251, 5950-5955. Oski. F.A.. Naiman, J.L., Allen, D.M. & Diamond, L.K. (1962) Leukocyte inclusions-Doehle bodies-associated with platelet abnormality (The May-Hegglin anomaly). Report of a family and review of the literature, Blood, 20, 657-667. Phillips, D.R. & Agin. P.P. (1977) Platelet plasma membrane glycoproteins. Evidence for the presence of nonequivalent disulfide bonds using nonreduced-reduced two-dimensional gel electrophoresis. Journal of Biological Chemistry, 252, 2121-2126. Presant, C.A. & Kornfeld. S. (1972) Characterization of the cell surface receptor for the Agaricus bisporus haemagglutinin. Journal of Biological Chemistry, 247, 6937-6945. Prokop, O., Uhlenbruck. G. & Koehler, W. (1968) A new source of antibody-like substances having anti-blood group specificity. Vox Sanguinis. 14, 321-323. Solum, N.O., Aakhns. A.M., Pedersen, T.M. &Gaudernack, G. (1989) The PhastSystem equipment used for crossed immunoelectrophoresis combined with immunoblotting of coprecipitated monoclonal antibodies as studied with platelet membrane receptor proteins. Electrophoresis, 10, 752- 7 58. Solum, N.O.. Hagen, I., Filion-Myklebust, C. & Stabaek. T. (1980a) Platelet glycocalicin: its membrane association and solubilization in aqueous media. Biochirnica et Biophysica Acta. 597, 2 3 5-246. Solum. N.O.. Hagen. I. & Gjemdal. T. (1977) Platelet membrane

329

glycoproteins and the interaction between bovine factor VIII related protein and human platelets. Thrombosis and Haemostasis. 38, 914-923. Solum. N.O., Hagen. I. & Sletbakk, T. (1980b) Further evidence for glycocalicin being derived from a larger amphiphilic platelet membrane glycoprotein. Thrombosis Research, 18, 773-785. Solum, N.O. & Olsen, T.M. (1984) Glycoprotein Ib in the Tritoninsoluble (cytoskeletal) fraction of blood platelets. Biochimica et Biophysica Acta, 799, 209-220. Solum. N.O. & Olsen,T.M. (1985 ) Effects of diamide and dibucaine on platelet glycoprotein Ib. actin-binding protein and cytoskeleton. Biochirnica et Biophysica Acta, 817, 249-260. Thorsen, L.I.. Gaudernack, G., Brosstad, F.. Pedersen. T.M. & Solum. N.O. (1987) Identification of platelet antigens by monoclonal antibodies using crossed immunoelectrophoresis with immunoblotting of the monoclonal antibody. Thrombosis and Haemostasis, 57.212-216. Towbin. H.. Staehelin. T. & Gordon, J. (1979)Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets. Procedure and some applications. Proceedings of the National Academyofsciencesofthe UnitedStatesofArnerica. 76,4350-4354. Tsuji, T., Tsunehisa, S., Watanabe, Y.. Yamamoto, K., Tohyama. H. & Osawa, T. (1983) The carbohydrate moiety of human platelet glycocalicin.Journal of Biological Chemistry. 258, 6335-6339. Uhlenbruck. G.. Pardoe, G.I. &Bird. G.W.G. (1969) On the specificity of Iectins with a broad agglutination spectrum. 11. Studies on the nature of the T-antigen and the specific receptors for the lectin of Arachis hypogea (ground-nut). Zeitschrift fiir Iinmiinitaets~orscfrirng und Kliriische Immunologie, 138, 423-433. Warren, L. (1959) The thiobarbituric acid assay of sialic acid. Journal of Biological Chemistry, 234, 1971-1975.