Inheritance of the human platelet alloantigen,PlA1, in type I Glanzmann's thrombasthenia.

T J Kunicki, … , A T Nurden, J P Caen

J Clin Invest. 1981;67(3):717-724. https://doi.org/10.1172/JCI110088.

The hereditary of the human platelet alloantigen, PlA1, has been studied in Glanzmann'sthrombasthenia. The PlA1 content of platelets from three patients, 20 kindred of thesepatients, including parents and siblings, and 15 unrelated normal individuals wasdetermined using immunologic techniques based on the release of 51Cr from labeledplatelets. The amount of membrane glycoproteins (GP) IIb and IIIa in the platelets of theseindividuals was determined by quantitative crossed immunoelectrophoresis of Triton X-100soluble proteins using a multispecific rabbit antibody raised against normal platelets.Platelets from the three thrombasthenic patients contained neither detectable GP IIb and GPIIIa nor detectable PlA1 antigen. Platelets from seven kindred with normal amounts of GP IIband GP IIIa contained PlA1 antigen levels identical to those detected in platelets of normalindividuals. Platelets from 13 kindred, including each parent studied, were shown to containan amount of GP IIb and GP IIIa equivalent to 53% of that amount detected on normalplatelets. Platelets from the same individuals expressed amounts of PlA1 antigen that wereeither 54.0 +/- 4.1 (mean +/- SD) or 28.0 +/- 2.7% of that present on platelets of normalindividuals homozygous for the Al allele. The results presented in this report provideevidence that the expression of the thrombasthenic glycoprotein abnormality and theinheritance of PlA1 antigen are controlled by […]

Research Article

Find the latest version:

http://jci.me/110088-pdf

Inheritance of the Human Platelet Alloantigen,

PlA1 in Type I Glanzmann's Thrombasthenia

THOMASJ. KUNICKI, DOMINIQUEPIDARD, JEAN-PIERRE CAZENAVE,ALANT. NURDEN,and JACQUESP. CAEN, Institut National de la Sante et de la Recherche Medicale,Hopital Lariboisie're, Paris, 10, France; Centre de Transfusion Sanguine, 67085Strasbourg Cedex, France

A B S T RA C T The heredity of the human platelet allo-antigen, PlAl, has been studied in Glanzmann's throm-basthenia. The PlAl content of platelets from three pa-tients, 20 kindred of these patients, including parentsand siblings, and 15 unrelated normal individuals wasdetermined using immunologic techniques based onthe release of 51Cr from labeled platelets. The amountof membrane glycoproteins (GP) lIb and Illa in theplatelets of these individuals was determined byquantitative crossed immunoelectrophoresis of TritonX-100 soluble proteins using a multispecific rabbit anti-body raised against normal platelets.

Platelets from the three thrombasthenic patients con-tained neither detectable GP IIb and GP IlIa nor de-tectable PlAl antigen. Platelets from seven kindred withnormal amounts of GP IIb and GP IIla contained PIAlantigen levels identical to those detected in plateletsof normal individuals.

Platelets from 13 kindred, including each parentstudied, were shown to contain an amount of GP IIband GPIlla equivalent to 53%of that amount detectedon normal platelets. Platelets from the same individualsexpressed amounts of PlAl antigen that were either54.0±4.1 (mean±SD) or 28.0±2.7% of that present onplatelets of normal individuals homozygous for the Alallele.

The results presented in this report provide evidencethat the expression of the thrombasthenic glycoproteinabnormality and the inheritance of PlA' antigen arecontrolled by different genes. These results further sug-gest that lack of expression of the PlAl antigen on throm-basthenic platelets results from the decrease or absenceof the glycoprotein carrier of the PlA' determinant, pre-viously shown to be GP IIIa.

Address reprint requests to Dr. Kunicki at The Blood Centerof Southeastern Wisconsin, Milwaukee, Wis.

Received for publication 15 August 1980 and in revisedform 10 November 1980.

INTRODUCTION

Glanzmann's thrombasthenia (GT)' is an inheriteddisorder of platelet function characterized, in vitro, bythe absence of aggregate formation in the presence ofthe physiologic agents, ADP, collagen, epinephrine,or thrombin, as well as absent or significantly impairedclot retraction, and, in vivo, by impaired thrombusformation (1-3).

Comparisons of the protein and glycoprotein compo-sition of normal and thrombasthenic platelets bysodium dodecyl sulfate (SDS- polyacrylamide gel elec-trophoresis (PAGE) have demonstrated a significant re-duction in the periodic acid-Schiff reagent stain densityof two bands known as glycoprotein (GP) Ilb and GPIlIa2 in thrombasthenic platelet samples (6-8).Furthermore, decreased levels of radioactivity in theseregions of acrylamide gels were observed on analysisof thrombasthenic platelets whose surface proteinswere labeled with 1251 by the lactoperoxidase-catalyzedprocedure (8, 9), and an analysis of thrombasthenicplatelets by two-dimensional reduced-nonreducedSDS-PAGE confirmed an absent Coomassie Blue R(CBR)-staining in the position of GPIlb and GPIIIa (8).

Using crossed immunoelectrophoresis (CIE), Hagenet al. (10) showed that the most prominent CBR-stainedimmunoprecipitate of normal platelet extracts, which isalso labeled with 1251 in samples from platelets io-dinated by the lactoperoxidase method, was absent orsignificantly reduced in extracts from thrombasthenicplatelets. SDS-PAGE of the antigens present in this

1Abbreviations used in this paper: CBR, Coomassie blue R;CIE, crossed immunoelectrophoresis; GP, glycoprotein; GT,Glanzmann's thrombasthenia; PRP, platelet-rich plasma;SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel elec-trophoresis; TG, 0.038 M Tris, 0.10 M glycine, pH 8.7; TX,Triton X-100.

2The nomenclature for platelet membrane glycoproteinsused in this report is adapted from Phillips and Poh Agin (4) aspreviously described (5).

717J. Clin. Invest. © The American Society for Clinical Investigation, Inc. * 0021-9738/81/03/0717/08 $1.00Volume 67 March 1981 717-724

precipitate given by samples from normal '25I-labeledplatelets demonstrated the presence of two 1251-labeledglycoprotein bands that migrate as GPlIb and GPIIIa.

A deletion of the platelet-specific alloantigen, PlAl,on platelets of five patients with thrombasthenia wasfirst reported by Kunicki and Aster (11). This observa-tion was subsequently confirmed by Muller et al. (12)and van Leeuwen et al. (13) who studied 12 differentthrombasthenic patients. Recently Kunicki and Aster(5) demonstrated that the PIAl determinant is eitherclosely associated with or is an integral part of the struc-ture of membrane GP lIla, thus establishing a link be-tween the glycoprotein abnormality characteristic ofthrombasthenic platelets and the abnormal expressionof the PlAl alloantigen.

We now report further studies on the platelets ofthree thrombasthenic patients and 20 members of theirimmediate families (kindred). With the use of CIE forthe determination of levels of platelet membrane pro-teins or glycoproteins (10), we have been able to dis-tinguish those kindred who are heterozygous for themembrane GP Ilb + GP Illa abnormality from thosewho are unaffected. These results combined with de-termination of the amount of PlAl on platelets of theseindividuals, by inhibition of 51Cr platelet lysis, providefurther information regarding the relationship betweenthe inheritance of GT and the expression of pIA' allo-antigen activity.

METHODSPatients. Each of the three related patients are members

of two large families, H. and W., which include descendantsof the Manouches gypsy tribe. This tribe had previously mi-grated throughout the Germanic countries and has recentlysettled, in part, within the Alsace region of France. These pa-tients (J.H., S.H., and F.W.) were previously shown by Levyet al. (14) to fulfill the diagnostic criteria of GT. A fourth pa-tient, previously diagnosed as having thrombasthenia (14),and who was a cousin of F.W., is deceased.

Preparation of platelets. The preparation of platelet-richplasma (PRP) from EDTA-anticoagulated whole blood, plate-let aggregation studies, clot retraction assays, and humanleukocyte antigen (HLA)-A, B, and Cantigen typing were per-formed in Strasbourg. On a given day, four numbered butotherwise unidentified samples of freshly prepared PRPfromany combination of thrombasthenic patients, GT kindred, ornormal individuals were shipped by train on ice to Paris whereCIE and inhibition of 5'Cr lysis assays were performed. Afifth sample of PRP from a normal individual was preparedsimultaneously in Paris and stored under similar conditions.The identity of the platelet donors from Strasbourg as well asthe pedigrees of families H. and W. were not divulged untilthe completion of all CIE and inhibition of 51Cr lysis as-says and the quantitative analysis of data from each individual.

Upon arrival in Paris, platelets were isolated from PRPandwashed free of residual plasma, erythrocytes, and leukocytesby differential centrifugation in 0.01 M Tris-HCl, 0.001 MEDTA, 0.145 MNaCl, pH 7.4, as previously described (5).

Inhibition of 51Cr release. The selection of antibodies re-active with the PlAl antigen, the choice of quinine- and quini-dine-dependent antiplatelet antibodies, the techniques for

51Cr-labeling of platelets, the 5"Cr-release assay, and the assayof antibody-reactive sites by inhibition of 5'Cr release havebeen previously described (5).

In a single experiment, washed platelets from four Stras-bourg donors to be tested for antibody-reactive sites and froma normal donor known to be homozygous for the PIA1 antigenwere resuspended in phosphate-buffered saline and analyzed.Several dilutions of each platelet suspension to be assayedwere prepared in phosphate-buffered saline to contain a finalprotein concentration, based upon the method of Markwellet al. (15), ranging from 100 gg to 2.0 mg/ml.

CIE. Triton X-100 (TX)-soluble platelet protein was elec-trophoresed against multispecific rabbit anti-human plateletantibody preparations essentially as described by Hagen et al.(10) with minor variations that are described in detail else-where.3

Antisera against whole human platelets were preparedin rabbits using the immunization scheme described by Bjer-rum and Bog-Hansen (16). Sera collected during 3-mo periodswere pooled and IgG was isolated by ammonium sulfateprecipitation followed by DEAE-cellulose chromatography,as described by Harboe and Ingold (17). Three pools of puri-fied IgG, obtained sequentially from the same sensitized rab-bits, were used in these experiments, and were obtained 6, 9,and 12 mo, respectively, after the primary immunization.

Washed platelets from Strasbourg donors and known normalindividuals were resuspended in 0.038 M Tris, 0.10 M gly-cine (TG) to a final platelet concentration of 4-6 x 109/mland solubilized by addition of TX to a final concentration of1% (vol/vol). Following centrifugation at 80,000 g for 1 h at4°C, the protein concentration of the TX-soluble supernateswas determined by the method of Markwell et al. (14).

15-t,l samples of TX-soluble protein containing 70-130 ,gof platelet protein were electrophoresed at 10 V/cm for 60min in first dimension gels consisting of 1%agarose and 0.5%(vol/vol) TX in TG. Second dimension electrophoresis wasperformed at 2 V/cm for 18 h into a biphasic gel system thatconsisted of (a) an intermediate gel containing 1% agaroseand 0.5% (vol/vol) TX in TG, and (b) an upper gel containingthe same suspension plus the IgG fraction isolated from rabbitantiplatelet antisera (750-800 gg/cm2). All electrophoreticoperations were performed at 15°C using a temperature-regu-lated circulating water bath (Haake, Inc., Saddle Brook, N. J.model NK 22). After electrophoresis, immunplates werewashed and stained with CBR(16). In a single experiment,platelet samples from four Strasbourg donors were alwaysanalyzed along with a sample from a known normal Parisdonor.

After electrophoresis, an enlarged image (four to five times)of each 5 x 7-cm dried CBR-stained immunplate was tracedonto drawing paper with the aid of a Leitz (E. Leitz, Inc., Rock-leigh, N. J.) photographic enlarger. The area beneath selectedprecipitates in the subsequent tracing was determined byplanimetry. In a previous study3 two precipitates were shownto contain GP lIb + IlIa and GP IIIb, respectively (shown inFig. 1). The calculated area beneath each of these precipitateson immunplates derived from increasing dilutions of an identi-cal platelet extract was directly proportional to the amount oftotal protein applied to the respective immunplate.

Additional methods. The relationships between the threethrombasthenic patients and 20 members of their immediate

3 Kunicki, T. J., A. T. Nurden, D. Pidard, and J. P. Caen.Different hereditary disorders of platelet function arecharacterized by different molecular and antigenic abnor-malities as shown by crossed immunoelectrophoresis. Sub-mitted for publication.

718 T. J. Kunicki, D. Pidard, J-P. Cazenave, A. T. Nurden, and J. P. Caen

FIGURE 1 Crossed immunoelectrophoresis of TX solubilizedproteins from (A) normal platelets, from (B) type I thrombas-thenic platelets (F.W.), and (C) from the platelets of a siblingof thrombasthenic patient F.W. 100 ug of platelet proteinssolubilized in 0.038 MTris, 0.1 Mglycine, pH 8.7 containing1% (vol/vol) TX were electrophoresed against a multispecificrabbit anti-human platelet antibody preparation (purified IgG;750 ,um/cm2). No antibodies were present in the intermediategel. Electrophoresis in the first dimension (left to right) was

performed at 10 V/cm for 1 h; in the second dimension (bot-tom to top), at 2 V/cm for 18 h. After electrophoresis, immun-plates were washed, dried, and stained with CBR. Theprecipitates containing fibrinogen (F) and membrane glyco-proteins Ib, IIb + IIIa, IIIb, and glycocalicin (Glyc) are indi-cated. Note the difference in area beneath the GP IIb + IIIaand the fibrinogen precipitates when comparing Figs. 1Aand 1C, and the absence of these precipitates in Fig. 1B(open arrows).

families, who all live in the area of Strasbourg, was confirmedby HLA-A, -B, and -C typing by the microlymphocytotoxicitymethod of Terasaki and McClelland (18).

Samples of 0.5 ml of citrated PRPwere used to study plate-let aggregation by ADPand collagen using a turbidometricmethod (19). Clot retraction was measured semiquantitativelyby weighing the serum expressed from non-anticoagulatedblood after 3 h of incubation (19).

RESULTSCIE analysis. A typical CIE profile of a platelet pro-

tein sample from a normal individual is shown in Fig.

1A. The platelet protein and glycoprotein antigen in 11major precipitates have been identified.3 Theseidentifications were based upon the results of competi-tive precipitation with monospecific antibodies, andSDS-PAGEanalysis of individual precipitates excisedfrom immunplates, and were aided by comparisons ofCIE profiles of normal platelet extracts and those givenby platelets from patients with GT, the Bernard-Souliersyndrome, the Gray Platelet syndrome, and von Wille-brand's disease. For the purposes of this report, onlythose precipitates containing fibrinogen, glycocalicin,and membrane GP Ib, Ilb + Illa, and IITb are indi-cated. No major variations were observed in the pro-files of platelet extracts from the 15 normal individuals.When soluble platelet protein from each of the threethrombasthenic subjects was analyzed (Fig. 1B), thecomplete absence of precipitates containing fibrinogenand GP IIb + IIIa was evident. This profile is charac-teristic of that previously observed with soluble plate-let protein from two type I thrombasthenic patients(10). Whensoluble protein samples from the platelets ofthe GT kindred were analyzed, 7 were shown to beidentical to that of normal samples (Fig. 1A), while 13exhibited a significant and similar reduction of the GPIlb + IIIa precipitate (Fig. 1C).

The levels of precipitates containing other mem-brane glycoproteins that are not affected by the throm-basthenic defect, such as GP Ib, glycocalicin, and GPITIb, appeared normal in samples from all thrombas-thenic patients and their kindred studied (Figs. 1Band 1C).

Quantitative analysis of GP IIb and IIIa by CIE.Taking advantage of the fact that the area beneath agiven precipitate is proportional to the quantity ofantigen present (in the presence of a constant amountof antibody), an attempt was made to determine therelative amounts of GP Ilb + IlIa in platelet samplesfrom GT kindred compared with normal plateletsamples.

To minimize the variability in the absolute area ofprecipitates on immunplates developed on differentdays, the area beneath the GP IIb + IIIa precipitatewas determined relative to the area on the same immun-plate beneath a second precipitate (GP IIIb) not affectedby the thrombasthenic defect. The ratio of the area ofthe GP IIb + IIIa precipitate to that of the GP IIIbprecipitate in a single normal platelet sample electro-phoresed on five different occasions was 2.24±0.17(mean±SD). For a second normal platelet sample, thevalue was calculated to be 2.23+0.21 (mean+SD,n = 5).

The area beneath the precipitates containing GPIlb+ III and GP ITIb, respectively, in platelet samplesfrom the 15 normal individuals and 20 GT kindredstudied in this report was calculated (Fig. 2). The re-sults were seen to be divided into two groups, desig-

Platelet Alloantigen and Type I Thrombasthenia

B lb GJyc

A0,-Oft f

C lb Glyc

filb . 4U+1|111

00plw f ^

719

normal subjects was assayed by determining their abil-ity to compete for PlAl_specific antibody and therebyinhibit the lysis of 5'Cr-labeled target platelets addedsubsequently (5). As shown in Fig. 3, platelets fromeight normal subjects and six GT kindred (group A)expressed an amount of PIAl antigen equivalent to thatcontained in platelets from normal individuals who arehomozygous for the Al allele. Platelets from each of thesix GTkindred in group A also contained a normal com-plement of GP IIb + IIIa. The final platelet proteinconcentration required to inhibit by 50%the maximumimmune lysis observed (ID50) was calculated to be22.5±1.5 (mean±SD) ,ug/ml. The amount of PlAlantigen expressed on platelets of individuals withinthis group is arbitrarily designated as 100%.

Platelets from two normal individuals and five GTkindred (group B) expressed an average of 54% of the

1 12 14 18 20

Glycoprotein I lIb (cm2)

FIGURE 2 Quantitation of the platelet GP lIb and GP IIIacontent. The area (square centimeters) beneath precipitatescontaining GP IIb + IIIa (ordinate) and GP IIIb (abscissa)for each platelet sample studied is shown. The area of theprecipitates in enlarged images of CBR-stained immunplatessuch as those shown in Figs. 1A and 1C was determined byplanimetry. TX-soluble protein from 20 kindred of the throm-basthenic patients (0; 0) and from 15 normal individuals (C)was analyzed, generating data that fell into two distinct groupsdesignated A and B. The ratio of the area beneath the GPIIb + IIIa precipitate to that beneath the GP IIlb precipitatewas calculated to be: for group A, 2.21±0.20 (mean±SD);for group B, 1.18±0.11.

nated A and B. Within a threefold range of absolute pre-

cipitate areas, a significant correlation exists betweenthe area of the precipitate containing GP IIb + Illaand that containing GPIlIb among samples in group A(r = 0.9610; P < 0.001) or group B (r = 0.9574; P< 0.001). In group A, the ratio of the area of the GPIIb + IIIa precipitate to that of the GPIlIb precipitatewas 2.21±0.20 (mean+SD, n = 22); in group B, thisvalue was 1.18 ±0.11 (n = 13). Group A included plate-let protein samples from all of the normal individualsstudied and samples from seven of the GT kindred.Group B consisted of samples from the remaining 13GT kindred. Based upon the mean ratios determinedfor each group, it can be calculated that platelets repre-

sented in group B contain, on the average, 53% of theamount of GPIIb + IIIa present in platelets of group A.These results provide quantitative evidence to stronglysuggest that 13 of the 20 GTkindred studied by CIE inthis report (see Fig. 1C and Fig. 2, group B) are hetero-zygous for the thrombasthenic abnormality.

Expression of PlA1 antigen". The PIA1 content ofplatelets from thrombasthenic patients, kindred, and

100*

c

0

I

20

D

Final Platelet Protein Concentraton (pg ml-1)

FIGURE 3 Inhibition of the lytic activity of anti-PlAl antibodyby nonlabeled platelets from normal subjects, thrombasthenicpatients, and kindred of thrombasthenic patients. 0.02 ml ofnonlabeled platelets in 0.013 M phosphate buffer, 0.145 MNaCl, pH 7.4, was incubated with 0.02 ml of anti-PIAP antibodyfor 2 h at 37°C. Complement and 51Cr-tagged, PlA-positivetarget platelets were then added, and the percent immunelysis was determined after an additional 2 h of incubation at37°C. At each final platelet protein concentration tested(abscissa), the percent inhibition of maximum immune lysisobserved in the absence of inhibitor platelets was determined(ordinate). Those kindred who were determined to beunaffected by the thrombasthenic defect (0) are distinguishedfrom those determined to be carriers of the defect (0) (seelegend to Fig. 2). Inhibition curves fell into four groups whichincluded platelets from: (A), eight controls (-) and sixthrombasthenia kindred (0); (B), two controls (U) and fivethrombasthenia kindred (0,O); (C), nine thrombastheniakindred alone (0) and (D), one PlAl-negative control (-) andthree thrombasthenic patients ([1). The delineated areasrepresent the mean±+SDfor each group. The ID50 (microgramsper milliliter) for groups A, B, and C was calculated to be22.5±+1.5 (mean+SD), 41.8 +3.2, and 81.1+7.8, respectively.

720 T. J. Kunicki, D. Pidard, J-P. Cazenave, A. T. Nurden, and J. P. Caen

.

.

AUU

4I

42'

E 34-

' 30

26-

a 22''5

14'

10'

.0

0

maximum level of PIA' antigen, an amount equivalentto that observed on platelets of normal individualsheterozygous for the Al allele. The ID50 for plateletswithin this group was calculated to be 41.8±3.2(mean±SD). Of the five kindred within this group, theplatelets of one individual were shown to contain a nor-mal complement of GPIIb + Illa, while those of the re-maining four individuals were shown to contain one-half the normal level of these GP.

Group C included platelets from nine GT kindred,all of which express one-half normal levels of GP IIb+ IIIa. Based upon an ID50 value of 81.1±7.8 (mean±SD) for this group, the platelets of these individualsexpress -28% of the amount of PIAl antigen found onplatelets of normal individuals homozygous for theAl allele.

Finally, platelets from one PlAl-negative normal indi-vidual and the three thrombasthenic subjects (group D)were incapable of inhibiting lysis induced by the PIAlantibody, even as final platelet protein concentrationsas high as 200 ,ug/ml. Because platelets from these threepatients contained no detectable GP Ilb + IIIa (seeFig. 1B), these individuals can be considered PIAnull (11).

The specificity of the thrombasthenic abnormalitywas again confirmed by the fact that platelets fromall of the individuals studied in this report (thrombas-thenics, kindred, and normal individuals) expressed anormal amount of the receptor for quinine- and quini-dine-dependent platelet-specific antibodies, a receptorpreviously shown to be normally expressed on theplatelets of five different thrombasthenic patients (11).

Inheritance of the PJA1 antigen and the thrombas-thenic abnormality. The results of the quantitation ofGPIIb + IIIa and of the PlAl antigen on platelets of themembers of the two thrombasthenic families, H. andW., are summarized in the family pedigrees shown inFig. 4.

In family H., subject II-2 is the thrombasthenic pa-tient J.H. Her mother (I-1) is, as expected, a carrier ofthe thrombasthenic abnormality. A brother (II-3) wasalso identified as a carrier, whereas the remaining sib-lings are unaffected. The father is deceased. The throm-basthenic patient, S.H. (II-1) is a first cousin of J.H.

In family W., two thrombasthenic patients (II-2 andII-9) were identified: subject II-2 is patient F.W.; sub-ject II-9 is deceased. Again, the mothers of both throm-basthenic patients (I-1 and I-4)- are carriers, as is theonly father studied (I-5). Among the remaining mem-bers of family W., eight individuals are carriers and fourare unaffected.

DISCUSSION

The pIA (Zw) alloantigen system was establishedthrough the combined studies of von Loghem et al.

Fem;ly H. I

Family W.

I i 4 3 (i a3 CZ

FIGURE 4 PlA and thrombasthenia genotypes of families H.and W. Romannumerals describe the generations within eachfamily. Males are represented by square symbols; females, bycircle symbols. Black symbols represent individuals homozy-gous for the thrombasthenia defect; half-black symbols, indi-viduals heterozygous for this defect. White symbols representindividuals homozygous for the PlA' antigen; half-white sym-

bols, individuals heterozygous for this antigen. The stippledsymbols represent the corresponding absence of expression ofthe P1 lu antigen. As an example of the interpretation of thesepedigrees, subject I-2 in family W. is both heterozygous for thethrombasthenia defect and heterozygous for the P1AM antigen.Platelets from I-2, therefore, express 25%of the amount of PlAlantigen expressed on platelets of subject II-8 in the same

family, who is homozygous for the P1 antigen and unaffectedby the thrombasthenia defect. Familial relationships were

confirmed by analyses of lymphocyte HLA-A, -B, and -Cantigens of each subject.

(20), Shulman et al. (21), and Van der Weerdt et al.(22), and shown to be represented by the alleliccounterparts, PIA1 and PI"2 (also known as Zwa, andZwb, respectively). PlAl, present in 98% of the generalpopulation, has been implicated in the pathogenesisof the syndrome known as post-transfusion purpura (21)and appears to be the alloantigen that most often pro-vokes isoimmune neonatal thrombocytopenia (23, 24).Recent studies by Van Leeuwen et al. (13) have demon-strated that anti-PlAl antibodies can inhibit the in vitroaggregation of homozygous PlAl_positive platelets as in-duced by the physiologic agents, adenosine diphos-phate, and collagen. Taken together, these observationssuggest that PlA is, clinically, an important platelet allo-antigen system, and tiat the PIAl antigenic determinantis associated with a platelet surface component of po-

tential significance in normal platelet function.PIAl antigenic activity was shown by Kunicki and

Aster (5) to be associated with a major platelet mem-

brane glycoprotein, IIIa, using three independent iso-lation procedures. The precise role of GPlIla in plate-let function remains to be elucidated, but the severe

dysfunction of platelets that lack both GP IIIa and GPIIb in GT suggests that these glycoproteins are criticalto platelet-platelet cohesion and, possibly, otheraspects of platelet function (7).

A better understanding of the genetic basis of thethrombasthenic defect, and of its relationship to the ex-

pression of the PlIA alloantigen, has awaited improvedmethods for the precise quantitation of platelet mem-

brane glycoproteins and antigens. This capability is re-

Platelet Alloantigen and Type I Thrombasthenia 7!21

quired for the detection of individuals who are hetero-zygous for either or both the thrombasthenia gene andthe Al allele. Using a complement fixation assay and anantibody, IgG L., present in the serum of a polytrans-fused thrombasthenic patient that was shown to recog-nize an unidentified 120,000-mol wt antigen presenton normal platelets, Degos et al. (25) were able to detectintermediate levels of the antigen on platelets of indi-viduals presumably heterozygous for the thrombas-thenic defect. No attempt was made to quantitate theseintermediate levels. Subsequent studies have shownthat this antigen is located on GP Ilb and/or GP IlIa(10). Recently, McEver et al. (26) demonstrated thatplatelets from five obligate and four presumed throm-basthenia heterozygotes bound roughly 62% of thatamount of a 1251-labeled monoclonal anti-GP IIb + GPIlIa antibody that bound to normal platelets.

In the present study, we have used CIE of TX solubleplatelet protein against a multispecific rabbit anti-human platelet antibody (Fig. 1) to quantitate the levelsof GP Ilb + GP IlIa in platelets from several normalindividuals, three thrombasthenic patients, and 20members of the immediate families (kindred) of thesepatients. Our results (Fig. 2) demonstrate that CIE,when used in the quantitative manner described in thisstudy, is both precise and reproducible (SE ±8%). Tri-ton extracts from the platelets of each of the threethrombasthenic patients contained no trace of GP IIb+ GP IIIa. Such results are similar to those previouslyshown for two other type I thrombasthenic patients(10). Triton extracts from the platelets of seven kindredcontained an amount of GP lIb + GP IlIa identical tothat observed in normal platelets. These individualswere apparently unaffected by the thrombasthenic ab-normality. Triton extracts from the platelets of the re-maining 13 kindred expressed levels of GP Ilb + GPIIIa that were calculated to be, on the average, 53% ofthose observed in extracts of normal platelets. Indi-viduals in this group, which included all parents of thepatients studied, are suggested to be carriers of thethrombasthenic abnormality.

It is probable that the glycoprotein abnormality ofthrombasthenia is a direct effect of the inheritance ofan altered gene or genes that normally control theexpression of GPIlb + GPIlla. The presence of 0, 50,or 100% of normal levels of GP Ilb + Illa on plateletsof thrombasthenic patients or their kindred arguesagainst a role for proteolysis in the etiology of throm-basthenia, as recently suggested by Shuman andKarpatkin (27). To our knowledge, the molecular ab-normality of GTis unique among such abnormalities in-volving other human cell systems, in that two ap-parently different membrane glycoproteins, lIb andIlla, are consistently reduced or absent. No evidencehas yet been reported that the decrease in one of theseglycoproteins is unaccompanied by an equivalent de-

crease in the other regardless of the absolute levels ofboth glycoproteins observed.

The expression of the PlA' alloantigen on plateletsfrom the three patients and 20 kindred was preciselydetermined (SE±6%) by inhibition of 5'Cr-labeledplatelet lysis (Fig. 3). As expected, platelets from thethree patients contained negligible amounts (<5%) ofPlA' alloantigen. Platelets from those kindred that werenormal, with respect to GP lIb + GP IlIa content, ex-pressed either 100 or 50% of the maximum level ofPlAl (i.e., the amount detectable on platelets of nor-mal individuals homozygous for the Al allele). Theseindividuals can be classified as homozygous andheterozygous, respectively, for the Al allele.

Platelets from those kindred previously determinedto be carriers of the thrombasthenic abnormality ex-pressed roughly 50 or 25% of the maximum level ofPlAl. The logical interpretation of these results is that(a) these individuals are, in fact, homozygous andheterozygous, respectively, for the Al allele, and(b) the relative amount of Al antigen expressed isdiminished by a factor of two as a result of the presenceof one-half the normal content of GPIIIa, the carrier ofthe Al determinant, on platelets of these individuals.It follows from these results that if the genes controllingthe expression of PlA were the same as those controllingexpression of the thrombasthenic defect, individualswho were double heterozygotes would express either50%of maximal amounts of PlAl or no PlAl whatsoever.That individuals whose platelets contained 25% ofmaximum levels were consistently found strongly sug-gests that the genes controlling these two traits are notthe same.

In early studies, Caen (3) differentiated type I throm-basthenia, in which clot retraction is absent and plate-lets do not contain detectable fibrinogen, from typeII thrombasthenia, in which clot retraction is nearlynormal and fibrinogen is detectable. Hagen et al. (10),using CIE, demonstrated that the platelets of a type IIthrombasthenic patient contained 13% of the normalamount of GP IIb + IIIa. The residual GPIlb and GPIlIa in the platelets of this patient were antigenicallyindistinguishable from the same glycoproteins on nor-mal platelets by crossed immunoelectrophoresisagainst rabbit antibody and by crossed affino immuno-electrophoresis against IgG L. (28). Using the samemethods reported in that study, we observed an amountof GPlIb + IlIa in the platelets of a different type IIthrombasthenic patient that was 15%of that detected innormal platelets.3 At the same time, the levels of PIAlantigen in the platelets of the same two thrombasthenicpatients were calculated to be 13 and 11%, respectively,of that detected on platelets from individuals homozy-gous of the Al allele.4

4Kunicki, T. J. Unpublished observations.

722 T. J. Kunicki, D. Pidard, J-P. Cazenave, A. T. Nurden, and J. P. Caen

At the present time, we have no explanation for thepresence of residual amounts of GP IIb and IIIa andPlAl antigen in platelets of type II patients in the etiologyof this variant of thrombasthenia. In light of these ob-servations, it would be of interest to compare moreprecisely the molecular structure and antigenicity of GPIlb and GP IIIa isolated individually from normalplatelets, to compare these with the same glycoproteinsisolated from type II thrombasthenic patients, and todetermine the inheritance of the membrane glyco-protein abnormality in the kindred of type II thrombas-thenic patients.

In the CIE analyses of the platelets from each of threetype I thrombasthenic patients studied in this report,the precipitate containing fibrinogen was not detected.A similar observation was made by Hagen et al. (10)who studied two different type I thrombasthenic pa-tients. In contrast, platelets from two type II thrombas-thenic patients (10)3 contained elevated but ap-parently subnormal amounts of fibrinogen. Due tothe flat shape of the fibrinogen precipitate, estimatesof the amount of fibrinogen in samples from type Ithrombasthenic patients (Fig. IC) could not be madefollowing CIE and will be the subject of a subsequentinvestigation.

The results reported here suggest that the gene con-trolling the inheritance of the allelic forms of the plAalloantigen system is not the same gene as that (those)involved in the inheritance of the thrombasthenic de-fect. Type I thrombasthenic individuals can correctlybe classified as PlA null since, regardless of their geno-type with respect to the PlA system, these alloantigenswhich are normally associated with GP IIIa (5) cannotbe expressed.

ACKNOWLEDGMENTSThe authors are indebted to Professor J. M. Levy and Dr. P.Lutz (Strasbourg, France) for permission to study their pa-tients; to Dr. R. H. Aster (Milwaukee, Wisc.) for his gift ofanti-PlAl and quinine- and quinidine-dependent antiplateletantibodies; to Professor S. Mayer and M. M. Tongio (Stras-bourg) for HLA-typing of the individuals studied in this re-port; to M. M. A. Sutter and J. Launay (Strasbourg), for theirtechnical assistance; and to MmeI. Delcroix for the typing ofthis manuscript.

This work was supported by grant number CRL78.5.128.1from the Institut National de la Sant6 et de la RechercheMedicale, France, and by a grant from the Thyssen Founda-tion. Dr. Kunicki is a recipient of a National Research ServiceAward (HL 05825) from the National Heart, Lung, and BloodInstitute, a fellowship from Institut National de la Sant6 etde la Recherche Medicale (INSERM) and an award from thePhillipe Foundation, Paris. J-P. Cazenave is Charg6 deRecherche a l'INSERM and is supported by grants 78.7.2602from DGRSTand CRL 79.4.261.5 from INSERM.

REFERENCES1. Caen, J. P., P. A. Castaldi, J. C. Leclerc, S. Inceman, M. J.

Larrieu, M. Probst, and J. Bernard. 1966. Congenital

bleeding disorders with long bleeding time and normalplatelet count. I. Glanzmann's thrombasthenia (Reportof 15 patients). Am. J. Med. 41: 4-26.

2. Hardisty, R. M. 1977. Disorders of platelet function. Br.Med. J. 33: 207-212.

3. Caen, J. P. 1972. Glanzmann's thrombasthenia. Clin.Haematol. 1: 383-392.

4. Phillips, D. R., and P. Poh Agin. 1977. Platelet plasmamembrane glycoproteins. Evidence for the presence ofnonequivalent disulfide bonds using nonreduced-reduced two-dimensional gel electrophoresis. J. Biol.Chem. 252: 2121-2126.

5. Kunicki, T. J., and R. H. Aster. 1979. Isolation andimmunologic characterization of the human platelet allo-antigen, PlAl. Mol. Immunol. 16: 353-360.

6. Nurden, A. T., and J. P. Caen. 1974. An abnormal plateletglycoprotein pattern in three cases of Glanzmann's throm-basthenia. Br. J. Haematol. 28: 253-260.

7. Nurden, A. T., and J. P. Caen. 1979. The different glyco-protein abnormalities in thrombasthenic and Bernard-Soulier platelets. Semin. Hematol. 16: 234-250.

8. Phillips, D. R., and P. Poh Agin. 1977. Platelet membranedefects in Glanzmann's thrombasthenia. J. Clin. Invest.60: 535-545.

9. Hagen, I., and N. 0. Solum. 1978. Further studies on theprotein composition and surface structure of normal plate-lets and platelets from patients with Glanzmann's throm-basthenia and Bernard-Soulier syndrome. Thromb. Res.13: 845-855.

10. Hagen, I., A. Nurden, 0. J. Bjerrum, N. 0. Solum, and J.Caen. 1980. Immunochemical evidence for protein ab-normalities in platelets from patients with Glanzmann'sthrombasthenia and Bernard-Soulier syndrome. J. Clin.Invest. 65: 722-731.

11. Kunicki, T. J., and R. H. Aster. 1978. Deletion of theplatelet-specific alloantigen PlAl from platelets in Glanz-mann's thrombasthenia. J. Clin. Invest. 61: 1225-1231.

12. Muller, J. Y., C. Patereau, and J. P. Soulier. 1978. Throm-basthenie de Glanzmann: antigene PlAl et anticorps anti-Glanzmann. Rev. Fr. Transfus. Immuno-hematol. 21:1069-1078.

13. Van Leeuwen, E. F., G. T. E. Zonneveld, L. E. von Riesz,C. S. P. Jenkins, J. A. van Mourik, and A. E. G. Kr. vondem Borne. 1979. Absence of the complete platelet-specific alloantigens Zw (PlAl) on platelets in Glanz-mann's thrombasthenia and the effect of anti-Zwa anti-body on platelet function. Thromb. Haemnostasis. 42: 422.(Abstr.)

14. Levy, J-M., G. Mayer, R. Sacrez, R. Ruff, J-J. Francfort,and L. Rodier. 1971. Thrombasth6nie de Glanzmann-Naegeli. Etude d'un groupe ethnique a forte endogamie.Ann. Pediatr. (Paris). 47: 129-137.

15. Markwell, M. A. K., S. M. Haas, L. L. Bieber, and N. E.Tolbert. 1978. A modification of the Lowry procedure tosimplify protein determination in membrane and lipo-protein samples. Anal. Biochem. 87: 206-210.

16. Bjerrum, 0. J., and T. C. Bog-Hansen. 1976. Immuno-chemical gel precipitation techniques for analysis ofmembrane proteins. In Biochemical analysis of Mem-branes. A. H. Maddy, editor. Chapman & Hall, Inc.,London. 379-426.

17. Harboe, N., and A. Ingild. 1973. Immunization, isolationof immunoglobulins, estimation of antibody titre. Scand.

J. Immunol. 2 (Suppl. 1): 161-169.18. Terasaki, P. I., and J. D. McClelland. 1964. Microdroplet

assay of human serum cytotoxins. Nature (Lond.). 204:998-1000.

19. Cazenave, J-P., M. A. Guccione, M. A. Packham, and J. F.

Platelet Alloantigen and Type I Thrombasthenia 723

Mustard. 1977. Effects of cephalothin and penicillin Gonplatelet function in vitro. Br. J. Haematol. 35: 135-152.

20. Van Loghem, J. J., Jr., H. Dorfmeijer, and M. van derHart. 1959. Serological and genetical studies on a plateletantigen (Zw). Vox Sang. 4: 161-169.

21. Shulman, N. R., R. H. Aster, A. Leitner, and M. C. Hiller.1961. Immunoreactions involving platelets. V. Post-trans-fusion purpura due to a complement-fixing antibodyagainst a genetically controlled platelet antigen: a pro-posed mechanism for thrombocytopenia and its rele-vance in "autoimmunity."J. Clin. Invest. 40: 1597-1620.

22. Van der Weerdt, Ch. M., L. E. Veenhovern von-Reiz, L. E.,Nijenhuis, and J. J. Van Loghem. 1963. The Zw bloodgroup system in platelets. Vox Sang. 8: 513-530.

23. Aster, R. H. 1977. Thrombocytopenia due to enhancedplatelet destruction. In Hematology. W. J. Williams, E.Beutler, A. J. Erslev, and R. W. Rundles, editors.McGraw-Hill, Inc., NewYork 1326-1360.

24. Shulman, N. R., V. J. Marder, M. D. Hiller, and E. M.Collier. 1964. Platelet and leukocyte isoantigens and their

antibodies: serologic physiologic and clinical studies.Prog. Hematol. 4: 222-304.

25. Degos, L., A. Dautigny, J. C. Brouet, M. Colombani, N.Ardaillou, J. P. Caen, and J. Colombani. 1975. A molecu-lar defect in thrombasthenic platelets. J. Clin. Invest. 56:236-240.

26. McEver, R. P., N. L. Baenziger, and P. W. Majerus. 1980.Isolation and quantitation of the platelet protein lacking inthrombasthenia using a monoclonal hybridoma antibody.Clin. Res. 28: 495A. (Abstr.)

27. Shulman, S., and S. Karpatkin. 1980. Crossed immuno-electrophoresis of human platelet membranes. Dimin-ished major antigen in Glanzmann's thrombasthenia andBernard-Soulier syndrome. J. Biol. Chem. 255: 4320-4327.

28. Hagen, I., A. T. Nurden, 0. J. Bjerrum, N. 0. Solum, andJ. P. Caen. 1979. Immunochemical difference betweentype I and type II thrombasthenic platelets. In Collo-quium Protides of the Biological Fluids. XXVII Col-loquium May 1979. H. Peeters, editor. Pergamon Press,Oxford, England. 875-878.

724 T. J. Kunicki, D. Pidard, J-P. Cazenave, A. T. Nurden, and J. P. Caen