ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 20, No. 6Copyright © 1990, Institute for Clinical Science, Inc.

Anti-Thymocyte Globulin Induced Thrombocytopenia

JOHN LAZARCHICK, M .D.* RENE RUSSELL, MLT.

and BRENDA HORN, MT.

D epartment o f Pathology/Laboratory Medicine Medical University o f South Carolina

Charleston, SC 29425

ABSTRACT

Anti-thymocyte globulin is strongly reactive with platelets as m easured by flow cytometric analysis. This reactivity appears to be independent of human leukocyte antigen (HLA) phenotype, is dose-dependent with satu­ration of platelet binding sites readily achieved with concentrations of horse anti-thymocyte globulin (ATG) at 0.25 mg per ml, and cannot be blocked with human anti-PLA1 antibody nor with heat-aggregated human IgG. Results of immunoblot studies after electrophoresis of platelet m em ­brane proteins are consistent with the proposal that horse ATG contains antibodies specific for platelet glycoproteins V and Ha.

Introduction

Anti-thymocyte globulin has proven to be effective in treating aplastic anemia and related disorders either alone or as an adjunct to achieve im m unosuppres­sion prior to bone marrow transplanta­tion.10 One of the adverse side effects of th is th e ra p e u tic ag en t has b e e n the developm ent of thrombocytopenia. The involvement of platelet glycoproteins in th e d e v e lo p m e n t o f d ru g - in d u c e d im m une th rom bocy topen ia has been well docum ented for certain m edica­tions.6 This study sought to determ ine if the throm bocytopenia seen with an ti­

* Address reprint requests to: John Lazarchick, M .D ., Department of Pathology/Lab Medicine, 171 Ashley Avenue, Charleston, SC 29425.

thymocyte globulin was on an immune basis and, if so, to determ ine if platelet specific antigens were involved.

M aterials and M ethods

F l o w C y t o m e t r ic A n a l y s is

Horse anti-human platelet antibodies were detected using a modification of the flow cytometric analysis procedure pre­viously described.5 In brief, 0.01 ml of target platelets was incubated with 0.1 ml of horse an ti-thym ocy te g lobulin (ATGAM)* diluted with Dulbecco’s buff­e re d saline co n ta in in g 0 .05 p e rc e n t sodium ethylenediam ine tetraacetic acid (EDTA), 0.02 percent sodium azide and

* Upjohn Company.

3730091-7370/90/1100-0373 $00.90 © Institute for Clinical Science, Inc.

3 7 4 LAZARCHICK, RUSSELL, AND HORN

one percent bovine serum albumin, pH 7.5 or with normal non-im m une horse serum for 60 m inutes at room tem pera­tu re . The m ixtures w ere then ex ten ­sively washed, centrifuged and the plate­le t p e lle ts in cu b a ted w ith 0.1 ml of fluorescein-conjugated goat (Fab')j anti­horse IgG anti serum for 15 m inutes at room tem perature. After repeat centrifu­gation and washing, the platelet pellets w ere resuspended in phosphate-buff­ered saline to a final volume of two ml. The am ount of fluo resce in -con juga te found in the p late le t suspensions was then m easured using flow cytom etric analysis. For this assay, region A on the spectrum III was defined as consisting of all 255 fluorescent intensity channels. The intensity of fluorescence, expressed as the mean value of region A(MCHA), is proportional to the amount of horse IgG bound to the platelet m em brane. The m ean channel A readings for the test serum plus platelet mixture and the con­trol serum plus platelet m ixture were determ ined. A ratio of relative fluores­cence was then calculated. A ratio of > 1 .3 of test serum/control serum is com­pa tib le w ith the p resen ce of p la te le t d irec ted an tibody in the tes t serum . W ashed H L A -ty p e d h um an p o o led platelets were used as the target cells for these experim ents. To assess for HLA specificity , sing le donor H L A -typed washed platelet preparations were used where indicated.

Experim ents were also performed to determ ine if horse ATG binding to plate­let glycoproteins could be inhibited by preincubation of the target platelets with specific anti-human platelet antibodies. The platelet antibodies were obtained using sera from a patient with idiopathic thrombocytopenic purpura (ITP), and a patient with anti-PLA1 antibody. In addi­tion, aggregated hum an IgG was p re ­pared by heating a 10 mg per ml of solu­tion o f IgG at 63°C for 20 m inu tes . Dilution of these IgG sources were incu­

bated for 30 m inutes at room tem pera­ture with the HLA-typed platelet prepa­rations as d escrib ed previously . The m ix tu res w ere th en w ash ed , c e n tr i ­fuged, and the platelet pellets reacted w ith dilu tions of e ith e r norm al horse serum or anti-thym ocyte globulin and the flow cytometric analysis perform ed as previously described.

P l a t e l e t P r e p a r a t io n / E l e c t r o p h o r e s is / I m m u n o b l o t t in g

P la te le t rich plasm a was p rep a re d from blood collected in EDTA tubes by centrifugation at 150 g for 20 minutes. The supernatant was then centrifuged at 1300 g for 10 minutes. The platelet bu t­ton was then resuspended in R inger’s- EDTA containing 10 mM benzamidine HCL and 100 p-g per ml of soybean tryp­sin inhibitor. After several washings, suf­ficient volume was added to achieve a platelet concentration of 1 X 109 ml. To this final suspension was then added 1:40 volume of 40 percent Triton X-100. The platelet suspension was then incubated at 37°C for one hour and refrigerated at 4°C overnight.

Sodium dodecyl sulfate po lyacry l­am ide gel electrophoresis was carried out as described by Laemmli using a five to 20 percent gradient gel as the separat­ing gel and five percent polyacrylamide as the stacking gel.4 Twenty-five |xl of p late le t suspension (1 X 109 platelets) were mixed with 25 |jl1 of a solution con­taining 0.002 percent bromphenol blue, four percent sodium dodecyl sulfate and0.125 M Tris, and 20 percent glycerol, PH 6.0, and incubated for five m inutes at 100°C prior to loading. Electrophore­sis was carried out for 16 hours at 20 volts. The voltage was increased to 80V for two additional hours.

P roteins w ere transferred from the po lyacry lam ide gel to n itro ce llu lo se m em brane was described by Towbin et al.8 The portion of the paper carrying the

PLATELET SPECIFIC ANTIBODIES 375

molecular weight standards was stained with 0.125 percent Coomasie blue, and the rem ainder was cut into strips corre­sponding to the e lectrophoresis lanes and placed in blocking buffer containing five percent non-fat dry milk in 0.01 M Tris, 0.17 M sodium chloride, pH 7.5.

After an overnight incubation at room tem perature, the strips were extensively washed and then incubated with appro­priate antibody-containing preparations. These included mouse monoclonal anti­human GP I lia antibody diluted 1:500 in Tris buffered saline. The normal horse serum and horse ATG were both diluted 1:500 in the same buffer. The strips were th en in cuba ted at room tem p e ra tu re overnight on a shaker bath. After wash­ing, a solution of peroxidase-conjugated goat antibody to e ither mouse IgG or horse IgG was added to the appropriate strips and the strips incubated for one hour at room tem perature . The strips w ere then developed using 4-choral- 1-naphthol as an indicator reagent (fig­ure 1).

Results

T he in itia l experim en ts w ere p e r­formed to determ ine if horse IgG in the an ti-thym ocyte globulin p repara tion

(ATGAM) would bind to human platelets in a dose-responsive manner. For those experim ents, both horse ATG and the normal horse serum were diluted to give final concentrations of IgG of approxi­mately 0.005, 0.01, and 0.05 mg per ml prior to incubation with the HLA-typed platelet source. The intensity of fluores­cence of the mean channel of A for the no rm al h o rse serum was e ssen tia lly u n ch an g ed b e in g 7.9, 7 .8 , and 8 .1 , respectively, at these concentrations. In contrast, the MCHA of the horse ATG p re p a ra t io n sh o w ed a p ro g re s s iv e increase from 10.3, 12.2, to 14.2 respec­tively. This dose-response pattern was present irrespective of the presence or absence of human serum from controls or from the patient who had developed marked throm bocytopenia after receiv­ing h o rse a n ti- th y m o c y te g lo b u lin (ATG).

Pre-incubation of the platelets with e ither normal hum an serum or normal horse serum did not inhibit subsequent binding with horse serum nor the extent of binding, i.e ., similar M CHA values w ere ob ta ined . The concen tra tion of immunoglobulin from these sources was increased to 0.25 and 1.0 mg per ml. The amount of non-specific binding using the control horse serum minimally increased

Count Count1250 . 1250

A1000 1000 -

750 750

500 500

250

V250

50 100 150 200 250Gr_FI

Ch D D is t Region A

Tot Count 7033.0Mean Ch 7.9

j \

B

50 100 150 200 250Gr_Fl

Ch D H is t Region A

Tot Count 4036.0Mean Ch 22.8

F i g u r e 1. A compari­son of the green fluores­cent histograms generated on assay mixtures of HLA- ty p ed p o o led p la te le ts incubated with either nor­mal horse serum (A) or the ATGAM preparation (B) is shown. The IgG concen­trations in each prepara­tion was approxim ately 0.25 mg per ml. The mean channel A ratio for the A T G A M /norm al h o rse serum was 2.9.

376 LAZARCHICK, RUSSELL, AND HORN

to give a M CHA value of 12.2. The MCHA value for the horse ATG dilu­tions increased also but plateaued with a value of 36. F u r th e r increases in the concentration of the horse ATG prepa­ra tio n d id n o t show any a d d itio n a l increased binding suggesting that the IgG binding sites on the platelets were fluorescence histogram obtained w ith washed HLA-typed platelets after incu­bation w ith norm al horse serum and horse ATG.

To d e te rm in e w h e th e r or no t th is binding of IgG in the horse ATG prepa­ration was specific for HLA antigens, sim ilar studies w ere perform ed using single donor HLA-typed platelets from three individuals of variable haplotypes at the A and B loci. These data are shown in table I. N on-specific b inding with normal horse serum was relatively con­stant among the th ree platelet sources and gave sim ilar values to the HLA- typed p la te le t pool. The IgG binding with the horse ATG preparation, in con­tras t, was g rea tly increased , and the MCHA ratios w ere similar among the th re e sing le d o n o r p la te le t sources. These results are strongly suggestive that the reactivity noted is independent of the HLA loci b u t since single loci were shared among this patient group,

T A B L E I

Immunoreactivity of Anti-thymocyte Globulin With Variable Platelet HLA* Antigens

MCHAf MCHA R a t io

P l a t e l e t Source NHSt ATGAM§ ATGAM (NHS)

H L A - t y p e d pool 13.1 31.1 2.4D o n o r 1(A2/3 : B60,39) 12.0 38.5 3.2

D o n o r 2(A2,30 : B35,Y) 12.4 37.3 3.0

D o n o r 3(All/ 24: B60,62) 11.3 35.8 3.2

* Human l e u k o c y t e - a s s o c i a t e d antigens. ■{■Mean v a l u e o f r e g i o n A.$N o n n a l h o r s e serum.§A n t i - t h y m o c y t e globulin.

i.e ., A2 in donors 1 and 2, and B60 in donors 2 and 3, it d id not absolutely exclude this possibility.

N ext, it was sough t to d e te rm in e w hether or not horse ATG binding could be inhibited by pre-incubating the HLA- p latelet pool source with specific an ti­p la te le t an tibod ies ob tained from an individual with idiopathic thrombocyto­penic pu rpura and from a patient who had anti-Plal antibody in her serum. In addition, heat aggregated hum an IgG was used to block platelet Fc receptors on the prem ise that the IgG in the horse serum preparation was binding to these receptors. The results of these studies are shown in table II. Both sources of hum an a n ti-p la te le t an tibody had no effect on the extent of horse ATG bind­ing as ju d g ed by the M CHA values. Preincubation of the platelets with non­aggregated IgG had minimal inhibitory effect on horse ATG binding. The p re­incubation with heat aggregated human IgG did show an increased MCHA value when the platelets were subsequently in cu b a ted w ith norm al horse serum , compatible with IgG aggregated binding to the Fc receptor; however, the MCHA values in the second incubation with horse ATG were similar to the nonaggre­gated IgG result.

To determ ine the specificity of horse ATG binding, sodium dodecyl sulfate (SDS) gel e lec tropho resis of p la te le t preparations and immunoblots with the norm al horse serum and horse ATG w ere p e rfo rm e d . T h ese re s u lts a re shown in figure 2. Normal horse serum showed no reactivity with any specific p latelet glycoprotein. In contrast, two bands are noted w ith the horse ATG preparation. The major band is with a platelet glycoprotein of approximately83,000 kd and the lighter staining band is with a glycoprotein of approximately 125—130 kd. The 83,000 kd band is the same molecular weight as glycoprotein V: th e second band has a m olecular weight similar to glycoprotein Ha.

PLATELET SPECIFIC ANTIBODIES 377

Effect of Platelet Pre-incubation With Human Immunoglobulin Sources

on Anti-thymocyte Globulin

TABLE I I

MCHA* MCHA R at io

P l a t e l e t Source NHS'f ATGAM% ATGAM (NHS)

HLA§-pool + Normal serum 7.1 30.1 4.2

HLA-pool + ITP serum 6.9 31.6 4.6

HLA-pool + Anti-PLAl serum 7.7 30.5 4.0

HLA-pool + Non-aggregated IgG 8.7 28.3 3.3

HLA-pool + Heat-aggregated IgG 12.9 28.4 2.2

*Mean value of region A. fNormal horse serum.¿Anti-thymocyte globulin.§Human leukocyte-associated antigens.

Discussion

Anti-thymocyte globulin has a num ber of adverse effects associated with its use including throm bocytopenia. That this side effect may be on an immune basis was first suggested by Rosenberg and co­workers who proposed several possible mechanisms to account for the throm bo­cytopenia including platelet activation as a result of im m une complex formation betw een the anti-thym ocyte globulin and lym phocy tes. T hey n o ted w hen horse ATG was added to platelet rich p lasm a in v i t r o , s ign ifican t p la te le t aggregation and release took place. They suggested that immune complex forma­tion was with a platelet m em brane com­ponen t ra the r than with ano ther cell type. S ubsequen tly G reco e t al in a study of horse ATG in te rac tion w ith lym phocytes, granulocytes, platelets, and erythroid bone marrow were able to dem onstrate that horse ATG was equally reactive with distinct antigens on each type of cell but that the reaction with platelets was only of low intensity, i.e., mean intensity of fluorescence was mini­m al.2 This was true for six different lots of horse ATG tested. W echter et al per­

formed a variety of immune assays on 33 different lots of horse ATG including a complement fixation assay for anti-plate­let antibodies. All of these lots showed anti-platelet reactivity.9

Our studies confirm and extend these reports. Several features of this immune reaction are notew orthy: (1) previous sensitization is not required for the bind­ing of horse ATG to platelet surface pro­teins since this occurred in the presence or absence of the donor’s plasma; (2) the interaction is dose-dependent and can be saturated im plying there are specific receptor proteins on the platelet surface with which the immunoglobulin is bind­ing; (3) platelet glycoproteins V and Ha appear to be the reactive antigens, at least with the lot of horse ATG employed in our studies. No hemostatic function is known for glycoprotein Ha. Glycopro­tein V, however, may be the thrombin receptor on the surface of intact plate­lets. 1 O ther studies question the biologi­cal significance of this proposal since chymotrypsin digestion of glycoprotein

A B C

F i g u r e 2. Transblot analysis following sodium dodecyl sulfate polyacrylamide gel electrophoresis of HLA-typed pooled platelet preparation. The electro- phoretically separated platelet proteins were reacted with: (A) monoclonal mouse anti-human glycoprotein Ilia, (B) ATGAM, (C) normal horse serum.

96 kOa - 83 kDa —

378 LAZARCHICK, RUSSELL, AND HORN

V or blocking of this protein by specific an tibody does no t im pair th ro m b in induced p la te le t aggregation; (4) the reaction is not inhibited by platelet anti­body to o ther glycoproteins, i.e ., I lia nor by blocking p late le t Fc receptors with aggregated human IgG.

These studies clearly dem onstrate the presence of antiplatelet specific antibody in horse ATG preparation as was shown by the results with flow cytometric anal­ysis. This assay, however, does not dis­tinguish betw een reactivity with HLA antigens on the p latelet surface versus reactivity with platelet specific antigens. The immunoblot studies docum ent the latter but do not exclude the additional presence of anti-HLA reactivity against the platelet since HLA antigens can not be detected on this type of immunoblot analysis.

References

1. G e o r g e , J. N .: The role of membrane proteins in platelet function. Trans. Med. Rev. 1 :4 3-4 6 ,1987.

2. G r e c o , B ., B ielo ry , L ., Steph a n y , D ., H s u ,S. M ., G a sc o n , P., N ie n h u is , A., an d Yo u n g , N .: A n ti-th y m o cy te g lobu lin reac ts w ith m any

normal human cell types. Blood 62 :104 7 -1 0 5 4 , 1983.

3. K u n ic k i , T. J . , F u i h a t a , K ., B o l d , B ., and N u g e n t , D. J. : The immunogenicity of platelet m em brane glycoproteins. Trans. M ed. Rev. 2 :2 1 -3 3 , 1987.

4. L a e m m li, U. K. : Cleavage of structural proteins during the assem bly o f the head of bacterio­phage T4. Nature 2 2 7 :6 8 1 0 -6 8 1 5 , 1970.

5. L a z a r c h ic k , J ., J o n e s , T. J., R u s s e l l , R. J ., and H a l l , S. A.: Platelet antibody detection using frozen pooled human lymphocytes anti­gen-type platelets as target cells. Diag. Clin. Immunol. 5 :3 7 7 -3 8 0 , 1988.

6. P f e u l l e r , S. L ., B i l s to n , R. A., L o g a n , D ., G ib so n , J. M., and F irk in , B. G .: Heterogene­ity of drug-dependent antigens and their anti­bodies in quinine and quinidine-induced throm­bocytopenia: Involvement of glycoproteins lb, lib , Ilia, and IX. Blood 72:1 1 5 5 -1 1 6 2 , 1988.

7. R o s e n b e rg , J. C., E k is , M. L ., Lysz, K ., and M o r r e l l , R. : Effect o f anti-thrombocytoglobu- lin and other immune reactions on human plate­lets. Surgery 77 :5 2 0 -5 2 9 , 1975.

8. T o w b in , H ., S t a e h e l i n , T ., and G o r d o n , J.: Electrophoresis transfer of proteins from poly­acrylamide gels to nitrocelluous sheets. Proce­dures and some applications. Proc. Natl. Acad. Sci. USA 7 6 :4 3 5 0 -4 3 5 4 , 1979.

9. W e c h t e r , W. J., N e ls o n , J., P e r p e r , R. J., P a r c e l l s , A. J., R ie b e , K. W ., E v a n s , J. S ., S a to h , P. S ., and Ko, H .: Manufacture of anti­thymocyte globulin (ATGAM) for clinical trials. Transplantation 2 8 :3 0 3 -3 0 7 , 1979.

10. Y o u n g , N ., G r i f f i t h , P., B r i t t a in , E ., E l f e n - b e in , G . , G a r d n e r , F ., et al: A multi-center trial of anti-thymocyte globulin in aplastic anemia in related diseases. Blood 7 2 :1861-1869 , 1988.