CD32)Phenotype and Level of Expression by Yoshiaki Tomiyama, Thomas J

Home , Fc receptor

Response of Human Platelets to Activating Monoclonal Antibodies: Importance of FcyRII (CD32)Phenotype and Level of Expression By Yoshiaki Tomiyama, Thomas J. Kunicki, Theodore F. Zipf, Sheila 6. Ford, and Richard H. Aster

Certain monoclonal antibodies (MoAbs) specific for platelet basis of MoAb binding studies to be heterozygous for the membrane glycoproteins are known to be capable of activat- two alleles of FcyRIIA. In contrast to their varying sensitivity ing platelets, and it is generally thought that platelets from to lgGl MoAbs, members of the platelet panel responded normal subjects are equally susceptible to stimulation by equally well to 50H.19, an lgGzaMoAb specific for CD9, and such MoAbs. We found that platelets from 20 normal donors these responses could not be blocked by MoAb IV.3 in the varied significantly in their sensitivity to three lgGl murine presence of plasma. This appears to be because of dual MoAbs specific for membrane glycoproteins CD9, GPIV actions of 50H.19 on platelets: one FcR-dependent and the (CD36). and the GPIIb/ llla complex (CD41). respectively. The other complement-dependent. Our findings confirm previous response of platelets to these MoAbs was blocked by prior reports that certain lgG1 MoAbs activate platelets through addition of MoAb IV.3 specific for the FcyRll receptor, indicat- binding of their Fc domains to FcyRll receptors and demon- ing that activation was Fc receptor mediated. Platelets that strate that this response is influenced both by Fhpll pheno- responded poorly to these MoAbs failed to bind the MoAb type and (in the case of the anti-CD41 MoAb) by the number

41H.16, specific for the ”responder“ form of FcyRII, but of FcyRll receptors expressed. The failure of nonresponding Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 platelets that responded well reacted with this MoAb. The platelets to bind detectable amounts of MoAb 41H.16, which average number of FcyRll receptors on platelets from is thought to recognize all FcyRll receptors except for one “responders” and “non-responders“ was approximately the allele of the FcyRllA gene, is consistent with the possibility same. However, the number of FcyRll receptors expressed that FcyRllA gene products, but not FcyRllB or FcyRllC gene influenced sensitivity of a subgroup of ”responder” platelets products, are expressed on platelets. to the anti-CD41 MoAb. These platelets were judged on the o 1992 by The American Society of Hematology.

ERTAIN MURINE monoclonal antibodies (MoAbs) extracellular and transmembrane domains of the mature C induce release and aggregation on binding to human proteins determined by these three genes differ only by a platelets.l-ll Some MoAbs specific for glycoproteins IIb or small number of amino acids, but the FcyRIIB gene IIIa act directly on the GPIIb/IIIa complex to induce products differ extensively from those of FcyRIIA and conformational changes leading to expression of the fibrin- FcyRIIC in their cytoplasmic domains. Presumably, this ogen receptor.3psJ2 More often, MoAbs specific for these variety of receptors allows for the induction of different or other membrane glycoproteins trigger activation and signals on binding of ligands to FcyRII. Which FcyRII gene release through conventional signal transduction path- products are expressed on platelets has not yet been For reasons not yet clear, MoAbs specific for defined. CD9 or GPIV (CD36) are the most consistent platelet The FcyRIIA gene appears to be diallelic. Monocytes activat~rs.~J~In general, MoAb-induced transmembrane homozygous for one of these alleles are unable to induce signalling requires an intact IgG Fc domain and can be proliferation of T cells sensitized with murine IgGl MoAbs blocked by MoAb IV.3 specific for the FcyRII receptor specific for CD3,22-24apparently because of their failure to (CD32), implying that binding of monoclonal Fc to that interact with murine IgGl Fc sufficiently well to induce receptor is important in the process of a~tivation!,~-~~It has clustering of CD3. Monocytes with this phenotype are been observed that platelets from normal subjects differ in designated “low responders” (LR). Monocytes bearing the their susceptibility to activation by murine MoA~s,~but the second FcyRIIA gene product produce good stimulation of basis for this has not been established. sensitized T cells and are designated “high responders” Three families of receptors with specificity for the Fc (HR).= Monocytes and neutrophils from LR and HR region of human IgG have been described in humans.lsJ6 FcyRI receptors bind monomeric IgG Fc with high affinity. FcyRII and FcyRIII receptors bind monomeric IgG Fc From The Blood Research Institute, The Blood Center of Southeast- weakly, but bind multimeric IgG Fc and IgG-containing em Wisconsin, Inc, Milwaukee; the Departments of Anatomy, Cell Biology, Medicine, Microbiology, and Pathology, the Medical College immune complexes tightly. FcyRII molecules have the of Wisconsin; and the Divisions of Pediatrics and Laboratory Medi- widest tissue distribution of the three FcyR families, being cine, The University of Tms, M.D. Anderson Cancer Center, Hous- expressed on all FcyR-bearing blood cells except NK ton. cells.lS Only FcyRII receptors have been demonstrated Submitted December 18, 1991; accepted June 30, 1992. unequivocally on platelets. Supported by Grants HL-13629 and POI-HL-44612 from the FcyRII receptors in humans are the products of at least National Heart, Lung, and Blood Institute and by The Kelcie Margaret three different genes, designated A, B, and C, each coding Kana Fund. for mature transmembrane proteins of about 40 Kd molec- Address reprint requests to Richard H. Aster, MD, President, The ular mass.1s-21At least six different mRNA transcripts and Blood Center of Southeastem Wuconsin, 1701 W Wisconsin Ave, four mature proteins are derived from these three genes. Milwaukee, WI 53233. The publication costs of this article were defrayed in part by page The FcyRIIAgene produces two transcripts as the result of charge payment. This article must therefore be hereby marked alternative polyadenylation, the FcyRIIB gene produces “advertisement” in accordance with 18 U.S.C.section 1734 solely to three or four transcripts by alternative splicing of exons indicate this fact. coding for cytoplasmic and signal sequences, and the 0 1992 by The American Society of Hematology. FcyRIIC gene appears to produce a single transcript. The 0006-4971 /92/8009-0021$3.00~0

Blood, Vol80, No 9 (November l),1992: pp 2261-2268 2261 2262 TOMIYAMA ET AL individuals can also be differentiated by their ability to form Platelets were separated from freshly collected blood anticoagu- rosettes and mediate antibody-dependent cellular cytotoxic- lated with EDTA and were washed three times in platelet ity (ADCC) against target cells sensitized with murine suspension buffer (PSB) (0.01 mol/L phosphate-buffered isotonic IgG126and by the isoelectric focusing patterns of their saline, 1% bovine serum albumin, 4.0 mmol/L EDTA, pH 6.5). FcyRII receptor^.^' IEF patterns of monocytes and plate- Final preparations contained fewer than one white blood cell per lets from the same individuals are concordant.28 10,000 platelets. Platelets, 1 x lo*, were suspended in 300 pL of buffer containing 2 bg of lz5I-Fab fragments. Preliminary studies Two well-studied MoAbs, IV.3 and 41H.16, appear to demonstrated that this quantity of Fab was in excess of the amount bind to the products of all three FcyRII genes.15 41H.16 needed to achieve saturation binding. After incubation at room binds strongly to the HR allele of FcyRIIA, but weakly or temperature for 1 hour, 50-pL aliquots were layered in triplicate not at all to the product of the LR allele.25Strong binding of on 200 FL of 20% Percoll in 400 pL polypropylene microcentrifuge 41H.16 to the HR gene product was recently shown to be tubes. The tubes were centrifuged at 10,000g for 4 minutes to allow determined by an arginine, rather than a histidine, residue pelleting of platelets and platelet-bound Fab. The tubes were then at position 131 in the second extracellular d~main.~~-~lclamped, the tips cut off, and the pelleted radioactivity was Monocytes from about 25% of the general population carry measured in a gamma scintillation counter. Molecules of Fab the form of FcyRIIA that fails to bind 41H.16, ie, has a bound were calculated from the formula: Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 histidine residue at position 131. (cpm bound) x x A In a number of clinical disorders, platelet destruction is SA x No. of Platelets x MW thought to be mediated by IgG Fc-dependent mechanism~.~*~~~A fuller understanding of platelet Fc receptors where A = Avogadro's number, SA = specific activity of Fab (cpm could provide clues to the pathogenesis of some of these per microgram), and MW = 50,000. conditions. Therefore, we characterized the FcyRII recep- Determination of IK3 and 41H.16 binding by flow cytometry. Platelet-rich plasma was prepared from freshly collected, citrated tors of platelets using MoAbs IV.3 and 41H.16 and studied whole blood using excess acid-citrate-dextrose (ACD) anticoagu- the relationship between FcyRII phenotype and level of lant to achieve a pH of about 6.4. The platelets were pelleted at expression and susceptibility of platelets to MoAb-induced 3,OOOg and washed twice in Ringer's citrate dextrose (RCD) buffer activation. containing 2 mmol/L EDTA and 50 ng/mL prostaglandin E1 (Sigma Chemical Co, St Louis, MO), pH 6.5. Platelets, 2 x lo7, MATERIALS AND METHODS were resuspended in 100 pL of RCD buffer containing 1 pg/mL of Murine MoAbs. Characteristics of the murine MoAbs used are MoAb IV.3 or 41H.16 and incubated for 60 minutes at room summarized in Table 1. ALB-6 (a-CD9) was obtained from temperature. These quantities were more than 10-fold the amounts AMAC, Inc (Westbrook, ME); 8A6 (a GPIV, subclone D9) from required to achieve saturation binding. After two washes in RCD Dr John Barnwell (New York University); 50H.19 (a-CD9) from buffer, the platelets were incubated with a 1:40 dilution of Dr A.R.E. Shaw (University of Alberta, Canada); and IV.3 affinity-purified F(ab')z fragments of fluorescein isothiocyanate (a-FcyRII) from Dr Clark Anderson (Ohio State University). (F1TC)-conjugated goat antimouse IgG (Jackson Immunore- Literature citations are provided in Table 1. Monoclonal IgG was search, West Grove, PA) for 30 minutes. The labeled platelets purified from ascites fluid by affinity chromatography on protein A were washed twice and fixed with 1% paraformaldehyde. Bound Sepharose CL4B (Pharmacia, Piscataway, NJ). FITC was analyzed by flow cytometry (FACStar-Plus; Becton Preparation and radiolabeling of Fab fragments. Fab fragments Dickinson, Mountain View, CA). Antibody binding was expressed of MoAb IV.3 were produced by papain digestion of purified IV.3 as mean fluorescence intensity in the linear mode.@ Background IgG followed by chromatographic purification on Protein A values determined with an irrelevant monoclonal (HB-57, an IgGl Sepharose CL-4B." The final preparation yielded bands of molec- MoAb specific for IgM) were subtracted from experimental values ular weight (MW) approximately 50 Kd (nonreduced) and 25 Kd to obtain net fluorescence intensity. (reduced) on sodium dodecyl sulfate (SDS) gel electrophoresis. Source of platelets. Platelets were obtained from normal unre- The Fab fragments were labeled with lZ5Iby the lactoperoxidase/ lated persons ranging in age from 21 to 60 years with informed glucose oxidase procedure using Enzymobeads (BioRad, Rich- consent. Donors had not ingested aspirin or other drugs known to mond, CA). Specific activity of the radiolabeled Fab was deter- inhibit platelet function for 1 week before study. mined by comparing protein content with counts per minute (cpm) Platelet aggregation. Platelet-rich plasma was prepared from determined in a gamma scintillation counter. By trichloroacetic freshly collected blood anticoagulated with one-tenth part of 3% acid precipitation, it was found that more than 95% of the sodium citrate. Donors had not ingested aspirin or other drugs radioactivity was protein-bound. known to inhibit platelet function for 1 week before testing. The Binding of radiolabeled IV3 Fab toplatelets. Binding of lZ51-Fab concentration of platelets was adjusted to about 300,00O/pL by from MoAb IV.3 to washed platelets was determined by the adding platelet-poor plasma from the same donor. The response to method of Court and LoB~glio~~with minor modification^.^^ MoAbs and other platelet agonists was measured using a Bio/Data PAP-4 aggregometer (Bio/Data Corp, Hatboro, PA) at 37°C with a Table 1. MoAbs Used stirring rate of 1,000 rpm. Reactions were terminated after 20 MoAb Target GP Subclass Reference minutes when platelets failed to aggregate. For reasons to be described (Results), we found it useful to measure the lag time OP-G2 CD41 (GPllb/llla) lgGi 34 required for platelets to respond to activating monoclonals. This ALE-6 CD9 lgG, 4.6 was defined as the time required for light transmission to reach one 50H.19 CD9 Wza 3, 35 half of its maximum value. CD36 (GPIV) lgGi 36 8A6 Release of 14C-serotoninfrom labeled platelets. Release of 14C- IV.3 CD32 (FcyRII) IgGzb 22,27 serotonin from platelets in plasma-free medium was determined by 41H.16 CD32 (FcyRII) Wza 28,37 a modification of the method of Sheridan et aL41Serotonin-labeled PLATELET ACTIVATING MONOCLONAL ANTIBODIES 2263 platelets were suspended in albumin-free Tyrode's buffer at a concentration of 3 x 105/pL and 75-pL aliquots were pipetted into the wells of a polystyrene microtiter tray (Dynatech, Chantilly, VA). Twenty-five microliters of test material or buffer was then added, and the trays were sealed and agitated gently on a horizontal rotator for 60 minutes at room temperature. Phosphate- buffered isotonic saline, 0.1 mL, pH 7.4 containing 0.5% EDTA was then added and the trays were spun at 2,500 rpm to pellet the crpz platelets. Radioactive content of 100-pLaliquots of the superna- tants was determined in a beta scintillation counter, and results crP3 were expressed as percent of total serotonin (determined from a detergent-lysed aliquot of labeled platelets). Release in control Fig 2. Variation in the aggregation response of platelets from tubes containing labeled platelets and an irrelevant MoAb was less three normal subjects (designated Grp 1, Grp 2, Grp 3) to MoAb OP-G2 than 5%. Release of 14C-serotoninfrom platelets in platelet-rich at 10 pg/mL. plasma was determined according to Griffith et al?5 Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 RESULTS aggregated only after a delay of 5 or 10 minutes. The extent of aggregation and the maximum slope of the aggregation MoAb-induced platelet aggregation. In preliminary stud- curve were generally less in platelets requiring a longer time ies, it was found that maximum aggregation of normal to respond. Further studies were conducted to determine platelets could be induced by 10 kg/mL of monoclonal the basis for these differences. OP-G2 (a GPIIb/IIIa), 20 pg/mL of ALB-6 (a-CD9), 5 Quantitation of FcyRII receptors on platelets. Measure- kg/mL of 8A6 (aGPIV), and 5 pg/mL of 50H.19 (a-CD9). ment of the binding of MoAb IV.3 Fab fragments to Higher incubations of MoAbs failed to further shorten the platelets of 20 unrelated normal subjects (10 male, 10 lag time of individual platelet preparations. The response of platelets to various amounts of MoAb 8A6 is shown in female) demonstrated a range of binding sites from 1,660 to Fig 1A. Qualitatively similar dose-response curves were 4,610 per platelet. The specificity of binding was confirmed obtained with MoAbs OP-G2, ALB-6, and 50H.19. The by showing that platelet-associated radioactivity was re- response of platelets to each of the IgGl MoAb (8A6, duced by more than 99% when platelets were preincubated ALB-6, OP-G2) was totally inhibited by prior addition of with a 50-fold excess of nonlabeled IV.3 Fab. The number MoAb IV.3 or Fab from IV.3 (shown for MoAb OP-G2 in of IV.3 Fab binding sites was stable (212% about the Fig 1B). The response to the IgGP, MoAb 50H.19 (a-CD9) mean) in four individuals, two with relatively high and two was not inhibited by IV.3 (see below). with relatively low FcyRII expression, who were studied on Platelets from different normal subjects varied signifi- four occasions over a 6-month period. Platelets from males cantly in their response to the three activating MoAbs of and females did not differ significantly in the quantities of the IgGl subclass (Fig 2). Platelets from some donors IV.3 Fab bound. Comparison ZK.3 and 4IH.16 binding. responded within 1or 2 minutes; others never responded or of Binding of the a-FcyRII MoAbs IV.3 and 41H.16 to platelets of the same group of normal subjects was measured by flow cytometry. Platelets from different donors varied considerably in the quantity of IV.3 bound as expected from the IV.3 Fab binding studies (see above). MoAb 41H.16 exhibited a different binding pattern in that no detectable amounts were bound to platelets from five members of the panel, while platelets from the remaining 15 donors bound variable amounts. When the ratio of 41H.16 fluorescence to iaal -lmin IV.3 fluorescence was determined, the value was approximately zero for the five platelet preparations that failed to bind 41H.16. The binding ratios for the remaining 15 donors allowed them to be segregated into two additional groups with ratios ranging from 0.19 to 0.39 (10 donors) and I \ from 0.58 to 0.63 (five donors) (Fig 3). The three groups will be referred to as groups 1, 2, and 3, respectively, in the following discussion. Varying ratios of 41H.16:IV.3 binding to platelets were also observed by Gosselin et a125 in six selected normal subjects. Response of platelets to the &GI MoAbs OP-G2, 8A6, and 100 1 ALB-6. Platelets in the three groups defined by the 41H.16:IV.3 binding ratios (Fig 3) differed in respect to the Fig 1. (A) Aggregation of platelets in response to different concentrations of MoAb 8A6 (a-GPIV). (6) Aggregation induced by MoAb time required for MoAbs OP-G2 (a-GPIIb/IIIa), 8A6 OP-G2 (anti-GPllb/llla) at concentration 20 pg/mL was inhibited by (a-GPIV), and ALB-6 (a-CD9) to induce an aggregation Fab fragmentsfrom MoAb IV.3 (crFcyRII). response (lag time) as shown in Fig 4. Platelets from donors TOMIYAMA ET AL

0.8 in group 1 responded slowly, or not at all, while those in group 3 all responded rapidly. The response of platelets from donors in group 2 varied between the two extremes. Grp 3 The magnitude of the aggregation responses and the 0.6 maximum slopes of the aggregation curves were also significantly lower with group 1 than with group 3 platelets (not shown). However, the distinction between groups 1 and 3 was clearest when lag times were compared. Aggregation . responses to each MoAb were completely inhibited by prior addition of MoAb IV.3 at 10 kg/mL (not shown). Platelets .. Grp 2 in the three groups responded equally well to the conven- m . tional agonists adenosine diphosphate (ADP) and collagen. m. The absolute number of FcyRII molecules per platelet, . determined by binding of IV.3 Fab, was slightly higher in group 3 (average 3,510 ? 570 molecules) than in group 1 Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 platelets (average 2,830 2 910 molecules) and was interme- Grpl diate in group 2 platelets (2,940 2 560 molecules). How- ever, these differences were not statistically significant. Among platelets in group 2, sensitivity to MoAb OP-G2 (a-GPIIb/IIIa) correlated with the number of FcyRII receptors expressed (3= .756, P < .01) (Fig 5). No relation- 0.2 ! I I 0 10 20 ship was found between FcyRII expression and the response of group 2 platelets to MoAbs 8A6 and ALB-6. Platelet Donor Panel Thus, responsiveness of platelets in group 2 to OP-G2, but not to 8A6 or ALB-6, is influenced by the number of Fig 3. Ratio of 41H.16:IV.3 binding in platelets from 20 normal FcyRII receptors. subjects. MoAb binding at saturation was determined by flow cytometry using FITC-labeled Flab'), specific for mouse IgG as a secondary Response of platelets to the ZgGr, MoAb, 50H.19. In antibody. Values shown are the average of two or three independent contrast to their differing sensitivity to the IgGl MoAbs measurements with each MoAb on platelets of each donor. Donors OP-G2, ALB-6, and 8A6, platelets in groups 1, 2, and 3 were divided into groups 1, 2, and 3 on the basis of 41H.16:IV.3 aggregated equally well in response to the IgG2, MoAb binding ratios. 50H.19 which, like ALB-6, is specific for CD9. The responses to 50H.19 could not be inhibited by prior addition of MoAb IV.3 at concentrations as high as 50 yg/mL. . . . . _L. . 9. a0-

"I - - -- ~,-pi 0rp2 G-3 Owl OW2 Grp3 OW OW2 OW3

OP-02 ALE-6 8A6 0-r

Fig 4. Aggregation responses of platelets from 20 normal subjects to platelet-activating MoAbs OP-G2 (10 pg/mL), ALB-6 (20 pg/mL), wa flab) Blndlng SIIOS pr maem and 8A6 (5 pg/mL). Platelets were assigned to groups 1, 2, and 3 on the basis of MoAb 41H.16:IV.3 binding ratios (see Fig 3 and text). Fig 5. Aggregation responses of platelets from group 2 donors to Ordinate depicts inverse lag time of aggregation response in min- MoAb OP-G2 (a-GPllb/llla) at 10 pg/mL. Ordinate depicts inverse lag utes-' x 102. A value of 0 indicates an "infinite" lag time, ie, no time of aggregation response in minutes-' x 102. Response was response within 20 minutes. Values shown are the average of two or roughly proportional to the number of binding sites for IV.3 Fab, ie, three independent measurements on platelets from each donor. the number of FcyRll receptors (P e .01) (y = .02x - 34.8, fl = .756). Horizontal bars indicate average values for each group. Groups 1and Similar, but less striking relationships were observed between FoyRll 3 differed significantly from each other for each of the three MoAbs expression and response to MoAbs 8A6 (a-GPIV) and ALB-6 (a-CDS) (P < .Ol). (not shown). PLATELET ACTIVATING MONOCLONAL ANTIBODIES 2265

Scrotonin relcasc was also induccd by MoAb 50H.19 (IgGl, a-CD9). This response was completely blockcd by ... MoAb IV.3, in contrast to aggregation induced by this MoAb (Fig 7A). Bccausc the scrotonin rcleasc studies were pcrformcd in thc absence and thc aggrcgation studics in thc prcscnce of plasma, it sccmcd possible that plasma affcctcd ... the ability of IV.3 to inhibit activation by this MoAb. This was confirmcd by dcmonstrating that IV.3 failed to inhibit scrotonin rclcasc triggcred by 50H.19 in platelct-rich plasma (Fig 7B). In contrast, IV.3 completely inhibited scrotonin rclcasc induced by the IgG, MoAbs ALB-6, 8A6, and OP-G2 in platelet-rich plasma (not shown). Plasma and serum hcated to 56°C for 30 minutes and unheated plasma containing 10 mmol/L EDTA failed to support serotonin Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021

_II_

Group 3 PImlelels

75 .. -1 .. T

11 0.5 0.5 11 0.5 0.5

0 10 0 10 0 10 0 10

6 1 0,s 0.26 0.126

-r*m o( ALe4 (ughn~)

Fig 6. Release of lT-serotonin from platelets of a donor from group 3 (r;) and a donor from group 1 (a)in response to different amounts of MoAk OPG2 (A) and ALB-6 (B) (average of three determinations). Error bars indicate 21 SD. Significance of differences is *** <.001, ** e.01, <.M.Results shown are typical of three paired studies in which responses of group 1 and group 3 platelets were compared. Similar results were obtained with MoAb 8A6 (not shown).

Induction of serotonin release hy MoAbs. t4C-scrotonin relcasc from platclcts was induccd by about one tenth the amount of MoAb rcquircd to induce an aggregation rc- sponsc. As shown in Figs 6, A and B, at low conccntrations, thc IgGl MoAbs caused relcase of I4C-scrotonin from 54(.19(UPrmO S s 11 55 $1 group 3 platclcts morc readily than from group 1 platelets. w.3 cuo" 0 IO 0 10 0 10 0 10 At highcr conccntrations of MoAb, group 1 platclcts also rclcascd scrotonin. As with aggregation, scrotonin rclcase Fig 7. Serotonin release induced by MoAb 50H.19 In group 1 and group 3 platelets in plasma-free (A) and plasma-containing media (B). was blocked by prior addition of MoAb IV.3 (a FcyRII) at Release was inhibited by MoAb IV.3 in buffer (A), but not in plasma 10 wg/mL (not shown). le). 2266 TOMIYAMA ET AL release in response to 50H.19 in the presence of IV.3. responded similarly to HAIgG, indicating that this particu- These findings suggest that complement activation by lar response is not affected by FcyRII phenotype. Similar 50H.19 accounts for its ability to induce the release reaction observations were made by Looney et alez8 in platelets whose FcyRII receptors are blocked by IV.3. Target molecules recognized by the three IgGl MoAbs used in our study are expressed in different numbers on the DISCUSSION platelet surface (CD36 about 20,000, CD41 about 40,000 to Our studies show that platelets from different normal 50,000, and CD9 about 65,000 molecules per platelet). subjects vary greatly in their responsiveness to IgGl MoAbs However, MoAb 8A6 specific for CD36 produced a maxi- specific for membrane glycoproteins CD9, GPIV, and mum platelet response at a concentration lower than that GPIIb/IIIa. Inhibition of the action of these MoAbs by required for OP-G2 or ALB-6. Therefore, within the range MoAb IV.3 specific for the FcyRII receptor is confirmatory from 20,000 to 65,000, the number of target molecules per of previous report^^.^-^^ and consistent with the view that the platelet does not appear to be an important determinant of Fc domains of activating MoAbs react with platelet FcyRII response. CD9, CD36, and CD41 are not known to vary receptors to induce release and aggregation. Recent re- significantly on platelets of normal subjects except in rare Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 ports indicate that this occurs both on an intra-platelet and individuals with genetically determined deficiencies of inter-platelet basi~.~J~ CD3645 or CD41 (type I Glanzmann’s thrombasthenia). Our quantitative studies of FcyRII expression on plate- The concordant behavior of group 1 platelets in response to lets using radiolabeled Fab from MoAb IV.3 showed stable all three of the IgGl MoAbs essentially rules out the differences in the number of FcyRII expressed on platelets possibility that their poor responses were caused by a of normal individuals, in confirmation of the findings of deficiency of target protein. Rosenfeld et al,42who used intact IV.3. Platelets from the As noted, the FcyRII gene family encodes a number of 20 subjects studied bound 1,660 to 4,610 IV.3 Fab (average mature proteins expressed differently in various cell types. 3,060), about twice the value (1,350) obtained by Karas et al 41H.16 is thought to bind to all FcyRII gene products using a dimeric IgG pr0be.4~With intact IV.3, we obtained except the allelic form of FcyRIIA that contains a histidine, values for FcyRII expression about half as great as those rather than an arginine residue, at position 131.15 There- found with Fab (data not shown). King et aPalso found fore, the failure of group 1 platelets to bind detectable that IV.3 Fab binding to platelets was about twice that of amounts of 41H.16 is consistent with the possibility that intact IV.3. The total number of binding sites per platelet only products of the FcyRII gene and not those of the they identified using labeled IV.3 Fab (4,700) was about FcyRII B and C genes are expressed on platelets. twice the number we measured. However, their binding Activation of platelets by the IgG2, MoAb 50H.19 spe- studies were performed at low ionic strength, which they cific for CD9 requires special comment because this MoAb, found doubled the apparent number of Fc receptor sites. in contrast to OP-G2, ALB-6, and 8A6, failed to differenti- Lower values for Fc receptors obtained with intact MoAbs ate between HR and LR platelets and because, in a plasma than with Fab may reflect bivalent binding of the intact medium, activation by 50H.19 could not be blocked by IV.3, molecule to FcyRII. even at concentrations as high as 50 pg/mL. However, IV.3 On the basis of the 41H.16:1V.3 binding ratios, it was readily blocked the action of 50H.19 on platelets in a possible to divide our panel into three groups (Fig 3). Using plasma-free medium (Fig 7). The IgG2, subclass of murine nomenclature applied to monocyte^,^^-^^ these appear to Ig is a potent activator of the classical complement pathway, correspond to homozygotes for the LR form of FcyRII in contrast to the IgGl subclass,46 and it is known that (group l), homozygotes for the HR form (group 3), and release can be triggered by assembly of the terminal LR/HR heterozygotes (group 2). Platelets judged to be HR complement components in the platelet membrane.47These homozygous on this basis all responded well to activating facts, together with our finding that heated plasma and MoAbs of the IgGl subclass, and those judged to be LR serum and plasma containing 10 mmol/L EDTA were homozygous all responded poorly (Fig 4). However, plate- unable to support 5OH.19-induced release in the presence lets from apparent heterozygotes (group 2) varied widely in of IV.3, strongly suggest that 50H.19 activates platelets by their sensitivity to MoAbs. Responsiveness of group 2 two mechanisms: one complement-dependent and the other platelets to OP-G2 correlated fairly well with the number of dependent on binding of its Fc region to FcyRII. Griffith et FcyRII molecules expressed (Fig 5). No such relationship a135 have recently shown that immobilized F(ab’)z from was observed with 8A6 or ALB-6. These findings indicate 50H.19, but not soluble F(ab’)z, can activate platelets by a that both FcyRII phenotype and the number of FcyRII direct action on CD-9.35 molecules carried on the platelet surface influence the Together with previous reports, our findings indicate that responsiveness of platelets from heterozygous (group 2) MoAbs activate platelets by several different mechanisms: individuals. The variation in response of group 2 platelets to (1) Fc-mediated activation of FcyRII; (2) complement 8A6 and ALB-6 implies that other, unrecognized factors activation; and (3) in the case of certain MoAbs,1,s.12direct are also important. Rosenfeld et a142 found that the re- activation via the target protein to which they bind. Plate- sponse of washed platelets to heat-aggregated IgG (HAIgG) lets from normal subjects vary greatly in their susceptibility correlated with the number of FcyRII receptors expressed. to Fc-mediated activation by IgGl MoAbs, and failure to We confirmed this finding with our platelet panel (data not respond to these MoAbs is associated with homozygosity shown). However, platelets from donors in groups 1 and 3 for one of the two defined alleles of the FcyRIIA gene (Figs PLATELET ACTIVATING MONOCLONAL ANTIBODIES 2267

3 and 4). In the absence of plasma, “non-responder” studies of platelet FcR may be relevant to platelet destruc- platelets can be induced to release serotonin by increasing tion mediated by alloantibodies reactive with Class I HLA the concentration of stimulating MoAbs (Fig 6). However, alloantigens. A number of reports of platelet activation and high concentrations of MoAbs did not cause group 1 thrombosis induced by platelet-reactive autoantibodies have platelets to aggregate in platelet-rich plasma. These obser- appeared,52-ssand it seems possible that platelet activation vations suggest that the two alleles of FcyRII differ in their occurs via FcR in some of these patients. In one study, a affinity for murine IgGl Fc and that high levels of mono- mild hemostatic defect was observed in three patients meric IgG in human plasma modulate the response to these whose platelets responded poorly to an MoAb specific for MoAbs by competing with their Fc domains for binding to CD9.56Finally, heparin-induced thrombocytopenia/ throm- FcyRII. Thus, in the presence of plasma, only the allele of bosis appears to be mediated by immune complexes reac- FcyRII with higher affinity for IgGl Fc is activated. tive with platelet FcR~~and platelets from different normal Although FcR of classes I, 11, and I11 appear to be subjects differ widely in their sensitivity to heparin- redundant in their functions, it has been argued that dependent antibodie~.~~,~~Further studies to determine FcyRII has special functional Whether the whether the FcyRII receptor phenotype of platelets predis- FcyRII phenotype of platelets is relevant to platelet physi- poses to any of these conditions appear warranted. Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 ology or pathophysiology in humans therefore deserves consideration. A recent report showed that the LR form of ACKNOWLEDGMENT FcyRIIA preferentially binds multimers of human I~Gz,~~ We are indebted to Drs Clark Anderson, A.R.E. Shaw, and John the IgG subtype mainly involved in the human immune Barnwell for generously supplying monoclonal antibodies used in response to saccharide antigens. Alloantibodies specific for these studies; to Kurt Wolfmeyer for excellent technical assistance; class I HLA antigens are capable of inducing platelet to Dr Peter Newman for helpful advice; and to the Word activation and aggregati~n~~,~~and may bind to platelets at a Processing Department of The Blood Center for preparation of site close to the FcyRII receptor.s* Therefore, further this manuscript.

REFERENCES 1. Morel MC, Lecompte T, Champeix P, Favier R, Potevin F, JC, Kunicki TJ: The Arg-Gly-Asp (RGD) recognition site of Samama M, Salmon C, Kaplan C: PL2-49, a monoclonal antibody platelet glycoprotein IIb-IIIa on nonactivated platelets is accessi- against glycoprotein IIb which is a platelet activator. Br J Haematol ble to high affinity macromolecules. Blood 79:2303,1992 7157, 1989 12. Frelinger AL 111, Du X, Plow EF, Ginsberg MH: Mono- 2. Slupsky JR, Seehafer JG, Tang SC, Masellis-Smith A, Shaw clonal antibodies to ligand-occupied conformers of integrin ARE: Evidence that monoclonal antibodies against CD9 antigen (glycoprotein IIb-IIIa) alter receptor affinity, specificity, and func- induce specific association between CD9 and the platelet glycopro- tion. J Biol Chem 266:17106, 1991 tein IIb-IIIa complex. J Biol Chem 264:12289, 1989 13. Ockenhouse CF, Magowan C, Chulay JD: Activation of 3. Gulino D, Ryckewaert JJ, Andrieux A, Rabiet MJ, Marquerie monocytes and platelets by monoclonal antibodies or malaria- G: Identification of a monoclonal antibody against platelet GPIIb infected erythrocytes binding to the CD36 surface receptor in vitro. that interacts with a calcium-binding site and induces aggregation. J Clin Invest 84:468,1989 J Biol Chem 265:9575,1990 14. Anderson GP, Anderson CL: Signal transduction by the 4. Worthington RE, Carroll RC, Boucheix C: Platelet activation platelet Fc receptor. Blood 76:1165, 1990 by CD9 monoclonal antibodies is mediated by the FcyII receptor. 15. Ravetch JV, Kinet J-P: Fc receptors. Annu Rev Immunol Br J Haematol74:216,1990 9:457,1991 5. Kornecki E, Walkowiak B, Naik UP, Ehrlich YH Activation 16. van de Winkel JGJ, Anderson CL: Biology of human of human platelets by a stimulatory monoclonal antibody. J Biol immunoglobulin G Fc receptors. J Leuk Biol49:511,1991 Chem 265:10042,1990 17. Hibbs ML, Bonadonna L, Scott BM, McKenzie IFC, Ho- 6. Carroll RC, Worthington RE, Boucheix C Stimulus-response garth PM: Molecular cloning of a human immunoglobulin G Fc coupling in human platelets activated by monoclonal antibodies to receptor. Proc Natl Acad Sci USA 85:2240,1988 the CD9 antigen, a 24 kDa surface-membrane glycoprotein. 18. Stengelin S, Stamenkovic I, Seed B: Isolation of cDNAs for Biochem J 266527,1990 two distinct human Fc receptors by ligand affinity cloning. EMBO J 7. Jennings LK, Fox CF, Kouns WC, McKay CP, Ballou LR, 7:1053,1988 Schultz HE: The activation of human platelets mediated by 19. Stuart SG, Simister NE, Clarkson SB, Kacinski BM, Shapiro anti-human platelet p24/CD9 monoclonal antibodies. J Biol Chem M, Mellman I: Human IgG Fc receptor (hFcRII; CD32) exists as 265:3815,1990 multiple isoforms in macrophages, lymphocytes and IgG-transport- 8. Kouns WC, Wall CD, White MM, Fox CF, Jennings LK A ing placental epithelium. EMBO J 8:3657,1989 conformation-dependent epitope of human platelet glycoprotein 20. Brooks DG, Qui WQ, Luster AD, Ravetch JV: Structure IIIa. J Biol Chem 265:20594,1990 and expression of human IgG FcRII (CD32). Functional heteroge- 9. Horsewood P, Hayward CPM, Warkentin TE, Kelton JG: neity is encoded by the alternatively spliced products of multiple Investigation of the mechanisms of monoclonal antibody-induced genes. J Exp Med 170:1369,1989 platelet activation. Blood 78:1019,1991 21. Qiu WQ, de Bruin D, Brownstein BH, Pearse R, Ravetch 10. Anderson GP, van de Winkel JGJ, Anderson CL: Anti- JVOrganization of the human and mouse low-affinityFcyR genes: GPIIb/IIIa (CD41) monoclonal antibody-induced platelet activa- Duplication and recombination Science 248:732, 1990 tion requires Fc receptor-dependent cell-cell interaction. Br J 22. Looney RJ, Abraham GN, Anderson CLHuman monocytes Haematol79:70,1991 and U937 cells bear two distinct Fc receptors for IgG. J Immunol 11. Tomiyama Y, Tsubakio T, Piotrowicz RS, Kurata Y, Loftus 136:1641,1986 2268 TOMIYAMA ET AL

23. Tax WJM, Willems HW, Reekers PPM, Capel PJA, Koene 41. Sheridan D, Carter C, Kelton JG: A diagnostic test for RAP: Polymorphism in mitogenic effect of IgGl monoclonal heparin-induced thrombocytopenia. Blood 67:27,1986 antibodies against +3 antigen on human T cells. Nature 304:445, 42. Rosenfeld SI, Ryan DH, Looney RJ, Anderson CL, Abra- 1983 ham GN, Leddy JP: Human Fcy receptors: Stable inter-donor 24. Tax WJM, Spits H, Hermes HFM, Willems HW, Capel PJA, variation in quantitative expression on platelets correlates with Koene RAP: Polymorphism of human Fc receptors for mouse IgGl functional responses. J Immunol138:2869,1987 and IgG2b. Transplant Proc 17:794,1985 43. Karas SP, Rosse WF, Kurlander RJ: Characterization of the 25. Gosselin EJ, Brown MF, Anderson CL, Zipf TF, Guyre PM: IgG-Fc receptor on human platelets. Blood 60:1277,1982 The monoclonal antibody 41H16 detects the Leu 4 responder form 44. King M, McDermott P, Schreiber AD: Characterization of of human FcyRII. J Immunol144:1,1990 the Fcy receptor on human platelets. Cell Immunol 128:462, 1990 26. Boot JHA, Geerts MET, Aarden LA Functional polymor- phisms of Fc receptors in human monocyte-mediated cytotoxicity 45. Yamamoto N, Ikeda H, Tandon NN, Herman J, Tomiyama towards erythrocytes induced by murine isotype switch variants. J Y, Mitani T, Sekiguchi S, Lipski R, Kralisz U, Jamieson GA A Immunol142:1217,1989 platelet membrane glycoprotein (GP) deficiency in healthy blood 27. Anderson CL, Ryan DH, Looney RJ, Leary PC: Structural donors: Nak"-negative platelets lack detectable GPIV (CD36). Blood 76:1698,1990

polymorphism of the human monocyte 40 kilodalton Fc receptor Downloaded from http://ashpublications.org/blood/article-pdf/80/9/2261/608561/2261.pdf by guest on 28 September 2021 for IgG1. J Immunol138:2254,1987 46. Hardy RR: Complement fixation by monoclonal antigen- 28. Looney RJ, Anderson CL, Ryan DH, Rosenfeld SI: Struc- antibody complexes, in Weir DM (ed): Handbook of Experimental tural polymorphism of the human platelet Fcy receptor. J Immunol Immunology, vol. 1 Immunochemistry. Oxford, UK, Blackwell 141:2680,1988 Scientific, 1986 29. Clark MR, Clarkson SB, Ory PA, Stollman N, Goldstein IM: 47. Sims PJ, Wiedmer T: The response of human platelets to Molecular basis for a polymorphism involving Fc receptor I1 on activated components of the complement system. Immunol Today human monocytes. J Immunol143:1731,1989 12:338,1991 30. Warmerdam PAM, van de Winkel JGJ, Gosselin EJ, Capel 48. Tax WJM, van de Winkel JGJ: Human Fcy receptor 11: A PJA Molecular basis for a polymorphism of human Fcy receptor I1 standby receptor activated by proteolysis? Immunol Today 11:308, (CD32). J Exp Med 172:19,1990 1990 31. Warmerdam PAM, van de Winkel JGJ, Vlug A, Westerdaal 49. Holmund G, Tiilikainen A, Penttinen K The effect of HLA NAC, Capel PJA A single amino acid in the second Ig-like domain antibodies on platelet aggregation. Scand J Immunol6:157, 1977 of the human Fcy receptor I1 is critical for human IgG2 binding. J 50. Heinrich D, Stephinger U, Mueller-Eckhardt C: Specific Immunol147:1338,1991 interaction of HLA antibodies (eluates) with washed platelets. Br J 32. Chong BH, Fawaz I, Chesterman CN, Berndt MC Heparin- Haematol35:441, 1977 induced thrombocytopenia: Mechanism of interaction of the heparin-dependent antibody with platelets. Br J Haematol 73:235, 51. Endresen GKM: Studies on the steric relationship between 1989 HLA antigens, betaz-microglobulin and the Fc receptor on human 33. Aster RH: The immune thrombocytopenias, in Kunicki TJ, platelets. Thromb Res 47:441,1987 George JN (eds): Platelet Immunology. Philadelphia, PA, Lippin- 52. Sugiyama T, Okuma M, Ushikubi F, Sensaki S, Kanaji K, cott, 1989, p 387 Uchino H: A novel platelet aggregating factor found in a patient 34. Tomiyama Y, Honda S, Furubayashi T, Mizutani H, Tsu- with defective collagen-induced platelet aggregation and autoim- bakio T, Kurata Y, Yonezawa T, Tarui S: A new monoclonal mune thrombocytopenia. Blood 69:1712,1987 antibody, OP-G2, against platelet glycoprotein IIb/IIIa that in- 53. Jackson SP, Jane SM, Mitchell CA, Cortizo WF, Hau L, duces platelet aggregation. Blood and Vessel 19:257,1988 Pfueller SL, Salem HH: Arterial thrombosis associated with 35. Griffith L, Slupsky J, Seehafer J, Boshkov L, Shaw ARE: immune thrombocytopenia: Presence of a platelet aggregating IgG Platelet activation by immobilized monoclonal antibody: Evidence synergistic with thrombin and adrenalin. Thromb Haemost 622346, for a CD9 proximal signal. Blood 78:1753, 1991 1989 36. Barnwell JW, Asch AS, Nachman RL, Yamaya M, Aikawa 54. Pfuller SL, David R, Firkin BG, Bilston RA, Cortizo WF, M, Ingravallo P: A human 88-kD membrane glycoprotein (CD36) Rains G: Platelet aggregating IgG antibody to platelet surface functions in vitro as a receptor for a cytoadherence ligand on glycoproteins associated with thrombosis and thrombocytopenia. plasmodium falcipamm-infected erythrocytes. J Clin Invest 84:765, Br J Haematol74:336,1990 1989 55. Yanabu M, Nomura S, Fukuroi T, Soga T, Kondo K, Sone N, 37. Zipf TF, Lauzon GJ, Longenecker BM: A monoclonal Kitada C, Nagata H, Kokawa T, Yasunaga K Synergistic action in antibody detecting a 39,000 MW molecule that is presenting on B platelet activation induced by an antiplatelet autoantibody in ITP. lymphocytes and chronic lymphocytic leukemia cells but is rare on Br J Haematol78:87,1991 acute lymphocytic leukemia blasts. J Immunol131:3064, 1983 56. Jennings LK, Chesney CM, Fox CF, Mauer AM: A platelet 38. Court WS, LoBuglio AF: Measurement of platelet surface- aggregation defect which suggests the interaction of p24/CD9 with bound IgG by a monoclonal Iz5I-anti-IgG assay. Vox Sang 50:154, GPIb. Blood 72:325a, 1988 (abstr, suppl) 1986 39. Janson M, McFarland JG, Aster RH: Quantitative determi- 57. Salem HH, van der Weyden MB: Heparin induced throm- nation of platelet allo-antigens using a monoclonal probe. Hum bocytopenia. Variable platelet-rich plasma reactivity to heparin Immunol15:251,1986 dependent platelet aggregating factor. Pathology 15:297,1983 40. Visentin GP, Wolfmeyer K, Newman PJ, Aster RH: Detec- 58. Pfueller SL, David R: Different platelet specificities of tion of drug-dependent, platelet-reactive antibodies by antigen- heparin-dependent platelet aggregating factors in heparin- capture ELISA and flow cytometry. Transfusion 30694,1990 associated immune thrombocytopenia. Br J Haematol64:149,1986