Justine S. Liepkalns, Eldad A. Hod, Sean R. Stowell...

I M M U N O H E M A T O L O G Y

Biphasic clearance of incompatible red blood cells through anovel mechanism requiring neither complement nor Fcg

receptors in a murine model_3647 2631..2645

Justine S. Liepkalns, Eldad A. Hod, Sean R. Stowell, Chantel M. Cadwell, Steven L. Spitalnik,

and James C. Zimring

BACKGROUND: Antibody binding to red blood cells(RBCs) can induce potentially fatal outcomes, includinghemolytic transfusion reactions (HTRs), hemolyticdisease of the fetus and newborn, and autoimmunehemolytic anemia. The mechanism(s) of RBC destruc-tion following antibody binding is typically thought torequire complement activation and/or the involvement ofFcg receptors (FcgRs). In the current report, we ana-lyzed mechanisms of HTRs during incompatible transfu-sions of murine RBCs expressing human glycophorin A(hGPA) into mice with anti-hGPA.STUDY DESIGN AND METHODS: C3 and Fcg receptorknockout, splenectomized, Fcg receptor blockingantibody–treated, and clodronate-treated mice werepassively immunized with anti-hGPA (10F7 or 6A7) andtransfused with RBCs expressing the hGPA antigen.Posttransfusion blood and serum were collected andanalyzed via flow cytometry and confocal microscopy.RESULTS: This HTR model results in both rapid clear-ance and cytokine storm. Neither complement norFcgRs were required for RBC clearance; in contrast,FcgRs were required for cytokine storm. Circulatingaggregates of hGPA RBCs were visible during the HTR.Splenectomy and phagocyte depletion by clodronatehad no effect on acute RBC clearance; however, incom-patible RBCs reentered over 24 hours in clodronate-treated mice.CONCLUSION: These data demonstrate a biphasicHTR, the first phase involving sequestration of incom-patible hGPA RBCs and the second phase involvingphagocytosis of sequestered RBCs. However, themechanism(s) of phagocytosis in the second phaserequired neither C3 nor FcgRs. These findings demon-strate novel mechanistic biology of HTRs.

Crossmatch-incompatible transfusions consistof infusing donor red blood cells (RBCs) into arecipient who has antibodies against antigenson the donor RBCs. Except for naturally occur-

ring blood group antibodies (e.g., the ABO system),alloantibodies to RBC antigens are generated throughprior exposure to allogeneic RBCs, typically by transfusionor pregnancy.1 In general, incompatible transfusions arestrictly avoided, as hemolysis of the transfused RBCs canoccur with potentially fatal outcomes, known as hemolytictransfusion reactions (HTRs).2,3 HTRs can occur duringamnestic antibody responses or as a result of clerical errorand mistransfusion.4,5 Moreover, in other instances,crossmatch-incompatible RBCs may be purposefullytransfused if the immediate risks of hypoxia (e.g., due toanemia) outweigh the potential damage from an inducedHTR. Alloantibodies against fetal RBCs can also causehemolytic disease of the fetus and newborn, resulting in

ABBREVIATIONS: DiI = 1,1′-dioctadecyl-3,3,3′3′-tetramethylindocarbocyanine perchlorate; DiO = 3,3′-dihexadecyloxacarbocyanine perchlorate; FcgRs = Fcg receptors;

hGPA = human glycophorin A; HTR(s) = hemolytic transfusion

reaction(s); MAC = membrane attack complex.

From the Center for Transfusion and Cellular Therapies, Depart-

ment of Pathology and Laboratory Medicine, Emory University,

and the Aflac Cancer Center, Department of Pediatrics, Emory

University School of Medicine, Atlanta, Georgia; and the

Department of Pathology and Cell Biology, Columbia University

College of Physicians and Surgeons–New York Presbyterian Hos-

pital, New York, New York.

Address reprint requests to: James C. Zimring, MD, PhD,

Puget Sound Blood Center, 1551 Eastlake Ave E, Suite 100,

Seattle, WA 98102; e-mail: [email protected].

Received for publication November 29, 2011; revision

received February 8, 2012, and accepted February 15, 2012.

doi: 10.1111/j.1537-2995.2012.03647.x

TRANSFUSION 2012;52:2631-2645.

Volume 52, December 2012 TRANSFUSION 2631

substantial morbidity and/or mortality.6 Finally, autoanti-bodies bound to RBCs can cause autoimmune hemolyticanemia.7 Thus, an understanding of the mechanisms bywhich antibody-bound RBCs are cleared from the circula-tion and the pathophysiology that ensues is important forseveral disease settings.

Antibody-induced hemolysis is generally thought tooccur by one of two mechanisms.8 First, RBCs can undergointravascular hemolysis when complement is activated toform the membrane attack complex (MAC). This is typi-cally due to immunoglobulin (Ig)M binding to the RBCsurface, but also occasionally occurs with IgG.9,10 RBCs canalso be opsonized and ingested by phagocytes, a processreferred to as extravascular hemolysis. If complement acti-vation does not lead to MAC formation to the extent thatrapid intravascular hemolysis occurs, then C3 depositedon the RBC surface may be converted to C3b and iC3b. Inthis case, complement receptors on phagocytes (i.e., CR1,CR2, and CR3) can consume C3b-coated RBCs. Second,antibody-induced opsonization can occur as a result of Fcdomains of IgG bound to RBCs, which are recognized byFcg receptors (FcgRs) on phagocytes extravascularly. Thereare substantial and significant data, generated mostly inanimal models, that demonstrate the existence of each ofthese pathways of RBC clearance after antibody binding.8,11

In contrast to the canonical pathways described,alternative mechanisms of IgG-mediated RBC clearancehave been suggested, which involve neither complementnor FcgR pathways. These include direct effects on theRBC by antibody binding. Destabilization of the RBCmembrane can induce programmed RBC death (erypto-sis).12,13 Antibodies have also been shown to directlyinduce phosphatidylserine expression on the RBC surface,which can lead to phagocytosis as a result of ligating phos-phatidylserine scavenger receptors on phagocytes.12

Moreover, mechanical induction of Ca2+ influx into RBCs,which leads to eryptosis, has also been demonstrated asa direct effect of antibody surface binding.14,15 Thesepathways are not mutually exclusive and may overlap.Although these alternative pathways raise importantmechanistic questions regarding HTRs, they have beenidentified in vitro; therefore, it is unclear if these findingsare relevant to authentic HTRs in vivo. Finally, RBCremoval from circulation in vivo by sequestration in thespleen and liver has been observed in the context of IgGand IgA autoimmune hemolytic anemia, but a role forsequestration has not been reported in the context ofincompatible transfusion.16,17

Herein, we utilize an in vivo murine model of incom-patible RBC transfusion involving IgG antibodies in therecipient against human glycophorin A (hGPA) epitopeson donor RBCs. The 6A7 and 10F7 mouse monoclonalantibodies (MoAbs) recognize the antithetical M humanblood group antigen or a nonpolymorphic common hGPAepitope, respectively.18-20 Our data demonstrate clearance

of RBCs by a mechanism requiring neither complementnor FcgRs. To our knowledge, this is the first report iden-tifying this alternative pathway of IgG antibody–inducedhemolysis in vivo in the setting of crossmatch-incompatible transfusion. In addition, our findingsmechanistically dissociate RBC clearance from phagocyteingestion and cytokine storm. Together, these findingsprovide insights into a novel HTR mechanism in mice, theprevalence of which is unknown in human HTRs.

MATERIALS AND METHODS

MiceC57BL/6 (B6), FVB,21 and C3 knockout mice were pur-chased from The Jackson Laboratories (Bar Harbor, ME).Both B6 and FVB mice represent long-term lines of inbredanimals routinely utilized in murine studies. FVB mice areparticularly advantageous in transgenic work due to theirlarge pronucleus, facilitating DNA microinjections, andlarge litters.21 FcgR KO mice (Fcer1g) were purchased fromTaconic Farms, Inc. (Hudson, NY). All mice were used at 8to 16 weeks of age. Transgenic mice expressing the Mvariant of the hGPA on RBCs alone were obtained from theNew York Blood Center (New York, NY).22 HOD mice con-sisted of transgenic animals expressing a fusion proteinhen egg lysozyme, ovalbumin, and Duffy b under an RBCspecific promoter.23 Both hGPA and HOD mice are on anFVB background. HOD and hGPA mice were crossed togenerate HOD ¥ hGPA F1 mice that express both trans-genes. All breeding (including FcgR KO and C3 KO mice)was performed by the Emory University Department ofAnimal Resources Husbandry Services and all procedureswere performed according to approved IACUC protocols.

Antibodies and passive immunizationAnti-hGPA (6A7 and 10F7) mouse MoAbs are of the IgG1subtype; 6A7 is specific for the M variant of hGPA,18-20

whereas 10F7 recognizes both M and N variants.18,19 Anti-Fy3 (MIMA29) is an IgG2a antibody that binds to the Duffyportion of the HOD antigen on the third extracellular loop(Fy3)24 and was a gift from M. Reid and G. Halverson at theNewYork Blood Center. In some experiments, 2.4G2, an Fcgreceptor–blocking antibody,25 was administered to recipi-ents intraperitoneally 24 hours before passive immuniza-tion with anti-hGPA. All above antibodies were purified byprotein G chromatography (Bio X Cell,West Lebanon, NH).Recipients were passively immunized by tail vein injectionwith 500 mg of 10F7, 25 mg of 6A7, or 200 mg of MIMA29 in atotal volume of 500 mL of phosphate-buffered saline (PBS),at least 2 hours before transfusion; control mice received500 mL of PBS alone. 10F7 was used at a concentrationoptimized in a previous study by Schirmer and col-leagues;26 doses of 6A7 and MIMA29 were optimized basedupon potency in inducing RBC clearance.

LIEPKALNS ET AL.

2632 TRANSFUSION Volume 52, December 2012

Fluorescent labeling, transfusion, and survivalmonitoring of murine RBCsRBC fluorescent labeling, treatment with phenylhydra-zine, transfusion, survival determination, and splenecto-mies were performed as previously described.27-29

hGPA ¥ HOD, hGPA, or HOD RBCs were labeled with 1,1′-dioctadecyl-3,3,3′3′-tetramethylindocarbocyanine per-chlorate (DiI) and control (FVB) blood was labeled with3,3′-dihexadecyloxacarbocyanine perchlorate (DiO). Thesurvival of DiI-labeled RBCs was calculated by normaliz-ing it to DiO-labeled wild-type FVB RBCs within the sameanimal. These resulting ratios from passively immunizedmice were then normalized to those of PBS-treatedanimals and graphed as percentage of hGPA ¥ HOD,hGPA, or HOD RBC survival.

Fluorescent microscopyBlood samples were collected after transfusion andsmears were prepared on glass slides. Smears were air-dried overnight to increase RBC adhesion to the slide (inthe dark to minimize photobleaching). Slides were thenmounted with permanent mounting medium (Vecta-Mount, Vector Laboratories, Burlingame, CA) and visual-ized with a point scanning laser confocal microscope(LSM 510, Zeiss, Thornwood, NY).

Clodronate treatment of recipient miceC57BL/6 mice were treated with plain (empty) liposo-mes (9.4 mg/mL l-a-phosphatidylcholine, 2.1 mg/mLcholesterol) or liposomes filled with clodronate(dichloromethylene-bisphosphonate [Cl2MBP]) intrave-nously (IV) 24 hours before passive immunization andtransfusion. Each mouse received 1.5 mg of clodronate in600 mL total volume. Control mice received the total600 mL volume in one setting, whereas the experimentalmice received clodronate-filled liposomes slowly (approx.10 mL/sec, in two doses of 300 mL each).

Cytokine measurementsPlasma interleukin (IL)-6, tumor necrosis factor (TNF)-a,monocyte chemoattractant protein-1, keratinocyte-derived chemokine, and IL-10 were quantified using thecytometric bead array mouse flex kit (BD Biosciences, SanDiego, CA).

Statistical analysisSurvival graphs were all calculated and created with com-puter software (Prism, GraphPad, San Diego, CA). All errorbars represent one standard deviation. p values were cal-

culated, via log transformation, using a two-way ANOVAwith Bonferroni posttest using Prism software.

RESULTS

RBC clearance during crossmatch-incompatibletransfusion for the M and Fy3 bloodgroup antigenshGPA mouse RBCs express transgenic hGPA and HODmouse RBCs express a triple fusion transgenic proteinconsisting of hen egg lysozyme, ovalbumin, and humanDuffy. Both hGPA and HOD mice express their respectivetransgenes in an RBC-specific fashion.22,23 To generateRBC donors with multiple known RBC antigens, hGPA andHOD mice were crossed to yield F1 progeny (hGPA ¥ HODmice) expressing both the hGPA and the HOD transgeneson RBCs (see Fig. 1A for depiction and antibody-bindingsites). In some cases the single transgenic parental strains,hGPA or HOD, were used as donors (as indicated by singleantigen nomenclature). C57BL/6 recipients were pas-sively immunized with monoclonal anti-hGPA (10F7 or6A7) or monoclonal anti-Fy3 (MIMA29), followed bytransfusion with a mixture of wild-type FVB RBCs (labeledwith DiO) and hGPA ¥ HOD RBCs (labeled with DiI).hGPA ¥ HOD mice are on an FVB background. Analysis ofperipheral blood from control animals that were trans-fused, but not passively immunized, demonstrated thatthe hGPA (10F7 and 6A7) and Fy3 epitopes are expressedand detectable on hGPA ¥ HOD, but not on cotransfusedcontrol wild-type RBCs (Fig. 1A).

Survival of transfused RBCs was determined by enu-merating labeled RBCs by flow cytometry and normalizingthe circulating incompatible RBCs as a function of wild-type RBCs within the same recipient (representative flowplots shown in Fig. 1B). Both anti-hGPA (10F7 and 6A7)induced rapid clearance of hGPA ¥ HOD RBCs, withapproximately 80% of the RBCs becoming undetectable inperipheral blood by 2 hours posttransfusion; only smallamounts of additional clearance were observed afterthe 2-hour time point (Fig. 1C). Anti-Fy3 induced a similarfinal degree of clearance, but with slower kinetics(Fig. 1C). Together, these data define a model ofcrossmatch-incompatible transfusion using two distinctantigens on the same donor RBCs and three monoclonalantibodies recognizing different epitopes.

Role of complement and FcgRs in clearance ofincompatible RBCsUpon activation of the complement pathway by antibodybinding, C3 becomes covalently attached to both theinciting antibodies and the surrounding cellular proteinsvia a thioester bond.30 In this process, C3 is converted toC3b and iC3b, which are ligands for complement recep-

NOVEL HEMOLYTIC TRANSFUSION REACTION


tors CR1, CR3, and CR4, which then promote phagocyto-sis.8 In addition, C3 is generally required for thecomplement cascade to proceed to complete MAC assem-bly.31 To assess the role of complement in clearance ofincompatible RBCs, we utilized recipients with a deletionof the C3 gene (C3 KO mice). Antibody-mediated clear-ance of transfused hGPA ¥ HOD RBCs was unaltered in C3KO mice, compared to wild-type mice for 10F7 (Fig. 2A),6A7 (Fig. 2B), and anti-Fy3 (Fig. 2C). The phenotype of theC3 KO mice was confirmed via a complement fixationassay; normal C3 was present in C57BL/6 mice and wasundetectable in C3 KO animals (data not shown). These

data demonstrate that C3 is not required for clearance ofincompatible RBCs in this model.

To examine the role of FcgRs in clearance of incom-patible RBCs, we used mice with a deletion of thecommon g chain, which is required for expression andfunction of the three murine FcgRs known to participate inphagocytosis (FcgRI, FcgRIII, and FcgRIV).32,33 A small, butsignificant (p < 0.001), decrease in clearance was observedin FcgR KO mice passively immunized with 10F7 or 6A7compared to wild-type mice (Figs. 2D and 2E). In contrast,clearance of hGPA ¥ HOD RBCs induced by anti-Fy3 wasabrogated in FcgR KO mice (Fig. 2F). These data indicate

Fig. 1. Analysis of RBC clearance during crossmatch-incompatible transfusions. (A) Schematic of the hGPA ¥ HOD RBCs used for

incompatible transfusion studies. Histograms represent staining with anti-hGPA (10F7 or 6A7) or anti-Fy3 (MIMA29) MoAbs of DiI-

labeled hGPA ¥ HOD RBCs (black line) or DiO-labeled control wild-type FVB RBCs (shaded gray) after circulation in PBS treated

C57BL/6 mice. (B) Representative dot plots of RBC survival at the 2-hour time point in mice passively immunized with anti-hGPA

or anti-Fy3 or controls treated with PBS alone. (C) Wild-type mice were passively immunized with either anti-hGPA 10F7 ( ), 6A7

(�), or anti-Fy3 ( ) MoAbs; control mice received PBS alone ( ). All recipients were then transfused with a mixture of DiI-labeled

hGPA ¥ HOD RBCs and DiO-labeled wild-type FVB RBCs. Percent survival of hGPA ¥ HOD RBCs in passively immunized animals

was normalized to hGPA ¥ HOD RBC survival in PBS-treated animals. Graph includes combined data from six independent experi-

ments, each with three mice per group.

LIEPKALNS ET AL.


that FcgRs are required for clearance of incompatibleRBCs by anti-Fy3, but have only a small effect on clearanceby anti-hGPA.

The above data demonstrate that neither C3 norFcgRs are individually required for RBC clearance by anti-hGPA; however, this does not address the possibility thatC3 and FcgRs represent redundant pathways (i.e., twopathways simultaneously involved in the same biology).To test for this redundancy, we used MoAb 2.4G2, whichhas been reported to block both FcgRII and FcgRIII bybinding a conserved epitope.25,34,35 In wild-type recipients,injection of purified 2.4G2 had no appreciable effect onclearance of hGPA ¥ HOD RBCs by 10F7; however, a subtlebut significant difference was observed with 6A7-inducedclearance (Fig. 3A). As predicted, based on results withFcgR KO mice, 2.4G2 significantly inhibited clearance ofhGPA ¥ HOD RBCs by anti-Fy3 at 2 hours posttransfusion(Fig. 3A). This blocking effect was progressively lost overtime (see below), which may result from incompleteblockage and/or gradual clearance of the 2.4G2 antibody.

Nevertheless, these data confirm that 2.4G2 functionallyinhibits clearance of RBCs by an FcgR-dependent anti-body (anti-Fy3) at early time points; therefore, it is a usefultool for assessing the FcgR dependence of RBC clearancein this model.

To assess the possibility that C3 and FcgRs areredundant pathways in anti-hGPA–mediated clearance,C3 KO mice were treated with 2.4G2 to inhibit both path-ways simultaneously. These mice were then passivelyimmunized with anti-hGPA (10F7 or 6A7) or anti-Fy3 andsubsequently transfused. Antibody 2.4G2 had a smalleffect on clearance by 6A7 (similar to the FcgR KO recipi-ents passively immunized with 6A7) and no effect onclearance by 10F7 (Figs. 3B and 3C). Similar to wild-typerecipients, 2.4G2 inhibited clearance by anti-Fy3 at earlytime points, but this effect diminished over 48 hours(Fig. 3D). Based on these data, we reject the hypothesisthat C3 and FcgRs represent redundant pathways in thismodel and conclude that neither C3 nor FcgRs arerequired for clearance by these anti-hGPA.

Fig. 2. RBC clearance in C3 and FcgR KO mice. RBC survival was compared in wild-type C57BL/6 mice versus C3KO mice (A-C) or

versus FcgR KO mice (D-F). Before transfusion, mice received either a control PBS injection ( , �) or were passively immunized

with the indicated antibodies ( , �). In all cases, KO mice are indicated by open symbols and dashed line, whereas C57BL/6 mice

are indicated with closed symbols and a solid line. All recipients were transfused with a mixture of DiI-labeled hGPA ¥ HOD RBCs

and DiO-labeled wild-type FVB RBCs. Samples were collected 2 hours, 20 to 24 hours, and 2 days posttransfusion and analyzed via

flow cytometry. Graphs include combined data from three independent experiments, each with three mice per group.



Phagocytic cells are not required for earlyclearance of hGPA and HOD RBCs but play a rolein preventing return to circulation

Phagocytes are known to use multiple scavenger receptors(in addition to Fc and complement receptors) to removedamaged cells.36 Therefore, we tested the hypothesis thatphagocytes are required for IgG-coated RBC clearance bytreating C57BL/6 recipients with clodronate (a toxic elec-tron transport chain decoupling biphosphonate). Clod-ronate was targeted to phagocytes by IV injection ofclodronate encapsulated in liposomes, which are selec-tively consumed by phagocytic cells.37,38 Control micewere injected with liposomes of the same compositionbut without clodronate (i.e., empty or “plain” liposomes).

Mice were treated with IV liposomes 24 hours beforepassive immunization and transfused with a mixture oftransgenic (hGPA or HOD) and control wild-type RBCs. Inrecipients immunized with anti-hGPA (10F7 or 6A7), hGPARBCs were cleared to the same extent within clodronateand plain liposome-treated animals at the first time point(2 hours posttransfusion; Figs. 4A and 4B, respectively).However, starting at 18 hours, most incompatible hGPARBCs returned to circulation, leading to more than 50%survival of incompatible RBCs at 2 days posttransfusion(Figs. 4A and 4B). In mice passively immunized with anti-Fy3, liposomal clodronate infusion prevented initial HODRBC clearance, confirming that clodronate was effectiveand resulted in sufficient phagocyte depletion to preventFcgR-dependent RBC clearance.

Fig. 3. RBC clearance in C57BL/6 and C3 KO mice treated with an FcgR-blocking antibody. Mice received control PBS or an anti-

FcgR-blocking antibody (2.4G2) before passive immunization with anti-hGPA (10F7 or 6A7) or anti-Fy3 (MIMA29) ( , �) or infusion

of PBS alone ( , �) and transfusion of hGPA ¥ HOD and wild-type FVB RBCs labeled in different colors (as previously described).

(A) 2-hour posttransfusion survival of hGPA ¥ HOD RBCs in wild-type C57BL/6 recipient with (black solid) or without (dashed

gray) 2.4G2 anti-Fcg receptor antibody and indicated subsequent treatments. (B-D) Kinetics of RBC survival were compared in C3

KO mice receiving 2.4G2 (open symbols, dashed lines) or PBS (closed symbols, solid lines) and the indicated subsequent treat-

ments. hGPA ¥ HOD RBC survival was calculated as a function of wild-type FVB RBC survival. Samples were collected 2 hours, 20 to

24 hours, and 2 days posttransfusion and analyzed via flow cytometry. Graphs include combined data from three independent

experiments, each with three mice per group.

LIEPKALNS ET AL.


Fig. 4. The role of the spleen and phagocytes on RBC clearance and reappearance. (A-C) C57BL/6 mice were treated with either

clodronate-filled liposomes (�, �, dashed lines) or plain liposomes ( , , solid lines). Recipients were either treated with PBS alone

( , �) or passively immunized with the indicated antibody ( , �) before a transfusion of a mixture of DiI-labeled hGPA RBCs or

HOD RBCs and DiO-labeled wild-type FVB RBCs. (D-F) C57BL/6 were either splenectomized (�, �, dashed line) or underwent sham

surgery ( , , solid line). Recipients were either treated with PBS alone ( , �) or passively immunized with the indicated antibody

( , �) before a transfusion of a mixture of DiI-labeled hGPA ¥ HOD RBCs and DiO-labeled wild-type FVB RBCs. hGPA ¥ HOD, hGPA,

and HOD RBC survivals were calculated as a function of wild-type FVB RBC survival. The same plots for PBS alone are represented

in all three panels to allow comparison. Data are presented separately for each antibody: (A, D) 10F7; (B, E) 6A7; and (C, F) anti-Fy3.

Samples were collected at 2 hours, 20 to 24 hours, and 2 days posttransfusion and analyzed via flow cytometry. Graphs include com-

bined data from three independent experiments, each with three mice per group. (G) Circulating RBCs in the indicated groups and

time points were stained with fluorescently labeled anti-mouse immunoglobulins. (H) Circulating RBCs in the indicated groups and

time points were stained with indicated antibodies followed by fluorescently labeled anti-mouse immunoglobulins.



Histologic analysis of cross-sections of spleens fromclodronate-treated animals demonstrated a large portionof splenic architecture had been compromised (data notshown). We therefore could not distinguish between theinterpretations that the increased survival of incompatiblehGPA RBCs was due to phagocyte depletion or to reducedsplenic blood filtration. Therefore, we tested the hypoth-esis that the spleen plays a central role in incompatiblehGPA ¥ HOD RBC clearance. To this end, recipient micewere splenectomized before passive immunization andtransfusion with hGPA ¥ HOD and wild-type FVB RBCs.Control mice underwent sham surgery. IncompatibleRBCs were cleared with a similar magnitude and kineticsin both sham-operated and splenectomized mice whenpassively immunized with 10F7 (Fig. 4D), 6A7 (Fig. 4E), oranti-Fy3 (Fig. 4F). These data do not rule out that thespleen is involved; however, these findings do reject thehypothesis that a spleen is required for either anti-hGPA–or anti-Fy3–mediated clearance of incompatible RBCs.Additional sites of clearance may include the liver or othercompartments in which phagocytes reside.

To investigate the persistence of antibody binding toincompatible RBCs that continue to circulate, peripheralblood was stained with an anti-mouse globulin reagent(direct antiglobulin test [DAT]) and analyzed by flowcytometry. In both 10F7- and 6A7-treated animals, DATlevels decreased over time in mice treated with clod-ronate. In contrast, antibody binding was stable in micetreated with plain liposomes (Fig. 4G). Antigen persis-tence was determined by staining RBCs with 10F7 or 6A7,followed by anti-mouse globulin. In contrast to DATs,there was only a subtle decrease in antigen levels over time(Fig. 4H). It is important to note that in the presence ofclodronate, the majority of RBCs at late time points havebeen sequestered and then returned to circulation. In thiscontext, these data suggest that the RBCs that return tocirculation have done so in part by breaking free of thebound antibodies, without lysing or losing their surfaceantigen. Also of interest, the few RBCs that survive in pas-sively immunized control animals are not escaping clear-ance due to low levels of antibody binding, but due tosome other property. Together, these data support a sce-nario in which phagocytes are not required for initial RBCclearance, but play an essential role in removing seques-tered RBCs, which otherwise return to circulation in theabsence of sufficient phagocytic activity.

Anti-hGPA induces aggregation of transfusedincompatible RBCs in vivoTo test the hypothesis that aggregation of RBCs in vivo isresponsible for anti-hGPA–mediated clearance, incompat-ible transfusions were performed and peripheral bloodwas analyzed during the rapid clearance phase (i.e., at timepoints before 2 hours posttransfusion).To characterize and

enumerate agglutination, differential size analysis was per-formed by flow cytometric light scatter analysis of post-transfusion samples. Scatter plots identified a populationwith increased size and complexity, which was only presentunder conditions of hGPA ¥ HOD RBC transfusion intomice passively immunized with anti-hGPA, but was absentin both PBS-treated controls and anti-Fy3–injected animals(Fig. 5A). Gating on the large complexes by forward andside scatter and then back reflection on the entire popula-tion (side scatter by DiI staining) demonstrated a substan-tial increase in the DiI signal (Figs. 5B and 5C) but not theDiO signal (data not shown).This finding indicated that thelarge aggregates were composed of hGPA ¥ HOD RBCs, butnot control RBCs. These data are consistent with the inter-pretation that the large complexes constitute selectiveagglutination in vivo of incompatible RBCs. These largercomplexes were also enumerated over time and correlatedwith clearance; the complexes were detectable early in thereaction, rapidly decreased during the clearance phase,and ceased to be detectable soon after transfusion(Fig. 5D). The large complexes were not an artifact of spon-taneous aggregation, as no significant complexes wereobserved in either control mice receiving PBS (i.e., compat-ible transfusion) or in the clearance of RBCs by anti-Fy3

(i.e., an FcgR-dependent HTR; Fig. 5E).To visualize incompatible RBCs directly during clear-

ance, the same early time point specimens analyzed abovewere smeared on microscope slides and analyzed by con-focal microscopy using filters that selectively visualize theemission spectra of DiI (hGPA ¥ HOD RBCs) and DiO(wild-type FVB RBCs). All slides were interpreted byviewers blinded to sample identity. DiI+ RBC clusters werevisible in recipients that had been passively immunizedwith anti-hGPA (either 10F7 or 6A7), but not in unimmu-nized controls or animals passively immunized with anti-Fy3 (Fig. 5F). In contrast, no clusters of DiO+ RBCs (wildtype) were seen in anti-hGPA, anti-Fy3, or control unim-munized mice (Fig. 5F). Together, these data show that thepresence of DiI-labeled hGPA ¥ HOD RBC clusters (byflow cytometry or microscopy) correlated with the clear-ance kinetics of hGPA ¥ HOD RBCs by anti-hGPA (10F7 or6A7) but not with their anti-Fy3–treated counterparts.

Increased inflammatory cytokine secretion inC57BL/6 mice, but not in FcgR KO animals duringincompatible transfusion with hGPA or HOD RBCsPhagocytes not only remove pathogens from circulationbut also secrete inflammatory cytokines, which can betriggered by FcgR cross-linking by antibody-coatedRBCs.39,40 We previously reported that incompatible trans-fusion of hGPA RBCs into mice passively immunized with10F7 or 6A7 results in a cytokine burst not seen duringcompatible transfusion.40 The data above indicate thatFcgRs are not required for clearance of hGPA RBCs. To test

LIEPKALNS ET AL.


Fig. 5. Anti-hGPA induces agglutination of circulating incompatible RBCs in vivo. (A) Representative ungated dot plots are shown of

samples collected from mice treated with PBS alone, anti-Fy3, or anti-hGPA (10F7 or 6A7) immediately after transfusion and were

analyzed by flow cytometry. Gates are drawn on larger populations of events only seen in the presence of 10F7 or 6A7. (B, C) Gated

large events are back-reflected on the total population, indicating that large events are predominantly composed of DiI + RBCs. (D)

Percent survival of hGPA ¥ HOD RBCs, normalized to hGPA ¥ HOD RBC survival in PBS-treated animals. (E) Percentage of large

complexes during early time points (immediately, 1 min, 10 min, and 2 hr) posttransfusion in animals treated with the indicated

antibodies based upon gates shown in (A). (D, E) = PBS alone; = anti-hGPA (10F7); � = anti-hGPA (6A7); = anti-Fy3 (MIMA29).

(F) Peripheral blood smears were produced from samples collected immediately after transfusion and were analyzed by confocal

microscopy visualizing DiI-labeled hGPA ¥ HOD RBCs (red) and DiO-labeled wild-type FVB RBCs (green). Representative confocal

images and flow cytometric plots are shown. Confocal microscopy fields were chosen by an observer blinded to the identity of the

specimens. The experiment was repeated three times with similar results. Graphs include combined data from three independent

experiments, each with three mice per group.



the role of FcgRs in cytokine burst, serum cytokines weremeasured in C57BL/6 and FcgR KO mice 2 hours afterincompatible transfusion, during which a robust responsehas been shown to occur in C57BL/6 mice.40 Clearance at2 hours posttransfusion was similar to that previouslyobserved (data not shown). Sera from C57BL/6 miceundergoing hGPA incompatible transfusion (i.e., passivelyimmunized with 10F7 or 6A7) had significantly elevatedlevels of inflammatory cytokines (IL-6, TNF-a, monocytechemoattractant protein-1, and keratinocyte-derivedchemokine) compared to compatible transfusion (miceinfused with PBS; Fig. 6A). Likewise, incompatible trans-fusions of HOD RBCs induced a cytokine burst in C57BL/6mice (Fig. 6B). In contrast to wild-type animals, no signifi-cant difference in cytokines was seen in FcgR KO micereceiving incompatible, compared to compatible, transfu-sions, in either the hGPA or the HOD systems (Figs. 6A and6B). IL-10 levels were also assessed; however, no changes

were observed between any of the groups (data notshown). To assess if the targeted deletion of FcgRs resultedin a phenotype incapable of cytokine burst, FcgR micewere transfused with phenylhydrazine-treated RBCs,which we previously reported induces cytokine storm inwild-type mice.28 A cytokine burst was observed in theFcgR KO mice transfused with phenylhydrazine-treatedRBCs (although it was substantially lower in magnitudecompared to wild-type mice). Thus, although blunted,FcgR KO mice are capable of releasing a cytokine burst inresponse to damaged RBCs, indicating that the lack ofcytokine burst in response to incompatible transfusionwas not due to an inability to release cytokines. Takentogether, these results indicate that FcgRs are required forcytokine burst 2 hours posttransfusion with incompatiblehGPA RBCs, even though FcgRs are not required for eitherrapid or sustained clearance of RBCs expressing the hGPAantigen.

Fig. 6. HTR-induced inflammatory cytokine secretion is abrogated in FcgR KO mice compared to C57BL/6 mice. Mice were

passively immunized with either 10F7 or 6A7 (A) or anti-Fy3 (B) before a transfusion of a mixture of DiI-labeled hGPA RBCs and

DiO-labeled control FVB RBCs. Clearance patterns were determined to confirm proper functioning of antibodies (data not

shown). Additional control mice were transfused with RBCs treated with phenylhydrazine before transfusion. All recipients were

exsanguinated 2 hours posttransfusion; sera were collected. Cytokines were quantified via flow cytometry. Experiment was per-

formed three times, each with three mice per group. KC = keratinocyte-derived chemokine; MCP = monocyte chemoattractant

protein.

LIEPKALNS ET AL.


DISCUSSION

RBC transfusions are typically performed only with bloodthat is compatible for clinically significant antibodies.However, when mistransfusions occur, or when no otheroption is available, incompatible RBCs may be infused,which puts the patient at risk of experiencing an HTR.Moreover, antibody-bound RBCs can be cleared from cir-culation during an amnestic response. Currently, there areno evidence-based therapeutic interventions available toprevent the clinical consequences of HTRs after incom-patible transfusion. Understanding the mechanismsunderlying HTRs is essential for providing a rationale forsuch interventions.

Canonically, mechanisms explaining IgG-mediatedHTRs are divided into two types, involving either comple-ment or FcgRs. In contrast, this study reveals that forhGPA-encoded antigens, clearance of incompatible RBCsrequires neither C3 nor FcgRs. Moreover, as the simulta-neous absence of both C3 and FcgR function has no effecton clearance, it is unlikely that C3 and FcgRs representredundant pathways. Rather, these data suggest that, inthis setting, C3 and FcgRs are neither required norinvolved. It is worth noting that 2.4G2 does not blockFcgRIV function; thus, we cannot unequivocally ruleout redundancy between C3 and FcgRIV; nonetheless,because FcgR KO mice lack functional FcgRI, FcgRIII, andFcgRIV, we can conclude that FcgRIV is not required forclearance. It is also worth noting that the lack of require-ment for C3 does not unequivocally reject a role forcomplement in general, because C5 can be activated bythrombin in the absence of C3.41 Thus, it remains theoreti-cally possible that the MAC could still form in some cases;however, clearance of hGPA RBCs by 10F7 and 6A7 occursnormally in DBA.2 (data not shown) and FVB mice,26 bothof which lack C5. Perhaps more importantly, RBCs lysedby the MAC would not be able to return to circulation inthe absence of clodronate-sensitive phagocytes, as MACassembly results in RBC lysis.

A mechanism of sequestration has previously beenproposed involving complement receptor binding toRBC-bound C3b, which is cleaved and the RBC is subse-quently released back into circulation. Such appears notto be the case in this study due to the results with C3KOmice; however, it is important to note that C4b can alsobind CR1. As C4b precedes C3 in the complementcascade, we cannot rule out a role for C4b in theobserved biology.

Taken together, the data presented herein allow theconclusion that hGPA RBCs are cleared by IgG alloanti-bodies using a mechanism(s) independent of C3 or FcgRs.The existence of such pathways for IgG-mediated HTRshas been postulated by others,12,14,15,17,42 but to our knowl-edge has not been demonstrated in vivo. Agglutination invivo was observed in a murine model of autoimmune

hemolytic anemia induced by IgG MoAbs, although everycirculating RBC was antigen positive in this case, in con-trast to incompatible transfusions, where antigen-positiveRBCs are more dispersed.17 In this model, the visualizationof aggregate formation by flow cytometry and confocalmicroscopy suggests that agglutination in vivo is a likelymechanism of initial clearance. Aggregation is not aninevitable outcome of murine HTR models, becauseagglutination did not correlate with incompatible RBCclearance by an anti-Fy3. Further study of different bloodgroup systems will be required to determine if this mecha-nism is found in additional settings or is restricted to HTRsin the hGPA system. The distinct mechanism of clearanceof hGPA or HOD RBCs (by 10F7, 6A7, anti-Fy3) may be dueto differences in the number of copies expressed, the IgGsubtype used during passive immunization, or the natureof the membrane protein and/or protein associations.

It is unclear to what extent these findings predictmechanisms of HTRs in humans. Sequestration of incom-patible cells outside of the circulation has been observedin human studies of transfused radioactively labeledRBCs43-45 including a release of a small number of seques-tered incompatible RBCs.43 However, since the humanstudies were performed using radioactivity it is difficult todistinguish between sequestration of intact RBCs versusconsumption of radiolabeled RBCs. It is likewise difficultto distinguish between release of previous sequesteredRBCs versus circulation of phagocytes that have con-sumed radiolabeled RBCs. Use of flow cytometry in thecurrent studies avoids these ambiguities.

Several important characteristics can be concludedfrom the current studies. First, as splenectomy does notaffect clearance, removal of the incompatible RBCs canoccur in extrasplenic sites. The spleen is therefore notrequired but may be involved in clearance of incompatibleRBCs. Second, although it is not clear whether sensitivityto clodronate varies with different organs, clodronatetreatment results in a return of incompatible RBCs to thecirculation over time. This observation suggests thatphagocytes are required to prevent initially cleared RBCsfrom reentering the vasculature, presumably by ingestingthe antibody-bound RBCs. However, as no return to thecirculation is seen in C3 KO or FcgR KO mice, the signal foringestion appears to be something other than opsoniza-tion with complement or IgG. The identity of this addi-tional signal is unclear from the current data; however,Brain and colleagues15 showed that binding of polyclonalhuman IgG antibodies (from a patient with an HTR due toanti-Pr) can directly damage RBCs, including membranedistortions, opening of Ca2+ channels, and exposure ofphosphatidylethanolamine. Thus, it is possible that 10F7and 6A7 induce expression of an “eat me signal,” such asphosphatidylserine or others, which is then recognized byscavenger receptors on clodronate-sensitive phagocytes.It seems unlikely that 10F7 or 6A7 induce direct eryptosis



of most cleared incompatible RBCs, as eryptotic RBCswould not reenter circulation.

The current data differ somewhat from a previouscharacterization of clearance of hGPA RBCs by the 10F7antibody.26 Both the current and the previous reportsshowed some effect of FcgR function on clearance;however, in the previous report FcgR played a significantlylarger role in clearance. More importantly, the previousreport indicated that clearance of hGPA RBCs by 10F7 wasdecreased in C3 KO mice. It is also worth noting that inthe previous article, 20% clearance still occurred in(C3 ¥ FcgR) double KO mice. Thus, in agreement with thecurrent findings, the previous study showed that someclearance was independent of either C3 or FcgRs; however,in the current studies, the vast majority of clearance wasindependent of either C3 or FcgRs. As neither flow cytom-etry nor fluorescence microscopy were used in the previ-ous study, it is unclear if aggregation was occurring inthat setting. Although it is unclear why clearance wasdecreased in C3 KO mice in the previous study, but not in

the current one, several factors may be relevant. First, theprevious report modeled intraoperative transfusion inmice that were anesthetized and transfused through a sur-gically exposed jugular vein; in contrast, the current studyused tail vein infusion in unanesthetized mice. Inflamma-tion can modify levels of circulating complement, FcgRfunction, and phagocyte function; in addition, inflamma-tion can increase antibody-mediated RBC clearance.46

Thus, surgery-induced inflammation may have altered thebiology of clearance in the previous report.26 Second, RBCclearance was measured using 51Cr labeling and samplingat a single posttransfusion time point; the current reportuses two-color fluorescent labeling and measurements byflow cytometry over 48 hours. Although the 51Cr labelingmethod is analogous to the approach used in humanstudies, there is no control RBC population (e.g., such asthe DiO-labeled RBCs in this report) to control for differ-ences in infusion volume or bleeding. Moreover, therewere methodologic differences in antibody purification,evaluation of purity, and administration. Thus, the differ-

Fig. 7. Schematic of proposed mechanism of biphasic HTR involving neither C3 nor Fc domains of bound antibodies binding to

FcgRs.

LIEPKALNS ET AL.


ences seen in the current system and the previous reportmay be due to methodologic variation. Nonetheless, thecurrent report demonstrates that conditions exist underwhich anti-hGPA–mediated HTRs can proceed indepen-dent of both C3 and FcgR function.

Overall, our data are consistent with a model thatentails a two-phased clearing process. During the firstphase, incompatible hGPA RBCs are rapidly sequesteredoutside the circulation after transfusion of incompatibleRBCs. During the second phase, which occurs after 2hours but before 24 hours, the “cleared” incompatibleRBCs are ingested, thus making their clearance perma-nent (see Fig. 7 for diagram). During the first phase ofclearance, there is also a burst of serum cytokines.Although FcgRs are not required for hGPA RBC clearance,they are required for cytokine release, indicating that RBC-bound IgG antibodies are ligating surface FcgRs and sig-naling into the cytokine-secreting cells. Thus, when asequestered IgG-coated RBC encounters a phagocyte, twoseparate signaling events occur: 1) FcgRs are ligated result-ing in cytokine burst and 2) another signal (not throughFcgR or C3R) induces phagocytosis.

Several therapeutics, targeting complement or FcgRs,have been proposed to inhibit RBC clearance. Althoughcomplement fixation on incompatible RBCs is well estab-lished in human transfusion biology,47 and althoughFcgRs polymorphisms alter the efficiency of clearance ofRBCs opsonized with anti-D in humans,48 the causal roleof complement and FcgRs has not been rigorously testedin the human setting of incompatible transfusions andfor various blood group antigens. Accordingly, to theextent that HTRs occur in humans independent of C3 orFcgRs, such therapies would likely fail to prevent removalof incompatible RBCs. Importantly, failure to preventRBC clearance may not indicate complete lack of efficacy.For example, in the current system, although an FcgR-blocking agent would not prevent RBC clearance, it coulddecrease morbidity and/or mortality due to cytokinestorm. Such an approach may have significant benefit inthe context of autoimmune hemolytic anemia, eventhough it may not affect hematocrit. Thus, prevention ofRBC clearance alone may be an overly simplified end-point in drug trials in these settings. Accordingly,although C3 and FcgRs likely play important roles inmany human HTRs, studies of alternative pathways arerequired to define the biology of HTRs completely; thiswill allow development of appropriate therapeutic inter-ventions for HTRs, which may all have similar clinicalpresentations, but have distinct cellular and molecularmechanisms.

ACKNOWLEDGMENTS

We thank Nicole Smith for her technical assistance. We also thank

Kathryn Girard and Krystal Hudson for helping to blind micros-

copy specimens and Chris Gilson for assisting in splenectomy of

mice.

CONFLICT OF INTEREST

The authors have no conflicts of interest to declare.

REFERENCES

1. Zimring JC, Spitalnik SL. Alloimmunization to red cell anti-

gens and management of alloimmunized patients. In:

Mintz PD, editor. Transfusion therapy clinical principles

and practice. Bethesda (MD): American Association of

Blood Banks; 2011. p. 631-42.

2. Davenport RD. Management of transfusion reactions. In:

Mintz PD, editor. Transfusion therapy clinical principles

and practice. Bethesda (MD): American Association of

Blood Banks; 2011. p. 757-84.

3. Vamvakas EC, Blajchman MA. Transfusion-related

mortality: the ongoing risks of allogeneic blood transfusion

and the available strategies for their prevention. Blood

2009;113:3406-17.

4. Linden JV, Paul B, Dressler KP. A report of 104 transfusion

errors in New York State. Transfusion 1992;32:601-6.

5. McClelland DB, Phillips P. Errors in blood transfusion in

Britain: survey of hospital haematology departments. BMJ

1994;308:1205-6.

6. Bowman J. The management of hemolytic disease in the

fetus and newborn. Semin Perinatol 1997;21:39-44.

7. Gehrs BC, Friedberg RC. Autoimmune hemolytic anemia.

Am J Hematol 2002;69:258-71.

8. Engelfriet CP. The immune destruction of red cells. Trans-

fus Med 1992;2:1-6.

9. Schreiber AD, Frank MM. Role of antibody and comple-

ment in the immune clearance and destruction of erythro-

cytes. I. In vivo effects of IgG and IgM complement-fixing

sites. J Clin Invest 1972;51:575-82.

10. Schreiber AD, Frank MM. Role of antibody and comple-

ment in the immune clearance and destruction of erythro-

cytes. II. Molecular nature of IgG and IgM complement-

fixing sites and effects of their interaction with serum.

J Clin Invest 1972;51:583-9.

11. Petz LD, Garratty G. Immune hemolytic anemias. 2nd ed.

Philadelphia (PA): Churchill Livingstone; 2004.

12. Attanasio P, Shumilina E, Hermle T, Kiedaisch V, Lang PA,

Huber SM, Wieder T, Lang F. Stimulation of eryptosis by

anti-A IgG antibodies. Cell Physiol Biochem 2007;20:591-

600.

13. Chadebech P, Habibi A, Nzouakou R, Bachir D, Meunier-

Costes N, Bonin P, Rodet M, Chami B, Galacteros F, Bier-

ling P, Noizat-Pirenne F. Delayed hemolytic transfusion

reaction in sickle cell disease patients: evidence of an

emerging syndrome with suicidal red blood cell death.

Transfusion 2009;49:1785-92.

14. Brain MC, Pihl C, Robertson L, Brown CB. Evidence for a



mechanosensitive calcium influx into red cells. Blood Cells

Mol Dis 2004;32:349-52.

15. Brain M, Ruether B, Valentine K, Brown C, ter Keurs H. Life-

threatening hemolytic anemia due to an autoanti-Pr cold

agglutinin: evidence that glycophorin A antibodies may

induce lipid bilayer exposure and cation permeability inde-

pendent of agglutination. Transfusion 2010;50:292-301.

16. Chadebech P, Michel M, Janvier D, Yamada K, Copie-

Bergman C, Bodivit G, Bensussan A, Fournie JJ, Godeau B,

Bierling P, Izui S, Noizat-Pirenne F. IgA-mediated human

autoimmune hemolytic anemia as a result of hemaggluti-

nation in the spleen, but independent of complement acti-

vation and Fc alpha RI. Blood 2010;116:4141-7.

17. Shibata T, Berney T, Reininger L, Chicheportiche Y, Ozaki

S, Shirai T, Izui S. Monoclonal anti-erythrocyte autoanti-

bodies derived from NZB mice cause autoimmune

hemolytic anemia by two distinct pathogenic mechanisms.

Int Immunol 1990;2:1133-41.

18. Bigbee WL, Langlois RG, Vanderlaan M, Jensen RH.

Binding specificities of eight monoclonal antibodies to

human glycophorin A—studies with McM, and MkEn(UK)

variant human erythrocytes and M- and MNV-type chim-

panzee erythrocytes. J Immunol 1984;133:3149-55.

19. Bigbee WL, Vanderlaan M, Fong SS, Jensen RH. Mono-

clonal antibodies specific for the M- and N-forms of

human glycophorin A. Mol Immunol 1983;20:1353-62.

20. Blackall DP, Ugorski M, Pahlsson P, Shakin-Eshleman SH,

Spitalnik SL. A molecular biologic approach to study the

fine specificity of antibodies directed to the MN human

blood group antigens. J Immunol 1994;152:2241-7.

21. Taketo M, Schroeder AC, Mobraaten LE, Gunning KB,

Hanten G, Fox RR, Roderick TH, Stewart CL, Lilly F,

Hansen CT, Overbeek PA. FVB/N: an inbred mouse strain

preferable for transgenic analyses. Proc Natl Acad Sci U S A

1991;88:2065-9.

22. Auffray I, Marfatia S, de Jong K, Lee G, Huang CH, Paszty

C, Tanner MJ, Mohandas N, Chasis JA. Glycophorin A

dimerization and band 3 interaction during erythroid

membrane biogenesis: in vivo studies in human glycoph-

orin A transgenic mice. Blood 2001;97:2872-8.

23. Desmarets M, Cadwell CM, Peterson KR, Neades R,

Zimring JC. Minor histocompatibility antigens on trans-

fused leukoreduced units of red blood cells induce bone

marrow transplant rejection in a mouse model. Blood

2009;114:2315-22.

24. Wasniowska K, Lisowska E, Halverson GR, Chaudhuri A,

Reid ME. The Fya, Fy6 and Fy3 epitopes of the Duffy blood

group system recognized by new monoclonal antibodies:

identification of a linear Fy3 epitope. Br J Haematol 2004;

124:118-22.

25. Unkeless JC. Characterization of a monoclonal antibody

directed against mouse macrophage and lymphocyte Fc

receptors. J Exp Med 1979;150:580-96.

26. Schirmer DA, Song SC, Baliff JP, Harbers SO, Clynes RA,

Krop-Watorek A, Halverson GR, Czerwinski M, Spitalnik SL.

Mouse models of IgG- and IgM-mediated hemolysis. Blood

2007;109:3099-107.

27. Gilson CR, Zimring JC. Alloimmunization to transfused

platelets requires priming of CD4+ T cells in the splenic

microenvironment in a murine model. Transfusion 2012;

52:849-59.

28. Hendrickson JE, Hod EA, Cadwell CM, Eisenbarth SC,

Spiegel DA, Tormey CA, Spitalnik SL, Zimring JC. Rapid

clearance of transfused murine red blood cells is associ-

ated with recipient cytokine storm and enhanced alloim-

munogenicity. Transfusion 2011;51:2445-54.

29. Zimring JC, Hair GA, Chadwick TE, Deshpande SS, Ander-

son KM, Hillyer CD, Roback JD. Nonhemolytic antibody-

induced loss of erythrocyte surface antigen. Blood 2005;

106:1105-12.

30. Gros P, Milder FJ, Janssen BJ. Complement driven by con-

formational changes. Nat Rev Immunol 2008;8:48-58.

31. Linton S. Animal models of inherited complement defi-

ciency. Mol Biotechnol 2001;18:135-48.

32. Takai T, Li M, Sylvestre D, Clynes R, Ravetch JV. FcR

gamma chain deletion results in pleiotrophic effector cell

defects. Cell 1994;76:519-29.

33. Nimmerjahn F, Bruhns P, Horiuchi K, Ravetch JV.

FcgammaRIV: a novel FcR with distinct IgG subclass speci-

ficity. Immunity 2005;23:41-51.

34. Kurlander RJ, Ellison DM, Hall J. The blockade of Fc

receptor-mediated clearance of immune complexes in vivo

by a monoclonal antibody (2.4G2) directed against Fc

receptors on murine leukocytes. J Immunol 1984;133:855-

62.

35. Bonnerot C, Daeron M. Biological activities of murine low-

affinity Fc receptors for IgG. Immunomethods 1994;4:41-7.

36. Delemarre FG, Kors N, Kraal G, van Rooijen N. Repopula-

tion of macrophages in popliteal lymph nodes of mice

after liposome-mediated depletion. J Leukoc Biol 1990;47:

251-7.

37. Miller MR, Thompson WR, Casella JF, Spevak PJ. Antibody-

mediated red blood cell agglutination resulting in sponta-

neous echocardiographic contrast. Pediatr Cardiol 1999;20:

287-9.

38. Waterbury L, Parnes H, Katz RS. Fatal warm antibody

autoimmune hemolysis with marked erythrocyte autoag-

glutination and liver necrosis. South Med J 1986;79:646-7.

39. Cutbush M, Mollison PL. Relation between characteristics

of blood-group antibodies in vitro and associated patterns

of red cell destruction in vivo. Br J Haematol 1958;4:115-37.

40. Cutbush M. Observations on the antihuman globulin test.

Bibl Haematol 1958;7:185-9.

41. Huber-Lang M, Sarma JV, Zetoune FS, Rittirsch D, Neff TA,

McGuire SR, Lambris JD, Warner RL, Flierl MA, Hoesel LM,

Gebhard F, Younger JG, Drouin SM, Wetsel RA, Ward PA.

Generation of C5a in the absence of C3: a new comple-

ment activation pathway. Nat Med 2006;12:682-7.

42. Garratty G. A new mechanism for immune destruction of

red blood cells? Transfusion 2010;50:274-7.

LIEPKALNS ET AL.


43. NIH conference. Pathophysiology of immune hemolytic

anemia. Ann Intern Med 1977;87:210-22.

44. Jandl JH, Kaplan ME. The destruction of red cells by anti-

bodies in man. III. Quantitative factors influencing the

patterns of hemolysis in vivo. J Clin Invest 1960;39:1145-56.

45. Jones NC, Mollison PL, Veall N. Removal of incompatible

red cells by the spleen. Br J Haematol 1957;3:125-33.

46. Meite M, Leonard S, Idrissi ME, Izui S, Masson PL, Coute-

lier JP. Exacerbation of autoantibody-mediated hemolytic

anemia by viral infection. J Virol 2000;74:6045-9.

47. Hillyer CD, Silberstein LE, Ness PM, Anderson KC, Roback

JD. Blood banking and transfusion medicine. 2nd ed.

Philadelphia (PA): Churchill Livingstone; 2007.

48. Miescher S, Spycher MO, Amstutz H, De Haas M, Kleijer M,

Kalus UJ, Radtke H, Hubsch A, Andresen I, Martin RM,

Bichler J. A single recombinant anti-RhD IgG prevents

RhD immunization: association of RhD-positive red

blood cell clearance rate with polymorphisms in the

FcgammaRIIA and FcgammaIIIA genes. Blood 2004;103:

4028-35.



Justine S. Liepkalns, Eldad A. Hod, Sean R. Stowell...

Documents

Transcript of Justine S. Liepkalns, Eldad A. Hod, Sean R. Stowell...