Experimental Hematology 34 (2006) 1393–1402

Evaluation of liposomal clodronatein experimental spontaneous autoimmune hemolytic anemia in dogs

Mark Mathesa, Michael Jordanb, and Steven Dowa,c

aDepartments of Clinical Sciences; cMicrobiology, Immunology, and Pathology, Colorado State University, Ft. Collins,

Colo., USA; bDepartment of Pediatrics, University of Cincinnati College of Medicine and Cincinnati Children’s Hospital, Cincinnati, Ohio, USA

(Received 3 November 2005; revised 6 April 2006; accepted 17 May 2006)

Objectives. Liposomal clodronate (dichloromethylene diphosphonate) has been used to de-plete macrophages and block clearance of opsonized cells in mouse models of autoimmune dis-ease. However, liposomal clodronate (LC) has not been previously evaluated in a large-animalspontaneous autoimmune disease model. Therefore, the safety and efficacy of LC treatmentwas assessed in normal dogs and in dogs with spontaneous autoimmune hemolytic anemia(AIHA).

Methods. LC was administered intravenously first to healthy dogs and then to dogs with spon-taneous, severe AIHA to determine if the treatment was safe and could block clearance ofopsonized red blood cells (RBCs) in vivo. Studies were also conducted to assess the in vitroeffects of LC on dog macrophages and dendritic cells.

Results. Intravenous infusion of low doses of LC was well tolerated and blocked clearance ofopsonized RBCs in normal dogs in vivo. LC was taken up by splenic macrophages and den-dritic cells in vivo, and induced killing of macrophages and dendritic cells in vitro. Seven dogswith severe, spontaneous AIHA were treated with LC in a pilot study. Treatment was well tol-erated, 2 of 7 LC-treated dogs with AIHA had a decrease in RBC clearance, and LC-treateddogs had significantly increased survival times compared to historical control dogs matchedfor disease severity.

Conclusions. These results indicate that LC can be safely administered intravenously to dogs andthat even relatively low doses are capable of blocking RBC clearance and improving outcomes ina spontaneous large-animal model of AIHA. Therefore, additional studies of LC for treatment ofautoantibody-mediated cytopenias in dogs and humans may be warranted. � 2006 Interna-tional Society for Experimental Hematology. Published by Elsevier Inc.

Binding of pathogenic autoantibodies to the surface of redblood cells (RBCs) triggers premature RBC destruction andthe development of immune-mediated hemolytic anemia(AIHA). In some patients with AIHA, the disease can de-velop suddenly and become life-threatening, whereas in otherpatients the onset of disease is more gradual [1–4]. Mostcases of AIHA in humans are idiopathic, though the diseasecan also occur in association with neoplasia and infection, orin association with systemic autoimmune disease such as sys-temic lupus erythematosus (SLE) [5]. The disease can be me-diated by RBC autoantibodies of the IgM or the IgG class,though IgG antibodies are thought to be the most clinicallyimportant autoantibodies [6]. The binding of autoantibodies

Offprint requests to: Steven Dow, D.V.M., Ph.D., Department of Micro-

biology, Immunology, and Pathology, Colorado State University, Ft Col-

lins, CO 80523; E-mail: sdow@colostate.edu

0301-472X/06 $–see front matter. Copyright � 2006 International Society fo

doi: 10.1016/j.exphem.2006.05.014

to the surface of RBC leads to RBC destruction, primarilyvia removal of the opsonized RBC by splenic and hepaticmacrophages via Fc receptors [7,8]. Macrophages thereforeplay a key role in the pathogenesis of AIHA. Conventionaltreatment of AIHA is directed primarily at reducing macro-phage-mediated destruction of opsonized RBC. The two pri-mary forms of treatment are splenectomy or high-dosecorticosteroid therapy [1,3].

Autoimmune hemolytic anemia is a spontaneous diseaseof dogs that shares many features with the disease in hu-mans [9–13] and is arguably the best large-animal modelof this disease. Most cases of AIHA in dogs are idiopathic,though as in humans some cases may also be associatedwith malignancy or other systemic autoimmune diseases[11]. Most dogs with AIHA have RBC autoantibodies ofthe IgG subclass, occasionally occurring together withIgM autoantibodies [14]. Treatment typically involves

r Experimental Hematology. Published by Elsevier Inc.

1394 M. Mathes et al. / Experimental Hematology 34 (2006) 1393–1402

administration of high doses of corticosteroids, togetherwith other immunosuppressive agents such as azathioprine[11,15–17]. The mortality rate associated with AIHA indogs is very high and approaches 50 to 70% in some studies[9–11,13,18,19]. Death may occur due to complications ofmultiple blood transfusions and immunosuppressive medi-cations, and a large number of dogs are euthanized due toinability to control the disease. In addition, the developmentof thromboembolic disease is relatively common in dogswith AIHA [10,12,20,21]. At present, there are no effectivetreatments for quickly controlling the acute RBC destructionthat typically occurs in dogs with AIHA.

Clodronate (dichloromethylene diphosphonate) incorpo-rated into liposomes offers several potential advantages asa novel agent for treatment of AIHA. Numerous studies per-formed in mice and rats over nearly two decades have dem-onstrated the impressive macrophage-depleting propertiesof liposomal clodronate (LC) [22–25]. When clodronate isincorporated into liposomes, the drug is preferentiallyphagocytosed by macrophages and some dendritic cells(DC), leading to rapid apoptosis of the cells that phago-cytose the liposomes [25,26]. Systemic injection of LCcan induce very efficient depletion of splenic and hepaticmacrophages within 24 hours [25]. An additional advantageis that macrophage depletion using LC does not lead toliberation of pro-inflammatory cytokines by the dyingmacrophages [27]. Previous studies have demonstrated theeffectiveness of systemic administration of LC for treatmentof several different autoimmune diseases in mouse models,including immune thrombocytopenia and immune-mediatedarthritis [28–31].

We hypothesized therefore, based on recent studies ina mouse model of AIHA, that LC could be used to blockclearance of opsonized RBC in dogs and as a therapeuticagent for treatment of dogs with spontaneous AIHA [32].Since LC has never been administered systemically in ani-mals other than rodents, we first conducted studies to deter-mine a safe and effective dose of LC in normal, healthydogs. Next, LC was administered to a small series ofdogs with acute, severe AIHA to assess clinical safetyand effectiveness.

We report here that intravenous infusion of LC was welltolerated in normal dogs and induced a dose-dependentblockade of clearance of opsonized autologous RBC.In vitro, LC was found to induce rapid killing of caninesplenic macrophages and dendritic cells. Seven dogs withspontaneous AIHA were treated with LC and the effectson RBC clearance were assessed, as well as the effects onoverall survival time, as compared to a control populationof dogs with AIHA not treated with LC. Clearance ofRBC was decreased in 2 of the 7 dogs following LCtreatment and overall survival of LC-treated dogs wassignificantly (p 5 0.035) increased compared to historicalcontrol animals. These results suggest that LC can be usedsafely in dogs with severe, spontaneous AIHA and may be

a therapeutic method of rapidly suppressing autoantibody-mediated RBC destruction.

Materials and methods

Preparation of clodronate liposomesClodronate liposomes were prepared as previously described[23,25]. Briefly, phosphatidylcholine and cholesterol (Avanti PolarLipids, Alabaster, AL, USA) and p-amino phenyl mannopyrano-side (Sigma-Aldrich, St. Louis, MO, USA) were dissolved in chlo-roform (5:8:6 weight ratio) and placed in a round-bottom tube anddried by lyophilizing overnight. The lipid film was resuspended ina 0.7 M solution of clodronate (dichloromethylene diphosphonate,a generous gift of Roche Diagnostics, Mannheim, Germany), fol-lowed by gentle sonication. The liposomes were allowed to‘‘swell’’ overnight, then were washed 3 times in phosphate-buff-ered saline (PBS) to eliminate unincorporated clodronate, thenstored under argon at 4�C until used. Empty (PBS) liposomeswere prepared under the same conditions using a 10� stock ofPBS to resuspend the lipid film instead of clodronate solution.Prior to use, all liposomes were passed through a 12-um filter toeliminate large aggregates. For preparation of dye-labeled clodro-nate liposomes, the preformed liposomes were incubated with DiD(1,10-dioctadecyl-3,3,30,30-tetramethylindodicarbocyanine, 4 chloro-benzenesulfonate; Molecular Probes, Eugene, OR, USA) for 15minutes at 37�C, then washed in PBS to remove unincorporated dye.

Preparation of rabbit anti-dogRBC antisera and opsonization of dog RBCTo prepare anti-dog RBC antisera, rabbits were immunized 3times with 5 mL washed canine RBCs mixed with a liposomal ad-juvant at 2-week intervals. Serum was collected prior to immuni-zation and at day 60 after immunization and the IgG fraction waspurified using a commercial IgG column (Pierce Biotechnology,Rockford, IL, USA). The anti-canine RBC antiserum was titratedto determine the amount that resulted in maximum RBC binding(as assessed by flow cytometry, data not shown) without inducingRBC agglutination. Canine RBC were opsonized for both in vitroand in vivo experiments using 8.5 ug rabbit IgG per 5 � 109 ca-nine RBCs. After 30 minutes incubation at room temperature,the RBCs were washed three times to remove unbound antibody.

Macrophage and dendritic cell cytotoxicity assaysSplenic macrophages and DC from healthy dogs were obtained byplastic adherence, using spleen tissues from purpose-bred dogs orpet dogs euthanized for other reasons at the Colorado State Uni-versity Veterinary Teaching Hospital. Spleen tissues were teasedapart with 18-gauge needles, then filtered through 70-um strainers(3M, Minneapolis, MN, USA). Mononuclear cells were isolatedby Ficoll density-gradient centrifugation. Mononuclear cellswere then resuspended in complete medium (MEM supplementedwith essential and nonessential amino acids (Invitrogen, SanDiego, CA, USA) and 10% fetal bovine serum (FBS; Hyclone, Lo-gan, UT, USA) at a concentration of 4 � 106 cells/mL. The cellswere then incubated in triplicate wells of 24-well plates for 2hours at 5% CO2 at 37�C, then the nonadherent cells were re-moved by gently washing with PBS. The remaining adherent cellswere then incubated with the indicated concentration of clodronate

1395M. Mathes et al./ Experimental Hematology 34 (2006) 1393–1402

or PBS liposomes diluted in complete tissue culture medium for 4hours.

Following treatment, remaining adherent cells were detachedusing trypsin-EDTA solution (Sigma), then washed in PBS, andthe number of viable cells determined by manual counting and try-pan blue exclusion. The cells were then immunostained with un-labeled mAbs to canine CD11b or canine CD11c (Serotec,Raleigh, NC, USA), followed by secondary PE-conjugated anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA).In other cases, cells were labeled with directly conjugated anti-bodies to canine CD4 or B cells (Serotec). Immunostained cellswere evaluated by flow cytometry, using a Cyan MLE flow cytom-eter (DakoCytomation, Fort Collins, CO, USA). At least 10,000events were collected for each sample and data was analyzedusing Summit Software (DakoCytomation).

Preparation of dye-labeled RBCCanine RBC were prepared from whole blood collected into EDTAby jugular venipuncture or via cephalic catheter. The RBC werewashed twice in PBS. For labeling with CFSE (carboxyfluoresceinsuccimidyl ester; Molecular Probes, Eugene, OR, USA), 109 RBCswere incubated with a 50-uM solution of CFSE in PBS for 15 min-utes, followed by two washes in PBS to remove unincorporateddye. For labeling RBC with DiD (1,10-dioctadecyl-3,3,30,30-tetra-methylindodicarbocyanine, 4 chlorobenzenesulfonate; MolecularProbes) or DiI (1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocya-nine perchlorate; Molecular Probes), the RBC were incubated in50 uM dye diluted in 280 mM sucrose solution. The labeledRBC were washed twice in PBS following dye loading. These pro-cedures resulted in nearly 100% labeling of RBC (data not shown).For dye labeling of opsonized RBCs, the RBC were dye labeledfirst and then opsonized.

Flow cytometric evaluationof clearance kinetics of dye-labeled RBCThe distribution of dye-labeled RBC in dogs was assessed by flowcytometry. For most dogs, a 5-mL aliquot of whole blood was usedfor labeling and in some cases opsonization with anti-RBC anti-sera. This amount of blood was calculated to result in labelingof approximately 1 to 2% of total circulating RBC. The RBCwere washed and labeled as indicated above, then infused intrave-nously via an indwelling cephalic vein catheter. At the indicatedtime points postinfusion, small samples of blood were obtainedfor flow cytometric analysis. RBC were washed twice in PBS,then diluted to a final concentration of 1 � 107 RBC per mL ofFACS buffer (PBS plus 2% FBS plus 0.1% sodium azide) priorto analysis. Dye-labeled RBC were detected in gated RBC popu-lations using appropriate lasers and photomultipler tubes for eachdye. The number of dye-labeled RBC as a percentage of total RBCwas calculated and plotted vs time postinfusion.

Infusion of LC and clinical monitoringAll animal protocols for these studies were approved by the Insti-tutional Animal Care and Use Committee at Colorado State Uni-versity. For the initial safety and toxicity studies, 4 Beagle dogswere purchased from a commercial vendor. Clodronate liposomeswere administered by slow intravenous infusion in nonanesthe-tized dogs. Briefly, the LC solution was infused at a constantrate into an indwelling peripheral intravenous catheter over

a 90-minute period, using an infusion pump (Medfusion 2001pump; Medex, Inc, Carlsbad, CA, USA). During the infusion andfor 24 hours thereafter, animals were monitored at 1- to 6-hour in-tervals for changes in heart rate, respiration, and body temperature.In addition, a complete blood count and serum biochemical panelwas evaluated prior to treatment, at 24 hours after the infusion,and again at 7 days after the clodronate infusion. The dose of LCadministered in the initial safety studies varied from 0.1 mL/kgto 1.0 mL/kg in normal purpose-bred Beagle dogs. For the clinicalpilot trial in dogs with spontaneous AIHA, a dose of 0.5 mL/kg wasselected based on the findings from the studies conducted in normaldogs.

Uptake of dye-labeled LC by spleen cellsLC were labeled with DiD as described above and infused ata dose of 0.5 mL/kg into 2 normal purpose-bred dogs. Four hoursfollowing intravenous infusion, the dogs were euthanized andspleen tissue was collected. A control dog was given an infusionof PBS only. Single cell suspensions of spleen mononuclear cellswere prepared as described above, followed by lysis of RBC usingammonium chloride. The spleen cells were then immunostainedwith antibodies to canine CD11b, CD11c, CD4, and B cells as de-scribed above and then analyzed by multicolor flow cytometry af-ter gating on live cells based on forward- and side-scattercharacteristics. Flow cytometry was used to identify DiDþ cells,using antibodies to CD11b, CD11c, and CD21. The percentageof each cell type that contained DiDþ liposomes was calculated.

LC treatment of dogs with spontaneous AIHAA clinical trial was conducted to assess the effects of LC infusionon RBC clearance kinetics, clinical response parameters, andoverall survival times in 7 dogs with spontaneous AIHA. Thepet dogs that entered the study were drawn from the referral pa-tient population treated at the CSU Veterinary Teaching Hospital.The design and conduct of these studies was approved by the An-imal Care and Use Committee at Colorado State University. Aclinical diagnosis of AIHA was made if dogs had regenerative ane-mia plus the presence of anti-erythrocyte antibodies as detected byflow cytometry, plus at least 3 of 5 of the following hematologicalabnormalities: spherocytosis, polychromasia, hyperbilirubinemia,positive saline autoagglutination test, and positive direct antiglob-ulin test. Exclusion criteria included nonregenerative anemia, re-generative anemia resulting from causes other than primaryAIHA (e.g., neoplasia, renal disease, GI bleeding), lack of detect-able anti-erythrocyte antibodies, the presence of anti-RBC anti-bodies of only the IgM subclass, prior treatment with intravenoushuman or equine immunoglobulin or bovine oxyglobin (BiopureCorporation, Cambridge, MA, USA), or serious concurrent renalor hepatic disease. All dogs enrolled in the clinical trial receivedthe same initial treatment, which included administration of immu-nosuppressive doses of prednisone (2.2 mg/kg/day), azathioprine(2.2 mg/kg/day), together with anticoagulation with heparin (150to 200 U/kg, administered subcutaneously three times daily). Iftransfusions were necessary, dogs were cross-matched prior totransfusion to assure that incompatible transfusions were not ad-ministered and the transfusions were administered prior to infusionof LC and were not repeated until the final RBC clearance studyhad been completed at 72 hours.

Immediately after entry into the study, RBC clearance kineticswere determined in dogs with AIHA. Autologous RBC were

1396 M. Mathes et al. / Experimental Hematology 34 (2006) 1393–1402

labeled with CFSE and infused without ex vivo opsonization.Peripheral blood samples were collected at 1- to 2-hour intervalsover the next 24 hours and analyzed by flow cytometry to deter-mine the percentage of CFSE-labeled RBC. Twenty-four hourslater, LC was infused intravenously at a dose of 0.5 mL/kg by con-tinuous infusion over a 90-minute period. Temperature, pulse, andrespiration were monitored 2 hours for the next 24 hours. At 48hours, RBC clearance kinetics were again assessed, this time usingDiI-labeled autologous RBC. After the completion of the secondRBC clearance study (72 hours), the clinical trial ended anddogs were treated and discharged when their condition warranted.Follow-up evaluations were done at the CSU Veterinary TeachingHospital or by telephone.

For statistical analysis of survival times, a historical controlgroup of dogs with AIHA that were not treated with LC butwere otherwise matched for disease severity and treatment proto-col was identified by searching medical records of 464 dogs withanemia treated at the CSU Veterinary Teaching Hospital over a 12-year period. From this group, 31 dogs were identified that had di-agnosis of severe AIHA and were also treated with prednisone,azathioprine, and heparin.

Statistical analysesComparisons between two treatment groups were done by Stu-dent’s paired t-test. For comparison of multiple treatment groups,ANOVA was used, followed by Tukey-Kramer multiple meanscomparison test. For comparison of survival times, Kaplan-Meieranalysis was done, followed by c2 test. Analyses were done usingGraphPad Prism software. Differences were considered statisti-cally significant for p values less than 0.05.

Results

Intravenous infusion of LC resultsin uptake by splenic macrophages anddendritic cells following intravenous infusion in dogsPrevious studies in mice have examined the distribution ofLC following intravenous administration and found that theprimary sites of uptake were the spleen and liver [33].Therefore, we conducted distribution studies in healthydogs to identify the cells responsible for uptake of clodro-nate liposomes in the spleen. LC was labeled with DiD asdescribed in Methods and 0.5 mg/kg was infused intrave-nously in two healthy purpose-bred dogs. A third dogserved as an untreated control. Four hours after infusion,the dogs were euthanized and spleen tissues were collectedand single cell suspensions were prepared. The spleen cellswere immunostained for identification of CD11bþ andCD11cþ cells and analyzed by flow cytometry (Fig. 1).We observed that the majority of cells containing DiDþ li-posomes were either CD11bþ macrophages (average 46%)or CD11cþ dendritic cells (average 14%) in the spleen oftreated dogs. The remainder of DiDþ liposomes were con-tained in cells that did not stain with available antibodies(data not shown). Thus, splenic macrophages and splenicDC appeared to be the primary antigen-presenting cells re-

sponsible for rapid uptake of LC after intravenous infusionin dogs.

LC induces in vitro killing of splenicmacrophages and dendritic cells from dogsLC has been reported to induce rapid apoptosis of macro-phages in vivo and in vitro in mice, but the effects of LCon macrophages in other species have not been examined[24,26]. Moreover, little is known about the effects of LCon splenic DC in other species [34]. Therefore, the effectsof LC on canine splenic macrophages and DC were as-sessed in vitro. For these studies, macrophages and DCwere enriched from normal dog spleen mononuclear cellsby plastic adherence, as described in Methods. Serial dilu-tions of LC or PBS (control) liposomes were added to theadherent cells in medium for 4 hours, at which point the re-maining adherent cells were detached and immunostainedfor quantitation of CD11bþ and CD11cþ cells by flow cy-tometry (Fig. 2). Addition of LC induced a rapid anddose-dependent cytotoxicity of both splenic macrophages(Fig. 2A) and splenic DC (Fig. 2B). In contrast, control(PBS) liposomes did not induce significant cytotoxicity ex-cept at the highest doses tested. Thus, splenic macrophagesand DC from dogs were very sensitive to killing by LC.

Assessment of safety andefficacy of LC treatment in normal dogsNext, studies were conducted to evaluate the safety and tox-icity of LC in normal dogs and to assess the effects of

Figure 1. Uptake of labeled LC in the spleen following intravenous infu-

sion in normal dogs. Uptake of LC by spleen cells after i.v. infusion in

dogs was assessed by administration of LC labeled with the fluorescent

dye DiD, as described in Methods. Labeled LC was infused intravenously

at a dose of 0.5 mL/kg into 2 normal purpose-bred dogs, while a third un-

treated dog served as a control. Four hours after infusion, single cell sus-

pensions of spleen cells were prepared and immunostained with mAbs to

canine CD11b (top panels) or CD11c (bottom panels) and analyzed by

flow cytometry. Approximately 50% of the labeled liposomes were con-

tained within CD11bþ macrophages, whereas approximately 14% of la-

beled liposomes were contained within splenic CD11cþ DC. The

remaining cell-associated labeled liposomes were contained primarily

within unstained mononuclear cells and CD21þ B cells (data not shown).

Similar results were obtained in one additional experiment.

1397M. Mathes et al./ Experimental Hematology 34 (2006) 1393–1402

treatment on clearance of opsonized RBC. Increasing dosesof LC were infused in normal purpose-bred Beagle dogs toassess safety and biological activity. Studies in mice typi-cally use LC doses that range from 60 to 120 mL/kgbody weight. However, these doses are much higher thanwould be practical for use in dogs. However, our recentstudies with intravenous infusion of liposome complexesin dogs suggest that a slow rate of infusion can result ingreater biological activity than bolus injections [35]. There-fore, three relatively low doses of LC (0.1 mL/kg, 0.3 mL/kg, and 1 mL/kg) were administered by slow intravenousinfusion in three normal dogs to assess safety and efficacy.LC was administered by constant rate infusion over a

Figure 2. Effects of in vitro treatment with LC on dog splenic macro-

phages and dendritic cells. To assess the ability of LC to kill dog splenic

macrophages, decreasing concentrations of LC or PBS liposomes (PBS)

were added to adherent spleen mononuclear cells and incubated for 4 hours

and the effects on CD11bþ and CD11cþ cells was assessed by flow cytom-

etry, as described in Methods. (A) Incubation of canine adherent spleen

cells with LC (filled bars) induced a significant, dose-dependent decrease

in the percentages of viable CD11bþ macrophages, compared to untreated

control cells (open bar) or cells treated with control (PBS; light gray bars)

liposomes. (B) Incubation of canine adherent spleen cells with LC (filled

bars) also induced a significant, dose-dependent decrease in the percent-

ages of CD11cþ dendritic cells, compared to untreated control cultures

or cultures treated with control (PBS) liposomes. These data represent

pooled data obtained from 5 independent experiments. (* denotes p !0.05 compared to control cultures.)

90-minute period to each dog (one dog per dose evaluated)and changes in heart rate and respiratory rate and body tem-perature were monitored over the next 24 hours. Changes incomplete blood count and serum biochemistry were also as-sessed at 24 hours and 7 days after infusion. We found thatthe 0.1 mL/kg and 0.3 mL/kg doses of LC were well toler-ated clinically and did not induce detectable clinicalchanges or changes in complete blood count or serum bio-chemistry values (data not shown). However, the 1 mL/kgdose induced transient diarrhea and a mild increase in se-rum alanine aminotransferase and serum alkaline phospha-tase activity within 24 hours after infusion (data notshown).

RBC clearance was assessed by labeling an aliquot ofautologous RBC (the number of RBC labeled in each dogwas calculated to represent approximately 1% of the totalnumber of RBC in circulation) in vitro with the fluorescentdyes CFSE, DiD, or DiI. Labeled RBC were then washedand reinfused intravenously. Samples of blood were col-lected at 2- to 4-hour intervals over the first 24 hours afterinfusion, then daily for 72 hours, and analyzed by flow cy-tometry for the presence of dye-labeled RBC (Fig. 3).Clearance of nonopsonized RBC was assessed first and inthese dogs, there was no appreciable clearance of non-opsonized RBC over a 72-hour period. However, whendye-labeled RBC were first opsonized with rabbit anti-dog RBC antiserum and then reinfused, the RBC were rap-idly cleared from circulation, with nearly complete removalby 24 hours after infusion.

The efficacy of LC treatment was assessed next by mea-suring the effects on RBC clearance kinetics in vivo. Clear-ance of opsonized and nonopsonized RBC was assessed ineach dog prior to LC treatment. Twenty-four hours later, the

Figure 3. Effects of LC treatment on clearance of opsonized RBC in vivo

in normal dogs. Experiments were conducted to determine whether intra-

venous infusion of LC could block clearance of opsonized RBC in dogs in

vivo. Clearance of opsonized and nonopsonized autologous RBC was as-

sessed prior to and 24 hours after LC infusion in a healthy dog treated

with 1 mL/kg LC, using flow cytometry as described in Methods. Prior

to treatment with LC, control RBC (nonopsonized) were not cleared

from circulation, whereas opsonized RBC were rapidly cleared from circu-

lation within 24 hours. Treatment with LC blocked clearance of opsonized

RBC at all time points evaluated in the 24 hours after infusion, compared

to clearance of opsonized RBC clearance determined prior to LC

treatment.

1398 M. Mathes et al. / Experimental Hematology 34 (2006) 1393–1402

dogs were treated by intravenous infusion of LC, and 24hours after LC treatment, the dogs were infused againwith opsonized RBC labeled with a third dye and clearancekinetics were again assessed. Treatment with LC blockedclearance of opsonized RBC in dogs in an apparent dose-dependent fashion. At the 1 mL/kg dose of LC, RBC clear-ance was completely blocked (Fig. 3). At the 0.1 mL/kgdose, clearance of opsonized RBC was reduced by approx-imately 40%, while at the 0.3 mL/kg dose, clearance ofopsonized RBC was reduced by 60%. These results indi-cated that doses of LC between 0.3 and 1 mL/kg could sub-stantially inhibit clearance of opsonized RBC in dogs.However, clinical monitoring indicated induction of mildtoxicity (diarrhea, elevation in liver enzymes) at the 1mL/kg dose. Therefore, we selected a mid-range dose of0.5 mL/kg for subsequent studies of clodronate liposomesin clinical studies in dogs with AIHA. Retreatment of thedogs 3 months later using the 0.5 mL/kg dose of LC alsodemonstrated approximately 80% inhibition of clearanceof opsonized RBC (data not shown).

RBCs are rapidly clearedin dogs with spontaneous AIHAImmune-mediated hemolytic anemia in dogs is a severe dis-ease characterized by rapid and massive RBC destruction.Treatment with high doses of corticosteroids and other im-munosuppressive agents such as azathioprine often fails tostop RBC destruction for several days in these patients[15,36]. Therefore, we conducted a pilot study to assessthe effects of treatment with LC on RBC clearance, hema-tologic parameters, and disease outcomes in dogs with clin-ically severe AIHA. Criteria for entry into the studyincluded a clinical diagnosis of AIHA, which includedpresence of regenerative anemia, spherocytosis, in-salineagglutination, and RBC surface-bound IgG, plus the ab-sence of intercurrent diseases such as renal disease or can-cer (see Materials and Methods for additional details).Seven client-owned dogs with AIHA met the entry criteriaand were enrolled in this pilot study (Table 1).

RBC clearance kinetics in AIHA dogs were assessedprior to LC treatment. In this portion of the study, exvivo opsonization of the RBC was not required since the

majority of RBC from each dog already had detectable sur-face-bound IgG, as detected by flow cytometry (data notshown). After infusion of dye-labeled RBC, samples werecollected every 2 hours for the first 6 hours, then again at12 hours and 24 hours, and the percentage of dye-labeledRBC remaining in circulation was determined. Remarkably,autologous RBC in dogs with severe AIHA were clearedextremely rapidly (Fig. 4). For all 7 dogs, there was signif-icant (p ! 0.001) clearance of dye-labeled RBC within 24hours of infusion. These results suggested that RBC de-struction and turnover was rapid and extensive in dogswith severe AIHA.

Effects of LC treatment on RBCclearance kinetics and hematologicparameters in dogs with spontaneous AIHAThe effects of treatment with LC on RBC clearance andclinical and hematologic parameters were assessed indogs with acute AIHA. Hematologic abnormalities presentin the 7 patients enrolled in this study are summarized asfollows: mean RBC was 1.68 cells � 106/uL (range 0.8to 2.69; normal range, 5.5 to 8.5 cells � 106/uL); meanHct was 16.2 % (range 8 to 20%; normal range, 40 to55%); mean Hb was 4.66 g/dL (range 2.6 to 6.7 g/dL; nor-mal range 13 to 20 g/dL), mean reticulocyte count was288,000/uL (range 94,000 to 575,000/uL; normal range,0 to 60,000/uL); mean platelet count was 201 � 103/uL(range 165 to 405 � 103/uL; normal range, 200 to500 � 103/uL); WBC 23 � 103/uL (range 15.2 to 37.3 �103/uL; normal 4.5 to 15 � 103 cells/uL).

Dogs enrolled in the study were all treated with the sameimmunosuppressive protocol (prednisone, 2.2 mg/kg/dayplus azathioprine, 2.2 mg/kg/day), which was institutedon admission. In addition, each dog enrolled in the trialwere treated with heparin (unfractionated heparin, 150to 200 IU/kg given subcutaneously three times daily),based on the high reported incidence of spontaneous throm-boembolism in dogs with AIHA [21,37]. Pretreatment RBCclearance kinetics were assessed in the first 24 hours afterentry into the study. Next, LC was administered intrave-nously at a dose of 0.5 mL/kg. Twenty-four hours afterthat, the RBC clearance kinetics were determined again,

Table 1. Clinical data for 7 study dogs with AIHA treated with LC

Dog Breed Age Signs Pre Hct Post Hct Pre Hb Post Hb Pre Plat Post Plat RBC clearance

1 Cocker Spaniel 1 Icterus 18 40 6.7 13.2 336 1100 NC

2 Cocker Spaniel 8 Lethargy 14 27 4.3 8.9 168 100 NC

3 Shih Tzu 5 Lethargy 13 21 4.5 7.3 202 110 NC

4 Cocker Spaniel 10 Icterus 9 20 2.2 6.8 223 139 NC

5 Miniature Poodle 8 Lethargy 9 18 1.8 5.8 405 160 D

6 Mixed breed 4 Lethargy 27 27 9.5 9.9 165 280 NC

7 OESD 11 Lethargy 17 ND 5.6 ND 324 ND D

Pre, pretreatment; post, 72 hours after treatment initiated; Hct, hematocrit (%); Hb, hemoglobin (G/dL); Plat, platelet count (� 103/uL); ND, not done; NC,

clearance not changed; D, decreased clearance; OESD, Old English Sheep Dog.

1399M. Mathes et al./ Experimental Hematology 34 (2006) 1393–1402

using RBC labeled with a different fluorescent dye. Com-plete blood counts and biochemical analyses were done at24-hour intervals for the first 72 hours of the study. Afterinfusion of LC, standard immunosuppressive treatmentand heparin continued until the patient was dischargedfrom the hospital.

Treatment with LC resulted in greatly reduced RBCclearance in 2 of 7 treated dogs, when compared to pretreat-ment clearance values (Fig. 4). For example, in the 2 dogsin which RBC clearance was suppressed, it was suppressedby between 80% and 90% compared to pretreatment values.Statistically, there was a significant difference (p 5 0.001)when 1-hour and 24-hour labeled RBC values were com-pared in AIHA dogs prior to LC treatment. However, afterLC treatment, there was no longer a significant difference

Figure 4. RBC clearance and the effects of LC treatment in dogs with

AIHA. RBC clearance in 7 dogs with spontaneous AIHA was assessed

by flow cytometry (see Methods) prior to treatment and again 24 hours af-

ter treatment with LC and the mean (� SE) percentage of labeled RBC

present in circulation at each time point postinfusion was determined.

(A) Prior to treatment, there was a significant decrease (p ! 0.001) in

the percentage of labeled RBC present in circulation 24 hours after infu-

sion of autologous RBC, compared with the percentage of RBC present

in circulation 1 hour after infusion, as assessed by t-test. (B) RBC clear-

ance kinetics were determined again 24 hours after treatment with 0.5

mL/kg LC. In 5 dogs (Nonresponders), there was no change in RBC clear-

ance compared to clearance prior to treatment with LC, whereas in 2 dogs

(Responders), RBC clearance was markedly reduced.

(p 5 0.16) in 1-hour vs 24-hour labeled RBC values. Wealso compared pretreatment and 72-hour posttreatment he-matologic parameters (Hct and platelet counts) between the7 study dogs and 31 disease-matched historical controldogs. The Hct increased 31% in historical control dogs at72 hours after treatment started, compared to a 58% in-crease in Hct in study dogs. Platelet values were also com-pared and we noted a 230% increase in platelet count inhistorical control dogs at 72 hours, whereas in study dogsthe platelet count increased by 163%. Due to the smallnumbers of dogs in the study population, firm conclusionsbased on statistical differences could not be drawn fromthese data. Overall, these results obtained in dogs with se-vere AIHA indicated that LC infusion was well toleratedclinically and appeared to be associated with reducedRBC clearance in a subset of 2 of the 7 treated dogs, thoughwe cannot exclude the possibility that the decrease in RBCclearance was due to the effects of prednisone therapy.

Survival is increased in dogswith AIHA treated with LC comparedto historical disease-matched control dogsImmune-mediated anemia is a life-threatening disease indogs and despite aggressive treatment is associated withmortality rates from 50 to 75% [9,10,13,20]. Most dogswith AIHA die or are euthanized within a week of diagno-sis. Death can occur due to severe anemia or complicationsof treatment, or in some cases due to thromboembolic dis-ease [20]. Therefore, we compared survival times in the 7dogs with AIHA treated with LC to a historical control pop-ulation of 31 dogs with severe AIHA treated with the sameimmunosuppressive protocol, but without LC. The controlpopulation was matched for diagnosis, disease severity,and treatment protocol (prednisone, azathioprine, and hep-arin). We found a significant (p 5 0.035) increase in sur-vival times at 1 week for dogs treated with LC comparedto historical control dogs (Fig. 5). One of 7 dogs in theLC-treated group died in the first week due to complica-tions associated with gastrointestinal thromboembolism.This was compared to 15 of 31 historical control dogsthat died in the first week of treatment. Perhaps more im-portantly, all 6 dogs in the LC-treated group that survivedto 1 week also had long-term survival, to at least 6 months,as ascertained for follow-up telephone contact (data notshown). Though a randomized clinical trial will be neces-sary to confirm these findings, these preliminary resultssuggest that treatment with LC, in conjunction with conven-tional immunosuppressive therapy, may be an effective ad-dition to medications available for management of severeAIHA.

DiscussionSeveral important findings emerged from these studies eval-uating the use of LC for treatment of dogs with AIHA. For

1400 M. Mathes et al. / Experimental Hematology 34 (2006) 1393–1402

one, we found that intravenous-infusion LC was well toler-ated by healthy dogs and by dogs with severe AIHA, withonly mild side effects being observed at the highest dosesevaluated. The drug appeared to be taken up primarily bysplenic macrophages and DC following intravenous infu-sion and in vitro induced rapid killing of both cell types.Remarkably, infusion of even relatively low doses of LCrelative to those used previously in studies in mice blockedclearance of opsonized RBC in normal dogs (Fig. 3) and in2 of 7 dogs with severe AIHA (Fig. 4). In addition, treat-ment of dogs with severe AIHA with LC was associatedwith a significant increase in long-term survival time, com-pared to historical control animals with AIHA (Fig. 5).These are the first studies to our knowledge to report theuse of intravenous infusion of LC in a large-animal sponta-neous autoimmune disease model. Thus, these findings arepotentially of substantial clinical relevance to the manage-ment of AIHA in humans.

LC has been evaluated extensively as a macrophage-de-pleting agent in a variety of mouse models [22,25–27]. Lo-cal administration of LC for treatment of arthritis has alsobeen investigated in rabbits and sheep [30,38,39]. By en-capsulating the clodronate in neutral liposomes, uptake isdirected primarily to phagocytic cells of the macrophagelineage [25]. Once the drug is internalized, it is releasedinto the cytoplasm where it then induces rapid apoptosis[26]. Following intravenous delivery, the drug induces se-lective macrophage depletion, primarily of marginal zonemacrophages [40]. Recent reports and our own unpublisheddata suggest that LC can also induce depletion of certainsubpopulations of DC in mice [34].

Macrophage depletion is an attractive option for treat-ment autoimmune diseases, particularly AIHA and im-mune-mediated thrombocytopenia, where macrophages

Figure 5. Comparison of survival times in AIHA dogs treated with LC

and historical control dogs with AIHA. Survival times at 1 week were

compared between 7 dogs with AIHA treated with LC and 31 historical

control dogs with AIHA matched for disease severity and immunosup-

pressive treatments. Survival times for LC-treated dogs was significantly

longer (p 5 0.035) than for historical control dogs, as assessed by

Kaplan-Meier analysis. In addition, the 6 surviving LC-treated dogs all

survived to at least 6 months, indicative of a long-term survival benefit,

compared to control dogs receiving only conventional immunosuppressive

treatment.

are implicated in extravascular destruction of opsonized tar-get cells. For example, LC was shown previously to be veryeffective in treatment of immune-mediated thrombocytope-nia and AIHA in mouse models [28,29,32]. The currentstudies represent therefore an important extension of thoseearlier studies to a spontaneous large-animal disease model.Autoimmune hemolytic anemia in dogs shares several im-portant features with AIHA in humans. For example, thedisease often occurs spontaneously in dogs and similarRBC determinants are reported to be targeted by patho-genic autoantibodies in dogs with AIHA [41,42]. The trig-gers for development of AIHA in dogs are poorlyunderstood, though it has been reported that recent vaccina-tion may trigger disease onset in some animals [43]. A ge-netic component also appears to exist, inasmuch as CockerSpaniel dogs appear to be at significant risk for developingthe disease [10,19]. Most dogs develop anti-RBC antibodiesof the IgG subclass, either alone or in combination withIgM anti-RBC antibodies [14]. Though it is reported thatsome dogs may develop only IgM antibodies againstRBC, in our experience this is extremely rare (Olver C etal; manuscript in preparation). This is important since mac-rophage depletion would be unlikely to be of much benefitin anemia mediated by IgM antibodies, where the destruc-tion of RBC is more likely to be mediated by complementlysis.

The tempo of RBC destruction in dogs with AIHA ap-pears to be much greater than previously appreciated. Forexample, in the 7 dogs with AIHA evaluated in this report,all had essentially complete clearance of labeled autologousRBC within 24 hours of infusion (Fig. 4). This marked de-gree of RBC destruction may account therefore for the in-ability of LC infusion to stop RBC clearance in all treateddogs. Nonetheless, LC treatment was able to block RBCclearance in 2 of the 7 treated dogs. Higher doses of LCor repeated treatments may be necessary to completelysuppress RBC destruction in dogs with very severe AIHA.

LC were efficient in inducing killing of splenic macro-phages from dogs in vitro (Fig. 2). Less has been reportedregarding the effects of LC on DC killing. It has been re-ported previously that LC treatment can induce DC deple-tion in mice in vivo and our in vitro data are consistent withthe idea that LC can also induce killing of dog DC [34,44].Therefore, it is possible that some of the immunologicaleffects of treatment with LC may reflect in part the effectsof DC depletion.

Because all the dogs with AIHA evaluated in this studywere also receiving concurrent treatment with other immu-nosuppressive drugs, it is possible that there may be impor-tant synergistic interactions between high-dose prednisonetherapy and LC treatment. Therefore, based on the resultsof this study we cannot attribute the effects of LC treatmenton RBC clearance and survival times in dogs with AIHAsolely to the effects of LC. Nonetheless, the results reportedhere are intriguing and suggest that treatment with LC

1401M. Mathes et al./ Experimental Hematology 34 (2006) 1393–1402

warrants further investigation in both dogs and humans withsevere or refractory immune-mediated cytopenias wheremacrophages are thought to play an important role. Insuch diseases, the major role of LC treatment may be torapidly stop the ongoing destruction of RBC or platelets,while gaining time for the more slowly acting standardimmunosuppressive drugs to become effective.

AcknowledgmentsThe authors wish to acknowledge the assistance of Ms. BrianaHarris, Dr. David Twedt, Dr. M. Salman and the CSU VeterinaryTeaching Hospital CCU for their help with these studies. Thiswork was supported by a grant from the Morris Animal Foundation.

References1. Rosse WF, Hillmen P, Schreiber AD. Immune-mediated hemolytic

anemia. Hematology (Am Soc Hematol Educ Program). 2004;48–62.

2. Axelson JA, LoBuglio AF. Immune hemolytic anemia. Med Clin

North Am. 1980;64:597–606.

3. Dhaliwal G, Cornett PA, Tierney LM Jr. Hemolytic anemia. Am Fam

Physician. 2004;69:2599–2606.

4. Izui S. Autoimmune hemolytic anemia. Curr Opin Immunol. 1994;6:

926–930.

5. Schreiber AD. Systemic lupus erythematosusdhematological aspects.

J Rheumatol. 1980;7:395–397.

6. Izui S, Berney T, Shibata T, Fulpius T, Fossati L, Merino R. Molecular

and cellular basis for pathogenicity of autoantibodies. Tohoku J Exp

Med. 1994;173:15–30.

7. Fossati-Jimack L, Ioan-Facsinay A, Reininger L, et al. Markedly dif-

ferent pathogenicity of four immunoglobulin G isotype-switch variants

of an antierythrocyte autoantibody is based on their capacity to inter-

act in vivo with the low-affinity Fcg receptor III. J Exp Med. 2000;

191:1293–1302.

8. Fossati-Jimack L, Reininger L, Chicheportiche Y, et al. High patho-

genic potential of low-affinity autoantibodies in experimental autoim-

mune hemolytic anemia. J Exp Med. 1999;190:1689–1696.

9. Reimer ME, Troy GC, Warnick LD. Immune-mediated hemolytic ane-

mia: 70 cases (1988–1996). J Am Anim Hosp Assoc. 1999;35:384–391.

10. Scott-Moncrieff JC, Treadwell NG, McCullough SM, Brooks MB. He-

mostatic abnormalities in dogs with primary immune-mediated hemo-

lytic anemia. J Am Anim Hosp Assoc. 2001;37:220–227.

11. McCullough S. Immune-mediated hemolytic anemia: understanding

the nemesis. Vet Clin North Am Small Anim Pract. 2003;33:1295–

1315.

12. Carr AP, Panciera DL, Kidd L. Prognostic factors for mortality and

thromboembolism in canine immune-mediated hemolytic anemia:

a retrospective study of 72 dogs. J Vet Intern Med. 2002;16:504–509.

13. Klag AR, Giger U, Shofer FS. Idiopathic immune-mediated hemolytic

anemia in dogs: 42 cases (1986–1990). J Am Vet Med Assoc. 1993;

202:783–788.

14. Quigley KA, Chelack BJ, Haines DM, Jackson ML. Application of

a direct flow cytometric erythrocyte immunofluorescence assay in

dogs with immune-mediated hemolytic anemia and comparison to

the direct antiglobulin test. J Vet Diagn Invest. 2001;13:297–300.

15. Day TK, Macintire DK, Murtaugh RJ, Mathews KA. Differing opin-

ions on treatment of immune-mediated hemolytic anemia. J Am Vet

Med Assoc. 2001;218:1414–1415.

16. Mason N, Duval D, Shofer FS, Giger U. Cyclophosphamide exerts no

beneficial effect over prednisone alone in the initial treatment of acute

immune-mediated hemolytic anemia in dogs: a randomized controlled

clinical trial. J Vet Intern Med. 2003;17:206–212.

17. Weinkle TK, Center SA, Randolph JF, Warner KL, Barr SC, Erb HN.

Evaluation of prognostic factors, survival rates, and treatment proto-

cols for immune-mediated hemolytic anemia in dogs: 151 cases

(1993–2002). J Am Vet Med Assoc. 2005;226:1869–1880.

18. Grundy SA, Barton C. Influence of drug treatment on survival of dogs

with immune-mediated hemolytic anemia: 88 cases (1989–1999).

J Am Vet Med Assoc. 2001;218:543–546.

19. Miller SA, Hohenhaus AE, Hale AS. Case-control study of blood type,

breed, sex, and bacteremia in dogs with immune-mediated hemolytic

anemia. J Am Vet Med Assoc. 2004;224:232–235.

20. McManus PM, Craig LE. Correlation between leukocytosis and nec-

ropsy findings in dogs with immune-mediated hemolytic anemia: 34

cases (1994–1999). J Am Vet Med Assoc. 2001;218:1308–1313.

21. Klein MK, Dow SW, Rosychuk RA. Pulmonary thromboembolism as-

sociated with immune-mediated hemolytic anemia in dogs: ten cases

(1982–1987). J Am Vet Med Assoc. 1989;195:246–250.

22. van Rooijen N, van Nieuwmegen R. Elimination of phagocytic cells in

the spleen after intravenous injection of liposome-encapsulated di-

chloromethylene diphosphonate. An enzyme-histochemical study.

Cell Tissue Res. 1984;238:355–358.

23. Claassen I, Van Rooijen N, Claassen E. A new method for removal of

mononuclear phagocytes from heterogeneous cell populations in vitro,

using the liposome-mediated macrophage ‘‘suicide’’ technique. J Im-

munol Methods. 1990;134:153–161.

24. Van Rooijen N, Sanders A. Liposome mediated depletion of macro-

phages: mechanism of action, preparation of liposomes and applica-

tions. J Immunol Methods. 1994;174:83–93.

25. Van Rooijen N. The liposome-mediated macrophage ‘‘suicide’’ tech-

nique. J Immunol Methods. 1989;124:1–6.

26. Naito M, Nagai H, Kawano S, et al. Liposome-encapsulated dichloro-

methylene diphosphonate induces macrophage apoptosis in vivo and

in vitro. J Leukoc Biol. 1996;60:337–344.

27. van Rooijen N, Sanders A. Elimination, blocking, and activation of

macrophages: three of a kind? J Leukoc Biol. 1997;62:702–709.

28. Alves-Rosa F, Stanganelli C, Cabrera J, et al. Rapid recovery of plate-

let count following administration of liposome-encapsulated clodro-

nate in a mouse model of immune thrombocytopenia. Br J

Haematol. 2002;116:357–366.

29. Alves-Rosa F, Stanganelli C, Cabrera J, van Rooijen N, Palermo MS,

Isturiz MA. Treatment with liposome-encapsulated clodronate as

a new strategic approach in the management of immune thrombocyto-

penic purpura in a mouse model. Blood. 2000;96:2834–2840.

30. Ceponis A, Waris E, Monkkonen J, et al. Effects of low-dose, noncy-

totoxic, intraarticular liposomal clodronate on development of ero-

sions and proteoglycan loss in established antigen-induced arthritis

in rabbits. Arthritis Rheum. 2001;44:1908–1916.

31. Richards PJ, Williams AS, Goodfellow RM, Williams BD. Liposomal

clodronate eliminates synovial macrophages, reduces inflammation

and ameliorates joint destruction in antigen-induced arthritis. Rheuma-

tology (Oxford). 1999;38:818–825.

32. Jordan MB, van Rooijen N, Izui S, Kappler J, Marrack P. Liposomal

clodronate as a novel agent for treating autoimmune hemolytic anemia

in a mouse model. Blood. 2003;101:594–601.

33. Buiting AM, Zhou F, Bakker JA, van Rooijen N, Huang L. Biodis-

tribution of clodronate and liposomes used in the liposome medi-

ated macrophage ‘‘suicide’’ approach. J Immunol Methods. 1996;

192:55–62.

34. Leenen PJ, Radosevic K, Voerman JS, et al. Heterogeneity of mouse

spleen dendritic cells: in vivo phagocytic activity, expression of mac-

rophage markers, and subpopulation turnover. J Immunol. 1998;160:

2166–2173.

35. Dow S, Elmslie R, Kurzman I, MacEwen G, Pericle F, Liggitt D.

Phase I study of liposome-DNA complexes encoding the interleu-

kin-2 gene in dogs with osteosarcoma lung metastases. Hum Gene

Ther. 2005;16:937–946.

1402 M. Mathes et al. / Experimental Hematology 34 (2006) 1393–1402

36. Scott-Moncrieff JC, Reagan WJ, Snyder PW, Glickman LT. Intrave-

nous administration of human immune globulin in dogs with im-

mune-mediated hemolytic anemia. J Am Vet Med Assoc. 1997;210:

1623–1627.

37. Johnson LR, Lappin MR, Baker DC. Pulmonary thromboembolism in

29 dogs: 1985–1995. J Vet Intern Med. 1999;13:338–345.

38. Kinne RW, Schmidt-Weber CB, Hoppe R, et al. Long-term ameliora-

tion of rat adjuvant arthritis following systemic elimination of macro-

phages by clodronate-containing liposomes. Arthritis Rheum. 1995;

38:1777–1790.

39. van Lent PL, van den Bersselaar L, van den Hoek AE, et al. Reversible

depletion of synovial lining cells after intra-articular treatment with li-

posome-encapsulated dichloromethylene diphosphonate. Rheumatol

Int. 1993;13:21–30.

40. van Rooijen N, Kors N, Kraal G. Macrophage subset repopulation in

the spleen: differential kinetics after liposome-mediated elimination.

J Leukoc Biol. 1989;45:97–104.

41. Barker RN, Gruffydd-Jones TJ, Stokes CR, Elson CJ. Identification of

autoantigens in canine autoimmune haemolytic anaemia. Clin Exp Im-

munol. 1991;85:33–40.

42. Barker RN, Elson CJ. Red blood cell glycophorins as B and T-cell

antigens in canine autoimmune haemolytic anaemia. Vet Immunol

Immunopathol. 1995;47:225–238.

43. Duval D, Giger U. Vaccine-associated immune-mediated hemolytic

anemia in the dog. J Vet Intern Med. 1996;10:290–295.

44. Ciavarra RP, Taylor L, Greene AR, et al. Impact of macrophage and

dendritic cell subset elimination on antiviral immunity, viral clearance

and production of type 1 interferon. Virology. 2005;342:177–189.