Defective Association of the Platelet Glycoprotein Ib–IX Complex with the Glycosphingolipid-Enriched Membrane Domain Inhibits Murine Thrombus and Atheroma Formation

Defective Association of the Platelet Glycoprotein Ib−IX Complex with the Glycosphingolipid-Enriched Membrane Domain Inhibits Murine Thrombus and This information is current as Atheroma Formation of September 24, 2021. Hao Zhou, Yali Ran, Qi Da, Tanner S. Shaw, Dan Shang, Anilkumar K. Reddy, José A. López, Christie M. Ballantyne, Jerry Ware, Huaizhu Wu and Yuandong Peng J Immunol 2016; 197:288-295; Prepublished online 20 May Downloaded from 2016; doi: 10.4049/jimmunol.1501946 http://www.jimmunol.org/content/197/1/288 http://www.jimmunol.org/

References This article cites 41 articles, 20 of which you can access for free at: http://www.jimmunol.org/content/197/1/288.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 24, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2016 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Defective Association of the Platelet Glycoprotein Ib–IX Complex with the Glycosphingolipid-Enriched Membrane Domain Inhibits Murine Thrombus and Atheroma Formation

Hao Zhou,*,†,1 Yali Ran,†,1 Qi Da,†,1 Tanner S. Shaw,† Dan Shang,†,‡ Anilkumar K. Reddy,† Jose´ A. Lo´pez,x Christie M. Ballantyne,† Jerry Ware,{ Huaizhu Wu,† and Yuandong Peng†

Localization of the platelet glycoprotein Ib–IX complex to the membrane lipid domain is essential for platelet adhesion to von Willebrand factor and subsequent platelet activation in vitro. Yet, the in vivo importance of this localization has never been addressed. We recently found that the disulfide linkage between Iba and Ibb is critical for the association of Iba with the glycosphingolipid-enriched membrane domain; in this study, we established a transgenic mouse model expressing this mutant Downloaded from 2/2 human Iba that is also devoid of endogenous Iba (HaSSMa ). Characterization of this model demonstrated a similar dissociation of Iba from murine platelet glycosphingolipid-enriched membrane to that expressed in Chinese hamster ovary cells, which correlates well with the impaired adhesion of the transgenic platelets to von Willebrand factor ex vivo and in vivo. Furthermore, 2/2 2/2 2/2 2/2 we bred our transgenic mice into an atherosclerosis-prone background (HaSSMa ApoE and HaWTMa ApoE ). We observed that atheroma formation was significantly inhibited in mutant mice where fewer platelet-bound CD11c+ leukocytes were circulating (CD45+/CD11c+/CD41+) and residing in atherosclerotic lesions (CD45+/CD11c+), suggesting that platelet-mediated http://www.jimmunol.org/ adhesion and infiltration of CD11c+ leukocytes may be one of the mechanisms. To our knowledge, these observations provide the first in vivo evidence showing that the membrane GEM is physiologically and pathophysiologically critical in the function of the glycoprotein Ib–IX complex. The Journal of Immunology, 2016, 197: 288–295.

he platelet glycoprotein (GP) Ib–IX complex is comprised interaction can be regulated by various mechanisms; the local- of three type I transmembrane proteins, GP Iba,GPIbb, ization of the GP Ib–IX complex in reference to the cell mem- and GP IX, with a stoichiometry of 1:2:1, in which two brane on either platelets or exogenous cell lines has recently

T 122 484 by guest on September 24, 2021 GP Ibb (Cys ) molecules link with one GP Iba (Cys and been shown to play an important regulatory role (6–8). This Cys485) through two extracellular membrane-proximal disulfide special localization is conferred by the association of the GP Ib– bonds (1), and interact with GP IX in a noncovalent fashion (2, 3). IX complex with a specialized membrane domain enriched with Lack of or dysfunction in GP Ib–IX causes Bernard–Soulier syn- glycosphingolipids, best known as glycosphingolipid-enriched mem- drome in humans, a bleeding disorder that is also recapitulated in branes (GEMs) or raft. Lack of or dysfunction in this association several mouse models expressing no or mutant GP Iba (4, 5). It caused by a disruption of platelet GEM structure (e.g., MbCD has been known that the GP Ib–IX–von Willebrand factor (vWf) treatment) (6) or introduction of a loss-of-association mutation to GP Iba expressed in Chinese hamster ovary (CHO) cells (7) elim- *Department of Hospital Infection Management of Nanfang Hospital, Southern Med- inates or inhibits the function of the GP Ib–IX complex, in particular, ical University, Guangzhou 510515, China; †Cardiovascular Research Section, De- the high shear-induced vWf binding of and signal transmission by ‡ partment of Medicine, Baylor College of Medicine, Houston, TX 77030; Department of the GP Ib–IX complex. Nevertheless, the physiological or path- Vascular Surgery, Union Hospital, Tongji Medical College, Huazhong University of Sci- ence and Technology, Wuhan 430022, China; xPuget Sound Blood Center, Division of ophysiological relevance of the GEMs in the function of the GP Hematology, Department of Medicine, University of Washington, Seattle, WA 98195; { Ib–IX complex has never been addressed and established. Department of Physiology and Biophysics, University of Arkansas for Medical Sciences, We recently demonstrated that disrupting the a/b disulfide link- Little Rock, AR 72205 age decreased the GP Iba–GEM association level in CHO cells, 1H.Z., Y.R., and Q.D. contributed equally to this paper. resulting in a marked inhibition of vWf binding at high shear (7). ORCIDs: 0000-0003-1871-7264 (Y.R.); 0000-0003-4585-2744 (D.S.); 0000-0003- 0359-0029 (Y.P.). In this study, we expressed this GEM association dysfunctional GP Iba in mice, and employed a ferric chloride-induced throm- Received for publication September 1, 2015. Accepted for publication April 29, 2016. bosis and a high cholesterol diet (HCD)–induced atherosclerosis This work was supported by American Heart Association Scientist Development model to investigate the physiological and pathophysiological roles Grant 0635155N (to Y.P.), National Institutes of Health Grants HL095676 (to Y.P.) of the GEMs in the function of the GP Ib–IX complex. and HL098839 (to H.W.), the Alkek Endowment Fund (to Y.P.), and by the Fondren Foundation (to Y.P.). Address correspondence and reprint requests to Dr. Yuandong Peng, Cardiovascular Materials and Methods Research Section, Department of Medicine, Baylor College of Medicine, One Baylor Mice and diets Plaza, Alkek N1317.06, Houston, TX 77030. E-mail address: [email protected] The transgenic mouse expressing the disulfide linkage-deficient human GP Abbreviations used in this article: CHO, Chinese hamster ovary; DPBS, Dulbecco’s a PBS; GEM, glycosphingolipid-enriched membrane; GP, glycoprotein; HCD, high Ib was generated in the Transgenic and Stem Cells Core Facility at the cholesterol diet; vWf, von Willebrand factor. University of Texas Health Science Center-Houston. The mice used in our thrombosis and atherosclerosis models were age- and gender-matched Copyright Ó 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00 littermates. In brief, we injected a ∼6-kb EcoRI fragment possessing the www.jimmunol.org/cgi/doi/10.4049/jimmunol.1501946 The Journal of Immunology 289 entire cassette for human GP Iba expression in mice (9) into mouse zy- and lamellipodia. The spreading was imaged with a Nikon E800 fluores- gotes (C57BL/6J strain) and implanted pseudopregnant females with the cence microscope. Each spreading experiment was performed at least three fertilized embryos. After birth, we performed a PCR-based approach to times, and at least 10 fields of view were counted to calculate the average screen the offspring for the human GP Iba gene in the genome. Further number of adherent platelets. breeding of these positive mice to the wild-type animals demonstrated that Ferric chloride thrombosis model and tail-bleeding time 5 of the 12 mice in the F1 generation stably carried the human GP Iba gene and 3 expressed the human GP Iba mRNA. We then crossed these 3 mice a a2/2 The right common carotid artery of an anesthetized (1.5% isofluorane) with the GP Ib -deficient animals (M ) to remove the coding sequence murine mouse at age 6 mo was exposed to a 10 3 10-mm strip of 3MM for the mouse GP Iba polypeptide. One resulting line was chosen 2/2 Waterman filter paper saturated in a 10% FeCl3 solution for 3 min. After (HaSSMa ) because of its lack of endogenous murine GP Iba and ex- a rinsing three times with PBS, the blood flow and electrocardiogram signal pression of human GP Ib on the platelet surface in a level comparable were monitored using a computer-based high-speed real-time Doppler to that in the wild-type GP Iba-expressing transgenic murine platelets a a2/2 signal-processing system (12–15). The occlusion time was counted as the (H WTM ) (10). In our atherosclerosis study, we took a two-step ap- period from removal of the filter paper to the time when the Doppler signal proach to breed our transgenic mice to an atherosclerosis-prone back- went to nearly zero. Mouse tail-bleeding times were determined by re- ground. First, we crossed the murine GP Iba-deficient mice (Ma2/2) with 2/2 moving 3 mm distal tail tissue and immediately immersing the tail into 37˚C ApoE knockout mice (ApoE ) to obtain mice that lack both murine GP 13 PBS. The time from incision to cessation of blood was defined as the Iba and ApoE (Ma2/2ApoE2/2); second, we crossed these ApoE2/2Ma2/2 2/2 tail-bleeding time. mice with the human GP Iba-expressing transgenic mice (HaWTMa a a2/2 and H SSM ), respectively, to obtain two novel atherosclerosis-prone Oil red staining and quantification of atherosclerotic lesions in a a2/2 2/2 a a2/2 2/2 mouse lines (H WT/M ApoE and H SS/M ApoE ). The de- whole aortas ficiency of the murine ApoE2/2, murine GP Iba (Ma2/2), and the ap- pearance of the human GP Iba mRNA (Ha) was verified either by a PCR After 16 wk on HCD, complete mouse aortas were isolated from the with specific primers for murine ApoE and human GP Iba or by flow ascending aorta to the iliac bifurcation, and then stained with oil red, as Downloaded from cytometry analysis of the surface expression of both human and mouse GP previously described (16). In brief, mouse aortas were cut open and pinned Iba. Mice were fed a normal diet or a HCD (21% fat [w/w], 0.15% cho- on a blue wax tissue-processing disc (Braintree Scientific). After quickly lesterol [w/w], Dyet 112734; Dyets), beginning at the age of 6 mo, and rinsing with ddH2O and then 70% isopropanol, a ∼1.8% solution (w/v) of maintained on a HCD for 16 wk. All animal experiments were done upon oil red (Sigma-Aldrich) in a solvent mixture of isopropanol:water (3:2) an approval from the Institutional Animal Care and Use Committee of was added and incubated with the tissues for 30 min. After staining, the Baylor College of Medicine or National Institutes of Health Office of aortas were washed with 70% isopropanol, followed by ddH2O. Digital Laboratory Animal Welfare. images of the stained aortas were captured using a dissection microscope http://www.jimmunol.org/ attached to a Canon camera, and surface area of the atherosclerotic lesions Sucrose density gradient centrifugation was quantified using Image J software (17, 18). Resting murine platelets were lysed with 1.6 ml ice-cold Brij 35 MES- Ab and flow cytometry analysis of circulating buffered saline (25 mM MES [pH 6.5], 150 mM NaCl, 23 proteinase inhibitor [Roche Diagnostics]) on ice for 1 h (7). The sample was then leukocyte–platelet aggregates (19) 3 mixed with 1.6 ml 80% sucrose in MES, transferred to the bottom of a 14 Two hundred microliters of facial blood were obtained from a mouse and 95-mm centrifuge tube (Seton Scientific), and gently overlaid with 4.8 ml mixed with 20 ml 3.8% sodium citrate. The blood/sodium citrate mixture 30% and then 1.6 ml 5% sucrose in MES. The gradient was then centrifuged was then fixed in 1% paraformaldehyde at room temperature for 10 min. at 34,000 rpm (∼146,000 3 g) in a SW40.Ti rotor (Beckman) for 18 h at 4˚C, Following fixation, 8 ml RBC lysing buffer (Sigma-Aldrich) was added followed by fractionation (from the top of the gradient) into twelve 800 ml 3 and incubated on ice for 10 min. After incubation, 1 ml 10 DPBS by guest on September 24, 2021 samples. Equal volumes (15 ml) of each fraction were resolved using a SDS- (Invitrogen) was added, mixed thoroughly, and then centrifuged at 1800 3 g PAGE gel, and transferred to a polyvinylidene difluoride membrane. GP Iba for 2 min. The pellet was then resuspended in 13 DPBS. The WBCs were was detected by Western blotting. The GEM fractions were identified by the then blocked with rat whole IgG (Jackson ImmunoResearch) and hamster presence of flotillin, a known GEM-specific marker. whole IgG (BioLegend) at 4˚C for 5 min. After blocking, the following Abs Flow chamber assay were used to identify interactions: PE/Cy5-CD45 (BioLegend), PE-CD11c (BioLegend), and FITC-CD41 (BD Biosciences). For controls, the following Sodium citrate anticoagulated murine whole blood was perfused over isotype-matched Abs were used: PE/Cy5-rat IgG2b (BioLegend), PE- k human immobilized vWf (10 mg/ml) in a parallel-plate flow chamber at hamster IgG1 (BioLegend), and FITC-rat IgG1 (BioLegend). The mixtures shear rates of either 1,500 s21 or 15,000 s21. The experiments were recorded of Abs and WBCs were then incubated on a rotator for 30 min at room tem- in real time by a high-speed digital camera (Model Quantix; Photometrics, perature, measured with a Coulter Epics XL-MCL flow cytometer, and ana- Tucson, AZ) connected to an inverted-stage microscope (Eclipse TE300; lyzed using EXPO32 ADC software (Beckman Coulter). Nikon, Garden City, NY). Flow cytometry analysis for the infiltration of leukocytes into Ristocetin-induced vWf binding to transgenic murine platelets mouse aorta Murine whole blood was collected from the inferior vena cava and mixed Mice were sacrificed after feeding HCD for 4 mo and cleared of blood by with acid-citrate dextrose solution (85 mM trisodium citrate, 71 mM citric flushing the system with 30 ml 13 DPBS. The aorta was then dissected out, acid, and 111 mM dextrose) in a 6:1 ratio. After diluting it with an equal cut into 1- to 2-mm pieces, and digested with a mixture of 125 U/ml colla- volume of 13 Dulbecco’s PBS (DPBS; Invitrogen), the blood was centri- genase XI (Sigma-Aldrich), 450 U/ml collagenase type 1 (Sigma-Aldrich), fuged at 90 3 g for 20 min, and the platelet-rich plasma was collected and 60 U/ml hyaluronidase type 1 (Sigma-Aldrich), and 60 U/ml DNase I (Roche) centrifuged at 750 3 g for 10 min in the presence of 50 ng/ml prostacyclin. in a DPBS/20 mM HEPES buffer (pH 7.2) (19). The tube was mixed and The pelleted platelets were then suspended and washed twice with 13 PBS placed into a 42˚C water bath for 1 h and mixed every 10 min. Following di- in the presence of 50 ng/ml prostacyclin. After the final wash, platelets gestion, the mixture was passed through a 70-mm cell strainer (Falcon) and were resuspended in 13 PBS buffer to a concentration of 2.0 3 108/ml. then centrifuged at 850 3 g for 2 min. After centrifugation, the cells were The aggregometry analyses of vWf binding to either wild-type or mutant fixed in 1% paraformaldehyde for 10 min at room temperature, washed, platelets (2 3 105/ml) under stirring conditions were performed in the and resuspended with 13 DPBS. The cells were then incubated with the presence of 1.5 mg/ml ristocetin (Sigma-Aldrich) and 10 mg/ml human same blocking agents, Abs, and isotype control Abs for 30 min at room vWf (Sekisui Diagnostics) with an eight-channel Bio/Data PAP-8C temperature prior to flow cytometry analysis. The same Abs and proce- aggregometer (Biodata). dures in the leukocyte–platelet aggregation assay were used to identify CD45, CD11c, and CD41. Platelet spreading on immobilized vWf Statistical analysis One hundred microliters of washed murine platelets (2 3 108/ml) were gently mixed with 1.5 mg/ml ristocetin and incubated on immobilized human vWf SPSS 21.0 was used for statistical analyses. Values are presented as mean 6 (10 mg/ml) in the presence of 0.8 mM Ca2+/1.2 mM Mg2+ for 1, 2, and 3 h SEM. Student t tests (for comparison between two groups) or one-way at 37˚C. After washing, the adherent platelets were fixed with 4% para- ANOVA (for comparisons of $3 groups) with Newman–Keuls multiple formaldhyde for 15 min at room temperature and stained with a rhodamine- comparison or nonparametric tests were used for statistical analyses: *p , labeled phalloidin (11) to reveal the morphological changes, that is, filopodia 0.05, ***p , 0.01. 290 IN VIVO ROLE OF LIPID DOMAIN IN GP Ib–IX FUNCTION

Results To further address the physiological relevance of these ex vivo The physiological role of the GEMs in the function of the GP findings, we went on to examine the transgenic mice in a ferric Ib–IX complex chloride-induced thrombosis model (Fig. 3), a common model used in the research of the GP Ib–IX complex (20–23). We treated Because platelets are anucleate, it is not possible to manipulate mouse carotid arteries with 10% FeCl (Fig. 3A, upper panel) and platelet gene structure and expression by employing traditional 3 monitored the blood flow with a Doppler system (Fig. 3A, lower approaches. We therefore went on to generate transgenic mice panel) (12–15). We found that the mice expressing wild-type GP expressing the GEM association dysfunctional human GP Iba Iba occluded their vessels faster than the mutant mice, where the (Ha ) (7). As shown in Fig. 1A, both wild-type and mutant hu- SS blood flow in 2 of 14 mutant mice did not even stop (Fig. 3B, open man GP Iba express on the murine platelets with comparable levels; circle). Importantly, we never observe the reperfusion of blood also, no endogenous murine GP Iba was detected, as shown by a flow in either mouse line 30 min after first occlusion, suggesting FITC-labeled rat anti-mouse GP Iba-specific Ab. Of note, a com- that the dysfunction in the GP Ib–IX–GEM association did not parable level of mutant human GP Iba can only be achieved and abolish the GP Iba function, instead only inhibited it. In agree- maintained after the mice have reached the age of 4 mo. Even ment, we found that the tail-bleeding times were comparable be- though we do not know the underlying mechanism for the delayed tween these two transgenic mouse lines, demonstrating that, under expression, consistent with our previous observations in CHO low shear stresses, mutations we introduced to the GP Iba do not cells (7), we observed significantly less percentage of the mutant significantly impact the vWf/GP Iba-mediated hemostasis in a transgenic GP Ib that localizes to the platelet GEMs when com- both mice (Fig. 3C). Taken together, by comparing the two trans- a 6 pared with the wild-type transgenic GP Ib (Fig. 1B, 1C; 33 1.8% genic platelets in several ex vivo and in vivo assays, we demon- Downloaded from 6 versus 50 4.2%). Furthermore, this reduction does not affect the strated that the disulfide linkage deficiency inhibits GP Iba static binding of human vWf to the transgenic platelets induced by localization to platelet GEMs, leading to an impairment of the GP ristocetin, a modulator that can only promote bond formation Iba function, in particular, high shear–induced vWf binding and between human GP Iba and vWf (Fig. 2A). In contrast, both the aIIbb3 activation. interaction of GP Iba with immobilized vWf at high shear (Fig. 2B) The pathophysiological role of the GEMs in the function of the and aIIbb3-mediated platelet morphological changes (Fig. 2C) http://www.jimmunol.org/ were attenuated upon the removal of the a/b disulfide linkage. GP Ib–IX complex In the former, we observed that the mutant platelets rolled at a Several lines of investigation have suggested that vWf partici- velocity ∼2 times faster than the wild type when perfused at a pates in atherogenesis through mediating platelet adhesion to athero- 2 2 shear rate of 15,000 s 1. However, at a shear rate of 1,500 s 1 the sclerotic lesions (24–26). Compared with an acute release (within difference was much less prominent, indicating that the changes minutes) of vWf induced by vascular injury, for example, FeCl3 in the multivalency of the GP Iba–vWf bond due to a/b disulfide treatment, atherogen stimulates endothelial cells to release vWf linkage deficiency cannot be revealed until high shear force is chronically (from weeks to months). Because of the evident im- imposed. To the latter, we found that, upon adhesion to immobi- pairment of vWf binding and the signal-transmitting function of lized vWf, both transgenic platelets were able to undergo a mor- the GP Ib–IX complex in our mutant transgenic mice, we postu- by guest on September 24, 2021 phological change, progressively switching from filopodia protrusion lated such dysfunction may compromise atheroma formation due to lamellipodia spreading (Fig. 2C, left panel). However, com- to an inhibition of the vWf–GP Ib–IX interaction. We bred our pared with the wild type in which .80% of the platelets fully transgenic and GP Iba-deficient mice to an atherosclerosis-prone spread (lamellipodia) within 3 h, only ∼60% of mutant platelets background (ApoE2/2) and fed them a HCD starting at the age of did so, indicative of a much slower aIIbb3 activation in the mutant 6 mo. After 16 wk on HCD, the wild-type human GP Iba transgenic platelets (Fig. 2C, right panel). mice developed extensive atherosclerotic lesions. In contrast, the

FIGURE 1. Disruption of a/b disulfide linkage inhibits the association of transgenic GP Iba with the murine platelet GEMs. (A) Murine whole blood was incubated with either a PE-labeled mouse anti-human GP Iba or a FITC-labeled rat anti-mouse GP Iba mAb. (B) The GP Iba GEM associations were revealed by sucrose gradient density fractionation and Western blotting (n = 8). The GEM fractions are identified by the presence of flotillin, a known GEM- specific marker. (C) The GEM association level of GP Iba is presented as the percentage of GEM-associating GP Iba in respect to the total GP Iba across all sucrose density fractions. All experiments and measurements were done at least three times (***p , 0.01). The Journal of Immunology 291

FIGURE 2. Transgenic GP Ib–IX function is impaired due to a GEM association dysfunction. (A) The aggregometry analyses of vWf binding to transgenic platelets (2 3 105/ml) were performed in the presence of 1.5 mg/ml ristocetin. (B) The murine whole blood was perfused over a human vWf-coated surface in a parallel- plate flow chamber at shear rates of either 1,500 s21 or 15,000 s21.(C) The rhodamine-conjugated phalloidin staining showed that transgenic platelets changed their morphology from Downloaded from filopodia protrusion to lamellipodia upon binding to immobilized vWf. Each experiment was performed at least three times, and the error bars in rolling experiments were calculated from the mean rolling veloci- http://www.jimmunol.org/ ties of at least 100 cells in five different view fields (*p , 0.05, ***p , 0.01). by guest on September 24, 2021

GP Iba mutant and deficient ApoE mice developed apparent, but cells to infiltrate into atherosclerotic lesions and contribute to the significantly fewer severe atherosclerotic lesions in their aortas, as development of atherosclerosis. revealed by oil red staining (Fig. 4A). Quantification of the lesion It has been reported that in patients platelet reactivity and cir- areas in these mice showed a clear and quantifiable amount of culating leukocyte–platelet aggregates are increased with coronary variation in severity (Fig. 4B). In addition, we also included the artery diseases (28–31). In animals, platelet–leukocyte interac- ApoE2/2 only mice in our experiment as a control, and found that tions have also been found to be critical for the initiation and these mice formed atherosclerotic lesions with comparable sizes progression of atherosclerosis. For instance, in hypercholesterol- to the transgenic ApoE2/2 mice expressing wild-type human GP emic animals (with a similar experimental setting as ours), the en- Iba. This result suggests that the genetic manipulation of GP Iba dothelium is deposited with platelet–leukocyte aggregates, which has a minor effect, if any, on the severity of the lesions we ob- makes it more prone to plaque formation (32). Because we pre- served in this study. By preparing single-cell suspensions from viously demonstrated that the CD11c+ monocytes play a critical these atherosclerotic aortas and labeling them with a PE-Cy5– role in atherogenesis in ApoE2/2 mice fed HCD, in this study we labeled Ab against CD45, a general leukocyte marker, we found examined whether platelets may bind to these cells to promote that the number of infiltrated CD45+ leukocytes in atherosclerotic their atherogenic function. To achieve this, we first stained the vessels was progressively reduced, with the most being found in pooled whole-blood samples with the same PE-Cy5–labeled anti- GP Iba wild type, fewer in the mutated version, and the lowest CD45 Ab and then analyzed the CD45-gated CD11c and CD41 being found in GP Iba knockout mice (Fig. 4C). These data triple-positive cells (leukocyte–platelet aggregates) (Fig. 5A). We correlated well with the variation of the lesion sizes in these an- found that the wild-type atherosclerotic mice have a significantly imals, indicating that the alleviated atheroma formation resulted higher number of CD45+CD41+ cells when compared with the specifically from a dysfunction or a loss of function in the GP Ib– mutant mice (Fig. 5B). To our surprise, as shown in Fig. 5A, we IX complex. Consistent with our previous finding on the critical found that nearly all (.98%) of these platelet-bound leukocytes role of CD11c+ cells in atherogenesis (27), we found that there are also CD11c positive (CD45+CD41+CD11c+). Interestingly, in was a significantly reduced portion of CD11c+ cells in the lesions circulation, there appears to be an accumulation of the free CD11c+ of mutant and GP Iba-deficient mice than in the wild-type animals cells in the knockout mice when compared with wild-type and (Fig. 4D). Interestingly, the infiltrated CD11c+ cells were reduced mutant animals (Fig. 5C). This suggests that, in the absence of to a greater extent than other CD45+ cell types in mutant and GP platelets, infiltration of CD11c+ leukocytes is greatly attenuated, Iba-deficient mice relative to the wild-type animals (Fig. 4E), leading to a significant decrease in atherosclerosis in these mice. indicating that platelets may preferentially facilitate these CD11c+ Furthermore, we have previously reported that .85% of these 292 IN VIVO ROLE OF LIPID DOMAIN IN GP Ib–IX FUNCTION

FIGURE 3. Dysfunction in the GP Ib–IX–GEM association delays the formation of a FeCl3-induced thrombus in the carotid artery. (A) Murine right common carotid artery was exposed to a 10 3 10-mm strip of 3MM Waterman filter paper saturated in a 10% FeCl3 solution (upper panel). The blood flow and electrocardiogram signals were monitored using a computer-based high-speed real-time Doppler signal-processing system (lower panel). The occlusion time was counted as the period from the removal of the filter paper to the time when the Doppler signal went to nearly zero. (B) The average oc- Downloaded from clusion time for the wild-type mice

(HaWT, n =8)was∼1 min shorter than that for the mutant animals (HaSS, n = 14). (C) Mouse tail-bleeding time was deﬁned as the time from incision to cessation of blood. http://www.jimmunol.org/

CD11c+ cells in the circulation of this mouse model are lipid- atherosclerotic vessel walls and are selectively localized in ath- laden foamy monocytes rather than lymphocytes and neutrophils, erosclerotic lesions (27, 34, 35). Our results suggest that GP Iba- which do not express CD11c (27, 33, 34). They inﬁltrate into the harboring platelets are important in facilitating the adhesion and by guest on September 24, 2021

FIGURE 4. GEM association dysfunctional GP Iba mitigates atheroma formation in ApoE-deficient animals fed HCD. All samples were collected from mice on HCD for 16 wk starting at the age of 6 mo. (A) Representative oil red staining of the aortas of mice (n = 12, *p , 0.05) fed either HCD or normal diet (ND). (B) Quantification of atherosclerotic lesions in whole aortas was done using Image J software. Mouse aortas were digested with a mixture of enzymes, and the CD45+ (C) and CD11c+ (D) leukocytes were counted in aortas of 16-wk HCD- fed animals (n = 8 of each group, age and gender matched, *p , 0.05, ***p , 0.01). (E) The CD11c+ cells are reduced to a greater extent than other CD45+ cell types in mutant and GP Iba-deficient mice when compared with the wild-type animals. The Journal of Immunology 293

FIGURE 5. CD11c+ leukocytes are the dominant cell type in circulation that can aggregate with platelets in atherosclerotic mice. (A) PE-Cy5–labeled anti-mouse CD45 Ab was used to gate the leukocyte population in whole-blood samples pooled from atherosclerotic mice (n = 12 of each group, age and gender matched). The percentages of the CD45 cells in complex with the platelet (CD41) (B) or coex- pressing CD11c Ag (C) were measured by flow cytometry. All experiments and measurements were done at least three times, and the error bars were calculated from the mean percentage values acquired from all experiments performed. ***p , 0.01. Downloaded from http://www.jimmunol.org/ infiltration of these CD11c-positive monocytes during atheroma the platelet membrane GEM domain to provide sufficient avidity formation (Fig. 6). for the GP Iba–vWf bond strength as well as act as a platform for the assembly of downstream signaling molecules to activate Discussion platelet integrins. Although the first observations on the role of It is well known that the interaction between GP Ib–IX and GEMs in GP Ib–IX function were made from an in vitro experi- vWf captures flowing platelets, initiates the slow translocation of ment .10 y ago, the physiological or pathophysiological rele- platelets (36, 37), and activates platelet integrins (e.g., aIIbb3 and vance of the GEMs in the function of this complex has never been a b a2b1), leading to firm arrest of the tethered platelets (36–40). addressed. We recently observed that disrupting the / disulfide by guest on September 24, 2021 Current belief is that the GP Ib–IX complex needs to be present in linkage markedly decreased the amount of GP Iba instead of b/IX

FIGURE 6. A hypothesized model for the platelet-mediated recruitment and infiltration of CD11c+ leukocytes to atherosclerotic vessel walls. Upon uptake of oxidized low-density lipoprotein (ox-LDL), the circulating leukocytes differentiate from a CD11c2 to a CD11c+ phenotype and increase the expression of activated CD11b (or other GP Ib–IX complex ligands) on their surfaces. In circulation, these CD11c+CD11b+ leukocytes may be captured by platelets that are anchored by endothelial vWf (Interaction I), or prebound with flowing platelets (Interaction II), and then recruited by endothelial vWf (InteractionIII)ina shear-dependent manner. Dysfunction in the GEM association may inhibit these GP Ib–IX complex-mediated interactions, which cause inefficient recruitment and infiltration of the CD11c+ leukocytes to the atherosclerotic vessel wall. Of note, this illustration is revised from an image published previously (27). 294 IN VIVO ROLE OF LIPID DOMAIN IN GP Ib–IX FUNCTION that is associated with the GEMs in CHO cells, and this partial (e.g., CD11b) upon GP Iba mutation (Interaction II). Because lack dissociation inhibited the function of the GP Ib–IX complex in of the a/b disulfide linkage in the GP Ib–IX complex impairs the vWf binding at high shear (7). In this study, we report a novel spreading of the transgenic platelets on an immobilized vWf transgenic mouse model expressing a/b disulfide linkage-deficient surface (Fig. 2C), one would expect that wild-type platelets, once GP Iba. This mutant GP Iba is also dysfunctional in the local- adhered to the vWf-bound atherosclerotic vessel walls (Interaction ization to the murine platelet GEM domain, which caused several I), will spread better than the mutant platelets and provide a better apparent inhibitory effects, including impaired shear-induced surface to recruit and facilitate more inflammatory leukocytes to platelet tethering to surface-bound vWf at high shear, significantly infiltrate into atherosclerotic lesions through an interaction be- delayed morphological changes of the cells, and a longer vessel tween GP Iba and its ligands on the leukocyte cells (Interaction occlusion time upon FeCl3-induced vascular injury in vivo. In II). Our data suggest that GP Iba plays a critical role in the re- contrast, we did notice that the difference between these two mice cruitment and infiltration of CD11c+ monocytes to the athero- in their shear-induced vWf binding can only be revealed at a shear sclerotic vessel wall, a process that is inhibited by dysfunction in rate of 15,000 s21, a condition that is generally not seen in vivo, or lack of GP Iba, thus alleviating atherosclerosis in the mutant suggesting that the major impact upon a/b disulfide linkage de- and knockout animals. Although the mechanism through which ficiency was on the GP Ib–IX–vWf binding-induced aIIbb3 acti- this impairment functions to prohibit atherosclerosis will continue vation. Further investigations are needed to identify the structural to be investigated in future studies, these data demonstrate a novel necessity of the b/IX complex to achieve a better inhibition or role of GP Iba in facilitating the adhesion and infiltration of the even an elimination of the GP Ib–IX–GEM association in which CD11c+ monocytes. both the shear-induced binding function and the vWf binding- Taken together, our study shows that interference in the function Downloaded from induced signaling function of the GP Ib–IX complex could be of the GP Ib–IX complex by inhibition of GP Iba localization to simultaneously abolished in vivo (7, 41). the GEM domain cannot only prolong thrombotic vessel occlusion To investigate whether a dysfunction in the GEM association upon vascular injury, but also inhibits inflammatory atherosclerosis, affects the GP Ib–IX complex-mediated platelet adhesion in a GP two settings that we used in this study to address the physiological Ib–IX-related disease setting (25, 42), we bred our transgenic and pathophysiological relevance of the GEM domain in the func-

mouse model with an atherosclerosis-prone mouse. In comparison tion of the GP Ib–IX complex. In addition, because the mutations we http://www.jimmunol.org/ with the wild-type animals, we found that atheroma formation was introduced into GP Iba are confined to a region that does not par- attenuated in the mutant animals after feeding them a HCD for ticipate in the bindings of any identified GP Ib–IX ligands, we be- 4 mo in which the number of infiltrated leukocytes in atherosclerotic lieve that our newly generated transgenic mouse expressing a GEM vessels is reduced in mutant and knockout mice, indicating that association dysfunctional GP Iba is a unique model that allows a our findings were specific to the role of the GP Ib–IX complex in detailed mechanistic investigation into the specific roles of either the atherogenesis. Interestingly, because we also found a similar de- GEMs in the function of the GP Ib–IX complex or the GP Ib–IX crease in the percentage of the platelet–leukocyte aggregates in complex in various physiological and pathological processes. circulation, our data suggest that the attenuation was imposed by guest on September 24, 2021 even prior to the infiltration of inflammatory cells into atherogen- Disclosures stimulated vessel walls. The authors have no financial conflicts of interest. Based on previous investigations and the data generated from our thrombosis and atherosclerosis models in this study, it is possible that the alleviation of atherosclerosis upon GP Iba dysfunction References a results from a combined impairment of GP Ib -mediated inter- 1. Luo, S. Z., X. Mo, V. Afshar-Kharghan, S. Srinivasan, J. A. Lo´pez, and R. Li. actions (Fig. 6), including the following: 1) impaired interaction 2007. Glycoprotein Ibalpha forms disulfide bonds with 2 glycoprotein Ibbeta between GP Iba with vWf, causing inefficient spreading of ad- subunits in the resting platelet. Blood 109: 603–609. 2. Berndt, M. C., C. Gregory, A. Kabral, H. Zola, D. Fournier, and P. A. Castaldi. herent platelets (Interaction I) and fewer platelet-bound leukocytes 1985. Purification and preliminary characterization of the glycoprotein Ib to adhere to the atherosclerotic vessel walls (Interaction III); and complex in the human platelet membrane. Eur. J. Biochem. 151: 637–649. 2) impaired interaction between GP Iba and its ligands on the 3. Mo, X., N. X. Nguyen, P. A. McEwan, X. Zheng, J. A. Lo´pez, J. Emsley, and R. Li. 2009. Binding of platelet glycoprotein Ibbeta through the convex surface of leukocyte surface (e.g., CD11b, Interaction II) causes fewer leucine-rich repeats domain of glycoprotein IX. J. Thromb. Haemost. 7: 1533–1540. platelets to interact with leukocytes. Because the shear force on 4. Lo´pez, J. A., R. K. Andrews, V. Afshar-Kharghan, and M. C. Berndt. 1998. a Bernard-Soulier syndrome. Blood 91: 4397–4418. the vWf–GP Ib bond will be largely increased when the platelets 5. Bergmeier, W., A. K. Chauhan, and D. D. Wagner. 2008. Glycoprotein Ibalpha are prebound by leukocytes (1, 2), the mutant GP Iba–vWf bonds and von Willebrand factor in primary platelet adhesion and thrombus formation: are weaker when compared with the wild type in supporting lessons from mutant mice. Thromb. Haemost. 99: 264–270. 6. Shrimpton, C. N., G. Borthakur, S. Larrucea, M. A. Cruz, J. F. Dong, and constant contact between platelet-bound leukocytes and the ath- J. A. Lo´pez. 2002. Localization of the adhesion receptor glycoprotein Ib-IX-V erosclerotic vessels, which causes a decrease in the number of complex to lipid rafts is required for platelet adhesion and activation. J. Exp. leukocytes that are captured onto the vessel walls and infiltrated Med. 196: 1057–1066. 7. Geng, H., G. Xu, Y. Ran, J. A. Lo´pez, and Y. Peng. 2011. Platelet glycoprotein Ib into the atherosclerotic lesions in the mutant animals (Interaction beta/IX mediates glycoprotein Ib alpha localization to membrane lipid domain III). In circulation, we observed that the platelet–leukocyte ag- critical for von Willebrand factor interaction at high shear. J. Biol. Chem. 286: + + ∼ 21315–21323. gregate counts (CD45 CD41 ) are decreased by 40% in the 8. Jin, W., O. Inoue, N. Tamura, K. Suzuki-Inoue, K. Satoh, M. C. Berndt, mutant animals relative to the wild-type animals (Fig. 5B), and M. Handa, S. Goto, and Y. Ozaki. 2007. A role for glycosphingolipid-enriched nearly all these CD45+CD41+ aggregates are also CD11c positive microdomains in platelet glycoprotein Ib-mediated platelet activation. J. Thromb. + Haemost. 5: 1034–1040. (Fig. 5A), demonstrating that fewer numbers of CD11c cells are 9. Ware, J., S. R. Russell, P. Marchese, and Z. M. Ruggeri. 1993. Expression of human bound by the mutant platelets than the wild-type platelets (the platelet glycoprotein Ib a in transgenic mice. J. Biol. Chem. 268: 8376–8382. counts of the circulating CD11c+ cells are comparable between 10. Ware, J., S. Russell, and Z. M. Ruggeri. 2000. Generation and rescue of a murine model of platelet dysfunction: the Bernard-Soulier syndrome. Proc. Natl. Acad. these two animals) (Fig. 5C). Thus, our results suggest that the Sci. USA 97: 2803–2808. binding of the platelets to the CD11c+ cells is impaired in the 11. Gu, M., X. Xi, G. D. Englund, M. C. Berndt, and X. Du. 1999. Analysis of the roles of 14-3-3 in the platelet glycoprotein Ib-IX-mediated activation of integrin mutant animals, which is possibly due to an inhibition of the in- a( b( ) using a reconstituted mammalian cell expression model. J. Cell Biol. + IIb) 3 teraction between GP Iba and its ligands on the CD11c cells 147: 1085–1096. The Journal of Immunology 295

12. Hartley, C. J., A. K. Reddy, S. Madala, B. Martin-McNulty, R. Vergona, Functional role of CD11c+ monocytes in atherogenesis associated with hyper- M. E. Sullivan, M. Halks-Miller, G. E. Taffet, L. H. Michael, M. L. Entman, and cholesterolemia. Circulation 119: 2708–2717. Y. X. Wang. 2000. Hemodynamic changes in apolipoprotein E-knockout mice. 28. Furman, M. I., S. E. Benoit, M. R. Barnard, C. R. Valeri, M. L. Borbone, Am. J. Physiol. Heart Circ. Physiol. 279: H2326–H2334. R. C. Becker, H. B. Hechtman, and A. D. Michelson. 1998. Increased platelet 13. Li, Y. H., A. K. Reddy, G. E. Taffet, L. H. Michael, M. L. Entman, and reactivity and circulating monocyte-platelet aggregates in patients with stable C. J. Hartley. 2003. Doppler evaluation of peripheral vascular adaptations to coronary artery disease. J. Am. Coll. Cardiol. 31: 352–358. transverse aortic banding in mice. Ultrasound Med. Biol. 29: 1281–1289. 29. Furman, M. I., M. R. Barnard, L. A. Krueger, M. L. Fox, E. A. Shilale, 14.Reddy,A.K.,G.E.Taffet,Y.H.Li,S.W.Lim,T.T.Pham,J.S.Pocius, D. M. Lessard, P. Marchese, A. L. Frelinger, III, R. J. Goldberg, and M. L. Entman, L. H. Michael, and C. J. Hartley. 2005. Pulsed Doppler signal A. D. Michelson. 2001. Circulating monocyte-platelet aggregates are an early processing for use in mice: applications. IEEE Trans. Biomed. Eng. 52: marker of acute myocardial infarction. J. Am. Coll. Cardiol. 38: 1002–1006. 1771–1783. 30. Sarma, J., C. A. Laan, S. Alam, A. Jha, K. A. Fox, and I. Dransfield. 2002. 15. Reddy, A. K., S. Madala, A. D. Jones, W. A. Caro, J. F. Eberth, T. T. Pham, Increased platelet binding to circulating monocytes in acute coronary syndromes. G. E. Taffet, and C. J. Hartley. 2009. Multichannel pulsed Doppler signal Circulation 105: 2166–2171. processing for vascular measurements in mice. Ultrasound Med. Biol. 35: 31. Keating, F. K., D. A. Whitaker, S. S. Kabbani, M. A. Ricci, B. E. Sobel, and 2042–2054. D. J. Schneider. 2004. Relation of augmented platelet reactivity to the magnitude 16. Beattie, J. H., S. J. Duthie, I. S. Kwun, T. Y. Ha, and M. J. Gordon. 2009. Rapid of distribution of atherosclerosis. Am. J. Cardiol. 94: 725–728. quantification of aortic lesions in apoE(-/-) mice. J. Vasc. Res. 46: 347–352. 32. Huo, Y., A. Schober, S. B. Forlow, D. F. Smith, M. C. Hyman, S. Jung, 17. Nicoletti, A., S. Kaveri, G. Caligiuri, J. Barie´ty, and G. K. Hansson. 1998. Im- D. R. Littman, C. Weber, and K. Ley. 2003. Circulating activated platelets ex- munoglobulin treatment reduces atherosclerosis in apo E knockout mice. J. Clin. acerbate atherosclerosis in mice deficient in apolipoprotein E. Nat. Med. 9: 61–67. Invest. 102: 910–918. 33. Wu, H., X. D. Perrard, Q. Wang, J. L. Perrard, V. R. Polsani, P. H. Jones, 18. Johansson, M. E., X. Y. Zhang, K. Edfeldt, A. M. Lundberg, M. C. Levin, C. W. Smith, and C. M. Ballantyne. 2010. CD11c expression in adipose tissue J. Boreń,W.Li,X.M.Yuan,L.Folkersen,P.Eriksson,etal.2014.Innate and blood and its role in diet-induced obesity. Arterioscler. Thromb. Vasc. Biol. immune receptor NOD2 promotes vascular inflammation and formation of 30: 186–192. lipid-rich necrotic cores in hypercholesterolemic mice. Eur. J. Immunol. 44: 34. Xu, L., X. Dai Perrard, J. L. Perrard, D. Yang, X. Xiao, B. B. Teng, S. I. Simon, 3081–3092. C. M. Ballantyne, and H. Wu. 2015. Foamy monocytes form early and contribute 19. Galkina, E., A. Kadl, J. Sanders, D. Varughese, I. J. Sarembock, and K. Ley. to nascent atherosclerosis in mice with hypercholesterolemia. Arterioscler. Downloaded from 2006. Lymphocyte recruitment into the aortic wall before and during develop- Thromb. Vasc. Biol. 35: 1787–1797. ment of atherosclerosis is partially L-selectin dependent. J. Exp. Med. 203: 35. Foster, G. A., L. Xu, A. A. Chidambaram, S. R. Soderberg, E. J. Armstrong, 1273–1282. H. Wu, and S. I. Simon. 2015. CD11c/CD18 signals very late antigen-4 acti- 20. Liu, J., M. E. Fitzgerald, M. C. Berndt, C. W. Jackson, and T. K. Gartner. 2006. vation to initiate foamy monocyte recruitment during the onset of hypercho- Bruton tyrosine kinase is essential for botrocetin/VWF-induced signaling and lesterolemia. J. Immunol. 195: 5380–5392. GPIb-dependent thrombus formation in vivo. Blood 108: 2596–2603. 36. Savage, B., E. Saldı´var,andZ.M.Ruggeri.1996.Initiation of platelet ad- 21. Bergmeier, W., C. L. Piffath, T. Goerge, S. M. Cifuni, Z. M. Ruggeri, J. Ware, hesion by arrest onto fibrinogen or translocation on von Willebrand factor.

and D. D. Wagner. 2006. The role of platelet adhesion receptor GPIbalpha far Cell 84: 289–297. http://www.jimmunol.org/ exceeds that of its main ligand, von Willebrand factor, in arterial thrombosis. 37. Savage, B., F. Almus-Jacobs, and Z. M. Ruggeri. 1998. Specific synergy of Proc. Natl. Acad. Sci. USA 103: 16900–16905. multiple substrate-receptor interactions in platelet thrombus formation under 22. Guerrero, J. A., G. Shafirstein, S. Russell, K. I. Varughese, T. Kanaji, J. Liu, flow. Cell 94: 657–666. T. K. Gartner, W. Baümler, G. E. Jarvis, and J. Ware. 2008. In vivo relevance for 38. Kroll, M. H., T. S. Harris, J. L. Moake, R. I. Handin, and A. I. Schafer. 1991. von platelet glycoprotein Ibalpha residue Tyr276 in thrombus formation. J. Thromb. Willebrand factor binding to platelet GpIb initiates signals for platelet activation. Haemost. 6: 684–691. J. Clin. Invest. 88: 1568–1573. 23. Joglekar, M. V., J. Ware, J. Xu, M. E. Fitzgerald, and T. K. Gartner. 2013. 39. Savage, B., S. J. Shattil, and Z. M. Ruggeri. 1992. Modulation of platelet Platelets, glycoprotein Ib-IX, and von Willebrand factor are required for FeCl function through adhesion receptors: a dual role for glycoprotein IIb-IIIa (3)-induced occlusive thrombus formation in the inferior vena cava of mice. (integrin a IIb b 3) mediated by fibrinogen and glycoprotein Ib-von Wille- Platelets 24: 205–212. brand factor. J. Biol. Chem. 267: 11300–11306. 24. Methia, N., P. Andre´, C. V. Denis, M. Economopoulos, and D. D. Wagner. 2001. 40. Cruz, M. A., J. Chen, J. L. Whitelock, L. D. Morales, and J. A. Lo´pez. 2005. The

Localized reduction of atherosclerosis in von Willebrand factor-deﬁcient mice. platelet glycoprotein Ib-von Willebrand factor interaction activates the collagen by guest on September 24, 2021 Blood 98: 1424–1428. receptor alpha2beta1 to bind collagen: activation-dependent conformational 25. Massberg, S., K. Brand, S. Gruner,€ S. Page, E. Muller,€ I. Muller,€ W. Bergmeier, change of the alpha2-I domain. Blood 105: 1986–1991. T. Richter, M. Lorenz, I. Konrad, et al. 2002. A critical role of platelet adhesion 41. Xu, G., and D. Shang. Z. zhang, T. S. Shaw, Y. Ran, J. A. Lopez, and Y. Peng. in the initiation of atherosclerotic lesion formation. J. Exp. Med. 196: 887–896. 2015. The transmembrane domains of b and IX subunits mediate the localization 26. Theilmeier, G., C. Michiels, E. Spaepen, I. Vreys, D. Collen, J. Vermylen, and of the platelet glycoprotein Ib-IX complex to the glycosphingolipid-enriched M. F. Hoylaerts. 2002. Endothelial von Willebrand factor recruits platelets to membrane domain. J. Biol. Chem. 290: 22155-22162. atherosclerosis-prone sites in response to hypercholesterolemia. Blood 99: 4486– 42. Koltsova, E. K., P. Sundd, A. Zarpellon, H. Ouyang, Z. Mikulski, A. Zampolli, 4493. Z. M. Ruggeri, and K. Ley. 2014. Genetic deletion of platelet glycoprotein Ib 27.Wu,H.,R.M.Gower,H.Wang,X.Y.Perrard,R.Ma,D.C.Bullard, alpha but not its extracellular domain protects from atherosclerosis. Thromb. A. R. Burns, A. Paul, C. W. Smith, S. I. Simon, and C. M. Ballantyne. 2009. Haemost. 112: 1252–1263.