Von Willebrand Factor Receptor GPIbα is Expressed by Human Factor XIIIa-Positive Dermal Dendrocytes and is Upregulated by Mast Cell Degranulation

Marcia R. Monteiro, Sandor S. Shapiro,* Toshiro Takafuta,* Diana W. Menezes, and George F. Murphy Jefferson Medical College, Department of Pathology, Philadelphia, Pennsylvania, U.S.A.; *Cardeza Foundation for Hematologic Research, Philadelphia, Pennsylvania, U.S.A.

GPIbα, a glycoprotein component of the GPIb-IX- dermal interstitium are demonstrated to represent V complex, serves as a platelet membrane receptor factor XIIIa-positive dermal dendrocytes. In organ that mediates adhesion to von Willebrand factor cultures of neonatal human foreskin, mast cell normally present in the vascular subendothelium. degranulation induced by either substance P or com- Recent data have demonstrated that GPIbα is not pound 48/80 resulted in transiently increased GPIbα restricted to platelets, but is also expressed by endo- expression by dermal dendrocytes. Because the GPIb- thelium in vitro. In this study, we describe the expres- IX-V complex plays a part in regulating hemostasis sion and distribution of GPIbα in normal adult and and may be important for cellular interactions with neonatal human skin. GPIbα is present, as detected extracellular matrix molecules, these data provide by immunohistochemistry, on endothelial cells and on additional insight into the potential function of FXIIIa- highly dendritic cells localized within the perivascular positive dermal dendrocytes in skin remodeling and space, dermal–epidermal junction, and reticular repair. Key words: glycoprotein GPIbα/mast cell secretagog- dermis. By dual-labeling immunofluorescence and ues/recombinant tumor necrosis factor-α. J Invest Dermatol confocal microscopy, GPIbα-positive cells within the 113:272–276, 1999

he glycoprotein GPIb-IX-V complex is a membrane (Ruggeri, 1991; Jandrot-Perrus et al, 1996), whereas the cytoplasmic receptor present on the surface of platelets. It is region of GPIbα interacts with the cell cytoskeleton through at composed of four leucine-rich glycoproteins: the least two actin-binding proteins (Andrews and Fox, 1992; Takafuta disulfide-linked 135 kDa Ibα and 25 kDa Ibβ chains; et al, 1998). Taken collectively, these observations suggest that the and GPIX (22 kDa) and GPV (83 kDa), which are GPIb-IX-V complex is important in mediating platelet adhesion Tnoncovalently associated (Berndt et al, 1985; Lopez and Dong, to von Willebrand factor at subendothelial sites affected by vessel 1997). Glycoproteins Ib and IX are present in equal numbers on injury, and may be important in thrombin-induced platelet the surface of intact platelets, whereas GPV is represented in the aggregation. complex in only half the quantity of the other components Recently, all four GPIb-IX-V components also have been (Modderman et al, 1992). Whether GPIX interacts primarily with documented to be expressed by human umbilical vein and tonsillar GPIbβ, or with GPIbα, remains controversial (Lopez et al, 1994; endothelial cells (Sprandio et al, 1988; Konkle et al, 1990; Wu et al, Wu et al, 1996). 1997). The complex is capable of mediating adhesion of endothelial Initial insight into the functional significance of the GPIb-IX-V cells to a von Willebrand factor substratum (Beacham et al, complex came from studies of the Bernard–Soulier syndrome, 1997). Moreover, certain breast carcinoma cell lines are capable of where the membrane complex is congenitally absent or structurally expressing a GPIbα-related polypeptide (Oleksowicz et al, 1997). defective (Jenkins et al, 1976; Roth, 1996). Affected individuals We now report the novel findings that the GPIb-IX-V complex demonstrate a bleeding disorder characterized by thrombocyto- component GPIbα is expressed by extravascular dermal cells penia, enlarged platelets with markedly decreased adhesion to rabbit with phenotypic characteristics of dermal dendrocytes. We also aortic subendothelium (Weiss et al, 1974; Caen et al, 1976), and demonstrate that mast cell secretion appears to regulate this decreased responsiveness to thrombin (DeMarco et al, 1991). The expression. amino-terminal extracytoplasmic region of GPIbα displays closely associated binding sites for both von Willebrand factor and thrombin MATERIALS AND METHODS Skin specimens Normal adult skin (breast, n ϭ 4, abdomen, n ϭ 2, and face, n ϭ 1) was obtained from elective plastic surgery, and neonatal Manuscript received December 17, 1998; revised April 26, 1999; human foreskin (n ϭ 10) was obtained from elective circumcisions. accepted for publication May 2, 1999. Reprint requests to: Dr. George F. Murphy, Jefferson Medical College, Antibodies, reagents, and cytokines IB1, a mouse monoclonal anti- 1020 Locust St, Rm 545, Philadelphia, PA 19107, U.S.A. body directed against glycoprotein GPIbα, was kindly provided by Dr. E-mail: murphy6@jeflin.tju.edu Zaverio Ruggeri (Scripps Institute, La Jolla, CA). Other antibodies used Abbreviations: FXIIIa, factor XIIIa; rTNF-α, recombinant tumor were: rabbit polyclonal antibody against human factor XIIIa (Calbiochem- necrosis factor α. Novabiochem, San Diego, CA); rabbit polyclonal control IgG (Biogenex,

0022-202X/99/$14.00 · Copyright © 1999 by The Society for Investigative Dermatology, Inc. 272 VOL. 113, NO. 2 AUGUST 1999 EXPRESSION BY DERMAL DENDROCYTES 273

Figure 1. Characterization of GPIbα-reactive dendritic cells in normal adult human skin. Immunohistochemistry for GPIbα demonstrates numerous positive cells with highly dendritic contours (arrows) within the superficial dermis (A), with concentration directly beneath the epidermis (e) (B) and about superficial vessels (v) lined by GPIbα-positive endothelium (A, C). By transmission electron microscopy (D) and correlative immunoelectron microscopy (E), GPIbα reactivity is restricted to interstitial cells characterized by elongate membranous flaps which in cross-section resembled dendrites and averaged 20–30 µm in length (arrowheads). Inset shows typical dendritic profile at higher magnification (insets: D, X12,000; E, X16,500). Note cytoplasmic and possible cell membrane immunoreactivity in E (unstained; Nu ϭ nucleus). Cells similar in contour to GPIbα-reactive dermal dendrocytes (F) also react for FXIIIa in adjacent tissue sections (G); arrowheads denote elongated dendrite-like extensions. Co-expression of GPIbα and FXIIIa reactivity in dermal dendrocytes was achieved using confocal microscopy (H ϭ GPIbα channel, I ϭ FXIIIa channel). Scale bars:(A–C)25µm; (D)2µm; (E)1µm; (F, G)10µm; (H, I)5µm.

San Ramon, CA); mouse monoclonal antibody against: (i) CD31/PECAM- avidin-conjugated Texas Red. For factor XIIIa colocalization, rabbit platelet endothelial cell adhesion molecule (Immunotech, Westbrook, ME); polyclonal antibody against human factor XIIIa (1 : 500) was added for (ii) CD34/human progenitor cell antigen (Becton Dickinson Immunocyto- 1 h at room temperature followed by fluorescein-conjugated donkey anti- metry Systems, San Jose, CA); (iii) E-selectin/CD 62E (Becton Dickinson); rabbit IgG for 30 min at room temperature. When both antibodies to be (iv) human mast cell tryptase (Chemicon, Temecula, CA); and (v) irrelevant colocalized were mouse monoclonal antibodies, i.e., GPIbα and CD31 or control mouse IgG1 antibody (Dako, Carpinteria, CA). GPIbα and CD34, mouse monoclonal antibody against GPIbα was always Fluorescent antibodies and reagents consisted of fluorescein-conjugated incubated overnight at 4°C, followed by a rabbit anti-mouse polyclonal donkey anti-rabbit IgG, Texas Red-conjugated donkey anti-rabbit IgG, antibody (Dako) and finally a donkey anti-rabbit Texas Red conjugate was fluorescein-conjugated donkey anti-mouse IgG (Accurate Scientific, added. Then, the second mouse antibody (CD31 or CD34) was incubated Westbury, NY), avidin-conjugated fluorescein, and avidin-conjugated Texas for an hour at room temperature, followed by a fluorescein-conjugated Red (Leinco Technologies, Ballwin, MO). donkey anti-mouse antibody. Mast cell degranulating agents consisted of compound 48/80 and Negative controls consisted of substitution of primary antibody steps substance P (Sigma, St Louis, MO). Recombinant human tumor necrosis with mouse and rabbit irrelevant antibodies to rule out potential nonspecific factor (TNF)-α (R&D Systems, Minneapolis, MN) was used in the primary antibody binding. Potential cross-reactive binding of secondary following concentrations: 10, 100, and 1000 U per ml. biotinylated and/or fluorescent antibodies was ruled out by switching of these labels, i.e., mouse Ibα antibody was followed by fluorescein- Immunohistochemistry Acetone-fixed cryosections (5 µm) from conjugated donkey anti-rabbit and human factor XIIIa antibody was normal adult skin and neonatal foreskin were stained using a three-step followed by biotinylated horse anti-mouse and avidin-fluorescein. When immunoperoxidase method employing avidin–biotin horseradish peroxidase primary antibodies of the same species were used, the possibility of (Vector Laboratories, Burlingame, CA). Briefly, tissue sections were secondary antibody exchange was ruled out by substitution of one of the incubated with either primary monoclonal antibodies (anti-human GPIbα primary antibodies with an antibody that would bind to a completely antibody concentration 10 µg per ml) or isotype-matched control antibody different site in the skin, i.e., GPIbα and CD1a. overnight at 4°C. After washing in phosphate-buffered saline, a species- specific biotinylated antibody was applied for 30 min at room temperature Confocal microscopy Dual labeling immunofluorescence for GPIbα as the secondary layer, followed by an avidin–biotin horseradish peroxidase and factor XIIIa was performed on normal adult skin using 20 µm sections, complex for 30 min at room temperature. Immunoreactivity was then as described above, and the results were analyzed by a BioRad MRC 600 revealed using 3,3Ј diaminobenzidine (Sigma) with light counterstaining confocal microscope. Three images were consolidated over this range to using Gill’s Hematoxylin No. 1 (Fisher Scientific, Malvern, PA). enhance appreciation of dendritic cell contour.

Dual-labeling immunofluorescence Mouse monoclonal antibody Electron microscopy Routine electron microscopy was done as previ- against GPIbα (40 µg per ml) was incubated on acetone-fixed cryostat ously described (Sueki et al, 1993). For immunoelectron microscopy, fresh sections overnight at 4°C. After thorough washing, biotinylated horse anti- frozen normal adult skin (5 µm sections) was fixed for 10 min with 4% mouse polyclonal antibody (Vector Labs) was then added, followed by paraformaldehyde in 0.1 M phosphate buffer pH 7.4 and washed in 274 MONTEIRO ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Areas were selected which contained the most abundant and prominently stained dendrocyte populations. Only cells containing a nucleus stained by hematoxylin surrounded by a positive rim of immunohistochemical reactivity were enumerated. The significance of differences was determined by a paired Student’s t test (two-tailed). RESULTS Distribution and phenotypic identification of GPIbα in normal human skin In normal adult skin and in neonatal foreskin, GPIbα reactivity was restricted to interstitial cells with highly dendritic contours, as well as to endothelium lining dermal vessels (Fig 1A–C) within the superficial dermis. These dendritic cells tended to be most concentrated directly beneath the epidermal layer (Fig 1B) and about microvessels of the superficial plexus (Fig 1C). The cells consistently showed delicate dendrite-like extensions, 20–30 µm in length, that originated from their cell bodies. Transmission electron microscopy examining serial sections demonstrated these extensions to represent cross-sections through thin membranous flaps which at their thinnest dimension averaged 100 nm in thickness (Fig 1D), and thus had characteristics previously described in detail for ‘‘dermal dendrocytes’’ (Sueki et al, 1993, 1995). Moreover, GPIbα immunoreactivity was confirmed to be cytoplasmic and possibly membrane-associated only in cells with this ultrastructural appearance (Fig 1E). Adjacent sections immunostained for GPIbα (Fig 1F) or FXIIIa (Fig 1G) showed apparent coexpression by dermal dendrocytes, and this impression was confirmed using confocal microscopy (Fig 1H, I), where consolidated images over a 20 µm span enhanced appreciation of dendritic cell contour. GPIbα/FXIIIa-positive dermal dendrocytes also showed variable reactivity for CD34 (data not shown). No positively stained epidermal dendritic cells were observed. Expression of GPIbα-positive cells in organ culture after mast cell degranulation FXIIIa-positive dermal dendrocytes coexpressing GPIbα, as confirmed by dual-label immunofluores- cence (Fig 2A, B), were next evaluated to determine the effects of mast cell degranulation, which has been previously shown to Figure 2. Effects of mast cell degranulation on dermal dendritic induce FXIIIa expression in these cells (Sueki et al, 1993). Mast cells coexpressing Factor XIIIa and GPIbα. Co-expression of these cell degranulation elicited by either compound 48/80 or substance molecules is observed in dendritic cells (arrows) in the perivascular interstitium; α P and confirmed immunohistochemically at 1 h by tryptase note that endothelial cells within vessels (v) also stain for GPIb (B) but not depletion, resulted in increased number and staining intensity of for factor XIIIa (A). Immunohistochemistry for GPIbα in neonatal foreskin α organ culture at baseline (C) and 24 h after substance P-induced mast cell immunoreactive GPIb -positive dermal dendrocytes by 6 h. This degranulation (D). Immunoreactivity for GPIbα on perivascular and effect reached a maximum by 24 h (Fig 2C, D), and returned to interstitial dermal dendritic cells, as compared with baseline and control baseline by 48 h (data not shown). At 24 h, the number of positive explants, is significantly increased 24 h after mast cell degranulation. Scale bar: cells in control explants averaged 33 Ϯ 6.5 (range 24–38), whereas 20 µm(A–D). the number in experimental explants averaged 48.8 Ϯ 8.9 (range 38–58) (p ϭ 0.024). Control cultures failed to demonstrate similar changes at any time-point. Addition of various concentrations of phosphate-buffered saline. Immunohistochemistry for GPIbα was per- rTNF-α, also previously shown to increase FXIIIa expression in formed as described above and after 3,3Ј diaminobenzidine development, dermal dendrocytes by immunohistochemistry (Sueki et al, 1993), the slides were refixed in 2% glutaraldehyde at 4°C overnight and then failed to alter immunoreactivity for GPIbα at any timepoint processed for electron microscopy using inverted beem capsules. examined, despite induction of endothelial E-selectin expression, Organ culture experiments Explants 2–3 mm wide were prepared indicating its biologic efficacy (Walsh et al, 1991). from neonatal foreskin specimens (n ϭ 10) and normal adult skin (abdomen n ϭ 1 and breast n ϭ 1) and incubated in RPMI 1640 media containing DISCUSSION 2 mM L-glutamine, 10% fetal bovine serum, 100 IU penicillin per ml and In this study, we demonstrate the novel findings that GPIbα is (i) 100 µg streptomycin per ml in a Falcon 12 well plate with the epidermis facing up. A total of six replicate experiments were performed to evaluate expressed by FXIIIa-positive and CD34-positive subpopulations of the effects of (i) mast cell degranulation, and (ii) recombinant TNF-α on dermal dendrocytes in normal human skin, and (ii) is transiently the expression of GPIbα. upregulated as a consequence of mast cell degranulation by a For studies of mast cell degranulation, explants were cultured for different mechanism potentially independent of TNF-α. periods of time (1, 6, 24, and 48 h) with medium containing the FXIIIa-positive dermal dendrocytes are now recognized to be degranulating agents: compound 48/80 (0.1 and 1 mg per ml) and distinctive dermal cells separate from truly dendritic monocyte- substance P (5 ϫ 10–4 M final concentration). For studies with recombinant macrophages primarily involved in antigen presentation (Sontheimer TNF-α (rTNF-α), explants were exposed to different concentrations of and Tharp, 1991; Murphy and Liu, 1997). Indeed, three-dimensional the cytokine (10, 100, 1000 U per ml) for varying times (6, 24, 48 h). To reconstructions have shown that FXIIIa-positive dermal dendrocytes evaluate mast cell degranulation and TNF-α effect, immunohistochemistry was performed to detect mast cell tryptase content and E-selectin expression, display dendritic contours only in two-dimensional planes, as a result respectively, as previously described (Kaminer et al, 1991; Walsh et al, 1991). of perpendicular sectioning of thin membrane bound flaps which protrude from the cell body (Sueki et al, 1993, 1995). Moreover, Cell quantitation Five square millimeters of dermal area were evaluated perivascular FXIIIa-positive dermal dendrocytes are intimately for each experimental condition in the mast cell degranulation experiments. associated with mast cells (Sueki et al, 1993) which we have shown VOL. 113, NO. 2 AUGUST 1999 EXPRESSION BY DERMAL DENDROCYTES 275

Figure 3. Schematic representation of GPIb-IX-V complex and its components (modified from Lopez and Dong, 1997). Note the leucine-rich domain (LRD) located nearby the ligand- binding sites (LBS) for both von Willebrand factor and thrombin in the amino-terminal extracytoplasmic region of the GPIbα molecule.

to synthesize, store, and secrete TNF-α upon degranulation (Klein et al, 1989; Walsh et al, 1991). Increased FXIIIa expression by dermal Supported in part by NIH grants CA 40358 (G.F.M.), HL 09163 (S.S.S.), and dendrocytes is observed in skin explants within 24 h after mast cell by fellowship support from the Brandywine Valley Hemophilia Foundation and the degranulation or addition of rTNF-α, suggesting the possibility of Delaware Valley Chapter of the National Hemophilia Foundation (T.T.). mast cell-induced dendrocyte activation (Sueki et al, 1993). These observations have led to the speculation that, despite expression of certain monocyte lineage epitopes and ability to present antigens in REFERENCES in vitro conditions (Nestle et al, 1998), dermal dendrocytes may be Andrews RK, Fox JEB: Identification of a region in the cytoplasmic domain of the functionally heterogeneous and play a potentially important part in platelet membrane glycoprotein Ib-IX complex that binds to purified actin- collaboration with mast cells in cutaneous hemostasis and repair via binding protein. J Biol Chem 267:18605–18611, 1992 production of FXIIIa that stabilizes fibrin and fibronectin macro- Beacham DA, Tran LP, Shapiro SS: Cytokine treatment of endothelial cells increases α molecular complexes (Sueki et al, 1993). glycoprotein Ib -dependent adhesion to von Willebrand factor. Blood 89: α 4071–4077, 1997 We have shown that mast cells contain and release TNF- , and Berndt MC, Gregory C, Kabral A, Zola H, Fournier D, Castaldi PA: Purification and that rTNF-α, in addition to its ability to influence FXIIIa reactivity preliminary characterization of the glycoprotein Ib complex in the human platelet in dermal dendrocytes, results in GPIbα mRNA and protein upregul- membrane. Eur J Biochem 151:637–649, 1985 ation in cultured umbilical vein endothelial cells (Konkle et al, 1990; Caen JP, Nurden AT, Jeanneau C, Michel H, Tobelem G, Levy-Toledano S: Bernard– α Soulier syndrome: a new platelet glycoprotein abnormality. Its relationship with Rajagopalan et al, 1992). In this study, however, rTNF- did not platelet adhesion to subendothelium and with the factor VIII von Willebrand result in similar effects on FXIIIa-positive dermal dendrocytes in the protein. J Lab Clin Med 87:586–596, 1976 concentrations employed. In addition to the possibility that dermal DeMarco L, Mazzucato M, Masotti A, Fenton JW, Ruggeri ZM: Function of α α dendrocytes and microvascular endothelium may respond glycoprotein Ib in platelet activation induced by -thrombin. J Biol Chem α α 266:23776–23783, 1991 differently to TNF- with regard to GPIb expression, technical Jandrot-Perrus M, Bouton MC, Lanza F, Guillin MC: Thrombin interaction with factors inherent in organ culture vs. cell culture may also have platelet membrane glycoprotein Ib. Semin Thromb Hemostas 22:151–156, 1996 influenced these results. Jenkins CS, Phillips DR, Clemetson KJ, Meyer D, Larrieu MJ, Luscher EF: Platelet α membrane glycoproteins implicated in ristocetin-induced aggregation. Studies The documentation of GPIb expression by dermal dendrocytes of the proteins on platelets from patients with Bernard–Soulier syndrome and and its modulation by mast cell degranulation adds further credence von Willebrand’s disease. J Clin Invest 57:112–124, 1976 to the notion that under conditions such as inflammation and repair, Kaminer MS, Lavker RM, Walsh LJ, Whitaker D, Zweiman B, Murphy GF: these cells may interact with von Willebrand factor and, possibly, Extracellular localization of human connective tissue mast cell granule contents. J Invest Dermatol 96:857–863, 1991 thrombin within the extracellular matrix. Because mast cells are Klein LM, Lavker RM, Matis WL, Murphy GF: Degranulation of human mast cells known to degranulate avidly during the early stages of inflammation induces an endothelial antigen central to leukocyte adhesion. Proc Natl Acad Sci (Waldorf et al, 1991) and wound healing (Murphy et al, 1990), the USA 86:8972–8976, 1989 observations reported herein further support their functional linkage Konkle BA, Shapiro SS, Asch AS, Nachman RL: Cytokine-enhanced expression of α glycoprotein Ibα in human endothelium. J Biol Chem 265:19833–19838, 1990 to dermal dendrocytes. Although GPIb is an integral membrane Lopez JA, Dong JF: Structure and function of the glycoprotein Ib-IX-V complex. Curr protein (Fig 3), little is known regarding the cell biology of its Opin Hematol 4:323–329, 1997 expression and regulation. The finding of cytoplasmic GPIbα Lopez JA, Weisman S, Sanan DA, Sih T, Chambers M, Li CQ: Glycoprotein (Gp) Ibβ staining by immunoelectron microscopy raises the possibility of cyto- is the critical subunit linking Gp Ibα and Gp IX in the Gp Ib-IX complex. J Biol Chem 269:23716–23721, 1994 plasmic to cell membrane transport of this molecule, particularly Modderman PW, Admiraal LG, von Sonnenberg A, dem Borne AE: Glycoproteins V when skin is activated during inflammation and repair. Major ques- and Ib-IX form a noncovalent complex in the platelet membrane. J Biol Chem tions remain unanswered regarding the biosynthetic capability of 267:364–369, 1992 dermal dendrocytes for GPIbα production, and factor(s) responsible Murphy GF, Liu V. The dermal immune system. In: Bos JD (ed.). The Skin Immune System (SIS): Cutaneous Immunology and Clinical Immunodermatology, 2nd edn. Boca for mast cell-induced upregulation of this membrane receptor on Raton: CRC Press, 1997, pp 347–364 dermal dendrocytes. Studies are underway to address these import- Murphy GF, Orgill DP, Yannas IV: Partial dermal regeneration is induced by ant issues. biodegradable collagen-glycosaminoglycan grafts. Lab Invest 63:305–313, 1990 276 MONTEIRO ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Nestle FO, Filgueira L, Nickoloff BJ, Burg G: Human dermal dendritic cells process dendrocytes and mast cells in human skin. Implications for hemostasis and matrix and present soluble protein antigens. J Invest Dermatol 110:762–766, 1998 repair. Lab Invest 69:160–172, 1993 Oleksowicz L, Dutcher JP, Deleon-Fernandez M, Paietta E, Etkind P: Human breast Sueki H, Telegan B, Murphy GF: Computer-assisted three-dimensional reconstruction carcinoma cells synthesize a protein immunorelated to platelet glycoprotein-Ibα of human dermal dendrocytes. J Invest Dermatol 105:704–708, 1995 with different functional properties. J Lab Clin Med 129:337–346, 1997 Takafuta T, Wu G, Murphy GF, Shapiro SS: Human beta-filamin is a new protein that α Rajagopalan V,Essex DW, Shapiro SS, Konkle B: Tumor necrosis factor-α modulation interacts with the cytoplasmic tail of glycoprotein Ib . J Biol Chem 273:175311– of glycoprotein Ibα expression in human endothelial and erythroleukemia cells. 117538, 1998 Blood 80:153–161, 1992 Waldorf HA, Walsh LJ, Schechter NM, Murphy GF: Early cellular events in evolving Roth GJ: Molecular defects in the Bernard–Soulier syndrome: assessment of receptor cutaneous delayed hypersensitivity in humans. Am J Pathol 138:477–486, 1991 genes, transcripts and proteins. Comptes Rendus de l’Academie des Sciences—Serie Walsh LJ, Trinchieri G, Waldorf HA, Whitaker D, Murphy GF: Human mast cells contain and release tumor necrosis factor α, which induces endothelial leukocyte Iii, Sciences de la Vie 319:819–826, 1996 adhesion molecule 1. Proc Natl Acad Sci USA 88:4220–4224, 1991 Ruggeri ZM: The platelet glycoprotein Ib-IX complex. Prog Hemost Thromb 10: WeissHJ, Tschopp TB, Baumgartner HR, Sussman II, Johnson MM, Egan JJ: Decreased 35–68, 1991 adhesion of giant (Bernard-Soulier) platelets to subendothelium. Further Sontheimer RD, Tharp MD: Immunological repertoire of the dermal microvascular implications on the role of the von Willebrand factor in hemostasis. Am J Med unit. In: Nickoloff BJ (ed.) Dermal Immune System. Boca Raton: CRC Press, 57:920–925, 1974 1991, pp 117–148 Wu G, Meloni FJ, Shapiro SS: Platelet glycoprotein (Gp) IX associates with GpIbα in Sprandio JD, Shapiro SS, Thiagarajan P, McCord S: Cultured human umbilical vein the platelet membrane GpIb complex. Blood 87:2782–2787, 1996 endothelial cells contain a membrane glycoprotein immunologically related to Wu G, Essex DW, Meloni FJ, Takafuta T, Fujimura K, Konkle BA, Shapiro SS: Human platelet glycoprotein Ib. Blood 71:234–237, 1988 endothelial cells in culture and in vivo express on their surface all four components Sueki H, Whitaker D, Buchsbaum M, Murphy GF: Novel interactions between dermal of the glycoprotein Ib/IX/V complex. Blood 90:2660–2669, 1997