Regulated surface expression and shedding support a dual role for semaphorin 4D in platelet responses to vascular injury Li Zhu*, Wolfgang Bergmeier†, Jie Wu*, Hong Jiang*, Timothy J. Stalker*, Marcin Cieslak*, Ran Fan*, Laurence Boumsell‡, Atsushi Kumanogoh§, Hitoshi Kikutani§, Luca Tamagnone¶, Denisa D. Wagner†, Marcos E. Milla储**, and Lawrence F. Brass*†† *Departments of Medicine, Pharmacology, 储Biochemistry, and Biophysics, University of Pennsylvania, Philadelphia, PA 19104; †Center for Blood Research, Institute for Biomedical Research and the Department of Pathology, Harvard Medical School, Boston, MA 02115; ‡Institut National de la Sante´ et de la Recherche Me´dicale U659, 94010 Creteil, France; §Department of Molecular Immunology and Core Research for Evolutional Science and Technology Program of Japan Science and Technology Corporation, Research Institute for Microbial Diseases, Osaka University, Osaka 560-0043, Japan; and ¶Institute for Cancer Research and Treatment, University of Turin, 10060 Candiolo, Torino, Italy Edited by Barry S. Coller, The Rockefeller University, New York, NY, and approved December 3, 2006 (received for review July 27, 2006)
signaling 兩 thrombosis 兩 metalloprotease 兩 CD72 兩 plexin-B1
P
latelet activation typically begins with the exposure of collagen within a damaged vessel wall or the local generation of thrombin, but the establishment of a stable thrombus requires the recruitment of additional platelets and the development of stable contacts between platelets (1). Platelet activation also results in the release from platelets of molecules that can affect nearby cells, including endothelial cells and leukocytes as well as other platelets. In a continuing search for molecules that might contribute to contact-dependent events during thrombus formation, we screened human platelets for members of the semaphorin family. Although sempahorins are best known as regulators of neurite outgrowth and vascular development, individual family members have been shown to participate in a variety of events. Class IV semaphorin [semaphorin 4D (sema4D; CD100)] is a type I integral membrane protein first reported on T cells where it supports B cell development by binding to CD72 (2–4). However, sema4D receptors are not limited to B cells. Prior work has shown that a soluble sema4D extracellular domain fragment can activate endothelial cells by its other known receptor, plexin-B1. This causes endothelial migration, actin rearrangement, and the formation of tube-like structures in vitro, responses that are relevant for wound healing and angiogenesis (5–11). Soluble sema4D has also been shown to inhibit monocyte (12) and dendritic cell (13) migration. These studies establish www.pnas.org兾cgi兾doi兾10.1073兾pnas.0606344104
potential roles for sema4D but leave open the cellular origin of soluble sema4D, particularly at sites of vascular injury. Here we demonstrate that platelets express sema4D, plexin-B1, and CD72 and show that mice that lack sema4D have impaired platelet function in vitro and in vivo. We also show that the extracellular domain of sema4D is shed from the platelet surface as thrombus formation proceeds and provide evidence that the metalloprotease ADAM17 is required for this to occur. These results suggest a dual mechanism in which sema4D initially attached to the platelet surface can serve as a ligand for platelet–platelet interactions that favor thrombus formation and then, after being shed from the platelet surface, serve as a bioactive soluble molecule capable of interacting with receptors on endothelial cells and monocytes as well as on nearby platelets. Results Sema4D is an integral protein with a large N-terminal -propeller ‘‘sema’’ domain, followed by an Ig-like domain, a lysine-rich domain, a transmembrane domain, and a cytoplasmic tail with consensus tyrosine and serine phosphorylation sites (14, 15). Western blots of human platelet lysates show that sema4D migrates at ⬇150 kDa under reducing conditions and 300 kDa under nonreducing conditions, the difference presumably reflecting the homodimerization previously reported in lymphocytes (Fig. 1A; ref. 16). The antibody used in these studies, Abm30, is directed against the sema4D C terminus. The same protein was observed with two antibodies (BB18 and BD16) directed against the sema4D extracellular domain (Fig. 1B). The sequence of sema4D RNA amplified from platelets and megakaryoblastic HEL cells matched that reported previously in lymphocytes (not shown). Contrary to an earlier report (16), sema4D was detectable by FACS on the surface of resting platelets. Phorbol 12-myristate 13-acetate (PMA) caused a transient doubling of surface expression, followed by a decline to nearly undetectable levels (Fig. 1 C and D). In theory, the decline in sema4D surface expression seen with prolonged platelet activation could be due to loss of the protein or internalization. Because sema4D cleavage occurs in activated T cells (17), we asked whether the loss of platelet sema4D might also Author contributions: L.Z., W.B., T.J.S., D.D.W., M.E.M., and L.F.B. designed research; L.Z., W.B., J.W., H.J., T.J.S., M.C., R.F., and M.E.M. performed research; W.B., A.K., H.K., L.T., D.D.W., and M.E.M. contributed new reagents/analytic tools; L.Z., W.B., T.J.S., L.T., D.D.W., M.E.M., and L.F.B. analyzed data; and L.Z., W.B., T.J.S., L.B., D.D.W., and L.F.B. wrote the paper. The authors declare no conflict of interest. This article is a PNAS direct submission. Abbreviations: PMA, phorbol 12-myristate 13-acetate; PGI2, prostaglandin I2. **Present address: Roche Pharmaceuticals, Palo Alto, CA 94304. ††To
whom correspondence should be addressed. E-mail:
[email protected].
© 2007 by The National Academy of Sciences of the USA
PNAS 兩 January 30, 2007 兩 vol. 104 兩 no. 5 兩 1621–1626
MEDICAL SCIENCES
Semaphorin 4D (sema4D; CD100) is an integral membrane protein and the ligand for two receptors, CD72 and plexin-B1. Soluble sema4D has been shown to evoke angiogenic responses from endothelial cells and impair monocyte migration, but the origin of soluble sema4D, particularly at sites of vascular injury, has been unclear. Here we show that platelets express sema4D and both of its receptors and provide evidence that these molecules promote thrombus formation. We also show that the surface expression of sema4D and CD72 increases during platelet activation, followed by the gradual shedding of the sema4D extracellular domain. Shedding is blocked by metalloprotease inhibitors and abolished in mouse platelets that lack the metalloprotease ADAM17 (TACE). Mice that lack sema4D exhibit delayed arterial occlusion after vascular injury in vivo, and their platelets show impaired collagen responses in vitro. In resting platelets, as in B lymphocytes, CD72 is associated with the protein tyrosine phosphatase SHP-1. Platelet activation causes dissociation of the complex, as does the addition of soluble sema4D. These findings suggest a dual role for sema4D in vascular responses to injury. As thrombus formation begins, platelet-associated sema4D can bind to its receptors on nearby platelets, promoting thrombus formation. As thrombus formation continues, sema4D is shed from the platelet surface and becomes available to interact with receptors on endothelial cells and monocytes, as well as continuing to interact with platelets.
Fig. 1. Platelets express sema4D. (A) Immunoblots with antibody Abm30 directed to the sema4D C terminus. (B) Proteins were precipitated from platelet lysate by using an anti-sema4D antibody (Abm30, BB16, or BD16) or nonimmune IgG and then probed with Abm30. (C and D) Platelets were incubated with 100 nM PMA for 15 (C) or 5–90 (D) min, stained with either a FITC-conjugated N-terminal sema4D antibody (A8) or FITC-conjugated isotype-matched mouse IgG, and then analyzed by flow cytometry. (D) The results are expressed relative to the amount of sema4D present on resting platelets (mean ⫾ SEM, n ⫽ 3).
reflect cleavage. In an initial experiment, platelets were stimulated with collagen. Immunoblots showed a complete loss of the fulllength protein and the concomitant appearance of a 25- to 30-kDa C-terminal fragment that was not present in resting platelets (Fig. 2A). Thrombin and PMA also caused loss of the intact protein, typically within 30–60 min (Fig. 2B). To determine whether the extracellular domain of sema4D is shed into the fluid phase, platelets were incubated with a nonpermeable biotinylation reagent and then stimulated with the PAR-1 (thrombin receptor) agonist peptide, SFLLRN. A 130-kDa fragment was detected in the supernatant of activated but not resting platelets (Fig. 2C). These results show that sema4D undergoes regulated recruitment to the platelet surface in response to agonists, after which the extracellular domain is gradually shed. Cleavage leaves behind a fragment recognizable with a C-terminal antibody, large enough to include the transmembrane domain as well. To determine whether platelet aggregation affects the rate of sema4D cleavage, experiments were performed in which platelets were incubated with PMA, SFLLRN, or collagen and simultaneously studied with or without the stirring needed for large platelet aggregates to form. In each case, cleavage of sema4D accelerated when aggregation was allowed to occur. Representative results obtained with PMA are shown in Fig. 2D. Note that the completion of cleavage lagged behind the completion of aggregation. The acceleration of cleavage required platelet–platelet contact, not just the mechanical effects of stirring; when aggregation was inhibited by blocking fibrinogen binding to ␣IIb3 with the peptide RGDS, cleavage resembled that seen in unstirred platelets (not shown). Role of ADAM17. Although cleavage of sema4D has been observed
in T cells (17, 18), the protease responsible has not been identified. The observation that sema4D cleavage can occur in washed platelets implies that the protease involved is intrinsic to the platelets and not a plasma protein. Platelets express a number of proteases, including several metalloproteases. Consistent with a role for one of these proteases, sema4D cleavage was inhibited by EDTA (data 1622 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0606344104
Fig. 2. Platelet activation induces sema4D cleavage. (A) Washed platelets were stimulated with collagen (50 g/ml), stirred, and allowed to aggregate at 37°C. Lysates were immunoblotted with the C-terminal sema4D antibody, Abm30. (B) Gel-filtered platelets were incubated with thrombin (0.5 unit/ml) or PMA (100 nM) at 37°C without stirring. Sema4D was precipitated and detected with Abm30. (C) Platelets were surface-labeled with biotin-7-NHS and incubated with SFLLRN (10 M) for 15 min without stirring. (Left) Sema4D in the fluid phase was precipitated with BB18 and detected with antibiotin. (Right) Platelet lysates were precipitated with antibiotin and blotted with Abm30. (D) Washed platelets were stimulated with 100 nM PMA in an aggregometer cuvettes at 37°C, half with stirring and half without. Sema4D was precipitated and detected with Abm30. The extent of aggregation in the stirred samples is indicated. The results are representative of three similar experiments.
not shown). It was also blocked by the metalloprotease inhibitors TAPI-2 (19) and GM6001 (20), which inhibited cleavage stimulated by PMA (Fig. 3 A and B), thrombin, and SFLLRN (not shown) and prevented loss of surface expression. Recent reports have shown that the metalloprotease ADAM17, or TACE, is required for cleavage of two platelet membrane proteins, GP Ib␣ (21) and GP V (22). ADAM17 is a Zn-dependent protease that is synthesized as an inactive precursor. Cleavage of the prodomain renders it active (23, 24). Immunoblots with an antibody against the ADAM17 catalytic domain show the zymogen (or immature) form of ADAM17 is present in both resting and activated platelets (Fig. 3C), consistent with an earlier report (22). A band corresponding to the mature (active) form of ADAM17 is also detectable but was less abundant than the immature form. There was a transient increase in the mature form of ADAM17 during platelet activation but no change in surface expression (Figs. 3 C and D). Additional genetic and pharmacologic approaches were used to determine whether ADAM17 is essential for sema4D cleavage. Because a complete loss of functional ADAM17 is a perinatal lethal (25), we performed studies on chimeric mice expressing a catalytically inactive mutant form of ADAM17 [denoted ADAM17(⌬Zn)] in blood cells after transplantation of ADAM17(⌬Zn/⌬Zn) or WT fetal liver into irradiated WT mice (21). Full-length sema4D was present in platelets from either chimera (Fig. 3E). However, cleavage occurred only in the WT chimeras (i.e., when functional ADAM17 was present). This result shows that ADAM17 is necessary for regulated cleavage of sema4D in mouse platelets. This also appears to be the case in human platelets, because appearance of the C-terminal fragment was inhibited by the ADAM17-selective inhibitor, BMS-561392 (Fig. 3F; ref. 26). Regulated Expression of Sema4D Receptors on Platelets. Sema4D is the ligand for two receptors: plexin-B1, a high-affinity receptor expressed on endothelial cells (5, 9), and CD72, a lower-affinity Zhu et al.
receptor expressed on B cells (27). Although not previously reported, CD72 was detectable in human platelet lysates, and plexin-B1 was detectable in lysates from both human and mouse platelets (Fig. 4A). Platelet plexin-B1 comigrated with plexin-B1 in human endothelial cells and transfected HEK-293 cells. CD72 migrated at ⬇45 kDa, which is the same size as in the human BJAB cell line.‡‡ CD72 was also detectable on the surface of human platelets by FACS (Fig. 4B) and immunofluorescence (not shown). Addition of PMA caused a 6.8 ⫾ 0.4-fold increase in CD72 surface expression after 15 min (mean ⫾ SEM, n ⫽ 3, P ⬍ 0.01). This increase was sustained for at least 60 min, in contrast to the transient 2-fold increase in sema4D expression seen in the same experiments. In lymphocytes, CD72 is associated with the protein tyrosine phosphatase, SHP-1, holding SHP-1 near several of its substrates (28). Coprecipitation studies showed that a CD72/SHP-1 complex ‡‡It
is uncertain at present whether mouse platelets also express CD72. Immunoblots from some, but not all, murine platelet lysates show a band of reactivity that comigrates with CD72 in spleen lysates and is absent from samples obtained from CD72 null mice (not shown).
Zhu et al.
Fig. 4. CD72 and plexin-B1 in platelets. (A) Immunoblots of cell lysates, including HEK-293 cells transfected with plexin-B1 as indicated. (B Upper) Representative flow cytometry data showing CD72 surface expression on resting platelets and on platelets activated with 100 nM PMA for 15 min. (Lower) CD72 surface expression after incubation with PMA expressed as the fold increase in mean fluorescence intensity relative to unstimulated platelets (mean ⫾ SEM, n ⫽ 3). (C) Platelet lysates were immunoprecipitated with anti-CD72 or nonimmune rabbit Ig and then blotted for SHP-1 and CD72. Where indicated, the platelets were preincubated with 100 nM PMA or with recombinant soluble sema4D (1 g/ml).
is present in resting platelets (Fig. 4C). Addition of PMA or recombinant sema4D caused the dissociation of the complex, providing an insight into at least one mechanism by which sema4D can affect platelet responsiveness. Impaired
Platelet
Responses
in
Mice
That
Lack
Sema4D.
Sema4D(⫺/⫺) mice were developed previously and found to have a defect in the development of a subset of B cells (29). To determine whether loss of sema4D expression affects platelet function, platelets were isolated from sema4D(⫺/⫺) and WT mice produced by crossing heterozygotes, and their responses to agonists were compared. Fig. 5A shows representative results obtained with collagen. At suboptimal concentrations the extent of collagen-initiated aggregation was reduced by ⬇75% in the sema4D(⫺/⫺) platelets (mean of pooled data from four separate studies performed in duplicate at either 3 or 5 g/ml collagen; P ⬍ 0.0002; Fig. 5B). There was also a defect in aggregation in response to the glycoprotein (GP) VI collagen receptor ligand, convulxin (not shown), and in collagen-induced phosphorylation of the serine/threonine kinase, Akt (Fig. 5C). Higher concentrations of collagen were required to produce detectable Akt phosphorylation than were required to cause aggregation, but in both cases, a defect seen at lower concentrations was overcome by adding more collagen. In contrast, there was no apparent aggregation defect when sema4D(⫺/⫺) platelets were stimulated with either ADP or the PAR4 agonist peptide, AYPGQV (Fig. 5D). There was also no consistent defect in collagen-induced aggregation when human platelets were studied in the presence of the ADAM17 inhibitor, BMS-561392 (Fig. 5E). These observations suggest that (i) the defect caused by loss of PNAS 兩 January 30, 2007 兩 vol. 104 兩 no. 5 兩 1623
MEDICAL SCIENCES
Fig. 3. ADAM17 and sema4D cleavage. (A) Gel-filtered platelets were preincubated with 20 M TAPI-2, 10 M GM6001, or 10 M GM6001NC (negative control) for 60 min and then stimulated with 100 nM PMA at 37°C for 30 min. Sema4D was precipitated and detected with Abm30. Ig light and heavy chains are indicated. (B) Platelet suspensions pretreated with and without TAPI-2 were incubated with PMA, stained with FITC-A8, and analyzed by flow cytometry (mean ⫾ SEM, n ⫽ 3). (C) Anti-ADAM17 (Tc3-7.49) immunoblot of lysates from human platelets that had been incubated with 100 nM PMA for the times indicated. Note that a longer exposure was used to make the mature protein more readily visible. (D) Surface expression of ADAM17 on human platelets detected by flow cytometry. The platelets were incubated with collagen (10 g/ml) or PMA (100 nM) for 15 min. Results are representative of five experiments. (E) Platelets from irradiated WT mice reconstituted with fetal liver derived from WT or ADAM17/TACE⌬⌮n/⌬⌮n embryos were stimulated with 100 nM PMA for 30 min at 37°C. Lysates were immunoblotted with anti-sema4D antibody, Abm30. (F) Immunoblot for sema4D in lysates prepared from human platelets incubated with PMA following a 15-min preincubation with 1 M BMS-561392 to inhibit ADAM17.
Fig. 6. Consequences of eliminating sema4D expression on thrombus formation in vivo. (A) Time to first occlusion after FeCl3-induced carotid artery injury in WT, heterozygous, and sema4D-null (KO) mice produced by crossing heterozygotes. Black bar indicates mean value. The number of mice is indicated at the bottom. (B) Fraction of injuries resulting in the formation of occlusive thrombi in cremaster muscle arterioles in mice injected with Rose Bengal and exposed to 550-nm green light. The studies were performed with five WT and four KO mice, each of which was studied several times over a period of 2 h.
Fig. 5. Consequences of eliminating sema4D expression or inhibiting sema4D cleavage on platelet function ex vivo. (A) Aggregation in response to collagen using platelets from matched WT and sema4D-null mice. (B) Maximum extent of platelet aggregation in response to collagen (pooled data at 3 and 5 g/ml from four separate studies performed in duplicate; mean ⫾ SEM; n ⫽ 8). (C) Phosphorylation of Akt in response to collagen (representative of three experiments). (D) Aggregation in response to ADP and the PAR4 agonist peptide, AYPGQV. (E) Human platelets pretreated with the ADAM17 inhibitor BMS-561392 (1 M) or vehicle were stimulated with collagen and allowed to aggregate. Afterward, lysates were prepared and immunoblotted with anti-sema4D (Abm30).
sema4D expression is in the collagen response pathway, and (ii) whereas sema4D expression is required, cleavage is not. Two different vascular injury models were used to determine whether loss of sema4D expression also leads to a defect in platelet activation in vivo. In both models, the mice studied were produced by crossing sema4D heterozygotes, and the operators performing the studies were not informed of the genotype of the mice. In the first model, the time to FeCl3-induced carotid artery occlusion was measured. Although there was overlap in some of the individual results, the mean time to occlusion increased by 35%, from 6.4 ⫾ 0.6 min in the WT to 8.7 ⫾ 0.6 min in the knockouts (mean ⫾ SEM; P ⫽ 0.015). Three heterozygotes studied at the same time gave an intermediate result (Fig. 6A). In a second thrombosis model performed in vivo, Rose Bengal and green light were used to produce a local injury to arterioles exposed in an exteriorized cremaster muscle preparation. Platelets were identified by infusing a nonblocking, Alexafluor-647-labeled CD41 antibody. The resultant fluorescent thrombi were detected by using digital videomicroscopy, as described in Methods. Although thrombi formed in both WT and knockout mice, stable arteriolar occlusion was observed in 47% of the injuries in WT mice but none of the knockouts (P ⬍ 0.02, Fig. 6B). Discussion The onset of thrombosis at sites of vascular injury brings platelets into close proximity with each other. The contacts that develop allow interactions to occur among molecules on the surface of 1624 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0606344104
adjacent platelets. A well described example involves the integrin ␣IIb3, but contact-dependent events can also occur when a ligand on the surface of one platelet binds to a receptor on another, as when ephrinB1 binds to EphA4 and EphB1 (30, 31). The present studies evolved from an ongoing effort to identify additional molecules on the platelet surface that can participate in communication between platelets once platelet activation has begun. A screen for candidates showed that human platelets express sema4D and its known receptors. As in T cells (27), sema4D is expressed as a disulfide-linked homodimer on the surface of resting platelets. Once platelet activation has occurred, additional sema4D is rapidly recruited to the platelet surface, leading to a transient increase in surface expression that gradually declines as sema4D is cleaved and shed. Cleavage of sema4D occurs with all of the platelet agonists that we have tested to date. In some respects, the recruitment and subsequent shedding of sema4D as a bioactive fragment resemble the behavior of CD40L, which is also shed from the surface of activated platelets (32). An earlier study showed that sema4D cleavage in Jurkat cells is inhibited by EDTA and EGTA but failed to detect inhibition by the metalloprotease inhibitor, GM6001 (17). The present studies show that in platelets cleavage is blocked by GM6001 or TAPI-2, and both pharmacologic and genetic evidence points toward ADAM17, either serving as the sema4D sheddase or providing a necessary step before the actual sheddase. The apparent discrepancy with the earlier report on GM6001 may be explained by the higher concentration of the inhibitor used here or by differences between platelets and Jurkat cells. How platelet agonists initiate sema4D cleavage is not entirely clear. The process could theoretically involve either the activation of ADAM17 or an increase in substrate availability (or both). Based on immunoblots, resting platelets express predominantly the immature form of ADAM17. We were unable to detect a net change in ADAM17 surface expression during platelet activation, but there was a transient increase in the mature form of the protease. The presence of sema4D and its receptors on platelets inevitably raises questions about their role. Platelets from sema4D(⫺/⫺) mice have an aggregation defect that appears to be limited to the GP VI ligands collagen and convulxin. The mice also display resistance to thrombosis in two distinct vascular injury models, one in a mediumsized artery (the carotid) exposed externally to FeCl3 and the other in a cremaster muscle arteriole exposed from within to light-excited Rose Bengal. The thrombotic response to injuries caused by FeCl3 has been shown to be collagen-dependent (33). In lymphocytes, sema4D on T cells promotes B cell receptor signaling by binding to CD72. Our data show that human platelets express both CD72 and Zhu et al.
Methods Reagents and Mice. GM6001, prostaglandin I2 (PGI2), apyrase,
PMA, thrombin, and ADP were from Sigma-Aldrich (St. Louis, Zhu et al.
MO). SFLLRN and AYPGQV were synthesized at the University of Pennsylvania. TAPI-2 was from CalBiochem (San Diego, CA). BMS-561392 was provided by Bristol-Myers Squibb (New York, NY). The extracellular domain of human sema4D (amino acids 1–734) was amplified by PCR from a full-length cDNA (OriGene Technologies, Rockville, MD), cloned into pSecTag2B, and expressed in FreeStyle HEK-293-F cells (Invitrogen, Carlsbad, CA). Soluble His-tagged sema4D was purified by using TALON affinity resin (Clontech, Palo Alto, CA). Sema4D(⫺/⫺) mice (29) have been backcrossed into a C57 BL/6 at least 10 times, but all comparisons were made between mice from heterozygous crosses. Antibodies. Monoclonal sema4D antibody Abm30 was from Re-
search Diagnostics (Flanders, NJ) and BD Biosciences (San Diego, CA). Sema4D antibodies BB18 and BD16 were described previously (16). Antibody to the sema4D extracellular domain (A8FITC) was from BD Biosciences. Monoclonal antibody Tc3-7.49 recognizes the ADAM17 catalytic domain (41). FITC-conjugated antibody to the ADAM17 extracellular domain was from R&D Systems (Minneapolis, MN). Antiplexin B1 (A8) is a monoclonal antibody against residues 771-1070 of human plexin B1 (Santa Cruz Biotechnology, Santa Cruz, CA). Antiplexin B1 (IC2) is a rabbit polyclonal antibody against intracellular domain of human plexin B1 (42). FITC-conjugated mouse anti-human CD72 (CBL460F) was from Chemicon International (Hampshire, U.K.). Platelet Isolation and Aggregation. Human blood obtained from
healthy donors was anticoagulated 1:5 with ACD (65 mM Na3 citrate/70 mM citric acid/100 mM dextrose, pH 4.4) and centrifuged at 129 ⫻ g for 20 min to obtain platelet-rich plasma (PRP). Gel-filtered platelets were prepared as described (31). Washed platelets were prepared by centrifuging PRP that had been diluted with HEN (150 mM NaCl/1 mM Na2EDTA/10 mM Hepes, pH 6.5) containing 1 M PGI2 and 0.5 units/ml apyrase at 341 ⫻ g for 15 min. Platelets were resuspended in modified Tyrode’s buffer (137 mM NaCl/20 mM Hepes/5.6 mM glucose/1 g/liter BSA/1 mM MgCl2/2.7 mM KCl/3.3 mM/NaH2PO4, pH 7.4). Mouse blood was collected from the inferior vena cava with a heparinized syringe and diluted (1:1) with Tyrode’s buffer. PRP was isolated by centrifugation at 129 ⫻ g for 10 min. The platelet count was adjusted to 2 ⫻ 108 per ml by using autologous platelet poor plasma. Aggregation studies were performed in a ChronoLog (Havertown, PA) aggregometer. Preparation of ADAM17Zn⌬/Zn⌬ Platelets. ADAM17Zn⌬/⫹ mice were
provided by Amgen Biologicals (Thousand Oaks, CA) (25). Fetal liver cells were isolated from ADAM17⫹/⫹ and ADAM17Zn⌬/Zn⌬ sibling embryos at day 16.5 of development and injected into irradiated C57BL/6J recipient mice (1,250 rad; 1 ⫻ 107 cells per mouse). The genotype of the embryos was initially identified phenotypically (open eyes at birth) and subsequently verified by PCR (21). For the platelet studies, the mice were bled under isoflurane anesthesia (IsoFlo, Abbott, North Chicago, IL) from the retroorbital plexus. Blood (7 volumes) was collected into a tube containing 30 units/ml heparin in PBS, pH 7.4 (3 volumes). Plateletrich plasma was obtained by centrifugation at 300 ⫻ g for 10 min then centrifuged at 1,000 ⫻ g in the presence of PGI2 (0.1 g/ml) for 7 min at room temperature. After one washing step, the platelets were resuspended in modified Tyrode’s–Hepes buffer (137 mM NaCl/0.3 mM Na2HPO4/2 mM KCl/12 mM NaHCO3/5 mM Hepes/5 mM glucose, pH 7.3) containing 0.35% BSA and PGI2. Platelet suspensions were incubated with or without 100 nM PMA at 37°C for 30 min and lysed with 2⫻ RIPA buffer containing protease inhibitors [1⫻ RIPA buffer (1% Triton X-100/1% deoxycholate/0.1% SDS/158 mM NaCl/10 mM Tris, pH 7.2/5 mM Na2EDTA)]. PNAS 兩 January 30, 2007 兩 vol. 104 兩 no. 5 兩 1625
MEDICAL SCIENCES
plexin-B1, and that mouse platelets express at least plexin-B1 and may express CD72 as well. These receptors are thought to evoke very different signaling mechanisms. CD72 is a type II integral membrane protein with a cytoplasmic ITIM domain. The binding of T cell sema4D to B cell CD72 releases an interaction between CD72 and the tyrosine phosphatase, SHP-1, allowing SHP-1 to fold into an inactive conformation (34, 35). Similarly, we found that CD72 is associated with SHP-1 in resting human platelets, forming a complex that dissociates when platelets are activated. This observation suggests a model in which an interaction between sema4D and CD72 could serve to release the tonic inhibitory effects of SHP-1 on phosphotyrosine-based signaling downstream of the collagen receptor, GP VI. Plexin-B1, the other sema4D receptor, is thought to signal by at least two mechanisms. Plexin-B1 is a transmembrane protein whose C terminus can bind to cytosolic proteins with appropriate PDZ domains. Interactions with sema4D cause Rho activation by plexin-B1 and the Rho exchange factors PDZ-RhoGEF and LARG (6, 7, 36). Plexin-B1 is also able to regulate Ras (37) and Rac (38) and to activate MAPK, phosphatidylinositol 3-kinase, and Akt (7, 8), which could account for the impairment of Akt activation that we observed in the sema4D-null mice. In cells other than platelets, the receptor tyrosine kinase Met has been shown to form part of a signaling complex with plexin-B1 (9, 11). Although platelets express Met (39), its role in sema4D signaling remains to be established. If interactions between sema4D and its receptors promote platelet activation, is it platelet-associated or soluble sema4D that is required? Although a late interaction with soluble sema4D may also occur, the present data suggest that much of the initial sema4D contribution takes place while sema4D is still associated with the platelet surface. Furthermore, unlike the loss of sema4D expression, inhibition of sema4D cleavage did not inhibit collagen-induced platelet aggregation. This result suggests that, as platelet aggregation begins, sema4D on the platelet surface serves as a direct ligand for receptors on the surface of adjacent platelets. At the same time, there is a large (and stable) increase in the surface expression of CD72, enhancing the ability of platelets to respond to sema4D. Finally, the ability of platelets to shed sema4D provides an intriguing complement to reports showing that endothelial cells can respond to soluble sema4D. Soluble recombinant sema4D causes actin reorganization, cell migration, and tube formation by endothelial cells (5, 9), as does sema4D shed from tumor cell lines (40). These responses suggest a new way that platelets can promote wound healing by favoring the recovery of the endothelial monolayer and promoting the growth of new vessels. Before the start of our studies, the only known potential source of soluble sema4D in blood was lymphocytes. In the context of vascular injury, activated platelets seem a more logical source. Using a conservative estimate of 1,000 copies of sema4D per platelet, a platelet concentration of 3 ⫻ 108 per ml, and a mass of 130 kDa for the exodomain fragment yields a final concentration 65 ng/ml. For comparison, the effects of soluble sema4D on monocytes are reportedly half-maximal at 5–10 ng/ml (12), and stimulation of endothelial cell migration is halfmaximal at ⬇200 ng/ml (5). In conclusion, we propose that platelet sema4D has at least two roles in the response to vascular injury. As platelet activation begins, sema4D that is still on the platelet surface is able to interact directly with receptors on nearby platelets, promoting thrombus formation on exposed collagen. As this continues, ADAM17-dependent cleavage of sema4D releases a soluble fragment that can contribute to wound healing through it effects on endothelial cells and may also affect monocyte activation. These results predict an increase in soluble sema4D levels in patients with thrombotic disorders or on cardiac bypass.
Biotinylation. Washed human platelets (5 ⫻ 108) in Tyrode’s buffer without Hepes and BSA were surface-biotinylated by using biotin7-NHS (Roche Diagnostics, Mannheim, Germany). Antibiotin was from Molecular Probes (Eugene, OR). Immunoprecipitation. Platelets were lysed in either RIPA or Non-
idet P-40 buffer (1% Nonidet P-40/50 mM Tris/150 mM NaCl) with protease inhibitors. After centrifugation at 16,000 ⫻ g for 15 min at 4°C, supernatants were precleared with protein A agarose (for rabbit antibodies) or protein G agarose (for mouse antibodies) and incubated with antibody overnight at 4°C. Protein/antibody complexes were isolated with protein A agarose or protein G agarose for 2 h at 4°C. After three washes with lysis buffer, the beads were boiled in sample buffer (2% SDS/1% 2-mercaptoethanol/0.008% bromophenol blue/80 mM Tris, pH 6.8/1 mM EDTA), and the proteins were resolved by SDS/PAGE and transferred to PVDF membranes. The membranes were blocked with 5% milk, probed as indicated, and visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech, Uppsala, Sweden). Flow Cytometry. Platelets were stained at 37°C for 20 min with
FITC-conjugated mouse anti-human specific antibodies against the extracellular domains of the proteins of interest. Isotype-matched FITC-conjugated mouse IgG was used as a negative control.
was exteriorized under anesthesia, cleaned of connective tissue, and spread on the pedestal of a chamber in which it was continuously superfused with bicarbonate buffer warmed to 37°C and bubbled with 95%N2/5%CO2. After 10 min, Alexafluor-647-labeled CD41 antibody (240 g/kg) was administered by the jugular vein, followed by Rose Bengal (40 mg/kg, Sigma). After an additional 5 min, thrombus formation was induced by exposing a selected arteriole to green light (555 nm) projected through the microscope objective. Use of a high-speed wavelength changer (Lambda DG-4, Sutter Instrument, Novato, CA) resulted in pulsed exposure of the vessel to the green light (10-ms pulses five times per second) throughout the duration of image capture (4.5 min). The resulting injury was limited to an area comparable to the microscope objective field of view, allowing for the induction of multiple injuries in a single mouse. Bright-field and fluorescence images were captured by using a digital CCD camera (SensiCam, Cooke, Auburn Hills, MI) coupled to Slidebook 4.0 image acquisition software (Intelligent Imaging Innovations, Denver, CO). The data are reported as the percentage of injuries that resulted in occlusive thrombi. Injuries that resulted in obvious artery spasm were excluded. Fisher’s exact test was used for statistical analysis.
Vascular Injury. The right common carotid artery of 8- to 12-weekold mice was exposed by blunt dissection and placed in contact for 2 min with a strip of No. 1 Whatman filter paper soaked with 10% FeCl3 (Sigma, St. Louis, MO). After rinsing with saline, blood flow in the artery was observed with a Doppler flow probe for 30 min and the time to first occlusion determined. Thrombus formation in the mouse cremaster muscle microcirculation was visualized by using a modified version of Falati et al. (43). The cremaster muscle
We thank Roy Black and Amgen Biologicals (Thousand Oaks, CA) for making the ADAM17Zn⌬/⫹ mice available; Carl DeCicco and Jim Trzaskos (Bristol-Myers Squibb, New York, NY) for providing BMS-561392; and Bruce Furie, Barbara Furie, Glen Merrill-Skoloff (Beth Israel Deaconess Medical Center, Cambridge, MA), Morty Poncz, and Michael Neyman (University of Pennsylvania) for helping us perform intravital videomicroscopy. These studies were supported by National Institutes of Health [Grants HL40387 and HL81012 (to L.F.B.) and HL56949 (to D.D.W.)], the Japan Ministry of Education, Culture, Sports, Science and Technology and the CREST program of Japan Science and Technology Corporation (A.K. and H.K.), the Italian Association for Cancer Research (L.T.), and American Heart Association Postdoctoral (T.J.S.) and Scientist Development (W.B.) grants.
Brass LF, Zhu L, Stalker TJ (2005) J Clin Invest 115:3385–3392. Bismuth G, Boumsell L (2002) Sci STKE 2002:RE4. Kumanogoh A, Kikutani H (2003) Adv Immunol 81:173–198. Herold C, Dastot H, Roy JP, Bensussan A, Boumsell L (1995) in Leucocyte Typing V, eds Schlossman SF, Boumsell L, Gilks W, Harlan JM, Kishimoto T, Ritz J, Shaw S, Silverstein R, Springer T, Tedder TF, Todd RF (Oxford Univ Press, Oxford), pp 288–289. Basile JR, Barac A, Zhu T, Guan KL, Gutkind JS (2004) Cancer Res 64:5212–5224. Perrot V, Vazquez-Prado J, Gutkind JS (2002) J Biol Chem 277:43115–43120. Aurandt J, Vikis HG, Gutkind JS, Ahn N, Guan KL (2002) Proc Natl Acad Sci USA 99:12085–12090. Basile JR, Afkhami T, Gutkind JS (2005) Mol Cell Biol 25:6889–6898. Conrotto P, Valdembri D, Corso S, Serini G, Tamagnone L, Comoglio PM, Bussolino F, Giordano S (2005) Blood 105:4321–4329. Barberis D, Artigiani S, Casazza A, Corso S, Giordano S, Love CA, Jones EY, Comoglio PM, Tamagnone L (2004) FASEB J 18:592–594. Giordano S, Corso S, Conrotto P, Artigiani S, Gilestro G, Barberis D, Tamagnone L, Comoglio PM (2002) Nat Cell Biol 4:720–724. Delaire S, Billard C, Tordjman R, Chedotal A, Elhabazi A, Bensussan A, Boumsell L (2001) J Immunol 166:4348–4354. Chabbert-de Ponnat I, Marie-Cardine A, Pasterkamp RJ, Schiavon V, Tamagnone L, Thomasset N, Bensussan A, Boumsell L (2005) Int Immunol 17:439–447. Suzuki K, Kumanogoh A, Kikutani H (2003) Cytokine Growth Factor Rev 14:17–24. Love CA, Harlos K, Mavaddat N, Davis SJ, Stuart DI, Jones EY, Esnouf RM (2003) Nat Struct Biol 10:843–848. Bougeret C, Mansur IG, Dastot H, Schmid M, Mahouy G, Bensussan A, Boumsell L (1992) J Immunol 148:318–323. Elhabazi A, Delaire S, Bensussan A, Boumsell L, Bismuth G (2001) J Immunol 166:4341–4347. Wang X, Kumanogoh A, Watanabe C, Shi W, Yoshida K, Kikutani H (2001) Blood 97:3498–3504. Crowe PD, Walter BN, Mohler KM, Otten-Evans C, Black RA, Ware CF (1995) J Exp Med 181:1205–1210. Galardy RE, Grobelny D, Kortylewicz ZP, Poncz L (1992) Matrix Suppl 1:259–262.
21. Bergmeier W, Piffath CL, Cheng G, Dole VS, Zhang Y, von Andrian UH, Wagner DD (2004) Circ Res 95:677–683. 22. Rabie T, Strehl A, Ludwig A, Nieswandt B (2005) J Biol Chem 280:14462–14468. 23. Leonard JD, Lin F, Milla ME (2005) Biochem J 387:797–805. 24. Gonzales PE, Solomon A, Miller AB, Leesnitzer MA, Sagi I, Milla ME (2004) J Biol Chem 279:31638–31645. 25. Peschon JJ, Slack JL, Reddy P, Stocking KL, Sunnarborg SW, Lee DC, Russell WE, Castner BJ, Johnson RS, Fitzner JN, et al. (1998) Science 282:1281–1284. 26. Grootveld M, McDermott MF (2003) Curr Opin Invest Drugs 4:598–602. 27. Kumanogoh A, Kikutani H (2001) Trends Immunol 22:670–676. 28. Adachi T, Flaswinkel H, Yakura H, Reth M, Tsubata T (1998) J Immunol 160:4662–4665. 29. Shi W, Kumanogoh A, Watanabe C, Uchida J, Wang X, Yasui T, Yukawa K, Ikawa M, Okabe M, Parnes JR, et al. (2000) Immunity 13:633–642. 30. Prevost N, Woulfe DS, Jiang H, Stalker TJ, Marchese P, Ruggeri Z, Brass LF (2005) Proc Natl Acad Sci USA 102:9820–9825. 31. Prevost N, Woulfe D, Tanaka T, Brass LF (2002) Proc Natl Acad Sci USA 99:9219–9224. 32. Andre P, Prasad KS, Denis CV, He M, Papalia JM, Hynes RO, Phillips DR, Wagner DD (2002) Nat Med 8:247–252. 33. Dubois C, Panicot-Dubois L, Merrill-Skoloff G, Furie B, Furie BC (2006) Blood 107:3902–3906. 34. Nakayama E, von Hoegen I, Parnes JR (1989) Proc Natl Acad Sci USA 86:1352–1356. 35. Wu Y, Nadler MJ, Brennan LA, Gish GD, Timms JF, Fusaki N, Jongstra-Bilen J, Tada N, Pawson T, Wither J, et al. (1998) Curr Biol 8:1009–1017. 36. Swiercz JM, Kuner R, Behrens J, Offermanns S (2002) Neuron 35:51–63. 37. Oinuma I, Ishikawa Y, Katoh H, Negishi M (2004) Science 305:862–865. 38. Vikis HG, Li W, Guan KL (2002) Genes Dev 16:836–845. 39. Pietrapiana D, Sala M, Prat M, Sinigaglia F (2005) FEBS Lett 579:4550–4554. 40. Basile JR, Castilho RM, Williams VP, Gutkind JS (2006) Proc Natl Acad Sci USA 103:9017–9022. 41. Moss ML, Jin SLC, Milla ME, Burkhart W, Carter HL, Chen WJ, Clay WC, Didsbury JR, Hassler D, Hoffman CR, et al. (1997) Nature 385:733–736. 42. Artigiani S, Barberis D, Fazzari P, Longati P, Angelini P, van de Loo JW, Comoglio PM, Tamagnone L (2003) J Biol Chem 278:10094–10101. 43. Falati S, Gross P, Merrill-Skoloff G, Furie BC, Furie B (2002) Nat Med 8:1175–1180.
1. 2. 3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13.
14. 15. 16. 17. 18. 19. 20.
1626 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0606344104
Zhu et al.
