主题：vi

【关键词】 Conformational

The present study characterizes the methanethiosulfonate (MTS) inhibition profiles of 26 consecutive cysteine-substituted mutants comprising transmembrane (TM) helix 6 of the human apical Na+-dependent bile acid transporter (SLC10A2). TM6 is linked exofacially to TM7 via extracellular loop 3. TM7 was identified previously as lining part of the substrate permeation path (Mol Pharmacol 70: 1565, 2006[Abstract/Free Full Text]). Most TM6 cysteine replacements were well tolerated, except for five residues with either severely hampered (I229C, G249C) or abolished (P234C, G237C, G241C) activity. Disruption of protein synthesis or folding and stability may account for lack of activity for mutant P234C. Subsequent Pro234 amino acid replacement reveals its participation in both structural and functional aspects of the transport cycle. Application of polar MTS reagents (1 mM) significantly inhibited the activity of six mutants (V235C, S239C, F242C, R246C, A248C, and Y253C), for which rates of modification were almost fully reversed (except Y253C) upon inclusion of bile acid substrates or removal of Na+ from the MTS preincubation medium. Activity assessments at equilibrative [Na+] revealed numerous Na+-sensitive residues, suggesting their proximity in or around Na+ interaction sites. In silico modeling reveals the intimate and potentially cooperative orientation of MTS-accessible TM6 residues toward functionally important TM7 amino acids, substantiating TM6 participation during the transport cycle. We conclude a functional requirement for helical flexibility imparted by Pro234, Gly237, and Gly241, probably forming a "conformational switch" requisite for substrate turnover; meanwhile, MTS-accessible residues, which line a helical face spatially distinct from this switch, may participate during substrate permeation.

By coupling bile acid movement to the passive flow of Na+ ions down their concentration gradient, the human apical Na+-dependent bile acid transporter (ASBT; SLC10A2) concentrates bile acids within the cell interior. Viewed from a physiological perspective, ASBT effectively conserves the body's recirculating bile acid pool (Trauner and Boyer, 2003) in tandem with numerous active transporters expressed along the enterohepatic pathway. Because cholesterol provides the precursor molecule in FXR- and hepatic CYP7A-mediated bile acid synthesis (Chiang et al., 2001; Pauli-Magnus et al., 2005), ASBT also constitutes a key modulator of cholesterol homeostasis. Numerous studies have recently underscored the exploitive potential of ASBT in cholesterol-lowering therapies (Oelkers et al., 1997; Izzat et al., 2000; Huff et al., 2002; Li et al., 2004) and emphasizing the usefulness of this high-capacity, high-affinity transporter in prodrug targeting (Swaan et al., 1997; Balakrishnan and Polli, 2006; Geyer et al., 2006). Consequently, ASBT's unique pharmaceutical relevance coupled to the absence of a crystal structure has provided a strong impetus toward elucidation of its structure/function relationships.

Using cysteine mutagenesis and thiol modification (SCAM), our previous studies identified transmembrane (TM) domain 7 in forming part of the putative substrate permeation pathway (Hussainzada et al., 2006) with extracellular loop (EL) 3 containing Na+ and bile acid interaction sites (Banerjee et al., 2008). We continue SCAM analysis along TM6 based on the following rationale: 1) our topology model published previously predicts that TM6 lies adjacent to and may interact with TM7 in forming a putative translocation pathway (Zhang et al., 2004); 2) EL3 amino acids link TM6 and TM7 membrane-spanning segments along the exofacial matrix; 3) the highly conserved nature of TM6 amino acids corroborate a potential role during transport; 4) presence of the charged, conserved Arg246, which could potentially participate in electrostatic interactions implicated previously during ligand binding (Banerjee et al., 2008); and finally, 5) the presence of two conserved proline residues (Pro234, Pro251), which have been shown in other membrane-bound carriers to provide cation binding sites and enable formation of conformational switches essential for substrate translocation (Deber et al., 1990; Sansom and Weinstein, 2000; Pajor and Randolph, 2005). As in our previous studies, the C270A mutant provides the scaffold for subsequent cysteine introduction as a result of its insensitivity to methanethiosulfonate (MTS) reagents. Therefore, the present study assesses MTS sensitivity of 26 consecutive cysteine mutants introduced along TM6 of hASBT, thereby providing novel insight into the molecular workings of the ASBT translocation cycle. We demonstrate a functional prerequisite for TM6 helical flexibility in global conformational changes to protein structure, leading to substrate turnover and the putative involvement of TM6 amino acids in lining portions of the permeation pathway.

Materials. [3H]Taurocholic acid (0.2 Ci/mmol) was purchased from American Radiolabeled Chemicals, Inc., (St. Louis, MO); Taurocholic acid (TCA) and glycodeoxycholic acid (GDCA) were from Sigma (St. Louis, MO); and sulfosuccinimidyl-2 (biotinamido)ethyl-1,3-dithiopropionate (sulfo-NHS-SS-biotin) was from Pierce Chemical Co. (Rockford, IL). MTS reagents (2-aminoethyl)-methanethiosulfonate (MTSEA), [2-(trimethylammonium) ethyl] methanethiosulfonate (MTSET), and methanethiosulfonate ethylsulfonate (MTSES) were from Toronto Research Chemicals, Inc. (North York, ON, Canada). Cell culture media and supplies were obtained from Invitrogen (Carlsbad, CA). All other reagents and chemicals were of highest purity available commercially.

Cell Culture and Transient Transfections. COS-1 cells (American Type Culture Collection, Manassas, VA) were maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 4.5 g/l glucose, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen, Carlsbad, CA) at 37°C in a humidified atmosphere with 5% CO2. Transient transfections were performed as described previously (Banerjee et al., 2005).

Site-Directed Mutagenesis. Site-directed mutations were incorporated into hASBT cDNA using the Quik Change site-directed mutagenesis kit from Stratagene (La Jolla, CA) and mutagenesis primers custom-synthesized and purchased from Sigma Genosys (St. Louis, MO). Plasmid purifications were performed using a kit from QIAGEN (Valencia, CA) and amino acid substitutions confirmed via DNA sequencing using an ABI 3700 DNA analyzer (Applied Biosystems, Foster City, CA) at the Plant-Microbe Genomics Facility of the Ohio State University (Columbus, OH).

Uptake Assay and Protein Membrane Expression. Initial rates of transport for each mutant were determined in transiently transfected COS-1 cells incubated in modified Hanks' balanced salt solution (MHBSS), pH 7.4, uptake buffer containing 5.0 µM[3H]TCA at 37°C for 12 min. We have demonstrated that this uptake period ensures linear steady-state kinetics in conjunction with an optimal signal-to-noise ratio for subsequent [3H]TCA analysis via liquid scintillation counting (Banerjee et al., 2005; Banerjee and Swaan, 2006; Hussainzada et al., 2006). Uptake was halted by a series of washes with ice-cold Dulbecco's phosphate-buffered saline, pH 7.4, containing 0.2% fatty acid free bovine serum albumin and 0.5 mM TCA. Cells were lysed in 350 µl of 1 N NaOH and subjected to liquid scintillation counting using an LS6500 liquid scintillation counter (Beckmann Coulter, Inc., Fullerton, CA) and total protein quantification using the Bradford protein assay (Bio-Rad Laboratories, Hercules, CA). Uptake activity was calculated as picomoles of [3H]TCA internalized per minute per milligram of protein.

Protein expression was determined by washing transiently transfected COS-1 cells in PBS followed by lysis in 0.2 ml of lysis buffer B (25 mM Tris, pH 7.4, 300 mM NaCl, 1 mM CaCl2, 1% Triton X-100, and 0.5% Sigma Protease Inhibitor Cocktail). Cell lysates were separated on a 12.5% SDS-polyacrylamide gel and transferred onto an Immuno-Blot polyvinylidene difluoride membrane (Bio-Rad Laboratories). Blots were probed with rabbit anti-ASBT primary antibody (1:1000) and visualized using goat anti-rabbit IgG/horseradish peroxidase-conjugated secondary antibody with chemiluminescent detection (ECL Plus Western Blot kit; GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK). Levels of cell surface protein expression were measured via biotin labeling, wherein transiently transfected COS-1 cells were incubated with sulfo-NHS-SS-biotin for 30 min at room temperature (Wong et al., 1995; Mitchell et al., 2004). After several washes with PBS containing 0.1 mM CaCl2 and 1.0 mM MgCl2, cells were disrupted with lysis buffer B at 4°C for 20 min (Zhang et al., 2004), and biotinylated proteins were recovered overnight at 4°C using 100 µl of streptavidin agarose beads. Samples were eluted with SDS-polyacrylamide gel electrophoresis buffer and immunoblotting performed as described above. Blots were probed for positive and negative controls, the plasma membrane marker -integrin (150 kDa) and the endoplasmic reticulum membrane protein calnexin (90 kDa), respectively, to assess the integrity of the biotinylation procedure (calnexin; data not shown). Relative hASBT membrane expression was standardized to integrin expression and quantified via densitometry as described previously (Hussainzada et al., 2006).

MTS Inhibition Studies. Sensitivity of mutants to charged, membrane-impermeant MTS reagents was determined by preincubation of transiently transfected COS-1 cells with either 1 mM MT-SES, MTSET, or MTSEA for 10 min at room temperature. After MTS treatment, cells were washed twice in modified Hanks' balanced salt solution (Sigma) followed by [3H]TCA uptake as described above. All MTS solutions were freshly prepared before each study because of the short aqueous half-life of these MTS reagents.

Cation and Substrate Protection Assays. To determine whether the presence or absence of Na+ and/or bile acid substrates alters MT-SEA labeling, transiently transfected COS-1 cells were washed twice in 1x PBS, pH 7.4, followed by coincubation with equal concentrations of MTSEA and GDCA (1 mM) prepared either in MHBSS or Na+-free buffer (MHBSS except choline chloride entirely substitutes NaCl) for 10 min at room temperature. After preincubation treatments, cells were washed twice in either MHBSS, pH 7.4, or Na+-free buffer and additionally equilibrated for 15 min at 37°C in these buffers followed by determination of [3H]TCA uptake as described above. All control wells were treated identically. For each mutant transporter, uptake values were determined by taking a ratio of mutant uptake at each experimental condition versus mutant uptake for its respective unmodified control. We normalize mutant ratios to C270A by expressing mutant ratios for each condition as a percentage of C270A ratios calculated in the same manner.

Sodium Activation. Measurement of [3H]TCA uptake at equilibrative extracellular Na + concentrations (12 mM; i.e., at equilibrium with cytosolic [Na+]) was performed (uptake conducted as described above; choline chloride used as equimolar NaCl replacement) and expressed as a ratio of uptake at physiological (137 mM) Na+ concentrations to determine overall sensitivity of each mutant to the presence/absence of Na+. In theory, Na+ ratios equal to 1 imply little measurable difference in transporter activity despite the scarcity of Na+ ions, whereas fractions less than 1 indicate a greater necessity for physiological Na+ concentrations for proper transport function of a mutant transporter.

Data Analysis. For each mutant, data are represented as mean ± S.D. of at least three different experiments with triplicate measurements. Data analysis was performed with Prism 4.0 (GraphPad Software, Inc., San Diego, CA) using analysis of variance with Dunnett's post hoc test. Data were considered statistically significant at p 0.05.

Fig. 1. Multiple sequence alignment of TM6 amino acids. A, secondary structure model of the last three transmembrane domains (TMDs) of hASBT according to the 7TM model. Roman numerals indicate flanking TMD, whereas TMD 6 amino acids are represented by gray circles inscribed with amino acid identity and position. Phospholipids of plasma membrane represented by circle (polar phosphate head group) with two tails (hydrophobic lipids). Top, exofacial; bottom, cytosolic. B, sequence alignment of amino acids 227 to 253, putatively forming TM6, for all known ASBT paralogs. Sequences were retrieved from GeneBank and aligned via the MULTALIN routine with annotation performed via the MPSA program. Shaded regions denote complete amino acid conservation among all species. Amino acid positioning relative to human ASBT is indicated by numbering on top. Bottom line indicates primary consensus for the TM6 region.

Cysteine Scan of TM6. Based on our topology model published previously (Zhang et al., 2004; Banerjee and Swaan, 2006), residues spanning Trp227 to Tyr253 are predicted to constitute TM6 of hASBT (Fig. 1A). High-sequence homology is observed among various evolutionarily diverse species for this protein region (Fig. 1B), which lies in intimate proximity to critical protein regions described previously (Hallén et al., 2000, 2002; Kramer et al., 2001; Hussainzada et al., 2006). Systematic cysteine substitutions were incorporated along TM6 followed by structural and functional analysis of mutant transporters.

Transport Activity and Membrane Expression of Cysteine Mutants. Because of its low background levels of bile acid transport (Hussainzada et al., 2006), the COS-1 cell line was used to transiently express all TM6 mutant transporters. Surface biotin labeling of membrane-expressed proteins was accomplished using the membrane-impermeant sulfo-NHS-SS-biotin and quantified via densitometry of protein bands (Fig. 2B). ASBT bands for each sample were standardized to an internal control (-integrin) and expressed as a percentage of C270A intensity (Fig. 2C). Initial transport activities (Fig. 2A) were then normalized to relative membrane expression for each mutant transporter.

Fig. 2. [3H]TCA uptake activity and membrane expression of TM6 cysteine mutants. A, uptake of [3H]TCA was measured in COS-1 cells as described under Materials and Methods and expressed as a percentage of the parental transporter C270A. B, intact transfected COS-1 cells were treated withsulfo-NHS-SS-biotin as described under Materials and Methods followed by Western blot processing. Blots were probed with the anti-hASBT antibody (1:30,000 dilution) followed by horseradish peroxide-linked anti-rabbit immunoglobin (1:2000 dilution). Each blot was probed for the internal plasma membrane marker -integrin (150 kDa) and the absence of calnexin (90 kDa) (data not shown), an endoplasmic reticulum membrane protein representing the negative control in the biotinylated fractions. Marker lanes are shown on the left side of the individual blots. Mature glycosylated hASBT visualizes as the 41-kDa band, whereas the lower 38-kDa band (not shown) represents the unglycosylated species. C, densitometric analysis for cysteine mutants normalized to internal marker (-integrin) and represented as a percentage of C270A parent. D, [3H]TCA uptake activity normalized to relative cell surface expression. Bars represent mean ± S.D. of three separate experiments with ***, p 0.001; **, p 0.01; and *, p 0.05, respectively, using analysis of variance with Dunnett's post hoc analysis.

After data normalization (Fig. 2D), most TM6 mutants retained appreciable levels of activity, except for five mutants either severely hampered (I229C, G249C) or inactivated (P234C, G237C, G241C) upon cysteine substitution. Only P234C lacked expression in biotinylated (Fig. 2B) and whole-cell (data not shown) extracts. This may be due to disruptions in protein synthesis, but more likely, alterations in protein folding and stability occur that induce rapid protein degradation via endoplasmic reticulum-associated machinery. It is interesting that all five residues are conserved among known species of ASBT (Fig. 1B), suggesting primary roles in transport function that necessitate preservation of these amino acids. Because of their low activity levels, these mutants were excluded from further studies.

TM6 Mutants Demonstrated Substantial Na+ Sensitivity. AsaNa+ cotransporter, ASBT activity relies on proper recognition, binding, and translocation of two Na+ ions per one bile acid molecule (Weinman et al., 1998). Thus, we examined the consequences of equilibrative extracellular Na+ concentrations upon mutant activity. For each mutant, the ratio of transport rates at equilibrative (12 mM) versus physiological (137 mM) Na+ concentrations was calculated and expressed as a percentage of the C270A Na+ ratio. This experimental scheme may uncover hidden functional defects in mutants otherwise unaffected by cysteine mutation. Therefore, the C270A parental construct displays a Na+ ratio of 0.70 ± 0.04 (data not shown), indicating minimal consequences to transporter function upon alanine substitution at the native cysteine residue. In contrast, significant decreases in activity were observed for the majority (64%) of TM6 cysteine mutants. Of 21 mutants assayed, 14 demonstrated hampered uptake rates at equilibrative [Na+] (Fig. 3). EL3 cysteine mutants from our earlier study also exhibited similarly extensive Na+ sensitivity (Banerjee et al., 2008), wherein uptake activities of 90% of assayed mutants were susceptible to equilibrative [Na+]. Because EL3 residues putatively form Na+ interaction sites (Banerjee et al., 2008), the widespread Na+-dependence observed for TM6 mutants implies their close proximity to such Na+ interaction sites and lends credence to TM6 participation during Na+ permeation.

Fig. 3. Sodium sensitivity of cysteine mutants. COS-1 cells expressing mutant transporters were incubated in uptake medium (5 µM[3H]TCA) containing low (12 mM) or physiological (137 mM) Na+ concentrations as described under Materials and Methods. Sodium ratios were calculated for each mutant as the quotient of activity at 12 versus 137 mM [Na+] and expressed as a percentage of C270A. Bars represent mean activity ± S.D. (n = 3). *, p 0.05.

Substrate and Cation Binding Modulate Accessibility of Cysteine Mutants to MTS Modification. Both positively and negatively charged MTS reagents were used to probe the solvent accessibility of 21 cysteine mutants demonstrating measurable uptake activity. Intact monolayers of COS-1 cells expressing mutant transporters were preincubated with 1.0 mM concentration of either MTSES, MTSET, or MTSEA followed by uptake assessments. MTSET (109 Å3) and MTSES (90 Å3) exhibited similar inhibition profiles, in which activities of only mutants V235C, S239C, F242C, and R246C were significantly reduced (data not shown), suggesting minimal electrostatic effects in accessibility at those sites. The remainder of the TM6 mutants assayed were either inaccessible to these MTS reagents, or their modification was functionally silent. Incubation with the relatively smaller MTSEA (69 Å3) inhibited uptake at sites accessed by the larger MTS reagents (i.e., V235C, S239C, F242C, and R246C; Fig. 4) and at two additional sites (A248C and Y253C; Fig. 4).

Fig. 4. Cation and substrate protection of TM6 cysteine mutants. Transiently transfected COS-1 cells expressing TM6 cysteine mutants were preincubated in buffer, pH 7.4, containing 1 mM MTSEA and either 137 mM NaCl (), 137 mM NaCl and 1 mM GDCA (gray bar); 137 mM choline chloride (dark gray bar); or 137 mM choline chloride and 1 mM GDCA () and followed by [3H]TCA uptake as described under Materials and Methods. Choline chloride does not activate the transporter and provides equimolar replacement for NaCl. All control wells were treated identically. Bars represent mean ± S.D. of at least three separate measurements. Data are expressed as a percentage of C270A values for each condition as described under Materials and Methods. Student's t test analysis performed with *, p < 0.05, and **, p < 0.01.

Because MTSEA (1 mM) application produced the most pronounced inhibition of the three MTS reagents used (similar to our previous study with TM7; Hussainzada et al., 2006), the effects of Na+ and bile acid substrate on MTSEA accessibility of mutants V235C, S239C, F242C, R246C, A248C, and Y253C was evaluated. Our previous studies with EL3 and TM7 have shown that coincubation of MTS reagents with bile acids and/or removal of Na+ from the preincubation buffer caused a reversal of the inhibitory effect observed with MTS incubation alone (Hussainzada et al., 2006; Banerjee et al., 2008). Likewise, in the present study, all TM6 mutants inhibited by MTSEA (1 mM) demonstrated significant uptake recovery when MTSEA incubation was performed in the absence of Na+ and/or presence of 1.0 mM glycodeoxycholic acid (GDCA) (Fig. 4). In particular, the removal of Na+ from the MTSEA preincubation medium significantly restored transport activity for mutants V235C, S239C, F242C, R246C, and A248C, whereas coincubation with GDCA (Km = 2.0 ± 0.4 µM) significantly protected mutants F242C and A248C. Concurrent removal of Na+ and the addition of the high-affinity substrate GDCA (1 mM) resulted in significant MTSEA protection for mutants V235C, S239C, F242C, and R246C. In all cases, mutant activities were restored to control (C270A) levels (within standard deviation; Fig. 4).

Although mechanistic details of the ASBT translocation cycle are as yet unresolved, ordered binding of ligands followed by translocation probably occurs, similar to many other Na+-coupled transporters (Quick and Jung, 1997; Jung, 2001; Pajor and Randolph, 2005; Zhang and Rudnick, 2005). In this scenario, the ordered binding of Na+ and bile acids would trigger the protein to assume various discrete structural conformations, eventually leading to carrier reorientation within membrane leaflets and substrate turnover. Within TM6, the lack of Na+ binding events (simulated by substitution of Na+ with choline+ in preincubation buffers) significantly decreased MT-SEA modification rates for all sites (Fig. 4), suggesting that protein conformational states assumed before the binding of Na+ occlude these thiol groups from subsequent modification. Furthermore, the binding of bile acid substrate (GDCA) in either the presence or absence of Na+ also triggers protein conformations that significantly decrease MTSEA access to all sites (Fig. 4). Because of the close association between TM6 and protein regions previously implicated during ligand binding and translocation (i.e., TM7 and EL3), the observed substrate protection may result from 1) occlusion of these sites via conformational changes; 2) the physical presence of substrates preventing access; or 3) a combination of both scenarios. The alternating accessibility of TM6 sites to thiol modification suggests the first scenario, whereas the restoration of mutant activity to control levels via substrate protection infers the second scenario. It is likely that a union of both situations prevails, in which TM6 residues may line portions of the permeation pore and also transduce conformational changes resulting from ligand interactions at adjacent protein regions (TM7, EL3), although further studies are needed to unequivocally conclude the origin of substrate protection. However, it is noteworthy that our data highlight a trend toward protection from MTS modification in the absence of Na+, which is entirely plausible given that EL3 residues contain putative Na+ interaction sites (Banerjee et al., 2008). It may be that the lack of Na+ binding at EL3 regions prevents downstream conformational changes that "open" TM6 residues to MTS modification. Overall, we conclude that TM6 amino acids probably line portions of the substrate permeation route due to their spatial proximity with EL3 and TM7 residues and their solvent-accessibility profile.

Mutation of Pro234 Affects Both Transporter Expression and Function. Because the P234C double mutant (C270A/P234C) lacked expression both at the plasma membrane (Fig. 2B) and in whole-cell extracts (data not shown), additional replacements were made at this site to determine whether Pro234 makes functional contributions during the hASBT transport cycle. Using the wild-type (WT) species as the scaffold, alanine, glycine, and cysteine replacements were incorporated and analyzed with respect to uptake activity and membrane expression. As expected from our results using the kinetically similar C270A template (Banerjee et al., 2005), the P234C mutant constructed against the WT background lacked expression both in membrane (Fig. 5B) and whole-cell (Fig. 5C) extracts, confirming that cysteine replacement at this position affects transporter expression levels irrespective of the mutational template used. Glycine replacement (P234G) also seems to disrupt protein expression, resulting in minimal transporter expression in membrane (Fig. 5B) and whole-cell (Fig. 5C) extracts. In contrast, the alanine mutant (P234A) displayed plasma membrane expression, albeit at reduced levels (30% of C270A levels; Fig. 5A). After normalization to cell surface expression levels, uptake function of the P234A mutant remained severely inhibited (Fig. 5A), suggesting functional and structural impairments to transporter function. Defective trafficking to the plasma membrane may account for the lowered membrane expression of the P234A mutant, because the ratio of its plasma membrane expression to whole-cell expression is approximately half of the similar ratio for the WT species (i.e., 0.452 versus 0.814); however, further studies are needed for unequivocal evidence. We conclude that Pro234 participates in both protein expression and transporter function, confirming the overall importance of this atypical amino acid.

Fig. 5. [3H]TCA uptake and membrane expression of Pro234 mutants. A, the initial [3H]TCA uptake (gray bars), relative intensity of immunoblotting (dark gray bars), and normalized uptake activities (black bars) for Pro234 single and double mutants and their parental templates are depicted. Bars represent mean ± S.D. of at least three separate measurements. Data are expressed as a percentage of C270A values for each condition. Student's t test analysis was performed with *, p < 0.05, and **, p < 0.01. Activities of the Pro234 constructs are statistically different from both WT and C270A (p < 0.01); however, asterisks (**) have been omitted for visual clarity. B, intact transfected COS-1 cells were treated with sulfo-NHS-SS-biotin as described under Materials and Methods followed by Western blot processing. Marker lanes are shown on the left side of the individual blots. Mature glycosylated hASBT visualizes as the 41-kDa band, whereas the lower 38-kDa band (not shown) represents the unglycosylated species. C, immunoblotting of whole-cell extracts from transfected COS-1 cells as described under Materials and Methods showing glycosylated (41 kDa) and unglycosylated (38 kDa) hASBT species.

作者单位：Department of Pharmaceutical Sciences, University of Maryland, Baltimore, Maryland

【摘要】 Myosin VI is an actin motor that moves to the minus end of the polarized actin filament, a direction opposite to all other characterized myosins. Using expression microarrays, we identified myosin VI as one of the top genes that demonstrated cancer-specific overexpression in clinical prostate specimens. Protein expression of myosin VI was subsequently analyzed in arrayed prostate tissues from 240 patients. Notably, medium-grade prostate cancers demonstrated the most consistent cancer-specific myosin VI protein overexpression, whereas prostate cancers associated with more aggressive histological features continued to overexpress myosin VI but to a lesser extent. Myosin VI protein expression in cell lines positively correlated with the presence of androgen receptor. Small interference RNA-mediated myosin VI knockdown in the LNCaP human prostate cancer cell line resulted in impaired in vitro migration and soft-agar colony formation. Depletion of myosin VI expression was also accompanied by global gene expression changes reflective of attenuated tumorigenic potential, as marked by a nearly 10-fold induction of TXNIP (VDUP1), a tumor suppressor with decreased expression in prostate cancer specimens. These results support that myosin VI is critical in maintaining the malignant properties of the majority of human prostate cancers diagnosed today.
--------------------------------------------------------------------------------
Myosins are defined as actin-dependent Mg2+ ATPases that use the energy derived from ATP hydrolysis to move along the actin filaments within the cell.1 Structurally, myosins have a common domain organization consisting of a conserved N-terminal actin binding and ATPase domain (motor or head domain), a neck region containing IQ motifs that bind to myosin light chains, and a C-terminal tail domain for specific cargo binding.2 In the human genome, there are 40 myosin genes, representing 12 classes of actin motors that mainly participate in actin-based cellular processes.1,2 Only the class II myosins are known to form bipolar filaments that are essential for well-characterized contractile functions. The remaining classes of myosins are so-named unconventional myosins1,2 that are generally thought to function in nonmuscle cells as actin-bound monomers or dimers. Although not well characterized in terms of the precise mechanism, unconventional myosins have been implicated in F-actin-mediated cellular functions such as cell motility, vesicular trafficking, intracellular transport of macromolecules, and possibly regulation of signal transduction.2,3
The class VI unconventional myosin was initially identified and partially characterized in Drosophila and pig.4,5 In most organisms including human, a single gene encodes the class VI unconventional myosin. Myosin VI is a unique member of the myosin superfamily.6,7 Primarily because of a 53-amino acid insertion between the motor and the neck domain, myosin VI moves to the pointed/minus end of the polarized actin filament, a direction opposite to all other myosins characterized to date.8,9 Because actin filaments are believed to orient their pointed/minus ends away from the plasma membrane and internal organelles,6 the unique motor direction of myosin VI is potentially linked mechanistically to its functional roles in endocytosis (transport of vesicles away from the plasma membrane),10 secretion (transport of vesicles away from the Golgi),11 and cell migration (pushing of the barbed/plus ends of F-actin against the cell mem- brane).12-14
Although the role of actin motors (myosins) in human cancer is generally poorly documented, an intriguing connection between myosin VI and human cancer was recently reported.14 Based on the initial observation that myosin VI is required in border cell migration during Drosophila ovary development,12 Yoshida and colleagues14 examined protein expression of myosin VI in human ovarian cancers and discovered a functional link between myosin VI expression and aggressive ovarian cancer. In the present study, we initially discovered an unusually consistent cancer-specific overexpression of myosin VI mRNA through global gene expression analysis that emphasized the comparison between normal prostate epithelium and cancerous acini. Further, the role of myosin VI in human prostate cancer was investigated through immunohistochemical analysis in a cohort of 240 patients, as well as functional studies in human prostate cancer cell lines.

【关键词】 prostate

Materials and Methods

Human Prostate Tissues for Expression Microarrays

Prostate tissue samples used for cDNA microarray analysis were fresh frozen specimens collected at the time of prostate surgery from 1993 to 2000 at the Johns Hopkins Hospital. Tissue specimens used in this study were from nine patients undergoing surgery for symptomatic benign prostatic hyperplasia (BPH) and 25 patients undergoing radical prostatectomy for prostate cancer. Established procedures15 were followed for sample selection and processing. A total of 59 specimens were processed because normal-tumor paired tissues were retrieved from each of the 25 radical prostatectomy cases. Cryosections were cut from trimmed blocks enriched for tissues of interest before downstream RNA extraction. The first and last section from each sample was reserved for pathological confirmation and visual estimation of the percentage of epithelium. This study was approved by the Institutional Review Board at Johns Hopkins Medical Institutions.

Human Prostate Tissues for Immunohistochemistry

All prostate specimens used for immunohistochemical analysis were radical prostatectomy samples selected from the surgical pathology files at the Johns Hopkins Department of Pathology with Institutional Review Board approval. Tissue microarrays (TMAs) were constructed as previously described.16 Six high-density TMAs, each containing surgical prostate tissues from 40 cases (240 cases in total), were used for immunohistochemical staining. Each case was represented by eight cores (0.6 mm in diameter) that were predominantly matched normal and cancer tissues but may also have been high-grade prostatic intraepithelial neoplasia (HGPIN) and proliferative inflammatory atrophy (PIA) lesions.16 Standard tissue sections were selected and processed also as previously described.16

Expression Microarrays

Printed glass cDNA microarrays were used throughout the study. For prostate tissue profiling, microarrays containing 11,904 human expression sequence tags were used. Expression sequence tags were selected from human IMAGE clone plate sets, based on relative enrichment of annotated genes within the plate, and supplemented by six plates (576 clones) of custom arrayed IMAGE clones selected based on relevance to prostate biology after an extensive literature search. For profiling in cell lines, a recent version of cDNA microarrays containing 20,344 human expression sequence tags was used, after integration of additional plates enriched for annotated genes.

Gene Expression Analysis

The experimental design, total RNA extraction, labeling, hybridization, image analysis, and data analysis were modified based on the protocols described previously.15 Total RNA samples extracted from tissues or cultured cells were amplified once using the MessageAmp aRNA kit (Ambion, Austin, TX) using an input of 500 ng of total RNA, and labeled by direct incorporation of Cy3-dUTP (Amersham Pharmacia, Piscataway, NJ) in a reverse transcription reaction using random primers and Superscript II reverse transcriptase (Invitrogen, Carlsbad, California). For prostate tissue profiling, expression profiles were generated by co-hybridization of each of the 59 Cy3-labeled probes with a Cy5-labeled common reference sample, prepared from a pool of two BPH specimens as described15 and its RNA similarly amplified. For expression profiling of cultured cells, a common reference of nontreated LNCaP cells was used. The expression profile for each sample was represented as normalized ratios of sample/reference for all genes represented on the array. For expression data from tissues, genes associated with unreliable data points, defined as a mean fluorescence intensity less than 1000, were excluded from further analysis. To select genes whose expression varied most across the 59 samples, we applied a stringent filtration procedure based on the criteria of at least twofold expression change relative to the median in at least 15 samples to yield a list of 275 genes. An agglomerative hierarchical two-way clustering algorithm based on Euclidean distance measures15 was used to cluster the samples and the 275 genes. Statistical analyses of the differentially expressed genes were performed on expression data derived from the 59 tissue specimens and downloaded prostate tissue expression data from Lapointe and colleagues,17 using weighted gene analysis as described.15 For expression data derived from cultured cells, we first excluded unreliable data points by the same cutoff at mean intensity of 1000. The weighted gene analysis based on a modified distance-based w metric15 was again used to determine the extent of differential expression between siRNA treated samples and nontreated cells.

Immunohistochemical Staining

Immunohistochemical staining was performed using the Envision+ kit (DAKO Corp., Carpinteria, CA) as described.18 For myosin VI staining, a 1:400 dilution of the primary antibody (a gift from Mark Mooseker, Yale University, New Haven, CT)5 was used. For TMAs, we also performed keratin 8 staining to assist in automated scoring analysis as described.18 Sections of TMAs adjacent to those stained for myosin VI were stained for keratin 8 using a 1:800 dilution of the anti-CK8 antibody (InnoGenex, San Ramon, CA). For immunohistochemical staining in standard tissue slides, double labeling of -methylacyl-CoA racemase and p63 (AMACR/p63) were performed as described16 in sections adjacent to those used for myosin VI staining.

TMA Analysis

To avoid human bias during the assessment of immunohistochemical staining, we used the Chromavision ACIS II system (Clarient, Inc., San Juan Capistrano, CA), for semiautomated scoring.18 This approach uses two adjacent TMA slides in which one slide is stained with keratin 8 to determine epithelial content and the other is stained for myosin VI. For automated analysis we excluded TMA spots with a mixed diagnosis (mixed epithelial cells of normal/cancer/other lesions). Expression level of myosin VI within each individual TMA spot was evaluated by automatic and parallel calculation of pixel numbers in three staining categories (weak, moderate, and strong staining), yielding a composite score based on a previously described formula16 for each spot. The scores were normalized to the total brown pixel numbers for keratin 8 in the adjacent section to account for differential epithelial content across the TMA spots. Tissue histology in all TMA spots was re-examined by a pathologist (A.M.D.) in the adjacent hematoxylin and eosin (H&E)-stained TMA slides. Only spots annotated as containing a single diagnosis (no mixed normal/tumor/other lesions) were selected for further analysis. A nonparametric Wilcoxon??s rank-sum test was performed to test the statistical significance in expression levels between groups of interest. For Figure 2d , because multiple array spots are represented for each tissue type (normal or tumor) from each patient, the averaged values were used.

Figure 2. TMA analysis of myosin VI. a: Confirmation of antibody specificity. Lanes 1 and 3: normal prostate tissues; lanes 2 and 4: prostate cancer tissues; lane 5: liver tissue as negative control; and lane 6: kidney tissue as positive control. b: Representative staining pattern of myosin VI in normal, Gleason grade 3 (medium grade), Gleason grade 4 (high grade), and prostate tissues with mixed normal and cancerous epithelial cells, as individually annotated. c: Box plots of myosin VI staining scores in normal (n = 665), PIA (n = 76), HGPIN (n = 18), and cancer tissues (n = 592). The score values were normalized to the epithelial content in each spot and displayed on the y axis. Each box is lined at lower quartile, median, and upper quartile score values for each group. The + symbols mark data values beyond the ends of the whiskers. d: Box plots of myosin VI staining scores. Patients were stratified by pathological Gleason scores (x axis), and myosin VI staining scores between normal and cancer samples were compared within each group. Number of patients represented in each category (normal versus cancer): Gleason 6 (80 versus 71), Gleason 7 (81 versus 61), Gleason 8 to 10 (34 versus 34).

Myosin VI Knockdown

The target sequences used to silence myosin VI expression were MYO6-siRNA-1, 5'-CCGCAAAAGTCCTGAGTAC-3', and MYO6-siRNA-2, 5'-AGCTTGATCTCTTCCGGGT-3' (Qiagen-Xeragon, Germantown, MD). The target sequence of nonsilencing control siRNA was 5'-TTCTCCGAACGTGTCACGT-3' (Qiagen-Xeragon). LNCaP cells were transfected with siRNA duplexes by using Lipofectamine 2000 reagents (Invitrogen Corp.). Efficiency of myosin VI knockdown was tested by Western blot at different concentrations and various time points. Optimal gene knockdown conditions in LNCaP cells were achieved using 120 nmol/L siRNA at 96 hours after transfection.

Western Blot Analysis

Cultured cells and frozen human prostate tissues were subjected to standard Western blot analysis as described.16 For myosin VI detection, a polyclonal rabbit antibody (1:1000) raised against a C-terminal myosin VI peptide (Sigma, St. Louis, MO) was used. For VDUP1 detection, a monoclonal antibody (1:1000) was used (MBL International Cooperation, Woburn, MA). A monoclonal antibody (clone 36) for E-cadherin (BD Biosciences, San Jose, CA) was used at 1:3000 dilution. ß-Actin was detected using a monoclonal antibody (AC-15) at 1:5000 dilution (Sigma).

Proliferation Assay

LNCaP cells that had been transfected 24 hours earlier with siRNA or without siRNA were seeded into a 96-well plate (8000 cells/well). The number of viable cells was determined daily with CellTiter 96 Aqueous nonradioactive cell proliferation assay (Promega, Madison, WI). In brief, 20 µl of the combined MTS/PMS solution was added to each well of the 96-well assay plate containing cells in 100 µl of culture medium. Optical density at 490 nm was recorded after 2 hours using an enzyme-linked immunosorbent assay plate reader.

Cell Migration Assay

For the in vitro migration assay, 24-well Costar transwell chambers (Corning Inc., Corning, NY) with 8-µm pore membrane were used. The under surface of the membrane was coated with fibronectin. LNCaP cells that had been transfected 96 hours earlier with and without siRNA were seeded (5 x 104/well) to the upper chambers and allowed to migrate for 16 hours at 37??C. At the end of the assay, after removal of nonmigratory cells on the upper surface, the migrated cells on the under surface were fixed and stained for 20 minutes with 0.5% crystal violet in 10% ethanol. Stained cells were eluted with 10% acetic acid, and the absorbance was determined. One-tailed Student??s t-test was used to assess the statistical significance (P < 0.05 considered to be significant).

Soft Agar Assay

The soft agar assay tests the anchorage-independent growth in vitro. In brief, 1 x 104 LNCaP cells that had been transfected 24 hours earlier with or without siRNA were resuspended with 3 ml of 0.3% agar (Invitrogen Corp.) in RPMI 1640 containing 10% fetal bovine serum. The cell-agar mixture was immediately seeded into six-well plates coated with 0.6% agar in RPMI 1640 with 10% fetal bovine serum. Culture media was replaced every 3 days. Colonies were stained with crystal violet as described above, at 2 weeks after seeding.

Myosin VI mRNA Overexpression in Human Prostate Cancer

We generated gene expression profiles from 59 histologically characterized human prostate tissues (raw data available at http://www.oncomine.org). To highlight the expression differences across the samples, we applied an unbiased/unsupervised procedure (see Materials and Methods) to select 275 genes with expression that varied most across the 59 samples. A two-way clustering analysis was performed using this set of genes across the 59 samples, breaking down to 9 BPH (B1 to B9), 25 normal (N1 to N25), and 25 prostate cancer tissues (T1 to T25) that were matched with the normal prostate samples by number (Figure 1a) . As shown, samples formed clusters based on their identities with few exceptions, and genes formed clusters based on differential expression patterns across the samples. We highlighted the identities of a cluster of 21 genes that demonstrated cancer-specific overexpression patterns (Figure 1a , fully annotated heatmap in Supplemental Figure 1 at http://ajp.amjpathol.org). Myosin VI clustered with many previously characterized prostate cancer markers, including prostate cancer antigen 3 (PCA3, DD3),19 AMACR,16 single-minded 2 (SIM2),20 hepsin,21 and TARP.22 Comparison of myosin VI expression ratios across the samples showed all but one of the 25 paired normal/cancer samples with higher expression in the cancer sample than in the paired normal sample (Figure 1b) . On average, cancer samples (4.37 ?? 2.05) showed a 3.7-fold higher expression of myosin VI mRNA when compared with the normal samples (1.20 ?? 0.24), and a 4.6-fold increase when compared with the BPH samples (0.94 ?? 0.10).

Figure 1. Myosin VI is a novel prostate cancer marker identified in microarray analysis of surgical human prostate specimens. a: Heatmap representation of gene expression data for 59 histologically characterized samples. Columns represent samples, including 25 normal (N1 to N25), 9 BPH (B1 to B9), and 25 cancerous prostate tissues (T1 to T25). Rows represent genes. Normalized expression ratios for each gene are represented by red-green color scale, with red indicating overexpression relative to the median and green indicating underexpression relative to the median. The color bar above the color matrix denotes the sample identity, with blue marking BPH samples, green marking normal samples, and red marking cancer samples. A subcluster of genes representing those specifically overexpressed in cancer samples was shown in relation to their relative position in the color matrix. b: Comparison of myosin VI mRNA expression in 25 normal-tumor pairs. Normalized expression ratios of sample/reference were extracted from the microarray data and displayed for each individual pair of normal-tumor samples from each of the 25 cases. Green bars, normal; red bars, cancer; x axis, cases; y axis, expression ratios normalized to the common BPH reference denominator.

To further demonstrate the consistency and extent of myosin VI overexpression in human prostate cancers, we performed weighted gene analysis15 using two independent data sets, one from this study (raw data available at http://www.oncomine.org) and the other from Lapointe el al.17 As shown in Table 1 , myosin VI was consistently identified as one of the top genes based on the test scores (w scores) comparing normal and cancerous human prostate tissues, with a P value less than 10C10 in both data sets (data not shown).

Table 1. Top Ranked Genes Overexpressed in Human Prostate Cancer

TMA Analysis of Myosin VI Expression

An affinity purified polyclonal antibody against the tail domain of porcine myosin VI was used5 for immunohistochemical (IHC) analysis of myosin VI expression in human prostate cancer tissues. The antibody recognized a major band of 150-kd human myosin VI (Figure 2a) in prostate cancer tissues and a kidney tissue sample (positive control, 5) but not in normal prostate tissues or a liver sample (negative control, 5), thus confirming the binding specificity of the antibody and suitability for tissue staining. High-density TMAs were used for immunohistochemical analysis of myosin VI expression. A visual evaluation of stained TMAs confirmed the strongly positive myosin VI staining in the majority of cancerous epithelial cells but generally negative or weak staining in normal epithelium and negative staining in stromal components, as shown in representative array spots (Figure 2b) .

Semiautomated scoring analysis18 was performed for IHC data from six TMAs. We first focused on the comparison of myosin VI protein expression among four histological lesions of interest: normal epithelium, PIA, HGPIN, and cancer epithelium. After histological evaluation of individual array spots by a pathologist (A.M.D.) and exclusion of array spots with poor quality and mixed diagnosis, IHC scores were obtained from 665 normal, 76 PIA, 18 HGPIN, and 592 cancer lesions. As shown in Figure 2c , cancer tissues had significantly higher myosin VI protein expression when compared with normal and PIA lesions (P < 10C10 and P < 10C5, respectively). Interestingly, when compared with the normal tissue, myosin VI protein expression is statistically higher in the two putative premalignant lesions, PIA and HGPIN (P < 10C6 and P < 10C5, respectively), suggesting that overexpression of myosin VI is an early event during prostate carcinogenesis.

Comparative analysis of myosin VI protein expression levels between normal and cancerous tissues was performed in three groups of patients stratified by pathological Gleason scores (Figure 2d) . The majority of prostate cancers diagnosed today present Gleason scores of 6 or 7, typically containing a predominant component of grade 3 cancer that is characterized by infiltrative growth of well-formed acini (Figure 2b, B and C) . As shown in Figure 2d , patients in these categories (Gleason scores 6 or 7) demonstrated the most consistent overexpression of myosin VI in the cancer tissues when compared with normal tissues (P < 10C10 and P < 10C6, respectively). Within the group of Gleason 6 patients in particular, the median score of the cancer samples was six times higher than the median score for the normal samples. High-grade prostate cancers (Gleason score 8 to 10) typically present back-to-back fused glands or loss of glandular differentiation (Figure 2b, D) . These histologically more aggressive cancers (Gleason score 8 to 10) also showed marked overexpression of myosin VI when compared with the normal tissues (Figure 2d) (P < 0.02), although there was a decreased overall extent of cancer-specific myosin VI overexpression in comparison to medium-grade cancers (Gleason score 6 and 7) (P < 0.01). Consistent with its decreased cancer-specific expression in more aggressive cancer lesions, myosin VI levels were negatively correlated with the presence of seminal vesicle invasion and pelvic lymph node metastasis (P < 0.03) (data not shown).

IHC Analysis Using Standard Slides

Histologically defined prostate cancer presents an invasive phenotype characterized by the absence of basal cells and local stromal invasion by the cancerous acini.23 Combined staining for cytoplasmic AMACR and basal cell-specific nuclear protein p63 can be used to reliably detect such cancer lesions.16 To illustrate the spatial pattern of myosin VI protein expression in relation to the cancerous histology as well as histological details surrounding the lesions of interest, we performed AMACR/p63 and myosin VI staining in adjacent cuts of standard sections (as opposed to arrayed tissues) from cases that were myosin VI-positive. As shown in Figure 3 , myosin VI staining patterns were highly correlated with a readily discernible cancerous morphology, in tissues where normal and cancerous histology are both present (Figure 3, A and B) and even adjoined within the same acini (Figure 3, C and D) . Intense myosin VI staining (Figure 3, B and D) was invariably seen in cancer lesions, as marked by positive cytoplasmic AMACR and negative nuclear p63 staining in adjacent sections (Figure 3, A and C) , whereas normal epithelial cells with intact basal cell layer and negative AMACR staining (Figure 3, A and C) were weakly positive or negative for myosin VI (Figure 3, B and D) .

Figure 3. Myosin VI staining correlates with cancer morphology and stromal invasion in standard prostate tissue sections. Sections were double stained for AMACR/p63 (A and C), and adjacent cuts from the same paraffin blocks were stained for myosin VI (B and D). Black arrows: positive p63 nuclear staining that is specific for normal basal cells and absent in cancerous lesions; red arrows: positive cytoplasmic AMACR staining that is highly specific for prostate cancer cells.

Western Blot Analysis of Myosin VI in Cell Lines

To establish an in vitro cell line model for functional studies, we examined protein expression of myosin VI in a panel of five human prostate cancer cell lines (Figure 4a) . LNCaP cells were originally isolated from pelvic lymph node metastases of human prostate cancer. These cells retain many biological features of human prostate cancer including relatively slow growth and androgen sensitivity. As shown in Figure 4a , the LNCaP cell line expressed the most abundant myosin VI protein expression, followed by two other androgen receptor-positive lines (LAPC-4 and CWR22Rv1) that were derived from xenographs of locally advanced human prostate cancer. PC-3 and Du145 lines were established from androgen-refractory distant metastasis of human prostate cancer and expressed less myosin VI than the androgen-sensitive cancer cell lines. The expression pattern of myosin VI in cultured human prostate cells is again in line with the expression changes observed in clinical tissue specimens, in which there was a general trend of decreased cancer-specific myosin VI expression in more aggressive cancers.

Figure 4. Myosin VI expression and function in prostate cancer cell lines. a: Protein expression of myosin VI in five commonly used human prostate cancer cell lines. Protein levels of ß-actin were examined in the same blot and serve as loading controls. b: Myosin VI expression was inhibited by two siRNA duplexes (lanes 3 and 4), but not affected by control siRNA treatment (lane 2). Protein levels of ß-actin were examined in the same blot and serve as loading controls. c: Bar graph showing impaired LNCaP cell migration after myosin VI knockdown (siRNA-MYO6-1, siRNA-MYO6-2). Data were compiled from three replicates for each treatment conditions. *Significantly lower number of migratory cells when compared with the sham-transfected cells. d: Cell proliferation curve in a span of 5 days. Data were complied from five replicates for each treatment and time point. e: Decreased soft agar colony formation after inhibition of myosin VI expression. Data were complied from four replicates for each treatment condition. *Significantly less colonies when compared with the sham-transfected cells.

Functional Roles of Myosin VI

Because LNCaP human prostate cancer cells demonstrated the most abundant expression of myosin VI, we performed in vitro functional assays after inhibition of myosin VI expression in these cells. As shown in Figure 4b , myosin VI protein expression was dramatically decreased by both siRNA duplexes designed to target the specific degradation of myosin VI RNA (target sequences are myosin VI-specific sequences in the motor domain) but was not affected by control nonsilencing siRNA under identical conditions. No gross morphological changes were observed in cultured cells after siRNA treatment. Consistent with previous findings,14 the inhibition of myosin VI expression resulted in impaired cell migration (Figure 4c) but did not affect the proliferation rate of cells in the culture medium (Figure 4d) . However, experimental knockdown of myosin VI significantly reduced the number of soft agar colonies 14 days after inoculation (Figure 4e) , suggesting a role of myosin VI in anchorage-independent growth, a hallmark of transformed phenotype.

Global Expression Changes after Inhibition of Myosin VI Expression

Additional clues regarding the biological impact of myosin VI expression was examined by cDNA microarray analysis after myosin VI knockdown in LNCaP cells. We compared expression differences between two siRNA-transfected samples and the two control cell samples (including cells treated with nonsilencing control siRNA). Genes were ranked based on a w metric15 that measures the extent of gene expression change as a function of myosin VI knockdown (Supplemental Figure 2 at http://ajp.amjpathol.org). After myosin VI inhibition, the majority (13 of 15) of the genes (Supplemental Figure 2 at http://ajp.amjpathol.org) showed expression suppression by approximately twofold. The list of suppressed genes included myosin VI (ranked no. 5) (Supplemental Figure 2 at http://ajp.amjpathol.org), the intended target of siRNA-mediated knockdown. Exceptionally, myosin VI knockdown resulted in a nearly 10-fold increased expression for TXNIP,24 whereas no other genes in the whole dataset consistently demonstrated more than threefold expression changes in either direction.

Validation of TXNIP/VDUP1 Expression

TXNIP (thioredoxin-interacting protein 1), also named VDUP1 (vitamin D3 up-regulated protein 1), is a tumor suppressor that also participates in transcriptional repression to inactivate oncogenic signals.24 The protein expression of TXNIP was dramatically increased after inhibition of myosin VI expression in both LNCaP (Figure 5a) and CWR22Rv1 cells (data not shown), as validated by Western blot analysis. In addition, protein expression of myosin VI and TXNIP (Figure 5b) appeared to be inversely correlated in unperturbed androgen receptor-positive cell lines (LNCaP, CWR22Rv1, LAPC-4), whereas the AR-negative PC-3 and DU-145 cells did not express higher levels of TXNIP despite lower expression of myosin VI. Protein expression of TXNIP was subsequently examined in five paired normal and tumor samples from radical prostatectomy specimens. Despite the heterogeneity of the overall expression pattern, TXNIP was generally decreased in the cancer specimens when compared with their matched normal counterparts (Figure 5c) .

Figure 5. Inverse correlation between myosin VI and VDUP1 protein expression. a: Western blot analysis of TXNIP after inhibition of myosin VI expression in LNCaP cells. b: Western blot analysis of TXNIP and myosin VI in the cell lines. c: Western blot analysis of TXNIP and myosin VI in five normal-tumor pairs of human prostate specimens. Expression levels of ß-actin were used as loading controls in all analysis.

In the human prostate, normal ducts and acini are lined by a double cell layer: a flat basal cell layer oriented parallel to the basement membrane and a secretory tall columnar luminal cell layer. A defining histological feature of human prostate cancer is the complete absence of basal cells and local stromal invasion/infiltration by the cancerous acini.23 Gain of invasive potential, therefore, is required for the establishment of histologically defined human prostate cancer. Global expression analysis emphasizing the comparison between normal prostate epithelium and cancerous acini15,17,25-31 may reveal molecular alterations accompanying this critical gain of function. In this follow-up study on a top gene identified in global transcriptome analysis of surgical prostate specimens, we uncovered a novel connection between an actin motor, myosin VI, and human prostate cancer. Although myosin VI may participate in diverse cellular functions,6,7 the observed association of myosin VI expression with histological characteristics of human prostate cancer may be linked to an established role of myosin VI in cell migration.12-14 When combined with pericellular proteolysis and proliferative force, an enhanced migratory potential may facilitate cancer cell/acini invasion in the tissue environment.32

Stromal invasion is a hallmark of virtually all human cancers of epithelial origin.33 Transcriptional up-regulation of myosin VI, however, does not appear to be a universal phenomenon for all carcinomas. Based on mRNA expression data in the public database, cancer-specific myosin VI overexpression is primarily restricted in human prostate and breast cancers (http://genome-www5.stanford.edu/cgi-bin/source/sourceSearch and http://www.oncomine.org). It is unclear whether steroid hormone receptors play a role in regulating myosin VI mRNA expression. Although expression levels of myosin VI appear to correlate with androgen receptor status (Figure 4a) , it is not regulated by synthetic DHT analog R1881 (unpublished observation) nor is it affected by complete knockdown of androgen receptor in human prostate cancer cells (unpublished observation). The regulatory mechanism accounting for myosin VI overexpression in human prostate cancer is currently unknown.

A hypothesis-driven approach has found elevated myosin VI protein levels in ovarian cancers as well as a positive correlation with ovarian cancer aggressiveness both in vitro and in vivo.14 We did not observe a similar correlation between myosin VI expression and any of the clinical and pathological indicators of prostate cancer aggressiveness. There was instead a general trend of slightly decreased cancer-specific myosin VI expression in prostate cancer cases with aggressive histological and clinical features. Therefore, human cancers of different tissue origin may display different modes of regulation and different patterns of alteration in myosin VI expression. It is worth noting that advanced human prostate cancers may have acquired other properties, such as enhanced pericellular proteolysis,34 for invasion-associated functions and may have thus become less reliant on the participation of myosin VI. Moreover, more dedifferentiated prostate cancers often express reduced levels of E-cadherin,35 which may lead to decreased myosin VI expression as previously suggested.12 Therefore, it is reasonable to speculate that myosin VI may regulate coordinated movement of a cluster of cells as seen in well-differentiated, E-cadherin-positive prostate cancer lesions but may not be as critical in advanced cancers in which E-cadherin-mediated cell adhesion is perturbed.

The aforementioned emphasis on cell migration should not prelude a role of myosin VI in other cellular processes that may also contribute to the development of human prostate cancer. In studies unrelated to human cancer, myosin VI was found to play critical roles in spermatogenesis,36 inner ear hair cell differentiation,37 asymmetric stem cell division,38 endocytosis,10 and secretion.13 Although these seemingly diverse functions may have a common underlying mechanism linked to the unique myosin VI motor direction, they appear to be species-, organ-, and tissue-specific and possibly depend on specific isoforms of myosin VI as well as the presence of critical myosin VI binding partners.6 In this study, inhibition of myosin VI expression in human prostate cancer cells resulted in reduced anchorage-independent growth (Figure 4) , as well as a nearly 10-fold induction of the tumor suppressor TXNIP (VDUP1) (Figure 5b) , suggesting a key role for myosin VI in maintaining the malignant phenotype of human prostate cancer cells. TXNIP may play a key role in regulating oncogenic signaling because expression analysis also identified dramatically reduced expression of TXNIP after transfection of an oncogenic ETS transcription factor.39 Despite the fact that the TXNIP expression may be regulated by the ischemic conditions encountered during the surgical tissue collection process and that the protein product is very labile,40,41 examination in clinical specimens indeed revealed a generally decreased pattern of TXNIP expression in cancer samples (Figure 5c) . This novel observation should be followed up pending the availability of an antibody suitable for immunohistochemical analysis.

Our functional studies primarily relied on siRNA technology because of lack of expression constructs for human myosin VI. Off-target effects, which lead to changes in expression in genes other than the target gene, cannot be efficiently controlled unless a rescue construct is available.42 It remains to be definitively determined whether the pattern of gene expression alterations (Supplemental Figure 2 at http://ajp.amjpathol.org) was a direct response to loss of myosin VI, or a result of off-target gene regulation by synthetic siRNA duplexes. A recent study,43 however, revealed distinctive, nonoverlapping patterns of off-target gene suppression among experiments targeting seven different locations of the same MAPK14 transcript, suggesting that off-target effects are specific to the target sequence but not to the target gene. Therefore our observation that the two different myosin VI siRNA sequences led to almost identical gene suppression patterns (Supplemental Figure 2 at http://ajp.amjpathol.org) argued against an off-target effect. In addition, the observed gene expression alterations were specific to myosin VI knockdown because these alterations were not observed in our expression analysis after knockdown of other genes (J.L., unpublished observation). Therefore, in this study, it is unlikely that the off-target effects played a dominant role in regulating gene expression and mediating the biological effect after myosin VI gene knockdown.

Because 80 to 90% of human prostate cancers diagnosed today present a pathology (Gleason score 6 and 7) that highly correlates with cancer-specific myosin VI overexpression (Figure 2) , relevant studies may have an impact in clinical management of human prostate cancer. However, unlike AMACR,16 we do not expect myosin VI to be useful as a tissue marker for prostate cancer diagnosis by IHC given that many of the normal and atrophy lesions were also positive for staining and a subset of the cancers were negative or weak (Figure 2, c and d) . Inherited inactivating deletions and mutations of myosin VI gene in both mice and humans result in hearing loss but do not affect viability,6 suggesting that myosin VI may be amenable to therapeutic intervention. Myosin VI function may be mediated by its interaction with multiple binding partners6 through its tail domain. Detailed structural and functional studies in the context of molecular interactions may help to identify specific therapeutic targets.

In summary, we discovered a novel connection between myosin VI and human prostate cancer. Myosin VI is one of the top genes and also the only myosin gene that has demonstrated cancer-specific overexpression in our expression data, shedding light on the nature and scale of dysregulated myosin VI expression in human prostate cancer. Previously characterized as a backward motor, myosin VI moves toward the minus end of the actin track, a direction opposite to all other known myosin members. Myosin VI may have unique properties and functions that are yet to be fully characterized, particularly in the context of human cancer. This novel connection should stimulate a thorough investigation of the unique structural and functional properties of myosin VI in a broader context.

We thank Dr. Mark Mooseker (Yale University, New Haven, CT) for providing the myosin VI antibody for immunohistochemical analysis, and Dr. Denise Montell (The Johns Hopkins University, Baltimore, MD) for critical reading and discussion of the manuscript.

【参考文献】
Berg JS, Powell BC, Cheney RE: A millennial myosin census. Mol Biol Cell 2001, 12:780-794

Krendel M, Mooseker MS: Myosins: tails (and heads) of functional diversity. Physiology 2005, 20:239-251

Wu X, Jung G, Hammer JA, III: Functions of unconventional myosins. Curr Opin Cell Biol 2000, 12:42-51

Kellerman KA, Miller KG: An unconventional myosin heavy chain gene from Drosophila melanogaster. J Cell Biol 1992, 119:823-834

Hasson T, Mooseker MS: Porcine myosin-VI: characterization of a new mammalian unconventional myosin. J Cell Biol 1994, 127:425-440

Buss F, Spudich G, Kendrick-Jones J: Myosin VI: cellular functions and motor properties. Annu Rev Cell Dev Biol 2004, 20:649-676

Frank DJ, Noguchi T, Miller KG: Myosin VI: a structural role in actin organization important for protein and organelle localization and trafficking. Curr Opin Cell Biol 2004, 16:189-194

Wells AL, Lin AW, Chen LQ, Safer D, Cain SM, Hasson T, Carragher BO, Milligan RA, Sweeney HL: Myosin VI is an actin-based motor that moves backwards. Nature 1999, 401:505-508

Menetrey J, Bahloul A, Wells AL, Yengo CM, Morris CA, Sweeney HL, Houdusse A: The structure of the myosin VI motor reveals the mechanism of directionality reversal. Nature 2005, 435:779-785

Hasson T: Myosin VI: two distinct roles in endocytosis. J Cell Sci 2003, 116:3453-3461

Warner CL, Stewart A, Luzio JP, Steel KP, Libby RT, Kendrick-Jones J, Buss F: Loss of myosin VI reduces secretion and the size of the Golgi in fibroblasts from Snell??s waltzer mice. EMBO J 2003, 22:569-579

Geisbrecht ER, Montell DJ: Myosin VI is required for E-cadherin-mediated border cell migration. Nat Cell Biol 2002, 4:616-620

Buss F, Luzio JP, Kendrick-Jones J: Myosin VI, an actin motor for membrane traffic and cell migration. Traffic 2002, 3:851-858

Yoshida H, Cheng W, Hung J, Montell D, Geisbrecht E, Rosen D, Liu J, Naora H: Lessons from border cell migration in the Drosophila ovary: a role for myosin VI in dissemination of human ovarian cancer. Proc Natl Acad Sci USA 2004, 101:8144-8149

Luo J, Duggan DJ, Chen Y, Sauvageot J, Ewing CM, Bittner ML, Trent JM, Isaacs WB: Human prostate cancer and benign prostatic hyperplasia: molecular dissection by gene expression profiling. Cancer Res 2001, 61:4683-4688

Luo J, Zha S, Gage WR, Dunn TA, Hicks J, Bennett CJ, Ewing CM, Platz EA, Ferdinandusse S, Wanders RJ, Trent JM, Isaacs WB, De Marzo AM: Alpha-methylacyl-CoA racemase: a new molecular marker for prostate cancer. Cancer Res 2002, 62:2220-2226

Lapointe J, Li C, Higgins JP, van de Rijn M, Bair E, Montgomery K, Ferrari M, Egevad L, Rayford W, Bergerheim U, Ekman P, DeMarzo AM, Tibshirani R, Botstein D, Brown PO, Brooks JD, Pollack JR: Gene expression profiling identifies clinically relevant subtypes of prostate cancer. Proc Natl Acad Sci USA 2004, 101:811-816

Faith DA, Isaacs WB, Morgan JD, Fedor HL, Hicks JL, Mangold LA, Walsh PC, Partin AW, Platz EA, Luo J, De Marzo AM: Trefoil factor 3 overexpression in prostate carcinoma: prognostic importance using tissue microarrays. Prostate 2004, 61:215-227

Bussemakers MJ, van Bokhoven A, Verhaegh GW, Smit FP, Karthaus HF, Schalken JA, Debruyne FM, Ru N, Isaacs WB: DD3: a new prostate-specific gene, highly overexpressed in prostate cancer. Cancer Res 1999, 59:5975-5979

DeYoung MP, Tress M, Narayanan R: Identification of Down??s syndrome critical locus gene SIM2-s as a drug therapy target for solid tumors. Proc Natl Acad Sci USA 2003, 100:4760-4765

Klezovitch O, Chevillet J, Mirosevich J, Roberts RL, Matusik RJ, Vasioukhin V: Hepsin promotes prostate cancer progression and metastasis. Cancer Cell 2004, 6:185-195

Wolfgang CD, Essand M, Vincent JJ, Lee B, Pastan I: TARP: a nuclear protein expressed in prostate and breast cancer cells derived from an alternate reading frame of the T cell receptor gamma chain locus. Proc Natl Acad Sci USA 2000, 17:9437-9442

DeMarzo AM, Nelson WG, Isaacs WB, Epstein JI: Pathological and molecular aspects of prostate cancer. Lancet 2003, 361:955-964

Jeon JH, Lee KN, Hwang CY, Kwon KS, You KH, Choi I: Tumor suppressor VDUP1 increases p27(kip1) stability by inhibiting JAB1. Cancer Res 2005, 65:4485-4489

Dhanasekaran SM, Barrette TR, Ghosh D, Shah R, Varambally S, Kurachi K, Pienta KJ, Rubin MA, Chinnaiyan AM: Delineation of prognostic biomarkers in prostate cancer. Nature 2001, 412:822-826

Magee JA, Araki T, Patil S, Ehrig T, True L, Humphrey PA, Catalona WJ, Watson MA, Milbrandt J: Expression profiling reveals hepsin overexpression in prostate cancer. Cancer Res 2001, 61:5692-5696

Welsh JB, Sapinoso LM, Su AI, Kern SG, Wang-Rodriguez J, Moskaluk CA, Frierson HF, Jr, Hampton GM: Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer. Cancer Res 2001, 6:5974-5978

LaTulippe E, Satagopan J, Smith A, Scher H, Scardino P, Reuter V, Gerald WL: Comprehensive gene expression analysis of prostate cancer reveals distinct transcriptional programs associated with metastatic disease. Cancer Res 2002, 62:4499-4506

Singh D, Febbo PG, Ross K, Jackson DG, Manola J, Ladd C, Tamayo P, Renshaw AA, D??Amico AV, Richie JP, Lander ES, Loda M, Kantoff PW, Golub TR, Sellers WR: Gene expression correlates of clinical prostate cancer behavior. Cancer Cell 2002, 1:203-209

Ernst T, Hergenhahn M, Kenzelmann M, Cohen CD, Bonrouhi M, Weninger A, Klaren R, Grone EF, Wiesel M, Gudemann C, Kuster J, Schott W, Staehler G, Kretzler M, Hollstein M, Grone HJ: Decrease and gain of gene expression are equally discriminatory markers for prostate carcinoma: a gene expression analysis on total and microdissected prostate tissue. Am J Pathol 2002, 160:2169-2180

Luo JH, Yu YP, Cieply K, Lin F, Deflavia P, Dhir R, Finkelstein S, Michalopoulos G, Becich M: Gene expression analysis of prostate cancers. Mol Carcinog 2002, 33:25-35

Kohn EC, Liotta LA: Molecular insights into cancer invasion: strategies for prevention and intervention. Cancer Res 1995, 55:1856-1862

Clark WH: Tumour progression and the nature of cancer. Br J Cancer 1991, 64:631-644

Mareel M, Leroy A: Clinical, cellular, and molecular aspects of cancer invasion. Physiol Rev 2003, 83:337-376

De Marzo AM, Knudsen B, Chan-Tack K, Epstein JI: E-cadherin expression as a marker of tumor aggressiveness in routinely processed radical prostatectomy specimens. Urology 1999, 53:707-713

Hicks JL, Deng WM, Rogat AD, Miller KG, Bownes M: Class VI unconventional myosin is required for spermatogenesis in Drosophila. Mol Biol Cell 1999, 10:4341-4353

Self T, Sobe T, Copeland NG, Jenkins NA, Avraham KB, Steel KP: Role of myosin VI in the differentiation of cochlear hair cells. Dev Biol 1999, 214:331-341

Petritsch C, Tavosanis G, Turck CW, Jan LY, Jan YN: The Drosophila myosin VI Jaguar is required for basal protein targeting and correct spindle orientation in mitotic neuroblasts. Dev Cell 2003, 4:273-281

Myers-Irvin JM, Van Le TS, Getzenberg RH: Mechanistic analysis of the role of BLCA-4 in bladder cancer pathobiology. Cancer Res 2005, 65:7145-7150

Xiang G, Seki T, Schuster MD, Witkowski P, Boyle AJ, See F, Martens TP, Kocher A, Sondermeijer H, Krum H, Itescu S: Catalytic degradation of vitamin D up-regulated protein 1 mRNA enhances cardiomyocyte survival and prevents left ventricular remodeling after myocardial ischemia. J Biol Chem 2005, 280:39394-39402

Dutta KK, Nishinaka Y, Masutani H, Akatsuka S, Aung TT, Shirase T, Lee WH, Yamada Y, Hiai H, Yodoi J, Toyokuni S: Two distinct mechanisms for loss of thioredoxin-binding protein-2 in oxidative stress--induced renal carcinogenesis. Lab Invest 2005, 85:798-807

Kittler R, Pelletier L, Ma C, Poser I, Fischer S, Hyman AA, Buchholz F: RNA interference rescue by bacterial artificial chromosome transgenesis in mammalian tissue culture cells. Proc Natl Acad Sci USA 2005, 102:2396-2401

Siolas D, Lerner C, Burchard J, Ge W, Linsley PS, Paddison PJ, Hannon GJ, Cleary MA: Synthetic shRNAs as potent RNAi triggers. Nat Biotechnol 2005, 23:227-231

作者单位：From the Departments of Urology* and Pathology, The Johns Hopkins University School of Medicine, Baltimore, Maryland; the Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; and the Cancer Genetics Branch, National Human Genome Research Institute, Bethes

　　本文分为两部分。第一部分提出了一种利用泥土（以及其中的腐殖质）还原和吸附铬的实验方法，讨论了这种方法在实验室含铬(VI)废水的处理中的可行性及可操作性。第二部分主要从综合处理的角度，讨论了实验室中含铬、汞、铅等废水的处理方法，提出一种综合处理、反复利用的思路。

关键词　实验室、废水处理、无机化学、环境保护

一、导言

　　在浦口校区大学一年级的学生实验中，含有重金属离子及其配合物的废水是最主要的污染物。目前，这些废水未经任何处理即直接排放，对周边环境造成了不小的损害。
　　我们认为，在建立一套较为完善的废水处理系统之前，尝试以可行性强、操作简单的化学方法降低重金属污染是值得考虑的。对此，主要的思路有两条，一是降低污染物毒性后排放，二是将金属回收利用。本文从这两个角度出发，分为两部分。第一部分针对铬这一最主要的污染物，尝试了以含腐殖质的泥土还原并吸附铬(VI)，将其排放形式转变为低毒的铬(III)的实验方法。第二部分则论述了具体的实施方法，希望能尽量减少排放物的污染，或者利用不同实验的废料废水相互作用，创造各种金属的回收条件。

二、淤泥处理铬(VI)废水的实验方法

　　泥土中所含的腐殖质能将六价铬还原为三价，并与之形成有机配合物而吸附[1]。为此我们设计了如下的实验：
目的：验证泥土对含铬(VI)的废水中铬(VI)的去除能力
原理：（可能之原理）在酸性条件下，利用铬(VI)的氧化性将泥土中的还原性有机物氧化，使之转化为铬(III)。铬(III)又能与泥土中的某些成分络合继而被泥土吸附。最终排放的废水中铬(VI)含量显著减少。
原料：淤泥二份（分别取自明湖湖底以及运动馆前水渠），实验室重铬酸钾回收液（约0.016M），硫酸及氢氧化钠溶液。
仪器：722型分光光度计，实验室常用无机玻璃仪器
步骤：
淤泥在90℃下烘干4小时备用。
把重铬酸钾回收液稀释50倍左右备用。此时重铬酸钾浓度约为0.094mg/L。
取100mL稀释液，置于250mL锥形瓶中，并用硫酸调整pH值至1左右。在不同的条件下还原吸附稀释液中的铬(VI)，然后分析溶液中剩余的铬(VI)的含量。
分析方法：滤出还原吸附后的溶液，用氢氧化钠溶液调整pH值至8左右，过滤除去沉淀，然后以分光光度法，在366nm波长处测定铬(VI)的含量。处理条件及测定结果见下表。
表 1 明湖淤泥处理效果　

条件	吸光度	去除率
2g/0.5h	1.287	12.3%
2g/1.0h	1.236	15.7%
2g/2.0h	1.216	17.1%
8g/2.0h	0.098	93.3%
原始液(~36mg/L)	1.467

表 2 效果对比

条件	吸光度	去除率
2g(1)	1.284	13.2%
4g(1)	0.980	33.8%
4g(2)	1.200	18.9%
8g(1)	0.050	96.6%
8g(2)	0.049	96.7%
原始液(~36mg/L)	1.480

标注(1)的样品取自水渠，标注(2)的样品取自明湖。处理时间均为1小时。
结果显示：
1、泥土对铬(VI)确有去除作用，但对其去除铬(VI)的具体机理尚不清楚，我们认为可能的机理是泥土中的还原性物质（可能主要是腐殖质）在酸性条件下还原了铬(VI)，同时泥土中另一些物质（可能是有机物）与铬(III)形成了易被吸附的配合物。
2、就相同的淤泥来说，处理时间的不同将导致结果的差异，但时间的影响并不十分显著。
3、不同来源的泥土在相同条件下处理的结果存在差异。从水渠中取得的淤泥处理效果稍好。从淤泥形成的环境来看，该样品（取自芦苇丛底部，有较多有机物沉积）腐殖质含量较高，还原能力较强。
4、泥土的质量并不与铬(VI)的去除率呈线性关系，但可以观察到的是，泥土的用量对处理效果有决定性影响。用量越大，其对铬(VI)的去除率也就显著升高。
5、在泥土过量(8g)的情况下，两种样品均能取得令人满意的去除率。可见，该方法对泥土的要求并不会太高，从而具有较强的可行性。
特别的是，我们尝试以10克泥土直接处理50毫升未经稀释（0.016M）的废水，以目测（浓度太高无法分光）判断，至少去除了50%。据此推测，可能存在这样一个平衡，即去除的铬(VI)的量与原始废水浓度正相关。
6、我们曾对处理后的废水进行测试，结果显示，除铬(VI)几乎被除尽外，水中铬(III)含量也很少，我们推测，剩余的铬进入了泥土中。

三、实验室废水处理过程

1、排放
　　排放是一种较为方便的处理方式。优点是操作简单，设备以及条件要求不高，故经济性较好。相应的缺点在于，虽然可以很大程度上减少污染，但无法完全消除。
　　以铬(VI)为例，前一部分已经说明淤泥处理重铬酸钾污染的可行性。据我们统计，浦口校区大一实验共计600人左右，使用后排放的洗液以及滴定剂共含有2~2.5千克重铬酸钾。按照实验结果的标准，8克泥土可以处理含约10~20毫克重铬酸钾的废水，一年的泥土需求量将在2~2.5吨（约1~2立方米）之间。为此，可以建设小规模的处理池，首先收集重铬酸钾废液，贮于池中，再投入足量的淤泥（由实验数据可见，为保证效果，且鉴于淤泥易于获得，应予过量投放）。加入适量硫酸酸化，再放置一定时间（由于一学年的废水可以同时处理，故处理时间十分充裕，可以在长期放置的情况下使之完全反应）。
　　基于另一实验事实，即处理效果与初始浓度正相关，铬浓度越高，相同质量的淤泥对其处理效果就相对越好。为此，我们在实际处理中可以不对废水进行如实验般的稀释，而可以采取多级处理的方案，逐步降低废水中铬浓度，以取得更佳的效果。
　　关于使用硫酸可能造成成本过高的问题，我们认为，由于铬(VI)在酸性条件下方显强氧化性，故任何以化学处理（还原方式）为主的处理方法都有一定的耗酸量，所以这方面的成本是难以避免的。
　　另一相关问题在于此法实施以后产生的含铬泥土如何处理。此种泥土含有较多的铬。大部分铬(VI)已被还原，故毒性已大大降低，污泥的总量大概二至三吨。由于其为固体形态，量又不大，便于集中和运输，可以直接交由南京的专业污染物处理点进一步处理。

2、回收

　　以实验室现有的条件，较简便的金属回收方法是将金属离子以氢氧化物的形式沉淀分离。这就要求与上述淤泥处理完全不同的方法。
首先考察各种金属离子的排放形式：铬（重铬酸钾，硫酸铬）；汞（氯化汞，氯化亚汞）；铅（EDTA合铅(II)）；铜（EDTA合铜，硫酸铜），等等。其中，氯化汞和硫酸铬属于共同排放。
通过计算得知，每年实验中排放氯化汞（重铬酸钾法测铁）约0.5千克，排放铅离子（锡青铜中铅锡的测定）1~2千克，数量也相当可观。
　　总体的处理思路是，对于高价阴离子，先将其还原为低价阳离子；而对EDTA配离子则可先行置换。为此我们考虑以硫酸亚铁胺为还原剂——在大一上期的化学制备实验中，产生了大量的硫酸亚铁胺。由于纯度的原因用途十分有限。因此可以用来还原重铬酸钾。还原后的溶液中含有铁(III)及铬(III)离子。从它们氢氧化物的溶度积可以知道，铁(III)及铬(III)离子的沉淀条件分别是PH=3~4以及PH=8~9，因此可以使用廉价的石灰调整PH值，先将高铁沉淀分离（待作他用），再将铬(III)沉淀回收。
　　由此产生的氢氧化铁以盐酸溶解后，可以用于置换EDTA合铅、铜中的铅和铜。这里，EDTA合铁(III)的稳定常数是EDTA金属配合物里最高的，所以置换可以完成。而且由于铁本身的毒性极小，几乎不造成污染，故EDTA合铁可以直接排放。而置换出的铅、铜同样以沉淀的形式回收。
至于硫酸铜、氯化汞、硫酸铬，都具备直接沉淀的条件，不再赘述。
回收的各种金属可以再度利用。
　　总的来说，沉淀回收法的原理较为简单，可操作性也很强，对污染的消除效果相当不错。成本虽然较排放法为高，但考虑到金属的回收再利用，以及消除环境污染的具体效果，这些支出还是可以接受的。

3、处理以外的一些要求

　　为达到降低以及消除污染的目的，首先必须将实验产生的各种金属离子尽量分类集中。这个工作比较繁。我们认为，可以在实验室建立一套相应的制度，例如：要求同学们在实验过程中自觉将各种废水分类集中，将工作量分摊，就成为一件易于办到的事。而与之对应的，需要实验室提供收集、贮存各种废水的容器和场所。每学期或者学年结束后，可以开展学生实践，由学生处理本学期或学年收集的废水，教学和实践、探索相结合。减少污染，保护环境，需要老师和同学共同努力。

四、结语

　　本文对重金属废水的处理提出了一些建议和思路。虽然这些方法在理论上是基本可行的，但具体实施起来可能还有我们没有考虑到的问题或困难，还须多作探讨。废水处理是一个复杂的问题，方法还要在实践中不断完善。

    续断为川续断Dipsacus asperoides C.Y. Cheng et T.M.Ai的干燥根。2005版《中国药典》一部采用高效液相色谱法测定其川续断皂苷VI的含量。文献报道的续断及其制剂的含量测定方法，一般为测定续断总皂苷的含量，川续断皂苷VI的含量测定方法报道较少，一般为药典方法。
    2005版《中国药典》续断含量测定项下：
    色谱条件与系统适用性试验：以十八烷基硅烷键合硅胶为填充剂；以乙腈-水（30：70）为流动相；检测波长为212nm。理论板数按川续断皂苷VI峰计算应不得低于3000。
    供试品溶液的制备：取本品细粉约0.5g，精密称定，置具塞锥形瓶中，精密加入甲醇25m，密塞，称定重量，超声处理（功率100W，频率40kHz）30分钟，放冷，再称定重量，用甲醇补足减失的重量，摇匀，滤过。精密量取续滤液5ml，置50ml量瓶中，加流动相稀释至刻度，摇匀，即得。

¹ Deutsches Herzzentrum und 1. Medizinische Klinik, Technische Universität München, D-80636 München, Germany
² Rudolf Virchow Center for Experimental Biomedicine, University of Würzburg, 97078 Würzburg, Germany
³ GSF National Research Center for Environment and Health, Institute of Pathology, D-85764 Neuherberg, Germany

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion References

Platelet adhesion and aggregation at sites of vascular injuryis crucial for hemostasis but may lead to arterial occlusionin the setting of atherosclerosis and precipitate diseases suchas myocardial infarction. A current hypothesis suggests thatplatelet glycoprotein (GP) Ib interaction with von Willebrandfactor recruits flowing platelets to the injured vessel wall,where subendothelial fibrillar collagens support their firmadhesion and activation. However, so far this hypothesis hasnot been tested in vivo. Here, we demonstrate by intravitalfluorescence microscopy of the mouse carotid artery that inhibitionor absence of the major platelet collagen receptor, GPVI, abolishesplatelet–vessel wall interactions after endothelial denudation.Unexpectedly, inhibition of GPVI by the monoclonal antibodyJAQ1 reduced platelet tethering to the subendothelium by 89%.In addition, stable arrest and aggregation of platelets wasvirtually abolished under these conditions. Using differentmodels of arterial injury, the strict requirement for GPVI inthese processes was confirmed in GPVI-deficient mice, whereplatelets also failed to adhere and aggregate on the damagedvessel wall. These findings reveal an unexpected role of GPVIin the initiation of platelet attachment at sites of vascularinjury and unequivocally identify platelet–collagen interactions(via GPVI) as the major determinant of arterial thrombus formation.

Key Words: arterial thrombosis • collagen • receptor • GPVI • mouse

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

Platelet adhesion and aggregation is essential to limit bloodloss at sites of vascular injury but may also lead to arterialocclusion and irreversible tissue damage after disruption ofthe atherosclerotic plaque. The first platelet response to vascularinjury is adhesion to the exposed subendothelial matrix, whichtriggers subsequent platelet aggregation. A current hypothesissupported by numerous in vitro studies suggests that the interactionof glycoprotein (GP)Ib-V-IX with von Willebrand factor (vWf)recruits flowing platelets to the injured vessel wall (1), wherethey interact with exposed extracellular matrix proteins resultingin firm adhesion and thrombus growth (2, 3). Although severalof the macromolecular components of the subendothelial layersuch as laminin, fibronectin, and vWf all provide a suitablesubstrate for platelet adhesion, fibrillar collagen is consideredthe most thrombogenic constituent of the vascular subendotheliumas it not only supports platelet adhesion but also acts as astrong activator of platelets in vitro (3, 4). However, thein vivo relevance of platelet–collagen interactions inthe setting of arterial thrombosis has not been established.This might be explained by the complexity of the platelet–collageninteraction, which involves a variety of different receptorsand signaling pathways making the in vivo inhibition of thisprocess very difficult. Besides GPIb-V-IX and integrin _IIbß₃,which interact indirectly with collagen via vWf (5), a largenumber of collagen receptors have been identified on platelets,including most importantly integrin ₂ß₁ (6), GPV (7),and GPVI (8).

Only recently, GPVI has been established as the central plateletcollagen receptor that is essential for platelet adhesion andaggregation on immobilized collagen in vitro, as it mediatesthe activation of different adhesive receptors, including integrins_IIbß₃ and ₂ß₁ (9–12). GPVI is a 60–65-kDtype I transmembrane GP belonging to the immunoglobulin superfamily(13, 14) that forms a complex with the FcR chain at the cellsurface in human and mouse platelets (9, 10). Signaling throughGPVI occurs via a pathway similar to that used by immunoreceptorsas revealed by the tyrosine phosphorylation of the FcR chainimmunoreceptor tyrosine-based activation motif by a src-likekinase (15). The mAb JAQ1 (10) blocks the major collagen bindingsite on mouse GPVI and inhibits firm platelet adhesion to collagenunder low and high shear flow conditions (12). In vivo applicationof JAQ1 induces virtually complete internalization and degradationof GPVI on mouse platelets resulting in a "GPVI knockout"–likephenotype for at least 2 wk. Such GPVI-depleted mice have significantlyprolonged bleeding times and their platelets fail to respondto collagen but not to other agonists (11). Despite its essentialrole in collagen-induced activation of platelets, there hasbeen only very limited evidence for a role of GPVI as a directadhesion receptor (14, 16).

In this study we investigated the in vivo significance of platelet–collageninteractions in the dynamic process of platelet adhesion andaggregation at sites of arterial injury. We show that inhibitionor deletion of GPVI virtually abrogates stable platelet adhesionand aggregation after endothelial denudation of the carotidartery in mice. Very unexpectedly, we found that tethering/slowsurface translocation of platelets was also strongly inhibitedin the absence of functional GPVI. These findings reveal a crucialrole of GPVI in the initiation of platelet recruitment at sitesof vascular injury and provide the first in vivo evidence thatplatelet–collagen interactions are of paramount importanceduring arterial thrombus formation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

Animals.
Specific pathogen-free C57BL/6J mice were obtained from .For experiments, 12-wk-old male mice wereused. All experimental procedures performed on animals wereapproved by the German legislation on protection of animals.

mAbs.
mAbs against GPVI (JAQ1) and GPIb (p0p/B) were generated aspreviously described (10). Fab fragments from JAQ1 and p0p/Bwere also generated as previously described (11). Irrelevantcontrol rat IgG and FITC-conjugated hamster anti–ß1integrin (Ha31/8) was obtained from . The followingantibodies were produced and modified in our laboratory (17)and used for flow cytometry: JON1-PE (anti-GPIIb/IIIa), p0p4-PE(anti-GPIb), p0p6-FITC (anti-GPIX), DOM1-FITC (anti-GPV), andLEN1-FITC (anti-GPIa).

Generation of GPVI-deficient Mice.
To generate mice lacking GPVI, C57BL6/J wild-type mice wereinjected with 100 µg JAQ1 intracardially (11). Animalswere used for in vivo assessment of platelet adhesion on day5 after mAb injection. Absence of GPVI expression on plateletswas verified by Western blot analysis and flow cytometry.

Flow Cytometry.
Heparinized whole blood, obtained from wild-type C57BL6/J orGPVI-depleted mice was diluted 1:30 with modified Tyrodes-Hepesbuffer (134 mM NaCl, 0.34 mM Na₂HPO₄, 2.9 mM KCl, 12 mM NaHCO₃,20 mM Hepes, 5 mM glucose, and 1 mM CaCl₂, pH 6.6). The sampleswere incubated with fluorophore-labeled antibodies for 10 minat room temperature and directly analyzed on a FACScalibur^TM( ).

Preparation of Platelets for Intravital Microscopy.
Wild-type or GPVI-deficient platelets were isolated from wholeblood as previously described (18) and labeled with 5-carboxyfluoresceindiacetate succinimidyl ester (DCF). The DCF-labeled plateletsuspension was adjusted to a final concentration of 200 x 10⁶platelets/250 µl. Where indicated, fluorescent wild-typeplatelets were preincubated with 50 µg/ml anti-GPVI (JAQ1)Fab fragments, which do not induce any detectable platelet signaling(19). In a separate set of experiments, platelets were preincubatedwith 50 µg/ml anti-GPIb (p0p/B) Fab fragments for 10 minto examine the role of GPIb for platelet recruitment after endothelialdenudation. The pretreated platelets together with the Fab fragmentswere infused into wild-type recipient mice and platelet adhesionwas assessed before and after carotid injury by in vivo videomicroscopy, as described below.

Carotid Ligation and Assessment of Platelet Adhesion and Aggregation by Intravital Microscopy.
Wild-type C57BL6/J or GPVI-deficient mice were anesthetizedby intraperitoneal injection of a solution of midazolame (5mg/kg body weight; Ratiopharm), medetomidine (0.5 mg/kg bodyweight; Pfizer), and fentanyl (0.05 mg/kg body weight; CuraMedPharma GmbH). Polyethylene catheters (Portex) were implantedinto the right jugular vein and 200 x 10⁶/250 µl fluorescentplatelets were infused intravenously. The right common carotidartery was dissected free and ligated vigorously near the carotidbifurcation for 5 min to induce vascular injury. Before andafter vascular injury, the fluorescent platelets were visualizedin situ by in vivo video microscopy of the right common carotidartery. Platelet–vessel wall interactions were monitoredusing a Zeiss Axiotech microscope (20x water immersion objective,W 20x/0.5; Carl Zeiss MicroImaging, Inc.) with a 100-W HBO mercurylamp for epi-illumination. All videotaped images were evaluatedusing a computer-assisted image analysis program (Cap Image7.4; provided by Dr. Zeintl; references 18 and 20). Tetheredplatelets were defined as all cells establishing initial contactwith the vessel wall, followed by slow surface translocationat a velocity significantly lower than the centerline velocity,or by firm adhesion. Their numbers are given as cells per mm²endothelial surface. The number of adherent platelets was assessedby counting the cells that did not move or detach from the endothelialsurface within 10 s. The number of platelet aggregates at thesite of vascular injury was also quantified and is presentedper mm².

Scanning Electron Microscopy.
After intravital video fluorescence microscopy, the carotidartery was perfused with PBS at 37°C for 1 min, followedby perfusion fixation with phosphate-buffered glutaraldehyde(1% vol/vol). The carotid artery was excised, opened longitudinally,additionally fixed by immersion in 1% PBS-buffered glutaraldehydefor 12 h, dehydrated in ethanol, and processed by critical pointdrying with CO₂. The carotid artery specimens were then orientedwith the lumen exposed, mounted with carbon paint, sputter coatedwith platinum, and examined using a field emission scanningelectron microscope (JSM-6300F; Jeol Ltd.).

Assessment of Arterial Thrombosis after Ferric Chloride Exposure.
Vascular injury of the carotid artery was induced by local applicationof ferric chloride (FeCl₃) essentially as described by Fay etal. (21). In brief, control or GPVI-depleted C57BL/6J mice wereanesthetized and fluorescence-tagged control or GPVI-deficientplatelets were infused intravenously into the jugular vein ofuntreated or GPVI-depleted recipients, respectively. Thereafter,the common carotid artery was dissected free and a filter paper(0.5 x 1.0 mm) saturated with 10% FeCl₃ was applied to the adventitialsurface of the vessel, as previously described (21). The timeto thrombotic occlusion of the carotid artery downstream ofthe site of injury (n = 10 per group) was defined as the timerequired for complete arrest of blood flow in the center ofthe vessel (platelet flow velocity 0 m/s) after removal of thefilter paper.

Determination of Platelet Recruitment after Wire-induced Arterial Denudation.
Wire-induced endothelial disruption was performed accordingto a method described by Lindner et al. (22). In brief, GPVI-depletedC57BL/6J mice were generated as described above (n = 12). Micepretreated with 100 µg irrelevant control IgG (n = 6)or PBS (n = 12) served as controls. 5 d after mAb injection,the animals were anesthetized and platelets were isolated fromcontrol or GPVI-depleted mice and labeled with DCF (see above).In the recipient mice, the right carotid artery was exposedvia a midline neck incision. The common, external, and internalcarotid arteries were identified, the right internal carotidartery was looped proximally and tied off distally with 8–0silk suture (Ethicon). Additional 8–0 silk ties were loopedround the common and external carotid arteries for temporaryvascular control during the procedure. A transverse arteriotomywas made in the right internal carotid artery and a 0.014-inflexible angioplasty guidewire was introduced and advanced 1cm to the aortic arch. Endothelial denudation injury of theright common carotid artery was performed by wire withdrawalwith rotating motion to ensure uniform and complete endothelialdenudation. After removal of the wire, the right internal carotidartery was tied off and platelet–vessel wall interactionswere visualized at the site of injury by in vivo video fluorescencemicroscopy as described above.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

To test the biological significance of platelet–collageninteractions in the processes of adhesion and aggregation invivo, we assessed platelet–vessel wall interactions aftervascular injury of the mouse carotid artery. Vigorous ligationof the carotid artery consistently caused complete loss of theendothelial cell layer and initiated platelet adhesion at thesite of injury, as assessed by scanning electron microscopy( a). Next, we used in vivo fluorescence microscopy(18, 20) to directly visualize and quantify the dynamic processof platelet accumulation after vascular injury. Numerous plateletswere tethered to the vascular wall within the first minutesafter endothelial denudation (4725 ± 239 platelets/mm²).Virtually all platelets establishing contact with the subendotheliuminitially exhibited a slow surface translocation of the "stop-start"type (23). As we observed transition from initial slow surfacetranslocation to irreversible platelet adhesion in 88% of allplatelets (4.182 ± 253 platelets/mm²; b), plateletarrest remained transient in only 12% (543 ± 32 platelets/mm²).Once firm arrest was established, adherent platelets recruitedadditional platelets from the circulation, resulting in aggregateformation ( c). Similar characteristics of platelet recruitmenthave been obtained earlier with immobilized collagen in vitro(5). In contrast, only few platelets were tethered to the intactvascular wall under physiological conditions (P < 0.05 vs.vascular injury) and 100% of these platelets were displacedfrom the vascular wall without firm arrest (P < 0.05 vs.vascular injury; , a–c).

fig-ommitted

Figure 1. Platelet adhesion and aggregation after vascular injury of the common carotid artery in C57BL/6J mice in vivo. (a) Scanning electron micrographs of carotid arteries before (left) and 2 h after vascular injury (right). Endothelial denudation induces platelet adhesion and aggregation, resulting in the formation of a platelet-rich (lower right) thrombus. (b) Platelet–endothelial cell interactions 5 min after vascular injury were investigated by in vivo fluorescence microscopy of the common carotid artery in situ (solid columns). Animals without vascular injury served as controls (open columns). The left and right panels summarize platelet tethering and firm platelet adhesion, respectively, of eight experiments per group. Platelets were classified according to their interaction with the endothelial cell lining as previously described (refer to Materials and Methods) and are given per mm² of vessel surface. Mean ± SEM. *, significant difference compared with control, P < 0.05. (c) Platelet aggregation after vascular injury was determined by fluorescence microscopy in vivo (solid columns). Animals without vascular injury served as controls (open columns). Mean ± SEM and n = 8 each group. *, significant difference compared with wild-type mice, P < 0.05. The microphotographs (right) show representative in vivo fluorescence microscopy images in control animals (top) or after vascular injury (bottom). White arrows indicate adherent platelets. Bars, 50 µm.

Subendothelial fibrillar collagen has been proposed to be ofmajor importance for platelet adhesion and aggregation at sitesof vascular injury (2, 4, 24) as in vitro it strongly supportsplatelet activation and adhesion. However, this hypothesis hasnot been tested in vivo where various other agonists and adhesionmolecules might be involved in thrombus formation. To directlytest the in vivo relevance of platelet–collagen interactionsin arterial thrombus formation, we inhibited or deleted GPVIin vivo. The mAb JAQ1 (10, 25) blocks the major collagen-bindingsite on mouse GPVI and almost completely inhibits firm plateletadhesion to immobilized fibrillar collagen under high shearflow conditions in vitro (12). To study the effect of GPVI inhibitionin arterial thrombus formation, mice received syngeneic, fluorescence-taggedplatelets preincubated with JAQ1 Fab fragments or isotype-matchedcontrol IgG and carotid injury was induced as described above.Very unexpectedly, platelet tethering/slow surface translocationat sites of endothelial denudation, a process thought to bemediated solely by GPIb interaction with immobilized vWf (1,3, 26, 27), was reduced by 89% (P < 0.05 vs. control IgG; a) in the presence of JAQ1 Fab fragments. In addition,stable platelet arrest was reduced by 93% in the presence ofJAQ1 ( a). We observed transition from initial tethering/slowsurface translocation to irreversible platelet adhesion in only58% of those platelets establishing initial contact with thesubendothelial surface compared with 89% with control IgG–pretreatedplatelets (P < 0.05; b). Aggregation of adherent plateletswas virtually absent after pretreatment of platelets with JAQ1Fab fragments but not in the controls (P < 0.05 vs. control;). The unanticipated inhibitory effect of GPVIblockade on tethering/slow surface translocation prompted usto examine the role of GPIb in this process. Mice received fluorescence-taggedplatelets preincubated with Fab fragments of a function blockingantibody against GPIb (p0p/B) and carotid injury was inducedas described above. As shown in e, this treatment resultedin a similarly profound reduction in platelet tethering andfirm adhesion (and consequently also in aggregate formation)as anti-GPVI treatment (see above) confirming the crucial roleof GPIb for platelet attachment to the damaged vascular wallunder conditions of arterial shear. This finding strongly suggestedthat both GPVI and GPIb are required to recruit platelets tothe injured arterial wall in vivo.

fig-ommitted

Figure 2. Inhibition of GPVI abrogates platelet adhesion and aggregation after vascular injury. (a) Platelet adhesion after vascular injury was determined by intravital video fluorescence microscopy. Fluorescent platelets were preincubated with 50 µg/ml anti-GPVI (JAQ1) Fab fragments or control rat IgG. Platelets without mAb preincubation served as control. The left and right panels summarize transient and firm platelet adhesion, respectively. Mean ± SEM and n = 8 each group. *, significant difference compared with control, P < 0.05. (b) The percentage of platelets establishing irreversible adhesion after initial tethering/slow surface translocation is illustrated. (c) Platelet aggregation after vascular injury in vivo. Aggregation of platelets preincubated with tyrode, irrelevant rat IgG, or anti-GPVI (JAQ1) was assessed by fluorescence microscopy as previously described. Mean ± SEM and n = 8 each group. *, significant difference compared with control, P < 0.05. (d) The photomicrographs show representative in vivo fluorescence microscopy images illustrating platelet adhesion in the absence or presence of anti-GPVI Fab (JAQ1) or control IgG. Bars, 30 µm. (e) Inhibition of GPIb abrogates platelet recruitment after vascular injury. Platelets were incubated with 50 µg/ml anti-GPIb Fab fragments (p0p/B) for 10 min. Platelets without mAb preincubation served as control. The left and right panels summarize transient and firm platelet adhesion, respectively. Mean ± SEM and n = 6 each group. *, significant difference compared with control, P < 0.05.

Together, the results described above demonstrated for the firsttime that direct platelet–collagen interactions are essentialfor initial platelet tethering and subsequent stable plateletadhesion and aggregation at sites of arterial injury. In addition,these data identify GPVI as a key regulator in this processwhereas other surface receptors, most importantly GPIb-V-IXand ₂ß₁, are not sufficient to initiate platelet adhesionand aggregation on the subendothelium in vivo.

The profound inhibition of platelet tethering by GPVI blockadewas surprising and suggested a previously unrecognized functionof this receptor in the very initial phase of thrombus formation.To exclude the possibility that this effect was based on stericimpairment of other receptors, e.g. GPIb-V-IX, by surface-boundJAQ1, we generated GPVI-deficient mice by injection of JAQ15 d before vascular injury. As reported previously, such treatmentinduces virtually complete internalization and proteolytic degradationof GPVI in circulating platelets, resulting in a GPVI knockout–likephenotype for at least 2 wk (11). As illustrated in Fig. 3 a, GPVI was undetectable in platelets from JAQ1-treated miceon day 5 after injection of 100 µg/mouse JAQ1 but notcontrol IgG, whereas surface expression and function of allother tested receptors, including GPIb-V-IX, _IIbß₃,and ₂ß₁ was unchanged in both groups of mice, confirmingearlier results (Fig. 3 a and Table I; reference 11).

fig-ommitted

Figure 3. Platelet adhesion after endothelial denudation in GPVI-deficient mice. (a) JAQ1-treated mice lack GPVI. On the top, platelets from mice pretreated with irrelevant control IgG or anti-GPVI (JAQ1) were stained for GPVI and GPIIb/IIIa (top) or GPIa and GPIb (bottom) and directly analyzed on a FACScalibur^TM is shown. Representative dot plots of six mice per group are presented. The expression levels of GPIIb/IIIa, GPIb-V-IX, and GPIa/IIa were not significantly different between the two groups of mice (refer to Table I). On the bottom, whole platelet lysates from three control IgG or JAQ1-treated mice separated by SDS-PAGE under nonreducing conditions and immunoblotted with FITC-labeled JAQ1, followed by incubation with horseradish peroxidase–labeled rabbit anti–FITC antibody is shown. (b) Scanning electron micrographs of carotid arteries 2 h after vascular injury in control animals or GPVI depleted. Endothelial denudation induced platelet adhesion and platelet aggregation in control animals. In contrast, only very few platelets attached along the damaged vessel wall in GPVI-depleted mice. Subendothelial collagen fibers are visible along the denuded area. (c) Platelet tethering and firm platelet adhesion, (d) transition from initial tethering to stable arrest (percentage of tethered platelets), and (e) platelet aggregation after vascular injury of the carotid artery was determined in GPVI-deficient (JAQ1-pretreated mice) or control IgG–pretreated mice (for details refer to Materials and Methods). The panels summarize platelet adhesion (tethering and firm adhesion) and platelet aggregation in eight experiments per group. Mean ± SEM. *, significant difference compared with control IgG, P < 0.05. (f) The photomicrographs show representative in vivo fluorescence microscopy images illustrating platelet adhesion in GPVI-deficient (JAQ1) and control IgG–treated mice. Bars, 30 µm.

fig-ommitted

Table I. Surface Expression of GPs on Platelets from JAQ1-treated Mice

As shown by scanning electron microscopy, platelet adhesionand aggregation after endothelial denudation of the common carotidartery were virtually absent in GPVI-deficient, but not in IgG-pretreated,mice ( b). Next, in vivo video fluorescence microscopywas used to define platelet adhesion dynamics after vascularinjury in GPVI-deficient mice (). The lossof GPVI profoundly reduced tethering/slow surface translocationof platelets at the site of vascular injury by 83% comparedwith IgG-pretreated mice (P < 0.05). This GPVI-independentslow surface translocation required vWf-GPIb–interactionas it was abrogated by preincubation of the platelets with Fabfragments of p0p/B (anti-GPIb), confirming the critical roleof GPIb in this process (not depicted). In the absence of GPVI,stable platelet adhesion was reduced by 90% compared with the(IgG-treated) control, whereas aggregation of adherent plateletswas virtually absent (). We saw transitionfrom platelet tethering to stable platelet adhesion in only58% of all platelets initially tethered to the site of injurycompared with 89% with control mAb–pretreated platelets(P < 0.05; d), indicating that GPIb-dependent surfacetranslocation is not sufficient to promote stable platelet adhesionand subsequent aggregation.

To further substantiate the role of GPVI in the process of plateletrecruitment after endothelial disruption, we next examined plateletadhesion/aggregation using two additional models of arterialthrombosis. First, arterial injury was induced in control orGPVI-depleted mice by local administration of ferric chlorideto the adventitial surface of the carotid artery as previouslydescribed (21). Time to arterial occlusion was monitored byin vivo fluorescence microscopy. As shown in , FeCl₃exposure resulted in a rapid thrombotic response in controlanimals. 9 out of 10 carotid arteries showed complete occlusionafter 235 ± 33 s. In contrast, arterial thrombus formationwas dramatically retarded in GPVI-deficient mice (P < 0.05vs. control mice). In fact, 6 out of 10 GPVI-deficient micedid not show arterial occlusion until 600 s after removal ofthe FeCl₃-saturated filter paper. In the remaining vessels,occlusion was markedly delayed (356 ± 55 s). These resultsfurther support a crucial role of GPVI in the process of arterialthrombus formation.

fig-ommitted

Figure 4. Role of GPVI in arterial thrombosis after ferric chloride exposure. Vascular injury of the carotid artery was induced by local application of ferric chloride on the carotid artery in GPVI-deficient or control mice. The time to thrombotic occlusion of the carotid artery downstream of the site of injury (n = 10 per group) was assessed in vivo by video fluorescence microscopy. Each symbol represents one experiment.

Next, we assessed platelet recruitment in the carotid arteryafter wire-induced endothelial disruption (22). As reportedearlier by Zhu et al. (28) and Lindner et al. (22), vascularinjury with a flexible wire consistently caused complete endothelialdenudation (unpublished data). In untreated control animalsand mice pretreated with irrelevant control IgG, disruptionof the endothelial surface initiated platelet tethering andadhesion as assessed in vivo by video fluorescence microscopy() . Numerous platelets were tethered to the vascularwall within the first minute after endothelial denudation (11.495± 1.283 tethered platelets/mm²). 46% of all plateletsestablishing contact with the subendothelium showed transitionfrom initial slow surface translocation to irreversible plateletadhesion (5.266 ± 915 firmly adherent platelets/mm²).Platelet adhesion at the site of injury was associated withthe formation of platelet aggregates attached to the site ofinjury. Platelet adhesion dynamics in mice pretreated with irrelevantIgG did not differ significantly from untreated control animals(13.521 ± 2.519 and 5.474 ± 1.575 tethered andfirmly adherent platelets/mm², respectively). In contrast tocontrol animals, platelet tethering/slow surface translocationand firm adhesion at sites of wire-induced endothelial denudationwere reduced by 90 and 95% in GPVI-depleted mice (P < 0.05vs. control mice; ). We observed transition from initialtethering/slow surface translocation to irreversible plateletadhesion in only 24% of those platelets establishing initialcontact with the subendothelial surface compared with 46% withcontrol animals (P < 0.05). Aggregation of adherent plateletswas virtually absent in GPVI-deficient mice (P < 0.05 vs.control; ). Together, these data add additional strongevidence to the concept that GPVI-mediated direct platelet–collageninteractions are essential for initial platelet tethering andsubsequent stable platelet adhesion and aggregation at sitesof arterial injury.

fig-ommitted

Figure 5. Role of GPVI in the regulation of platelet recruitment after wire injury of the carotid artery. Wire-induced endothelial denudation of the carotid artery was induced in GPVI-deficient mice. Untreated animals served as controls. The left shows representative in vivo fluorescence microscopy images illustrating the time course of platelet recruitment to the site of injury in control animals or GPVI-deficient mice (x500). The right summarizes platelet tethering, firm adhesion, and aggregate formation. Mean ± SEM. *, significant difference compared with control, P < 0.05.

Fibrillar collagen is a major constituent of the normal vesselwall but also of atherosclerotic lesions (29). In the processof atherogenesis, enhanced collagen synthesis by intimal smoothmuscle cells and fibroblasts has been shown to significantlycontribute to luminal narrowing (30). Plaque rupture or fissuring,either spontaneously or after balloon angioplasty, results inexposure of collagen fibrils to the flowing blood but theircontribution to arterial thrombus formation has been elusive.Platelets express a large number of different collagen receptors,which made it very difficult to identify the role of each ofthese receptors in the processes of adhesion and activationin vitro. In addition, reagents suitable for specific inhibitionof individual collagen receptors in vivo have not been available.Only recently has GPVI been identified as the central plateletreceptor that is essential for both adhesion and activationof platelets on collagen in vitro (12). In contrast, the absenceof other major collagen receptors such as integrin ₂ß₁or GPV only results in more subtle defects in the platelet–collageninteraction (7, 12, 31), suggesting that inhibition or deletionof GPVI, but no other collagen receptor, is required to abrogateplatelet collagen–interactions in vivo.

The results of this study provide the first definitive evidencethat subendothelial collagens are the major trigger of arterialthrombus formation and reveal an unexpected function of GPVIin platelet recruitment to the injured vessel wall. The processesof platelet tethering and slow surface translocation under conditionsof elevated shear are known to largely depend on GPIb interactionwith immobilized vWf (1). In addition, a number of studies haveshown that GPIb or even its NH₂-terminal 45-kD domain, whichcarries the binding site for vWf, mediates tethering of cellsor coated beads, respectively, to a vWf-coated surface underhigh shear flow conditions (32, 33). Together, these findingssuggested that GPIb–vWf interactions might be sufficientto establish the initial contact and slow surface translocationof platelets at sites of vascular injury. However, the resultspresented here demonstrate that tethering/slow surface translocationof platelets at sites of arterial injury in vivo is largelyinhibited in the absence of functional GPVI although expressionand function of GPIb-V-IX is not altered under these experimentalconditions ( and 3; reference 11). On the other hand,inhibition of the vWf binding site on GPIb by Fab fragmentsof the p0p/B mAb also virtually abrogated platelet adhesionto the injured vessel wall, confirming the strict requirementfor this interaction under conditions of high shear in mice( e). Thus, it appears that GPIb and GPVI act in concertto recruit platelets to the subendothelium in vivo by yet undefinedmechanisms. This strongly suggests that presentation of vWfon the extracellular matrix of the damaged vessel wall may differsignificantly from the conditions found in vitro when it ishomogeneously coated to a glass surface. At sites of vascularinjury, vWf is thought to become immobilized mostly on fibrillarcollagen (1, 5). Based on our results, one may speculate thatthe vWf layer on collagen fibrills might be inhomogeneous andfrequently interrupted making efficient interactions betweenGPIb and vWf impossible unless a second receptor interacts withthe "gaps," i.e., collagen not covered with vWf. GPVI is knownto be a low affinity collagen receptor mediating loose, butnot firm adhesion that may support this hypothesis (14, 16).Another point in favor of the idea that GPIb and GPVI act inconcert is the recent identification of different snake venom–derivedproteins that interact with platelets specifically through bothGPIb and GPVI, indicating that these two receptors might bephysically and functionally linked (34, 35).

During platelet tethering, ligation of GPVI can shift _IIbß₃and ₂ß₁ integrins from a low to a high affinity state(12). Both _IIbß₃ and ₂ß₁ then act in concertto promote subsequent stable arrest of platelets on collagen(5, 12) whereas _IIbß₃ is essential for subsequentaggregation of adherent platelets. Thus, ligation of GPVI duringthe initial contact between platelets and subendothelial collagenprovides an activation signal that is essential for subsequentstable platelet adhesion and aggregation. Our results suggestthat occupation or lateral clustering of GPIb (during GPIb-dependentsurface translocation), which has been shown to induce low levelsof _IIbß₃ integrin activation in vitro (32), may notbe sufficient to promote platelet adhesion in vivo.

This revised model of platelet attachment to the subendotheliumhighlights a central role of GPVI–collagen interactionsin all major phases of thrombus formation, i.e., platelet tethering,firm adhesion, and aggregation at sites of arterial injury (e.g.,during acute coronary syndromes). Although the data obtainedin mice cannot be directly extrapolated to the situation inhumans, the profound antithrombotic protection that was achievedby inhibition or depletion of GPVI strongly indicates that aselective pharmacological modulation of GPVI–collageninteractions may become a promising strategy to control theonset and progression of pathological arterial thrombosis.

	Acknowledgments

Scanning electron microscopy was performed with the skillfulhelp of Helga Wehnes.

This work was supported by grants Ni 556/4-1 to B. Nieswandtand Ga 481/4-1 to M. Gawaz from the Deutsche Forschungsgemeinschaft(DFG). B. Nieswandt is a Heisenberg Fellow of the DFG.

Submitted: June 10, 2002
Revised: September 16, 2002
Accepted: November 4, 2002

	Footnotes

S. Massberg, M. Gawaz, and S. Grüner contributed equallyto this work.

	References

日期：2007年5月18日 - 来自[2003年第162卷第1期]栏目

Principal Role of Glycoprotein VI in 2?1 and IIb?3 Activation During Collagen-Induced Thrombus Formation

From the Departments of Biochemistry (C.L., M.J.E.K., J.W.M.H.), Molecular Cell Biology and Genetics (J.L.V.B.), and Biophysics (M.A.M.J.V.), CARIM, Maastricht University, The Netherlands; the Laboratory for Thrombosis Research (A.S., K.V., H.D.), KU Leuven, Campus Kortrijk, Belgium; and E348 Institut National de la Santé et de la Recherche Médicale (C.L., M.J.-P.), Faculté Xavier Bichat, Université Paris, France.

ABSTRACT

Objective— High-shear perfusion of blood over collagen results in rapid platelet adhesion, aggregation, and procoagulant activity. We studied regulation of 2?1 and IIb?3 integrin activation during thrombus formation on collagen.

Methods and Results— Blockade of glycoprotein (GP) VI by 9O12 antibody or of P2Y purinergic receptors permitted platelet adhesion but reduced aggregate formation, fibrinogen binding, and activation of 2?1 and IIb?3, as detected with antibodies IAC-1 and PAC1 directed against activation-dependent epitopes of these integrins. Combined blockade of GPVI and P2Y receptors and thromboxane formation abolished integrin activation but still allowed adhesion of morphologically unstimulated, nonprocoagulant platelets. Exogenous ADP partly restored the suppressive effect of GPVI blockade on integrin 2?1 and IIb?3 activation. Adhesion was fully inhibited only with simultaneous blocking of GPVI and 2?1, indicating that the integrin can support platelet–collagen binding in the absence of its activation. Blockade or absence of GPIb only moderately influenced integrin activation and adhesion unless GPVI was inhibited.

Conclusions— GPVI- and autocrine-released ADP induce affinity changes of 2?1 and IIb?3 during thrombus formation on collagen under flow. These integrin changes are dispensable for adhesion but strengthen platelet–collagen interactions and thereby collagen-induced platelet activation.

Integrin activation during thrombus formation on collagen was studied using fluorescent-labeled antibodies IAC-1 and PAC1 directed against activation-dependent epitopes of 2?1 and IIb?3 integrin, respectively. Glycoprotein VI blockade by 9O12 antibody or P2Y ADP receptors reduced integrin activation along with aggregate formation and fibrinogen binding but not 2?1-dependent adhesion.

Key Words: ADP ? collagen ? glycoprotein VI ? integrins ? platelets ? thrombus

Introduction

Platelet integrins are critical in hemostasis. Abundantly expressed at the platelet surface, integrins are required for platelet interactions with subendothelial matrix components and for platelet–platelet interactions leading to aggregate and thrombus formation.1 Integrin 2?1 plays a role in platelet adhesion to collagen under static2,3 and flow conditions.4,5 Integrin IIb?3 allows platelets to bind to fibrinogen and von Willebrand factor (vWF) present on collagen and other platelets.6 This leads to stable platelet adhesion and aggregate formation.7

On resting platelets, these integrins are considered to be present in a low-affinity state. Intracellular signaling or ligand binding results in conformational changes of the integrins with a switch to higher-affinity states.6 Agonists such as thrombin, collagen, ADP, and vWF induce IIb?3 activation and platelet aggregation.8,9 Full integrin activation with ADP requires the P2Y1 and P2Y12 purinergic receptors.10,11 Recent studies show that integrin 2?1 can also be activated by inside-out signaling.12,13 Thrombin and collagen turn this integrin into a high-affinity form, whereas ADP changes it to intermediate affinity.13 Although much is known of the affinity and avidity changes of IIb?3 on isolated platelets especially,8,14 regulation of integrin activation during thrombus formation is incompletely understood.

In vivo studies as well as ex vivo experiments, in which blood was allowed to flow over collagen under arterial shear conditions, have indicated that glycoprotein (GP) VI is a principal receptor responsible for collagen-induced activation of platelets during thrombus formation.5,15–21 The 2?1contribution to platelet–collagen interaction has been debated extensively.16 Current evidence with murine and human platelets shows that this integrin functions to reinforce the activating effect of GPVI to produce stable, nonembolizing thrombi.17,22–24 Integrin 2?1, putatively in its activated form, synergizes with GPVI to stimulate Ca2+ signaling, granule secretion, and subsequent aggregate formation.4,15,22,24,26 It also assists GPVI in triggering of the procoagulant platelet response (ie, by stimulating surface exposure of phosphatidylserine ).3,5,17,22 This procoagulant phospholipid strongly potentiates local formation of thrombin and, hence, coagulation.27 In flowing human (but less clearly so in murine) blood, GPVI blockade still allows 2?1-dependent platelet adhesion to collagen.5,22 This raises the question of how the activation state of this integrin relates to its adhesive and signaling function. Under flow over collagen, this may involve the GPIb-V-IX complex, which is another receptor implicated in integrin IIb?3 activation on interaction with vWF.28,29

By using novel antibodies against GPVI and against activation-dependent epitopes on 2?1, we investigated 2?1 and IIb?3 activation during human platelet interaction with collagen and subsequent thrombus formation under flow. We found that GPVI and P2Y receptor stimulation caused activation of either integrin, whereas GPIb contributed to a lesser extent. Surprisingly, we found significant integrin-dependent adhesion in the absence of its activation.

Methods

Materials, blood preparation, and experimental design are available online at http://atvb.ahajournals.org.

Thrombus Formation Under Flow Conditions

Flow experiments over collagen were performed with D-phenylalanyl-L-prolyl-L-arginine chloromethylketone (PPACK)-anticoagulated blood, as described.22 Briefly, whole blood was perfused for 4 minutes over a collagen-coated coverslip through a parallel-plate transparent flow chamber at a wall-shear rate of 150 to 1000 s–1. PS exposure was detected by postperfusion with rinsing Hepes buffer, pH 7.45 (in mmol/L: 136 NaCl, 10 glucose, 5 Hepes, 2.7 KCl, 2 MgCl2, and 2 CaCl2, plus 0.1% BSA and 1 U/mL heparin), containing 1 μg/mL OG488-annexin-A5. Integrin activation was monitored by adding fluorescein isothiocyanate (FITC)-coupled IAC-1 (10 μg/mL) or PAC1 (1:50) to blood before perfusion. Where indicated, probes were added to the rinsing buffer.

High-resolution phase-contrast and fluorescent images were recorded in real time with intensified cameras as described.17 Using Quanticell software (Visitech), fluorescence images (average 32 camera images) were corrected for background fluorescence by subtraction of the mean gray level from an adjacent site of the coverslip. Area coverage by fluorescent platelets was determined with Quanticell software. Area coverage from phase-contrast images was analyzed offline using ImagePro software (Media Cybernetics) by automated threshold settings and application of double masks. Area coverage data were used to determine platelet deposition on collagen because they could be collected simultaneously with, and thus compared with, fluorescence data.

Confocal and 2-Photon Laser Scanning Microscopy

For confocal microscopy, coverslips with thrombi stained with FITC-IAC-1, Gi9-FITC, or MOPC-21-FITC were observed with a Bio-Rad laser-scanning microscope. Where indicated, thrombi were fixed with 2% formaldehyde and blocked with 15% BSA in PBS, pH 7.5. After permeabilization with 0.005% sodium dodecyl sulfate, actin cytoskeletal fibers were labeled with Texas-red phalloidin. An MRC600 laser-scanning microscope system was used, equipped with argon–krypton and red diode lasers. Optical sections (4 to 8 scans) were recorded in Kalman filtering mode.

Two-photon laser scanning microscopy (TPLSM) was with a Bio-Rad 2100 multiphoton system. PPACK blood containing 30% calcein-labeled platelets17 was perfused over collagen, and z-series of scans were recorded during perfusion. Alternatively, blood was supplemented with OG488-fibrinogen (0.25 mg/mL). Excitation was by a Spectra-Physics Tsunami Ti:Sapphire laser, tuned and mode-locked at 800 nm, producing pulses of 100 fs wide (repetition rate 82 MHz). Fluorescence was detected at 508 to 523 nm.

Results

Suppressed Aggregation and PS Exposure by GPVI and ADP Receptor Blockade but Unchanged Platelet Adhesion to Collagen

The monoclonal antibody (mAb) 9O12, directed against the GPVI collagen-binding site, acts as an efficient GPVI antagonist in vitro.30 At 25 μg/mL, antigen-binding fragments (Fabs) of 9O12 specifically blocked platelet adhesion to collagen under static conditions. The fragments also inhibited collagen-induced platelet aggregation and granule secretion. These effects were accompanied by suppression of collagen-induced protein tyrosine phosphorylation (Figure I, available online at http://atvb.ahajournals.org). Surprisingly, 9O12 stimulated phosphorylation of unknown protein bands at 25 and 36 kDa. The mAb binds to the first Ig domain of GPVI, distal from the membrane (C. Lecut and M. Jandrot-Perrus, unpublished data, 2003).

The ability of 9O12 Fab to block GPVI in blood flowing over type-I collagen fibers was investigated at intermediate shear stress (1000 s–1). Under control conditions, perfusion resulted in rapid platelet deposition on collagen, which increased linearly in time for at least 5 minutes, as recorded from phase-contrast images.17,22 In blood supplemented with calcein-loaded platelets, the 3D build-up of thrombus formation was followed by recording of stacks of fluorescence images using the high-resolution technique of TPLSM, which provides higher sensitivity and less bleaching than conventional microscopy. After an initial phase of 2 minutes of adhesion to the collagen, flowing platelets preferentially incorporated into aggregates (Figure II, available online at http://atvb.ahajournals.org). In the presence of 9O12 Fab (50 μg/mL), aggregate formation but not initial adhesion was severely impaired, leaving only 1 to 2 layered platelet groups. As a result, thrombus volume was reduced greatly with 9O12.

Phase-contrast images, recorded after an arbitrary end point of 4 minutes of flow (Figure 1), indicated that surface area covered by aggregated platelets under control conditions was 16.4±1.3% on average (n=11 subjects), with a variation between donors from 9.9% to 25.6%. With 9O12 present, only small aggregates were observed together with numerous single platelets, whereas the area covered by platelets remained unchanged (94.9±7.5% of control condition; Figure 2A). This is an underestimate of the platelet deposition reduction, bearing in mind the abolishment of aggregate formation. Higher concentrations of 9O12 up to 200 μg/mL gave similar results.

Figure 1. Effect of GPVI and P2Y receptor blockade on platelet aggregation, PS exposure, and integrin activation. Whole blood was perfused over collagen at 1000 s–1 for 4 minutes; preincubation with vehicle (control condition), anti-GPVI 9O12 (50 μg/mL), or P2Y blockers (40 μmol/L MRS2179 and 20 μmol/L AR-C69931MX). A, Representative phase-contrast images (120x120 μm) after perfusion. Note the presence of collagen fibers. Representative fluorescence images (150x150 μm) after perfusion with: B, OG488-annexin-A5 to measure PS exposure; C, FITC-IAC-1 to detect activated 2?1; or D, FITC-PAC1 to detect activated IIb?3.

Figure 2. Quantitative effect of GPVI and P2Y receptor blockade on platelet aggregation, PS exposure, and integrin activation. Whole blood was perfused over collagen, and platelets were stained with fluorescent annexin-A5, IAC-1, or PAC-1 (Figure 1). Blood was untreated (control) or treated with anti-GPVI Fab 9O12 (50 μg/mL). Alternatively, blood was treated with P2Y blockers (40 μmol/L MRS2179 and 20 μmol/L AR-C69931MX) in combination with anti-GPVI Fab and ASA; with P2Y blockers alone; or with P2Y blockers and ASA. A, Surface area coverage of all platelets and PS-exposing platelets. B, Surface area coverage of FITC-IAC-1 and PAC-1 staining detecting activated 2?1 and IIb?3, respectively. Per parameter, data were expressed as percentages of respective control condition set at 100% (mean±SEM; n=4 to 5; *P<0.05; **P<0.01 vs control).

GPVI-induced PS exposure of the platelets on collagen was monitored by staining with OG488-annexin-A5.22,31 Under control conditions, many platelets bound annexin-A5, giving an area coverage with fluorescence of 10.5±1.5% (n=8). These platelets had a round, blebbing structure, as described previously.17 With 50 μg/mL 9O12 Fab present, annexin-A5 staining reduced greatly to 15% of the control situation (Figures 1B and 2A), along with platelet blebbing. Thus, 9O12 suppressed collagen-induced aggregate formation and PS exposure under flow (both effects of GPVI-induced platelet activation) but not platelet adhesion to collagen. This was in agreement with earlier results obtained with the anti-GPVI scFv antibody 10B12.17

Blockers of P2Y1 (MRS2179) and P2Y12 (AR-C6991MX) receptors25 were used to evaluate the contribution of autocrine ADP in platelet deposition under flow. With the P2Y blockers present, platelets deposited as single cells or as small 2-layered aggregates (Figure 1A). Surface area covered with platelets was increased slightly to 131.0±19.9% of control (Figure 2A), likely because of increased contacts of platelets with collagen fibers. Annexin-A5 fluorescence was increased similarly to 128.4±29.5% of control (Figures 1B and 2A). In the presence of P2Y blockers, adhesion and PS exposure were about halved when the blood was also pretreated with acetylsalicylic acid (ASA) to block thromboxane formation (Figure 2A), indicating that released thromboxane is involved in platelet adhesion.

Platelet deposition was reduced further when 9O12 was combined with P2Y blockers and ASA. This resulted in adhesion of merely single platelets (28.9±4.8% of control), whereas annexin-A5 staining was abolished completely (Figure 2A). This extends earlier observation14 and indicates that human GPVI, together with the secondary mediators ADP and thromboxane, is responsible for aggregate formation and PS exposure but is dispensable for platelet adhesion.

Suppressed 2?1 and IIb?3 Activation by Blockade of GPVI or ADP Receptors

To study 2?1 activation, a new antibody IAC-1 was used, which specifically recognizes high-affinity forms of this integrin.32 IAC-1 does not bind to resting platelets but readily binds to a neoepitope in the 2 I-domain that become exposed during platelet activation with ADP, thromboxane, or thrombin. Activation uncovers an I-domain region at amino acids 199 to 201, which is located at the opposite site of the metal ion-dependent adhesion site domain that is involved in binding to collagen. IAC-1 is thus of little effect on collagen binding.32 FITC-labeled IAC-1 binds to platelets on immobilized convulxin (data not shown), indicating that 2?1is activated after platelet adhesion via GPVI.

When added to blood at 10 μg/mL, FITC-labeled IAC-1 did not inhibit platelet deposition on collagen under flow (surface area coverage with all platelets 13.2±1.4% versus 13.6±1.1% in the absence of IAC-1; n=7). Yet, FITC-IAC-1 avidly bound to the platelets adhering to collagen under control conditions but not in the presence of platelet inhibitors (Figure III, available online at http://atvb.ahajournals.org). Other experiments were performed with FITC-Gi9, a mAb recognizing all 2?1 forms, and with FITC-MOPC-21, a mAb that does not bind to platelets. FITC-Gi9 stained control and inhibited platelets, but FITC-MOPC-21 failed to give detectable staining (Figure III).

During perfusions under control condition, fluorescence staining of FITC-IAC-1 increased in time; it covered 9.5±2.1% (n=8 subjects) of the surface area after 4 minutes. Addition of 9O12 (50 μg/mL) greatly reduced staining to 23±1.9% of the control condition, despite the presence of many adherent platelets (Figures 1C and 2B). Addition of P2Y blockers also reduced IAC-1 staining to 41.2±22.1% of control. Blockade of GPVI and P2Y and presence of ASA resulted in almost complete suppression of IAC-1 fluorescence (Figure 2B).

To ensure that the reduced IAC-1 binding was not caused by a low detection level of the fluorescence camera, platelets were counterstained for actin with fluorescent phalloidin and examined by confocal microscopy. Under control conditions, individual platelets in aggregates were strongly labeled with FITC-IAC-1, as apparent from overlays of IAC-1 and phalloidin images (Figure IV, available online at http://atvb.ahajournals.org). With 9O12 present, FITC-IAC-1 fluorescence of collagen-adherent platelets was reduced greatly, in contrast to the still bright phalloidin staining.

Because 9O12 inhibited platelet aggregation, its effect was studied on signaling toward integrin IIb?3 using fluorescent-labeled PAC1, which is a mAb against activated IIb?3.33 Under control conditions, addition of FITC-PAC1 to blood or postperfusion with FITC-PAC1 gave fluorescent-labeled platelet aggregates on collagen (Figure 1D). After 4 minutes of perfusion, staining with FITC-PAC1 was 7.5±2.1% (n=8) of the surface. Blocking of GPVI or P2Y receptors decreased PAC1 staining (Figure 2B). Similarly, as observed with IAC-1, the combination of GPVI and P2Y blockers plus ASA almost completely suppressed PAC-1 staining.

To verify that ADP could activate integrins on collagen-adherent platelets in the absence of GPVI activity, 9O12-treated platelets were postperfused with 10 μmol/L ADP. Perfusion with ADP but not with vehicle gave a substantial 5-fold increase in IAC-1-labeling or PAC1-labeling of adherent platelets (Figure 3).

Figure 3. Effect of ADP addition on integrin activation of GPVI-inhibited platelets. Blood containing 50 μg/mL 9O12 was perfused over collagen, and FITC-PAC1 or FITC-IAC-1 binding was measured. Thereafter, vehicle buffer or ADP (10 μmol/L) with same fluorescent antibody was superfused at 1000 s–1 for 1 minute. Percentages of area coverage with fluorescence are given at respective control condition (mean±SEM; n=3; *P<0.05 vs 9O12 alone).

Addition of OG488-labeled human fibrinogen to the blood provided another means to measure IIb?3 activation on flow. TPLSM with low detection threshold showed that platelet aggregates on collagen were labeled diffusely with OG488-fibrinogen (Figure 4A and 4B). In the presence of 9O12, only some of the single, collagen-adherent platelets showed fluorescence. Total surface area coverage with OG488-fibrinogen decreased greatly to 13.0±2.8% of control (Figure 4C). P2Y blockers were less effective in reducing OG488-fibrinogen binding (ie, to 42.4±5.0% of control); many individual platelets still bound fibrinogen (Figure 4B). With 9O12, ASA, and P2Y blockers present, fibrinogen binding was no longer detected (data not shown).

Figure 4. Effect of GPVI and P2Y receptor blockade on platelet–fibrinogen binding. Blood with OG488-labeled fibrinogen (0.25 mg/mL) was perfused for 4 minutes over collagen in the presence of vehicle (control), anti-GPVI 9O12, or P2Y blockers (Figure 1). A, Phase-contrast images (120x120 μm) and B, merged stacks of TPLSM images (110x110 μm) after perfusion. C, Quantitative effect of receptor blockade on OG488-fibrinogen fluorescence. Data are percentages of control (mean±SEM; n=3; **P<0.01).

These results indicate that antagonism of GPVI or the P2Y receptors severely impaired exposure of activation-dependent epitopes on 2?1 and IIb?3. Only combined blockade of GPVI and secondary mediators resulted in full inhibition of integrins. This confirms that autocrine ADP and thromboxane play key roles in thrombus build-up24,25 and reveals that these mediators, along with GPVI, mediate 2?1 and IIb?3 activation.

GPIb Involved in Collagen Platelet Adhesion but not in Integrin Activation

GPIb-IX-V mediates platelet adhesion to collagen under shear. Its role in integrin activation was studied using the anti-GPIb mAb 12G1, which specifically hinders shear-dependent adhesion to vWF.34 At maximally effective concentration of 40 μg/mL, 12G1 F(ab')2 completely blocked adhesion of platelets to immobilized vWF. However, the F(ab')2 only partially reduced adhesion to collagen fibers at 1000 s–1 (Figure V, available online at http://atvb.ahajournals.org). Although less platelet-collagen contacts were formed, aggregate formation was not prevented; the area covered by platelets remained 72.2±10.6% (n=7) of control. GPIb blockade with 12G1 inhibited staining with OG488-annexin-A5 to 41.7±3.3% of control (Figure V). This intervention reduced staining with FITC-PAC1 and FITC-IAC-1 only moderately to 70.7±11.4% and 73.8±16.3% of control, respectively. However, with 9O12 present, 12G1 markedly inhibited platelet deposition to 12.0±5.3% of control.

The results with blocking antibody were corroborated by studies with blood from a patient with Bernard-Soulier syndrome, displaying GPIb-deficient platelets (Figure 5A). After 4 minutes of perfusion, surface area covered with patient platelets was 7.6±0.9% (ie, slightly lower than the averaged value for healthy subjects ). With 9O12, adhesion of patient platelets was almost abolished to 8.2±1.8% of control conditions (Figure 5A and 5B). Thus, at this shear rate, GPIb and GPVI together determine adhesion to collagen, but GPIb does not have a major role in integrin activation.

Figure 5. Abolished adhesion by blockade of GPVI of platelets from Bernard-Soulier (BSS) patient. Blood from a healthy control subject or a patient displaying BSS was perfused over collagen with/without 9O12 (50 μg/mL). A, Representative phase-contrast images (120x120 μm) of deposited platelets; inserts are 3-fold magnifications showing giant size of BSS platelets. B, Quantitative effect on platelet deposition. Values are percentage of surface area coverage (mean±SEM; n=3; *P<0.05; **P<0.01 vs absence of 9O12).

Functional Importance of 2?1 Activation

Under control conditions, most platelets on collagen displayed pseudopods, lamellipods, or blebs. When 9O12 was combined with ASA and P2Y blockers, platelets adhered individually (coverage 28.9±4.8% of control) and did not show morphological signs of activation (Figure 6). The remaining adhesion was integrin-dependent because blocking anti-2?1 mAb 6F1 (10 μg/mL) severely abrogated adhesion with 9O12 (18.0±3.5% versus control; n=4), and extra addition of IIb?3-blocking arg-gly-asp-ser (400 μmol/L) eliminated all platelet deposition (<5%). Thus, integrins participated in platelet adhesion in the absence of detectable binding of IAC-1, PAC1, or fibrinogen.

Figure 6. Platelet shape and activation of integrin 2?1. Blood was perfused over collagen (Figure 2). Blood was treated with anti-2?1 6F1 (10 μg/mL), anti-GPVI 9O12 (50 μg/mL), or P2Y blockers (MRS2179+AR-C69931MX) plus ASA, as indicated. Shown are representative phase-contrast images (bar=10 μm). White arrows indicate pseudopods; black arrows, blebbing platelets.

Experiments in which 6F1 was added to the blood informed on the functional importance of (activated) 2?1. Blocking of 2?1 with 6F1 notably reduced pseudopod formation but still allowed bleb formation (Figure 6). The 6F1 effect was complete with P2Y blockers present as well (only blebs formed because of GPVI activation). Pseudopod and lamellipod formation was restored when 9O12-treated platelets were postperfused with ADP (data not shown). This suggested that 2?1-dependent pseudopod formation, which correlated with IAC-1 binding, was responsible for increased platelet–collagen contact.

Discussion

In this study, we used newly developed tools to determine the role of human GPVI and ADP in integrin activation during collagen-induced thrombus formation under flow. The O12 mAb, directed against the collagen-binding site of human GPVI, was used to block GPVI activity. This inhibited formation of platelet aggregates and staining with annexin-A5 (detecting PS exposure) and IAC-1 (detecting activated 2?1), as well as fibrinogen and PAC1 (detecting activated IIb?3). FITC-IAC-1 is the first described mAb to specifically recognize high activation states of 2?1.32 Blockade of the P2Y1 and P2Y12 receptors partially inhibited binding of IAC-1, PAC1, and fibrinogen to platelets, but blockade of GPVI and P2Y receptors in combination with ASA treatment was needed to abolish all binding completely. Conversely, postperfusion with ADP resulted in increased IAC-1 and PAC1 binding to GPVI-inhibited platelets. Together, these results provide direct evidence for a role of GPVI and autocrine ADP in inside-out integrin signaling. The inhibitory effects of 9O12 are consistent with studies using isolated platelets showing that stimulation with GPVI agonists results in integrin activation6,13 and exposure of procoagulant PS.27 In addition, they extend recent evidence that GPVI and 2?1 contribute to human thrombus formation.17,23

Although anti-GPVI 9O12 efficiently suppressed aggregation, procoagulant activity, and integrin activation under flow, it did not abolish platelet adhesion to collagen, even not at a high dose of 200 μg/mL. This is remarkably similar to the effects of anti-human GPVI scFv, 10B12, which is also directed against the collagen-binding domain of GPVI.17 Thus, 2 different antibodies against human GPVI appear to suppress platelet activation under shear but not adhesion to collagen. However, we find that combined blockade of human GPVI and ADP/thromboxane effects does lower the adhesion. For mouse blood, this is less clear because blockade or absence of murine GPVI has been found to either abolish5,18,22 or still permit21 platelet–collagen adhesion under flow. This discrepancy is probably attributable to different experimental conditions.

At the moderately high shear rate of 1000 s–1 used, the anti-GPIb mAb 12G1 only partially reduced platelet adhesion to collagen/vWF. When given alone, 12G1 inhibited adhesion slightly and hardly influenced binding of IAC-1 and PAC1. However, in combination with GPVI blockade, GPIb antagonism or absence of GPIb (in a Bernard-Soulier patient) completely abolished platelet adhesion to collagen. This indicates that GPIb-V-IX only partially contributes to integrin activation under conditions in which GPVI and P2Y receptors are also signaling.

Adhesion of platelets treated with GPVI and P2Y antagonists was mostly blocked when anti-2?1 mAb 6F1 was added. Furthermore, staining of platelets with IAC-1 or PAC1 correlated with 2?1-dependent pseudopod formation of the platelets on collagen. These observations suggest that platelet adhesion to collagen can occur under conditions in which the IAC-1/PAC1 epitopes are not or only partially exposed (ie, with no or local integrin activation), basically in agreement with earlier suggestions by Inoue et al.35 In conclusion, our results indicate that GPVI is responsible for integrin affinity regulation on platelet adhesion to collagen under high shear. Furthermore, autocrine released ADP and subsequent engagement of P2Y receptors play assisting roles. Thus GPVI- and P2Y-coupled signaling act synergistically to achieve full integrin activation and thereby stable thrombus formation.

Acknowledgments

C.L. was supported by a Marie-Curie Fellowship from the European Community. We acknowledge grant 902-16 to 276 from the Netherlands Organization for Scientific Research. We thank AgroBio for production of 9O12 ascites.

References

Ruggeri ZM. Platelets in atherothrombosis. Nat Med. 2002; 8: 1227–1234.

Moroi M, Jung SM. Platelet receptors for collagen. Thromb Haemost. 1997; 78: 439–444.

Siljander P, Farndale RW, Feijge MAH, Comfurius P, Kos S, Bevers EM, Heemskerk JWM. Platelet adhesion enhances glycoprotein VI-dependent procoagulant response. Arterioscler Thromb Vasc Biol. 2001; 21: 618–627.

Savage B, Ginsberg MH, Ruggeri ZM. Influence of fibrillar collagen structure on the mechanisms of platelet thrombus formation under flow. Blood. 1999; 94: 2704–2715.

Nieswandt B, Brakebusch C, Bergmeier W, Schulte V, Bouvard D, Mohtari-Nejad R, Lindhout T, Heemskerk JWM, Zirngibl H, F?ssler R. Glycoprotein VI but not 2?1 integrin is essential for platelet interaction with collagen. EMBO J. 2001; 20: 2120–2130.

Shattil SJ, Kashiwagi H, Pampori N. Integrin signaling, the platelet paradigm. Blood. 1998; 91: 2645–2657.

Savage B, Sixma JJ, Ruggeri ZM. Functional self-association of von Willebrand factor during platelet adhesion under flow. Proc Natl Acad Sci U S A. 2002; 99: 425–430.

Shattil SJ, Ginsberg MH. Integrin signaling in vascular biology. J Clin Invest. 1997; 100: 1–5.

Kasirer-Friede A, Ware J, Leng L, Marchese P, Ruggeri ZM, Shattil SJ. Lateral clustering of platelet GP Ib-IX complexes leads to up-regulation of adhesive function of integrin IIb?3. J Biol Chem. 2002; 277: 11949–11956.

Daniel JL, Dangelmaier C, Jin JG, Kim YB, Kunapuli SP. Role of intracellular signaling events in ADP-induced platelet aggregation. Thromb Haemost. 1999; 82: 1322–1326.

Eckly A, Gendrault JL, Hechler B, Ceazenave JP, Gachet C. Differential involvement of P2Y1 and P2YT receptors in morphological changes of platelet aggregation. Thromb Haemost. 2001; 85: 694–701.

Wang Z, Leisner TM, Parise LV. Platelet 2?1 integrin activation: contribution of ligand internalization and the 2-cytoplasmic domain. Blood. 2003; 102: 1307–1315.

Jung SM, Moroi M. Platelet collagen receptor integrin 2?1 activation involves differential participation of ADP-receptor subtypes P2Y1 and P2Y12. Eur J Biochem. 2001; 268: 3513–3522.

Takagi J, Petre BM, Walz T, Springer TA. Global conformational rearrangements in integrin extracellular domains in outside-in and inside-out signaling. Cell. 2002; 110: 599–511.

Goto S, Tamura N, Handa S, Arai M, Kodama K, Takayama H. Involvement of glycoprotein VI in platelet thrombus formation on both collagen and von Willebrand factor surfaces under flow conditions. Circulation. 2002; 106: 266–272.

Nieswandt B, Watson SP. Platelet collagen interaction: is GPVI the central receptor? Blood. 2003; 102: 449–461.

Siljander PRM, Munnix ICA, Smethurst PA, Deckmyn H, Lindhout T, Ouwehand WH, Farndale RW, Heemskerk JWM. Platelet receptor interplay regulates collagen-induced thrombus formation in flowing human blood. Blood. 2004; 103: 1333–1341.

Massberg S, Gawaz M, Grüner S, Schulte V, Konrad I, Zohlh?fer D, Heinzmann U, Nieswandt B. A crucial role of glycoprotein VI for platelet recruitment to the injured arterial wall in vivo. J Exp Med. 2003; 197: 41–49.

Chen H, Locke D, Liu Y, Liu C, Kahn ML. The platelet receptor GPVI mediates both adhesion and signaling responses to collagen in receptor density-dependent fashion. J Biol Chem. 2002; 277: 3011–3019.

Best D, Senis YA, Jarvis GE, Eagleton HJ, Roberts DJ, Saito T, Jung SM, Moroi M, Harrison P, Green FR, Watson SP. GPVI levels in platelets: relationship to platelet function at high shear. Blood. 2003; 102: 2811–2818.

Kato K, Kanaji T, Russell S, Kunicki TJ, Furihata K, Kanaji S, Marchese P, Reininger A, Ruggeri ZM, Ware J. The contribution of glycoprotein VI to stable platelet adhesion and thrombus formation illustrated by targeted gene deletion. Blood. 2003; 102: 170–1707.

Kuijpers MJE, Schulte V, Bergmeier W, Lindhout T, Brakebusch C, Offermanns S, F?ssler R, Heemskerk JWM, Nieswandt B. Complementary roles of glycoprotein VI and 2?1 integrin in collagen-induced thrombus formation in flowing whole blood ex vivo. FASEB J. 2003; 17: 685–687.

Chen H, Kahn ML. Reciprocal signaling by integrin and nonintegrin receptors during collagen activation of platelets. Mol Cell Biol. 2003; 23: 4764–4777.

Suzuki-Inoue K, Inoue O, Frampton J, Watson SP. Murine GPVI stimulates weak integrin activation in PLC2 –/– platelets. Blood. 2003; 102: 1367–1373.

Turner NA, Moake JL, McIntire LV. Blockade of ADP receptors P2Y12 and P2Y1 is required to inhibit platelet aggregation in whole blood under flow. Blood. 2001; 98: 3340–3345.

Remijn JA, Wu YP, Heninga EH, IJsseldijk MJW, van Willigen G, de Groot PG, Sixma JJ, Nurden AT, Nurden P. Role of ADP receptor P2Y12 in platelet adhesion and thrombus formation in flowing blood. Arterioscler Thromb Vasc Biol. 2002; 22: 686–691.

Heemskerk JWM, Bevers EM, Lindhout T. Platelet activation and blood coagulation. Thromb Haemost. 2002; 88: 186–193.

Gu MY, Xi XD, Englund GD, Berndt MC, Du XP. Analysis of the roles of 14-3-3 in platelet glycoprotein Ib-IX-mediated activation of integrin IIb?3 using a reconstituted mammalian cell expression model. J Cell Biol. 1999; 147: 1085–1096.

Yap CL, Hughan SC, Cranmer SL, Nesbitt WS, Rooney MM, Giuliano S, Kulkarni S, Dopheide SM, Yuan YP, Salem HH, Jackson SP. Synergistic adhesive interactions and signaling mechanisms operating between platelet glycoprotein Ib/IX and integrin IIb?3. J Biol Chem. 2000; 275: 41377–41388.

Lecut C, Feeney LA, Kingsbury G, Hopkins J, Lanza F, Gachet C, Villeval JL, Jandrot-Perrus M. Human platelet glycoprotein VI function is antagonised by monoclonal antibody-derived Fab fragments. J Thromb Haemost. 2003; 1: 2653–2662.

Heemskerk JWM, Vuist WMJ, Feijge MAH, Reutelingsperger CPM, Lindhout T. Collagen but not fibrinogen surfaces induce bleb formation, exposure of phosphatidylserine and procoagulant activity of adherent platelets. Blood. 1997; 90: 2615–2625.

Schoolmeester A, Vanhoorelbeke K, Depraetere H, Feys HB, Heemskerk JWM, Hoylaerts M, Deckmyn H. Monoclonal antibody IAC-1 is specific for activated 2?1 and binds to amino acid 199–201 of the integrin 2 I-domain. Blood 2004;in press.

Abrams C, Deng YJ, Steiner B, O’Toole T, Shattil SJ. Determinants of specificity of a baculovirus-expressed antibody Fab fragment that binds selectively to activated form of integrin IIb?3. J Biol Chem. 1994; 269: 18781–18788.

Cauwenberghs N, Vanhoorelbeke K, Vauterin S, Westra DF, Romo G, Huizinga EG, Lopez JA, Berndt MC, Harsfalvi J, Deckmyn H. Epitope mapping of inhibitory antibodies against platelet glycoprotein Ib reveals interaction between the leucine-rich repeat N-terminal and C-terminal flanking domains of glycoprotein Ib. Blood. 2001; 98: 652–660.

Inoue O, Suzuki-Inoue K, Dean WL, Frampton J, Watson SP. Integrin 2?1 mediates outside-in regulation of platelet spreading on collagen through activation of Src kinases and PLC2. J Cell Biol. 160; 769–780.

日期：2007年5月18日 - 来自[2004年第24卷第9期]栏目

A Crucial Role of Glycoprotein VI for Platelet Recruitment to the Injured Arterial Wall In Vivo

¹ Deutsches Herzzentrum und 1. Medizinische Klinik, Technische Universität München, D-80636 München, Germany
² Rudolf Virchow Center for Experimental Biomedicine, University of Würzburg, 97078 Würzburg, Germany
³ GSF National Research Center for Environment and Health, Institute of Pathology, D-85764 Neuherberg, Germany

Top
Abstract
Introduction
Materials and Methods
Results and Discussion
References

Ruggeri, Z.M. 1997. Mechanisms initiating platelet thrombus formation. Thromb. Haemost. 78:611–616.
van Zanten, G.H., S. de Graaf, P.J. Slootweg, H.F. Heijnen, T.M. Connolly, P.G. de Groot, and J.J. Sixma. 1994. Increased platelet deposition on atherosclerotic coronary arteries. J. Clin. Invest. 93:615–632.
Clemetson, K.J., and J.M. Clemetson. 2001. Platelet collagen receptors. Thromb. Haemost. 86:189–197.
Baumgartner, H.R. 1977. Platelet interaction with collagen fibrils in flowing blood. I. Reaction of human platelets with alpha chymotrypsin-digested subendothelium. Thromb. Haemost. 37:1–16.
Savage, B., F. Almus-Jacobs, and Z.M. Ruggeri. 1998. Specific synergy of multiple substrate-receptor interactions in platelet thrombus formation under flow. Cell. 94:657–666.
Santoro, S.A. 1986. Identification of a 160,000 dalton platelet membrane protein that mediates the initial divalent cation-dependent adhesion of platelets to collagen. Cell. 46:913–920.
Moog, S., P. Mangin, N. Lenain, C. Strassel, C. Ravanat, S. Schuhler, M. Freund, M. Santer, M. Kahn, B. Nieswandt, et al. 2001. Platelet glycoprotein V binds to collagen and participates in platelet adhesion and aggregation. Blood. 98:1038–1046.
Moroi, M., S.M. Jung, M. Okuma, and K. Shinmyozu. 1989. A patient with platelets deficient in glycoprotein VI that lack both collagen-induced aggregation and adhesion. J. Clin. Invest. 84:1440–1445.
Gibbins, J.M., M. Okuma, R. Farndale, M. Barnes, and S.P. Watson. 1997. Glycoprotein VI is the collagen receptor in platelets which underlies tyrosine phosphorylation of the Fc receptor gamma-chain. FEBS Lett. 413:255–259.
Nieswandt, B., W. Bergmeier, V. Schulte, K. Rackebrandt, J.E. Gessner, and H. Zirngibl. 2000. Expression and function of the mouse collagen receptor glycoprotein VI is strictly dependent on its association with the FcRgamma chain. J. Biol. Chem. 275:23998–24002.
Nieswandt, B., V. Schulte, W. Bergmeier, R. Mokhtari-Nejad, K. Rackebrandt, J.P. Cazenave, P. Ohlmann, C. Gachet, and H. Zirngibl. 2001. Long-term antithrombotic protection by in vivo depletion of platelet glycoprotein VI in mice. J. Exp. Med. 193:459–469.
Nieswandt, B., C. Brakebusch, W. Bergmeier, V. Schulte, D. Bouvard, R. Mokhtari-Nejad, T. Lindhout, J.W. Heemskerk, H. Zirngibl, and R. Fassler. 2001. Glycoprotein VI but not alpha2beta1 integrin is essential for platelet interaction with collagen. EMBO J. 20:2120–2130.
Clemetson, J.M., J. Polgar, E. Magnenat, T.N. Wells, and K.J. Clemetson. 1999. The platelet collagen receptor glycoprotein VI is a member of the immunoglobulin superfamily closely related to FcalphaR and the natural killer receptors. J. Biol. Chem. 274:29019–29024.
Jandrot-Perrus, M., S. Busfield, A.H. Lagrue, X. Xiong, N. Debili, T. Chickering, J.P. Le Couedic, A. Goodearl, B. Dussault, C. Fraser, et al. 2000. Cloning, characterization, and functional studies of human and mouse glycoprotein VI: a platelet-specific collagen receptor from the immunoglobulin superfamily. Blood. 96:1798–1807.
Watson, S.P., N. Asazuma, B. Atkinson, O. Berlanga, D. Best, R. Bobe, G. Jarvis, S. Marshall, D. Snell, M. Stafford, et al. 2001. The role of ITAM- and ITIM-coupled receptors in platelet activation by collagen. Thromb. Haemost. 86:276–288.
Chen, H., D. Locke, Y. Liu, C. Liu, and M.L. Kahn. 2002. The platelet receptor GPVI mediates both adhesion and signaling responses to collagen in a receptor density-dependent fashion. J. Biol. Chem. 277:3011–3019.
Nieswandt, B., W. Bergmeier, K. Rackebrandt, J.E. Gessner, and H. Zirngibl. 2000. Identification of critical antigen-specific mechanisms in the development of immune thrombocytopenic purpura in mice. Blood. 96:2520–2527.
Massberg, S., G. Enders, R. Leiderer, S. Eisenmenger, D. Vestweber, F. Krombach, and K. Messmer. 1998. Platelet-endothelial cell interactions during ischemia/reperfusion: the role of P-selectin. Blood. 92:507–515.
Schulte, V., D. Snell, W. Bergmeier, H. Zirngibl, S.P. Watson, and B. Nieswandt. 2001. Evidence for two distinct epitopes within collagen for activation of murine platelets. J. Biol. Chem. 276:364–368.
Massberg, S., G. Enders, F.C. Matos, L.I. Tomic, R. Leiderer, S. Eisenmenger, K. Messmer, and F. Krombach. 1999. Fibrinogen deposition at the postischemic vessel wall promotes platelet adhesion during ischemia-reperfusion in vivo. Blood. 94:3829–3838.
Fay, W.P., A.C. Parker, M.N. Ansari, X. Zheng, and D. Ginsburg. 1999. Vitronectin inhibits the thrombotic response to arterial injury in mice. Blood. 93:1825–1830.
Lindner, V., J. Fingerle, and M.A. Reidy. 1993. Mouse model of arterial injury. Circ. Res. 73:792–796.
Savage, B., E. Saldivar, and Z.M. Ruggeri. 1996. Initiation of platelet adhesion by arrest onto fibrinogen or translocation on von Willebrand factor. Cell. 84:289–297.
Baumgartner, H.R., R. Muggli, T.B. Tschopp, and V.T. Turitto. 1976. Platelet adhesion, release and aggregation in flowing blood: effects of surface properties and platelet function. Thromb. Haemost. 35:124–138.
Bergmeier, W., K. Rackebrandt, W. Schroder, H. Zirngibl, and B. Nieswandt. 2000. Structural and functional characterization of the mouse von Willebrand factor receptor GPIb-IX with novel monoclonal antibodies. Blood. 95:886–893.
Goto, S., Y. Ikeda, E. Saldivar, and Z.M. Ruggeri. 1998. Distinct mechanisms of platelet aggregation as a consequence of different shearing flow conditions. J. Clin. Invest. 101:479–486.
Sixma, J.J., G.H. van Zanten, J.D. Banga, H.K. Nieuwenhuls, and P.G. de Groot. 1995. Platelet adhesion. Semin. Hematol. 32:89–98.
Zhu, B., D.G. Kuhel, D.P. Witte, and D.Y. Hui. 2000. Apolipoprotein E inhibits neointimal hyperplasia after arterial injury in mice. Am. J. Pathol. 157:1839–1848.
Rekhter, M.D. 1999. Collagen synthesis in atherosclerosis: too much and not enough. Cardiovasc. Res. 41:376–384.
Opsahl, W.P., D.J. DeLuca, and L.A. Ehrhart. 1987. Accelerated rates of collagen synthesis in atherosclerotic arteries quantified in vivo. Arteriosclerosis. 7:470–476.
Holtkotter, O., B. Nieswandt, N. Smyth, W. Muller, M. Hafner, V. Schulte, T. Krieg, and B. Eckes. 2002. Integrin alpha 2-deficient mice develop normally, are fertile, but display partially defective platelet interaction with collagen. J. Biol. Chem. 277:10789–10794.
Kasirer-Friede, A., J. Ware, L. Leng, P. Marchese, Z.M. Ruggeri, and S.J. Shattil. 2002. Lateral clustering of platelet GP Ib-IX complexes leads to up-regulation of the adhesive function of Integrin alphaIIbbeta3. J. Biol. Chem. 277:11949–11956.
Marchese, P., E. Saldivar, J. Ware, and Z.M. Ruggeri. 1999. Adhesive properties of the isolated amino-terminal domain of platelet glycoprotein Ibalpha in a flow field. Proc. Natl. Acad. Sci. USA. 96:7837–7842.
Du, X., E. Magnenat, T.N. Wells, and K.J. Clemetson. 2002. Alboluxin, a snake C-type lectin from Trimeresurus albolabris venom is a potent platelet agonist acting via GPIb and GPVI. Thromb. Haemost. 87:692–698.
Dormann, D., J.M. Clemetson, A. Navdaev, B.E. Kehrel, and K.J. Clemetson. 2001. Alboaggregin A activates platelets by a mechanism involving glycoprotein VI as well as glycoprotein Ib. Blood. 97:929–936.

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion References

Platelet adhesion and aggregation at sites of vascular injuryis crucial for hemostasis but may lead to arterial occlusionin the setting of atherosclerosis and precipitate diseases suchas myocardial infarction. A current hypothesis suggests thatplatelet glycoprotein (GP) Ib interaction with von Willebrandfactor recruits flowing platelets to the injured vessel wall,where subendothelial fibrillar collagens support their firmadhesion and activation. However, so far this hypothesis hasnot been tested in vivo. Here, we demonstrate by intravitalfluorescence microscopy of the mouse carotid artery that inhibitionor absence of the major platelet collagen receptor, GPVI, abolishesplatelet–vessel wall interactions after endothelial denudation.Unexpectedly, inhibition of GPVI by the monoclonal antibodyJAQ1 reduced platelet tethering to the subendothelium by 89%.In addition, stable arrest and aggregation of platelets wasvirtually abolished under these conditions. Using differentmodels of arterial injury, the strict requirement for GPVI inthese processes was confirmed in GPVI-deficient mice, whereplatelets also failed to adhere and aggregate on the damagedvessel wall. These findings reveal an unexpected role of GPVIin the initiation of platelet attachment at sites of vascularinjury and unequivocally identify platelet–collagen interactions(via GPVI) as the major determinant of arterial thrombus formation.

Key Words: arterial thrombosis • collagen • receptor • GPVI • mouse

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

Platelet adhesion and aggregation is essential to limit bloodloss at sites of vascular injury but may also lead to arterialocclusion and irreversible tissue damage after disruption ofthe atherosclerotic plaque. The first platelet response to vascularinjury is adhesion to the exposed subendothelial matrix, whichtriggers subsequent platelet aggregation. A current hypothesissupported by numerous in vitro studies suggests that the interactionof glycoprotein (GP)Ib-V-IX with von Willebrand factor (vWf)recruits flowing platelets to the injured vessel wall (1), wherethey interact with exposed extracellular matrix proteins resultingin firm adhesion and thrombus growth (2, 3). Although severalof the macromolecular components of the subendothelial layersuch as laminin, fibronectin, and vWf all provide a suitablesubstrate for platelet adhesion, fibrillar collagen is consideredthe most thrombogenic constituent of the vascular subendotheliumas it not only supports platelet adhesion but also acts as astrong activator of platelets in vitro (3, 4). However, thein vivo relevance of platelet–collagen interactions inthe setting of arterial thrombosis has not been established.This might be explained by the complexity of the platelet–collageninteraction, which involves a variety of different receptorsand signaling pathways making the in vivo inhibition of thisprocess very difficult. Besides GPIb-V-IX and integrin _IIbß₃,which interact indirectly with collagen via vWf (5), a largenumber of collagen receptors have been identified on platelets,including most importantly integrin ₂ß₁ (6), GPV (7),and GPVI (8).

Only recently, GPVI has been established as the central plateletcollagen receptor that is essential for platelet adhesion andaggregation on immobilized collagen in vitro, as it mediatesthe activation of different adhesive receptors, including integrins_IIbß₃ and ₂ß₁ (9–12). GPVI is a 60–65-kDtype I transmembrane GP belonging to the immunoglobulin superfamily(13, 14) that forms a complex with the FcR chain at the cellsurface in human and mouse platelets (9, 10). Signaling throughGPVI occurs via a pathway similar to that used by immunoreceptorsas revealed by the tyrosine phosphorylation of the FcR chainimmunoreceptor tyrosine-based activation motif by a src-likekinase (15). The mAb JAQ1 (10) blocks the major collagen bindingsite on mouse GPVI and inhibits firm platelet adhesion to collagenunder low and high shear flow conditions (12). In vivo applicationof JAQ1 induces virtually complete internalization and degradationof GPVI on mouse platelets resulting in a "GPVI knockout"–likephenotype for at least 2 wk. Such GPVI-depleted mice have significantlyprolonged bleeding times and their platelets fail to respondto collagen but not to other agonists (11). Despite its essentialrole in collagen-induced activation of platelets, there hasbeen only very limited evidence for a role of GPVI as a directadhesion receptor (14, 16).

In this study we investigated the in vivo significance of platelet–collageninteractions in the dynamic process of platelet adhesion andaggregation at sites of arterial injury. We show that inhibitionor deletion of GPVI virtually abrogates stable platelet adhesionand aggregation after endothelial denudation of the carotidartery in mice. Very unexpectedly, we found that tethering/slowsurface translocation of platelets was also strongly inhibitedin the absence of functional GPVI. These findings reveal a crucialrole of GPVI in the initiation of platelet recruitment at sitesof vascular injury and provide the first in vivo evidence thatplatelet–collagen interactions are of paramount importanceduring arterial thrombus formation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References