Plasmin-Cleaved ß--Glycoprotein Is an Inhibitor of Angiogenesis

【摘要】  ß-2-Glycoprotein 1, an abundant plasma glycoprotein, binds anionic cell surfaces and functions as a regulator of thrombosis. Here, we show that cleavage of the kringle domain at Lys317/Thr318 switches its function to a regulator of angiogenesis. In vitro, the cleaved protein specifically inhibited the proliferation and migration of endothelial cells. The protein was without effect on preformed endothelial cell tubes. In vivo, the cleaved protein inhibited neovascularization into subcutaneously implanted Matrigel and Gelfoam sponge implants and the growth of orthotopically injected tumors. Collectively, these data indicate that plasmin-cleaved ß-2-glycoprotein 1 is a potent antiangiogenic and antitumor molecule of potential therapeutic significance.
--------------------------------------------------------------------------------
There is increasing evidence of a strong interrelationship between thrombosis, fibrinolysis, and angiogenesis that is controlled by synchronized cross talk between zymogens and their cleavage products.1 For example, during coagulation, the proteolytic activation of prothrombin produces thrombin and prothrombin fragments I and II, which triggers clotting and regulates endothelial cell (EC) growth,2 respectively. In fibrinolysis, fibrin-catalyzed cleavage of plasminogen produces clot-digesting plasmin and the antiangiogenic molecule angiostatin.3,4
ß2-Glycoprotien 1 (ß2GP1), also known as apolipoprotein H, is a single-chain plasma glycoprotein composed of 326 amino acid residues that forms four complement control protein modules (domains I through IV) and a distinct C-terminal kringle domain (domain V).5-8 Kringle domain V carries a lysine-rich sequence motif (C281KNKEKKC288) that binds negatively charged lipids9,10 and a hydrophobic loop (313LAFW316) that embeds the protein into anionic lipid-containing target membranes.6,11-13 Because of these properties, the protein inhibits ADP-induced platelet aggregation14-16 and competes for the assembly of coagulation cascade proteins on procoagulant cell surfaces.17-23 Other studies have shown that it binds EC24 and protects cells against nitric oxide-mediated apoptosis25 and atherosclerosis.26,27 Interestingly, varying levels of a proteolytically cleaved form (Lys317/Thr318 cleavage site) of the protein (nicked ß2GP1) have been found in the plasma of leukemia patients28 and patients treated with streptokinase.29 Because cleavage at Lys317/Thr318 abrogates the protein??s ability to bind anionic surfaces,9,10 a decrease in the ratios of intact to nicked forms of ß2GP1 might influence the thrombotic events commonly seen in these patients. Many enzymes involved in coagulation and fibrinolysis (factor Xa, factor XI, plasmin, and elastase) cleave ß2GP1 at Lys317/Thr318, suggesting that activation of fibrinolysis contributes to an increasingly diminished role of ß2GP1 in thrombosis.29,30 On the other hand, plasmin cleavage of the intact protein (iß2GP1) to the nicked form (nß2GP1) results in a gain of function that also regulates thrombus formation by accelerating thrombin-dependent factor XI activation23,31 and fibrinolysis by inhibiting plasminogen/tissue plasminogen activator (t-PA)-mediated activation of plasminogen.32 These findings, together with observations on the relationship between kringle structures and antiangiogenic activity,33 raise the possibility that iß2GP1-to-nß2GP1 transitions result in a kringle domain alteration that dramatically switches its function from regulating thrombosis to regulating fibrinolysis and angiogenesis. Indeed, recent studies raised the possibility that nß2GP1 functions as an antiangiogenic molecule in vivo.34
It this article, we demonstrate that nß2GP1 inhibits EC proliferation in vitro, inhibits neovascularization into subcutaneously implanted Matrigel and Gelfoam plugs, and blocks tumor growth in a mouse model system. Taken together, these data provide evidence in support of the concept that nß2GP1 plays a regulatory role in EC physiology and angiogenesis.

【关键词】  plasmin-cleaved --glycoprotein inhibitor angiogenesis

Materials and Methods

Animals, Cells, and Reagents

Male C57Bl/6 and BALB/c mice were purchased from the National Cancer Institute-Frederick Cancer Research Facility (Frederick, MD). Tramp C2RE3 prostate adenocarcinoma cells (TRAMP) were provided by J. Killion, M. D. Anderson Cancer Center. These cells were derived from TRAMP C3 cells35 by selection for aggressively growing tumors after repeated orthotopic injections. The cells were maintained in vitro in minimal essential media containing 10% fetal bovine serum (Invitrogen, Carlsbad, CA). Bovine aortic endothelial cells (BAECs) were cultured in bovine endothelial growth medium (Cell Applications, Inc., San Diego, CA). Human umbilical vein endothelial cells (HUVECs) were cultured in Medium 200 containing low serum growth supplement (Cascade Biologicals, Portland, OR). Plasmin and its chromogenic substrate, S-2251, were purchased from Chromogenix (Lexington, MA). Human serum albumin (HSA) was from Alpha Therapeutics (Los Angeles, CA), and annexin 2 antibodies (clone 5) and Matrigel were from BD Biosciences (Bedford, MA). Mouse CD31 antibodies (clone CO.3R1D4) were from Serotec (Raleigh, NC). Other chemicals and chromatographic media were from Sigma-Aldrich (St. Louis, MO). Rabbit anti-human angiostatin (Oncogene Research Products, San Diego, CA) exhibited significant cross-reactivity with plasminogen and plasmin. Proteins used in this study were routinely tested to ensure the absence of lipopolysaccharide contamination using the Pyrochrome LAL reagent (Associates of Cape Cod Inc., East Falmouth, MA) assay. Immobilized plasmin was prepared by incubating 1 mg of plasmin in ice-cold PBS with 1 ml of Affi-Gel 10 (Bio-Rad Laboratories, Hercules, CA). The coupling was allowed to proceed at 4??C for 2 hours, after which uncoupled reagents were removed by repeated washings with PBS. Polyclonal antibodies to iß2GP1 were produced in rabbits by multiple intradermal injections of 0.5 mg of ß2GP1 in complete Freund??s adjuvant in multiple intradermal sites, followed by two boosters (0.25 mg of protein) at 2-week intervals in incomplete Freund??s adjuvant. The rabbits were bled 2 weeks after the last injection. IgG was purified from the immune serum by protein G affinity chromatography.

51Cr-Labeled Mouse Red Blood Cells

Syngeneic mouse red blood cells were labeled with 51Cr by incubation at 37??C for 4 hours with 0.25 mCi of Na51chromate (GE Healthcare, Little Chalfont, Buckinghamshire, UK) in Hepes-buffered saline (pH 7.4) containing 30 mmol/L glucose. Unbound 51Cr was removed by repeated washings with the same buffer. The cells were resuspended to a 25% hematocrit in the same buffer before injection.

Purification of iß2GP1

Intact ß2GP1 was purified from pooled human plasma as described previously.36,37 In brief, whole blood collected from healthy volunteers (Gulf Coast Regional Blood Center, Houston, TX) was centrifuged at 2500 x g for 10 minutes to sediment the blood cells. The supernatant (plasma) was then chilled on ice, and perchloric acid was added dropwise with continuous stirring. The plasma was incubated on ice for 15 minutes, followed by centrifugation at 20,000 x g for 15 minutes to sediment the precipitated proteins. The supernatant containing ß2GP1 was brought to pH 7.0 with saturated sodium bicarbonate and dialyzed against Tris buffer (50 mmol/L Tris, pH 8.0) containing 20 mmol/L NaCl. The dialysate was passed over a DEAE-Sephacel column equilibrated with the same buffer. The flow-through was collected and passed over a Hi-Trap Heparin-Sepharose affinity column. The column was washed with Tris buffer containing 20 mmol/L NaCl, and the bound ß2GP1 was eluted with the same buffer containing 250 mmol/L NaCl. Purity was assessed by gel electrophoresis and Western blotting with rabbit anti-human iß2GP1. The identity of the protein was confirmed by N-terminal sequencing.

Preparation of nß2GP1

Intact ß2GP1 was incubated with immobilized plasmin at 37??C for 17 hours. The beads were removed by centrifugation and the supernatant recovered. Cleavage was verified by an electrophoretic shift under reducing conditions and by N-terminal sequencing, which revealed a second N terminus corresponding to the Lys317/Thr318 cleavage site. Western blotting of the purified product indicated that the nß2GP1 preparations were plasmin-free and did not contain autoproteolytic products (no reactivity with plasmin or angiostatin antibodies).

Immunofluorescence Analysis of ß2GP1 Binding to EC

BAECs and TRAMP cells were incubated with iß2GP1 or nß2GP1 (4 µmol/L) on ice for 30 minutes. The cells were then washed with PBS and incubated for an additional 30 minutes on ice with 2 µg of biotinylated rabbit anti-human ß2GP1 IgG, followed by incubation with 50 ng of fluorescein isothiocyanate (FITC)-streptavidin. Binding was determined by flow cytometric analysis using cells incubated only with the primary antibody and FITC-streptavidin as negative controls. For the competition experiments with annexin 2 antibody, BAECs were cultured on glass coverslips for 24 hours and incubated on ice for 1 hour with iß2GP1 or nß2GP1 (4 µmol/L) in the absence or presence of annexin II or CD31 (negative control) antibodies (0.33 µmol/L). The cells were then washed, fixed with 2% paraformaldehyde, and stained with biotinylated rabbit anti-human ß2GP1 IgG (2 µg), followed by phycoerythrin-conjugated streptavidin (100 ng).

Assay for Plasmin Activity

BAECs or HUVECs were cultured to 80% confluence. One milliliter of conditioned or fresh (negative control) medium was transferred to cuvettes, and the change in absorbance at 405 nm was recorded following the addition of the chromogenic plasmin substrate S-2251 (0.3 mmol/L).

Cell Proliferation Assay

Thymidine Incorporation

BAECs and TRAMP C2RE3 cells were cultured in complete medium containing 0.5 µCi of thymidine and 4 µmol/L HSA (control), iß2GP1, or nß2GP1. After 72 hours, the cells were washed three times with PBS, twice with 5% trichloroacetic acid, and solubilized in 0.2% SDS. The cell lysate was resuspended in 5 ml of scintillation cocktail for liquid scintillation counting.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Bromide Assay

HUVECs were cultured in 96-well plates in medium containing 4 µmol/L HSA (control), iß2GP1, or nß2GP1 (4 µmol/L). After 72 hours, 25 µl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (2.5 mg/ml) was added to each well and incubated for 2 hours at 37??C. The medium was then removed and the formazan crystals solubilized in 50 µl of dimethyl sulfoxide before spectrophotometric quantification (A = 560 nm). Cell proliferation was expressed as the percentage of controls.

Migration Assays

Boyden Chamber Assay

Cells were plated at 70% confluency on 6.5-mm Transwell polycarbonate membranes (8-µm pore size; Corning, Acton, MA). Vascular endothelial growth factor (VEGF) (25 ng/ml) was added to the lower chamber, and iß2GP1 or nß2GP1 (4 µmol/L) was added to the upper chamber. After 5 hours at 37??C, the cells on the upper surface were removed by scraping. The polycarbonate filters were then stained with Hema-Diff reagent (StatLab Medical Products, Inc. Lewisville, TX). Results are expressed as the mean ?? SD of 10 individual experiments.

Scratch Assay

Cells were cultured on 24-well tissue culture plates to confluency. Cells in the center of the wells were removed by scratching with a 1-ml pipette tip.38 The remaining adherent cells were washed twice with PBS, incubated with iß2GP1 or nß2GP1 (4 µmol/L) in BAEC medium for 8 hours, and photographed. Motility was measured by counting the number of cells that repopulated the cleared area. Results are expressed as the number of cells/mm2 ?? SD and are the mean of four individual experiments.

Tube Disruption Assay

Forty-eight-well tissue culture plates were coated with 250 µl of Matrigel (8.9 mg/ml) for 2 hours at 37??C. Cells (7.5 x 104) were plated in BAEC medium for 96 hours to allow for tube formation. The preformed tubes were then incubated with HSA, iß2GP1, or nß2GP1 (4 µmol/L) for 24 hours and assessed for tube integrity by microscopy.

Neovascularization Assays

The effect of iß2GP1 and nß2GP1 on neovascularization was determined by two independent assays.

Gelfoam Implant

Sterile Gelfoam absorbable sponges (Pharmacia & Upjohn, Peapack, NJ) were cut into 5 x 5 x 7-mm pieces and hydrated overnight with PBS. Agarose (0.4%, 100 µl) containing VEGF (2 pmol/implant) and nß2GP1 (0.2 µmol) or HSA (0.2 µmol, control) was pipetted onto each sponge. After 1 hour at room temperature, the gel foams were placed into a subcutaneous pocket as described previously.39 Vascularization into the implants was quantified after 2 weeks by assessing blood volume after i.v. injection of 51Cr-labeled syngeneic red blood cells several minutes before recovery of the implants. Blood volume was calculated from the specific activity of the blood (cpm/µl blood/g implant).

Matrigel Plug

Matrigel (1.5 ml) was mixed on ice with VEGF (0.7 pmol) in the presence or absence of iß2GP1 or nß2GP1 (6 nmol). BALB/c mice (three per group) were injected intradermally with 0.5 ml of the Matrigel. Two weeks later, 51Cr-labeled syngeneic red blood cells were injected i.v. several minutes before recovery of the implants. Blood volume was calculated as described for the gel foams.

Murine Prostate Cancer Model

TRAMP C2RE3 cells (2 x 104) were implanted orthotopically into the prostate of 6-week-old, C57Bl/6 mice. The mice were randomly assigned to different groups. nß2GP1 or iß2GP1 (1.7 mg/0.2 ml pump) was administered to the mice with Alzet 2002 mini-osmotic pumps (delivery rate of 3.6 mg/kg/day; Durect Corp., Cupertino, CA) that were implanted i.p. and s.c. on days 1 and 14, respectively (spent pumps were not removed). Mice in the chemotherapy and combination therapy groups were also administered docetaxel intraperitoneally at 8 mg/kg once a week for 4 weeks beginning on day 3. Animals were sacrificed on day 28, and the tumors were harvested, weighed, and quick-frozen for immunohistochemistry. Frozen sections were stained for CD31-positive EC with rat anti-CD31 antibody (BD Biosciences) followed by Texas Red-conjugated goat anti-rat IgG (Jackson ImmunoResearch). Apoptotic cells were detected by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining performed according to manufacturer??s (Promega Corporation, Madison, WI) instructions.

Analysis of ß2GP1 Degradation Products

BAECs were grown for 72 hours in the absence or presence of intact or nß2GP1 (4 µmol/L). The supernatants were centrifuged to remove cell debris and incubated with rabbit anti-human ß2GP1. The antibody and bound antigens were concentrated by pull-down with protein G-Sepharose beads. The beads were washed and solubilized with SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer and resolved by gel electrophoresis. For Western blotting, the proteins were transferred to polyvinylidene difluoride membranes and probed with the same antibodies followed by peroxidase-conjugated anti-rabbit IgG.

Immunoprecipitation

One microgram of Glu-plasminogen, plasmin, or angiostatin was incubated with 10 µg of iß2GP1 or nß2GP1 for 1 hour at 20??C, followed by incubation for 1 hour with control rabbit IgG or rabbit anti-ß2GP1 IgG (20 µg) and protein G-Sepharose (20 µl). Captured IgG (and bound proteins) were centrifuged, and the supernatant (unbound protein) was mixed with SDS-PAGE sample buffer for Western blot analysis. The protein G beads were washed twice with PBS, centrifuged through 30% sucrose in PBS, and resuspended in SDS-PAGE sample buffer.

Results

Purification of iß2GP1 and Preparation of nß2GP1

Intact ß2GP1 was purified from pooled human plasma by perchloric acid treatment followed by ion exchange and heparin affinity chromatography as described in Materials and Methods. SDS-PAGE and Western blot analysis showed that the purity of the protein was >98% (Figure 1) . Nicked ß2GP1 was prepared by incubating iß2GP1 for 17 hours at 37??C with immobilized plasmin. SDS-PAGE analysis of the product under reducing conditions showed that >98% of the protein was cleaved (Figure 1A) . N-Terminal sequencing revealed that the protein was cleaved at amino acids 317 and 318 (Lys-Thr). Western blotting of the final preparation with plasmin antibodies indicated that the nß2GP1 was plasmin-free. Solid-phase enzyme-linked immunosorbent assay binding assays were performed to determine the binding of the nicked isoform to anionic phosphatidylserine (PS). Figure 1B shows that the ability of iß2GP1 to bind PS was lost after plasmin treatment. Control experiments showed that neither isoform bound neutral phosphatidylcholine.

Figure 1. Analysis of purified iß2GP1 and nß2GP1. A: SDS-PAGE analysis of iß2GP1 and nß2GP1; lane 1, Coomassie Brilliant Blue staining of iß2GP1; lane 2, Western blot analysis of iß2GP1 with polyclonal anti-ß2GP1 IgG; lanes 3 to 6, Western blot analysis of nß2GP1 (lanes 3 and 5) and iß2GP1 (lanes 4 and 6) in reducing (lanes 3 and 4) and nonreducing (lanes 5 and 6) SDS-PAGE; lane 7, Western blotting of purified nß2gp1 with mouse monoclonal plasmin antibodies; and lane 8, plasmin positive control. Arrowheads represent molecular mass markers (250, 150, 100, 75, 50, 37, 20, and 10 kDa). B: Binding of intact and nß2GP1 to PS. Ninety-six-well enzyme-linked immunosorbent assay plates were coated with phosphatidylcholine or PS in ethanol (20 µg/ml) and incubated with serial dilutions of intact and nß2GP1. Protein binding was assessed with rabbit anti-ß2GP1 and peroxidase-conjugated anti-rabbit IgG. and PC; • and PS; circles, iß2GP1; triangles, nß2GP1. C: N-Terminal sequence analysis of iß2GP1 and nß2GP1. The amino acids in bold show the plasmin-cut site in iß2GP1.

ß2GP1 Binds EC and Inhibits Proliferation and Migration in Vitro

Previous studies have shown that iß2GP1 binds to the surface of EC through annexin 2 expressed at the cell surface.24 To determine the binding of iß2GP1 and nß2GP1, BAECs were incubated with the proteins for 30 minutes. Fluorescence-activated cell sorting analysis after incubation of the cells with biotinylated anti-human-ß2GP1 followed by FITC-conjugated streptavidin showed that both iß2GP1 and nß2GP1 bound to >75% of the EC (Figure 2) . Unlike EC, only the intact protein bound to TRAMP C2RE3 cells.

Figure 2. Binding of iß2GP1 and nß2GP1 to endothelial cells. BAECs and TRAMP cells were incubated with iß2GP1 or nß2GP1 on ice for 30 minutes followed by incubation with biotinylated anti-ß2GP1 IgG and FITC-streptavidin as described in Materials and Methods. Control cells were incubated only with primary antibody and FITC-streptavidin. Data are representative of three independent experiments.

Several studies were performed to determine the influence of iß2GP1 and nß2GP1 on EC function. First, the effect of ß2GP1 on EC migration was determined by assessing cell mobility using the Boyden chamber transwell migration assay. Figure 3, A and B , shows that nß2GP1, but not iß2GP1, significantly inhibited the migration of cells to the lower chamber. Additional evidence supporting this observation was obtained with the in vitro wound healing model system.38 Confluent EC monolayers incubated with iß2GP1 or nß2GP1 were scraped, and migration of cells into the denuded sections was monitored at various time intervals. Figure 3, C and D , shows that, whereas control and iß2GP1-treated ECs repopulated the denuded area within 8 hours, incubation in the presence of nß2GP1 resulted in 60% inhibition in cell migration. Control experiments showed that nß2GP1 did not inhibit migration of TRAMP cells, suggesting that the inhibitory effect of nß2GP1 was EC-specific (Figure 3C) . In contrast to the apparent specific effect of nß2GP1 on BAEC migration, both iß2GP1 and nß2GP1 significantly inhibited BAEC proliferation (Figure 4A) . Growth of TRAMP C2RE3 cells (and PC3, LnCap, MCF7, MDA-MB-231, and MDA-MB-435 cells, data not shown) was unaffected when incubated with either protein (Figure 4C) . In contrast to BAECs, HUVECs were sensitive to the inhibitory effect of the nß2GP1, but not iß2GP1 (Figure 4B) .

Figure 3. nß2GP1 inhibits EC migration. A: BAECs were incubated with iß2GP1 or nß2GP1 (4 µmol/L) for 5 hours, and cells that migrated to the VEGF-containing lower side of the polycarbonate membrane were counted. Results are expressed as the mean ?? SD of 10 experiments. B: Representative photomicrographs of migrated BAECs (top panels) and TRAMP C2RE3 (bottom panels) cells. C: Confluent BAECs and TRAMP C2RE3 monolayers were denuded by scraping and incubated for 5 hours with 4 µmol/L nß2GP1 (shaded bars) or iß2GP1 (hatched bars). Results are expressed as the mean ?? SD of four experiments. D: Typical photomicrograph of denuded BAEC monolayers at 0 hours (top panels) and 8 hours (bottom panels).

Figure 4. ß2GP1 inhibits EC proliferation. BAECs (A), HUVECs (B), or TRAMP C2RE3 (C) cells were cultured for 72 hours in complete medium in the presence of HSA (control), iß2GP1, or nß2GP1. Cell proliferation was assessed by thymidine incorporation (A and C) or 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (B) as described in Materials and Methods. Results are expressed as the mean ?? SD of four experiments. D: Western blot analysis of iß2GP1 and nß2GP1 (4 µmol/L) incubated with BAECs for 72 hours. The medium was centrifuged to remove cell debris, and ß2GP1 in the supernatant was captured with anti-ß2GP1/protein G beads. Lane 1: nß2GP1, lane 2: iß2GP1, lane 3: nß2GP1 incubated with BAECs, and lane 4: iß2GP1 incubated with BAECs. Arrowheads represent molecular mass markers (100, 75, 50, 37, 25, 15, and 10 kDa). E: Plasmin activity (A405 nm) in conditioned media from BAEC () and HUVEC () measured with S-2251 (300 µmol/L).

Considering the specific inhibition of nß2GP1 on EC migration, the inhibitory effect of the intact protein on the proliferation of BAECs (but not HUVECs) was unexpected. Because the proliferation experiments were performed over prolonged incubation times, we tested the BAEC supernatants for potential ß2GP1 degradation products by Western blotting using ß2GP1 antibodies. Figure 4D shows that both the nicked and intact proteins were degraded into several distinct products. The nicked protein produced two major degradation products at 37 and 15 kDa, with a minor product at 30 kDa. With the exception of a band consistent with the nicked isoform at 40 kDa, the intact protein produced degradation products identical to those formed by the nicked protein. It seems likely, therefore, that the ability of iß2GP1 to inhibit BAEC proliferation was due to its in situ conversion to the nicked isoform (together with other degradation products) as a result of BAEC-derived proteolytic activity. Indeed, analysis of culture supernatants revealed significant plasmin activity in BAEC- but not HUVEC-conditioned medium (Figure 4E) . Taken together, these data suggest that the inability of iß2GP1 to inhibit the proliferation of HUVECs was because the protein was not cleaved to the active nicked isoform. Thus, the inability of the intact protein to inhibit BAEC migration (Figure 3A) was probably due to the very limited proteolysis of the protein within the time frame of these experiments (5 hours) as opposed to the extended incubation times in the proliferation assay (72 hours). Taken together, these data suggest that nß2GP1 specifically functions as an inhibitor of EC growth and migration. It is important to note that neither protein was able to disrupt preformed EC tubes in vitro (Figure 5) .

Figure 5. ß2GP1 does not disrupt pre-existing EC tubes. EC tubes formed on Matrigel-coated plates were incubated with iß2GP1 or nß2GP1 (4 µmol/L) for 24 hours. Note that the tube structure remains largely intact.

Inhibition of Neovascularization by nß2GP1

Because nß2GP1 inhibits EC migration and proliferation in vitro, we tested its potential to function as an inhibitor of angiogenesis in vivo. Intact and nß2GP1 was incorporated together with VEGF into Matrigel (Figure 6A) or Gelfoam plugs (Figure 6B) and implanted subcutaneously into BALB/c mice. Two weeks later, the mice were injected with 51Cr-labeled syngeneic red blood cells to quantify blood volume within the implants. Figure 6 shows that control implants seemed to be highly vascularized, whereas both the nß2GP1- and iß2GP1-containing implants were relatively clear. Indeed, assessment of blood volume within the implants showed that vascularization was reduced 10-fold and threefold in the Gelfoam and Matrigel implants, respectively. The absence of neovascularization in the iß2GP1-containing implants suggests that the protein was cleaved to the nicked isoform in situ.

Figure 6. nß2GP1 inhibits neovascularization into subcutaneous implants. Matrigel (A) or Gelfoam sponge (B) containing VEGF and HSA or nß2GP1 were placed s.c. in BALB/c mice (10 mice/group). The implants were recovered 2 weeks later. Blood volume (µl/g implant) was calculated from the specific activity of 51Cr-labeled red blood cells in the implants. The data shown are the mean ?? SD of three experiments.

Inhibition of Orthotopically Implanted TRAMP Prostate Cancer Tumor Growth by nß2GP1

To test the ability of ß2GP1 to inhibit tumor growth, TRAMP C2RE3 prostate carcinoma cells were injected orthotopically. Groups of animals were treated by i.p. implantation of 14-day Alzet pumps containing HSA (control group), iß2GP1, or nß2GP1. To ensure continued release of the test proteins over the time course of the study, an additional pump was implanted s.c. at day 14. The experiment was terminated at day 28, and tumor growth was assessed. The data presented in Figure 7 show a 56% reduction in tumor volumes in mice treated with nß2GP1. The extent of inhibition was comparable with that obtained with docetaxel alone. CD31 and TUNEL staining of thin sections from tumors recovered from the nß2GP1-treated animals revealed significant tumor cell apoptosis (Figure 7 , photomicrograph). Although treatment with docetaxel resulted in inhibition in tumor growth, multiple thin sections through the residual tumor failed to reveal large numbers of apoptotic cells. The reason for this is unclear but could be related to the route of drug administration (i.p.), which would affect its bioavailability. Similar to recent data obtained with doxorubicin,40 our results suggest that inhibition of tumor growth was due to cytostasis, not apoptosis. Irrespective of the treatment group, multiple serial sections through the tumors did not reveal apoptotic ECs. Taken together with the inhibition of angiogenesis in the in vivo Matrigel and Gelfoam assays, these data raise the possibility that the reduction in tumor growth is a result of inhibition of vascular expansion.

Figure 7. nß2GP1 reduces tumor burden in an orthotopic murine prostate cancer model. A, B, C: Mice (10 mice/group/experiment) injected orthotopically with TRAMP C2RE3 (2 x 104 cells) were treated as indicated. iß2GP1 or nß2GP1 was administered through Alzet mini osmotic pumps implanted on days 1 and 14. Docetaxel was administered once a week i.p. Results are mean ?? SD of three experiments. Photomicrographs: Representative thin sections were stained for CD31 (red), TUNEL (green), and total cells with 4,6-diamidino-2-phenylindole (DAPI) (blue).

Binding of iß2GP1 and nß2GP1 to Annexin 2, Plasminogen, and Its Cleavage Products

It has been previously shown that annexin 2 on EC binds t-PA and plasminogen,41-43 a process that is critical to the localized production of plasmin for dissolution of fibrin clots and wound healing. Interestingly, plasmin also catalyzes proteolysis of the extracellular matrix, a process that is crucial to EC growth and expansion. Because annexin 2 also binds ß2GP1,24,44 it is possible that iß2GP1 and nß2GP1 regulate the binding of plasminogen to annexin 2 and/or the activity of plasmin. To determine whether ß2GP1 binds to plasminogen or its cleavage products, iß2GP1 and nß2GP1 were incubated with plasminogen, plasmin, and angiostatin. Complexes were captured with anti-ß2GP1 and protein G-Sepharose. Western blot analysis of the supernatants and pellets showed that nß2GP1 bound to plasminogen exclusively, whereas the intact isoform did not bind plasminogen or any its cleavage products (Figure 8A) .

Figure 8. Binding of iß2GP1 and nß2GP1 to annexin 2 and plasminogen and its cleavage products. A: iß2GP1 and nß2GP1 were incubated with the target proteins, plasminogen, plasmin, or angiostatin. The mixture was then incubated with anti-ß2GP1 and Protein G-Sepharose to capture ß2GP1 and any bound target proteins. Proteins in the supernatant and protein G bead pellets were resolved by SDS-PAGE and analyzed by Western blotting using plasminogen, plasmin, or angiostatin antibodies. Odd-numbered lanes represent incubation with anti-ß2GP1, and even-numbered lanes represent incubation with control IgG. Lanes 1 and 2: Plasminogen, lanes 3 and 4: plasmin, and lanes 5 and 6: angiostatin. Data are representative of three independent experiments. B: BAECs were plated on glass coverslips overnight and incubated on ice for 1 hour with iß2GP1 or nß2GP1 (4 µmol/L) in the absence or presence of anti-annexin II or anti CD31 (0.33 µmol/L). The cells were then washed, fixed with 2% paraformaldehyde, and stained with biotinylated rabbit anti-human ß2GP1 IgG/streptavidin-PE.

Previous studies have shown that iß2GP1 binds to HUVECs through annexin 2. To determine whether annexin 2 also serves as a "receptor" for nß2GP1, we tested the ability of annexin 2 antibodies to block the binding of iß2GP1 and nß2GP1 to BAECs. Figure 8B shows that annexin 2 antibodies significantly reduced the ability of both proteins to bind BAECs. Control experiments using anti-CD31 (Figure 8B) or nonspecific mouse IgG (not shown) did not inhibit binding.

Discussion

There is increasing evidence that proteins that participate in coagulation and fibrinolysis also regulate angiogenesis.1 Many of these "dual function" proteins are conformationally altered cleavage products of parent molecules that regulate thrombosis. For example, during coagulation, the zymogen prothrombin is cleaved to produce thrombin and prothrombin fragments (I and II), which promote clotting and arrest EC growth, respectively.2 Likewise, high-molecular weight kininogen plays a role in contact activation of coagulation,45 whereas its kallikrein cleavage product promotes EC apoptosis.46 Similar properties have been found for apolipoprotein (a)38,47,48 and histidine proline-rich glycoprotein.49-51 During fibrinolysis, the zymogen plasminogen is activated to produce plasmin, which initiates dissolution of fibrin and revascularization. Concurrent with EC proliferation (vascularization), autocatalytic inactivation of plasmin produces angiostatin that inhibits further angiogenesis.4 Interestingly, all of these antiangiogenic fragments are characterized by one or more disulfide-linked kringle domain structures.33

ß2GP1 (apolipoprotein H) is a kringle domain-containing plasma glycoprotein that binds to EC through annexin 224 and functions as an EC survival factor.25 Similar to other multifunctional coagulation proteins, cleavage of ß2GP1 induces a conformational change52 that alters the regulatory properties of the protein.32 Because of the structural similarities in the kringle domain of ß2GP1 with other antiangiogenic fragments, we investigated the potential function of this protein as a regulator of angiogenesis. Our results show that nß2GP1 is a potent inhibitor of EC proliferation and migration in vitro. This property was dependent on the nß2GP1 isoform because prolonged incubation of iß2GP1 with BAECs gave similar results due to the in situ generation of nß2GP1 by the action of cell-derived proteases secreted into the medium.

These data raised the possibility that nß2GP1 might function as an endogenous regulator of angiogenesis. To test this, two independent in vivo experiments were performed: i) VEGF-dependent neovascularization into subcutaneously implanted Matrigel and Gelfoam plugs, and ii) the growth of orthotopically implanted prostate cancer cells. The data presented in Figure 6 show that the inclusion of both the intact and nicked isoforms of ß2GP1 into the Matrigel and Gelfoam plugs significantly inhibited vascularization. In contrast, only the nicked isoform inhibited tumor growth. Based on our in vitro data, the ability of the Matrigel/Gelfoam-embedded iß2GP1 to inhibit neovascularization was probably due to localized production of plasmin at the implant/wound site, which led to the in situ generation of the active nicked isoform. For the tumor inhibition studies, iß2GP1 was implanted into 14-day Alzet osmotic pumps where the diffusion port was placed distal to the incision site. Pump placement, combined with the fact that solutes (including plasmin) cannot diffuse into the pumps, probably precluded plasmin-dependent cleavage of iß2GP1 to nß2GP1. Immunohistochemical studies of tumor thin sections revealed numerous TUNEL-positive tumor cells in the nß2GP1 treatment groups (Figure 7) . Interestingly, only a small fraction of the CD31-positive EC appeared also to be TUNEL-positive. This suggests that the observed antitumor effect of nß2GP1 is because of inhibition of vascular expansion that is critical to tumor survival and not ablation of pre-existing vasculature.

Because the antiangiogenic activity of ß2GP1 is dependent on site-specific cleavage within domain V, its activity probably requires a conformationally altered kringle domain that binds to an EC-specific cell surface moiety/receptor. The data shown in Figure 8 provide evidence that, similar to angiostatin,53 both iß2GP1 and nß2GP1 bind to EC through annexin 2.24,34,44 Because only the nicked isoform inhibited angiogenesis (Figure 6) , it is possible that nß2GP1 competes with iß2GP1 for binding to annexin 2 on the EC surface. We also show that nß2GP1, but not iß2GP1, binds plasminogen. Because annexin 2 regulates plasmin production by forming an annexin 2/t-Pa/plasminogen complex,41-43 the binding of nß2GP1 to annexin 2 on EC could initiate a negative feedback loop that blocks assembly of the annexin 2/t-Pa/plasminogen complex, thereby precluding plasmin production and concomitant neovascularization.32 In principle, such a mechanism provides a self-limiting autoregulatory negative feedback loop in situ. This model is in agreement with recent data indicating an important role for annexin 2 in neoangiogenesis.54 In addition to its possible role in inhibiting the generation of plasmin at the EC expansion front, nß2GP1 might also exert direct inhibitory effects on EC growth through (unidentified) downstream events that alter cell cycle regulatory pathways.34

Collectively, the data reported here support the concept that nß2GP1 functions as an antiangiogenic molecule that likely suppresses tumor expansion through a specific EC-dependent pathway. The apparent resistance of pre-existing vasculature to the protein makes it a potential candidate for antiangiogenic therapy that warrants further investigation.

Acknowledgements

We thank Asa J. Ramoth for expert technical assistance.

【参考文献】
  Browder T, Folkman J, Pirie-Shepherd S: The hemostatic system as a regulator of angiogenesis. J Biol Chem 2000, 275:1521-1524

Rhim TY, Park CS, Kim E, Kim SS: Human prothrombin fragment 1 and 2 inhibit bFGF-induced BCE cell growth. Biochem Biophys Res Commun 1998, 252:513-516

Gately S, Twardowski P, Stack MS, Cundiff DL, Grella D, Castellino FJ, Enghild J, Kwaan HC, Lee F, Kramer RA, Volpert O, Bouck N, Soff GA: The mechanism of cancer-mediated conversion of plasminogen to the angiogenesis inhibitor angiostatin. Proc Natl Acad Sci USA 1997, 94:10868-10872

Soff GA: Angiostatin and angiostatin-related proteins. Cancer Metastasis Rev 2000, 19:97-107

Steinkasserer A, Estaller C, Weiss EH, Sim RB, Day AJ: Complete nucleotide and deduced amino acid sequence of human ß2-glycoprotein I. Biochem J 1991, 277:387-391

Schwarzenbacher R, Zeth K, Diederichs K, Gries A, Kostner GM, Laggner P, Prassl R: Crystal structure of human beta 2-glycoprotein I: implications for phospholipid binding and the antiphospholipid syndrome. EMBO J 1999, 18:6228-6239

Bouma B, de Groot PG, van den Elsen JH, Ravelli RG, Schouten A, Simmelink MA, Derksen RM, Kroon J, Gros P: Adhesion mechanism of human ß2-glycoprotein I to phospholipids based on its crystal structure. EMBO J 1999, 18:5166-5174

Kato H, Enjyoji K: Amino acid sequence and location of the disulfide bonds in bovine beta 2 glycoprotein I: the presence of five Sushi domains. Biochemistry 1991, 30:11687-11694

Hunt J, Krilis S: The fifth domain of ß2-glycoprotein I contains a phospholipid binding site (Cys281-Cys288) and a region recognized by anticardiolipin antibodies. J Immunol 1994, 152:653-659

Hunt JE, Simpson RJ, Krilis SA: Identification of a region of ß2-glycoprotein I critical for lipid binding and anti-cardiolipin antibody cofactor activity. Proc Natl Acad Sci USA 1993, 90:2141-2145

Wang SX, Cai GP, Sui SF: Intrinsic fluorescence study of the interaction of human apolipoprotein H with phospholipid vesicles. Biochemistry 1999, 38:9477-9484

Wang SX, Cai GP, Sui SF: The insertion of human apolipoprotein H into phospholipid membranes: a monolayer study. Biochem J 1998, 335:225-232

Hammel M, Schwarzenbacher R, Gries A, Kostner GM, Laggner P, Prassl R: Mechanism of the interaction of beta(2)-glycoprotein I with negatively charged phospholipid membranes. Biochemistry 2001, 40:14173-14181

Schousboe I: Binding of ß2-glycoprotein I to platelets: effect of adenylate cyclase activity. Thromb Res 1980, 19:225-237

Nimpf J, Wurm H, Kostner GM: ß2-Glycoprotein-I (apo-H) inhibits the release reaction of human platelets during ADP-induced aggregation. Atherosclerosis 1987, 63:109-114

Nimpf J, Wurm H, Kostner GM: Interaction of ß2-glycoprotein-I with human blood platelets: influence upon the ADP-induced aggregation. Thromb Haemost 1985, 54:397-401

Schousboe I, Rasmussen MS: Synchronized inhibition of the phospholipid mediated autoactivation of factor XII in plasma by beta 2-glycoprotein I and anti-ß2-glycoprotein I. Thromb Haemost 1995, 73:798-804

Schousboe I: ß2-Glycoprotein I: a plasma inhibitor of the contact activation of the intrinsic blood coagulation pathway. Blood 1985, 66:1086-1091

Shi W, Chong BH, Hogg PJ, Chesterman CN: Anticardiolipin antibodies block the inhibition by beta 2-glycoprotein I of the factor Xa generating activity of platelets.

Brighton TA, Hogg PJ, Dai YP, Murray BH, Chong BH, Chesterman CN: ß2-Glycoprotein I in thrombosis: evidence for a role as a natural anticoagulant. Br J Haematol 1996, 93:185-194

Nimpf J, Bevers EM, Bomans PH, Till U, Wurm H, Kostner GM, Zwaal RF: Prothrombinase activity of human platelets is inhibited by ß2-glycoprotein-I. Biochim Biophys Acta 1986, 884:142-149

Goldsmith GH, Pierangeli SS, Branch DW, Gharavi AE, Harris EN: Inhibition of prothrombin activation by antiphospholipid antibodies and beta 2-glycoprotein 1. Br J Haematol 1994, 87:548-554

Shi T, Iverson GM, Qi JC, Cockerill KA, Linnik MD, Konecny P, Krilis SA: beta(2)-Glycoprotein I binds factor XI and inhibits its activation by thrombin and factor XIIa: loss of inhibition by clipped beta(2)-glycoprotein I. Proc Natl Acad Sci USA 2004, 101:3939-3944

Ma K, Simantov R, Zhang JC, Silverstein R, Hajjar KA, McCrae KR: High affinity binding of beta 2-glycoprotein I to human endothelial cells is mediated by annexin II. J Biol Chem 2000, 275:15541-15548

Lin KY, Wang HH, Lai ST, Pan JP, Chiang AN: ß2-Glycoprotein 1 protects J774A. 1 macrophages and human coronary artery smooth muscle cells against apoptosis. J Cell Biochem 2005, 94:485-496

Gomes LF, Alves AF, Sevanian A, Peres CA, Cendoroglo MS, Mello-Almada C, Quirino LM, Ramos LR, Junqueira VB: Role of ß2-glycoprotein I. LDL-, and antioxidant levels in hypercholesterolemic elderly subjects Antioxid Redox Signal 2004, 6:237-244

Lin KY, Pan JP, Yang DL, Huang KT, Chang MS, Ding PY, Chiang AN: Evidence for inhibition of low density lipoprotein oxidation and cholesterol accumulation by apolipoprotein H (ß2-glycoprotein I). Life Sci 2001, 69:707-719

Itoh Y, Inuzuka K, Kohno I, Wada H, Shiku H, Ohkura N, Kato H: Highly increased plasma concentrations of the nicked form of beta(2) glycoprotein I in patients with leukemia and with lupus anticoagulant: measurement with a monoclonal antibody specific for a nicked form of domain V. J Biochem 2000, 128:1017-1024

Horbach DA, van Oort E, Lisman T, Meijers JC, Derksen RH, de Groot PG: ß2-Glycoprotein I is proteolytically cleaved in vivo upon activation of fibrinolysis. Thromb Haemost 1999, 81:87-95

Ohkura N, Hagihara Y, Yoshimura T, Goto Y, Kato H: Plasmin can reduce the function of human beta2 glycoprotein I by cleaving domain V into a nicked form. Blood 1998, 91:4173-4179

Shi T, Giannakopoulos B, Iverson GM, Cockerill KA, Linnik MD, Krilis SA: Domain V of ß2-glycoprotein I binds factor XI/XIa and is cleaved at Lys317-Thr318. J Biol Chem 2005, 280:907-912

Yasuda S, Atsumi T, Ieko M, Matsuura E, Kobayashi K, Inagaki J, Kato H, Tanaka H, Yamakado M, Akino M, Saitou H, Amasaki Y, Jodo S, Amengual O, Koike T: Nicked beta(2)-glycoprotein I: a marker of cerebral infarct and a novel role in the negative feedback pathway of extrinsic fibrinolysis. Blood 2004, 103:3766-3772

Cao Y, Cao R, Veitonmaki N: Kringle structures and antiangiogenesis. Curr Med Chem Anticancer Agents 2002, 2:667-681

Beecken WDC, Engl T, Ringel EM, Camphausen K, Michaelis M, Jonas D, Folkman J, Shing Y, Blaheta RA: An endogenous inhibitor of angiogenesis derived from a transitional cell carcinoma: clipped ß2-glycoprotein-I. Ann Surg Oncol 2006, 13:1241-1251

Foster BA, Gingrich JR, Kwon ED, Madias C, Greenberg NM: Characterization of prostatic epithelial cell lines derived from transgenic adenocarcinoma of the mouse prostate (TRAMP) model. Cancer Res 1997, 57:3325-3330

Polz E, Kostner GM: The binding of beta 2-glycoprotein-I to human serum lipoproteins: distribution among density fractions. FEBS Lett 1979, 102:183-186

Wurm H: ß2-Glycoprotein-I (apolipoprotein H) interactions with phospholipid vesicles. Int J Biochem 1984, 16:511-515

Kim JS, Yu HK, Ahn JH, Lee HJ, Hong SW, Jung KH, Chang SI, Hong YK, Joe YA, Byun SM, Lee SK, Chung SI, Yoon Y: Human apolipoprotein(a) kringle V inhibits angiogenesis in vitro and in vivo by interfering with the activation of focal adhesion kinases. Biochem Biophys Res Commun 2004, 313:534-540

McCarty MF, Baker CH, Bucana CD, Fidler IJ: Quantitative and qualitative in vivo angiogenesis assay. Int J Oncol 2002, 21:5-10

Demidenko ZN, Vivo C, Halicka HD, Li CJ, Bhalla K, Broude EV, Blagosklonny MV: Pharmacological induction of Hsp70 protects apoptosis-prone cells from doxorubicin: comparison with caspase-inhibitor- and cycle-arrest-mediated cytoprotection. Cell Death Differ 2006, 13:1434-1441

Hajjar KA, Jacovina AT, Chacko J: An endothelial cell receptor for plasminogen/tissue plasminogen activator. I. Identity with annexin II. J Biol Chem 1994, 269:21191-21197

Cesarman GM, Guevara CA, Hajjar KA: An endothelial cell receptor for plasminogen/tissue plasminogen activator (t-PA). II. Annexin II-mediated enhancement of t-PA-dependent plasminogen activation. J Biol Chem 1994, 269:21198-21203

Kang HM, Choi KS, Kassam G, Fitzpatrick SL, Kwon M, Waisman DM: Role of annexin II tetramer in plasminogen activation. Trends in Cardiovascular Medicine 1999, 9:92-102

Zhang J, McCrae KR: Annexin A2 mediates endothelial cell activation by antiphospholipid/anti-beta2 glycoprotein I antibodies. Blood 2005, 105:1964-1969

Colman RW, Schmaier AH: Contact system: a vascular biology modulator with anticoagulant, profibrinolytic, antiadhesive, and proinflammatory attributes. Blood 1997, 90:3819-3843

Zhang JC, Claffey K, Sakthivel R, Darzynkiewicz Z, Shaw DE, Leal J, Wang YC, Lu FM, McCrae KR: Two-chain high molecular weight kininogen induces endothelial cell apoptosis and inhibits angiogenesis: partial activity within domain 5. FASEB J 2000, 14:2589-2600

Kim JS, Chang JH, Yu HK, Ahn JH, Yum JS, Lee SK, Jung KH, Park DH, Yoon Y, Byun SM, Chung SI: Inhibition of angiogenesis and angiogenesis-dependent tumor growth by the cryptic kringle fragments of human apolipoprotein (a). J Biol Chem 2003, 278:29000-29008

Yu HK, Kim JS, Lee HJ, Ahn JH, Lee SK, Hong SW, Yoon Y: Suppression of colorectal cancer liver metastasis and extension of survival by expression of apolipoprotein(a) kringles. Cancer Res 2004, 64:7092-7098

Doñate F, Juarez JC, Guan X, Shipulina NV, Plunkett ML, Tel-Tsur Z, Shaw DE, Morgan WT, Mazar AP: Peptides derived from the histidine-proline domain of the histidine-proline-rich glycoprotein bind to tropomyosin and have antiangiogenic and antitumor activities. Cancer Res 2004, 64:5812-5817

Juarez JC, Guan X, Shipulina NV, Plunkett ML, Parry GC, Shaw DE, Zhang JC, Rabbani SA, McCrae KR, Mazar AP, Morgan WT, Donate F: Histidine-proline-rich glycoprotein has potent antiangiogenic activity mediated through the histidine-proline-rich domain. Cancer Res 2002, 62:5344-5350

Olsson AK, Larsson H, Dixelius J, Johansson I, Lee C, Oellig C, Bjork I, Claesson-Welsh L: A fragment of histidine-rich glycoprotein is a potent inhibitor of tumor vascularization. Cancer Res 2004, 64:599-605

Matsuura E, Inagaki J, Kasahara H, Yamamoto D, Atsumi T, Kobayashi K, Kaihara K, Zhao D, Ichikawa K, Tsutsumi A, Yasuda T, Triplett DA, Koike T: Proteolytic cleavage of beta(2)-glycoprotein I: reduction of antigenicity and the structural relationship. Int Immunol 2000, 12:1183-1192

Tuszynski GP, Sharma MR, Rothman VL, Sharma MC: Angiostatin binds to tyrosine kinase substrate annexin II through the lysine-binding domain in endothelial cells. Microvasc Res 2002, 64:448-462

Ling Q, Jacovina AT, Deora A, Febbraio M, Simantov R, Silverstein RL, Hempstead B, Mark WH, Hajjar KA: Annexin II regulates fibrin homeostasis and neoangiogenesis in vivo. J Clin Invest 2004, 113:38-48

作者单位：From the Department of Cancer Biology, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas