Selectins and fibrin(ogen) play key roles in the hematogenous dissemination of tumor cells, and especially of colon carcinomas. However, the fibrin(ogen) receptor(s) on colon carcinoma cells has yet to be defined along with its relative capacity to bind fibrinogen versus fibrin under flow. Moreover, the functional P-selectin ligand has yet to be validated using intact platelets rather than purified selectin substrates. Using human CD44-knockdown and control LS174T cells, we demonstrate the pivotal involvement of CD44 in the P-selectin-mediated binding to platelets in shear flow. Quantitative comparisons of the binding kinetics of LS174T versus P-selectin glycoprotein ligand-1 (PSGL-1)-expressing THP-1 cells to activated platelets reveal that the relative avidity of P-selectin-CD44 binding is more than sevenfold lower than that of P-selectin-PSGL-1 interaction. Using CD44-knockdown LS174T cells and microspheres coated with CD44 immunoprecipitated from control LS174T cells, and purified fibrin(ogen) as substrate, we provide the first direct evidence that CD44 also acts as the major fibrin, but not fibrinogen, receptor on LS174T colon carcinoma cells. Interestingly, binding of plasma fibrin to CD44 on the colon carcinoma cell surface interferes with the P-selectin-CD44 molecular interaction and diminishes platelet-LS174T heteroaggregation in the high shear regime. Cumulatively, our data offer a novel perspective on the apparent metastatic potential associated with CD44 overexpression on colon carcinoma cells and the critical roles of P-selectin and fibrin(ogen) in metastatic spread and provide a rational basis for the design of new therapeutic strategies to impede metastasis.
- shear stress
- LS174T colon carcinoma cells
- THP-1 monocytic cells
blood-borne metastasis is a highly regulated and dynamic process, in which cancerous cells separate from a primary tumor and migrate across blood vessel walls into the bloodstream where they interact extensively with various host cells before they lodge in the target organ and form secondary metastatic colonies. Mounting evidence suggests that platelets facilitate the hematogenous dissemination of tumor cells. The most compelling evidence is the inhibition of metastasis by either pharmacological (12) or genetic depletion (9) of platelets and the restoration of metastatic potential by platelet infusion in a mouse model (23). Previous studies suggest that platelets, by adhering to tumor cells, provide a protective shield against the cytotoxic activity of natural killer cells (39, 42). Alternatively, platelets may facilitate tumor extravasation from the vasculature by potentiating tumor cell adhesion to the endothelial cell-coated vessel wall (7). Activated platelets may release a number of growth factors, which stimulate the growth of highly metastatic colon carcinoma cells (17) and contribute to tumor-induced angiogenesis (44).
Several studies have disclosed the critical involvement of P-selectin and fibrin(ogen) in the facilitation of blood-borne metastasis (5, 6, 9, 26, 33, 40–42). Microscopic observations of tumor cells trapped in the lung vasculature of wild-type mice reveal the presence of a dense coat of platelets surrounding them as well as their intimate association with fibrin(ogen) (5, 11, 18, 29). In contrast, the colon carcinoma cells in P-selectin-deficient mice had a looser and more limited platelet coat (5). The initial seeding and subsequent lodging of metastatic cells in target organs was attenuated in P-selectin-knockout mice compared with wild-type controls (5, 6, 26). Many cancer patients, including those with disseminated colon cancer (55), have detectable abnormalities of blood coagulation such as elevated levels of fibrinogen and fibrinopeptide A (30). Increased plasma levels of soluble fibrin have also been observed at the time of recurrence of colorectal cancer (19). The most convincing evidence for the direct role of fibrin(ogen) in metastatic spread is the profound inhibition of experimental and spontaneous metastasis in fibrinogen-deficient mice compared with wild-type controls (9, 40–42). It is believed that platelet/fibrin(ogen) clots surrounding tumor cells may protect them from immunological and physiological stresses in the bloodstream and facilitate their lodging to the pulmonary vasculature (39). This hypothesis is corroborated by recent in vivo data, which disclose that in mice lacking functional natural killer cells, fibrinogen deficiency was no longer a significant determinant of metastatic potential (42). In contrast with selectins (5, 6, 26, 33), fibrinogen does not play a role in the initial binding/seeding of tumor cells within the pulmonary vasculature (40). Instead, it appears to facilitate metastasis by mediating the sustained adhesion and survival of tumor cells in the high shear environment of target organs (40, 42).
Using purified P-selectin-IgG Fc as a substrate, we recently discovered that sialofucosylated CD44 variant isoforms (CD44v) represent the major functional P-selectin ligands on a variety of colon carcinoma cells (14, 38). However, the physiological significance of P-selectin-CD44v binding using free-flowing intact cells has yet to be demonstrated. Most importantly, the fibrin(ogen) receptor in colon carcinoma cells remains elusive. Although fibrin(ogen) can bind growth factors, plasma proteins, and integrins such as αIIbβ3, αvβ3, and αMβ2 as well as intercellular adhesion molecule 1, none of these adhesion molecules are expressed on the surface of a variety of colon carcinoma cells, such as LS174T, Colo205, or T84 cells, as probed by flow cytometry (22, 36). By identifying and characterizing the functional fibrin(ogen) receptor(s) on colon carcinoma cells, we will provide insights for developing novel therapeutic strategies that will selectively block colon carcinoma cell binding to fibrin(ogen) and thus interfere with metastatic spread. Although several in vivo studies have documented the pivotal role of fibrin(ogen) in blood-borne metastasis (9, 40–42), the relative capacity of colon carcinoma cells to bind fibrinogen versus fibrin under physiological flow conditions has not been previously determined.
The present study was undertaken to elucidate the dynamics and precise molecular requirements of colon carcinoma cell binding to platelets in free-cell suspensions subjected to well-defined, controlled levels of shear stress. Using CD44-knockdown LS174T colon carcinoma cells, we assessed the functional involvement of CD44 in P-selectin-mediated binding to activated platelets under flow. We also provide detailed quantitative comparisons of the binding kinetics of LS174T versus THP-1 monocytic-like cells to platelets in shear flow to compare the relative avidity of P-selectin-CD44 interaction to that of P-selectin glycoprotein ligand-1 (PSGL-1)-P-selectin binding. Furthermore, we investigate in a comprehensive and systematic manner the effects of plasma proteins in the modulation of these heterotypic adhesive interactions, and we demonstrate for the first time that CD44 functions as a fibrin, but not fibrinogen, receptor. Collectively, our data bring together disparate observations about the contributions of selectins, fibrin(ogen), and CD44 to metastasis in a unifying mechanistic interpretation and provide a rational basis for the design of novel therapeutic strategies to impede metastasis.
MATERIALS AND METHODS
Reagents and monoclonal antibodies.
An anti-P-/E-selectin MAb EP5C7 was generously provided by Dr. Cary L. Queen (Protein Design Labs, Fremont, CA). The Fab anti-αIIbβ3 MAb c7E3 was from Centocor (Malvern, PA). The nonpeptide small-molecule platelet αIIbβ3 antagonist XV454 (1) was a kind gift of Dr. Shaker A. Mousa (Albany College of Pharmacy, Albany, NY). Purified von Willebrand factor (vWf) (essentially factor VIII free) was purchased from Haematologic Technologies (Essex Junction, VT). The fibrinogen (vWf-, plasminogen-, and fibronectin-free) was a kind gift from Dr. Owen McCarty (Oregon Health & Science University, Portland, OR). The anti-human fibrin MAb MH-1 was generously provided by Dr. James McLinden (American Biogenetic Sciences, Copiague, NY). The seminaphtharhodafluor-1 (SNARF-1), carboxylic acid, acetate, succinimidyl ester red fluorescent mitochondrial dye was from Invitrogen (Carlsbad, CA). The chimeric form of P-selectin-IgG Fc (P-selectin) was a generous gift from Dr. Ray Camphausen of Wyeth External Research (Cambridge, MA). The phycoerythrin (PE)-conjugated anti-CD44 (515), PE-conjugated anti-CD162 (anti-PSGL-1), and isotype control antibodies were from BD Biosciences (San Jose, CA). All other reagents were from Sigma Chemical (St. Louis, MO).
Cell culture and staining.
LS174T human colon adenocarcinoma cells and THP-1 monocytic cells were obtained from the American Type Culture Collection (Manassas, VA) and were cultured in the recommended medium. CD44-knockdown (38) or control LS174T cells were harvested by mild trypsinization (0.25% trypsin/EDTA for 5 min at 37°C) and subsequently incubated at 37°C for 2 h to allow regeneration of surface glycoproteins (8, 20, 32). During this time, the carcinoma cell suspensions (107 cells/ml) were incubated with 5 μM SNARF for 60 min at 37°C. LS174T cells were then washed once to remove excess dye, resuspended in Dulbecco's phosphate-buffered saline (DPBS) containing Ca2+/Mg2+, and stored at 4°C for no longer than 4 h before use in aggregation assays, flow chamber, or flow cytometry. THP-1 cells were grown in suspension (1) and stained with 5 μM SNARF.
Human blood samples were drawn by venipuncture from healthy volunteers into heparin (10 units/ml) anticoagulant after written informed consent was obtained. Platelet-rich plasma (PRP) was prepared by centrifugation of whole blood at 160 g for 15 min. PRP was subjected to a further centrifugation (1,100 g for 15 min) in the presence of 2 μM PGE1, and the platelet pellet was resuspended in HEPES-Tyrode buffer containing 5 mM EGTA and 2 μM PGE1 (34). Thereafter, platelets were washed once via centrifugation at 1,100 g for 10 min, resuspended at 2 × 108 platelets/ml in HEPES-Tyrode buffer, and kept at room temperature (RT) for no longer than 4 h before use in rheometric assays.
Platelet-poor plasma (PPP) was obtained by centrifugation (1,900 g for 15 min) of the blood remaining after PRP removal. The final platelet count was adjusted to the desired levels by dilution with either HEPES-Tyrode buffer or PPP. These studies have been reviewed and approved by the Institutional Review Board of the Johns Hopkins University.
Cone-and-plate rheometry assays.
Platelet and SNARF-stained LS174T colon carcinoma or THP-1 cell suspensions were allowed to equilibrate separately to 37°C for 2 min. Thereafter, 50 μl of LS174T or THP-1 cells (2 × 107 cells/ml) along with 150 μl of platelets (2 × 107 platelets/ml) were placed onto the stationary plate of a cone-and-plate rheometer (RS150; Haake, Paramus, NJ) to achieve the desired ratio of platelets to LS174T or THP-1 cells (3:1). Shear rates were varied from 100 s−1 to 5,000 s−1 (microcirculation to pathological conditions) for 30 or 60 s. Static conditions were achieved by setting the shear rate to 0 s−1. The 0.5° cone and plate of the rheometer were maintained at 37°C during the entire experiment. Upon termination of shear, samples were obtained by pipette and instantly fixed with 1% formaldehyde (35). Immediately thereafter, specimens were allowed to incubate with a FITC-labeled platelet-specific MAb directed against platelet αIIbβ3 (5 μg/ml c7E3-FITC) for 30 min in the dark at RT. The labeling reaction was then stopped by further dilution with 1% formaldehyde, and specimens were subsequently analyzed in a FACSCalibur flow cytometer (Becton Dickinson).
Cell treatment with thrombin, MAbs, and enzymes.
To potentiate platelet activation, platelet specimens were incubated for 10 min before shear exposure with thrombin (2 U/ml) in the presence of the fibrin polymerization inhibitor glycine-proline-arginine-proline (GPRP)-NH2 (2 mM). The inclusion of GPRP-NH2 prevented not only fibrin polymerization but also the formation of homotypic platelet aggregates, even after the exposure of specimens to thrombin and/or relatively high shear rates up to 5,000 s−1 (35). Purified fibrinogen (1 mg/ml final concentration), vWf (7.5 μg/ml), human IgG (5 mg/ml), or human albumin (5 mg/ml) was added to the platelet suspension before thrombin/GPRP-NH2 incubation. In select experiments, fibrinogen (1 mg/ml) was added to the cell suspension on the plate 1 s before the onset of shear.
For inhibition studies, platelets were pretreated for 10 min with EP5C7 (50 μg/ml) and/or XV454 (100 nM), which were kept present during the aggregation assays. In parallel, control experiments were performed in which platelets and tumor cells were treated exactly as stated above but in the absence of any function-blocking inhibitor.
Quantitation of heteroaggregation and ligand site density by flow cytometry.
The particle distribution and cellular composition of stable aggregates generated in the rheometric assay were determined by a dual-color flow cytometric methodology. In brief, SNARF-stained cells and FITC-labeled platelets were identified on the basis of their characteristic forward-scatter, side-scatter, and fluorescence profiles in a FACSCalibur flow cytometer. Acquisition and processing of 3,000 SNARF-stained LS174T (or THP-1) cell events were then used to determine 1) the percentage of platelet-LS174T or THP-1 cell heteroaggregation and 2) the population distribution of bound platelets to the tumor cell surface, as previously described (2, 35). The site density of CD44 on LS174T cells and PSGL-1 on THP-1 cells was quantified by flow cytometry, using the PE-conjugated anti-human CD44 (515) or CD162 (KPL-1) MAb, respectively, in conjunction with the Quantibrite PE Phycoerythrin Fluorescence Quantitation Kit (BD Biosciences) to create a calibration curve to relate the geometric mean fluorescence intensity to the number of PE molecules bound per cell (14).
Colon carcinoma cell lysis and immunoprecipitation of CD44.
LS174T whole cell lysate was prepared by membrane disruption using 2% Nonidet P-40 followed by differential centrifugation (14, 38). CD44 was immunoprecipitated from colon carcinoma cell lysate with an anti-CD44 MAb, 2C5, using recombinant protein G agarose beads (Invitrogen) (14, 38).
Preparation of CD44-coated microspheres.
Immunoprecipitated CD44 from LS174T whole cell lysate or human IgG was diluted to desired concentrations with binding buffer (0.2 M carbonate/bicarbonate buffer, pH 9.2) and was incubated with 10-μm polystyrene microspheres (2.5 × 107 microspheres/ml; Bangs Labs, Fishers, IN) overnight at 4°C with constant rotation (14, 38). Microspheres were washed twice with DPBS and subsequently blocked with DPBS-1% BSA for 30 min at RT. Microspheres were resuspended (2 × 106 microspheres/ml) in DPBS-0.1% BSA for use in flow cytometric and flow chamber assays.
Flow-based adhesion assays.
Fibrinogen-coated surfaces were prepared by layering 1 mg/ml fibrinogen (vWf-, plasminogen-, and fibronectin-free) on 35-mm polystyrene suspension dishes and allowing it to coat overnight at 4°C. Fibrin-coated surfaces were prepared by two different previously published protocols: 1) by incubating fibrinogen-coated dishes with thrombin (2 U/ml) for 2 h at 37°C (31) and 2) by incubating purified fibrinogen in solution with 2 mM GPRP-NH2 and 0.25 U/ml thrombin for 10 min at 37°C before layering it on dishes (4) and incubating overnight at 4°C. After being washed three times with Ca2+/Mg2+-free DPBS, dishes were incubated with DPBS-1% BSA for 1 h at RT before use in adhesion assays. In select experiments, the fibrin-coated surfaces were further treated for 1 h at 37°C with 75 U/ml Streptomyces hyaluronlyticus hyaluronidase (EMD, San Diego, CA) (10) before their use in adhesion assays. Fibrin(ogen) coating was visualized by phase contrast microscopy. Suspensions of either CD44-knockdown or control LS174T cells or THP-1 cells (106 cells/ml) or CD44-coated or IgG-coated microspheres (2 × 106 beads/ml) were perfused for 5 min at 37°C over the fibrin- or fibrinogen-coated dishes using a parallel-plate flow chamber (250-μm channel depth and 5.0-mm channel width) (14, 36, 38). The extent of adhesion was quantified by averaging the total number of bound cells over multiple ×10 fields of view (0.55 mm2 each) at the end of the 5-min perfusion run for each condition.
SDS-PAGE and Western blotting.
Purified fibrinogen and freshly prepared monomeric fibrin in solution were diluted with DPBS and NuPAGE LDS sample buffer and separated via electrophoresis using a 4–12% Bis-Tris NuPAGE gel (Invitrogen) (56). The gel was either fixed with 50% methanol-10% acetic acid followed by staining with GelCode Blue Coomassie reagent (Pierce) or resolved proteins were transferred to Sequi-blot nitrocellulose membrane (Bio-Rad) and blocked with StartingBlock (Pierce) for 15 min. Immunoblots were stained with the human anti-fibrin MAb MH-1 and rinsed with Tris-buffered saline-0.1% Tween 20. Subsequently, blots were incubated with the appropriate horseradish peroxidase-conjugated secondary antibody and developed with SuperSignal West Pico chemiluminescent substrate (Pierce).
Determination of adhesion efficiency of platelet binding to tumor cells.
Platelet-LS174T (or THP-1) cell adhesion efficiency is defined as the fraction of heterotypic shear-induced collisions that result in stable heteroaggregate formation, and it is determined by the intercellular collision frequency and the capture efficiency of these collisions. This index was determined by fitting the aggregation data over the first 30 s after application of shear with a mathematical model based on the Smoluchowski two-body collision theory, as previously described (22, 35, 43). The radii of single LS174T and THP-1 cells were calculated to be 6.5 and 7 μm, respectively, by image processing, whereas the corresponding radius for a platelet singlet was set to 1.34 μm (50).
Data are expressed as means ± SE. Statistical significance of differences between means was determined by ANOVA or Student's t-test wherever appropriate. If means were shown to be significantly different, multiple comparisons were performed by the Tukey test. Probability values of P < 0.05 were selected to be statistically significant.
Comparative binding kinetics of LS174T colon carcinoma cells versus THP-1 cells to platelets in shear flow.
Previous work has established that platelets bind to colon carcinoma LS174T cells under dynamic flow conditions in blood plasma (35), which contains proteins such as fibrinogen and vWf that are capable of supporting adhesive interactions. To investigate how the hydrodynamic shear environment of the circulation modulates these heterotypic adhesive events in the absence of plasma proteins, washed platelets in HEPES-Tyrode buffer, stimulated for 10 min with thrombin (2 U/ml) and the fibrin polymerization inhibitor GPRP-NH2 (2 mM), were combined with SNARF-stained LS174T cells in the cone-and-plate rheometer and subjected to prescribed levels of shear for 30 or 60 s. A ratio of three platelets (1.5 × 107 platelets/ml) to one LS174T (5 × 106 cells/ml) was maintained in all experiments. The subphysiological concentration of platelets was selected to prevent homotypic platelet aggregation, which interferes with platelet-LS174T heteroaggregation (35). Under these conditions, platelet-platelet aggregation was minimal, with less than ∼5% of platelets in homotypic aggregates, and LS174T cells did not aggregate homotypically. A baseline level of platelet-LS174T cell binding (12.4 ± 2.0% LS174T cells in heteroaggregates, n = 5) was detected under static no-flow (0 s−1) conditions at 30 s, and it remained unaltered at 60 s (Fig. 1A). Application of increasing levels of hydrodynamic shear progressively augmented the extent of heterotypic aggregation, up to a maximum of 27.0 ± 5.3% at 800 s−1 at the 30-s time point (Fig. 1A). The extent of platelet recruitment by LS174T cells also tended to increase with increasing shear exposure time over a wide range of shear rates (Fig. 1A). At shear rates above 2,500 s−1, the extent of platelet-LS174T heteroaggregation decreased with increasing shear (Fig. 1A). For comparison purposes, we examined the adhesive behavior of PSGL-1-expressing THP-1 monocytic-like cells with washed platelets in shear flow. Our data reveal that the extent of platelet recruitment by THP-1 cells increased with increasing shear rate, plateaued at 800 s−1, and remained essentially unaltered up to 5,000 s−1 (Fig. 1A). To quantify cell-cell interactions independent of physical parameters such as cell size and intercellular collision frequency, we estimated the efficiency of platelet capture by LS174T or THP-1 cells by fitting the aggregation data over the first 30 s after application of shear to a mathematical model based on Smoluchowski's two-body collision theory (2, 21, 22, 35, 43). Maximal adhesion efficiencies were observed at the lowest shear rate of 100 s−1, at which ∼18 versus ∼9 of 100 collisions led to stable platelet-THP-1 versus platelet-LS174T cell aggregate formation. In contrast with the monotonic decrease of the platelet-LS174T adhesion efficiency with increasing shear, the efficiency of platelet recruitment by THP-1 cells was essentially constant between 400 and 800 s−1 before decreasing modestly at higher shear rates (Fig. 1B). Our data also show that THP-1 relative to LS174T cells displayed a higher efficacy to capture thrombin-activated platelets especially in the high shear regime (Fig. 1B), which can be explained by the increased percentage of THP-1 cells with adherent platelets (Fig. 1A) and the higher number of platelets bound per THP-1 cell (Fig. 1, C and D). It is noteworthy that application of shear alone (100–5,000 s−1, 60 s) in the absence of any exogenously added chemical agonist failed to induce aggregation between washed platelets and LS174T or THP-1 cells beyond the basal levels detected at 0 s−1 (data not shown).
CD44 is the primary P-selectin ligand on LS174T carcinoma cells.
Ensuing experiments focused on determining the molecular constituents, which are involved in platelet binding to LS174T colon carcinoma versus THP-1 monocytic-like cells under dynamic shear conditions. In view of our observations showing that significant heteroaggregation is detected only upon washed platelet activation with thrombin/GPRP-NH2, platelet P-selectin and/or αIIbβ3-integrins emerge as plausible candidates, since both the active conformation of the αIIbβ3-integrins and the surface expression of P-selectin and the αIIbβ3-integrins are induced by platelet stimulation (27). Furthermore, prior work has implicated these molecules in LS174T cell adhesion to activated platelets in plasma at moderate (800 s−1) but not low (100 s−1) shear conditions (35). As shown in Fig. 2A, individual blockade of either platelet P-selectin or αIIbβ3-integrin function was equally effective in suppressing the extent of platelet recruitment by LS174T colon carcinoma cells at all shear rates, except for 100 and 5,000 s−1, at which heteroaggregation in the absence of MAbs was near basal levels. No statistically significant additive effect was noted when P-selectin and αIIbβ3-integrin antagonists were used simultaneously, although this intervention consistently reduced the extent of heterotypic aggregation down to baseline levels (Fig. 2A). Given the distinct characteristics of selectin-ligand versus integrin-ligand binding kinetics (48), these data suggest that platelet P-selectin and αIIbβ3-integrins act primarily in a sequential manner to mediate maximal heterotypic binding under shear. In contrast, the use of the platelet αIIbβ3-specific antagonist XV454 (1) alone failed to impair the extent of platelet-THP-1 cell binding at all shear rates (Fig. 2B). Antibody interference assays also reveal that P-selectin is necessary and sufficient to support adhesion of activated washed platelets to THP-1 cells at 5,000 s−1 (Fig. 2B). At lower shear rates (100–2,500 s−1), blocking P-selectin partially repressed heteroaggregation (Fig. 2B). Interestingly, simultaneous blockade of P-selectin and αIIbβ3-integrin function abolished platelet-THP-1 cell binding (Fig. 2B), suggesting an auxiliary role for αIIbβ3-integrins in this adhesion process. These data are in concert with previous observations disclosing that 1) αIIbβ3-integrins stabilize the adhesion of free-flowing THP-1 cells to immobilized platelets in shear flow (1, 37) and 2) P-selectin binding to PSGL-1 is solely responsible for maximal adhesion of activated platelets to unstimulated blood monocytes in shear flow (2).
Using purified P-selectin-IgG Fc as a substrate, we recently discovered that sialofucosylated CD44v represents the major functional P-selectin ligand on a variety of colon carcinoma cells, including LS174T (14, 38). However, the physiological significance of P-selectin-CD44 binding using free-flowing intact cells has yet to be established. To this end, thrombin-GPRP-NH2-activated washed platelets were combined with stable CD44-knockdown LS174T cells (38) in the cone-and-plate rheometer and were subjected to prescribed levels of shear for 60 s. As a control, activated platelets were mixed in the linear shear field of a cone-and-plate rheometer with stable LS174T transfectants generated by using a mammalian scramble control plasmid (38). As shown in Fig. 2C, the CD44-knockdown LS174T cells displayed a markedly reduced capacity relative to mammalian scramble controls to recruit activated platelets in free-cell suspensions, disclosing that CD44 is the predominant ligand for platelet bound P-selectin.
Effect of plasma proteins on platelet binding to LS174T colon carcinoma cells in shear flow.
We next aimed to delineate the effects of blood plasma on the binding interaction of activated platelets with LS174T cells under dynamic flow conditions. As a first step, washed platelets, resuspended in either HEPES-Tyrode buffer or PPP and stimulated with thrombin/GPRP-NH2 for 10 min, were combined with LS174T cells and subjected to controlled levels of hydrodynamic shear for 60 s. Our data indicate that the presence of plasma markedly affected the platelet-LS174T heteroaggregation process. More specifically, peak heteroaggregation in the presence and absence of blood plasma was detected at 100 s−1 and 800 s−1, respectively (Fig. 3A). Most importantly, the extent of heteroaggregation was significantly lower in plasma-containing specimens at all shear rates except at the low level of 100 s−1, at which the extent of platelet recruitment by LS174T cells was modestly higher in the presence rather than the absence of plasma (Fig. 3A). As a control experiment, we also examined the capacity of LS174T cells to capture platelets in PRP. Figure 3A shows that the extent and pattern of heteroaggregation between LS174T cells and PRP or washed platelets resuspended in PPP as a function of hydrodynamic shear were nearly identical. Cumulatively, these data suggest that a plasma constituent(s) exerts an inhibitory effect on platelet-LS174T colon carcinoma cell binding at shear rates ≥400 s−1.
To identify the potential inhibitory effect of key plasma proteins in this heteroaggregation process, purified fibrinogen (1 mg/ml), vWf (7.5 μg/ml), IgG (5 mg/ml), or human serum albumin (HSA; 5 mg/ml) was added to washed platelet suspensions in HEPES-Tyrode buffer during the 10-min incubation with thrombin/GPRP-NH2. Because thrombin cleaves fibrinogen to fibrin, in select experiments, purified fibrinogen was added to the cell suspension 1 s before the onset of shear application to minimize its cleavage. At the shear rate of 800 s−1, exogenously added purified vWf, IgG, and HSA did not interfere with the extent of platelet capture by LS174T cells. In distinct contrast, the presence of fibrin alone markedly suppressed heteroaggregation down to the levels detected in blood plasma-containing specimens (Fig. 3, A and B). Interestingly, the extent of heteroaggregation remained intact when purified fibrinogen was added to the platelet-LS174T cell suspensions just before the onset of shear (Fig. 3B). We also examined the effects of all the aforementioned exogenously added purified proteins on platelet-LS174T heterotypic aggregation at 100 s−1. In marked contrast with the data acquired in the high shear regime, the presence of fibrin did not alter the heteroaggregation levels at 100 s−1 (Fig. 3B). Moreover, exogenously added IgG and HSA mildly increased the extent of platelet recruitment by LS174T colon carcinoma cells at low shear conditions, whereas the addition of vWf and fibrinogen did not have any effect on this process (Fig. 3B). For comparison reasons, we assessed the effects of these purified proteins on platelet-THP-1 heteroaggregation in shear flow. As shown in Fig. 3C, neither of these exogenously added proteins significantly affected the extent of platelet capture by THP-1 cells at 100 or 800 s−1. Taken together, our data disclose that the presence of fibrin, but not fibrinogen or any other plasma proteins, selectively interferes with platelet binding to LS174T colon carcinoma cells at elevated levels of hydrodynamic shear.
CD44 is the primary fibrin receptor on LS174T colon carcinoma cells.
We next aimed to provide a mechanistic interpretation for the diminished levels of platelet-LS174T cell heteroaggregation in the presence of fibrin in shear flow. In view of our observations showing that platelet P-selectin and αIIbβ3-integrins act primarily in a sequential manner to mediate maximal heterotypic binding under shear (Fig. 2A), we hypothesized that fibrin, by binding to the major functional P-selectin ligand, CD44, interferes with the P-selectin-CD44 molecular recognition. To this end, we examined the relative adhesion levels of CD44-expressing and CD44-knockdown LS174T colon carcinoma cells to immobilized fibrin under shear. Fig. 4A shows that CD44-knockdown LS174T cells relative to mammalian scramble or untreated controls displayed a markedly reduced capacity (∼25% of control) to tether and adhere to purified fibrin at a wall shear stress of 0.5 dyn/cm2. Interestingly, LS174T colon carcinoma cells interacted minimally with immobilized fibrinogen in shear flow (Fig. 4A). As a control for proper fibrin(ogen) coating, free-flowing THP-1 monocytic-like cells interacted effectively with both fibrinogen and fibrin under dynamic flow conditions (Fig. 4A).
To verify that LS174T CD44 interacts directly with fibrin, but not fibrinogen, we used a cell-free flow-based adhesion assay, in which polystyrene beads, coated with CD44 immunopurified from LS174T cells, were perfused over immobilized fibrin and fibrinogen under controlled kinematic conditions. Our data reveal that CD44-coated beads interacted avidly and extensively with immobilized fibrin but not fibrinogen at 0.25 and 0.5 dyn/cm2 (Fig. 4B). To document the specificity of CD44-fibrin interaction in this assay, beads coated with human IgG (10 mg/ml) were perfused over fibrin- and fibrinogen-coated surfaces. As shown in Fig. 4B, IgG-coated beads interacted minimally with immobilized fibrin and fibrinogen in shear flow.
To examine whether small amounts of hyaluronic acid (HA) possibly contaminating the fibrin(ogen) preparations may be playing a role in CD44-mediated LS174T cell adhesion to fibrin, fibrin-coated surfaces were treated with hyaluronidase before the perfusion of intact LS174T colon carcinoma cells. Our data show that this enzymatic treatment did not interfere with the extent of LS174T cell binding to fibrin under flow (39 ± 3 vs. 39 ± 1 LS174T cells/mm2 in the presence or absence of hyaluronidase at 0.5 dyn/cm2). Further support for the absence of any HA contamination is provided by the fact that LS174T cells and CD44-coated microspheres fail to bind fibrinogen (see Fig. 4, A and B).
To confirm the purity of fibrinogen and fibrin as well as the generation of fibrin, we performed SDS-PAGE followed by immunoblot analysis using an anti-fibrin MAb, MH-1. Fig. 4C discloses that the molecular masses of fibrinogen and fibrin detected by SDS-PAGE match very well with previously published data (56). Our immunoblot data also confirm the generation of soluble fibrin upon fibrinogen incubation with thrombin/GPRP-NH2.
Taken altogether, our data provide the first direct evidence that CD44 is the primary fibrin, but not fibrinogen, receptor on LS174T colon carcinoma cells. Moreover, platelet P-selectin and fibrin compete for CD44 binding, a process that dictates the extent of platelet-colon carcinoma heterotypic cell aggregation.
The major findings of this work are as follows. 1) CD44 on LS174T colon carcinoma cells is the major functional ligand for platelet bound P-selectin. 2) Using stable CD44-knockdown LS174T cells as well as polystyrene beads coated with CD44 immunopurified from control (untreated) LS174T cells, we provided the first direct evidence that CD44 acts as the primary fibrin, but not fibrinogen, receptor on colon carcinoma cells. 3) We assessed the influence of plasma proteins in the modulation of platelet-LS174T cell heteroaggregation under shear and discovered that the presence of fibrin, but not fibrinogen, diminishes the extent of heteroaggregation. 4) We provided a mechanistic interpretation for the reduced platelet-LS174T heteroaggregation under shear in the presence of fibrin by showing that fibrin interferes with the platelet P-selectin-CD44 molecular recognition. To further support the validity of our concept, we used THP-1 monocytic cells as a negative control, which are known to interact with P-selectin via PSGL-1, in a CD44-independent manner. 5) LS174T relative to THP-1 cells display a lower efficacy to capture activated platelets in shear flow due to the lower affinity/avidity of the P-selectin-CD44 versus P-selectin-PSGL-1 binding interaction.
Platelet P-selectin binding to LS174T CD44 followed by αIIbβ3-integrin involvement regulates the extent of platelet-LS174T cell heteroaggregation.
Unactivated washed platelets in buffer failed to aggregate with LS174T colon carcinoma cells in shear flow beyond the basal levels detected at 0 s−1. In contrast, thrombin/GPRP-NH2-activated platelets were efficiently captured by LS174T cells when subjected to hydrodynamic shear levels >100 s−1. Application of increasing levels of shear progressively augmented the extent of platelet-LS174T heterotypic aggregation, up to a maximum at 800–2,500 s−1, and then decreased down to near basal levels at 5,000 s−1. Inclusion of exogenous thrombin in the medium degrades plasma and platelet-bound fibrinogen into fibrin, and it stimulates platelets by increasing the affinity of platelet receptors (e.g., αIIbβ3) for their respective ligands and inducing the release of adhesion molecules from intracellular pools onto the cell surface (e.g., P-selectin and αIIbβ3). In view of this evidence, we examined the potential role of these adhesion molecules in platelet-LS174T heteroaggregation. Blocking either platelet P-selectin or αIIbβ3-integrin function alone was equally effective in repressing the extent of platelet recruitment by LS174T colon carcinoma cells. No statistically significant additive effect was noted when P-selectin and αIIbβ3-integrin antagonists were used simultaneously. Selectin-ligand bonds have been reported to have high tensile strength and fast molecular association and dissociation rates (13, 15, 47). Moreover, integrin-ligand bonds cannot be formed at high shear and corresponding short contact times in the absence of selectin contribution (48). In light of these observations, our functional data suggest that P-selectin and αIIbβ3-integrins act primarily in a sequential manner to mediate maximal binding of activated platelets to LS174T colon carcinoma cells under dynamic flow conditions. On the other hand, the molecular interaction between platelet P-selectin and PSGL-1 on THP-1 cells is necessary and sufficient for maximal heteroaggregation at 5,000 s−1, although an auxiliary role for αIIbβ3-integrins cannot be excluded at lower rates. Interestingly, THP-1 relative to LS174T cells display a markedly increased efficacy to capture activated platelets in shear flow. Given the predominant role of platelet P-selectin in these adhesion events, we examined whether P-selectin ligands are expressed at substantially higher levels on THP-1 than LS174T cells. Using CD44-knockdown LS174T cells, we first established that CD44 is the primary physiological ligand for platelet bound P-selectin. Moreover, quantitative flow cytometric analysis (14) reveals that the site density of CD44 molecules (233 ± 52 molecules/μm2) on the LS174T cell surface is profoundly higher than that of PSGL-1 (66 ± 6 molecules/μm2) on THP-1 cells, suggesting that the higher levels of platelet-THP-1 heteroaggregation are not due to an overabundance of PSGL-1 versus CD44 expression, but rather due to the higher affinity of P-selectin-PSGL-1 than P-selectin-CD44 interactions. Taking into account that the adhesion efficiency of platelet capture by THP-1 cells is two- to sixfold higher than LS174T cells along with the differential expression levels of PSGL-1 and CD44 on the respective cell surfaces, we conclude that the relative avidity of P-selectin-PSGL-1 interaction is 7- to 21-fold higher than that of P-selectin-CD44 binding.
CD44 on LS174T colon carcinoma cells acts as the primary fibrin, but not fibrinogen, receptor.
Prior work using an anti-CD44 whole IgG MAb, BU52, revealed a partial reduction (∼50%) in the extent of human lung fibroblast adhesion to both fibrinogen and polymerized fibrin gels under static conditions (49), suggesting that CD44 may support fibroblast binding to fibrin(ogen). However, as the authors stated in their article “the issue of whether CD44 interacts directly or indirectly with fibrin(ogen) remains to be defined” (49). Using stable CD44-knockdown LS174T colon carcinoma cells, we provide the first direct evidence that CD44 serves as the primary functional fibrin, but not fibrinogen, receptor on colon carcinoma cells in shear flow. Further direct support to this concept is provided by cell-free flow-based adhesion assays that disclose that microspheres, coated with CD44 immunopurified by LS174T colon carcinoma cells, interact extensively with immobilized fibrin, but not fibrinogen, in shear flow. The fact that CD44-expressing LS174T cells or CD44-coated beads bind fibrin, but not fibrinogen, in appreciable levels suggests that the binding sites are cryptic in fibrinogen but become exposed as a consequence of fibrin formation (52).
Fibrin, by interfering with the platelet P-selectin-LS174T CD44 molecular recognition, mitigates the extent of platelet-LS174T cell heteroaggregation under elevated shear conditions.
We have previously shown that resting platelets suspended in plasma readily aggregated with LS174T colon carcinoma cells when subjected to well-defined hydrodynamic shear conditions in the absence of any exogenously added chemical stimulus (35). Interestingly, peak adhesion efficiency was detected at low shear (20 to 50 s−1), was significantly potentiated by thrombin activation, and decreased with increasing shear rate, reaching a background value at 1,000 s−1 irrespective of the state of platelet activation (35). In distinct contrast, resting platelets suspended in buffer in the absence of any plasma proteins failed to aggregate with LS174T colon carcinoma cells in shear flow beyond the basal levels detected at 0 s−1. Moreover, significant binding of thrombin-activated washed platelets to LS174T cells in buffer was noted at shear rates >100 s−1 and up to 2,500 s−1. Cumulatively, these observations disclose the dual role of plasma proteins in the platelet-colon carcinoma heteroaggregation process in shear flow. More specifically, at low shear rates (up to 100 s−1) plasma proteins including IgG and HSA augment the extent of platelet capture by LS174T cells, possibly via a P-selectin-independent/RGD-dependent pathway (35). At moderate and high shear levels, platelet P-selectin and αIIbβ3-integrins act mainly in a sequential manner to mediate maximal heterotypic binding. Given the dual role of CD44 as a P-selectin ligand and fibrin receptor, we speculate that CD44 binding to P-selectin initiates colon carcinoma cell tethering to activated platelets, which may be stabilized, among others, via the subsequent molecular recognition of platelet αIIbβ3-bound fibrin by CD44. Interestingly, the presence of plasma fibrin, but not fibrinogen or any other plasma proteins, selectively interferes with the platelet P-selectin-CD44 binding interaction via competitive inhibition and drastically diminishes the extent of platelet recruitment by LS174T colon carcinoma cells at elevated levels of shear. To further support the validity of this concept, we used THP-1 monocytic cells as a negative control, which interact with P-selectin via PSGL-1, in a CD44-independent manner. Although THP-1 cells are capable of interacting with both fibrinogen and fibrin, neither of these proteins interfered with the P-selectin-PSGL-1 molecular recognition and thus did not alter the extent of platelet recruitment by THP-1 cells.
The role of CD44 in hematogenous metastasis.
This is the first evidence that CD44 has a dual role in mediating tethering and rolling through its binding to P-selectin (38) and firm adhesion via its interaction with fibrin. Taking into account the previously established HA binding function of CD44 (3) and its role in migration (51), the multitasking functions of CD44 in tethering/rolling, firm adhesion, and migration emerge from this and other prior studies.
Mounting evidence suggests that CD44 is aberrantly expressed in many human tumors (45). In certain cases, such as with colorectal carcinomas, the expression of CD44 confers metastatic potential in vivo (16, 28, 46), and it results in poor prognosis (53). Interestingly, the upregulation of CD44 expression appears to be an early event in colon carcinogenesis (24) and requires adenomatous polyposis coli gene inactivation (54). Knocking down CD44 expression in colon carcinoma cells drastically reduces their metastatic capacity in a mouse model (16). These observations coupled to the well-established HA binding function of CD44 have led to the hypothesis that CD44-mediated tumor cell adhesion to HA is a dominant factor regulating metastasis (25). However, our data revealing the P-selectin and fibrin binding function of CD44 bring another dimension to the existing dogma and offer a novel, unifying perspective on the apparent metastatic potential associated with CD44 overexpression on colon carcinoma cells and the critical role of selectins and fibrin in metastatic spread. Our findings support further research to investigate CD44 as a potential therapeutic target to combat metastasis and contribute to the complexity of the possible functions from this ubiquitous adhesion molecule.
This work was supported by National Cancer Institute Grant R01-CA101135 from the National Institutes of Health (to K. Konstantopoulos), the Masson-Agarwal Faculty Scholar award (to K. Konstantopoulos), a National Science Foundation Graduate Research Fellowship (to S. N. Thomas), and a predoctoral fellowship from the American Heart Association (to C. S. Alves).
↵* C. S. Alves and M. M. Burdick contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society