INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DURING BLOOD-BORNE METASTASIS cancerous cells undergo extensive interactions with various host cells, including polymorphonuclear leukocytes (PMNs), in the circulatory system before they eventually colonize a distant tissue. Several lines of evidence suggest that these adhesive interactions may affect the ability of tumor cells to metastasize. Some studies indicate that PMNs may exert a direct cytotoxic effect on tumor cells (7), and their tumoricidal ability relies on intimate contact between PMNs and tumor cells (6, 19). However, several independent studies have shown that PMNs may enhance tumor metastasis. This has been demonstrated by experiments showing that PMNs facilitate tumor cell extravasation in vitro (35, 38) and promote the arrest and deposition of tumors in the microvasculature of target organs in animal models (3, 11, 35). Light and electron microscopic studies also reveal close association of PMNs with metastatic cancer cells in vivo (5). Hence, whether PMNs abet hematogenous metastasis or are cytotoxic to tumor cells remains controversial. Nevertheless, PMN-tumor cell binding appears to be critical to both these processes, and the molecular mechanisms mediating these cell-cell interactions require further investigation.

Published data indicate that PMNs can bind certain melanoma, neuroblastoma, and colon adenocarcinoma cells through the CD18 (₂) integrin receptor on PMN surface in the absence of any selectin contribution under static conditions (16, 22, 26). Alternatively, PMNs have been reported to interact with certain colon carcinomas via PMN CD62L (L-selectin) in a CD18-independent manner at 4°C, a temperature that renders integrins inactive (20). However, static binding assays fail to replicate the hydrodynamic shear environment of the vasculature in which PMN-tumor cell binding typically occurs and, thus, may not be sufficient to determine the precise molecular requirements of this adhesion process. It is now well established that fluid shear affects the binding kinetics and receptor specificity of PMN homotypic aggregation (37) as well as PMN heterotypic interactions with platelets (15) and endothelial cells (13, 34). Furthermore, we recently showed that CD11b (Mac-1) is sufficient to mediate binding of chemotactically stimulated PMNs to CD54 (ICAM-1)-negative/sLe^x-bearing LS174T colon carcinoma cells at low shear (100 s¹) (12). In marked contrast, L-selectin, CD11a (lymphocyte function-associated antigen LFA-1), and CD11b are all requisite for optimal PMN-LS174T binding under high shear conditions (800 s¹) (12). Along these lines, the CD54-expressing/sLe^x-low HCT-8 colon carcinoma cells fail to aggregate with PMNs at 800 s¹. However, a detailed quantitative analysis of PMN binding to LS174T vs. HCT-8 colon carcinoma cells as well as the relative contributions of L-selectin, CD11a, CD11b, and their respective ligands in mediating these heterotypic adhesive interactions as a function of fluid shear and time have yet to be investigated.

In the present study, we utilized a rheometric-flow cytometric methodology to compare the aggregation kinetics, stability, and molecular requirements of PMN binding to LS174T cells with that to HCT-8 cells under defined shear conditions in the presence of the chemotactic agent N-formyl-methionyl-leucyl-phenylalanine (fMLP). To quantify cell-cell interactions independent of physical parameters such as shear rate, cell size, and initial cell concentration, we calculated the efficiency of PMN homotypic and PMN-colon carcinoma heterotypic aggregation by using a mathematical model based on Smoluchowski's two-body collision theory (33, 36, 37). Moreover, the contributions of intercellular contact duration (dictated by shear rate) and tensile forces (exerted by shear stress) on receptor-ligand bonds mediating PMN-colon carcinoma heteroaggregation were differentiated by varying the viscosity of the cell suspension (10, 29, 37).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents. The IgG murine monoclonal antibodies (MAbs) 6.7(blocking anti-CD18), HI111 (blocking anti-CD11a), and ICRF44(44) (blocking anti-CD11b) were purchased from BD Pharmingen (San Diego, CA). The blocking F(ab')₂ anti-CD54 MAb MHCD54F was from Caltag (Burlingame, CA). Red-out and PMN isolation media were obtained from Robbins Scientific (Sunnyvale, CA). Citrate-phosphate-dextrose (CPD) solution, fMLP, chymotrypsin, trypsin, benzyl-2-acetamido-2-deoxy--D-galactopyranoside (Bzl GalNAc), tunicamycin, dextran sulfate, and fucoidan were from Sigma-Aldrich (St. Louis, MO). Heparin was from Elkins-Sinn (Cherry Hill, NJ). Dulbecco's phosphate-buffered saline (D-PBS) and trypsin/EDTA were acquired from Life Technologies (Gaithersburg, MD). Phosphatidylinositol-specific phospholipase C (PI-PLC) was purchased from Glyko (Novato, CA), and Vibrio cholerae neuraminidase was from Roche Molecular Biochemicals (Indianopolis, IN). d,l-Threo-1-phenyl-2-amino3-morpholino-1-propanol hydrochloride (threo-PPPP) was procured from Matreya (State College, PA). CellTracker orange dye CMTMR [5(6)-(4-chlormethylbenzoylamino)tetramethylrhodamine] and green dye CFDA-SE [5(6)-carboxyfluorescein diacetate-succinimidyl ester] were purchased from Molecular Probes (Eugene, OR). CMTMR and CFDA-SE are excited at 488 nm by the argon laser of a flow cytometer, and their emission spectra are well separated (570 nm for CMTMR and 515 nm for CFDA-SE), thereby allowing simultaneous two-color immunofluorescence measurements.

Colon carcinoma cell culture and staining. The LS174T and HCT-8 human colon adenocarcinoma cell lines were obtained from American Type Culture Collection (Manassas,VA) and cultured in the recommended media. Before each experiment, LS174T or HCT-8 cells were detached from the tissue culture substrate by mild trypsinization (0.25% trypsin/EDTA for 2 min at 37°C). The colon carcinoma cells were resuspended at a concentration of 1 × 10⁷ cells/ml in the recommended media and incubated for 2 h at 37°C to regenerate surface glycoproteins (12, 20, 21). During this period, the cells were incubated with 1 µM CMTMR (orange dye) at 37°C for 1 h. The cells were washed once to remove excessive dye and resuspended at a concentration of 2.5 × 10⁶ cells/ml in D-PBS containing Ca²⁺/Mg²⁺/0.1% BSA (Sigma-Aldrich) and maintained at 4°C for no longer than 3 h before being used in binding assays or flow cytometry.

PMN isolation and labeling. Human blood was drawn by venipuncture from healthy volunteers into a sterile syringe containing CPD solution (2.8 ml CPD/20 ml blood). A red blood cell agglutinating agent (Red-out; 1/100 volume) was added to the blood to minimize erythrocyte contamination in PMNs (12, 14). The blood was gently mixed, layered over the PMN isolation medium, and centrifuged to allow for density separation of cell populations. The PMN layer was carefully aspirated and washed once, and the PMNs were resuspended in Ca²⁺/Mg²⁺-free D-PBS at a concentration of 1 × 10⁷ cells/ml. The PMNs were then incubated with 0.1 µM CFDA-SE at 4°C for 1 h, washed once to remove excess dye, resuspended in Ca²⁺/Mg²⁺-free D-PBS containing 0.1% BSA at a concentration of 0.5 × 10⁷ cells/ml, and maintained at 4°C for no longer than 3 h before being used in aggregation assays. L-selectin was preserved during PMN isolation and staining with CFDA-SE as assessed by flow cytometry (12).

PMN-colon carcinoma cell aggregation assay. CFDA-SE-labeled PMNs and CMTMR-stained colon carcinoma cells were mixed at final cell concentrations of 1 × 10⁶ and 2 × 10⁶ cells/ml, respectively, and allowed to equilibrate at 37°C for 2 min. This mixed suspension (500 µl) was then placed on the plate of a cone-and-plate rheometer (RS150, 60-mm-diameter plate, 0.5°Cone angle; Haake, Paramus, NJ) and stimulated with 1 µM fMLP 1 s before initiation of shear at 37°C. The suspension was subjected to well-defined hydrodynamic conditions with shear rates ranging from 50 to 1,200 s¹ for durations of 10, 30, 60, 120, 180, and 300 s. Upon termination of shear, 100-µl aliquots were immediately fixed with 1% formaldehyde (final concentration) at room temperature (RT) and subsequently analyzed in a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA), as previously described (12). The fixative was kept in the specimens during the flow cytometric analysis to preserve aggregate integrity (12). In selected experiments, 6% Ficoll was used to increase the viscosity of the cell suspension (10, 37).

Cell treatment with MAbs and enzymes. For some blocking experiments, PMNs (1 × 10⁷ cells/ml) were pretreated for 10 min at 37°C with anti-CD11a and/or anti-CD11b MAbs (20 µg/ml), which were kept present during the aggregation assays. In some studies, the HCT-8 cells were incubated with F(ab')₂ anti-CD54 MAb (10 µg/ml) for 10 min at 37°C before being mixed with PMNs in the cone-and-plate rheometer. For selected inhibition studies, PMNs were incubated with chymotrypsin (1 U/ml) for 20 min at RT to cleave L-selectin [and P-selectin-glycoprotein ligand-1 (PSGL-1)] from the PMN surface and washed once before being used in aggregation assays (12). For others, heparin (100 U/ml) (25), dextran sulfate (100 µg/ml) (32), or fucoidan (10 µg/ml) (8) were added to the suspension just before the application of shear. For experiments with biosynthetic inhibitors, LS174T cells were cultured for 48 h in the presence of either Bzl GalNAc (2 mM) to inhibit O-linked glycosylation or tunicamycin (200 ng/ml) to inhibit N-linked glycosylation of surface glycoproteins before being used in aggregation assays (30). In other studies, LS174T cells were cultured with 5 µM threo-PPPP for 96 h, which blocks the transfer of UDP-glucose to ceramide, thereby blocking the biosynthesis of glycosphingolipids having a glucosylceramide core (1). Colon carcinoma cells treated with the aforementioned biosynthetic inhibitors were washed twice before being used in aggregation assays (22). In selected experiments, LS174T cells were incubated at 37°C for 30 min with either 20 µg/ml trypsin to cleave cell surface glycoproteins or 0.1 U/ml Vibrio cholerae neuraminidase to cleave terminal cell surface sialic acid residues, or for 1 h with 1 U/ml PI-PLC, which cleaves glycosyl phosphatidylinositol (GPI)-linked surface glycoproteins (21), and washed once (twice in the case of trypsin) before being mixed with PMNs. Cell viability was consistently >97% after treatment with each of the aforementioned enzymes and inhibitors, as detected by the trypan blue exclusion assay.

Quantitation of receptor expression. For indirect immunofluorescence, colon carcinoma cells were incubated with saturating concentrations of primary MAbs directed against either sLe^x or CD54 for 30 min at 4°C and then washed once with D-PBS/0.1%BSA. After an additional 30-min incubation with 15 µg/ml FITC-labeled IgM or phycoerythrin-labeled IgG (Vector Laboratories, Burlingame, CA), the specimens were washed again, fixed with 1% formaldahyde, and analyzed by flow cytometry (21). Appropriate isotype-matched MAbs were also included for background fluorescence determination.

Flow cytometric analysis of PMN-colon carcinoma aggregation. A dual-color flow cytometric methodology (FACSCalibur flow cytometer) was employed to analyze the sheared samples for size distribution and cellular composition of formed aggregates. The PMN and colon carcinoma cell populations were identified on the basis of their forward-scatter, side-scatter, and fluorescence profiles. Single PMN (P) and single tumor cell (T) populations were identified on the green (FL1) vs. orange (FL2) fluorescence dot plots, and the aggregates were resolved as integral multiples of the mean fluorescence values of single P and T populations (12). The homotypic PMN aggregates were resolved into doublets (P₂) and higher order aggregates (P₃₊), whereas heterotypic aggregates consisted of a single colon carcinoma cell adhering to one (TP), two (TP₂), and three or more (TP₃₊) PMNs. The aggregation was quantified as the fraction of total PMN in homotypic or heterotypic aggregates as described previously (12).



Aggregates with more than three PMN were <10% of the total neutrophil population and resulted in <3% error in estimation of percent aggregation.

Estimation of PMN homotypic and PMN-colon carcinoma cell heterotypic adhesion efficiencies. PMN homotypic and PMN-colon carcinoma heterotypic aggregation observed in the cone-and-plate rheometer result from intercellular collisions induced by subjecting cell suspensions to fluid shear. The aggregation process may be expressed as a system of chemical reactions, where single PMN and colon carcinoma cells represent two monomer species and (k,l) denotes a general aggregate comprised of k PMN and l colon carcinoma cells. Each interaction between (k,l) and another aggregate (i,j) represents an elementary reaction step, and K_{[(i,j),(k,l)]} is the aggegation rate coefficient for this reaction. The two-species aggregation kinetics may then be expressed in the form of a population balance equation (PBE) (17)


where C(k,l) is the concentration of the aggregate and _i,j is the Kronecker delta function. The first term on the right-hand side of the PBE represents the formation of the aggregate (k,l) by the combination of two smaller aggregates, whereas the second term represents the consumption of the aggregate (k,l), resulting in the formation of a larger aggregate. k_max and l_max denote the maximum number of PMNs and colon carcinoma cells, respectively, that can be found in an aggregate. For our system, l_max = 1 because colon carcinoma cells do not aggregate homotypically under the experimental conditions of this study, and k_max = 3 because aggregates containing more than three PMNs are rare events.

1

Statistics. Data are expressed as means ± SE. Statistical significance of differences between means was determined by one-way ANOVA. Posttests were performed by using the Tukey method. Values of P < 0.05 were selected to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Comparative efficiency of heteroaggregation of LS174T and HCT-8 colon carcinoma cells with fMLP-stimulated PMNs under hydrodynamic shear. At low shear rates (50-200 s¹), PMN-LS174T heteroaggregation increased with time until 120 s, when as many as 60% PMNs were in heterotypic aggregates (Fig. 1A). Partial disaggregation of these heterotypic aggregates was observed at longer periods (180-300 s) of shear exposure for all shear rates examined. Similarly, HCT-8 cells aggregated extensively with PMNs under these low shear conditions, with maximal values 75% after 120 s of shear exposure (Fig. 1B). Partial disaggregation of heterotypic aggregates followed at later times for shear rates of 100-200 s¹. In marked contrast to LS174T cells, PMN-HCT-8 cell heteroaggregates did not undergo disaggregation at 50 s¹ (Figs. 1, A and B).

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Kinetics of polymorphonuclear leukocyte (PMN)-colon carcinoma cell and PMN-PMN aggregation. CMTMR (orange dye)-labeled colon carcinoma cells (LS174T or HCT-8, 2 × 106 cells/ml) were incubated with CFDA-SE (green dye)-labeled PMNs (1 × 106 cells/ml) at 37°C for 2 min, stimulated with 1 µM fMLP, and sheared at prescribed shear rates and times. Sheared specimens were immediately fixed with 1% formaldehyde and analyzed with a FACSCalibur flow cytometer. A and C: percentage of total PMNs in heterotypic and homotypic aggregates in the presence of LS174T cells, respectively. B and D: percentage of total PMNs in heterotypic and homotypic aggregates in the presence of HCT-8 cells, respectively. Values are means ± SE (n = 3-13).

1

1

1

1

1

1

1

1

1

1

View larger version (31K): [in this window] [in a new window]   Fig. 2.   PMN homotypic and PMN-colon carcinoma heterotypic adhesion efficiencies as a function of shear rate. The percentages of PMNs in homotypic (A) and heterotypic (C) aggregates after 30 s of shear exposure of either pure PMN or PMN-colon carcinoma cell suspensions were determined as a function of hydrodynamic shear by a dual-color flow cytometric methodology. The adhesion efficiencies for homotypic (B) and heterotypic (D) aggregation were calculated by using a mathematical model based on Smoluchowski's theory. The experimental protocol was similar to that mentioned in the legend of Fig. 1. Values are means ± SE (n = 3-6).

1

1

1

Relative contribution of CD11a, CD11b, and CD54 to PMN-colon carcinoma cell adhesion. Function-blocking MAbs were used against PMN CD11a and CD11b to investigate their relative roles in PMN-LS174T heteroaggregation in the shear range of 100-800 s¹. Blocking CD11b function abrogated PMN binding to LS174T cells under all shearing conditions (Fig. 3A). In distinct contrast, use of an anti-CD11a MAb did not appreciably affect PMN-LS174T heteroaggregation in the low-shear regime (100-200 s¹). However, PMN binding to LS174T cells was reduced by 40 and >60% in the presence of anti-CD11a MAb relative to control values at shear rates of 400 and 800 s¹, respectively. Together, our data illustrate that CD11b is requisite for PMN-LS174T binding over the entire range of shear rates examined, whereas CD11a contribution becomes evident only in the high-shear regime.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   The relative contributions of CD11a, CD11b, CD54, and L-selectin to PMN-colon carcinoma cell heteroaggregation. CFDA-SE-labeled PMNs (1 × 106 cells/ml) were incubated with anti-CD11a (20 µg/ml) and/or anti-CD11b (20 µg/ml) function-blocking MAbs at 37°C for 10 min before being mixed with CMTMR-labeled colon carcinoma cells (2 × 106 cells/ml) at 37°C for 2 min. The cell mixture was then stimulated with 1 µM fMLP 1 s before application of shear. Sheared samples were immediately fixed with 1% formaldehyde and analyzed by flow cytometry. A: the extent of PMN-LS174T heteroaggregation after shearing for 30 s at prescribed shear rates. B: the extent of PMN-HCT-8 cell heteroaggregation after 30 s of shearing at 100 s1. C: the extent of PMN-HCT-8 cell heterotypic aggregation after 300 s of shear exposure at 100 or 400 s1. In selected experiments, PMNs were treated with 1 U/106 cells chymotrypsin for 20 min at room temperature before being mixed with LS174T cells and subjected to shear. In some experiments, HCT-8 cells were incubated with an anti-CD54 function-blocking MAb at 37°C for 10 min before being mixed with PMNs. Values are means ± SE (n = 4-10). *P < 0.05 with respect to control (untreated) samples. #P < 0.05 with respect to CD54-treated samples.

1

1

1

1

1

1

1

1

Role of L-selectin in PMN-LS174T adhesion and characterization of L-selectin ligand on colon carcinoma cells. We next wanted to systematically investigate the relative contribution of L-selectin to the PMN-LS174T adhesion process over a wide range of shear rates (100-800 s¹) and to characterize the L-selectin ligand on LS174T cells. Chymotrypsin treatment (1 U/ml) has been shown to cleave L-selectin from PMN surface (12). In the low-shear regime (100-200 s¹; Fig. 3A), the extent of LS174T cell binding to chymotrypsin-treated PMNs was similar to that to untreated PMNs. In marked contrast, chymotrypsin treatment reduced PMN-LS174T heteroaggregation by 50% at 400 s¹ and by 85% at 800 s¹. These data suggest that the L-selectin contribution in mediating adhesion becomes progressively larger with increasing shear. Fucoidan, which has been shown to block PMN binding to L-selectin (8), abrogated PMN-LS174T heteroaggregation at 800 s¹ (Fig. 4), providing further evidence for L-selectin involvement in this process under high-shear conditions.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Characterization of L-selectin ligand on LS174T cells. CFDA-SE-labeled PMNs (1 × 106 cells/ml) were incubated with CMTMR-labeled LS174T cells (2 × 106 cells/ml) for 2 min at 37°C, stimulated with 1 µM fMLP, and sheared at 800 s1 for 30 s. Sheared samples were immediately fixed with 1% formaldehyde and analyzed with a FACSCalibur flow cytometer. In selected experiments, LS174T cells were cultured in the presence of either benzyl-2-acetamido-2-deoxy--D-galactopyranoside (Bzl GalNAc; 2 mM) or tunicamycin (200 ng/ml) for 48 h or with d,l-threo-1-phenyl-2-amino3-morpholino-1-propanol hydrochloride (threo-PPPP; 5 µM) for 96 h. Some experiments were conducted with LS174T cells treated with either trypsin (20 µg/ml for 30 min) or phosphatidylinositol-specific phospholipase C (PI-PLC; 1 U/ml for 1 h). Others were conducted in the presence of dextran sulfate (100 µg/ml), heparin (100 U/ml), or fucoidan (10 µg/ml). Values are means ± SE (n = 3-5). *P < 0.05 with respect to control (untreated) samples.

1

1

1

Effect of shear stress on PMN-colon carcinoma cell aggregate formation. Exposure of cell suspensions to increasing levels of shear rate increases the frequency of cell-cell collisions but decreases the intercellular contact duration. Concomitantly, there is an increase in the magnitude of tensile forces acting on the adhesive contact region to dissociate intercellular bonds. To differentiate the contributions of contact duration and tensile forces on cell aggregation, we used 6% Ficoll to double the viscosity of the suspending medium, thereby increasing the shear stress at a constant shear rate or contact time (10, 29, 37). PMN homotypic aggregation was unaffected by the increase in viscosity in the presence of LS174T (Fig. 5A) as well as HCT-8 cells (data not shown). Similarly, increasing the viscosity did not appreciably affect PMN-LS174T binding over the entire range of shear rates examined (Fig. 5B), suggesting that these interactions are dependent on contact duration and that the aggregates are sufficiently stable to resist the increased shear forces. However, in the case of HCT-8 cells, heterotypic aggregation at 30 s remained ~45% at low shear rates (50-200 s¹) for 0.8-cP suspensions, whereas heteroaggregation decreased sharply for 1.6-cP suspensions from ~55% at 50 s¹ to 22% at 200 s¹ (Fig. 5C). The significant reduction in PMN-HCT-8 heterotypic adhesion observed at 1.6 cP compared with that at 0.8 cP at a shear rate of 200 s¹ suggests that the adhesive bonds are not strong enough to resist the increased shear force acting on them. For shear rates >400 s¹, PMN-HCT-8 cell heteroaggregation was close to background levels in 0.8-cP as well as 1.6-cP suspensions (Fig. 5C).

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Effect of shear stress on PMN-colon carcinoma cell aggregate formation. Neuraminidase (0.1 U/ml)-treated or untreated CMTMR-labeled colon carcinoma cells (LS174T or HCT-8, 2 × 106 cells/ml) were incubated with CFDA-SE-labeled PMNs (1 × 106 cells/ml) at 37°C for 2 min. In selected experiments, the viscosity of the suspending medium was increased from 0.8 cP to 1.6 cP at 37°C with 6% Ficoll. The cell mixture was stimulated with 1 µM fMLP and sheared for 30 s at prescribed shear rates. Sheared specimens were immediately fixed with 1% formaldehyde and analyzed with a FACSCalibur flow cytometer. A and B: percentage of PMNs in homotypic and heterotypic aggregates, respectively, in the presence of untreated LS174T cells. C and D: percentage of PMNs in heterotypic aggregates in the presence of HCT-8 cells and neuraminidase-treated LS174T cells, respectively. Values are means ± SE (n = 3-4). *P < 0.05 with respect to the aggregation in 0.8-cP suspensions.

1

1

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The main findings of this work are as follows: 1) the adhesion efficiency of PMN homotypic aggregation is not affected by the presence of colon carcinoma cells over the entire range of shear rates examined in this work; 2) the efficiency of PMN-colon carcinoma cell heterotypic aggregation decreases with increasing shear, with PMN binding to CD54-bearing HCT-8 cells being more efficient than that to CD54-negative LS174T cells at low shear; 3) under these low-shear conditions, both CD11a and CD11b contribute to PMN-HCT-8 cell aggegation, with CD54 on HCT-8 cells acting as a CD11a ligand only at early time points; 4) CD11a involvement becomes progressively larger with increasing shear by binding to a yet unidentified ligand distinct from CD54; and 5) in the high-shear regime, only PMN-LS174T cell aggregation occurs, which is initiated by PMN L-selectin binding to a sialylated, O-linked, protease-sensitive ligand on LS174T cells. Thus we provide evidence that both fluid shear and shear exposure time modulate the molecular interactions between PMNs and colon carcinoma cells.

PMN homotypic adhesion efficiency is unaffected by the presence of colon carcinoma cells. PMN homotypic aggregation in PMN-colon carcinoma cell suspensions was significantly lower than that in pure PMN suspensions whenever appreciable PMN-colon carcinoma heteroaggregation occurred. We argue that the reduction in PMN homotypic aggregation occurred because of recruitment of PMNs by colon carcinoma cells and not because of the release of any inhibitory enzyme. This concept is corroborated by the fact that PMN homotypic adhesion efficiency estimated in the presence of colon carcinoma cells did not differ significantly from that calculated for pure PMN suspensions. PMN homotypic adhesion efficiency decreased with increasing shear in the low-shear regime. This decrease is attributed to the lower probability of CD18 integrin receptors to mediate stable adhesion at increased shear rates and correspondingly shorter intercellular contact durations in the absence of any selectin contribution. Nevertheless, the efficiency is nearly unchanged at higher shear rates (200 s¹) because of the involvement of L-selectin, which binds to its counterreceptor, PSGL-1, and mediates transient tethering between PMNs, thereby increasing the collisional contact duration and allowing the CD18 integrins to mediate stable adhesion (37). The PMN homotypic adhesion efficiency in our studies with pure suspensions was consistently lower than that previously reported (37). This discrepancy may be ascribed to three reasons. First, the coefficient for collision frequency used by Taylor et al. (37) was in error by a factor of 2, as acknowledged by the authors in a later publication (10). Second, omission of (1 + _i,jk,l) from the stoichiometric coefficients in the PBE (17) results in a significant error, as shown in Table 2. Last, the extent of PMN homotypic aggregation, which varies from donor to donor, was observed to be lower in our study than that reported by Taylor et al. (37).

The molecular mechanisms of PMN-binding to LS174T and HCT-8 colon carcinoma cells are shear and time dependent. The extent of PMN binding to HCT-8 cells was consistently higher than that to LS174T cells at low shear (50-200 s¹) and could be at least partially attributed to the presence of CD54 on the HCT-8 carcinoma cell surface. Under these low shear conditions, CD11a and CD11b are involved in PMN-HCT-8 binding at both short and long shear exposure times. This finding is in contrast to PMN homotypic aggregation (24) as well as PMN binding to CD54-transfected mouse cells (23), where adhesion was almost entirely CD11b dependent at long shear exposure times. However, this discrepancy may be attributed to the presence of distinct CD11a ligands with markedly different binding kinetics in each of these cell types. The lack of any additive inhibitory effect upon simultaneous use of CD11a (on PMNs) and CD54 (on HCT-8 cells) suggests that PMN CD11a binds to CD54 expressed on HCT-8 colon carcinoma cells at short shear exposure times. However, at longer shear exposure, the inability of anti-CD54 mAb alone to substantially reduce PMN-HCT-8 cell adhesion suggests the presence of another ligand for CD11a with a higher affinity than CD54. Furthermore, CD11b appears to interact with a yet unidentified ligand on HCT-8 cells, rather than CD54 at all shear rates and shear exposure times examined here. Along these lines, PMN CD11b binds to CD54-negative LS174T colon carcinoma cells and mediates heteroaggregation under shear (12). It is noteworthy that at low shear (100 s¹), CD11a does not seem to be involved in PMN binding to LS174T cells. This is in accord with previous studies performed under static conditions showing that transmigration of fMLP-activated PMNs across CD54-negative T84 human colon adenocarcinoma cell layers is mediated by CD11b and is independent of CD11a (26-28). However, the contribution of CD11a to PMN-colon carcinoma heteroaggregation becomes evident (LS174T cells) or even more pronounced (CD54-expressing HCT-8 cells) in the high-shear regime (400 s¹). In distinct contrast, a previous study shows that PMN homotypic aggregation via CD11a alone is barely detectable at 400 s¹ and becomes entirely CD11a-independent/CD11b-dependent at even higher shear rates (24). The partial disaggregation of PMN-colon carcinoma cell aggregates observed at long shear exposure times may reflect a decay in the avidity of PMN CD11a and CD11b (24) as well as of their respective counterreceptors on the carcinoma cell surface.

L-selectin-mediated PMN-colon carcinoma cell binding depends on intercellular contact duration and is shear stress resistant. PMN homotypic and PMN-LS174T heterotypic aggregation were unaffected when the shear stress was doubled independently of the shear rate by increasing the viscosity of the medium from 0.8 to 1.6 cP. However, the twofold increase in shear stress caused a significant reduction of PMN binding to sLe^x-low HCT-8 cells and neuraminidase-treated (sLe^x-low) LS174T cells at 200 s¹. These data suggest that L-selectin-ligand interactions increase the contact duration between PMNs and sLe^x-expressing LS174T cells, thereby allowing a sufficient number of CD18 integrin bonds to form that can withstand the increase in shear stress by elevating the viscosity of the suspending medium. PMN L-selectin is not involved in PMN-LS174T heteroaggregation at low shear rates, but its contribution increases progressively with increasing shear rate, which is in agreement with previously published data on PMN homotypic aggregation (37). However, unlike PMN homotypic adhesion efficiency, the PMN-LS174T cell heterotypic efficiency falls with increasing shear (200-800 s¹), even though L-selectin is involved. This decrease may occur because the L-selectin ligand on LS174T cells may not be expressed in large numbers or may have a lower binding affinity for L-selectin. On the other hand, PMNs expressing L-selectin as well as its counterreceptor, PSGL-1, may preferentially bind other PMNs, rather than LS174T cells, which express only the L-selectin ligand and are devoid of L-selectin. The L-selectin ligand on LS174T colon carcinoma cells appears to be a sialylated, O-glycosylated, protease-sensitive molecule distinct from PSGL-1 (12). Moreover, the lack of any inhibitory effects on PMN binding to LS174T cells at high shear with the use of certain specific enzymes and biosynthetic inhibitors suggests that the L-selectin ligand is neither a glycosphingolipid nor a GPI-anchored molecule and does not require N-glycans for binding to L-selectin.

1