Department of Veterinary Molecular Biology, Marsh Laboratory, Montana State University, Bozeman
Correspondence: Steve D. Swain, Ph.D., Veterinary Molecular Biology, Marsh Laboratory, P.O. Box 173610, Montana State University, Bozeman, MT 59717. E-mail: uvsss{at}montana.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
was supported differentially by various ECMs, and PAF and IL-8 "priming" had strikingly different effects, depending on the ECM present. Although PAF and IL-8 inhibited TNF--induced ROS production in neutrophils adhered to collagen, fibronectin, and laminin, PAF enhanced ROS production strongly in HSP-adherent cells. This study illustrates how neutrophils can integrate multiple stimuli, resulting in complex modulation of their functional response.
Key Words: platelet activating factor • interleukin-8 • tumor necrosis factor- • reactive oxygen species
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Neutrophils must integrate a number of environmental signals, resulting in modulation of the type and amplitude of a given inflammatory response. For example, neutrophils in suspension typically exhibit different responses (e.g., respiratory burst kinetics) than neutrophils adherent to a physiological surface [8
], and this modulation seems to result from distinct signal-transduction processes within the neutrophil [9
, 10
]. One facet of the cellular environment that is thought to contribute significantly to neutrophil responsiveness is the extracellular matrix (ECM). Neutrophils encounter ECM proteins as they exit the vasculature, particularly at sites of tissue damage, which would seem to be a logical location for regulatory cues. Indeed, it has been suggested that through interactions with cytokines and enzymes, ECM proteins play a specialized role in providing signals that coordinate all leukocyte behaviors, thereby regulating inflammation [11
, 12
]. Adhesion to ECM proteins is especially important in modulation of neutrophil function. Neutrophil functions such as chemotaxis [13
, 14
], degranulation [15
, 16
], adhesion [17
], and phagocytosis [18
] have all been shown to be modulated by ECM proteins. Although most research into neutrophil/ECM interactions has focused on the necessity of ECM interactions to support the tumor necrosis factor- (TNF-)-induced oxidative burst [8
, 19
20
21
22
], it is clear that ECM proteins can have profound effects on neutrophil function in response to a number of inflammatory mediators.
Two inflammatory mediators that are implicated in regulating neutrophil function in vivo are interleukin-8 (IL-8) and platelet-activating factor (PAF). IL-8 is a proinflammatory cytokine produced by a wide range of cells, including monocytes, granulocytes, and epithelial and endothelial cells [23
]. Most commonly known as a chemotactic factor for neutrophils, IL-8 has also been shown to activate, degranulate, and prime neutrophils for subsequent stimulation by other activators such as TNF- [24
, 25
]. PAF is a phospholipid produced by platelets, neutrophils, monocytes, and endothelial cells [26
]. Neutrophils exhibit a wide range of inflammatory responses to PAF, including changes in intracellular calcium levels, membrane potential, intracellular pH, and actin polymerization [27
28
29
]. It is interesting that IL-8 and PAF can elicit different responses in neutrophils, depending on the cellular environment. For example, in an in vivo model of mastitis in cattle, IL-8 does not seem to attract neutrophils, and IL-8 is strongly chemotactic for bovine neutrophils in vitro [30
]. Conversely, the same study suggested that PAF was implicated in the recruitment of neutrophils in vivo, whereas it was not chemotactic for neutrophils in vitro. Currently, little is know about the mechanisms behind these differential responses, however there is some evidence that ECM proteins can have complex modulating effects on neutrophil responses to these agents. For example, heparan sulfate has been shown to enhance neutrophil responses to IL-8 [31
] but inhibit neutrophil responses to PAF [32
]. Clearly, further studies on the mechanisms involved in integration of environmental inputs will be essential to understand how neutrophils moderate their responses to provide the appropriate outcome to a given stimulus.
In the present study, we characterized functional responses of bovine neutrophils stimulated with IL-8 and PAF and determined if these responses were altered by interactions with a variety of relevant ECM proteins, including collagen IV, laminin, fibronectin, thrombospondin, and heparan sulfate proteoglycan (HSP). Our results indicate that interaction of neutrophils with extracellular matrix proteins can differentially modulate this cell’s responses to IL-8 and PAF. Cellular adhesive properties, F-actin polymerization, intracellular Ca2+ changes, and degranulation of neutrophils adherent to ECM proteins were distinct from those responses in cells adherent to plastic, and there were also some subtle differences between individual ECMs. The neutrophil response most sensitive to modification by concurrent stimulation with IL-8 or PAF and adhesion to ECM proteins was the TNF--induced respiratory burst. The fact that different combinations of ECM protein and IL-8 or PAF can have totally opposite effects on this response suggests that further study is warranted into the signal-transduction mechanisms involved in this message integration.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
were from Calbiochem-Novabiochem (San Diego, CA). Fluo-3 AM, methylumbelliferyl phosphate and BODIPY-phallacidin were from Molecular Probes (Eugene, OR). Lyso-PC was purchased from Avanti Polar-Lipids (Alabaster, AL). Dulbecco’s phosphate-buffered saline without calcium or magnesium (DPBS) was from Gibco-BRL (Grand Island, NY), and RPMI 1640 was purchased from BioWhittaker (Walkersville, MD). Fluoro-Nunc module microwell plates and Lab-tek chambered permanox slide systems were from Nalge Nunc International (Naperville, IL). All other reagents, including laminin, collagen type IV, heparan sulfate proteoglycan, phorbol myristate acetate (PMA), luminol (used at pH 9.0 in 0.2 M borate), and fatty acid-free bovine serum albumin (BSA) were from Sigma Chemical Co. (St. Louis, MO). IL-8 and PAF were diluted in DPBS containing 0.2% fatty acid-free BSA. HEPES-buffered saline (HBS) was made with 20 mM HEPES (pH 7.4) along with 125 mM NaCl, 5 mM KCl, 0.62 mM MgCl2, 1.8 mM CaCl2, and 6 mM glucose. Endotoxin-free water was used for all solutions to which cells were exposed.
Neutrophil isolation
Blood from Holstein calves (6 and 18 months of age) was collected into tubes containing 5 mM ethylenediaminetetraacetate. Neutrophils were isolated by hypotonic lysis of red blood cells followed by separation from mononuclear cells on a two-step Histopaque gradient, as described previously [33
]. Neutrophils purified with this technique were 95% pure by flow cytometric analysis and Wright staining and were 
98% viable, as determined by trypan blue exclusion.
Coating 96-well plates with ECM proteins
Ninety-six-well microtiter plates were rinsed twice with DPBS and coated with 2 µg/ml extracellular matrix proteins. Briefly, fibronectin, laminin, thrombospondin, and heparan sulfate proteoglycan were diluted to 20 µg/ml in RPMI, whereas collagen IV was diluted to 20 µg/ml in sterile H2O. The diluted proteins were then added to the wells and incubated at 37°C for 1 h. The coated plates were rinsed three times in DPBS and finally with injectable grade water to remove any nonbinding proteins or salts. Plates were wrapped in Parafilm and stored at 4°C for up to 2 weeks.
Measurement of neutrophil intracellular calcium flux
Changes in intracellular Ca2+ following treatment with IL-8 or PAF were measured using the Ca2+-sensitive probe, Fluo-3 AM [34
]. Isolated neutrophils were loaded with 3 µM Fluo-3 in the dark with rocking at 24°C for 30 min. After washing in DPBS, 5 x 105 cells (in 50 µl DPBS) were added to sets of ECM-coated strips (one strip each of ECM-coated and uncoated wells) containing 150 µl HBS. Plates were incubated for 1 h at 37°C. Using the Fluoroskan Ascent FL plate reader (Thermo Labsystems, Helsinki, Finland), the baseline level of fluorescence was measured for 50 s. IL-8 (20 µl; 1x10-8 M final concentration) or PAF (20 µl; 1x10-7 M final concentration) was then injected into the wells, and the subsequent fluorescent change (reflecting the intracellular Ca2+ response) was recorded for 75 s (0.5-s intervals). Results are expressed as the maximum change in fluorescence upon addition of PAF or IL-8, with results pooled over four separate experiments.
Measurement of neutrophil filamentous actin polymerization
The relative dynamics of F-actin polymerization were visualized in adherent neutrophils using a fluorescently labeled F-actin-binding probe (phallacidin) and fluorescent microscopy. Permanox slide systems were coated with individual ECM proteins as described for the 96-well plates. After rinsing with DPBS, 2 x 106 neutrophils in 200 µl DPBS containing 1 mM CaCl2 were added to each chamber and incubated 1 h at 37°C. The wells then received a treatment of IL-8 (1x10-8 M final concentration) or PAF (1x10-7 M final concentration). After addition of the stimulus, 200 µl 7.4% formaldehyde was added to fix individual wells of cells at 0 (no treatment), 30, or 120 s. After 15 min at room temperature, the chambers were rinsed twice with DPBS. The fixed cells were permeabilized and stained with 4.8 x 10-8 µM BODIPY-phallacidin and 6.4 µg/ml lysophosphatidylcholine of DPBS. The chambers were incubated for 15 min at 37°C, rinsed, and examined with fluorescent microscopy. Images were recorded using a Spot digital camera (Diagnostic Instruments, Sterling Heights, MI). Image analysis was performed using a PC version of the NIH Image program (Scion Image). Our initial observations indicate that treatment with PAF or IL-8 caused a rapid distribution of polymerized actin at the periphery of the cells but that there was some difference in the progression of this phenomenon, depending on the ECM being used. To provide a semi-quantitative assessment of this, the following procedure was used: Three transects were made across each cell image (at 0°–180°, 90°–270°, and 150°–330°). The pixel intensity of the brightest point in the outer 10% of the cell (A) and the dimmest point in the inner 90% of the cell (B) was recorded for each transect. A ratio of these values was determined using the formula (A-B)/B. Values for the three transects for each cell were averaged. These values were determined for five cells in each image and were averaged. Finally, the values for the same treatment (activator, ECM, and time) for three separate experiments were pooled.
Measurement of neutrophil degranulation
Degranulation of neutrophil-specific and azurophil granules was determined in response to IL-8 or PAF. Neutrophils (5x105) in 50 µl DPBS were allowed to adhere to ECM-coated wells containing 130 µL HBS for 1 h at 37°C. Wells received IL-8 (1x10-8 M final concentration), PAF (1x10-7 M final concentration), or buffer control or a combination of "primary" (IL-8 or PAF) and secondary "activation" treatments of TNF- (100 ng/ml final concentration). All treatments were added as 20 µl from a 10x stock solution with the exception of TNF-, which was added as 2 µl from a 100x solution. Cells treated with ionomycin (10 µM final concentration) were included as controls. Plates were incubated for 30 min at 37°C and then centrifuged at 4°C for 5 min with no brake at 350 g. The supernatant was collected and used for the granule assays: To determine specific granule degranulation, the amount of lactoferrin released was measured by a single-antibody-antigen enzyme-linked immunosorbent assay (ELISA), exactly as described previously [35
]. Azurophil granule degranulation, as determined by myeloperoxidase activity, was measured initially using the substrates pyrogallol and H2O2, as described previously [36
]. However, no detectable release of myeloperoxidase was determined using this technique. To verify that no myeloperoxidase was present in the samples and that assay sensitivity was not the limiting factor, a variation of luminol-enhanced chemiluminescence was also used. Because luminol chemiluminescence is dependent on myeloperoxidase [37
], we incubated samples with excess endogenous oxidant (H2O2) and luminol to probe for myeloperoxidase activity. These assays were performed using 50 µl samples of supernatant fluids in a 96-well plate, with luminol and H2O2 provided by the Kirkegaard & Perry LumiGlo chemiluminescent kit reagents (Gaithersburg, MD), and chemiluminescent measurements were made using the Fluoroskan Ascent FL. Myeloperoxidase in the supernatant fluids was also undetectable using this method, whereas measurable amounts were determined in supernatants from neutrophils treated with 10 µM ionomycin.
Measurement of neutrophil adhesion
The ability of neutrophils to form firm adherence with a substrate was determined using crystal violet staining. Briefly, isolated neutrophils (5x105) in 180 µl buffer were applied to ECM-coated or uncoated wells and allowed to adhere for 1 h at 37°C. Treatments of 10-8 M IL-8 or 10-7 M PAF, 100 ng/ml PMA or 100 ng/ml TNF-, were applied. Other wells were treated with combinations of IL-8 or PAF and TNF- at the above concentrations. Plates were then incubated at 37°C for 30 min. The percentage of adhered cells was then determined exactly as described [38
]. Triplicate wells of untreated cells were fixed with formaldehyde before the first washing to estimate maximum adherence.
Measurement of neutrophil respiratory burst activity
The oxidative burst of adherent neutrophils was measured using luminol-enhanced chemiluminescence [39
]. ECM-coated 96-well plate strips were rinsed with DPBS three times, and 5 x 105 neutrophils (in 50 µl RPMI) were added to each well. After incubation at 37°C for 1 h to allow the neutrophils to adhere, each well was supplemented with 10 µl luminol (150 µM final concentration). Two rounds of treatments were then applied to each well: first, a "priming" treatment consisting of IL-8 (1x10-8 M final concentration), PAF (1x10-7 M final concentration), or buffer control; second, an "activation" treatment consisting of PMA (100 ng/mL final), TNF- (100 ng/ml final), or buffer control. All treatments were added as 20 µl from a 10x stock solution. The PMA was added last, and data collection began immediately after addition of the activating agent using the Fluoroskan Ascent FL microtiter plate reader at 37°C with 30-s data intervals for 60 min. Results are expressed as total chemiluminescence measured over time (i.e., the area under the curve for a given time period).
Statistical analysis
One-way analysis of variance, followed by Tukey’s post-hoc testing, was used to test for statistical significance among treatment groups using Graph Pad Prism.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (25K): [in a new window] |
Figure 1. Suppression of plastic-induced activation of bovine neutrophils. Isolated neutrophils were applied to polystyrene wells containing RPMI and luminol at 37°C, and chemiluminescence was monitored for up to 60 min. The upper panel shows a representative tracing of chemiluminescence in relative light units (RLU) produced by cells on uncoated polystyrene. In the lower panel, wells were coated with the indicated amounts of ECM proteins prior to application of cells. Total ROS production (integrated chemiluminescence) of neutrophils is expressed as the percentage of that generated by neutrophils on uncoated polystyrene. Collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP). Results are pooled from five separate experiments; n = 6 for each bar (mean±SEM). All values shown are significantly different from those measured in uncoated wells (P 0.01).
|
 View larger version (83K): [in a new window] |
Figure 2. Intracellular calcium changes in ECM-adherent neutrophils. Isolated neutrophils were loaded with Fluo-3, adhered to the indicated ECM protein or plastic for 45–60 min at 37°C, and stimulated with 10-7 M PAF (upper panel) or 10-8 M IL-8 (lower panel). Fluorescence was measured at 0.5-s intervals for 75 s, and the maximum fluorescence was determined as the peak fluorescence minus the prestimulus baseline fluorescence. Substrates are plastic (PLA), collagen type IV (COL), fibronectin (FIB), laminin (LAM), thrombospondin (THR), or heparan sulfate proteoglycan (HSP). The data are expressed as mean ± SEM of four pooled experiments with five replicates on each ECM within each experiment. *, A statistically significant difference compared with cells adhered to plastic (P0.05).
|
 View larger version (44K): [in a new window] |
Figure 3. Actin polymerization in adherent neutrophils. Neutrophils were incubated on ECM-coated chamber slides for 60 min at 37°C and then stimulated with IL-8 (10-8 M) or PAF (10-7 M). At the indicated time intervals, the neutrophils were fixed and permeabilized with lysophosphatidylcholine, and intracellular actin was stained with BODIPY-phallacidin. The 0 time point was obtained with immediate fixation after adding a buffer control. After washing, fluorescent images were recorded with 100x original magnification. Representative results on plastic and three ECM proteins are shown. Neutrophils adherent to thrombospondin or collagen are not shown, however their F-actin polymerization was similar to that seen in fibronectin-adherent cells. Fibronectin (FIB); heparan sulfate proteoglycan (HSP); laminin (LAM). The data are representative of 3 (ranging from 3 to 6) experiments.
|
 View larger version (30K): [in a new window] |
Figure 4. Analysis of actin polymerization in adherent neutrophils. Images of adherent neutrophils stained for intracellular actin were analyzed as described in Materials and Methods. Values represent a ratio of the increase in pixel intensity of the outer 10% of the neutrophil as compared with the inner 90% of the same cell. Cells were adherent to the indicated ECM protein and were stained at the indicated times after 10-7 M PAF stimulation (upper panel) or 10-8 M IL-8 stimulation (lower panel; see Fig. 3
). Plastic (PLA); collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP). The data are pooled from three separate experiments, with 15 data points per ECM protein per experiment expressed as the mean ± SEM.
|
 View larger version (41K): [in a new window] |
Figure 5. Release of lactoferrin by adherent neutrophils. Neutrophils were allowed to adhere to ECM-coated wells for 1 h at 37°C and were then stimulated with buffer (control, open bars), 10-7 M PAF (hatched bars), and 10-8 M IL-8 (solid bars). After 30 min at 37°C, the plates were centrifuged, and lactoferrin was measured in the supernatants by ELISA. Results are expressed as the % of lactoferrin released by PAF- or IL-8-stimulated adherent bovine neutrophils compared with unstimulated cells in the same experiment. Plastic (PLA); collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP). Results are pooled from five independent experiments (mean±SEM); n = 3 within each experiment. *, Statistically significant differences from unstimulated cells; P 0.05; **, P 0.01.
|
. For comparison purposes, neutrophils were treated with TNF-, which has been shown to increase neutrophil adhesion to some substrates [16
, 20
]. Prior to any treatment, 60–70% of isolated neutrophils adhered to plastic, collagen, fibronectin, and thrombospondin, and treatment with PAF or IL-8 induced little additional adhesion to these matrices (Fig. 6
). In contrast, only small numbers of untreated neutrophils (10–15%) adhered to laminin and HSP, and this was not enhanced much by PAF or IL-8 treatment (Fig. 6)
. Although TNF- treatment of neutrophils did not enhance adhesion to fibronectin, collagen, or thrombospondin, it did induce a dramatic increase in neutrophil adhesion to laminin and HSP (Fig. 6)
. When PAF or IL-8 was added in combination with TNF-, adhesion was not affected on plastic, collagen, fibronectin, or thrombospondin, compared with TNF- alone. It is interesting that a combination of PAF or IL-8 with TNF- caused decreased neutrophil adhesion to HSP as compared with TNF- treatment alone. PAF also decreased the neutrophil adhesion seen with TNF--treated cells on laminin (Fig. 6)
.
 View larger version (98K): [in a new window] |
Figure 6. Adhesion of stimulated neutrophils to ECM proteins. Neutrophils were incubated in ECM-coated wells for 60 min at 37°C and were then treated with 10-7 M PAF, 10-8 M IL-8, 100 ng/ml TNF-, or a combination of PAF or IL-8 with TNF-, as indicated. After 30 min at 37°C, the wells were washed and fixed, and adherent cells were stained with crystal violet. The solubilized, stained cells were quantified using absorbance at 550 nm and were compared with wells where cells were fixed prior to washing (100% adherence). Plastic (PLA); collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP); control (CON). Results are expressed as the percent of cells adhered to each substrate (mean±SEM). The data are representative of three independent experiments; n = 3 for each ECM. *, P 0.05; **, P 0.01, indicating a statistically significant difference compared with untreated cells on the same ECM.
|
 View larger version (35K): [in a new window] |
Figure 7. ROS production in adherent neutrophils. Isolated neutrophils were allowed to adhere to ECM-coated wells for 60 min at 37°C and were then treated with buffer (control, open bars), 100 ng/ml PMA (hatched bars), and 100 ng/ml TNF- (solid bars). Luminol-enhanced chemiluminescence was measured for 60 min, and values shown represent the integrated total chemiluminescence of the 60-min measurement period. Relative light units (RLU); plastic (PLA); collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP). The data are means ± SEM of pooled experiments ranging from 6 to 20 experiments with each ECM; n = 3 replicates for each ECM in each experiment. *, P 0.05; **, P 0.01, indicating a statistically significant difference compared with cells adhered to plastic and subjected to the same stimulation.
|
[8
, 20
]. We confirmed this response in bovine neutrophils and show further that ECM proteins had a differential ability to support TNF--induced ROS production (Fig. 7)
. It is interesting that the pattern of ROS production induced by TNF- was very different than that seen when PMA is used. For example laminin supported one of the strongest TNF--induced responses but one of the lowest PMA-induced responses. Additionally, cells adherent to plastic or collagen exhibited a minimal TNF-induced ROS production, whereas these substrates supported maximal responses when activated by PMA.
Despite their ability to act as priming agents for neutrophils in suspension, PAF and IL-8 actually inhibited TNF--induced ROS production in neutrophils adherent to collagen, fibronectin, and laminin (Fig. 8
). In sharp contrast to the results observed with these ECM proteins, PAF and to a lesser extent IL-8 increase the TNF--induced ROS production in neutrophils adherent to HSP (Fig. 8)
. Thus, it is apparent that among all of the functional responses we analyzed, ROS production was most susceptible to differential modulation by ECM proteins and inflammatory agents.
 View larger version (35K): [in a new window] |
Figure 8. Effects of costimulation with IL-8 or PAF on the TNF--induced respiratory burst in adherent neutrophils. Neutrophils were allowed to adhere to ECM-coated wells for 60 min at 37°C and were then stimulated with 100 ng/ml TNF- or TNF- combined with 10-7 M PAF (upper panel) or 10-8 M IL-8 (lower panel). Total chemiluminescence over a 60-min measurement period was determined as in Figure 3
, and cotreatment (TNF- with PAF or IL-8) values were converted to the percent of the average chemiluminescence of TNF- alone-treated cells on the same ECM in the same experiment. Plastic (PLA); collagen type IV (COL); fibronectin (FIB); laminin (LAM); thrombospondin (THR); heparan sulfate proteoglycan (HSP). Results are expressed as mean ± SEM and are pooled from 2 to 11 experiments on each ECM; n = 3 on each ECM for each experiment. *, P 0.05; **, P 0.01, indicating a statistically significant difference compared with cells adhered to plastic.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Moreover, we demonstrate that different ECM proteins induce differential responses in these cells, depending on the type of ECM protein and inflammatory agents present. Clearly, ECMs have complex effects on neutrophil responses to IL-8 and PAF, and among the different ECMs, there are similarities and differences in their effects. Although we see, for example, that some responses of neutrophils adherent to collagen IV are lower than cells adhered to other ECMs, it would be a gross oversimplification to believe that one ECM enhances inflammatory responses to agents such as IL-8 and PAF and that another dampens those responses. Rather, the neutrophil appears to integrate the ECM interaction as one of multiple messages (including IL-8 and PAF) to modulate certain functional responses selectively. The neutrophil response that is the most upstream among those we measured, i.e., changes in intracellular Ca2+ concentration, showed the least variability. This would suggest that under most conditions, the initial neutrophil responses to IL-8 and PAF are not regulated differentially. Other neutrophil responses that are more downstream, such as degranulation, actin polymerization, adhesion, and ROS production, demonstrated an increasingly more diverse set of responses, depending on the combination of ECM and stimulant.
The neutrophil function that exhibited the most differential response to ECM and IL-8 or PAF interactions was clearly the production of ROS. With neutrophils in suspension, IL-8, PAF, and even TNF- are normally considered priming agents, that is, they potentiate ROS production to a second stimulus but are not themselves able to activate ROS production [50
, 52
]. However, when neutrophils are adherent to certain ECM proteins, TNF- becomes a potent, direct stimulant of ROS production [8
, 20
]. This TNF--induced ROS production occurs after a 15- to 30-min lag period [8
, 20
], is dependent on an intact actin cytoskeleton [8
], and is also believed to be dependent on interactions of the CD11a/CD18b integrin with the substrate [55
, 56
]. Although very little is known about how priming agents such as IL-8 and PAF affect the TNF- oxidative burst, there is one study in which IL-8 actually decreases TNF--induced ROS production of fibronectin-adherent human neutrophils [57
]. Our study with bovine neutrophils adhered to several ECMs demonstrates that TNF--induced ROS production is highly variable, depending on the type of ECM used, the presence of a priming agent such as IL-8 or PAF, and the interaction of both factors. This is most evident in the case where PAF costimulation can have totally opposite effects, depending on the type of ECM used: inhibitory when the ECM is fibronectin or laminin and enhancing when the ECM is HSP. Furthermore, this effect seems to be separate from other functional responses that we measured. This is most notable with regard to adhesion. Although the TNF--induced oxidative burst is considered to be adhesion-dependent, there is no correlation between the effects we observed on firm adhesion and on ROS production. For example, PAF decreased TNF--stimulated adhesion to laminin and HSP. However, ROS production in response to TNF- decreased on laminin but increased on HSP when PAF was present.
Previous studies showing that human neutrophils adherent to ECM proteins responded to TNF- with sustained ROS production led to speculation that neutrophil-ROS production can be partitioned physiologically (i.e., they would respond after they had left the circulation and encountered high concentrations of ECM) [14
]. Indeed, the interstitial space is often considered to be the battleground where neutrophils and other host-defense cells encounter microorganisms that have breeched the initial barrier of skin or epithelium [58
]. Although we confirmed that bovine neutrophils also exhibit high ROS production in the permissive company of ECM proteins, the heterogeneity with which different ECM proteins supported this response raises additional questions. One possibility to consider is whether differential responses to various ECM proteins reflect physiological partitioning on an even smaller scale. Distribution of ECM proteins in tissues is extremely complex, with heterogeneous topographic distribution of different proteins [59
60
61
62
63
]. For example, the basement membrane of some endothelia is constructed predominately of collagen type IV [60
]. Therefore, the relatively low TNF--induced ROS production observed when neutrophils were adherent to collagen type IV may represent a protective mechanism to preserve endothelial integrity. Conversely, the relatively high TNF--induced ROS production observed when neutrophils are adherent to fibronectin may be related to a different interstitial distribution of that ECM protein. Furthermore, local abundance of some ECM proteins can change during disease or inflammation [64
65
66
], so neutrophil ROS production could be altered by these changes. These interpretations are complicated, however, by our observations of the effects of PAF or IL-8 on TNF--induced ROS production. Although these agents are important for the recruitment of neutrophils to inflammatory sites, they also dampen ROS production when fibronectin and laminin, among others, are present. Thus, our results suggest the possibility that HSP interactions with neutrophils may have an overriding role in permitting ECM-dependent, TNF--induced ROS production to proceed in the presence of IL-8 and/or PAF. However, the widespread distribution of HSP moieties, in the extracellular matrix and on the surface of other cells, complicates any idea that different ECM-mediated responses represent a form of microscale physiological partitioning of neutrophil responses. Indeed, although our study shows strong, complex effects that ECM interactions can have on neutrophil functional responses, extrapolation of these results to in vivo situations should still be considered speculative. In vivo, neutrophils probably encounter multiple ECM moieties in a short time period, as well as multiple, soluble pro- and anti-inflammatory agents. In this situation, many different types of signal integration could occur. Thus, further studies with more complex representations of the ECM are clearly needed to determine what these interactions may be in vivo.
Conversely, the simplistic ECM interaction models used here would appear to be very useful in studying the mechanisms of interaction of signal-transduction systems in neutrophils and perhaps in designing pharmacological agents that could modulate neutrophil responses differentially. For example, TNF--induced ROS production is initially dependent on the integration of two separate signals: from TNF-, which probably involves multiple, nonreceptor tyrosine kinases, and from a cell-surface ligand, which interacts with the appropriate ECM. The simplest mechanistic explanation of differential responses to ECMs would involve each ECM interacting with the neutrophil via unique cell-surface ligands. Because of their known interactions with extracellular ECM proteins and the intracellular cytoskeleton and their ability to modulate several signal-transduction pathways, neutrophil integrins are the most attractive candidates as the necessary link [67
, 68
]. The question that comes to mind is whether there is sufficient specificity of integrin-ECM interaction to account for the differential response we see. The dominant neutrophil integrin is the ß2 integrin Mß2 (CD11b/CD18, CR3, Mac-1), which has been implicated in adhesion to fibronectin, collagen, and to a lesser extent, laminin [15
, 69
, 70
]. Although ß2 integrin-adhesive interactions are believed to be necessary for some functional responses, including the TNF--induced oxidative burst of adherent neutrophils [18
], they do not show discriminate binding to any one ECM. Additionally, some adhesion-dependent neutrophil responses have been shown to be CD11b/CD18-independent [17
]. Unlike ß2 integrins, ß1 integrins are believed to mediate only cell-ECM interactions. However, there is considerable controversy regarding which types of ß1 integrins are expressed on the surface of neutrophils. Although evidence has been proposed for 5ß1 [71
], 2ß1 [72
], 6ß1 [73
], and 9ß1 [71
] integrins on human neutrophils, there is no consensus among these studies. Whether ß1 integrins discriminate between different ECM ligands is unclear as well. For example, laminin is often cited as an ECM with a specific integrin-binding relationship with 6ß1 [74
]. However, laminin can also bind 3ß1, vß3, and 2ß1 integrins, which are also shown to be on neutrophils [75
].
Furthermore, other nonintegrin cell-surface molecules can bind laminin [76
, 77
]. The question of what cell-surface receptor(s) mediates neutrophil/HSP interactions is even less clear. Although HSP can interact with the Mß2 integrin, it may also interact with a multitude of other molecules, including L-selectin, coagulation factors, enzymes, other ECM proteins, growth factors, and cytokines [78
]. Understanding these interactions is complicated by the fact that HSP-like molecules are contained within ECM domains and expressed on the surface of many cells [79
]. Research into HSP effects on cells has focused on their ability to potentiate growth factor actions [80
] and modulate other signaling molecule actions on cells [81
], and little is known about any direct effects HSP may have on leukocytes. That our observations about HSP effects contradict a published study on HSP actions on IL-8-induced calcium-concentration changes [31
] and a study on HSP-induced inhibition of the neutrophil-oxidative burst [32
] suggests that this may be a very complex ECM/cell interaction. Therefore, although it might be appealing to look for straightforward, specific ECM/ligand interactions to explain these differential responses—the fact that each ECM can bind many integrin and nonintegrin molecules [82
] and that some integrins can bind to more than one ECM [11
]—suggest that this view may be too simplistic. Although the outcome (ROS production) in our experiments is likely the result of the integration of TNF- and integrin signaling, inhibition and enhancement of ROS production can occur as a result of a third, G-protein-coupled receptor (GPCR) message, such as PAF.
Moreover, the effect of this third signal is in turn dependent on the type of the integrin (or other ECM binding protein)-mediated signal. Indeed, research in various cell types has demonstrated recently how GPCR-mediated messages integrate with other signals, such as mitogen-activated protein kinase cascades [83 , 84 ], Ras-dependent signaling [85 ], receptor tyrosine kinases [86 ], and the focal adhesion-related kinase Pyk2 [86 , 87 ]. The latter studies showing the interactions of GPCR with focal adhesions also implicate integrins in these signaling interactions. Clearly, one or more integrative events are occurring in the relevant signal-transduction pathways of neutrophils in our experimental system. It would be worthwhile to determine, for example, at which step a GPCR message acts to inhibit neutrophil-ROS production, especially if other neutrophil responses are not hindered by this treatment. At this time, we know very little about where in the signal-transduction cascades these messages are integrating, although it is probably downstream or on a divergent pathway from changes in intracellular calcium concentrations. If the actual steps where these messages are integrated could be elucidated, they could be attractive targets for new therapeutic agents that could potentially differentially regulate deleterious neutrophil responses (such as excessive ROS production) without inhibiting other important host-defense functions.
Received July 19, 2001; revised December 27, 2001; accepted January 2, 2002.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and fMLP in human neutrophils adherent to fibrinogen FEBS Lett. 451,33-38[Medline] potentiates neutrophil degranulation in response to host defense cytokines neutrophil-activating peptide 2 and IL-8 by modulating intracellular cyclic AMP levels J. Immunol. 149,1356-1364[Abstract]9ß1 Integrin is expressed on human neutrophils and contributes to neutrophil migration through human lung and synovial fibroblast barriers J. Leukoc. Biol. 66,809-816[Abstract]6ß1 heterodimer J. Leukoc. Biol. 53,576-582[Abstract]This article has been cited by other articles:
 |
|
 |
|
  J. A. Wofford and J. R. Wright Surfactant protein A regulates IgG-mediated phagocytosis in inflammatory neutrophils Am J Physiol Lung Cell Mol Physiol, December 1, 2007; 293(6): L1437 - L1443. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Illario, A. L. Cavallo, K. U. Bayer, T. Di Matola, G. Fenzi, G. Rossi, and M. Vitale Calcium/Calmodulin-dependent Protein Kinase II Binds to Raf-1 and Modulates Integrin-stimulated ERK Activation J. Biol. Chem., November 14, 2003; 278(46): 45101 - 45108. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Casillan, N. C. Gonzalez, J. S. Johnson, D. R. S. Steiner, and J. G. Wood Mesenteric microvascular inflammatory responses to systemic hypoxia are mediated by PAF and LTB4 J Appl Physiol, June 1, 2003; 94(6): 2313 - 2322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. Cho, Z. Yao, S.-W. Wang, R. F. Alban, R. G. Barbers, S. W. French, and C. K. Oh Regulation of Activin A Expression in Mast Cells and Asthma: Its Effect on the Proliferation of Human Airway Smooth Muscle Cells J. Immunol., April 15, 2003; 170(8): 4045 - 4052. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. GOTTE Syndecans in inflammation FASEB J, April 1, 2003; 17(6): 575 - 591. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
